Electrochemical DNA Sensors Based on Enzyme Dendritic

Anal. Chem. 2004, 76, 3132-3138

Electrochemical DNA Sensors Based on Enzyme Dendritic Architectures: an Approach for Enhanced Sensitivity Elena Domı´nguez, O Ä scar Rinco´n, and Ara´ntzazu Narva´ez*

Department of Analytical Chemistry, Faculty of Pharmacy, University of Alcala´ , Alcala´ 28871, Spain

The modification of enzymes with multiple single-stranded oligonucleotides opens up a new concept for the development of DNA sensors with enhanced sensitivity. This work describes the generation of reporter sequences labeled with an enzyme for the demonstration of their ability to specifically hybridize and to permit signal amplification by successive hybridization steps. The synthetic pathway for the labeling of GOx with oligonucleotide sequences is based on the oxidation of the glycosidic residues of the enzyme and their covalent binding with 5′-end aminemodified oligonucleotides. Spectrophotometric characterization of these functionalized sequences results in an average number of three linked oligonucleotides per enzyme molecule. Their specificity is demonstrated in both a direct and a sandwich-type hybridization assay. The transduction of the enzyme-linked DNA sensors is based on self-assembled multilayers, including a chemically modified anionic horseradish peroxidase electrochemically connected to a water-soluble cationic poly[(vinylpyridine)Os(bpy)2Cl] redox polymer in an electrostatic ordered assembly. The sensing layer is constructed by the covalent binding of the DNA probe over the redox polymer through the 3′-phosphate group, enabling the capture of the target sequence. Upon addition of glucose, hybridization results in the production of H2O2, which readily diffuses to the electrocatalytic assembly, giving rise to a cathodic current at 100 mV vs Ag/AgCl. Hybridization is always performed at room temperature, and after 30 min of incubation, an amperometric response is obtained that is proportional to DNA concentration. The simultaneous sandwich assay enables the quantification of a free-label 44-mer oligonucleotide at 1 nM concentration. Signal amplification is realized by a new hybridization step over the free sequences, giving rise to a dendritic architecture that accumulates enzyme molecules per hybridization event. In the past decade, intensive research has been performed that is aimed at the transduction of hybridization events in a more efficient manner than the time-consuming and laborious traditional DNA sequencing protocols. A multitude of transduction schemes is nowadays available, offering different merits as well as concerns, not only specificity and sensitivity, but also practicability. In * Corresponding author: [email protected].

3132 Analytical Chemistry, Vol. 76, No. 11, June 1, 2004

general, electrochemical transduction offers high sensitivity, costand time-effectiveness, minimal power requirements, and the required portability for a wide-scale or decentralized genetic testing in areas of application as diverse as medicine and environmental and forensic science.1 The native electroactivity of the nucleic acids and its different electrochemical influence on the DNA configuration (i.e., accessibility of the residues) opens up the simplest and most direct strategy to transduce hybridization events.2 However, this labelfree protocol suffers from the irreversible redox behavior of the nucleobases and the use of high overpotentials, resulting in significant background currents and, thus, limited sensitivity.3,4 The electrocatalytic oxidation of guanine moieties within DNA by redox mediators5-7 or through the oxidation products of the adenine residues by NADH,8 although very elegant routes to overcome these problems, still do not result in a sequence independent approach. Hybridization has also been directly transduced through the dielectric properties9 or charge migration through DNA,10 extending rather formally the possibilities of labelfree transduction schemes. Electrochemical detection of double-strand formation can also be realized indirectly by the use of redox indicators11-13 or threading intercalators;14,15 however, the selectivity and sensitivity of this approach are very much dependent on the intrinsic characteristics of the indicator.1 Alternatively and in search of a more versatile configuration, either the target sequence or a (1) Wang, J. Anal. Chim. Acta 2002, 469, 63-71. (2) Palecek, E.; Fojta, M. Anal. Chem. 2001, 73, 74A-83A. (3) Palecek, E.; Jelen, F. Crit. Rev. Anal. Chem. 2002, 32, 261-270. (4) Drummond, T. G.; Hill, M. G.; Barton, J. K. Nat. Biotechnol. 2003, 21, 11921199. (5) Johnston, D. H.; Glasgow, K. C.; Thorp, H. H. J. Am. Chem. Soc. 1995, 117, 8933-8938. (6) Yang, I. V.; Thorp, H. H. Anal. Chem. 2001, 73, 5316-5322. (7) Armistead, P. M.; Thorp, H. H. Bioconjugate Chem. 2002, 13, 172-176. (8) Lobo-Castanon, M. J.; Miranda-Ordieres, A. J.; Tun ˜o´n-Blanco, P.; De-losSantos-Alva´rez, P. Anal. Chem. 2002, 74, 3342-3347. (9) Berney, H.; West, J.; Haefele, E.; Alderman, J.; Lane, W.; Collins, J. K. Sens. Actuators, B 2000, 68, 100-108. (10) Boon, E. M.; Ceres, D. M.; Drummond, T. G.; Hill, M. G.; Barton, J. K. Nat. Biotechnol. 2000, 18, 1096-1100. (11) Millan, K. M.; Saraullo, A.; Mikkelsen, S. R. Anal. Chem. 1994, 66, 29432948. (12) Wang, J.; Rivas, G.; Cai, X. Electroanalysis 1997, 9, 395-398. (13) Marrazza, G.; Chianella, I.; Mascini, M. Anal. Chim. Acta 1999, 387, 297307. (14) Takenaka, S.; Yamashita, K.; Takagi, M.; Uto, Y.; Kondo, H. Anal. Chem. 2000, 72, 1334-1341. (15) Takagi, M. Pure Appl. Chem. 2001, 73, 1573-1577. 10.1021/ac0499672 CCC: $27.50

© 2004 American Chemical Society Published on Web 05/01/2004

reporter probe can be labeled with electroactive markers that result in redox currents when they are brought close to the electrode surface upon hybridization. Different alternatives have been described for the modification of the oligonucleotide, ranging from postsynthetic labeling with derivatized sequences16 to the incorporation of the redox label at various positions during the chemical synthesis of the oligonucleotide.17 However, the use of redox-labeled oligonucleotides does not offer direct means of signal amplification. Willner et al. have described a polymeraseinduced generation of a redox-active DNA replica of the target DNA. Further coupling of the redox-active DNA replica with glucose oxidase enhances sensitivity through bioelectrocatalytic cascades.18 Many applications of DNA sensing require extremely high sensitivity because few hybridization events, corresponding to the extremely small number of target sequences, have to be detected. Most commonly, signal amplification of sensing devices is realized by the enzymatic catalysis of the transformation of multiple substrate molecules. In 1996, Heller pioneered this idea for the detection of hybridization by labeling the target oligonucleotide with peroxidase.19 Three years later and in a simple capture assay, the authors demonstrated a single-base mismatch in an 18-mer oligonucleotide at high temperature.20 The enzyme, soybean peroxidase, was electrochemically communicated with a redox polymer-coated microelectrode. A practical improvement of this idea was further demonstrated with a sandwich assay because, unlike the previous format, the target sample does not require chemical modification.21 β-Galactosidase,22 alkaline phospatase23,24 and glucose dehydrogenase (PQQ dependent)25 have also been used for oligonucleotide labeling with electrochemical transduction of hybridization. Although the labeling of either the target or the reporter sequence with an enzyme can achieve selectivity and sensitivity in a single hybridization step,2 further enhancement of sensitivity could be achieved if various (optimally, many) enzyme molecules could account per hybridization event. This has been approached by Willner’s group using liposomes labeled with biotin and peroxidase and detecting specific sequences through the formation of an insoluble product measured by Faradaic impedance spectroscopy.26 Using this technique, the same group has described an amplified detection of DNA based on alkaline phosphatase functionalized with multiple oligonucleotides.24 Other approaches for DNA signal amplification include the use of (16) Ihara, T.; Nakayama, M.; Murata, M.; Nakano, K.; Maeda, M. Chem. Commun. 1997, 17, 1609-1610. (17) 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. (18) Patolsky, F.; Weizmann, Y.; Willner, I. J. Am. Chem. Soc. 2002, 124, 770772. (19) de Lumley-Woodyear, T.; Campbell, C. N.; Heller, A. J. Am. Chem. Soc. 1996, 118, 5504-5505. (20) Caruana, D. J.; Heller, A. J. Am. Chem. Soc. 1999, 121, 769-774. (21) Campbell, C. N.; Gal, D.; Cristler, N.; Banditrat, C.; Heller, A. Anal. Chem. 2002, 74, 158-162. (22) Wang, J.; Kawde, A. N.; Musameh, M.; Rivas, G. Analyst 2002, 127, 12791282. (23) Aguilar, Z. P.; Fritsch, I. Anal. Chem. 2003, 75, 3890-3897. (24) Patolsky, F.; Lichtenstein, A.; Willner, I. Chem.sEur. J. 2003, 9, 11371145. (25) Ikebukuro, K.; Kohiki, Y.; Sode, K. Biosens. Bioelectron. 2002, 17, 10751080. (26) Alfonta, L.; Singh, A. K.; Willner, I. Anal. Chem. 2001, 73, 91-102.

nanoparticle-labeled reporter sequences27-30 and branched DNA.31 They all have demonstrated successful ways for signal amplification, although it would be desirable to have a signal amplification pathway with no need of intensive synthetic work and more simple means of transduction. This work describes the use of multiple oligonucleotide sequences linked to an enzyme for the detection of specific hybridization. The rationale behind this concept relies on the ability of these sequences to hybridize successively and, thus, the building up of a dendritic architecture that accumulates catalytic molecules for enhancement of sensitivity. As will be described, the general concept of enzyme-linked DNA sensors can be readily implemented with this approach, enabling an easy and direct way of signal amplification with the additional advantage of minimal synthetic work because only two different and complementary oligonucleotide sequences labeled with the enzyme are required. Glucose oxidase constitutes the basis of this dendritic enzyme concept being transduced through the diffusion of hydrogen peroxide into a bioelectrocatalytic assembly based on polyelectrolyte multilayers. EXPERIMENTAL SECTION Reagents. Sodium salt of 3-mercapto-1-propane sulfonic acid (MPS), 2-4,dinitrofluorobenzene, sodium periodate, 1-ethyl-3(dimethylaminopropyl) carbodiimide (EDC), and streptavidin from Streptomyces avidinii salt-free powder were purchased from Sigma. Sodium cyanoborohydride was obtained from Fluka. The cationic poly[(vinylpyridine)Os-(bpy)2Cl] redox polymer partially quaternized with bromoethylamine (RP) was synthesized as described elsewhere.32 Horseradish peroxidase (EC 1.11.1.7, 290 U mg-1 solid; type VI), glucose oxidase, from Aspergillus niger (EC 1.1.3.4, 1687 U mL-1 in 0.1 M sodium acetate buffer solution) and glucose oxidase biotinamidocarpoyl-labeled, from Aspergillus niger, (EC 1.1.3.4, 217 U mg-1 protein, 6.9 mol biotin/mol protein) were purchased from Sigma. D-Glucose, hydrogen peroxide (30%), and all buffer salts were obtained from Merck. All other chemicals used were of analytical grade. Water was obtained with a Millipore Milli-Q system. The oligonucleotide sequences were synthesized by TIB MOLBIOL Syntheselabor (Berlin, Germany). The capture probes, termed as AP and BP (3′-end phosphate modified) were 19-mer oligonucleotides for the direct and the sandwich-type assays, respectively. The reporter sequences, AT-amine and XT-amine were also 19-mer oligonucleotides with a 5′-end primary amine for its chemical modification with GOx, as detailed below. Both sequences were complementary and noncomplementary, respectively, to AP capture probe and to positions 25-44 at the 5′ end of the AB target sequence (44-mer). Other oligonucleotides (44-mer) were ABX used as negative control (noncomplementary) and AB4, (27) Patolsky, F.; Ranjit, K. T.; Lichtenstein, A.; Willner, I. Chem. Comun. 2000, 12, 1025-1026. (28) Wang, J.; Li, J. H.; Baca, A. J.; Hu, J. B.; Zhou, F. M.; Yan, W.; Pang, D. W. Anal. Chem. 2003, 75, 3941-3945. (29) Mirkin, C. A.; Letsinger, R. L.; Mucic, R. C.; Storhoff, J. J. Nature 1996, 382, 607-609. (30) Elghanian, R.; Storhoff, J. J.; Mucic, R. C.; Letsinger, R. L.; Mirkin, C. A. Science 1997, 277, 1078-1081. (31) Collins, M. L.; Irvine, B.; Tyner, D.; Fine, E.; Zayati, C.; Chang, C.; Horn, T.; Ahle, D.; Detmer, J.; Shen, L. P.; Kolberg, J.; Bushnell, S.; Urdea, M. S.; Ho, D. D. Nucleic Acids Res. 1997, 25, 2979-2984. (32) Katakis, I.; Heller, A. Anal. Chem. 1992, 64, 1008-1013.

Analytical Chemistry, Vol. 76, No. 11, June 1, 2004

3133

Table 1. Oligonucleotide Sequences name

oligonucleotide sequence (5’-3’)

AP BP AT-amine XT-amine AP-b AT-b AB ABX AB4

GGCGTAATAGCGAAGAGGC CGTTGCAAGCCATACAGCG GCCTCTTCGCTATTACGCC CGGCTCGTATGTTGTGTGG GGCGTAATAGCGAAGAGGC GCCTCTTCGCTATTACGCC CGCTGTATGGCTTGCAACGAAAAAAGGCGTAATAGCGAAGAGGC TCTGATACTGAGCCTGGACAAAAAACTGACCAGGTTAGTCCCGA CGCTGTATGGCTTGCAACGAAAAAAGGCCTAATAACTAAGATGC

including a four-base-mismatched sequence for the sandwich-type assay. AP-b and AT-b were 5′-end-biotinylated oligonucleotides (19mer) for perfect match and negative control, respectively, with AT and used for signal amplification. All the oligonucleotide sequences are listed in Table 1. Equipment. Electrochemical measurements were performed with a computer-controlled BAS CV-50W voltammetric analyzer (Bioanalytical Systems, West Lafayette, IN) using a conventional three-electrode electrochemical cell. Potentials were measured versus a potassium chloride-saturated silver/silver chloride electrode (Ag/AgCl, KClsat) and a platinum wire as counter electrode. Spectrophotometric measurements were performed with a Shimadzu UV-160A recording spectrophotometer (Shimadzu Corporation, Kyoto, Japan). Electrode Modification. Au wires (0.5-mm diameter, geometrical area 0.08 cm2, 99.99% purity), used as working electrodes, were treated with freshly prepared piranha solution (7:3 mixture of concentrated H2SO4 and 30% H2O2; caution: piranha reacts violently with organic compounds) for 30 min and with a boiling saturated KOH solution for 2 h. The cleaned wire electrodes were stored in concentrated H2SO4. Prior to use, the electrodes were dipped in a concentrated HNO3 solution for 10 min, and thoroughly washed with water. The cleaned gold surface was checked by cyclic voltammetry in 1 M H2SO4, between 0 and 1.7 V until an ideal redox wave of polycrystalline gold in H2SO4 was obtained. The negatively charged surface (Au/MPS) was prepared by immersing the cleaned gold wire into a 1 mM ethanolic solution of MPS for 12 h and then rinshing with pure ethanol. Assembly of the Bioelectrocatalytic Interface (Au/MPS/ RP/HRPm/RP). The buildup of the bioelectrocatalytic interface was based on sequential deposition of different polyelectrolytes onto the negatively charged electrode surface (Au/MPS) by alternate immersion in the corresponding aqueous solutions at room temperature and under stirring conditions. The redox (RP) and catalytic (HRPm) layers were deposited for 2 h from a 100 µg mL-1 polymer aqueous solution and from a 1 mg mL-1 enzyme solution in 0.1 M phosphate buffer pH 7.0, respectively. Each deposition was followed by thorough washing. A second layer of RP was deposited to ensure a more efficient electrochemical communication of HRP and to expose amino groups prone to link the DNA (see bellow). Electrostatic self-assembly of the different layers was guaranteed by charge reversal with the positively charged redox polymer and the anionic peroxidase. This enzyme was prepared, as previously described, from commercially available horseradish peroxidase.33 The electrocatalytic efficiency of this assembly was characterized by cyclic voltammetry. 3134 Analytical Chemistry, Vol. 76, No. 11, June 1, 2004

modification 3’-phosphate 5’-amine 5’-amine 5’-biotin 5’-biotin 4 mismatched

Preparation of the DNA Sensing Layer. The probe oligonucleotides (AP or BP) were covalently linked to the previously described bioelectrocatalytic assembly through phosphoramide formation between amino and phosphate groups of the redox polymer and the oligonucleotides, respectively.34 A 50 µg mL-1 (8 µM) oligonucleotide buffered solution containing 0.1 M carbodiimide (EDC) was in contact with the modified electrodes for 12 h and under vigorous stirring. The noncovalently bound oligonucleotides were removed by thorough washing with buffer solution. Preparation of the Enzyme Functionalized Oligonucleotides. The 19-mer AT and XT amine-modified oligonucleotides were labeled with the enzyme (GOx) as follows. First, 2 mL of a commercial GOx solution (5.7 mg mL-1) was ultrafiltered through 30 000-MWCO membranes (Millipore) and extensively washed with 0.01 M carbonate buffer, pH 8.7. The enzyme was recovered in ∼3 mL, and then 100 mL of 1% 2,4-dinitrofluorobenzene in ethanol was added. This mixture was gently rotated for 1 h at room temperature, enabling the chemical protection of the enzyme amine groups. Then 22 g of sodium periodate was added to the solution for oxidation of enzyme glycosidic residues during 30 min in darkness and rotating conditions. The excess of periodate was eliminated by addition of 1 mL of 0.16 M aqueous solution of ethylene glycol. This activated enzyme solution was ultrafiltered through 30 000-MWCO membranes and extensively washed with carbonate buffer until complete elimination of the free aldehydes, as shown by testing with the 3-methyl-2-benzothiazolinone hydrazone assay.35 The concentration of the resulting oxidized enzyme was determined by the BioRad protein assay. Oligonucleotides were then bound to this activated GOx solution through Shiff base formation. A 5-fold molar excess of amine-modified oligonucleotides was added to 0.5 mL of enzyme solution and rotated for 3 h in darkness. Reduction of the formed imine groups was achieved by addition of 5 µL of 1 M sodium cyanoborohydride in 1 M NaOH to the DNA-enzyme solutions and left to react for 12 h at 4 °C. Finally, the samples were ultrafiltered by centrifugation at 4000 rpm through 30 000-MWCO membranes and extensively washed with 0.1 M phosphate buffer pH 7.0 until complete elimination of unattached oligonucleotides. This was tested spectrophotometrically until disappearance of the DNA peak in the ultrafiltration wastes, which was ascribed to a successful elimination of free (33) Narva´ez, A.; Sua´rez, G.; Popescu, I. C.; Katakis, I.; Domı´nguez, E. Biosens. Bioelectron. 2000, 15, 43-52. (34) Fan, C. H.; Li, G. X.; Gu, Q. R.; Zhu, J. Q.; Zhu, D. X. Anal. Lett. 2000, 33, 1479-1490. (35) Sawicki, E.; Hauser, T. R.; Stanley, T. W.; Elbert, W. Anal. Chem. 1961, 33, 93-96.

oligonucleotides. The resulting GOx-labeled oligonucleotides, AT-GOx and XT-GOx, were characterized spectrophotometrically and stored at 4 °C until use. In addition, an equally treated GOx but in the absence of DNA was also prepared (GOxCTRL) for control experiments and evaluation of nonspecific adsorption. Amperometric Detection of Hybridization. Amperometric detection of hybridization events with the previously described Au/MPS/RP/HRPm/RP-ssDNA electrodes has been evaluated with two different assay formats. To check the performance of the dendritic enzyme concept, a direct hybridization assay was developed. In this case, the functionalized oligonucleotides were directly incubated on the electrode modified with the capture probe (AP). Two controls were also tested, a noncomplementary functionalized sequence (XT-GOx) and a unmodified GOx (GOxCTRL). Signal amplification was evaluated by a subsequent hybridization step with a perfectly matched biotinylated sequence. The electrodes were then incubated for 1 h in the presence of a GOxbiotin-streptavidin complex for signal acquisition. For the preparation of this complex, a solution of biotinylated GOx (2.5 mg L-1) was mixed with streptavidin (10 mg L-1) for 40 min at 4 °C under stirring in 0.1 M phosphate buffer pH 7.0 containing 5 mM NaCl. In the sandwich assay, the capture probe was BP, and hybridization was performed in one single step by incubation of the target sequences (AB, AB4, or ABX as control) with the reporter dendritic sequence (AT-GOx). All hybridization reactions were performed at room temperature in 0.1 M (or 0.01 M) phosphate buffer pH 7.2 and under stirring conditions in a 500-µL-volume cell. After hybridization, the electrodes were extensively washed with buffer solution for elimination of the nonspecific adsorbed oligonucleotides. The steady-state amperometric measurements were recorded at 0.1 V vs Ag/AgCl, KClsat at room temperature in 4 mL of 0.1 M phosphate buffer pH 6.0. After current stabilization, a glucose solution was injected up to a concentration of 25 mM that resulted in the production of H2O2 that readily diffuses to the RP/HRPm/ RP layers, giving rise to a cathodic current that detects hybridization. To ensure that this current was not limited by the electrochemical communication between the modified HRP and the redox polymer layers, the cathodic current obtained after addition of a saturating H2O2 concentration (400 µM) was also recorded. RESULTS In 1995, the general utility of polyelectrolyte multilayers (PEMs) based on electrostatic interactions for the assembly of proteins was first demonstrated by Lvov et al.36 Since then, most analytical applications refer to the use of this technology for the assembly of catalytic proteins33 and antibodies,37 enabling the construction of reliable sensing devices based on amperometric transduction schemes. In these assemblies, the use of positively charged osmium-based redox chemistry has proved to fulfill the analytical requirements for the nondiffusional electrochemical communication with peroxidases.33 Nevertheless, when this configuration scheme is directly used for the transduction of nonfaradic events (immuno or hybridization reactions), the reporter enzyme, peroxidase, has to be electrically connected with the redox polymer through an insulator biorecognition layer. This has (36) Lvov, Y.; Ariga, K.; Ichinose, I.; Kunitake, T. J. Am. Chem. Soc. 1995, 117, 6117-6123. (37) Danilowicz, C.; Manrique, J. M. Electrochem. Commun. 1999, 1, 22-25.

not limited the demonstration of hybridization reactions with oligonucleotides labeled with peroxidase.19,38,39 However, this scheme of transduction based on peroxidase-labeled oligonucleotides or any other affinity molecule does not permit the introduction of new peroxidase catalytic cascades because the imposed distance to the electrode surface may ruin the electrocatalytic communication. The rationale behind this work relies on the use of enzymes modified with multiple oligonucleotides enabling direct transduction of hybridizations events and also direct means of signal amplification by subsequent hybridization steps. For this reason, GOx has been chosen as reporter enzyme, resulting in the production of hydrogen peroxide that readily diffuses through the sensing and highly permeable transducing polyelectrolyte multilayers.33 Consequently, the efficiency of this configuration depends on the electrocatalytic conversion of a diffusional product not being limited by the distance to the electrode surface. A scheme of this concept is depicted in Figure 1A for the direct hybridization showing the dendritic architecture that results after a second hybridization step. The sandwich configuration is shown in Figure 1B. The electron-transfer reactions in the electrocatalytic layers are summarized in Figure 1C. Performance of the Bioelectrocatalytic Interface. The bioelectrocatalytic interface includes a layer of anionic horseradish peroxidase sandwiched between two monolayers of osmium-based redox polymer to ensure a more efficient electrocatalytic communication. Both osmium redox layers showed a quasireverible electrochemical behavior with surface coverage, estimated by integration of the charge passing through the cathodic peak at 50 mV s-1, of 1.2 × 10-10 and 3.6 × 10-10 mol cm-2 for the first and second layers, respectively (Figure 2A). The relative standard deviation remained below (5% (n ) 12), indicating not only a highly reproducible bioelectrocatalytic interface but also the provision of a rather constant surface, exposing anchoring amine groups for further modification with oligonucleotides. Previous experiments show that in the absence of the catalytic layer, the electrochemical reduction of hydrogen peroxide with the Os polymer was negligible.33 Figure 2B shows the electrocatalytic currents obtained with one and two layers of redox polymer after addition of a saturating concentration of hydrogen peroxide. As seen in the figure, the adsorption of a second redox layer improves the electrocatalytic efficiency, resulting in a better communication of the sandwiched peroxidase. Although much higher current densities have been reported when this enzyme is entrapped in a thick redox polymer film,40 this ordered assembly seems more suitable for this work because it renders a more permeable structure, ensuring an efficient diffusion and conversion of the catalytic product. Detection of DNA Hybridization with Functionalized Oligonucleotides. The feasibility of the concept of dendritic enzymes for the transduction of DNA hybridization demands the modification of the catalytic molecule with several oligonucleotide sequences and the demonstration of their specific hybridization. If both requirements are accomplished, then signal amplification can (38) Zhang, Y. C.; Kim, H. H.; Mano, N.; Dequaire, M.; Heller, A. Anal. Bioanal. Chem. 2002, 374, 1050-1055. (39) Zhang, Y. C.; Kim, H. H.; Heller, A. Anal. Chem. 2003, 75, 3267-3269. (40) Calvente, J. J.; Narva´ez, A.; Domı´nguez, E.; Andreu, R. J. Phys. Chem. B 2003, 107, 6629-6643.

Analytical Chemistry, Vol. 76, No. 11, June 1, 2004

3135

Figure 1. Representation of the amperometric DNA sensor using functionalized oligonucleotides. (A) Direct capture assay with the amplification step that results in an enzyme dendritic architecture. (B) Sandwich-type assay configuration for the detection of label-free sequences. (C) Scheme of the electron-transfer pathway for the electrocatalysis of H2O2 produced by glucose oxidase (GOx).

Figure 2. (A) Voltammetric response of two monolayers of cationic redox polymer in the electrode: Au/MPS/RP/HRPm/RP. Experimental conditions: 0.1 M phosphate buffer (pH 7.0); starting potential, +600 mV vs Ag/AgCl; KClsat; and scan rate, 50 mV s-1. (B) Dependence of the electrocatalytic response on the number of redox polymer layers: (1) Au/MPS/RP/HRPm/RP shown as control when no H2O2 is added, (2) Au/MPS/RPm/HRP, and (3) Au/MPS/RP/HRPm/RP. Experimental conditions: addition of 0.4 mM H2O2 in 0.1 M phosphate buffer (pH 7.0); starting potential, +600 mV vs Ag/AgCl; KClsat; and scan rate, 2 mV s-1.

be easily achieved by subsequent hybridization, enabling the introduction of new enzyme molecules. The oligo dendritic enzymes, AT-GOx and XT-GOx, were characterized spectrophometrically together with an unmodified GOx (GOxCTRL). The corresponding UV spectra clearly showed the presence of two peaks at 260 and 280 nm for AT-GOx and XT-GOx, while the first peak was not present in the GOxCTRL, demonstrating the successful binding of ssDNA to the enzyme (data not shown). The concentration of protein was 2.80 ( 0.01, 2.97 ( 0.01, and 2.16 ( 0.01 µM, for AT-GOx, XT-GOx, and GOxCTRL, respectively. The ratio between the absorbance at 260 and 280 nm was used to determine the concentration of DNA in the samples, resulting in 9.1 ( 0.3 and 8.6 ( 0.2 µM for AT-GOx and XT-GOx, respectively. From these values, it was estimated that approximately three of the initial five sequences were successfully attached per enzyme molecule. When these functionalized sequences were incubated with the ssDNA modified electrodes, Au/MPS/RP/HRPm/RP-AP, a cathodic current was obtained upon addition of glucose. Figure 3A shows the steady-state response profile of the obtained currents after 30 min of incubation at room temperature with 2.8 nM (protein concentration) of AT-GOx or XT-GOx. The specificity of 3136 Analytical Chemistry, Vol. 76, No. 11, June 1, 2004

Figure 3. (A) Amperometric response profiles for the direct hybridization assay of AP probe modified electrodes incubated with 2.8 nmol L-1 of AT-GOx (1) or XT-GOx (2) upon addition of 25 mM glucose. (B) Increase of the electrocatalytic current after addition into the cell of 0.4 mM H2O2. Hybridization conditions: 30 min under stirring in phosphate buffer 0.1 M (pH 7.2) at room temperature.

the signal is demonstrated by the almost undetectable signal of the noncomplementary XT-GOx, while the hybridization of AT-GOx results in a H2O2 flux that is converted into electron fluxes through the electrocatalytic layers. Also in Figure 3B, it is clear that the hybridization response is not limited by the electrocatalysis of H2O2 because a significant increase in the cathodic current was obtained when a saturating concentration of hydrogen peroxide (400 µM) was directly injected. This is important to ensure the transduction of new molecules of GOx in the dendritic architecture. Figure 4 shows the dependence of the amperometric signal on the hybridization time. Under conditions identical to those described in Figure 3 and varying the time between 2 min and 1 h, the cathodic current increased with the hybridization time up to a plateau at 30 min. It is somewhat difficult to compare this time dependence with other configurations reported in the literature, because hybridization rates depend on oligonucleotide length, temperature, and salt concentration.41 Nevertheless, this time seems to be twice as long as that required with soybean peroxidase (SBP)-labeled oligonucleotides.21 Most likely, the intrinsic nature of the functionalized sequences used in this work (i.e., three oligonucleotides per enzyme molecule) together with (41) Dai, H.; Meyer, M.; Stepaniants, S.; Ziman, M.; Stoughton, R. Nucleic Acids Res. 2002, 30, e86.

Figure 4. Dependence of the amperometric current on the hybridization time. Experimental conditions as in Figure 3. Standard deviations are obtained with a minimum of three electrodes.

Figure 5. Dependence of the amperometric current on the concentration of the functionalized oligonucleotides in direct assay: (l) response for the complementary sequence AT-GOx; (m) response for the noncomplementary sequence XT-GOx. Hybridization conditions: 30 min under stirring in 0.1 M phosphate buffer (pH 7.2) at room temperature. Standard deviations were obtained with a minimum of three electrodes.

the higher molecular weight of GOx, as compared to SBP, and the fact that no spacer is used between GOx and the oligonucleotide sequences, can contribute to steric hindrances diminishing hybridization rates. The response obtained in these assays is also proportional to the concentration of functionalized oligonucleotides. As seen in Figure 5, increasing concentrations of AT-GOx results in a concomitant current increase up to 2 µA cm-2. Parallel experiments performed with the noncomplementary functionalized sequence (XT-GOx) also resulted in increasing values of crosshybridization although a clear distinction between specific and nonspecific response is always attainable. For the highest concentration assayed, close to 30 nM, the nonspecific response reaches a value of 0.5 µA cm-2, which represents 25% of the total specific response; however, the contribution of the nonspecific response in this direct assay can decrease to ∼2% if the dendritic enzyme concentration remains bellow 5 nM. This response is mainly due to nonspecific adsorption of GOx, because similar currents were obtained at these concentrations when the GOxCTRL was assayed. Once the specificity of these functionalized oligonucleotides has been proved, the critical issue will be the demonstration of further hybridization of the free anchored sequences. If so, this double hybridization will permit the construction of a dendritic architecture (see Figure 1A) with enhanced sensitivity. A direct demonstration of this concept was performed by a second hybridization step using a biotinylated sequence and avidinmodified GOx. Then, it was observed that the specific response was doubled as a consequence of an accumulation of enzyme molecules. A control experiment with a noncomplementary bioti-

Figure 6. (A) Amperometric response profiles in the sandwich hybridization assay of different concentrations of AB target sequence: (A) 0, (B) 5, (C) 10, (D) 20, (E) 30, (F) 40, (G) 50 nmol L-1, and (H) 0.1 µmol L-1 of ABX as negative control. (B) Dependence of the amperometric response on the concentration of AB in the sandwich hybridization assay. Standard deviations are obtained with a minimum of three electrodes. Sandwich hybridization conditions: simultaneous incubation of the target with 11.2 nmol L-1 of the complementary dendritic enzyme (AT-GOx) for 30 min under stirring in 0.1 M phosphate buffer (pH 7.2) and at room temperature.

nylated sequence did not result in any current increase. Overall, these results show that the functionalized oligonucleotides can specifically hybridize and that the remaining attached sequences are prone to signal amplification by successive hybridization steps, resulting in a dendritic architecture. As a proof of concept, the amplification has been tested here with a biotinylated sequence, although incubation with complementary functionalized sequences opens up the way of multiple and subsequent amplification steps. Detection of DNA Hybridization with Functionalized Oligonucleotides in a Sandwich-Type Assay. Most frequently, the target sequence is substantially longer than the probe, imposing difficulties related to cross-hybridization.42 To minimize false positive signals by improving the selectivity, a DNA capture sandwich-type assay can be performed. In addition, this assay format avoids the need to label the target sequence and thus, offers a more general way of specific DNA detection. Consequently, the concept of enzyme dendritic sequences has also been checked in this sandwich-type configuration. The modified electrodes, Au/MPS/RP/HRP/RP-BP, were hybridized with the target sequence (AB) in the presence of 11.2 nM (protein concentration) of GOx-AT for 30 min at room temperature in a one-step assay. Figure 6 shows the response profiles and the dependence of these steady-state currents on the concentration of the target sequence. The specificity of this configuration is demonstrated by the very low responses obtained in the absence of target oligonucleotide and, more importantly, in the presence of an excess of a negative control (ABx). Quantitatively, the nonspecific response represents 6% of the total signal. In search of reducing the incubation steps (i.e., practicability of the assay), the target sequence has been simultaneously incubated with the functionalized oligonucleotide on the DNAmodified electrode surface. Considering that one GOx molecule (42) Palecek, E.; Kizek, R.; Havran, L.; Billova, S.; Fojta, M. Anal. Chim. Acta 2002, 469, 73-83.

Analytical Chemistry, Vol. 76, No. 11, June 1, 2004

3137

Figure 7. Amperometric responses in the sandwich hybridization assay of 40 nM (1) AB, complementary sequence; and (2) AB4, four mismatched oligonucleotide. Sandwich hybridization conditions: simultaneous incubation of AB or AB4 with 11.2 nmol L-1 of the complementary functionalized sequence (AT-GOx) for 30 min, under stirring, at room temperature, and in 0.1 M (A) or 0.01 M (B) phosphate buffer pH 7.2.

contains an average of three oligonucleotide sequences, this onestep format may result in multiple hybridizations of the target sequence, since both are in a homogeneous phase. Thus, the amount of target DNA prone to hybridization with the capture probe at the heterogeneous phase is decreased, diminishing the number of GOx molecules per hybridization event. Overall, these results indicate that the functionalized sequences can also be used in a sandwich assay, although the one-step format may be detrimental for its sensitivity, and higher sensitivity could be achieved if sequential hybridization were performed. The specificity of these functionalized oligonucleotides is challenged by the detection of mismatched sequences. For this reason, a four-base mismatch was introduced in the segment of the target DNA complementary to the functionalized sequence. The long target (AB4) is shown in Table 1. As expected from the previous results for a completely negative control and under identical hybridization conditions, the detection of mutations is compromised by cross-hybridization. Figure 7A shows the responses obtained for the perfectly matched sequence and an equal concentration of the 4-base mismatch; however, the specificity is clearly enhanced when the astringency is increased by dilution of the buffer to 0.01 M. Under these conditions and at room temperature, the 4-base mismatch can be clearly detected, as seen in Figure 7B. Consequently, the specificity of the functionalized sequence can be modulated by the astringency of the hybridization, and a mere dilution of the buffer concentration enables the detection of 4-base mutations. The concentration of the functionalized oligonucleotides used in these experiments was higher than that estimated in a direct assay if total absence of nonspecific response is pursued. This is of critical importance because it determines the limit of quantification for the target DNA. Figure 8 shows the responses obtained when this one-step sandwich assay is performed with a concentration of the reporter-functionalized sequence of 2.8 nM. Under these conditions, cross-hybridization and nonspecific adsorption is completely suppressed, as demonstrated by the absence of current when an excess of noncomplementary sequence is assayed. This permits the specific quantification of 1 nM long target sequences at room temperature and in 30 min. CONCLUSIONS The transduction of affinity reactions through enzyme functionalized oligonucleotides permits the design of simple ampero3138 Analytical Chemistry, Vol. 76, No. 11, June 1, 2004

Figure 8. Amperometric responses in the sandwich hybridization assay of 1 nmol L-1 AB and 10 nmol L-1 ABX. Sandwich hybridization conditions: simultaneous incubation of AB or ABX with 2.8 nmol L-1 AT-GOx for 30 min, under stirring, in 0.1 M phosphate buffer (pH 7.2) and at room temperature.

metric sensors with the additional advantage of introducing a catalytic reaction. Traditionally, one enzyme molecule accounts for each affinity event. The presented work describes a new concept for the amperometric transduction of hybridization reactions based on the modification of an enzyme with several oligonucleotide sequences.24 The rationale behind this concept relies on the ability of this functionalized oligonucleotide to specifically detect hybridization events and to further permit signal amplification by successive hybridization steps. In this manner, each hybridization event results in an accumulation of catalytic molecules and a concomitant increase in sensitivity. Glucose oxidase has been used as a reporter enzyme, because if successive hybridization steps are performed for signal amplification, the imposed distance to the electrode surface does not risk electrochemical communication. A highly permeable polyelectrolyte multilayer assembly is responsible for the electrocatalytic conversion of hydrogen peroxide fluxes into a cathodic current measured at 100 mV vs Ag/AgCl. The ability of these functionalized oligonucleotides to specifically transduce hybridizations reactions has been demonstrated in two different assays, always performed at room temperature and during 30 min of incubation. As described for the direct assay, the sensitivity of the assay is doubled after a second hybridization step. Although this work has been demonstrated for amperometric DNA sensors, other means of transduction are also applicable. Moreover, this scheme of transduction and using different oxidases opens up the route for the design of affinity-based simple amperometric multidetection systems with no need for spatial resolution and direct signal amplification. Overall, this approach does not require intensive synthetic work because only two different and complementary oligonucleotide sequences, linked to a catalytic molecule, are required for signal transduction and amplification. ACKNOWLEDGMENT Financial support from the Spanish Ministry of Science and Technology is gratefully acknowledged (Project No. BIO99-1213). A.N. and O.R. acknowledge a postdoctoral research contract from the Community of Madrid and a predoctoral fellowship from the Ministry of Education and Science, respectively. Received for review January 5, 2004. Accepted March 16, 2004. AC0499672