Anal. Chem. 2005, 77, 2624-2631
5-Hydroxytryptophan as a Precursor of a Catalyst for the Oxidation of NADH Noemı´ de-los-Santos-A Ä lvarez,†,‡ Marı´a Jesu´s Lobo-Castan˜o´n,‡ Arturo J. Miranda-Ordieres,‡ Paulino Tun˜o´n-Blanco,‡ and He´ctor D. Abrun˜a*,†
Department of Chemistry & Chemical Biology, Baker Laboratory, Cornell University, Ithaca, New York 14853-1301, and Departamento de Quı´mica Fı´sica y Analı´tica, Universidad de Oviedo, Julia´ n Claverı´a 8, 33006 Oviedo, Spain
Following oxidation of 5-hydroxytryptophan (5-HTPP) at a pyrolytic graphite electrode at pH 7.5, two quasireversible redox couples emerge at -0.170 and +0.032 V, respectively, due to oxidation products strongly adsorbed to the electrode surface. These redox processes have been electrochemically and kinetically characterized in terms of the dependence of the formal potential (E°′) with pH, variation of the current density with scan rate, operational stability, and electron-transfer rate constant (ks). The wave centered at +0.032 V could mediate the oxidation of NADH, exhibiting a strong and persistent electrocatalytic response. A quinone-imine structure has been proposed as the electrocatalytically active species. The kinetics of the reaction between the mediator and NADH has been characterized via rotating disk electrode voltammetry, and it has been found that the rate constant for the reaction is dependent on the solution concentration of NADH. 5-HTPP modified electrodes could be employed in the amperometric detection of NADH with a limit of detection in the nanomolar range. Moreover, 5-HTPP modified electrodes retain their electrocatalytic activity for at least one week. The potential application of these electrodes to amperometric biosensor is demonstrated. The use of electron-transfer mediators along with the development of new electrode materials with improved electrochemical properties1-3 are some of the most frequently used strategies to overcome the inherent difficulties that the oxidation of NADH exhibits at solid electrodes. On one hand, high overpotentials are required (over the range of 0.8-1.6 V depending on the electrode material), which limits the selectivity of the determination in real samples. On the other hand, the strong adsorption of not only the reduced coenzyme itself but also of its oxidation products causes fouling of the electrode and, as a consequence, loss of sensitivity, reproducibility, and operational lifetime. The importance of solving these drawbacks lies in the fact that NAD+/NADH is a cofactor for over 300 dehydrogenase enzymes. * To whom correspondence should be addressed. Fax: (+1) 607-255-9864. E-mail:
[email protected]. † Cornell University. ‡ Universidad de Oviedo. (1) Wang, J.; Musameh, M. Anal. Chem. 2003, 75, 2075-2079. (2) Ben-Ali, S.; Cook, D. A.; Evans, S. A. G.; Thienpont, A.; Bartlett, P. N.; Kuhn, A. Electrochem. Commun. 2003, 5, 747-751. (3) Ramesh, P.; Sampath, S. Anal. Chem. 2000, 72, 3369-3373.
2624 Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
Therefore, a single device would enable the determination all of these substrates, as well as the enzymatic activity of the enzymes by only changing the enzyme or the substrate employed, respectively. The potential versatility of such a device justifies the great deal of effort that has been devoted to this research area, especially in the search for improved mediators to accelerate the interfacial electron transfer between NADH and a solid electrode. Although mediators are still being employed in solution,4,5 most of them are immobilized on or incorporated into the electrode material. For this application, one-electron-transfer mediators such as ferrocene and its derivatives6,7 or hexacyanoferrate8 are considered to be less effective than two-electron mediators. In fact, the value of the apparent second-order rate constant between the reduced cofactor and this type of catalysts is very low (8 M-1‚s-1 for ferrocenecarboxylic acid).9 This is one of the reasons why research has focused mainly on the latter. The numerous materials employed as NADH mediators include compounds with o-quinone functionalities,10-14 redox dyes with phenoxazine, phenothiazine, or phenazine structures in monomeric3,15-17 or polymeric18,19 forms, transition metal complexes,20,21 organic22 and (4) Schro¨eder, I.; Steckhan, E.; Liese, A. J. Electroanal. Chem. 2003, 541, 109115. (5) Tham, S. Y.; Pearson, J. E.; Kane, J. W.; Treloar, P. H.; Vadgama, P. M. Sens. Actuators, B 1998, 50, 204-209. (6) Bu, H.; Mikkelsen, S. R.; English A. M. Anal. Chem. 1998, 70, 4320-4325. (7) Pessoˆa, C. A.; Gushikem, Y.; Kubota L. T. Electrochim. Acta 2001, 46, 24992505. (8) Eftekhari A. Sens. Actuators, B 2001, 80, 283-289. (9) Antiochia, R.; Lavagnini, I.; Magno F. Electroanalysis 1999, 11, 129-133. (10) Pariente, F.; Lorenzo, E.; Abrun ˜a, H. D. Anal. Chem. 1994, 66, 4337-4344. (11) Golabi, S. M.; Zare, H. R.; Hamzehloo, M. Electroanalysis 2002, 14, 611618. (12) Ramesh, P.; Sivakumar, P.; Sampath, S. J. Electroanal. Chem. 2002, 528, 82-92. (13) Ciszewski, A.; Milczarek, G. Anal. Chem. 2000, 72, 3203-3209. (14) Tang, J.; Wu, Z.; Wang, J.; Wang, E. Electrochem. Commun. 2000, 2, 796799. (15) Pessoa, C. A.; Gushikem, Y.; Kubota, L. T.; Gorton, L. J. Electroanal. Chem. 1997, 431, 23-27. (16) Munteanu, F. D.; Dicu, D.; Popescu, I. C.; Gorton, L. Electroanalysis 2003, 15, 383-391. (17) Santos, A. de S.; Gorton, L.; Kubota, L. T. Electrochim. Acta 2002, 47, 33513360. (18) Gao, Q.; Wang, W.; Ma, Y.; Yang, X. Talanta 2004, 62, 477-482. (19) Zhou, D.; Sun, J. J.; Chen, H. Y.; Fang, H. Electrochim. Acta 1998, 43, 18031809. (20) Wu, G.; Maskus, M.; Pariente, F.; Ferna´ndez, V. M.; Lorenzo, E.; Abrun ˜a, H. D. Anal. Chem. 1996, 68, 3688-3696. (21) Popescu, I. C.; Domı´nguez, E.; Narva´ez, A.; Pavlov, V.; Katakis, I. J. Electroanal. Chem. 1999, 464, 208-214. 10.1021/ac048554y CCC: $30.25
© 2005 American Chemical Society Published on Web 03/16/2005
inorganic23 conducting salts, and conducting polymers.24-26 More recently, two new families of precursors of NADH mediators have been described: nitroaromatic compounds27-29 and N9-substituted adenine derivatives.30-33 These require the previous reduction or oxidation, respectively, to form the electrocatalytically active species. In the case of N9-substituted adenine derivatives, the catalytic species has been attributed to an oxidation product having a quinone-imine structure. Despite the extensive work carried out, some difficulties remain. In many cases, the desorption of the mediator limits the long-term stability of the device. Other mediators exhibit formal potentials too high to prevent interferences or too low to significantly catalyze the oxidation of NADH. Finally, many catalysts or matrices where they are immobilized such as silica gel are unstable in alkaline solutions, in which some dehydrogenase enzymes present their maximum activity. It is well known that the electrooxidation of 5-hydroxytryptophan (5-HTPP) gives rise to a tryptophan-4,5-dione34-38 product that strongly adsorbs to carbon electrodes. In principle, this compound, with an o-quinone structure, is expected to act as a catalyst for the oxidation of NADH. We have recently found, however, that the oxidation of 5-HTPP in neutral solutions yields an additional surface-bound redox-active species. This oxidation product exhibits a strong and persistent electrocatalytic effect toward the oxidation of NADH at low potentials, which is of great interest to minimize interferences and for biosensor development. We herein describe the electrochemical and kinetic characterization of 5-HTPP modified pyrolytic graphite electrodes and their electrocatalytic activity toward the oxidation of NADH. In addition, the kinetics of the electrocatalytic process has been investigated. Moreover, the analytical features of such modified electrodes designed for the amperometric detection of NADH have also been studied. EXPERIMENTAL SECTION Materials. 5-Hydroxy-L-tryptophan and β-nicotinamide adenine dinucleotide reduced form (NADH) (disodium salt, 98%) were (22) Kataky, R.; Bryce, M. R.; Goldenberg, L.; Hayes, S.; Nowak, A. Talanta 2002, 56, 451-458. (23) Cai, C. X.; Xue, K. H.; Zhou, Y. M.; Yang, H. Talanta 1997, 44, 339-347. (24) Toh, C.-S.; Bartlett, P. N.; Mano, N.; Aussenac, F.; Kuhn, A.; Dufourc, E. J. Phys. Chem. Chem. Phys. 2003, 5, 588-593. (25) Bartlett, P. N.; Simon, E. Phys. Chem. Chem. Phys. 2000, 2, 2599-2606. (26) Lobo, M. J.; Miranda, A. J.; Lo´pez-Fonseca, J. M.; Tun ˜o´n, P. Anal. Chim. Acta 1996, 325, 33-42. (27) Mano, N.; Kuhn, A. J. Electroanal. Chem. 1999, 477, 79-88. (28) Mano, N.; Thienpont, A.; Kuhn, A. Electrochem. Commun. 2001, 3, 585589. (29) Casero, E.; Darder, M.; Takada, K.; Abrun ˜a, H. D.; Pariente, F.; Lorenzo, E. Langmuir 1999, 15, 127-134. (30) AÄ lvarez Gonza´lez, M. I.; Saidman, S. B.; Lobo Castan ˜o´n, M. J.; Miranda Ordieres, A. J.; Tun ˜o´n Blanco, P. Anal. Chem. 2000, 72, 520-527. (31) de los Santos AÄ lvarez, N.; Lobo Castan ˜o´n, M. J.; Miranda Ordieres, A. J.; Tun ˜o´n Blanco, P. J. Electroanal. Chem. 2001, 502, 109-117. (32) de-los-Santos-AÄ lvarez, N.; Lobo-Castan ˜o´n, M. J.; Miranda-Ordieres, A. J.; Tun ˜o´n-Blanco, P. Anal. Chim. Acta 2004, 504, 271-277. (33) de-los-Santos-AÄ lvarez, P.; Molina, P. G.; Lobo-Castan ˜o´n, M. J.; MirandaOrdieres, A. J.; Tun ˜o´n-Blanco, P. Electroanalysis 2002, 14, 1543-1549. (34) Humphries, K. A.; Wrona, M. Z.; Dryhurst, G. J. Electroanal. Chem. 1993, 346, 377-403. (35) Humphries, K.; Dryhurst, G. J. Pharm. Sci. 1987, 76, 839-847. (36) Humphries, K.; Dryhurst, G. J. Electrochem. Soc. 1990, 137, 1144-1149. (37) Cohen, J. L.; Widera, J.; Cox, J. A. Electroanalysis 2002, 14, 231-234. (38) Verbiese-Genard, N.; Kauffmann, J. M.; Hanocq, M.; Molle, L. J. Electroanal. Chem. 1984, 170, 243-254.
obtained form Sigma-Aldrich (Madrid, Spain) and used as received. Other chemicals employed were of analytical grade. Water was purified with a MilliQ system (Millipore). Apparatus. Cyclic voltammetric measurements were carried out using a computer-controlled Autolab Pgstat-10 potentiostat (EcoChemie). Rotating disk electrode (RDE) studies were performed with a Methrom 628-10 RDE system. All measurements were made using a conventional three-electrode system. A platinum wire served as auxiliary electrode and a Ag|AgCl|KCl(sat) electrode as reference. The working electrode was a 3-mmdiameter pyrolytic graphite rod (Goodfellow, UK) sealed into a homemade polyethylene holder (ADEPRO, Avile´s, Spain) in either the static or rotating designs. Electrical contact was made through a two-part silver epoxy resin. To obtain reproducible surfaces, the pyrolytic graphite electrode (PGE) was polished on sandpaper (600 grit, Buehler), followed by sonication in purified water for 2 min. Amperometric measurements were carried out with a stationary PGE immersed into a magnetically stirred solution. A fixed potential of +0.130 V was applied using a PAR M-400 (EG&G USA) electrochemical detector, whose output was connected to a Metrohm E586 Labograph strip chart recorder. Procedures. Electrode Modification with 5-HTPP. Unless otherwise stated, the PGE was immersed into a 0.1 M KCl solution containing 0.4 mM 5-HTPP for 90 s, rinsed with water, and then placed in a pH 7.5 0.1 M phosphate buffer solution, where the potential was cycled between -0.3 and +0.9 V five times at 50 mV‚s-1. The solution was previously deoxygenated by bubbling a stream of N2 for at least 10 min, and the oxidation was carried out under an N2 atmosphere. Surface coverages (Γ) for the modified electrodes were determined by integration of the area under the corresponding voltammetric wave obtained at 50 mV‚s-1 and assuming a twoelectron process. The surface coverage was normalized to the effective electrode area estimated by chronoamperometric measurements using 0.1 M [Fe(CN)6]-4 in 1 M KCl solution. A value of 6.3 × 10-6 cm2/s was used for the diffusion coefficient of [Fe(CN)6]-4 at 25 °C. The average area of the PGE was 0.075 cm2. RESULTS AND DISCUSSION Electrochemical Characterization of the Oxidation Products of 5-HTPP. Initially, the oxidation of 5-HTPP was carried out by cyclic voltammetry with a PGE placed in 0.1 M KCl solutions containing 0.4 mM 5-HTPP. However, it has been reported that the oxidation of concentrated solutions of 5-HTPP results in the buildup of a product film on the electrode surface34 or even passivation.37,38 For that reason, the compound was adsorbed for 30 s from 0.1 M KCl solutions containing 0.4 mM 5-HTPP, and subsequently, the electrode was transferred into a pH 7.5 0.1 M phosphate buffer solution, where the oxidation was performed as described in the Experimental Section. As can be seen in Figure 1A, on the first voltammetric scan, a sharp anodic peak (Ia) is evident at +0.345 V. Upon scan reversal at +0.430 V, a small and broad cathodic wave (IIIc) at about +0.051 V (indistinct in Figure 1A but evident in Figure 1B) and a sharp cathodic peak (IIc) at -0.175 V were observed. On subsequent potential scans, three additional anodic peaks appeared at -0.158 (IIa), +0.082 (IIIa), and +0.221 V (IVa). Peaks IIa and IIIa form Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
2625
Figure 1. (A) Cyclic voltammograms obtained in 0.1 M phosphate buffer pH 7.5 at 50 mV‚s-1 with a PGE previously immersed into 0.1 M KCl solution containing 0.4 mM 5-HTPP for 30 s, (dashed line) first scan and (solid line) fourth scan. Inset: Enlargement of fourth scan. (B) After 30 cycles between -0.3 and +0.43 V.
quasi-reversible redox couples with the cathodic peaks observed. If the potential is cycled between -0.3 and +0.43 V, there is a continuous decrease in the intensity of the initial anodic peak, Ia, along with an increase in the magnitude of redox couples II and III. It is interesting to note that whereas process II exhibits a symmetrical and sharp wave shape, process III is broader and seems to be made up of several overlapping redox processes with very similar formal potentials. After several scans (typically 30) between -0.3 and +0.43 V, the wave ascribed to process III becomes sharper and the formal potential is slightly shifted toward more negative values (Figure 1B). Peak Ia has been previously ascribed to the oxidation of 5-HTPP to a very reactive C4-centered carbocation that rapidly leads to several products, one of them being responsible for redox process IIa/IIc. 2626 Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
Redox process IIa/IIc corresponds to the reduction of tryptophan-4,5-dione (A) to 4,5-dihydroxytryptophan (B) in a 2 e-, 2 H+ reversible reaction (see Scheme 1).34 Peak IVa is due to the oxidation of another minor product generated after oxidation of peak Ia. When the potential was cycled within the limits of -0.30 and +0.23 V, i.e., when the potential was reversed only 10 mV beyond peak IVa, it was observed that the intensity of peak IVa decreased whereas the peak intensity of redox couple II increased slightly. This suggests that, after oxidation, this product also yields A. This behavior is consistent with the electrochemical features of the previously described dimer D.39 All of the structures involved are depicted in Scheme 1, where reactions have been simplified for clarity. To our knowledge, there are no reports concerning the appearance of process III during the oxidation of this compound in neutral solutions. It has, however, been mentioned when the oxidation takes place in acid solutions.35,36 In this case, the authors suggest that C is responsible for the redox activity of process III. Compound C is a proposed intermediate of a dimer with a quinoline structure. This material has been isolated and identified as the ultimate product of an exhaustive oxidation at pH 2. Although this intermediate (C) could not be isolated, an analogue of it was identified in the electrolyzed solution of two metabolites of 5-HTPP, 5-hydroxytryptamine (5-HT) at pH 2,40 and 5-hydroxytryptophol (5-HTOL) at physiological pH.41 These intermediates can be oxidized to a quinone-imine in a 2 e-, 2 H+ reversible reaction with E°′ of +0.3 V (SCE) at pH 240 or -0.042 V (SCE) at pH 7.4.41 In the case of the oxidation of 5-HTOL, up to four other different oxidation products were said to contribute, to varying degrees, to this redox process. Despite the fact that Dryhurst et al. did not find evidence for the formation of either the quinoline or its precursor C in electrolyzed solutions of 5-HTPP at neutral pH, we suggest that the intermediate C is a minor oxidation product, which might be generated and stabilized by strong adsorption onto graphite electrodes. As a result, little if any of either the intermediate compound or the product with a quinoline structure is observed in solution. If the potential is reversed at +0.9 V, a new anodic wave is observed at +0.850 V, which corresponds to the oxidation of tryptophan-4,5-dione (A) causing the amplitude of process II to dramatically diminish and, eventually, disappear after a few scans. The stability of both processes (II and III) was studied in acid and neutral solutions. After modification, the electrodes were rinsed with water and immersed into pH 3 or pH 7.5 phosphate solutions. The decay of the surface coverage with time was then determined under continuous cycling at 50 mV‚s-1 between -0.050 and +0.500 V and -0.300 and +0.300 V, respectively. As can be seen in Figure 2, in all cases, a significant decay in coverage was observed during the initial scans (typically 20 min) likely due to the loss of weakly bound material. Afterward, the rate of desorption slowed, except for process II at pH 3, where a continuous loss of material occurred. The surface coverage of process II (Figure 2A) that remained after 50 min of continuous cycling was higher at pH 3 (60%) than at pH 7.5 (23%). This is in good agreement with the reported instability of A above pH 7.2.34 (39) Wrona, M. Z.; Dryhurst, G. J. Electroanal. Chem. 1990, 278, 249-267. (40) Cheng, F. C.; Wrona, M. Z.; Dryhurst, G. J. Electroanal. Chem. 1991, 310, 187-218. (41) Wu, Z.; Shen, X. M.; Dryhurst, G. Bioorg. Chem. 1995, 23, 227-255.
Scheme 1. Reactions Involved in the Oxidation of 5-HTPP
In contrast, the loss of material for process III (Figure 2B) was almost equal in both solutions (13% vs 17%). From these data it is clear that the species responsible for process III is much more stable than that for process II. It is noteworthy that a better fit to the data was found for 1/coverage versus time plots than for ln coverage versus time plots in all cases except for process II in acid solution (insets in Figure 2). Therefore, the desorptive losses seem to obey secondorder kinetics except for process II at pH 3, where first-order kinetics were observed. The reason for this difference in behavior is not clear at this time. From the structures attributed to compounds related to process II (o-quinone) and III (p-quinone-imine) redox responses were expected to be pH dependent. To verify this, cyclic voltammograms using modified electrodes in solutions of varying pH from 2 to 9 were recorded at 50 mV‚s-1. For both redox couples, the E°′ shifted to more negative potentials with an increase in pH according to the following equations:
E (II)/V ) -0.058 pH + 0.264
r ) 0.9995
n)9
E (III)/V ) -0.059 pH + 0.478
r ) 0.9991
n)9
A slope of ∼59 mV/pH unit is the anticipated Nernstian value for a process involving an identical number of protons and electrons. The width at half-height (Ewhh) of the anodic peak IIa was 51 mV at 50 mV‚s-1. This value is very close to the one expected for a two-electron reaction (90/n) as in the case for the redox activity
of o-quinones in aqueous media. However, the Ewhh for process III is higher (typically 80 mV) suggesting, as stated earlier, that it is probably composed of several redox processes with very similar E°′ values. Nevertheless, after several scans, the Ewhh converged to 50 mV (Figure 1B), suggesting that two electrons are also involved in this process as expected for a quinone-imine reaction. Both processes exhibited the typical behavior of a surfaceimmobilized redox couple, i.e., symmetrical wave shape and small peak separation (Figure 1B). Additionally, the currents were directly proportional to the rate of potential sweep over the range of 20-200 mV‚s-1 for process II (I/µA ) 0.109 v/(mV‚s-1) - 2.2; r ) 0.999 n ) 11) and 20-500 mV‚s-1 for process III (I/µA ) 0.01567 v/(mV‚s-1) - 0.11; r ) 0.9998 n ) 25). Kinetics of Electron Transfer of Adsorbed Oxidation Products. The electron-transfer rate constant (ks) for a surface electrochemical reaction can be determined by cyclic voltammetry (CV) or by square wave voltammetry (SWV). The former method is based on measuring the shift of the peak potentials with scan rate, according to the general expressions derived by Laviron.42 So-called “Laviron plots” present Epeak - E° ′ versus log v. At high sweep rates, these plots become linear and from the slope and intercepts one can obtain the values of the transfer coefficient (R) and the rate constant (ks), respectively. A Laviron plot for redox couple II at pH 7.5 is shown in Figure 3A, and as is evident, for scan rates above 4 V‚s-1, the values of Epeak - E°′ are proportional to the logarithm of the scan rate. E°′ was taken as the average of the cathodic and anodic peak potentials at low scan rate. Using (42) Laviron, E.; J. Electroanal. Chem. 1979, 101, 19-28.
Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
2627
Figure 3. (A) Dependence of Epeak - E°′ of redox couple II on the logarithm of scan rate (Laviron plot) obtained at pH 7.5. (B) Dependence of the ratio between anodic peak current and the frequency at pH 3 for process III by SWV. Figure 2. Fractional coverage decay under continuos scanning at 50 mV‚s-1 for (A) process II and (B) process III (b) at pH 7.5 (O) at pH 3. Insets: first- and second-order plots for the variation in fractional coverage as a function of time.
Table 1. Values of Electron-Transfer Rate Constant (ks, s-1) and Transfer Coefficient (r) Obtained Using Two Methods in pH 3 and 7.5 from Three Determinations with Different Modified Electrodes pH 3
such plots, values of ks and R were obtained and they are presented in Table 1. The second method is based on the measurement of the SWV peak currents at different frequencies. Theoretically, the SWV peak current/frequency ratio (I/f) depends parabolically on the frequency. If the adsorption of both the product and the reactant is equally strong and the transfer coefficient is 0.5, the maximum value of the I/f ratio appears at a frequency approximately equal to the rate constant of the redox reaction (ks). This theory can be extended to asymmetric redox reactions and different adsorption constants.43 In such cases, the critical frequency (fmax) for which the I/f ratio reaches its maximum value is related to ks by the equation ks ) kmaxfmax (1), where kmax is a kinetic parameter that depends on R. The average value for kmax is 1.18 ( 0.05 for values of R between 0.25 and 0.85. Taking into account the values of R previously obtained by using the Laviron method, the average value of kmax was used. An I/f versus f plot for process III at pH (43) Komorsky-Lovric, S.; Lovric, M.J. Electroanal. Chem. 1995, 384, 115-122.
2628 Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
process II
process III
pH 7.5 process II
process III
SWV method 9.44 ( 0.01 20.0 ( 1.2 12.6 ( 0.7 24 ( 2 Laviron’s method 11.9 ( 0.5 20.4 ( 1.0 15.7 ( 1.8 23 ( 2 R 0.53 ( 0.08 0.48 ( 0.03 0.54 ( 0.02 0.49 ( 0.01
3 is shown in Figure 3B. From such plots and using eq 1, values of ks were determined and were consistent with those obtained using the classical CV method (Table 1). The rate of electron transfer is higher in neutral solutions for both compounds. An identical trend has been reported for adenine derivative-based catalysts.31 Furthermore, the surface electron-transfer reaction due to the species responsible for process II is slower than for the compound related to process III. Electrocatalytic Activity toward Oxidation of NADH. As mentioned before, not only o-quinones but also p-quinone-imines are well-known catalytic structures for the oxidation of NADH, as has been recently proposed for catalysts derived from N9adenine derivatives.30-33 To test the potential electrocatalytic behavior of the adsorbed compounds, cyclic voltammograms were
Figure 4. Cyclic voltammograms obtained in 0.1 M phosphate pH 7.5 at 10 mV‚s-1 with a PGE modified (A) as in Figure 1 and (B) after cycling the potential up to +0.9 V (dashed line) in the absence and (solid line) in the presence of 0.1 mM NADH. Inset: cyclic voltammograms obtained as in (B) with a bare PGE.
recorded at 10 mV‚s-1 in pH 7.5 phosphate solution in the absence and in the presence of 0.1 mM NADH, and the results are presented in Figure 4. When the solution contains the reduced coenzyme (NADH), an enhancement in the anodic peak current was observed at potentials close to the formal potential of redox couple III along with a decrease of the cathodic current. Concomitantly, a significant decrease in the amplitude of redox couple II occurred, which could suggest that this compound might also exhibit some electrocatalytic response. With the intent of ascertaining whether both or only one of the adsorbed species is able to act as catalyst of the oxidation of NADH, the dione was oxidized by cycling the potential up to +0.9 V. As described earlier, this gives rise to the disappearance of wave II so that only redox couple III could be observed in the absence of NADH (Figure 4B, dashed line). When 0.1 mM NADH was added to the solution, a dramatic enhancement of the anodic peak current was evident. This increase was much larger than that one obtained in the presence of adsorbed dione A. This behavior could be rationalized based on the large amount of
tryptophan-4,5-dione generated during the oxidation of 5-HTPP, which could inhibit the electrocatalytic effect probably by blocking the electrode surface. In fact, inhibition of the electron transfer resulting from electrode surface covering by the oxidized products of 5-HTPP has been previously reported.37,38 It should be mentioned that we cannot completely rule out some contribution of the dione to the electrocatalytic current since a small shoulder, at more negative potentials than the catalytic peak potential, could be (at times) observed when most (but not all) of the dione A was oxidized. No current peak was observed at potentials at which the direct uncatalyzed oxidation of NADH takes place, indicating the high efficiency of the electrocatalytic process. The catalytic efficiency for the oxidation of NADH can be evaluated from the ratio between the catalytic current and the current of the catalyst. Its value at an NADH concentration of 0.1 mM increased from 1.3 to 6.1 after removing redox couple II. To attempt to rule out the contribution, to the catalytic current, of surface quinones that might be formed during the oxidation step, a bare PGE was subjected to the identical oxidation step except in the absence of 5-HTPP. It was immersed into a 0.1 M KCl solution containing no 5-HTPP for 90 s, and then, it was transferred to pH 7.5, 0.1 M phosphate buffer solution, where the oxidation was performed as previously indicated. Afterward, NADH was added, but there was no evidence of any catalytic oxidation of NADH, unambiguously establishing that the electrocatalytic effect arises from the adsorbed compound (inset Figure 4B). With the aim of developing an optimum NADH sensing interface, modified electrodes, prepared by cycling the potential between -0.3 and +0.9 V, were selected to optimize other parameters affecting the catalytic current. The time of adsorption in 0.1 M KCl solution containing 0.4 mM 5-HTPP affects the surface coverage of the catalyst adsorbed on the electrode surface and, consequently, the electrocatalytic behavior. It was varied from 30 to 120 s. The surface coverage and the catalytic response to 0.1 mM NADH were measured. Both Γ and the catalytic current increased with the time of adsorption up to 90 s. Longer periods of time did not result in higher surface coverages, and in fact, the magnitude of the electrocatalytic current decreased. These findings suggest that not all of the deposited material is electrocatalytically active. The inactive material could impede the diffusion of NADH to the active centers of the catalyst adsorbed directly on the electrode surface, inhibiting the catalytic reaction. Thus, modified electrodes prepared after 90 s of adsorption were used in all further investigations. The pH of the detection solution could be changed in order to obtain better catalytic currents, diminish the overvoltage required for the oxidation of NADH, or both. Taking into account that the potential of redox process III was pH dependent, a similar trend would be expected for the catalytic peak potential. Nonetheless, although the formal potential shifted to more negative potentials in alkaline media (-50 mV when measured at pH 9), the magnitude of the catalytic current decreased. This trend has also been reported for other mediators with a quinone-imine structure.33,44 Therefore, pH 7.5 solutions were selected as more appropriate for NADH detection. (44) de-los-Santos-AÄ lvarez, N.; Lobo-Castan ˜o´n, M. J.; Miranda-Ordieres, A. J.; Tun ˜o´n-Blanco, P. Electroanalysis 2003, 15, 441-446.
Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
2629
Figure 6. Dependence of second-order rate constant (k1) and (inset) 1/k1 on the NADH concentration at pH 7.5.
Figure 5. (A) Levich and (B) Koutecky-Levich plots for the electrocatalyzed oxidation of (b) 1 and (O) 0.2 mM NADH obtained in 0.1 M phosphate buffer pH 7.5 with electrodes modified with 5-HTPP.
Kinetics of the Electrocatalytic Oxidation of NADH. Rotating disk electrodes were used to evaluate the kinetics of the reaction between the NADH and the catalyst immobilized on PGE. The catalytic current was measured in pH 7.5 at +0.130 V as a function of rotation speed. As can be seen in Figure 5A (Levich plot), the limiting current (Ilim) for the oxidation of 0.2 and 1 mM NADH, respectively, increased with increasing rotation rate (ω). A lack of linearity was, however, evident suggesting kinetic limitations. Slow electron transfer between the adsorbed mediator (catalyst) and the electrode can be ruled out as rate-limiting taking into account the values determined earlier. Under these conditions, a Koutecky-Levich plot (Ilim-1 vs ω-1/2 plot) can be used to determine the apparent rate constant (k1) for the catalytic reaction between the oxidized catalyst and NADH. The expected linearity was obtained for several NADH concentrations (Figure 5B). From the intercepts of the linear plots, the values of k1 for NADH concentrations over the range of 0.2-1 mM were determined and the results are shown in Figure 6. As can be seen, the value of k1 strongly depends on the NADH concentration. Such behavior has been previously reported for other mediators of the oxidation of NADH and supports an electrocatalytic reaction that proceeds through a mechanism similar to the Michaelis-Menten model. The catalyst and NADH form an intermediate complex 2630 Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
that dissociates to give rise to the reduced catalyst and NAD+. This is the rate-limiting step of the overall chemical reaction.45 A linear relationship between the reciprocal apparent constant and the concentration of NADH was obtained. From the slope and the intercept of this plot, the values of the apparent Michaelis-Menten constant (KM) and the rate constant for the complex dissociation reaction k+2 were found to be 4.1 × 10-5 M and 0.84 s-1, respectively. To compare the value of k1 with those reported in the literature, a linear extrapolation to zero NADH concentration was performed. The value obtained was 2.1 × 104 M-1‚s-1, which is of the same order as those of other mediators with the same catalytic functionality.31 Amperometric Detection of NADH. Modified electrodes with 5-HTPP were tested for the amperometric detection of NADH at an applied potential of +0.130 V in a magnetically stirred pH 7.5, 0.1 M phosphate buffer solution. The current increased with concentration over the range from 5 × 10-8 to 1 × 10-3 M. At higher NADH concentrations, there was a leveling off of the amperometric response as would be anticipated for a MichaelisMenten type process. A linear response was found between 5 × 10-8 and 6 × 10-6 M. From such a plot, a value of KM of 2.12 × 10-4 M was obtained. The discrepancy with the value obtained from rotating disk experiments could arise from the uncertainty in the measurement of the surface coverage of modified electrodes and in the value of the intercept of the 1/k1 versus NADH concentration plot. The relative standard deviation between electrodes was estimated from the slope of the calibration plot of five different and freshly prepared modified electrodes, and the value obtained was 4.7%. The limit of detection was estimated to be 27 nM. A mean response time of 9 s for the 95% of the steady-state signal was measured for a set of six different electrodes. Finally, the stability of the modified electrodes was assessed. A modified electrode was prepared, and several calibrations were (45) Gorton, L.; Torstensson, A.; Jaegfeldt, H.; Johansson, J. J. Electroanal. Chem. 1984, 161, 103-120.
carried out. The modified PGE was kept immersed in background solution when not in use. During the first day of operation, the amperometric response remained within 95% of the initial signal. Even though a slow loss in activity was detected during the following days, the modified electrode could still be used after a week of repetitive NADH additions. It is worth noting that the linear range did not decrease during the course of a week, although a slight increase in the response time was observed. Longer periods of time were not evaluated because the ease of modification allowed one to rapidly prepare a freshly modified electrode daily. CONCLUSIONS We have described the generation of a very efficient catalyst toward the oxidation of NADH derived from the electrochemical oxidation of 5-HTPP at PGE in neutral pH media. During such oxidation, several products are generated, two of them adsorb strongly onto the electrode surface. The electron-transfer rate constant, determined for both surface-bound compounds, is higher
in neutral solutions than in acidic media. For the electrocatalytically active species, higher values of ks have been found in both media and a quinone-imine structure is proposed. In fact, values of the kinetic constants of the electrocatalytic oxidation of NADH are very similar to those obtained with other catalysts with such a catalytic functionality. The modified electrodes exhibited very good analytical characteristics in terms of sensitivity, rapidity, stability, and reproducibility, making them attractive amperometric transducers in biosensors for substrates of dehydrogenase enzymes or for the measurement of enzymatic activity. ACKNOWLEDGMENT N.S.-AÄ . thanks the Ministerio de Educacio´n, Cultura y Deportes (Spain) for FPU grant.
Received for review September 28, 2004. Accepted February 10, 2005. AC048554Y
Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
2631
