Charge Transfer through DNA: A Selective Electrochemical DNA

The key novelty of the in situ experiment is that the redox intercalator (AQMS) is ... in environment of the AQMS upon intercalation results in a shif...
0 downloads 0 Views 247KB Size
Anal. Chem. 2006, 78, 2138-2144

Articles

Charge Transfer through DNA: A Selective Electrochemical DNA Biosensor Elicia L. S. Wong and J. Justin Gooding*

School of Chemistry, The University of New South Wales, Sydney, NSW 2052, Australia

The charge-transfer properties of DNA duplexes are exploited to produce a fast, simple, sensitive, and selective DNA biosensor by exposing the DNA recognition interface to a sample containing target DNA and the redox-active intercalator, anthraquinonemonosulfonic acid (AQMS). Electrochemistry from electron transfer through the DNA to AQMS intercalated into DNA duplexes can be differentiated from electrochemistry due to direct access of the AQMS to the electrode surface due to the difference in the environment of the AQMS giving a shift in the potential at which the molecule is reduced. The ability to distinguish between the two electrochemical signals enables DNA hybridization to be monitored in real time. This in situ detection scheme has good selectivity, being able to differentiate between a complementary target DNA sequence and one containing either C-A or G-A singlebase mismatches. The concentration detection limit of the biosensor is 0.5 nM (1 pmol) with an assay time of 1 h. The fact that the end user is only required to simultaneously add the sample containing the target DNA and AQMS gives a DNA biosensor that is highly compatible with PCR on chip technologies. Handheld DNA biosensor technologies that focus on the direct detection of nucleic acids are currently an area of tremendous interest as they play a major role in clinical, forensic, and pharmaceutical applications.1,2 For this technology to continue to emerge as commercially viable, there are several challenges that current research aims to overcome. Some major challenges include higher selectivity, higher sensitivity, shorter assay times, and greater simplicity in performing the assay. The latter is motivated by the need to integrate DNA biosensors with PCR on chip devices as at present there is no DNA detection technology sufficiently sensitive to avoid an amplifications step. Simplicity in operating a DNA biosensor becomes paramount upon integration with a PCR reactor on a chip as every step that requires addition of a reagent or washing greatly increases the complexity of the final chip. * Corresponding author. E-mail: [email protected]. (1) Wang, J. Nucleic Acids Res. 2000, 28, 3011-3016. (2) Niemeyer, C. M.; Blohm, D. Angew. Chem., Int. Ed. Engl. 1999, 38, 28652869.

2138 Analytical Chemistry, Vol. 78, No. 7, April 1, 2006

With regard to simplicity, label-free approaches such as surface plasmon resonance and the quartz crystal microbalance are attractive since the final DNA biosensor only needs to be exposed to the sample. However, both techniques, in common with most label-free biosensors, respond to any change at the interface and hence suffer from a lack of selectivity due to nonspecific adsorption.3-5 Using a label, either in solution attached to the DNA sequence being analyzed (the target) or to the surface immobilized recognition sequence (the probe) provides enhanced selectivity but at the cost of simplicity in performing the assay or preparing the sample. Labeling techniques have been used with all the common transducing elements for DNA biosensors including optical,6-11 mass-sensitive,3 and electrochemical12-16 devices. Of these, electrochemical DNA biosensors are perhaps the most promising transduction approach for integration with PCR reactors on a chip because miniaturization of electrochemical sensors is compatible with advanced microfabrication technology.17 Furthermore, electrochemical approaches can offer simple detection of target DNA with good sensitivity and unrivaled specificity.18 With regard to specificity, Barton and co-workers have developed an electrochemical DNA detector that can detect single-base pair mismatches without requiring any stringent washings. The concept relies on long-range charge transfer through DNA (3) Okahata, Y.; Matsunobu, Y.; Ijiro, K.; Mukae, M.; Murakami, A.; Makino, K. J. Am. Chem. Soc. 1992, 114, 8299-8300. (4) Peterson, A. W.; Wolf, L. K.; Georgiadis, R. M. J. Am. Chem. Soc. 2002, 124, 14601-14607. (5) Wong, E. L. S.; Chow, E.; Gooding, J. J. Langmuir 2005, 21, 6957-6965. (6) Nilsson, P.; Persson, B.; Uhlen, M.; Nygren, P. A. Anal. Biochem. 1995, 224, 400-408. (7) Piunno, P. A. E.; Krull, U. J.; Hudson, R. H. E.; Damha, M. J.; Cohen, H. Anal. Chem. 1995, 67, 2635-2643. (8) Abel, A. P.; Weller, M. G.; Duveneck, G. L.; Ehrat, M.; Widmer, H. M. Anal. Chem. 1996, 68, 2905-2912. (9) Pilevar, S.; Davis, C. C.; Portugal, F. Anal. Chem. 1998, 70, 2031-2037. (10) Fang, X. H.; Li, J. W. J.; Perlette, J.; Tan, W. H.; Wang, K. M. Anal. Chem. 2000, 72, 747a-753a. (11) Fang, X. H.; Liu, X. J.; Schuster, S.; Tan, W. H. J. Am. Chem. Soc. 1999, 121, 2921-2922. (12) Boal, A. K.; Barton, J. K. Bioconjugate Chem. 2005, 16, 312-321. (13) Xu, C.; He, P. G.; Fang, Y. Z. Anal. Chim. Acta 2000, 411, 31-36. (14) Millan, K. M.; Mikkelsen, S. R. Anal. Chem. 1993, 65, 2317-2323. (15) Kelley, S. O.; Barton, J. K.; Jackson, N. M.; Hill, M. G. Bioconjugate Chem. 1997, 8, 31-37. (16) Wong, E. L. S.; Gooding, J. J. Anal. Chem. 2003, 75, 3845-3852. (17) Gooding, J. J. Electroanalysis 2002, 14, 1149-1156. (18) Liu, X. J.; Tan, W. H. Anal. Chem. 1999, 71, 5054-5059. 10.1021/ac0509096 CCC: $33.50

© 2006 American Chemical Society Published on Web 02/28/2006

duplexes.15,19-21 Redox-active intercalators were used as labels either in solution (methylene blue) or attached to the probe DNA (daunomycin). Electrons have been shown to transfer through DNA duplexes, but not through single stranded DNA, to the intercalator at the distal end of the DNA.20 Any perturbation of the base-pair stacking (e.g., mismatched base pairing) greatly decreased the efficiency of charge transfer with a concomitant diminution in electrochemical current.15,16,19-22 As a consequence, such an approach allows the detection of all the different singlebase pair mismatches, with the exception of G-A mismatches, without further processing steps.15,20 G-A mismatches could, however, be detected by adding ferricyanide to the measurement solution, which electrocatalytically turned over the methylene blue and in the process amplified the difference between current from the perfect complement and the G-A mismatches sequence.19 The one limitation of the technology is a close-packed array of DNA molecules is required to ensure the electrochemical response was due to charge transfer through the DNA. The need for a close-packed array of DNA duplexes compromises the robustness of the assay as well as limiting both the assay time and the detection limit (a recent report by Inouye et al.23 on a DNA biosensor based on the same long-range charge-transfer principle using the same DNA recognition interface but using ferrocene-labeled probe DNA required hybridization time of ∼90 min with concentrations detection limit). Difficulties arise with an interface composed of a close-packed array of probe DNA because target DNA approaching the interface is repelled from the highly negatively charged surface. Furthermore, a crowded interface also limits the configuration freedom of the DNA required to form the duplex.24,25 We have extended the idea of DNA transduction via long-range charge transfer using an anionic intercalator, 2,6-anthraquinonedisulfonic acid (AQDS) to overcome limitations related to the design of the DNA interface.16 The DNA recognition interface consisted of mixed monolayer of loosely packed thiolated ss-DNA probes and a diluent component of 6-mercapto-1-hexanol (MCH) on a gold electrode. The purpose of the diluent was to prevent nonspecific adsorption of DNA and AQDS to the electrode surface as well as to space the probe molecules apart. The measurement was performed in a sequential manner with the DNA recognition interface first exposed to complementary target DNA. After the formation of DNA duplex, it was incubated in an AQDS solution and finally transferred to a solution with no intercalator for electrochemical measurement. With this procedure, no electrochemical current was observed prior to hybridization when only ss-DNA was present on the electrode surface. This absence of signal is due to the electrostatic repulsion between the anionic AQDS molecules and the alcohol-terminated MCH diluent layer. (19) Kelley, S. O.; Boon, E. M.; Barton, J. K.; Jackson, N. M.; Hill, M. G. Nucleic Acids Res. 1999, 27, 4830-4837. (20) Kelley, S. O.; Jackson, N. M.; Hill, M. G.; Barton, J. K. Angew. Chem., Int. Ed. Engl. 1999, 38, 941-945. (21) Boon, E. M.; Ceres, D. M.; Drummond, T. G.; Hill, M. G.; Barton, J. K. Nat. Biotechnol. 2000, 18, 1096-1100. (22) Wong, E. L. S.; Erohkin, P.; Gooding, J. J. Electrochem. Commun. 2004, 6, 648-654. (23) Inouye, M.; Ikeda, R.; Takae, M.; Tsuri, T.; Chiba, J. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 11606-11610. (24) Steel, A. B.; Herne, T. M.; Tarlov, M. J. Anal. Chem. 1998, 70, 4670-4677. (25) Peterson, A. W.; Heaton, R. J.; Georgiadis, R. M. Nucleic Acids Res. 2001, 29, 5163-5168.

As a result, any AQDS current observed after exposure to complementary target can be attributed to long-range charge transfer through the DNA. Not only does this approach allow the detection of DNA hybridization but it has the ability to discriminate both C-A and G-A mismatches without further processing steps19 and the detection of a target sequence from a cocktail of other noncomplementary DNA molecules.26 Despite its selectivity and robustness, the DNA biosensor based on long-range charge transfer using AQDS still has limitations with regard to detection limit (100 nM) and a long assay time (5 h).22 Furthermore, in common with the Barton approach (when methylene blue was used as the intercalator),15,19,21 assay simplicity is not optimal as the redox intercalators were added after hybridization had occurred, which gives an additional step when performing the measurement. The purpose of this paper is to present a new experimental strategy that is able to detect DNA hybridization in real time using long-range charge transfer with an approach we refer to as the in situ assay. We report the scope of the in situ methodology to (i) simplify the detection of DNA hybridization such that it is compatible with the design of PCR on the chip, (ii) detect DNA hybridization in subnanomolar concentration within 1 h, and (iii) detect single-base mismatch without further processing steps. The key novelty of the in situ experiment is that the redox intercalator (AQMS) is present in the sample solution containing the target DNA. Hence, the assay can be conducted in a single step rather than in a sequential manner.16 This is possible because the change in environment of the AQMS upon intercalation results in a shift in the reduction potential of the redox intercalator. Therefore, current due to charge transfer through the DNA duplexes can be differentiated from AQMS, which is reduced by diffusing directly to the electrode surface. Furthermore, as the redox-active intercalator and the target DNA coexist in the same solution, the in situ approach allows the DNA hybridization to be monitored in real time. The ability to monitor DNA hybridization in real time is desirable for the detection of RNA within living cells.27,28 EXPERIMENTAL SECTION Materials. MCH, 2-mercapto-1-ethanol (MCE), and AQMS were purchased from Aldrich Chemicals (Sydney, NSW, Australia). Reagent grade K2HPO4, KH2PO4, NaCl, KCl, NaOH, HCl, and absolute ethanol were purchased from Ajax (Sydney, Australia). All the chemicals were used as received without further purification. The 20-mer deoxyoligonucleotides were purchased from Proligo Pty. Ltd. (Sydney, NSW, Australia) with purification by HPLC. The base sequences of the deoxyoligonucleotides were as follows: thiolated DNA probe (20-base sequence DNA P1), 5′-GGGGCAGAGCCTCACAACCT-p-(CH2)3-SH-3′; complementary DNA target to P1 (20-base sequence DNA 2), 5′-AGGTTGTGAGGCTCTGCCCC-3′; C-A mismatch DNA target to P1 (20-base sequence DNA 3), 5′-AGGTTGTGAGGCCCTGCCCC-3′; G-A mismatch DNA target to P1 (20-base sequence DNA 4), 5′AGGTTGTGAGGCGCTGCCCC-3′; noncomplementary DNA tar(26) Wong, E. L. S.; Mearns, F. J.; Gooding, J. J. Sens. Actuators, B 2005, 111, 515-521. (27) Kostrikis, L. G.; Tyagi, S.; Mhlanga, M. M.; Ho, D. D.; Kramer, F. R. Science 1998, 279, 1228-1229. (28) Piatek, A. S.; Tyagi, S.; Pol, A. C.; Telenti, A.; Miller, L. P.; Kramer, F. R.; Alland, D. Nat. Biotechnol. 1998, 16, 359-363.

Analytical Chemistry, Vol. 78, No. 7, April 1, 2006

2139

get to P1 (20-base sequence DNA 5), 5′-GGATGGACGAAGCGCTCAGG-3′. All deoxynucleotides stock solutions were prepared with TrisHCl (pH 8.0) and stored in a -80 °C freezer until use. Solution Preparation. All solutions were prepared with Milli-Q water (18 MΩ cm, Millipore, Sydney, Australia) unless stated otherwise. Several buffers were used in this work: immobilization buffer contained 1 M KH2PO4 (pH 4.5); hybridization buffer contained 10 mM Tris-HCl, 1 M NaCl (pH 7.0); and phosphate buffer contained 0.05 M K2HPO4/KH2PO4, 0.3 M NaCl (pH 7.0). Stock solutions of 100 µM AQMS were prepared in 50 mM K2HPO4/KH2PO4, 0.2 M KCl (pH 7.0). The pH was adjusted with either NaOH or HCl solution. Milli-Q water and all buffers were sterilized using an autoclave. Instrumentations. All electrochemical measurements were performed with a BAS 100B electrochemical analyzer (Bioanalytical Systems Inc., Lafayette, IN). Cyclic voltammetry (CV) and Osteryoung square wave voltammetry (OSWV) were carried out in a single-compartment cell with a 2-mL volume. A conventional three-electrode system, comprising a bare or modified gold working electrode, a platinum flag auxiliary electrode, and an Ag|AgCl|3.0 M NaCl reference electrode (Bioanalytical Systems Inc.), was used for the measurement. All potentials are reported versus Ag|AgCl reference at room temperature. Solutions were degassed with argon for ∼15 min prior to performing the electrochemical measurements and were blanketed under an argon atmosphere during the entire experimental period. CVs were measured between 0 and -700 mV unless stated otherwise. OSWVs were conducted with a pulse amplitude of 25 mV, a step of 4 mV, and a frequency of 10 Hz. OSW voltammograms were measured between 0 and -700 mV. Preparation of the DNA Recognition Interface. Before all the immobilization of probe ss-DNA, the gold working electrodes were cleaned and prepared as described previously.29 The electrochemical area of the gold electrode was determined from the reduction of gold oxide using a conversion factor of 482 µC cm-2.30 All current densities quoted are relative to the electrochemically determined surface area. The cleaned gold electrodes were immersed in 4 µM thiolated DNA P1 for 90 min in immobilization buffer (1 M KH2PO4, pH 4.5). Subsequently, the P1 modified gold surfaces were incubated in 1 mM alkanethiol solution for 30 min, followed by rinsing with phosphate buffer (the modified electrode is referred to as the DNA recognition interface, which can be either P1/MCH or P1/MCE modified electrode). Hybridization and Detection of Immobilized DNA. Hybridization and detection of the DNA duplexes was conducted simultaneously in the same solution by exposing the DNA recognition interface to the electrochemical measuring solution (0.05 M K2HPO4/KH2PO4, 1.0 M NaCl, pH 7.0) containing target DNA (typically 4 µM for most experiments presented here) and 25 µM AQMS intercalator. Note that AQMS is used in this study rather than AQDS to prevent the intercalation being the ratelimiting step as AQMS has been shown to intercalate more rapidly than AQMS.22 (29) Gooding, J. J.; Erokhin, P.; Hibbert, D. B. Biosens. Bioelectron. 2000, 15, 229-239. (30) Hoogvliet, J. C.; Dijksma, M.; Kamp, B.; Bennekom, W. P. v. Anal. Chem. 2000, 72, 2016.

2140

Analytical Chemistry, Vol. 78, No. 7, April 1, 2006

RESULTS AND DISCUSSION The procdure for the detection of DNA hybridization via longrange charge transfer used in this work is depicted in Scheme 1. Detection of hybridization was conducted in the presence of target DNA and the redox intercalator in what we call the in situ assay. This is distinct from the sequential method described previously where hybridization, intercalation, and electrochemical measurement were performed in three separate steps.16 The in situ method is possible because of the observation made when performing the sequential method that upon DNA hybridization a distinct shift in the reduction electrode potential, Ec, of the anthraquinone at the ds-DNA-modified electrode occurs relative to a bare or MCHmodified-gold electrode.16,22 In Situ Detection. The Ec of AQMS, for example, in the solution recorded at the bare gold electrode as determined by CV, was found to be -540 mV (Figure 1a). In contrast, the Ec for intercalated AQMS at the ds-DNA-modified electrode was at -480 mV (Figure 1b). The anodic shift in reduction potential indicates a stabilization of the reduced form of the redox intercalator when intercalated into the DNA duplexes. The difference in reduction potential provides the ability to distinguish between the background current due to the AQMS directly accessing the electrode and the current from the AQMS molecules intercalated in the DNA duplex. This is important because the redox species in solution can still access the electrode directly despite the presence of the diluent layer (part of the DNA recognition interface), as the short alkanethiol provides an insufficient barrier to block electron transfer. Thus, the difference in redox measurement potential provides the opportunity to dispense with the three separate steps of hybridization, intercalation, and measurement.16 The preliminary experiment for in situ detection using AQMS was performed by recording an OSWV of the DNA recognition interface in an electrochemical measuring solution (0.05 M KH2PO4/K2HPO4, 1.0 M NaCl, pH 7.0) containing 25 µM AQMS and 4 µM complementary target DNA. The same DNA recognition interface used in our previous study,16 comprising a mixed selfassembly monolayer of mercaptopropyl-terminated probe ss-DNA tethered at the 3′ end of the DNA (P1) and a diluent layer of MCH, was employed. The OSWV was performed at 15-min intervals. During the first 60 min of the OSWV experiments, a single reduction peak was observed at -550 mV that is similar to the potential obtained for AQMS in solution as shown in Figure 1a. After 60 min, a shoulder peak began to appear at around -470 mV (Figure 2). Similar results were observed using CV, but because of the inferior sensitivity of CV, relative to OSWV, the peaks were less well resolved (CV not shown). The shoulder is attributed to current arising from the intercalated AQMS molecules. Thus, the presence of the shoulder allows the transduction of DNA hybridization. The magnitude of the OSWV current of the species contributing the shoulder peak is demonstrated by a background subtraction of the major peak resulting from the direct access of the AQMS to the electrode (inset of Figure 2). This electrochemistry was only observed when duplexes are assembled on the electrode surface. The shoulder peak increased in magnitude over a period of 2 h. These observations led us to conclude that the electrochemistry is derived from AQMS associated with the DNA duplex. Further evidence for this comes from the

Scheme 1. Schematic Representation of DNA Transduction via Long-Range Electron Transfera

a Through the (a) in situ approach, where the electrochemical measurement was performed in the presence of target DNA and intercalator and (b) sequential electrochemical approach, where the electrochemical measurement was performed after the (i) hyridization and (ii) intercalation events.

Figure 1. (a) CV of the bare gold electrode in 1 mM AQMS intercalator solution, containing 0.05 M phosphate buffer, 1.0 M NaCl (pH 7.0). The scan rate was 100 mV s-1. (b) CV of the DNA recognition interface after exposure to 4 µM complementary target in 0.05 M phosphate buffer, 1.0 M NaCl (pH 7.0) after incubation in 1 mM AQMS solution for 3 h and rinsing with phosphate buffer. The scan rate was 100 mV s-1. The reduction currents of AQMS in the solution and once intercalated are distinctively different so that current arising from nonspecific interaction can be distinguished from the intercalation current.

exposure of the DNA recognition interface to noncomplementary target DNA. Upon exposure of the DNA recognition interface in the measuring solution containing 4 µM noncomplementary target DNA, DNA 5, and 25 µM AQMS solution, the shoulder peak was not observed even several hours after the addition of target DNA and AQMS (Figure 3(i)).

The inset in Figure 2 also demonstrates the ability of the in situ approach to monitor DNA hybridization in real time. With measurements recorded at regular time intervals, the time to form a complete duplex with the P1/MCH interface is 2 h. As longrange charge transfer requires the formation of complete duplexes to give an electrochemical signal, this measurement informs on Analytical Chemistry, Vol. 78, No. 7, April 1, 2006

2141

Figure 2. OSWVs of the DNA recognition interface after exposure to complementary target ss-DNA. OSWV were performed in 25 µM AQMS solution in 0.05 M phosphate buffer, containing 1.0 M NaCl (pH 7.0). The step is 4 mV, with pulse amplitude of 25 mV and frequency of 10 Hz. The appearance of shoulder peaks was attributed to the intercalation of AQMS into the DNA duplexes and was used to transduce hybridization. The inset showed the resulting backgroundsubtracted shoulder peak with time scale indicating the growth of this shoulder peak with time.

Figure 3. OSW voltammograms of the subtracted shoulder peak of the DNA recognition interface after exposure to (i) noncomplementary, (ii) complementary, (iii) C-A mismatch, and (iv) G-A mismatched target DNA. The current densities were obtained from the subtracted shoulder peak where the OSWVs were performed in 0.05 M phosphate buffer containing 1.0 M NaCl (pH 7.0) at step of 4 pulse amplitude of 25 mV and frequency of 10 Hz in the presence of 4 µM complementary target DNA and 25 µM AQMS. A significant diminution in current was observed after exposure to single-base mismatches.

the time taken for all the bases in the duplex to form WatsonCrick pairs. Note that the 2-h hybridization time we report here, although consistent with other electrochemical DNA chargetransfer-based biosensors,15,23 is contrary to the hybridization time of a few minutes reported using techniques such as quartz crystal 2142

Analytical Chemistry, Vol. 78, No. 7, April 1, 2006

microbalance (QCM)31 and surface plasmon resonance (SPR).4,32 We believed that these surface binding techniques (QCM and SPR) are probing the kinetics of bringing the target DNA in contact with the surface rather than the formation of a completed DNA duplex. We have demonstrated this previously by performing kinetics measurements on different DNA recognition interfaces using both the electrochemical and QCM approaches.5 Detection of a Single-Base Pair Mismatch. In the previous studies of DNA biosensors based on long-range charge transfer, direct evidence for electrons transferring through the DNA was attained by demonstrating the modified electrodes could discriminate between a complementary target sequence and one containing a single-base pair mismatch by attenuation in current. Although the in situ assay is simpler and faster than the previously published sequential assay,16 the ability to detect mismatches should not be compromised. The detection of mismatches with the in situ assay was investigated using target DNA sequences 3 (a C-A mismatch) and 4 (a G-A mismatch). Figure 3 shows background-subtracted OSWVs obtained using the in situ assay upon exposure to 4 µM target DNA 2 (Figure 3ii), 3 (Figure 3iii), and 4 (Figure 3iv). The figure clearly illustrates the ability of AQMS to differentiate between a perfectly matched and a mismatched target DNA where approximately 93 and 73% diminutions in current were observed for C-A and G-A mismatched target DNA, respectively, compared to the complementary target. The hybridization efficiencies between the probe and the complementary, C-A and G-A mismatched target DNA were calculated via a chronocoulometry approach24 and were found to be 78 ( 4 (n ) 4, 95% confidence interval), 75 ( 5 (n ) 4, 95% confidence interval), and 75 ( 4% (n ) 4, 95% confidence interval), respectively. Therefore, the diminution in current observed with the C-A and G-A mismatches did not result from lower hybridization efficiency between the immobilized probe and the C-A and G-A mismatched target sequences. Instead, the decrease in current is caused by a decreased charge transfer from the intercalated AQMS to the electrode surface through the perturbed DNA duplex. The ability of AQMS to detect both C-A and G-A singlebase mismatches is consistent with our previous study using the sequential assay.26 Furthermore, the sensitivity of the shoulder peak to mismatches provides strong evidence for long-range charge transfer induced by the intercalated AQMS. The ability to differentiate between a complementary target DNA and a C-A mismatch target sequence is also consistent with the observations where daunomycin or methylene blue19 is used to investigate the C-A mismatch via a similar charge-transfer approach. However, the ability of an in situ assay to detect a G-A mismatch by a large diminution in current represents an improvement in selectivity relative to previous studies.19 Both the C-A and G-A single-base pair mismatches occur in the middle of the base sequence, and therefore, the intercalator must intercalate above the mismatch in order to be able to perturb the charge transfer through the DNA duplex. We propose that the AQMS intercalated at the very top of the DNA in the in situ assay because of charge repulsion as seen previously with the sequential AQDS system.16 (31) Satjapipat, M.; Sanedrin, R.; Zhou, F. M. Langmuir 2001, 17, 7637. (32) Tsuruoka, M.; Murano, S.; Okada, M.; Ohiso, I.; Fujii, T. Biosens. Bioelectron. 2001, 16, 695-699.

Figure 4. Cathodic current density vs complementary target DNA concentration (/ M) plot. The inset showed the resulting background-subtracted shoulder peak as seen in the OSW voltammogram’s peak of Figure 2. The current densities were obtained from the subtracted shoulder peak where the OSWVs were performed in 0.05 M phosphate buffer containing 1.0 M NaCl (pH 7.0) at step of 4 mV, pulse amplitude of 25 mV, and frequency of 10 Hz in the presence of 4 µM complementary target DNA and 25 µM AQMS.

Effect of Target DNA Concentration. Next, the analytical performance of the DNA biosensor was assessed using different concentrations of target DNA ranging from 0.5 nM to 4 µM. The target DNA calibration curve, measured using backgroundsubtracted OSWVs, is shown in Figure 4. The target DNA calibration followed a Langmuir-like adsorption curve. Therefore, the binding constant Ka for the DNA hybridization was calculated using the linearized form of the Langmuir model as defined by

[T] [T] 1 + ) I Isat KaIsat

(1)

where [T] is the target concentration, I the current density, Isat the saturated current density, and Ka the binding constant. The linearized Langmuir plot of the isotherm gives the equation [T]/ I ) 2.4964e - 01[T] + 5.48202e - 8, and hence, the saturated current density of complementary target is 4.01 µA cm-2. This is consistent with the calibration curve (Figure 4) where the current density was found to plateau at 3.8 µA cm-2. The Ka values calculated using eq 1 for in situ detection (4.55 × 106 M-1) is 14 times greater than the sequential assay (3.16 × 105 M-1).16 We attribute this observation to perturbation of the equilibrium in the sequential approach as the measurement is performed in the absence of target DNA in solution. Besides the greater binding constant, the sensitivity for in situ detection (10.45 × 107 µA cm-2 M-1) is also greater than the sequential system (5.36 × 106 µA cm-2 M-1)16 using the same recognition interface. The greater Ka values, and the higher sensitivity, obtained with the in situ assay results in a decrease in the lowest detectable concentration of target DNA from 500 nM (for the sequential assay) down to 10

nM. The reproducibility of the in situ assay was obtained through independent measurements of four individual DNA-modified electrodes and has a relative standard deviation between 2.5 and 9.5% (n ) 4, 95% confidence interval) over the concentration range of the calibration curve. Further Improvement. With the new in situ approach, improvements with regard to assay time, detection limit, and sensitivity are achieved without compromising the selectivity advantage of the long-range charge-transfer strategy. With regards to the assay time, using the in situ approach, the assay time is 2 h compared with the 5-8-h assay time for the sequential assay.16 Further decrease in the in situ assay time can only be achieved by decreasing the hybridization time since this is the rate limiting step.5 With the DNA interface used (P1/MCH) in this study, the MCH diluent (6C) is longer than the length of the probe linker (3C); see Scheme 1. Thus, some of the DNA bases at the proximal end of the probe ss-DNA may be hindered by the diluent layer during hybridization. An alternate DNA recognition interface (P1/ MCE) was investigated, where the length of the diluent is less than the link to the DNA. A decrease in assay time is observed with the current reaching steady value within 1 h upon exposure to 100 nM complementary target 2 via the in situ approach (data not shown). With regard to detection limit, a previous study has also shown that the length of the diluent layer affects the thermodynamics of hybridization and, hence, the hybridization efficiency.5 Using the P1/MCE DNA recognition interface, a Ka value of 10.9 × 106 M-1 was obtained (from a calibration curve similar to Figure 4), which is higher than the Ka value of P1/MCH interface. As a result of the increase in Ka, the lowest detectable concentration of target DNA using the new interface is 0.5 nM. This detection limit is in the same order of magnitude as many of the other reported electrochemical DNA hybridization biosensors17 but is higher than the detection limits achieved using nanoparticle approaches developed by Wang and co-workers.33,34 The advantage of the in situ assay over other electrochemical biosensor technologies, however, is simplicity and selectivity, making it more compatible with integrating the biosensor with an on-chip PCR reactor. CONCLUSION Following our initial investigation on the transduction of DNA hybridization via long-range charge transfer conducted in sequential steps, we have made significant improvements with regard to sensitivity and ease of use via a single-step in situ electrochemical approach. Using the in situ approach, the DNA biosensor is able to detect target DNA in the subnanomolar range within 1 h. Furthermore, the in situ detection scheme is also able to differentiate between complementary, noncomplementary target DNA and even target DNA with single-base pair mismatches, including the most thermodynamically stable G-A mismatch, without requiring any additional stringency steps. This new approach also has the ability of studying biological processes in real time and thus allows the kinetic processes to be monitored. The good sensitivity, excellent selectivity, and simplicity of use (33) Wang, J.; Xu, D. K.; Kawde, A. N.; Polsky, R. Anal. Chem. 2001, 73, 55765581. (34) Wang, J.; Liu, G. D.; Merkoci, A. J. Am. Chem. Soc. 2003, 125, 32143215.

Analytical Chemistry, Vol. 78, No. 7, April 1, 2006

2143

of the DNA biosensor makes it more compatible with integrating with on-chip PCR reactors than other DNA biosensors we are aware of. ACKNOWLEDGMENT We thank the University of New South Wales and the Australian Research Council (ARC) for funding aspects of this

2144

Analytical Chemistry, Vol. 78, No. 7, April 1, 2006

research. We also thank Dr. Katharina Gaus for advice on the preparation of the manuscript. Received for review May 25, 2005. Accepted January 30, 2006. AC0509096