Direct electron transfer from alcohol dehydrogenase

Abstract

This study employed hydrodynamic cyclic voltammetry (HCV) with a glassy carbon (GC) rotating disk electrode (RDE), cyclic voltammetry (CV), and electrochemical impedance spectroscopy (EIS) to investigate the direct electron transfer (DET) behavior of alcohol dehydrogenase (ADH, EC 1.1.1.1) embedded or otherwise in polymyxin (PM) and adsorbed concomitantly with cofactor NAD⁺ or NADH, or none on GC and covered with Nafion®. The hybrid GC/PM-ADH-NAD⁺/Nafion electrode thus constructed (and visualized by scanning electron microscopy (SEM)) persistently bioelectrocatalyzed ethanol oxidation and performed DET involving baker's yeast ADH, GC, and ethanol—a key finding in the present study—making this system a promising anode for use in biofuel cells. A rate constant (k_s) of 0.82 s⁻¹ was obtained for this electrode at a potential scan rate (ν) of 120 mV s⁻¹. EIS experiments, particularly those conducted after higher potential HCV scans, allowed resistance to electron hopping (R_eh) between redox centers in ADH and between these centers and ethanol molecules to be estimated as 84 kΩ lower for the GC/PM-ADH-NAD⁺/Nafion electrode during ethanol oxidation than for bare GC in the presence of ethanol. Nuclear magnetic resonance (NMR) unequivocally confirmed acetaldehyde production from ethanol oxidation by ADH DET.

---

1 Introduction

About 4800 enzymes have been classified to date,^1,2 roughly 1400 (ref. 1) of which are redox enzymes (also known as oxidoreductases), whose activity depends on non-proteinaceous redox cofactors.³ Since redox enzymes catalyze oxidation/reduction reactions, their coupling with electrodes may seem obvious, yet in most cases the electronic coupling is not straightforward, by reason of kinetic restrictions and/or unfavorably long distances between the electrode surface and the redox active site in cofactor-bound enzymes.³ About 17% of all classified enzymes are NAD(P)-dependent, or NAD(P)-linked, dehydrogenases,^3,4 whose activity depends on pyridine nucleotides,^3,4 which occur in two biologically active forms—viz. β-nicotinamide adenine dinucleotide (NAD) and β-nicotinamide adenine dinucleotide phosphate (β-NADP). Ubiquitous in all living systems, these non-proteinaceous coenzymes largely act on oxidoreductase-catalyzed reactions as acceptors or donors of the equivalent to a hydride ion (H⁻) from a substrate in a reversible manner, thus playing a key role in biological electron transfer reactions and pathways.^3,4 In contrast to other oxidoreductases, NAD(P)-dependent dehydrogenases are dependent on a soluble cofactor.^3,4 Other terms, they operate in nature by a diffusional route in which the oxidized cofactor penetrates into the protein, accepts the electron (as a hydride equivalent), and diffuses from the protein matrices,⁵ barring a few exceptions.^3,6–8

An extensive research literature is devoted to biosensors based on NAD(P)-dependent dehydrogenases (see ref. 3, 4, and 9, as well as the sources given in ref. 4 and 9). Examples include the oxidized form NAD⁺, which, immobilized on a carbon paste (CP) electrode, is converted to NADH by oxidation of ethanol catalyzed by alcohol dehydrogenase (ADH), with both ethanol and ADH in solution;¹⁰ an ethanol sensor based on CP chemically modified with ADH, NAD⁺, and 3-β-naphthoyl – brilliant cresyl blue (as a mediator);¹¹ and sensors constructed by co-immobilizing yeast ADH (EC 1.1.1.1) and NAD⁺ on CP by adding polyethylenimine (a polymer-bound mediator) to the reaction mixture;¹² based on yeast ADH, NAD⁺, and methylene blue (as a mediator) incorporated into a CP matrix and covered with poly(ester sulfonic acid);¹³ based on poly(o-phenylenediamine) electropolymerized on the surface of ADH from a baker's-yeast-NAD⁺-modified CP electrode;¹⁴ constructed by incorporating ADH and electroactive materials (tetracyanoquinodimethane, tetrathiafulvalene, and dimethyl ferrocene) within graphite paste electrodes;¹⁵ by co-immobilizing ADH, NAD⁺, and the phenoxazine compound Meldola blue (as a mediator) on carbon nanotube paste;¹⁶ by sequential immobilization of baker's yeast ADH,¹⁷ aldehyde dehydrogenase, and poly(diallyldimethylammonium chloride)-covered multiwalled carbon nanotubes (MWCNTs) onto a glassy carbon (GC) electrode substrate;¹⁸ in ionic liquid (IL)–graphene/chitosan/ADH (from a Saccharomyces cerevisiae (Sce))-modified GC electrode;¹⁹ on a graphene–Au nanorods–ADH (from Sce)-modified GC electrode;²⁰ on an ADH (from Sce)–NAD/TBO (toluidine blue O)/Nafion-modified GC electrode;²¹ on a Nafion/ADH (from Sce)/graphene/GC electrode;²² and on a MWCNTs/IL/electrogenerated NAD⁺ oxidation products (Ox-P(NAD⁺))-modified GC electrode.²³

Another important aspect addressed in the literature is related to the oxidation/reduction of NAD(P)H/NAD(P)⁺. For instance, Elving and co-workers, investigating the electrochemical oxidation of NADH at pH 7.0 using different (e.g., GC) conditioned, pretreated electrodes under diverse experimental conditions, observed pre-wave or pre-peak effects resulting from adsorption onto the electrode of NAD⁺ produced during the oxidation process.²⁴ They also investigated the adsorption of NAD⁺, NADH, and the NAD–NAD dimer (produced by NAD⁺ reduction on a GC electrode) onto, e.g., a pretreated GC electrode,²⁵ in addition to proposing a model redox mechanism for the NAD⁺/NADH couple, also involving mediation by surface species and modification by the environment.^26,27 Gorton and Bartlett's³ and Gorton and Domínguez's⁴ studies also reviewed direct electrochemical oxidation of NAD(P)H, in addition to mediators for electrocatalytic NAD(P)H oxidation. Wu et al.²⁸ surveyed the principal methods for regeneration of NAD(P)H/NAD(P)⁺ coenzymes, including enzymatic, chemical, electrochemical, and photochemical approaches. Kochius et al.²⁹ critically reviewed immobilized redox mediators for electrochemical NAD(P)⁺ regeneration. Radoi and Compagnone³⁰ reviewed strategies to produce surface-modified electrodes for NADH sensing, surface redox-mediated NADH probes, and bulk-modified electrodes for NADH electrocatalytic oxidation. Pumera et al.³¹ suggested that NAD⁺ adsorption onto sp² carbon materials is made possible by oxygen-containing groups (mainly carboxylic groups) formed at the edges and edge-like defects of graphene sheets. Li et al.³² reported high oxidative conversion of NADH to bioactive NAD⁺ by electropolymerizing methylene green on carboxylated-carbon-nanotube-modified carbon paper, which acted as an electrode of large surface area. Punckt et al.^33,34 showed that NADH oxidation on graphene-based electrodes, resulting in electrocatalytic behavior, involved not only differences in chemical structure, but also differences in electrode morphology (e.g., graphene electrode porosity, as effected by packing morphology; functional group; lattice defect concentration).

The literature reports direct electron transfer (DET) involving pyrroloquinoline quinone (PQQ)-dependent dehydrogenases (PQQ-DHs)—e.g., Ramanavicius et al.³⁵ and Razumiene et al.³⁶ reported DET processes between a polypyrrole-entrapped quinohemoprotein–ADH (QH–ADH) film from Gluconobacter sp. 33 (EC 1.1.99.8) and a platinum electrode, a phenomenon taking place via the conducting polymer network,³⁵ with QH–ADH adsorbed onto carbon electrodes allowing direct electrocatalytic oxidation of ethanol³⁶—and the use of these enzymes in biofuel cells.^5,37–43 There are also reports of NAD⁺-dependent dehydrogenases applied in biofuel cells.^44,45

“Promoter” molecules (also known as “facilitators”) are used to ease protein immobilization onto a carbon surface.⁴⁶ Promoters are usually small molecules, including oligosaccharides, multivalent metal ions, or organic multivalent ions such as aminoglycosides (e.g., neomycin) and peptides (e.g., polymyxin, polylysin) that are not redox-active but modify the surface characteristics of carbon so that a stable protein film is formed on the electrode.⁴⁶ Ideally, a promoter should stabilize the protein onto the electrode, prevent its denaturation, and promote electron transfer by optimizing protein orientation.⁴⁶ The orientation of negatively charged proteins for direct electrochemistry at negatively charged electrodes (e.g., GC) can be optimized by adding multivalent positive ions.⁴⁶ The cations bridge the negatively charged surfaces of both electrode and protein,⁴⁶ making the interactions between protein and electrode surface very similar to those occurring at biological membranes.⁴⁷ The negatively charged electrode surface is optimal for direct electrochemistry of positively charged proteins.⁴⁶ Recently, we successfully employed a simple approach based on horseradish peroxidase (HRP) adsorption in the presence of polymyxin (PM) to form Nafion-covered HRP-PM films on a GC surface, resulting in effective DET between heme Fe³⁺/Fe²⁺ redox centers in HRP enzymes and GC and between these centers and H₂O₂ and O₂ molecules.⁴⁸ Also, we showed that a GC/PM-GOx/Nafion electrode—constituted of glucose oxidase (GOx) embedded in PM adsorbed onto GC and covered with Nafion—was capable of persistently bioelectrocatalyzing O₂ reduction and performing DET between native GOx, GC, and O₂. DET, however, was not detected between native GOx and glucose or H₂O₂.⁴⁷

The primary purpose of the present study was to employ hydrodynamic cyclic voltammetry (HCV), cyclic voltammetry (CV), and electrochemical impedance spectroscopy (EIS) to investigate DET and provide a kinetic analysis of the Nafion-covered ADH system in the presence (or absence) of PM containing (or not) NAD⁺ or NADH upon adsorption onto a GC surface. Scanning electron microscopy (SEM) was also employed for physical characterization of the GC/PM-ADH-NAD⁺ (or NADH)/Nafion surface. The electrocatalytic behavior of these electrodes in promoting ethanol oxidation and DET between ADH and ethanol was also investigated so as to be corroborated by acetaldehyde production—identified by nuclear magnetic resonance (NMR)—from the oxidation of ethanol resulting from ADH DET. Experimental DET between ADH and GC had been clearly observed using these modified electrodes, running counter to the established theory that views NAD(P)-dependent dehydrogenases as dependent on diffusional cofactors—i.e., that oxidized cofactors mediate electron transfer (as hydride equivalents).

2 Experimental section

2.1 Instruments and reagents

HCV, CV, and EIS measurements were performed in a three-electrode glass cell with a working electrode consisting of a GC disk (0.20 cm² geometric area) embedded in Teflon (Pine Research Instrumentation). A Pt plate (Degussa) was employed as the counter-electrode, and a saturated calomel electrode (SCE) served as the reference electrode. All experiments were carried out at room temperature (around 25 °C). Before use, the GC electrode surface was polished to a mirror finish by abrasion with emery paper, sequential polishing in 1.0 and 0.05 μm alumina slurries, and subjection to a final cleaning step by sonication twice (15 min each time) in Milli-Q water (Millipore). No additional preconditioning or pretreatment of the GC electrode was used to increase oxygen functional groups associated with protein faradaic activity.⁴⁹

Alcohol dehydrogenase from baker's yeast (EC 1.1.1.1) (Sigma-Aldrich); β-nicotinamide adenine dinucleotide sodium salt from yeast (NAD⁺) (Sigma-Aldrich); β-nicotinamide adenine dinucleotide, reduced disodium salt hydrate (NADH) (Sigma-Aldrich); polymyxin B sulfate salt (Fluka BioChemika); 5 wt% Nafion perfluorinated resin solution in lower aliphatic alcohols and water (Aldrich); ethanol (Merck); acetaldehyde (Fluka); and K₃Fe(CN)₆ (Dinâmica) were all used as received. A 0.1 M KH₂PO₄ (Vetec) solution was adjusted to pH 6.5 with 2 M KOH (Vetec). The solutions were prepared with Milli-Q water and purged for 30 min with pure nitrogen (Air Liquid) prior to each experiment performed in the presence 100 mM of ethanol or 1 mM K₃Fe(CN)₆, or none.

2.2 Electrode preparation

The electrode named GC/PM-ADH-NAD⁺ (or NADH)/Nafion was prepared using a freshly polished GC surface to which 22 μL of ADH solution in water (prepared with 5.5 μL of 20 mg mL⁻¹ ADH, 5.5 μL of 2 mg mL⁻¹ PM, and 11 μL of either NAD⁺ or NADH at 10 mM) was applied and allowed to dry at room temperature. The electrode was then covered with 11 μL of Nafion solution at 5 wt% in lower aliphatic alcohols and water, and again dried at room temperature.^47,48 Additional electrode variations were prepared according to the same principle—e.g., a GC/ADH-NAD⁺/Nafion electrode required 22 μL of ADH solution prepared with 5.5 μL of 20 mg mL⁻¹ ADH, 5.5 μL of water, and 11 μL of 10 mM NAD⁺, followed by the above-described steps.

The electrode denominated GC/VC XC-72 was prepared by dripping an aqueous solution of 97.3 μg mL⁻¹ Vulcan XC-72 carbon on a GC surface to produce a thin Vulcan XC-72 carbon film with a loading of 14.6 μg cm⁻², followed by drying at room temperature,⁵⁰ without additional preconditioning or pretreatment of the GC/VC XC-72 electrode, and the above-described steps to produce, e.g., the GC/VC XC-72/ADH/Nafion modified electrode.

2.3 Apparatus and measurements

A PGSTAT128N potentiostat/galvanostat (Autolab) equipped with a FRA2.X module was used in EIS and other electrochemical experiments. For HCV, the potentiostat/galvanostat was employed in conjunction with an AFMSRX rotation rate module (Pine Instruments). The EIS experiments were conducted at fixed potentials—namely, at an open-circuit potential (OCP) of around 0.18 V (on average) in the presence of Fe(CN)₆³⁻ and 0.65 V in the presence of ethanol, with potential perturbation of 25 mV (rms) within a frequency range of 10 mHz to 100 kHz. Care was taken to ensure that AC impedance data corresponded to the interfaces being investigated at high frequencies—namely, the potentiostat employed had faster performance; the highest cutoff frequency was limited to 160 kHz for the FRA2.X module, which used a fixed filter when the frequency applied exceeded 19 kHz; cables were short and, akin to the connections, were shielded; the electrochemical cell was placed inside a Faraday cage; the working electrode was positioned in front of and close to a larger Pt plate; and the SCE was placed near the working electrode. Nova 1.10 Autolab (2013) software was used to simulate the behavior of equivalent circuits of the interface in the presence of different redox probes, and the parameters of these circuits were fitted to the measured spectra using a nonlinear least-squares program.

The microstructure of the GC/PM-ADH-NAD⁺ (or NADH)/Nafion electrode was coated with a deposited ultrathin gold coating to be visually characterized by SEM, performed on a JSM-6380LV field-emission scanning electron microscope (Jeol). An accelerating voltage of 15 kV was applied.

The ¹H NMR experiments were recorded on a 7.05-T Avance DPX-300 spectrometer (Bruker).

3 Results and discussion

3.1 HCV responses from GC and modified GC electrodes in an N₂-saturated solution

Fig. 1 shows HCV responses describing the redox behavior of (A) GC/ADH/Nafion, (B) GC/PM-ADH/Nafion, and (C) GC/NAD⁺/Nafion and GC/PM-ADH-NAD⁺/Nafion electrodes for different numbers of HCV scans. Use of the GC/ADH/Nafion electrode revealed an increase in anodic currents after 0.35 V (positive-going scan) during the first scan (Fig. 1A, red curve), with an anodic current peak positioned at around 0.62 V (current intensity close to 2.4 μA, without current background subtraction), in comparison with bare GC, which only became sufficiently decreased during the second HCV scan (Fig. 1A, green curve), with no shoulders. No redox currents or peaks were detected during the 10th and subsequent HCV scans (Fig. 1A, magenta curve). Use of the GC/PM-ADH/Nafion electrode revealed an increase in anodic currents after 0.37 V (positive-going scan) during the first scan (Fig. 1B, red curve), with an anodic current shoulder positioned at around 0.62 V (current intensity close to 1.7 μA, without current background subtraction), in comparison with bare GC, which only became sufficiently decreased during the second HCV scan (Fig. 1B, green curve), with no shoulders. No redox currents or peaks were detected during the 20th and subsequent HCV scans (Fig. 1B, dark yellow curve). Using the GC/NAD⁺/Nafion electrode, no redox currents were detected during the first scan (red curve; see also inset to Fig. 1C) until near −0.87 V (negative-going scan), but a large cathodic current (reduction current) was observed, in comparison with bare GC. The second scan (Fig. 1C, green curve) revealed increased anodic currents after 0 V (positive-going scan), with an anodic current peak positioned at around 0.33 V, slightly decreasing in intensity and displacing towards a more positive potential (0.50 V) during the 30th HCV scan (Fig. 1C, cyan curve). A similar slight decrease in current intensity was observed at potentials more negative than −0.87 V (negative-going scan) as the number of HC scans was increased. Differently, use of the GC/PM-ADH-NAD⁺/Nafion electrode revealed an increase in anodic current after 0.38 V (positive-going scan) during the first scan (Fig. 1C, magenta curve), with an anodic current peak positioned at around 0.69 V (current intensity close to 7.8 μA, without current background subtraction) and an increase in cathodic current after −0.87 V (negative-going scan), yet lower than that observed during the 30th HCV scan (Fig. 1C, cyan curve) using the GC/NAD⁺/Nafion electrode. During the second HCV scan (Fig. 1C, dark yellow curve), a current peak was detected at 0.38 V, along with a shoulder at 0.75 V, both with currents of lower intensity than the current peak observed during the first HCV scan (Fig. 1C, cyan curve). In the third HCV scan (Fig. 1C, navy curve) the current peak was displaced towards a more positive potential (0.44 V), with a shoulder at around 0.75 V, both with slightly less intense currents, compared with the second HCV scan (Fig. 1C, dark yellow curve). During the second and third HCV scans, the currents after −0.87 V (negative-going scan) were also decreased. Finally, during the 30th HCV scan (Fig. 1C, purple curve), no well-defined redox currents or peaks were observed.

Fig. 1 Hydrodynamic cyclic voltammograms for (A) bare GC (a′, first; a, second scan) and GC/ADH/Nafion (b, first; c, second; d, third; e, fifth; f, 10th; g, 20th; h, 30th scan), (B) bare GC (a, second scan) and GC/PM-ADH/Nafion (b, first; c, second; d, third; e, fifth; f, 10th; g, 20th; h, 30th scan), and (C) bare GC (a, second scan), GC/NAD+/Nafion (b, first; c, second; d, 10th; e, 30th scan), and GC/PM-ADH-NAD⁺/Nafion electrodes (f, first; g, second; h, third; i, 30th scan) in N₂-saturated 0.1 M KH₂PO₄ (pH 6.5) (inset: the same curves, restricted to −0.4–0.8 V). First scans started at −0.4 V in the positive-going direction. ω = 200 rpm, ν = 10 mV s⁻¹.

Both GC/NAD⁺/Nafion and GC/NADH/Nafion electrodes (Fig. S1 (ESI†), red, green, blue, cyan, magenta, and dark yellow curves), as well as GC/PM-NAD⁺/Nafion and GC/PM-NADH/Nafion electrodes (data not shown), failed to yield well-defined redox currents or peaks. Other electrode combinations, such as GC/PM, GC/Nafion, and GC/PM/Nafion, failed to detect any redox currents or peaks under the conditions reported in Fig. S1, ESI.†^47,48

Fig. S2 (ESI†) shows the redox behavior of GC/ADH-NAD⁺/Nafion, GC/ADH-NADH/Nafion, and GC/PM-ADH-NAD⁺/Nafion electrodes for different numbers of HCV scans. All three modified electrodes exhibited a well-defined anodic peak (Fig. S2 (ESI†), red, cyan, and navy curves), with anodic currents increasing after 0.25 V. The GC/ADH-NADH/Nafion electrode provided the highest current peak (close to 32 μA, without current background subtraction). A current peak potential of 0.65 V on average was observed for the three modified electrodes, whose anodic current peaks drastically decreased (Fig. S2 (ESI†), green, magenta, and purple curves) only during the second HCV scan, without exhibiting any redox currents or peaks (Fig. S2 (ESI†), blue, dark yellow, and wine curves) during the 30th HCV scan.

For HCV scans in a narrower potential scan window (Fig. S3 (ESI†)), the GC/ADH-NAD⁺/Nafion electrode showed the same well-defined anodic peak (Fig. S3 (ESI†), red curve) previously described, with anodic currents increasing after 0.30 V, a higher current peak value (close to 42 μA, without current background subtraction), and a 0.65 V current peak potential, which drastically decreased (Fig. S3 (ESI†), green curve) only during the second HCV, exhibiting no redox currents or peaks (Fig. S3 (ESI†), cyan curve) during the fifth and subsequent HCV scans.

In addition, a well-defined shoulder at 0.78 V was observed for the GC/VC XC-72/PM-ADH/Nafion electrode (Fig. S4 (ESI†), red curve), with current intensity close to 28 μA (without current background subtraction) and increased anodic currents after 0.30 V, which only became sufficiently decreased during the second HCV scan (Fig. S4 (ESI†), green curve), with no shoulders. No redox currents or peaks were detected during the 20th and subsequent HCV scans (Fig. S4 (ESI†), dark yellow curve).

The results above clearly suggest the occurrence of DET between ADH and GC while using these modified electrodes, allowing us to assume the ADH oxidation reaction (DET from ADH to the GC surface) to be:

ADH − 2e⁻ − 2H⁺ → ADH_mod (1)

This assumption was also based on computational studies of the mechanism of proton and hydride transfer in liver alcohol dehydrogenase, described by Hammes-Schiffer and co-workers,^51,52 in which case we are assuming that DET takes place in the region comprising protein residues Cys-46, His-67, and Cys-174 and the catalytic zinc ion^51,52 (see also ref. 53 and 54), and hydrogen proton release and modification in ADH probably occur at protein residue His-67 (Scheme 1). To our knowledge, DET occurrence on electrode surfaces has been reported for PQQ–DHs,^5,35–41 but not for ADH (EC 1.1.1.1)—the most important achievement of the present investigation. ADH surface concentration (Γ_ADH) values are determinable by integrating the graph area under the HCV oxidation peaks (or shoulder) (Γ = Q/nFA,⁵⁵ where Q is the charge consumed, n is the number of electrons involved (n = 2, assuming the occurrence of reaction 1), F is the Faraday constant, and A is the geometric surface area) (Fig. 1), yielding values of 1.4, 1.2, and 3.7 nmol cm⁻² for GC/ADH/Nafion, GC/PM-ADH/Nafion, and GC/PM-ADH-NAD⁺/Nafion electrodes, respectively.

Scheme 1 The active site of ADH^51–54 during DET to GC.

In addition, NADH is neither oxidized nor reduced at the GC/NADH/Nafion and GC/PM-NADH/Nafion electrodes in the potential range of −1.0 to 0.8 V, most probably because no additional preconditioning or pretreatment was performed on GC electrode employed to increase the number of oxygen functional groups^24–27,49 or because the immobilization was not of the same type used for the CP electrode,¹⁰ or even because no NADH was employed in the solution.^{3,4,24–27,33,34} Our results differ from those obtained by Yao and Musha,¹⁰ who found a well-defined anodic peak at around 0.4 V vs. SCE, attributed to oxidation of NADH immobilized on the CP electrode surface. Elving and co-workers^24–27 showed that, using conditioned and pretreated GC electrodes, NADH in solution exhibited a wave during HCV, with increased anodic currents after 0.1 V vs. SCE and wave current limit starting in the range of 0.5 to 0.7 V vs. SCE. (The investigators also combined NAD⁺ in solution to minimize a pre-wave pattern caused by NAD⁺ adsorption after NADH oxidation.) Elving and co-workers proposed the following NADH oxidation mechanism:^24–27

NADH − e⁻ → NADH˙⁺ − H⁺ → NAD˙ − e⁻ ⇌ NAD⁺ (2)

Gorton and Bartlett³ and Gorton and Domínguez⁴ reviewed and discussed this and other mechanisms, and recorded E⁰′ for the NADH˙/NADH redox couple as 0.78 V vs. SCE.

For the GC/NAD⁺/Nafion and GC/PM-ADH-NAD⁺/Nafion electrodes, only at potentials more negative than −0.87 V (negative-going scan) was it possible to detect large reduction currents in comparison with bare GC in the HCV responses (Fig. 1C), as similarly observed by Yao and Musha¹⁰ as E_p/2 = −1.08 V vs. SCE, and by Elving and co-workers^25,26 as cathodic peaks or waves at potentials more negative than −1.0 V, allowing us to assume that the following mechanisms^3,4,25–27 are involved in these reduction currents:

NAD⁺ + e⁻ → NAD˙ + H₃O⁺ + e⁻ → NADH + H₂O (3)

or

2NAD˙ → (NAD)₂ (4)

Gorton and Bartlett³ and Gorton and Domínguez⁴ reported E⁰′ for the NAD⁺/NAD˙ redox couple as −1.16 V.

Elving and co-workers^25,26 pointed out that oxidation of the (NAD)₂ dimer ((NAD)₂ – 2e⁻ → 2NAD⁺) occurs at ca. −0.4 V vs. SCE. Our HCV results (Fig. 1C) for the GC/NAD⁺/Nafion electrode revealed an anodic current peak displacing from 0.33 to 0.50 V from the second to the 30th HCV scan, a displacement also detected during the second and third HCV scans (0.38 to 0.44 V) using the GC/PM-ADH-NAD⁺/Nafion electrode. In our study, this anodic peak can be attributed to the mechanism described by eqn (2), rather than to oxidation of the (NAD)₂ dimer.

However, ADH oxidation occurred more or less catalytically, depending on the starting scan potential and the substances that modify the electrodes. The GC/ADH/Nafion electrode, for instance, showed increased anodic currents after 0.35 V (Fig. 1A, red curve), of around 2.4 μA (without current background subtraction) at 0.62 V (anodic current peak). The GC/PM-ADH/Nafion electrode showed increased anodic currents after 0.37 V (Fig. 1B, red curve), with around 1.7 μA (without current background subtraction) at 0.62 V (anodic current shoulder), yet the GC/VC XC-72/PM-ADH/Nafion electrode exhibited a shoulder at 0.78 V (Fig. S4 (ESI†), red curve), with current intensity close to 28 μA (without current background subtraction)—the difference stemming from the presence (or absence) of PM, the starting scan potential, and the presence of a VC XC-72 film on the GC surface. For the GC/PM-ADH-NAD⁺/Nafion electrode, the current peak at around 0.69 V (Fig. 1C, red curve and, Fig. S2 (ESI†), navy curve) appeared as a shoulder at 0.75 V (Fig. 1C, dark yellow curve) during the second HCV scan, while for the GC/ADH-NAD⁺/Nafion electrode the anodic peak appeared at around 0.65 V, with a slightly higher current peak seen in Fig. S3 (ESI,† red curve) than in Fig. S2 (ESI,† red curve), and an even greater current peak for the GC/ADH-NADH/Nafion electrode (Fig. S2 (ESI†), cyan curve), here involving differences in starting scan potential and compounds (NADH or NAD⁺, and/or PM surrounding ADH). Variability in current peak values can be attributed to rearrangements or conformational changes in ADH structure when in the presence of NADH or NAD⁺, and/or PM surrounding ADH, being also dependent on the starting scan potential.

3.2 HCV responses from GC and modified GC electrodes in a N₂-saturated solution containing 100 mM ethanol

Fig. 2 shows HCV responses describing the redox behavior of (A) GC/ADH/Nafion, (B) GC/PM-ADH/Nafion, and (C) GC/NAD⁺/Nafion and GC/PM-ADH-NAD⁺/Nafion electrodes in the presence of 100 mM ethanol for different numbers of HCV scans. When the GC/ADH/Nafion electrode was employed in the presence of ethanol (Fig. 2A), a visible increase in currents occurred after ca. 0.35 V, with an anodic current shoulder positioned at around 0.62 V (current intensity close to 2.8 μA, without current background subtraction), in comparison with bare GC. This shoulder, however, was higher than the anodic current peak for GC/ADH/Nafion in the absence of ethanol (Fig. 1A). The 30th HCV scan (Fig. 2A, navy curve) revealed no redox currents or peaks for the GC/ADH/Nafion electrode, a behavior also observed during the 10th HCV scan using the GC/ADH/Nafion electrode in the absence of ethanol (Fig. 1A, magenta curve). When the GC/PM-ADH/Nafion electrode was employed in the presence of ethanol (Fig. 2B), a visible increase in currents occurred after ca. 0.35 V, with an anodic current peak positioned at around 0.60 V (current intensity close to 1.6 μA, without current background subtraction), in comparison with bare GC. The current intensity, however, was very close to the anodic current shoulder for GC/PM-ADH/Nafion in the absence of ethanol (Fig. 1B). The 30th HCV scan (Fig. 2B, navy curve) revealed no redox currents or peaks for the GC/PM-ADH/Nafion electrode, a behavior also observed during the 20th HCV scan using the GC/PM-ADH/Nafion electrode in the absence of ethanol (Fig. 1B, dark yellow curve). When the GC/PM-ADH-NAD⁺/Nafion electrode was employed in the presence of ethanol (Fig. 2C), the first scan (red curve) revealed increased anodic currents after 0.47 V (positive-going scan), with an anodic current peak positioned at around 0.65 V, and increased cathodic currents after −0.87 V (negative-going scan)—both types of current, however, being lower than those obtained during the first HCV scan using the GC/PM-ADH-NAD⁺/Nafion electrode in the absence of ethanol (Fig. 1C, magenta curve). During the second HCV scan (green curve), two anodic current peaks were detected, at 0.37 and 0.68 V, both with currents similar to those observed using the GC/PM-ADH-NAD⁺/Nafion electrode in the absence of ethanol (Fig. 1C, dark yellow curve). Finally, during the 30th HCV scan (navy curve), the currents remained increased after −0.1 V (positive-going scan), with a well-defined anodic current peak at 0.74 V and no redox currents after −0.87 V (negative-going scan). Redox currents were also absent from the 30th scan using the GC/PM-ADH-NAD⁺/Nafion electrode in the absence of ethanol (Fig. 1C, purple curve).

Fig. 2 Hydrodynamic cyclic voltammograms for (A) bare GC (a′, first; a, second scan) and GC/ADH/Nafion (b, first; c, second; d, third; e, fifth; f, 10th; g, 20th; h, 30th scan), (B) bare GC (a, second scan) and GC/PM-ADH/Nafion (b, first; c, second; d, third; e, fifth; f, 10th; g, 20th; h, 30th scan), and (C) bare GC (a, second scan) and GC/PM-ADH-NAD⁺/Nafion electrodes (b, first; c, second; d, third; e, fifth; f, 10th; g, 20th; h, 30th scan) in N₂-saturated 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol. First scan started at −0.4 V in the positive-going direction. ω = 200 rpm, ν = 10 mV s⁻¹.

Γ_ADH values determinable by integrating the graph under the HCV oxidation peaks (or shoulder) (Fig. 2) yielded values of 1.9, 0.9, and 0.8 nmol cm⁻² for GC/ADH/Nafion, GC/PM-ADH/Nafion, and GC/PM-ADH-NAD⁺/Nafion electrodes in the presence of ethanol, respectively.

A similar behavior was observed for the GC/NAD⁺/Nafion electrode in the presence of 100 mM ethanol (Fig. S5A, ESI†), but with smaller currents, when compared with the GC/NAD⁺/Nafion electrode in the absence of ethanol (Fig. 1C)—the main difference being a displacement of the anodic current peak to around 0.57 V, and its complete disappearance during the 10th and subsequent HCV scans (Fig. S5A (ESI†), magenta curve). With the GC/NADH/Nafion electrode, currents were visibly increased after 0.37 V (positive-going scan), but sufficiently decreased during the second and third HCV scans (Fig. S5A (ESI†), purple, wine, and olive curves) and entirely absent after the fifth HCV scan. The GC/NAD⁺/Nafion electrode failed to detect well-defined redox currents or peaks in the presence of ethanol (Fig. S5B, ESI†), a behavior also depicted in Fig. S1 (ESI†) for GC/NAD⁺/Nafion in the absence of ethanol, when −0.80 V was set as a lower potential limit.

Fig. S6 (ESI†) depicts the redox behavior of GC/ADH-NAD⁺/Nafion and GC/ADH-NADH/Nafion electrodes for different numbers of HCV scans. Both modified electrodes showed a well-defined anodic peak (Fig. S6 (ESI†), red curves), with anodic currents increasing after 0.32 V, a higher current peak value (close to 33 μA, without current background subtraction) for GC/ADH-NADH/Nafion, and an average current peak potential of 0.66 V for both electrodes, whose anodic current peaks were sufficiently decreased only during the second HCV scan (Fig. S6 (ESI†), green curves), with some cathodic currents (see eqn (3) and (4) for the reduction mechanisms involved) after −0.87 V (negative-going scan) for GC/ADH-NAD⁺/Nafion electrode. Slightly more anodic currents were detected during the 30th HCV scan (Fig. S6 (ESI†), navy curves) for GC/ADH-NAD⁺/Nafion, compared with GC/ADH-NADH/Nafion. A noteworthy feature is that even though GC/ADH-NAD⁺/Nafion exhibited lower anodic current peaks than those detected by GC/ADH-NAD⁺/Nafion in the absence of ethanol during the first HCV scan (compare Fig. S6 and S2 in the ESI,† red curves), the GC/ADH-NADH/Nafion electrode showed around the same anodic current peaks in the presence or absence of ethanol for the first HCV scan (compare Fig. S6 (ESI†), red curve, with Fig. S2 (ESI†), cyan curve). Anodic currents (positive-direction scan) were observed for GC/ADH-NAD⁺/Nafion and GC/ADH-NADH/Nafion electrodes in the presence of ethanol, but not in its absence, during the 30th HCV scan (compare Fig. S6 (ESI†), navy curves with Fig. S2 (ESI†), blue and dark yellow curves).

Fig. S7 (ESI†) depicts the redox behavior of GC/ADH-NAD⁺/Nafion, GC/ADH-NADH/Nafion, GC/PM-ADH-NAD⁺/Nafion, and GC/PM-ADH-NADH/Nafion electrodes in the presence of ethanol for different numbers of HCV scans. All four modified electrodes exhibited a well-defined anodic peak (Fig. S7, ESI†), with anodic currents increasing after ca. 0.27 V. Higher current peak values were observed for GC/ADH-NAD⁺/Nafion and GC/ADH-NADH/Nafion electrodes (Fig. S2, S3, and S6B in the ESI†), with an average current peak potential of 0.65 V for the four modified electrodes. Their anodic current peaks underwent drastic decreases only during the second HCV scan (Fig. S7A, S7C, and S7D in the ESI,† green curves), but the anodic currents (positive-going scan) were conserved during the 30th HCV scan for the GC/ADH-NAD⁺/Nafion electrode, whereas GC/PM-ADH-NAD⁺/Nafion exhibited a well-defined shoulder (Fig. S7A and S7B (ESI†), respectively).

Using the GC/VC XC-72/PM-ADH/Nafion electrode in the presence of ethanol, a well-defined shoulder at 0.80 V (Fig. S8A and S8B (ESI†), cyan and red curves, respectively) was observed, with increased anodic currents after 0.32 V. Only during the second HCV scan (Fig. S8B (ESI†), green curve) were these currents sufficiently decreased and devoid of shoulders. No redox currents or peaks were detected during the 20th and subsequent HCV scans (Fig. S8B (ESI†), dark yellow curve), a behavior comparable to that of GC/VC XC-72/PM-ADH/Nafion in the absence of ethanol (Fig. S4, ESI†). In the presence of ethanol and within a potential limit of 0.8 V (Fig. S8A, ESI†), however, the GC/VC XC-72/PM-ADH/Nafion electrode detected anodic currents (positive-going scan) during the 30th HCV scan. Using the GC/VC XC-72/PM-ADH-NAD⁺/Nafion electrode in the presence of ethanol, a sharply defined current peak was observed during the first HCV scan at around 0.62 V (Fig. S8C (ESI†), red curve) with anodic currents increasing after 0.28 V, but sufficiently decreased only during the second HCV scan (Fig. S8C (ESI†), green curve), with no shoulders. No redox currents or peaks were detected during the 20th and subsequent HCV scans (Fig. S8B (ESI†), dark yellow curve).

The results described above suggest the occurrence of persistent, structurally catalyzed ADH-mediated ethanol oxidation (with ADH oxidation involving DET from ADH to the GC surface—Scheme 1) enhanced by NAD⁺ presence and PM-embedding (Fig. 2C) and set forth by reaction 1, followed by (Scheme 2):

ADH_mod + CH₃CH₂OH → ADH + CH₃CHO (5)

Scheme 2 The modified active site of ADH after DET to GC, combined with ethanol oxidation.

NAD⁺ and PM⁴⁶ stabilized ADH on the GC surface (at least during 30 HCV scans), preventing its denaturation and promoting electron transfer by optimizing protein orientation (achieved by adding multivalent positive ions (NAD⁺ and PM) that bridge between the negatively charged surfaces of both electrode and protein), as previously observed by our group for HRP⁴⁸ and GOx⁴⁷ embedded in PM. The persistent ADH-mediated ethanol oxidation (ADH DET) suggests that PM-ADH-NAD⁺/Nafion films are promising materials for use as anodes in biofuel cells.^5,37–45

Also, when NAD⁺ was reduced (see eqn (3) and (4)) at sufficiently negative potential limits in the presence of ethanol, the resulting NADH was oxidized during only ten HCV scans (Fig. S5A, ESI†). In the presence of ethanol, NADH was oxidized during only five HCV scans (Fig. S5A), appearing to behave like NADH in solution^{3,4,24–27,33,34} since this film is not as stable on the GC surface. ADH alone and PM-embedded ADH proved catalytic towards ethanol oxidation (Fig. 2A and B). ADH trapped along with NAD⁺ catalyzed ethanol oxidation more persistently than ADH trapped along with NADH (Fig. S6, ESI†). ADH trapped along with NAD⁺ and embedded along with PM catalyzed ethanol oxidation more effectively than unembedded ADH trapped along with NAD⁺ and unembedded or PM-embedded ADH trapped along with NADH (Fig. S7, ESI†). Competition appears to occur between acetaldehyde output and ethanol input at the GC/ADH-NAD⁺/Nafion electrode, slowing the ADH-mediated oxidation of ethanol (note, in the narrow potential scan window, that the anodic current peaks drastically decrease during the HCV scans; compare Fig. S7A with inset A, ESI†). This competition appears to be the principal disadvantage with the GC/ADH-NAD⁺/Nafion and GC/PM-ADH-NAD⁺/Nafion electrodes. Finally, PM-embedded ADH under a VC XC-72 film appears to catalyze ethanol oxidation more effectively when the potential is not much higher than 0.8 V (positive-going scan). This oxidation is instead catalyzed by the structural arrangement caused by ADH trapping along with NAD⁺ (Fig. S8, ESI†), differently from the mechanism proposed by Punckt et al.,^33,34 since no catalytic effects can be attributed to differences in VC XC-72 film chemical structure or electrode morphology (e.g., porosity of the VC XC-72 film on the GC surface).

Fig. 3 depicts the first HCV scans for bare GC at 0.02 V s⁻¹ and for GC/PM-ADH-NAD⁺/Nafion in N₂-saturated 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol, at various potential scan rates.

Fig. 3 Hydrodynamic cyclic voltammograms (first scans) for bare GC at 0.02 V s⁻¹ (black curve) and for a GC/PM-ADH-NAD⁺/Nafion electrode in N₂-saturated 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol, at various potential scan rates. Scans started at 0.2 V in the positive-going direction. ω = 200 rpm. Inset: I_apν⁻¹ vs. ν.

Anodic current peaks (I_ap) were proportional to potential scan rates (ν) (inset to Fig. 3), characterizing diffusionless electrochemical systems^47,48,56 and DET from ADH to GC, resulting in Γ_ADH values determinable by integrating the graph under the HCV oxidation peaks (Fig. 3), yielding a value of 3.6 ± 0.6 nmol cm⁻² for the GC/PM-ADH-NAD⁺/Nafion electrode. Clearly, this value is sufficiently decreased by conformational structure changes as the number of HCV scans is increased, as previously commented (e.g., after five HCV scans, Γ_ADH reaches 0.3 ± 0.1 nmol cm⁻² for the GC/PM-ADH-NAD⁺/Nafion electrode).

The diagnostic criteria for ideal irreversible reactions (Butler–Volmer kinetics) proposed by Honeychurch and Rechnitz⁵⁷ are in agreement with our results shown in Fig. 3 – namely, (a) a linear plot of I_ap versus ν intercepting the origin was observed in our case using data from Fig. 3; (b) the peak widths at half-peak heights, W_1/2, where W_1/2 = 62.5/(1 − α)n mV, had an average value of 90 mV, and α an average value of 0.65, close to the α value obtained from (c) the slope 2.3RT/(1 − α)nF of a linear plot of anodic potential peaks (E_ap) versus log ν, using data from Fig. 3.

The rate constant (k_s) for reaction 1 can be determined from:^56,57

log k_s = α log(1 − α) + (1 − α)log α − log(RT/nFν) − [α(1 − α)nFΔE_p/2.3RT] (6)

when ΔE_p > 200/n mV, and where we assumed ΔE_p = (E_ap − E^o′_aa) − (E_cp − E^o′_ac) ≈  E_ap − E^o′_aa ≈ W_1/2. In our case, W_1/2 approached 90 mV, close to the expected ΔE_p > 200/n mV when n = 2, assuming the occurrence of reaction 1.

The k_s value of 0.82 s⁻¹ obtained for ν = 120 mV s⁻¹ is close to the k_s value obtained for a GC/PM-HRP/Nafion or a GC/PM-GOx/Nafion electrode.^47,48

3.3 EIS responses from GC and modified GC electrodes in 0.1 M KH₂PO₄ N₂-saturated solution (after HCV scans in the absence or presence of 100 mM ethanol) concomitantly with 1 mM K₃Fe(CN)₆ (including CV responses)

Fig. S9 (ESI†) depicts the CV responses obtained from bare GC and also after 30 HCV scans in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol, for bare GC, GC/ADH/Nafion (in the absence of ethanol), GC/PM-ADH/Nafion (in the absence of ethanol), GC/NAD⁺/Nafion (in the absence of ethanol), GC/PM-ADH-NAD⁺/Nafion (in the absence of ethanol), GC/ADH/Nafion, GC/PM-ADH/Nafion, and GC/PM-ADH-NAD⁺/Nafion electrodes in the presence of 1 mM K₃Fe(CN)₆. The bare GC electrode response is characteristic of Fe(CN)₆³⁻ redox probes on bare GC.^47,48,58,59 Redox peaks are centered at 0.19 V. The current peaks related to the Fe(CN)₆³⁻ redox probe are decreased for bare GC after 30 HCV scans in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol, and quite absent for GC/ADH/Nafion (in the absence of ethanol), GC/PM-ADH/Nafion (in the absence of ethanol), GC/NAD⁺/Nafion (in the absence of ethanol), GC/PM-ADH-NAD⁺/Nafion (in the absence of ethanol), GC/ADH/Nafion, GC/PM-ADH/Nafion, and GC/PM-ADH-NAD⁺/Nafion electrodes after 30 HCV scans in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol, indicating a strong blocking effect of these devices on the Fe(CN)₆³⁻ charge transfer process.

In order to better elucidate the electrochemical responses depicted in Fig. S9 (ESI†) we conducted EIS experiments at OCP, which was 0.11 V on average for bare GC, and after 30 HCV scans in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol for bare GC, GC/ADH/Nafion (in the absence of ethanol), GC/PM-ADH/Nafion (in the absence of ethanol), GC/NAD⁺/Nafion (in the absence of ethanol), GC/PM-ADH-NAD⁺/Nafion (in the absence of ethanol), GC/ADH/Nafion, GC/PM-ADH/Nafion, and GC/PM-ADH-NAD⁺/Nafion electrodes in the presence of K₃Fe(CN)₆ (Fig. 4).

Fig. 4 Impedance plane plots for bare GC (□) and drawn after 30 HCV scans in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol for bare GC (○), GC/ADH/Nafion (in the absence of ethanol) (△), GC/PM-ADH/Nafion (in the absence of ethanol) (▽), GC/NAD⁺/Nafion (in the absence of ethanol) (◇), GC/PM-ADH-NAD⁺/Nafion (in the absence of ethanol) (◁), GC/ADH/Nafion (▷), GC/PM-ADH/Nafion ([hexagon open, point down]), and GC/PM-ADH-NAD⁺/Nafion ([pentagon open]) electrodes in N₂-saturated 0.1 M KH₂PO₄ (pH 6.5) containing 1 mM K₃Fe(CN)₆. Potential perturbation: 25 mV (rms). Frequency range: 100 kHz–10 mHz. Constant potential for EIS acquisition: OCP (0.11 V vs. SCE on average). Lines represent spectra (adjusted) calculated using a non-linear least-squares program, conforming to the equivalent circuit R_s[Q_dl(R_ctW_lf)], or R_s[Q_dl(R_ctQ_lf)]. Calculated average values: R_s = 80 Ω, Q_dl = 6.0 μF sⁿ⁻¹, n = 0.9, and W_lf = 75 μF. Inset: the same curves, restricted to 5 kΩ.

Fig. 4 shows typical EIS readings for complex plane impedance plots combining regions of mass transfer and kinetic control at low and high frequencies, respectively.⁵⁵ When the electrochemical system is kinetically sluggish, high R_ct values (charge transfer resistance) are found within a well-defined semicircular region, displaying a limited frequency range in which mass transfer is a significant factor⁵⁵ (see curve ○ in Fig. 4). When R_ct is low, the system is so kinetically facile that mass transfer always plays a role, and the semicircular region is not well defined⁵⁵ (see curve □ in Fig. 4).

The simplest equivalent circuit of an electrochemical cell is a Randles equivalent circuit composed of resistors and capacitors.⁵⁵ This type of equivalent circuit was used in the present study and perfectly fitted the EIS results obtained (lines on the EIS curves shown). The equivalent circuit employed can be represented as R_s[Q_dl(R_ctW_lf)] ,⁵⁹ where R_s stands for the solution resistance, Q_dl for the constant phase element involving an n exponent to represent C_dl, W_lf for a Warburg impedance indicating a purely diffusion-controlled process at the low-frequency limits,⁶⁰ and Q_lf for a constant phase element at the low-frequency limits.

The results obtained at OCP (Table 1) (see ref. 47, 48, 58 and 59 for the assumed conditions to obtain EIS results and the relation between k⁰ (the standard heterogeneous rate constant) (or k⁰_app) and R_ct) reveal that k⁰ for bare GC in the presence of 1 mM K₃Fe(CN)₆ has a high value (0.0018 cm s⁻¹), even if compared to bare GC after 30 HCV scans in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol (0.000035 cm s⁻¹) or to a GC surface covered with a PM-ADH/Nafion film after 30 HCV scans in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol (charge transfer rate 1070 times lower (k⁰_app approaches 0.000002 cm s⁻¹) than for bare GC in the presence of 1 mM K₃Fe(CN)₆), which decreases the rate of charge transfer to the Fe(CN)₆³⁻ probe at OCP. The charge transfer rates are slightly lower for GC/ADH/Nafion (1.3 times; k⁰_app approaches 0.000012 cm s⁻¹) and GC/PM-ADH/Nafion (1.8 times; k⁰_app approaches 0.000002 cm s⁻¹) after 30 HCV scans in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol than for the same electrodes and conditions in the absence of ethanol (Table 1). They were, however, slightly higher (7 times; k⁰_app approaches 0.000017 cm s⁻¹) for GC/PM-ADH-NAD⁺/Nafion after 30 HCV scans in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol than for the same electrodes and conditions in the absence of ethanol (Table 1). This behavior is fully consistent with the pronounced decrease in Fe(CN)₆³⁻ CV current peaks observed for the GC/ADH/Nafion, GC/PM-ADH/Nafion, and GC/PM-ADH-NAD⁺/Nafion electrodes (Fig. S9, ESI†), reinforcing the view that the GC/ADH/Nafion, GC/PM-ADH/Nafion, and GC/PM-ADH-NAD⁺/Nafion electrodes strongly block charge transfer to the Fe(CN)₆³⁻ probe. These electrodes exhibited k⁰_app as low as and Γ_ADH values as high as those of a GC/PM-HRP/Nafion or GC/PM-GOx/Nafion electrode,^47,48 and as a result the ADH/Nafion, PM-ADH/Nafion, and PM-ADH-NAD⁺/Nafion films should be much more densely packed than PM-HRP/Nafion or PM-GOx/Nafion films, most probably because of ADH or combined PM-ADH or PM-ADH-NAD⁺ structural features.

Table 1 Approximate R_ct and k⁰ (or k⁰_app) values obtained from non-linear least-squares calculations for elements of the equivalent circuit R_s[Q_dl(R_ctQ_lf)], or R_s[Q_dl(R_ctQ_lf)], adjusted for EIS responses (Fig. 4) provided by bare GC and, after 30 HCV scans in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol, by bare GC, GC/ADH/Nafion (in the absence of ethanol), GC/PM-ADH/Nafion (in the absence of ethanol), GC/NAD⁺/Nafion (in the absence of ethanol), GC/PM-ADH-NAD⁺/Nafion (in the absence of ethanol), GC/ADH/Nafion, GC/PM-ADH/Nafion, and GC/PM-ADH-NAD⁺/Nafion electrodes in N₂-saturated 0.1 M KH₂PO₄ (pH 6.5) containing 1 mM K₃Fe(CN)₆

Electrode	Rct (kΩ)	K^0 or k^0app^a (cm s^−1 × 10^5)
a k^0 (or k^0app) values obtained from Rct as described in ref. 47, 48, 58 and 59.
Bare GC	0.75	180
Bare GC after 30 HCV scans in 0.1 M KH2PO4 (pH 6.5) containing 100 mM ethanol	39	3.5
GC/ADH/Nafion after 30 HCV scans in 0.1 M KH2PO4 (pH 6.5)	88	1.5
GC/PM-ADH/Nafion after 30 HCV scans in 0.1 M KH2PO4 (pH 6.5)	450	0.3
GC/NAD^+/Nafion after 30 HCV scans in 0.1 M KH2PO4 (pH 6.5)	135	1.0
GC/PM-ADH-NAD^+/Nafion after 30 HCV scans in 0.1 M KH2PO4 (pH 6.5)	535	0.25
GC/ADH/Nafion after 30 HCV scans in 0.1 M KH2PO4 (pH 6.5) containing 100 mM ethanol	111	1.2
GC/PM-ADH/Nafion after 30 HCV scans in 0.1 M KH2PO4 (pH 6.5) containing 100 mM ethanol	802	0.2
GC/PM-ADH-NAD^+/Nafion after 30 HCV scans in 0.1 M KH2PO4 (pH 6.5) containing 100 mM ethanol	78	1.7

---

The CV responses obtained after 30 HCV scans (see Fig. S1 in the ESI† for HCV responses) in 0.1 M KH₂PO₄ (pH 6.5) using GC/NAD⁺/Nafion, and GC/NADH/Nafion electrodes in the presence of 1 mM K₃Fe(CN)₆, depicted in Fig. S10 (ESI†), combined with the EIS responses obtained at OCP for these electrodes, depicted in Fig. S11 (ESI†), reinforce the view that the GC/NADH/Nafion electrode blocks charge transfer to the Fe(CN)₆³⁻ probe (k⁰_app approaches 0.00010 cm s⁻¹, Table S1, ESI†) more strongly than the GC/NAD⁺/Nafion electrode (k⁰_app approaches 0.00028 cm s⁻¹, Table S1, ESI†). Similarly, the CV responses obtained after 30 HCV scans (see Fig. S2 and S4 (ESI†), for HCV responses) in 0.1 M KH₂PO₄ (pH 6.5) for GC/ADH-NAD⁺/Nafion, GC/ADH-NADH/Nafion, GC/PM-ADH-NAD⁺/Nafion, and GC/VC XC-72/PM-ADH/Nafion electrodes in the presence of 1 mM K₃Fe(CN)₆, depicted in Fig. S12 (ESI†), combined with the EIS responses obtained at OCP for these electrodes, depicted in Fig. S13 (ESI†), reinforce the view that GC/ADH-NAD⁺/Nafion and GC/ADH-NADH/Nafion block charge transfer to the Fe(CN)₆³⁻ probe (k⁰_app values approach 0.000003 and 0.000006 cm s⁻¹, respectively; Table S1, ESI†) more strongly than GC/PM-ADH-NAD⁺/Nafion and GC/VC XC-72/PM-ADH/Nafion (k⁰_app values approach 0.000018 and 0.00014 cm s⁻¹, respectively; Table S1, ESI†).

The CV responses obtained after 30 HCV scans (see Fig. S5 and S6 (ESI†), for HCV responses) in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol using GC/NAD⁺/Nafion, GC/NADH/Nafion, GC/ADH-NAD⁺/Nafion, and GC/ADH-NADH/Nafion electrodes in the presence of 1 mM K₃Fe(CN)₆, depicted in Fig. S14 (ESI†), combined with the EIS responses obtained at OCP for these electrodes, depicted in Fig. S15 (ESI†), reinforce the view that GC/ADH-NAD⁺/Nafion and GC/ADH-NADH/Nafion block charge transfer to the Fe(CN)₆³⁻ probe (k⁰_app values approach 0.000026 and 0.00001 cm s⁻¹, respectively; Table S1, ESI†) slightly less strongly than the same modified electrodes in the absence of ethanol (k⁰_app approaches 0.000003 cm s⁻¹; Table S1, ESI†). The CV responses obtained after 30 HCV scans (see Fig. S7 (ESI†), for HCV responses) in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol using GC/ADH-NAD⁺/Nafion (also in the narrow potential scan window), GC/PM-ADH-NAD⁺/Nafion, GC/ADH-NADH/Nafion, and GC/PM-ADH-NADH/Nafion electrodes in the presence of 1 mM K₃Fe(CN)₆, depicted in Fig. S16 (ESI†), combined with the EIS responses obtained at OCP for these electrodes, depicted in Fig. S17 (ESI†), reinforces the view that GC/ADH-NAD⁺/Nafion and GC/PM-ADH-NAD⁺/Nafion block charge transfer to the Fe(CN)₆³⁻ probe (k⁰_app approaches 0.000006 cm s⁻¹; Table S1, ESI†) more effectively than GC/ADH-NAD⁺/Nafion after 30 HCV scans (see Fig. S6 (ESI†), for HCV responses) in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol (k⁰_app approaches 0.000026 cm s⁻¹; Table S1, ESI†). The CV responses obtained after 30 HCV scans (see Fig. S8 (ESI†), for HCV responses) in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol for the GC/VC XC-72/PM-ADH-NAD⁺/Nafion electrode in the presence of 1 mM K₃Fe(CN)₆, depicted in Fig. S18 (ESI†), combined with the EIS responses obtained at OCP for these electrodes, depicted in Fig. S19 (ESI†), reinforce the view that GC/VC XC-72/PM-ADH-NAD⁺/Nafion blocks charge transfer to the Fe(CN)₆³⁻ probe (k⁰_app approaches 0.000093 cm s⁻¹; Table S1, ESI†) more effectively than the same electrode after 30 HCV scans (see Fig. S4 (ESI†), for HCV responses) in 0.1 M KH₂PO₄ (pH 6.5) and in the absence of ethanol (k⁰_app approaches 0.00014 cm s⁻¹; Table S1, ESI†).

The CV responses obtained after 5 HCV scans (see Fig. 3 for HCV responses) in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol using bare GC and GC/PM-ADH-NAD⁺/Nafion electrodes in the presence of 1 mM K₃Fe(CN)₆, depicted in Fig. S20 (ESI†), combined with the EIS responses obtained at OCP for these electrodes, depicted in Fig. S21 (ESI†), reinforce the view that GC/PM-ADH-NAD⁺/Nafion blocks charge transfer to the Fe(CN)₆³⁻ probe (k⁰_app approaches 0.000002 cm s⁻¹; Table S2, ESI†) more effectively than bare GC (k⁰_app approaches 0.000057 cm s⁻¹; Table S2, ESI†).

Finally, EIS experiments (Fig. 5) were conducted at 0.65 V (choice of this value was based on the current peak potential for ADH oxidation at a GC/PM-ADH-NAD⁺/Nafion electrode—see Fig. 2 and 3) in the presence of 100 mM ethanol in N₂-saturated 0.1 M KH₂PO₄ (pH 6.5) solution, revealing a noteworthy behavior in terms of R_ct (Table 2).

Fig. 5 Impedance plane plots drawn after 5 HCV scans in N₂-saturated 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol for bare GC at 20 mV s⁻¹ (□) and GC/PM-ADH-NAD⁺/Nafion electrodes at 20 (○), 50 (△), 80 (▽), and 100 mV s⁻¹ (⋄). Potential perturbation: 25 mV (rms). Frequency range: 100 kHz–10 mHz. Constant potential for EIS acquisition: 0.65 V vs. SCE. Lines represent spectra (adjusted) calculated using a non-linear least-squares program, conforming to the equivalent circuit R_s[Q_dl(R_ctW_lf)], or R_s[Q_dl(R_ctQ_lf)]. Calculated average values: R_s = 96 Ω, Q_dl = 3.3 μF sⁿ⁻¹, n = 0.9, and W_lf = 39 μF. Inset: the same curves, restricted to 50 kΩ.

Table 2 Approximate R_ct values obtained from non-linear least-squares calculations for elements of the equivalent circuit R_s[Q_dl(R_ctW_lf)], or R_s[Q_dl(R_ctQ_lf)], adjusted for EIS responses at 0.65 V (Fig. 5) provided by bare GC and GC/PM-ADH-NAD⁺/Nafion electrodes after five HCV scans in N₂-saturated 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol

Electrode	Rct (kΩ)
Bare GC (20 mV s^−1)	120
GC/PM-ADH-NAD^+/Nafion (20 mV s^−1)	115
GC/PM-ADH-NAD^+/Nafion (50 mV s^−1)	74
GC/PM-ADH-NAD^+/Nafion (80 mV s^−1)	60
GC/PM-ADH-NAD^+ Nafion (100 mV s^−1)	36

---

R_ct values for GC/PM-ADH-NAD⁺/Nafion in the presence of ethanol were lower than R_ct for bare GC (Fig. 5 and Table 2) and can possibly be attributed to effective electron hopping^{47,48,58,61–64} not only between the region comprising three protein residues (Cys-46, His-67, and Cys-174)—including the catalytic zinc ion—and enzymes, but also between these regions and ethanol molecules.

The difference in R_ct values (assumed as resistance to electron hopping, R_eh (ref. 47 and 48)) between bare GC and GC/PM-ADH-NAD⁺/Nafion electrodes can be 84 kΩ lower for the GC/PM-ADH-NAD⁺/Nafion electrode during ethanol oxidation than for bare GC in the presence of ethanol (see values in Table 2: 36 and 120 kΩ), depending on how much faster the potential scan rate was before EIS measurements. These values lend support to the reactions proposed above for ADH and ethanol oxidation, which are supposed to sufficiently bioelectrocatalyze electron hopping (DET) between ADH and the GC surface and between ADH and ethanol.

3.4 Physical characterization of bare GC and GC/PM-ADH-NAD⁺ (or NADH)/Nafion electrodes

SEM images acquired ex situ show GC surfaces (Fig. 6A) to be smoother than their counterparts on modified GC electrodes. Presence of a smooth film is suggested on the GC/PM-ADH-NAD⁺ electrode (Fig. 6B and D), whereas the GC/PM-ADH-NADH electrode exhibits a cracked film (Fig. 6C and E).

Fig. 6 SEM images of different electrodes: bare GC (A), 1000× magnification; GC/PM-ADH-NAD⁺/Nafion (B), 100×; GC/PM-ADH-NADH/Nafion (C), 100×; GC/PM-ADH-NAD⁺/Nafion (D), 1000×; GC/PM-ADH-NADH/Nafion (E), 1000×; GC/PM-ADH-NAD⁺/Nafion (F), 25 000×; and GC/PM-ADH-NADH/Nafion (G), 25 000×. Images are representative of five regions on each electrode.

Small granules were detected on GC/PM-ADH-NAD⁺ (Fig. 6F), in contrast with the larger granules found on GC/PM-ADH-NADH (Fig. 6G). These SEM images corroborate the assumption that PM-ADH-NAD⁺ and PM-ADH-NADH films on the GC surface strongly block charge transfer to the Fe(CN)₆³⁻ probe, as previously discussed using CV and EIS results.

3.5 NMR analysis for acetaldehyde production from ethanol oxidation by DET from ADH

Production of acetaldehyde from ethanol oxidation by ADH DET was elucidated by NMR spectroscopy of hydrogen signals observed in the 1D spectrum at 300 MHz using a 300 μL aliquot (inserted in the NMR tube with a few drops of deuterated water) collected directly from the electrochemical cell solution in the vicinity of the GC/ADH/Nafion electrode after 5 HCV scans (−0.8–0.8 V), the GC/PM-ADH/Nafion electrode after 5 HCV scans (−0.8–0.8 V), and the GC/PM-ADH-NAD⁺/Nafion electrode after 30 HCV scans (−0.8–0.8 V) in 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol. An aliquot from the same system, collected before performing the HCV scans, was used for comparison (Fig. 7).

Fig. 7 ¹H NMR spectrum for an aliquot of 0.1 M KH₂PO₄ (pH 6.5) containing 100 mM ethanol before (A) and after (B) the GC/ADH/Nafion electrode underwent 5 HCV scans, (C) the GC/PM-ADH/Nafion electrode underwent 5 HCV scans, and (D) the GC/PM-ADH-NAD⁺/Nafion electrode underwent 30 HCV scans.

The stronger band located at around δ 4.65 (Fig. 7, S22 and S23, ESI†) was attributed to hydrogen from water, being the only band present in an aliquot of 0.1 M KH₂PO₄ (pH 6.5) (Fig. S22A, ESI†). For 100 mM ethanol in 0.1 M KH₂PO₄ (pH 6.5) collected before the electrodes had undergone HCV scans (Fig. 7A, S22B and S23D, ESI†), a triplet was observed at around δ 0.98, attributed to a CH₃ group adjacent to a CH₃ group, and a quadruplet was detected at around δ 3.46, attributed to CH₂ adjacent to a CH₂ group characteristic of ethanol. The same peaks, however, were visible before and after performing 5 or 30 HCV scans for 100 mM ethanol in 0.1 M KH₂PO₄ (pH 6.5) using GC/ADH/Nafion, GC/PM-ADH/Nafion, and GC/PM-ADH-NAD⁺/Nafion electrodes (Fig. 7, S22B and S23, ESI†). In addition, a quadruplet was observed at δ 9.55 (see Fig. 7B–D and S23A, ESI†), corresponding to a CH adjacent to an acetaldehyde CH₃ group. For 20 mM acetaldehyde in 0.1 M KH₂PO₄ (pH 6.5) (see enlarged details in Fig. S22C, ESI†), a quadruplet was detected at δ 9.50, corresponding to a CH adjacent to a CH₃ group, as well as a doublet at δ 2.06, attributed to a CH₃ group adjacent to an acetaldehyde CH group, a quadruplet at δ 5.08, from a CH group adjacent to a CH₃ group, and a doublet at δ 1.16, from a CH₃ group adjacent to a CH group pertaining to a hydrated form of acetaldehyde detectable because of the equilibrium established between acetaldehyde and its hydrated form in the presence of water.⁶⁵ The small peak found at δ 1.75 (Fig. S22C, ESI†) can be attributed to OH groups of hydrated acetaldehyde.

Because acetaldehyde is not as stable in aqueous solutions (e.g., equilibrium between acetaldehyde and its hydrated form in the presence of water), we were not able to detect acetaldehyde from ethanol oxidation by DET from ADH present in the GC/ADH/Nafion and GC/PM-ADH/Nafion electrodes after 30 HCV scans (Fig. S23B and C, ESI†), coincidentally with an absence of redox currents or peaks for these electrodes after 30 HCV scans (Fig. 2A and B). In contrast, we detected acetaldehyde from ethanol oxidation by DET from ADH present in the GC/PM-ADH-NAD⁺/Nafion electrode after 30 HCV scans (Fig. 7D), coincidentally with the presence of an anodic peak at 0.74 V for this electrode after 30 HCV scans (Fig. S7B, ESI†). However, we did not detect acetaldehyde from the GC/PM-ADH-NAD⁺/Nafion electrode after it was subjected to ω = 200 rpm for 3 hours (Fig. S23D, ESI†), although acetaldehyde was detected from this electrode when it underwent 5 HCV scans after 3 hours under ω = 200 rpm (Fig. S23A, ESI†).

The ¹H NMR results unequivocally demonstrate the production of acetaldehyde from ethanol oxidation by DET from ADH present in the GC/ADH/Nafion, GC/PM-ADH/Nafion, and GC/PM-ADH-NAD⁺/Nafion electrodes, and the persistent acetaldehyde production from the last electrode.

4 Conclusions

A Nafion-covered ADH or PM-ADH or PM-ADH-NAD⁺ (or NADH) film produced on a GC surface proved capable of DET from ADH to GC and from ADH to ethanol—findings previously unreported in the research literature. The PM-ADH-NAD⁺/Nafion film was rich in ADH enzymes with electroactive configuration. This configuration was retained even after 30 continuous HCV scans in the presence of ethanol, demonstrating the importance of NAD⁺ and PM as promoter molecules—i.e., stabilizing ADH, preventing its denaturation, and promoting electron transfer by optimizing ADH orientation (achieved by adding multivalent positive ions (NAD⁺ and PM) that bridge the negatively charged surfaces of both electrode and protein). This film-modified electrode proved sufficiently persistent to ensure bioelectrocatalytic ethanol oxidation, making this system a promising resource for use as an anode in biofuel cells. Also, for this modified electrode a rate constant (k_s) of 0.82 s⁻¹ was obtained for ν = 120 mV s⁻¹. Resistance to electron hopping taking place between the region comprising three protein residues (Cys-46, His-67, and Cys-174)—including the catalytic zinc ion—probably involving mostly the His-67 residue from the ADH enzymes and occurring between that region and ethanol molecules, was estimated by EIS experiments as 84 kΩ lower for the GC/PM-ADH-NAD⁺/Nafion electrode during ethanol oxidation than for bare GC in the presence of ethanol, depending on how much faster the potential scan rate was before EIS measurements. Charge transfer to the Fe(CN)₆³⁻ probe was strongly blocked by the presence of PM-ADH, ADH-NAD⁺, and PM-ADH-NAD⁺ films on the GC surface. Ethanol oxidation by DET from ADH present in the GC/ADH/Nafion, GC/PM-ADH/Nafion, and GC/PM-ADH-NAD⁺/Nafion electrodes was confirmed by production of acetaldehyde, which was unquestionably elucidated by the ¹H NMR results.

Acknowledgements

The authors wish to thank PROPP-UFMS, FUNDECT-MS (grant 23/200.155/2009), and CNPq (grants 301403/2011-2 and 473991/2012-8) for their financial support. F.L. thanks CAPES for the fellowship. Thanks are also given to L. M. Ravaglia for conducting the NMR spectra acquisition.

References

1. X. D. Benetton, S. Srikanth, Y. Satyawali, K. Vanbroekhoven and D. Pant, J. Microb. Biochem. Technol., 2013, S6:007.
2. Y.-H. P. Zhang, Biotechnol. Bioeng., 2010, 105, 663–677.
3. L. Gorton and P. N. Bartlett, in Bioelectrochemistry: Fundamentals, Experimental Techinques and Applications, ed. P. N. Bartlett, John Wiley & Sons, England, 2008, ch. 4, pp. 157–198.
4. L. Gorton and E. Domínguez, in Encyclopedia of Electrochemistry, Bioelectrochemistry, ed. A. J. Bard, M. Stratmann and G. S. Wilson, Wiley-VCH, Weinheim, 2002, vol. 9, ch. 4, pp. 67–143.
5. I. Willner and E. Katz, Angew. Chem., Int. Ed., 2000, 39, 1180–1218.
6. P. Schenkels, S. de Vries and A. J. J. Straathof, Biocatal. Biotransform., 2001, 19, 191–212.
7. S. R. Piersma, S. de Vries and J. A. Duine, in Enzymology and Molecular Biology of Carbonyl Metabolism 6, ed. H. Weiner, R. Lindahl, D. W. Crabb and T. G. Flynn, Plenum Press, New York, 1997, ch. 48, pp. 425–434.
8. D. Gamenara, G. A. Seoane, P. S.-Méndez and P. D. de María, Redox Biocatalysis – Fundamentals and Applications, John Wiley & Sons, New Jersey, 2013, p. 119.
9. M. J. Lobo, A. J. Miranda and P. Tuñón, Electroanalysis, 1997, 9, 191–202.
10. T. Yao and S. Musha, Anal. Chim. Acta, 1979, 110, 203–209.
11. L. Gorton, G. Bremle, E. Csöregi, G. J. Pettersson and B. Persson, Anal. Chim. Acta, 1991, 249, 43–54.
12. E. Domínguez, H. L. Lan, Y. Okamoto, P. D. Hale, T. A. Skotheim, L. Gorton and B. H. Hägerdal, Biosens. Bioelectron., 1993, 8, 229–237.
13. Q. Chi and S. Dong, Anal. Chim. Acta, 1994, 285, 125–133.
14. M. J. L. Castañón, A. J. M. Ordieres and P. T. Blanco, Biosens. Bioelectron., 1997, 12, 511–520.
15. P. C. Pandey, S. Upadhyay, B. C. Upadhyay and H. C. Pathak, Anal. Biochem., 1998, 260, 195–203.
16. R. Antiochia and I. Lavagnini, Anal. Lett., 2006, 39, 1643–1655.
17. B. Kowalewska and P. J. Kulesza, Electroanalysis, 2012, 24, 254–263.
18. B. Kowalewska and P. J. Kulesza, Anal. Chem., 2012, 84, 9564–9571.
19. C. Shan, H. Yang, D. Han, Q. Zhang, A. Ivaska and L. Niu, Biosens. Bioelectron., 2010, 25, 1504–1508.
20. L. Li, H. Lu and L. Deng, Talanta, 2013, 113, 1–6.
21. A. P. Periasamy, Y. Umasankar and S.-M. Chen, Talanta, 2011, 83, 930–936.
22. K. Guo, K. Qian, S. Zhang, J. Kong, C. Yu and B. Liu, Talanta, 2011, 85, 1174–1179.
23. H. Teymourian, A. Salimi and R. Hallaj, Talanta, 2012, 90, 91–98.
24. J. Moiroux and P. J. Elving, Anal. Chem., 1978, 50, 1056–1062.
25. J. Moiroux and P. J. Elving, J. Electroanal. Chem., 1979, 102, 93–108.
26. P. J. Elving, W. T. Bresnahan, J. Moiroux and Z. Samec, Bioelectrochem. Bioenerg., 1982, 9, 365–378.
27. Z. Samec and P. J. Elving, J. Electroanal. Chem., 1983, 144, 217–234.
28. H. Wu, C. Tian, X. Song, C. Liu, D. Yang and Z. Jiang, Green Chem., 2013, 15, 1773–1789.
29. S. Kochius, A. O. Magnusson, F. Hollmann, J. Schrader and D. Holtmann, Appl. Microbiol. Biotechnol., 2012, 93, 2251–2264.
30. A. Radoi and D. Compagnone, Bioelectrochemistry, 2009, 76, 126–134.
31. M. Pumera, R. Scipioni, H. Iwai, T. Ohno, Y. Miyahara and M. Boero, Chem.–Eur. J., 2009, 15, 10851–10856.
32. H. Li, K. E. Worley and S. C. Barton, ACS Catal., 2012, 2, 2572–2576.
33. C. Punckt, M. A. Pope, J. Liu, Y. Lin and I. A. Aksay, Electroanalysis, 2010, 22, 2834–2841.
34. C. Punckt, M. A. Pope and I. A. Aksay, J. Phys. Chem. C, 2013, 117, 16076–16086.
35. A. Ramanavicius, K. Habermüller, E. Csöregi, V. Laurinavicius and W. Schuhmann, Anal. Chem., 1999, 71, 3581–3586.
36. J. Razumiene, M. Niculescu, A. Ramanavicius, V. Laurinavicius and E. Csöregi, Electroanalysis, 2002, 14, 43–49.
37. R. L. Arechederra and S. D. Minteer, Fuel Cells, 2009, 9, 63–69.
38. A. Ramanavicius and A. Ramanaviciene, Fuel Cells, 2009, 9, 25–36.
39. I. Willner, Y.-M. Yan, B. Willner and R. Tel-Vered, Fuel Cells, 2009, 9, 7–24.
40. M. Zayats, B. Willner and I. Willner, Electroanalysis, 2008, 20, 583–601.
41. J. A. Cracknell, K. A. Vincent and F. A. Armstrong, Chem. Rev., 2008, 108, 2439–2461.
42. M. Falk, Z. Blum and S. Shleev, Electrochim. Acta, 2012, 82, 191–202.
43. M. J. Cooney, V. Svoboda, C. Lau, G. Martin and S. D. Minteer, Energy Environ. Sci., 2008, 1, 320–337.
44. G. T. R. Palmore, H. Bertschy, S. H. Bergens and G. M. Whitesides, J. Electroanal. Chem., 1998, 443, 155–161.
45. M. N. Arechederra, P. K. Addo and S. D. Minteer, Electrochim. Acta, 2011, 56, 1585–1590.
46. H. A. Heering and G. W. Canters, Engineering the Bioelectronic Interface: Applications to Analyte Biosensing and Protein Detection, ed. J. Davis, Royal Society of Chemistry Publishing, Cambridge, UK, 2009, ch. 5, pp. 119–152.
47. A. H. H. Tanji, F. Lima, S. R. Santos and G. Maia, J. Phys. Chem. C, 2012, 116, 18857–18864.
48. S. R. Santos and G. Maia, Electrochim. Acta, 2012, 71, 116–122.
49. J. E. Frew and H. A. O. Hill, Eur. J. Biochem., 1988, 172, 261–269.
50. G. V. Fortunato, L. B. Venarusso and G. Maia, ChemElectroChem, 2014, 1, 625–636.
51. P. K. Agarwal, S. P. Webb and S. H. Schiffer, J. Am. Chem. Soc., 2000, 122, 4803–4812.
52. S. H. Schiffer and S. J. Benkovic, Annu. Rev. Biochem., 2006, 75, 519–541.
53. V. Leskovac, S. Trivić and D. Peričin, FEMS Yeast Res., 2002, 2, 481–494.
54. E. G. Weinhold, A. Glasfeld, A. D. Ellington and S. A. Benner, Proc. Natl. Acad. Sci. U. S. A., 1991, 88, 8420–8424.
55. A. J. Bard and L. R. Faulkner, Electrochemical Methods: Fundamentals and Applications, John Wiley & Sons Inc, New York, 2nd edn, 2001.
56. E. Laviron, J. Electroanal. Chem., 1979, 101, 19–28.
57. M. J. Honeychurch and G. A. Rechnitz, Electroanalysis, 1998, 10, 453–457.
58. L. B. Venarusso, K. Tammeveski and G. Maia, Electrochim. Acta, 2011, 56, 8926–8933.
59. F. Lima, G. V. Fortunato and G. Maia, RSC Adv., 2013, 3, 9550–9560.
60. S. Krause, Encyclopedia of Electrochemistry – Instrumentation and Electroanalytical Chemistry, Wiley-VCH, Weinheim, 2003, vol. 3, ch. 2.6, pp. 196–229.
61. D. Chen, G. Wang and J. Li, J. Phys. Chem. C, 2007, 111, 2351–2367.
62. D. H. Murgida and P. Hildebrandt, Phys. Chem. Chem. Phys., 2005, 7, 3773–3784.
63. D. N. Blauch and J.-M. Savéant, J. Am. Chem. Soc., 1992, 114, 3323–3332.
64. D. N. Blauch and J.-M. Savéant, J. Phys. Chem., 1993, 97, 6444–6448.
65. E. Lombardi and P. B. Sogo, J. Chem. Phys., 1960, 32, 635–636.

---

Footnotes

† Electronic supplementary information (ESI) available: The electronic supplementary information contains figures and tables concerning supplementary results, and references. See DOI: 10.1039/c4ra02946a

‡ The authors contributed equally to this work.

---