nII

,

Lisica,

idad

a r t i c l e i n f o

Article history:Received 1 July 2010Received in revised form 17 November 2010Accepted 19 November 2010

Journal of Inorganic Biochemistry 105 (2011) 410419

Contents lists available at ScienceDirect

Journal of Inorgan

l se1. Introduction

The therapy for metal overload pathologies usually involves theadministration of suitable chelators to selectively remove the metal fromthe body. Regarding copper(II) and nickel(II) metal ions, there is still aneed for safe and efcient chelating agents, as the existing ones havea number of drawbacks such as toxic side effects and controversialefciency [1].

Copper as an essential element is a component of many metallopro-teins and enzymes and plays a vital role in electron transfer reactionsof many cellular processes. However, excessive copper can be verytoxic resulting in severe diseases [2]. Certain chelating agents have been

treatment with D-penicillamine, trien [N,N-bis(2-aminoethyl)ethane-1,2-diamine] and ammonium tetrathiomolybdate are consideredsafer alternatives. Trien is a lesser active agent for copper(II) removalin biological media than D-penicillamine, and although both chelatorshave similar toxicity, side effects are less frequent and generallymilder with D-penicillamine. Ammonium tetrathiomolybdate, actingdifferently from both D-penicillamine and trien, has been used due toits lower toxic prole, but it is still an experimental drug and its long-term efcacy is unknown [4].

Copper(II) chelation therapy attracts also attention in recentinvestigations and treatment of neurodegenerative disorders, such asAlzheimer, Parkinson, and CreutzfeldtJakob [5]. Furthermore, anshown to bind copper with high afnity. Prevchelation agents has focused on Wilson's diseametabolic disease of copper toxicity that is fD-Penicillamine has been one of the most coagents for treatment of this disease. When th

Corresponding author. Fax: +351 217 946 470.E-mail address: jcosta@ff.ul.pt (J. Costa).

0162-0134/$ see front matter 2010 Elsevier Inc. Aldoi:10.1016/j.jinorgbio.2010.11.014distorted square pyramidal environment. 2010 Elsevier Inc. All rights reserved.Available online 13 December 2010

Keywords:Macrocyclic compoundsStability constantsChelation therapyCopper(II) complexNickel(II) complexa b s t r a c t

Two pentaaza macrocycles containing pyridine in the backbone, namely 3,6,9,12,18-pentaazabicyclo[12.3.1]octadeca-1(18),14,16-triene ([15]pyN5), and 3,6,10,13,19-pentaazabicyclo[13.3.1]nonadeca-1(19),15,17-triene ([16]pyN5), were synthesized in good yields. The acidbase behaviour of these compounds wasstudied by potentiometry at 298.2 K in aqueous solution and ionic strength 0.10 M in KNO3. The protonationsequence of [15]pyN5 was investigated by 1H NMR titration that also allowed the determination ofprotonation constants in D2O. Binding studies of the two ligands with Ca2+, Ni2+, Cu2+, Zn2+, Cd2+, and Pb2+

metal ions were performed under the same experimental conditions. The results showed that all thecomplexes formed with the 15-membered ligand, particularly those of Cu2+ and especially Ni2+, arethermodynamically more stable than with the larger macrocycle. Cyclic voltammetric data showed that thecopper(II) complexes of the two macrocycles exhibited analogous behaviour, with a single quasi-reversibleone-electron transfer reduction process assigned to the Cu(II)/Cu(I) couple. The UVvisible-near IRspectroscopic and magnetic moment data of the nickel(II) complexes in solution indicated a tetragonaldistorted coordination geometry for the metal centre. X-band EPR spectra of the copper(II) complexes areconsistent with distorted square pyramidal geometries. The crystal structure of [Cu([15]pyN5)]

2+ determinedby X-ray diffraction showed the copper(II) centre coordinated to all ve macrocyclic nitrogen donors in aious work on copper(II)se, which is an inheritedatal if left untreated [3].mmonly used chelatinge patient cannot tolerate

excess of copperMoreover, high lincluding prostatherapeutic valuin the treatmentmixtures of coppefcient proteascancer cells [7].

l rights reserved.Two macrocyclic pentaaza compounds cochelating agents in copper(II) and nickel(

Ana S. Fernandes a, M. Ftima Cabral a, Judite Costa a,

Michael G.B. Drew d, Vitor Flix e

a iMed.UL, Faculdade de Farmcia, Universidade de Lisboa, Av. Prof. Gama Pinto, 1649-003b Instituto de Tecnologia Qumica e Biolgica, Universidade Nova de Lisboa, Av. da Repblc Instituto Superior Tcnico, Av. Rovisco Pais, 1049-001 Lisboa, Portugald School of Chemistry, University of Reading, Whiteknights, Reading, RG6 6AD UKe Departamento de Qumica, CICECO, and Seco Autnoma de Cincias da Sade, Univers

j ourna l homepage: www.etaining pyridine evaluated as novel) overload

Matilde Castro a, Rita Delgado b,c,

boa, Portugal2780-157 Oeiras, Portugal

e de Aveiro, 3810-193 Aveiro, Portugal

ic Biochemistry

v ie r.com/ locate / j inorgb ioappears to be an essential co-factor for angiogenesis.evels of copper were found in many human cancers,te, breast, colon, lung, and brain. Consequently, thee of copper(II) chelators as anti-angiogenic moleculesof these cancers has been reported [6]. More recently,er(II) chelators and copper salts were found to act asome inhibitors and apoptosis inducers, specically in

On the other hand, human exposure to nickel occurs primarily viainhalation and ingestion, through occupational exposure anddiet, leadingto adverse effects on human health. Nickel allergy in the form of contactdermatitis is the most common and well-known reaction. Although theaccumulation of nickel in the body through chronic exposure can causelung brosis, kidney, and cardiovascular diseases, the most seriousconcerns relate to nickel's carcinogenic activity. Epidemiological studieshave clearly implicated nickel compounds as human carcinogens [8,9].All nickel compounds, except for metallic nickel, were classied ascarcinogenic to humans in 1990 by the International Agency for Research

used as supplied without further purication. Organic solvents werepuried or dried by standard methods [15].

Caution: Althoughnoproblemswere found in thiswork, perchloratesin the presence of organic matter are potentially explosive and should beprepared in small quantities.

2.2. Synthesis of the macrocycles

2.2.1. Synthesis of the macrocycle [15]pyN5

411A.S. Fernandes et al. / Journal of Inorganic Biochemistry 105 (2011) 410419on Cancer (IARC) [10].Over the years, various chelators have been investigated for their

ability to bind nickel. The most effective ones are EDTA, DTPA(diethylenetriaminepentaacetic acid), diethyldithiocarbamate, tet-raethylthiuram disulde, and clioquinol (5-chloro-8-hydroxy-7-iodoquinoline), all of them presenting considerable side effects[8,11,12].

Therefore, the development of novel chelators selective for nickel(II)and for copper(II), and exhibiting minor side effects is an imperativeresearch. This led us to investigate the possible use of macrocycliccompounds. In fact,metal chelates ofmacrocycles often showpropertiesthat are particularly different from those of analogous open chainchelators. Macrocycles having more rigid structures can impose speciccoordination geometry to the metal ion, whereas open chain chelatorsadapt more easily to the geometric requirements of the metal centre[13]. In the present work, the synthesis and characterization of twopentaaza macrocyclic compounds containing pyridine in the backbone,[15]pyN5 (3,6,9,12,18-pentaazabicyclo[12.3.1]octadeca-1(18),14,16-tri-ene) and [16]pyN5 (3,6,10,13,19-pentaazabicyclo[13.3.1]nonadeca-1(19),15,17-triene), cf. Scheme 1, as well as the study of their copper(II)and nickel(II) complexes are reported, in order to evaluate their possibleuseas chelatingagents. Toaccomplish this aim, theacidbasebehaviourofthese two macrocycles was studied and their ability to coordinate Cu2+

and Ni2+ and other divalent metal ions (Ca2+ and Zn2+ are included dueto their essential role in living organisms) was evaluated. The adoptedstructures of the Cu(II) and Ni(II) complexes were also studied byspectroscopicmethods in solution, and the single crystal X-ray diffractionof [Cu[15]pyN5](PF6)2 was determined. Finally, due to the important roleof the redox behaviour of the copper(II) complexes in biology somevoltammetric studies were carried out.

2. Experimental section

2.1. General procedures

Elemental analysis was performed on a VarioEL CHNS analyserfrom vacuum-dried powder samples. Melting points were determinedwith a Kpffer Melting Point apparatus.

2.1.1. Reagents2,6-Pyridinedimethanol, N,N-bis(2-aminoethyl)ethane-1,2-diamine

and N,N-bis(2-aminoethyl)1,3-propanediamine were purchased fromAldrich. 2,6-Pyridinedicarbaldehyde was prepared by publishedmethods[14]. All the commercially available chemicals were of reagent grade andSchemeTo a stirred solution of freshly prepared 2,6-pyridinedicarbalde-hyde (2.34 g, 18 mmol) in methanol (40 mL) was added a solution ofPb(NO3)2 (6.1 g, 18 mmol) in water (80 mL). To the resulting solutionwas added dropwise, with rapid stirring, a solution of N,N-bis(2-aminoethyl)ethane-1,2-diamine (3.35 g, 18 mmol) in methanol(40 mL) over a period of 3 h. The solution was stirred while heatingunder reux for 7 h, during which time an intense deep red colourdeveloped. After reux the solution was cooled to 5 C, and sodiumborohydride (4.36 g, 45.5 mmol) was added in small portions over60 min. The yellow solution obtainedwas stirred at room temperaturefor 30 min and then heated on a hot water-bath at 60 C for 30 min,before being left overnight at room temperature. Leadwas removed bytreating the mixture with Na2S.9H2O (10 g, 42 mmol) followed byheating on a hot water-bath for 30 min. The solution was then cooled,and lead(II) sulphidewas removed byltration through a bed of Celite.The ltrate was extracted with dichloromethane (450 mL), thecombined extracts were dried with anhydrous MgSO4, and the dichlor-omethane was removed with a rotary evaporator to leave a light yellowoil. This oil was dissolved in methanol and 37% hydrochloric acid wasadded until pH2. During the addition, an off-white solid precipitated,whichwas identied as the pure desired compound. Yield: 85%.Mp 280282 C (decomp.). 1H NMR (D2O, pD=5.10): 3.21 (4H, s (singlet), NCH2), 3.41 (4H, t (triplet), NCH2CH2N), 3.51 (4H, t, NCH2CH2N),4.57 (4H, s, NCH2py), 7.53 (2H, d (doublet), py) and 7.99 (1H, t, py)ppm. 13C NMR (D2O, pD = 5.10): 43.92 (NCH2CH2N), 45.64 (NCH2CH2N), 46.33 (NCH2), 49.92 (NCH2py), 124.07 (py), 140.33 (py)and 150.94 (py) ppm. Found: C, 37.03; H, 7.16; N, 16.56. Calc. forC13H23N54HCl1.5H2O: C, 36.98, H, 7.16, N, 16.59%.

2.2.2. Synthesis of the macrocycle [16]pyN5A procedure analogous to that described for [15]pyN5 was used,

replacing N,N-bis(2-aminoethyl)ethane-1,2-diamine by N,N-bis(2-aminoethyl)1,3-propanediamine. The product was obtained as a thickyellow oil, which was puried by passing through a neutral aluminacolumn (2.530 cm) and eluting with dichloromethanemethanol(10:0.5 v/v). The pure compound was dissolved in methanol and 37%hydrochloric acidwas added until pH2. An off-white salt precipitated.Yield: 46%. Mp 266268 C (decomp.). 1H NMR (D2O, pD=2.55): 2.21(2 H, q (quintuplet), CH2CH2N), 3.37 (4H, t, CH2CH2N), 3.65 (4H, t,NCH2CH2N), 3.71 (4H, t, NCH2CH2N), 4.63 (4H, s, NCH2py),7.56 (2H, d, py) and 8.00 (1H, t, py) ppm. 13C NMR (D2O, pD = 2.55): 21.30 (CH2CH2N), 42.25 (NCH2CH2N), 43.20 (NCH2CH2N),44.14 (CH2CH2N), 51.35 (NCH2py), 124.50 (py), 140.44 (py) and150.68 (py) ppm. Found: C, 37.59; H, 7.80; N, 15.37. Calc. for C14H25-N54HCl2H2O: C, 37.80, H, 7.50, N, 15.70%.1.

412 A.S. Fernandes et al. / Journal of Inorganic Biochemistry 105 (2011) 4104192.2.3. Synthesis of the copper(II) complex [Cu[15]pyN5](PF6)2An aqueous solution of Cu(ClO4)2.6H2O (0.150mmol, 0.056 g) was

added to a stirred solution of [15]pyN5 (0.150 mmol, 0.0593 g) dissolvedin theminimumvolume of water ( 1 mL). Then 0.0489 g (0.300 mmol)of NH4PF6was added and themixturewas stirred at 60 C for 1 h. The pHof the solutionwas increased to 6.2 by addition of a solution of KOH0.1 M,and the solventwas evaporatedunder vacuum. The residuewas taken in aminimum amount of methanolacetonitrile (10:2.5). Blue crystals wereformed in about 4 weeks by slow evaporation of the solvent mixture at4 C. Yield: 80%.

2.3. Potentiometric measurements

2.3.1. Reagents and solutionsStock solutions of the ligands were prepared at ca. 2.50103 M.

Metal ion solutions were prepared at about 0.025 to 0.050 M from nitratesalts (analytical grade) in demineralized water (from a Millipore/Milli-Qsystem) and were standardized by titration with Na2H2EDTA [16].Carbonate-free solutions of the titrant, KOH, were prepared at ca.0.10 M by dilution of a commercial ampoule of Titrisol (Merck) withdemineralized water under a stream of pure argon gas. These solutionswere discarded every time carbonate concentrationwas about 0.5% of thetotal amountofbase. For theback titrations, a0.100 Mstandard solutionofHNO3 prepared from a Merck ampoule was used. The titrant solutionswere standardized (tested by Gran method) [17]. For the competitiontitrations, a standard K2H2EDTA aqueous solution was used.

2.3.2. Equipment and work conditionsThe potentiometric setup for conventional titrations consisted of a

50 mL glass-jacketed titration cell sealed from the atmosphere andconnected to a separate glass-jacketed reference electrode cell by aWilhelm-type salt bridge containing 0.10 M KNO3 solution. An Orion720A+measuring instrument tted with a Metrohm 6.0150.100 glasselectrode and a Metrohm 6.0733.100 AgAgCl reference electrode wasused for themeasurements. The ionic strength was kept at 0.100.01 Mwith KNO3, temperature was controlled at 298.20.1 K by circulatingwater through the jacketed titration cell using a Huber Polystat cc1thermostat, and atmospheric CO2 was excluded from the titration cellduring experiments by passing argon across the top of experimentalsolution. Titrant solutionswere added through capillary tips at the surfaceof the experimental solution by a Metrohm Dosimat 765 automaticburette. Titration procedure was automatically controlled by softwareafter selection of suitable parameters, allowing for long unattendedexperimental runs.

2.3.3. MeasurementsThe [H+] of the solutions was determined by the measurement of

the electromotive force of the cell, E=E'o+Q log[H+]+Ej. The termpH is dened as log [H+]. E'o and Q were obtained by titrating asolution of known hydrogen-ion concentration at the same ionicstrength, using the acid pH range of the titration. The liquid-junctionpotential, Ej, was found to be negligible under our experimentalconditions. The value of Kw was determined from data obtained in thealkaline range of the titration, considering E'o and Q valid for the entirepH range and found to be equal to 1013.80 M2. The potentiometricequilibrium measurements were carried out using 20.00 mL of ca.2.50103 M ligand solutions diluted to a nal volume of 30.00 mL,in the absence of metal ions and in the presence of each metal ion forwhich the CM:CL ratio was 1:1. For the reactions of Cu2+ with bothligands, competition titrations were performed. K2H4EDTA was usedas the reference ligand, for which values of protonation and stabilityconstants were determined before under the same experimentalconditions: log K1H=10.22, log K2H=6.16, log K3H=2.71, log K4H =2.0,log KCuEDTA=19.23, log KCuHEDTA=3.06, log KCuEDTAOH=11.33 [18].Ratios of 0.75:1:1 and 1:1:1 (CL:CL':CCu) were used for L=[15]pyN5

and [16]pyN5, respectively, and L=EDTA. The competition reactionsreached equilibrium upon 15 to 20 min at each point in the pH rangewhere the competition reaction took place. The same values for thestability constants were obtained in both directions of the reaction,the direct curve titrating with KOH and the back titration with HNO3.

2.3.4. Calculation of equilibrium constantsOverall equilibrium constants iH and MmHhLl (being MmHhLl=

[MmHhLl]/[M]m[H]h[L]l) were calculated by tting the potentiometricdata from protonation or complexation titrations with the HYPERQUADprogram [19]. Species distribution diagrams were plotted from thecalculated constants with the HYSS program [20]. Only mononuclearspecies, ML, MHL, and MH-1L were found for the metal complexes of theligands (beingMH-1L=MLOHKw).Differences, in logunits, between thevalues of MHL (or MH-1 L) and ML constants provide the stepwisereaction constants. The species considered in a particular model werethose that could be justied by the principles of coordination chemistry.The errors quoted are the standard deviations of the overall stabilityconstants given directly by the program for the input data, which includeall the experimental points of all titration curves, and determined by thenormal propagation rules for the stepwise constants.

Protonation constants were obtained from ca. 180 experimentalpoints, and stability constants for each metal ion were determinedfrom 120 to 180 experimental points (2 or 3 titration curves).

2.4. NMR measurements

2.4.1. Characterization of the macrocyclesThe 1H (400.13 MHz) and 13C NMR (100.62MHz) spectra were

recorded on a Bruker Avance-400 spectrometer at 294 K probetemperature. Chemical shifts () were given in ppm and couplingconstants (J) in Hz. The NMR spectra were performed in CDCl3 ( ppm1H: 7.26; 13C: 77.16) or in D2O. The reference used for the 1H NMRmeasurements in D2O was 3-(trimethylsilyl)propionic acid-d4-sodiumsalt (DSS) and in CDCl3 the solvent itself (at 7.26 ppm). For 13C NMRspectra 1,4-dioxane ( ppm: 1H: 3.75; 13C: 67.20) was used as internalreference. 2D NMR spectra correlation spectroscopy (COSY), hetero-nuclear multiple quantum coherence (HMQC), and heteronuclearmultiple bond correlation (HMBC) were acquired using gradient pulseprograms from Bruker library. Phase-sensitive nuclear Overhauser effectspectroscopy (NOESY) was performed using a mixing time of 1.5 s. Twoand monodimensional FIDs were processed using the TopSpin softwareversion1.3 fromBruker. Peakassignmentswerebasedonpeak integrationand multiplicity for 1H spectra and on 2D experiments for 13C spectra.

2.4.2. NMR titration measurementsThe titration of [15]pyN5 (0.010 M in D2O) was carried out in the

NMR tube. The pD values were adjusted by adding DCl or CO2-freeKOD solutions. Thelog [H] was measured directly in the NMR tubewith a combined glass AgAgCl microelectrode (Mettler-ToledoU402-M3-S7/200) coupled with an Orion 3 Star pH meter. Theelectrode was previously standardized with commercial aqueousbuffer solutions, and the pD values were calculated according to theequation pD=pH+(0.400.02), where pH* is the direct pHreading [21]. The dissociation constants in D2O (pKD) were calculatedfrom the NMR titration data, using a non-linear least-squares curve-tting procedure that minimizes the sum of the squares of thedeviations of the observed and calculated values of the chemicalshifts. These pKD values were converted to pKH values obtained inwater by the equation pKD=0.11+1.10pKH [21].

2.4.3. Magnetic momentsMagnetic moments were measured at 294 K using solutions of Ni

[15]pyN52+ (2.38102 M, pH 6.45) and Ni[16]pyN52+ (2.23102 M,6.32) in D2O. The 1HNMR spectra of the solutionswith DSS, as internal

reference, were acquired in a tube containing an internal capillary

413A.S. Fernandes et al. / Journal of Inorganic Biochemistry 105 (2011) 410419lled with D2O and DSS, and the corresponding magnetic momentscalculated from the shift () between both reference signals [22].

2.5. Spectroscopic studies

Electronic spectra were recorded with a UNICAM model UV-4(UVvisible) or a Shimadzu model UV-3100 (UVvisible-near IR)spectrophotometers using aqueous solutions of Ni2+ and Cu2+

complexes of both macrocycles (1.0102 to 1.0103 M) at pHs6.63 to 7.05.

EPR spectroscopy measurements of copper(II) complexes of [15]pyN5 and [16]pyN5 were recorded at 99 K with a Bruker EMX 300spectrometer equipped with continuous-ow cryostats for liquidnitrogen, operating at X-band. The complexes were prepared at about1.0103 M and pH values of 4.96, 7.23, and 9.79 for Cu[15]pyN52+, and5.04, 7.23, and 9.92 for Cu[16]pyN52+, in 1 M NaClO4 aqueous solution.

2.6. Electrochemical studies

A BAS CV-50W Voltammetric Analyzer connected to BAS/Win-dows data acquisition software was used. Cyclic voltammetricexperiments were performed in a glass cell MF-1082 from BAS in aC-2 cell enclosed in a Faraday cage, at room temperature, under argon.The reference electrode was AgAgCl (MF-2052 from BAS) lled withNaCl3 Minwater,standardizedfortheredoxcoupleFe(CN)63/Fe(CN)64. The auxiliary electrode was a 7.5-cm platinum wire (MW-1032 from BAS) with a gold-plated connector. The working electrodewas a glassy carbon (MF-2012 from BAS).

Copper(II) complexes of [15]pyN5 and [16]pyN5 (1.63103 M;pH=7.05 and 1.46103 M; pH=7.09, respectively) were preparedin 0.1 M KNO3 in water. The solutions were deaerated by an argonstream prior to all measurements and were kept under argon duringthe measurements. Between each scan, the working electrode waselectrocleaned by multi-cycle scanning in the supporting electrolytesolution, polished on diamond 1 m and on alumina 0.3 m, cleanedwith water and sonicated before use, according to standardprocedures.

Cyclic voltammograms with sweep rate ranging from 25 to1000 mV s1 were recorded in the region from +1.2 to 1.2 V. Atthis potential range the ligands were found to be redox inactive. Thehalf-wave potentials, E1/2, were obtained by averaging the anodic andcathodic peak potentials. All potential values are reported relative tothe AgAgCl reference electrode and the E1/2 andEp of the Fe(CN)63/Fe(CN)64 couple, under our experimental conditions, were 196 mVand 73 mV, respectively.

2.7. X-ray crystallography

Blue crystals of [Cu[15]pyN5](PF6)2 with suitable quality for singlecrystal X-ray diffraction determination were grown up from metha-nolacetonitrile solution.

Crystal data: C13H21CuF12N5, Mr=600.83; monoclinic, spacegroup P21/c, Z=4, a=8.8619(9) b =14.9388(14), c=16.6689(16), =103.674(9), U=2144.2(4) 3, (calc)=1.861 Mg m3, (Mo-K) = 1.283 mm1.

X-ray datawere collected at 150(3) K on a CCDX-calibur plate systemusing graphite monocromatized Mo-K radiation (=0.71073 ) atReading University. The selected crystal was positioned at 50 mm fromthe CCD, and the frames were taken using a counting time of 2 s. Theprocessing of the data was carried out with the Crysalis program [23].Intensities were corrected for empirical absorption effects with theABSPACK program [24]. The structure was solved by direct methods andbysubsequentdifferenceFourier synthesesand renedby fullmatrix leastsquares on F2 using the SHELX-97 suite [25]. Anisotropic thermalparameters were used for the non-hydrogen atoms. The hydrogen

atoms bonded to carbon and nitrogen atomswere included in renementin calculated positions with isotropic parameters equivalent to 1.2 timesthose of the atom to which they were attached. The nal renement of298 parameters converged to nal R and Rw indices R1=0.0467 andwR2=0.1028 for 2861 reections with IN2(I) and R1=0.0992 andwR2=0.1078 for all 6262 hkl data. Molecular diagrams presented aredrawn with graphical package software PLATON [26].

3. Results and discussion

3.1. Synthesis and characterization of the macrocycles

Compounds [15]pyN5 and [16]pyN5 were prepared in good yieldby [1:1] condensation of 2,6-pyridinedicarboxaldehyde and N,N-bis(2-aminoethyl)ethane-1,2-diamine (trien) and N,N-bis(2-ami-noethyl)1,3-propanediamine, respectively, using Pb2+ as the tem-plate ion, followed by reduction of the resulting tetraimines withsodium borohydride. The pure products were obtained as tetrahy-drochloride salts in 85% and 46% yields, respectively. The lower yieldof the later compound results from the unfavourable adoptedgeometry of the lead(II) complex during the cyclization reaction [27].However, Ca2+ or Ba2+ did not lead to better yields.

Both macrocycles were synthesized by different and more timeconsuming procedures [2830]. Stetter et al. [28] prepared [15]pyN5following amodied Richman and Atkinsmethod [31] in 78% yield, andRiley et al. [29] followed the same procedure with minor changes.Kimura et al. [30] prepared [16]pyN5, in unspecied yield, by reuxingthe bisdiethyl esters of pyridine-2,6-dicarboxylic acid and N,N-bis(2-aminoethyl)1,3-propanediamine in ethanol and high dilution followedby reduction of the resulted diamide with diborane in tetrahydrofuran.

1D and 2D NMR spectroscopy were used for characterization of [15]pyN5 and [16]pyN5. The chemical shifts and the corresponding assign-ments were accomplished by 1H, 13C, COSY, HMQC, HMBC, and NOESY atpD 5.10 and 2.55, respectively, as described in Appendix A of theSupplementary material (cf. Table S1 and Figs. S1S5).

3.2. Acidbase behaviour of the ligands

The acidbase behaviour of [15]pyN5 and [16]pyN5 was studied bypotentiometry in water at 298.2 K and ionic strength 0.10 M in KNO3.The former compound was also studied by 1H NMR spectroscopy. Thedetermined protonation constants are collected in Table 1 togetherwith the values of the related [15]aneN5 and [16]aneN5 compounds(cf. Scheme 1) for comparison. Both compounds have ve basiccentres; however, only three constants for [15]pyN5 and four for [16]pyN5 could be accurately determined by potentiometry and one morefor [15]pyN5 was obtained by 1H NMR. The two compounds exhibithigh and fairly high values, respectively, for the rst two protonationconstants corresponding to the protonation of nitrogen atoms inopposite positions, minimizing the electrostatic repulsion betweenpositive charges of the ammonium groups formed. The third andfourth constants are much lower due to the stronger electrostaticrepulsions as they correspond to protonation of nitrogen atoms atshort distances from already protonated ones and to the limitedmotion allowed in the ring backbone. The increase in basicity of thesetwo last centres in [16]pyN5 is correlated with the increase of thelength of the chain between contiguous nitrogen atoms. The valuesreported before (in NaClO4 medium) [2933] shown in Table 1 differslightly from ours; however, for the rst time, we were able toaccurately determine the fourth protonation constant.

The overall basicity and all the stepwise protonation constants of [15]pyN5and [16]pyN5 (see Scheme1andTable1) are smaller than that of thecorresponding macrocycles without pyridine as expected taking intoaccount the electron withdrawing effect of the pyridine ring.

1H NMR spectroscopic titration of [15]pyN5 was carried out inorder to understand its protonation sequence and to determine the

lower protonation constants. In Fig. 1 is shown the spectrum of the

ligand at pD 5.10 and the titration curves for all resonances. The 1HNMR spectrum exhibits six resonances in the 1.247.04 pD region, butfor higher pD values, Hd and He resonances overlap. The resonances at

7.99 and 7.53 ppm were assigned to Ha and Hd protons, the twosinglets at 4.57 and 3.21 ppm to Hc and Hf protons and the triplets at3.51 and 3.41 ppm to Hd and He protons, respectively.

Table 1Stepwise protonation constants (log KiH) of [15]pyN5, [16]pyN5 and other similar compounds for comparison.a T=298.2 K; I=0.10 M in KNO3.

Equilibrium quotient [15]pyN5 [16]pyN5 [15]aneN5b [16]aneN5b

[HL+]/[H+][L] 9.616(8) 9.43c; 9.11d 9.71(1) 9.48e 10.85 10.64[H2L2+]/[HL+][H+] 8.67(1) 8.80c; 8.82d 8.32(3) 8.56e 9.65 9.49[H3L3+]/[H2L2+][H+] 5.33(2) 5.28c; 5.27d 5.56(5) 5.83e 6.00 7.28[H4L4+]/[H3L3+][H+] 1.4(2)f 2.37(8) b2e 1.74 1.71[H5L5+]/[H4L4+][H+] 1.16 1.45

[H4L4+]/[L][H+]4 25.02 25.96 28.24 29.12

a Values in parentheses are standard deviations on the last signicant gure.b T=298.2 K; I=0.2 M in NaClO4; ref. [33].c T=298.2 K; I=0.1 M in NaClO4; ref. [29].d T=298.2 K; I=0.1 M NaClO4; ref. [32].e T=298.2 K; I=0.2 M in NaClO4; ref. [30].f Determined in this work by 1H NMR spectroscopy, using the calculated value of pKD4 and the equation pKD=0.11+1.10pKH [21].

414 A.S. Fernandes et al. / Journal of Inorganic Biochemistry 105 (2011) 410419Fig. 1. (a) 1H NMR titration curves for [15]pyN5, chemical shift H (ppm) in function of pD; (b) 1H NMR spectrum of [15]pyN5 (D2O, pD 5.10).

ML2+ and MHL3+ are formed, but MLOH+ species were also found forCu2+ with [15]pyN5 and [16]pyN5 and Pb2+ with [16]pyN5. In allcases, the proposed model was accepted by the HYPERQUAD program[19] using all data points from all titration curves, with good statisticalparameters. The complexes of Co2+ with both ligands wereimpossible to study due to their fast oxidation, which occurred evenunder argon atmosphere owing to small oxygen impurities in thepurge gas. The very low value for the Ca[16]pyN52+ was also

Table 4pM values for [15]pyN5, H4EDTA, and H5DTPA with some divalent metal ions.

Ion [15]pyN5a H4EDTAb H5DTPAb

Ca2+ 5.00 7.89 6.55Ni2+ 18.19 15.68 16.01Zn2+ 14.24 13.84 14.44

a Calculated from the constants in Tables 1 and 2.b Calculated from the values of the protonation constants and of the stability

constants reported in refs. [18,38]. All the values calculated for 100% excess of freeligand at physiological conditions, pH=7.40; CM=1.0105 M, CL=2.0105 M,using the Hyss program [20].

Table 2Stepwise stability constants (log units) of the complexes of [15]pyN5, [16]pyN5, andother related ligands with several metal ions.a T=298.2 K; I=0.10 M.

Equilibrium quotient [15]pyN5 [16]pyN5 [15]aneN5 [16]aneN5

[CaL2+]/[Ca2+][L] 3.21(2) [NiL2+]/[Ni2+][L] 21.71(1) 16.81(1) 18.1b 18.1c

[NiHL3+]/[NiL2+][H+] 2.88(2) [CuL2+]/[Cu2+][L] 23.31(3) 20.86(4) 28.3d 27.1d

[CuHL3+]/[CuL2+][H+] 2.31(4) 3.59(6) 4.3d [CuL2+]/[CuLOH+][H+] 10.06(6) 11.83(7) [ZnL2+]/[Zn2+][L] 17.76(2) 15.30(1) 19.1e 17.9 e

[ZnHL3+]/[ZnL2+][H+] 2.68(3) 3.28(3) 3.1e 3.7 e

[CdL2+]/[Cd2+][L] 16.853(7) 15.532(4) 19.2e 18.1e

[CdHL3+]/[CdL2+][H+] 2.64(4) 3.05(2) 3.4e 3.9e2+ 2+ e e

415A.S. Fernandes et al. / Journal of Inorganic Biochemistry 105 (2011) 410419The 1HNMR titration curves show the effect of successive protonationof basic centres of the molecule. The rst equivalent protonates mainlyN3 since the downeld shift of Hf resonance is larger, followed by the shiftof He and Hd protons. The Hc resonance has a very small shift in this pDrange,meaning a small percentageof protonationofN2 atoms. The secondacid equivalent added protonates mainly the N2 centre, as Hc, Hb, andHa resonancesmovedowneld. SimultaneouslyN3 centrewasprotonatedto a low degree, as the Hf, He, and Hd resonances show a slight downeldshift. The third acid equivalent (pD 7.213.12) continues protonatingthe N2 centre, since c and d resonances shift downeld. A slight shiftof He and Hf resonances reveals a small percentage of protonation on N3

centres. The addition of one more equivalent of acid (pD 3.121.24) onlyprotonates N3 atoms as Hf, He, and Hd resonances move downeld. Theabsence of any change on Ha, Hb, and Hc resonances suggests that N1 isnot protonated, even at very low pD values.

The 1H NMR titration also allowed the determination of the pro-tonation constants in D2O for [15]pyN5: pKD1=10.61(7), pKD2=9.6(l),pKD3=6.29(7), and pKD4=1.6(2). These values are in agreement withthe equation for the correlation between the protonation constantsdetermined in H2O and in D2O for polyaza and polyoxapolyazamacrocyclic compounds: pKD=0.11+1.10pKH [21].

3.3. Thermodynamic stability of metal complexes

[PbL ]/[Pb ][L] 15.44(2) 12.932(6) 17.3 14.3[PbHL3+]/[PbL2+][H+] 3.4(3) 3.61(5) 3.8e 5.0e

[PbL2+]/[PbLOH+][H+] 10.66(2)

a Values in parentheses are standard deviations on the last signicant gure.b T=308.2 K; ref. [35].c Ref. [36].d T=298.2 K; I=0.2 M; polarographic method; ref. [37].e T=298.2 K; I=0.2 M; ref. [37].The stability constants of [15]pyN5 and [16]pyN5 with Ca2+, Ni2+,Cu2+, Zn2+, Cd2+, and Pb2+, determined by potentiometric titrationsat the experimental conditions already indicated for the protonationconstants, are collected in Table 2 together with those of the relatedmacrocycles [15]aneN5 and [16]aneN5 taken from the literature forcomparison. Only mononuclear species (1:1, metal-to-ligand ratio)were found for the complexes of bothmacrocycles. In most cases, only

Table 3pM valuesa calculated for [15]pyN5, [16]pyN5, and other similar ligands with severaldivalent metal ions.

Ion [15]pyN5 [16]pyN5 [15]aneN5 [16]aneN5

Ca2+ 5.00 Ni2+ 18.19 13.52 12.38 12.52Cu2+ 19.79 17.57 22.78 21.52Zn2+ 14.24 12.01 13.38 12.32Cd2+ 13.34 12.25 13.48 12.52Pb2+ 11.92 9.65 11.58 8.72

a Calculated from the constants in Tables 1 and 2 for 100% excess of free ligand atphysiological conditions, pH=7.40; CM=1.0105M, CL=2.0105M, using theHyss program [20].impossible to obtain by the method used. Direct determinations ofthe stability constants of Cu[15]pyN52+ and Cu[16]pyN52+ were notpossible as ML2+ was completely formed in the beginning of thetitration (pH2.2) and, consequently, reliable values for theconstants were obtained through a competition with a second ligand,for which the protonation and stability constants are accuratelyknown [34]. Among the various ligands tried, H4EDTA was chosen asthe best second ligand. In spite of the higher overall basicity of [16]pyN5, this ligand forms ML2+ complexes less stable than those of [15]pyN5 (variations of 1.32 to 4.9 log units), being the largest decrease forthe nickel(II) complexes followed by the copper(II) ones. However,contrary to [15]pyN5, [16]pyN5 forms ML2+ complexes with Cd2+

slightly more stable than with Zn2+. Differences in the cavity size ofboth ligands are responsible for this different behaviour.

The comparisonof stability constant values of themetal complexes ofthe ligandswith andwithout pyridine (cf. Scheme1 andTable 2) [3538]reveals that the former complexes present lower values, except for Ni[15]pyN52+. However, stability constants do not provide directlycomparable basis for the measuring total ion sequestering abilities ofthe ligands at physiological conditions (pH 7.4) and therefore they wereused to calculate the pM values, dened aslog [M2+ ] (cf. Table 3). Theadvantage of comparing pM values rather than stability constants is thatthe pM values reect the inuence of ligand basicity and metal chelateFig. 2. Species distribution curves for aqueous solutions containing Ni2+, Zn2+, Cd2+,Pb2+, and [15]pyN5 (L) at 1:1:1:1:1 molar ratio. Percentages are given relative to thetotal amount of [15]pyN5 at an initial value of 1.67103 M.

protonation. The pM values for the nickel(II) complexes of [15]pyN5 and[16]pyN5 are signicantly higher than those of [15]aneN5 and [16]aneN5(differences are, in log units, 5.81 for the 15-memberedmacrocycles and1.0 for the 16-membered ones). The zinc(II), cadmium(II), and lead(II)complexes with both series of macrocycles have pM values of the sameorder of magnitude, while for copper(II) complexes, the values aremarkedly higher for ligands without pyridine (differences, in log units,are 2.99 for the 15-membered and 3.95 for the 16-memberedmacrocycles). The last pM differences can, in part, be related to theconformation adopted by the ligands upon complexation. Nevertheless,a polarographic technique was used for stability constants determina-tions of Cu[15]aneN52+ and Cu[16]aneN52+, and additionally, nocompetition reaction with a second ligand was employed. We areconvinced that those values should be conrmed for denitiveconclusions to be drawn.

The very high stability constant value of Ni[15]pyN52+ led us toevaluate the potential role of [15]pyN5 as chelator for removal ofexcess of nickel(II) from the body. In Table 4 are collected the pMvalues for nickel(II) and other two important biological metal ions,calcium(II) and zinc(II), for our chelator together with clinically usedones, namely H4EDTA and H5DTPA. The [15]pyN5 presents not onlythe highest pNi value but also the largest selectivity (differencebetween pM values) towards zinc(II) and calcium(II).

Additionally, [15]pyN5 can be used in the quantitative determina-

Table 5Spectroscopic UVvisible-near IR data and magnetic moments () for the Ni(II) complexes of [15]pyN5 and [16]pyN5.

Complex; color pH UVvisible-near IRa max/nm (, M1 cm1) (MB)

Ni[15]pyN52+ (yellow) 6.69 1150 (4.7), 930 (10.8), 800 (sh. 9.0), 600 (10.0), 530 (11.2), 306 (172.6), 262 (2.48103) 3.24Ni[16]pyN52+ (blue) 6.86 1148 (18.7), 1060 (sh. 29.3), 940 (41.8), 880 (sh. 37.3), 810 (sh. 27.1), 625 (17.8) 403 (18.2), 345 (82.0), 309 (176.7), 262 (2.16103) 3.37

a sh.=shoulder.

416 A.S. Fernandes et al. / Journal of Inorganic Biochemistry 105 (2011) 410419Fig. 3. X-band EPR spectra of Cu[15]pyN52+ (a) and Cu[16]pyN52+ complexes (b) in anaqueous solution of 1.0103 M and in 1.0 M of NaClO4, both recorded at 99 K andat pH 7.23. Microwave power of 2.0 mW, modulation amplitude of 1.0 mT, and thefrequency () was of 9.41 GHz. The simulated spectra are shown in gray, theexperimental ones are in black.tion of Ni2+ in solutions containing also zinc(II), cadmium(II), andlead(II) (in similar amounts), as can be observed by the speciesdistribution diagram in Fig. 2.

3.4. Spectroscopic studies

3.4.1. Nickel(II) complexesThe UVvisible-near IR spectra for Ni[15]pyN52+ and Ni[16]pyN52+

complexes were recorded in water solutions at pH values 6.69 and 6.86,respectively (cf. Table 5). The electronic spectrum of the yellow Ni[15]pyN52+ exhibits two absorption bands of low intensities at 530 and930 nm, and the charge transfer band, at 262 nm. The Ni[16]pyN52+

complex is blue, and the spectrum also exhibits one intense peak that isascribed to a charge-transfer absorption band, and bands at 625 and940 nm. The 1.75 and 1.50 ratios between the near IR (1) and the visible(2) bands and the corresponding magnetic moments of 3.24 BM and of3.37 BM calculated for the two complexes, respectively, are characteristicof high-spin six coordinate nickel(II) centre in distorted tetragonalsymmetry [39], where the ve positions are occupied by nitrogen atomsof the backbone and the last position by the solvent (water or metalcounter-ion nitrate). Therefore, in solution both complexes exhibitstructures that are not quite different, even though the Ni[15]pyN52+

presents stronger equatorial eld and Ni[16]pyN52+ a more distortedgeometry.

Following considerations of Busch and co-workers [40], weassigned the visible-near IR bands to 3B1g3B2g, directly related to10Dqxy and 3B1g 3Ega, equal to the difference between 10Dqxy and35/4Dt transitions. The values of the equatorial and axial ligand eldwere calculated based on these assignments: Dqxy=1887 cm1 andDqz=260 cm1 for Ni[15]pyN52+ and Dqxy=1600 cm1 andDqz=525 cm1 for Ni[16]pyN52+. Therefore, Dqz is strongly inuencedby the in-plane ligand eld and decreases as Dqxy increases, as found inother cases [41]. Similar geometry was described for Ni(Me2[15]pyN5)2+

(Me2[15]pyN5=2,13-dimethyl-3,6,9,12,18-pentaazabicyclo[12.3.1]-octadeca-1(18),14,16-triene) based on spectroscopic studies in solutionand supported by molecular models [42].

3.4.2. Copper(II) complexesThe Cu[15]pyN52+ and Cu[16]pyN52+ exhibit broad bands in the

visible region, due to the copper dd transitions with max at 610 and646 nm, respectively. The corresponding X-band EPR spectra exhibitthe four expected lines at low eld, due to the interaction of theunpaired electron spin with the copper nucleus, and a strongunresolved band at high eld, see Fig. 3. Bands in the visible region(max) and the hyperne coupling constants Ai (i=x, y, and z) and gvalues obtained by the simulation of the spectra [43] are shown inTable 6Spectroscopic X-band EPR data for the Cu(II) complexes of [15]pyN5 and [16]pyN5.

Complex Visible band, max/nm(, M1 cm1)

EPR parametersAi104 cm1

gx gy gz Ax Ay Az

Cu[15]pyN52+ 610 (150) 2.035 2.070 2.210 26.9 40.0 170.6Cu[16]pyN52+ 646 (143) 2.038 2.077 2.222 23.2 33.5 163.8

Table 6. These values are characteristic of mononuclear copper(II)complexes in rhombic symmetry with elongation of the axial bondsand a dx2y2 ground state. Elongated octahedral or distorted squarepyramidal stereochemistries are consistent with these data, buttrigonalbipyramidal or tetragonal geometries involving compressionof axial bonds can be excluded [4446].

The g and Ai parameters are related to the electronic transitions bythe factors derived from ligand eld theory [4750]: the g valuesincrease and the Az value decreases as the planar ligand eld becomesweaker or as the axial ligand eld becomes stronger, and this occurswith the simultaneous red-shift of the dd absorption bands in theelectronic spectra. This sequence, in principle, parallels the degree ofdistortion from square-planar to square pyramidal, and then to

that can be assigned to the Cu(II)/Cu(I) couple. Upon repetitive cycling,the voltammetric response remained essentially unchanged. This feature

Table 9Dimensions of the NHF hydrogen bonds for [Cu[15]pyN5](PF6)2.

Table 8Selected bond distances () and angles () in the copper(II) coordination sphere.

CuN(4) 2.060(3)CuN(7) 1.921(3) CuN(10) 2.011(2)CuN(1) 2.229(3) CuN(13) 2.034(3)N(7)CuN(10) 82.5(1) N(1)Cu-N(4) 81.7(1)N(7)CuN(13) 156.1(1) N(10)CuN(13) 85.4(1)N(7)CuN(4) 82.4(1) N(10)CuN(4) 164.5(1)N(13)CuN(4) 107.3(1) N(7)CuN(1) 120.9(1)N(10)CuN(1) 109.3(1) N(13)CuN(1) 82.6(1)

Table 7Cyclic voltammetric data for [15]pyN5 and [16]pyN5 copper(II) complexes.a

Complex Epc/mV Epa/mV Ep/mV E1/2/mV

Cu[15]pyN52+ 749 673 76 711Cu[16]pyN52+ 649 569 80 609a Scan rate=100 mV s1. E1/2 values (vs. AgAgCl) were taken as the average of the

anodic (Epa) and the cathodic peak potentials (Epc). Ep=|EpaEpc|.

417A.S. Fernandes et al. / Journal of Inorganic Biochemistry 105 (2011) 410419octahedral or tetragonal geometries [5153]. In agreement with this,the Cu[15]pyN52+ complex exhibits the lowest gz value and the largestAz and simultaneously its absorption band is blue shifted in relation tothat of Cu[16]pyN52+, pointing to a stronger equatorial ligand eld,indicating similar structures for the complexes of both macrocycles,consistent with distorted square pyramidal geometry, as alsoobserved in the crystal [Cu[15]pyN5](PF6)2 (vide infra).

3.5. Cyclic voltammetry studies

The redox behaviour of Cu[15]pyN52+ and Cu[16]pyN52+ wasinvestigated by cyclic voltammetry in water. In Table 7 are depictedtheir electrochemical data, where Epa and Epc are the anodic and thecathodic peak potentials, respectively, and Ep=|EpaEpc|. As can beseen, the two complexes exhibit analogous electrochemical behaviourshowing a single quasi-reversible one-electron transfer reductionprocess at half-wave potential values, E1/2, (vs. AgAgCl) of 711 mV(EpaEpc=76 mV) and 609 mV (EpaEpc=80mV), respectively,Fig. 4. Molecular structure of [Cu[15]pyN5]2+ complex with atomic labelling schemeadopted.indicates that the initial copper complexes are regenerated during thepotential scan. For both copper(II) complexes, the E1/2 values wereindependent when the scan rate () was varied between 25 and1000 mV s1, theEp values increased and thepeak current ratio (Ipa/Ipc)was slightlydifferent but close tounity. Furthermore, a linear relationshipbetween the peak currents and the square root of the (1/2) wasobserved. This fact implies that these electrochemical processes aremainly diffusion-controlled.

The Cu[16]pyN52+ yields a E1/2 value that is shifted to less negative,indicating a easier reduction to Cu(I), than the corresponding valueobserved for Cu[15]pyN52+. This difference, which is in agreementwith the stability constants discussed before, can be rationalized interms of exibility and size of the macrocyclic cavities in bothcomplexes, the geometric requirements and the size of the metal ionin different oxidation states. The reduction of Cu(II) (d9) to Cu(I) (d10)involves a drastic increase in the metal radius and a geometric changefrom pyramidal to tetrahedral. Obviously, the larger and more exiblecavity of [16]pyN5 compared to that of [15]pyN5 tends to stabilizebetter the copper(I) complex.

3.6. X-ray structure of the copper(II) complex

The single crystal structure of [Cu[15]pyN5](PF6)2 was determinedby X-ray diffraction. The molecular structure of [Cu[15]pyN5]2+

presented in Fig. 4 shows the metal centre coordinated by the venitrogen donor atoms from [15]pyN5. Selected bond distances andangles given in Table 8 indicate that the copper(II) centre hasdistorted square pyramidal geometry. The basal plane is formed bythe nitrogen atoms N(3), N(4), N(7), and N(10) with the trans anglesN(7)CuN(13) and N(10)CuN(4) of 156.1(1) and 164.5(1),respectively. The apical position is occupied by the remainingnitrogen donor N(1), which is 2.263(3) from the least squaresplane dened by the basal nitrogen donors. The copper centre is 0.238(1) from this plane towards the apical site leading to a CuN(1)distance of 2.229(3) . On the other hand, the N(7)CuN(1) angle of120.9(1) seems to indicate a tendency of the metal coordinationd(HF)/ d(NF)/ bNHF/

N(1)H(1) 2.44 3.319(4) 163 F(21)N(4)H(4) 2.30 3.034(3) 138 F(12) [x+1, y+1, z]N(4)H(4) 2.49 3.186(3) 133 F(26) [x, y+1/2, z1/2]N(10)H(10) 2.28 3.129(3) 156 F(13)

[x+1, y1/2,z+1/2]N(10)H(10) 2.43 3.061(3) 127 F(11)

[x+1, y1/2,z+1/2]N(10)H(10) 2.43 2.975(4) 119 F(22)N13H(13) 2.37 3.084(3) 136 F(12) [x+1, y+1, z]N(13)H(13) 2.53 3.296(3) 142 F(16) [x+1, y+1, z]

theis r

418 A.S. Fernandes et al. / Journal of Inorganic Biochemistry 105 (2011) 410419environment for a trigonal bipyramidal geometry. However, thetrigonal index dened as t=()/60 (where and are the largestangles in the metal coordination sphere with N) [54], is only 0.13indicating undoubtedly the existence of a copper square pyramidalmetal coordination sphere, taking into account that this angularparameter has an ideal value of 0 for the square pyramidal geometrywith C2v symmetry and 1 for the trigonal bipyramidal geometry withD3h symmetry. Furthermore, the basal nitrogen donors display anaverage tetrahedral distortion of0.123(1) , which is consistentwith the spectroscopic data of the complex reported above.

To achieve the geometric arrangement described, the macrocycleis folded through the axis dened by the nitrogen atoms contiguous tothe pyridine ring (N(4) and N(10)) leading to a dihedral anglebetween the equatorial plane and the plane dened by the nitrogenatoms N(1), N(4), and N(10) of 64.80(9).

The CuN(sp2) distance is shorter than the remaining four CuN(sp3)distances following the usual structural trend found for copper complexesof tetraazamacrocyles incorporating the NCH2(pyridine)CH2N frag-ment [55]. Furthermore, the CuN distances of [Cu[15]pyN5]2+ comparewellwith those found for the related 15-memberedmacrocyclic complex,[Cu(Me2[15]pyN5)]2+,whichexhibits a similar coordinationenvironment[56].

InTable 9 are gathered theNHFhydrogenbonds found in the solidstate for [Cu[15]pyN5](PF6)2. The [Cu[15]pyN5]2+ cations and PF6 anionsare assembled into 1D innite chains by multiple NHF hydrogenbonds along the [001] crystallographic direction. Furthermore, one ofthese chains presented in Fig. 5 shows that there are eight independenthydrogen bonds derived from the interaction of all NHbinding groups ofeach [Cu[15]pyN5]2+ complex with two independent counter-ions. Inaddition two of these intermolecular bonding interactions are bifurcated,one trifurcated andone almost linearwith anNFdistance of 3.318(4)and an NHF angle of 163.

4. Conclusions

Two macrocyclic ligands having ve donor nitrogen atoms, one ofthem being a pyridine, [15]pyN5 and [16]pyN5, have been synthe-

Fig. 5. Crystal packing diagram showing the 1D chain formed by the interaction betweenred lines). (For interpretation of the references to color in this gure legend, the readersized. In spite of being known for several years, scarce quantitativeevaluation of the binding ability of these macrocycles to chelate metalions has been carried out until the present work. Here the acidbasereactions of both macrocycles have been studied and their stabilityconstants with several metal ions of biological relevance or ability toact as toxic agents were determined by accurate techniques. Theincrease of the cavity size of the macrocycles, from 15 to 16 members,led to a decrease of all the stability constants without any specialincrease of selectivity. Therefore, from both chelators, the [15]pyN5 isthe more promising for the aimed medical applications. However, acomparison of this ligand with [15]aneN5 revealed that the lattermacrocycle is a better chelator for copper(II), although a denitiveconclusion implies the redetermination of the stability constant of Cu[15]aneN52+ using accurate methods. Nevertheless, the pCu valuecalculated for [15]pyN5 of 19.79 (cf. Table 3) is much higher than the16.36 value determined under the same conditions for the clinicallyavailable copper(II) chelator trien [57].

Concerning nickel(II), [15]pyN5 is a very strong chelator and selectivetowards zinc(II) and calcium(II) essential metal ions and therefore itsevaluation for chelation therapy is pertinent. These encouraging chemicalresults warrant further studies.

Abbreviations[15]pyN5 3,6,9,12,18-pentaazabicyclo[12.3.1]octadeca-1

(18),14,16-triene[16]pyN5 3,6,10,13,19-pentaazabicyclo[13.3.1]nonadeca-1

(19),15,17-triene[15]aneN5 1,4,7,10,13-pentaazacyclopentadecane[16]aneN5 1,5,8,11,14-pentaazacyclohexadecaneDTPA diethylenetriaminepentaacetic acidDSS 3-(trimethylsilyl)propionic acid-d4-sodium saltpM concentration of free metal ion in solutiontrien N,N-bis(2-aminoethyl)ethane-1,2-diamine

Acknowledgments

The authors acknowledge the nancial support from Fundao para aCincia e a Tecnologia (FCT), with co-participation of the EuropeanCommunity fund FEDER (project no. PTDC/QUI/67175/2006). The authorswish to thank the Elemental Analysis Service Unit of ITQB-UNL forproviding analytical data. A.S.F. acknowledges Fundao para a Cincia e aTecnologia, Portugal, for the nancial support (PhD grant SFRH/BD/28773/2006). We also thank the EPSRC (U.K.) and the University ofReading for funds for the diffractometer.

Appendix A. Supplementary data

Crystallographic data for the structure [Cu[15]pyN5](PF6)2 in thispaper have been deposited with the Cambridge Crystallographic DataCentre as supplementary publication number CCDC782478. Copy of thedata can be obtained, free of charge, on application to CCDC, 12 Union

PF6 counter-ions and [Cu[15]pyN5]2+ complexes via NHF hydrogen bonds (dashedeferred to the web version of this article.)Road, Cambridge, CB2 1EZ, UK [fax: +44(0) 1223 336033 or e-mail:deposit@ccdc.cam.ac.uk]. Supplementary data to this article can befound online at doi:10.1016/j.jinorgbio.2010.11.014.

References

[1] M. Blanua, V.M. Varnai, M. Piasek, K. Kostial, Curr. Med. Chem. 12 (2005) 27712794.[2] T. Wang, Z. Guo, Curr. Med. Chem. 13 (2006) 525537.[3] K. Camphausen, M. Sproull, S. Tantama, S. Sankineni, T. Scott, C. Mnard, C.N.

Coleman, M.W. Brechbiel, Bioorg. Med. Chem. 11 (2003) 42874293.[4] S. Bolognin, D. Drago, L. Messori, P. Zatta, Med. Res. Rev. 29 (2009) 547570.[5] E. Gaggeli, H. Kozlowsi, D. Valensin, G. Valensin, Chem. Rev. 106 (2006) 19952044.[6] F. Tisato, C.Marzano,M. Porchia,M. Pellei, C. Santini,Med. Res. Rev. 30 (2010) 708749.[7] K.G. Daniel, P. Gupta, R.H. Harbach, W.C. Guida, Q.P. Dou, Biochem. Pharmacol. 67

(2004) 11391151.[8] O. Andersen, Chem. Rev. 99 (1999) 26832710.[9] E. Denkhaus, K. Salnikow, Crit. Rev. Oncol. Hematol. 42 (2002) 3556.

[10] International Agency for Research on Cancer, IARC monographs on the evaluationof carcinogenic risks to humans, IARC, Lyon, 1990.

[11] J. Saary, R. Qureshi, V. Palda, J. DeKoven, M. Pratt, S. Skotnicki-Grant, L. Holness, J. Am.Acad. Dermatol. 53 (2005) 845855.

[12] J.P. Thyssen, T. Menn, Chem. Res. Toxicol. 23 (2010) 309318.[13] R.D. Hancock, A.E. Martell, Chem. Rev. 89 (1989) 18751914.[14] J. Costa, R. Delgado, Inorg. Chem. 32 (1993) 52575265.[15] D.D. Perrin, W.L.F. Armarego, Purication of Laboratory Chemicals, 3rd ed.

Pergamon, Oxford, 1988.[16] G. Schwarzenbach, W. Flaschka, Complexiometric Titrations, Methuen & Co, London,

1969.[17] F.J. Rossotti, H.J. Rossotti, J. Chem. Educ. 42 (1965) 375378.[18] R. Delgado, M.C. Figueira, S. Quintino, Talanta 45 (1997) 451462.[19] P. Gans, A. Sabatini, A. Vacca, Talanta 43 (1996) 17391753.[20] L. Alderighi, P. Gans, A. Ienco, D. Peters, A. Sabatini, A. Vacca, Coord. Chem. Rev. 184

(1999) 311318.[21] R. Delgado, J.J.R. Frasto da Silva, M.T.S. Amorim, M.F. Cabral, S. Chaves, J. Costa,

Anal. Chim. Acta 245 (1991) 271282.[22] D.F. Evans, J. Chem. Soc. (1959) 20032005.[23] CRYSALIS, Oxford Diffraction Ltd, 2005.[24] ABSPACK, Oxford Diffraction Ltd, 2005.[25] G.M. Sheldrick, Acta Cryst. A64 (2008) 112122.[26] A.L. Spek, PLATON, A Multipurpose Crystallographic Tool, Utrecht University,

Utrecht, The Netherlands, 2010.[27] N.V. Gerbeleu, V.B. Arion, J. Burgess, Template Synthesis of Macrocyclic

Compounds, Wiley-VCH, Weinheim, 1999.[28] H. Stetter, W. Frank, R. Mertens, Tetrahedron 37 (1981) 767772.[29] D.P. Riley, S.L. Henke, P.J. Lennon, R.H. Weiss, W.L. Neumann, W.J. Rivers, K.W.

Aston, K.R. Sample, H. Rahman, C. Ling, J. Shieh, D.H. Busch, W. Szulbinski, Inorg.Chem. 35 (1996) 52135231.

[30] E. Kimura, M. Kodama, R. Machida, K. Ishizu, Inorg. Chem. 21 (1982) 595602.[31] J.E. Richman, T.J.J. Atkins, Am. Chem. Soc. 96 (1974) 22682270.[32] A. Dees, A. Zahl, R. Puchta, N.J.R. E-Hommes, F.W. Heinemann, I. Ivanovic-

Burmazovic, Inorg. Chem. 46 (2007) 24592470.

[33] M. Kodama, E. Kimura, Dalton Trans. (1978) 104110.[34] J. Costa, R. Delgado, M.G.B. Drew, V. Flix, Dalton Trans. (1998) 10631071.[35] M. Kodama, E. Kimura, S. Yamaguchi, Dalton Trans. (1980) 25362538.[36] M. Kodama, T. Koike, N. Hoshiga, R. Machida, E. Kimura, Dalton Trans. (1984) 673678.[37] M. Kodama, E. Kimura, Dalton Trans. (1978) 10811085.[38] L.D. Pettit,H.K.J. Powell, IUPACStabilityConstantsDatabase, AcademicSoftware, Timble,

2003.[39] X. Cui, M.J. Calhorda, P.J. Costa, R. Delgado, M.G.B. Drew, V. Flix, Helv. Chim. Acta

87 (2004) 26132628.[40] L.Y. Martin, C.R. Sperati, D.H. Busch, J. Am. Chem. Soc. 99 (1977) 29682981.[41] L. Sacconi, F. Mani, A. Bencini, in: G. Wilkinson, R.D. Gillard, J.A. McCleverty (Eds.),

Comprehensive Coordination Chemistry, Pergamon Press, Oxford, 1987.[42] M.C. Rakowski, M. Rycheck, D.H. Busch, Inorg. Chem. 14 (1975) 11941200.[43] F. Neese, Diploma Thesis, University of Konstanz, Germany, June 1993[44] J. Costa, R. Delgado, M.C. Figueira, R.T. Henriques, M. Teixeira, Dalton Trans. (1997)

6573.[45] M.C. Styka, R.C. Smierciak, E.L. Blinn, R.E. DeSimone, J.V. Passarielo, Inorg. Chem.

17 (1978) 8286.[46] B.J. Hathaway, Coord. Chem. Rev. 52 (1983) 87169.[47] H.R. Gersmann, J.D. Swalen, J. Chem. Phys. 36 (1962) 32213233.[48] H. Yokoi, M. Sai, T. Isobe, S. Ohsawa, Bull. Chem. Soc. Jpn 45 (1972) 21892195.[49] P.W. Lau, W.C. Lin, J. Inorg. Nucl. Chem. 37 (1975) 23892398.[50] Y. Li, Bull. Chem. Soc. Jpn 69 (1996) 25132523.[51] A.W. Addison, M. Carpenter, L.K.-M. Lau, M. Wicholas, Inorg. Chem. 17 (1978)

15451552.[52] M.J. Maroney, N.J. Rose, Inorg. Chem. 23 (1984) 22522261.[53] P. Barbaro, C. Bianchini, G. Capannesi, L. Di Luca, F. Laschi, D. Petroni, P.A. Salvadori,

A. Vacca, F. Vizza, Dalton Trans. (2000) 23932401.[54] A.W. Addison, T.N. Rao, J. Reedjik, J. van Rijn, G.C. Verschoor, Dalton Trans. (1984)

13491356.[55] F.H. Allen, Acta Cryst. B58 (2002) 380388.[56] M.G.B. Drew, S. Hollis, P.C. Yates, Dalton Trans. (18291834).[57] R. Delgado, S. Quintino, M. Teixeira, A. Zhang, Dalton Trans. (1996) 5563.

419A.S. Fernandes et al. / Journal of Inorganic Biochemistry 105 (2011) 410419

Two macrocyclic pentaaza compounds containing pyridine evaluated as novel chelating agents in copper(II) and nickel(II) ove...IntroductionExperimental sectionGeneral proceduresReagents

Synthesis of the macrocyclesSynthesis of the macrocycle [15]pyN5Synthesis of the macrocycle [16]pyN5Synthesis of the copper(II) complex [Cu[15]pyN5](PF6)2

Potentiometric measurementsReagents and solutionsEquipment and work conditionsMeasurementsCalculation of equilibrium constants

NMR measurementsCharacterization of the macrocyclesNMR titration measurementsMagnetic moments

Spectroscopic studiesElectrochemical studiesX-ray crystallography

Results and discussionSynthesis and characterization of the macrocyclesAcidbase behaviour of the ligandsThermodynamic stability of metal complexesSpectroscopic studiesNickel(II) complexesCopper(II) complexes

Cyclic voltammetry studiesX-ray structure of the copper(II) complex

ConclusionsAcknowledgmentsSupplementary dataReferences