Anal. Chem. 2005, 77, 4286-4289
Flavin Adenine Dinucleotide As Precursor for NADH Electrocatalyst Noemı´ de-los-Santos-A Ä lvarez, Patricia de-los-Santos-A Ä lvarez, M. Jesu´s Lobo-Castan˜o´n, Arturo J. Miranda-Ordieres, and Paulino Tun˜o´n-Blanco*
Departamento de Quı´mica Fı´sica y Analı´tica, Universidad de Oviedo, 33006 Oviedo, Spain
The generation of a new electrocatalytic system for NADH after oxidizing flavin adenine dinucleotide (FAD) is shown. The oxidation is performed in alkaline medium until +1.4 V (Ag/AgCl) at graphite electrodes. The catalytic activity is ascribed to the electrooxidized moiety of FAD and not to quinone surface groups. A comparison between this catalyst and that attributed to poly(FAD) (Karyakin, A. A.; Ivanova Y. N.; Revunova, K. V.; Karyakina, E. E. Anal. Chem. 2004, 76, 2004-2009.) is presented. It is concluded that the surface quinone groups generated during the strong anodization of the electrode in acidic medium at 2-2.5 V and not the poly(FAD) are responsible for the catalytic activity described in the above mentioned work. The use of NAD+-dependent dehydrogenase enzymes for biosensing has generated a great deal of interest. Typically, the detection is performed by measuring the amount of NADH produced in the enzymatic conversion of the analyte. As NADH is an electrochemically active species, one of the alternatives is the use of electrochemical detection systems. However, the oxidation of the cofactor at unmodified carbon, gold, or platinum electrodes occurs at high applied potentials and gives rise to different oxidation products, which are adsorbed on the electrode surface causing a gradual decrease in the electrochemical signal.1-3 Many investigators have thus turned their attention to find different catalysts that could reduce the high overvoltage needed for NADH oxidation and the surface fouling effects.1-4 Quinones, quinoneimines, and phenylenediimines are typical catalytic functionalities. Compounds containing these functionalities have been immobilized on the electrode surface by adsorption, film entrapment, covalent attachment, or electropolymerization. The prosthetic group, flavin adenine dinucleotide (FAD), is one of the molecules involved in the electron-transfer chain in living organisms. Until now, two different ways of obtaining NADH electrocatalysts from FAD have been described, both involving the riboflavin moiety of the molecule. The first one was proposed by Gorton et al.5 and relies on the surface reversible redox process * Corresponding author: (e-mail)
[email protected]. (1) Gorton, L.; Domı´nguez, E. In Encyclopedia of Electrochemistry. Vol. 9: Bioelectrochemistry; Wilson G. S., Ed.; Wiley-VCH: Weinheim, 2002; Chapter 4, pp 67-143. (2) Gorton, L.; Domı´nguez, E. Rev. Mol. Biotechnol. 2002, 82, 371. (3) Lobo, M. J.; Miranda, A. J.; Tun ˜o´n, P. Electroanalysis 1997, 9, 191-202. (4) Valentini, F.; Salis, A.; Curulli, A.; Palleschi, G. Anal. Chem. 2004, 76, 32443248.
4286 Analytical Chemistry, Vol. 77, No. 13, July 1, 2005
yielded by monomeric adsorbed FAD on carbon fiber electrodes. The authors used TiO2-modified electrodes to shift the formal potential of this redox couple, thus improving the kinetics of the catalytic reaction. The second one has been recently proposed by Karyakin et al.,6 and it is described as an electropolymerization reaction leading to a polyazine7 with high electrocatalytic activity toward oxidation of NADH. The electropolymerization reaction occurs at extreme positive potentials in aqueous acidic medium, yielding a surface redox pair with a formal potential of 0 V at pH 6.0. However, they did not consider that at so high an applied potential the adenine moiety of FAD is irreversibly oxidized. According to our previous results,8-10 the oxidation of the adenine moiety in N9-substituted adenine compounds (adenosine, adenosine mono- di- or triphosphate, or even NAD+) at pyrolytic graphite electrodes in alkaline solutions leads to the corresponding 2,8dioxoadenine derivative, which strongly adsorbs on the electrode surface. These oxidation products, with a quinoneimine structure, act as excellent catalysts for oxidation of NADH. A similar behavior would be expected for FAD, which is a flavin but also an adenine compound containing a large N9 substituent (Figure 1). In this correspondence, we examine the possibility of generating a new electrocatalytic system for NADH after oxidizing the adenine moiety of FAD. A comparison between this new catalyst and those involving the riboflavin function of the molecule5,6 is presented. EXPERIMENTAL SECTION Materials. FAD, flavin adenine mononucleotide (FMN), adenosine diphosphate (ADP), and nicotinamide adenine dinucleotide, reduced form (β-NADH) were purchased from Sigma (Madrid, Spain) and were used as received. Other chemicals were of analytical grade. Electrochemical Measurements. Voltammetric measurements were carried out with a potentiostat Autolab Pgstat-10 (EcoChemie). A conventional three-electrode electrochemical cell (5) Kubota, L. T.; Gorton, L.; Roddick-Lanzilotta, A.; McQuillan, A. J. Bioelectrochem. Bioenerg. 1998, 47, 39-46. (6) Karyakin, A. A.; Ivanova Y. N.; Revunova, K. V.; Karyakina, E. E. Anal. Chem. 2004, 76, 2004-2009. (7) Ivanova, Y. N.; Karyakin, A. A. Electrochem. Commun. 2004, 6, 120-125. (8) AÄ lvarez-Gonza´lez, M. I.; Saidman, S. B.; Lobo-Castan ˜o´n, M. J.; MirandaOrdieres, A. J.; Tun ˜o´n-Blanco, P. Anal. Chem. 2000, 72, 520-527. (9) de los Santos AÄ lvarez, N.; Mun ˜´ız Ortea, P.; Montes Pan ˜eda, A.; LoboCastan ˜o´n, M. J.; Miranda-Ordieres, A. J.; Tun ˜o´n-Blanco, P. J. Electroanal. Chem. 2001, 502, 109-117. (10) de los Santos AÄ lvarez, P.; Lobo-Castan ˜o´n, M. J.; Miranda-Ordieres, A. J.; Tun ˜o´n Blanco, P. Anal. Chem. 2002, 74, 3342-3347. 10.1021/ac048545p CCC: $30.25
© 2005 American Chemical Society Published on Web 05/18/2005
Figure 1. Structure of FAD.
was used with a Ag|AgCl|KClsat electrode as the reference electrode and a platinum wire as the auxiliary electrode. Glassy carbon (GC) and pyrolytic graphite (PG) electrodes were used as working electrodes. GC disk electrodes (diameter 3 mm, Metrohm) were polished successively with 1- and 0.3-µm alumina slurries (Buehler). Electrodes were then sonicated in water for 5 min. PG electrodes were homemade using a 3-mm-diameter pyrolytic graphite rod (Goodfellow), and the renewal of the electrode surface was achieved by polishing on sandpaper and washing with purified water in an ultrasonic bath. After polishing, working electrodes were cycled, between -0.2 and +1.4 V in phosphate 0.1 M, pH 9, at 50 mV s-1 for graphite electrodes, and between -0.8 and +0.8 V in 0.02 M H2SO4 at 50 mV s-1 for glassy carbon electrodes, until a stable voltammogram was obtained. RESULTS AND DISCUSSION In neutral and alkaline media, FAD undergoes an irreversible oxidation process at graphite electrodes with a peak potential of +1.2 V at pH 9. This process is very close to that obtained for other adenine compounds8 and it is not observed for the oxidation of FMN under the same experimental conditions. It is attributed to the oxidation of the adenine moiety in the prosthetic group. After oxidation of FAD until +1.4 V at pH 9, a new reversible process, associated with an oxidation product, is observed at potentials close to 0 V. This product is strongly adsorbed on the electrode surface as was demonstrated by transfer experiments. As it is shown in Figure 2A (dashed line), this reversible process remains when the corresponding modified electrode is transferred into a fresh solution without FAD. The formal potential for this process is 0.009 V at pH 9, which is very close to those obtained for other adenine derivatives (for example, 0.003 V for ADP, Figure 2B, dashed line), which suggests that a common electroactive species is formed during the electrooxidation of these molecules. The electrochemical oxidation of NADH at electrodes modified with oxidation products of FAD is shown in Figure 2A (solid line). In the presence of NADH, a distinct increase in the oxidation
Figure 2. (A) Cyclic voltammograms traced with PG electrodes modified with oxidation products of FAD. Dashed line curve corresponds to the pure background electrolyte. Solid line curve corresponds to a 0.1 mM solution of NADH. Scan rate: 50 mV s-1; Background electrolyte: 0.1 M phosphate, pH 9.0. Modification was done in a 0.1 mM FAD solution at pH 9.0 by five repeated potential scans between -0.2 and 1.4 V. (B) Cyclic voltammograms traced with PG electrodes modified with oxidation products of ADP. Dashed line curve corresponds to the pure background electrolyte. Solid line curve corresponds to a 0.1 mM solution of NADH. Electrode modification was done in a 0.1 mM ADP solution. Other conditions as in (A). For comparison, dotted line was traced in 0.1 mM NADH solution with a bare PGE preanodized in a pure background electrolyte, using the same potential program.
current at potentials close to 0.009 V is observed, which is in good agreement with an electrocatalytic effect. The uncatalyzed oxidation of NADH also appears at 0.3 V, meaning an efficiency of the catalysis lower than that obtained with electrodes modified with the oxidation products of other adenine nucleotides such as ADP (see Figure 2B for comparison). It is well known that anodization of high-density carbons, such as pyrolytic graphite or glassy carbon, produces quinoidal-like surface functionalities that act as electron-transfer mediators for the oxidation of NADH.11-14 To exclude any significant contribution of surface quinone groups to the observed catalytic effect, we have evaluated the voltammetric response of NADH at a PG electrode pretreated with exactly the same procedure, i.e., oxidation until +1.4 V in phosphate solutions, pH 9, but in the absence of FAD (Figure 2B, dotted line). At this electrode, only the direct noncatalyzed oxidation of NADH is observed. With PG electrodes (11) Blaedel, W. J.; Jenkins, R. A. Anal. Chem. 1975, 47, 1337-1343. (12) C ˇ e¨nas, N.; Rozgaite¨, J.; Pocius, A.; Kulys, J. J. Electroanal. Chem. 1983, 154, 121-128. (13) Engstrom, R.; Strasser, V. A. Anal. Chem. 1984, 56, 136-141. (14) Wang, J.; Tuzhi, P. Anal. Chem. 1986, 58, 1787-1790.
Analytical Chemistry, Vol. 77, No. 13, July 1, 2005
4287
anodized in the presence of FMN, the oxidation of NADH also results in a broad peak, with a peak potential above 0.3 V (data not shown), similar to that obtained at unmodified electrodes. Thus, it is clear that the catalytic activity described above is only due to the oxidation product of the adenine moiety of FAD. These results confirm a general property previously reported for different N9-substituted adenine compounds.8-10,15-18 According to this model, one of the routes of oxidation of FAD leads to the formation of a surface reversible redox couple ascribed to the 2,8dihydroxyadenine derivative, which is the observed NADH electrocatalyst. In the paper entitled “Electropolymerized Flavin Adenine Dinucleotide as an Advanced NADH Transducer”,6 the authors described a different NADH electrocatalyst derived from FAD via oxidation of the riboflavin moiety at high applied potentials on GC electrodes in acidic medium. With the aim to exclude the possible contribution of the oxidation of the adenine moiety to the described electrocatalyst, we tried to reproduce their results, preparing modified GC electrodes following the same procedure they described. After oxidation of FAD at GC electrodes in acidic medium until 2 V and transferring the modified electrode to a phosphate solution pH 6, the voltammogram shown in Figure 3A is obtained. Two different reversible processes are recorded. The first one, with a formal potential of -0.38 V at pH 6, corresponds to the riboflavin moiety of monomeric FAD. The second one has a formal potential of 0.06 V at pH 6 and has been previously attributed to poly(FAD).6 Both processes are also observed after oxidation of FMN under the same experimental conditions. The compound responsible for the last process catalyses the oxidation of NADH. Taking into account that anodic treatments of carbon electrodes in highly acidic media have been reported to produce the highest density of oxygen funcionalities,19 which could mediate electron transfer from NADH,11-14 it seems reasonable to exclude the contribution of these surface groups to the observed catalytic effect. Control experiments were conducted to assess whether the oxidation conditions (and not the presence of FAD) are responsible for the catalytic behavior. For this purpose, GC electrodes were cycled between -0.5 and +2.0 V in 0.1 M HCl/ KCl solutions in the absence of flavine nucleotides. After removing the electrodes from the acidic solution, these were placed in a 0.05 M phosphate, 0.1 M KCl, pH 6.0, solution and the voltammogram shown in Figure 3B (dashed line) was obtained. A broad surface-bound reversible redox process with a formal potential of 0.06 V appears, which is identical to that attributed to poly(FAD). This redox system can only be attributed to surface quinone groups generated during the strong anodic treatment of the electrode.11-14 The species responsible for this process also acts as NADH catalyst as shown in Figure 3B. For both anodized FADmodified and preanodized GC electrodes, the catalytic current (15) de los Santos AÄ lvarez, P.; Molina, P. G.; Lobo Castan ˜o´n, M. J.; Miranda Ordieres, A. J.; Tun ˜o´n Blanco, P. Electroanalysis 2002, 14, 1543-1549. (16) 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. (17) 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. (18) de los Santos AÄ lvarez, N.; Lobo-Castan ˜o´n, M. J.; Miranda-Ordieres, A. J.; Tun ˜o´n Blanco, P. Electroanalysis 2004, 16, 881-887. (19) Kozlowski, C.; Sherwood, P. M. A. J. Chem. Soc., Faraday Trans. 1 1985, 81, 2745-2756.
4288 Analytical Chemistry, Vol. 77, No. 13, July 1, 2005
Figure 3. Cyclic voltammograms obtained in 0.05 M phosphate, 0.1 M KCl, pH 6.0, with modified GC electrodes. Solid lines: with NADH (curve 1) 0.05 and (curve 2) 0.15 mM. Dashed line: background electrolyte. Scan rate: 20 mV s-1. Modification of the electrodes: 26 potential scans between -0.5 and +2.0 V at 50 mV s-1 in a 0.1 M KCl, 0.1 M HCl solution (A) with 1 mM FAD and (B) without FAD.
increases linearly with NADH concentration. For FAD-modified electrodes (Figure 3A) the least-squares line was Icat/A ) (1.5 ( 0.7) × 10-7 + (0.0036 ( 0.0002) [NADH]/M (r ) 0.999; n ) 4), and the corresponding for the preanodized GCE (Figure 3B) was Icat/A ) (1.9 ( 0.8) × 10-7 + (0.0039 ( 0.0002) [NADH]/ M (r ) 0.999; n ) 4). Note that the slopes of the response of both preanodized and FAD-modified GCE are the same within the experimental error. In addition, when GC electrodes were pretreated by anodization in a 0.1 M HCl/KCl solution containing no FAD, then immersed for 2 min in a stirred 0.1 M HCl/KCl solution of 0.1 mM FAD at open circuit, and, finally, placed in a 0.05 M phosphate, 0.1 M KCl, pH 6.0, solution, we obtained the same CV response (Figure 4A) as that previously described in Figure 3A. This means that the anodization of GC did yield modified electrodes that were subsequently capable of incorporating FAD and thus showing that oxidation of FAD until 2 V is not needed for forming the redox process that appears at ∼0 V. Even more, we have observed that when FAD is incorporated by adsorption at open circuit to a previously anodized GC electrode surface, the resulting adsorbate layer is by far more stable (longer desorption time under ultrasonic radiation) than FAD incorporated during the anodization step (Figure 4). This is not the usual behavior for an electropolymerized layer, which is expected to be more stable than a monomeric adsorbate. The stronger
Figure 4. Cyclic voltammograms of FAD-modified GC electrodes in 0.05 M phosphate, 0.1 M KCl, pH 6.0, after sonication in water for (solid line) 2 or (dashed line) 60 min. Modification step: (A) 26 cycles between -0.5 and +2.0 V in a pure 0.1 M HCl/KCl background electrolyte followed by 2 min of FAD preconcentration at open circuit in a stirred 1 mM FAD solution; (B) 26 cycles between -0.5 and +2.0 V in a 1 mM FAD solution in 0.1 M HCl/KCl background electrolyte.
adsorption of FAD on preanodized GC surfaces was already reported by Bowers and Yenser.20 These results clearly indicate that the new set of peaks that appears at potentials near 0 V in cyclic voltammograms during oxidation of FAD in acidic medium, and that authors attributed to polymer-type redox activity,6 arises exclusively from these surface quinone groups, which are the true catalysts for the oxidation of NADH. The catalytic activity of these surface quinone groups toward the oxidation of NADH has already been characterized by Cˇ e¨nas et al.12 with kinetic and electrochemical characteristics very similar to those described in the works of Karyakin and co-workers.6,7 The reversible redox process corresponding to the adsorbed 2,8-dihydroxyadenine derivative, which at pH 6 should be at potentials of ∼0.2 V, does not appear. This fact is consistent with our previous observations of highest efficiency in the generation of catalyst from N9-substituted adenine derivatives when the oxidation step is performed in alkaline solutions.9,18 Therefore, (20) Bowers, M.; Yenser, A. Y. Anal. Chim. Acta 1991, 243, 43-53.
under the conditions used by Karyakin et al.6 to prepare the modified GC electrode, no contribution of the 2,8-dihydroxyadenine FAD derivative to the catalytic activity for the oxidation of NADH could be observed. The amperometric response of PG electrodes, modified with the mediator arising from the oxidized adenine moiety of FAD, to increasing concentrations of NADH at 0.1 V in stirred alkaline solution (pH 9) was investigated. The modified electrode responds linearly to NADH concentrations between 2.5 × 10-8 and 1 × 10-5 M with a sensitivity of 0.05 A M-1 cm-2. The relative standard deviation for six successive measurements of 7.4 × 10-7 M NADH was 4.8%. At pH 7.4 and +0.125 V, a similar linear range was obtained with a sensitivity of 0.04 A M-1 cm-2. The limit of detection, estimated from the value of three times the standard deviation of the background noise, is in both cases 4 × 10-8 M. Comparing with other catalysts arisen from adenine derivatives, FAD gives rise to the lowest sensitivity (for example, for ADPmodified electrodes at pH 10, a sensitivity of 0.35 A M-1 cm-2 and a limit of detection of 2.5 × 10-9 M were obtained9). This fact is probably related to the larger molecular size of the N9 substituent of adenine in FAD. As most dehydrogenases shows their maximum activity in alkaline media where most NADH catalysts decrease their activity, the use of the catalyst derived from adenine derivatives could be advantageous. At concentrations up to 1 × 10-5 M, the response levels off following MichaelisMenten kinetics. The operational stability of the electrodes was tested through successive amperometric measurements of 7.5 × 10-7 M NADH at 0.1 V. After 2 h and more than 10 successive measurements, a decrease of 14% in the response was obtained. After 4 h of continuous operation, the response remained at the 80% level, indicating an acceptable operational stability for the modified electrodes, not very different from that previously reported for other FAD-modified electrodes.6 CONCLUSION Only two different catalysts for the oxidation of NADH can be obtained from FAD: one related to the riboflavin moiety of the molecule5 and the other one, described in this study, ascribed to the product generated after oxidizing the adenine moiety of FAD in alkaline solution. It is verified that this catalyst is a species different from the one described previously arisen from the oxidation of FAD in acidic medium.6 We demonstrate that the catalytic activity of the latter is in fact due to surface quinone groups generated during the oxidation process until 2-2.5 V. ACKNOWLEDGMENT This research was supported by FICYT project PB-EXP01-28. Received for review September 30, 2004. Accepted April 27, 2005. AC048545P
Analytical Chemistry, Vol. 77, No. 13, July 1, 2005
4289
