Detection of Attomole Quantitites of DNA Targets on Gold

Omotayo A. Arotiba , Anna Ignaszak , Rehana Malgas , Amir Al-Ahmed , Priscilla G.L. Baker , Selwyn F. Mapolie , Emmanuel I. Iwuoha. Electrochimica Act...
0 downloads 0 Views 138KB Size
Download PDF
Anal. Chem. 2003, 75, 6586-6592

Detection of Attomole Quantitites of DNA Targets on Gold Microelectrodes by Electrocatalytic Nucleobase Oxidation Mitchell R. Gore, Veronika A. Szalai, Patricia A. Ropp, Ivana V. Yang, Joel S. Silverman, and H. Holden Thorp*

Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-3290

The electrochemical detection of nucleic acid targets at low concentrations has a number of applications in diagnostics and pharmaceutical research. Self-assembled monolayers of alkanethiol-derivatized oligonucleotides on gold electrodes provide a useful platform for such detectors, and the electrocatalytic oxidation of nucleobases included in the DNA targets is a particularly sensitive method of electrochemical detection. A strategy has been developed for combining these two aspects by substituting either 7,8-dihydro-8-oxoguanine (8G) or 5-aminouridine (5U) into DNA targets. Upon hybridization of targets containing these modified nucleobases, electrocatalytic signals at probe-modified gold electrodes are observed in the presence of Os(bpy)32+, which oxidizes both 8G and 5U upon oxidation to the Os(III) state. Self-assembled monolayers were prepared on both macro (1.6 mm) and micro (25 µm) gold electrodes using published procedures involving C6-terminated alkanethiol oligonucleotides and mercaptohexanol as the diluent. The extent of electrode modification by the modified probe was assessed using radiolabeling and a standard chronocoulometry method; both approaches gave loading levels within expected ranges ((1-6) × 1012 molecules/cm2). Hybridization of the modified targets where the non-native nucleobase was incorporated by solid-phase synthesis produced electrocatalytic signals from strands that were independently detected using radiolabeling and chronocoulometry. This result was used as a basis to develop an on-electrode amplification scheme where Taq polymerase was used to extend the immobilized DNA probes from solution-phase polymeric templates using modified nucleotriphosphates. This reaction produced an electrode that was modified with extended DNA containing the appropriate modified nucleotide. Radiolabeled nucleotide triphosphates were used to confirm the desired onelectrode DNA synthesis. When these electrodes were cycled in the presence of Os(bpy)32+, electrocatalytic signals were observed when as little as 40 amol (400 fM) of the desired target was present in the hybridization solution.

* Corresponding author. E-mail: [email protected].

6586 Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

The development of microarrays for detection of nucleic acid sequences has a number of important applications in diagnosis and in pharmaceutical research.1-7 The availability of arrays that are inexpensive to manufacture and simple to use would enable further adoption of microarray techniques in drug discovery and in the clinic.8-12 Use of electrochemical techniques instead of fluorescence can allow for simpler, smaller detectors.5,12-14 The simplest such system would involve a solid electrode modified with an oligonucleotide probe that produces a novel electrochemical signal upon hybridization of the immobilized probe to a specific target. A number of such systems are under development.1,8,12 We have developed an electrocatalytic system for DNA detection based on the oxidation of guanine by Ru(III) according to15-17

Ru(bpy)32+ f Ru(bpy)33+ + e-

(1)

Ru(bpy)33+ + DNA f Ru(bpy)32+ + DNAox

(2)

where DNAox is DNA where a guanine base has been oxidized by a single electron. This system has been used in a variety of detection systems by our group18,19 and others.11,20,21 Recent studies outside our laboratory11,20 have confirmed that the catalytic (1) Popovich, N.; Thorp, H. H. Interface 2002, 11, 30-34. (2) Mikkelsen, S. R. Electroanalysis 1996, 8, 15-19. (3) Wang, J.; Polsky, R.; Tian, B.; Chtrathi, M. P. Anal. Chem. 2000, 72, 52855289. (4) Tarlov, M. J.; Steel, A. B. In Biomolecular Films: Design, Function, and Applications; Rusling, J. F., Ed.; Marcel Dekker: New York, 2003; pp 545608. (5) Kuhr, W. G. Nat. Biotechnol. 2000, 18, 1042-1043. (6) Palecek, E.; Fojta, M. Anal. Chem. 2001, 73, 74A-83A. (7) Heller, M. J. Annu. Rev. Biomed. Eng. 2002, 4, 129-153. (8) Fan, C.; Plaxco, K. W.; Heeger, A. J. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 9134-9137. (9) Boon, E. M.; Ceres, D. M.; Drummond, T. G.; Hill, M. G.; Barton, J. K. Nat. Biotechnol. 2000, 18, 1096-1100. (10) 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. Diagn. 2001, 3, 74-84. (11) Dennany, L.; Forster, R. J.; Rusling, J. F. J. Am. Chem. Soc. 2003, 125, 5213-5218. (12) Park, S. J.; Taton, T. A.; Mirkin, C. A. Science 2002, 295, 1503-1506. (13) Willner, I. Science 2002, 298, 2407-2408. (14) Fritz, J.; Cooper, E. B.; Gaudet, S.; Sorger, P. K.; Manalis, S. R. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 14142-14146. (15) Thorp, H. H. Trends Biotechnol. 1998, 16, 117-121. (16) Napier, M. E.; Thorp, H. H. Langmuir 1997, 13, 6342-6344. (17) Armistead, P. M.; Thorp, H. H. Anal. Chem. 2000, 72, 3764-3770. 10.1021/ac034918v CCC: $25.00

© 2003 American Chemical Society Published on Web 10/23/2003

electrochemistry in eqs 1 and 2 is a particularly sensitive method of electrochemical detection of DNA on surfaces. The detection of native guanine bases by electrocatalysis requires Ru(bpy)32+ as a mediator, because the potential of the guanine oxidation is 1.07 V (vs SCE).22,23 Detection of the electrocatalysis from native guanine therefore requires the use of electrodes that can be cycled to relatively high positive potentials in water, such as indium tin oxide24 and carbon. However, we have studied a number of substituted nucleobases that exhibit much lower redox potentials, such as 8-oxoguanine and 5-aminouridine, which exhibit potentials closer to 0.5 V.23 These bases can be oxidized with metal complexes of lower potential, such as Os(bpy)33+. The immobilization of derivatized thiols on gold to form selfassembled monolayers has enabled many new technologies involving gold electrodes and particles.25,26 The modification of electrodes with thiol-derivatized DNA probes has been used to create numerous kinds of hybridization sensors where binding of the complementary target can be read either electrochemically or using a number of other methods.4,8-10,27 The studies of Steel et al. provide simple protocols for producing self-assembled monolayers of thiol-modified DNA probes that are amenable to electrochemical characterization.28 We report here on a new system for using catalytic electrochemistry to detect DNA targets hybridized to thiol-modified oligonucleotide probes on gold electrodes. In this system, a modified nucleobase is included in the DNA target. This modified nucleobase, typically 7,8-dihydro-8-oxoguanine (8-oxoguanine, 8G) or 5-aminouridine (5U), must exhibit a low enough potential to be oxidized by Os(bpy)33+.23 These modified nucleobases can be inserted into the target DNA by direct synthesis using modified phosphoramidites. However, an advantageous system is one in which the modified nucleobase is added to the target by a polymerase after the target has been hybridized to the electrode surface. Such a scheme can be effected using the nucleotide triphosphate of the given derivative and a polymerase that accepts the modification (Scheme 1). Willner et al. have used primer extension on electrodes modified with DNA in a strategy for detecting single-base mismatches after elongation with one base.29 We have used a our elongation strategy in combination with thermal denaturation on the electrode to detect as little as 40 amol of a given input target. EXPERIMENTAL SECTION Materials. All solutions were made with deionized water (18 mΩ‚cm resistivity) from a Millipore MilliQ system. DNA oligo(18) Yang, I. V.; Ropp, P. A.; Thorp, H. H. Anal. Chem. 2002, 74, 347-354. (19) Yang, I. V.; Thorp, H. H. Anal. Chem. 2001, 73, 5316-5322. (20) Zhou, L. P.; Rusling, J. F. Anal. Chem. 2001, 73, 4780-4786. (21) Wang, J.; Zhou, F. J. Electroanal. Chem. 2002, 537, 95-102. (22) Johnston, D. H.; Glasgow, K. C.; Thorp, H. H. J. Am. Chem. Soc. 1995, 117, 8933-8938. (23) Baik, M.-H.; Silverman, J. S.; Yang, I. V.; Ropp, P. A.; Szalai, V. A.; Yang, W.; Thorp, H. H. J. Phys. Chem. B 2001, 105, 6537-6444. (24) Armstrong, N. R.; Lin, A. W. C.; Fujihira, M.; Kuwana, T. Anal. Chem. 1976, 48, 741-750. (25) Taton, T. A.; Mirkin, C. A.; Letsinger, R. L. Science 2000, 289, 1757-1760. (26) Smith, E. A.; Kyo, M.; Kumasawa, H.; Nakatani, K.; Saito, I.; Corn, R. M. J. Am. Chem. Soc. 2002, 124, 6810-6811. (27) Hashimoto, K.; Ito, K.; Ishimori, Y. Anal. Chem. 1994, 66, 3830-3833. (28) Steel, A. B.; Herne, T. M.; Tarlov, M. J. Anal. Chem. 1998, 70, 4670-4677. (29) Patolsky, F.; Lichtenstein, A.; Willner, I. Nat. Biotechnol. 2001, 19, 253257.

Scheme 1

Table 1. Oligonucleotide Sequences 1-5′-ACGTACCAATTTTTGATGATGAACTTCATATCCTGAG TTATGTCGGAATTCTGCACCATCTTCACTTCAGAGATCTC CTCCGTCTTGATATTTGTCAACCCAGAAC-3′a,e 2-5′-ATAA8AC8TCTTACA8A88CCAA-3′b 3-5′-GTTCTGGTTTGACAAATATCAAGACGGAGGAGATCTC TGAAGTGAAGATGGATCAGAATTCCGACATAACTCAGGAT ATGAAGTTCATCATCAAAAATTGGTACGT-3′c 4-5′-PCB-GTTCTGGTTTGA-3′d 5-pUC19 plasmid linearized sense strand 6-5′-CCTCTTCGCTAT-3′ 7-5′-CGCTCACTGCCC-3′ 8-5′-GACGAGCATCAC-3′ 9-5′-GTAAGACACGAC-3′ 10-5′-GCTCAGTGGAAC-3′ 11-5′-GATTTATCAGCA-3′ 12-5′-GCCGCAGTGTTA-3′ 13-5′-TCTTTTACTTTC-3′ a 1-SH is 1 containing a 5′-(CH ) -thiol. b 8 is 8-oxoguanine. c 3-8G 2 6 is 3 in which all G’s are replaced by 8-oxoguanine, 3-5U is 3 in which all T’s are replaced by 5-aminouracil. d PCB is a 5′ photocleavable biotin. e 1primer-SH is the first 12 bases of the 5′ end of 1-SH.

nucleotides were purchased from the UNCsChapel Hill Lineberger Comprehensive Cancer Center Nucleic Acids Core Facility (Chapel Hill, NC). The probe DNA strand used (1; see Table 1) was a 107-bp segment (antisense strand) of the Homo sapiens amyloid precursor protein gene (chromosome 21), D678N mutant (National Center for Biotechnology Information, http://www.ncbi.nlm.nih.gov/, gi:116904654:1-107). The noncomplementary target DNA control 2 was a 23-mer sequence with 8-oxoguanine at positions 5, 8, 16, 18, and 19. Hexaammineruthenium(III) chloride (99%) was purchased from Strem Chemical and used as received. Os(bpy)3Cl2 was purchased from Sigma-Aldrich and used as received. 8-Oxoguanosine triphosphate was purchased from Tri-link Biotechnologies, Inc. (San Diego, CA) and used as received. 5-Aminouridine (Tri-link Biotechnologies, Inc.) was phosphorylated using the procedure developed by Hoheisel and Lehrach.30 The buffers used were the same as those used by Steel et al.28 DNA deposition was carried out in D-BFR, 1.0 M potassium phosphate buffer, pH 7; electrode rinsing was carried out in R-BFR, 10 mM NaCl, 5 mM Tris buffer, pH 7.4; hybridization was carried out in H-BFR, 1.0 M NaCl, 10 mM Tris buffer, pH 7.4, 1 mM (30) Hoheisel, J. D.; Lehrach, H. FEBS Lett. 1990, 274, 103-106.

Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

6587

Figure 1. Voltammetric behavior of Os(bpy)32+ at DNA-modified electrodes. The data correspond to electrodes modified with 1-SH and exposed to 3 (solid), 2 (dashed), 3-8G (dotted), or 3-5U (dashdotted). The concentration of Os(bpy)32+ was 100 µM in E-BFR/0.1 M NaCl. The electrode area was 0.3 cm2, and the sweep rate was 1 V/s.

EDTA; electrochemical experiments were carried out in E-BFR, 10 mM Tris buffer, pH 7.4. Electrode Preparation. Gold macroelectrodes were prepared by evaporation of a 200-Å chromium adhesion layer followed by a 2000-Å gold layer (both 99.99% purity) onto clean 1 cm × 1 cm glass squares; these larger electrodes were used in an O-ring cell that produces an electrode area of 0.32 cm2.31 Pencil-type gold macroelectrodes (1.6-mm diameter) were purchased from BAS. Gold microelectrodes (25-µm diameter) were purchased from BAS. Before each experiment, electrodes were cleaned by immersion in warm piranha solution (70% concentrated sulfuric acid, 30% hydrogen peroxide solution (30% in H2O)) for 15 min followed by 5% aqueous hydrofluoric acid for 30 s. The electrodes were then rinsed thoroughly with deionized water and immersed in the DNA deposition solutions while still wet. WARNING: Pirhana solution reacts violently with organic solvents and should not be stored in tightly capped containers. The modified electrodes were prepared using the procedure of Steel et al.28 The clean gold electrode was immersed in a 1.0 µM solution of probe oligonucleotide (1-SH or 1primer-SH) in D-BFR for 2 h, rinsed with R-BFR for 5 s, immersed in a 1.0 mM 6-mercapto-1-hexanol solution (MCH) solution in deionized water for 1 h, and rinsed for 5 s with R-BFR. Hybridization was performed at 35 °C for 60 min in H-BFR. In Figure 1 the concentration of complementary target and noncomplementary target (for nonspecific adsorption controls) was 0.1 µM and the solution volume was 1 mL. In Figure 2 the concentration of complementary target and noncomplementary target (for nonspecific adsorption controls) was 0.1 µM and the solution volume was 100 µL. After removal from the hybridization solution, electrodes were rinsed with R-BFR for 5 s. Measurements. Cyclic voltammetry and chronocoulometry were performed with a BAS100B (macroelectrodes) or Axon Geneclamp 500 amplifier/Digidata 1200 (microelectrodes) electrochemical analyzer with a single-compartment voltammetric cell equipped with a 1 cm × 1 cm gold wafer macroelectrode, 1.6-mm gold pencil macroelectrode, or 25-µm-diameter gold wire micro(31) Willit, J. L.; Bowden, E. F. J. Phys. Chem. 1990, 94, 8241-8246.

6588 Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

Figure 2. Voltammetric behavior of Os(bpy)32+ at DNA-modified electrodes. The data correspond to electrode modifications of 1-SH (circle), 3 (square), 2 (diamond), 3-8G (cross), or 3-5U (plus). The concentration of Os(bpy)32+ was 100 µM in E-BFR/0.1 M NaCl. The electrode area was 4.9 × 10-6 cm2, and the sweep rate was 1 V/s. Apparent hysteresis observed near the switching potential is due to artifacts from the background subtraction. The volume of the hybridization solution was 100 µL.

electrode as the working electrode, a Pt wire counter electrode, and a Ag/AgCl reference electrode. The Ag/AgCl electrode was purchased from BAS, Inc. For solutions containing Os(bpy)32+, cyclic voltammograms were taken from 0 to 0.8 V. For solutions containing Ru(NH3)63+, chronocoulometric response curves were collected by stepping the potential from 0.3 to -0.6 V. Background scans of buffer alone were collected for both techniques and subtracted from scans of metal complexes and immobilized DNA. All experiments were carried out at laboratory ambient temperature (21-25 °C). Target Oligonucleotide Synthesis. The target oligonucleotides (3, 3-8G, 3-5U) were synthesized using Klenow (exo-) DNA polymerase and a standard primer extension procedure. The 107mer 1 was used as the template and the 12-mer 4 as the primer. The 107-mer targets (3, 3-8G, or 3-5U) were then isolated using Roche strepavidin magnetic particles followed by 20% denaturing polyacrylamide gel electrophoresis. The amount of DNA isolated was determined by UV-visible at 260 nm, and the respective target oligonucleotides (3, 3-8G, 3-5U) were brought to a final concentration of 0.1 µM oligonucleotide in H-BFR. Quantification of DNA. The probe DNA surface density at the electrode surface was calculated using both radiolabeling and the chronocoulometric method devised by Steel et al.28 The radiolabeling procedure entailed radiolabeling either the 5′-end (for target DNA) or the 3′-end (for probe DNA) using γ-phosphorylated [32P]-dATP and polynucleotide kinase for 5′-labeling or dideoxy-[32P]-dATP and terminal deoxynucleotidal transferase for 3′-labeling. The labeled, immobilized oligonucleotides on the gold electrodes were then phosphorimaged and compared to phosphorimaged control spots of known amounts of the appropriate labeled oligonucleotide solutions on filter paper, and the amount of immobilized DNA was determined using the Imagequant software package. The chronocoulometric method has been described elsewhere.28 Briefly, the DNA-modified electrode is first immersed in E-BFR and chronocoulometric data are collected by stepping the potential from 0.3 to -0.6 V for 230 ms. The electrode is then

rinsed and immersed in a solution of 100 µM Ru(NH3)63+ in E-BFR, and chronocoulometric data are collected using the same procedure. The chronocoulometric response curves are converted to Anson plots by plotting charge versus (time)1/2. The linear part of the Anson plot is then extrapolated back to time zero to obtain the intercept for the plot in the presence and absence of Ru(NH3)63+. Since Ru(NH3)63+ in a low-ionic strength buffer essentially completely exchanges with the native charge-compensating cation of the immobilized DNA, the amount of chargecompensating Ru(NH3)63+ trapped at the electrode surface can be measured by chronocoulometry, and the surface density of the immobilized DNA is obtained from the integrated Cottrell expression via the equations

Q ) (2nFAD01/2C0/π1/2)t1/2 + Qdl +nFAΓ0

(3)

ΓDNA)Γ0(z/m)(NA)

(4)

where Q is charge (in C), t time (in s), n the number of electrons per reduced molecule, F the Faraday constant (C/equiv), A the electrode area (cm2), D0 the diffusion coefficient (cm2/s), C0 the bulk concentration (mol/cm2), Qdl the capacitive charge (C), nFAΓ0 the charge from the reduction of Γ0 (mol/cm2) of adsorbed redox marker, m the number of bases in the probe or target DNA, z the charge of the redox molecule, NA Avogadro’s number, and ΓDNA the surface density of the DNA in molecules/cm2. The surface excess of redox marker is then calculated as the difference in the chronocoulometric intercepts in the absence and presence of Ru(NH3)63+. The procedure was carried out after DNA probe deposition and again after DNA target hybridization in order to determine the hybridization efficiency for the experiment. On-Electrode Amplification. Gold electrodes (either 1.6 mm or 25 µm) were cleaned with piranha solution and rinsed with distilled water. The electrodes were then derivatized with a 5′thiolated primer (1primer-SH) and mercaptohexanol by the method described above. The electrodes were then incubated in a solution of the complementary sequence in hybridization buffer. The amplification solution for the macroelectrodes consisted of 360 µL of distilled water, 10 µL of 30 mM MgCl2, 1 µL each of R-[32P]-dATP (3.4 µM), dCTP (10 mM), dTTP (10 mM), and 8-oxodGTP (10mM), 1 µL of Taq polymerase, and 40 µL of 10× PCR buffer. The amplification solution for the microelectrodes consisted of 90 µL of distilled water, 2.5 µL of 30 mM MgCl2, 0.25 µL each of R-[32P]-dATP (3.4 µM), dCTP (10 mM), dTTP (10 mM), and 8-oxo-dGTP (10 mM), 1 µL of Taq polymerase (5 units/µL), and 10 µL of 10× PCR buffer. The hybridized electrode was rinsed, placed in the amplification solution, and incubated for 3 min at 72 °C, 3 min at 94 °C, and 3 min at 55 °C. This process was repeated for 20 cycles, with Geiger counter readings and cyclic voltammograms collected after 0, 1, 10, and 20 cycles. The electrode was rinsed thoroughly with R-BFR before Geiger readings and voltammograms were collected. RESULTS AND DISCUSSION Detection of Electrochemically Labeled Oligonucleotides. The system design in Scheme 1 was first investigated using synthetic oligonucleotide targets that contained the either 5-aminouracil (3-5U) or 8-oxoguanine (3-8G). These experiments were

first performed using 1 cm × 1 cm gold wafer electrodes. Mixed monolayers of 1-SH and MCH on gold were formed by immersion of the gold electrodes in the respective thiol solutions for 2 h using procedures adapted from the literature.28 The SAM-modified electrode was then immersed in a 0.1 µM solution of the noncomplementary oligonucleotide 2 or the complementary oligonucleotide 3 with only native bases or substituted with 8-oxoguanine (3-8G) or 5-aminouridine (3-5U). Cyclic voltammograms of the respective films were then collected in a solution of 100 µM Os(bpy)32+ at a scan rate of 1 V/s. Such relatively high scan rates are used to optimize the catalytic current, as we have discussed in detail elsewhere.17,32 As Figure 1 shows, hybridization of the attached probe to 3-8G or 3-5U produces an ∼50% enhancement in the peak oxidative current compared to that obtained after hybridization to the complementary oligonucleotide that does not contain guanine (3). Having shown that hybridization to the appropriately derivatized oligonucleotides produces a detectable current enhancement, we next sought to determine whether appreciable nonspecific binding of the target to the electrode was occurring. As shown in Figure 1, the current produced after exposure to the noncomplementary control target 2 is similar to that obtained after hybridization to the complementary native target 3, indicating that there is little, if any, nonspecific adsorption to the gold surface. This finding is consistent with the observations of Steel et al. where both chronocoulometry and X-ray photoelectron spectroscopy indicated that the diluent MCH thiol prevents the nonspecific adsorption of noncomplementary DNA to the modified gold surface.28 The precision of the detection system was assessed by interrogation of multiple films. The averages of the oxidative peak currents with standard deviations obtained for five different films of each type were as follows: 3, -2.3 ( 0.04 µA; 2, -1.74 ( 0.08 µA; 3-8G, -3.47 ( 0.06 µA; and 3-5U, -2.87 ( 0.1 µA. These values illustrate that the detection of the catalytic current for the modified-base containing films is both consistent and reproducible. The successful detection of electrochemically labeled nucleic acids on relatively large gold macroelectrodes with saturating target concentrations led us to investigate the extension of the same method to 25-µm gold microelectrodes with lower hybridization volumes. Use of miniaturized electrodes reduces the absolute amount of target detected in the reaction and should therefore lead to a decrease in the limit of detection of the method.1,27,33 Mixed monolayers of 1-SH and MCH were formed by immersion of the 25-µm gold electrode in the respective thiol solutions for 2 h. In the case of the microelectrodes, detection of each subsequent target was carried out with the same probe monolayer film following removal of the previous target sequence by denaturation of the film at 90 °C for 5 min in 1 M sodium phosphate buffer (pH 7.4). These studies also provide an assessment of the stability of the DNA-modified electrode and the robustness of the electrochemical detection system. Since the gold-thiol bond is relatively thermostable,34 we expected that the electrode could be regenerated by heat denaturation after each hybridization, which would mean that any current increase observed would come solely from the presence of the target strand. (32) Armistead, P. M.; Thorp, H. H. Anal. Chem. 2001, 73, 558-564. (33) Szalai, V. A.; Thorp, H. H. J. Phys. Chem. B 2000, 104, 6851-6859. (34) Delong, H. C.; Donohoe, J. J.; Buttry, D. A. Langmuir 1991, 7, 2196-2202.

Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

6589

Table 2. Extent of DNA Surface Modification × 10-12 Molecules/cm2 1.6-mm electrodes

1-SH

3

radiolabeling chronocoulometry 25-µm electrodes radiolabeling chronocoulometry

5.9 ( 0.5 3.6 ( 0.5 1-SH 0.8 ( 0.1 2.7 ( 0.6

2.7 ( 0.5 1.2 ( 0.1 3 0.5 ( 0.3 1.9 ( 0.8

hybridization efficiency (%) 46 33 63 70

The cyclic voltammograms for the modified electrodes are shown in Figure 2. As with the large electrodes, the peak current values for the mixed-monolayer probe film are very similar to those obtained following hybridization to the unmodified target 3 or exposure to the noncomplementary target 2. When the electrochemically active target strands 3-8G and 3-5U are hybridized to the probe monolayers, the oxidative peak current is increased by over 50%. After the electrochemically active target strands are heat denatured, the catalytic current disappears, indicating that the current is solely a result of the presence of the labeled oligonucleotide and that the mixed-monolayer film is indeed regenerated intact by the heat denaturation. The enhancement in the reductive peak is likely due to depletion of the modified base on the oxidative scan, which allows the oxidized mediator to accumulate at the electrode. In addition, the formation of duplex DNA on the surface often produces detectable increases in the capacitive current, which is evident in Figure 2. In the on-electrode amplification examples shown below, there is less evidence of increases in the capacitive current, probably because these reactions involve the addition of single-stranded DNA to the electrode. Quantification of Immobilized DNA. The quantities of probe and target nucleic acid on the electrode surface were determined using both radiolabeled nucleic acids and the chronocoulometry method of Steel et al.28 Table 2 shows the observed probe and target surface densities for the gold macroelectrodes and gold wire microelectrodes. All of the values are within the range ((1-10) × 1012 molecules/cm2) observed by Steel et al.28 and are below the maximum value imposed by the physical dimensions of the DNA double helix itself.35 There is some variance between the values obtained for the two electrode sizes and between the values obtained using the two different methods. However, the hybridization efficiencies are similar for the 1.6-mm electrodes for each method (34 vs 47%) and also very similar for the 25-µm electrodes (64 and 69%). Most important, the independent quantitation by two different methods confirms that the current enhancements shown in Figures 1 and 2 are due to hybridization of modified DNA targets. On-Electrode Amplification. The quantitation results and the current enhancements observed in Figures 1 and 2 clearly demonstrate that there is a high specificity in the detection system for the modified bases. This result suggested to us that a system where a polymerase was used to incorporate the modified base would allow for high target selectivity with improved sensitivity. Further, because we had shown that the probe-modified electrode (35) Nadassy, K.; Tomas-Oliveira, I.; Alberts, I.; Janin, J.; Wodak, S. J. Nucleic Acids Res. 2001, 29, 3362-3376.

6590 Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

Figure 3. (A) A mixed 1primer-SH/mercaptohexanol monolayer is formed on the gold electrode surface. (B) Annealing (55 °C): 1primerSH, acting as both probe and primer, captures its target sequence. (C) Extension (72 °C): Taq polymerase extends 1primer-SH using the target sequence as a template and incorporating the electrochemically labeled 8-oxo-dGTP into the probe strand. (D) Denaturation (94 °C): The target sequence is heat denatured from the electrochemically labeled probe strand and is now free to anneal to a new, unextended probe during the next annealing cycle. (E) After numerous cycles, the target sequence has acted as a template for the extension and electrochemical labeling of many probe strands.

was stable at relatively high temperatures, we could use thermal cycling to add modified bases catalytically and lower the detection limits. To realize such a system, we exposed the electrode that was hybridized to the native target to a solution of Taq polymerase and the nucleotide triphosphates (dNTPs) of the bases where one of the dNTPs is substituted by either 8G or 5U. This step elongates the attached probe with a sequence that is complementary to that of the target but that is modified with the electrochemically active base. Following release of the target by heat denaturation and rehybridization, additional probe strands can be elongated with the modified nucleobase. This procedure leads to amplification of the target strand and production of an electrode covalently modified with single-stranded probes that are specifically labeled when the target strand is present. The pertinent details of this system are shown in Figure 3. To test the system shown in Figure 3, a mixed monolayer of 1primer-SH and MCH was formed on a 25-µm gold electrode.

Figure 4. (A) Voltammetric behavior of Os(bpy)32+ at 1primer-SHmodified electrodes during on-electrode amplification using 3 as the target. The data correspond to cycle 0 (solid), cycle 1 (dash-dotted), cycle 10 (dashed), and cycle 20 (dotted). The concentration of Os(bpy)32+ was 100 µM in E-BFR/0.1 M NaCl. The electrode area was 4.9 × 10-6 cm2, and the sweep rate was 100 mV/s. (B) DNA synthesized at 1primer-SH-modified electrodes detected during onelectrode amplification. The data correspond to 1 as the target (solid) or 3 as the target (dashed).

Figure 5. (A) Voltammetric behavior of Os(bpy)32+ at 1primer-SHmodified electrodes during on-electrode amplification using 3 as the target. The data correspond to cycle 0 (solid), cycle 1 (dash-dotted), cycle 10 (dashed), and cycle 20 (dotted). The concentration of Os(bpy)32+ was 100 µM in E-BFR/0.1 M NaCl. The electrode area was 4.9 × 10-6 cm2, and the sweep rate was 100 mV/s. (B) DNA synthesized at 1primer-SH-modified electrodes during on-electrode amplification. The data correspond to 1 as the target (solid) or 3 as the target (dashed).

The electrode was then incubated in a 1 µM solution of either noncomplementary oligonucleotide 1 or target 3 for 1 h. The experimental and control electrodes were then placed in the amplification solution and subjected to thermal cycling as described in the Experimental Section. The amount of DNA synthesized (Figures 4B, 5B, and 6B) was then determined using a Geiger counter by measuring the radioactivity at the electrode in counts per second. The amount of radioactivity was then converted to moles of DNA by determining the counts per second emitted by a known quantity of R-[32P]-dATP and adjusting for the number of A’s per synthesized DNA strand. As shown in Figure 4A, there is a 2-fold increase in the oxidative current for the electrode hybridized to 3 after 1 temperature cycle; this enhancement increases to 4-fold after 20 cycles. No such increase in signal was obtained for the electrode exposed to 1. Parallel experiments were conducted using radioactive dATP during the amplification so that the number of nucleotides added to the probe during amplification could be determined. As shown in Figure 4B, the amount of radiation at the electrode surface increases 10fold in the presence of the target, from 2.5 counts/s after 1 cycle to 25 counts/s after 20 cycles. This result indicates that increasing amounts of radiolabeled dATP are incorporated into the extended thiolated primers during the amplification reaction, as expected. In contrast, the amount of radiation at the electrode exposed to

noncomplementary 1 remains essentially unchanged during the thermal cycling. This background level of radiation indicates that some dATP adheres nonspecifically to the electrode but that no DNA synthesis takes place in the absence of the correct target sequence, even after multiple cycles. In Figure 4, large concentrations of target were present in experiments designed initially to validate the on-electrode amplification concept. We then sought to limit the concentration of target in solution to demonstrate that the on-electrode amplification provided the desired decrease in the limit of detection. The experiment was repeated using the 25-µm electrode in the presence of only 40 amol of 3 (400 fM). As shown in Figure 5A, there was again a 2-fold increase in oxidative current after 1 cycle that increased to almost 4-fold after 20 cycles. Figure 5B also shows that the amount of radiation at the electrode surface increased almost 4-fold in the presence of the desired target 3 but remained relatively constant when the target was not present. As expected, the increase in the amount of DNA synthesized is lower than in Figure 4B due to the lower amount of available target. We next sought to demonstrate that the on-electrode amplification system could be used to selectively detect a scarce target in a complex mixture. In this experiment, monolayers of 1primerSH and MCH were formed on two 1.6-mm gold electrodes. The Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

6591

in the amplification solutions and subjected to thermal cycling. As shown in Figure 6A, there is no catalytic current when 3 is absent from the mixture, even after 20 cycles, indicating that no DNA is synthesized and no 8-oxo-dGTP is incorporated. In contrast, Figure 6B shows that when the complementary target 3 is present in the complex mixture, a 1.5-fold increase in oxidative current is observed after 20 cycles. This indicates that, even in the presence of a 1000-fold excess of noncomplementary targets, the correct target hybridizes to the film and acts as a template for the extension of the immobilized probe. The result in Figure 6C confirms that the current enhancement in Figure 6B is due to the synthesis of the desired DNA.

Figure 6. (A) Voltammetric behavior of Os(bpy)32+ at 1primer-SHmodified electrodes during on-electrode amplification using 1, 2, and 5-13 as the targets. The data correspond to cycle 0 (solid), cycle 1 (dashed), cycle 10 (dotted), and cycle 20 (dash-dotted). The concentration of Os(bpy)32+ was 100 µM in E-BFR/0.1 M NaCl. The electrode area was 0.02 cm2, and the sweep rate was 100 mV/s. (B) Voltammetric behavior of Os(bpy)32+ at 1primer-SH-modified electrodes during on-electrode amplification using 1, 2, 3, and 5-13 as the targets. The data correspond to cycle 0 (solid), cycle 1 (dashed), cycle 10 (dotted), and cycle 20 (dash-dotted). The concentration of Os(bpy)32+ was 100 µM in E-BFR/0.1 M NaCl. The electrode area was 0.02 cm2, and the sweep rate was 100 mV/s. (C) DNA synthesized at 1primer-SH-modified electrodes during on-electrode amplification. The data correspond to 1, 2, and 5-13 as the targets (solid line) or 1, 2, 3, and 5-13 as the targets (dashed line).

experimental electrode was then incubated in a solution containing 40 fmol of 3 (0.4 nM) and 40 pmol (0.4 µM) each of 11 different noncomplementary sequences (1, 2, 5-13). The control electrode was then incubated in a solution containing only 40 pmol (0.4 µM) each of the noncomplementary sequences. Note that the concentrations are higher in these experiments because the 1.6-mm electrodes were used. The electrodes were then placed 6592

Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

CONCLUSION Gold electrodes functionalized with mixed monolayers of thiolmodified DNA and MCH produce excellent platforms for hybridization of target DNA. These films exhibit little binding of noncomplementary targets and are thermally robust up to temperatures where hybridized target DNA can be removed by thermal denaturation. Detection of target DNA using electrocatalytic nucleobase oxidation can be realized on gold electrodes using Os(bpy)33+/2+ as the mediator and modified nucleobases, such as 8-oxoguanine or 5-aminouridine, as the oxidized base. This system allows for the sensitive and selective detection of target oligonucleotides on the gold monolayer platform. We have exploited the thermal stability and low nonspecific binding of the gold monolayer electrodes to develop a system for on-electrode amplification of DNA targets. In this system, DNA is hybridized to the probe-modified electrode, which is then treated with a thermostable polymerase and a mixture of nucleotide triphosphates where one nucleotide is replaced with either 8-oxoguanine or 5-aminouridine. Upon thermal cycling, the probe strands immobilized on the electrode become elongated with sequences complementary to the target template with the desired modified base inserted. These electrodes can then be electrocatalytically oxidized in the presence of Os(bpy)32+. The method has been used to detect template strands with as little as 40 amol present in the solution. In addition, DNA was successfully synthesized and detected from complex mixtures containing 11 sequences at 1000-fold excess compared to the desired target. ACKNOWLEDGMENT The authors gratefully acknowledge Jon Howell for custom fabrication of the 25-µm gold microelectrode and Dr. Alex Marzec for assistance with the Axon potentiostat. This research was supported in part by Clinical Micro Sensors, Inc., A Motorola Company. M.R.G. gratefully acknowledges the National Science Foundation and Dr. Henry Frierson for financial support. M.R.G. was supported by the RES Program at UNCsChapel Hill and an NIH predoctoral fellowship (1-F31-HG02520-01). We thank Professor Royce Murray and his co-workers for assistance in preparing the gold macroelectrodes.

Received for review August 6, 2003. Accepted September 22, 2003. AC034918V