0 The Origin of Bioelectrochemistry: An Overview

1

0The Origin ofBioelectrochemistry: AnOverview

Pierre BiancoBIP – CNRS, Marseille, France

0.1 Electrokinetic Phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.2 Membrane Phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60.3 Electron Transfer Reactions in Biological Compounds . . . . . . . . . . 60.4 Transmission of Information in Living Organisms . . . . . . . . . . . . 8

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3

Electrical phenomena known in antiquitywere lightning, the attraction of lightbodies, and the discharges delivered bythe torpedo fish. No relation amongthese seems to have been suspected atthe time. Maybe the most intriguingnatural phenomenon further related tobioelectrochemistry was the productionof electricity by a living organism suchas a fish (not only the torpedo fish butalso the gymnotus, or eel of Surinam,and the electric catfish living in theNile waters). Nevertheless, no attemptsto explain electrical phenomena wereundertaken during the long period thatextends from antiquity to the beginning ofthe modern era, that is, the seventeenthcentury. It is accepted that the science ofelectricity began in 1600 with the treatiseDe Magnete by William Gilbert. The first‘‘electric machine’’ producing electricityon demand was devised by Otto vonGuericke in 1672; the first ‘‘condensers’’known as Leyden jars were constructedby Ewald Georg von Kleist and Pietervan Musschenbroeck in the middle of theeighteenth century. At that time, electricitywas considered a ‘‘fluid’’ that flowed withinliving organisms like water in pipes.

The earliest observations on the chem-ical effects produced by electricity werereported by Beccaria (ca. 1750), who ob-served gas evolution on passing electric

sparks through water placed in a tube. Thefirst experiments on the effect of electriccurrents on living organisms were per-haps described by Johann Georg Sulzer atthe same time. Sulzer found that when aplate of lead and another of silver laidon the tongue touched one another, a‘‘vitriolic taste’’ was perceived by the ex-perimenter. Unfortunately, Sulzer did notconnect the observed phenomenon withelectricity. The possibility of using Leydenjars opened new ways for investigating theeffects of electricity on living organisms.In 1752, Leopoldo Caldani, an Italian phys-iologist from Bologna, concluded from hisexperiments on the crural nerves of variousanimals that ‘‘electric matter is the mosteffective of stimulating agents’’ in livingorganisms. From the rapidity of the mus-cular responses to repetitive stimuli viathe nervous system, it became generallyaccepted that a ‘‘nerval fluid’’ or ‘‘ani-mal spirit’’ must exist in living organisms,rapidly identified as ‘‘animal electricity’’.In October 1786, Luigi Galvani, then pro-fessor of anatomy in Bologna, gave anaccount about his famous experiments onthe muscular contractions that convulsedfrog’s legs when touched with a metallicarc, or better a composite (iron + copperor silver) arc (Fig. 1). Galvani supposedthat the observed phenomena was due tothe ‘‘animal electricity’’, the interior and

4 0 The Origin of Bioelectrochemistry: An Overview

Fig. 1 The first experiments on ‘‘animal electricity’’ by Luigi Galvani described in De ViribusElectricitatis in Motu Musculari Commentarius (published in 1792).

exterior muscles of the frog’s legs forminga kind of small Leyden jar, and the nervebehaving as a conductor for the jar.

Volta, then professor of physics at theUniversity of Pavia, first shared Galvani’sviews. During 1790 to 1792, he carriedout a series of careful experiments dealingwith the effect of the electric dischargeson the convulsion of frog’s legs, varyingthe nature and the combination of metalsby using rather sophisticated heteroge-neous metallic arcs. In 1794, he took adifferent stance, concluding that the ‘‘gal-vanism [. . .] is nothing but an artificialelectricity set in motion by the contactof heterogeneous conductors’’. In conclu-sion, he claimed that there was only onetype of electricity. The construction of the‘‘pile’’ (1799) derives, in fact, from Volta’sexperiments on contact between metals.

This discovery can be regarded as the realfoundation of the science of electrochem-istry and, indirectly, of bioelectrochem-istry, considered as an ‘‘affiliate’’ of elec-trochemistry. In reality, the controversiesthat have arisen from the different inter-pretations given by Galvani and Volta – theexistence of an animal electricity and theassignment of muscular contraction to thecontact potential difference between un-like metals, partly right and partly wrong,respectively – highlighted an undeniablefeature, that is, that both assumptionswere incomplete. Of course, there existsone type of electricity resulting from themovement of electrons but Galvani andVolta are pardonable because they wereunaware of the electron concept. Never-theless, Volta wrongly supposed that thecontact of the two metals was also the

0.1 Electrokinetic Phenomena 5

source of current, the production of whichwould be, as he himself said, a ‘‘perpetualmotion’’. A relatively satisfactory explana-tion was suggested by Johann WilhelmRitter (ca. 1800) who claimed that the con-tact potential difference between unlikemetals could not be a source of energy:the source of energy required to contractthe muscles of the frog’s legs should orig-inate from chemical reactions providingthe necessary energy.

Volta’s breakthrough stimulated a highlevel of research activity on the effect ofthe electric current. Davy spoke of thevoltaic battery as ‘‘an alarm-bell to ex-perimenters in every part of Europe’’.The development of electrochemistry pro-gressed throughout the nineteenth andtwentieth centuries, thus allowing severalfundamental as well as applied aspects tobe investigated, leading to highly valuablediscoveries and theories. Three main av-enues were opened, including studies onnew power supplies (development of cellsand batteries), the problem of the trans-mission of charge through solutions, andthe transfer of charge at metal–solutioninterfaces (Faraday’s laws). During thistime, bioelectrochemistry developed moreslowly, without striking discoveries com-pared to those that have marked theevolution of ‘‘mother’’ electrochemistry.It has benefited, however, from increas-ing knowledge in different areas cover-ing biochemistry, membrane phenomena,electrophysiology, medical diagnosis, andso on. All these investigations have con-tributed to revealing of the electrochemicalnature of a large variety of biological phe-nomena. Though it is not possible toencompass the totality of the numerousadvances resulting from research in thefield of bioelectrochemistry, we proposeto examine the most important branchesand experimental implementations gained

from the encounter between ‘‘pure’’ elec-trochemistry and biological and physiolog-ical phenomena. Progress in the under-standing of biological and physiologicalprocesses using electrochemistry can beexamined on the basis of two distinct setsof phenomena related to (1) the existenceof potential differences at phase bound-aries and (2) the electrical polarization andthe electron-exchange process.

0.1Electrokinetic Phenomena

At the end of the nineteenth century,research was undertaken on electrocap-illary and electrokinetic phenomena thatarise from matter in motion. Three re-lated effects were distinguished, namely,electroosmosis, streaming potentials, andelectrophoresis.

The existence of electroosmosis wasconfirmed experimentally shortly after thediscovery of the decomposition of waterby means of electric current. It is one ofthe earliest-known electrochemical effects.Studying the effect of electroosmosis hasproven to be useful for the understandingof the metabolism of cells. For example,it was established by Blinks in themiddle of the last century [1] that trueelectroosmosis is negligible compared toosmotic pumping in some large plant cellsof fresh water and marine algae.

Streaming potentials are produced by aflow of liquid forced through a capillarysystem. In 1943, Miller and Dent [2] pre-sented experimental evidence that stream-ing potentials are the cause of the waveportions of an electrocardiogram. Later,several experimenters attempted to corre-late electrocardiogram profiles with differ-ent physiological events, such as the effectof the pulsatile flow of saline electrolytes


that can generate electrokinetic potentialsremarkably similar to in vivo electrocardio-grams (e.g. Findl and Kurtz, [3]).

Electrophoresis is concerned with themigration of particles under the influenceof an external field. First, electrophoreticvelocities were measured by the sameexperimental methods as those used byHittorf (around 1853) for transport num-bers. The method was remarkably de-veloped and refined by Tiselius (1937,Nobel Prize in 1947). Since then, therehave been a large number of applica-tions in biology and medicine, in par-ticular in the analysis of sera and theseparation of amino acids and proteins.Recently, the technique was improvedby coupling capillary electrophoresis toan electrochemical detector (CEEC) thatcan be implanted in vivo for studyingmicrodialysis [4].

0.2Membrane Phenomena

Membrane phenomena cover an extremelybroad field. Membranes are organizedstructures especially designed to performseveral specific functions. They act as abarrier in living organisms to separate tworegions, and they must be able to controlthe transport of matter. Moreover, alter-ation in transmembrane potentials canhave a profound effect on key physiologicalprocesses such as muscle contraction andneuronal activity. In 1875, Gibbs statedthe thermodynamic relations that formthe basis of membrane equilibria. Thetheory of ionic membrane equilibriumwas developed later by Donnan (1911).From theoretical considerations, Donnanobtained an expression for the electric po-tential difference, commonly known as themembrane potential between two phases.

Results from Scatchard in 1953 [5] haveindicated that ion-exchange membranesmay be useful as electrodes at whichthere is no reduction or oxidation andno restriction to special classes of ionsexcept size. The first ion-selective elec-trodes were constructed in 1936 and thenlargely developed in the 1960s, whenEisenman [6] established the relationshipgiving the electrical potential differencebetween two aqueous solutions separatedby an ion-exchange membrane. Of particu-lar interest for measurements in biologicalmedia was the construction of ion-selectiveelectrodes capable of detecting calciumions. In human serum and other bio-logical fluids, calcium is partly bound tosubstances such as proteins. The detec-tion of free calcium became possible byusing a calcium-selective electrode. After-ward, other membrane electrodes wereconstructed (e.g. the urea electrode, byGuilbault and Montalvo, [7]).

Electric fields have been shown to influ-ence the conformation of various naturaland synthetic polynucleotides in solution.In 1958, Hill [8] calculated that high elec-tric fields could bring about separation ofthe two molecule chains of nucleotides inDNA. It has even been suggested that theelectric fields and their variations at bio-logical interfaces might act as the triggerfor division of genetic material in the cellprior to self-duplication. On the basis of theanalogy between a cell surface–biologicalfluid interface and an electrode–solutioninterface, investigations were carried outaround the 1970s using differential capac-itance measurements (ac polarographicmethod) in conjunction with ellipsome-try measurements. The dependence onapplied potential of the adsorption of sev-eral biomolecules on mercury electrodeswas investigated, thus giving insights intothe different orientations and bindings

0.3 Electron Transfer Reactions in Biological Compounds 7

of biological molecules at solid–solutioninterfaces (1977; e.g. [9]).

0.3Electron Transfer Reactions in BiologicalCompounds

Curiously, it is also from the study of cap-illary phenomena (developed by Lippmanin 1873) that a new field of research wasopened up toward the study of electrontransfer reactions through the discoveryof polarography by Heyrovsky in 1927(Nobel Prize in 1959). This popular tech-nique provided an extraordinary amountof data. To the question ‘‘What can electro-chemical studies tell one about biologicalelectron transfer and related processes?’’,Dryhurst [10] replies by giving a definiteset of similarities between electrochemicaland biological reactions, comparing the in-terface electrode–solution to the interfaceenzyme–solution working in very simi-lar conditions of pH, ionic strength, andtemperature.

A large series of compounds of biologi-cal interest have been investigated throughthe second and third quarter of the lastcentury using polarographic methods, forexample, purines, pyrimidines, vitaminB12 and related cobalamines, nucleic acids,pteridines, flavins and flavin nucleotides,porphyrins, cytochromes, and so on. As apractical example, the tests for cystine andproteins introduced by Brdicka (1933) [11],which are of importance in clinical anal-ysis, deserve to be mentioned. Brdickaobserved that catalytic activity existed notonly for cystine but also for proteins con-taining −SH and −S−S− groups togetherwith amino groups. These catalytic wavescould be detected in solutions of cobalt(II)but were virtually absent in solutions ofcobalt(III). The great merit of the catalytic

double wave detected in ammoniacal solu-tions containing cobalt(II) and cobalt(III)was based on the fact that the number ofactive groups that govern the height of thecatalytic double wave differs from normalto pathological sera. Such a test has beenshown to be efficient for detecting hepati-tis cases and possible cancers in more than90% of proved cases.

During the second half of the twentiethcentury, electroanalysis has enjoyed a re-naissance because of the development ofseveral new technologies (e.g. single andcyclic voltammetry, pulse voltammetry,etc.) resulting from an enormous ex-ploratory effort in the theory and method-ology of electrochemical techniques servedby progress in electronics. Bioelectro-chemistry has benefited from these newtechnologies that allowed other (e.g. solid)macro- and microelectrodes to be used. Mi-croelectrodes were pioneered in the 1940sto measure oxygen concentrations insidebiological tissues [12]. In 1969, Adams, apioneer in the field of solid electrodeswrote [13]: ‘‘There is every reason to believethat the current interest in solid electrodevoltammetry will continue and perhaps in-crease in the near future.’’ It was more thana vision for bioelectrochemistry wheresolid electrodes have been extensively usedfor several decades, in particular for di-rect observations in vivo. As an example,cyclic voltammetry at the platinum elec-trode has been shown to be suitable foranalyses in blood serum by Koryta [14]. Themeasurement of the time dependence ofascorbic acid concentration in the cortex ofan isolated kidney gave information aboutcirculation within the organ (1973). Adamssucceeded in probing the concentration ofneurotransmitters by directly implantingmicroelectrodes inside the living brain of arat [15]. Thereafter, microelectrodes wereminiaturized to ultramicroelectrodes to be


used to probe chemical reactions insideeven single biological cells [16].

Another major success for bioelectro-chemistry was achieved in the field ofredox protein electrochemistry. The firstwork using this approach was reportedin 1965 by Griggio and Pinamonti [17];concerning cytochrome c. In the 1980s,reports on different families of metal-loproteins were published concomitantlyby several groups [18–24], thus demon-strating that the direct electrochemistryof redox proteins could be observed pro-vided that well-defined experimental con-ditions were fulfilled. A new avenue wasthus opened toward the understandingof electron transfer processes throughoutelectron-carrier chains using the electro-chemical model.

0.4Transmission of Information in LivingOrganisms

This is, perhaps, the most complicatedand sophisticated approach that has beentempted as a challenge for bioelectrochem-istry: several reactions are involved in theoverall process controlling the transmis-sion of information in living organisms,including electron transfer, ion transportthrough membranes, and so on. Oneof the most important mechanisms con-sists of electrical signals transmitted afterelaboration to the organs, which have toperform the required action. Various kindsof electrical conduction are involved inthe overall process. Several approacheshave contributed to unravel such a com-plicated process, for example, the work ofKoryta [25], who studied a series of macro-cyclic ligands with alkali metals and theirproperties in regard to membrane trans-port and phosphorylation uncoupling.

An – even shorter – overview on the de-velopment of bioelectrochemistry throughits two centuries of existence illustrateswell the kinds of biological and physiolog-ical problems that are being studied usingelectrochemical concepts and techniques.Several avenues have been opened towardthe study of electron transfer processes inliving organisms, the analysis of biologi-cal fluids, the control of the composition ofthe intracellular medium (as established byNeher and Sakmann, Nobel Prize winnersin 1991), the electrochemical detection ofimmunological reactions, the constructionof biosensors (which constitute a signifi-cant portion of the total effort), and so on.From the examination of the number ofpapers that have been published in the lastten years, it can be concluded that bioelec-trochemistry is thriving with increasingvitality.

References

1. T. Shedlovsky, Electrochemistry in Biology andMedicine, John Wiley & Sons, New York,1955.

2. J. R. Miller, R. F. Dent, Lab. Clin. Med. 1943,28, 168.

3. E. Findl, R. J. Kurtz in Electrochemical Studiesof Biological Systems, ACS Symposium Series(Ed.: D. T. Sawyer), American ChemicalSociety, Washington, DC, 1977, p. 180.

4. R. Weinberger, Practical Capillary Elec-trophoresis, Academic Press, New York, 1993.

5. G. Scatchard, J. Am. Chem. Soc. 1953, 75,2883.

6. S. Ciani, G. Eisenman, G. Szabo, J. Membr.Biol. 1969, 1, 1.

7. G. G. Guilbault, J. Montalvo, J. Am. Chem.Soc. 1969, 91, 2164.

8. T. L. Hill, J. Am. Chem. Soc. 1958, 80, 2142.9. H. Kinoshita, S. D. Christian, M. H. Kim

et al. in Electrochemical Studies of Biologi-cal Systems, ACS Symposium Series (Ed.:D. T. Sawyer), American Chemical Society,Washington, DC, 1977, p. 118.

10. G. Dryhurst, Electrochemistry of BiologicalMolecules, Academic Press, New York, 1977.

0.4 Transmission of Information in Living Organisms 9

11. R. Brdicka, Collect. Czech. Chem. Commun.1933, 5, 148.

12. P. W. Davies, F. Brink, Rev. Sci. Instrum.1942, 13, 524.

13. R. N. Adams, Electrochemistry at Solid Elec-trodes, Marcel Dekker, New York, 1969.

14. J. Pradac, J. Koryta, Ber. Bunsen-Ges. Phys.Chem. 1973, 77, 808.

15. R. N. Adams, Prog. Neurobiol. 1990, 35, 297.16. J. B. Chien, R. A. Wallingford, A. G. Ewing,

J. Neurochem. 1990, 54, 633.17. L. Griggio, S. Pinamonti, Atti dell’Istituto

Veneto di Scienze, lettere ed Arti 1965, 124,15.

18. P. Yeh, T. Kuwana, Chem. Lett. 1977, 1145.

19. M. J. Eddowes, H. A. O. Hill, J. Chem. Soc.,Chem. Commun. 1977, 771.

20. K. Niki, T. Yagi, H. Inokuchi et al., J. Elec-trochem. Soc. 1977, 124, 1889.

21. P. Bianco, J. Haladjian, Biochim. Biophys.Acta 1979, 545, 86.

22. E. E. Bancroft, H. N. Blount, F. M. Hawk-ridge, Biochem. Biophys. Res. Commun. 1981,101, 1331.

23. I. Taniguchi, K. Toyosawa, H. Yamaguchiet al., J. Chem. Soc., Chem. Commun. 1982,1032.

24. F. A. Armstrong, H. A. O. Hill, N. J. Walton,FEBS Lett. 1982, 145, 241.

25. J. Koryta, Chem. Listy 1973, 67, 897.

11

1Voltammetry of Proteins

Fraser A. ArmstrongDepartment of Chemistry, Oxford University, Oxford, UK

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

1.2 Electrodes for Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

1.3 Voltammetric Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161.3.1 Protein Sample Contained in Solution . . . . . . . . . . . . . . . . . . . . . 161.3.2 Protein Sample Confined to Electrode . . . . . . . . . . . . . . . . . . . . . 181.3.3 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

1.4 Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191.4.1 Observing Active Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191.4.1.1 Iron-sulfur Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191.4.1.2 Flavins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211.4.1.3 Highly Oxidizing Intermediates: Fe(IV)=O . . . . . . . . . . . . . . . . . 22

1.5 Resolving Complex Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . 231.5.1 Coupled and Gated Electron Transfer . . . . . . . . . . . . . . . . . . . . . 231.5.2 Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

1.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

13

1.1Introduction

Direct electrochemical methods for study-ing redox-active centers in proteins arethose in which the exchange of electronswith an electrode is direct and does notinvolve electroactive mediators. To be use-ful, the interaction between the proteinand the electrode should fulfill the follow-ing criteria: (1) interfacial electron transfershould be fast (reversible), which requiressmall reorganization energy requirementsand good electronic coupling between theactive site and the electrode surface; (2)the protein should not become denatured,and characteristic properties such as cat-alytic activity should not be impaired.Voltammetric methods offer some impor-tant advantages over potentiometry, whichis the traditional method for measuringthe electrochemical properties of redox co-factors in proteins.

Compared with potentiometry, voltam-metric methods provide more rapid anddirect measurements of redox properties,and a wide range of electrode potentialscan be applied, often extending well be-yond the thermodynamic limits of water(<−0.4 V and >0.8 V vs the StandardHydrogen Electrode (SHE), at pH 7). Ac-tive sites having very high or very lowpotentials may be difficult to study by

potentiometric methods, either becausethey are themselves very reactive in theirhighly oxidized or reduced states and thusdecompose water, or the titrants or media-tors that are used to elicit their redox chem-istry also react with water. These problemscan be overcome by direct electrochemicalmethods [1–3]. As outlined below, exam-ples include the highly oxidizing catalyticintermediates of heme-containing perox-idases that exhibit reduction potentialshigher than 0.75 V, and reduced statesof certain Fe-S clusters that have reduc-tion potentials below −0.55 V. (The term‘‘reduction potential’’ defines the potentialfor a half-cell reaction written in the orderO + ne− = R, in which O and R are oxi-dized and reduced forms of the redox cou-ple, respectively. Although this potential iscommonly referenced against the SHE (asthe other half cell), the nonstandard condi-tions under which biological systems arestudied makes it appropriate to refer in-stead to a formal potential, denoted Eo′

andqualified for a particular set of conditions,where the temperature, pH, ionic strength,and specific ions present are specified.)The technical problems of measuring howreduction potentials vary over a wide tem-perature range, and the effects of high pres-sure can each be accommodated throughvoltammetric experiments [4]. Voltamme-try can be used even if the different redox

14 1 Voltammetry of Proteins

states of an active site lack a spectroscopichandle, such as a strong and characteristicelectronic transition or unpaired electrons;instead, the current peaks represent sig-nals that can be used to label species andmonitor the reactions of species that havebeen identified independently by spectro-scopic techniques [2].

Importantly, and unlike potentiometry,voltammetric methods are dynamic andgive information on kinetics, that is, ratesof electron transfer and coupled (EC)reactions: the latter include those in whichelectron transfer drives a reaction suchas ion/proton transfer, or is ‘‘gated’’, thatis, the case in which the electron-transferevent is controlled by a preceding chemicalprocess. Redox reactions can be quantifiedin both the potential and time domains,and these may be separated and resolved:for example, steady-state catalytic studiesof adsorbed enzymes reveal how catalyticelectron transport varies as a function ofpotential, which can be important if therate is sensitive to the oxidation state of aparticular site in the molecule [1].

Direct electrochemical methods canoften be integrated with spectroscopy.This is commonly carried out using anoptically transparent cell (such as anOTTLE), which views the light absorbedby species in solution close to a gold mini-grid or conducting glass electrode. To citean example, the intense Soret absorptionband (390–450 nm), which is a keycharacteristic of heme groups, provides agood marker for studying cytochromes [5].Improvements in sensitivity are makingpossible studies on proteins that are boundtightly (adsorbed) at the electrode – a goodexample being potential (AC)-modulatedelectroreflectance spectroscopy [6]. Morespecific information on the structuralchanges that accompany electron transfercan be obtained with techniques such

as surface-enhanced Raman spectroscopy(SERS) and surface-enhanced resonanceRaman spectroscopy, particularly for hemeproteins adsorbed on bare and modified Agelectrodes [7].

1.2Electrodes for Proteins

Success in protein voltammetry dependscritically upon the electrode and how itis prepared and modified. The current re-sponse may stem from protein moleculesfree in solution and undergoing a re-action upon diffusing to the electrodesurface, or it may stem from moleculesthat are already bound tightly (adsorbed) tothe electrode. Quasi-reversible diffusion-controlled electrochemistry has been doc-umented for a wide range of proteins,mostly the smaller variety (molecularmass <15 kDa) that function as mobileelectron carriers [3]. Diffusion-controlledelectrochemistry requires that the proteininteracts with the electrode in a tran-sient manner, that is, weakly, so that theelectrode does not become blocked. In-creasingly, however, attention has turnedto electrodes that bind protein moleculestightly, so that the sample is studied as astable monolayer that typically comprisesless than a picomole [1].

Since the pioneering studies of Hill [8]and Kuwana [9], the most successful elec-trodes for proteins have been noble metals(Au or Ag) modified with various adsor-bates, or materials such as carbon or metaloxides that have natural surface function-alities [10–18]. Conducting metal oxidesare often optically transparent, and thusprovide additional possibilities for spectralstudies, while a further development hasbeen the modification of electrode surfaceswith surfactant films [19–21]. Examples of

1.2 Electrodes for Proteins 15

(a) (b)

(c) (d)

Y

X

Y

X

Y

X

Y

X

Y

X

Y

X

Y

X

Y

X

Y

X

Y

X

Y

X

Y

X

O OO O O O O O

O− O− OO−

Fig. 1 Electrodes for protein voltammetry: (a) noble metals (Au, Ag) modifiedwith a SAM. Group X is typically sulfur, while Y is a functionality, such as −CH3,COO−, CH2OH, the variety and mixture of which can be designed to optimizethe interaction with the protein. Examples are described in Refs. [10–13]; (b) ametal oxide electrode. Examples are described in Refs. [9, 15–18]; (c) a carbonelectrode, typically pyrolytic graphite with the ‘‘edge’’ surface projected to thesolution. Protein adsorption is often optimized by inclusion of polycations.Examples are described in Refs. [1, 14]; (d) An electrode coated with a surfactantlayer within which the protein is confined. Examples are described inRefs. [19–21].

the electrodes that have been used suc-cessfully for protein studies are depictedin Fig. 1. It is important that the electrodebinds the protein in an orientation suit-able for fast electron transfer and that itdoes not cause the protein to denature;also it should not adsorb other moleculesthat block the surface. A problem withcommonly used metal electrodes, suchas Ag, Au, Pt, and Hg, is that they leadto denaturation and irreversible adsorp-tion of the resulting inactive protein, andthey are easily fouled by contaminants (thewater molecules that are normally bound

at the electrode/electrolyte interface areeasily displaced). This is demonstrated bythe observation that an unmodified ‘‘pris-tine’’ Ag surface, prepared by treatmentwith a hydrogen flame, gives a reversiblebut short-lived voltammetric response withcytochrome c [22].

Many of the adsorbates used to modifymetal electrodes produce self-assembledmonolayers (SAMs), in which the ter-minal functional groups provide goodbinding sites for proteins. For example,cytochrome c, which is positively chargedand contains an excess of lysine residues


in the region surrounding the heme edge,forms a stable electroactive layer at an Auelectrode modified with a monolayer of n-alkanethiols that terminate in −COO− ormixtures of COO− and −OH [11]. This in-teraction is weakened if the ionic strengthis raised. Likewise, the ‘‘blue’’ Cu proteinazurin, in which the surface over the activesite consists of a patch of hydrophobicamino acids, forms a stable electroac-tive layer at alkane thiols terminating in−CH3 groups, which is stable at high ionicstrength [12, 13]. One pertinent questionthat can be addressed is how electrochem-ical ET rates depend upon the nature ofthese terminal groups or the length of the(−CH2−)n spacer. For cytochrome c, anexponential variation of rate constant (k0)with distance is observed for medium-to-long chain SAMs, as expected from theory;however, reports suggest that as the chainlength is shortened, the electron-transferrate becomes limited by other factors. Itis not yet established why this shouldoccur, but one explanation proposed byNiki and coworkers is that the protein israrely in the right orientation for fast elec-tron transfer and must first reorganizeto achieve this [23]. Another interestingobservation is that SAMs composed of mix-tures of spacers or end functional groupsyield protein layers that exhibit faster ETrates [11].

Metal oxide electrodes have been usedwith or without modification [9, 15–18].Tin oxide and indium oxide are semi-conductors, while ruthenium dioxide isa metallic conductor. All of these yielddirect electrochemistry of cytochrome c,while modification with amine groups pro-duces a positively charged surface that canbe used for negatively charged proteins,such as the small Fe-S proteins known asferredoxins [18].

Carbon (pyrolytic graphite or glassy car-bon) is a very convenient electrode materialfor a wide range of proteins. The simpleact of polishing with abrasives, such asalumina and diamond paste, generates hy-drophilic surface oxide groups so that thereis some similarity with conducting metaloxide electrodes [14]. Oxide formation isenhanced if the electrode is oriented sothat the ‘‘edge’’ plane contacts the surface(pyrolytic graphite ‘‘edge’’; PGE). The sur-face of PGE is rough and can be modifiedfurther, either by covalent attachment offunctionalities or by adsorption of agents(coadsorbates) that assist the binding ofprotein molecules. For proteins with neg-atively charged surfaces, polycations areoften effective: these range from sim-ple metal cations to organic moleculessuch as aminocyclitols, polymyxin andpolylysine, and they probably form cross-linkages between the protein and electrodesurface, and between adjacent proteinmolecules [1].

Surfactant films accommodate proteinsin a more hydrophobic environment.These films seem to stabilize many pro-teins, and often enhance their electron-transfer and catalytic properties. Surfac-tant films have been effective for pro-teins ranging from myoglobin to largemembrane-bound systems such as thephotosynthetic reaction center [19–21].

1.3Voltammetric Methods

1.3.1Protein Sample Contained in Solution

The most widely used method is DCcyclic voltammetry, although others suchas square-wave voltammetry present use-ful advantages for particular problems.The disadvantages of cyclic voltammetry

1.3 Voltammetric Methods 17

are its relative lack of sensitivity and thefrequent appearance of broad ill-definedwaves from which potential values are dif-ficult to ascertain. The reason for the poorresponse with proteins does not neces-sarily stem from poor electrode kinetics,but from the manner in which the pro-tein diffuses between bulk solution andinteraction sites on the electrode. Thisis illustrated in Fig. 2. In cyclic voltam-metry, the commonly held belief is thatthe current response should be a pair ofpeaks, separated by 59/n mV (where n isthe number of electrons transferred inthe electrode reaction) whose amplitudevaries as the square root of scan rate; how-ever, this is true only for a uniform planarsurface at which diffusion is linear and

is not the case if conditions are such asto provide a steady state current [24, 25].The natural selectivity that proteins exhibitfor their biological reaction partners is ex-pected to extend to electrode interactions,and it may be that only certain zones onan electrode produce active encounters. Asshown in Fig. 2, if these zones are smallor widely separated, the electrode will be-have instead like a microelectrode array:diffusion to isolated sites on the electrodethen gives rise to steady state voltammetryhaving a sigmoidal waveform. The famil-iar peaklike voltammograms appear onlyat scan rates that are sufficiently slow toallow diffusion fields to overlap. Figure 2also shows voltammograms obtained forsolutions of cytochrome c obtained at

Electrode site density Expected CV response CVs for cytochrome cat two types of graphite

+100

0

0.5

+250

Potential vs NHE[mV]

Cur

rent

[µA

]

+400

Basal

Edge

Fig. 2 Reversible voltammograms for: (top row) a diffusing redox couple reactingat a planar macroelectrode at which the entire surface is interactive; (bottom row) adiffusing couple reacting at a microelectrode, or a macroelectrode at which most ofthe surface is blocked to protein interaction. Theoretical voltammograms areshown at the center, while the right hand side shows actual results obtained forcytochrome c at a polished pyrolytic graphite edge plane (top) or basal planeelectrode, (bottom) showing the effect of the density of interactive sites on theelectrode.


electrodes differing greatly in their den-sity of interaction sites [14, 24]. The effectof microscopic domains does not arise forprotein molecules that are adsorbed (orfor thin-layer voltammetry) but can stillbecome important for enzymes if a cat-alytic current is being measured that is dueto turnover of substrates diffusing fromsolution to isolated enzyme molecules ad-sorbed on the electrode [26].

By contrast, SW and DP voltammetries,aside from being inherently more sensi-tive than cyclic voltammetry, produce apeaklike current response regardless ofthe electrode conditions and diffusionalgeometry. Caution should be used, how-ever, since the ‘‘pleasing’’ results that canbe obtained often mask important mecha-nistic information and do not provide suchimmediate visualization of EC reactionsas cyclic voltammetry. However, they dooffer the means to drive reactions moreeffectively by providing large potential per-turbations; thus, as well as being useful

for probing the driving force dependenceof electron transfer, inherently slower re-actions may be detected [27, 28].

1.3.2Protein Sample Confined to Electrode

Redox-active species that are immobilizedon the electrode and undergo simple re-versible electron transfer give rise to apeak-type cyclic voltammetry response thatis not influenced by diffusion effects,but which is, instead, much more sen-sitive to the characteristic properties ofthe protein [1]. Provided the sample ishomogeneous and there are no interac-tions between molecules in the layer, thepeaks for oxidation and reduction shouldhave the Nernstian characteristics definedby Laviron; that is, they comprise finitepassed charge with no tailing current, witha peak separation close to zero [29]. This isdepicted in Fig. 3(a), where the capacitivebackground that is observed in real ex-periments is not shown. The peak widths

(a)

(b)

= 91/n mV at 25 ˚C

7.5

5.0

2.5

0

−2.5

−5.0

−7.5−0.2 −0.1 0 0.1

E vs SCE[V]

j[µ

A c

m−2

]

0.2 0.3 0.4

i α n2

δ

Fig. 3 Electrochemistry for a proteinsample adsorbed on an electrode:(a) theoretical reversible cyclicvoltammogram; (b) voltammograms(scan rate 2 V s−1) of a film of azurinadsorbed on an Au electrode modifiedwith a decane-thiol SAM, recorded onthe fiftieth, hundredth, two-hundredth,and three-hundredth cycles, showingthat the electrochemistry is stable(Q. J. Chi, J. D. Zhang, J. E. T. Andersenet al., J. Phys. Chem. B 2001, 105,4669–4679, with permission of theAmerican Chemical Society). In thesystem studied, k0 is 290 s−1. Rateconstants over 2000 s−1 have beenmeasured for azurin adsorbed on a PGEelectrode, at which the ET distance ismuch shorter.

1.4 Case Studies 19

at half height (δ) should be 91/n mV at25 C (or 84/n mV at 0 C) and the peakcurrents should vary as n2; thus redox cen-ters undergoing a cooperative two-electronreaction (such as flavin in many pro-teins) will appear particularly prominentin the voltammogram. Having the pro-tein molecules already arranged at theelectrode means that the response is notinfluenced by sluggish diffusion, so thatmuch faster reactions can be studied. Thisincludes coupled ‘‘EC’’ reactions in whichthe chemical process is usually the one ofinterest. For an enzyme, analysis of the cat-alytic electron-transport kinetics is aidedby rotation of the electrode to control masstransport of substrate. It is also possible totransfer the coated electrode into a cryosol-vent, and hence study electron transfer andcoupled reactions at low temperatures [30].

For cyclic voltammetry, kinetic informa-tion can be obtained from the behavior ofthe peaks as the scan rate is increased. Insquare-wave voltammetry, the variable pa-rameters are pulse height and frequency,and kinetics are extracted from appro-priate models [27, 28]. Staircase cyclicvoltammetry combines the attributes ofpotential-step (digital) methods with the vi-sual qualities of cyclic voltammetry, whichconventionally involves a linear (analogue)potential ramp [31]. Electron-transfer ki-netics can be analyzed either in terms ofthe Butler-Volmer model (which yields val-ues of k0, the electron exchange constant)or in terms of the Marcus theory, in whichthe system is defined by the parametersλ (reorganization energy) and kmax (thelimiting rate constant at high overpoten-tial). The Butler-Volmer model resemblesthe limit of the Marcus theory as λ → ∞.Figure 3(b) shows voltammetry of a film ofazurin adsorbed on an Au electrode mod-ified with a decane-thiol SAM. The scanswere taken at different times, showing

that the layer is stable, whereas scan-ning tunneling microscopy on the samesystem shows clearly discernable proteinmolecules that are well ordered [12]. Theexchange rate constant k0 is 290 s−1 forthe decane thiol SAM/Au electrode, whilerate constants over 2000 s−1 have beenmeasured for azurin adsorbed on a PGEelectrode, at which the ET distance is muchshorter.

1.3.3General

Other methods that have been applied toprotein electrochemistry include chrono-amperometry, which records the timecourse of the decaying current producedfollowing a step in potential, and ACimpedance methods, which can mea-sure the rate of electron exchange atlow driving force. AC-potential-modulatedimpedance has been used to measurek0 values >104 s−1 in the case of cy-tochrome c adsorbed at short chain-lengthSAMs on Au [23]. The electron-transferreaction of cytochrome c at bare andSAM-modified Ag electrodes has also beenstudied by time-resolved surface-enhancedresonance Raman spectroscopy, monitor-ing the changes in characteristic spectralfeatures of oxidized and reduced hemegroups following a potential jump [32].

1.4Case Studies

1.4.1Observing Active Sites

1.4.1.1 Iron-sulfur ClustersIron-sulfur (Fe-S) clusters are amongthe most common redox-active cofactorsin Biology, yet they lack a distinctivechromophore that enables them to be


observed under ambient conditions and in‘‘real’’ time. The established methods forstudying Fe-S clusters are Electron Para-magnetic Resonance Spectroscopy (EPR)[including ENDOR (electron-nuclear dou-ble resonance) and ESEEM (electron spinecho envelope modulation spectroscopy)],Mossbauer, and MCD, all of which requirecryoscopic temperatures, whereas CD andNMR have been used with varying successto examine clusters at ambient tempera-tures. Another problem is that Fe-S clus-ters tend to have very negative reductionpotentials, so that potentiometric titra-tions lasting several hours must be carriedout under rigorously anaerobic conditions.Each of these problems can be overcome byvoltammetry [33]. Once a particular clus-ter has been identified, its voltammetryprovides both a measure of the reductionpotential and a ‘‘handle’’ for the differentreactions that the cluster may undergo.

Figure 4 shows the voltammetry of asmall protein (a ferredoxin) that containsa [3Fe-4S] and a [4Fe-4S] cluster [34]. Bycorrelation with spectroscopic studies, thetwo redox signals that are observed athigher potential have been assigned to

the well-established [3Fe-4S]+/0 and [4Fe-4S]2+/+ redox couples. The third signal atmore negative potential is only observedin proteins containing a [3Fe-4S] cluster,and inspection of the peak shapes andpH dependence show the reaction toinvolve two electrons in a cooperativemanner coupled to the uptake of at leasttwo protons. This signal is, therefore,assigned to the couple [3Fe-4S]0/2−, whichproduces an unusual ‘‘hyper-reduced’’cluster in which all Fe atoms are formallyin the 2+ oxidation state. Consequently,a [3Fe-4S] cluster in a protein can beidentified by the appearance of two signals;one at modest potentials correspondingto the [3Fe-4S]+/0 couple, and anothermuch sharper signal at more negativepotentials corresponding to the hyper-reduction reaction involving [3Fe-4S]0/2−.

In many proteins, clusters can intercon-vert between different structures, for exam-ple, between [3Fe-4S] and [4Fe-4S] forms.These transformations may be importantphysiologically since they produce changesin reduction potentials, ligand binding,and catalytic activities; and importantly,they depend on the particular oxidation

[3Fe-4S]0/2−

[4Fe-4S]2+/+

−0.5

−0.4

−0.3

−0.2

−0.1

0

0.1

0.2

0.3

0.4

−1 −0.5 0 0.5

E vs SHE[V]

i[µ

A]

[3Fe-4S]+/0−

Fig. 4 Cyclic voltammogram of a filmof ferredoxin that contains one [4Fe-4S]and one [3Fe-4S] cluster. The threeredox couples observable in Sulfolobusacidocaldarius 7Fe ferredoxin have beenassigned on the basis of otherevidences, mostly spectroscopy. SeeRef. [34].

1.4 Case Studies 21

levels of the clusters and, therefore, onthe potential that is applied. Voltammetry,particularly of a film of protein adsorbedon an electrode, allows these reactions tobe visualized and controlled. Metals otherthan Fe may be incorporated, resultingin heterometal clusters [M3Fe-4S], andFig. 5 shows the use of voltammetry tofollow the fast reaction of a [3Fe-4S] clus-ter with Fe2+, Zn2+, and Cd2+ to form[4Fe-4S], [Zn3Fe-4S], and [Cd3Fe-4S] clus-ters respectively [35, 36]. This particularexample, ferredoxin III isolated from asulfate reducing bacterium Desulfovibrioafricanus, contains two Fe-S clusters, a[4Fe-4S] cluster that is inert and a [3Fe-4S] cluster that is reactive and has a highaffinity for various metal ions. The pro-tein is adsorbed on a PGE electrode in thepresence of polymyxin, and on addition ofmetal ions (M2+) the reaction is observedsimply by cycling the potential. By employ-ing periods of time in which the electrodeis polarized at different potentials, it ispossible to determine the stabilities andmetal ion affinities of different oxidationlevels. A number of different heterometalderivatives [M3Fe-4S] (M2+ = Zn, Cd, Co,Cu, Pb; and M+ = Tl and Cu) have beenprepared ‘‘on the electrode’’ in adsorbedproteins and their stabilities, and redoxproperties have been compared [33, 35, 36].

1.4.1.2 FlavinsFlavins (FAD and FMN) are importantcofactors in enzymes, where they often oc-cur together with various metal centers.In many cases, they undergo cooperativetwo-electron transfers (the semiquinoneradical is usually unstable) so that in a pro-tein that is adsorbed on an electrode theycan be distinguished by a much sharpersignal than that displayed by one-electronsites [37–39]. This is illustrated in Fig. 6,which shows the result obtained for afilm of E. coli fumarate reductase (molec-ular mass 93 kDa) that contains an FADcofactor and three Fe-S clusters. From the

Fig. 5 Reaction of a [3Fe-4S] clusterwith metal ions, as observed byvoltammetry of a film of ferredoxin fromDesulfovibrio africanus. Only theoxidative sweep is shown. The scan rateis 200 mV s−1. Note the decrease inintensity (↓) of the two signals from[3Fe-4S]+/0 and [3Fe-4S]0/2− redoxcouples, while the signal due to[M3Fe-4S]2+/+ grows (↑). (J. N. Butt,F. A. Armstrong, J. Breton et al., J. Am.Chem. Soc. 1991, 113, 6663–6670, withpermission of the American ChemicalSociety.)

[3Fe-4S]0 [M3Fe-4S]2+

MM2+

−1.0

Fe

1 µA

Zn

Cd

−0.5

+

+

+

0E vs SHE

[V]


−600

(b)

(a)

−50

−20

−10

0

10

20

0

50

−400 −200

E vs SHE[mV]

i[n

A]

i[n

A]

0 200

−600 −400 −200 0 200

No substrate

1.2mM succinate4µM fumarate

Fum

arat

ere

duct

ion

Succ

inat

eox

idat

ion

Red

uctio

nO

xida

tion

Fig. 6 Voltammetry of E. coli fumarate reductase:(a) voltammogram in the absence of substrate;(b) voltammogram obtained in the presence of fumarate andsuccinate. Note that the oxidation and reduction currents areequal when the succinate/fumarate ratio is 300. (C. Leger,K. Heffron, H. R. Pershad et al., Biochemistry 2001, 40,11 234–11 245, with permission of the American ChemicalSociety.)

areas under the signals, the electroactivecoverage is approximately equivalent to amonolayer. The signal due to the FADis clearly observable and can be followedeven at a scan rate of 100 V s−1, whichis fast enough to outrun catalysis when asubstrate molecule is bound [38, 39].

1.4.1.3 Highly Oxidizing Intermediates:Fe(IV)=OFe(IV) is an important intermediate insuch heme enzymes as cytochrome c ox-idase and cytochrome P450, as well asthe nonheme enzymes methane monooxy-genase and ribonucleotide reductase.Voltammetric studies on yeast cytochrome

c peroxidase and various mutant formshave provided insight into the redox prop-erties of the catalytic intermediate (Com-pound I) that contains the ferryl group(Fe(IV)=O) and a cation radical. Whenadsorbed at a PGE electrode, cytochromec peroxidase catalyzes reduction of hydro-gen peroxide at potentials (>0.7 V) that aremuch higher than required for the naturalelectron donor, cytochrome c (0.26 V) [26].The pyrolytic graphite electrode can betaken to quite high potentials withoutexcessive water oxidation or surface degra-dation. An example is shown in Fig. 7.In the absence of substrate, a reversibleredox couple is observed, the half-height

1.5 Resolving Complex Reactions 23

Fig. 7 Voltammetry of a film of yeastcytochrome c peroxidase adsorbed on apyrolytic graphite edge electrode.(A) Voltammogram obtained in theabsence of substrate, after adsorbingthe enzyme from dilute solution at 0 C.Both reduction and oxidation peakwidths are below 84 mV. (B) In thepresence of H2O2: the electrode isrotating at 400 rpm, but note that thereduction current in the negativedirection still retains a peaklike feature,suggesting that the rotation rate is nothigh enough to ensure steady stateconditions. (M. S. Mondal, H. A. Fuller,F. A. Armstrong, J. Am. Chem. Soc.1996, 118, 263–264, with permission ofthe American Chemical Society.)

Potential vs SHE

Cur

rent

[µA

]

0.1

0.0

−0.1

0.4 0.6 0.8 1.0

0.2

0.0

−0.2

−0.4

AB

peak widths of which (both oxidation andreduction) are significantly narrower thanthe one-electron value. This suggests thatthe two-electron transition between Fe(III)and Compound I is a cooperative pro-cess [31, 40]. The potential at pH 5 is760 mV for WT and 880 mV for a mutantW51F in which a distal-pocket tryptophanhas been replaced by phenylalanine [41].On addition of H2O2, the peaks transformto an amplified catalytic wave (dashed line),which also commences at high potential(see later). Voltammetry is, therefore, use-ful in determining the factors that stabilizeFe(IV) and in relating the redox energeticsto catalytic activity.

1.5Resolving Complex Reactions

1.5.1Coupled and Gated Electron Transfer

The heme groups in many enzymes canadopt five- and six-coordinations, with thefive-coordinate center being available for

substrate binding and turnover. Smallproteins, including cytochrome c and site-directed variants, provide good examplesof redox-dependent changes in axial liga-tion that are relevant to the larger enzymes,and which are ‘‘EC’’ reactions particularlysuited to study by voltammetry. In mito-chondrial cytochrome c, the Fe is axiallyligated by a histidine and a methioninethioether. However, oxidation at high pHis followed by a ligand exchange (the me-thionine dissociates, to be replaced by anearby lysine), which can be observed bycyclic voltammetry. The ‘‘high pH’’ formhas a much lower reduction potential, asexpected because the methionine stabi-lizes the Fe(II) form. The interconversionis quite slow, and can be studied with cy-tochrome c solutions that yield reversibleET at modified Au electrodes [42]. A partic-ularly good example is a variant of yeast cy-tochrome c, in which phenylalanine-82, aligand close to the axial ligand methionine-80, is replaced by a histidine [43]. Uponoxidation to Fe(III), methionine-80 isdisplaced spontaneously by histidine-82,


(a)

(b)

0.30 0.00

i[µ

A]

E vs Ag/AgCl[V]

−0.30−6

−3

0

3

6

−0.60

His-82

His-82 O' R'

RO

C0 CR

E2

E1

Met-80

Met-80

His-18

Fe(III)

Fe(III) Fe(II)

Fe(II)

His-18

O → R

O' → R'

R' → O'

R → O

His-82

His-82

Met-80

Met-80

His-18

His-18

Fig. 8 (a) square scheme depicting redox-coupled ligandexchange at a mutant form of yeast cytochrome c (F82H), in whicha phenylalanine residue close to the Fe has been replaced by ahistidine. In the oxidized state (Fe(III)) the new ligand competeseffectively with methionine-80 that normally ligates the Fe;(b) voltammogram of a solution of F82H cytochrome c reacting atan Au electrode modified with a layer of bis(4-pyridyl) disulfide.Scan rate is 50 mV s−1. (B. A. Feinberg, X. Liu, M. D. Ryan et al.,Biochemistry 1998, 37, 13 091–13 101 with permission of theAmerican Chemical Society.)

while reduction is followed by spontaneousrecombination with methionine-80. Thereactions are represented by a squarescheme, which is shown along with theresulting EC voltammetry in Fig. 8. At ascan rate of 50 mV s−1, the oxidation peak

appears at high potential, while reductionis seen at low potential. From the way thatthe voltammetry changes with scan rate,the rate constants for the ligand exchangereactions can be determined. In this casethe reactions are quite slow.


Various coupled reactions occurring atFe-S clusters have been investigated. Asan example, a detailed study of protontransfer has been carried out on thewell-characterized 7Fe ferredoxin fromAzotobacter vinelandii, which is structurallydefined to 1.4 A resolution, and exhibitsfast electron transfer at a PGE electrode,similar to that shown in Fig. 4 [44, 45].In many proteins, [3Fe-4S] clusters binda proton in their one-electron reducedstates, a scheme for which is shownin Fig. 9. The oxidized form is a veryweak base, expected to have a very highreduction potential, so it can be ignored. InAzotobacter vinelandii ferredoxin, the [3Fe-4S] cluster is buried below the surface,and the proton must transfer through an‘‘anhydrous’’ barrier. How this transferoccurs, and the nature by which long-range proton transfer is coupled with

electron transfer are both relevant tothe wider question of the mechanismof enzymes known as proton pumps. Amutant (D15N) in which proton transferis retarded provides an example of thedetection and quantitative investigation of‘‘gating’’. In this case, electron transferto the cluster draws a proton into theprotein, whereas the reoxidation processis gated by proton release. In D15N,an aspartate residue whose carboxylateis exposed to the surface above the[3Fe-4S] cluster has been replaced byasparagine.

Figure 9 shows plots of the positionsof oxidation and reduction peaks observedwhen the pH of the contacting electrolyte isabove and below the pK of the reduced clus-ter. Under the conditions pH pK, theelectrode reaction involves only electrontransfer; therefore, the peaks separate as

Fig. 9 Plots of peak positions againstscan rate for the redox reaction of theburied [3Fe-4S] cluster in a D15Nmutant of ferredoxin I from Azotobactervinelandii. Under solution conditionspH > pK, only an electron is transferred;by contrast, for pH < pK, protontransfer also occurs, but at a rate that issufficiently slow so that it becomesuncoupled at high-scan rates and protonrelease gates the oxidation process (nopeak is observed) at intermediate scanrates. See Refs. [44 and 45]. −0.1

−0.05

0

0.05

0.1

0.15

0.01 0.1 1 10 100

pH << pK

pH >> pK

Peak potential vs Ealk[V]

Scan rate[V s−1]

[3Fe-4S]+ [3Fe-4S]0Ealk

k0

[3Fe-4S]0−H+

H+

K = koffkon


the scan rate is raised, giving rise to a ap-proximately symmetrical trumpet-shapedplot. Analysis of this plot in terms of theButler-Volmer model gives a standard rateconstant of 550 s−1. By contrast, at lowpH, three distinct regions of the plot areobserved. First, at slow scan rates, the ox-idation and reduction peak positions areclose together; but at a much higher po-tential, as expected since the reductionpotential increases as the pH is lowered.Second, at fast scan rates, the peak posi-tions overlay those observed at high pH:this is because (commencing the cyclefrom the oxidative potential limit) the elec-tron exchange occurs in both directionsbefore the proton can transfer to the clus-ter. Finally, in the intermediate region,no oxidation peak is observed because theproton now has time to arrive at the clus-ter, but on the return scan the electron istrapped (species [3Fe-4S]0-H+) until theproton has been removed; the electrontransfer is thus ‘‘gated’’ by the release ofthe proton (k = 2.5 s−1). The plot, there-fore, divides the EC reaction into threedistinct time domains. The native proteinexhibits more complex kinetics, in whichthe rate constants kon and koff are muchhigher, assisted by the mobile aspartateside chain that binds a proton reversiblyand serves as a courier.

1.5.2Catalysis

For enzymes adsorbed on an electrode,both nonturnover and catalytic studiescan be carried out. Nonturnover mea-surements are essentially the same asmentioned earlier, and focus on the ac-tive site redox transitions occurring in theabsence of substrate. Addition of the sub-strate to the electrolyte results in catalytic

activity, which can be studied under tran-sient or steady state conditions. Providedelectron transfer is faster than turnover,the transformation to nonturnover voltam-metry as the scan rate is increased (to atime domain too short to allow turnover tooccur) provides information on the redoxproperties of the enzyme-substrate com-plex. For fumarate reductase (and otherflavoenzymes), the prominence of the FADsignal over the other centers in the en-zyme enables its oxidation and reductionpeak positions to be observed, even at100 V s−1; so that the reduction poten-tial can be measured in the absence orpresence of substrate [38, 39].

In steady-state voltammetry experi-ments, enzyme activity is viewed in the‘‘potential domain’’ that can pinpoint therole of centers as electron relays, or revealthe presence of internal control mecha-nisms, such as a redox transformationthat causes the enzyme to ‘‘switch off’’at a certain potential. Such studies canalso reveal and quantify how an enzymeis ‘‘redox-biased’’ to favor catalysis in aparticular direction. Figure 6(b) shows thevoltammetry of a film of fumarate reduc-tase obtained in the presence of a lowconcentration of fumarate and a high con-centration of succinate, from which it iseasily seen how the catalytic activity of theenzyme is biased heavily in the direction offumarate reduction [38]. This experimenthas been carried out with a rotating discelectrode. The current for succinate ox-idation is independent of rotation rate,while that for fumarate reduction is verysensitive because the reaction is diffusioncontrolled.

Figure 10 shows the catalytic voltam-metry of a film of nitrate reductase, amembrane-bound enzyme that containsa Mo active site and Fe-S clusters [46]. Theenzyme is adsorbed on a PGE electrode;


−0.4

−4

1000 A

45

4

[NO3]µM

−2

0

NarI ElectrodeMembrane

[3Fe-4S]3[4Fe-4S]

Mo-bis-MGD

NO3− + 2H+

2H+

QQH2

NO2− + H2O NO3

− + 2H+ NO2− + H2O

Mo-bis-MGD

[3Fe-4S]3[4Fe-4S]

cyt b

cyt b2 × 1e−

2 × 1e−

NarH

NarG

−0.2 0.0

E vs SHE[V]

Cur

rent

[µA

]

0.2 0.4

Fig. 10 Catalytic voltammetry of a film of nitrate reductase:(a) cartoon showing the membrane-extrinsic sub-complexNarGH bound to the membrane-intrinsic NarI (left), andadsorbed instead on an electrode (right); (b) catalytic volta-mmetry of a film of NarGH adsorbed on a PGE electrode inthe presence of different concentrations of nitrate (voltammo-gram without substrate is also shown). Note (arrowed) thatthe catalytic current for the experiment at low nitrate concen-tration passes through a maximum. (L. J. Anderson, D. J. Rich-ardson, J. N. Butt, Biochemistry 2001, 40, 11 294–11 307, withpermission of the American Chemical Society.)

although in this case, unlike that for fu-marate reductase, the electroactive cover-age is too low to observe signals under non-turnover conditions. An interesting obser-vation is that for low levels of nitrate the

activity (reduction current) passes througha maximum value at a particular poten-tial, below which it decreases even thoughthe driving force is raised. It is knownthat during the catalytic cycle, the Mo


center operates between Mo(VI), Mo(V),and Mo(IV) oxidation states, and one in-terpretation is that substrate binds moretightly to the intermediate Mo(V) formthan to Mo(IV). This idea is supported byextensive measurements of the substrateand pH dependencies of voltammograms,and correlations with EPR-potentiometricdata [46]. Applying a more negative po-tential causes the activity to drop, so thatnitrate reductase is an example of an en-zyme that displays a potential dependent‘‘switch’’. Similar properties are observedfor another Mo enzyme, DMSO reductasefrom E. coli [47] and for succinate dehy-drogenase from beef heart or E. coli [48].These effects are difficult to observe byconventional experiments, but are easilydetected by voltammetry and may be im-portant physiologically in the regulation ofmetabolism.

1.6Conclusion

The study of protein molecules at sur-faces is a field in its infancy, for whichdynamic electrochemical methods alreadyplay an important role. Contrary to thebeliefs once held, protein molecules do‘‘talk’’ to electrodes. Unique insight isbeing gained into complex problems offundamental interest, and new techno-logical applications are being identified.The methodology is interactive: electrontransfer and coupled reactions are in-duced under precise conditions; and then,simultaneously, these reactions can bemonitored, deconvoluted, and quantified.Future efforts are expected to focus moreon larger and more complex enzymes, par-ticularly those that exist, in vivo, associatedwith membranes. The problem of design-ing membrane-mimetic enzyme systems

on electrodes while retaining fast electrontransfer will provide important challengesfor the future.

References

1. F. A. Armstrong, H. A. Heering, J. Hirst,Chem. Soc. Rev. 1997, 26, 169–179.

2. F. A. Armstrong in Bioelectrochemistry ofBiomacromolecules: Bioelectrochemistry: Prin-ciples and Practice (Eds.: G. Lenaz, G. Milazo),Birkhauser Verlag, Basel, 1997, pp. 205–255.

3. F. A. Armstrong, G. S. Wilson, Electrochim.Acta 2000, 45, 2623–2645.

4. L. D. Gilles de Pelichy, E. T. Smith, Biochem-istry 1999, 38, 7874–7880.

5. A.-E. Nassar, Z. Zhang, N. Fu et al., J. Phys.Chem. B 1997, 101, 2224–2231.

6. Q. Feng, S. Imabayashi, T. Kakiuchi et al.,J. Electroanal. Interfacial Electrochem. 1995,394, 149–154.

7. S. Lecompte, H. Wackerbarth, T. Souli-mane et al., J. Am. Chem. Soc. 1998, 120,7381–7382.

8. M. J. Eddowes, H. A. O. Hill, J. Chem. Soc.Chem. Commun. 1977, 771–772.

9. P. Yeh, T. Kuwana, Chem. Lett. 1977,1145–1148.

10. P. M. Allen, H. A. O. Hill, N. J. Walton,J. Electroanal. Interfacial Electrochem. 1984,178, 69–86.

11. A. El-Kasmi, J. M. Wallace, E. F. Bowdenet al., J. Am. Chem. Soc. 1998, 120, 225–226.

12. Q. J. Chi, J. D. Zhang, J. E. T. Andersen et al.,J. Phys. Chem. B 2001, 105, 4669–4679.

13. L. J. C. Jeuken, F. A. Armstrong, J. Phys.Chem. B 2001, 105, 5271–5282.

14. F. A. Armstrong, P. A. Cox, H. A. O. Hillet al., J. Electroanal. Interfacial Electrochem.1987, 217, 331–366.

15. E. F. Bowden, F. M. Hawkridge, H. N.Blount, J. Electroanal. Interfacial Electrochem.1984, 161, 355–376.

16. M. A. Harmer, H. A. O. Hill, J. Electroanal.Interfacial Electrochem. 1985, 189, 229–246.

17. I. Taniguchi, K. Watanabe, M. Tominaga,F. M. Hawkridge, J. Electroanal. InterfacialElectrochem. 1992, 333, 331–338.

18. I. Taniguchi, Y. Hirakawa, K. Iwakiriet al., J. Chem. Soc. Chem. Commun. 1994,953–954.

19. J. F. Rusling, Acc. Chem. Res. 1998, 31,363–369.

1.6 Conclusion 29

20. J. L. Kong, Z. Q. Lu, Y. M. Lvov et al., J. Am.Chem. Soc. 1998, 120, 7371–7372.

21. P. J. Farmer, R. Lin, M. Bayachou, CommentsInorg. Chem. 1998, 20, 101–120.

22. D. E. Reed, F. M. Hawkridge, Anal. Chem.1987, 59, 2334–2339.

23. A. Avila, B. W. Gregory, K. Niki et al., J. Phys.Chem. B 2000, 104, 2759–2766.

24. F. A. Armstrong, A. M. Bond, H. A. O. Hillet al., J. Phys. Chem. 1989, 93, 6485–6493.

25. F. A. Armstrong, A. M. Bond, H. A. O.Hill et al., J. Am. Chem. Soc. 1989, 111,9185–9189.

26. F. A. Armstrong, A. M. Bond, F. N. Buchiet al., Analyst 1993, 118, 973–978.

27. T. M. Saccucci, J. F. Rusling, J. Phys. Chem.B 2001, 105, 6142–6147.

28. L. J. C. Jeuken, J. P. McEvoy, F. A. Arms-trong, J. Phys. Chem. B 2002, 106, 2304–2313.

29. E. Laviron in Electroanalytical Chemistry (Ed.:A. J. Bard), Marcel Dekker, New York, 1982,pp. 53–157, Vol. 12.

30. J. P. McEvoy, F. A. Armstrong, J. Chem. Soc.Chem. Commun. 1999, 1635–1636.

31. H. A. Heering, M. S. Mondal, F. A. Arm-strong, Anal. Chem. 1999, 71, 174–182.

32. D. H. Murgida, P. Hildebrandt, J. Am. Chem.Soc. 2001, 123, 4062–4068.

33. F. A. Armstrong in Advances in InorganicChemistry (Eds.: A. G. Sykes, R. Cam-mack), Academic Press, New York, 1993,pp. 117–163, Vol. 38.

34. J. L. C. Duff, J. L. J. Breton, J. N. Butt et al.,J. Am. Chem. Soc. 1996, 118, 8593–8603.

35. J. N. Butt, F. A. Armstrong, J. Breton et al.,J. Am. Chem. Soc. 1991, 113, 6663–6670.

36. J. N. Butt, S. E. J. Fawcett, J. Breton et al.,J. Am. Chem. Soc. 1997, 119, 9729–9737.

37. A. Sucheta, R. Cammack, J. Weiner et al.,Biochemistry 1993, 32, 5455–5465.

38. C. Leger, K. Heffron, H. R. Pershad et al.,Biochemistry 2001, 40, 11 234–11 245.

39. A. K. Jones, R. Camba, G. A. Reid et al.,J. Am. Chem. Soc. 2000, 122, 6494–6495.

40. M. S. Mondal, H. A. Fuller, F. A. Armstrong,J. Am. Chem. Soc. 1996, 118, 263–264.

41. M. S. Mondal, D. B. Goodin, F. A. Arm-strong, J. Am. Chem. Soc. 1998, 120,6270–6276.

42. P. D. Barker, A. G. Mauk, J. Am. Chem. Soc.1992, 114, 3264–3619.

43. B. A. Feinberg, X. Liu, M. D. Ryan et al.,Biochemistry 1998, 37, 13 091–13 101.

44. J. Hirst, J. L. C. Duff, G. N. L. Jameson et al.,J. Am. Chem. Soc. 1998, 120, 7085–7094.

45. K. Chen, J. Hirst, R. Camba et al., Nature2000, 405, 814–817.

46. L. J. Anderson, D. J. Richardson, J. N. Butt,Biochemistry 2001, 40, 11 294–11 307.

47. K. Heffron, C. Leger, R. A. Rothery et al.,Biochemistry 2001, 40, 3117–3126.

48. H. R. Pershad, J. Hirst, B. Cochran et al.,Biochim. Biophys. Acta 1999, 1412, 262–272.

31

2Single Cell Electrochemistry

Jonathan M. CooperUniversity of Glasgow, Glasgow, United Kingdom

Sung-Kwon JungBiocurrents Research Center, Woods Hole, Massachusetts

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

2.2 Advantages of Microelectrodes in Single Cell Studies . . . . . . . . . . . 35

2.3 Fabrication of Planar Microelectrodes Using Photolithography . . . . 362.3.1 Functional Three-dimensional Micromachined Electrochemical

Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382.3.2 Advantages of Using Three-dimensional Microstructures . . . . . . . . 392.3.3 Biosensors, Microfluidics, Microarrays, and Lab-on-a-Chip . . . . . . . 40

2.4 Self-referencing Microelectrodes . . . . . . . . . . . . . . . . . . . . . . . . 412.4.1 Ion-selective Microelectrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . 432.4.2 Self-referencing Amperometric Microelectrodes . . . . . . . . . . . . . . 43

2.5 Scanning Probe Microscopy and Single Cell Measurement . . . . . . . 44

2.6 Semiconductor Devices for Biological Sensing . . . . . . . . . . . . . . . 46

2.7 Corroborative Measurements for Single Cell Electrochemistry . . . . . 47

2.8 Future Prospects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

33

2.1Introduction

There is a huge diversity of cell types avail-able for the biologist to study, rangingin size from a few microns in diameterto several tens of centimeters in length.The variation in size is only matched bythe diversity of structure, illustrated byeukaryotic animal cells (enveloped by acell membrane) and plant cells (enclosedby a robust cell wall). The greatest inter-est in single cell study has resulted fromits relevance to human medicine and hasincluded examples from the prokaryotes(such as the bacteria) and from the eu-karyotes (e.g. mammalian cells). As thesmallest unit of sustainable life, the singlecell has provided a unique understand-ing of more complex biological systems.Analysis of metabolic events at this levelhas provided fundamental informationabout a wide range of important pro-cesses including cell signaling, cell–druginteractions, and disease mechanisms (in-cluding ischemia and cell death). There isnow not only the possibility of obtainingdata concerning biochemical processes,but also detail on physiological functional-ity, including for example, the generationof the action potential in neurons andthe propagation of cell contraction inmyocytes.

In terms of understanding complex bi-ological systems, the study of the singlecell has provided a number of clear ad-vantages, particularly the deconvolution ofsignaling and metabolic events withoutthe need to be concerned by the his-tory, distribution, integrity, and activityof neighboring cells. In order to providesuitable analytical tools there has beenextensive research into a variety of opti-cal and electrochemical techniques witha correspondingly large literature. In thefield of optical sensing, both natural andsynthetic fluorescent probes have beenintroduced into cells through molecularbiology, electroporation, and absorptionin order to study both fundamental andapplied aspects of the cellular physiologyand biochemistry, as well as disease pro-cesses and drug activity. Generally, suchfluorescent probes are ‘‘environmentallysensitive’’ such that their emission prop-erties change as either the intracellularor the local conditions change (e.g. pH,ionic concentration). More recently, thegenes for a variety of naturally occurringfluorochromes (such as green fluorescentprotein) have been introduced at strategicpoints in the cell’s genome, such that it isnow possible to tell if proximal genes havebeen activated (e.g. by the deliberate addi-tion of a drug). Single cell fluorescence hasalso been greatly enhanced by advances

34 2 Single Cell Electrochemistry

in organic chemistry (in producing newfluors), in the production of ultrasensitiveimaging devices (most notably the chargecoupled device, CCD) and new forms ofmicroscopy, such as confocal imaging andnear field scanning optical microscopy.

As an alternative to electroanalyticalmethods optical sensing have been devel-oped over the last 20 years and are still usedextensively in research and industry [1–3].One important technique, known as patchclamping or voltage clamping, involvesdrawing a glass microcapillary and us-ing it to capture a portion of the cellmembrane, close to a potentiometric (ion)sensor. Preferably, a single ion channelis trapped within the low volume of thecapillary, while the cell maintains its in-tegrity and viability. As a method, it is ableto detect quantal events in ion transport,resulting from the opening and closingof single channels. Although this chapterdoes not focus in detail on either patchclamping or fluorescence, both techniquesare referred to again, later, as they provideimportant corroborative techniques, whenused in conjunction with new electroana-lytical methods.

Despite the predominant use of fluores-cence and/or patch clamp techniques insingle cell measurements, there has beena steady increase in the demand for newelectroanalytical tools applicable to singlecell studies [4]. Traditionally, such meth-ods have been confined to the developmentand production of hand crafted sensorsincluding the aforementioned glass capil-laries [1–3] for patch clamping, as well asconical microelectrodes for scanning elec-trochemical microscopy (SECM) [5, 6] andcarbon fiber microelectrodes to measurefor example, the release of neurotransmit-ter from single neurons [7, 8].

More recently, techniques adapted di-rectly from the semiconductor indus-try [9–13] have been used to manufacturesensors with more closely defined physicalcharacteristics. In this context, the singlecell, as a subject of study is of a compa-rable size to a variety of Bio-microelectro-mechanical systems (Bio-MEMS) devicesthat can now be produced through micro-fabrication and micromachining. Indeed,despite the small size of the single cell, rela-tively large signals can be readily obtained,ranging from several tens or hundredsof picoamperes in case of amperometricmeasurements, and across several decadesof millivolts in the case of intracellularmembrane potential measurements. Inthe latter cases, there is often the needto functionalize the electrode with an en-zyme, thus creating a microbiosensor.One result of this, which presents a dif-ficulty in making robust measurements, isthat current densities tend to be smaller,often a result of the low catalytic activ-ity of the enzyme and the small fluxesof analyte, either within or proximal tothe cell.

In all cases, when low volume, singlecell electrochemical measurements are be-ing made on single cells, care must betaken to consider the fate of all speciesinvolved both in the electrochemical andthe cell reactions (including the availabil-ity of oxygen, the levels of potentially toxicbyproducts of cell metabolism; the buffer-ing capacity of the system; the availabilityof nutrients; and the gradients of pro-tons, either at the working or counterelectrodes). In this latter context, Lab-on-a-Chip methods, described later maybe used to create a microincubator sys-tem to constantly replenish the cell inculture.

In this chapter, we set out to reviewthe current state of the art in single

2.2 Advantages of Microelectrodes in Single Cell Studies 35

cell electrochemical techniques, datingfrom early techniques that involved thedirect electrochemical measurement ofdopamine and the catecholamines [7, 8].More recently, there have been a vari-ety of methods [9–11], in which enzymeassays have been linked to electrochem-ical measurements, to provide a degreeof specificity and enabled the determina-tion of purines and lactate [13, 14]. Whilethe use of Bio-MEMS methods has sig-nificantly lowered the absolute detectionlimit through minimal analyte dilution[9–11, 13, 14], the self-referencing tech-nique, which is based on a differencemeasurement between two locations ina gradient, has permitted determinationof real-time flux values with minimal im-pact of sensor drift or noise [15–19]. Thislatter approach belongs to a family of tech-niques employing position modulation ofmicroprobes in order to enhance detec-tion. The methods aim to negate smallbut irregular drift, which often compro-mises electrochemical measurements atthe single cell level, by assuming thetime-dependent drift is common to bothphysical locations of the sensor. Thus,the detection limit can be further reducedby two orders of magnitude below typ-ical applications of microelectrodes. Thechapter finally concludes with a brief re-view of aspects of probe microscopy in cellanalysis.

2.2Advantages of Microelectrodes in SingleCell Studies

In general, the advantages of using mi-croelectrodes in electroanalysis (in whichone physical dimension is below 100 µm)are already well documented [11, 13, 14,

20–28]. Most significantly, in all casesminiaturization of the electrode changesthe nature of the profile of the diffusiongradient of the analyte between the bulk so-lution and the sensor surface (where oftenthe process of measurement is changingthe redox nature of the species under inves-tigation). The exact nature of the diffusionprofile is critically dependent on the pre-cise geometry of the microelectrode, andmay most often be approximated as a point,a disc, a cylinder, or a band. Regardlessof the exact dimensions or geometry, amicroelectrode benefits from more effi-cient transport of analyte to the sensorsurface, resulting in higher fluxes of re-dox species and faster response times toachieve steady state equilibrium. Thereis a consequent improvement in ratio ofthe Faradaic current (the signal) to thenon-Faradaic charging current (the latterdecreasing as the double layer capaci-tance of the sensor falls as a function ofthe reduced surface area). Improvementsin amplifiers, and their ready incorpora-tion into instrumentation, now make lownoise measurements routine, althoughclearly both proper shielding and ground-ing remain essential. Measurements ofpicoampere currents (typically equivalentto measurement of submicrovolt sizedvoltages) are commonplace, and at thestate of the art, tens of nanovolts can beresolved.

In principle, both the spatial resolu-tion of the measurement (i.e. the abilityto position the electrode) and the sig-nal to noise of the analytical signal willboth improve with further miniaturiza-tion of the probe. While both of theseattributes should be of significant impor-tance in single cell measurements, thereare, however, practical limits to the sizeof sensor, which ultimately defines the


ease with which the electrode can be fab-ricated and manipulated. The dimensionof the microelectrode will also determineits mechanical robustness, and the sizeof the absolute signal being recorded.Static ‘‘probe’’ microelectrodes have gen-erally been limited to a diameter ofca. 10 µm [7, 8, 20–22, 25–28], while ‘‘nan-odes’’ (in which one dimension <1 µm)have been mounted and rastered acrossa surface to produce very high-resolutiontwo-dimensional maps (or micrographs)of ion- and redox fluxes, a technique moregenerally termed scanning electrochemi-cal microscopy.

Thus electrochemical microelectrodes,providing good spatial and fast tempo-ral resolution can be easily positioned ata predefined distance from either a con-fluent cell culture or a single cell. Aslong ago as the mid-1970s Wightmanused single carbon fiber amperometricmicroelectrodes implanted (via a cannula)to measure dopamine release from ratsin vivo [7]. Since then, similar methodshave been used to determine a varietyof electroactive extracellular species suchas tryptophan and the catecholamines [8].These sensors can also been positioned inconjunction with an optical fiber close tothe cell to obtain electrochemical measure-ments, while simultaneously measuringchanges in ion fluxes, such as calcium,using fluorescent indicator dyes [28]. Sim-ilar electronalytical approaches have beenused very successfully to monitor cel-lular oxidative bursts from single hu-man fibroblasts [20, 22], including forexample, the measurement of superox-ide and peroxynitrite. In related work,peptides secreted from melanocytes andserotonin secreted from pancreatic β-cellshave been monitored by electrochemicaldetection at a carbon fiber microelec-trode [29, 30].

2.3Fabrication of Planar Microelectrodes UsingPhotolithography

Whether considering a static or a scanned‘‘probe’’ electrode, the size of the signal,and hence the interpretation of the data,critically depends on being able to faith-fully reproduce the active measurementarea. For hand crafted electrodes, the fabri-cation procedures for these electrodes areoften labor-intensive and when the sizebecomes less than 10 µm, calibration is re-quired because of the irregularity in size.In order to fabricate identical microelec-trodes, methods are being adapted fromthe electronics industry in order to processdevices with high throughputs and lowcosts. By using photolithography [9–11,13, 14, 31] (literally writing with light), it ispossible to precisely control the depositionof both metals and insulators on a surface,and hence define planar microelectrode ar-rays, Fig. 1. The technique makes use of amask, a (UV) light source and a spun pho-toactive polymer (or photoresist) to transferthe pattern (from the mask) onto a surface,as a template for the subsequent deposi-tion of metals or dielectrics. For example,metals such as gold or platinum, whichprovide suitable electroanalytical surfaces,can be readily deposited over the templateand the geometry of the device revealed bydissolving the polymer using a techniqueknown as ‘‘lift-off’’. The process, which in-volves a number of sequential steps, canbe repeated many times to reproduce iden-tical microstructures.

At the center of the microfabricationprocedure is the need to make a very accu-rate representation of the microelectrode(as the mask), for it is this ‘‘original’’ thatwill determine the quality of all subse-quent devices. In addition, in order toensure a faithful reproduction, careful

2.3 Fabrication of Planar Microelectrodes Using Photolithography 37

UV light

Quartz

Chromium

(a)

(b)

(c)

(d)

(e)

54.74°

<111>

Photoresist

Substrate

Photoresistremoved

Metal

Mask

Fig. 1 Schematic of the fabrication processes used in planar microfabrication. Themethods have been adapted to micromachining of surfaces to produce microstructuredsensors, particularly those involved in producing gold planar microelectrodes on glass,used in electronic Petri dishes. (The schematic is reproduced from Ref. [31], although moredetails of the processes involved can be found in Ref. [9].)

consideration has to be given to the lengthof the exposure, the thickness of the poly-mer, the size of features on the mask,and the optical arrangement used. In thesame manner in which metals can bepatterned, dielectrics (insulators), includ-ing silicon nitride, photoactive glasses,

and thin and thick film polymers canalso be deposited likewise, and used toisolate wires and delineate an active sens-ing area.

The process of photolithography andlift-off is highly reproducible, and is readilycapable of defining sensor surface area


with feature sizes of ca. 2 µm. The limitof a feature, which can be fabricated byphotolithography is a consequence of thewavelength of the light used in patterntransfer and the quality of the optics.Feature sizes as small as 150 nm, can befabricated, although, in practice the costof the instrumentation and the yield ofproduction limit the size. Although the de-tailed methods of planar microfabricationtechnologies used in producing devicesfor bioanalysis are essentially the same asthe core technology used in the micro-electronics industry, there are a numberof important differences in the materi-als used [31]. As there often is a need toview the cell(s) under a microscope, the‘‘substrate’’ on which devices are fabri-cated is most often glass, which is opticallytransparent. Likewise, noble metals, par-ticularly gold, are used throughout thedevice to prevent corrosion and providea suitable electrode surface. Finally pack-aging, including encapsulation of activeelectronics or wires from corrosive bio-logical solutions is required. The result isthat a device made by photolithographicpatterning of metals, and packaged pro-vides a means for the production of awide variety of two- or three-dimensionalelectrode arrays, which can be presented

in a similar manner to that shown inFig. 2.

2.3.1Functional Three-dimensionalMicromachined Electrochemical Devices

One further application of planar, two-dimensional microfabrication is that themethods can be used as the foundationto develop three-dimensional structures, amethodology known as micromachining.By providing a volume above the electrodearray there is the possibility of being ableeither to constrain the cell in a fixedvolume, so preventing analyte dilutioninto the ‘‘bulk’’, or of creating networksof fluidic channels to enable either thecell or the fluid to be moved relative toeach other (ultimately comprising Lab-on-a-Chip and microfluidic techniques, seefollowing discussion).

Micromachining has been used in theproduction of electroanalytical structuresthat have ultralow volumes, typically inthe subnanoliter range [10, 11, 13, 14],and are thereby compatible with singlecell methods. Broadly speaking, two dif-ferent methods have been described forproducing such picoliter-scale devices: em-bossing (or stamping) of microstructures

Fig. 2 A planar 64-microelectrodearray, similar to those used in thedevelopment of the electronic Petri dish,with a series of connectors allowingelectrical signals to be obtained fromthe central cellular playground. Thedevice was produced using the methodsshown in Fig. 1 in order to developstructures of the type shown in Fig. 3.

2.3 Fabrication of Planar Microelectrodes Using Photolithography 39

into plastics [10]; or incorporating sensorswithin a low volume chamber made fromthick (40 µm) photoactive polymers [11,13, 14]. In the latter case, the addi-tional fabrication step of using a thickfilm layer, such as polydimethylsiloxane(PDMS) or SU8 [13], has enabled pla-nar microelectrodes to be integrated intomicromachined chambers in order to per-form a variety of single cell analyses. Forexample, such devices with volumes inthe range of 100 pL to 1 nL, have enableddetailed single heart cell analyses to be per-formed [13, 14]. In this work, the principlehas been to use a cascade of solution-phaseenzyme-linked assays to determine femto-mole amounts of the purines, includingadenosine and inosine [11, 14], and lac-tate [13], Fig. 3. More detailed studies caninvolve introduction of chemicals to af-fect the disruption of the cell membraneusing detergents (to measure intracellularconcentrations), or the addition of respi-ratory uncouplers, in order to simulateischemia. Typically, the cell measurementdevice, Fig. 2, with detail in Fig. 3, will bemounted on a microscopic setup to enablemeasurements, Fig. 4.

2.3.2Advantages of Using Three-dimensionalMicrostructures

Apart from providing new formats, tomake novel measurements, such minia-turized devices offer additional electro-analytical advantages. For example, tra-ditionally, one perceived disadvantage ofsingle cell analysis has been the diffi-culty of mitigating for biological variation,resulting in the need to repeat the mea-surement on many different individualcells. Micromachining and microfabrica-tion offer the possibility of developingarray technologies and the ability to col-lect data from multiple single cell sens-ing sites. More systematic advantages,which occur directly from miniaturizationwithin a three-dimensional chamber in-clude greatly reduced amounts of materialused in each assay (with reduction in costs)as well as the ability to constrain the vol-ume surrounding the cell. In this lattercase, although the flux of ions or metabo-lites from single cells may be small, thediffusion distance to the electrode is short(10–100 µm). The result is that responses

Fig. 3 A single heart cell within a picoliter-scalemicrochamber, see Refs. [11, 13, 14] for moredetails, which has been used for the ultrasensitiveamperometric measurement of purines (the cell isca. 120 µm in length). The view is a plane, with thethree electrodes of the amperometric device,fabricated on a glass microscope slide. Thephotograph shows clearly the large outer counterelectrode, the inner, concentric working electrode (atwhich the active sensing process occurs), and finallythe small (white) Ag/AgCl reference electrode,against which the working potential is measured.The cell is placed by micropipette into the chamber.The device operates as a miniaturized Clarkelectrode, measuring hydrogen peroxide producedas a consequence of an enzyme cascade, in solution.Purine and lactate release can be detected in realtime from the single cell in femtomolar amounts [11,13, 14].


Sensor

Microinjectionsystem

Left micro-manipulator

Digital CCDvideo cameraMonitor

Potentiostat

PC

Microscope

Right micro-manipulator

Faradaycage

Fig. 4 High-sensitivity electrical measurements can be made by designing analyticalinstrumentation with suitable electrical shielding close to the single cell measurementchamber. The device shown has an integrated single cell analytical chamber, similar tothe type shown in Fig. 3, together with shielded amplifiers and a PC interface. Thesetup is common to the measurements from a family of single sensors, currentlybeing produced in the author’s laboratories.

are fast (typically less than 10 s). For ex-ample, employing the Einstein equation,where l is distance; D, diffusion coeffi-cient; and t , time:

l = (2Dt)1/2 (1)

It takes only ∼1.9 and ∼7.5 s for oxy-gen and glucose, respectively, to diffuse100 µm within such a device, with noneof the analyte lost to the bulk solution. In-deed, historically, one approach for singlecell measurements has been to enhancemass detection limit by minimizing the an-alyte dilution, a technology that resulted inthe development of the Cartesian divers forthe measurement of oxygen consumptionin pancreatic islet cells [32]. One future ad-vantage of a reduced electrochemical pathlength is that the solution resistance fallscorrespondingly, and the measurements

of fast electrochemical reactions such asheterogeneous or homogeneous reactionbecome feasible.

2.3.3Biosensors, Microfluidics, Microarrays,and Lab-on-a-Chip

The application of enzyme-linked assayswith low-volume devices is analogous tothe development of a microbiosensor, andhas the ability to greatly extend the range ofmetabolites that can be measured from sin-gle cells [13, 14]. A further example of theuse of enzyme assays involves the oxidase-linked enzyme-catalyzed oxidation of glu-cose in the presence of molecular oxygenby the flavoprotein, glucose oxidase [33]from groups of isolated kidney cells. Aselsewhere [13, 14], the measurement pro-cedure involves a Clark type oxygen (or

2.4 Self-referencing Microelectrodes 41

hydrogen peroxide electrode) ‘‘biosensor’’,using an anode (or cathode), poised againsta suitable reference electrode, enabling thedetermination of glucose flux as a functionof distance from a free-standing micro-electrode. Such studies illustrate the widevariety of both oxidase and dehydrogenaseenzymes, which can now be readily linkedto electrochemical assays. This offers theprospect of being able to perform wholecell analyses, using enzyme-linked assays,measuring different metabolites, at differ-ent stages of the cell cycle, a technique thatwas recently named ‘‘metabolomics’’.

With recent advancement of micro-fabrication technologies in semiconduc-tor industry, the construction of three-dimensional devices with electrodes in-corporated within chambers and channelsthat can be used to constrain low vol-umes (typically nanoliters or below), andprovide flow rates of less than micro-liters per minute. This technology, whichhas become synonymous with the termLab-on-a-Chip, enables a variety of ana-lytical measurements, including samplepreparation, chromatographic separation,and detection (including electrochemi-cal analysis) to be performed on thesame device [34]. The technology pro-vides the possibility of fast, cheap, andintegrated analysis within closed fluidicsystems. Fluids or cells can be movedby micropumps, or through a varietyof techniques associated with electro-chemical phenomena, including electroos-mosis and dielectrophoresis, the latterbeing capable of moving fluids and dif-ferent cell types independently of eachother.

As an alternative to the closed flu-idic systems described earlier, open mi-croarrays structures can enable parallelmeasurements in which multiple manip-ulations are required for high throughput

cell-based assays and screens. Under suchcircumstances, it may be more realisticto introduce analytes and/or biologicalreagents by micropipetting, either manu-ally [10, 11, 14] or by using either balancedpressure [13] or piezoelectric ‘‘ink-jet’’ sys-tems [35]. In either case, there is the abilityto dispense volumes of less than 10 pL(although there are significant differencesin throughput). In these open-array experi-ments, in which the volume to be analyzedis <1 nL, there is an associated problembecause of the loss of solvent by evapora-tion [10, 11, 14] (which can rapidly alter theeffective concentration in a microchamberby an order of magnitude). The use ofhumidified environments and the abilityto dispense and aspirate fluids through alayer of mineral oil have both been used toovercome such problems. At the limit ofminiaturization, in order to explore diffu-sion as well as both electronic and chemicalcross talk, 10 independently fabricated pla-nar devices with volumes as low as ∼4 pLhave been used to provide possible in-sights into the limits of three-dimensionalelectroanalytical arrays [36].

2.4Self-referencing Microelectrodes

Cells constantly, either actively or pas-sively, transport chemicals across theplasma membrane and communicate withsurrounding cells, generating chemicaldiffusion gradients around them. Mea-surement of such activities within gra-dients enables a better understanding ofthe cellular transport mechanisms. Forexample, calcium is regulated to nanomo-lar values in the cytosol by a complexinteraction of cellular and plasma mem-brane pumps, reporters, channel activ-ity, regional sequestration, and chemical


buffering. Such changes in ionic flux,although small, can have significant ef-fects in the context of time-dependent driftof a local reference electrode. In orderto overcome this, the technique of self-referencing microelectrodes [15–19] hasbeen developed and applied to a number of

single cell methods. Central to the methodis the need to use translational motioncontrol systems comprising translationstages arranged in an orthogonal arrayand driven by linear stepper motors [37],thereby providing nanometer resolutionof the microelectrode relative to the cell.

Glucose source0.1 or0.2 Hz

Step motor controller

Current amplifier

IonView

2.5 s 2.5 s

a

D S

Near position

Far position

0.2 s

10 or 20 µm

(a)

(b)

Fig. 5 (a) Illustration of the operation of the self-referencing glucose microsensor. Apersonal computer precisely controls the translational square wave frequency of sensorpositioning allowing IonView software to report current changes referred to sensormovements in a gradient. The positioning of the sensor and its motion relative to thetarget site of interest is guided by observation with a video camera and a monitor.(b) Illustration of time-dependent sensor location and data processing in IonViewsoftware. The time corresponds to a translational square wave frequency of 0.2 Hz and anexcursion distance of 10 µm. IonView automatically subtracts current values obtained atfar position from those at near position. Whereas the first half of the values (denoted asD) at each position is discarded to eliminate the artifact from sensor motion and sensorresponse time, the second half (denoted as S) is sampled for mathematical subtraction.The speed of sensor movement is set by angle, α, and optimized to be 50 µm s−1.

2.4 Self-referencing Microelectrodes 43

In the self-referencing mode, the micro-electrode is laterally oscillated betweentwo points 5–20 µm apart at a squarewave frequency of 0.1–0.3 Hz and dataare collected at 100 Hz. The system setupand data processing are illustrated inFig. 5 [15–19]. Data are reported as differ-ences of voltage or current at two positions.The convention of the difference valueis defined as [(V , inear) − (V , ifar)], where(V, inear) and (V, ifar) are the voltages orthe currents at the near and far pole, re-spectively, from the cell. The differencevalues are directly related to fluxes, whichis discussed later.

2.4.1Ion-selective Microelectrodes

Ionic gradients can be measured po-tentiometrically with ion-selective micro-electrodes [37–39] using, for example, aborosilicate capillary of 2–4-µm diam-eter, functionalized with an ionophorecocktail. The self-referencing ion-selectiveelectrode reports differential voltages mea-sured between two known positions. Thisinformation can be used to calculate theflux by employing Fick’s first law, as

J = −D

[Cav10(Cf SV ) − Cav

x

](2)

where S is the inverse of the Nernstslope in (mV)−1, Cav is the averageconcentration between two positions, V

is the difference in sensor response (mV),and x is the distance between the positionof the two measurements. In a steepgradient generated from an artificial ionsource, it is possible to observe a consistentdisparity between concentration ratio andV , which is because of the bandwidthof the system, particularly with regard togradient reestablishment. This disparity

can be compensated for to obtain thevalues consistent with those obtained fromstationary electrodes.

Self-referencing ion-selective electrodeshave seen attractive applications in a diver-sity of medical problems. For example, thevacuolar-type H+-ATPase has been shownto play an important role in the acidi-fication of the lumen of the proximal vasdeferens, part of the male reproductive sys-tem. An acidic luminal fluid is required forthe maintenance of sperm quiescence andfor the prevention of premature activationof acrosomal enzymes during their storagein the epididymis and vas deferens. Protonsecretion in the proximal vas deferens hasbeen measured with the self-referencingtechnique (Fig. 6) [40]. Likewise, alteredpotassium homeostasis is indicative ofdying cells, as the transplasma mem-brane potential is no longer maintained,as measured in both viable and nonviableembryos [41].

2.4.2Self-referencing AmperometricMicroelectrodes

The principle of the self-referencing tech-nique can be easily applied to the measure-ments of redox active molecules. As therelationship between signal and concen-tration is mostly linear with amperometricmicroelectrodes, their fluxes around cellsare calculated as follows:

J = −DCf i

Sx(3)

where i is the difference current be-tween two self-referencing positions andS is the sensitivity of a microelectrodein pA mM−1, whereas other notations arethe same as in Eq. (2). Self-referencingoxygen electrodes have been used formeasurements in single pancreatic HIT


DPC

0

20

40

60

Rel

ativ

e si

gnal

[%]

80

100

120

140

160

Time[min:sec]

SITS

Bafilamycin

00.0

001

.53

03.4

805

.43

07.3

609

.31

12.3

214

.27

16.2

018

.15

20.1

022

.03

23.5

825

.51

27.4

630

.10

32.0

534

.00

35.5

337

.48

39.4

141

.36

44.0

546

.00

47.5

549

.48

51.4

353

.36

Fig. 6 Representative trace showing effects ofdiphenylamine-2-carboxylate (DPC),4-acetamido-4′-isothiocyanostilbene-2,2′-disulfonic acid (SITS), and bafilomycin on rate ofproton secretion. Addition of 0.5 mM DPC hadno effect, and 1 mM SITS strongly inhibited

proton secretion to a level that was not furtherreduced by 1 µM bafilomycin. An initialdisturbance of proton gradient resulted fromaddition of SITS. Signal is expressed relative tocontrol value.

cells [15, 16], shown in Fig. 7, as wellas single neurons and single plantcells [17–19], and individual mouse em-bryos [42]. Most recently, using a sizeexclusion-based carbon fiber microelec-trode, nitric oxide effluxes from singlemacrophages and damaged neural tissuehave been measured [18].

The self-referencing technique can alsobe further expanded to the measurementof electrochemically dormant molecules, ifenzymes are suitably incorporated withinthe microelectrodes, as biosensors. Ox-idase enzymes are of special interestbecause the enzymes produce, in the pres-ence of oxygen, hydrogen peroxide, whichcan be readily detected at an electrode.Under such circumstances, a glucose mi-croelectrode should respond rapidly to

changing glucose concentrations whenoperated in a self-referencing mode [33].Information about glucose consumptionby pancreatic cells is of great medi-cal importance because insulin secretionis metabolically driven in the pancreas,which is responsible for glucose home-ostasis (e.g. using pancreatic HIT cellclusters [44]).

2.5Scanning Probe Microscopy and Single CellMeasurement

The principle of rastering a microsensoracross the surface and collecting high-resolution data is one which is now famil-iar, and is reviewed in detail in Volume X.

2.5 Scanning Probe Microscopy and Single Cell Measurement 45

−10

0

10

20

30

40

0 200 400 600

Time[s]

∆ i

[fA]

11 mM glucose

10 µM rotenone

−80

−70

−60

−50

−40

0

(a)

(b)

100 200 300 400 500

Time[s]

i [

pA]

−2.0

−1.0

0.0

1.0

∆ i

[pA

]

3040

30

30–4040–50

50–6060–70

100–110

i ∆ i

Stationary Self referencing mode

Fig. 7 (a) Real-time trace (i) and difference current (i) of oxygen current as a functionof distance from an oxygen source (diameter ∼10 µm) measured with a stationary and aself-referencing oxygen microsensor. In the self-referencing mode, the oxygen sensor wasoscillated over a distance of 10 µm at a frequency of 0.2 Hz. The numbers are the relativedistances from the oxygen source in microns. (b) Influence of rotenone on oxygenconsumption by a single pancreatic HIT cell. The trace is an average representative of fivemeasurements. The oxygen sensor was positioned 10 µm from the plasma membrane ofthe cell and oscillated in a square wave between 10 and 40 µm at a frequency of 0.2 Hz.Addition of rotenone significantly inhibited the mitochondrial respiration. A potential of−600 mV versus Ag/AgCl was applied to the oxygen sensor in a modifiedHEPES (4-(2-hydroxyethyl)-1-piperazine-ethane-sulphonic acid) buffer at 37 C. (Datafrom Ref. [43].)

Information concerning topology, heat,light, or the nature of the ionic speciespresent can be readily obtained. The char-acterization and imaging of single cellswith electrochemistry, known as SECMhas recently been reviewed extensively [45]in order to explore the application ofnanodes (or ultramicroelectrodes) in mi-croscopy. In addition, there have beena variety of references to work that has

looked at model systems, to help interpretthe results [46]. The technique of SECMis able to explore the flux and hence, con-centration of a redox species by exploringthe oxidation or the reduction current at afixed distance (z) above the source (in thiscase, the cell or cells). The electrode is thenscanned in the x and y directions in orderto obtain a chemical image. Clearly addi-tional data can be obtained by repeating


the scan while changing the distance (z)above the redox source. A variety of studieshave looked at redox changes according ei-ther to their consumption of oxygen, wherethe tip acts as a Clark electrode; for exam-ple, Ref. [47], where it is possible to relatethe consumption of oxygen with individualcell size, and cell growth within an embryo.Such work can be readily adapted and ex-panded in order to include mediators thatare able to probe the transmembrane po-tential of bacteria [48].

In an alternative iteration of SECM, scan-ning ion-conductance microscopy (SICM)has also been used in order to image liv-ing cells in aqueous environments, usinga microcapillary in order to probe ion flux[43, 49–51]. In such experiments, it is pos-sible to gather information on the topologyof the cell (as a function of its distance, z)from the microcapillary, leading to veryhigh z-spatial resolution of the mousemelanocyte [43]. Such a technique read-ily leads to the possible estimation of thevolume of individual cells including kid-ney, heart muscle, and cancerous breastcells [49]. In addition, small cellular struc-tures, such as lamelopodia, dendrites, andmicrovilli have been resolved. The tech-nique can readily be corroborated [50] byits use with simultaneous high-resolutionoptical microscopy involving scanned nearfield optical microscopy, SNOM (that op-erates as evanescent field imaging, belowthe limits of the diffraction limitation).Under such circumstances, it is possibleto sarcomeric striations of the cardiomy-ocyte, indicating micron-scale resolutionof subcellular organelles. This techniquehas recently been extended, such that whena single myocyte was patch-clamped [1–3]and imaged using SICM, it was possible toimage a single ion channel, within a singlecell [51].

2.6Semiconductor Devices for BiologicalSensing

In addition to using free-standing orscanned ion-selective electrodes basedupon capillaries for ion measurements,as described earlier, photolithographictechniques have been used to adapt sil-icon devices (such as field-effect tran-sistors) for use in biological measure-ments [9]. The techniques differs fromthose described for three-dimensional mi-cromachining, in so much as silicon isbeing used as an active sensor. Thevast majority of such devices are basedupon the field-effect transistor, whichare subsequently adapted as either ion-selective field-effect transistors (ISFETs)and light-addressable potentiometric sen-sors (LAPS). They are used to mea-sure extracellular ion fluxes either incell culture systems or clinical diagnos-tic tests [52–54]. Both ISFET and LAPStechnologies have their strengths in thegeneric nature of their measurement sys-tems (which also enable them to be usedin the wider context of chemical imag-ing) and in the compatibility of theirfabrication methods with industry stan-dards. Devices used for cell screeningeither involve simple measurement ofions, or in a rather more complex man-ner, by measuring metabolic condition asa function of proton flux in LAPS tech-nology [52–54]. As an alternative method,adapted metal oxide semiconductor field-effect transistors (MOSFETs), with thegate removed, have been used as amethod to probe signaling in a varietyof single cells, including both inverte-brate and mammalian systems [55–58].In a series of elegant experiments ithas proved possible to use capacitativemeasurements to probe changes in the

2.8 Future Prospects 47

cell’s charged state, providing a methodnot only to stimulate but also collect in-formation across real (in vitro) neuralnetworks.

2.7Corroborative Measurements for Single CellElectrochemistry

In many cases, there is a need or a de-sire to be able to corroborate the singlecell electrochemical measurement by useof an associated technique that can pro-duce independent biological information.In some cases, this may be simple di-rect observation, so that events such ascell death or rigor (e.g. in the heart cell)that can be related directly to observedmetabolic events [11, 13, 14]. Alternatively,fluorescence measurements involving theuse of an optical fiber have been used forthe measurement of ions, while simul-taneous electrochemical measurementshave greatly enhanced the biological in-formation [28]. The marriage of oxygenor glucose microelectrodes to calciumfluorescence measurement in pancreaticislets has been observed [33]. The self-referencing electrochemical technique canalso be linked to fluorescence measure-ments or used in conjunction with patchclamp [55–58]. Alternatively, the voltageclamping of a cell has been combinedSICM to provide evidence for the spa-tial location of individual ion channelsin the cell membrane, while the cell isheld under electrochemical control [51].Sub-single cell information has also beenresolved using either SNOM [50] andsubsequently related to electrochemicalmeasurements of comparable resolution.Likewise, SECM data has been closely re-lated to the topology of the observed cellstructures [45].

2.8Future Prospects

The further miniaturization of the elec-troanalytical sensors within chambers isinevitable. In future assays volumes of10–500 pL will become more routine, andultimately, the volume in an analyticalchamber may be reduced sufficiently suchthat it may be possible to implementthe application of single molecule elec-trochemical detection [59, 60] within thesestructures. Reduction in the size of theanalytical chamber [11, 13, 14] and intro-duction of a variety of self-referencing tech-niques [15, 16, 18] have made it possible tomake reliable single cell measurements,which carry forward the basic conceptsfrom patch clamping [1–3], but enable avariety of metabolites to be measured. Vol-umes have been greatly reduced [36, 61]in such measurement systems, therebyenabling increasingly sensitive absolutemeasurements to be made. By reduc-ing the time base between measurementpulses within the microsecond regime, itwill be possible in the future to constrainthe effective diffusion length within the so-lution, and thereby make measurementsin volumes as small as 20 fL [62].

Coupled to the development of smallersensors using nanotechnology, with in-creased functionality using biosensors,and with the development of methods forconstraining lower volumes, the presentchallenge is the integration of fast, high-bandwidth, high-sensitivity instrumenta-tion close to the sensors, for parallelrecording from single cells. In this re-spect silicon may provide one possibleroute forward [55–58]. Undoubtedly, inthe future, the greatest challenges will in-volve the development of techniques thatcombine miniaturization and molecularcounting with integration of microfluidics


and biosensor technology and hence theautomation of ultrasensitive single cellassays. Such technology would have sig-nificant impact in drug discovery andbiotechnology as a whole, in providinga method to probe metabolic profiles ofsingle cells in a high throughput format.

References

1. O. P. Hammill, A. Marty, E. Neher et al.,Pflugers Arch. 1981, 391, 85–100.

2. B. S. Khakh, X. R. Bao, C. Labarca et al., Nat.Neurosci. 1999, 2, 322–330.

3. B. Sakmann, E. Neher, Annu. Rev. Physiol.1984, 46, 455–472.

4. D. M. Cannon Jr., N. Winograd, A. G. Ewing,Annu. Rev. Biophys. Biomol. Struct. 2000, 29,239–263.

5. K. McCormack, R. Davies, Pain 1996, 68,5–11.

6. C. Kranz, G. Wittstock, H. Wohlschlageret al., Electrochim. Acta 1997, 42, 3105–3111.

7. R. M. Wightman, E. Strope, P. M. Plotskyet al., Nature 1976, 262, 145–147.

8. R. M. Wightman, Anal. Chem. 1981, 53,A1125–A1134.

9. M. Lambrechts, W. Sansen, Biosensors: Mi-croelectrochemical Devices, Institute of PhysicsPublications, New York.

10. R. A. Clark, P. B. Hietpas, A. G. Ewing, Anal.Chem. 1997, 69, 259.

11. C. D. T. Bratten, P. H. Cobbold, J. M. Coo-per, Anal. Chem. 1997, 69, 253–258.

12. D. O. Wipf, A. J. Bard, Anal. Chem. 1992, 64,1362–1367.

13. X. Cai et al., Anal. Chem. 2002; accepted.14. C. D. T. Bratten, P. H. Cobbold, J. M. Coo-

per, Anal. Chem. 1998, 70, 1164–1170.15. S. -K. Jung, K. Hammar, P. J. S. Smith, Biol.

Bull. 2000, 199, 197–198.16. D. M. Porterfield, R. F. Corkey, R. H. Sanger

et al., Diabetes 2000, 49, 1511–1516.17. D. M. Porterfield, J. D. Laskin, S. -K. Jung

et al., Am. J. Physiol. 2001; in press.18. S. C. Land, D. M. Porterfield, R. H. Sanger

et al., J. Exp. Biol. 1999, 202, 211–218.19. S. M. Kumar, D. M. Porterfield, K. J. Muller

et al., J. Neurosci. 2001, 21, 215–220.20. C. Amatore, S. Arbault, D. Bruce et al., Fara-

day Discuss. 2000, 116, 319–333.

21. R. M. Wightman, D. O. Wipf in Electroan-alytical Chemistry (Ed.: A. J. Bard), Mar-cel Dekker, New York, 1989, pp. 267–353,Vol. 15.

22. S. Arbault, P. Pantano, J. A. Jankowski et al.,Anal. Chem. 1995, 67, 3382.

23. Y. Y. Lau, T. Abe, A. G. Ewing, Anal. Chem.1992, 64, 1702–1705.

24. G. Chen, D. A. Gutman, S. E. Zerby et al.,Brain Res. 1996, 733, 119–124.

25. Q. Xin, R. M. Wightman, Brain Res. 1997,776, 126–132.

26. M. A. Bunin, C. Prioleau, R. B. Mailmanet al., J. Neurochem. 1998, 70, 1077–1087.

27. P. A. Garris, J. R. C. Christensen, G. V. Re-bec et al., J. Neurochem. 1997, 68, 152–161.

28. Q. Xin, R. M. Wightman, Anal. Chem. 1998,70, 1677–1681.

29. C. D. Paras, R. T. Kennedy, Electroanalysis1997, 9, 203–208.

30. C. A. Aspinwall, L. Huang, J. R. T. Lakeyet al., Anal. Chem. 1999, 71, 5551–5556.

31. G. T. A. Kovacs, K. Petersen, M. Albin, Anal.Chem. 1996, 68, 407A–412A.

32. C. Hellerstrom, Endocrinology 1967, 81,105–112.

33. S.-K. Jung, L. M. Kauri, W.-J. Qian et al., J.Biol. Chem. 2000, 275, 6642–6650.

34. J. Harrison et al., Anal. Chem., submitted.35. A. V. Lemmo, J. T. Fisher, H. M. Geysen

et al., Anal. Chem. 1997, 69, 543–551.36. P. Yu, G. S. Wilson, J. Chem. Soc., Faraday

Discuss. 2000, 116, 305–317.37. P. J. S. Smith, K. Hammar, D. M. Porterfield

et al., Microsc. Res. Tech. 1999, 46, 398–417.38. P. J. S. Smith, J. R. Trimarchi, Am. J. Physiol.

Cell Physiol. 2001, 280, C1–C11.39. P. J. S. Smith, R. H. Sanger, L. F. Jaffe, Meth-

ods in Cell Biology, Academic Press, NewYork, 1994, pp. 115–134, Vol. 40.

40. S. Breton, K. Hammar, P. J. S. Smith et al.,Am. J. Physiol. 1998, 275, C1134.

41. J. R. Trimarchi, L. Liu, D. M. Porterfield et al.,Zygote 2000, 8, 15–24.

42. J. R. Trimarchi, L. Liu, D. M. Porterfield et al.,Biol. Reprod. 2000, 62, 1866–1874.

43. Y. E. Korchev, M. Milovanovic, C. L. Bash-ford et al., J. Microsc. 1997, 188, 17–23.

44. S.-K. Jung, R. H. Sanger, P. J. S. Smith et al.,Anal. Chem. 2001; in press.

45. T. Yasukawa, T. Kaya, T. Matsue, Electroanal-ysis 2000, 12, 653–659.

46. C. A. Wijayawardhana, G. Wittstock, H. B.Halsall et al., Anal. Chem. 2000, 72, 333–338.

2.8 Future Prospects 49

47. H. Shiku, T. Shiraishi, H. Ohaya et al., Anal.Chem. 2001, 73, 3751–3758.

48. C. Cai, B. Liu, M. V. Mirkin et al., Anal.Chem. 2001; accepted in press.

49. Y. E. Korchev, J. Gorelik, M. Lab et al., Bio-phys. J. 2000, 78, 451–457.

50. Y. E. Korchev, M. Raval, M. Lab et al., Bio-phys. J. 2000, 78, 2675–2679.

51. Y. E. Korchev, Y. A. Negulyaev, C. R. W. Ed-wards et al., Nat. Cell Biol. 2000, 2, 616–619.

52. P. Skladal, Electroanalysis 1997, 9, 737–745.53. A. Fanigliulo, P. Accossato, M. Adami et al.,

Sens. Actuators, B 1996, 32, 41–48.54. H. Uchida, W. Y. Zhang, T. Katsube, Sens.

Actuators, B 1996, 34, 446–449.55. P. Fromhertz, A. Offenhauser, T. Vetter et al.,

Science 1991, 252, 1290, 1293.56. D. Braun, P. Fromhertz, Phys. Rev. Lett. 2001,

86, 2905–2908.57. G. Zeck, P. Fromhertz, Proc. Natl. Acad. Sci.

U.S.A. 2001, 98, 10 457–10 462.

58. S. Vassanelli, D. Fromhertz, J. Neurosci.1999, 19, 6767–6773.

59. F. F. Fan, J. Kwak, A. J. Bard, J. Am. Chem.Soc. 1996, 118, 9669–9675.

60. F. F. Fan, A. J. Bard, Science 1995, 267,871–874.

61. O. Shirihai, P. Smith, K. Hammar et al., Glia1998, 23, 339–348.

62. M. M. Collison, R. M. Wightman, Science1995, 268, 1883–1885.

63. Y. T. Kim, D. M. Scarnulis, A. G. Ewing,Anal. Chem. 1986, 58, 1782–1786.

64. R. Nuccitelli, Vibrating probe technique forstudies of ion transport in NoninvasiveTechniques in Cell Biology (Eds.: J. K. Foskett,S. Grinstein), Wiley-Liss, New York, 1990,pp. 273–310.

65. P. J. Kinlen, V. Menon, W. Ding, J. Elec-trochem. Soc. 1999, 146, 3690–3695.

66. C. A. Siedlecki, R. E. Marchant, Biomaterials1998, 19, 441–454.

51

3Bioelectronics

Christiane ZieglerDepartment of Physics, University of Kaiserslautern, Kaiserslautern, Germany

3.1 Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

3.2 Brief Summary of Electronic Signal Transduction in the Neural System 53

3.3 Biology Meets Electronics: Applications . . . . . . . . . . . . . . . . . . . . 573.3.1 Bioelectronic Noses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 573.3.1.1 Comparison between Biological and Electronic Noses . . . . . . . . . . 573.3.1.2 Specific Example: Neural Network Biosensors . . . . . . . . . . . . . . . 593.3.2 Biocomputer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 643.3.3 Short Summary of In Vivo Applications . . . . . . . . . . . . . . . . . . . . 65

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

53

3.1Scope

No overall accepted definition for bioelec-tronics exists so far. In Ref. [1], one canfind the following list of topics that may berelated to bioelectronics:

• Biological materials for electronics;• Biological-inorganic hybrids of rele-

vance for electronics;• Biocompatible electronic devices such

as pacemakers and other electronicimplants;

• In vivo sensors and other sensors forbiologically relevant processes;

• Sensors based on biological materials orbiological processes;

• Artificial sensing devices such as eyes,ears, noses, and so on;

• Materials for electronics synthesizedby biological processes rather than byconventional synthesis;

• Materials inspired by biology and usefulfor electronics;

• Algorithms inspired by biology;• . . . .

Often, only the first two – and some-times, additionally, biosensors based onnerve cells or nerve cell tissues – are in-cluded in the definition. In this shortreview article, bioelectronics will be re-stricted to in vitro electronic devices

in which neural cells are involved asbiological component, that is, the cou-pling of an electronically active biolog-ical system to an artificial electronics.After a short summary of the basicsof electronic signal transduction in thenervous system (Sect. 3.2), applicationssuch as natural neural network sen-sors and computer devices (Sect. 3.3) arediscussed.

3.2Brief Summary of Electronic SignalTransduction in the Neural System

Signals from the outer world are perceivedby animals through the nervous system.Receptor cells in the sensing organs (eye,ear, nose, tongue, skin) convert outside sig-nals (i.e. light, noise, odorant molecules,taste, pressure, temperature) into electri-cal signals, which are transported by neuralcells to the brain (sensory system). Fromthe brain, electrical signals are then trans-mitted to muscle cells through the motornervous system.

The nervous system is divided into twoparts: the central nervous system (CNS),which comprises the brain and the spinalcord, and the peripheral nervous system(PNS), which includes the rest of thenervous tissue.

54 3 Bioelectronics

Neural cells consist of four distinctregions: the cell body, the dendrites,the axon, and the synapses (Fig. 1). Thecell body is the metabolic center of thecell, that is, it keeps the cell living

and functioning. The dendrites are smallbranches extending from the cell bodyand are the main receptive part of thecell. The axon is a long tube (up to1 m in humans), sometimes covered by

Dendrite

Terminal

Axon

Pre

syna

ptic

cel

lP

osts

ynap

tic c

ells

Node of Ranvier

Axon hillock

Basal dendritesAxon (initialsegment)

Excitatory terminalfiber of an axon

Inhibitory terminalfiber of an axon

Myelin sheath

Presynaptic terminal

Synaptic cleft

Postsynaptic dendrite

Apical dendrites

NucleusPerikaryon Cell body

Fig. 1 Schematic view of the morphology of a typical neuron and its functional regions [2].

3.2 Brief Summary of Electronic Signal Transduction in the Neural System 55

an insulating myelin sheath, which isregularly interrupted by the so-callednodes of Ranvier in which almost all ofthe ion channels (see the following text)are located. The synapses extend from theend of the axon and form the connectionsto other nerve or to muscle cells. Inthe membrane of the neural cell, thereare proteins forming membrane-spanningion channels for Na+, K+, Ca2+, andCl−, which are either voltage- or ligand-gated. Voltage-gated means that changesin potentials over the membrane causeconformational changes in the proteinstructure and hence to an opening orclosing of an inner transport channel(pore). Ligand-gated channels do the sameupon binding a molecule, the ligand, to itsreceptor site.

The ion concentrations inside and out-side the cell are vastly different for thesingle species. The K+ concentration ismuch higher inside the cell, whereas theNa+ concentration is much lower. Al-though there is electroneutrality insidethe cell as well as outside, these differentconcentrations and also the different per-meabilities through the membrane leadto an electrochemical potential across themembrane. For an exact calculation of themembrane potential, the concentrationsand permeabilities of all ions as well as thediffusion currents have to be taken intoaccount. The resting potential (i.e. whenall channels are closed and hence no sig-nal is transported along the axon) is −60to −70 mV between the inner and outercell compartment. This potential is nearthe resting potential for K+ alone, that is,without taking the potentials of the otherions into account, because of the highpermeability of the membrane for K+.

Because of a signal, from the outer worldor from another nerve cell, (see in follow-ing text) a so-called action potential can

be initiated at the axon hill near the cellbody (Fig. 2). During this action poten-tial, the membrane potential is switched toaround +30 to +40 mV for a short time(1 ms). This is generated because voltage-gated Na+ channels (see in following text)open for a short time and therefore pos-itive sodium ions can flow into the cell(along its concentration gradient), whichraises the membrane potential. Becauseof this voltage change, the Na+ chan-nels close again and K+ channels openup, which cause potassium to flow out ofthe cell (along the concentration gradientfor K+), which restores the negative in-ner potential. This action potential acts onother voltage-gated Na+ channels locatedfurther down the axon where the next ac-tion potentials are generated in the sameway with a frequency of about 1 kHz. Themyelin sheath serves as an electrical isola-tion and prevents the potential from dyingout due to membrane leakage. Therefore,in particular, long axons are covered by amyelin sheath. As soon as the action po-tential reaches the synapses, voltage-gatedCa2+ channels are opened. This leads tothe excretion of membrane vesicles con-taining small chemical substances, theneurotransmitters, into the synaptic cleftbetween the (presynaptic) nerve cell andthe next (postsynaptic) cell. These neu-rotransmitters diffuse through the cleftand bind to receptors that are usually lo-cated on the dendrites of the postsynapticcell. Ligand-gated Na+ channels open andlead to a membrane potential. If severalof such channels are open, the mem-brane potential can become large enoughto generate an action potential throughopening of voltage-gated Na+ channels(see preceding text). There are not onlyexcitatory synapses, as described above,but also inhibitory ones that do not trans-mit, but block action potentials between

56 3 Bioelectronics

Modelneuron

Secretion

Excitatorypostsynapticpotential

Inhibitorypostsynapticpotential

Grade inputpotential

Action potentialat trigger zone

Conductedaction potential

Secretorypotential

Fig. 2 Signals arising along the functional regions of a neuron [2].

two cells through the generation of anegative membrane potential. This inhibi-tion is important to filter the informationflooding between our sensing organs andour brain. All information coming fromthe different synapses is integrated bythe cell. This means that there are twotypes of electrical signals by which in-formation can be coded and transported:There are graduated membrane potentialsbuilding up near the synapses or at re-ceptor cells. Graduated means that theycan have any potential, which results froman addition of all positive (excitatory) ornegative (inhibitory) potentials formed atthe synapses. If a certain threshold po-tential is reached, the voltage-gated Na+channels at the axon hill can open andan action potential is initiated. The ac-tion potentials themselves are relativelystereotypic, that is, they exhibit always thesame shape. Information is only coded

in the number and frequency of actionpotentials.

Details of signal transduction by neuralcells can be found in a large variety oftextbooks [2].

Nerve cells do not divide, and cutinterconnections between nerve cells inthe CNS are not able to regenerate, whichleads to dysfunctions after brain or spinalcord injuries. This makes it also difficultto utilize these cells in technical devicesbecause no cell lines can be established.This means that only primary cell cultures,that is, cells that come directly from aliving animal, can be used. There aresome neuron-like cell lines in whichhybridoma cells between a type of cancercells and a neural cell are used. Cancercells divide easily and hence cell lines canbe established. However, these hybridomacell lines do not easily form synapticjunctions. Because without synapses there

3.3 Biology Meets Electronics: Applications 57

is no signal transduction between cells,such single cells are limited in their use inbioelectronic applications. All applicationsthat are referred to in Sect. 3.3 are thereforebased on primary cell cultures.

3.3Biology Meets Electronics: Applications

3.3.1Bioelectronic Noses

3.3.1.1 Comparison between Biologicaland Electronic NosesOur sense of smell is able to recognizeand discriminate, with great sensitivity andaccuracy, thousands of volatile moleculesof diverse structure. Odorant stimuli atconcentrations as low as a few partsper trillion can be detected. The generalsensation of odors in biology, however,not only includes human odor sensationbut also the odor sensation of animalsin air and in water with their completelydifferent sensitivity and selectivity patternsto detect chemical species in the gas andliquid state. Since we do not know detailsof odor perception in the various animals,the most general definition of a noseis as a detection system to sense anymolecule in the gas or liquid state. Theodor patterns generated by such odorant-detector systems evidently depend on theirbiological and biochemical architecture(see schematic Fig. 3a). These also dependon the individual training, learning, andpreconditioning of the systems, on theenvironment, and so on.

In the first step of odor sensation, thegaseous analytes ©1 , which are typicallysmall, mostly hydrophobic molecules withmasses up to 300 amu, reach the olfactoryregion at the posterior of the nose.Sampling ©2 is done by inhalation with

different speeds. Dust and other particlesare filtered ©3 in the nostrils. A secondstep of filtering and sampling is performedby the mucous membrane, which iscovered with a thin film of water. Thehydrophobic odorants have to pass thisbarrier, which is probably promoted by so-called odorant binding proteins to reachthe surface of an olfactory signal cellin the epithelium. The interaction withthe odorant binding proteins may alsolead to some preconditioning ©4 , becausedifferent interactions may occur. Thefollowing steps all happen on the surfaceor in the olfactory cell. The odorantmolecules will interact specifically with(in humans) about 800 different receptorproteins embedded in the membrane ofthe 6 million neurons. This molecularrecognition can be depicted as the essentialstep of sensing ©5 . The binding of theodorants triggers complex intracellularsignal cascades, which can lead to actionpotentials. This step can be described asa transduction ©5 of the former chemicalinformation into an ionic, that is, anelectrical information. In the further datatreatment, the signals coming from thishuge number of different olfactory cellsmust be guided by their axons to theso-called glomeruli in which a largenumber of axons meet. This is a kindof signal pretreatment ©6 , because thesignals coming from one type of receptorsare sent only to one (or two) of thedifferent glomeruli. Or, in other words,axons of one type of receptor meet inone glomerulus. (This means that thenumber of glomeruli reflects the numberof different receptors in a specific animal.)The feature extraction ©8 is done by abackpropagation from the human brain byspecial centrifugal fibers, manipulating allsteps of signal processing. The comparisonwith memorized data ©7 , for example, with

58 3 Bioelectronics

Mucous layer with(?) odorant binding

proteins (OBP)Brain

Backpropagation

Odor

Ana

lyte

s

Axons, glomeruli, mitral, and granule cellsReceptors R with ion

channel (Na+) and amplifi-cation cascade (G, AC, ...)

Pat

tern

rec

ogni

tion

Res

ults

Sam

plin

g

F

ilter

ing

Pre

cond

ition

ing

Sen

sors

and

tran

sduc

ers

Ioni

cs a

nd s

igna

l

p

retr

eatm

ent

Feat

ure

ext

ract

ion

Com

paris

on w

ith

m

emor

ized

dat

a,

m

emor

y, n

ew in

put

Triggeredion channel

Receptorsignal "i"

Water

Ri

cAMP

PKAATP

GDPGTP

AC

G

GRj

Na+

Epithelium

Ana

lyte

s

Sam

plin

gF

ilter

ing

Sen

sors

and

tran

sduc

ers

Ele

ctro

nics

and

dat

a pr

etre

atm

ent

Feat

ure

ext

ract

ion

Pat

tern

rec

ogni

tion

Res

ults

Pre

cond

ition

ing

- Specific molecules- Combustible gases- Air quality- Odor characterization- Food identification- Toxic / healthy- Ok / not ok- ...

Analytical chemistry dataHuman odor panelsBiological function testsProcess parameter tests...

7 Comparison with calibration data:

(a)

(b)

1 2 3 4 5 6 8 7 9 10

1 2 3 4 5 6 8 9 10

Fig. 3 (a) Schematic representation of the signal cascade in the human nose, which is involved inrecognizing odor molecules (adapted from Ref. [3]). (b) Schematic representation of electronic nosesystems, which are realized by modular sensor systems (adapted from Ref. [3]).

previously obtained odor sensations, andthe pattern recognition ©9 is done in thebrain. This leads to a perception of theodor as the final result ©10 , which should

be related directly to the molecules at thebeginning of the odor sensation process.(For a more detailed review, see Ref. [3]and references therein.)


An ‘‘electronic nose’’ is a technicalchemical sensor system, which showsthe same 10 principal steps of odorrecognition (Fig. 3b). However, althoughgeneral similarities may be seen, anymore detailed view illustrates drasticdifferences between the biological andthe technical system (see Refs. [4, 5] andreferences therein).

The most important part is the singlesensor element, which consists of a chem-ically sensitive coating enabling molecularrecognition and of transducers enablingsignal conversion. The specific recogni-tion site on a certain transducer is thekey component of the total sensor system.Evidently, a large amount of different ma-terials with different recognition sites canbe utilized already today for chemical andbiochemical sensor systems. The simplerthe structure of these materials is, the eas-ier it is to optimize their long-term stability.The design of bioelectronic noses, whichutilize biological function units, is evi-dently more complex. Usually, their stabil-ities are limited and hence time-dependentsignal outputs cause serious problems incalibration procedures. As a result, corre-sponding systems are not yet commerciallyavailable. Significant progress is expectedby screening systematically new materi-als with stabilized biological functions thatmay be either biomimetic recognition sitesor biological systems and by screening sys-tematically new materials, which interfacethe biological function units with inor-ganic transducer substrates. So far it is notclear whether biomimetic structures withtheir enhanced stability will be superiorto the evolutionary optimized biologicalstructures that, however, rely on self-repairof their inherently unstable molecules.

Commonly used transducers monitorchemical compositions by monitoring phe-nomenological sensor properties such as

changes in mass, conductivity, capacity,optical parameters, or temperature. Ex-amples include interdigital structures (forcomplex impedance measurements), ther-mopiles (for temperature measurements),piezoelectric oscillators based upon bulk,surface, or plate waves (for mass mea-surements), optics components (for opti-cal measurements), multielectrode arrays(for electrochemical measurements or forelectrical connections to axons in nervesystems), or arrays of capacitive sen-sor elements.

Table 1 illustrates, very schematically,the hierarchy in the description of bioelec-tronic noses with an increasing complexityconcerning structures and functions. Abioelectronic nose will not compete withour human nose but will fulfill dissimi-lar functions: The detection of moleculesthat are toxic, of molecules that do notsmell, and, most importantly, an odor de-tection that is independent of individualinfluences on a living being. On the otherhand, a bioelectronic nose (if compared toan electronic nose with simple inorganicor organic recognition structures) can takeadvantage of some of the properties of thebiological system if it utilizes living cells,in particular the amplification of smallsignals and self-repair mechanisms. Thismakes the use of cells superior to theuse of single biomolecules, however, atthe expense of an enhanced complexity ofthe system. In the following section, onespecific example will be shown, which isa bioelectronic system even in a narrowdefinition of bioelectronics (cf. Sect. 3.1).

3.3.1.2 Specific Example: Neural NetworkBiosensorsA real biohybrid nose would consist ofodorant receptor cells coupled to an elec-tronic read-out unit. However, owing tothe difficulties in establishing highly stable

60 3 Bioelectronics

Tab. 1 Molecular hierarchy in the complexity ofstructures and functions, which may be used inthe design of bioelectronic noses. The respectiveparts of the biological nose are given in italics

• inorganic structures and organic molecules• biomimetic recognition sites• biological recognition sites

odorant binding proteinsolfactory receptors

• recognition sites embedded in biologicalmembranes

olfactory receptors in cilia• recognition sites in membranes with

subsequent signal amplification• whole cells

olfactory cells• cell arrays• neural tissue

olfactory mucosa• neural tissue with subsequent signal

processing stepolfactory mucosa and olfactory bulb

• brainolfactory mucosa, olfactory bulb, and olfactory

cortex• animals• humans

odor sensation of distinguished test persons

odorant receptor cell cultures, nonsensorynerve cells have to be used today as an alter-native. Although, with some exceptions, nospecific odor response can be expected bysuch networks, also nonsensory neuronsare very sensitive to a large variety of neu-roactive compounds added to the culturemedium [6]. Neuroactive compounds canbe defined as water-soluble molecules thatcan influence the sensitive electrophys-iological mechanisms of nerve cells [7].Neuroactivity can have a variety of reasonsbecause all the steps of action potentialgeneration and transduction as describedin Sect. 3.2 may be influenced. Further-more, changes in the metabolism of thecell, which are brought about not onlyby chemicals but also physical parameters

like the temperature, can cause changes inthe electrical activity of the neuron. Manyinteresting neuroactive compounds eitheract as ion channel blocker or block re-ceptors in the synaptic cleft, mimickingneurotransmitters.

Nerve cells from the CNS of mammalianembryos grown in culture on micro-electrode arrays [8–11] or on field-effecttransistors [12] show spontaneous nativeactivity patterns (i.e. action potential gen-eration) after several days in culture. Theseneurons from embryos still have the abilityto form synaptic connections (cf. Sect. 3.2)and can form nerve cell networks.

The effects of an altered environmentcan be studied by detecting changes in thespontaneous native action potential pat-terns. These changes are often substanceand concentration specific (see below) and,most importantly, histiotypic.

As in nature, networks are relatively faulttolerant concerning, for example, changesin synaptic connections. All these effectscan be measured by the change of mem-brane potential during an action potential(cf. Sect. 3.2). This potential has a directinfluence on the gate of a field-effect tran-sistor, or, in another device, it influencesthe capacity between a microelectrode andthe axon, which can be measured witha.c.-coupled amplifiers with high inputimpedances. All measurement conditionshave to be chosen so that no electrochem-ical reaction takes place at the electrodesurface in order to avoid the formation ofpoisoning chemicals.

The contact between the cells and theseelectrodes is one of the very crucial pointsin such devices. Neural cells do not forma direct contact to metals or to silicon.These technical surfaces have thereforeto be coated with a biocompatible over-layer. There are different concepts that in-clude positive charges provided by amino


silanes [13], artificial amino acids such aspoly-D-lysin, or biological molecules suchas laminin, which is an anchor proteinin the extracellular matrix of cells. Be-cause a tight contact of the cell to theelectrode is advantageous as it minimizessignal loss into the electrolyte, new con-cepts aim at using only the short activeamino acid sequences, the so-called epi-topes, of laminin [14, 15] or even simplermolecules [16]. Such molecules can beused as derivatives with an electropoly-merizable group such as hydroxyphenylor pyrrol. In this case, such adhesionmolecules can be immobilized directly atthe electrode surfaces by applying a po-tential, therefore restricting the nerve cellsonto the regions of measurement. How-ever, there is still a huge amount of workto be done because each type of cell mayreact differently to the artificial environ-ment so that no general concept for neuralcell adhesion is at hand today. And restrict-ing the adhesion area is, in most cases, notstable over a longer period, that is, the cellsovergrow other regions of the substrate.

As it is now possible to maintain neuralnetworks in electrophysiologically activeand pharmacologically responsive statesfor over 9 months in vitro [17, 18], suchsystems have become reliable candidatesfor the detection and characterization ofbiologically significant effects of a great va-riety of chemical substances. An overviewof tests performed till 1997 is given inRefs. [7, 18], a recent example (testing ofcannabinoid agonists) can be found inRef. [19]. Although still at an experimentalstage, instruments for extracellular record-ing are commercially available [20–22].Reported applications are fundamental as-pects of cell networks [23], as well aspharmacological [18, 19] and toxicologicalresponse studies [18, 24], qualitative [7, 8],and even quantitative biosensing purposes

if artificial neural networks (ANNs) areused to evaluate the data from the naturalneural network [25, 26]. Pharmacologicalresponse has also been studied in heartmuscle cells, which give larger signals ascompared to neural cells and are henceeasier to study [27].

Some examples shall illustrate the con-cept of using natural neural networks foran analytical purpose. Figure 4 shows thetoxicological effects of two neurotoxins,trimethyltin chloride (TMTC), and lead ac-etate (PbAc) [18]. The presented data aretwo different variables that can be de-rived from the measured voltage spikesof the network in percent of the valueof the native network without any addedneurochemicals. These variables are thetotal spike activity (corresponding to thenumber of action potentials) and the burstrate (i.e. the number per minute of so-called bursts, in which high spike activityis recorded, for details of signal process-ing, see Refs. [7, 18]). The activity of thenetwork is completely inhibited after addi-tion of 4 µM of TMTC or 2 mM of PbAc,but is restored after washing. This clearlyshows that these chemicals act reversiblyat these concentrations in the nervoussystem. Such studies are of general inter-est when new substances are developed,which can come into contact with thehuman nervous system, such as phar-maceuticals, body care products, but alsopesticides and others.

Figure 4 also shows that taking differentaspects (variables) of the network activityinto account, one may get fingerprints ofthe different substances because they showdifferent concentration dependencies.

The same holds for other neurochem-icals. Bicuculline and strychnine bothblock inhibitory receptors in the synap-tic cleft (see Sect. 3.2). This induces a

62 3 Bioelectronics

0

20

40

60

% N

ativ

e ac

tivity 80

100

120Native

KG86cortex

Total spike production

TMTC

Burst rate

Wash

0.1 0.2 0.3 0.4 0.5 1 2 3 4

TMTC[µM]

0

50

100

% N

ativ

e ac

tivity

150

200Native

KGL16cortex

Total spike production

Burst rate

Wash

0.001 0.01 0.1 0.5 1 2

Lead acetate[mM]

(a)

(b)

Fig. 4 Different recordings from networks of auditory cortex tissue uponapplication of trimethyltin (TMTC) (a) and lead acetate (b). For details, seetext and Ref. [16].

higher network activity (because the in-hibition does not work properly) at lowconcentrations and an oscillatory networkresponse (similar to an epileptic attack)at higher concentrations. However, theoscillation starts at different absolute con-centrations for the two substances and alsothe concentration dependency of the dif-ferent measurable variables is different.Such differences can be evaluated in a

variety of data-processing methods, for ex-ample, in a so-called principal componentanalysis [28]. In the principal componentanalysis, a problem of high dimensional-ity is projected to only two dimensions sothat a visualization of the results is pos-sible. This means here that instead of an11-dimensional coordinate system [11 dif-ferent measurement variables (such as thenumber of action potentials per minute)


Fig. 5 Principal component analysisPCA of data recorded on embryonicmice spinal cord cultures in the nativestate, upon application of 60-µMbicuculline and, after a washing cycle, of25-µM strychnine. PC1,2 are theprincipal components 1 and 2.Classification of data is best if datapoints are close together within onecluster and clusters are well separated.For details, see text and Ref. [25]. −4

−3

2

1

0

−1

−2

−3 −2 −1 0 1 2 3 4 5 6 7

Bicuculline

Strychnine

Native activity

PC2

PC1

for both neurochemicals, measured at afixed concentration] only 2 coordinates areused. Each of the coordinates contains in-formation of all of the 11 variables. Ascan be seen from Fig. 5, all data pointscoming from one substance cluster in asmall area that can clearly be distinguishedfrom the area in which the data points ofthe other substance cluster. The nativeactivity before addition of the neurochemi-cals, however, was widespread. Therefore,the two substances can easily be distin-guished. Or, simplified: If a data point of anew measurement (which either contains

strychnine or bicuculline) falls within thecluster of the strychnine measurements,this sample contained strychnine.

The quantitative analysis of strychnineis presented in Fig. 6. Here, the differentconcentration dependencies of the differ-ent variables were evaluated by an ANN. Ina neural network analysis, in a first step,the algorithm calculating the results hasto be trained for the specific task. This isdone by feeding data into the network andtelling the system the correct result. TheANN calculates a result with the input vari-ables, getting a completely false result at

0.01 0.01

0.1

1

10

100

0.01 0.1 1 10 100 0.01 0.1 1 10 100

0.1

1

10

100

(a) (b)

Pre

dict

ed c

once

ntra

tion

[µM

]

Pre

dict

ed c

once

ntra

tion

[µM

]

Adjusted concentration[µM]

Adjusted concentration[µM]

Fig. 6 Predicted vs. adjusted (real)concentration of strychnine in embryonic micespinal cord cultures, evaluated with abackpropagation artificial neural network [22, 23,25]. (a) The net was trained with 2/3 of the data

and tested with the remaining 1/3 of the dataand (b) training was performed with data from afirst concentration series, test with data from asubsequent series performed 30 min afterthe first.

64 3 Bioelectronics

the very beginning of the training phase.By changing the weights of different partsof the calculation, the network trains it-self to reproduce the correct result. If lateron, new, that is, unknown data are fedinto the ANN, it is able to predict the cor-rect answer by using the now optimizedcalculation parameters.

Shown in Fig. 6 is the correlation of realconcentrations of strychnine and thoseconcentrations that are predicted by theANN analysis of experimental data ofstrychnine exposure [25, 26]. In the leftpart of the figure, the good correlationwithin the same experiment is shown.This means that a part of the data ofone experiment was chosen to train theANN, whereas the other data of the sameexperiment was used as the independenttest data.

However, for real biosensor applica-tions, the calibration and training of theANN is done in a first, independent step.In a second step, that is, after days orweeks, an unknown analyte is studied. Theresult of such an experiment is shown onthe right side of Fig. 6. Although still cor-related, the prediction error is now large.In this experiment, the results used fortraining were obtained immediately be-fore the sensor test was performed. Onereason for the deviations is therefore cer-tainly the poisoning of the network by thehigh strychnine concentrations and a tooshort recovery time between the trainingand analysis sequences. But these resultsshow that, in principle, a quantificationof biosensor data based on natural neu-ral networks after evaluation by ANNs willbe possible.

One trend in neuron-based biosensorsconcerns the utilization of whole tissue.Tissue may be even more organo-typicthan in vitro grown cell networks. Thisis particularly important for drug tests

in the pharmaceutical industry. However,tissue cultures are in most cases difficultto keep for longer time, although newculture techniques may help promotelongevity [29, 30].

Taking whole animals or at least com-plete sensory organs such as insect an-tenna comprises another relatively newtrend, which was introduced by Rechnitzand coworkers in 1986 [31]. New experi-ments on potato beetles or their antennashow a sensitive response to broken greenleaves [32], diseased potato tubers [33], orsmouldering odorants [34]. These exam-ples are nearest to a bioelectronic nose inits original sense.

3.3.2Biocomputer

Parallel computing will be a feature ofparticular importance for future computerarchitectures. However, the trend towardsmassively parallel computers, that is, withthousands and more computing elements,such as in the connection machine, haslost impetus because of several problemsinherent to silicon devices on the hardwareand the software level. These problemsinclude the connectivity problem, thecomplexity of an appropriate routing,and the difficulty to write parallel codes.On the other hand, parallel processingis performed in a very efficient wayin biological systems, far superior toexisting silicon-based devices. The otherfeature that one may overcome withbiocomputers is the fact that siliconcomputers need absolutely correct inputdata to find the correct solution. One canhope that a biological computer will arriveat the correct answer based on partialinformation by filling in the gaps by itself.Therefore, the combination of naturalneurons with silicon technology could lead


to bioelectronic circuits with fascinatingnew data processing properties. However,like in neural network-based biosensors,a tremendous amount of work has tobe done before one can think of a real‘‘biocomputer’’.

A recent example in this direction wasdemonstrated by Ditto and coworkerswith the so-called leech-ulator, in whichtwo leech neurons were used for asimple mathematical addition [35]. Leechneurons are used because they havebeen extensively studied and are wellunderstood. The neurons are electricallystimulated by inserted microelectrodes.Each neuron has its own electrical activityand responds in its own way to theelectrical stimulus. These features can beused to make each neuron represent anumber. Calculations are then performedby linking up the individual neurons.Using principles of chaos theory, the twoneurons are selectively stimulated. Fromthe signal cascade that follows, the PCcan extract the correct answer to a simpleaddition problem.

Because the leech neurons are ableto form their own connections fromone to another, the biological computerworks out its own way of solving theproblem. This approach to computingis particularly suited to pattern recogni-tion tasks such as reading handwriting,which would take enormous amounts ofpower to be performed on a conventionalcomputer.

However, intracellular electrodes reducethe longevity of the neurons to extremelyshort periods, which limits the use ofthis approach for practical applications.The concept of extracellular detectiondescribed in Sect. 3.1 has therefore tobe used for a real device. In particular,the use of CMOS devices in whichdata preprocessing can be performed in

the direct vicinity of the cells may bepromising, hence reducing the problemof the still bad signal-to-noise ratio ofextracellular recording devices [36].

3.3.3Short Summary of In Vivo Applications

In vivo applications of bioelectronic de-vices, such as artificial limbs, cochlear orretina implants, will not be stressed here indetail. Many of the problems that have to besolved for in vitro devices, such as a stableneuron/electrode contact, do also matterfor in vivo applications. However, for thelatter, much more difficult requirementshave to be met, such as biocompatibil-ity not only against the neural cells butthe whole body, including resistance tobody reactions against the foreign devicesuch as inflammation or scar formation,mechanical stability in a moving system(muscle, eye, . . . ), long-term stability overyears, as well as practical requirementssuch as easy implantation. Despite allthese difficulties, there are systems such aspacemakers and cochlear implants alreadyon the market [37]. Retina implants areunder development [38]. And first studiesare made with intelligent artificial limbs.Therefore one can hope that in the twenty-first century many of the above-mentionedproblems will be solved.

References

1. S. Roth, One-Dimensional Metals, Wiley-VCH, Weinheim, Germany, 1995.

2. E. R. Kandel, J. H. Schwartz, Principles ofNeural Science, 3rd ed., Elsevier, New York,1991.

3. W. Gopel, Ch. Ziegler (coordinating au-thors), H. Breer, D. Schild (biological sec-tion), R. Apfelbach, J. Joerges, R. Malaka,(contributing authors), Biosens. Bioelectron.1998, 13, 479–493.

66 3 Bioelectronics

4. Ch. Ziegler, W. Gopel (coordinating au-thors), H. Hammerle, H. Hatt, G. Jung,L. Laxhuber, H.-L. Schmidt, S. Schutz,F. Vogtle, A. Zell (contributing authors),Biosens. Bioelectron. 1998, 13, 539–571.

5. W. Gopel, Sens. Actuators, B 1998, 52,125–142.

6. Ch. Ziegler, Fresenius Z. Anal. Chem. 2000,366, 552–559.

7. G. W. Gross, A. Harsch, B. K. Rhoades et al.,Biosens. Bioelectron. 1997, 12, 373–393.

8. G. W. Gross, B. K. Rhoades, R. J. Jordan,Sens. Actuators 1992, 6, 1–8.

9. G. W. Gross, H. M. E. Azzazy, M.-C. Wuet al., Biosens. Bioelectron. 1995, 12, 373–393.

10. U. Egert, B. Schlosshauer, S. Fennrich et al.,Brain Res. Protocols 1998, 2, 229–242.

11. S. M. Potter, Distributed processing incultured neuronal networks in Advancesin Neural Population Coding (Ed.: M. A.L. Nicolelis), Progress in Brain Research,Elsevier, New York, 2001, pp. 49–62,Vol. 130.

12. P. Fromherz, A. Offenhauser, Th. Vetteret al., Science 1991, 252, 1290–1293.

13. J. J. Hickmann, D. A. Stenger, Interactionsof cultured neurons with defined surfacesin Enabling Technologies for Cultured NeuralNetworks (Eds.: D. A. Stenger, T. M. McKenna), Academic Press, New York, 1994,pp. 51–76.

14. J. Graf, Y. Iwamoto, M. Sasaki et al., Cell1987, 48, 989–996.

15. P. Heiduschka, W. Gopel, W. Beck et al.,Chem.–Eur. J. 1996, 2, 667–672.

16. A. Blau, C. Weinl, J. Mack et al., J. Neurosci.Methods 2001, 112, 65–73.

17. G. W. Gross, Internal dynamics of ran-domized mammalian neuronal networks inculture in Enabling Technologies for CulturedNeural Networks (Eds.: D. A. Stenger, T. M.Mc Kenna), Academic Press, New York,1994, pp. 277–317.

18. G. W. Gross, S. Norton, K. Gopal et al.,Cellul. Eng. 1997, 2, 138–147.

19. S. I. Morefield, E. W. Keefer, K. D. Chapmanet al., Biosens. Bioelectron. 2000, 15, 383–396.

20. http://www.multichannelsystems.com.21. http://www.cnns.org/.22. http://www.panasonic.com/medical

industrial/medsys.asp.23. http://www.caltech.edu/∼pinelab/steve.

html.24. A. Gramowski, D. Schiffmann, G. W. Gross,

NeuroToxicology 2000, 21, 331–342.25. A. Harsch, Ch. Ziegler, W. Gopel, Biosens.

Bioelectron. 1997, 12, 827–835.26. A. Harsch, Ch. Ziegler, W. Gopel, Sens. Ac-

tuators, B 2000, 65, 160–162.27. A. Mohr, W. Finger, K. J. Fohr et al., Sens.

Actuators, B 1996, 34, 265–269.28. A. Harsch, Ph.D. Thesis, University of

Tubingen, Germany, 1997.29. H. Kamioka, Y. Jimbo, P. J. Charlety et al.,

Cellul. Eng. 1997, 2, 148–153.30. P. Thiebaud, C. Beuret, M. Koudelka-Hep

et al., Biosens. Bioelectron. 1999, 14, 61–65.31. S. Belli, G. Rechnitz, Anal. Lett. 1986, 19,

403–416.32. P. Schroth, M. J. Schoning, P. Kordos et al.,

Biosens. Bioelectron. 1999, 14, 303–308.33. S. Schutz, B. Weißbecker, U. T. Koch et al.,

Biosens. Bioelectron. 1999, 14, 221–228.34. S. Schutz, B. Weißbecker, H. E. Hummel

et al., Nature 1999, 398, 298–299.35. http://www.neuro.gatech.edu/acl/.36. INPRO (Information Processing by Natural

Neural Networks), European Union ProjectIST-2000-26463, for more information con-tact the author.

37. http://www.medizin.uni-tuebingen.de/hno/cochlearimplant.htm and links given there.

38. http://www.uak.medizin.uni-tuebingen.de/subret and http://www.nero.uni-bonn.de/projekte/ri/ri-index-en.htm and links giventhere.

67

4Electrochemistry ofNAD(P)+/NAD(P)H

Lo GortonDepartment of Analytical Chemistry, Lund University, Lund, Sweden

Elena DomınguezDepartment of Analytical Chemistry, Faculty of Pharmacy,University of Alcala, Madrid, Spain

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 694.1.1 NAD(P)+/NAD(P) in Living Systems . . . . . . . . . . . . . . . . . . . . . 694.1.2 Formal Potential of the NAD+/NADH Redox Couple . . . . . . . . . . . 724.1.3 Focus and Scope of the Chapter . . . . . . . . . . . . . . . . . . . . . . . . . 73

4.2 Direct Electrochemical Oxidation of NAD(P)H . . . . . . . . . . . . . . . 764.2.1 General Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 764.2.2 Effect of Adsorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 764.2.3 Mechanism and Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

4.3 Homogeneous Oxidation of NAD(P)H by Oxidizing Redox Compounds 814.3.1 One-electron No-proton Acceptors . . . . . . . . . . . . . . . . . . . . . . . 824.3.2 Two-electron-proton Acceptors . . . . . . . . . . . . . . . . . . . . . . . . . . 844.3.2.1 o-quinones and p-quinones . . . . . . . . . . . . . . . . . . . . . . . . . . . . 844.3.2.2 Aromatic Diimines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

4.4 Electrocatalytic Oxidation of NAD(P)H at Mediator-modified Electrodes 894.4.1 General Remarks of CMEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 894.4.1.1 o-Quinones and Phenylenediimine Derivatives . . . . . . . . . . . . . . . 894.4.2 Other Mediating Functionalities and Metal-coated Electrodes . . . . . 1154.4.3 Catalytic NADH Oxidation at CMEs Based on Polymers . . . . . . . . . 1164.4.4 Mechanistic and Kinetic Aspects . . . . . . . . . . . . . . . . . . . . . . . . 118

4.5 CMEs Based on NADH-oxidizing Enzymes . . . . . . . . . . . . . . . . . 123

68 4 Electrochemistry of NAD(P)+/NAD(P)H

4.6 Amperometric Biosensors Based on NAD(P)-dependentDehydrogenase Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124

4.7 Direct Electrochemical Reduction of NAD(P)+ . . . . . . . . . . . . . . . 1244.7.1 General Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1244.7.2 Adsorption Phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1264.7.3 Mechanism and Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

4.8 Electrocatalytic Reduction of NAD(P)+ . . . . . . . . . . . . . . . . . . . . 1314.8.1 Nonenzymatic Electroreduction of NAD(P)+ . . . . . . . . . . . . . . . . 1314.8.2 Enzymatic Electroreduction of NAD(P)+ . . . . . . . . . . . . . . . . . . . 133

Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

69

4.1Introduction

4.1.1NAD(P)+/NAD(P) in Living Systems

The pyridine nucleotides, NAD (nicoti-namide adenine dinucleotide) and NADP(nicotinamide adenine dinucleotide phos-phate), are ubiquitous in all living sys-tems as they are required for the reac-tions of more than 400 oxidoreductase-denoted dehydrogenases (250 of thesedepend on NAD and around 150 onNADP). This number represents 17% ofall classified enzymes and consequently,these nucleotides are responsible formore enzymatic reactions than any othercoenzyme [1]. Largely, the role of thesenonproteinaceous coenzymes in enzyme-catalyzed reactions is to serve as theacceptor/donor of what is equivalent toa hydride ion (H−) from the substrate in areversible manner, thus playing a key rolein biological electron transport [2]. Thismeans that they have necessarily evolvedto be specific in their redox reactions. Itis of utmost importance for the function-ing of the living cell system that theseredox coenzymes are inherently fastidi-ous in their choice of redox partners andtherefore recognize and undergo rapid re-action with their desired redox partners

and at the same time, do not react at anyappreciable rate with thermodynamicallyfavorable but undesirable side reactions.Unlike many other oxidoreductases de-pending on other redox cofactors, suchas flavin nucleotides, pyrroloquinolinequinone (PQQ), heme, and iron-sulfurclusters, in the nicotinamide-dependentdehydrogenases, the cofactor is not per-manently bound within the enzyme butacts as a soluble cosubstrate, that is,there exists a 1 : 1 stoichiometric rela-tionship between the cofactor and thesubstrate. It should be noted that a fewother nonredox enzymes, such as adeno-sylhomocysteinase (EC 3.3.1.1), urocanatehydratase (EC 4.2.1.49), or myo-inositol-1-phosphate synthase (EC 5.5.1.4), containvery tightly bound NAD with no net changeat the end of the reaction, but actingas a true coenzyme and serving catalyti-cally as the hydrogen acceptor and donorfor the intermediates in the reaction se-quence [3].

There are two different forms of thenicotinamide coenzymes: β-NAD+ andβ-NADP+. They have closely related struc-tures, both revealed in Fig. 1, and uniqueelectrochemical properties. In solution,NAD+ acquires a folded conformationas suggested from circular dichroism [4],fluorescence spectroscopy [5], NMR [6–8],and X-ray crystallography [9] experiments,


(a)

N+H

H2NOC

R*

NN

N

N

OO

PO

PO

O O

HO

OH OH

OH

O N

HO OH

H2N

OH2N

NAD+(b)

N

H2NOC

SH

RHR*

1,4-NADH

+2e−+H+

si -face si -face

re-face re-face

R*

Fig. 1 (a) Formulas of enzymatically active 1,4-NADH (i.e.β-NAD reduced form). The isomer, an α-glycosidicnicotinamide-ribose linkage, is not enzymatically active. In thephosphorylated coenzyme, a PO(OH)2 group replaces theindicated H; (b) stereospecific redox reaction between NAD(P)+and NAD(P)H.

and also confirmed by molecular dynamicssimulation [10]. While a consensus atomic-level model has not emerged from thesedata, and the interpretation of some NMRdata have been questioned [11], NAD+ insolution is thought to be a mixture offolded and unfolded forms with the aro-matic rings in close proximity in the foldedform. However, the nature and extent ofthe interaction between the aromatic ringsin the folded position still remains con-troversial. Some groups have proposedthat parallel-ring stacking with an inter-ring distance of less than 0.39 nm is thehallmark of the folded form of NAD+ [4,12–14], while others have proposed aless restrictive [6, 7] conformation of thefolded form with an interring distancegreater than 0.45 nm and the aromaticrings not perfectly stacked in parallel [8].Molecular dynamics calculations, consis-tent with NMR relaxation data, result ina conformation with an average interringdistance of 0.52 nm, an average interring

angle of 148, nearly parallel glycosyl bondvectors, and the nicotinamide si side facingthe adenine [10]. Additional discrepancy isalso observed in the relative proportionsreported for the extended and folded con-formation. A range within 15% and 60%has been reported for the folded conforma-tion [14–16]. Regardless of how compactthe folded conformation is in solution,this clearly contrasts with the extendedunfolded configuration that NAD+ offerswhen attached to an enzyme [17].

The redox reaction between the oxidizedand reduced forms involves two electronsand one proton and can formally beconsidered as a hydride (H−) transfer;

NAD(P)+ + H+ + 2e− ⇐⇒ NAD(P)H(1)

As in the case of biological systems, thehydride transfer takes place at the C-4 po-sition of the nicotinamide ring and thus,a basic understanding of the electrochem-ical behavior of the NAD(P)+/NAD(P)H

4.1 Introduction 71

redox pair may lead to a more compre-hensive overview of biological electrontransfer mechanisms. Equally unique andone of the most stringently conservedproperties of NAD(P) is the absolutestereospecificity of the dehydrogenases forthese coenzymes. Some of them can onlytransfer the R hydrogen from the dihy-dronicotinamide to their substrates or theS hydrogen (sometimes also denoted asA and B), and the same stereospecificityis kept for the reduction of NAD(P)+introducing the hydrogen either to there-face of the trigonal C-4 or to the si-face, respectively, (Fig. 1b). Additionally,most NAD(P)-linked enzymes are alsostereospecific for the hydrogen transferof the substrate allowing stereochemicalchoices in biosynthetic work. The stere-ospecific mechanism of these reactionswas clearly elucidated in the early 1950s bythe group of Westheimer at the Universityof Chicago. This was made possible by theuse of deuterated forms of coenzymes andsubstrates [18–21].

The basic redox reaction catalyzed bya NAD(P)-dependent dehydrogenase fol-lows according to reaction (2):

substrate + NAD(P)+ ⇐⇒ product

+ NAD(P)H + H+ (2)

However, more complex reaction schemesprevail, for example, the redox intercon-version of L-glutamate and α-ketoglutaratecatalyzed by L-glutamate dehydrogenase.Most of these dehydrogenases are specificto either NAD+ or NADP+; however, somecan make use of both, although usuallywith different reaction rates. As a rule ofthumb, the NAD+-dependent dehydroge-nases are involved in catabolic reactions,whereas the NADP+-dependent ones areinvolved in anabolic reactions in conjunc-tion with, for instance, photosynthesis [22].

NAD+/NADH or NADP+/NADPH alsoparticipate in a number of other enzyme-catalyzed redox reactions, which do notinvolve dehydrogenases in the classicalsense as described earlier. These redoxenzymes contain bound cofactors suchas flavins, heme, and iron-sulfur clusters.One group of these enzymes is usuallydenoted diaphorase (DI) when used inanalytical systems and includes bothflavin-containing lipoamide and lipoyldehydrogenases (EC 1.8.1.4).

NADH + diaphoraseox ⇐⇒NAD+ + diaphorasered (3)

Both flavin adenine dinucleotide (FAD)and flavin adenine mononucleotide(FMN)-containing DIs are found, somespecific for NAD, whereas othersfor NADP. They have been usedfor staining dehydrogenase-rich tissuesand in attempted applications for theregeneration of the cofactor such asin bioelectroanalytical devices and flowsystems (see following text) and inbioorganic synthetic reactions.

Examples of other redox enzymes in-clude:

1. NADH-FMN oxidoreductase (or cy-tochrome c reductase, EC 1.6.99.3)catalyzing the reduction of FMNby NADH,

NADH + FMNNADH-FMN

oxidoreductase⇐========⇒NAD+ + FMNH2 (4)

2. NADH peroxidase (EC 1.11.1.1) catalyz-ing the reduction of hydrogen peroxideby NADH,

NADH + H2O2NADH peroxidase−−−−−−−−−−→

NAD+ + H2O (5)


3. NADH and NADPH oxidase (EC1.6.99.1), a flavin enzyme using oxygenas electron acceptor for the oxidation ofNADH with the production of hydro-gen peroxide,

NADH + O2NADH oxidase−−−−−−−−−→

NAD+ + H2O2 (6)

4. Proton translocating transhydrogenase(EC 1.6.1.1) found in bacteria and an-imal mitochondria that couples thetransfer of reducing equivalents be-tween NAD(H) and NADP(H) tothe translocation of protons acrossthe membrane, the equilibrium beingshifted toward the inner transport

NADH + NADP+ + Hout+ ⇐⇒

NAD+ + NADPH + Hin+ (7)

This enzyme is believed to play animportant role in the redox balance ofNADPH in the cell.

5. Hydrogenases containing iron-sulfurclusters (EC 1.12.1.2) and producedby a few microorganisms catalyze thereduction of NAD+ into NADH bymolecular hydrogen,

NAD+ + H2 ⇐⇒ NADH + H+ (8)

In biological electron-transfer pathways,such as the respiratory chain in the mi-tochondria, there is a switch betweencharge-transfer reactions involving one ortwo electrons. In the respiratory chain,the starting point is the donation of a hy-dride equivalent from cytosolic NADH toa membrane-bound flavoprotein (NADHdehydrogenase), which in turn delivers itscharge to an iron-sulfur cluster containingprotein, and then via ubiquinone throughthe cytochrome chain to cytochrome oxi-dase, and finally to molecular oxygen. The

charge-transfer reactions taking part priorto the reduction of NAD+ to NADH in thecytosolic solution involve two electrons,whereas those occurring within the respi-ratory chain between the flavoprotein andcytochrome oxidase are one-electron reac-tions [22]. It has therefore been and still isof great interest and great controversy howthe charge transfer occurs between NADHand the oxidized flavin within the flavopro-tein NADH dehydrogenase as somewherefrom NADH to ubiquinone, the reactionswitches from being a two-electron- to aone-electron-transfer reaction. Similarly,but in the reverse order, the photosyntheticelectron-transfer pathway starts with a se-ries of one-electron-transfer reactions andends up with a two-electron one-protonreduction of NADP+ to NADPH.

4.1.2Formal Potential of the NAD+/NADHRedox Couple

The generally accepted formal potential,Eo′

, of the NAD+/NADH redox couple atpH 7.0 (25 C) is −315 mV versus normalhydrogen electrode (NHE)(−560 mV vs.saturated calomel electrode (SCE) [23, 24].From thermal data and the equilibriumconstants of the ethanol/acetaldehydeand 2-propanol/acetone reactions cat-alyzed by alcohol dehydrogenase, a valueof −320 mV was calculated, which waslater recalculated to be −315 ± 5 mV ver-sus NHE by Clark [23]. Through directpotentiometric titrations using several dif-ferent mediators and xanthine oxidase ascatalyst, Rodkey [25, 26] obtained an Eo′

value of −311 mV versus NHE (25 C)and a temperature variation of the Eo′

of −1.31 mV/C in the range of 20 to40 C. A variation of the Eo′

with pH of−30.3 mV/pH (30 C) was found, which

4.1 Introduction 73

is in good agreement with the theoreti-cal value of −30.1 mV/pH. Rodkey andDonovan [27] also investigated the Eo′

ofNADP+/NADPH and found that this re-dox couple at pH 7 is very close to thatof NAD+/NADH with a maximum dif-ference at the same pH of 4 mV andthe variation with pH being the sameas that of the NAD+/NADH redox cou-ple. The Eo′

of the NAD+/NADH redoxcouple remains more negative than theEo′

of the many substrate/product cou-ple depicted in reaction (2), being obviousthat the catalytic oxidation of NADH intoNAD+ is thermodynamically favored, un-less a second reaction (catalytic, chemical,or electrochemical) takes place making thereduction of NAD+ virtually favorable.

The inherent nature of the redox re-action makes it natural to couple thesedehydrogenase-catalyzed reactions to elec-trochemical methods. The transfer ofelectron(s) from a substrate to an elec-trode (or the reverse) may then take placevia electrochemical redox reactions of thecoenzymes, as depicted in Fig. 2. Theutility of combining enzymes and elec-trochemical methods for electroanalyticalapplications was predicted by Clark andLyons by an enzyme electrode [28] and byShaw in energy production by a biofuelcell anode [29] in the early 1960s.

NAD(P) in either redox form ispoorly dissolved in nonaqueous solu-tions. A series of NAD+/NADH-model

compounds have therefore additionallybeen chemically and electrochemicallyinvestigated in both protic and apro-tic media. Some of the commonlystudied are nicotinamide, nicotinamidemononucleotide (NMN+), 1-methyl-3-car-bamoylpyridinium ion (MCP+), 1-benzyl-3-carbamoylpyridinium ion (BNA+), 10-methylacridinium (MA+), 1-benzyl-3-cya-noquinolininium (BQCN+), 1-benzyl-3-carbamoylquinolinium (BQA+), and thecorresponding 1,4-dihydro forms anddimers [30–71]. Some of these analoguesare presented in Fig. 3. Biomimetic ana-logues of NAD+, comprising a nicoti-namide functionality coupled via anatrazine ring to a dibenzenesulphonic acidunit, have demonstrated their potentialuse as artificial and stable coenzymes withdifferent dehydrogenases [72, 73] with re-oxidation of the reduced biomimetic coen-zyme by N-methylphenazinium [74].

4.1.3Focus and Scope of the Chapter

The basic direct electrochemistry ofNAD(P)+/NAD(P)H has been studiedsince the early 1960s [75] and is sum-marized. Overall, this chapter covers thecharge-transfer reactions of both redoxforms occurring not only at naked but alsoat chemically modified electrodes (CMEs)facilitating the interconversion of one re-dox form into the other. Following the

−560

e−

NAD+ / NADH S/P Mediator Electrode poised at around 0 mV

Eo' vs. SCE pH 7.0[mV]

0

Fig. 2 Electron-transfer pathway in dehydrogenase-catalyzed reactions coupled to electrodesvia electrochemical redox reactions of the coenzymes. Electron fluxes represent the oxidation ofthe substrate (e.g. Pyruvate/Lactate Eo′ = −435 mV vs. SCE, pH 7.0).


OH

N+

CNH2

O

O

OH

CH2O

P

O

OHO

P

O

OHO

ON

OH

N

N

N

H2N

CH2

OH

NMN+

Nicotinamide

Fig. 3 Formulas of some NAD+/NADH analogues.

Tab. 1 Summary of the main reactions in the electrochemistry of NAD+/NADH

Overall redox processNAD+ + H+ + 2e− ⇐⇒ NADH

Electrochemical oxidation of NADH Simple ElectronTransfer (ECE mechanism)

Electrochemical reduction of NAD+

NADH−e−

−−−→ NADH•+ −H+

−−−→ NAD• −e−⇐==⇒ NAD+ NAD+ +e−⇐==⇒ NAD

• +H+ + e−−−−−−−−→ NADH

DisproportionationNADH

•+ + NAD• ⇐⇒ NADH + NAD+

Dimerization2 NAD

• ⇐⇒ (NAD)2Simple One Step Hydride Transfer

NADH + Q −−−→ NAD+ + QH−NAD+ + RH −−−→ NADH + R

Sequential Hydride Transfer within a complex (pathway I)NADH + Q ⇐⇒ [NADH

•+Q−] −−−→ [NAD+QH−] −−−→ NAD+ + QH−Sequential Hydride Transfer within a complex (pathway II)

NADH + Q −−−→ [NAD•QH

•] −−−→ [NAD+QH−] −−−→ NAD+ + QH−

Electrocatalytic oxidation through the formation of a charge transfer complexNADH + Medox ⇐⇒ [CT] −−−→ NAD+ + Medred

In detail:NADH + Medox ⇐⇒ [NADH − Med] ⇐⇒ [NADH

•+Med•−] −−−→

[NAD MedH•] −−−→ NAD+ + MedH−

Electrocatalytic oxidation through a charge transfer-complex with innersphere electron transfer∗PyH2 + Q ⇐⇒ [PyH2Q] ⇐⇒ [PyH

•+2 Q

•−] −−−→ [PyH•QH

•] −−−→ PyH+ + QH−

∗PyH2 corresponds to dihydropyridine compounds.

4.1 Introduction 75

authors’ view, the description of electrodesmodified with inorganic and organic me-diators and biochemical catalysts is alsogiven. The homogeneous charge-transferreactions with other redox compounds asthe basis for the understanding of the re-action sequence occurring at electrodeschemically modified with compounds sim-ilar to those used as modifiers are also dis-cussed. A summary of the main reactionsdescribed in this chapter is presented inTable 1. Sch. 1 depicts the electrochemicalpathways for the oxidation and reductionof NADH and NAD+, respectively.

Much of the driving force for these stud-ies has been, and still is, with the hope toincrease the fundamental understandingof biological electron-transfer reactions,and also with the hope that the electro-chemical reactions of NAD(P)+/NAD(P)Hcould be used for practical applications

such as in (electro)analytical chemistry [75]including amperometric biosensors [76,77] and biofuel cells [78–80] and in elec-troorganic chiral synthesis [24, 81–84].

The general chemistry of NAD(P)+ andNAD(P)H will not be covered or furthercommented on in this chapter, except thatNAD(P)H is relatively stable in aqueoussolutions at pHs more alkaline than pH 7and NAD(P)+ at pHs more acidic thanpH 7. The stability is very much depen-dent on the buffer constituents and ionicstrength. The pH where both redox formstogether exhibit minimal destruction dueto acid/base decomposition is found be-tween pH 7 and 8, depending on whetherthe aqueous medium is unbuffered andwhen buffered, on the buffer constituents.In general, the stability of NADPH is lessdependent on the buffer than is that ofNADH. The reader is advised to refer to

NADH

OXIDATION

NADHX NADH•++NAD•

NAD•

+e−

+e− −e−−e−

−e−

−H+

−x−

+HX

(NAD)2

(NAD)2•+NAD+

(NAD)2

REDUCTION

NADH + NAD+

E o' = −1.16 V

E o' = 0.81 V

E≈−0.4 V

kd≈106−107M−1s−1

pKa≈ −4kH+>106s−1

Scheme 1 Pathways for the electrochemistry of NAD+/NADH redox pair.


excellent books and reviews [24, 85–88] forfurther information.

4.2Direct Electrochemical Oxidation ofNAD(P)H

4.2.1General Observations

Studies on the electrochemical oxidationof NADH have been made using cyclicvoltammetry (CV), potential step chrono-amperometry, constant electrode potentialcoulometry, and rotating disk electrode(RDE) methodology. Most commonly, avariety of carbonaceous electrode materialshave been used including glassy carbon(GC) and pyrolytic graphite (PG) [75,89–99], and carbon fibers [100–105]. Ofthe metal electrodes, platinum (Pt) [30,75, 91, 94, 96, 98, 106–109] and gold(Au) electrodes [98, 110–115] are the mostcommon electrode materials, althoughother metals, for instance, silver [116, 117],have also been used.

A poorly defined oxidation wave ofNADH at a Pt electrode around 1 V ver-sus NHE was observed in an initial studyby Burnett and Underwood [118] reflectingthe large over voltage of the electro-chemical NADH oxidation. The electrodematerial has a significant effect on theover voltage. The oxidation of NADH in anaqueous solution, seen as a single peak inCV, takes place at potentials of ≈0.4, ≈0.7,and ≈1 V at C, Pt, and Au electrodes, re-spectively [75, 90, 97, 98]. As a commonobservation of all bare electrode materials,it was early recognized, as in the case ofthe reduction of NAD+ (see in the follow-ing text), that the electrochemical reactionresults in electrode fouling, necessitat-ing careful pretreatment and conditioningof the electrodes to obtain reproducible

results between runs [30, 90, 94, 109]. CVoften gives values of the number of elec-trons participating in the electrochemicalprocess (n) close to the expected value oftwo, while lower values of n are usuallyfound in coulometric studies. In contin-ued coulometric studies by Coughlin andcoworkers [106, 107] and by Jaegfeldt andcoworkers [108], the investigators founda recovery of 99.3% enzymatically activeNAD+, using low concentrations of the co-factor, a pretreated fast rotating Pt gauzeelectrode to minimize the adsorption, andcorrecting the decomposition of the coen-zyme in solution [24]. From the fact thatthe major product of the electrochemicaloxidation of NADH in aqueous solutionwas NAD+ in combination with the factthat n is equal to two, it follows that thenet reaction can be summarized as [90, 91,119–121]:

NADH −−−→ NAD+ + H+ + 2e− (9)

4.2.2Effect of Adsorption

It was early recognized that the elec-trochemical oxidation of NADH sufferedfrom severe effects of adsorption. Early in-vestigations at C and Pt electrodes showedthat the adsorption of NAD+ and possiblyother unknown species occur at positivepotentials [89–91, 93–95, 109], with indi-cations of the desorption of NAD+ fromGC electrodes at a potential of 0 V [93].In a thorough investigation by Samec andElving, the oxidation of NADH at GC, Pt,and Au electrodes was studied and theresults obtained at the different electrodematerials with CV and RDE were com-pared [98]. The influence of the preadsorp-tion of NAD+, NADH, NMN+, NMNH,nicotinamide, adenine, and adenosine be-fore investigating the electrochemistry of

4.2 Direct Electrochemical Oxidation of NAD(P)H 77

NADH, pointed to the fact that boththe adenine and nicotinamide moietiesare involved in adsorption at Pt and Auelectrodes. NADH was also shown to ad-sorb onto C, Pt, and Au electrodes. Thiswas demonstrated with electrodes exposedto NADH solutions at open circuit, fol-lowed by extensive cleaning with waterand buffer. Then, CV experiments withthese electrodes in buffers not contain-ing NADH revealed anodic waves becauseof the oxidation of the adsorbed NADH.It was concluded that one of the majordifferences between the three electrodematerials arises because of differences inthe adsorption of the coenzyme at the elec-trode surface. At Pt and Au electrodes,NADH was strongly adsorbed, whereasat GC, it was the oxidation product NAD+that was most strongly adsorbed. Addition-ally, it was also observed at the Pt and Auelectrodes, in parallel with the two-electronoxidation, there is a further oxidationprocess of the adsorbed NADH leadingto unspecified products and presumably,poisoning the metal surface. RDE inves-tigations with GC, Pt, and Au electrodesat concentrations of NADH below 2 mMrevealed that the limiting current was lin-early dependent on the square root of theangular velocity (ω1/2) and independentof scan direction. However, at higher con-centrations, deviations from linearity werenoticed for all the three-electrode materi-als, reflecting the influence on electrodefouling by the concentration of NADH inthe solution.

The adsorption of NAD+ on Au hasbeen further studied during recent yearsusing fourier transform surface-enhancedRaman scattering (FT-SERS) [113–115].The surface-enhanced Raman scattering(SERS) of NAD+ shows a strong potentialdependence in the non-Faradaic regions.Either the adenine or the nicotinamide

moiety may change its adsorption stateduring the potential scanning process.In regions of positive electrode poten-tial, only the bands responsible for theadenine and nicotinamide moieties canbe observed. In contrast, with a negativeshift of the potential, several additionalstrong bands representing the ribose andphosphate moieties are also evident. Thestacked NAD+ molecule is consideredto be opened to some extent on anelectrically charged electrode. Specifically,under sufficiently negative potential, theNAD+ molecule appears to exist in awell-extended state on the gold electrode,leading to the tight adsorption of the entireNAD+ molecule on the electrode.

In a work from the 1980s, Blankespoorand Miller investigated the influence ofthe adsorption of NAD+ on the electrodesurface on the electrochemistry of NADH.Potential step chronoamperometry [99] of1.0 mM NADH at pretreated GC elec-trodes with preadsorbed NADH (same asin [96]) was performed. At long timescales(t > 0.1 s), the current showed Cottrellbehavior for a two-electron process. Atshort timescales (t < 0.1 s), however, thecurrent was significantly less than thatexpected for a two-electron process andapproached the Cottrell behavior for a one-electron process. If the same experimentwas performed in the presence of 8 mMNAD+ in the contacting solution even atvery short times (≈ 2 ms), linear Cottrellbehavior for a two-electron process wasachieved. For pretreated uncoated GC withno NAD+ in the contacting solution attimes less than 7 ms, the current was evengreater than expected for a two-electronprocess. The investigators found clear ev-idence that NAD+ formed as the productwhen oxidizing NADH at carbon elec-trodes is adsorbed on the electrode surfaceand inhibits further oxidation of NADH. It


is also known that when oxidizing NADHat clean GC electrodes, a prewave appearson the NADH oxidation because of theweak adsorption of NADH and the strongadsorption of NAD+ [97]. The normalNADH anodic wave appears at concen-trations of NADH exceeding 0.1 mM. Theprewave can be eliminated by first saturat-ing the electrode surface with NAD+, forinstance, by adding NAD+ to the solutionand waiting, or by electrolytically generat-ing NAD+ [94, 95]. At high concentrationsof NAD+ (19 mM) in the contacting buffer,reproducible cyclic voltammograms ofNADH oxidation could be obtained [99].

4.2.3Mechanism and Kinetics

To shed further light on the mechanismof NADH oxidation, Blaedel and Haas [30]oxidized NADH model compounds in ace-tonitrile and observed two main oxidationsteps in the absence of a base, clearlydemonstrating the stepwise oxidation ofNADH-analogues. When oxidizing NADHin aqueous solutions, as mentioned ear-lier, only a single wave is observed andno waves due to the rereduction of in-termediates have been observed in CVeven at fast sweeps (30 V/s) [91], indi-cating a high chemical irreversibility ofthe reaction. A potential variation (Epor E1/2) with pH for the overall electro-chemical NADH oxidation of −30 mV/pHmay be expected if the limiting reactioninvolves a proton-transfer step, but hasnot been observed. Various results havebeen reported such as −17 [90], −11 [91],+35 mV/pH [106], and no pH dependenceat all [120]. An electrochemical-chemical-electrochemical (ECE) mechanism for theelectrochemical oxidation has thereforebeen proposed in several studies [30, 96,97, 99, 109, 120, 122].

NADH−e−

−−−→ NADH•+

−H+−−−→ NAD

• −e−⇐⇒ NAD+ (10)

However, different views of the rate ofthe individual steps in the reaction mech-anism and influences of the concurrentreactions have been discussed.

A key detail in the NADH oxidationpattern is the deprotonation step and itsrelation to the initial potential-determiningelectron-transfer step. Kinetic studies onthe electrochemical oxidation of NADHwere presented by Moiroux and Elving [96]and Jaegfeldt [109] almost simultaneously.Jaegfeldt reported the involvement ofsecond-order pH dependence for whichthere is still no satisfactory explanation.Moiroux and Elving estimated the rateconstant of the second step, kH+ , inreaction (10)

NADH•+ kH+−−−→ NAD

• + H+ (11)

to be 60 s−1 at NAD+-covered GC elec-trodes, assuming the rate to be initiallycomparable with the overall rate-limitingfirst step. The possibility of a dimerizationof NAD• radicals has also been suggestedas known to occur when electrochemicallyreducing NAD+ [75, 123],

2 NAD• kd−−−→ (NAD)2 (12)

with a very high dimerization rate con-stant, kd, in the order of 106 –107 M−1

s−1 [98, 124–126], followed by the forma-tion of NAD+ from the oxidation of thedimer:

(NAD)2 −−−→ 2 NAD+ + 2e− (13)

However, as the oxidation occurs positiveto 0.2 V (oxidation of (NAD)2 to NAD+

4.2 Direct Electrochemical Oxidation of NAD(P)H 79

starts already at ≈ −0.4 V [75]), oxidationof NAD• to NAD+ (last step in reaction(10)) would be sufficiently rapid to outrunthe dimerization, so that practically nodimer is produced.

Blankespoor and Miller [99] later reex-amined the results and demonstrated thatNAD+ is an inhibitor of the oxidation pro-cess and that the oxidation is of first orderin NADH in the presence of a large excessof NAD+. Investigations by pulse radioly-sis [127] indicated a deprotonation rate ofNADH•+, reaction (11), much higher thanthe value estimated by Moiroux and Elv-ing and greater than 106 s−1, a value laterconfirmed by Matsue and coworkers to be6 106 s−1 [128].

An extension of reaction (10) has beenfurther suggested by also taking thedisproportionation reaction into account.

NADH•+ + NAD

• ⇐⇒ NADH + NAD+(14)

A chemical one-electron oxidation byferrocenium salts in buffered aque-ous propanol [129, 130] (see also fol-lowing text) gave an estimated valueof the formal potential, Eo′

, for theNADH•+/NADH redox couple of 0.81 V.This value was later reexamined by Mat-sue and coworkers who found it to be0.78 V versus SCE (at pH 7) [128]. Al-though reaction (14) is highly energeti-cally favorable

(Eo′

NAD+/NAD• = −1.16 V

),

Blankespoor and Miller referred to stud-ies by Amatore and Saveant [131], whoused relatively comparable parameters ina theoretical treatment of a concurrentECE and disproportionation mechanism,suggesting that more than 95% of the prod-uct is formed through the ECE pathway.

The high over-potential observed in thedirect electrochemical oxidation of NADHis thus caused by the very high potentialof the NADH

•+/NADH redox couple (first

reaction in the reaction sequence outlinedearlier, reaction 10).

NADH −−−→ NADH•+ + e− (15)

Thus, in light of an initial rate-limitingelectron-transfer step and the knowledgeof the Eo′

of the produced radical (0.78 Vvs. SCE) [128–130], its fast deprotona-tion rate (6 106 s−1) [127, 128, 130] andits estimated high acidity (pKa ≈ −4) [39],a pH dependence of the oxidation ofNADH, could not be observed. Thus,the reaction paths suggested by Elvingand coworkers [96, 97] assuming an ECEmechanism, reaction (10), and further sup-ported by experiments by Blankespoor andMiller [99] and calculations by Amatoreand Saveant [131] may be supplementedas the most probable reaction sequenceoccurring (Sch. 1). However, there still re-main unanswered questions that are ableto give a full satisfactory explanation to allthe observed results.

With strong support that the majorroute for direct oxidation of NADH oc-curs according to an ECE mechanism asoutlined in reaction (10) and in Sch. 1,some results remain unexplained. InRef. [93], Samec and Elving argue that

in the reaction sequence NADH−e−

−−−→NADH

•+ −H+−−−→ NAD

• −e−⇐⇒ NAD+ the

first reaction NADH−e−

−−−→ NADH•+ can

be assumed to be rate-determining. Asthe oxidation occurs positive to +0.2 V,oxidation of NAD• to NAD+ would besufficiently rapid to outrun the dimer-ization, so that practically no dimer isproduced. Under these circumstances, asingle anodic two-electron wave should beobserved at a potential that corresponds tothe initial irreversible one-electron NADHoxidation and that is independent of thenature of the electrode material, NADH


concentration, and solution pH as well.However, as all three effects are, at leastto some extent, involved in NADH ox-idation at solid electrodes, modificationof the mechanism outlined is required.The effect of electrode material is obvi-ously significant and it may underlie theeffects of both NADH concentration andsolution pH. The correspondence betweenthe rate of NADH oxidation at solid elec-trodes and the state of the electrode surfacecan be reasonably explained on the basisof intimate involvement of surface oxy-gen species in the rate-determining stepof the overall reaction. Samec and El-ving [97, 98] suggested in line with thework by Blaedel and Jenkins [90], thatsurface oxygen species are implicated inthe reaction assisting in carbon hydrogenbond cleavage (Fig. 4). Even at electrodesnot covered with adsorbed NAD+, theinitial step in the NADH oxidation pro-ceeds to at least some extent throughredox mediator systems located close tothe electrode surface such as the redoxcouples formed by oxygen adsorbed at Auand Pt surfaces, for example, OH

•ads/H2O

and Oads/OH•ads and by organic func-

tionalities resulting from oxidation of acarbon surface, such as for instance,quinone/semiquinone/hydroquinone sys-tems. The possible involvement of twosurface oxygen redox systems in the firststage of the NADH oxidation is schemat-ically depicted in Fig. 5 [97]. The electrontransfer path involves electron exchangebetween energy levels located at the sur-face atom and in the electrode, coupledwith electron exchange between energylevels of the surface oxygen atom and the

NADH molecule. The proton transfer pathinvolves transfer of the proton bound toC(4) of NADH to a third species, which isthe proton acceptor, for instance, a watermolecule, with possible intermediate for-mation of a bond to the surface oxygenatom (Fig. 4).

These findings suggest that the pretreat-ment of the electrode surface in some wayincreases the number of surface oxygengroups causing the oxidation reaction tobecome kinetically more rapid. Studiesin this direction have been carried out,however, mainly on carbon electrodes,where the effect is most pronounced [90,132–137], both with electrochemical pre-treatment and with other means such asradio frequency oxygen plasma treatment.Although the existence of functional oxideson carbon is well established, the type andquantity of functional groups varies greatlywith carbon material and pretreatment

Fig. 4 Proton transfer path onoxidation of NADH involving movementof H at the C4 of the nicotinamide ringto a proton acceptor such asH2O [97, 98].

N

H2NOC

H

HR

NADH

N+

H2NOC

H3O+

H2O

O

O

O

H

R

NAD+

+

+

A

B

C

4.3 Homogeneous Oxidation of NAD(P)H by Oxidizing Redox Compounds 81

NADH

H2O

H2O

H3O+

e−

NAD•

OH•ads

OH•ads

Substrate Mediator

A

NADH

H2O

H3O+

e−

NAD•

Oads

Substrate Mediator

B

Fig. 5 Schematic representation ofpossible surface oxygen redox systemsas mediators in the oxidation of NADHat solid electrodes [97, 98].

history [138]. Oxygen functional groups oncarbon have been studied and identifiedby infrared and Raman spectroscopy, wetchemical analysis, electron spectroscopyfor chemical analysis (ESCA), and so forth.Most probably, the presence of quinoneson the carbon electrode surface causes arather drastic decrease in the over voltage

compared with an untreated carbon sur-face. However, the long-term stability ofthe electrocatalytic effect of these elec-trodes for continued NADH oxidation isvery restricted, probably as a result ofblocking by NADH oxidation products atthe electrode surface (see also followingtext under CMEs for NADH oxidation).

On the basis of recent findings byTunon-Blanco and coworkers [139], onecould also speculate that the adeninemoiety of adsorbed NAD+ may undergoan oxidative reaction at high anodicpotentials, forming a strongly mediatingfunctionality on the electrode surfaceand thus facilitating the oxidation ofNADH at potentials below the Eo′

of theNADH•+/NADH redox couple, (Fig. 6).

4.3Homogeneous Oxidation of NAD(P)H byOxidizing Redox Compounds

As mentioned earlier, the pyridinecofactors have critical positions in

Fig. 6 Proposed structural formulasand redox reaction of the electrooxidixedadenine moiety of NAD+ acting as anefficient mediator for NADHoxidation [139].

N

N N

N

NH2

O

R

O

N

NH

N

HN

NH2

O

R

O+ 2e− + 2H+

N

N N

N

NH2

HO

R

OH


biological electron-transfer pathwayslocated at the switching point betweenone and two electron reactions. It istherefore of great scientific interest tofind evidence of either of the tworeaction pathways when studying theoxidation of NAD(P)H or the reductionof NAD(P)+ in model reactions eitheroccurring directly at an electrode surfaceor between the NAD(P) counterpartand another redox molecule in solution.Many investigations have therefore beenpursued with the objective of elucidatingthis important redox reaction proceedingin either direction. Experiments have beenperformed on NADH and NADPH inaqueous solutions and also on a greatnumber of model compounds in bothprotic and aprotic media with organic andinorganic oxidants. Today there is stillno general agreement on the true natureof the charge transfer reaction. Theseinvestigations are also of great importancefor the understanding of the reactionsequence occurring at mediator-modifiedelectrodes studied for the possible practicalapplications of catalytic NADH oxidation.A brief discussion of some of the mostimportant investigations in this directionfollows.

4.3.1One-electron No-proton Acceptors

An extensive investigation of the reac-tion mechanism and kinetics betweenNADH and a series of ferrocenium hex-afluorophosphate derivatives was madeby Miller and coworkers [129, 130]. Thisinvestigation was pursued because the re-action between NADH and one-electronno-proton acceptors is abnormal in com-parison with most biological reactions withNADH where a net transfer to the oxidiz-ing agent of a hydride equivalent from

the four-position of the dihydropyridinering occurs. With one-electron no-protonacceptors, the four-hydrogen from NADHis transferred to the solvent, not to theoxidant, and in this respect, it mimics thereaction occurring at a naked electrode sur-face. The reaction sequence when studiedwith CV proceeds as follows. In a first elec-trochemical step, ferrocene (Fc) is oxidizedto ferrocenium (Fc+);

Fc ⇐⇒ Fc+ + e− (16)

which in turn is then followed by theabstraction of the first electron fromNADH to ferrocenium;

NADH + Fc+ k1⇐⇒k−1

NADH•+ + Fc (17)

In the following reaction, the acid proton isdonated to a component of the solvent (B).

NADH•+ + B

k2⇐⇒k−2

NAD• + BH+ (18)

Finally, the second electron is donated to asecond ferrocenium

NAD• + Fc+ k3⇐⇒

k−3NAD+ + Fc (19)

Miller and coworkers found that thereaction was first order with respect to both[Fc+] and [NADH] (second order overall)and independent of pH, added NAD+, orFc. In the reaction, two equivalents ofFc+ and one equivalent of NADH, bothreactants, were completely consumed andFc and enzymatically active NAD+ wereisolated in greater than 75% yield.

In addition, they investigated a possibleisotope effect with NADH deuterated inthe four-position. As no primary isotopeeffect was found, it was thus establishedthat for one-electron no-proton acceptors,the reaction proceeds with the first-electron donation as the rate-limiting step


(k1) (c.f. above the direct oxidation at nakedelectrodes), and therefore, H−, H+, orH• are not involved in the rate-limitingstep. This was further supported by theobservation that changes in pH or bufferconcentration did not affect the rate ofoxidation of NADH by Fc+. Thus, theycould conclude that all their observationswere consistent with rate-limiting one-electron transfer from NADH to Fc+.They further concluded that k2[B] must begreater than k−1[Fc] for the reaction to berate-limited by the first electron donation.

The investigation was performed in anaqueous one-propanol solution in the pres-ence of perchlorate, which affects theEo′

of the ferrocenes and the reactionrate with NADH. To be able to drawkinetic-thermodynamic conclusions, val-ues for the second-order rate constant,kobs, and the Eo′

values were extrapolatedto a perchlorate concentration equal tozero. The Eo′

s of the investigated ferroce-nium/ferrocene derivatives range between+0.203 and 0.371 V versus SCE. Plots oflog kobs versus Eo′

revealed a straight linefor the ferrocene derivatives with less pos-itive Eo′

-values with a slope of 16.0 V−1

close to that predicted 16.6 V−1 for aprocess limited by the separation of thereactants, whereas for the most oxida-tive ferrocenes, deviations from linearitywere observed. The interpretation given isthat within the initial one-electron transfer,there are actually three processes that oc-cur, namely, (1) the reactants must diffusetogether to form an encounter complex,(2) an electron can then hop within thecomplex, and (3) finally, there must be dif-fusional separation of the products. For theweakest oxidizing ferrocenes, the actualrate-limiting process is therefore the sep-aration of the Fc/NADH+ pair, whereasfor the ferrocenes with more positive Eo′

s,

the electron hopping is the rate-limitingprocess.

In this investigation, an approximatevalue of the Eo′

of the NADH•+/NADHredox couple was also evaluated. Extrap-olating the linear log kobs versus Eo′

relationship to the diffusion-controlledrate-limited region, a value equivalent to0.808 V versus SCE was estimated for thissolvent. Recalculating the Eo′

value foraqueous solution, a value of 0.688 V ver-sus SCE could be given. With this valuein mind, the authors stated that the verypositive potential found is indicative of theinstability of NADH

•+. Only very powerfulone-electron oxidants will be able to oxidizeNADH by a mechanism pathway involvingone-electron transfer. Pulse radiolysis andlaser flash photolysis studies [42, 127] fur-ther support this as NADH•+ is completelydeprotonated within 1 µs, suggesting thatk2[B] is greater than 106 s−1 at pH 7 inaqueous solution. In addition, the pKa forNADH•+ could be estimated to be −3.5,which is in agreement with that estimatedfrom photochemical data in acetonitrileequal to −4 [39]. The value for deproto-nation has later been determined to be6 106 s−1 by Matsue and coworkers [128]who used pure aqueous buffers insteadof a mixture of water and one-propanol.These authors evaluated the Eo′

of theNADH•+/NADH to be 0.78 V versus SCE.

From these data, it is clear that theequilibrium of reaction (17) favors the left-hand side, but the existence of the fastirreversible reaction (18) allows reaction(17) to proceed in the forward direction.Reaction (19) is a homogeneous electron-transfer reaction. It proceeds irreversiblyas the potential of the NAD•/NAD+ coupleis at least 1.3 V more negative [140] thanthat of the Fc+/Fc couple, which in turnindicates that the equilibrium constant forreaction (19) is less than equal to 10−20.


4.3.2Two-electron-proton Acceptors

4.3.2.1 o-quinones and p-quinonesMiller and coworkers have also investi-gated in a series of papers the homoge-nous reaction with two-electron-protonacceptors, quinones and quinoneimines,both with electrochemistry and spec-troscopy [141–143]. Initially, they startedby reporting on the reaction betweenNADH and quinoneimines with electro-chemical techniques (see in the followingtext). The UV/vis spectroscopy was alsoemployed to examine the reaction kinet-ics and mechanism between NADH and7 different o-quinones and 14 p-quinonesin a pH interval between 6 and 8 [143].As these two-electron-proton acceptors arecapable of both electron transfer and hy-dride routes, the hope was to find evidencefor either of the two basically differentreaction pathways. Experiments run withdeaereated and oxygen-saturated solutionsyielded the same results revealing thatmolecular oxygen did not affect the reac-tion at all. They also compared the reactionrate with α- and β-NADH and two differentquinones and found it to be in agree-ment with earlier investigations [144] thatα-NADH has a higher reaction rate withthese mediators than β-NADH, attributedto differences in conformation between α-and β-NADH. In summary, the reactionof NADH with all quinones investigatedwas first order in each reactant with a sto-ichiometry of 1 : 1. Using singly or doublydeuterated NADH, no reduced quinonewas found with a deuterium incorporatedonto a hydroquinone carbon; however,there was a primary isotope effect and asubstantial substitution effect on the rate.Thus, there is strong evidence that the hy-drogen is being exchanged between carbon(NADH) and oxygen (hydroquinone).

To answer the question whether the re-action between NADH and quinone typeacceptors occurs according to a singlestep 1H− transfer or sequentially eitheras H − 1e− or as 1e− − H+ − 1e−, hasnot been easy to resolve. Miller andcoworkers [143] found that in general, theo-quinones reacted about 100 times morerapidly than p-quinones with comparableEo′

-values. No variation of the second-order reaction rate constant (kobs) with pHcould be found. However, there was a lin-ear correlation between log kobs and the Eo′

value of the Q/QH− redox couple at pH 7,Eo′

Q/QH−,pH7, rather than with the Q/QH2

redox couple for both o- and p-quinones.The slopes of the log kobs versus [NADH]for the two groups of quinones wereclose to identical (o-quinones; 16.4 V−1, p-quinones; 16.9 V−1), reflecting that whenthe reaction in both cases (o- and p-quinones) becomes more exothermic, one-half of the change in the thermodynamicvalue is reflected in the activation free en-ergy, which in using electrochemical termsis equal to α = 0.5. This could explain whyno variation of the reaction rate was foundwith pH as the Eo′

s of both NAD+/NADHand the Q/QH− redox couples will movewith 30 mV/pH-unit and thus the thermo-dynamic driving force will remain constantwith a change in pH. Thus, strong evidencewas presented showing that the initialstep was not electron transfer generatingNADH+ or proton transfer to solution butrather transfer of the NADH hydrogen tothe quinone oxygen in the rate-limitingstep. However, the investigation did notdecide whether the charge transfer oc-curred in a single one-hydride step or ina stepwise mode within an intermediatecharge-transfer complex.

At pH 7, the simple one-step hydridetransfer would occur according to thefollowing. In the first step, the hydride


is transferred from NADH to the quinoneyielding NAD+ and QH−.

NADH + Q −−−→ NAD+ + QH− (20)

This is followed by rapid protonation ofQH− .

QH− fast−−−→H+

QH2 (21)

A net hydride transfer could also occur se-quentially but within a complex accordingto two alternative reaction pathways pro-posed. In the first one [reactions (22)–(24)],the reaction starts with complex formationand donation of the first electron

NADH + Q ⇐⇒ [NADH•+Q−] (22)

which is followed in a rate-limiting step bythe transfer of the hydrogen atom

[NADH•+Q−]

rls−−−→ [NAD+QH−] (23)

and finally, by the dissociation of thecomplex;

[NAD+QH−] −−−→ NAD+ + QH− (24)

In an alternative pathway [reactions(25)–(27)], the formation of the complexis followed immediately by the transfer ofa hydrogen atom

NADH + Qrls−−−→ [NAD

•QH

•] (25)

followed by the transfer within the complexof the second electron

[NAD•QH

•] −−−→ [NAD+QH−] (26)

and finally, the formation of the products

[NAD+QH−] −−−→ NAD+ + QH− (27)

The general conclusion drawn by Millerand coworkers is that the charge trans-fer from NADH to the two-electron-proton acceptors does not occur ac-cording to an outer sphere stepwise

one-electron-transfer reaction but ratheraccording to a hydride transfer mecha-nism. However, as there is definite prooffor an intermediate charge transfer com-plex being formed between the two reac-tants (see the following text), the hydrideequivalent can be transferred stepwisewithin the complex and rate-limiting hy-drogen atom transfer followed by electrontransfer (H, e−) could not be ruled out.

In another series of detailed reports,Fukuzumi and coworkers [43–45, 47, 48,50, 55] have tried to find evidence foran e− − H+ − e− sequence occurring be-tween NADH analogues and quinones.They investigated the reaction betweena great number of dihydropyridine com-pounds (PyH2), being NADH modelcompounds, with p-benzoquinone deriva-tives (Q) in acetonitrile in the presenceand absence of Mg2+. In their inves-tigations, they have proven that whenPyH2 reacts with Q, a charge transfercomplex is formed [145,146]. Such com-plexes have been isolated and also theirspectra have been detected. The inves-tigated NADH model compounds couldbe divided into two separate groups:those equal to 1-(substituted benzyl)-1,4-dihydronicotinamide (X-BNAH) deriva-tives and N-methylacridan (AcH2). Severalpossible reaction steps were examinedthat would yield the strongest correlationbetween the evaluated second-order rateconstant, kobs, and the G through linear(log kobs vs. G) relationships. This neces-sitated the calculation of G for a numberof possible reaction steps.

In Sch. 2, possible reaction steps arerevealed for which expressions for G

needed to be found. The Goet denotes the

Gibbs free energy change for the transferof the first electron, Go

H+ for the transferof the proton, Go

et′ for the transfer ofthe second electron, Go

H for the transfer


PyH2 + Q

PyH2•+ + Q•−

PyH+ + QH−

PyH• + QH•∆G°H+

∆G°H−

∆G°H∆G°et′∆G°et

Scheme 2 Proposed reactionpathways between dihydro-pyridine compounds (PyH2)

used as NADH modelcompounds and quinonederivatives.

of the hydrogen atom, and GoH− for the

transfer of the hydride. The expressionsrelating these parameters are given below:

Goet

F= Eo′

(PyH

•+2

PyH2

)− Eo′

(Q

Q•−)(28)

Goet′

F= Eo′

(PyH+

PyH•

)− Eo′

(QH•

QH−)(29)

GoH+ = 2.3RT

[pKa

(PyH

•+2

PyH•

)

− pKa

(QH

•

Q•−)]

(30)

The Eo′values of the PyH2

•+/PyH2and PyH+/PyH• were determined byCV, whereas the Eo′

values for Q/Q•−and QH•/QH− were taken from theliterature data. The addition of Mg2+was motivated as these ions can formcomplexes with either PyH2 and/or Q, thusaffecting the oxidation potential of PyH2(PyH2

•+/PyH2) or reduction potential of(Q/Q•−). Necessary pKa values could alsobe calculated. The final missing G

values, GoH and Go

H− for correlationsfor hydrogen atom and single hydridetransfer can then easily be calculated:

GoH = Go

et + GoH+ (31)

GoH− = Go

et + GoH+ + Go

et′ (32)

When plotting log kobs versus GoH−

(equivalent to a hydride transfer), two lin-ear relationships were found with differentintercepts: one for the X-BNAH deriva-tives and another for AcH2. The slopes ofthese straight lines were almost equal andshow that as the reaction becomes moreexothermic, about half of the change inGo

H− is reflected in the activation bar-rier of the hydride-transfer reaction, thatis, α = 0.5, which is in accordance withthe results of Miller and coworkers (seepreceding text). On the other hand, whenplotting log kobs versus Go

H, only onesingle straight correlation was found forall investigated X-BNAH derivatives in-cluding AcH2. The slope was larger bya factor of 2 than the slopes for the logkobs versus Go

H− plots, indicating thatthe change in Go

H is directly reflectedin the activation barrier of the hydride-transfer reactions, that is, α is close tounity (0.91). Thus, a single correlation oflog kobs versus Go

H for both all X-BNAHderivatives and AcH2, in combination withthe different intercepts between X-BNAHand AcH2 in the plots of log kobs versusGo

H− (= GoH + Go

et′ ), reveals that theactivation barrier of the hydride-transferreactions from PyH2 to Q is dependenton only Go

H and thereby independentof Go

et′ ; the different correlations of thelog kobs versus Go

H− plots between X-BNAH and AcH2 are caused mainly by the


difference in the reduction potentials be-tween X-BNAH/X-BNAH• (−1.08 V) andAcH/AcH• (−0.43 V). The apparent cor-relation of log kobs versus Go

H− mayhold, provided that Go

et′ is correlatedwith Go

H in a homologous series of re-actions. Thus, as Go

H consists of Goet

and GoH+ , the mechanistic question is

reduced to (I) a direct transfer of a hy-drogen atom from PyH2 to Q, followedby an exothermic electron transfer fromPyH• to QH− or (II) an electron transferfrom PyH2 to Q in the activation processand the subsequent proton transfer fromPyH2

•+ to Q•−, followed by an exother-mic electron transfer from PyH• to QH•.To answer that question, Fukuzumi andcoworkers investigated whether there wereprimary kinetic isotope effects (kH/kD) asthey may provide useful information onwhether an actual transfer of hydrogennucleus occurs as a form of proton orhydrogen atom. Indeed, primary isotopeeffects were found. In the case of theelectron-transfer activation mechanism, akinetic isotope effect is expected to be at-tributed only to the proton transfer fromPyH2

•+ to Q•− as it is not expected for theelectron-transfer process from PyH2 to Qor from PyH• to QH•. Strong linear cor-relation was found between log (kH/kD)

and (GoH+ /F)2 in agreement with the

Marcus equation, whereas no linear cor-relation was found between log (kH/kD)

and (GoH/F)2, indicating that the pri-

mary isotope effect was not correlated withGo

H. A hydride transfer would requirethat kH/kD should remain constant witha linear change in Go

H− for the casewhere α = 0.5. It could thus be concludedthat an actual transfer of hydrogen nucleusoccurs as a form of proton and the electron-transfer activation mechanism seems to bethe most plausible one to explain boththe variation of the rate constants and

the primary kinetic isotope effects in thehydride-transfer reactions from PyH2 toQ. Furthermore, it was shown that a singlecorrelation between log kobs and Go

et/Ffor different X-BNAH, AcH2, and NADH(whether in acetonitrile or water) in theabsence and presence of Mg2+ prevails in-dicating that the activation barrier of thehydride-transfer reactions is well corre-lated with the energetics of the electrontransfer Go

et. However, the investiga-tion also included studies of one-electronnonproton acceptors and the results fromprevious investigations on the use of suchoxidants were also used for calculations. Asa general remark, comparing the values ofkobs obtained for one-electron-nonprotonacceptors and for the quinones (hydride ac-ceptors) with similar Eo′

values, one findsthat it is greater for the quinones. Thisfact has been used as evidence againstan electron-transfer process involvementin the hydride-transfer reaction by assum-ing that the work term, ωp, of the radicalion pair produced upon electron transfercan be neglected. However, as the reactionbetween NADH and its analogues occurswithin a CT-complex through an innersphere electron-transfer mechanism, thework term, ωp, of the radical ion pair ofopposite charges is expected to be muchlarger than for an outer sphere reactionmechanism and thus cannot be neglected.Sch. 3 reveals the difference between anouter sphere and an inner sphere reac-tion mechanism and also the relationshipbetween Go

et and Go23, that is, the free

energy change for the initial electron trans-fer occurring within the CT-complex, thesecond step in reaction (33). ωp and ωr

are the work terms required to bring theproducts (red•+ and ox•−) and the reac-tants (red∗ and ox) together to the meanseparation in the activated complex, whichare largely coulombic. ωr may be neglected


red* + ox

[red*ox]

red•* + ox•−

[red•*ox•−]∆G23

∆G23 = ∆G°et + ωp − ωr

∆G°et

ωpωr

Scheme 3 Reactionmechanisms and relationshipbetween the Go

et and Go23

within the CT-complex.

as the reactants in the investigations in-cluded neutral species.

An expression for the rate constant, kobs,based on these assumptions could thusbe derived and when compared with ex-perimental values, a good correlation wasfound.

The authors thus concluded that themost plausible reaction path would be ac-cordingly what is outlined in reaction (33):

PyH2 + Qk12⇐⇒k21

[PyH2Q]k23⇐⇒k32

[PyH2•+Q

•−]kH+−−−→ [PyH

•QH

•]

fast−−−→ PyH+ + QH− (33)

Miller and Valentine [147] in a later paperwere sceptical of all the conclusions drawnby Fukuzumi and coworkers [47] and sug-gested that an appropriate and challengingview is that one-step hydride transfershould not be replaced by the more compli-cated three-step electron-proton-electroninner sphere mechanism until such evi-dence appears that unequivocally discrim-inates between the two pathways.

4.3.2.2 Aromatic DiiminesIn addition to the extensive investigationsof the reaction between NADH and vari-ous quinones and one-electron acceptorsin solution, Miller and coworkers also

investigated the reaction between NADHand 14 diaminobenzenes and 7 diaminopy-rimidines [141, 142]. These investigationswere performed in aqueous solutions andwith GC electrodes. The background andinterest lie in the expected high reactiv-ity of oxidized diamines as iminium ionsare known to be better hydride acceptorsthan ketones and the mechanism has anal-ogy in the flavin oxidation of NADH. Theinvestigation concluded that the numberof electrons transferred was equal to two,that 1,2-diimines are faster oxidants forNADH than 1,4-diimines, and that thetransfer of the 4-hydrogen of NADH occursto a protonated diimine. No real correla-tion could, however, be found betweenlog kobs and the Eo′

for these mediators.Also of significance is the finding that pro-tonated diimines (i.e., positively charged)are more reactive than the correspondingneutral diimines. This is the first instanceof positively charged mediators that aredemonstrated superior to the neutral oneshaving otherwise the same catalytic func-tionality and Eo′

.It is thus possible to come up with

some general remarks on the reac-tivity of possible oxidants for NADH.Two-electron-proton acceptors are betteroxidants than one-electron-nonproton ac-ceptors, aromatic diimines are better thanquinones and oxidized aminophenols,oxidants with the catalytic functionality

4.4 Electrocatalytic Oxidation of NAD(P)H at Mediator-modified Electrodes 89

located at the 1,2-position seem better thanthe corresponding ones having the cat-alytic functionality located in 1,4-position,and finally, positively charged mediatorsare better than neutral ones.

The need to be able to catalyze the elec-trochemical reactions of both NAD+ andNADH at substantially lower applied po-tentials than at naked electrodes and withless or preferably no side reactions forany practical applications (analysis, fuelcells, electroorganic synthesis) has put fo-cus on finding catalytic redox molecules(mediators) that can accomplish this goal.The use of mediators in bioelectrochem-istry is today well established both forfundamental and applied purposes [23,148–150]. One can in principle, dividethe group of redox mediators into twoseparate groups: one-electron-nonprotonacceptors/donors and two-electron-protonacceptors/donors. In recent years, in ap-plied aspects of bioelectrochemistry, forexample, in enzyme-based amperometricbiosensors, focus has been on the use ofthe one-electron acceptors/donors ratherthan on the two-electron-proton accep-tors/donors to facilitate shuttling of theelectrons from/to the cofactor of redox en-zymes and electrodes for the followingreasons [150]: (1) as no protons partici-pate in their redox conversion, the Eo′

remains constant with a change in pH and(2) there are no radical intermediates inthe redox reaction prone to abortive sidereactions. In contrast, the two-electron-proton acceptors/donors suffer from bothof these two drawbacks in aqueous solu-tions. Thus, it is easily seen in the Marcusexpression [151, 152] that a change in pHcan greatly affect the second-order rateconstant (kET) between the redox media-tor of the two-electron-proton type and theenzyme cofactor, which in turn would inthe case of an enzyme-based biosensor,

affect the response signal.

kET = exp−β(d−d0) exp−(Go+λ)/4RTλ (34)

where β is the attenuation factor, d isredox-to-redox center distance, d0 is equalto 3.6 A, λ is the reorganization factor,and Go the free energy change betweenthe two reactants (in this case, betweenmediator and NAD+ or NADH). How-ever, neither the mediated oxidation ofNADH nor the mediated reduction ofNAD+ (see in the following text) canbe accomplished rapidly and continuouslywith redox compounds catalyzing a sim-ple electron-transfer reaction in contrastto, for instance, flavin-containing oxidasesand dehydrogenases, where the radical in-termediates of the bound flavin moleculeare stabilized within the protein globuleby adjacent amino acid residues. Thus, forboth electrocatalytic NADH oxidation andNAD+ reduction, one has to look for redoxmediators, which are able to accept/donatea hydride equivalent, either in one sin-gle reaction step or sequentially with thecofactor, and moreover, are able to be elec-trochemically converted into a sequentialmode (e− − H+ − e−) with the electrodeand the contacting solution as the electrodeis not likely to provide or to consume theproton. However, surface functionalitiesmay facilitate proton transfer, see Fig. 4.

4.4Electrocatalytic Oxidation of NAD(P)H atMediator-modified Electrodes

4.4.1General Remarks of CMEs

4.4.1.1 o-Quinones and PhenylenediimineDerivativesThe use of mediators as ‘‘depolarizing cat-alysts’’ in electrochemistry goes back to


the beginning of this century. The drivingforce to study the electrocatalytic oxidationreactions of NAD(P)H at electrodes inten-tionally modified, starting in the late 1970s,was based on several observations. The de-liberate immobilization of compounds onelectrode surfaces so that after immobiliza-tion, the electrode not only possesses itsoriginal properties but also those of the im-mobilized species was a relatively new areaof research [153]. Bioelectrochemistry andelectrochemistry in biology/medicine alsoexperienced a dramatic increase in interestderived from the concept of enzyme-basedamperometric biosensor in the 1960s. In1977, the first two publications on ef-ficient direct electron-transfer reactionswith a small redox protein, cytochrome c,were reported [154, 155], rapidly followedby indirect electrochemical observation ofdirect electron transfer also with largerredox proteins with enzyme activity, lac-case [156, 157], and HRP (horseradishperoxidase) [158] and the simultaneous in-terest in and increased knowledge of redoxmediators to study biological redox sys-tems [148]. It is therefore not surprisingthat during the same period, the first CMEthat was able to drastically reduce the largeover voltage for electrochemical NADH ox-idation appeared [159]. Since then, it hasbeen the topic of many reports summa-rized in a number of review papers [76, 77,149, 160–166]. Table 2 summarizes theextensive work done on CMEs for electro-catalytic NAD(P)H oxidation.

The first paper on a CME for elec-trocatalytic NADH oxidation reported byTse and Kuwana in 1978 [159] was basedon two primary amine containing o-quinone derivatives, dopamine or 3,4-dihydroxybenzylamine, that could be co-valently immobilized onto the surface ofcyanuric chloride–activated GC electrodesforming a monolayer on the electrode

surface. When immobilized, these redoxcompounds revealed from CV, Eo′

valuesof around +0.160 V versus SCE at pH 7and the anodic peak potential in the pres-ence of NADH at around +0.2 V. Thus,the over voltage could be reduced by some0.4 V. When running CV, an ECE mech-anism was shown for the electrocatalyticreaction. What was also important in theirstudy was that they could confirm thatenzymatically NAD+ was produced as theproduct of the electrocatalytic reaction. Theelectrode also showed a high catalytic effi-ciency for the electrooxidation of ascorbate.However, there was a substantial differ-ence in the stability of the surface-tetheredo-quinone when catalyzing the oxidationof NADH or the oxidation of ascorbate.For NADH, the electrocatalytic propertyvanished after a few cycles, whereas forascorbate, it seemed stable for a long term.

In a follow-up paper from Kuwana’sgroup [167], an o-quinone derivative at-tached to a larger aromatic derivative (4-(2-(1-pyrenyl)vinyl)catechol) that stronglyadsorbed on graphite was studied. Whenthe mediator was kept in its reduced state(catechol), it was virtually stable for along term. When kept in its oxidized state(o-quinone), some minor deactivation oc-curred with time, whereas when kept ata potential close to the Eo′

, a much morerapid loss of electroactivity was noticed, ex-plained by abortive side reactions causedby an initial reaction between catecholand o-quinone. In the presence of NADH,an even much faster deactivation processoccurs, presumably caused by an interme-diate semiquinone assumed to react witha radical intermediate of the coenzyme,NADH•+, to give an inactive compound,which poisons and blocks off the surface(Fig. 7).

Since the first paper by Tse and Kuwana,there have been numerous publications


Tab.

2C

hem

ical

lym

odifi

edel

ectr

odes

for

elec

troc

atal

ytic

NA

D(P

)Hox

idat

ion

base

don

inor

gani

can

dor

gani

cm

edia

tors

Med

iati

ngsc

hem

eE

o′ /V

vs.r

ef.

Imm

obili

zati

onR

ate

cons

tant

Sens

itiv

ity

Ref

eren

ces

M−1

s−1

µAm

M−1

cm−2

CA

RB

ON

ELEC

TRO

DES

Qui

none

s0

–0.

25V

vs.S

CE,

pH7

Oxi

dize

dgr

aphi

te,g

lass

yca

rbon

800

132,

134

–13

7,16

9–

174

Qui

none

s0.

2–

0.4

Vvs

SCE

Pret

reat

edca

rbon

fiber

101

–10

5,17

5So

l-gel

deri

ved,

cera

mic

-car

bon

elec

trod

es0.

15V

vs.

Ag[

AgC

l]KC

l(s

atd.

),pH

7.3

176

Alc

ohol

dehy

drog

enas

e(A

DH

)-N

AD

+ -mod

ified

carb

onbl

ack

−0.0

4V

vs.S

CE,

pH7.

6A

DH

+N

AD

+ad

sorb

edon

carb

onbl

ack

177,

178

o-Q

UIN

ON

ES

3,4-

Dih

ydro

xybe

nzyl

amin

e(e

ugen

ol)

0.17

Vvs

.SC

E,pH

7C

oval

ent

ongl

assy

carb

on7.

710

4(h

omog

eneo

us)

159,

179

Dop

amin

e0.

13V

vs.S

CE,

pH7

Cov

alen

ton

glas

syca

rbon

3.6

104

(hom

ogen

eous

)15

9D

opam

ine

0.19

Vvs

.Ag/

AgC

l,pH

7El

ectr

oche

mic

alim

mob

iliza

tion

ongl

assy

carb

on

180

Dop

amin

e0.

2V

vs.A

g/A

gCl,

pH7

Inco

rpor

ated

inlip

idfil

mca

ston

togl

assy

carb

on18

1

Imin

ede

rive

dfr

om3,

4-di

hydr

oxyb

enza

ldeh

yde

and

4-am

inop

yrid

ine

0.34

Vvs

.Ag/

AgC

l,pH

7A

dsor

ptio

n-se

lfas

sem

bly

onPt

0.2

mM

to2.

0m

M18

2

Dop

amin

e/3,

4-di

hydr

oxyb

enzo

icac

id0.

15V

vs.S

CE,

pH7

Cov

alen

tlybo

und

tose

lf-as

sem

bled

cyst

eam

ine

mon

olay

ers

onA

u

0.1

mM

to1.

0m

M18

3,18

4

(con

tinue

dov

erle

af)

92 4 Electrochemistry of NAD(P)+/NAD(P)HTa

b.2

(con

tinue

d)

Med

iati

ngsc

hem

eE

o′ /V

vs.r

ef.

Imm

obili

zati

onR

ate

cons

tant

Sens

itiv

ity

Ref

eren

ces

M−1

s−1

µAm

M−1

cm−2

Poly

(met

hacr

yloy

lchl

orid

e)w

ithdo

pam

ine

0.16

0V

vs.S

CE,

pH6.

35C

ast

ongl

assy

carb

on18

5

Poly

(acr

yloy

lchl

orid

e)or

poly

(met

hacr

yloy

lchl

orid

e)w

ithdo

pam

ine

0.17

Vvs

.SC

E,pH

7C

ast

ongl

assy

carb

on18

6

Dop

amin

eco

ntai

ning

poly

mer

0.17

Vvs

.SC

E,pH

7C

ast

ongl

assy

carb

onD

iffus

ion

cont

rolle

d18

7

Poly

(3,4

-Dih

ydro

xybe

nzal

dehy

de)

poly

(2,3

-DH

B)

0.12

–0.

17V

vs.

SSC

E,pH

7el

ectr

opol

ymer

ized

onG

C,

4.3

103

(3,4

-DH

B,

pH7)

0.01

–1.

2,0.

01–

0.9

mM

188

–19

2

carb

onfe

lt/ep

oxy

com

posi

te,e

ffect

ofM

g2+,

Ca2+

1.3

103

(3,4

-DH

B,

pH8.

5)

6.2

103

(3,4

-DH

B,

pH8.

5,20

mM

Mg2+

)10

.210

3(3

,4-D

HB

,pH

8.5,

20m

MM

g2+)

4-M

ethy

l-o-q

uino

ne0.

07V

vs.S

CE,

pH8.

2A

dsor

bed

onsc

reen

prin

ted

carb

on19

3

p-(4

-Vin

ylpy

ridy

l)-N

-(3,

4-di

hydr

oxyb

enzy

liden

e)am

inob

enze

ne

0.31

Vvs

.SC

E,pH

∼9.2

.In

corp

orat

edin

toth

em

erca

ptan

mon

olay

eron

the

elec

trod

e

0.1

–2

mM

194

Chl

orog

enic

acid

(1,3

,4,5

-tet

rahy

dro-

xycy

cloh

exan

e-ca

boxy

licac

id3-

(3,4

-dih

ydro

xyci

nnam

ate)

)

0.22

3V

vs.A

g/A

gCl,

pH7

Elec

trod

epos

ited

onG

C8.

510

3(0

.05

mM

NA

DH

)0.

1–

1.0

mM

195


4-[2

-(2-

Nap

hthy

l)vi

nyl]

cate

chol

4-[2

-(9,

10-

etha

noan

thra

cen-

9-yl

)vi

nyl]c

atec

hol

0.17

Vvs

.SC

EA

dsor

bed

ongr

aphi

te2

106

196

4-[2

-(1-

Pyre

nyl)

viny

l]cat

echo

l,4-

[2-(

1-py

reny

l)et

hano

]cat

echo

l0.

10V

vs.S

CE

Ads

orbe

don

grap

hite

110

416

7

1,2-

Nap

htho

quin

one

−0.1

5V

vs.S

CE

Mix

edw

ithca

rbon

past

e19

72,

3-D

ichl

oro-

1,4-

naph

thoq

uino

neco

vale

ntly

boun

dto

N-s

ubst

itute

spy

rrol

e

−0.1

5V

vs.S

CE,

pH7

Elec

trop

olym

eriz

edon

carb

on16

4,19

8

Mer

capt

ohyd

roqu

inon

epH

6.35

Elec

trop

olym

eriz

edon

togl

assy

carb

on19

9

Pyrr

oloq

uino

line

quin

one

(PQ

Q)

−0.1

5V

vs.S

CE

Cov

alen

tbi

ndin

gto

cyst

amin

ean

dcy

stea

min

em

odifi

edgo

ld/N

ε-2

,4-d

initr

ophe

nyl

lysi

neon

gold

2.3

102

(no

Ca2+

)20

0–

203

3.4

104

(20

mM

Ca2+

)PQ

Q−0

.15

Vvs

.SC

E,pH

7PQ

Q-s

piro

pyra

nm

onol

ayer

onA

u20

4,20

5

PQQ

∼0V

vs.S

CE,

pH8.

1Im

mob

iliza

tion

ofPQ

Qin

elec

trop

olym

eriz

edfil

ms

ofam

inob

enze

neis

omer

son

gold

,pl

atin

uman

dca

rbon

206

(con

tinue

dov

erle

af)


b.2

(con

tinue

d)

Med

iati

ngsc

hem

eE

o′ /V

vs.r

ef.

Imm

obili

zati

onR

ate

cons

tant

Sens

itiv

ity

Ref

eren

ces

M−1

s−1

µAm

M−1

cm−2

PQQ

-NA

D+

−0.1

6V

vs.S

CE,

pH8

Cov

alen

tat

tach

men

tof

PQQ

toa

cyst

amin

em

onol

ayer

onA

u,fo

llow

edby

cova

lent

linka

geof

N-(

2-am

inoe

thyl

)-N

AD

+

203,

207

–20

9

Tran

sitio

nm

etal

com

plex

esco

ntai

ning

1,10

-Phe

-na

nthr

olin

e-5,

6-di

one

Liga

nds

0V

vs.S

CE

Ads

orbe

don

glas

syca

rbon

(6.2

±0.

6)10

321

0

[Re(

phen

-dio

ne)

(CO

) 3C

l]an

d[F

e(ph

en-d

ione

) 3](

PF6)

2

−0.0

5V

vs.S

SCE,

pH8,

Mix

edw

ithca

rbon

past

e21

1

−0.0

3V

vs.S

SCE,

pH8

Os(

4,4′

-dim

ethy

l-2,2

′ -bip

yrid

ine)

2(1

,10-

phen

anth

rolin

e-5,

6-di

one)

0.03

2V

vs.A

g/A

gCl,

pH7

Ads

orbe

don

grap

hite

1.9

103,a

tpH

6.1

212

Os(

4,4′

-dim

ethy

l-2,2

′ -bip

yrid

ine)

2(1

,10-

phen

anth

rolin

e-5,

6-di

one)

0V

vs.S

CE,

pH6

Mix

edw

ithca

rbon

past

e14

0µA

cm−2

Imm

ol−1

213

Hom

olep

ticca

tech

ol-p

enda

ntte

rpyr

idin

eco

mpl

exes

[M(L

2)2]

2+(M

=C

o,C

r,Fe

,N

i,R

u,an

dO

s,an

dL2

=4′

-(3,

4-di

hydr

oxyp

heny

l)-

2,2′

:6′

,2′′

-ter

pyri

dine

)

∼0.1

0.3

Vvs

.A

g/A

gCl,

pH7

Elec

trod

epos

ition

onG

Can

dPt

214


p-Q

UIN

ON

E

Chl

oran

il+0

.050

Vvs

.SC

E,pH

7C

oval

ent

ongr

aphi

te16

4,21

5

Chl

oran

ilco

vale

ntly

boun

dto

N-s

ubst

itute

spy

rrol

e+0

.1V

vs.S

CE

Elec

trop

olym

eriz

atio

non

carb

on16

4,19

8

p-B

enzo

quin

one

+0.1

15V

vs.

Ag/

AgC

l,pH

7.3

Imm

obili

zed

ongr

aphi

te2.

910

61

–9

mM

216

p-PH

ENYL

ENED

IIM

INE

p-Ph

enyl

ened

iam

ine,

N,N

,N′ ,N

′ -te

tram

ethy

l-p-

phen

ylen

edia

min

e

0.10

Vvs

.SC

EM

ixed

into

carb

onpa

ste

217

Var

iam

ine

blue

B0.

095

Vvs

.Ag/

AgC

l,pH

7A

dsor

bed

onto

para

ffin

impr

egna

ted

grap

hite

0.00

910

421

8

Mel

dola

blue

−0.1

75V

vs.S

CE

Ads

orbe

don

grap

hite

310

416

0,21

9–

225

Mel

dola

blue

−0.1

2V

vs.

Ag/

AgC

l,pH

7A

dsor

bed

onto

para

ffin

impr

egna

ted

grap

hite

3.9

104

218

Mel

dola

blue

−0.1

32V

vs.

Ag/

AgC

l,pH

7.4

Imm

obili

zed

asdi

amm

inet

etra

(iso

thio

-cy

anat

o)ch

rom

ates

(Rei

neck

ates

)in

grap

hite

-epo

xyco

mpo

site

elec

trod

es

5.83

103

0.5

µM–

3m

M22

6

Mel

dola

blue

−0.0

7V

vs.S

CE,

pH7

Mix

edin

toca

rbon

past

e22

7,22

8

Mel

dola

blue

−0.1

4–−0

.085

Vvs

.Ag/

AgC

lpH

7.5

Imm

obili

zed

inel

ectr

opol

ymer

ized

film

sof

poly

pyrr

ole

with

enzy

me

(AD

H,L

DH

)an

dN

AD

+

229,

230

Mel

dola

blue

−0.0

5V

Thic

k-fil

mca

rbon

elec

trod

e23

1,23

2

(con

tinue

dov

erle

af)


Tab.

2(c

ontin

ued)

Med

iati

ngsc

hem

eE

o′ /V

vs.r

ef.

Imm

obili

zati

onR

ate

cons

tant

Sens

itiv

ity

Ref

eren

ces

M−1

s−1

µAm

M−1

cm−2

Mel

dola

blue

−0.0

3V

vs.S

CE,

pH7

Imm

obili

zed

ontit

aniu

mph

osph

ate

coat

edon

toa

silic

age

lsur

face

mix

edw

ithca

rbon

past

e

1.0

10−5

–5.

010

−5M

233

Mel

dola

blue

∼−0.

1V

vs.S

CE

Imm

obili

zed

intit

ania

-,zi

rcon

ia-s

ol-g

elca

rbon

com

posi

teel

ectr

odes

210

−6–

110

−323

4

Mel

dola

blue

0.11

Vvs

.Ag/

AgC

l,pH

7.3

Imm

obili

zed

inso

l-gel

,m

ixed

into

carb

onpa

ste

0.5

–10

mM

176

Mel

dola

blue

−0.0

9–

0.05

Vvs

.SC

E,pH

7A

dsor

bed

onZ

r-ph

osph

ate

mix

edw

ithca

rbon

past

e39

mA

M−1

(10

–20

0µM

)23

5

Ran

dom

bloc

km

ethy

l-silo

xane

poly

mer

cont

aini

ngM

eldo

labl

ue

−0.1

5V

vs.S

CE,

pH7

Cas

ton

togr

aphi

te40

µA cm−2

mM

−116

2,23

6,23

7

3-A

nilin

o-M

eldo

abl

ue−0

.36

Vvs

.SC

E,pH

7A

dsor

bed

ongr

aphi

te3

103

224

Azu

reI

−0.2

25V

vs.

Ag/

AgC

l,pH

7A

dsor

bed

onto

para

ffin

impr

egna

ted

grap

hite

0.63

104

218

Azu

reA

∼−0

.01

Vvs

.A

g/A

gCl,

pH7

Cov

alen

tlybo

und

onto

cyst

amin

ean

dcy

stei

nm

odifi

edgo

ld

238

Azu

reA

0.21

Vvs

.Ag/

AgC

l,pH

7C

oval

ent

boun

dto

self-

asse

mbl

edm

onol

ayer

of3,

3′-d

ithio

bis(

succ

inim

i-dy

lpro

pion

ate)

onA

u

5.5

104

239


Poly

(Azu

reA

)/po

ly(A

zure

I)0.

01V

vs.A

g/A

gCl,

pH7/

0.04

Vvs

.A

g/A

gCl,

pH5.

4

Elec

trop

olym

eriz

edon

GC

0.3

AM

−124

0–

242

Azu

reII

−0.2

20V

vs.

Ag/

AgC

l,pH

7A

dsor

bed

onto

para

ffin

impr

egna

ted

grap

hite

0.64

104

218

Azu

reC

0.23

Vvs

.Ag/

AgC

l,pH

7C

oval

ent

boun

dto

self-

asse

mbl

edm

onol

ayer

of3,

3′-d

ithio

bis(

succ

ini-

mid

ylpr

opio

nate

)on

Au

9.5

104

239

Met

hyle

negr

een

−0.1

15V

vs.

Ag/

AgC

l,pH

7A

dsor

bed

ongr

aphi

te/A

dsor

bed

onto

para

ffin

impr

egna

ted

grap

hite

1.9

104

218,

243

Met

hyle

negr

een

−0.1

Vvs

.SC

EM

ixed

into

carb

onpa

ste

228,

241,

242

Met

hyle

negr

een

0.12

Vvs

.SC

E,pH

7A

dsor

bed

onZ

r-ph

osph

ate

mix

edin

toca

rbon

past

e23

5

Poly

(Met

hyle

negr

een)

0.11

2V

vs.A

g/A

gCl,

pH5.

4El

ectr

opol

ymer

ized

onG

C1.

1A

M−1

241,

242,

244

Nile

blue

−0.4

20V

vs.S

CE,

pH7

Ads

orbe

don

GC

/gr

aphi

te/a

dsor

bed

onA

g

1016

0,22

1,22

4,24

5–

247

Nile

blue

−0.4

05V

vs.

Ag/

AgC

l,pH

7A

dsor

bed

onto

para

ffin

impr

egna

ted

grap

hite

0.03

410

421

8

Nile

blue

−0.0

2–−0

.04

Vvs

.SC

E,pH

7(Z

r),

−0.0

3V

vs.S

CE,

pH7

Ads

orbe

don

Zr-

,Ti

-pho

spha

tem

ixed

into

carb

onpa

ste

11m

AM

−1(1

0–

120

µM)

235,

248

–25

0

(con

tinue

dov

erle

af)


Tab.

2(c

ontin

ued)

Med

iati

ngsc

hem

eE

o′ /V

vs.r

ef.

Imm

obili

zati

onR

ate

cons

tant

Sens

itiv

ity

Ref

eren

ces

M−1

s−1

µAm

M−1

cm−2

Nile

blue

∼−0

.12

Vvs

.A

g/A

gCl,

pH7

Cov

alen

tlybo

und

onto

cyst

amin

ean

dcy

stei

nm

odifi

edgo

ld

238

3-β

-Nap

htho

yl-n

ilebl

ue−0

.220

Vvs

.SC

E,pH

7A

dsor

bed

ongr

aphi

te5

104

164,

224,

245,

251

–25

4

Bis

(3,3

-nile

blue

)-te

reph

thoy

l−0

.20

Vvs

.SC

E,pH

7A

dsor

bed

ongr

aphi

te2

104

255,

256

Poly

(nile

blue

)−0

.07

Vvs

.SC

E,pH

6.8

Elec

trop

olym

eriz

edon

GC

6.3

102

257

3-α

-Pyr

enyl

iden

e-ni

lebl

ueA

dsor

bed

ongr

aphi

te25

5Ph

enot

hiaz

ine

0.28

Vvs

.Ag/

AgC

l,pH

7A

dsor

bed

onto

para

ffin

impr

egna

ted

grap

hite

0.71

104

218

Tolu

idin

ebl

ueO

−0.2

85V

vs.S

CE,

pH7

Ads

orbe

don

grap

hite

/ano

dize

dgl

assy

carb

on/w

axim

preg

nate

dgr

aphi

te

221,

224,

245,

258

–26

0

Tolu

idin

ebl

ueO

−0.2

1V

vs.

Ag/

AgC

l,pH

7A

dsor

bed

onto

para

ffin

impr

egna

ted

grap

hite

1.01

104

218

Tolu

idin

ebl

ueO

−0.0

6−

0.04

Vvs

.SC

E,pH

7(Z

r),

Ads

orbe

don

Zr-

,Ti-

phos

phat

em

ixed

with

carb

onpa

ste

12m

AM

−123

5,24

9,25

0,26

1,26

2

−0.0

6V

vs.S

CE,

pH7

(Ti)

(10

−20

0µM

)

Tolu

idin

ebl

ueO

−0.2

5V

vs.S

CE,

pH7

Cov

alen

tlybo

und

onto

carb

onfib

re4.

010

−5−

1.5

10−3

M26

3


Tolu

idin

ebl

ueO

∼−0

.2V

vs.S

CE,

pH7

Mix

edin

toca

rbon

past

e0.

05–

1.25

mM

264,

265

Tolu

idin

ebl

ueO

Mon

omer

:−0.

125

Vvs

.Ag/

AgC

l,pH

7C

oval

ently

boun

don

togo

ld23

8,26

6/20

0

(dim

er:+

0.13

5V

vs.A

g/A

gCl,

pH7)

Poly

(tol

uidi

nebl

ueo)

0.05

Vvs

.Ag/

AgC

l,pH

5.4

Elec

trop

olym

eriz

edon

GC

/gra

phite

241,

242,

267,

268

Poly

mer

cont

aini

ngco

vale

ntly

boun

dto

luid

ine

blue

O−0

.1±

0.1

Vvs

.SC

E,pH

7R

VC

,gra

phite

,mix

edin

toca

rbon

past

e26

9–

277

3-β

-Nap

htho

yl-t

olui

dine

blue

O−0

.135

Vvs

.SC

E,pH

7A

dsor

bed

ongr

aphi

te1.

410

422

4,24

5

Met

hyle

nebl

ue−0

.21

Vvs

.A

g/A

gCl,

pH7

Ads

orbe

don

para

ffin

impr

egna

ted

grap

hite

0.8

104

218

Met

hyle

nebl

ue0.

08–

0.1

Vvs

.SC

E,pH

7(Z

r),

Ads

orbe

don

Zr-

,Ti-

phos

phat

em

ixed

with

carb

onpa

ste

28m

AM

−123

5,24

8,24

9

0.02

Vvs

.SC

E,pH

7(T

i)(1

0–

200

µM)

Poly

(met

hyle

nebl

ue)

0.25

Vvs

.SC

E,pH

7C

oele

ctro

poly

mer

ized

with

pyrr

ole

onG

CN

AD

H2.

5×

10−8

278

NA

DPH

4.0

×10

−8Po

ly(m

ethy

lene

blue

)−0

.04

Vvs

.SC

E,pH

7El

ectr

opol

ymer

ized

onto

ala

poni

tege

l27

9

Poly

(met

hyle

nebl

ue)

0.01

5V

vs.A

g/A

gCl,

pH5.

4El

ectr

opol

ymer

ized

onG

C0.

6A

M−1

241,

242

(con

tinue

dov

erle

af)


Tab.

2(c

ontin

ued)

Med

iati

ngsc

hem

eE

o′ /V

vs.r

ef.

Imm

obili

zati

onR

ate

cons

tant

Sens

itiv

ity

Ref

eren

ces

M−1

s−1

µAm

M−1

cm−2

Poly

(met

hyle

nebl

ue)

0.1

Vvs

.SC

E,pH

9El

ectr

opol

ymer

ized

onsc

reen

prin

ted

gold

280

Bri

llian

tcre

sylb

lue

−0.3

4V

vs.S

CE,

pH7

Ads

orpt

ion

ongr

aphi

te22

4,25

9

Bri

llian

tcre

sylb

lue

−0.2

80V

vs.

Ag/

AgC

l,pH

7A

dsor

bed

onpa

raffi

nim

preg

nate

dgr

aphi

te0.

024

104

218

Bri

llian

tcre

sylb

lue

−0.0

2V

vs.S

CE,

pH7

(Zr)

,A

dsor

bed

onZ

r-,T

i-ph

osph

ate

mix

edw

ithca

rbon

past

e

235,

249

−0.1

30V

vs.S

CE,

pH7

(Ti)

Bri

llian

tcre

sylb

lue

∼−0

.035

Vvs

.A

g/A

gCl,

pH7

Cov

alen

tlybo

und

onto

cyst

amin

ean

dcy

stei

nm

odifi

edgo

ld

238

Poly

(bri

llian

tcre

sylb

lue)

0.04

8V

vs.A

g/A

gCl,

pH7.

4El

ectr

opol

ymer

ized

onG

C0.

2A

M−1

241,

242,

281

3-β

-Nap

htho

yl-b

rilli

antc

resy

lblu

e−0

.180

Vvs

SCE,

pH7

Ads

orbe

don

grap

hite

310

422

4

3-β

-Nap

htho

yl-b

rilli

antc

resy

lblu

e0

Vvs

.SC

EM

ixed

with

carb

onpa

ste

282

Met

hyle

nevi

olet

0.00

5–

0.02

Vvs

.SC

E,pH

7A

dsor

bed

onZ

r-ph

osph

ate

mix

edw

ithca

rbon

past

e23

5


Thio

nine

−0.0

2V

vs.S

CE,

pH7

(Zr)

,A

dsor

bed

onZ

r-,

Ti-p

hosp

hate

mix

edw

ithca

rbon

past

e

235,

249

0.08

Vvs

.SC

E,pH

7(T

i)Th

ioni

ne0.

14V

vs.S

CE,

pH7

Cov

alen

tlybo

und

tose

lf-as

sem

bled

cyst

eam

ine

mon

olay

er

2.0

10−5

–1.

010−

2M

283,

284

Thio

nine

0.24

Vvs

.Ag/

AgC

l,pH

7C

oval

ent

boun

dto

self-

asse

mbl

edm

onol

ayer

of3,

3′-d

ithio

bis(

succ

inim

i-dy

lpro

pio

nate

)on

Au

1.2

105

239

Cro

sslin

ked

thio

nine

0.07

Vvs

.Ag/

AgC

l,pH

7R

eact

ion

ofth

ioni

new

ithto

luen

edi

isoc

yana

te(t

riis

ocya

nate

)on

glas

syca

rbon

/gra

phite

7.0

10−7

–1.

810−

3M

285,

286

Poly

(thi

onin

e)0

Vvs

.Ag/

AgC

l,pH

7El

ectr

opol

ymer

ized

ongr

aphi

te/g

old/

glas

syca

rbon

/In-

Snox

ide

cond

uctin

ggl

ass

2.1

103

5.0

10−6

–1.

010−

3M

287

–29

0

Met

hylv

iole

t0.

02V

vs.S

CE,

pH7

(Zr)

,A

dsor

bed

onZ

r-,

Ti-p

hosp

hate

mix

edw

ithca

rbon

past

e

235,

249

0.03

Vvs

.SC

E,pH

7(T

i)R

esor

ufin

−0.2

6V

vs.S

CE,

pH7

Ads

orbe

don

grap

hite

110

316

0

Gal

locy

anin

e−0

.2V

vs.S

CE,

pH7

Ads

orbe

don

grap

hite

110

216

0

Met

hylc

apri

blue

−0.3

Vvs

.SC

E,pH

7A

dsor

bed

ongr

aphi

te20

160

(con

tinue

dov

erle

af)


b.2

(con

tinue

d)

Med

iati

ngsc

hem

eE

o′ /V

vs.r

ef.

Imm

obili

zati

onR

ate

cons

tant

Sens

itiv

ity

Ref

eren

ces

M−1

s−1

µAm

M−1

cm−2

Ethy

lcap

ribl

ue−0

.31

Vvs

.SC

E,pH

7A

dsor

bed

ongr

aphi

te10

160

o-PH

ENYL

ENED

IIM

INE

N-m

ethy

lphe

nazi

nium

−0.1

6V

vs.S

CE,

pH7

(Ele

ctro

)ads

orbe

don

togr

aphi

te29

1–

295

N-e

thyl

phen

azin

ium

−0.2

1V

vs.S

CE,

pH7

1-M

etho

xy-N

-met

hylp

hena

zini

um−0

.15

VV

vs.S

CE

N-m

ethy

l-phe

nazi

nium

−0.1

10V

vs.

Ag/

AgC

l,pH

7A

dsor

bed

onto

para

ffin

impr

egna

ted

grap

hite

0.42

104

218

N-m

ethy

l-phe

nazi

nium

,1-

met

hoxy

-N-m

ethy

l-ph

enaz

iniu

m

−0.1

2V

vs.

Ag/

AgC

l,pH

7.4

Imm

obili

zed

asdi

amm

inet

e-tr

a(is

othi

ocya

nato

)ch

rom

ates

(Rei

neck

ates

)in

grap

hite

-epo

xyco

mpo

site

elec

trod

es

310

3(N

MP+

)6.

510

3

(M-N

MP+

)0.

5µM

−3

mM

226

N-m

ethy

l-phe

nazi

nium

−0.1

Vvs

.SC

EIm

mob

ilize

din

elec

trop

olym

eriz

ed1,

2-,1

,3-,

1,4-

diam

inob

enze

ne(D

AB

),py

rrol

e-2-

carb

oxyl

icac

id(P

Y-2-

CO

OH

)an

d4,

4′-

dihy

drox

yben

zoph

enon

e(D

HB

)on

Au,

Pt,c

arbo

nel

ectr

odes

10−6

−10

−2M

296


N-m

ethy

lphe

nazi

um+

HR

P+

poly

viny

lpyr

idin

eco

mpl

exed

[Os(

bpy)

2C

l]3+/2+

-pol

ymer

0–

0.1

Vvs

.SC

E,pH

71

Acm

−2M

−1,

110

−7–

210

−4M

297

Poly

(o-p

heny

lene

diam

ine)

(PPD

)0

Vvs

.SC

EEl

ectr

opol

ymer

ized

onca

rbon

past

e26

4,29

8,29

9

p-A

MIN

OPH

ENO

L

2,6-

Dic

hlor

ophe

nol-i

ndop

heno

l0.

055

Vvs

.Ag/

AgC

l,pH

6.5

Ads

orpt

ion

onto

grap

hite

elec

trod

ean

don

togr

aphi

teel

ectr

odes

pret

reat

edw

ithLa

(NO

3) 3

orTh

(NO

3) 4

0.00

8–

0.2

mM

300

9H-b

enzo

phen

oxaz

in-9

-one

−0.2

Vvs

.SC

E,pH

7A

dsor

bed

ongr

aphi

te1.

110

316

0,30

1

Poly

(o-a

min

ophe

nol)

(PA

P)0

Vvs

.SC

E26

4,29

8Po

ly(a

nilin

e)-p

oly(

viny

lsul

fona

te)

0.1

Vvs

.SC

E,pH

7El

ectr

opol

ymer

ized

onG

C30

2Po

ly(i

ndol

e-5-

carb

oxyl

icac

id)

Gla

ssy

carb

on30

3Po

ly(p

heno

safr

anin

e)−0

.4–−0

.2V

vs.

Ag/

AgC

l,pH

7El

ectr

opol

ymer

ized

ongr

aphi

te30

4

FLA

VIN

Rib

oflav

in−0

.430

Vvs

.A

g/A

gCl,

pH7

Ads

orbe

don

topa

raffi

nim

preg

nate

dgr

aphi

te0.

083

104

218

Rib

oflav

in−0

.220

Vvs

.SC

E,pH

7(Z

r),

−0.3

3V

vsSC

E,pH

7(T

i)

Ads

orbe

don

Zr-

,Ti

-pho

spha

tem

ixed

with

carb

onpa

ste

810

223

5,24

9,30

5,30

6

FAD

−0.4

3V

vs.S

CE,

pH7

Cov

alen

tlybo

und

toG

C30

7

FAD

and

FMN

−0.2

50V

vs.S

CE,

pH7

Ads

orbe

don

TiO

2m

odifi

edca

rbon

fiber

s1

–8

mM

308,

309

(con

tinue

dov

erle

af)


Tab.

2(c

ontin

ued)

Med

iati

ngsc

hem

eE

o′ /V

vs.r

ef.

Imm

obili

zati

onR

ate

cons

tant

Sens

itiv

ity

Ref

eren

ces

M−1

s−1

µAm

M−1

cm−2

10-(

3′-M

ethy

lthio

prop

yl)-

isoa

lloxa

ziny

l-7-c

arbo

xylic

acid

−0.3

6V

vs.

Ag/

AgC

l,pH

7A

dsor

bed

onA

uN

oca

taly

sis

310

OTH

ERC

ATA

LYST

S

Nap

htho

lgre

enB

−0.0

5V

vs.S

CE,

pH7

Elec

trop

olym

eriz

edon

glas

syca

rbon

910

231

1

2,7-

Din

itro-

9-flu

oren

one,

2,4,

7-tr

initr

o-9-

fluor

enon

e,2,

4,5,

7-te

tran

itro-

9-flu

oren

one

−0.0

5V

vs.

Ag/

AgC

l,pH

8El

ectr

oche

mic

alre

duct

ion

ofad

sorb

edX

-nitr

o-9-

fluor

enon

eon

GC

yiel

dsth

ehy

drox

ylam

ine

5.2

104

5µM

–2

mM

312

2,4,

7-Tr

initr

o-9-

fluor

enon

e,2,

5,7-

trin

itro-

9-flu

oren

one-

4-ca

rbox

ylic

acid

−0.0

45V

vs.

Ag/

AgC

l,pH

8(−

0.7

Vvs

.A

g/A

gCl,

0.2

MC

a2+)

Elec

troc

hem

ical

redu

ctio

nof

adso

rbed

X-n

itro-

9-flu

oren

one

onG

Cyi

elds

the

hydr

oxyl

amin

e

313

5,5′

-Dith

iobi

s(2-

nitr

oben

zoic

acid

)−0

.04

Vvs

.SSC

E,pH

7El

ectr

oche

mic

alre

duct

ion

ofad

sorb

ed5,

5′-

dith

iobi

s(2-

nitr

oben

zoic

acid

)on

Au

yiel

dsth

ehy

drox

ylam

ine

314

NA

D+

0.13

Vvs

.Ag/

AgC

l,pH

7El

ectr

opol

ymer

izat

ion

ofN

AD

+m

ixed

with

carb

onpa

ste

2.5

105

110

−6–

110

−5M

139


Var

ious

benz

imid

azol

e,2-

benz

oim

idaz

olin

one,

1,2,

3,4-

tetr

ahyd

roqu

inoo

xalin

one-

2,1,

5-be

nzod

iaze

pine

deri

vativ

es

∼0.1

Vvs

.Ag/

AgC

l,pH

7A

dsor

bed

onpr

etre

ated

carb

onor

Pt31

5

Tetr

athi

oful

vale

ne(T

TF)

0.3

Vvs

.Ag/

AgC

l,pH

7.8

Mix

edw

ithca

rbon

past

e31

6–

318

Tetr

acya

noqu

inod

imet

han

(TC

NQ

)0.

38V

vs.S

CE

Ads

orpt

ion

ongr

aphi

te1.

510

631

9,32

0

0.36

Vvs

.SC

EEl

ectr

oads

orpt

ion

ongr

aphi

te4.

810

6

TCN

Q0.

22V

vs.S

CE,

pH7.

5M

ixed

with

carb

onpa

ste

316

–31

8,32

1

CO

ND

UC

TIN

GSA

LTS

NM

A+ T

CN

Q−

+0.3

Vvs

.Ag/

AgC

l,pH

322

NM

P+TC

NQ

−−0

.2–+0

.07

Vvs

.A

g/A

gCl,

pH7

Dep

ositi

onof

cond

uctin

gsa

lton

glas

syca

rbon

orm

ixed

with

carb

onpa

ste

10−5

–10

−3M

322

–32

9

TTF.

TCN

Q0.

1–

0.4

Vvs

.SC

E,pH

7D

epos

ition

ofco

nduc

ting

salt

ongl

assy

carb

onor

mix

edw

ithca

rbon

past

e

10−5

–10

−3M

327,

330

Hex

amet

hyle

nete

trat

ellu

ra-

fulv

alen

ete

trac

ya-

noqu

inod

imet

hane

(HM

TTeF

-TC

NQ

)

330,

331

OTH

ERM

EDIA

TOR

S

MnII

I -mes

o-te

rtap

heny

lpor

phin

e−0

.67

Vvs

.A

dsor

ptio

n33

2Po

ly(m

etal

loph

thal

ocya

nine

)N

i,C

o,Z

n0.

01V

vs.S

CE,

pH7.

4El

ectr

opol

ymer

ized

ongl

assy

carb

on20

µMto

3.0

mM

333

(con

tinue

dov

erle

af)


Tab.

2(c

ontin

ued)

Med

iati

ngsc

hem

eE

o′ /V

vs.r

ef.

Imm

obili

zati

onR

ate

cons

tant

Sens

itiv

ity

Ref

eren

ces

M−1

s−1

µAm

M−1

cm−2

Tetr

arut

hena

ted

coba

lt-po

rphy

rin

com

plex

0.7

Vvs

.Ag/

AgC

lFi

lms

ppb

NA

DH

334

NaN

iFe(

III)

(CN

) 6+0

.2V

vs.A

g/A

gCl

pH7.

5El

ectr

odep

ositi

onon

Ni

335

CoF

e(II

I)(C

N) 6

0.42

Vvs

.El

ectr

odep

ositi

onon

Au

0.5

–6.

0m

M33

6H

ighl

ybo

ron-

dope

dco

nduc

tive

diam

ond

elec

trod

es0.

58V

vs.S

CE,

pH7.

10.

1–

0.5

µM33

7,33

8

p-Fe

rroc

enyl

anili

nefil

ms

0.25

Vvs

.Ag/

AgC

l,pH

7.2

Elec

trod

epos

ited

and

adso

rbed

onca

rbon

0.25

339

Dim

ethy

lferr

ocen

e0.

3V

vs.A

g/A

gCl,

pH7.

8M

ixed

with

carb

onpa

ste

316,

317

Poly

(3-m

ethy

lthio

phen

e)0.

45V

vs.A

g/A

gCl,

pH7

Elec

trod

epos

ition

onPt

10pp

b34

0,34

1

Os(

bpy)

2(P

VI)

10C

l]Clp

olym

er-

0.23

Vvs

.Ag/

AgC

l,pH

7Po

lym

erm

odifi

edca

rbon

fiber

342

Met

alliz

edca

rbon

0–

0.2

Vvs

.SC

E,pH

734

3–

346


P O

O H

HN R

H2NOC

P O

O H

N R

H2NOC

H

∗

Fig. 7 Deactivation mechanism proposed for an adsorbed mediator ofthe o-quinone type and NADH [167].

O

N

(H3C)2N+ +N

N

CH3

NC

NC CN

CN S

S S

S

(a) (b) (c) (d)

Fig. 8 Structural formulas of some commonly used mediators for catalyticNADH oxidation: (a) Meldola blue (p-phenylenediimine) [220, 252];(b) N-methylphenazinium (o-phenylenediimine) [291]; (c) TCNQ [319]; and(d) TTF [316].

on CMEs for electrocatalytic NADH oxi-dation (Table 2). The structural formulasof the most common mediators are pre-sented in Fig. 8. Analyzing these CMEs,a number of principally different evolu-tion lines can be identified: (1) Searchfor a catalytic functionality other than aneutral o-quinone derivative that couldfurther reduce the over voltage and atthe same time, serve as a long-term sta-ble catalyst. (2) Immobilization chemistryother than covalent binding to the elec-trode surface based on functionalizationof solid electrodes. (a) Mediator deriva-tives that form strong interactions withthe electrode surface such as adsorptionof extended aromatic ring systems on car-bon (graphite) and mediator-thiol deriva-tives forming self-assembled monolayerson gold. (b) Mediator derivatives that canbe electropolymerized or trapped withinelectropolymerized layers onto the elec-trode surface. (c) Covalent introduction of

the mediator molecule into a polymericbackbone that can be cast onto solid elec-trodes. Other alternatives have been themixing of monomeric or polymeric me-diators into composite electrodes such ascarbon paste [168] or the use of NAD(P)H-oxidizing enzymes immobilized on theelectrode surface (Table 3).

The demands on the perfect mediatorare very high. A successful transducerhas to meet a number of demands.1) The major object has been (and stillis) to be able to substantially reduce theover voltage but at the same time, re-tain an acceptable reaction rate with atleast a second-order rate constant of 106

to 107 M−1 s−1, preferentially higher, ap-proaching a diffusion-controlled process.For sensor purposes, it would be highly de-sirable to be able to apply a potential of theelectrode between 0 V and ≈ −0.2 V, thatis, within ‘‘the optimal potential range’’,where contributions to the response from


Tab.

3C

hem

ical

lym

odifi

edel

ectr

odes

for

elec

troc

atal

ytic

NA

D(P

)Hox

idat

ion

base

don

enzy

mes

Subs

trat

eEn

zym

ean

dm

edia

tor

App

lied

pote

ntia

land

pHIm

mob

iliza

tion

Line

arre

spon

sera

nge

Ref

eren

ces

NA

DH

Dia

phor

ase

Dir

ecte

lect

ron

tran

sfer

with

elec

trod

e,0

Vvs

.A

g/A

gCl,

pH7.

5–

10.5

Entr

appe

don

the

surf

ace

ofC

past

ean

dgl

assy

Cel

ectr

odes

byco

veri

ngw

ithdi

alys

ism

embr

anes

.

10–

80µM

347

NA

DH

Dia

phor

ase

+fe

rroc

ene

0.25

Vvs

.SC

E,pH

7A

min

ofer

roce

ne,

2-am

inoe

thyl

ferr

ocen

e,or

alky

lfe

rroc

ene-

CH

2N

H(C

H2) n

NH

2,

(n=

4–

12)

was

bond

edw

ithdi

apho

rase

onto

poly

(acr

ylic

acid

)fil

mco

ated

grap

hite

felt

elec

trod

e

0.01

–0.

1m

M34

8–

351

NA

DH

Dia

phor

ase

+fe

rroc

enem

etha

nol

0.25

Vvs

.SC

E,pH

7.5

Imm

obili

zed

atth

em

onol

ayer

leve

lat

aA

uel

ectr

ode

self-

asse

mbl

edw

ith2-

amin

oeth

anet

hiol

352

NA

DH

Enzy

me

imm

obili

zed

onel

ectr

ode

353

Dia

phor

ase

+2-

ferr

ocen

ylet

hano

l(2-

FEA

)or

vita

min

K3

NA

DH

0.3

Vvs

.SC

E,pH

935

4D

iaph

oras

e+

hexa

cyan

ofer

rate

(III

)


NA

DH

Dia

phor

ase

+fe

rroc

ene,

para

-ben

zoqu

inon

e,ca

tech

ol,

para

-am

inop

heno

l,he

xacy

anof

erra

te,

dich

loro

phen

ol-in

doph

enol

,an

thra

quin

one

sulfo

nate

,vi

tam

inK

3,fl

avin

es,

−0.1

5–

0V

vs.A

g/A

gCl,

pH7

–8.

5En

zym

eon

top

ofN

AD

+an

dm

edia

tor

cont

aini

ngca

rbon

past

e

355

–35

9

NA

DH

Dia

phor

ase

+fe

rroc

enyl

met

hano

l

0.4

Vvs

.SC

E,pH

8Im

mob

ilize

don

gold

-pla

ted

poly

este

rcl

oth

360

NA

DH

0.2

Vvs

.SC

E,pH

7Im

mob

ilize

don

elec

trod

e0.

005

–0.

125

mM

361

Dia

phor

ase

+fe

rroc

enyl

met

hano

l

NA

DH

Dia

phor

ase

+el

ectr

opol

ymer

ised

met

hyle

nebl

ue

0V

vs.S

CE,

pH7

Imm

obili

zatio

nof

diap

hora

sew

ithin

ala

poni

tege

lad

sorb

edon

elec

trod

esu

rfac

e

11.2

mA

M−1

cm−2

362

NA

DH

Dia

phor

ase

+fe

rric

yani

de0.

3V

vs.S

CE,

pH7

Imm

obili

zed

usin

gPV

A-S

bQon

cylin

dric

alca

rbon

orPt

mic

roel

ectr

ode

363

NA

DH

Dia

phor

ase

+m

ethy

lene

gree

nor

Mel

dola

blue

0–

0.25

Vvs

,SC

E,pH

7.3

Enzy

me

and

med

iato

rm

ixed

with

carb

onpa

ste

Mel

dola

blue

:0.

20m

Am

M−1

cm−2

Met

hyle

negr

een:

0.24

mA

mM

−1cm

−2

228

(con

tinue

dov

erle

af)


Tab.

3(c

ontin

ued)

Subs

trat

eEn

zym

ean

dm

edia

tor

App

lied

pote

ntia

land

pHIm

mob

iliza

tion

Line

arre

spon

sera

nge

Ref

eren

ces

NA

DH

Dia

phor

ase

+C

o(ph

en) 3

3+Po

lypy

rrol

edi

aphr

agm

elec

trod

eco

ated

with

diap

hora

sean

dC

o(ph

en) 3

3+

364

NA

DH

Dia

phor

ase

+fe

rric

yani

de0.

25V

vs.S

CE,

pH7.

5/0.

3V

vs.S

CE

Enzy

me

held

behi

nda

dial

ysis

mem

bran

e–

entr

appe

din

aph

otoc

ross

linka

ble

PVA

bear

ing

styr

ylpy

ridi

nium

grou

ps/I

mm

obili

zed

onN

AD

+m

odifi

edca

rbon

past

e

365

–36

8/36

9

NA

DH

Dia

phor

ase

+cy

toch

rom

ec

−0.2

Vvs

.SC

ED

iaph

oras

eel

ectr

osta

tical

lybo

und

tocy

toch

rom

ec

imm

obili

zed

In2O

3/S

nO2

(ITO

)el

ectr

ode

coat

edw

ithpo

lygl

utam

icac

idde

riva

tives

370

NA

DH

Lipo

amid

ede

hydr

ogen

ase

(dia

phor

ase)

+fe

rroc

ene

0.4

Vvs

.Ag/

AgC

l,pH

6–

7R

edox

gelf

orm

edfr

omco

poly

mer

izat

ion

ofvi

nylfe

rroc

ene

with

acry

lam

ide

and

N,

N′ -m

ethy

lene

bisa

cryl

amid

ean

den

zym

eon

ferr

ocen

em

ixed

with

carb

onpa

ste

?−3

mM

0.67

µAm

M−1

371

NA

DH

dehy

drog

enas

e+

Ferr

ocen

ylm

etha

nol,

ferr

ocen

ecar

boxy

licac

id,

ferr

ocen

ebut

yric

acid

0.1

Vvs

.SC

E,pH

6.8

Entr

appe

din

apo

lypy

rrol

em

embr

ane

onG

C,d

irec

tET

with

poly

pyrr

ole

372

4.4 Electrocatalytic Oxidation of NAD(P)H at Mediator-modified Electrodes 111N

AD

H−0

.5V

vs.A

g/A

gCl,

pH7.

4En

zym

esm

ixed

into

carb

onpa

ste

1–

50µM

373,

374

Salic

ylat

ehy

drox

ylas

e+

tyro

sina

se

NA

DH

0.15

Vvs

.SC

E,pH

7D

ialy

sis

mem

bran

e37

5

NA

DH

oxid

ase

+1′

,3-d

imet

hylfe

rroc

ene

etha

nola

min

e

NA

DH

oxid

ase,

RuC

l 62−

/RuC

l 63−

0.15

V1

–10

mM

NA

DH

376

NA

DH

NA

DH

oxid

ase

+H

RP

+O

s-PV

P

−0.1

–0

Vvs

.Ag/

AgC

l,pH

7R

edox

gelf

orm

edon

elec

trod

e25

nM–

10µM

377

NA

DPH

Ferr

edox

in-N

AD

Pre

duct

ase

Dir

ecte

lect

ron

tran

sfer

,−0

.5–

0V

vs.A

g/A

gCk,

pH7.

5

Ads

orbe

don

glas

syC

,Au,

orA

gel

ectr

odes

10–

40µM

347

NA

DPH

Glu

tath

ione

redu

ctas

e+

ferr

ocen

e

0.4

Vvs

.Ag/

AgC

l,pH

6–

7R

edox

gelf

orm

edfr

omco

poly

mer

izat

ion

ofvi

nylfe

rroc

ene

with

acry

lam

ide

and

N,

N′ -m

ethy

lene

bisa

cryl

amid

ean

den

zym

eon

ferr

ocen

em

ixed

with

carb

onpa

ste

?−3

mM

1.05

µAm

M−1

371

NA

DH

,NA

DPH

−0.1

Vvs

.SC

E,pH

7El

ectr

opol

ymer

izat

ion

ofen

zym

ew

itham

phip

hilic

pyrr

ole

onel

ectr

ode

29m

AM

−1cm

−2,0

.2µM

(NA

DH

),37

8–

380

Flav

inre

duct

ase

+ri

bofla

vin

15.8

mA

M−1

cm−2

,0.4

µM(N

AD

PH)


easily oxidizable species, for example,ascorbate, urate, and acetaminophen, arenegligible, where molecular oxygen is notelectrochemically reduced, and where thepotential of zero charge is found for mostelectrode materials resulting in low back-ground currents and noise. In biofuelcell applications, the Eo′

of the mediatorshould approach that of NAD+/NADH, soas not to lose any energy. 2) The mediatorelectrode should reveal long-life stability(weeks to months). The immobilization ofthe mediator should be irreversible. Thechemical stability of the mediating func-tionality (hydrolysis, light decomposition,chemical oxidation), the electrochemicalstability, and the stability in the pres-ence of NADH (no radical side reactions)should be very high. 3). All reaction ratesshould be very high, and fast electrontransfer rate between the electrode andthe immobilized mediator, fast charge-transfer rate within the film (the latterapplying to polymeric and multilayer coat-ings), fast reaction rate between NADHand the immobilized mediator shouldprevail. The mediator should be prefer-entially selective for NAD(P)H oxidation,have a well-defined stoichiometry withNAD(P)H, and finally, yield enzymaticallyactive NAD+ as the end product. In linewith the work on homogeneous oxida-tion of NADH with oxidants referred toearlier, most work on CMEs for NADHoxidation has focused on the use of two-electron-proton acceptors. However, somework has been devoted to the use ofone-electron-nonproton acceptor type me-diators (Table 2). As anticipated, CMEsbased on such types of mediators havebeen less successful in decreasing the over-voltage in reaching high reaction rates withNADH, and currently, there are no reportson the products formed as a result of thecatalytic oxidation.

In agreement with the findings by Millerand coworkers [141], mediators incorpo-rating a positively charged phenylenedi-imine functionality have shown somevery promising properties. Monomericphenylenediimines are neither for a longterm chemically nor electrochemically sta-ble, they cannot be easily immobilized ontoan electrode surface, and the Eo′

is toohigh for any real applications. However,when introduced into a larger aromaticnucleus, the chemical and electrochemi-cal stability is much increased and at thesame time, the Eo′

is decreased by sev-eral hundred millivolts. Extended aromaticmolecules are also strongly adsorbed ontocarbon electrodes, especially onto graphite.Representative examples of such medi-ators are N-methylphenazinium (NMP+,‘‘phenazine methosulfate’’) incorporatinga positively charged o-phenylenediiminefunctionality and 7-dimethylamino-1,2-benzophenoxazinium (‘‘Meldola blue’’),a positively charged p-phenylenediiminefunctionality (Fig. 8). These moleculeshave been studied since the early 1980sas electrode modifiers for NADH oxida-tion on the basis of their strong adsorptiononto graphite [220, 252, 291]. In contrast tothe earliest work on CMEs for NADH oxi-dation based on o-quinone modified elec-trodes, these modifiers allowed catalysis ofNADH oxidation much below −100 mVversus SCE at pH 7 and with equal or evenhigher reaction rates with NADH thanthe o-quinones (Table 2). However, three-to four-membered aromatic ring systemssuch as NMP+ and Meldola blue have re-stricted adsorption stability on graphite,and these mediators do suffer from chem-ical instability, NMP+ being light-sensitiveand easily demethylated and Meldola bluedecomposing at pHs above 7.5 as it isunderivatized in position 3. Numerouscommercially available phenoxazine and


phenothiazine derivatives (common dyes,for example, Nile blue, methylene blue,and thionine) are derivatized in both posi-tion 3 and 7 with amine functionalities (atleast one of which is a primary amine)making them alkaline-stable. However,as a result of the two ‘‘competing’’ p-phenylenediimine functionalities withinthese dyes, the positive charge is delo-calized over the entire molecule with theresult that the Eo′

is too low to result inhigh reaction rates with NADH. However,by coupling the primary amine function-ality at position 3 or 7 of the commerciallyavailable phenoxazine and phenothiazinederivatives with, for example, an aro-matic aldehyde or acid chloride, severalbeneficial new properties are donated tothe original dye. The Eo′

of the originalmolecule at pH 7 is increased to a valueclose to that of Meldola blue (−175 mV vs.SCE, pH 7) as a result of the localizationof the p-phenylenediimine functionalitywithin the molecule and the number ofaromatic rings can be increased, therebystabilizing the CME [160, 219, 224, 245,255, 256]. Still, these synthesized deriva-tives suffer from having pKa values around

8 to 9, transferring the positively chargedmediator into a neutral one at pHs higherthan the pKa (see Fig. 9), with the resultthat the reaction rate drastically drops, asalso the long-term stability in the presenceof NADH [162, 224]. The reason for thishas not been elucidated but it could bespeculated to be of the same origin as thatof the o-quinones (Fig. 7).

A further support to the belief that pos-itively charged mediators are superior toneutral ones with respect to reaction ratewas shown in a work by Katz and cowork-ers [200]. PQQ (Fig. 10), was covalently im-mobilized onto a thiol derivative–modifiedgold electrode. The PQQ incorporates an o-quinone functionality and when immobi-lized, has an Eo′

value at pH 7 of −0.125 Vversus SCE. In the presence of NADH,immobilized PQQ shows some moder-ate catalytic activity for NADH oxidation.However, when Ca2+ was added to thecontacting solution, a much higher (≈ 10times) catalytic activity was revealed by CV,although the addition of Ca2+ had virtuallyno effect on the Eo′

of PQQ; thus, the ther-modynamic driving force (Eo′

) remainedconstant. Addition of Ca2+ has also a

Fig. 9 Structural formulas of3-β-naphthoyl-Nile blue (with apositively charged catalytic functionalityhaving high reaction rate with NADH)and its imino form (with a neutralcatalytic functionality having lowreaction rate with NADH) [224].

N

O(C2H5)2N+ N

H O

−H+

N

O(C2H5)2N N

O


N

C HN

O

OC

CHN(CH2)2S

O

O−

Ca2+

N

C HN

OH

OH

C

HN(CH2)2S

O

O−

O

Ca2+

−2e−−2H+

O

O

−OC

O

−O

Fig. 10 Interaction of added Ca2+ ions with (PQQ) mediator covalently attached onto a thiolderivative gold electrode. The catalytic activity for NADH oxidation is increased ≈ 10 times withno effect on the Eo′

of PQQ [200].

positive effect on the electron-transfer ratebetween the modifier and the electrode,increasing the rate constant from 3.3 to18.7 s−1. Other divalent metal ions such asMg2+ and Ba2+ were also shown to havea positive effect on the reaction rate withNADH. Immobilized PQQ in the pres-ence of Ca2+ have been used in a variety ofbiosensor prototypes. Of special interest isthe further covalent binding of the surface-tethered PQQ with an enzymatically ac-tive NAD-derivative to form a coenzyme-mediator arrangement directly at the elec-trode surface [202, 203, 207–209].

Other o-quinone derivatives have alsoshown a drastic increase of the reac-tion with NADH in the presence ofCa2+ or Mg2+. Electrodeposited 3,4-dihydroxybenzaldehyde on GC was shownto have a pKa of around 7 [190]. At pHshigher than the pKa value, the predom-inant form of the reduced form of themediator is then QH−, whereas below thepKa, it is QH2. Additions of Mg2+ or Ca2+were shown to increase the reaction ratewith NADH only at pHs above the pKa,reflecting the binding of the divalent iononly for the QH− but not for the QH2 formof the mediator.

Other larger aromatic ring systemsincorporating o-quinone functionalitieshave also been investigated, especially1,10-phenanthroline-5,6-dione. This com-pound has the ability to form strong and

stable complexes with a wide variety oftransition metal ions such as Fe, Ru, Co,Cr, Ni, and Os, thus making the mediatormolecule positively charged. Immobiliza-tion on carbon electrodes has been onthe basis of adsorption [212], electropoly-merization [210], or mixing with carbonpaste [213]. The resulting CMEs showedgood electrocatalytic properties for NADHoxidation.

One of the main drawbacks with thebest modifiers for NADH oxidation is asmentioned earlier: the variation of theEo′

of the mediator with pH. However,in the last few years, some reportshave been published where a variety ofquinoic type mediators (phenoxazines,phenothiazines, phenazines, flavins) havebeen chemisorbed onto finely dispersedZr- and Ti-phosphates, followed by mixingthe resulting mediator-transition metalion-phosphate complex into carbon pasteelectrodes [233, 235, 248–250, 261, 305,306]. The resulting electrodes revealedhigh electrochemical activity of the boundmediator. The Eo′

of the bound mediatorsdid not or to a very low extent varywith the pH of the contacting solution.Although the Eo′

remained constant witha change in pH, the Eo′

value of theimmobilized mediator could be somewhat(50–100 mV) influenced by the bufferconstituents. When comparing the Eo′

values of all bound mediators with their


Eo′values in solution at pH 7, drastic

shifts in the positive direction variedfrom about 50 to over 400 mV, therebydrastically increasing the reaction ratebetween the mediator and NADH (see alsothe following text).

4.4.2Other Mediating Functionalities andMetal-coated Electrodes

Until a few years ago, most studied andefficient mediators used to incorporate ei-ther a o-quinone or a p-phenylenediiminefunctionality. There are, however, someother types of mediating structures knownto have high reaction rates with NADH,for example, tetrathiafulvalene (TTF) andtetracyanoquinodimethane (TCNQ) (Fig. 8and Table 2). Electrode materials basedon acceptor/donor radical salts suchas N -methylphenazinium tetracyanoquin-odimethane (NMP-TCNQ) attracted muchattention in the early 1980s. The conduc-tivity of these materials is similar to that of

graphite and they have a working potentialwindow of a few hundred millivolt around0 V versus SCE. Oxidation of NADH onNMP-TCNQ radical salt electrodes was ob-served at −0.2 V [322–324], and thus thehalf-wave potential on the organic salt isshifted toward more negative potentials by0.4 to 0.6 V as compared with carbon orplatinum electrodes. The reaction rate forNADH oxidation was estimated to be closeto that of Meldola blue and therefore, itwould fall on the upper linear relationshipin Fig. 11 (see further discussion). Therehas been some controversy regarding thereaction mechanism for NADH oxidationas to whether it occurs directly at the solidNMP-TCNQ (or TTF-TCNQ) electrode sur-face or is caused by the dissolution of someminor NMP+ (or TTF+) known as efficientoxidants for NADH.

In 1998 to 1999, some new efficientmediator functionalities were published.Aromatic molecules derivatized with ni-tro substituents can be electrochemicallyreduced to form hydroxylamine groups,

−0.6−2

−1

0

1

2

3

4

5

6

−0.4 −0.2

y = 3.3428 + 6.9504 × R = 0.95627

a

a a a

a a

ae e

e

c

c

b

f

g

h

i i j

kk kd

y = 5.5273 + 10.317 × R = 0.92782

Eo' vs. SCE

log

k obs

0 0.2

Fig. 11 Dependence of log kobs,[NADH] = 0 on the Eo′at pH 7.0 for

surface-immobilized mediators. Values taken from (a) [160], (b) [305],(c) [200], (d) [312], (e) [210] (pH 7.2), (f) [290], (g) [212], (h) [167],(i) [191], (j) [195] (calculated), (k) [239].


which in turn can be oxidized tonitroso groups by a two-electron-two-proton process [311–314]. In analogy tothe quinone/hydroquinone system, thecatalytic cycle involves the nitroso/hydroxy-lamine couple, as outlined in Fig. 12.Currently reported mediators show highreaction rates with NADH at low poten-tials (Table 2). The positive effect on thereaction rate in the presence of Ca2+ (upto 5 times) was also shown for some ofthese new types of mediators [313].

A drastic oxidation of carbon electrodeswill introduce in a rather unselective way,oxygen-containing functionalities on itssurface [90, 132–137, 172, 381–383]. Base-pretreated GC electrodes have been usedto decrease the overpotential to around350 mV, allowing LC-EC detection at 0.5 Vversus Ag/AgCl with NADH limit of detec-tion in the order of fmol [173]. Carbon-fibermicroelectrodes, after electrochemical pre-treatment [101–105, 175], have been usedfor NADH detection using fast scan(100 V/s) conditions to discriminate be-tween NADH and other compounds thatare also oxidized. The long-term stabilityof the pretreated carbon electrode surfaceis, however, far from great.

Various metal coatings on carbonelectrodes have also been used to re-duce the overvoltage for NADH oxida-tion [343–346]; however, the reduction ofthe overpotential is usually not selectivefor NADH but a general phenomenonfor several other electrochemical reactionssuffering from high overvoltages. Coatingof porous titanium with binary Pt-Pd orternary Pt-Pd-Rh/Ir alloys has also beenreported to decrease the overvoltage forNADH oxidation at pH 9 [384]. Recently, itwas reported that diamond electrodes [337,338, 385] show much higher stability com-pared with other types of carbon electrodesfor NADH oxidation, although the appliedpotential is rather high.

4.4.3Catalytic NADH Oxidation at CMEs Basedon Polymers

Polymer-coated electrodes for NADH ox-idation can be divided into three ma-jor groups: those made from elec-tropolymerization of monomers withno mediating properties, electropolymer-ization of monomers with mediatingproperties, and premade polymers into

O

+N

−O

O O

NH

OH

2 e− + 2 H+

O

NO

4 e− + 4 H+

(a)

(b)

H2O

Fig. 12 Structural formulas of 2,4,7-trinitro-9-fluorenone (a) and electrochemicalformation of its catalytic active form (b) [312].


which mediating functionalities are cova-lently bound.

Electrodes modified by electrodeposi-tion of poly(3-methylthiophene) [340] andpoly(indole-5-carboxylic acid) films [376]show a rather nonselective catalytic effectfor NADH with concomitant oxidation of,for example, dopamine, epinephrine, andacetaminophen. However, the sensitivityfor NADH is increased up to 10 timesand interferences such as ascorbate couldbe minimized by charge-selective mem-branes. Also, poly(aniline)-poly(vinylsulfo-nate) coated GC electrodes were shownto give stable and reproducible electrocat-alytic responses to NAD(P)H in citrate-phosphate buffer at pH 7 [302].

A number of o-quinone [164, 180,188–192, 195, 198, 199], phenoxazine,and phenothiazine derivatives [229, 230,240–242, 244, 257, 267] have been elec-tropolymerized onto electrodes with vary-ing degrees of success for practical applica-tions (Table 2). The electropolymerizationprocess leads in many cases to better stabil-ity of the modified layer and, of course, to ahigh loading of catalytic groups on the elec-trode surface. Electropolymerization has,in general, not largely affected the Eo′

of the o-quinone derivatives as comparedwith their monomer counterparts as theyare based on a bifunctional monomer; onemoiety contains the o-quinone and anotherthe electropolymerizable moiety, whereasfor the phenoxazines and phenothiazines,a drastic shift to the positive direction isusually registered. As mentioned above,many of the phenoxazine and phenthi-azine dyes contain primary or secondaryamine groups both in position 3 and 7,and are thus prone to electrochemical oxi-dation to form radicals. This process leadsto the formation of a polymer and alsocauses localization of the positive charge toone of the remaining amine groups, thus

shifting the Eo′value to a more positive

region, with the result that a much higherreaction rate with NADH is obtained. Thedrawback is, however, that NADH does notreadily penetrate and diffuse into the poly-mer making the response at these CMEsrather low.

The introduction of the mediatingmolecule into a polymeric backbone thatcould be cast onto a solid electrode ormixed into a composite electrode wouldin principle, be very beneficial as the sur-face concentration on the electrode surfacecan be much higher than that based onadsorption or covalent binding [153, 386].The success, however, of this approachvery much depends on several factors suchas the partition coefficient of the analytebetween the aqueous solution and thepolymer phase, the diffusion coefficientof the analyte and the possible necessarycounterions within the polymer film, andthe redox self-exchange between adjacentredox sites within the polymer. In com-parison with the progress that has beenobtained with polymers incorporating one-electron-transfer mediators/donors thathave shown great progress for redox en-zymes with bound cofactors [150], thecorresponding progress for mediator con-taining polymers for catalytic NADH oxi-dation has been less successful. The firstattempts in this direction from Miller andcoworkers based on o-quinones [185–187]showed that the redox propagation withinthe polymer was very sluggish and thatNADH did not really penetrate the polymerbulk. The variation of the response currentto NADH with the thickness of the de-posited film at such electrodes showedthat only for an initial increase in thecoverage, there was an increase in theresponse. A continued increase in poly-mer deposition was followed by a decreasein response reflecting the low partition


coefficient and the low diffusion coeffi-cient of NADH within the polymer film.The choice of o-quinone was also neitherbeneficial for the decrease in overvoltagenor for the stability of the mediator inthe presence of NADH, as commented onearlier. Further attempts in this directionhave been made with improved catalyticfunctionalities, for example, phenoxazineand phenothiazine derivatives [162, 236,237, 269, 270, 272–274], but in principle,facing the same problems with redox self-exchange restrictions within the film, suchas partition coefficient for NADH, anddiffusion coefficient of NADH within thefilm. The polymer best suited for NADHpenetration is based on a hydrophilic, per-meable film of HRP covalently bound toa three-dimensional epoxy network hav-ing polyvinyl pyridine (PVP)-complexed[Os(bpy)2Cl]3+/2+ redox centers. However,this approach suffers from the necessity ofadding a soluble quinoid compound thatin a first reaction is reduced by NADHand then in turn is reoxidized by molecu-lar oxygen to form the hydrogen peroxidethat in turn can start the reaction se-quence with the covalently bound HRP and[Os(bpy)2Cl]3+/2+ redox centers resultingin the response current [292].

4.4.4Mechanistic and Kinetic Aspects

The decrease in overvoltage is closelyrelated to the Eo′

of the mediator on theelectrode surface. In principle, this meansthat the lower the Eo′

of the mediatoris, the more is the overvoltage decreased.This decrease is usually measured either asthe difference between the peak potentialof cyclic voltammograms (Ep) or thedifference between the half wave potential(E1/2) of RDE scans for the catalyzed anduncatalyzed waves. For the simplest case

(monolayer coverage) where only the masstransport of NADH and/or the second-order rate constant (kobs) between NADHand the mediator can limit the reaction, Ep

and E1/2 for the catalyzed waves are givenby Eqns. (35) and (36).

Ep = Eo′ + RT

nF

[0.78 + ln

(D

1/2NADH

kobs

)

+ ln(

nFυ

RT

)1/2]

(35)

E1/2 = Eo′ + RT

nFln

(1 + ik

id(L)

)(36)

where

ik = nFAkobsCNADH (37)

and

id(L) = 0.620nFAD2/3NADHν−1/6CNADHω1/2

(38)

and where DNADH is the diffusion co-efficient, the surface coverage of themediator, A the electrode surface area, υ

the sweep rate used for CV, CNADH thebulk concentration of NADH, ν the kine-matic viscosity, and ω the angular velocityof the RDE.

As can be seen, factors other than Eo′

affect the effective decrease in overvoltage,for example, sweep rate (υ), coverage (),rotational speed (ω), and so forth. Thepicture becomes even more complicatedwhen all other steps that may contributeto the kinetics of the charge transportbetween the NADH in solution and theelectrode are considered.

It is of course of importance that theelectron transfer rate between mediatorand the electrode (ks) does not set a limitto the rate of the overall reaction. Theapparent electron transfer rate (ks) of asurface-immobilized redox compound can


be evaluated in CV from the variation of thepeak potentials of the oxidation and reduc-tion waves with the sweep rate [254, 387,388]. However, variations of the peak po-tential with sweep rate can also be causedby a number of other factors such as un-compensated resistance, local pH changes,and movements of counter ions [388]. Thesweep method will therefore only give thelowest value of the rate constant. Typical re-ported values are 1 s−1 for oxidized carbonblack [136], 4 to 17 s−1 for adsorbed phe-noxazines and phenothiazines, dependingon pH [220, 243, 254, 301], 3 to 6 s−1 foradsorbed o-quinones [167, 196], 3.3 s−1 forcovalently bound PQQ on gold [200], 22 to65 s−1 for electrodeposited dihydroxyben-zaldehydes on GC [190, 191], 70 s−1 for2-nitro-9-fluorenone on GC [312], 450 and600 s−1 for 5,5′-dithiobis(2-nitrobenzoicacid) and 5,5′-dithiobis(2-nitropyridine),respectively, on gold [314].

When the mediator is adsorbed or co-valently tethered in multilayers, or whenincorporated into a polymeric backbone orfilm, the charge propagation in the filmmay contribute to the rate-limiting kinet-ics in the catalytic cycle of a CME [153,386, 389–391]. The combined effects ofelectron and counterion transport as wellas conformational changes and solventmolecule movements in the film duringan electrochemical cycle are collectivelydenoted charge transport. For instance,in redox polymer-coated electrodes, thecharge transport usually obeys Fick’s lawof diffusion and hence can be defined bya diffusion coefficient. The partition coef-ficient of NADH between the contactingsolution and the polymer film, the diffu-sion coefficient of NADH in the polymer,and the reaction rate between NADH andthe mediator in the film may also con-tribute to the rate-limiting kinetics.

The rate constant of the chemicalreaction between NADH and the surface-attached mediator (kobs) can be evaluatedby CV [392] but is more appropriately donewhen a rotating electrode is used [153].The situation is straightforward at CMEsbearing monolayers or less of immobilizedmediator. The general equation for thecatalytic current (icat) is, if is less thanequal to the monolayer and the appliedstationary potential is greater than equalto 0.12 V more positive than the Eo′

of themediator,

icat = nFAkobsCNADHD2/3NADHω1/2

D2/3NADHω1/2 + 1.61v1/6kobs

(39)

Equation (39) reverts to the Levich equa-tion for an irreversible reaction when theheterogeneous rate constant kf is replacedby kobs. Inversion of Eq. (39) thus givesthe corresponding Koutecky-Levich equa-tion,

1

icat= 1

nFAkobsCNADH

+ 1

0.620nFAD2/3NADHv1/6CNADH

1

ω1/2

(40)

Equation (40) shows that a plot of 1/icat

versus 1/ω1/2, extrapolating to 1/ω1/2 →0, gives the electrocatalytic rate fromthe intercept expressed as the productkobs. The coverage is evaluated from theintegration of the area of the anodic orcathodic wave of a cyclic voltammogram.Care must be taken when evaluating thesurface coverage as the effective area underthe voltammetric wave can be largelyinfluenced by the sweep rate if the electrontransfer rate, ks, has a limited value [254].

For many CMEs used for the electrocat-alytic oxidation of NAD(P)H, it has beenpostulated that a charge transfer complex(CT) is formed in the reaction sequence


between NAD(P)H and the mediator be-cause kobs was found to decrease in adistinct pattern with an increase in the bulkconcentration of NAD(P)H [160, 200, 212,220, 224, 235, 245, 249, 301, 305]. This isalso in agreement with the findings of thehomogeneous reaction between NAD(P)Hand its analogues with various electronacceptors [145, 146] (see preceding text).

NADH + Medoxk+1⇐⇒k−1

[CT]

k+2−−−→ NAD+ + Medred (41)

Combination of the rate constants k+1,k−1, and k+2 yields (c.f. Michaelis-Mentenkinetics)

KM = k−1 + k+2

k+1(42)

Now, the overall second-order rate con-stant, kobs, can be expressed as

kobs = k+2

KM + CNADH(43)

Inversion of Eq. (43) gives

1

kobs= KM

k+2+ CNADH

k+2(44)

and substitution of kobs, Eq. (44), intoEq. (40) gives

1

icat= 1

nFA(

k+2

KM + CNADH

)CNADH

+ 1

0.620nFAD2/3NADHv−1/6CNADH

1

ω1/2

(45)

or

1

icat= 1

nFAk+2+

(KM

nFAk+2

+ 1.61v1/6

nFAD2/3NADHω1/2

)1

CNADH(46)

If this assumption of a charge-transfercomplex holds true, plots of 1/kobs ver-sus CNADH and 1/icat versus 1/CNADHwould result in straight lines. Moreover,the intercept of Eq. (46) should be indepen-dent of CNADH and ω. Experimental resultsfrom many different laboratories confirmEqs. (44) to (46) [160, 200, 210, 212, 218,220, 235, 239, 249, 301, 305] and it can belooked upon as a general observation withconcluding evidence that a charge-transfercomplex is formed between NAD(P)H andthe mediator.

The rate constant, kobs, can thus beobtained either through the use of CVor the use of a RDE. The close agreementbetween the results obtained with bothCV or RDE [210, 312] further support themechanism given in reaction (41).

If kobs in Eq. (40) is replaced byk+2/(KM + CNADH) from Eq. (43), then

icat =nFA[(k+2/(KM + CNADH)]

×CNADHD2/3NADHω1/2

D2/3ω1/2 + 1.61ν1/6

×[k+2/(KM + CNADH)]

(47)

This equation tells us that for a constantCNADH (up to about KM), icat will varylinearly with for small coverages andwill be independent of for high cov-erages. Experimental evidence has beenreported [220]. This means that if the initialsurface coverage of the mediator is suffi-ciently high, some inactivation or decom-position of the mediator will pass unno-ticed with regard to the response current.

By 1986 [160], it was noticed thatfor a number of adsorbed phenoxazinederivatives, there exists a linear correlationbetween the experimentally evaluated logkobs and the Eo′

of the immobilized medi-ator. Actually, it was found that obviouslythere existed two different linear corre-lations, one with mediators with higher


reaction rates and another with lower re-action rates. In Fig. 11 is shown log kobsversus Eo′

plots of reliable values foundso far in the literature. It is obvious thatwhat was indicated in Ref. [160] still holdstrue, that is, there exists two different lin-ear correlations. It is interesting to notethat for immobilized PQQ [200], the eval-uated kobs falls on the lower line but inthe presence of Ca2+ ions increasing thereaction rate of PQQ with NADH about 10times, this kobs falls on the upper line. Katzand coworkers [200] analyzed in detail theeffect of Ca2+ on k+2 and KM and surpris-ingly, equal values of k+2 were obtained inthe absence and presence of Ca2+, whereasthe KM value was very much affected byCa2+. In the absence of Ca2+, the KMvalue was 109 mM, whereas in the pres-ence of 20 mM Ca2+, the KM value waslowered to 0.73 mM. This means that withCa2+ addition, the equilibrium betweenNADH and PQQ is shifted dramatically tothe intermediate complex formation butthe decay rate of the CT-complex leadingto the formation of NAD+ is not changedat all. Therefore, the [Ca2+··· PQQ] com-plex initially formed (Fig. 10) has a muchhigher affinity for NADH than PQQ it-self. Probably, the positive charge causedby the complex formation of Ca2+ withPQQ functioning as promoters for theNADH oxidation provide a favorable ori-entation of the NADH molecules for thisredox process. It is also interesting tonote that the highest kobs value evalu-ated for the upper line, 5.2 104 M−1 s−1,was obtained with a new type of mediatorhaving a nitroso/hydroxylamine catalyticfunctionality (Fig. 12) with an Eo′

value atpH 7 of −0.05 V versus Ag/AgCl. Similaror even higher values of kobs were found forsome phenothiazines (thionine, Azure C,and Azure A) covalently bound to thiolderivative–modified gold electrodes [239],

however, at much higher applied po-tentials (≈ +210–230 mV vs. Ag/AgCl).Thus, these mediators fall on the lowerlinear line. The perfect mediator for ana-lytical purposes has not yet been identified,but some guidelines for future work can,however, be seen. One should focus ontrying to make new mediator derivativesfalling on the upper curve but having ahigher Eo value. To be able to reach akobs value of 107 M−1 s−1, the Eo′

at pH 7should be around +100 mV versus SCE.

It has been found that kobs is very much afunction of pH. It is still not clear whetherthis is a result of the direct influence ofpH on reaction kinetics or whether the Eo′

values of both the NAD+/NADH and themediator vary with pH. As an example,Meldola blue, kobs increased from 3 104 atpH 7 to 8 104 M−1 s−1 at pH 6, indicatingthat pH must have an effect on the reac-tion rate as the Eo′

of Meldola blue moves30 mV per pH unit and thus Eo’ shouldbe constant as the Eo′

of NAD+/NADHalso shifts 30 mV per pH unit. A similarincrease in reaction rate with a decrease inpH was found for another benzophenox-azine derivative (9H-benzophenoxazin-9-one) adsorbed on graphite [301] andfor 3,4-dihydroxybenzaldehyde electrode-posited on GC [191]. However, for boththese mediators, the Eo′

will move be-low pH 7 at 60 mV per pH unit resultingin an increased Eo′

with a changein pH to more acidic values. For 9H-benzophenoxazin-9-one, both k+2 and KM

increased for lower pHs, although witha net increase in kobs, whereas for 3,4-dihydroxybenzaldehyde, a decrease in pHled to a slight increase in k+2 but a moredrastic decrease in KM, thus causing anet increase in kobs. Further investigationsare indeed needed to be able to come upwith a more detailed explanation for suchfindings.


The lower reaction rate at higher pHin conjunction with less stable mediatorproperties have been discussed [160, 162,301] in relation to the mechanism occur-ring in the solution proposed by Fukuzumiand coworkers [44, 145, 393]. In a first re-action step, NADH forms a CT-complexwith the mediator. Next, the first electronis donated to form [NADH•+ Med•−].

NADH + MedoxkCT⇐⇒

[NADH-Med]k+1⇐⇒k−1

[NADH•+Med

•−]

(48)

The major reaction route would then bethat the acid proton of NADH•+ withinthe complex should be transferred to themediator part within the complex (Med•−),followed by a rapid transfer of the secondelectron, and finally, a decay of the complexand the release of enzymatically activeNAD+ and reduced mediator.

[NADH•+Med

•−]kH⇐⇒

[NAD MedH•]

fast⇐⇒ NAD+ + MedH−

(49)

Depending on whether the mediator isa one- or two-proton acceptor, a possiblefollow-up process can then occur

MedH− + H+ −−−→ MedH2 (50)

which in a protic medium should be veryfast.

At high pH, hydroxyl ions could possiblyinfluence the reaction in a similar mode aswas proposed in aprotic media when a basewas added. Then, a reaction competingwith reaction (49) could take place

[NADH•+Med

•−] + OH−

kB⇐⇒ NAD• + Med

•− + H2O (51)

leading to free radicals, which in turn maycause abortive side reactions, with the re-sult that the mediator slowly decomposes,possibly followed by electrode fouling. Noinvestigations so far have been made withrotating ring-disk electrodes trying withelectrochemical means to identify the re-action products formed at CMEs. Suchinvestigations should be done in the fu-ture to possibly shed further light into thereaction mechanism.

From a fundamental point of view, thereaction between NADH and flavin is themost interesting. As free flavins have atpH 7 Eo′

values around −0.43 V versusSCE, the resulting Eo′

between NADHand the flavin will not allow the reactionto be studied with electrochemical meansas the resulting currents are too small.When adsorbed or covalently immobi-lized onto ordinary solid electrodes, flavinderivatives more or less retain their solu-tion values [307, 394]. Recently, however,it was reported that when immobilizingriboflavin onto Zr- or Ti-phosphate, fol-lowed by mixing with carbon paste [235,249, 305, 306], the Eo′

for riboflavin wasfound to be around −0.2 V versus SCEand virtually independent of pH. Thus,the reaction rate with NADH was sub-stantially increased. It could be confirmedthat for riboflavin too the reaction occursin a CT-complex [305]. In the presence ofCa2+ or Mg2+, a further shift in the Eo′

value of riboflavin was noticed, as also adrastic increase in the reaction rate withNADH. The possibility of changing theEo′

of adsorbed riboflavin with the com-position of the contacting solution [235,305] for a constant pH and to changethe pH of the solution while keeping theEo′

of adsorbed riboflavin constant givespossibly a unique possibility to separatelyinvestigate the influence of pH and theinfluence of the difference in Eo′

values

4.5 CMEs Based on NADH-oxidizing Enzymes 123

between NADH and the flavin on the re-action rate between these two compounds.Moreover, a basic understanding of theinteraction between quinoid-type media-tors and Zr- and Ti-phosphate may lead tothe fact that tailored modified electrodescan be designed for biosensor purposes,where the previous drawback with usingelectron-proton acceptor/donor type me-diators having pH-dependent Eo′

valueswill be circumvented. A similar effect onthe Eo′

value of immobilized flavins wasfound for TiCl4-treated carbon fibers [309].The treatment causes the formation of aTiO2 layer on the carbon fiber surface,which has a strong affinity for phosphategroups. Both FAD and FMN strongly bindto these electrodes, revealing Eo′

valuesclose to −0.2 V versus SCE at pH 7 withonly a slight variation with pH and clearcatalytic oxidation currents in the presenceof NADH.

4.5CMEs Based on NADH-oxidizing Enzymes

A number of CMEs for biocatalytic NADHand NADPH oxidation have also been de-scribed in the literature (Table 3). Mostcommonly, various preparations of DIhave been used for NADH. Direct electrontransfer between DI and electrodes seemspossible [347], although the efficiency ofdirect electron transfer between the re-duced enzyme and the electrode (eithercarbon paste or GC) is not very efficient.However, the electron transfer from re-duced DI to the electrode can be mediatedusing both one-electron-nonproton andtwo-electron-proton acceptors (Table 3).The potential at which NADH is oxi-dized is therefore very much dependenton the mediator employed. A thermostableDI has been intensively studied by Ikeda

and coworkers [358, 359] and was used tooxidize NADH electrocatalytically, duringwhich a variety of quinone compoundsand several kinds of flavins were studiedas mediators. The DI-catalyzed electrolyticoxidation, that is, bioelectrocatalytic oxida-tion of NADH, proceeds very rapidly withquinones and flavins as mediators. Anal-ysis of the bioelectrocatalytic current bythe theory of steady-state catalytic currentsreveals that the bimolecular reaction ratesbetween the enzyme and the quinoneswhose redox potentials are more positivethan −0.28 V versus Ag/AgCl at pH 8.5are as high as 108 M−1 s−1, suggesting thereactions to be diffusion-controlled. Theredox potential of DI was determined tobe more positive than NAD+/NADH by42 mV at pH 8.5 using a spectroscopicmethod. By changing the pH of the so-lution to more alkaline values, the Eo′

of DI will be more negative than that ofNAD+/NADH, making it possible to runthe reaction backward, that is, the bioelec-trocatalytic reduction of NAD+ is predictedto be favorable under alkaline conditions(see in the following text).

The electrochemical oxidation of NADHcatalyzed by NADH dehydrogenase en-trapped in a polypyrrole (PPy) membraneon an electrode surface has also been in-vestigated [372]. The results suggest thatdirect electron transfer occurs between theenzyme and PPy without the aid of smallmediators.

The NADH oxidase has also beenused to oxidize NADH. The reaction canbe based on mediated electron transferfrom reduced NADH oxidase to theelectrode [375, 376]. The reaction can alsobe based on the formation of hydrogenperoxide, which in turn, reacts with HRPcovalently bound within an Os-containinghydrogel that efficiently mediates thereduction of reduced peroxidase [377].


For NADPH, efficient catalytic oxidationhas been reported using three differ-ent redox enzymes, namely, ferredoxin-NADP-reductase [347], glutathione reduc-tase [371], and flavin reductase [378, 379],using different types of mediators.

4.6Amperometric Biosensors Based onNAD(P)-dependent DehydrogenaseEnzymes

Given the difficulties associated with thecatalytic reduction of NAD+, all amper-ometric biosensors studied on the basisof NAD-dependent dehydrogenases havebeen focused on measuring NADH, usu-ally its production, but in some instances,also its consumption, in conjunction witha dehydrogenase [76, 77]. However, fora successful approach, one must keepin mind that most of the reactions cat-alyzed by these dehydrogenases have Eo′

values more positive than that of theNAD+/NADH redox couple. Only for afew reactions, for example, the oxidationof monosaccharides (glucose, galactose) oraldehydes by their corresponding dehy-drogenases, does the equilibrium of thereaction (reaction (2), see preceding text)actually favor the production of NADH be-cause the Eo′

s of these reactions have lowervalues than that of the NAD+/NADH re-dox couple. To be able to construct abiosensor on the basis of a dehydrogenaseby measuring the production of NADH,it is very essential that the initial NADHformed be instantaneously consumed bythe mediator (or possibly directly at theelectrode surface); otherwise, equilibriumin reaction (2) will be reached and fur-ther production of NADH will cease. Thereduced mediator in turn must also bereoxidized rapidly to regenerate its active

oxidized form. In essence, this means thatall three reaction steps (the enzymatic, themediated, and the electrochemical) needto occur very close in space for a success-ful approach. This is outlined in Fig. 2where it is shown that in the initial step,the electrons have to move against thethermodynamic driving force through anintimate coupling with the mediator andthe electrode in order to realize a netthermodynamic driving force for NADHproduction. It is therefore very importantthat the mediated reaction is as rapid aspossible motivating further search for find-ing the optimal mediator. Any biosensorbased on an NAD-dependent dehydroge-nase with the ambition of reaching themarket must resolve the problem of immo-bilization of sufficient amounts of NAD+within the biosensor format. A very el-egant approach has been suggested byWillner and coworkers through their cova-lent coupling of an enzyme-active NAD+analogue with a mediator (PQQ), which inturn is tethered to a gold electrode [202,203, 208, 209]. However, in light of therestricted lifetime of both NAD+ andNADH because of their susceptibility to de-composition in aqueous environment [24],a surplus of cofactor seems necessaryand suggested coimmobilization of NAD+into composite electrodes [168] or in elec-tropolymerized layers [229] together withmediator and enzyme may be a promisingapproach to solve this problem.

4.7Direct Electrochemical Reduction ofNAD(P)+

4.7.1General Observations

Studies on the electrochemical reductionof NAD+ have been made using normal

4.7 Direct Electrochemical Reduction of NAD(P)+ 125

DC polarography and phase-selective ACpolarography at dropping-mercury elec-trodes, CV at the hanging-mercury dropelectrode, potential-step chronoamperom-etry, constant electrode potential coulome-try, and RDE methodology. Platinum, gold,PG, and GC electrodes have also beenused.

The electrochemical reduction of NAD+in aqueous solutions is well documentedincluding studies on the adsorption ofproduct, intermediates, and NAD+ itselfon electrode surfaces [75, 97, 118, 124,125, 395, 396]. In an initial polarographicstudy by Kaye and Stonehill [397], thereduction of NAD+ at the mercury elec-trode was observed as a single wave withan E1/2 of ≈ −0.9 V in neutral aqueoussolution. Burnett and Underwood [118] re-solved this single wave into two cathodicwaves, a first one at E1/2 ≈ −1.1 V anda second at E1/2 ≈ −1.7 V, by recordingthe polarogram in an alkaline buffer pH 9to 10 to minimize the hydrogen evolu-tion and chemical decomposition of thenucleotide and in the presence of tetraalky-lammonium salt to depress adsorption ofNAD+ and related species. Appropriateconditions, including ionic strength andsurface activity of the background elec-trolyte and initial potential and potentialscan rate, have to be chosen to diminishthe adsorption effects [118], ascribed to bemainly caused by the adenine moiety [75,124, 396, 398, 399] of the species partici-pating in the overall reduction of NAD+(see the following text).

The polarographic patterns of bothwaves have been extensively discussedby Elving and coworkers [75] includingNAD+ and analogues (Fig. 3). A sim-ilar pattern is observed for all theone-substituted nicotinamides, includingNAD+, showing the two well-separated1e− waves at alkaline pHs, the intensity

of the second wave being less than 10% ofthe first. Contrary to this pattern, nicoti-namide shows, between pH 3 and 7, twoadjacent cathodic waves of equal height,closely followed by a hydrogen dischargewave. Regardless of this different pattern,the first wave corresponds to a reversible1e− charge transfer (see following text),followed by an irreversible chemical dimer-ization of the radical formed, which isuncharged in the case of NAD

•+ (NAD,reaction 12) but negatively charged for thenicotinamide [400]. Exhaustive controlledelectrolysis of the one-substituted nicoti-namides has provided an insight of thespecies involved [75]. In brief, after elec-trolysis at a potential corresponding to thefirst wave plateau (−1.25 V), both cathodicwaves disappear and an anodic wave corre-sponding to the neutral dimer of the firstwave product appears at ≈ −0.4 V (reac-tion 13). By comparison of the faradaiccurrents and the spectral characteristicsof the final products of this electrolysiswith the obtained under conditions of fa-vored adsorption, Elving and coworkersdemonstrated that the number of elec-trons in both cases is very close to 1(0.96 ± 0.01) and that adsorption has verylittle effect on the nature of the electrol-ysis products. If potential is then shiftedto −1.8 V, no further current flow is ob-served, indicating that the final productof the first wave is neither an intermedi-ate in the formation of the second wave,nor is it further reduced to a differentproduct. However, a small cathodic wavehas also been reported to appear [401],which has been attributed to the freenicotinamide present as an impurity andslight NAD+ decomposition during ex-tensive electrolysis at alkaline conditions.After exhaustive electrolysis at −1.80 V,the average value for n results to be veryclose to 2 (1.98 ± 0.11) [401] and no waves


are seen at mercury electrodes because,as already described, oxidation of thedihydropyridine product occurs at morepositive potentials than that for the oxida-tion of mercury [75]. On PG electrodes,after electrolysis of MCP+ (1-methyl-3-carbamoylpyridinium ion) at −1.80 V, alarge anodic peak is observed at −0.04 Vas a result of the oxidation of the 2e−product; subsequent potential sweep to-ward more negative regions produces acathodic peak at −1.18 V corresponding tothe reduction of the MCP+ formed duringthe anodic scan. Finally, if the potential isthen scanned in the positive region, the ex-pected dimer oxidation is seen at −0.34 V,followed by the peak at −0.04 V [75].

Taking into account these considera-tions and the appearance of the secondwave at a potential where the dimer isnot reduced (see following text), in addi-tion to the high dimerization rate (reaction12 and Sch. 1), and cyclic voltammetricpatterns, Elving and coworkers concludedthat the second electron transfer has tobe faster than the dimerization and thus,NAD+ is directly reduced to the dihydropy-ridine at the second wave in an overall 2e−process. The rationale for involvement ofa proton in the overall NAD+-reductionprocess derives from the formation of en-zymatically active NADH. Two differentsequences were thus postulated by theseauthors: (1) e−, e−, H+ or (2) e−, H+, e−.If protonation of the free radical formedon addition of the first electron is veryrapid, as compared with its dimerizationrate constant, either sequence could becompetitive with the dimerization reactionin producing the dihydropyridine deriva-tive; otherwise, the first sequence wouldbe more likely [75]. This uncertainty wasclearly solved by the fact that the neu-tral radical NAD• is not further reducedin aprotic media and thus, protonation

occurs prior to or concurrent with chargetransfer [400].

Overall, the cathodic reduction of NAD+to NADH at mercury electrodes proceedsthrough two discrete 1e− steps, wellseparated in potential, and through theprotonation of the free radical producedon the first step prior to or concurrent withthe second electron transfer (Sch. 1).

NAD+ + H+ + 2e− −−−→ NADH (52)

4.7.2Adsorption Phenomena

Special attention has been paid to the im-portance of adsorption and conformationfactors in the one-electron reduction ofNAD+ on different electrode surfaces [97,124]. On mercury electrodes, the area oc-cupied per NAD+ molecule reaches anexperimental limiting value of 130 A2,which calls for a folded configuration of thecoenzyme with the nicotinamide and ade-nine rings parallel and close to each otherand with the adenine ring flat on the elec-trode surface. This assumption is basedon the theoretical areas for an extended orfolded NAD+ configuration as the mini-mum calculated areas result in 190 A2 forthe extended configuration and 125 A2 and85 A2 for the folded configurations (par-allel and perpendicular to the electrodesurface, respectively) [97]. The theoreticalareas have been deduced from projec-tions of CPK space-filling models of NAD+for monolayer coverage. The folded paral-lel configuration at the electrode/solutioninterface calls then for an electron-transferpathway from the electrode through theadenine ring to the nicotinamide moi-ety [97]. The possibility of adenine servingas a mediator for nicotinamide reduc-tion has been demonstrated by pulse


radiolysis, fluorescence, and electrochem-ical experiments [399, 402]. Similarly, aspreviously described, the mediation ofadenine for NADH oxidation has beensuggested [134]. On GC electrodes, NAD+covers an area of 90 A2 per molecule andmost probably, a perpendicular configura-tion can be assumed, which agrees wellwith the proposed adsorption sequencefor NAD+ produced by anodic NADH ox-idation at clean GC electrodes [95]. Undersuch conditions, NAD+ is rapidly adsorbedin a parallel orientation, followed by arelatively slow reorientation to a perpen-dicular orientation where the adsorbateis more tightly bound than the planar-oriented adsorbate [95]. On PG electrodes,a parallel orientation of folded NAD+may be favored because of the observed110 A2 occupied per molecule and theplanar-oriented hexagonal rings of thesesurfaces [97]. Negligible NAD+ adsorptionis observed at platinum electrodes in thefaradaic regions. Similar studies to the pre-viously described using FT-SERS in goldelectrodes are not possible in the faradaicregion because all changes on the SERSbands occur or begin in the non-faradaicregions [113–115].

As previously discussed, NAD+ in so-lution is partially in a folded configura-tion [14–16]. It then seems clear that thefolded conformation is preferentially ad-sorbed within the faradaic region, anddifferences may remain in the specific ori-entation, either parallel or perpendicular atthe electrode/solution interface, depend-ing on the electrode material.

4.7.3Mechanism and Kinetics

The first electron transfer to NAD+ isclaimed to be reversible based on alog i-E plot [118] and cyclic voltammetric

curves [126]. However, unequivocal re-versibility was further demonstrated bySchmakel and coworkers [401]. In their in-vestigations, the slope of the first NAD+wave was greater than that for a reversible1e− transfer and E1/2 became more pos-itive with the increase in concentrationor drop time. These observations and thecase of a reversible electrode process inthe absence of adsorption are indicative ofan irreversible dimerization subsequent tocharge transfer, which was further sup-ported by the difference in the cyclicvoltammetric patterns at slow and rapidscan rates and the increase in the ipa/ipc ra-tio with increasing scan rate. Additionally,the 60-mV separation of peak potentialsfurther supports the reversible 1e− na-ture of the redox couple [401, 75]. Theheterogeneous rate constant, ks, of thisreaction:

NAD+ + e− ks⇐⇒ NAD•

(53)

has been estimated to exceed 1 cm s−1 andan Eo′

of −1.155 V at pH 9.1 (25 C) wasfound for the NAD+/NAD• couple [125].The reversibility of reaction (53) is thusvirtual because it is coupled to a fastdimerization reaction already describedunder the section of direct electrochemicaloxidation of NADH.

2 NAD• kd−−−→ (NAD)2 (12)

Different values of kd have been derivedfrom electrochemical measurements,for example, 2 106 [401], 3 106 [124],8 106 [126], and 8 107 M−1 s−1 [125]. Be-cause of the rather complex mechanismof NAD+ reduction including a rapidfollow-up chemical reaction and adsorp-tion processes depending on experimentalconditions (see following text), this con-stant should only be calculated underconditions in which the electron transfer


(NAD)2

III II

IIV

Et4NCl Et4NCl

Et4NCl

KCl

Slo

w s

can

rate

and

any

Ei

orFa

st s

can

rate

and

Ei =

−0.

9 V

400

mM

Et 4

NC

l ≅100 mM

Et4 N

Cl

400 mM Et4NClFast scan rate and Ei = −0.1 V

63–400 mM KCl

NAD•

NAD+

(NAD)2NAD•NAD+

(NAD)2

(NAD)2

(NAD)2

NAD−

NAD• NAD•

NAD+

(NAD)2NAD•NAD+

NAD+

e−

e−

e−

e−

(NAD)2NAD•NAD+e−

e− e−

Hg

Fig. 13 Mechanistic pathways for the first charge transfer inNAD+ reduction under different experimental conditions. Et4NC1,tetraethylammonium chloride; Ei, initial scan potential. Shadedarea around Hg corresponds to the adsorbed layer. (Redrawn fromW. T. Bresnahan, P. J. Elving, J. Am. Chem. Soc. 1981, 103,2379–2386.)

is strictly diffusion-controlled. A rangeof kd = 6 − 8 107 M−1 s−1 [402, 403] wasobtained from pulse radiolysis and spec-trophotometric measurements. The dimeris not further reduced in the potentialrange down to −1.8 V (foot of the dis-charge current for hydrogen on mercury).However, when the potential is shifted intothe background discharge region (−1.85to −1.90 V), slow increase of absorptionat 340 nm is observed, suggesting indirectreduction of the dimer to the dihydropyri-dine by interaction of the dimer with re-duction products of background dischargesuch as hydrogen radicals. Alternatively,NAD dimers decompose at slightly alka-line conditions in the absence of oxygento NAD+, which can be directly reducedin a 2e− process at −1.8 V to dihydropy-ridines [75]. On the contrary, the oxidationof the dimer to NAD+ is clearer starting ata potential of ≈ −0.4 V [75].

(NAD)2 −−−→ 2 NAD+ + 2e− (13)

A disproportionation reaction of the dimerresulting in NAD+ and NADH has beenalso reported [404], which is at least for-mally analogous to the disproportionationof a semiquinone [97]:

(NAD)2 + H+ −−−→ NAD+ + NADH(54)

From a mechanistic point of view,the reduction of NAD+ to NAD• in-cludes: (1) competing adsorption, (2) thetimescale of the experiment, (3) the ini-tial state of the electrode surface, and(4) the background solution ionic strengthand surface activity [124]. The nature ofthe buffer is not expected to affect thisfirst wave because its pH independenceis above pH 5 [401]. A general overviewof the mechanistic first-reduction pathwaydepending on experimental conditions isdepicted in Fig. 13. In the absence of


surfactant, NAD+ undergoes both adsorp-tion and diffusion-controlled reduction inKCl solutions. The former predominateson the timescale of fast-scan CV experi-ments, whereas diffusion-controlled cur-rents are observed on slow timescale suchas DC polarography and slow scan rateCV. The absence of pre- or postprocessesis ascribed to nearly equal adsorption ofNAD+ and NAD•. The dimer is also ad-sorbed and it only desorbs at potentialsmore negative than −1.20 V and −1.32 Vin 63 mM and 400 mM KCl solutions, re-spectively. Oxidation of adsorbed NAD

•is

not expected because the shape of the an-odic peak on fast-scan CV indicates that itis diffusion-controlled. In the presence oflow surfactant concentration (≈100 mM),two processes are seen on both slow andfast scan rates; one process involves re-duction to soluble products of NAD+ asit diffuses to the electrode and the otherto products that are adsorbed at the in-terface. As it remains uncertain whetherNAD• dimerizes in the interphase or in thesolution, the dashed arrows in Fig. 13 in-dicate that dimerization in the interphasewould be followed by rapid desorption. Inthe presence of high surfactant concen-tration (≈0.4 M), NAD+ can be reducedunder adsorption and diffusion control de-pending on both initial potential and scanrate. At starting potentials in which NAD+is not adsorbed (−0.9 V), only diffusion-controlled currents are observed on bothslow and fast scan rates. On the con-trary, starting at −0.1 V, where NAD+is adsorbed, and on rapid scan rate, re-duction occurs under both adsorption anddiffusion control. The former situation re-mains rather similar to that observed inthe presence of KCl with the adsorptionof both NAD+ and NAD• but with the ad-sorption of (NAD)2, at most transient, asno AC tensammetric peak appears [124].

The formation of the dimer is notstereospecific [75, 123, 395]. Solutions ofNAD+, electrolyzed at −1.1 V with mer-cury electrode produced a dimer mixturein a 90% yield consisting of two sets ofstereoisomers, 4,4′-dimers (90%) and 4,6′-dimers (10%), where each set consistedof three different stereoisomers [123]. Therelative abundance of the dimers was inde-pendent of pH (7–10) and no 6,6′-dimerscould be indentified by high-performanceliquid chromatography (HPLC) and NMRstudies. This contrasts with the predom-inant formation of 6,6′-dimers in theelectrochemical reduction of NAD ana-logues, particularly with small (e.g. H orCH3) one-substituents or in the presenceof DMSO (dimethyl sulfoxide) or ace-tonitrile [400]. When solutions of NAD+were reduced at −1.8 V, that is, at thelimiting current plateau of the secondwave, the result was the formation of 1,4-NADH (50–60%), 1,6-NADH (15–30%),and dimers (10–20%) [75, 123]. The ob-tained amount of 1,4-NADH increasedwhen pH was decreased from 10 to 7 [123].Consequently, and unlike enzymatic re-duction where stereoselectivity is inherentwith the addition of hydrogen at the 4 po-sition of the pyridine ring and from onlyone face of the plane of the ring (either reor si face), electrochemical reduction is notstereospecific.

The oxidation of the dimer (reaction13) has been thermodynamically and ki-netically studied with synthetic analoguesof NAD dimers at gold and platinumelectrodes [405] because (NAD)2 shows in-terest as a potential two-electron donoracting with no involvement of proton trans-fer, and its oxidation may provide a way ofgenerating the NAD• radical under milderconditions than that depicted in Sch. 1.Additionally, (NAD)2 may be able to par-ticipate in enzymatic reactions [406–408].


The NAD+ and (NAD)2 analogs form achemically reversible redox couple but witha large separation (0.7 − 1.2 V) betweenthe electrochemical reduction of NAD+and the oxidation of (NAD)2. It has beensuggested that the reason for this peak sep-aration lies in the different rate-limitingstep for both processes. In fact, the re-duction of NAD+ to (NAD)2 requires theirreversible radical–radical dimerizationof two NAD• radicals, whereas the oxi-dation of (NAD)2 into 2NAD+ demandsthe cleavage of the dimerization radical(NAD)2

•+ into NAD and NAD+ (Sch. 1).The reduction of NAD+ is kinetically con-trolled by the dimerization of the NAD•

radicals, while the oxidation of (NAD)2 in-volves mixed kinetic control by electrontransfer and cleavage of the (NAD)2

•+ rad-ical or total kinetic control by electrontransfer [405].

The reduction at the second polaro-graphic wave, resulting in the formationof NADH from NAD+ by the net trans-fer of two electrons and a proton, is notas clearly understood as the first wave.Studies in organic media, acetonitrile, andDMSO that are poor proton donors havebeen carried out to elucidate the secondreduction step. Two well-defined wavesare only observed with nicotinamide asa model compound, while NAD+ showsa 1e− wave at E1/2 ≈ −1.0 V in DMSO.A second wave only appears in the pres-ence of a proton donor at −1.99 V, theintensity of which can be equal to thefirst wave when increasing proton concen-tration [75, 400]. In aqueous media, thepH range over which this second wavecan be studied is restricted to pH 7 to 10,as a result of the background dischargecurrent and hydrolysis of the coenzyme.Within this range, the difference foundbetween the empirical variation of poten-tial with pH, in the order of 20 mV, and

the theoretical value of 59 mV indicatesthat H3O+ is not the only proton donorand/or buffer compounds must also actas proton donors [97]. This is closely re-lated to the very weak base character ofthe free radical NAD•, initially produced inthe first wave, which is not protonatedat pH 0.4 [409]. The reaction sequencefor this second wave has been suggestedas [97]:

NAD• + HX ⇐⇒

[NAD•−H−X]

e−−−−→

[NADH−X−]−X−

−−−→ NADH (55)

where HX is a proton donor. The laststep, which is a rapid first-order chemicalreaction, may involve rearrangement ofan initial dihydro product to 1,4-NADHand some 1,6-NADH, similar to theobservation with NAD analogues [410].The character of the proton donor stronglyinfluences the NAD• reduction as it hasbeen demonstrated by replacing K+ ionsin the background electrolyte by NH4

+ atpH 9 [97]. Two possible explanations havebeen described: first, the positive chargein the nicotinamide moiety for the NAD•

- NH4+ adduct favors radical reduction in

contrast to the noncharged NAD• –H2O,and second, the stronger acid characterof NH4

+ (pKa 9.2) versus H2O (pKa 15.7)would favor the last dissociation step ofEq. (55), shifting the equilibrium towardsNADH formation [97].

Complete electrochemical conversion ofNAD+ into NADH has been reported us-ing electrochemically produced potassiumamalgam in a specially designed reactor(10 M KOH, 1.5 A) and by electrochem-ically oxidizing NADH and undesirableproduced species, mainly NAD dimers,back to active NAD+ via a nickel foam an-ode [411]. However, the reduced coenzyme

4.8 Electrocatalytic Reduction of NAD(P)+ 131

proved to be only 70% biologically activeowing to the presence of species otherthan the desired 4-dihydro-isomer (27%6-dihydro, 1% 2-dihydro, and 2% mixeddimers) [412]. Higher yields, up to 95%in the absence of O2, are obtained byconstant-current electrolysis (0.3 mA) ofNAD-alginic acid covalent adducts [413,414]. An anion-charged membrane hasbeen used to divide an electrochemicalcell into two compartments where elec-trolysis at −0.9 V versus Ag/AgCl yielded85% conversion in 5 h [415]. As previ-ously reported [413, 414], the activity ofthe resulting NADH is strongly depen-dent on the cathodic overpotential owingto the favored dimerization rather thanthe formation of inactive NADH isomers.Dimerization is claimed to be hindered atcholesterol-modified gold amalgam elec-trodes, but the turnover value for NAD+remains below 0.5 h−1 [416] or at elec-trodes modified with a cholesteryl oleatelayer in the presence of divalent cationswith approximately 100% conversion into1,4-dihydro isomers [417].

4.8Electrocatalytic Reduction of NAD(P)+

The complications involved in the di-rect electrochemical reduction of NAD+has led to different mediated attemptsbased on nonenzymatic electrocatalysisand bioelectrocatalysis for the regenera-tion of enzymatically active NADH. How-ever, one-electron mediators face the samelimitations as described for direct reduc-tion. Dimerization can be only preventedif an additional regeneration enzyme isused that is able to accept two elec-trons in two steps from the mediator,and then to transfer one electron pair toNAD+ [418].

4.8.1Nonenzymatic Electroreduction ofNAD(P)+

Just as with NADH oxidation with media-tors, there has been a great controversyas to whether the reduction of NAD+occurs in a single one-hydride step orsequentially.

Photochemical enzyme-free methodshave been reported for the reductionof NAD(P) [419–421]. The Rh(bpy)3

3+derivatives have been successfully usedas reduction catalysts, enabling electro-chemical reduction [421–423]. The bet-ter stability of the bis-(tridentate) ligandRh(terpy)2

3+ allowed a mechanistic studyof the photoinduced reduction of NAD+,which seems to proceed by a hydride-transfer route according to the dependenceof NAD+ reduction on pH [424]. The in-termediate Rh hydride complex was identi-fied by 1H NMR spectroscopy [422]. A sim-ilar conclusion is drawn when the processis studied at modified electrodes with elec-tropolymerized pyrrole-substituted Rh(III)complexes in which hydrogenation in-volves the formation of an electrogeneratedhydride complex, followed by the insertionof NAD+ in the metal–hydride bond [425,426]. Electrochemical and spectroscopicstudies have demonstrated that Rh(III)complexes are more efficient than theCo(II) or Ir(III) partners for hydride trans-fer catalysis [427].

The electrochemical pathway for themediated reduction of NAD(P)+ througha hydride transfer mechanism with (pen-tamethylcyclopentadienyl-2,2′-bipyridinechloro) rhodium(III), (Cp∗RhIII(bpy)L)2+,has been described in detail [418, 422, 425].The rhodium(III) complex is reduced torhodium(II) and (I) with the concomitantloss of one monodentate ligand that greatly


facilitates the disproportionation of the in-termediary Rh(II) [428, 429]:

[Cp∗RhIII(bpy)L]2+ + e−

⇐⇒ [Cp∗RhII(bpy)L]+ (56)

[Cp∗RhII(bpy)L]+

⇐⇒ [Cp∗RhII(bpy)]+ + L (57)

[Cp∗RhII(bpy)L]+ + e−

⇐⇒ [Cp∗RhI(bpy)] + L (58)

[Cp∗RhII(bpy)]+ + e−

⇐⇒ [Cp∗RhI(bpy)] (59)

[Cp∗RhII(bpy)L]+ + [Cp∗RhII(bpy)]+

⇐⇒ [Cp∗RhIII(bpy)L]2+

+ [Cp∗RhI(bpy)] (60)

The Rh(I) complex can then incorporatea proton into its ligand sphere resultingin the hydrido complex that can transferits hydride (2e− + H+) to NAD(P)+ asfollows [418,425]:

[Cp∗RhI(bpy)] + H3O+ −−−→[Cp∗RhIH(bpy)]+ + H2O (61)

[Cp∗RhIH(bpy)]+ + NAD(P)+ + L −−−→[Cp∗RhIII(bpy)L]2+ + NAD(P)H (62)

The regioselective hydride transfer from[Cp∗Rh(bpy)H]+ to NAD+ to give 1,4-NADH has been appointed to the abil-ity of the amide group to coordi-nate to the ring-slipped Cp∗Rh metalcenter [70].

Similar requirements to those previouslyenumerated for the mediated oxidation ofNADH apply to the reduction of NAD+.In general, redox catalysts for the electro-chemical reduction of NAD+ should be

selective, stable (both chemically and elec-trochemically), with high reaction ratesand a redox potential that such directcathodic reduction of NAD+ leading toNAD dimers does not take place. Most ofthe nonenzymatic electrocatalytic reduc-tion of NAD+ is based on the mentionedrhodium (III) chemistry as two electronor hydride transfer selective mediatorsbecause direct hydride transfer to car-bonyl groups of dehydrogenase substratesis always much slower than the trans-fer to the NAD(P)+ [418] and additionally,they also show regioselectivity at the 4-position at ≈ −0.7 V versus SCE [430,431]. The efficiency of these catalysts forthe continuous regeneration of NADH isdemonstrated by the enzymatic reductionof pyruvate to D-lactate in homogeneousconditions producing an enantiomericexcess of 94% [430, 432], which com-pares very well with direct dehydrogenase-based catalytic reduction of NAD+ (re-action 2). Similar results are obtainedwith poly[pentamethylcyclopentadienyl-2,2′-bipyridine-chloro-rhodium(III)] mod-ified graphite foil electrodes [431] or PPyRh(terpy)2

3+ reticulated vitreous carbonelectrodes [433]. However, the reactionrate in these systems remains rather lowwith a maximum turnover for NAD+ of2 h−1, whereas strictly catalytic and homo-geneous turnover rates easily reach valuesof 2000 h−1 [434, 435].

The only exception to the rhodiumchemistry for the nonenzymatic electrocat-alytic reduction of NAD+ recalls the struc-ture of flavins and it was demonstratedwith aminopteridone derivatives [436]. Inthe same line, the low formal potential ofneutral red (NR) and the possibility of elec-tropolymerization resulted in poly(NR)-modified electrodes able to regenerateNADH at −0.6 versus Ag/AgCl andpH 6 [437].


4.8.2Enzymatic Electroreduction of NAD(P)+

As previously mentioned (reaction 8),hydrogenases catalyze the reduction ofNAD+ with molecular hydrogen achievingmaximum turnover values of 150 h−1 instrictly catalytic redox systems [438–441].Direct electrocatalytic reduction of NAD+by two different Alcaligenes eutrophusspecies and Desulfovibrio vulgaris hydro-genases was simultaneously suggested bythree groups in 1992 in GC [442], Pt gridscontaining 10% rhodium [443], and basalplane PG electrodes [444]. More recently,direct electron communication of Alcali-genes eutrophus H16 hydrogenase has alsobeen reported [445].

More efficient systems can be de-signed by mediated bioelectrocatalysisthat involve ferredoxin-NADP+ reductase(FNR)[81, 446–448], FMN-reductase [448],DI or lipoamide dehydrogenase [82, 358,359, 447, 449–452], viologen-acceptingpyridine nucleotide oxidoreductase [362,453], enoate reductase [454], formate de-hydrogenase [448, 455], and hydrogenaseas biocatalysts [442, 456]. Even though forsome of these enzymes direct electrontransfer with electrode surfaces is success-ful to some extent (DI [347, 457]) differentartificial or natural electron donors havebeen used as electron transfer media-tors to increase the charge transfer ef-ficiency. Quinones [358, 359, 447], methylor propyl viologen derivatives [81, 362, 442,446, 450–453, 456], dithiothreitol [82], andflavins [358, 359, 448, 449, 455] are used aselectron donors.

The overall equation can be depicted as:

NAD+ + Mredkred⇐⇒kox

NADH + Mox (63)

where the rate of NAD+ reduction, Rred,can be given by:

Rred = kcat,red[E]

1 + KM,NAD+

[NAD+]+ KM,Mred

[Mred]

(64)

Being kcat,red the catalytic constant forthe NAD+ reduction, [E] the enzyme con-centration, [NAD+], KM,NAD+ and [Mred],KM,Mred , the concentrations and Michaelis-Menten constants of NAD+ and mediatorin the reduced form, respectively. As pre-viously described, some NAD(P)-relatedflavoenzymes (e.g. DI, FNR) can also cat-alyze the oxidation of NAD(P)H and asimilar expression can also be applied.In the mediated bioelectroreduction ofNAD(P)+, Mred shuttles electrons from theoxidized enzyme (Eox) to electrodes obey-ing a ping-pong mechanism and then, Mox

is electrochemically reduced to Mred at ap-propriate electrode potentials. The electrontransfer from DI to NAD+ is thermody-namically unfavorable because the Eo′

ofthe flavoenzyme remains 42 mV more pos-itive at pH 8.5 than the Eo′

of NAD+ [358],see also above under Sect. 4.5. Redox cou-ples with Eo′

sufficiently lower than thatof the NAD(P)+/NAD(P)H pair could beefficient electron donors to drive the re-action toward NAD(P)+ reduction. Somequinones (e.g. adriaycin or alizarin redS) [358, 359, 447] and reduced methylviologen derivatives (MV•+) [451] renderreaction rates sufficiently high to pro-duce measurable electrocatalytic currentson the usual timescale of CV. This rateconstant has been reported in the rangeof 7 103 M−1 s−1 [452]. The redox kinet-ics between DI and the mediators hasbeen expressed by a Butler-Volmer typeequation, indicating a nonspecific and rel-atively simple electron transfer where thelog(kcat,red/KM,Mred ) or log(kcat,ox/KM,Mox )versus Eo′

M plots follow a linear Gibbs en-ergy relationship, at least from −0.7 to−0.25 V versus Ag/AgCl [359].


Special attention should be given to theuse of flavin-based mediators for the bio-electrocatalytic reduction of NAD(P)+. Infact, FAD or FMN spontaneously oxidizesNAD(P)H in living systems according tothe standard Gibbs energy variation atpH 7.0 between NADH and FAD, Go =−20.3 kJ mol−1. The reverse reaction isthen rather forbidden and according to theGibbs energy variation for the reduction ofNAD+ by FADH2:

Go = 60.8 − 5.79pH

+ RT ln[FAD][NADH]

[FADH2][NAD+](65)

the ratio [FAD][NADH]/[FADH2][NAD+]should be lower than 3 10−4 at pH 7.0 inorder to make the reaction thermodynam-ically favorable [455]. This has only beenachieved in thin-layer electrochemical ex-periments with a large electrode surfacearea in relation to the cell volume [455]. Re-duction of NAD(P)+ by FADH2 or FMNH2

in the presence of formate dehydrogenase,ferredoxin-NADP+-reductase, and FMN-reductase has been demonstrated underthese experimental conditions [448, 455,458]. Obviously, other alternatives to drivethis unfavorable equilibrium is the cou-pling of a dehydrogenase-based reaction toremove NADH in the vicinity of the elec-trode surface. In fact, steady state electro-catalytic currents with flavin nucleotidesand DI for the reduction of NAD+ havebeen only demonstrated electrochemicallyin the presence of lactate dehydrogenaseand an excess of pyruvate [359, 449].

Acknowledgment

Lo Gorton and Elena Domınguez thankThe Swedish Natural Science ResearchCouncil (NFR) and The Spanish Ministryof Science and Technology, respectively,

for financial support and Prof. GeorgeS. Wilson, University of Kansas, Lawrence,KS, USA for valuable discussions.

References

1. H. B. White, III Evolution of Coenzymes andthe Origin of Pyridine Nucleotides, AcademicPress, New York, 1982.

2. F. H. Westheimer, H. F. Fisher, E. E.Conn et al., J. Am. Chem. Soc. 1951, 73,2403–2408.

3. K.-S. You, CRC Crit. Rev. Biochem. 1985, 17,313–451.

4. D. W. Miles, D. W. Urry, J. Biol. Chem.1968, 243, 4181–4188.

5. S. F. Velick, J. Biol. Chem. 1958, 233,1455–1467.

6. A. P. Zens, T. J. Williams, J. C. Wisowatyet al., J. Am. Chem. Soc. 1975, 97,2850–2857.

7. R. M. Riddle, T. J. Williams, T. A. Brysonet al., J. Am. Chem. Soc. 1976, 98,4286–4290.

8. A. P. Zens, T. A. Bryson, R. B. Dunlap et al.,J. Am. Chem. Soc. 1976, 98, 7559–7564.

9. J. J. Tanner, S.-C. Tu, L. J. Barbour et al.,Protein Sci. 1999, 8, 1725–1732.

10. P. E. Smith, J. J. Tanner, J. Am. Chem. Soc.1999, 121, 8637–8644.

11. J. Jacobus, Biochemistry 1971, 10, 161–164.12. R. H. Sarma, V. Ross, N. O. Kaplan,

Biochemistry 1968, 7, 3052–3062.13. N. J. Oppenheimer, L. J. Arnold, N. O.

Kaplan, Proc. Natl. Acad. Sci. U.S.A. 1971,68, 3200–3205.

14. G. McDonald, B. Brown, D. Hollis et al.,Biochemistry 1972, 11, 1920–1930.

15. O. Jardetzky, N. G. Wade-Jardetzky, J. Biol.Chem. 1966, 241, 85–91.

16. R. H. Sarma, R. J. Mynott, J. Am. Chem. Soc.1973, 95, 7470–7480.

17. C. E. Bell, T. O. Yeates, D. Eisenberg,Protein Sci. 1997, 6, 2084–2096.

18. H. F. Fisher, E. E. Conn, B. Venneslandet al., J. Biol. Chem. 1953, 202, 687–692.

19. F. A. Loewus, F. H. Westheimer, B. Ven-nesland, J. Am. Chem. Soc. 1953, 75,5018–5021.

20. F. A. Loewus, P. Offner, H. F. Fisher et al.,J. Biol. Chem. 1953, 202, 699–706.

21. B. Vennesland, F. H. Westheimer in TheMechanisms of Enzyme Action (Eds.: W. D.


McElroy, B. Glass), Johns Hopkins Press,Baltimore, 1954, pp. 357–410.

22. G. Dryhurst, K. M. Kadish, F. Scheller et al.,Biological Electrochemistry, Academic Press,New York, 1982.

23. W. M. Clark, Oxidation-Reduction Potentialsof Organic Systems, Robert E. KriegerPublishing, Huntington, 1972.

24. H. K. Chenault, G. M. Whitesides, Appl.Biochem. Biotechnol. 1987, 14, 147–197.

25. F. L. Rodkey, J. Biol. Chem. 1955, 213,777–786.

26. F. L. Rodkey, J. Biol. Chem. 1959, 234,188–190.

27. F. L. Rodkey, J. A. Donovan, J. Biol. Chem.1959, 234, 677–680.

28. L. C. Clark, C. Lyons, Ann. N.Y. Acad. Sci.1962, 102, 29–45.

29. I. M. Shaw, Biochemical Fuel Cells, Proc. Am.Power Sources Conf. (1963) 17, pp. 53–56.

30. W. J. Blaedel, R. G. Haas, Anal. Chem. 1970,42, 918–927.

31. D. Thevenot, G. Hammouya, Experientia,Suppl. 1971, 18, 631–645.

32. H. Hanschmann, Stud. Biophys. 1974, 45,183–194.

33. I. Carelli, M. E. Cardinali, M. G. Bonicelli,Annali di Chimica 1977, 67, 89–96.

34. F. M. Martens, J. W. Verhoeven, R. A. Gaseet al., Tetrahedron 1978, 34, 443–446.

35. A. Kitani, K. Sasaki, Nippon Kagaku Kaishi1978, (6) 817–821.

36. Y. Ohnishi, M. Kitami, Bull. Chem. Soc. Jpn.1979, 52, 2674–2677.

37. I. Carelli, M. E. Cardinali, F. M. Moracci,J. Electroanal. Chem. 1980, 107, 391–404.

38. K. Sasaki, A. Kitani, A. Kunai et al., Bull.Chem. Soc. Jpn. 1980, 53, 3424–3429.

39. F. M. Martens, J. W. Verhoeven, Recl. Trav.Chim. Pays-Bas 1981, 100, 228–236.

40. M. Miller, B. Czochralska, D. Shugar,Bioelectrochem. Bioenerg. 1982, 9, 287–298.

41. M. F. Powell, T. C. Bruice, J. Am. Chem. Soc.1982, 104, 5834–5836.

42. F. M. Martens, J. W. Verhoeven, C. A. G. O.Varma et al., J. Photochem. 1983, 22,99–113.

43. S. Fukuzumi, K. Hironaka, N. Nishizawaet al., Bull. Chem. Soc. Jpn. 1983, 56,2220–2227.

44. S. Fukuzumi, Y. Kondo, T. Tanaka, J. Chem.Soc., Perkin Trans. 1984, (2) 673–680.

45. S. Fukuzumi, Y. Kondo, T. Tanaka, J. Chem.Soc., Chem. Commun. 1985, (15) 1053–1054.

46. N. W. Koper, S. A. Jonker, J. W. Verhoeven,Recl. Trav. Chim. Pays-Bas 1985, 104,296–302.

47. S. Fukuzumi, S. Koumitsu, K. Hironakaet al., J. Am. Chem. Soc. 1987, 109, 305–316.

48. S. Fukuzumi, M. Ishikawa, T. Tanaka,J. Chem Soc., Perkin Trans. 1989, 2,1811–1816.

49. S. Fukuzumi, S. Mochizuki, T. Tanaka,J. Am. Chem. Soc. 1989, 111, 1497–1499.

50. S. Fukuzumi, S. Mochizuki, T. Tanaka,Inorg. Chem. 1990, 29, 653–659.

51. J. C. Lepretre, D. Limosin, G. Pierre et al.,J. Electroanal. Chem. 1990, 286, 63–74.

52. J. Klippenstein, P. Arya, D. D. M. Wayner,J. Org. Chem. 1991, 56, 6736, 6737.

53. M. Okamura, T. Kashiwagi, Y. Mikata et al.,Tetrahedron Lett. 1991, 32, 1475–1478.

54. M. Okamura, T. Kashiwagi, Y. Mikata et al.,Chem. Lett. 1992, (7) 1247–1250.

55. S. Fukuzumi, Y. Tokuda, Chem. Lett. 1992,(9) 1721–1724.

56. J. C. Lepretre, D. Limosin, G. Pierre,J. Electroanal. Chem. 1992, 324, 115–125.

57. V. Glezer, J. Stradins, B. Turovska et al.,Electrochim. Acta 1992, 37, 277–279.

58. A. Anne, P. Hapitot, J. Moiroux et al., J. Am.Chem. Soc. 1992, 114, 4694–4701.

59. A. Anne, J. Moiroux, J. M. Saveant, J. Am.Chem. Soc. 1993, 115, 10 224–10 230.

60. M. Angulo, R. Marin Galvin, M. RuizMontoya et al., J. Electroanal. Chem. 1993,348, 303–315.

61. A. Anne, J. Moiroux, Can. J. Chem. 1995,73, 531–538.

62. A. Anne, S. Fraoua, P. Hapiot et al., J. Am.Chem. Soc. 1995, 117, 7412–7421.

63. I. Prieto, M. Angulo, J. M. RodrıguezMellado, J. Electroanal. Chem. 1995, 399,135–139.

64. M. Angulo, I. Prieto, J. M. RodrıguezMellado, J. Electroanal. Chem. 1995, 399,141–146.

65. T. Kajiki, N. Tamura, T. Nabeshima et al.,Chem. Lett. 1995, (11) 1063–1064.

66. E. Exposito, M. R. Montoya, M. An-gulo et al., J. Electroanal. Chem. 1996, 402,211–215.

67. A. Anne, S. Fraoua, V. Grass et al., J. Am.Chem. Soc. 1998, 120, 2951–2958.

68. J. C. Lepretre, D. Limosin, G. Pierre et al.,Eur. J. Org. Chem. 1998, (10) 2237–2243.

69. A. G. Rosales, M. R. Montoya, R. M. Galvinet al., Electroanalysis 1999, 11, 32–36.


70. H. C. Lo, O. Buriez, J. B. Kerr et al., Angew.Chem. Int. Ed. Engl. 1999, 38, 1429–1432.

71. S. Yasui, M. Okamura, M. Fujii, Rev.Heteroatom Chem. 1999, 20, 145–165.

72. R. J. Ansell, D. A. P. Small, C. R. Lowe,J. Mol. Catal. B: Enzyme 1997, 3, 239–252.

73. R. J. Ansell, D. A. P. Small, C. R. Lowe,J. Mol. Recognit. 1999, 12, 45–56.

74. R. J. Ansell, C. R. Lowe, Appl. Microbiol.Biotechnol. 1999, 51, 703–710.

75. P. J. Elving, C. O. Schmakel, K. S. V. San-thanam, Crit. Rev. Anal. Chem. 1976, 6,1–67.

76. I. Katakis, E. Domınguez, Mikrochim. Acta1997, 126, 11–32.

77. M. J. Lobo, A. J. Miranda, P. Tunon,Electroanalysis 1997, 9, 191–202.

78. B. Persson, L. Gorton, G. Johansson et al.,Enzyme Microb. Technol. 1985, 7, 549–552.

79. B. Persson, L. Gorton, G. Johansson,Bioelectrochem. Bioenerg. 1986, 16, 479–483.

80. G. Palmore, R. Tayhas, H. Bertschy et al.,J. Electroanal. Chem. 1998, 443, 155–161.

81. R. DiCosimo, C.-H. Wong, L. Daniels et al.,J. Org. Chem. 1981, 46, 4622,4623.

82. Z. E. Shaked, J. J. Barber, G. M. Whitesides,J. Org. Chem. 1981, 46, 4100,4101.

83. L. G. Lee, G. M. Whitesides, J. Am. Chem.Soc. 1985, 107, 7008–7018.

84. W. Hummel, M. R. Kula, Eur. J. Biochem.1989, 184, 1–13.

85. J. Everse, B. Anderson, K.-S. You, ThePyridine Nucleotide Coenzymes, AcademicPress, New York, 1982.

86. D. Dolphin, R. Poulson, O. Avramovic,Coenzymes and Cofactors, Wiley & Sons, NewYork, 1987.

87. D. Dolphin, R. Poulson, O. Avramovic,Pyridine Nucleotide Coenzymes: Chemical,Biochemical, and Medical Aspects, Part A,Wiley & Sons, New York, 1987.

88. D. Dolphin, R. Poulson, O. Avramovic,Pyridine Nucleotide Coenzymes: Chemical,Biochemical, and Medical Aspects, Part B,Wiley & Sons, New York, 1987.

89. W. J. Blaedel, R. A. Jenkins, Anal. Chem.1974, 46, 1952–1955.


91. R. D. Braun, K. S. V. Santhanam, P. J. Elv-ing, J. Am. Chem. Soc. 1975, 97, 2591–2598.

92. M. A. Jensen, P. J. Elving, Bioelectrochem.Bioenerg. 1978, 5, 526–534.

93. J. Moiroux, P. J. Elving, Anal. Chem. 1978,50, 1056–1062.


95. J. Moiroux, P. J. Elving, J. Electroanal. Chem.1979, 102, 93–108.

96. J. Moiroux, P. J. Elving, J. Am. Chem. Soc.1980, 102, 6533–6538.

97. P. J. Elving, W. T. Bresnahan, J. Moirouxet al., Bioelectrochem. Bioenerg. 1982, 9,365–378.

98. Z. Samec, P. J. Elving, J. Electroanal. Chem.1983, 144, 217–234.

99. R. L. Blankespoor, L. L. Miller, J. Elec-troanal. Chem. 1984, 171, 231–241.

100. M. F. Suaud-Chagny, F. G. Gonon, Anal.Chem. 1986, 58, 412–415.

101. P. Pantano, W. G. Kuhr, Anal. Chem. 1993,65, 623–630.

102. W. B. Nowall, W. G. Kuhr, Anal. Chem.1995, 67, 3583–3588.

103. W. B. Nowall, W. G. Kuhr, Electroanalysis1997, 9, 102–109.

104. M. A. Hayes, E. W. Kristensen, W. G. Kuhr,Biosens. Bioelectron. 1998, 13, 1297–1305.

105. M. A. Hayes, W. G. Kuhr, Anal. Chem. 1999,71, 1720–1727.

106. M. Aizawa, R. W. Coughlin, M. Charles,Biochim. Biophys. Acta 1975, 385,362–370.

107. R. W. Coughlin, M. Aizawa, B. F. Alexan-der et al., Biotechnol. Bioeng. 1975, 17,515–526.

108. H. Jaegfeldt, A. Torstensson, G. Johansson,Anal. Chim. Acta 1978, 97, 221–228.

109. H. Jaegfeldt, J. Electroanal. Chem. 1980, 110,295–302.

110. K. Takamura, A. Mori, F. Kusu, Bioelec-trochem. Bioenerg. 1981, 8, 229–238.

111. A. Silber, C. Braeuchle, N. Hampp,J. Electroanal. Chem. 1995, 390, 83–89.

112. X. Xing, M. Shao, C.-C. Liu, J. Electroanal.Chem. 1996, 406, 83–90.

113. Y. J. Xiao, J. P. Markwell, Langmuir 1997,13, 7068–7074.

114. Y. J. Xiao, T. Wang, X. Q. Wang et al.,J. Electroanal. Chem. 1997, 433, 49–56.

115. Y. J. Xiao, Y.-F. Chen, X. X. Gao, Spec-trochim. Acta, A, Mol. Biomol. Spectrosc.1999, 55, 1209–1218.

116. G. Li, J. Zhu, H. Fang et al., J. Electrochem.Soc. 1996, 143, L141, L142.

117. G. Li, Q. Gu, C. Fan et al., Anal. Lett. 1998,31, 1703–1715.


118. J. N. Burnett, A. L. Underwood, Biochem-istry 1965, 4, 2060–2064.

119. P. Leduc, D. Thevenot, J. Electroanal. Chem.1973, 47, 543–546.

120. P. Leduc, D. Thevenot, Bioelectrochem.Bioenerg. 1974, 1, 96–107.

121. R. W. Coughlin, B. F. Alexander, Biotechnol.Bioeng. 1975, 17, 1379–1382.

122. J. Ludvik, J. Volke, Anal. Chim. Acta 1988,209, 69–78.

123. H. Jaegfeldt, Bioelectrochem. Bioenerg. 1981,8, 355–370.

124. W. T. Bresnahan, P. J. Elving, J. Am. Chem.Soc. 1981, 103, 2379–2386.

125. M. A. Jensen, P. J. Elving, Biochim. Biophys.Acta 1984, 764, 310–315.

126. A. J. Cunningham, A. L. Underwood, Bio-chemistry 1967, 6, 266–271.

127. J. Grodowski, P. Neta, B. W. Carlson et al.,J. Phys. Chem. 1983, 87, 3135–3138.

128. T. Matsue, M. Suda, I. Uchida et al.,J. Electroanal. Chem. 1987, 234, 163–173.

129. B. W. Carlson, L. L. Miller, J. Am. Chem.Soc. 1983, 105, 7453,7454.

130. B. W. Carlson, L. L. Miller, P. Neta et al.,J. Am. Chem. Soc. 1984, 106, 7233–7239.

131. C. Amatore, J. M. Saveant, J. Electroanal.Chem. 1981, 123, 189–201.

132. J. F. Evans, T. Kuwana, M. T. Henne et al.,J. Electroanal. Chem. 1977, 80, 409–416.

133. J. Scheurs, J. Van den Berg, A. Wonderset al., Recl. Trav. Chim. Pays-Bas 1984, 103,251–259.

134. L. Falat, H. Y. Cheng, J. Electroanal. Chem.1983, 157, 393–397.

135. N. Cenas, J. Rozgaite, A. Pocius et al.,J. Electroanal. Chem. 1983, 154, 121–128.

136. N. Cenas, J. Kanapieniene, J. Kulys,J. Electroanal. Chem. 1985, 189, 163–169.

137. K. Ravichandran, R. P. Baldwin, J. Liq.Chromatogr. 1984, 7, 2031–2050.

138. R. L. McCreery in Electroanalytical Chemistry(Ed.: A. J. Bard), Marcel Dekker, New York,1991, pp. 221–374.

139. M. I. Alvarez-Gonzalez, S. B. Saidman,M. J. Lobo-Castanon et al., Anal. Chem.2000, 72, 520–527.

140. W. T. Bresnahan, J. Moiroux, Z. Samecet al., Bioelectrochem. Bioenerg. 1980, 7,125–152.

141. A. Kitani, Y.-H. So, L. L. Miller, J. Am.Chem. Soc. 1981, 103, 7636–7641.

142. A. Kitani, L. L. Miller, J. Am. Chem. Soc.1981, 103, 3595–3597.

143. B. W. Carlson, L. L. Miller, J. Am. Chem.Soc. 1985, 107, 479–485.

144. J. Hajdu, D. S. Sigman, Biochemistry 1977,16, 2841–2846.

145. S. Fukuzumi, N. Nishizawa, T. Tanaka,J. Org. Chem. 1984, 49, 3571–3578.

146. S. Fukuzumi, T. Tanaka, Chem. Lett. 1982,(10) 1513–1516.

147. L. L. Miller, J. R. Valentine, J. Am. Chem.Soc. 1988, 110, 3982–3989.

148. R. Szentrimay, P. Yeh, T. Kuwana inElectrochemical Studies of Biological Systems(Ed.: D. T. Sawyer), ACS, Washington, D.C., 1977, pp. 143–169.

149. P. N. Bartlett, P. Tebbutt, R. G. Whitaker,Prog. React. Kinet. 1991, 16, 55–155.

150. A. Heller, J. Phys. Chem. 1992, 96,3579–3587.

151. R. A. Marcus, Angew. Chem. Int. Ed. Engl.1993, 32, 1111–1121.

152. R. A. Marcus, N. Sutin, Biochim. Biophys.Acta 1985, 811, 265–322.

153. R. W. Murray in Electroanalytical Chemistry(Ed.: A. J. Bard), Marcel Dekker, New York,1984, pp. 191–368.

154. M. J. Eddowes, H. A. O. Hill, J. Chem. Soc.,Chem. Commun. 1977, (21) 771–772.

155. P. Yeh, T. Kuwana, Chem. Lett. 1977, (10)1145–1148.

156. I. V. Berezin, V. A. Bogdanovskaya, S. D.Varfolomeev et al., Dokl. Akad. Nauk SSSR(in Russian) 1978, 240, 615–618.

157. M. R. Tarasevich, A. I. Yaropolov, V. A. Bog-danovskaya et al., Bioelectrochem. Bioenerg.1979, 6, 393–403.

158. A. I. Yaropolov, V. Malovik, S. D. Var-folomeev et al., Dokl. Akad. Nauk. SSSR (inRussian) 1979, 249, 1399–1401.

159. D. C.-S. Tse, T. Kuwana, Anal. Chem. 1978,50, 1315–1318.

160. L. Gorton, J. Chem. Soc., Faraday Trans. 11986, 82, 1245–1258.

161. L. Gorton, E. Csoregi, E. Domınguez et al.,Anal. Chim. Acta 1991, 250, 203–248.

162. L. Gorton, B. Persson, P. D. Hale et al., ACSSymp. Ser. 1992, 487, 56–83.

163. P. N. Bartlett in Biosensor, Fundamentalsand Applications (Eds.: A. P. F. Turner,I. Karube, G. S. Wilson), Oxford SciencePublishing, Oxford, 1987, pp. 211–246.

164. W. Schuhmann, H.-L. Schmidt in Advancesin Biosensors (Ed.: A. P. F. Turner), JAIPress, London, 1992, pp. 79–130.


165. W. Schuhmann, J. Huber, H. Wohlschlae-ger et al., J. Biotechnol. 1993, 27, 129–142.

166. E. Lorenzo, F. Pariente, L. Hernandez et al.,Biosens. Bioelectron. 1998, 13, 319–332.

167. H. Jaegfeldt, T. Kuwana, G. Johansson,J. Am. Chem. Soc. 1983, 105, 1805–1814.

168. L. Gorton, Electroanalysis 1995, 7, 23–45.169. R. C. Engstrom, V. A. Strasser, Anal. Chem.

1984, 56, 136–141.170. R. C. Engstrom, Anal. Chem. 1982, 54,

2310–2314.171. K. Ravichandran, R. P. Baldwin, Anal.

Chem. 1984, 56, 1744–1747.172. J. Wang, T. Peng, Anal. Chem. 1986, 58,

1787–1790.173. E. J. Eisenberg, K. C. Cundy, Anal. Chem.

1991, 63, 845–847.174. G. Palleschi, H. S. Rathore, M. Mascini,

Electroanalysis 1989, 1, 199–203.175. W. G. Kuhr, V. L. Barrett, M. R. Gagnon

et al., Anal. Chem. 1993, 65, 617–622.176. S. Sampath, O. Lev, J. Electroanal. Chem.

1998, 446, 57–65.177. E. A. Yastrebova, I. V. Osipov, S. D. Var-

folomeev et al., Zhur. Anal. Khim. (Engl.Transl.) 1982, 37, 1278–1283.

178. E. A. Yastrebova, S. D. Varfolomeev, I. V.Osipov et al., Dokl. Akad. Nauk. SSSR (Engl.Transl.) 1982, 266, 681–684.

179. C. Ueda, D. C.-S. Tse, T. Kuwana, Anal.Chem. 1982, 54, 850–856.

180. R. S. Deinhammer, M. Ho, J. W. Anderegget al., Langmuir 1994, 10, 1306–1313.

181. Z. Wu, W. Jing, E. Wang, Electrochem.Commun. 1999, 1, 545–549.

182. E. Lorenzo, L. Sanchez, F. Pariente et al.,Anal. Chim. Acta 1995, 309, 79–88.

183. J.-J. Sun, J.-J. Xu, H.-Q. Fang et al.,Bioelectrochem. Bioenerg. 1997, 44, 45–50.

184. J.-J. Sun, H.-Q. Fang, H.-Y. Chen, Chem.Res. Chin. Univ. 1998, 14, 96 100.

185. C. Degrand, L. L. Miller, J. Am. Chem. Soc.1980, 102, 5728–5732.

186. M. Fukui, A. Kitani, C. Degrand et al., J. Am.Chem. Soc. 1982, 104, 28–33.

187. A. N. K. Lau, L. L. Miller, J. Am. Chem. Soc.1983, 105, 5271–5277.

188. F. Pariente, E. Lorenzo, H. D. Abruna, Anal.Chem. 1994, 66, 4337–4344.

189. F. Pariente, E. Lorenzo, F. Tobalina et al.,Anal. Chem. 1995, 67, 3936–3944.

190. F. Pariente, F. Tobalina, M. Darder et al.,Anal. Chem. 1996, 68, 3135–3142.

191. F. Pariente, F. Tobalina, G. Moreno et al.,Anal. Chem. 1997, 69, 4065–4075.

192. F. Tobalina, F. Pariente, L. Hernandez et al.,Anal. Chim. Acta 1998, 358, 15–25.

193. M. J. Batchelor, M. J. Green, C. L. Sketch,Anal. Chim. Acta 1989, 221, 289–294.

194. M. Kunitake, K. Akiyoshi, K. Kawatanaet al., J. Electroanal. Chem. 1990, 292,277–280.

195. H. R. Zare, S. M. Golabi, J. Electroanal.Chem. 1999, 464, 14–23.

196. H. Jaegfeldt, A. Torstensson, L. Gortonet al., Anal. Chem. 1981, 53, 1979–1982.

197. K. Ravichandran, R. P. Baldwin, J. Elec-troanal. Chem. 1981, 126, 293–300.

198. W. Schuhmann, R. Lammert, M. Haem-merle et al., Biosens. Bioelectron. 1991, 6,689–697.

199. G. Arai, M. Matsushita, I. Yasumori, NipponKagaku Kaishi 1985, (5) 894–897.

200. E. Katz, T. Loetzbeyer, D. D. Schlereth et al.,J. Electroanal. Chem. 1994, 373, 189–200.

201. E. Katz, I. Willner, J. Electroanal. Chem.1996, 418, 67–72.

202. I. Willner, A. Riklin, Anal. Chem. 1994, 66,1535–1539.

203. A. Bardea, E. Katz, A. F. Buckmann et al.,J. Am. Chem. Soc. 1997, 119, 9114–9119.

204. M. Lion-Dagan, E. Katz, I. Willner, J. Am.Chem. Soc. 1994, 116, 7913,7914.

205. E. Katz, B. Willner, I. Willner, Biosens.Bioelectron. 1997, 12, 703–719.

206. A. Curulli, I. Carelli, O. Trischitta et al.,Biosens. Bioelectron. 1997, 12, 1043–1055.

207. E. Katz, S. V. Heleg, A. Bardea et al., Biosens.Bioelectron. 1998, 13, 741–756.

208. I. Willner, E. Katz, B. Willner et al., Biosens.Bioelectron. 1997, 12, 337–356.

209. I. Willner, E. Katz, B. Willner, Electroanalysis1997, 9, 965–977.

210. Q. Wu, M. Maskus, F. Pariente et al., Anal.Chem. 1996, 68, 3688–3696.


212. I. C. Popescu, E. Domınguez, A. Narvaezet al., J. Electroanal. Chem. 1999, 464,208–214.

213. M. Hedenmo, A. Narvaez, E. Domınguezet al., Analyst 1996, 121, 1891–1895.

214. G. D. Storrier, K. Takada, H. D. Abruna,Inorg. Chem. 1999, 38, 559–565.

215. H. Huck, H. L. Schmidt, Angew. Chem. Int.Ed. Engl. 1981, 20, 402,403.


216. A. S. N. Murthy, J. Sharma, Talanta 1998,45, 951–956.

217. K. Ravichandran, R. P. Baldwin, Anal.Chem. 1983, 55, 1586–1591.

218. Q.-J. Chi, S.-J. Dong, J. Mol. Catal. A: Chem.1996, 105, 193–201.

219. H. Huck, Fresenius’ J. Anal. Chem. 1982,313, 548–552.

220. L. Gorton, A. Torstensson, H. Jaegfeldtet al., J. Electroanal. Chem. 1984, 161,103–120.

221. V. Laurinavicius, B. Kurtinaitiene, V. Gure-viciene et al., Anal. Chim. Acta 1996, 330,159–166.

222. G. Nagy, I. Kapui, L. Gorton, Sens. Actuators,B 1995, B24, 323–327.

223. G. Nagy, I. Kapui, L. Gorton, Anal. Chim.Acta 1995, 305, 65–73.

224. B. Persson, L. Gorton, J. Electroanal. Chem.1990, 292, 115–138.

225. S. A. Wing, J. P. Hart, Analyst 1992, 117,1215–1229.

226. B. Gruendig, G. Wittstock, U. Ruedel et al.,J. Electroanal. Chem. 1995, 395, 143–157.

227. G. Bremle, B. Persson, L. Gorton, Electro-analysis 1991, 3, 77–86.

228. J. Kulys, G. Gleixner, W. Schuhmann et al.,Electroanalysis 1993, 5, 201–207.

229. S. Yabuki, H. Shinohara, Y. Ikariyama et al.,J. Electroanal. Chem. 1990, 277, 179–187.

230. Y. Ikariyama, T. Ishizuka, H. Sinohara et al.,Denki Kagaku 1990, 58, 1097–1102.

231. S. D. Sprules, J. P. Hart, S. A. Wring et al.,Analyst 1994, 119, 253–257.

232. S. D. Sprules, J. P. Hart, R. Pittson et al.,Electroanalysis 1996, 8, 539–543.

233. L. T. Kubota, F. Gouvea, A. N. An-drade et al., Electrochim. Acta 1996, 41,1465–1469.

234. J. Wang, P. V. A. Pamidi, M. Jiang, Anal.Chim. Acta 1998, 360, 171–178.

235. L. T. Kubota, L. Gorton, Electroanalysis 1999,11, 719–728.

236. P. Hale, H.-S. Lee, Y. Okamoto, Anal. Lett.1993, 26, 1073–1085.

237. B. Persson, H. S. Lee, L. Gorton et al.,Electroanalysis 1995, 7, 935–940.

238. D. D. Schlereth, E. Katz, H. L. Schmidt,Electroanalysis 1995, 7, 46–54.

239. M. Ohtani, S. Kuwabata, H. Yoneyama,J. Electroanal. Chem. 1997, 422, 45–54.

240. C.-X. Cai, K.-H. Xue, J. Electroanal. Chem.1997, 427, 147–153.

241. A. A. Karyakin, E. E. Karyakina, W. Schuh-mann et al., Electroanalysis 1999, 11,553–557.

242. A. A. Karyakin, E. E. Karyakina, H. L.Schmidt, Electroanalysis 1999, 11, 149–155.

243. J.-L. Han, A.-M. Yu, H.-Y. Chen, HuaxueXuebao (Acta Chim. Sinica) 1995, 53,362–368.

244. D.-M. Zhou, H.-Q. Fang, H.-Y. Chen et al.,Anal. Chim. Acta 1996, 329, 41–48.

245. B. Persson, J. Electroanal. Chem. 1990, 287,61–80.

246. F. Ni, H. Feng, L. Gorton et al., Langmuir1990, 6, 66–73.

247. H. X. Ju, L. Dong, H. Y. Chen, Chem.J. Chin. Univ. (in Chinese) 1995, 16,1200–1203.

248. C. A. Pessoa, Y. Gushikem, L. T. Kubotaet al., J. Electroanal. Chem. 1997, 431, 23–27.

249. L. T. Kubota, F. Muteanu, A. Roddick-Lanzilotta et al., Quımica Analıtica 2000, 19(Suppl. 1), 15–27.

250. A. Malinauskas, T. Ruzgas, L. Gorton,J. Electroanal. Chem. 2000, 448, 55–63.

251. H. Huck, A. Schelter-Graf, J. Danzer et al.,Analyst 1984, 109, 147–150.

252. H. Huck, Frezenius’ J. Anal. Chem. 1982,313, 548–552.

253. H. Huck, A. Schelter-Graf, H. L. Schmidt,Bioelectrochem. Bioenerg. 1984, 13, 199–209.

254. H. Huck, Phys. Chem. Chem. Phys. 1999, 1,855–859.

255. L. Gorton, B. Persson, M. Polasek, G.Johansson in Contemporary ElectroanalyticalChemistry (ElectroFinnAnalysis) (Eds.: A.Ivaska, A. Lewenstam, R. Sara), PlenumPublishing, New York, 1991, pp. 183–189.

256. M. Polasek, L. Gorton, R. Appelqvist et al.,Anal. Chim. Acta 1991, 246, 283–292.

257. C. X. Cai, K. H. Xue, Anal. Chim. Acta 1997,343, 69–77.

258. L. I. Boguslavsky, L. Geng, I. Kovalev et al.,Biosens. Bioelectron. 1995, 10, 693–704.

259. Q. Chi, S. Dong, Electroanalysis 1995, 7,147–153.

260. Q. Shi, Q. Cheng, P. Zhang, Fenxi Huaxue1997, 25, 690–692.

261. A. Malinauskas, T. Ruzgas, L. Gorton et al.,Electroanalysis 2000, 12, 194–198.

262. A. Malinauskas, T. Ruzgas, L. Gorton,J. Colloid Interface Sci. 2000, 224, 325–332.

263. H. X. Ju, L. Dong, H. Y. Chen, Talanta 1996,43, 1177–1183.


264. M. J. Lobo-Castanon, S. L. Alvarez-Crespo,M. I. Alvarez-Gonzalez et al., Sci. Pap. Univ.Pardubice, Ser. A 1998, 3, 17–29.

265. C. Ramirez Molina, M. Boujtita, N. El Murr,Anal. Chim. Acta 1999, 401, 155–162.

266. D. D. Schlereth, E. Katz, H. L. Schmidt,Electroanalysis 1994, 6, 725–734.

267. C.-X. Cai, K.-H. Xue, Talanta 1998, 47,1107–1119.

268. D.-M. Zhou, J.-J. Sun, H.-Y. Chen et al.,Electrochim. Acta 1998, 43, 1803–1809.

269. Y. Okamoto, T. Kaku, R. Shundo, Pure Appl.Chem. 1996, 68, 1417–1421.

270. F. Torabi, K. Ramanathan, P.-O. Larssonet al., Talanta 1999, 50, 787–797.

271. V. U. Spohn, Nova Acta Leopold., Suppl.1998, 15, 155–175.

272. B. Persson, H. L. Lan, L. Gorton et al.,Biosens. Bioelectron. 1993, 8, 81–88.

273. Z. Huan, B. Persson, L. Gorton et al.,Electroanalysis 1996, 8, 575–581.

274. E. Domınguez, H. L. Lan, Y. Okamoto et al.,Biosens. Bioelectron. 1993, 8, 167–175.

275. E. Domınguez, H. L. Lan, Y. Okamoto et al.,Biosens. Bioelectron. 1993, 8, 229–237.

276. H.-C. Shu, B. Mattiasson, B. Persson et al.,Biotechnol. Bioeng. 1995, 46, 270–279.

277. H.-C. Shu, L. Gorton, B. Persson et al.,Biotechnol. Bioeng. 1995, 46, 280–284.

278. Q. Chi, S. Dong, Analyst 1994, 119,1063–1066.

279. S. Cosnier, K. Le Lous, J. Electroanal. Chem.1996, 406, 243–246.

280. A. Silber, N. Hampp, W. Schuhmann,Biosens. Bioelectron. 1996, 11, 215–223.

281. Y. Wang, D.-M. Zhou, H.-Y. Chen, Chem.Res. Chin. Univ. 1997, 13, 276–281.

282. L. Gorton, G. Bremle, E. Csoeregi et al.,Anal. Chim. Acta 1991, 249, 4354.

283. J. J. Xu, H. Q. Fang, H. Y. Chen, Chem. J.Chin. Univ. (in Chinese) 1997, 18, 706–710.

284. H.-Y. Chen, D.-M. Zhou, J.-J. Xu et al.,J. Electroanal. Chem. 1997, 422, 21–25.

285. K. Hajizadeh, H. T. Tang, H. B. Halsallet al., Anal. Lett. 1991, 24, 1453–1469.

286. A.-M. Yu, J.-L. Han, K.-S. Yang et al.,Gaodeng Xuexiao Huaxue Xuebao (Chem.J. Chin. Univ.) (in Chinese) 1995, 16,1204–1206.

287. K. Tanaka, S. Ikeda, N. Oyama et al., Anal.Sci. 1993, 9, 783–789.

288. K. Tanaka, S. Ikeda, N. Oyama et al., Anal.Sci. 1993, 9, 783–789.

289. C.-X. Cai, H.-X. Ju, H.-Y. Chen, GaodengXuexiao Huaxue Xuebao (Chem. J. Chin.Univ.) (in Chinese) 1995, 16, 368–372.

290. T. Ohsaka, K. Tanaka, K. Tokuda, J. Chem.Soc., Chem. Commun. 1993, (3) 222–224.

291. A. Torstensson, L. Gorton, J. Electroanal.Chem. 1981, 130, 199–207.

292. Y. Kimura, K. Niki, Anal. Sci. 1985, 1,271–274.

293. O. Miyawaki, T. Yano, Enzyme Microb.Technol. 1992, 14, 474–478.

294. O. Miyawaki, T. Yano, Enzyme Microb.Technol. 1993, 15, 525–529.

295. I. Carelli, I. Chiarotto, A. Curulli, Curr. Top.Electrochem. 1994, 3, 141–157.

296. A. Curulli, I. Carelli, O. Trischitta et al.,Talanta 1997, 44, 1659–1669.

297. M. Vreeke, R. Maidan, A. Heller, Anal.Chem. 1992, 64, 3084–3090.

298. M. J. Lobo, A. J. Miranda, J. M. Lopez-Fonseca et al., Anal. Chim. Acta 1996, 325,33–42.

299. M. J. L. Castanon, A. J. M. Ordieres, P.Tunon Blanco, Biosens. Bioelectron. 1997,12, 511–520.

300. A. B. Florou, M. I. Prodromidis, M. I.Karayannis et al., Electroanalysis 1998, 10,1261–1268.

301. L. Gorton, G. Johansson, A. Torstensson,J. Electroanal. Chem. 1985, 196, 81–92.

302. P. N. Bartlett, P. R. Birkin, E. N. K. Wallace,J. Chem. Soc., Faraday Trans. 1997, 93,1951–1960.

303. M. Somasundrum, J. V. Bannister, J. Chem.Soc., Chem. Commun. 1993, (21) 1629–1631.

304. K. Tanaka, K. Tokuda, T. Ohsaka, J. Chem.Soc., Chem. Commun. 1993, (23) 1770–1772.

305. L. T. Kubota, L. Gorton, J. Solid StateElectrochem. 1999, 3, 370–379.

306. A. Malinauskas, T. Ruzgas, L. Gorton,Bioelectrochem. Bioenerg. 1999, 49, 21–27.

307. O. Miyawaki, L. B. Wingard, Biochim.Biophys. Acta 1985, 838, 60–68.

308. H. Shinohara, Proc. SPIE-Int. Soc. Opt. Eng.1996, 2716, 177–182.

309. L. T. Kubota, L. Gorton, A. Roddick-Lanzilotta et al., Bioelectrochem. Bioenerg.1998, 47, 39–46.

310. K. J. Stine, D. M. Andrauskas, A. R. Khanet al., J. Electroanal. Chem. 1999, 472,147–156.

311. C. X. Cai, K. H. Xue, Microchem. J. 1998, 58,197–208.


312. N. Mano, A. Kuhn, J. Electroanal. Chem.1999, 477, 79–88.

313. N. Mano, A. Kuhn, Electrochem. Commun.1999, 1, 497–501.

314. E. Casero, M. Darder, K. Takada et al.,Langmuir 1999, 15, 127–134.

315. V. Laurinavicius, B. Kurtinaitiene, V. Li-auksminas et al., Monatshefte fur Chemie1999, 130, 1269–1281.

316. P. C. Pandey, S. Upadhyay, B. C. Upadhyayet al., Anal. Biochem. 1998, 260, 195–203.

317. P. C. Pandey, Trans. Indian Inst. Met. 1998,51, 319–325.

318. P. C. Pandey, S. Upadhyay, H. C. Pathaket al., Anal. Lett. 1998, 31, 2327–2348.

319. A. S. N. Murthy, Anita, Bioelectrochem.Bioenerg. 1994, 33, 71–73.

320. A. S. N. Murthy, Anita, R. L. Gupta, Anal.Chim. Acta 1994, 289, 43–46.

321. P. C. Pandey, Anal. Biochem. 1994, 221,392–396.

322. J. Kulys, Biosensors 1986, 2, 3–13.323. J. Kulys, Enzyme Microb. Technol. 1981, 3,

344–352.324. W. J. Albery, P. N. Bartlett, J. Chem. Soc.,

Chem. Commun. 1984, (4) 234–236.325. W. J. Albery, P. N. Bartlett, A. E. G. Cass

et al., J. Chem. Soc., Faraday Trans. 1 1986,82, 1033–1050.

326. W. J. Albery, P. N. Bartlett, A. E. G. Casset al., J. Electroanal. Chem. 1987, 218,127–134.

327. K. McKenna, S. E. Boyette, A. Brajter-Toth,Anal. Chim. Acta 1988, 206, 75–84.

328. J. J. Kulys, U. Bilitewski, R. D. Schmid,Anal. Lett. 1991, 24, 181–189.

329. C. J. Stanley, R. B. Cox, M. F. Cardosi et al.,J. Immunol. Methods 1988, 112, 153–161.

330. S. Zhao, U. Korell, L. Cuccia et al., J. Phys.Chem. 1992, 96, 5641–5652.

331. S. Zhao, R. B. Lennox, J. Electroanal. Chem.1993, 346, 161–173.

332. J. Wang, T. Golden, Anal. Chim. Acta 1989,217, 343–351.

333. F. Xu, H. Li, S. J. Cross et al., J. Electroanal.Chem. 1994, 368, 221–225.

334. L. Angnes, C. M. N. Azevedo, K. Araki et al.,Anal. Chim. Acta 1996, 329, 91–96.

335. B. F. Yon Hin, C. R. Lowe, Anal. Chem.1987, 59, 2111–2115.

336. C.-X. Cai, K.-H. Xue, Y.-M. Zhou et al.,Talanta 1997, 44, 339–347.

337. T. N. Rao, I. Yagi, T. Miwa et al., Anal.Chem. 1999, 71, 2506–2511.

338. A. Fujishima, T. N. Rao, E. Popa et al.,J. Electroanal. Chem. 1999, 473, 179–185.

339. A. D. Ryabov, V. S. Kurova, V. N. Goralet al., Chem. Mater. 1999, 11, 600–604.

340. N. F. Atta, A. Galal, E. Karagozler et al.,J. Chem. Soc., Chem. Commun. 1990, (19)1347–1349.

341. N. F. Atta, A. Galal, A. E. Karagozler et al.,Biosens. Bioelectron. 1991, 6, 333–341.

342. H. Ju, D. Leech, Anal. Chim. Acta 1997, 345,51–58.

343. J. Wang, F. Lu, L. Angnes et al., Anal. Chim.Acta 1995, 305, 3–7.

344. J. Wang, Q. Chen, M. Pedrero et al., Anal.Chim. Acta 1995, 300, 111–116.

345. J. Wang, P. V. A. Pamidi, C. L. Ren-schler et al., J. Electroanal. Chem. 1996, 404,137–142.

346. C. J. McNeil, J. A. Spoors, J. M. Cooperet al., Anal. Chim. Acta 1990, 237, 99–105.

347. D. Kobayashi, S. Ozawa, T. Mihara et al.,Denki Kagaku 1992, 60, 1056–1062.

348. H. Ukeda, M. Imabayashi, K. Mat-sumoto et al., Agric. Biol. Chem. 1989, 53,2909–2915.

349. M. J. Green, H. A. O. Hill, J. Chem. Soc.,Faraday Trans. I 1986, 82, 1237–1243.

350. Y. Kashiwagi, T. Osa, Chem. Lett. 1993, (4)677–680.

351. Y. Kashiwagi, Q. Pan, Y. Yanagisawa et al.,Denki Kagakuoyobi Kogyo Butsuri Kagaku1994, 62, 1240–1246.

352. T. Sawaguchi, T. Matsue, I. Uchida,Bioelectrochem. Bioenerg. 1992, 29, 127–133.

353. H. C. Chang, A. Ueno, H. Yamada et al.,Denki Kagaku oyobi Kogyo Butsuri Kagaku1990, 58, 1211–1212.

354. M. Comtat, M. Galy, P. Goulas et al., Anal.Chim. Acta 1988, 208, 295–300.

355. K. Miki, T. Ikeda, S. Todoriki et al., Anal.Sci. 1989, 5, 269–274.

356. S. Todoriki, K. Miki, T. Ikeda et al., DenkiKagaku oyobi Kogyo Butsuri Kagaku 1990,58, 1089–1096.

357. S. Ozawa, T. Ikeda, M. Senda, Anal. Sci.1991, 7, 1689–1692.

358. Y. Ogino, K. Takagi, K. Kano et al.,J. Electroanal. Chem. 1995, 396, 517–524.

359. K. Takagi, K. Kano, T. Ikeda, J. Electroanal.Chem. 1998, 445, 211–219.

360. S. Kunugi, K. Ikeda, T. Nakashima et al.,Polym. Bull. 1990, 24(2), 247–250.

361. H. C. Chang, A. Ueno, H. Yamada et al.,Analyst 1991, 116, 793–796.


362. S. Cosnier, K. LeLous, Talanta 1996, 43,331–337.

363. I. Karube, K. Yokoyama, E. Tamiya, Biosens.Bioelectron. 1993, 8, 219–228.

364. Y. Miwa, M. Nishizawa, T. Matsue et al.,Denki Kagaku oyobi Kogyo Butsuri Kagaku1994, 62, 1256,1257.

365. M. Montagne, H. Durliat, M. Comtat, Anal.Chim. Acta 1993, 278, 25–33.

366. T. Noguer, J. L. Marty, Enzyme Microb.Technol. 1995, 17, 453–456.

367. M. Montagne, H. Erdmann, M. Com-tat et al., Sens. Actuators, B 1995, B27,440–443.

368. T. Noguer, J. L. Marty, Anal. Chim. Acta1997, 347, 63–69.

369. J. Katrlik, A. Pizzariello, V. Mastihuba et al.,Anal. Chim. Acta 1999, 379, 193–200.

370. T. Suzuki, K. Yamamoto, Y. Tanaka et al.,Maku 1989, 14, 319–328.

371. H.-Z. Bu, S. R. Mikkelsen, A. M. English,Anal. Chem. 1998, 70, 4320–4325.

372. T. Matsue, N. Kasai, M. Narumi et al.,J. Electroanal. Chem. 1991, 300, 111–118.

373. T. Huang, A. Warsinke, T. Kuwana et al.,Anal. Chem. 1998, 70, 991–997.

374. G. de Oliveira Neto, J. Rover, L. T. Kubota,Electroanalysis 1999, 11, 527–533.

375. C. J. McNeil, J. A. Spoors, D. Cocco et al.,Anal. Chem. 1989, 61, 25–29.

376. M. Somasundrum, J. Hall, J. V. Bannister,Anal. Chim. Acta 1994, 295, 47–57.

377. Z. Liu, O. Niwa, T. Horiuchi et al., Biosens.Bioelectron. 1999, 14, 631–638.

378. S. Cosnier, M. Fontecave, D. Limosin et al.,Anal. Chem. 1997, 69, 3095–3099.

379. S. Cosnier, M. Fontecave, C. Innocent et al.,Electroanalysis 1997, 9, 685–688.

380. S. Cosnier, J.-L. Decout, M. Fontecave et al.,Electroanalysis 1998, 10, 521–525.

381. J. M. Laval, C. Bourdillon, J. Moiroux, J. Am.Chem. Soc. 1984, 106, 4701–4706.

382. J. Wang, M. S. Lin, Anal. Chem. 1988, 60,499–502.

383. J. Wang, E. Gonzalez-Romero, Electroanaly-sis 1993, 5, 427–430.

384. L. Campanella, T. Ferri, M. P. Sammartinoet al., J. Mol. Catal. 1987, 43, 153–159.

385. A. Fujishima, E. Popa, Z. Wu et al. in NovelTrends in Electroorganic Synthesis (Ed.: F.Torii), Springer, Tokyo, 1998, pp. 421–424.

386. R. W. Murray, Molecular Design of ElectrodeSurfaces, Wiley-Interscience, New York,1992.

387. E. Laviron, J. Electroanal. Chem. 1979, 101,19–28.

388. E. Laviron in Electroanalytical Chemistry(Ed.: A. J. Bard), Marcel Dekker, New York,1982, pp. 53–151.

389. W. Albery, A. R. Hillman, J. Electroanal.Chem. 1984, 170, 27–49.

390. C. P. Andrieux, J. M. Dumas-Bouchiat,J. M. Saveant, J. Electroanal. Chem. 1984,169, 9–21.

391. C. P. Andrieux, J. M. Saveant, J. Electroanal.Chem. 1984, 171, 65–93.

392. C. P. Andrieux, J. M. Saveant, J. Electroanal.Chem. 1978, 93, 163–168.

393. S. Fukuzumi, N. Nishizawa, T. Tanaka,J. Chem. Soc., Perkin Trans. 2 1985,371–378.

394. L. Gorton, G. Johansson, J. Electroanal.Chem. 1980, 113, 151–158.

395. A. L. Underwood, J. N. Burnett in Electro-analytical Chemistry (Ed.: A. J. Bard), MarcelDekker, New York, 1973, pp. 1–85.

396. M. A. Jensen, W. T. Bresnahan, P. J. Elving,Bioelectrochem. Bioenerg. 1983, 11, 299–306.

397. R. C. Kaye, H. I. Stonehill, J. Chem. Soc.1952, 56, 3244–3247.

398. J. Moiroux, S. Deycard, T. Malinski,J. Electroanal. Chem. 1985, 194, 99–108.

399. C. O. Schmakel, M. A. Jensen, P. J. Elving,Bioelectrochem. Bioenerg. 1978, 5, 625–634.

400. K. S. V. Santhanam, P. J. Elving, J. Am.Chem. Soc. 1973, 95, 5482–5490.

401. C. O. Schmakel, K. S. V. Santhanam, P. J.Elving, J. Am. Chem. Soc. 1975, 97,5083–5092.

402. E. J. Land, A. J. Swallow, Biochim. Biophys.Acta 1968, 162, 327–337.

403. B. H. J. Bielski, P. C. Chan, J. Am. Chem.Soc. 1980, 102, 1713–1716.

404. W. T. Bresnahan, P. J. Elving, Biochim.Biophys. Acta 1981, 678, 151–156.

405. A. Anne, P. Hapiot, J. Moiroux et al.,J. Electroanal. Chem. 1992, 331, 959–970.

406. B. Czochralska, M. Szweykowska, D.Shugar, Arch. Biochem. Biophys. 1980, 199,497–505.

407. V. Carelli, F. Liberatore, A. Casini et al.,Bioorg. Chem. 1980, 9, 342–351.

408. L. Avigliano, V. Carelli, A. Casini et al.,Biochim. Biophys. Acta 1983, 723, 372–375.

409. E. M. Kosower, A. Teuerstein, H. D. Burowset al., J. Am. Chem. Soc. 1978, 100,5185–5190.


410. Y. Ohnishi, Y. Kikuchi, M. Kitami, Tetrahe-dron Lett. 1979, 3005–3008.

411. F. G. Drakesmith, A Process and Apparatusfor the Electrochemical Regeneration of Co-enzymes; Eur. Pat. Appl. 275649, A1, 1988.

412. F. G. Drakesmith, B. Gibson, J. Chem. Soc.,Chem. Commun. 1988, (22) 1493–1494.

413. M. Aizawa, R. W. Coughlin, M. Charles,Biochim. Biophys. Acta 1976, 440, 233–240.

414. M. Aizawa, R. W. Coughlin, M. Charles,Biotechnol. Bioeng. 1976, 18, 209–215.

415. S.-E. Yun, M. Taya, S. Tone, Biotechnol. Lett.1994, 16, 1053–1058.

416. S. H. Baik, C. Kang, C. Jeon et al., Biotechnol.Tech. 1999, 13, 1–5.

417. M. Aizawa, S. Suzuki, M. Kubo, Biochim.Biophys. Acta 1976, 444, 886–892.

418. E. Steckhan, Top. Curr. Chem. 1994, 170,83–111.

419. D. Mandler, I. Willner, J. Am. Chem. Soc.1984, 106, 5352, 5353.

420. P. Cuendet, M. Graetzel, Photochem. Photo-biol. 1984, 39, 609–612.

421. R. Wienkamp, E. Steckhan, Angew. Chem.Int. Ed. Engl. 1983, 22, 782,783.

422. E. Steckhan, S. Herrmann, R. Ruppert et al.,Organometallics 1991, 10, 1568–1577.

423. Y. Shimizu, A. Kitani, S. Ito et al., DenkiKagaku 1994, 62, 1233,1234.

424. K. Umeda, A. Nakamura, F. Toda, Bull.Chem. Soc. Jpn. 1993, 66, 2260–2267.

425. E. Hoefer, E. Steckhan, B. Ramos et al.,J. Electroanal. Chem. 1996, 402, 115–122.

426. C. Caix, S. Chardon-Noblat, A. Deronzieret al., J. Organomet. Chem. 1997, 540,105–111.

427. W. Kaim, R. Reinhardt, E. Waldhor et al.,J. Organomet. Chem. 1996, 524, 195–202.

428. U. Kolle, M. Grutzel, Angew. Chem. Int. Ed.Engl. 1987, 26, 567–570.

429. U. Koelle, A. D. Ryabov, Mendeleev Com-mun. 1995, (5) 187–189.

430. R. Ruppert, S. Herrmann, E. Steckhan,Tetrahedron Lett. 1987, 28, 6583–6586.

431. S. Cosnier, H. Gunther, J. Electroanal.Chem. 1991, 315, 307–312.

432. R. Ruppert, M. Franke, S. Herrmann et al.,DECHEMA-Monogr. 1989, 112, 13–23.

433. M. Beley, J. P. Collin, J. Mol. Catal. 1993,79, 133–140.

434. A. Nakamura, H. Minami, I. Urabe et al.,J. Ferment. Technol. 1988, 66, 267–272.

435. M. Persson, M. O. Mansson, L. Bulow et al.,Bio-Technology 1991, 9, 280–284.

436. S. Kwee, H. Lund, Bioelectrochem. Bioenerg.1974, 1, 87–95.

437. A. A. Karyakin, O. A. Bobrova, E. E.Karyakina, J. Electroanal. Chem. 1995, 399,179–184.

438. B. Danielsson, F. Winqvist, J. Y. Malpoteet al., Biotechnol. Lett. 1982, 4, 673–678.

439. K. Otsuka, S. Aono, I. Okura, Chem. Lett.1987, (10) 2089–2090.

440. M. Takeuchi, I. Okura, F. Hasumi, J. Mol.Catal. 1991, 68, L21–L23.

441. Q. H. Wong, L. Daniels, W. H. Orme-Johnson et al., J. Am. Chem. Soc. 1981, 103,6227,6228.

442. D. D. Schlereth, V. M. Fernandez, M.Sanchez-Cruz et al., Bioelectrochem. Bioen-erg. 1992, 28, 473–482.

443. J. Cantet, A. Bergel, M. Comtat et al., J. Mol.Catal. 1992, 73, 371–380.

444. T. Sagara, H. Hirayama, K. Akutsu et al.,Bioelectrochem. Bioenerg. 1992, 28, 191–204.

445. K. Delecouls, P. Saint-Aguet, C. Zaboroschet al., J. Electroanal. Chem. 1999, 468,139–149.

446. M. Ito, T. Kuwana, J. Electroanal. Chem.1971, 32, 415–425.

447. K. Kano, K. Takagi, Y. Ogino et al., Chem.Lett. 1995, (7) 589–590.

448. A. Bergel, M. Comtat, Bioelectrochem. Bioen-erg. 1992, 27, 495–500.

449. K. Takagi, K. Kano, T. Ikeda, Chem. Lett.1996, (1) 11–12.

450. T. Matsue, H. C. Chang, I. Uchida et al.,Tetrahedron Lett. 1988, 29, 1551–1554.

451. H. C. Chang, T. Matsue, I. Uchida et al.,Chem. Lett. 1989, (7) 1119–1122.

452. D. D. Schlereth, V. M. Fernandez, Biotech-nol. Tech. 1990, 4, 201–204.

453. H. Gunther, A. S. Paxinos, M. Schulz et al.,Angew. Chem. Int. Ed. Engl. 1990, 29,1053–1055.

454. H. Simon, H. Gunther, J. Bader et al.,Angew. Chem. Int. Ed. Engl. 1981, 20,861–863.

455. A. Bergel, M. Comtat, J. Electroanal. Chem.1991, 302, 219–231.

456. D. D. Schlereth, V. M. Fernandez, Biotech-nol. Lett. 1989, 11, 407–410.

457. T. Larsson, A. Lindgren, T. Ruzgas, Bioelec-trochemistry 2001, 53, 243–249.

458. A. Bergel, A. Courteix, J. Cantet et al., Re-cents Prog. Genie Procedes 1993, 7, 231–236.

145

5Electrochemical Immunoassay

C. Ajith Wijayawardhana, H. Brian Halsall, and William R. HeinemanUniversity of Cincinnati, Cincinnati, Ohio

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

5.2 Enzyme Electrochemical Immunoassays . . . . . . . . . . . . . . . . . . . 1485.2.1 Electrochemical Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1485.2.2 The E-S-P System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1585.2.3 Heterogeneous Enzyme Immunoassay . . . . . . . . . . . . . . . . . . . . 1595.2.4 Ultrasensitive Heterogeneous Enzyme Assays in Small Volumes . . . 1615.2.5 Homogeneous Enzyme Immunoassay . . . . . . . . . . . . . . . . . . . . 163

5.3 Nonenzyme Electrochemical Immunoassays . . . . . . . . . . . . . . . . 1645.3.1 Nonenzyme Homogeneous Immunoassays . . . . . . . . . . . . . . . . . 1645.3.2 Heterogeneous Nonenzymatic Immunoassays . . . . . . . . . . . . . . . 164

5.4 Electrochemical Immunosensors . . . . . . . . . . . . . . . . . . . . . . . . 166

5.5 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170


147

5.1Introduction

The development of immunoassay is oneof the great success stories in bioanalyti-cal chemistry. Its success is due largely tothe extraordinarily selective and versatilereagent provided by nature in the form ofantibody (Ab). As part of the immune de-fense system in animals, antibodies withhigh specificity can be synthesized in rea-sonable quantity by an organism withinweeks of being injected with a foreignspecies called an antigen (Ag). It is es-timated that an animal can synthesizeup to 107 to 108 antibodies of differentspecificities [1, 2]. This variability is thesource of the remarkable versatility of im-munoassay in detecting a broad range ofanalytes.

Traditionally, antibodies for immunoas-says have been obtained from the antiseraof animals immunized with the Ag of in-terest. Such antisera contain a collectionof different antibodies called polyclonalantibodies, each specific for a differentsite (epitope) of the Ag. An Ag typi-cally needs to be greater than approxi-mately 1500 Daltons (Da) to elicit efficientAb formation [1]. However, antibodies forsmaller molecules, such as therapeuticdrugs, can be obtained by covalently at-taching the smaller molecule (hapten) to

some carrier Ag, usually a large protein or amultiple antigenic peptide (MAP) [3]. Theresulting conjugate elicits the formationof antibodies, some of which are directedtoward the hapten. Use of polyclonal an-tibodies, however, is complicated by theinconsistency in the polyclonal antiserafrom different animals or even amongdifferent batches from the same animal.This led to the development of mono-clonal antibodies [4], which have a known,reproducible specificity and affinity. Mon-oclonal antibodies are obtained in largequantities from cell cultures grown fromcells retrieved from the immunized animaland selected for producing the desired Ab.The more recently developed recombinantantibodies using molecular engineeringmethods are also expected to be widelyused in the future because of their low costand improved or novel specificities [5, 6].

The origins of using Ab electrochemicalimmunoassay (ECI) can be traced to thepioneering efforts of Breyer and Radcliff,who in 1951 used polarography to fol-low the interaction of azo-labeled proteinwith specific antiserum [7]. Although thispredated radioimmunoassay – commonlymisperceived to be the first type ofimmunoassay – by several years, it tookmore than a quarter of a century for de-velopments in biochemistry, electronics,immunology, and analytical techniques to

148 5 Electrochemical Immunoassay

fuse into a methodology now capable of al-most unrivaled sensitivity. Also, ECI is freefrom the potentially hazardous radioac-tive labels used in radioimmunoassays,and can be applied directly to coloredand/or turbid samples that are difficultfor immunoassays with optical detection.Another advantage of ECI is the ability towork with extremely small samples be-cause electrochemistry is an interfacialprocess where the important electrode pro-cesses occur at the electrode interface, notin bulk solution. This can be very impor-tant in situations where minimizing thesample volume is critical, as in the clinicaltesting of neonates, the blood analysis of illor seriously injured patients, and the test-ing of cerebral and spinal fluids [8, 9]. Also,working in small volumes is of paramountimportance in enzyme immunoassays, themost common type of immunoassay, be-cause the enzymatic product is brought todetectable levels faster in small volumeswhere product dilution is restricted. Thishas been exploited in immunoassays inmicrocapillaries where analyte amounts ofa few zeptomoles (10−21 moles) have beendetected [10].

In this chapter, we first present the dif-ferent formats used in ECI according tothe type of label used. Since enzymes areby far the most common label type, weorganize this in two parts according towhether an enzyme label is used or not.Each of these is further divided as hetero-geneous, which refers to assays that requirethe separation of the Ab-bound Ag fromthe free Ag, and homogeneous, which refersto assays with no separation. Becauseof their importance to ECI, the differentenzyme-substrate-product (E-S-P) systemsand their properties are tabulated inTable 1. The sections on assay formats arefollowed by a discussion on the implemen-tation of ECI in the form of miniaturized

and simple-to-use systems called electro-chemical immunosensors. Next, a briefdiscussion on applications of ECI in clin-ical, pharmaceutical, environmental, andfood analysis is given with the aid of thesummary Table 2. The chapter concludeswith an outlook to the future.

5.2Enzyme Electrochemical Immunoassays

In enzyme ECI, the Ag or a second Abis labeled with an enzyme that catalyzesthe production of an electrochemicallydetectable product and the rate of prod-uct formation is used to quantitate Ag.Therefore, enzyme ECI relies on the Abfor specificity and the enzyme label forsensitivity through chemical amplification.Chemical amplification here refers to pass-ing a substance through a catalytic, cycling,or multiplication mechanism to generatea relatively large amount of product [96,97]. In this way, a trace concentration ofanalyte may result in significantly higherproduct concentration, which for analyti-cal purposes can be more easily measuredthan the analyte itself. Before describingthe different assay formats, we choose todiscuss two key issues of enzyme ECI,electrochemical detection and the choiceof E-S-P system.

5.2.1Electrochemical Detection

The vast majority of electrochemical-detection techniques in immunoassay arebased on voltammetry, the branch ofelectroanalysis that involves applying apotential to an electrochemical cell andmeasuring the current that results fromoxidation or reduction at the electrode [98].Of the many techniques of voltammetry,

5.2 Enzyme Electrochemical Immunoassays 149

Tab.

1C

omm

onE-

S-P

syst

ems

inEC

I

Enzy

me

labe

lSu

bstr

ate

Prod

uct

Det

ecti

onm

etho

dLi

mit

ofde

tect

ion

Ref

eren

ce

Alk

alin

eph

osph

atas

e(A

LP)

4-A

min

ophe

nylp

hosp

hate

(PA

PP)

4-A

min

ophe

nol

(PA

P)FI

A-E

C,G

C,0

.30

Vvs

.Ag/

AgC

lE:

25fM

(500

zmol

e),

T=

60m

inE′

:4am

ole

11

’’’’

FIA

-EC

,GC

,0.3

25V

vs.A

g/A

gCl

P:9

nM12

’’’’

IDA

P:1

µM2

13

’’’’

RD

E,G

C,0

.3V

vs.A

g/A

gCl

E′:3

47am

ole2

,T

=20

sP:

5nM

214

1-N

apht

hylp

hosp

hate

1-N

apht

hol

FIA

-EC

,GC

,0.5

5V

vs.A

g/A

gCl

E:5.

1fM

,T=

20m

inP:

300

pM15

’’’’

DPV

,SPE

imm

unos

enso

r,0.

275

Vvs

.Ag

pseu

dore

f.E:

51fM

(770

zmol

e),

T=

20m

inP:

100

nM

16

Glu

cose

-6-p

hosp

hate

Glu

cose

Glu

cose

sens

or,0

.60

Vvs

.A

g/A

gCl

E:∼0

.02

IU,

T=

30m

in17

4-H

ydro

xyna

phth

yl-1

-ph

osph

ate

(HN

P)D

ihyd

roxy

naph

thal

ene

FIA

-EC

,GC

,0.3

0V

vs.A

g/A

gCl

E:30

0fM

(6am

ole)

,T

=20

min

18

3-In

doxy

lpho

spha

teIn

digo

blue

AC

V,C

past

eim

mun

osen

sor,

−0.4

Vvs

.Ag/

AgC

lP:

50nM

19

Phen

ylph

osph

ate

Phen

olLC

-EC

,Cpa

ste

imm

unos

enso

r,0.

85V

vs.A

g/A

gCl

P:10

nM20

5-B

rom

o-4-

chlo

ro-3

-indo

lyl

phos

phat

ees

ter

H2O

2H

RP-

mod

ified

Cim

mun

osen

sor,

0.0

Vvs

.Ag/

AgC

l55

0fM

,T=

5m

in21

6-(N

-ferr

ocen

oyla

min

o)2,

4-di

met

hylp

heny

lpho

spha

te6-

(N-fe

rroc

enoy

l-am

ino)

-2,4

-di

met

hylp

heno

l

Nafi

on-m

odifi

edG

C,0

.6V

vs.

Ag/

AgC

lE:

6.3

pM,

T=

15m

in22

(con

tinue

dov

erle

af)


Tab.

1(c

ontin

ued)

Enzy

me

labe

lSu

bstr

ate

Prod

uct

Det

ecti

onm

etho

dLi

mit

ofde

tect

ion

Ref

eren

ce

ALP

+bi

enzy

mat

icbi

osen

sor

(tyr

osin

ase

&G

DH

)

Phen

ylph

osph

ate

Phen

ol1

Bie

nzym

eca

taly

ticbi

osen

sor

with

aC

lark

O2

elec

trod

e,0.

60m

Vvs

.Ag/

AgC

l

E:3.

2fM

(320

zmol

e),

T=

57m

in23

ALP

+bi

enzy

mat

icsy

stem

(NA

DH

oxid

ase

&al

coho

ldeh

ydro

gena

se(A

DH

))

NA

DP+

NA

D+1

Mem

bran

e-co

vere

dPt

elec

trod

e,0.

60V

vs.A

g/A

gCl

E:50

amol

e,T

>20

min

24

HR

P3,

3’,5

,5’-T

MB

TMB

(ox)

FIA

-EC

,GC

,0.1

0V

vs.A

g/A

gCl

E:85

fM,T

=15

min

25

Hyd

roqu

inon

eB

enzo

quin

one

FIA

-EC

,GC

,−0.

25V

vs.A

g/A

gCl

E:1.

7pM

,T

=15

min

25

Red

oxO

s+2-b

ased

poly

mer

Os+

3R

edox

poly

mer

-coa

ted

imm

unos

enso

rE:

0.01

µgm

L−1

26

Glu

cose

-6-p

hosp

hate

dehy

drog

enas

eN

AD

++

gluc

ose-

6-ph

osph

ate

NA

DH

LC-E

C,C

past

eel

ectr

ode,

0.8

Vvs

.Ag/

AgC

lP:

0.1

µM27

Gal

acto

sida

se4-

amin

ophe

nyl-b

eta-

D-

gala

ctop

yran

osid

e(P

APG

)PA

PFI

A-E

C,G

C,0

.20

Vvs

.Ag/

AgC

lE:

100

fM,T

=2

hP:

50nM

28

Not

e:H

RP:

Hor

sera

dish

pero

xida

se;T

MB

:Tet

ram

ethy

lben

zidi

ne;A

CV

:Alte

rnat

ing-

curr

entv

olta

mm

etry

;am

ole:

1e-1

8m

oles

;DPV

:Diff

eren

tialp

ulse

volta

mm

etry

;E:D

etec

tion

limit

for

unco

njug

ated

enzy

me;

E′:D

etec

tion

limit

for

conj

ugat

ed(A

b/A

g)en

zym

e;FI

A-E

C:F

low

-inje

ctio

nan

alys

isw

ithel

ectr

oche

mic

alde

tect

ion;

GC

:Gla

ssy

carb

on;L

C-E

C:L

iqui

dch

rom

atog

raph

yw

ithel

ectr

oche

mic

alde

tect

ion;

P:D

etec

tion

limit

for

final

enzy

mat

icpr

oduc

t;P’

:Det

ectio

nlim

itfo

rst

anda

rds

ofP;

RD

E:R

otat

ing

disk

elec

trod

e;SP

E:Sc

reen

-pri

nted

elec

trod

e;SQ

V:S

quar

e-w

ave

volta

mm

etry

;zm

ole:

1e-2

1m

oles

.1

Follo

w-u

pam

plifi

catio

nst

ep/s

invo

lvin

gre

cycl

ing

enzy

mes

.2

Uno

ptim

ized

dete

ctio

nlim

it.


Tab.

2A

pplic

atio

nsof

Elec

troc

hem

ical

Imm

unoa

ssay

Ana

lyte

Ass

ayty

peLa

bel&

subs

trat

eD

etec

tion

sche

me

Lim

itof

dete

ctio

nR

efer

ence

and/

orra

nge

Bio

med

ical

and

phar

mac

euti

cala

naly

sis

ALP

He,

EA

LP/P

PFI

A-E

C40

–50

0U

L−1

29

Am

phet

amin

eH

o,N

,CC

obal

toce

nium

ion

Nafi

on-m

odifi

edel

ectr

ode

2.5

nM–

1.0

µM30

Apo

lipop

rote

inE

He,

E,S,

IA

LP/P

APP

Mem

bran

eim

mun

osen

sor

50–

1000

µgL−

121

Bio

tinH

o,E,

C,I

ALP

/PA

PPM

icro

poro

usA

u-im

mun

osen

sor

1nM

31

’’H

e,E,

C,I

‘‘wir

ed’’

HR

P/H

2O

2‘‘W

ired

’’-en

zym

eim

mun

osen

sor

10nM

–1

mM

26

Car

cino

embr

yoni

cA

gH

e,E,

SH

RP/

H2O

2/f

erro

ceny

lmet

hano

l(FM

A)

SEC

M∼1

04m

olec

ules

32

Coc

aine

He,

E,N

ALP

/PP

Bie

nzym

atic

recy

clin

gbi

osen

sor

with

FIA

-im

mun

ocol

umn

380

pM(3

8fm

ole)

33

’’H

o,E,

C,I

HR

P/H

2O

2G

raph

ite-im

mun

osen

sor

0.1

–10

µM34

Cor

tisol

Ho,

E,C

,IH

RP/

H2O

2M

embr

ane-

O2

elec

trod

e-im

mun

osen

sor

0.1

µM35

Dig

oxin

Ho,

E,C

ALP

/PA

PPM

icro

poro

usA

u-im

mun

osen

sor

100

pM31

’’H

e,E,

CA

LP/P

PFI

A-E

C50

pgm

L−1

36

Dig

oxin

He,

E,S

ALP

/PA

PPC

apill

ary-

FIA

-EC

10–

1000

pgL−

137

Dig

itoxi

nH

e,E,

CA

LP/P

APP

Mic

ropo

rous

Au-

imm

unos

enso

r10

nM31

Estr

iol

Ho,

N,C

Elec

troa

ctiv

e–

NO

2gr

oups

Diff

eren

tialp

ulse

pola

rogr

aphy

0.02

–2.

2µg

mL−

138

Fact

orV

III

He,

E,C

,IH

RP/

hydr

oqui

none

/H

2O

2

GC

-imm

unos

enso

r1

µgm

L−1

39

’’H

e,E,

CA

LP/g

luco

se-6

-pho

spha

teG

luco

seel

ectr

ode

1.56

–10

0U

mL−

117

(con

tinue

dov

erle

af)


Tab.

2(c

ontin

ued)

Ana

lyte

Ass

ayty

peLa

bel&

subs

trat

eD

etec

tion

sche

me

Lim

itof

dete

ctio

nR

efer

ence

and/

orra

nge

Fatt

yac

id–

bind

ing

prot

ein

He,

E,S,

IA

LP/P

APP

Gra

phite

-imm

unos

enso

r5

–30

µgm

L−1

40

Hep

atiti

s-B

surf

ace

Ag

HeE

,SH

RP/

o-ph

enyl

ened

iam

ine

Line

ar-s

wee

ppo

laro

grap

hy0.

1–

5ng

mL−

141

’’H

e,E,

S,I

GO

x/gl

ucos

eM

embr

ane-

O2

elec

trod

e-im

mun

osen

sor

0.1

–10

0µg

mL−

142

Gra

nulo

cyte

mac

roph

age

colo

ny-s

timul

atin

gfa

ctor

(GM

-CSF

)

He,

E,C

,IA

LP/P

APP

SPE

imm

unos

enso

r0.

10µg

mL−

1,

1–

30µg

mL−

143

Hum

anal

pha-

feto

prot

ein

He,

E,C

,IC

atal

ase/

H2O

2M

embr

ane-

O2

elec

trod

e-im

mun

osen

sor

0.05

–50

ngm

L−1

44

’’H

e,E,

SA

LP/P

APP

FIA

-EC

0.16

3µg

L−1,

0.31

6–

100

µgL−

145

Hum

anch

orio

nic

gona

dotr

opin

(HC

G)

Ho,

E,S

ALP

/PA

PPM

icro

poro

usA

u-im

mun

osen

sor

2.5

units

L−1

46

’’H

e,E,

C,I

Cat

alas

e/H

2O

2M

embr

ane-

O2

elec

trod

e-im

mun

osen

sor

0.02

–10

0IU

mL−

147

Hum

anIg

GH

e,E,

C,S

Ure

ase

Impe

danc

em

easu

red

onpo

lym

er-c

oate

del

ectr

ode

0.00

01–

100

µgm

L−1

48

Hum

an-I

gMH

e,E,

S,I

HR

P/I 2

Dis

pers

edca

rbon

mat

eria

l-im

mun

osen

sor

1nM

49

Hum

anlu

tein

izin

gho

rmon

eH

o,E,

S,I

HR

P/ H2O

2

‘‘Enz

yme-

chan

nelin

g’’

imm

unos

enso

r1

ngm

L−1

50

Hum

anse

rum

albu

min

(HSA

)H

e,M

,C,I

Bi3+

SPE

imm

unos

enso

r0.

2µg

L−1

(90

fmol

e),

0.3

–30

µgm

L−1

51

’’H

e,M

,CC

u2+(c

atal

yst)

/o-p

heny

lene

diam

ine

Line

ar-s

wee

ppo

laro

grap

hy10

–50

0ng

mL−

152

’’H

e,M

,CN

i2+D

iffer

entia

lpul

sepo

laro

grap

hy33

3nM

53


Hyd

roxy

coum

arin

(um

belli

fero

ne)

He,

E,C

,IH

RP/

hydr

oqui

none

/H

2O

2

GC

-Im

mun

osen

sor

24µM

54

’’H

e,E,

C,I

HR

P,H

2O

2,F

e(C

N) 6

4−R

enew

able

imm

unos

enso

r20

–10

00µM

55

Isoe

nzym

ela

ctat

ede

hydr

ogen

ase-

1(L

DH

-1)

Ho,

E,I

LD-1

/Lac

tate

/N

AD

+M

embr

ane

imm

unos

enso

r0.

005

–0.

12U

mL−

156

Lido

cain

eH

o,E,

C,I

GO

x/gl

ucos

e/m

edia

tor

Au-

base

dim

mun

osen

sor

5–

50µM

57

’’H

o,E,

C,

Glu

cose

-6-p

hosp

hate

dehy

drog

enas

e/N

AD

+A

ssay

kit,

Ptel

ectr

ode

1–

10ng

mL−

158

Mou

seIg

GH

e,E,

SA

LP/P

APP

IDA

elec

trod

e10

–10

00ng

mL−

113

’’H

e,E,

SA

LP/P

APP

FIA

-EC

0.81

pgm

L−1,6

orde

rsm

agni

tude

59

’’H

e,E,

SA

LP/P

APP

RD

E50

–50

00ng

mL−

160

’’H

e,E,

SA

LP/P

APP

SEC

M5

–25

00ng

mL−

161

Oro

som

ucoi

d(O

MD

,α

1-a

cid

glyc

opro

tein

)

He,

E,C

ALP

/PP

LC-E

C(c

arbo

npa

ste

elec

trod

e)1.

0–

10.0

ngm

L−1

20

Phen

obar

bita

lH

e,E,

CA

LP/P

APP

Cap

illar

y-FI

A-E

C30

–30

00µg

L−1

62

Phen

ytoi

nH

o,E,

CG

luco

se-6

-pho

spha

tede

hydr

ogen

ase/

NA

D+ /

2,6-

dich

loro

indo

phen

ol(D

CIP

)

FIA

-EC

2.5

–30

µgm

L−1

63

’’H

o,N

,CC

obal

toce

nium

Nafi

on-lo

aded

Cpa

ste

elec

trod

e50

–80

0nM

64

’’H

e,E,

CA

LP/(

N-fe

rroc

enoy

l)-6

-am

ino-

2,4-

dim

ethy

lpho

spha

teN

afion

film

mod

ified

elec

trod

e10

nM22

PSA

He,

E,S

ALP

/PA

PPFI

A-E

C0.

008

µgm

L−1,

0.02

–1

µgm

L−1

65

’’H

o,E,

SA

LP/P

APP

Mic

ropo

rous

Au-

imm

unos

enso

r0.

8–

20µg

mL−

146

(con

tinue

dov

erle

af)


Tab.

2(c

ontin

ued)

Ana

lyte

Ass

ayty

peLa

bel&

subs

trat

eD

etec

tion

sche

me

Lim

itof

dete

ctio

nR

efer

ence

and/

orra

nge

Rab

bit

IgG

He,

E,C

,IA

LP/P

PIm

mun

osen

sor

with

cond

uctiv

ebi

ocom

posi

tes

0–

14µg

mL−

166

’’H

e,E,

S,A

LP/P

PC

apill

ary-

FIA

-EC

5600

mol

ecul

es10

’’H

o,E,

S,I

Cho

line

oxid

ase/

HR

P/C

holin

e‘‘W

ired

’’-en

zym

eim

mun

osen

sor

2–

1000

ngm

L−1

50

’’H

e,E,

SA

LP/n

apht

hylp

hosp

hate

Sol-g

elba

sed

imm

unos

enso

r50

–50

00ng

mL−

167

Red

bloo

dce

lls(H

uman

)H

e,E,

S,I

HR

P/H

2O

2/F

e(C

N) 6

4−FI

A-im

mun

osen

sor

(1–

30)1

08ce

llsm

L−1

68

Theo

phyl

line

Ho,

E,C

Glu

cose

-6-p

hosp

hate

dehy

drog

enas

e/N

AD

+ /2,

6-D

CIP

FIA

-EC

2.5

–40

mg

mL−

169

’’H

e,E,

C,I

ALP

/PA

PPFI

A-E

C25

ngm

L−1

70

Thyr

oid

stim

ulat

ing

horm

one(

TSH

)H

e,E,

SA

LP/P

APP

FIA

-EC

0.02

–10

0m

IUL−

171

’’H

e,E,

S,I

ALP

/NA

DP+

Mem

bran

eim

mun

osen

sor

0.2

–10

0m

IUL−

124

Mul

tian

alyt

ede

tect

ion

inbi

omed

ical

and

phar

mac

euti

cala

naly

sis

Folli

cle

stim

ulat

ing

horm

one

(FSH

)an

dlu

tein

izin

gho

rmon

e(L

H)

He,

E,S,

IH

RP/

H2O

2/f

erro

cene

Chr

onoa

mpe

rom

etry

2.1,

1.8

Ul−

1:F

SH,

LH72

HC

Gan

dhu

man

plac

enta

lla

ctog

enH

e,E,

SH

RP/

H2O

2/f

erro

ceny

lmet

hano

lSE

CM

0.1

IUm

L−1,

3ng

mL−

1:H

CG

,H

PL

73

HC

Gan

dPS

AH

e,E,

SA

LP/P

APP

Mic

ropo

rous

Au-

imm

unos

enso

r0.

4,0.

5µg

mL−

1:

PSA

,HC

G46

HSA

and

hum

anIg

GH

e,M

,CB

i3+an

dIn

3+A

nodi

cst

ripp

ing

volta

mm

etry

(ASV

)1.

8,0.

6µg

mL−

1:

HA

S,Ig

G74


Spat

ially

sepa

rate

dm

ouse

IgG

He,

EA

LP/P

APP

Chr

onoa

mpe

rom

etry

NA

75

Envi

ronm

enta

ltes

ting

Ala

chor

He,

C,L

,Li

poso

me-

ferr

ocya

nide

FIA

-EC

25–

300

µgm

L−1

76

Atr

azin

eH

o,E,

C,I

GO

x/H

RP/

gluc

ose

SPE

imm

unos

enso

r12

ngL−

177

’’H

e,E,

CA

LP/P

APP

Cap

illar

y-FI

A-E

C0.

10–

10.0

µgL−

178

2,4-

Dic

hlor

ophe

noxy

acet

icac

id(2

,4-D

)H

e,E,

C,I

GO

x/gl

ucos

e/H

2O2

SPE

imm

unos

enso

r0.

21pp

m79

’’H

e,E,

C,I

Ace

tylc

holin

este

rase

/ac

etyl

chol

ine

SPE

imm

unos

enso

r10

ngL−

180

’’H

e,E,

CA

LP/p

hosp

hori

ces

ter

of[[(

4-hy

drox

yphe

nyl)

amin

o]ca

rbon

yl]

coba

ltoce

nium

hexa

fluro

phos

phat

e

Nafi

on-c

oate

dSP

E0.

01–

100

µgL−

181

’’H

e,E,

Cβ

gala

ctos

idas

e/p-

AP

–β

gala

ctos

idas

eB

ienz

ymat

icre

cycl

ing

bios

enso

r5

pgm

L−1

(5am

ol),

0.02

–10

0ng

mL−

182

’’H

e,E,

CA

LP/P

PB

ienz

ymat

icre

cycl

ing

bios

enso

r0.

001

–10

00µg

L−1

23

Indo

le-3

-ace

ticac

idH

e,E,

CA

LP/P

APP

Cap

illar

y-FI

A-E

C3

pgµL

−183

PCB

sH

e,E,

CA

LP/1

-nap

hthy

lpho

spha

teSP

Eim

mun

osen

sor

0.01

–10

µgm

L−1

16

’’H

e,E,

CH

RP/

ferr

ocen

eace

ticac

id/

H2O

2FI

A-E

C0.

1–

50µg

mL−

184

Food

chem

istr

y

Esch

eric

hia

coli

He,

E,S,

IH

RP/

H2O

2FI

A-im

mun

osen

sor

50–

200

cells

mL−

185

’’H

e,E,

SA

LP/1

-nap

hthy

lpho

spha

teor

PAPP

Mag

netiz

edC

elec

trod

e47

00ce

llsm

L−1

86

’’H

e,E

Glu

cose

/Kre

bsgl

ycol

ysis

cycl

e/m

edia

tor

FIA

-EC

105−

108

cfu

mL−

187

Lute

iniz

ing

horm

one

(chi

cken

)H

e,E,

SA

LP/P

APP

FIA

-EC

2.5

pgm

L−1

12

(con

tinue

dov

erle

af)


Tab.

2(c

ontin

ued)

Ana

lyte

Ass

ayty

peLa

bel&

subs

trat

eD

etec

tion

sche

me

Lim

itof

dete

ctio

nR

efer

ence

and/

orra

nge

Prog

este

rone

He,

E,C

,IA

LP/P

APP

SPE

imm

unos

enso

r5

ngm

L−1

88

’’H

e,E,

C,I

ALP

/nap

hthy

lPho

spha

teSP

Eim

mun

osen

sor

0–

50ng

mL−

189

Salm

onel

laH

e,E,

S,I

ALP

/PA

PPG

C-im

mun

osen

sor

5000

cells

mL−

190

’’H

e,E,

SA

LP/N

AD

P+/A

DH

/D

iaph

oras

e/fe

rric

yani

dePt

elec

trod

e10

00cf

um

L−1

91

’’H

e,E,

S,I

HR

P/H

2O

2FI

A-I

mm

unos

enso

r50

–20

0ce

llsm

L−1

85

Stap

hylo

cocc

usau

reus

(pro

tein

A-b

eari

ng)

He,

E,S

ALP

/NA

DP/

AD

H/

diap

hora

sePt

elec

trod

e10

pgm

L−1

prot

ein

A92

’’H

o,E,

C,I

HR

P/H

2O

2‘‘E

nzym

e-ch

anne

ling’

’im

mun

osen

sor

1000

cells

mL−

193

Mis

cella

neou

s

Afr

ican

swin

evi

rus

(1)

He,

E,C

HR

P/H

2O

2/h

ydro

quin

one

FIA

-EC

1–

80ng

mL−

194

Ale

utia

ndi

seas

e(M

ink

auto

imm

uno

dise

ase)

Ho,

E,C

,IC

holin

este

rase

/but

yryl

thio

chol

ine

Hg-

film

elec

trod

e1

–7

nM,s

peci

ficA

b95

Not

e:PC

Bs:

Poly

chlo

rina

ted

biph

enyl

s;C

:Com

petit

ive

assa

y;G

DH

:Glu

cose

dehy

drog

enas

e;G

Ox:

Glu

cose

oxid

ase;

He:

Het

erog

eneo

us;H

o:H

omog

eneo

us;I

:Im

mun

osen

sor;

L:Li

poso

mal

;N:N

onen

zym

e;PP

:Phe

nylp

hosp

hate

;S:S

andw

ich

assa

y;PS

A:P

rost

rate

-spe

cific

antig

en;S

ECM

:Sc

anni

ngel

ectr

oche

mic

alm

icro

scop

y;ID

A:I

nter

digi

tate

dar

ray.


amperometry has been the most popularwith ECI. In amperometry, the electrodeis held at a fixed potential and the cur-rent produced from a redox event atthe electrode surface is measured. If de-tected in a convective media, as whenusing an RDE or a flow-through elec-trochemical cell, the method is termedhydrodynamic chronoamperometry. Theseare very sensitive methods, with typicaldetection limits in the nanomolar levels.Because detection can be done in verysmall sample volumes (less than 10 µL),the absolute amount of analyte in sam-ple can be as low as 10−14 moles orless [99].

In amperometry, the optimum electrodepotential for detection is chosen by ob-taining the current response produced bythe analyte as a function of the appliedpotential. This current response for anelectroactive species in solution usuallyhas three distinctive regions of behavior,

as shown by Curve A in Fig. 1. One, a re-gion of potential where the compound isnot electroactive and current is negligible.Two, a region of rising current responsedefined by the Nernst equation, if the sys-tem is reversible. Third, a limiting currentplateau that is independent of potential.The best potential for detection is alongthis limiting current plateau where an-alyte is being electrolyzed at the limitof mass transport to the electrode andsmall changes in the applied potential donot significantly affect the current mea-surement. The current response at thelimiting plateau is directly proportionalto the analyte concentration and is givenby the equation I = nFADCo/d, where I

is the current, n the number of electronsinvolved in the redox reaction, F the Fara-day constant, A the electrode surface area,D the diffusion coefficient, Co the bulkanalyte concentration, and d the thicknessof the diffusion layer.

∆E

Optimum Efor detection

E[V]

Cur

ve A

Cur

ve B

Current

Potential

0.00−0.20 −0.10 0.10 0.20 0.30 0.40 0.50

I = nFADCo/d

Fig. 1 A hypothetical hydrodynamic voltammogram for enzyme substrate (Curve B) andproduct (Curve A). The plateau current for product is given by I = nFADCo/d (seeSect. 5.2.1 for details).


5.2.2The E-S-P System

The choice of the E-S-P system is a criticalfactor in enzyme ECI. It is important thatthe enzymatic reaction be rapid and thatthe substrate S be redox inactive over somepotential range where the product P isactive. This potential window for the hypo-thetical redox curves of P and S in Fig. 1 isdenoted by E. The optimum potential fordetection within E is at its low end wheredeleterious effects on the detection limitcaused by the background noise and inter-ferences from electroactive impurities areminimal. For the same reason, it is also al-ways advantageous to have a P that reachesthe limiting current plateau at a smaller Evalue. Reversibility of P, which had notbeen important in most applications thusfar, is likely to become important withthe growing interest in using recyclingelectrodes such as the IDA electrodes forenhanced sensitivity [13, 100].

Many of the E-S-P systems used inECI and the detection limits obtainedare given in Table 5.1. The most com-monly used enzyme label of these isALP. In the early days of ECI, PP wasused as a substrate [20], but high oxida-tion potentials and electrode poisoningeffects of P (phenol) have limited itsuse. The most popular enzyme-substrate(E-S) pair currently is ALP-PAPP, usedfirst in ECI by Tang and coworkers in1988 [101]. Its enzymatic product PAP hasa low oxidation potential (i.e. 0.18 V vs.Ag/AgCl at a GC electrode in pH 7 buffer)and causes no electrode poisoning. Itsreversibility has been very successfully ex-ploited in electrochemical detection withIDA electrodes [13, 100]. A detection limitof 500 zeptomoles for ALP, the lowestreported for any single E-S pair for ECI

applications, was obtained with the ALP-PAPP pair [11]. The disadvantages of PAPPare its high cost, instability at pH 7, andnonenzymatic hydrolysis during substrateincubation [102]. Where low sensitivity isadequate, the less costly and more sta-ble ALP substrate 1-naphthyl phosphatehas been used [15, 103]. This E-S pair hasshown to be especially good with screen-printed electrode-based immunosensorsdiscussed in Sect. 5.4. Other enzymes usedin ECI include HRP [26, 104], galactosi-dase [28], glucose-6-phosphate dehydroge-nase (G6PDH) [27, 63], catalase [47], andacetylcholinesterase [80]. Of these, HRPhas been used recently with separation-free immunosensors (Sect. 5.4) and galac-tosidase used to catalyze the formation ofPAP from a substrate less prone to nonen-zymatic hydrolysis than PAPP [28].

Although using a single E-S pair is themost common practice in enzyme im-munoassays, there has been some interestlately in using multienzyme systems tofurther enhance assay sensitivity. This isbest illustrated in a recent report where abienzymatic biosensor was used to detectphenol produced in the reaction involvingALP and PP [23]. The detection scheme,as shown in Fig. 2a, involves using tyrosi-nase to oxidize phenol first to catechol andthen to o-quinone. o-Quinone becomesa mediator in the enzymatic dehydro-genation of glucose and is reconvertedto catechol. Quantitation of ALP is doneindirectly by measuring the loss of O2

in the oxidation of phenol. The bienzy-matic recycling resulted in a 350-fold signalamplification, thus enabling the detec-tion of 320 zeptomoles of ALP. Similarly,the E-S pair ALP-NADP+ has been usedwith bienzymatic recycling of the productNAD+ using ADH and either NADH ox-idase [24] or diaphorase [92], as shown inFig. 2(b).


Phenol(P)

Phenylphosphate(S)

Substratereaction

ALP(E)

Phenol

Tyrosinase

o-QuinoneCatechol

Glucose dehydrogenase

Glucose

O2 O2

Tyrosinase

Oxygen electrode (measure O2 consumption)

Biosensor

(a)

(b)

NADP+

AL

P(E

)

P

NAD+

NADH

Dia

phor

ase

Ferricyanide

FerrocyanideH2O2

O2 Ethanol

Acetaldehyde

Alc

ohol

dehy

drog

enas

e

NA

DH

oxi

dase

DetectDetect

Fig. 2 Multienzyme cycling for enhanced assay sensitivity using (a) phenol recyclingbiosensor (b) NAD+-NADH cycling.

5.2.3Heterogeneous Enzyme Immunoassay

Heterogeneous enzyme immunoassay isby far the most common assay formatused in ECI. Most of the heterogeneousassays are based on the enzyme-linkedimmunosorbent assay (ELISA) techniquewhere Ab is immobilized on the wallsof small (<500 µL) microtiter wells orcuvettes [105]. The general procedure ofheterogeneous immunoassay is outlined

in Fig. 3 for determining Ag, which repre-sents the analyte in this and other exam-ples of enzyme ECI. The reaction surface,say the inside walls of a cuvette, is first pre-pared by attaching specific Ab to the wallsby passive adsorption or covalent bond-ing. The cuvettes are then rinsed with anonionic surfactant such as Tween 20 anda ‘‘blocker’’ protein such as bovine serumalbumin [105], casein [106], or gelatin [107]to cover any exposed surface between Abmolecules to which proteins can bind


AbAbAbAb

AbAbAbAb

AbAbAbAb

Ab:Ag∗Ab:Ag

Rinse

AddS

AddS

Rinseo

Rinse AgAg∗

AddAb∗

RinseAg

RinseAb∗

Addo

Preparation ofcuvette

Reagentcuvette

Add AgAg∗

AbAb:AgAbAb:Ag

AddAg

AbAb:Ag:Ab∗AbAb:Ag:Ab∗

Sandwich assay

Competitive assay

Ab:Ag∗Ab:Ag

Ag*Ag

Ag

AgAg*

Ag∗ Ab:Ag∗Ab:Ag

Ab:Ag∗Ab:Ag Detect P

AgAgAg

Ag

Ag

Ab∗Ab ∗

Ab∗

Ab∗ AbAb:Ag:Ab∗AbAb:Ag:Ab∗

AddAb Ab

Ab

Ab

Ab

PS

P

S

oAbAb*Ag*

SAgP

BlockerAntibodyEnzyme-labeled antibodyAntigenEnzyme-labeled antigenSubstrateProduct

Cuvette wall

Ab

PS

P

S

Ag∗

Fig. 3 General protocol for heterogeneous enzyme immunoassay.

nonspecifically (see Sect. 5.2.4). Cuvettesthus prepared are used in either competi-tive or sandwich immunoassays.

In competitive immunoassays, a stan-dard of enzyme-labeled Ag∗ is added to thesample for competitive equilibration withthe Ab. Depending on the analyte and theconfiguration of the reaction vessel used,equilibration can take several minutes tohours, but adequate results may be ob-tainable before this. The unbound Ag andAg∗ are then rinsed from the tubes andthe enzyme substrate S added. At a fixedtime, the sample is analyzed for the elec-troactive product (P), whose concentrationshall be proportional to the Ag∗ in thewell if the enzymatic reaction is carriedout under substrate saturation conditions.Because of the competitive binding, a typ-ical standard plot of the current versus theconcentration of Ag has an inverse, lin-ear relationship. The assays of this type inTable 2 are denoted by E (enzyme) and C(competitive) under the assay type.

The sandwich immunoassay is theother widely used form of heterogeneousenzyme immunoassay. As shown in Fig.3,

this assay derives its name from the factthat the analyte is ‘‘sandwiched’’ betweentwo different antibodies, one of which isenzyme-labeled. It can be applied onlyto molecules that are sufficiently largeand have an appropriate topographicaldistribution of epitopes to accommodatetwo antibodies simultaneously. The Abspecific for Ag is first immobilized onthe cuvette walls, and the sample is thenadded and incubated to allow Ag binding.Following this, the cuvette is washed withbuffer and a second antibody, Ab∗, alsospecific for Ag and labeled with an enzymelabel, is added. After its incubation, theunbound Ab∗ is rinsed from the cuvetteand S added for electrochemical detectionof P. Because more Ab∗ is retained inthe well at high concentrations of Ag, thesignal due to P here is directly proportionalto the concentration of Ag. In Table 2, thesandwich assays are denoted by E (enzyme)and S (sandwich) under assay format.

Electrochemical detection in a heteroge-neous immunoassay, whether competitiveor sandwich, can be done either directlyin the container in which the assay was


done or in an external electrochemical cellby transferring an aliquot to it. Althoughthe former has been done [90, 92], it in-volves the tedious process of transferringthe electrodes from one cuvette to anotherwith rinsing in between. More common,therefore, is the latter, and its most com-mon form is FIA-EC. FIA, a powerfulmicroanalytical method in batch analysis,involves introducing a sample to a mo-bile phase buffer via an injection valve andmeasuring the analyte at a detector down-stream [108]. The rapidly moving buffer inFIA allows nearly instantaneous detection,and because of its favorable hydrodynamicprofiles, amperometric detection with FIAis one of the most sensitive electroanalyti-cal methods [109]. These properties makeFIA ideal for immunoassays where thereare usually several µL-volume samples totest, and its popularity in ECI is evidentby the number of FIA-EC-based assays inTable 2. Where necessary, a liquid chro-matography column can be added to theFIA system to remove substances that mayinterfere in the detection. Such systemsare denoted in Table 2 by LC-EC. In mostof the FIA-based ECI, a thin-layer electro-chemical cell consisting of a glassy carbonworking electrode operated amperometri-cally has served as the detector.

An adaptation of FIA for ECI is to use anin-line immunocolumn, with the columneffluent allowed to pass through an electro-chemical cell for detection. The first of thistype, reported by Wilson and coworkers,used a high-performance chromatogra-phy column where Ab was immobilizedfor a competitive enzyme assay [110]. Theimmunocolumn was renewed in prepara-tion for the next sample by passing acidicbuffer through it to displace bound Ag andAg∗. In a slightly different design, a proteinA immunocolumn was used to bind the Abfor an assay of theophylline [70]. Because

protein A can bind to most immunoglob-ulins with high affinity, this system can,in principle, be adapted for a variety ofassays. However, these types of assaysare generally time-consuming, and often,the harsh conditions used to renew theimmunocolumn decrease its lifetime. Re-cently, Scheller and coworkers reported anoncompetitive immunoassay for cocaineusing an immunocolumn where theseproblems were significantly reduced [33].The immunocolumn contained a perfu-sion chromatography carrier modified bya large number of cocaine molecules. Thesample was introduced into the columnafter mixing it with an excess of Ab∗ con-taining ALP. The excess Ab∗ was boundby cocaine in the column, but not theAg-Ab∗ whose Ab∗ binding sites are occu-pied and unavailable for further binding.The Ag-Ab∗ was collected in the effluent,the enzyme substrate PP added, and theproduct phenol detected at a bienzymaticbiosensor identical to the one describedearlier. By packing the column with alarge excess of cocaine, more than 200assays were performed before the columnrequired regeneration.

5.2.4Ultrasensitive Heterogeneous EnzymeAssays in Small Volumes

Minimizing the reaction cell volume hasa profound effect on the sensitivity andspeed of an assay [99, 111]. This is duemainly to the fact that in smaller volumes,where the dilution of P is less, the Pconcentration can be brought to detectablelevels more rapidly than would be possibleotherwise. Another way that small volumesaffect ECI is by shortening the diffusionaldistances that species in solution haveto travel to react at the surface. Shorterdiffusional distances result in reduced


incubation times and hence, to a reducedtotal assay time. However, minimizing thereaction cell volume has to be done withas little loss of the total surface area aspossible in order to maintain an adequatenumber of binding sites. Microcapillaries,with their high surface-area-to-volumeratio, provide just that type of reaction cell.However, when using capillaries and otherlarge surface area formats, nonspecificadsorption (NSA) of proteins becomesan issue of paramount importance. Asnoted earlier, rinsing with a simplemixture of a detergent and a blockerprotein can reduce NSA to requiredlevels when working with cuvettes orwells. In capillaries, on the other hand,where high assay sensitivity is soughtand the high surface-area-to-volume ratioleads to a correspondingly high NSA,more sophisticated blocking techniquesare called for. Halsall and coworkershave found that NSA in capillaries canbe drastically reduced using a reagentmixture consisting of Tween 20 and ablocker protein as above to minimizeNSA arising from hydrophobic forces, andion-pairing reagents to minimize NSAarising from electrostatic forces [10]. Thecapillaries were also treated to yield anamino group–rich surface for attachingthe Ab using carbodiimide coupling. Thecapillaries thus prepared and incorporatedinto FIA systems have enabled some of themost sensitive assays in immunoassay [10,37]. A detection limit of 5600 moleculesand a linear range of four orders ofmagnitude were reported for a sandwichassay of rabbit IgG in human serum in 70-µl capillaries [10]. Similarly, 260 attomolesof digoxin have been detected using a 20-µL capillary [37]. Capillary immunoassayshave also been used in competitive enzymeimmunoassays to detect indole-3-aceticacid in tomato embryos [83], phenobarbital

in serum [62], and atrazine [78] in riverwater.

Minimizing the reaction cell volume,however, can also be counterproductiveto the assay sensitivity in situations wherethere is significantly more available samplethan can be tested in the reaction volume.For example, say a 1-ml sample, whichin fact contains 100 zeptomoles of Ag, isavailable for a sandwich assay. If a 20-µlcapillary were used, only two zeptomolesof enzyme label will at best remain inthe capillary for detection. If, however, thefull sample could be assayed while keep-ing the final detection volume at 20 µl, thechallenge would be less daunting with 100instead of 2 zeptomoles of enzyme labelsto detect. One way to optimize the assaysensitivity in terms of both the samplingcapacity and the enzyme substrate incuba-tion volume is to use an immunoreceptorthat can be dispersed during sampling,but concentrated in a small volume forsubstrate incubation and detection. Para-magnetic microimmunobeads, sold by anumber of companies [112], hold muchpromise for such ultrasensitive assays.These beads, typically 1 to 3 µm in diam-eter, can be dispersed by gentle shakingduring sampling, and separated and heldagainst the reaction cell with a magnetfor the various rinsing and washing stepsof an assay. More importantly, the beadscan be transferred to and mixed in anelectrochemical cell. This provides muchflexibility in choosing what enzyme sub-strate volume to mix in as well as the type ofelectrochemical detection method to use.Recently, we exploited this concept usingan RDE in an assay for mouse IgG [14,60]. In this work, RDE amperometry wasadapted to place the rotating electrodedirectly on a microdrop of enzyme sub-strate to minimize product dilution. Theexcellent sensitivity of RDE amperometry


and the thorough mixing of the beadsprovided by the electrode enabled the de-tection of 413 attomoles of mouse IgG inless than 10 seconds [60]. This detectiontime is 30- to 120-fold faster than in capil-lary assays. Considering that NSA was themain factor in determining the detectionlimit in the bead-based assay and that nospecial treatment of the beads was done tominimize NSA, it is likely that even lowerdetection limits could be attainable withproper NSA blocking and using longerincubation times if necessary.

5.2.5Homogeneous Enzyme Immunoassay

Homogeneous enzyme immunoassaysuse a competitive assay format that re-lies on a reduction in the rate of en-zyme catalysis of Ag∗ as it binds to theAb. The general equilibrium scheme forhomogeneous immunoassays, which alsoholds true for the nonenzymatic onesdiscussed later, is shown in Fig. 4. The

strength of the homogeneous assay liesin its ability to distinguish between Ag∗and Ab : Ag∗, which simplifies an assay bymaking a separation step unnecessary.

The first homogenous enzyme ECI wasdeveloped by Heineman and coworkersfor phenytoin using glucose-6-phosphatedehydrogenase (G-6-PD) as the enzyme la-bel and NAD+ as the substrate [27]. Theproduct, NADH, was determined by ox-idation using LC-EC with a thin-layercell containing a GC electrode held at0.75 V versus Ag/AgCl. A reversed-phaseC-18 LC precolumn was necessary to re-move other proteins in the sample thatwould otherwise have passivated the elec-trode at such a high potential. The useof an LC column, however, was avoidedin later work by allowing NADH to re-act with 2,6-DCIP mixed into the sample,and detecting the product DCIPH2 at themuch lower electrode potential of 0.20 Vversus Ag/AgCl [63]. This improved assayprotocol was applied to the detection oftheophylline in whole blood [113] as well

Ag

+

+Ab

Ab:Ag

Ab:Ag∗/ x

Signal for label changed bybinding with Ab

Rate of NADH formationdiminished by binding with Ab

Current due to reduction/oxidationof label decreased by binding withAb

NADH

Ag∗/ x

∗ : Enzyme labelExample:

where ∗: glucose-6-phosphatasedehydrogenase

x: Nonenzyme labelExamples:- Metal ions (Pb2+, Zn2+, Co2+, Bi2+);electroactive compounds (cobaltoceniumion, mercuric acetate); -nitro groups

NAD+ *

Fig. 4 General reaction scheme for both enzyme and nonenzyme homogeneousimmunoassay.


as in hemolyzed, lipemic, and ictericserum samples [69]. The ability to performimmunoassays directly on such samplesmarks a significant advantage over assaysbased on optical detection where detectionis severely hampered by spectral interfer-ences due to turbidity and/or the intensecolor of the sample [114]. Homogeneousenzyme assays in Table 2 are denoted byE (enzyme) and Ho (homogeneous) un-der assay type. Another important area ofECI where homogeneous immunoassaysare becoming very important is in electro-chemical immunosensors discussed later.

5.3Nonenzyme ElectrochemicalImmunoassays

A second class of ECI involves labelingan Ag with a nonenzymatic electroactivelabel. We shall denote this kind ofconjugate by Agx to differentiate it fromAg∗, the enzyme-labeled Ag. Nonenzymeimmunoassays developed thus far havebeen based on competitive binding of Agand Agx to Ab. Electrochemical detectionof the label is done while it is conjugatedor after its release by adding a suitablecleaving reagent. Detection in virtuallyall these assays has been done at aHg electrode. Although Hg electrodeshave lost some popularity because of thetoxicity of Hg, under proper handling Hgelectrodes can give excellent sensitivity inelectroanalysis [115]. This is particularlytrue for ASV in which the metal ionis preconcentrated in the Hg electrodeprior to detection. Detection limits in thenanomolar range are obtained routinelywith ASV. Modern advances, both inthe design of Hg electrodes and inthe automation of systems using Hgelectrodes, are making it possible to useHg electrodes safely [116].

5.3.1Nonenzyme Homogeneous Immunoassays

As described in Sect. 5.2.5, homogeneousassays are based on the competitive bind-ing of Ag and the labeled Ag to a limitedamount of Ab. The labels that have beenused in nonenzyme homogeneous assaysare metal ions and electroactive function-alities or molecules. These immunoassaysare based on the reduction of the elec-troactivity of the label, as measured by adrop in the current when Ag∗ binds to Ab.It is believed that this drop is due to thedecrease in the diffusion coefficient (seeSect. 5.2.1) of Ag∗ as it binds to Ab and/orto the sequestration of Ag∗ in the bindingpocket of Ab, making it less available to theelectrode [117]. Therefore, these assays arebest suited for smaller Ags where the frac-tional change in the diffusion coefficientand/or the sequestration is greater uponbinding to Ab.

This approach was first demonstratedusing the electroactive functionality mer-curic acetate in an assay for estriol [118].Since then, assays for estriol usingNO2 groups [38], morphine using fer-rocene [119], and HSA using metal ionsPb2+ [120], Co2+ [121], and Zn2+ [122]have been reported. However, homoge-neous assays of this kind have not shownany significant advantage over other types,and their use is rare. Assays of this typein Table 2 are denoted by N (nonenzyme)and He (heterogeneous) under assay type.

5.3.2Heterogeneous NonenzymaticImmunoassays

The assays in this class have for themost part used electroactive metal ionsfor labels. The first such immunoassaywas reported by Heineman and Hal-sall for the analyte HSA using the label

5.3 Nonenzyme Electrochemical Immunoassays 165

In3+ [53]. The Agx was made by couplingIn3+ to Ag molecules using the bifunc-tional chelate diethylenetriamine penta-acetic acid (DTPA). Because In3+ is notfound in human tissue, interference fromthe clinical sample itself is avoided. Theassay protocol used is outlined in Fig. 5.Following competitive equilibration of Agand Agx with Ab (adsorbed to insolubilizedprotein A), the Ab-Agx complex was sep-arated by centrifugation. The In3+ labelof Agx was then released by acidificationand determined by differential pulse an-odic stripping voltammetry (DP-ASV). The

peak height of the stripping wave for In3+was shown to be inversely proportional tothe HSA in the sample, as expected froma competitive assay.

An exciting possibility of heterogeneousimmunoassay with ASV detection is itsextension to multianalyte detection usingmultiple metal labels. This concept wassuccessfully applied by Hayes and cowork-ers in the simultaneous detection of HSAand human IgG in perhaps what is the firstmultianalyte ECI [74]. Because ASV is anextremely sensitive method that can detectup to about six metals simultaneously, it

Centrifuge

CentrifugeSeparatesupernate

DetermineIn3+

by DPASV−600−400 −800

E vs. Ag/AgCl[mV]

ip

HSA sample ( )

Add Ag-In3+( ),add Ab, mix,incubate

Separate supernate,rinse, resuspend pelletin acid to release In3+

Fig. 5 Heterogeneous nonenzyme assay protocol for determining HSAwith releasable In3+ label by ASV.


is possible to detect as many analytes, inprinciple [116].

Two other types of nonenzymatic het-erogeneous immunoassay have been re-ported. One uses liposomes containingthe electroactive marker ferrocyanide asa label [76]. Following the competitivebinding of Agx in an immunocolumn,ferrocyanide is released by flushing withdetergent, and detected at a downstreamelectrode. An alternative to the direct mea-surement of a metal ion was reportedrecently in a competitive immunoassay forHSA where Cu2+ catalyzed the conversionof o-phenylenediamine to the electroac-tive product 2,3-diaminophenaline, whichwas detected with linear-scan polarogra-phy [52]. Nonenzyme heterogeneous as-says in Table 2 are denoted by N (nonen-zyme) and He (heterogeneous) underassay type.

5.4Electrochemical Immunosensors

The immunoassay techniques discussedso far have all been best suited for a labenvironment because of both the type ofequipment required and the experimentalconditions used. In many instances, in-cluding virtually all clinical applications,this is not a problem since the tests needto be done in a lab environment any-way for a host of other reasons. However,methodology for less complicated and ‘‘on-the-spot’’ testing is finding an increasinglyimportant niche in immunoassay [123].This is particularly true for field-testing inenvironmental applications, home-baseddiagnostic kits, and testing in emergencysituations, as in an ambulance or an oper-ation theater. To adapt ECI to a system for‘‘on-the-spot’’ analysis, the system must

be small, easy to handle and operate, capa-ble of working without additional reagents,and renewable where use in a disposablemanner is prohibitively expensive. Thedevelopment of such systems is the fo-cus of the rapidly developing area of ECIloosely termed electrochemical immunosen-sors, which is discussed in this section.

To the best of our knowledge, the idealreagent-free electrochemical immunosen-sor for ‘‘on-the-spot’’ analysis has yet to bedeveloped. Systems reported in the liter-ature as electrochemical immunosensors,some of which are listed and denoted by Iunder assay type in Table 2, are based pri-marily on one factor: the immobilizationof the sensing Ab or Ag directly on the elec-trode [124, 125]. The schematic of an elec-trochemical immunosensor, thus defined,is shown in Fig. 6. It should be noted thatelectrochemical immunosensors are oftencategorized as electrochemical biosensors,which covers any type of electrochemi-cal sensor consisting of a biorecognitionelement placed directly over or in closeproximity to the electrode [126].

The earliest works in electrochemicalimmunosensors were directed at usingsimple electrodes with passively adsorbedAb to detect Ag based on any changesin the electrode potential due to Ag bind-ing [127]. Their promise for simple, label-free, and direct immunosensors, however,was short-lived because of poor sensitiv-ity due to nonspecific binding. Since then,potentiometric immunosensors using ion-and gas-sensing electrodes for enzymeassays [128, 129], pH electrodes, and field-effect transistors (FETs) [130] have beenreported. These immunosensors too havehad limited success because of poor se-lectivity due mainly to NSA [124, 131]. Incontrast, a far greater degree of successhas been achieved with amperometric im-munosensors.

5.4 Electrochemical Immunosensors 167

Electricalsignal

ElectrodeLayer withAb/Ag

detectionelement

Signalprocessor

Sample

Analyte

Interferents

Fig. 6 The basic design of an electrochemical immunosensor.

The first amperometric immunosensordeveloped by Aizawa in 1979 for detectingHCG [47] consisted of an oxygen probecovered with a membrane coated withAb. Detection was based on a competi-tive format using catalase to decomposeH2O2 and the oxygen probe to measurethe product O2 as it diffused through themembrane. Similar membrane-based elec-trochemical immunosensors have beenused to detect human α-fetoprotein [44],thyrotropin [24], and hepatitis-B surfaceantigen (HBSA) [42]. However, replace-ment of the membranes with each use is adisadvantage. The alternative of renewingthe sensor surface by cleaving the Ab–Agbond in acidic media has been applied toglassy carbon-based immunosensors [54,68], but this too is tedious and renewabil-ity is never total. Such problems greatlyhampered the growth of electrochemi-cal immunosensors until the early 1990swhen technologies to make disposableimmunosensors were developed. For im-munosensors to be disposable, they must

be inexpensive, made reproducibly in largenumbers, and in this regard, the develop-ment of screen printed electrodes (SPEs)has been very important for electrochemi-cal immunosensors.

SPEs, fabricated with thick-film technol-ogy [132] using a graphite powder-basedink to print electrodes on a polystyrenesurface, are adapted for immunosensorschiefly by passive adsorption of antibod-ies to the electrode surface. SPE-basedimmunosensors have been used in anumber of applications including the de-tection of the herbicide 2,4-D [79, 81, 133],PCBs [134], the biochemical messengerGM-CSF [43], and the hormone proges-terone [87, 88, 135]. All these applicationsuse enzyme labels in a heterogeneousassay format. An exception to using anenzyme is an SPE immunosensor forHSA reported recently by Wang andcoworkers where a competitive and metalion label format similar to the one de-scribed in Sect. 5.3.2 was used [51]. Themercury electrode necessary for stripping


electroanalysis was formed by in situdeposition of Hg2+ on a SPE carbon elec-trode. SPE immunosensors have also beenused in semi-homogeneous enzyme as-says, which are discussed shortly.

Two other recent technologies appliedto electrochemical immunosensors, onebased on thin-film technology (lithogra-phy) [136] and the other on the conceptof ‘‘wired’’ redox enzyme electrodes [137,138], are also likely to become important inthe future. There has been much interestin using photolithographically patternedIDA electrodes in immunoassay follow-ing the first report by Heineman andcoworkers [13]. Recently, a dual-analyteimmunosensor made by thin-film tech-nology was used to detect FSH andLH using a sandwich assay format [72].The Ab immobilization was done on athiol-modified gold electrode treated withavidin-photobiotin. Photolithography wasapplied to activate photobiotin in definedpatterns for binding Neutravidin conju-gated antibodies.

The concept of ‘‘wired’’ enzymes,borrowed from electrochemical biosen-sors [126], was recently applied to make animmunosensor for HRP [26]. Here, rab-bit Ab for HRP was co-immobilized withthe redox polymer [Os(bipyridyl)2poly-4-vinylpyridine)10Cl]Cl onto a GC electrode.Upon capture at the immunosensor sur-face, the redox centers of the HRPmolecules become electrically wired to theelectrode surface via the redox polymer.Therefore, when H2O2 is added to theimmunosensor following the sample incu-bation and wash, the HRP that is oxidizedduring catalysis can be electrochemicallyreduced (and detected) to its original formby the wired electrode. The detection ofHRP in this way was shown to have theclassical features of a fast redox couplestrongly bound to an electrode surface.

The immunosensors described thus farall use a heterogeneous assay formatin contrast to the ideal ‘‘on-the-spot,’’reagent-free, disposable or reusable, andone-step analysis envisioned with im-munosensors. To the best of our knowl-edge, no present system is capable ofall these, but the ‘‘separation-free’’ im-munosensors that have appeared in theliterature recently come a step closer to-ward this goal. These immunosensors aresimpler to use because they do not re-quire the separation of the enzyme-labeledAb/Ag in the bulk from the surface-bound. One approach taken in doing thisinvolves immobilizing the enzyme labelalong with the Ab in making the im-munosensor [34, 35]. The electrochemicaldetection was based on the enzyme inhi-bition that follows Ag binding. For moresensitive assays, however, it is necessaryto work with enzyme-labeled Ag or Abin solution, and a method has to be de-vised to differentiate the enzyme labels inthe bulk from the surface-bound. Meyer-hoff and coworkers described an elegantmethod to do this, in which a microporousnylon membrane coated on one side withgold served to immobilize Ab as well asact as an electrode [46]. After incubationfor the competitive binding of the enzyme-labeled Ag (Ag∗), the enzyme substratewas added from the backside of the porousmembrane. Because of proximity and floweffects, the enzymatic product formed atthe immunosensor surface was preferen-tially detected, thus differentiating it fromthe product formed in the bulk and al-lowing a separation-free assay. Anotherapproach to discriminating the productformed in the bulk has been proposedwhere a reagent is added to scavenge theproduct formed in the bulk but not thatformed near the electrode surface, which

5.4 Electrochemical Immunosensors 169

Product 1

Substrate

Enzyme 2

Product 2

e−

Enzyme 1

Ab

Ag∗

Electrode

Fig. 7 The principle of enzyme-channeling immunoassay using an electrode-boundenzyme to make the substrate for the enzyme label on the Ag∗.

is oxidized by the electrode sooner thancould be scavenged [77].

Enzyme-channeling immunoassay in-troduced by Litman and coworkers [139]gives another promising way to makeseparation-free immunosensors. The basicprinciple of enzyme channeling is illus-trated in Fig. 7. As shown, the substratefor the enzyme label of Ag∗ is produced insitu by a second enzyme co-immobilizedwith Ab on the immunosensor surface.Because of the distances involved, the en-zymatic product of the second enzymeis ‘‘channeled’’ to the Ab-bound Ag∗,and hardly any to that in the bulk. Inthe area of ECI, enzyme channeling wasfirst exploited by Brown and coworkersin 1991 to develop a potentiometric im-munosensor [140]. Since then, Rishponand coworkers have developed ampero-metric immunosensors based on enzymechanneling for Staphylococcus aureus [93],rabbit IgG, and human LH [50]. For theseapplications, GOx was used as the secondenzyme and HRP as the Ag∗ enzyme label.Hydrogen peroxide, the by-product in theoxidation of glucose at the second enzyme,becomes a co-substrate with iodide in theenzyme catalysis at Ag∗, and its product

I2 is detected at the electrode. Recently,Heller and coworkers used the sameprinciple with an osmium-containing re-dox hydrogel where choline oxidase wasused for the second enzyme to catalyzethe oxidation of choline to produce betainealdehyde and H2O2 [141]. The H2O2 pro-duced is channeled to the HRP labels onthe Ab∗ where it is reduced. Because of thehydrogel, the HRP labels are ‘‘wired’’ tothe electrode and detected as in the earlierexample.

The development of electrochemical im-munosensors has increased rapidly overthe last five years, and as the technolo-gies that made this possible keep evolving,we can expect even more growth in thefuture. With respect to other types, elec-trochemical immunosensors can easilybe miniaturized, and being insensitiveto light, they can be used in opaqueor turbid media. Also, since the signaltransducer is electrical, it needs no signalconversions as with optical transducers.However, for their use as one-step and‘‘on-the-spot’’ immunosensors, it will benecessary to find a way to incorporateall the reagents in the immunosensorsin some way.


5.5Applications

In Table 2, we give a comprehensive list ofECI applications. The table is arranged inthree parts: biomedical and pharmaceuti-cal analysis, environmental analysis, andfood chemistry. Each entry in the table isfurther categorized according to the as-say type, label and substrate, detectionscheme, and the limit of detection and/orrange obtained. Assay type tells whether itis homogeneous or heterogeneous, if anenzyme label is involved, and whether theassay is done using an immunosensor ornot. The detection limits (and units) aregiven as reported in the original publica-tions.

5.6Conclusion

Electrochemical immunoassay has ad-vanced greatly over the last two decades.Some of the most sensitive assays in im-munoassay have been obtained with ECI.Being insensitive to the color and to alarge degree the turbidity of the sample,ECI has a significant advantage over im-munoassay with optical detection for directapplications to samples. The other asset ofECI likely to play a key role in the fu-ture is its easy adaptability to miniaturizedimmunoassays. Such assays are obviouslyimportant where minimizing sample vol-ume is crucial as in the testing of neonatesand critically injured patients. Further-more, there has been much interest latelyin miniaturized assays for multianalytedetection using a platform of denselypatterned arrays of antibodies [8, 9]. Sev-eral technologies for patterning biochem-ically active arrays have been developedin the last ten years including ink-jet technology [142], screen printing [143],

photoresist/lift-off methodology [144], mi-crocontact printing [145], and local pho-tochemical activation of preimmobilizedcoupling agents [146]. There have alreadybeen some successes in multianalyteECI [72, 73, 75], and as the earlier-mentioned technologies advance and be-come more accessible, we can expectresearch in this area to grow significantly.SECM, the tool of choice for probingthe microscale electrochemistry of sur-faces [147], might become very useful inmultianalyte detection because it can ad-dress different regions of a surface by scan-ning it in two dimensions, thus making theuse of multiple electrodes unnecessary.The feasibility of SECM for multianalytedetection in enzyme immunoassay has al-ready been successfully demonstrated [32,61, 73, 148, 149]. Micrototal analysis sys-tems (µTAS) [150], where the goal is a‘‘lab-on-a-chip’’ that integrates microfab-ricated parts including control electronicsfor total analysis in a miniaturized sensor,is another technology likely to play a keyrole in the future of ECI. If successful,µTAS can revolutionize ECI with minia-turized immunosensors used in a host ofapplications ranging from field-deployedsensors that continuously monitor airor water quality to implantable sensorsthat monitor various physiological con-ditions of the human body. Because ofits amenability to miniaturization, ECIarguably has the best chance of immunoas-say types to exploit these technologies. Asnoted in the introduction, ECI lay dormantfor nearly 30 years before developmentsin technology and fundamental sciencecould catapult it to the popular bioanalyt-ical method it has become. Perhaps theearlier-mentioned technologies of todaymay usher a second growth of ECI witha whole set of new possibilities.

5.6 Conclusion 171

References

1. I. Roit, Essential immunology, 4th ed., Black-well Scientific, Boston, 1980.

2. C. P. Price, D. J. Newman (Eds.), Principlesand Practice of Immunoassay, 2nd ed.,Macmillan, London, 1997.

3. D. N. Posnett, J. P. Tan, Methods Enzymol.1989, 178, 739–746.

4. G. Kohler, C. Milstein, Nature 1975, 256,495–497.

5. G. Winter, C. Milstein, Nature 1991, 349,293–299.

6. A. J. T. George, in Principles and Prac-tice of Immunoassay (Eds.: C. P. Price,D. J. Newman), 2nd ed., Macmillan, Lon-don, 1997, pp. 65–98.

7. B. Breyer, R. Radcliff, Nature 1951, 167,79–81.

8. L. J. Kricka, Clin. Chem. 1998, 44/9,2008–2014.

9. P. R. Ekins, Clin. Chem. 1998, 44/9,2015–2030.

10. S. H. Jenkins, H. B. Halsall, W. R. Heine-man, J. Clin. Immunoassay 1990, 13,99–104.

11. R. Q. Thompson, M. Porter, C. Stuver et al.,Anal. Chim. Acta 1993, 271, 223–229.

12. Y. Qu, L. R. Berghman, F. Vandesande,Anal. Biochem. 1998, 259, 167–175.

13. O. Niwa, Y. Xu, H. B. Halsall et al., Anal.Chem. 1993, 65, 1559–1563.

14. C. A. Wijayawardhana, S. Purushothama,M. A. Cousino et al., J. Electroanal. Chem.1999, 468, 2–8.

15. M. Cardosi, S. Birch, J. Talbot et al., Electro-analysis 1991, 3, 169–176.

16. M. Del Carlo, I. Lionti, M. Taccini et al.,Anal. Chim. Acta 1997, 342, 189–197.

17. R. Renneberg, W. Schossler, F. Scheller,Anal. Lett. 1983, 16, 1279–1289.

18. M. Masson, T. Haruyama, E. Kobatake et al.,Anal. Chim. Acta 1999, 402, 29–35.

19. C. Fernandez-Sanches, Anal. Chim. Acta1999, 402, 119–127.

20. M. J. Doyle, H. B. Halsall, W. R. Heineman,Anal. Chem. 1984, 56, 2355–2360.

21. W. O. Ho, D. Athey, C. J. McNeil, Biosens.Bioelectron. 1995, 10, 683–691.

22. A. La Gal Lal Salle, B. Limoges, C. Degrand,Anal. Chem. 1995, 67, 1245–1253.

23. C. G. Bauer, A. V. Eremenko, E. Ehren-treich-Forster et al., Anal. Chem. 1996, 68,2453–2458.

24. D. Athey, C. J. McNeil, J. Immunol. Methods1994, 176, 153–162.

25. G. Volpe, D. Compagnone, R. Draisci et al.,Analyst 1998, 123, 1303–1307.

26. B. Lu, E. I. Iwuoha, M. R. Smyth et al., Anal.Chim. Acta 1997, 345, 59–66.

27. H. M. Eggers, H. B. Halsall, W. R. Heine-man, Clin. Chem. 1982, 28/29, 1848–1851.

28. M. Masson, Z. Liu, T. Haruyama et al., Anal.Chim. Acta 1995, 304, 353–359.

29. S. D. Jackson, H. B. Halsall, A. J. Pesceet al., Fresenius J. Anal. Chem. 1993, 346,859–862.

30. B. Limoges, C. Degrand, P. Brossier et al.,Anal. Chem. 1993, 65, 1054–1060.

31. M. W. Ducey Jr., A. M. Smith, X. Guo et al.,Anal. Chim. Acta 1997, 357, 5–12.

32. C. A. Wijayawardhana, G. Wittstock, H. B.Halsall et al., Anal. Chem. 2000, 72,339–342.

33. C. G. Bauer, A. V. Eremenko, A. Kuhn et al.,Anal. Chem. 1998, 70, 4624–4630.

34. A. A. Suleiman, Y. Xu, Electroanalysis 1998,10, 240–243.

35. Y. Xu, A. A. Suleiman, Anal. Lett. 1997, 30,2675–2689.

36. K. R. Wehmeyer, H. B. Halsall, W. R.Heineman et al., Anal. Chem. 1986, 58,135–139.

37. N. Kaneki, Y. Xu, A. Kumari et al., Anal.Chim. Acta 1994, 287, 253–258.

38. K. R. Wehmeyer, H. B. Halsall, W. R.Heineman, Clin. Chem. 1982, 28,1968–1972.

39. F. Manning, C. O’Fagain, R. O’Kennedyet al., Anal. Proc. Anal. Commun. 1994, 31,13–15.

40. G. Key, A. Schreiber, R. Feldbrugge et al.,Clin. Biochem. 1999, 32, 229–231.

41. J. Xu, J. Song, W. Guo, Anal. Lett. 1996, 29,565–573.

42. J. L. Boitieux, D. Thomas, G. Desmet, Anal.Chim. Acta 1984, 163, 309–313.

43. E. Crowley, C. O’Sullivan, G. G. Guibault,Anal. Chim. Acta 1999, 389, 171–178.

44. M. Aizawa, A. Morioka, S. Suzuki, Anal.Chim. Acta 1980, 115, 61–67.

45. Y. Xu, H. B. Halsall, W. R. Heineman, Clin.Chem. 1990, 36/11, 1941–1944.

46. M. W. Ducey Jr., A. M. Smith, X. Guo et al.,Anal. Chim. Acta 1997, 357, 5–12.

47. M. Aizawa, A. Morioka, S. Suzuki et al.,Anal. Biochem. 1979, 94, 22–28.


48. C. J. McNeil, D. Athey, M. Ball et al., Anal.Chem. 1995, 67, 3928–3935.

49. C. A. Wijayawardhana, G. Wittstock, H. B.Halsall et al., Electroanalysis 2000, 12,640–644.

50. D. Ivnitski, J. Rishpon, Biosens. Bioelectron.1996, 11, 409–416.

51. J. Wang, B. Tian, K. R. Rogers, Anal. Chem.1998, 70, 1682–1685.

52. W. Guo, J. Song, M. Zhao et al., Anal.Biochem. 1998, 259, 74–79.

53. M. J. Doyle, H. B. Halsall, W. R. Heineman,Anal. Chem. 1982, 54, 2318–2322.

54. B. Deasy, E. Dempsey, M. R. Smyth et al.,Anal. Chim. Acta 1994, 294, 291–297.

55. B. Lu, M. R. Smyth, J. Quinn et al., Electro-analysis 1996, 8, 619–622.

56. S. Kelly, D. Compagnone, G. Guilbault,Biosens. Bioelectron. 1998, 13, 173–179.

57. K. Di Gleria, H. A. Hill, C. J. McNeil et al.,Anal. Chem. 1986, 58, 1203–1205.

58. G. A. Broyles, G. A. Rechnitz, Anal. Chem.1986, 58, 1241–1245.

59. Y. Xu, H. B. Halsall, W. R. Heineman, J.Pharm. Biomed. Anal. 1989, 7, 1301–1311.

60. C. A. Wijayawardhana, H. B. Halsall, W. R.Heineman, Anal. Chim. Acta 1999, 399,3–11.

61. G. Wittstock, K. Yu, H. B. Halsall et al.,Anal. Chem. 1995, 67, 3578–3582.

62. J. Zhang, W. R. Heineman, H. B. Halsall, J.Pharm. Biomed. Anal. 1999, 19, 145–152.

63. H. T. Tang, H. B. Halsall, W. R. Heineman,Clin. Chem. 1991, 37/2, 245–248.

64. S. Rapicault, B. Limoges, C. Degrand, Anal.Chem. 1996, 68, 930–935.

65. S. F. Chen, Y. Xu, M. Po-Chee, Clin. Chem.1997, 43/8, 1459–1461.

66. M. Santandreu, F. Cespedes, S. Alegretet al., Anal. Chem. 1997, 69, 2080–2085.

67. J. Wang, P. V. A. Pamidi, K. R. Rogers,Anal. Chem. 1998, 70, 1171–1175.

68. B. Lu, M. R. Smyth, R. O’Kennedy et al.,Anal. Chim. Acta 1997, 340, 175–180.

69. H. Yao, S. H. Jenkins, A. J. Pesce et al., Clin.Chem. 1993, 39/7, 1432–1434.

70. D. A. Palmer, T. E. Edmonds, J. J. Seare,Anal. Lett. 1993, 26, 1425–1439.

71. Z. Yu, Y. Xu, M. P. C. Ip, J. Pharm. Biomed.Anal. 1994, 12, 787–793.

72. D. J. Pritchard, H. Morgan, J. M. Cooper,Anal. Chim. Acta 1995, 310, 251–256.

73. H. Shiku, Y. Hara, T. Matsue et al., J. Elec-troanal. Chem. 1997, 438, 187–190.

74. F. J. Hayes, H. B. Halsall, W. R. Heineman,Anal. Chem. 1994, 66, 1860–1865.

75. Y. Ding, L. Zhou, H. B. Halsall et al., J.Pharm. Biomed. Anal. 1999, 19, 153–161.

76. A. J. Edwards, R. A. Durst, Electroanalysis1995, 7, 838–845.

77. R. W. Keay, C. J. McNeil, Biosens. Bioelec-tron. 1998, 13, 963–970.

78. T. Jiang, H. B. Halsall, W. R. Heineman, J.Agric. Food. Chem. 1995, 43, 1098–1104.

79. S. Kroger, S. J. Setford, A. P. F. Turner,Anal. Chem. 1998, 70, 5047–5053.

80. T. Kalb, P. Skladal, Electroanalysis 1997, 9,293–297.

81. M. Dequaire, C. Degrand, B. Limoges, Anal.Chem. 1999, 71, 2571–2577.

82. F. F. Bier, E. Ehrentreich-Forster, C. G.Bauer et al., Fresenius J. Anal. Chem. 1996,354, 861–865.

83. H. Gao, T. Jiang, W. R. Heineman et al.,Fresenius J. Anal. Chem. 1999, 364, 170–174.

84. M. Del Carlo, M. Mascini, Anal. Chim. Acta1996, 336, 167–174.

85. A. G. Gehring, J. D. Brewster, P. L. Irwinet al., J. Electroanal. Chem. 1999, 469, 27–33.

86. F. G. Perez, M. Mascini, I. E. Tothill et al.,Anal. Chem. 1998, 70, 2380–2386.

87. R. M. Pemberton, J. P. Hart, J. A. Foulkes,Electrochim. Acta 1998, 43, 3567–3574.

88. R. M. Pemberton, J. P. Hart, P. Stod-dard et al., Biosens. Bioelectron. 1999, 14,495–503.

89. J. D. Brewster, A. D. Gehring, R. S.Mazenko et al., Anal. Chem. 1996, 68,4153–4159.

90. J. L. Brooks, B. Mirhabibollahi, R. G. Kroll,J. Appl. Bacteriol. 1992, 73, 189–196.

91. I. Abdel-Hamid, D. Ivnitski, P. Atanasovet al., Anal.Chim. Acta 1999, 399, 99–108.

92. J. L. Brooks, B. Mirhabibollahi, R. G.Kroll, Appl. Environ. Microbiol. 1990, 56,3278–3284.

93. J. Rishpon, D. Ivnitski, Biosens. Bioelectron.1997, 12, 195–204.

94. M. Steine, U. Bilitewski, Analyst 1997, 122,155–159.

95. S. S. Babkina, E. P. Medyantseva, H. C.Budnikov et al., Anal. Chim. Acta 1993, 273,419–424.

96. W. R. Heineman, H. B. Halsall, Anal.Chem. 1985, 57, 1321A–1331A.

97. W. J. Blaedel, R. C. Boguslaski, Anal. Chem.1978, 50, 1026–1032.

5.6 Conclusion 173

98. W. R. Heineman, P. T. Kissinger, inLaboratory Techniques in Electroanalyt-ical Chemistry, (Eds.: P. T. Kissinger,W. R. Heineman), 2nd ed., Marcel Dekker,New York, 1996, pp. 51–123.

99. M. A. Cousino, T. B. Jarbawi, H. B. Halsallet al., Anal. Chem. 1997, 69, 544A–549A.

100. U. Wollenberger, M. Paeschke, R. Hintsche,Analyst 1994, 119, 1245–1249.

101. H. T. Tang, C. E. Lunte, H. B. Halsall et al.,Anal. Chim. Acta 1988, 214, 187–195.

102. I. Rosen, J. Rishpon, J. Electroanal. Chem.1989, 258, 27–39.

103. M. Del Carlo, I. Lionti, M. Taccini et al.,Anal. Chim. Acta 1997, 342, 189–197.

104. G. Volpe, D. Compagnone, D. Draisci et al.,Analyst 1998, 123, 1303–1307.

105. B. R. Clark, E. Engvall, in Enzyme Im-munoassay, (Ed.: E. T. Maggio), CRC Press,Boca Raton, 1981, pp. 167–179.

106. R. F. Vogt Jr., D. L. Phillips, L. O. Hender-son et al., J. Immunol. Methods 1987, 101,43–50.

107. K. Kato, Y. Umedo, F. Suzuki et al., Clin.Chim. Acta 1980, 102, 261–265.

108. J. Ruzicka, E. H. Hansen, Flow InjectionAnalysis, Wiley & Sons, New York, 1981.

109. H. B. Halsall, W. R. Heineman, J. Int. Fed.Clin. Chem. 1990, 2, 179–187.

110. G. S. Sittampalam, G. S. Wilson, TrendsAnal. Chem. 1984, 3, 96–99.

111. G. Wittstock, S. H. Jenkins, H. B. Halsallet al., Nanobiology 1998, 4, 153–162.

112. The companies include Dynal Inc., ofNorway and Bangs Inc., of Indiana, USA.

113. H. Yao, H. B. Halsall, W. R. Heinemanet al., Clin. Chem. 1995, 41/4, 591–598.

114. S. G. Thompson, in Clinical Chemistry,(Eds.: L. A. Kaplan, A. J. Pesce), Mosby, St.Louis, 1989, pp. 191–206.

115. J. Wang, Stripping Analysis, VCH, DeerfieldBeach, 1985.

116. J. Wang, in Laboratory Techniques in Electro-analytical Chemistry, (Eds.: P. T. Kissinger,W. R. Heineman), 2nd ed., Marcel Dekker,New York, 1996, pp. 719–737.

117. Y. Xu, H. B. Halsall, W. R. Heineman, inImmunochemical Assays and Biosensor Tech-nology for the 1990s, (Eds.: R. M. Nakamura,Y. Kasahara, G. A. Rechnitz), American So-ciety for Microbiology, Washington, D.C.,1992, pp. 291–309.

118. W. R. Heineman, C. W. Anderson, H. B.Halsall, Science 1979, 204, 865–866.

119. S. G. Weber, W. C. Purdy, Anal. Lett. 1979,12, 1–9.

120. I. A. Alam, G. D. Christian, Fresenius Z.Anal. Chem. 1982, 15, 1449–1456.



123. C. L. Morgan, D. J. Newman, C. P. Price,Clin. Chem. 1996, 42/2, 193–209.

124. P. Skladal, Electroanalysis 1997, 9, 737–745.125. P. Treloar, J. Kane, P. Vadgama, in Prin-

ciples and Practice of Immunoassay, (Eds.:C. P. Price, D. J. Newman), 2nd ed., Mac-millan, London, 1997, pp. 483–509.

126. B. R. Eggins, Biosensors: an Introduction,Wiley-Tubner, New York, 1996.

127. J. Janata, J. Am. Chem. Soc. 1975, 97,2914–2916.

128. L. Engel, W. Baumann, Fresenius J. Anal.Chem. 1993, 346, 745–751.

129. N. Yamamoto, Y. Nagasawa, M. Sawai et al.,J. Immunol. Methods 1978, 22, 309–313.

130. M. Grotoh, E. Tamiya, M. Suzuki et al.,J. Mol. Catal. 1989, 53, 285–289.

131. C. L. Morgan, D. J. Newman, C. P. Price,Clin. Chem. 1996, 42/2, 193–209.

132. H. D. Goldberg, R. B. Brown, D. P. Liuet al., Sens. Actuators, B 1994, 21, 171–183.

133. T. Kalab, P. Skladal, Anal. Chim. Acta 1995,304, 361–368.

134. M. D. Carlo, M. Mascini, Field Anal. Chem.Technol. 1999, 3, 179–184.

135. J. P. Hart, R. M. Pemberton, R. Lux-ton et al., Biosens. Bioelectron. 1997, 12,1113–1121.

136. C. J. Zhong, M. D. Porter, Anal. Chem.1995, 67, 709A–715A.

137. A. Heller, Acc. Chem. Res. 1990, 23,128–148.

138. I. Katakis, A. Heller, in Frontiers in Biosen-sorics 1, (Eds.: F. W. Scheller, F. Schubert,J. Fedrowitz), Birkhauser Verlag, Boston,1997, pp. 229–241.

139. D. J. Litman, T. M. Hanlon, E. F. Ullman,Anal. Biochem. 1980, 106, 223–229.

140. D. V. Brown, M. E. Meyerhoff, Biosens. Bio-electron. 1991, 6, 615–622.

141. C. N. Campbell, T. de Lumley-Woodyear,A. Heller, Fresenius J. Anal. Chem. 1999,364, 165–169.

142. J. D. Newman, A. F. P. Turner, G. Mar-razza, Anal. Chim. Acta 1992, 262, 13–17.


143. J. P. Hart, S. A. Wring, Electroanalysis 1994,6, 617–624.

144. P. Connolly, TIBTECH 12, 1994, 123–127.145. J. Lahiri, E. Ostuni, G. M. Whitesides, Lang-

muir 1999, 15/5, 2055–2060.146. D. J. Pritchard, H. Morgan, J. M. Cooper,

Angew. Chem., Int. Ed. Engl. 1995, 34, 91–93.147. Section 9.10 discusses the applications

of SECM in bioelectrochemistry. Also,the reader is referred to the original

paper on SECM, A. J. Bard, R.-R. F. Fan,D. T. Pierce et al., Science 1991, 254, 68–74.

148. H. Shiku, T. Matsue, I. Uchida, Anal. Chem.1996, 68, 1276–1278.

149. A. L. Ghindilis, R. Krishnan, P. Atanasovet al., Biosens. Bioelectron. 1997, 12,415–423.

150. A. van der Berg, P. Bergveld (Eds.), Micro-total Analysis Systems, Kluwer AcademicPress, Boston, 1995.

175

6Electrochemistry ofMetalloporphyrins inNonaqueous Media

Karl M. Kadisha and Eric Van Caemelbeckea,b

aUniversity of Houston, Houston, TexasbHouston Baptist University, Houston, Texas

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1776.1.1 Porphyrin Electrochemistry in the Early 1960s and the 1970s . . . . . 1786.1.2 Porphyrin Electrochemistry in the 1980s and the 1990s . . . . . . . . . 1796.1.3 Selection of Appropriate Solvent and Supporting Electrolyte . . . . . . 1796.1.4 Selection of Appropriate Supporting Electrolyte . . . . . . . . . . . . . . 180

6.2 Effect of Macrocycle Structure on Potentials . . . . . . . . . . . . . . . . . 181

6.3 Effect of Axial Ligation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186

6.4 Periodic Table of Metalloporphyrins . . . . . . . . . . . . . . . . . . . . . . 1916.4.1 Groups 1–4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1916.4.2 Groups 5 and 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1926.4.3 Group 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1936.4.4 Group 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1956.4.4.1 Ruthenium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1956.4.4.2 Osmium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1966.4.4.3 Iron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1966.4.4.3.1 Reaction of Heme-thiolate Proteins . . . . . . . . . . . . . . . . . . . . . . . 2036.4.4.3.2 Mechanism of Electron-transfer in Peroxidases . . . . . . . . . . . . . . . 2046.4.4.3.3 Reactions of Nitrite Reductions . . . . . . . . . . . . . . . . . . . . . . . . . 2056.4.4.3.4 Reactions of Nitric Oxide Reductases . . . . . . . . . . . . . . . . . . . . . 2056.4.5 Group 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2066.4.6 Group 10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2096.4.7 Group 11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2116.4.8 Group 12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2136.4.9 Group 13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2146.4.10 Group 14 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2156.4.11 Group 15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216

176 6 Electrochemistry of Metalloporphyrins in Nonaqueous Media

6.5 Concluding Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218

177

6.1Introduction

Thousands of different metalloporphyrinshave been electrochemically investigatedin close to three dozen different nonaque-ous solvents over the last three decades.These include compounds with hundredsof different macrocycles and coordinatedaxial ligands, and 70 different metal ornonmetal ions, of which some have beenshown to exist in three or four differentoxidation states (see the ‘‘Periodic Tableof Metalloporphyrins’’ in Fig. 1 and ThePorphyrin Handbook, Volumes 1–10 [1]).

Despite a large variety of known com-pounds, or perhaps because of it, previousreviews on metalloporphyrin electrochem-istry have concentrated for the mostpart on describing the behavior of ‘‘sim-ple’’ model compounds with octaethyl-porphyrin (OEP) or tetraphenylporphyrin(TPP) [2–10] macrocycles, and were mostoften arranged according to a specific ele-ment (Fe, for example), group of elements,the Periodic Table of the elements or thenature of the metal–ligand bond; examplesin the latter case include porphyrins withmetal–carbon bonds [3, 4, 11, 12] andthose with metal–metal bonds [4, 11, 13].Unfortunately, these approaches and per-spectives are truly useful only if the readeris made aware of how changes in structure

and reactivity of a given compound may berelated to the well-studied OEP and TPPderivatives via linear free energy relation-ships. This point was elucidated in a majorreview by Kadish, Van Caemelbecke, andRoyal [7], which recently appeared in theliterature.

This present review will not attempt toprovide a comprehensive description of allknown porphyrin electrochemistry in non-aqueous media, but will concentrate in parton specific types of porphyrin macrocycles,in part on specific groups of metallopor-phyrins, and in part on guiding the readerthrough the vast array of electron-transfermechanisms that can exist for a relatedseries of compounds under a given set ofexperimental conditions. It is hoped thatthis approach will answer the majority ofthe reader’s questions as to what has beendone in the past, while at the same timeenabling the reader to utilize the data inthe literature to predict what might beobserved in future studies involving theelectrochemistry of yet-to-be synthesizedmetalloporphyrin complexes.

Our current discussion will, therefore,begin with a brief historical overview ofmetalloporphyrin electrochemistry and alist of previously utilized nonaqueous sol-vents and supporting electrolytes, whichhave been used in porphyrin studies. Itwill be followed by a general description


1 18

H 2 13 14 15 16 17 He

Li Be B C N O F Ne

Na Mg 3 4 5 6 7 8 9 10 11 12 Al Si P S Cl Ar

K Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn Ga Ge As Se Br Kr

Rb Sr Y Zr Nb Mo Tc Ru Rh Pd Ag Cd In Sn Sb Te I Xe

Cs Ba La Hf Ta W Re Os Ir Pt Au Hg Tl Pb Bi Po At Rn

Fr Ra Ac

Lanthanides Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

Actinides Th Pa U Np Pu Am Cm Bk Cf Es Fm Md No Lr

Fig. 1 Periodic Table of metalloporphyrins. Shaded elements indicate specific elementsthat have been incorporated into a given porphyrin macrocycle. Extensive tabulations ofredox potentials may be found in Ref. [21].

of metalloporphyrin redox behavior, ar-ranged first according to the type ofmacrocycle and then according to the typeof central metal or nonmetal ion in thecomplex. We will not include in this reviewa detailed discussion of solvent or sub-stituent effects on metalloporphyrin redoxreactions since both topics have recentlybeen covered in great detail [7].

6.1.1Porphyrin Electrochemistry in the Early1960s and the 1970s

The electrochemistry of metalloporphyrinsat the start of the 1960s involved, in largepart, measurements of standard redox po-tentials for naturally occurring complexesin aqueous buffered media [14]. The choiceof an aqueous solvent was often dictated bythe biological relevance of the compoundsavailable for study, while the choice ofthe measurement technique (potentiome-try or polarography at a dropping mercuryelectrode) was necessitated by the type ofavailable electrochemical instrumentation,virtually all of which was homemade and

limited almost exclusively to electrochem-ical laboratories. However, the situationbegan to change in the mid-1960s due tothree main factors. The first was the pub-lication of relatively easy-to-understandpapers on cyclic voltammetric theory [15,16], which led to the popularization of thistechnique as a rapid and efficient methodfor obtaining reversible redox potentials.The second was the increased use ofnonaqueous solvents for studying electro-chemical reactions [17–20], and the thirdwas the increased ‘‘availability’’ of easilysynthesized tetraphenyl and octaethylpor-phyrins with a wide variety of differentcentral ions.

By the beginning of the 1970s, the major-ity of electrochemical studies on syntheticmetalloporphyrins was being carried outin nonaqueous media using the techniqueof cyclic voltammetry. However, mostutilized instrumentation was still home-made and only a handful of laboratorieswere actually making the measurements.An overview of the situation at this pe-riod is provided in several independentreviews [2, 6, 7, 9, 21].

6.1 Introduction 179

6.1.2Porphyrin Electrochemistry in the 1980sand the 1990s

A number of papers on the electrochem-istry of synthetic porphyrins in the 1980swere involved with the following top-ics: (1) porphyrins with metal–carbon andmetal–metal bonds [11–13, 22], (2) por-phyrins with unusual structures, (3) por-phyrins with central metal ions invery high [23] or very low oxidationstates, (4) porphyrins able to react withsmall molecules and (5) porphyrins ableto carry out a specific catalytic func-tion [24–26].

A number of porphyrins with newand/or ‘‘unusual’’ macrocycles were syn-thesized in the 1980s and 1990s andthe redox properties of these compoundswere often compared to potentials forthe oxidation or reduction of analogouscompounds containing the same metalion and the well-known TPP or OEPmacrocycle. Examples include, but are notlimited to, highly halogenated metallopor-phyrins [27–39], metalloporphyrins withhighly distorted macrocycles [37, 40–45],and metalloporphyrins linked to one ormore other porphyrins or to different redoxactive molecules [46].

The Periodic Table of Metallopor-phyrins, presented first by Buchler [47] andthen by Kadish [2], was expanded in the1980s to include most of the Main Group,lanthanide, and actinide elements, as wellas a number of the Group 15 nonmetals. Atthe same time, the range of known metalor nonmetal oxidation states that couldbe accommodated by a given metallopor-phyrin was expanded in part by the useof ‘‘novel’’ axial ligands or macrocycles,in part by the use of different solventsor solvent conditions (i.e. low tempera-ture), and in part by the application of

new instrumental techniques that wereable to identify previously unobserved orunreported transient intermediates in thevarious redox reactions.

The electrochemistry of metallopor-phyrins in the 1980s and 1990s was nolonger limited by the need to constructappropriate instrumentation, which hadbecome commercially available at rela-tively low cost and was becoming standardanalytical equipment in a large numberof nonelectrochemical laboratories aroundthe world. The reporting of redox po-tentials for newly synthesized porphyrinsthus became routine and, more oftenthan not, most studies of metallopor-phyrin electrochemistry had as their mainfocus the use and application of electro-chemical techniques toward the solvingof chemical problems. Several key areasof chemical problems that were examinedin recent years include (1) the effect ofmacrocycle distortion on porphyrin redoxpotentials [30, 42, 45, 48], (2) quantitatingrelationships between structure and chem-ical or electrochemical reactivity with theuse of linear free energy relationships,that is, substituent effects, (3) the useof metalloporphyrins to activate smallmolecules [24–26], (4) the elucidation ofrelationships between redox potentials andcatalytic properties of a given metallopor-phyrin or group of metalloporphyrins [24],(5) the study of donor–acceptor inter-actions between two linked porphyrinsor between a metalloporphyrin and an-other redox active center (e.g. such as afullerene [46]).

6.1.3Selection of Appropriate Solvent andSupporting Electrolyte

The earliest electrochemical studiesof metalloporphyrins in aprotic media


utilized dimethylformamide (DMF) ordimethylsulfoxide (DMSO) for reductionsand benzonitrile (PhCN), acetonitrile(CH3CN) or butyronitrile (BuCN) foroxidations. Electrochemical data are nowavailable in close to three dozen differentnonaqueous solvents with the majorityof data appearing in the nonbindingsolvent dichloromethane (CH2Cl2) [21].One advantage of CH2Cl2 is its largecathodic and anodic potential range, whichcan span from −1.9 to +1.9 V vs asaturated calomel electrode (SCE), thusenabling one to examine both oxidationsand reductions of a given compound underthe same single set of solution conditions.

Table 1 lists the most common nonaque-ous solvents that have previously beenutilized for the electrochemistry of met-alloporphyrins. Purification procedures,potential limits and physical characteris-tics of these solvents are given in theliterature [17–20, 49, 50].

The selection of a specific nonaqueoussolvent from the list in Table 1 was basedin many cases only on the ‘‘habit’’ ofthe individual laboratory, but in others itvery much depended on the requirementsof the individual experiment. The firstrequirement is, of course, solubility andthe electrochemical parameters that oneis investigating. One must also considerease of purification, the chemical reactivityof the solvent, its ability to stabilizeπ -anion or π -cation radicals, and itsoverall potential range for both oxidationand reduction, the latter of which willdepend in part on the type of electrodematerial (Hg for example cannot beused for oxidations, while Ag and Auboth have a limited positive range insolvents containing some anions.). Otherpractical factors include the cost of thesolvent, its toxicity, and its general ease ofhandling.

Tab. 1 List of solvents used in electrochemicalstudies of metalloporphyrins

Solvent Nameabbreviation

1,1,1-TCE 1,1,1-Trichloroethane1,2-DBE 1,2-Dibromoethane1,2-DME 1,2-Dimethoxyethane2-MeTHF 2-MethyltetrahydrofuranC6H5Cl ChlorobenzeneC2H2Cl2 DichloroethaneC2H2Cl4 1,1,2,2-TetrachloroethaneC6H4Cl2 1,3-DichlorobenzeneC6H6 BenzeneCH2Br2 DibromomethaneCH2Cl2 Dichloromethane

(methylene chloride)CHCl3 TrichloromethaneCl-Naph 1-ChloronaphthaleneDMA N, N-DimethylacetamideDMF N, N-DimethylformamideDMSO DimethylsulfoxideEtOH EthanolEtONa Sodium ethoxideHMP HexamethylphosphoramideHQ HydroquinoneMe2CO AcetoneMe2NH DimethylamineMeCN AcetonitrileMeCO2Et Ethyl acetateMe-Naph 1-MethylnaphthaleneMeNO2 NitromethaneNMA N-MethylacetamideNMF N-MethylformamidePC Propylene carbonatePhCN BenzonitrilePrCN n-ButyronitrilePy PyridineTHF Tetrahydrofuran

6.1.4Selection of Appropriate SupportingElectrolyte

A list of the supporting electrolytes thathave been used for studies of metal-loporphyrin electrochemistry is given inTable 2 [21]. The majority of electrochemi-cal studies involving metalloporphyrins in

6.2 Effect of Macrocycle Structure on Potentials 181

Tab. 2 List of supporting electrolytes used inmetalloporphyrin studies arranged by size of thecation

Electrolyte Nameabbreviation

KCN Potassium cyanideLiBr Lithium bromideLiCl Lithium chlorideLiClO4 Lithium perchlorateNaClO4 Sodium perchlorateNH4PF6 Ammonium

hexafluorophosphateTMACl Tetramethylammonium

chlorideTMAP Tetramethylammonium

perchlorateTMAPF6 Tetramethylammonium

hexafluorophosphateTEABF4 Tetraethylammonium

tetrafluoroborateTEACl Tetraethylammonium

chlorideTEAP Tetraethylammonium

perchlorateTEAPF6 Tetraethylammonium

hexafluorophosphateTPrABF4 Tetra-n-propylammonium

tetrafluoroborateTPrAP Tetra-n-propylammonium

perchlorateTPrASO3CF3 Tetra-n-propylammonium

trifluoromethylsulfonateTBABF4 Tetra-n-butylammonium

tetrafluoroborateTBABr Tetra-n-butylammonium

bromideTBACl Tetra-n-butylammonium

chlorideTBAOH Tetra-n-butylammonium

hydroxideTBAOTeF5 Tetra-n-butylammonium

pentafluorooxotellurateTBAP Tetra-n-butylammonium

perchlorateTBAPF6 Tetra-n-butylammonium

hexafluorophosphateTBASO3CF3 Tetra-n-butylammonium

trifluoromethylsulfonateTHAP Tetra-n-hexylammonium

perchlorateTHASbF6 Tetra-n-hexylammonium

hexafluoroantimonate(V)

nonaqueous media have utilized tetraalky-lammonium salts as supporting elec-trolytes, the most common of whichhave been the tetrabutylammonium per-chlorates and tetraethylammonium per-chlorates (abbreviated TBAP and TEAP).Several studies have also utilized tetraalky-lammonium salts of BF4

− or PF6−, (i.e.

TBABF4 or TBAPF6). The selection of onesupporting electrolyte over another maydepend on the cost and ease of purification,or it may have been the need for a labora-tory to select a salt that had a minimumor a maximum coordination ability withthe investigated porphyrin in its neutral,electrooxidized or electroreduced form.Usually, ClO4

−, BF4− and PF6

− can beconsidered as nonbinding or very weaklybinding anions and this sometimes hasbeen a major factor in their selection forstudies of electrooxidation processes.

Finally, there is the question of support-ing electrolyte concentration. Most mea-surements have been made in solutionscontaining 0.1 M TBAP (the most oftenutilized salt), but others have utilized solu-tions with 0.2 M supporting electrolyte, es-pecially in the case of spectroelectrochem-ical measurements. Attention should bepaid to this fact since redox potentials mea-sured in solutions of 0.1 M TBAP are notalways identical to those measured withTBAP concentrations of 0.01 or 1.0 M. Theexperimentally obtained differences in po-tential may amount to several hundredmillivolts and will vary as a function of thespecific metalloporphyrin and the specificelectrode reaction examined [21].

6.2Effect of Macrocycle Structure on Potentials

All known metalloporphyrins are elec-troactive (see summary of selected redox


potentials [21]), and virtually all are ableto undergo three or more electron-transferreactions in nonaqueous media, with theexact number of processes depending onthe potential range of the utilized electro-chemical solvent, the type of macrocycle,the type of central metal ion, and/or thetype of axially coordinated ligands.

The first synthetic porphyrins whoseelectrochemistry was studied in nonaque-ous solvents were largely those with TPPor OEP macrocycles. The electrochemicalbehavior of the TPP and OEP complexesare generally similar to each other, but themore basic OEP derivatives are almost al-ways easier to oxidize and harder to reducethan the TPP complexes containing thesame central metal ion and the same setof axial ligands as shown in Fig. 2(a) for(TPP)Zn and (OEP)Zn.

The absolute potential difference be-tween half-wave potentials for oxidationand reduction of a given (OEP)M or(TPP)M complex was initially claimed tobe a constant value of 2.25 ± 0.15 V [51,52], but there are now many exam-ples in the literature where this is notthe case.

Several important trends in the elec-trochemical behavior of synthetic metal-loporphyrins have been pointed out overthe last 35 to 40 years. The first is thatmetalloporphyrins can be reduced by two,and only two, electrons at the conjugatedmacrocycle to give porphyrin π -anion rad-icals and dianions. The same compoundscan also be oxidized by two, and onlytwo, electrons at the conjugated macrocy-cle to give π -cation radicals and dications.The porphyrins may also undergo one ormore metal-centered reactions, some ofwhich have been unambiguously assignedas involving the metal orbitals and oth-ers of which are open to continuing

debate as to the site of electron trans-fer. Most metal-centered redox processeshave been reported to occur at poten-tials that are located between E1/2 forformation of the porphyrin π -cation andπ -anion radical, although examples havebeen reported where this seems not to bethe case.

Early electrochemical studies on met-alloporphyrins containing OEP or TPPmacrocycles have led to often-quoted di-agnostic criteria for differentiating reac-tions, which occur at the π -conjugatedmacrocycle as opposed to those that oc-cur at the central metal ion [51, 52].These include a constant HOMO-LUMOgap (2.25 ± 0.15 V [51, 52]) (highest occu-pied molecular orbital-lowest unoccupiedmolecular orbital) independent of themetal oxidation state and a constant dif-ference between E1/2 for the two stepwisereversible one-electron reductions or twostepwise reversible one-electron oxidationsat the porphyrin π -ring system. However,more recent investigations of porphyrinswith macrocycles other than OEP or TPPhave shown that those diagnostic crite-ria are not always followed, and in manycases, the values of E1/2 seem to berelated to the planarity of the macrocy-cle, the nature of the metal ion, and/orthe presence or absence of specific axiallycoordinated axial ligands, which preferen-tially stabilize one oxidation state over theother.

A schematic representation of severalporphyrin macrocycles, which have beenstudied with regard to their electrochem-ical properties in different metallatedforms, is given in Fig. 3 and abbreviationsfor a much larger group of investigatedcompounds is given in Table 3.

The electrochemistry of (TMP)M, (R8

TPP)M, (DPP)M, (Br8TPP)M, (Br8TMP)M,


(a)

(b)

1.02 0.67

1.140.82

2.00 1.50 1.00 0.50 0.00 −0.50 −1.00

−1.58

−1.32 −1.70

−1.50 −2.00

Potential[V vs SCE]

(TPP)Zn

(OEP)Zn

1.2

0.96

0.82

−1.32

−0.82

0.4E

[V vs SCE]

(Br8TPP)Zn

(TPP)Zn

−0.4 −1.2

Fig. 2 Cyclic voltammograms of (a) (TPP)Zn and (OEP)Zn and (b) (TPP)Zn and(Br8TPP)Zn.

(F20TPP)M, and other macrocycles de-scribed in Fig. 3 has been discussed ina recent review [7]. All of the porphyrinsexhibit reductions and oxidations at the π -conjugated macrocycle, but the potential atwhich these redox reactions are located willdepend on the planarity and basicity of themacrocycle that can effect each oxidationand reduction to a different degree, thusleading to substituent effects that can dif-fer substantially depending on the specificsite of electron transfer.

For example, the oxidation, but notthe reduction, of porphyrins with highlydistorted macrocycles, such as R8TPP(R = an alkyl group), DPP, Br8TPP, andF20Br8TPP, are all generally easier (oc-cur at a more negative potential) than isobserved in the case of analogues hav-ing an OEP or a TPP macrocycle, andthe same metal ion/axial ligand combi-nation [7]. The easier oxidation has beenexplained theoretically as resulting froma destabilization of the HOMO due to


N N

N NM

F5

F5

F5

N N

N NM

SO3−

SO3−

SO3−

−O3S

Br

N N

Br

BrBr

Br

N N

Br

BrBr

M

N

N N

N

N NN NM

Me

Me

Me

Me

+

+

+

+

F5

(F20TPP)M

[(TPPS)M]4−

(Br8TPP)M

[(TMpyP)M]4+

Fig. 3 Selected porphyrin macrocycles.

macrocycle distortion [35, 42, 43]. In thecase of (Br8TPP)M and (Br8TMP)M, thereis a systematic anodic shift in E1/2 forthe reduction upon going from TPP toBr8TPP [29, 30, 35, 36] or from TMPto Br8TMP [7] as a macrocycle, and thiscontrasts with much smaller differencesbetween E1/2 for the first oxidation of thesame compounds or series of compounds.This is illustrated in Fig. 2(b), which com-pares cyclic voltammograms of (TPP)Znand (Br8TPP)Zn under the same solu-tion conditions. As seen in this figure, thedifference in reduction potentials amounts

to 500 mV, while the oxidations are sepa-rated by only 140 mV, thus leading to asubstantially reduced HOMO-LUMO gapin the case of (Br8TPP)Zn.

Also, it should be pointed out thatsome (Br8TPP)M derivatives sometimesundergo additional electroreductions thatare not observed for (TPP)M, and theseinvolve, in at least one case, a stepwiseelimination and electroreduction of the Brgroups on the porphyrin macrocycle [34].

Porphyrins with F20TPP macrocycles areall easier to reduce and more difficultto oxidize than the same compounds


Tab. 3 List of selected porphyrin macrocycles (see Fig. 3 for selected structures ofthese macrocycles)

Abbreviation Common name

TPP TetraphenylporphyrinOEP OctaethylporphyrinTMP TetramesitylporphyrinTBP TetrabenzoporphyrinBr8TPP OctabromotetraphenylporphyrinT(F5)PP Tetrakis-(pentaphenyl)porphyrinDPP DodecaphenylporphyrinMe8TPP OctamethyltetraphenylporphyrinT(p-Me)PP Tetrakis-(p-methylphenyl)porphyrintriPh((o-NH2)Ph)P 5-(o-Aminophenyl)triphenylporphyrin(NO2)TPP 2-Nitro-tetraphenylporphyrin(2-O)OEP Oxophlorin(5-NO2)OEP 5-Nitro-octaethylporphyrinT(CF3)P Tetrakis(perfluoromethyl)porphyrinTPPS(4) Tetrakis(p-(sodiosulfonato)phenyl)porphyrin[(TMpy(4)P)]4+ 5,10,15,20-Tetrakis(1-methylpyridinium-4-yl)porphyrin(N-Ph)TPP (N-Phenyl)-tetraphenylporphyrinDPIXDME Deuterioporphyrin IX dimethyl esterEtio EtioporphyrinHPIXDME Hematoporphyrin IX dimethyl esterMPIX Mesoporphyrin IXMPIXDME Mesoporphyrin IX dimethyl esterPPIX Protoporphyrin IXPPIXDME Protoporphyrin IX dimethyl esterT(o-piv)PP Meso- α, α, α, α-tetrakis(o-pivamidophenyl)porphyrinTPP(a-(C12)2-CT) α-5, α-15, β-10, β-20-Bis[2,2′-dodecane-

diamidodiphenylene]porphyrin

with a TPP macrocycle, and potentialdifferences of 440 to 570 mV have beenreported between compounds in these twoseries [7, 21].

Several porphyrins with Br8F20TPP,Cl8F20TPP or Me8F20TPP macrocycleshave been investigated for their elec-trochemical properties [7, 21, 44] and,as expected on the basis of the highlyelectron-withdrawing halogen groups, dif-ferences of up to 800–900 mV havebeen reported between E1/2 for reduc-tion of the (TPP)M and (X8F20TPP)Mcomplexes with the same metal ion. Theelectron-withdrawing halogen groups onthe porphyrin macrocycles of Br8F20TPP

and Cl8F20TPP not only induce a pos-itive shift in the potentials but alsosometimes lead to a merging of thetwo one-electron ring-centered oxidations,thus giving what appears to be a singleoverall two-electron-transfer process. De-tails of this interesting electrochemicalfeature have been presented in the liter-ature [7]. A now well-known, but perhapsunusual, electrochemical behavior is seenfor porphyrins having positively chargedmacrocycles, as in the case of theTMpyP derivatives (see structure in Fig. 3).The reduction of most (TMpyP)MII,(TMpyP)MIII and (TMpyP)MIV complexesin nonaqueous media generally occurs in


0.00 −0.30 −0.60

E[V vs SCE]

[(TMpyP)Zn]4+

[(TMpyP)Cu]4+

−0.90 −1.20

Fig. 4 Cyclic voltammograms of [(TMpyP)M]4+ where M = Cu(II)or Zn(II) in DMF, 0.1 M TBAP. (Adapted from K. M. Kadish,C. Araullo, G. B. Maiya, D. Sazou, J.-M. Barbe, R. Guilard, Inorg.Chem. 1989, 28, 2528–2533.)

multiple overlapping two-electron-transfersteps [53–55] (Fig. 4) and a discussion ofthe electron-transfer mechanism has re-cently been reviewed [7].

6.3Effect of Axial Ligation

A large number of different axial lig-ands have been complexed to metallopor-phyrins, and 252 examples taken from theliterature are summarized in Table 4 [21].Each ligand has been bound to one ormore complexes in nonaqueous mediaand a summary of potentials, along with

abbreviations of the ligands is found inRef. [21].

Metalloporphyrins containing differentσ -bonded alkyl or aryl groups [3, 45,56–64] or the diatomic molecules NO [27,65–77] or CO (see following sections)have been electrochemically investigatedand several reviews of the general elec-trochemical behavior of these compoundshave been published [7, 12, 78]. Metallo-porphyrins with 29 different central ionsare now known to form carbon σ -bondedcomplexes (see Fig. 5) [12] and the elec-trochemical behavior of these compoundshas been shown to depend on the centralmetal ion.

6.3 Effect of Axial Ligation 187

Tab. 4 List of axial ligands which have been coordinated to metalloporphyrins

:CH=CHPh− (:CH2CH2Ph)− -(CH2Ph)CH-(CO)2 (CO)3 (NPh)2

2−:C(NHCH2Ph)2 :CHCH2Ph 1-(MeCO)Im1,2-Me2Im 1-MeIm 2-CNPy2-MeIm 3-(MeCO)Py 3-(NH2)Py3-(OH)Py 3-(OMe)Py 3,4-Me2Py3,5-Cl2Py 3-BrPy 3-ClPy3-CNpy 3-MePy 4-(Me2N)Py4-(MeCO)Py 4-(MeO)Py 4-(NH2)Py4-(NMe2)Py 4-(OMe)Py 4,4′-bipy4-CNPy 4-MeIm 4-MePy4-NH2py 4-OH-pip 4-PhIm4-PhPy 5,6-Me2BzIm acac−AuCl4− azpy BF4

−biPy Br− BuBuCN BzIm C=C(p-C6H4Cl)2C=CPh2 C=CPh2 C2H2PhC2H2Ph C2H3 C2H3O2

−C2Ph C3H6Br C3H6ClC3H6I C4H8Br C4H8ClC4H8I C5H10Br C5H10ClC5H10I C5H11 C6F4HC6F5 C6H11 C6H12IC6H13 C6H4CN C6H4MeC6H4SO3

− C8H13 cat2−CCl3 CH=C(p-C6H4Cl)2 CH=CPh2CH2Br CH2Cl CH2ICH2Ph CH2Ph CHCl2CHI2 CHMe2 CHO2

−CI3 Cl− ClO4

−CMe3 cMU CN−CNCH2Ph CO Co(CO)4CO2Et COEt COMeCOPr Cr(CO)3Cp CSCSe dabco DMFDMS DMSO dppeEt EtNH2 EtOHF− facam− FcFe(CO)4 H− HSO4

−I− Im m-(OH)C6H4NH2m, m-C6F2H3 Me Me2NHMeCN Mn(CO)5 Mo(CO)3CpN3

− N4(CH=CH2) N4(CH=CHCN)N4C(CH=CH2) N4C(CH=CHCN) N4C(CMe3)N4C(m-C6H4Me) N4C(p-C6H4(NO2)) N4C(p-C6H4Me)N4C2(CO2Me)2 N4CEt N4CMeNCS− NH2OH NH3NHMe2 NMe3 NONO2

− NO3− NPh2

−

(continued overleaf )


Tab. 4 (continued)

NS o, m, p-C6H2F3 o, o, p-C6H2F3O−•

2 O2− O22−

O2C(3-ClC2H4)− O2C(o-C6H3Cl)− O2CCCl3−

O2CCF3− O2CCH2Cl− O2CCH2Ph−

O2CCHCl2− O2CCMe3− O2CEt−

O2CH− O2C-m-C6H3NO2− O2CMe−

O2CPr− OBu− OC2F5−

OC3HF6− OC6H2(NO2)3

− o-C6H4EtOCH2Ph− OCMe3 OCN−O2CPh− o-cresol− OEt−OH− OH2 O-m, m-Bu2, o-(OH)C6H2

−O-m, p-C6H3Me2

− OMe− O-o, p-C6H3(NO2)2−

O-o, p-C6H3Me2− O-p-C6H3CN− O-p-C6H3NO2

−O-p-C6H4Me− O-p-C6H4NO2

− OPh−OPh− O-p-PhO-o, p-(NO2)2Ph− O-p-PhO-p-(NO2)Ph−OPr− OPr− OTeF5

−P(OEt)3 P(OMe)3 p-C6H4Brp-C6H4Me p-C6H4Cl p-C6H4Fp-C6H4NO2 p-C6H4OMe p-cresol−PEt3 PF3 PF6

−Ph Ph PhCH2NH2PhNH2 pip PPh2MePPh3 PPhMe2 PrPrOH Py PyanPyen PyPh Pyrazolepyz Quinine QuinuclidineRe(CO)5 S(CH2)4 S2−SbF6

− SC6F4H− SCN−Se2− SH− SnPh3

−S-o-(CF3CONH)C6H4 S-o, o-C6H3(CF3CONH)2 SO3CF3

−SO3(p-Me)C6H4

− SO3Me− SO3Ph−SO4

2− S-p-C6H4Me− SPh−tdt2− THF THFtMU trans-1,2-(4-py)2C2H2 W(CO)3Cp

For example, σ -bonded porphyrins withRh, P, As, or Sb central ions undergoreversible oxidations and reductions at theπ -conjugated macrocycle and there is littleeffect of the σ -bonded axial ligand on theporphyrin electrochemical behavior [12,21]. However, derivatives with Al, Ga, Inor Tl central metals all undergo a rapidcleavage of the metal-carbon bond after aone-electron oxidation of the compound.In contrast, the Fe, Ru, and Co σ -bondedmetalloporphyrins (see further sections ofthe review) will undergo a metal-centered

oxidation followed by a migration of theaxial ligand to one of the four nitrogens ofthe porphyrin macrocycle.

The σ -bonded Fe and Co complexesare generally unstable upon reductionat the metal center and this electrodereaction may be followed by a cleavageof the metal-carbon bond [12]. This is notthe case for the osmium sigma-bondedcomplex (OEP)Os(Ph)2, which undergoesboth oxidations and reductions at themetal center without loss of the axialligand [79].

6.3 Effect of Axial Ligation 189

1 18

2 13 14 15 16 17 He

B C N O F Ne

3 4 5 6 7 8 9 10 11 12 Al Si P S Cl Ar

Cu Zn Ga Ge As Se Br Kr

Pd Ag Cd In Sn Sb Te I Xe

Pt Au Hg Tl Pb Bi Po At Rn

H

Li Be

Na Mg

K Ca Sc Ti V Cr Mn Fe Co Ni

Rb Sr Y Zr Nb Mo Tc Ru Rh

Cs Ba La Hf Ta W Re Os Ir

Fr Ra Ac

Fig. 5 Periodic Table of sigma-bonded metalloporphyrins. All 27 shaded elementshave been synthesized but only the 16 lightly shaded ones have been electrochemicallyinvestigated.

A significant number of nitrosyl metal-loporphyrins has been synthesized overthe past two decades [78, 80]. (see Pe-riodic Table of nitrosyl porphyrins inFig. 6) and the most well-characterizedof these complexes have been thosewith Fe(II) [5, 66, 67, 71, 76, 77, 81–89]or Co(II) [27, 67, 70, 90–93] centralmetal ions.

Porphyrin-nitrosyl complexes with sixother metal ions are also known, and allbut one of which has been electrochemi-cally investigated. These are: Ru [69, 73,94–96], Os [5], Rh [97], Cr [98], Mo [99]and Mn [100]. Some nitrosyl metallopor-phyrins can be reversibly reduced or oxi-dized by one or two electrons without lossof the NO ligand and this generally occurswhen the electrode reactions involve theπ -conjugated macrocycle; in the case ofa metal-centered reduction or oxidation,however, the electron-transfer reactionswill most often be accompanied by a lossof the NO ligand, resulting in an irre-versible oxidation as shown in Fig. 7 for thecase of (TPP)Cr(NO) and (TPP)Mn(NO) inCH2Cl2.

In this regard, it should be noted that theloss of NO may be sufficiently slow so thatthe electrode reaction appears reversibleon the cyclic voltammetry and/or thin-layer

timescales of 0.1 to 15 s but not on the bulk-electrolysis timescale, which could be aslong as 20–30 minutes depending on thedesign of the utilized electrochemical cells.The stability of the metal-NO bond maybe related to the site of electron additionor electron abstraction (ring, metal orNO axial ligand) but it may be alsorelated to the nature of the electrochemicalsolvent (bonding vs nonbonding) that willcompete with NO for an axial coordinationposition on the metal.

The actual site of electron transferupon reduction or oxidation of porphyrinswith bound NO groups has not beenwell established except for compoundswith Fe(II) or Co(II) central metals,where an oxidation of the former complexinvolves the Fe(II)/Fe(III) redox coupleand an oxidation of the latter involveselectrogeneration of a Co(II) porphyrin π -cation radical [7].

Several types of metalloporphyrins withcarbonyl ligands have been electrochemi-cally investigated. The first is representedby (P)[Rh(CO)2]2 [101] and (P)[Ir(CO)2]2[102] where P = TPP or OEP, but the onesmost often studied for their electrochem-istry have been the air-stable porphyrinswith a single CO axial ligand such as(P)MII(CO) and (P)MII(CO)(L) where L is a


1 18

2 13 14 15 16 17 He

B C N O F Ne

3 4 5 6 7 8 9 10 11 12 Al Si P S Cl Ar

Cu Zn Ga Ge As Se Br Kr

Pd Ag Cd In Sn Sb Te I Xe

Pt Au Hg Tl Pb Bi Po At Rn

H

Li Be

Na Mg

K Ca Sc Ti V Cr Mn Fe Co Ni

Rb Sr Y Zr Nb Mo Tc Ru Rh

Cs Ba La Hf Ta W Re Os Ir

Fr Ra Ac

Fig. 6 Periodic Table of nitrosyl metalloporphyrins. All 9 shaded elements have beensynthesized but only the 7 lightly shaded ones have been electrochemically investigated.

0.80 0.00

Co

Fe

Mn

Cr

Potential[V vs SCE]

−0.80 −1.60

Fig. 7 Cyclic voltammograms of(TPP)MII(NO) (M = Cr, Mn, Fe or Co)in CH2Cl2 containing 0.1 M TBAP.(Adapted from S. L. Kelly, D. Lancon,K. M. Kadish, Inorg. Chem. 1984, 23,1451–1458.)

nitrogeneous base or solvent molecule andM = Ir [103] or Ru [92, 104–113]. A thirdclass of porphyrins with carbonyl ligandsare those that coordinate CO moleculesafter in situ electrogeneration of the por-phyrin in a specific metal oxidation state.Examples in this category include deriva-tives of (P)FeII(CO)x and [(P)CoIII(CO)x ]+where x = 1 or 2 [36, 114, 115].

The porphyrins with nonlabile CO ax-ial ligands undergo almost exclusivelymacrocycle-centered reductions and oxi-dations, which is not the case when thereis a metal-centered reaction as is observedfor (P)FeII(CO)x and [(P)CoIII(CO)x ]+. Forinstance, Fe(II) porphyrins coordinate oneor two CO molecules, but no coordination

is seen for the Fe(III) and Fe(I) formsof the porphyrin. In a similar manner,Co(III) porphyrins coordinate either one ortwo CO ligands depending on the solutionconditions, but neither the Co(II) nor Co(I)forms of the porphyrin bind CO. Thus, theFe(II) and Co(III) porphyrin carbonyl com-plexes lose their axial CO molecules whenconverted electrochemically to another ox-idation state.

Electrogenerated Ni(I) porphyrins havealso been proposed to bind CO [116] but astable product has never been isolated inthe solid state.

Finally, it should be noted that Rh(II)porphyrins with PF3 [117] or thiocar-bonyl axial ligands [118] have also been

6.4 Periodic Table of Metalloporphyrins 191

examined and the electrochemical datafor these complexes have been com-pared to the results for related (P)Rh(CO)derivatives under similar experimentalconditions.

6.4Periodic Table of Metalloporphyrins

Most free-base porphyrins that have beeninvestigated to date undergo two reversibleone-electron oxidations and two reversibleone-electron reductions, but in some casesthe electrogeneration of a monoanion ordianion is also accompanied by additionof protons on the macrocycle [119, 120].The electrochemistry of metalloporphyrinswith metals from Group 1 through 15of the Periodic Table has recently beenreviewed [7], and only a brief summary ofthe most important results will be given inthe present chapter.

6.4.1Groups 1–4

No electrochemistry of Group 1 porphyrinshas yet been reported in the literature. TheGroup 2 porphyrins usually exhibit elec-trode reactions that involve formation ofπ -cation radicals and dications upon oxi-dation and π -anion radicals and dianionsupon reduction. However, in some casesthe second reduction is shifted beyond thenegative potential limit of the solvent, thusprecluding observation of this reaction.

With the exception of the Sc(III) com-plexes, the electrochemistry of Group3 metalloporphyrins is limited in largepart to double- and triple-decker deriva-tives of the type (P)2M, (P)(Pc)La and(P)La(Pc)M(P), where Pc represents aphthalocyanine and M = Gd(III), Y(III)or Lu(III) [121–126]. These species all

have a rich electrochemistry that was re-cently reviewed by Buchler and Ng [7].The electrochemically examined mononu-clear Sc(III) porphyrins are representedby (OEP)Sc(OH), which shows straight-forward redox behavior involving the por-phyrin macrocycle [52, 128].

The Group 4 porphyrins, representedby complexes with M = Ti, Zr, andHf, have been studied in detail, butagain, with the exception of the Ticomplexes, most of these porphyrins ex-ist as (P)2M and (P)(Pc)M complexes,where P = OEP or TPP and M = Zr orHf [127]. A number of Ti porphyrins withoxo, peroxo [129–132], halogen [133–134]or chalcogen [135–137] axial ligands havebeen synthesized and many of these com-pounds have been studied as to theirelectrochemical properties. The Ti por-phyrins can exist with the metal ion ina +4, +3 or +2 oxidation state, and inmost cases, other than for the titanyl por-phyrins, the electrode reactions of thesecompounds are accompanied by chemi-cal reactions involving a loss of the axialligand [7].

The electrochemistry of (P)2Zr involvesthe conjugated π -ring system. The twomacrocycles interact with each otherthrough the bridging Zr ion and thehalf-wave potentials for their oxidationand reduction occur at different half-wavepotentials [7, 127], as is expected whena molecule contains two equivalent andinteracting redox centers. The (P)2Zr com-plexes should, in principle, show up tofour reductions and four oxidations (as-suming that the redox reactions of thetwo macrocycles occur at different halfwave potentials as a result of the in-teracting equivalent redox centers), butfewer redox processes are actually observedwithin the potential window of the sol-vent/supporting electrolyte system, which


ranges in the most optimal of cases be-tween +2.0 V and −2.0 V versus SCE inone or more solvents.

6.4.2Groups 5 and 6

The most studied of the Group 5porphyrins have been compounds withvanadyl and niobium metal ions; no elec-trochemical data have yet been reportedfor tantalum porphyrins. Vanadium por-phyrins have been synthesized as bothV(IV) and V(II) complexes. The vanadylderivatives, represented as (P)VO, gener-ally undergo well-defined reduction andoxidation reactions involving the por-phyrin, macrocyle; the lower oxidationstate V(II) complexes have been preparedas (P)VII(L) and (P)VII(L)2 derivatives,where P = T(p-Me)PP or OEP and L =THF or PPhMe2, but these species havenot been examined as to their electrochem-ical properties [7].

The synthesis of several niobium por-phyrins has been described in theliterature, but the electrochemistry ofthese compounds is limited to mononu-clear derivatives of (P)NbVO(O2CMe) and(P)NbO, where P = OEP or TPP [138–140]and dinuclear derivatives of [(P)Nb]2O3

where P = T(p-CH3)PP or OEP [141].Porphyrins with all of the Group 6

metals have been investigated as to theirelectrochemistry, but most data have beenobtained for the Cr and Mo complexes [7].The only porphyrins with a tungstenmetal ion whose electrochemistry has beenreported are the oxo/hydroxo derivatives(P)WO(OH) [52, 142]. Both metal andring-centered processes are observed uponreduction of these W(V) porphyrins, anda similar electrochemical behavior hasgenerally been reported for most Cr(V)

and Mo(V) porphyrins having the same setof axial ligands.

The earliest electrochemistry of chro-mium porphyrins involved Cr(III) deriva-tives with anionic axial ligands and/orcoordinated nitrogeneous bases, but anumber of studies have since been car-ried out with complexes having Cr(II),Cr(IV) or Cr(V) central metal ions. Thelatter series of compounds are exemplifiedby the oxo-Cr(IV), oxo-Cr(V), and nitrido-Cr(V) [143–147] derivatives. Cr(IV) µ-oxodimers are also known [148, 149].

A one-electron oxidation of the oxo-Cr(IV) porphyrin will lead to an oxo-Cr(V)species, and this result contrasts with thenitridochromium(V) porphyrins [150], allof which exhibit only macrocycle-centeredelectron transfers upon oxidation.

Solvent effects on the redox poten-tials and electroreduction mechanisms of(TPP)CrCl have been discussed by Bas-solo and Hoffman [151, 152], as well asby Bottomley and Kadish [153]. The for-mer research group demonstrated that(TPP)CrCl can readily coordinate withLewis bases containing oxygen, sulfuror nitrogen donor atoms to form six-coordinate Cr(III) species of the type(TPP)CrCl(L), while the latter group in-dicated that the Cr(III)/Cr(II) process wasaccompanied by chemical reactions cou-pled with reversible electron transfers. Thecurrent-voltage curves led the authors topropose the electron-transfer mechanismshown in Sch. 1.

The redox reactions of (TPP)CrClO4,(TPP)CrClO4(L) and [(TPP)Cr(L)2]+ werealso investigated and correlations betweenE1/2 for reduction of (TPP)CrClO4 andthe Gutman solvent donor number [154],as well as between E1/2 for reductionof (TPP)CrClO4 or (TPP)CrII and thesolvent dielectric constant, were also ex-amined [98].


Scheme 1 ‘‘Box mechanism’’for the first reduction of(P)CrX(L), where X is a halideand L a solvent molecule or anitrogeneous base.

(TPP)CrCl(L) [(TPP)CrCl(L)]−

[(TPP)Cr(L)2]+ (TPP)Cr(L)2

Cl− L

e−

−e−

Cl− L

Molybdenum porphyrins containing me-tal ions in +2, +4 or +5 oxidationstates are known, but only the complexeswith a +4 or +5 metal oxidation statehave been investigated as to their elec-trochemical properties. Electrochemicallyexamined porphyrins with Mo(IV) ionshave been of the type (P)MoO(X), whereX = F−, Cl−, Br−, NCS−, OH−, OMe−or ClO4

− and P = OEP or TPP [155, 156],while the Mo(VI) porphyrins are repre-sented by bis-peroxo complexes of the type(P)Mo(O2)2, where P = T(p-Me)PP, T(m-Me)PP and TPP [157]. The first type ofmolybdenum porphyrins have been shownto undergo only macrocycle-centered re-ductions, but both macrocycle- and metal-centered redox processes were reportedfor oxidation of the same compounds.Detailed electrochemical studies of ox-omolybdenum(V) porphyrins have beenreported in the literature [2, 7], andpotentials of the Mo(V)/Mo(IV) couplewere shown to vary with the nature ofthe anionic axial ligands on (P)MoO(X)as well as with the covalent charac-ter of the MoV-X bond. The bis-peroxo-molybdenum(VI) porphyrins of the type(P)Mo(O2)2 were studied for their elec-trochemical properties in CH2Cl2 0.1 MTBAP, and the compounds were shownto undergo one macrocycle-centered ox-idation and two metal-centered reduc-tions [157].

Cis-dioxomolybdenum(VI) porphyrinshave also been examined after theaddition or abstraction of electrons.The reduction [158] of (TPP)Mo(O)2

yields (TPP)MoO, while oxidation of thecompound [159] leads to [(TPP)MoO]+.

6.4.3Group 7

The most extensively studied of the Group7 porphyrins are the maganese complexesthat can exist in up to four differentoxidation states [23]. In fact, no electro-chemistry of technetium porphyrins hasyet been reported, and the electrochemistryof rhenium porphyrins has been limited tocompounds of the type (P)HRe(CO)3 and(P)[Re(CO)3]2 where P = TPP, TMTAA orTAA [160–162].

Although manganese porphyrins havebeen characterized as stable derivativeswith metal oxidation states of +2, +3, +4and +5, most electrochemistry has beencarried out with the Mn(III) derivativescontaining anionic axial ligands [7]. Highvalent porphyrins containing manganesewith a +4 or +5 oxidation state have beendescribed for their spectroscopic [23] andcatalytic properties [24], but relatively littleelectrochemical data have been publisheddue to the high reactivity of these com-pounds.

All Mn(III) porphyrins examined to dateare easily reduced to their Mn(II) form,and this electrode reaction is usually fol-lowed at more negative potentials by theformation of a Mn(II) π -anion radical anddianion, the latter of which may or may notbe observed depending on the cathodic po-tential window of the solvent and basicityof the porphyrin macrocycle. The Mn(III)


porphyrins can be oxidized at the metalcenter or at the porphyrin π -conjugatedsystem with the exact site of electron trans-fer depending on the nature of the axialligands. Either a Mn(IV) porphyrin or aMn(III) porphyrin π -cation radical is gen-erated in the first oxidation of the Mn(III)complex. The formation of an Mn(IV) por-phyrin should be followed at more positivepotentials by oxidation of the macrocycle togive a Mn(IV) porphyrin π -cation radical(assuming that the Mn(IV) form of the por-phyrin does not react with the solvent or anadded component), whereas the formationof a Mn(III) porphyrin π -cation radical isgenerally followed at more positive poten-tials by a second electron abstraction fromthe macrocycle to give a Mn(III) porphyrindication.

The specific pathway by which a Mn(III)porphyrin is oxidized will depend on thetype of macrocycle and the specific set ofaxial ligands. The site of electron transferand the potentials for oxidation may alsobe a function of the porphyrin ring basicityor its planarity, the nature of the solvent,the type of counterion on Mn(III), andthe basicity and/or steric effects of anybound axial ligands [7, 24]. A summary ofthe possible reactions that might occur areillustrated in Sch. 2.

Five major types of manganese por-phyrins have been electrochemically char-acterized to date. These are (1) (P)MnIIIX,where X is a halide ion or an anionicspecies such as C2H5O2

−, N3−, OCN−,

OH− or SCN−, (2) (P)MnIII(N4CR)

where N4CR is a tetrazolato axial lig-and [163], (3) (P)MnII(NO) [100], (4) oxo-Mn(IV) complexes, (P)MnIVO [23, 164,165], and (5) bimetallic (P)Mn complexeswith µ-nitrido bridges.

The first type of porphyrins have beencharacterized for their Mn(III)/Mn(II)

electrode reactions under a variety of dif-ferent solution conditions [7, 21]. Earlyelectrochemical studies by Boucher andGarber focused on E1/2 values of theMn(III)/Mn(II) reaction as a function ofthe porphyrin macrocycle [167], and thiswork was followed several years laterby studies of electrochemical substituenteffects on a series of (T(p-X)PP)MnCl com-plexes [166]. The effect of counterion [168]and the effect of solvent and/or supportingelectrolyte [166] on the Mn(III)/Mn(II) re-action has also been reported. The Mn(II)porphyrins formed upon reduction of(P)MnIIIX are usually stable under a ni-trogen atmosphere, but this may not bethe case in the presence of oxygen thatmay bind to the metal center.

Manganese porphyrins containing non-planar macrocycles such as DPP and Fx

DPP (x = 4, 8, 20 or 28), T(2,6-(OMe)2, 3,5-Cl2)PP have been synthesized [169] andinvestigated electrochemically [170, 171],as have Mn(III) complexes with a posi-tively charged macrocycle [53] such as inthe case of [(TMpyP)MnCl]4+. The DPPderivatives show electrochemical behaviorsimilar to manganese porphyrins with pla-nar macrocycles (such as TPP), but thestability of the reduced and oxidized formsof the porphyrins as well as the potentialsat which they are electrogenerated will varywith the type, number, and position of theadded substituents.

Manganese porphyrins with a nitrosylaxial ligand have also been examined fortheir electrochemical properties. The oxi-dation of (TPP)Mn(NO) to its Mn(III) formis accompanied by a loss of the NO lig-and (Fig. 7). However, the two reductionsof the compound are reversible and step-wise generate a Mn(II) nitrosyl porphyrinπ -anion radical and a nitrosyl porphyrindianion at more negative potentials [100].High valent manganese porphyrins have


[(P)MnIII]2+•(P)MnII

[(P)MnII]2−

[(P)MnIII]3+ [(P)MnII]−[(P)MnIV]3+[(P)MnV]3+ •

•

[(P)MnIII]+

[(P)MnIV]2+

−e− −e− −e− e−

e−e−

−e−

−e− −e−e−

[(P)MnV]2+[(P)MnV]4+ •

[(P)MnV]+[(P)MnV]5+

•

e−−e−

Scheme 2 Electron-transfer reactions of manganese(III) porphyrins.

been used as oxidation catalysts [24, 172,173], and studies [174] have focused on theformation of these species via an electro-chemical oxidation of the Mn(III) complex.The substitution of Cl− by OH− as a coun-terion on the manganese porphyrin canlead to a Mn(IV) species as the singly oxi-dized product and a similar site of electrontransfer may be seen upon oxidation of(P)MnIIICl complexes with specific typesof porphyrin macrocycles (such as DPPor substituted DPP), or when the electro-chemistry of the porphyrin is carried outat low temperature [174]. A Mn(IV) specieshas also been proposed for the singly elec-trooxidized product of the µ-nitrido dimer(TPP)Mn-N-Fe(Pc) in CH2Cl2 [175].

6.4.4Group 8

6.4.4.1 RutheniumThe first ruthenium porphyrins studiedwere of the type (P)RuII(CO). The electron-transfer mechanism for reduction andoxidation of these compounds has beenexamined under different solution con-ditions [104, 105], and the nature of the

singly reduced or singly oxidized producthas been shown to depend on the solventin which the electrochemistry was inves-tigated. For example, a Ru(I) porphyrinwas proposed as a reduction product of(OEP)Ru(CO) in MeCN, PhCN or PrCN,but a Ru(II) porphyrin π -anion radical wasgenerated upon reduction of the samecompound in DMSO or Py [176]. Ruthe-nium porphyrins of the type (P)Ru(L) or(P)Ru(L)2, where L is a thioether, sulfox-ide or benzoate axial ligand have also beenexamined for their electrochemical prop-erties [177]. Those compounds usually un-dergo two one-electron oxidations and asingle one-electron reduction. The first ox-idation was proposed to be metal-centered,while the second oxidation was shownto involve the porphyrin macrocycle. Afew examples of trans-dioxoruthenium(VI)porphyrins containing TMP, OEP, TPP,T(p-Cl)PP, T(p-Me)PP, or (T(p-OMe)PP)macrocycles have been reported and stud-ies of these compounds have shown thatthe compounds usually undergo one oxi-dation and one reduction [178, 179]. Theoxidation was proposed to involve the por-phyrin π -ring system while the reduction,


which was irreversible, was assigned asinvolving a conversion of Ru(VI) to Ru(V).

The electrochemistry of ruthenium por-phyrins coordinated to small molecules,such as CS [118], PF3 [117], NO [69, 72, 73,96], or bound by a σ -bonded alkyl or arylgroup [57, 180–182], has also been doc-umented in the literature. The electrore-duction of (T(p-Me)PP)RuII(CS)(L) whereL = EtOH, CN, Im or Py has been studiedin THF, DMF or Et2Cl2, and a mech-anism involving a two-electron-transferprocesses, accompanied by an uptake ofprotons, has been proposed by Latos-Grazynski and coworkers [118]. Kadishand coworkers [117] reported the electro-chemistry of (P)Ru(PF3), where P = TPP,T(p-Br)PP, T(p-Me)PP, T(p-Et)PP or OEP.The authors pointed out similarities be-tween the electrochemistry of the CO andPF3 derivatives in CH2Cl2, but differencesin electrochemical behavior were observedin other nonaqueous solvents.

The electrochemistry of ruthenium ni-trosyl porphyrins has been investigatedby several research groups [69, 72, 73,96]. The electrochemistry of sigma-bondedruthenium(III) porphyrins [57, 180–182]parallels that reported for sigma-bondediron(III) porphyrins [12]. A migration ofthe axial ligand follows the one-electron,or in some cases, two-electron oxidationof the sigma-bonded iron(III) complex [2,12, 45, 56, 183, 184]; and a migration ofthe axial ligand from the metal of thesingly oxidized porphyrin to one of thefour nitrogens of the macrocycle has alsobeen observed for σ -bonded Ru(III) por-phyrins [12].

6.4.4.2 OsmiumOsmium porphyrins can exist with themetal ion in one of several different ox-idation states and derivatives of Os(II),Os(III), Os(IV), Os(V), or Os(VI) have

been studied for their electrochemicalproperties [7]. (OEP)OsII(CO)(Py) can beconverted to its Os(III) form upon aninitial one-electron oxidation and to itsOs(III) porphyrin π -cation radical formafter the abstraction of a second elec-tron at more positive potentials [110]. Os-mium(III) derivatives can be reduced andoxidized at the metal center [5, 110, 185].The first oxidation of Os(IV) porphyrins orthe first reduction of Os(V) porphyrins ofthe type (P)Os(L)2 where P = MPIXDME,TPP, OEP, T(p-OMe)PP, T(p-Me)PP, orT(p-Cl)PP and L = OEt−, OMe−, OiPr−,OPh−, SPh−, SC6F4H−, SC6H4Me−, Br−or Ph are also metal-centered [79, 186].A summary of the various redox mecha-nisms and half-wave potentials for selectedredox reactions is given in the litera-ture [7, 21].

6.4.4.3 IronThe electrochemistry of iron porphyrinsin nonaqueous media has been discussedin several reviews [2, 7, 10, 12], andonly a few of the major trends ofiron porphyrin electrochemistry will besummarized in the current paper. Bothhigh and low oxidation states of themetal ion can be accessed upon reductionor oxidation of iron porphyrins andthe overall electron-transfer mechanismof these metalloporphyrins is shown inSch. 3, where [(P)FeIII]+ represents theinitial compound in the absence of anassociated anionic ligand.

In general, most synthetic iron por-phyrins are able to undergo three orfour electron-transfer reactions in a va-riety of nonaqueous solvents, with theexact number of redox reactions depend-ing on the type of porphyrin macrocycle,the type of axial ligands, and the sol-vent/supporting electrolyte mixture used


[(P)FeIII]+

[(P)FeIII]2+ [(P)FeIV]2+•

[(P)FeI]−

[(P)FeI]2−

[(P)FeIV]3+[(P)FeIII]3+

−e− −e− e−

e−

•

[(P)FeIV]4+

−e−

[(P)FeII]−•

e− (?)

[(P)FeII]2−•

−e−−e−e−

e−

e−

(P)FeII

Scheme 3 Electron-transfer reactions of iron(III) porphyrins.

to make the electrochemical measure-ment. (TPP)FeX, where X is a halide ion,undergoes three well-defined reductionsand two well-defined oxidations in mostsolvents [187], while (TPP)Fe(NO) [68] and(TPP)Fe(C6H5) [188] show two reductionsand two oxidations, respectively. The elec-trode reactions of these compounds aregenerally all reversible and all involve ei-ther the conjugated macrocycle or the cen-tral metal ion to give porphyrins in Fe(IV),Fe(III), Fe(II) or Fe(I) oxidation states.

The type of axial ligand(s) coordinated tothe iron center will determine not only theredox potentials, rates and mechanismsfor electron transfer but will also stronglyinfluence the oxidation state and spin stateof the central iron ion and the associatedchemistry of the neutral, electrooxidized,and electroreduced forms of the porphyrin.Virtually, all monomeric iron porphyrinswith halide or perchlorate axial ligandscontain Fe(III) in their air-stable form, buta conversion of Fe(III) to Fe(II) is readilyaccomplished at potentials generally lo-cated between +0.12 and −0.5 V vs SCE.

The Fe(III)/Fe(II) electrode reactions ofsynthetic (P)FeX complexes have been ex-tensively investigated over the last 30 yearsand have been the subject of several de-tailed reviews [2, 7, 10]. The effect of

solvent and counterion on half-wave po-tentials of iron porphyrins was examinedby several research groups between 1975and 1981, but the most systematic studyof how solvent and axially coordinatedmonovalent anions will affect the redox po-tentials was undertaken by Bottomley andKadish, who characterized reductions of(TPP)FeX with five different anions in 12different nonaqueous solvents [187]. Theresults of this study showed that the bind-ing strength of the counterion to the Fe(III)porphyrin in CH2Cl2 increased in the or-der: ClO4

− < Br− < Cl− < N3− < F−. A

similar stabilization of Fe(III) over Fe(II)by the counteranion was also observed inPhCN and DMF, but the effect was lesspronounced in the latter two solvents.

The half-wave potentials for the Fe(III)/Fe(II) reaction of (TPP)FeX varied little asa function of the counteranion in DMSOor Py, and this was accounted for by adisplacement of the halide axial ligandby a bound solvent molecule leading to[(TPP)FeIII(S)2]+ and (TPP)FeII(S)2 in so-lution, where S = DMSO or Py. The effectof solvation on the Fe(III)/Fe(II) reactionof (TPP)FeX was quantitated by Bottom-ley and Kadish who correlated E1/2 valuesfor reduction with the Gutmann solventdonor number [187]. Half-wave potentials


for reduction of Fe(II) porphyrins are alsorelatively insensitive to the counterion onthe Fe(III) form of the complex, with theonly exceptions being compounds withX = OH− or OMe−, where E1/2 is shiftednegatively (from compounds with Cl− orClO4

− counterions) as a result of an en-hanced stabilization of the Fe(II) form ofthe porphyrin, which maintains the anionas a bound axial ligand [189, 190].

Fe(III) and Fe(II) porphyrins may befour, five, or six coordinate in solutionand will contain a high, intermediateor low-spin iron center depending onthe type of axial ligand [80]. Most six-coordinate Fe(III) porphyrins with axiallycoordinated nitrogeneous bases are lowspin, whereas most six-coordinate Fe(III)porphyrins coordinated with oxygen donorligands are high spin. A variety of electron-transfer mechanisms is possible uponconversion of the initial Fe(III) complex toits Fe(II) form, and selected examples thatare most often observed are summarizedin Sch. 4.

The relationship between ligand-bindingstrength of different nitrogenous bases toiron(II) or iron(III) porphyrins and E1/2

for reduction of these compounds was re-ported in independent studies by Constantand Davis [191] and Kadish and Bottom-ley [192, 193]. Iron(II) complexes of theform (TPP)Fe(L)2, where L was one ofeleven different substituted Pys were elec-trochemically investigated by Bottomleyand Kadish, who showed that E1/2 forthe Fe(II)/Fe(I) reaction was directly pro-portional to the aqueous pKa value of thePy ligand [192, 193]. The thermodynamicsof ligand addition to Fe(II) porphyrins, asmeasured by electrochemical techniques,has also been reported [194, 195].

The effect of ortho phenyl substituentson the spectroscopic, redox properties andaxial ligand binding constants for a series

of iron(III) substituted TPPs was examinedby Walker and coworkers [196]. The au-thors concluded that the ortho-halogens orortho-CF3 groups were electron donatingas a result of a direct overlap between theelectron cloud of the substituents and theπ system of the porphyrin. Several theo-retical papers have discussed the influenceof substituents at the meta, para, and orthopositions of the meso-phenyl groups onthe ionization potentials of ortho, meta,and para -substituted TPPs [197–200].Other studies have examined the electron-transfer rate constants of iron porphyrinsand looked at relationships between rateconstants and axial coordination [2, 10].

The effect of substituents on electron-transfer rate constants has also beenreported [201], and a summary of thesestudies have been given in several re-views [2, 7, 10]. The previously open-to-question assignment of the Fe(II) reduc-tion product as an Fe(I) species or anFe(II) porphyrin π -anion radical [202] nowseems definitive in many cases [203, 204],but the exact site of electron transfer willvary as a function of the solution condi-tions as well as the porphyrin macrocycle.

The reduction of Fe(II) porphyrinsappeared from most early studies tobe relatively insensitive to changes insolvent [187], and this was due to theweakly complexing nature of the Fe(II)form of the porphyrin with virtually allutilized electrochemical solvents exceptfor Py. However, the complete lack ofsolvent binding by Fe(I) porphyrins wasless than clear-cut and it is now knownthat some singly reduced iron(II) por-phyrins can bind a Py molecule at lowtemperature under certain solution con-ditions [204]. Recent electrochemical stud-ies on nonplanar iron(III) porphyrins ofthe type (FxDPP)FeCl [203], where x = 0,12, 20, 28 or 36 and (T(C3F7)P)FeCl or


e−

FeIII

X

X−

X− FeIII

XL

FeIII

L+

or FeIII

L

XL

X− FeIII

L+

L

FeII

X−

X− FeIIL

FeII

L

e−

LFeII

L

L

e−e−

e−e−

e−e−

e−Le−

L

FeII − •

FeI− or

e−

X−

Scheme 4 Electroreduction of iron(III) porphyrins in coordinating media.

(T(C3F7)P)Fe(py)2 [205], have revealed thatthe site of reduction in iron(II) porphyrinswill depend not only on the specific por-phyrin macrocycle but also on the type ofaxial ligation to iron(II).

Some iron(III) porphyrins can be oxi-dized at the metal center to give an iron(IV)species [23], while others are oxidized atthe porphyrin π -conjugated system to pro-duce an iron(III) porphyrin π -cation radi-cal. Early studies involving solvent and/orcounterion effects of iron porphyrins indi-cated that compounds of the type (P)FeX,where P = OEP or TPP, were oxidizedat the porphyrin ring [187, 206, 207], andsimilar conclusions were also reached bythe groups of Goff and Reed [208–211]on the basis of IR spectroscopy. Sev-eral iron(III) porphyrins bound to weak-field ligands, such as ClO4

−, SO3CF3−

or C(CN)3−, were also examined as to

their electrochemical properties and both(TPP)FeClO4 and (TPP)FeSO3CF3 wereproposed to produce iron(III) porphyrinπ -cation radicals upon oxidation [212].

Fuji examined a series of (P)FeCl com-plexes containing Cl groups at either themeso or the β-pyrrole positions of themacrocycle [213]. All of the compounds

were oxidized at the porphyrin macrocy-cle and E1/2 values for these electrodereactions were located between 1.08 and1.45 V, depending on the specific por-phyrin macrocycle. The half-wave poten-tials were shifted toward more positivepotentials with increase in the electron-withdrawing affinity of the substituents onthe macrocycle, but the magnitude of theshift in E1/2 was larger for those com-pounds that had substituents at the mesopositions than for those that were sub-stituted at the β-pyrrole positions of themacrocycle.

Most iron(III) porphyrins with axiallybound halide groups such as Cl− areinitially oxidized at the porphyrin π -ring system, independent of the type ofporphyrin macrocycle. However, a switchin the electron-transfer site from the ringto the metal may occur when iron(III)porphyrins are axially bound by hydroxy,methoxy, or σ -bonded groups such asC6H5, or C6F4H [7, 12].

The direct electrochemical oxidation ofan iron(III) porphyrin to give an iron(IV)complex has been observed on a numberof occasions, but the high oxidation stateproduct is often quite reactive [23, 24,26, 189, 214–218]. Thus, the oxidations


of (TMP)Fe(OH) and (TMP)Fe(OMe) areirreversible at room temperature (due toa chemical reaction involving the Fe(IV)species), but both oxidized products wereformulated as Fe(IV) derivatives at lowtemperature [189, 214, 215, 217, 218]. Thereaction was shown to involve an ECprocess in which the electrochemical stepis the oxidation of the macrocycle and thechemical step the elimination of HClO4

and the following rearrangement of themolecule to its oxo form (Sch. 5) [214].

Iron(IV) porphyrins can be generatedupon oxidation of σ -bonded iron(III)porphyrins [7, 12, 62, 184, 188, 219]with the stability of the electrooxidizedproduct, depending in large part on theporphyrin macrocycle and the type of σ -bonded axial ligand [12]. The chemical orelectrochemical oxidation of a bis-fluoroFe(III) complex, [(T(p, m-F2)PP)Fe(F)2]−,has been examined by Nanthakumar andGoff [220, 221] and the iron oxidation statewas assigned as Fe(III) or Fe(IV), on thebasis of NMR and UV-visible spectroscopy.

Kadish and coworkers examined theeffect of porphyrin ring distortion on theredox potentials of (BrxTPP)FeCl, wherex varied from 0 to 8 [7, 29, 30]. Thefirst reduction of the compounds variedlinearly with the number of Br groups onthe macrocycle, but not the first oxidation(Fig. 8).

A nonlinear relationship between E1/2

for oxidation of (BrxTPP)FeCl and x wasobtained and was interpreted in terms oftwo opposite effects; one was the electron-withdrawing effect of the Br substituents,

which stabilized the HOMO, and the otherwas the increased nonplanarity of themacrocycle, which resulted in a destabi-lization of the HOMO [29]. Similar effectsof macrocycle distortion on the electro-chemistry of other nonplanar porphyrinshave been reported [7, 37, 42, 45, 170, 171,184, 219, 222–225]. Several picket fenceand basket-handle porphyrins have alsobeen investigated for their electrochemicalproperties [226–236].

Several iron porphyrins bound to di-atomic molecules, such as CO, NO, CS,CSe, and O2, have also been examinedas to their electrochemistry in nonaque-ous media. Fe(II) porphyrins can coor-dinate CO to give mono- and bis-COderivatives [237–239], and the electrooxi-dation of these species by cyclic voltamme-try results in irreversible waves becauseof a rapid loss of CO upon formation ofFe(III) [7, 30, 240]. Studies of (TPP)FeCland (TPP)FeClO4 in Py and CH2Cl2/Pymixtures under a CO atmosphere indica-ted that the following five types ofiron(II) porphyrins could be formed:(TPP)Fe, [(TPP)FeCl]−, [(TPP)Fe(CO)Cl]−,(TPP)Fe(py)2, and (TPP)Fe(CO)(py), andthat these could be electrochemically con-verted into two types of iron(I) porphyrins,namely [(TPP)Fe]− and [(TPP)Fe(CO)(py)]− [240].

Studies of (BrxTPP)FeCl under a COatmosphere revealed that the numberof CO molecules bound to the iron(II)form of the porphyrin depends on thenumber of Br groups on the macrocy-cle [30]. Plots of νCO vs x were constructed

(TMP)FeIII(OH)−e−

(TMP+ )FeIII(OH)(CIO4)

(TMP)FeIV(OH)(CIO4)

•

(TMP)FeIV = O−HClO4

−HClO4

Scheme 5 Electrooxidation of hydroxyiron(III) porphyrins.


0 2 4

Number of bromides6 8 0 2 4

Number of bromides6 8

1.600.15

0.05

−0.05

−0.15

−0.25

−0.35

Only inductive effect

First oxidation First reduction

1.50

1.40

1.30

E1/

2(V

vs

SC

E)

E1/

2(V

vs

SC

E)

1.20

1.10

Fig. 8 Plot of E1/2 versus number of Br groups on (BrxTPP)FeCl in PhCN, 0.1 MTBAP. (Adapted from K. M. Kadish, F. D’Souza, A. Villard, M. Autret, E. VanCaemelbecke, P. Bianco, A. Antonini, P. Tagliatesta, Inorg. Chem. 1994, 33, 5169,5170.)

and the resulting nonlinear correlationinterpreted in terms of two competingeffects, one of which was the electron-withdrawing effect of the Br substituents,and the other the nonplanarity of themacrocycle [30].

Stable Fe(II) porphyrins complexed withNO have been characterized as (P)Fe(NO),(P)Fe(NO)2, and (P)Fe(NO)(L) [80], al-though the original formulation of (P)Fe(NO)2 has now been called into questionin favor of an (P)Fe(NO)(ONO) assign-ment [76, 77]. Lancon and Kadish investi-gated the electrochemistry of (P)Fe(NO)and (P)Fe(NO)(L) in 10 different elec-trochemical solvents [68]. The electrooxi-dation of (P)Fe(NO) reversibly leads to[(P)Fe(NO)]+ in a nonbinding solvent,whereas the electroreduction reversiblyleads to [(P)Fe(NO)]−. The initial oxida-tion of (P)Fe(NO) was proposed (on the

basis of IR and electron spin reso-nance (ESR) spectroscopy) to give thebis-nitrosyl complex [(P)Fe(NO)2]+ClO4

−under an NO atmosphere [83], but morerecent studies by Lorkovic and Ford sug-gest the formation of (P)Fe(NO)(ONO) [76,77]. Ryan and coworkers [66] reportedthat (P)Fe(NO) (P = TPP or OEP) un-dergoes three one-electron reductionsin nonaqueous solvents and the prod-ucts of the first two one-electron reduc-tions were characterized by UV-visiblespectroelectrochemistry.

A large number of redox processesmay occur for iron nitrosyl complexes inbinding solvents such as DMSO or Py,and examples of several pathways, whichhave been characterized or proposed aresummarized in Sch. 6.

The reversible half-wave potentials foroxidation of iron(II) porphyrins bound to


L

Fe

L

NO

e− e− e−e− e−L

LNO

Fe

L

NO

Fe

L

Fe

L

X

X

LFe

L

L

Fe

NO

Fe

NO

Fe

X

NO

NO

Fe

NO

NO

Fe

NO

NO

Fe

NO

X

NO

ll ll llL

−NO

Ill Ill Ill Il

+

NOX

Ill

+

Ill

+

LIll

+

Ill

+

L

Scheme 6 Electrooxidation of nitrosyl iron(II) porphyrins.

NO, CS, and CSe are shifted positivelywith respect to the values of E1/2 forreduction of the same porphyrin inthe absence of the axial ligand [241].The product formed after electrochemi-cal oxidation of (OEP)Fe(O2) has beenidentified as an η2-dioxygen adduct [7].Fe(II) porphyrins complexed by CS, CSe,and NO show no loss of the diatomicmolecule upon oxidation. Six-coordinate(P)Fe(CS)(py) [242] and (P)Fe(CS)(py) [243]are reversibly oxidized by one electron atE1/2 = 0.73 and 0.70 V, respectively, inCH2Cl2, and these potentials may be com-pared to an E1/2 = 0.52 V for oxidation of(P)Fe(NO)(py) [71] under the same solu-tion conditions.

Two iron porphyrins may be bridgedthrough a single bridging atom X to forma binuclear metalloporphyrin of the type(P)Fe-X-Fe(P), and the electrochemistryof these compounds where X = O, N,or C has been recently reviewed [7]. Theformal oxidation states of the two ironatoms are +3 for the µ-oxo complexes,+3.5 for the µ-nitrido compounds and

+4 for the µ-carbido derivatives, butnone of these assignments is unambigu-ous, given the existing spectroscopic data.On the other hand, the electrooxidationof µ-oxo, µ-nitrido, and µ-carbido por-phyrins is relatively straightforward, andwith one exception [244], has been shownto occur in multiple reversible single-electron-transfer steps, in which each ofthe two porphyrin macrocycles is oxidizedat a separate half-wave potential consistentwith an interaction of the two macrocyclesthrough the bridging ligand. The µ-oxoporphyrin dimer [(T(p-Et2N)PP)Fe]2O canbe oxidized by a total of four electrons inCH2Cl2 [244]. Two one-electron oxidationsof the compound occur at E1/2 = 0.34 and0.47 V, while a subsequent two-electronoxidation step is seen at E1/2 = 0.76 V.The oxidation of one porphyrin macro-cycle in [(P)Fe]2O is more difficult thanthe oxidation of the other, and the differ-ence in E1/2 can be accounted for by anelectrostatic interaction between the twoelectroactive sites of the molecule.


The µ-nitrido complex, [(TPP)Fe]2N, isstable in CH2Cl2, benzonitrile or Py, andup to four electrons can be abstractedfrom the compound without a loss of thenitrido-bridged structure [245]. A nitrido-bridged Fe(IV)-Fe(IV) dimer was proposedfor the singly oxidized compound while thedoubly oxidized product was described asa µ-nitrido Fe(IV)-Fe(IV) π -cation radical.

[(TPP)Fe]2C in CH2Cl2 can be oxi-dized by four electrons in four successiveone-electron-transfer steps, but the num-ber of oxidations varies with the natureof the nonaqueous solvent [246]. Neu-tral [(TPP)Fe]2C was shown to containtwo Fe(IV) units by Mossbauer spec-troscopy [247], and the four redox pro-cesses of this porphyrin most likely corre-spond to a stepwise oxidation of the π -ringsystem to form initially two individual π -cation radicals (one on each macrocycle)and then two porphyrin dications at morepositive potentials [246].

A comprehensive review of σ -bondediron porphyrin electrochemistry has re-cently been published [12], and results onthese compounds will not be discussedin the present paper. Several reviews havebeen published on the redox tuning of ironporphyrins over the last 20 years [2, 7, 10,192] and this topic will also not be coveredin the present paper. The exact potentialfor the Fe(III)/Fe(II) reaction will dependon the type of axial ligand coordinatedto the Fe(III) or Fe(II) forms of the por-phyrin. Axial ligands such as NO, C6H5

and ClO4− will change drastically the po-

tential at which the Fe(III)/Fe(II) redoxcouple is observed, but shifts of E1/2 forthis redox reaction will occur upon solventbinding to the Fe(III) and/or Fe(II) formof the compound. The basicity of the por-phyrin macrocycle will also influence E1/2

for the Fe(II)/Fe(III) electrode process.

The Fe(III)/Fe(II) reactions are not theonly porphyrin redox processes, whichmay be tuned by a change in the macro-cycle, and half-wave potentials for thering-centered redox reactions will alsovary with the planarity of the macrocy-cle [42]. For example, (Et8TPP)FeCl, whichhas a nonplanar macrocycle [21], is oxi-dized to its Fe(III) π -cation radical form atE1/2 = 0.66 V [21], but the same electrodereactions of (TPP)FeCl and (OEP)FeCl(two porphyrins with planar macrocycles)occur at E1/2 = 1.14 and 1.08 V, respec-tively [21]. The Fe(III)/Fe(II) reaction of[(P)Fe(HIm)2]+X− shifts in potential ow-ing to hydrogen bonding of the boundimidazole, and several studies of these hy-drogen binding interactions as they applyto redox potentials, have been reported inthe literature [286–289].

6.4.4.3.1 Reaction of Heme-thiolate Pro-teins In their resting states, the P450sand NOSs exist as a mixture of a hexacoor-dinated low-spin state Fe(III) species witha water molecule trans- to one proximalcysteinate ligand and a pentacoordinatehigh-spin state species with only cysteinateas an axial ligand [248]. The proteins can bereduced to their iron(II) states and transferelectrons to particular substrates as well asbind various ligands such as CO or O2.The binding of O2 will yield an Fe(II)-O2 species that is a very poor oxidizingagent, except toward particularly reactivesubstrates. The next step in the catalyticactivation of the substrate by the P450sor NOS involves a one-electron reductionof the Fe(II)-O2 species and the forma-tion of several intermediates that are lesswell characterized [249]. The first productis a ferric peroxo complex of the typeFe(II)-O−O− that can be protonated toyield Fe(III)-O−OH. The last intermediatein the cycle is obtained via a heterolytic


cleavage of the O−O bond after protona-tion of the terminal O(H) atom and a lossof H2O [250–252]. This high valent iron-oxo porphyrin has been described as anFe(IV)=O porphyrin radical, an Fe(V)=Ospecies or a Fe(IV)-O

•species and should

react with electron-rich centers of the sub-strate. All of the reactive states from Fe(II)to the high valent oxo species are involvedin the catalytic cycle of substrate monooxy-genation by the P450s and NOSs; theirfunction in biological systems has recentlybeen reviewed [248].

Heme-thiolate proteins have been re-ported to catalyze a large number of reac-tions, including diverse types of oxidations,reductions, isomerizations and dehydra-tions [248]. The nature and mechanismof these reactions have been discussed indetail [248], and the reactions have beenshown to be associated with various reac-tive states under which the heme-thiolateprotein can exist. The nature of the ironporphyrin intermediate in the protein thatis involved in the activation of a substratewill depend on the reactivity of the sub-strate. It has recently been shown [253,254] that in the mutant P450s, the ratiobetween different reactions that occur onthe same substrate will be a function ofthe rates of proton delivery to the activesite of P450, suggesting that the presenceof distal amino acid residues also affectsthe rate of transformation of Fe(II)-O2 toFe(III)-OO− and Fe(V)=O.

6.4.4.3.2 Mechanism of Electron-transferin Peroxidases Peroxidases have beendivided into three categories [255]. Theseare Class I, the intercellular peroxidases,Class II, the extracellular peroxidases andClass III, the well-known extracellularplant peroxidases. The peroxidases in ClassI, II, and III differ in the number ofdisulfide bonds and bound calcium ions,

but all of the peroxidases use the catalyticcycle described in Sch. 7 [256].

The first step of the cycle [Sch. 7, (1)]involves the removal of two electronsfrom the protein, one coming from themetal ion and the other usually com-ing from the porphyrin moiety, whichcontains the iron [257]. The intermediatethus generated is known as Compound I.This species that contains two oxidizingequivalents of the protein can oxidizetwo substrate molecules. The first step,which is described in Eq. (2) of Sch. 7,involves formation of Compound II viareduction of the porphyrin π -cation rad-ical, while the second step leads to theresting state of the protein via a reac-tion between Compound II and a secondsubstrate molecule [Sch. 7, (3)]. Both Com-pounds I and II are sufficiently stable sothat several spectral methods have beenused to determine the electronic struc-ture of these species [258–262]. Kineticstudies for formation of these compoundshave also been carried out [263–269]. Therate-limiting step under steady state con-ditions is most often the reduction ofCompound II [266–271], a result that isaccounted for by an easier reduction of theporphyrin radical as compared with theFe(IV) center in Compound II. Traditionalsubstrates for peroxidases are phenolsand other aromatic dyes, although in cy-tochrome c peroxidase the substrate isferro-cytochrome c [272, 273].

Electron transfer between the por-phyrin radical and the substrate is read-ily achieved, while the aromatic sub-strate is unable to transfer directly anelectron to the iron because of sterichindrance within the active site pocket,thus slowing down the third electrodereaction described in Sch. 7. Further de-tails on the electron-transfer reactions ofperoxidases, including cytochrome and


Scheme 7 Mechanism ofelectron-transfer in Peroxidases.

FeIV = OR

FeIIIRH2O2

FeIV = OR

+ S

+ S FeIIIR

FeIV = OR

FeIV = OR

Resting state

Compound I

Compound II

+ S

+ S

(1)

(2)

(3)

•

• •

•

their enzyme-substrate complexes, havebeen reported in a recent review [256].

6.4.4.3.3 Reactions of Nitrite ReductionsTwo major pathways have been shownto exist in nitrite reduction [274]. In thefirst pathway, nitrite is reduced to NO,while in the second there is a direct con-version of nitrite to NH3 or NH4

+. Twoclasses of nitrite reductase (NIR), namelythe cytochromes cd [274], and the coppernitrite reductase [274], have been identi-fied for the first pathway and two classesof enzyme, namely the siroheme nitritereductase and cytochrome c nitrite re-ductase, have been proposed to followthe second pathway. The mechanism ofthese four enzymes has been recently re-viewed [274], and only a brief summaryof the electron-transfer reactions of cy-tochrome cd nitrite reductase will be givenhere. The initial step in the conversion ofNO2 to NO involves a binding of the ni-trite ion to the metal of the reduced heme.This first step is followed by the uptakeof two protons and the loss of one watermolecule to yield an electrophilic ferrousFe2+-NO+ species, also formulated as aFe3+-NO• complex. The dissociation of NOfrom this species produces the ferric hemed, which is in turn reduced back to its orig-inal state by heme c. Why the enzyme doesnot reduce the nitrosyl species, FeII-NO+or FeIII-NO to its FeII-NO form, prior todissociation of NO in the heme, has beendiscussed in the literature [274], and may

involve a switch in the nature of the axialligand.

6.4.4.3.4 Reactions of Nitric Oxide Reduc-tases The bacterial nitric oxide reductase(NOR) is able to reduce NO to N2Oduring bacterial denitrification [275–278].A review on the topic has recently beenpublished [274] and only the electron-transfer mechanism of the enzyme isdiscussed in the present review.

NOR has been suggested to contain aheme c and two heme b groups [279–285],so that this type of enzyme has been de-fined as a cytochrome cbbNO complex withthe heme bNO at a binuclear catalytic sitealong with a nonheme iron [276]. It hasbeen proposed [281], by similarities be-tween NOR and the family of cytochromeoxidase, that the catalytic site for NO bind-ing and activation is the dinuclear highspin heme b and the nonheme contain-ing iron (III), and that low-spin hemes areused as electron sinks.

In the first step of the reaction, twoelectrons from two low-spin hemes areprovided to the catalytic site, and thereduction of both the heme b and thenonheme Fe is accompanied by a releaseof a water molecule [274]. In the next step,two molecules of NO and one proton areadded to the active site to yield two FeII-NOcomplexes, one from the heme b and theother from the nonheme. In the last stepof the mechanism, one molecule of N2Ois produced by the active site. The first


step in the catalytic cycle is also proposedto be accompanied by axial ligation of theproximal histidine to the Fe(III) form ofthe heme b, but this ceases when NOcoordinates to the FeII center of heme b.

6.4.5Group 9

Iridium porphyrins have been the least-studied among the Group 9 metallopor-phyrins. Only a few types of iridiumporphyrins have been investigated fortheir electrochemical properties, includ-ing (P)Ir(CO)Cl [102], (P)[Ir(CO)3]2 [103]and [(P)IrCl2]dppe [290], where P =OEP or TPP. Metal-centered and ring-centered processes are observed for(P)[Ir(CO)3]2 [102], while (P)Ir(CO)Cl and[(P)IrCl2]dppe are characterized by onlymacrocycle-centered processes.

A number of studies have been carriedout on the air-stable Rh(III) porphyrinsrepresented by (P)Rh(R), (P)RhX and(P)RhX(L), where R is an alkyl group, X ahalide ion, and L a nitrogeneous base [183,291–301]. A few air-stable monomeric Rhporphyrins of the type (P)Rh(NO) and(P)Rh(NO)2 have also been electrochem-ically characterized [298].

The electrooxidation of Rh(III) por-phyrins is relatively straightforward andconsists for the most part of two suc-cessive one-electron transfers, both ofwhich involve the porphyrin π -ring sys-tem. An assignment of the electron-transfer site has often been made onthe basis of ESR and/or UV-visiblespectra of the singly oxidized prod-ucts. Other types of rhodium porphyrins,which have been electrochemically ex-amined include [(TPP)Rh(L)2]+Cl− and(TPP)Rh(L)Cl, where L = NMe2, both ofwhich were proposed to form a dimericRh(II) species after an initial one-electron

reduction at the Rh(III) center [295]. How-ever, more recent studies by the group ofSaveant [300, 301] suggested that the re-action proceeds via a more complicatedmultielectron transfer and conversion to[(P)RhI]− .

(TPP)Rh(COMe) was shown to undergotwo one-electron reductions, the first ofwhich was proposed to be macrocycle-centered on the basis of UV-visibledata [296]. (TPP)Rh(O2) also exhibits tworeductions with the product depending onthe type of utilized solvent [302]. A Rh(II)porphyrin dimer was electrogenerated inPhCN or THF, but a σ -bonded product,(P)RhIII(CH2Cl), was formed in CH2Cl2.

Cobalt porphyrins have been by farthe most well-studied of the Group 9porphyrins. The electrochemistry of cobaltporphyrins is characterized by both metal-and ring-centered electrode processes. Themost often characterized metal-centeredprocesses have been the Co(III)/Co(II)and Co(II)/Co(I) transitions, which occurat potentials easily accessible in mostnonaqueous electrochemical solvents. Thetwo ring-centered redox processes are onlyobserved for easily reducible complexes(usually those with multiple electron-withdrawing groups on the macrocycle,such as in the case of CN4TPP or Br8TPP)and involve the conversion of a Co(I)porphyrin to its Co(I) π -anion radical formfollowed at more negative potentials by aconversion of the Co(I) π -anion radical toits dianionic form. Oxidation of the Co(III)complexes will give a Co(III) porphyrin π -cation radical and then a dication at morepositive potentials. A summary of theseelectrode processes is shown in Sch. 8.

As shown in Sch. 8, the oxidation ofcobalt(II) porphyrins can also proceedvia the initial formation of a Co(II) π -cation radical (as opposed to a Co(III)species), and this has been shown to


[(P)CoIII]+ [(P)CoII]+

−e−

[(P)CoIV]2+ [(P)CoIII]2+•

−e−

[(P)CoIII]3+

−e−−e−

−e− −e− e− e−

[(P)CoII]−•

[(P)CoI]−

e−e−

[(P)CoI]2−•

e−

[(P)CoI]3−

(P)CoII

Scheme 8 Electron-transfer reactions of cobalt(II) porphyrins.

occur for (TPP)Co in dry noncoordinatingsolvents. An oxidation of the sigma-bonded(P)CoIII(R) complexes to their Co(IV) formhas also been proposed for derivativescontaining an alkyl axial ligand [303].

Cobalt porphyrins with several types ofmacrocycle have been electrochemicallyinvestigated, but the most often andmost well-studied of the cobalt porphyrinshas been (TPP)Co, which was initiallycharacterized in the mid-1960s [51, 304]and has since been cited and reexaminedin literally hundreds of publications overthe next three decades [7, 21]. A largenumber of electrochemical studies havefocused on the effect of solvent, axialligation, and supporting electrolyte orcounterion on the redox reactions of (P)Coand (P)CoCl. Sch. 9 summarizes severalpossible electron-transfer pathways, whichcan occur for the Co(II)/Co(III) transition.

Cobalt(II) porphyrins can be tetra-,penta- or hexacoordinated, while theCo(III) complexes are generally 5- or 6-coordinate, depending on the solutionconditions. Basolo and coworkers showedthat the potential for oxidation for a Co(II)porphyrin or related macrocycle could berelated linearly to the base strength of thebound axial ligand [305, 306], and a similarrelationship has been observed between

E1/2 for the Co(II)/Co(III) reaction of thecomplex and the coordinating ability ofthe aprotic solvent in which the electro-chemistry was carried out [2, 307, 308].Oxidation potentials of Co(II) porphyrinsalso depend on the type and concentrationof the supporting electrolyte, with the mostdifficult oxidations occurring in solutionsof TBABF4 and the easiest in solutions ofTBACl [307, 309, 310].

(TPP)Co in a low dielectric constantsuch as toluene or benzene [311] under-goes two one-electron reductions and onlytwo oxidations are seen as opposed tothree in most other solvents. The first ox-idation under these conditions involvesthe simultaneous abstraction of two elec-trons from (TPP)Co to give [(TPP)Co]2+,while the first reduction leads to a stableCo(I) complex formulated as [(TPP)Co]−.Other studies relating to the effect ofcounterion and solvent binding on the(P)CoX electrochemistry [310] have beendiscussed in a recent review [7].

The type of porphyrin macrocycle will in-fluence both the half-wave potentials andthe overall redox behavior of cobalt por-phyrins [34, 55, 308, 312–314]. Porphyrinswith macrocycles, which vary substantiallyin basicity and planarity have been studied,including T(p-X)PP, (CN)xTPP, BrxTPP,


CoIIL

Co

L

IIL

Co

L

II

L

CoIIIL

Co

L

IIIL

Co

L

III

L

+

+ +

e− e− e−e− e−

e−

Scheme 9 Electrooxidation ofcobalt(II) porphyrins incoordinating media.

TMpyP, and TMP (see Table 3 for macrocy-cle abbreviations and Fig. 3 for structuresof selected macrocycles).

The Co(III)/Co(II) and Co(II)/Co(I) re-actions of (T(p-X)PP)Co were investigatedin butyronitrile, and plots of E1/2 for thesetwo electrode reactions versus 4σ (the sub-stituent constant) gave linear relationshipswith slopes of 0.034 and 0.041 V, re-spectively [308]. The electron-withdrawingeffect of substituents on E1/2 was alsoevaluated for ((CN)xTPP)Co [312, 313],where x = 1 to 4 and the largest effectwas seen for ((CN)4TPP)Co, where thefirst reduction was 0.55 V more positivethan the analogous metal-centered reduc-tion of (TPP)Co [313]. The electrochemicaland spectroelectrochemical properties of((CN)4TPP)Co were investigated in severaldifferent nonaqueous solvents [313], andthree well-defined one-electron reductionswere generally observed within the po-tential range of each solvent. The first

one-electron reduction was proposed tooccur at the porphyrin π -ring system [313](as opposed to the metal center that oc-curs for the case of (TPP)Co), while thesecond and third one-electron reductionswere proposed to involve the stepwiseformation of a Co(I) π -anion radical fol-lowed by a Co(I) dianion at more negativepotentials.

(BrxTPP)Co complexes with x = 0 to8 were shown to undergo three one-electron oxidations and up to nine one-electron reductions in PhCN containingTBAP as supporting electrolyte [34]. Theconversion of the compounds to theircobalt(I) forms was reversible, but thesecond electron addition, which shouldproduce the cobalt(I) π -anion radical formof these compounds, was shown to becoupled with a stepwise elimination of Brgroups as shown in Sch. 10.

Araullo-McAdams and Kadish investi-gated the electrochemistry of [(TMpyP)

e−(BrxTPP)CoII [(BrxTPP)CoI]− [(BrxTPP)CoI]2−e−

[(Brx-1TPP)CoI]−

[(TPP)CoI]−

(x-1)e−

−(x-1)Br−+(x-1)H

−e−

−Br

+H

(TPP)CoII

Scheme 10 Electroreduction of (BrxTPP)CoII.


CoII]4+ (see Fig. 3 for structure of theTMpyP macrocycle) in DMF, DMSO, andPy [55]. A single oxidation and up tofour reductions were observed; some re-actions were shown to be metal-centeredand involved only a single electron, whileothers were centered at the porphyrinπ -conjugated system or the four 1-methyl-4-pyridyl substituents and involved twoelectrons.

One important characteristic of cobaltporphyrins is their ability to bind or re-act with small molecules, such as NO [27,67, 70, 91, 93, 100], CO [36, 114, 115],O2 [314–320], or CO2 [321], and severalstudies have focused on the chemicaland/or electrochemical reactivity of (P)Cotoward these small molecules. The in-teraction of cobalt porphyrins with NOand the electrochemical properties ofthe resulting cobalt-nitrosyl porphyrinshave been investigated by several researchgroups [7]. (TPP)Co(NO) exhibits two ox-idations and three reductions at a micro-electrode in CH2Cl2 [90]. The NO groupremains coordinated after electrooxidationand the initial electron abstraction from(TPP)Co(NO) was proposed to involve theporphyrin π -ring system. Other electrodereactions were accompanied by a disso-ciation of NO from the compound andthe site of electron transfer could not bedetermined.

(OEP)Co(NO) was shown by Fajerand coworkers to be oxidized to its π -cation radical form without a loss ofNO [67] and the oxidized product wasassigned as an a2u type radical on thebasis of IR spectroscopy [70]. The elec-trochemistry of (P)Co(NO) has been re-ported in several coordinating mediaand these results have been discussedin a recent review [7]. Derivatives of(T(p-X)PP)Co(NO) have also been elec-trochemically investigated [70, 93] and an

electron-transfer mechanism similar tothat of (TPP)Co(NO) [91] or (OEP)Co(NO)was observed. (TPPBr4NO2)Co(NO) wasalso electrochemically investigated [27]and showed what was proposed to be atransient Co(I)-nitrosyl complex.

The electrochemistry of cobalt por-phyrins has been examined under bothCO [36, 114, 115] and O2 atmospheres. Thestudies under O2 were to examine the ap-plication of cobalt porphyrins as electrocat-alysts for the four-electron reduction of O2to H2O [314, 316, 317, 322]. The oxidationof (P)Co under a CO atmosphere has beenreported for P = TPP [115], OEP [114] andBrxTPP (x = 0 to 8) [29]. (OEP)Co is con-verted to [(OEP)CoIII(CO)]+ in CH2Cl2under CO [114] and these results contrastwith what was observed for oxidation of thesame compound under N2, where a Co(II)porphyrin π -cation radical was proposed asthe singly oxidized product. The electroox-idation of (BrxTPP)Co under CO leads toboth mono and bis-CO complexes in so-lution with the ratio of mono/bis adductsdepending on the number of Br groups onthe porphyrin macrocycle [29].

6.4.6Group 10

Very little electrochemistry has been pub-lished on Pt or Pd porphyrins. In con-trast, the electrochemistry of nickel por-phyrins with numerous different macro-cycles has been reported in the literatureand the most commonly observed electron-transfer mechanisms that occur uponreduction or oxidation of the Ni(II) deriva-tives are shown in Sch. 11.

The air-stable form of the neutralporphyrins contain Ni(II). In some cases,derivatives with Ni(I) or Ni(III) centralmetal ions can be electrogenerated, butin others only Ni(II) π -cation radicals and


[(P)NiII]+ [(P)NiI]−

[(P)NiII]2+

[(P)NiII]−

[(P)NiIII]3+

[(P)NiI]2−[(P)NiIII]2+ [(P)NiII]2−

[(P)NiI]3−[(P)NiIV]3+

••

•

•

•

(P)NiII

[(P)NiIII]+

−e−

−e−

−e−

?

e− e−

e−

−e− −e− e− e−

Scheme 11 Electron-transfer reactions of nickel(II) porphyrins.

dications or Ni(II) π -anion radicals anddianions are observed as electron-transferproducts [116, 323–326]. The exact siteof electron transfer will depend in parton the specific porphyrin macrocycle,in part on the presence or absence ofany coordinated axial ligands, in parton the properties of the solvent and inpart on the type of supporting electrolyteas well as on the temperature of themeasurement [7].

Some nickel(II) porphyrins are con-verted to a Ni(II) porphyrin π -dicationin two well-separated one-electron-transfersteps [7], while other Ni(II) porphyrinsexhibit two, closely spaced and often over-lapped one-electron-transfer steps uponoxidation [7]. The occurrence or absenceof overlapped one-electron oxidations willdepend on both the solution conditionsand the type of porphyrin macrocycle, butin most cases the final oxidation prod-uct can be described as a [(P)NiIII]3+species.

Nickel porphyrins with Me8TPP,TC6TPP, Et8TPP or DPP macrocyles(see macrocycle abbreviations in Table 2)have all been examined for theirelectrochemical behavior [7, 21] and, in

agreement with theoretical calculations,the nonplanar nickel porphyrins areusually easier to oxidize than thecomplexes with a planar macrocycle [42].The electrochemistry of [(TMpyP)Ni]4+in DMF differs significantly from theother nickel(II) porphyrins, in that boththe neutral and reduced forms of thecompound have been shown to existin a monomer-dimer equilibrium [327].The two electron oxidation product of(TpivPP)Ni was proposed to contain aNi(IV) π -cation radical [328], but a laterstudy by Saveant and coworkers suggestedan oxidation of the picket fence [234].

Both low and high oxidation states ofnickel porphyrins are easily accessiblein many nonaqueous solvents and theexact site of electron transfer (metal vsring) upon reduction or oxidation hasbeen the topic of numerous studies [7].Bocian and coworkers [325], as well asConnich and Macor [326], showed that aNi(II) porphyrin π -cation radical can beconverted to a Ni(III) porphyrin in bindingsolvents such as Py, THF or MeCN. Thetype of Ni(II) porphyrin π -cation radical,that is, a1u or a2u, will depend on the type ofporphyrin macrocycle and a Ni(II) π -cation


radical with a mixed a1u and a2u characterhas been described in the literature [329].

The electroreduction of Ni(II) por-phyrins and related macrocycles wasinvestigated by the groups of Stolzen-berg [330–332], Fajer [333], Kadish [116,323, 328, 334, 335], Saveant [324] andmore recently, by Fajer, Mansuy andcoworkers [336], who showed that the veryelectron-deficient ((NO2)7T(oxo-Cl2)PP)Niwas reduced in three reversible one-electron reductions. Earlier work hadshown that Ni(II) chlorins or isobacteri-ochlorins could be reduced to their Ni(I)state [330, 331, 333], as opposed to por-phyrins, where a Ni(II) porphyrin π -anionradicals was generally formed upon re-duction [7]. However, the singly reducedproducts of Ni(II) porphyrins have beendescribed as Ni(I) porphyrins, Ni(II) por-phyrin π -anion radicals or Ni(II) porphyrinπ -anion radicals with some Ni(I) charac-ter [116]. The exact site of electron transferand formulation of the reduction prod-uct will depend on several factors, two ofwhich include the solution conditions andthe nature of the porphyrin macrocycle [7].

Several fluorine-substituted (DPP)Niderivatives (see Fig. 9) have also beeninvestigated as to their electrochemicalproperties in nonaqueous solvents and asummary of the data has been discussed ina recent review that compared the electro-chemical behavior of planar and nonplanarNi(II) porphyrins [7].

Nickel porphyrins with nonplanar ma-crocycles had a smaller to slightly smallerHOMO-LUMO gap than nickel porphyrinswith planar macrocycles (such as TPP), butthe difference in HOMO-LUMO gap ongoing from a planar to a nonplanar macro-cycle varies with the type of substituentson the nonplanar macrocycle. For example,the HOMO-LUMO gap of (Et8F20TPP)Niis 60 mV smaller than that of (TPP)Ni, but

the HOMO-LUMO gap of (Br8F20TPP)Niis 410 mV smaller than that of (TPP)Ni [7].The absolute potential separation betweenthe first and second oxidations of the sub-stituted (DPP)Ni derivatives increases onchanging the supporting electrolyte fromTBAP to TBAPF6 [7]. This result can be ac-counted for by a stronger axial binding ofClO4

− over PF6− to [(P)Ni]+, along with a

larger binding constant of ClO4− to doubly

oxidized [(P)Ni]2+ as compared with singlyoxidized [(P)Ni]+.

6.4.7Group 11

Copper and silver porphyrins exist asstable M(II) complexes, while gold por-phyrins contain a M(III) metal oxida-tion state in their air-stable form. TheAg(II) porphyrins can be oxidized totheir Ag(III) form before formation ofa π -cation radical and dication at morepositive potentials [337, 338]. Reduction ofthese compounds involves formation of aAg(I) porphyrin, followed by electrogen-eration of a π -anion radical and dianionat more negative potentials [337]. How-ever, the latter two ring-centered processesare most often not observed because ofdemetalation of the electrogenerated Ag(I)porphyrin [339–341].

The electrochemistry of (P)CuII in-volves macrocycle-centered electrode re-actions in nonaqueous media, althougha Cu(II)/Cu(I) reaction has been pro-posed for a few complexes, most notably((CN)4TPP)Cu [342] and (N-CH3TPP)CuCl[343]. The electrochemical generationof a Cu(I) porphyrin in the case of((CN)4TPP)Cu occurs only after formationof the dianionic form of the complex [342].

Several Cu(II) porphyrins with non-planar macrocycles have been examinedfor their electrochemical properties, one


(F28

DP

P)M

NN

NN

FF

FF

FF

F F

MF

F

F

F

NN

NN

F

FF

F F

FF

F

F

F

F

FM

(F12

DP

P)M

(F4D

PP

)M

NN

NN

F

FF

FM

NN

NN

F

FF

F F

FF

F

M

(F8D

PP

)M

NN

NN

M

(F20

DP

P)M

F5

F5

F5

F5

NN

NN F

F

FF

FF

FFM

(F8D

PP

)M (

mes

o)

NN

NN

OM

e

OM

e

OM

e

MeO

MeO

OM

eO

Me

OM

e

OM

e

OM

e

OM

e

OM

eM

eOM

eO

MeO MeO

MeO

MeO

MeO

OM

e

M

[(O

Me)

20D

PP

)M

M

FF

F

F

F

F

FF

FF

F F

FF

F

NN

NN

F

F

F

F

F

(F36

DP

P)M

Fig.

9St

ruct

ures

offlu

orin

ated

subs

titut

edD

PP.


of which is (Et8TPP)Cu, which exhibitstwo reversible one-electron oxidations innonaqueous media [344, 345]; the singlyoxidized product, [(Et8TPP)Cu]+, wasformulated as an a2u porphyrin π -cationradical on the basis of its IR spectrum [344]and this formulation was also confirmedby resonance Raman studies [345]. An-other nonplanar porphyrin, (DPP)Cu, wasshown to undergo two successive ring-centered one-electron reductions and twosuccessive ring-centered one-electron oxi-dations [346]. Both oxidations of the com-pound were easier (occurred at morenegative potentials) than the respective ox-idations of (TPP)Cu and this result wasaccounted for by a destabilization of theHOMO owing to the nonplanarity of themacrocycle.

The electrochemistry of gold porphyrinshas been limited to [(TPP)Au]+AuCl4−,MPIX dimethylester, and Etio Au(II)complexes [347a]. The compounds wereoxidized via a single one-electron-transferstep to their π -cation radical form andwere reduced by three steps, the first ofwhich was proposed to generate an Au(III)π -anion radical. However, more recentstudies by the groups of Kadish, Fukuzumiand Crossley show the unambiguousformation of a Au(II) porphyrin in thefirst electron addition [347b].

6.4.8Group 12

The electrochemistry of zinc, cadmium,and mercury porphyrins is uncomplicatedand only macrocycle-centered electrodereactions are observed in nonaqueous sol-vents [7, 21]. Metalloporphyrins with thistype of metal ion have been studied insome detail, but the best characterizationof the electrochemical reaction productshas been obtained with zinc porphyrins

whose site of electron transfer is very welldefined [2, 7]. In fact, (P)Zn complexeshave been used as standard for comparisonwith other oxidized or reduced metallopor-phyrins, especially the iron(III) porphyrinsthat can undergo either metal- or ring-centered electrode processes [7].

Most electrochemical studies on zincporphyrins have involved planar com-pounds such as (OEP)Zn or (TPP)Zn, butsome electrochemical data on nonplanarzinc porphyrins such as (X8F20TPP)Zn,where X = F [32], Br [44], Cl [44] or Me [44]have also been published. The firstand second oxidations of (TPP)Zn and(OEP)Zn are well separated in potentialsin most nonaqueous solvents, and both[(TPP)Zn]+ and [(OEP)Zn]+ have been ex-amined by ESR spectroscopy [348–350].A dimerization of [(OEP)Zn]+ was pro-posed [348, 349], but no evidence of dimerformation was found in the case of[(TPP)Zn]+ or other [(P)Zn]+ complexes.

The oxidative behavior of (TPP)Zn hasbeen examined in the presence of nitroge-neous bases by Kadish and coworkers [351,352], and an electrochemical study of(TPP)Zn and [(TPP)Zn(L)]+ yielded thefirst quantitative measurements of nitroge-neous base addition to porphyrin π -cationradicals.

Only a few nonplanar zinc porphyrinshave been electrochemically exami-ned [7, 21]. The electrochemical behav-ior of these compounds resemble thatof the planar compounds in that tworing-centered one-electron reductions andtwo ring-centered one-electron oxidationsare obtained. However, in the case of(X8F20TPP)Zn, (X = Cl, Br, Me) [44] thetwo oxidations of the nonplanar com-pounds are overlapped in potential givingwhat appears to be a single two-electron-transfer process.


6.4.9Group 13

The first electrochemistry of metallopor-phyrins with a Group 13 metal ion involvedOEP derivatives of the type (OEP)M(OH),where M = Al, Ga, In or Tl [52]. Thesestudies were later followed by an elec-trochemical characterization of (P)MCl,(P)MClO4, (P)Tl(O2CCF3), and (P)M(R),where P = OEP or TPP [7, 353–356].

Four classes of Group 13 metallopor-phyrins have thus far been character-ized for their electrochemical propertiesin nonaqueous media. These include(1) compounds with anionic axial ligands,represented as (P)MX, where X is usu-ally a halide or ClO4

−, (2) compoundswith carbon sigma-bonded ligands, repre-sented as (P)M(R), where R is an alkylor an aryl group [12], (3) metal–metalbonded complexes of the type (P)MM′(L),where M is the metal ion within theporphyrin and M′(L) the metalate ax-ial ligand represented in some cases byRe(CO)5, Mn(CO)5 or Fe(CO)4 and in oth-ers by M′(CO)3Cp where M′ = Cr, Mo orW [13], (4) compounds with a tetrazolatoor a triazolato axial ligand represented as(P)In(N4CR) and (P)In(N3CR) [357, 358].

The porphyrins in groups (1) and (2)above have been characterized with Al(III),Ga(III), In(III), and Tl(III) metal ions,while those in groups (3) are limitedto compounds with In(III) and Tl(III)metal ions. The triazolato and tetrazolato

compounds have been mainly those withIn(III) metal ions [358, 359]. The electro-chemistry of metal–metal bonded [13] andGroup 13 sigma-bonded [12] porphyrinshas been discussed in a recent review andthe present chapter will only summarizethe most important electrochemical behav-ior of the compounds.

The Group 13 porphyrins usuallyshow similar electrochemical behaviorupon going down the Periodic Table,although differences do exist in thestability of the oxidized and reducedforms of the compounds upon goingfrom Al(III) to Tl(III). All complexes ofthe type (P)MX, where M = Al(III) [360],Ga(III) [355], In(III) [356], or Tl(III) [353],and P = TPP or OEP are oxidized at theporphyrin macrocycle to give π -cation rad-icals and dications, and all but those inthe thallium series [353] are converted totheir π -anion radicals and dianions uponreduction. The Tl(III) complexes undergowhat appears in many cases to be a metal-centered two-electron reduction prior todemetallation [339].

The σ -bonded porphyrins of the type(P)M(R) are characterized by two reversiblemacrocycle-centered reductions for com-pounds with M = Al, Ga, In, or Tl centralions, but as shown in Sch. 12, the oxidationpathway and stability of the electrogen-erated species will depend on both thespecific metal ion and the type of σ -bonded axial ligand. Further details forthe electrochemistry of these compounds

−e−

(P)M(R) [(P)M(R)]+•

[(P)M(R)]2+−e−

[(P)M]+

R•

R•

[(P)M]2+−e− −e−

[(P)M]3+

Scheme 12 Electrooxidation of (P)M(R), where R is an alkyl or arylgroup and M is an element from Group 13 of the Periodic Table.


and other types of porphyrins with Group13 metal ions have been discussed in arecent review [12].

6.4.10Group 14

Most electrochemical data on Group 14metalloporphyrins involves compoundswith Ge(IV) [58, 64, 361, 362], Sn(IV)[363–365] or Sn(II) [366] metal ions andonly a few examples of Si(IV) [59] andPb(II) [367] porphyrin electrochemistryhave been reported.

The characterized Si complexes arerepresented as (P)Si(R)2, (P)Si(R)X, and(P)SiX2, where P = TPP, OEP or T(p-Me)PP, R = Ph or Me and X = OH−,ClO4

−, F−, Cl− or O3SCF3−, while

the electrochemistry of Pb porphyrinshas most often been described for(OEP)PbII [52, 367] and (TPP)PbII [52, 59,368, 369].

The electroreduction of Si porphyrins in-volves reactions at the π -ring system [51,304]. Si porphyrins with an OEP macro-cycle have been shown to undergo oneor more coupled chemical reactions fol-lowing electrooxidation [59], while Si por-phyrins with a TPP or substituted TPPmacrocycle exhibit two reversible one-electron oxidations centered at the con-jugated macrocycle [368].

A number of Ge(IV), Sn(II) and Sn(IV)porphyrins has been investigated for theirelectrochemical properties. The Ge por-phyrins are represented as (P)Ge(R)2and (P)Ge(R)X [58], where P = TPP orOEP, R = Me, CH2C6H5 or C6H5 andX = Cl−, OH− or ClO4

−. The Sn(II)porphyrins [365, 366] are represented as(P)Sn, where P = OEP, T(p-Me)PP), T(m-Me)PP or TMP, while the Sn(IV) por-phyrins are represented as (P)Sn(L), whereP = TPP, T(m-Me)PP, T(p-Me)PP or OEP

and L = S2− or Se2− or (P)SnX2, whereP = T(p-Me)PP or T(m-Me)PP and L =ClO4

−, Br−, Cl−, F− or OH−. A fewSn(IV) porphyrins with carbon σ -bondedaxial ligands have also been examined fortheir electrochemistry [365] and these areof the type (P)SnMeI, where P = OEP,T(p-Me)PP, T(m-Me)PP or TMP.

The electrochemistry of (P)Ge(R)2 and(P)Ge(R)X are characterized by one ortwo reversible single-electron additionsleading to porphyrin π -anion radicals anddianions. The compounds also undergotwo oxidations, the first of which isalways reversible in the case of the(P)Ge(R)2 complexes [58]. A rapid cleavageof the Ge-carbon bond was proposedto follow electron transfer and the finalproduct of the oxidation was formulated as(P)Ge(R)(ClO4) [58].

The mechanism for electroreduction of(P)GeX2 was shown to depend on thetype of bound anion [362]. (P)Ge(OH)2 isreduced in two successive reversible one-electron transfers involving the porphyrinπ -ring system and a similar site of electrontransfer was proposed for the electroreduc-tion of (P)Ge(ClO4)2; the reduced ClO4

−complex is unstable and reacts with waterin the solvent or water molecules boundto the compound to give (P)Ge(OH)(ClO4)during the timescale of the measurement.

(P)SnII can be converted electrochem-ically to [(P)SnIV]2+, but the reduction ofSn(IV) to Sn(II) has not been observed, andthe overall electrochemical behavior of themonomeric species [(P)SnIV]2+ occurs asshown in Sch. 13 [366].

The reduction of Sn(IV) porphyrinsoccur at the π -ring system and are usu-ally reversible, while the oxidation usuallyinvolve the axial ligand and are irreversibleon the cyclic voltammetry timescale.For example, (P)SnS [363], (P)SnSe [363]and (P)SnFe(CO)4 [370, 371], where P =


e−

[(P)SnIV]2+

[(P)SnIV]3+ •[(P)SnIV]+

•

(P)SnIV

(P)SnII

•[(P)SnII]−

e−

e−

[(P)SnII]2−

−e− e− 2e−

Scheme 13 Electron-transferreactions of tin(IV) porphyrins.

TPP, T(m-Me)PP, T(p-Me)PP or OEPall undergo two reversible one-electronreductions centered at the porphyrinmacrocycle, but only a single irreversibleoxidation [363, 372]. In each case, the irre-versibility of the oxidation was accountedfor by a cleavage of the tin-axial ligandbond following electron transfer.

The electrochemical properties of Sn(II)porphyrins in coordinating media havebeen discussed in a recent review [7]. Thecompounds were proposed to undergotwo reversible one-electron macrocycle-centered reductions and an irreversibletransfer of two electrons upon oxidation.The oxidation yielded ultimately a productformulated as [(P)SnIV(X)S]+, where S =py, THF or PhCN and X = PF6

− or ClO4−.

6.4.11Group 15

Most examined Group 15 porphyrinscontain a M(V) ion, but derivatives witha +3 metal ion are also known andconversion between the M(V) and M(III)form of the porphyrin via a stable ortransient M(IV) species has been reportedin the literature [373–375]. The P(V),As(V), and Sb(V) porphyrins can be dividedinto three groups according to the nature oftheir axial ligands [7]. One group contains

two σ -bonded axial ligands (such as CH3 orC6H5), another contains two anionic axialligands (such as Cl−, OH− or OMe−),and the third contains one σ -bonded axialligand and one anionic axial ligand. TheM(V) porphyrins contain an overall chargeof +1 in their [(P)M(R)2]+, [(P)M(R)(X)]+or [(P)M(X)2]+ oxidation states, and thuseach complex must be associated withanother anion in the solid state or insolution [7].

The majority of electrochemical data forGroup 15 porphyrins has been for com-pounds with a phosphorus [376–380] orantimony [373] ion. The electrochemistryof arsenic [381] and bismuth [382] por-phyrins is limited to [(OEP)As(R)(R′)]+,[(OEP)As(R)(X)]+ and [(OEP)As(X)(X′)]+(R and R′ are sigma-bonded ligands, X andX′ are anionic axial ligands) in the arsenicseries and (P)Bi(X), where P = TPP TMP,T(p-Me)PP or OEP and X = NO3

− orSO3CF3

− for the bismuth porphyrins.The As(V) complexes undergo two oxi-

dations and two reductions, all of whichinitially occur at the porphyrin π -ring sys-tem [7]. An additional electrode processis seen for [(OEP)As(F)2]+PF6

−, whichis associated with reduction of an As(III)porphyrin species formed in solution aftergeneration of the As(V) porphyrin dianion.


The bismuth porphyrins studied by thegroups of Kadish and Guilard undergotwo macrocycle-centered oxidations, andunlike what is observed for the otherGroup 15 porphyrins, none of the bismuthporphyrins shows evidence for conversionfrom the M(III) to the M(V) form of thecomplex. The authors pointed out that theoxidation potentials of (P)Bi(SO3CF3) wererather insensitive to the type of porphyrinmacrocycle.

The electrochemically examined phos-phorus porphyrins are represented by[(P)PV(L)2]+X−, where P = TPP, OEPor T(p-Me)PP; L = OH−, OR− or Cl−and X− = Cl− or OH−, while the anti-mony porphyrins, which have been studiedfor their electrochemical properties are(OEP)Sb and (OEP)SbClO4 for derivativeswith a Sb(III) ion and are represented by(P)Sb(X)(X′), (P)Sb(R)(R′) and (P)Sb(R)(X),where P = TPP, OEP or T(p-Me)PP, R orR′ is a sigma-bonded ligand such as CH3

and X or X′ is an anionic axial ligand suchas OH−, OMe−, Cl− or F− for those por-phyrins that contain a Sb(V) central ion.

Earlier electrochemical investigations of[(TPP)P(Cl)2]+Cl− showed that the com-pound could be reduced in two one-electron-transfer steps, both of whichoccur at the porphyrin π -ring system.The reduced products were relatively un-stable and a mechanism involving aninternal electron transfer between theπ -ring system and the central ion wassuggested to account for the electrochem-ical behavior of this compound as well

as those of the type (P)P(X)2, whereX = OMe− or Cl−. The electrochemistryof [(P)P(OR)2]+OH−, where OR− is aphenoxide or substituted phenoxide an-ion, was characterized by two successiveone-electron-transfer steps centered at theporphyrin macrocycle [383].

Antimony porphyrins have been pre-pared with central ions in both +5 and+3 oxidation states, but most electro-chemical data on these compounds havebeen obtained for derivatives containingSb(V). The Sb(V) porphyrins are all ex-tremely easy to reduce with the firstreduction occurring at E1/2 = −0.26 Vfor [(TPP)Sb(Cl)2]+ and at −0.42 V for[(TPP)Sb(OMe)2]+. The mechanism forelectroreduction of Sb(V) porphyrins hasbeen discussed [7] for compounds of thetype [(T(p-Me)PP)Sb(X)2]+Cl−, where X =OMe− or Cl−. Each compound undergoestwo one-electron reductions on the cyclicvoltammetry timescale to give porphyrinπ -anion radicals and dianions, but the sta-bility of the electroreduced products varieswith the nature of the axial ligand [373,374]. The compounds with X = OMe−were relatively stable after reduction, butthe compounds with X = Cl− were shownto undergo a series of chemical reactionsfollowing electron transfer. The ultimatestable porphyrin product of the reduc-tion was proposed to contain a Sb(III)ion. The electroreduction mechanism of[(P)Sb(X)2]+ is illustrated in Sch. 14.

As shown in this scheme, the internalcharge transfer from the porphyrin ring to

e− e−

−2X Internal electron transfer

e− e−−2X

e−

(P)SbVX2

(P)SbIII

[(P)SbVX2]+ [(P)SbVX2]−

[(P)SbIII]+ [(P)SbIII]−X = Cl− or OMe−

Scheme 14 Electroreduction of [(P)SbVX2]+ where X = Cl− or OMe−.


the central Sb ion occurs after reductionof the compounds by one or two electronsand is accompanied in each case by a lossof the two axial ligands.

A few antimony(V) porphyrins withone of two sigma-bonded ligands (CH3or C2H5) have been electrochemicallyinvestigated. The compounds are usuallyextremely stable and are characterized byfacile reductions coupled with chemicalreactions involving a loss of the axialligands. The [(P)Sb(Me)(X)]+ complexesundergo one reversible oxidation withinthe positive potential limit of the solvent,and E1/2 for these reactions were foundto be directly related to the strength of theaxial ligand trans- to the Me group [373].

6.5Concluding Statement

The present review describes the electro-chemistry of synthetic metalloporphyrinsin nonaqueous media. This work does notinclude metalloporphyrin electrochem-istry in aqueous media and discussesonly briefly the electrochemistry of someiron porphyrins in real biological systems.We have not included the electrochem-istry of porphyrin-like complexes such asmetallocorroles, metalloporphycenes andmetallochlorins, or the electrochemistry ofpigments of biological relevance, such asvitamin B12 derivatives. For a coverage ofthese topics, the reader is referred to recentmajor reviews in The Porphyrin Handbook.

Acknowledgments

K. M. K. acknowledges support of theRobert A. Welch foundation for continu-ous support of his research on porphyrinelectrochemistry over the last 25 years un-der Grant E-680.

References

1. K. M. Kadish, R. Smith, Guilard, The Por-phyrin Handbook, Academic Press, SanDiego, 2000, Vol. 1–10.

2. K. M. Kadish, Prog. Inorg. Chem. 1986, 34,435–605.

3. K. M. Kadish in Redox Mechanisms andInterfacial Properties of Molecules of BiologicalImportance, (Eds.: G. Dryhurst, K. Niki),Plenum Press, New York, NY, 1988,pp. 27–45.

4. K. M. Kadish, R. Guilard, Chem. Rev. 1988,88, 1121–1146.

5. J. W. Buchler, W. Kokisch, P. D. Smith,Struct. Bonding 1978, 34, 79.

6. D. G. Davis in The Porphyrins, (Ed.: D. Dol-phin), Academic Press, New York, 1978,Vol. 5.

7. K. M. Kadish, E. Van Caemelbecke,G. Royal, Electrochemistry of Metallopor-phyrins in Nonaqueous Media, in ThePorphyrin Handbook, (Eds.: K. M. Kadish,K. Smith, R. Guilard), Academic Press, SanDiego, 2000, pp. 1–114, Vol. 8.

8. J.-H. Fuhrhop in Porphyrins and Metallopor-phyrins, (Ed.: K. Smith), Elsevier, New York,1975.

9. R. H. Felton in The Porphyrins, (Ed.:D. Dolphin), Academic Press, New York,1978, Vol. 3.

10. K. M. Kadish in Iron Porphyrins, (Eds.:A. B. P. Lever, H. B. Gray), Addison Wes-ley, Reading, 1983, pp. 161–249, Vol. 2.

11. R. Guilard, C. Lecomte, K. M. Kadish, Struct.Bonding 1987, 64, 205–268.

12. R. Guilard, E. Van Caemelbecke, A. Tabard,K. M. Kadish, Synthesis, Spectroscopy andElectrochemical Properties of Porphyrinswith Metal-Carbon Bonds, in The PorphyrinHandbook, (Eds.: K. M. Kadish, K. Smith,R. Guilard), Academic Press, San Diego,2000, pp. 295–345, Vol. 3.

13. J.-M. Barbe, R. Guilard, Synthesis, Spectro-scopic and Structural Studies of Metal–Metal-bonded Metalloporphyrins, in ThePorphyrin Handbook, (Eds.: K. M. Kadish,K. Smith, R. Guilard), Academic Press, SanDiego, 2000, pp. 211–244, Vol. 3.

14. J. E. Falk, Porphyrins and Metalloporphyrins,Elsevier Publishing, New York, 1964.

15. R. S. Nicholson, I. Shain, Anal. Chem. 1964,36, 706.

16. R. S. Nicholson, Anal. Chem. 1965, 37, 1351.

6.5 Concluding Statement 219

17. C. K. Mann in Electroanalytical Chemistry,(Ed.: A. J. Bard), Marcel Dekker, New York,1969, Vol. 3.

18. C. K. Mann, R. K. Barnes, ElectrochemicalReactions in Nonaqueous Systems, MarcelDekker, New York, 1970, Vol. 3.

19. D. T. Sawyer, J. L. Roberts, ExperimentalElectrochemistry for Chemists, John Wiley &Sons, New York, 1974.

20. D. T. Sawyer, A. Sobkowiak, J. L. Roberts,Electrochemistry For Chemists, 2nd ed., JohnWiley & Sons, New York, 1995.

21. K. M. Kadish, E. Van Caemelbecke,G. Royal, Metalloporphyrins in Nonaque-ous Media: Database of Redox Poten-tials, in The Porphyrin Handbook, (Eds.:K. M. Kadish, K. Smith, R. Guilard), Aca-demic Press, San Diego, 2000, pp. 1–219,Vol. 9.

22. K. M. Kadish, R. Guilard, Comments Inorg.Chem. 1988, 7, 287–305.

23. R. Weiss, A. Gold, A. X. Trautwein, J. Ter-ner, High-Valent Iron and ManganeseComplexes of Porphyrins and RelatedMacrocycles, in The Porphyrin Handbook,(Eds.: K. M. Kadish, K. Smith, R. Guilard),Academic Press, San Diego, 2000.

24. B. Meunier, A. Robert, G. Pratviel, J. Berna-dou, Metalloporphyrins in Catalytic Oxida-tions and Oxidative DNA Cleavage, in ThePorphyrin Handbook, (Eds.: K. M. Kadish,K. Smith, R. Guilard), Academic Press, SanDiego, 2000, pp. 121–187, Vol. 4.

25. K. S. Suslick, Shape-selective Oxidationby Metalloporphyrins, in The PorphyrinHandbook, (Eds.: K. M. Kadish, K. Smith,R. Guilard), Academic Press, San Diego,2000, pp. 41–63, Vol. 3.

26. J. T. Groves, K. Shalyaev, J. Lee, Oxometal-loporphyrins in Oxidative Catalysis, in ThePorphyrin Handbook, (Eds.: K. M. Kadish,K. Smith, R. Guilard), Academic Press, SanDiego, 2000, pp. 17–40, Vol. 3.

27. K. M. Kadish, Z. Ou, X. Tan, T. Boschi,D. Monti, V. Fares, P. Tagliatesta, J. Chem.Soc., Dalton Trans. 1999, 10, 1595–1602.

28. P. Ochsenbein, K. Ayougou, D. Mandon,J. Fischer, R. Weiss, R. N. Austin, K. Jaya-raj, A. Gold, J. Terner, J. Fajer, Angew.Chem., Int. Ed. Engl. 1994, 33, 348–350.

29. K. M. Kadish, F. D’Souza, A. Villard,M. Autret, E. Van Caemelbecke, P. Bianco,A. Antonini, P. Tagliatesta, Inorg. Chem.1994, 33, 5169, 5170.

30. P. Tagliatesta, J. Li, M. Autret, E. Van Cae-melbecke, A. Villard, F. D’Souza, K. M.Kadish, Inorg. Chem. 1996, 35, 5570–5576.

31. T. Wijesekera, D. Dupre, M. S. R. Cader,D. Dolphin, Bull. Soc. Chim. Fr. 1996, 133,765–775.

32. E. K. Woller, S. G. DiMagno, J. Org. Chem.1997, 62, 1588–1593.

33. M. Autret, Z. P. Ou, A. Antonini, T. Boschi,P. Tagliatesta, K. M. Kadish, J. Chem. Soc.,Dalton Trans. 1996, 2793–2797.

34. F. D’Souza, A. Villard, E. Van Caemelbec-ke, M. Franzen, T. Boschi, P. Tagliatesta,K. M. Kadish, Inorg. Chem. 1993, 32,4042–4048.

35. F. D’Souza, M. E. Zandler, P. Tagliatesta,Z. Ou, J. Shao, E. Van Caemelbecke, K. M.Kadish, Inorg. Chem. 1998, 37, 4567–4572.

36. K. M. Kadish, J. Li, E. Van Caemelbecke,Z. P. Ou, N. Guo, M. Autret, F. D’Souza,P. Tagliatesta, Inorg. Chem. 1997, 36,6292–6298.

37. M. W. Grinstaff, M. G. Hill, E. R. Birnbaum,W. P. Schaefer, J. A. Labinger, H. B. Gray,Inorg. Chem. 1995, 34, 4896–4902.

38. G. Hariprasad, S. Dahal, B. G. Maiya,J. Chem. Soc., Dalton Trans. 1996,3429–3436.

39. J. G. Goll, K. T. Moore, A. Ghosh, M.J. Therien, J. Am. Chem. Soc. 1996, 118,8344–8354.

40. M. Ravikanth, D. Reddy, A. Misra, T. K.Chandrashekar, J. Chem. Soc., Dalton Trans.1993, 1137–1141.

41. D. Reddy, M. Ravikanth, T. K. Chandra-shekar, J. Chem. Soc., Dalton Trans. 1993,3575–3580.

42. M. Senge, Highly Substituted Porphyrins,in The Porphyrin Handbook, (Eds.: K. M.Kadish, K. Smith, R. Guilard), AcademicPress, San Diego, 2000, pp. 240–347, Vol. 1.

43. J. A. Schelnutt, X.-Z. Song, J.-G. Ma, S.-L. Jia, W. Jentzen, C. J. Medforth, Chem.Soc. Rev. 1998, 27, 31–41.

44. J. A. Hodge, M. G. Hill, H. B. Gray, Inorg.Chem. 1995, 34, 809–812.

45. K. M. Kadish, E. Van Caemelbecke,F. D’Souza, C. J. Medforth, K. M. Smith,A. Tabard, R. Guilard, Inorg. Chem. 1995,34, 2984–2989.

46. D. Gust, Intermolecular Photoinduced Elec-tron-Reactions of Metalloporphyrins, in ThePorphyrin Handbook, (Eds.: K. M. Kadish,


K. Smith, R. Guilard), Academic Press, SanDiego, 2000, pp. 153–205, Vol. 8.

47. J. W. Buchler in Porphyrins and Metallopor-phyrins, (Ed.: K. Smith), Elsevier, New York,1975.

48. J. A. Shelnutt, X.-Z. Song, J.-G. Ma, S.-L. Jia, W. Jentzen, C. J. Medford, Chem. Rev.1998, 27, 31–41.

49. D. D. Perrin, W. L. F. Armarego, D. R. Per-rin, Purification of Laboratory Chemicals,Pergamon Press, New York, 1980.

50. J. F. Coetzee, Recommended Methods for Pu-rification of Solvents and Tests for Impurities,Pergamon Press, New York, 1982.

51. R. H. Felton, H. Linschitz, J. Am. Chem.Soc. 1966, 88, 1113–1116.

52. J.-H. Fuhrhop, K. M. Kadish, D. G. Davis,J. Am. Chem. Soc. 1973, 95, 5140–5147.

53. E. Van Caemelbecke, W. Kutner, K. M.Kadish, Inorg. Chem. 1993, 32, 438–444.

54. K. M. Kadish, C. Araullo, G. B. Maiya, D.Sazou, J.-M. Barbe, R. Guilard, Inorg. Chem.1989, 28, 2528–2533.

55. C. Araullo-McAdams, K. M. Kadish, Inorg.Chem. 1990, 29, 2749–2757.

56. K. M. Kadish, E. Van Caemelbecke,E. Gueletii, S. Fukuzumi, K. Miyamoto,T. Suenobu, A. Tabard, R. Guilard, Inorg.Chem. 1998, 37, 1759–1766.

57. J. W. Seyler, C. R. Leidner, J. Chem. Soc.,Chem. Commun. 1989, 794.

58. K. M. Kadish, Q. Y. Xu, J.-M. Barbe, J. E.Anderson, E. Wang, R. Guilard, J. Am.Chem. Soc. 1987, 109, 7705–7714.

59. K. M. Kadish, Q. Y. Xu, J.-M. Barbe, Inorg.Chem. 1988, 27, 1191–1198.

60. K. M. Kadish, Q. Y. Xu, J. E. Anderson, ACSSymp. Ser. 1988, 378, 451–465.

61. K. M. Kadish, F. D’Souza, E. Van Caemel-becke, A. Villard, J.-D. Lee, A. Tabard,R. Guilard, Inorg. Chem. 1993, 32,4179–4185.

62. D. Lancon, P. Cocolios, R. Guilard, K. M.Kadish, Organometallics 1984, 3, 1164–1170.

63. G. B. Maiya, B. C. Han, K. M. Kadish, Lang-muir 1989, 5, 645–650.

64. Q. Y. Xu, J. M. Barbe, K. M. Kadish, Inorg.Chem. 1988, 27, 2373–2378.

65. R. Guilard, G. Lagrange, A. Tabard, D.Lancon, K. M. Kadish, Inorg. Chem. 1985,24, 3649–3656.

66. I. K. Choi, D. Feng, K. J. Paeng, M. D. Ryan,Inorg. Chem. 1991, 30, 1832–1839.

67. E. Fujita, C. K. Chang, J. Fajer, J. Am. Chem.Soc. 1985, 107, 7665–7669.

68. K. M. Kadish, D. Lancon, P. Cocolios,R. Guilard, Inorg. Chem. 1984, 23, 2372,2373.

69. K. M. Kadish, V. A. Adamian, E. Van Cae-melbecke, Z. Tan, P. Tagliatesta, P. Bianco,T. Boschi, G. B. Yi, M. A. Khan, G. B. Rich-ter-Addo, Inorg. Chem. 1996, 35, 1343–1348.

70. A. D. Kini, J. Washington, C. P. Kubiak,B. H. Morimoto, Inorg. Chem. 1996, 35,6904–6906.

71. D. Lancon, K. M. Kadish, J. Am. Chem. Soc.1983, 105, 5610–5617.

72. I. M. Lorkovic, K. M. Miranda, B. Lee,S. Bernhard, J. R. Schoonover, P. C. Ford,J. Am. Chem. Soc 1998, 120, 11 674–11 683.

73. I. M. Lorkovic, P. C. Ford, Inorg. Chem.1999, 38, 1467–1473.

74. M. Hoshino, K. Yasufuku, S. Konishi,M. Imamura, Inorg. Chem. 1984, 23, 1982.

75. L. E. Laverman, M. Hoshino, P. C. Ford,J. Am. Chem. Soc. 1997, 119, 12 663–12 664.

76. I. Lorkovic, P. C. Ford, J. Am. Chem. Soc.2000, 122, 6516–6517.

77. I. Lorkovic, P. C. Ford, Inorg. Chem. 2000,39, 632, 633.

78. L. Cheng, G. Richter-Addo, Binding andActivation of Nitric Oxide by Metallo-porphyrins and Heme, in The PorphyrinHandbook, (Eds.: K. M. Kadish, K. Smith,R. Guilard), Academic Press, San Diego,2000, pp. 219–290, Vol. 4.

79. W. H. Leung, T. S. M. Hun, K. Y. Wong,W. T. Wong, J. Chem. Soc., Dalton Trans.1994, 2713–2718.

80. W. R. Scheidt, Systematics of the Stereo-chemistry of Porphyrins and Metallopor-phyrins, in The Porphyrin Handbook, (Eds.:K. M. Kadish, K. Smith, R. Guilard), Aca-demic Press, San Diego, 2000, pp. 49–112,Vol. 3.

81. D. S. Bohle, C.-H. Hung, J. Am. Chem. Soc.1995, 117, 9584, 9585.

82. K. M. Kadish, J. Electroanal. Chem. 1984,168, 261–274.

83. L. W. Olson, D. Schaeper, D. Lancon, K. M.Kadish, J. Am. Chem. Soc. 1982, 104,2042–2044.

84. B. B. Wayland, L. W. Olson, J. Am. Chem.Soc. 1974, 96, 6037–6041.

85. I. K. Choi, M. D. Ryan, Inorg. Chim. Acta1988, 153, 25–30.


86. X. H. Mu, K. M. Kadish, Inorg. Chem. 1990,29, 1031–1036.

87. J. W. Buchler, W. Kokisch, P. D. Smith, Z.Naturforsch. B 1978, 33b, 1371.

88. E. Fujita, J. Fajer, J. Am. Chem. Soc. 1983,105, 6743–6745.

89. M. K. Ellison, W. R. Scheidt, J. Am. Chem.Soc. 1997, 119, 7404, 7405.

90. K. M. Kadish, X. H. Mu, Inorg. Chem. 1990,29, 1031–1036.

91. K. M. Kadish, X. H. Mu, X. Q. Lin, Inorg.Chem. 1988, 27, 1489–1492.

92. K. M. Kadish, X. H. Mu, Pure Appl. Chem.1990, 62, 1051–1054.

93. G. B. Richter-Addo, S. J. Hodge, G.-B. Yi,M. A. Khan, T. Ma, E. Van Caemelbecke,N. Guo, K. M. Kadish, Inorg. Chem. 1996,35, 6530–6538.

94. D. S. Bohle, C.-H. Hung, B. D. Smith, In-org. Chem. 1998, 37, 5798–5806.

95. D. S. Bohle, P. A. Goodson, B. D. Smith,Polyhedron 1996, 15, 3147–3150.

96. D. S. Bohle, C.-H. Hung, A. K. Powell, B.D. Smith, S. Wocadlo, Inorg. Chem. 1997,36, 1992–1993.

97. B. B. Wayland, R. A. Newman, Inorg. Chem.1981, 20, 3093–3097.

98. S. L. Kelly, K. M. Kadish, Inorg. Chem. 1984,23, 679–687.

99. T. Diebold, M. Schppacher, B. Chevrier,R. Weiss, J. Chem. Soc., Chem. Commun.1979, 693, 694.

100. S. L. Kelly, D. Lancon, K. M. Kadish, Inorg.Chem. 1984, 23, 1451–1458.

101. C.-L. Yao, J. E. Anderson, K. M. Kadish, In-org. Chem. 1987, 26, 2725–2727.

102. K. M. Kadish, Y. J. Deng, C.-L. Yao, J. E. An-derson, Organometallics 1988, 7, 1979–1983.

103. C. Swistak, J.-L. Cornillon, J. E. Anderson,K. M. Kadish, Organometallics 1987, 6,2146–2150.

104. K. M. Kadish, D. Chang, Inorg. Chem. 1982,21, 3614–3618.

105. K. M. Kadish, D. J. Leggett, D. Chang, In-org. Chem. 1982, 21, 3618–3622.

106. K. M. Kadish, J. L. Cornillon, C.-L. Yao,T. Malinski, G. Gritzner, J. Electroanal.Chem. 1987, 235, 189–207.

107. S. Takagi, T. K. Miyamoto, M. Hamagushi,Y. Sasaki, Inorg. Chim. Acta 1990, 173,215–221.

108. E. R. Birnbaum, W. P. Schaefer, J. A. Labin-ger, J. E. Bercaw, H. B. Gray, Inorg. Chem.1995, 34, 1751–1755.

109. K. M. Kadish, P. Tagliatesta, Y. Hu, Y. J.Deng, X. H. Mu, L. Y. Bao, Inorg. Chem.1991, 30, 3737–3743.

110. G. M. Brown, F. R. Hopf, T. J. Meyer, D. G.Whitten, J. Am. Chem. Soc. 1975, 97,5385–5390.

111. G. M. Brown, F. R. Hopf, J. A. Ferguson,T. J. Meyer, D. G. Whitten, J. Am. Chem.Soc. 1973, 95, 5939–5942.

112. T. Boshi, G. Bontempelli, G.-A. Mazzochin,Inorg. Chim. Acta 1979, 37, 155–160.

113. D. P. Rillema, J. K. Nagle, L. F. Barringer,T. J. Meyer, J. Am. Chem. Soc. 1981, 103,56–62.

114. Y. Hu, B. C. Han, L. Y. Bao, X. H. Mu, K. M.Kadish, Inorg. Chem. 1991, 30, 2444–2446.

115. X. H. Mu, K. M. Kadish, Inorg. Chem. 1989,28, 3743–3747.

116. K. M. Kadish, M. M. Franzen, B. C. Han,C. Araullo-McAdams, D. Sazou, J. Am.Chem. Soc. 1991, 113, 512–517.

117. K. M. Kadish, Y. Hu, P. Tagliatesta, T. Bos-chi, J. Chem. Soc., Dalton Trans. 1993,1167–1172.

118. K. Rachlewicz, M. Grzeszczuk, L. Latos-Grazynski, Polyhedron 1993, 12, 821–829.

119. G. Peychal-Heiling, G. S. Wilson, Anal.Chem. 1971, 43, 545–550.

120. G. Peychal-Heiling, G. S. Wilson, Anal.Chem. 1971, 43, 550–556.

121. D. Chabach, A. De Cian, J. Fischer, R. Wei-ss, M. E. Bibout, Angew. Chem., Int. Ed.Engl. 1996, 35, 898, 899.

122. J. W. Buchler, P. Hammerschmitt, I. Kau-feld, J. Loffler, Chem. Ber. 1991, 124,2151–2159.

123. J. W. Buchler, B. Scharbert, J. Am. Chem.Soc. 1988, 110, 4272–4276.

124. J. K. Duchowski, D. F. Bocian, Inorg. Chem.1990, 29, 4158–4160.

125. J. W. Buchler, M. Kihn-Botulinski, J.Loffler, B. Scharbert, New J. Chem. 1992,16, 545–553.

126. D. Chabach, M. Tahiri, A. De Cian, J. Fisc-her, R. Weiss, M. E. Bibout, J. Am. Chem.Soc. 1995, 117, 8548–8556.

127. J. Buchler, D. K. P. Ng, Metal TetrapyrroleDouble- and Triple-Deckers with SpecialEmphasis on Porphyrin Systems, in ThePorphyrin Handbook, (Eds.: K. M. Kadish,K. Smith, R. Guilard), Academic Press, SanDiego, 2000, pp. 246–294, Vol. 3.


128. J. W. Buchler, G. Eikelmann, L. Puppe,K. Rohbock, H. H. Schneehage, D. Weck,Liebigs Ann. Chem. 1971, 745, 135–151.

129. T. Malinski, D. Chang, J. M. Latour, J.-C.Marchon, M. Gross, A. Giraudeau, K. M.Kadish, Inorg. Chem. 1984, 23, 3947–3955.

130. A. R. Miksztal, J. S. Valentine, Inorg. Chem.1984, 23, 3548–3552.

131. P. Friant, J. Goulon, J. Fischer, L. Ricard,M. Schappacher, R. Weiss, M. Momenteau,Nouv. J. Chim. 1985, 9, 33–40.

132. R. Guilard, J.-M. Latour, C. Lecomte, J.-C.Marchon, J. Protas, D. Ripoll, Inorg. Chem.1978, 17, 1228–1237.

133. J.-M. Latour, J. C. Marchon, N. Nakajima,J. Am. Chem. Soc. 1979, 101, 3974–3976.

134. M. Nakajima, J.-M. Latour, J. C. Marchon,J. Chem. Soc., Chem. Commun. 1977, 763.

135. C. Ratti, P. Richard, A. Tabard, R. Guilard,J. Chem. Soc., Chem. Commun. 1989, 69–70.

136. R. Guilard, C. Ratti, A. Tabard, P. Richard,D. Dubois, K. M. Kadish, Inorg. Chem. 1990,29, 2532–2540.

137. L. K. Woo, J. A. Hays, V. G. Young, C. L. Day,C. Caron, F. D’souza, K. M. Kadish, Inorg.Chem. 1993, 32, 4186–4192.

138. P. Richard, R. Guilard, J. Chem. Soc., Chem.Commun. 1983, 1454.

139. R. Guilard, P. Richard, M. El Borai, E. Lavi-ron, J. Chem. Soc., Chem. Commun. 1980,516–518.

140. J. E. Anderson, Y. H. Liu, R. Guilard, J.-M.Barbe, K. M. Kadish, Inorg. Chem. 1986, 25,3786–3791.

141. J. E. Anderson, Y. H. Liu, R. Guilard, J.-M.Barbe, K. M. Kadish, Inorg. Chem. 1986, 25,2250–2255.

142. C. M. Newton, D. G. Davis, J. Magn. Reson.1975, 20, 446–457.

143. J. W. Buchler, C. Dreher, K.-L. Lay, Y. J. A.Lee, W. R. Scheidt, Inorg. Chem. 1983, 22,888.

144. J. W. Buchler, C. Dreher, Z. Naturforsch. B.1984, 39, 222.

145. L. A. Bottomley, F. L. Neely, Inorg. Chem.1997, 36, 5435–5439.

146. F. L. Neely, L. A. Bottomley, Inorg. Chim.Acta 1992, 192, 147–149.

147. J. T. Groves, T. Takahashi, W. M. Butler,Inorg. Chem. 1983, 22, 884–887.

148. P. J. Nichols, J. D. Fallon, B. Moubaraki,K. S. Murray, B. O. West, Polyhedron 1993,12, 2205.

149. K. A. Macor, T. G. Spiro, J. Am. Chem. Soc.1983, 105, 5601–5607.

150. L. A. Bottomley, F. L. Neely, Inorg. Chem.1990, 29, 1860–1865.

151. F. Basolo, R. D. Jones, D. A. Summerville,Acta Chem. Scand. 1978, A32, 771.

152. D. A. Summerville, R. D. Jones, B. M. Hoff-man, F. Basolo, J. Am. Chem. Soc. 1977, 99,8195–8202.

153. L. A. Bottomley, K. M. Kadish, Inorg. Chem.1983, 22, 342–349.

154. V. Gutmann, The Donor-Acceptor Approachto Molecular Interactions, Plenum Press,New York, 1978.

155. T. Malinski, P. M. Hanley, K. M. Kadish,Inorg. Chem. 1986, 25, 3229–3235.

156. J. Topich, N. Berger, Inorg. Chim. Acta 1982,65, L131–L134.

157. K. M. Kadish, D. Chang, T. Malin-ski, H. Ledon, Inorg. Chem. 1983, 22,3490–3492.

158. H. Ledon, F. Varescon, T. Malinski, K. M.Kadish, Inorg. Chem. 1984, 23, 261–263.

159. T. Malinski, H. Ledon, K. M. Kadish,J. Chem. Soc., Chem. Commun. 1983,1077–1079.

160. D. D. Axtell, G. R. Miller, T. H. Ridgway,M. Tsutsui, J. Coord. Chem. 1978, 8,113–115.

161. K. M. Kadish, D. Schaeper, L. A. Bottomley,M. Tsutsui, R. L. Bobsein, J. Inorg. Nucl.Chem. 1980, 42, 469–474.

162. K. M. Kadish, L. A. Bottomley, D. Shaefer,M. Tsutsui, R. L. Bobsein, Inorg. Chim. Acta1979, 36, 219.

163. R. Guilard, N. Jagerovic, J.-M. Barbe, Y. H.Liu, I. Perrot, C. Naillon, E. Van Caemel-becke, K. M. Kadish, Polyhedron 1995, 14,3041–3050.

164. R. S. Czernuszewicz, Y. O. Su, M. K. Stern,K. A. Macor, D. Kim, J. T. Groves, T. G.Spiro, J. Am. Chem. Soc. 1988, 110,4158–4165.

165. J. T. Groves, M. K. Stern, J. Am. Chem. Soc.1988, 110, 8628–8638.

166. L. J. Boucher, H. K. Garber, Inorg. Chem.1970, 9, 2644–2649.

167. K. M. Kadish, M. M. Morrison, Bioelectro-chem. Bioenerg. 1977, 3, 480.

168. S. L. Kelly, K. M. Kadish, Inorg. Chem. 1982,21, 3631–3639.

169. K. Perie, J.-M. Barbe, P. Cocolios, R. Guil-ard, Bull. Soc. Chim. Fr. 1996, 133, 697–702.


170. R. Guilard, K. Perie, J. M. Barbe, D. J. Nurco,K. M. Smith, E. Van Caemelbecke, K. M.Kadish, Inorg. Chem. 1998, 37, 973–981.

171. S. Fukuzumi, I. Nakanishi, J.-M. Barbe,R. Guilard, E. Van Caemelbecke, N. Guo,K. M. Kadish, Angew. Chem., Int. Ed. Engl.1999, 38, 964–966.

172. R. D. Arasasingham, T. C. Bruice, Inorg.Chem. 1990, 29, 1422–1427.

173. R. D. Arasasingham, G.-X. He, T. C. Bruice,J. Am. Chem. Soc. 1993, 115, 7985–7991.

174. K. R. Rodgers, H. M. Goff, J. Am. Chem.Soc. 1988, 110, 7049–7060.

175. C. Ercolani, J. Jubb, G. Pennesi, U. Russo,G. Trigiante, Inorg. Chem. 1995, 34,2535–2541.

176. Y. J. Deng, X. H. Mu, P. Tagliatesta, K. M.Kadish, Inorg. Chem. 1991, 30, 1957–1960.

177. A. Pacheco, B. R. James, S. J. Retting, Inorg.Chem. 1995, 34, 3477–3484.

178. W. H. Leung, C. M. Che, J. Am. Chem. Soc.1989, 111, 8812–8818.

179. C. Ho, W. H. Leung, C. M. Che, J. Chem.Soc., Dalton Trans. 1991, 2933–2939.

180. J. W. Seyler, C. R. Leidner, Inorg. Chem.1990, 29, 3636–3641.

181. J. W. Seyler, L. K. Safford, P. E. Fanwick,C. R. Leidner, Inorg. Chem. 1992, 31,1545–1547.

182. J. W. Seyler, L. K. Safford, C. R. Leidner, In-org. Chem. 1992, 31, 4300–4307.

183. R. Guilard, K. M. Kadish, Chem. Rev. 1988,88, 1121–1146.

184. K. M. Kadish, E. Van Caemelbecke,F. D’Souza, C. J. Medforth, K. M. Smith, A.Tabard, R. Guilard, Organometallics 1993,12, 2411–2413.

185. J. P. Collman, D. S. Bohle, A. K. Powell, In-org. Chem. 1993, 32, 4004–4011.

186. C. M. Che, W. H. Leung, W. C. Chung,Inorg. Chem. 1990, 29, 1841–1846.

187. L. A. Bottomley, K. M. Kadish, Inorg. Chem.1981, 20, 1348–1357.

188. D. Lancon, P. Cocolios, R. Guilard, K. M.Kadish, J. Am. Chem. Soc. 1984, 106,4472–4478.

189. C. Swistak, X. H. Mu, K. M. Kadish, Inorg.Chem. 1987, 26, 4360–4366.

190. K. M. Kadish, R. K. Rhodes, Inorg. Chem.1983, 22, 1090–1094.

191. L. A. Constant, D. G. Davis, Anal. Chem.1975, 47, 2253–2260.

192. L. A. Bottomley, L. Olson, K. M. Kadish,Adv. Chem. Ser. 1982, 201, 279–311.

193. K. M. Kadish, L. A. Bottomley, Inorg. Chem.1980, 19, 832–836.

194. K. M. Kadish, L. K. Thompson, D. Beroiz,L. A. Bottomley, ACS Symp. Ser. 1977, 38,51.

195. K. M. Kadish, D. Schaeper, J. Chem. Soc.,Chem. Commun. 1980, 1273–1275.

196. R. Koerner, J. L. Wright, X. D. Ding,M. J. M. Nesset, K. Aubrecht, R. A. Watson,R. A. Barber, L. M. Mink, A. R. Tipton,C. J. Norvel, K. Skidmore, U. Simonis,F. A. Walker, Inorg. Chem. 1998, 37,733–745.

197. A. Ghosh, J. Phys. Chem. 1994, 98, 11 004.198. A. Ghosh, J. Am. Chem. Soc. 1995, 117,

4691–4699.199. A. Ghosh, Acc. Chem. Res. 1998, 31,

189–198.200. A. Ghosh, Quantum Chemical Studies of

Molecular Structures and Potential EnergySurfaces of Porphyrins and Hemes, in ThePorphyrin Handbook, (Eds.: K. M. Kadish,K. Smith, R. Guilard), Academic Press, SanDiego, 2000, pp. 1–38, Vol. 7.

201. K. M. Kadish, M. M. Morrison, L. A. Const-ant, L. Dickens, D. G. Davis, J. Am. Chem.Soc. 1976, 98, 8387–8390.

202. C. A. Reed in ACS Advances in Chem-istry Series, (Ed.: K. M. Kadish), AmericanChemical Society, Washington, D.C., 1982,Vol. 201.

203. K. M. Kadish, E. Van Caemelbecke, F.D’Souza, M. Lin, D. J. Nurco, C. J. Medford,T. P. Forsyth, B. Krattinger, K. M. Smith,S. Fukuzumi, I. Nakanishi, J. A. Shelnutt,Inorg. Chem. 1999, 38, 2188–2198.

204. R. J. Donohoe, M. Atamian, D. F. Bocian,J. Am. Chem. Soc. 1987, 109, 5593–5599.

205. K. T. Moore, J. T. Fletcher, M. J. Therien,J. Am. Chem. Soc. 1999, 121, 5196–5209.

206. M. A. Phillippi, E. T. Shimomura, H. M.Goff, Inorg. Chem. 1981, 20, 1322–1325.

207. A. Wolberg, Isr. J. Chem. 1974, 12,1031–1035.

208. C. A. Reed, T. Mashiko, S. P. Bentley,M. E. Kastner, W. R. Scheidt, K. Spartalian,G. Lang, J. Am. Chem. Soc. 1979, 101,2948–2958.

209. H. M. Goff, M. A. Phillippi, A. D. Boersma,A. P. Hansen, Electrochemical and Spectro-scopic Studies on Biological Redox Compo-nents, ACS Advances in Chemistry Series,American Chemical Society, Washington,D.C., 1982, pp. 357–376, Vol. 201.


210. M. A. Phillippi, H. M. Goff, J. Am. Chem.Soc. 1982, 104, 6026–6034.

211. T. E. Shimomura, M. A. Phillippi, H. M.Goff, W. F. Scholz, C. A. Reed, J. Am.Chem. Soc. 1981, 103, 6778–6780.

212. A. D. Boersma, H. M. Goff, Inorg. Chem.1984, 23, 1671–1676.

213. H. Fujii, J. Am. Chem. Soc. 1993, 115,4641–4648.

214. J. T. Groves, J. A. Gilbert, Inorg. Chem.1986, 25, 123–125.

215. T. S. Calderwood, W. A. Lee, T. C. Bruice,J. Am. Chem. Soc. 1985, 107, 8272, 8273.

216. J. T. Groves, Z. Gross, M. K. Stern, Inorg.Chem. 1994, 33, 5065–5072.

217. T. S. Calderwood, T. C. Bruice, Inorg. Chem.1986, 25, 3722–3724.

218. W. A. Lee, T. S. Calderwood, T. C. Bruice,Proc. Natl. Acad. Sci. U.S.A. 1985, 82,4301–4305.

219. S. Fukuzumi, I. Nakanishi, K. Tanaka,T. Suenobu, A. Tabard, R. Guilard, E. VanCaemelbecke, K. M. Kadish, J. Am. Chem.Soc. 1999, 121, 785–790.

220. A. Nanthakumar, H. M. Goff, J. Am. Chem.Soc. 1990, 112, 4047–4049.

221. A. Nanthakumar, H. M. Goff, Inorg. Chem.1991, 30, 4460–4464.

222. L. D. Sparks, C. J. Medforth, M. S. Park,J. R. Chamberlain, M. R. Ondrias, M. O.Senge, K. M. Smith, J. A. Shelnutt, J. Am.Chem. Soc. 1993, 115, 581–592.

223. R.-J. Cheng, P. Y. Chen, P.-R. Gau, C.-C. Chen, S.-M. Peng, J. Am. Chem. Soc.1997, 119, 2563.

224. R. A Sheldon, (Ed.), Metalloporphyrins inCatalytic Oxidations, Marcel Dekker, NewYork, 1994.

225. M. W. Grinstaff, M. G. Hill, J. A. Labinger,H. B. Gray, Science 1994, 264, 1311.

226. D. Lexa, P. Maillard, M. Momenteau, J. M.Saveant, J. Am. Chem. Soc. 1984, 106,6321–6323.

227. C. Gueutin, D. Lexa, J.-M. Saveant, D.-L.Wang, Organometallics 1989, 8, 1607–1613.

228. D. Lexa, J. Mispelter, J.-M. Saveant, J. Am.Chem. Soc. 1981, 103, 6806–6812.

229. D. Lexa, J.-M. Saveant, D. L. Wang, Organo-metallics 1986, 5, 1428–1434.

230. D. Lexa, M. Momenteau, J.-M. Saveant,F. Xu, Inorg. Chem. 1985, 24, 122–127.

231. C. Gueutin, D. Lexas, M. Momenteau, J.-M. Saveant, F. Xu, Inorg. Chem. 1986, 25,4294–4307.

232. D. Lexa, M. Momenteau, P. Rentien,G. Rytz, J.-M. Saveant, F. Xu, J. Am. Chem.Soc. 1984, 106, 4755–4765.

233. D. Lexa, M. Momenteau, J. M. Saveant,F. Xu, J. Am. Chem. Soc. 1986, 108,6937–6941.

234. A. El-Kasmi, D. Lexa, P. Maillard, M.Momenteau, J.-M. Saveant, J. Am. Chem.Soc. 1991, 113, 1586–1595.

235. J. H. Cameron, S. C. Turner, Polyhedron1993, 12, 1675–1680.

236. J. H. Cameron, S. C. Turner, J. Chem. Soc.,Dalton Trans. 1993, 1941–1945.

237. S. H. Strauss, R. H. Holm, Inorg. Chem.1982, 21, 863–868.

238. D. V. Stynes, B. R. James, J. Chem. Soc.,Chem. Commun. 1973, 325.

239. B. B. Wayland, L. F. Mehne, J. Swartz,J. Am. Chem. Soc. 1978, 100, 2379–2383.

240. C. Swistak, K. M. Kadish, Inorg. Chem.1987, 26, 405–412.

241. D. Dolphin, B. R. James, H. C. Welborne,ACS Adv. Chem. Ser. 1983, 201, 563.

242. L. A. Bottomley, M. R. Deakin, J.-N. Gorce,Inorg. Chem. 1984, 23, 3563–3571.

243. J.-N. Gorce, L. A. Bottomley, Inorg. Chem.1985, 24, 1431–1436.

244. D. Chang, P. Cocolios, Y. T. Wu, K. M.Kadish, Inorg. Chem. 1984, 23, 1629–1633.

245. K. M. Kadish, R. K. Rhodes, L. A. Bottom-ley, H. M. Goff, Inorg. Chem. 1981, 20,3195–3200.

246. D. Lancon, K. M. Kadish, Inorg. Chem. 1984,23, 3942–3947.

247. D. R. English, D. N. Hendrickson, K. S.Suslick, Inorg. Chem. 1983, 22, 367, 368.

248. D. Mansuy, P. Battioni, Diversity of Reac-tions Catalyzed by Heme-Thiolate Proteins,in The Porphyrin Handbook, (Eds.: K. M.Kadish, K. M. Smith, R. Guilard), AcademicPress, San Diego, 2000, pp. 1–15, Vol. 4.

249. P. R. O. de Montellano, (Ed.), CytochromeP450: Structure, Mechanism and Biochem-istry, Plenum Publishing, New York, 1995.

250. J. T. Groves, Y. Z. Han in Cytochrome P450:Structure, Mechanism, and Biochemistry,(Ed.: P. R. O. de Montellano), 2nd ed.,Plenum Publishing, New York, 1995, p. 3.

251. P. R. Ortiz de Montellano in CytochromeP450: Structure, Mechanism, and Biochem-istry, (Ed.: P. R. O. de Montellano), 2nded., Plenum Publishing, New York, 1995,p. 245.


252. E. J. Mueller, P. J. Loida, S. G. Sligar in Cy-tochrome P450: Structure, Mechanism, andBiochemistry, (Ed.: P. R. O. de Montellano),2nd ed., Plenum Publishing, New York,1995, p. 83.

253. A. D. N. Vaz, D. F. Mc Ginnity, M. J. Coon,Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 3555.

254. A. D. N. Vaz, S. J. Pernecky, G. M. Raner,M. J. Coon, Proc. Natl. Acad. Sci. U.S.A.1996, 93, 4644.

255. K. G. Welinder, Curr. Biol. 1992, 2, 388.256. T. L. Poulos, Peroxidase and Cytochrome

P450 Structures, in The Porphyrin Hand-book, (Eds.: K. M. Kadish, K. M. Smith,R. Guilard), Academic Press, San Diego,2000, pp. 189–218, Vol. 4.

257. D. Dolphin, A. Forman, D. C. Borg, J. Fajer,R. H. Felton, Proc. Natl. Acad. Sci. U.S.A.1971, 68, 614.

258. R. Rutter, M. Valentine, M. P. Hendrich,L. P. Hager, P. G. Debrunner, Biochemistry1983, 22, 4769.

259. B. Chance, L. Powers, Y. Ching, T. Poulos,G. R. Schonbaum, I. Yamazaki, K. G. Paul,Arch. Biochem. Biophys. 1984, 235, 596.

260. V. Fulop, R. P. Phizackerley, S. M. Soltis, I.J. Clifton, S. Wakatuski, J. Erman, J. Hajdu,S. L. Edwards, Structure 1994, 2, 201.

261. S. L. Edwards, H. X. Nguyen, R. C. Hamlin,J. Kraut, Biochemistry 1987, 26. 1503.

262. P. Gouet, H. M. Jouve, P. A. Williams,I. Anderson, P. Andreoletti, L. Nussaume,J. Hajdu, Nat. Struct. Biol. 1996, 3, 951.

263. H. B. Dunford in Peroxidases in Chemistryand Biology, (Eds.: J. Everse, K. E. Everse,M. B. Grisham), CRC Press, Boca Raton,1991, pp. 1–24, Vol. 2.

264. H. K. Baek, H. E. Van Wart, Biochemistry1989, 28, 5714.

265. D. Job, H. B. Dunford, Eur. J. Biochem.1976, 66, 607.

266. H. B. Dunford, M. L. Cotton, J. Biol. Chem.1975, 250, 2920.

267. I. C. Kuan, K. A. Johnson, M. Tien, J. Biol.Chem. 1993, 268, 20 064.

268. R. S. Koduri, M. Tien, Biochemistry 1994,33, 4225.

269. H. Wariishi, J. Huang, H. B. Dunford,M. H. Gold, J. Biol. Chem. 1991, 266, 20 694.

270. J. E. Critchlow, H. B. Dunford, J. Biol.Chem. 1972, 247, 3703.

271. J. E. Critchlow, H. B. Dunford, J. Biol.Chem. 1972, 247, 3714.

272. A. M. Altschul, R. Abrams, T. R. Hogness,J. Biol. Chem. 1940, 136, 777.

273. T. Yonetani, Enzymes 1976, 13, 345.274. L. Cheng, G. B. Richter-Addo, Binding and

Activation of Nitric Oxide by Metallopor-phyrins and Heme, in The Porphyrin Hand-book, (Eds.: K. M. Kadish, K. M. Smith,R. Guilard), Academic Press, San Diego,2000, pp. 189–218, Vol. 4.

275. B. A. Averill, Chem. Rev. 1996, 96, 2951.276. W. G. Zumft, H. Korner, Antonie van

Leewenhock 1997, 71, 43.277. B. C. Berks, S. J. Ferguson, J. W. B. Moir,

Biochim. Biophys. Acta 1995, 1232, 97.278. T. C. Hollocher in Nitric Oxide. Principles

and Actions, (Ed.: J. Lancaster Jr.), AcademicPress, San Diego, 1996, pp. 289–344.

279. T. Fujiwara, Y. Fukumori, J. Bacteriol. 1996,178, 1866.

280. P. Girsch, S. de Vries, Biochim. Biophys.Acta 1997, 202, 1318.

281. P. Moenne-Loccoz, S. de Vries, J. Am.Chem. Soc. 1998, 120, 5147.

282. M. R. Cheesman, W. G. Zumft, A. J. Thom-son, Biochemistry 1998, 37, 3994.

283. N. Sakurai, T. Sakurai, Biochemistry 1997,36, 13 809.

284. N. Sakurai, T. Sakurai, Biochem. Biophys.Res. Commun. 1998, 243, 400.

285. T. Sakurai, N. Sakurai, H. Matsumoto,S. Hirota, O. Yamauchi, Biochem. Biophys.Res. Commun. 1998, 251, 248.

286. P. O’Brien, D. A. Sweigart, Inorg. Chem.1985, 24, 1405–1409.

287. M. M. Doeff, D. A. Sweigart, P. O’Brien, In-org. Chem. 1983, 22, 851–852.

288. R. Quinn, M. Nappa, J. S. Valentine, J. Am.Chem. Soc. 1982, 104, 2588–2595.

289. R. Quinn, J. Mercer-Smith, J. N. Burstyn,J. S. Valentine, J. Am. Chem. Soc. 1984, 106,4136–4144.

290. K. M. Kadish, Y. J. Deng, J. D. Korp, Inorg.Chem. 1990, 29, 1036–1042.

291. J. E. Anderson, C.-L. Yao, K. M. Kadish, In-org. Chem. 1986, 25, 718,719.

292. J. E. Anderson, C.-L. Yao, K. M. Kadish,Organometallics 1987, 6, 706–711.

293. J. E. Anderson, C.-L. Yao, K. M. Kadish, J.Am. Chem. Soc. 1987, 109, 1106–1111.

294. J. E. Anderson, Y. H. Liu, K. M. Kadish, In-org. Chem. 1987, 26, 4174–4179.

295. K. M. Kadish, C.-L. Yao, J. E. Anderson,P. Cocolios, Inorg. Chem. 1985, 24,4515–4520.


296. K. M. Kadish, J. E. Anderson, C. L. Yao,R. Guilard, Inorg. Chem. 1986, 25,1277–1280.

297. K. M. Kadish, W. Koh, P. Tagliatesta, D.Sazou, R. Paolesse, S. Licoccia, T. Boschi,Inorg. Chem. 1992, 31, 2305–2313.

298. K. M. Kadish, Y. Hu, T. Boschi, P. Taglia-testa, Inorg. Chem. 1993, 32, 2996–3002.

299. V. Grass, D. Lexa, J.-M. Saveant, J. Am.Chem. Soc. 1997, 119, 7526–7532.

300. V. Grass, D. Lexa, M. Momenteau, J.-M. Saveant, J. Am. Chem. Soc. 1997, 119,3536–3542.

301. D. Lexa, V. Grass, J.-M. Saveant, Organome-tallics 1998, 17, 2673–2676.

302. J. E. Anderson, C.-L. Yao, K. M. Kadish, In-org. Chem. 1986, 25, 3224–3228.

303. S. Fukuzumi, K. Miyamoto, T. Suenobu,E. Van Caemelbecke, K. M. Kadish, J. Am.Chem. Soc. 1998, 120, 2880–2889.

304. A. Stanienda, G. Biebl, Z. Phys. Chem. 1967,52, 254–275.

305. M. J. Carter, L. M. Engelhardt, D. P. Rill-ema, F. Basolo, J. Chem. Soc., Chem. Com-mun. 1973, 810–812.

306. M. J. Carter, D. P. Rillema, F. Basolo, J. Am.Chem. Soc. 1974, 23, 392–400.

307. L. A. Truxillo, D. G. Davis, Anal. Chem.1975, 47, 2260–2267.

308. F. A. Walker, D. Beroiz, K. M. Kadish,J. Am. Chem. Soc. 1976, 98, 3484–3489.

309. J. Huet, A. Gaudemer, C. Boucly-Goester,P. Boucly, Inorg. Chem. 1982, 21,3413–3419.

310. K. M. Kadish, X. Q. Lin, B. C. Han, Inorg.Chem. 1987, 26, 4161–4167.

311. X. H. Mu, X. Q. Lin, K. M. Kadish, Electro-analysis 1989, 1, 113–116.

312. A. Giraudeau, H. J. Callot, J. Jordan,I. Ezhar, M. Gross, J. Am. Chem. Soc. 1979,101, 3857–3862.

313. X. Q. Lin, B. Boisselier-Cocolios, K. M. Kad-ish, Inorg. Chem. 1986, 25, 3242–3248.

314. C. Shi, F. C. Anson, Inorg. Chem. 1998, 37,1037–1043.

315. C. Shi, F. C. Anson, Inorg. Chem. 1992, 31,5078–5083.

316. B. Steiger, C. Shi, F. C. Anson, Inorg. Chem.1993, 32, 2107–2113.

317. M. Yuasa, F. C. Anson, J. Porphyrins Ph-thalocyanines 1997, 1, 181–188.

318. F. D’Souza, Y.-Y. Hsieh, G. R. Deviprasad,J. Electroanal. Chem. 1997, 426, 17–21.

319. F. D’Souza, Y.-Y. Hsieh, G. R. Deviprasad,Chem. Commun. (Cambridge) 1998, 1027,1028.

320. Y. Le Mest, C. Inisan, A. Laouenan,M. L’Her, J. Talarmin, M. El Kalifa, J.-Y. Saillard, J. Am. Chem. Soc. 1997, 119,6095–6106.

321. M. Tezuka, M. Iwasaki, Chem. Lett. 1993, 3,427–430.

322. H.-Z. Yu, J. S. Baskin, B. Steiger, F. C. An-son, A. H. Zewail, J. Am. Chem. Soc. 1999,121, 484, 485.

323. K. M. Kadish, M. M. Franzen, B. C. Han,C. Araullo-McAdams, D. Sazou, Inorg.Chem. 1992, 31, 4399–4403.

324. D. Lexa, M. Momenteau, J. Mispelter, J.-M.Saveant, Inorg. Chem. 1989, 28, 30–35.

325. J. Seth, V. Palaniappan, D. F. Bocian, Inorg.Chem. 1995, 34, 2201–2206.

326. P. A. Connick, K. A. Macor, Inorg. Chem.1991, 30, 4654–4663.

327. K. M. Kadish, D. Sazou, Y. M. Liu, A. Saoi-abi, M. Ferhat, R. Guilard, Inorg. Chem.1988, 27, 686–690.

328. K. M. Kadish, D. Sazou, G. B.Maiya, B. C. Han, Y. H. Liu, A. Saoiabi,M. Ferhat, R. Guilard, Inorg. Chem. 1989,28, 2542–2547.

329. C. Y. Lin, S. Hu, T. Rush, III, T. G. Spiro,J. Am. Chem. Soc. 1996, 118, 9452, 9453.

330. A. M. Stolzenberg, M. T. Stershic, Inorg.Chem. 1987, 26, 3082, 3083.

331. A. M. Stolzenberg, M. T. Stershic, J. Am.Chem. Soc. 1988, 110, 6391–6402.

332. A. M. Stolzenberg, M. T. Stershic, J. Am.Chem. Soc. 1988, 110, 5397–5403.

333. M. W. Renner, L. R. Furenlid, K. M. Barki-gia, A. Forman, H. K. Shim, D. J. Simpson,K. M. Smith, J. Fajer, J. Am. Chem. Soc.1991, 113, 6891–6898.

334. K. M. Kadish, D. Sazou, Y. H. Liu, A. Saoi-abi, M. Ferhat, R. Guilard, Inorg. Chem.1988, 27, 1198–1204.

335. D. Chang, T. Malinski, A. Ulman, K. M.Kadish, Inorg. Chem. 1984, 23, 817–824.

336. K. Ozette, P. Leduc, M. Palacio, J. Bartoli, K.M. Barkigia, J. Fajer, P. Battioni, D. Mansuy,J. Am. Chem. Soc. 1997, 119, 6442, 6443.

337. K. M. Kadish, X. Q. Lin, J. Q. Ding,Y. T. Wu, C. Araullo, Inorg. Chem. 1986, 25,3236–3242.

338. S. E. Jones, H. N. Po, Inorg. Chim. Acta1980, 42, 95–99.


339. A. Giraudeau, A. Louati, H. J. Callot,M. Gross, Inorg. Chem. 1981, 20, 769–772.

340. A. Kumar, P. Neta, J. Phys. Chem. 1981, 85,2830.

341. M. Krishnamurthy, Inorg. Chem. 1978, 17,2242–2245.

342. A. Giraudeau, A. Louati, M. Gross, H. J.Callot, L. K. Hanson, R. K. Rhodes, K. M.Kadish, Inorg. Chem. 1982, 21, 1581–1586.

343. D. Kuila, A. B. Kopelove, D. K. Lavallee, In-org. Chem. 1985, 24, 1443–1446.

344. M. W. Renner, K. M. Barkigia, Y. Zhang,C. J. Medforth, K. M. Smith, J. Fajer, J. Am.Chem. Soc. 1994, 116, 8582–8592.

345. S. A. Sibilia, S. Hu, C. Piffat, D. Melamed,T. G. Spiro, Inorg. Chem. 1997, 36,1013–1019.

346. J. Takeda, M. Sato, Chem. Lett. 1995, 939,940.

347. (a) M. E. Jamin, R. T. Iwamoto, Inorg. Chim.Acta 1978, 27, 135–143. (b) K. M. Kadish,Wenbo E., Z. Ou, J. Shao, P. J. Sintic,K. Ohkubo, S. Fukuzumi, M. J. Crossley, J.Chem. Soc. Dalton, Chem. Commun, 2002,356–357.

348. J.-H. Fuhrhop, D. Mauzerall, J. Am. Chem.Soc. 1969, 91, 4174–4181.

349. J. Fajer, D. C. Borg, A. Forman, D. Dolphin,R. H. Felton, J. Am. Chem. Soc. 1970, 92,3451–3459.

350. R. Felton, D. Dolphin, D. C. Borg, J. Fajer,J. Am. Chem. Soc. 1969, 91, 196–198.

351. K. M. Kadish, L. R. Shiue, R. K. Rhodes, In-org. Chem. 1981, 20, 1274–1277.

352. K. M. Kadish, L. R. Shiue, Inorg. Chem.1982, 21, 3623–3630.

353. K. M. Kadish, A. Tabard, A. Zrineh, M.Ferhat, R. Guilard, Inorg. Chem. 1987, 26,2459–2466.

354. K. M. Kadish, J.-L. Cornillon, A. Coutso-lelos, R. Guilard, Inorg. Chem. 1987, 26,4167–4173.

355. K. M. Kadish, B. Boisselier-Cocolios,A. Coutsolelos, P. Mitaine, R. Guilard,Inorg. Chem. 1985, 24, 4521–4528.

356. K. M. Kadish, J.-L. Cornillon, P. Cocolios,Inorg. Chem. 1985, 24, 3645–3649.

357. R. Guilard, I. Perrot, A. Tabard, P. Richard,C. Lecomte, Y. H. Liu, K. M. Kadish, Inorg.Chem. 1991, 30, 27–37.

358. R. Guilard, N. Jagerovic, A. Tabard, C. Nail-lon, K. M. Kadish, J. Chem. Soc., DaltonTrans. 1992, 1957–1966.

359. R. Guilard, N. Jagerovic, A. Tabard, P. Rich-ard, L. Courthaudon, A. Louati, C. Lecomte,K. M. Kadish, Inorg. Chem. 1991, 30, 16–27.

360. R. Guilard, A. Zrineh, A. Tabard, A. Endo,B. C. Han, C. Lecomte, M. Souhassou, A.Habbou, M. Ferhat, K. M. Kadish, Inorg.Chem. 1990, 29, 4476–4482.

361. K. M. Kadish, Q. Y. Xu, J.-M. Barbe, Inorg.Chem. 1987, 26, 2565, 2566.

362. K. M. Kadish, Q. Y. Xu, J.-M. Barbe, J. E.Anderson, E. Wang, R. Guilard, Inorg.Chem. 1988, 27, 691–696.

363. R. Guilard, C. Ratti, J.-M. Barbe, D. Dubois,K. M. Kadish, Inorg. Chem. 1991, 30,1537–1542.

364. K. M. Kadish, Q. Y. Xu, G. B. Maiya, J.-M. Barbe, R. Guilard, J. Chem. Soc., DaltonTrans. 1989, 1531–1536.

365. K. M. Kadish, D. Dubois, S. Koeller, J.-M.Barbe, R. Guilard, Inorg. Chem. 1992, 31,3292–3294.

366. K. M. Kadish, D. Dubois, J.-M Barbe, R. Gui-lard, Inorg. Chem. 1991, 30, 4498–4501.

367. J. A. Ferguson, T. J. Meyer, D. G. Whitten,Inorg. Chem. 1972, 11, 2767–2772.

368. C. R. Lorenz, H. D. Dewald, F. R. Lemke,J. Electroanal. Chem. 1996, 415, 179–181.

369. K. M. Kadish, D. G. Davis, Ann. N. Y. Acad.Sci. 1973, 206, 495–503.

370. K. M. Kadish, B. Boisselier-Cocolios,C. Swistak, J.-M. Barbe, R. Guilard, Inorg.Chem. 1986, 25, 121–122.

371. K. M. Kadish, C. Swistak, B. Boisselier-Cocolios, J.-M. Barbe, R. Guilard, Inorg.Chem. 1986, 25, 4336–4343.

372. R. Guilard, J.-M. Barbe, M. Fahim,A. Atmani, G. Moninot, K. M. Kadish, NewJ. Chem. 1992, 16, 815–820.

373. K. M. Kadish, M. Autret, Z. P. Ou, K. Akiba,S. Masumoto, R. Wada, Y. Yamamoto, In-org. Chem. 1996, 35, 5564–5569.

374. Y. H. Liu, M. F. Benassy, S. Chojnacki,F. D’souza, T. Barbour, W. J. Belcher, P. J.Brothers, K. M. Kadish, Inorg. Chem. 1994,33, 4480–4484.

375. T. Barbour, W. J. Belcher, P. J. Brothers,C. E. F. Rickard, D. C. Ware, Inorg. Chem.1992, 31, 746–754.

376. C. A. Marrese, C. J. Carrano, J. Chem. Soc.,Chem. Commun. 1983, 1279.

377. S. Mangani, E. F. Meyer, D. F. Cullen,M. Tsutsui, C. Carrano, Inorg. Chem. 1983,22, 400.


378. C. A. Marrese, C. J. Carrano, J. Chem. Soc.,Chem. Commun. 1982, 1279.

379. C. A. Marrese, C. J. Carrano, Inorg. Chem.1983, 22, 1858–1862.

380. K.-y. Akiba, R. Nadano, W. Satoh, Y. Yama-moto, S. Nagase, Z. Ou, X. Tan, K. M.Kadish, Inorg. Chem., 2001, 40, 5553–5567.

381. K. M. Kadish, Z. Ou, X. Tan, W. Satoh,Y. Yamamoto, K. Akiba, manuscript inpreparation.

382. L. Michaudet, D. Fasseur, R. Guilard, J. Por-phyrins Phthalocyanines 2000, 4, 261–270.

383. T. A. Rao, B. G. Maiya, Inorg. Chem. 1996,35, 4829–4836.

229

7Electrochemical Measurementsof Nitric Oxide in BiologicalSystems

Tadeusz MalinskiOhio University, Athens, Ohio

7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2317.1.1 The Role of Nitric Oxide in Biological Systems . . . . . . . . . . . . . . . 2317.1.1.1 NO Release in Biological Systems . . . . . . . . . . . . . . . . . . . . . . . . 2317.1.1.2 NO as a Regulator of the Cardiovascular System . . . . . . . . . . . . . . 2337.1.1.3 NO in the Nervous System . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2347.1.1.4 NO as a Part of the Immune System . . . . . . . . . . . . . . . . . . . . . . 2347.1.1.5 Pathology of NO Release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235

7.2 Electrochemical Methods for NO Detection . . . . . . . . . . . . . . . . . 2357.2.1 Electrochemical Oxidation of NO . . . . . . . . . . . . . . . . . . . . . . . . 2357.2.2 Preparation of Porphyrinic Sensor . . . . . . . . . . . . . . . . . . . . . . . 2387.2.2.1 Operation Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2427.2.3 Clark Probe for NO Detection . . . . . . . . . . . . . . . . . . . . . . . . . . 243

7.3 Measurement of NO in Biological Systems . . . . . . . . . . . . . . . . . . 2447.3.1 Measurement of NO in a Single Cell . . . . . . . . . . . . . . . . . . . . . . 2447.3.2 In vitro Measurement of NO in Tissue . . . . . . . . . . . . . . . . . . . . 2467.3.3 In vivo Measurements of NO in the Kidney . . . . . . . . . . . . . . . . . 2487.3.4 In vivo Measurements of NO in the Beating Heart . . . . . . . . . . . . 2497.3.5 Measurement of NO in the Brain . . . . . . . . . . . . . . . . . . . . . . . . 2527.3.5.1 In vitro Measurement with Clark Probe . . . . . . . . . . . . . . . . . . . . 2527.3.5.2 In vivo Measurement with Porphyrinic Sensor . . . . . . . . . . . . . . . 2527.3.6 Measurements of NO in Human Beings . . . . . . . . . . . . . . . . . . . 255

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

231

7.1Introduction

7.1.1The Role of Nitric Oxide in BiologicalSystems

Nitric oxide (NO) has been implicated inthe pathogenesis of several diseases. A de-ficiency of NO may play a role in somediseases (hypertension, hyperglycemia,atherosclerosis, Parkinson’s disease andAlzheimer’s disease) whereas an increaseof NO may participate in others (arthritis,reperfusion injury, cancer) [1–11]. Thus,from a biochemical as well as a medicalperspective, it is important to quantify thedetails of NO production in abnormal andnormal tissues, including direct measure-ments. NO is stable in oxygen-free andcell-free solutions. However, NO reactsrapidly with several components in vitroor in vivo, producing protein nitrosylationas well as reacting with hemoglobin andoxygen. In the presence of superoxide, NOis rapidly converted to peroxynitrite. Con-sequently, NO has a half-life of two to sixseconds in vivo, and detection of NO inbiological systems has been proven to betechnically difficult [12, 13].

The currently used instrumentaltechniques for NO measurementsare spectroscopic and electrochemical

methods [14, 15]. Mass spectrometry andgas chromatography have been usedoccasionally for NO detection but are muchless sensitive. Spectroscopic methodsare based on the detection of theproducts of NO oxidation: NO2

− orNO3

− (UV-visible spectroscopy, electronspin resonance (ESR) spectroscopy, orspin trapping). Electrochemical methodsinclude voltammetry, amperometry, andcoulometry. Electrochemical methodsoffer several features that are notavailable from analytical spectroscopicmethods. Most important is the capabilityafforded by the use of microelectrodesfor direct in situ measurements ofNO in single cells near the sourceof NO synthesis (nitric oxide synthaseNOS) [16–18]. Electrochemical methodsdetect NO directly and are based onelectron exchange between NO and anelectrode. All spectroscopic methods detectNO indirectly. Therefore, electrochemicalmethods are more suited than thespectroscopic methods for in situ, invitro, and in vivo monitoring of NOconcentration in biological systems.

7.1.1.1 NO Release in Biological SystemsBefore 1987, NO biosynthesis was thoughtto be restricted to bacteria engaged innitrification or denitrification reactions.However, we have recently learned that NO

232 7 Electrochemical Measurements of Nitric Oxide in Biological Systems

can be released by many different cells inmammalian systems [19–21]. NO can playdifferent physiological roles, dependingon the place of its release. It can bea neurotransmitter when (for seconds)a small puff of it is generated by theneurons of central and peripheral nervoussystem. It can regulate blood pressureand inhibit blood coagulation when (forminutes) a larger burst of it is generatedin the endothelium, a monolayer of cellslining the cardiovascular system. Also,NO can act as a cytostatic agent when(for hours) a continuous blast of it isbiosynthesized by the immune system;its presence may halt the proliferation ofcancer and pathogens [21].

NOS is the enzyme that biosynthe-sizes the cosubstrates L-arginine and O2into the coproducts NO and L-citrulline(Fig. 1) [22]. NOS generates NO by catalyz-ing the oxidation of a guanidine nitrogenon the amino acid L-arginine. The en-zyme is stereospecific for the L-isomer, asD-arginine is not a substrate. This pro-cess is an overall five-electron oxidationof nitrogen with little or no chemical orbiochemical precedence. Surprisingly, thismultistep, odd-electron transfer is carriedout by a single protein. There are two con-stitutive nitric oxide synthase (cNOS) iso-forms, neuronal (nNOS), and endothelial(eNOS), and one inducible NOS isoform(iNOS). All three iso-enzymes are homod-imers. Each monomer has a molecularweight ranging from 130 000 to 150 000

daltons, containing four prosthetic groups:flavin-adenine dinucleotide (FAD): flavinmononucleotide (FMN), (6R)-5,6,7,8-tetra-hydrobiopterin (BH4), and iron proto-porphyrin IX (Heme) [23]. The turnoverrate of NO production is about sevenmolecules of NO per eNOS monomer andabout forty molecules of NO per nNOS.Since NO is hydrophobic (solubility inwater is only 1.82 mmol/L) and some-what lipophilic (Kow > 6.5 at 37 C), itdiffuses rapidly through the hydropho-bic environment of cell membranes likeO2 and N2. In the aqueous phase ofthe cytoplasm, the diffusion coefficient ofNO is 3.6 × 10−5 cm2 s−1 [24]. Biosynthe-sized within the cell, NO may react witha select few types of molecules insidethe cell, or outside (after free diffusionthrough the cell membrane). The mostrapid scavenger of NO is superoxide (O2

−,k = 6.7 × 109 M−1 s−1). The peroxynitrite(OONO−) that is formed in the reactionbetween O2

− and NO is quite stable, butwhen protonated (pKa = 6.8), usually itquickly rearranges to H+ and NO3

−. Re-action of NO with O2 is much slower andleads to production of NO+/NO2

−, fol-lowed by further oxidation to NO3

−. NOmay also react with a few metal ions (iron,copper or manganese), which are usuallybound to proteins. The selective reactivityof NO with such proteins and its reactionwith O2

− and O2 dominate the chemistryof NO in biological systems.

NOS

O2

L-Arginine

L-Citruline

NO

NADPH

BH4

Fig. 1 NOS catalyzedproduction of NO fromL-arginine. Tetrahydrobiopterin(BH4) is an essential cofactor;NADPH acts as the source ofelectrons for oxygenreduction-activation.


To date, a cNOS and an iNOS have beenfound in many cell types in various partsof the body (Table 1). The monolayer ofcells lining the cardiovascular system (theendothelium) contains mainly the eNOSand produces the largest amount of NO.The endothelium should be consideredone of the largest specialized organs inthe body; the total weight of eNOS cellsin the human body is about 1.5 kg and iscomparable with the weight of the liver. Inthe brain, NO is produced by both nNOSand eNOS. nNOS is also found in myocytesand in the skeletal muscles. NO can alsoact as a cytostatic agent in the immunesystem. NO from iNOS can be produced(after induction) by almost every cell inthe human body, where it plays a primaryrole in host defense. In this role, NO maydefend the body against invading bacteria,viruses, and even cancer.

7.1.1.2 NO as a Regulator of theCardiovascular SystemVascular endothelial cells contain calcium-dependent constitutive eNOS. In order tomaintain normal blood pressure, eNOSsynthesizes NO in bursts lasting a few

Tab. 1 NO synthases

Constitutive NOS Inducible NOS

Vascularendothelium

Vascularendothelium

Brain Vascular smoothmuscle

Platelets MacrophagesAdrenal glands Kupffer cellsPeripheral nerves HepatocytesMast cells EndocardiumMasangial cells Masangial cellsMyocardium LymphocytesMegakaryocytes Chondrocytes

FibroblastsNeutrophilsMegakaryocytes

minutes. The synthesis of NO is stimulatedby chemical agonists such as bradykinin,acetylcholine, ATP, and several otheragents that stimulate Ca2+ flux. However,physical agonists such as shear stress, flow,electrical current, light, and electromag-netic fields can also stimulate NO releasein the cardiovascular system by causingthe release of Ca2+ from internal storesof cytoplasm or by opening ion channelsto allow the relatively large extracellularCa2+ concentration into the cell [25]. Af-ter eNOS is turned on by Ca2+ flux andbiosynthesizes NO for about a minute, itis turned off by the phosphorylation of oneof its serine residues. The production ofNO from eNOS is also inhibited by NOitself [26].

NO synthesized by eNOS cells, diffusesout in all directions. About 80 to 90%of NO released by the endotheliumin the cardiovascular system is washedaway by the blood, where it is usedto prevent platelet aggregation and thesubsequent formation of blood clots. Theremaining amount of NO diffuses tothe wall (smooth muscle) of arteriesand veins and triggers a cascade ofevents leading to the production of cyclicGMP and smooth muscle relaxation.Relaxation of the smooth muscle allowsthe blood vessel to dilate (increase ofvessel diameter), resulting in loweredblood pressure. The cardiovascular systemmaintains a constant level of NO at agiven blood flow [27]. When blood flowincreases, the endothelium releases moreNO to maintain its constant concentrationin the blood stream. When this normallevel is not produced, either becauseproduction is blocked by administrationof eNOS inhibitors or by pathologicalstates such as deposition of cholesterol onthe wall of the arteries (atherosclerosis),the vascular muscles do not relax to the


appropriate degree, and vasoconstrictionresults [28, 29]. Vasoconstriction increasesblood pressure, decreases flow, and isresponsible for hypertension [30].

Platelets in the blood can also releaseNO [26]. One platelet can produce about10−17 moles of NO in a single burst last-ing a few minutes, when all its eNOS areactivated. NO released by platelets preventsblood coagulation, formation of thrombi,and subsequent blockage of arteries. Thepathology of this process leads to coro-nary thrombosis and is a major cause ofstroke. The heart itself releases a signif-icant amount of NO, on a beat-to-beatbasis, during the systolic (compression)as well as the diastolic (decompression)period [10]. Under normal conditions NOis produced in the beating heart by eNOScells, which are always located in closeproximity 10 to 20 µm from cardiac musclecells (myocytes). The basal concentrationof NO in the heart is maintained at a rela-tively high level about 400 to 500 nmol/L,which changes gradually with the chang-ing intensity of preload forces on theheart. Without the mechanical transduc-tion of NO production, cardiac outputwould change abruptly as preload changedabruptly. Under conditions when heart hasto provide extensive work (pumping bloodunder heavy stress, exercise, etc.), NO isproduced by myocytes in addition to eNOScells. Myocyte NO production is not stim-ulated by mechanical forces of the heartbut by norepinephrine [27].

7.1.1.3 NO in the Nervous SystemnNOS, like cNOS, is turned on by Ca2+flux, but only produces NO for a fewseconds before being shut off by thebinding of NO to the heme in the activesite; the nNOS then rests for a few minutesuntil the bound NO disassociates. NOgenerated by nNOS in certain neurons

in the peripheral nervous system servesas a neurotransmitter to help control thecardiovascular, respiratory and digestivesystems [31–35].

In the central nervous system, NO acts asa neurotransmitter in the cerebellum. Itsaction resembles the interaction betweeneNOS cells and smooth muscle cells in ablood vessel. NO released from the postsy-naptic neurons (in response to activation ofreceptor molecule N -methyl-D-aspartate)stimulates neighboring neurons to pro-duce cyclic GMP. NO has also been im-plicated in another part of the brain – thehippocampus, which is involved in learn-ing and the formation of memory. NOappears to be essential for establishinglong-term potentiation memory but notshort-term potentiation memory. In long-term potentiation, the strength of synapticcontact increases as a consequence of thefrequent use of memory. Some recent find-ings support the possibility that NO can actas a retrograde messenger in this process.The inhibition of NO release in hippocam-pus prevents long-term potentiation anddecreases learning capabilities.

7.1.1.4 NO as a Part of the ImmuneSystemThe role of NO in the immune system ismuch different from its role in the car-diovascular or the nNOS system [36–42].In humans, almost every type of cell inthe body can express iNOS. In contrast tonNOS and eNOS, the iNOS is calcium-independent. The signal to translate theDNA sequence for the iNOS into theamino acid sequence for the iNOS enzymecomes from certain cytokines, which areproduced by the infected cells. Any kind ofinfection (including bacterium, viruses, orcancer) will lead to the production of cy-tokines. Cytokines carry the message of theinfectious state to the surrounding cells,

7.2 Electrochemical Methods for NO Detection 235

which will start to produce iNOS. Imme-diately after translation is complete andprosthetic groups are in place, iNOS willcontinuously produce large amounts ofNO for an extended period (several hours).The total NO production will therefore bemuch higher than that produced by eNOS.

iNOS produces a sufficient concentra-tion of NO to locally inhibit ribonucleotidereductase, the enzyme that converts ri-bonucleotides to the deoxyribonucleotidesnecessary for DNA synthesis, hence theprofound cytostatic effect of NO on the pro-liferation of rapidly dividing tumor cellsor pathogens. In addition, DNA synthe-sis is a fundamental step in normal cellproliferation. Even normally high NO con-centrations from cNOS can inhibit theribonucleotide reductase, thus halting theproliferation of smooth muscles aroundmajor arteries and cardiac myocytes. NOis not toxic even at higher biological con-centrations. Therefore, it is unlikely thatNO can actually kill tumor cells but justlimit their proliferation.

7.1.1.5 Pathology of NO ReleaseA pathology of NO production can leadto many diseases (Table 2). In most

Tab. 2 Pathology of NO release

NO concentration

Too low Too high

Hypertension Septic shockAtherosclerosis HypotensionDiabetes Excessive bleedingIschemia (stroke, heart

attack)Meningitis

Alzheimer’s disease Rheumatoid arthritisParkinson’s diseaseFibrosisCancer

life-threatening diseases such as hyper-tension, atherosclerosis, and diabetes, thenet concentration of NO is lower than ina healthy system. This does not neces-sarily mean that the expression of eNOSor nNOS is lower. In some of thesediseases, eNOS expression is even higher(hypertension).

7.2Electrochemical Methods for NO Detection

7.2.1Electrochemical Oxidation of NO

Electrochemical methods for NO determi-nation offer several features that are notavailable with spectroscopic approaches.Perhaps the most important is the capabil-ity of microelectrodes to directly measureNO in single cells in situ, in close proximityto the source of NO generation. Figure 2shows sensors that have been developedfor the electrochemical measurement ofNO. One is based on the electrochemicaloxidation of NO on a platinum electrode(the classical Clark probe for detection ofoxygen) and operates in the amperometricmode [17]. The other is based on the elec-trochemical oxidation of NO on conductivepolymeric porphyrin (porphyrinic sen-sor) [24]. The Clark probe uses a platinumwire as a working electrode (anode) and asilver wire serves as the counterelectrode(cathode). The electrodes are mounted ina capillary tube filled with a sodium chlo-ride/hydrochloric acid solution separatedfrom the analyte by a gas-permeable mem-brane. A constant potential of 0.9 V isapplied, and direct current (analytical sig-nal) is measured from the electrochemicaloxidation of NO on the platinum anode. Inthe porphyrinic sensor, NO is catalyticallyoxidized on a polymeric metalloporphyrin


Carbon fiber

Conductivepolymeric porphyrin e

NO Anions(ascorbic,NO2

−, etc)

Copper lead

Silver epoxy

Ag wirePt wire

Glass capilary

Gas permeablemembrane

NO

NO

NO2−

(a) (b)

e

Nafion NO→NO+

NO+

Fig. 2 Schematic diagram of the porphyrinic sensor (counterelectrode is not shown) (a) andClark-type probes, as adopted for determination of NO (b).

(n-type semiconductor) on which the oxi-dation reaction occurs at 610 mV (versusa saturated calomel electrode (SCE)), thatis, about 200 mV lower than the potentialrequired for NO oxidation on platinum orcarbon electrodes. The current efficiencyfor the reaction on porphyrinic film ismuch higher than on platinum or carbon,even at the physiological pH of 7.4.

Generally, the oxidation of NO on solidelectrodes proceeds via ‘‘EC mechanisms’’:electrochemical oxidation (1) followed bya chemical reaction (2) [24, 43, 44]. Thefirst electrochemical step is a one-electrontransfer from an NO molecule to the elec-trode, resulting in the formation of NO+

NO −−−→ NO+ (1)

NO+ is a relatively strong Lewis acid and,in the presence of OH−, is converted to ni-trite (NO2

−) in the fast, highly irreversiblechemical step:

NO+ + OH− −−−→ HNO2 (2)

Since the oxidation potential of nitrite inan aqueous solution is only 60 to 80 mV,more positive than that of NO, oxidation of

NO on solid electrodes utilizing scannedpotentials, that is, with electrodes oper-ated in the voltammetric mode, results infurther oxidation of nitrite to nitrate, a pro-cess in which two additional electrons aretransferred. Currents resulting from theone-electron oxidation of NO to NO2

−, andtwo-electron oxidation of NO2

− to NO3−,

are highly overlapped.Using a bare metal, carbon, or porphyrin

electrode, it is impossible to differentiatenitrite produced electrochemically thanfrom that derived from chemical NOoxidation. Therefore, a barrier has to beplaced between the electrode and theanalyte to prevent access of nitrite to theelectrode surface. This has been achievedin the porphyrinic sensor by depositing alayer of cation exchanger (Nafion), whichrepels negatively charged species suchas NO2

− (Fig. 2a). Nafion also preventsthe formation of NO2

− in the film afterone electron oxidation of NO to NO+. Inthe Clark-type probe, the analyte solutioncontaining NO and nitrite are separatedfrom the inner electrode solution by a gas-permeable membrane.


Oxidation of NO on classical conductivematerials such as noble metals (platinum,gold, etc.) or carbon, which are usedas electrodes, produces a relatively lowcurrent at neutral pH. This is due to astrong absorption of NO to the electrodesurface and a slow rate of electron transferbetween NO and the electrode. Typicaldifferential pulse voltammograms (DPV)of NO on carbon fiber covered withNafion, and carbon fiber covered withporphyrinic film and Nafion are shown inFig. 3. There is about a 190 mV differencebetween the oxidation potential of NOon carbon fiber and porphyrinic film. Aconcentration of 0.1-µM NO produces avery small current on the carbon fiberelectrode operating in the DPV mode(Fig. 3a). However, this same carbon fibercovered with a layer of polymeric porphyrinproduces a much larger current (Fig. 3b)for NO oxidation. The current generatedon polymeric porphyrin is mass transportcontrolled and is linearly proportional tothe concentration of NO. The linearity isobserved over four orders of magnitude ofNO concentration [45].

The sensitivity and selectivity of the poly-meric porphyrin depends not only on thepotential of NO to oxidize but also on thefast process of electrochemical NO oxida-tion, which generates the high current.In addition, surface effects, axial ligationto the central metal in the porphyrin, na-ture of the central metal (iron ≥ nickel >

cobalt zinc, copper), gas permeabilitythrough sensor layer(s), and fast removalof NO+ by Nafion are all important inpromoting fast and selective oxidation

Fig. 3 DPV’s obtained from oxidationof 0.1-µM NO on a carbon fiberelectrode covered with Nafion andcarbon fiber electrode covered withpolymeric (TMHPP)Ni porphyrin andNafion.

NO

EP=0.63 V

(a)

(b)

0.1 nA

0.40 0.60 0.80 1.00

Potential[V]

NO

EP=0.82 V

of NO. Typical current density for NOoxidation on a porphyrinic sensor is 0.3 to1.8 mA cm−2 µM−1, and is at least four toseven times higher than that which can beobtained on activated carbon fiber coveredwith Nafion. Current density for NO oxi-dation on the porphyrinic sensor dependson film quality, and should be at least0.4 mA cm−2 µM−1 for measurements inbiological environments. Sensors based onhigh-quality polymeric films will show cur-rent densities of 1.5–1.8 mA cm−2 µM−1

and a detection limit approaching 10−9 MNO. A current density comparable to thatof carbon fiber indicates poor quality (highimpedance) of the porphyrinic film. More-over, the design of the porphyrinic sensorresults in a high selectivity for NO becausethe current generated by NO oxidation isusually several orders of magnitude higherthan the current produced by other speciesoxidized at a similar potential. A typicalexample of such a potential interferantis dopamine, which can be oxidized ona carbon fiber covered with Nafion at a


0.40 0.50 0.60 0.70 0.80 0.90

(a)

(b)

(c)

100 pA

DopamineNO

Potential[V]

Fig. 4 DPV’s of 50-nM NOsolution recorded in theabsence (a) and in the presenceof 1 µM of dopamine (b) and0.2 mM of dopamine (c).

potential of 0.56 V. However, oxidation ofdopamine on a porphyrinic sensor occursat 0.80 V with the current density, whichis about four orders of magnitude lowerthan that observed for the oxidation of NO(Fig. 4). Dopamine at a concentration of0.1 µM produces no measurable current,and at 0.2 mM produces the current com-parable to that produced by the oxidationof 50 nM NO.

7.2.2Preparation of Porphyrinic Sensor

Microsensors are produced by threadinga carbon fiber (diameter 7 µm) throughthe pulled end of a capillary tube withabout 1 cm left protruding. Nonconductiveepoxy is put at the glass–fiber interface.After the epoxy cement drawn into thetip of the capillary has cured, the car-bon fiber is sealed in place. The carbonfiber is then sharpened by gradual burning(propane-air microburner, 1300–1400).The sharpened fiber is immersed in melt-ing wax-rosin (5 : 1) at a controlled temper-ature for 5 to 15 sec, and after cooling issharpened again. The flame temperature

and the distance of the fiber from the flameneed to be carefully controlled. The result-ing electrode is a slim cylinder with a smalldiameter (0.5–2 µm) rather than a short ta-per, a geometry that aids in implantationand increases the active surface area. Thetip (length 2–6 µm) is the only electroac-tive part of the carbon fiber. For the sensorto be implanted into a cell, this lengthmust be less than the cell thickness. Theunsharpened end of the fiber is attachedto a copper wire lead with silver epoxycement. The sharpened tip of the carbonfiber is covered with a film of conductivepolymeric film (Fig. 5).

Molecules that undergo fast randompolymerization on the electrode surface,like tetraphenylporphyrins with aminoor pyrrole substituents on the phenylring, will form films with a well-developed surface but poor catalytic prop-erties for NO oxidation and low elec-trical conductivity. Out of more thanforty different porphyrinic and nonpor-phyrinic and organic and inorganic elec-trocatalysts tested for these properties,only three show the desired characteris-tics: tetrakis (3-methoxy-4-hydroxyphenyl)


(a) (b)20 µm5 µm

Fig. 5 Scanning electron micrograph of a thin film of polymeric (TMHPP)Ni (a) and the tipof a thermally pointed carbon fiber sensor covered with polymeric (TMHPP)Ni and Nafionfilm (b).

porphyrin (TMHPP), meso-tri (N-methyl-4-pyridinium)-p-phenylene-5′-O-2′, 3′-O-isopropylideneuridine-porphyrin (PUP)and N, N′-bis[5-p-phenylene-10,15,20-tris(3-methoxy-4-hydroxyphenyl)-porphy-rin]-1,10-phenanthroline-4,7-diamide (H2

(1,10-phen)(TMHPP)2), with Ni(II), Co(II)or Fe(II) as central metals [46–51]. Thecontinuous-scan cyclic voltammogramsshowing formation of the films from(TMHPP)Ni and (PUP)Co are depictedin Fig. 6. The detection limit for NO de-pends on film quality and varies between5 × 10−8 and 10−9M. The best sensitiv-ity is usually obtained using a PUP film.However, the stability of TMHPP films orH2(1,10-phen)(TMHPP)2 film (about 60days) is much better than the stability ofPUP films (about 7 days).

As described previously, the porphyrinicfilm is covered with Nafion. This serves

to attract NO+ from the underlyingpolymeric porphyrin surface and preventsits further oxidation to NO2

−, as well asto prevent access of NO2

− and other com-mon biological anions such as ascorbateto the porphyrinic film. The current dueto NO oxidation on polymeric porphyrininitially increases with thin Nafion cover-age. However, thicker films (>1 µm) causethe NO current to decrease. In order totest the integrity of the Nafion film, thecurrent due to the Ni(II)-Ni(III) reactioncan be measured. Oxidation of Ni(II) toNi(III) in the film occurs only if diffusionof the OH− counterion into the film oc-curs. Because a Nafion film with sufficientthickness (without pinholes) will preventOH− diffusion, a lack of Ni(II)-Ni(III) con-version in the voltammogram (obtainedin 0.1 M NaOH) is a good method forverifying film integrity. Figure 7 shows


(a)

Ia

Ic

Cur

rent

−0.2 0.5 1.2 −1.0 0.0 1.2

(b)Potential

[V]Potential

[V]

10 nA

Fig. 6 Continuous-scan cyclic voltammograms obtained by a polymerization of(TMHPP)Ni (a) and (PUP)Co (b) on a carbon electrode.

(a) (b) (c) (d)

Cur

rent

100 nA

Potential[V]

0.0 0.7

Fig. 7 Cyclic voltammograms of Ni(II)/Ni(III) couplerecorded at the different coverage of (TMHPP)Ni film withNafion.

cyclic voltammograms obtained for thissame poly-(TMHPP)Ni film covered withNafion films of different thickness, aftersequential coating by dipping the carbonfiber for 5 s into a 1.25% ethanolic solutionof Nafion. Each coat of Nafion decreases

Ni(II) and Ni(III) peaks by a factor of aboutthree. The thickness of Nafion obtained af-ter three coatings (total 15s) is sufficient toprevent diffusion of high concentrationsof small anions. Sensors have to be furthertested for their response to nitrite in a


buffer solution at pH 7.4. A lack of re-sponse to 20-µM NO2

−, a concentration 1to 2 orders of magnitude higher than thatexpected in plasma, indicates that Nafioncoating is sufficient for NO measurementsin this medium. It should be noted thatNO2

− concentrations of ≥10 µM are foundin airway lining fluid and saliva, represent-ing a potential problem. Nafion coatings,therefore, may need to be individualizedfor different biological systems.

Multiple porphyrinic electrodes (an elec-trode array) for studies of NO release inlarge cells and tissues can be constructed inthe disk form using glass microcapillaries.A bundle of 2 to 6 fibers mounted in a glasscapillary in a similar way to that describedfor a single fiber can be effective in manylarge cell or tissue experiments [6]. The useof large electrodes (diameter > 100 µm)does not improve the performance of thesensor. This is due to an increase of thebackground current (capacity current), andthe inability to place the sensor at the sitewhere the highest concentration of NO isto be expected (the cell membrane). Thislast limitation can be overcome by usinga large surface sensor on which cells can

be grown directly. In this case, the sensorconsists of vitreous carbon support coveredwith a polymeric film and Nafion (Fig. 8).The detection limit of this sensor is simi-lar to that obtained from the carbon fibersensor. This sensor provides a convenientway to measure NO release in cell culturesunder various conditions.

Carbon fiber–based sensors are toofragile to be used in deep-tissue mea-surements. However, these sensors can beprotected by insertion into an intravenouscatheter, which can be implanted in tissuesor blood vessels. To make the catheter-protected sensor, a bundle of (5–7) carbonfibers are mounted inside a truncatedneedle (Fig. 9). After curing, the shaftof the truncated needle is coated with anonconducting epoxy. Then, conductiveporphyrinic film and Nafion are depositedon the protruding carbon fiber. A punctureneedle is used to place a Teflon catheterwith a perforated tip deep inside the tissueof a blood vessel. The Teflon catheter’s po-sition is secured and the puncture needleis removed. Then, an NO sensor mountedat the end of a truncated needle shaft is in-serted into the hollow Teflon catheter. The

Fig. 8 Schematic diagram of aporphyrinic sensor adapted to detectNO released from cells grown directlyon its surface (a); photograph of eNOScells (human umbilical vein eNOS cells)grown directly on a porphyrinicsensor (b).

RVC

Human endothelial cells

Cells

Nafion

TMHPPNifilm(a)

(b)

EpoxidesInsulatedwire

RVC


Needle

Catheter

Porphyrinic sensor Copper wire

Gold crimp pinEpoxy insulation

10-µm pores

Truncated needle

Fig. 9 Schematic diagram of an intravenous catheter–protected porphyrinic sensor for invivo measurement.

NO sensor should be just short enough tofit inside the perforated catheter and pro-tect the NO sensor tip from mechanicaldamage.

7.2.2.1 Operation ModesDPV or chronoamperometry can be usedto monitor analytical signals (current is lin-early proportional to NO concentration).In DPV, a potential modulated with 40-mV rectangular pulses is linearly scanned

0 1 2 3Time

[s]

(a)

(b)

(c)

Cur

rent

NO Fig. 10 Amperometric curves showingporphyrinic sensor responses to NO inhomogenous static solution (a),homogenous flowing solution, and aheterogeneous flow solution (blood).

from 0.4 to 0.8 V. The resulting voltammo-gram (alternating current versus voltageplot), contains a peak due to NO oxidation.The peak current should be observed at apotential of 0.63 to 0.67 V (using 40-mVpulse amplitude), which is the charac-teristic potential for NO oxidation on aporphyrinic-Nafion sensor. In chronoam-perometric measurements, a constant po-tential between 0.65 and 0.70 V is keptconstant, and a plot of current versus timeis recorded. This amperometric method(with a response time better than 0.1 ms),provides rapid quantitative response tominute changes in NO concentration.DPV, which also provides quantitative in-formation but requires approximately 5 to40 s for the voltammogram to be recorded,is used mainly for qualitative analysis(i.e. verification of specificity). Severalother electroanalytic techniques, includingnormal pulse voltammetry, square wavevoltammetry, and coulometry, can be usedto measure NO with a porphyrinic sen-sor. Amperograms and voltammogramscan be recorded with three-electrode sys-tems. Such three-electrode systems consist


of (1) an NO sensor working electrode,(2) a platinum wire (0.25 mm) counter-electrode, and (3) a silver/silver chlorideelectrode or calomel electrode as thereference electrode. A typical amperomet-ric response of the sensor under varyingflow conditions is shown in Fig. 10. Thecurrent measured with both techniques is(at constant NO concentration) controlledby mass transport (laminar flow or diffu-sion) of NO to the electrode. Therefore, foraccurate determination of NO concentra-tion, the sensor has to be calibrated undermass transport conditions similar to thoseused for in vivo or in vitro measurements.The flow rate of blood at which the NOis monitored has to be measured or calcu-lated based on the diameter of the bloodvessel and the blood pressure. A flow sys-tem of the mimetic solution of blood (e.g.an aqueous solution of dextran, MW 70 000with a viscosity of 2.8 cP at 37 C) can beused to calibrate the NO sensor responseoutside a biological system.

7.2.3Clark Probe for NO Detection

The Clark probe was originally designedfor the detection of oxygen. The probe con-sists of a glass pipette whose opening issealed with a thick, gas-permeable rubbermembrane. Only gases readily diffuse intothe glass pipette through the membraneand are then oxidized or reduced at thesurface of the metal electrode (workingelectrode). In the Clark probe for oxygendetection, a working electrode (platinum)is polarized with a potential of −0.9 Vversus the counter/reference electrode (sil-ver), and a current due to the reduction ofoxygen is observed. The reduction of oxy-gen involves a four-electron transfer yield-ing superoxide (O2

−) and peroxide (O2−2)

intermediates, with the final reduction

product being water. For the detection ofNO with the Clark probe, the polarizationof the electrodes should be +0.9 versusthe silver electrode (vis-a-vis −0.9 for oxy-gen) [17]. A mixture of NaCl and HCl canbe used as the electrolyte; the probe doesnot respond to NO in basic solution.

The Clark probe operates in the am-perometric mode. As an analytical signal,current is measured at a constant potentialof 0.9 V. For a Clark probe with a glasspipette diameter of 300-µm platinum an-ode (50 µm) and silver cathode (200 µm),the current is 3 to 100 pA/µM. This currentis relatively small because of the low NOconcentration achieved within the glasspipette. The detection limit of the elec-trode is about 10−8 M when determinedin a homogeneous solution, where a con-stant concentration of NO can be achievedon both sides of the gas-permeable mem-brane. However, when the probe is usedfor measurements of low concentrationsof NO in a biological environment (hetero-geneous NO solution), several problemsmay arise that limit this application.

The response of the electrode is relativelyslow, with a half-rise time of 1.4 to 3.2 s.This is about three orders of magnitudeslower than the typical response time ofporphyrinic sensors. The slow responsetime is mainly due to the long diffusiondistance to the electrode. Because of itsrelatively large size, the electrode cannotbe placed exactly at the site where NOconcentration is highest (e.g. membraneof the cell). Furthermore, a fraction ofNO that reaches the probe chamber(capillary) may be lost inside due toreaction with oxygen, which also readilydiffuses through the membrane. Thus,while the gas-permeable membrane willprevent diffusion of NO2

− from thesample solution to the capillary chamber, itwill not prevent the formation of NO2

− and


NO3− in the chamber resulting from the

reactions of NO with O2 and O2−; the latter

is an intermediate product of electrodicreduction of oxygen in the silver cathode.Accordingly, the concentration of NOmeasured with a Clark probe in a biologicalmedium will usually be one to two ordersof magnitude lower than the concentrationmeasured with a porphyrinic sensor. Thehighest sensitivity for NO is obtained at apotential of approximately 0.9 V.

7.3Measurement of NO in BiologicalSystems

7.3.1Measurement of NO in a Single Cell

By use of a manual or motorized computer-controlled micromanipulator with 0.2-µmx-y-z- resolution, the porphyrinic sensorcan be implanted into a single cell, placedon the surface of the cell membrane, orkept at a controlled distance (0.2–10 µm)from an NO generating cell [16, 25, 52–58].

When the tip of the sensor touches thecell membrane, a transient small electricalnoise is observed. This is a good indicatorof zero distance from the cell, and fromthis point, the sensor can be moved outfrom the surface by 0.2-mm incrementscontrolled by a computer. Injection of NOSagonists can be done with micro-, nano-,or femto injectors. Injection of largervolumes of agonists with a microinjectorwill cause a ‘‘jet effect,’’ and the initialrelease of NO will be due to a mechanicalforce agonist (shear stress). The shearstress peak of NO will be followed byan NO peak resulting from the effect ofthe chemical agonist of NOS. Injection ofthe agonist with a nano- or femto injectorwill reduce significantly the jet effect andsubsequent NO release due to shear stress.

Figure 11 shows the pattern of NOrelease from a single human umbilicalvein endothelial cell (HUVEC) isolatedfrom the culture, and the pattern of NOrelease from a single cell still in the culture.The sensor was placed in close proximityto the cell membrane (4 ± 1 µm). This

Time[s]

Time[s](a) (b)

200 nmol/LNO

CaI CaI

0 1 2 3 0 1 2 3

Fig. 11 Amperogram of NO recorded with porphyrinic sensor (diameter0.75 µm) placed 4 ± 1 µm from the surface of the isolated single eNOScell (a) and placed 4 ± 1 µm from the surface of single cell in the cellculture (6 × 106 cells) (b). NO release was stimulated from humanumbilical vein eNOS cells with calcium ionophore (CaI).

7.3 Measurement of NO in Biological Systems 245

0 50 100 150

0.0

0.2

0.4

0.6

0.8

Distance[µm]

NO

[µM

]

Fig. 12 Exponential decrease of NO concentration with increasingdistance of porphyrin sensor from the membrane of a single isolated eNOScell (solid line), a single eNOS cell in cell culture and a single eNOS cell inculture (3 × 104 cells). NO release was stimulated from HUVEC with CaI.

figure shows that the peak concentrationof NO on the membrane of a single cellisolated from a given eNOS cell culturewill be the same as the peak concentrationof NO measured on the membrane of asingle cell in the same culture. However,the duration of the plateau is significantlyextended when measurements are done ina cell culture. This difference is due to thehorizontal gradient in NO concentrations,which is much smaller in a cell culture.The depletion of NO from the membraneis slower when the cell is surrounded byother cells that release NO in a culture.In a single isolated cell, the gradient ofNO concentration between the surface andbulk solution is high all around the cell.Therefore, the process of NO depletion isrelatively rapid.

The height of the peak of NO releasedepends on the distance of porphyrinic

sensor from a membrane surface (Fig. 12).The highest NO concentration is ob-served on the HUVEC membrane (750 ±20 nM), with concentration decreasingexponentially with the distance from thecell membrane. At about 60 µm from thecell membrane, the NO concentration was80 nM (single isolated cell) and at a dis-tance greater than 70 ± 15 µm from thecell membrane, NO is not detectable bythe porphyrinic sensor. The decrease ofNO concentration from the cell surface isnot as rapid when the cell is surrounded byanother cell from cell culture. However, ata distance greater than 170 ± 15 nm fromthe cell membrane (in cell culture) NOis also not detectable by the porphyrinicsensor.

From an analytical standpoint, thedetection of NO at the site of thehighest concentration, the surface of the


eNOS cell membrane, is the most conve-nient and accurate method for measure-ment of endogenous NO. Because of thehydrophobic properties of NO (partitioncoefficient between nonaqueous/aqueousphase = 6.5), the membrane is a storagereservoir for NO [24]. Therefore, a smallvolume membrane develops a relativelyhigh steady state concentration of NOwithin a short period of time after acti-vation of NOS. During the diffusion of NOthrough the aqueous phase, significant di-lution occurs. Thus, in situ measurementof NO released from a single isolated cell,from a group of tissue culture cells, orfrom an isolated artery, requires the posi-tioning of the electrochemical porphyrinicsensor on the membrane surface of theendothelium or in close proximity. As canbe seen in Fig. 12, NO decreased exponen-tially with distance from the eNOS cell;therefore, even under static conditions ofeNOS cell culture medium, it would beimpossible to detect NO by the sensorat a distance greater than 170 µm fromthe surface of endothelium. The diameterof the sensor (smaller than, or compa-rable to the diameter of the eNOS cell)allows its placement on or close to themembrane surface. These features of elec-trochemical detection with the sensor offersignificant practical advantages over UV-visible spectroscopy, ESR spectroscopy, orbiochemical assays in real-time detectionof NO [15].

An additional feature of electrochemicaldetection is the direct measurement ofnet NO (biologically active) concentrationand not a total concentration of NOproduced by the endothelium, whichincludes its oxidation products NO3

−,NO2

−, or peroxynitrite (ONOO−). The netNO concentration measured by the sensoris that concentration detectable after abouta millisecond time (the response time of

the sensor). This concentration of NO hasa real physiologic meaning as the NO canrapidly diffuse to smooth muscle cells andcause their relaxation.

7.3.2In vitro Measurement of NO inTissue

NO concentration can be measured withelectrochemical sensors in tissue slices assmall as 50 µm [9, 10, 28, 54, 59]. NOconcentration was measured in vitro inthe carotid artery of normotensive Wistar-Kyoto (WKY), spontaneously hypertensiverats (SHR), and spontaneously hyperten-sive rat-stroke prone (SHR-SP) using aporphyrinic sensor placed near the cellsurface (10 ± 2 µm). Amperometric curvesshowing the change of NO concentra-tion with time were recorded in theabsence and presence of membrane su-peroxide dismutase (SOD) with attachedpolyethylene glycol 400 (PEG-SOD activ-ity 100 U ml−1) (Fig. 13). Since PEG-SODrapidly dismutates superoxide (O2

−), thisindirect approach was used to estimate pro-duction of O2

− at the time of NO release.After addition of CaI, a rapid increase ofNO concentration was observed. Peak NOconcentration was higher for WKY thanSHR and SHR-SP rats (530, 320, and290 nM, respectively). In the presence ofPEG-SOD, an increase of peak NO con-centration was observed for both WKY andhypertensive rats. In the WKY strain, SODtreatment increased the peak NO releaseby 8%. However, for hypertensive rats theincrease of NO peak concentration wasmuch higher (40–60%) in the presenceof PEG-SOD than for normotensive rats.This finding indicates that a significantconcentration of O2

− is generated in thecarotid artery of hypertensive rats, and thisO2

− consumed a major portion of NO in a


(a) (b) (c)

0 10 20 0 10 20 0 10 20

Time[s]

Time[s]

Time[s]

CaI CaI CaI

NO 200 nM

Fig. 13 Typical amperograms showing changes of NO concentration near thesurface (10 ± 3 µm) of the endothelium of the carotid artery of WKY rat (a)SHR-SP rats (b) and diabetic rats (c). NO was measured with a porphyrinic sensor(diameter 3 µm) in the presence (dotted line) and absence (solid line) ofPEG-SOD.

fast chemical reaction. This process leadsto the decrease of the biologically activeNO concentration, followed by vasocon-striction and hypertension.

This is a typical example of the uniqueapplication of the electrochemical sensorfor the detection of NO release from nor-motensive and hypertensive rats [6, 29, 30,55, 60–62]. It has been reported, based onspectroscopic measurements, that the en-dothelium of hypertensive rats producedmore NO2

−/NO3− than the endothelium

of WKY rats [63]. However, these reportswere in contradiction to the data obtainedon smooth muscle relaxation, hindered inhypertensive rats [62]. This means that theendothelium of hypertensive rats shouldproduce less NO. Electrochemical mea-surements with the porphyrinic sensorclearly show that the biologically activenet concentration of NO produced by theendothelium of hypertensive rats is lowerthan that produced by the endothelium ofWKY rats. These results correlate well withpreviously reported smooth muscle relax-ation data. Total production of NO by theendothelium of hypertensive rats is higher

than in WKY rats. However, the endothe-lium of hypertensive rats also generatedsignificant amounts of superoxide, whichrapidly reacts with NO to produce thestable product OONO−. Therefore, thenet NO concentration as detected by theporphyrinic sensor is much lower in hy-pertensive rats as compared with WKYrats. In the presence of membrane perme-able PEG-SOD, an effective dismutation ofO2

− occurred, followed by a large increaseof net NO concentration in SHR rats, fi-nally exceeding that observed for WKY ratsin the absence of PEG-SOD. Peroxynitritewhen protonated (pKa = 6.8) to HOONOusually undergoes isomerization (t1/2 <

1 s) to form hydrogen cation and nitrate an-ion [64]. However, at high concentrationsof NO and O2

−, large concentrations ofHOONO can be formed. Under these con-ditions, HOONO may undergo cleavage toa hydroxyl free radical (OH•), a nitrogendioxide free radical (NO2

•), or nitron-ium cation (NO2

+) and hydroxide anion(OH−). Three of these cleavage products(OH•, NO2

−, NO2+) are among the most

reactive and damaging species in biological


systems, and may be major contributorsto the severe damage of the endotheliumand the cardiovascular system occurringin hypertension. A similar increase in O2

−production by dysfunctional endotheliumis observed during endotoxemia.

7.3.3In vivo Measurements of NO in theKidney

Figure 14(a) shows the amperometriccurve measured in vivo during endotox-emia with a porphyrinic sensor placed inthe rat kidney. During administration oflipopolysaccharide (LPS), an increase of

NO production from its basal concentra-tion was observed. The concentration ofNO reached a peak of 210 nM after 240 s,persisting for 15 min before decaying ata rate of 0.7 n Ms−1. After 45 min, theNO concentration started to rise again, butat a much slower rate. A plateau of NOrelease was established 80 ± 15 min afterLPS administration.

UV-visible spectroscopy (Griess reagent)was used to determine a change ofNO2

−/NO3− concentration in blood dur-

ing endotoxemia [65]. Both NO2− and

NO3− are the end products of NO oxi-

dation in biological systems (Figure 14b).The concentration of NO2

− and NO3−

(b)

(a)

0 1 30

1

2

3

4

5

6

7

8

Time[hrs]

(NO

2− ) +

(NO

3− )[µ

M]

0 20 40 60 80 100

Time[min]

LPS

NO 100 nM

Fig. 14 (a) NO release recorded in the kidney of a rat during septicshock; (b) concentration of NO in the blood measured by UV-visiblespectroscopy (Griess reagent) during septic shock.


significantly increased from basal 1.3to 6.3 µM after 6 h of endotoxemia.The most dramatic rate of increasewas observed after 3 h of endotoxemia.The concentration of O2

− (measured invitro during stimulation with A23187using a chemiluminescence method)also increased significantly during en-dotoxemia [9]. During the first hourof endotoxemia, the O2

− concentrationincreased threefold relative to the control.

Induction of iNOS by bacterial toxinsdirectly or through cytokines leads to thegeneration of NO in arterial walls. Admin-istration of toxin (LPS) also generated NOby calcium-dependent eNOS in the en-dothelium during endotoxemia. The highproduction of NO by the endothelium isobserved only in the early acute phase ofendotoxemia (first 20 min); however, it hasa profound effect on the late chronic phaseof endotoxemia when iNOS becomes themain generator of NO. The porphyrinicNO sensor measured only free NO; that is,the net NO concentration not consumedin the extremely fast chemical reactionwith O2

− and other redox centers in thetissue. This free NO, when measured in-termittently, remained approximately thesame during the second chronic phase ofendotoxemia, in spite of high NO pro-duction by iNOS (after 45 min from LPSinfusion), measured indirectly by assayingthe accumulation of the NO decay prod-ucts (NO2

−; NO3−) in blood plasma by the

Griess method. NO measurement with theporphyrinic sensor clearly indicates thatthe net concentration of NO is only twotimes higher during the late chronic phaseof endotoxemia, than the observed preen-dotoxemic basal concentration. This netNO concentration is much lower than thetotal NO produced during this period, butstill sufficient to account for the coincidentchronic hypotension observed. However,

low-net NO concentration cannot be ac-countable for the sudden death of animalsafter about 6 h of endotoxemia. Direct elec-trochemical measurements suggest thatthe dysfunction of eNOS cells eventuallyleading to the death of the animal is trig-gered by O2

−/OONO−, rather than by NOitself during endotoxemia.

7.3.4In vivo Measurements of NO in theBeating Heart

Measurement of NO amidst the dynamicin vivo conditions of cyclic breathing andheart beating is a challenging task [10]. Inorder to overcome these potential interfer-ences and to record a reliable NO signal,the catheter-protected porphyrinic sensoris slightly modified in two ways. First, theactive sensor tip is shortened to 50 to 60 µmfrom 0.3 to 0.5 mm usually used for in vivomeasurements. In addition, the truncatedneedle from which the active sensor tipemerges is cut 50- to 60-µm shorter thanits protective catheter, so that the tip ofthe sensor is completely recessed withinthe ventilated catheter tip, rather than pro-truding from an unventilated catheter tip.The sensor has to be tested under differentexperimental conditions to ensure that thesensor did not generate an analytical sig-nal when subjected to conditions of fluidpressure, mechanical deformation of thesensor tip, or intrinsic electrical activitywithin the heart.

In order to estimate potential piezo-electric interference due to mechanicaldeformation of the sensor tip, a mi-cromanipulator can be used to deformthe sensor tip at physiologically relevantfrequency (1–3 Hz). As can be seen inFig. 15, the deformation of the sensorgenerates small electrical perturbations,six to ten times smaller than the signal


50 nMNO

(a)

50 nA

Bending

0 1 2

Time[s]

(b)

Standardprecordialelectrode

Porphyrinicsensor

Time

0.5 µMNO0.5 µA

Fig. 15 Response of the catheter-protected porphyrinic sensor tomechanically bending the active tip of the sensor at a frequency of 3 Hz,and to an ECG signal. The vertical bars represent the equivalent currentthat would be generated by the porphyrinic sensor in the 50 nM NO[(a), lower tracing] or 0.5-µM NO (b).

1.0

0.0

Nitr

ic o

xide

[µM

]

7.0 7.2 7.4

298 ms

7.6 7.8 8.0

Time[s]

0.5

Fig. 16 Changes of NO concentration recorded in a beating heart. The porphyrinicsensor was implanted in the myocardium (left ventricle, rabbit).


recorded by the sensor at a low (20–50 nM)concentration of NO. To eliminate thepossibility that the sensor might respondto the electrical current generated withinthe extensive cardiac conduction systemor within the depolarizing myocardium,the sensor can be used as the exploringelectrode connected to a standard electro-cardiography (ECG). As is apparent fromFig. 15, when used in this manner, the por-phyrinic sensor is barely able to detect theECG signal. In these experiments, observ-ing the analytic response of the sensor topotential piezoelectric or in vivo electricalcurrent interferants, the measured peakof NO was at least 30 to 50 times largerand temporarily shifted from the conser-vatively estimated background noises.

An anesthetized rabbit was used tomeasure local fluctuations in NO con-centration in the apical left ventricularmyocardium. Rapid changes in cardiacNO concentration related to the cardiaccycle are observed (Fig. 16). In the rabbitheart (endocardium), each cardiac cycle(period about 300 ms) begins and endswith an intercycle NO concentration ofabout 100 nM. During early systole, NOconcentration reaches a basal of 0.2 µM fol-lowed by a slow increase to a semi-plateau.Early diastolic filling is accompanied by arapid increase of NO concentration witha peak diastolic about 1.1 µM that is at-tained at 240 ms into the cardiac cycle.After this peak, there is a sharp decay inthe intercycle NO concentration. The NOsignal recorded by the sensor disappearswhen monitored at a potential of 0.40 V(which is 230 mV below the peak potentialof NO oxidation), clearly indicating that thesignal measured at 0.67 V is due to NO,

not mechanical noises, which are inde-pendent of potential. To demonstrate therelationship between the ECG signal, leftventricular volume, and instantaneous NOconcentration, simultaneous recordings ofeach of these were performed (Fig. 17).

During ischemia of the heart, a signif-icant decrease of the mechanically stim-ulated NO release is observed (Fig. 18).After 240 s of ischemia, the peak of NOis 400 nM (about 60% lower than the con-trol). NO is released in a pulsatile fashionfrom the beating heart and its synthesisis directly related to ventricular loadingconditions in vivo. Direct measurementsof NO with the porphyrinic sensor inthe beating heart may help to explain

Fig. 17 NO signal, ECG signal, andventricular volume recordedsimultaneously in the left ventricle in thebeating heart (rabbit).

0 200 400

Time[ms]

NO,1 µmol/L

ECG

NO

Ven

tric

ular

vol

ume

[ml]

8

7

6

5

4


NO 500 nmol/L

(a)

(b)

300.0 300.4 300.8 301.2 301.6

Time[s]

0.0 0.4 0.8 1.2 1.6

Fig. 18 NO concentration recorded in the wall of the left ventricle of thebeating heart (rabbit) under normal (a) and ischemic conditions (b), (after240 s of ischemia).

certain aspects of the beat-to-beat regu-lation of cardiac performance and alsoprovide insight into the pathophysiologyof diseases associated with increased my-ocardial distention, such as valvular heartdisease or heart failure [10]. Cells withinthe heart are subjected to tremendous me-chanical deformation during filling andbeating. NO affects mechanical propertiesof cardiac myocytes by increasing cGMPto facilitate relaxation and to mediate anacetylcholine stimulated decrease in con-tractility. Amperometric detection with aporphyrinic sensor is the only analyticalmethod currently available to measure NOconcentration in the beating heart withsufficient time resolution and sensitivity.

7.3.5Measurement of NO in the Brain

7.3.5.1 In vitro Measurement with ClarkProbeFigure 19 presents an amperometric curveobtained at different potentials for NO

released from the cellular layer of a rat’scerebellar slices [17]. The white matterwas stimulated with biphasic pulse-trains(intensity ±600 µA, frequency 20 Hz, du-ration of pulse-train 5 s). The electricalstimulation of the white matter evoked in-creases in NO concentration between 20and 75 nM. Figure 19(b) shows an amper-ogram obtained in the presence of 100-nMNO at different potentials. The applied po-tential results in current increases, whichare maximal at a potential of about 0.9 V.

7.3.5.2 In vivo Measurement withPorphyrinic SensorFigure 20(a) presents a typical amperomet-ric curve obtained from in vivo measure-ments of NO concentrations in a rat’sbrain during a middle cerebral artery occlu-sion (MCAO). The porphyrinic sensor wasstereotoxically implanted perpendicularlyinto the ipsilateral parietal cortex at coordi-nates 0.8-mm posterior and 4.5-mm lateralto the bregma, and 3 mm below dura.An increase in NO concentration (about


(a)

0.9 V

0.7 V

0.5 V

0.5 pA

(b)

0.9 V

0.7 V

0.5 V

2 pA

5 s

Fig. 19 NO release in the molecular layer of a rat’s cerebellar slices measuredduring white-matter stimulation (20 Hz for 5 s); traces were recorded at differentplatinum anode potentials. (b) Currents measured in the presence of 100-nM NOat different potentials. (Reproduced from K. Shibuki et al., Nature 1991, 349, 326,with permission.)

NO 2 µM

MCAO

Time[min]

0 5 10 40

(a)

35

MCAOL-NAME

Time[min]

0 5 10 40

(b)

35

NO 2 µM

Fig. 20 Continuous recording of the changes of NO concentration in the ischemic brain (rat)in the absence (a) and in the presence of an NOS inhibitor (L-NAME).


2.2 µM) was established after 3 min. After6 min of ischemia, NO concentration grad-ually decreased, after 20 min it decreasedbelow basal level of 40 nM, and after35 min was undetectable by the sensor. Inthe presence of L-NAME (cNOS inhibitorinfused 5 min before MCAO), the NO con-centration also increased but its maximumwas 0.9 µM (about 60% lower than the con-trol). After 40 min, the NO concentration

was slightly below (30 nM) the basal leveland was detectable by the sensor. The ki-netics of NO release and its concentrationproduced during ischemia is related to thebrain damage. Initial high production ofNO during the acute phase of ischemiaplays a beneficial role in preventing braindamage [66]. However, after prolonged is-chemia cNOS starts also to produce O2

−,which consumes NO. The generation of

(a)

(b)

Fig. 21 Necrosis of ischemicbrain; normal (a) and treatedwith L-NAME (b).

146 nM NO

120

50 nM

Bradykinin L-NMMA+bradykinin

L-arginine+bradykinin

360 120 360 600

26 nM NO

160 nM NO

Time[s]

Fig. 22 In vivo measurement of NO concentration in human beings.


high concentration of ONOO initiates acascade of events leading to the productionof highly oxidative radicals (OH•, NO2),and subsequent brain damage. The low-ering of NO production during ischemiaprevents the generation of O2

−, ONOO−OH• and NO2

•, and mollifies ischemiadamage (Fig. 21). The necrosis of the braindecreased significantly in the presence ofL-NAME (control 7% of necrosis L-NAME3% of necrosis).

7.3.6Measurements of NO in HumanBeings

Catheter-protected porphyrinic sensorscan be sterilized with ethylene oxide andused for NO measurement in humans [67].For measurements of NO in veins or arter-ies, an intervenous catheter is used. Twocannulae are inserted into a vein. A 24-G catheter is inserted retrogradely, anda 23 G butterfly needle is positioned an-terogradely with its tip 10 to 15 mm fromthe end of the catheter. The catheter isflushed with 0.5-ml heparin (5000 U/ml),and a NO sensor is mounted on a 24-Gneedle and is placed so its tip protruded 1to 2 mm beyond the end of the catheter.A platinum wire counterelectrode and sil-ver/silver chloride reference electrode canbe placed on the skin adjacent to the veinand covered with conductive gel. Hep-arin solution physiological saline or drugscan be infused continuously through thebutterfly needle. Infusion of bradykinincauses a dose-dependent NO signal, at-tenuated by coinfusion of inhibitor ofNOS L-NMMA (Fig. 22). Infusion of L-arginine with bradykinin restored NOconcentration.

The detection of a signal for NO inresponse to bradykinin provides further

evidence in man that the endothelium-derived relaxing factor is NO. Functionalstudies suggest that the alteration inthe L-arginine/NO pathway occurs inhypertension, diabetes, and hypercholes-terolemia, and decreased production of NOmight be a link between risk factors andatherogenesis [68]. The ability to assesslocal generation of NO with porphyrinicsensor should facilitate exploration of therole of this mediator in physiological orpathophysiological processes.

References

1. D. S. Bredt, Free Radical Res. 1999, 31, 577.2. V. L. Dawson, T. M. Dawson, Proc. Soc. Exp.

Biol. Med. 1996, 211, 33.3. S. Grunfeld, C. A. Hamilton, S. Measaros

et al., Hypertension 1995, 26, 854.4. T. Malinski, M. Kapturczak, J. Dayharsh

et al., Biophys. Res. Commun. 1993, 194, 654.5. J. P. Cooke, V. J. Dzan, Circ. Res. 1997, 96,

379.6. M. Tschudi, S. Mesaros, T. Luscher et al.,

Hypertension 1996, 27, 1.7. K. S. Christopherson, D. S. Bredt, J. Clin.

Invest. 1997, 100, 2424.8. C. Ladecola, Trends Neurosci. 1997, 20, 132.9. R. J. Gryglewski, P. P. Wolkow, W. Uracz

et al., Circ. Res. 1998, 82, 819.10. D. J. Pinsky, M. Kapturczak, E. Block et al.,

Circ. Res. 1997, 81, 372.11. N. Hill-Kapturczak, M. Kapturczak, E. R.

Block et al., C. J. Am. Soc. Nephrol. 1999,3, 481.

12. J. S. Beckman, W. H. Koppenol, Am. J. Phys.1996, 271, C1 424.

13. F. Kiechle, T. Malinski, Ann. Clin. Lab. Sci.1996, 26, 501.

14. T. Malinski, S. Mesaros, P. Tomboulian,Methods Enzymol. 1996, 268, 58.

15. T. Malinski, E. Kubaszewski, F. Kicchle,Methods Neurosci. 1996, 31, 14.

16. T. Malinski, Z. Taha, Nature 1992, 358, 656.17. K. Shibuki, O. Okada, Nature 1991, 349, 326.18. T. Malinski, E. Kubaszewski, F. Licchle et al.,

The Biology of Nitric Oxide, Portland Press,London, 1994.

19. R. F. Furghott, J. V. Zawadzki, Nature 1980,288, 373.


20. R. M. J. Palmer, A. G. Ferrige, S. Moncada,Nature 1987, 327, 524.

21. R. M. J. Palmer, D. S. Ashton, S. Moncada,Nature 1988, 333, 664.

22. R. M. J. Palmer, E. A. Higgs, S. Moncada,Pharm. Rev. 1991, 43, 1991.

23. C. Nathan, J. FASEB 1992, 6, 3051.24. T. Malinski, Z. Taha, S. Grunfeld et al.,

Biochem. Res. Biophys. Res. Commun. 1993,193, 1076.

25. L. A. Blatter, Z. Taha, S. Mesaros et al., Circ.Res. 1995, 76, 922.

26. T. Malinski, M. Radomski, Z. Taha et al.,Biochem. Biophys. Res. Commun. 1993, 194,960.

27. A. Kanai, H. C. Strauss, A. Truskey et al.,Circ. Res. 1995, 77, 284.

28. T. Malinski, M. Kapturczak, J. Dayharshet al., Biochem. Biophys. Res. Commun. 1993,194, 654.

29. W. Linz, T. Jessen, R. H. A. Becker et al.,Circulation 1997, 96, 3164.

30. F. Cosentino, S. Patton, V. d’Uscio et al., J.Clin. Invest. 1998, 101, 1530.

31. D. S. Bredt, S. H. Snyder, Proc. Natl. Acad.Sci. U.S.A. 1990, 87, 682.

32. Z. G. Zhang, M. Chopp, F. Bailey et al.,J. Neurol. Sci. 1995, 128, 22.

33. L. J. Forman, P. Liu, R. G. Nagele, Neu-rochem. Res. 1998, 23(2), 141–148.

34. P. Klatt, K. Schmidt, G. Uray et al., J. Biol.Chem. 1994, 268, 14 781.

35. J. B. Hibbs, R. R. Traintor, Z. Vavrin, Science1987, 235, 473.

36. D. H. Stuehr, H. J. Cho, N. S. Kwon et al.,Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 7773.

37. S. Archer, FASEB L. 1993, 7, 349.38. Y. -J. Sung, J. H. Hotchkiss, R. F. Austic

et al., Biochem. Biophys. Res. Commun. 1992,184, 36.

39. L. R. Cantilena, R. P. Smith, S. Frasur, J. Lab.Clin. Med. 1992, 51, 496.

40. R. F. Furchgott, P. M. Vanhoutte, Faseb J.1987, 3, 2007.

41. K. M. K. Rao, J. Padmanabhan, D. L. Kilby,J. Leukocyte Biol. 1992, 51, 496.

42. R. G. Bogle, S. Mancada, C. D. Pearson et al.,E. Br. J. Pharmacol. 1992, 105, 768.

43. T. Malinski, Z. Taha, S. Grunfeld et al., Anal.Chim. Acta 1993, 279, 135.

44. T. Stephane, F. Bediour, F. Devynek, Talanta1996, 43, 303.

45. S. Mesaeros, S. Grunfeld, A. Mesarosovaet al., Anal. Chim. Acta 1997, 339, 265.

46. L. Czuchajowski, E. J. Bennett, S. Goszczyn-ski et al., J. Am. Chem. Soc. 1989, 111, 607.

47. J. E. Bennett, T. Malinski, Chem. Mater.1991, 3, 490.

48. T. Malinski, A. Ciszewski, J. Bennet et al.,J. Electrochem. Soc. 1991, 138, 2008.

49. T. Malinski, J. Bennett, E. Redox Chemistryand Interfacial Behavior of Biological Molecules,Plenum Press, New York, 1988, pp. 87–107.

50. J. R. Fish, E. Kubaszewski, A. Peat et al.,Chem. Mater. 1992, 4, 795.

51. T. Malinski, J. Bennet, F. Bailey et al., Pro-ceedings of the 12th international conferenceof the IEEE Engineering in Medicine and Biol-ogy Society, IEEE, Philadelphia, 1990, Vol. 12,p. 1691.

52. F. L. Kiechle, T. Malinski, Am. J. Clin. Pathol.1993, 100, 567.

53. J. Balligand, D. Ungureanu-Lorgrois, W.Simmons et al., J. Biol. Chem. 1994, 269,27 580.

54. R. Cohen, F. Plane, S. Najibi et al., Proc.Natl. Acad. Sci. U.S.A. 1997, 94, 4193.

55. G. Wiemer, B. Pierchala, S. Mesaros et al.,Endothelium 1996, 4, 119.

56. H. Heitsch, S. Brovkovych, T. Malinsky et al.,Hypertension 2001, 37, 72.

57. L. Vergnani, S. Hatrik, F. Ricci et al., Circu-lation 2000, 101, 1261.

58. S. F. Silverton, O. A. Adebanjo, B. S.Moonga et al., Biochem. Biophys. Res. Com-mun. 1999, 259, 73.

59. V. Brovkovych, S. Patton, S. Brovkovychet al., J. Phys. Pharm. 1997, 48 663.

60. G. Wiemer, B. Pierchola, S. Mesaros et al.,Endothelium 1996, 4, 119.

61. T. Malinski, M. Kapturczak, T. Dayharshet al., Biochem. Biophys. Res. Commun. 1993,194, 654.

62. S. Gruntfeld, C. A. Hamilton, S. Mesaroset al., Hypertension 1995, 26, 1995.

63. L. Linder, W. Kiowski, F. Buhler et al., Circu-lation 1990, 81, 1769.

64. J. S. Beckman, Nitric Oxide: Principles andActions, Academic Press, San Diego, 1996.

65. H. J. Gyllenhammer, Immunol. Methods1987, 97, 209.

66. T. Malinski, F. Bailey, Z. G. Zhang et al.,J. Cereb. Blood Flow Metab. 1993, 13, 335.

67. P. Valance, S. Patton, K. Bhagat et al., Lancet1995, 345, 153.

68. R. G. Knowles, M. Palacious, R. M. J. Palmeret al., Proc. Natl. Acad. Sci. U.S.A. 1989, 86,5159.

257

8Scanning ElectrochemicalMicroscopy Applied to BiologicalSystems

Hitoshi Shiku, Hiroaki OhyaYamagata Public Corporation for the Development of Industry,Yamagata 990-2473, Japan

Tomokazu MatsueDepartment of Biomolecular Engineering, Graduate School of Engineering,Tohoku University, Sendai 980-8579, Japan

8.1 SECM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2598.1.1 Feedback Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2598.1.2 Generation/Collection Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . 261

8.2 Applications to Biological Systems . . . . . . . . . . . . . . . . . . . . . . . 2628.2.1 Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2668.2.2 Antigen–Antibodies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2678.2.3 Local Fluxes Through Biological Materials . . . . . . . . . . . . . . . . . . 2688.2.4 Liquid–Liquid Interfaces, Liquid–Air Interfaces . . . . . . . . . . . . . . 2698.2.5 Planar BLMs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2708.2.6 Cells, Tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273

259

8.1SECM

SECM (Scanning electrochemical mi-croscopy) is a technique to characterizethe local electrochemical nature of variousmaterials by scanning a probe microelec-trode [1, 2]. The spatial resolution of SECMis inferior to the conventional scanningprobe microscopes such as scanning tun-neling microscopy (STM) and atomic forcemicroscopy (AFM) as the fabrication ofthe probe, microelectrode, with nanome-ter sizes is quite difficult and the faradaiccurrent of the microprobe is very small(often picoamps or less). However, SECMhas unique characteristics that cannot beexpected for STM and AFM: SECM can im-age localized chemical reactions and it alsocan induce localized chemical reactions ina controlled manner.

SECM has been widely applied tonumerous fields involving electrochem-istry, such as electrode surfaces, poly-mers, biomaterials, and liquid–liquid in-terfaces [3–12]. The probe current reflectsthe electrochemical processes occurring inthe small space surrounded by the probeand the substrate. The electron transferat the probe and substrate (if conductive)and mass transfer across the solution af-fect the probe current. The probe currentis also influenced if chemical reactions

or adsorption–desorption processes occurwithin the gap solution or at the electrodesurface. Therefore, the phenomena in the‘‘ultramicro electrochemical cell’’ can bequantitatively understood by analyzing theprobe current response [13]. The volume ofthe small space and the transient time for amolecule to transfer from the probe to sub-strate are precisely controlled by changingthe probe-sample separation, which offersnew research fields for characterizing sin-gle molecules [14, 15] or highly reactiveradical species [16, 17].

8.1.1Feedback Mode

One of the most remarkable features ofmicroelectrode measurements is that onecan observe steady state redox currents inrelatively short time domains. When thecurrent response at a microelectrode is ina steady state, a hemispherical diffusionregion of the electrogenerated species isformed around the microelectrode probe(Fig. 1a). The size of the diffusion regionlargely relies on the probe radius and thesteady state current is expressed by thefollowing equation:

i = 4nFDCa (1)

where F is Faraday’s constant, D and C

are the diffusion coefficient and the bulk

260 8 Scanning Electrochemical Microscopy Applied to Biological Systems

Insulator

Tip

Conductor

Tip

(a) (b) (c)

Fig. 1 Schematic illustration of the feedback mode SECM.

concentration of the reagent, and a is theradius of the microelectrode probe. If aplanar wall exists within the hemisphericalregion, the distribution of the specieschanges due to geometrical perturbation.

When the wall is a conductor, the gradi-ent becomes steeper compared with that inthe bulk solution and therefore the probecurrent increases (Fig. 1b). The situationshowing a larger faradaic current com-pared to the steady state current in thebulk solution is called a positive feedbackmode [3–7]. For example, an electrode re-action at the probe converts a species inits oxidized state into the correspondingreduced species, which diffuses onto theconducting substrate to regenerate the ox-idized species. This redox cycling betweenthe probe and substrate increases the re-dox response at the probe. The oppositecase is called the negative feedback mode.When the wall is an insulator, the currentdeclines, reflecting the broader concentra-tion profile at the probe surface (Fig. 1c). Inthe feedback mode, either positive or neg-ative, the probe current largely depends onthe probe-substrate distance. This is themost fundamental characteristic of SECMand therefore the probe can be preciselypositioned at the desirable point by moni-toring probe current. For both conductiveand insulating samples, SECM can providetopographic information on the samplesurfaces by scanning the microelectrodeprobe.

The feedback mode SECM is an experi-mental setup to measure faradaic current

profiles along the sample surface (x-yplane) or normal to the surface (z di-rection) by scanning the probe withinthe very small probe-sample separationwhere the probe current deviates fromthe steady state value. Voltammetry andchronoamperometry at fixed positions arealso available. The currents reflect entireevents including mass transfer, chargetransfer, and chemical reactions (if any)in the small space between the probe andsubstrate. For all circumstances, the steadystate current observed in the bulk solu-tion is a very important parameter thatis considered as a standard response forquantitative analysis.

However, it is not easy to calculate theexact relation between the probe currentand the probe-sample distance in the feed-back mode due to the geometric complexityin the experimental system. For quantita-tive analyses, digital simulation has beenused in the feedback mode SECM bothfor conductor and insulator substrates [18].Digital simulation is a powerful techniquefor analysis of relatively complicated elec-trochemical systems such as SECM. Inthe simulation, the space between theprobe and the substrate are divided intosmall volume elements. Mass transferbased on diffusion for each volume ele-ment is calculated. If one chooses suitableinitial and boundary conditions, the ex-perimental situation including the masstransfer and various chemical reactionscan be simulated. From the simulation,one can determine the heterogeneous rate

8.1 SECM 261

constants [19–21], homogeneous reactionconstants [16, 17, 22–24], local fluxes, dif-fusion coefficients [25], and so forth. Thisquantitative analysis was introduced in1989 for SECM studies and to date this isstill a most reliable method to quantify var-ious parameters in a specific experimentalsystem.

8.1.2Generation/Collection Mode

In the General/Collection (G/C) mode ex-perimental set up, the potentials of the twoworking electrodes were controlled with abipotentiostat [26]. A species electrogener-ated at an electrode (the generator) surfacediffuses into the solution and is detected atthe other working electrode (the collector).This mode is applicable to the studied withrotating ring-disk electrodes or microarrayelectrodes. In 1986, Engstrom and cowork-ers [27–30] reported the G/C experimentalsystems with an amperometric ultrami-croelectrode probe. They placed a probemicroelectrode very close to a macrosizeelectrode and detected ferricyanide, whichwas generated at the large electrode. Gath-ering a set of transient current responsesrecorded at several probe-substrate dis-tance, the concentration profiles near thegenerator electrode have been visualizedwith tens of ms time resolution [27]. In thesame article, they demonstrated the capa-bility of the experimental system to imagesurface reactivity with 20-µm spatial res-olution and also detected short lifetimeactive species such as NAD• radical. They

successfully applied their experimentalsystem to mapping of concentration pro-files for epinephrine and determination ofthe second-order rate constant for the re-action between epinephrine quinone andleucoadrenochrome [28].

G/C mode experiments are basically to-pography independent in estimating localconcentration profiles only if the probe-substrate distance is far enough from thesample surface. However, when the probeis placed very close to the sample surface,the geometrical influence of the micro-electrode probe in the experiment systemshould be considered carefully. Figure 2shows two typical situations in the G/Cmode experiments, in which amperomet-ric probe is inserted into the diffusionlayer of the substrate electrode. In the firstcase shown in Fig. 2(a), the probe-sampleseparation is significantly large and thereis no crossover of the diffusion regionsfrom the probe and the sample surface. Lo-cal concentration can be directly estimatedfrom the probe current by using Eq. (1), al-though the concentration around the probeis not exactly homogeneous. There are sev-eral studies [31, 32] done in the G/C modeto estimate the fluxes by recording the con-centration profile around the substrate. Inthese studies, the surrounding concentra-tion profile of the species was comparedwith theoretical ones calculated based ondiffusion equations. The latter case shownin Fig. 2(b), the probe diffusion layer hasreached the sample surface. This situa-tion is more likely in the feedback mode.Eq. (1) cannot be used for estimation of the

Fig. 2 Schematic illustration ofthe G/C mode SECM.

Collector

(a) (b)

Generator


local concentration. The digital simulationtechnique mentioned earlier is still appli-cable [33, 34], but the calculation becomesmuch more complicated because the con-centration profile formed at the substrate iswide spread. In addition, when the probe islaterally scanned for imaging, convolutionof the topography and original concen-tration profile makes the interpretationof the images unclear. A study of two-dimensional mapping of the local fluxeshas been reported [35].

There are several ways to avoid thefeedback effect. One is to use transienttechniques such as fast-scan cyclic voltam-metry that have been vigorously carriedout for monitoring single cell events bymeans of microelectrode [36]. A pure con-centration profile will be obtained withina short time scale before the diffusionlayer at the probe reached the samplesurface. The other is to utilize poten-tiometric probes that have been alreadydemonstrated in G/C mode SECM imag-ing [37–39]. As the potentiometric probedoes not consume or generate electroac-tive species, original concentration profilesaround the sample are less perturbed.There have been a lot of studies in-volving mapped concentration profiles ofCl−, H+, Na+, K+, NH4

+, and so on.A study of mapping of local chemi-cal reagent distribution by scanning apotentiometric probe was reported in1972 [40].

8.2Applications to Biological Systems

SECM is applicable to many surfaces thatgenerate electroactive substances (G/Cmode) or that regenerate the redox speciesconsumed at the probe reaction (feed-back mode). Therefore, SECM has high

potential to examine biological systemsunder physiological conditions with highspatial resolution. For instance, enzyme-catalyzed reactions, electron and ion trans-port through biomembranes, and chem-ical communication within a cell andbetween cells have been investigated withSECM. In the SECM studies focusingon biological systems, base substratesare basically not electrodes and there-fore, the experimental setups are some-what different from those in the stan-dard SECM experiments. Typically, thereare three types of sample configurations:(1) Biomaterials immobilized on solid sup-ports.(2) Membranes or interfaces sand-wiched by two liquid phases for analyzingflux through the boundary due to electronor mass transfer. (3) Living samples forintact observation.

The first category includes enzyme-catalyzed reactions or metabolic reactionsat solid substrates. Table 1 is a list ofthe studies that characterize immobilizedbiomaterials by SECM. Glucose oxidaseis a major enzyme used as a target todemonstrate various detection schemes.So far, imaging in the G/C and feedbackmodes, and the detection of the biocatalyticproducts with a specially fabricated probehave been reported [41–45]. Diaphorase isthe only enzyme routinely visualized bythe feedback mode SECM imaging at themonolayer surface coverage [46–48]. Al-kaline phosphatase and HRP have beenused for the labeled enzyme to charac-terize antigen–antibody binding [49–53].Mitochondria and yeast have been visual-ized by SECM, on the basis of metabolicreactions catalyzed by enzymes existingin the samples. To date, all the experi-ments were carried out in excess enzymesubstrate conditions.

Table 2 introduces studies for analyzinglocal fluxes due to electron transfer

8.2 Applications to Biological Systems 263

Tab. 1 SECM studies on immobilized enzymes, immobilized biomaterials

Sample Monitoring reagents Reference and notes

Plant tissue containingenzymes

Norepinefirine, dopamine 41

Glucose oxidase H2O2 41 Collecting enzymaticallygenerated electroactive species

FcCOOH, Fe(CN)64−, HQ 42 Determining enzymatic kinetic

parametersHydroquinone 43 Imaging enzymatic activity and

topographyH2O2 44 Probing with enzyme electrode

Glucose oxidasemonolayer

H2O2 45

Diaphorase monolayer FcCH2OH, Co(phen)32+ 46 Determining enzymatic kinetic

parametersFcCH2OH 47, 48

Mitochondria TMPD 43Urease, Yeast H+ 37 Using potentiometric pH

selective tipAlgae Fe3+ 41 Monitoring metal uptakeAnti-digoxin p-Aminophenol 49 By use of alkaline phosphatase

labeled antigenCEA FcCH2OH 50 By use of HRP labeled antibodyHCG, HPL FcCH2OH 51 Dual immunoassayCEA, HCG, HPL, AFP FcCH2OH 52, 53 Multiimmunoassay

Notes: HRP, horseradish peroxidase; TMPD, N,N,N′,N′-tetramethyl-p-phenylenediamine; CEA,carcinoembrionic antigen; HCG, human chorionic gonadotropin; HPL, human placental lactogen;AFP, α-fetoprotein.

or mass transfer through the bound-ary of porous membranes, liquid–liquidinterfaces, liquid–gas interfaces, and pla-nar bilayer lipid membranes (BLMs).Fluxes through porous biological mem-branes have been characterized in theG/C mode by SECM imaging. Hydro-dynamic pressures, osmotic pressures,electric fields, and concentration gradientsdrive the fluxes [31, 35, 54–61]. Elec-tron [62–68] and mass transfer [69–74] atliquid–liquid and liquid–air interface arecharacterized by the approach character-istics of a probe microelectrode to theinterface. In Table 2, the microelectrodeprobe was placed in the left side of the liq-uid phase. The potential between the two

liquid phases is generally unbiased and twoelectrodes configuration is available. BLMsare formed in a pinhole made in the Teflonsheet where the membranes are orientedvertically [75–79] or horizontally [80].

Finally, studies on living samples suchas cells or tissues are shown in Table 3.SECM imaging capability offers bothtopography of the sample surface andthe distribution of a specific speciesgenerated from the cells or tissues [81–87].Time course recording with a probefixed very close to the sample is anotherimportant experimental setup [88–102].The chronoamperometric responses arecritically influenced by the probe-sampleseparation.


Tab. 2 Studies on analysis of local flux across membranes or interfaces


Characterization of local flux through biological materialsPorous mica Fe(CN)6

4− 31, 54Hairless mouse skin Fe(CN)6

4− 55–58 Driven by membrane potentialDentine slice Fe(CN)6

4− 59, 60 Driven by hydrostatic pressureLaryngeal cartilage Ru(NH3)6

3+, O2 35 Driven by osmotic pressureHuman dentine Fe(CN)6

4− 61 Driven by concentration gradientElectron transfer at liquid–liquid interfaceR(water)/Fc(nitrobenzene) 62 R = FcCOO−,Ru(bpy)3

3+63 R = InCl63−, Fe(CN)6

3− ,Fe(phen)3

2+, Mo(CN)84−

Fe(CN)63− (water)/TCNQ(DCE) 64

Vitamin B12(water)/DBCH(Benzonitrile) 65 Investigating surfactant effectsZnPor (benzene)/Ru(CN)6

4− (water) 66Electron transfer across monolayer at liquid–liquid interfaceZnPor(benzene)/R (water) 67 Monolayer of saturated

phosphocholineR = Ru(CN)6

4−, Mo(CN)84−,

Fe(CN)64− , [Co(II)sepulchrate]2+,

V2+68 Monolayer of a mixture of

conjugated and saturatedphospholipids R = Fe(CN)6

4−Mass transfer through liquid/liquid or liquid–gas interfaceWater/heptane Cu2+ 69Water/DCE Cu2+ 69 Cu2+ in DCE was chelated by

oxymeO2 70Br2, Fe(CN)6

4− 71 Double potential stepchronoamperometry

K+ 72 K+ in DCE was transported withDB18C6

Utilizing a micropipette baseelectrode

Water/nitrobenzene O2 70FcCH2OH,Co(phen)3

2+ 73Water/air Br2 71 Double potential step

chronoamperometryMass transfer through monolayer at air–water interfaceWater/air O2 74Mass and charge transfer through BLMBLM H+ 75 Using potentiometric pH

microelectrodeH+, NH4

+ 76 Using NH4+ selective

microelectrode


Tab. 2 (continued)


H+, K+, Na+, Ca2+ 77 With double-barreledmicroelectrodes

FcCH2OH,FcCOOH/FcCOO−

78 Estimation of permiationcoefficients

BLM incorporated alamethicin ion channelRu(NH3)6

3+,Fe(CN)6

3−, I−79

BLM doped with iodine I− 80 Study of charge transfer throughBLM

Notes: TCNQ, 7,7,8,8-tetracyanoquinodimethane; DCE, 1,2-dichloroethane;Por, 5,10,15,20-tetraphenyl-210H,23H-porphine; DBCH, trans-1,2-dibromocyclohexane;DB18C6, dibenzo-18-crown-6.

Tab. 3 Studies on cells and tissues, including studies of extracellular time courserecording with microelectrodes

Sample Monitoring reagents Reference

Lingstrum sinenis leaf Fe(CN)64− 81

O2 81Elodea leaf O2,Ru(NH3)6

3+ 81Guard cells in tradescantia

fluminesisO2 82

Algal protoplast O2 83, 84O2, Fe(CN)6

4− 85Fe(CN)6

4−/3−, Co(phen)32+ ,

FcCH2OH, HQ86

p-BQ, O2 32SW-480(epithelial-like) O2 87Simian virus

40-transformedhuman fibroblast

H2O2, O2 88

Pancreatic islet O2 89Chinese hamster ovarian

cancer cellDoxorubicin 90

Detection of exocytosis at single cellsBovine adrenal medullary

cellEpinephrine, Norepinephrine 36, 91–97

PC12 (ratpheochromocytoma)

Dopamine 98

Mast cell Serotonin 99Pancreatic β-cell Insulin 100, 101Rat melanotrophs α-Melanocyte stimulating

hormone102

BQ, benzoquinone; HQ, hydroquinone.


8.2.1Enzymes

For SECM applications to biological sys-tems, the two electrode configuration (witha working and an auxiliary electrode) isavailable and the sample substrates aregenerally unbiased. The basics for theSECM operation described in the previoussection are valid and experiments are car-ried out in the feedback or the G/C mode.Typical characteristics for both modes canbe seen through the studies on immobi-lized enzymes.

A kinetic analysis of immobilized en-zymes has been carried out in the feedbackmode SECM. Bard and coworkers stud-ied kinetics between glucose oxidase andmediators at the enzyme-immobilized sur-faces [42]. When the probe approaches theenzyme-immobilized surface in the pres-ence of excess glucose, oxidation currentfor the mediator is enhanced due to theregeneration of the mediator at the sur-face. Utilizing digital simulation, approachcurves reflecting catalytic current increasedue to immobilized glucose oxidase wereanalyzed for a couple of mediators andfor several immobilizing procedures. Inthe simulation, a Michaelis–Menten typeequation expressing the reaction betweenthe mediator and the enzyme was adoptedfor the boundary condition at the biocat-alytic substrate surface. The equation wasfurther simplified to two limiting cases:zero- and first-order electron-transfer ki-netics for large and small mediator con-centrations, respectively. In the literature,they indicated a detection limit criterionfor feedback mode SECM [42, 43]:

Jmed > 10−3DC/a (2)

where Jmed is the flux for the mediator atthe enzyme-immobilized substrate surface

under assumption of a zero-order hetero-geneous enzymatic reaction, D and C arethe diffusion coefficient and bulk concen-tration profiles of the reagent, and a isthe radius of the probe microelectrode.This expression will be useful for judg-ing the applicability of the feedback modeoperation for the specific experimental cir-cumstance. The expression (2) suggeststhat the smaller probe radius for highresolution imaging in the feedback modeSECM could be canceled out by the lowersensitivity for the local enzymatic activity.Lower mediator concentration is desirableto detect the catalytic enhancement orig-inated from redox cycling between theprobe and the immobilized enzyme.

Yamada and coworkers carried out char-acterization of a diaphorase monolayerat a glass surface in the same opera-tion mode SECM [46]. Diaphorase purifiedfrom Bacillus stearothermophilus is a mem-brane protein, which catalyzes the oxida-tion of NADH in the presence of an elec-tron mediator. Because of the high activityof diaphorase, a Michaelis–Menten typeequation was applicable without furthersimplification for the boundary conditionat the substrate surface. In the follow-ing studies [47], diaphorase patterns withmonolayer-level coverage at flat glass sub-strates were visualized by feedback modeimaging.

Feedback mode SECM imaging wassuccessfully employed to map metabolicactivities with high spatial resolution. Het-erogeneous distributions of active NADHcytochrome reductase within individualmitochondria membrane surfaces were vi-sualized [43]. However, as suggested byBard’s group, this mode will not generallybe accessible for surfaces with relativelylower enzymatic activities. On the otherhand, the G/C mode may offer wider appli-cability for imaging biocatalytic reactions


at the sample surface. In fact, the first re-port on mapping of biocatalytic reactions atsolid supports was done by the G/C modeprocedure, in 1989 [41]. In this study, glu-cose oxidase and plant tissues containingenzymes were incorporated into carbonpaste. Recently, a glucose oxidase coat-ing on an alkanethiolate-gold substratewith monolayer-level coverage was visu-alized by monitoring the oxidation currentof H2O2 continuously generated by thesurface immobilized enzyme [45]. The res-olution was improved by adding catalase,which effectively eliminates H2O2 in thebulk solution. Enzyme electrodes [44] orpotentiometric pH probes [37] have beenapplied to observe enzymatically activeregions in samples, that is another advan-tage of the G/C mode over the FB modeoperation.

8.2.2Antigen–Antibodies

Electrochemical detection in immunoas-says has already been a great success byutilizing labeled enzymes to enhance sig-nals originating from antigen–antibodybinding. SECM strengthens the protocoldue to its capability of two-dimensionalvisualization of the distribution of anti-gen–antibody complexes with high spatialresolution. Because of the small vol-ume and small area required for anassay, total throughput can be drasti-cally improved compared with the cur-rent assays running in a 96-well plate.A study characterizing antigen–antibodycomplexes on a solid support was re-ported by Wittstock and coworkers [49].Alkaline phosphatase-labeled antigen wasused for sensing antigen–antibody bind-ing. The antigen-coated reaction zoneswere prepared at a glass substrate by

using the enzyme-labeled and the en-zyme free antigen solutions. In aqueoussolutions, the probe is laterally scannedalong the substrate surface to detect oxida-tion current of p-aminophenol generatedby the enzyme-catalyzed hydrolysis of p-aminophenyl phosphate. As the probedirectly detects the enzymatic product, noincubation time for the enzyme reaction isnecessary.

The combination of SECM withthe enzyme-linked immunosorbent assay(ELISA) offers a novel assay systemof biologically important materials. Wehave characterized CEA microspottedon a glass substrate [50]. The samplepreparation was basically similar to aconventional sandwich method utilizingHRP-labeled antibody, but analyte solutioncontaining CEA was microspotted in a20-µm-radius region with a tapered glasscapillary. In the presence of H2O2 andferrocenylmethanol, the probe detected thereduction current of ferriciniumethanolproduced by HRP-catalyzed reactionat the substrate. In a single SECMimage of 400 µm × 500 µm, ten spots forimmobilized CEA were clearly visualizedas the area of large reduction currentand, as a result, as low as 10 000 CEAmolecules per spot were detectable. Wehave extended this SECM/ELISA formultianalyte sensing system. In additionto CEA, HCG, HPL, and α-fetoproteinwere applied for dual- and multianalyteimmunoassays [51–53]. Figure 3 shows aset of SECM images taken in the dual-immunoassay of HCG and HPL [51].On each substrate, anti-HCG and anti-HPL were microspotted at the spatiallyaddressable reaction zone, namely, theleft for anti-HCG and the right for anti-HPL. Then, an analyte solution containingHCG and/or HPL was dropped onto thesubstrate, and then the substrate was


2 pA

1 pA

1 pA

X[µm]

4003002001000

(a)

(b)

(c)

HPLHCG

Fig. 3 SECM images and their cross-sections of the substrates for thedual-immunoassay. The substrate microspotted with anti-HCG (left arrow) and antiHPL (right arrow) were prepared with analyte solution containing HPL (a), HCG (b)or a mixture of HPL and HCG (c) (from ref. [51]. Copyright 1997 Elsevier Science.).

treated with a solution containing thetwo HRP-labeled antigens. The locationof signal at the reaction zones (Fig. 3)indicates the type of antigen in theanalyte solution. The calibration curvesfor both analytes were also shown inthe literature. This assay protocol will beapplicable to simultaneous detection ofvarious biomolecules.

8.2.3Local Fluxes Through Biological Materials

SECM affords a microscopic view ofheterogeneous flux distribution acrossporous membranes. The probe senses a

particular redox species by choosing asuitable electrode potential and, more-over, the local flux can be quantitativelydetermined by analyzing SECM images.White and coworkers studied the SECMimaging of transport of redox speciesthrough porous membranes [31, 54–58].They used a galvanostatic mode to main-tain a steady state membrane current. Thelocal fluxes of a particular redox speciesthrough the membrane pores were mea-sured with a probe located close to themembrane surface. The G/C mode SECMimaging allows quantitative estimation ofboth the flux through individual pores andthe pore size. The experimental data can


be fitted with those calculated based onsemi-spherical diffusion equation to obtainfluxes. As samples, the hairless and nudemouse skin demonstrated measurementof the fluxes of ferrocyanide ion transferthrough the pores by SECM [58]. The over-all rate of ferrocyanide transport acrossthe skin was analyzed by either induc-tively coupled plasma-mass spectroscopyor atomic adsorption spectroscopy. The ra-tio of the ferrocyanide ion transfer acrossthe pores to that through the other partof the skin was discussed on the ba-sis of solution compounds or surfactanteffect [58].

Unwin and coworkers measured osmot-ically driven redox fluxes through porouslaryngeal cartilage [35]. In their experi-ments, the probe-sample distance wassignificantly small to allow high spatialresolution mapping of the flux distribu-tion, but, at the same time, the currentprofiles were affected by the probe-sampleseparation at each data point. Thus, theSECM image was a convolution of the lo-cal flux and topographical information atthe sample surface. In a separate measure-ment, they imaged the same area withoutan osmotic driving force to obtain a puretopography. Then, the topographic imagewas subtracted from the original imageto yield the two-dimensional flux distri-butions. One point should be mentioned,however, that their calculation was doneunder assumptions of the homogeneity ofthe flux at the sample surface within thearea of probe size and the axial symmetryto the probe center.

8.2.4Liquid–Liquid Interfaces, Liquid–AirInterfaces

Recently, electron transfer and masstransfer across liquid–liquid or liquid–air

interface have been vigorously studiedby SECM [6, 7]. Bard and Mirkin andcoworkers applied feedback mode SECMto analyze the electron transfer at theinterface between two immiscible elec-trolyte solutions (ITIES) [62]. The potentialbetween the two liquid phases is unbi-ased and controlled by changing the ratioof the concentration of supporting elec-trolyte in the two liquid phases. Becauseof the novel approach to direct detec-tion of local reactions occurring at theinterface, SECM/ITIES has made a re-markable impact overcoming traditionalapproaches.

Tsionski and coworkers extended thestudy to the electron transfer through alipid monolayer at a benzene–water inter-face [67]. The electron transfer reactionsbetween the oxidized form of zinc por-phyrin (ZnPor+) in a benzene phase andthe reduced form of a metal complex (R,R = Ru(CN)6

4−, Mo(CN)84−, Fe(CN)6

4−,and so forth (Table 2)) in an aqueousphase have been surveyed. At the probemicroelectrode surface, ZnPor0 was ox-idized to ZnPor+. When the probe ispositioned close to the benzene–water in-terface, ZnPor+ is reduced back to ZnPor0

by accepting an electron from R in theaqueous phase at the liquid–liquid in-terface. In the experiment, the drivingforce was controlled with two parame-ters; the difference in standard potentialsof the redox mediators in benzene andin water (E0), and the interfacial po-tential drop (o

w), which is controllableby varying the concentration ratio of abase electrolyte such as ClO4

− in thetwo liquids. The driving force dependenceon the electron transfer rate at the liq-uid–liquid interface has been shown inthe literature in the absence and pres-ence of the monolayer. The existence ofthe monolayer lowers the electron transfer


rate through the liquid–liquid interface.Interestingly, they observed an ‘‘invertedregion’’ where electron transfer rate de-clines as the overpotential is increased,which may be explainable by Marcustheory.

Unwin and coworkers studied oxygentransfer across a monolayer dispersed atan air–water interface [74]. They have con-structed an SECM measurement systemin a Langmuir trough. A monolayer of 1-octadecanol was formed on the aqueoussubphase containing potassium nitrateand the surface pressure of the mono-layer was controlled with a compressionbarrier. A U-shaped microelectrode probewas placed in the aqueous solution andmoved up to the air–water interface. Whenthe probe approaches an air–water inter-face in the absence of the monolayer, thereduction current for O2 is enhanced bypositive feedback, due to fast translocationof O2 from the gas phase to the solution.They recorded a set of approach curveswhile controlling the monolayer compres-sion. A plot of the oxygen transfer rate asa function of the molecular area has beenshown.

The same group also reported thefirst work that applied SECM for study-ing mass transport across an air–waterinterface [71]. By utilizing double poten-tial step amperometry, bromine transferthrough the air–water interface was in-vestigated. An anodic pulse of 1.20 Vversus AgQRE and 10-ms width was ap-plied in a solution containing bromideanion to generate bromine at a probe-interface separation of 2.9 µm, then, theprobe potential was stepped to 0.70 V toreduce bromine in the solution. On thebasis of digital simulation analysis, thebromine transfer at the interface is diffu-sion controlled in the time scale of theexperiment.

8.2.5Planar BLMs

Because of the structural similarity tobiomembranes, the planar BLM is anattractive experimental system to char-acterize ion and electron transport bymeasuring membrane potential or ion con-ductivity. The BLM is also used as a matrixfor carriers and ion channels that promotehighly selective and effective transport.The direct detection scheme of the SECMto probe local concentration shifts near theBLM supplements the information thatconventional impedance measurementsdid not support. Antonenko and coworkersmeasured local pH changes near a BLM bymeans of a potentiometric pH microelec-trode [75]. The pH profile evolves duringthe transport of weak bases such as ammo-nia through the BLM. Recently, they haveelaborated on this study by utilizing an am-monium ion-selective microelectrode. Themembrane permeability of ammonia wasdetermined from the direct measurementsof both H+ and NH4

+ concentration pro-files perpendicular to the BLM surface [76].Furthermore, double-barreled microelec-trodes responsive to H+, K+, Na+, or Ca2+were fabricated to simultaneously measuretwo cation concentrations within the un-stirred layer face at the BLM surfaces [77].Using a similar experimental procedure,they also studied permeation mediated byion carriers such as nigericin or tributyltinacross BLMs and estimated the exchangerates of K+/H+ and Cl−/OH− occurringnear BLM surfaces [75].

Yamada and coworkers studied BLMpermeation for hydrophobic ferrocene (Fc)derivatives by microvoltammetry [78]. Lin-ear sweep voltammetry has been carriedout with a microelectrode positioned closeto the BLM. Two voltammograms for the


reduction of Ru(NH3)63+ and for the ox-

idation of ferrocenecarboxylic acid wererecorded at the same probe-BLM sep-aration. The deviation behavior of thevoltammograms normal to that taken apartfrom the BLM has been found to bedifferent. Namely, the reduction currentfor Ru(NH3)6

3+ evidently decreased com-pared with that taken far from the BLMsurface. On the contrary, the decline of theoxidation current for ferrocenecarboxylicacid normal to that measured in the bulksolution was slight. The results obviouslyreflect the difference of the BLM perme-ability for the two redox species. Digitalsimulation was applied to analyze the per-meation phenomena considering severalprocesses including the transmembranemass transfer and the diffusion in thevicinity of the BLM surface. The perme-ation coefficients for ferrocenecarboxylicacid were quantitatively estimated fromthe analysis. They also discussed the pHeffect on BLM permeability for the fer-rocenecarboxylic acid. Recently, the sameexperimental procedure has been appliedto estimate permeability of several redoxspecies through a cell membrane of analgal protoplast [86].

Matsue and coworkers studied ion trans-portation process across voltage-gatedalamethicin ion channels incorporated

into a BLM [79]. Alamethicin, a polypep-tide antibiotic, forms barrel-type ion chan-nels penetrating through the BLM whenthe membrane potentials are applied.Figure 4 shows a schematic illustration ofthe experimental system. A redox speciessuch as Ru(NH3)6

3+ was added in thecis side. A microelectrode (W1) soakedin the trans side was positioned close tothe BLM incorporating alamethicin chan-nels. Then a positive membrane potentialpulse was applied at the working elec-trode 2 (W2), where the total ionic currentwas measured. Simultaneously, the probemicroelectrode detected the reduction cur-rent for Ru(NH3)6

3+, which permeatedfrom the cis to the trans side through thealamethicin ion channels. The chronoam-perometric responses were recorded atthe microelectrode and analyzed by digi-tal simulation to obtain the permselectivityfor several redox species through the chan-nels. Fe(CN)6

4− transportation throughthe channels have found to be restricteddue to its negative charge compared withcationic Ru(NH3)6

3+.

8.2.6Cells, Tissues

In this section, we note studies on moni-toring of intact biological samples such as

Fig. 4 Schematic illustration of theexperimental setup for characterizationof the redox permeability acrossalamethicin ion channels in a planarbilayer lipid membrane (from ref. [79].Copyright 1994 American ChemicalSociety.).

Ox

Red

e

W2 RE CE

Bipotentiostat

Ion channelCis Trans

A−

BLM

C+

W1


living cells or tissues. This application ofSECM has been receiving a great dealof attention because it can afford vastknowledge, which has not been gainedby the studies of living samples bymeans of microelectrodes [88–102]. Mi-croelectrodes have been applied to in vivo,intra- and extra-cellular electrochemistrysince 1970s that will be described else-where in this volume. The extracellulardetection scheme shares its technical ad-vantages and limitations with SECM tech-nique [88–102]. The probe-sample sepa-ration is a critical parameter for spatialand temporal resolutions of the exper-imental systems. Preclusion due to theconsumption of electrolysis at the mi-croelectrode probe should be consideredcarefully.

Exocytotic secretion of catecholaminesfrom a single adrenal medullary cell hasbeen investigated [36, 91–97]. A carbon-fiber microelectrode positioned adjacent tothe cell records a series of current spikesdue to the oxidation of catecholaminesreleased from individual vesicles aftertheir fusion with the cell membrane. Thespike-like responses were compared withthe data obtained by fast-scan voltam-metry [36]. The shape of the spikes alsolargely depends on the distance betweenthe probe and the cell surface. The smallerprobe-sample separation affords shortertime resolution and was found to resolvea rate-determining step in the secretionof catecholamines from the vesicle at thecell membrane. Spatially resolved currentresponses were obtained within the singlecell surface by changing lateral positionof a pair of microelectrodes [95]. To date,various studies characterized single cellresponses by positioning a microelectrodevery close to the surface have been re-ported [96–102].

SECM provides direct visual informationon topography and local electrochemicalactivity of biological materials. The im-ages of plant tissues have been reportedin an early SECM study to demonstratethe capability of SECM to explore livingsamples [81]. Recently, intact leaves havebeen studied. SECM measurements ex-hibited the topography of stomata andimages of O2 evolution from stomataand a single stomatal complex [82]. Pho-tosynthesis and respiration activities ofsingle cells were quantitatively evaluatedby monitoring the O2 concentration pro-file around a single plant cell [83]. The O2

consumption/generation rate of a singlealgal protoplast evidently depends on theillumination power of light irradiation. Inthe dark, the O2 concentration profile de-clines when approaching the cell surfacedue to O2 uptake by the respiration ofthe single cell. When light irradiated, thecell shows a positive gradient, due to O2

evolution by photosynthesis. Oxygen evo-lution was saturated with light intensityof more than 15 klux. From the concen-tration profile around the single cell, thetotal consumption/evolution rate for oxy-gen was quantitated. The electron transferfrom the photosynthetic electron-transportchain to an electron acceptor, known asthe Hill reaction, was also studied. Theinfluence of BQ on the respiratory andphotosynthetic activity of a single proto-plast was investigated in detail by usingSECM. BQ interacts with the respira-tory and photosynthetic electron-transportchains to accept electrons and, as a con-sequence, BQ itself is reduced to HQ.The analysis of localized concentrationsof BQ and HQ yielded the fluxes. The in-tracellular redox reactions were discussedbased on these fluxes and oxygen genera-tion/consumption rates [32].


References

1. A. J. Bard, F.-R. F. Fan, J. Kwak et al., Anal.Chem. 1989, 61, 132–138.

2. R. C. Engstrom, C. M. Pharr, Anal. Chem.1989, 61, 10 99A–11 04A.

3. A. J. Bard, G. Denuault, C. Lee et al., Acc.Chem. Res. 1990, 23, 357–363.

4. A. J. Bard, F.-R. F. Fan, D. T. Pierce et al.,Science 1991, 254, 68–74.

5. M. Arca, A. J. Bard, B. R. Horrocks et al.,Analyst 1994, 119, 719–726.

6. M. V. Mirkin, Anal. Chem. 1996, 68,177A–182A.

7. A. L. Barker, M. Gonsalves, J. V. Macpher-son et al., Anal. Chim. Acta 1999, 385,223–240.

8. L. A. Bottomley, Anal. Chem. 1998, 70,425R–475R.

9. L. A. Bottomley, J. E. Coury, P. N. First,Anal. Chem. 1996, 68, 185R–230R.

10. R. M. Nyffenegger, R. M. Penner, Chem.Rev. 1997, 97, 1195–1230.

11. H. Takano, J. R. Kenseth, S. Wong et al.,Chem. Rev. 1999, 99, 2845–2890.

12. http://www.msstate.edu/Dept/Chemistry/dow1/secm/secm bib.html.

13. A. J. Bard, F.-R. F. Fan, Faraday Discuss.1992, 94, 1–22.

14. F.-R. F. Fan, A. J. Bard, Science 1995, 267,871–874.

15. F.-R. F. Fan, J. Kwak, A. J. Bard, J. Am.Chem. Soc. 1996, 118, 9669–9675.

16. F. Zhou, A. J. Bard, J. Am. Chem. Soc. 1994,116, 393, 394.

17. D. A. Treichel, M. V. Mirkin, A. J. Bard,J. Phys. Chem. 1994, 98, 5751–5757.

18. J. Kwak, A. J. Bard, Anal. Chem. 1989, 61,1221–1227.

19. D. O. Wipf, A. J. Bard, J. Electrochem. Soc.1991, 138, 469–474.

20. A. J. Bard, M. V. Mirkin, P. R. Unwin,J. Phys. Chem. 1992, 96, 1861–1868.

21. M. V. Mirkin, T. C. Richards, A. J. Bard,J. Phys. Chem. 1993, 97, 7672–7677.

22. P. R. Unwin, A. J. Bard, J. Phys. Chem.1991, 95, 7814–7824.

23. F. Zhou, P. R. Unwin, A. J. Bard, J. Phys.Chem. 1992, 96, 4917–4924.

24. C. Demaille, P. R. Unwin, A. J. Bard, J. Phys.Chem. 1996, 100, 14 137–14 143.

25. R. D. Martin, P. R. Unwin, J. Electroanal.Chem. 1997, 439, 123–136.

26. A. J. Bard, L. R. Faulkner, ElectrochemicalMethods, Wiley & Sons, New York, 1980.

27. R. C. Engstrom, M. Weber, D. J. Wunderet al., Anal. Chem. 1986, 58, 844–848.

28. R. C. Engstrom, T. Meaney, R. Tople et al.,Anal. Chem. 1987, 59, 2005–2010.

29. R. C. Engstrom, R. M. Wightman, E. R.Kristensen, Anal. Chem. 1988, 60, 652–656.

30. R. C. Engstrom, B. Small, L. Kattan, Anal.Chem. 1992, 64, 241–244.

31. B. D. Bath, R. D. Lee, H. S. White et al.,Anal. Chem. 1998, 70, 1047–1058.

32. T. Yasukawa, I. Uchida, T. Matsue, Bio-phys. J. 1999, 76, 1129–1135.

33. R. D. Martin, P. R. Unwin, Anal. Chem.1998, 70, 276–284.

34. R. D. Martin, P. R. Unwin, J. Chem. Soc.Faraday Trans. 1998, 94, 753–759.

35. J. V. Macpherson, D. O’Hare, P. R. Unwinet al., Biophys. J. 1997, 73, 2771–2781.

36. R. M. Wightman, T. J. Schroeder, J. M.Finnegan et al., Biophys. J. 1995, 68,383–390.

37. B. R. Horrocks, M. V. Mirkin, D. T. Pierceet al., Anal. Chem. 1993, 65, 1213–1224.

38. G. Denuault, M. H. T. Frank, L. M. Peter,Faraday Discuss. 1992, 94, 23–35.

39. C. Wei, A. J. Bard, G. Nagy et al., Anal.Chem. 1995, 67, 1346–1356.

40. H. S. Issacs, G. Kissel, J. Electrochem. Soc.1972, 119, 1628–1632.

41. J. Wang, L.-H. Wu, D. Li, J. Electroanal.Chem. 1989, 272, 285–292.

42. D. T. Pierce, P. R. Unwin, A. J. Bard, Anal.Chem. 1992, 64, 1795–1803.

43. D. T. Pierce, A. J. Bard, Anal. Chem. 1993,65, 3598–3604.

44. B. R. Horrocks, D. Schmidtke, A. Helleret al., Anal. Chem. 1993, 65, 3605–3614.

45. G. Wittstock, W. Schuhmann, Anal. Chem.1997, 69, 5059–5066.

46. H. Yamada, H. Shiku, T. Matsue et al.,Bioelectrochem. Bioenerg. 1994, 33, 91–93.

47. H. Shiku, T. Takeda, H. Yamada et al.,Anal. Chem. 1995, 67, 312–317.

48. H. Shiku, I. Uchida, T. Matsue, Langmuir1997, 13, 7239–7244.

49. G. Wittstock, K. Yu, H. B. Halsall et al.,Anal. Chem. 1995, 67, 3578–3582.

50. H. Shiku, T. Matsue, I. Uchida, Anal. Chem.1996, 68, 1276–1278.

51. H. Shiku, Y. Hara, T. Matsue et al.,J. Electroanal. Chem. 1997, 438, 187–190.


52. H. Balets, W. Gopel, J. Hess, (Eds.), SensorsUpdate, vol.6, Chap. 12, Wiley-VCH, Wein-heim, 2000.

53. H. Shiku, Y. Hara, T. Takeda et al., ACSSymp. Ser. 1997, 656, 202–209.

54. E. R. Scott, H. S. White, J. Membr. Sci. 1991,58, 71–87.

55. E. R. Scott, H. S. White, Solid State Ionics1992, 53–56, 176–183.

56. E. R. Scott, H. S. White, J. B. Phipps, Anal.Chem. 1993, 65, 1537–1545.

57. E. R. Scott, A. I. Laplaza, H. S. White et al.,Pharm. Res. 1993, 10, 1699–1707.

58. E. R. Scott, J. B. Phipps, H. S. White,J. Invest. Dermatol. 1995, 104, 142–145.

59. J. V. Macpherson, M. A. Beeston, P. R.Unwin et al., J. Chem. Soc. Faraday Trans.1995, 91, 1407–1410.

60. J. V. Macpherson, M. A. Beeston, P. R.Unwin, Langmuir 1995, 11, 3959–3963.

61. S. Nugues, G. Denuault, J. Electroanal.Chem. 1996, 408, 125–140.

62. C. Wei, A. J. Bard, M. V. Mirkin, J. Phys.Chem. 1995, 99, 16 033–16 042.

63. Y. Selzer, D. Mandler, J. Electroanal. Chem.1996, 409, 15–17.

64. T. Solomon, A. J. Bard, J. Phys. Chem. 1995,99, 17 487–17 489.

65. Y. Shao, M. V. Mirkin, J. F. Rusling, J. Phys.Chem. B 1997, 101, 3202–3208.

66. M. Tsionsky, A. J. Bard, M. V. Mirkin,J. Phys. Chem. 1996, 100, 17 881–17 888.

67. M. Tsionsky, A. J. Bard, M. V. Mirkin,J. Am. Chem. Soc. 1997, 119, 10 785–10 792.

68. M.-H. Delville, M. Tsionsky, A. J. Bard,Langmuir 1998, 14, 2774–2779.

69. C. J. Slevin, J. A. Umberes, J. H. Athertonet al., J. Chem. Soc. Faraday Trans. 1996, 92,5177–5180.

70. A. L. Barker, J. V. Macpherson, C. J. Slevinet al., J. Phys. Chem. B 1998, 102,1586–1598.

71. C. J. Slevin, J. V. Macpherson, P. R. Unwin,J. Phys. Chem. B 1997, 101, 10 851–10 859.

72. Y. Shao, M. V. Mirkin, J. Electroanal. Chem.1997, 439, 137–143.

73. H. Yamada, S. Akiyama, T. Inoue et al.,Chem. Lett. 1998, 147–148.

74. C. J. Slevin, S. Ryley, D. J. Walton et al.,Langmuir 1998, 14, 5331–5334.

75. Y. N. Antonenko, A. A. Bulychev, Biochim.Biophys. Acta 1991, 1070, 279–282,474–480.

76. Y. N. Antonenko, P. Pohl, G. A. Denisov,Biophys. J. 1997, 72, 2187–2195.

77. P. Pohl, S. M. Saparov, Y. N. Antonenko,Biophys. J. 1998, 75, 1403–1409.

78. H. Yamada, T. Matsue, I. Uchida, Bio-chem. Biophys. Res. Commun. 1991, 180,1330–1334.

79. T. Matsue, H. Shiku, H. Yamada et al.,J. Phys. Chem. 1994, 98, 11 001–11 003.

80. M. Tsionski, J. Zhou, S. Amemiya et al.,Anal. Chem. 1999, 71, 4300–4305.

81. C. Lee, J. Kwak, A. J. Bard, Proc. Natl. Acad.Sci. U.S.A. 1990, 87, 1740–1743.

82. M. Tsionsky, Z. G. Cardon, A. J. Bard et al.,Plant Physiol. 1997, 113, 895–901.

83. T. Yasukawa, I. Uchida, T. Matsue, DenkiKagaku 1998, 66, 660–661.

84. T. Yasukawa, T. Kaya, T. Matsue, Chem.Lett. 1999, 975–976.

85. T. Yasukawa, T. Kaya, T. Matsue, Anal.Chem. 1999, 71, 4637–4641.

86. T. Yasukawa, I. Uchida, T. Matsue, Biochim.Biophys. Acta 1998, 1369, 152–158.

87. T. Yasukawa, Y. Kondo, I. Uchida, T. Mat-sue, Chem. Lett. 1998, 767, 768.

88. S. Arbault, P. Pantano, J. A. Jankowski et al.,Anal. Chem. 1995, 67, 3382–3390.

89. S.-K. Jung, W. Gorski, C. A. Aspinwall et al.,Anal. Chem. 1999, 71, 3642–3649.

90. H. Lu, M. Gratzl, Anal. Chem. 1999, 71,2821–2830.

91. D. J. Leszczyszyn, J. A. Jankowski, O. H.Viveros et al., J. Biol. Chem. 1990, 265,14 736–14 737.

92. R. M. Wightman, J. A. Jankowski, R. T.Kennedy et al., Proc. Natl. Acad. Sci. U.S.A.1991, 88, 10 754–10 758.

93. K. T. Kawagoe, J. A. Jankowski, R. M.Wightman, Anal. Chem. 1991, 63,1589–1594.

94. T. J. Schroeder, J. A. Jankowski, K. T.Kawagoe et al., Anal. Chem. 1992, 64,3077–3083.

95. T. J. Schreoder, J. A. Jankowski, J. Seny-shyn et al., J. Biol. Chem. 1994, 269,17 215–17 220.

96. R. H. Chow, L. von Ruden, E. Neher, Na-ture 1992, 356, 60–63.

97. R. H. Chow, J. Klingauf, E. Neher, Proc. Natl.Acad. Sci. U.S.A. 1994, 91, 12 765–12 769.

98. T. K. Chen, G. Luo, A. G. Ewing, Anal.Chem. 1994, 66, 3031–3035.


99. G. Alvarez de Toledo, R. Fernandez-Chacon, J. M. Fernandez, Nature 1993, 363,554–558.

100. R. T. Kennedy, L. Huang, M. A. Atkinsonet al., Anal. Chem. 1993, 65, 1882–1887.

101. L. Huang, H. Shen, M. A. Atkinson et al.,Proc. Natl. Acad. Sci. U.S.A. 1995, 92,9608–9612.

102. C. D. Paras, R. T. Kennedy, Anal. Chem.1995, 67, 3633–3637.

277

9Ion-Selective Electrodes forMeasurements in BiologicalFluids

Eric BakkerAuburn University, Auburn, Alabama

Mark E. MeyerhoffThe University of Michigan, Ann Arbor, Michigan

9.1 Zero-current Electrochemical Cells Incorporating Ion-selectiveMembranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279

9.2 Selectivity of Potentiometric Membrane Electrodes . . . . . . . . . . . . 284

9.3 Glass Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286

9.4 Solid-state Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290

9.5 Liquid Membrane Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . 2919.5.1 Ion-exchangers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2929.5.2 Neutral Ionophores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2939.5.3 Charged Ionophores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297

9.6 Gas-sensing Probes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298

9.7 Ion-selective Electrode-based Biosensors . . . . . . . . . . . . . . . . . . . 302References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306

279

9.1Zero-current Electrochemical CellsIncorporating Ion-selective Membranes

Ion-selective electrodes (ISEs) are rela-tively simple yet important electroanalyt-ical devices that are now used routinelyto quantitate a large number of differ-ent analytes in complex biological sam-ples. In many instances, these sensorscan be miniaturized to yield micrometer-sized sensing tips for intracellular sensingapplications [1], for the mapping of chem-ical gradients [2], and for scanning probeelectrochemical microscopy [3]. The timeresponse of ISEs is often rapid (withinmilliseconds in some cases), and theyare characterized as nonperturbing mea-surement devices because the analyte isnormally not consumed in the course ofthe measurement. Further, they are amongthe very few analytical methods capable ofmeasuring ion activities, rather than totalconcentrations [4].

Measurements with ISEs are carried outusing an electrochemical cell similar tothe basic design depicted in Fig. 1. Theactual selective ion-sensing process is con-fined to the ion-selective membrane, whichcontacts the sample on one side and aninternal filling solution on the other. Inplanar miniaturized systems, the latteroften consists of a hydrophilic polymer

(hydrogel) doped with suitable electrolytesand saturated with water [5]. An internalreference electrode, usually silver–silverchloride, completes the measuring cellwithin the ion-selective electrode. A similarreference electrode also contacts the sam-ple through a so-called bridge electrolyte,an aqueous solution or hydrogel with anelectrolyte composition that is compatiblewith the sample to be measured. This junc-tion between the bridge electrolyte and thesample phase yields a small, ideally invari-ant liquid junction potential as the samplecomposition is changed owing to differ-ences in the mobilities of the ions on eachside of this liquid–liquid junction [6, 7].

As Fig. 2 illustrates, the observed zero-current cell potential (electromotive force,emf) is a sum of each individual poten-tial drop that occurs across the entirecell. Indeed, each vertical line in Fig. 2represents an interface where a bound-ary potential develops. It is the potentialacross the ion-selective membrane that willbe directly dependent on the ion compo-sition of the sample and the contactinginner electrolyte. Therefore, all other po-tential contributions must be reversibleand constant, so that observed cell poten-tial (Ecell) changes can be attributed tochanges in the membrane potential only,which in turn are indicative of changesin the ion activity in the sample phase.

280 9 Ion-Selective Electrodes for Measurements in Biological Fluids

Innerelectrolyte

Ag/AgCl

Ion-selectivemembrane

EMF

Bridge electrolyte

Capillary

Frit

Referenceelectrolyte

Ag/AgCl

Ion-selective electrode Reference electrode

Aqueoussample

Fig. 1 Experimental assembly for measuring ISEs in a zero current galvanic cell.

SampleExternalreferenceelectrode


Internalsolution

Internalreferenceelectrode

A B C D E

1 2 3 4

Fig. 2 Basic cell notation of ISE cell assembly. Numbers and vertical bars denoteboundary potentials. The sum of all individual boundary potentials make up theobserved cell potential.

The electrode that houses the ion-selectivemembrane, the inner electrolyte, and theinner reference electrode is called the in-dicator electrode or working electrode. Ifthe external reference electrode containsan interchangeable bridge electrolyte, it isoften termed a double junction referenceelectrode to indicate that two separate liq-uid junction potentials are present (onebetween the reference electrolyte and thebridge electrolyte, and the second be-tween the bridge electrolyte and the samplesolution). The use of combination pH glass

electrodes is commonplace (that is, indica-tor and reference electrode are combinedinto one housing), although two separateelectrode housings are still often usedwhen working with most other ISE in-dicator electrodes.

As mentioned earlier, the observed cellpotential Ecell is the sum of all potentialcontributions in the cell. Therefore, ac-cording to Fig. 2, Ecell is written as a func-tion of the Galvani potential differencesbetween the individual solutions andelectrode materials as follows:

9.1 Zero-current Electrochemical Cells Incorporating Ion-selective Membranes 281

Ecell = (φE − φD) + (φD − φC)

+ (φC − φB) + (φB − φA)

= E4 + E3 + E2 + E1 (1)

If all potential changes are constant, withexception of the phase boundary potentialE2, Eq. (1) simplifies to:

Ecell = K + E2 (2)

where K = E1 + E3 + E4. The boundarypotential E3 between the ion-selectivemembrane and the inner electrolyte so-lution is typically constant as the ioniccomposition of this inner electrolyte doesnot change in most practical situations.The boundary potential at the mem-brane/sample interface, E2, is in idealcases a direct function of the sample ionactivity. It is assumed that sample ionscan rapidly exchange with the ion-selectivemembrane material so that interfacialequilibrium must exist. Any ion will spon-taneously partition in the space chargeregion according to its standard chemi-cal potential difference and to its so-calledfree ion activity in each phase. Hydrophilicions prefer the aqueous phase, whereasions with a high affinity for the mem-brane material will rather distribute to themembrane phase. Since these ions carry acharge, this process spontaneously formsa space charge region, very locally at theinterface, and yields a phase boundary po-tential that perfectly counterbalances thechemical tendency of each ion to partitioninto one or the other phase. In practice,therefore, each bulk phase remains essen-tially electroneutral. This process can beexpressed mathematically as follows:

The electrochemical potential for ion Iin the aqueous phase, µI, is as:

µI(aq) = µI(aq) + zFφ(aq) = µ0I (aq)

+ RT ln aI(aq) + zFφ(aq) (3)

and for the contacting ion-selective mem-brane phase:

µI(mem) = µI(mem) + zFφ(mem)

= µ0I (mem) + RT ln aI(mem)

+ zFφ(mem) (4)

where µI is the chemical potential ofion I (µ0

I under standard conditions), z

is the charge, and aI the activity of theuncomplexed ion I, φ is the electricalpotential, and R, T , and F are the universalgas constant, the absolute temperature,and the Faraday constant. Since it isassumed that the interfacial ion transferand complexation processes are fast andthat, therefore, equilibrium holds at theinterface, the electrochemical potentialsfor both phases must be equal. This leadsto the following expression for the phaseboundary potential EPB (which is identicalto E2 shown in Eqs. 1 and 2):

EPB = φ = −µ0I (mem) − µ0

I (aq)

zF

+ RT

zFln

aI(aq)

aI(mem)(5)

The term comprising the standard chem-ical potentials is often combined to thesymbol kI, that is, kI = exp(µ0

I (aq) −µ0

I (org)/RT ). In this case, Eq. (5) sim-plifies to:

EPB = RT

zFln

kIaI(aq)

aI(mem)(6)

The concept of spontaneous interfacial iondistribution in the space charge regionand its influence on the observed bound-ary potential is mathematically formulatedin Eq. (6). Ions with a higher free en-ergy of transfer from the sample to themembrane phase, a higher sample ac-tivity, and/or a lower membrane activityobviously lead to a more positive boundary


potential (when I = cation). This potentialcounteracts the tendency of the ions tofurther distribute unequally across theinterface. Since aI(mem) is usually con-stant, as is kI, Eq. (2), taken together withEq. (6) reduces to the well-known Nernstequation describing the response of theelectrochemical cell toward ion I:

Ecell = E0I + RT

zFln aI(aq) (7)

Note that in this equation all of the sample-independent emf contributions are in-cluded into one constant potential term(E0

I ). In practice, ISEs obey this equationwhen the ion-selective membrane interactsreversibly with the specific ion of interest;that is, a rapid ion transfer must occurbetween the sample and the sensing mem-brane (high exchange current density). Onthe other hand, this partitioning processmust have no substantial effect on thechemical composition of the ion-selectivemembrane (that is, aI(mem) must remainconstant to obtain Nernstian response).Silver halide precipitate membranes, forexample, can be used as ion-selectivesensing materials as long as the surfacecomposition remains intact and does notbecome fouled by other silver precipitates.

A Nernstian response to silver ions can beobserved since silver can reversibly interact(precipitate and dissolve) with the mem-brane and the concentration of silver inthe membrane is invariant (unit activity forsolid phase). Analogous cases can be madefor other suitable membrane materials.

As can be seen in Eq. (7), the observedpotential change is proportional to thelogarithm of the sample ion activity andto the absolute temperature and inverselyproportional to the charge of the analyteion. For measurements performed at25 C, Eq. (7) reduces to the following ifpotentials are measured in millivolts:

Ecell(mV) = E0I + 59.2

zlog aI(aq) (8)

The sensitivity of the analytical techniqueis therefore mainly dependent on thecharge z of the analyte. For monovalentor divalent cations, for example, a 10-foldconcentration change will yield +59.2 mVor +29.6 mV change in the observed cellpotential. Anionic analytes (with negativez values) will induce negative poten-tial changes. Figure 3 shows a typicalcalibration curve for a calcium-selectiveelectrode based on a liquid membrane

−12

50 mV

EM

F

−10 −8

log aCa

−6 −4

Fig. 3 Potentiometric response of ahighly selective calcium electrode on thebasis of the lipophilic calciumionophore ETH 129 (see Ca2+-1 inFig. 4), measured in calcium bufferedsolutions with a physiological electrolytebackground.

9.1 Zero-current Electrochemical Cells Incorporating Ion-selective Membranes 283

doped with the neutral ionophore ETH 129(see structure Ca2+-1 in Fig. 4) [8]. Thesamples containing low sample activitiesare buffered with the calcium chelatorethylenediamine tetraacetate (EDTA). Thiscalcium buffer eliminates the influence ofcalcium ions leaching from the membrane

into the sample, which ordinarily dictatesthe lower detection limit. ISEs are capa-ble of assessing free rather than total ionconcentrations or activities. This experi-mental distinction between free and com-plexed forms of the analyte makes themquite useful for bioavailability studies.

N

O O

H+- 2

O O

O

O

OO

OO

O

O

OO

NH4+- 1

N

OO

O

N

Ca2+- 1

OO

NHO

O

O

NHO

K+- 1

N

H+- 1

O OC2H5 O OC2H5

CH2

OO

O O

O O

N

S

S S

S

N N

O

N

O

N

O O

N N

O O O

N

O

H

OO

N

H

Zn2+- 1Ag+- 1

Mg2+- 1

Li+- 1

Na+- 1

3

Fig. 4 Structures of typical cation-selective ionophores used in organic ISE membranes.


In routine blood analysis of electrolytes,where ISEs are used nearly universally,sometimes extremely small concentrationchanges are assessed with direct poten-tiometry. This requires potential stabil-ities and reproducibilities in the 10 to100 microvolt range, which is achieved intemperature-controlled flow-through cellsand with frequent, automated recalibra-tions between measurements. In batchmode benchtop analyses with ISEs, sucha high precision is often not observed. Inaddition, accuracies are mainly limited byvariations in the liquid junction potentialbetween the calibration and sample phasesand by interferences from other sam-ple ions, temperature fluctuations, and ifconcentrations rather than activities aredesired, variations in activity coefficients.The latter reflects the well-known relation-ship between the sample activity aI and itsconcentration cI:

aI = γIcI (9)

where γI is the activity coefficient. Manypossible experimental biases can be min-imized in benchtop analysis by adding anionic strength adjusting solution to thesample prior to measurement. ISEs arealso routinely integrated into automatictitration instruments where they makeup excellent endpoint indicators. In thesecases, the accuracy and precision of the cellpotential readings affect the accuracy of theanalytical measurement to a much smallerextent than in direct potentiometry, wherethe observed Ecell is used to quantitate theanalyte level in the sample phase.

9.2Selectivity of Potentiometric MembraneElectrodes

For ISEs, interferences by other sam-ple ions are mainly dictated by their

competitive partitioning into the ion-selective membrane phase. Often, there-fore, ISEs are well understood in thesense that their response behavior is usu-ally fully predictable from thermodynamicconstants, such as the free enthalpies oftransfer of the uncomplexed ions fromthe sample into the membrane. The se-lectivity can be characterized with onethermodynamically founded potentiomet-ric selectivity coefficient K

potIJ for each

interfering ion. It is obtained from thesimple relationship, which directly relatesthe selectivity coefficient to the differencein the two E0 values for both mea-sured ions:

KpotIJ = exp

E0

J − E0I

RTzIF

(10)

where zI is the charge of the so-called pri-mary ion I and J is any interfering ion.The subscript IJ in the symbol K

potIJ de-

scribes the chosen primary and interferingions. For example, the selectivity coeffi-cient for a sodium-selective electrode overinterfering calcium ions is characterizedwith the symbol K

potNa,Ca. Evidently, the se-

lectivity coefficient is a direct function ofthe differences of the individual potentialsextrapolated to 1-M activity for the ions Iand J (E0

J − E0I ); that is, the vertical shift

between the single calibration curves at1M sample activities. Correct K

potIJ values

are obtained from adequate, unbiased E0

measurements for each ion.Historically, potentiometric selectivity

coefficients were used to assess the extentof interference in a mixed sample byuse of the so-called Nicolsky-Eisenmanequation [9]:

Ecell = E0I + RT

zIFln

(aI +

∑K

potIJ a

zI/zJ

J

)(11)

9.2 Selectivity of Potentiometric Membrane Electrodes 285

The activity term in the Nernst Eq. (7) isextended by a sum of selectivity-weightedactivities of interfering ions. In essence,the selectivity coefficient must be as smallas possible for each interferent ion sothat Eq. (11) reduces toward Eq. (7). TheNicolsky-Eisenman equation has been ex-tensively used in the literature for the pastthree decades. Unfortunately, it fails to giveaccurate mixed potential predictions if thecharges of the primary and interfering ionsare not identical [10, 11]. Nonetheless, it isstill valuable to give rough estimates of theextent of interference, and is accurate forthe response regions where only one ionis the predominant potential determiningspecies. For liquid membrane electrodes(see following text), the following equationhas been developed that offers more accu-rate predictions of the cell emf value in asample containing a mixture of monova-lent and divalent ions [12]:

Ecell = E0I + RT

Fln

1

2

∑i1

Kpot1/zI

I,i1 ai1(aq)

+

√√√√√√√√√

(1

2

∑i1

Kpot1/zI

I,i1 ai1(aq)

)2

+∑

i2

Kpot2/zI

I,i2 ai2(aq)

(12)

The indices i1 and i2 under the sumsin Eq. (12) indicate that only the sampleactivities of monovalent (i1) or divalentions (i2) are summed. The summationsinvolve all extractable ions, including theanalyte ion I zI , in which case K

potI,I = 1.

Equation (12) is valid for any numberof monovalent and divalent ions in thesample. It has been extended to cover ionsof higher valency as well [12].

In general, it is desired to use ion-selective electrode membranes in biologi-cal samples that exhibit very small selectiv-ity coefficients to ensure accurate analyticalresults. Before one embarks on the op-timization of ISE selectivity, however, itis crucial to evaluate the expected com-position of the samples to be measured.Often, seemingly nonideal selectivities areirrelevant for a practical application if theprimary ion is relatively concentrated andthe interferent is dilute or virtually nonex-istent. The following equation has beendeveloped for liquid membrane ISEs topredict required maximum selectivity co-efficients if any one interferent ion is notallowed to give a larger than 1% interfer-ence [10]:

KpotIJ (required) = aI

(100aJ)zI/zJ

(13)

Consider, for example, the measurementof a sample that contains equal 1.0-mMactivities of NaCl, KCl, and CaCl2. Ac-cording to Eq. (13), a sodium-selectiveelectrode must show the following se-lectivity coefficients to reliably quantitatesodium in this sample: K

potNa,K < 0.010,

and KpotNa,Ca < 0.0031. On the other hand,

a calcium-selective electrode must onlyexhibit Kpot

Ca,Na = KpotCa,K < 0.10. If the sam-

ple were to be diluted 10-fold, the re-quirements would change to K

potNa,K <

0.010, and KpotNa,Ca < 0.0010, and K

potCa,Na =

KpotCa,K < 1.0. These examples show that re-

quired selectivity coefficients involving twoions of the same charge are simple to pre-dict, but Eq. (13) must be used to predictcases with ions of different charges. Theeffect of sample dilution on the requiredselectivity coefficients can be easily under-stood. ISE responses to divalent ions showhalf the slope of monovalent ions (Eq. 8).Sample dilution will therefore decrease the


cell potential for the monovalent ion morestrongly than for the divalent ion, therebyrequiring a less selective electrode if thedivalent ion is to be assessed.

Table 1 tabulates the normal measuringranges of some common cations in blood,and required maximum selectivity coeffi-cients of ISEs used to assess these ionsin undiluted and diluted blood samples.Required K

potIJ are calculated according to

Eq. (13) by using the minimum primaryion and maximum interfering ion activi-ties shown in the Table.

Analogous predictions can be made forother target samples, and the requiredselectivity coefficients form a basis for ISEevaluation and optimization.

As shown specifically in the follow-ing sections, different ion-selective mem-brane materials exhibit selectivities accord-ing to fundamentally different processes,including adsorption, precipitation, liq-uid–liquid partitioning, and complexa-tion. These processes can be optimizedto yield ISEs that satisfy a given analyticalrequirement.

9.3Glass Electrodes

The most widely used ISE is, withoutdoubt, the pH glass membrane electrode.Nearly one hundred years ago, Cremerreported its basic behavior [13]. pH glasselectrodes have since evolved into highlyreliable sensors as they show a wide dy-namic measurement range, high selectiv-ity, and good temperature resistance [14].The pH response of glass electrodes arisesfrom a hydrated surface layer that containsfixed SiO− ion-exchanger sites that arehighly selective to the hydrogen ion ac-tivity. pH glass electrodes often displayextraordinary potentiometric selectivitiesfor H+ over other cations (Table 2), whichis largely due to the favorable equilibriumfor the primary equilibrium reaction occur-ring within the hydrated layer of the glass:

SiO− + H+ ←−→ SiOH

Interestingly, there are still some de-bates concerning the exact response

Tab. 1 Normal concentration range of some cations in blood and required maximumselectivity coefficients Kpot

IJ of ISEs for their reliable assessment by direct potentiometry inundiluted and 10-fold diluted blood samples, as calculated according to Eq. (13)

Ion I aI(min) aI(max) LogKpotIJ (required) in blood

Interferent J Undiluted 10-fold diluted

Na+ 1.02 × 10−1 1.12 × 10−1 K+ −0.6 −0.6Mg2+ −0.2 −0.7Ca2+ −0.3 −0.8

K+ 2.60 × 10−3 3.70 × 10−3 Na+ −3.6 −3.6Mg2+ −1.8 −2.3Ca2+ −1.9 −2.4

Mg2+ 1.60 × 10−4 2.80 × 10−4 Na+ −5.9 −4.9K+ −2.9 −1.9Ca2+ −2.4 −2.4

Ca2+ 3.40 × 10−4 4.10 × 10−4 Na+ −5.6 −4.6K+ −2.6 −1.6Mg2+ −1.9 −1.9

9.3 Glass Electrodes 287

Tab. 2 Membrane compositions and selectivities of important ISEsa

Analyte Membrane type Membrane composition LogKpotIJ

H+ Glass 72.2% SiO2, 6.4% CaO, 21.4% Na2O (mol%) Na+ : −11,K+ : −11

H+ Polymer Tri-n-dodecylamine, KTpClPB, PVC,bis-2-ethylhexylsebacate

Na+: −10.4,K+ : −9.8,Ca2+: <−11.1

H+ Polymer Octadecyl isonicotinate, KTpClPB, PVC,ortho-nitrophenyloctylether

Li+ : −6.9,Na+: −5.6,K+: −4.4

Li+ Polymer 7-tetradecyl-2,6,9,13-tetraoxa-tricyclo[12.4.4.01.14] docosane, KTpClPB,PVC, bis(benzylphenyl)adipate

Na+: −3.1,K+: −3.6,NH4

+: −3.8,Ca2+: <−5.0

Na+ Glass 11% Na2O, 18% Al2O3, 71% SiO2 K+: −2,Ag+: +2.6,NH4

+: −4.2,H+: 1 to 2.5

Na+ Polymer Calix[4]arenecrown-4 ionophore, KTpClPB,PVC, ortho-nitrophenyloctylether

Li+: −2.8,K+: −5.0,NH4

+: −4.4,Mg2+: −4.5,Ca2+: −4.4

K+ Polymer Valinomycin, NaTFPB, PVC,bis-2-ethylhexylsebacate

Na+: −4.5,Mg2+: −7.5,Ca2+: −6.9

NH4+ Polymer Nonactin/Monactin, KTpClPB, PVC,

ortho-nitrophenyloctyletherLi+: −2.9,Na+: −2.3,K+: −1.1,Mg2+: −4.0,Ca2+: −4.0

Mg2+ Polymer Double-armed dizazacrown ether ionophore(see Mg2+-1 in Fig.4), KTpClPB, PVC,ortho-nitrophenyloctylether

Li+: −3.7,Na+: −3.2,K+: −1.4,NH4

+: −2.0,Ca2+: −2.5

Ca2+ Polymer N,N,N′,N′-tetracyclohexyl-3-oxapentanediamide (ETH 129), KTpClPB,PVC, ortho-nitrophenyloctylether

Na+: −8.3,K+: −10.1,Mg2+: −9.3

Ag+ Solid-State Ag2S Cu2+: −6,Pb2+: −6 to −9,H+: −5,Hg2+: −2

Ag+ Polymer Methylenebis(diisobutyldithiocarbamate),NaTFPB, PVC, bis-2-ethylhexylsebacate

Na+: −8.7,K+: −8.2,Ca2+: −11.0,Cu2+: −10.5,Pb2+: −10.3

(continued overleaf)


Tab. 2 (continued)


Zn2+ Polymer N-benzyliminodiacetic acidbis(N-ethyl-N-cyclohexylamide, KTpClPB,PVC, ortho-nitrophenyloctylether

Li+, Na+: −2.7,K+: −2.5,NH4

+: −3.1,Ca2+: −2.8,Cd2+: −3.6,Pb2+: −2.1

F− Solid-State Single LaF3 crystal OH−: −1,Br−: −4,Cl−: −4,HCO3

−: <−3Cl− Polymer 2,7-di-tert-butyl-9,9-dimethyl-4,5-

xanthenediamine, MTDDACl, PVC,ortho-nitrophenyloctylether

Sal−: +1.8,SCN−: +1.6,NO3

−: +0.7,HCO3

−: −2.6I− Solid-State 50 mol% Ag2S, 50 mol% AgI Cl−: −4,

Br−: −7,SCN−: −4,S2−: >10

S2− Solid-State Ag2S Br−: −25,I−: −18,Cl−: −30

HSO3− Polymer Guanidinium derivative ionophore, PVC,

ortho-nitrophenyloctyletherClO4

−: −2.2,Cl−: <−3.0,Sal−: −2.3

SCN− Polymer chloro[5,10,15,20-tetrakis[4-(hexyloxy-carbonyl)phenyl] porphyrinato]manganese(III), PVC, ETH 469

ClO4−: −2.0,

I−: −2.3,NO3

−: −3.6,NO2

−: −3.0,Cl−: −3.4,HCO3

−: −5.1CO3

2− Polymer N, N-dioctyl-4-trifluoroacetylbenzamide,MTDDACl, PVC, bis-2-ethylhexylsebacate

SCN−: +1.0,NO3

−: −1.6,Br−: −3.5,Cl−: −5.0,Sal−: +3.3

NO2− Polymer 2,9,16,23 tetra-tert-butylphthalocyanine)-

cobalt(III), hexadecyltrioctylammoniumiodide, PVC, dibutylphthalate

SCN−: −1.0,I−: −1.6,NO3

−: −3.1,Cl−: −3.5,Br−: −2.9

Phosphate Polymer 3-decyl-1,5,8-triazacyclodecane-2,4-dione,PVC, Dibutylsebacate

SCN−: −2.3,NO3

−: −2.8,Cl−: −2.3,OAc−: −3.2

(continued overleaf)

9.3 Glass Electrodes 289

Tab. 2 (continued)


Salicylate Polymer (2,9,16,23-tetra-tert-butylphthalocyanine)tin(IV), PVC, dinonyl sebacate

ClO4−: −3.3,

SCN−: −2.9,Cl−: −4.8,OAc−: −3.4

asee [15, 16] and references cited therein. Abbreviations: KTpClPB, potassiumtetrakis(4-chlorophenyl)borate; PVC, poly(vinyl chloride); NaTFPB, sodiumtetrakis[3,5-bis(trifluoromethyl)phenyl]borate; MTDDACl, methyl tridodecyl ammonium chloride.

mechanism of the pH glass electrode.Some researchers favor viewing it as anadsorption process, rather than simple ion-exchange, responsible for the observed pHresponse [17].

A variety of optimized glass composi-tions exist that are suited for a varietyof applications. The classical sodium-containing glass has been largely replacedby lithium glass. High content of Li2O fa-vors a low membrane resistance and low al-kali error; that is, a larger measuring range.Today, glass electrodes that are small, haveextremely large measuring ranges of pH 0to 14, and show low resistances even withthe relatively thick membranes requiredto achieve acceptable robustness can bemanufactured. In the early years of pHglass electrode development, researcherstended to blow their own glass electrodesakin to Christmas tree bulbs. These elec-trodes were extremely fragile and had tobe handled with great care. Today, glassmembrane thicknesses can exceed 1 mmand can therefore be sufficiently strongto break a laboratory beaker without dam-aging the glass membrane itself. Glasseswith extremely high Li2O content, how-ever, tend to crystallize easily and cantherefore not be handled by a glass blower.Moreover, they have shorter lifetimes, cor-rode more easily, and are not suited for

high temperature applications (includingthe sterilization cycles for certain biologicalmeasurements; for example, monitoringthe pH of cell cultures, etc.). Therefore,a variety of more rugged and extremelystable high temperature glasses are com-mercially available. Although they exhibitemf response over a reduced pH range,they function reliably at high tempera-ture without the need for intermittentcalibration.

Since pH glass membrane electrodesare used in a variety of applications,manufacturers of pH electrodes have de-voted much effort in designing combina-tion pH electrodes in many shapes andsizes, for use in NMR test tubes, with flatsurfaces for paper and cheese pH measure-ments, pressure-resistant electrodes forreactor applications, and a large variety oflaboratory pH electrodes. Beyond the type,size, and shape of the pH sensitive glass,an essential component of the combina-tion pH electrode is the external referenceelectrode. Depending on the applicationof the pH electrode, one can choose fromrefillable electrolytes in single- and double-junction designs, as well as low viscositymaintenance-free solid polymers and gelelectrolytes that can sustain high externalpressures. In addition, available junctionmaterials range from simple ceramic frits,


flat, circular ceramic frits for surface pHapplications, Teflon sleeve junctions, freediffusion liquid junctions, and hole junc-tions for polymeric electrolytes. Somereference electrolytes can be pressurized inspecial chambers, while others are factoryprepressurized or sustain high pressureswithout additional treatment. The excel-lent reliability, lifetime, and analytical per-formance of pH glass electrodes/referencecombinations set an extremely high stan-dard that few, if any other ISEs, can match.

Some specialized glass formulationshave been found to be sensitive to ionsother than H+ as well. Glass membraneswith a high content of Na+, for example,are known to be more Na+ responsive [14].While they are still selective for H+,they can be used to assess Na+ activitiesin many biological samples because atphysiological pH, the activity of H+islow compared to Na+ levels (140 mM inblood), and thus does not contribute tothe measured Ecell value. Indeed, suchglass electrodes are still used today in anumber of clinical blood analyzers, despitethe availability of much more selectiveionophore-based sodium electrodes (seefollowing text). A different class of glasselectrodes, chalcogenide glasses, has beenfound to respond to a variety of heavy metalions, including lead and cadmium [18].

The main limitation of glass membraneelectrodes in biological applications is theirtendency to be fouled by strongly adsorb-ing proteins, and by the limited numberof ions that can be sensed potentiometri-cally via glass membranes. Indeed, glassmembranes cannot be utilized to senseanions, and thus do not provide a genericapproach for ion sensing. While the pro-tein adsorption problems can be reducedwith repetitive washing cycles, difficultiesin manufacturing miniaturized versions ofglass electrodes to be used in conjunction

with other types of ISEs in planar sensorarrays, and so on have motivated signif-icant research on different pH selectivematerials. Most notable are pH electrodesbased on polymeric membranes dopedwith amine functional ionophores [19],and solid-state iridium oxide membranematerials [20].

9.4Solid-state Electrodes

ISEs based on solid-state membrane mate-rials use sparingly soluble inorganic saltsas the membrane matrix. Interesting ex-amples of this family of electrodes includethe silver precipitate-based membranes.Pressed pellets of silver sulfide (Ag2S)are known to respond to silver ions be-cause the membrane is a conductor forsilver cations [21–23]. Interestingly, thismembrane is also sensitive to sulfide ionsbecause any change in the sample sulfideactivity affects the available silver activityat the surface of the electrode membraneaccording to the following dissolutionequilibrium:

Ag2S(s) = 2Ag+(aq) + S2−(aq)

This electrode shows a very high selectivityto both silver and sulfide (both ions cannotbe present in solution in large quantitiesbecause of the small solubility product ofAg2S). The only substantial interferent isthe mercury(II) ion owing to the extremelylow Ksp for HgS. The pressed pelletrequires periodic polishing to removesurface adsorbates and other precipitates,but shows otherwise very long lifetimes.This principle has been extended toother silver salts, especially silver halides.Silver sulfide pellets doped with CuS willrespond to Cu2+ because of the largerdissolution equilibrium of CuS relative toAg2S. A silver sulfide pellet doped with

9.5 Liquid Membrane Electrodes 291

AgBr will be bromide selective, and soon. The selectivity observed with suchsilver halide precipitates follows exactly thesolubility product of the respective silversalt. Consequently, the following sequenceis always observed for any given silverprecipitate membrane:

S2− > I− > Br− > Cl−

A silver chloride precipitate membrane inprolonged contact with an I−-containingsolution will therefore eventually becomean iodide-sensitive electrode since all sur-face bound chloride will exchange with io-dide. However, this process can take sometime in dilute solutions, during which theelectrode remains essentially responsiveto chloride [24]. Unfortunately, silver pre-cipitate membranes are often unsuitablefor use in biological samples as (1) thiol-containing molecules (proteins containingcysteines) may foul the surface of theion-selective membrane and (2) the AgClmay dissolve because silver ions are com-plexed by protein amine functions [25].Chloride measurements in such samplesare therefore normally not performedwith AgCl-based membranes. Analyticalproperties similar to pressed pellet mem-branes have also been observed with silversalts embedded in a silicone membrane,with ionophore-based silver-selective liq-uid membranes, and with silver–silverhalide wires. The latter class is usuallyless preferred for practical measurementssince any exposed metal may induce someredox species cross-sensitivity of the elec-trode, which does not usually occur withion-conducting membrane-based indica-tor electrodes.

Another solid-state membrane of ex-tremely high selectivity and applicabilityis the single crystal LaF3 membrane elec-trode [26]. The crystal is doped with eu-ropium to lower its electrical resistance

and it acts as an effective pure F−conductor. When used as a fluoride ion-selective electrode material, a large mea-suring range of about 6 orders of mag-nitude is observed (10−6 M to 1 M F−),with Nernstian response slopes. The onlysignificant interferent is hydroxide (due tothe low solubility of La(OH)3). Such elec-trodes are typically used under strict pHcontrol between 5 and 6 to avoid hydroxideinterference at higher pH and the for-mation of HF at lower pH, which woulddecrease the free activity of F−. Suitableionic strength adjustment buffers for flu-oride measurements typically also containcomplexing agents such as citrate to re-move cations such as aluminum and ironfrom the sample that have a tendency tocomplex fluoride. The LaF3 electrode hasan extremely long lifetime under normaluse. Its main disadvantage is its cost dueto the necessity of using a polished singlecrystal.

9.5Liquid Membrane Electrodes

Organic liquid membrane ISEs are cer-tainly the most versatile group of ISEs, andhave been especially successful for mea-surements in biological samples. Theirresponse is dependent on liquid–liquidpartitioning principles. The ion-selectivemembrane is either a highly plasticized hy-drophobic polymer, a liquid organic phase,or a plasticizer-free polymer with a lowglass transition temperature. With the firstsystem, the plasticizer to polymer ratio isoften large (2 : 1 by weight), and the mainfunction of the polymer was originally toprovide the membrane with required me-chanical stability. The choice and concen-tration of plasticizer dictates to a large partthe polarity of the membrane, membrane


selectivity, and diffusion coefficients of ac-tive components in the membrane. Forbiomedical applications, it is often moredesirable to use systems without any leach-able components, and plasticizer-free poly-mers are becoming increasingly available.

A variety of components are either freelydissolved in this hydrophobic matrix orcovalently anchored onto the polymericbackbone of the membrane. These mem-brane components mediate the selectiveextraction of many analytes and also makesure that the ISE membrane exhibitsion-exchanger properties. Thus far, liq-uid/polymer membrane ISEs for morethan five-dozen analytes have been de-scribed [15, 27, 28]. They are routinelyused in clinical analysis for the directpotentiometric detection of many an-ions and cations, and their applicationis steadily broadening with the adventof more selective membrane materials,advances in miniaturization, and the avail-ability of more rugged sensors. Two mainclasses of liquid membrane ISEs canbe distinguished: one that contains anion-exchanger without molecular receptorproperties, and the other that is based onhighly selective ionophores. While modernchemical research is mainly directed to theimprovement of the second class, manycommercial ISEs are still based on the first.

9.5.1Ion-exchangers

Liquid or solvent polymeric membranesmust exhibit ion-exchanger properties toproperly function. It was mentioned pre-viously that the concentration of analyteion in the membrane phase must re-main approximately invariant in the courseof an experiment. In a classical two-phase partitioning experiment, however,a concentration increase of an electrolyte

in the sample would lead to a pro-portional increase in the organic phase.Therefore, according to Eq. (6) such aprocess would lead to no change inthe phase boundary potential since theratio aI(aq)/aI(mem) is constant. Conse-quently, liquid membranes are routinelydoped with a salt of a lipophilic ion,for example, tridodecylmethylammoniumchloride for anion-selective electrodes andpotassium tetrakis(4-chlorophenyl)boratefor cation-selective electrodes. Numerousother ion-exchangers have been reportedin the literature, but their main function isalways the same. Prior to use, the liq-uid membrane is allowed to conditionin a solution that contains a high con-centration of the cation or anion to bemeasured. During this conditioning pe-riod, the hydrophilic counterion of theion-exchanger in the membrane is re-placed with the analyte ion of interest. Theselectivity of such membranes is a directfunction of the free energy of hydrationof the measured ions. For anion-selectivemembranes (containing the anion salt of aquaternary ammonium ion, for example),the observed selectivity sequence alwaysfollows the so-called Hofmeister sequence:

R− > ClO4− > I− > NO3

−

> Br− > Cl− > F−

where R− denotes an organic anion. Bycomplete analogy, the selectivity sequencefor ion-selective membranes incorporatinga cation-exchanger is:

R+ > Cs+ > Rb+ > K+ > Na+ > Li+

Consequently, ion-exchanger-based ISEshave historically been used for the detec-tion of perchlorate and nitrate, as well asa host of lipophilic organic ions includingcertain drugs and cationic/anionic surfac-tants. Indeed, the use of such electrodes for


rapid quantitation of charged organic drugspecies is the subject of a monograph andseveral reviews [29, 30]. Given the limitedselectivity over other lipophilic ions (bothorganic and inorganic), such drug sen-sors are best used for monitoring the drugcontent of pharmaceutical formulations,where the presence of such interferentspecies are not typically present.

Another more recent application of ion-exchanger-based membrane electrodes re-lates to their utility for sensing polyionicspecies, including the anticoagulant hep-arin, and its antidote polycationic pro-tamine [31]. With the appropriate choiceof the organic ion-exchanger in the mem-brane (tridodecylmethylammonium chlo-ride in the case of polyanion sensing, anddinonylnapthalene sulfonate in the case ofpolycation sensing) 0.1–10 µg mL−1 lev-els of polyions can be sensed in biologicalsamples. These new sensors, however,function via a nontraditional nonequilib-rium response mechanism, based on thecooperative extraction of the polyion intothe membrane to form strong ion-pairswith the lipophilic exchanger. Indeed, nouseful potentiometric response is observedif the membrane is equilibrated withthe given polyion for extended periodsto completely convert the ion-exchangerto the polyion complex form. However,if operated always in a nonequilibriummode, these ion-exchanger-based sensorscan provide a simple means to measurecertain high charge density polyions insamples as complex as whole blood.

Interestingly, anion-exchanger-basedmembrane electrodes are also used rou-tinely to assess chloride in extracellularfluids. While there are some interferencesfrom thiocyanate (for patients who smokecigarettes) and bicarbonate, as well asfrom some large anionic molecules suchas heparin, calibration in samples with

very similar background electrolytes canminimize these effects. In recent years,however, ionophore-based chloride sen-sors that offer potential advantages over theion-exchanger-based system, most notablywith respect to selectivity, have becomeavailable [15].

It is interesting to note that the listedselectivity sequences are thermodynamicand are not always observed under prac-tical measuring conditions. If a stronglyinterfering ion is present but relativelydilute (typically at less than 10−4 M lev-els), its effect on the cell potential is oftenmuch smaller than predicted on the basisof the Nicolsky equation and its modifi-cations [32]. Such low concentrations leadto a local depletion of these ions at themembrane surface and to an incompleteion exchange, even if thermodynamicallyfavored. For short exposure times, there-fore, the electrode can often still reliablybe used to assess the analyte. After pro-longed contact with a sample containinga strong interferent, the electrode does,however, completely recondition and failsto respond to the primary analyte. Theseeffects have historically been exploited forsensing applications in samples for whichno thermodynamic selectivity was avail-able. It requires careful reconditioningbetween measurements, and is not rec-ommended if intrinsically better selectivitycan be achieved with another ISE. This ef-fect is perhaps an explanation as to whysome manufacturers offer ion-exchanger-based ISEs for a variety of ions, althoughthe basic membrane compositions are es-sentially identical.

9.5.2Neutral Ionophores

Lipophilic, electrically neutral ionophores,also called ion carriers because of their


capability of selectively transporting ionsacross artificial membranes, are chem-ical components that are essential toachieving high-sensing selectivity with liq-uid or polymer membrane–based ISEs.Neutral carrier-based membranes also re-quire the addition of a lipophilic ion-exchanger for proper functioning. Theion-exchanger forms the counterion ofthe complexed analyte in the mem-brane, and is typically less concentratedthan the ionophore. For example, cation-selective membranes may contain theionophore and the lipophilic tetraphenylb-orate derivative cation-exchanger potas-sium tetrakis(4-chlorophenyl)borate, whileanion-selective membranes may be dopedwith tridodecylmethylammonium chlo-ride as anion-exchanger in addition tothe ionophore. While ionophore-free ion-exchanger-based membranes always showthe same selectivity pattern that follows thehydration energies of the ions, ionophore-based membranes show selectivities thatare significantly different. This stems fromthe formation of strong complexes be-tween the extracted analyte ion and theionophore in the membrane. Complex for-mation constants can vary widely, and havebeen reported to be around 108 M−1 formonovalent to about 1020 M−2 for divalentions. Sensor selectivity is now dictated bythe free energy of transfer of the ions fromthe aqueous to the membrane phase, thecomplex formation constants between theextracted ions and the ionophore, and theconcentrations of ionophore and lipophilicion-exchanger (also called ionic site) inthe membrane. Because hydration ener-gies are still important, given that an ionextraction process is involved, it is typicallymuch more difficult to design ISEs forhydrophilic ions than it is for hydropho-bic ones. On the other hand, it is often achallenge to design selective receptors for

extremely large, bulky ions. Consequently,ionophore-based ISEs for potassium wererealized quite early on, while truly selectivesensors for magnesium, lithium, sodium,and small anions such as chloride andphosphate have only been developed fairlyrecently and are still topics of current re-search. Ionophore-based ISEs for bulkyanions such as perchlorate are virtuallyunknown.

Figures 4 (see earlier) and 5 illustratethe structures of a small selection ofsuccessful ionophores used to prepare liq-uid/polymeric membrane ISEs. Table 2summarizes the typical membrane com-positions of the corresponding ISEs andobserved selectivity coefficients. Observedselectivities are in many cases extremelyhigh, and contribute to the great successthat such sensors have enjoyed for practi-cal measurement applications in biologicalsamples.

Ionophores can be developed onthe basis of a variety of recogni-tion principles. They include simplecrown ethers, bis-crown ionophores,crowns with bulky side groups to pro-hibit intermolecular sandwich forma-tion, noncyclic amide and thioamideionophores, basket-shaped calix[4] areneand calix[6] arene ionophores, calixareneswith crown bridges, thiocarbamates,lipophilic amine bases as H+-ionophores,guanidinium derivatives, multitopic hy-drogen bond–forming ionophores forselective anion recognition, metallopor-phyrins, corrins, and phthalocyanines withdifferent metal centers and a variety ofaxial ligands, and aromatic trifluoroacetylderivatives for the recognition of hy-drophilic nucleophiles. It should be notedthat some of these ionophores are electri-cally charged in their uncomplexed form,and this special class of ion carriers isdiscussed separately later.


N

N

N

NN

H H

NH NH

OO

NH

+

Sulfite-1

Phosphate-1

N

N

N

M

N

N

NN N

Cl

Cl

Salicylate-1 (M = Sn(IV))

O

N NH

NH

SH

S

NH

CF3

O

O

N

N

N

NCo(III)

N

N

NN N

N

COOC6H13 COOC6H13

N N

N

COOC6H13 COOC6H13

MCl

Cl−- 1 CO32−- 1

SCN−- 1 (M = Mn(III))NO2−- 1

Fig. 5 Structures of typical anion-selective ionophores used in ion-selective electrode membranes.

To function properly in ion-selectivemembranes, ionophores typically share anumber of important characteristics [33].They should be highly lipophilic so thatthey are strongly retained within the

hydrophobic membrane phase to en-sure a long lifetime of the sensor [34].This is most often achieved by addinglong alkyl chains, cyclohexyl or adamantylsubstituents to the molecular backbone.


They should have a polar moiety or a set ofpolar functional groups that is responsiblefor the ion recognition process. The rest ofthe ionophore molecules should containhydrophobic regions that are compatiblewith the surrounding membrane matrix.The historical argument that an ionophoremolecule must also exhibit a certain mobil-ity within the membrane has been largelydisproved by the comparable analyticalperformance of a number of membranematerials where the ionophore is cova-lently anchored onto the polymeric back-bone [35]. It seems beneficial, however, toat least either have mobile ionophores ormobile ionic sites to guarantee an accept-ably low membrane resistance.

Membrane selectivities for any givenionophore can vary substantially. Mem-branes of relatively high polarity aretypically preferred for the development ofdivalent ISEs and many anion-selectiveelectrodes, while nonpolar membranesare generally more suited for monovalentcations. However, the role of membranepolarity on the membrane selectivity hasbeen overrated in the past as there seemsto be no direct correlation when a large

number of plasticizers are compared [36].Indeed, many other parameters appear toinfluence selectivity equally as well, in-cluding the tendency to form ion pairs,the availability of functional groups onthe plasticizer that can compete with theionophore, and indirect variations of com-plex stoichiometries of the ionophore indifferent solvent environments. Therefore,optimization of ISE selectivity is to a largedegree still an empirical process. However,if the complex stoichiometries are known,optimum membrane concentration ratiosbetween ionophore and lipophilic ionicsites can be effectively predicted. A ther-modynamic analysis of the correlation be-tween expected selectivity coefficient andmembrane composition reveals that selec-tivity maxima do indeed exist [37]. Table 3shows the optimum concentration ratiosfor a select number of assumed complexstoichiometries and charges of the twocompared ions. Since some ionophoresare capable of forming mixed complexes,it is advisable to use these values asfirst guesses only and to evaluate theselectivity for a wide range of ionic siteconcentrations.

Tab. 3 Optimum concentration ratio of lipophilic anionic sites to neutralionophore, giving highest selectivity for the analyte ion over an interfering ion.The optimum ratios are dependent on the charges of both ions and therespective complex stoichiometries (number of ligands per ion) in themembrane

Charge of cation Stoichiometry of ligand–ion complex

I (analyte) J (interfering I (analyte) J (interfering Optimum ratio ofzi ion) zj ni ion) nj sites to ionophore

2 2 1 2 1.412 2 2 3 0.772 2 3 4 0.542 1 1 1 1.622 1 2 2 0.732 1 3 3 0.461 1 1 2 0.71


Microelectrode forms of neutral carrier-based membrane electrodes with tip diam-eters of 1 µm or less have been used byphysiologists for many years to probe thefree concentration of potassium, calcium,sodium, and proton activity within singlecells [38]. Such electrodes can be preparedconveniently by incorporating the appro-priate membrane components (includingrequired lipophilic sites) either in a wetorganic liquid or polymer form, withinthe distal end of pulled glass pipet tipsin which the inner walls at the distalend have been treated with an organosi-lane reagent to render them hydrophobic.The inner filling solution and Ag/AgClwire complete the microelectrode assem-bly. For intracellular measurements, anexternal reference electrode can be placedoutside the cell, provided that correctionfor the cell membrane potential is made viaseparate experiments with two Ag/AgClelectrode–based pipets filled with KCl, oneinside and the other outside a separate cellbathed in the same solutions. Alternately,dual-barrel micropipets can be prepared, inwhich one barrel contains the ISE mem-brane and the other is filled with KCl (anda Ag/AgCl reference wire) to serve as theexternal reference within the biological cellbeing probed.

9.5.3Charged Ionophores

Most ionophores used in liquid/polymermembrane ISEs today are electricallyneutral in their uncomplexed form andassume the charge of the analyte ion whencomplexed. In the membrane, therefore,the following complexation equilibriumexists between ionophore L and theanalyte I+:

L (org) + I+(org) = LI+(org)

Consequently, ISE membranes contain-ing such ionophores must also containa lipophilic ion-exchanger whose chargeis opposite that of the analyte ion, whichdictates the concentration of LI+ complexin the membrane through the electroneu-trality condition. However, Fig. 5 alsoshows the structures of a small selec-tion of ionophores that are electricallycharged in their uncomplexed form (seethe metalloporphyrin and phthalocyaninestructures in Fig. 5). With the exception ofthe successful dialkylphosphate calciumcarriers [39], such ionophores have beensomewhat neglected in the field sincethere has been evidence that the bind-ing selectivity of the ionophore cannot befully translated into a high-sensing selec-tivity for a corresponding ISE preparedwith these compounds. In recent years,however, this view has been modifiedand electrically charged ionophores arebeing evaluated more aggressively, espe-cially in the design of receptors for anionicspecies. An additional electrical charge onthe coordinating functional group of theionophore seems especially attractive tofurther strengthen otherwise weak com-plexes with additional coulombic forces,which can be particularly useful in thecase of anion recognition.

Functional ISEs based on charged car-riers can be fabricated with membranesthat contain just the salt of a chargedionophore, since the ionophore has bothionophoric and ion-exchanger properties.However, it has been shown that the cor-responding sensing selectivities are thenoften less than ideal [40]. Consider, forexample, a membrane with a chargedionophore selective for a monovalent an-ion. The concentration of uncomplexedionophore in the membrane is ordinar-ily small and dictated by the dissociationconstant of the complex:


LA(org) = L+(org)+A−(org)

If the ISE membrane were to be condi-tioned in a solution of an interferent whosecomplex with the ionophore is weaker, theconcentration of uncomplexed ionophorein the membrane would be larger sincethe dissociation constant is now larger aswell. In a mixed sample situation whereboth ions are present, an intermediatesituation would be observed, that is, ahigher interfering ion level would lead toa higher concentration of uncomplexedionophore in the membrane. This depen-dency between sample composition anduncomplexed ionophore concentration isnonideal, since it effectively favors theformation of the complex with a weakerbinding analyte. Optimum selectivitiescan be achieved by incorporating an ion-exchanger into the membrane that has thesame charge as the analyte, and whichforms the counterion of the uncomplexedionophore L+. That concentration mustnow be invariant of the type of sampleion extracted, and theory predicts the ISEselectivity to be dependent on the bindingselectivity of the charged ionophore in thevery same way as for membranes contain-ing neutral ionophores [40]. In effect, thepresence of the lipophilic additive servesto maintain a fixed level of uncomplexedcharged carrier L+, thereby buffering the

membrane phase with respect to the freeprimary ion activity in the membrane,which further serves to enhance selectivity.

Interestingly, suitable extensions of thistheory have allowed researchers to ex-plain other peculiar effects involving somecharged carrier-based ISEs, for exam-ple, apparently super-Nernstian responseslopes that are sometimes observed [41].While the details of this effect are be-yond the scope of this chapter, it shouldbe emphasized that such behavior can beexplained with thermodynamic processesoccurring at the sample–membrane inter-face. Undoubtedly, charged carrier–basedISEs, especially for anions, will be a majorresearch direction in years to come.

9.6Gas-sensing Probes

Several of the ion-selective membrane elec-trodes described in the previous sectioncan also be employed as transducers todevise highly selective potentiometric gas-sensing probes. As shown in Table 4,equilibrations of gases in aqueous solu-tion yield ionic species. Hence, a numberof relatively simple and analytically usefulprobes can be constructed by incorpo-rating ISE-based electrochemical cells asdetectors behind outer gas permeablemembranes, and choosing appropriate

Tab. 4 Some potential analyte gases and their relevant solution phase equilibria thatcan be employed to devise potentiometric gas-sensing probes based on innerion-selective membrane electrodes

Analyte gas Equilibrium reactions Inner ion electrode

CO2 CO2 + H2O ←−→ H+ + HCO3− H+, CO3

2−SO2 SO2 + H2O ←−→ H+ + HSO3

− H+, HSO3−, SO3

2−NO2 NO2 + H2O ←−→ NO3

− + NO2− + H+ H+, NO2

−, NO3−

HOAc (acetic acid) HOAc ←−→ H+ + OAc− H+, OAc−NH3 NH3 + H2O ←−→ NH4

+ + OH− H+, NH4+

9.6 Gas-sensing Probes 299

inner electrolyte solutions into which theanalyte gas can equilibrate [42].

In 1958, Severinghaus and Bradley orig-inally proposed the classical design forsuch devices [43] for the measurementof CO2 in blood and this is shownin Fig. 6. The inner ion-selective mem-brane in this configuration is usually apH-sensitive glass membrane, althoughother pH-sensitive ISEs including neu-tral carrier-based polymeric membraneelectrodes can also be employed. Thegas-sensing probe is designed in such a

way that the inner membrane electrodetransducer presses tightly against an outergas permeable membrane barrier, typicallysilicone rubber or microporous Teflon ma-terials, creating a very thin layer of innerelectrolyte between the inner ion-sensingmembrane and outer gas-permeable film.Note that the external reference electrodefor the ISE is also placed behind the gas-permeable membrane in the bulk of theinner filling solution; hence the entire elec-trochemical cell is confined to this fillingsolution, and thus the complete device

Ecell

Gas-permeablemembrane


Internal fillingsolution

Referenceelectrodes

Internalsolution ofISE

Gas + H2O Ions

Gas

Fig. 6 Schematic of typical potentiometric gas-sensing probe. Inner electrode is aglass membrane pH sensor for a conventional Severinghaus style device. Other ISEscan be used in a similar geometric arrangement to detect ionic forms of gases withinthe thin layer of the inner filling electrolyte solution at the distal end of the probe.


should be termed a gas-sensing ‘‘probe’’ or‘‘sensor,’’ not a gas-sensing ‘‘electrode.’’

For the conventional Severinghaus stylesensor, equilibration of analyte gas fromthe sample phase through the outergas-permeable membrane determines thepH of the inner thin film of electrolyteat the distal end of the probe. For a CO2sensor, as an example, equilibration of theanalyte in aqueous solution yields protonsand bicarbonate ions (Table 4). Using apH sensor as the inner transducer, theelectrochemical potential of the cell can bewritten as:

Ecell = K + 0.059 log aH+(film) (14)

where aH+ is the activity of hydrogen ions

in the thin film of inner electrolyte, and K

is the sum of all additional constant emfvalues of the cell, including the voltage ofthe inner reference electrode of the pH sen-sor and the external reference electrode.Note that there are no junction potentialsconsidered within this K term, since theinner reference electrode, typically a bareAg/AgCl contact, is in direct ionic contactwith the same inner electrolyte solution asthe ion-sensing membrane. The equilib-rium constant for CO2 equilibration canbe written as:

KCO2 = aH+ × aHCO3

PCO2

(15)

where PCO2 is the partial pressure of CO2.When the PCO2 is equal on both sides of thegas-permeable membrane, the thin film ofelectrolyte is in complete equilibrium withthe sample phase (e.g. blood) and the aH+within the thin film is directly proportionalto the PCO2 in the sample:

aH+(film) = KCO2PCO2

aHCO3(film)(16)

Substitution of aH+ (film) from Eq. (16)into Eq. (14) yields an expression for the

voltage of the gas sensor as function of thePCO2 in the sample phase:

Ecell = K + 0.059 logKCO2PCO2

aHCO3(film)(17)

If the activity of bicarbonate is keptconstant within the film, which can beachieved by using a fairly concentratedNaHCO3 solution (0.1 M) as the innerfilling solution (with added KCl or NaClto poise the potential of the Ag/AgClreference electrode in contact with thebulk of this solution, see Fig. 6), and giventhat KCO2 is a constant, Eq. (16) can berewritten as:

Ecell = K ′ + 0.059 log PCO2 (18)

Thus, the gas-sensing probe responds likeany ISE with a Nernstian slope toward theanalyte gas.

By analogy [42], sensors for any of theother acidic analyte gases listed in Table 4can be prepared by using a solution ofthe conjugate base anion as the innerfilling solution (with added chloride salt).In each case, increasing the level of gasin the sample phase lowers the pH ofthe thin electrolyte film, which is senseddirectly by the pH glass or other pH-sensitive membrane electrode. A sensorfor ammonia gas can be prepared by usingan NH4Cl solution as the inner electrolyte.Here, the conjugate acid, NH4

+, is heldconstant, and the equation for the outputvoltage of the probe is:

Ecell = K ′ − 0.059 log PNH3 (19)

It should be noted that the response ofthese gas-sensing probes is also directlyproportional to the dissolved concentrationof the analyte gas in accordance withHenry’s law: Pgas = H × [gas], where His Henry’s constant for the given gas at agiven temperature.

9.6 Gas-sensing Probes 301

The response time of these gas-sensingprobes is generally on the order of 3 to5 min at very low dissolved concentrationsof the analyte gas (<10−4 M), and approx.1 min at higher levels (>10−3 M). Further,selectivity over ions in the sample phaseis excellent because such ions cannot passthrough the outer gas-permeable mem-brane and influence the output voltageof the inner potentiometric measurementcell. One could therefore measure low lev-els of ammonium ions in the presenceof very high levels of potassium and/orsodium by adjusting the pH of the samplephase to a value greater than 11, whereall ammonium will be in the form offree ammonia gas, and then determiningthis dissolved gas level with the ammoniagas-sensing probe. However, selectivity ofthese sensors is by no means absolute.Other volatile gases or neutral organicacids or bases that can permeate the outer‘‘gas’’-permeable membrane can yield sig-nificant interference problems [44]. Forexample, using the conventional Sever-inghaus design, measurement of CO2 inthe presence of SO2 would be nearly im-possible, since SO2 is a stronger acid,and its diffusion into the inner electrolytewould also decrease the pH of the filmof bicarbonate filling solution within theCO2 sensor. In the case of NH3 gas sen-sors, volatile amines, such as methylamineand ethylamine, pose similar interferenceproblems.

One can enhance the selectivity of poten-tiometric gas sensors by using a bufferedinner filling solution, and using an innerISE detector that responds with selec-tivity toward an ionic form of the gas.For example, instead of using an innerpH sensor, an alternate design for anammonia-sensing probe would make useof a nonactin-based (see compound NH4

+-1 in Fig. 4) polymer membrane type ISE as

the inner transducer in conjunction withhighly buffered (at pH 7.0 or so) inner fill-ing solution. Ammonia diffusion throughthe outer gas permeable membrane will re-sult in the formation of ammonium ionswithin the thin film that can be sensed bythe inner electrode. Although amines canalso diffuse into the film, the nonactin-based sensor exhibits high selectivity forammonium over these protonated amines,and hence little or no interference willbe observed [45]. The disadvantage of thisconcept, relative to the conventional Sever-inghaus design, is that the resulting devicehas only a very limited dynamic measure-ment range where Nernstian response isobserved. At increasing gas levels, the pHof the inner buffer solution will be altered,and the fraction of analyte gas in the ionicform at equilibrium will decrease, therebyyielding a significant nonlinear emf re-sponse as a function of increasing thegas level in the sample. Sensors of thistype have also been demonstrated for de-termination of SO2, CO2 and NO2 usinginner sulfite/bisulfite, carbonate, and ni-trate–nitrite selective polymer membranetype ISEs, respectively.

Another alternate potentiometric gas-sensing design, specifically developed tomeasure PCO2 in whole blood sampleswithin modern point-of-care biomedicalinstruments [46], relies on a novel differ-ential measurement approach, using twoidentical membrane electrodes. Figure 7illustrates this configuration. The two iden-tical polymeric membranes are doped witha neutral proton carrier (e.g. tridodecy-lamine; see compound H+-1 in Fig. 4),rendering these membranes pH selectivebut also gas permeable (owing to theirpolymeric nature). The electrodes onlydiffer with respect to the compositionof their inner reference electrolyte solu-tions, one being strongly buffered with


CO2

CO2 + H2O HCO3− + H+

PolymericpH-sensingmembranesBuffered

internal

Solidsubstrate

Bicarbonateinternal

Ag/AgClwires

Samplephase

Ecell

Fig. 7 Illustration of a differential sensor design for measuring carbon dioxide levels in biologicalfluids using two identical planar polymer membrane pH sensors with different internal solutions.

respect to pH, and the other composedof the typical NaHCO3/NaCl electrolyteemployed as the inner filling solutionwithin the conventional Severinghaus de-sign. Measuring the cell voltage of theright electrode (without buffer) versus theleft electrode (with buffer) will be insen-sitive to the pH of the sample, since theouter surfaces of both membranes willrespond equally but in opposite direc-tions to the proton activity in the samplephase (i.e. pH responses on the sam-ple side of the membranes cancel eachother). Carbon dioxide diffusion into theinner reference electrolyte of the left elec-trode does not change the pH of thebuffer, but the equilibration of CO2 withinthe right electrode alters the pH of theinner solution, yielding a change in themembrane potential. The net result isthat such a differential design will yield

emf response to only PCO2 in the sam-ple via exactly the same expression asEq. (18), except the voltage will becomemore negative rather than positive, withincreasing CO2 levels. In practice, as de-picted in Fig. 7, these differential CO2

sensors are constructed as thin layer filmson inexpensive planar substrates (plastic,ceramic), creating essentially disposablechip type devices.

9.7Ion-selective Electrode-based Biosensors

Beyond gas sensors, ISEs can also be uti-lized to devise novel biosensors, capableof detecting many important charged andneutral molecules in complex biologicalsamples. The most widely studied anduseful design for such biosensors is the

9.7 Ion-selective Electrode-based Biosensors 303

enzyme electrode [47], which couples thebiological specificity of an enzyme cat-alyzed reaction with the inherent selectivityof the ISE to create a single probe capa-ble of responding reversibly to biologicalmolecules. Selectivity of the final analyticalbiosensor is controlled by the additive se-lectivities of these two components of themeasurement system. Consequently, ionicinterferences are often avoided by employ-ing potentiometric gas-sensing probes asthe underlying transducer for fabricationof such enzyme-based biosensors.

To fabricate an enzyme electrode, a givenhighly purified enzyme is immobilizedeither physically (entrapment behindsemipermeable membrane) or chemically(e.g. crosslinked as a thin layer) at thesurface of an appropriate ISE or gassensor. For example, a sensor for ureacan be prepared by immobilizing theenzyme urease at the surface of an ammo-nium ion-selective membrane electrode,prepared by incorporating the antibioticnonactin (structure NH4

+-1 in Fig. 4) intoa PVC or other polymeric membrane.Figure 8 illustrates such a device for mea-suring urea in whole blood samples. Urea(an analyte measured in blood to deter-mine kidney function) diffuses from theblood sample into the thin layer of im-mobilized urease at the surface of theion-sensing membrane. In this thin biocat-alytic layer, the following reaction occurs:UREA + H2O → 2NH3 + CO2. At a phys-iological pH of 7.4, the vast majority ofammonia gas is in the form of protonatedammonium ions (NH4

+) and CO2 is inthe form of bicarbonate ions (HCO3

−).The ammonium ions generated within thislayer can be detected potentiometrically bythe polymeric membrane doped with non-actin. A steady state level of ammoniumions exists in this thin layer when therate of ammonium ion generation from

the enzymatic reaction is equal to the ratethat this product can diffuse away fromthe layer into the bulk of sample solution.An analogous urea sensor can be preparedusing an ammonium selective glass mem-brane electrode as the transducer ontowhich the urease is immobilized. Usingthin immobilized enzyme layers, responsetimes for such devices can be 1 min or lessat mM concentrations of urea.

Detection limits for enzyme-based po-tentiometric biosensors are governed bythe innate detection limits of the mem-brane electrode toward the product of theenzyme reaction, as well as the kinetics ofthe enzymatic reaction (Michaelis-Mentenconstant and turnover number of the en-zyme), and mass transfer rate of substrateinto the enzymatic layer. Carr and Bow-ers developed in detail the theory of howsuch parameters affect the response curvesof such sensors [48]. Generally, enzyme-based potentiometric biosensors respondto target substrates over the concentra-tion range of 0.01 mM to 10 mM. Inthe case of the urea example mentionedearlier, limited selectivity of the nonactin-based ammonium membrane electrodeover potassium ions (Kpot

NH4,K= 0.1) influ-

ences not only the detection limits but theaccuracy of urea measurements in samplescontaining significant levels of potassium.For measurement of blood urea levels,typically in the range of 3 to 10 mM, vari-ations in background levels of potassium(normally 2 to 6 mM) can influence theemf value, and therefore reliable measure-ments can only be made by measuringK+ levels separately with a K+-selectivemembrane electrode (usually based onvalinomycin; see K+-1 in Fig. 4) and cor-recting the output voltage of the biosensorfor these variations.

Another option for overcoming suchionic interferences is to immobilize the


mV

Referenceelectrode

Immobilizedurease

Internal electrolyte

Nonactin

Blood sample

Urea

NH4+

Urea

Urease

K+

K+

K+

Urea

Urea

rbc

rbc

rbc

rbc

rbc

Cross-linkedenzyme layer

Ammonium selectivepolymeric membrane

(nonactin)

+ + + +− − − −

2NH4 + CO2 +

Fig. 8 Design of potentiometric urea sensor based on the immobilized enzymeurease on the surface of a polymer membrane type ammonium ISE.

enzyme at the surface of gas-sensingprobes. For example, in the case of ureameasurements, the problem of potassiuminterference can be eliminated completelyby fabricating the biosensor with an am-monia gas sensor as the underlying trans-ducer. Ammonia generated from ureadiffusion into the enzyme layer, passesthrough the outer gas-permeable mem-brane of the gas sensor, increasing the pHof the thin film of inner NH4Cl filling solu-tion within such devices, which is detectedby the inner glass electrode (if traditional

Severinghaus gas-sensor design is uti-lized). It should be noted that such deviceswill still function at physiological pH, eventhough the vast majority of total ammoniagenerated enzymatically under such condi-tions will be in the ammonium ion form.However, the 1% that is gas is still ad-equate to provide a significant responseto the substrate in the 0.1- to 10-mMrange. Biosensors based on decarboxylat-ing enzymes that liberate carbon dioxideimmobilized on CO2 gas-sensing probescan be devised in analogous manner.


Tab. 5 Examples of some bioanalytically useful potentiometric enzyme-based biosensors that canbe prepared for monitoring important biomolecules using appropriate immobilized enzymes andISE or gas sensing detectors

Analyte Enzyme Reaction ISE detectors

Urea Urease Urea −−−→ 2NH3 + CO2 NH4+-glass or

polymerNH3-gas

sensorH+-glass or

polymerCreatinine Creatininase Creatinine −−−→

N-methylhydantoin + NH3

NH4+-glass or

polymerNH3-gas

sensorL or D-amino

acidsL or D amino

acid oxidaseL (D) AA −−−→

RCOCOOH + NH3 + H2O2

NH4+-glass or

polymerNH3-gas

sensorL-glutamine Glutaminase L-glutamine −−−→

glutamic acid + NH3

NH4+-glass or

polymerNH3-gas

sensorAdenosine Adenosine Adenosine −−−→ inosine + NH3 NH4

+ -glassor polymer

deaminase NH3-gassensor

L-glutamate Glutamatedecarboxylase

L-glutamate −−−→ GABA + CO2 CO2-gassensor

Amygdalin β-Glucosidase amygdalin −−−→HCN + 2C6H12O6+ benzaldehyde

CN−-solid-state

Glucose Glucoseoxidase

glucose + O2 −−−→gluconic acid + H2O2

H+-glass orpolymer

Penicillin Penicillinase Penicillin −−−→ penicilloic acid H+-glass orpolymer

Many potentiometric enzyme electrodesfor widely different analyte species can beprepared merely by choosing the appro-priate ISE transducer and immobilizedenzymatic reagent. Table 5 summarizesthe enzymes and ISEs/gas sensors used toconstruct biosensors for a number of im-portant biomolecules, ranging from ureaand creatinine to amino acids, nucleotides,and even glucose and penicillin. In the

latter two cases, the relevant enzymescatalyze a reaction that generates acidproducts, resulting in pH changes withinthe thin film of immobilized biocatalyst.Hence, simple glass or polymeric mem-brane pH electrodes can be employedas transducers for such devices. How-ever, detection of pH changes requiresthat the samples being analyzed pos-sess the same level of buffering capacity


as the standards used to calibrate suchsensors. Thus, practical application of thepH electrode–based biosensors normallymandates dilution of the sample witha defined buffer reagent before accuratemeasurements can be made.

In practice, potentiometric biosensorsare used just like any other membrane elec-trodes or gas-sensing probe. Calibrationplots are prepared using known standardsof the biomolecule analytes (usually pre-pared in a given buffer), and exposing thesensors to these solutions until a steadystate emf value can be recorded (usu-ally in 1 to 2 minutes). Plots of emfversus log [substrate] are prepared andused to obtain concentration data for un-known samples (either diluted in buffer ornondiluted) that are exposed to the samesensor under exactly the same measure-ment conditions.

References

1. D. Ammann Ion-Selective Microelectrodes,Springer, Berlin, 1986.

2. M. A. Pineros, J. E. Shaff, L. V. Kochian,Plant Physiol. 1998, 116, 1393.

3. B. R. Horrocks, M. V. Mirkin, D. T. Pierceet al., Anal. Chem. 1993, 65, 1213.

4. W. E. Morf, The Principles of Ion-Selective Elec-trodes and of Membrane Transport, Elsevier,New York, 1981.

5. E. Lindner, V. V. Cosofret, R. P. Buck et al.,Electroanalysis 1995, 7, 864.

6. D. P. Brezinski, Analyst 1983, 108, 425.7. R. E. Dohner, W. E. Morf, W. Simon et al.,

Anal. Chem. 1986, 58, 2585.8. U. Schefer, D. Ammann, E. Pretsch et al.,

Anal. Chem. 1986, 58, 2282.9. G. G. Guilbault, R. A. Durst, M. S. Frant

et al., Pure Appl. Chem. 1976, 48, 127.10. E. Bakker, R. K. Meruva, E. Pretsch et al.,

Anal. Chem. 1994, 66, 3021.11. W. E. Morf, W. Simon in Ion-Selective

Electrodes in Analytical Chemistry (Ed.: H.Freiser), Plenum Press, New York, 1978.

12. M. Nagele, E. Bakker, E. Pretsch, Anal. Chem.1999, 71, 1041.

13. M. Cremer, Z. Biol. (Munich) 1906, 47, 562.

14. G Eisenman, (Ed.), Glass Electrode for Hydro-gen and Other Cations, Marcel Dekker, NewYork, 1967.

15. P. Buhlmann, E. Pretsch, E. Bakker, Chem.Rev. 1998, 98, 1593.

16. M. E. Meyerhoff, Trends Anal. Chem. 1993,12, 257.

17. C. Huang, Y. C. Jean, K. L. Cheng, F. C.Chang, J. Electrochem. Soc. 1995, 142, L175.

18. A. E. Owen, J. Non-Cryst. Solids 1980, 35, 36,999.

19. U. Oesch, D. Ammann, Z. Brzozka et al.,Anal. Chem. 1986, 58, 2285.

20. S. A. M. Marzouk, S. Ufer, R. P. Buck et al.,Anal. Chem. 1998, 70, 5054.

21. E. Pungor, K. Toth, Analyst 1970, 95, 625.22. E. Pungor, K. Toth, Pure Appl. Chem. 1973,

34, 105.23. W. E. Morf, G. Kahr, W. Simon, Anal. Chem.

1974, 46, 1538.24. A. Hulanicki, T. Sokalski, A. Lewenstam,

Mikrochim. Acta 1988, 111, 119.25. M. L. Hitchman, A. Aziz, D. D. H. Chin-

gakule et al., Anal. Chim. Acta 1985, 171,141.

26. M. S. Frant, J. W. Ross, Science 1966, 154,1553.

27. E. Bakker, P. Buhlmann, E. Pretsch, Electro-analysis 1999, 11, 915.

28. M. S. Frant, Analyst 1994, 119, 2293.29. V. V. Cosofret, R. P. Buck, Pharmaceutical

Applications of Membrane Sensors, CRC Press,Boca Raton, 1992.

30. V. V. Cosofret, R. P. Buck, Crit. Rev. Anal.Chem. 1993, 24, 1.

31. M. E. Meyerhoff, B. Fu, E. Bakker et al.,Anal. Chem. 1996, 68, 168A.

32. M. Maj-Zurawska, T. Sokalski, A. Hulanicki,Talanta 1988, 35, 281.

33. E. Pretsch, M. Badertscher, M. Welti et al.,Pure Appl. Chem. 1988, 60, 567.

34. O. Dinten, U. E. Spichiger, N. Chaniotakiset al., Anal. Chem. 1991, 63, 596.

35. E. J. R. Sudholter, P. D. Vanderwal, M. Skow-ronskaptasinska et al., Anal. Chim. Acta1990, 230, 59.

36. R. Eugster, T. Rosatzin, B. Rusterholz et al.,Anal. Chim. Acta 1994, 289, 1.

37. R. Eugster, P. M. Gehrig, W. E. Morf et al.,Anal. Chem. 1991, 63, 2285.

38. D. Ammann, Ion-Selective Microelectrodes,Springer, Berlin, 1986.

39. J. W. Ross, Science 1967, 156, 1378.


40. U. Schaller, E. Bakker, U. E. Spichiger et al.,Anal. Chem. 1994, 66, 391.

41. S. Amemiya, P. Buhlmann, Y. Umezawa,Anal. Chem. 1998, 70, 445.

42. J. W. Ross, J. H. Riseman, J. A. Krueger, PureAppl. Chem. 1973, 35, 473.

43. W. Severinghaus, A. F. Bradley, J. Appl. Phys-iol. 1958, 13, 515.

44. R. K. Kobos, S. J. Parks, M. E. Meyerhoff,Anal. Chem. 1982, 54, 1976.

45. Y. M. Fraticelli, M. E. Meyerhoff, Anal. Chem.1981, 53, 1857.

46. M. E. Meyerhoff, Clin. Chem. 1990, 36, 1567.47. G. G. Guilbault, Analytical Uses of Immobi-

lized Enzymes, Marcel Dekker, New York,1984.

48. P. Carr, L. D. Bowers, Immobilized Enzymesin Analytical and Clinical Chemistry: Funda-mentals and Applications, Wiley & Sons, NewYork, 1980.

309

10Electrochemistry in Bioanalysis

Susan R. MikkelsenUniversity of Waterloo, Waterloo, Canada

10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31110.1.1 Bioanalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31110.1.2 Electroanalytical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311

10.2 Potentiometric Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31210.2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31210.2.2 Zero-Current Potentiometry . . . . . . . . . . . . . . . . . . . . . . . . . . . 31310.2.2.1 ISEs in Bioanalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31410.2.2.2 Enzymes, Antibodies, and Nucleic Acids with Zero-Current

Potentiometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31510.2.3 Dynamic Potentiometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31610.2.4 Emerging Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31710.2.5 Commercialization of Potentiometric Devices . . . . . . . . . . . . . . . . 318

10.3 Amperometric Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32010.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32010.3.2 Direct Amperometric Methods . . . . . . . . . . . . . . . . . . . . . . . . . . 32110.3.3 Amperometric Measurements with Enzymes . . . . . . . . . . . . . . . . 32310.3.4 Amperometric Immunoassays and Immunosensors . . . . . . . . . . . 32710.3.5 Amperometric Nucleic Acid Assays and Sensors . . . . . . . . . . . . . . 32910.3.6 Emerging Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33010.3.7 Commercialization of Amperometric Devices . . . . . . . . . . . . . . . . 330

10.4 Impedimetric Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33210.4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33210.4.2 Impedimetric Assays and Sensors . . . . . . . . . . . . . . . . . . . . . . . 33310.4.3 Commercially Available Instruments for Impedance Measurements . 334

310 10 Electrochemistry in Bioanalysis

10.5 Future Directions and Perspectives . . . . . . . . . . . . . . . . . . . . . . . 334Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335

311

10.1Introduction

10.1.1Bioanalysis

The American Association of Bioanalystsconsiders the field of bioanalysis to in-clude qualitative and quantitative analyt-ical methods in bacteriology (includingmycobacteriology), mycology, parasitol-ogy, virology, immunology, haematology(including flow cytometry), chemistry (in-cluding urinalysis), endocrinology, toxicol-ogy, andrology, embryology, and clinicalmolecular biology. As diverse as these ar-eas evidently are, they have at least onefactor in common: analytes are alwayspresent in a complex biological matrix.Thus, special demands are placed on theselectivity of a method (i.e. its ability todiscriminate analyte from possible inter-fering species), if the method is to be usefulin practice. The electroanalytical methodsdetailed below achieve the required selec-tivity through the measurement method(see also Volume 3), through electrode sur-face modification (see also Volume 10),or by the addition of selective chemi-cal or biochemical reagents to the assaymixture.

This chapter surveys applications ofelectroanalytical methods to practical

problems in bioanalysis and considersthe state of commercialization of tradi-tional and promising new methods andinstrumentation. Every effort has beenmade to cite recent review articles, so thatreaders may readily find more detailed in-formation. The subsections are organizedaccording to the instrumental method;hopefully, the diversity of applications willstimulate further innovations in method-ology and instrumentation for bioanalysis.

10.1.2Electroanalytical Methods

Three broad classifications of electrochem-ical methods are used in this chapter. Po-tentiometric methods include zero-currentpotentiometry and methods in which cur-rent of controlled magnitude is appliedto the working electrode, such as inpotentiometric stripping analysis (PSA).Amperometric methods consider all tech-niques in which current is measured;these include constant-potential amper-ometry and amperometric measurementsmade in response to a variety of ap-plied potential waveforms in voltammetricmethods. Impedimetric methods com-prise a final classification; in these meth-ods, faradaic currents are generally absent,and impedance, conductance, or capaci-tance is the measured property.


Descriptions of the underlying prin-ciples and practical examples of thesemethods may be found in many text-books. Excellent textbooks include thewidely cited Electrochemical Methods: Fun-damentals and Applications [1], and Labo-ratory Techniques in Electroanalytical Chem-istry [2]. The novice may find the AnalyticalChemistry by Open Learning series [3–5]and the primer Electroanalysis [6] par-ticularly useful. More specialized textsthat focus on bioanalytical electrochem-istry [7–11] and biosensors [12, 13] are alsoavailable.

Volume 3 of this series may be con-sulted for a survey of electrochemical in-strumentation and electroanalytical chem-istry. In addition, several chapters in thisvolume contain detailed information onmethods of importance to bioanalysis.In particular, Chapter 17 (mediated elec-tron transfer), Chapter 7 (electrochemistryof nitric oxide), Chapter 12 (electrochem-istry of nucleic acids), Chapter 13 (enzymeelectrodes), Chapter 14 (in vivo electro-chemistry), Chapter 5 (electrochemical im-munoassays), Chapter 2 (single cell elec-trochemistry), and Chapter 9 (ion-selectiveelectrodes) provide more details on thefundamental processes underlying the ap-plications to bioanalysis that are describedin this chapter.

10.2Potentiometric Methods

10.2.1Introduction

Potentiometry may be conducted atzero applied current or at controllednonzero current values; in both cases,electrochemical potential is measured andrelated to analyte concentration. Under

ideal conditions, a metallic indicator elec-trode in contact with a solution con-taining its cation will have a poten-tial described by the Nernst equation(Eq. 1):

Ecell = E − Eref

= E0 +(

RT

nF

)× ln[Mn+] − Eref , (1)

where Ecell is the measured cell potential,E is the indicator electrode potential,Eref is the potential of the referenceelectrode, E0 is the standard potentialand R, T , n, and F have their usualmeanings. If no current is applied tothe cell, and if no chemical reactioninvolving Mn+ occurs at the electrodesurface, [Mn+] represents the metal ionconcentration in the bulk of the solution.When the experiment generates a differentconcentration of metal ion at the surfaceof the working electrode, either by currentflow or by chemical reaction, [Mn+]represents the metal ion concentration(actually an activity) at the electrodesurface.

This general expression has been elab-orated to describe the response of ion-selective electrodes (ISEs) in which anactivity gradient exists for an ion, a, acrossa membrane that separates an externalfrom an internal reference electrode:

Ecell = constant +(

RT

zF

)

× ln(Aa,external/Aa,internal), (2)

where Aa,external and Aa,internal representthe activities of ion a in the external (sam-ple) and internal (reference) solutions,respectively, and constant includes the dif-ference in potentials of the two referenceelectrodes and their liquid junction po-tentials. In practice, activity coefficientsare assumed equal in the external and

10.2 Potentiometric Methods 313

internal solutions and concentrations aresubstituted for activities.

When the electrode is selective for butnot specific to the ion of interest (theprincipal ion), the Nicolsky–Eisenmanequation applies (Eq. 3):

E = constant +(

RT

ziF

)

× ln(ai +

∑Kij a

zi/zj

j

), (3)

where E is the measured potential, R, T ,and F have their usual meanings, zi andzj are the charges on the analyte and in-terfering ions, respectively, ai and aj arethe activities of the analyte and interferingions, respectively, and Kij is the poten-tiometric selectivity coefficient [14]. Thisempirical equation is now known to be in-accurate when two ions of different charge

contribute significantly to the measuredpotential [15], and detailed procedures toovercome this problem have been sug-gested [16].

10.2.2Zero-Current Potentiometry

Nonselective, metallic indicator electrodeshave been used for potentiometric mea-surements in complex biological media,for example, Pt electrodes have been usedto monitor the redox potential of fermenta-tion broths as cultures grow [17]. However,zero-current potentiometry more often in-volves ISEs based on solid membranes,composed of a sparingly soluble salt ofthe ion of interest or liquid membranes, inwhich an ion-selective reagent is dissolved,with the membrane separating reference

Internal Ag/AgClreference electrode

Plasticbody

(a) (b)

(d)(c)

Metal wire (e.g. Cu)

P-type Si

N-type drainChannel

N-Type source

Gate insulator

Encapsulation

Ion-selective membrane

Internal KCI solutionwith principal ion

Plasticbody

Internal Ag/AgClreference electrode

Porous supportimpregnated withhydrophobic ion-exchanger

IdPlasticized PVC withhydrophobic ion-exchanger, multiplydip-coated

Internal KCl solutionwith principal ion

Solid crystallinemembrane, e.g.Ag2S/AgCl

Fig. 1 ISEs: (a) The solid membrane; (b) liquid membrane; (c) coated-wire electrode;(d) field-effect transistor. The analyte solution is in contact with the ion-selective layers;potentials for A–C are measured using an external reference electrode, while the draincurrent, Id, is monitored using a current-to-voltage converter in D.


from the analyte solutions; more recently,solid-state devices have been introducedthat are based on coated metallic con-ductors or field-effect transistors (FETs).Examples of these potentiometric elec-trodes are shown in Fig. 1. An excellentintroduction to the principles of their op-eration is given by Cunningham [13].

10.2.2.1 ISEs in BioanalysisWell-established potentiometric devicesused in bioanalysis are based on macro-scopic electrodes that use gas-permeablepolymer, solid or liquid membranes se-lective to CO2, NH3, pH, or variousalkali, alkaline earth, or halide ions [7].The membranes separate the external an-alyte solution from the inner referencesolution, and these electrodes are usedin conjunction with an integral or exter-nal reference electrode. The selectivities ofthese membranes can be excellent, for ex-ample, the main interferent in the fluorideISE that uses a solid LaF3/EuF2 membraneis hydroxide, with a KF/OH value of 0.1;thus, over the pH range 0–8.5, the elec-trode responds in a Nernstian fashion tofluoride concentrations above 10−6 M [9].Solid Ag2S membranes doped with AgX,where X is Cl−, Br−, I−, CN−, or SCN−,are used for the determination of theseanions.

The oldest clinical applications of po-tentiometry involved blood gas analysis,and this method is still used to quanti-tate dissolved carbon dioxide: a glass pHelectrode is housed behind a membranepermeable to CO2, and a thin solutionlayer of hydrogen carbonate separates thepH electrode from the membrane. This de-vice, called the Stow-Severinghaus electrode,remains unchanged in principle from itsoriginal version reported in 1958 [18]. Themeasurement is based on the hydrolysis ofCO2 to HCO3

− and H+ in the thin solution

layer, and the measured local pH indicatesthe bulk pCO2 values. The ammonia gas-sensing electrode is based upon similarmeasurement principles: ammonium ionformation causes a local decrease in pHin the intermembrane space; althoughammonia itself is not clinically relevant,the ammonia gas-sensing electrode hasbeen incorporated into enzyme-based sen-sors for other, clinically important analytessuch as urea and creatinine [19].

ISEs for Li+, K+, Ca2+, and Mg2+ havebeen developed on the basis of liquidmembranes that contain ionophores: theseare hydrophobic chelating agents thatcontain selective binding sites for theion of interest. While the structures ofionophores used in commercially availabledevices are often proprietary, examples ofwell-studied ISE ionophores include 14-crown-4 ether for Li+ [20] and valinomycinfor K+ [21]. Valinomycin is 5000 timesmore selective toward K+ over Na+ and18 000 times more selective toward K+over H+.

Efforts to produce a phosphate-selectiveISE have been hindered by its diverse spe-ciation and lability in biological samples;a recent review describes various poten-tiometric and amperometric approachesto this problem [22]. Of the potentiomet-ric approaches, selectivity is most oftenachieved using inorganic or organometal-lic extracting agents such as organotincompounds in liquid-membrane ISEs,cobalt complexes or metallic cobalt incoated-wire and metallic electrodes, re-spectively. Nickel phosphate, silver phos-phate, and mixtures of lead precipitateshave also been used in phosphate ISEs.All of these sensors suffer from lim-ited selectivity. Enzyme-based sensors forphosphate have also been a topic of re-search; as yet, however, no commercialphosphate sensor exists [22].


The difference between free and totalconcentrations of ions can be significant inbiological samples and is particularly im-portant for calcium and magnesium [23].ISEs are known to respond to free ion activ-ity (Eq. 3), but it has recently been shownthat direct potentiometric measurementscan lead to total ionic concentrations ofcalcium if the composition of the innersolution induces a strong zero-current fluxof the primary ion toward the inner com-partment [24].

10.2.2.2 Enzymes, Antibodies, and NucleicAcids with Zero-Current PotentiometryPotentiometric electrodes have been usedas transducers in biosensors in which athin layer containing enzymes, antibodies,or whole cells separates the electrode sur-face from the analyte solution and providesselective recognition properties. Thin lay-ers containing enzymes or cells catalyzereactions that cause the localized produc-tion or consumption of the species towhich the electrode responds. Antibodies,incorporated into immunosensors, havebeen shown to cause small but measurablesignals upon antigen binding, but usuallyrequire a label (such as an enzyme) thatallows catalysis of a reaction that can befollowed potentiometrically [13].

Potentiometric enzyme electrodes havebeen demonstrated for many clinically im-portant analytes, including glucose, urea,creatinine, amino acids, uric acid, oxalate,and penicillin. Urea sensors, for example,have been developed that use immobilizedurease to produce HCO3

− and NH4+,

and this reaction has been detected usingpCO2 and pNH3 gas-sensing electrodes,glass pH electrodes, glass- or liquid-membrane ammonium ISEs, and ion-selective FETs. Chapter 4 of Volume 6 con-cerns the grafting of molecular propertiesonto semiconductor surfaces. Reviews of

enzyme-based potentiometric biosensorsare available [25–29], and detection limitsfor substrate are generally of the orderof 10−5 M. Further examples includea glutamate sensor based on glutamatedecarboxylase and a CO2 gas-sensing elec-trode [30], a urea sensor using urease and aliquid-membrane ISE for ammonium ionthat contains the ionophore nonactin [31],a disposable urea sensor based on ure-ase and an ammonium-sensitive polymermatrix [32], an organophosphate sensorusing a hydrolase enzyme in conjunctionwith a pH electrode [33], a lysine sensorusing lysine oxidase and a solid-state am-monium electrode [34], a glucose-sensitiveenzyme-modified FET coated with Nafionand applied to serum determinations [35],a cyanide sensor with a detection limitof 10−10 M based on enzyme inhibi-tion at a peroxidase-modified FET coatedwith poly(4-vinylpyridine-co-styrene) [36].Applications of enzyme-based potentio-metric biosensors in flow injection analy-sis have revealed only millimolar detectionlimits for urea and penicillin [37]. Intact,viable cells of the Hansenula polymor-pha (a source of alcohol oxidase) andGluconobacter oxidans (containing a xy-lose dehydrogenase) were immobilized onpH-sensitive FETs in sensors for formalde-hyde [38] and xylose [39], respectively.

Binding between antibodies or anti-gens, with one partner immobilized ona membrane, has been monitored poten-tiometrically; this kind of ‘‘direct’’ im-munosensor was originally proposed byJanata and demonstrated by the bindingof mannan to concanavalin A, a lectin [40].Aizawa’s group showed potentiometric re-sponses to antigen–antibody interactionson membranes [41, 42]. Baumann’s grouphas shown that solid-state coated-wire elec-trodes made of titanium or graphite canalso be used to detect immunological


binding reactions directly [43, 44]. Morerecently, a chloride ISE based on Ag2S hasbeen modified with antibody to thyroid-stimulating hormone (TSH) in a nonla-beled immunosensor for TSH [45]. In allof these sensors, the change in potentialthat occurs with analyte concentration issmall but reproducible. This topic has beenreviewed by Bergveld [46].

Enzymes are commonly used as la-bels in potentiometric immunosensorsto improve the detectability of theantibody–antigen interaction. The en-zyme label is bound either to the an-alyte species and used as a reagentin a competitive immunoassay scheme,or to a second antibody, used to la-bel antigen (analyte) that binds to aprimary antibody immobilized on thetransducer surface. Antibody-modifiedLangmuir-Blodgett films [47] and ion-selective polyvinylchloride (PVC) films [48]have been suggested for potentiometricimmunosensor design. Urease [48–50],glucose oxidase [50], β-lactamase [51], andperoxidase [52] have been used as labelsin immunosensors for model analytessuch as human and rabbit IgG, antihu-man IgG, Salmonella typhimurium., andα-2-interferon.

FETs have also been directly coupled tobiological processes, in so-called BioFETs,to record electrical signals from surface-adherent mammalian hippocampal neu-rons on an array of 16 micro-FETs [53] andto monitor the response of intact insectantennae to the plant odor component,Z-3-hexen-1-ol [54].

Total DNA has been quantitated usinga silicon light-addressable potentiometricsensor (LAPS) modified with a biotinylatednitrocellulose membrane to capture DNAlabeled with streptavidin (through a DNA-binding protein) and urease (through

an anti-DNA antibody) [55]. Sequence-selective detection of polymerase chainreaction products has been carried out in asimilar manner [56], and in both cases, thepH change due to reaction of the ureaselabel is the measured signal.

10.2.3Dynamic Potentiometry

PSA has been introduced as a methodfor the detection of trace metals and nu-cleic acid sequences in biological samples,and specialized instrumentation for PSAis commercially available (Radiometer).The method is based on monitoring po-tential changes at an electrode to which(a) constant current is applied [57] or (b) achemical oxidant is introduced [58]; thesehave been described in more detail ina monograph [59]. Both methods haverecently been applied to the determina-tion of copper, zinc, and selenium inhuman plasma and urine, using a mer-cury film working electrode [60]; chemicalstripping was effective for copper andzinc quantitation, using mercuric ionsas the oxidant in both cases, whereasconstant-current oxidation was used forthe determination of selenium via oxi-dation to hydrogen selenide. Cadmiumand lead have also been determined inblood using PSA [61]. Disposable, screen-printed electrodes for lead quantitation byPSA have been reported [62], and a modelimmunosensor has been proposed thatuses screen-printed electrodes and PSAdetection for monitoring the interactionof human serum albumin (HSA) with theHSA antibody, using Bi3+ as a markerion [63].

PSA was introduced as a bioanalyti-cal method for nucleic acid analysis byWang [64], who showed that transitionmetal complexes that are reversibly bound


to an electrode-immobilized double-stran-ded (ds) DNA layer can be stripped bythe application of a constant reducingcurrent to the electrode surface. Peaksin the chronopotentiograms correspond-ing to the maximum differential potentialchange that occurs as the complexes arereduced can be correlated, by height orarea, with the quantity of dsDNA on theelectrode surface. A variety of carbon-basedelectrode materials, including carbon pasteand screen-printed graphite inks for dis-posable sensors, were examined usingan adsorptive DNA immobilization pro-cedure. Peptide nucleic acids have beenused instead of native sugar-phosphatebackbone DNA to aid the detection ofpoint mutations [65–67], and applicationshave included the detection of the humanimmunodeficiency virus and the point mu-tation in the p53 gene associated withcancer [68, 69].

Recently, Mascini’s group has also usedthis method, in combination with dispos-able graphite screen-printed electrodes,initially for model DNA sequence detec-tion [70] and more recently for the quan-titation of human apolipoprotein E geno-types [71]; in both cases, daunomycin wasused as an indicator species. Daunomycinis a DNA intercalant bearing both quinoneand hydroquinone functionalities, andMascini’s group used constant-current ox-idation of the hydroquinone moiety for de-tection. Their detection limit of 1 µg mL−1

target DNA resulted in the need for PCRamplification prior to DNA detection.

10.2.4Emerging Technology

A fundamental breakthrough, first re-ported in 1992, has occurred in themeasurement of polyions by potentiom-etry [72]. In principle (Eqs. 2 and 3),

the determination of analytes possessinghigh charge would be expected to pro-ceed with extremely poor sensitivity, asthe calibration slope RT /zF is inverselyrelated to analyte ion charge. However, ithas recently been shown that much moresensitive devices for heparin, a polyan-ion, and protamine, a polycation, canbe constructed using PVC membranesdoped with tridodecylmethyl-ammoniumchloride, an anion exchanger, or di-nonylnaphthalene sulfonate, a cation ex-changer [73–75]. The selectivity and sensi-tivity of these membranes is truly remark-able: a potential difference of −50 mV wasobserved upon changing the porcine hep-arin (z = −70) concentration from 0.04 to0.4 µM, and even greater sensitivity wasobserved toward beef lung heparin, whichpossesses higher charge density, althoughEq. (1) predicts a potential change of lessthan 1 mV with a tenfold concentrationchange of species possessing such highcharge. The sensitivity and selectivity ofthese devices are believed to result froma nonequilibrium extraction process thatleads to tight ion pairs with the exchangerin the PVC membrane, causing chargeseparation at the sample-membrane in-terface. Applications of polyion-selectiveelectrodes to enzymatic assays (involvingproteases and their inhibitors) and modelimmunoassays (involving the avidin-biotinreaction) have been reviewed [76].

Further recent advances in potentiomet-ric bioanalysis include the invention ofself-plasticizing ion-selective membranes,in which the ionophores are covalently im-mobilized and no leachable plasticizer isincorporated into the photocurable mem-branes; near-Nernstian responses wereobserved for potassium, sodium, cal-cium, and pH sensors constructed usingcommercially available ionophores [77].


Microfabrication [78], and now nanofab-rication [79], of ISEs and FETs promises toenable multianalyte assays in continuous-readout implantable devices or for in vitrotesting with immunosensor arrays.

10.2.5Commercialization of PotentiometricDevices

Microfabrication has been the topic of arecent review in which thin-film (<1 µm,based on vacuum evaporation, sputteringor chemical vapor deposition) and thick-film (>10 µm, based on screen printingor lamination) technologies are describedfor the mass production of potentiomet-ric sensors and sensor arrays [80]. Currentchallenges include the cost of fabrication,especially for thin-film devices, the con-trol of physical dimensions of the sensingelements, the incorporation of liquid reser-voirs, and the stability of the integratedreference electrodes.

Many ISEs based on solid or liquid mem-branes and ion-selective FETs possesssufficiently small Kij values (Eq. 1) to-ward common interferents, and have beendeveloped into commercial benchtop ana-lyzers that allow the measurement of Na+,K+, Ca2+, Cl−, HCO3

−, and other ions,in instruments that are designed solely forISE measurements or in combination withother integrated devices (such as amper-ometric sensors and spectrophotometers)to measure other parameters [7]. Applica-tions of these devices to blood gas analysishave been reviewed recently [81]. Com-panies that produce this instrumentationfor clinical chemistry applications includeAbbott Diagnostics, AVL, Bayer Diagnos-tics, Beckman Coulter, Nova Biomedical,Ortho-Clinical Diagnostics, Boehringer-Mannheim/Roche, and Radiometer [82].Many instruments allow measurements to

be made directly on serum or whole blood,as these matrices possess relatively con-stant ionic strength (activity coefficientsvary with ionic strength and temperature);urine samples are generally diluted withionic strength buffer. Matrix effects havebeen a concern for ISEs that are usedfor whole blood assays, and the need forprotein-containing quality control mate-rials has been the topic of one recentreview [83].

Clinical interest in Mg2+ has increasedrecently, with the commercialization ofMg2+ ISEs based on neutral carrierionophores for benchtop clinical elec-trolyte instruments such as AVL andNova [84]. Magnesium is important forneuronal activity, cardiac excitability, mus-cular contraction, and blood pressure,among other physiological roles; it existsbound to proteins or to small anions andas free Mg2+, but has traditionally beenmeasured as total Mg. The recent avail-ability of ISEs combined with the knowndecrease in dietary magnesium intake inthe western world has stimulated clinicalresearch that may have important implica-tions in cardiac care [85] and diabetes [86],as it is now known that ‘‘ionized Mg’’ (freeMg2+) levels can vary, whereas total Mgremains constant.

Significant recent advances in microflu-idics and sensor miniaturization haveresulted in a new class of portable, point-of-care blood chemistry analyzer. The i-STATPortable Clinical Analyzer is perhaps themost developed of these devices. Figure 2shows a diagram of the i-STAT disposablecassette that contains thin-film potentio-metric ISEs for sodium, potassium, chlo-ride, ionized calcium, pH, pCO2, and urea(via urease and an ammonium ISE) in ad-dition to amperometric sensors (discussedin Sect. 10.3) for oxygen and glucose (viaglucose oxidase with hydrogen peroxide


Air bladder

Cartridge base

Puncturing barb

Calibrant pouch

Biosensor chips

Tape gasket

Sample entry well

Cartridge cover

Fluid channel

Sample entrywell gasket

Cartridge label

(a) (b)

Fig. 2 Exploded view of an i-STAT Cassette (a)and the incorporated biosensor chips (b). Bloodsamples are introduced through the sampleentry well gasket and flow through the fluidchannel, where they pass over the sensingelements of the biosensor chips. These elementsmeasure sodium, potassium, chloride, ionizedcalcium, pH, and pCO2 using ISEs.

Enzyme-containing elements measure urea,which is hydrolyzed to ammonium ions that aremeasured potentiometrically, and glucose, whichis measured amperometrically via enzymaticproduction of hydrogen peroxide. The pO2 ismeasured amperometrically and hematocrit isdetermined conductometrically. (Reproducedfrom company literature.)

detection) and a conductometric sensorfor hematocrit (see Sect. 10.4); once filledwith whole blood, the cassette is con-nected to a hand-held, battery-operatedunit containing measurement electronics,a liquid-crystal display (LCD) and, if de-sired, a small, detachable printer. Thisdevice has been tested for use in criti-cal and neonatal care and in hemodialysisunits with excellent results [87–89].

The LAPS represents another majoradvance in commercially available po-tentiometric bioanalysis [90]. In this de-vice, a pH-sensitive silicon nitride insu-lator functions as a phototransistor gate.

Changes in pH at the gate-solution in-terface cause titration of surface hydroxyland amino functions, causing a change ingate potential. Cellular metabolism causesacidification of the surrounding medium,and this device, commercialized as theCytosensor Microphysiometer (MolecularDevices), has been used to monitor cellularmetabolism via pH change under a vari-ety of stressing situations. The Cytosen-sor, shown schematically in Fig. 3, hasbeen used for the determination of com-pound toxicities [91], the study of recombi-nant adrenoceptor subtypes [92], and hasbeen suggested for the determination of


(a) (b)

Fig. 3 The Cytosensor Microphysiometer.(a) Flow chamber containing cells retained in adisk-shaped region between two microporouspolycarbonate membranes; the top of thesensing surface contacts the lower membraneand is illuminated by an LED; culture mediumflows across the outer surface of the upper

membrane. (b) Instrumentation electronicscontrol illumination and data acquisition, andthe periodic flow of culture medium; a completedevice contains eight flow chambers managed byone computer. (Reproduced from companyliterature.)

antibiotic susceptibilities [93]. A differentLAPS device, also based on pH measure-ment, has been used in an immunosensorfor the detection of pathogenic bacteria;in this device, antibodies are used to selec-tively detect E. coli O157 : H7 and to capturethe organisms on a membrane; detectionoccurs via a urease label that causes acidifi-cation of the solution during reaction withurea [94].

10.3Amperometric Methods

10.3.1Introduction

Amperometric methods include all thosein which current is the measured parame-ter. Arguably, the simplest of these meth-ods uses the fuel cell in which no potentialis applied: current is generated in responseto a thermodynamically favorable overallcell reaction that involves two electrodesin a solution; although uncommon as ananalytical method, the fuel cell approach

has been applied to the determination ofbiomass during cell cultivation for a varietyof organisms, and detection limits below105 cells mL−1 have been reported forP. aeruginosa, F. arbrescens, and E. coli [95,96].

The most established amperometricmethods for bioanalysis involve constant-potential amperometry, in which mass-transport-controlled oxidation or reduc-tion is used to quantitate the product(or substrate) of an enzymatic reaction.For rapid (reversible) electrochemical re-actions, Eq. (4) describes the current ob-tained at a planar working electrode:

i = (nFADC)

δ, (4)

where i is the limiting current, measuredat a potential where all analyte is oxidizedor reduced, n and F have their usual mean-ings, D is the diffusion coefficient of theelectroactive species, C is its bulk solu-tion concentration, and δ is the thicknessof the stagnant layer immediately adja-cent to the electrode surface (the Nernst

10.3 Amperometric Methods 321

diffusion layer). If transport of analyte tothe electrode surface occurs solely by dif-fusion, δ = (πDt)1/2, and Eq. (4) becomesthe well-known Cottrell equation. If theworking electrode is rotated (or if thesolution is stirred) at a fixed rate, δ be-comes constant and the current is invariantwith time.

Although time-dependent signals areinconvenient for bioassays that are notautomated, commercial instruments com-monly measure current at defined timesor integrate current for a defined period(in chronocoulometry) following the ap-plication of the potential: in both cases,the magnitude of the resulting signal(current difference or charge) is directlyproportional to the concentration of theredox-active species.

Cyclic, square wave, ac, and differentialpulse voltammetry have also been used forbioanalysis, although commercializationof specialized bioassay instruments thatexploit the increased selectivity of thesemethods has not yet occurred. Figure 4shows the applied waveforms and (re-versible) voltammetric responses for eachof these techniques. Equations describingthe peak currents may be found in mosttexts; of analytical importance is the di-rect proportionality between peak currentmagnitude and analyte concentration forall four techniques.

10.3.2Direct Amperometric Methods

This section considers analytes that maybe directly quantitated via their selectiveoxidation or reduction under appropriateconditions. The amperometric quantita-tion of dissolved oxygen, the oldest bioana-lytical application of amperometry, exploitsa gas-permeable membrane for selectivity;oxygen diffuses across this membrane andis reduced at an unmodified, internal elec-trode held at constant potential [81]. Theoriginal Clark electrode used a polyethy-lene membrane and a Pt internal workingelectrode [97], and this device replacedWarburg manometry for routine oxygenquantitation. A historical review of bio-logical oxygen measurements has beenpublished [98], and this focuses mainly onthe oxygen electrode.

Fig. 4 Voltammetric waveforms andresponses for a reversibly redox-activeanalyte: (a) cyclic voltammetry;(b) square wave voltammetry;(c) alternating current voltammetry; and(d) differential pulse voltammetry. Thearrows in the waveforms representtimes at which current is measured,while double-headed arrows in theresponse curves show the measuredparameters that are directly proportionalto analyte concentration.

Time

E

Time

E

E

Iac

Time

E

If

Ir

Time

E

If

Ir

E

∆i =

If −

I r

E

I net

= I

f − I r

E (−)I

(a)

(b)

(c)

(d)


These devices have been miniaturizedand extensively studied for in vivo appli-cations [99, 100]. The no-net-flux methodof in vivo calibration has been intro-duced for sensors housed in microdialysisprobes [100], but calibration is generallydone before implantation and after ex-plantation. Recently, the double potentialstep technique has been applied to thein vivo chronocoulometric quantitation ofoxygen using a gold working electrode andan activated carbon reference electrodeimplanted in the venous bloodstream ofanimals (8 dogs, 1 sheep, and 1 pig); al-though the working electrodes were indirect contact with blood, this study foundthat some biofouling by blood proteinsoccurred over an implantation time of 4years (9 of the 10 implanted electrodesshowed no tissue overgrowth), but oxygenmeasurements were not affected when theelectrodes were tested in buffer after theexplantation [101].

Catecholamines and other neurotrans-mitters, such as dopamine, 5-hydroxyin-dole-3-acetic acid, homovanillic acid, dihy-droxyphenylacetic acid, and 5-hydroxytryp-tophan (serotonin), are oxidizable atcarbon electrodes. Over the past twodecades, chromatographic and capillaryelectrophoretic techniques have been usedin conjunction with electrochemical de-tection for the separation, identification,and quantitation of these compoundsand their bioconjugates in body fluidsand microdialysates [102–107, and ref-erences therein]. A recent advance in-cludes the use of a redox-cycling electro-chemical cell for high-performance liquidchromatography (HPLC) detection that,through repeated oxidation and reductionof the reversibly electroactive dopamine,allows detection limits of 116 pM to beachieved [106]. A carbon fiber workingelectrode, positioned at the outlet of a

30-µm-diameter capillary, was electricallyisolated from the electrophoretic separa-tion using an integrated decoupler, whichallowed nanomolar detection limits to beachieved for dopamine in brain micro-dialysate [107]. Capillary electrophoresiswith electrochemical detection has allowedmicrosampling of neurotransmitters fromsingle neurons of the pond snail Planorbiscorneus, with which microelectrochemistryat single cells has also been done, using5- to 10-µm-diameter carbon fiber micro-electrodes sealed in glass to expose onlythe tip, which is beveled flat at a 45angle [108]. Carbon fiber working micro-electrodes have also been used for thereal-time in vivo detection of neurotrans-mission events [109]. Temporal resolutionis provided by high-speed chronoamper-ometry or fast-scan cyclic voltammetry atelectrodes coated with Nafion, a cation ex-changer that rejects ascorbate, which istypically present at concentrations 2 to 3orders of magnitude higher than thoseof dopamine, norepinephrine, and sero-tonin [109].

Exocytosis that results in the secretion of5-hydroxytryptophan and insulin has beenmonitored from single pancreatic β-cells(human, porcine, canine, mouse, and cul-tured tumor cells), using 9-µm-diametercarbon fiber electrodes [110]. Unmodi-fied electrodes allowed 5-HT quantitation,while modification with a ruthenium ox-ide/cyanoruthenate film allowed selectivequantitation of insulin at electrodes posi-tioned 1 µm away from cell surfaces.

An interdigitated gold microarray elec-trode coated with Nafion has recently beenused to detect serotonin in human wholeblood samples (20 µL) to monitor aller-gic response, during which serotonin issecreted into the bloodstream [111]. Re-sults showed excellent agreement betweenelectrochemically monitored serotonin


levels and IgE levels determined byimmunoassay.

Nitric oxide has been of considerablerecent interest because of its signalingproperties and its importance in patho-physiology [112, 113]. Methods for theelectrochemical detection of NO have beenreviewed [114, 115]. The two principal am-perometric methods have recently beencompared for in vivo NO detection: directoxidation at Pt/Ir electrodes, with selec-tivity provided by membrane coatings, oroxidation at Ni-porphyrin modified carbonfiber microelectrodes coated with Nafionfor anion rejection; of these, only the sec-ond method yielded signals that could beattributed to NO [116].

The redox behavior of nucleic acids, es-pecially at mercury and carbon electrodes,has been studied for decades, and a vari-ety of quantitative methods exist [117–120,and references therein]. Recently, under-ivatized nucleic acids in solution havebeen selectively quantitated by ac voltam-metry at copper electrodes [121], whileDNA bound to gold electrode surfacesvia gold-thiol chemisorption has beenquantitated by chronocoulometry usingcharge-compensating redox markers [122].Derivative square wave voltammetry hasbeen used to examine chemically in-duced DNA damage following reactionwith styrene oxide [123].

Cyclic voltammetry at a graphite work-ing electrode has been used directly onwhite and red blood cells from humanbone marrow; one of the three irreversibleanodic peaks observed with a bone mar-row blood cell mixture was only present inleukemic patients [124]. Cyclic voltamme-try has also been used directly on E. coli,rat jejunal mucosal tissue, and the enzymelactate dehydrogenase to evaluate oxidativedamage [125]. An earlier study showed thatbacteria can be classified as gram-positive

or gram-negative on the basis of peakpositions and currents in cyclic voltammo-grams performed at basal plane pyrolyticgraphite electrodes; the voltammetric re-sponse is believed to result from coenzymeA oxidation [126].

10.3.3Amperometric Measurements withEnzymes

Since the introduction of the Clark oxygenelectrode more than four decades ago [97],measurements of oxidase enzyme activi-ties and oxidase substrate concentrationshave been performed using this device ora similar device in which the oxidase prod-uct, hydrogen peroxide, is reoxidized tomolecular oxygen. It has also been shownthat oxidase enzymes, although quiteselective toward their electron-donatingsubstrates (glucose, lactate, cholesterol),are not very selective toward oxygen asan electron-acceptor: rapid reduction ofalternate species such as ferricyanide, fer-rocene derivatives, quinone derivatives,and the phenoxazine/phenothiazine classof dyes also occurs if these species pos-sess a formal potential that is morepositive than that of the enzyme’s ac-tive site [127–130]. These mediator com-pounds then act as shuttles to transferelectrons from the reduced enzyme to anamperometric electrode.

The first biosensor was reported in1962 [131, 132], and consisted of glucoseoxidase bound to a membrane coveringan amperometric oxygen electrode. Thiswas the first device that could be usedto measure the concentration of an en-zyme substrate without adding enzymeas a reagent to the analyte solution.The term biosensor now applies to anycombination of transducer (electrochem-ical, optical, piezoelectric, thermoelectric)


with biological recognition agent (enzyme,antibody, chemoreceptor, nucleic acid)that results in a device for the selectivemeasurement of an analyte species [133].Chapters 4 and 6 of Volume 10 considerelectrochemical biosensors and electro-catalysis at modified electrodes.

Several recent reviews have discussedthe coupling of selective biological re-dox processes with amperometric elec-trodes. A review of bioelectrochemistrywith emphasis on amperometric biosen-sors has been published as a sectionof the biannual review of dynamic elec-trochemistry [134], and reviews coveringrecent developments in clinical applica-tions of electroanalysis and biosensorsare also available [135–137]. A survey ofthe mechanistic aspects of amperometricbiosensors has appeared [138]; this reviewconsiders direct electron transfer from en-zyme to electrode, mediation by freelydiffusing species such as ferrocenes, medi-ation by electron-hopping through redox-active gels, mediation through conduct-ing polymers, mediator-modified enzymesand mediator-modified electrodes, andconcludes that electrochemically depositedpolymers are promising materials for inex-pensive, reproducible mass production ofminiaturized sensors and electrode arrays.Two recent reviews focus on electrosynthe-sized polymers as immobilization matricesfor enzymes and other biomolecules [139,140]. A very comprehensive review ofimmobilization techniques and media-tion mechanisms has appeared that em-phasizes communication between redoxproteins and conducting supports for bio-electronic applications [141]; these authorsfocus on the advantages and implicationsof tailored, organized protein architecturesin ordered and defined nanostructureswith the aim of improving electroniccoupling between redox proteins and

transducers (see also Chapter 9 of Vol-ume 10). More specialized reviews ofglucose sensors [142] and the applicationof biosensors as selective detectors forchromatographic and electrophoretic in-struments [143] are also available.

Challenges facing the development ofin vitro amperometric biosensors (inter-ference rejection, rapid response, repro-ducibility, response range) have been metin many cases, and commercially avail-able devices based on disposable teststrips that incorporate miniature two-or three-electrode electrochemical cellsare available for a variety of analytes(see Sect. 10.3.7). Thin-film and thick-filmtechnology [80] have been used to mass-produce reproducible sensing elements,and amperometric detection in oxidase-based devices occurs by peroxide oxidationor the oxidation of freely diffusing medi-ators such as ferricyanide and ferrocenederivatives. The screen-printing processfor disposable sensor preparation has alsobeen reviewed [144].

New developments in this area includeuric acid sensors based on the mediationof urate oxidase by a novel redox polymer,poly(N-methyl-o-phenylenediamine) [145],and by the freely diffusing mediator 1-methoxy-5-methylphenazinium [146], con-tinued research on the direct amperomet-ric detection of NADH [147] and the use ofredox mediators [148] for dehydrogenaseenzymes, to allow practical sensors thatexploit this large class of enzymes, and theuse of cytochrome P450-modified glassycarbon electrodes as drug metabolismbiosensors [149].

The area of long-term, continuous, invivo measurement using enzyme-basedamperometric biosensors has encoun-tered some unique challenges in addi-tion to the problems of biocompatibil-ity, miniaturization, and the need for


telemetric data transmission that havealso been encountered during the devel-opment of in vivo amperometric oxygensensors [150]. Implantable sensors mustideally be ‘‘reagentless’’, that is, the ad-dition of reagents, either enzymes orcosubstrates, is not possible once thedevices have been implanted. Glucose sen-sors in which the enzyme layer can berecharged, or replaced every ten days, havebeen proposed [142, 151], but this is notideal for the user and may result in non-compliance. In addition, chemicals fromthe sensor must not escape into the sur-rounding environment.

Two approaches to reagentless invivo amperometric glucose sensors haveemerged, and these devices are illustratedin Fig. 5. In the first, immobilized glucoseoxidase catalyzes the oxidation of glucosewith concomitant reduction of molecu-lar oxygen; no artificial electron acceptorsare used and the hydrogen peroxide isdetected at a platinum [152] or platinum-iridium anode [153, 154]. Although thisdesign has been criticized because of theinherently variable in vivo oxygen levels,Wilson’s group has shown that oxygenlevels of 10 mm Hg and above yield sen-sor responses to 10 mM glucose that are90% of those achieved with high oxy-gen levels; with 2 mM glucose, virtuallyno oxygen dependence was observed at8 mm Hg and above [154]. The main ad-vantage of this design is the absence ofartificial mediators that could leach intothe surrounding environment. However,peroxide oxidation requires an applied po-tential of 600 to 700 mV vs Ag/AgCl, and atthese potentials, oxidation of interferencessuch as ascorbate and urate may occur.Membrane materials such as Nafion, acation exchanger that prevents anion trans-port to the anode, combined with celluloseacetate, have been used to minimize this

kind of interference [154, 155]. Nafion hasbeen shown to mineralize via deposition ofcalcium phosphate, both in vitro (cell cul-ture medium) and in vivo in subcutaneousimplants in rats [156], but in the glucosesensors, the Nafion-cellulose acetate inter-ference rejection membranes are placedbetween the immobilized enzyme layerand the Pt/Ir anode to prevent direct con-tact with the biological environment [154,155]. An outer layer of mixed polysiloxaneand polyurethane is used to provide bio-compatibility. With these devices, sensitiv-ities of 3 to 4 nA mM−1 are achieved [154].

The second approach employs artifi-cial electron acceptors, and in vivo ex-periments have involved sensors thatcross-link glucose oxidase in redox-activehydrogels used as coatings on goldanodes [157, 158]. These hydrogels areblock copolymers formed by reaction ofpoly[(1-vinylimidazole)osmium(4,4′-dime-thylbipyridine)2Cl] with poly(ethylene gly-col) diglycidyl ether 400 in the presenceof glucose oxidase, and form the firstof the three layers on the gold elec-trode surface; mediation occurs by thereduction of Os(III) by the enzyme, andelectron-hopping through the hydrogel al-lows Os(III) regeneration by oxidationat the gold electrode. Interference rejec-tion and glucose flux restriction occursat the second layer, composed of Nafionand cellulose acetate, whereas the thirdlayer is a biocompatible polyethylene ox-ide film. In air, these sensors show 2to 3 nA mM−1 sensitivity, but under ar-gon, the sensitivity improves to 4 to5 nA mM−1 because the native reactionof glucose oxidase with oxygen can occur,but the peroxide produced by this reactionis not detected at the applied potential of200 mV versus saturated calomel electrode(SCE) [157]. These sensors have beenimplanted in jugular veins of rats, where


1. Sensing layer (PVI-DME-Os/GOX/PEG)

2. Dynamic Range Expanding layer (CA:Nafion:PVPA/PAZ)

3. Biocompatible layer (PEO)

Outer layer

0.09 mm

0.30 mm1 2 3Polyimide

insulation 0.25 mm Gold

Connectingpoint

Enzyme layer

Cellulose acetate membrane

Epoxy

0.11

0.3 0.5

0.11

0.1625 − 30

Nafion membrane

Pt-Ir wire

Teflon coating

Inner layer

Ag/AgCl reference

(a)

(b)

Fig. 5 The two main approaches to amperometric enzyme-based biosensors for invivo measurements: (a) Wilson’s group uses a Pt-Ir wire electrode coated with Nafionand cellulose acetate for interference rejection, followed by the enzyme layer and abiocompatible outer layer; the enzyme generates hydrogen peroxide, which is oxidizedat the electrode (Reproduced from Y. Hu, G. S. Wilson, J. Neurochem. 1997, 68,1745–1752). (b) Heller’s group uses a gold electrode modified first with anenzyme/mediator polymer layer, followed by a dynamic range expanding layer that alsoaids in interference rejection, and finally a biocompatible outer layer; once reduced byglucose, the enzyme glucose oxidase is reoxidized by Os(III), and the generated Os(II)is oxidized at the gold surface to yield the measured current. (Reproduced fromJ. G. Wagner, D. W. Schmidtke, C. P. Quinn et al., Proc. Natl. Acad. Sci. U.S.A. 1998, 95,6379–6382.)

oxygen concentrations are relatively lowand invariant [157], and subcutaneouslyin chimpanzees [158]; comparison of thecontinuous in vivo sensor measurementswith periodic in vitro blood glucosemeasurements was performed in response

to various stimuli, including insulinand glucose administration, and goodagreement between these methods wasobserved.

Calibration of implanted devices re-mains an issue to be resolved. In research


trials, sensors are calibrated in vivo bycomparison of sensor readings with ac-cepted in vitro blood glucose measure-ments (e.g. finger-stick measurements),and can be calibrated in vitro before im-plantation and after explantation [150, 154,157]. Long-term implants suffer from bio-fouling, and a continuous decrease insensitivity occurs [159, 160]. New meth-ods for in vivo calibration have beenproposed [161, 162], and work is contin-uing in this area. Implants are ideallymade subcutaneously, and analyte levelsin the interstitial fluid, rather than bloodor plasma levels, are measured. It has re-cently been shown, over a three-day trialin dogs, that subcutaneous amperomet-ric peroxide-based glucose measurementsparallel plasma glucose levels with onlya slight delay (<10 min), occurring whenblood levels fluctuate [163].

Further reports of importance to thearea of in vivo sensing include the use ofvoltammetric waveforms for interferencediscrimination in amperometric glucoseoxidase–modified, peroxide-based carbonfiber microelectrodes [164], the modifi-cation of carbon fiber microelectrodeswith osmium-based redox polymer con-taining glutamate oxidase and ascorbateoxidase for in vivo rat brain glutamate andascorbate measurements [165], the devel-opment of lactate oxidase-peroxide-baseddevices for in vivo lactate measurementsin rat brain and subcutaneous tissue [166,167], and the integration of lactate andglucose sensors onto a single silicon con-tainment sensor [168].

10.3.4Amperometric Immunoassays andImmunosensors

The use of antibodies as analyticalreagents allows strong binding and se-lective recognition of analytes. Antibodies

are glycoproteins that possess two iden-tical antigen-binding sites and have amolecular weight of about 150 kD. The an-tibody–antigen reaction itself is not easilymonitored with readily available equip-ment, so a label is generally bound toeither the antibody or the antigen. Ei-ther the antibody or the antigen may bethe analyte species, but more commonly,it is the antigen; the antigen may bea small molecule (a hapten) or a largeprotein or glycoprotein in which an epi-tope, or surface structure, combines withone of the antibody paratopes or bind-ing sites. Immunoassays are described ascompetitive, where analyte competes witha labeled analyte reagent for a limitednumber of binding sites, or noncompet-itive, where quantitative binding of analyteto excess antibody is detected using asecond labeled antibody. Typically, anti-bodies have antigen association constantsbetween 105 and 1012 M−1 [169], so that, inprinciple, competitive immunoassays arecapable of detecting picomolar concentra-tions of analytes, whereas noncompetitiveassays are potentially capable of detect-ing concentrations that are several ordersof magnitude lower [170]. In heteroge-neous immunoassays, the labeled speciesreact with binding partners that are im-mobilized on a solid support, such as amicrotiter plate or an electrode surface,and excess labeled reagent is removed bya rinse step. Homogeneous assays relyon a change in the properties of the la-bel that occurs upon binding between thelabeled reagent and the binding partner;as the properties of the label change ina detectable manner, separation of thebound from the free labeled reagent isunnecessary.

Amperometric methods have been usedto detect labels that are simple redox-active molecules or enzymes that generate


redox-active species as a reaction product.Enzymatic labels have the inherent ben-efit of amplification: a single enzymelabel can produce thousands of detectablemolecules. In this section, amperometricimmunoassays and immunosensors areconsidered together because immunosen-sors are simply heterogeneous immunoas-says in which the working electrode itselffunctions as the solid support. Recentreviews of electrochemical immunoas-says [171], immunoassays [172], and im-munosensors in clinical analysis [173]have been published; all of these have somediscussion devoted to amperometric mea-surements. Advantages of amperometricmeasurements over more traditional ab-sorbance or fluorescence measurementsinclude speed, accuracy, precision, and theability to make measurements on turbidor intensely colored samples.

One review focuses on methods forattaining zeptomole detection limits inamperometric immunoassays [174]. Enzy-matic amplification schemes have beenproposed for absorbance-based immuno-assays in which a labeling enzyme, alka-line phosphatase, dephosphorylates NADPto begin a second amplification stageinvolving alcohol dehydrogenase and di-aphorase; this ‘‘cascade’’ scheme wastested in a noncompetitive immunoas-say for TSH and found to have a de-tection limit of 10 zmol with an in-tensely colored formazan product [175,176]. Cascade schemes have also beentested in electrochemical immunoassays;detection limits in the picomolar to fem-tomolar ranges have been found [177,and references therein]. Enzymes thathave now been tested in amperomet-ric immunoassays include alkaline phos-phatase [178–181], peroxidase [182–189],glucose oxidase [190–192], glucose-6-phosphate dehydrogenase [193], and

catalase [194]. The microporous gold elec-trode used by Meyerhoff’s group allowssubstrate diffusion from the back sideof the electrode, with the alkaline phos-phatase label bound on the front side of theelectrode; this scheme allows separation-free immunoassays because both productformation and detection occur at the sur-face of the electrode [180].

Chemical amplification schemes havealso been tested in amperometric im-munoassays. For example, Guo andcoworkers used a copper ion label in a6 : 1 ratio to the labeled species, HSA, ina heterogeneous competitive immunoas-say format. Following the rinse step, Cu2+was released into the solution by lower-ing the pH, and the increased temperatureallowed the copper-catalyzed conversionof o-phenylenediamine to the electroactiveproduct, 2,3-diaminophenazine, and thedetection limit for albumin was loweredby two orders of magnitude to 7 ng mL−1,compared with direct metal ion label-ing [195].

Further methodological improvementshave been shown using alkaline phos-phatase as a labeling enzyme. Capil-lary electrochemical enzyme immunoas-say, in which an inner capillary wall isused as the solid support and detectionoccurs downstream in an amperomet-ric flow cell, has been introduced byHeineman’s group and shown to havea 3-pmol detection limit for phenobar-bital in serum [196]. Alternate substratesto replace 4-aminophenylphosphate havebeen investigated: 1-naphthylphosphatewas shown to be preferred for the de-termination of progesterone in milk [197]and 3-indoxyl phosphate, which gener-ates indigo dimer that is readily detectedby ac voltammetry, yielded a detectionlimit for human IgG of 0.1 nM [198]. Im-munoassay methods are not limited to


the detection of molecular analytes, andseveral amperometric immunoassays havenow been reported for the detection ofpathogens. The detection of Salmonellaspecies is of importance in food pro-cessing, and amperometric enzyme-linkedassays have been described for its quan-titation at levels as low as 8000 cellsmL−1 [197, 199]. Staphylococcus aureus isalso of interest in food production, andan amperometric immunosensor has beenreported for its detection [200]. Entero-hemorrhagic Escherichia coli O157 : H7has become a worldwide health problem,and amperometric filtration-capture [201]and immunomagnetic flow injection im-munoassays [202] have been developedwith detection limits of 5 × 103 and1 × 105 cells mL−1, respectively. Differ-ential pulse voltammetry has been usedin an enzyme-linked immunoassay forHelicobacter pylori–specific IgG; the sub-strate 3,3′-5, 5′-tetramethylbenzidine wasused for the peroxidase label, and thismethod showed a sevenfold improvementin detection limit compared with spec-trophotometric detection [203].

10.3.5Amperometric Nucleic Acid Assays andSensors

In 1987, Bard’s group reported that thebinding of small, redox-active transitionmetal polypyridine complexes with highmolecular weight nucleic acids in solu-tion can be studied by voltammetry; thedecrease in the apparent diffusion coef-ficient that occurs upon binding led tosignificantly decreased voltammetric peakcurrents, allowing the determination of as-sociation constants [204]. Since then, sev-eral groups have studied electrode-boundnucleic acids and oligonucleotides in aneffort to develop sequence-selective DNA

biosensors [205–213] and sensors for thestudy of small molecule interactions withDNA [214–216]. A short review of electro-chemical DNA sensors was published in1996 [217], and significant developmentshave occurred since that time.

Covalently immobilized calf thymusDNA on glassy carbon electrodes was de-tected using tris(2,2′-bipyridyl)cobalt(III)and tris(1,10-phenanthroline)cobalt(III),complexes that are reversibly electroac-tive at moderate applied potentials [205].Immobilization using carbodiimide andN-hydroxysuccinimide reagents to gener-ate NHS esters on the carbon surfaceoccurs at the guanosine residues [206],and this allowed immobilization of short,single-stranded oligonucleotides possess-ing a (dG)n tail [206, 207]. The two cobaltcomplexes associate selectively with thedouble-stranded form of DNA and bindonly weakly to single-stranded forms;thus, the recognition of complementaryDNA that results in the hybridizationon the electrode surface between surface-bound probe and soluble target sequencesyields in significantly higher voltammet-ric peak currents for the reduction ofthe cobalt complexes. The detection of amodel sequence and the F508 deletionsequence associated with cystic fibrosishave been accomplished with syntheticoligonucleotides [206, 207] and with 400-base PCR products amplified from humanDNA [208].

Oligonucleotides modified with a termi-nal thiol group can be chemisorbed ontogold electrodes and, after hybridizationwith a soluble complementary sequence,detected voltammetrically using the redox-active bis (benzimide) dye Hoechst33 258 [209]. This group also showed thatvoltammetric peak potentials for acri-dine orange [210] and daunomycin [211]


could be used to distinguish surface-bounddouble- and single-stranded DNA.

Electrocatalytic oxidation of guanosineresidues by electrogenerated tris(2,2′-bipyridyl)ruthenium(III) has recently beenshown to provide a useful transduc-tion mechanism in DNA biosensors [212].Probe strands in which inosine wassubstituted for guanosine were immobi-lized on an electrode surface, and thesehybridized selectively with targets thatcontained oxidizable guanosine residues;PCR-amplified genomic DNA from herpessimplex virus type II, Clostridium perfrin-gens, and human immunodeficiency viruswere detected by this method.

Recent work with DNA-modified goldelectrodes has shown that single-base mis-matches (point mutations) can be detectedin immobilized double-stranded oligonu-cleotides using redox-active moleculesthat intercalate between the stacked basepairs [213]. In this work, charge trans-port from daunomycin, methylene blue,and an iridium polypyridine complex (allintercalators) through synthetic 15-baseoligodeoxynucleotides was significantly at-tenuated when single-base mismatcheswere present at any of the three locationsin the hybridization product; a noninterca-lating DNA-binding complex, rutheniumpentamine chloride, showed unaffectedvoltammetric signals in the presence ofthe mismatches.

Voltammetry at modified electrodes hasalso been used to investigate the interac-tions of redox-active DNA-binding agentswith immobilized DNA. Examples includethe survey done by Hashimoto and cowork-ers [211], an investigation of methyleneblue [214], a comparison of interactions ofredox cations with surface-bound and solu-tion DNA [215], and an investigation of thekinetics and mechanism of mitoxantrone

interaction with DNA immobilized onglassy carbon electrodes [216].

10.3.6Emerging Technology

Several reviews of new materials forchemical and biochemical sensors haveappeared recently [218–221]. New appli-cations of sol-gels (see also Volume 1,Chapter 8), zeolites, organic polymers, andconducting composites to sensing devicesinclude small inorganic ions, gases, smallorganic molecules, proteins, and DNA.Sensors and assays employing molecu-larly imprinted polymers have also beenreviewed [222, 223]. Extensive fundamen-tal research on alkanethiol self-assembledmonolayers is expected to yield repro-ducibly fabricated enzyme electrodes inwhich control of molecular architecture isreadily achieved [224].

Fabrication of sensors and sensor ar-rays is in a state of rapid development.The groups led by Kuhr [225, 226] andHeineman [227, 228] have made signifi-cant advances in microfabrication of sen-sor arrays for simultaneous multianalyteamperometric assays and sensors; bothhave applied their arrays to immunoas-says [226, 227]. Further research effortstoward controlled-release microchip fabri-cation [229] and nanoscale chemical sen-sors [230] have also shown promise andwill aid in the development of sensorsand sensing arrays for in vitro and in vivomeasurements.

10.3.7Commercialization of AmperometricDevices

Biosensors based on hydrogen peroxidedetection have been sold by Yellow SpringsInstruments (YSI) for many years, and


the YSI 2700 is their most advancedinstrument [231]. The YSI biosensors arebased on a Pt electrode, covered witha cellulose acetate dialysis membrane,an enzyme layer, and a polycarbonatemembrane that limits substrate diffu-sion, so that the enzyme itself doesnot limit the reaction rate. This instru-ment is marketed for food analysis andbiotechnology and allows the following ox-idase substrates to be determined: choline(to 450 mg L−1), D-glucose (to 9 g L−1

with one membrane and 25 g L−1 witha second type), ethanol (to 3.2 g L−1),galactose (to 25 g L−1), hydrogen perox-ide (to 600 mg L−1), lactose (to 25 g L−1),L-glutamate (to 10 mM), L-glutamine (to8 mM), L-lactate (to 2.67 g L−1), methanol(to 2.50 g L−1), and sucrose (to 25 g L−1).A simple flow injection system is used totransport analyte into the flow cell in whichthe biosensors are housed. Nova manufac-tures the BioProfile series, designed forcell culture applications, and allows hydro-gen peroxide–based biosensor detectionof glucose, lactate, glutamine, and gluta-mate. Many commercially available clinicallaboratory analyzers that use ISEs forpotentiometric blood gas and electrolytequantitation also employ amperometricmeasurements for routine assays of dis-solved oxygen, glucose, and lactate inblood, urine, and other body fluid sam-ples [82].

Point-of-care testing has now becomefeasible with the introduction of thehandheld i-STAT instrument and simi-lar portable devices, such as the opticallybased Reflotron. The i-STAT (Fig. 2) con-tains both potentiometric and amperomet-ric sensors integrated into sensor arraysthat are included within disposable cas-settes [87–89]. Eight tests are possible withthe most advanced cassette, the EC8+, andthese are sodium, potassium, chloride,

urea nitrogen, glucose, pH, pCO2, andhematocrit (five additional values are cal-culated from these results, including bicar-bonate, total carbon dioxide, base excess,anion gap, and hemoglobin).

Devices designed for use by individualswith no technical training have tradition-ally focused on the need for regular glucosemeasurements by diabetic patients. Sincethe first introduction of the ExacTech am-perometric glucose-monitoring device byMediSense Inc. (which was taken over byAbbott in 1996), several other electrochem-ical glucose meters have been introduced.In 1998, the market for such deviceswas dominated by four companies: LifeS-can (42%), Boehringer-Mannheim/Roche(18%), Abbott/MediSense (16%), andBayer (12%); all others together held12% of the total market [232]. Of thefour main companies, all now marketan amperometric glucose self-testing de-vice. LifeScan recently introduced theFastTake, Boehringer-Mannheim/Rocheprovides the Accu-Chek Advantage andAccu-Chek Complete devices, Abbott pro-duces the Precision Q.I.D. and the Pre-cision Q.I.D. Pen (the ExacTech line isnow sold only in Europe), and Bayer of-fers the Glucometer Elite. Clinical testingresults for these devices have been re-ported [233–235].

The monitoring of blood lactate levelsin high-performance athletics has led tothe introduction of handheld devices, sim-ilar to the devices used for blood glucosemonitoring by diabetics [236]. The AnaloxP-LM-5 (Analox Instruments, England)uses a 5-µL blood sample and produces avalue within 20 s of application of the sam-ple to the lactate oxidase/peroxide-basedtest strip. The Lactate Pro (Arkray Inc.,Japan) is another lactate oxidase–based,amperometric, handheld device for lactatemonitoring, and is marketed for athletes


who desire peak performance during com-petitions. These devices compete with theRoche Accusport, which is based on re-flectance photometry, and are sold forabout US $400.

An exciting development in individ-ual monitoring is occurring with thecommercial introduction of subcutaneousimplantable glucose sensors for contin-uous and long-term monitoring. It hasbeen shown that subcutaneous glucosemeasurements reflect plasma glucose lev-els, but the subcutaneous implants donot cause the sensor fouling that hasbeen observed with blood measurements.The MiniMed Continuous SubcutaneousBlood Glucose (CSBG) Monitoring Sys-tem (Applied Medical Technology Co., UK)was approved for physician-supervised useby the US FDA in February, 1999. It isbased on glucose oxidase/hydrogen per-oxide detection technology and consistsof a three-electrode cell housed in flexi-ble polymer tubing with a side windowto expose the active electrode area thatis covered by a polyurethane membrane.The miniaturized sensor is embedded in asplit 22-gauge needle that is removed afterthe sensor is inserted into the abdominalregion. The sensor is connected to an ex-ternal pager-sized meter by wires, allowingdata collection every 10 s, and an averagevalue calculated every minute [237]. A sec-ond CSBG instrument is being developedby Synthetic Blood International (Ohio)for similar applications; their device usesa silicone-sheathed Pt electrode with theglucose oxidase enclosed in a semiperme-able cellulose acetate membrane, coupledwith a titanium-encased battery and mi-croprocessor [238].

Electrochemical DNA sensors and sen-sor arrays are being developed by at leastthree companies in the United States. Clin-ical Microsensors, Inc. (Pasadena CA) is

interested in direct electron transfer fromferrocene-labeled DNA to gold electrodes,while Xanthon, Inc. (Research TrianglePark, NC) is developing sensors on thebasis of electrocatalytic guanine oxida-tion [239, 240]. AndCare Inc. (Durham,NC) is developing a multichannel elec-trochemical system for quantitative mon-itoring of PCR amplification using inter-mittent pulse amperometry and a labelingenzyme (peroxidase) mediated by tetram-ethylbenzidine [241].

10.4Impedimetric Methods

10.4.1Introduction

Impedimetric measurements are basedon the nonfaradaic response observed be-tween two electrodes immersed in a sam-ple solution when a high frequency (1 to300 kHz) alternating potential is applied.The amplitude of the applied waveform issmall, and this, combined with the highfrequency relative to other electrochemi-cal methods, effectively prevents faradaicreactions from occurring to any signifi-cant extent. In contrast to potentiometricand amperometric methods, impedimetricmethods are not analyte-selective; instead,bulk properties of the solution and the elec-trode–solution interfaces are monitored.

The impedance cell can be thought ofas a simple electrical circuit, composed oftwo capacitors, that represent the two elec-trode–solution interfaces, separated by aseries of three resistors that symbolize thepolarization resistance at each electrodesurface, with the electrolyte (solution) re-sistance in the center, as shown in Fig. 6.The overall impedance of the cell Z con-sists of an overall capacitance C and a total

10.4 Impedimetric Methods 333

Fig. 6 Equivalent circuit representationof a typical impedance cell used inmicrobiology. The symbols Ri and Cirepresent the interfacial resistances andcapacitances, respectively, while Rm isthe resistance of the medium separatingthe two electrodes.

1 2

21

Ci

Ri RiRm

Ci

resistance R, and depends on the angu-lar frequency ω of the applied potentialwaveform, as shown in Eq. (5):

Z =(

R2 + 1

C2ω2

)1/2

. (5)

Instrumentation allows the measure-ment of capacitance, resistance, or itsinverse, conductance, and is based onWheatstone bridge electronics. The firstcommercial instrument for bioanalyticalmeasurements consisted of a referencecell and a sample cell placed in oppos-ing arms of the bridge; this device, theStrattometer, was used to study blood co-agulation and microorganism growth inculture media [242].

10.4.2Impedimetric Assays and Sensors

A review of the diverse food hygieneapplications of impedance microbiologyhas recently been published [243]. The de-tection of Enterobacteriaceae, particularlySalmonella spp., the determination of to-tal bacterial counts, and the examinationof antibiotics and food additives for an-timicrobiological effects are important.Another short review describes a modelexplaining the effects of microbial growthand metabolism on capacitance at elec-trodemedium interfaces [244].

Recently reported impedimetric assaysalso focus on the detection and enumer-ation of microorganisms: the detectionof E. coli in potable water [245] and the

quantitation of bacterial content in milksamples [246]. Antimicrobial susceptibilitytesting has also been performed on gram-negative bacilli by an impedance as-say [247]. Dielectric spectroscopy has beenused to study the internal structures ofcells [248], and has been used in con-junction with multivariate calibration andartificial neural networks to quantitate themetabolic substrates in biological cell sus-pensions [249].

Impedimetric measurements are inher-ently nonselective, but the modificationof electrode surfaces with selective recog-nition agents has resulted in selectivesensing devices based on impedimetrictransducers. Reviews of immunosensorsand cell-based biosensors include sectionson impedance transduction [177, 250, andreferences therein]. A review of bioelec-tronic noses, devices intended to mimicthe human olfactory system in detectingand identifying odors, includes interdigitalstructures for complex impedance mea-surements and arrays of capacitive sensorelements [251].

Hendji and coworkers studied inter-digitated pairs of gold film electrodesprepared by micromachining [252]. Thisdevice records a differential signal be-tween an enzyme-modified surface anda control surface modified with an in-ert protein; the studied enzymes in-cluded urease, glucose oxidase, andacetylcholinesterase. Darbon and cowork-ers [253] and Souteyrand and cowork-ers [254] studied antigen–antibody inter-actions, and this measurement method


was included in Bergveld’s review [46].Nucleic acids have also been detectedby impedimetric methods, using single-stranded DNA-modified electrodes [255].A self-assembled peptide monolayer hasbeen used for the impedimetric detectionof viruses [256].

Polymer membranes have shown greatpromise for impedimetric sensors. Shep-pard and coworkers coated interdigitatedplatinum electrodes with a pH-sensitivehydrogel that swelled with a change inpH [257]. McNeil and coworkers exploiteda polymer that dissolves at pH values above7; in their urea sensor designed with abase transducer of screen-printed gold inkelectrodes, urease was incorporated as anouter layer over this polymer, and thepresence of urea in solution resulted inthe dissolution of the inner polymer layerwith very large concomitant impedancechanges [258]. Screen-printed interdigi-tated carbon electrodes coated with alayer of poly(methylvinyl ether)/maleic an-hydride modified by esterification withn-octanol were further modified with lay-ers of urease or creatinine deaminase andthen used to monitor urea or creatininelevels in serum [259]; in this work, alkalinepH caused ester hydrolysis, and this led tomarked changes in capacitance.

10.4.3Commercially Available Instruments forImpedance Measurements

At present, four companies provide spe-cialized impedance measurement instru-mentation for clinical microbiology orfood hygiene applications. BioMirieuxmanufactures the Bactometer, which canmonitor conductance, capacitance, orimpedance in up to two temperature-controlled incubators, each containing 256samples. Malthus Instruments produces

the System V, which monitors only con-ductance, but can handle up to 1200samples. Don Whitley Scientific manu-factures the ‘‘Rapid Automated BacterialImpedance Technique’’ (RABIT) instru-ment, which is capable of measuringconductance in a maximum of 16 incu-bators, each containing 32 individuallytemperature-controlled electrode tubes.Sy-Lab offers the BacTrac, which mon-itors total impedance in 6 individuallytemperature-controlled incubators, eachcontaining a maximum of 40 impedancetubes. All of these instruments monitorrelative or absolute changes in impedanceor conductance at regular intervals (e.g.every 6 min) [243].

10.5Future Directions and Perspectives

Electroanalytical methods have become in-creasingly important for practical assays incomplex biological media. The commer-cial promise of ISEs and electrochemicalbiosensors has now been realized for thein vitro quantitation of many clinicallyimportant analytes, and new devices forin vivo quantitation are being commer-cialized. Point-of-care and physician officelaboratory testing is now a reality, but thefundamental problem of quality controlmust be addressed. The American Asso-ciation of Bioanalysts recently conducteda large-scale comparison of clinical testresults obtained from the physician of-fice laboratories and specialized off-siteclinical laboratories, and very large discrep-ancies were found [260]. Devices designedfor this type of application, and for useby nonprofessionals, must be designed foruser-independent results.

The i-STAT Portable Clinical Ana-lyzer [87–89] is one example of the recent

10.5 Future Directions and Perspectives 335

trend toward the incorporation of electro-chemical sensing elements into arrays forsimultaneous multianalyte analysis. Otherexamples of this kind of microfabricated to-tal analysis systems, or µTAS devices, thatincorporate electrochemical transducerssuch as ISEs and amperometric enzymeelectrodes have been introduced and areexpected to gain commercial importancein clinical settings [261, 262].

Acknowledgment

The Natural Sciences and EngineeringResearch Council of Canada is gratefullyacknowledged for the ongoing financialsupport of the author’s research program.

References

1. A. J. Bard, L. R. Faulkner, ElectrochemicalMethods, Fundamentals and Applications,2nd Ed., Wiley & Sons, New York, 2001.

2. P. T. Kissinger, W. R. Heineman, Labora-tory Techniques in Electroanalytical Chem-istry, 2nd ed., Marcel Dekker, New York,1996.

3. T. Riley, C. Tomlinson, Principles of Electro-analytical Methods, Wiley & Sons, New York,1987.

4. A. Evans, Potentiometry and Ion SelectiveElectrodes, Wiley & Sons, New York, 1987.

5. T. Riley, A. Watson, Polarography and OtherVoltammetric Methods, Wiley & Sons, NewYork, 1987.

6. C. M. A. Brett, A. M. O. Brett, Electroanaly-sis, Oxford University Press, Oxford, 1998.

7. G.-A Junter, (Ed.), Electrochemical DetectionTechniques in the Applied Biosciences: Volume1: Analysis and Clinical Applications, EllisHorwood, Chichester, 1988.

8. G.-A. Junter, (Ed.), Electrochemical DetectionTechniques in the Applied Biosciences: Volume2: Fermentation and Bioprocess Control,Hygiene and Environmental Sciences, EllisHorwood, Chichester, 1988.

9. J. Wang, Electroanalytical Techniques in Clin-ical Chemistry and Laboratory Medicine, VCHPublishers, New York, 1988.

10. W. F. Smith, Voltammetric Determination ofMolecules of Biological Significance, Wiley &Sons, New York, 1992.

11. J. P. Hart, Electroanalysis of Biologically Im-portant Compounds, Ellis Horwood, Chich-ester, 1990.

12. D. G. Buerk, Biosensors: Theory and Appli-cations, Technomic Publishing, Lancaster,1993.

13. A. J. Cunningham, Introduction to Bioana-lytical Sensors, Wiley & Sons, New York,1998.

14. B. P. Nicolsky, M. M. Schultz, A. A. Beli-justin in Glass Electrodes for Hydrogen andOther Cations, (Ed.: G. Eisenman), MarcelDekker, New York, 1967.

15. E. Bakker, R. K. Meruva, E. Pretsch et al.,Anal. Chem. 1994, 66, 3021–3030.

16. E. Bakker, E. Pretsch, P. Buehlmann, Anal.Chem. 2000, 72, 1127–1133.

17. M. J. Silva, J. L. Wong, Bioelectrochem.Bioenerg. 1995, 37, 141–148.

18. J. W. Severinghaus, A. F. Bradley, J. Appl.Physiol. 1958, 13, 515–520.

19. L. Campanella, M. P. Sammartino, M. Tom-assetti, Analyst 1990, 115, 827–830.

20. V. P. Y. Gadzekpo, J. M. Hungerford, A. M.Kadry et al., Anal. Chem. 1985, 57, 493–495.

21. A. Malinowska, M. E. Meyerhoff, Anal.Chem. 1998, 70, 1477–1488.

22. S. O. Engblom, Biosens. Bioelectron. 1998,13, 981–994.

23. B. M. Altura, B. T. Altura, Scand. J. Clin.Lab. Invest. 1996, 56 (Suppl. 224), 211–234.

24. A. Ceresa, E. Pretsch, E. Bakker, Anal.Chem. 2000, 72, 2050–2054.

25. I. Karube, K. Ikebukuro, Y. Murakami et al.,Ann. N.Y. Acad. Sci. 1995, 750, 101–108.

26. H. S. Yim, C. E. Kibbey, S. C. Ma et al.,Biosens. Bioelectron. 1993, 8, 1–38.

27. J. M. Kauffman, G. G. Guilbault, BioprocessTechnol. 1991, 15, 63–82.

28. J. Janata, Principles of Chemical Sensors,Plenum Press, New York, 1989.

29. D. Monroe, Crit. Rev. Clin. Lab. Sci. 1989,27, 109–158.

30. G. V. Diaz, L. H. el-Issa, M. A. Arnold et al.,J. Neurosci. Methods 1988, 23, 63–69.

31. M. H. Gil, A. P. Piedade, S. Alegret et al.,Biosens. Bioelectron. 1992, 7, 645–652.

32. C. Eggenstein, M. Borchardt, C. Diekmannet al., Biosens. Bioelectron. 1999, 14, 33–41.

33. A. Mulchandani, P. Mulchandani, I. Kanevaet al., Anal. Chem. 1998, 70, 4140–4145.


34. J. Saurina, S. Hernandez-Cassou, S. Alegretet al., Biosens. Bioelectron. 1999, 14, 67–75.

35. S. V. Dzyadevich, Y. I. Korpan, V. N.Arkhipova et al., Biosens. Bioelectron. 1999,14, 283–287.

36. V. Volotovsky, N. Kim, Biosens. Bioelectro.1998, 13, 1029–1033.

37. R. Koncki, I. Walcerz, E. Leszczynska,J. Pharm. Biomed. Anal. 1999, 19, 633–638.

38. Y. I. Korpan, M. V. Gonchar, A. A. Sibirnyet al., Biosens. Bioelectron. 2000, 15, 77–83.

39. A. N. Reshetilov, M. V. Donova, D. V.Dovbnya et al., Biosens. Bioelectron. 1996,11, 401–408.

40. J. Janata, J. Am. Chem. Soc. 1975, 97,2914,2915.

41. M. Aizawa, A. Morioka, S. Suzuki et al.,Anal. Biochem. 1979, 94, 22–28.

42. N. Yamamoto, Y. Nagasawa, M. Sawai et al.,J. Immunol. Methods 1978, 22, 309–317.

43. U. Pfeifer, W. Bauman, Fresenius J. Anal.Chem. 1992, 343, 541–549.

44. L. Engel, W. Bauman, Fresenius J. Anal.Chem. 1994, 349, 447–450.

45. Z.-H. Lin, G.-L. Shen, Q. Miao et al., Anal.Chim. Acta 1996, 325, 87–92.

46. P. Bergveld, Biosens. Bioelectron. 1991, 6,55–72.

47. T. Dubrovsky, S. Vakula, C. Nicolini, Sens.Actuators, B 1994, 22, 69–73.

48. R. Koncki, A. Owczarek, W. Dzwolak et al.,Sens. Actuators, B 1998, 47, 246–250.

49. K. Dill, L. H. Stanker, C. R. Young, J. Bio-chem. Biophys. Methods 1999, 41, 61–67.

50. L. Campanella, R. Attioli, C. Colapicchioniet al., Sens. Actuators, B 1999, 55, 23–32.

51. T. A. Sergeyeva, A. P. Soldatkin, A. E.Rachkov et al., Anal. Chim. Acta 1999, 390,73–81.

52. A. L. Ghindilis, P. Atanasov, E. Wilkins,Sens. Actuators, B 1996, 34, 528–532.

53. A. Offenhausser, C. Sprossler, M. Mat-suzawa et al., Biosens. Bioelectron. 1997, 12,819–826.

54. P. Schroth, M. J. Schoning, P. Kordos et al.,Biosens. Bioelectron. 1999, 14, 303–308.

55. V. T. Kung, P. R. Panfili, E. L. Sheldon et al.,Anal. Biochem. 1990, 187, 220–227.

56. J. D. Olson, P. R. Panfili, R. F. Zuk et al.,Mol. Cell. Probes 1991, 5, 351–358.

57. D. Jagner, Trends Anal. Chem. 1983, 2,53–56.

58. S. Bruckenstein, J. W. Bixler, Anal. Chem.1965, 37, 786–790.

59. J. Wang, Stripping Analysis: Principles, In-strumentation and Applications, VCH, Deer-field Beach, 1985.

60. M. L. Gozzo, L. Colacicco, C. Calla et al.,Clin. Chim. Acta 1999, 285, 53–68.

61. P. Ostapczuk, Clin. Chem. 1992, 38,1995–2001.

62. J. Wang, B. Tian, Anal. Chem. 1993, 65,1529–1532.

63. J. Wang, B. Tian, K. R. Rogers, Anal. Chem.1998, 70, 1682–1685.

64. J. Wang, C. Xiaohua, G. Rivas et al., Anal.Chim. Acta 1996, 326, 141–147.

65. J. Wang, Biosens. Bioelectron. 1998, 13,757–762.

66. E. Palecek, M. Fojta, M. Tomschik et al.,Biosens. Bioelectron. 1998, 13, 621–628.

67. J. Wang, E. Palecek, P. E. Nielsen et al.,J. Am. Chem. Soc. 1996, 118, 766–770.

68. J. Wang, X. Cai, G. Rivas et al., Anal. Chem.1996, 68, 2629–2634.

69. J. Wang, G. Rivas, X. Cai et al., Anal. Chim.Acta 1997, 344, 111–118.

70. G. Marrazza, I. Chianella, M. Mascini,Biosens. Bioelectron. 1999, 14, 43–51.

71. G. Marrazza, G. Chiti, M. Mascini et al.,Clin. Chem. 2000, 46, 31–37.

72. S. Ma, V. C. Yang, M. E. Meyerhoff, Anal.Chem. 1992, 64, 694–697.

73. B. Fu, E. Bakker, V. C. Yang et al., Macro-molecules 1995, 28, 5834–5840.

74. J. H. Yun, V. C. Yang, M. E. Meyerhoff,Anal. Biochem. 1995, 224, 212–220.

75. L.-C. Chang, M. E. Meyerhoff, V. C. Yang,Anal. Biochem. 1999, 276, 8–12.

76. S. Dai, J. M. Esson, O. Lutze et al., J. Pharm.Biomed. Anal. 1999, 19, 1–14.

77. L. Y. Heng, E. A. H. Hall, Anal. Chem. 2000,72, 42–51.

78. A. Uhlig, E. Lindner, C. Teutloff et al., Anal.Chem. 1997, 69, 4032–4038.

79. M. Versen, B. Klehn, U. Kunze et al., Ultra-microscopy 2000, 82, 159–163.

80. E. Lindner, R. P. Buck, Anal. Chem. 2000,72, 336A–345A.

81. C. E. W. Hahn, Analyst 1998, 123, 57R–86R.82. Internet addresses for these com-

panies are: www.abbottdiagnostics.com,www.avlmed.com, www.bayerdiag.com, www.beckman.com, www.boehringer-mannheim.com, www.novabiomedical.com, and www.orthoclinical.com; a review of a recently-introduced point-of-care blood gas andelectrolyte analyzer from Radiometer can


be found in J. Lindemans, P. Hoefkens,A. L. van Kessel et al., Clin. Chem. 1999,45, 111–117.

83. B. Maas, R. Sprokholt, A. Maas et al., Scand.J. Clin. Lab. Invest. Suppl. 1996, 224,179–186.

84. R. J. Elin, E. N. Hristova, S. A. Cecco et al.,Scand. J. Clin. Lab. Invest. Suppl. 1996, 224,203–210.

85. B. M. Altura, B. T. Altura, Scand. J. Clin.Lab. Invest. Suppl. 1996, 224, 211–234.

86. N. Mikhail, K. Ehsanipoor, South Med. J.1999, 92, 1162–1166.

87. D. Bingham, J. Kendall, M. Clancy, Ann.Clin. Biochem. 1999, 36, 66–71.

88. J. N. Murthy, J. M. Hicks, S. J. Soldin, Clin.Biochem. 1997, 30, 385–389.

89. M. H. Gault, C. E. Harding, S. Duffett et al.,Nephron 1998, 80, 344–348.

90. H. M. McConnell, J. C. Owicki, J. W. Parceet al., Science 1992, 257, 1906–1912.

91. D. Cooke, R. O’Kennedy, Anal. Biochem.1999, 274, 188–194.

92. M. Pihlavisto, M. Scheinin, Eur. J. Pharma-col. 1999, 385, 247–253.

93. G. T. Baxter, L. J. Bousse, T. D. Dawes et al.,Clin. Chem. 1994, 40, 1800–1804.

94. A. G. Gehring, D. L. Patterson, S. I. Tu,Anal. Biochem. 1998, 258, 293–298.

95. S. Nishikawa, S. Sakai, I. Karube et al.,Appl. Environ. Microbiol. 1982, 43, 814–818.

96. H. P. Bennetto, J. L. Stirling, K. Tanakaet al., Biotechnol. Bioeng. 1983, 25, 559–568.

97. L. C. Clark, Trans. Am. Soc. Artif. Intern.Organs 1956, 2, 41–47.

98. J. W. Severinghaus, P. B. Astrup, J. Clin.Monitor. 1986, 2, 174–189.

99. D. G. Buerk, A. G. Tsai, M. Intaglietta et al.,Microcirculation 1998, 5, 219–225.

100. X. Z. Liang, Y. Zhang, C. E. Lunte, J. Pharm.Biomed. Anal. 1998, 16, 1143–1152.

101. N. Holmstroem, P. Nilsson, J. Carl-sten et al., Biosens. Bioelectron. 1998, 13,1287–1295.

102. J. Cummins, L. M. Matheson, J. F. Smyth,J. Chromatogr. 1990, 528, 43–53.

103. C. D. Forster, I. A. Macdonald, Biomed.Chromatogr. 1999, 13, 209–215.

104. C. Holmes, G. Eisenhofer, D. S. Goldstein,J. Chromatogr., B 1994, 653, 131–138.

105. P. Tuamainen, P. T. Mannisto, Eur. J. Clin.Chem. Clin. Biochem. 1997, 35, 229–235.

106. J. K. Cullison, J. Waraska, D. J. Buttaro et al.,J. Pharm. Biomed. Anal. 1999, 19, 253–259.

107. J. Qian, Y. Wu, H. Yang et al., Anal. Chem.1999, 71, 4486–4492.

108. B. B. Anderson, A. G. Ewing, J. Pharm. Bio-med. Anal. 1999, 19, 15–32.

109. D. J. Michael, R. M. Wightman, J. Pharm.Biomed. Anal. 1999, 19, 33–46.

110. C. A. Aspinwall, L. Huang, J. R. T. Lakeyet al., Anal. Chem. 1999, 71, 5551–5556.

111. M. Okochi, H. Yokouchi, N. Nakamuraet al., Biotechnol. Bioeng. 1999, 65, 480–484.

112. L. Liaudet, F. G. Soriano, C. Szabo, Crit.Care Med. 2000, 28, N37–N52.

113. F. L. Kiechle, T. Malinski, Am. J. Clin.Pathol. 1993, 100, 567–575.

114. T. Malinski, S. Mesaros, P. Tomboulian,Methods Enzymol. 1996, 268, 58–69.

115. D. Christodoulou, S. Kudo, J. A. Cook et al.,Methods Enzymol. 1996, 268, 69–83.

116. N. Villeneuve, F. Bedioui, K. Voituriez et al.,J. Pharmacol. Toxicol. Methods 1998, 40,95–100.

117. V. Brabec, Biopolymers 1979, 18, 2397–2404.118. D. Krznavic, B. Cosovic, Anal. Biochem.

1986, 156, 454–462.119. E. Palecek, Anal. Biochem. 1988, 170,

421–431.120. C. Teijeiro, K. Nejedly, E. Palecek, J. Biomol.

Struct. Dyn. 1993, 11, 313–331.121. P. Singhal, W. Kuhr, Anal. Chem. 1997, 69,

4828–4832.122. A. B. Steel, T. M. Herne, M. J. Tarlov, Anal.

Chem. 1998, 70, 4670–4677.123. J. Mbindyo, L. Zhou, Z. Zhang et al., Anal.

Chem. 2000, 72, 2059–2065.124. H.-N. Li, Y.-X. Ci, J. Feng et al., Bioelec-

trochem. Bioenerg. 1999, 48, 171–175.125. R. Kohen, J. Pharmacol. Toxicol. Methods

1993, 29, 185–193.126. T. Matsunaga, T. Nakajima, Appl. Environ.

Microbiol. 1985, 50, 238–242.127. A. E. G. Cass, G. Davis, G. D. Francis et al.,

Anal. Chem. 1984, 56, 667–671.128. S. A. Jaffari, J. C. Pickup, Biosens. Bioelec-

tron. 1996, 11, 1167–1175.129. T. Kaku, H. I. Karan, Y. Okamoto, Anal.

Chem. 1994, 66, 1231–1235.130. M. L. Fultz, R. A. Durst, Anal. Chim. Acta

1982, 140, 1–18.131. L. C. Clark, C. Lyons, Ann. N.Y. Acad. Sci.

1962, 102, 29–45. A later version, usinghydrogen peroxide oxidation, is describedin (132).

132. S. J. Updike, G. P. Hicks, Nature 1967, 214,986–988.


133. A. P. F. Turner, I. Karube, G. S Wilson,(Eds.), Biosensors: Fundamentals and Appli-cations, Oxford University Press, Oxford,1987. pp. v–vii.

134. J. L. Anderson, L. A. Coury, J. Leddy, Anal.Chem. 1998, 70, 519R–589R.

135. J. Wang, Anal. Chem. 1999, 71, 328R–332R.136. J. Wang, J. Pharm. Biomed. Anal. 1999, 19,

47–53.137. E. Magner, Analyst 1998, 123, 1967–1970.138. K. Habermueller, M. Mosback, W. Schuh-

mann, Fresenius J. Anal. Chem. 2000, 366,560–568.

139. E. Palmisano, P. G. Zambonin, D. Cen-tonze, Fresenius J. Anal. Chem. 2000, 366,586–601.

140. S. Cosnier, Biosens. Bioelectron. 1999, 14,443–456.

141. I. Willner, E. Katz, Angew. Chem., Int. Ed.Engl. 2000, 39, 1181–1218.

142. E. Wilkins, P. Atanasov, Med. Eng. Phys.1996, 18, 273–288.

143. H. A. Fishman, D. R. Greenwald, R. N. Zare,Annu. Rev. Biophys. Biomol. Struct. 1998, 27,165–198.

144. J. P. Hart, S. A. Wring, Trends Anal. Chem.1997, 16, 89–103.

145. T. Nakaminami, S.-I. Ito, S. Kuwabata et al.,Anal. Chem. 1999, 71, 1928–1934.

146. T. Nakaminami, S.-I. Ito, S. Kuwabata et al.,Anal. Chem. 1999, 71, 4278–4283.

147. M. A. Hayes, W. G. Kuhr, Anal. Chem.1999, 71, 1720–1727.

148. M. Mayer, M. Genrich, W. Kunnecke et al.,Anal. Chim. Acta 1996, 324, 37–45.

149. E. I. Iwuoha, S. Joseph, Z. Zhang et al., J.Pharm. Biomed. Anal. 1998, 17, 1101–1110.

150. C. Henry, Anal. Chem. 1998, 70,594A–598A.

151. P. Atanasov, S. Yang, C. Salehi et al.,Biosens. Bioelectron. 1997, 12, 669–680.

152. W. K. Ward, J. E. Troupe, ASAIO J. 1999,45, 555–561.

153. J. P. Lowry, M. Miele, R. D. O’Neill et al.,J. Neurosci. Methods 1998, 79, 65–74.

154. Y. Hu, G. S. Wilson, J. Neurochem. 1997, 68,1745–1752.

155. Y. Zhang, Y. Hu, G. S. Wilson et al., Anal.Chem. 1994, 66, 1183–1188.

156. R. C. Mercado, F. Moussy, Biosens. Bioelec-tron. 1998, 13, 133–145.

157. J. G. Wagner, D. W. Schmidtke, C. P. Quinnet al., Proc. Natl. Acad. Sci. U.S.A. 1998, 95,6379–6382.

158. D. W. Schmidtke, A. Heller, Anal. Chem.1998, 70, 2149–2155.

159. N. Wisniewski, F. Moussy, W. M. Reichert,Fresenius J. Anal. Chem. 2000, 366, 611–621.

160. V. Thome-Duret, M. N. Gangnerau,Y. Zhang et al., Diabetes Metab. 1996, 22,174–178.

161. B. Aussedat, V. Thome-Duret, G. Reachet al., Biosens. Bioelectron. 1997, 12,1061–1071.

162. J. P. Lowry, R. D. O’Neill, M. G. Boutelleet al., J. Neurochem. 1998, 70, 391–396.

163. K. Rebrin, G. M. Steil, W. P. Van Antwerpet al., Am. J. Physiol. 1999, 277, E561–E571.

164. L. I. Netchiporouk, N. F. Shram, N. Jaff-rezic-Renault et al., Anal. Chem. 1996, 68,4358–4364.

165. N. V. Kulagina, L. Shankar, A. C. Michael,Anal. Chem. 1999, 71, 5093–5100.

166. Y. Hu, G. S. Wilson, J. Neurochem. 1997, 69,1484–1490.

167. D. Pfeiffer, B. Moeller, N. Klimes et al.,Biosens. Bioelectron. 1997, 12, 539–550.

168. J. Perdomo, C. Sundermeier, H. Hinkerset al., Biosens. Bioelectron. 1999, 14, 27–32.

169. B. Harlow, D. Lane, Antibodies: A LaboratoryManual, Cold Spring Harbor Laboratory,Cold Spring Harbor, 1988.

170. R. Ekins, Nature 1989, 340, 256–258.171. A. Warsinke, A. Benkert, F. W. Scheller,

Fresenius J. Anal. Chem. 2000, 366, 622–634.172. D. S. Hage, Anal. Chem. 1999, 71,

294R–304R.173. R.-I. Stefan, J. F. K. van Staden, H. Y.

Abdoul-Enein, Fresenius J. Anal. Chem.2000, 366, 659–668.

174. M. A. Cousino, W. R. Heineman, H. B.Halsall, Ann. Chim. 1997, 87, 93–101.

175. C. J. Stanley, A. Johansson, C. H. Self, J.Immunol. Methods 1985, 83, 89–95.

176. A. Johannsson, D. H. Ellis, D. L. Bates et al.,J. Immunol. Methods 1986, 87, 7–11.

177. A. L. Ghindilis, P. Atanasvo, M. Wilkinset al., Biosens. Bioelectron. 1998, 13,113–131.

178. R. Renneberg, W. Schoessler, F. Scheller,Anal. Lett. 1983, 16, 1279–1289.

179. S. H. Jenkins, H. B. Halsall, W. R. Heine-man, Anal. Biochem. 1988, 168, 292–299.

180. M. W. Ducey, A. M. Smith, X. Guo et al.,Anal. Chim. Acta 1997, 357, 5–12.

181. Y. Qu, L. R. Berghman, F. Vandesande,Anal. Biochem. 1998, 259, 167–175.


182. I. Abdel-Hamid, A. L. Ghindilis, P. Atanasovet al., Anal. Lett. 1999, 32, 1081–1094.

183. J. Rishpon, D. Ivnitski, Biosens. Bioelectron.1997, 12, 195–204.

184. D. Ivnitski, T. Wolf, B. Solomon et al., Bio-electrochem. Bioenerg. 1998, 45, 27–32.

185. J. M. F. Romero, M. Stiene, R. Kast et al.,Biosens. Bioelectron. 1998, 13, 1107–1115.

186. M. Santadreu, A. Alegret, L. Fabregas, Anal.Chim. Acta 1999, 396, 181–188.

187. F. Wendzinski, B. Gruendig, R. Renneberget al., Biosens. Bioelectron. 1997, 12, 43–52.

188. A. F. Chetcuti, D. K. Y. Wong, M. C. Stuart,Anal. Chem. 1999, 71, 4088–4094.

189. M. A. Lopez, F. Ortega, E. Dominguez et al.,J. Mol. Recognit. 1998, 11, 178–181.

190. G. A. Robinson, V. M. Cole, G. C. Forrest,Biosensors 1988, 3, 147–160.

191. A. Benkert, F. Scheller, W. Schoessler et al.,Anal. Chem. 2000, 72, 916–921.

192. R. W. Keay, C. J. McNeil, Biosens. Bioelec-tron. 1998, 13, 963–970.

193. D. Athley, C. J. McNeil, W. R. Bailey et al.,Biosens. Bioelectron. 1993, 8, 415–419.

194. B. Mirhabibollahi, J. L. Brooks, R. G. Kroll,J. Appl. Bacteriol. 1990, 68, 577–585.

195. W. Guo, J.-F. Song, M.-R. Zhao et al., Anal.Biochem. 1998, 259, 74–79.

196. J. Zhang, W. R. Heineman, H. B. Halsall,J. Pharm. Biomed. Anal. 1999, 19, 145–152.

197. J. L. Brooks, B. Mirhabibollahi, R. G. Kroll,J. Appl. Bacteriol. 1992, 73, 189–196.

198. C. Fernandez-Sanchez, M. B. Gonzalez-Garcia, A. Costa-Garcia, Biosens. Bioelectron.2000, 14, 917–924.

199. B. Mirhabibollahi, J. L. Brooks, R. G. Kroll,J. Appl. Bacteriol. 1990, 68, 577–585.

200. A. G. Gehring, C. G. Crawford, R. S. Ma-zenko et al., J. Immunol. Methods 1996, 195,15–25.

201. R. S. Mazenko, F. Rieders, J. D. Brewster,J. Microbiol. Methods 1999, 36, 157–165.

202. F. G. Perez, M. Mascini, I. E. Tothill et al.,Anal. Chem. 1998, 70, 2380–2386.

203. Y.-N. He, H.-Y. Chen, J.-J. Zheng et al.,Talanta 1997, 44, 823–830.

204. M. T. Carter, M. Rodriguez, A. J. Bard,J. Am. Chem. Soc. 1989, 111, 8901–8911.

205. K. M. Millan, A. Spurmanis, S. R.Mikkelsen, Electroanalysis 1992, 4, 929–932.

206. K. M. Millan, S. R. Mikkelsen, Anal. Chem.1993, 65, 2317–2323.

207. K. M. Millan, A. Saraullo, S. R. Mikkelsen,Anal. Chem. 1994, 66, 2943–2948.

208. K. M. Millan, S. R. Mikkelsen, presented atACS National Meeting, San Francisco, CA,April 15, 1997.

209. K. Hashimoto, K. Ito, Y. Ishimori, Anal.Chem. 1994, 66, 3830–3833.

210. K. Hashimoto, K. Miwa, Y. Ishimori,Supramol. Chem. 1993, 2, 265–268.

211. K. Hashimoto, K. Ito, Y. Ishimori, Anal.Chim. Acta 1994, 286, 219–224.

212. M. E. Napier, C. R. Loomis, M. F. Sistareet al., Bioconjugate Chem. 1997, 8, 906–913.

213. S. O. Kelley, E. M. Boon, J. K. Barton et al.,Nucleic Acids Res. 1999, 27, 4830–4837.

214. S. O. Kelley, J. K. Barton, Bioconjugate Chem.1997, 8, 31–37.

215. A. B. Steel, T. M. Herne, M. J. Tarlov, Bio-conjugate Chem. 1999, 10, 419–423.

216. A. M. Oliviera Brett, T. R. A. Macedo,D. Raimundo et al., Biosens. Bioelectron.1998, 13, 861–867.

217. S. R. Mikkelsen, Electroanalysis 1996, 8,15–19.

218. M. E. Tess, J. A. Cox, J. Pharm. Biomed.Anal. 1999, 19, 55–68.

219. H. Mkuguruma, I. Karube, Trends Anal.Chem. 1999, 18, 62–68.

220. F. Cespedes, S. Alegret, Trends Anal. Chem.2000, 19, 276–285.

221. F. Cespedes, E. Martinez-Fabregas, S. Ale-gret, Trends Anal. Chem. 1996, 15, 296–304.

222. T. Takeuchi, J. Haginaka, J. Chromatogr., B1999, 728, 1–20.

223. K. Yano, I. Karube, Trends Anal. Chem.1999, 18, 199–204.

224. J. J. Gooding, D. B. Hibbert, Trends Anal.Chem. 1999, 18, 525–533.

225. S. E. Rosenwald, N. Dontha, W. G. Kuhr,Anal. Chem. 1998, 70, 1133–1140.

226. N. Dontha, W. B. Nowall, W. G. Kuhr,J. Pharm. Biomed. Anal. 1999, 19, 83–91.

227. Y. Ding, L. Zhou, H. B. Halsall et al.,J. Pharm. Biomed. Anal. 1999, 19, 153–161.

228. C. A. Wijayawardhana, G. Wittstock, H. B.Halsall et al., Anal. Chem. 2000, 72,333–338.

229. J. T. Santini, M. J. Cima, R. Langer, Nature1999, 397, 335–338.

230. J. Kong, N. R. Franklin, C. Zhou et al., Sci-ence 2000, 287, 622–625.

231. The internet address for Yellow SpringsInstruments is www.ysi.com.

232. The marketing firm of Frost & Sullivan,San Jose, CA performed this research andpublished it on their website, www.frost.com.


233. M. E. Collison, P. J. Stout, T. S. Glushkoet al., Clin. Chem. 1999, 45, 1665–1673.

234. A. G. Glasmacher, W. Brennemann,C. Hahn et al., Exp. Clin. Endocrinol. Dia-betes 1998, 106, 360–364.

235. G. J. Kost, H.-T. Vu, J. H. Lee et al., Crit.Care Med. 1998, 26, 581–590.

236. Internet addresses for Analox Instru-ments and Arkray Inc. are www.analox.comand www.arkray.co.jp. See also N. Shimojo,K. Naka, H. Uenoyama et al., Clin. Chem.1993, 39, 2312–2314, for a description ofthe Lactate Pro.

237. The internet address for Applied MedicalTechnology Co. is www.minimed.com. Seealso J. J. Mastrototaro, J. Pediatr. Endocrinol.Metab. 1999, 12 (Suppl. 3), 751–758.

238. The internet address for Synthetic BloodInternational is www.sybd.com..

239. The internet addresses for Clinical Mi-crosensors and Xanthon are www.microsen-sor.com. and www.xanthoninc.com.

240. E. K. Wilson, Chem. Eng. News 1998, 76,47–49.

241. M. Wojciechowski, R. Sundseth, M. Morenoet al., Clin. Chem. 1999, 45, 1690–1693.

242. G. A. Junter, in Electrochemical DetectionTechniques in the Applied Biosciences, Vol. 1,(Ed.: G. A Junter), Ellis Horwood, Chich-ester, 1988, pp. 139–156.

243. M. Wawerla, A. Stolle, B. Schalch et al.,J. Food Prot. 1999, 62, 1488–1496.

244. P. A. Noble, J. Microbiol. Methods 1999, 37,45–49.

245. K. O. Colquhoun, S. Timms, C. R. Fricher,J. Appl. Bacteriol. 1995, 79, 635–639.

246. C. J. Felice, R. E. Madrid, J. M. Olivera et al.,J. Microbiol. Methods 1999, 35, 37–42.

247. A. H. Huang, J. J. Wu, Y. M. Weng et al.,J. Clin. Microbiol. 1998, 36, 2882–2886.

248. K. Asami, T. Yonezawa, H. Wakamatsuet al., Bioelectrochem. Bioenerg. 1996, 40,141–145.

249. A. M. Woodward, A. Jones, X.-Z. Zhanget al., Bioelectrochem. Bioenerg. 1996, 40,99–132.

250. C. Ziegler, Fresenius J. Anal. Chem. 2000,366, 552–559.

251. C. Ziegler, W. Goepel, H. Haemmerle et al.,Biosens. Bioelectron. 1998, 13, 539–571.

252. A. M. N. Hendji, N. Jaffrezic-Renault,C. Martelet et al., Sens. Actuators, B 1994,21, 123–129.

253. P. Darbon, V. Michel, F. Math et al., Anal.Chem. 1998, 70, 5072–5078.

254. E. Souteyrand, J. R. Martin, C. Martelet,Sens. Actuators, B 1994, 20, 63–69.

255. E. Souteyrand, J. P. Cloarec, J. R. Mar-tin et al., J. Phys. Chem. B 1997, 101,2980–2985.

256. M. Knichel, P. Heiduschka, W. Beck et al.,Sens. Actuators, B 1995, 28, 85–94.

257. N. F. Sheppard, M. J. Lesho, P. McNallyet al., Sens. Actuators, B 1995, 28, 95–102.

258. C. J. McNeil, D. Athey, M. Ball et al., Anal.Chem. 1995, 67, 3928–3935.

259. W. O. Ho, S. Krause, C. J. McNeil et al.,Anal. Chem. 1999, 71, 1940–1946.

260. J. Hurst, K. Nickel, L. H. Hilborne, J. Am.Med. Assoc. 1998, 279, 468–471.

261. E. Pearson, A. Gill, P. Vadgama, Ann. Clin.Biochem. 2000, 37, 119–145.

262. I. Poels, S. Picioreanu, L. J. Nagels et al.,Biomed. Chromatogr. 2000, 14, 30–31.

341

11Interfacial Properties ofProteins/SpectroelectrochemicalStudies

Katsumi NikiIllinois State University, Normal, Illinois

11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343

11.2 Methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34411.2.1 Surface-enhanced Resonance Raman Scattering Spectroscopy . . . . . 34411.2.2 Electroreflectance (Potential-Modulated UV-vis Reflectance)

Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34511.2.3 Infrared Reflection Absorption Spectroscopy . . . . . . . . . . . . . . . . 34611.2.4 Absorption Linear Dichroism and Total Internal Reflection

Fluorescence Spectroscopies . . . . . . . . . . . . . . . . . . . . . . . . . . . 346

11.3 Redox Behavior of Heme- and Flavoproteins at Unmodified ElectrodeSurfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346

11.3.1 Gold Electrode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34611.3.1.1 Cytochrome c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34611.3.1.2 Cytochrome c3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34711.3.2 Silver Electrode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34711.3.2.1 Cytochrome c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34711.3.2.2 Cytochrome c3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35011.3.2.3 Cytochrome c552 (Thermus thermophilus) . . . . . . . . . . . . . . . . . . . 35111.3.2.4 Cytochrome b5 and Hemoglobin . . . . . . . . . . . . . . . . . . . . . . . . 35111.3.2.5 Myoglobin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35211.3.2.6 Flavoproteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35211.3.2.7 Cytochrome P-450 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353

11.4 Electrode Reaction of Electron Transfer Proteins at Modified Electrodes 35311.4.1 Cytochrome c at Pyridine Derivative-modified Electrodes . . . . . . . . 35411.4.1.1 ER Spectroscopic Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35411.4.1.2 SERRS Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35611.4.1.3 Stability of Bis-(4pyridyl) Disulfide and 4-Mercaptopyridine at Gold

Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357

342 11 Interfacial Properties of Proteins/Spectroelectrochemical Studies

11.4.2 Cytochrome c at ω-Functional Alkanethiol Modified Electrodes . . . . 35711.4.3 Cytochrome c at Iodine-modified Electrodes . . . . . . . . . . . . . . . . . 35811.4.4 Cytochrome c3 at 4,4

′-Bipyridyl and Carboxylic Acid-terminated

Alkanethiol Modified Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . 35811.4.5 Proteins Cast in Lipid Bilayer . . . . . . . . . . . . . . . . . . . . . . . . . . . 359

11.5 Molecular Orientation of Proteins on Solid Substrates. Cytochrome c

Immobilized on Self-assembled Monolayers . . . . . . . . . . . . . . . . . 35911.5.1 IRRAS Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35911.5.2 Conformation of Cytochrome c on Carboxylic Acid-terminated SAM 359

11.6 Electron Transfer Kinetics of Electron Transfer Proteins at ElectrodeSurfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360

11.6.1 Electron Transfer Kinetics of Cytochrome c at CarboxylicAcid-terminated Alkanethiol SAM . . . . . . . . . . . . . . . . . . . . . . . 360

11.6.2 Electron Transfer Kinetics of Cytochromes c and c552 at SilverElectrode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361

11.6.3 Electron Transfer Kinetics of Azurin at Methyl-terminated AlkanethiolSelf-assembled Monolayers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362Recent review article related to this chapter . . . . . . . . . . . . . . . . . 362References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362

343

11.1Introduction

It is well known that there are electrostaticinteractions between redox proteins in thephysiological electron transfer (ET) chains.The initial step in the formation of a pro-tein complex is a nonspecific associationbetween two proteins, followed by a rota-tional diffusion on the molecular surfaceto reach a proper configuration for the ETreaction. Details of the molecular structureof such complexes and intermolecular ETreactions have not been fully understoodby using physiological redox complexesbecause of the complexity of biologicalsystems.

In electrochemical systems, on the otherhand, electrodes act as both electron ac-ceptors and electron donors, and areconsidered a simple model system formimicking a charged interface of thephysiological binding domain. The het-erogeneous ET reactions between elec-trodes and various ET proteins in so-lutions have been extensively studied,as described in previous chapters. Theelectrode reactions of cytochrome c at mer-cury, platinum, silver, and gold electrodeshave been reported to be irreversible.On the other hand, the electrode reac-tions of cytochromes c3 (cyt. c3) have

been found to exhibit a reversible elec-trode reaction at both mercury and solidelectrodes without surface modifiers. Theaddition of 4,4′-bipyridyl leads to a re-versible voltammetric behavior of horsecytochrome c (cyt. c: hereafter cyt. c rep-resents horse heart cytochrome c unlessotherwise stated) on gold and platinumelectrodes. Since then, numerous surfacemodifiers that facilitate direct electrochem-istry of various ET proteins have beenproposed.

In this chapter, electrochemical prop-erties of ET proteins at electrode in-terfaces studied by spectroelectrochem-ical techniques are described. In situspectroelectrochemical techniques at well-defined electrode surfaces are suffi-ciently selective and sensitive to dis-tinguish not only steady state struc-tures and oxidation states of adsorbedspecies but also dynamics of reactants,products, and intermediates at elec-trode surfaces on a monolayer level.The spectroelectrochemical techniquesused in studies of ET proteins in-clude IR reflection-absorption, potential-modulated UV-vis reflectance (electrore-flectance), surface-enhanced Raman scat-tering (SERS) and surface plasmon res-onance, total internal reflection fluo-rescence, (TIRF) and absorbance lineardichroism spectroscopies.


11.2Methodologies

Detailed account of experimental setupsfor spectroelectrochemistry can be foundelsewhere. A brief summary is provided inthis chapter.

11.2.1Surface-enhanced Resonance RamanScattering Spectroscopy

The technique of SERS spectroscopy isbased on the enhancement of Ramansignals for those species adsorbed on sub-microscopically roughened noble metalelectrodes or colloids (particularly silver,gold and copper). The interaction of ad-sorbates with the metal surfaces resultsin a large enhancement of the Ramanscattering cross section. The enhance-ment factor compared to the dissolvedmolecules can be up to six orders ofmagnitude. If the excitation line is in res-onance with an electronic transition of theadsorbed molecules, the molecular reso-nance Raman and the surface-enhancedRaman effects can combine and sensitivityis further increased. The SERS observedunder the resonance condition is calledthe surface-enhanced resonance Ramanscattering (SERRS). The enhancements ofvibrational signals of chromophores makeresonance Raman attractive as a methodof selective analysis of the chromophorewithin a peptide matrix on a monolayer orsubmonolayer level. In the cases of biolog-ical molecules, the advantages of high sen-sitivity and quenching of fluorescence bythe metal surface are attractive. In addition,the small scattering cross section of waterallows the observation of SERS (SERRS) ofthe adsorbates in aqueous solutions.

Cotton and coworkers were the firstto demonstrate that SERS and SERRS

spectroscopies are applicable to obtainingstructural-functional information on bio-logical macromolecules (cyt. c and myo-globin, Mb) adsorbed on a silver elec-trode [1]. In the case of cyt. c adsorbed onthe silver electrode, strong Raman bandsoriginating from the heme chromophorewere observed when the adsorbed cyt. c

was irradiated by a laser excitation wave-length in the vicinity of the electronictransition energy in cyt. c [the Soret band(B band) at 410 nm, for both ferro- andferri-forms and Q bands (α- and β-bands)at 550 nm and 520 nm for the ferro-form],whereas the Raman bands originatingfrom the protein matrix remained too weakto be detected in the resonance Ramanscattering (RRS) signals [2–5].

Argon ion (514.5, 488.0, 476.5, 496.5 nm),krypton ion (674.1, 530.9, 568.2 nm) andtunable dye lasers have commonly beenused as excitation energy sources. TheSERS spectra from gold electrodes are onlyobtainable with excitation energies smallerthan 570 nm. The Raman scattering spec-tra are typically detected in one of twomodes: A photodiode array or CCD cameraenables us to measure a large frequencyrange in a short time, but generally lacksthe spectral resolution required for de-tailed analysis [6]. A monochromatic detec-tion (double monochrometer and photon-counting system) is time-consuming touse because it requires scanning over thefrequency range of interest. However, itoffers much better resolution and can beeasily extended to dual-channel Ramandifference spectroscopy [7]. Degradation ofcyt. c caused by a prolonged exposure tolaser irradiation is prevented by usinga rotating electrode, but photoreductionof ferri-cyt. c has been observed in somecases [8].

The retention of native conformations isconfirmed by the comparison of SERRS

11.2 Methodologies 345

of the adsorbed species with RRS of thesolution species and of the redox potentialsin adsorbed with solution state [7, 9–18]or the enzymatic activity of the adsorbedspecies [19, 20].

11.2.2Electroreflectance (Potential-ModulatedUV-vis Reflectance) Spectroscopy

Potential-modulated UV-vis reflectancespectroscopy, often referred to as electrore-flectance (ER), was originally developed insolid-state physics to characterize surfacesand was applied to studies of the elec-tronic band structure of semiconductors.The ER technique has also been used tocharacterize metal electrode surfaces inthe absence and presence of adsorbates.The reflectivity of metal electrodes is afunction of the surface charge density ofthe electrodes. ER technique has also beenused to investigate electrode reactions oforganic species adsorbed on the electrodesurfaces. Several review articles on ER areavailable [21–24].

Hinnen and coworkers first applied theER technique to elucidate the electrodereaction mechanism (formal potentials

of adsorbed species) of cyt. c and cyt. c3

adsorbed on a gold electrode [25]. Sagaraand coworkers [26] and Feng and cowork-ers [27] showed that the ER techniqueenabled one to measure the ET rates ofthe adsorbed species at electrode surfacesup to 104 s−1, whereas traditional elec-trochemical techniques only enabled oneto measure the ET rates up to severalhundreds times per second. In the ER mea-surements, the electrode surface is cov-ered by the adsorbed layer (monolayer orsubmonolayer coverage) of electrochem-ically active species, which undergoes areversible (rapid) redox reaction and is re-garded as a mirror in the wavelength rangeof interest. When the electrode potential ismodulated by an ac potential in the vicin-ity of the formal potential of the adsorbedspecies, the spectrum of the reflected light(ER spectrum) corresponds to the differ-ence spectrum between the oxidized andreduced forms of the adsorbed species [24,26]. Figure 1 shows the ER spectrum ofthe cyt. c3 adsorbed on the gold electrodein 30-mM phosphate solution at pH 7.0.The structure of the ER spectrum accordswith the difference spectrum of cyt. c3 inthe solution (broken line).

Fig. 1 ER spectrum of cyt. c3 adsorbedon gold electrode at its formal potentialin 30-mM phosphate buffer solution atpH 7.0. Eac = 70 mV, ω = 90 s−1. Thebroken line represents the differencespectrum of cyt. c3 in solution phase.(This is not always true. The ERspectrum of dye molecules directlyadsorbed on electrode surfaces isdifferent from the difference spectrumof the solution species, but the signalintensities against electrode potentialare Nernstian [27]).

300−2

−1

0

1

2

400

λ[nm]

(∆R

/R)1

0−4

Abs

orba

nce

diffe

renc

e[a

.u.]

500 600


When the electrode potential is scannedat the maximum of the ER spectrum(λ = 420 nm), an ER voltammogram isobtained with the peak at its formalpotential. The ET rates of the ad-sorbed species can be calculated fromthe modulation frequency dependence ofthe peak heights (real and imaginaryparts) [28].

11.2.3Infrared Reflection AbsorptionSpectroscopy

The polarization-modulation Infrared Re-flection Absorption Spectroscopy (IRRAS)technique at a high-incident angle of theparallel polarized IR beam is used toinvestigate the structures of cyt. c andsurface modifiers at electrode interfaces.The electrical double-layer structure of thegold electrode covered by cyt. c with asurface modifier at the monolayer levelremains the same when the electrodeis emersed from the electrolyte solutionat controlled potential. In ex situ IR-RAS measurements, an interference byelectrolytes at the emersed electrode canbe minimized. Monolayers of adsorbateswith a dipole moment perpendicular tothe electrode surface can absorb on theorder of 103 of the p-polarized IR radi-ation at high angles of incidence. Theinteraction of the adsorbate with thes-polarized IR radiation, on the otherhand, is negligibly small. Thus, the ab-sorption of radiation by the adsorbateat the electrode surface can be detectedby measuring the intensity of the re-flected p-polarized light with respect tothe reflected s-polarized light (measur-ing intensities of the reflected p- ands-polarized components alternatively byusing the polarization-modulation tech-nique) [29].

11.2.4Absorption Linear Dichroism and TotalInternal Reflection FluorescenceSpectroscopies

It has been known that the configura-tion of protein layers at synthetic mem-branes such as self-assembled monolayers(SAM) at electrode surfaces play a crucialrole in the ET characteristics at pro-tein/membrane interfaces and biologicalfunctions of the protein molecules. Differ-ences in molecular orientation give rise todifferences in ET activities and biofunc-tion of the protein layers. Saavedra andhis coworkers reported a method for deter-mining the dipole orientation distributionof molecules in a thin film using a combi-nation of two techniques: absorption lineardichroism measured in a planar integratedoptical waveguide-attenuated total reflec-tion (IOW-ATR) geometry, and emissionanisotropy, measured in a TIRF geome-try [30, 31].

11.3Redox Behavior of Heme- and Flavoproteinsat Unmodified Electrode Surfaces

11.3.1Gold Electrode

11.3.1.1 Cytochrome cCytochrome c in the solution often exhibitsan irreversible voltammetric response atgold electrodes, with the exception of acarefully prepared gold electrode [32]. Astrong adsorption of cyt. c on an elec-trode surface is considered to block theET reaction of cyt. c in the solution. Hin-nen and coworkers, and Hinnen andNiki measured the formal potential ofhorse heart cyt. c adsorbed on gold, ruthe-nium, and glassy carbon electrodes by

11.3 Redox Behavior of Heme- and Flavoproteins at Unmodified Electrode Surfaces 347

ER voltammograms and found that theformal potentials are around −0.25 V ver-sus NHE (normal hydrogen electrode) in20-mM KClO4, which depend on the sup-porting electrolyte but nearly independentof the electrode materials [25, 33]. Theformal potential of horse heart cyt. c ad-sorbed on a polycrystalline gold electrodeis estimated from the peak potential ofthe ER voltammogram to be −0.18 V in30 mM phosphate buffer at pH 7.0, whichis 0.44 V more negative than that of thenative one [34]. It is interesting to notethat cyt. c adsorbed directly on the goldelectrode takes part in a reversible (rapid)ET process at its formal potential. Thevoltammetric reduction peak of cyt. c in thesolution at the gold electrode agrees withthe formal potential of the adsorbed cyt. c,but no reoxidation peak is observed [33].The oxidation potential of the adsorbedferro-cyt. c is so negative that it cannot me-diate the reoxidation of ferro-cyt. c in thesolution. The shift of the formal potentialis probably due to an unfolding of the pro-tein structure of cyt. c upon the adsorptionon the gold electrode surface.

11.3.1.2 Cytochrome c3Cytochromes c3 are C-type tetra hemepro-teins isolated from the respiratory chainof sulfate reducing bacteria, Desulfovib-rio, which contain about 110 amino acidresidues in the molecule and their molec-ular weights range 13 to 14 kD. This classof proteins is distinguished from otherC-type cytochromes by their unique struc-ture and redox properties. The fifth andsixth axial ligands are imidazoles of thehistidine residue and the hemes are cova-lently bound to a single peptide throughthioether linkages. The formal potentialsof four hemes in cyt. c3 are closely spaced(spread in the range of 100–200 mV) andtheir average is about 0.5 V more negative

than that of mitochondrial cyt. c [35]. Theformal potential of cyt. c3 from D. vulgaris,Miyazaki strain, adsorbed on a polycrys-talline gold electrode is determined to be−0.26 V versus NHE, which is practicallythe same as that of the native species [25].The adsorbed cyt. c3 mediates the ET re-action of cyt. c3 in the solution, whichcan take place in a direct electrode reactionwithout ET mediators or surface modifiers.

11.3.2Silver Electrode

11.3.2.1 Cytochrome cCotton and coworkers first reported thatthe surface-bounded horse heart cyt c isreoxidized at 0.04 V versus NHE by SERRSmeasurements [1], which has a 0.22 Vmore negative potential than the formalpotential of the native one [37]. Niki andcoworkers showed from SERRS measure-ments that the formal potential of cyt. c ad-sorbed on a silver electrode is −0.05 V [38].Hildebrandt and Stockburger have exten-sively studied the conformation and redoxproperties of horse heart cyt. c adsorbedon a silver electrode by using the SERRStechnique [7, 9, 39–42]. Cytochrome c ex-hibits various conformational states uponadsorption on the silver electrode [39–42].The RRS of ferro- (cyt. c2+) and ferri-cyt. c

(cyt. c3+) in an aqueous solution at pH7.0 and SERRS of cyt. c2+ and cyt. c3+ ad-sorbed on the silver electrode at −0.16 Vversus NHE, and SERRS of cyt. c2+ andcyt. c3+ adsorbed on the silver electrode at0.34 V are shown in Fig. 2 [41, 42].

Most of the SERRS bands are well cor-related with the RRS bands obtained in anaqueous solution, in which cyt. c is presentin the six-coordinate low spin (6cLS) con-figuration and the axial ligands are me-thionine and histidine in the nativestate. The marker bands ν4 at 1360 cm−1


1400

14671491 14891503v3

v3

v3v2

v3 v2

v10

v10

v4 v4

v3

v3

−0.5 V

−0.4 V+0.1 V

1588

1589

1591

1584

1582

1500

1500

1633

1633

1620

1490

1491

r

1620r

1624

1636

1583

cyt 2+

5cHS(e)

(c)

(a)

(f)

(d)

(b)

1357

1359 1369

1360 1370

1371 0.0 V

SERR,adsorbedat +0.1 V

SERR,adsorbedat −0.4 V

RR,aqueoussolution

6cLS

6cLS

6cLS 6cLS

6cLS

5cHS 6cLS

∆v/cm−1

1600 1400

∆v/cm−1

1600

Ncyt 3+

N

cyt 3+Icyt 2+

I

cyt 2+II cyt 3+

II

Fig. 2 RRS and SERRS spectra of cyt. c excited at 413 nm.(a) RRS of native cyt. c2+ in aq. solution. (b) RRS of nativecyt. c3+ in aq. solution. (c) SERRS of cyt. cI

2+ measured at −0.4 Vafter adsorption at −0.4 V. (d) SERRS of cyt. cI

3+ measured at+0.1 V after adsorption at −0.4 V. (e) SERRS of cyt. cII

2+measured at −0.5 V after adsorption at +0.1 V. (f) SERRS ofcyt. cII

3+ measured at 0.0 V after adsorption at +0.1 V. Thecyt. cII

3+ and cyt. cI2+ contributions are substracted from (d) and

(e), respectively. Electrolyte: (0.05 M Na2SO4, 0.05 Mtris/cacodylic acid buffer at pH 7.0. Electrode potentials are givenagainst saturated calomel electrode, which is 0.244 V versusNHE. (P. Hildebrandt, M. Stockburger, Biochemistry, 1989, 28,6710–6721, Fig. 2 in page 6712.)

(cyt. c2+) and 1370 cm−1 (cyt. c3+) or ν3at 1491 cm−1 (cyt. c2+) and 1500 cm−1

(cyt. c3+) in Figs. 2(a) and (b) are the6cLS configuration. The SERRS spectrumshown Fig. 2(c) was obtained from cyt. c

adsorbed on the silver electrode at −0.16 V

and is very similar to the RRS spectrumof cyt. c2+. When the electrode poten-tial is shifted from −0.16 V to 0.34 V,the observed SERRS spectrum (Fig. 2d)is very similar to cyt. c3+ of the 6cLS state.However, the SERRS spectrum at 0.34 V


is time-dependent and the spectrum be-comes essentially the same as that shownin Fig. 2(f) after about two hours. Anadsorption-induced partial transformationfrom the 6cLS state to the five-coordinatehigh spin state (5cHS) is observed. TheSERRS spectra of cyt. c on the silver elec-trode adsorbed at 0.34 V are shown inFigs. 2(e) and (f). The spin-state markerbands observed at 0.24 V appear as dou-blets (mixtures of the 6cLS and 5cHSstates), for example, ν3 at 1489 cm−1

(5cHS) and 1503 cm−1 (6cLS) or ν10 at1624 cm−1 (5cHS) and 1636 cm−1 (6cLS)for cyt. c3+ [43]. When the electrode po-tential is shifted from 0.24 V to −0.26 V,cyt. c3+ is reduced to cyt. c2+ and the ν3bands are observed at 1467 cm−1 (5cHS)and 1491 cm−1 (6cLS), respectively, andthe SERRS spectrum shown in Fig. 2(e)approaches that shown in Fig. 2(c) (6cLSstate) after several hours. [41].

Hildebrandt and Stockburger found thatthere are two conformational states ofcyt. c upon adsorption on a silver elec-trode [38–40]. Cytochrome c adsorbed onthe silver electrode at negative potentials(<0.04 V) exhibits the state I (cyt. cI) con-formation and that adsorbed at positiveelectrode potentials (>0.04 V) exhibits thestate II (cyt. cII) conformation. The struc-ture of the RR spectrum of cyt. c inthe solution is fully maintained in thestate I, which must be the same for thewhole cyt. c molecules. In the state II,the spin state of the heme iron is inthe mixed 5cHS and 6cLS conformation.The conformational states I and II areat potential-dependent equilibrium. Themost stable species of ferro-cyt. c is thereduced form of state I (cyt. cI

2+) in theelectrode potential range, more negativethan 0.2 V and that of ferri-cyt. c is theoxidized form of state II (cyt. cII

3+) inthe electrode potential range more positive

than 0.2 V. The stability of cyt. cII dependson the composition and concentration ofthe supporting electrolyte. When the elec-trode potential is maintained at morenegative potentials than −0.34 V, bothcyt. cI

2+ and cyt. cII2+ are considerably

irreversibly denatured (cannot be reox-idized reversibly). The formal potentialof cyt. cI

2+/3+ is estimated to be 0.25 V,which is close to that of the native valuein solution. On the other hand, the for-mal potentials of cyt. cII

2+/3+(5cHS) andcyt. cII

2+/3+ (6cLS) are estimated to be−0.07, which agrees with the results ob-tained by Niki and coworkers [11, 38], and−0.17 V, respectively. At potentials morepositive than the point of zero charge (pzc)of the silver electrode (−0.41 V versusNHE) [44], the surface charge density inthe double layer is negative in sign (moreanions than cations) and a rapid changefrom the state I to II (denaturation or un-folding) is noted. These equilibria stronglydepend on the properties of anions andthe charge distribution at the silver elec-trode/electrolyte interface, suggesting thatthe changes in the structure of the hememoiety is induced by the electrostatic in-teractions between the electrode and cyt. c.At more positive potentials than the pzc,the interaction between the silver electrodeand lysine residues of cyt. c may be weakbecause of the ionic association of excessanions in the double layer and the lysineresidues. The overall reaction scheme ofthe potential induced transformation ofcyt. c adsorbed on the silver electrode isshown in Fig. 3 [41].

In the state II, the axial Met-Fe bondis weakened, leading to a thermal equi-librium between the 6cLS and 5cHSconfigurations. These structural changesare induced by the electrostatic inter-actions between the basic protein andcharged electrode surfaces. It was also


0.0 V

cyt 3+ (acid) 6cHS ?

cyt 3+5cHS cyt 2+5cHS

cyt2+6cLS

cytdenat

pKa ∼4.5 <−0.1 V

<−0.6 V

<−0.6 V>−0.2 V

−0.31 V

+0.02 V

<−0.2 V

cyt 3+6cLSI

II II

cyt 3+6cLS cyt 2+6cLS−0.41 V

II II

I

Fig. 3 Overall reaction scheme of thepotential induced transformation ofcyt. c adsorbed on silver electrode.Electrode potentials are given againstsaturated calomel electrode, which is0.244 V versus NHE. (P. Hildebrandt,M. Stockburger, Biochemistry, 1989, 28,6710–6721, Fig. 15 in page 6719.)

demonstrated that similar conformationaltransition may occur at charged sur-faces other than the electrode surface [7].The docking of cyt. c onto the negativelycharged binding domains of cyt. oxidaseand cyt. reductase induces the formationof the conformational states I and II.

11.3.2.2 Cytochrome c3

Niki and coworkers studied the redoxproperties of cyt. c3 from Desulfovibriovulgaris, Miyazaki and Hildenboroughstrains, and Desulfovibrio desulfuricans,Norway strain, adsorbed on a silverelectrode by the SERRS technique [11].The average formal potential of thehemes in cyts. c3 adsorbed on a silverelectrode is determined by voltammetricmeasurements and is nearly the sameas that in the native state. On the otherhand, the redox potentials of the adsorbedcyts. c3 monitored by the shift of theoxidation state marker band of SERRSare about 100 mV more positive thanthose measured by voltammetry. The redoxtransition of the oxidation state marker

band takes place in the vicinity of theformal potential of the most positiveredox site among the four hemes. Thisis probably due to the fact that cyts. c3are specifically adsorbed on the silverelectrode through the lysine residueswith the most positive formal potentialsurrounding the heme. The heme locatedclosest to the electrode is only responsiveto the excitation wavelength of 514.5 nmand the enhancement diminishes rapidlywith distance from the electrode surface.In addition, the bonds that align inparallel with the surface field are selectivelyenhanced relative to the vibrations that areperpendicular to the surface field. Vermaand coworkers demonstrated that cyt. c3from D. vulgaris, Miyazaki, adsorbed ona silver colloid (citrate-reduced) could bereduced by hydrogen in the presence ofhydrogenase, physiological redox partnerof cyt. c3 [19]. These results showed thatthe enzymatic activity of cyt. c3 is preservedeven in its adsorbed state on a citrate-reduced silver colloid. An adsorption-induced partial transformation from thelow-spin (LS) to the high-spin (HS) state isobserved on the silver colloid. Eng andcoworkers studied RRS and SERRS ofcyt. c3 isolated from D. desulfricans [46]. Acomparison of the protein in a solutionwith that on a silver colloid (citrate-reduced) shows that the native structureof cyt. c3 is retained at the SERRS-activesubstrate. No indication of the HS stateof the adsorbed cyt. c3 was observed. The


difference between two research groups isprobably due to the difference either inpreparation of the silver colloid or in thecytochrome strain [19, 46].

Hobara and coworkers reported thatthe conformation of ferri-cyt. c3 (D. desul-furicans, Norway strain) changes uponadsorption on a bare silver electrode fromthe 6cLS to the mixed 5cHS and 6cLSstate [47]. The fraction of the 5cHS state in-creases upon reduction to the ferro-form.The change in the spin state between theferri- and ferro-cyt. c3 is reversible. The for-mal potentials of cyt. c3 adsorbed on thebare silver electrode was monitored by theshift of the oxidation state marker bandand is −0.12 V in the forward scan (re-duction) and −0.07 V versus NHE in thereverse scan (reoxidation). These resultsare consistent with the formal potential ofthe heme in cyt. c3 with the most positivepotential, which is closest to the silver elec-trode surface and is therefore subject to thelargest enhancement. The ER voltammo-gram of cyt. c3 adsorbed directly on thebare silver electrode surface exhibits sig-nificant hysteresis and a low voltammetricsignal on the reverse scan, probably be-cause of a slow reorganization of the hememoiety of the adsorbed cyt. c3.

11.3.2.3 Cytochrome c552 (Thermusthermophilus)The potential-dependent processes of theET heme protein cyt. c552 from Thermusthermophilus adsorbed on a silver electrodeis studied by using SERRS [48]. In thereduced state at −0.16 V versus NHE,the SERRS spectrum of cyt. c552 is verysimilar to the RRS spectrum in thesolution, suggesting that the adsorbedcyt. c552 retains its native structure in the6cLS state. When the electrode potentialis changed from −0.16 to 0.24 V, theadsorbed cyt. c552 is oxidized to the

ferri-form and exists in the mixed 5cHSand 6cLS state, which is different fromthat of the ferri-form in the solution. Whenthe adsorbed cyt. c552 is rereduced to theferro-form at −0.16 V, no evidence forthe formation of the HS state, that is,the conformational state of the adsorbedcyt. c552, is reversible. The formal potentialof the adsorbed cyt. c552 is determinedby monitoring the intensity of SERRSsignals to be 0.144 V for cyt. c552

2+/3+(6cLS), which is a little more negativethan the value of the solution species(0.228 V) [49]. The discrepancy in theformal potentials between the native andadsorbed cyt. c552 and a non-Nenstianbehavior of the adsorbed cyt. c552 canbe attributed to the coupling of the ETreaction and conformational transitions.

11.3.2.4 Cytochrome b5 and HemoglobinSERS spectra of cyt. b5 and of oxyhe-moglobin (HbO2) adsorbed on a silvercolloid (borohydride reduced) showed dis-tinct changes relative to the solutionRRS [50]. The reduced form of these pro-teins showed SERRS spectra similar to thesolution RRS spectra, although some con-version to FeIII was generally observed. Inthe case of deoxy Hb, the Fe-imidazolestretching band appears to shift from215 to 200 cm−1 in the SERRS spectrum,suggesting some perturbation of the heme-protein linkage. These results indicate thatthe heme attraction to the silver surfaceis facilitated under an oxidizing condition,perhaps via increased surface charge onthe silver surface.

De Groot and Hester demonstrated thatthe SERRS spectrum of oxyhemoglobin(HbO2) gives no indication of the presenceof the HS component (no µ-oxo dimeris present on the silver surface) onsilver colloids (citrate-reduced) [51], incontrast to the results of Smulevich and


Spiro (on the borohydride reduced silvercolloids) [50]. The presence of only LSheme strongly supports a retention ofthe native state on the silver surface. DeGroot and coworkers demonstrated thatthe SERRS spectrum of hemoglobin (Hb)retains its native state on the citrate-reduced silver colloids and concludedthat the difference between the twogroups is due to the preparation of silvercolloids [20].

11.3.2.5 MyoglobinCommercially available Mb does not showwell-defined cyclic voltammograms at anyelectrode. However, purified samples ofboth horse heart and sperm whale met-myoglobins have been found to showstable redox waves at a highly hydrophilicsurface of In2O3 electrodes [52, 53].

Mb adsorbed on silver colloids (boro-hydride reduced) showed the same SERScharacteristics as those of Hb. Met-aquoMb showed facile µ-oxo dimer formation,whereas deoxy Mb did not, but its Fe-imidazole stretching band was lowered to200 cm−1, as in the case of deoxy Hb [50].The correspondence between SERRS ofthe adsorbed Mb (whale) and RRS of thesolution Mb for both oxidized and reducedforms is best if the values for the LS stateMb are used in the comparison [1]. How-ever, Mb in solution is present in the HSaquamet Fe(III) state, which suggests thatMb undergoes a change in spin state uponadsorption on a silver electrode. Anotherpossibility is that Mb retains its HS statebut the frequencies are shifted because of adirect interaction of the heme group withthe electrode. It is likely that the reduc-tion of Fe(III) to Fe(II) occurs at −0.36 Vversus NHE for Mb adsorbed on the sil-ver electrode followed by its reoxidationat 0.04 V, suggesting that the formal po-tential of the adsorbed Mb on the silver

electrode is shifted toward a more neg-ative potential than the formal potentialof the solution Mb at Eo′

(Mb2+/3+) =0.055 ± 0.003 V [50, 51]. The possibility ofa loss of the heme or protein denaturationappears more probable in the case of Mb.

11.3.2.6 FlavoproteinsIt was shown that there is one-to-onecorrespondence of all bands of SERS atsilver colloids (borohydride reduced) andRRS in solution of a riboflavin-bindingproteins, several of which show the samefrequency and relative intensity in thetwo spectra [54]. The SERS spectra ofriboflavin-binding proteins and glucoseoxidase (GOx) show very similar featuresbut differ significantly from the RRSspectrum of the Ag+-complexed flavin.It was shown that the activity of GOxadsorbed on silver colloid remained nearlyintact. On the other hand, Lee andcoworkers reported that the SERRS signalsof the riboflavin-binding proteins on acolloidal silver (EDTA reduced) are shownto arise from free flavin extracted fromthe riboflavin-binding proteins. No spectrawith the flavin incorporated in the proteinsare observed [55]. Thus, the spectra aremost probably due to free flavin as animpurity in the preparations.

Holt and Cotton concluded that a carefulpurification of GOx results in extremelyweak and, under some circumstances,undetectable SERRS spectra at a silverelectrode [56, 57]. GOx shows a gradualdisruption of the protein matrix onthe electrode surface as Flavin adeninedinucleotide (FAD) is released from theGOx molecule onto the electrode surfaceand denaturation of the GOx structureallows direct contact between FAD andthe electrode surface. The flavoproteinssuffer loss of flavin upon contact withSERS active silver surfaces in all cases

11.4 Electrode Reaction of Electron Transfer Proteins at Modified Electrodes 353

reported to date [58]. The flavin dissociatedfrom GOx tends to form a complex withAg+, and this may be a factor leading tothe loss of the prosthetic group.

SERRS of flavin mononucleotide (FMN)and flavodoxin (Fld) from D. vulgaris,Hildenborough strain, was recorded by de-positing on silver colloids (borohydride re-duced) [59]. The SERS signals are sensitiveto the electrode potential and agree verywell with those obtained at a silver elec-trode. The formal potentials of both FMN(Eo′ = −0.245 V versus NHE) and Fld(around −0.22 V) in the solution are closeto those of adsorbed species evaluated fromcyclic voltammograms at a basal planegraphite (BPG) electrode. The SERRS sig-nal of FMN is very weak at a potential inthe vicinity of −0.25 V because the reducedform is SERRS inactive (no resonance ef-fect is involved). The SERRS intensities ofthe oxidized form of Fld are about one-tenth of those of FMN. This is probablydue to the fact that the protein matrix ofFld prevents a direct interaction betweenthe chromophore and a silver electrode. Aslow dissociation of Fld takes place underan irradiation of the laser beam and gener-ates monomeric FMN. The SERRS signalsof Fld, which is adsorbed on BPG and cov-ered by silver colloids, are almost the sameas those observed on the silver electrode.

11.3.2.7 Cytochrome P-450It was shown from the SERRS measure-ment on silver colloids (citrate-reduced)that drug-induced rat liver cyt. P-450 pos-sesses different relative populations of theLS/HS character in their native states.Denaturation from the active form (P-450) to an inactive form (P-420) or lossof the heme on the silver colloid is pre-vented if careful control of pH and colloidpreparation were made [10]. The ability ofa particular colloid to support an active

enzyme is due to a layer of citrate ions,which forms a coating on the colloidsurface, providing a spacer between thesilver surface and the protein, and protect-ing it from silver-induced reactions [18].The change from the LS to the mixedLS/HS state is observed in the SERRSspectra of cyt. P-450 LM2 (rabbit liver)upon the adsorption on silver colloids.Both substrate-induced spin-state changesin the oxidized P-450 and an effect ofthe thiolate ligand on the oxidation statemarker band in the reduced P-450 were ob-served in the SERS spectra of the adsorbedenzyme [12]. These findings indicate thatthe protein structure near the substratebinding site and the coordination by thio-late are not affected by the interaction withthe metal surface. It is assumed that thecatalytic activity of the adsorbed enzyme ispreserved.

The SERRS of P-450 (rat liver) adsorbedonto a phosphatidyl-choline/silver colloid(citrate-reduced) substrate leads to a low tohigh spin-state conversion. This spin-statemarker band shift is ascribed to a stronginteraction of P-450 with the phospholipidcoating [17].

11.4Electrode Reaction of Electron TransferProteins at Modified Electrodes

Since Eddowes and Hill first reportedthat 4,4′-bipyridyl (4,4′-bipy) modified goldand platinum electrodes, which facilitatea reversible (rapid) electrode reaction ofcyt. c, numerous surface modifiers havebeen reported [60]. Among these sur-face modifiers, carboxylic acid–terminatedalkanethiol SAM are most attractivefor mimicking physiological ET part-ners with positively charged bindingsites [61].


11.4.1Cytochrome c at PyridineDerivative-modified Electrodes

11.4.1.1 ER Spectroscopic StudiesHinnen and Niki [33], and Sagara andcoworkers [34] studied the interfacial prop-erties of cyt. c at a gold electrode in thepresence of various surface modifiers by anER technique and found that there are twotypes of surface modifiers. Cytochrome c

is immobilized on a bis(4-pyridyl) disul-fide (4-PySSPy) or 4-mercaptopyridine(PySH) modified gold electrode and theelectrode reaction of cyt. c takes placethrough the film formed by the sur-face modifier. Figure 4 shows the ERvoltammograms (potential dependence atconstant wavelength) and the differentialcapacity curves and voltammograms ofcyt. c adsorbed on bare and 4-PySSPymodified gold electrodes. The same results

−0.5 0

0/V (SCE)

Potential vs. MSE[V]

Cur

rent

µA[c

m−2

](∆

R/R

) ×

103

Cap

acity

µF[c

m−2

]

−1

−10

0 0

(b)

(a)

(c)

10

20

30

10

−1

0

Fig. 4 (a) ER voltammograms at λ = 405 nm and (b) The differentialcapacity curves of cyt. c on bare (broken line) and 4-PySSPy modified(solid line) gold electrodes. (c) Cyclic voltammograms of 4-PySSPymodified (broken line) and cyt. c immobilized on 4-PySSPy modified(solid line) gold electrodes. Electrode potential is given againstHgSO4/Hg in saturated K2SO4, which is 0.65 V versus NHE. (C. Hinnen,K. Niki, J. Electroanal. Chem. 1989, 264, 157–165, Fig. 4 on page 162.)


are obtained with the PySH-modified goldelectrode. The formal potential of cyt. c

immobilized on the surface modifier isslightly more negative than that of thenative one, which is about 450 mV morepositive than that of cyt. c adsorbed on abare gold electrode. Figure 5 shows the ERvoltammogram of cyt. c adsorbed on thegold electrode in the presence of 4,4′-bipy

in the solution. The peak potential, whichcorresponds to the formal potential of cyt. c

coadsorbed with 4,4′-bipy, is 0.03 V ver-sus NHE. It is about 0.23 V more negativethan that of the native one and 0.21 V morepositive than the one adsorbed on the baregold electrode. It is also interesting to notethat the electrode reaction of cyt. c coad-sorbed with 4,4′-bipy is reversible (a rapid

−10

5

00

Potential vs. MSE[V]

Cur

rent

µA[c

m−2

](∆

R/R

) ×

103

Cap

acity

µF[c

m−2

]

0 (V/SCE)

−0.5−1.0(c)

(b)

(a)

AB

10

0

10

−2

−1

0

Fig. 5 (a) ER voltammogram at λ = 405 nm, (b) the differential capacitycurve and (c) cyclic voltammogram of cyt. c at 4,4′-bipy modified electrode.Arrows A and B indicate the formal potentials of native cyt. c and cyt. cadsorbed on a bare gold electrode, respectively. Electrode potential is givenagainst HgSO4/Hg in saturated K2SO4, which is 0.65 V versus NHE.


electrode reaction). The ER voltammetricpeak is shifted to −0.18 V after rinsing theelectrode with the supporting electrolyte,which is identical to the formal potentialof cyt. c on the bare gold electrode. Theseresults suggest that 4,4′-bipy is washedout and cyt. c remained adsorbed on thegold electrode. The formal potential ofcyt. c coadsorbed with 4,4′-bipy on the goldelectrode is strongly dependent on the con-centration of 4,4′-bipy in the solution andvaries from −0.18 V (without 4,4′-bipy) to0.03 V (saturated 4,4′-bipy solution) [62].Models of the interfacial structure of cyt. c

in the presence of surface modifiers areproposed as shown in Fig. 6 [34]. The typeI represents cyt. c adsorbed directly on thegold electrode. The type II represents cyt. c

coadsorbed with a surface modifier such as4,4′-bipy on the gold electrode. The type IIIrepresents cyt. c immobilized on a surfacemodifier the gold electrode.

11.4.1.2 SERRS StudiesTaniguchi and coworkers studied the bind-ing of surface modifiers on gold and silverelectrodes using SERS (SERRS) [63–66].

It was shown that cleavage of the S−Sbonding of 4-PySSPy takes place upon ad-sorption on both gold and silver electrodes,and a stable chemisorbed film is formedthrough sulfur to these electrodes [63, 65].Cytochrome c adsorbed on a gold or silverelectrode is displaced entirely by 4-PySSPyand the electrode reaction of cyt. c in thesolution takes place through the PyS- filmon the electrode surface [63, 64]. Addedpurine partially displaces cyt. c from a sil-ver electrode surface and a mixed adsorbedlayer of purine and cyt. c is formed, atwhich a reversible electrode reaction ofcyt. c takes place [63].

Fan and coworkers studied using SERSthe conformation of pyridine derivativesat a silver electrode in the presence ofcyt. c [67]. When the 4,4′-bipy modifiedsilver electrode is transferred in a cyt. c

solution, 4,4′-bipy partially displaces cyt. c

from the electrode and a mixed adsorbedlayer of 4,4′-bipy and cyt. c is formed. Thespin state of the coadsorbed cyt. c is amixture of the 5cHS and 6cLS states. Theformal potential of the coadsorbed cyt. c

estimated from the shift of the oxidation

−180 mV+30 mV

+260 mV + 260 mV

AuAu

Type I Type II Type III

Au

Fig. 6 Schematic representations of cyt. c at gold electrode in theabsence and in the presence of surface modifier. Type I representscyt. c adsorbed on a bare gold electrode. Type II represents cyt. ccoadsorbed with the surface modifier such as 4,4′-bipy on a goldelectrode. Type III represents cyt. c immobilized on the surfacemodifier on a gold electrode. The large and small circles representcyt. c and surface modifier molecules, respectively. Semicirclerepresents an unfolded cyt. c. Electrode potentials are given againstNHE. (T. Sagara, K. Niwa, A. Sone et al., Langmuir, 1990, 6,254–262, Fig. 12 in page 262.)


state marker band of SERRS is 0.05 Vversus NHE, which is 0.21 V more negativethan that of the native one and 0.20 V morepositive than that of the adsorbed cyt. c

on a bare silver electrode. 2,2′-Bipyridylforms a silver chelate upon adsorption ona silver electrode. When the silver electrodecovered by cyt. c layer is transferred into2,2′-bipy solution, a slow displacement ofcyt. c by 2,2′-bipy takes place and bothmolecules coexist at the electrode surface.

When the silver electrode covered by acyt. c layer is transferred into 4-PySSPysolution, the SERRS signals of cyt. c

disappear because cyt. c is completelydisplaced by 4-PySSPy. However, cyt. c

is subsequently immobilized on top ofthe 4-PySSPy modified electrode and arapid electrode reaction takes place atthe formal potential of the native one.The adsorption of bis-(2-pyridyl) disulfide(2-PySSPy) and cyt. c is competitive. Bis-(2-pyridyl) disulfide coadsorbs with cyt. c onthe silver electrode and the electrochemicalfunction of this adsorbed layer is similarto that of the film formed by 4,4′-bipycoadsorbed with cyt. c.

An adsorbed layer of 1,2-Bis(4-pyridyl)ethylene (4-PyCH) on a silver electrodeis stable and is hard to displace bycyt. c. Cytochrome c is not immobilizedon the PyCH layer, but cyt. c in thesolution undergoes a rapid electrodereaction through this layer.

The SERS signals of the cyt. c/4-PyS/Auelectrode excited by 647.1-nm laser arealmost identical to those of the 4-PyS/Auelectrode, but the SERRS of the cyt. c/4-PyS/Ag excited by a 413.1-nm laser givesrise to the strong enhancement of thecyt. c signals without interference by thesignals of the surface modifier [68, 69].The spin state of cyt. c immobilized on the4-PyS modified silver electrode is in the6cLS state suggesting that cyt. c is in the

native configuration. This fact, togetherwith a similarity of the low frequencyregion of the SERR spectrum to the Ramanresonance spectrum of the protein insolution, provides strong evidence that theheme environment of cyt. c is unperturbedin the adsorbed state on the 4-PyS modifiedsilver electrode. Photodegradation of cyt. c

adsorbed on the 4-PyS-modified electrodeis noticed after a prolonged irradiation ofthe sample at moderately high laser power.

11.4.1.3 Stability of Bis-(4pyridyl) Disulfideand 4-Mercaptopyridine at Gold ElectrodesA structural instability of the 4-PyS modi-fied electrode prepared from the solutionsof 4-PySSPy and 4-PySH has been re-ported [70]. This instability manifests itselfas a decrease in the ability of the modifiedsurfaces to facilitate the electrode reactionof cyt. c with an increase of immersiontime in the precursor solutions. The mod-ified surfaces spontaneously decomposeto yield an adlayer composed largely ofadsorbed atomic and oligomeric sulfur.

11.4.2Cytochrome c at ω-Functional AlkanethiolModified Electrodes

The SERRS spectrum of cyt. c immo-bilized on a carboxylic acid–terminatedalkanethiol modified gold or silver elec-trode reveals that both oxidized and re-duced forms of cyt. c are in the 6cLSstate [68, 69] and its formal potential is0.26 V versus NHE in 10-mM phosphatebuffer solution at pH 7.0 [71, 72]. Theseresults strongly support the retention ofthe native state of cyt. c on the carboxylicacid-terminated alkanethiol modified elec-trodes. Song and coworkers reported theformal potential of cyt. c of the system,cyt. c/HOOC(CH2)nS/Au (n = 5, 10 and


15), to be 0.215 V in 4.4-mM phosphatebuffer solution at pH 7.0 [73].

11.4.3Cytochrome c at Iodine-modified Electrodes

Cytochrome c is found to exhibit aquasireversible voltammetric response atan iodine-modified silver electrode. SERRSspectroscopy indicates that cyt. c immobi-lized on the iodine-modified silver elec-trode is in the 6cLS state [74, 75]. The shiftof the oxidation state marker band wasplotted against the electrode potential andthe formal potential of cyt. c immobilizedon the iodine-modified silver electrode isestimated to be 0.19 V versus NHE, whichis somewhat negative relative to that of thenative state. A similar shift is also seenwhen cyt. c interacts with mitochondrialmembranes (50–60 mV) [76, 77]. It can beconcluded that the SERRS results indicatethat the adsorbed cyt. c is structurally sim-ilar to the native protein in solution. TheSERRS spectrum of cyt. c adsorbed on theiodine-modified gold electrode is differentfrom that on the iodine-modified silverelectrode. The SERRS spectrum of cyt. c

(ox) excited by 550 nm at open-circuit po-tential reveals that cyt. c is present in themixed 5cHS and 6cLS state [78].

11.4.4Cytochrome c3 at 4,4′-Bipyridyl andCarboxylic Acid-terminated AlkanethiolModified Electrodes

SERRS spectroscopic and ER voltam-metric techniques were used to investi-gate the effect of surface modifiers, 11-mercaptoundecanoic acid (11-MUDA) and4,4′-bipy, on the structure and redox prop-erties of cyt. c3 from D. desulfuricans, Nor-way strain, at a silver electrode [47]. The ERvoltammograms of cyt. c3 immobilized on

11-MUDA exhibit a small hysteresis be-tween the forward and backward potentialscans, suggesting that the redox processis reversible and that cyt. c3 retains its na-tive structure. The SERRS spectra of cyt. c3(both oxidized and reduced forms) immo-bilized on 11-MUDA clearly indicate thatthe hemes are in the 6cLS state. The for-mal potential of cyt. c3 immobilized onthe 11-MUDA modified silver electrodewas monitored by the shift of the oxidationstate marker band and is −0.20 V versusNHE, which is about 0.10 V more nega-tive than that observed on a bare silverelectrode [11].

The SERS bands attributable to bothcyt. c3 and 4,4′-bipy suggest that 4,4′-bipyis coadsorbed with cyt. c3 on the silverelectrode surface [47]. The spin state offerri-cyt. c3 is in the mixed 6cLS and5cHS state, which is similar to thatobserved in the absence of 4,4′-bipy. Thetransformation from the oxidized form(the mixed spin state) to the reducedform (6cLS state) is completed at 0.0 V,which is 0.2 V more positive than theformal potential of the most positiveredox site of the native cyt. c3. Theintensity of the SERRS spectrum of thecyt. c3 coadsorbed with 4,4′-bipy in thelow-frequency region (370–430 nm) isconsiderably weak compared with thatof RRS in solution. This result supportsthe idea that 4,4′-bipy has a significanteffect on the heme environment of theadsorbed cyt. c3. The formal potentialof cyt. c3 coadsorbed with 4,4′-bipy ismonitored by the shift of the oxidationstate marker band that becomes morepositive with increasing concentration of4,4′-bipy. The shift of the formal potentialcan be explained in terms of the decreasein the heme exposure to the solvent byshielding the exposed heme edge by 4,4′-bipy molecules.

11.5 Molecular Orientation of Proteins on Solid Substrates. Cytochrome c Immobilized on SAM 359

11.4.5Proteins Cast in Lipid Bilayer

Rusling and his collaborators have ob-tained reversible voltammetric responsesof cyt. P-450cam, hemoglobin, Mb, cyt. c,and chlorella ferredoxin cast in liquid crys-tal films or composites of liquid crystaland Nafion [79–81] on gold electrodes.The structure of the cast films was char-acterized by low angle x-ray diffraction.The structures of Mb and cyt. P-450cam inthe cast film were characterized by UV-vis, Electron Spin Resonance (ESR), andreflectance Fourier Transform Infrared(FT-IR) spectroscopic techniques, and areshown to be similar to the native confor-mations. Visible linear dichroism and ESRanisotropy showed that Mb is specificallyoriented in the static films. The orientationof the heme plane of Mb averaged 60 withrespect to the normal to the film plane, anddistributions are rather broad.

11.5Molecular Orientation of Proteins on SolidSubstrates. Cytochrome c Immobilized onSelf-assembled Monolayers

As a result of the heterogeneous distri-butions of electrical charge and chemicalfunctionalities present on the surface ofmost proteins, the adsorption of proteinsto solid substrates of differing surfaceproperties may produce different molec-ular orientations. The optical thickness ofprotein layers can be evaluated by using el-lipsometry, surface plasmon resonance, orguided wave perturbation. When a chro-mophore of the protein is used as aprobe, polarized spectroscopic techniquesare found to be applicable to measure thetilt angle of the heme moiety in proteinfilms. Macdonald and Smith used SERRS

spectroscopy to measure the conformationof cyt. c adsorbed on a citrate-coated silvercolloid and found that the heme orienta-tion was influenced by the surface coverageof protein [82].

11.5.1IRRAS Studies

The IRRAS spectrum of a gold electrodeemersed from the solution containingcyt. c and 4-PySSPy reveals that the 4-PySSPy molecules are oriented perpendic-ularly to the electrode surface (as PyS-), andthat cyt. c is immobilized on the 4-PySSPylayer [83]. On the other hand, 4,4′-bipy ad-sorbed on a gold electrode gives rise toweak IRRAS signals, suggesting that thedipole moment of the adsorbed 4,4′-bipymolecule is parallel to the electrode sur-face. The amount of cyt. c coadsorbed with4,4′-bipy is estimated from the intensity ofthe IRRAS signals of cyt. c to be about 70%of those observed at the gold electrode withmonolayer coverage. The amount of thecoadsorbed cyt. c with 4,4′-bipy-measuredvoltammetry is estimated to be about 50%of a monolayer.

11.5.2Conformation of Cytochrome c onCarboxylic Acid-terminated SAM

Edmiston and coworkers studied the ori-entation distribution of the heme groupin the protein films by using a com-bination of absorption linear dichroismand fluorescence anisotropy [30, 31, 84].Electrostatic adsorption of the positivelycharged protein to the negatively chargedhead group of archidic acid deposited ona Langmuir-Blodgett film produces a nar-row orientation distribution of the heme(a tilted angle from z-axis is 46 ± 6).


11.6Electron Transfer Kinetics of ElectronTransfer Proteins at Electrode Surfaces

11.6.1Electron Transfer Kinetics of Cytochrome cat Carboxylic Acid-terminated AlkanethiolSAM

The charging current of the electrical dou-ble layer at the electrode interface limitsthe electrode reaction rate measurementsby traditional electrochemical techniques.The potential-modulated ER spectroscopictechnique, on the other hand, involves themeasurement of the faradaic current as adifference in the spectrum between oxi-dized and reduced forms at the electrodesurface generated by an ac modulationof the electrode potential. This techniqueenables the measurement of electrode re-action rates up to approximately 104 s−1

because the effect of the double-layercharging current can be minimized [28].

The ET reaction rates between cyt. c

(horse heart) and a gold (111) electrodethrough the carboxylic acid–terminatedalkanethiol, HOOC(CH2)nSH, SAM aremeasured by the ER technique as afunction of the chain length of alkanethiols(n = 2–11) [71, 72, 85]. The ET reaction

rate constant decreases exponentially withthe chain length as shown in Fig. 7,and the exponential decay factor givenby the Marcus theory is 1.09 ± 0.02 per

methylene group (0.71 ± 0.01 A−1

).The ET rates through the long alka-

nethiol SAM are controlled by the ET ratesthrough the alkanethiol bonds. The ETreaction rates through the short-chain alka-nethiol monolayers, on the other hand, arenearly independent of the chain length.It is assumed that there is a configura-tional rearrangement of cyt. c on the SAMprior to the ET reaction: a thermodynam-ically stable adsorbed structure of cyt. c

(ox) (I), which is formed upon the ad-sorption of cyt. c from the solution to thecarboxylate termini, resulting in a configu-rational rearrangement to cyt. c (ox) (II), atwhich the most efficient ET reaction takesplace. The ET reaction is followed by a sec-ond configurational rearrangement givenby Eq. (3) to form a thermodynamicallystable binding state, cyt. c (red)(I). Therate-controlling step of the ET reactionthrough a short alkanethiol chain is verylikely to be the configurational rearrange-ment of cyt. c on the SAM surface givenby Eq. (1) and the transformation rate con-stant k1 of the forward reaction is estimated

0−3−2

−1

0

1

2

3

4

5

6

7

2 4

Slope = −1.09 per CH2

6 8 10n

log

(kap

p/s

−1)

12 14 16 18 20

Fig. 7 Logarithmic plots of theapparent standard rate constant of cyt. cas a function of a number of methylenegroups of alkanethiol SAM. (•) obtainedby ER technique and () obtained by acimpedance technique. (Z.-Q. Feng,S. Imabayashi, T. Kakiuchi et al., J.Chem. Soc., Faraday Trans. 1997, 93,1367–1370, Fig. 3 on page 1369.)

11.6 Electron Transfer Kinetics of Electron Transfer Proteins at Electrode Surfaces 361

to be 2.6 × 103 s−1.

cyt. c(ox)(I)k1−−−→←−−−k2

cyt. c(ox)(II) (1)

cyt. c(ox)(II) + e− kf−−−→←−−−kb

cyt. c(red)(II)

(2)

cyt. c(red)(II)k3−−−→←−−−k4

cyt. c(red)(I) (3)

The ET rate through longer chain alka-nethiols decreases with increasing ionicstrength or decreasing pH of the solu-tion and is markedly suppressed withincreasing the solution viscosity [83]. Therate-limiting ET step through short alkylchains results from a configurational rear-rangement process of cyt. c from a stablebinding form on the carboxylic acid ter-minus to a configuration that facilitatesthe most efficient ET pathway repre-sented by Eq. (1) [85]. The ET rate on themixed SAM, which was formed in themixed solution of HS(CH2)10COOH andHS(CH2)9CH3, reaches a maximum whenthe solution contains 20% HS(CH2)9CH3and is about six times faster than that onthe SAM of the single component [86].Collinson and coworkers measured theET rate of cyt. c covalently attachedto the carboxylic acid–terminated alka-nethiol (16-mercaptohexadecanoic acid)-modified gold by cyclic voltammetry andfound that the ET rate is similar tothat of cyt. c immobilized electrostaticallyon the same electrode [87]. The cyclicvoltammetric peak broadening of cyt. c

immobilized electrostatically on the car-boxylic acid–terminated alkanethiol SAMwas explained in terms of inhomogene-ity in the formal potential and kinetics ofcyt. c on the SAM [88–90].

Cytochrome c extracted from yeast(iso-1 cyt. c) exhibits a different kinetic

behavior from that of horse heart cyt. c onthe carboxylic acid–terminated alkanethiolmodified gold electrode. The exponentialdecay factor of the ET rates throughlonger alkyl chains (n = 7 and 10) issimilar to that of horse heart cyt. c,but the ET rates are about three ordersof magnitude smaller [90]. The ET ratesthrough the mixed SAM formed in the 1 : 1mixed solution of HS(CH2)10COOH andHS(CH2)7OH (or HS(CH2)8OH) are 2000to 2500 times faster than those throughpure carboxylic acid–terminated SAM. Onthe contrary, the ET rate of horse heartcyt. c through the mixed SAM is five timesfaster than that through the pure carboxylicacid–terminated SAM [90].

Ruzgas and coworkers measured theET reaction rate of cyt. c at a goldelectrode through N -acetylcysteine SAMby the ER technique and proposed amodified approach to evaluate the ET rateconstants [91]. However, the kinetic resultsobtained are unreasonably small.

11.6.2Electron Transfer Kinetics of Cytochromes cand c552 at Silver Electrode

Both traditional electrochemical and ERtechniques cannot provide insight intomolecular processes at the electrode inter-face, which includes conformational trans-formations of the electroactive speciescoupled with the ET reactions as has beenreported for cyt. c [71, 72, 85]. SERRS canprovide both structural and kinetic infor-mation on the ET process of cyt. c atan electrode surface. The time-resolvedSERRS technique was developed to eluci-date the heterogeneous ET rate of cyt. c

with different conformations at a sil-ver electrode surface by monitoring thechange in the SERRS intensity with timeafter applying a potential step [92, 93]. The


time dependence of the SERRS signalsof iso-1 cyt. c adsorbed on the silver elec-trode were measured in a time domainbetween 45 and 175 ms after a potentialstep from 0.16 to +0.29 V versus NHEby using a rotating electrode to avoidphotodegradation. Both the ET reactionand the conformational transition ratesof cyt. c at the silver electrode have beenmeasured to be 10.3 s−1 and 4.3 s−1, re-spectively. The dynamics of cyt. c552 fromThermus thermophilus at the silver elec-trode was also measured by time-resolvedSERRS [94].

11.6.3Electron Transfer Kinetics of Azurin atMethyl-terminated AlkanethiolSelf-assembled Monolayers

Ulstrup and his coworkers showed thatazurin can be assembled directly ongold (111) by adsorption via the surfacedisulfide group [95, 96]. The copper redoxcenter is located opposite to the electrodein this orientation so that the distanceof the ET pathway is too large to exhibita well-defined voltammogram (the ETrate is very slow). On the other hand,the immobilization of azurin is achievedthrough hydrophobic interactions betweenthe hydrophobic area around the copperredox center and methyl-terminus ofalkanethiol SAMs. Gaigalas and coworkersshowed that the blue copper protein,azurin, can be immobilized on the methyl-terminated hexanethiol SAM and azurinexhibits quasi-reversible electron exchangewith gold electrode through the SAM [97].Ulstrup and his coworkers measured thedistance dependence of the ET rate byusing alkanethiols with different chainlengths. The ET rate is almost independentof the chain length up to ca. 9 methyleneunits but follows exponential distance

decay with a decay constant (β) of1.03 ± 0.02 per CH2 unit at longer chainlengths [98].

Recent review article related to this chapter

Extensive review article was published onthe redox reaction of cyt. c through SAMon electrodes [99].

References

1. T. M. Cotton, S. G. Schultz, R. P.Van Duyne, J. Am. Chem. Soc. 1980, 102,7960–7962.

2. T. M. Cotton in Spectroscopy of Surfaces (Eds.:R. J. H. Clark, R. E. Hester), Wiley & Sons,New York, 1988, pp. 91–153.

3. T. M. Cotton, J.-H. Kim, G. D. Chumanov,J. Raman Spectrosc. 1991, 22, 729–742.

4. T. M. Cotton, J.-H. Kim, R. E. Holt, Adv.Biophys. Chem. 1992, 2, 115–147.(JAI Press).

5. P. Hildebrandt in Cytochrome c: Multidis-ciplinary Approach (Eds.: R. A. Scott, A.G. Mauk), University Science Books, Sousal-ito, 1996, pp. 285–314.

6. P. Hildebrandt, G. J. Pielak, R. J. P. William,Eur. J. Biochem. 1991, 201, 211–216.

7. P. Hildebrandt, M. Stockburger, Biochem-istry 1989, 28, 6722–6728.

8. P. Hildebrandt, K. A. Macor, R. C. Czernus-zewicz, J. Raman Spectrosc. 1988, 19, 65–69.

9. P. Hildebrandt, M. Stockbsurger, J. Phys.Chem. 1986, 90, 6017–6024.

10. K. Kelley, B. N. Rospendowski, W. E. Smithet al., FEBS Lett. 1987, 222, 120–124.

11. K. Niki, Y. Kawasaki, Y. Kimura et al., Lang-muir 1987, 3, 982–986.

12. P. Hildebrandt, R. Greinert, A. Stier et al.,FEBS Lett. 1988, 227, 76–80.

13. C. R. Wolf, J. S. Miles, S. Seilman et al., Bio-chemistry 1988, 27, 1597–1603.

14. T. M. Cotton, V. L. Schlegel, R. E. J. Holtet al., Proc. SPIE-Int. Soc. Opt. Eng. 1989,1,055, 263–270.

15. T. M. Cotton, B. N. Rospendowski, V. L.Schlegel et al., Proc. SPIE-Int. Soc. Opt. Eng.1991, 1,403, 93–96.

16. S. Hashimoto, R. Nakajima, I. Yamazakiet al., FEBS Lett. 1989, 248, 205–209.

11.6 Electron Transfer Kinetics of Electron Transfer Proteins at Electrode Surfaces 363

17. B. N. Rospendowski, V. L. Schlegel, R. E.Holt et al., in Charge and Field Ef-fects in Biosystems-2 (Eds.: A. M. J. Allen,S. F. Clearry, F. M. Hawkridge), PlenumPress, New York, 1989, pp. 43–58.

18. B. N. Rospendowski, K. Kelley, C. R.Wolf et al., J. Am. Chem. Soc. 1991, 113,1217–1225.

19. A. L. Verma, K. Kimura, T. Yagi et al., Chem.Phys. Lett. 1989, 159, 189–192.

20. J. de Groot, R. E. Hester, S. Kaminaka et al.,J. Phys. Chem. 1988, 92, 2044–2048.

21. D. M. Kolb in Spectroelectrochemistry: Theoryand Practice (Ed.: R. J. Gale), Plenum, NewYork, 1988, pp. 78–188.

22. W. Plieth in Spectroscopic and DiffractionTechniques in Interfacial Electrochemistry(Eds.: C. Gutierrez, C. Melendres), KluwerAcademic Publishing, The Netherlands,1990, pp. 223–260.

23. W. Plieth, W. Kozlowski, T. Twomey in Ad-sorption of Molecules at Metal Electrodes (Eds.:J. Lipkowski, P. N. Ross), VCH, New York,1992, pp. 239–284.

24. T. Sagara, Recent Res. Dev. Phys. Chem. 1998,2, 159–173.

25. C. Hinnen, R. Parsons, K. Niki, J. Electroanal.Chem. 1983, 147, 329–337.

26. T. Sagara, S. Igarashi, H. Sato et al., Lang-muir 1991, 7, 1005–1012.

27. T. Sagara, J. Iizuka, K. Niki, Langmuir 1992,8, 1018–1025.

28. Q. Feng, T. Sagara, K. Niki, Anal. Chem.1995, 67, 3564–3570.

29. O. Hofmann, K. Doblhofer, H. Gerischer,J. Electroanal. Chem. 1984, 161, 337–344.

30. P. L. Edmiston, L. L. Wood, J. E. Lee et al.,J. Phys. Chem. 1996, 100, 775–784.

31. P. L. Edmiston, J. E. Lee, S.-S. Cheng et al.,J. Am. Chem. Soc. 1997, 119, 560–570.

32. E. F. Bowden, F. M. Hawkridge, H. N. Blo-unt, J. Electroanal. Chem. 1984, 161, 355–376.

33. C. Hinnen, K. Niki, J. Electroanal. Chem.1989, 264, 157–165.

34. T. Sagara, K. Niwa, A. Sone et al., Langmuir1990, 6, 254–262.

35. K. Niki, Y. Kobayashi, H. Matsuda, J. Electro-anal. Chem. 1984, 168, 275–286.

36. F. M. Hawkridge, T. Kuwana, Anal. Chem.1973, 45, 1021–1027.

37. W. R. Heineman, B. J. Norris, J. F. Goelz,Anal. Chem. 1975, 47, 70–84.

38. K. Niki, Y. Kawasaki, C. Hinnen et al. inFrontiers of Bioinorganic Chemistry (Ed.:

A. V. Xavier), VCH, Weinheim, 1983, pp.622–630.

39. P. Hildebrandt, M. Stockburger in Spec-troscopy of Biological Molecules: Proceedings ofthe First European Conference on Spectroscopyof Biological Molecules (Eds.: A. J. P. Alix,L. Bernard, M. Manfait), John Wiley & Sons,Chichester, 1985, pp. 25–30.

40. P. Hildebrandt, M. Stockburger in RamanSpectroscopy: Sixty Years On VibrationalSpectra and Structure, (Eds.: H. D. Bist,J. R. Durig, J. F. Sullivan), Elsevier SciencePublishing, The Netherlands, 1989,pp. 443–446, Vol. 17A.

41. P. Hildebrandt, M. Stockburger, Biochem-istry 1989, 28, 6710–6721.

42. P. Hildebrandt, J. Mol. Struct. 1991, 242,379–395.

43. N. Parthasarathi, C. Hansen, S. Yamaguchiet al., J. Am. Chem. Soc. 1987, 109,3865–3871.

44. G. Valette, A. Hamelin, J. Electroanal. Chem.1973, 45, 301–319.

45. P. Hildebrandt, T. Heimburg, D. Marsh, Eur.Biophys. J. 1990, 18, 193–201.

46. L. H. Eng, V. Schlegel, D.-L. Wang et al.,Langmuir 1996, 12, 3055–3059.

47. D. Hobara, K. Niki, T. M. Cotton, Biospec-troscopy 1998, 4, 161–170.

48. S. Lecomte, H. Wackerbarth, P. Hildebrandtet al., J. Raman Spectrosc. 1998, 26, 687–692.

49. K. Hon-Nami, T. Oshima, J.Biochem. (Tokyo)1977, 82, 769–776.

50. G. Smulevich, T. G. Spiro, J. Phys. Chem.1985, 89, 5168–5173.

51. J. de Groot, R. E. Hester, J. Phys. Chem. 1987,91, 1693–1696.

52. I. Taniguchi, K. Watanabe, M. Tomi-naga et al., J. Electroanal. Chem. 1992, 333,331–338.

53. M. Tominaga, T. Kumagai, S. Takita, Chem.Lett. 1993, 1771–1774.

54. R. A. Copeland, S. P. A. Fodor, T. G. Spiro,J. Am. Chem. Soc. 1984, 106, 3872–3874.

55. N.-S. Lee, Y.-Z. Hsieh, M. D. Morris et al.,J. Am. Chem. Soc. 1987, 109, 1358–1363.

56. R. E. Holt, T. M. Cotton, J. Am. Chem. Soc.1987, 109, 1841–1843.

57. R. E. Holt, T. M. Cotton in Redox Chemistryand Interfacial Behavior of Biological Molecules(Eds.: G. Dryhurst, K. Niki), Plenum Press,New York, 1988, pp. 217–228.

58. R. E. Holt, T. M. Cotton, J. Am. Chem. Soc.1989, 111, 2815–2821.


59. V. Brabec, K. Niki, Chem. Lett. 1988,1445–1448.

60. M. A. Eddowes, H. A. O. Hill, J. Chem. Soc.Chem. Commun. 1977, 771, 772.

61. M. J. Tarlov, E. F. Bowden, J. Am. Chem. Soc.1991, 113, 1847–1849.

62. T. Sagara, H. Murakami, S. Igarashi et al.,Langmuir 1991, 7, 3190–3196.

63. I. Taniguchi, M. Iseki, T. Eto et al., Bioelec-trochem. Bioenerg. 1984, 13, 373–383.

64. I. Taniguchi, M. Iseki, H. Yamaguchi et al.,J. Electroanal. Chem. 1984, 175, 341–348.

65. I. Taniguchi, M. Iseki, H. Yamaguchi et al.,J. Electroanal. Chem. 1985, 186, 299–307.

66. I. Taniguchi in Redox Chemistry and Inter-facial Behavior of Biological Molecules (Eds.:G. Dryhurst, K. Niki), Plenum Publishing,New York, 1988, pp. 113–123.

67. K.-J. Fan, I. Satake, K. Ueda et al. inRedox Chemistry and Interfacial Behaviorof Biological Molecules (Eds.: G. Dryhurst,K. Niki), Plenum Publishing, New York,1988, pp. 125–138.

68. D. Hobara, K. Niki, T. M. Cotton, DENKIKAGAKU 1993, 61, 776, 777.

69. D. Hobara, K. Niki, C.-I. Zhou et al., ColloidsSurf. A: Physicochem. Eng. Aspects 1994, 93,241–250.

70. B. D. Lamp, D. Hobara, M. D. Porter et al.,Langmiur 1997, 13, 736–741.

71. Z.-Q. Feng, S. Imabayashi, T. Kakiuchi et al.,J. Electroanal. Chem. 1995, 394, 149–154.

72. Z.-Q. Feng, S. Imabayashi, T. Kakiuchi et al.,J. Chem. Soc., Faraday Trans. 1997, 93,1367–1370.

73. S. Song, R. A. Clark, E. F. Bowden et al.,J. Phys. Chem. 1993, 97, 6564–6572.

74. C.-I. Zhou, T. M. Cotton, X.-G. Qu et al. inRedox Mechanisms and Interfacial Proper-ties of Molecules of Biological Importance(Eds.: F. A. Schultz, I. Taniguchi), The Elec-trochemical Society of Pennington, NewJersey, 1993, pp. 63–74.

75. T.-H. Lu, X.-J. Yu, S.-J. Dong et al., J. Electro-anal. Chem. 1994, 369, 79–86.

76. P. L. Dutton, D. F. Wilson, C.-P. Lee, Bio-chemistry 1970, 9, 5077–5082.

77. J. Vanderkooi, M. Erecinska, B. Chance,Arch. Biochem. Biophys. 1973, 157, 531–540.

78. D. Hobara, K. Niki, G. Chumanov et al., un-published results.

79. J. F. Rusling, Interface 1997, 6(4), 26–31.80. J. F. Rusling, Acc. Chem. Res. 1998, 31,

363–369.81. Y. M. Lvov, Z.-Q. Lu, J. B. Schenkman et al.,

J. Am. Chem. Soc. 1998, 120, 4073–4080.82. I. D. G. Macdonald, W. E. Smith, Langmuir

1996, 12, 706–713.83. K. Niwa, M. Furukawa, K. Niki, J. Electroanal.

Chem. 1988, 245, 275–285.84. J. E. Lee, S. S. Saavedra, Langmuir 1996, 12,

4025–4032.85. A. Avila, W. Gregory, K. Niki et al., J. Phys.

Chem. B 2000, 104, 2759–2766.86. S. Arnold, Z.-Q. Feng, T. Kakiuchi et al.,

J. Electroanal. Chem. 1997, 438, 91–97.87. M. Collinson, E. F. Bowden, M. J. Tarlov,

Langmuir 1992, 8, 1247–1250.88. T. M. Nahir, E. F. Bowden, J. Electroanal.

Chem. 1996, 410, 9–13.89. R. A. Clark, E. F. Bowden, Langmuir 1997,

13, 559–565.90. A. E. Kasmi, J. M. Wallace, E. F. Bow-

den et al., J. Am. Chem. Soc. 1998, 120,225–226.

91. T. Ruzgas, L. Wong, A. K. Gaigalas et al.,Langmuir 1998, 14, 7298–7305.

92. S. Lecomte, H. Wackerbarth, T. Soulimaneet al., J. Am. Chem. Soc. 1998, 120,7381–7382.

93. H. Wackerbarth, U. Klar, W. Gunther et al.,Appl. Spectrosc. 1999, 53, 283–291.

94. S. Lecomte, P. Hildebrandt, T. Soulimane,J. Phys. Chem. B 1999, 103, 10 053–10 064.

95. A. K. Gaigalas, G. Niaura, J. Colloid InterfaceSci. 1997, 193, 60–70.

96. Q. Chi, J.-D. Zhang, E. P. Friis, J. E. T. Ander-sen, J. Ulstrup, Electrochem. Commun. 1999,I, 91–96.

97. Q. Chi, J.-D. Zhang, J. U. Nealsen, E. P. Friis,Ib. Chorkendorff, G. W. Canters, J. E. T.Andersen, J. Ulstrup, J. Am. Chem. Soc.,2000, 122, 4047–4055.

98. Q. Chi, J.-D. Zhang, J. E. T. Andersen, J. Uls-trup, J. Phys. Chem. B, 2001, 105, 4669–4679.

99. M. Fedurco, Coordin. Chem. Rev., 2000, 209,263–331.

365

12Electrochemical Analysis ofNucleic Acids

Emil Palecek, Miroslav Fojta, Frantisek Jelen, and Vladimır VetterlInstitute of Biophysics, Brno, Czech Republic

12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36912.1.1 History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36912.1.2 Nucleic Acid Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370

12.2 Electrochemical Behavior of NA Components . . . . . . . . . . . . . . . . 37012.2.1 Adsorption/Desorption Behavior . . . . . . . . . . . . . . . . . . . . . . . . 37012.2.1.1 Mercury Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37012.2.1.1.1 Oligonucleotides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37412.2.1.2 Solid Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37512.2.2 Reduction and Oxidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37512.2.3 Microanalysis of Nucleic Acid Components by Stripping Techniques 37612.2.3.1 Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37612.2.3.2 Reactions of Pyrimidine and Purine Bases with the Electrode Mercury 37612.2.3.3 Unusual Bases and Nucleosides . . . . . . . . . . . . . . . . . . . . . . . . . 37712.2.3.3.1 5-Ribosyluracil (pseudouridine) . . . . . . . . . . . . . . . . . . . . . . . . . 37712.2.3.3.2 Methylated Adenines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37812.2.3.3.3 5-Fluorouracil (5-FU) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37812.2.3.4 Sparingly Soluble Compounds of Nucleic Acid Components with

Copper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379

12.3 Adsorption/Desorption Behavior of NAs . . . . . . . . . . . . . . . . . . . 37912.3.1 Mercury Dropping Electrode . . . . . . . . . . . . . . . . . . . . . . . . . . . 37912.3.1.1 Adsorption of Double-stranded (Native) DNA . . . . . . . . . . . . . . . . 38112.3.1.2 Adsorption of Single-stranded (Denatured) DNA . . . . . . . . . . . . . . 38112.3.2 Adsorption Kinetics at Mercury Dropping and Hanging Electrodes . 38112.3.3 Electrochemical Impedance Spectroscopy (EIS) . . . . . . . . . . . . . . . 38212.3.3.1 Other Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38412.3.4 Adsorption of NAs on Other Electrodes . . . . . . . . . . . . . . . . . . . . 38412.3.4.1 DNA Adsorption to Charged Lipid Membranes . . . . . . . . . . . . . . . 385

366 12 Electrochemical Analysis of Nucleic Acids

12.4 Reduction and Oxidation of NAs on Different Electrodes . . . . . . . . 38512.4.1 Mercury Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38512.4.1.1 Reduction of Adenine and Cytosine Residues . . . . . . . . . . . . . . . . 38512.4.1.2 Anodic Signal of Guanine Residues . . . . . . . . . . . . . . . . . . . . . . 38612.4.2 Carbon Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38712.4.3 Other Solid Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38912.4.4 Analysis of NAs by Different Electrochemical Techniques . . . . . . . . 389

12.5 Relations Between Structures and Electrochemical Responses of DNA 391

12.6 DNA Structure on Electrode Surfaces . . . . . . . . . . . . . . . . . . . . . 39312.6.1 Dependence of the dsDNA Signals at the HMDE on Potential Scanning

Direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39412.6.2 Opening of the DNA Double Helix Around −1.2 V (Region U) . . . . 39512.6.2.1 Opening of dsDNA at Acid pHs in a Wider Potential Range T . . . . . 398

12.7 Interactions of NAs with Small Molecules . . . . . . . . . . . . . . . . . . 39912.7.1 Reversible (Noncovalent) Interactions . . . . . . . . . . . . . . . . . . . . . 39912.7.1.1 Inorganic Cations and Simple Metal Complexes . . . . . . . . . . . . . . 40012.7.1.2 Organic Metal Chelates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40012.7.1.3 Other Noncovalent DNA Binders . . . . . . . . . . . . . . . . . . . . . . . . 40112.7.2 Covalent Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40312.7.2.1 Electroactive Markers of NAs . . . . . . . . . . . . . . . . . . . . . . . . . . . 40312.7.2.2 Other Nucleic Acid Modifications . . . . . . . . . . . . . . . . . . . . . . . . 405

12.8 DNA Conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40612.8.1 Application of Electrodes in DNA Conductivity Studies . . . . . . . . . 406

12.9 Analytical Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40712.9.1 Sensors for DNA Hybridization . . . . . . . . . . . . . . . . . . . . . . . . . 40712.9.1.1 Immobilization of DNA on the Electrode . . . . . . . . . . . . . . . . . . . 40812.9.1.2 Detection of the Hybridization Event . . . . . . . . . . . . . . . . . . . . . . 40812.9.1.3 Redox Indicators Covalently Bound to DNA . . . . . . . . . . . . . . . . . 41012.9.1.4 Indicator-free Detection Systems. Intrinsic Electroactivity of DNA . . 41012.9.1.5 Changes in Interfacial Properties and DNA Conductivity . . . . . . . . 41012.9.1.6 Blocking and Interfacing the Transducer . . . . . . . . . . . . . . . . . . . 41112.9.1.7 Electrocatalytic Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41212.9.1.8 Detection of Point Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . 41212.9.2 Sensors for DNA Damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41212.9.2.1 Detection of DNA Strand Breaks . . . . . . . . . . . . . . . . . . . . . . . . 41312.9.2.2 Damage to DNA Bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41312.9.2.3 Detection of Damaging Agents Specifically Interacting with DNA . . 41412.9.2.4 DNA Cleavage Controlled by Electrochemical Reactions . . . . . . . . . 41412.9.3 Other Determinations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415

12 Electrochemical Analysis of Nucleic Acids 367

12.9.3.1 Determination of ssDNA in an Excess of dsDNA . . . . . . . . . . . . . . 41512.9.3.2 Determination of RNA Traces in DNA Solutions . . . . . . . . . . . . . . 41512.9.3.3 Determination of Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 416

12.10 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417Addendum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419

369

12.1Introduction

12.1.1History

To our knowledge, the first paper dealingwith electrochemical analysis of nucleicacids (NAs) was published by Berg in1957 [1]. He used a supporting electrolytecontaining cobalt ions to determine pro-teins in RNA and DNA samples, andunder the given conditions, found that NAswere polarographically inactive species.The ability of DNA and RNA to yield reduc-tion and oxidation signals upon interactionwith electrodes was discovered by Palecekin 1958 and in the years [2–5]. It wasthe so-called oscillographic polarographyat controlled a.c. (OP) [6] that proved to bebetter suited for the analysis of NAs thanthe d.c. polarography popular at that time.Using OP, it was shown that in addition tothe reduction of adenine (A) (whose reduc-tion at dropping mercury electrode (DME)in strongly acidic medium was describedalready in 1945 [7]); cytosine (C) was alsoreducible at a DME at neutral pH [2, 8],and that guanine (G) produced a specificanodic signal, later explained by the oxida-tion of the DNA reduction product formedat highly negative potentials [2, 3, 9]. A, C,and G residues yielded their signals, not

only in nucleosides and nucleotides butalso in RNA and DNA [2–5]; and NA sig-nals were strongly influenced by the NAstructures [4, 5, 10].

In the following decade, electrochemi-cal techniques, and particularly differentialpulse polarography (DPP), produced earlyevidence of DNA premelting and polymor-phy of the DNA double helix (reviewedin [11]). The results of conventional elec-trochemical analysis of NAs with mercuryor carbon electrodes immersed in the NAsolution during the electrochemical mea-surements were thoroughly reviewed [4, 5,10–18]. In the last 15 years, new trendsappeared oriented toward the immobiliza-tion of the NA at the electrode surface. Itwas found that DNA and RNA could beeasily immobilized at mercury and carbonelectrodes, simply by immersing the elec-trode in a drop of NA solution for a shorttime. As a result of strong adsorption ofDNA or RNA a stable layer was formed atthe electrode surface; the electrode wasthen washed and electrochemical mea-surements were performed in solutionsnot containing any NA. By this method,the volume of the analyte was reduced by2 to 3 orders of magnitude. The techniquewas called adsorptive transfer strippingvoltammetry (AdTSV). Development ofsensors (detectors) for DNA hybridizationand for DNA damage applying adsorptive


or covalent immobilization of DNA at theelectrode (transducer) to create recognitionlayers at their surfaces, became popular inthe years that followed.

12.1.2Nucleic Acid Samples

Currently, an electrochemist can analyzea number of different kinds of highlypurified synthetic and natural DNA andRNA molecules. In addition to highly poly-merized chromosomal DNA samples thatwere available even 50 years ago, betterdefined plasmid DNAs and viral DNAsand RNAs have become available (Fig. 1).Shorter or longer DNA fragments can beprepared by cleavage of the viral and plas-mid DNAs by restriction endonucleases.Amplification of pieces of any DNA canbe performed by the Polymerase ChainReaction (PCR) from minute amounts ofNAs. Biosynthetic polynucleotides withmonotonous or random nucleotide se-quences, as well as synthetic oligonu-cleotides, whose nucleotide sequence canbe programmed at the automatic synthe-sizer, are commercially available. RNAoligonucleotides are, however, more ex-pensive than oligodeoxyribonucleotides(ODNs). Oligonucleotides with chemi-cally modified bases and/or the sugar-phosphate backbone, as well as end-labeledDNAs, can be prepared or purchased.

In this paper, we briefly summarizethe basic electrochemical properties ofNAs and their components, and surveythe recent trends in the electrochemicalanalysis of NAs, including labeling of NAswith electroactive markers, covalent andnoncovalent immobilization of NA at theelectrode surfaces, and development ofsensors for DNA hybridization and DNAdamage. We wish to show that electrodesare important tools, useful in biochemical

analysis, which may help solve somespecific biological problems and contributeto better understanding of NA interactionswith electrically charged surfaces.

12.2Electrochemical Behavior of NAComponents

12.2.1Adsorption/Desorption Behavior

12.2.1.1 Mercury ElectrodesThe encounter of molecules at biologicalinterfaces, such as cell membranes andnuclear matrix, is the initial step inbiomolecular processes and a prerequisitefor the manifestation of biological effectsof biopolymers in living cells [19–21]. Ithas been known that electric fields havingmagnitudes equivalent to those existingat a charged cell surface/biological fluidinterface affect the conformation of DNAin solution [22–24]. As a rough modelof a biological surface/biological fluidinterface, an electrolyte solution/electrodeinterface can be employed to study theinterfacial behavior of NAs. Mercuryelectrodes, which have an atomicallysmooth surface, the charge of which caneasily be changed and controlled in arelatively wide range, have proved to bewell suited for such experiments [25].

NAs, as well as NA bases, nucleosides,and nucleotides, are strongly adsorbed atthe mercury electrodes [26–63]. Amongbases, G is most strongly adsorbed at theseelectrodes [28]. The adsorption can be fol-lowed by measurement of the impedanceand/or differential capacitance of the elec-trode double layer [26, 27, 31, 32, 34–39,41, 42, 47]. In 1965, it was found byone of us (VV) [26, 27] that NA basespossess an extraordinarily high ability of

12.2 Electrochemical Behavior of NA Components 371

GENOMIC (chromosomal)ds, molecularly polydisperse, nucleotide sequence usually unknown

Plasmid or viralmonodisperse, nucleotide sequence known

Biosynthetic polynucleotidespolydisperse, simple repeated sequence motifs or homopolymers

Synthetic oligonucleotidesmonodisperse, programmed nucleotide sequence

PCR products of various lengths and defined sequences

(b1) Restriction fragments

DNA molecules

usually 3-4 kbpm.w. around 2 × 108

ds

ds

rel sc oc lin

ss

ATATATATATATATATATATATT A T A T A T A T A T A T A T A T A T A T A T T T T T T T T T T T T T T T T T T T T T T

ss CCCCCCCCCCCCCCCCC

GCGCATTTCCGG

ss

CGCGATATCGCGCGCGTATAGCGC

ds

TTTTTTTTTTTTTTTTTTT

average m.w. 105 − 106

usual lengths 10-20 nucleotides

Chemically modified oligonucleotides, including end-labeling with fluorescein, −SH groups, etc.

AAAAAAAAAAAAAAAAAAA

(a)

(b)

(c)

(d)

(e)

Fig. 1 Nucleic acid samples available for electrochemical experiments. (a,b) naturally occurring DNAs; (b1) dsDNA fragments of defined lengths andnucleotide sequences can be conveniently prepared by cleavage with restrictionendonucleases; (c) NAs (both DNA and RNA) synthesized by enzymes; (d) fullysynthetic DNAs and RNAs of limited lengths; (e) PCR, can amplify the desiredDNA segment from template DNA. ds, double-stranded; ss, single-stranded,kbp, kilobase pairs. Covalently closed circles of sc, supercoiled and rel, relaxedDNA. oc, open circular DNA (containing at least one interruption of thesugar-phosphate backbone); lin, linear DNA. See text for more details.

self-association at the electrode surfaceand undergo a two-dimensional (2-D) con-densation forming a monomolecular layer(self-assembled monolayer (SAM), a com-pact film). By this high condensationability, NA bases differ from most of theother purine and pyrimidine derivatives,

which currently do not occur in NAs.The two-dimensional condensation wasalso observed in some of the halogen-,aza-, and methyl derivatives of commonNA bases [28–31] and in most of thenucleosides [41, 43–46] and nucleotides[47–49, 64] commonly occurring in NAs.


The formation of a compact film at theelectrode is characterized by the appear-ance of a well-defined capacitance ‘‘pit’’on the capacitance-potential (C-E) curves(Fig. 2).

The orientation of bases and nucleosidesin the compact film [50–52, 54–56], the ef-fect of ions of the solvent and substituentsof bases on the film formation [28–31, 34,50–52, 54, 55, 65, 66], the energy of theinteraction between bases in the compactfilm [34, 54, 55, 67, 68], and the kineticsof the two-dimensional condensation ofbases at the electrode surface [43, 46, 54,57–59, 65, 69–83] were investigated. Inneutral bases the capacitance pit is usuallyobserved near the potential of the elec-trocapillary maximum (potential of zerocharge, p.z.c.), with the exception of baseswith a large electric permanent dipole

moment [84] like C, which forms the pitat negative potentials [27, 47, 57, 85–87].The halogen ions can induce a new poten-tial region of condensation with C [86] andwith A [59].

We have studied the effect of brominesubstituents on the two-dimensional con-densation of C, U, and uridine. Thepotential of maximum adsorption, andthus the potential of the capacitance pit,depends on the mutual competition be-tween the electrostatic and nonelectrostaticadsorption forces and the forces that re-pulse the adsorbed molecules from theelectrode surface. In neutral bases and nu-cleosides, the nonelectrostatic adsorptionusually prevails and the pit appears nearthe p.z.c. With C at pH 5.0, the electrostaticadsorption on the negatively charged elec-trode surface via the positive charge on N -3

E1 E0 E2

Cs

EE −max

a

b

c

Fig. 2 Dependence of differential capacitance Cs on the electrode potential E.(a) capacitance curve of the background electrolyte measured at hanging mercurydrop electrode (HMDE) in the absence of surface active substances,(b) capacitance curve of electrolyte solution with surface active substance,(c) capacitance curve characteristic for the two-dimensional condensation ofadsorbed molecules on the electrode. E−

max, potential of cathodic tensammetricmaximum. Between the potentials E1 and E2, a compact film is formed on theelectrode. (a) background electrolyte: 2 M NaCl, pH 7; (b) 20 mM thymine (T),pH 7, 40 C; (c) 20 mM T, pH 7, 17 C. [V. Drazan and V. Vetterl, unpublished].


and positive end of the permanent dipolemoment prevails over the nonelectrostaticadsorption, and the pit appears at very neg-ative potentials. The bromine substituentincreases the polarizability of C and thusincreases the nonelectrostatic dispersionforces between the C molecule and theelectrode surface. With 5-Br-C the non-electrostatic adsorption is stronger thanthe electrostatic one and the pit occurs nearthe p.z.c [31, 88]. The bromine substituentusually decreases the solubility of basesor nucleosides and increases the stan-dard free energy of adsorption. Thus thebromine substituent can either decreasethe tendency to two-dimensional conden-sation due to the lower solubility of thesubstituted derivative as was observed with5-bromouracil [31], or can increase the 2D-condensation ability due to the higher ad-sorption energy of the substituted deriva-tive as observed with 5-Br-uridine.

We have shown that the capacitance pitcorresponding to the compact layer formedat the electrode surface is observed in the Csolutions in 0.5 M NaCl at pH 5.5, but notat pH 7.0 [31, 85, 86]. The two-dimensionalcondensation of C molecules at themercury electrode surface is thus obviouslysupported by protonation of a part of theC molecules. Ab initio quantum chemicalcalculations have proved that protonationsignificantly increases the stabilizationenergy of both stacked and hydrogen-bonded C dimers, which may support thetwo-dimensional condensation [89].

From the temperature dependence ofthe capacitance pit or electrocapillary mea-surements, the surface concentration ofthe adsorbed molecules and the area Arequired for an adsorbed molecule at theelectrode surface can be determined [34,43–45, 50–52, 55]. Using data obtainedfor the crystal structure of bases thearea that would be occupied by one

adsorbed molecule in different surfaceorientations can be evaluated and com-pared to the experimentally determinedarea A. From these calculations, it hasbeen concluded that at low surface con-centrations (the so-called dilute adsorptionregion) the adsorbed bases lie flat at theelectrode surface. In the compact layer, theadsorbed bases seem to adopt a perpendic-ular surface orientation [50, 60]. Similarreorientation from flat to perpendicularposition at higher surface concentrationshas been observed with several nucle-osides and nucleotides [48, 51, 52, 54,61]. Some nucleosides, such as adeno-sine, can probably adopt two differentperpendicular orientations at the electrodesurface [29, 41, 42, 50, 62, 63]. De Levieand Wandlowski [90] have suggested thatthe compact film of U is a planar arrayof hydrogen-bonded molecules, similar tothat found in the solid state. The adsorp-tion energy and the energy of the lateralinteractions between bases in the compactlayer can be estimated, either from thecourse of adsorption isotherms [27, 28, 50,60–62] or from the temperature depen-dence of the capacitance pit [34, 54, 55,91]. For A in a neutral solvent this energywas −4.7 kJ mol−1 [55].

Direct observation of an ordered phaseof NA bases on solid electrodes bytechniques, such as scanning tunnelingmicroscopy (STM) and atomic force mi-croscopy (AFM), may help determine theorientation of the molecules in the com-pact film [83, 92–98]. These techniqueswere also recently applied to the surfaceof mercury [99–101]. It was found thatcationic detergent benzalkonium chloride(BAC), used for DNA spreading on micain scanning force microscopy, forms acondensed film at the mercury electrodesurface. The corresponding pit on C-Ecurves resembled the pits of bases and


nucleosides [102]. A similar cationic de-tergent cetyl-dimethyl-benzylammoniumchloride formed capacitance pits on C-Ecurves as well [103].

12.2.1.1.1 Oligonucleotides Adsorptionof oligonucleotides on mercury elec-trodes was studied by a.c. polarog-raphy [48, 49, 104, 105], surface ten-sion measurements [106], DPP, and ad-sorptive stripping cyclic voltammetry(CV) [107]. The self-complementary de-camer d(CCAGGCCTGG) produced ca-thodic and anodic signals. By measur-ing the anodic peak G (resulting fromG residues), it was possible to detectthe decamer at subnanomolar concen-trations [107]. With oligoriboadenylates, itwas found that the shorter molecules(dinucleotides and trinucleotides) were ad-sorbed with all A residues oriented flat atthe electrode surface and with all sugaror sugar-phosphate residues close to thesurface. The tetranucleotides and longeroligomers seemed to be adsorbed with amaximum of three A rings directly an-chored to the electrode surface [106].

Impedance measurements of peptidenucleic acid (PNA) and DNA decamers(GTAGATCACT and complementary se-quences) were performed at HMDE [108](Fig. 3). From the calculated degree ofthe electrode coverage by adsorbed PNAand DNA decamers it was found thatwith DNA decamers the degree of themaximum coverage was reached near thepotential (potentials are given against thesaturated calomel electrode, if not statedotherwise) of −0.6 V, where the elec-trode surface is neutral and the decamersare adsorbed via their hydrophobic bases.The dependence of the electrode coveragedegree by PNA molecules on the poten-tial showed two maxima, at −1.2 V and−0.5 V. At −1.2 V, the electrode had anegative charge, and the PNA decamerswere not electrostatically repulsed fromthe electrode as DNA decamers (in contrastto DNA the PNA backbone is electricallyneutral [109, 110]). The second maximumat −0.5 V corresponded to the PNA ad-sorption by its hydrophobic bases [108].Prolonged exposure of PNA to highly neg-ative potentials did not result in PNA

DNA decamer

DNA

Electrolyte

Histone

PNA

PNA decamer

scDNA

p.z.c.

1.2

0.8

C/1

0−7F

0.4

−1.5 −1.0E vs SCE

[V]

−0.5

Histone

Fig. 3 AC impedance response of PNAand DNA decamers (nucleotidesequence GTAGATCACT), sc plasmidDNA (scDNA) and a histone at HMDE.Concentrations: DNA and PNAdecamers 10 µg ml−1, pUC19 DNA110 µg ml−1, histone 30 µg ml−1; Thesamples were adsorbed underconditions securing full coverage of theelectrode at neutral pH and roomtemperature. (Adapted from M. Fojta,V. Vetterl, M. Tomschik et al., Biophys. J.1997, 72, 2285–2293.)


desorption, whereas almost all of the DNAwas removed from the surface at thesepotentials.

12.2.1.2 Solid ElectrodesRecently, the use of solid electrodes [43,50, 64, 72, 93, 94, 96, 111–127], insteadof mercury, has opened new perspectivesfor the understanding of the factors thatgovern the formation of self-organizedmonolayers at the electrochemical in-terface and widened the experimentalfield, namely by STM and AFM imag-ing [64, 83, 93–96, 112, 113, 116–118, 126,127]. Single crystal gold electrodes havebeen increasingly used in electrochemicalstudies so that reliable results are ob-tained on a well-defined surface structure,which now allows a correct assessment ofthe surface specificity of interfacial pro-cesses [124]. It was shown that U anduridine may form condensed layers, notonly at the mercury electrode interface [31,43, 46, 71] but also at a gold single crys-tal interface [72, 112]. Gold electrodes havebeen selected because they display a broaddouble layer region and have been the ob-ject of many investigations [114, 119–122].Recently it was found that C and cytidineform condensed layers at the single crystalgold electrode as well [116, 128]. Both PNAand DNA oligomers displayed a strongadsorption onto the carbon electrode [129].Potential-controlled release of DNA fromsolid electrodes has been utilized in thedevelopment of gene carriers for futuregene therapies [130, 131].

12.2.2Reduction and Oxidation

Among simple NA constituents, A and C(and their nucleosides and nucleotides)can be reduced at mercury electrodes

in aqueous solutions. Both bases pro-duce well-developed pH-dependent polaro-graphic waves involving base protonation(reviewed in [132, 133]). G is reduced atthe mercury electrode at highly nega-tive potentials close to background dis-charge, yielding 7,8-dihydrogen G [134,135] (Fig. 4). This reaction is chemicallyreversible. Oxidation of the G reductionproduct can be observed in cyclic orin anodic stripping modes if the elec-trode is shortly exposed to highly negativepotentials prior to scanning to positivepotentials [136–138]. The mechanism ofelectroreduction of inosine (nucleoside ofhypoxanthine, a deamination product ofA) was recently studied using elimina-tion polarography [139]. Reduction of Tand U at a mercury electrode can be ob-served in nonaqueous media [140, 141].At carbon electrodes, purine bases pro-duce well-defined oxidation peaks within awide pH range (0–12.5) [142, 143]. Purinenucleosides and nucleotides are oxidizedat potentials more positive than theparent bases [144]. Signals correspond-ing to the oxidation of purine bases,nucleotides, and nucleotides have alsobeen obtained using chemically modi-fied carbon electrodes [145, 146] (for moredetails see Sect. 12.4.3). Recently, Caiand coworkers [147] proposed a methodfor trace A determination using anelectrochemically/chemically modified (inalkaline sodium nitrate solution) car-bon paste electrode (CPE). Pyrimidinesare considered to be electroinactive oncarbon electrodes; however, Oliveira-Brettand Matysik recently reported [148] spe-cific anodic peaks observed in solu-tions of T and C bases (but not theirnucleosides). Sugar components of nu-cleotides can be oxidized at copper elec-trodes [149].


NN

N

OR

H

H

NN

N

H

OR

H O

NN

N

N

R

H N

H

H

NN

N

N

R

O

H

NH

H

TT-A

N

NN

N

N

R

HH

NN

O

H

OR

CH3

NN

O

H

OR

CH3 H N

NN

N

N

R

H A

Minor groove

Major groove

CC-G

G

Fig. 4 Scheme of electroactive sites in NA bases. Adenine (A) andcytosine (C) are reducible at mercury electrodes. Guanine (G) undergoesa chemically reversible reduction at the mercury electrodes, yielding ananodic peak G due to oxidation of its reduction product. A and G can beoxidized at carbon electrodes. Watson–Crick base pairs TA and CG areshown without the reduction sites involved in hydrogen bonding. Arrowsindicate schematically electrode reduction or oxidation: ( ),reduction at mercury electrodes; ( ), oxidation at carbon electrodes;

( ), chemically reversible reduction/oxidation of G at mercuryelectrodes; for more details see the text.

12.2.3Microanalysis of Nucleic Acid Componentsby Stripping Techniques

12.2.3.1 PrinciplesCathodic stripping voltammetry (CSV)can be used to determine low concen-trations of substances that form spar-ingly soluble compounds with mercuryof the electrodes [150–152]. The processof deposition (first step) can be gener-ally expressed as production of mercuryions and formation of a sparingly sol-uble film. The second step (stripping)takes place in a negative-going potentialscan, reducing the deposited salt backinto the solution, stripping the film. The

relations between the peak potential atequilibrium on one side, and the pH, an-ion concentration and solubility on theother, were published [152–154]. In gen-eral, the peak potential becomes morenegative with increasing pH, increasinganion concentration and decreasing solu-bility of the compound, and more positivewith the higher association constant ofthe acid.

12.2.3.2 Reactions of Pyrimidine andPurine Bases with the Electrode MercuryIt was first shown by Revenda [155] thatsome inorganic anions form sparinglysoluble salts with mercury to produceanodic polarographic waves. Later, it


was found that similar anodic currentswere produced by sulfur-containing or-ganic substances [156], and by organiccompounds such as derivatives of bar-bituric acid, including U [157, 158].Depolarization of the mercury electrodewith all DNA bases was demonstrated byPalecek [2, 9], using the OP method. AllNA bases produced in an alkaline mediumindentation on the OP curves dE/dt againstE, probably caused by formation of com-pounds with the electrode mercury. At thetime, these findings were little exploitedfor analytical purposes because of a lowsensitivity of the available methods. At mil-limolar concentrations, it was, however,possible to detect changes in the indenta-tions of T, U, 5-hydroxymethyluracil andother bases because of UV-irradiation oftheir solutions [159]. Later, we showedthat pyrimidine and purine bases com-monly occurring in NAs (U, T, C, A, G)and a number of their derivatives yieldedsimilar anodic waves using DC polarogra-phy (DCP) and normal pulse-polarography(NPP), and peaks using DPP [160, 161].These signals appeared close to 0 V be-cause of the formation of sparingly solublecompounds with mercury. The behaviorof NA bases was studied at various con-centrations of the bases, pH values, pulseamplitudes, and so on, by NPP and DPP.The DPP behavior of C was similar tothat of U; detection limit of C was about5 µM. Differences between the behaviorof U and T were large at higher concen-trations at which the electrode was fullycovered [160]. DPP rendered it possible todetermine A and G at concentrations of 1to 10 µM [161].

Earlier, it was assumed that the useof CSV in organic analysis is lim-ited mainly to sulfur-containing sub-stances [162–164]. Using the HMDE, itwas demonstrated that pyrimidine and

purine bases, and their derivatives, canbe deposited on a mercury electrode andthen stripped out by scanning to nega-tive potentials [160, 161, 165, 166]. UsingDPCSV, purine and pyrimidine baseswere analyzed at very low concentrations(Table 1). The sensitivity of the determina-tion depended on several parameters, suchas the potential and time of deposition,electrode size and scan rate. Pyrimidinebases could be determined at concentra-tions down to 10 to 100 nM, and purinebases at concentrations as low as 1 to10 nM (the limit of A detection was about2 nM at 6 min waiting time) [161, 166].Nucleosides and nucleotides derived frompyrimidine bases were inactive and did notsubstantially interfere with the determina-tion of bases [160].

For CSV analysis of purine and pyrim-idine bases, we also applied the transferstep (Sect. 12.4.4.) to find out whether thelayers of mercury compounds were suffi-ciently stable to remain on the electrodeafter its washing and transfer to anothersolution. (This method was called cathodictransfer stripping voltammetry CTSV). Wefound that A yielded a well-developed peakwhose height only slightly decreased as aresult of medium exchange [167]. In con-trast to A, C produced no CTSV signalwhile showing a peak in conventional CSV.This result suggested that by using CTSVit would be possible to analyze some mix-tures of bases that differ in their ability toundergo transfer of their deposited layers.

12.2.3.3 Unusual Bases and Nucleosides

12.2.3.3.1 5-Ribosyluracil (pseudouridine)5-Ribosyluracil (pseudouridine) is oneof the family of unusual nucleosidescontained in transfer ribonucleic acid(tRNA), and its determination in urine


Tab. 1 The ability of pyrimidine and purine derivatives to form sparingly soluble compoundswith the electrode mercury

Pyrimidine derivatives Purine derivatives

Cytosine + [160,166] Adenine + [161,166]Thymine + [160,166] Guanine + [161,166]Uracil + [160,166] Xanthine + [161]Cytidine − [160] Hypoxanthine + [161]Thymidine − [160] Adenosine + [161]Uridine − [160] Guanosine + [161]Pseudouridine + [169] Xanthosine + [161]5-Bromouracil + [166,176] Inosine + [161]5-Chlorouracil + [166,176] 1-Methyladenine + [170]5-Fluorouracil + [166,176] 3-Methyladenine + [170]5-Acetyluracil + [176] 6-Methyladenine + [170]5-Formyluracil + [176] 2-Aminopurine + [166]5-Nitrouracil + [176] 8-Oxyadenine + [166]5-Azauracil + [176] 6-Benzyladenine + [166]6-Azauracil + [176] Uric acid + [170]6-Chlorouracil + [166]6-Methylthymine + [176]2-Thiouracil + [166]Orotic acid + [166]

Sparingly soluble compound with the electrode mercury is produced (+), not produced (−).

and blood is used in cancer diagnos-tics [168]. Pseudouridine as well as uri-dine and deoxyuridine were investigatedby DCP, DPP, and NPP [169]. Amongthese U nucleosides, only pseudouridineyielded anodic polarographic currents inalkaline medium, whereas uridine and de-oxyuridine were inactive under the sameconditions [169]. The inactivity of thesesubstances was caused by substitution ofthe base residue at position N1 by thesugar, while in the case of the uridine iso-mer, pseudouridine and the ribose residuewere bound to C5 of the pyrimidinering. Using DPCSV in connection witha HMDE, it was shown that pseudouridinecould be determined in concentrationsdown to 20 to 60 nM.

12.2.3.3.2 Methylated Adenines Adenineand its methylated derivatives (1-CH3-Ade,

N -CH3-Ade and N ,N ′-diCH3-Ade) werestudied in alkaline solutions using sev-eral electrochemical methods, includingCSV [170]. Some of these substances rep-resent rare NA constituents. It was shownthat the 6-aminogroup of A is the mer-cury binding site, and CSV could be usedfor the determination of methylated As.The detection limit for N -CH3-Ade wascomparable to that for A [170]. Methylsubstituents affected the adsorption andtwo-dimensional condensation of A [30].

12.2.3.3.3 5-Fluorouracil (5-FU) 5-Fluo-rouracil (5-FU) is frequently used in thetreatment of a wide variety of carcinomas.It is well known that many anticancerdrugs are toxic to the treated organismand that the actual dosage must be care-fully controlled. It was demonstrated that5-FU reacts with the mercury electrode

12.3 Adsorption/Desorption Behavior of NAs 379

[forming a sparingly soluble compound]and can be determined by CSV at lowconcentrations [166]. Later, it was foundthat in a slightly alkaline medium 5-FUcould be monitored [171] at pH 10 (boratebuffer) or at pH 7.6 (borax with 0.1 MKNO3) [172], at concentrations down to50 nM. CSV was also applied to determine5-FU in human serum with the previ-ous solvent extraction procedure to preventthe interference of other electrochemicallyactive substances. The limit of detectionwas 5 µM and mean recovery 43% [171].Recently, CSV and DPCSV studies of de-termination of 5-FU in urine [173] and inblood serum [174] were published. UsingDPCSV, and following a simple pretreat-ment procedure with trichloroacetic acid,5-FU was determined with almost 100%recovery [175, 176]. It was also shown thata flow-injection system with mercury elec-trode as a detector could provide sensitiveCSV determination of 5-FU [177, 178].

12.2.3.4 Sparingly Soluble Compounds ofNucleic Acid Components with CopperIn the presence of A and adeno-sine, the copper(II)/copper(Hg) couplesplit to the copper(II)/copper(I) andcopper(I)/copper(Hg) couples [179, 180].Sparingly soluble compounds of copper(I)with A and its ribonucleoside were ac-cumulated on the electrode, either byreduction of the Cu(II) ions or by oxida-tion of the copper amalgam electrode. Thecopper(I) A deposit was stripped eithercathodically or anodically. The strippingpeaks obtained for copper complexes hadhigher detection limits, but appeared over awider range of pH and at more negative po-tentials than the peaks related to mercurycompounds [161]. It was shown that in ad-dition to A, other purine bases, such as G,hypoxanthine, xanthine, and their nucleo-sides (guanosine and inosine) [181–183],

as well as pyrimidine base C and cyti-dine [66] produced similar effects in thepresence of copper; these compoundswere detectable at low concentrations bystripping techniques.

12.3Adsorption/Desorption Behavior of NAs

On interacting with electrodes, DNA andRNA are usually strongly adsorbed, un-dergoing charge transfer reactions in theiradsorbed states and sudden desorption innarrow potential ranges, resulting in ten-sammetric signals [10, 12, 14]. Both thefaradaic and tensammetric signals can pro-vide information, not only about the kindand about the concentration of the ana-lyzed NA but also about the changes in theNA structure, and about the interaction ofthe NAs with various compounds. Studiesof DNA and RNA adsorption/desorptionproperties are, therefore, of importance fora better understanding of different types ofinteractions of these biomacromoleculeswith electrodes. Miller measured the dif-ferential capacitance of the DME in 1961and showed that ss and dsDNA as well asRNA are adsorbed in a potential range ofabout 0 to −1.1 V [37, 38].

12.3.1Mercury Dropping Electrode

The adsorption of NAs can be followedby a number of physical and electrochem-ical methods such as voltammetry [184],a.c. polarography [31, 40, 185–187], mea-surements of the surface tension [40, 188,189], and of the impedance and/or differ-ential capacitance of the electrode doublelayer [31, 37, 38, 88]. Differential capac-itance C of the electrode double layeris a sensitive indicator of the adsorp-tion. When the NAs are adsorbed at the


Base residues

Distorted double-stranded regions

Sugar-phosphate backbone

(screened with counterions) Electrostatic attraction(unscreened phosphates)

I

3

2 1

el

0a

b

cp.z.c.

(b)

(a)

−1.5 −1.0 −0.5

−1.6 −1.1 −0.6E[V]

E[V]

ssss

ds ds

3

1

a

c

3

2 1

10 µΩ

d1

b13

Fig. 5 Adsorption/desorption behavior of double- or single-stranded DNAat mercury electrodes (at weakly alkaline pH’s). (a) scheme of ACpolarographic curves of DNA obtained with DME. (a) dsDNA, low ionicstrength; (b) dsDNA, moderate ionic strength; (c) ssDNA; (el) backgroundelectrolyte; p.z.c. See text for more details. (b) AC admittance curvesmeasured at HMDE. (a, b) ssDNA; (c, d) dsDNA; (a, c) initial potential−0.5 V, potential scanned from positive to negative values; (b, d) initialpotential −1.7 V, potential scanned from negative to positive values[F. Jelen and P. Belusa, unpublished]; see text for more details.

electrode surface, they remove from thesurface the molecules and ions of thesolvent and lower the value of the differ-ential capacitance of the electrode doublelayer, because the solvent has usuallymuch higher dielectric permittivity thanthe NAs. The potential of maximum ad-sorption of NAs is usually close to the p.z.c.

At desorption potentials the adsorption-desorption (tensammetric) peaks appearon the C-E curves, resulting from suddenchanges of the surface charges and/or sur-face coverage within a narrow potentialrange.

It was shown [184, 186, 190–197] that alltypes of residues–purine and pyrimidine


bases, ribose and/or deoxyribose, andphosphoric acid participated in the adsorp-tion of NAs. The extent of participation ofthese components in the adsorption ofNAs depended on pH and ionic strengthof the solvent and on the electric chargeof the electrode surface given by the elec-trode potential E (Fig. 5). At neutral pH andmoderate ionic strength, at the electrodepotentials close to the p.z.c., hydropho-bic bases were adsorbed most strongly. Atlow ionic strengths and on the positivelycharged electrode surface the NAs were ad-sorbed electrostatically by their negativelycharged phosphate groups.

12.3.1.1 Adsorption of Double-stranded(Native) DNAAt moderate ionic strengths (µ ∼ 0.3 M),the phosphate charges are screened anddsDNA is adsorbed as an electroneutralcompound [186, 190], predominantly viathe sugar-phosphate backbone. The ten-sammetric (adsorption/desorption) peakappears on a.c. polarograms and/or C-Ecurves at about −1.1 V and is denotedas peak 1 in Fig. 5. If dsDNA containssome ss or distorted regions in whichthe hydrophobic bases can come intocontact with the electrode surface, an-other tensammetric peak 2 appears onC-E curves around a potential of −1.3 V.This peak corresponds to the desorption(reorientation) of distorted regions ofdsDNA that are adsorbed firmly via bases(Fig. 5).

At low ionic strengths (µ < 0.1 M)dsDNA is electrostatically adsorbed viacharged phosphate groups at potentialscorresponding to the positively chargedelectrode surface. Unscreened phosphatecharges are repulsed [186, 198] fromthe negatively charged surfaces, and ds-DNA can be only weakly adsorbed viabases (coming predominantly from the

molecule ends in shorter intact dsDNAs),if available. The corresponding tensam-metric (reorientation) peak occurs aroundthe p.z.c. and is denoted as peak 0 (Fig. 5).

12.3.1.2 Adsorption of Single-stranded(Denatured) DNAIn the range of ionic strengths 0.01to 1.0 M single-stranded (ss) DNA isadsorbed as an apparently electroneutralmolecule via sugar-phosphate backboneand (more strongly) via bases [10, 184,186, 190–197]. The tensammetric peakat about −1.1 V corresponding to thedesorption of sugar-phosphate backbonewas denoted as peak 1. The peak around−1.4 V corresponding to the desorption ofbases was denoted as peak 3 (Fig. 5).

12.3.2Adsorption Kinetics at Mercury Droppingand Hanging Electrodes

Adsorption of DNA is diffusion-control-led [37]. If insufficient time is allowedfor adsorption, the surface will only bepartly covered and the amount adsorbedt (surface concentration) at a time t willbe given by the Koryta expression [199]

t = kc√

Dt (1)

where k = 0.745 for the DME and k = 1.13for the HMDE [200, 201], c is the bulkconcentration, and D is the diffusion co-efficient. The differential capacitance ofthe electrode double layer C depends ont and thus on concentrations c, if mea-surements are made at constant t . At lowc Eq. (1) is valid and the dependence ofthe capacity C on c is a straight line. Athigh concentration c the surface concen-tration t reaches the saturated value s

and the capacity C reaches almost constantvalue Cs. The intercept of the linear plot of


capacity C against concentration c for di-lute solutions with the almost constant Cs

values at high concentrations gives the lim-iting concentration cs at which the surfacebecomes fully covered at the given time t .From the values cs the saturated surfaceconcentration s (fully covered surface)and the area A occupied per one nucleotideon the electrode surface A = M/Ns

(M is the mean molecular weight of anucleotide and N is the Avogadro number,N = 6.023 × 1023g−1 mol−1) can be cal-culated [37]. Miller [37, 38] found that withherring sperm denatured DNA adsorbedat DME (dropping time t = 14.5 s) the sat-urated value of the surface concentrations was reached at cs = 75 µg ml−1. FromEq. (1) it was calculated that s = 6 ×10−5mg cm−2. The corresponding area Aoccupied per one nucleotide on the elec-trode surface was A = 93 A

2[37]. Later,

Janik and Sommer [191] found about85 A

2per nucleotide for poly(U) contain-

ing about 100 nucleotides; the area Adecreased with the increasing length of

the polynucleotide chain down to 75 A2

(with polynucleotides of about 1500 nu-

cleotides). Coverage of about 110 A2

permolecule of tRNA (about 75 nucleotides)was observed [202]. These results suggestthat in longer ss NA chains the ratioof three-dimensional loops/adsorbed seg-ment trains may increase, allowing smallerfraction of the nucleotide residues to comeinto direct contact with the surface.

We studied the time dependence of theimpedance of the electrode double layeraround the potentials of the tensammet-ric peaks and around the potential ofmaximum adsorption of native and de-natured calf thymus DNA (100 µg ml−1

in 0.3 M NaCl with 0.05 M Na2HPO4,pH 8.6). At potentials of maximum ad-sorption (around −0.7 V) the differential

capacitance of the double layer decreasedfaster with denatured DNA than with na-tive DNA as a result of the higher diffusioncoefficient of denatured DNA (as shown inprevious studies) [10, 203]. The final valueof the differential capacity 11 µF cm−2

was reached with denatured DNA afterabout 4 s, with native DNA after 30 s [88].Polyadenylic acid [poly(A)] can assume (de-pending on pH and ionic strength) eitherss or ds form [186, 187, 204–206]. Themore flexible ss poly(A) has a higher dif-fusion coefficient than the double-helicalone [206] so that the diffusion transport tothe electrode of ss poly(A) is faster thenthe adsorption of the ds one [194].

12.3.3Electrochemical Impedance Spectroscopy(EIS)

Useful information about the adsorptionkinetics, mobility of the adsorbed polynu-cleotide segments, and mechanism ofelectrode processes can be obtained bymeasurement of the frequency depen-dence of the impedance of the electrodedouble layer (EIS) [31, 88, 207–209]. Ifthe adsorption/desorption process is slowwith respect to the period of the a.c.potential used for the impedance measure-ment, the measured capacitance valuesdecrease with increasing frequency (dis-persion of the capacity). The frequencyeffect is most remarkable around the po-tentials of adsorption/desorption peaks.With more flexible ss polynucleotides, thefrequency effect is larger than with themore rigid ds ones [210].

It was shown [91, 208, 211, 212] that formonomeric surface-active substances, dif-fusion to the surface is the slowest stepand, therefore, the rate-determining fac-tor in the adsorption mechanism at the


surface. Consequently, at high bulk con-centration of the surface-active material,when the diffusion transport is fast as perFick’s First law, the height of the tensam-metric peak depends less on frequencythan at low bulk concentration (slow dif-fusion transport). On the other hand, thebulk concentration of a polymeric sub-stance has sometimes no influence on thefrequency dependence of tensammetricpeaks. Complete lack of frequency de-pendence of the tensammetric peaks onthe bulk concentration indicates that nosignificant diffusion of whole polymericmolecules takes place during the varia-tion in potential of the imposed alternatingvoltage. Hence, the frequency effect mustdepend on the rate of migration of ad-sorbable segments from the surface phaseinto the adsorption layer and vice versa.Migration of segments can occur onlyin conjunction with rearrangement of thewhole polymeric molecule; consequently,the temperature dependence of the disper-sion is expected to be far stronger than inthe case of unhindered diffusion of smallmolecules [211, 212].

The lack of any dispersion of the ca-pacity in the peaks of polylysine [212]shows that the adsorption/desorption pro-cess that gives rise to the peaks is notdiffusion controlled, in spite of the fact thatthe diffusion of the polymeric moleculesto the surface is a relatively slow process,requiring times on the order of seconds toreach equilibrium at polylysine concentra-tions employed. This effect is explained byassuming that there exists at the interfacea region of high polymer concentration,known as the surface phase, and that thepeaks result from the adsorption and des-orption of segments of the polymer fromthe surface phase. This process does notneed to depend on the diffusion of wholepolymer molecules from the solution, and

is fast enough at elevated temperatures tokeep pace with a signal of low audio fre-quency [208, 212]. Similar behavior mightbe expected with other biopolymers, in-cluding DNA.

From the frequency dependence of theimpedance of the electrode double layerrepresented in a complex impedance plot(the imaginary component Z′′ is plottedagainst the real component Z′, Cole-Cole,or Nyquist plot), the electric equivalentcircuit of the electrode covered with anadsorbed layer can be determined. Fromsuch a circuit, the physical parametersof the layer, such as the effective thick-ness and the degree of molecular order ofthe layer, can also be evaluated [213–221].The complex plane impedance plots forDNA exhibited arc shapes, from whichthe apparent resistance R2 of the layerand the solution resistance R1 were deter-mined. The resistance R2 represents thedielectric losses of the capacitance of theelectrode double layer, that is, the energylost as heat, which arises because of thefriction of charged DNA segments forcedto move in a viscous solvent by an a.c.electric field [88]. It was found that withDNA solutions at the potentials where des-orption takes place, the dielectric lossesare higher (i.e. the resistance R2 is lower)than at the potentials of maximum ad-sorption. Desorption of denatured ssDNAis accompanied by higher dielectric lossesthan desorption of native dsDNA. Withdenatured DNA desorption of more firmlybound bases is accompanied by higherdielectric losses than desorption of thesugar-phosphate backbone [88]. In the po-tential region of desorption the resistivitycomponent of the impedance increasesmore steeply in ssDNA than in dsDNA.This can be explained by a higher flexibil-ity of the denatured DNA compared withthe dsDNA resulting in higher dielectric


losses during the adsorption/desorptionprocess [88].

12.3.3.1 Other TechniquesThe ellipsometric method was usedto study adsorbed DNA for the firsttime by Humphreys and Parsons [222].Since that time the sensitivity of thismethod has appreciably increased [223].The measurement of x-ray reflectionenabled studies of the structure of verythin layers (down to several A) and themorphology of the interface [224].

12.3.4Adsorption of NAs on Other Electrodes

Electrochemical techniques, scanningprobe microscopy (STM, AFM), and el-lipsometry can be used for the investiga-tions of layers of NA bases, nucleosides,and polynucleotides adsorbed and/or de-posited on metallic and semiconductorsubstrates. Hinnen and coworkers [225]studied the adsorption of ds and ssDNAon gold electrodes by CV and differen-tial capacitance measurements. From thegold surface, the DNA was completelydesorbed at −0.8 V. The ellipsometric ob-servation [222] and impedance measure-ments [108] showed complete desorptionof ssDNA from the Hg surface at about−1.6 V. Thus, the DNA on the Hg surfacewas still adsorbed at the potentials of re-duction of base residues, while from thegold surface DNA was desorbed (at −0.8 V)before the reduction of base residuescould take place [225]. For a potential stepfrom −0.8 V to a potential near the p.z.c.(around 0 V), the time required to ob-tain a nearly constant value of differentialcapacitance using a gold electrode wasabout 7 min for native DNA and 2 min fordenatured DNA; the final value of the dif-ferential capacity being about 12 µF cm−2

both for ds and ssDNA. The final valueusing a Hg electrode was 11 µF cm−2 andthe time required to obtain this nearly con-stant value was shorter than with a goldelectrode.

The adsorption of DNA on gold wasalso studied by STM and AFM tech-niques [226, 227]. NAs were adsorbed ina broad range of potentials on carbon andsilver electrodes [126, 129, 188, 189, 228,229]. Gold surfaces modified with thiol-derivatized DNA duplexes were studied asa function of potential using AFM [230].The duplexes either stood up (up to about+0.45 V against a Ag wire) or were flat (atmore positive potentials) on the gold sur-face, depending on the electrode potentialrelative to p.z.c. At open circuit mono-layers of well-packed DNA helices with afilm depth of about 45 A corresponding toan average ∼45 orientation of the helicalaxis with respect to the gold surface wereformed. The voltage-induced morphologychanges were reversible and constituted ananoscale mechanical switch. New insightinto the mechanism of biopolymer adsorp-tion at charged surfaces can be obtained bythe study of surface properties using opti-cal methods [84, 223, 231–235] and x-rayreflection [224].

Adsorption of organic molecules on thesurface of a semiconductor influences theregion of the space charge in it. The effectof the internal electric field of the semi-conductor on the optical properties of itssurface is well established [236]. A num-ber of significant studies of organic thinfilms have been carried out using ellip-sometry [237], showing high sensitivity ofthis technique to the interfaces with or-ganic species, and to very thin films ofbiological materials [234, 235, 237]. Ellip-sometry has experienced a rapid evolutionin the last two decades. The extension ofthe technique into the infrared region was

12.4 Reduction and Oxidation of NAs on Different Electrodes 385

possible by coupling the ellipsometer toa Fourier-transform infrared spectrome-ter (FTIR) [223]. This combination pro-vided significant advantages in studyingthe optical response of complex thin-film structures, such as high-temperaturesuperconductors [232] or anisotropic ma-terials [231, 233].

Adsorption of ssDNA on gallium ar-senide (GaAs) surfaces was studied, andits reactions with complementary andnoncomplementary strands probed [238].The adsorption of DNA on GaAs sur-faces was studied, both on a GaAs waferand on the molecular controlled semi-conductor resistor (MOCSER) [239] usingFTIR spectroscopy, X-ray PhotoelectronSpectroscopy (XPS) measurements, andelectrical measurements. It was found thatDNA is bound directly to the GaAs sur-face through the phosphate group, withoutany modification of the surface or of theadsorbates. The hybridization process ofthe adsorbed ssDNA with its complemen-tary strands was performed both on aGaAs wafer and on the MOCSER andthen compared to the reaction with thenoncomplementary strands. Both the IRspectroscopy and the MOCSER indicatedselectivity in the reactions [238].

12.3.4.1 DNA Adsorption to Charged LipidMembranesThe negatively charged phosphate groupsalong the DNA backbone strongly interactwith cationic lipid membranes, to whichDNA readily adsorbs forming highly or-dered two-dimensional aggregates. Theseaggregates can be imaged with AFM. Itwas found that the DNA adsorption de-pends on the surface charge density andon the size of the cationic lipid bilayer ar-eas [240]. A condensed phase of DNA ona lipid membrane was also observed bySpector [241] using AFM.

12.4Reduction and Oxidation of NAs onDifferent Electrodes

NAs can be reduced and oxidized at elec-trodes yielding faradaic signals that canbe useful in biochemical analysis (re-viewed in [14, 15, 242]). Perhaps the largestnumber of results has been obtained withmercury and carbon electrodes.

12.4.1Mercury Electrodes

In agreement with the electrochemical be-havior of the monomeric units of theNAs (Sect. 12.2), A and C residues insingle-stranded (ss) NAs were reducibleat mercury electrodes (Figs. 4, 6). Thepresence of G in DNA and RNA wasmanifested by an anodic peak G incyclic modes (Figs. 4, 6a). T and Uresidues were inactive in aqueous solu-tions; in nonaqueous solutions (such asdimethyl formamide with 0.1 M tetrabutylammonium perchlorate) reduction of Uresidues was observed in poly(U) [243].With ssNAs, qualitatively similar resultswere obtained, both on the DME and onthe HMDE; this was, however, not thecase with dsNAs where significant (rela-tively slow) conformational changes tookplace at the HMDE surface under cer-tain conditions (Sect. 12.6). Reduction ofNAs on mercury electrodes involved pro-tonation of the NA base residues [10],similar to protonation of free bases(Sect. 12.2).

12.4.1.1 Reduction of Adenine andCytosine ResiduesIn contrast to free A, which is reducedat acid but not at neutral pHs, reductionof A residues in ssDNA and ssRNA canbe observed even at neutral pH, when a


(a)+0.1

−0.6

I[µ

A]

−1.4

(b)

G (c)

(d)

G

500 nA

20 nA

II

III

20 nA

500 nA

d.c.P DPP

−0.8 −1.0E

[V]

E[V]

Al

Cl-2

Cl-1

E[V]

1,2

1

2

2 135

0.1 µA

4

2

1

CA −1.6

−1.5 −1.0 −0.5

−1.5

−0.8 −1.0E[V]

E[V]

−1.6 −1.5

Fig. 6 Schemes of reduction and oxidation signals of DNA at mercury electrodes. (a) cyclicvoltammograms of ssDNA obtained upon repeated potential cycling. (CA), peak due to reductionof C and A; (G), peak due to G; inset, detail of peak G; the succeeding scans are numbered.(b) OP at controlled AC: polarogram dE/dt = f(E) of ssDNA. The indentations CI-2 and AIcorrespond to peak CA and peak G, respectively. (c, d), d.c. and DP polarograms of (c), ds or (d),ssDNA. DPP peak III corresponds to the CV peak CA. (Adapted from E. Palecek, in Topics inBioelectrochemistry and Bioenergetics (Ed.: G. Milazzo), John Wiley & Sons, Chichester, 1983,pp. 65–155, Vol. 5; F. Jelen, E. Palecek, Biophys. Chem. 1986, 24, 285–290; L. Havran, M. Fojtaand E. Palecek, unpublished.)

suitable salt (such as 0.3 M CsCl or 0.3 Mammonium formate) efficiently screeningthe negatively charged phosphates of theNA (and facilitating the nucleic acid ad-sorption) is used in a buffered backgroundelectrolyte. Under these conditions, A andC residues in DNA and RNA are re-duced in an adsorbed state, usually ina single peak (wave) at potentials closeto −1.4 V. For more details, see previ-ous reviews [4, 5, 10, 31] and referencestherein.

12.4.1.2 Anodic Signal of GuanineResiduesThe ability of DNA and RNA to producean anodic signal due to G residues wasrecognized about 40 years ago by meansof OP [2, 3, 9] (Fig. 6b). It was shown thatat highly negative potentials of the back-ground electrolyte decomposition a G re-duction product was formed, the oxidationof which was probably responsible for theobserved anodic signal [245, 246]. Then,for almost two decades, this phenomenon


was little studied in connection with re-placement of OP by differential (derivative)pulse polarography as a preferred methodfor the NA electrochemical analysis.

Application of CV to studies of G signalsin guanosine [134, 135], synthetic polynu-cleotides [244], and DNA [135], as well asidentification of the G reduction productby means of macroscale electrolysis [135],helped to understand better the electrodeprocesses to which G residues were sub-jected on the mercury electrode and to findconditions suitable for the analysis of G-containing compounds [136]. It has beenshown that ODNs [107, 247], syntheticpolynucleotides, DNAs, and RNAs [136,138] as well as PNA [247] containing Gdisplay in CV (or in other methods work-ing in the cyclic mode) an anodic peakclose to −0.3 V (peak G) (Fig. 6a), whileNAs not containing G do not produce thispeak. A condition for the appearance ofpeak G is the previous polarization of themercury electrode to sufficiently negativepotentials (around −1.8 V). At these po-tentials reduction of G residues takes placeinvolving the 7,8 double bond of the im-idazole ring in G as a primary reductionsite (Fig. 4). In the anodic process reoxida-tion of the G reduction product back to Goccurs. This chemically reversible processtakes place even when a NA contains Aand/or C residues in addition to G, whereA and C reduction leads to the formationof products blocking the electrode surface(Fig. 6a) [244]. The electrode process of Gresidues involves protonation [135, 136],and close to neutral pH it requires thepresence of some salts in the backgroundelectrolyte, such as 0.6 M ammonium for-mate, 0.1 M MgCl2, and Mg(ClO4)2 [136],to provide a well-developed peak G. PeakG is highly symmetric (suggesting in-volvement of adsorption in the electrodeprocess), offering a better possibility for the

determination of NAs at concentrationsbelow ppm (by stripping techniques) thanthe highly asymmetric reduction peak ofC and A residues formed at potentials tooclose to the background discharge. Squarewave voltammetric (SWV) stripping [248]and constant current chronopotentiomet-ric stripping analysis (CPSA) [249] wereapplied to study peak G, at concentrationsbelow 1 ppm both techniques produced apeak G better developed than that obtainedwith CV. In CPS and SWVs analyses, theelectrode was exposed to negative poten-tials (around −1.8 V) for at least 1 s; pro-longed contact of ssDNA at these potentialswas favorable for the accumulation of theG reduction product. At these potentials adisturbance of the structure of dsDNA onthe electrode surface may, however, takeplace (Sects. 12.5 and 12.6).

Recently mercury film electrodes (MFE)have been employed in NA analysis. Usingmercury film on a silver electrode, cathodicsignals of electroreducible nucleosidesand high concentrations (hundreds ofµg ml−1) of denatured and degraded calfthymus DNA were obtained [250]. Weused mercury film on a glassy carbonelectrode (GCE) for the measurements ofboth redox and tensammetric response ofDNA, RNA, synthetic polynucleotides, andPNA [251]. Yeast tRNA and calf thymusssDNA were detected at concentrations of50 and 100 ng ml−1, respectively, at 180 saccumulation, when peak G was measuredusing CPSA at the MFE. Moreover, MFE(like the HMDE) was capable of detectingcleavage of DNA in solution and on theelectrode surface [252] (Sect. 12.9).

12.4.2Carbon Electrodes

Oxidizability of A and G residues (Fig. 4)in polynucleotides was demonstrated by


the end of the 70 s [243, 253]. Linear sweepvoltammetric (LSV) signals were, however,poorly developed and the improvementachieved by application of DPV [253] wasnot sufficient to obtain sensitivity compa-rable to that obtained with mercury elec-trodes. Reports of well-developed peaksproduced by ssDNA on GCE [254] wereprobably due to the presence of free purinebases released from DNA by treatmentwith concentrated perchloric acid [248].

Recently, application of CPSA with so-phisticated baseline correction dramati-cally improved the sensitivity of the NAanalysis on carbon electrodes [255–257].Poorly developed peaks and inflectionsproduced by NAs at concentrations ofhundreds or tens of micrograms per mlwith voltammetric techniques turned intowell-developed peaks Gox and Aox (dueto oxidation of G and A, respectively)detectable at concentrations 2 to 3 or-ders of magnitude lower (Fig. 7). It wasshown that both the interfacial and redoxproperties of RNA and DNA are stronglydependent on the nature of the carbonelectrode materials [258]. The trend inthe sensitivity: CPE > pyrolytic graphite >

highly oriented pyrolytic graphite > car-bon strip (showing detection limits for(G)20 ODN of 20, 30, 40, and 50 ng perml, respectively, using 5-min accumula-tion); at these ODN concentrations, noresponse was obtained with glassy car-bon and carbon fiber electrodes. Carefulsearching for a suitable carbon fiber ma-terial in J. Wang’s laboratory resulted in afinding that carbon fibers are suitable asmicroelectrodes for determination of NAsat high sensitivity in unstirred solutions atvery low ionic strength [259]. CPSA of NAswith carbon electrodes now offers sensi-tivities comparable to those obtained withmercury electrodes [247, 257, 260]. Recentdata suggest that even the voltammetricstripping on carbon electrodes can matchthe sensitivity of CPSA, if proper baselinecorrection is applied to compensate forthe very high charging currents of carbonelectrodes [247, 261]. Renewable graphitepencil electrodes were demonstrated tobe an excellent material for trace mea-surements of NAs, showing low detectionlimits and good surface-to-surface repro-ducibility [262]. Membrane-covered elec-trodes were used for the analysis of

400 s V−1

0.2 s V−1

1.0 s V−1

Gox

1.0

2.0 µA

1

2

1

2

3

4

1.2

0.6 1.2

Potential[V]

0.6 1.2

Aox

(c)

(a)

(b)

Fig. 7 CPSA of DNA and RNA at carbonelectrodes. (a) ssDNA decamer AAAAGGAGAGat a relatively high concentration of 1 µg ml−1

(310 nM) showing oxidation peaks of G (Gox)and A (Aox). Voltammetric (b), andchronopotentiometric (c) G (oxidation)stripping peaks of tRNA at concentrations of10 ng ml−1 at 5-min accumulation time (B1,C3,) and 5 ng ml−1 (200 pM) at 10-minaccumulation (C1). Corresponding backgroundelectrolytes (B2, C2 and C4). (Adapted fromM. Tomschik, F. Jelen, L. Havran et al.,J. Electroanal. Chem. 1999, 476, 71–80; J. Wang,X. Cai, J. Wang et al., Anal. Chem. 1995, 67,4065–4070.)


short ODNs in the presence of long NAmolecules [259, 263].

12.4.3Other Solid Electrodes

Solid electrodes modified with variouscompounds have been used to improveDNA oxidation response. Siontorou andcoworkers [264] obtained peaks Aox andGox of degraded DNA at GCE modifiedwith self-assembled bilayer lipid mem-brane. Using GCE modified with Nafion-ruthenium oxide pyrochlore, enhance-ment of oxidation peaks of both peaks,Gox and Aox, was achieved [145]. Thorp’sgroup investigated DNA oxidation re-sponse at GC and indium-tin oxide (ITO)electrodes modified with self-assembleddicarboxylate monolayers [265], and withnitrocellulose and nylon membranes [266].In these experiments, DNA was attachedto the electrode either covalently or viaadsorption forces in the modifier layer;bare ITO surface did not adsorb DNA.Oxidation of DNA was mediated by aredox metal chelate [Ru(bipy)3], whichshuttled electrons to the electrode sur-face from DNA in solution or attachedat the modifier film [265, 266]. Electro-catalytic oxidation of DNA was observedalso when a redox mediator was immobi-lized on the electrode surface, for example,on ITO modified with electropolymerizedpoly[Ru(bipy)3] film [146].

Adsorption and electrooxidation of ssNAs on a silver electrode were studiedby electrochemical methods and surface-enhanced Raman spectroscopy [189, 228].Using the latter electrode, Fan and cowork-ers [229] observed an anodic signal insolutions of DNA. This signal was at-tributed to redox reactions of purine bases,and provided a convenient way to deter-mine DNA. Oxidation of purine bases

on the gold electrode (where measure-ment of the respective anodic currentsis complicated by simultaneous formationof gold oxide) was discussed by Hinnenand coworkers [225]. Pang and cowork-ers electrooxidized native and denaturedDNA and purine nucleotides on a gold mi-croelectrode [267]. A Copper electrode inconnection with sinusoidal voltammetrywas utilized by Singhal and Kuhr [149, 268]for determination of nucleotides, oligonu-cleotides, and DNA based on oxidation ofthe sugar moiety. This detection approachis universal to all types of nucleotides andis highly sensitive for both ss and dsDNA.

12.4.4Analysis of NAs by DifferentElectrochemical Techniques

Various electrochemical methods havebeen applied for the analysis of NAs,including DPP [5, 11] and DPV [13, 269,270], linear sweep and CV [13, 271]square wave [138] and a.c. voltamme-try [272–274], and recently constant cur-rent chronopotentiometry [249, 255–257,275, 276] and elimination voltamme-try [139, 277–279]. DPP was applied forthe analysis of DNA in 1966 [280], and ina short time, it replaced OP and d.c. po-larography used in the early NA studies [4,5]. The main advantage of DPP is its bet-ter sensitivity and resolution of peaks. Calfthymus ssDNA produced a well-developedDPP peak III (Fig. 6d) at concentrationsof about 10 to 20 µg ml−1, while dsDNAwas inactive at the same concentration.At higher concentrations (hundreds ofµg ml−1), dsDNA produced peak II at po-tentials by about 70 mV more positive thanpeak III (Fig. 6c). For years, DPP was themost sensitive instrumental method of de-termination of traces of ssDNA in dsDNAsamples [5].


Adsorption Washing Measurement

Mediumexchange

sc DNA

oc DNA

scDNA

ocDNA

Carbonelectrode

Mercuryelectrode

Ionizingradiation

Radicals

Nucleases

Carbonelectrode

Mercuryelectrode

Single-strandbreak 3

No peak IIIGox

Gox

Hydroxylradical

OH

OH

OH OH OHOHOH OHOH

OHOH

OHOHOH

OHOH

OH

OHOH

OHOH

Mediumexchange

(a)

(c)

(d)

(b) (b1)

Fig. 8 (a) Schematic representation of theadsorptive transfer stripping procedure. NA (orprotein) is strongly adsorbed at the electrodefrom a small (3–5 µl) drop of solution. Then theDNA-modified electrode is washed, followed bytransfer into a usual electrochemical cell withblank background electrolyte, where themeasurement is performed. Multiple mediumexchanges are possible, and this allows the studyof interactions of surface-confined DNA withsubstances in solution. (b, c) Detection of DNAstrand breaks (sbs) with DNA-modified mercuryelectrodes. (b) Scheme of AC voltammetricresponse of sc and open circular (oc) DNA atmercury electrodes. ScDNA does not yield

peak III. Introduction of a sb in scDNA results ina better accessibility of bases in the vicinity of thesb and in formation of peak III. (b1), Unlikemercury electrodes, carbon electrodes are littlesensitive to formation of sb in DNAs.(c) Voltammetric detection of formation of DNAsb in solution. ScDNA is treated with aDNA-cleaving agent followed by DNA adsorptionat the electrode. After washing, the electrode istransferred into a cell and DNA voltammetricresponse is measured. (c) Detection ofDNA-damaging agent with scDNA-modifiedelectrode. Intact scDNA adsorbed at theelectrode serves as a sensitive layer of a sensorfor DNA damage.

Tens to hundreds of micrograms ofDNA, required for DPP and for other abovementioned polarographic and voltammet-ric techniques, were acceptable in the

analysis of chromosomal DNAs, but theseDNA amounts were too large for the anal-ysis of plasmid and viral DNAs, DNAfragments, and synthetic oligonucleotides,

12.5 Relations Between Structures and Electrochemical Responses of DNA 391

whose preparation is laborious and/orexpensive. Application of adsorptive strip-ping techniques in 1986 [203] helped todecrease bulk concentrations of DNA byabout two orders of magnitude and to re-duce the lowest amounts of the analyte tohundreds of ng of the NAs. Shortly after-ward, these amounts were decreased by2 to 3 orders of magnitude, using a sim-ple way of preparation of DNA-modifiedelectrode [281]. Instead of performing thevoltammetric analysis with the HMDE im-mersed in the analyte, the DNA-modifiedelectrode was prepared by immersing theelectrode into a drop of the analyte (about3–5 µl) for a short period of time (usu-ally 30–300 s), during which the DNA wasirreversibly adsorbed and firmly attachedto the electrode (Fig. 8a). The HMDE wasthen washed and transferred to the back-ground electrolyte (not containing anyDNA) in which the voltammetric mea-surements were performed. This medium-exchange procedure was called AdTSV [13,281–283].

DNA, RNA, and protein-modified elec-trodes can be prepared using both carbonand mercury surfaces [283]. Stability ofimmobilization of NAs at HMDE andgraphite electrodes is very good [270].AdTSV has been widely applied to vari-ous kinds of NA and protein studies [13,15, 249, 270, 281, 284, 285]. Comparedwith conventional voltammetry, AdTSVhas many advantages that are mainly dueto the separation of the biomacromoleculeadsorption from the electrode processes.These advantages include (1) reduction ofthe sample volume to 3 to 10 micro-liters, (2) elimination of interferences bylow molecular mass substances that arewashed off in AdTSV, (3) adsorption ofthe biomacromolecule on the electrodefrom media not suitable for the conven-tional voltammetric analysis, (4) in studies

of interactions of DNA immobilized at theelectrode surface with other compoundsin solution such as specific proteins.During the measurement, the results ofAdTSV are influenced neither by the DNAinteractions in the bulk of solution norby transport of DNA to the electrodefrom the solution. More details aboutanalytical applications and the use of NA-modified electrodes are given in Sects. 12.7and 12.9.

12.5Relations Between Structures andElectrochemical Responses of DNA

In early d.c. polarographic and OP stud-ies, marked differences in the responsesof native (double-stranded, ds) and de-natured (single-stranded, ss) DNA wereobserved [4, 11] (Figs. 5, 6). Large signalsof ssDNA as compared with much smalleror no signals produced by dsDNA wereexplained by decreased accessibility of theelectroactive sites in dsDNA (Fig. 4) [4, 5,10, 11, 14, 15]. The primary reduction sitesof C and A (on mercury electrodes) arelocated in the interior of the DNA du-plex structure, where they form a part ofthe Watson-Crick hydrogen bonding sys-tem (Fig. 4). The primary reduction siteof G is located closer to the surface ofthe molecule in the major groove and it isnot involved in the Watson-Crick hydrogenbonding. The primary oxidation sites (oncarbon electrodes) are also not involved inthis hydrogen bonding.

In agreement with the location of theA and C reduction sites in dsDNA, theDNA reduction signals (obtained withmercury electrodes) showed a high sen-sitivity to changes in the DNA structure.In the d.c. polarographic mode (withDME) at neutral pH-values the dsDNA


was inactive, while the ssDNA produceda polarographic reduction wave (Fig. 6d)at concentrations of hundreds of µg perml [5]. DPP peak II produced by dsDNAwas highly sensitive to small changes inthe DNA structure induced by chemicals,enzymes, and radiations. In intact chro-mosomal dsDNA this peak was about a

hundred times smaller than peak III ofssDNA. Both ss and dsDNAs produced aDP (non-faradaic) peak I, which was lesssensitive to changes in the DNA struc-ture and was, therefore, not used in theDNA structure studies. At neutral pH, theDPP was capable of reflecting changes inthe DNA structure in the bulk of solution

20 8060

0.1

0.2

I

Temperature[°C]

40

Premelting region

Melting region

20

20

1.0

1.2

A

60

I

40°C

Fig. 9 Premelting changes in DNA conformation detected byUV-adsorption (•), and electrochemical methods (). Temperaturedependence of the heights of DPP peak II of dsDNA was measured in4-M NaClO4, 0.05-M sodium phosphate, pH 6.5 at temperaturesindicated in the graph; absorbance at 260 nM. Inset, temperaturedependence of the heights of CV peak G of dsDNA obtained in 0.3-Mammonium formate, 0.05-M sodium phosphate, pH 6.9 by (),conventional voltammetry performed at the given temperatures; (),AdTSV: (, •, ) DNA adsorption at the given temperatures, CVmeasurement at room temperature (after medium exchange); (),adsorption at room temperature, measurement at the temperaturesgiven in the graph; (, •, ), dsDNA; () ssDNA. Premelting changesin dsDNA conformation occurring at elevated temperatures (, ) arefixed at the electrode surface and manifested by CV measurements atroom temperature (•). No such changes occur in ssDNA (). WhenCV measurements are performed at elevated temperatures, the heightof peak G decreases with temperature because of the shift in potentialof G reduction (at highly negative potentials). (Adapted fromE. Palecek, Bioelectrochem. Bioenerg. 1988, 20, 171–194; E. Palecek,I. Fric, Biochem. Biophys. Res. Commun. 1972, 47, 1262–1269.)

12.6 DNA Structure on Electrode Surfaces 393

(Fig. 9), and its results agreed well withthose of optical methods [10, 11]. In ad-dition, the non-faradaic capacity signals ofDNA obtained by a.c. polarography (Fig. 5)were highly sensitive to changes in DNAstructure. A.c polarographic peak III pro-vided information about accessibility ofbases for interaction with the electrodesurface and its changes were in mostcases qualitatively similar to changes infaradaic signals. Strong influence of DNAstructure on electrochemical signals mademercury electrodes suitable for studies ofDNA structural transitions and detectionof local conformational changes. Mer-cury electrodes sensitively reflected singlestrand interruptions in linear and circu-lar DNA molecules [286], transition fromright-handed B-DNA to left handed Z-DNA double-helix [287], and differences inthe superhelix density of sc DNAs (Fig. 8)and superhelix density-dependent struc-tural transitions in DNA [10, 273]; someexamples are given in Sect. 12.9.

A comparison between the responses ofss and dsDNA, obtained by DPP (withDME) on one hand, and CV and SWV(with HMDE, at full and partial elec-trode coverages) on the other, is shownin Fig. 10(a, b). A qualitative differencebetween ss and dsDNA is observed onlyby DPP (ssDNA produces peak III whiledsDNA yields peak II but not peak III).At the HMDE the faradaic responses dif-fer only quantitatively and ssDNA alwaysshows higher signals than the parentdsDNA. The difference between the ssand dsDNA signals is higher at a par-tial electrode coverage (faster diffusion ofssDNA contributes to this difference) andin cathodic signals (Fig. 10a). At full elec-trode coverage, peak G of calf thymusdsDNA (Fig. 10b) corresponded approxi-mately to 50% of the height of this peakproduced by ss form of this DNA. The

difference between the CV anodic peakG heights of ss and dsDNA was thussmaller (Fig. 10b) than in the cathodicpeak CA, in agreement with the locationof the G reduction site outside the hydro-gen bonding system, close to the DNAmolecule surface. On carbon electrodesthe difference in the heights of the oxi-dation peaks of ss and dsDNAs (Fig. 10c)was even smaller, reflecting probably onlythe different flexibilities of ss and dsDNAmolecules adsorbed at relatively rough car-bon surfaces [31].

With polarographic methods workingwith large potential excursions duringthe drop lifetime or with voltammetricmethods using HMDE the electrochemicalresponses have not always correspondedto the DNA structure in solution [10]. Thisproblem, which is connected with slowprocesses involving dsDNA structure onthe electrode surface, is discussed in thefollowing section.

12.6DNA Structure on Electrode Surfaces

A good correlation between polarographicresponses (obtained with methods work-ing with small potential excursions duringthe drop lifetime) and optical methods(Sect. 12.5) suggests that the structureof dsDNA is not significantly changedbecause of the adsorption of dsDNAat DME in the potential range of themeasured DNA reduction signals (i.e.around −1.4 V). It is shown that in anarrow potential window, close to the po-tential of the a.c. polarographic peak I(Sect. 12.3), large changes in the interfa-cial properties of dsDNA may occur andinvolve major parts of the adsorbed DNAmolecules.


−5

10

25

40

55

70

85

100

ss CV CV CV CV CPSA CPSADPP DPPSWVf f fp p f pII III

Pea

k he

ight

[%]

(a) (b) (c)

Fig. 10 Relative heights of faradaic peaks of calfthymus ss and dsDNAs at full (f) and partial (p)electrode coverage obtained by (a, b) cyclicvoltammetric (CV) and SWV stripping analysis(at HMDE) and by (DPP at DME) at neutral pHand (c) by CPSA at pyrolytic graphite electrodes.(a, b). Signals of dsDNA are displayed relativelyto the signals of ssDNA that were taken as 100%(for each given experimental conditions). Atpartial electrode coverage of HMDE (10 µg DNAml−1, 60 s waiting) the surface concentration ofssDNA was higher than that of dsDNA (the latterDNA has a lower diffusion coefficient). To obtainfull electrode coverage dsDNA was at aconcentration of 200 µg ml−1 and ssDNA at40 µg ml−1). At HMDE some secondary changesin the structure of dsDNA may occur (Sect. 12.6)whose extent is influenced by the time for whichdsDNA is exposed to potentials of the region U

(Fig. 11). These changes cannot influence theDPP signals of dsDNA (obtained with DME);consequently, the DPP peak III (characteristic forssDNA) is completely absent in intact dsDNA(Figs. 4 and 6 for the location of the reductionsites in DNA and other details). (a) DPP peak II(about 70 mV less negative than peak III) issensitive to damage of dsDNA. Signals obtainedwith HMDE: CA, cathodic peak (due to reductionof A and C residues) (b) G anodic peak (due tooxidation of the G reduction product). (a,b) Background electrolyte: 0.3 M ammoniumformate, 0.05 M sodium phosphate, pH 6.9.(c) DNA was adsorbed from O.2 M NaCl, 50 mMphosphate, pH 7 and CPSA performed in 0.2 Msodium acetate pH 5 after medium exchange;CPSA oxidation peak of G, Gox is shown. ([10];F. Jelen, L. Havran and E. Palecek, unpublished).

12.6.1Dependence of the dsDNA Signals at theHMDE on Potential Scanning Direction

In 1961, Miller [37] concluded on thegrounds of his differential capacity mea-surements (Sect. 12.3.) that in the region

of positive potentials partial unwind-ing of dsDNA took place on the mer-cury electrode. Later, Flemming [17, 290]did not confirm Miller’s conclusionand assumed that DNA preserved itsdouble-helical structure over the wholerange of potentials. Flemming’s [290] a.c.


voltammetric measurements of dsDNA(with the HMDE) agreed qualitatively witha.c. polarographic results (using DME),only if the potential was scanned fromnegative to positive values (Fig. 5b). Whenscanned in the opposite direction dsDNAyielded (at the HMDE) a ‘‘sharp peak’’at about −1.5 V (corresponding to a.c.voltammetric peak III, characteristic forssDNA) (Fig. 5b). On the other hand, ss-DNA produced the ‘‘sharp peak’’ indepen-dently of the potential scanning direction.To explain his observations Flemming as-sumed that at potentials more positive thanp.z.c., the dsDNA surface concentration (atfull electrode coverage) was low, while atmore negative potentials the concentra-tion and thickness of the adsorption layerincreased and intermolecular interactionsbetween the dsDNA segments (extendedinto the solution) took place. Flemmingspeculated that the ‘‘round peak’’ (corre-sponding to peak I) is due to desorption ofisolated dsDNA molecules and the ‘‘sharppeak’’ to desorption of associates of thedsDNA molecules [17, 18, 290]

Shortly afterward, it was shown [184,197, 291] that keeping dsDNA on the mer-cury electrode at potentials around −1.2 Vat neutral pH resulted in electrochemi-cal responses characteristic for ssDNA; nodifference whether(not clear) capacitive orfaradaic signals were measured [10, 12].These results suggested that formation ofthe capacitive ‘‘sharp peak’’ by dsDNA(when scanning to the negative poten-tials) could hardly be explained only byassociation of dsDNA molecules. To over-come this difficulty, Berg [18] suggested acomplicated model for DNA adsorptionand electron exchange based on a so-called potential induced π -state of dsDNAinvolving B-A transition at the surfaceand hypothetical changes of other DNAproperties, as well as electron hopping

to bases inside the dehydrated parts ofadsorbed DNA molecules. The idea ofelectron hopping in dsDNA is now be-ing intensively studied and is discussedin Sect. 12.8. The experimental evidencepresented by Berg [18] in support of hismodel was limited to the observation thatthe ‘‘sharp peak’’ decreased (‘‘peak fad-ing’’ in Berg’s terminology), either as aresult of waiting at the potential of thispeak (about −1.4 V) or due to scanningover this potential [18]. This phenomenoncan, however, be explained by two well-known facts: (1) at about −1.4-V reductionof A and C residues takes place, produc-ing the DNA reduction product that blocksthe electrode surface [184, 197, 244, 292](Fig. 6, Sect. 12.4); such electrode blockingwith intact dsDNA is negligible, but it maygain importance at longer waiting timesif dsDNA is degraded or contaminatedwith ssDNA. (2) At the same potentialthe unreduced dsDNA should graduallydesorb from the surface [108] (Sect. 12.3).Thus, no new special properties of DNAare necessary to explain the ‘‘peak fading.’’

12.6.2Opening of the DNA Double Helix Around−1.2 V (Region U)

Detailed studies based on both faradaicand capacitive DNA signals showed thatat neutral and weakly alkaline pH-valuesthe prolonged contact of dsDNA with thesurface of the mercury electrode chargedto potentials around −1.2 V (region U)resulted in striking changes in the elec-trochemical responses of DNA (Fig. 11).These changes, characteristic for ssDNA,were interpreted as the opening of theDNA double helix at the electrode sur-face [10, 15], and consequently, to anincreased accessibility of bases for their


−1.5

100

50

I[%

]

−1.0

Eexp[V]

−0.5

W U T

Fig. 11 Changes in the dsDNA structure at the HMDE surface atpotentials around −1.25 V (region U) at neutral pH. The graphshows the dependence of relative heights of the cathodic peak CAof (•,) ds and (, ×) ssDNA on the electrode potential. (,),conventional voltammetry (DNA was adsorbed at the electrode atthe given potentials); (•, ×), AdTSV (DNA was adsorbed at−0.1 V followed by medium exchange. Then, the surface-anchoredDNA was exposed to the potentials Eexp given in the graph for100 s, followed by CV scan). Heights of peak of ssDNA obtainedby conventional voltammetry for accumulation potential −0.1 Vwas taken as 100%. The helix opening in the region U wasrelatively slow (about 1 min was necessary to open major portionof chromosomal dsDNA molecules). No structural changes wereobserved in the region W. No extensive DNA opening was takingplace in the region T; fast opening limited to the vicinity of theDNA strand ends could, however, occur in this potential region.At acid pH’s DNA opening in the region T gained importance andclose to pH 5 the peak heights of dsDNA almost corresponded tothose of ssDNA. See text for more details. (Adapted fromE. Palecek, Bioelectrochem. Bioenerg. 1992, 28, 71–83.)

interaction with the electrode. It was sug-gested ([10, 293] and referenced therein)that the opening of the dsDNA is due tostrains in the DNA molecule, resultingfrom a strong repulsion of the negativelycharged DNA phosphates from the elec-trode surface to which the DNA was firmlyadsorbed via hydrophobic bases (providedfor example by single-strand interruptions,the DNA molecule ends, and transiently

opened DNA regions) (Fig. 11). It wasshown that the DNA opening was relativelyslow (tens to hundreds of s were necessaryto open about 90% of a chromosomal DNAon the electrode surface) and its rate in-creased with the potential shift to morenegative values [293]. The duplex open-ing was partially irreversible [293, 294]and depended on the DNA nucleotide se-quence [295]. With calf thymus DNA, both


AT and GC pairs were involved in the earlystage of the opening process [293]. Open-ing of dsDNA on the mercury electrode atpotentials of the region U was observedat ionic strengths between about 0.1 and1.0. At low ionic strength, dsDNA wasadsorbed at negative potentials of the mer-cury electrode very weakly [294] becauseof strong repulsion of the unscreenednegative charges of the DNA phosphatesfrom the electrode surface and becausethe above-mentioned phenomena couldnot come into play. In 10-mM KClO4,where probably only electrostatic inter-actions were involved, no differences inthe differential capacity of ss and soni-cated dsDNA were observed [225]. Underconditions of moderate ionic strengths atpotentials positive (region T) or negative(region W) to the region U (Fig. 11), noextensive surface changes in DNA confor-mation (comparable to those observed inregion U) were detected [10, 296].

It does not appear probable that theDNA opening on the electrode surfacecorresponds fully to the known DNAdenaturation in solution; some special fea-tures and/or limitations of the openingprocess can be expected with the DNAimmobilized at the surface. For example,formation of ‘‘ladder DNA’’ [12] or someother ds form in which bases are accessiblefor the interaction with electrode surfacemight also be compatible with the experi-mental data [10, 293, 295]. Increase in oxi-dation peaks of A and G were observed ongraphite electrodes at sufficiently negativepotentials (between −0.4 and −0.8 V) [31,190]. These peak changes were larger inAT-rich DNAs and were interpreted interms of DNA unwinding on a nega-tively charged graphite surface. More workwith carbon electrodes will be necessary tobetter understand this phenomenon. Liter-ature about the problem of DNA structure

on mercury electrodes was thoroughly re-viewed [10, 12, 18, 31]. Further, we wishto limit ourselves mainly to more recentpapers dealing with the subject.

Recent results obtained with a DNA-modified HMDE [293] unambiguously ex-cluded the explanation offered by Flem-ming (reviewed in [17]). DsDNA was at-tached at the HMDE at potential Ea −0.1 V(i.e. at a potential more positive thanthe p.z.c.); the electrode was washed andtransferred to the background electrolytenot containing DNA and exposed to poten-tials Ebx (varying from −0.1 to −1.55 V)for 100 s prior the CV measurements [293].In these experiments, a distinct region Uwas observed (showing a steep increase inpeaks CA and G) similar to that obtainedin a conventional way with the HMDEimmersed in a DNA solution during themeasurements. In experiments with theDNA-modified HMDE, the amount ofDNA attached to the electrode surfacecould not increase (because there was noDNA in the bulk of solution to diffuse tothe electrode and to reoccupy the emptysurface resulting from reorientation of theDNA molecules), and the increase of CVpeaks could not thus be due to a higherDNA surface concentration at negative po-tentials.

The irreversibility of the opening of ds-DNA [293] is in agreement with the findingthat the electrode tends to fix DNA in thespatial arrangement in which the moleculewas adsorbed at the electrode surface [282,283, 288]. For example, signals of dsDNA(but not of ssDNA) showed typical premelt-ing changes when dsDNA was adsorbedat different temperatures (in an AdTSVexperiment with the HMDE) and voltam-metric measurements were carried out(after the medium exchange) at room tem-perature [288] (Fig. 9). Similarly, effects ofcomposition and pHs of the medium from


which DNA was adsorbed effectively influ-enced the signals measured (in AdTSV) inthe usual background electrolyte at neu-tral pH [282, 283, 288]. Our studies of a.c.voltammetric behavior of DNA complexeswith intercalators [297] (Sect. 12.7.1) sug-gested that an altered DNA conformationmay be conserved after DNA adsorptionat the HMDE surface and the intercalatorremoval.

A strong support for opening of theDNA double helix in the region U hasrecently been obtained from our experi-ments with covalently closed circular (ccc)DNAs ([272, 284] (Fig. 8b)). These DNAsdo not contain any molecular ends andstrand interruption, and their extensiveunwinding is prevented for topological rea-sons [298, 299]. Exposition of the cccDNAs(plasmid sc DNAs) to the potentials of theregion U at the HMDE surface resultedin little or no detectable DNA opening,as indicated either by faradaic [284] or ca-pacitive [272] signals, in agreement withthe limitations in the duplex unwind-ing of these DNA molecules [298, 299].Studies of DNA adducts with platinumdrugs, such as bifunctional cis- and trans-diaminedichloroplatinum(II) and mono-functional diethylenetriaminedichloroplat-inum(II) (dien-Pt), showed that the ef-fects observed in region U are inhibitedby interstrand crosslinks (produced bythe bifunctional compounds), but not byother types of adducts formed in DNAby dien-Pt, in agreement with the con-cept of the DNA surface opening, whichshould be inhibited by the DNA inter-strand crosslinks [300].

12.6.2.1 Opening of dsDNA at Acid pHs ina Wider Potential Range TDecrease of pH (from neutral) resultedin an increase of the reduction signals ofdsDNA in the region T; at about pH 5

the signals of dsDNA were close to thoseof ssDNA and a distinct region U wasno longer observable. This effect was ex-plained by protonation of bases at acid pHsconnected with destabilization of the DNAstructure [10]. It was assumed that par-tially protonated DNA was adsorbed in theregion T via protonated regions, with fur-ther protonation and destabilization takingplace at the surface in the neighborhoodof the adsorbed segments. This processwas much faster as compared with DNAopening observed at neutral pH [10].

It has been shown above that theHMDE is a useful tool in the study ofinterfacial properties of DNA, includingconformational changes of DNA occurringat the surface. In cases when such changesshould be avoided, polarographic methodsworking with small voltage excursionsduring the (DME) drop lifetime should beapplied [10]. Even in experiments with theHMDE, the surface changes in the DNAstructure can be minimized by choosingproper conditions, including fast voltagescanning. The possibility of potential-controlled opening of dsDNA appearsvery attractive in connection with recentattempts to create DNA biotechnologieson chips involving fast DNA sequencing(Sect. 12.9).

Opening of the DNA double helix onthe surface was recently observed with theatomic force microscope (AFM). Individ-ual dsDNA molecules attached to an AFMtip and a gold surface were overstretched,and the mechanical stability of theDNA double/helix was investigated [301].Stretching experiments with single DNAmolecules revealed a highly cooperativetransition, where the natural B-DNA wasconverted into a new overstretched con-formation called S-DNA [302]. In λ-phageDNA the B-S transition at 65 piconewtons

12.7 Interactions of NAs with Small Molecules 399

(pN) was followed by a second conforma-tional transition at 150 pN, during whichthe DNA double helix melted into twosingle strands [301]. On relaxation, thetwo single strands recombined to the dsconformation. Both the B-S and melt-ing transitions occurred at significantlylower forces in poly(dA-dT), comparedwith poly(dG-dC) [303].

12.7Interactions of NAs with Small Molecules

12.7.1Reversible (Noncovalent) Interactions

A broad range of low-molecular massinorganic and organic species interact re-versibly with NAs. Basically, there arethree different modes of small moleculebinding to duplex DNA [304] (Fig. 12):(1) binding of cations along the outersurface of DNA double helix, primarilyvia electrostatic interactions with phos-phate anions. This mode of binding is notrestricted to dsDNA. In addition to nonspe-cific condensation-type interactions [305],

transition metals can also form site-specific complexes with base residues;(2) groove binding interactions involvingdirect contacts of the interacting moleculeswith inner surfaces of major or minorgrooves of DNA double helix, includ-ing edges of base pairs in these grooves(depending on the nature of the inter-acting species, hydrogen bonding, electro-static and/or van der Waals contacts maytake part in this mode of binding); and(3) intercalation of planar condensed aro-matic ring systems between adjacent basepairs, primarily involving stacking inter-actions. Interaction of DNA binders withDNA results in changes of electrochem-ical responses of the former substancesdue to altered mass transport and/ordecreased accessibility of the binder elec-troactive moiety. In addition, alterationsof DNA electrochemical behavior uponinteraction with the ligands can be de-tected. Electrochemical techniques havebeen shown to be capable of determina-tion of binding constants, binding modesas well as binding site sizes of associationof DNA with ligands. Differential bindingof some substances to ss and dsDNA can

Electrostatic GrooveBinding

Mg2+

Mg2+Na+

N+

+

+

+

+

+

N+

Na+

Intercalation

Fig. 12 Scheme of three general modes of interactions of dsDNA withsmall molecules. See the text for details.


be utilized in electrochemical detectionof DNA hybridization or degradation [306](Sect. 12.9).

12.7.1.1 Inorganic Cations and SimpleMetal ComplexesConcentrations of monovalent (sodium,potassium, ammonium) ions significantlyaffect adsorption of DNA on an elec-trode (Sect. 12.3). This effect is relatedto shielding of negative charges ofDNA sugar-phosphate backbone by thecounterions. Interactions of magnesiumions with DNA (which play a unique rolein many biological processes) were uti-lized by Maeda and coworkers [307] ina Mg2+ ion-selective electrode. Associa-tion of divalent transition metals (cad-mium, lead) with DNA was studied bySequaris and coworkers [308, 309], usingcyclic and alternating current voltamme-try and chronocoulometry at the HMDE.Enhanced adsorption of Pb2+ ions at theHMDE as a result of the presence ofDNA, and compaction of DNA from anextended coil to a condensed state in thepresence of lead ions were observed [309].Assembly of silver ions along duplex DNAmolecule, followed by chemical reductionof the metal, was used by Braun andcoworkers [310] to create an electricallyconductive nanowire. Differential behav-ior of large DNA molecules and smallerRNA ones upon interaction with triva-lent [Co(NH3)6]3+ complex was utilizedby us in a method for RNA determinationin DNA samples [269]. While long DNAmolecules were precipitated by 50 mMcobalt complex, RNA molecules remainedin solution and were detected using AdTSV(Sect. 12.9). Binding of [Ru(NH3)6]3+/2+complex to thiol-derivatized ODNs, an-chored on gold electrodes [311, 312] orto DNA electrostatically immobilized atfunctionalized 4-thiopyridine monolayer

on gold surface [313], was employed fordetection and quantification of DNA onthe electrodes, and for detection of DNAhybridization [311]. On the DNA-modifiedelectrode, reversibility of reduction of thepositively charged ruthenium(III) complex(attracted to the surface-confined DNA)was enhanced relative to bare (DNA-free)electrode, while reversibility of reductionof negatively charged ferricyanide (repelledfrom the DNA) was lowered [311]. Somemetal ions exhibited an ability to form acoordinate bond with oxygen and nitrogenatoms of base residues. Free DNA basesand purine nucleotides form sparinglysoluble salts with mercury (Sect. 12.2).Accumulation of mercury (II) ions inpolyuridylic acid adsorbed at the HMDEsurface was reported by Johnston andcoworkers [314]. Copper(I) ions were sta-bilized upon binding to purine or C basesand nucleosides, and resulted in the for-mation of insoluble species at the HMDEor GCE and in splitting of Cu2+ reduc-tion at the HMDE into two one-electronsteps [66, 183]. Purine bases, denaturedand native DNA bound nickel(II) ions andcatalyzed electroreduction of the metal atthe mercury electrode [315].

12.7.1.2 Organic Metal ChelatesOrganic metal chelates possess severaladvantages in studies of specific DNA in-teraction because of their well-pronounced(and often reversible) electrochemistry,and the possibility to control the modeof DNA binding by choosing the propermetal chelating ligands. Bard and cowork-ers [316, 317] observed that tris-chelatedruthenium, cobalt and iron complexeswith 1,10-phenanthroline (phen) inter-acted with duplex DNA via intercalation.However, the mode of binding dependedon the metal redox state, for (2+) ionsexhibiting more favorable hydrophobic


interaction with DNA than for the (3+)ones. Analogous 2, 2′-bipyridine (bipy)chelates of ruthenium and osmium dis-played predominantly electrostatic bind-ing [317]. The cathodic and anodic CVpeak potentials of the electrostatic binders(e.g. [Os(bipy)3]3+/2+) shifted to more neg-ative values, while a shift to more positivevalues was observed upon intercalativebinding (e.g. [Co(phen)3]3+/2+) [316, 318].Normal pulse voltammetry on an ITOelectrode was used to determine bind-ing constants and distribution of (bipy)3

and (phen)3 metal chelates bound toDNA [319]. Association of [Ru(phen)3]2+or [Os(phen)3]2+ with DNA was detectedvia the measurements of electrogeneratedchemiluminescence (ECL) of the metalchelate in the presence of oxalate on gold,GC, or platinum electrodes [320, 321].Intercalation of the ruthenium complexresulted in a stronger decrease of the ECLsignal than external binding of the osmiumchelate. On the other hand, an ECL signalappeared on association of [Ru(phen)3]2+with duplex DNA at aluminum(III)-alkanebisphosphonate thin film [318]. Amicroscale method for the study of theinteractions of DNA with other redox-active molecules (e.g. intercalative cobaltchelates) based on DNA-modified goldelectrodes was proposed by Pang andAbruna [322]. Selective binding of the Co,Os, or Ru (phen)3 complexes to duplex NAswas utilized in discrimination between ssand dsDNA on the electrode, and detec-tion of DNA hybridization [275, 323–326](Sect. 12.9).

Single- and double-stranded DNA oligo-mers containing G displayed catalyticenhancement of the oxidation of ruthe-nium [265, 266, 327] or rhenium [328]complexes on carbon or ITO electrodes.These chelates were capable of one-electron oxidation of G residues [146, 265,

266, 327, 328]. Reduced forms of the com-plexes shuttled electrons from DNA to theelectrode, where the oxidized form waselectrochemically regenerated. Using thiseffect, Johnston and coworkers [329] stud-ied solvent accessibility of nucleobasesin ss and duplex DNA. Formation ofDNA duplex precluded direct collision of[Ru(bipy)3]3+, with the G residue and theelectrons had to tunnel through a finite dis-tance. This distance was lower when G wasin a mismatch and the oxidation rate fol-lowed the trend G(single-strand) > GA >

GG > GT > GC. [Os(bipy)3]3+ exhibitedanalogous electrocatalytic effects in thepresence of 8-oxo-G (8-OG), which haslower redox potential than G [330]. DNAinteractions with metalloporphyrins (MPs)containing copper, nickel, zinc and cad-mium ions were studied by Qu andLi [331], using CV on an HMDE. MPslacking an axial ligand were capable ofintercalating into DNA double helix, re-sulting in a deep decrease of the MP signal.MPs containing the axial ligand exhibitedonly outside DNA binding, resulting ina weaker current change. Binding con-stants of copper complexes with phen anda macrocyclic ligand tetrabenzo-[b,f,j,n] [1,5, 9, 13]-tetraazacyclohexadecane were de-termined by means of CV at GCE [306].These chelates caused DNA cleavage un-der aerobic conditions in the presence ofreducing agents (Sect. 12.9).

12.7.1.3 Other Noncovalent DNA BindersSpecific interaction of a minor-groovebinding dye, Hoechst 33258, with ds-DNA on a gold electrode resulted inan enhancement of the dye oxida-tion signal. This effect was utilized byHashimoto and coworkers [332] to de-tect DNA hybridization (Sect. 12.9). Wangand coworkers [333] studied interaction ofdaunomycin (DM) with DNA in solution


100 nA

12

3

a

ab

b

c

0.0

−1.6 −1.2 −0.8

Gox

bc a

Gox

5 s/Vδ

δ

δ

1.2E

[V]

E[V]

c

(a)

(b)

dtdE

I

b

Fig. 13 Examples of electrochemicaldetection of reversible interactions ofduplex DNA with small molecules. (A),AdTS AC voltammetric behavior ofdsDNA adsorbed at HMDE in thepresence of chloroquine (CQ). (a), noCQ; (b), 10 µM CQ; (c), 50 µM CQ.Upon binding of intercalators, DNApeak 2 increases while peak 3 decreases.(B), interaction of DNA with DMfollowed by CPSA at CPE. As a result ofDM binding to DNA, peak δ

corresponding to free DM decreasesand peak δb attributed to the bounddrug appears; (a), DNA; (b), DM; (c),DNA-DM. (Adapted from M. Fojta,L. Havran, J. Fulneckova et al.,Electroanalysis 2000, 12, 926–934;J. Wang, M. Ozsoz, X. H. Cai et al.,Bioelectrochem. Bioenerg. 1998, 45,33–40.)

and on the electrode surface by meansof CPSA at CPE (Fig. 13b). DNA-boundDM was oxidized at more positive poten-tials than the free drug. Moreover, lowconcentration of the drug in the bulk of so-lution induced conformational changes inDNA adsorbed on the electrode. Mascini’slaboratory used DM as an electroactiveindicator of polynucleotide duplex forma-tion [334]. Using reversible electrochem-istry of DM, site-specifically coupled toan ODN, Kelley and coworkers [335] stud-ied long-range electron transfer through dsODNs anchored on a gold electrode via endthiol groups. It was shown that perfectlymatched DNA double helix effectively con-ducted electrons from DM bound to theODN, at a distant site relative to the elec-trode surface. Disturbing effects of single-base mismatches on electron transferthrough base stack (Sect. 12.8) could be de-tected by measuring voltammetric signalsof DNA intercalators (DM, methylene blueor a metallointercalator), but not of grooveor electrostatic binders [336]. Association

of another DNA-binding drug, distamycin,with a manganese porphyrin derivativeand binding of the complex to DNA wasfollowed by Rodriguez and Bard [337].Pandey and Weetal [338] proposed a tech-nique of electrochemical detection of DNAintercalation involving a photochemical re-action of anthraquinone derivatives. Theseintercalators were photochemically acti-vated and then reduced by an electrondonor, followed by electrooxidation of thereduction product at a modified CPE. Inthe presence of DNA, the anodic currentdecreased because of a limited transportof the DNA-bound anthraquinone to theelectrode. Aromatic amines (vigorous car-cinogens and important environmentalpollutants) were sensitively detected us-ing dsDNA-modified CPE. The techniqueproposed by Wang and coworkers [339]was based on selective accumulation ofthe analyte in the DNA sensing layer viaintercalative binding. By measuring DNAAC voltammetric signals at an HMDE,we [297] monitored DNA conformational


changes induced by various intercalators(Fig. 13a). In the presence of chloroquine,doxorubicin, 9-amino acridine, and Co, orRu metal chelates during DNA adsorptionon the electrode and subsequent removalof these substances, changes in DNA peaksindicating distortions of the DNA doublehelix and a reduced ability of the DNA to beunwound at the HMDE surface (Sect. 12.6)were observed.

A sensor for DNA-binding drugs wasdeveloped by Maeda and coworkers [340,341]. This technique is based on thiol-derivatized DNA-modified gold electrode.The DNA layer strongly suppressed CVresponse of the ferrocyanide/ferricyanideredox pair. On binding of cationic inter-calative drug (e.g. quinacrine or acridineorange), electrostatic repulsion betweensurface-confined DNA and the anionic de-polariser was decreased, and the electrodereaction became more feasible, display-ing increasing current response with thedrug concentration. Takenaka and cowork-ers studied electrochemical behavior ofDNA complexes with bis-intercalators (e.g.a bis-9-acridinyl derivative with a viologenlinker) [342] and threading intercalators(e.g., a ferrocenyl-modified naphthalenediimide) [343, 344]. Using the threadingintercalator, a sensor capable of discrim-ination between single-stranded, duplex,and hairpin DNA on the gold elec-trode surface and a sensor for DNAhybridization [345] was proposed. The fer-rocenyl naphthalene diimide derivativecomplexed with the surface-attached DNAfacilitated electron transfer from glucoseoxidase-catalyzed reaction to the elec-trode [346]. A fullerene derivative carry-ing a cationic intercalative moiety (pyri-dinium cation) exhibited a specific CVresponse on binding to duplex DNA [347].Brett and coworkers [348–352] developed

a GCE-based DNA sensor, useful in stud-ies of electrochemical behavior of drugs,such as metronidazole [348–350], nitroim-idazoles [351] or mitoxantrone [352]. A nat-ural quinoxaline antibiotic echinomycin(EM) was electrochemically active andyielded several CV signals applicablefor its determination at submicromolarconcentrations [F. Jelen, A. Erdem, andE. Palecek, unpublished]. Interaction ofEM with dsDNA attached to the HMDEresulted in specific DNA and EM signalsin agreement with the strong binding ofEM to dsDNA by bis-intercalation. Underthe same conditions, interaction of EMwith ssDNA resulted in high DNA, butvery small or no EM signals suggestedonly very weak binding of EM to ssDNAon the electrode surface. EM, thus, appearsto be a good candidate for a redox indica-tor in electrochemical DNA hybridizationsensors (Sect. 12.9.1).

12.7.2Covalent Interactions

Derivatization and complexation proce-dures are commonly used in electrochem-ical analysis if the analyte is inherentlynot electroactive [16, 353]. Natural NAs areelectroactive, but their oxidation and re-duction on electrodes is electrochemicallyirreversible taking place at highly positiveor highly negative potentials [10, 132, 353].

12.7.2.1 Electroactive Markers of NAsComplexation (see the preceding para-graph) and derivatization of NAs has beenused to obtain reversible electrode pro-cesses at less extreme potentials; catalyticsignals can afford higher sensitivity andimproved selectivity for the DNA structure.The first electroactive markers covalentlybound to DNA were investigated in the be-ginning of the 1980s [242, 354–358]. They


2

2

1 1

1Os

2 µA

200 nA

III

(b)

(a)

−0.75 −0.5

−1.0 −0.5

−0.25E[V]

E[V]

HN

N

O

O

R

CH3

O

Os

OO

O

N

N

I

Fig. 14 DNA modified with osmium tetroxide, 2, 2′-bipyridine (Os,bipy). (A), structure of the Os,bipy adduct with T residue. (B), AdTSDP voltammograms of (1), Os,bipy-modified DNA and (2),unmodified ssDNA; Os, catalytic peak yielded by the DNA adduct;III, peak due to reduction of A and C. Inset, cyclic voltammogram ofthe (1). modified and (2), unmodified DNA showing reversiblefaradaic peaks of the DNA-Os,bipy adduct. Background electrolyte:0.3 M ammonium formate, 50 mM sodium phosphate, pH 6.9.([242, 359]; R. Kizek, L. Havran, M. Fojta, and E. Palecek,Bioelectrochemistry, in press).

were based on osmium tetroxide com-plexes with nitrogen ligands (Os, L) andfulfilled all the above requirements [355,356, 359]; they produced a reversible cou-ple at about −0.6 V, yielded a catalyticsignal at about −1.2 V on the mercuryelectrode (Fig. 14), and showed useful se-lectivity for ssDNA. Os,L binds to C5-C6double bond of pyrimidines, showing astrong preference for Ts (Fig. 14). Osmiumtetroxide, 2, 2′-bipyridine (Os,bipy), and

some other Os,L are highly selective forssDNA [360, 361]; Os,bipy and other Os,Lcomplexes have been widely applied asprobes of the DNA structure, in vitro and invivo, in connection with nonelectrochem-ical methods, including DNA sequencingtechniques and immunoassays [299, 360,362, 363]. Most of the DNA-Os,L adductscan be determined electrochemically atppb concentrations [242]; in AdTSV exper-iments with 5-µl analyte volume tens of


pg are thus sufficient (corresponding toseveral femtomoles of a ss 20-mer ODN).

Quite recently, in an attempt to de-velop a universal, sensitive, and convenientmethod of DNA or RNA detection, elec-troactive oligonucleotides were preparedby covalent linkage of a ferrocenyl groupto the amino hexyl-terminated ODN [364].Using HPLC equipped with an electro-chemical detector, DNA and RNA weredetermined at femtomole level [364]. CVwith a Pt disk electrode showed a reversiblecouple around 0.4 V [364] in 50 mMKCl, 50 mM MgCl2 and 50 mM Tris-HCl (pH 8.0). To design a DNA biosen-sor Korri-Yousoufi and coworkers [365]synthesized an electroactive polypyrrolefunctionalized with ODN, which pro-duced an oxidation peak at −0.2 V as-sociated with the oxidation (or doping)of the polypyrrole chains (indicating thatthe ODN-substituted polypyrrole film washighly electroactive). (For further de-tails Sect. 12.9). Functionalization of otherconducting polymers-polythiophenes withindividual nucleobases yielded selectivenucleobase-responsive material potentiallyuseful in the development of DNA hy-bridization sensors [366].

12.7.2.2 Other Nucleic Acid ModificationsGold surfaces modified with thiol-derivati-zed ODNs were investigated by variousmethods [230, 311, 312, 367–370], mainlyin connection with the development of theDNA hybridization sensors (Sect. 12.9).The purpose of end labeling of ODNswith thiol or disulfidic groups was notto introduce in DNA an electroactivemarker but to attach DNA on gold viaa sulfur-gold linkage. On mercury sur-faces these groups can, however, producespecific signals and turn into efficientelectroactive markers of DNA [E. Palecekand L. Havran, unpublished data]. Many

biologically important substances bind co-valently to DNA [304]. In recent yearselectrodes were successfully applied instudies of DNA adducts with some anti-cancer drugs, such as mitomycin C andthiotepa, by Marin and coworkers [285,371–374]. Great potentialities of electro-chemical analysis in the research of a largenumber of DNA-drug adducts have been,however, little utilized.

Peptide (or polyamide) nucleic acid(PNA) is one of the candidates for di-agnostic and therapeutic applications inmedicine of the 21st century [375–377].In PNA the entire sugar-phosphate isreplaced by (N -(2-amino-ethyl) glycineunits. In contrast to DNA and RNA (withnegatively charged backbones), PNA hasan electrically neutral backbone. Electro-chemical responses of PNA were similarto DNA and RNA (i.e. A, C and Gwere reduced on mercury electrodes; Gproducing an anodic CV peak due to ox-idation of the G reduction product; andA and G were oxidized on carbon elec-trodes) [247]. Peak potentials of ssPNAat the HMDE were shifted to negativevalues as compared with ssDNA. Differ-ences in backbones of PNA and DNAwere manifested by the different adsorp-tion behavior of these two compounds, asdetected by a.c. impedance [108] on mer-cury (Fig. 3), and by chronopotentiometricmeasurements on carbon electrodes [129].At higher surface concentrations, associ-ation of PNA molecules on the mercurysurface was observed; prolonged exposureof PNA to highly negative potentials didnot result in PNA desorption under con-ditions, when almost all DNA moleculeswere removed from the surface. Adsorp-tion of PNA increased with decreasing saltconcentration; in contrast, adsorption ofDNA decreased under the same condi-tions [108]. In DNA hybridization sensors,


PNA has proved to be a better probe thanDNA [375] (Sect. 12.9).

12.8DNA Conductivity

Studies of DNA conductivity have a longhistory [378–382]. Recent attempts to drawconclusions about DNA electron transferfrom fluorescence quenching measure-ments on DNA strands containing donorand acceptor molecules have spurred adebate over the question of whether or notthe DNA double helix is able to conductelectrical charges (reviewed in [383–386]).Other approaches, including applicationof electrodes [335, 387–393], have alsobeen applied. Jortner and coworkers [394]and Okada and coworkers [395] suggestedpossible long-range carrier transport bya hopping mechanism. It was shown byGiese [396, 397] that when nucleobase rad-ical ions were formed within the DNAduplex, a long-range charge transfer thatrequired properly spaced G residues couldoccur. Importance of G residues, and par-ticularly of GG steps, was demonstratedby Schuster and coworkers [398–401].Okahata and coworkers [402] observedhigh anisotropic conduction in oriented,densely packed DNA helices along theirlongitudinal axis, but not in the plane per-pendicular to it. Passage of electron orhole current through a cluster of dsDNAmolecules with characteristics resemblingthose of a semiconductor was reported byFink and Schoneberger [403].

12.8.1Application of Electrodes in DNAConductivity Studies

Berg [18] (Sect. 12.6) and Barker [389–391]were probably the first who attempted to

draw conclusions about the DNA conduc-tivity from their polarographic/voltammet-ric studies. According to Barker [389–391],DNA loops extending to the solutionfrom chromosomal ssDNA adsorbed onthe mercury electrode may capture hy-drated photoelectrons that migrate alongthe DNA strands by hopping. This processis pH-dependent and occurs at weakly al-kaline pHs. Attachment of a thiol groupto one end of an ODN made it possibleto form SAMs of ds or ssDNAs on goldsurfaces [230, 311, 312, 368, 370]. Redoxproperties of methylene blue intercalatedin DNA helices attached to gold electrodesvia 5′-end thiol-terminated linkers wereinvestigated by Kelley and coworkers [369].To locate exactly the redox center withinthe duplexes, DM was covalently boundto G residue in 15-mer DNA duplexes,and efficient reduction of DM on a goldelectrode regardless of the DM positionin the duplex was observed [335, 392]. CAmismatch between DM and the electrodesurface completely abolished the electrore-duction of DM. These results indicatedthat electron transfer was blocked by per-turbation in base stacks induced by basemismatches, and supported the concept ofthe long-range (wire-like) charge transferwithin the π -stack of the DNA duplex. Itwas shown by Heller and coworkers [388]that in contrast to a conducting poly-mer the disordered chromosomal dsDNAfilm was not able to ‘‘wire’’ the soybeanperoxidase to induce electroreduction ofH2O2 to water. On the other hand, in or-dered solid 12-mer dsDNA film (alignedto the gold electrode via the end thiol)dsDNA displayed semiconductor charac-teristics that were disturbed by occurrenceof mismatched base pairs in the DNA du-plex and did not occur in ssDNA. It wasconcluded that randomly oriented dsDNAdid not conduct electrons or holes.

12.9 Analytical Applications 407

The results of Heller and cowork-ers [388], Okahata and coworkers [402],and Fink and Schoneberger [403] appearto be consistent with a doped ionic semi-conductor model, in which the cause ofcarrier mobility and donor ionization is thereduced difference between the static andhigh frequency longitudinal dielectric con-stants resolved in the direction of the longaxes of the helices in aligned dsDNA films.Porath and coworkers [393] measuredelectrical transport through individualpoly(dG)/poly(dC) molecules connected totwo metal nanoelectrodes. They obtained,in air as well as in vacuum, nonlinearcurrent-voltage curves that exhibited a volt-age gap at low applied bias, indicatinglarge-bandgap semiconductor behavior. Inrecent years, substantial progress in stud-ies of the DNA conductivity has beenmade, but many questions remain unan-swered. One of the most exciting discover-ies is the change in the electrical propertiesof dsDNA on introduction of a base mis-match, demonstrated by a number ofauthors using different approaches [230,275, 335, 375, 392, 397, 404–406]. Thisfinding, as well as future elucidation ofDNA conductivity, may have a tremendousimpact on our understanding of manybiological processes, including mutagen-esis, carcinogenesis, and DNA repair, andcan be utilized in biotechnologies for theconstruction of DNA sensors and variouselectronic devices.

12.9Analytical Applications

12.9.1Sensors for DNA Hybridization

DNA renaturation and hybridization, thatis, the ability of DNA to reform its

double-helical structure from its comple-mentary single strands, was discoveredabout 40 years ago by J. Marmur andP. Doty [407]. Shortly afterward, the abilityof electroanalysis to follow DNA renatura-tion was demonstrated [11, 408]. The prin-ciple of DNA hybridization was utilizedin many molecular-biological methods,including nucleotide sequencing. Rapidtesting of nucleotide sequences is requiredin different fields, including diagnosticsof various diseases, genetic testing, foren-sic medicine, rapid analysis of biologicalwarfare agents, environmental testing,and so on. Development of an inex-pensive, easy-to-use, fast response deviceremains the focus of interest for manyscientists. Compared with devices basedon optical transducers, the electrochem-ical devices are much cheaper, simpler,smaller, and more modest in their powerrequirements.

At present, the development of the elec-trochemical biosensors using ssDNA as arecognition layer is a rapidly developingfield [327, 334, 361, 375, 387, 409–413]. Insuch a biosensor, a short ssODN (DNAprobe) is immobilized on an electrode(transducer) to create the recognition layer.The probe-modified electrode is then im-mersed into a solution of target DNA totest its nucleotide sequence. When thesequence of target DNA exactly matchesthat of the probe DNA (based on thecomplementarity principle stating that Apairs with T and G with C), a hybrid(probe-target) duplex DNA is formed at theelectrode surface. The following sectionswill focus on two most important stepsin the detection of the DNA nucleotidesequence: the formation of the DNA recog-nition layer and hybrid duplex DNA, andthe transformation of the latter event intoan electrical signal.


12.9.1.1 Immobilization of DNA on theElectrodeBoth non-covalent and covalent bind-ings of the probe to the electrode sur-face have been used. Adsorption forceswere utilized for binding DNA to car-bon [414–416], mercury [108, 281, 326],and gold [322, 417] surfaces. Very strongbinding of DNA to mercury surfacesis due to hydrophobic interactions ofbases with the surface, preventing underusual conditions, efficient DNA hybridiza-tion [326]. Electrostatic binding of DNA tothe positively charged carbon electrode issufficiently strong and bases are accessi-ble for the specific interaction with targetDNA [129, 275, 410, 418]. Marazza andcoworkers [419] compared the probe im-mobilization by adsorption at controlledpotential with avidin-biotin probe immo-bilization on disposable graphite screen-printed electrodes, using DM as a redoxindicator. They concluded that simple ad-sorption is more reproducible than theavidin-biotin procedure and CPSA is moreefficient than DPV. Rather high concen-trations of target DNA were necessary forthe analysis (1–4 µg ODN ml−1).

Various kinds of covalent bindingof DNA to carbon, gold, mercury,and ITO [146, 327] surfaces were used(Sect. 12.4). Random covalent binding ofDNA to electrode surfaces involving chem-ical modification of bases [299] (such aswith carbodiimide derivatives, applied forimmobilization of the probe DNA tocarbon and other electrodes [409]) may de-crease the specificity of the recognitionlayer and thus cannot be recommended.Significantly, better results can be ex-pected with immobilization of the probevia one end of the DNA molecule notinvolving damage to DNA bases (suchas with thiolated ODNs which can eas-ily be immobilized to the gold [230, 311,

312, 335, 367–370, 387, 392] and mercury[L. Havran and E. Palecek, unpublished]surfaces). Thiol-terminated aliphatic link-ers are frequently used to place theprobing sequence at a proper distancefrom the electrode surface. Covalentlybound probes provide better possibilitiesfor easy removal of non-specifically boundmolecules. A microfabricated disposableDNA sensor, based on a gold electrode anda DNA probe with a mercaptohexyl groupat the 5′-end, for the detection of hep-atitis virus B DNA was introduced [420].Hoechst 33258 was used as an indicatordetecting 104 to 106 copies of DNA per ml.The sensor was applied for the analysisof DNA extracted from blood of patients.Various techniques of pretreatment of thegold electrode and thiol-linked probe im-mobilization involving tricky aspects havebeen employed in many papers [311, 332,421–423]. Recently, a procedure based ontreatment of the electrode with boilingKOH followed by concentrated nitric acid,has been developed and provides report-edly reproducible results [345].

12.9.1.2 Detection of the HybridizationEventIn the early studies, redox indicators inter-acting preferentially with dsDNA, such assimple intercalators (some metal interca-lators, DM, etc.) and minor-groove binders(e.g. Hoechst 33258), were applied to dif-ferentiate between ds and ssDNA [420,424, 425] (reviewed in [409, 410, 418]).These indicators interacted not only withdsDNA but also with ssDNA, and duplexformation was detected from the signalincrease (compared to the probe alone).Redox indicators with higher selectivity forduplex DNA are currently being sought.

Bis-intercalators and threading interca-lators are particularly interesting as redoxindicators. The latter intercalator, which


has substituents on opposite sites of theintercalating aromatic ring system, mustthread one of the substituents between thebase pairs at the intercalation site. Thesetypes of molecules usually have high DNAbinding constants, indicating that once thesubstituent slides between the base pairs,a very stable complex is formed. Nat-ural and synthetic bis-intercalators havetwo intercalating rings covalently linked

with a connecting chain. Recently, electro-chemical studies of the synthetic thread-ing intercalator ferrocenyl naphthalenediimide [426] and the naturally occurringnogalamycin [427] and bis-intercalator EM(antibiotic and antitumor agent) wereundertaken (Fig. 15). Both intercalatortypes bound to dsDNA more tightlythan usual intercalators, showing al-most no response with ssDNA-modified

(b) (c)

(a)

200 nAG

a

a

b

Ea

E C

E[V]

−0.3 −0.1

O OO

O

O

OO O

C

C

H

N

N C N CH

CH2

O

C N CH C N CH C

OO

H

H

CH3 O

Me

O

N CH

CH3

CH

C

O

CH3

O

CH3

HC

CNCHCNCHCNCHCO

H

CH3O

CH3

O

CH3

OCH

CH3 CH3

NC

ON

NH

CH3

D-Ser

L-AlMeValL-Ala

L-AlMeCysL-Ala

D-Ser

L-AlMeCys

R

Fig. 15 Chemical structures of (a) a bis-intercalator EM; (c) a threading intercalator ferrocenylnaphthalene diimide; (b) sections of cyclic voltammograms of EM with DNA. (a) EM+dsDNA;(b) EM+ssDNA. In contrast to a mixture of EM with ssDNA, complex of EM with dsDNA yields aspecific reversible pair of peaks (F. Jelen, A. Erdem and E. Palecek, Bioelectrochemistry, in press).


electrodes. Bis-9-acrinidyl derivatives con-taining linker chains with one or moreviologen linkers were synthesized [342,343, 346] as potential electrochemical indi-cators of dsDNA. Ferrocenyl naphthalenediimide was used to detect (dT)20, us-ing (dA)20 as a probe immobilized onthe gold electrode via adsorption [417].By this method, (dT)20 was detected atsub-attomole levels. Even higher sensi-tivity was obtained with thiolated probechemisorbed on a gold electrode using fer-rocenyl naphthalene diimide (Fig. 15) asa redox indicator [345]. By this approach,it was possible to detect the yeast cholinetransport gene in plasmid DNA. Studies ofthe ability of this method to discriminateDNAs with one or more mismatches areunder way.

12.9.1.3 Redox Indicators CovalentlyBound to DNAIn principle, target DNA can be cova-lently end-labeled with a redox indicatorproducing a signal on binding the tar-get DNA to the probe. It appears moreconvenient to bind a redox indicator toan ODN complementary to target DNA(reporter probe) on the site next to that hy-bridized to the probe [428]. Various kindsof covalent modifications of the target orreporter probe NAs can be used (Sect. 12.7)to detect the hybridization event throughthe signal of the electroactive marker.Electroactive markers producing catalyticsignals may yield high sensitivity for theanalysis, and thus appear very promising.ODNs labeled with ferrocene were used todetect DNA sequences by HPLC with elec-trochemical detection [429] at femtomolelevels. Combination of this techniquewith PCR [430], using ferrocene-modifiedODNs as primers, resulted in a sensitivityincrease down to the sub-attomole level.

These principles can be utilized in DNAsensor development.

12.9.1.4 Indicator-free Detection Systems.Intrinsic Electroactivity of DNAAny electrode can be used to distin-guish the probe and the target DNAs iftheir base contents are sufficiently dif-ferent. For example, the absence of anelectroactive G in the probe, but not inthe target, was used to detect DNA hy-bridization on ITO [146, 265, 266, 330,431] and carbon electrodes [326, 327, 432].To overcome the limitations of the probesequences (absence of G), guanines inthe probe sequence were substituted byhypoxanthine residues (pairing with C’s)and the hybridization detected throughthe target DNA G signal [327, 432]. Suchsubstitution may, however, decrease thestability of the duplex and the specificity ofthe hybridization.

12.9.1.5 Changes in Interfacial Propertiesand DNA ConductivityLarge differences between the interfacialproperties of ds and ssDNAs observedearlier by capacitance measurements [10,37] suggested that a.c. impedance mea-surements could be used to detect DNAhybridization on electrodes [433, 434](Sect. 12.8.). A three-component ODN sys-tem on a gold electrode (involving avidin-biotin interactions) was used to detect spe-cific DNA sequences by means of faradaicimpedance spectroscopy [435]. Impedancespectroscopy does not seem, however, tobe the most convenient method for theDNA biosensor; faster and simpler voltam-metric or chronopotentiometric methodswill probably be more convenient. Con-ductivity of the perfect DNA, contrastingwith a loss of conductivity in duplexeswith mismatched bases, may be of use in


the development of the DNA hybridizationsensors (Sect. 12.8).

12.9.1.6 Blocking and Interfacing theTransducerIn contrast to model studies using ODNsas target DNAs, in experiments with realDNA samples, nonspecific interactionson the electrode surfaces (involving DNAimpurities, long ssDNA chains, etc.) canobscure the hybridization signals. Efficientinterfacing between the DNA system andthe electrode surface is thus necessary.Thiols and conducting polymers have beenused for this purpose.

Individual chains of organic conju-gated polymers possess a high intrinsicconductivity [436, 437]. These conductingpolymers can be synthesized chemically(for example, polyphenylenes) or elec-tropolymerized as thin films onto anelectrode (e.g. polythiophenes and polypyr-roles). Experimentally obtained conductiv-ities of conjugated polymers are between 1and 104 S/cm, several orders of magnitudelower than those of metals, but sufficientlyhigh to consider these macromolecularchains (several micrometers in length)as molecular wires [436]. Deposited asthin films on electrodes, the reversibleredox processes in these polymers canbe potential-controlled and cyclic voltam-mogram can provide an electrochemicalsignature of the given polymer. This sig-nature is very sensitive to the nearbyenvironment; any modification of the pen-dent groups along the chain is electricallytransduced to the electrode and manifestedby modified polymer electrochemical sig-nature. Different functionalities can beinserted in the polymer known to showselectivity for solution species such asantibodies, enzymes, and NAs. Specificchemical recognition resembling affinitychromatography can thus be built through

proper functionalization of the conductingpolymers.

Conducting polymers, such as copoly-mer functionalized with an osmium com-plex [438–442], polyazines, polyanilines,polypyrroles [365, 413, 436, 443], and poly-thiophenes [366] (reviewed in [444–446]),may be used for blocking and interfac-ing the transducer, for modulation ofthe DNA interactions at surfaces and forgenerating signals resulting from suchinteractions. Conducting polypyrrole func-tionalized with bulky ODN remainedelectroactive in aqueous media; on in-teracting with the complementary butnot the noncomplementary ODN, a de-crease in the voltammetric current at−0.2 V was observed [365]. Adsorption ofNAs on a polypyrrole-coated GCE wasutilized for amperometric ODN detec-tion in flowing streams [443]. Doping ofthe NA probes within electropolymer-ized polypyrrole films and monitoringthe current changes provoked by the hy-bridization event appears to be a promisinglabel-free biosensing strategy [447].

Two component films were prepared,containing in addition to a thiol-derivedssDNA probe a diluent thiol, mercapto-hexanol, to prevent nonspecific adsorptionof ssDNA on the gold surface [423]. Thedielectric constant and thickness of thefilm were measured by two-color surfaceplasmon resonance and the amount ofDNA tethered to the surface was quan-tified. The kinetics of hybridization andthermally induced dehybridization weremeasured, indicating a high efficiency ofthe hybridization process. Self-assembledmonolayer containing a viologen groupwas formed on a gold electrode via gold-S bonds [448]. Binding of dsDNA to thislayer resulted in a positive shift of the redoxpotential of the viologen centers, indicat-ing hydrophobic interactions. Alkanethiol


monolayers can easily be assembled notonly on gold but also at mercury sur-faces [449–453].

12.9.1.7 Electrocatalytic ReactionsElectrode processes in experiments withredox indicators involved one or a few elec-trons and were, therefore, inherently low-yield reactions. Recently catalytic processeshave been used to collect as many electronsas possible ([327, 454, 455]; E. Palecek,M. Fojta, and L. Havran, unpublished).Thorp [327] used a soluble mediator thatmoved close to G residues present onlyin target DNA (but absent in the probe)and shuttled electrons to the polymer-modified ITO electrode. The reducedform of the mediator [Ru(bipy)3]2+ wasoxidized by holding the electrode at a suf-ficiently positive potential. The oxidizedform of the mediator removed electronsfrom G residues, generating reduced[Ru(bipy)3]3+ and completing a catalyticcycle. About 100 electrons per hybridizedG could be collected under favorable con-ditions. Horseradish peroxidase coupledto target DNA was applied to detect thehybridization by electrocatalytic reductionof hydrogen peroxide

H2O2 + 2e− −−−→ 2OH− (2)

The enzyme molecule turned over about1800 times per second, producing about3600 electrons in 1 s [456]. This approachwas also used in connection with thedetection of point mutations (see below).

12.9.1.8 Detection of Point MutationsMany diseases are connected with a singlebase mutation (point mutation) at spe-cific sites of the genome. Detection of achange in a single nucleotide in the DNAis rather difficult and requires highly spe-cific methods. The first electrochemical

detection of a point mutation (single basemismatch) was achieved by using PNAprobe instead of DNA [275]. This methodwas applied to detect the mutation hotspot in the p53 gene [406]. A relativelysimple technique using a CPE and a sim-ple redox indicator was used to obtainthese results. Caruana and Heller [455]detected a single base mismatch in an 18-mer ODN, using a redox polymer-coatedmicroelectrode and thermostable soybeanperoxidase-labeled target DNA. They ob-tained excellent discrimination between aperfectly matched duplex and a single basemismatch at elevated temperature, whileat room temperature such discriminationwas not possible. Specificity and sensitivityare, perhaps, the most important featuresof a DNA hybridization sensor. Caruanaand Heller [455] offered a method supe-rior in both sensitivity and specificity,representing significant progress towardsa practical DNA hybridization sensor.

12.9.2Sensors for DNA Damage

DNA damage may cause serious distur-bances of cell life. Chemical changes inthe DNA bases may lead to alterations inbase pairing followed by mutations [304].Formation of DNA double-sb’s frequentlyresults in mitotic faults and chromosomeaberrations. Accumulation of mutationsand/or other kinds of DNA damage repre-sents serious carcinogenic and teratogenicrisks. Development and improvement ofanalytical techniques, capable of rapidand sensitive detection of various typesof DNA damage (including DNA biosen-sors) is, therefore, the focus in manylaboratories. The sensitivity of electro-chemical methods to changes in DNAstructure has been utilized in these in-vestigations. Three basic approaches can


be distinguished: (1) detection of DNA sbs(Fig. 8), (2) detection of base damage, and(3) detection of substances that specifi-cally interact with DNA (covalently and/ornon-covalently) that are electroactive andyield specific electrochemical signals. Elec-trochemical detection is also frequentlyemployed in methods of the determinationof modified nucleobases (such as 8-OG orother adducts) based on DNA hydrolysis,followed by HPLC separation [457]; theseapproaches are, however, out of the scopeof this chapter.

12.9.2.1 Detection of DNA Strand BreaksGenerally, DNA sb’s are formed by thedirect hydrolysis (enzymatic or chemical)of the phosphodiester bond, on damageto the deoxyribose moiety (usually after aradical attack), or as a consequence of somekinds of chemical damage to DNA basesfollowed by a destabilization of the DNAsugar-phosphate backbone [304]. In 1960sand 1970s we showed that DNA sb couldbe detected using DPP. The polarographicpeak II (Sect. 12.4) increased markedly ontreatment of dsDNA with DNase I [458],with ionizing radiation and/or with ul-trasound [458, 459]. Later, voltammetrictechniques, including LSV [460, 461] ora.c. voltammetry [462], were used to studyelectrochemical behavior of sonicated orγ -irradiated DNA. A highly sensitive DNAbiosensor has been developed in our labo-ratory [272, 273, 276, 463, 464]. This sensoris based on qualitative differences be-tween the behavior of sc DNA, and DNAmolecules containing free ends (oc, seeFig. 8, or linear). On mercury electrodes,the latter DNAs produce non-faradaic (ten-sammetric) peak 3 and a well-developedfaradaic peak CA (Sect. 12.4), because ofpartial unwinding of the double helix onthe electrode surface in the vicinity of thestrand ends (Sect. 12.6). On the contrary,

in scDNA lacking sbs, the nucleobases can-not interact with the electrode surface thatresults in the absence of peak 3 and in re-stricted electroreducibility of C and A at theHMDE (Sects. 12.3 and 12.4; Fig. 8). Thismethod made possible the detection of asingle chain interruption among 2.5 × 105

phosphodiesteric bonds [272]. The HMDEand MFE [252] modified with adsorbedscDNA have been used for the detectionof hydroxyl radicals formed through Fen-ton chemistry in laboratory-prepared solu-tions, and in various samples of naturaland industrial water or food [273]. ScDNAcan also be cleaved at the HMDE sur-face enzymatically. Deoxyribonuclease Icleaved electrode surface-confined scDNAin a remarkably potential dependent man-ner [463]. On the contrary, measurementsof peak Gox on carbon electrodes exhibitedno significant differences between the sig-nals of sc and linear or oc DNAs [252, 257](Fig. 8). A deep degradation of DNA dueto acid hydrolysis [248] or oxidative dam-age mediated by copper complexes [306]was required to observe increased inten-sity of DNA signals on carbon electrodes.Such changes were also accompanied bya decreased binding of an intercalativeredox indicator [306], suggesting a signif-icant loss of double-stranded character ofthe DNA.

12.9.2.2 Damage to DNA BasesChemical damage to DNA base residuesmay change (or abolish) their intrinsicelectroactivity (Sect. 12.4). Moreover, dam-aged bases can be released from DNAmolecules, which may be monitored bytransfer stripping techniques (Sect. 12.4)and/or by CSV (Sect. 12.2) [248]. Peak Gmeasured at the HMDE (Sect. 12.4) wasused as the transduction signal in sys-tems detecting DNA alkylating or acylatingagents. Methylation of the DNA G residues


by dimethyl sulfate resulted in a decreaseof peak G [138]. ssDNA was modified, ei-ther in solution followed by AdTS (DNAwas separated from the reaction mixture bywashing the DNA-modified HMDE) or onthe electrode surface (intact DNA was im-mobilized at the HMDE and the electrodewas immersed in dimethyl sulfate for 20 to30 min). Decrease of peak G was observedalso in DNA interacting with mitomycin C,a drug attacking primarily G residues [285,372, 374, 465], and in another antineo-plastic drug thiotepa [371]. On the basisof a similar principle, the DNA-modifiedgraphite electrode was used as a sen-sor for the detection of antitumor plat-inum drugs [466]. Using DNA-modifiedCPE, Wang and coworkers [467] devel-oped a highly sensitive (at µg L−1 levels)biosensor for the detection of hydrazinederivatives usable in environmental anal-ysis. Interaction of the hydrazines withimmobilized DNA resulted in a decreaseof peak Gox. Thorp and coworkers [330]utilized differences in the redox potentialsof G and 8-OG (a highly abundant G oxida-tion product formed in vivo upon oxidativestress) in the development of a methodfor the detection of 8-OG. [Os(bipy)3]3+complex was capable of shuttling electronsfrom 8-OG (but not from unmodified G)to the ITO electrode, exhibiting electrocat-alytic effect (Sect. 12.7). Changes in DNAelectrochemical behavior due to damage toDNA bases by UV radiation were detectedby DPP [468], and by CPSA using a screen-printed electrode-based DNA sensor [469].Some kinds of base damage or abasic sitescan be converted into sb or ssDNA re-gions by using DNA repair enzymes. We[M. Fojta and E. Palecek, unpublished] de-tected apurinic sites in DNA using E. coliexonuclease III, an enzyme that cleaves ds-DNA at abasic sites followed by exonucle-olytic degradation of one DNA strand. The

resulting ssDNA regions were sensitivelydetected by measuring AC voltammetricpeak 3 (Sect. 12.3). To detect base dam-age, mismatch-sensitive DNA hybridiza-tion techniques can be used [329, 336, 405].

12.9.2.3 Detection of Damaging AgentsSpecifically Interacting with DNACovalent adducts and/or noncovalent com-plexes of DNA with some chemicals, in-cluding carcinogens or cytostatics, producespecific electrochemical signals. Appear-ance of specific peaks was observed onDNA interactions with mitomycin C, os-mium tetroxide complexes, DM, aromaticamines, and a variety of other DNA binders(Sect. 12.7).

12.9.2.4 DNA Cleavage Controlled byElectrochemical ReactionsIn vivo, many DNA-damaging processesinvolve redox reactions. For example,redox-active metals, such as copper, ironor manganese, can mediate reactionsyielding reactive oxygen species that arevigorous DNA-damaging agents. Elec-trodes have been utilized to control re-dox states of the metals and to modu-late subsequent DNA damage. Rodriguezand coworkers [470] electrochemically acti-vated Mn(III) and Fe(III) complexes with aporphine derivative in the presence of oxy-gen. As a result, cleavage of sc DNA presentin the bulk of the solution took place, andwas detected by means of gel electrophore-sis. We [464] electrochemically modulatedcleavage of scDNA adsorbed at the sur-face of the HMDE in the presence of ironor copper complexes and of chromiumcompounds, using in situ electrochemi-cal detection of the DNA sbs formation(see above). Both laboratories [464, 470]detected a certain extent of DNA cleavagein the absence of the metal complexes,suggesting a role of oxygen reduction


products on the platinum and mercuryelectrodes. Thorp’s group [328] generatedboth single-sbs and piperidine-labile sitesby controlled-potential electrolysis of DNA(sc plasmid or an oligonucleotide, respec-tively) in the presence of trans-[Re(O)2(4-OMe-py)4]+. Piperidine cleavage revealeddamage specifically to G residues, suggest-ing a primary mechanism of one-electronoxidation of G mediated by the metal com-plex (Sect. 12.7).

12.9.3Other Determinations

12.9.3.1 Determination of ssDNA in anExcess of dsDNAThe specificity and high sensitivity of DPpolarographic peak III to ss (denatured)DNA (Sect. 12.4) was utilized in methodsof determination of trace amounts of ss-DNA in an excess of ds (native) DNA [280].An analogous approach was applied forthe determination of ssRNA in excess ofbacteriophage f2 sus 11 double-strandedRNA [471]. DPP measurements were ableto detect <1% of ssDNA in dsDNA. More-over, using DPP peak III intramolecular ssregions generated by exonuclease III (anenzyme degrading one strand of dsDNA)in dsDNA molecules were detected [472].CV or AC voltammetric measurementsfollowing heat treatment of plasmidDNAs were employed to determine smallamounts of the nicked (oc) form of theplasmid in excess of scDNA [272, 473]. At92C, only the nicked (or linear) moleculeswere irreversibly denatured, but not cova-lently closed sc circles. The amount ofocDNA was calculated from the amountof ssDNA (which appeared after the sam-ple heating) manifested by an increase ofCV peak G [473] or tensammetric peak3 [272]. The same principle was applied,

for example, to monitor nicking of scDNAupon ionizing irradiation.

12.9.3.2 Determination of RNA Traces inDNA SolutionsDNA samples are usually contaminatedby small portions of RNA and determina-tion of the RNA content in DNA samplesis frequently necessary. Optical methodsused for this purpose have low sensitivityand specificity. Recently, highly sensi-tive and specific electrochemical methodshave been proposed [269, 270]. NaturalRNAs produced qualitatively the same re-dox response as DNA at the HMDE [271](Fig. 16a), MFE [251], and carbon elec-trodes [255] at neutral and weakly acidicpH’s. On the other hand, in weakly alkalinemedia RNAs yielded a single tensammet-ric peak R (at about −1.3 V, Fig. 16),whose potential differed from peak po-tentials of both DNA peaks (peak 1 at−1.2 V and peak 3 at −1.45 V; in thereferences [269–271], peak 1 and peak 3are denoted as peak D1 and peak D2,respectively) [269–271] (Fig. 16b). This dif-ference made it possible to detect simul-taneously DNA and RNA in a mixture,and to determine traces of RNA in sam-ples of DNA. Using AdTS DPV at theHMDE, picomole amounts of tRNA ortotal bacterial RNA were determined inalmost 200-fold excess of native chro-mosomal or plasmid DNA [270]. Such adirect approach was, however, ineffectivewhen RNA in more than tenfold excessof single-stranded DNA was to be deter-mined (e.g. DNA of some viruses andbacteriophages such as M13 are natu-rally single-stranded). To overcome thislimitation, we developed [269] a procedurecombining selective precipitation of ss ordsDNA by [Co(NH3)6]3+ and AdTS DPV.Short molecules of RNA with globulartertiary structure (e.g. tRNAs) remained


(a)

(c)

(b)

R

G G0.2 µA

20 nA

ssDNA

ssDNA

500 mV

CA

−1.5 −0.5 −1.5 −0.5

CA

R 3

31NaCl [Co[NH3]6]Cl3

R

tRNA

tRNA

tRNA+ssDNA

E[V]

E[V]

Fig. 16 (a) AdTS cyclic voltammograms of yeast transfer RNA(tRNA, left) and of ssDNA (right) obtained in 0.3 M ammoniumformate, 50 mM sodium phosphate, pH 6.9. Potentials of peak CAand peak G of DNA and DNA do not differ. (b) AdTS DPvoltammograms of tRNA (top) and ssDNA (bottom) obtained in0.3 M sodium bicarbonate, pH 9.6; under these conditions, tRNAyields a single capacitive peak R at potential differing form those ofDNA peaks 1 and 3. (c) AdTS DP voltammograms of a 1 : 8 (w/w)mixture of tRNA + ssDNA obtained at pH 9.6 after adsorption of theNAs from 0.2-M NaCl (left) or 50-mM [Co(NH3)6]Cl3. DNA isselectively precipitated by the cobalt complex. (Adapted fromM. Fojta, R. Doffkova, E. Palecek, Electroanalysis 1996, 8, 420–426;E. Palecek, M. Fojta, Anal. Chem. 1994, 66, 1566–1571; M. Fojta,C. Teijeiro, E. Palecek, Bioelectrochem. Bioenerg. 1994, 34,69–76.)

in solution (yielding peak R), while longfilamentous DNA molecules were pre-cipitated on addition of the cobalt com-plex (Fig. 16c). Possible coprecipitation ofpolydisperse total cellular RNA was elim-inated by partial degradation of RNA intoshort oligonucleotides prior to addition of[Co(NH3)6]3+ [269].

12.9.3.3 Determination of ProteinsKnowing the protein content is impor-tant in characterizing the purity of the

NA sample. DPP determination of cys-tine/cysteine containing proteins based oncatalytic currents in cobalt solutions [474]was, for many years, one of the mostsensitive nonradioactive methods of pro-tein determination. Using this technique,it was possible to determine about 1%of protein in a few micrograms ofDNA [475]. Recently described methodsof protein determination, by means ofCPSA on mercury [249] and carbon elec-trodes [476], will probably offer even bettersensitivities.

12.10 Conclusion 417

12.10Conclusion

Since its beginning by the end of the 1950sand thereafter for about two decades, elec-trochemical analysis of NAs developed asa typical basic research activity. In thattime, fundamental information about theadsorption/desorption properties, as wellas reduction and oxidation of DNA atthe mercury and solid electrodes, was ob-tained [4, 5]. High sensitivity of mercuryelectrodes for changes in DNA structurewas utilized in biochemical analysis [5, 11].In the 1980s, this research was almostabandoned because the sensitivity of theconventional voltammetric analysis wastoo low to compete with methods such asgel electrophoresis, commonly used in theNA laboratories. In that decade, only a fewlaboratories continued their electrochem-ical studies of NAs. Nevertheless, it wasduring that period when new electrochem-ical approaches were designed, resultingin a tremendous increase in the sensitivityof the DNA electrochemical analysis, inthe invention of the DNA-modified elec-trodes [281, 288], and in the introductionof the first electroactive markers intothe DNA molecules [242, 354–358]. Thesenew approaches represented the necessarybackground for development of DNA sen-sors. In the first half of the 1990s, morelaboratories entered the field in an attemptto develop sensors for DNA hybridiza-tion [409]. Although some of the earlypapers suffered from a lack of knowledgeof the DNA chemistry, they, neverthe-less, represented an important stimulusin the development of the field. Since themid-nineties, the creation of an electro-chemical DNA hybridization sensor hasbecome a reality, and in recent years wehave witnessed an increased interest inbiochemists and biotechnology companies

in DNA electrochemistry and sensor devel-opment [477–480]. Attempts to exploit theresults of the basic research for commer-cial purposes represented, on one hand,another stimulation of DNA electrochem-istry research; and on the other, resultedin delay in the publication of a few pa-pers because of the patenting of newprinciples and procedures. At present,there are indications that the first simpleelectrochemical DNA hybridization sensorapplicable for practical purposes might beavailable in the near future.

In recent years, attempts have beenmade to develop microfabrication tech-nologies for integrated NA analysis [477–483]. Electrodes have been used, not onlyin the development of the DNA sensors butalso in developing bioelectronic chips forsample preparation, and this may becomea part of integrated systems for DNA hy-bridization. Using such a system, bacterialcells can be separated by means of di-electrophoresis and subjected to electroniclysis, followed by proteolytic digestion ona single bioelectronic chip [477, 483]. NAsreleased from the cells can then be testedon a separate DNA hybridization chip.Miniaturized DNA chips, incorporating alarge number of oligomer probes or evenlonger DNAs (500 to 5000 nucleotides inlength), to increase the accuracy and ca-pacity of the sensor were created. Highprobe densities of small silicon chips wereachieved up to approximately 105 sitescm−2 [436, 484]. In addition to silicon,glass, and some polymers with flat orporous surfaces were used as substrates.ODN probes were anchored on the solidsubstrates in various ways, including poly-acrylamide gels; deposited as small dotson a glass substrate [485], and photochem-ically addressed as sensing dots, associatedwith in situ synthesis of the desired ODNprobe [484, 486]. Potentialities of the DNA


chips have already been largely confirmed,but detection of the hybridization eventstill raises some questions. In fact, the cur-rently used DNA chips require fixation ofa fluorescence or radioactive tag to the tar-get DNA, costing time and effort, and thesubsequent reading of the chip requiresexpensive equipment. It is believed thatapplication of electrochemical principlesdescribed in this review may offer a simpleand effective solution to this problem.

Addendum

After finishing this paper an importantprogress in the development of the electro-chemical DNA hybridization sensors wasachieved [487–520] bringing the researchcloser to its goal, that is to the design ofa new type of a DNA biosensor for theanalysis of real DNA samples. This re-search is in part reflected in the specialissue of Talanta devoted to electrochem-istry of NA’s and DNA biosensors [497].Most of the papers on the developmentof the electrochemical DNA hybridizationsensors discussed in Sect. 12.9.1 dealt withmodel systems using relatively short targetODNs, while experiments working withreal DNA samples were rather rare becauseof problems with nonspecific adsorption ofnoncomplementary (and therefore nonhy-bridized) DNA. All systems relied on solidelectrodes with immobilized short single-stranded probe DNA, on which both thehybridization and detection steps were per-formed [487]. Recently a new method hasbeen proposed [488–490, 513] in which theDNA hybridization is performed at onesurface (surface H, optimized for DNA hy-bridization; commercially available mag-netic beads were used as surface H) andelectrochemical detection on another sur-face, the detection electrode (DE). Owingto minimum nonspecific DNA adsorption

at the surface H very high specificity of theDNA hybridization can be reached. The DEcan be chosen only with respect to the elec-trode process securing high sensitivity ofthe analysis. High sensitivity in the detec-tion of relatively long target DNAs has beenobtained (1) by using label-free methods,namely, cathodic stripping voltammetry atmercury or solid mercury amalgam elec-trodes for the determination of purinebases, released from DNA by acid treat-ment [488, 489] or oxidation of DNA gua-nine residues at carbon electrodes [491]and (2) by enzyme-linked immunoassay oftarget DNA modified by osmium tetroxide,2, 2′-bipyridine (DNA-Os,bipy) at carbonelectrodes [490] (3) by enzyme-linked sand-wich assay based on streptavidin-biotininteractions [492] and (4) by direct deter-mination of DNA-Os,bipy at mercury orcarbon electrodes [493, 494], and (5) byprecipitation of silver on gold nanoparti-cle tags followed by CPSA determinationof the silver at carbon electrodes [495].Studies of DNA hybridization kinetics ongold electrodes by optical surface plasmonresonance showed that the d.c. field can en-hance or retard DNA hybridization and canalso denature surface-immobilized DNAduplexes [496]. The latter phenomenonmay correspond to the opening of theDNA duplexes summarized in Sect. 12.6and can be utilized in the developmentof the DNA hybridization sensors. Incor-poration of the above principles into amicrofluidic system may soon result in anew device for the sequence analysis oflong natural DNA molecules.

Acknowledgment

This work was supported by a grant ofthe Grant Agency of the Czech RepublicNo. 204/97/K084 (to E.P.), by grants ofthe Grant Agency of the Academy of


Sciences of the Czech Republic Nos.S5004107 (to V.V.), A4004801 (to M.F),A4004901 (to F.J.), and A4004002 (toV.V.) and by institutional grants K4055109and Z5004920. The authors are gratefulto Drs. M. Heyrovsky and Z. Pechan forcritical reading of the manuscript.

References

1. H. Berg, Biochem. Z. 1957, 329, 274–276.2. E. Palecek, Naturwissenschaften 1958, 45,

186, 187.3. E. Palecek, Nature 1960, 188, 656, 657.4. E. Palecek, in Progress in Nucleic Acid

Research and Molecular Biology (Eds.: J. N.Davidson, W. E. Cohn), Academic Press,New York 1969, pp. 31–73, Vol. 9.

5. E. Palecek, in Methods in Enzymology:Nucleic Acids, part D (Eds.: L. Grossman,K. Moldave), Academic Press, New York,1971, pp. 3–24, Vol. 21.

6. R. Kalvoda, Techniques of Oscillographic Po-larography, Elsevier, New York, 1965.

7. J. Heath, Nature 1946, 158, 23.8. E. Palecek, B. Janik, Arch. Biochem. Biophys.

1962, 98, 527.9. E. Palecek, Collect. Czech. Chem. Commun.

1960, 25, 2283–2289.10. E. Palecek, in Topics in Bioelectrochemistry

and Bioenergetics (Ed.: G. Milazzo), John Wi-ley & Sons, Chichester, 1983, pp. 65–155,Vol. 5.

11. E. Palecek, Prog. Nucleic Acid Res. Mol. Biol.1976, 18, 151–213.

12. H. W. Nurnberg, P. Valenta, in Ions inMacromolecular and Biological Systems (Eds.:D. H. Everett, B. Vincent), Scientechnica,Bristol, 1977, pp. 201–236.

13. E. Palecek, Bioelectrochem. Bioenerg. 1986,15, 275–295.

14. E. Palecek, in Encyclopedia of AnalyticalScience (Ed.: A. Townshend), AcademicPress, London, 1995, pp. 3600–3609, Vol. 6.

15. E. Palecek, Electroanalysis 1996, 8, 7–14.16. J. M. Sequaris, in Wilson and Wilson’s

Comprehensive Analytical Chemistry (Eds.:G. Svehla, J. G. Vos), Elsevier, Amsterdam,1992, pp. 115–156, Vol. XXVII.

17. H. Berg, in Topics in BioelectrochemistryBioenergetics (Ed.: G. Milazzo), John Wiley& Sons, London, 1976, pp. 39–104, Vol. 1.

18. H. Berg, in Comprehensive Treatise of Elec-trochemistry of Biopolymers Vol. 10. Elec-trochemistry of Biopolymers (Eds.: S. Srini-vasan, Y. Chizmaddzhev, J. Bockris et al.,),Plenum, New York, 1985, pp. 189–229.

19. R. Berezney, D. S. Coffey, Biochem. Biophys.Res. Commun. 1974, 60, 1410–1417.

20. E. Neumann, in Ions in Macromolecularand Biological Systems (Eds.: D. H. Everett,B. Vincent), Scientechnica, Bristol, 1978,pp. 170–191.

21. S. M. Gasser, U. K. Laemmli, Cell 1986, 46,521–530.

22. D. Porschke, M. Jung, J. Biomol. Struct.Dyn. 1985, 6, 1173–1184.

23. E. Neumann, Prog. Biophys. Mol. Biol. 1986,47, 197–231.

24. E. Neumann, Bioelectrochem. Bioenerg. 1986,16, 565–567.

25. S. Trasatti, A. Petrii, Pure Appl. Chem. 1991,63, 711–734.

26. V. Vetterl, Experientia 1965, 21, 9–11.27. V. Vetterl, Collect. Czech. Chem. Commun.

1966, 31, 2105–2126.28. V. Vetterl, Abhandlungen der DAW, Berlin

1966, 4, 493–500.29. V. Vetterl, E. Kovarikova, R. Zaludova, Bio-

electrochem. Bioenerg. 1977, 4, 435–444.30. J. Jursa, V. Vetterl, J. Electroanal. Chem.

1989, 289, 237–244.31. V. Brabec, V. Vetterl, O. Vrana, in Ex-

perimental Techniques in Bioelectro-chemistry (Eds.: V. Brabec, D. Walz,G. Milazzo), Birkhauser Verlag, Basel, 1996,pp. 287–359, Vol. 3.

32. R. deLevie, Chem. Rev. 1988, 88, 599–609.33. B. Janik, P. J. Elving, J. Am. Chem. Soc.

1970, 92, 235–243.34. V. Drazan, V. Vetterl, Collect. Czech. Chem.

Commun. 1998, 63, 1977–1993.35. M. S. Ibrahim, M. E. Ahmed, A. M. Kawde

et al., Analysis 1996, 24, 6–9.36. M. E. Ahmed, M. S. Ibrahim, Y. M. Temerk

et al., Electrochim. Acta 1996, 41, 2883–2892.37. I. R. Miller, J. Mol. Biol. 1961, 3, 229–240.38. I. R. Miller, J. Mol. Biol. 1961, 3, 357–361.39. M. Rueda, A. Mota, M. L. S. Goncalves

et al., J. Electroanal. Chem. 1997, 431,257–267.

40. H. Jehring, Elektrosorptionalyse mit der Wech-selstrompolarographie., Akademie, Berlin,1974.

41. V. Vetterl, J. Electroanal. Chem. 1968, 19,169–173.


42. V. Vetterl, Biophysik 1968, 5, 255–260.43. C. Buess-Herman, Prog. Surf. Sci. 1994, 46,

335–375.44. C. Buess-Herman, C. Franck, L. Gierst,

J. Electroanal. Chem. 1992, 329, 91–103.45. J. Lipkowski, C. Buess-Herman, J. P. Lam-

bert et al., J. Electroanal. Chem. 1986, 202,169–189.

46. M. Scharfe, C. Buess-Herman, J. Elec-troanal. Chem. 1994, 366, 303–310.

47. V. Vetterl, J. Pokorny, Bioelectrochem. Bioen-erg. 1980, 7, 517–526.

48. D. Krznaric, P. Valenta, H. W. Nurnberget al., J. Electroanal. Chem. 1978, 93, 41–56.

49. P. Valenta, D. Krznaric, J. Electroanal. Chem.1977, 75, 437–454.

50. V. Brabec, S. D. Christian, G. Dryhurst,J. Electroanal. Chem. 1977, 85, 389–405.

51. Y. M. Temerk, M. M. Kamal, M. E. Ahmedet al., Bioelectrochem. Bioenerg. 1986, 16,497–507.

52. C. Mousty, G. Quarin, Electrochim. Acta1990, 35, 1291–1302.

53. U. Retter, J. Electroanal. Chem. 1982, 136,164–167.

54. U. Retter, H. Lohse, J. Electroanal. Chem.1982, 134, 243–250.

55. U. Retter, V. Vetterl, J. Jursa, J. Electroanal.Chem. 1989, 274, 1–9.

56. T. Wandlowski, M. Heyrovsky, L. Novotny,Electrochim. Acta 1992, 37, 2663–2672.

57. U. Retter, J. Electroanal. Chem. 1980, 106,371–375.

58. R. Sridharan, R. deLevie, J. Electroanal.Chem. 1987, 218, 287–296.

59. V. Vetterl, R. de Levie, J. Electroanal. Chem.1991, 310, 305–315.

60. V. Brabec, M. H. Kim, S. D. Christian et al.,J. Electroanal. Chem. 1979, 100, 111–133.

61. V. Brabec, S. D. Christian, G. Dryhurst,Bioelectrochem. Bioenerg. 1978, 5, 635–649.

62. V. Brabec, S. D. Christian, G. Dryhurst,Biophys. Chem. 1978, 7, 253–268.

63. G. Quarin, Electrochim. Acta 1984, 29,1707–1714.

64. N. J. Tao, Z. Shi, J. Phys. Chem. 1994, 98,7422–7426.

65. Z. A. Ahmed, M. E. Ahmed, M. S. Ibrahimet al., Bioelectrochem. Bioenerg. 1995, 38,359–365.

66. M. M. Correia dos Santos, P. M. P. Sousa,A. M. M. Modesto et al., Bioelectrochem. Bio-energ. 1998, 45, 267–273.

67. R. Guidelli, M. L. Foresti, J. Electroanal.Chem. 1986, 197, 103–121.

68. E. V. Stenina, B. B. Damaskin, J. Elec-troanal. Chem. 1993, 349, 31–40.

69. L. Pospisil, M. Svestka, J. Electroanal. Chem.1997, 426, 47–53.

70. T. Wandlowski, L. Pospisil, J. Electroanal.Chem. 1989, 258, 179–192.

71. H. Francois, M. Scharfe, C. Buess-Herman,J. Electroanal. Chem. 1990, 296, 415–428.

72. M. Scharfe, A. Hamelin, C. Buess-Herman,Electrochim. Acta 1995, 40, 61–67.

73. T. Wandlowski, R. de Levie, J. Electroanal.Chem. 1993, 349, 15–30.

74. T. Wandlowski, G. B. Jameson, R. de Levie,J. Phys. Chem. 1993, 97, 10 119–10 126.

75. T. Wandlowski, G. B. Jameson, R. de Levie,J. Electroanal. Chem. 1994, 379, 215–222.

76. C. Donner, H. Baumgartel, L. Pohlmannet al., Ber. Bunsen-Ges. Phys. Chem. 1996,100, 403–412.

77. B. Kurtyka, R. de Levie, J. Electroanal. Chem.1995, 397, 311–314.

78. T. Wandlowski, M. Hromadova, R. de Levie,Langmuir 1997, 13, 2766–2772.

79. F. David, H. Ouguenoune, A. Bolyos et al.,Anal. Chim. Acta 1994, 292, 297–304.

80. L. Pospisil, M. Svestka, J. Electroanal. Chem.1994, 366, 295–302.

81. L. Pospisil, S. Zalis, N. Fanelli, J. Chem.Educ. 1995, 72, 997–1002.

82. A. Avranas, N. Papadopoulos, Langmuir1992, 8, 2804–2809.

83. R. Srinivasan, R. de Levie, J. Electroanal.Chem. 1986, 206, 307–312.

84. V. Brabec, V. Kleinwachter, V. Vetterl,in Bioelectrochemistry of Biomacromolecules(Eds.: G. Lenaz, G. Milazzo), BirkhauserVerlag, Basel, 1997, pp. 1–104, Vol. 5.

85. V. Vetterl, Bioelectrochem. Bioenerg. 1976, 3,338–345.

86. J. Jursa, V. Vetterl, Bioelectrochem. Bioenerg.1984, 12, 137–146.

87. J. Jursa, V. Vetterl, Studia Biophys. 1986,114, 75–82.

88. V. Vetterl, N. Papadopoulos, V. Drazanet al., Electrochim. Acta 2000, 45, 2961–2971.

89. J. Sponer, J. Leszczynski, V. Vetterl et al.,J. Biomol. Struct. Dyn. 1996, 13, 695–766.

90. R. de Levie, T. Wandlowski, J. Electroanal.Chem. 1994, 366, 265–270.

91. V. I. Melik-Gaikazyan, J. Phys. Chem.(USSR) 1952, 26, 560–580.


92. W. M. Heckl, D. P. Smith, G. Binnig et al.,J. Proc. Natl. Acad. Sci. 1991, 88, 8003–8005.

93. R. Srinivasan, J. C. Murphy, N. Pat-tabiraman, Ultramicroscopy 1991, 42–44,453–459.

94. R. Srinivasan, P. Gopalan, J. Phys. Chem.1993, 97, 8770–8775.

95. N. J. Tao, J. A. de Rose, S. M. Lindsay,J. Phys. Chem. 1993, 97, 910–919.

96. N. J. Tao, Z. Shi, J. Phys. Chem. 1994, 98,1464–1471.

97. T. Wandlowski, D. Lampner, S. Lindsay,J. Electroanal. Chem. 1996, 404, 215–226.

98. M. Holzle, T. Wandlowski, D. M. Kolb, Surf.Sci. 1995, 335, 281–290.

99. C. Bruckner-Lea, J. Janata, J. Conroy et al.,Langmuir 1993, 9, 3612–3617.

100. J. Janata, C. Bruckner-Lea, J. F. T. Conroyet al., in Interfacial Design and ChemicalSensing (Eds.: T. E. Mallovk, D. J. Harrison),ACS, Washington DC, 1994, pp. 175–184.

101. J. F. T. Conroy, K. Caldwell, C. Bruckner-Lea et al., Electrochim. Acta 1995, 40,2927–2934.

102. F. Jelen, V. Vetterl, A. Schaper et al.,J. Electroanal. Chem. 1994, 377, 197–203.

103. S. Sotiropoulos, P. Nikitas, N. Papadopou-los, J. Electroanal. Chem. 1993, 356,225–243.

104. B. Janik, P. J. Elving, J. Am. Chem. Soc.1973, 95, 8495–8502.

105. D. Krznaric, P. Valenta, H. W. Nurnberg,J. Electroanal. Chem. 1975, 65, 863–881.

106. V. Brabec, A. P. Kavunenko, J. Electroanal.Chem. 1987, 237, 261–267.

107. E. Palecek, V. Kolar, F. Jelen et al., Bioelec-trochem. Bioenerg. 1990, 23, 285–299.

108. M. Fojta, V. Vetterl, M. Tomschik et al.,Biophys. J. 1997, 72, 2285–2293.

109. A. D. Mesmaeker, K.-H. Altmann, A. Wald-ner et al., Curr. Opin. Struct. Biol. 1995, 5,343–355.

110. P. E. Nielsen, M. Egholm, O. Buchard, Bio-conjugate Chem. 1994, 5, 3–7.

111. A. Popov, R. Naneva, N. Dimitrov et al.,Electrochim. Acta 1992, 37, 2369–2371.

112. T. Dretschkow, A. S. Dakkouri, T. Wand-lowski, Langmuir 1997, 13, 2843–2856.

113. T. Dretschkow, T. Wandlowski, Electrochim.Acta 1998, 43, 2991–3006.

114. K. Takehara, S. Yamada, Y. Ide, J. Elec-troanal. Chem. 1992, 333, 339–344.

115. H. Bare, C. Buess-Herman, Physicochem.Eng. Aspects 1998, 134, 181–191.

116. T. Wandlowski, D. Lampner, S. M. Lindsay,J. Electroanal. Chem. 1996, 404, 215–226.

117. M. Kasaya, H. Tabata, T. Kawai, Surf. Sci.1995, 342, 215–223.

118. M. Kasaya, H. Tabata, T. Kawai, Surf. Sci.1998, 406, 302–311.

119. K. Takehara, Y. Ide, Bioelectrochem. Bioen-erg. 1992, 27, 501–507.



122. B. Roelfs, E. Bunge, C. Schroter et al.,J. Phys. Chem. B 1997, 101, 754–765.

123. Z. Yang, I. Engquist, B. Liedberg et al.,J. Electroanal. Chem. 1997, 430, 189–195.

124. E. Lust, A. Janes, P. Miidla et al., J. Elec-troanal. Chem. 1997, 425, 25–37.

125. Z. Yang, I. Enqquist, M. Wirde et al., Lang-muir 1997, 13, 3210–3218.

126. G. M. Brown, D. P. Allison, R. J. Warmacket al., Ultramicroscopy 1991, 38, 253–264.

127. R. Srinivasan, J. C. Murphy, N. Pattabira-man, Ultramicroscopy 1992, 42, 453–459.

128. S. Hason, V. Vetterl, Bioelectrochem. 2001,in press.

129. J. Wang, G. Rivas, X. H. Cai et al., Electro-analysis 1997, 9, 120–124.

130. J. Wang, G. Rivas, X. Zhang et al., Langmuir1999, 15, 6541–6545.

131. J. Wang, X. Zhang, C. Parrado et al., Elec-trochem. Commun. 1999, 1, 197–202.

132. B. Janik, P. J. Elving, Chem. Rev. 1968, 68,295–319.

133. J. W. Webb, B. Janik, P. J. Elving, J. Am.Chem. Soc. 1973, 95, 991–1003.

134. L. Trnkova, M. Studnickova, E. Palecek,Bioelectrochem. Bioenerg. 1980, 7, 644–658.

135. M. Studnickova, L. Trnkova, J. Zetek et al.,Bioelectrochem. Bioenerg. 1989, 21, 83–86.

136. E. Palecek, F. Jelen, L. Trnkova, Gen. Phys-iol. Biophys. 1986, 5, 315–329.

137. E. Palecek, M. Tomschik, V. Stankova et al.,Electroanalysis 1997, 9, 990–997.

138. F. Jelen, M. Tomschik, E. Palecek, J. Elec-troanal. Chem. 1997, 423, 141–148.

139. L. Havran, L. Trnkova, O. Dracka, J. Elec-troanal. Chem. 1998, 454, 65–73.

140. T. E. Cummings, P. J. Elving, J. Electroanal.Chem. 1978, 94, 123.

141. T. E. Cummings, P. J. Elving, J. Electroanal.Chem. 1979, 102, 237–248.

142. G. Dryhurst, P. J. Elving, J. Electrochem. Soc.1968, 115, 1014–1022.


143. G. Dryhurst, G. F. Pace, J. Electrochem. Soc.1970, 117, 1259–1265.

144. G. Dryhurst, Anal. Chim. Acta 1971, 57,137–149.

145. J. M. Zen, M. R. Chang, G. Ilangovan, Ana-lyst 1999, 124, 679–684.

146. A. C. Ontko, P. M. Armistead, S. R. Kircuset al., Inorg. Chem. 1999, 38, 1842–1846.

147. X. Cai, B. Ogorevc, K. Kalcher, Electroanaly-sis 1995, 7, 1126–1131.

148. A. M. O. Brett, F. M. Matysik, J. Electroanal.Chem. 1997, 429, 95–99.

149. P. Singhal, W. G. Kuhr, Anal. Chem. 1997,69, 3552–3557.

150. K. Brainina, Talanta 1971, 18, 513–520.151. K. Brainina, E. Neyman, Electroanalytical

Stripping Methods, Wiley & Sons, New York,1993.

152. T. M. Florence, J. Electroanal. Chem. 1979,97, 219–236.

153. J. Heyrovsky, J. Kuta, Principles of Polarog-raphy, Czechoslovak Academy of Sciences,Prague, 1965.

154. F. Vydra, K. Stulik, E. Julakova, Electrochem-ical Stripping Analysis, Wiley & Sons, NewYork, 1973.

155. J. Revenda, Collect. Czech. Chem. Commun.1934, 6, 453–467.

156. I. M. Kolthoff, C. Barnum, J. Am. Chem.Soc. 1940, 62, 3061–3066.

157. P. Zuman, J. Koryta, R. Kalvoda, Collect.Czech. Chem. Commun. 1953, 18, 350–365.

158. O. Manousek, P. Zuman, Chem. Listy 1955,49, 668–678.

159. D. Kalab, Chem. Zvesti 1964, 18, 435–439.160. E. Palecek, F. Jelen, Collect. Czech. Chem.

Commun. 1980, 45, 3472–3481.161. E. Palecek, Anal. Biochem. 1980, 108,

129–138.162. M. R. Smyth, J. G. Osteryoung, Anal. Chem.

1977, 49, 2310–2314.163. T. M. Florence, J. Electroanal. Chem. 1978,

97, 237–255.164. Y. Vaneesorn, W. F. Smyth, Anal. Chim.

Acta 1980, 117, 183–191.165. E. Palecek, Anal. Lett. 1980, 13, 331–345.166. E. Palecek, F. Jelen, M. A. Hung et al., Bio-

electrochem. Bioenerg. 1981, 8, 621–631.167. E. Palecek, F. Jelen, I. Postbieglova, Studia

Biophys. 1989, 130, 51–54.168. E. Borek, O. K. Scharma, T. P. Waalkes,

in Modified Nucleosides and Cancer(Ed.: G. Nass), Springer-Verlag, Heildeberg1983, pp. 301–315.

169. E. Palecek, Anal. Chim. Acta 1985, 174,103–113.

170. E. Palecek, J. Osteryoung, R. A. Osteryoung,Anal. Chem. 1982, 54, 1389–1394.

171. A. J. M. Ordieres, M. J. G. Gutierrez, A. C.Garcia et al., Analyst 1987, 112, 243–251.

172. J. Wang, S. M. Lin, V. Villa, Analyst 1987,112, 247–251.

173. M. Khodari, M. Ghandour, A. M. Taha, Ta-lanta 1997, 44, 305–310.

174. J. Yan, C. Zhu, G. Pu, Bioelectrochem. Bioen-erg. 1993, 29, 347–355.

175. Y. Jiangli, Z. Chongjie, P. Guogong et al.,Bioelectrochem. Bioenerg. 1993, 29, 347–355.

176. B. Bouzid, A. M. G. Mac Donald, Anal.Chim. Acta 1988, 211, 155–173.

177. B. Bouzid, A. M. G. Mac Donald, Anal.Proc. 1986, 23, 295–297.

178. B. Bouzid, A. M. G. Mac Donald, Anal.Chim. Acta 1988, 211, 175–193.

179. S. Glodowski, R. Bilewicz, Z. Kublik, Anal.Chim. Acta 1987, 201, 11–22.

180. S. Glodowski, R. Bilewicz, Z. Kublik, Anal.Chim. Acta 1986, 186, 39–47.

181. X. Zhao, W. R. Jin, Y. Wang, Electrochim.Acta 1996, 41, 887–893.

182. X. Zhao, W. R. Jin, J. Tang, Electroanalysis1996, 8, 370–374.

183. M. M. Correia dos Santos, C. M. L. F.Lopes, M. L. Simoes-Goncalves Bioelec-trochem. Bioenerg. 1996, 39, 55–60.

184. P. Valenta, H. W. Nurnberg, P. Klahre, Bio-electrochem. Bioenerg. 1974, 1, 487–505.

185. E. Palecek, V. Vetterl, Biopolymers 1968, 6,917–928.

186. V. Brabec, E. Palecek, Biopolymers 1972, 11,2577–2589.

187. V. Brabec, E. Palecek, Z. Naturforsch. 1973,28c, 685–692.

188. V. Brabec, G. Dryhurst, J. Electroanal. Chem.1978, 91, 219–229.

189. V. Brabec, K. Niki, Biophys. Chem. 1985, 23,63–70.

190. V. Brabec, V. Glezers, V. Kadysh, Col-lect. Czech. Chem. Commun. 1983, 48,1257–1271.

191. B. Janik, R. G. Sommer, Biopolymers 1973,12, 2803–2822.

192. H. Berg, G. Horn, J. Flemming, in DynamicAspects of Biopolyelectrolytes and Biomem-branes, Elsevier Press, New York, 1982,pp. 181–184.

193. J. Flemming, Biopolymers 1973, 9,1975–1988.


194. B. Janik, R. G. Sommer, Bioelectrochem.Bioenerg. 1976, 3, 622–633.

195. B. Malfoy, J. M. Sequaris, P. Valenta et al.,J. Electroanal. Chem. 1977, 75, 455–469.

196. P. Valenta, P. Grahmann, J. Electroanal.Chem. 1974, 49, 41–53.

197. P. Valenta, H. W. Nurnberg, Biophys. Struct.Mech. 1974, 1, 17–26.

198. V. Brabec, Biophys. Chem. 1980, 11, 1–7.199. J. Koryta, Collect. Czech. Chem. Commun.

1953, 18, 206.200. J. Koryta, Collect. Czech. Chem. Commun.

1953, 18, 206–213.201. P. Delahay, I. Trachtenberg, J. Am. Chem.

Soc. 1957, 79, 2355–2362.202. M. K. Kaisheva, M. Matsumoto, Y. Kita

et al., Langmuir 1988, 4, 762–765.203. E. Palecek, P. Boublikova, F. Jelen, Anal.

Chim. Acta 1986, 187, 99–107.204. W. Guschlbauer, V. Vetterl, FEBS Lett.

1969, 4, 57–60.205. V. Vetterl, W. Guschlbauer, Arch. Biochem.

Biophys. 1972, 148, 130–140.206. W. Guschlbauer, Nucleic Acid Structure,

Springer Verlag, New York, 1976.207. C. M. A. Brett, A. M. O. Brett, Electrochem-

istry. Principles, Methods, and Applications,Oxford University Press, Oxford, 1993.

208. I. R. Miller, D. C. Grahame, J. Am. Chem.Soc. 1957, 79, 3006–3012.

209. M. Sluyters-Rehbach, J. H. Sluyters, J. Elec-troanal. Chem. 1982, 136, 39–58.

210. L. Hanak, V. Vetterl, Bioelectrochem. Bioen-erg. 1998, 46, 9–13.

211. A. N. Frumkin, B. B. Damaskin, in ModernAspects of Electrochemistry (Eds.: J. O. M.Bockris, B. E. Conway), Butterworth, Lon-don, 1964, pp. 149–223.

212. I. R. Miller, D. C. Grahame, J. Am. Chem.Soc. 1956, 78, 3577–3585.

213. E. Sabatani, J. Cohen-Boulakia, M. Bruen-ing et al., Langmuir 1993, 9, 2974–2981.

214. E. Sabatani, Y. Gafni, I. Rubinstein, J. Phys.Chem. 1995, 99, 12 305–12 311.

215. R. P. Janek, W. R. Fawcet, A. Ulman,J. Phys. Chem. 1997, 101, 8550–8558.

216. R. P. Janek, W. R. Fawcett, A. Ulman, Lang-muir 1998, 14, 3011–3018.

217. P. Diao, D. L. Jiang, X. L. Cui et al., Bioelec-trochem. Bioenerg. 1998, 45, 173–179.

218. P. Diao, D. L. Jiang, X. L. Cui et al., Bioelec-trochem. Bioenerg. 1999, 48, 469–475.

219. A. E. Vallejo, C. A. Gervasi, L. M. Gassa,Bioelectrochem. Bioenerg. 1998, 47, 343–348.

220. H. O. Finklea, D. A. Snider, J. Fedyk et al.,Langmuir 1993, 9, 3660–3667.

221. X. Cui, D. Jiang, P. Diao et al., J. Electroanal.Chem. 1999, 470, 9–13.

222. M. W. Humphreys, R. Parsons, J. Elec-troanal. Chem. 1977, 75, 427–436.

223. A. Roeseler, Infrared Spectroscopic Ellipsom-etry, Akademie-Verlag, Berlin, 1990.

224. O. M. Magnussen, B. M. Ocko, M. Deutschet al., Nature 1996, 384, 250–252.

225. C. Hinnen, A. Rousseau, R. Parsons et al.,J. Electroanal. Chem. 1981, 125, 193–203.

226. T. Boland, B. D. Ratner, Biophysics 1995, 92,5297–5301.

227. R. M. Zimmerman, E. C. Cox, Nucleic AcidsRes. 1994, 22, 492–497.

228. E. Koglin, J.-M. Sequaris, in Topics in Cur-rent Chemistry, Springer-Verlag, New York,1986, pp.1–, Vol. 4.

229. C. Fan, H. Song, X. Hu et al., Anal. Biochem.1999, 271, 1–7.

230. S. O. Kelley, J. K. Barton, N. M. Jacksonet al., Langmuir 1998, 14, 6781–6784.

231. J. Humlicek, A. Roeseler, Thin Solid Film1993, 234, 332–336.

232. J. Humlicek, C. Thomsen, M. Cardonaet al., Physica C 1994, 222, 166–172.

233. J. Humlicek, Philos. Mag. B 1994, 70,699–710.

234. A. Rothen, C. Mathot, Immunochemistry1969, 6, 241–251.

235. A. Rothen, in Surface and Membrane Science,(Ed.: D. A. Cadonhead), Academic Press,New York, 1974, Vol. 8.

236. M. Cardona, Modulation Spectroscopy, Aca-demic Press, New York, 1969.

237. R. M. A. Azzam, N. M. Bashara, Ellipsome-try and Polarized Light, North-Holland, NewYork, 1977.

238. R. Artzi, Feinberg Graduate School, Weiz-mann Institute of Science, Rehovot, 1999.

239. A. Vilan, R. Ussyshkin, K. Gartsman et al.,J. Phys. Chem. B 1998, 102, 3307–3312.

240. H. Clausen-Schaumann, H. E. Gaub, Lang-muir 1999, 15, 8246–8251.

241. M. S. Spector, J. M. Schnur, Science 1997,275, 791–792.

242. E. Palecek, M. A. Hung, Anal. Biochem.1983, 132, 236–242.

243. V. Brabec, Gen. Physiol. Biophys. 1983, 2,193–199.

244. F. Jelen, E. Palecek, Biophys. Chem. 1986,24, 285–290.


245. B. Janik, E. Palecek, Z. Naturforsch. 1966,21b, 1117, 1118.

246. B. Janik, E. Palecek, Elektrochemische meth-oden und prinzipien in der molekular-biologie, III. Jenaer symp., Academic-Verlag,Berlin, 1966, Abhandlungen der DAW,pp. 513–518.

247. M. Tomschik, F. Jelen, L. Havran et al.,J. Electroanal. Chem. 1999, 476, 71–80.

248. F. Jelen, M. Fojta, E. Palecek, J. Electroanal.Chem. 1997, 427, 49–56.

249. M. Tomschik, L. Havran, M. Fojta et al.,Electroanalysis 1998, 476, 71–80.

250. J. Wu, Y. Huang, J. Zhou et al., Bioelec-trochem. Bioenerg. 1997, 44, 151–154.

251. T. Kubicarova, M. Fojta, J. Vidic et al., Elec-troanalysis 2000, 12, 1390–1391.

252. T. Kubicarova, M. Fojta, J. Vidic et al., Elec-troanalysis 2000, 12, 122–142.

253. V. Brabec, Bioelectrochem. Bioenerg. 1981, 8,437–449.

254. C. M. A. Brett, A. M. O. Brett, S. H. P.Serrano, J. Electroanal. Chem. 1994, 366,225–231.

255. J. Wang, X. Cai, J. Wang et al., Anal. Chem.1995, 67, 4065–4070.

256. J. Wang, X. Cai, C. Jonsson et al., Electro-analysis 1996, 8, 20–24.

257. X. Cai, G. Rivas, P. A. M. Farias et al., Bio-electrochem. Bioenerg. 1996, 401, 41–47.

258. C. Cai, G. Rivas, P. A. M. Farias et al., Elec-troanalysis 1996, 8, 753–758.

259. J. Wang, X. Cai, J. Fernandez et al., Anal.Chem. 1997, 69, 4056–4059.

260. J. Wang, Anal. Chem. 1995, 67, R487–R492.261. J. Wang, S. Bollo, J. L. L. Paz et al., Anal.

Chem. 1999, 71, 1910–1913.262. J. Wang, A. N. Kawde, E. Sahlin, Analyst

2000, 125, 5–7.263. J. Wang, X. H. Cai, J. R. Fernandes et al.,

Anal. Chem. 1997, 69, 4056–4059.264. C. G. Siontorou, A. M. O. Brett, D. P.

Nikolecis, Talanta 1996, 43, 1137–1144.265. M. E. Napier, H. H. Thorp, Langmuir 1997,

13, 6342–6344.266. M. E. Napier, H. H. Thorp, J. Fluorescence

1999, 9, 181–186.267. D.-W. Pang, Y.-P. Qi, Z.-L. Wang et al., Elec-

troanalysis 1995, 7, 774–776.268. P. Singhal, W. G. Kuhr, Anal. Chem. 1997,

69, 4828–4832.269. M. Fojta, R.,Doffkova, E. Palecek, Electro-

analysis 1996, 8, 420–426.

270. E. Palecek, M. Fojta, Anal. Chem. 1994, 66,1566–1571.

271. M. Fojta, C. Teijeiro, E. Palecek, Bioelec-trochem. Bioenerg. 1994, 34, 69–76.

272. M. Fojta, E. Palecek, Anal. Chim. Acta 1997,342, 1–12.

273. M. Fojta, V. Stankova, E. Palecek et al., Ta-lanta 1998, 46, 155–161.

274. M. Fojta, R. P. Bowater, V. Stankova et al.,Biochemistry 1998, 37, 4853–4862.

275. J. Wang, E. Palecek, P. E. Nielsen et al.,J. Am. Chem. Soc. 1996, 118, 7667–7670.

276. M. Fojta, L. Havran, E. Palecek, Electroanal-ysis 1997, 9, 1033, 1034.

277. O. Dracka, J. Electroanal. Chem. 1996, 402,19–28.

278. L. Trnkova, O. Dracka, J. Electroanal. Chem.1993, 348, 265–271.

279. L. Trnkova, O. Dracka, J. Electroanal. Chem.1996, 413, 123–129.

280. E. Palecek, B. D. Frary, Arch. Biochem. Bio-phys. 1966, 115, 431–436.

281. E. Palecek, I. Postbieglova, J. Electroanal.Chem. 1986, 214, 359–371.

282. E. Palecek, Anal. Biochem. 1988, 170,421–431.

283. E. Palecek, F. Jelen, C. Teijeiro et al., Anal.Chim. Acta 1993, 273, 175–186.

284. C. Teijeiro, K. Nejedly, E. Palecek, J. Biomol.Struct. Dyn. 1993, 11, 313–331.

285. C. Teijeiro, P. Perez, D. Marin et al., Bioelec-trochem. Bioenerg. 1995, 38, 77–83.

286. M. Vojtiskova, E. Lukasova, F. Jelen et al.,Bioelectrochem. Bioenerg. 1981, 8, 487–496.

287. J. M. Sequaris, M. L. Kaba, P. Valenta, Bio-electrochem. Bioenerg. 1984, 13, 225–227.


289. V. Brabec, E. Palecek, J. Electroanal. Chem.1978, 88, 373–385.

290. J. Flemming, H. Berg, Bioelectrochem. Bioen-erg. 1974, 1, 459–465.

291. E. Palecek, Collect. Czech. Chem. Commun.1974, 39, 3449–3455.

292. H. W. Nurnberg, P. Valenta, Croat. Chem.Acta 1976, 48, 623–641.


294. V. Brabec, E. Palecek, Studia Biophys. 1976,60, 105–110.

295. F. Jelen, E. Palecek, Gen. Physiol. Biophys.1985, 4, 219–237.


296. E. Palecek, in Proc. Electroanalysis in Hy-giene, Environmental, Clinical and Phar-maceutical Chemistry (Ed.: W. F. Smyth),Elsevier, Amsterdam 1980, pp. 79–99.

297. M. Fojta, L. Havran, J. Fulneckova et al.,Electroanalysis 2000, 12, 926–934.

298. A. D. Bates, A. Maxwell, DNA Topology,Oxford University Press, Oxford, 1993.

299. E. Palecek, Crit. Rev. Biochem. Mol. Biol.1991, 26, 151–226.

300. D. Kasparova, O. Vrana, V. Kleinwachteret al., Biophys. Chem. 1987, 28, 191–197.

301. H. Clausen-Schaumann, M. Rief, C. Tolks-dorf et al., Biophys. J. 2000, 78, 1997–2007.

302. D. Bensimon, V. Simon, V. Croquette et al.,Phys. Rev. Lett. 1995, 74, 4754–4757.

303. M. Rief, H. Clausen-Schaumann, H. E.Gaub, Nat. Struct. Biol. 1999, 6, 346–349.

304. M. G. Blackburn, M. J. Gait, Nucleic Acids inChemistry and Biology, IRL Press, New York,1990.

305. G. S. Manning, Q. Rev. Biophys. 1978, 11,179–246.

306. J. Labuda, M. Buckova, M. Vanickova et al.,Electroanalysis 1999, 11, 101–107.

307. M. Maeda, K. Nakano, S. Uchida et al.,Chem. Lett. 1994, 1805–1808.

308. J.-M. Sequaris, M. Esteban, Electroanalysis1990, 2, 35–41.

309. J.-M. Sequaris, J. Swiatek, Bioelectrochem.Bioenerg. 1991, 26, 15–28.

310. E. Braun, Y. Eichen, U. Sivan et al., Nature1998, 391, 775–778.

311. A. B. Steel, T. M. Herne, M. J. Tarlov, Anal.Chem. 1998, 70, 4670–4677.

312. A. B. Steel, T. M. Herne, M. J. Tarlov, Bio-conjugate Chem. 1999, 10, 419–423.

313. M. Aslanoglu, A. Houlton, B. R. Horrocks,Analyst 1998, 123, 753–757.

314. R. F. Johnston, D. M. Lewis, J. Q. Cham-bers, J. Electroanal. Chem. 1999, 466, 2–7.

315. J. L. M. Alvarez, J. A. G. Calzon, J. M. L.Fonseca, J. Electroanal. Chem. 1998, 457,53–59.

316. M. T. Carter, A. J. Bard, J. Am. Chem. Soc.1987, 109, 7528–7530.

317. M. T. Carter, M. Rodriguez, A. J. Bard,J. Am. Chem. Soc. 1989, 111, 8901–8911.

318. X. H. Xu, A. J. Bard, J. Am. Chem. Soc. 1995,117, 2627–2631.

319. T. W. Welch, H. H. Thorp, J. Phys. Chem.1996, 100, 13 829–13 836.

320. M. Rodriguez, A. J. Bard, Anal. Chem. 1990,62, 2658–2662.

321. M. T. Carter, A. J. Bard, Bioconjugate Chem.1990, 1, 257–263.

322. D. W. Pang, H. D. Abruna, Anal. Chem.1998, 70, 3162–3169.

323. A. Erdem, B. Meric, K. Kerman et al., Elec-troanalysis 1999, 11, 1372–1376.

324. K. M. Millan, A. Saraullo, R. Mikkelsen,Anal. Chem. 1994, 66, 2943–2948.

325. Y. Mishima, J. Motonaka, S. Ikeda, Anal.Chim. Acta 1997, 345, 45–50.

326. X. Cai, G. Rivas, H. Shiraishi et al., Anal.Chim. Acta 1997, 344, 64–76.

327. H. H. Thorp, Tibtech 1998, 16, 117–121.328. D. H. Johnston, C.-C. Cheng, K. J. Campbell

et al., Inorg. Chem. 1994, 33, 6388–6390.329. D. H. Johnston, K. C. Glasgow, H. H.

Thorp, J. Am. Chem. Soc. 1995, 117,8933–8938.

330. P. A. Ropp, H. H. Thorp, Chem. Biol. 1999,6, 599–605.

331. F. Qu, N.-Q. Li, Electroanalysis 1997, 9,1348–1352.

332. K. Hashimoto, K. Ito, Y. Ishimori, Anal.Chem. 1994, 66, 3830–3833.

333. J. Wang, M. Ozsoz, X. H. Cai et al., Bioelec-trochem. Bioenerg. 1998, 45, 33–40.

334. S. Palanti, G. Marrazza, M. Mascini, Anal.Lett. 1996, 29, 2309–2331.

335. S. O. Kelley, N. M. Jackson, M. G. Hillet al., Angew. Chem., Int. Ed. Engl. 1999,38, 941–945.

336. S. O. Kelley, E. M. Boon, J. K. Barton et al.,Nucleic Acids Res. 1999, 27, 4830–4837.

337. M. Rodriguez, A. J. Bard, Inorg. Chem.1992, 31, 1129–1135.

338. P. C. Pandey, H. H. Weetall, Anal. Chem.1994, 66, 1236–1241.

339. J. Wang, G. Rivas, D. B. Luo et al., Anal.Chem. 1996, 68, 4365–4369.

340. M. Maeda, Y. Mitsuhashi, K. Nakano et al.,Anal. Sci. 1992, 8, 83, 84.

341. K. Nakano, M. Maeda, S. Uchida et al.,Anal. Sci. 1997, 13(Suppl S), 455–456.

342. S. Takenaka, T. Ihara, M. Takagi, J. Chem.Soc., Chem. Commun. 1990, 1485–1487.

343. S. Takenaka, Y. Uto, H. Saita et al., Chem.Commun. 1998, 1111, 1112.

344. S. Takenaka, K. Yamashita, Y. Uto et al.,Denki Kagaku 1998, 12, 1329–1334.

345. S. Takenaka, K. Yamashita, M. Takagi et al.,Anal. Chem. 2000, 72, 1334–1341.

346. S. Takenaka, Y. Uto, M. Takagi et al., Chem.Lett. 1998, 989–990.


347. S. Takenaka, K. Yamashita, M. Takagi et al.,Nucleic Acids Symp. Ser. 1999, 42, 149, 150.

348. A. M. O. Brett, S. H. P. Serrano, I. Gutzet al., Bioelectrochem. Bioenerg. 1997, 42,175–178.

349. A. M. O. Brett, S. H. P. Serrang, S. I. G. R.Gutz, Electroanalysis 1997, 9, 110–114.

350. A. M. O. Brett, S. H. P. Serrano, I. Gutzet al., Bioelectrochem. Bioenerg. 1997, 42,1132–1137.

351. A. M. O. Brett, S. H. P. Serrano, I. Gutzet al., Electroanalysis 1997, 9, 1132–1137.

352. A. M. Brett, T. R. Macedo, D. Raimundoet al., Biosens. Bioelectron. 1998, 13, 861–867.

353. J. Kuta, E. Palecek, in Topics Bioelectrochem.Bioenerg. (Ed.: G. Milazzo), John Wiley &Sons, Chichester, London, 1983, pp. 1–63,Vol. 5.

354. E. Palecek, E. Lukasova, F. Jelen et al., Bio-electrochem. Bioenerg. 1981, 8, 497–506.

355. E. Lukasova, F. Jelen, E. Palecek, Gen. Phys-iol. Biophys. 1982, 1, 53–70.

356. E. Lukasova, M. Vojtiskova, F. Jelen et al.,Gen. Physiol. Biophys. 1984, 3, 175–191.

357. E. Palecek, M. Vojtiskova, F. Jelen et al., inCharge and Field Effects in Biosystems (Eds.:M. J. Allen, P. N. R. Usherwood), Abacuspress, Tonbridge, 1984, pp. 397–404.

358. E. Palecek, M. Vojtiskova, F. Jelen et al.,Bioelectrochem. Bioenerg. 1984, 12, 135, 136.

359. F. Jelen, P. Karlovsky, P. Pecinka et al., Gen.Physiol. Biophys. 1991, 10, 461–473.

360. E. Palecek, F. Jelen, E. Minarova et al.,Structural Tools for the Analysis of Protein-Nucleic Acid Complexes, Birghauser Verlag,Basel, 1992, pp. 1–22.

361. J. M. Hall, J. Moore-Smith, V. Bannisteret al., Biochem. Mol. Biol. Int. 1994, 32,21–28.

362. E. Palecek, Methods Enzymol. 1992, 212,139–155.

363. E. Palecek, in Nucleic Acids and MolecularBiology (Eds.: F. Eckstein, D. M. J. Lilley),Springer Verlag, Berlin, 1994, pp. 1–13,Vol. 8.

364. T. Ihara, Y. Maruo, S. Takanaka et al., Nu-cleic Acids Res. 1996, 24, 4273–4280.

365. H. Korri-Youssoufi, F. Garnier, P. Srivas-tava et al., J. Am. Chem. Soc. 1997, 119, 7388,7389.

366. P. Bauerle, A. Emge, Adv. Mater. 1998, 3,324–330.

367. M. J. Tarlov, E. F. Bowden, J. Am. Chem.Soc. 1991, 113, 1847–1849.

368. T. M. Herne, M. J. Tarlov, J. Am. Chem. Soc.1997, 119, 8916–8920.

369. S. O. Kelley, J. K. Barton, N. M. Jacksonet al., Bioconjugate Chem. 1997, 8, 31–37.

370. R. Levicky, T. M. Herne, M. J. Tarlov et al.,J. Am. Chem. Soc. 1998, 120, 9787–9792.

371. D. Marina, R. Valera, E. de la Red et al.,Bioelectrochem. Bioenerg. 1997, 44, 51–56.

372. D. Marin, P. Perez, C. Teijeiro et al., Bio-phys. Chem. 1998, 75, 87–95.

373. D. Marin, C. Teijero, P. Perez et al., RecentRes. Dev. Electrochem. 1998, 1, 31–43.

374. P. Perez, C. Teijeiro, D. Marin, Chem. Biol.Interact. 1999, 117, 65–81.

375. J. Wang, Biosens. Bioelectron. 1998, 13,757–762.

376. E. Uhlmann, A. Peyman, G. Breipohl et al.,Angew. Chem. Int. Ed. Engl. 1998, 37,2797–2823.

377. P. E. Nielsen, Acc. Chem. Res. 1999, 32,624–630.

378. D. D. Eley, D. I. Spivey, Trans. Faraday Soc.1962, 58, 411.

379. D. Dee, M. E. Baur, J. Chem. Phys. 1974, 60,541–560.

380. T. A. Hofmann, J. Ladik, Adv. Chem. Phys.1964, 7, 84.

381. S. Suhai, J. Chem. Phys. 1972, 57,5599–5603.

382. V. Mikac-Dadic, V. Pravdic, A. Rupprecht,Bioelectrochem. Bioenerg. 1974, 1, 364–369.

383. T. J. Meade, in Metal Ions in BiologicalSystems (Eds.: A. Siegel, H. Siegel), MarcelDekker, New York, 1996, pp. 453–478.

384. B. Norden, P. Lincoln, B. Akerman et al.,in Metal Ions in Biological Systems (Eds.:A. Sigel, H. Sigel), Marcel Dekker, NewYork, 1996, pp. 177–252, Vol. 33.

385. E. Tuite, in Organic and Inorganic Pho-tochemistry (Eds.: V. Ramamurthy, K. S.Schanze), Marcel Dekker, New York, 1998,pp. 55–74.

386. R. E. Holmlin, P. J. Dandliker, J. K. Barton,Angew. Chem. Int. Ed. Engl. 1997, 36,2714–2730.

387. S. O. Kelley, J. K. Barton, in Metal Ions inBiological Systems. Interactions Between FreeRadicals and Metal Ions in Life Processes(Eds.: A. Sigel, H. Sigel), 1999, pp. 211–249,Vol. 36.

388. G. Hartwich, D. J. Caruana, T. deLumley-Woodyear et al., J. Am. Chem. Soc. 1999,121, 10 803–10 812.


389. G. C. Barker, J. Electroanal. Chem. 1986,214, 373–390.

390. G. C. Barker, J. Electroanal. Chem. 1987,226, 171–192.

391. G. C. Barker, A. W. Gardner, Analyst 1992,117, 1811–1828.

392. S. O. Kelley, J. K. Barton, Science 1999, 283,375–381.

393. D. Porath, A. Bezryadin, S. de Vries et al.,Nature 2000, 403, 635–638.

394. J. Jortner, M. Bixon, T. Langenbacher et al.,Proc. Natl. Acad. Sci. 1998, 95,12 759–12 765.

395. A. Okada, V. Chernyak, S. Mukamel, J. Phys.Chem. A 1998, 102, 1241–1251.

396. B. Giese, S. Wessely, M. Spormann et al.,Angew. Chem. Int. Ed. Engl. 1999, 38,996–998.

397. E. Meggers, M. E. Michel-Beyerle, B. Giese,J. Am. Chem. Soc. 1998, 120, 12 950–12 955.

398. B. Armitage, D. Ly, T. Koch et al., Proc. Natl.Acad. Sci. 1997, 94, 12 320–12 325.

399. S. M. Gasper, G. B. Schuster, J. Am. Chem.Soc. 1997, 119, 12 762–12 771.

400. Y. Kan, G. B. Schuster, J. Am. Chem. Soc.1999, 121, 11 607–11 614.

401. D. Ly, L. Sanii, G. B. Schuster, J. Am. Chem.Soc. 1999, 121, 9400–9410.

402. Y. Okahata, T. Kobayashi, K. Tanaka et al.,J. Am. Chem. Soc. 1998, 120, 6165, 6166.

403. H.-W. Fink, C. Schonenberg, Nature 1999,398, 407–410.

404. T. de Lumley-Woodyear, D. J. Caruana,C. N. Campbell et al., Anal. Chem. 1999,71, 394–398.

405. J. Wang, P. E. Nielsen, M. Jiang et al., Anal.Chem. 1997, 69, 5200–5212.

406. J. Wang, G. Rivas, X. H. Cai et al., Anal.Chim. Acta 1997, 344, 111–118.

407. J. Marmur, R. Rownd, C. L. Schildkraut, inProgress in Nucleic Acid Research (Eds.:J. N. Davidson, W. E. Cohn), AcademicPress, London, 1963, pp. 231–300,Vol. 1.

408. E. Palecek, Die Polarographie in Chemother-apie, Biochemie und Biologie. I. Jenaer symp.,Academic Verlag, Berlin, 1964, Abhandlun-gen der DAW, pp. 270–274.

409. S. R. Mikkelsen, Electroanalysis 1996, 8,15–19.

410. J. Wang, G. Rivas, X. Cai et al., Anal. Chim.Acta 1997, 347, 1–8.

411. E. Palecek, M. Fojta, M. Tomschik et al.,Biosens. Bioelectron. 1998, 13, 621–628.

412. J. Wang,Anal. Chem. 1999, 71, 328R–332R.413. J. Wang, Chem.-A Eur. J. 1999, 5,

1681–1685.414. J. Wang, G. Rivas, X. H. Cai, Electroanalysis

1997, 9, 395–398.415. J. Wang, J. R. Fernandes, L. T. Kubota, Anal.

Chem. 1998, 70, 3699–3702.416. J. Wang, G. Rivas, J. R. Fernandes et al.,

Electroanalysis 1998, 10, 553–556.417. K. Yamashita, S. Takenaka, M. Takagi, Nu-

cleic Acids Symp. Ser. 1999, 42, 185, 186.418. J. Wang, X. H. Cai, G. Rivas et al., Anal.

Chem. 1996, 68, 2629–2634.419. G. Marrazza, I. Chianella, M. Mascini,

Biosens. Bioelectron. 1999, 14, 43–51.420. K. Hashimoto, K. Ito, Y. Ishimori, Sens.

Actuators 1998, 46, 220–225.421. S. O. Rolley, J. K. Barton, N. M. Jackson

et al., Bioconjugate Chem. 1997, 8, 31.422. Y. Okahata, Y. Matsunobu, K. Ijiro et al.,

J. Am. Chem. Soc. 1992, 114, 8299, 8300.423. K. A. Peterlinz, R. M. Georgiadis, J. Am.

Chem. Soc. 1997, 119, 3401, 3402.424. K. Hashimoto, K. Miwa, M. Goto et al.,

Supramol. Chem. 1993, 2, 265–270.425. K. Hashimoto, K. Ito, Y. Ishimori, Anal.

Chim. Acta 1994, 286, 219–224.426. S. Takenaka, M. Takagi, Y. Uto et al., Nu-

cleic Acids Symp. Ser. 1998, 39, 107, 108.427. S. Takenaka, M. Takagi, Bull. Chem. Soc.

Jpn. 1999, 72, 327–337.428. T. Ihara, M. Nakayama, M. Murata et al.,

Chem. Commun. 1997, 1609, 1610.429. S. Takenaka, Y. Uto, H. Kondo et al., Anal.

Biochem. 1994, 218, 436–443.430. Y. Uto, H. Kondo, M. Abe et al., Anal.

Biochem. 1997, 250, 122–124.431. M. E. Napier, C. R. Loomis, M. F. Sistare

et al., Bioconjugate Chem. 1997, 8, 906–913.432. J. Wang, G. Rivas, J. R. Fernandes et al.,

Anal. Chim. Acta 1998, 375, 197–203.433. C. Berggren, P. Stalhandske, J. Brundell

et al., Electroanalysis 1999, 11, 156–160.434. E. Souteyrand, J. P. Cloarec, J. R. Mar-

tin et al., J. Phys. Chem. B 1997, 101,2980–2985.

435. A. Bardea, F. Patolsky, A. Dagan et al.,Chem. Commun. 1999, 21–22.

436. F. Garnier, in Biomedical Chemistry: Ap-plying Principles to the Understanding andTreatment of Disease (Ed.: P. F. Torrence),John Wiley & Sons, Chichester, 2000,pp. 349–369.


437. T. A. Skotheim, Handbook of ConductingPolymers, Marcel Dekker, New York, 1986.

438. T. de Lumley-Woodyear, C. N. Campbell,A. Heller, J. Am. Chem. Soc. 1996, 118, 5504,5505.

439. T. de Lumley-Woodyear, P. Rocca, J. Lindsayet al., Anal. Chem. 1995, 67, 1332–1338.

440. Y. Degani, A. Heller, J. Phys. Chem. 1987,91, 1285–1289.


442. I. Katakis, A. Heller, in Frontiersin Bioelectronics I. Fundamental Aspects(Eds.: F. W. Scheller, F. Schubert,J. Fedrowitz), Birkhauser Verlag, Basel,1997, pp. 229–241.

443. J. Wang, M. Jiang, B. Mukherjee, Anal.Chem. 1999, 71, 4095–4099.

444. S. Cosnier, Biosens. Bioelectron. 1999, 14,443–456.

445. S. Cosnier, Electroanalysis 1997, 9, 894–902.446. F. Palmisano, G. Zambonin, D. Cen-

toze, Fresenius’ J. Anal. Chem. 2000, 366,586–601.

447. J. Wang, M. Jiang, A. Forbes et al., Anal.Chim. Acta 1999, 402, 7–12.

448. J. Li, G. Cheng, S. Dong, Electroanalysis1997, 9, 834–837.

449. K. Slowinski, R. V. Chamberlein, R. Bile-wicz et al., J. Am. Chem. Soc. 1996, 118,4709–4710.

450. K. Slowinski, R. V. Chamberlain, C. J.Miller et al., J. Am. Chem. Soc. 1997, 119,11 910–11 919.

451. D. Mandler, I. Turyan, Electroanalysis 1996,8, 207–213.

452. N. Muskal, I. Turyan, D. Mandler, J. Elec-troanal. Chem. 1996, 409, 131–136.

453. N. Muskal, D. Mandler, Electrochim. Acta1999, 45, 537–548.

454. T. Lumley-Woodyear, D. J. Caruana, C. N.Campbell et al., Anal. Chem. 1999, 71,394–398.

455. (a) D. J. Caruana, A. Heller, J. Am. Chem.Soc. 1999, 121, 769–774; (b) E. Palecek,R. Kizek, L. Havran et al., Anal. Chim. Acta2002, in press.

456. T. Lumley-Woodyear, C. N. Campbell, A.Heller, J. Am. Chem.Soc. 1996, 118, 5504,5505.

457. H. J. Helbock, K. B. Beckman, M. K. Shige-naga et al., Proc. Natl. Acad. Sci. 1998, 95,288–293.

458. E. Palecek, Biochim. Biophys. Acta 1967, 145,410–417.

459. J. Puranen, M. Forss, Strahlentherapie 1983,159, 505–507.

460. J. M. Sequaris, P. Valenta, H. W. Nurnberg,Int. J. Radiat. Biol. Related Stud. Phys. Chem.Med. 1982, 42, 407–415.

461. J.-M. Sequaris, P. Valenta, H. W. Nurnberget al., Bioelectrochem. Bioenerg. 1978, 5,483–503.

462. D. Krznaric, B. Cosovic, J. Stuber et al.,Chem.-Biol. Interact. 1990, 76, 111–128.

463. M. Fojta, T. Kubicarova, E. Palecek, Electro-analysis 1999, 11, 1005–1012.

464. M. Fojta, T. Kubicarova, E. Palecek, Biosens.Bioelectron. 2000, 15, 107–115.

465. R. Buresova, Ph.D. Thesis, Masaryk Univer-sity, Brno, 1997.

466. V. Brabec, Electrochim. Acta 2000, 45.467. J. Wang, M. Chicharro, G. Rivas et al., Anal.

Chem. 1996, 68, 2251–2254.468. M. Vorlickova, E. Palecek, Int. J. Radiat.

Biol. 1974, 26, 363–372.469. J. Wang, G. Rivas, M. Ozsos et al., Anal.

Chem. 1997, 69, 1457–1460.470. M. Rodriguez, T. Kodadek, M. Torres et al.,

J. Bioconjugate Chem. 1990, 1, 123–131.471. E. Palecek, J. Doskocil, Anal. Biochem. 1974,

60, 518–530.472. E. Lukasova, M. Vojtiskova, E. Palecek, Bio-

electrochem. Bioenerg. 1980, 7, 671–684.473. P. Boublikova, M. Vojtiskova, E. Palecek,

Anal. Lett. 1987, 20, 275–291.474. E. Palecek, Z. Pechan, Anal. Biochem. 1971,

42, 59–71.475. M. Vorlickova, E. Palecek, Biochim. Biophys.

Acta 1973, 331, 276–282.476. X. Cai, G. Rivas, P. A. M. Farias et al., Anal.

Chim. Acta 1996, 332, 49–57.477. M. J. Friedrich, Lab. Med. 1999, 30,

181–189.478. J. Hodgson, Nat. Biotechnol. 1998, 16,

725–727.479. R. F. Service, Science 1998, 282, 396–399.480. E. K. Wilson, Chem. Eng. News 1998, 76,

47–49.481. W. Gopel, P. Heiduschka, Biosens. Bioelec-

tron. 1994, 9, 3–13.482. D. T. Burke, M. A. Burns, C. Mastrangelo,

PCR Methods Appl. 1997, 7, 189–197.483. J. Cheng, E. L. Sheldon, L. Wu et al., Nat.

Biotechnol. 1998, 16, 541–546.484. E. L. Sheldon, Clin. Chem. 1993, 39,

718–722.


485. G. Yershov, V. Barsky, E. Belgovskiy et al.,Proc. Natl. Acad. Sci. 1996, 93, 4913.

486. U. Maskos, E. M. Southern, Nucleic AcidsRes. 1992, 20, 1675–1684.

487. E. Palecek, M. Fojta, Anal. Chem. 2001, 73,74A–83A.

488. E. Palecek, S. Billova, L. Havran et al., Ta-lanta 2002, in press.

489. E. Palecek, M. Fojta, F. Jelen, Bioelectrochem.2001, in press.

490. E. Palecek, R. Kizek, L. Havran et al., Anal.Chim. Acta 2001, in press.

491. J. Wang, A.-N. Kawde, A. Erdem et al., An-alyst 2001, 126, 2020–2024.

492. J. Wang, A. Xu, A. Erdem et al., Talanta,2002, in press.

493. R. Kizek, L. Havran, M. Fojta. etal., Bioelec-trochem. 2001, in press.

494. M. Fojta, L. Havran, R. Kizek. etal., Talanta,2002, in press.

495. J. Wang, R. Polsky, D. Xu, Langmuir 2001,17, 5739–5741.

496. R. J. Heaton, A. W. Peterson, R. M. Geor-giadis, Proc. Natl. Acad. Sci. USA 2001, 98,3701–3704.

497. Electrochemistry of nucleic acis and devel-opment of electrochemical DNA sensors,Talanta, Special issue 2002.

498. L. Alfonta, A. K. Singh, I. Willner, Anal.Chem. 2001, 73, 91–102.

499. P. M. Armistead, H. H. Thorp, Anal. Chem.2000, 72, 3764–3770.

500. P. M. Armistead, H. H. Thorp, Anal. Chem.2001, 73, 558–564.

501. L. Authier, C. Grossiord, P. Brossier, Anal.Chem. 2001, 73, 4450–4456.

502. F. Azek, C. Grossiord, M. Joannes etal.,Anal. Biochem. 2000, 284, 107–113.

503. P. K. Bhattacharya, J. K. Barton, J. Am.Chem. Soc. 2001, 123, 8649–8656.

504. G. Bidan, M. Billon, K. Galasso etal., Appl.Biochem. Biotechnol. 2000, 89, 183–193.

505. E. M. Boon, D. M. Ceres, T. G. Drummondetal., Nature Biotechnol. 2000, 18,1096–1100.

506. A. Erdem, K. Kerman, B. Meric etal., Anal.Chim. Acta 2000, 422, 139–149.

507. C. H. Fan, G. X. Li, Q. R. Gu etal., Anal.Lett. 2000, 33, 1479–1490.

508. S. Hason, J. Dvorak, F. Jelen etal., Talanta2002, in press.

509. G. Marrazza, G. Chiti, M. Mascini etal.,Clin. Chem. 2000, 46, 31–37.

510. G. Marrazza, S. Tombelli, N. Mascini etal.,Clin. Chim. Acta 2001, 307, 241–248.

511. M. Mascini, I. Palchetti, G. Marrazza, Fres.J. Anal. Chem. 2001, 369, 15–22.

512. D. T. Odom, J. K. Barton, Biochemistry 2001,40, 8727–8737.

513. E. Palecek, Talanta 2002, in press.514. F. Patolsky, A. Lichtenstein, I. Willner,

Angew. Chem. Int. Ed. Engl. 2000, 39,940–943.

515. M. I. Pividori, A. Merkoci, S. Alegret, Ana-lyst 2001, 126, 1551–1557.

516. E. Souteyrand, C. Chen, J. P. Cloarecetal., Appl. Biochem. Biotechnol. 2000, 89,195–207.

517. S. Takenaka, K. Yamashita, M. Takagi etal.,Anal. Chem. 2000, 72, 1334–1341.

518. S. Takenaka, Bull. Chem. Soc. Japan 2001,74, 217–224.

519. A. Tani, A. J. Thomson, J. N. Butt, Analyst2001, 126, 1756–1759.

520. C. Xu, H. Cai, P. G. He etal., Analyst 2001,126, 62–65.

431

13Enzyme Electrodes

Frieder W. Scheller and Ulla WollenbergerUniversity of Potsdam, Golm, Germany

13.1 Coupling of Enzyme-Catalyzed Reactions with ElectrochemicalIndication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433

13.2 Biochemical Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43313.2.1 Catalytic Action of Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . 43313.2.2 Classification of Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43513.2.2.1 Oxidoreductases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43513.2.2.2 Transferases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43513.2.2.3 Hydrolases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43613.2.2.4 Lyases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43613.2.2.5 Isomerases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43613.2.2.6 Ligases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436

13.3 Configurations of Enzyme Electrodes . . . . . . . . . . . . . . . . . . . . . 437

13.4 Immobilization of Enzymes on Electrode Surfaces . . . . . . . . . . . . 43813.4.1 Methods of Immobilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43813.4.1.1 Adsorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43813.4.1.2 Gel Entrapment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43813.4.1.3 Covalent Coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43913.4.1.4 Crosslinking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43913.4.2 Immobilization Effects in Enzyme Electrodes . . . . . . . . . . . . . . . . 439

13.5 Types of Enzyme Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . 441

13.6 Performance Parameters of Mono-Enzyme Electrodes [13] . . . . . . . 44613.6.1 Concentration Dependence of the Signal (Measuring Range) . . . . . 44613.6.2 pH Dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44713.6.3 Temperature Dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447

432 13 Enzyme Electrodes

13.7 Coupled Enzyme Reactions in Electrochemical Enzyme Sensors . . . 44713.7.1 Sequential Enzyme Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . 44813.7.2 Enzyme Competition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45113.7.3 Accumulation of Intermediate . . . . . . . . . . . . . . . . . . . . . . . . . . 45213.7.4 Amplification by Analyte Recycling . . . . . . . . . . . . . . . . . . . . . . . 45213.7.4.1 Enzymatic Recycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45313.7.4.1.1 Linear Recycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45313.7.4.1.2 Exponential Recycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45613.7.4.2 Bioelectrochemical Recycling . . . . . . . . . . . . . . . . . . . . . . . . . . . 45613.7.5 Sequential Activation of Enzymes . . . . . . . . . . . . . . . . . . . . . . . . 456

13.8 Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 457References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 458

433

13.1Coupling of Enzyme-Catalyzed Reactionswith Electrochemical Indication

Traditionally, enzymes are used as an-alytical reagents to measure substratemolecules by catalyzing the turnover ofthese species to detectable products. Inaddition, compounds modifying the rateof the enzyme reaction, such as acti-vators, prosthetic groups, inhibitors andenzymes themselves, are accessible to themeasurement [1]. Owing to their excellentchemical specificity, enzymes allow thedetermination of minute amounts in com-plex media and thus avoid the need ofhighly sophisticated instrumentation. Fur-thermore, when enzymes are employedas labels in binding assays using antibod-ies, binding proteins, lectins, and so on,the inherent chemical amplification prop-erties of the enzyme’s catalytic activity canbe exploited to realize extremely sensitiveassay methods (Table 1).

Finally, biologically related parameters,for example, taste, odor, fatigue sub-stances, mutagenicity, and nutritivity arequantifiable by using multienzyme sys-tems, intact organelles or cells [2, 3].

Electrochemical sensors are well-estab-lished tools in the determination ofgases ion activities, and oxidizable andreducible organic substances down to

the submicromolar concentration range.The analysis of many other importantsubstances by electrochemical sensorsrequires the coupling with an enzymaticreaction which involves an electroactivespecies (Table 2).

13.2Biochemical Fundamentals

The electrodes determine mainly the out-put of the electroenzymatic process. Incontrast, the analytical selectivity is de-termined by the specificity of the signal-producing interaction of the enzyme withthe analyte. Moreover, the properties of theenzyme, such as its specific activity, influ-ence the dynamic range, and the sensitivityof biosensors. In this respect, account hasto be taken of the fact that enzymatic pro-cesses are susceptible to deviations fromtheir optimal environmental conditions;in particular, their thermal and chemicalstability is limited. These peculiarities deci-sively determine the limits of applicabilityof enzymes.

13.2.1Catalytic Action of Enzymes

A prerequisite for the catalytic function ofan enzyme is its native structure, which is


Tab. 1 Analytical potential of enzymes

• Efficient substrate conversion: up to 105 molecules per sec• Sensitivity for substrates: Millimolar–Nanomolar• Sensitivity for inhibitors: Millimolar–Picomolar• High chemical selectivity• Measurable substances: Substrates

CosubstratesEnzymes activityProsthetic groupsInhibitors/Activators

• Signal generation and amplification in binding assaysImmunoassayDNA hybridization

Tab. 2 Coupling of enzymes with electrochemical sensors in enzyme electrodes

Biocomponents

Oxidoreductases Hydrolases:Dehydrogenases ProteasesOxidases EsterasesPeroxidases GlycosidasesElectron-transferases

Transferases:KinasesTransaminases

Indicated species

Cosubstrates:NAD(P)HO2/H2Mediators

Products: Products:Phenols H+Redox dyes HCO3

−NH4

+

Prosthetic groups: I−Heme F−PQQFADCu2+

Electrode type Amperometric electrodes Potentiometric electrodes

Note: PQQ: Pyrroloquinolinequinone tricarboxylic acid.

13.2 Biochemical Fundamentals 435

determined by the number and sequenceof amino acids (primary structure) formingthe molecule. Favored by hydrogen bonds,parts of the polypeptide chain exist in a α-helical or a β-sheet structure (secondarystructure). Most enzymes are globularproteins, the tertiary structure of whichmay be fixed by disulfide bonds betweencysteine residues.

In spite of these stabilizing interactions,the ordered three-dimensional structure isonly stable below 50 C and at mediumpH. Exceptions are enzymes from mi-croorganisms that have adapted to extremeenvironmental conditions, for example,temperatures up to 90 C.

Within the mostly spherical, ellipsoidal,or kidney-shaped protein molecules a localcavity with a characteristic constitution andstereoconfiguration forms the catalyticallyactive center, where a chemically andspatially congruent substrate (‘‘lock-and-key principle’’) is converted to a product.To a limited extent, the protein structureis capable of adapting conformationally tothe substrate.

Enzymes accelerate the equilibriumformation of chemical reactions by a factorof 108 –1020 as compared with uncatalyzedreactions. Thus urea is hydrolyzed at pH 8and 20 C in the presence of urease about1014 times faster than without catalysis,and the splitting of H2O2 is acceleratedby a factor of 3 × 1011 in the presence ofcatalase [4].

13.2.2Classification of Enzymes

The roughly 3000 enzymes currentlyknown are grouped into six main classesaccording to the type of the reactioncatalyzed [5]. At present only a limitednumber are used for analytical purposes.

13.2.2.1 Oxidoreductases(EC 1.X.X.X) catalyze oxidation and reduc-tion reactions by transfer of hydrogen orelectrons. The following are of analyticalimportance:

1. dehydrogenases catalyze hydride transferfrom the substrate, S, to an acceptor, A(which is not molecular oxygen), or viceversa:

SH2 + A ⇐⇒ S + AH2

Example: lactate dehydrogenase (EC1.1.1.27)

L-lactate + NAD+ ⇐⇒pyruvate + NADH + H+

2. oxidases catalyze hydrogen transferfrom the substrate to molecular oxygen:

SH2 + 12 O2 ===⇒ S + H2O or

SH2 + O2 ===⇒ S + H2O2

Example: glucose oxidase (EC 1.1.3.4)

β-D-glucose + O2 ===⇒gluconolactone + H2O2

3. peroxidases catalyze oxidation of a sub-strate by hydrogen peroxide:

2 SH + H2O2 ===⇒ 2 H2O + 2 S

Example: horseradish peroxidase (EC1.11.1.7)

H2O2 + 2[Fe(CN)6]4− + 2H+ ===⇒2[Fe(CN)6]3− + 2H2O

13.2.2.2 Transferases(EC 2.X.X.X) transfer C-, N-, P-, or S-containing groups (alkyl, acyl, aldehyde,


amino, phosphate, glycosyl) from one sub-strate to another. Transaminases, trans-ketolases, transaldolases and transmethy-lases belong to this class:

AX + B ⇐⇒ A + BX

Example: hexokinase (EC 2.7.1.1)

D-hexose + ATP ⇐⇒D-hexose-6-phosphate + ADP

13.2.2.3 Hydrolases(EC 3.X.X.X) catalyze cleavages or thereverse fragment condensation. Accordingto the type of bond cleaved, a distinctionis made between peptidases, esterases,glycosidases, phosphatases, and so on.

Examples: cholesterol esterase (EC 3.1.1.13)

Cholesterol ester + H2O ===⇒cholesterol + fatty acid,

alkaline phosphatase (EC 3.1.3.1)

orthophosphate monoester + H2O ===⇒alcohol + orthophosphate

13.2.2.4 Lyases(EC 4.X.X.X) nonhydrolytically removegroups from their substrates under for-mation of double bonds, or add groups todouble bonds. Only a few enzymes of thisclass are used in analysis.

13.2.2.5 Isomerases(EC 5.X.X.X) catalyze intramolecular rear-rangements and are subdivided into race-mases, epimerases, mutases, cis-trans-isomerases, and so on.

Examples: glucose isomerase (EC 5.3.1.5)

D-glucose ⇐⇒ D-fructose

Mutarotase (aldose-1-epimerase) (EC 5.1.3.3)

α-D-glucose ⇐⇒ β-D-glucose

13.2.2.6 Ligases(EC 6.X.X.X) split C−C, C−O, C−N, C−S,and C-halogen bonds without hydrolysisor oxidation, mostly with the concomitantconsumption of high-energy compoundslike adenosine triphosphate (ATP) andother nucleoside triphosphates.

Example: pyruvate carboxylase (EC 6.4.1.1)

pyruvate + HCO3− + ATP ⇐⇒

oxaloacetate + ADP + Pi

Several components are required forthe catalytic process, which are eitherdirectly involved in catalysis or influencethe formation of the enzyme-substratecomplex. They are designated coenzymes,prosthetic groups, and effectors.

Coenzymes receive redox equivalents,protons, or chemical groups from the sub-strate during the enzymatic reaction. Sincecoenzymes readily dissociate from the en-zyme they can act as group mediatorsbetween different enzyme molecules (e.g.coenzyme A). In case the factor is not re-converted to its original state by the sameenzyme, it is called cosubstrate. An exam-ple would be cleavage of the energy-richATP to adenosine diphosphate (ADP) andphosphate during energy-consuming sub-strate conversion and the regeneration ofthe ATP by adenylate kinase. Many vitaminderivatives, such as coenzyme A, pyridox-alphosphate, thiamine pyrophosphate, andcobalamine (Vitamin B12) are coenzymes

13.3 Configurations of Enzyme Electrodes 437

of enzymatic reactions. Oxidative coen-zymes with a defined redox potential serveas hydrogen or electron carriers duringoxidoreduction reactions.

Prosthetic groups have the same func-tion as coenzymes but are tightly boundto the enzyme. When they are split off,the protein is mostly denatured. Flavin nu-cleotides and heme are the most importantprosthetic groups.

Effectors accelerate (activators) or block(inhibitors) the catalytic process. Many ofthem are metal ions, for example, Mg++,Ca++, Zn++, K+, and Na+, which ei-ther form stoichiometric complexes withthe substrate, stabilize an optimal proteinconformation, or effect the association ofsubunits. These inorganic complementsof enzyme reactions are frequently sub-sumed together with coenzymes as cofac-tors.

13.3Configurations of Enzyme Electrodes

An enzyme electrode (Fig. 1) is a densepackage of dialyzer, enzyme reactor, anddetector. A typical example would be a

glucose or lactate electrode comprisingthe appropriate oxidase entrapped in orbound to a membrane that is fixed atan oxygen or hydrogen peroxide detectingelectrode and covered by a semipermeablemembrane [6].

The membrane directed towards the (usu-ally ideally mixed or flowing) measuringsolution fulfils a number of functions.Firstly, it provides the sensor with a cer-tain degree of selectivity. The pore size and,perhaps, charge permits the exclusion ofdeleterious or interfering molecules, suchas proteins or electroactive compoundsand may provide useful partitioning ofother compounds. Furthermore, the thick-ness and pore size permit us to affect themeasuring range of the sensor by control-ling the actual rate of reagents reachingthe reaction layer.

After permeation of analyte, cosub-strates, and effectors through the mem-brane to the underlying enzyme layer theanalyte is converted therein under for-mation or consumption of a detectablespecies. In the illustrated case this isthe formation of reduced mediator or co-substrate (e.g. hydrogen peroxide fromoxygen) that is oxidized at the electrode

Selectivitysensitivity

response time

ne−

Measuringrange

Analyte

Potentiometric

Electrode Enzyme layerPermselective

membrane

Amperometric

E

Mo

Mred

Product

H+

Fig. 1 Scheme of an enzyme electrode.


or the generation of protons. The enzymemembrane is characterized by an enzymeloading reflecting the interplay of enzymekinetics and mass transport. The loadingis crucial for the response characteristicsand the stability of the sensor whereas thechoice of enzyme determines the chemicalselectivity of the measurement.

The choice of the indicator electrodeis largely determined by the speciesinvolved in the sensing reaction. Oxy-gen and H2O2, which are the cosub-strate and product of oxidases, as wellas NAD(P)H, the cosubstrates of about300 pyridine nucleotide-dependent dehy-drogenases, can be determined amper-ometrically. Hydrolases are mostly cou-pled to ion selective electrodes. Based onthese principles, many enzyme sensorshave been developed and commercial-ized [2, 3].

13.4Immobilization of Enzymes on ElectrodeSurfaces

13.4.1Methods of Immobilization

For the repeated use of enzymes in an-alytical devices, numerous techniques forfixing them to carrier materials includingmembranes and electrode surfaces havebeen developed [3]. Immobilization bringsabout a number of further advantagesfor their application in analytical chem-istry.

• In many cases the enzyme is stabilized;• The immobilized enzyme may be easily

separated from the sample;• The stable and largely constant enzyme

activity renders the enzyme an integralpart of the analytical instrument.

The techniques for immobilizing en-zymes comprise physical and chemi-cal methods as well as combinationsof both [7]. The main physical methodsare adsorption to water-insoluble carriersor surfaces and entrapment in water-insoluble polymeric gels. Chemical immo-bilization is effected by covalent couplingto derivatized carriers or by intermolecularcrosslinking of the biomolecules. The suit-ability of a method for a particular task is atpresent still being empirically elucidated.However, some generally valid aspects willbe outlined below.

13.4.1.1 AdsorptionAdsorption of biomolecules at the elec-trode surface is the simplest method ofimmobilization. An aqueous solution ofthe biomolecules is contacted with the ac-tive surface for defined period. Thereafterthe molecules that are not adsorbed areremoved by washing.

Since the adsorption of a protein toa functionalized surface is a reversibleprocess, changes of pH, ionic strength,substrate concentration, temperature, andso on may detach the biomolecule. In addi-tion to the simplicity of the procedure, theadvantage of adsorptive immobilization isthat it does not need nonphysiological cou-pling conditions or chemicals potentiallyimpairing enzyme or cell functions.

13.4.1.2 Gel EntrapmentEntrapment in polymeric gels preventsthe biomolecules from diffusing from thereaction mixture. On the other hand,small substrate and effector moleculescan easily permeate. Gel entrapment isa milder procedure than adsorption, thatis, the biomolecules are not covalentlybound to the matrix, membrane or toeach other. The method is therefore

13.4 Immobilization of Enzymes on Electrode Surfaces 439

widely employed. The most importantmatrices used are hydrogels such asalginate, carageenan, collagen, cellulosetriacetate, polyacrylamide, gelatin, agar,silicone rubber, and poly(vinyl alcohol).

13.4.1.3 Covalent CouplingCovalently coupled enzymes are eitherreacted with the activated (e.g. function-alized) surface or copolymerized with areactive monomer. The reaction should in-volve only groups that are not essential forthe biological activity of the biomolecule.Chemically reactive sites of a protein maybe amino groups, carboxyl groups, phenolresidues of tyrosine, sulfhydryl groups orthe imidazole group of histidine. The im-mobilization is conducted in three steps:activation of the carrier, coupling of thebiomolecule, and removal of adsorbed butnot covalently attached biomolecules. Adisadvantage of covalent coupling is thefrequently occurring loss of activity [8].

13.4.1.4 CrosslinkingEnzymes may be intermolecularly cross-linked by bi- or multifunctional reagents.The protein molecules may be crosslinkedwith each other or with another, function-ally inert protein (e.g. albumin or gelatin).The biomacromolecules can also be ad-sorbed to a carrier or entrapped in a geland than crosslinked. Among others, glu-taraldehyde, biisocyanate derivatives, andbisdiazobenzidine are being used as bi-functional reagents [8].

The advantages of crosslinking are thesimple procedure and the strong chemicalbinding of the biomolecules. Furthermore,the choice of the degree of crosslinkingpermits the physical properties to beinfluenced. The main drawback is thepossibility of activity losses due to chemical

alterations of the catalytically essential sitesof the protein.

13.4.2Immobilization Effects in EnzymeElectrodes

Both for economic reasons and in orderto achieve a high sensitivity and functionalstability, immobilization methods havinga high activity yield are desirable forbiosensors [2, 8].

Whilst in homogenous solution the ini-tial rate of substrate conversion riseslinearly with enzyme concentration. Withimmobilized enzymes, the measured reac-tion rate depends also on substrate masstransfer.

In enzyme electrodes the biocatalystand the signal transducer are spatiallycombined, that is, the enzyme reactionproceeds in a layer separated from themeasuring solution. The substrates reachthe membrane system of the biosensorby convective diffusion from the solution.The rate of this external transport processdepends essentially on the degree ofmixing. In the multilayer system infront of the sensor the substrates andproducts are transferred by diffusion. Slowmass transfer to and within the enzymematrix leads to different concentrations ofthe reaction partners in the measuringsolution and in the matrix. Diffusionand the enzyme reaction do not proceedindependently of one another; they arecoupled in a complex manner.

Usually in the operation of biosensorsthe flow conditions are adjusted to providea mass transfer rate from the solutionto the membrane system which is fastas compared with the internal masstransfer (exception: implanted sensors).On the other hand, variations of thediffusion resistance of the semipermeable


membrane are being used to optimizethe sensor performance. A semipermeablemembrane with a molecular weight cutoffof 10.000 and a thickness of 10 µm onlyslightly influences the response time andsensitivity. In contrast, thicker membranesincrease the measuring time, but mayalso lead to an extension of the linearmeasuring range.

The ratio of the rate of the enzymaticreaction to that of diffusion indicateswhether the process in an enzyme layeris determined by enzyme kinetics or bysubstrate diffusion. At low enzyme activity,the process is kinetically controlled. Inthis case the substrate concentrationdoes not become zero in any part ofthe enzyme layer, that is, the enzymesensor signal is mainly a function of the‘‘active’’ enzyme concentration. Therefore,effectors (activators, inhibiting factors,including H+, and OH−) and, the amountof enzyme in front of the tranducer,as well as the time-dependent enzymeinactivation, may all directly effect themeasuring signal.

At high enzyme loadings internal dif-fusion control is reached. Any substratemolecule diffusing into the enzyme layeris immediately converted therein; onlypart of the enzyme is acting catalytically.Diffusion controlled sensors exhibit thefollowing characteristics as long as an en-zyme reserve is present:

• the sensitivity remains constant;• the sensitivity does not depend on

inhibitors and pH variation;• the temperature is of minor influence

since the activation energy of diffusion(∼2.5%/ C) is lower than that of theenzyme reaction (10%/ C).

At high substrate concentration theenzyme reaction rate attains a limiting

value. Therefore the enzyme sensor sig-nal reaches a concentration-independentvalue corresponding to the product con-centration at the transducer surface.

From the analysis of the coupling ofenzyme reaction and mass transfer, thefollowing conclusions may be drawn forthe design of biosensors. The substrateconcentration at which deviations from theanalytically usable linear measuring rangeoccur depends on the extent of diffusionlimitation. Under kinetic control, a lineardependence may only be expected forvery low substrate concentrations. Underdiffusion control, the decrease of substrateconcentration in the enzyme layer causedby slow substrate diffusion results inan extended linear range. It has to beconsidered, however, that for two-substratereactions deviations from linearity mayalso be produced by cosubstrate limitation.

At low substrate concentration, the sen-sitivity of kinetically controlled sensorsincreases linearly with the enzyme load-ing. When the amount of enzyme becomessufficiently high as to provide completesubstrate conversion the system passesover to diffusion control. Under theseconditions, a decrease of the diffusion re-sistance by decreasing the layer thicknessresults in an increased sensitivity. Nev-ertheless, a membrane-covered enzymeelectrode is only 10 to 50% as sensitive as abare electrode for an analogous electrode-active substance.

The variation of the enzyme loading isa means of determining the minimumamount of enzyme required for maximumsensitivity. Furthermore, this test revealsthe magnitude of the enzyme reserve ofdiffusion-controlled sensors.

Owing to the excess of enzyme in themembrane, a diffusion limited enzymesensor has a higher functional stabilitythan a kinetically controlled one. With

13.5 Types of Enzyme Electrodes 441

the former, 2000 to 10 000 measurementsper enzyme membrane can be performed,while kinetically controlled sensors typ-ically permit only 200 to 500 measure-ments.

The response time is determined bythe mass transfer. Using fast-respondingtransducers in stationary measurements, astable signal is obtained within one secondup to a few minutes.

In summary, it may be concluded thatoptimal sensitivity and response time canbe achieved by applying high enzymeactivity in thin membranes [2].

13.5Types of Enzyme Electrodes

The oxygen electrode according to Clarkand its version modified for H2O2 detec-tion are the most widely used transducersin biosensors [3]. The electrode potentialis crucial for the selectivity of the sensor.Any electroactive substance being con-verted at lower potential contributes to thetotal current. Thus at an electrode poten-tial of +600 mV for H2O2 measurement,ascorbic acid, uric acid, or paracetamol(acetaminophen) are oxidized as well.

A way to reduce interferences by cooxi-dizable sample constituents is by keepingthe applied electrode potential as low aspossible. Therefore, a reaction partner ischosen to be electrochemically indicatedthat is converted at low potential. Forthis purpose, the natural electron accep-tors of many oxidoreductases have beenreplaced by redox-active dyes or other re-versible electron mediators. Among themare the ferricyanide/ferrocyanide cou-ple, N -methylphenazinium sulfate, fer-rocenes, and benzoquinone. With thesemediators an electrode potential around+200 mV can be applied, which decreases

electrochemical interferences and permitsus to apply such enzymes coupled withelectrodes in oxygen-free solution. In anal-ogy to the natural cosubstrate the mediatoris often added to the sample solution [3, 9].However, homogeneous reaction of themediators with endogenous electron ac-ceptors such as ascorbate must be avoided.

The integration of redox enzymes at elec-trode surfaces together with such mediatorcompounds to act like electron trans-fer systems in biomembranes, enables areagentless measuring regime. The cova-lent binding or adsorption of mediator andenzyme to the electrode or their integra-tion into the electrode body itself, as incase of carbon paste electrodes, have beenshown to be successful concepts that leadto functioning glucose sensors [9–11].

Adsorption of redox polymers at car-bon electrodes result in the catalysis ofthe electron transfer by wiring the en-zyme molecules to the electrode. In thismanner sensors for glucose, hydrogenperoxide, and NAD(P)H were developedby wiring enzyme to glassy carbon elec-trodes, with an osmium complex. It wasshown that such a network is capable ofconnecting FADH2/FAD centers of glu-cose oxidase and pyrroloquinoline quinonecenters of glucose dehydrogenase to theelectrode [10, 11].

The rate of enzymatic reactions canalso be established by potentiometric mea-surement of product formation using anion selective electrode. The most impor-tant ion selective electrode is the glasselectrode for pH measurement. Despitetheir outstanding selectivity for H+ ions,glass electrodes are used only seldom inenzyme electrodes because their sensi-tivity is affected by the buffer capacityof the sample matrix or sensor fillersolution.

442 13 Enzyme ElectrodesTa

b.3

Mon

o-en

zym

eel

ectr

odes

for

som

ere

leva

ntan

alyt

es

Ana

lyte

Enzy

me

Indi

cate

dsp

ecie

s

O2

H2O

2M

edia

tor

Pros

the-

H+

NA

D(P

)HO

ther

sti

cgr

oup

Ace

tylc

holin

e,C

arba

mat

es,

Org

anop

hosp

hate

s

Ace

tylc

holin

ees

tera

seEC

3.1.

1.7

XX

XX

Alc

ohol

sA

lcoh

olox

idas

eEC

1.1.

3.13

XX

X

Alc

ohol

dehy

drog

enas

eEC

1.1.

3.13

(NA

D(P

)+-d

ept.)

X

D-A

min

oac

ids

D-A

min

oac

idox

idas

eEC

1.4.

3.3

XX

X

L-A

min

oac

ids

L-A

min

oac

idox

idas

eEC

1.4.

3.2

XX

XN

H4+ I

−

Asc

orbi

cac

idA

scor

bate

oxid

ase

EC1.

10.3

.3X

L-A

spar

agin

eA

spar

agin

ase

EC3.

5.1.

1X

Am

ygda

linβ

-Glu

cosi

dase

EC3.

2.1.

21C

N−

Bile

acid

β-H

ydro

xyst

eroi

dde

hydr

ogen

ase

EC1.

1.1.

51

X

Cho

lest

erol

Cho

lest

erol

oxid

ase

EC1.

1.3.

6X

X


Cho

line

Cho

line

oxid

ase

EC1.

1.3.

17X

XX

Cre

atin

ine

Cre

atin

ine

imin

ohyd

rola

seEC

3.5.

4.21

NH

3

D-F

ruct

ose

Fruc

tose

dehy

drog

enas

eEC

1.1.

99

XX

Gal

acto

seG

alac

tose

oxid

ase

EC1.

1.3.

9X

X

Glu

cose

Glu

cose

oxid

ase

EC1.

1.3.

4X

XX

XX

Glu

cose

dehy

drog

enas

eEC

1.1.

1.17

(NA

D+ -

dept

.)

X

Glu

cose

dehy

drog

enas

eEC

1.1.

99.1

7(P

QQ

-dep

t)

X

L-G

luta

mic

acid

Glu

tam

ate

dehy

drog

enas

eEC

1.4.

1.3

X

Glu

tam

ate

oxid

ase

EC1.

4.3.

11X

XX

L-G

luta

min

eG

luta

min

ase

EC3.

5.1.

2N

H3

Gly

cero

lG

lyce

rol

dehy

drog

enas

eEC

1.1.

1.16

X

Gly

cola

teG

lyco

late

oxid

ase

EC1.

1.3.

1X

X

Hyd

roge

nH

ydro

gena

seEC

1.9.

8.11

X

(con

tinue

dov

erle

af)


Tab.

3(c

ontin

ued)

Ana

lyte

Enzy

me

Indi

cate

dsp

ecie

s

O2

H2O

2M

edia

tor

Pros

the-

H+

NA

D(P

)HO

ther

sti

cgr

oup

Hyd

roge

npe

roxi

dePe

roxi

dase

EC1.

11.1

.7X

X

Cat

alas

eEC

1.11

.1.6

XIs

ocitr

ate

Isoc

itrat

ede

hydr

ogen

ase

EC1.

1.1.

42

X

Lact

icac

idLa

ctat

eox

idas

eEC

1.1.

3.2

XX

X

Cyt

ochr

ome

b 2EC

1.1.

2.3

XX

Lact

ate

dehy

drog

enas

eEC

1.1.

1.27

(NA

D+ -

dept

.)

X

Lact

ate

mon

ooxy

gena

seEC

1.13

.12.

4

XC

O2

Mal

ate

Mal

ate

dehy

drog

enas

eEC

1.1.

1.37

X

Nitr

iteN

itrite

redu

ctas

e(N

AD

(P)H

)EC

1.6.

6.4

XN

H3

Oxa

licac

idO

xala

teox

idas

eEC

1.2.

3.4

X

Peni

cilli

nβ

-Lac

tam

ase

peni

cilla

seEC

3.5.

2.6

X

Peni

cilli

nam

idas

eEC

3.5.

1.11

X


Phen

ols

Lacc

ase

EC1.

10.3

.2X

quin

one

Tyro

sina

seEC

1.14

.18.

1X

X

Pero

xida

seEC

1.11

.1.7

X

Phos

phat

ePy

ruva

teox

idas

eEC

1.2.

3.3

XX

Pyru

vate

Lact

ate

dehy

drog

enas

eEC

1.1.

1.27

(NA

D+ -

dept

.)

X

Pyru

vate

oxid

ase

EC1.

2.3.

3X

XX

Salic

ylat

eSa

licyl

ate

hydr

oxyl

ase

EC1.

14.1

3.1

XX

CO

2

Succ

inic

acid

Succ

inat

ede

hydr

ogen

ase

EC1.

3.99

.1

X

Sulfa

teA

ryls

ulfa

tase

EC3.

1.6.

1X

Supe

roxi

deSu

pero

xide

dism

utas

eEC

1.15

.1.1

XX

X

Trig

lyce

ride

sLi

pase

EC3.

1.1.

3X

Ure

aU

reas

eEC

3.5.

1.5

XN

H3C

O2

Uri

cac

idU

rate

oxid

ase

(uri

case

)EC

1.7.

3.3

XX

Xan

thin

e,H

ypox

anth

ine

Xan

thin

eox

idas

eEC

1.2.

3.2

XX


The selectivity of the glass electrode forNH3 and CO2 may be improved overthat of pH measurement by inclusionof a gas permeable membrane betweenthe enzyme layer and the pH electrode.At constant solution, pH a defined rela-tion exists between the potential of theglass electrode and the concentration ofthe gas-forming ions HCO3

− or NH4+

which are formed in the enzyme reac-tion. Maximum sensitivity of the electrodeis reached when the H+ concentrationin the solution is sufficient to ensuremaximum conversion of the weak elec-trolyte into its undissociated form, thatis CO2 or NH3. With NH3 this occursat pH > 10 and with CO2 at pH > 5.Generally, these pH values differ substan-tially from the pH optima of deaminaseand decarboxylase enzymes; therefore, forthe respective enzyme electrodes a com-promise pH has to be found. To ob-tain optimal conditions for both steps,the enzyme reaction is often separatedfrom the potentiometric indication anda pH change is included between thesestages. This setup is termed a reactorelectrode.

The semipermeable membrane maybe replaced by an air gap between themeasuring solution and the pH elec-trode. This increases the measuring ratebut affects the electrolyte layer andthus the reproducibility of the measure-ment.

Other pH-sensing transducers used inbiosensors are metal oxide electrodes.Beside the common antimony oxide elec-trode, palladium oxide and iridium oxideprobes have been coupled with immo-bilized enzymes. These sensors may beminiaturized by using chemical vapor de-position technology. Moreover, they aremechanically more stable than glass elec-trodes. Unfortunately response of metal

oxide electrodes is affected by redox activesubstances [3].

Furthermore, the gate area of pH-sensitive ISFET’s (ion-sensitive field-effect transistor) has been covered withpH changing enzymes, such as ure-ase and β-lactamase [12]. Furthermore,ATPase, glucose oxidase, and trypsinwere used for ATP, glucose, and pep-tide ENFET’s (enzyme-field-effect transis-tor).

Representative examples of monoen-zyme electrodes are presented in Table 3.

13.6Performance Parameters of Mono-EnzymeElectrodes [13]

13.6.1Concentration Dependence of the Signal(Measuring Range)

The linear measuring range of enzymeelectrodes extends over 2 to 5 decades ofconcentration. The lower detection limit ofsimple amperometric enzyme electrodesis about 100 nmol l−1 whereas potentio-metric sensors may be only applied downto 100 µmol l−1. This shows that the de-tection limit is affected not only by theenzyme reaction but also by the trans-ducer.

The measuring signal of the ampero-metric glucose electrode increases withincreasing substrate concentration andreaches a concentration-independent sat-uration corresponding to the maximumrate. The substrate concentration givingrise to the half-maximum current in airsaturated solution is between 0.5 and1.0 mmol l−1 glucose.

The linear range extends to 2 mmol l−1

glucose in the measuring cell. At lowglucose concentration the cosubstrate

13.7 Coupled Enzyme Reactions in Electrochemical Enzyme Sensors 447

concentration (ca. 200 µmol l−1 at air sat-uration) influences the response onlyslightly.

13.6.2pH Dependence

With a high enzyme excess in themembrane, pH variations should haveonly a minor influence on the measuringprocess. Therefore the pH profile underdiffusion control should be substantiallyless sharp than those of the respectiveenzyme in solution. The results obtainedwith GOD-membranes agree with thisassumption.

13.6.3Temperature Dependence

The rate of enzyme reactions rises withtemperature up to a certain optimum.Above that, the effect of the thermalinactivation dominates over that of theincrease of the collision frequency.

For immobilized enzymes, frequently anincrease of the temperature optimum forsubstrate conversion is observed. Becauseenzyme kinetics and diffusion are super-imposed, the higher activation energy re-sults in a predominant acceleration of theenzyme reaction with rising temperature.

13.7Coupled Enzyme Reactions inElectrochemical Enzyme Sensors

The number of substances that can bemeasured by monoenzymatic approachesin electrochemical biosensors is limited,because in the majority of biocatalytic reac-tions electrochemically active compoundsare not involved. To form readily detectablespecies, different enzymatic reactions haveto be coupled, as is already routine in wetbiochemical analysis [13]. This couplingcan be accomplished in ways analogous tothose present in a living cell. Here, natureprovides us a variety of ways of regulatingmetabolic pathways. Thus, like in nature,catalytic activities of different enzymes canbe combined in biosensors either in se-quence, competing pathways, or in cycles(Table 4).

In general, enzymes to be used inmultienzyme sensors should fulfill thefollowing requirements:

• Their pH optima should be reasonablyclose to each other.

• They should not be inhibited by co-factors, effectors, or intermediates re-quired for sensing.

• Their cofactors or effectors should notreact with each other.

Tab. 4 Coupling principles for design of sensor performance

Principle ofcoupling

Special cases Effect on performance

Sequence – New analyte specificityAccumulation Sensitivity specificityCascade New analyte sensitivity

Parallel reaction Competition elimination,antiinterference

New analyte specificitymutiparameter

Cycle Multiple cycles nonlinearcycle

Sensitivity


The most effective sensors work indiffusion control with respect to a singlereaction.

In conjunction with appropriate mea-suring regimes not only does a muchwider range of analyte species becomesin this way accessible to measurement bythe bioelectroanalytical approach also theselectivity and sensitivity of the biosen-sor may be enhanced through the ap-propriate choice of the coupling strat-egy [2, 14].

13.7.1Sequential Enzyme Reactions

An enzyme sequence electrode is a biosen-sor, where at least two enzymes metabo-lize a substrate in consecutive reactionswith the formation of a measurable sec-ondary product or consumption of cosub-strate. Obviously, the number of enzymes(E1–En) in such a reaction chain can beincreased as long as recognition of the pri-mary substrate S ends up in a detectablemetabolite P* (Fig. 2). The same is trueif a detectable cosubstrate is consumedin the final reaction. On this basis, fam-ilies of electrodes have been developed,which combine, for example, glucose-, lactate-, and glutamate-generating pri-mary enzyme reactions with the respectiveoxidases, that is, glucose oxidase, lac-tate oxidase, and glutamate oxidase (seeSect. 13.5.). The measurement of oxygenconsumption of the NAD(P)H oxidationby salicylate hydroxylase (E.C. 1.14.13.1)opened the way to the large group of

dehydrogenase substrates [14]. Enzyme se-quence electrodes given in Table 5 showthat indicator reaction based on oxidasesand dehydrogenases (see Table 3) are cou-pled with enzymes such as hydrolase,lyases, and transferases. Sensors for ox-idase substrates have also been studiedusing coimmobilized peroxidase for hydro-gen peroxide transformation [16]. Theseelectrodes work at lower electrode po-tential than is necessary for peroxideoxidation and are therefore less affected byelectrochemically interfering substances.In principle, all peroxide forming oxi-dases (E.C. 1.x.3.x) can be combined withperoxide detectors based on peroxidases.Soluble mediators, ‘‘electrically wired’’ en-zymes, and mediator-modified electrodesare used. In addition, also the apparent di-rect electron transfer between electrodeand peroxidase that occurs at a favor-able low electrode potential is exploitedfor the indication [14]. Thus, the influenceof interfering electrochemical reactions isminimized.

In general, the sensor performance iscomparable to those of monoezyme elec-trodes. In ideal cases, diffusion limitationis achieved by immobilizing a sufficientlyhigh amount of enzyme [17].

The enzyme sequence sensor approachwill be illustrated for the determination ofcitrate. Here three enzymes are immobi-lized on an oxygen sensor and the linearreaction sequence is employed:

Citrate lyase:

Citrate ===⇒ oxaloacetate + acetate

S P1 = S2

E1A C

P2 = Sn

E2

Electrode

A C

P*

EnA C

Fig. 2 Scheme of a sequentialenzyme sensor. A and Crepresent coreactants. Up to nenzymes can be used. P* iselectrochemically indicated.


Tab.

5En

zym

ese

quen

ceel

ectr

odes

with

oxid

ase

indi

cato

ren

zym

e

Ana

lyte

Enzy

me

sequ

ence

Indi

cato

ren

zym

e

Aux

iliar

yen

zym

esEC

num

ber

Enzy

me

ECnu

mbe

r

Sacc

aros

eIn

vert

ase

3.2.

1.26

Glu

cose

oxid

ase

1.1.

3.4

+m

utar

otas

e5.

1.3.

3La

ctos

eß

-gal

acto

sida

se3.

2.1.

23G

luco

seox

idas

e1.

1.3.

4M

alto

seG

luco

amyl

ase

3.2.

1.3

Glu

cose

oxid

ase

1.1.

3.4

orm

alta

se3.

2.1.

20+

mut

arot

ase

5.1.

3.3

Glu

cosi

nola

teM

yros

inas

e3.

2.3.

1G

luco

seox

idas

e1.

1.3.

4St

arch

Alp

ha-a

myl

ase

3.2.

1.1

Glu

cose

oxid

ase

1.1.

3.4

Glu

coam

ylas

e3.

2.1.

3Fr

ucto

seG

luco

seis

omer

ase

5.3.

1.18

Glu

cose

oxid

ase

1.1.

3.4

+m

utar

otas

e5.

1.3.

3G

luco

seG

luco

seox

idas

e1.

1.3.

4Pe

roxi

dase

1.11

.1.7

Glu

cose

dehy

drog

enas

e1.

1.1.

47La

ctat

em

onoo

xyge

nase

1.13

.12.

4+

lact

ate

dehy

drog

enas

e1.

1.1.

27Py

ruva

teLa

ctat

ede

hydr

ogen

ase

1.1.

1.27

Lact

ate

mon

ooxy

gena

se1.

13.1

2.4

AD

PPy

ruva

teki

nase

2.7.

1.40

Lact

ate

mon

ooxy

gena

se1.

13.1

2.4

+la

ctat

ede

hydr

ogen

ase

1.1.

1.27

Glu

cose

-6-p

hosp

hate

Glu

cose

-6-p

hosp

hate

dehy

drog

enas

e1.

1.1.

49La

ctat

em

onoo

xyge

nase

1.13

.12.

4

+la

ctat

ede

hydr

ogen

ase

1.1.

1.27

Alk

alin

eph

osph

atas

e3.

1.3.

1G

luco

seox

idas

e1.

1.3.

4A

TPH

exok

inas

e2.

7.1.

1G

luco

se-6

-pho

spha

tede

hydr

ogen

ase

1.1.

1.49

Mal

ate

Mal

ate

dehy

drog

enas

e,de

carb

oxyl

atin

g1.

1.1.

40Py

ruva

teox

idas

e1.

2.3.

3

Mal

ate

dehy

drog

enas

e1.

1.1.

37Py

ruva

teox

idas

e1.

2.3.

3+o

xala

ceta

tede

carb

oxyl

ase

4.1.

1.3

Mal

ate

dehy

drog

enas

e1.

1.1.

40Sa

licyl

ate

hydr

oxyl

ase

1.14

.13.

1

(con

tinue

dov

erle

af)

450 13 Enzyme ElectrodesTa

b.5

(con

tinue

d)

Ana

lyte

Enzy

me

sequ

ence

Indi

cato

ren

zym

e

Aux

iliar

yen

zym

esEC

num

ber

Enzy

me

ECnu

mbe

r

Oxa

lace

tate

oxal

acet

ate

deca

rbox

ylas

e4.

1.1.

3Py

ruva

teox

idas

e1.

2.3.

3C

itrat

eC

itrat

ely

ase

4.1.

3.6

Pyru

vate

oxid

ase

1.2.

3.3

+ox

alac

etat

ede

carb

oxyl

ase

4.1.

1.3

Gly

cero

lG

lyce

rold

ehyd

roge

nase

1.1.

1.6

Dia

phor

ase

1.6.

4.3

1.1.

1.72

Inos

ine

Nuc

leos

ide

phos

phor

ylas

e2.

4.2.

1X

anth

ine

oxid

ase

1.2.

3.2

Ace

tylc

holin

eA

cety

lcho

line

este

rase

3.1.

1.7

Cho

line

oxid

ase

1.1.

3.17

Cre

atin

eC

reat

inea

mid

inoh

ydro

lase

3.5.

3.3

Sarc

osin

eox

idas

e1.

5.3.

1C

reat

inin

eC

reat

inin

eam

idoh

ydro

lase

3.5.

2.10

Sarc

osin

eox

idas

e1.

5.3.

1+

crea

tine

amid

ino-

hydr

olas

e3.

5.3.

3

Cre

atin

ine

amid

ohyd

rola

se3.

5.2.

10G

luta

mat

ede

hydr

ogen

ase

1.4.

1.1/

2/3

Glu

tam

ine

Glu

tam

inas

e3.

5.1.

2G

luta

mat

eox

idas

e1.

4.3.

11Et

hano

lA

lcoh

olde

hydr

ogen

ase

1.1.

1.1

hydr

ogen

ase

1.12

.1.2

Cho

lest

erol

este

rC

hole

ster

oles

terh

ydro

lase

3.1.

1.13

Cho

lest

erol

oxid

ase

1.1.

3.6

Phen

ylph

osph

ate

Alk

alin

eph

osph

atas

e3.

1.3.

1ty

rosi

nase

1.14

.18.

1Ph

osph

ate

Nuc

leos

ide

phos

phor

ylas

e2.

4.2.

1X

anth

ine

oxid

ase

1.2.

3.2

Gly

coge

nph

osph

oryl

ase

b2.

4.1.

1G

luco

seox

idas

e1.

1.3.

4+

alka

line

phos

phat

ase

3.1.

3.1

+m

utar

otas

e5.

1.3.

3A

MP

Gly

coge

nph

osph

oryl

ase

a2.

4.1.

1G

luco

seox

idas

e1.

1.3.

4+

alka

line

phos

phat

ase

3.1.

3.1

+mut

arot

ase

5.1.

3.3


Oxalacetate decarboxylase:

Oxaloacetate ===⇒ pyruvate + CO2

Pyruvate oxidase:

Pyruvate + phosphate + O2 ===⇒acetyl phosphate + CO2 + H2O2

Citrate conversion by citrate lyase doesnot yield in a directly detectable species.Therefore the enzymatic decarboxylationof its product oxaloacetate is appended.The product of that secondary reactionis easily measurable by pyruvate oxidase.Hence, the concentration of citrate is re-lated to oxygen consumption (or hydrogenperoxide formation) in the pyruvate oxi-dase indicator reaction (see Table 3). Inaddition, carbondioxide formation may bemeasured potentiometrically.

The main hindrance to practical appli-cation of those multienzyme sensors isthat they respond to all substrates of thesequence.

However, this may be advantageous ifnot only the initial substrate is to bemeasured.

13.7.2Enzyme Competition

Another coupling principle uses the com-petitive action of two enzymes on the samesubstrate, where one of these enzymesproduces the electrochemical signal. Anexample is the ATP-sensor with hexoki-nase and glucose oxidase coimmobilized

on an oxygen electrode [2]. For in situ andin vivo application, analyte measurementsmust proceed at concentrations that exceedthe measuring range of the sensor. Alsohere competitive reactions that decreasethe amount of analyte are used.

Introduction of an additional (compet-ing) enzyme into the biocatalytic layercapable of filtering chemical signals byeliminating constituents of the sample in-terfering with either the enzymatic or theelectrochemical reaction leads to improvedselectivity. In this case a (interfering) sub-stance is completely converted to noninter-fering products in the diffusion-controlledenzyme membrane. This is in particularimportant for the development of sensorswell suited for measurement of real sam-ples.

Figure 3 illustrates the basic set-up. En-zyme E2 is immobilized in the layer toprevent the interfering compound I fromreaching the detector surface or the indi-cator enzyme E1 that is responsible foranalyte conversion. For example, ascorbicacid and acetaminophen, which interferewith the anodic hydrogen peroxide ox-idation, can be transformed into inertproducts by reaction with an eliminator en-zyme, for example, ascorbate oxidase [18],laccase [19] or tyrosinase [20] or peroxi-dase [21].

For the effective elimination of glucose,glucose oxidase, and catalase, but alsohexokinase can be used. Other antiinter-ference systems have been devised to elim-inate lactate (with lactate monooxygenase

Fig. 3 Schematic illustration ofenzymatic elimination of interferingconstituents in biosensors. A and C arecoreactants to the indicator enzyme E1and the eliminator enzyme E2. I*represents the interfering compound.

S

E1

A CE2

Electrode

A C

P∗

PI∗


in pyruvate measurements), ascorbic acid(with ascorbate oxidase for catecholaminedetection at a graphite electrode) [22], am-monia [23] or glutamate (with glutamatedehydrogenase) [24].

13.7.3Accumulation of Intermediate

The sensitivity of electroanalytical mea-surements can be enhanced by accumula-tion of the electrochemically active analyteat the electrode before measurement (strip-ping analysis) [25]. This principle adaptedto enzyme membrane electrodes relies onsequentially acting enzymes with the aimof developing sensitive and selective sen-sors.

The sensors combine preconcentrationof an intermediary product with a bio-catalytic indicator system. Oxygen probesas well as chemically modified electrodesare the base sensors. The principle of themeasurement is illustrated in Fig. 4. Inthe first step of the measurement the re-action of the analyte S1 and a saturatingconcentration of appropriate cosubstrate Aproceeds for a certain time during whichan intermediate product I is formed by thegenerator enzyme E1. The intermediate isaccumulated in the enzyme membrane,due to its slow diffusion. When this reac-tion approaches equilibrium, the secondstep, the actual measurement, is triggeredby injection of an excess of substrate (S2)of the indicator enzyme (E2), which con-verts the accumulated intermediate under

formation of an electroactive product P∗.This method yields amplification of theanalytical signal between twofold and 60-fold. For example, for the determinationof glycerol, glycerol dehydrogenase is usedfor oxidation of the analyte with forma-tion of NADH [26]. The indicator sequencelactate dehydrogenase/lactate monooxyge-nase senses NADH with resulting oxygenconsumption. Pyruvate acts as initiatorsubstance and reacts with the preconcen-trated NADH.

A limit of accumulation is set by theequilibrium of the reaction. Differencesin the amplification factors are mainlyattributed to the equilibrium constants ofthe generator enzyme and the diffusionbehavior of intermediate and initiatorsubstrate. The largest enhancement dueto accumulation is obtained when theintermediate is a large molecule and theinitiator a small molecule.

13.7.4Amplification by Analyte Recycling

Electroanalytical techniques are fairly sen-sitive and currents as low as 10−10 Acan be recorded with commercial de-vices.

The introduction of a layer incorporat-ing the enzyme over the surface of theelectrode decreases the sensitivity of theelectrode by one to two orders of mag-nitude, due to the additional diffusionresistance. Therefore, for the measure-ment of analyte concentrations in the

E1

S1

S2

A C

E2

Electrode

C

P∗

I

Fig. 4 Scheme of anintermediate accumulatingbiosensor. A and C representcoreactants; I is theaccumulated intermediate, thatis stripped after the substrate S2is added to initiate enzymereaction E2.


nanomolar range an increase of sensitivityof the enzyme electrode is required. Oneway to solve this problem is the continu-ous regeneration of the analyte in cyclicreactions.

The combination of the electrochemicaldetection principle and the recycling of theanalyte can be performed as is illustrated inFig. 5, the bienzymatic system possessingthe potential of the highest amplificationrate [14].

13.7.4.1 Enzymatic RecyclingIn the bienzymatic approach, the sensitiv-ity enhancement is provided by shuttlingthe analyte between enzymes acting incyclic series of reactions accompanied bycosubstrate consumption and accumula-tion of by-products (Fig. 5a).

The target analyte S can be substrateor coenzyme of the respective enzyme.Assuming a sufficiently high activity ofenzyme E1 in the presence of its cosub-strate C1 and an analyte at a concentrationfar below its Michaelis constant, amplifica-tion is achieved by turning-on the secondenzyme (E2) through addition of its cosub-strate, C2.

By measuring the concentration changeof one of these coreactants directly or inan additional analytical step, the recyclingsystem is used as a biochemical amplifierfor the analyte (S = S1 or S2).

13.7.4.1.1 Linear Recycling When onemolecule of product is formed per sub-strate molecule, the total concentration ofintermediate substrates (S1 + S2) remainsconstant and the concentrations of thecoreactants increase or decrease linearlywith time. Then the number of cycles inwhich the substrate is turned over in agiven time is a function of the substrateconcentration. For this case the ampli-fication factor, G, is under steady stateconditions:

G = k1k2L2

2D(k1 + k2)

where ki is the first order rate constant, L

the membrane thickness, and D the diffu-sion coefficient. At high activities of bothenzymes immobilized into the enzymelayer with high characteristic diffusiontime (L2/D), the possible amplificationis very large.

Fig. 5 Illustration of analyterecycling schemes:(a) bienzymatic analyterecycling. C representscosubstrate of the enzymaticreactions; (b) bioelectrocatalyticanalyte regeneration.

S

(a)

S1 = P2

E1

P1 = S2

E2

C1

C2

Electrode

P∗

(b)

S = M∗

E

M∗ox M∗

red

C

Electrodee−


This concept of linear enzymatic sig-nal amplification has been realized bycoupling dehydrogenases with oxidasesor transaminases, or by coupling kinaseswith each other (Table 6). When oxidasesare coupled with their respective dehydro-genases, electrode detectable species areincluded in the reaction system. Therefore,the change of coreactant concentrationcan be measured directly at the electrodeonto which the recycling enzyme pair isimmobilized. In most cases, oxygen con-sumption has been followed. Productionof protons, peroxide, and ammonia havealso been monitored.

Depending on the enzymes and mem-brane materials used, amplification byseveral orders of magnitude has been real-ized. Detection limits down to 70 pmol l−1

have been achieved [27]. When dealingwith extremely high amplification one hasto bear in mind, however, that the sen-sor signal becomes highly susceptible tominute amounts of contaminants affect-ing the enzyme reactions or the diffusionof the reactants.

A well-studied enzymatic recyclingpair is lactate oxidase/lactate dehydroge-nase.

Lactate dehydrogenase:

Pyruvate + NADH ===⇒ lactate + NAD+

Lactate oxidase:

Lactate + O2 ===⇒ pyruvate + H2O2

The oxygen consumption in the mem-brane bearing both enzymes is enhancedin the presence of NADH, yielding anincrease in the sensitivity to lactate. Thelactate signal can be enhanced by a factorof several thousands and lactate concen-trations as low as 1 nmol l−1 are measur-able [28].

Further examples of recycling enzymepairs include copper enzymes such aslaccase and tyrosinase which oxidizes awide range of substances including cat-echolamines, phenols, and redox dyesby dissolved oxygen in combination withflavoenzymes, haem-, and pyrroloquino-line quinone containing (NADH indepen-dent) dehydrogenases.

An ultrasensitive biosensor for thesephenolic substances and redox dyes hasbeen created with the quinoprotein glucosedehydrogenase and laccase, both of highspecific activity coentrapped in a gelmatrix in front of a Clark-type oxygenelectrode.

(PQQ)glucose dehydrogenase:

glucose + Mox ===⇒gluconolactone + Mred

Laccase:

Mred + O2 ===⇒ Mox + H2O

The extraordinary efficiency of the am-plification sensor is based on the excess ofenzyme molecules compared with the con-centration of the analyte molecule withinthe reaction layer. The current density ofthe membrane-covered sensor is almostthree orders of magnitude higher than isthe bare electrode with a lower limit ofdetection as low as 70 pmol l−1 [27]. As analternative to the oxygen sensor the re-action can be followed with an antimonyelectrode or an ISFET indicating the pHshift during the recycling process [29]. Theaccumulation of H+ is not as pronouncedas the diminution of the dissolved oxygenconcentration. Therefore, the response ofthe potentiometric sensors is somewhatsmaller than that of the amperometricsystem.


Tab. 6 Enzymatic analyte recycling for signal enhancement in biosensors

Analyte Enzyme couple Transducer

Glucose Glucose oxidase/glucosedehydrogenase

Oxygen electrode

Lactate Lactate oxidase/lactatedehydrogenase

Oxygen electrode

Lactate/pyruvate Cytochrome b2/lactatedehydrogenase

Pt-electrode

NADH/NAD+ Peroxidase/glucosedehydrogenase

Oxygen electrode

NADHoxidase/alcoholdehydrogenase

Oxygen electrode

Glycerol dehydrogenase/diaphorase

Oxygen electrode

p-hydroxybenzoatehydroxylase/glucose-6-phosphatedehydrogenase

Oxygen electrode

Glutamate Glutamate dehydrogenase/alanine aminotransferase

Modified carbon electrode

Glutamate oxidase/Glutamatedehydrogenase

Oxygen electrode

Glutamate oxidase/Alanineaminotransferase

Hydrogen peroxide electrode

L-Leucine Leucine dehydrogenase/amino acid oxidase

Oxygen electrode

ADP/ATP Hexokinase/pyruvate kinase Oxygen electrode with lactatedehydrogenase/lactate mono-oxygenase modified carbon-electrode with glucose-6-phosphate dehydrogenase

ADP Myokinase/pyruvate kinase Oxygen electrode with pyruvateoxidase

Ethanol Alcohol oxidase/alcoholdehydrogenase

Oxygen electrode

AdrenalineAminophenol

Glucose dehydrogenase/laccase Oxygen electrode antimony pHelectrode

Dopamine, Glucose dehydrogenase/tyrosinase

Oxygen electrode

Ferrocene deriv. Tyrosinase/diaphorase Oxygen electrode, with glucosedehydrogenase forNADH-regeneration

Oligosaccharidedehydrogenase/laccase

Oxygen electrode

Benzoquinone Fructose dehydrogenase/laccase Oxygen electrodeHydroquinone Cytochrome b2/laccase Oxygen electrodeMalate/oxalacetate Malate dehydrogenase/Lactate

monooxygenaseOxygen electrode

Phosphate Nucleoside phosphorylase/alkaline phosphatase

Oxygen electrode with xanthineoxidase

Maltose phosphorylase/acidphosphatase

Hydrogen peroxide electrode withglucose oxidase


Oligosaccharide: acceptor oxidase (oligo-saccharide dehydrogenase), cytochromeb2, diaphorase, and fructose dehydroge-nase can be employed in place of theabove mentioned glucose dehydrogenaseand laccase has been replaced by otherphenol oxidizing enzymes [30].

Recycling systems are not necessarilylimited to reactions in which electrochem-ically active compounds are produced. Inthose cases, the recycling enzyme pair iscombined with an indicator enzyme (orsequence) transforming one of the cyclecoreactants (mostly a product) into a mea-surable species (see Table 5).

13.7.4.1.2 Exponential Recycling Enor-mous signal amplification is expected ifin the cycling reaction more than one ana-lyte molecule is regenerated. Here the totalamount of intermediates and by-productsis increasing exponentially with time. Theconcentration of any of the cycling inter-mediates or byproducts at any given timeis a linear function of the initial substrateconcentration. An example illustrating thisprinciple is the ADP/ATP cycling systemmyokinase/pyruvate kinase [31]. In a sin-gle cycle two molecules of ADP are formedby myokinase per molecule ATP derivedfrom the phosphorylation of ADP by pyru-vate kinase.

Pyruvate kinase:

phosphoenolpyruvate + ADP ===⇒pyruvate + ATP

Myokinase:

ATP + AMP ===⇒ 2 ADP

The amount of ATP and ADP initiallypresent in a very low concentration in-creases exponentially with cycling time,

when AMP and phosphoenolpyruvate con-centrations are high enough to ensure nocosubstrate limitation. The abundance ofpyruvate formed in the cycle is manifestedby the oxygen consumption in the pyruvateoxidase layer.

13.7.4.2 Bioelectrochemical RecyclingIn the bioelectrocatalytic approach the tar-get analyte is recycled between electrodeand redox centre of the enzyme, thusmediating the charge transfer to the elec-trode. Therefore, the enzyme product hasto be essentially electroactive (Fig. 5b).In ideal cases sufficient (co)substrate ofthe enzyme is present, the overpoten-tial required for regeneration is low,and the analyte is stable in both redoxstates.

Vital for a reaction cycle is the closecontact of enzyme and electrode mate-rial. Starting from surface immobilizationusing adsorption, covalent binding, andentrapment of the redox enzyme, bulkmodification procedures have been estab-lished, the latter appearing to be the mosteffective way.

The immobilization of PQQ-dependentglucose dehydrogenase onto the surfaceof a glassy carbon electrode [32] or athick-film electrode [33] results in a verysensitive detector for various quinoid com-pounds. Also, tyrosinase-based sensors areused for monitoring a spectrum of mono-and diphenolic compounds that are ofparticular interest for environmental con-trol [34].

13.7.5Sequential Activation of Enzymes

An enzymatic reaction cascade can bedefined as sequential activation of a seriesof enzymes triggered, for example, byligand binding to a receptor and resulting

13.8 Application 457

in a large-scale amplification of the initialstep or signal.

Figure 6 illustrates the application ofthis principle for signal generation inbiosensors. The binding of the activator(initiator) transforms the inactive enzymeE1i to its active form E1a. This enzymeactivates a second inactive enzyme andso forth. At the end of the cascadean enzyme is activated which producesthe product P∗ indicated by the trans-ducer.

Phosphorylase b:

glycogenn + Pi ⇐⇒ glycogenn−1

+ α-glucose-1-phosphate

Muscle glycogen phosphorylase b activ-ity is controlled by the concentration ofAMP. The effect of AMP is enzyme acti-vation. Therefore, AMP response reflectsan amplification of the probe sensitivity,which is considerably higher than thatfor glucose-1-phosphate. The product ofthe reaction α-glucose-1-phosphate is in-dicated in a reaction sequence of alkalinephosphatase, mutarotase, and glucose oxi-dase [35].

13.8Application

A relevant aspect in biosensor researchis the simplification of operation, themore so as test strips are at presentstill superior in this respect. The savingsof reagents provided by reusable sensorsshould not be exceeded by the expensesnecessary for sensor maintenance. Thetask is to combine reusability and simplehandling.

Whereas in traditional enzymatic anal-ysis spectrophotometric methods domi-nate, test strips and biospecific electrodesare at the leading edge in the analyticalapplication of immobilized enzymes.

Between 15 and 20 analyzers basedon enzyme electrode are on the marketworldwide [36]. They are one-parameterinstruments for the measurement ofglucose, galactose, uric acid, choline,ethanol, lysine, lactate, pesticides, sucrose,lactose, and the activity of α-amylase.They provide for a negligible enzymeconsumption of less than 1 µg per sample.

The selective determination of bloodglucose is of the utmost importance forthe screening and treatment of diabetes.The normal concentration of glucose in

Fig. 6 Schematic illustration ofsequential activation ofenzymes in biosensors. A andrepresent coreactants.

S

E2i E2a

E1i E1a

E3i E3a

A C

Electrode

A C

P∗

A C


blood serum ranges between 4.2 and5.5 mmol l−1, pathological situations maycause an increase up to more than30 mmol l−1.

Clinical chemists are interested in auto-analyzers characterized by high measuringfrequency as well as in portable bedside-type analytical devices with short lag timebetween sample withdrawal and avail-ability of the result. Therefore, enzymeelectrode-based analytical systems for theapplication of highly diluted as well asundiluted media have been developed andcommercialized.

Glucose analyzers based on enzymeelectrodes have been brought onto themarket in the United States, Japan, France,Lithuania, and Germany. As comparedwith the convential enzymatic analysis,the main advantages of such analyzers arethe extremely low enzyme demand (a fewmilliunits per sample), the simplicity ofoperation, and the high analytical quality.

The first commercial electrochemicalglucose sensor for patient self-monitoringwas introduced by Medisense, UnitedKingdom. There are now a number ofsystems that measure very rapidly andrequire less than 1-µl sample (Therasense,USA).

The sensor is based on a ferrocene-modified glucose oxidase electrode strip.For glucose determination, a drop of bloodis transferred to a disposable enzymeelectrode strip, which is then insertedinto a pen-sized readout instrument. Theresponse is more rapid than that ofphotometric test strips. Venous as wellas capillary blood may be used as samplematerial. There are a number of systemsthat measure very rapidly and require lessthan 1-µl sample.

The invasive application of biospecificelectrodes for direct analysis of substances

like glucose is one of the most impor-tant challenges of biosensor designers.Encouraging results have been obtainedfor subcutaneous tissue as the samplingsite because of the lower risk of infectionand blood clotting as compared to that afterintravascular implantation.

Nevertheless, no truly reliable im-plantable sensor has yet reached the mar-ket. In the meantime, bedside-type discreteanalyzers seem to offer the most economicand reliable means for critical care. Inaddition, ex vivo on-line analysis basedon microdialysis sampling seems to behelpful.

References

1. H. J. Bergmeyer, M. Graßl, Methods of En-zymatic Analysis, VCH Verlagsgesellschaft,Weinheim, 1986.

2. F. W. Scheller, F. Schubert, Biosensor, Tech-niques and Instrumentation in Analyti-cal Chemistry, Elsevier, Amsterdam, 1992,Vol. 11.

3. E. A. H. Hall, Biosensors, Prentice Hall Ad-vance Reference Series, Engineering, OpenUniversity Press, Milton Keynes, England,1991.

4. M. Dixon, E. C. Webb, (Eds.), Enzymes, Long-mans, London, 1967.

5. Enzyme nomenclature, Recommendations(1972) of the International Union of Pure andApplied Chemistry and the International Unionof Biochemistry, Elsevier, Amsterdam, 1973.

6. L. C. Clark, C. Lyons, Ann. N. Y. Acad. Sci.1962, 102, 29.

7. K. Mosbach, Immobilized enzymes andcells, in Methods in Enzymology (Eds.: S. P.Colowick, N. O. Kaplan), Academic Press,San Diego, 1988, Vol. 137.

8. P. W. Carr, L. D. Bowers in Immobilized En-zymes in Analytical and Clinical Chemistry,Chemical Analysis, Fundamentals and Appli-cations (Eds.: P. J. Elving, J. D. Winefordner,I. M. Kolthoff), John Wiley & Sons, NewYork, 1982, Vol. 56.

9. A. E. G. Cass, Biosensors, in A PracticalApproach, The Practical Approach Series

13.8 Application 459

(Eds.: D. Rickwood, B. D. Hames), OxfordUniversity Press, Oxford, 1990.

10. A. Heller, Acc. Chem. Res. 1990, 23, 128–134.11. I. Willner, E. Katz, Angew. Chem., Int. Ed.

2000, 39, 1181–1218.12. B. W. van der Schoot, P. Bergveld, Biosensors

1998, 3, 161.13. A. P. F. Turner, I. Karube, G. S. Wilson,

Biosensors, Fundamentals and Applications,Oxford University Press, Oxford, 1987.

14. F. W. Scheller, F. Schubert, J. Fedrowitz,Frontiers in Biosensorics, I Fundamentals;II Practical Applications, Birkhauser Verlag,Basel, 1997.

15. F. W. Scheller, D. Pfeiffer, F. Lisdat et al.,Enzyme biosensors based on oxygen detec-tion, in Methods in Biotechnology, Enzymeand Microbial Biosensors: Techniques and Pro-tocols (Eds.: A. Mulchadani, K. R. Rogers),Humana press, Totowa, 1998, Vol. 6.

16. P. Bartlett, P. Tebutt, R. G. Whitaker, Prog.React. Kinet. 1991, 16, 55.

17. N. Gajovic, A. Warsinke, F. W. Scheller, J.Chem. Tech. Biotechnol. 1995, 63, 337–344.

18. J.-I. Anzai, H. Takeshita, Y. Kobayashi et al.,Anal. Chem. 1998, 70, 811–817.

19. U. Wollenberger, F. W. Scheller, D. Pfeiffer,Anal. Chim. Acta 1986, 187, 39–45.

20. J. Wang, N. Naser, U. Wollenberger, Anal.Chim. Acta 1993, 281, 19–24.

21. R. Maidan, A. Heller, Anal. Chem. 1992, 64,23, 2889–2896.

22. G. Nagy, M. E. Rice, R. N. Adams, Life Sci.1982, 31, 2611–2616.

23. K. Kihara, E. Yasukawa, Anal. Chim. Acta1986, 183, 75–80.

24. M. B. Madaras, R. B. Spokane, J. M. Johnsonet al., Anal. Chem. 1997, 69, 18, 3674–3678.

25. P. T. Kissinger, W. R. Heineman, LaboratoryTechniques in Electroanalytical Chemistry,Marcel Dekker, New York, 1996.

26. F. Schubert, J. Lutter, F. W. Scheller, Anal.Chim. Acta 1991 243, 17–21.

27. A. L. Ghindilis, A. Makower, B. G. Baueret al., Anal. Chim. Acta 1995, 304, 25–31.

28. U. Wollenberger, F. Schubert, F. W. Schelleret al., Studia Biophysica 1987, 119, 167–170.

29. A. L. Ghindilis, A. Makower, F. W. Scheller,Anal. Chim. Acta 1995, 304, 25–31.

30. U. Wollenberger, F. Lisdat, F. W. Scheller inFrontiers in Biosensorics, I Fundamentals; IIPractical Applications (Eds.: F. W. Scheller,F. Schubert, J. Fedrowitz), Birkhauser Ver-lag, Basel, 1997.

31. D. Pfeiffer, F. W. Scheller, C. McNeil et al.,Biosens. Bioelectron. 1995, 10, 169–180.

32. A. F. Eremenko, A. Makower, W. Jin et al.,Biosens. Bioelectron. 1995, 10, 717–722.

33. A. Rose, F. W. Scheller, U. Wollenbergeret al., Fresenius J. Anal. Chem. 2001, 369,145–152.

34. F. Ortega, E. Dominuez, G. Jonsson-Petters-son et al., J. Biotech. 1993, 31, 289–300.

35. U. Wollenberger, F. W. Scheller, Biosens.Bioelectron. 1993, 8, 291–297.

36. D. Pfeiffer in Frontiers in Biosensorics, IFundamentals; II Practical Applications (Eds.:F. W. Scheller, F. Schubert, J. Fedrowitz),Birkhauser Verlag, Basel, 1997.

461

14Carbon Fiber Microelectrodes forthe in vivo Measurement ofNeurotransmitters: A Close UpLook at Neurochemical Activityin the Brain

Jennifer L. Peters, Nadezhda V. Kulagina, Hua Yang, and Adrian C. MichaelUniversity of Pittsburgh, Pittsburgh, Pennsylvania

14.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463

14.2 A Mathematical Description of Extracellular Neurochemistry . . . . . 465

14.3 Histology of Microelectrode and Microdialysis Probe Tracks in vivo . 468

14.4 Monitoring Electrically Evoked Dopamine Release with Microelectrodes 469

14.5 In vivo Electrochemistry with Very High Spatial Resolution . . . . . . 471

14.6 Monitoring Spontaneous, as Opposed to Evoked, Dopamine Release 474

14.7 Electrochemical Sensors for in vivo Measurement . . . . . . . . . . . . . 478


463

14.1Introduction

Motivated by the pioneering work of RalphAdams [1] and his group at the Universityof Kansas, electroanalytical chemists inseveral countries have developed a pow-erful set of microelectrode-based meth-ods that are now making significantcontributions to our understanding of theneurochemical activity of the living mam-malian brain. Numerous electrochemi-cally detectable substances play importantfunctional roles in the central nervoussystem, and there is intense interest inobtaining more information about theconcentration, lifetime, and regulation ofthese substances in the extracellular spaceof the living brain. Although early workin this area focused on the neurotransmit-ter, dopamine, the list of substances thathave now been monitored electrochemi-cally in vivo is quite long and still growing.Strategies have been demonstrated for theselective in vivo electrochemical detectionof dopamine [2–4], norepinephrine [5–7],serotonin [8–10], glutamate [11–13], cho-line [14–16], nitric oxide [17–19], ascorbicacid (vitamin C) [11], glucose [14, 16], lac-tate [20, 21], hydrogen peroxide [22], andoxygen [23, 24]. Ultimately, the goal ofin vivo electrochemical monitoring isto aid in the elucidation of the role

of neurochemistry, not only in normalbrain function but also in the numer-ous pathological conditions that afflictthe human brain, such as Parkinson’sdisease, Alzheimer’s disease, schizophre-nia, ischemia, traumatic brain injury, anddrug abuse. The electroanalytical tech-niques available for in vivo monitoringprovide access to unique neurochemicalinformation that is presently inaccessibleby other in vivo analytical methods, such asmicrodialysis and spectroscopic imaging[e.g. positron-emission tomography (PET),single-photon-emission computed tomog-raphy (SPECT), and magnetic resonanceimaging (MRI)]. Taken in combination,electrochemical methods, microdialysis,and spectroscopic imaging comprise apowerful array of tools for unlocking thechemical secrets of the brain.

The electrochemical methods that haveevolved for neurochemical applicationshave several advantages that make themideally suited to the task for which theyare intended: the methods are selective,sensitive, and rapid. Nevertheless, the keyadvantages that will be highlighted in thischapter are derived from the micrometerphysical dimensions of the microelec-trodes themselves (Fig. 1). Today, themajority of in vivo electrochemistry isconducted in the brain with microelec-trodes constructed with individual, or a

464 14 Carbon Fiber Microelectrodes for the in vivo Measurement of Neurotransmitters

GlassCarbon fiberEpoxy

10 µm

20 µm

Glass

Carbonfiber

Fig. 1 Electron micrograph of carbonfiber microdisk and microcylinderelectrodes.

few, carbon fibers. Carbon fibers withdiameters less than 10 µm are readily avail-able; they are quite easy to use for electrodefabrication, and despite their small size aresurprisingly robust. The tiny dimensionsof these electrodes have two immediateand immensely important consequencesfor in vivo analysis: first, measurementswith micrometer spatial resolution are pos-sible [25]; second, the electrodes producevery little physical disruption of the deli-cate architecture of the brain tissue. Thelatter point can be qualitatively appreci-ated just by considering the dimensionsof structures found in the brain. For ex-ample, the body of most neuronal cellsin the mammalian brain is typically onthe order of 10 to 20 µm in diameter [26],while blood capillaries in the brain areseparated from each other by distancesof 20 to 100 µm [27]. Evidence from lightand electron microscopy, discussed laterin the chapter, confirms that implantationof these tiny electrodes inflicts essentiallyimperceptible brain injury.

The minimal amount of brain injurythat microelectrodes cause renders themuniquely well suited to monitoring ongo-ing neurochemical events in the brain.Carbon fiber electrodes can be implantedto within a few micrometers of viable nerveterminals, which comprise the loci of theneurochemical events we desire to mon-itor. The intimate proximity of the elec-trodes to nerve terminals is of critical im-portance to the quality of the neurochem-ical information that is obtained. Mostneurotransmitters are rapidly cleared fromthe extracellular space by either transportermechanisms or metabolic processes [28].These clearance events limit the lifetimeof neurotransmitters in the extracellularspace to the millisecond regime, whichprovides the neurotransmitter moleculeswith very limited opportunity to diffusethe necessary distance to reach an im-planted sensor. Hence, if the sensor is notextremely close to functional nerve termi-nals, the sensor may not be capable of pro-viding accurate neurochemical informa-tion about events that occur in the imme-diate vicinity of neuronal terminals. Thefull significance of this critical point hasonly recently begun to be fully appreciated.

Although the limited extracellular life-time of neurotransmitters represents oneof the main challenges for in vivo neu-rochemical analysis, it should also bepointed out that information about ex-tracellular neurochemistry is extremelyvaluable to our understanding of brainfunction, not to mention dysfunction as-sociated with disease states of the brain.The role of neurotransmitters is to conveymessages between neighboring neurons,and that role is fulfilled by diffusion of

14.2 A Mathematical Description of Extracellular Neurochemistry 465

neurotransmitter molecules through theextracellular space [29]. Neurotransmittersare stored in neuronal terminals withinsmall organelles called synaptic vesicles.The arrival of an action potential, the elec-trical signal originating at the cell body,triggers the fusion of synaptic vesicles withthe terminal membrane, allowing a smallpore to form through which the contents ofthe vesicle escape to the extracellular space.While in the extracellular space, the neu-rotransmitter may encounter and bind toa receptor on the surface of a neighboringcell, thereby conveying neuronal informa-tion between the releasing and receivingcell (which may be one and the same, by theway). Hence, by monitoring the extracellu-lar concentration of neurotransmitters, asopposed to their intracellular or total tissueconcentration, the measurement specifi-cally focuses on the small fraction of theneurotransmitter that is at any one timeengaged in neurotransmission. The res-olution of information about the activefraction of neurotransmitter is a powerfulaspect of methods that specifically monitorextracellular events.

Although electrochemistry is not theonly method that has been developed forin vivo neurochemical monitoring, the in-formation it provides is highly unique. Forexample, microdialysis sampling is widelyused as another approach [30]. But, thesampling probes available for microdial-ysis are larger than the microelectrodesavailable for in vivo electrochemistry.Whereas microelectrodes may have diame-ters of 10 µm or less, the smallest availablemicrodialysis probes have diameters ofabout 200 µm, and occupy approximatelya 100 00-fold greater volume. Because oftheir size, the implantation of microdial-ysis probes into the brain inflicts severeinjury on the surrounding tissue. A recentstudy that used electron microscopy to

examine the tissue near a microdialysisprobe implantation site found severe lossof neuronal structure in the immediatevicinity of the probe and less severe dam-age extending as far as 1.4 mm from theprobe site [31]. Recently, it has becomeapparent that many of the differences be-tween results obtained with microdialysisand in vivo electrochemistry can be at-tributed to the differences between themagnitudes of damage that each method-ology inflicts on the brain tissue underinvestigation.

Spectroscopic imaging technologies ha-ve also been developed as tools formonitoring ongoing chemical activity ofthe brain [32, 33]. Because imaging tech-niques, such as PET and SPECT, for exam-ple, involve no implantation of sensing orsampling devices into the brain itself, theycompletely eliminate brain injury. A pow-erful aspect of these methods, therefore, istheir suitability for use in human patientsor volunteers. In contrast, although theuse of microdialysis in humans is emerg-ing as a tool for critical patient care, theapplication of in vivo electrochemistry hasbeen limited to date to laboratory animals.But, it should be mentioned that spectro-scopic imaging can only be applied to thestudy of radiolabeled imaging agents anddo not provide direct information aboutthe unlabeled endogenous substances thatcan be monitored with implantable micro-electrodes. With this in mind, it is fair tosay that spectroscopy and electrochemistryprovide complementary information.

14.2A Mathematical Description of ExtracellularNeurochemistry

In the introduction, we qualitatively ex-plained that the proximity of an implanted


sensor to viable neuronal terminals in thebrain has a major impact on the abil-ity of the sensor to provide informationabout neurochemical events occurring inthe vicinity of the terminals themselves.Because this issue has enormous impactboth on the design of in vivo experimentsand on the interpretation of their outcome,it is valuable to examine this issue in asomewhat more quantitative fashion be-fore proceeding to describe the actual invivo experiments. Thus, in this section, wepresent a mathematical discussion of thediffusion processes that deliver substancesto implanted electrodes.

The release of neurotransmitters fromneuronal terminals in the brain can beviewed as a quantal event. When an actionpotential arrives at a nerve terminal, a fi-nite number of synaptic vesicles fuse withthe terminal membrane and release theircontents into the extracellular space. Thisrelease event occurs very rapidly: studiesof exocytotic events in single isolated cellsshow that they occur on a sub-millisecondtimescale [34, 35]. Furthermore, the eventis spatially discrete, since the nerveterminals themselves have dimensionsof just a few hundred nanometers [36,37]. With this description in mind, it isreasonable to regard each nerve terminal as

a small, spatially discrete diffusion sourceand to regard each release event as aninstantaneous activation of that diffusionsource, as depicted in Fig. 2. The concen-tration profile plotted in the bottom portionof Fig. 2 represents a finite initial sourceof a diffusible substance: the question wewish to address here is how the concentra-tion of that substance evolves in both spaceand time after the initial release event.

While neurotransmitters diffuse in theextracellular space, they are subject toclearance processes that either transportthem back to nerve terminals or me-tabolize them to inactive products inthe extracellular space. For example, themembranes of many nerve terminals areequipped with transporters, which aretransmembrane proteins that transportneurotransmitter from the extracellularspace into the cytoplasm of the terminal.On the hand, as in the case of acetylcholine,the extracellular space may contain en-zymes that rapidly metabolize the neuro-transmitter. The impact of these clearanceprocesses on the concentration of the dif-fusing substance can be considered withan equation of the following type:

∂C(x, t)

∂t= D

∂2C(x, t)

∂x2 − kC(x, t) (1)

0−h +h

C0

Fig. 2 Schematicrepresentation of a finitediffusion source resulting fromneurotransmitter release from aneuronal terminal.

14.2 A Mathematical Description of Extracellular Neurochemistry 467

where C(x, t) is the concentration, D isits diffusion coefficient in the extracellularspace, k is the rate constant of the clearanceprocess, x is the distance from the centerof the initial diffusion source, and t isthe time that has elapsed since the initialrelease event [38]. The first term on theright hand side of Eq. (1) accounts for theimpact of diffusion on the concentration ofthe diffusing substance, while the secondterm accounts for the impact of theclearance process.

Hence, by solving Eq. (1) we can gaininsight into how the combined influenceof the diffusion and clearance processesdetermine the concentration of the dif-fusing substance that evolves from theinitial condition proposed in Fig. 2. Withthe initial condition of Fig. 2, the solutionto Eq. (1) is:

C(x, t) = C0

2

[erf

(h − x

2√

Dt

)

+ erf(

h + x

2√

Dt

)]exp(−kt) (2)

where ‘‘erf’’ is the error function [39, 40].Figure 3 shows plots of Eq. (2) calculatedat various times after the initial releaseevent. Figure 3(a) shows plots calculatedwith a clearance rate constant of zero(no clearance) while Fig. 3(b) shows plotscalculated with a rate constant of 10 s−1,which is consistent with the value reportedfor the pseudo first-order rate constant forthe dopamine transporter in the striatalregion of the rat brain [41]. The otherparameters used to obtain Eq. (3) werealso consistent with values for the striataldopamine system of the rat: the valueof h was 100 nm [36, 37] and D was2.4 × 10−6 cm2 s−1 [42].

Figure 3(a) shows how the substance be-comes increasingly diluted with time asis it diffuses into an ever-increasing vol-ume of space. In this case, because the

original diffusion source is so small, thedilution effect is quite extreme, even af-ter a relatively short amount of time: aftera diffusion time of 50 ms the maximumconcentration is less than 2% of the origi-nal value, C0. Just as importantly, Fig. 3(a)shows that over the time interval consid-ered here, the concentration profile of thediffusing substance does not extend morethan a few micrometers in either directionfrom the diffusion source. This illustratesthe point that as the distance between thesource and an implanted sensor increases,the measurement is increasingly affectedby the dilution phenomenon.

Figure 3(b) shows how the introductionof the clearance process changes the con-centration behavior predicted by Eq.(2).The clearance process imposes a finitelifetime on the diffusing substance. Thus,not only is the substance diluted by thediffusion process, it is also lost to the clear-ance process. After a time correspondingto approximately three times the reac-tion half-life, the substance has almostcompletely disappeared. The added con-tribution of the clearance process limits toan even greater extent the maximum traveldistance of the diffusing substance fromthe source. Figures 3(a) and (b) illustratehow the combined effects of diffusionallyinduced dilution and the clearance processaffect the manner in which concentrationvaries as function of distance from the dif-fusion source. If the distance between thediffusion source and a nearby sensor istoo great, the sensor will not be able toreport concentration events occurring inthe vicinity of the source itself.

The most obvious message to be takenfrom Fig. 3 is that the proximity of animplanted device to viable nerve terminalshas a considerable impact on what canbe observed with that device. A devicethat disrupts nearby terminals when it


C/C

0

0.00

0.01

0.02

(a)

C/C

0

0.00

0.01

0.02

(b)

−50 −25 0 25 50

NoKinetics

50 ms

150 ms

300 ms

Distance[µm]

Distance[µm]

−50 −25 0 25 50

k = 10 s−1

50 ms

150 ms

300 ms

Fig. 3 Theoretical concentrationprofiles obtained by plotting Eq. (2) atvarious values of t, as indicated on thefigure. Panel A shows the result for thepure diffusion case. Panel B includes thekinetics of a coupled chemical reaction.Values of h, k, and D are discussed inthe text.

is introduced into brain tissue, therebyrendering the nearby terminals nonviable,will observe a smaller fraction of theextracellular concentration of substancesreleased by terminals positioned furtheraway from the device. On the other hand,if the device can be introduced withoutcausing disruption, then that device willprovide a closer view of events occurringin the vicinity of the viable terminals.

14.3Histology of Microelectrode andMicrodialysis Probe Tracks in vivo

The previous section pointed out thecritical importance of being able to placea sensor as close as possible to viable

neuronal terminals in order to accuratelymonitor the release of neurotransmittersfrom those terminals. This is one ofthe factors that motivated the use ofextremely small electrodes fashioned fromcarbon fibers for in vivo neurochemicalmeasurements. These electrodes, withdiameters of less than 10 µm in some cases(Fig. 1), are literally smaller than most ofthe cells found in the mammalian braintissue [26]. Recently, we have attemptedto use histological methods to directlyexamine the level of tissue disruptioncaused by implantation of carbon fiberelectrodes into brain tissue [43]. As it turnsout, this has been difficult to do becausethe amount of damage caused by theseelectrodes is too small to detect.

Carbon fiber microcylinder electrodes(7 µm in diameter and 400 µm in length)were implanted vertically into the brains ofanesthetized rats and were used to mon-itor electrically evoked dopamine release,as described in detail in the next section.After the electrodes were removed fromthe animal, the brain tissue was collectedand thinly sliced in the plane perpendic-ular to the axis of the electrodes. Whenwe examined the tissue sections throughwhich the carbon fiber passed, we couldfind no evidence of the electrode track withthe light microscope. This shows that thetissue damage caused by the fiber itself isminimal. On the other hand, tracks couldbe found in tissue sections through whichthe glass barrel of the electrode had passedbecause the barrels were tapered and their

14.4 Monitoring Electrically Evoked Dopamine Release with Microelectrodes 469

diameters were larger than the diameterof the fibers. Tissue sections taken just be-fore the electrode track disappeared werefurther examined by electron microscopy.The electron micrographs clearly showedthe presence of intact axonal terminalswithin micrometer distances of the elec-trode barrel. This study confirms that themicroelectrodes, by virtue of their abilityto penetrate brain tissue without inflictinginjury, gain very close proximity to in-tact axonal terminals once implanted intoliving brain tissue.

The microscopy of microelectrodestracks stands in stark contrast to imagesof tissue in the vicinity of the implantationsites of microdialysis probes, as recentlyreported by Drew and coworkers [31]. Thetissue in the immediate vicinity of themicrodialysis probe suffered a severe lossof synaptic junctions. This confirms thegeneral idea that implantation of a largerdevice leads to further tissue disruption,thereby directly increasing the separationdistance between that larger device and vi-able neuronal terminals. As discussed infurther detail in the following section, dif-ferences between in vivo results obtainedwith microelectrodes and microdialysisprobes can be attributed to differences inthis separation distance.

14.4Monitoring Electrically Evoked DopamineRelease with Microelectrodes

While electron microscopy reveals thepresence of intact synaptic terminals inthe immediate vicinity of microelectrodes,it does not demonstrate that those termi-nals are viable. Evidence that the nearbyterminals are viable is available in theform of a variety of experimental results.For example, carbon fiber microelectrodes

have been used in numerous laboratoriesto monitor electrically evoked dopaminerelease in the rat striatum [2, 44, 45], inaddition to several other brain regionsof interest [46, 47]. Electrically evoked re-lease of dopamine refers to the release ofdopamine upon electrical stimulation ofdopaminergic axons that pass through abrain region called the medial forebrainbundle (MFB). For this type of experi-ment, a bipolar stimulating electrode ispositioned just above the MFB and smallcurrent pulses are passed between thetwo poles of the electrode. These cur-rent pulses stimulate action potentials thattravel along the axons and evoke dopaminerelease upon reaching the axon terminalsin the striatum. The stimulus responses inFig. 4(a), which were recorded in the stria-tum with a carbon fiber microcylinder elec-trode, show that the electrochemical signalessentially changes as soon as the stimu-lation begins. Usually, the signal changeswithin the first 100 ms of the stimulation.This confirms that the evoked release takesplace very near to the microelectrode, sincedopamine can only diffuse a few microm-eters in such a short interval of time.

Several lines of evidence have beengathered to confirm that the substancereleased into the extracellular space dur-ing MFB stimulation is in fact dopamine.These have been discussed many timesin the literature and so will be men-tioned only briefly here. The first lineof evidence is electrochemical in na-ture. The stimulus responses in Fig. 4(a)were obtained by repetitively performingcyclic voltammetry at a sweep rate of300 V s−1 and with an interval of 100 ms.A background subtracted voltammogramof the substance that changed concentra-tion during the stimulation is obtainedby taking the difference of the current-potential response recorded during the


Microelectrode adjacentto microdialysis probe

0 20 40 60 80

Time[s]

0 20 40 60 80

Carbon fibermicroelectrode

20 nA 20 nA

0 20 40 60 80

Carbon fiberbundle electrode

10 nA

(a) (b) (c)

Fig. 4 Stimulated dopamine release measured in striatum by several analyticaldevices. Stimulation responses were recorded 5 min before (solid lines) and25 min after (dotted line) the systemic administration of 20 mg kg−1 nomifensine.Electrical stimulation was applied to the MFB bundle with the followingparameters: 45 Hz, 10 s (a & b) or 25 s (c) and 50 µA rms. Open circles denote thebeginning and end of the electrical stimulation.

stimulation and just before stimulationbegan. The background-subtracted voltam-mograms obtained in vivo compare verywell with those obtained during calibra-tion of the microelectrode in authenticdopamine after removal of the electrodefrom the animal. Anatomical evidence isalso available: both the stimulation and thevoltammetric microelectrode must be cor-rectly positioned or no stimulus responsewill be obtained [46]. Finally, pharma-cological evidence exists: the effects ofmany drugs that interact selectively withdopamine neurons have been shown to ex-ert predictable effects on the amplitude andduration of the stimulus responses [45,48, 49].

Several times in the discussion pre-sented so far, we have emphasized theimportance of the proximity of the sen-sor to viable axon terminals in the brain.Microdialysis probes, which are consider-ably larger than carbon fiber microelec-trodes, have been confirmed by electron

microscopy to cause severe damage tobrain tissue [31]. In a series of experimentsthat made combined use of voltammetricmicroelectrodes and microdialysis probes,we investigated the magnitude of the effectof this tissue damage on the neurochem-ical results obtained by microdialysis [44,50–52]. In this work, a carbon fiber micro-electrode was mounted directly onto theouter surface of a 220-µm diameter micro-dialysis probe. The combined device waslowered into the striatum of anesthetizedrats, where evoked dopamine release hasbeen measured before, as described inthe preceding paragraphs. However, whenelectrical stimulation was applied to theMFB, no evoked dopamine release couldbe observed at the carbon fiber micro-electrode that was mounted onto themicrodialysis probe. This result demon-strates that the tissue in the immediatevicinity of the microdialysis probe is devoidof viable dopamine terminals, consistentwith electron microscopy findings of Drew

14.5 In vivo Electrochemistry with Very High Spatial Resolution 471

and coworkers [31]. To date, the only waywe have been able to observe evokeddopamine release with a microelectrodemounted onto a microdialysis probe hasbeen to make use of a drug, nomifensine,which inhibits the dopamine transporter(Fig. 4b). Consistent with Fig. 3, a drugthat decreases the effective rate constantof the transport of dopamine from the ex-tracellular space is expected to increasethe distance over which dopamine can dif-fuse from viable terminals. Blockade ofthe transporter, however, does not elim-inate the diffusionally induced dilutionphenomenon, so even after uptake inhi-bition, we observe smaller signals at themicroelectrodes adjacent to the microdial-ysis probes that at microelectrodes placedthemselves into brain tissue.

Consistent with the difficulty we expe-rienced in observing evoked dopaminerelease in the vicinity of microdialysisprobes, we have also recently found thatit is similarly difficult to observe evokeddopamine release when a larger-sizedelectrode is used [53]. Most in vivo elec-trochemistry is conducted with microelec-trodes constructed with a single carbonfiber. But, it is also feasible to makelarger electrodes by bundling many fiberstogether. In this work, we constructedelectrodes with a bundle of fibers with a di-ameter of about 100 µm, that is, a diameterapproaching that of a microdialysis probe.Although these larger electrodes work justas well as single-fiber electrodes for de-tecting dopamine in standard solutions,we could not observe evoked dopaminerelease in the rat striatum without theaid of the dopamine uptake inhibitor,nomifensine (Fig. 4c). Again, this resultconfirms the importance of the overallsize of the device implanted into the brainfor neurochemical monitoring. If the de-vice is too large, regardless of whether

it is an electrode or a sampling probe,implantation will lead to tissue damagethat increases the distance between the de-vice and viable axon terminals, which inturn significantly impacts the ability of thedevice to provide information on chemi-cal events taking place in the immediatevicinity of viable terminals.

14.5In vivo Electrochemistry with Very HighSpatial Resolution

Microelectrodes constructed with singlecarbon fibers come in two basic versions,the microcylinder and the microdisk, asdepicted in Fig. 1. All the measurementsdiscussed earlier were performed with mi-crocylinder electrodes, which are smallenough in diameter to be implanted withminimal tissue damage. Although theseelectrodes gain close proximity to viableneuronal terminals, they provide informa-tion about extracellular neurotransmitterconcentrations as averaged over the lengthof the cylinder. On the other hand, the mi-crodisk electrodes provide a much higherspatial resolution and provide much morehighly localized information. For exam-ple, when two microcylinder electrodesare implanted fairly close to each otherin the rat striatum, electrical stimulationof the MFB produces very similar re-sponses at each electrode [44]. If insteadtwo microdisk electrodes are implanted,MFB stimulation often produces very dif-ferent responses [44]. Alternatively, severalgroups have demonstrated that differentstimulus responses are obtained when anindividual microdisk electrode is used torecord stimulus responses at several loca-tions within the striatal region of the samerat [44, 54]. The contrast between the re-sults obtained at the two styles of electrode


show that the microdisks report local vari-ations in the extracellular response duringstimulation, while the microcylinders re-port the spatial average of that response.

Figure 3, above, explains the originof the local variations in extracellulardopamine concentrations exhibited in theresults obtained with microdisk electrodes.When placed randomly into the tissue, amicrodisk electrode sometimes finds itselfvery near to a site of active dopaminerelease (a ‘‘hot spot’’), while on otheroccasions will find itself in a locationwhere dopamine release does not occur(a ‘‘cold spot’’). The concept of hot andcold spots is consistent with the knownarchitecture of the striatal region of thebrain, which contains numerous bundlesof myelinated nerve fibers that do notcontain dopamine terminals: if the smallelectrode were to be implanted in such abundle, then indeed very little dopaminerelease would be observed at that location.

Numerous studies have examined theways in which drugs that selectively act ondopamine neurons alter evoked dopamineresponses as measured with microelec-trodes [45, 48, 49]. Several drugs are wellknown to increase the amplitude of thestimulus response, while others decreasethe amplitude. Some drugs also affect thetime course of the response. At microcylin-der electrodes, the effects of drugs on thestimulus responses at different electrodesare generally similar, implying again thatthese electrodes spatially average out localvariations in the response. At microdiskelectrodes, however, this may not be thecase. For example, in a recent study, wereported that two drugs gave very differ-ent responses, depending on whether themicrodisk electrode was in a hot spot ora cold spot for dopamine release [25]. Inexperiments involving the dopamine up-take inhibitor, nomifensine, we found that

the drug had a rather small effect in hotspots, but had a pronounced effect in coldspots. In contrast, when we gave rats a doseof L-DOPA, the biosynthetic precursor ofdopamine, we found that the drug had themost pronounced effect in hot spots and amild effect in cold spots.

We have been interested in understand-ing the local variations in drug-inducedmodifications of the stimulus responsesbecause they may be relevant to un-derstanding the mechanisms underlyingthe functional effects of these drugs.Nomifensine, for example, is a psychoac-tive drug [55, 56], while L-DOPA is usedin the therapy of Parkinson’s disease [57,58]. Our strategy was to compare the ex-perimentally observed effects of the drugswith predictions obtained by modelingdiffusion and clearance processes, in sim-ilar fashion to that used to obtain Fig. 3.For this work, however, we modeled theclearance process with Michaelis-Mentenkinetics dopamine uptake by the trans-porter is a saturable process and sincenomifensine is a competitive uptake in-hibitor:

∂C(x, t)

∂t= D

∂2C(x, t)

∂x2 − Vmax · C(x, t)

Km + C(x, t)

(3)

where Vmax and Km are the maximalvelocity and the Michaelis constant, re-spectively. Eq. (3) does not have a simpleanalytical solution, so a numerical solu-tion method was used. Stimulus responseswere simulated by obtaining solutions toEq. (3) while linearly ramping the concen-tration at the diffusion source for a finiteamount of time. The value of Km wasincreased to simulate the effect of compet-itive uptake inhibition after nomifensineadministration, while the rate at which thesource concentration was ramped was in-creased to simulate the effect of increasing

14.5 In vivo Electrochemistry with Very High Spatial Resolution 473

Distance from diffusion source (dimensionless)0.0 0.5 1.0

Con

c. (

dim

ensi

onle

ss)

0

1

Distance from diffusion source (dimensionless)0.0 0.5

Con

c. (

dim

ensi

onle

ss)

0

1

EnhancingRelease:

InhibitingUptake:

Time [s]

0 30

i [n

A]

0

1

Time [s]

0 30

i [n

A]

0

1

Time [s]

0 30

i [n

A]

0

1

Time[s]

0 30

i [n

A]

0

3

Time[s]

0 30

i [n

A]

0

3

Time[s]

0 30

i [n

A]

0

3

(a)

(b)

Fig. 5 Theoretical and experimental descriptions of the impact ofuptake inhibition (a) and enhancement of release (b) of the responsesrecorded at microdisk electrodes. Theoretical curves: Numericalsolutions of Eq. (3) were used to generate predicted concentrationprofiles at various times during a simulated period of stimulation. Thecalculated concentration profiles shown in the main panel of the top andbottom portion of this figure were obtained at the end of the simulatedstimulus. The top panel shows how an increase in the Michaelisconstant (Km) changes the concentration profile, while the bottompanel shows the effect of an increase in the magnitude the simulatedstimulus (further details can be found in Ref. [25]). Stimulationresponses: The inset panels show experimental stimulus responsesrecorded in the rat brain with microdisk electrodes. Open circles denotethe beginning and end of the electrical stimulation. Predrug responses(solid lines) were recorded prior to systemic administration of either20 mg kg−1 nomifensine (a) or 250 mg kg−1 L-DOPA (b). Postdrugresponses (dotted lines) were recorded 25 min after nomifensineadministration or 55 min after L-DOPA administration. Note that thetrends in the amplitude of the experimental signals correspond very wellto those apparent in the theoretical concentration profiles.


the biosynthesis of dopamine after L-DOPA administration. The time courseof the simulated change in concentrationat locations near the diffusion source wastaken to represent the calculated responsesin hot spots, while the time course atlarger distances from the source was takento represent the calculated responses incold spots. Figure 5 shows the correlationbetween the simulated and the exper-imentally observed stimulus responses.Hence, the combined actions of diffu-sion and clearance appear to provide asound explanation for the local variationsin the stimulus response observed withmicrodisk electrodes, and in drug-inducedvariations in those responses.

At present, the functional significanceof the local variations revealed by thesemicrodisk-based measurements remainsto be fully explored. Nevertheless, it is cer-tainly intriguing to speculate that theselocal variations may be of great functionalsignificance, considering that the mam-malian brain contains multiple receptorsubtypes for many neurotransmitters, in-cluding dopamine. In fact, five geneticallydistinct dopamine receptors have beenidentified [59–65]. Most of the dopaminereceptors in the striatum, the brain re-gion in which most of our work has beencarried out, are of the so-called D1 andD2 type. One of the main distinctions be-tween these receptors is their affinity fordopamine. The D1 receptor has a rela-tively low affinity for dopamine, with aKD of 1–5 µm, while the D2 receptorhas about a 10-fold higher affinity, witha KD of 100 to 500 nM [66]. Apparently,these two dopamine receptor types aredesigned to respond to quite differentdopamine concentrations: we speculatethat the mechanisms giving rise to the localvariations in stimulus responses observedat microdisk electrodes, namely, diffusion

and clearance, may also be responsible forproducing the different dopamine concen-trations for which the D1 and D2 receptorsare designed.

14.6Monitoring Spontaneous, as Opposed toEvoked, Dopamine Release

Many of the applications of in vivo voltam-metry have focused on monitoring therelease of dopamine, and other neuro-transmitters, as evoked by the electricalstimulation of axons. The use of electri-cal stimulation is convenient because itevokes dopamine release at a specific andknown time. Moreover, the combination ofvoltammetry and electrical stimulation hasrevealed a large amount of useful informa-tion about the regulation of dopamine re-lease, the kinetics of dopamine uptake, andthe diffusion of dopamine in the extracel-lular space [67]. Nevertheless, the study ofevoked dopamine release does not provideinformation about spontaneous dopaminerelease, that is, dopamine release triggeredby the endogenous neuronal activity of thebrain. Such information is also of greatinterest.

Although some reports on the useof voltammetry to monitor spontaneousdopamine release do exist [68, 69], thereis a widely held perception that rest-ing extracellular dopamine levels in thebrain are simply too low to be mea-sured with voltammetry. Numerous stud-ies based on microdialysis, for exam-ple, suggest that the resting extracellularconcentration of dopamine in the ratstriatum is in the single-digit nanomo-lar range [70–72], while the detectionlimit of voltammetric methods is rarelymuch below 100 nM. But, there are twoproblems with the quantitative results


obtained with microdialysis. First, themicrodialysis-based estimates of dopamineconcentration do not take into account themanner in which diffusion and clearanceprocesses affect that concentration [73, 74].Considering Fig. 3, it seems reasonable tohypothesize that the concentrations ob-served by microdialysis are lower thanthose in closer vicinity to viable axonalterminals, where measurements can beperformed with microelectrodes. Second,it is difficult to rationalize the differencebetween the nanomolar concentrations ofdopamine reported by microdialysis andthe micromolar dopamine affinity valuesof the dopamine receptors, as mentionedat the end of the previous section. The D1receptor in particular, with its KD valueof 1 to 5 µm [66], would have little chanceto play any functional role in the brainif the extracellular dopamine concentra-tion were three orders of magnitude lowerthan its affinity for dopamine. On the otherhand, there is abundant evidence that D1

receptors do have functional significancein the brain [75], implying that there mustbe sufficient dopamine in the extracellu-lar space to cause their activation. Hence,there is good reason to expect that restingdopamine levels in the brain are substan-tially higher than existing estimates, andmay even be sufficiently high to detect byvoltammetric means.

Nevertheless, reports on the use ofvoltammetry to measure resting dopaminelevels in the brain are scarce. This ismainly because of the large contributionof background currents to the restingvoltammetric signal [2]. Especially whenfast-scan techniques are used, whichare highly selective toward dopamine,background current may comprise the vastmajority of the total signal observed inthe brain. Furthermore, the backgroundsignal recorded in brain tissue might be

quite different from that observed at thesame electrode in a calibration buffer.Hence, with the microelectrode implantedin the brain, it is essentially impossible toascertain what fraction of the resting signalis due to the oxidation of extracellulardopamine. A strategy that has allowed us toat least partially circumvent this problemis to monitor voltammetric signals inthe brain under conditions that lead tothe suppression of spontaneous dopaminerelease. The idea in this case is to monitorthe decrease in extracellular dopaminelevels expected when spontaneous releaseis suppressed.

The striatal region of the rat brain re-ceives both a dopamine projection fromthe midbrain and a glutamatergic projec-tion from the cortex. Although it is a matterof some controversy, several lines of evi-dence suggest that spontaneous release ofdopamine in the striatum is regulated bythe glutamatergic projection [76–83]. Glu-tamate is an amino acid neurotransmitterthat is recognized by several different re-ceptors found in the brain. Brain glutamatereceptors are divided into two main sub-types: the ionotropic and metabotropicglutamate receptors. Kynurenic acid is anaturally occurring substance that actsas a broadly selective antagonist of theionotropic glutamate receptors: in otherwords, it blocks those receptors fromactivation by glutamate. Recently, we con-ducted a series of experiments in whichkynurenic acid was infused directly intobrain tissue via a micropipet placed about100 µm from a carbon fiber microcylinderelectrode [84]. The infusion of kynurenicacid caused a sudden and clear decreasein the voltammetric signal at the car-bon fiber electrode (Fig. 6), and severallines of supporting evidence were consis-tent with the idea that this was due to adecrease in the extracellular concentration


Time[min]

0 4 8 12 16

Cha

nge

in o

xida

tion

curr

ent

[nA

]

−8

−4

0

4 180 nM

300 mV2 nA

Fig. 6 The effect of local injection of200 nL of 1 mM kynurenate on theresting dopamine signal measured instriatum with a carbon fibermicroelectrode. The kynurenate wasdelivered via a small pipet placed in thetissue about 100 µm from the electrode.The horizontal bar denotes the durationof the injection. The concentration scalebar was calculated from postcalibrationresults. Background subtractedvoltammogram (inset) was calculatedby subtracting the voltammogramscollected for 50 s at the beginning of theexperiment from the voltammogramscollected for 50 s at the end.

of dopamine. According to postcalibrationof the microelectrode, the magnitude ofthe kynurenate-induced decrease in thevoltammetric signal corresponded to a 450-nM decrease in the extracellular dopamineconcentration. Since we are not certainthat the infusion of kynurenic acid into thestriatum caused a complete suppressionof spontaneous dopamine release, this re-sult allows us to conclude that the restingdopamine concentration in the vicinityof the carbon fiber electrodes is at least450 nM. Nevertheless, this value, almostfivefold higher than the detection limitof the electrode for dopamine, is approx-imately two orders of magnitude higherthan the concentration estimates obtainedby microdialysis, and is consistent withthe dopamine affinities of the D1 and D2

receptors found in this brain structure.The results described in the previ-

ous paragraph suggest that extracellulardopamine levels in the rat striatum areindeed higher than the detection lim-its available with carbon fiber electrodes.Under resting conditions, however, thevoltammetric signal from dopamine isobscured by a large background current.Nevertheless, this result implies that theremay be many more opportunities than we

presently realize for examining changesin spontaneous dopamine release rates.So, we have started to explore this issuein rather more detail than before [85]. Wechose to examine the effects of drugs thatblock the dopamine uptake transporter,which include cocaine, nomifensine, am-phetamine, and Ritalin. Interest in themechanism of action of these drugs isderived from both their illicit and med-ical uses. Several studies have reportedthe effects of uptake inhibitors on thevoltammetric responses observed duringelectrical stimulation of dopamine axonsin the brain [45, 48], but these reportsrarely describe any effect of the drugson the voltammetric signal recorded inthe absence of the electrical stimulation.This has been a point of some confu-sion for a considerable amount of time,because the actions of these drugs arewidely attributed to their ability to in-crease extracellular dopamine levels [70,86]. We have been interested in under-standing why this expected increase isusually not observed when voltammetricmicroelectrodes are used.

These so-called uptake inhibitors, itturns out, do not just inhibit uptake. Infact, they indirectly change the rate of


spontaneous dopamine release. This isknown from studies in which electrophys-iological techniques have been used tomonitor the frequency at which dopamineneurons spontaneously generate actionpotentials, the electrical signal that trig-gers the release of neurotransmitters fromtheir storage vesicles. After the administra-tion of uptake inhibitors, the frequency atwhich dopamine neurons spontaneouslygenerate action potentials drops substan-tially [87, 88]. Thus, these drugs not onlydecrease the rate of dopamine uptake,they also decrease the rate of spontaneousdopamine release. In essence, this appearsto be a homeostatic response designedto regulate the extracellular dopamineconcentration. We hypothesized that thishomeostatic response is the reason that

dopamine levels, as measured with car-bon fiber microelectrodes, have not beenobserved to increase following pharmaco-logical uptake inhibition.

To test this hypothesis, we adminis-tered an uptake inhibitor, nomifensine, torats that had been pretreated with a drug,sulpiride, which blocks the D2 dopaminereceptor. In animals pretreated with thisdrug, the uptake inhibitors are less ef-fective at decreasing the action potentialfrequency in dopamine neurons, stronglyimplying that the D2 receptors mediate thehomeostatic response by providing neg-ative feedback information to dopamineneurons [85]. In the sulpiride-pretreatedrats, an increase in dopamine was ob-served with carbon fiber microelectrodesimmediately after the administration of

Time[min]

0 10 20 30 40 50

Cha

nge

in o

xida

tion

curr

ent

[nA

]

0

5

10

15

Nomifensineinjection

300 mV

5 nA

0.5 µM

Fig. 7 Change in resting dopamine signal after systemic administration of20 mg kg−1 nomifensine 60 min after administration of 100 mg kg−1 sulpiride. Thescale bar was determined from postcalibration results. Background subtractedvoltammogram (inset) was calculated by subtracting the voltammograms collectedfor 2 min at the beginning of the experiment from the voltammograms collected for2 min at the end.


nomifensine (Fig. 7), clearly showing thatsulpiride interferes with the homeostaticregulation that would have otherwiseprevented the increase in extracellulardopamine.

These results on the change in extra-cellular dopamine levels following phar-macological manipulations begin to showthat microelectrode-based methods for invivo neurochemical analysis have much tooffer. These microelectrodes, by virtue oftheir ability to get so close to viable axonalterminals in the brain, provide a brand newview of neurochemical events in the brain.In comparison to microdialysis, for exam-ple, the microelectrode not only providesa totally different estimate of extracellulardopamine levels (as discussed earlier) butalso provides new insights into the regula-tion of those levels. Microdialysis studies,for instance, did not reveal the abilityof kynurenic acid to decrease extracellu-lar dopamine levels [89–91]. Microdialysisstudies also did not reveal the homeostaticregulation of extracellular dopamine levelsby D2 receptors after the administrationof uptake inhibitors. The contrast betweenthe results obtained with microelectrodesand microdialysis probes can generally beattributed to the issue raised previously inthis chapter, namely, how close do the de-vices get to viable axonal terminals in thebrain.

14.7Electrochemical Sensors for in vivoMeasurement

While much of this chapter has focusedon the in vivo detection of the neuro-transmitter dopamine, which has beenthe paradigm in this field, there is alsoa great deal of interest in the develop-ment of electrochemical sensors for the

in vivo measurement of other substancesof neurochemical interest, including gluta-mate, choline, and glucose. Such sensorsare desirable because they can be fabri-cated from carbon fiber microelectrodes,affording them the spatial resolution re-quired to measure neurotransmitter re-lease in the vicinity of nerve terminals.There are several important considera-tions, however, for the development ofsensors for in vivo use. Care must betaken that the sensors are biocompatibleand that they are not susceptible to ex-cessive fouling by the brain. Additionally,in vivo electrochemical sensors must re-spond selectively to the neurotransmitterof interest.

Many different schemes have been usedin the design of sensors, several ofwhich make use of the selectivity affordedby the incorporation of enzymes. Themajority of enzyme sensors that havebeen successfully used in vivo rely onoxidase enzymes that generate hydrogenperoxide. While hydrogen peroxide can beoxidized directly at the electrode surface,this approach is inconvenient because itrequires the use of large positive potentialswhere many compounds oxidize. Thefollowing example demonstrates one wayin which these problems can be overcome.

Our group has designed electrochem-ical sensors for glutamate, choline, andglucose [11, 14–16]. These sensors areconstructed by coating carbon fiber mi-croelectrodes (r = 5 µm, l = 300–400 µm)with the oxidase enzyme, horseradish per-oxidase (HRP) and redox polymer (Fig. 8)containing an Os2+/Os3+ redox couple,and then applying a cross-linking agent.Ascorbate oxidase (AAox) can also be in-corporated to facilitate ascorbate rejection.The sensors are then cured, and a thin layerof Nafion is applied to increase the selectiv-ity for cations over anions. These sensors

14.7 Electrochemical Sensors for in vivo Measurement 479

N+N N N

CH2

CH2

NH2

OO

O

O

( )

Os

)( m

9

Redox polymer Cross linker

N Nbpy =

2+/3+

Os2+

Os3+e−

H2O

Electrode

H2O2

O2

HRP = horseradish peroxidaseGOx = glutamate oxidase

glu

αkg + NH3

(bpy)2Cl

HRP GOx

(a)

(b)

Fig. 8 (a) Scheme for the electrochemical detection ofglutamate. (b) Structure of the redox polymer and cross linker.

work according to the following scheme,diagramed in Fig. 8: the substrate is se-lectively oxidized by the oxidase enzyme,producing peroxide, which in turn, is re-duced by HRP. HRP is reduced by theOs2+/Os3+ redox couple, which is subse-quently reduced at the electrode surface,generating the electrochemical signal. Anadvantage of this sensor construction isthat interference from easily oxidizablecompounds, such as ascorbate, are elim-inated by operation of the sensor at lowpotentials (−100 mV vs. Ag/AgCl), inclu-sion of AAox, and coating with Nafion.

We have recently demonstrated in vitrothat these sensors respond robustly to thesubstrate of interest and exhibit no re-sponse to, or interference from, a varietyof substances present in vivo, includingascorbate, dopamine and dihydroxypheny-lacetic acid (DOPAC) [11, 14–16]. Further-more, we have verified pharmacologicallythat the sensors respond in vivo to several

compounds that are expected to alter extra-cellular levels of the substrate of interest.One such substance that is of particular im-portance neurochemically is tetrodotoxin(TTX), as will be discussed next.

TTX is a compound that halts thepropagation of action potentials and thesubsequent release of neurotransmitter byblocking sodium channels. TTX sensitivityhas long been recognized as a marker ofneuronal activity, and is particularly rele-vant in the case of glutamate, for whicha significant metabolic pool exists. Micro-dialysis measurements show no changesin extracellular glutamate upon local infu-sion of TTX, revealing that microdialysismeasures only the TTX insensitive, ormetabolic pool of glutamate [92–95]. Glu-tamate sensors, fabricated as describedearlier, measure a decrease in extracellu-lar glutamate upon TTX injection (Fig. 9),providing the first reported in vivo mea-surement of the neuronal component


Time[min]

0 5 10 15

Cur

rent

[pA

]

0

15

30

45

10 µM

Glu sensor

H2O2 sensor

Fig. 9 Signals recorded in vivo with glutamate microsensorand glutamate oxide–free background microsensor duringthe local injection of 200 nL of 100-µM TTX via a micropipet.The horizontal bar indicates the duration of the injection. Thescale bar was calculated based on postcalibration results.

of the extracellular glutamate concentra-tion. This result offers further evidencethat microelectrodes provide informationabout neuronally relevant events becausethey closely approach synaptic terminals.Local TTX injection was also monitored byglucose and choline sensors, which showno decrease in signal and a modest de-crease in signal, respectively, upon TTXinjection. These are the expected results:since glucose is not a neurotransmitter, itis not released by an impulse-dependentmechanism. Some component of the extra-cellular choline concentration results fromthe metabolism of the neurotransmitteracetylcholine, so it is reasonable that thecholine signal be partially TTX dependent.The specificity of the TTX response at thesesensors confirms the selectivity of sensorsconstructed with oxidase enzymes for thesubstrate of interest.

Glutamate is the major excitory neu-rotransmitter used by the brain, makingthe development of a method that reli-ably measures the neuronal component ofthe extracellular glutamate concentration

extremely desirable. Having done so, weare now able to make use of these sensorsto address various issues of neurochemicalinterest. One such question is: what isthe basal extracellular glutamate concen-tration? Our results provide a value instriatum of 30 µM. This is higher thanthe values of 1 to 4 µM provided by micro-dialysis measurements [92–98], consistentwith the idea that microsensors measureclose to neuronal terminals where concen-trations are highest. Another area of inter-est is the interaction between glutamateand other brain neurotransmitters. Wehave just begun to study the interaction ef-fect of dopamine and dopaminergic drugson extracellular glutamate concentration,and we are able to observe an interaction.

An additional application of sensorsbased on oxidase enzymes is that they canbe modified to respond only to changes inextracellular peroxide, a species of interestin the study of oxidative stress. Many neu-rodegenerative diseases, including Parkin-son’s disease, are thought to result at leastin part, from oxidative stress, a condition

14.8 Conclusion 481

characterized by the excessive productionof reactive oxygen species, including hy-drogen peroxide (H2O2) [99–105]. Sincedopamine is susceptible to autooxida-tion and to oxidative deamination bymonoamine oxidase (MAO) during itsmetabolism, and both of these oxidativeprocesses lead to the formation of per-oxide, an increase in dopamine releasemay induce an overproduction of peroxide.Therefore, the direct in vivo measurementsof peroxide would potentially provide in-sight into the dopaminergic induction ofoxidative stress.

Recently, we have used microsensors se-lective for peroxide to monitor the in vivoformation of peroxide in rat striatum fol-lowing electrical and pharmacological ma-nipulations of the dopamine system [22].These studies reveal a biphasic increasein peroxide in striatum upon brief (10 s)electrical stimulation of the MFB (Fig. 10),demonstrating that our sensors are able tomonitor the real-time formation of perox-ide in vivo. The combination of selectiveand sensitive hydrogen peroxide microsen-sors with selective axonal stimulation pro-vided a valuable new approach to the inves-tigation of the neurodegenerative mecha-nisms associated with Parkinson’s disease.

14.8Conclusion

In this chapter, we have tried to illustratethe power of microelectrode-based tech-niques that have been developed for mon-itoring ongoing neurochemical events inthe living brain. We have not dwelt onthe technical details of the electroanalyticalmethods themselves: these are both wellknown to the likely readers of a volumesuch as this and are readily found in theoriginal citations we have provided. Rather,our goal has been to focus attention onthe unique capabilities of microelectrode-based measurements, which are mainlyderived from the minimal tissue disrup-tion that the electrodes inflict on the brain.The minimal tissue disruption makes itpossible to implant these microelectrodesto within micrometer, maybe even sub-micrometer, distances of intact, viableaxonal terminals. Hence, microelectrodesallow concentrations of neurochemicallyactive substances to be observed in themost important location within the tis-sue, namely, right at the site of action ofthe neurochemicals themselves. To date,no other technique permits monitoring inthis prime location.

Fig. 10 Signals recorded in vivo withperoxide microsensor and HRP freebackground sensor during an electricalstimulation of the MFB. Open circlesdenote the beginning and end of theelectrical stimulation. The scale bar wascalculated based on postcalibrationresults.

Time[min]

0 4 8 12 16 20

Cur

rent

[pA

]

0

4

8

12

160.5 µM

HRP free sensor

H2O2 sensor


References

1. P. T. Kissinger, J. B. Hart, R. N. Adams,Brain Res. 1973, 55, 209–213.

2. J. E. Baur, E. W. Kristensen, L. J. May et al.,Anal. Chem. 1988, 60, 1268–1272.

3. M. E. Rice, C. Nicholson, Anal. Chem. 1989,61, 1805–1810.

4. D. J. Wiedemann, K. T. Kawagoe, R. T.Kennedy et al., Neuroscience 1988, 25,513–524.

5. P. Brun, M. F. Suaud-Chagney, F. Gononet al., Eur. J. Pharmacol. 1993, 235, 205–210.

6. P. Brun, M. F. Suaud-Chagney, F. Gononet al., Neuroscience 1993, 52, 961–972.

7. P. Bickford-Wimer, K. Pang, G. M. Roseet al., Brain Res. 1991, 558, 305–311.

8. L. C. Daws, G. A. Gerhardt, A. Frazer, Neu-rosci. Lett. 1999, 266, 165–168.

9. L. C. Daws, G. M. Toney, G. A. Gerhardtet al., J. Pharmacol. Exp. Ther. 1998, 286,967–976.

10. J. A. Stamford, Z. L. Kruk, J. Millar, BrainRes. 1990, 515, 173–180.

11. N. V. Kulagina, L. Shankar, A. C. Michael,Anal. Chem. 1999, 71, 5093–5100.

12. Y. Hu, K. M. Mitchell, F. N. Albahadilyet al., Brain Res. 1994, 659, 117–125.

13. J. J. Burmeister, G. A. Gerhardt, Anal.Chem. 2001, 73, 1037–1042.

14. J. Cui, N. V. Kulagina, A. C. Michael, J. Neu-rosci. Methods 2001, 104, 183–189.

15. M. G. Garguilo, A. C. Michael, Anal. Chem.1994, 66, 2621–2629.

16. M. G. Garguilo, N. Huynh, A. Proctor et al.,Anal. Chem. 1993, 65, 523–528.

17. C. Desvignes, F. Robert, C. Vachette et al.,Neuroreport 1997, 8, 1321–1325.

18. M. N. Friedemann, S. W. Robinson, G. A.Gerhardt, Anal. Chem. 1996, 68, 2621–2628.

19. H. Tu, J. Xue, X. Cao et al., Analyst 2000,125, 163–167.

20. C. Meyerhoff, F. Bischof, F. J. Mennel et al.,Biosens. Bioelectron. 1993, 8, 409–414.

21. M. Mascini, S. Fortunati, D. Moscone et al.,Clin. Chem. 1985, 31, 451–453.

22. N. V. Kulagina, A. C. Michael, unpublishedobservations.

23. R. T. Kennedy, S. R. Jones, R. M.Wightman, Neuroscience 1992, 47, 603–612.

24. J. B. Zimmerman, R. M. Wightman, Anal.Chem. 1991, 63, 24–28.

25. J. L. Peters, A. C. Michael, J. Neurochem.2000, 74, 1563–1573.

26. R. J. Preston, G. A. Bishop, S. T. Kitai,Brain Res. 1980, 183, 253–263.

27. J. L. Peters, A. C. Michael, S. R. Sesack, un-published observations.

28. J. R. Cooper, F. E. Bloom, R. H. Roth, TheBiochemical Basis of Neuropharmacology,Oxford University Press, Oxford, 1991.

29. K. Fuxe, L. F. Agnati, (Eds.), Volume Trans-mission in the Brain, Raven Press, New York,1991.

30. T. E. Robinson, J. B. Justice Jr., (Eds.),Techniques in the Behavioral Neural Sciences,Elsevier, Amsterdam, 1991, Vol. 7.

31. K. L. Clapp-Lilly, R. C. Roberts, L. K. Duffyet al., J. Neurosci. Methods 1999, 90,129–142.

32. R. E. Carson, M. E. Daube-Witherspoon,P. Herscovitch, (Eds.), Quantitative Func-tional Brain Imagining with Positron Emis-sion Tomography, Academic Press, SanDiego, 1998.

33. M. Laruelle, J. Cereb. Blood Flow Metab.2000, 20, 423–451.

34. Z. Zhou, S. Misler, Proc. Natl. Acad. Sci.1995, 92, 6938–6942.

35. G. Y. Chen, P. F. Gavin, G. A. Luo et al.,J. Neurosci. 1995, 15, 7747–7755.

36. P. M. Groves, J. C. Linders, S. J. Young,Neuroscience 1994, 58, 593–604.

37. V. M. Pickel, S. C. Beckley, T. K. Joh et al.,Brain Res. 1981, 225, 373–385.

38. C. Nicholson, Biophys. J. 1995, 68,1699–1715.

39. J. Crank, The Mathematics of Diffusion,Clarendon, Oxford, 1975.

40. H. Yang, J. L. Peters, C. Allen et al., Anal.Chem. 2000, 9, 2042–2049.

41. K. T. Kawagoe, P. A. Garris, D. J. Wiede-mann et al., Neuroscience 1992, 51, 55–64.

42. C. Nicholson, M. E. Rice in VolumeTransmission in the Brain (Eds.: K. Fuxe,L. F. Agnati), Raven Press, New York, 1991,pp. 279–294.

43. J. L. Peters, A. C. Michael, S. R. Sesack, un-published observations.

44. Y. Lu, J. L. Peters, A. C. Michael, J. Neuro-chem. 1998, 70, 584–593.

45. L. J. May, W. G. Kuhr, R. M. Wightman,J. Neurochem. 1988, 51, 1060–1069.

46. P. A. Garris, R. M. Wightman, J. Neurosci.1994, 14, 442–450.

47. S. R. Jones, P. A. Garris, C. D. Kilts et al.,J. Neurosci. 1995, 64, 2581–2589.

14.8 Conclusion 483

48. R. M. Wightman, J. B. Zimmerman, BrainRes. Rev. 1990, 15, 135–144.

49. W. G. Kuhr, J. G. Bigelow, R. M. Wight-man, J. Neurosci. 1986, 6, 974–982.

50. H. Yang, A. C. Michael, unpublished obser-vations.

51. H. Yang, J. Qian, A. C. Michael, unpub-lished observations.

52. H. Yang, J. L. Peters, A. C. Michael, J. Neu-rochem. 1998, 71, 684–692.

53. C. Allen, J. L. Peters, A. C. Michael, unpub-lished observations.

54. L. J. May, R. M. Wightman, Brain Res. 1989,487, 311–320.

55. C. Spyraki, H. C. Fibiger, Science 1981, 212,1167, 1168–.

56. S. Fielding, M. R. Szewczak J. Clin. Psychi-atry 1984, 45, 12–20.

57. A Barbeau, F. H McDowell, (Eds.), L-Dopaand Parkinsonism, F. A. Davis, Philadelphia,1970.

58. D. B. Calne, M. Sandler, Nature 1970, 226,21–24.

59. J. R. Bunzow, H. H. M. van Tol, D. K.Grandy et al., Nature 1988, 336, 783–787.

60. A. Dearry, J. A. Gingrich, P. Falardeauet al., Nature 1990, 347, 72–76.

61. F. J. Monsma, L. C. Mahan, L. D. McVittieet al., Proc. Natl. Acad. Sci. U.S.A. 1990, 87,6723–6727.

62. P. Sokoloff, B. Giros, M. P. Martres et al.,Nature 1990, 347, 146–151.

63. R. K. Sunahara, H. C. Guan, B. F. O’Dowdet al., Nature 1991, 350, 614–619.

64. H. H. M. van Tol, J. R. Bunzow, H. C. Guanet al., Nature 1991, 350, 610–614.

65. Q. Y. Zhou, D. K. Grandy, L. Thambi et al.,Nature 1990, 347, 76–80.

66. C. Missale, S. R. Nash, S. W. Robinsonet al., Physiol. Rev. 1998, 78, 189–225.

67. R. M. Wightman, C. Amatore, R. C. Eng-strom et al., Neuroscience 1988, 25, 513–522.

68. A. Gratton, R. Wise, J. Neurosci. 1994, 14,4130–4144.

69. E. A. Kiyatkin, Eur. J. Neurosci. 1993, 5,284–291.

70. R. J. Olson, J. B. Justice Jr., Anal. Chem.1993, 65, 1017–1022.

71. P. M. Sam, J. B. Justice Jr., Anal. Chem.1996, 68, 724–728.

72. L. H. Parsons, J. B. Justice Jr., J. Neurochem.1992, 58, 212–218.

73. J. L. Peters, H. Yang, A. C. Michael, Anal.Chim. Acta 2000, 412, 1–12.

74. J. L. Peters, A. C. Michael, J. Neurochem.1998, 70, 594–603.

75. C. R. Gerfen, T. M. Engber, L. C. Mahanet al., Science 1990, 250, 1429–1432.

76. A. Cheramy, R. Romo, G. Godeheu et al.,Neuroscience 1986, 19, 1081–1090.

77. D. W. Clow, K. Jhamandas, J. Pharmacol.Exp. Ther. 1988, 248, 722–728.

78. M. M. Iravani, Z. L. Kruk, J. Neurochem.1996, 66, 1076–1085.

79. S. Jin, B. B. Fredholm, Br. J. Pharmacol.1997, 121, 1269–1276.

80. D. Martinez-Fong, M. G. Rosales, J. L.Gongora-Alfaro et al., Brain Res. 1992, 595,309–315.

81. G. Segovia, A. Del Arco, F. Mora, J. Neuro-chem. 1997, 69, 1476–1483.

82. Y. Wu, S. M. Pearl, M. J. Zigmond et al.,Neuroscience 2000, 96, 65–72.

83. K. D. Youngren, D. A. Daly, B. Moghad-dam, J. Pharmacol. Exp. Ther. 1993, 264,289–293.

84. N. V. Kulagina, M. J. Zigmond, A. C.Michael, Neuroscience 2001, 102, 121–128.

85. H. Yang, A. C. Michael, unpublished obser-vations.

86. A. D. Smith, J. B. Justice Jr., J. Neurosci.Methods 1994, 54, 75–82.

87. N. B. Mercuri, F. Stratta, P. Calabresi et al.,Neurosci. Lett. 1991, 131, 145–148.

88. L. C. Einhorn, P. A. Johansen, F. J. White,J. Neurosci. 1988, 8, 100–112.

89. K. A. Keefe, M. J. Zigmond, E. D. Aber-crombie, Neuroscience 1992, 47, 325–332.

90. D. W. Miller, E. D. Abercrombie, Brain Res.Bull. 1996, 40, 57–62.

91. B. Moghaddam, R. J. Gruen, Brain Res.1991, 544, 329, 330.

92. M. W. Lada, T. W. Vickroy, R. T. Kennedy,J. Neurochem. 1998, 70, 617–625.

93. M. Herrera-Marschitz, Z. B. You, M. Goinyet al., J. Neurochem. 1996, 66, 1726–1735.

94. J. Semba, S. Kito, M. Toru, J. Neural Transm.1995, 100, 39–52.

95. M. Shiraishi, Y. Kamiyama, P. C. Hutte-meier et al., Brain Res. 1997, 759, 221–227.

96. E. Zilkha, T. Obrenovitch, A. Koshy et al.,J. Neurosci. Methods 1995, 60, 1–9.

97. O. Niwa, K. Torimitsu, M. Morita et al.,Anal. Chem. 1996, 68, 1865–1870.

98. M. W. Lada, T. W. Vickroy, R. T. Kennedy,Anal. Chem. 1997, 69, 4560–4565.

99. S. Fahn, G. Cohen, Ann. Neurol. 1992, 32,804–812.


100. H. S. Maker, C. Weiss, D. J. Silides et al.,J. Neurochem. 1981, 36, 589–593.

101. B. Halliwell, J. Neurochem. 1992, 59,1609–1623.

102. J. T. Coyle, P. Puttfarcken, Science 1993,262, 689–695.

103. M. E. Gotz, G. Kunig, P. Riederer et al.,Pharmacol. Ther. 1994, 63, 37–122.

104. P. Jenner, C. W. Olanow, Neurology 1996,47, 161S–170S.

105. N. A. Simonian, J. T. Coyle, Ann. Rev. Phar-macol. Toxicol. 1996, 36, 83–106.

485

15Potentiometric Measurements ofProteins

Jackson Pellett and Marian StankovichChemistry Department, University of Minnesota, Minneapolis, MN, USA

15.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48715.1.1 What Electrochemical Studies Can Tell Us . . . . . . . . . . . . . . . . . . 48715.1.2 Direct versus Indirect Electrochemistry . . . . . . . . . . . . . . . . . . . . 488

15.2 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49015.2.1 Mediator Titrant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49015.2.2 Redox Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491

15.3 Apparatus – Glassware Components for Anaerobiosis . . . . . . . . . 49215.3.1 Argon Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49215.3.2 The Cell Parts: Preparation of Auxiliary, Reference, and Working

Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493

15.4 Methods – Setting up the Mediated Electron Transfer Experiment . 49715.4.1 The Coulometric Titration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49715.4.2 The Potentiometric Titration . . . . . . . . . . . . . . . . . . . . . . . . . . . 501

15.5 Data Analysis and Sample Systems . . . . . . . . . . . . . . . . . . . . . . . 50315.5.1 Coulometric Titration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50315.5.2 Potentiometric Titration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 509

487

15.1Introduction

Why is the redox potential measurementof enzymes important?

Many enzymes use redox centers to storeand transfer electrons during catalysis.These redox centers can be composed ofmetals such as iron or cobalt, or organiccofactors such as quinones, amino acidradicals, or flavins. In order to fully ap-preciate the catalytic mechanisms of theseenzymes, it is often necessary to deter-mine the free energy required to reduceor oxidize their protein redox centers. Thisis called the redox potential. The mea-surement of enzyme redox potentials canbe performed by either direct or indi-rect electrochemical methods. The typeof electrochemistry suitable for a partic-ular protein system is simply dictated bythe accessibility of its redox center to theelectrode surface. Because most reactionscatalyzed by enzymes occur within hy-drophobic pockets of the protein, the redoxsites are often far from the surface of theprotein. Unless an electron transfer pathexists from the protein surface to the re-dox center, it is not feasible to use directelectrochemistry to measure the redox po-tential. Since only a few enzymes (mostnotably certain heme-containing enzymes)have such electron transferring paths and

are thus amenable to direct electrochemi-cal study, this review will focus primarilyon indirect electrochemical methods forstudying redox active enzymes. Unless oth-erwise stated, the potentials determined inthis work are reported as formal potentialsat pH 7.0 (E′

) versus the standard hydro-gen electrode (SHE). This nomenclatureis frequently used by biochemists, as thephysiologically relevant pH for biologicalmolecules is generally close to neutral. Inreporting a formal potential for a partic-ular protein, we ascertain that it obeysthe Nernst equation under a set of veryspecific conditions of ionic strength, pH,temperature, and protein concentration.

Although the formal potential is re-ported versus the SHE, the measurementsare made using a silver–silver chloride ref-erence electrode that is calibrated versusferri/ferrocyanide.

Before discussing the specific electro-chemical methods used to measure pro-tein redox potentials, we will examine thetype of information that an enzymologistcan obtain from electrochemical studies.

15.1.1What Electrochemical Studies Can Tell Us

It is crucial that the redox potentials ofenzymes that catalyze reactions involvingelectron transfer be rigidly controlled.

488 15 Potentiometric Measurements of Proteins

This is true because the interactionsbetween the protein and redox centerenable enzymes to modulate the redoxproperties of the limited number ofmetals and cofactors available in nature.This regulation arises from the tighterbinding by the apoprotein of one oxidationstate of the redox active cofactor relativeto the other oxidation state. This isillustrated in Fig. 1 for the noncovalentbinding of flavin adenine dinucleotide(FAD) to porcine medium-chain acyl-CoAdehydrogenase (MCAD). As a result ofthe tighter binding of the two-electron-reduced FAD (FADH2−red) to MCAD(Ka = 2 × 106 vs Ka = 6 × 108), the redoxpotential for the two-electron reduction ofFAD is shifted negative by 74 mV uponprotein binding [1]. Several factors canaccount for the preferential binding ofone oxidation state of a metal or cofactorover another. These include the formationof covalent and hydrogen bonds as wellas electrostatic and solvent interactionswith the redox center. Since changesin these interactions can have profound

effects on the redox potentials of thespecies involved, the redox properties ofenzymes can also serve as sensitive probesof the protein environment surroundingthe redox center. For instance, changesin the ionization state of amino acids inthe vicinity of redox centers may perturbthe redox potential of the enzyme. Theseredox-linked ionizations have been usedsuccessfully to assign pKas of certainresidues in the active sites of severalenzymes [1–3].

15.1.2Direct versus Indirect Electrochemistry

The electrochemical properties of mostof the metals and cofactors used by en-zymes can be characterized using standarddirect voltammetric electrochemical meth-ods (see Chapter 1 by Fraser Armstrongand Chapter 11 by Katsomi Niki). Whenfree in solution, these species interact wellwith a variety of different electrodes andusually exhibit Nernstian behavior uponoxidation or reduction. The incorporation

O

NHNH

N N O

O

NH

O

N

N N

O

NH

O

N

N N

O

NHNH

N N ONO

O

OH

O

−O

NO

- -

+2e−

+1H++2e−

+2H+

Kox

Kred

E°′(bound)−219 mV (pH 7)

E°′(free)

E°′(bound) E°′(free)KredKox

RTnF

= + 2.303 log

Fig. 1 Regulation of redox potential.


of these species into the protein matrix,however, usually results in unfavorableelectron transfer properties between theredox center and most electrodes. In somecircumstances, it is possible to constructmodified electrode surfaces that do interactwell with certain proteins [4, 5].

One advantage of being able to performthe voltammetric analysis of enzymes isthat it is possible to obtain informationabout the kinetics of electron transfer byvarying the scan rate. It is crucial, how-ever, that proper control experiments beperformed to demonstrate that interac-tions between the modified electrode andenzyme are not perturbing the electro-chemical properties of the enzyme. In thelikely event that voltammetric analysis isnot feasible for a particular protein system,indirect electrochemical methods are oftensuccessful.

An advantage of dynamic techniquessuch as cyclic voltammetry is that po-tentials of unstable states can be morereadily evaluated. They are most applicableto proteins whose function is to transferelectrons (cytochromes, ferredoxins, etc.)rather than catalyze redox reactions. Flavo-proteins and dinuclear iron proteins arein this category: they have the equivalentof very low self-exchange rates, meaningthat they cannot readily transfer electronsto each other, and they also do not interactwell with electrodes. This is because theiractive sites are only accessible to smallmolecules. Accessibility to active sites ofenzymes is also a topic that has been ad-dressed by Willner in Chapter 17 in thisvolume. In our flavoprotein and dinucleariron cluster work, we will consider onlythose proteins that have the dual functionof transferring electrons and selectivelycatalyzing reactions.

Four different indirect methods thatrely on UV-visible spectroscopy have been

successfully used to determine the redoxpotentials of proteins. In the first methodto be developed, sodium dithionite is usedas a reductant in an anaerobic cell, withthe spectral changes of enzymes and re-dox indicator dyes monitored as a functionof the amount of reductant added [6–9].Massey devised a second method, whichuses 5-deazaflavin and ethylenediaminete-traacetic acid (EDTA) light to generate thepowerful reductant, the deazaflavin radi-cal [10]. Spectral changes are monitoredas a function of irradiation time. A thirdmethod was later developed by Massey, inwhich a large amount of xanthine (the re-ductant) and a small amount of xanthineoxidase (the catalyst) are added initially toa solution of enzyme and redox indicatordye. Note that in contrast to other meth-ods, the reductant is not added stepwisebut batchwise at the beginning of the ex-periment. The amount of xanthine oxidasecatalyst used is also very small to allow thereduction to proceed relatively slowly, al-lowing sufficient time for the enzyme andmediator to equilibrate during the entirecourse of the reduction. Spectra are takenduring the course of the reduction. Sinceboth the enzyme and dyes are spectrallyactive, the position of equilibrium at eachpoint of the reduction is used to calculatethe redox potential of the enzyme [11].

The redox potential of the mediatordye–enzyme system is not measureddirectly using reference and indicatorelectrodes in any of the methods justdescribed. In contrast, our spectroelec-trochemical method [12], which is basedon the methods of Kuwana [13] and Wil-son [14], is still an indirect method. Weuse mediator dyes, which are in equi-librium with the enzyme and have theadvantage of measuring the potential thatis associated with each equilibrium po-sition of the dye and enzyme as both


become reduced in a reductive titration.In order to do this, we have designedan anaerobic cell in which both spectraland electrochemical measurements maybe made. All the components – custommade electrodes, reference, auxiliary, andworking – are designed to fit into the cell.A spectrophotometer is required, as wellas a potentiostat to electrochemically gen-erate a titrant and measure equilibriumredox potentials.

15.2Materials

This chapter provides the reader with thebackground necessary to determine theredox potentials of proteins using goldelectrodes whose surfaces are unmodified,so that the potential measurements relyon the equilibration of enzymes and re-dox mediators. The electrodes sense thepotential of the mediator dyes, not theenzyme; that is why this is termed an indi-rect electrochemical approach. This indirectmethod is useful because the active sitesof most enzymes are buried and do notinteract well with modified or unmodifiedelectrodes. Although this indirect methodwas originally designed to work on flavo-proteins, it has been applied to enzymescontaining a variety of redox and spectrallyactive cofactors. The chemical reductionthat takes place is very similar to what wefind in a dithionite titration, except thatin this case the titrant is generated in situelectrochemically. The titrant then trans-fers electrons quantitatively to the enzyme.When the potential is measured, a secondmolecule, a redox indicator is present toequilibrate with the partially reduced en-zyme and the electrode, which now actsas a redox sensor. It can now be seenthat the working electrode plays two roles

in these experiments: first, its potential iscontrolled at a negative value to reducethe mediator titrant, and second, when theenzyme and mediator dye are allowed toequilibrate at open circuit, the electrodemerely reports the potential achieved atequilibrium.

This method was developed to overcomeseveral disadvantages of the earlier dithion-ite titration method, the most severe beingthe complexation of the flavoprotein ox-idases by bisulfite, an oxidation productof dithionite [15]. The present methodpresents additional advantages: it is ableto generate many redox titrants in situwithout standardization by titration, andhas a more flexible experimental design,as indicated in the following section.

15.2.1Mediator Titrant

On the basis of the work of Szentirmayand Kuwana [13], methyl viologen (MV++)was chosen as the electrogenerated medi-ator titrant. MV++ has a negative redoxpotential, ( −450 mV) and once reducedat the electrode, quantitatively transferselectrons to most flavoproteins and otherredox enzymes, with redox potentials inthe range of −100 to −350 mV. MV++has other advantages as a mediator titrant(Table 1): (1) its oxidized form does not ab-sorb visible light, so its spectrum does notinterfere with that of the enzyme; (2) it re-acts quickly with the electrode and is ableto transfer electrons quickly and quantita-tively to many enzyme active sites, even tothose of enzymes that cannot be reduced bydithionite. Methyl viologen’s unique reac-tivity may be due to its aromatic characterand the fact that it is positively chargedeven in the reduced state, in contrast tomost other dyes that are negatively charged

15.2 Materials 491

Tab. 1 Selected redox potentials of mediator/titrants

E′ (mV vs SHE)a Compound Reference

465 1,1′-bis(hydroxymethyl)ferrocene 13425 Ferri/ferrocyanide 17422 Ferrocene 13

92 Phenazine methosulfate 660 1,4 napthoquinone 655 Phenazine methosulfate 65 Methylene blue 60 Duroquinone 6

−3 5-hydroxy-1, 4-naphthoquinone 13−10, −19 Pyocyanine 6, 13

−46 Indigo-tetrasulfonic acid 6−116 Indigo-disulfonate 6−133 2-amino-1,4 napthoquinone 13−139 2-hydroxy-1,4-naphthoquinone 6−150 8-Cl-riboflavin 18−208 Riboflavin, FAD, FMNb 6−225 Anthraquinone -2-sulfonate 6−244 Phenosafranin 6−325 Neutral red 6−350 Benzyl viologen 6−446 Methyl viologen 6, 13

aPotentials measured at pH 7.bFlavin mononucleotide.

or neutral; (3) it is a good redox indica-tor for enzymes with negative potentials;and (4) the characteristic spectrum of re-duced methyl viologen (MV+•) signals theend point of the titration. The advantageof using MV++ is that it can be presentat high concentrations (0.1 mM) to rapidlytransfer electrons to the enzyme (10 µM).Speed is important in ‘‘coulometric titra-tion’’ where we are counting the numberof electrons required to reduce the enzymein order to minimize the time in whichoxygen leaks can occur.

While benzyl viologen is similar instructure to MV++, its potential is morepositive (E′ = 350 mV). It can be usedas an alternative mediator titrant for anenzyme if binding to an enzyme by MV++is suspected. A reductive titration withdithionite can be used as a control to test

to see if MV++ is binding to the enzymeunder study.

Ferrocene is used as an oxidativetitrant – its potential is much more pos-itive than those of the enzymes (+0.5 V),so it can be used as a chemically generatedoxidant, useful for reversing the reductionreaction.

15.2.2Redox Indicators

These dyes are spectrally and redox active,and must have redox potentials near thatof the enzyme to function as a ‘‘redoxbuffer’’. The potential of the enzyme ismeasured at equilibrium after a certainnumber of reducing equivalents have beentransferred. The dye equilibrates with thepartially reduced enzyme at open circuit to


produce partially reduced dye and partiallyreduced enzyme. The working electrode,acting as an indicator electrode, reportsthe potential of the dye–enzyme mixture.Since the indicator dye is the link tothe electrode in the equilibrium mixture,these dyes must (1) be able to equilibrateboth with the enzyme active site and theelectrode, (2) exhibit reversible electrontransfer behavior themselves, and (3) bestable under experimental conditions.They must not bind to the enzyme. Afterequilibrium is reached, more reducingequivalents are added, so that a Nernstplot can be constructed and E

′and n can

be obtained. (Equilibrium is determinedto be the point at which the potential andspectra of enzyme and dye stop changing.)

Extensive lists of redox indicator dyesare given in Clark [6], Kuwana [13], andthe CRC Handbook of Biochemistry [16].Some of the best indicator dyes forflavoproteins and dinuclear iron proteinsare pyocyanine, indigo disulfonate (IDS),8-chlororiboflavin, and lumiflavin acetate.Redox indicator dyes must be titrated, andtheir spectra and potentials determinedunder conditions of temperature, pH,and buffer identical to those in theenzyme experiments to allow for spectralcorrection.

15.3Apparatus – Glassware Componentsfor Anaerobiosis

In order to perform any of these studies,the system has to be made oxygen-free andmaintained in this state if quantitative elec-tron transfer to the proteins and mediatorsystem is to occur, making it necessary toinstall an argon/vacuum line, with scrub-bers; special spectroelectrochemical cellsare needed as well. This elaborate setup

for oxygen removal and anaerobic main-tenance is necessary because the simplermethods of oxygen removal, such as bub-bling argon or electrochemical scrubbing,cannot be used with proteins because theylead to denaturation.

Since the system design is driven bythe requirement for anaerobiosis, the elec-trodes must fit inside a spectrophotometriccuvette whose contents have previouslybeen made anaerobic by the proceduredescribed in the following section. Theelectrodes must also be custom made tofit inside without leaking oxygen, eitheraround the joints or through the tips. Theymust also be stored anaerobically betweenexperiments (see following section).

15.3.1Argon Line

Figure 2 illustrates a typical anaerobic trainthat can be used to remove dissolvedoxygen from buffers and protein solu-tion. Pure argon (99.99%) is passed firstthrough an Oxyclear disposable cartridge(Fisher) and then over activated Ridox cata-lyst (Fisher) to remove all residual oxygen.While the Ridox scrubber can be usedat room temperature, unlike some otherscrubbers, it is important that it be ac-tivated according to the manufacturer’sinstructions before it is first used. This isdone by heating the scrubber to 215 Cwhile a 10% hydrogen and argon mix-ture is passed over the Ridox, liberatingwater. Once activated, the Ridox turns red-dish. After the activated Ridox is cooledby flowing argon, the line is ready to run.The pressure is maintained at 5 psi, us-ing a regulator to keep excess pressurefrom building up in the line. The sec-ond side of the nitrogen line is attachedto a vacuum pump, making it possible toswitch between positive gas pressure and a

15.3 Apparatus – Glassware Components for Anaerobiosis 493

To argon

Trap

Trap

Bubbler

Bubbler

To vacuum

To vacuum

Neoprene tubing Neoprene tubingto spec cell

Neoprene tubingto spec cell

Ridox

Fig. 2 Argon line.

vacuum by turning the two-way stopcock.This will be necessary when degassingprotein solution (see following text). Neo-prene rubber tubing is used to connect thecells to the argon line because the oxygenpermeability of neoprene is much lowerthan that of other polymers, for example,Tygon. Once in operation, the argon linecan be maintained to give 1-ppm O2.

15.3.2The Cell Parts: Preparation of Auxiliary,Reference, and Working Electrodes

Working, auxiliary, and reference elec-trodes are custom made to fit individualspectroelectrochemical cells. Three factorsshould be considered when designing

these electrodes. First, all the electrodesmust be long enough to make contact withthe protein solution. Second, the tips ofthe electrodes must be thin enough to fitcomfortably inside a 1 cm × 1 cm squarecuvette. Third, the electrodes must capableof being made anaerobic.

The components of a silver–silver chlo-ride reference electrode are shown in Fig. 3.First, cleaned silver wire is epoxied intoa 7/25 female ground glass joint. A layerof AgCl is then deposited on the silverby dipping the wire into a HCl : HNO3

(3 : 1) solution. A small thirsty glass plug(Bioanalytical Systems; BAS) is attachedwith heat-shrink tubing to the end of theslender glass sleeve (made by the glassblower) that is topped by back-to-back male


Silver wireFrit

Glass sleeve

Reference electrode

Fig. 3 Reference electrodecomponents and assembly.

7/25 ground glass joints. This thirsty glasscapped sleeve provides electrical contactbetween the electrode and the solutionwithout allowing the analyte solution tomix with the electrolyte. The sleeve isfilled with a 0.10 M KCl solution satu-rated with AgCl. The silver wire is insertedand the 7/25 ground glass joint is sealedwith Apiezon N stopcock grease. The sil-ver–silver chloride reference electrode isnow complete. The equation that governsthe potential of this reference electrode is:

AgCl(s) + e− = Ag0 + Cl− (0.1 M) (1)

Because the electrode’s potential is de-pendent on the type of crystal structureof the AgCl(s) that is used, as well asother factors, the potential is standardizedby measuring it versus that of standardferri/ferrocyanide solution, as describedbelow [17]. It remains sealed and is notremade frequently. Since this electroderemains sealed, its potential remains rel-atively constant. However, its potential ischecked before each experiment. The po-tential difference is about −144 mV, whichindicates that the reference electrode is144 mV more negative than the +425 mVof the ferri/ferrocyanide, or +281 mV ver-sus SHE. This is close to what one wouldcalculate for a silver–silver chloride elec-trode with 0.1 M chloride.

The ferri/ferrocyanide system must bekept out of light and free of oxygen inorder to be stable. The solution shouldalso be made fresh daily.

The auxiliary electrode has an unusualcomposition, designed to prevent the gen-eration of oxygen. This is important be-cause most of the reactions taking place atthe working electrode are reductions; thus,equal and opposite reactions (oxidation)will be taking place in the auxiliary elec-trode. If oxygen were generated, it coulddiffuse into the solution, reoxidizing theenzyme analyte. To prevent oxygen beinggenerated from the water as would oc-cur at a typical auxiliary electrode in anaqueous solution containing a supportingelectrolyte, the composition of the auxiliaryelectrode is modified.

2H2O ===⇒ 4H+ + O2 + 4e− (2)

The modified auxiliary electrode is asilver wire in contact with 0.1 M KCl. Inthis case, since the oxidation describedin Reaction (3) is easier to carry out thanthat described in the oxidation of water(Reaction 2), Reaction (3) would occurpreferentially.

Ag0 + Cl− ===⇒ AgCl(s) + e− (3)

To achieve this end, the compositionof the auxiliary electrode is very similar

15.3 Apparatus – Glassware Components for Anaerobiosis 495

Fig. 4 Auxiliary electrodecomponents and assembly.

Coiled silver wire

Copper lead

Glass sleeve Auxiliary electrode

to that of the reference electrode, exceptthe electrode area is much larger, equal tothat of the working electrode (Fig. 4). Inthe auxiliary electrode, silver wire is en-cased (enclosed) in a glass sleeve tippedwith thirsty glass to allow electrical con-tact with the solution, while preventingthe electrode compartment contents frommixing with the solution. This glass sleeveis filled with 0.1 M KCl. Sometimes anoxygen scrubber such as glucose–glucoseoxidase is added to the auxiliary elec-trode. The electrode must be remadebefore each experiment because, unlikethe reference electrode, its compositionchanges with time as oxidation and reduc-tion occur, and products buildup in thesolution.

Finally, the working electrode is a coiledgold wire. As shown in Fig. 5, gold wire is

sealed into a female 7/25 ground glass jointwith epoxy. The gold wire is then cleanedwith a slurry of ethanol and silica beforebeing coiled. The coiled gold electrode isrinsed between experiments.

All three electrodes, the reference, theauxiliary, and the working, are fine-tunedto fit into the cramped spectroelectrochem-ical cell. None of these electrodes are in theoptical path. The current path is not opti-mized inside the cell because the IR dropis not large in aqueous solutions.

The electrode storage cell (Fig. 6) is usedto store the reference and auxiliary elec-trodes anaerobically between experimentsto prevent oxygen from being absorbed bythe thirsty glass and then diffusing intothe solution during experiments. The elec-trodes are transferred from the storage cellto the degassed spectroelectrochemical cell

Fig. 5 Working electrode. Coiled gold wire

Copper lead


Stir bar

Port for WE

Port for RE

Port for AE

Fig. 6 Storage cell.

under positive argon pressure (see follow-ing text).

The storage cell contains MV++ inbuffer, which has been degassed by 10 minof argon bubbling. The electrodes areplaced in the cell under positive Argonpressure. The potential of the workingelectrode (another gold electrode in thestorage cell) is then controlled at a negativevalue (−550 mV vs SHE) and the MV++solution is reduced until it turns blue. Thisway the electrodes are stored in reduced

solution.

MV++ + e− ===⇒ MV+• (blue) (4)

In Fig. 7, the spectroelectrochemical celland stir bar are shown without electrodes,with ground glass plugs in the 7/25 and10/25 female joints designed to hold theelectrodes. The cell is configured as itwould be for anaerobiosis. The cell is madeof glass or quartz; and no graded seals thatwould increase the cell volume are used.The bottom part is a glass cuvette made

Glass-enclosedstir bar

Dithionitesyringe

Port for WE

Port for RE

Port for AE

Vacuum

Argon

Fig. 7 Spectroelectrochemicalcell without electrodes. Ports aresealed using ground glass plugsbefore anaerobic procedure.

15.4 Methods – Setting up the Mediated Electron Transfer Experiment 497

up of square glass tubing, onto which asuperstructure is attached: three femaleground glass joints or ports for electrodes,and a long glass sidearm (extension)topped by a stopcock and ground glassjoint for connection to the argon line.The ports are adjusted for fitting theelectrodes. There is a small glass sidearmnear the top of the extension (sidearm)that provides the option of adding otherreagents after the cell containing theenzyme solution has been made anaerobic.There is a glass-encased stirrer for mixingduring electrochemical reduction. Someresearchers add scrubbers to the sidearmor to the analyte solution. The cell isoptimized so that it will sit in thespectrophotometer to be used in theexperiments.

15.4Methods – Setting up the MediatedElectron Transfer Experiment

Two types of reductive titration experi-ments, namely, coulometric and potentio-metric, will be described. This is becausetwo kinds of experiments are requiredto completely characterize an enzyme,namely, measurement of n and of E

′

since mediators and other conditions dif-fer for the two measurements. Even so,the cell used and the anaerobic require-ments are the same: in both experiments abulk electrolytic reduction is performed ina spectrophotometric cell and the spectralchanges are monitored as a function of thenumber of reducing equivalents added.

15.4.1The Coulometric Titration

First, the enzyme solution is prepared byadding concentrated enzyme solution and

MV++ to buffer solution in the open aero-bic spectroelectrochemical cell. The bufferalso serves as a supporting electrolyte. Thestirrer is added. The volume of the en-zyme solution is typically 4.3 ml, whichis sufficient for immersing the electrodes(which are out of the optical path) andfilling the optical path. The enzyme con-centration is determined by measuring theabsorbance of the enzyme solution in theopen spectroelectrochemical cuvette. Theenzyme concentration required for the ex-periment is dictated by the absorbancechange desired, about 0.1–0.2 absorbanceunits during the entire coulometric titra-tion. This requires a solution of about10 µM for an enzyme that has a molar ab-sorptivity of about 10 000 M−1 cm−1. Theviologen concentration is about 100 µM.Oxidized MV++ does not absorb in thevisible region, so it does not interferewith the absorbance reading for the en-zyme. A buffer solution is chosen in whichthe enzyme will be stable, for example,50 mM phosphate at controlled pH. Addi-tional supporting electrolyte is not addedbecause many ions may bind to enzymeor interfere with the binding of other lig-ands, which are under investigation. Thecell temperature is maintained at a temper-ature at which the enzyme is most stable.Glycerol can be added to the buffer tostabilize protein, and if necessary, to pre-vent protein precipitation out of solution,since glycerol does not seem to inter-fere with electrochemical measurementsthemselves, although it can influence theslope of Nernst plots. The spectrum ofthe enzyme solution in the spectroelec-trochemical cell is taken before and afterdegassing in order to establish whethersolvent evaporation occurred during de-gassing, leading to a change in volumeand concentration.


After the spectrum is recorded, the cellis made anaerobic as follows:

1. An elaborate procedure is required fordegassing protein solutions, becauseordinary bubbling of argon throughthe solution causes bubbling andfoaming of the protein, leading to pro-tein denaturation. Since anaerobiosiscannot be induced with the electrodesin place, the electrode ports are sealedby inserting matched ground glassplugs to which Apiezon N grease hasbeen applied (Fig. 7). In order to makea protein solution anaerobic, the lowerpart of the spectroelectrochemical cellis immersed in an ice bath (to lowerthe vapor pressure of the solution andto help stabilize the enzyme). The cellis then attached to the argon line viathe Neoprene tubing, the stopcock atthe top is opened, and a vacuum isapplied. As the first bubbles start toform in the enzyme solution, the cell isslowly back-filled with argon. It is thenturned, and the solution is tipped intothe large sidearm extension of the cell

to maximize the surface area of the en-zyme solution. Next, the cell is gentlyrocked to promote gas exchange withthe argon of the atmosphere. The cellis rocked for about 5–10 min, reim-mersed in the ice bath, held upright,and a vacuum applied again. This cy-cle is repeated about 10 times over thecourse of 1.5 h.

2. When anaerobiosis is complete, thespectroelectrochemical cell is clampedin position under positive argon pres-sure.

3. The auxiliary and reference electrodesare removed from their anaerobic stor-age cell, the plugs are removed fromthe spectroelectrochemical cell, andthe electrodes are quickly inserted intothe spectroelectrochemical cell underpositive argon pressure (the referenceelectrode needs to be standardizedagainst ferri/ferrocyanide, either be-fore the start of the titration or at theend of the experiment, depending onthe experiment (see following text)).

4. The gold working electrode does notneed to be stored anaerobically, so it

Glass-enclosedstir bar

Dithionitesyringe

Ag/AgCl referenceelectrode

Gold workingelectrode

Auxiliaryelectrode

Vacuum

Argon

Fig. 8 Spectroelectrochemicalcell with electrodes.


is just greased and inserted into thespectroelectrochemical cell.

5. The spectroelectrochemical cell issealed by turning the stopcock at thetop of the cell, so that the cell isshut off from the argon line. The cellis then removed and placed in thespectrophotometer, which is equippedwith a stirrer and a temperaturecontrol apparatus. Figure 8 shows theassembled cell.

6. A spectrum is recorded; then appro-priate leads from the potentiostat areconnected to the reference, working,and auxiliary electrodes. The stirrer isturned on.Several good potentiostats with cur-rent integration and the ability tomeasure open circuit potentials areavailable. One is produced by BAS(West Lafayette, IN), the CV-50 W($12 000), and two by CH Instruments(Austin, TX), the CHI600A for $6000,and possibly the CHI1200 for $2500.

7. In order to perform electrochemicalreduction, the potential of the work-ing electrode is set and maintainedat a value at least 100 mV more neg-ative than that of the MV++ redoxpotential (−450 mV vs SHE), so thatthe viologen is reduced to produceMV+• radical, which is highly colored(ε at 600 nm is 12 000 M−1 cm−1).MV++ is reduced at the working elec-trode (as shown in Eq. 4) and transferselectrons quantitatively to the protein(flavoprotein, EFlox), which is now un-dergoing a two-electron transfer, asshown below:

2MV+• + EFlox + H+

= EFlred H− + 2MV++ (5)

The potential of the working elec-trode should not be controlled at

too negative a value for two reasons.First, a second electron will transfer toMV+•, forming MV, an insoluble ma-terial. Second, the hydrogen ions maybe reduced to hydrogen gas, a pro-cess that utilizes electrons, but not forthe process of interest. The reversiblepotential for hydrogen evolution is−420 mV, a value more positive thanour controlled potential value. The re-action is negligible at this potentialbecause the gold electrode has a smalloverpotential for hydrogen evolution,so hydrogen evolution does not occurat the thermodynamic value.

8. The current is monitored and in-tegrated, and the enzyme spectrummonitored at an appropriate wave-length as the reduction proceeds.Since MV+• quantitatively transferselectrons to enzymes with more posi-tive potentials and/or to oxygen in thesystem, the spectrum of reduced MVdoes not appear until the end point ofthe titration.

9. The number of coulombs, namely, theintegrated current (Q), expressed ascoulombs (C) required to completelyreduce the protein, must be calculatedso that the appropriate number ofpoints can be taken in the reductivetitration:

Q = nFM (6)

where M is the number of moles ofprotein in solution, n, the numberof electrons required to reduce onemole of enzyme, and F , the Faradayconstant. Because of the presence ofresidual oxygen, 3–5 mC of chargemay be consumed before the enzymespectrum starts to change. Oxygen isreduced, either by the reduced enzymeor by methyl viologen. In most cases,protein reduction does not occur until


after the oxygen has been reduced.After the reduction of the enzymestarts (denoted by spectral changes),the solution may be shaken to scrubthe oxygen out of the atmosphere tomake sure the oxygen is reduced.(Sometimes the oxygen is removedwith an internal scrubber.)

10. After the absorbance becomes stable,the spectrum is recorded, without stir-ring; the stirring is then restartedand a few millicoulombs of chargeare added to generate MV+•. Theabsorbance change at a selected wave-length is monitored. For example,for a flavoprotein at a concentrationof 10−5 M and a volume of 4.3 mL,8.3 mC of charge would be requiredfor the full two-electron reduction. Onthe basis of this number, one shouldcalculate the charge required for 20,40, 60, and 80% reduction. This in-cremental addition of reductant willprovide a titration with at least sixpoints. The solution should be stirredduring the electrochemical reduction.The current should be monitored andintegrated. After an appropriate num-ber of millicoulombs are added, forexample, at 20% reduction, the re-duction is stopped, and when theabsorbance at 450 nm stabilizes, stir-ring is stopped and the UV-visiblespectrum is recorded using a double-beam instrument; then the additionof reducing equivalents is continued.The end of the experiment is signaledwhen the spectrum of MV+• is ob-served, indicating titrant excess.

The absorbance at 450 nm is plotted as afunction of the number of millicoulombsor reduction equivalents (Eq. 6) added tothe sample. In the case of a simultaneoustwo-electron transfer to a flavoprotein, the

A450 versus n plot should be linear withan intersection on the x axis at n = 2. Thiswavelength (450 nm) is chosen because thegreatest difference in molar absorptivitybetween the oxidized and reduced form offlavin occurs at this wavelength.

Flavoprotein reduction can alternativelytake place in two single-electron steps,resulting in stabilization of a significantamount of the one-electron reduced form,EFl−• (‘‘red’’ anionic semiquinone) orEFlH• (‘‘blue’’ neutral semiquinone) dur-ing the coulometric reductive titration.These spectra are distinct from each otherand from the oxidized and two-electronreduced forms of the flavin. EFl– • hasdistinctive large absorbances at 530 and370 nm, whereas EFlH• has a broad ab-sorbance in the range 600–700 nm. Ifthese spectra are seen, the absorbancesA630, A530, or A370 can be plotted as afunction of n. If the potentials of the two-electron transfers are widely separated,that is, if the first electron is easier to trans-fer and is quantitatively transferred beforethe second, almost quantitative amounts ofsemiquinone are stabilized. Under theseconditions, the plots of A versus n arelinear and maximize at n = 1. From theextrapolated portions, the molar absorp-tivity of semiquinone at that wavelengthcan be determined. In this case – one re-action is proceeding in the first part of thetitration and another at the second – thereshould be two separate sets of isosbesticpoints in the first and second halves of thetitration. For an enzyme exhibiting suchbehavior, one would expect that in the po-tentiometric titration, each electron willbe transferred at a separate potential, andpotential-pH behavior determined for eachelectron separately.

In contrast, if less than 50% semiqui-none is stabilized, the A versus n plots canbe curved, whereas the spectra exhibit no


tight isosbestic points because all threeoxidation states are present during themajority of the titration. If the molarabsorptivity of the particular semiquinoneis known, the maximal amount at 50%titration can be determined. From theamount of semiquinone (M), the potentialseparation of the two electrons can becalculated

E′1 − E

′2 =

(4.606

RT

F

)log

[2M

1 − M

](7)

In many cases, the midpoint potentialis determined for the two-electron transferat half reduction (Em) via the potentio-metric titration described later, and thepotentials of the individual electrons arecalculated from that value and the percentsemiquinone. Thus, the individual E′

canbe determined in this manner by usingEqs. (7 and 8).

E′1 − E

′2 = 2Em (8)

If semiquinone is stabilized, we candetermine its protonation state by notingthe spectral properties of the semiquinone,and thus we can determine if the firstelectron is accompanied by a proton toform EFlH• as indicated below:

EFlox + e− + H+ = EFlH• (9)

Such predictions from spectral changesobserved on coulometric titrations mustbe verified by potential/pH studies usingpotentiometric titrations (described below)to determine if one or both electrontransfers are transferred with a proton andwhere pKas may occur.

In the case in which no semiquinoneis stabilized, the entire process is treatedas a two-electron transfer. Therefore, acoulometric titration can predict whichwill happen in the potentiometric titration.

The two are used together to define thebehavior of the enzymes.

Usually, the spectroelectrochemical celland anaerobic technique are tested beforeuse on an enzyme by reducing an indicatordye alone, making sure reduction is takingplace without oxygen leakage. The criteriaare that (1) the correct number of reducingequivalents be transferred and (2) thecorrect potential obtained for the dye, aswell as good Nernst plots, namely linearover a wide range and with correct slopes.Because the dyes can react directly withthe electrode, one experiment should beenough to provide all this information. Itis important to titrate the dyes under thesame conditions as for the enzyme.

15.4.2The Potentiometric Titration

The potentiometric titration is performedin a similar way, except that there is anadditional mediator or mediators present.The potentiometric titration gives optimalresults when performed in the presence oftwo kinds of mediators, a mediator titrantsuch as MV++ and a redox indicator orindicators. These redox indicators musthave redox potentials very near that of theprotein, because when in equilibrium withthe protein, the redox indicator is poisingthe potential of the indicator (workingelectrode).

The gold working electrode has tworoles in this experiment: first, with itspotential controlled at −0.550 mV, it is thecathode at which the MV++ is reduced.After transferring an aliquot of charge,which is transferred to the protein and themediator, the potentiostat is turned off,and the cell is held at open circuit in orderfor the contents of the cell and the goldelectrode to equilibrate at open circuit. Thepotential of the electrode and the spectrum


are monitored: when both stop changingand are constant for 10 min, equilibriumhas been achieved. After equilibrium hasbeen achieved, the electrode acts as anindicator electrode, sensing the potentialof the redox indicator couple, which isalso the potential of the enzyme and thepotential of the system, Ecell.

These concepts and procedures areshown in an example in which we assumeelectrons from MV+• are transferred first tothe protein (EFlox) (Eq. 10) and then to theredox indicator indigo disulfonate (IDSox).As the concentration of the two-electronreduced form of the enzyme EFlredH−increases, it equilibrates with the redoxindicator, for example, indigo disulfonate(IDS) (Eq. 11)

2MV+• + EFlox + H+

= EFlredH− + 2MV++ (10)

EFlredH− + IDSox

= 2EFlox + IDSred + H+ (11)

The titrant of Eq. (10) is chosen so thatthe position of equilibrium of this reactionis far to the right, whereas for Eq. (11),the mediator is chosen so that there areapproximately (within a factor of 10) equalamounts of the two forms of the mediator,IDSox and IDSred.

At each point in the titration, thesystem is allowed to equilibrate. Thenboth the potential and the spectrumof the system are recorded. This givesthe measured potential in the followingEq. (12), which represents the equilibriumbetween the enzyme and the dye, which iscommunicated to the electrode by the dye.

Equilibration times range from 10 to30 min, and are considered reached whenthe potential does not drift more than amillivolt in 10 min. The potential is thenmeasured and the spectrum taken. Several

more aliquots of reducing equivalents aretransferred, further reducing the enzyme.A Nernst plot covering as broad a rangeas possible is constructed from the pointsof the reductive titration. The range of theNernst plot may be limited by the potentialrange of the mediator dye.

The concentrations of all species arecalculated from the spectra and the molarabsorptivities. E

′and n are calculated

from the Nernst plot of the E measuredversus log [ox/red] for the enzyme. Thedyes must be titrated under the sameconditions as the enzyme, in order toobtain the spectrum of the enzyme alone.

A disadvantage of this method is thatit depends on good equilibration, mak-ing controls necessary. For example, tominimize the effect of dye binding to theenzyme, the indicator dyes are present insmall concentrations (about 3 µM) to pre-vent binding to the enzyme (10 µM). Inorder to remove possible effects from dyeinteraction, the potential of the enzyme isdetermined in two separate experimentsusing two different indicator dyes.

The slope and intercept of the Nernstplot are evaluated. It should be 29 or59 mV at room temperature, and a largerange of potentials should be covered.If the slope deviates from these values,this may indicate incomplete equilibration.The experiment can then be rerun usingdifferent dyes to alleviate the problem.If the Nernst plot has a good slope,the reaction is judged to be reversible.As a further test of the reversibility,some reoxidative points are obtained ina reductive titration. For example, if 3 mCof reductive equivalents were transferredto obtain a point in the titration, theenzyme will then be reoxidized by 0.5 mC.The potential and spectrum are measuredto make sure the data obtained from

15.5 Data Analysis and Sample Systems 503

reoxidation fits on the same Nernst plotas data obtained in the reductive direction.

The mediator dyes are the limitingfactor in making the redox potentialmeasurements. The dyes chosen musthave potentials very near the potentialof the enzyme; they must also be ableto interact with the active site of theenzyme quickly. One problem is thatthere are few good indicators in thepotential range between 0 and −150 mVat pH 7.0, which is the redox potentialregion for many flavoproteins, so onlypart of the potential region for theenzyme studied may be covered by thedyes available. Because of the lack ofgood redox indicators in the potentialregion of interest, it may not be possibleto repeat the potentiometric experimentwith dyes with different structures. Ifthis is done and the same potential isobtained, this is a clear indication that themeasurement is valid and independent ofdye properties.

15.5Data Analysis and Sample Systems

15.5.1Coulometric Titration

Several kinds of information can beacquired from this experiment such asthe total number of electrons transferredto the enzyme, the confirmation that aspectral intermediate is present along withits quantitation, and finally, an indicationof whether the process occurs by two singleelectron transfers or a simultaneous two-electron transfer. A plot of absorbanceat selected wavelengths versus numberof electrons transferred can be used toanalyze the data. Current efficiency iscalculated by comparing the number of

electrons required to reduce the enzyme,once reduction starts, to the theoreticalnumber of reducing equivalents (seeearlier text).

Such results are shown in Fig. 9for a flavoprotein (MCAD) bound tohexadienoyl-CoA (HD-CoA). HD-CoA wasdesigned to be both a Raman active probeand a thermodynamically stable productanalog. It was desired to verify that thisproduct analog did not accept electronsfrom the reduced enzyme, which woulddestroy its Raman active properties, mak-ing it unsuitable to probe the properties ofthe reduced enzyme. The best way to showthat no significant reduction was occurringto HD-CoA was to perform a coulometricreductive titration of the MCAD saturatedwith HD-CoA and confirm that the en-zyme was indeed 100% reduced before theanalog started to accept electrons. It wasestablished that the reducing equivalentswere transferred to the enzyme first by si-multaneously monitoring the number ofreducing equivalents transferred to the en-zyme–HD-CoA complex and the spectralchanges occurring at the enzyme.

The coulometric titration shown inFig. 9(a) was performed at an HD-CoAconcentration such that 95% of the MCADwas complexed. The titration consisted oftwo reductive events. The absorbance at456 nm decreased sharply (A = −0.20)during the first part of the titration be-cause the reduction of flavin (see inset). Anincrease in absorbance at 570 nm accom-panied flavin reduction, consistent withthe formation of a charge transfer complexbetween the dienoyl-CoA ligand and the re-duced enzyme. The absorbance at 570 nmreached a maximum between n = 2.6 and3 reducing equivalents, with the majorityof the 456 nm decrease in absorbance oc-curring during this part of the titration.From n = 3 to 11.8, the absorbance of


0.00

0.10

0.20

0.30

325 425

(a)

(b)

525 625 725 825

Wavelength[nm]

Abs

orba

nce

0.00

0.05

0.10

0.15

0.20

0.25

0.0 3.0 6.0 9.0 12.0n

Abs

orba

nce

[456

nm

]

0.00

0.01

0.02

0.03

0.04

Abs

orba

nce

[570

nm

]

1

7

0

6

12

18

24

0 1 2 3 4 5 6

Moles of HD-CoA permole of enzyme

n (T

otal

)

4e−

2e−

Fig. 9 (a) Coulometric reduction of MCAD (14.7 µM) in the presence of HD-CoA (80 µM)and MV++ (100 µM). Titrations were performed under anaerobic conditions at 25 C in50 mM potassium phosphate buffer, pH 7.6. Only selected spectra are shown for clarity.Curves 1–7: n = 0.0, 1.0, 1.8, 2.6 (bold), 4.7, 8.2, and 11.8, respectively. Inset: plot ofabsorbance at 456 nm () and 570 nm () as a function of the number of reducingequivalents (n) added during the titration and (b) correlation between the total number ofreducing equivalents (n) and the amount of HD-CoA in solution. Datum at 0 µM HD-CoA (•)was obtained from Ref. [1]. All other data (©) were obtained from coulometric titrations inthis work. Concentrations of HD-CoA were varied from 14.3 to 80 µM. Enzyme concentrationswere kept between 13.8 and 14.8 µM. Theoretical fits are shown for two ( ) and four(- - - - ) electron reduction of HD-CoA.

the charge transfer band decreased. Af-ter 11.8 reducing equivalents had beentransferred, reduced methyl viologen wasthermodynamically stabilized.

Since more than two reducing equiv-alents could be added to the system, itbecame apparent that both the flavin andthe HD-CoA were undergoing reduction.


To determine if one or both double bondsof HD were being reduced, coulometricreductions of HD-CoA were performed atseveral different HD-CoA concentrations.Figure 9(b) clearly indicates that the reduc-tion of HD-CoA is a two-electron process.

It is clear that 85% the MCAD reducesbefore the HD-CoA, so that the redoxpotential of the MCAD bound to HD-CoA can be measured. Thus, it should bepossible to correlate the Raman signal withthe redox potential of the ligand bound tothe enzyme.

15.5.2Potentiometric Titration

The measured potential, relative to thereference electrode employed (Ecell), rep-resents the equilibrium position of theenzyme and the dye, and is communicatedto the electrode by the dye:

Ecell = E′dye + 0.059 log

[ ox

red

]= E

′enzyme + 0.059 log

[ ox

red

](12)

The concentrations of all species arecalculated from the absorbance spectraand the molar absorptivities. E

′and n

are calculated from the Nernst plot of theEcell versus log [ox/red] for the enzyme.The dyes must be titrated under the sameconditions as the enzyme so that theirspectra can be subtracted and the spectrumof only enzyme obtained.

Potentiometric titrations can be usedto determine three things in addition tobasic protein characterization: first, theenzyme is being bound and regulatedby regulatory proteins or substrates; andsecond, protons are being transferredalong with the electrons, and the ionizationof acidic groups is occurring in the pHregion being studied. The case of ligands

that bind to the protein causing a changein its potential will be considered first.However, let us look back at the illustrationshown in Fig.1, in which the potentialof the redox-active cofactor, in this casethe flavin undergoes a change in redoxpotential (E′

) upon binding to apoprotein(ligand). In this case, the redox potential ofthe flavin at the active site of the enzyme(MCAD) shifts positive because of the ratioof the binding constants (shown in theinset). The new potential is −145 mV. Nowwe will see that the binding of a ligand (L),substrate or regulatory protein, to MCADcan cause the redox potential of the MCADenzyme (or the flavin at the active site)to change again. These redox potentialshifts present the clearest evidence thatbinding is occurring. The example usedfor illustration in Fig. 10 is of HD-CoA, aproduct analog, L, binding to the enzymeMCAD, designated E. It is assumed thata single ligand binds to a single flavinactive site, and that if there is more thanone active site per protein molecule, theyfunction independently. This concept ofbinding coupled to electron transfer isrepresented in the following series ofequations, which can be incorporated intoa thermodynamic box in which Eox =free oxidized enzyme, Ered = free reducedenzyme, Eox• L = oxidized enzyme boundto ligand, Ered• L = reduced enzyme boundto ligand

Eox + 2e− = Ered E′1 = −145 mV

(13)

Eox•L + 2e− = Ered•L E′2 = −52 mV

(14)

These values are related by (1) thebinding constant of the oxidized enzymeto the ligand, and (2) the binding of thereduced enzyme the reduced enzyme tothe ligand:


−85

−65

−45

−25

−0.8 −0.4 0.0 0.4 0.8

log(ox/red)

Mea

sure

d po

tent

ial

[mV

]

0.00

0.05

0.10

0.15

0.20

325 425 525 625 725 825

Wavelength[nm]

Abs

orba

nce

1

7

Fig. 10 Potentiometric titration of MCAD (10.7 µM) in the presence of HD-CoA(181 µM). MV++ (100 µM) serves as a redox mediator. Redox indicators arepyocyanine (5.0 µM) and indigo disulfonate (2.5 µM). Titration was performed underanaerobic conditions at 25 C in 50 mM potassium phosphate buffer, pH 7.6.Intermediate spectra have been removed for clarity. The spectrum of the oxidizedMCAD•HD-CoA complex is shown by curve 1. Curves 2–6 show the complex at E = −30,−44, −51, −62, and −79 mV versus SHE. Curve 7 is the spectrum of the fully reducedMCAD bound to HD-CoA. Inset: Nernst plot indicating an E′ = −0.052 V and n = 1.6.

L + Eox = L•Eox (15)

Kox = [L•Eox]

[L][Eox](16)

L + Ered = L•Ered (17)

Kred = [L•Ered]

[L][Ered](18)

Kdox = 1

Kox(19)

Kred = 1

Kd red(20)

By rearranging these equations andsubstituting into the Nernst equation forthe free enzyme, we can calculate thepotential of the bound enzyme:

E = E′free − 0.059

nlog

(Ered

Eox

)(21)

Rearranging the binding equation:

[Eox] = [L•Eox]

Kox[L](22)

[Ered] = [L•Ered]

Kred[L](23)

and substituting these values into theNernst equation, we obtain for 1 M ligand,and the unit activity of other species:

E′bound = E

′free + 2.303

RT

nFlog

Kred

Kox

(24)

So the magnitude of the potential shifton binding is related to the ratio of thebinding constants. We could modify itusing the dissociation constants. Likewise,if we know any three of the thermodynamic


constants, we can calculate the fourth. Wecan also represent this relationship as athermodynamic box, as for Fig. 1.

The redox potential of HD-CoA-comp-lexed MCAD was determined at severalpH values. A representative potentio-metric titration at pH 7.6 is shown inFig. 10. The midpoint potential is for theoxidized FAD/hydroquinone couple. AnE

′of −52 mV (inset, Fig. 10) for the

MCAD•HD-CoA complex is 93 mV morepositive than the midpoint potential of thefree MCAD (−145 mV) [1], indicating asignificant potential shift in the enzymeupon binding HD-CoA. Equation (24) canbe used to calculate a dissociation constantfor the analog and reduced enzyme (Kdred)of 2.5, a 1400-fold difference between Kdoxand Kdred. Tighter binding of product tothe reduced enzyme is consistent with pre-vious work on acyl-CoA dehydrogenases(ACDs).

Kdred = 10−[

(Em(bound)−Em(free))nF

2.303RT

]× Kdox

(25)

We can detect whether electron transferis coupled to the proton transfer reactionby measuring the redox potential overas wide a pH range as enzyme stabilityallows. If E

′is independent of pH, this

indicates that there is no proton transfercoupled to the electron transfer reaction,or that the simultaneous ionization ofthe oxidized and reduced forms offsetone another (see below). However, ifproton transfer is coupled to electrontransfer for a particular redox state, theelectron transfer will be dependent onpH. For enzymes, proton transfer usuallyaccompanies electron transfer in order tomaintain charge neutrality at the activesite. This is illustrated in the simplecase of oxidized free flavin, abbreviated(Flox). In order to understand how pKa

values can be determined from redoxpotential values determined over a broadrange of pH, consider the relatively simpleexample of the following half-reaction thatis true for free flavin (at low pH) and theaccompanying Nernst equation.

Flox + 2e− + 2H+ −−−−−− FlredH2 (26)

E′ = E

− 0.05916

2log

[FlredH2]

[Flox][H+]2

(27)

FlredH2 is an acid that can dissociate.We need to express the Nernst equation insuch a way that the log term contains theformal concentration of Flox and FlredH2.Flox is not an acid or a base, so itsformal concentration (F ) equals its molarconcentration:

FFlox = [Flox] (28)

For the diprotic acid FlredH2, we usethe following equation to express [FlredH2]in terms of the formal concentrationof FlredH2. To do this, the fractionalcomposition equation or (α equation) isused to convert the form of an acid or baseto its concentration in a particular form:

[FlredH2] = αFlredH2•FFlredH2

= [H+]2FFlredH2

[H+]2 + [H+]Ka1 + Ka1Ka2

(29)

substituting Eqs. (28 and 29) into Eq. (27)we obtain

E′ = E

− 0.05916

2

× log[H+]2FFlredH2

[H+]2 + [H+]Ka1 + Ka1Ka2

FFlox [H+]2

(30)


that can be rearranged to form

E = E − 0.05916

2

× log1

[H+]2 + [H+]Ka1 + Ka1Ka2︸︷︷︸E′

− 0.05916

2log

FFlredH2

FFlox

(31)

This first term is the formal potential. Inthis case, the potential 1/pH curve wouldhave three line segments as the pH, whichis initially below pKa1, is raised to valueshigher than all pKas. In the first case, inwhich pH is lower than pKa1 the reactionthat involves a two-electron, two-protontransfer is

Flox + 2e− + 2H+ −−−−−− FlredH2 (32)

The Nernst equation is

E = E − 0.05916

nlog

[FlredH2]

[Flox] [H+]2(33)

If the pH-dependent term is factoredout, we can see the potential/pH slope =59 mV, and thus this pH dependencebecomes a part of E

′:

E = E − 0.05916 log[H+] − 0.05916

n

× log[FlredH2]

[Flox](34)

In the second case, in which the pHvalues are between the two pKa values, orpKa1 < pH < pKa2 , the reaction becomesa two-electron, one-proton transfer:

Flox + 2e− + H+ −−−−−− FlredH− (35)

The Nernst equation becomes:

E = E − 0.05916

nlog

[FlredH]

[Flox] [H+](36)

If we separate out the pH dependence,we see that the pH-dependent term has a0.059/2 factor, giving a 29-mV slope forthis part of the pH range.

At pH < pKa2 , there is no pH depen-dence since no protons are transferredwith the electrons. The pH at which thepotential/pH slope changes correspondsto pKas for the reduced form.

Obviously if the oxidized flavin is boundto a protein, two things may occur, first, itspKas may be shifted from the free form,so that the redox data become influencedby the pKas of nearby acid residues. Forexample, there may be a positively chargedamino acid side chain that can hydrogenbond to the N(1) position of the flavin. Thepresence of the positive charge near theflavin may make electron transfer easier,thus causing the redox potential of theflavin to be shifted positive. This positivecharge near the flavin may cause the pKasof the flavin group to be shifted so that overmuch of the pH range, only one protonwould need to be transferred to maintaincharge neutrality. Second, the pKas ofthese amino acid groups of the proteinpart of the enzyme may be influencedby the redox state of the flavin, namely,the pKa of the same group could change.Therefore, for the flavoproteins, there maybe as many as three or four pKas thatcan influence redox behavior, typically twofrom the reduced flavin and two from theprotein. By measuring the redox potentialsof the enzyme over a wide range of pHvalues, we may be able to discover thepresence of those ionizable amino acidside chains that control the redox potentialof the bound flavin and determine thepKa values of those ionizable groups. Inthe more complex cases, the potential/pHbehavior must be modeled by fitting thedata to a redox-linked ionization model


described by Clark [6] as in Eq. (37):

E′ = E0 + 2.303

RT

nF

× log

[H+]3+[H+

]2Kred,1

+[H+]Kred,1Kred,2+Kred,1Kred,2Kred,3

[H+]3+[H+

]2Kox,1

+[H+]Kox,1Kox,2+Kox,1Kox,2Kox,3

(37)

where E0 is an extrapolated value becausethe potential at pH = 0 cannot be obtainedexperimentally. In the data whose pKas areestablished by curve fitting, E

′values at

a particular pH are determined from theintersection of the Nernst plot in whichby definition the log[(ox)/(red)] term isequal to zero. In this case, three pKas wereconsidered and used to fit the experimentalE

′data taken at various pH values.

It is important to point out, as Eq. (37)illustrates, the principle that the redox-linked ionizations will not be detected ifKox,1 = Kred,1 and so on, because then thesecond term of the equation will be zero.

References

1. J. D. Pellett, K. M. Sabaj, A. W. Stephenset al., Biochemistry 2000, 39, 13 982–13 992.

2. G. J. Mancini-Samuelson, V. Kieweg, K. M.Sabaj et al., Biochemistry 1998, 37,14 605–14 612.

3. L. H. Bradley, R. P. Swenson, Biochemistry1999, 38, 12 377–12 386.

4. H. A. Heering, J. Hirst, F. A. Armstrong, J.Phys. Chem. B 1998, 102, 6889–6902.

5. J. N. Butt, S. E. J. Fawcett, J. Breton et al., J.Am. Chem. Soc. 1997, 119, 9729–9737.

6. W. M. Clark, Oxidation-Reduction Potentialsof Organic Systems, Williams & Wilkins,Baltimore, Md., 1960.

7. D. B. Burleigh Jr., G. P. Foust, C. H. Will-iams Jr., Anal. Biochem. 1969, 27, 536–544.

8. G. P. Foust, D. B. Burleigh Jr., S. G. Mayhewet al., Anal. Biochem. 1969, 27, 530–535.

9. S. G. Mayhew, Methods in molecular biologyin Flavoprotein Protocols (Eds.: S. K. Chap-man, G. A. Reid), Humana Press, Totowa,N.J., 1999, Vol. 131.

10. V. Massey; P. Hemmerich, Biochemistry1978, 17, 9–16.

11. V. Massey in Flavins and Flavoproteins, 1990(Eds.: B. Curti, S. Ronchi, G. Zanetti), Walterde Gruyther, Berlin, 1991, pp. 59–66.

12. M. T. Stankovich, Anal. Biochem. 1980, 109,295–308.

13. R. Szentirmay, P. Yeh, T. Kuwana in Electro-chemical Studies of Biological Systems, ACSSymposium Series 38 (Ed.: D. T. Sawyer),American Chemical Society, Washington,DC, 1977, pp. 143–169.

14. G. S. Wilson, Methods in Enzymology, Aca-demic Press, New York, 1978, Vol. 54,pp. 396–410.

15. V. Massey, F. Muller, R. Feldberg et al., J.Biol. Chem. 1969, 244, 3999–4006.

16. G. Fasman, (Ed.), CRC Handbook of Bio-chemistry and Molecular Biology, Physical andChemical Data Vol. 1, 3rd ed., 1985, Vol. 1,pp. 122–130.

17. J. E. O’Reilly, Biochim. Biophys. Acta 1973,292, 509–515.

18. J. D. Pellett, D. F. Becker, A. K. Saengeret al., Biochemistry 2001, 40, 7720–7728.

511

16Membrane Electrochemistry

H. Ti Tien and Angelica OttovaPhysiology Department, Michigan State University, East Lansing, MI, USA

16.1 Introductory Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 513

16.2 Types of Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51316.2.1 Semipermeable Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . 51316.2.2 Ion-exchange Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51416.2.3 Microporous Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 514

16.3 The Lipid Bilayer Principle of Biomembranes . . . . . . . . . . . . . . . . 514

16.4 Classification of Biomembranes . . . . . . . . . . . . . . . . . . . . . . . . . 515

16.5 Composition, Structure and Function of Biomembranes . . . . . . . . 51716.5.1 Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51716.5.2 Structure and Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51716.5.3 The Ultrastructure of Biomembranes . . . . . . . . . . . . . . . . . . . . . 518

16.6 The Function of Biomembranes Pertaining to Electrochemistry . . . . 51816.6.1 Basics of Membrane Electrochemistry . . . . . . . . . . . . . . . . . . . . . 51916.6.1.1 Nernst–Planck Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51916.6.1.2 Thermodynamic Considerations . . . . . . . . . . . . . . . . . . . . . . . . . 52016.6.1.3 Nernst Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52316.6.1.4 Electroneutrality Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52316.6.1.5 Ion Selectivity, Ion Specificity, pH, and Cell Membranes . . . . . . . . 52316.6.1.6 The Henderson–Hasselbalch Equation . . . . . . . . . . . . . . . . . . . . 524

Ion-selective membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 525The Nicolsky equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 525

16.6.1.7 Origins of Membrane Potentials . . . . . . . . . . . . . . . . . . . . . . . . . 525

512 16 Membrane Electrochemistry

16.7 The Role of Electrical Double Layers (EDLs) and their BiologicalImplications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526

16.7.1 Electrokinetic Potential, Uk . . . . . . . . . . . . . . . . . . . . . . . . . . . . 527The Dorn effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 528Streaming potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 528Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 528Electroosmosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 528

16.7.2 Adsorption Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52916.7.3 Distribution Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52916.7.4 Diffusion Potentials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52916.7.5 The Gibbs–Donnan Potential . . . . . . . . . . . . . . . . . . . . . . . . . . 53016.7.6 Phase Boundary Potentials . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53216.7.7 Relationship among Various Potentials . . . . . . . . . . . . . . . . . . . . 533

16.8 Experimental Membrane Systems . . . . . . . . . . . . . . . . . . . . . . . 53516.8.1 Langmuir–Blodgett Films . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53516.8.2 Immiscible Liquid–Liquid (L–L) Interfaces (ITIES) . . . . . . . . . . . . 53516.8.3 Planar Lipid Bilayers and Liposomes . . . . . . . . . . . . . . . . . . . . . . 536

16.9 Bilayer Lipid Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53616.9.1 Formation Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53616.9.1.1 Conventional BLM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53616.9.1.2 BLMs via L–B Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53716.9.1.3 Supported BLMs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53716.9.2 Methods of Investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54016.9.2.1 Potentiometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54016.9.2.2 The Patch-Clamp Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . 54116.9.2.3 Cyclic Voltammetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54216.9.2.4 CV and BLM Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54216.9.2.5 Modeling of BLM Behavior by Electrochemical Equivalent Circuit . . 54416.9.3 Electrochemical Impedance Spectroscopy . . . . . . . . . . . . . . . . . . 54516.9.4 Photoelectrospectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 548

16.10 Electronic and Charge Transfer Processes in Membranes . . . . . . . . 55016.10.1 Properties of Iodine–Iodide-containing BLMs . . . . . . . . . . . . . . . 55016.10.2 Polypyrrole BLMs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 551

16.11 Photoelectric Effects in Membranes . . . . . . . . . . . . . . . . . . . . . . 551

16.12 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 552References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 556

513

16.1Introductory Remarks

There is at least one similarity betweenelectrochemistry and biological mem-branes (biomembranes); they are bothinterface science. Electrochemistry is awell-established discipline dealing withelectrical phenomena at, as well as across,the interfaces. The study of biomem-branes, on the other hand, is a relativelynew endeavor, whose investigations in-volve the use of powerful physicochemi-cal techniques based on electrochemistry.Hence, membrane electrochemistry maybe defined as the application of electro-chemistry to membrane studies, includingboth artificial and natural systems. Definedin another way, Membrane Electrochemistry,the title of this chapter, is a synonymto the origin of electrophysiology. Thisbrings us immediately to the often-citedstory about how Galvani observed thetwitching of a frog’s leg upon touchingby two dissimilar metals. Thus, the elec-trical properties of membranes, naturaland synthetic, have been investigated sinceGalvani and Volta’s time (1737–1798 and1745–1827, respectively). There are manydifferent types of membranes: semiperme-able, ion exchange, microporous, synthet-ics of all kinds, biomembranes of all sorts,and reconstituted biomembranes (bilayer

lipid membranes or BLMs). Among theunique features of these membranes aretheir selectivity and specificity; they limitunrestricted hydrodynamic flow. Today,membrane electrochemistry is a maturefield of endeavor. Among the aims ofmembrane electrochemistry are investi-gating charge generation, separation, andtransport in natural and model membranesystems at the molecular level, and us-ing the relevant knowledge thus gainedfor developing practical devices such assensors for chemical, biomedical, and en-vironmental applications.

16.2Types of Membranes

16.2.1Semipermeable Membranes

Generally speaking, the bilayer lipid mem-brane (BLM or planar lipid bilayer)is semipermeable, meaning that somemolecules are allowed to pass freely (dif-fuse) through the structure. The lipidbilayer is virtually impermeable to largemolecules, relatively impermeable to smallions and polar molecules. Specifically, os-mosis, a diffusion phenomenon, can beeasily observed in a device called osmome-ter in which a membrane separating two


aqueous chambers is permeable only towater. When solutions of different con-centration are present, water will flow fromthe dilute chamber into the concentratedone until equilibrium is reached, at whichthe extra pressure known as the osmoticpressure is produced. Van’t Hoff foundempirically that the osmotic pressure isdirectly proportional to the concentrationof solute for dilute solutions. In this con-nection, Pfeffer immersed plant cells ofChara in salt solutions of various con-centrations and observed that the cellsswelled in hypotonic (low salt) solutionsand shrank in hypertonic (high salt) so-lutions. To account for similarities in theosmotic behavior of plant cells and man-made osmometers, Pfeffer, recognizingthat the boundary of discontinuity betweenthe protoplasma and its environment mustconstitute an osmotically semipermeablemembrane, postulated the existence of aninvisible (under light microscope) plasmamembrane in the plant cell.

16.2.2Ion-exchange Membranes

Synthetic polymers containing fixed chargegroups (−COO−, −SO4

−, −PO4−2, etc.)

in the form of thin sheets are knownas ion-exchange membranes. The equi-librium selectivity coefficients for alkalications for polymers containing stronglyacidic groups follow the Hofmeister series(i.e. Li, Na, K, Rb, Cs), whereas the orderis reversed for ion-exchange membraneshaving weakly acidic groups [1]. The elec-trical potential difference between twosolutions separated by an ion-exchangemembrane in the absence of any cur-rent flowing through the membrane canbe measured. Carboxylated ion-exchangemembranes have been used as a substratefor ion sensors.

16.2.3Microporous Membranes

Polycarbonate membranes, known com-mercially as microporous Nucleopore

filters, are outstanding examples. Also in-cluded in this type of membranes arecellulose filters. With the former, they con-tain pores of known density and diameter.A novel BLM system with a longer lifetime,a larger surface area, and a greater stabilityin the face of chemical and mechanicaldisturbances has been reported [1]. Thepolycarbonate-BLM system is supportedin such a manner that an aqueous solu-tion could be easily added to both sides. Inan ideal situation, this Nucleopore-coatedmembrane may be visualized as tens ofthousands of micro-BLMs simultaneouslygenerated in situ.

16.3The Lipid Bilayer Principleof Biomembranes

All living beings are made of cells, fromsimple single-celled bacteria to complexmulticelled humans. The conclusion thatthe fundamental structure of biomem-branes is a lipid bilayer is based on threepivotal experimental findings: firstly, theelegant and simple experiment of estab-lishing the orientation of amphipathicmolecules at interfaces by Langmuir in1917. Secondly, using that method Gorterand Grendel in 1925 reported that the ex-tracted lipid molecules from the plasmamembrane of red blood cells (RBCs) oc-cupied the area on the surface of aLangmuir trough that was twice that of theoriginal membrane. Thirdly, the lipid bi-layer concept, as deduced from the above,was dramatically substantiated in 1961 by

16.4 Classification of Biomembranes 515

reconstituting a BLM from the lipids ex-tracted from the white matter of cow’sbrain [1, 2]. Electrical activities of the nerveand light-induced effects in photosynthe-sis have been elucidated with the aid ofexperimental planar BLMs and liposomes.In nerves, translocation of ions across pro-tein channels embedded in the lipid bilayerplay the pivotal role. In photosynthesis,light absorption by pigments, confinedin the lipid bilayer, initiates electroniccharge generation and separation, leadingeventually to redox reactions on oppositesides of the membrane. Presently, electronand charge transfer processes through thelipid bilayer are being actively investigated,from both theoretical and biotechnologicalviewpoints. The crucial role played by thelipid bilayer may be summarized as fol-lows. Living organisms are made of cellsbound by their membranes. They are self-assembling entities; each is organizing aparticular combination of phospholipidsin the form of a bilayer with other con-stituents (e.g. proteins) embedded in it.This lipid bilayer, existing in all biomem-branes, is most unique; it serves not merelyas a physical barrier but functions as a two-dimensional matrix for reactions. Also, thelipid bilayer acts as a conduit for ion trans-port, as a framework for antigen–antibodybinding, as a bipolar electrode for redoxreactions, as a reactor for energy conver-sion (e.g. light to electric to chemical).Further, a modified lipid bilayer performsas a transducer for signal transduction (i.e.sensing), and numerous other functionsas well. All these myriad activities requirethe ultrathin lipid bilayer. To study BLMsin detail, the task has been a dauntingone until a few years ago because a 5-nmBLM is an extremely labile structure withlimited lifetime. Planar BLMs can now beformed on various substrates with long-term stability, thereby opening the way for

basic research and development work inthe domain such as biotechnology, catal-ysis, electrochemistry, microelectronics,and membrane biophysics. Today, fromall lines of experimental findings, there islittle doubt that all biomembranes possessa lipid bilayer structure, thereby underly-ing the lipid bilayer principle of biomem-branes. The self-assembled lipid bilayer,the crucial component of most, if not allbiomembranes, is in a liquid-crystallineand dynamic state. A functional cell mem-brane system, based on self-assembledlipid bilayers, proteins, carbohydrates, andtheir complexes, should be considered inmolecular and electronic terms; it is ca-pable of supporting ion or/and electrontransport, and is the site of cellular ac-tivities in that it functions as a ‘‘device’’for either energy conversion or signaltransduction. Such a system, as we knowintuitively, must act as some sort of atransducer capable of gathering informa-tion, processing it, and then delivering aresponse based on this information. Allthese myriad activities are predicated onthe BLM of nanometer thickness, whoseelectrochemistry is the focus of this chap-ter (Fig. 1).

16.4Classification of Biomembranes

Broadly, biomembranes may be dividedinto two types: plant and animal. Plant cellmembranes generally contain little choles-terol. In terms of cells and organelles, thefollowing classification is usually acceptedaccording to their particular functions:

• the plasma membrane (e.g. red bloodcells);

• the chloroplast (e.g. thylakoid mem-branes);


Partition

WaterWater Water

P-G border

P-G border

Teflon coating Teflon tubing

s-BLM(b)

(a) A B C

Metal(e.g., Pt, stainless steel)

Hydrogel(e.g., agar, agarose)

sb-BLM

Water

Fig. 1 Two basic types of BLMs (planar lipid bilayers). (a) A conventional BLM separatingtwo aqueous solutions in formation [1] (a short movie illustrating the BLM-formingprocess) may be seen by visiting the URL: http://www.msu.edu/user/ottova/soap bubble.html; (b) supported BLMs; (c) a BLM on metal substrate (s-BLM); (d) asalt-bridge supported BLM (sb-BLM). P-G stands for Plateau-Gibbs border that supportsthe BLM [2, 3].

16.5 Composition, Structure and Function of Biomembranes 517

• the mitochondrion (e.g. cristae mem-branes);

• the nerve (e.g. axon membranes);• the eye (e.g. the visual receptor mem-

branes);• other membrane systems (muscles,

microtubules, cytoskeletal elements,etc.).

From the viewpoint of membrane electro-chemistry, some of these membranes havebeen reconstituted and characterized elec-trochemically via experimental BLMs (orplanar lipid bilayer). Monographs and re-views covering many aspects of membraneresearch are available [1–28].

16.5Composition, Structure and Functionof Biomembranes

16.5.1Composition

Overall, a cell membrane is a lipid bi-layer matrix modified by functional pro-teins, carbohydrates, and their complexes.Typically, cells are about 90% water. The

rest, on the dry weight basis, is ap-proximately: 50% protein, 15% each forcarbohydrate and nucleic acid, 10% lipid,and 10% miscellaneous compounds. Morespecifically, Table 1 presents the grosscomposition of biomembranes, along withtheir functions.

16.5.2Structure and Function

The general ultrastructure of the biomem-brane is that the lipids are in the formof a bilayer providing the framework forembedding proteins, carbohydrates, andother constituents. These membrane con-stituents are immersed to a varying degreein the lipid bilayer. The lipid bilayer ofthe membrane is fluidlike and has theconsistency of olive oil (viscosity ∼1 cP).The lipids, proteins, and other constituentsare thus perceived to be able to extendfreely within the confine of the lipid bi-layer. However, the picture is a dynamicone, in that phospholipids and cholesterolform a hydrophobic, fluid bilayer in whichfunctional entities such as receptors, ionchannels, pigments, proteins, and so on

Tab. 1 Gross composition and function of some basic types of membranesa

Lipids Proteins Carbohydrates Function

Plasma membrane 25–50 50–70 10 Diffusion barrier activetransport

(Gram + bacteria) 75 25 10 Antigenic propertiesCristae (mitochondrion) 20–40 60–80 2–4 Energy transduction ATP

synthesisThylakoid (chloroplast) 40 35–65 6 Sunlight conversion ATP

8% pigments synthesisVisual (visual receptor) 40 50 4 Light detection signal

4–10% pigments transductionNerve (axon) 80 20 3 Conduction of nervous

impulse

a(Expressed in % of dry weight, taken from p. 38 [3].)


could be embedded. That is, membraneconstituents are able to exchange and/orbe modified by compounds in contactwith the cell membrane. At the molecu-lar level, both lipids and proteins exhibitasymmetry; the composition of the in-side of a membrane is different fromthe outside. Since membrane function andstructure are the ‘‘two sides’’ of the samelipid bilayer, this must be so in orderto explain the active transport of speciesacross the membrane. Also, active sites ofmembrane-bound enzymes or immuno-logical determinants are found only onone side of a lipid bilayer.

16.5.3The Ultrastructure of Biomembranes

A brief account of the molecular struc-ture of biomembranes may be presentedfrom a chronological viewpoint. The recog-nition of the lipid bilayer as a principalelement for all biomembranes dates backonly to the second quarter of twentiethcentury. The origin of the lipid bilayer con-cept, however, is much older, which istraceable to more than three centuries! Itall began with the physicist and inventorRobert Hooke of Hooke’s Law fame, whoin 1665 coined the term cell to describethe tiny array of a cork slice after observ-ing it with a primitive microscope thathe had constructed. By happenstance, thesame Robert Hooke, along with Isaac New-ton, also studied mundane soap bubblesunder the microscope, and described theso-called black holes in soap films. Newtonestimated the thickness of ‘‘blackest’’ soapfilms to be 3/8 × 10−6 in. When convertedinto modern units, the thickness is about6 to 9 nm, which is in excellent agreementwith modern measurements [1, 13]. Theearly investigation of black soap films hada profound influence in the development

of the lipid bilayer concept of biomem-branes and its subsequent experimentalrealization in BLMs (planar lipid bilayersand spherical liposomes).

16.6The Function of Biomembranes Pertainingto Electrochemistry

At about the time while the lipid bilayerconcept was being proposed in the 1920s,Fricke, a physicist, carried out the follow-ing electrical measurements on RBCs, inanswering the question ‘‘How thick is theplasma membrane of red blood cells (RBC)’’?Using a Wheatstone bridge, the conduc-tivity and capacity of the RBC suspensionwere measured as a function of frequency.At low frequencies, the impedance of thesuspension of RBC is very high, whereas athigh frequencies the impedance decreasesto a low value. To explain his findings,Fricke proposed a model. That is RBCs aresurrounded by a thin layer of low dielectricmaterial electrically equivalent to a resistor(Rm) and a capacitor (Cm) in parallel. Thus,the lines of current flow around the RBC atlow frequencies. At very high frequencies,the resistance becomes very low becauseall the current is shunted through the ca-pacitor. In this connection, it should bepointed out, however, that Hober in 1910found that suspensions of intact RBCshave a high electrical resistance, while thecytoplasm has a conductivity similar tothat of physiological saline. From this fact,Hober concluded that the cell membranehas a high electrical resistance. Using theformula for a parallel plate, Fricke deter-mined the capacitance (Cm) of the RBCsto be 0.81 µF cm−2. For a parallel-platecondenser, the capacitance is given by

Cm = ε0εr

4πtm(1)

16.6 The Function of Biomembranes Pertaining to Electrochemistry 519

where Cm = membrane capacitance (µFcm−2), εr = relative dielectric constantof the membrane, and tm = membranethickness. Fricke calculated the thicknessof RBC membrane, tm, to be 3.3 and11 nm, assuming εr = 3 and 10, respec-tively. Indeed, modern measurements onartificial BLMs and biomembranes, fullyconfirmed Fricke’s estimation of the thick-ness of the basics of membrane plasmamembrane [3].

16.6.1Basics of Membrane Electrochemistry

In membrane electrochemistry, the mostuseful approach is based on theNernst–Planck (N–P) equation, as will bediscussed in this section. In membranes,transport processes are of paramount

importance. The key concept here is theelectrochemical potential (µi ), a quantitythat is related to the chemical potential(µ) and the electrical potential (φ).Other electrical parameters of membranesare resistance (Rm), capacitance (Cm),current/voltage (I/V ) characteristics, andbreakdown voltage (Vb). These electricalproperties of membranes are often usedin fundamental studies and in practicalapplications (Fig. 2).

16.6.1.1 Nernst–Planck EquationIonic gradients and ion transport aredeemed to be the principal processes re-sponsible for the formation and changesof the transmembrane electrical potential,affecting the membrane excitability, enzy-matic and immunological activities, and

Fig. 2 An ultrathin membrane (e.g.BLM) separating two aqueoussolutions. Biface is defined as twointerconnecting interfaces wherematerial and energy transport arepossible. Ef = /tm, = electricalpotential, = potential differenceacross the membrane = Em, tm = lipidbilayer thickness (estimate varies from 5to 6 nm) (see Fig. 1 and text for details).

mib, mi

b~ mia, mi

a~

BifaceInterface

Aqueous solution(left, outer side)

Aqueous solution(right, inner side)

Planar lipid bilayer(BLM)

Dipole layer

x = 05 nm

x = tmtm

Ψd

Interface

Ψs

Ef

∆Ψ


transducing function. All these phenom-ena are generally credited with metabolicprocesses being able to maintain the steadystate, nonequilibrium conditions acrossthe biomembranes by generating an activetransport of ions or electrons. For instance,in the case of excitable membranes, thistransport of ions across biomembranesgenerates and maintains nonequilibriumionic concentration gradients, allowing themembrane potential to shift back andforth between the equilibrium potentialsof different ionic species. The membranesystem under consideration is assumedto be of molecular thickness and of pla-nar configuration, and that transport ofcharged species (flux = Ji ) takes place inone dimension (the x-direction) perpen-dicularly to the membrane (Fig. 3). Thus,the flux is given by

Ji = −kUiCi

dµαi

dx(2)

where the electrochemical potential, µαi , is

defined in terms of chemical potential, µ,and electrical potential, φ. That is

µi = µi + ziFφ = µ0i + RT ln ai + ziFφ

(3)

and ai = γici , where γi is the activitycoefficient. We will assume that ai

∼= Ci

(i.e. ideal solution) in all further equationsof this chapter.

The driving force for moving ionicspecies across a planar lipid bilayer inthe x-direction may be expressed as anelectrochemical potential gradient, that is

dui

dx∼= RT d ln[Ci ]

dx+ ziF

dφ

dx(4)

The electrochemical potential is a funda-mental measure of irreversibility as may bededuced from the second law of thermo-dynamics in the following manner. Fromthe entropy S and dS ≥ 0, where dS is the

change in S for the reversible (S = 0)and irreversible processes (S > 0), re-spectively. With a BLM interposed betweentwo aqueous solutions, it can be shownthat, for electrical work (Wele) only,

T dS = −µ dn − dWele

= −(µαi − µ

βi ) dn − ziF (µα

i − µβi )

= −(µαi − µ

βi ) dn − zF (µα

i − µβi )

= −(µαi − µ

βi )

= −µαβi (5)

where superscripts α and β denote insideand outside of the membrane, respectively(Fig. 2). From Eq. (4), transport of ionsand other species is caused by a forceequivalent to −dµ/dx (Fig. 3).

16.6.1.2 Thermodynamic ConsiderationsA living cell is most unique in that itsplasma membrane separates two aqueoussolutions, which gives rise to a voltageor an electrical potential difference (PD)on the order of tens of millivolts (mV).This small PD is essential, however,for functioning of the cell. What is themeaning of this observed PD? From theviewpoint of thermodynamics, a cell is an‘‘open system’’ involving both transport ofmatter and energy. Thus, the answer to theposed question may be simply answered asfollows: the observed PD is a manifestationof electrical activities across the barriermembrane, which in turn is a measure ofthe Gibbs free energy change (G):

G = −nFEm (6)

where Em denotes the membrane poten-tial or voltage (PD), n is the number ofelectrons involved in the reaction, and F

is the Faraday constant. From Eq. (6) thefollowing can be obtained:


S m

OP C

R

E

Lipid bilayer

Out (o)

Co

~

Jo → i

Ci

CH

x = 0 x = m

In (i)

RE

mib ~mi

a

RE

(BLM)

tm

Fig. 3 An ultrathin membrane (e.g. BLM) separating two aqueoussolutions. (a) Electrical arrangement for measuring membranepotential and current. S = switch, m = voltage supply,OP = operation amplifier, C = current out, R = resistor (>1010 ),E = electrometer. (b) Experimental cell showing a BLM interposedbetween two aqueous solutions. RE = reference electrode (Ag/AgClor SCE saturated calomel electrode), µi = electrochemical potential,C = concentration, J = flux, CH = channel, tm = membranethickness, o and i denote, respectively, outside and inside bathingsolution [2, 3].

E0m = RT

nFln K (7)

where E0m is a potential at standard

conditions and K is the equilibriumconstant of a given reaction. It shouldbe pointed out that, when a pair ofreference electrodes is placed across the

cell membrane, one is actually measuringthe sum of two half-cells:

Ecell = Em = Ein − Eout (8)

where Ein and Eout denote the inside andthe outside of the cell, respectively. Byconvention, the inside of the cell is taken


as negative. In this connection, the poten-tial developed across the membrane resultsfrom a difference in electrochemical poten-tial resulting from the equilibrium of ion,I, partitioned between two phases:

Iα −−−−−− Iβ (9)

where Iα and Iβ represent species I inthe inner and outer phase, α and β,respectively. Through the Nernst equation,the values of half-cell potentials can berelated to the activities (or concentrations)of the species involved in the solutionsbathing the membrane. Then one obtains

Em = Ecell = E0 + RT

nFln

[Iα ]

[Iβ ](10)

where [Iα ] and [Iβ ] are the concentrationsof active species, in the two phases. Allactivity coefficients are taken as unity. Onthe basis of information given in Fig. 3,a PD of 58–60 mV would be expectedfor 10-fold difference in concentrationsat room temperature (22 ± 2 C). If themembrane is selective, for example, itprefers cations over anions, then themore concentrated side will be negative.Later, after completing our discussion ofthe N–P equation, we will return to theexperimental methods of monitoring thesemembrane potentials.

Since the N–P equation is central tomembrane biophysics (transport, electro-chemistry, and physiology), it is instructiveto present more details here. From Eq. (2),the flux (J ) across a membrane is equalto velocity times concentration (C), i.e.J = vC. The negative sign indicates thedirection of decreasing concentration. Theflux is also proportional to the mobility(U ) and concentration (C), and the drivingforce of the species involved. In Eq. (4),where k is a constant to be evaluated. Forsimplicity, we have dropped the subscript

i. By substituting µ (Eq. 3) into Eq. (2), wecan write:

J = −kUC

RT d(ln C)

dx+ zF

dφ

dx

(11)

The value of k is found by rewritingthe last term of Eq. (11) using for µ =v/(−zdφ/dx) because the mobility U ofthe species is defined as the ratio of thevelocity v of the ion and the potentialgradient of the field causing the ion’smotion. Since vC has the dimension ofthe flux J , it follows that k = 1/F . Eq. (11)can then be rewritten in the following way:

J = −UC

F

RT

d(ln C)

dx+ zF

dφ

dx

(12)

Equation (12) is the famous Nernst–Planck flux equation, upon which all basicequations relevant to membrane electro-chemistry and electrophysiology depend!For example, if the species being trans-ported is not charged (i.e. z = 0), we haveFick’s diffusion equation. Note here thatthe first term of the right-hand side of theabove Eq. (12) describes the diffusion forcecaused by concentration gradient, whilethe second term describes the electricalforce caused by the electrical gradient de-veloped across the membrane (Fig. 2). Inorder to integrate Eq. (12), a two-variabledifferential equation, many attempts havebeen made. In all cases an a priori as-sumption has been made, that is, that theindependent variable in the resulting so-lution is the concentration distribution,with potential gradient being its conse-quence. The first approach was made byHenderson in 1907, who assumed thatthe concentration gradient is linear withinthe membrane phase so that potential gra-dient is nonlinear. This is valid for theliquid junction potentials. However, in thecase of biomembranes, and particularly


for BLMs, the lipid core is a barrier withvery high energy for charge penetration.Thus, the assumption of a constant con-centration gradient is unrealistic. The bestknown is the second approach by Gold-man, who assumed linearity of potentialgradient across the membrane phase sothat the concentration gradient cannot belinear. It is worth noting that the Hender-son equation gives the better results fora system with only one simple electrolytepresent in different concentrations on bothsides of the membrane. However, for bio-logical systems with various permeant ionspecies on both sides of the membrane,the Goldman approach usually gives betterresults.

With the above-mentioned background,it is informative to describe an applicationof the N–P equation as applied to the nerveaxon, which is surrounded by a plasmamembrane. One of the main functions ofthe plasma membrane is to control thepassage of ions and molecules into andout of the cell. For most biomembranes,the intracellular [K+]i greatly exceedsextracellular [K+]o, and the opposite istrue for the extracellular [Na+]o and[Cl−]o. These concentration differencesare due to the active transport systemembedded in the lipid bilayer of the plasmamembrane [3].

16.6.1.3 Nernst EquationAt equilibrium, there is no net flow of ionsacross the membrane. Therefore

J = 0; RT d ln C = −zF dφ (13)

Upon integration of the above Eq. (13)across the membrane, with z = 1 forunivalent ions, it gives us

Eα − Eβ = Em = RT

Fln

[aβi ]

[aαi ]

(14)

where Em is the membrane potential.By substituting appropriate ionic concen-trations of the nerve axon, the Nernstpotentials Em = −88 mV for K+ andEm = 55 mV for Na+, are obtained.

16.6.1.4 Electroneutrality PrincipleIn a system consisting of an ultrathinmembrane (e.g. a planar BLM or a cellmembrane) separating two aqueous so-lutions, the law of electroneutrality mustbe obeyed. That is, the sum of positivelycharged species is equal to the sum ofnegatively charged species for the sys-tem as a whole. However, as pointedout by Guggenheim, charge separationon a microscopic scale may occur atmembrane/solution interfaces [7]. For ex-ample, to develop a potential difference(PD = Em) of 50 mV across a BLM of 5 nmthickness, the amount of charged speciesinvolved is far too small (10−10 M) tobe detected chemically, or by any othermeans, except electrically. Therefore, theEm across a modified BLM (or a biomem-brane, for that matter) cannot be simplyrelated to the composition of bulk solutionconcentrations. The pivotal role played bythe membrane is implicated, as describedin the following section.

16.6.1.5 Ion Selectivity, Ion Specificity,pH, and Cell MembranesOne of the main tasks of biomembranefunctions is ion or molecular recognition,which entails selectivity and specificity ofthe membrane. By selectivity for ions ismeant that positively charged species overthe negatively charged ones (e.g. Na+ overCl−), whereas ion specificity is defined thatone kind of ions over another (e.g. K+ overNa+). Thus, understanding the principlesthat lie behind the structural–functionalrelationship of biomembranes should help


in providing the insights of much of thecell’s membrane electrochemistry.

Let us return to the discussion of theobserved membrane potential (Em) ofthe cell; the magnitude of which is onthe order of millivolts. This brings usimmediately to the problem of how tomeasure and analyze this vital electricalparameter of the cell. Fortunately, electro-chemists have provided life scientists witha variety of reference electrodes such asAg/AgCl and Hg/Hg2Cl2. The purposeof a reference electrode is to maintaina constant potential without changing itwith respect to time. A good referenceelectrode should not be easily polarizedduring a measurement. Further, to pre-vent contaminations, a salt bridge made ofa piece of tubing filled with agar gel (e.g.0.3 g of agar/15 ml of 3 M KCl) is usu-ally attached to the electrode. For example,‘‘Hg/calomel Hg2Cl2/saturated KCl/agargel’’ represents a so-called saturatedcalomel electrode (SCE). The standard po-tentials for Ag/AgCl and Ag/Hg2Cl2 are0.222 and 0.268 mV, respectively (vs. NHE,normal hydrogen electrode). Thus, whena pair of these reference electrodes areplaced across the cell membrane (for thatmatter across any physical barriers withelectrical conductance), the resulting PDobserved is owing to the contribution of themembrane and its bathing solutions only,since the reference electrodes themselvesare invariant. It is worth remembering thatwe cannot measure a single potential of anelectrode but only the PD. The voltage (V )or the PD of a cell is measured, whennearly zero current is passing throughthe contacting reference electrodes. Nowa-days, it can be easily accomplished. Allone needs is a high impedance voltmeteror an electrometer. In this connection,we need to consider the measurement ofH3O+ or pH, since ionic species of all

kinds are involved in biomembrane func-tions. The hydrated H+ occupies a mostunique place, and its function in mem-brane electrochemistry deserves a specialconsideration, as investigated by Hender-son and Hasselbach, and by Nicolsky (seefollowing section and Ref. [3]).

16.6.1.6 The Henderson–HasselbalchEquationIn acid–base physiology, one is concernedwith the concentration of hydrogen ions(H3O+) in the bulk phase. The Hender-son–Hasselbalch equation of the formgiven below is of interest:

pH = pK + log[HCO3

−]

[H2CO3](15)

where pH is defined as − log[αH+ ], pK =3.6 for the bicarbonate reaction. For nor-mal healthy people, the pH values of theextracellular and intracellular fluids fall inthe range 7.35–7.45. When pH values areoutside the range between 7.1 and 7.7 forlong periods, the life as we know it is nolonger viable. This narrow 0.6 pH unitappears to be small but is actually mislead-ing. From a biophysical viewpoint, the pHvalues are best expressed in terms of H+ion concentration in moles per liter. Thus,a pH of 7.4 represents a 40 nM of H+ion concentration. A pH of 7.7 is equal to20 nM of H+ ion concentration. Evidently,a change of merely 0.3 pH unit representsa twofold change in H+ ion concentration!To better appreciate these numbers, one isreminded of the normal K+ ion concentra-tion in the extracellular fluid, which is onthe order of 5 mM, and is of crucial impor-tance in nerve functioning, whereas theH+ ion concentration is expressed in nM(1 nanomolar = 1 × 10−9 moles per liter).Thus, there is a 1 000 000 times differencein magnitude of these two ion concen-trations! Further, the vital role is played


by hydrogen ions in the body fluids, bothintracellular and extracellular, in cell func-tions (e.g. enzyme catalysis, periodontaldisease, etc.).

Ion-selective membranes A simple andreliable way to access pH is of evidentimportance. The problem for accuratepH measurements was solved success-fully in the 1930s, using a soft soda-lime glass (made of SiO2 : Na2O : CaO =72 : 22 : 6 wt.%). Owing to the high impe-dance of the glass used for pH sensing, anelectrometer was necessary. Today, withmodern electronics, this presents littleproblem, as already mentioned above. Itis worth noting, however, that a pH glasselectrode used today may be thought of as athin membrane separating two aqueous so-lutions, with ion selectivity and specificityfor H3O+. In this sense, certain dopedBLMs have been found to behave simi-larly. Therefore, equations developed forion-selective membranes are applicable toboth cases. We will discuss here only theion-selective BLM interposed between twosolutions.

A typical cell for BLM studies usuallyconsists of two silver–silver chloride (orSCE) reference electrodes arranged asfollows:

Ag| AgCl,Cl−(aq),left solution|modified BLM|right solution./

Cl−(aq),AgCl|Ag (16)

It is worth stressing that a modified BLMin the above cell arrangement behavesas a bipolar electrode or two workingelectrodes connected in series. The tworeference electrodes depicted in the setupmerely facilitate the connection to themeasuring instrument. The cell potential

difference (Em) observed across an ion-selective membrane (e.g. BLM) is called amembrane potential. In other words, mem-brane potentials are responsible for theoperation of the cell membranes in liv-ing organisms. Electrochemists make useof ion-selective membranes to constructchemical sensors for various ions suchas hydrogen, sodium, potassium, calcium,and fluoride ions, to name a few [7–9].

The Nicolsky equation With common pHinstruments, the Nernst equation canaccurately predict the pH of an unknownsolution, over the pH range 2 to 12. At moreextreme values of pH, some response toother ionic species in solution begins tobecome apparent. This is the so-calledsalt error. For example, a valinomycin-doped BLM for K+ suffers from someinterference because of Na+. A betterequation proposed by Nicolsky should beused. Accordingly,

Em = k + RT

ziFln

(ai + K

potij a

zi/zj

j

)(17)

where Kpotij stands for the selectivity

coefficient. The slope (RT/F ) is measuredin millivolts per decade.

16.6.1.7 Origins of Membrane PotentialsBy placing a pair of reference electrodes(e.g. SCE) on opposite sides of a biomem-brane, a PD (voltage, V ) on the order oftens of millivolts (mV) is usually detected.This voltage suggests an unequal distri-bution of charges across the membrane.These charges may be ions or free chargessuch as electrons and holes, or they maybe dipolar molecules [3]. This charge im-balance across the membrane is a result ofcharge separation, and may be due to oneor a combination of several of the followingsituations:


• preferential adsorption of ions;• orientation of dipolar molecules;• movement of ions from one side to the

other;• deformation of molecules;• oxidation–reduction (redox) reactions;• perturbation of the membrane by ex-

ternal sources (light, applied voltage,pressure or temperature difference, andothers).

In addition to the membrane voltage(V or Em), there are other electrical pa-rameters associated with the membrane.Prominent among these are membranecapacitance (Cm), resistance (Rm), current(Im), and breakdown voltage (Vb). Frombasic physics, V ∝ q (the fundamentalcharge associated with an electron), onecan write

q = CV (18)

where C is the proportionality constant,termed capacitance. Thus, we have thefamiliar formula:

C = q

V(19)

The unit for C in membrane electrochem-istry is defined in microfarad (µF) perunit area in cm2. Recalling Eq. (1), Frickedetermined the capacitance (Cm) of theRBC membrane using a parallel-plate con-denser formula. It should be noted herethat a capacitor can be discharged, andtherefore there is a capacitive current (ic)connected with it, which is given by

ic = dq

dt= C

dV

dt(20)

and ∫dV = V = 1

C

∫ic dt (21)

This capacitive current, ic, differs fromthe current, I , expressed in Ohm’s law

(E = IR). The capacitive current is adisplacement of charges on opposite sidesof the membrane, and therefore, it doesnot involve a direct translocation of ionsthrough the membrane. One needs toremember that the electrical current isexpressed in amperes (coulombs persecond).

16.7The Role of Electrical Double Layers (EDLs)and their Biological Implications

A separation of charges occurs at almostall interfaces; that is, interfaces are electri-fied. This leads to the formation of EDLsat the interface. This effect is of particularimportance for the systems of well-definedsurfaces relative to the volume or thick-ness of a given phase. This is because thepresence of EDL being the primary en-ergy barrier, is the factor governing theproperties of such systems [3, 5, 7].

The experimental evidence demon-strates that the EDL model elaboratedindependently by Gouy and Chapman atthe turn of twentieth century can de-scribe adequately the distribution of elec-trical potential adjacent to a BLM. In theGouy–Chapman theory, the Poisson equa-tion relates the electrostatic attraction ofthe counterions to the charged surface andthe Boltzmann relation – the tendency ofthe counterions to diffuse from a region ofhigh concentration. The Gouy–Chapmantheory does not account for the discrete-ness of fixed structural surface charges butconsidered them to be smeared uniformlyover membrane surface. Gouy and Chap-man assumed also that the aqueous phaseis structureless with a constant electricpermittivity coefficient and the ions arepoint charges. The Gouy–Chapman the-ory was derived for uniunivalent electrolyte

16.7 The Role of Electrical Double Layers (EDLs) and their Biological Implications 527

solutions. Physiological solutions usuallyare more complicated, containing bothmonovalent and divalent cations. In thiscase, the description given by Grahameand others is more correct. Moreover,some ions bind to the surface, pro-ducing an electrostatic potential at thissurface lower in magnitude than thosepredicted by Gouy–Chapman theory. Thesimplest way to include this processwas proposed by Stern who introducedthe Langmuir adsorption isotherm intothe Gouy–Chapman theory. Although theEDL theory for classical electrochemicalsystems (metal/electrolyte solution) wasdeveloped relatively early, it was not un-til 1935 that Theorell applied it to themembrane processes. He also introducedthe idea that the membrane potential is asum of an electrodiffusion potential acrossthe membrane proper and the effects oftwo boundary potentials at both mem-brane–solution interfaces. Since then,there has been a renaissance of activ-ity testing various models of EDL at themembrane–solution interface. The elec-trochemical behavior of biomembranesand their models is therefore largely af-fected by the presence of EDL. In thefollowing paragraphs, we will illustrate theorigins of various membrane potentials interms of BLMs (planar lipid bilayers).

The unique aspect of BLM, simi-lar to biomembranes, is the ultrathinlipid bilayer separating two aqueous so-lutions (Fig. 3). Membranes of these typeshave two interdependent interfaces (i.e.a biface, meaning the two coexistingsolution–membrane interfaces, throughwhich material, charge, and energy trans-fer are possible). The creation of a biface isthe result of two immiscible phases withvery different relative dielectric constants(εwr = 80; εmr = 2–5). At each interface,

as well as across the ultrathin lipid bi-layer (∼5 nm), powerful electric fieldsmay exist, which facilitate charge sepa-ration and transport, as discussed laterin conjunction with the origins of BLMpotentials.

An observed transmembrane PD (Em) ina BLM system may be due to one or a com-bination of the following potentials [3, 7]:

• diffusion (concentration) potential, φd;• adsorption (surface or interfacial) po-

tential, V ;• distribution (outer) potential, U ;• Galvanic (inner) potential, φ(= V + U );• Gouy (Donnan) potential, UG or φDon;• electrokinetic potential, Uk;• thermoelectric potential, Et;• redox (electrostenolytic) potential, ERX;• photoelectric potential, Ehν .

In the classification given above, redox,thermoelectric, and photoelectric poten-tials in the BLM are discussed in a latersection, whereas a brief summary will begiven on the electrokinetic potential rele-vant to membrane electrochemistry.

16.7.1Electrokinetic Potential, Uk

The electrokinetic potential manifests it-self when one phase is in motion relativeto the other. This relative movementmay be on account of (1) applying me-chanical forces acting on the liquid orsolid phase, and (2) applying an elec-tric field tangentially to the EDL. To thefirst group belong two electrokinetic phe-nomena: the Dorn effect and streamingpotential. The second group consists oftwo other phenomena: electrophoresis andelectroosmosis. These are summarized inTable 2.


Tab. 2 Electrokinetic phenomena at interfaces

Phenomena Stationary phase Moving phase Mechanical Potentialforce difference

Dorn effect Electrolyte solution Suspended particles Cause ResultStreaming potential Walls of capillary Electrolyte solution Cause ResultElectrophoresis Electrolyte solution Suspended particles Result CauseElectroosmosis Wall of capillary Electrolyte solution Result Cause

The Dorn effect It occurs when particlessuspended in the electrolyte solutionprecipitate under the influence of gravity(or centrifugal force), and results inpotential difference between the top andbottom of suspension.

Streaming potential When liquid flowsalong a narrow tube or capillary, or througha porous plug (a bundle of narrow tubesin parallel or channels in BLMs andbiomembranes) under a hydrostatic head,a difference of potential is set up betweenthe ends of the capillary or across theporous plug as the case may be.

Electrophoresis This is the migration oflarge molecules (e.g. proteins, polymers),microscopic aggregates of molecules (e.g.colloidal particles) or cells, or lipo-somes (e.g. erythrocytes) under the in-fluence of an electric field applied tothe medium in which the particles aresuspended.

Electroosmosis If a difference of potentialis deliberately set up along the axis ofnarrow tube or across a porous plug,liquid in the tube or plug flows untila hydrostatic head sufficient to preventcontinued flow has built-up to the valueof the electroosmotic pressure. In all thesephenomena, the EDL and potentials areinvolved as stated above.

Among electrokinetic phenomena, elec-trophoresis plays a prominent role. Inthe form of a special technique, itserves for analytical purposes and asa tool in the investigation of cell dis-eases and membrane phenomena. Foranalytical purposes, this technique hasbeen used in various forms of paperor gel electrophoresis (two-dimensionalelectrophoresis, immunoelectrophoresis,isotachoelectrophoresis, and others) forpreparative separation and identification offraction components. Besides this, as a mi-croelectrophoresis, the technique may beused to elucidate the nature, number, anddistribution of charge groups in the periph-eral zone of biological cells or to separatesubpopulations in a mixture of cells, orvesicles (continuous flow electrophoresis,electrofocusing). Microelectrophoresis hasbeen used for diagnostic tests in medicineas well as for investigations of cell dis-eases and membrane phenomena. It isalso a very important method in the in-vestigation of membrane structure andbioelectrochemical phenomena at the cellmembrane–electrolyte solution interface.

Under separate headings, the nature andorigin of other membrane potentials: diffu-sion (concentration) potential adsorption(surface or interfacial) potential, distri-bution (outer) potential, Galvani (inner)potential, and Gouy (Donnan) potentialwill be considered. These potentials are


best described in the language of inter-facial chemistry, and have been treatedin detail [3, 6, 7]. In order to apply theseconcepts relevant to membrane electro-chemistry, it is informative, first of all, todefine the terms and to summarize someof the general conclusions of these poten-tials [3, 7].

16.7.2Adsorption Potential

Also known as surface or interfacial po-tential (designated by symbol V ), it isopposite to that observed with the diffu-sion potential. For this interfacial potentialto appear, a difference in salt concentra-tions across the BLM is not required; itdepends upon the sorbed species and/orthe dipoles. With cationic interface-activecompounds, such as hexadecyltrimethy-lammonium bromide (HDTAB) and phos-phatidylethanolamine (PE) at low pH, anegative potential results and vice versawith anionic interface-active compounds,such as dodecyl acid phosphate (DAP) andphosphatidylserine (PS). The magnitudeof V decreases with increasing of the saltconcentration. Two other caveats should benoted about this potential: (1) it may haveonly transient existence; however, the timefor its decay can take a long time and (2)the law of electroneutrality at the interfacesis not obeyed [1].

16.7.3Distribution Potential

It is also called outer potential, symbolizedby U . The potential arises as a result of adifferential distribution of the oppositelycharged species across the interface, themagnitude of which is influenced by thediffusion of ions and their concentrations,

and increases with increasing salt concen-tration, but is independent of the nature ofthe common ion.

The sum of interfacial potential and dis-tribution potential is termed the inner orGalvani potential, designated by φ. It willbe discussed below that the observed trans-membrane potential difference, Em, is dueeither to V, U , or to V plus U ,as defined above. It should be remem-bered that there are two components of φ

(=V + U ). They are concerned with onlythe distribution potential. If compounds,such as phospholipids and interface-activeagents, are preferentially adsorbed at theinterface, the so-called adsorption poten-tial (V ) may also develop. As shown inFig. 2, both adsorbed fixed charge speciesand dipoles may contribute to the observedpotential. The nature and origin of the ad-sorption (or interfacial) potential can bediscussed in terms of the classical EDL the-ory of Gouy–Chapman–Stern–Graham[7, 15–19].

16.7.4Diffusion Potentials

In general, diffusion potentials appear al-ways when an electrolyte diffuses down itsconcentration gradient. However, one pre-requisite condition, namely, both cationicand anionic species must have differentmobilities. The magnitude and sign ofsuch potential depends on the differencein mobilities between cation and anion(the greater the mobility difference, thegreater the potential difference), with themore dilute region having the same signas the faster ion. In bulk solution, thesepotentials are readily disturbed by con-vection and may be neglected, but if twoelectrolyte solutions of different concen-trations are in contact in a way that avoids


convectional mixing but allows a diffu-sional flow, the diffusion potentials arestable. Such electrolyte contact may be ac-complished either in a direct way or byelectrolyte separation by semipermeablemembrane. The direct contact of two im-miscible solutions leads to the so-calledliquid junction potentials that are a pos-sible source of error whenever electrodeswith salt bridges are used. The electrolyteseparation by a permselective membraneis characteristic of biomembrane systems,and has already been discussed above. Forthe sake of simplicity, all ion fluxes areconstant and independent of time (steadystate conditions). This requires the stableproperties of the membrane and constantconcentrations at the membrane. The lat-ter may be achieved in several ways. Inthe case of experiments with BLMs, thevastness of the two compartments on bothsides of the membrane provides practicallyconstant concentrations.

16.7.5The Gibbs–Donnan Potential

The Gibbs–Donnan potential occurs whena nonpermeant ion (for biological sys-tems, usually a polyion such as a protein)is unequally distributed between the twoelectrolyte solutions separated by a perms-elective membrane, which allows certainelectrolyte ions to move freely betweenthe two solutions. The second law ofthermodynamics and the principle of elec-troneutrality restrict this movement. Thefirst restriction requires that each per-meant ion species moves only down itselectrochemical potential gradient, the lat-ter requires the sum of all positive charges(cations) to be equal to that of all negativecharges (anions) in each solution. Whenthe system reaches its equilibrium, there isno flux of any permeant species i, Ji = 0.

Eq. (12) then takes the form

ziFdφ

dx= −RT

d(ln Ci)

dx(22)

Integration of this equation over the wholemembrane thickness at the equilibriumconditions leads to the expression forelectrical potential difference developedacross the membrane, known as Donnanpotential,

φII − φI = IIDIφ = RT

ziFln

CIi

CIIi

(23)

where superscripts I and II denote the twosides of the membrane. This dependencemay also be obtained in another way. Themain condition of the equilibrium state isequilibrium of electrochemical potentialsof each permeant species on both sides ofthe membrane

µIi = µII

i (24)

which leads immediately to the depen-dence

RT ln CIi + ziFφI = RT ln CII

i + ziφII

orIIDIφ = RT

ziFln

CIi

CIIi

(25)

being identical with Eq. (25) may be usefulin the determination of nonpermeantpolyion concentration (see Eq. 13).

An electric potential difference is gen-erally established at a membrane (e.g.a BLM) and adjacent solution interface.There have been two entirely differentapproaches to describe this potential dif-ference, with respect to location of fixedmembrane charges and BLM permeabilityto ions. When a membrane being per-meable to ions contains homogeneouslydistributed fixed charge groups and is in


equilibrium with a symmetrical electrolytesolution, one can consider the existingpotential difference in terms of Donnanequilibrium potential. Donnan, in 1911,studied the electrical potential set up ata semipermeable membrane between twoelectrolytes, an effect of importance in liv-ing cells, known as the Donnan effect.Specifically, when a thin, selectively per-meable membrane separates two solutionsto small ions (e.g. Na+, K+, Cl−, etc.) butnot to large ionic species (e.g. proteins).The diffusion of oppositely charged, per-meable ions will be in equilibrium as soonas the concentration ratios for the ionshave been established. That is, for exam-ple, [Na+]out/[Na+lin = [Cl−lin/[Cl−lout =rD, where subscripts ‘‘out’’ and ‘‘in’’ de-note, respectively, the outer and innersolution, and rD is the so-called Donnanequilibrium constant (assuming of coursethe activity coefficients of respective ionsare unity). Owing to the concentration dif-ferences of permeable ions, an electrostaticpotential difference is established in addi-tion to a difference in osmotic pressureacross the membrane. This membranepotential difference, known as the Don-nan potential, may be described by theNernst equation. It should be stressedthat the Donnan phenomenon is the re-sult of certain ionic species, either fixedon or not able to, move away from themembrane. Further, the Donnan phe-nomenon may be discussed in its simplestform. The potential changes discontinu-ously at the boundary between the twoelectroneutral phases. However, replacingthe assumption of local electroneutrality,used in the classical approach, by the Pois-son–Boltzmann equation, under suitablemodification of BLM or electrolyte solutionwith lipophilic probe ions this situationmay be improved. If, on the other hand,one assumes that all charges are smeared

uniformly over the surface of impermeableBLM, the potential profile in the EDL isdescribed by the Gouy–Chapman theory.The potential distribution in the proxim-ity of membrane–electrolyte interface maybe regarded in fact as the two limitingcases of more general Poisson-Boltzmannequation. This equation may be derivedfor a model charged membrane, perme-able to some extent to ions, in which thefixed charges are uniformly distributedthrough a layer of finite thickness t . Italso includes the possibility of ion bind-ing to the fixed charge sites. By varyingthe thickness of charged layer, it has beenshown that the constant changes of po-tential profile occur across the membraneinterface, which for t l take the formof Gouy–Chapman theory, while for t > lis described by the Donnan theory. Thus,both the Gouy–Chapman theory and theDonnan theory can be considered as spe-cial cases of the Poisson–Boltzmann equa-tion. This model assumes the uniformand delocalized distribution of chargesfor BLMs containing certain amount ofcharged phospholipids (e.g. PA or PE). Forthis case, the negative charges of thesemolecules should be considered as local-ized, although they may still be distributedhomogeneously. Under the influence ofexternal field applied across the BLM,the surface charge density induced bymembrane polarization would be consid-erably high for discrete charges than for‘‘smeared out’’ ones. This should largelyaffect the BLM stability and conductanceproperties. The simplest way to treat theproblem of lateral discrete distribution ofcharges on the membrane surface is to re-place the continuous distribution functionby the sum over all discrete membranecharges fixed in a regular lattice, usuallysquare. From the solution to the Pois-son–Boltzmann equation, the electrostatic


potential at any point in the aqueousphase is calculated to be identical to thelattice sum over the Coulomb potentialsbecause of individual fixed charges. Thelocal potential maxima at all charged sitesin the lattice are taken into considera-tion. As a consequence of such maxima,this model predicts that the discrete na-ture of the surface charge should manifestitself in the discrete adsorption of ionsat the membrane surface. However, theeffects of discrete charges have to be con-sidered when charges are buried withinthe low dielectric interior of the BLM orthe lipids have multiple charges in theirheadgroups [3].

16.7.6Phase Boundary Potentials

These are also known as surface poten-tials. Phase boundary potentials, appearingat the interfaces between the membraneand adjacent electrolyte solutions, are at-tributed mainly to fixed charges within orattached to the membrane phase. Thesefixed charges that originated from ionicgroups are part of the membrane struc-ture and, under special conditions, maychange their number through the dissoci-ation of weak acidic groups or converselythrough association of ions with bindinggroups. The electric dipoles (e.g. head-groups of lipid matrix) also contribute tothese potentials. The phase boundary po-tentials do not usually directly add to thetransmembrane potential difference; how-ever, they do affect the electrical potentialprofile within the membrane. Therefore,their presence may alter the permeabilityproperties of membrane (e.g. by affect-ing organization of integral or peripheralproteins), indirectly affecting in this way,the transmembrane potentials. The rela-tionship between fixed charges and phase

boundary potentials can be illustrated byeither one of two simplified models. Thefirst model assumes that the fixed chargesare adjacent only to both surfaces of anonpolar membrane interior. This modelmay represent an artificial BLM. It is alsoclose to a biological membrane. In the sec-ond model, an assumption is made thatthe fixed charges are homogeneously dis-tributed over the whole membrane. Anion-exchange membrane may representthis model.

In the case of an ion-exchange mem-brane that is the second model mentionedabove, an assumption has been made thatits fixed, say negative charges, are homo-geneously distributed over the membranephase. In this situation, the appearingphase boundary potentials may be treatedas two separate Donnan equilibrium po-tentials at each interface. The role ofimpermeable polyion from the previoussection is now as the fixed charges fromthe membrane interior. If at the begin-ning an initial unequal distribution ofpermeant ions exists on both sides ofsuch membrane (let us assume that a lessconcentrated electrolyte solution is on theright side of the membrane), the whole sys-tem will proceed toward equilibrium. Asa consequence, at equilibrium, the elec-trochemical potentials must be identicalon both sides of each boundary. Since bi-ological membranes do not have a highdensity of fixed charges in their interior,the effects of Donnan distribution are ofsmaller magnitude than for ion-exchangemembrane. As stated at the beginning ofthis section, it appears that the perme-ability and conductance properties of themembrane are strongly influenced by theelectrical potential profile. Since the vari-ations of phase boundary potentials mayalter the behavior of pores, channels, andion carriers, they may express themselves


as alterations of current–voltage relation-ships, giving indirect information aboutpotential distribution within the mem-brane and at its interfaces. However, thephase boundary potentials are not usuallyreadily measurable with electrodes towingto the thickness of the membrane phase.

16.7.7Relationship among Various Potentials

Of immediate concern here, we havebeen discussing the observed membranepotential, Em, which is the sum ofall interfacial potentials and diffusion(concentration) potentials at the interfacesand within the membrane. Thus, wehave the appropriate name, the so-calledtransmembrane potential, Em, which isgiven by

Em = φ + φd (26)

orEm = V + U + φd (27)

or

Em = UG + Ep + U + φd (28)

For most of these individual potentials,the final expression takes the form ofthe Nernst equation (Eq. 14). Experimen-tal conditions or circumstances can besuch, that Em measures one or a com-bination of these potentials. Specifically:(1) for a symmetrical BLM, the polariza-tion potential, Ep, is unimportant, (2) foran unmodified BLM, whose interior isliquid-hydrocarbon-like, φ is not mea-sured, (3) for BLM with low interfacialcharge density (e.g. σ < 1, net charge per100 A2), φG contributes very little, (4) forcharged species with substantial solubilityin BLM, φ may be neglected, and (5) fora BLM with high interfacial charge density

(>1 e/100 A2), the diffusion potential, φd,is inconsequential.

It should be emphasized, however, thatthe above statements are offered as generalguidance only. Even for a relatively sim-ple BLM, the measured transmembranepotential (Em) could be a result of vari-ous potentials. The extent of contributionsof these potentials to the overall Em isdependent upon the nature and proper-ties of BLM-forming constituents, whosecharge densities, dipole orientations, andrelative permeabilities to ions are someof the factors that should be taken intoconsideration for a careful analysis of theobserved membrane potential [1, 3, 6–11,29, 30].

In concluding this section, mentionshould be made concerning the dipolesof membrane constituents in the BLM.The dipole flip-flop could be importantfor the voltage-dependent conductance inBLM (e.g. alamethicin-modified BLM).Derzhanski et al. have derived an equationfor the current as a function of frequen-cies. Pastushenko and Chizmadzhev haveconsidered the energetic profile of dipolemolecules in biomembranes in general,and in BLM in particular. This topic hasbeen reviewed in detail by Shchipunov andDrachev, and by de Levie [3, 15–24].

It is known that the value of dielectricconstant is a function of the local electricfield. Since the electric field near the in-terface is not a constant, but a functionof distance from it, the further refine-ment of EDL theory and potential profileat the BLM interface should include thispositional dependency of dielectric coef-ficient. The local value of this coefficientis of course a measure of the influenceof uncharged species such as water orlipid molecules on the interaction betweencharges in their vicinity. The dielectriccoefficient of bulk aqueous electrolyte at


293 K is about 80. Its value near theinterface is not precisely known; usuallyit is assumed to be about 2 to 6. How-ever, owing to finite size of ions andwater molecules, one may speculate thatthe value of dielectric coefficient in closeproximity of the surface would be aboutunity. The transition between these twolimiting values is obviously continuous inthe interfacial region.

Since conventional BLMs are usuallyformed as a symmetrically oriented dou-ble layer of lipid molecules on a hy-drophobic support separating two aqueousphases, these closely packed and alignedmolecules should generate surface dipoles,hence dipole potentials, at each of thetwo BLM–solution interfaces. Thus, allthe latest interest in evaluation of co-herent and precise theory of EDL at themembrane interface should not be un-derestimated. The electrostatic potentialsdue to the charges on lipids or proteinsare important in membrane functioning.They appear to influence the calcium effecton excitable membranes, channel conduc-tance in membranes, passive ion transport,function and structure of thylakoid mem-branes, kinetics and catalytic propertiesof membrane-bound enzymes, electrontransfer reactions, and the interaction ofsome electroactive compounds inside themembrane (e.g. porphyrins) with theirelectron exchange partners in aqueousphase. This influence may also be becauseof the simple electrostatic screening effectof membrane surface charge against thecoionic substrates. Such simple model ofthe interfacial region as Gouy–Chapman’scan provide the explanation for experi-ments carried out on conductance of BLMsmodified with channel forming ionophore,gramicidin [5–7, 13, 15–19, 29].

All dissolved molecules, including thosethat can readily pass through lipid

bilayers and those that transverse withdifficulty (lipid bilayer–hampered sub-stances), are capable of diffusion. Forlipid bilayer–impeded substances, diffu-sion across cell membranes may neverthe-less be highly probable but such diffusionmust be facilitated by lipid bilayer span-ning proteins. Such proteins and theirpolypeptides essentially form carriers orchannels (pores). In this case, they(1) ac-cept specific lipid bilayer–hampered sub-stances at one side, (2) allow the substanceto enter the lipid bilayer without dissolvingin it, and (3) then deliver the substance toeither side.

In the case of water flow acrossa semipermeable membrane separatingtwo electrolyte solutions, two electroki-netic phenomena may be demonstrated,namely, electroosmosis and streaming po-tential. The former associated with Jv,volume flow, is driven by an electric cur-rent, I , (because of ψ ), whereas the latteris owing to pressure difference (P ). Insum, we have

Jv = LPP + LPψψ (29)

and

I = LψPP + Lψψ (30)

where ψ , LψP , and LPψ are, respec-tively, potential, and phenomenologicalcoefficients. In other words, Jv and I

are both driven by P and ψ , whereLψP = LPψ (due to Onsager) is theso-called phenomenological (cross) coef-ficient. The most significant conclusionfrom the above equations is that the coef-ficients relate the volume flow and currentin a given system. Thus, experimentallyone needs only to measure one parameteror the other, which may not be directlydeterminable [3, 31].

16.8 Experimental Membrane Systems 535

16.8Experimental Membrane Systems

Owing to complex structural and environ-mental factors associated with biomem-branes, numerous investigators used dif-ferent techniques and carried out studieson model systems in order to understandthe fundamental life processes. Theseinclude ion accumulation or active trans-port, conduction of nerve impulses, energytransduction, protein synthesis, perme-ability barrier of ions and molecules, im-munological reactions, phagocytosis andpinocytosis, and so on, in physical andchemical terms [3]. Under separate head-ings below, different model systems willbe described.

16.8.1Langmuir–Blodgett Films

In 1917, Langmuir demonstrated in asimple apparatus (now known as theLangmuir trough) that fatty materials in-cluding lipids in organic solvents spread ina monomolecular (monolayer) way whenthey are placed at an air–water interface.From the known quantity of the materialused and measured area covered, the di-mensions of a molecule may be estimated.Later, the Langmuir–Blodgett (L–B) tech-nique and self-assembly of BLM methodswere developed for use in technology.However, in the case of the L–B tech-nique, the ‘‘pinhole defects’’ are difficultto avoid. There is a major difference be-tween the BLM and multilayers formed bythe L–B technique. A BLM, formed eitherby the conventional ‘‘painting’’ methodor self-assembled on a substrate (e.g.a freshly cleaved metallic wire or agargel [2, 3, 32] is a dynamic liquidlike struc-ture that is capable of accommodating ahost of modifiers. In contrast, an L–B

multilayer of bimolecular thickness, albeitmuch more stable than a BLM, usuallycontains pinholes and is in a solid state.The modifier molecules capable of electrontransfer through the film are frequentlyshort-circuited by the pinholes. Until now,however, the above approach seems to beapplied only to the UQ10 incorporated inthe L–B films at the electrode surface andconducting polymer nuclei deposited inthe alkanethiol monolayer modified elec-trodes [32].

16.8.2Immiscible Liquid–Liquid (L–L) Interfaces(ITIES)

The interface between two immisciblesolutions (e.g. water and nitrobenzene)containing dissolved species is a site ofan electric potential. By measuring thispotential difference at the aqueous elec-trolyte/solid electrolyte phase boundary,the phenomena taking place at the inter-face between two immiscible solutions orthe membranes of ion-selective electrodehave been studied. Changing the compo-sition of the solutions in contact can alterthis potential; or applied current can alterthe composition of the solutions. Thus,judicious choice of applied potential orcurrent can be used to study the structureof the interface. Since the interface is ‘‘ul-trathin’’ (1 nm), it cannot be observeddirectly. It can be, however, investigatedby electrochemical or optical methods [14,33]. The knowledge gained on the immisci-ble liquid interfaces can be applied also tothe studies between solution–metal, or so-lution–membrane interfaces. Further, bycontrolling the interfacial potential differ-ence between two immiscible solutionsrates of interfacial electron and ion transfermay be controlled. In the electron transfercase, the interest is in two-phase reactions


that are important in organic synthesis,whereas in the case of ion transfer, liq-uid membranes for metal recycling canbe successfully carried out. For example,AC impedance studies of the ion transferhave evidenced the transient adsorption ofxanthene-type anions and other ions hav-ing asymmetric charge distribution at thenitrobenzene/water interface. The phaseangle of the charge transfer admittance be-comes smaller than unity at the potentialsbeyond the midpoint potential of the iontransfer. This anomaly does not apprecia-bly depend on the sign of the ionic chargeor the location of the midpoint potential, aspredicted by the theory proposed recentlybased on the thermodynamic reasoningof the adsorption and ion partitioning ofsurface-active ions in electrochemical liq-uid–liquid two-phase systems [14, 33].

16.8.3Planar Lipid Bilayers and Liposomes

A good experimental model for biomem-branes should possess a lipid bilayerstructure, onto as well as into which func-tional entities can be embedded. Thus,since the 1960s, the two most widely usedmodel membranes have been BLMs, alsoreferred to as planar lipid bilayers andspherical liposomes. Planar BLMs andspherical liposomes are complementaryto each other, since both types are de-rived from common amphipathic lipidsand related compounds. Both are excel-lent model membrane systems, and havebeen extensively employed for investiga-tions into a variety of physical, chemical,and biological functions. In the remainderof this chapter on membrane electrochem-istry, the focus will be mainly on planarBLMs, because they are easily and havebeen investigated electrochemically (for li-posomes, see Refs. [3, 12, 34]). As a result

of experimental BLM studies, we now havea better knowledge in the following areasof biomembranes:

• structural–functional relationship oflipids, proteins, carbohydrates, pig-ments, and their complexes;

• mechanisms of transport, energy trans-duction, signal processing; and

• sensory perception such as vision,olfaction, hearing, taste, and touchingsensations.

16.9Bilayer Lipid Membranes

16.9.1Formation Techniques

16.9.1.1 Conventional BLMFor all conventional BLM experiments,a cell with two chambers is necessary.This is usually accomplished by having acup made of Teflon (or other highly in-sulating material) placed in a containermade of glass (or some other transpar-ent plastic material) for visualization. Thetwo chambers are electrically isolated fromeach other apart from a small aperture inthe wall of the Teflon cup, onto whichthe BLM is formed. The process of BLMformation is monitored by measuring theincrease in membrane capacitance (Cm)and also by visual observation. The areaof a BLM is measured using a graticleplaced inside the ocular piece of the mi-croscope. As illustrated in Fig. 3, the usualpicture of a BLM interposed between twoaqueous solutions consists of a liquid-hydrocarbon phase sandwiched betweentwo hydrophilic regions. The structure ofthe BLM is considered to be a thin slab ofliquid crystals in two dimensions, havinga fluid hydrocarbon core about 5 nm thick.This liquid-crystalline structure of BLM

16.9 Bilayer Lipid Membranes 537

is essentially an excellent insulator. Theelectrical properties of BLMs have beenextensively investigated, which usuallyentails measuring membrane resistance(Rm or conductance, Gm = 1/Rm), ca-pacitance (Cm), potential difference (Em),dielectric breakdown voltage (Vb), and cur-rent/voltage (I/V ) characteristics (voltam-mogram). Unmodified BLMs (i.e. BLMsformed from common phospholipids suchas lecithin or oxidized cholesterol dis-solved in an n-alkane solvent in 0.1 M KClsolution) has typical intrinsic values ofRm greater than 108 ohms cm2, Cm 0.5 µF cm−2, Em ∼ 0, Vb < 200 mV, withI/V curves obeying Ohm’s law. With a fewexceptions, the results of these measure-ments have been interpreted by treatingthe BLM electrically, in a system equiv-alent to that of a resistor connected inparallel with a capacitor Fig. 4.

16.9.1.2 BLMs via L–B TechniqueConventional BLMs can also be formed bythe L–B technique, [see Tien, 1974]. Essen-tially, it is a combination of the monolayertechnique and the dipping method. A parti-tion made from a sheet of Teflon (25–500-µm thick) with a hole (0.25–1 mm) at thecenter is placed between the two halves ofa trough, with the hole above the aque-ous phase. The trough is first filled withan aqueous solution. After the surface hasbeen cleared, a monolayer of a suitablelipid solution is spread in the surface inthe usual manner. After evaporation of thesolvent, the Teflon partition is then low-ered slowly into the aqueous solution. ABLM is thus formed over the hole (seeFig. 4). Conversely, with the Teflon parti-tion bearing a hole in place, the aqueoussolution on either side is covered by amonolayer. Then, the levels of aqueous so-lutions are raised simultaneously above the

hole thereby forming a BLM from the ap-position of the hydrocarbon chains of thetwo lipid monolayers. Another variationof the monolayer-dipping technique is byhaving a closed chamber used in the waterpermeability studies. The closed chamberwith an aperture on its side is held in place.The chamber, while immersed in aque-ous solution below its aperture, is filledcompletely with aqueous solution. A fewdrops of BLM-forming solution (n-hexaneor n-octane as lipid solvent) are introducedto the surface of the outer aqueous so-lution. After the solvent has evaporated,the level of outer aqueous solution coatedwith a lipid monolayer is slowly loweredbelow the aperture and then raised care-fully above it, thereby forming a BLM.It should be mentioned, however, thatthere is a major difference between theBLM and multilayers formed by the L–Btechnique. A BLM, formed either by theconventional ‘‘painting’’ method or self-assembled on a substrate (e.g. a freshlycleaved metallic wire or agar gel) is a dy-namic liquidlike structure that is capableof accommodating a host of modifiers. Inthis connection, efforts to stabilize BLMsby using polymerizable lipids have beensuccessful. However, the electrochemicalproperties of these BLMs were greatly com-promised [3].

16.9.1.3 Supported BLMsFigure 5 portrays the two most essentialsteps of our experimental procedure. Inthe first step as shown in Fig. 5, the tipof a Teflon-coated wire (platinum, silver,or stainless steel) is cut with a sharp knifeunder a lipid solution, for example, a 1%glycerol dioleate in squalene. When thenascent metal surface is exposed in a lipidsolution, a monolayer of lipid moleculesis irreversibly bound onto its surface. Theadsorbed lipid monolayer with unattached


BLM

BLM

Cs

Cm

Cm

Rm

Rm

Rn

E

VWE

A−

D+

D

A

[OX]1

[OX]2

y2y1

y2s y1

s

[OX]2

[RED]1[RED]2

[RED]1s

[OX]1s

[RED]2s

(Outside)(Inside)

E 10

BLM

BLMe

BLM

E 20

(SCE)

(a) (b)

(SCE)(RE)0

(SCE)(AE)

E A0E A

ME AI

<100 A

BLM

Fig. 4 Cyclic voltammetry of membranes.(a) Experimental details showing athree-dimensional view of a BLM modified withelectron mediators (such as TCNQ, TTF,ferrocene, iodine, fullerene C60). A

three-electrode system is illustrated. (b) A usualrepresentation of a BLM separating two aqueoussolutions. Electrically, a BLM is shown to beequivalent of a resistor (Rm) and a capacitor(Cm) connected in parallel [1–3, 20–24, 35].

hydrocarbon chains interacts with the hy-drophobic chains of other lipid molecules.A brief explanation of the first step is asfollows. A freshly cut metallic surface hasgreat affinity for lipid molecules. Amphi-pathic lipid molecules interact with thenascent metallic surface because they arehighly hydrophilic. In the second stepof self-assembly process, the lipid-coatedmetallic wire is immersed into a bathingsolution (e.g. 0.1 M KCl). The lipid layeradhering to the metal surface will attenuateto a thin lipid layer under favorable condi-tions. It is believed that the various forcesof interactions among lipid molecules lead

eventually to a spontaneous phase transi-tion to the lowest energetic equilibriumstate, that is, a lipid bilayer. The secondstep is illustrated in Fig. 5, which leadseventually (under favorable conditions) tothe spontaneous formation of an excep-tionally stable, self-assembled lipid bilayer,as shown in Fig. 5.

The precise arrangement and degree ofordering of the lipid molecules in the fi-nal structure shown in Fig. 4 is not knownfor certain. But it seems highly probablethat the bilayer nature of the assemblyis a consequence of the thermodynamicsof free-energy changes at the metal-lipid


−V0

−i

0

+i

V00

Teflon-coated metallic wire s-BLM

Switch

−

+

BLM

sim

ulat

or

ElectrometerFunctiongenerator

Computerprocessing

Rm

Rn

Cm

Cm

Rm

Triangle wave

Mem

bran

e cu

rren

t

−V0

−i

0

+i

V00

Triangle wave

Mem

bran

e cu

rren

t

Cm

Cs(b)

(a)

(i) (ii)

Rm

Rn

Fig. 5 Experimental setup for investigating supported BLMs. (a) Arrangement in details.(b) (i) scan result and its equivalent circuit and (ii) scan result and its representation [3, 35].

surface and at the lipid–aqueous solutioninterface. Our measurements of the elec-trical properties of supported bilayer lipidmembranes (s-BLMs) are consistent withthose of conventional BLMs and closelyrelated systems.

To further improve both the repro-ducibility of obtaining s-BLMs with similarelectrochemical properties and of formingeven longer lifetime of such s-BLMs, amodified version of the original methodof s-BLM formation has been developed.


For the best cutting of Pt wire, we con-structed a miniature guillotine in whichthe sharp knife is moved vertically ontothe wire placed on the flat base. The wireis then cut while immersed in a drop oflipid solution so that the initial contact ofthe newly exposed wire surface is with thelipid solution. The newly cut surface is al-most perpendicular to the length of wire.This lipid-coated wire is then inserted intoa small piece of Teflon tubing, which hasbeen filled with lipid solution. This assem-bled Teflon-covered Pt wire is transferredinto an aqueous bathing solution. Thisprocess is shown in Fig. 4(a). We then letthe system stand for a few minutes orlonger. Supplied by diffusion within thelipid solution, the nascent platinum sup-port surface will attract the polar groups oflipid molecules. In order to promote theprocess of self-assembling of BLM, oneof the following two methods has beenused: (1) by removing glycerol dioleate(GDO) solution from teflon (TEF) sleevesimple using capillary forces (Fig. 4b), or(2) by a piston system, which is shownschematically in Fig. 4(c). The final step isself-adjusting thinning and formation ofthe P-G border (Fig. 4d). During the timeof self-adjusting process, the change in ca-pacitance of a layer was monitored. After 5to 10 min, the capacitance reached a con-stant value and the bilayer structure. Usingthese methods we are able to control both,the rate of the BLM formation process andalso the ratio of BLM to the P-G border.These simple steps allowed us to producestable and reproducible s-BLMs.

16.9.2Methods of Investigation

16.9.2.1 PotentiometryUp to now, we have used mainly theclassical concepts of electrochemistry in

describing the various phenomena asso-ciated with membranes, involving onlyhydrated ions and ions surrounded byother ions (ionic cloud). For ions in thesolution, there is ceaseless motion (or ran-dom walk). Ions get nowhere by randomwalk alone (their time average displace-ment is zero). A net transport of ions ispossible by: (1) the presence of an electricalfield, and (2) a chemical potential gradi-ent. Thus, a net drift is superimposedon the random walk. When ions reachthe interface, properties change abruptly;the anisotropy of the forces compel ionsto adopt new configurations unknown inbulk solutions. Here, a variety of phe-nomena occur such as charge separation,potential gradient, adsorption, and ori-entation of H2O dipole, etc. At presentmodern electrochemistry is focused onthe transfer of charges across the inter-face, in particular, the electrified interface.The involvement of the EDL, therefore, isself-evident.

Of special interest to the above top-ics is the question: ‘‘Experimentally, howdoes one go about measuring membranepotential such as φ(=Em)?’’ One mayplace a pair of identical electrodes, suchas: (1) Ag/AgCl, (2) K+-sensitive glasselectrodes, or (3) saturated calomel elec-trodes (Hg/Hg2Cl2) with salt bridges,across the membrane with fixed negativecharges. When the concentration ratio ofthe solutions bathing opposite sides ofthe membrane is 10 : 1 (say, 0.1 M KClon the outside and 0.01 M KCl inside), anEm of about 116 mV would be measuredfor case (1), about 0 mV for case (2), and58 mV for case (3). Evidently, the choiceof the kind of electrodes used to mea-sure Em is of crucial importance. For BLMand most membrane studies, calomel elec-trodes with saturated KCl bridges havebeen extensively employed. A pertinent


interpretation to the above data is to likenthe membrane as a bipolar electrode ratherthan merely as a permeating barrier.

16.9.2.2 The Patch-Clamp TechniqueTo measure what is happening in a sin-gle, living cell, researchers use a so-calledthe patch-clamp technique (PCT), whichrequires an extremely fine pipette heldtightly against the cell membrane. By care-fully heating and pulling a small glasscapillary tube, a very fine pipette can beformed. When pulled by machine, the tipwill be much smaller than a human hairand the opening on the end of the pipettemay be only 1 µm in diameter. The im-petus for developing the PCT was therealization of the existence of ion chan-nels in biomembranes [3, 4]. In the 1970s,shortly after the single-channel activitieswere demonstrated in BLMs, the PCT wasthen developed. Briefly, the PCT involvesplacing a glass micropipette in contact witha patch of cell membrane so as to isolateone or several ion channels of the sametype. The membrane-coated pipette is thenused as an electrode to record the flowof ions between the inside of fluid-filledpipette and the other side of the mem-brane fragment. As in planar BLMs, tinyelectrical currents at the picoampere (10–9A) level are measured. The pipette is usu-ally filled with a KCl and pressed againstthe cell membrane (or BLM or giant lipo-some). A small suction is then applied sothat a tiny patch of the membrane formsa tight seal around the edge of the pipettetip. Thus, any current through the pipetteprobe must also flow through the patchedmembrane. The most crucial aspect ofthe technique is the formation of a tightseal between the pipette and membranepatch, which is usually on the order ofgigaohms (>1010 ). The technique callsfor an electronic feedback circuit, which

ensures that the voltage applied acrossthe patched membrane is held at a con-stant value, so that the monitored currentsare proportional to changes in the mem-brane resistance. If the patched membranecontains an ion channel, a brief pulse ofcurrent is observed each time the channelopens. The observed pulses are generallyof the same size, monitored as currentsteps, in a given experiment, suggestingthat ion channels are usually either openor closed, although the time for which thechannel is open varies from one openingto the next. From the observed currentsteps, which are on the order of 10−12 A,an opened ion channel allows a translo-cation of a few thousand ions across themembrane per millisecond. It should bepointed out that the openings and closingsof channels are random events; they canbe predicted only in statistical probabili-ties. The PCT provides experimenters witha picture at the molecular level how thou-sands of channels might behave in a givencell membrane. An ion channel consists ofa single protein molecule or a complex ofmolecules, which forms a mutable, water-filled pore through the lipid bilayer of cellmembrane. Some of the ion channels areoperated by an electrical signal (voltage-gated), whereas others are activated bychemical signal (ligand-gated). The chan-nel allows specific ion species to flood intoor out of the cell. It is believed that openingand closing of the channel is achieved by aminute change in the conformation of thechannel protein complex. Mention shouldbe made that prior to the PCT, intracel-lular electrodes used in the voltage-clamptechnique (VCT) measure the transmem-brane potential across the membrane andthrough it [3]. The influences of electricalstimuli or of neurotransmitters are de-scribed in terms of change in membraneconductance. It has been very successful


in accounting and describing the nerveaction potential. However, the VCT ap-proach, as practiced earlier, was incapableof giving direct information at a molecularlevel. In contrast, the PCT brings mem-brane biophysics to the molecular level,which enables electrophysiological studiesto be done on single channels embedded inBLMs, liposomes, or biomembranes [13].

16.9.2.3 Cyclic VoltammetryThe charges that we are concerned withhere are the electronic charges (electronsand holes). For charges of this type tobe transported across the interface, elec-trochemical reactions must take place.In the presence of a membrane that isimpermeable to ions, what will happenthen? Here, the membrane must serve atleast two functions: (1) pathway for elec-tronic charges, and (2) electrode surfacefor chemical transformation (reductionand oxidation or redox reactions). To probesuch a membrane, we will describe awell-established cyclic voltammetry (CV)method of electrochemists, which was ap-plied to the BLM system in 1984 [20–24,32, 36].

CV is an elegant and simple electro-chemical technique for studying redoxreactions at the metal electrode–solutioninterfaces and has become increasinglyemployed in all fields of chemistry. Un-til recently, however, the technique hasapparently not been applied to any mem-brane systems including the BLM separat-ing two aqueous solutions.

16.9.2.4 CV and BLM ExperimentsBefore discussing the results obtained ona BLM system using the powerful CVtechnique, a brief description of the setupused in this type of experiment is in order(Fig. 4). The Teflon cup will be referred

to as the inside, and the other chamberwill be referred to as the outside. We haveused a three-electrode system is used forobtaining voltammograms in the followingconfiguration: one calomel electrode (SCE)is placed in the Teflon cup and two othercalomel electrodes are on the outside.For membrane CV, that is, by applyingthe above voltammetric techniques to theBLM system, a Lucite block containing twoadjacent 2-cm diameter chambers (8 ml),one of which holds a 10-ml Teflon cuphaving a hole on its side, is used.This aperture or hole is most simplymade by punching into the wall of theTeflon cup. (usually less than 0.5 mmdiameter) should be smooth so that thestability of the BLM is enhanced. TheTeflon cup is referred to as the inside,and the other chamber is referred toas the outside. The voltammograms ofthe BLM are obtained using an X-Yrecorder fed by a picoammeter and thevoltage generator (e.g. CHI Workstation,CH Instruments, Austin, TX). The voltagefrom the programmer is applied throughthe potentiometer to the calomel electrode(SCE) immersed in the inside solution.Another SCE immersed in the outsidesolution is connected to the picoammeter.The important feature of the setup is a veryweak dependence of its input voltage onthe current being measured.

The significant advantages of the CVtechnique are, besides its simplicity andthe good precision of measurement, theinvolvement of the lipid bilayer and thecapability for future development. Sincethe electron transfer chain componentsare known to be closely associated with thelipid bilayer, the values thus determinedhitherto by the usual Pt electrode maybe quite different from their actual val-ues in the membrane. Conceivably, the


technique described offers a novel ap-proach to the determination of Eo′

ofmembrane-bound biomolecules such asthe cytochromes and other redox enzymes,using modified BLM as the working elec-trode. Further, this new type of electron-ically conducting BLMs coupled with theCV technique may be useful in the studyof membrane bioenergetics and in thedesigning of molecular electronic devicesbased on ultrathin films. For example, toimpart electronic properties, an organicsemiconductor such as TCNQ (7,7,8,8-tetracyanoquinodimethane) or TTF (tetra-thiafulvalene) has been embedded into theBLM, thereby making the lipid bilayer elec-tron conducting.

In order to apply the powerful CV tech-nique to the BLM system, a conceptualeffort has to be made, that is to con-sider one side of the BLM as the workingelectrode, while the other side is provid-ing the connection to the external circuit.An unmodified BLM behaves essentiallyas an excellent insulator (specific resistiv-ity >1014 ) and does not function as aworking electrode. However, upon incor-poration of TCNQ into the BLM and inthe presence of ascorbic acid in the outersolution and equal molar (0.001 and 0.01)K3Fe(CN)6/K4Fe(CN)6 in the other inside,the observed electrical properties are such(Rm decreasing from 108 to 106 ohms cm2,Cm increasing from 0.4 to 0.5 F cm−2, Em

developing to about 180 mV with the ascor-bic acid side positive. In the absence ofTCNQ in BLM with other experimentalconditions the same, the I/V character-istic is linear; it practically coincides withthe x-axis on the scale used. In the absenceof redox couples in the bathing solution,but with TCNQ in the BLM, the I/V curveremains linear. To show more clearly thatthe redox reactions are occurring acrossthe TCNQ-containing BLM the following

experiment has been carried out. To oneside of a TCNQ-containing BLM, a concen-tration of equal molar ferri–ferrocyanidesolution was added, whereas to the otherside, aliquots of ascorbic acid of knownconcentration were introduced. Concur-rently, the membrane potentials (Em) aftereach addition of ascorbic acid were mea-sured. It is most interesting to note thatthe extrapolated value of Em at the equalmolar concentrations of the two redoxcouples agrees very well with the differ-ence of the two standard redox potentials(Eo′ ∼ 300 mV). The highly asymmetri-cal I/V curves are reminiscent of thoseof a p–n junction diode, which permitsan electron current flow in a forward-bias direction only. In the TCNQ–BLMsystem, oxidation occurs at the mem-brane–solution interface contacting theascorbic acid solution and reduction offerricyanide to ferrocyanide takes place onthe other side of the BLM that has a nega-tive polarity. Implicit in this interpretationis the transmembrane movement of elec-trons via the TCNQ molecules imbeddedin the lipid bilayer; that is, the whole sys-tem has the properties of a p–n junctionwith the TCNQ–BLM acting as a rectifier.

To demonstrate that the TCNQ–BLMbehaves like a semiconductor electrode(e.g. Pt, which is frequently used in CV), acomparative experiment was carried out, inthat cyclic voltammograms of quinhydronewere obtained using either Pt or TCNQ-containing BLM under very similar condi-tions. In particular, the cathodic portionsof the voltammograms are quite alike,thus substantiating that the TCNQ–BLMfunctioned as a working electrode. Inthis connection, a cyclic voltammogramof horse-heart ferricytochrome c was alsoobtained. Thus, in order to be consistentwith the theory of CV, the TCNQ–BLM


must function as a redox electrode in theusual practice of electrochemistry [20–24].

Under normal physiological conditions,there exists a transmembrane potentialdifference of about 50 mV, which meansthat the electrical field is on the orderof 100 000 V cm−1, which is close to thecritical value and may lead to mechanicalrupture of the membrane. In this con-nection, the effect of magnetic fields onthe conductance characteristics of A pro-tein channels embedded in BLMs shouldbe examined in future studies. Thus, adetailed understanding of the electricalparameters of the membrane has greatbiological significance and is importantin a wide scope of applications. Further,the properties of experimental BLMs onthe whole, and the interactions betweentheir components, are better understoodthan those of the more complicated in-tact biological membranes. For example,studies on membrane transport phenom-ena in BLMs provide a frame of referencefor understanding material transport inbiomembranes. As such, biomembranes,including the mitochondrial cristae andphotosynthetic thylakoid membranes thatefficiently transduce energy into the chem-ical free energy on which almost all formsof life depend, are far too complex to be un-derstood in physicochemical terms [1–3].Therefore, the real interest in the conven-tional BLM system lies in its usefulness forthe interpretation of results from biologi-cal membranes. Indeed, studies of BLMsfacilitate the initial testing of working hy-potheses that may generate guidelines fora better choice of appropriate in vivo andreconstituted membrane experiments.

16.9.2.5 Modeling of BLM Behavior byElectrochemical Equivalent CircuitThe possible ionic and electronic mech-anisms of charge transfer across the

membrane interior for their I/V char-acteristics have been investigated. Thesetwo types of mechanism give practicallyidentical voltammogram shapes in thepresence of the same redox couples inthe aqueous phases. However, the ionicmodel differed from the electronic onein the characteristic potentials of peakcurrents, which were shifted along thepotential axis toward more negative (ca-thodic peak) and more positive (anodicpeak) values, when compared to the cor-responding peaks in the electronic model.This is mainly due to the difference inthe resistance of the membrane interior inthese two models. Moreover, both sides ofthe membrane separating the two redoxcouples in aqueous solutions parametersof the membrane has great biological sig-nificance and is important in a wide scopeof applications. Further, the propertiesof BLMs on the whole, and the inter-actions between their components, arebetter understood than those of the morecomplicated intact biological membranes.For example, studies on membrane trans-port phenomena in BLMs provide a frameof reference for understanding materialtransport in biomembranes. As such,biomembranes, including the mitochon-drial cristae and photosynthetic thylakoidmembranes that efficiently transduce en-ergy into the chemical free energy onwhich almost all forms of life depend,are far too complex to be understoodin physico-chemical terms. Therefore, thereal interest in the BLM system lies inits usefulness for the interpretation of re-sults from biological membranes. Indeed,studies of BLMs facilitate the initial testingof working hypotheses that may generateguidelines for a better choice of appropri-ate in vivo and reconstituted membraneexperiments [1–3, 13].


16.9.3Electrochemical Impedance Spectroscopy

For the investigation of the properties ofBLMs, electrical methods have been ap-plied at the very beginning. In additionto the CV technique, other methods suchas electrical impedance spectroscopy (EIS)have been applied. Shortly after the discov-ery of the BLM system, Hanai and Haydonreported the thickness measurement of aplanar lipid bilayer using the impedancetechnique [1–3]. Their results are in accordwith the value obtained on RBC, estimatedby Fricke (see Eq. 1). The impedance tech-nique, nowadays also known as EIS, hassubsequently used by many others. Thebasis of the technique is that a small alter-nating current (AC) of known frequencyand amplitude is applied to the system(e.g. a BLM). The resulting amplitude andphase difference that develop across theBLM are monitored. For a BLM of cross-sectional area (A), and thickness tm, theability of the BLM to conduct and to storeelectrical charges are described by the fol-lowing:

G = σA

tmC = εA

tm(31)

where G and C denote, respectively,the conductance and capacitance of thesystem. σ and ε are constant, representingthe electrical conductance and dielectricpermittivity, respectively. Note that thesetwo elements, G and C, are connected inparallel. Thus, the impedance (Z) of thesystem may be expressed as follows:

z(f ) = 1

G + C(32)

where f is the frequency. Therefore, ameasurement of Z provides estimates ofG and C, and is given by

G = 1

Zcos θ and C = − 1

f Zsin θ (33)

where θ is the phase angle. It can be shownthat Z will disperse as a function of fre-quency, f . The above Eq. (33) reveals thatthe dispersion becomes most conspicu-ous for f greater than G/C. Typical plots,known as Bode plots (Z vs f ; θ vs f )are made, from which the first insights ofa BLM separating two aqueous solutionsconcerning its thickness and infrastruc-ture are obtained. Additional commentsare given below to illustrate the usefulnessof EIS as applied to the BLM system [3, 5,36, 37].

The BLM system, electrically speaking,has two distinct regions, with particularelectrical characteristics. These are: (1)the BLM itself, which behaves as aresistor (Rm) connected in parallel with acapacitor (Cm), and (2) the two contactingBLM–solution interfaces, which can actas an interfacial capacitor (Ci). Whenan alternating voltage is applied betweenthe two reference electrodes, immersedin the bathing solutions, and across theBLM, ions flow toward and away from theelectrodes, depending on their polarity. Atlow frequency, the time the ions flow (inone direction, before reversing their flow),is long enough to drive the ions throughembedded channels, if present, in theBLM, and into the bathing solution on theopposite side. The interfacial capacitance,Ci, may or may not be of consequence;it depends on size when compared withCm. Here, Rm is the controlling factor.At medium frequency, more ions get achance to pass through the channels, ifpresent. Once on the other side, the rateof arrival of ions into the BLM–solutioninterface is slow enough that they caneasily disperse away from their entry pointsat the channel. At medium frequency, the


distinguishing characteristic of the BLMcircuit is the membrane resistance (Rm).At high frequency, the movement of ionsis most affected by the capacitance (Cm) ofthe BLM. It is informative to mention herethe treatment of data as obtained by EIS,using the Bode plot [3].

In an idealized impedance spectrum (ad-mittance or Z vs f ) where admittance isdefined as the rate at which charges arestored in the capacitor per cycle. The be-havior of a pure capacitor, in a Bode plot,forms a line at 45, whereas a pure resistorforms a horizontal line. The Bode plot isa graph derived from plotting frequencyagainst the admittance of an electrical cur-rent. The Bode plot enables us to track theresponses of a diverse electrical system byapplying and following a range of frequen-cies. For example, the phase angle betweenthe electrical excitation and the resultantelectrical flow is measured at a range offrequencies. The significance of the phasediagram is that if one plots phase shiftsover the same range of frequencies thatone uses to generate the Bode plot, onecan detect when the different electricalcharacteristics of the circuit are important.On a phase diagram, the behavior of a purecapacitor shows a 90 phase angle, whilethe more the circuit behaves as a resistorthe more it will show up as approach-ing 0 phase angle. Thus, if one makes acomposite of these graphs, the dip in thephase corresponds to the part of the Bodeplot that looks like a resistor trace. Froman analysis of impedance data, one maybe able to find out specifically the electri-cal behavior of the BLM, which correlateswith the behavior of the modifier (e.g. ionchannel) at Rm, and search out specificallychanges in the impedance/admittance ofthe BLM at the point at which the modi-fier behavior is controlling. The impedancespectra of the base system, as typified by

the so-called Bode plots, including theBLM and its adjacent aqueous solutionsexhibited for all tested cations a single ca-pacitive contribution. The correspondingtime constant was discussed in terms of amodel, which includes a dielectric disper-sion in the BLMs. This phenomenon wasdetected by using either DC or AC volt-age perturbations [35, 36]. Data obtainedusing EIS techniques show two time con-stants, which can be assigned to the lipidbilayer and to the channel/BLM contribu-tions, respectively. In connection to theliposome study using EIS, Karolis andcolleagues have reported low-frequencyimpedance measurements of pure egglecithin (PC) BLMs have revealed thepresence of four layers that can be at-tributed to the acyl chain, carbonyl, glycerolbridge, and phosphatidylcholine regions ofthe lecithin molecule. Measurements onbilayers formed in the presence of unoxi-dized cholesterol revealed that cholesterolmolecules were located in the hydrocar-bon region of the bilayer with its hydroxylgroups aligned with the carbonyl regionof the lecithin molecules. Measurementsof oxidized cholesterol–lecithin BLMs re-vealed that these molecules protruded lessinto the hydrocarbon region and their po-lar hydroxyl group aligned with the glycerolbridge region of the lecithin molecule [13].

Generally, either AC (e.g. EIS) or DCis used for investigating BLMs. Recently,a simple setup for measurements of elec-trical properties of supported planar lipidbilayers (s-BLMs), using a complementaryAC/DC method has been reported [3, 13].The results obtained demonstrated theusefulness of such an approach for study-ing BLMs. The frequency dependence ofresistance and capacitance makes it pos-sible to compare different published dataobtained by AC at different frequenciesor DC. In some experiments, capacitance


increases more sharply with the lower-ing of frequency. As determined, cer-tain s-BLMs modified by antraquinone-2-sulfonic acid (AQS) or TCNQ can transferelectrons readily from the bathing solutionvia the BLM to the supporting Pt surface.Concerning the analysis and interpretationof the experiments, certain assumptionsare made about the dielectric constant(ε) and the thickness of the BLM (tm).For example, a planar lipid bilayer systemaqueous solution|BLM|aqueous solution), isrepresented by the equivalent circuitshown in Fig. 5, in which the capacitance(Cm) has been found to be independentof frequency in the range from DC to ACabout 10 MHz, directly proportional to theBLM area, and apparently dependent onlyon the dielectric constant (ε) layer of thehydrophobic interior of the membrane.The Cm of the lipid bilayer, accordingto the parallel plate condenser equationis followed. Monitoring the BLM forma-tion, thinning, perturbation, and rupturemay be achieved by recording a func-tion of membrane impedance using theequivalent circuit shown in Fig. 5. Essen-tially, a BLM separating two interphases:(aqueous solution|BLM|aqueous solution oraqueous solution|BLM|metal support, oraqueous solution|BLM|hydrogel substrate, isrepresented by three Rm − Cm domainsconnected in series. Each domain has itsown impedance due to a difference in con-stituents and physical attributes. Thus, theBLM domain, owing to its ultrathinnessand low dielectric constant, is predominantover the other two adjacent domains. Forexample, the contribution of the EDL (theso-called Gouy–Chapman layer), whichmay be two orders of magnitude higherthan the lipid bilayer, may be neglectedin a serial arrangement. It is worth notinghere that, when modifiers are present, theymay affect the Rm/Cm signals differently

as well as selectively, which may be fre-quency dependent.

In general, it is simpler to use the DCmethod in assessing the planar lipid bi-layer (BLM) properties. However, it causesa transient response, owing to the Cm.The advantage of the EIS method is thatit permits observation of fast changeswhen investigating interactions betweenthe BLM and its modifiers. It should befurther pointed out that, in using theEIS method, the applied frequency mustbe carefully chosen so that the effect ofcapacitive shunting is minimized. More-over, the selected frequency must not beclose to the line frequency (i.e. 50, 60,or 400 Hz). Some preliminary measure-ments indicate that a frequency around350 Hz is a suitable one to use. Theother advantage with the EIS method isthat membrane/electrode polarization isless likely to occur. Perhaps, a comple-mentary AC–DC mode is a method ofchoice [3]. Insofar as the technique is con-cerned, it seems evident, on the basis ofexperimental findings [3, 37], that it mayopen new vistas in basic membrane bio-physics studies, and offer fresh prospectfor biotechnological exploitation (see be-low on Applications).

It is worth pointing out that, althoughDC measurements have the inherent ad-vantages such as simple instruments andstraightforward interpretation, the dynam-ics of transport processes are not read-ily accessible. Therefore, AC impedancespectroscopy methods have also beenused. For example, a frequency range of0.2–1000 kHz, a capacitance bridge is cou-pled to an external oscillator. Sometimeselectrical connection to the aqueous solu-tions is made via two sheet Pt electrodes(1 cm2) coated with Pt black, or Ag/AgClelectrodes. Impedance spectroscopy, alsoreferred to as EIS, is a well-established


technique for investigating the dynam-ics of membranes including planar lipidbilayers and spherical liposomes. The prin-ciple of the EIS technique is based on thechoice of an appropriate equivalent cir-cuit, which represents the main featuresof the membrane system. The instru-mentation for AC impedance spectroscopyis a gain/phase analyzer (Solartron In-struments, UK) controlled by a personalcomputer.

In recent experiments using the EIStechnique, for example, a BLM-containinggramicidin-A is reported to exhibit twotime constants, one of which can beassigned to the BLM and the otherto the channel BLM contributions. Ear-lier, Yamada et al. have studied electrontransfer with three different redox cou-ples through a BLM-containing TCNQusing AC impedance spectroscopy andclaimed that the impedance spectroscopytechnique can resolve perturbations inthe molecular organization of mem-branes down to 0.1-nm resolution (seeRef. [3]), and may be used to monitorperturbations in the molecular struc-ture due to external influences (e.g.the presence of pharmacologically activemolecules in the external environmentadjacent to the membrane). This tech-nique is being used to study the dielectricproperties of proteins. Also applicableare examinations of the effects of in-tense electric fields on cells and cellmembranes, electro-mechanical proper-ties and stability of biomembranes, elec-trical breakdown phenomena, and elec-troporation impedance spectroscopy (EIS),reported the heterogeneous electron trans-fer rate constants for Co(phen)3

(2+/3+)

and Fe(CN)6(4−/3−) in BLMs saturated

with TCNQ) Related studies of supportedlipid membranes have been reported [13,36–39].

16.9.4Photoelectrospectrometry

In view of the sensitivity and sophisti-cation of spectroscopy techniques, andtheir utility in elucidating certain aspectsof the structure and dynamics of mem-branes, attempts have been made to applyabsorption, fluorescence, and impedancespectroscopy to the BLM system [25, 26, 38,40–42]. Additionally, by combining elec-trical methods with those of spectroscopy,a photoelectric action spectrum of a pig-mented BLM may be obtained. This tech-nique of measuring action spectra of pig-mented membranes, termed photoelectro-spectrometry, is several orders of magnitudemore sensitive than conventional absorp-tion spectroscopy. For applications of elec-trochemical and electro-optical methods,the BLM is a system of choice and has beenso employed. For example, pigmentedBLMs have been studied by photoelectro-spectrometry and shown to be capable oflight-induced electron transfer and redoxreactions (see Fig. 6). Owing to the im-portance of dyes, researchers have investi-gated systematically dye-sensitized BLMs.Redox reactions at the BLM–solution in-terfaces are most conveniently scrutinizedby this technique. Fifty-seven dyes of dif-ferent chemical groups have thus beenstudied. From the obtained voltammo-grams, five types of characteristic curveshave been established. Electronic pro-cesses in, as well as across, the BLMare considered. In the presence of lightof known wavelengths, a number of dye-sensitized BLM have also been investi-gated by photoelectrospectrometry. Ultra-violet light flashes can also induce voltagetransients across BLMs, when aromaticamino acids are absorbed to one side of themembrane. These photoeffects varied withthe chromophore structure, the aqueous


10

8

6

−2 0 2

Log(f )[Hz]

Log(

Z)

[Ω]

4

6

4

2

0 0

0

2

4

6

8

102.5

2.0

1.5

1.0

0.5

0.0

2

4

0

Z'[MΩ cm2]

Z"

[MΩ

cm

2 ]

Z '[MΩ cm2]

Z"

[MΩ

cm

2 ]

t[s]

Cm

[µF

cm

−2]

t[s]

Cm

[µF

cm

−2]

2 4 6 0

00

(c)

(a)

(a) (b)

(d)

(b)

100 200 300 400100 200 300 400

2 4 68 10

6

Fig. 6 (a) Bode plots of the impedance for s-BLM (•) and C60 modified s-BLM ().Impedance was measured in 0.1 M PBS at 0 V DC with respect to open-circuit potential andwith AC frequency between 0.01 Hz and 100 kHz. (b) (i) Nyquist complex plane plot for s-BLMand (ii) C60 modified s-BLM. (c) Time course of membrane capacitance Cm of s-BLM under100 Hz (•), 1 kHz (), and 10 kHz (). (d) Time course of membrane capacitance Cm of C60modified s-BLM under 100 Hz (•), 10 kHz (), and 100 kHz ().

solution salt concentration, pH, and oxy-gen partial pressure. These photoeffectsare attributed to the migration of electri-cally charged photochemical intermediatesin the BLM, and provide a new methodfor studying the effect of UV light on

membranes. When a s-BLM doped withZn-phthalocyanine was excited by light,a voltage and a current were recorded,with the action spectra paralleled closelyto that of the absorption spectrum of thephotoabsorber [13, 26, 40, 43, 44].


MonochromatorXenonAPC lamp

ShutterPolarizer

Collimator

Teflon cup

Glass chamber

BLMLens

Lens

Microscope

Observer

R1YX

R2E

FFPA

RVG

SW

SCE

PCFaraday cage

L S

MX

XL F

Heatfilter

Fig. 7 Setup for studying light-induced effectsin p-BLMs. (a) Setup used for obtaining actionspectra of BLMs containing pigments or dyes.The setup is also useful for investigating thestructure of lipid bilayers [2, 15–19, 40, 41].(b) Arrangement for studying p-BLMs with redox

species in the bathing solutions XL – xenonlamp, L – lens, SCE – saturated calomelelectrode, PC – transparent chamber,SW – switch, FFPA – Amplifier, E – electrometer,and R – recorder [15–19].

16.10Electronic and Charge Transfer Processes inMembranes

Although unmodified BLMs are excel-lent insulators, incapable of either ion orelectron conduction, a suitable modifica-tion makes them ion selective or electronconducting. Presently, it is evident thata variety of compounds can be incor-porated into BLMs giving them designtypes of membrane conductivity. A greatdeal of experimental work has also beendone, which proves that BLMs modifiedwith suitable materials show an electronicconductivity with redox reaction takingplace at both interfaces of the membrane.

With these types of BLMs, an oxidationtakes place on the side of the membranefacing the negative electrode whereas re-duction occurs on the side facing theanode. A transverse movement of elec-trons occurs across the membrane andthe BLM functions essentially as a bipolarredox electrode [1–3, 32].

16.10.1Properties of Iodine–Iodide-containingBLMs

The conduction properties of bilayerlecithin membranes in iodine-containingsolutions have been examined from apotentiodynamic experimental approach.

16.11 Photoelectric Effects in Membranes 551

Voltammetric data obtained by using avariety of forms of iodide ions (derivedfrom charge transfer type interactions)implicated the I3

− ion as the chargecarrier, accounting for the diffusion-limited voltammetric response, whereasthe charge transport of I− seemed to belimited by transmembrane diffusion. Thedata were used to support one of many pro-posed mechanisms for the conductance ofiodide in membranes.

16.10.2Polypyrrole BLMs

Studies were performed on oscillationsacross a polypyrrole-lecithin BLM sepa-rating two electrolyte solutions, one con-taining an electron-acceptor reagent andthe other an electron-donor reagent. Sys-tems studied showed sustained oscilla-tions of electrical potential of about 15 mVwith an interval of about 40 s. The pro-posed mechanism of these oscillations wasbased on the polypyrrole oxidation-dopingreduction-doping process. The equivalentelectrical circuit of the investigated os-cillating system was presented. It wassuggested that the mechanism for this phe-nomenon could be useful in interpretationof some oscillations observed in biologicalsystems [25, 26, 38, 42]. It may be notedthat polypyrrole can be synthesized anddoped in situ, which has been shown capa-ble of electronic conduction. The possibleapplication of the described system to thedevelopment of molecular electronic de-vices has been also suggested.

16.11Photoelectric Effects in Membranes

The use of pigmented bilayer lipid mem-branes (p-BLMs) has yielded a wealth of

information concerning the permeabilityof the lipid bilayer to ions, electrons andholes. The light-driven transmembranecharge separation and electron transferprocesses in pigmented BLMs involvingchlorophylls, porphyrins, and their deriva-tives have been of particular interest tophotobiochemistry and photobiophysics.In this connection, mentioned must bemade of experiments using lipid vesicles.For example, it has been found that thereis a clear difference between transmem-brane electron transfer and transmem-brane molecular diffusion. The p-BLMshave been investigated since the late 1960sas an experimental model for the pho-tosynthetic thylakoid membrane of thechloroplast and the visual receptor mem-brane of the eye. For the investigationof light-induced effects in p-BLMs, atechnique termed photoelectrospectrom-etry has been used, as described in thepreceding section. The photoactive BLMstransduce the light energy into electronmovement across the membrane. Such amodel of a molecular photosensors basedupon BLM structure has been expandedby using the thermotropic liquid crystalmaterials as a BLM. In the thermotropicliquid crystal bilayers, the light not onlyinduces both photocurrent and photovolt-age, but also the change of a membranecapacitance. In this connection, the vari-ous methods for measuring photoelectriceffects in p-BLMs have been comparedand concluded that, insofar as the twomeasured parameters, open-circuit volt-age (Voc) and Vfb, are concerned, bothof which result from asymmetrical chargedisplacements across the BLM followinglight excitation and can provide kinetic in-formation on membrane reactions. Voc,denoting the photovoltage measured un-der current clamped conditions, offersadvantages of detecting photoeffects from


charge displacements that have long timeconstants, and producing data that canbe used to reconstruct (membrane cur-rent) waveforms, which under favorablecircumstances can accurately resolve timeconstants from about 10 ns to 5 s. Asshown earlier, Im with three different timeconstants can be resolved in single Voc. Theresolution of long time constants is limitedby the BLM’s RC time constants, which canbe readily determined on Voc apparatus. Itshould be mentioned that common diffi-culties with Voc and Vfb are twofold: (1) theuse of small diameter BLMs (<0.1 mm),result in increased Rm values, which fur-ther limit resolution of photoelectric effecttime constants and increase noise, and (2)from a single experimental arrangement,it is not easy to distinguish various typesof waveforms. Thus, although no system-atic procedure has yet emerged, the moredetailed information provided by Voc hasproven nonetheless useful.

A p-BLM consists of an ultrathin(∼5 nm) insulating lipid bilayer withchlorophyll or porphyrin molecules sep-arating two redox solutions. For example,the photopotential of more than 150 mVhas been observed across a BLM formedfrom chloroplast has been depicted thata photoactive BLM is similar to that ofa barrier-layer type of photovoltaic cell,and that the observed photoelectric ef-fects have been explained in terms oflight-induced redox reactions on oppositesides of the membrane. In other words,the threshold for light absorption is givenby the embedded pigment and the exci-tons (electron-hole pairs) thus produced bylight absorption are separated to oppositesides of the BLM owing to the presence ofhigh electric field across the lipid bilayer.The separated excitons, that is, electronsand holes, move to opposite sides of the

BLM effecting reduction and oxidation, re-spectively. The characteristics of the actionspectra (either photovoltage or photocur-rent) are determined by its action [1–3,15–19, 25, 26, 41].

16.12Applications

The development of conventional BLMs(black lipid membranes, or planar lipidbilayers) and later s-BLMs and sb-BLMs,have made it possible for the first timeto study, directly, electrical properties andtransport phenomena across a 5-nm ultra-thin lamina separating two aqueous solu-tions. s-BLMs, formed on metallic wires,conducting glasses, and gel substrates,as well as on microchips, possess prop-erties resembling biomembranes. Theseself-assembled, s-BLMs, have opened re-search opportunities in studying hithertounapproachable phenomena at interfaces.As a specific example, in living organ-isms, the specificity is dictated by the DNAmolecule, which carries the organism’s ge-netic master design. Thus, a single-strandof DNA is complementary to its counter-part, thereby forming a double-strandedDNA (dsDNA) – a double helix. It seemsthat the first report relating DNA and BLMwas by Amao and Kumazawa in 1993,who reported the hydrophobic and electro-static effect of basic polyamino acid–DNApolyion complex in BLMs. Since then,other investigators have carried out manyinteresting experiments [45].

Some recent findings demonstrate thepotential for investigating processes atsolid–liquid interfaces. As a result ofthese studies, biomembranes have nowbeen recognized as the basic structureof Nature’s sensors and molecular de-vices. For example, the plasma membrane

16.12 Applications 553

of cells provides sites for a host of lig-and–receptor contact interactions such asthe antigen-antibody binding. To impartrelevant functions in BLMs, a variety ofcompounds such as ionophores, enzymes,receptors, pigments, and so on have beenincorporated. Some of these incorporatedcompounds cause the BLMs to exhibitnonlinear phenomena and photoelectriceffects. A modified or reconstituted BLMis viewed as a dynamic system that changesboth in time and in response to envi-ronmental stimuli. In the past, we werelimited by our lack of sophistication inmanipulating and monitoring such a lipidbilayer system. Today, membrane electro-chemistry is a mature field of research,

as a result of applications of many disci-plines and techniques including interfacialchemistry, electrochemistry, voltage- andPCTs, spectroscopy, and microelectron-ics. We now know a great deal about thestructure of biomembranes, ‘‘ion pumps’’,electroporation, membrane channels, etc.In membrane reconstitution experiments,for example, the evidence is that in-tracellular signal transduction begins atmembrane receptors. It should be reiter-ated that the research area covered in thispaper is highly interdisciplinary. Empha-sis has been placed on basic research. TheBLM research has been benefited by across-fertilization of ideas among variousbranches of sciences. It seems likely thatdevices based on ‘‘smart’’ materials may

GND

GND

Out

S

D

+−

T(See inset)

Inset

MeterV

S

SH

2 O

Aqueoussolution

G

RE

REWE

Fig. 8 Experimental setup for investigating s-BLMs. (a) Equivalent circuit and(b) showing an s-BLM on the gate of a FET device [1–3].


be constructed in the form of a hybridstructure, for example, utilizing both inor-ganic semiconducting nanoparticles andsynthetic lipid bilayers. The biomimetic ap-proach to practical applications is uniqueand full of exciting possibilities. We canglean the design principles from Nature’ssuccessful products and apply them toour research and development from whichadvanced sensors may ultimately depend(Figs. 8, 9, 10).

To date, planar lipid bilayers (s-BLMs) onsolid or hydrogel substrates are of great sci-entific interest and of practical merit due

to their ability to mimic biomembranes.Further, they provide a natural environ-ment for embedding proteins or othercompounds under nondenaturing condi-tions. In respect to practical applications,they permit the preparation of ultrathin,high-resistance films with well-defined ori-entation on metals or semiconductors,and the incorporation of receptor pro-teins into these insulating structures forthe design of biosensors and bioelectronicdevices using electrical methods of detec-tion. It is worth pointing out that, with allother systems the compound of interest is

Teflon coating Teflontubing

Teflon tubing

Metal

(e)

(a) (b)

(f)

(c)

(d)

Tip for s-BLMTip forsb-BLM

Ag/AgCl

Agar gel

Graphicdisplay

5 nm

Workingelectrode

Counterelectrode Pt

ReferenceelectrodeAg/AgCl

To instrument

S-BLM

Agar gel

Hydrogel(e.g., agar, agarose)

Metal(e.g., Pt, stainless steel)

or

P-Gborder

RL

Lens

hn

(Ag/AgCl)

Reference electrode

Fig. 9 Experimental arrangements forconventional and supported BLM studies. A andB are probes, respectively, for metal andhydrogel supported BLM experiments; C setupfor conventional BLM experiments; D setup for

s-BLM experiment using three-electrodesystem [45]; E setup for investigatinglight-induced effects in sb-BLMs; and Fillustrating ligand–receptor contact interactionin s-BLMs [1–3].


Cathode

Fe3+

Ceric ionProteins

Polypeptides

ReceptorsToxinsEnzymes

Channels

Porphyrins

Pigments

EIMFerroceneBiphenolsN-vinylcarbazoleChloranils

ValinomycinATPaseMonensinGramicidinCrown ethers

Glucose oxidaseAntigen

Chlorophyllsb-caroteneIodinePerylene

PhthalocyaninesCrystal violetRetinals (cis, rrans)RhodopsinTCNQ, TTFUbiquinoneMenadione

(nanoparticles)Silver halidesGraphiteFullerenes (C60, C70)

Semiconductors

ATP

Cystine

Methylene blue

Methyl viologen

NADAQS− (Na+)

NADP+

Na+

ThionineKI

Antibody

(left, outside)Extracellular domain

Ascorbic acid (V−c)Hydroquinone

NADH

Cysteine

Napththoquinone

CatechinFerrous ion (Fe2+)

Gentian violet

K+

Sodium dithionite

(right, inside)Intracellular domain

BLMLipid bilayer domain

Photocathode

(reduction)

Anode

PhotoAnode

(oxidation)

Fig. 10 To elicit desired effects, some of the compounds and materials have beenembedded into BLMs (planar lipid bilayers). Also listed are a few of the species in thebathing solutions that have been used in experiments [1–3, 13, 51].

immobilized in a rigid, solid-like structure,whereas in the BLM it is embedded. By em-bedding is meant that the compound(s)(membrane modifiers such as electron

acceptors, donors, mediators, polypep-tides, proteins, etc.) of interest in the lipidbilayer is relatively free to adapt to its sur-roundings. The functions of lipid bilayers


are mediated via specific modifiers, whichassume their active conformations only ina liquid-crystalline environment. Further,the presence of the lipid bilayer greatly re-duces the background noise (interference)and effectively excludes hydrophilic elec-troactive compounds from reaching thedetecting surface causing undesired re-actions. From specificity, selectivity, anddesign points of view, a supported planarlipid bilayer (s-BLM) is an ideal natu-ral environment for embedding a hostof materials of interest. Hence, the s-BLM system offers a wider opportunityfor biosensor development. Many inves-tigators have reported chemical sensingwith modified probes that mimic gating atbiomembranes, incorporating ion-channelreceptors [46–48]. These self-assembled,supported BLMs, not only have overcomethe long-term stability problem of con-ventional planar lipid bilayers, but havealso opened a range of possibilities formanipulating interfacial films as well asin developing practical biosensors. It isbeyond the scope of this chapter evento mention, let alone to describe any ofthese new and exciting developments (e.g.in semiconductor–septum electrochemi-cal photovoltaic – SC–SEP-cell). A num-ber of comprehensive reviews as well assome recent publications are available [3,16, 17, 27, 28, 32, 34, 37, 45–51].

References

1. H. T. Tien, Bilayer Lipid Membranes (BLM):Theory and Practice, Marcel Dekker, NewYork, 1974.

2. A. Ottova-Leitmannova, H. T. Tien, Prog.Surf. Sci. 1992, 41, 337–446.

3. H. T. Tien, A. L. Ottova, Membrane Bio-physics: As Viewed From Experimental BilayerLipid Membranes (planar lipid bilayers andspherical iposomes), Elsevier, Amsterdam,New York, 2000, p. 648.

4. J. R. Harris, A.-H. Etemadi, (Eds.), Artificialand Reconstituted Membrane Systems, Subcel-lular Biochemistry, Plenum Press, New York,1989, Chap. 3, Vol. 14.

5. M. Blank, (Ed.), Electrical Double Layers inBiology, Plenum Press, New York, 1986.

6. B. Ivanov, (Ed.), Thin Liquid Films: Fun-damentals and Applications, Marcel Dekker,New York, 1988, Chap. 14, 15.

7. T. Osa, J. L. Atwood, (Eds.), Inclusion Aspectsof Membrane Chemistry, Reidel Publishing,Boston, Mass., 1991, pp. 191–274.

8. A. A. Marino, (Ed.), Handbook of Bioelec-tricity, Marcel Dekker, New York, 1988,pp. 181–242.

9. M. Blank, E. Findl, (Ed.), Mechanistic Ap-proaches to Interactions of Electric and Electro-magnetic Fields with Living Systems, PlenumPress, New York, London, 1987.

10. G. Milazzo, (Ed.), Topics in Bioelectrochemistryand Bioenergetics, John Wiley and Sons, NewYork, 1983, pp. 157–224, Vol. 5.

11. G. Dryhurst, K. Niki, (Eds.), Redox Chemistryand Interfacial Behavior of Biological Molecules,Plenum Press, New York, 1988, p. 529.

12. A. D. Bangham, BioEssays 1995, 17, 1081.13. H. T. Tien, A. L. Ottova, J. Membr. Sci. 2001,

189, 83–117.14. J. Koryta, Ions, Electrodes and Membranes,

John Wiley and Sons, New York, 1982.15. S. G. Davison, (Ed.), Prog. Surf. Sci. 1985, 19.16. S. G. Davison, (Ed.), Prog. Surf. Sci. 1986, 23.17. S. G. Davison, (Ed.), Prog. Surf. Sci. 1989, 30.18. S. G. Davison, (Ed.), Prog. Surf. Sci. 1992, 41.19. S. G. Davison, (Ed.), Progress in Surface Sci-

ence, Pergamon Press, New York, 1985–1992.20. H. T. Tien, J. Phys. Chem. 1984, 88, 3172.21. H. T. Tien, Bioelectrochem. Bioenerg. 1978, 5,

318.22. H. T. Tien, Bioelectrochem. Bioenerg. 1982, 9,



19.25. M. Blank, (Ed.), Biomembrane Electrochem-

istry, Advances in Chemistry Series No. 235,American Chemical Society, Washington,DC, 1994, Chap. 24.

26. A. G. Volkov, D. W. Deamer, D. L. Tanelianet al., Liquid Interfaces in Chemistry andBiology, John Wiley and Sons, New York,1998.


27. F. L. Carter, R. E. Siatkowski, H. Wohltjen,(Eds.), Molecular electronic devices, Proceed-ings of Third International Symposium, NorthHolland, Amsterdam, 1988.

28. K. L. Mittal, D. O. Shah, (Eds.), Surfactantsin Solution, Plenum Press, New York, 1991,pp. 133–178, Vol. 11.

29. P. Krysinski, H. T. Tien, Prog. Surf. Sci. 1986,23, 317.

30. R. Guidelli, G. Aloisi, L. Becucci et al., J.Electroanal. Chem. 2001, 504, 1–28.

31. R. P. Rastogi, R. C. Srivastava, S. N. Singh,Chem. Rev. 1993, 93, 1945–1990.

32. P. Krysinski, H. T. Tien, A. Ottova, Biotech-nol. Prog. 1999, 15, 974–990.

33. T. Kakiuchi, J. Electroanal. Chem. 2001, 496,137.

34. M. Seitz, E. Ter-Ovanesyan, M. Hausch et al.,Langmuir 2000, 16, 6067–6070.

35. L.-Q. Gu, L. G. Wang, J. Xun et al., Bioelec-trochem. Bioenerg. 1996, 39, 275–283.

36. K. Asaka, A. Ottova, H. T. Tien, Thin SolidFilms 1999, 354, 201–207.

37. H. Gao, J. Feng, G.-A. Luo et al., Electroanal-ysis 2001, 13, 45–53.

38. V. J. Cunnane, D. J. Schiffrin, M. Fleisch-mann et al., J. Electroanal. Chem. 1988, 243,455.

39. Y.-F. Cheng, D. J. Schiffrin, J. Chem. Soc.,Faraday Trans. 1994, 90, 2517.

40. H. T. Tien, (Ed.), Photochem. Photobiol. 1976,24, 95–207.

41. G. J. Kavarnos, Fundamentals of PhotoinducedElectron Transfer, VCH Publishers, New York,1993.

42. J. O’M. Bockris, F. B. Diniz, J. Electrochem.Soc. 1988, 135, 1947.

43. J. Feng, C. Y. Zhang, A. L. Ottova et al., Bio-electrochemistry 2000, 51, 187–191.

44. H. Gao, G. A. Luo, J. Feng et al., J. Elec-troanal. Chem. 2001, 496, 158–161.

45. Y. L. Zhang, H. X. Shen, C. X. Zhang et al.,Electrochim. Acta 2001, 46, 1251–1257.

46. Y. S. Umezawa, Kihara, K. Suzuki et al.,(Eds.), Anal. Sci. 1998, 14, 1.

47. D. Ivnitski, E. Wilkins, H. T. Tien et al., Elec-trochem. Commun. 2000, 2, 457–460.

48. Y. L. Cheng, R. J. Bushby, S. D. Evans et al.,Langmuir 2001, 17, 1240–1242.

49. J.-M. Kauffmann, (Ed.), Bioelectrochem. Bio-energ. 1997, 42, 1–104.

50. C. M. A. Brett, A. M. Oliveira-Brett, (Eds.),Electrochim. Acta 1998, 43.

51. H. T. Tien, A. Ottova, Current Topics inBiophysics 2001, 25(1), 39–60.

559

17Mediated Electron-transferbetween Redox-enzymes andElectrode Supports

Eugenii Katz, Andrew N. Shipway, and Itamar WillnerThe Hebrew University of Jerusalem, Jerusalem, Israel

17.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 561

17.2 Electron-transfer Provided by Diffusional Mediators . . . . . . . . . . . 56217.2.1 Dissolved Enzymes Activated by Diffusional Mediators . . . . . . . . . 56317.2.2 Monolayer- or Multilayer-enzyme Electrodes Activated by Diffusional

Mediators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56417.2.3 Polymer- or Inorganic Matrix-immobilized Enzymes Activated by

Diffusional Mediators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 566

17.3 The Electrical Contacting of Dissolved Enzymes at Mediator-Functionalized Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 566

17.4 Chiroselective Electron-transfer-mediated Biotransformations . . . . . 56817.4.1 Chiral Diffusional Electron-transfer Mediators . . . . . . . . . . . . . . . 56917.4.2 Chiral Monolayer-immobilized Electron-transfer Mediators . . . . . . 569

17.5 The Electrical Contacting of Mediator-modified Enzymes . . . . . . . . 56917.5.1 Dissolved Redox Enzymes Functionalized with Electron-transfer

Mediators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57017.5.2 Monolayer- and Multilayer-enzyme Assemblies Functionalized with

Electron-transfer Mediators . . . . . . . . . . . . . . . . . . . . . . . . . . . . 573

17.6 Polymer- and Inorganic Matrix-bound Enzymes Contacted byCoimmobilized Mediators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 576

17.6.1 The Electrical Contacting of Enzymes in Mediator-functionalizedPolymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 576

17.6.2 The Electrical Contacting of Enzymes in Mediator-functionalizedSol–gel Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 580

17.6.3 The Electrical Contacting of Enzymes in Mediator-containing GraphitePaste Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 582

560 17 Mediated Electron-transfer between Redox-enzymes and Electrode Supports

17.7 The Electrical Contacting of FAD-enzymes by Mediator-functionalizedFAD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 583

17.7.1 Electrical Contacting of Enzymes by Reconstitution of Apo-flavoenzymes with Relay-FAD Cofactor Units . . . . . . . . . . . . . . . . 583

17.7.2 Electrical Contacting of Enzymes by Surface-reconstitution ofApo-flavoenzymes on Relay-FAD-functionalized Electrodes . . . . . . . 583

17.8 The Electrical Contacting of NAD(P)+-dependent Enzymes . . . . . . . 58717.8.1 The Electrochemical Regeneration of NAD(P)+-cofactors . . . . . . . . 58717.8.2 The Electrochemical Regeneration of NAD(P)H-cofactors . . . . . . . . 59017.8.3 The Association of NAD(P)+-dependent Enzymes with NAD(P)+

Cofactors by Covalent and Entrapment Methods . . . . . . . . . . . . . . 59417.8.4 The Integration of NAD(P)+-dependent Enzymes with Monolayer

Arrays of NAD+-cofactor and Redox-catalysts . . . . . . . . . . . . . . . . 597

17.9 Electrical Contacting by Interprotein Electron-transfer . . . . . . . . . . 59917.9.1 Soluble Cytochromes as Electron-transfer Mediators . . . . . . . . . . . 59917.9.2 Heme-protein Monolayers as Electron-transfer Mediators . . . . . . . . 60117.9.2.1 Microperoxidase-11 Monolayers . . . . . . . . . . . . . . . . . . . . . . . . . 60317.9.2.2 Heme-containing De novo Protein Monolayers . . . . . . . . . . . . . . . 60317.9.2.3 Cyt c-aligned Monolayers Associated with Cytochrome Oxidase . . . . 605

17.10 Applications of Enzymes Electrically Contacted by MediatedElectron-transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 606

17.10.1 Biosensors Based on Electrically ‘‘Wired’’ Enzyme Electrodes . . . . . 60617.10.2 Bioelectrocatalyzed Synthesis by ‘‘Wired’’ Enzyme Assemblies . . . . 60717.10.3 Biofuel Cells Based on ‘‘Wired’’ Enzyme Assemblies . . . . . . . . . . . 607

17.11 The External Control of the Electron-transfer Process . . . . . . . . . . . 61017.11.1 Photochemical Control by Enzyme-bound Photoisomerizable Units . 61017.11.2 Photochemical Control by Electrode-bound Photoisomerizable Units 61417.11.3 Photochemical Control by Mediator-bound Photoisomerizable Units 616

17.12 Conclusion and Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . 617Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 618References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 618

561

17.1Introduction

The redox-active component of most re-dox enzymes is encapsulated deep insidethe enzyme structure. This spatial isola-tion from the bulk environment helps inthe chemical and electrochemical isola-tion of the redox-center, allowing it tobe out of equilibrium with the protein’ssurroundings. The imbalances that resultfacilitate the selective and directed chem-ical and electrochemical processes thatsustain life. In an effort to gain under-standing of these processes, the distancedependence of electron-transfer rates inproteins has been extensively studied,both experimentally [1–7] and theoreti-cally [8–14]. According to electron-transfertheory, the electron-transfer rate constantbetween a donor–acceptor pair is givenby Eq. (1), where Go and λ correspondto the free energy and reorganization en-ergy accompanying electron-transfer andd0 and d are the Van der Waals distanceand actual distance separating the donorand acceptor centers. If one views the elec-trode and the enzyme redox-center as adonor–acceptor pair, it becomes clear thatthe thick protein layer surrounding theactive center provides an effective kineticbarrier to electron-transfer.

ket ∝ e[−β(d−d0)]e[−(Go+λ)2/4RT λ] (1)

The electrochemical insulation of theenzyme-active site by its protein or glyco-protein shell usually precludes the possi-bility of any direct electron-transfer withbulk electrodes [15]. However, under care-fully controlled conditions, some enzymescan exhibit direct, nonmediated electricalcommunication with electrode supports,and biocatalytic transformations can bedriven by these processes [16, 17]. Forexample, the direct electroreduction ofO2 and H2O2 biocatalyzed by laccase [18]and horseradish peroxidase (HRP) [19], re-spectively, have been demonstrated. Thisunusually facile electronic contacting isbelieved to be the consequence of incom-pletely encapsulated redox centers. Whenthese enzymes are properly orientatedat the electrode surface, the electrode-active site distance is short enough forthe electron-transfer to proceed relativelyunencumbered. Direct electron commu-nication between enzyme-active sites andelectrodes may also be facilitated by thenanoscale morphology of the electrode.The modification of electrodes with metalnanoparticles allows the tailoring of sur-faces with features that can penetrate closeenough to the enzyme active site to makedirect electron-transfer possible [20, 21].

The electrical contacting of redox en-zymes that defy direct electrical communi-cation with electrodes can be established


by using synthetic or biologically ac-tive charge-carriers as intermediates be-tween the redox-center and the electrode.These artificial electron donor or acceptormolecules (in case of reductive or oxidativeenzymes, respectively), usually referred toas electron-transfer mediators, can be ac-cepted by many redox enzymes in placeof their natural oxidants or reductants [22].They have a wide range of structures,and hence properties, including a rangeof redox potentials. The redox potentialof a suitable mediator should provide anappropriate potential gradient for electron-transfer between an enzyme active site andan electrode. The redox potential of themediator, Eo′

M, should be more positive ormore negative than the redox potential ofthe enzyme-active site, Eo′

E , in the case ofoxidative (Eo′

M > Eo′E ) and reductive (Eo′

M <

Eo′E ) bioelectrocatalysis, respectively. In the

process of shuttling charge between theredox-center and the electrode, the me-diator is cycled between its oxidized andreduced states. The mediator should bestable in both the reduced and oxidizedforms and any side reactions between themediator redox states and the enzyme orthe environment should be eliminated. Tobe effective in its role, the mediator mustoften compete with the enzyme’s naturalsubstrate (e.g. molecular oxygen in caseof oxidases), effectively and efficiently di-verting the flow of electrons to and fromthe electrode. An efficient mediator shouldprovide rapid reaction with the redox en-zyme, effectively oxidizing or reducing theenzyme-active center. A mediator shouldalso exhibit reversible electrochemistry (alarge rate constant (ket) for the interfacialelectron-transfer at the electrode surface).

The total efficiency of the electrontransport provided by mediators dependsnot only on the mediator properties, butalso on the whole system architecture. A

mediator that cannot compete with thenatural dioxygen electron acceptor when itis functioning via a diffusional route couldbe very efficient when it is included into anorganized supramolecular assembly andoperates in a nondiffusional mode. Sev-eral approaches have been developed andmany bioelectrochemical systems havebeen designed to enhance the electricalcontacting of redox enzymes at electrodesurfaces by the use of electron medi-ators. These methodologies range fromthe simple application of soluble enzymeswith diffusional electron mediators to sys-tems with very sophisticated biomoleculararchitectures composed of numerous com-ponents. Common to all these systems isthe application of a multistep mediatedelectron-transfer (MET) process, each stepof which proceeds over a short distance.The development of these efficient electri-cal contacting methodologies has resultedin the construction of numerous amper-ometric biosensors and bioelectrocatalyticsystems including bioreactors and biofuelcells. The contacted enzyme may also beused as part of a more complex assem-bly – the triggering of any electron-transferstep by an external signal (e.g. light) resultsin the control of the biocatalytic activity.

The aim of this paper is to review thevarious systems in which redox enzymesare electrically contacted via a MET processincluding those in which the mediatingprocess is triggered by external signals. Allpotentials in the text and figures are quotedversus saturated calomel electrode (SCE).

17.2Electron-transfer Provided by DiffusionalMediators

MET can be affected by a diffusionalmechanism (Fig. 1) in which the electron

17.2 Electron-transfer Provided by Diffusional Mediators 563

Fig. 1 The electrical ‘‘wiring’’of an oxidative redox enzyme viaa diffusional electron-transfermediator shuttling between theenzyme active center and theelectrode.

Substrate

Product

e−

Transduced current

R

R+ Enzyme

relay is either oxidized or reduced atthe electrode surface. Diffusional pene-tration of the oxidized or reduced relayinto the protein yields a sufficiently shortelectron-transfer distance for the electricalactivation of the biocatalyst. Penetrationof the mediator close to the enzyme-active center inside the protein matrix iscontrolled by the hydrophobic/hydrophilicproperties of the mediator and the enzyme,the size and shape of the mediator, and theelectrostatic charge interactions betweenthe mediator and the enzyme. The me-diators can diffusionally shuttle electronsbetween the electrode and enzymes in sev-eral configurations: soluble, immobilizedas monolayers (or multilayers), or incor-porated into porous matrices. In all thesecases, the medium needs to provide a freediffusion pathway between the conductivesupport (providing electrochemical regen-eration of the mediator) and the enzymemolecules (working as biocatalysts).

17.2.1Dissolved Enzymes Activated by DiffusionalMediators

Ferrocene derivatives, organic dyes, fer-ricyanide, Ru-complexes, and other elec-trochemically active substrates have beenemployed for diffusional MET and the elec-trical activation of soluble redox enzymes

lacking direct electrical contact with theconductive support [22]. Cyclic voltamme-try (CV) allows to assess the effectivenessof a particular enzyme/mediator combina-tion using the theory for the catalytic elec-trochemical processes. The second-orderrate constant for the reaction between theenzyme and the mediator can also be de-termined [23–25]. Soluble redox enzymeselectrically contacted by the use of dif-fusional electron-transfer mediators withvarious redox potentials (Eo′

M) providingdifferent rate constants of electron-transfer(kEM) have been extensively reviewed [22].

Comparison of the electron-transfer ef-ficiency provided by different mediatorsin the presence of the same redox en-zyme allows the definition of the im-portant parameters of MET. The reac-tion of glucose oxidase (GOx) has beenextensively studied with a number ofartificial electron acceptors including or-ganic dyes such as phenazine methosul-fate, 2,6-dichlorophenolindophenol, andN, N, N ′, N ′-tetramethyl-4-phenylenedia-mine [26]. However, these mediators havea number of limitations such as poorstability and the pH dependence oftheir redox potentials. Simple inorganicredox species such as hexacyanofer-rate [27], hexacyanoruthenate, and pen-taamine pyridine ruthenium [28] donot suffer from these problems. These


inorganic compounds have almost idealelectrochemistry and are more stable thanthe organic dyes. The application of inor-ganic mediators has been exemplified withother oxidases such as sarcosine oxidaseand lactate oxidase [29]. Inorganic media-tors are difficult to ‘‘tune’’ for solubilityand electrochemical properties though, asthey cannot be modified or derivatized aseasily as their organic counterparts.

The majority of these problems havebeen overcome by the use of ferrocenederivatives as electron acceptors for solubleoxidases (e.g. GOx). Ferrocene derivativesof varying charge and solubility, with re-dox potentials (Eo′

M) between +0.1 V and+0.4 V, have been shown to accept elec-trons from GOx [30]. The published val-ues [22] for the second-order rate constant(kEM) for the reaction of the reduced activecenter of GOx (FADH2) and an oxidizedferrocene derivative range from 0.26 × 105

to 5.25 × 105 M−1s−1. There is no simplecorrelation between kEM and Eo′

M; how-ever, it is clear that positively chargedferrocenes are favored for the mediatedelectron transport from GOx [31]. Thiseffect originates from the electrostatic at-traction between the positively charged ox-idized mediator and the negatively chargedGOx. Comparison of the mediating ef-ficiency of charged electron relays andcharged enzymes should always take elec-trostatic interactions into account [32–34].The size of the mediator is also importantand it has been shown that ferrocenesinside cyclodextrin cavities do not me-diate electron transport from GOx [35].To improve the contact between elec-tron relay molecules and redox enzymes,micellar systems composed of ferrocene-functionalized surfactants have been ap-plied [36–38]. GOx has been electricallycontacted using such redox-active micellesand an advantage in the electron-transfer

efficiency over homogeneous electron-relays was demonstrated.

17.2.2Monolayer- or Multilayer-enzyme ElectrodesActivated by Diffusional Mediators

Monolayers and multilayers of redoxenzymes (e.g. GOx [39], bilirubin oxi-dase [40]) have been organized on elec-trode surfaces using bifunctional reagents(producing covalent bonding between thelayers) [39, 40] or using bioaffinity interac-tions (for instance, biotin/avidin [41–43]or antigen/antibody [44, 45]). The enzymecontent in monolayers is low, however,and electrical contact in the presenceof a diffusional mediator does not usu-ally result in a detectable amperometricresponse. Thus, an increase of the en-zyme content is essential to obtain thedetectable current when diffusional medi-ators are applied. The stepwise depositionof a multilayer assembly results in theincrease of the enzyme content, leadingto a significantly larger current. The de-position of a controllable number of theenzyme layers also allows the ‘‘tuning’’of the enzyme electrode–amperometricoutput [46].

The enzyme content of monolayerassemblies may also be increased by theapplication of rough electrode surfaces.Treatment of Au surfaces with Hg resultsin a roughening of the conductive supportby the generation and dissolution of anAu-amalgam [47]. Typically, Au surfaceswith initial roughness factor of 1.2–1.5can be roughened to exhibit a roughnessfactor of 15–25. Multilayers of GOx werelinked to smooth and rough Au electrodesby coupling to cystamine-functionalizedsurfaces, and ferrocene monocarboxylicacid was applied as a diffusional mediatorto contact the enzymes [46]. The resulting

17.2 Electron-transfer Provided by Diffusional Mediators 565

currents of a four-layer roughened GOxelectrode are approximately sixfold greaterthan those for a smooth electrode.

The organization of ordered enzymemultilayer arrays opens the possibility ofassembling multilayers of two or moredifferent enzymes [46], one of which is aredox enzyme electrically contacted via dif-fusional MET. The other enzyme(s) mayturn the analyte substrate into a productthat is the substrate for the redox en-zyme. The resulting transduced current istherefore proportional to the concentration

of analyte. A multilayer enzyme electrodeconsisting of choline oxidase (ChO) andacetylcholine esterase (AChE) was tailoredaccording to this concept (Fig. 2a) [46].The analyte acetylcholine is hydrolyzedto choline in the presence of AChE andthe resulting product is bioelectrocatalyt-ically oxidized by ChO to betaine in thepresence of 2,6-dichlorophenolindophenol(1) as a diffusional mediator. The currentdeveloped by the biocatalytic electrode isproportional to the concentration of theprimary substrate acetylcholine (Fig. 2b).

Acetylcholine[mM]

I[µ

A]

0 20 40 600

0.4

0.8

(a)

(b)

O

Cl

Cl

N

ChO AChE

Acetylcholine

Choline

Betaine

e−

Transduced current

AChEChO

e−1

Fig. 2 (a) A bienzymatic network consisting of ChO and AChE forthe amperometric detection of acetylcholine; (b) dependence of theelectrocatalytic current produced by the assembly on theconcentration of acetylcholine. Applied potential 0.3 V, with 40-µM2,6-dichorophenolindophenol (1) as the electron mediator.


17.2.3Polymer- or Inorganic Matrix-immobilizedEnzymes Activated by Diffusional Mediators

Redox enzymes may also be electri-cally contacted with electrode supports bytheir entrapment in an electropolymerizedfilm [48–52]. One of the original motiva-tions for this approach was the possibilitythat direct electrical contact with the en-zyme by the conducting polymer might bepossible, but this question remains un-resolved. Although direct oxidation hasbeen claimed in some conductive polymers(e.g. polypyrrole (PPy) [52]), the balanceof the evidence indicates that if there isany direct electrical contact, the effect issmall. It is possible to enhance the elec-trical communication using diffusionalmediators shuttling electrons betweenthe conductive electrode support andthe enzymes incorporated into the poly-meric film. N -Methyl phenazinium [48],benzoquinone [48, 53], hydroquinonesul-fonate [54], and ferrocene monocarboxylicacid [50, 51] have been used for this pur-pose. The polymer porosity is an importantissue for this kind of enzyme electrical‘‘wiring’’ as it provides the route for themediator to transport the charge.

The electrical ‘‘wiring’’ of redox enzymesentrapped in inorganic (e.g. sol–gel) ma-trices can also be achieved by the ap-plication of solution-state electron-relaysthat can penetrate through the matrix.For example, HRP films deposited incarbon paste substrates and then cov-ered with a thin silica sol–gel layerhave been used to detect H2O2 in aque-ous media [55] and organic peroxides innonaqueous solutions [56] by the use of[Fe(CN)6]4− or ferrocene mediators, re-spectively. The extension of this electrodeconfiguration to other enzymes has ledto the development of new amperometric

sensors for glucose [57] and phenolic com-pounds [58].

It should be noted that all the systemsdiscussed above operate using low molec-ular weight, soluble, mediators that canleach out the electrode and be lost tothe bulk solution. This is a significantdisadvantage, particularly for in vivo appli-cations. Practically important systems canbe designed using similar immobilizationapproaches, but with the critically impor-tant difference that the mediator moleculesare also immobilized in the assembly andprovide the electron transport via a nondif-fusional route (Sect. 17.6).

17.3The Electrical Contacting of DissolvedEnzymes at Mediator-FunctionalizedElectrodes

The electrochemical contacting of solution-state enzymes with surface-immobilizedredox-mediators is of interest for study-ing the interfacial association affinityinteractions between enzymes and me-diators. The electrochemical kinetics ofelectrodes functionalized with layers ofvarious mediators (e.g. viologens [59], C60-derivatives [60], microperoxidase-11 [61])have been studied upon their interactionwith diffusionally free enzymes. Somerelay-functionalized interfaces demon-strate MET without a significant amountof association of the enzyme moleculeswith the relay-modified surfaces. For ex-ample, a C60 monolayer-functionalizedAu-electrode has been applied to contactsoluble GOx (Fig. 3a) [60]. CV revealed anefficient electrocatalytic current developedby the C60-modified electrode in the pres-ence of GOx and glucose (Fig. 3b) [60],dependent on the glucose concentration(Fig. 3b, inset). However, kinetic analysis

17.3 The Electrical Contacting of Dissolved Enzymes at Mediator-Functionalized Electrodes 567

a

b

c

d

−2

0

2

4

6

8

(a)

(b)

−0.2 0 0.4

E[V]

0.2 0.6

0 50 100 1500

2

4

I cat

[µA

]

Glucose[mM]

GOx

SHN

Glucose

Gluconicacid

e−

e−

C60

OS NH2 C60

OHO

I[µ

A]

Fig. 3 (a) Assembly of a C60 monolayer-functionalized-electrode for the bioelectrocata-lysed oxidation of glucose. (b) Voltammograms of the C60 monolayer electrode in thepresence of GOx (2 mg mL−1) and glucose at (a) 0, (b) 20, (c) 40, and (d) 100 mM. Mea-surements were performed in 0.1 M-phosphate buffer, pH 7.1, under Ar, at a scan rate of5 mV s−1. Inset: calibration curve for the concentration dependence of the catalytic current.

performed using a rotating disk electrodemodified with the C60-monolayer showedno temporary association of GOx withthe relay-modified surface. This observa-tion suggests that the interaction of theenzyme with the relay-interface and theresulting electron-transfer proceeds in ashort timescale and it is not the limitingstep in the bioelectrocatalytic process.

In the case of surfaces modified withbiological mediators such as microperoxi-dase-11 (MP-11) (2), the temporary as-sociation of enzyme molecules to therelay-functionalized electrode surface has

been demonstrated. The MP-11 monolay-ers can mediate electron-transfer to hemo-proteins (e.g. cytochrome c (Cyt c), myo-globin (Mb), and hemoglobin (Hb)) by theformation of affinity complexes at the inter-face [61]. The heme-containing monolayercan also stimulate MET to cytochrome-dependent enzymes. The electrocatalyticreduction of nitrate ions was achievedwhen the dissolved cytochrome-dependentnitrate reductase (NR) was electrically con-tacted by a monolayer of MP-11 (Fig. 4a).Figure 4(b) shows the cyclic voltammo-grams of the MP-11 monolayer-electrode


N N

NNFe

CH3

O OH

CH3

H3C SH3CS

CysAlaGln

HisThr

Val

Glu

LysGln

Val

Cys

N N

O OH

H3C

OHO

SHN

O

MP-11 NR

NO3−

e−e−

a

b

c

−0.6 −0.4 −0.2 0

−8

(b)

(a)

−4

I[µ

A]

E[V]

0

MP-11

NO2−

2

S NH2

Au

HO

O

MP-11

Fig. 4 (a) Bioelectrocatalytic reduction of nitrate by NR mediated by a monolayerimmobilized MP-11; (b) cyclic voltammograms of an MP-11-modified electrode (a) in0.05 M phosphate buffer, pH 7.0, (b) after the addition of NR (24 µM), and (c) after thesubsequent addition of KNO3 (5 mM). Recorded under Ar, potential scan rate 5 mV s−1.

(curve a) and with added NR (curve b).With NR, an electrocatalytic cathodic cur-rent is observed indicating the mediatedelectrochemical reduction of NR. Additionof NO3

− to the system results in a substan-tial enhancement in the cathodic current(curve c) resulting from the electrochem-ical reduction of the nitrate biocatalyzedby NR. The Michaelis-Menten analysis ofthe cathodic electrocatalytic currents devel-oped in the presence of different concen-trations of NO3

− results in the conclusionthat NR produces a temporary complexwith the MP-11-modified interface and theMET proceeds within this complex.

As these systems operate in thepresence of soluble enzymes, they havelittle potential for practical purposes.

However, the study of these systemshas revealed temporary affinity associationof the enzyme molecules with themediator-functionalized electrode surfaces(e.g. MP-11). This, in turn, has allowed thedevelopment of integrated biocatalytic sys-tems composed of mediators and enzymesby the lateral cross-linking of the enzyme-mediator affinity complex generated atthe electrode support (Cf. Sects. 17.8 and17.9.2).

17.4Chiroselective Electron-transfer-mediatedBiotransformations

The MET process involves the penetrationof an electron-relay into a protein matrix to

17.5 The Electrical Contacting of Mediator-modified Enzymes 569

reach a position close to the enzyme activesite. As the protein structure is chiral, itmight be expected that the ability of twoenantiomeric mediators to reach the activesite, affect efficient electron transfer, andexit back into the bulk solution could besignificantly different.

17.4.1Chiral Diffusional Electron-transferMediators

Chiroselective MET has been addressed ina series of studies using diffusional chi-ral electron mediators [62–64]. The enan-tiomeric electron relays (S)- and (R)-N, N -dimethyl-1-ferrocenyl-ethylamine were re-ported to stimulate chiroselective bioelec-trocatalyzed oxidation of glucose in thepresence of GOx [62]. The bioelectrocat-alyzed oxidation of glucose was approx-imately twofold enhanced in the pres-ence of the (S)-isomer as compared with(R)-isomer. Kinetic analysis of the bio-electrocatalyzed oxidation of glucose bythe two enantiomeric electron relays per-formed using a rotating disk electrodesuggests that the protein induces chiraldiscrimination because of diastereoiso-meric interactions during the mediator’spenetration and dissociation pathways.Chiral discrimination in MET and subse-quent bioelectrocatalyzed transformationshas also been observed for other diffu-sional electron relays and enzymes [63, 64].

17.4.2Chiral Monolayer-immobilizedElectron-transfer Mediators

The concept of enantioselective electri-cal contacting of redox enzymes andelectrode surfaces was further devel-oped by the organization of a chiral

electron-transfer mediator as a monolayeron an electrode surface [65]. The (R)- and(S)-2-Methylferrocene carboxylic acids (3and 4) were assembled as monolayers onAu-electrodes (Fig. 5a). The monolayer-mediated oxidation of glucose in thepresence of GOx showed an approximately1.9-fold enhancement at the electrodefunctionalized by the (S)-enantiomer (4)over the electrode functionalized by the(R)-enantiomer (3) (Fig. 5b). Rotating diskelectrode studies revealed that the METproceeds via the formation of a com-plex between the mediator monolayer andGOx. The electron-transfer rates withinthe complex from GOxred to the (3) and (4)differ substantially, 0.3 s−1 and 0.5 s−1, re-spectively, whereas the Michaelis-Mentenconstants characterizing the complex areidentical, KM = 1.2 × 10−5 M. Thus, theaffinity for the association of GOxred tothe chiral ferrocenyl cation monolayers aresimilar. However, the diastereoisomericnature of the complex yields differentelectron-transfer distances between the ox-idized mediator and the reduced flavinsites, thereby stimulating the differentelectron-transfer rates.

17.5The Electrical Contacting ofMediator-modified Enzymes

The chemical modification of redox en-zymes with electron relay groups canenable nondiffusional MET, often referredto as the electrical wiring of the pro-teins [66–69] (Fig. 6a). The covalent attach-ment of electron-relay units at the proteinperiphery and inner sites yields shortinterrelay electron-transfer distances. Elec-tron ‘‘hopping’’ or tunneling betweenthe periphery and the active site en-ables electrical communication between


(b)

(a)

SOO N

O

O 2O

O N

O

O

S

O

HN

S (CH2)5 NH2

H2N (CH2)5 NH2

Fc CH3

O

HO O

S HN

(CH2)5 NH

OFc CH3

EDC GOx

Glucose

Gluconicacid

e− e−

Au

3 or 4

2.0

1.5

1.0

0.5

0

b

a

0 0.2 0.4

E[V]

I[µ

A]

Fe

Fe

HO

O

HO

O

CH3

H3C

3

4

Fig. 5 The enantioselective bioelectrocatalyzed oxidation of glucose by GOx at an electrodemodified by a chiral electron-transfer mediator: (a) organization of the chiral ferrocenemonolayer-modified Au-electrode and its interaction with soluble GOx. (EDC = 1-(3-dime-thylaminopropyl)-3-ethylcarbodiimide hydrochloride); (b) cyclic voltammograms of theferrocene-modified-electrode (curves a and b for R-Fc (3) and S-Fc (4), respectively) in thepresence of 1 × 10−5 M GOx and 50 mM glucose. Recorded in 0.1 M phosphate buffer, pH 7.0,under Ar, potential scan rate 5 mV s−1, electrode area 0.26 cm2.

the redox enzyme and its environment.The simplest systems of this kind in-volve electron relay-functionalized en-zymes diffusionally communicating withelectrodes [66, 67], but more complex as-semblies include immobilized enzymes onelectrodes as integrated assemblies [70].

17.5.1Dissolved Redox Enzymes Functionalizedwith Electron-transfer Mediators

GOx has been covalently modifiedwith ferrocene electron-relay groups bythe carbodiimide coupling of ferrocene


Enzyme

Substrate

Product

R

R

R

e–e–

Fe

N (CH2)10

N

NNNN

Os

N

OH

O

Fe NH

(CH2)n NH2

N

RuN Cl

ClNN

H2N Lys

H2N Lys

His

H

(a)

(b)

(c)

(d)

(e)

(f)

O

HN

N

GOx

GOx

GOx

GOx

+

+

+

+

EDC

EDC

Lys

Lys

His

N

HN

N

GOx

GOx

GOx

GOx

Fe NH

O

N (CH2)10

N

NNNN Os

N

NH

O

Fe NH

(CH2)n

N

RuN

ClNN

O

OH

NH3

RuH3N H2O

NH3H3NNH3

+GOx

N

HO

NN

GOx

N

HO

NNNH3

RuH3N

NH3H3NNH3

n = 2, 3, 4, 6, 8, 10

5

6

7

8

Fig. 6 (a) A soluble enzyme electricallycontacted with an electrode surface viaredox-mediator groups covalently tethered to theprotein backbone. (b)–(f) Different kinds ofredox-mediators tethered by different means tothe protein backbone; (b) a ferrocene unit linkedto a lysine amino group via an amide bond;(c) an Os-complex linked to a lysine amino

group via an amide bond; (d) a ferrocene unitlinked via a Schiff-base to an aldehyde groupgenerated on the glycoprotein by oxidation withIO4

−; (e) Ru(bpy)2Cl2 linked to histidineresidues by the formation of a Ru(bpy)3His-complex; (f) a Ru-complex linked to apyridine group by ligand exchange.

carboxylic acid [66, 67] (Fig. 6b) or os-mium (bis-4,4′-diaminobutane-N ,N -bis-2-pyridyl-11-aminoundecanoic acid) (5) [36](Fig. 6c) to the amino functions of the

protein backbone. Other chemical proce-dures have also been used to produce relaygroups peripherally bound to proteins.For example, carbonyl functions generated


in glycoproteins (e.g. GOx) by oxida-tion with periodate have been coupled toan amino-functionalized ferrocene (6) viaSchiff bases (Fig. 6d) [68]. In a different ap-proach, ruthenium complexes (7, 8) wereformed on GOx using either natural his-tidine residues [71] (Fig. 6e) or artificially

introduced pyridine groups [67] (Fig. 6f) inthe protein backbone. The randomly dis-tributed ferrocene relay groups providedelectrical contact between the GOx redoxcenter and the unmodified electrode sur-face. It was also demonstrated that longerspacer groups bridging the electron relay

b

c

a

4

(a)

(b)

3

2

1

00 20 40 60 80 100

t[min]

[GS

H] ×

105

[M]

O

O N

O

O

S

Au

LysH2NO

SGR

LysNH

GR

N+

N+

(CH2)nOOH

O

SGR

LysNH

NH2Lys

NH2Lys

HNLys

n = 2, 5, 11

EDC

N+

N

(CH2)nO

GSSG

GSH

e−

e−

Glu Cys Gly

SH

Glu Cys GlySS

CysGlu Gly

GSSG

GSH

9

+

Fig. 7 (a) The assembly of an electrically contacted glutathione reductase monolayer on anelectrode; (b) the rate of bioelectrocatalyzed reduction of GSSG by the electrically contactedenzyme-electrode using chain-lengths of (a) n = 2, (b) n = 5, (c) n = 11 and upon theapplication of a potential corresponding to −0.72 V on the enzyme-electrode in the presence ofGSSG (10 mM). The experiments were performed in 0.1 M phosphate buffer, pH 7.2, under Ar.


groups and the enzyme provide highermobility, shorten the electron-transfer dis-tance, and thus enhance the enzyme’sbioelectrocatalytic activity [68]. Partial andreversible unfolding of the enzyme (usingurea) during the covalent modification ofthe protein with the relay is also importantas it allows the attachment of the electronmediator to inner positions close to theactive site of the enzyme [72].

While an increased loading of anelectron mediator on a protein enhancesthe effectiveness of electrical contacting,the enzyme activity suffers because ofchanges brought about in its structure. Thechemical modification of redox proteinswith synthetic electron mediators is alwaysaccompanied by this partial denaturingof the native biocatalyst, and so themodification must be carefully controlledto achieve the optimum effect. WithGOx modified by ferrocene units, theoptimal MET was found at a loadingcorresponding to 12–13 electron relayunits [72]. The rate constant of electron-transfer between the enzyme’s flavinadenine dinucleotide (FAD) site and thenearest electron relay group was foundto be approximately 0.9 s−1 (substantiallylower than that for electron-transfer to thenative dioxygen electron acceptor of theenzyme, proceeding with a rate constantof ca. 5 × 103 s−1) [72]. Improvements inthe efficiency of electrical contact may bepossible by the use of protein engineering,genetic manipulation, or relay units placedat optimal positions in the enzymestructure.

17.5.2Monolayer- and Multilayer-enzymeAssemblies Functionalized withElectron-transfer Mediators

Enzyme monolayers can be electricallycontacted by using electron mediators

covalently attached to the protein [73–75].In this case, electron mediators must beavailable close enough to both the en-zyme active site and the electrode tofacilitate an efficient electron-transfer pathbetween them. A glutathione reductasemonolayer assembled onto Au-electrodesthrough a thiolate bridging unit (Fig. 7a)was functionalized with N -methyl-N ′-carboxyalkyl-4,4′-bipyridinium salts (9),where the chain length tethering theelectron relay was systematically length-ened [73, 74]. Electroreduction of thebipyridinium unit was found to activatethe enzyme for the reduction of oxidizedglutathione (GSSG). The effectiveness ofGSSG reduction was controlled by thechain length tethering the electron relayto the protein, and longer bridging chainsenhanced the electrobiocatalyzed reduc-tion rate (Fig. 7b). This phenomenon wasrationalized in terms of shorter electron-transfer distances between the electron-relay sites and the enzyme redox-centerfor the systems with long-chain tetheredbipyridinium units. It was also found thatin order to attain electron-transfer commu-nication between the enzyme redox-centerand the electrode, it is important to modifythe protein with the bipyridinium com-ponents in the presence of urea. Addedurea partially unfolds the protein and en-ables the functionalization of inner lysineresidues. These inner electron relay unitsprovide a route for electron ‘‘hopping’’between the electrode and the active redox-center of the protein, and thus contributeto the electrical contacting of the enzymelayer with the electrode [73, 74].

Nonorganized multilayers of mediator-tethered enzymes have been constructedby cross-linking on an electrode sur-face [76]. A ferrocene-contacted GOx filmwas assembled on an Au-electrode by the

574 17 Mediated Electron-transfer between Redox-enzymes and Electrode SupportsA

u

S NH2

S NH2

GOxH2N Lys

H2NLys

NH2Lys

NH2

LysFe

H2N

H

O

H

O

S

S

Fe

LysLys

Lys

LysGOx

LysLys

Lys

LysGOx

LysLys

Lys

LysGOx

LysLys

Lys

LysLys

Lys

LysGOx

Fe

Fe

Fe

FeFe

LysNH

CHC

NLysLys

Lys

a

b

(a)

(b)

0 0.2 0.4 0.6

0

40

80

E[V]

i[m

A c

m−2

]

0 10 20 300

40

80

Glucose[mM]

i[µ

A cm

−2]

10

Fig. 8 (a) The preparation of a nonordered polymeric layer of GOxelectrically ‘‘wired’’ by ferrocene groups incorporated in the enzymenet; (b) cyclic voltammograms of the GOx/ferrocene-modifiedelectrode in the absence (a) and in the presence (b) of glucose,30 mM. Performed under argon, in phosphate buffer, pH 7, scan rate10 mV s−1. Inset: calibration curve for the amperometric response toglucose at 0.35 V. [Adapted from S. Kuwabata, T. Okamoto, Y. Kajiyaet al., Anal. Chem. 1995, 67, 1684–1690.]


N C SNH

HNS

SHN

NH

HNS

SS

GOxNH

Lys

H2NLys

NH2Lys

NH2

Lys

H2N Lys

H2NLys

NH2Lys

NH2

LysGOx

N C SNCSHN

NH

HNS

SS

GOxNH

Lys

H2NLys

Lys

NH2

LysN C SN

H

HN

S

(a)

(b)

(a),

(b), (a),

repeat

NH

NH

HNS

SS

GOxNH

Lys

H2NLys

Lys

NH2

LysNH

HN

S

NH

S

GOxNH

Lys

H2NLys

Lys

NH2

Lys

NH2

n

FeNH

(CH2)5O

HO

NH

NH

HNS

SS

GOxNH

LysLys

Lys

NH2

LysNH

HN

S

NH

S

GOxNH

LysLys

Lys

NH2

Lys

NH2

n

FeFe

SO3−

SO3−

Au

(a)

1.0

0.5

00 0.2 0.4 0.6

a

b

c

E [V]

I[µ

A] 0 10 20

0

40

80

120

Glucose[mM]

I [nA

]

(b)

Fig. 9 (a) The stepwise assembly and electricalcontacting of a crosslinked organized multilayerarray of GOx on an Au-electrode; (b) cyclicvoltammograms of the GOx/ferrocene-modifiedelectrode in the presence of glucose (20 mM) in

(a) 1, (b) 4 and (c) 8 layer configurations. Inset:amperometric responses of the 4-layer GOx arrayat 0.4 V as a function of glucose concentration.Recorded in 0.1 M phosphate buffer, pH 7.3,under argon.


cross-linking of GOx, (2-aminoethyl)ferro-cene (10), and a monolayer of p-mercaptoaniline with glutaric dialde-hyde [76] (Fig. 8a). The immobilized en-zyme was effectively contacted throughthe electron-relay groups, providing elec-tron propagation throughout the assem-bly. A calibration curve (Fig. 8b, inset)for the modified electrode in the pres-ence of glucose was derived from theanalysis of CV (Fig. 8b). More controlledsuperstructures can be constructed ina stepwise manner by the layer-by-layerdeposition of enzymes and the covalent at-tachment of relay units to the enzyme back-bone. A multilayer of GOx was modifiedwith (6-ferrocenemethylamino)hexanoicacid (Fig. 9a) [39]. The ferrocene units areoxidized by the electrode, and they inturn oxidize the enzyme redox-sites. Theamperometric responses of the layeredelectrodes were found to be controlled bythe number of protein layers, implyingthat all the enzyme-layers are electricallycontacted with the electrode. As the rate-limiting process in the oxidation of glucoseis the electron-transfer between the sub-strate and the enzyme redox-center, theamperometric response of the electrode isdependent on the glucose concentration.This dependence allows the application ofthe layered enzyme-electrode as a glucosebiosensor (Fig. 9b, inset)), whose sensitiv-ity is controlled by the number of enzymelayers in the assembly (Fig. 9b).

17.6Polymer- and Inorganic Matrix-boundEnzymes Contacted by CoimmobilizedMediators

The incorporation of enzymes into organicpolymer or inorganic composite matri-ces provides very convenient and stable

biocatalyst interfaces that have importantpractical applications. Many cheap andversatile matrices composed of differentkinds of polymers (e.g. PPy [48–52]) or in-organic components (e.g. sol–gel [77, 78],graphite paste [79, 80]) impregnated withelectron-transfer mediators, cofactors, andother materials have been investigated forbiosensor design.

17.6.1The Electrical Contacting of Enzymes inMediator-functionalized Polymers

Many examples of bioelectrocatalytic elec-trodes consisting of polymer-embeddedenzymes exist [81–83]. The electrical con-tacting of enzymes immobilized in poly-mer matrices is achieved either by virtueof a conducting polymer or by incorpo-rating electron relay groups within thepolymer, providing electron hopping be-tween the enzyme and the electrodesupport. Here, we will focus on the appli-cation of electron relay groups introducedinto polymer films to provide mediatedelectrical ‘‘wiring’’ of entrapped enzymemolecules [83].

The simplest way to achieve these assem-blies involves the incorporation of ionicredox-groups (e.g. ferrocyanide [84–86])by the anion-exchange properties of thepolymer. Consequently, the coentrapmentof enzymes and anionic redox-relays canbe employed for electrical’’ wiring’’ of thebiocatalysts [87, 88]. The anion-exchangeproperties of regular PPy has been usedto incorporate ferrocene carboxylate [50]and pyrroloquinoline quinone (PQQ) [89]for the electrical ‘‘wiring’’ of GOx dur-ing the electrogeneration of a PPy film.This approach does not offer, however,covalent binding of the mediator, re-sulting in its leakage from the poly-meric matrix. Biocatalytic electrodes based

17.6 Polymer- and Inorganic Matrix-bound Enzymes Contacted by Coimmobilized Mediators 577

on this technology therefore tend tobe unstable and cannot be used forin vivo measurements. To stabilize theelectron-relays inside the polymeric ma-trix, new amphiphilic pyrrole derivativescontaining redox-units (11–15) have beensynthesized (Fig. 10) [83, 90–92], allowingthe relay to be covalently incorporatedinto the polymer structure. For example,a ferrocene-containing pyrrole derivative(11) has been employed to electricallycontact GOx [83] and pyruvate oxidase [93],and a flavin analog-functionalized pyrrole(12) to contact flavin reductase [94]. En-zymes operating in reductive pathways(e.g. NR, pyridine-nucleotide oxidoreduc-tase) have been electrically contacted inpolymeric matrices by the use of theviologen-functionalized pyrrole derivatives(13–15) [90, 95–97]. NR, electrically con-tacted by the viologen units of a poly-mer matrix, has been used for the

Fig. 10 Structures of redox-functionalized pyrrole derivativesapplied for electropolymerization andthe entrapment of redox enzymes.

(CH2)12N+N N

+CH3

(CH2)3N+N N

+(CH2)17 CH3

(CH2)12N+N N

+(CH2)12 N

(CH2)12N+N

H3C

CH3Fe

(CH2)3N O PO

H

O (CH2)6 N N

NNH

O

O

13

14

15

11

12

amperometric detection of nitrate ions(Fig. 11a), leading to the calibration curvegiven in Fig. 11(b).

Another approach to the electrical con-tacting of polymer-bound enzymes in-volves the use of polymers that arefunctionalized with redox-units [98, 99].Polyelectrolytes represent the best choice

0 4 80

(b)(a)

2

4

Nitrate[mM]

i [µ

A c

m−2

]

10 nA2 min

x x x

x

x

y y x

a

b

t

Cat

hodi

c cu

rren

t

0 20 400

0.2

0.4

0.6

i [µ

A c

m–2

]

Nitrate[µM]

Fig. 11 (a) The amperometric responses of (a) a PPy-viologen-NR electrode, and (b) anidentical electrode constructed without the enzyme, in response to injections (x) increasingthe nitrate concentration by 3.5 µM, and (y) of buffer; (b) calibration curves (Inset: smallerconcentration range) for the response to nitrate of a PPy-viologen-NR electrode at −0.7 V.[Adapted from S. Cosnier, C. Innocent, Y. Jouanneau, Anal. Chem. 1994, 66, 3198–3201.]


EnzymeR

R

RR

R

R

RR

R

R

R

R

R

N N N+

Os2+/3+

(bpy)2Cl

NH2

y

16

x : y : z = 1 : 4 : 1.2

O O9

O

O

Polymer backbone

NH2O O

9

OHHO

HN

Polymer backbone2

HN

Polymer backbone1

17

zx

Fig. 12 A redox enzyme electrically ‘‘wired’’ to an electrode surface by flexible polymer chainsfunctionalized with redox-mediator groups and surrounding the enzyme at the electrode surface.

for the optimization of interactions withenzymes and electrodes. Hydrophilic,charged, flexible chains of polyelectrolytescan easily surround protein molecules,and even penetrate inside the protein ma-trix, providing good contact between theprotein structures and polymer backbone(Fig. 12). Each unit of a polyelectrolyte isweakly adsorbed on an electrode surface,but the cooperative effect of the entirepolymer chain leads to strong adsorption,while some parts of the chain remainunattached, providing binding domainsfor protein molecules. Three-dimensionalredox polyelectrolyte networks that elec-trically connect enzyme redox centersto electrodes have been formed in sev-eral systems [100–102], of which enzyme‘‘wired’’ hydrophilic epoxy cements arean excellent example [101, 102]. In thiscase, the polymeric chain consists ofa poly(vinylpyridine) backbone of whichapproximately one-sixth of the pyridineunits are complexed to [Os(bpy)2Cl]2+ andabout one-fifth of the pyridines have beenreacted with 2-bromoethylamine to form

pyridinium-N -ethylamine polycationic do-mains (16). This redox polyelectrolyte in-teracts with enzymes easily and ‘‘wires’’their redox centers by penetrating intothe protein shell (e.g. of lactate oxi-dase, glycero-3-phosphate oxidase, or cel-lobiose oxidase) [98, 103]. Although neg-atively charged enzymes can strongly in-teract with this polycationic polymer evenwithout cross-linking, cross-linking withthe water-soluble diepoxide poly(ethyleneglycol) diglycidyl ether (17) can further sta-bilize the system. For example, GOx wasincorporated into a cross-linked polymericmatrix containing Os(bpy)2 units coordi-nated with the pyridine groups in theprotein backbone. This enzyme-modifiedelectrode shows an electrocatalytic an-odic current in the presence of glucose(Fig. 13a), leading to the calibration curvegiven in Fig. 18(a, inset). A similar pos-itively charged copolymer of allylamineand ferrocene-functionalized acrylic acidcan interact with negatively charged pro-teins and be cross-linked with glutaricdialdehyde in the presence of GOx to


(a)

I [µ

A]

a

b0.2

0.1

0

0 0.2 0.4

0

0.4

0.8

0 100 200

i[m

A c

m−2

]

Glucose[mM]

0

(b)

0.5

1.0

200 400 600

I [µ

A]

a

b

c

0 40 80

Time[min]

0

0.4

0.8

I [µ

A]

E [V]

E [V]

12

34

567

8

910

Fig. 13 (a) Cyclic voltammograms of a graphite mirco-electrode (7µm diameter) modified with the Os complex-containing polymer 16,crosslinked with the bi-epoxy reagent 17 with embedded GOx: (a) inthe absense of glucose and (b) with glucose, 5 mM. Potential scanrate 5 mV s−1. Inset: calibration curve for the amperometricdetermination of glucose at 0.4 V; (b) cyclic voltammograms of theelectrode modified with ferrocene-functionalized crosslinkedpolyallylamine containing GOx in the polymer matrix: (a) in theabsense of glucose, (b) with glucose, 1 mM, and (c) with glucose,3 mM. Potential scan rate 5 mV s−1. Inset: amperometric responsesof the emzyme electrode (at 0.6 V) upon successive additions ofglucose. Numbers show glucose concentration in mM. [Adapted fromA. Heller, J. Phys. Chem. 1992, 96, 3579–3587, S. Koide, K. Yokoyama,J. Electroanal. Chem. 1999, 468, 193–201.]


yield stable electrically ‘‘wired’’ biocatalyticmatrices [104, 105]. These enzyme elec-trodes also demonstrate an electrocatalyticcurrent for glucose oxidation (Fig. 13b).Successive additions of glucose at a fixedoxidative potential result in increases inthe current (Fig. 13b, inset). A neutralpolyacrylamide chain carrying ferroceneunits has been used for the entrapment ofGOx molecules without the assistance ofelectrostatic interactions [99]. Many otherpolymers functionalized with differentredox groups (e.g. quinones [106]) havealso been employed for the electrical‘‘wiring’’ of entrapped redox enzymes. The

free-radical polymerization of redox hydro-gels (as an alternative to the electropoly-merization process) has also permitted theefficient entrapment of enzymes [107].

17.6.2The Electrical Contacting of Enzymes inMediator-functionalized Sol–gel Matrices

Sol–gel matrices are ideal candidatesfor the construction of enzyme-containing matrices. They are chemicallyinert, resist swelling, are processed atlow temperatures, and have tunableporosity [77]. To date, silicon alkoxide

Enzyme

Si(OMe)4

SiMe(OMe)3

Si(OMe)3(CH2)3NH2

FeNH

O

Si(OMe)3(CH2)3

MeOH

18

Fe

Fe

O O

Si

Si

O

O

O

O O

O

Si

O

OO

Si

Si

Si

O

O

O O

Si

Si

Si

OOOH

O

O

O

OO

Si

O

O

SiOO Si

O

MeO

HO

Si

OO

SiHO O

Si

Si

SiO

O

MeSi OH

O

O

SiHO O

O

Me

Si

OH

O

OSiOH

O Si Me

Si

O

O

O

Si

HOO

SiO

SiO

O Si

Si

OSi

OSiOO

O

O

MeO

Si

O Si

Si

Si

OOO

Si

OH

OO

Si

O

Si

SiO

OSi

O

O

SiO

OHO

SiSi

OO

O

O

Enzyme

Enzyme

Enzyme

Si(OMe)4

SiMe(OMe)3

Si(OMe)3(CH2)3NH2

FeNH

O

Si(OMe)3(CH2)3 MeOHFe

Fe

O O

Si

Si

O

O

O

O O

O

Si OO

Si

Si

Si

MeO

O

O O

Si

Si

Si

OOO

OO

OH

OO

Si

O

O

SiOO

Si OMe

OHO

Si

OO

Si

Si

Si

O

O

Si OH

O

O

SiHO O

O

Me

Si

OH

O

OSiOH

OSi Me

Si

O

O

O

Si

HOO

SiO

SiO

O Si

Si

Si

OSiOO

O

O

MeO

Si

HO Si

Si

Si

OOO

Si

OH

OO

SiO

Si

SiO

O

SiO

OHO

SiSi

OO

O

O

Enzyme

EnzymeElectroconductivepowder or nanoparticles

Me

O

O

NH2

H2N

NH2

(a)

(b)

Fig. 14 Encapsulation of enzyme molecules into a sol–gel matrix containing redox-mediatorgroups (a) and with the addition of conductive graphite (or metal) particles (b).


precursors have been most extensivelystudied as they are inexpensive and exhibitslow reaction kinetics. With these startingmaterials, one can readily prepare silicasol–gels that are doped with a widevariety of reagents (e.g. enzymes, electron-transfer mediators, cofactors, promoters,etc.), and the characteristics of the finalmatrix can be ‘‘tuned’’ by adjustingthe processing conditions (pH, precursorratios, etc.). GOx has been electricallycontacted in a sol–gel matrix along witha ferrocene mediator (Fig. 14a) [108]. Over80% of the GOx remained active in thesol–gel and the amperometric responseagreed well with theoretical predictions.The construction of sol–gel/enzymemultilayer matrices has allowed theamperometric sensing of glucose [109] andL-lactate [110] with linear responses withinlarge concentration ranges.

The electrical characteristics of thesefilms can be improved by the incorporationof electroconductive materials such ascarbon powder or metal particles intothe mixture (Fig. 14b). The electrodesprepared with these additions benefitfrom the porosity and rigidity ofthe silica matrix and from theelectrical conductivity of the additive [77].

Composite sol–gel electrodes containingGOx and coimmobilized redox mediatorshave been applied for the preparationof glucose biosensors [111–115]. In theseprocedures, the redox mediator is eitheradded during the gellification process(resulting in its physical entrapmentin the silicate structure) [112] oris chemically bound to the silicatenetwork (e.g. N -(3-trimethoxysilylpropyl)ferrocenylacetamide (18) may be usedas a functionalized comonomer) [114,115]. The electron relay may also betethered to the protein backbone priorto the synthesis of the matrix. Relay-functionalized GOx has been entrapped incomposite graphite/sol–gel matrices [113,116]. The presence of the relay facilitateselectron hopping between the enzyme-active site and the nearest graphiteparticle from where the graphite providesconductivity to the electrode. Ferrocene-mediated ‘‘wiring’’ of GOx resultedin an amperometric response of theelectrode in the presence of glucose [78](Fig. 15). Metallic components for theimprovement of matrix conductivity areoften composed of a graphite corecovered with a metal (e.g. palladium orrhodium) shell [117, 118]. More recently,

Fig. 15 The electrochemical responseof a GOx-ferrocene-sol-gel modifiedelectrode in the presence of glucose:cyclic voltammogram before (a) andafter (b) the addition of 10 mM glucosesolution; phosphate buffer, pH 5.6;potential scan rate, 10 mV s−1. [Adaptedfrom D. Avnir, S. Braun, O. Lev et al.,Chem. Mater. 1994, 6, 1605–1614.]

0.40.20−0.2

0

−40

40

80

E [V]

I [µ

A]

a

b


the carbon particles have been replacedby gold nanoparticles embedded togetherwith the enzyme in the porous silicatematrix [119].

17.6.3The Electrical Contacting of Enzymes inMediator-containing Graphite PasteComposites

Amperometric biosensors based on carbonpaste electrodes have received a virtual ex-plosion of reports in the last decade [79,80]. In essence, this approach allows thesimple mixing of an enzyme, an electron-relay, and a conductive powder to form abiosensing paste. The modification of anelectrode with a thick carbon paste layerallows the regeneration of a deactivatedsurface by polishing or cutting to expose anew and fully bioactive surface. This fea-ture is certainly attractive compared withthin film-modified electrodes, which haveto be discarded or at least remodifiedafter deactivation. The main challengein the construction of carbon pastebiosensors is the establishment of effi-cient electrical ‘‘wiring’’. Electron-transfermediators have been introduced intoGOx/graphite pastes, either as conjugateswith the particles or enzyme, or simplyas free molecules. Examples of mediatorsinclude ferrocene derivatives [120–135](monomeric or polymeric), benzoquinoneand benzoquinone-functionalized poly-mers [136–143], viologen derivatives [144],tetrathiafulvalene (TTF) [145, 146], tetra-cyanoquinodimethane (TCNQ) [147–149],TTF–TCNQ conducting salts [150], cobaltphthalocyanine [151], Meldola blue [152],methylene green [153], cupric hexacyano-ferrate [154], and others [79, 80]. Manyother redox enzymes have been incorpo-rated into graphite paste electrodes withappropriate electron-transfer mediators

and used as amperometric biosensors fortheir respective substrates [80]. It shouldbe noted that low potential redox-relays(particularly viologens [144]) provide ef-fective oxidation of the FADH2 cofactorunder potentials negative enough to pre-vent the nonspecific oxidation of inter-ferants. Also, the application of NAD+-dependent enzymes (e.g. glucose dehy-drogenase (GDH) [155], alcohol dehydro-genase (AlDH) [156]) requires the incor-poration of NAD+ cofactor and a redoxcatalyst for the oxidation of NADH.

Carbon paste electrodes have alsobeen used as supports for multienzymesystems. For example, AChE andChO have been coimmobilized incarbon pastes, either with monomericTTF [157] or flexible ferrocene-containingpolymers [158] as the electron mediator.The hydrolysis of acetylcholine isbiocatalyzed by AChE, and then the cholineproduced is oxidized by the electricallycontacted ChOx giving an analyticalamperometric signal corresponding to theacetylcholine concentration.

It should be noted, however, that carbonpastes do not offer a general route toany desired assembly. Some enzymes,for example, fructose dehydrogenase [136,159] and aldose dehydrogenase [160], havedefied all efforts at incorporation intothe matrix without loss of activity. It ispossible to introduce these enzymes tothe surface of preformed carbon pasteelectrodes while retaining their biocatalyticactivity, but this procedure does not exploitthe many advantages of the carbon pastetechnique [79, 80]. The construction andoptimization of multicomponent matricescomposed of graphite powder, organic oils,enzymes, redox mediators, cofactors, andother materials is still a subject of intensivestudy.

17.7 The Electrical Contacting of FAD-enzymes by Mediator-functionalized FAD 583

17.7The Electrical Contacting of FAD-enzymesby Mediator-functionalized FAD

We have already noted that the randomfunctionalization of the protein backbonewith electron relay groups leads to amixture of products, the average ofwhich is detected electrochemically. Inorder to accomplish the best possibleelectron contacting, the mediator shouldbe selectively placed in an optimumposition between the redox-center and theenzyme periphery. In the case of surface-confined enzymes, the orientation of theenzyme-mediator assembly with respectto the electrode should also be optimized.These requirements have been addressedby the use of mediators that are covalentlyattached to the enzyme cofactor or toboth the cofactor and to the electrodesurface. This specific covalent attachmentpositions the mediator at exactly thepoint that is required, leading to muchmore homogeneous behavior of the manyenzyme molecules in a sample.

17.7.1Electrical Contacting of Enzymes byReconstitution of Apo-flavoenzymes withRelay-FAD Cofactor Units

A novel means for the establishmentof electrical contact between the redox-center of enzymes and their environmentbased on a reconstitution approach hasrecently been demonstrated [161, 162].According to this method (Fig. 16a),the FAD-redox centers of GOx or D-amino acid oxidase (AOx) were removedto yield the respective apo-enzymes.The amino-functionalized semisyntheticN6-(2-aminoethyl)-FAD (19) was cova-lently linked to (6-ferrocenemethylamino)hexanoic acid, and the bifunctional

redox-active ferrocene-FAD cofactor pro-duced (20) was reconstituted into apo-GOxor apo-AOx. The resulting semisyntheticenzymes revealed bioelectrocatalytic fea-tures for the oxidation of glucose orD-alanine, respectively. Figure 16(b) showscyclic voltammograms recorded in a solu-tion of ferrocene-FAD-reconstituted GOxat a cystamine-modified electrode in thepresence of different concentrations ofglucose. The calibration plot (Fig. 16b, in-set) was derived from the respective cyclicvoltammograms. The bioelectrocatalyticfeatures of these ‘‘electroenzymes’’ orig-inate from the single ferrocene electron-relay group that electrically contacts theFAD center with the electrode surface.

17.7.2Electrical Contacting of Enzymes bySurface-reconstitution of Apo-flavoenzymeson Relay-FAD-functionalized Electrodes

The organization of a reconstitutedenzyme aligned on a catalyst-FADmonolayer was recently realized by thereconstitution of an apo-enzyme ona surface functionalized with a relay-FAD monolayer (Fig. 17a) [162, 163].PQQ, (21) was covalently linked to abase cystamine monolayer, and N6-(2-aminoethyl)-FAD (19) was then attachedto the PQQ redox-relay units. Followingthe construction of this organizedelectrode-PQQ-FAD assembly, apo-GOxwas reconstituted onto the semisyntheticFAD unit to yield an immobilizedbiocatalyst on the electrode with a surfacecoverage of 1.7 × 10−12 mole cm−2. Theresulting reconstituted enzyme revealsbioelectrocatalytic properties. Figure 17(b)shows cyclic voltammograms of theenzyme electrode in the absence and thepresence of glucose. When the substrate ispresent, an electrocatalytic anodic current


GOx

FAD

GOx GOx

FAD

FAD

Fc

FAD Fc

Native GOx Apo-GOx Reconstituted GOxS NH3

+

S NH3+

S NH3+

GOx

FADFc

Glucose

Gluconicacid

e−e−

0

2

4

6

(b)

(a)

I [µ

A]

E[V]

0 0.2 0.4 0.6

0 20 400

2

4

Glucose[mM]

I [µ

A]

a

b

c

d

20

N

N

NN

H HH OH

OHHH OHH H

OPOO−

O PO

O−O

NNN

NNH

O

O

O

NH

OHOHHH

H H

O NH Fe

FAD Fc

20

Fig. 16 (a) The preparation of an electrically ‘‘wired’’ enzymeby the reconstitution method, involving the removal of thenative FAD cofactor from the enzyme (e.g. GOx) and theincorporation of the artificial FAD-ferrocene dyad into theapo-enzyme; (b) cyclic voltammograms of a system consistingof ferrocene-FAD-reconstituted GOx (1.75 mg mL−1) at variousconcentrations of glucose: (a) 0, (b) 1, (c) 3, and (d) 20.5 mM.Experiments were performed in 0.1 M phosphate buffer, pH 7.3,at 35 C, using a cystamine-modified Au-electrode, potentialscan rate 2 mV s−1, under argon. Inset: calibration curve of thebiocatalytic current (0.5 V) at different glucose concentrations.

17.7 The Electrical Contacting of FAD-enzymes by Mediator-functionalized FAD 585

400

(a)

(b)

200

0

2

1

00 0.04 0.08

I [m

A]

Glucose[mM]

−0.6 −0.2 0.2

a

b

E [V]

I [µ

A]

SS

NH2

NH2S NH2

N

HNCO2H

OO

HO2C

HO2C

SHN21PQQ PQQ

SHN PQQ

GOx

Apo-GOxFAD

FADH2N

GOx

SHN PQQ FAD

Glucose

Gluconicacid

e−e−

EDC

19

Fig. 17 (a) The surface-reconstitution ofapo-GOx on a PQQ-FAD monolayer assembledon an Au-electrode; (b) Cyclic voltammogramsof the PQQ-FAD-reconstituted GOx on anAu-electrode. (a) In the absence of glucose.(b) With glucose, 80 mM. Recorded in 0.1 M

phosphate buffer, pH 7.0, under Ar, at 35 C,scan rate, 5 mV s−1. Inset: Calibration curvecorresponding to the amperometric responses(measured by chronoamperometry, E = 0.2 V)of the PQQ-FAD reconstituted GOx enzyme-electrode at different concentrations of glucose.

is observed, implying electrical contactbetween the reconstituted enzyme and theelectrode surface. The PQQ site locatedat the protein periphery is constantlyoxidized by the electrode, and the PQQ-mediated oxidation of the FAD center

activates the bioelectrocatalytic oxidationof glucose. The resulting electrical currentis controlled by the recycling-rate ofthe reduced FAD by the substrate.Figure 17(b, inset) shows the derivedcalibration curve for the amperometric


responses of the reconstituted enzymeelectrode at different concentrations ofglucose. The resulting current densitiesare unprecedentedly high (300 µA cm−2 at80 mM of glucose). Control experimentsrevealed that reconstituted GOx on anelectrode-FAD assembly (lacking the PQQcomponent) does not exhibit directelectron-transfer communication with theelectrode surface, demonstrating that thePQQ relay unit is indeed a key componentin the process [162, 163].

The electron-transfer turnover rate ofGOx with molecular oxygen as the electronacceptor corresponds to ca. 600 s−1 at25 C. Using an activation energy of7.2 kcal mole−1, the electron-transferturnover rate of GOx at 35 C is estimatedto be ca. 900 s−1 [162, 163]. A denselypacked monolayer of GOx (ca. 1.7 × 10−12

mole cm−2) that exhibits the theoreticalelectron-transfer turnover rate is expectedto yield an amperometric response ofca. 300 µA cm−2. This indicates thatreconstituted GOx on the PQQ-FADmonolayer exhibits an electron-transferturnover with the electrode of similareffectiveness to that observed for theenzyme with oxygen as a natural electronacceptor. Besides the high sensitivityof the resulting enzyme-electrode, theimpressive efficiency of electrical contacthas important consequences in thedesign of future enzyme-electrodes.Amperometric glucose-sensing electrodesgenerally suffer from the nonspecificoxidation of various interferants such asascorbic acid or uric acid. Also, oxygeninterferes with current transduction asa result of nonelectrochemical oxidationof the enzyme redox-site. The efficientelectrical contact of the biocatalyst suggeststhat the nonspecific oxidation of theinterfering substrates and the reactionof the biocatalyst with oxygen should

have little effect on the resultingcurrent. Indeed, it was found thatthe transduced amperometric responseat a glucose concentration of 5 mMwas almost unaffected in the presenceof oxygen or other interferants [162,163]. The resulting selectivity of thereconstituted enzyme electrode and thehigh current densities achieved havefurther importance in the applicationof these electrodes as invasive glucosesensors. These assemblies are sensitiveenough to function on microelectrodes atthe end of needles thin enough to be ofversatile medical use.

The intermediate location of a redox-relay between the electrode surface andthe cofactor unit embedded in theenzyme is of key importance for theestablishment of electrical contact betweenthe enzyme and the electrode. Forexample, a PQQ monolayer assembledonto an Au-electrode was employed toreconstitute the PQQ-dependent apo-GDH [164, 165]. In this case, the PQQplays the role of the embedded cofactor,and since no additional electron-relay wasimmobilized between PQQ-cofactor andthe electrode, the reconstituted enzymelacks the electrical contact with theelectrode. The electrochemical oxidationof glucose by the reconstituted biocatalystwas only stimulated in the presence of adiffusional electron-transfer mediator. Inother cases, however, the orientation ofthe protein with respect to the electrodeis sufficient to promote electron-transferwithout the need for a mediator. An Fe(III)-protoporphyrin IX complex was assembledas a monolayer on an Au-electrodeand apo-Mb was reconstituted with theheme-cofactor monolayer [166]. Althoughnative Mb usually lacks direct electricalcommunication with electrode supports asa result of insulation of the heme center,

17.8 The Electrical Contacting of NAD(P)+-dependent Enzymes 587

the surface reconstituted Mb revealedelectrical contact with the electrode. Thisproperty was attributed to the alignment ofthe heme center on the electrode surfacein a structural orientation that facilitateselectron-transfer with the electrode.

17.8The Electrical Contacting ofNAD(P)+-dependent Enzymes

The nicotinamide redox cofactors(NAD(P)+) play important roles in biolog-ical electron transport, acting as carriersof two electrons and one proton. It isof prime importance for the function-ing of the living systems that these redoxcofactors recognize and undergo rapid re-action with their desired redox partners(NAD(P)+-dependent enzymes), and atthe same time resist thermodynamicallyfavorable side reactions. The two nicoti-namide cofactors, nicotinamide adeninedinucleotide (NAD+) and nicotinamideadenine dinucleotide phosphate (NADP+),have closely related structures and electro-chemical properties. Their oxidation andreduction involves two electrons and aproton, and can formally be consideredto be a hydride transfer. In aqueous so-lution at pH 7.0, the thermodynamicredox potentials (Eo′

) for NAD+/NADHand NADP+/NADPH redox pairs are−0.561 V and −0.565 V, respectively [167].Electrochemistry of NAD(P)+/NAD(P)Hhas been extensively studied at differentelectrodes (e.g. gold, platinum, and glassycarbon), and it has been demonstrated thatthe electrochemical oxidation/reductionprocess is highly irreversible and proceedswith large overpotentials [168, 169]. De-hydrogenases, the vast majority of redoxenzymes, require NAD(P)+/NAD(P)H-coenzyme for their operation; thus, the

electrochemical regeneration of its ox-idized or reduced form (NAD(P)+ orNAD(P)H) is vitally important for theelectrical contacting of these enzymes. Inview of this great importance, electrocat-alysts for the oxidation and reduction ofNAD(P)+/NAD(P)H have been developed.

17.8.1The Electrochemical Regeneration ofNAD(P)+-cofactors

The oxidation of NAD(P)H appears tooccur as a single-step two-electron pro-cess at high overpotentials at bare elec-trodes [22, 170, 171] (ca. 0.4 V, 0.7 Vand 1 V at carbon, Pt, and Au elec-trodes, respectively [172, 173]), and theelectrolysis of NAD(P)H under appropriateconditions can regenerate NAD(P)+ withalmost 100% efficiency [174, 175]. How-ever, strong adsorption of NAD(P)H andNAD(P)+ (e.g. on platinum, gold, glassycarbon, and pyrolytic graphite) generallypoisons the electrode surface and inhibitsthe oxidation process [172, 173, 176–179].Furthermore, NAD(P)+ is an inhibitor ofthe direct oxidation of NAD(P)H [180], andadsorbed NAD(P)H can be oxidized tounwanted products (probably dimers [170]and/or stable adducts with the electrodesurface [181]). The differences in activity ofelectrode materials and the effects of elec-trode pretreatment on NAD(P)H oxidationoriginate from adsorption phenomena. Anappropriate choice of electrode materialcan significantly decrease the overpoten-tial and enhance the process [182] (e.g.the NAD(P)H oxidation proceeds at an Agelectrode at 0.23 V [183]). Special electrodepretreatment (e.g. electrochemical surfaceoxidation) can decrease the NAD(P)Hadsorption and enhance the NAD(P)+yield [184–186]. Moreover, oxidative pre-treatment of glassy carbon electrodes can


produce quinonoid groups on their sur-faces, which act as catalytic sites forNAD(P)H oxidation [186, 187].

For the efficient electrooxidation ofNAD(P)H, mediated electrocatalysis isnecessary [22, 170, 171], and a wide rangeof diffusional mediators has been stud-ied [188–193]. Organic compounds thatundergo two-electron reduction-oxidationprocesses and also function as protonacceptors-donors upon their redox trans-formations (such as ortho- and para-derivatives of quinones, phenylenedi-amines and aminophenols) have beenfound to be ideal for the mediationof NAD(P)H oxidation, although single-electron-transfer mediators (e.g. ferrocenederivatives) are also capable of oxidizingNAD(P)H [190, 191]. Some compoundsdemonstrate very high rates for the me-diated oxidation of NAD(P)H in aqueoussolutions [188, 189, 194, 195].

The use of electrode-immobilized medi-ators [22, 170, 171] (such as o-quinones[181, 193, 195–199], p-quinones [200,201], phenazine, phenoxazine and phe-nothiazine derivatives [202–210], Os-complexes [211], metallophthalocyani-nes [212], and organic conductivesalts [213, 214]) has also been applied forthe regeneration of NAD(P)+. A great vari-ety of immobilization techniques has beenapplied for the preparation of these mod-ified electrodes – the mediator moleculeshave been directly adsorbed onto electrodesurfaces [181, 196, 199, 202–205], incorpo-rated into polymer layers [197, 198] or cova-lently linked to functional groups on elec-trode surfaces [193, 195, 206, 207]. The co-valent coupling of redox mediators to self-assembled monolayers on Au-electrodesurfaces has an important advantage forthe preparation of multi-component orga-nized systems. For example, PQQ (21) has

N

HNCO2H

OO

HO2C

S

HN O

0

0.2

0.4

0.6

0.8

1

I [µ

A]

0−0.2

E [V]

a

b

c

Au

Fig. 18 Cyclic voltammograms of an Au-PQQ electrode with (a) 0.1 MTris-buffer, pH 7.0, (b) 10 mM NADH and (c) 10 mM NADH and 20 mMCa2+. Recorded at a scan rate of 1 mV s−1.


been covalently attached to amino groupsof a cystamine monolayer assembled ontoan Au surface, and the structure obtaineddemonstrated good electrocatalytic activity

for NAD(P)H oxidation, particularly inthe presence of Ca2+-cations as promot-ers (Fig. 18) [215]. The NADP+-dependentmalic enzyme was covalently linked to

e−

Transduced current

NADP+

Malicenzyme

NH

OPQQ

O

HN

NADPH

OHOH

OHO

O

OOH

O

CO2+

0

0.4

0.8

(a)

(b)

0 0.1−0.1−0.2

E [V]

a

b

c

I [µ

A]

500

400

300

200

100

0−4−5−6−7

log [Malate][M]

I [n

A]

22

Fig. 19 (a) Amperometric detection of malate by PQQ-malic enzymemonolayer electrode using diffusional NADP+ electrochemically regeneratedby the PQQ catalyst; (b) cyclic voltammograms of the PQQ-malic enzymeelectrode in the presence of malic acid at (a) 0, (b) 0.125, (c) 125 and(d) 1250 µM. Inset: calibration curve corresponing to the concentrationdependence of the biosensor. Recorded in 0.1 M phosphate buffer, pH 7.2,with 2 mM NADP+ and a scan rate of 2 mV s−1.


the PQQ monolayer assembled on theAu-electrode [216] (Fig. 19a). In the pres-ence of NADP+ as diffusional cofactor andmalic acid (22) as substrate, biocatalyzedoxidation of 22 yields the reduced cofactor,which is electrocatalytically reoxidized byPQQ. The resulting current is controlledby the concentration of NADPH, whilethe concentration of the reduced cofactoris determined by the analyte malic acid(Fig. 19b). This system is still not fully in-tegrated, however, as the cofactor is in adiffusional configuration.

NAD(P)+-dependent enzymes, electri-cally contacted with electrode surfaces,can provide efficient bioelectrocatalysisfor the NAD(P)H oxidation. For example,diaphorase (DI) was applied to oxidizeNADH, using a variety of quinonecompounds, several kinds of flavins,or viologens as mediators between theenzyme and electrode [217, 218]. Thebimolecular reaction rate constants be-tween the enzyme and mediators whoseredox potentials are more positive than−0.28 V at pH 8.5 can be as high as108 M−1 s−1, suggesting that the reac-tions are diffusionally controlled. The highstability of the enzyme GDH has even al-lowed the bioelectrocatalytic regenerationof NAD(P)+ on a preparative scale [219].

Electrocatalytically regenerated NAD(P)+has been coupled to secondary NAD(P)+-dependent enzymes, thus providing recy-cling of the cofactor for the biocatalytictransformations [220, 221]. For example,L-lactate was converted into D-lactate, witha yield better than 97% using L-lactate de-hydrogenase lactic dehydrogenase (LDH),relying on NAD+ regenerated by an elec-trocatalytic reaction [222]. The system in-volves stereospecific catalysis of L-lactateoxidation by the enzyme, with the elec-trochemical regeneration of NAD+ at the

anode and electrochemical reduction ofpyruvate at the cathode.

Electrocatalytic recycling of NAD(P)+has allowed the development of numer-ous amperometric biosensors based onNAD(P)+-dependent hydrogenases [221].Bioelectrocatalytic regeneration ofNAD(P)+ using electrically contacted DIhas been used to drive the biocatalyzed ox-idation of methanol to CO2, using NAD+-dependent dehydrogenases [218], and amethanol/dioxygen biofuel cell has beenassembled using these bioreactions in theanode compartment.

17.8.2The Electrochemical Regeneration ofNAD(P)H-cofactors

The electrochemical reduction of NAD(P)+has been studied in aqueous [223–225]and nonaqueous solutions [226]. It usu-ally proceeds stepwise at very negativepotentials (e.g. in aqueous solution at pH7.0 and at a mercury electrode, two sep-arated waves [224] are observed at −1.0 Vand −1.6 V); thus, the process requires alarge overpotential. The first step is a one-electron reduction of NAD(P)+, resultingin the neutral radical species NAD(P)•,which goes on to produce a nonenzymat-ically active dimer [227]. The second stepresults in the formation of the enzymati-cally inactive 1,6-reduction product. Onlya small amount of the enzymatically ac-tive 1,4-reduction product NAD(P)H canbe found in the product mixture, with ayield dependent on the electrolysis con-ditions. The initial state of the electrodesurface and the background electrolyte, aswell as the applied potential, strongly affectthe mechanism of the NAD(P)+ reductionand the composition of the product mix-ture. The electrode material has a dramaticeffect on the kinetics of the process and


an appropriate selection of electrode cangreatly improve the process efficiency. Forexample, a bare silver electrode providesa quasi-reversible reduction-oxidation ofNAD(P)+/NAD(P)H (i.e. cathodic and an-odic peak are observed at potentials 0.12 Vand 0.23 V, respectively [183]), althoughNAD(P)H is usually a minor product ofnoncatalytic electrochemical processes atbare electrodes [22].

Direct, nonmediated electrochemical re-duction of NAD(P)+ at modified electrodesurfaces has been used to produce theenzymatically active NAD(P)H and evento couple the NAD(P)H regeneration pro-cess with some biocatalytic reactions [228].The modifier molecules used for thesepurposes are not redox active and theydo not mediate the electron-transfer pro-cess between an electrode and NAD(P)+;however, they can effectively decrease therequired overpotential and prevent for-mation of the nonenzymatically activedimer product [228]. For example, theefficiency of the direct electrochemical re-generation of NADH from NAD+ wasenhanced by the use of a cholesterol-modified gold amalgam electrode thathinders the dimerization of the NAD-radicals on its modified-surface [228]. Thisdirect electrochemical NAD+ reductionprocess was used favorably to drive an en-zymatic reduction of pyruvate to D-lactatein the presence of lactate dehydrogenase.The turnover number for NAD+ was es-timated as 1400 s−1. Other modifiers thatenhance formation of the enzymaticallyactive NAD(P)H include L-histidine [229]and benzimidazole [230], immobilized asmonolayers on silver electrodes. Cyclicvoltammetric experiments demonstratedthat these modified electrodes can cat-alyze the reduction of NAD+ to enzy-matically active NADH at particularly lowoverpotentials.

The electrocatalysis of NAD(P)H regen-eration by the application of either non-biological redox materials or NAD(P)+-dependent enzymes can greatly improvethe process efficiency and avoid the forma-tion of enzymatically inactive byproducts.Complexes of Rh have been reported to beactive for the electrocatalytic reduction ofNAD(P)+ to NAD(P)H [231, 232]. SimilarRh complexes have been utilized in ho-mogeneous catalysis with chemical [233]and photochemical [234–236] activation.The electrocatalytic process includes theregioselective transfer of two electronsand a proton to NAD(P)+. In thesesystems, hydrido-rhodium species are as-sumed to be the active catalytic moiety.Tris(bipyridine)rhodium(III) [232, 234],tris(5-sulfo-2,2′-bipyridine)rhodium(III)[236], (pentamethylcyclopentadienyl-2,2′-bipyridine-chloro)rhodium(III) [231, 237],and chlorotris[diphenyl(m-sulfonatophe-nyl)phosphine]rhodium(I) [235] have beenused as homogeneous mediation of elec-trons to NAD(P)+. The catalytic efficiencyof a series of Rh-complexes has beenstudied [238] and it was shown that thecatalyst activity decreases in the presenceof electron-withdrawing substituents inthe 2,2′-bipyridine ligand and increaseswith electron-donating substituents. Sub-stituents in the 6-position of the ligandslow the catalytic reaction because ofsteric effects. Structure-activity relation-ships were found in the mechanismof the regioselective reduction of NAD+by Rh-complexes [239]. These examplesdemonstrate the ability of diffusionallyfree, catalytically active mediators to trans-port hydride equivalents to the solubleNAD(P)+ cofactor. The NADH cofactor,electrochemically regenerated using Rh-complexes, has been used to mediatebiocatalytic reactions [231].


Electrode modification with the Rh-complexes catalytically active in hydridetransfer (thus for the catalysis of NAD(P)+reduction) has brought important im-provements in the electrocatalytic regen-eration of NAD(P)H [240–243]. Rhodiumcomplexes have been immobilized atelectrode surfaces by their incorporationinto polymeric films organized on elec-trode surfaces. Films of substituted PPyswere prepared by the oxidative poly-merization of the corresponding pyrrole-ligand derivative. Immobilization of Rh-complexes into the polymeric films wasachieved by a ligand-exchange reaction ofthe complex with this ligand-containingpolymer-film [243]. Another method in-cludes the immobilization of copoly-mers of hydrophilic N -vinyl-pyrrolidoneand polymerizable derivatives of the Rh-complex (pentamethylcyclopentadienyl-2,2′-bipyridine-chloro)rhodium(III). Gam-ma-irradiation of these mixtures on anelectrode surface gives highly permeableredox active polymer films [240]. Thesefilms swell easily in water and thus providea three-dimensional reaction layer, whichis advantageous for electrocatalytic pro-cesses mediated by the surface-confinedrhodium complex. A quinone-modifiedelectrode surface has also been reportedto be electrocatalytically active for NAD+reduction [244].

The regeneration of NAD(P)H withthe participation of mediator-contactedenzymes ensures that NAD(P)+ reduc-tion proceeds selectively and that onlyenzymatically active NAD(P)H is pro-duced. Many enzymes have been usedin this context to provide the bio-electrocatalytic reduction of NAD(P)+,for example, ferredoxin-NADP+ reduc-tase (FNR) [245–249], lipoamide dehydro-genase [250–254], formate dehydrogenase

(FDH) [252, 255, 256], 2-oxocarboxylate re-ductase [257], enoate reductase [257, 258],DI [247, 259–262], AlDH [263], and hydro-genase [264, 265]. A variety of low poten-tial electron-transfer mediators have beenused to activate the reductive enzymes,for instance, viologen derivatives [246,248–250, 253, 254, 261, 262], flavins [255,256, 259, 264], quinones [247], and theredox protein ferredoxin [266]. Some re-dox enzymes can directly communicatewith electrode supports, and thus stim-ulate the regeneration of the NAD(P)Hcofactor. For example, hydrogenases (fromRhodococcus opacus and Atcaligenes eutro-phus H16) have been successfully appliedfor the bioelectrocatalytic regenerationof NAD+, without the application of aredox-mediator [267, 268]. However, theelectrocatalytic rates of these systems aregenerally too slow to produce observablecatalytic current on the cyclic voltammet-ric time scale (notable exceptions includedimethyl viologen (MV2+)/DI [253, 254,261, 262], MV2+/FNR [249], and ferre-doxin/FNR [266]). Application of a low po-tential quinone adriamycin as an electron-transfer mediator to DI or FNR allowshighly effective bioelectrocatalytic reduc-tion of NADP+, with an electrocatalyticcurrent appearing at potentials more neg-ative than −0.7 V. A high current yield(ca. 97%) was also achieved with solubleAlDH and diffusional acetophenone as anelectron mediator [263].

Immobilized low potential electron-transfer mediators (e.g. viologens) aremore promising than diffusional medi-ators for the practical regeneration ofNAD(P)H coupled with further biocat-alytic reactions. The immobilization ofviologens usually results in significantpositive potential shift of their redox po-tential [269–272], which however, badlyaffects their efficiency. The potential shift


of the immobilized viologen mediatorresults from dimer formation and inter-actions with the supporting organic matri-ces. A special immobilization technique,which includes the cross-linking of long-chain amino-viologen derivatives, FNR,and bovine serum albumin (BSA) with glu-taric dialdehyde on an electrode surface,has been developed to prevent such po-tential shifts, retaining the high reductivecapabilities of the viologen mediator [248].A cyclic voltammogram observed at thisviologen/FNR/BSA-modified electrode inthe presence of NADP+ shows high ca-thodic bioelectrocatalytic current at poten-tials more negative than −0.45 V. The bio-electrocatalytic regeneration of NAD(P)Hwith the use of the coimmobilized violo-gen mediator and DI can proceed with avery high current yield (97.8%) upon theapplication of a potential of −0.8 V [262].Therefore, this process was applied for am-perometric biosensing of NAD(P)+ [273].Sensitivities of 1.4 and 3.5 mA M−1 cm−2

were recorded for NAD+ and NADP+,respectively.

Bioelectrocatalytically regeneratedNAD(P)H has been coupled with fur-ther biocatalytic reactions, where theNAD(P)+-dependent enzymes utilize theregenerated NAD(P)H and perform thereduction of their respective substrates.For example, the reduction of acetoneto 2-propanol has been achieved in thepresence of AlDH upon the biocatalyticregeneration of NADPH in the presenceof MV2+ and FNR [263] (Fig. 20a). NADHregenerated by directly contacted hydroge-nase (from Alcaligenes eutrophus H16) hasbeen utilized to drive the transformation ofα-ketoglutarate into L-glutamate, catalyzedby an L-glutamate dehydrogenase [274].Turnover numbers of 450 h−1 and 207 h−1

were obtained in an electrochemical thin-layer cell and in a preparative scale-batchreactor, respectively. Coimmobilization oftwo enzymes, one of them catalyzing theregeneration of NAD(P)H and the second

Acetone

2-Propanol

AlcDHFNR

NADPH

NADP+MV·+

MV2+

e−

Substrate 2

Product 2

AlcDH

NADPH

NADP+Substrate 1

Product 1

(b)

(a)

e−

AlcDH

Fig. 20 Bioelectrocatalytic reduction systems using: (a) AlcDH and FNRfor the biocatalyzed transformation and the regeneration of NADPH,respectively, and; (b) using AlcDH for both processes, where thebiocatalytic cycle Substrate 1/Product 1 performs the mediating functionfor the NADPH regeneration and the second biocatalytic cycle results inthe formation of the aim product 2.


utilizing the NAD(P)H, allows NAD(P)H-driven biocatalytic transformations to beperformed in a single biocatalytic matrixin the presence of NAD+ and a mediator.Glutamate dehydrogenase (GluDH) andFNR coimmobilized on a glassy carbonelectrode surface have been shown tostimulate the biocatalytic reduction of α-ketoglutorate in the presence of diffusionalNADP+ and adriamycin [247]. In somecases, single enzymes (e.g. AlDH) canperform both functions: (1) regenerationof NAD(P)H in the presence of a me-diator providing electron transport fromthe electrode and (2) biocatalytic reductionof a ketone to an alcohol [263] (Fig. 20b).Stereoselective biocatalytic reactions per-formed on a preparative scale with highcurrent yield (ca. 98%) and utilizing bio-electrochemically regenerated NAD(P)Hand secondary enzymatic processes havebeen also reported [275, 276].

17.8.3The Association of NAD(P)+-dependentEnzymes with NAD(P)+ Cofactors byCovalent and Entrapment Methods

In view of the high cost of NAD(P)+/NAD(P)H cofactors, practical applicationsrequire their immobilization together withthe enzymes. The covalent coupling ofnatural NAD(P)+ cofactors to an organicsupport results, however, in a substan-tial decrease of their efficiency. Mobilityof the cofactor is vital for its efficientinteraction with enzymes, so serious at-tention has been paid to the synthesis ofartificial analogs of the NAD(P)+ cofac-tors carrying functional groups separatedfrom the bioactive site of the cofac-tor by spacers [277, 278]. The spacer isusually linked to N -6 position of theNAD(P)+ molecule, and should providesome flexibility for the bioactive part

of the cofactors, allowing them to beassociated with the enzyme molecules.Structure/activity relationships of the arti-ficial functionalized NAD(P)+-derivativeshave been studied with different enzymes,and the possibility to substitute the nat-ural NAD(P)+ cofactor with these artifi-cial analogs has been demonstrated [277,279]. The artificial NAD(P)+ derivatives(e.g. N6-aminoethyl-NAD+, (23)) havebeen covalently linked to insoluble ma-trices (e.g. Sepharose [280–282]) or water-soluble polymers (e.g. dextran [283–286],polyethyleneimine [281, 282, 287], polyly-sine [281], polyethyleneglycol [288–291]).In another study, a NAD+ derivative carry-ing a polymerizable acrylamide group wasprepared and subjected to radical polymer-ization to produce NAD+-functionalizedpolymers [292].

All the polymeric NAD(P)+-derivativeshave been checked for their cofactor ac-tivity and compatibility with enzymaticbiocatalytic processes. The polymer-linkedNAD(P)+-derivatives were associated withNAD(P)+-dependent enzymes such asAlDH [281, 282, 291, 292], lactate dehy-drogenase [286, 292], malate dehydroge-nase [288, 292] and aldehyde dehydroge-nase [287]. It was found that differentNAD(P)+-polymers are active as cofactorstowards different enzymes. For example,polyethyleneimine and polylysine boundNAD+-derivative revealed 60% and 25%activity, respectively, as compared with thenative NAD+ in the presence of rabbitmuscle lactate dehydrogenase, but onlyminute activity (ca. 2–7%) in the presenceof alanine dehydrogenase from Bacillussubtilis [281]. A comparative study of thecofactor activity with different enzymes isa subject of great interest. Even thoughseveral studies [279] attempt to predictthe structural/functional relationship forthe polymer-bound NAD(P)+-derivatives,


most of the NAD(P)+-dependent enzymesstill have unknown 3-D structures, andtheir compatibility with the polymeric co-factors can only be tested experimentally.

A polymer-bound NAD+-derivative(NAD+-alginic acid) was immobilized onan electrode surface. Despite the highmolecular weight of the NAD+-alginicacid, all NAD+-units were reduced elec-trochemically in the presence of MV2+and DI, providing NADH for couplingwith reductive enzymatic reactions [293].A reagentless ethanol biosensor was de-veloped using a NAD+-dextran conju-gate [284]. A mixture of AlDH, NADH-oxidase, and NAD+-dextran was incor-porated into a poly(vinylalcohol) matrixonto an electrode surface. In the presenceof ethanol, AlDH reduced the polymer-bound NAD+-units, producing NADH.The resulting NADH was then oxidizedby the NADH-oxidase in the presence ofoxygen, recycling the original NAD+ andproducing H2O2 that was finally detectedby the electrode. The system could beimproved further if a redox-mediator pro-vided electrical contacting of the NADH-oxidase instead of hydrogen peroxide gen-eration. The biocatalytic systems based onthe association between polymer-boundNAD(P)+-derivatives and the correspond-ing enzymes allow the performance ofbiocatalytic processes, including applica-tion in a model enzyme bioreactor [289].Polyethyleneglycol-bound NAD+ has beensuccessfully employed in the continuousproduction of L-amino acids from the cor-responding α-keto acids by stereospecificreductive amination [291]. (S)-1-Phenyl-2-propanol was obtained by reduction of thecorresponding ketone using AlDH(fromRhodococcus erythropolis) together witha polymer-bound NADH and a totalturnover of 8 × 104 was achieved [294].

A different approach to the associationof NAD(P)+ with enzymes involves thecovalent coupling of the functionalizedcofactor molecules to the enzyme back-bone [295–299] (Cf. coupling of redox-mediators with the enzyme backbone,Chapter 5). For example, AlDH carboxylgroups (aspartic and glutamic residuals)were activated with a carbodiimide andN -hydroxysuccinimide and then the activeester groups were reacted with N6-[N -(6-aminohexyl)carbamoylmethyl]-NAD+(24),resulting in an enzyme-cofactor com-plex [295] (Fig. 21a). This modificationtechnique resulted in the spacially ran-dom attachment of NAD+-units, with anaverage loading about 0.3–1.6 NAD+ perenzyme. To provide specific attachment, a44-cystein residue was introduced on thesurface of GDH using site-directed muta-genesis [300]. A synthetic analog of NAD+was site-specifically covalently linked tothe mutant by a disulfide bridge at the 44-cysteine residue (Fig. 21b). The enzyme-NAD+ complex demonstrated biocatalyticactivity that could be explained either by en-zyme activation with the NAD+ unit linkedto the same enzyme (intra-complex acti-vation), or by cross-reaction between theNAD+ units and enzymes from differentcomplexes (inter-complex activation). Thepossibility of the cross-reaction allows theuse of two different enzymes with NAD+-units linked only to one of them (Fig. 21c).A mixture of the NAD+-functionalizedGDH with the native nonfunctionalizedlactate dehydrogenase allowed the biocat-alytic processes of both enzymes, pro-ducing L-lactate and gluconic acid frompyruvate and D-glucose [300]. A turnoverof 45 cycles per minute for each NAD+molecule and total turnover per cofac-tor of 1.35 × 105 for the first 2.5 dayswere found in this system when the re-action was performed in a hollow fiber


NH

HN O

SS

NH

O

N

O

O

O

O

NH

HN O

NH

O

Enzyme

COOHN

O

O

HO Carbodiimidecoupling

OPOO−

O PO

O−O

N

NN

NNH

O

OH OHHH

H HO

HHN+

H HOHOH

NH2

O

HN O(CH2)n

NH2

NH

HN O

NS

S

O

O

N

O

ONS

S

NH

O

Enzyme

SH

(c)

(b)

(a)

+

n = 2 23n = 6 24

NAD+

NAD+

NAD+ NAD+

NAD+

Fig. 21 (a) Preparation of a randomly linkedenzyme-NAD+ complex by the use of asemisynthetic amino-NAD; (b) preparation of aspecifically linked enzyme-NAD+ complex by theuse of a genetically modified enzyme and

amino-NAD+; (c) the biocatalytic functions oftwo different enzymes activated by NAD+covalently tethered to only one of them via inter-and intra-molecular complexation.


NS

O

O

N S

S

S25Poly(ethylene glycol)-NAD+

NH2

Poly(ethylene glycol)-NAD+

O

N S

S

HN O

Poly(ethylene glycol)-NAD+

O

HN

HN O

H2N

EnzymeEnzyme

Fig. 22 Preparation of a malate dehydrogenase-polyethylene glycol-NAD complex.

reactor. The recycling of NAD+ in thesesystems was achieved either by a diffu-sional redox-mediator [299], or by a secondenzyme utilizing the reduced NADH fora secondary biocatalytic reaction [296] (Cf.Chapter 7.1).

Covalent coupling between syntheticNAD+-derivatives, water-soluble polymermatrices, and enzyme molecules hascombined both techniques mentionedabove. For example, polyethyleneglycolcarrying covalently bound NAD+-unitswas activated with a bifunctional reagent(25) and covalently attached to malate [288]or glucose dehydrogenase [290] (Fig. 22).

17.8.4The Integration of NAD(P)+-dependentEnzymes with Monolayer Arrays ofNAD+-cofactor and Redox-catalysts

Electrodes functionalized with monolay-ers of enzyme cofactors (e.g. NAD+-monolayers) demonstrate the ability toform stable affinity complexes with theirrespective enzymes [301]. These interfacialcomplexes can be further cross-linked toproduce integrated bioelectrocatalytic ma-trices consisting of the relay-units, thecofactor, and the enzyme molecules. Elec-trically contacted biocatalytic electrodesof NAD+-dependent enzymes have been

organized by the generation of affin-ity complexes between a catalyst/NAD+monolayer and the respective enzymes[302]. A PQQ monolayer covalently linkedto an amino-functionalized nicotinamideadenine dinucleotide, N6-(2-aminoethyl)-NAD+ (23), was assembled onto anAu-electrode. The resulting monolayer-functionalized electrode binds NAD+-dependent enzymes, such as lactate de-hydrogenase and AlDH, by affinity in-teractions between the cofactor and thebiocatalyst (Fig. 23a). These enzyme elec-trodes electrocatalyze the oxidation of theirrespective substrates (in these cases lacticacid and ethanol). The enzyme electrodereveals only temporary stability and ca.25% of the biocatalyst dissociates fromthe monolayer affinity-complex to theelectrolyte solution within 30 min. Two-dimensional cross-linking of the enzymelayer associated with the PQQ/NAD+-cofactor monolayer with glutaric dialde-hyde generates a stable, integrated, elec-trically contacted, cofactor-enzyme elec-trode. The electrical response of a cross-linked layered PQQ/NAD+-LDH electrodein the absence (curve a) and the pres-ence (curve b) of lactate and a calibrationcurve for the amperometric responses ofthe integrated LDH layered electrode tolactate are given in Fig. 23(b). This system


4

2

(a)

(b)

0

−2

4

2

00 10 20

I [µ

A]

Lactate[mM]

−0.4 −0.2 0

E [V]

I [µ

A]

a

b

SHN PQQ

PQQ

PQQ

PQQ

PQQ

PQQ

PQQ

NAD+

SHN

NAD+

LDH

SHN

NAD+

LDH

SHN

NAD+

SHN

NAD+

LDH

SHN

NAD+

LDH

Pyruvate

Lactate

e−

e−e−

N

N

LDH

LDH

H

O

H

O(Cross linking)

Fig. 23 (a) The assembly of an integrated lactate dehydrogenase monolayer-electrode by thecross-linking of an affinity complex formed between the enzyme and a PQQ-NAD+monolayer-functionalized Au-electrode; (b) cyclic voltammograms of the integratedcross-linked PQQ-NAD+/LDH electrode (roughness factor ca. 15): (a) In the absence oflactate. (b) With lactate, 20 mM. Recorded in 0.1 M Tris-buffer, pH 8.0, in the presence of10 mM CaCl2, under Ar, scan rate, 2 mV s−1. Inset: amperometric responses of the integratedelectrode at different concentrations of lactate upon application of potential 0.1 V.

17.9 Electrical Contacting by Interprotein Electron-transfer 599

exemplifies a fully integrated rigid bio-catalytic matrix composed of the enzyme,NAD+-cofactor, and catalyst. The com-plex between the NAD+-cofactor and LDHaligns the enzyme on the electrode sup-port, thereby enabling the effective electri-cal communication between the enzymeand the electrode, while the PQQ-catalyticsites provide the electrochemical regener-ation of NAD+.

17.9Electrical Contacting by InterproteinElectron-transfer

The electron-transfer cascades utilized bynature (e.g. in the respiratory and photo-synthetic schemes) tend to be very highyielding, suffering little loss through un-wanted side- and back-reactions. The key tothis efficiency lies in the insulating proteinshell and the ability of the charge-carriersto form highly selective complexes. Smallmolecular electron mediators offer little se-lectivity in their electronic reactions as theycan exchange electrons with any appropri-ate donor or acceptor that can approachthem. Proteins, on the other hand, donot allow the close contact of their redox-active site with anything except certainsubstrates. This allows their use as elec-tron mediators between specific speciesthat form complexes with them, even in thepresence of other redox-active molecules.Proteins have been used as electron me-diators in artificial systems, ranging fromsimple diffusional systems to complex im-mobilized multiprotein assemblies.

17.9.1Soluble Cytochromes as Electron-transferMediators

Cytochromes, particularly cytochrome c(Cyt c), are small heme-containing redox

proteins capable of mediating electrontransfer between redox enzymes with greatspecificity [303]. The electron-transferevent between Cyt c and enzymes oc-curs specifically in interprotein com-plexes [304].

Reversible electrochemistry of Cyt c,which includes a partially exposed heme-site, has been observed at Au-, Pt-, andAg-electrodes modified with monolayersof various promoters [305–307]. Gener-ally, these promoters are linked to theelectrode by a thiol or disulfide groupand include an organic functional unitthat interacts with the Cyt c backbone.The promoter monolayer prohibits di-rect contact between the protein and themetal electrode surface, thus preventingirreversible unfolding of the protein. Inaddition, interactions between the pro-moter and the protein molecules canresult in the specific alignment of theprotein at the electrode surface, providingshort electron-transfer distances. The mostcommon promoter for the activation ofthe interfacial electrochemistry of Cyt c isbis(4-pyridyl)disulfide [308–312], but sev-eral other thiol and disulfide derivatizedmolecules are also effective [313]. Aminoacid and oligopeptide monolayers havebeen used as promoters for enhanc-ing the interfacial electrochemistry ofCyt c [314, 315], as well as other cy-tochromes [316]. Other promoters, suchas imidazole [317], thiophene [318], andiodide [319] deposited onto Au or Ag-electrodes have also been used to facil-itate Cyt c electrochemistry. Thus, elec-trochemically contacted at a promoter-modified electrode surface, soluble Cyt ccan mediate specific electron transportevents between Cyt c-recognizing en-zymes and the electrode. Since the elec-trochemistry of Cyt c is reversible at thepromoter-modified electrodes and Cyt c


is able to donate and accept electronsto/from enzymes, both directions (ca-thodic and anodic) of electron transportare possible.

Reductive biocatalytic transformationsmediated by Cyt c have been exemplifiedin a variety of following systems. Cyt celectrochemically reduced at a promoter-modified electrode surface has mediatedthe electron-transfer to soluble laccase(from Coriolus hirsutus), which biocat-alyzed the reduction of O2 to water [320].

A similar O2 biocatalytic reduction hasbeen also demonstrated using a longersequence of inter-protein electron-transfersteps [321] (Fig. 24a). Figure 24(b), curvea, shows a cyclic voltammogram typi-cal for the reversible electrochemistry ofCyt c at an Au-electrode in the presenceof a promoter [321]. The addition of cy-tochrome c551 (Cyt c551) and cytochromeoxidase (COx), from Pseudomonas aerugi-nosa) under aerobic conditions results in asignificant increase in the cathodic current

(a)

(b)

1/2 O22 H+

H2OCOxox

COxred

2 Cyt c551 red

2 Cyt c551 ox

2 Cyt cox

2 Cyt cred

2e−

0−0.2 0.2

0

−0.4

−0.8

−1.2

I [µ

A]

E[V]

a

b

Fig. 24 (a) The bioelectrocatalyzed reduction ofdioxygen by COx, mediated by Cyt c and Cyt c551in a multistep electron-transfer process;(b) cyclic voltammograms of a promoter-modi-fied Au electrode (a) in the presence of Cyt c(5.3 mg mL−1), and (b) after the addition of

Cyt c551 (0.74 mg mL−1) and COx (770 nM).Recorded in 0.02 M phosphate buffer, pH 7.0,scan rate 1 mV s−1, in the presence of O2.[Adapted from D. A. Powis, G. D. Wattus, FEBSLett. 1981, 126, 282–284.]


(curve b), originating from inter-proteinelectron-transfer, resulting in the electro-catalytic reduction of O2. It should benoted that the addition of either Cyt c551 orCOx to the Cyt c alone does not yield theelectrocatalytic process, nor does the sys-tem generate an electrocatalytic currentin the absence of O2. Thus, consider-ing the sequence of electron transport innatural respiratory chains, one can con-clude that this artificial process proceedsin the following steps: (1) the interfa-cial electrochemical reduction of Cyt c atthe electrode, (2) the homogeneous inter-protein electron-transfer from reducedCyt c to Cyt c551, (3) the electron-transferfrom reduced Cyt c551 to COx, and fi-nally (4) electron-transfer from COx to O2

(Fig. 24a). This system illustrates that evenunder homogeneous conditions directedMET can be achieved by the exploitationof specific electron-transfer events. Appli-cation of nitrite reductase, instead of COx,results in a similar system, ultimately re-ducing NO2

− to NO [322].Oxidative biocatalytic processes medi-

ated by Cyt c have been exemplified in sev-eral systems. The electrochemical oxida-tion of Cyt c has been coupled to lactate de-hydrogenase, which catalyzes the oxidationof L-lactate [323]. Similar electron transportwas also provided by ferrocene mono-carboxylic acid mediating the electron-transfer from lactate dehydrogenase tothe electrode. Although the rate con-stants for electron-transfer between lactatedehydrogenase and Cyt c- and ferrocene-mediators are similar (6.7 × 106 M−1 s−1

and 5.0 × 106 M−1 s−1, respectively), theCyt c mediation process is much morespecific and is not affected by any inter-ferant that is not recognized by Cyt c. Thespecificity of this mediation process canbe further utilized to provide regenera-tion of NADP+, which can be connected

to any NADP+-dependent enzyme [324].In this system, NADPH-cytochrome-P450-reductase oxidizes NADPH, and later do-nates electrons to the oxidized Cyt c, whichis finally regenerated at the promoter-modified electrode.

17.9.2Heme-protein Monolayers asElectron-transfer Mediators

The first step toward the constructionof organized multiprotein/enzyme bio-electrocatalytic matrices is the monolayerdeposition of the protein that serves as abase, mediating electron-transfer to fur-ther enzyme-layers. The positively chargedCyt c molecules have been adsorbed as amonolayer by the electrostatic attractionto negatively charged electrode interfacesproduced by the self-assembly of thiolcarboxylic acids [324–330]. The specificalignment of nonsymmetrically chargedCyt c can be achieved when the Cyt c ad-sorption is performed under applicationof a potential on the electrode [331]. Af-ter immobilization, the Cyt c moleculesremain free for lateral and spinning move-ments on the electrode surface, whichaids their interaction with secondary en-zymes. Covalent binding of Cyt c to self-assembled monolayers [324, 332] usuallyresults in their random orientation, be-cause many functional groups around theCyt c backbone are involved in the immo-bilization reaction. In this case, the Cyt cmolecules are fixed on the surface andcannot adjust their orientation to com-municate with other enzymes. For exam-ple, Cyt c adsorbed onto a self-assembledmonolayer atop an electrode is capableof mediating electron-transfer from solu-ble NADPH-cytochrome-P450-reductase tothe electrode [324]. Cyt c molecules co-valently fixed at the interface did not


exhibit such properties. These results sug-gest that the electrostatically adsorbedCyt c has some mobility that allows re-orientation, as required to interact withthe electrode and with the enzyme. In

contrast, covalently bonded Cyt c has apermanent, random orientation with re-spect to the electrode surface and blocksthe Cyt c electron mediating process.Thus, alignment of the heme-protein is

SHN

H

O

H

O

O

SHN

O

NR

SHN

O

SHN

O

NR

NR

SHN

O

SHN

O

MP-11

MP-11

MP-11

MP-11 MP-11

MP-11

NR

NR

NO2− + H2O

NO3−

e−

N

N

NR

NR

0

(a)

(b)

−20

−40

−60

0−0.2−0.4−0.6

E [V]

40

20

00 20 40

I cat

[µA

]

(NO3−)

[mM]

a

b

e−

I [µ

A]

Fig. 25 (a) The assembly of an integrated nitrate sensor electrode by thecross-linking of a microperoxidase-11/NR affinity complex on an Au-electrode;(b) cyclic voltammograms of the integrated MP-11/NR monolayer-modifiedAu-electrode (roughness factor ca. 15). (a) 0.1 M phosphate buffer, pH 7.0, (b) inthe presence of KNO3, 20 mM. Potential scan rate 5 mV s−1. Inset: electrocatalyticcathodic currents (E = −0.6 V) transduced by the modified electrode at differentconcentrations of KNO3. Measurements were performed under argon.


highly important for it to interact withenzymes that require specific bindingsites.

17.9.2.1 Microperoxidase-11 MonolayersMicroperoxidase-11 (2) consists of theactive-site microenvironment of Cyt c [333,334], and is obtained by tryptic digestionof the native hemoprotein. Despite thestructural similarities between MP-11 andCyt c, the heme sites of the oligopeptideand the native protein differ substantiallyin their redox-potentials (Eo = −0.40 Vfor MP-11 [335] and Eo = +0.012 V forCyt c [305–307, 336]). Microperoxidase-11assembled [335, 337] as a monolayer onan Au-electrode provides affinity interac-tions with hemoproteins and cytochrome-dependent enzymes [61]. An MP-11 mono-layer was found to yield an affinitycomplex with the native cytochrome b5-dependent NR, (from Escherichia coli),giving an NR surface coverage estimatedat 3.8 × 10−12 mol cm−2 [337]. The asso-ciation constant of the MP-11/NR com-plex was determined by quartz crystalmicrobalance (QCM) measurements tobe Ka = 3.7 × 103 M−1. Cross-linking ofthis affinity complex layer with glutaricdialdehyde generates a stable electricallycontacted enzyme electrode that effects thebioelectrocatalyzed reduction of nitrate tonitrite (Fig. 25a). The bioelectrocatalyzedtransformation proceeds with a currentefficiency of ca. 85%, and the resultingenzyme-electrode can be employed as anamperometric nitrate sensor (Fig. 25b).

Cobalt(II)-protoporphyrin IX-reconstitu-ted Mb has also been found to forman affinity complex with MP-11 mono-layer electrodes [338, 339]. The associ-ation constant of the affinity complexbetween the reconstituted Mb and MP-11 corresponds to Ka = 1.6 × 105 M−1,and the electron-transfer rate constant in

the resulting supramolecular complex isket = 0.3 s−1. The MP-11 mediates theelectrocatalyzed reduction of Co(II)-Mband the resulting cobalt hydride hydro-genates alkynes (e.g. acetylene dicarboxylicacid to maleic acid). Cross-linking of theCo(II)-reconstituted Mb affinity complexwith glutaric dialdehyde generates a stableelectrode for the electrocatalytic hydro-genation of acetylene dicarboxylic acid witha current yield of ca. 80%.

17.9.2.2 Heme-containing De novoProtein MonolayersDe novo synthetic proteins containingheme-units and mimicking cytochrome bfunctions have been assembled onsolid supports [340, 341]. The controllablealignment of their structure and redoxproperties of their heme-units makesthem interesting for mediating electronsbetween enzymes and electrode supports.Recently, de novo synthesized four-helix polypeptides were applied to themimicking of cytochrome b functionsand to tailor layered cross-linkedelectrocatalytic electrodes. A four-helixbundle de novo protein (14,728 D), whichincludes histidine units in two of itshelices, was assembled onto Au-electrodes(Fig. 26a). Two Fe(III)-protoporphyrinIX molecules were reconstituted intothe assembly to yield a system thatdemonstrates a vectorial electron-transfercascade [342]. Thus, the application ofthis artificial heme-containing proteinas a mediator in supra-protein/enzymeassembly allows vectorial electron-transferto the biocatalytic redox enzyme. The denovo-synthesized protein assembly formsaffinity complexes with the cytochrome-dependent NR and with Co(II)-protoporphyrin IX-reconstituted Mb [343].The resulting layered complex of


B

B

B

(a)

(b)

0

−40

−80

−120

0−0.2−0.4

I[µ

A]

E [V]

0

60

120

0 40 80

I [µ

A]

(NO3−)

[mM]

a

b

c

d

e

Au

Au

Au

SHN

ON

O

O

S

SHN

ON

O

O

S

AB

A = BrAc-GGELRELHEKLAKQFEQLVKLHEERAKKLGGC-NH2B = Ac-LEELWKKGEELAKKLQEALEKGKKLAK(AcBr)-NH2

NR

SHN

ON

O

O

S

SHN

ON

O

O

S A

NR

NO3−

NO2−

e−

e−e−

PGCACPG

CAC

=

SHN

ON

O

O

S

SHN

ON

O

O

S A

NR H

O

H

O

= Fe(III) Protoporphyrin IX

A

AB

AB

Fig. 26 (a) The assembly of a nitrate sensing electrode by the cross-linking of an affinitycomplex formed between NR and a Fe(III)-protoporphyrin reconstituted de novo fourhelix-bundle protein; (b) cyclic voltammograms of the NR/two heme-reconstituted de novoprotein-layerd Au-electrode at nitrate concentrations of (a) 0, (b) 12, (c) 24, (d) 46 and(e) 68 mM. Inset: calibration curve for the amperometric response of the electrode at differentnitrate concentrations (at E = −0.48 V). Potential scan rate 5 mV s−1, 0.1 M phosphatebuffer, pH 7.0, under argon, electrode roughness factor ca. 20.


COx

COx

COx

COx

e−e−

H

O

H

O(Cross linking)

O2

H2O

SHN

O NO

O Cys

Cyt c

Cyt c

S

SHN

O NO

O CysS

SHN

O NO

O Cys

Cyt c

S

SHN

O NO

O Cys

Cyt c

S

Fig. 27 The cross-linking of COx in an affinity complex with asurface-aligned Cyt c-electrode to form an integrated assemblycapable of the bioelectrocatalytic reduction of O2.

Fe(III)-de novo protein/NR, or Fe(III)-de novo protein/Co(II)-reconstitutedMb were cross-linked with glutaricdialdehyde to yield electrically contactedelectrocatalytic electrodes. The Fe(III)-de novo protein/NR-electrode assemblywas applied for the electrocatalyzedreduction of NO3

− to NO2− and

acted as an amperometric NO3−

sensor (Fig. 26b), and the Fe(III)-de novo protein/Co(II)-reconstitutedMb integrated electrode stimulatedthe electrocatalyzed hydrogenation ofacetylene dicarboxylic acid to malic acid.

17.9.2.3 Cyt c-aligned MonolayersAssociated with Cytochrome OxidaseElectrical contacting of iso-2-cytochrome c(iso-2-Cyt c, from Saccharomyces cerevisiae)can be achieved by the covalent linkingof its cysteine residues to an electrode-bound maleimide monolayer [344]. Theassociation constant between the iso-2-Cyt c-monolayer and COx is very high(Ka = 1.2 × 107 M−1) [301], and an inte-grated iso-2-Cyt c/COx electrode is, there-fore, prepared with relative ease (Fig. 27).The iso-2-Cyt c monolayer electrode wasfirst interacted with COx to generate the


affinity complex on the surface with analmost close-packed surface density (ca.2 × 10−12 mole cm−2). The resulting lay-ered complex was cross-linked with glu-taric dialdehyde to yield the integratedelectrically contacted electrode. The spe-cific orientation of the mediator on thesurface along with the restrictions placedon the enzyme by its closely packedstate lead to a complex that is highlyoptimized for efficient electron-transfer.Noncatalyzed electrochemical reduction ofO2 at a bare Au electrode proceeds withan overpotential that is even larger whenthe electrode is covered with iso-2-Cyt c orCOx alone, because of the partial block-ing of the electrode surface. Thus, whenthe electrode surface covered with eitheriso-2-Cyt c or COx, the electocatalysis of O2

reduction is not achieved. For a layered iso-2-Cyt c/COx-cross-linked-electrode in thepresence of O2, an electrocatalytic O2-reduction-wave is observed at potentialsmore negative than −0.07 V, indicatingthat the iso-2-Cyt c/COx layer acts as abiocatalytic interface for the reduction ofdioxygen. Thus, the effective bioelectro-catalyzed reduction of O2 by the iso-2-Cyt c/COx interface originates from directelectrical communication from the elec-trode to iso-2-Cyt c to COx, which acts asan electron storage biocatalyst for the mul-tielectron reduction. Systems providingfour-electron transfer mechanisms for thereduction of O2 are scarce, yet they are ofextreme importance for the developmentof O2-based biofuel cells [345].

17.10Applications of Enzymes ElectricallyContacted by Mediated Electron-transfer

Although the study of enzyme elec-trochemistry is of great value for the

understanding of natural processes, italso opens up interesting prospectsfor the development of new technolo-gies. In the utilization of redox en-zymes for the construction of new de-vices, the biomolecules may be ap-plied, either for their natural functions(e.g. in biofuel cells and biocatalysis)or for functions that are not per-formed in nature (e.g. in quantitativesensors).

17.10.1Biosensors Based on Electrically ‘‘Wired’’Enzyme Electrodes

The electrochemical activation of enzymeelectrodes results in the bioelectrocat-alyzed oxidation or reduction of a substratespecific to the biocatalyst. Provided theMET is fast, it is controlled by the sub-strate concentration. Hence, the mediator-enzyme assemblies provide a basis forthe construction of quantitative analyticalbiosensors [346]. The continuous monitor-ing of endogenous compounds or drugsby implantable biosensors enables theclose surveillance of patients via a rapidreturn of clinical information [347, 348].Such real-time measurements are thushighly desired in intensive care units,during surgery and for the managementof diabetes, as they offer an early warn-ing of changes in a patient’s condition,allowing rapid corrective action to beundertaken. The analysis of blood glu-cose levels in diabetics is just one examplewhere such cheap and continuous moni-toring is particularly desirable [346]. Sev-eral other important metabolites can alsobe readily detected using amperometricbiosensors based on electrically ‘‘wired’’enzymes (e.g. biosensors for lactate [349,350], bilirubin [40, 351], amino acids [352],and peptides [353] have been developed

17.10 Applications of Enzymes Electrically Contacted by Mediated Electron-transfer 607

using lactate oxidase or lactate dehy-drogenase, bilirubin oxidase, amino acidoxidase, and tyrosinase, respectively). Am-perometric biosensors for the monitoringof drugs have also received considerable at-tention (e.g. for assay of theophylline [354]using theophylline oxidase). In addition tomedical applications, electrically ‘‘wired’’redox enzymes have found many uses infood technology and biotechnology (e.g. foranalysis of carbohydrates, organic acids, al-cohols, additives, pesticides and fish/meatfreshness [355, 356]), in environmentalmonitoring (e.g. for analysis of pollutants,pesticides [357–359]), and in defense ap-plications (e.g. for detection of toxins,pathogenic bacteria [360, 361]).

17.10.2Bioelectrocatalyzed Synthesis by ‘‘Wired’’Enzyme Assemblies

The coupling of enzymatic and electro-chemical reactions has provided efficienttools, not only for analytical but also forsynthetic purposes. In the latter field, thepossibilities of enzymatic electrocatalysis,for example, the coupling of glucose oxida-tion (catalyzed either by GOx or GDH)to the electrochemical regeneration ofa co-substrate (benzoquinone or NAD+)have been demonstrated [362–364]. Anelectroenzymatic reactor has also beendeveloped [363–364] to demonstrate theproduction of biochemicals on a laboratoryscale. NAD(P)+ derivatives immobilizedby covalent attachment to polymer ma-trices or protein backbones have beenused in enzyme reactors [365, 366]. An-other important coenzyme ubiquinone canbe regenerated at an electrode [367, 371]and applied to drive secondary enzymaticreactions with the participation of mem-brane enzymes (e.g. fumarate reductase),

resulting in the production of biologicallyimportant compounds [372].

A biosynthetic multienzyme reactionof particular interest involves carbondioxide fixation with the production ofmethanol [373, 374]. FDH catalyzes the re-duction of carbon dioxide to formate, andmethanol dehydrogenase (MDH) catalyzesthe reduction of formate to methanol. Bothof these enzymes require NAD+/NADH-cofactor, and in the presence of the reduceddimethyl viologen mediator (MV•+), theycan drive a sequence of enzymatic reac-tions. The cascade of biocatalytic reactionsresults in the reduction of CO2 to formatecatalyzed by FDH, followed by the reduc-tion of formate to methanol catalyzed byMDH. A more complex system composedof immobilized cells of Parococcus deni-trificans has been demonstrated for thereduction of nitrate and nitrite [375].

17.10.3Biofuel Cells Based on ‘‘Wired’’ EnzymeAssemblies

Considering the need to explore energyalternatives, the development of biofuelcells is an extremely attractive applicationof bioelectrocatalytic electrodes. Biofuelcells utilize biocatalysts for the conversionof abundant raw materials (e.g. methanol,glucose) to electrical energy usingmolecular oxygen or H2O2 as anoxidizer [376–379]. The fact that anelectrically ‘‘wired’’ enzyme can catalyzethe oxidation or reduction of substratesenables the use of enzyme-modifiedelectrodes as key elements of biofuel cells.With this possibility in mind, a biofuel cellelement based on the bioelectrocatalyticoxidation of glucose by H2O2 wasconstructed [380] (Fig. 28a). An electricallycontacted GOx monolayer (assembled on


(a)

(b)

+

−

−−

+

+ −

−+

+

−

+

GOx

FADPQQ

Glucose

Gluconicacid

e− e− H2O2

H2O

Anode Cathode

e−

Mem

brane

Electrolyte Electrolyte

V

A

Vcell

Icell

Rload

MP-11

Rcell

40

20

00 5 10

400

300

200

100

00 100 200 300 400

Vcell [mV]

I cel

l [µ

A]

R[kΩ−1 ]

Pow

er[µ

W−1

]

Fig. 28 (a) Schematic configuration of a biofuelcell employing glucose and H2O2 as fuel andoxidizer, and using PQQ-FAD/reconstituted GOxand MP-11-functionalized electrodes asbiocatalytic anode and cathode, respectively;

(b) current-voltage behavior of the biofuel cell atdifferent external loads. Inset: Electrical powerextracted from the biofuel cell at differentexternal loads.

an electrode by the reconstitution of apo-GOx on a PQQ-FAD layer, Fig. 17) wasused as the anode, and a microperoxidase-11-functionalized electrode was applied

as the cathode. The bioelectrocatalyzedoxidation of glucose proceeds at the anode,while the MP-11 catalyzed the reductionof H2O2 in the catholyte compartment.

17.10 Applications of Enzymes Electrically Contacted by Mediated Electron-transfer 609

(a)

(b)

GOx

FADPQQ

Glucose

Gluconicacid

e− e− O2

H2O

Anode Cathode

e−

Electrolyte

V

A

Vcell

Icell

Rload

COx

Cyt c

e−

0.10

0.08

0.06

0.04

0.02

00 40 80 120

Vcell [mV]

I cel

l [m

A] 4

3

2

1

00 1 2 3

R [kΩ]

Pow

er

[µW

]

Fig. 29 (a) Schematic configuration of anoncompartmentalized biofuel cell employingglucose and O2 as fuel and oxidizer, and usingPQQ-FAD/GOx- and Cyt c/COx-functionalizedelectrodes as biocatalytic anode and cathode,

respectively; (b) current-voltage behavior of thebiofuel cell at different external loads. Inset:Electrical power extracted from the biofuel cell atdifferent external loads.


The current-voltage behavior of the biofuelcell at different external loads and atoptimized concentrations of the oxidizerand fuel substrates are shown in Fig. 28(b).Up to 32 µW could be extracted at anexternal load of 3 k (Fig. 28b, inset). Theconfiguration of this biofuel cell representsa very general method for the developmentof future biofuel cell elements as it can beextended to other oxidative enzymes andfuel substrates, such as methanol oxidaseor lactate oxidase and the respective alcoholor lactic acid fuel substances.

The next generation of biofuel cellscould utilize complex, ordered enzymeor multienzyme systems immobilized atboth electrodes, even removing the needfor compartmentalization of the anodeand the cathode. In a working example,the anode described above (based on thereconstituted GOx), together with a cath-ode based on an aligned iso-2-Cyt c (fromSaccharomyces cerevisiae)/COx couple wasconstructed, providing the reduction ofO2 to water (Cf. Chapter 8.2.3) (Fig. 29a).Since the reconstituted GOx provides ex-tremely efficient biocatalyzed oxidation ofglucose that is unaffected by oxygen, theanode can operate in the presence of oxy-gen. Thus, this biofuel cell uses O2 as anoxidizer and glucose as a fuel without theneed for compartmentalization [345]. Thecell operation was studied at different ex-ternal loads (Fig. 29b), and achieved a fillfactor of ca. 40% with a maximum power4 µW at an external load of 0.9 k.

17.11The External Control of theElectron-transfer Process

We have seen how the functions of var-ious enzyme and multienzyme systemscan be persuaded to operate in synthetic

environments with a knowledge of designconsiderations and use of electron-transfermediators. The assemblies that have beendiscussed are of use, both for fundamentalresearch and for real-world applications,such as biosensors and biofuel cells. Whatthey all lack, however, is the ability to con-trol their activity after they have been putinto operation. In some cases, the electrodesensitivity can be ‘‘tuned’’, for instance, bythe number of biocatalyst layers, but thefinal structure cannot be changed. Assem-blies that can be ‘‘tuned’’ or turned onand off during their operation are highlydesirable. The use of such interfaces couldallow the construction of variable outputfuel cells or variable range sensors. It iseven possible that architectures with rudi-mentary information processing abilitiescould be assembled. For instance, photo-switchable biomaterials open a route tooptobioelectronic systems [381].

17.11.1Photochemical Control by Enzyme-boundPhotoisomerizable Units

The activity of biocatalysts is highly depen-dant on their ability to adopt their naturalconformation. We have already seen thatenzymes that are functionalized with toomany relay groups begin to lose their ac-tivity due to the disruption of their correctshape. Photoisomerizable molecules un-dergo reversible structural changes whenexposed to specific energies of light. Theirtwo states often differ considerably, for in-stance, in their charge, polarization, andshape. If an enzyme molecule is func-tionalized with photoisomerizable groups,then the local environment of that enzymewill depend on the state of those groups.As the conformation of the enzyme is verysensitive to its environment, the shape and,therefore, the activity of the enzyme may

17.11 The External Control of the Electron-transfer Process 611

Substrate

Product

e−

Enzyme A

Substrate

Enzyme Bh 1

h 2

"ON" "OFF"

S MON

NO2

+N

HO

NO2

Lys

Lys

Lys

LysS

Glucose

(a)

(b)

Gluconicacid

Fc-CO2HFc+-CO2H

e−

S

S

S

Lys

Lys

Lys

LysS

Glucose

Fc-CO2HFc+-CO2H

e−

M

M

M320 nm< <

360 nm

> 495 nm

Active GOx Inactive GOx

SS

O

ON

O

O

O

ON

O

O

S

O

ON

O

OLys

Lys

Lys

Lys

S

S

S

H2N

26a 26b

ν

ν

λ

λ

Fig. 30 (a) The assembly of anoptobioelectronic system by the immobilizationof an enzyme tethered with photoisomerizableunits; (b) a nitrospiropyran-functionalized GOx

layer for the photo-controlled oxidation ofglucose, with ferrocene carboxylic acid as adiffusional electron mediator.


be controlled by photonic input (Fig. 30a).If the enzyme is electrically contactedby an electron mediator, then there is apossibility that photoisomerization maycontrol the enzyme activity by changingthe ability of the mediator to interact with

the redox-center (as opposed to changingthe catalytic ability of the enzyme).

The enzyme GOx was transformedinto a photoswitchable enzyme by itschemical modification with nitrospiropy-ran photoisomerizable units [382]. The

GOx GOx

FAD

FAD FAD FAD

FAD

FAD FAD

FAD

FAD

FAD FAD

Native GOx Apo-GOx Reconstituted GOx

S

M

ON

NO2

+N

HO

NO2

S

e−

Glucose

Gluconicacid

Fc+

CO2H

FcCO2H

GlucoseFc+

CO2H

FcCO2H

S M

SS

NH2

NH2 O

H H

O

S M

S

S N NLys S

+ ON

NO2

O

HO

EDC

(a)

(b)

(c)

S M

380 nm> >

360 nmλ

λ > 475 nm

λ > 475 nm

380 nm> >

360 nmλ

λ >475 nm

320 nm> >

380 nmλ

19

27

28a 28b

Fig. 31 (a) The synthesis of a photoisomerizable semisynthetic FAD; (b) reconstitution of thesemisynthetic FAD into apo-GOx to give a photoisomerizable enzyme; (c) immobilization of thephotoisomerizable enzyme to give an assembly with photoswitchable bioelectrocatalyticfunctions.


nitrospiropyran units (26a) are photoiso-merized to the protonated merocyanine-state (26b) upon irradiation with 360 nm <

λ < 400-nm light. The opposite photoiso-merization from (26b) to (26a) proceedsupon illumination with visible light (λ >

475 nm). The structurally modified en-zyme was assembled as a monolayer on anAu-electrode (Fig. 30b). In the presence ofthe diffusional electron mediator ferrocenecarboxylic acid, the (26a)-functionalizedGOx was activated for the oxidation ofglucose, as evidenced by the high elec-trocatalytic current resulting from theoxidation of glucose. Photoisomerizationof the monolayer to the (26b)-state re-sults in a structural distortion of theprotein that prevents electrical contactwith the electron mediator, and a sub-stantially lower amperometric response.By the cyclic photoisomerization of the en-zyme monolayer between the (26a) and(26b) states, reversible ‘‘high’’ and ‘‘low’’amperometric responses of the functional-ized electrode are transduced. Incomplete‘‘on’’-‘‘off’’ bioelectrocatalytic functions ofthe enzyme-electrode were observed. Thiswas attributed to the insufficient distortionof the protein by photoisomerizable unitsattached randomly, far from the active site,thereby, perturbing only to limited extentthe biocatalytic activity of the enzyme.

To optimize the photoswitchable bio-electrocatalytic features of redox en-zymes, the site-specific functionalizationor mutation of the active-site microenvi-ronment is essential. This has been ac-complished by a semisynthetic approach,involving the reconstitution of the flavoen-zyme GOx with a semisynthetic photoi-somerizable FAD-cofactor [383, 384]. Ni-trospiropyran carboxylic acid (27) wascovalently coupled to N6-(2-aminoethyl)-FAD (19) to yield a photoisomeriz-able nitrospiropyran-FAD cofactor (28a)

(Fig. 31a). The native FAD cofactor wasextracted from GOx and the semisyn-thetic FAD cofactor was reconstituted intothe apo-GOx (apo-GOx) (Fig. 31b). Thisreconstituted enzyme includes a photoi-somerizable unit directly attached to theredox center of the enzyme, and hence, theenzyme is predisposed for optimized pho-toswitchable bioelectrocatalytic properties.The photoisomerizable enzyme was as-sembled on an Au-electrode as describedin Fig. 31(c). The bioelectrocatalytic oxi-dation of glucose was stimulated in thepresence of ferrocene carboxylic acid asa diffusional electron-transfer mediator.The (28a)-state of the reconstituted GOxwas inactive for the bioelectrocatalytictransformation, whereas photoisomeriza-tion of the enzyme to the (28b)-stateactivated the system (Fig. 32). By the cyclicphotoisomerization of the enzyme mono-layer between (28a) and (28b) states, thebioelectrocatalyzed oxidation of glucosewas cycled between the ‘‘off’’ and ‘‘on’’states, respectively (Fig. 32, inset). It wasalso found that the direction of the photo-bioelectrocatalytic switch of the (28a/28b)-FAD-reconstituted GOx is controlled bythe electrical properties of the diffusionalelectron-transfer mediator [385]. With fer-rocene dicarboxylic acid as a diffusionalelectron-transfer mediator, the enzyme inthe (28a)-state was found to correspond tothe switched ‘‘off’’ biocatalyst, while the(28b)-state exhibits switched ‘‘on’’ behav-ior. In the presence of the protonated 1-[1-(dimethylamino)ethyl]ferrocene, the direc-tion of the photobioelectrocatalytic switchis reversed. This control of the photoswitchdirection of the photoisomerizable GOxwas attributed to electrostatic interactionsbetween the diffusional electron-transfermediator and the photoisomerizable unitlinked to the FAD. The (28b)-state at-tracts the oxidized negatively charged


200

100

0

0 0.2 0.4

E [V]

I [n

A]

0

40

80

120I

[nA

]

a

b

c

d

ms s sm m

Fig. 32 Cyclic voltammograms of an electrode bearing amonolayer of GOx reconsitiuted with photoisomerizable dyad28(a/b) in the presence of glucose (50 mM), ferrocenecarboxylic acid (50 µM), and with the photoisomerizable units(a, c) in the spiropyran-state (28a), and (b, d) in themerocyanine-state (28b). Recorded in 0.01 M phosphate buffer,pH 7.3, scan rate 5 mV s−1. Inset: switching behavior of theelectrocatalytic current as a function of the state of thephotoisomerizable group: ‘‘s’’ and ‘‘m’’ represent thephotoisomerizable units in the spiropyran and the merocyaninestates, respectively.

electron-transfer mediator, but repels theoxidized positively charged relay. As a re-sult, this photoisomer state of the enzymeis switched ‘‘on’’ in the presence of neg-atively charged mediator, but exists inthe ‘‘off’’ state, using positively chargedelectron-transfer mediator.

17.11.2Photochemical Control by Electrode-boundPhotoisomerizable Units

Photonic control over electroactivated bio-catalytic processes can also be achievedby the use of electrode surfaces thatare modified by photoisomerizable units.These interfaces, whose state controlsthe ability of a substrate to interactwith them, are known as ‘‘command’’surfaces [382, 386–390]. In one example, a

mixed monolayer consisting of pyridine-4-thiol and nitrospiropyran thiol (29a)revealed ‘‘command’’ interface propertiesfor Cyt c electrochemistry [386, 388]. Ona gold surface, a pyridine-4-thiol mono-layer functions as a promoter for the Cyt celectrochemical reduction/oxidation pro-cess [305–307]. The monolayer composedof the pyridine-promoter groups andneutral spiropyran units (29a) providesefficient reduction and oxidation of Cyt c,allowing enzymatic reactions that requirethe participation of a Cyt c-dependent en-zyme. A biocatalytic cascade of reductivereactions was achieved in the presence ofCOx and O2 (Fig. 33a). Oxidative reactionswith the participation of lactate dehydroge-nase resulted in lactate oxidation (Fig. 33c).When the photoisomerizable groups aretransformed into the positively charged


ON

NO2

O

HNS

S N

+N

HO

NO2

O

HNS

S N

S N

S N

LDH(ox)

Cyt c(red)

Cyt c(ox)

COx(ox)

COx(red)

LDH(red)

+++

+++

e−

O2 + 4 H+

2 H2O

Cyt c(ox)

+++Repulsion

ON

NO2

O

HNS

S N

S N

Cyt c(red)

Cyt c(ox)

+++

+++

e−

Pyruvate

Lactate

29a

29b

29a

> 475 nmλ

> 475 nmλ

(b)

(c)

(a)

320 nm < < 380 nm λ

320 nm < < 380 nm λ

Fig. 33 Coupling of the photoswitchable interactions between Cyt c and a (29a)-pyridinemixed monolayer with: (a) the reduction of O2 by COx and (c) the oxidation of lactate byLDH; (b) when the electrode is in the cationic merocyanine state (29b), repulsiveinteractions disallow the functioning of bioelectrocatalytic processes.

protonated merocyanine state (29b), thepositively charged Cyt c molecules are re-pelled from the electrode surface andelectrical contact is no longer established(Fig. 33b). In this case, not only Cyt c elec-trochemistry is inhibited, but also any sec-ondary reactions. Switching between theelectrochemically contacted system andthe electrochemically inactive system wasstudied by CV for Cyt c only (Fig. 34a) andin the presence of the secondary enzyme

with the respective substrate (e.g. LDHand lactate) (Fig. 34b). Reversible switch-ing between electrochemically active andinactive states was obtained (Fig. 34a&b,insets). Related photoisomerizable com-mand interfaces have been used to con-trol the biocatalytic activity of GOx [387,390] and the electrocatalytic regenera-tion of NAD+-cofactor [391]. Commandsurfaces have also been used to detectvarious signals (temperature change [392],


2

1

0

−1

−2−0.4 0

(a) (b)

0.4

E [V]

E [V]

I [µ

A]

I [µ

A]

0

0.4

0.8

1.2I p

eak

[µA

]

I pea

k[µ

A]

0

−1

−2−0.2 0 0.2

a

b

0

0.4

0.8

1.2

a

b

Fig. 34 (a) Cyclic voltammetric response ofCyt c (0.1 mM) at a mixed (29(a/b)) and pyridinemonolayer-modified electrode in (a) the neutralspiropyran-state (29a) and (b) the positivelycharged merocyanine-state (29b), recorded at50 mV s−1. Inset: switching behaviour of theCyt c electrochemistry as a function of the stateof the photoisomerizable group: circles andsquares represent the photoisomerizable unitsin the spiropyran and the merocyanine states,respectively; (b) cyclic voltammetric response ofCyt c (0.1 mM) with COx (1 µM) at a mixed

(29(a/b)) and pyridine monolayer-modifiedelectrode in the presence of O2 and in (a) theneutral state (29a), and (b) the cationicmerocyanine state (29b). Inset: switchingbehaviour of the electrocatalytic current as afunction of the state of the photoisomerizablegroup: circles and squares represent thephotoisomerizable units in the spiropyran andthe merocyanine states, respectively. Allexperiments were performed in 0.1 M phosphatebuffer, pH 7.0.

pH change [393]) by the variation of redoxcofactor regeneration rates (therefore bythe control of enzymatic activity). Thus,they represent examples of biocatalyticswitches.

17.11.3Photochemical Control by Mediator-boundPhotoisomerizable Units

Photoswitchable electrical communicationbetween enzymes and electrodes hasalso been achieved by the application ofphotoisomerizable electron-transfer me-diators [386, 389] (Fig. 35). Diffusionalelectron mediators (viologen (30) or fer-rocene (31) derivatives) were function-alized with photoisomerizable spiropy-ran/merocyanine units. These mediatorscan be reversibly photoisomerized from

the spiropyran-state (30a, 31a) to themerocyanine-state (30b, 31b) (360 < λ <

380 nm) and back (λ > 475 nm). An en-zyme multilayer array composed of glu-tathione reductase or GOx was only elec-trically contacted when the photoactivegroup linked to the redox relay (violo-gen or ferrocene derivative, respectively)was in the spiropyran-state (30a, 31a)(Fig. 35a). Cyclic activation/deactivation ofthe enzyme arrays was achieved upon pho-tochemical isomerization of the electron-transfer mediators between (30a, 31a) and(30b, 31b) states (Fig. 35b). Thus, the appli-cation of photoisomerizable redox relaysprovides a novel means of controllingthe electrical communication of redox en-zymes with electrode surfaces. The lackof electrical interactions between the re-spective enzymes and the mediators in the

17.12 Conclusion and Perspectives 617

(b)

(a)

FeO

O

+N N

+CH3

R

"ON"

"OFF"0

2

4

6

v [m

M s

−1] 30

31

s m s m

M

Enzyme

e−

S

Enzyme

S Rred/ox

Rred/oxRox/red

Substrate

Product Enzyme

e−

Enzyme

30a or31a

30b or31b

> 475 nmλ

380 nm> >

360 nmλ

Fig. 35 (a) The electrochemical contacting of amultilayer array consisting of either GOx or GRin the presence of diffusional mediators (ferro-cene and viologen, respectively) tethered tophotoisomerizable units (30 and 31); (b) switch-ing behaviour of the bioelectrocatalytic function

of a multilayer array of GR as a function of thestate of a photoisomerizable group attached to adiffusional viologen mediator (30): ‘‘s’’ and ‘‘m’’represent the photoisomerizable units in thespiropyran (30a) and the merocyanine (30b)states, respectively.

merocyanine state could originate fromsteric constraints or electrostatic repulsiveinteractions between the isomers and en-zymes, preventing the intimate contact ofthe mediator and the enzyme redox centerrequired for electron-transfer.

17.12Conclusion and Perspectives

The electrical contacting of redox en-zymes or redox-active proteins with elec-trodes is a fundamental concept in therapidly developing field of bioelectronics.Molecular, macromolecular or biomolecu-lar electron-transfer mediators can act asdiffusional electron carriers that shuttle

charges between the protein redox-sitesand the electrodes. Electron mediatorsmay also act as electrocatalysts that shuttleelectrons between cofactors and electrodes(e.g. FADH2, NAD(P)H) or activate com-plex multi-enzyme biocatalytic cascades.The integration of electron mediators,redox-proteins, and conductive supports inrigidified assemblies leads to the organi-zation of bioelectronic devices. Nanoengi-neering of electrodes with electron-relayunits or electron-relay-functionalized pro-teins to yield electrically contacted enzymeelectrodes represents a major recent ad-vance in bioelectronics.

The tailoring of electrical contact be-tween redox enzymes and electrodesprovides a basis for electrochemical


biosensors. Besides the broad applicationof electrically contacted enzyme-electrodesfor the amperometric analysis of enzyme-substrates, one may apply electrically con-tacted enzymes as redox labels for thesensing of biorecognition events, suchas antigen-antibody [385, 394, 395] oroligonucleotide-DNA [396] complexation.The use of bioelectrocatalytic electrodesas active components of biofuel cells isalso appealing. In particular, the develop-ment of catalytic electrodes that can drivethe reduction of O2 to water may con-tribute important interfaces for biologicalor chemical fuel-cell elements [345].

Future efforts at improving and ex-tending the scope of electron-transfermediation may take any of several ap-proaches. The design of mediators maybe carefully tuned to accomplish ideal re-dox potentials and specific interactions,and methods of immobilization and coim-mobilization that minimize losses in ac-tivity may be developed. The modifica-tion of the enzymes themselves, eitherby mutagenesis, de novo synthesis orsite-specific reactions may lead to struc-tures with readily accessible active sitesor with mediators attached in optimizedpositions. The construction of carefully in-tegrated multicomponent systems couldlead to highly efficient biosensing, bio-catalytic, and biofuel devices. The controlof electron-transfer events may also seea growth of interest. Photochemical con-trol of bioelectrocatalysts has already beenestablished by the use of photoisomeriz-able units, photocommand interfaces, andphotoactive electron-relays. Other switch-able biocatalyzed transformation can beenvisaged by the application of magnetic,thermal or pH changes as external trigger-ing signals. These addressable assembliescould even find uses in information stor-age and processing.

Acknowledgment

This research is supported by the IsraeliMinistry of Science, Infrastructure Projectof Biomicroelectronics, and the MINERVAGerman-Israeli Program (DIP).

References

1. J. A. Cowan, H. B. Gray, Chem. Scr. 1988,28A, 21–26.

2. A. G. Sykes, Chem. Soc. Rev. 1985, 14,283–321.

3. S. S. Isied, Prog. Inorg. Chem. 1984, 32,443–517.

4. S. E. Peterson-Kennedy, J. L. McGourty,P. S. Ho et al., Coord. Chem. Rev. 1985, 64,125–133.

5. G. McLendon, T. Guarr, M. McGuire et al.,Coord. Chem. Rev. 1985, 64, 113.

6. T. Tanaka, K. Takehaka, H. Kawamura et al.,J. Biochem. (Tokyo) 1985, 99, 833.

7. J. R. Miller in Antennas and Reaction Centersof Photosynthetic Bacteria (Ed.: M. E. Michel-Beyerle), Springer-Verlag, Berlin, 1985,p. 234.

8. M. Bixon, J. J. Jortner, J. Phys. Chem. 1986,90, 3795.

9. R. A. Marcus, N. Sutin, Biochem. Biophys.Acta 1985, 811, 265.

10. D. DeVault, Quantum Mechanical Tunnelingin Biological Systems, 2nd ed., CambridgeUniversity Press, Cambridge, 1984.

11. A. K. Churg, R. M. Weiss, A. Warshel et al.,J. Phys. Chem. 1983, 87, 1683–1694.

12. S. Larson, J. Chem. Soc., Faraday Trans. 21983, 79, 1375.

13. J. J. Hopfield, Proc. Natl. Acad. Sci. U.S.A.1974, 71, 3640–3644.

14. J. N. Onuchic, D. N. Beratan, J. J. Hopfield,J. Phys. Chem. 1986, 90, 3707–3721.


16. A. L. Ghindilis, P. Atanasov, E. Wilkins,Electroanalysis 1997, 9, 661–674.

17. W. Schuhmann, Biosens. Bioelectron. 1995,10, 181–193.

18. I. V. Berezin, V. A. Bogdanovskaya, S. D.Varfolomeev et al., Dokl. Akad. Nauk SSSR1978, 240, 615–618.

19. A. I. Yaropolov, V. Malovik, S. D. Var-folomeev et al., Dokl. Akad. Nauk SSSR1979, 249, 1399–1401


20. J. Zhao, R. Henkens, J. Stonehuerner et al.,J. Electroanal. Chem. 1992, 327, 109–119.

21. S. Yabuki, F. Mitzutani, Electroanalysis 1997,9, 23–25.

22. P. N. Bartlett, P. Tebbutt, R. C. Whitaker,Prog. React. Kinetics 1991, 16, 55–155.

23. G. Davis, Biosensors 1985, 1, 161–178.24. G. Davis in Biosensors: Fundamentals and Ap-

plications, (Eds.: A. P. F. Turner, I. Karube,G. S. Wilson), OUP, Oxford, 1987, p. 247.

25. R. S. Nicholson, I. Shain, Anal. Chem. 1965,37, 178–190.

26. Y. A. Aleksandrovskii, L. V. Bezhikina, Y. V.Rodionov, Biokhimiya 1981, 46, 708–716.

27. J. Mahenc, H. Aussaresses, C. R. Hebd.Seances. Acad. Sci. Ser. C 1979, 289,357–359.

28. A. L. Crumbliss, H. A. O. Hill, D. J. Page,J. Electroanal. Chem. 1986, 206, 327–331.

29. I. Taniguchi, S. Miyamoto, S. Tomimuraet al., J. Electroanal. Chem. 1988, 240,333–339.

30. A. E. G. Cass, G. Davis, G. D. Francis et al.,Anal. Chem. 1984, 56, 667–671.

31. J. E. Frew, H. A. O. Hill, Philos. Trans. R.Soc. London B 1987, 316, 95–106.

32. B. A. Feinberg, M. D. Ryan in Topics inBioelectrochemistry and Bioenergetics, (Ed.: G.Milazzo), John Wiley & Sons, 1981, p. 225,Vol. 4.

33. S. Wherland, H. B. Gray, Proc. Natl. Acad.Sci. U.S.A. 1976, 73, 2950–2954.

34. H. C. Davis, L. Smith, A. R. Wasserman,Biochim. Biophys. Acta 1964, 85, 238–246.

35. A. D. Ryabov, E. M. Tyapochkin, S. D.Varfolomeev et al., Bioelectrochem. Bioenerg.1990, 24, 257–262.

36. S. M. Zakeeruddin, M. Gratzel, D. M.Fraser, Biosens. Bioelectron. 1996, 11,305–315.

37. A. D. Ryabov, A. Amon, R. K. Gor-batova et al., J. Phys. Chem. 1995, 99,14 072–14 077.

38. C. Deshaies, J. Chopineau, J. Moiroux et al.,J. Phys. Chem. 1996, 100, 5063–5069.

39. I. Willner, A. Riklin, B. Shoham et al., Adv.Mater. 1993, 5, 912–915.

40. B. Shoham, Y. Migron, A. Riklin et al.,Biosens. Bioelectron. 1995, 10, 341–352.

41. P. Pantano, T. H. Morton, W. G. Kuhr,J. Am. Chem. Soc. 1991, 113, 1832,1833.

42. T. Hoshi, H. Takeshita, J. Anzai et al., Anal.Sci. 1995, 11, 311,312.

43. X. Du, J. Anzai, T. Osa et al., Electroanalysis1996, 8, 813–816.

44. C. Bourdillon, C. Demaille, J. Gueriset al., J. Am. Chem. Soc. 1993, 115,12 264–12 269.

45. C. Bourdillon, C. Demaille, J. Moiroux et al.,J. Am. Chem. Soc. 1994, 116, 10 328,10 329.

46. A. Riklin, I. Willner, Anal. Chem. 1995, 67,4118–4126.

47. E. Katz, D. D. Schlereth, H.-L. Schmidt,J. Electroanal. Chem. 1994, 367, 59–70.

48. N. C. Foulds, C. R. Lowe, J. Chem. Soc.,Faraday Trans. 1 1986, 82, 1259–1264.

49. M. Umana, J. Waller, Anal. Chem. 1986, 58,2979–2983.

50. C. Iwakura, Y. Kajiya, H. Yoneyama,J. Chem. Soc., Chem. Commun. 1988,1019,1020.

51. F. Mizutani, M. Asai, Bull. Chem. Soc. Jpn.1988, 61, 4458–4460.

52. S.-I. Yabuki, H. Shinohara, M. Aizawa,J. Chem. Soc., Chem. Commun. 1989,945,946.

53. P. Janda, J. Weber, J. Electroanal. Chem.1991, 300, 119–127.

54. Y. Kajiya, H. Sugai, C. Iwakura et al., Anal.Chem. 1991, 63, 49–54.

55. J. Li, S. N. Tan, H. L. Ge, Anal. Chim. Acta1996, 335, 137–145.

56. J. Li, S. N. Tan, J. T. Oh, J. Electroanal.Chem. 1998, 448, 69–77.

57. J. Li, L. S. Chia, N. K. Goh et al., Sens.Actuators B 1997, 40, 135–141.

58. J. Li, L. S. Chia, N. K. Goh et al., Anal. Chim.Acta 1998, 362, 203–211.

59. V. J. Razumas, A. V. Gudavicius, J. J. Kulys,J. Electroanal. Chem. 1986, 198, 81–87.

60. F. Patolsky, G. Tao, E. Katz et al.,J. Electroanal. Chem. 1998, 454, 9–13.

61. A. Narvaez, E. Dominguez, I. Katakis et al.,J. Electroanal. Chem. 1997, 430, 217–233.

62. S. Marx-Tibbon, E. Katz, I. Willner, J. Am.Chem. Soc. 1995, 117, 9925,9926.

63. S. J. Sadeghi, G. Gilardi, G. Nicolosi et al.,Chem. Commun. 1997, 517,518.

64. A. D. Ryabov, Y. N. Firsova, V. N. Gozalet al., Chem. Eur. J. 1998, 4, 806–813.

65. G. Tao, E. Katz, I. Willner, Chem. Commun.1997, 2073,2074.

66. Y. Degani, A. Heller, J. Phys. Chem. 1987,91, 1285–1289.

67. Y. Degani, A. Heller, J. Am. Chem. Soc.1988, 110, 2615–2620.


68. W. Schuhmann, T. J. Ohara, H.-L. Schmidtet al., J. Am. Chem. Soc. 1991, 113,1394–1397.

69. P. N. Bartlett, R. G. Whitaker, M. J. Greenet al., J. Chem. Soc., Chem. Commun. 1987,1603,1604.

70. I. Willner, E. Katz, Angew. Chem., Int. Ed.Engl. 2000, 39, 1180–1218.

71. E. S. Ryabova, V. N. Goral, E. Csoregi et al.,Angew. Chem. Int. Ed. Engl. 1999, 38,804–807.

72. A. Badia, R. Carlini, A. Fernandez et al.,J. Am. Chem. Soc. 1993, 115, 7053–7060.

73. I. Willner, E. Katz, A. Riklin et al., J. Am.Chem. Soc. 1992, 114, 10 965,10 966.

74. I. Willner, N. Lapidot, A. Riklin et al., J. Am.Chem. Soc. 1994, 116, 1428–1441.

75. E. Katz, A. Riklin, I. Willner, J. Electroanal.Chem. 1993, 354, 129–144.

76. S. Kuwabata, T. Okamoto, Y. Kajiya et al.,Anal. Chem. 1995, 67, 1684–1690.

77. D. Avnir, S. Braun, O. Lev et al., Chem.Mater. 1994, 6, 1605–1614.

78. A. Walcarius, Electroanalysis 1998, 10,1217–1235.

79. K. Kalcher, J.-M. Kauffmann, J. Wang et al.,Electroanalysis 1995, 7, 5–22.

80. L. Gorton, Electroanalysis 1995, 7, 23–45.81. S. Cosnier, Electroanalysis 1997, 9, 894–902.82. C. Kranz, H. Wohlschlager, H.-L. Schmidt

et al., Electroanalysis 1998, 10, 546–552.83. S. Cosnier, Biosens. Bioelectron. 1999, 14,

443–456.84. H. Mao, P. G. Pickup, J. Electroanal. Chem.

1989, 265, 127–142.85. G. Lian, S. Dong, J. Electroanal. Chem. 1989,

260, 127–136.86. S. Dong, G. Lian, J. Electroanal. Chem. 1990,

291, 23–29.87. W. Schuhmann, Mikrochim. Acta 1995, 121,

1–29.88. S. Yabuki, H. Shinohara, M. Aizawa,

J. Electroanal. Chem. 1990, 277, 179–187.89. M. G. Loughram, J. M. Hall, A. P. F.

Turner, Electroanalysis 1996, 8, 870–875.90. S. Cosnier, C. Innocent, Y. Jouanneau, Anal.

Chem. 1994, 66, 3198–3201.91. S. Cosnier, L. Allien, L. Coche-Guerente

et al., Sens. Mater. 1996, 8, 169–177.92. S. Cosnier, B. Galland, C. Innocent,

J. Electroanal. Chem. 1997, 433, 113–119.93. N. Gajovic, K. Habermuller, A. Warsinke

et al., Electroanalysis 1999, 11, 1377–1383.

94. S. Cosnier, J.-L. Decout, M. Fontecave et al.,Electroanalysis 1998, 10, 521–525.

95. T. Parpaleix, J. M. Laval, M. Maida et al.,Anal. Chem. 1992, 64, 641–646.

96. I. Willner, E. Katz, N. Lapidot et al.,Bioelectrochem. Bioenerg. 1992, 29, 29–45.

97. G. Ramsay, S. M. Wolpert, Anal. Chem.1999, 71, 504–506.

98. A. Heller, J. Phys. Chem. 1992, 96,3579–3587.

99. H. Bu, S. R. Mikkelsen, A. M. English, Anal.Chem. 1995, 67, 4071–4076.

100. B. A. Gregg, A. Heller, Anal. Chem. 1990,62, 258–263.

101. B. A. Gregg, A. Heller, J. Phys. Chem. 1991,95, 5970–5975.

102. B. A. Gregg, A. Heller, J. Phys. Chem. 1991,95, 5976–5980.

103. I. Katakis, A. Heller, Anal. Chem. 1992, 64,1008–1013.

104. E. J. Calvo, C. Danilowicz, L. Diaz,J. Electroanal. Chem. 1994, 369, 279–282.

105. S. Koide, K. Yokoyama, J. Electroanal. Chem.1999, 468, 193–201.

106. G. Arai, M. Masuda, I. Yasumori, Bull.Chem. Soc. Jpn. 1994, 67, 2962–2966.

107. S. A. Emr, A. M. Yacynych, Electroanalysis1995, 7, 913–923.

108. P. Audebert, C. Demaille, C. Sanchez,Chem. Mater. 1993, 5, 911–913.

109. T.-M. Park, E. I. Iwuoha, M. R. Smith et al.,Anal. Commun. 1996, 33, 271–273.

110. T.-M. Park, E. I. Iwuoha, M. R. Smith et al.,Talanta 1997, 44, 973–978.

111. M. Tsionsky, G. Gun, V. Glezer et al., Anal.Chem. 1994, 66, 1747–1753.

112. I. Pankratov, O. Lev, J. Electroanal. Chem.1995, 393, 35–41.

113. S. Sampath, I. Pankratov, J. Gun et al.,J. Sol.-Gel Sci. Technol. 1996, 7, 123–128.

114. J. Gun, O. Lev, Anal. Chim. Acta 1996, 336,95–106.

115. J. Gun, O. Lev, Anal. Lett. 1996, 29,1933–1938.

116. S. Sampath, O. Lev, Electroanalysis 1996, 8,1112–1116.

117. S. Sampath, O. Lev, Anal. Chem. 1996, 68,2015–2021.

118. S. Sampath, O. Lev, J. Electroanal. Chem.1997, 426, 131–137.

119. S. Bharathi, O. Lev, Anal. Commun. 1998,35, 29–31.

120. B. Grundig, C. Krabisch, Anal. Chim. Acta1989, 222, 75–81.


121. H. Gunasingham, C.-H. Tan, T.-C. Aw,Anal. Chim. Acta 1990, 234, 321–330.

122. P. D. Hale, T. Inagaki, H. I. Karan et al.,J. Am. Chem. Soc. 1989, 111, 3482–3484.

123. L. Gorton, H. I. Karan, P. D. Hale et al.,Anal. Chim. Acta 1990, 228, 23–30.

124. J. Wang, L.-H. Wu, Z. L. Lu et al., Anal.Chim. Acta 1990, 228, 251–257.

125. P. D. Hale, L. I. Boguslavsky, T. Inagakiet al., Mol. Cryst. Liq. Cryst. 1990, 190,251–258.

126. P. D. Hale, L. I. Boguslavsky, T. Inagakiet al., Anal. Chem. 1991, 63, 677–682.

127. P. D. Hale, H. L. Lan, L. I. Boguslavskyet al., Anal. Chim. Acta 1991, 251, 121–128.

128. A. Amine, J.-M. Kauffmann, G. J. Patri-arche, Talanta 1991, 38, 107–110.

129. A. Amine, J.-M. Kauffmann, G. J. Patriarcheet al., Anal. Lett. 1991, 24, 1293–1315.

130. F. Mizutani, S. Yabuki, A. Okuda et al., Bull.Chem. Soc. Jpn. 1991, 64, 2849–2851.

131. S. Sakura, R. P. Buck, Bioelectrochem.Bioenerg. 1992, 28, 387–400.

132. P. D. Hale, H. S. Lee, Y. Okamoto, Anal.Lett. 1993, 26, 1–16.

133. A. Amine, J.-M. Kauffmann, G. G. Guilbaultet al., Anal. Lett. 1993, 26, 1281–1299.

134. I. Rosen-Margalit, J. Rishpon, Biosens.Bioelectron. 1993, 8, 315–323.

135. J. Wang, N. Naser, Electroanalysis 1994, 6,571–575.

136. T. Ikeda, Bull. Electrochem. 1992, 8,145–159.

137. T. Ikeda, I. Katasho, M. Senda, Anal. Sci.1985, 1, 455–457.

138. T. Ikeda, H. Hamada, K. Miki et al., Agric.Biol. Chem. 1985, 49, 541–543.

139. T. Ikeda, H. Hamada, M. Senda, Agric. Biol.Chem. 1986, 50, 883–890.

140. N. Motta, A. R. Guadalupe, Anal. Chem.1994, 66, 566–571.

141. T. Inagaki, H. S. Lee, P. D. Hale et al.,Macromolecules 1989, 22, 4641–4643.

142. H. I. Karan, P. D. Hale, H. L. Lan et al.,Polym. Adv. Technol. 1991, 2, 229–235.

143. T. Kaku, H. I. Karan, Y. Okamoto, Anal.Chem. 1994, 66, 1231–1235.

144. P. D. Hale, L. I. Boguslavsky, H. I. Karanet al., Anal. Chim. Acta 1991, 248, 155–161.

145. H. Gunasingham, C.-H. Tan, Analyst 1990,115, 35–39.

146. H. Gunasingham, C.-H. Tan, T. C. Aw, Clin.Chem. 1990, 36, 1657–1661.

147. U. Wollenberger, V. Bogdanovskaya, S. Bo-brin et al., Anal. Lett. 1990, 23, 1795–1808.

148. T. Tatsuma, T. Watanabe, J. Electroanal.Chem. 1993, 356, 245–253.

149. P. C. Pandey, A. M. Kayastha, V. Pandey,Appl. Biochem. Biotechnol. 1992, 33,139–144.

150. H. Gunashingham, C.-H. Tan, Anal. Chim.Acta 1990, 229, 83–91.

151. I. Rosen-Margalit, A. Bettelheim, J. Rish-pon, Anal. Chim. Acta 1993, 281, 327–333.

152. J. Kulys, H. E. Hansen, T. Buch-Rasmussenet al., Anal. Chim. Acta 1994, 288, 193–196.

153. J. Kulys, L. Wang, H. E. Hansen et al.,Electroanalysis 1995, 7, 92–94.

154. J. Wang, X. Zhang, M. Prakash, Anal. Chim.Acta 1999, 395, 11–16.

155. G. Bremle, B. Persson, L. Gorton, Electro-analysis 1991, 3, 77–86.


157. P. D. Hale, L. F. Liu, T. A. Skotheim,Electroanalysis 1991, 3, 751–756.

158. L. I. Boguslavsky, P. D. Hale, L. Geng et al.,Solid State Ionics 1993, 60, 189–197.

159. T. Ikeda, F. Matsushita, M. Senda, Biosens.Bioelectron. 1991, 6, 299–304.

160. M. Smolander, G. Marko-Varga, L. Gorton,Anal. Chim. Acta 1995, 303, 233–240.

161. A. Riklin, E. Katz, I. Willner et al., Nature1995, 376, 672–675.

162. E. Katz, A. Riklin, V. Heleg-Shabtai et al.,Anal. Chim. Acta 1999, 385, 45–58.

163. I. Willner, V. Heleg-Shabtai, R. Blonderet al., J. Am. Chem. Soc. 1996, 118, 10 321,10 322.

164. E. Katz, D. D. Schlereth, H.-L. Schmidtet al., J. Electroanal. Chem. 1994, 368,165–171.

165. H.-L. Schmidt, W. Schuhmann, Biosens.Bioelectron. 1996, 11, 127–135.

166. L.-H. Guo, G. McLendon, H. Razafitrimoet al., J. Mater. Chem. 1996, 6, 369–374.

167. A. L. Lehninger, Biochemistry, 2nd ed.,Worth, New York, 1975, p. 479.


169. L. Falat, H.-Y. Cheng, J. Electroanal. Chem.1983, 157, 393–397.

170. I. Katakis, E. Dominguez, Mikrochim. Acta1997, 126, 11–32.

171. L. Gorton, B. Persson, P. D. Hale et al.in Biosensors and Chemical Sensors, (Eds.:


P. G. Edelman, J. Wang), American Chem-ical Society, Washington, 1992, pp. 56–83,Chap. 6.



174. R. W. Coughlin, M. Aizawa, B. F. Alexanderet al., Biotechnol. Bioeng. 1975, 17, 515–526.

175. H. Jaegfeldt, A. Torstensson, G. Johansson,Anal. Chim. Acta 1978, 97, 221–228.


177. R. D. Braun, K. S. V. Santhanam, P. J. Elv-ing, J. Am. Chem. Soc. 1975, 97, 2591–2598.


179. J. Moiroux, P. J. Elving, J. Am. Chem. Soc.1980, 102, 6533–6538.

180. R. L. Blankespoor, L. L. Miller, J. Electroanal.Chem. 1984, 171, 231–241.

181. H. Jaegfeldt, T. Kuwana, G. Johansson,J. Am. Chem. Soc. 1983, 105, 1805–1814.

182. A. Silber, C. Brauchle, N. Hampp, J. Electro-anal. Chem. 1995, 390, 83–89.

183. G. X. Li, J. J. Zhu, H. Q. Fang et al.,J. Electrochem. Soc. 1996, 143, L141,L142.

184. N. Cenas, J. J. Kanapieniene, J. Kulys,J. Electroanal. Chem. 1985, 189, 163–169.

185. K. Ravichandran, R. P. Baldwin, Anal.Chem. 1984, 56, 1744–1747.

186. N. Cenas, J. Rozgaite, A. Pocius et al.,J. Electroanal. Chem. 1983, 154, 121–128.

187. W. B. Nowall, W. G. Kuhr, Anal. Chem.1995, 67, 3583–3588.

188. A. Kitani, L. L. Miller, J. Am. Chem. Soc.1981, 103, 3595–3597.

189. A. Kitani, Y.-H. So, L. L. Miller, J. Am.Chem. Soc. 1981, 103, 7636–7641.

190. B. W. Carlson, L. L. Miller, J. Am. Chem.Soc. 1983, 105, 7453,7454.

191. B. W. Carlson, L. L. Miller, P. Neta et al.,J. Am. Chem. Soc. 1984, 106, 7233–7239.

192. B. W. Carlson, L. L. Miller, J. Am. Chem.Soc. 1985, 107, 479–485.

193. L. L. Miller, J. R. Valentine, J. Am. Chem.Soc. 1988, 110, 3982–3989.

194. N. Cenas, J. J. Kanapieniene, J. Kulys,Biochim. Biophys. Acta 1984, 767, 108–112.

195. D. C.-S. Tse, T. Kuwana, Anal. Chem. 1978,50, 1315–1318.

196. H. Jaegfeldt, A. Torstensson, L. Gortonet al., Anal. Chem. 1981, 53, 1979–1982.

197. C. Degrand, L. L. Miller, J. Am. Chem. Soc.1980, 102, 5728–5732.

198. M. Fukui, A. Kitani, C. Degrand et al., J. Am.Chem. Soc. 1982, 104, 28–33.

199. E. Lorenzo, L. Sanchez, F. Pariente et al.,Anal. Chim. Acta 1995, 309, 79–88.

200. C. Ueda, D. C.-S. Tse, T. Kuwana, Anal.Chem. 1982, 54, 850–856.

201. H. Huck, H.-L. Schmidt, Angew. Chem., Int.Ed. Engl. 1981, 93, 421,422.

202. L. Gorton, A. Torstensson, H. Jaegfeldtet al., J. Electroanal. Chem. 1984, 161,103–111.

203. A. Torstensson, L. Gorton, J. Electroanal.Chem. 1981, 130, 199–207.

204. H. Huck, Fresenius J. Anal. Chem. 1982, 313,548–552.

205. L. Gorton, J. Chem. Soc., Faraday Trans. 11986, 82, 1245–1258.

206. D. D. Schlereth, E. Katz, H.-L. Schmidt,Electroanalysis 1994, 6, 725–734.

207. D. D. Schlereth, E. Katz, H.-L. Schmidt,Electroanalysis 1995, 7, 46–54.

208. L. Gorton, G. Johansson, A. Torstensson,J. Electroanal. Chem. 1985, 196, 81–92.

209. B. Persson, L. Gorton, J. Electroanal. Chem.1990, 292, 115–138.

210. B. Persson, J. Electroanal. Chem. 1990, 287,61–80.

211. T. Vering, W. Schuhmann, D. Seiwald et al.,J. Electroanal. Chem. 1994, 364, 277–279.

212. F. Xu, H. Li, S. J. Cross et al., J. Electroanal.Chem. 1994, 368, 221–225.

213. W. J. Albery, P. N. Bartlett, J. Chem. Soc.,Chem. Commun. 1984, 234–236.

214. J. Kulys, Enzyme Microb. Technol. 1981, 3,344–349.

215. E. Katz, T. Lotzbeyer, D. D. Schlereth et al.,J. Electroanal. Chem. 1994, 373, 189–200.

216. I. Willner, A. Riklin, Anal. Chem. 1994, 66,1535–1539.

217. Y. Ogino, K. Takagi, K. Kano et al.,J. Electroanal. Chem. 1995, 396, 517–524.

218. G. T. R. Palmore, H. Bertschy, S. H. Ber-gens et al., J. Electroanal. Chem. 1998, 443,155–161.

219. J. M. Obon, P. Casanova, A. Manjon et al.,Biotechnol. Prog. 1997, 13, 557–561.

220. M. Hedenmo, A. Narvaez, E. Dominguezet al., Analyst 1996, 121, 1891–1895.

221. M. J. Lobo, A. J. Miranda, P. Tunon, Electro-analysis 1997, 9, 191–202.

222. A. E. Biade, C. Bourdillon, J. M. Laval et al.,J. Am. Chem. Soc. 1992, 114, 893–897.


223. J. N. Burnett, A. L. Underwood, Biochem-istry 1975, 5, 2060–2066.

224. P. J. Elving, W. T. Bresnahan, J. Moirouxet al., Bioelectrochem. Bioenerg. 1982, 9,365–378.

225. C. O. Schmakel, K. S. V. Santhanam, P. J.Elving, J. Am. Chem. Soc. 1975, 97,5083–5092.

226. K. S. V. Santhanam, P. J. Elving, J. Am.Chem. Soc. 1973, 95, 5482–5490.

227. A. J. Cunningham, A. L. Underwood, Bio-chemistry 1967, 6, 266–271.

228. S. H. Baik, C. Kang, I. C. Jeon et al.,Biotechnol. Tech. 1999, 13, 1–5.

229. Y. T. Long, H. Y. Chen, J. Electroanal. Chem.1997, 440, 239–242.

230. G. X. Li, H. Y. Chen, D. X. Zhu, Chem. J.Chin. Univ. 1996, 17, 553,554.

231. R. Ruppert, S. Herrmann, E. Steckhan,Tetrahedron Lett. 1987, 28, 6583–6586.

232. R. Wienkamp, E. Steckhan, Angew. Chem.,Int. Ed. Engl. 1982, 21, 782,783.

233. R. Ruppert, S. Herrmann, E. Steck-han, J. Chem. Soc., Chem. Commun. 1988,1150–1151.

234. R. Wienkamp, E. Steckhan, Angew. Chem.,Int. Ed. Engl. 1983, 22, 497.

235. I. Willner, R. Maidan, M. Shapira, J. Chem.Soc., Perkin Trans. 2 1990, 559–564.

236. M. Franke, E. Steckhan, Angew. Chem., Int.Ed. Engl. 1988, 27, 265–267.

237. U. Koelle, A. D. Ryabov, Mendeleev Com-mun. 1995, 187–189.

238. E. Steckhan, S. Herrmann, R. Ruppert et al.,Organometallics 1991, 10, 1568–1577.

239. H. C. Lo, O. Buriez, J. B. Kerr et al., Angew.Chem., Int. Ed. Engl. 1999, 38, 1429–1432.

240. E. Hofer, E. Steckhan, B. Ramos et al.,J. Electroanal. Chem. 1996, 402, 115–122.



243. S. Chardonnoblat, S. Cosnier, A. Deronzieret al., J. Electroanal. Chem. 1993, 352,213–228.

244. H. X. Ju, W. Ma, C. X. Cai et al., Chem. J.Chin. Univ. 1995, 16, 1680–1684.

245. R. J. Day, S. J. Kinsey, E. T. Seo et al., Trans.N.Y. Acad. Sci. 1972, 34, 588–594.

246. M. Ito, T. Kuwana, J. Electroanal. Chem.1971, 32, 415–425.

247. K. Kano, K. Takagi, Y. Ogino et al., Chem.Lett. 1995, 589,590.

248. A. L. de Lacey, M. T. Bes, C. Gomez-Morenoet al., J. Electroanal. Chem. 1995, 390, 69–76.

249. S. Peguin, P. Soucaille, Biotechnol. Bioeng.1996, 51, 342–348.

250. R. DiCosimo, C.-H. Wong, L. Daniels et al.,J. Org. Chem. 1981, 46, 4622,4623.

251. Z. Shaked, J. J. Barber, G. M. Whitesides,J. Org. Chem. 1981, 46, 4100,4101.

252. S. Chao, M. S. Wrighton, J. Am. Chem. Soc.1987, 109, 5886–5888.

253. T. Matsue, H.-C. Chang, I. Uchida et al.,Tetrahedron Lett. 1988, 29, 1551–1554.

254. H.-C. Chang, T. Matsue, I. Uchida et al.,Chem. Lett. 1989, 1119–1122.

255. A. Bergel, M. Comtat, J. Electroanal. Chem.1991, 302, 219–231.

256. A. Bergel, M. Comtat, Bioelectrochem. Bioen-erg. 1992, 27, 495–500.

257. H. Simon, J. Bader, H. Gunter et al., Angew.Chem., Int. Ed. Engl. 1985, 24, 539–553.

258. H. Simon, H. Gunter, J. Bader et al., Angew.Chem. Int. Ed. Engl. 1981, 20, 861–863.

259. K. Takagi, K. Kano, T. Ikeda, Chem. Lett.1996, 11,12.

260. K. Takagi, K. Kano, T. Ikeda, J. Electroanal.Chem. 1998, 445, 211–219.

261. S. Kim, S. E. Yun, C. Kang, J. Electroanal.Chem. 1999, 465, 153–159.

262. Y. Kashiwagi, Y. Yanagisawa, N. Shibayamaet al., Chem. Lett. 1996, 1093,1094.

263. R. Yuan, S. Kuwabata, H. Yoneyama, Chem.Lett. 1996, 137,138.

264. J. Contet, A. Bergel, M. Comtat, Bioelec-trochem. Bioenerg. 1992, 27, 475–486.

265. K. Delecouls, P. Saint-Aguet, C. Zaboroschet al., J. Electroanal. Chem. 1999, 468,139–149.

266. K. Nishiyama, H. Ishida, I. Taniguchi,J. Electroanal. Chem. 1994, 373, 255–258.

267. P. Gros, C. Zaborosch, H. G. Schlegel et al.,J. Electroanal. Chem. 1996, 405, 189–195.

268. J. Cantet, A. Bergel, M. Comtat et al., J. Mol.Catal. 1992, 73, 371–380.

269. K. W. Willman, R. W. Murray, J. Electroanal.Chem. 1982, 133, 211–231.

270. C. Van Dijk, T. Van Eijs, J. W. Van Leeuwenet al., FEBS Lett. 1984, 166, 76–80.

271. Y. S. Obeng, A. Founta, A. J. Bard, New J.Chem. 1992, 16, 121–129.

272. E. Katz, N. Itzhak, I. Willner, Langmuir1993, 9, 1392–1396.

273. S. Cosnier, K. LeLous, Talanta 1996, 43,331–337.


274. J. Cantet, A. Bergel, M. Comtat, EnzymeMicrob. Technol. 1996, 18, 72–79.

275. N. Takano, K. Matuda, M. Itaya, DenkiKagaku 1998, 66, 86–91.

276. Y. Kashiwagi, Y. Yanagisawa, N. Shibayamaet al., Electrochim. Acta 1997, 42, 2267–2270.

277. M. Maurice, J. Souppe, New J. Chem. 1990,14, 301–304.

278. A. F. Buckmann, V. Wray, Biotechnol.Biochem. 1992, 15, 303–310.

279. J. Hendle, A. F. Buckmann, W. Aehle et al.,Eur. J. Biochem. 1993, 213, 947–956.

280. H.-L. Schmidt, G. Grenner, Eur. J. Biochem.1976, 67, 295–302.

281. P. Zappelli, A. Rossodivita, L. Re, Eur. J.Biochem. 1975, 54, 475–482.

282. P. Zappelli, A. Rossodivita, G. Prosperiet al., Eur. J. Biochem. 1976, 62, 211–215.

283. A. F. Buckmann, Biocatalysis 1987, 1,173–186.

284. B. Leca, J.-L. Marty, Biosens. Bioelectron.1997, 12, 1083–1088.

285. B. Leca, J.-L. Marty, Anal. Chim. Acta 1997,340, 143–148.

286. M. Montagne, J.-L. Marty, Anal. Chim. Acta1995, 315, 297–302.

287. P. Zappelli, R. Pappa, A. Rossodivita et al.,Eur. J. Biochem. 1977, 72, 309–315.

288. T. Eguchi, T. Iizuka, T. Kagotani et al., Eur.J. Biochem. 1986, 155, 415–421.

289. S. Furukawa, N. Katayama, T. Iizuka et al.,FEBS Lett. 1980, 121, 239–242.

290. A. Nakamura, I. Urabe, H. Okada, J. Biol.Chem. 1986, 261, 16 792–16 794.

291. S. A. M. Vanhommerig, L. A. A. E. Sluyter-man, E. M. Meijer, Biochim. Biophys. Acta1996, 1295, 125–138.

292. Y. Yamazaki, H. Maeda, Agric. Biol. Chem.1981, 45, 2277–2288.

293. Y. Nakamura, S. Suye, J. Kira et al., Biochim.Biophys. Acta 1996, 1289, 221–225.

294. U. Kragl, W. Kruse, W. Hummel et al.,Biotechnol. Bioeng. 1996, 52, 309–319.

295. M.-O. Mansson, P.-O. Larsson, K. Mosbach,Eur. J. Biochem. 1978, 86, 455–463.

296. M.-O. Mansson, P.-O. Larsson, K. Mosbach,FEBS Lett. 1979, 98, 309–313.

297. C. Woenckhaus, R. Koob, A. Burkhard et al.,Bioorg. Chem. 1983, 12, 45–57.

298. J. Kovar, K. Simek, I. Kucera et al., Eur. J.Biochem. 1984, 139, 585–591.

299. P. Goulas, Eur. J. Biochem. 1987, 168,469–473.

300. M. Persson, M.-O. Mansson, L. Bulow et al.,Biotechnology 1991, 9, 280–284.

301. A. B. Kharitonov, L. Alfonta, E. Katz et al.,J. Electroanal. Chem. 2000, 487. 133–141.

302. A. Bardea, E. Katz, A. F. Buckmann et al.,J. Am. Chem. Soc. 1997, 119, 9114–9119.

303. R. J. P. Williams, Electron Transfer in Biologyand the Solid State, American ChemicalSociety, Washington, D. C., 1990, pp. 1–23.

304. M. Brunori, Biosens. Bioelectron. 1994, 9,633–636.

305. F. A. Armstrong, H. A. O. Hill, N. J. Wal-ton, Q. Rev. Biophys. 1986, 18, 261–322.

306. F. A. Armstrong, H. A. O. Hill, N. J. Wal-ton, Acc. Chem. Res. 1988, 21, 407–413.

307. J. E. Frew, H. A. O. Hill, Eur. J. Biochem.1988, 172, 261–269.

308. I. Taniguchi, K. Toyosawa, H. Yamaguchiet al., J. Chem. Soc., Chem. Commun. 1982,1032,1033.

309. I. Taniguchi, S. Yoshimoto, K. Nishiyama,Chem. Lett. 1997, 353,354.

310. B. D. Lamp, D. Hobara, M. D. Porter et al.,Langmuir 1997, 13, 736–741.

311. H. A. O. Hill, D. J. Page, N. J. Walton,J. Electroanal. Chem. 1987, 217, 141–158.

312. Y. Xie, S. Dong, Bioelectrochem. Bioenerg.1992, 29, 71–79.

313. P. M. Allen, H. A. O. Hill, N. J. Walton,J. Electroanal. Chem. 1984, 178, 69–86.

314. Z.-X. Huang, M. Feng, Y.-H. Wang et al.,J. Electroanal. Chem. 1996, 416, 31–40.

315. R. Santucci, A. Faraoni, L. Campanella et al.,Biochem. J. 1991, 273, 783–786.

316. S. Bagby, P. D. Barker, K. DiGleria et al.,Biochem. Soc. Trans. 1988, 16, 958,959.

317. G. Li, H. Chen, D. Zhu, Anal. Chim. Acta1996, 319, 275, 276.

318. X. Qu, T. Lu, S. Dong et al., Bioelectrochem.Bioenerg. 1994, 34, 153–156.

319. T. Lu, X. Yu, S. Dong et al., J. Electroanal.Chem. 1994, 369, 79–86.

320. W. Jin, U. Wollenberger, F. F. Bier et al.,Bioelectrochem. Bioenerg. 1996, 39, 221–225.

321. D. A. Powis, G. D. Wattus, FEBS Lett. 1981,126, 282–284.

322. H. A. O. Hill, N. J. Walton, J. Am. Chem.Soc. 1982, 104, 6515–6519.

323. A. E. G. Cass, G. Davis, H. A. O. Hill et al.,Biochim. Biophys. Acta 1985, 828, 51–57.

324. W. Jin, U. Wollenberger, E. Kargel et al.,J. Electroanal. Chem. 1997, 433, 135–139.

325. M. J. Tarlov, E. F. Bowden, J. Am. Chem.Soc. 1991, 113, 1847–1849.


326. D. Hobara, K. Niki, C. Zhou et al., ColloidsSurf., A 1994, 93, 241–250.

327. T. M. Nahir, E. F. Bowden, J. Electroanal.Chem. 1996, 410, 9–13.

328. S. Song, R. A. Clark, E. F. Bowden et al.,J. Phys. Chem. 1993, 97, 6564–6572.

329. P. N. Bartlett, D. J. Caruana, Analyst 1992,117, 1287–1292.

330. M. Collinson, E. F. Bowden, M. J. Tarlov,Langmuir 1992, 8, 1247–1250.

331. B. A. Kuznetsov, N. A. Byzova, G. P. Shu-makovich, J. Electroanal. Chem. 1994, 371,85–92.

332. J. M. Cooper, K. R. Greenough, C. J.McNeil, J. Electroanal. Chem. 1993, 347,267–275.

333. P. A. Adams in Peroxidases in Chemistry andBiology (Eds.: J. Everse, K. E. Everse), CRCPress, Boston, 1991, pp. 171–200, Vol. 2,Chap. 7.

334. G. Ranghino, G. Antonini, P. Fantucci,Israel J. Chem. 1994, 34, 239–244.

335. T. Lotzbeyer, W. Schuhmann, E. Katz et al.,J. Electroanal. Chem. 1994, 377, 291–294.

336. H. A. O. Hill, N. J. Walton, I. J. Higgins,FEBS Lett. 1981, 126, 282–284.

337. F. Patolsky, E. Katz, V. Heleg-Shabtai et al.,Chem. Eur. J. 1998, 4, 1068–1073.

338. V. Heleg-Shabtai, E. Katz, I. Willner, J. Am.Chem. Soc. 1997, 119, 8121,8122.

339. V. Heleg-Shabtai, E. Katz, S. Levi et al.,J. Chem. Soc., Perkin Trans. 2, 1997,2645–2651.

340. A. E. Strong, B. D. Moore, J. Mater. Chem.1999, 9, 1097–1105.

341. D. L. Pilloud, F. Rabanal, B. R. Gibney et al.,J. Phys. Chem. B 1998, 102, 1926–1937.

342. E. Katz, V. Heleg-Shabtai, I. Willner et al.,Angew. Chim. 1998, 37, 3253–3256.

343. I. Willner, V. Heleg-Shabtai, E. Katz et al.,J. Am. Chem. Soc. 1999, 121, 6455–6468.

344. V. Pardo–Yissar, E. Katz, I. Willner et al.,Faraday Discuss. 2000, 116, 119–134.

345. E. Katz, I. Willner, A. B. Kotlyar, J. Elec-troanal. Chem. 1999, 479, 64–68.

346. J. Wang, J. Pharm. Biomed. Anal. 1999, 19,47–53.

347. M. E. Collison, M. E. Meyerhoff, Anal.Chem. 1990, 62, 425A.

348. S. A. Jaffari, A. P. F. Turner, Physiol. Meas.1995, 16, 1–15.

349. M. Kyrolainen, H. Hakanson, R. Ekrothet al., Anal. Chim. Acta 1993, 279, 149–153.

350. C. Meyerhoff, F. Bischof, F. J. Mennel, et al.,Biosens. Bioelectron. 1993, 8, 409–414.

351. J. Wang, M. Ozsoz, Electroanalysis 1990, 2,647–650.

352. V. Kacanicklic, K. Johansson, G. MarkoVarga et al., Electroanalysis, 1994, 6,381–390.

353. H. Bramwell, A. E. G. Cass, P. N. B. Gibbset al., Analyst 1990, 115, 185–188.

354. J. Wang, E. Dempsey, M. Ozsoz et al.,Analyst 1991, 116, 997–999.

355. A. Maines, D. Ashworth, P. Vadgama, FoodTechnol. Biotechnol. 1996, 34, 31–42.

356. A. S. Bassi, D. Q. Tang, E. Lee et al., FoodTechnol. Biotechnol. 1996, 34, 9–22.

357. C. Nistor, J. Emneus, Waste Manage. 1999,19, 147–170.

358. J. L. Marty, B. Leca, T. Noguer, Analysis1998, 26, M144–M149.

359. M. P. Marco, D. Barcelo, Measurement Sci.Technol. 1996, 7, 1547–1562.

360. D. Ivnitski, I. Abdel-Hamid, P. Atanasovet al., Biosens. Bioelectron. 1999, 14,599–624.

361. B. M. Paddle, Biosens. Bioelectron. 1996, 11,1079–1113.

362. C. Bourdillon, J. M. Laval, D. Thomas,J. Electrochem. Soc. 1986, 133, 706–711.

363. C. Bourdillon, R. Lortie, J. M. Laval, Biotech-nol. Bioeng. 1988, 31, 553–558.

364. A. Fassouane, J. M. Laval, J. Moiroux et al.,Biotechnol. Bioeng. 1990, 35, 935–939.

365. Y. Yamazaki, H. Maeda, Agric. Biol. Chem.1982, 46, 1571–1581.

366. J. R. Wykes, P. Dunnill, M. D. Lilly, Biotech-nol. Bioeng. 1975, 17, 51–68.

367. T. Erabi, H. Hiura, M. Tanaka, Bull. Chem.Soc. Jpn. 1975, 48, 1354–1356.

368. O. S. Ksenzhek, S. A. Petrova, M. V. Kolo-dyazhny, Biolectrochem. Bioenerg. 1982, 9,167–174.

369. R. S. Schrebler, A. Arratia, S. Sanchez et al.,Bioelectrochem. Bioenerg. 1990, 23, 81–91.

370. K. Takehara, H. Takemura, Y. Ide et al.,J. Electroanal. Chem. 1991, 308, 345–350.

371. E. Katz, Studia Biophysica 1988, 125,211–217.

372. K. T. Kinnear, H. G. Monbouquette, Biotech-nol. Bioeng. 1993, 42, 140–144.

373. S. Kuwabata, R. Tsuda, K. Nishida et al.,Chem. Lett. 1993, 1631–1634.

374. S. Kuwabata, R. Tsuda, H. Yoneyama,J. Am. Chem. Soc. 1994, 116, 5437–5443.


375. K. Takayama, K. Kano, T. Ikeda, Chem. Lett.1996, 1009,1010.

376. G. Prentice, CHEMTECH 1984, 14,684–701.

377. G. Tayhas, R. Palmore, G. M. Whitesides,Enzymatic Conversion of Biomass for Fu-els Production, ACS Symposium SeriesNo. 566, 1994, pp. 271–290, Chap. 14.

378. K. Kordesch, Ber. Bunsen-Ges. Phys. Chem.1990, 94, 902–907.

379. C. Van Dijk, C. Laane, C. Veeger, Recl. Trav.Chim., Pays-Bas 1985, 104, 245–249.

380. I. Willner, E. Katz, F. Patolsky et al., J. Chem.Soc, Perkin Trans. 2 1998, 1817–1822.

381. I. Willner, Acc. Chem. Res. 1997, 30,347–356.

382. M. Lion-Dagan, E. Katz, I. Willner, J. Am.Chem. Soc. 1994, 116, 7913,7914.

383. I. Willner, R. Blonder, E. Katz et al., J. AmChem. Soc. 1996, 118, 5310,5311.

384. I. Willner, E. Katz, B. Willner et al., Biosens.Bioelectron. 1997, 12, 337–356.

385. R. Blonder, E. Katz, I. Willner et al., J. Am.Chem. Soc. 1997, 119, 11 747–11 757.

386. I. Willner, M. Lion-Dagan, S. Marx-Tibbon et al., J. Am. Chem. Soc. 1995, 117,6581–6592.

387. I. Willner, A. Doron, E. Katz et al., Langmuir1996, 12, 946–954.

388. M. Lion-Dagan, E. Katz, I. Willner, J. Chem.Soc., Chem. Commun. 1994, 2741,2742.

389. M. Lion–Dagan, S. Marx-Tibbon, E. Katzet al., Angew. Chem., Int. Ed. Engl. 1995, 34,1604–1606.

390. E. Katz, B. Willner, I. Willner, Biosens.Bioelectron. 1997, 12, 703–719.

391. E. Katz, M. Lion-Dagan, I. Willner,J. Electroanal. Chem. 1995, 382, 25–31.

392. E. Katz, I. Willner, Electroanalysis 1995, 7,417–419.

393. E. Katz, M. Lion-Dagan, I. Willner, J. Elec-troanal. Chem. 1996, 408, 107–112.

394. R. Blonder, E. Katz, Y. Cohen et al., Anal.Chem. 1996, 68, 3151–3157.

395. F. Patolsky, B. Filanovsky, E. Katz et al.,J. Phys. Chem. B 1998, 102, 10 359–10 367.

396. F. Patolsky, E. Katz, A. Bardea et al.,Langmuir 1999, 15, 3703–3706.

0 The Origin of Bioelectrochemistry: An Overview

Documents

Transcript of 0 The Origin of Bioelectrochemistry: An Overview