Electrochemical DNA biosensors based on thin gold films sputtered on capacitive nanoporous niobium oxide

Available online at www.sciencedirect.com

Biosensors and Bioelectronics 23 (2008) 852–856

Electrochemical DNA biosensors based on thin gold films sputtered on capacitive nanoporous niobium oxide Sangchul Rho a , Deokjin Jahng a , Jae Hoon Lim b , Jinsub Choi b,∗ , Jeong Ho Chang b , Sang Cheon Lee b , Kyung Ja Kim b b

a Department of Environmental Engineering and Biotechnology, Myongji University, Yongin, Gyunggi-do 449-728, South Korea Nanomaterials Application Division, Korea Institute of Ceramic Engineering and Technology (KICET), Seoul 153-801, South Korea

Received 24 May 2007; received in revised form 3 September 2007; accepted 5 September 2007 Available online 11 September 2007

Abstract Electrochemical DNA biosensors based on a thin gold film sputtered on anodic porous niobium oxide (Au@Nb2 O5 ) are studied in detail here. We found that the novel DNA biosensor based on Au@Nb2 O5 is superior to those based on the bulk gold electrode or niobium oxide electrode. For example, the novel method does not require any time-consuming cleaning step in order to obtain reproducible results. The adhesion of gold films on the substrate is very stable during electrochemical biosensing, when the thin gold films are deposited on anodically prepared nanoporous niobium oxide. In particular, the novel biosensor shows enhanced biosensing performance with a 2.4 times higher resolution and a three times higher sensitivity. The signal enhancement is in part attributed to capacitive interface between gold films and nanoporous niobium oxide, where charges are accumulated during the anodic and cathodic scanning, and is in part ascribed to the structural stability of DNA immobilized at the sputtered gold films. The method allows for the detection of single-base mismatch DNA as well as for the discrimination of mismatch positions. © 2007 Elsevier B.V. All rights reserved. Keywords: Anodization; Niobium oxide; Porous oxide; Electrochemical biosensor; Capacitance

1. Introduction The research activities in the preparation of electrochemical biosensors for detecting DNA hybridization events have dramatically increased over the past decade, resulting in the development of new material design and novel fabrication processes (Katz et al., 2004; Wu et al., 2001; He et al., 2000; Kerman et al., 2004; Liua and Huo, 2005; Vlassiouk et al., 2005; Hang and Guiseppi-Elie, 2004; Yoon et al., 2002; Archer et al., 2004; Zhou et al., 2001; Su et al., 2004). Among the preparation steps for the DNA biosensor, immobilization of biomolecule probes on a desired substrate is a very important process since the sensitivity, the detection resolution and the reproducibility are significantly affected by the step. In general, the surface of the substrate and/or the terminus of DNA have to be modified for stable immobilization of DNA probes. Various surface modification methods, e.g., affinity binding such as avidin (or

∗

Corresponding author. Tel.: +82 2 3282 2457; fax: +82 2 3282 7769. E-mail address: [email protected] (J. Choi).

0956-5663/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2007.09.001

streptavidin) and biotin (Diamandis and Christopoulos, 1991; Marrazza et al., 1999; Park et al., 2004), and covalent attachment using water-soluble carbodiimide (EDC) (Millan et al., 1992; Millan and Mikkelsen, 1993; Liu et al., 1996; Bardea et al., 1999), have been developed for an effective immobilization of DNA probes. In this respect, gold substrates have attracted special attention as an electrode for electrochemical DNA biosensors since thiolated-DNA can be strongly bound at the surface of gold through Au–thiol binding (Herne and Tarlov, 1997; Levicky et al., 1998; Kelly and Barton, 1997; Hashimoto et al., 1994). Thiolated-DNA can be monolayered on gold by a self-assembly manner, which provides stable and structurally well-defined electrochemical interfaces (Hashimoto et al., 1994; Sun et al., 1998; Hong et al., 1999). One major drawback of the gold electrode is attributed to the cleaning step (Cooper and Cass, 2002): in order to obtain reproducible results, the gold electrode is mechanically polished and then electrochemically etched in weak H2 SO4 by cyclic voltammetry before immobilization. This time-consuming process determines a quality of electrochemical DNA biosensors. Alternatively, thin gold films deposited by low pressure gold sputtering or electrochemical deposition can

S. Rho et al. / Biosensors and Bioelectronics 23 (2008) 852–856

853

provide a highly sensitive and a reproducible electrode for DNA biosensors without the requirement of the cleaning step (Cooper and Cass, 2002). According to the Jiang group, Au–thiol binding is improved when thiolated-DNA probes are immobilized on gold nanoparticles that are electrochemically deposited on the gold electrode. Thus, an enhancement of DNA immobilization and hybridization can be achieved with a gold nanoparticlemodified electrode (Liu et al., 2005a,b). However, the thin gold film directly sputtered on a substrate very easily peels off during immobilization or electrochemical measurement, giving less reliability; usually a Cr layer is required for enhancing adhesion (Cooper and Cass, 2002). Recently, DNA detection by using capacitive metal oxides has been proposed to obtain amplified signals that allow easy distinguish between single-stranded DNA (ssDNA) and doublestranded DNA (dsDNA) (Moreno-Hagelsieb et al., 2007). In this article, we report on highly reliable and sensitive DNA biosensors using a thin gold film sputtered on capacitive anodic nanoporous niobium oxide: the nanoporous niobium oxide offers a good adhesion as well as an enhancement of redox signals by accumulation of charges in between the gold film and the niobium oxide. This method shows a highly reliable electrochemical redox signal with a three times higher sensitivity compared to a bulk gold electrode. The mechanism of enhancing the signal by the thin gold film on nanoporous niobium oxide is in part attributed to capacitive niobium oxide and is in part ascribed to the bridged thin gold film. 2. Experimental 2.1. Materials Methylene blue (MB) and 6-mercapto-1-hexanol (MCH) were purchased from Sigma (USA). Niobium foils (99.9% purity) and gold foils (99.999% purity) with a thickness of 0.25 mm were purchased from Goodfellow (England). Highly doped n-Silicon waters (ρ < 1  cm) were obtained from Universitywafer (USA). The synthetic oligonucleotides, corresponding to portions of 16S rDNA (E. Coli numbering: 190–208) of ammonia oxidizing ␤-proteobacteria, were synthesized by Metabion GmbH (Germany). Their base sequences were as following: • Thiolated-ssDNA probe: 5 -HS-(CH2 )6 –GGA GAA AAG CAG GGG ATC G-3 • Complementary ssDNA target: 5 -CGA TCC CCT GCT TTT CTC C-3 • Single-base mismatch ssDNA1: 5 -CGA TCC CCA GCT TTT CTC C-3 • Single-base mismatch ssDNA2: 5 -CGA TCC CCT GCT TTT CAC C-3 2.2. Preparation of anodic porous niobium oxide The preparation of porous niobium oxide by electrochemical anodization of high purity foils was described elsewhere in detail (Choi et al., 2006; Choi et al., 2007). The original diameter of

Fig. 1. Schematic diagram of various biosensor electrodes: (a) a bulk gold electrode (b) anodically prepared niobium oxide, (c) a thin gold film that is completely coated on the whole porous niobium oxide (C-@Nb2 O5 ), (d) a thin gold film that is partially coated at the center of porous niobium oxide (P-Au@Nb2 O5 ) and (e) gold-coated silicon wafer (Au@n-Si).

foils was 15 mm and the thickness was 0.25 mm. Anodization was carried out only at the center of the foil by using a mask with a 14.5 mm-hole diameter, thus remaining an unanodized metal area on a 0.5 mm-thick rim of the foil (see Fig. 1(b)). This metal layer will work as electrode connectors. The final thickness of anodized niobium oxide was adjusted to 120–180 nm and the pore diameter was around 10 nm. 2.3. Immobilization of ssDNA onto the electrode surface 2.3.1. Immobilization onto the gold surface The gold electrode was mechanically polished with a 1, 0.3 and 0.05 ␮m alumina (Al2 O3 ) slurry on micropads (Buehler, Lake Bluff, IL), and washed ultrasonically with deionized (DI) water (18 > M cm) to remove any traces of alumina. The electrode was cleaned by electrochemical etching in 0.05 M H2 SO4 by potential scanning between −0.3 and +1.5 V until a reproducible cyclic voltammogram was obtained. Then, the gold electrode was rinsed with copious amounts of DI water,

854

S. Rho et al. / Biosensors and Bioelectronics 23 (2008) 852–856

washed in ethanol, and finally dried with high-purity nitrogen before monolayer adsorption. The cleaned gold electrodes were reacted with 80 ␮L of 5 -thiolated-ssDNA probes (0–20 ␮g/mL) in immobilization buffer (1 M KH2 PO4 , pH 4.5) and incubated for 150 min at room temperature. Subsequently, the modified gold electrode was thoroughly rinsed with DI water to remove the weakly adsorbed ssDNA probes. In order to block the uncovered area, the electrode was immersed in 1 mM ethanolic MCH solution for 60 min at room temperature, followed by rinsing with absolute ethanol. This allows for the self-assembly of an electrochemically inactive MCH layer on the gold electrode.

potential. After accumulation of MB, the electrode was rinsed three times with 40 mM tris-Cl buffer (pH 7.4) to remove nonspecifically bound MB. MB has been widely used as an electroactive indicator of DNA hybridization events since it shows fast and reversible redox reaction. It is generally accepted that the MB reduction peak decreases after hybridization events since the guanine bases, which can be reacted with MB in the case of ssDNA, are firmly bound to the cytosine (Rohs et al., 2000; Gu et al., 2002).

2.3.2. Immobilization of ssDNA probes onto completely Au-coated niobium oxide (C-Au@Nb2 O5 ), partially Au-coated niobium oxide (P-Au@Nb2 O5 ) and Au-coated silicon wafers (Au@n-Si) Thin gold films were deposited onto porous niobium oxide and a piece of n-type silicon wafer by a magnetron sputter (MSP1S, Vacuum Device Inc., Japan), which is used for making an Au-coating layer for scanning electron microscope (SEM) measurements. The thickness of the gold layer was adjusted to 50 nm in all cases. We prepared two different types of gold-coated niobium oxide specimens: one is that the porous niobium oxide and unoxidized metal layer were completely covered with the gold layer (C-Au@Nb2 O5 , Fig. 1(c)). Thus, current can flow vertically through the thin gold film–niobium oxide–niobium metal (path 1) as well as flow along the surface of gold thin film (path 2). For the second sample, the gold film was deposited only at the center of the porous niobium oxide, thus the current can only flow along path 1 (P-Au@Nb2 O5 , Fig. 1(d)). The area of gold film sputtered is around 8 mm × 8 mm. Note that anodic porous niobium oxides before and after gold coating, respectively, are displayed in supplementary material (Fig. S1(a) and (b)). The prepared Au-coated substrates were reacted with 80 ␮L of thiolated-ssDNA probes (0–20 ␮g/mL) in immobilization buffer (1 M KH2 PO4 , pH 4.5) and incubated 150 min at room temperature. Subsequently, the modified electrodes were thoroughly rinsed with DI water to remove the weakly adsorbed ssDNA probes. The blocking procedure by the MCH layer is the same as that of the bulk gold electrode.

The electrochemical detections of ssDNA and dsDNA were carried out by using a potentiostat/galvanostat (AutoLab PGSTAT12, Eco Chemie) interfaced to a computer. The cell was a three-electrode system consisting of a Pt mesh acting as the counter electrode, Ag/AgCl/3 M KCl as the reference electrode and DNA immobilized substrates with a size of 1 cm2 as the working electrode. Note that all the electrochemical measurements (bulk gold, P-Au@Nb2 O5 , C-Au@Nb2 O5 and porous niobium) were carried out in the same area (1 cm2 ) by using a home-made Teflon holder. The redox signal of the accumulated MB on the electrode was measured by cyclic voltammetry (CV) in 40 mM tris-Cl buffer (pH 7.4) from −0.5 to 0.1 V at the scan rate of 100 mV s−1 .

2.6. Electrochemical detection

3. Results and discussion As reported by many research groups, distinct redox peaks of MB are observed when thiolated-ssDNA probes are anchored on the gold electrode (Zhu et al., 2004; Liu et al., 2005a,b; Jin et al., 2007). As exhibited in Fig. 2, redox peaks at −0.28 V (reduction) and −0.25 V (oxidation) are observed. The current density at the redox potential increases as the concentration of immobilized ssDNA probes increases. Note that the dramatic decrease of current density below −0.3 V indicates that there is strong hydrogen evolution. This would interfere with the precise detection of a hybridization event.

2.4. Hybridization The hybridization reaction was performed by pipetting 80 ␮L of different concentrations of DNA target solution (2 × SSC buffer; 8.765 g sodium chloride, 4.41 g sodium citrate, adjusted to pH 7.0 with NaOH) onto the probe-immobilized electrodes (C-Au@Nb2 O5 ) for 90 min at 37 ◦ C. Afterward, the electrodes were washed three times with 40 mM tris-Cl buffer (pH 7.4) to remove unhybridized DNA, resulting in the formation of hybrid-modified electrodes. 2.5. Intercalation of methylene blue Methylene blue was accumulated by immersing the electrodes into the stirred 40 mM tris-Cl buffer (pH 7.4) containing 1 mM MB with 20 mM NaCl for 5 min without applying any

Fig. 2. Cyclic voltammograms (CV) of methylene blue (MB) on the bulk gold electrodes where various concentrations of thiolated-ssDNA are immobilized: (a) 0.75 ␮M, (b) 1.50 ␮M, (c) 2.25 ␮M and (d) 3.00 ␮M.

S. Rho et al. / Biosensors and Bioelectronics 23 (2008) 852–856

Fig. 3. CV of MB on C-Au@Nb2 O5 where various concentrations of thiolatedssDNA are immobilized: (a) 0 ␮M, (b) 0.75 ␮M, (c) 1.50 ␮M, (d) 2.25 ␮M and (e) 3.00 ␮M.

On the contrary, in the case of ssDNA probes immobilized on C-Au@Nb2 O5 , a very reproducible plate current density is observed below −0.3 V (Fig. 3). This represents that the hydrogen evolution reaction is suppressed on C-Au@Nb2 O5 . Similar to the results shown in Fig. 2, the intensity of the redox peaks increases as the concentration of immobilized DNA probes increases. After 2.25 ␮M ssDNA probes are immobilized, the intensity does not significantly change, meaning that the immobilization of ssDNA is saturated on the active sites of the gold film. For the thin gold film sputtered on n-Si, the film peels off when it contacts the immobilization buffer solution (see the supplementary material, Fig. S2). Thus, we cannot further carry out experiments using Au@n-Si. Reduction peaks of MB on the bulk gold electrode and CAu@Nb2 O5 are compared as a function of the concentration of ssDNA probes (see the supplementary material, Fig. S3). It shows that C-Au@Nb2 O5 has a 2.4 times steeper slope, indicating that the detection resolution is greatly enhanced by our method. Fig. 4 shows that the reduction current den-

Fig. 4. Comparison of CV on (a) the bulk electrode, (b) P-Au@Nb2 O5 and (c) C-Au@Nb2 O5 . Note that the concentration of immobilized ssDNA probes is 1.50 ␮M in all cases. When niobium oxide is used as a substrate, current density is enhanced.

855

Fig. 5. Electrochemical detection of hybridization events. As the concentration of complementary DNA targets increases, the current peak decreases: note that the concentration of complementary target DNA is (a) 0 ␮M, (b) 1.5 ␮M, (c) 3 ␮M and (d) 4.5 ␮M. The inset shows a good linearity of the slope of the concentration of DNA targets vs. reduction current peak.

sity peak observed on C-Au@Nb2 O5 is three times higher than that on the bulk gold electrode. We ascribe the enhancement of the signal of C-Au@Nb2 O5 to the capacitive interface between Au and Nb2 O5 . The positive charge and negative charge, which are produced during cathodic and anodic scanning, respectively, are accumulated at the interface and the redox peaks are amplified. Additionally, it might be ascribed to the larger number of the immobilized ssDNA probes and their structural stability at the C-Au@Nb2 O5 than on the flat Au substrate. The idea on the capacitive enhancement is supported experimentally using P-Au@Nb2 O5 (Fig. 1 (d)). Note that in this case, current should flow along path 1 since the thickness of niobium oxide is much shorter than the distance between the gold layer and the rim of the niobium foil. If the thin gold film is not deposited, current cannot easy flow along path 1 since the niobium oxide is a non-conducting material (Eg = 3.4–5.4 eV) (Schultz and Lohrengel, 2000). In fact, there is almost no current flow observed when ssDNA is directly immobilized on anodic niobium oxide via biotin–streptavidin binding (see the supplementary material, Fig. S4). However, the enhancement of current density is observed when the thin gold film is deposited on the anodic niobium oxide. Interestingly, the CV behavior of P-Au@ Nb2 O5 resembles that of the bulk gold electrode. However, the CV curve does not show redox peaks (b of Fig. 4). In case of C-Au@ Nb2 O5 , the isolate gold films are bridged to the metal rim that work as electrode connectors. Thus, current flows not only through path 1 but also along path 2. In this case, we observe that current density is enhanced like P-Au@Nb2 O5 as well as redox peaks are exhibited. More detailed mechanisms on the signal enhancement are under investigation. After hybridization with ssDNA targets, electrochemical analysis was carried out on C-Au@Nb2 O5 . Note that the MB redox peaks are reduced when the hybridization event occurs. As shown in Fig. 5, the redox peaks reduce as the concentration

856

S. Rho et al. / Biosensors and Bioelectronics 23 (2008) 852–856

Acknowledgments Supports from the MOCIE (Ministry of Commerce, Industry and Energy) of Korea and KICET are gratefully acknowledged. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bios.2007.09.001. References

Fig. 6. CV measurements of the hybridization of DNA probes with (a) completely complementary DNA targets, (b) single-base mismatch ssDNA2 and (c) single-base mismatch ssDNA1. Fig. 6(d) shows CV before hybridization. Note that single-base mismatch ssDNA1 and ssDNA2 have different mismatch positions. MB is used as the intercalator.

of DNA targets increase. A good linearity of target concentration versus decrease in redox peak is exhibited in the inset of Fig. 5. The feasibility for electrochemical detection of single-base mismatch DNA by using C-Au@ Nb2 O5 is demonstrated in Fig. 6. We used single-base mismatch DNA targets that have different mismatch positions (see Section 2 in detail). Compared to complimentary ssDNA targets, single-base mismatch DNAs have a higher current peak. Note that if the mismatch site is in the middle (single-base mismatch DNA1), the possibility to bind with DNA probes is much less as compared to when the mismatch site is at the end (single-base mismatch DNA2). Therefore, a higher reduction peak is observed on single-base mismatch DNA1 compared to single-base mismatch DNA2. This result is in good agreement with previous reports (Yershov et al., 1996; Jin et al., 2007). 4. Conclusions We demonstrated that a thin gold film sputtered on capacitive porous niobium oxide can be applied for electrochemical biosensor platform technology for enhancing redox signal. By using this novel method, a 2.4 times higher detection resolution is exhibited as compared to that obtained by conventional bulk gold electrodes. In particular, the novel method does not require any time-consuming cleaning step in order to obtain reproducible results. We demonstrated that single-base mismatch DNAs, which have different mismatch positions, can be easily discriminated by the method. We speculated that the signal enhancement is ascribed to the capacitive interface between the gold film and the niobium oxide. More detailed mechanisms on the signal enhancement are under investigation in terms of dependence of nanopore size and gold film thickness on the signal enhancements.

Archer, M., Christophersen, M., Fauchet, P.M., 2004. Biomed. Microdevices 6, 203–211. Bardea, A., Dagan, A., Willner, I., 1999. Anal. Chim. Acta 385, 33–43. Choi, J., Lim, J.H., Kim, K.J., Lee, S.C., Chang, J.H., Cho, M.A., 2006. Electrochim. Acta 51, 5502–5507. Choi, J., Lim, J.H., Lee, J., Kim, K.J., 2007. Nanotechnology 18, 055603. Cooper, J., Cass, T. (Eds.), 2002. Biosensors. Oxford University Press, NewYork. Diamandis, E.P., Christopoulos, T.K., 1991. Clin. Chem. 37, 625–636. Gu, J.Y., Lu, X.J., Ju, H.X., 2002. Electroanalysis 14, 949–954. Hang, T.C., Guiseppi-Elie, A., 2004. Biosens Bioelectron. 19, 1537–1548. Hashimoto, K., Ito, K., Ishimori, Y., 1994. Anal. Chem. 66, 3830–3833. He, L., Musick, M.D., Nicewarner, S.R., Salinas, F.G., Benkovic, S.J., Natan, M.J., Keating, C.D., 2000. J. Am. Chem. Soc. 122, 9071–9077. Herne, T.M., Tarlov, M.J., 1997. J. Am. Chem. Soc. 119 (38), 8916–8920. Hong, H.G., Park, W., Yu, E.J., 1999. J. Am. Chem. Soc. 476, 177–181. Jin, Y., Yao, X., Liu, Q., Li, J., 2007. Biosens. Bioelectron. 22, 1126–1130. Katz, E., Willner, I., Wang, J., 2004. Electroanalysis 16, 19–44. Kelly, S.O., Barton, J.K., 1997. Bioconjug. Chem. 8, 31–37. Kerman, K., Kobayashi, M., Tamiya, E., 2004. Meas. Sci. Technol. 15, R1–R11. Levicky, R., Herne, T.M., Tarlov, M.J., Satija, S.K., 1998. J. Am. Chem. Soc. 120, 9787–9792. Liua, L., Huo, Q., 2005. Appl. Phys. Lett. 87, 133902. Liu, S., Ye, L., He, P., Fang, Y., 1996. Anal. Chim. Acta 335, 239–243. Liu, S.F., Li, X.H., Li, Y.C., Li, Y.F., Li, J.R., Jiang, L., 2005a. Electrochim. Acta 51, 427–431. Liu, S.F., Li, Y.F., Li, J.R., Jiang, L., 2005b. Biosens. Bioelectron. 21, 789–795. Marrazza, G., Chianella, I., Mascini, M., 1999. Biosens. Bioelectron. 14, 43–51. Millan, K.M., Spurmanis, A.L., Mikkelsen, S.K., 1992. Electroanalysis 4, 929–932. Millan, K.M., Mikkelsen, S.K., 1993. Anal. Chem. 65, 2317–2323. Moreno-Hagelsieb, L., Foultier, B., Laurent, G., Pampin, R., Remacle, J., Raskin, J.P., Flandre, D., 2007. Biosens. Bioelectron. 22, 2199–2207. Park, J.W., Lee, H.-Y., Kim, J.M., Yamasaki, R., Kanno, T., Tanaka, H., Tanaka, H., Kawai, T., 2004. J. Biosci. Bioeng. 97, 29–32. Rohs, R., Skelenar, H., Lavery, R., R¨oder, B., 2000. J. Am. Chem. Soc. 122, 2860–2866. Schultz, J.W., Lohrengel, M.M., 2000. Electrochim. Acta 45, 2499–2513. Su, X.D., Robelek, R., Wu, Y.J., Wang, G.Y., Knoll, W., 2004. Anal. Chem. 76, 489–494. Sun, X., He, P., Liu, S., Ye, J., Fang, Y., 1998. Talanta 47, 487–495. Vlassiouk, I., Takmakov, P., Smirnov, S., 2005. Langmuir 21, 4776–4778. Wu, G.H., Datar, R.H., Hansen, K.M., Thundat, T., Cote, Majumdar, A., 2001. Nat. Biotechnol. 19, 856–860. Yershov, G., Barsky, V., Belgovskiy, A., Kirillov, E., Kenindlin, E., Ivanov, I., Parinov, S., Guschin, D., Drobishev, A., Dubiley, S., Mirzabekov, A., 1996. Proc. Natl. Acad. Sci. U.S.A 93, 4913–4918. Yoon, H.C., Yang, H.S., Kim, Y.T., 2002. Analyst 127, 1082–1087. Zhou, X.C., Huang, L.Q., Li, S.F.Y., 2001. Biosens. Bioelectron. 16, 85–95. Zhu, N., Zhang, A., Wang, Q., He, P., Fang, Y., 2004. Anal. Chim. Acta 510, 163–168.