Electrochemical Detection of Single-Nucleotide Mismatches

The detection of a single-nucleotide mismatch in un- labeled duplex DNA by electrochemical methods is pre- sented. Impedance spectroscopy is used to ...
0 downloads 0 Views 179KB Size
Download PDF
Anal. Chem. 2004, 76, 4059-4065

Electrochemical Detection of Single-Nucleotide Mismatches: Application of M-DNA Yi-Tao Long,†,‡ Chen-Zhong Li,†,‡ Todd C. Sutherland,†,‡ Heinz-Bernhard Kraatz,*,† and Jeremy S. Lee*,‡

Department of Chemistry, University of Saskatchewan, 110 Science Place, Saskatoon, S7N 5C9 Saskatchewan, Canada, and Department of Biochemistry, University of Saskatchewan, 107 Wiggins Road, Saskatoon, S7N 5E5 Saskatchewan, Canada

The detection of a single-nucleotide mismatch in unlabeled duplex DNA by electrochemical methods is presented. Impedance spectroscopy is used to characterize a perfect duplex monolayer and three DNA monolayers differing in the position of the mismatch. The monolayers were studied as B-DNA (normal duplex DNA) and after conversion to M-DNA (a metalated duplex). Modeling of the impedance data to an equivalent circuit provides parameters that are useful in discriminating the four monolayer configurations. The resistance to charge transfer, RCT, was lower for all duplexes after conversion to M-DNA. Contrary to expectations, RCT was also found to decrease for duplexes containing a mismatch. However, RCT was found to be diagnostic for mismatch detection. In particular, the difference in RCT between B- and M-DNA (∆RCT) decreased from 190(22) Ω‚cm2 for a perfectly matched duplex to 95(20), 30(20), and 85(20) Ω‚cm2 for a mismatch at the top (distal), middle, and bottom (proximal) positions of the monolayer with respect to the gold surface. Further, a method to form loosely packed single-stranded (ss)-DNA monolayers by duplex dehybridization that is able to rehybridize to target strands is presented. Rehybridization efficiencies were in the range of 40-70%. Under incomplete hybridization conditions, the RCT was the same for matched and mismatched duplexes under B-DNA conditions. However, ∆RCT between B- and M-DNA, under incomplete hybridization, still provided a distinction. The ∆RCT for a perfect duplex was 76(12) Ω‚cm2, whereas a mismatch in the middle of the sequence yielded a ∆RCT value of 30(15) Ω‚cm2. The detection limit was measured and the impedance methodology reliably detected single DNA base pair mismatches at concentrations as low as 100 pM. DNA biosensors provide a powerful means of recognizing specific DNA sequences. The design of sequence-selective DNA biosensors, via hybridization, has received much attention in recent years.1-9 In particular, methods for the rapid identification * Authors to whom correspondence should be addressed. E-mail: kraatz@ sask.usask.ca (H.-B. Kraatz); [email protected] (J. S. Lee). † Department of Chemistry. ‡ Department of Biochemistry. (1) Willner, I. Science 2002, 298, 2407-2408. 10.1021/ac049482d CCC: $27.50 Published on Web 05/28/2004

© 2004 American Chemical Society

of base mutations or single nucleotide polymorphisms (SNPs) would prove useful for the diagnosis of many genetic diseases and in clinical pharmacology.10-15 Two classes of DNA mutation biosensors are commonly employed. The first requires covalent modification of the target DNA strands with groups including redox probes,16-19 fluorescent dyes, or radioactive markers.2,20,21 The second class employs reporter molecules that are not covalently bound to the DNA, such as biomolecular beacons22-24 and electrochemically active intercalators.25-29 The latter methods (2) Wang, D. G.; Fan, J. B.; Siao, C. J.; Berno, A.; Young, P.; Sapolsky, R.; Ghandour, G.; Perkins, N.; Winchester, E.; Spencer, J.; Kruglyak, L.; Stein, L.; Hsie, L.; Topaloglou, T.; Hubbell, E.; Robinson, E.; Mittmann, M.; Morris, M. S.; Shen, N. P.; Kilburn, D.; Rioux, J.; Nusbaum, C.; Rozen, S.; Hudson, T. J.; Lipshutz, R.; Chee, M.; Lander, E. S. Science 1998, 280, 1077-1082. (3) Fritz, J.; Cooper, E. B.; Gaudet, S.; Sorger, P. K.; Manalis, S. R. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 14142-14146. (4) Heaton, R. J.; Peterson, A. W.; Georgiadis, R. M. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 3701-3704. (5) Richter, J. Physica E 2003, 16, 157-173. (6) Venter, J.; Adams, M.; Sutton, G.; Kerlavage, A.; Smith, H.; Hunkapiller, M. Science 1998, 280, 1540-1542. (7) Bontidean, I.; Kumar, A.; Csoeregi, E.; Galaev, I. Y.; Mattiasson, B. Angew. Chem., Int. Ed. 2001, 40, 2676-2678. (8) Liu, X.; Tan, W. Anal. Chem. 1999, 71, 5054-5059. (9) Ferguson, J. A.; Boles, T. C.; Adams, C. P.; Walt, D. R. Nat. Biotechnol. 1996, 14, 1681-1684. (10) McCarthy, J. J.; Hilfiker, R. Nat. Biotechnol 2000, 18, 505-508. (11) Brookes, A. Gene 1999, 234, 177-186. (12) Wang, J. Chem.sEur. J. 1999, 5, 1681-1685. (13) Palecek, E.; Fojta, M.; Jelen, F.; Vetterl, V. In Encyclopedia of Electrochemistry; Bard, A. J., Stratmann, M., Eds.; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2002; Vol. 9, pp 365-429. (14) Aoki, H.; Buhlmann, P.; Umezawa, Y. Electroanalysis 2000, 12, 1272-1276. (15) DeWitt, N. Nat. Biotechnol. 2000, 18, 1027. (16) Yu, C. J.; Wan, Y. J.; Yowanto, H.; Li, J.; Tao, C. L.; James, M. D.; Tan, C. L.; Blackburn, G. F.; Meade, T. J. J. Am. Chem. Soc. 2001, 123, 1115511161. (17) Umek, R. M.; Lin, S. W.; Vielmetter, J.; Terbrueggen, R., H.; Irvine, B.; Yu, C. J.; Kayyem, J. F.; Yowanto, H.; Blackburn, G. F.; Farkas, D. H.; Chen, Y.-P. J. Mol. Diag. 2001, 3, 74-84. (18) Huang, T. J.; Liu, M.; Knight, L. D.; Grody, W. W.; Miller, J. F.; Ho, C.-M. Nucleic Acids Res. 2002, 30, e55. (19) Mao, Y.-D.; Luo, C.-X.; Ouyang, Q. Nucleic Acids Res. 2003, 13, DOI: 10.1093/nar/gng1108. (20) Hacia, J. G. Nat. Genet. 1999, 21, 42-47. (21) Lindblad-Toh, K.; Winchester, E.; Daly, M. J.; Wang, D. G.; Hirschhorn, J. N.; Laviolette, J. P.; Ardlie, K.; Reich, D. E.; Robinson, E.; Sklar, P.; Shah, N.; Thomas, D.; Fan, J. B.; Gingeras, T.; Warrington, J.; Patil, N.; Hudson, T. J.; Lander, E. S. Nat. Genet. 2000, 24, 381-386. (22) Tyagi, S.; Bratu, D. P.; Kramer, F. R. Nat. Biotechnol. 1998, 16, 49-53. (23) Bernacchi, S.; Mely, Y. Nucleic Acids Res. 2001, 29, e62/61-e62/68. (24) Tan, W.; Fang, X.; Li, J.; Liu, X. Chem.sA Eur. J. 2000, 6, 1107-1111.

Analytical Chemistry, Vol. 76, No. 14, July 15, 2004 4059

have the potential advantage that the target DNA does not require any post-isolational modification, thereby reducing the number of manipulations. Electrochemical methods for SNP detection are attractive because they allow a direct electrical readout, which again reduces the complexity of the assay.13,29-31 For example, Barton and colleagues have developed a chronoamperometric technique that is dependent on the intercalator Methylene Blue (MB+), which acts as a redox mediator between the DNA duplex attached to a gold electrode and the redox probe [Fe(CN)6]3-/4in solution. Duplex DNA containing a single mismatch has a lower rate of charge transport and, thus, can be distinguished from a perfect duplex.32,33 Heller and co-workers have used cyclic voltammetry (CV) to detect DNA mismatches at Au electrodes using a redox-active polymer adjacent to the Au surface and the covalent attachment of an enzyme to the target DNA sequence.34,35 The CV provides a readout of the enzyme-amplified signal and is able to discriminate a DNA mismatch from a match. Along the same lines but using impedance spectroscopy, Willner and co-workers were able to detect mismatches in B-DNA (normal doublestranded-(ds)-DNA) by enzymatic amplification.36,37 Again, this uniquely sensitive technique relies on the covalent modification of the DNA. The weakness of such techniques is that they are very dependent on hybridization efficiencies. Recently, another method was proposed38,39 to detect DNA hybridization and potentially DNA-mismatch detection. One end of a ferrocene-(Fc)DNA construct containing a self-complementary region was attached to an Au surface. Upon monolayer formation, the singlestranded-(ss)-DNA induces a hairpin, forcing the Fc group to be positioned close to the electrode surface. Hybridization with a complementary strand released the hairpin structure to form B-DNA, resulting in the Fc moiety increasing the distance to the electrode surface. Complementary sequences result in a change in the redox potential and kinetics of electron transfer of the Fc group, thus enabling an effective electrochemical hybridization sensor. The method was recently tested38 and proved very effective as a hydridization sensor. Furthermore, a recent study19 has successfully used a similar ss-DNA hairpin system to determine single base pair mismatches but again the hybridization efficiency (25) Kelley, S. O.; Jackson, N. M.; Hill, M. G.; Barton, J. K. Angew. Chem., Int. Ed. 1999, 38, 941-945. (26) Yamashita, K.; Takagi, M.; Kondo, H.; Takenaka, S. Anal. Biochem. 2002, 306, 188-196. (27) Hanafi-Bagby, D.; Piunno, P. A. E.; Wust, C. C.; Krull, U. J. Anal. Chim. Acta 2000, 411, 19-30. (28) Takenaka, S.; Yamashita, K.; Takagi, M.; Uto, Y.; Kondo, H. Anal. Chem. 2000, 72, 1334-1341. (29) Boon, E. M.; Kisko, J. L.; Barton, J. K. Methods Enzymol. 2002, 353, 506522. (30) Wang, J. Anal. Chim. Acta 2002, 469, 63-71. (31) Drummond, T. G.; Hill, M. G.; Barton, J. K. Nat. Biotechnol. 2003, 21, 11921199. (32) Boon, E. M.; Ceres, D. M.; Drummond, T. G.; Hill, M. G.; Barton, J. K. Nat. Biotechnol. 2000, 18, 1096-1100. (33) Kelley, S. O.; Boon, E. M.; Barton, J. K.; Jackson, N. M.; Hill, M. G. Nucleic Acids Res. 1999, 27, 4830-4837. (34) Hartwich, G.; Caruana, D. J.; de Lumley-Woodyear, T.; Wu, Y.; Campbell, C. N.; Heller, A. J. Am. Chem. Soc. 1999, 121, 10803-10812. (35) Caruana, D. J.; Heller, A. J. Am. Chem. Soc. 1999, 121, 769-774. (36) Patolsky, F.; Lichtenstein, A.; Willner, I. Nat. Biotechnol. 2001, 19, 253257. (37) Patolsky, F.; Weizmann, Y.; Katz, E.; Willner, I. Angew. Chem., Int. Ed. 2003, 42, 2372-2376. (38) Fan, C.-H.; Plaxco, K. W.; Heeger, A. J. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 9134-9137. (39) Hu, X. Ph.D. Dissertation, Duke University: Durham, NC, 2001.

4060

Analytical Chemistry, Vol. 76, No. 14, July 15, 2004

can lead to erroneous conclusions. In summary, it is highly desirable to develop an approach to single-nucleotide mismatch detection that does not require DNA labeling and does not depend on hybridization efficiency. Previously, we studied, by electrochemical impedance spectroscopy (EIS)40 and cyclic voltammetry,41 the electrochemical behavior of DNA duplexes attached to gold electrodes under Band M-DNA conditions. M-DNA is a metalated form of DNA which forms at pH 8.5 with Zn2+;42-44 see refs 45-47 for recent reviews. In EIS, the impedance of an electrode undergoing heterogeneous electron transfer through a self-assembled monolayer can be described using an equivalent electrical circuit consisting of resistance and capacitance elements, such as RS (the solution resistance), RCT (the charge-transfer resistance), CPE (the constantphase element), and a mass transfer element W (Warburg impedance). M-DNA formation causes significant changes in the electronic properties of the DNA that are readily detected by EIS. Nyquist plots show large differences between B- and M-DNA, which when analyzed by a modified Randle’s circuit demonstrate that the charge-transfer resistance through the DNA is decreased for M-DNA. The charge-transfer resistance increases with increasing duplex length for M-DNA, but in each case, the resistance for M-DNA is lower than its corresponding B-DNA construct. In addition, electron transfer is faster in M-DNA as compared to that in B-DNA.40,41 We hypothesized that any disruption of the electron transfer pathway in DNA, caused by bulges and kinks due to a mismatched sequence, would have an effect on the electronic and electron transfer properties of the DNA construct. As was previously reported, mismatches result in a larger motion of the DNA helix which will influence the electronic coupling between the base pairs and lead to a decrease in electron transfer rate.29 Single-nucleotide mismatches in B-DNA will affect the local environment of adjacent bases. This effect is dependent on the identity of the base and the position of the mismatch in the helix.48,49 EIS was deemed to be the most suitable technique to evaluate the effects of mismatches on the electronic and electron-transfer properties of DNA and has been used before in enzyme-amplified methods for mismatch detection.36 In this contribution, we demonstrate that the presence and position of a single-nucleotide mismatch in unlabeled M- and B-DNA monolayers give rise to characteristic changes in the impedance spectrum, which can be exploited for the detection of mismatches even under conditions of incomplete hybridization with a detection limit of 100 pM. This (40) Long, Y.-T.; Li, C.-Z.; Kraatz, H.-B.; Lee, J. S. Biophys. J. 2003, 84, 32183225. (41) Li, C.-Z.; Long, Y.-T.; Kraatz, H.-B.; Lee, J. S. J. Phys. Chem. B 2003, 107, 2291-2296. (42) Lee, J. S.; Latimer, L. J. P.; Reid, R. S. Biochem. Cell Biol. 1993, 71, 162168. (43) Aich, P.; Labiuk, S. L.; Tari, L. W.; Delbaere, L. J. T.; Roesler, W. J.; Falk, K. J.; Steer, R. P.; Lee, J. S. J. Mol. Biol. 1999, 294, 477-485. (44) Rakitin, A.; Aich, P.; Papadopoulos, C.; Kobzar, Y.; Vedeneev, A. S.; Lee, J. S.; Xu, J. M. Phys. Rev. Lett. 2001, 86, 3670-3673. (45) Carell, T.; Behrens, C.; Gierlich, J. Org. Biomol. Chem. 2003, 1, 22212228. (46) Robertson, N.; McGowan, C. A. Chem. Soc. Rev. 2003, 32, 96-103. (47) Bhalla, V.; Bajpai, R. P.; Bharadwaj, L. M. EMBO Rep. 2003, 4, 442-445. (48) Aboul-Ela, F.; Koh, D.; Tinoco, I. J. Nucleic Acids Res. 1985, 13, 48114824. (49) Ikuta, S.; Takagi, K.; Wallace, R. B.; Itakura, K. Nucleic Acids Res. 1987, 15, 797-811.

Chart 1. List of DNA Sequences Used for Monolayer Film Preparationa

a

Mismatched base pairs are indicated by the blocked characters in the sequence.

represents a significant improvement over other methods and demonstrates the utility of M-DNA. EXPERIMENTAL SECTION Materials. The 5′-disulfide-labeled and unlabeled oligonucleotide strands were synthesized by standard phosphoamidate solidphase DNA synthesis using a fully automated DNA synthesizer, purified by reversed-phase HPLC and then characterized by electrospray ionization mass spectrometry (see Supporting Information).40 The DNA sequences and position of the mismatches are shown in Chart 1. Monolayer Preparation. The freshly cleaned gold electrodes (BAS, 1.6-mm diameter) were incubated in 0.05-mM ss-DNA or ds-DNA B-DNA, 20-mM Tris-ClO4 buffer solution (pH 8.6) for 5 days. Then the electrodes were washed with Tris-ClO4 buffer and mounted into an electrochemical cell. Dehybridization and regeneration of the single-stranded probe electrode was achieved by denaturing the duplex DNA by soaking the electrode in a heated (60 °C) water/EtOH (60:40) bath for 10 min and then rinsing in room temperature 20-mM Tris-ClO4 buffer. Reproducible behavior was found for repeated measurements on different electrodes. Rehybridization was performed by exposing the ssDNA self-assembled monolayer (SAM) to SSC buffer (300-mM NaCl, 30-mM sodium citrate, pH 7) heated to 37 °C in the presence of target DNA for 10 min and then was allowed to cool to room temperature for an additional 3 h. B-DNA was converted to M-DNA by the addition of 0.4-mM Zn(ClO4)2‚6H2O for 2 h at pH 8.6.40,41 The formation of the monolayer was assessed by standard blocking studies with [Fe(CN)6]3-/4-, X-ray photoelectron spectroscopy (XPS), and EIS, as described previously.40 The blocking studies showed a decrease in peak current attributed to the reduced diffusion of the redox probe to the Au surface. The XPS data show the presence of a an Au-thiolate bond and a thickness of 44 Å for a 1:2 monolayer.40 Electrochemical Measurements. A conventional threeelectrode cell was used. All experiments were conducted at room temperature (22 °C). The cell was enclosed in a grounded Faraday cage. The reference electrode was constructed by sealing a Ag/ AgCl wire into a glass tube with a solution of 3 m KCl that was capped with a Vycor tip. The counter electrode was a platinum wire. Impedance spectra were measured using an EG&G 1025

frequency response analyzer interfaced to an EG&G 283 potentiostat/galvanostat. The ac voltage amplitude was 5 mV and the voltage frequencies used for EIS measurements ranged from 100 kHz to 100 mHz. The applied potential was 250 mV vs Ag/AgCl (formal potential, E°′, of the redox probe [Fe(CN)6]3-/4-. All measurements were repeated a minimum of five times with separate electrodes to obtain statistically meaningful results. RESULTS AND DISCUSSION Monolayers of fully matched B-DNA on gold were prepared from the oligonucleotide 1 and its fully matched complementary strand 2. The properties of the resulting 1:2 B-DNA surface compares well with those described before.40,41 To evaluate the effect of mismatches by EIS, we prepared three types of mismatched monolayers, each containing a single pyrimidine‚ pyrimidine mismatch in the complementary strand. Complementary mismatched strand 3 contains a mismatch in the second top base pair, resulting in a mismatch distal to the electrode surface. Complementary mismatched strand 4 contains a T instead of a G in position 11, giving a monolayer with the mismatch in the middle of the duplex. Complementary mismatched strand 5 possesses a C instead of an A in position 19, resulting in a mismatch proximal to the electrode surface. Mismatched B-DNA monolayers of 1:3, 1:4, and 1:5 were prepared in an analogous manner (see Figure 1). Impedance measurements were carried out on all monolayers in 20-mM Tris-ClO4 (pH 8.6) in the presence of a 4 mM [Fe(CN)6]3-/4- (1:1) mixture, as the solution-based redox probe. The B-DNA monolayers were then converted to M-DNA monolayers by the addition of 0.4 mM Zn2+ at pH 8.6 as described before.40,41 The impedance measurements were repeated under M-DNA conditions for all four monolayers. Control experiments were performed with longer incubation times with Zn2+ and repeated EIS measurements. Neither procedure produced changes in the impedance spectra, demonstrating that the DNA was not significantly damaged, for example, by oxidation due to electron transfer.25,31 Typical impedance spectra, in the form of Nyquist plots, for B-DNA and M-DNA monolayers of a perfectly matched duplex (1:2) and a duplex containing a mismatch in the middle of the helix (1:4) are shown in Figure 2. Each point represents a value of Zim and Zre measured at a particular ac frequency. The spectra show a lower impedance for M-DNA than for B-DNA, as would be expected from previous observations.40-44 Analytical Chemistry, Vol. 76, No. 14, July 15, 2004

4061

Figure 1. Proposed structures of the monolayers prepared in this study. (a) 1:2, the complementary ds-DNA monolayer. (b) 1:3, mismatch distal to the Au surface of the ds-DNA monolayer. (c) 1:4, mismatch in the middle of the ds-DNA monolayer. (d) 1:5, mismatch proximal to the Au surface of the ds-DNA monolayer. The arrow indicates the position of the mismatched base pair.

Figure 2. Nyquist plots (-Zim vs Zre) of the 20 base pair complementary B-DNA (O), middle mismatch B-DNA (0), complementary M-DNA (b), and middle mismatch M-DNA (9) assembled on gold in a 4 mM [Fe(CN)6]3-/4- (1:1) mixture as the redox probe in 20 mM Tris-ClO4 and 20 mM NaClO4 solution. Applied potential of 250 mV versus Ag/AgCl. [Zn2+] ) 0.4 mM; pH 8.6. In all cases the measured data points are shown as symbols with the calculated fit to the equivalent circuit as solid lines. Inset: The experimental data were fit to the equivalent circuit. RS, solution resistance; RX, monolayer pinhole/defect resistance; RCT, charge transfer resistance; CPE, constant phase element; W, Warburg impedance.

More importantly, the presence of a mismatch in the DNA duplex decreases the impedance of B-DNA while increasing the impedance of M-DNA. To provide a rationale for this behavior, the impedance spectra of all DNA films were analyzed with a modified Randles equivalent circuit.40 The circuit drawing is shown in Figure 2. The same model was used to fit all monolayers described here. The fit of the equivalence circuit to the experimental values is given as a solid line. This treatment allows the interpretation of the impedance data in terms of electronic circuit components, 4062 Analytical Chemistry, Vol. 76, No. 14, July 15, 2004

which are listed for all monolayers in Table 1. The equivalent circuit contains five elements that are described below. A solution resistance term, RS, remains constant at 5-6 Ω‚ cm2, as would be expected for measurements under identical conditions of supporting electrolyte concentration and temperature. The circuit contains a constant phase element (CPE) acting as a nonideal capacitor, which is commonly used instead of a capacitor to account for inhomogeneity on the electrode surface.51 Here, the CPE will be interpreted as a capacitor as described elsewhere in situations where the exponential modifier is greater than 0.9.50 This is the case for all monolayers presented in this paper. Monolayer composition and thickness are contributing factors to the CPE. The magnitude of the CPE for films of the matched duplex 1:2 and the two top and bottom mismatched duplexes 1:3 and 1:5 are in the range of 10-25 µF‚cm-2. However, for films of 1:4, B- and M-DNA containing the middle mismatch, a significantly higher capacitance of about 40(2) µF‚cm-2 was observed. A possible interpretation is a change in the monolayer thickness for films of 1:4 due to kinking of the helix, which is caused by the mismatch in the center of the DNA sequence. This kink will cause the duplex to bend toward the surface, resulting in a decrease in the film thickness. The top and bottom mismatch containing films, 1:3 and 1:5, respectively, yield CPE values that also indicate slight variations in the thickness of the monolayer due to fraying of the strands for both films. However, the effects of fraying for the top and bottom mismatch films are opposite. Fraying in the top of the helix will result in a better interaction between individual nonpaired base pairs and the solvent, thus leading to a thicker monolayer. In the case of the mismatch being proximal to the electrode surface, as for film 1:5, fraying is expected to enhance flexibility of the DNA-linker region, which may result in a lower film thickness by compaction. Effective monolayer thickness can be calculated assuming the CPE is an ideal capacitor and the permittivity of the DNA monolayer is 1 F‚m-1. Under these assumptions, the effective thicknesses for B-DNA monolayers of 1:2, 1:3, 1:4, and 1:5 are 120, 140, 40, and 70 Å, respectively. The RX component of the equivalence circuit can be attributed to pinholes in the monolayer structure. The value of RX is similar for each of the B-DNA monolayers; indicating the number and size of the pinholes does not change between monolayers. However, RX tends to decrease upon conversion to M-DNA. This behavior can be attributed to M-DNA having a slightly more compact structure than B-DNA due to the divalent metal ions reducing the repulsion between the phosphate backbone residues of adjacent helices.40-43 The Warburg impedance element, W, is dependent on the rate of diffusion of the [Fe(CN)6]3-/4- redox probe. The Warburg impedance is smallest for the perfect duplex in the B-DNA conformation, suggesting that this is the most ordered monolayer, which offers the least access of the solution electrophore through the DNA monolayer. The charge-transfer resistance term, RCT, comprises resistance terms resulting from (a) transfer of the electron from the [Fe(CN)6]3-/4- redox probe to the DNA monolayer, (b) the resistance to charge transfer between the base pairs of the DNA helix, and (c) from the helix to the surface of the gold electrode. (50) Creager, S. E.; Wooster, T. T. Anal. Chem. 1998, 70, 4257-4263. (51) Dijksma, M.; Boukamp, B. A.; Kamp, B.; van Bennekom, W. P. Langmuir 2002, 18, 3105-3112.

Table 1. Circuit Element Values for the Complementary DNA Monolayer and the Series of Mismatch DNA Monolayersa 1:2 circuit element /Ω‚cm2

RS RX/Ω‚cm2 CPEb/µF‚cm-2 RCT/Ω‚cm2 ∆RCT/Ω‚cm2 W/10-3 Ω‚s-1/2‚cm2

B-DNA

1:3 M-DNA

5.8(0.5) 6.0(0.6) 300(21) 245(18) 15.0(0.4) 15.7(0.5) 390(20) 200(10) 190(22) 1.5(0.3) 3.9 (0.4)

B-DNA

1:4 M-DNA

4.9(0.9) 4.8(0.7) 357(19) 319(16) 12.8(0.3) 10.9(0.2) 299(15) 204(12) 95(19) 3.8(0.6) 3.4(0.2)

B-DNA

1:5 M-DNA

6.5(0.7) 5.8(0.6) 351(17) 323(16) 42.1(0.9) 38.0(0.6) 258(14) 228(11) 30(18) 7.8(0.4) 8.1(0.5)

B-DNA

M-DNA

5.3(0.7) 6.1(0.8) 312(15) 255(12) 25.0(0.3) 23.1(0.2) 317(18) 232(10) 85(20) 3.0(0.3) 3.9(0.2)

a The values in parentheses represent the standard deviations from several electrode measurements (n g 5) not the non-linear curve fitting errors. b CPE and associated units are interpreted as a capacitor with an exponential modifier >0.9.50

As expected, for all monolayers, RCT is lower for M-DNA than for B-DNA. More importantly, RCT allows the discrimination between a single nucleotide mismatch and a perfectly matched DNA film. The presence of a mismatch causes a decrease in RCT for all films containing mismatches. This is contrary to expectations since the electron transfer (ET) kinetics through an unstacked region is retarded.25,32,33 Therefore, thickness and/or disorder in the monolayer, as signified by the changes in the CPE term, must dominate rates of ET rather than simple RCT. As far as mismatch detection is concerned, the evaluation of the difference in charge-transfer resistance, ∆RCT, between B- and M-DNA for a given film gives excellent discrimination between a perfect duplex and one containing a single mismatch at either the top or the middle positions of the duplex. Table 1 lists the ∆RCT for all films. ∆RCT for the perfectly matched duplex film 1:2 is 190 (22) Ω‚cm2 whereas for the mismatched films, ∆RCT is significantly smaller. Interestingly, ∆RCT for the top mismatch containing the film of 1:3 and the bottom mismatch (1:5) are similar (95(19) Ω‚cm2 for 1:3 and 85(20) Ω‚cm2 for 1:5). ∆RCT for the duplex containing the middle mismatch is much lower (30(18) Ω‚cm2 for 1:4). The use of ∆RCT was attractive from an application perspective because different electrode morphologies can yield different impedances but the comparative impedance measurements between B- and M-DNA are reproducible. To assess the ability for mismatch determination under nonideal conditions, we investigated the effect of rehybridization. In this format, the DNA probe sequence is washed across a ssDNA monolayer which may result in differences in hybridization. The direct formation of a ss-DNA monolayer yields a film in which the DNA strands are densely packed and may interfere with the binding of the complementary strand.52,53 Therefore, we decided to form a ds-DNA film and then dehybridize to a more loosely packed ss-DNA monolayer. In this way, rehybridization efficiencies for the target DNA in the range of 40-70% can be achieved.53,54 Figure 3a outlines the experimental procedure. Washing of a dsDNA film with a hot (60 °C) water/EtOH (60:40) bath followed by rinsing in room-temperature Tris-ClO4 buffer results in dehybridization and formation of a ss-DNA film consisting of DNA strand 1. This film is then exposed to solutions of complementary target ss-DNA and allowed to hybridize for 3 h. The heating could (52) Yang, M.-S.; Yau, H. C. M.; Chan, H. L. Langmuir 1998, 14, 6121-6129. (53) Peterson, A. W.; Wolf, L. K.; Georgiadis, R. M. J. Am. Chem. Soc. 2002, 124, 14601-14607. (54) Peterlinz, K. A.; Georgiadis., R. M. J. Am. Chem. Soc. 1997, 119, 34013402.

Figure 3. (a) Hybridization-dehybridization procedure. (i) Soaked in a water:EtOH (60:40) bath at 60 °C for 10 min and then rinsed with room temperature 20 mM Tris-ClO4 buffer. (ii) Add target ssDNA in SSC buffer and allow duplex formation to occur for 10 min at 37 °C followed by 3 h at room temperature. (b) Nyquist plot of fully hybridized “ideal” monolayer of 1:2 construct (O), ss-DNA monolayer of 1 (0) after dehybridization procedure, and rehybridized ds-DNA film of 1:2 (b) following the rehybridization procedure. The impedance spectra of the rehybridized film is different compared to that of the “ideal” 1:2 films, indicating the heterogeneity of the monolayer as a result of incomplete hybridization.

have deleterious effects on the monolayer; however, this is ruled out because the RX component has remained the same or increased, indicating that no new pinholes or defect sites were created. We expected from the results of Table 1 that the combination using the middle mismatch, strand 4, would give the smallest difference in ∆RCT under these conditions. Upon reformation of the ds-DNA film using 2 or 4 following the dehybridization-rehybridization procedure, the impedance signal does not return to the values for a perfect ds-DNA monolayer as shown by Figure 3b, indicating that the resulting film most likely consists of ds- and ss-DNA. Importantly, despite incomplete rehybridization the presence of a mismatch can still be detected as shown by the impedance spectra in Figure 4. After rehybridization of the ss-DNA film with a matching complementary strand, 2, and one containing a single mismatch in the middle of the strand, 4, all spectra were fit to the same equivalent circuit described above and the electronic circuit parameters are shown Analytical Chemistry, Vol. 76, No. 14, July 15, 2004

4063

Figure 4. Nyquist plots (-Zim vs Zre) of the rehybridized 20 base pair complementary B-DNA (O), middle mismatch B-DNA (0), complementary M-DNA (b), and middle mismatch M-DNA (9) assembled on gold with 4 mM Fe(CN)63-/4- (1:1) mixture as the redox probe in 20 mM Tris-ClO4 and 20 mM NaClO4 solution. Applied potential of 250 mV versus Ag/AgCl. [Zn2+] ) 0.4 mM; pH 8.6. In all cases the measured data points are shown as symbols with the calculated fit to the equivalent circuit as solid lines. Table 2. Fitted Impedance Values for the Rehybridized Complementary DNA Monolayer and the Rehybridized Middle Mismatch DNA Monolayera rehybridized 1:2 element /Ω‚cm2

RS RX/Ω‚cm2 CPEb/µF‚cm-2 RCT/Ω‚cm2 ∆RCT/Ω‚cm2 W/10-3 Ω‚s-1/2‚cm2

B-DNA

M-DNA

6.3 (0.6) 6.0 (0.5) 345 (11) 311 (15) 26 (1.5) 17 (0.6) 295 (11) 219 (5) 76 (12) 3.1 (0.1) 4.9 (0.3)

rehybridized 1:4 B-DNA

M-DNA

5.8 (0.6) 5.9 (0.8) 355 (17) 334 (9) 34 (0.9) 20 (1.5) 284 (12) 255(9) 29 (15) 3.9 (0.4) 5.2 (0.4)

a Values in parentheses represent the standard deviations from several electrode measurements (n g 5) not the nonlinear curve fitting errors. b See Table 1.

in Table 2. Apparent from Figure 4 and Table 2 is that the B-DNA films, which result from the hybridization of a matched and mismatched DNA target, are indistinguishable. However, the films clearly show a difference under M-DNA conditions. Again, RCT is used to discriminate between matched and mismatched DNA films. The difference in RCT between B- and M-DNA is consistently larger (76(12) Ω‚cm2) for a perfect duplex compared to a mismatched film in which ∆RCT decreases to 29(15) Ω‚cm2. The usefulness of M-DNA could be due to the inherent ability of Zn2+ to bind and electronically alter the duplex DNA without changing the electronic properties of ss-DNA monolayers. Thus, incomplete hybridization does not have a drastic effect on mismatch detection; rather a surface that contains more duplex DNA will simply result in a larger ∆RCT. In support of this hypothesis, a control experiment showing the small effect that Zn2+ addition has on ss-DNA is included in the Supporting Information. From a practical perspective, hybridization of a target strand is likely to be performed with a target DNA that is longer than the 20 bases of the probe. Therefore, strands 6 (match) and 7 (mismatch) were synthesized which contain three additional adenosines at both the 3′ and 5′ ends, creating overhangs. After hybridization of either 6 or 7 to strand 1, EIS was performed 4064 Analytical Chemistry, Vol. 76, No. 14, July 15, 2004

Figure 5. Nyquist plots (-Zim vs Zre) of the rehybridized 1:6 B-DNA (0), 1:7 B-DNA (O), 1:7 M-DNA (b), and 1:6 M-DNA (0) assembled on gold with a 4 mM Fe(CN)63-/4- (1:1) mixture as the redox probe in 20 mM Tris-ClO4 and 20 mM NaClO4 solution. Applied potential of 250 mV versus Ag/AgCl. [Zn2+] ) 0.4 mM; pH 8.6. In all cases the measured data points are shown as symbols with the calculated fit to the equivalent circuit as solid lines.

under standard conditions; the resulting Nyquist plot is shown in Figure 5. It is clear that under B-DNA conditions there is little difference between RCT for 1:6 (match with overhangs ) 510 Ω‚cm-2) and 1:7 (mismatch with overhangs ) 470 Ω‚cm-2) duplexes but after conversion to M-DNA this difference is enhanced (∆RCT is 160 and 70 Ω‚cm-2 for match and mismatch, respectively). Thus, as was the case for duplexes without overhangs, the measurement of ∆RCT gives good discrimination for targets with short overhangs. Control experiments were also performed where the rehybridization was carried out in either standard PCR buffer or in the presence of a 1000-fold excess of a noncomplimentary target sequence. In neither case was there a significant change in the value of ∆RCT (data not shown). Finally, to detect the sensitivity of the dehybridizationrehydridization method, a target concentration range from 10-5 to 10-12 M has been employed. Impedance spectra were recorded for matched and mismatched DNA in both B- and M-DNA conditions and the resulting impedance was fit to the equivalent circuit shown in Figure 2. As shown in Figure 6, ∆RCT remains relatively constant down to concentrations of 100 pM of target ss-DNA. It is important to emphasize that a clear discrimination between matched and mismatched DNA is obtained by the difference in RCT between B- and M-DNA. CONCLUSION Although this study is restricted to the detection of a pyrimidine‚ pyrimidine mismatch, purine‚purine mismatches should also be detectable since they cause similar disruptions to the helical stack of a DNA duplex.55 The presence of a mismatch causes a decrease in RCT regardless of the position of the mismatch. The difference in charge-transfer resistance, ∆RCT, between B- and M-DNA represents a reliable measure of the presence of a single nucleotide mismatch under ideal conditions. Even under conditions in which incomplete hybridization is observed, single (55) Yamashita, K.; Takagi, M.; Kondo, H.; Takenaka, S. Chem. Lett. 2000, 10381039.

from biological samples is required before this sensor can be utilized. In the long term, greater sensitivity can be achieved by reducing the size of the electrode. For example, if a macroelectrode can detect 100 pM of target, then a microelectrode with a corresponding reduction in target volume would have a sensitivity in the femtomolar range at which point direct detection without amplification of the target becomes realistic.

Figure 6. Determination of detection limits by monitoring the change in RCT between B-DNA and M-DNA as a function of target singlestranded DNA concentration. Complementary DNA strands (0) and middle mismatch DNA strands (O). Error bars are derived from a minimum of five electrodes.

nucleotide mismatches can be detected by the formation of M-DNA. Thus, these results present a significant step forward in the electrochemical detection of SNPs using an unlabeled DNA hybrid. A device with a detection limit of 100 pM can potentially be used for such applications as clinical diagnosis of mutations. However, an amplification step to increase DNA concentrations

ACKNOWLEDGMENT The authors wish to thank CIHR, NSERC, and UMDI for financial support, H.-B.K. is the Canada Research Chair in Biomaterials and J.S.L. is supported by a Senior Investigators Award from the Regional Partnership Program of CIHR. The authors also thank Don Schwab, the Plant Biotechnology Institute, Canada, for the preparation of DNA samples. SUPPORTING INFORMATION AVAILABLE HPLC and UV-visible spectra of 1, EIS mass spectrum of 1, and impedance spectra of ss-DNA monolayer and ss-DNA under M-DNA conditions. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review April 2, 2004. Accepted April 23, 2004. AC049482D

Analytical Chemistry, Vol. 76, No. 14, July 15, 2004

4065