Anal. Chem. 2000, 72, 3752-3756
Stripping Analysis of Nucleic Acids at a Heated Carbon Paste Electrode Joseph Wang,†,‡,* Peter Gru 1 ndler,§,|,* Gerd-Uwe Flechsig,†,§ Markus Jasinski,†,§ Gustavo Rivas,† † Eskil Sahlin, and Jose Luis Lopez Paz†
Department of Chemistry and Biochemistry, New Mexico State University, Las Cruces, New Mexico 88003, and Fachbereich Chemie, Universita¨t Rostock, D-18051 Rostock, Germany
A new electrically heated carbon paste electrode has been developed for performing adsorptive stripping measurements of trace nucleic acids. Such coupling of electrochemistry at electrically heated electrodes with adsorptive constant-current stripping chronopotentiometry offers distinct advantages for trace measurements of nucleic acids. The application of increased temperatures during the deposition step results in dramatic (4-34-fold, depending on temperature applied) enhancement of the stripping signal. Such improvement is attributed to the accumulation step at the heated electrode. Forced thermal convection near the electrode surface facilitates the use of quiescent solutions and hence of ultrasmall volumes. Using an electrode temperature of 32 °C and a quiescent solution during the 1-min accumulation, the response is linear over the 1-8 mg/L range tested, with a detection limit of 0.5 mg/L. Such electrode heating technology offers great promise for various applications involving thermal manipulations of nucleic acids. Since the electroactivity of DNA was discovered,1 there have been intensive efforts to apply modern electrochemical techniques in nucleic acid analysis and research.2 While early studies exploited the reduction of DNA and RNA at mercury electrodes,2,3 recent activity has shifted to solid-state electrochemical devices.4 The inherent miniaturization of solid electrodes, and their compatibility with advanced microfabrication technology, make these electrodes excellent candidates for DNA diagnostics. Adsorptive stripping protocols at carbon paste and strip electrodes have been particularly useful for detecting extremely low (subnanogram) levels of DNA and RNA.4,5 Monitoring the anodic signal associated with the guanine oxidation has also been used for the biosensing of DNA hybridization6 and the flow detection of nucleic acids.7 †
New Mexico State University. Corresponding author: (e-mail)
[email protected]; (fax) +1-(505) 6466033. § Universita ¨t Rostock. | Corresponding author: (e-mail)
[email protected]; (fax) +49-(0)381-498-1752. (1) Palecek, E. Nature 1960, 188, 656. (2) Palecek, E. Electroanalysis 1996, 8, 7. (3) Vojtiskova, M.; Lukasova, E.; Jelen, F.; Palecek, E. Bioelectrochem. Bioenerg. 1981, 8, 487. (4) Palecek, E.; Fojta, M. Anal. Chem. 1994, 66, 1566. (5) Anderson, J. L.; Coury, L. A.; Leddy, J. Anal. Chem. 1998, 70, 560R. ‡
3752 Analytical Chemistry, Vol. 72, No. 16, August 15, 2000
The present paper describes a new approach for nucleic acid stripping analysis based on the use of electrochemistry at electrodes with direct electrical (Joule) heating. For this technology, the term hot-wire electrochemistry has been introduced, due to the fact that it started with thin wire electrodes.8,9 It involves the use of heated electrodes to induce thermally efficient convection within a thin solution layer near the surface (with the bulk solution not being exposed to elevated temperatures). This is realized by integration of a symmetric working electrode arrangement in both the heating and potentiostatic circuitry and applying an alternate heating current. Such electrode-heating technology has already been shown very useful for a wide range of electroanalytical applications, including trace metal stripping analysis,10-13 voltammetry above boiling point,14 or amperometric monitoring of flowing streams.15 Electrochemistry at directly heated electrodes holds great promise for DNA diagnostics in view of the major role of elevated temperatures in nucleic acids interactions (e.g., hybridization). It was the idea of A. M. Oliveira Brett (Coimbra, Portugal) to apply the technique of hot-wire electrochemistry to the analysis of nucleic acids to investigate in situ melting of the DNA double strand directly at the electrode surface.16 Similar investigations with biologically active substances at a heated gold electrode have been made together with Oliveira Brett’s group.17 The present study is devoted to stripping analysis of nucleic acids and demonstrates that the heated electrode operation yields a substantial enhancement of the adsorptive stripping chronopotentiometric response in connection with the use of quiescent solutions (i.e., small sample volumes). While hot-wire electrochemistry (6) Wang, J.; Rivas, G.; Cai, X.; Palecek, E.; Nielsen, P.; Shiraishi, H.; Dontha, N., Luo, D.; Parrado, C.; Chicharro, M.; Farias, P.; Valera, F.; Grant, D.; Ozsoz, M.; Flair, M. Anal. Chim. Acta 1997, 347, 1. (7) Wang, J.; Chen, L.; Chicharro, M. Anal. Chim. Acta 1996, 319, 347. (8) Gru ¨ ndler, P. Fresenius' J. Anal. Chem., 1998, 362, 180. (9) Gru ¨ ndler, P.; Kirbs, A. Electroanalysis 1999, 11, 223. (10) Tadesse, Z.; Gru ¨ ndler, P. J. Electroanal. Chem. 1996, 415, 85. (11) Gru ¨ ndler, P.; Flechsig, G.-U. Electrochim. Acta 1998, 43, 3451. (12) Jasinski, M.; Kirbs, A.; Schmehl, M.; Gru ¨ ndler, P. Electrochem. Commun. 1999, 1, 26. (13) Wang, J.; Gru ¨ ndler, P.; Flechsig, G.-U; Jasinski, M.; Lu, J.; Wang, J.; Zhao, Z.; Tian, B. Anal. Chim. Acta 1999, 369, 33. (14) Gru ¨ ndler, P.; Kirbs, A.; Tadesse, Z. Analyst 1996, 121, 1805. (15) Wang, J.; Jasinski, M.; Flechsig, G.-U.; Gru ¨ ndler, P.; Tian, B. Talanta 2000, 50, 1205. (16) Oliveira Brett, A. M., personal communication. (17) Voss, T.; Gru ¨ndler, P.; Brett, C. M. A.; Oliveira Brett, A. M. J. Pharm. Biomed. Anal. 1999, 19, 127. 10.1021/ac000286q CCC: $19.00
© 2000 American Chemical Society Published on Web 07/06/2000
Figure 1. Schematic drawing of the carbon paste electrode: (a) carbon paste; (b) working electrode contact; (c) FEP plastic tube; (d) ac heating device. Shadowed areas represent epoxy resins. See text for details.
commonly employs metal (wire or layer) electrodes, its adaptation to nucleic acid stripping analysis has required the development of a new thermally modulated carbon paste electrode (Figure 1). Such a new electrically heated electrode extends the scope of hotwire electrochemistry to additional analytes. While the concept of nucleic acid analysis at such electrodes is demonstrated below in connection with adsorptive stripping measurements, the new technology holds great promise for diverse applications ranging from thermal control of DNA hybridization to on-chip flow detection of nucleic acids. The performance characteristics of the new adsorptive stripping protocol at the heated carbon paste electrode are reported in the following sections. EXPERIMENTAL SECTION Apparatus. A TraceLab potentiometric stripping unit (PSU20, Radiometer) in connection with an IBM PS/2 55SX personal computer and controlled by the TAP2 software was used for the constant-current stripping chronopotentiometric measurements. The differentiated signal (dt/dE) was recorded against the potential, and the peak area served as the analytical signal. Cyclic voltammetry was performed with a BAS voltammetric analyzer connected to a 386 PC and controlled by the CV-50W software. The open-circuit potential measurements during heating and subsequent cooling to room temperature were performed with the Autolab PGSTAT10 electrochemical analyzer (Eco Chemie BV) connected to a Compaq PS/2 55 SX computer and controlled by the GPES software. A pH meter (Fisher Scientific) was used for open-circuit potential measurements during the temperature calibration. All experiments (except temperature calibrations) were carried out in a glass cell containing 5.0 mL of solution. Temperature calibrations were done in a commercial 100-mL glass cell (Schott). The three-electrode system consisted of a heated or unheated carbon paste electrode, an Ag/AgCl (3M NaCl) reference electrode (model RE-1, BAS Inc.), and a platinum wire auxiliary electrode. The working electrode was heated by a laboratory-made sine-wave power generator that was connected to the assembly via a high-frequency transformer. The frequency of the alternating current was 100 kHz in all experiments. Detailed information about the heating devices and techniques can be found in references.8,11,12,18 Heating was turned on or off automatically with the TAP2 software. All water and pipet tips were sterilized by autoclaving for 30 min. (18) Tadesse Zerihun, Gru ¨ ndler, P. J. Electroanal. Chem. 1996, 404, 243.
Reagents. Double-stranded calf thymus DNA (dsDNA, Catalog No. D4522), single-stranded calf thymus DNA (ssDNA, Catalog No. D8899), transfer RNA (tRNA, Catalog No. R8759), diethyl pyrocarbonate (DEPC), sodium acetate buffer (3 M, certified free of DNase and RNase, Catalog No. S7855), potassium chloride, and potassium ferrocyanide and ferricyanide were received from Sigma. Oligo(dG)20 was obtained from Life Technologies, and deoxyguanosine 5′-monophosphate disodium salt ((dG)1) was obtained from Fluka. All stock solutions (1000 mg L-1) of dsDNA, ssDNA, and oligo(dG)20 were prepared in buffer (10 mM TrisHCl, 1 mM EDTA, pH 8.0). The tRNA stock solution (1000 mg L-1) was prepared in DEPC-treated water.19 The (dG)1 stock solution (0.1 M) was prepared in water. The ferro-/ferricyanide solution was prepared in 0.1 M KCl. All other solutions were prepared in 0.2 M acetate buffer (pH 5.0). Construction and Preparation of the Working Electrodes. To construct the carbon paste electrode, three electrical connections were glued into a FEP plastic tube (length 18 mm, outer diameter 3.2 mm, inner diameter 1.6 mm) with epoxy resin (Figure 1). A 7-mm-long segment was resected from the middle and the tube was filled with carbon paste. The carbon paste was prepared by hand mixing graphite powder (grade 38; Fisher Scientific) and mineral oil (free of DNase, RNase or protease; Sigma) with a mass ratio of graphite powder to mineral oil of 70: 30. Prior to each series of measurements, the carbon paste surface was renewed and carefully polished with weighing paper. Temperature calibration was performed before starting a set of experiments. The stationary temperature of the carbon paste electrode during permanent heating was measured following earlier procedures.18 The gold wire electrode was similar to that described earlier. 11 Procedure for Adsorptive Stripping Measurements. All measurements were performed in 0.2 M acetate buffer (pH 5.0). The electrode was pretreated by applying a potential of +1.70 V for 10 s followed by adsorptive accumulation of the nucleic acids at +0.50 V. Subsequent chronopotentiometric stripping measurements were performed with a constant oxidizing current of 5.0 µA and an initial potential of +0.50 V. In the case of heated measurements, the alternating heating current was applied only during accumulation at +0.50 V and turned off 5 s prior to stripping. Before evaluating the stripping curve, the curve data were filtered and baseline corrected using the procedures available in the TAP2 TraceLab software. All experiments were carried out without mechanical stirring and without removal of dissolved oxygen. RESULTS AND DISCUSSION The coupling of hot-wire electrochemistry and adsorptive stripping analysis of nucleic acids has been accomplished by constructing an electrically heated carbon paste electrode (Figure 1). Knowledge of the temperature profile is essential for the adaptation of the new electrodes for “hot-layer” electrochemical protocols. Figure 2 displays the open-circuit potentiometric response (which corresponds with the temperature difference) of the carbon paste electrode in comparison to a heated gold wire electrode. The wire electrode displays a much faster (5 times) (19) Maniati, T.; Fritzsch, E.; Sambrook, J. Molecular Cloning: A Laboratory Manual; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY, 1982.
Analytical Chemistry, Vol. 72, No. 16, August 15, 2000
3753
Figure 2. Temperature over room temperature (∆T) vs time profile during heating and after heating for (A) a gold wire electrode (heated to 85 °C) and (B) carbon paste electrode (heated to 39.5 °C). Heating current is turned on at (a) and turned off at (b).
Figure 3. Cyclic voltammograms for 10 mM Fe(CN)63-/4-, 0.1 M KCl obtained after heating the carbon paste electrode for 3 min at (a) 20, (b) 41.5, and (c) 50.5 °C. Scan rate, 100 mV/s. Table 1. Cyclic Voltammetric Data for 10 mM Fe(CN)63-/4-, 0.1 M KCl at the Carbon Paste Electrode for Different Surface Temperaturesa temp/°C
∆Ep/mV
E°/mV
ip,anodic/ip,cathodic
20 80 98
363 284 284
206 164 162
0.85 0.92 1.04
aScan
between -0.40 and +0.75 V. Scan rate, 100 mV/s.
response, due to its very low heat capacitance. The carbon electrode needs ∼5 s to reach its hot or cold temperature values. Note that the standardized (in relation to temperature) T vs time functions are equal to one another for a given electrode. The voltammetric behavior of the new heated carbon paste electrode configuration was tested first using the ferrocyanide redox marker. Such cyclic voltammograms recorded at different temperatures are shown at Figure 3. The drawn-out voltammograms observed at room temperature (∆Ep ) 330 mV, a) reflect the quasi-reversible redox process, associated with the composite character of carbon paste electrodes. Raising the temperature to 41.5 (b) and 50.5 °C (c) resulted in larger and sharper peaks and smaller peak separations (down to 230 mV). Temperature effects are summarized in Table 1. With carbon paste electrodes, contradictory influences affect the appearance of the curves. The inner resistivity of the electrode material is higher than with metallic electrodes. The iR drop at this inner resistance contrib3754 Analytical Chemistry, Vol. 72, No. 16, August 15, 2000
Figure 4. Cyclic voltammograms after background subtraction for 1 mM (dG)1 obtained at (a) 20, b) 28, and (c) 36 °C. Scan rate, 100 mV/s.
utes to peak separation. The latter should increase with electrolysis current magnitude. On the other hand, with increasing temperature, the specific resistance of carbon decreases. This should bring about smaller peak separation. There is, however, overlaid another mechanism. With elevated temperature, the carbon particles inside the paste tend to lose contact to some extent, due to volume extension with increasing temperature. This (again) means higher resistance and, consequently, increased peak separation. Accordingly, the improvements in Figure 3 can be attributed mainly to enhanced convective transport, but altogether, there is not a clear depiction about the temperature effects on CV curve shape. With reversible redox systems, the temperature effect on kinetics of charge transfer is not relevant. Thermal treatment was shown to enhance the electron exchange at carbon paste electrodes through the removal of the inhibitory oil layer from the graphite surface. 20 Yet, this possibility is ruled out in the present case, as the lowering the temperature back to 20 °C resulted in return to a drawn-out voltammogram (as in Figure 3a; i.e., no lasting surface pretreatment). The influence of the electrode heating on the cyclic voltammetric response to (dG)1 is shown in Figure 4. Besides the increase in peak area due to mass transport improvement, we can also observe a peak shift to lower potentials at higher temperatures. This means that the analyte is more easily oxidized at elevated temperatures. Figure 5 demonstrates the suitability of heated carbon paste electrodes for adsorptive constant-current stripping chronopotentiometry of nucleic acids. Such stripping potentiograms for ssDNA (A), oligo(dG)20 (B), and tRNA (C) were recorded sequentially by using electrode temperatures of 20 (a) and 32 °C (b). All nucleic acids displayed a small guanine peak following the 1-min accumulation at room temperature (a). These peaks are substantially (5-7-fold) enhanced by heating the electrode during the accumulation step (b). The background and corresponding noise are not affected by this heating. (Note that the stripping step has been carried out at room temperature.) The greatly improved signal-to-background characteristics associated with the hot-layer operation thus offers convenient quantitation of trace levels of nucleic acids following very short accumulation periods. No memory effects are observed upon alternating between the low (20) Rice, M.; Galus, Z.; Adams, R. N. J. Electroanal. Chem. 1983, 143, 89.
Figure 7. Influence of the temperature during the accumulation step on the stripping time for 5 mg L-1 dsDNA. Other conditions as in Figure 5.
Figure 5. Baseline-corrected stripping potentiograms for (A) 5.0 mg L-1 ssDNA, (B) 1.0 mgL-1 guanosine 5′-monophosphate (dG)20, and (C) 1.0 mg L-1 tRNA following sequential exposure of the carbon paste electrode to (a) 20 and (b) 32 °C during accumulation. Pretreatment at +1.70 V for 10 s followed by accumulation at +0.50 V for 1 min and stripping with an applied oxidizing current of 5.0 µA. Measurements performed in 0.2 M acetate buffer (pH 5.0).
Figure 8. Baseline-corrected stripping potentiograms for ssDNA from (a) 1.0, (b) 2.0, (c) 3.0, (d) 4.0 and (e) 5.0 mg L-1. Accumulation performed at 20 (dotted line) and 32 °C (solid line). Calibration plots for ssDNA with accumulation at (A) 20 and (B) 32 °C are shown as an inset. Other conditions as in Figure 5. Figure 6. Influence of the temperature during the accumulation step on the stripping time for (a) 1 mg L-1 tRNA and (b) 5 mg L-1 ssDNA. Other conditions as in Figure 5.
and high temperatures. This indicates that the main role of the high temperature is to enhance the preconcentration efficiency (e.g., through faster localized convection and faster kinetics) and not to preactivate the surface or only accelerate the guanine oxidation. The ability to eliminate the stirring of the solution also facilitates the use of small sample volumes. Figure 6 illustrates the effect of the electrode temperature (during the 1-min accumulation) upon the stripping response to tRNA (a) and ssDNA (b). The response of both nucleic acids increases slowly upon raising the temperature between 20 and 32 °C, faster between 32 and 36 °C, and very rapidly between 36 and 47 °C. The enhancement is 13-fold for tRNA and 20-fold for ssDNA. In Figure 7 the same effect is demonstrated in the case of dsDNA. Here we find the strongest temperature effect, which leads to a 34-fold enhancement. So one can say that the temperature effect depends on the chain length of the nucleic acid molecule. The reason could be a change of structure in the molecule which is suspected to be temperature dependent. Moreover, there are premelting effects at elevated temperatures in the case of dsDNA which have to be considered. Faster molecular movement and changes in structure could facilitate the adsorption. More electrochemical active sites of the nucleic acid molecules could come into close contact with the electrode surface. Then the signal would be much higher. In highly complex molecules such as dsDNA, this effect should be relatively strong compared to less complex ones such as tRNA.
Figure 8 displays a calibration series of stripping potentiograms for ssDNA in a range between 1 and 5 mg/L. The deposition step was carried out at 32 (solid line) and at 20 °C (dotted line), respectively. The corresponding calibration plots (1-8 mg/L) are shown as an inset. Both temperatures allow quantitation at these levels and result in linear calibration plots. Yet, the only moderately heated electrode offers a considerable improvement of the sensitivity (slopes of 1.15 vs 0.27 ms L/mg) and, hence, greatly improved signal-to-noise characteristics (note that the elevated temperature does not affect the background). Indeed, a detection limit of ∼0.5 mg/L (0.5 ppm) was estimated from the signal-tonoise characteristics (S/N ) 3) of the potentiometric stripping response to 1 mg/L ssDNA following a 1-min accumulation at 32 °C. In conclusion, we have demonstrated the dramatic temperature effect on adsorptive accumulation of nucleic acids at carbon and the design of a new heated carbon paste electrode. The effect strongly depends on the chain length and structure of the analyte molecules. This offers not only new methods to explore behavior and properties of nucleic acids but also new techniques to determine them. Such use of heated carbon electrodes extends the scope of hot-wire electrochemistry to a wide range of electroanalytical applications. Thermal manipulations beyond adsorptive accumulation, including controlled DNA hybridization and denaturation (including discrimination among hybrids with varying binding strengths), and on-chip flow detection of nucleic acids could be the subject of further investigations. The technology also offers instantaneous detection of thermally induced structural changes. An important advantage would be, for instance, the Analytical Chemistry, Vol. 72, No. 16, August 15, 2000
3755
possibility of proper adjusting of the electrode temperature for increasing the rate and efficiency of the DNA hybridization at the electrode surface. These new opportunities will be explored in our laboratories soon.
Forschungsgemeinschaft (DFG) to P.G. G.R. and E.S. acknowledge CONICET (Argentina) and STINT (Sweden) for financial support. The authors are grateful to A. M. Oliveira Brett (Coimbra, Portugal) for cooperation and support.
ACKNOWLEDGMENT This work was supported by grants from the National Institutes of Health (NIH Grant RR14549-01) to J.W., and from the Deutsche
Received for review March 10, 2000. Accepted May 31, 2000.
3756
Analytical Chemistry, Vol. 72, No. 16, August 15, 2000
AC000286Q
