Amplified Electrocatalysis at DNA-Modified Nanowires - Nano Letters

Zhichao Fang , Leyla Soleymani , Georgios Pampalakis , Maisa Yoshimoto , Jeremy A. Squire , Edward H. ..... Fang Wei , Peter B. Lillehoj , Chih-Ming H...
0 downloads 0 Views 164KB Size
NANO LETTERS

Amplified Electrocatalysis at DNA-Modified Nanowires

2005 Vol. 5, No. 6 1051-1055

Melissa A. Lapierre-Devlin, Camille L. Asher, Bradford J. Taft, Rahela Gasparac, Marcel A. Roberts, and Shana O. Kelley* Eugene F. Merkert Chemistry Center, Boston College, Chestnut Hill, Massachusetts 02467 Received March 12, 2005; Revised Manuscript Received April 30, 2005

ABSTRACT Arrayed gold nanowires are a novel and useful platform for electrochemical DNA detection. Pilot studies testing the use of these templated structures with an electrocatalytic reporter system revealed that very low detection thresholds for target DNA sequences can be obtained. One factor contributing to the heightened sensitivity is the high signal-to-noise ratio achieved with the large electrocatalytic signals observed at DNA-modified nanowires. Here, we explain the improved sensitivity with evidence illustrating that electrocatalysis at DNA-modified nanostructures generates amplified signals that are significantly larger than those observed at bulk gold surfaces. The results presented strongly suggest that the three-dimensional architectures of the nanowires facilitate the electrocatalytic reaction because of enhanced diffusion occurring around these structures. Effects unique to the nanoscale are shown to underlie the utility of nanowires for DNA biosensing.

Electrocatalysis is used effectively for many applications, including fuel production and organic synthesis.1,2 Electrocatalytic processes have also been harnessed as reporter systems for biosensing.3-8 DNA and other biological analytes can be detected with high sensitivity by monitoring catalytic reactions reporting on biomolecular complexation events. Electrocatalytic detection systems typically produce large signals that can be easily quantitated, leading to favorable sensitivity and accuracy. Recent studies in our laboratories have focused on an electrocatalytic DNA detection system7 that uses threedimensional (3D) gold nanowire ensembles (NEEs) as the electrode platform.8 Attomole sensitivity toward target DNA sequences was achieved using the NEEs as sensing elements.8 This detection limit constitutes a considerable improvement over the femtomole sensitivity obtained with gold macroelectrodes. One source of the improved sensitivity is the high signal-to-noise ratio achieved. While background capacitive currents at 3D NEEs are minimal, the faradaic currents measured are significantly larger than those observed at macroelectrodes with comparable surface areas. Here, we demonstrate that the heightened sensitivity reflects unique nanoscale effects, e.g., the enhanced diffusion that occurs at three-dimensional structures with dimensions that match the biological scale. Three-dimensional gold NEEs are generated by electroless deposition of gold into polycarbonate membranes,9,10 followed by oxygen plasma etching to expose the wires. The wires are approximately 15-20 nm in diameter and ap* Corresponding author. E-mail: [email protected]. 10.1021/nl050483a CCC: $30.25 Published on Web 05/20/2005

© 2005 American Chemical Society

proximately 150-250 nm in exposed length.11 Thiolated DNA is deposited,12-14 and films similar to those obtained with bulk gold surfaces are generated.8 An electrocatalytic reporter system, which uses Ru(NH3)63+ as a primary acceptor and Fe(CN)63- as a secondary acceptor, accurately reports on the amount of immobilized DNA (Scheme 1).7 Three-dimensional NEEs, both bare and modified with double-stranded DNA,15 yield similar electrochemical responses in solutions of Ru(NH3)63+ (Figure 1a,b) to those collected with macroelectrodes (Figure 1d,e).16 The calculated working area of the NEE (0.01 cm2)17 is slightly smaller than the macroelectrode (0.02 cm2), but is not drastically different. The similar magnitudes of the currents measured for Ru(III) indicates that the areas of the NEEs and macroelectrodes are comparable. However, when electrocatalytic currents are measured, the magnitude of the electrochemical signals is strikingly different. The NEE exhibits background-subtracted cathodic peak currents (ipc) of 8 µA (Figure 1c), while the macroelectrode ipc is 1 µA (Figure 1f).18 The observation of electrocatalysis at DNA-modified NEEs presents an interesting point of comparison to previous studies of bimolecular processes at microelectrodes where diminished reactivity was observed.19 At unfunctionalized microelectrodes studied by Wightman and co-workers,19 reactants diffuse away from the small working area before turnover and re-reduction can occur. At the DNA-modified NEEs described here, the oligonucleotide film serves to sequester bound ions at the surface. Therefore, electrocatalysis proceeds even though transport away from the nanoscale structures would occur faster than for microelectrodes.

Figure 1. Representative cyclic voltammograms of (a) 3 mM Ru(NH3)63+ at a bare NEE, (b) 27 µM Ru(NH3)63+ at a NEE modified with 30-mer duplex DNA, (c) 27 µM Ru(NH3)63+ and 2 mM Fe(CN)63- at a NEE modified with 30-mer duplex DNA, (d) 3 mM Ru(NH3)63+ at a bare macroelectrode, (e) 27 µM Ru(NH3)63+ at a macroelectrode modified with 30-mer duplex DNA, (f) 27 µM Ru(NH3)63+ and 2 mM Fe(CN)63- at a macroelectrode modified with 30-mer duplex DNA. Scan rate for all CV experiments is 100 mV/s. Scheme 1. (a) Schematic illustration of Ru(III)/Fe(III) electrocatalysis at a DNA-modified Au NEE. (b) Scanning electron micrograph of NEEs used in the studies described. The area shown represents a 5 µm × 5 µm area. The individual wires protrude by ∼200 ( 10 nm from the membrane surface.

Figure 2. Representative cyclic voltammograms for 18-mer duplex DNA-modified macroelectrode (a) and NEE (b) collected in solutions containing 2 mM Fe(CN)63- and 10 (blue), 20 (red), or 80 µM (black) Ru(NH3)63+. Scan rate for all CV experiments is 100 mV/s.

The efficiency of an electrocatalytic process should be strongly influenced by the concentrations of reactants and rates of diffusion.20 Surprisingly, the dependence of the electrocatalytic currents on the concentration of Ru(NH3)63+ at NEEs is markedly different than at macroelectrodes 1052

(Figure 2). At the macroelectrode, the addition of Ru3+ strongly attenuated the efficiency of electrocatalysis. As Ru(NH3)63+ is titrated into the immobilized DNA monolayer, the catalytic wave disappears while the direct Ru(III) reduction becomes more pronounced (Figure 2a). However, Nano Lett., Vol. 5, No. 6, 2005

Figure 3. Plot of (a) ipc vs (scan rate)1/2 and (b) ipc vs (scan rate) for macroelectrode (dashed) and NEE (solid) modified with 18mer duplex DNA in 20 µM Ru(NH3)63+.

at NEEs, electrocatalytic currents increase directly with Ru(NH3)63+ (Figure 2b). The behavior at the nanoelectrodes is consistent with the theoretical prediction that the inclusion of more catalyst should produce more current,20 while the results obtained at the macroelectrodes are unexpected. The different concentration dependences for electrocatalysis at macroelectrodes as compared to NEEs indicates that the mobilities of ions bound to these two substrates are different. The surface of an Au macroelectrode is accessible only by linear diffusion, and ions within the DNA film may be essentially trapped on the time scale of an electrochemical experiment. The three-dimensional structure of the NEEs opens the possibility that radial diffusion may occur, and also that the radius of curvature of the individual nanowires may lead to more efficient diffusion for ions binding in equilibrium with the DNA film. The observation that a higher concentration of Ru(NH3)63+ diminishes electrocatalysis efficiency at macroelectrodes is consistent with the idea that the added ions are not mobile within the film, and moreover, may even slow the diffusion of the ions that could participate in electrocatalysis when less Ru was present. Diffusion appears to be facile at the NEEs even when the Ru/DNA stoichiometry is high.21 The diffusional mobility of Ru3+ ions bound to DNA monolayers immobilized on bulk gold surfaces versus NEEs was investigated by analyzing the dependence of cathodic peak current (ipc) on scan rate for solutions containing Ru(NH3)63+ (Figure 3). Electrochemical processes involving surface-bound species exhibit currents depending directly on Nano Lett., Vol. 5, No. 6, 2005

Figure 4. Representative cyclic voltammograms for (a) 18-mer duplex DNA-modified macroelectrode and (b) NEE collected in solutions containing 40 µM Ru(NH3)63+ and 0 µM (blue) or 320 µM (red) Fe(CN)63-. Scan rate for all CV experiments is 100 mV/s. Background subtraction was performed so that scans could be directly compared.

the scan rate of a cyclic voltammetry experiment, while solution-borne or diffusible species will yield ipc values proportional to the square root of the scan rate.22 When this analysis was performed on macroelectrodes and NEEs, different dependences were observed. The macroelectrode currents displayed the direct scan rate dependence characteristic of surface-bound species (Figure 3b), while those measured at NEEs increased linearly with (scan rate)1/2, consistent with greater diffusion (Figure 3a). At the DNAmodified surfaces, it appears that the binding of ions is more static when a flat substrate is used relative to when threedimensional wires are the support for the film. Ru(III) is bound by the DNA film immobilized on NEEs, evidenced by the fact that electrochemical signals are not observed at bare NEEs (data not shown), but it appears that the binding does not preclude efficient diffusion. When electrocatalysis at DNA-modified macroelectrodes and NEEs was monitored as a function of Fe(CN)63concentration, the NEEs again showed much better performance with respect to electrocatalysis efficiency (Figure 4). For example, when DNA-modified macroelectrodes were 1053

References

Figure 5. Scan rate dependence of electrocatalytic turnover at 18mer duplex DNA- modified macroelectrodes (black) or NEEs (gray). Data shown represent averages of three independent trials. Turnover was calculated by measuring signals in solutions containing only 20 µM Ru(NH3)63+ and then solutions containing 20 µM Ru(NH3)63+ and 2 mM Fe(CN)63-. A ratio of catalyzed to uncatalyzed background subtracted cathodic peak currents was calculated representing turnover.

monitored for electrocatalysis in solutions containing 40 µM Ru(NH3)63+ and 320 µM Fe(CN)63-, very poor (