Electrochemistry Communications 11 (2009) 417–420

Contents lists available at ScienceDirect

Electrochemistry Communications

journal homepage: www.elsevier .com/locate /e lecom

Preferential binding specificity of silver cation to a single nucleobaseover base pairs evaluated by abasic site-containing DNA

Yong Shao a,*, Zhenjiang Niu a, Shiying Zou b

a Zhejiang Key Laboratory for Reactive Chemistry on Solid Surfaces, Institute of Physical Chemistry, Zhejiang Normal University, Yingbin Dadao 688, Jinhua, Zhejiang 321004, Chinab College of Materials and Chemical Engineering, Sichuan University of Science and Engineering, Zigong 643000, China

a r t i c l e i n f o

Article history:Received 18 November 2008Received in revised form 1 December 2008Accepted 2 December 2008Available online 10 December 2008

Keywords:DNAAbasic siteElectrochemistryInteractionSilver cationBinding

1388-2481/$ - see front matter � 2008 Elsevier B.V. Adoi:10.1016/j.elecom.2008.12.005

* Corresponding author. Tel.: +86 579 82282234; faE-mail address: yshao@zjnu.cn (Y. Shao).

a b s t r a c t

The binding specificity of silver cations to abasic (AP) site-containing DNA was electrochemically inves-tigated by comparison with the fully matched DNA without the AP site. AP site-containing DNA isdesigned in a way that only the nucleotide opposite the AP site is variable to allow for coexistence ofan unpaired nucleotide and a number of DNA base pairs. The surface of a gold electrode was modifiedby AP site-containing DNA duplex on which Ag+ binding specificity was evaluated. Electrochemical inves-tigations on the AP-DNA-modified electrodes reveal that Ag+ preferentially associates to the unpairednucleotides instead of the coexisted base pairs and shows sequence-dependant binding, especially stron-ger for purines than for pyrimidines. Additionally, the hydrogen bond pattern moieties of the unpairednucleotides should be involved in Ag+ binding evidenced by a decrease of the redox signal when introduc-ing a ligand with its hydrogen bond moiety complementary to the nucleotide deoxycytidine. This is thefirst attempt to make a comparison in one DNA molecule for metal ion binding to coexisted unpairednucleotide and DNA base pairs. The present method demonstrates an easy way for investigating bindingspecificity of heavy metal ions to AP site in the presence of coexisted DNA base pairs.

� 2008 Elsevier B.V. All rights reserved.

1. Introduction

Metal ion-DNA interaction is responsible for a wide range ofbiochemical processes such as DNA stability [1], initiation of anti-tumor agent therapeutics [2]. Metalation of DNA is an alteration forconstruction of nanomaterials, for example, silver (Ag) nanowires[3]. The interaction of Ag+ with DNA was convinced at very earlystage by potentiometric titrations and spectroscopy and at leastthree types of binding model (types I, II, and III) were suggested[4,5] (binding around base pairs for the type I, conversion of theN–H� � �N hydrogen bond of a complementary base pair to an N–Ag–N bond for the type II and binding at higher silver/DNA concen-tration ratio for the type III). DiRico et al. [6] reported that guanineinstead of cytosine or thymine in DNA was available for Ag+ bind-ing. However, binding of Ag+ to backbone phosphate groups wasexcluded. Arakawa et al. [7] suggested that Ag+ associated to guan-ine N7 and adenine N7. Nevertheless, the binding sites of nucleo-bases for Ag+ are not very clear [8]. Up to now, Ag+ bindingpreference to single nucleotide or base pairs is still not evaluated.Additionally, compared with that occurred to DNA base pairs, dif-ficulties arise when investigating the interaction of metal ion withfree single nucleobase because multiple sites [8] in nucleobase are

ll rights reserved.

x: +86 579 82282595.

available for metal ion binding and in some cases insoluble prod-ucts were formed [9].

Here, a series of abasic site (AP site)-containing DNAs are de-signed to allow for coexistence of an unpaired nucleotide and basepairs in one DNA molecule. The unpaired nucleotide is embeddedwithin the DNA helix and freely approaching each other in thepresence of Ag+ for formation of the insoluble polymeric products[9] is avoided. It is, therefore, convenient to compare the Ag+ bind-ing specificity within these two types of involved sites in this de-sign. AP site-containing DNA (AP-DNA) is also in vivo producedin cell by removal of a damaged nucleotide [10]. So the bindingspecificity of metal ions to the AP site should be clearly usefulfor evaluating unfound toxicity of metal ions in gene level. Wehave recently discovered that AP site binding of a ligand (2-ami-no-7-methyl-1,8-naphthyridine, AMND) with its hydrogen bondpattern complementary to the base opposite the AP site can facil-itate electron transfer through DNA [11]. In the present study,we describe the voltammetric studies of self-assembled monolay-ers of the AP-DNA duplexes upon Ag+ association (Scheme 1), byvariety of the nucleotide opposite the AP site and comparison withfully matched DNA without the AP site (FM-DNA).

2. Experimental

DNAs were synthesized by Shanghai Sangon BiotechnologyCo. Ltd. (Shanghai, China) and purified by HPLC. AMND was

Scheme 1. Demonstration for construction of DNA duplex-modified gold electrode for evaluation over the binding specificity of silver cation to AP-DNA and FM-DNA.

0.0 0.4 0.8

-2

0

2

0.0 0.1 0.2 0.3

-0.2

0.0

0.2

i / µ

A

E / V vs Ag / AgCl

i / µ

A

E / V vs Ag/AgCl

AP-[G] AP-[A] AP-[T] AP-[C] FM-DNA

a

b

418 Y. Shao et al. / Electrochemistry Communications 11 (2009) 417–420

synthesized according to the literature [12]. The DNA concentra-tion was measured by UV absorbance at 260 nm using extinctioncoefficient calculated by the nearest neighbor analysis [13]. TheDNA strand containing the AP site with (for electrode assembly)or without (for homogeneous solution investigation) a thiol linkerterminus at the 50 end (HS-(CH2)6-50-TCTGCGTCCAGXGCAACGCA-CAC-30, X = tetrahydrofuranyl residue as AP site model) was mixedin an equimolar amount with its complementary strand containingfour different bases opposite the AP site (50-GTGTGCGTTGCNCTG-GACGCAGA-30, N = A, C, G, or T) and annealed (for FM-DNA, X =G, N = C). Then 20 ll of 10 lM DNA duplex with the thiol linker ter-minus in 0.1 M phosphate buffer solution (PBS, pH 7.0) containing0.1 M MgCl2 with or without 50 lM AMND was dropped onto agold disk electrode (1.6 mm in diameter, BAS, USA) to form amonolayer and the electrode was kept under the saturated vaporpressure condition overnight. The electrode was again immersedinto a 1 mM 6-mercapto-1-hexanol (MCH) buffer solution for 1 h.After triply washing with 0.1 M PBS containing 1 mM EDTA (for re-moval of Mg2+) and 0.1 M PBS, the electrode modified by DNA du-plexes was immersed into a 0.1 M PBS containing desired Ag+

concentration (AgNO3, 99%, Sigma, St. Louis, USA) for 15 min. Afterwashing and immersing the electrodes in 0.1 M PBS 10 min, elec-trochemical experiments were performed using CHI 1030A (CHI,Austin, USA) at 25 �C in 0.1 M PBS free of Ag+. The gold electrodemodified by DNA duplexes with Ag+ loading, Ag/AgCl (sat. KCl,BAS), and platinum wire (Nilaco, Japan) were used as working, ref-erence, and counter electrodes, respectively. About 0.1 M PBS wasdeoxygenated by purging with purified nitrogen gas. For simplic-ity, the formed AP-DNA duplexes were referred to AP-[N] (N = A,C, G, or T) according to the bases opposite the AP site.

i / µ

A

0.0 0.4 0.8E / V vs Ag/AgCl

-8

-4

0

4

8

AP-[G] AP-[A] AP-[T] AP-[C] FM-DNA

Fig. 1. CVs of DNA-modified gold electrodes in a 0.1 M phosphate buffer solution(pH 7.0) after 5 lM (a) or 10 lM (b) AgNO3 pretreatment. Inset in (a) is the CV atFM-DNA-modified electrode partially presented for displaying the small peaks. Scanrate: (a) 0.1 V s�1; and (b) 1 V s�1.

3. Results and discussion

The binding of Ag+ to DNAs without the thiol linker terminuswas first analyzed by potentiometric titration [4] with freshly pre-pared silver wire and Ag/AgCl (sat. KCl, BAS) as indicator and refer-ence electrodes, respectively. No significant difference in potentialresponse was observed for AP-DNAs and FM-DNA in identical Ag+

concentrations (data not shown). This could be caused by the factthat the used DNA sequences are identical (for the total 22 basepairs) except the AP site and any weakly binding event will makethe specific binding at the AP site undetectable. So it is unsuitableat homogeneous aqueous solution to distinguish the Ag+ bindingpreference to the unpaired nucleotide from the coexisted basepairs. Therefore, the DNA self-assembled monolayer at electrodesurface was constructed by reaction of DNA thiol terminus withgold [14]. In order to minimize non-specific adsorption and passiv-ate occasional void gold surface between duplexes, the electrodewas again immersed into 1 mM 6-mercapto-1-hexanol (MCH)

buffer solution 1 h. As shown in Fig. 1a, after immersion of theDNA-loaded electrodes into 5 lM Ag+ buffer solution 15 min andwashing with PBS, almost symmetrical redox peaks clearly appearat AP-[N]-modified electrodes (scan rate 0.1 V s�1), while negligiblesmall peaks are observed for FM-DNA-modified electrode (the insetof Fig. 1a). The fact that no significant difference is observed for theredox peak potentials at FM-DNA- and AP-DNA-modified elec-trodes suggests that the possible difference in Ag+ binding sites

0.0 0.4 0.8

-1.0

-0.5

0.0

0.5

1.0

i / µ

A

E / V vs Ag/AgCl

AP-[C] AP-[C]-AMND

Fig. 2. CVs of AP-[C]-modified gold electrode with (red line) or without (black line)AMND loading in a 0.1 M phosphate buffer solution (pH 7.0) after 10 lM AgNO3

pretreatment. Scan rate: 0.1 V s-1. (For interpretation of the references to color inthis figure legend, the reader is referred to the web version of this article.)

Y. Shao et al. / Electrochemistry Communications 11 (2009) 417–420 419

on DNAs does not obviously change the electron transfer efficacy ofthe bound Ag+ through DNA to the electrode surface. In our condi-tions, only the type I binding occurs for FM-DNA and Ag+ bindingbefore the type I binding should occur for AP-DNA [7,9]. Therefore,it is reasonable to conclude that Ag+ electrostatic binding to phos-phate backbone is excluded [6] and the type I binding of Ag+ to basepairs [9] is weak enough to be mostly eliminated by buffer washingunder this condition. Interestingly, the binding of Ag+ to the AP-[N]is strong enough to be differentiated from the conventionally men-tioned type I binding. Additionally, the presence of an AP site inDNA would affect the duplex conformation mainly at the site [15]and unalterable conformation for the remaining base pairs as thatin FM-DNA should be expected. Therefore, the dependence of peakheights on the types of nucleotides opposite the AP site (AP-[G] > AP-[A] > AP-[T] > AP-[C] > FM-DNA) suggests that Ag+ shouldmainly bind to the nucleotide and not simply stack with the upand down Gs flanking the AP site. This point can be evidenced bythe unchanged DNA melting temperature in the presence of Ag+

(data not shown). If the oxidation peak areas in Fig. 1a were usedto estimate the binding capacities of Ag+, AP-[G]:AP-[A]:AP-[T]:AP-[C]:FM-DNA = 59:39:18:14:1 were obtained. Therefore, Ag+

binding to the unpaired nucleotide opposite the AP site must firsttake place before occurrence of Ag+ binding to the base pairs. Thedependence of electrochemical responses of Ag+ binding on thevariety of the unpaired nucleotides is surprisingly agreement withthe theoretical expectation for the interaction energies between Ag+

and nucleotides [16,17], but disagreement with the binding energycomparison for one nucleotide and its corresponding base pair [17],probably indicating that the stacking of base pairs in DNA makesthe sites for Ag+ binding sterically inaccessible in this Ag+

concentration.The sharp redox peaks observed in Fig. 1a are indicative of rapid

charge transfer [18]. According to the strongest binding case forAP-[G], the peak area under reduction peak is about 2.4 � 10�7 C.If assuming 1:1 binding of Ag+ to AP-[G], hypothesizing the DNAcoverage of about 41 pmol cm�2 [19] and roughly considering onlythe electrode geometric area, the theoretic charge needed forreduction of the associated Ag+ is about 7.9 � 10�8 C, just one thirdof the experimentally observed peak area. This observation showsthat more than one Ag+ are involved in the interactions with theguanine opposite the AP site, coincident with multiplicity of Ag+

binding sites to guanine [8]. Fig. 1b shows the results obtained at1 V s�1 by electrode treatment in 10 lM Ag+ solution. The bindingspecificity at the AP-[N]-modified electrodes is also kept at thiscondition. Additionally, increases of the potential width at halfheight of the peak current (DEp,1/2) in all cases are observed byincreasing the scan rate to 1 V s�1 from 0.1 V s�1 in all Ag+ concen-tration (Fig. 1b). This phenomenon should reflect the dominantrole of diffusion of counter-ions to the electrode surface [18] uponAg+ reduction that will affect the electrode process at high scanrate.

The clear redox responses at the AP-[N]-modified electrodes rel-ative to the small signal at FM-DNA-modified electrode predict thatAg+ should, at least partially, bind to the hydrogen bond patternmoiety of the nucleotide opposite the AP site. Previously, a relatedresearch [20] was measured by thermodynamic stability and fluo-rescence spectroscopy and proved that AMND binding occurredthrough hydrogen bonding to a target nucleobase opposite the APsite with selectivity in the order of C > T > G > A. This artificiallybase-pairing process was also believed to be accompanied byAMND stacking with nucleobases flanking the AP site evidencedby improved electron transfer through DNA [11]. Additionally,AMND is electroinactive during the potential range employed inthis study. Here AMND is also used to verify that the hydrogen bondpattern moiety of the nucleotide opposite the AP site is involved inAg+ binding. Before assembly on the gold electrode, AP-[C] duplex

was saturated by AMND beforehand to allow for AMND bindingat the AP site. After modification of the electrode by AMND-loadedAP-[C] duplex, same procedures were employed as mentioned inSection 2. As shown in Fig. 2, significantly decreased redox peaksare observed for the AMND treated electrode, compared with thatoccurred at the electrode without AMND loading. Because AMNDbinds only to the nucleotide opposite the AP site by hydrogen bondinteraction and not affect the remaining intact DNA structure [20],the occupied hydrogen bond pattern moiety of cytosine oppositethe AP site by AMND would make this moiety unavailable for Ag+

binding. Therefore, a decreased signal is expected. The peak poten-tial shift is also observed after AMND treatment. This fact demon-strates the presence of possible lateral interaction for theassembled duplexes during redox process, which should be depen-dent on Ag+ concentration in the monolayer. Kargov et al. also ob-served DNA lateral interaction when Ag+ associated to DNA thatwas immobilized in polyacrylamide gel [21].

4. Conclusions

Abasic site-containing DNA is designed to allow for coexistenceof an unpaired nucleotide and a number of DNA base pairs. It isvery convenient to investigate the binding specificity of metal ionson these two types of binding sites by electrochemistry based onelectrode assembly. Electrochemical measurements reveal thatAg+ preferentially associates to the unpaired nucleotides insteadof the coexisted base pairs and shows sequence-dependant bindingon the AP-[N]-modified electrodes, especially stronger for purinesthan for pyrimidines opposite the AP site. Additionally, the hydro-gen bond pattern moiety of the unpaired nucleotide should be in-volved in the binding of Ag+. This is the first attempt to make acomparison in one DNA molecule for the metal ion binding tothe coexisted unpaired nucleotide and DNA base pairs. Thisremarkable sensitivity signifies the suitability of the new schemefor monitoring the DNA binding specificity of heavy metal ionsand developing new aptamer for sensor design.

Acknowledgment

This work was supported by ZJNU Scientific Research Founda-tion for the Returned Overseas Chinese Scholars (Grant No.ZC304008066).

420 Y. Shao et al. / Electrochemistry Communications 11 (2009) 417–420

References

[1] M.T. Record, C.F. Anderson, T.M. Lohman, Q. Rev. Biophys. 11 (1978) 103.[2] B. Rosenberg, L. VanCamp, J.E. Trosko, V.H. Mansour, Nature 222 (1969) 385.[3] E. Braun, Y. Eichen, U. Sivan, G. Ben-Yoseph, Nature 391 (1998) 775.[4] M. Daune, C.A. Dekker, H.K. Schachman, Biopolymers 4 (1966) 51.[5] R.H. Jensen, N. Davidson, Biopolymers 4 (1966) 17.[6] D.E. DiRico Jr., P.B. Keller, K.A. Hartman, Nucleic Acids Res. 13 (1985) 251.[7] H. Arakawa, J.F. Neault, H.A. Tajmir-Riahi, Biophys. J. 81 (2001) 1580.[8] B. Lippert, Coord. Chem. Rev. 200–202 (2000) 487.[9] R.M. Izatt, J.J. Christensen, J.H. Rytting, Chem. Rev. 71 (1971) 439.

[10] J. Lhomme, J.F. Constant, M. Demeunynck, Biopolymers 52 (1999) 65.[11] Y. Shao, K. Morita, Q. Dai, S. Nishizawa, N. Teramae, Electrochem. Commun. 10

(2008) 438.

[12] E.V. Brown, J. Org. Chem. 30 (1965) 1607.[13] C.R. Cantor, M.M. Warshaw, H. Shapiro, Biopolymers 9 (1970) 1059.[14] D. Chen, J.H. Li, Surf. Sci. Rep. 61 (2006) 445.[15] M. Lukin, C. Santos, Chem. Rev. 106 (2006) 607. and references therein.[16] J.V. Burda, J. Šponer, P. Hobza, J. Phys. Chem. 100 (1996) 7250.[17] J.V. Burda, J. Šponer, J. Leszczynski, P. Hobza, J. Phys. Chem. B 101 (1997)

9670.[18] I.O. K’Owino, R. Agarwal, O.A. Sadik, Langmuir 19 (2003) 4344.[19] S.O. Kelley, J.K. Barton, Bioconjugate Chem. 8 (1997) 31.[20] K. Yoshimoto, S. Nishizawa, M. Minagawa, N. Teramae, J. Am. Chem. Soc. 125

(2003) 8982.[21] S.I. Kargov, N.I. Korolev, O.B. Stanislavski, I.A. Kuznetsov, Mol. Biol. (Mosk) 20

(1986) 1499.