Recent Progress in Applications of Enzymatic Bioelectrocatalysis

Abstract

Bioelectrocatalysis has become one of the most important research fields in electrochemistry and provided a firm base for the application of important technology in various bioelectrochemical devices, such as biosensors, biofuel cells, and biosupercapacitors. The understanding and technology of bioelectrocatalysis have greatly improved with the introduction of nanostructured electrode materials and protein-engineering methods over the last few decades. Recently, the electroenzymatic production of renewable energy resources and useful organic compounds (bioelectrosynthesis) has attracted worldwide attention. In this review, we summarize recent progress in the applications of enzymatic bioelectrocatalysis.

1. Introduction

Oxidoreductases catalyze redox reactions between two sets of redox substrate couples and are considered industrially useful catalysts due to their high activities and substrate specificities under mild conditions (room temperature, normal pressure, and neutral pH). However, most oxidoreductases, in addition to nicotinamide cofactor (NAD(P))-dependent enzymes, show low substrate specificities for one of the substrates. Such redox enzymes can accept or donate electrons from or to electrodes directly or via artificial redox mediators. The coupled reaction is called bioelectrocatalysis, and the catalytic function of the redox enzymes provides a variety of specific and strong catalytic activities to nonspecific electrode reactions. [1,2,3,4,5]. Bioelectrocatalysis provides a firm base for characterizing redox enzyme reactions and applying the concept and related technologies to useful bioelectrochemical devices such as biosensors [6,7,8,9,10,11,12,13,14], biofuel cells [11,15,16,17,18,19,20,21], biosupercapacitors [22], and other bioreactors [23].

Bioelectrocatalytic reactions are classified into two types according to the mode of the electron transfer described above: direct electron transfer (DET) and mediated electron transfer (MET), as shown in Figure 1. These reactions proceed in the following schemes in the simple case of the substrate’s oxidation. On the other hand, we note here that some redox enzymes that have a catalytic site alone, without any other redox site(s), are also able to show DET-type bioelectrocatalysis, e.g., cytochrome c peroxidase [24], horseradish peroxidase (HRP) [25,26,27], ferredoxin-NADP⁺ reductase [28], flavin adenine dinucleotide (FAD)-dependent glucose oxidase (FAD-GOD) [29,30], and FAD-dependent glucose dehydrogenase (FAD-GDH) [31].

The DET-type reaction is given by Equations (1) and (2):

$$\mathrm{S}+\frac{n_{\mathrm{S}}}{n_{\mathrm{E}}}{\mathrm{E}}_{\mathrm{O}}\stackrel{\mathrm{Enzyme}}{\to }\frac{n_{\mathrm{S}}}{n_{\mathrm{P}}}\mathrm{P}+\frac{n_{\mathrm{S}}}{n_{\mathrm{E}}}{\mathrm{E}}_{\mathrm{R}},$$

(1)

and

$${\mathrm{E}}_{\mathrm{R}}\begin{array}{c}\mathrm{Electrode}\\ \rightleftarrows \\ \end{array} {\mathrm{E}}_{\mathrm{O}}+n_{\mathrm{E}}{\mathrm{e}}^{-},$$

(2)

On the other hand, the MET-type reaction is given by Equations (3) and (4):

$$\mathrm{S}+\frac{n_{\mathrm{S}}}{n_{\mathrm{M}}}{\mathrm{M}}_{\mathrm{O}}\stackrel{\mathrm{Enzyme}}{\to }\frac{n_{\mathrm{S}}}{n_{\mathrm{P}}}\mathrm{P}+\frac{n_{\mathrm{S}}}{n_{\mathrm{M}}}{\mathrm{M}}_{\mathrm{R}},$$

(3)

and

$${\mathrm{M}}_{\mathrm{R}}\begin{array}{c}\mathrm{Electrode}\\ \rightleftarrows \\ \end{array}{\mathrm{M}}_{\mathrm{O}}+n_{\mathrm{M}}{\mathrm{e}}^{-},$$

(4)

where S, P, E, and M indicate a substrate, a product, an enzyme, and a mediator, respectively. n_X is the number of electrons of the chemical species X. X_O and X_R are the oxidized and reduced forms of X, respectively. In DET-type bioelectrocatalysis, it is easy to construct relatively simple systems with minimum overpotentials in the electron transfer between the electrode-active redox center of an enzyme and an electrode from the thermodynamic perspectives. In this review, the electrode-active redox center means the site that can directly communicate with electrodes and is assigned to the catalytic active site (especially for redox enzymes that have the catalytic site alone) or an electron-donating/accepting site (other than the catalytic center) that constitutes the intramolecular electron transfer. However, the reported redox enzymes enable DET-type reactions are increasing but still limited in number because the electrochemical communication between an enzyme and an electrode occurs only when the electrode-active cofactor of the enzyme is close in distance to the electrode surface [32,33,34,35]. In MET-type bioelectrocatalysis, on the other hand, a variety of enzymes can be used in principle, and a suitable selection of mediators makes it possible to construct realistic systems. In addition, once both enzymes and mediators are stably immobilized on electrodes, the measurement systems work as pseudo-DET-type systems [10,11,36]. Particularly, redox polymers anchoring osmium complexes [3,37,38,39,40,41,42,43,44,45], ferricyanide [46,47], metallocenes [48,49,50,51], and viologen units [52,53,54,55] are constructed as polymeric mediators immobilized on electrodes. In summary, a DET-type system is often more ideal than a MET-type system, whereas it seems to be practical to utilize an MET-type system for several objectives.

The overpotential in bioelectrocatalysis has two components: (1) thermodynamics; the difference in the formal potential between the substrate and the electron-donating site, and (2) kinetics; slow kinetics in the heterogeneous electron transfer (Figure 1). There is no way to avoid the problem concerning the first issue as long as one utilizes natural enzymes. Slow kinetics in heterogeneous electron transfer is compensated by the so-called overpotential of the electrode in the DET-type reaction and by the large driving force (that is, increased difference in the formal potential) between the enzyme and mediator. In this review, we will describe several techniques for improving the performance of enzymatic bioelectrocatalytic systems and summarize recent studies on their applications.

2. Tuning of Nanostructured Electrodes and Protein-Engineering of Redox Enzymes

2.1. Electrode Nanomaterials

In DET-type and enzyme/mediator-immobilized MET-type bioelectrocatalysis, immobilization of as many amounts of enzymes and mediators as possible on the electrode surface is, in principle, able to increase the current density of bioelectrocatalytic systems. Therefore, nanostructured electrodes with a large ratio of the effective surface area against the projective surface area are useful and frequently utilized for bioelectrocatalytic systems. In addition, it is suggested that the heterogeneous electron transfer kinetics at the top edge of the microstructures of these electrodes is accelerated by the electric field strengthened by the expansion of the electric double layer [56] and the charge accumulation as expected by the Poisson equation [4]. This effect is very useful to decrease the overpotential due to the heterogeneous electron transfer in the DET-type reaction.

Electrode nanomaterials utilized in bioelectrocatalysis are roughly divided into two classes: carbon and metal. Carbon nanomaterials, such as carbon nanotubes [57,58,59], carbon blacks [60], carbon cryogels [61,62], and MgO-templated carbons [43,63,64], have properties to physically adsorb a large amount of enzymes and mediators at hydrophobic sites and are generally used as platforms favorable for bioelectrocatalysis. On the other hand, nanoporous gold constructed by anodization [27,65,66,67] or dealloying [68,69,70] and metallic nanoparticles of gold [57,71,72,73,74,75,76,77,78], silver [79,80,81], platinum [29,30,31], titanium oxide (TiO₂) [80,81], iron oxide (Fe₂O₃) [82,83], and indium tin oxide (ITO) [84] are also widely used. Compared to carbon nanomaterials, the pore and particle sizes of metallic nanomaterials can be easily controlled according to several manufacturing methods. In addition, conductive supports, such as polymer hydrogels, act as nanostructured electrodes [85].

Furthermore, nanostructured electrodes exert another positive effect on DET-type bioelectrocatalysis from the viewpoint of the enzyme’s orientation. The limited catalytic current density (j_cat) in DET-type reactions is expressed by Equation (5):

$$j_{\mathrm{cat}}=\pm n_{\mathrm{E}}F{k}_{\mathrm{c},\mathrm{DET}}{\Gamma }_{\mathrm{E},\mathrm{eff}},$$

(5)

where $F$ is the Faraday constant, $k_{\mathrm{c},\mathrm{DET}}$ is the catalytic constant in DET-type bioelectrocatalysis (with $k_{\mathrm{c},\mathrm{DET}}=\left(n_{\mathrm{S}}/n_{\mathrm{E}}\right)k_{\mathrm{c},\mathrm{DET}\left(1\right)}$, $k_{\mathrm{c},\mathrm{DET}\left(1\right)}$ being the catalytic constant for single turnover of the enzyme), and ${\Gamma }_{\mathrm{E},\mathrm{eff}}$ is the surface concentration of the effective enzyme. Here, “effective” means being able to electrochemically communicate with electrodes. In other words, enzymes of which the electrode-active sites face away from the electrode are not included because the long-range electron transfer kinetic constant (k°) exponentially decreases with an increase in the distance between the electrode surface and the electrode-active site in an enzyme (d) [86,87,88], as described by Equation (6):

$$\mathrm{k}°=\mathrm{k}{°}_{\max }\exp \left(-\mathsf{\beta }\mathrm{d}\right),$$

(6)

where k°_max is the rate constant at the closest approach (d = d_min) and β is the decay coefficient. Based on the random orientation model in which enzymes are assumed to be randomly oriented on electrodes [32,89,90], the enzymes can penetrate into nanostructured electrodes with mesopores, and the mesoporous electrodes can adsorb a large amount of enzymes suitable for DET-type reactions compared with planar electrodes, as illustrated in Figure 2. This is called the curvature effect of mesoporous electrodes for DET-type bioelectrocatalysis and is also very effective in decreasing the overpotential of DET-type bioelectrocatalysis. In practical applications, it is necessary to select and optimize electrode nanomaterials based on the estimated size, shape, and hydrophobicity of enzymes.

On the other hand, chemical modifications of electrode surfaces are very effective in controlling the enzyme orientation by electrostatic or other specific interactions between the electrode-active site and the electrode surface; therefore, they are useful to increase the population of the enzyme orientations with minimum distances between the electrode-active site and the electrode surface [91]. Such favorite orientations also contribute to decreasing the overpotential in the DET-type reaction. For example, [NiFe]-hydrogenase (H₂ase) from Desulfovibrio vulgaris Miyazaki F and copper efflux oxidase (CueO) expressed in Escherichia coli showed strong DET-type bioelectrocatalysis activity at (positively charged) p-phenylenediamine-functionalized Ketjen-Black-modified glassy carbon electrodes compared with Ketjen-Black-modified electrode without any chemical functionalization and artificially introduced charged groups. The surfaces near the electrode-active sites of the enzymes are estimated to be negatively charged and the attractive electrostatic interaction with positively charged electrode surfaces increases the probability of enzyme orientations favorable for the interfacial electron transfer [92,93].

In contrast, bilirubin oxidase (BOD) from Myrothecium verrucaria showed strong DET-type bioelectrocatalytic activity at negatively charged electrodes. The surface near the electrode-active site of the enzyme is positively charged and the electrostatic interaction with negatively charged electrodes increases the probability of favorable orientations of the enzyme [94]. On the other hand, the DET-type bioelectrocatalysis of BOD was also improved by modifying an electrode with bilirubin (as the natural electron donor), which seemed to attractively interact with the electrode-active type I copper site as the electron-accepting site. The interaction increases the probability of favorable orientations of the enzyme [95]. In addition, membrane-bound d-fructose dehydrogenase (FDH) from Gluconobacter japonicus NBRC3260 showed strong DET-type bioelectrocatalytic activity at a 4-mercaptophenol-modified porous gold electrode, probably due to the stabilization of the enzyme layer by the hydroxy groups of 4-mercaptophenol [96]. FDH also showed the attractive and specific interaction with methoxy substituents on the electrode surface, which resulted in the favorable orientation [97].

In addition, gas-diffusion bioelectrodes are effective for enzymatic bioelectrocatalysis in which gaseous substrates such as dihydrogen, dioxygen, and carbon dioxide were used [98]. Since these gasses have low water solubility, the bioelectrocatalytic currents are often controlled and limited by diffusion processes of the substrates at usual electrodes. On the other hand, gas-diffusion bioelectrodes realize direct supplies of gaseous substrates from the gas phase to the reaction layer due to their suitable conductivity, hydrophobicity/hydrophilicity balance, and gas permeability, as illustrated in Figure 3.

2.2. Protein-Engineering Approaches

Bioelectrocatalysis can be improved not only by functionalizing electrodes but also by engineering enzymes. Based on the crystal structure of enzymes, various mutations can be introduced into redox enzymes by protein-engineering methods [21,99].

2.2.1. Formal Potential Shift of Electrode-Active Sites

The negative shift of the formal potential of the electron-donating site for the substrate oxidation process (vice versa for the substrate reduction) by mutation is very effective in reducing the thermodynamic overpotential in the intramolecular electron transfer in redox enzymes with multiredox sites (Figure 1). The formal potentials of metallic cofactors such as hemes and copper clusters are predominantly controlled by coordinated amino acid residues. Particularly, the point mutation of the axial ligands of hemes and blue copper clusters can easily tune their formal potentials. The direction of the potential shift depends in part on the electron-donating/accepting characteristics of the mutated amino acid side chains [100,101]. Briefly in general, an axial ligand with a relatively strong electron-donating character shifts the formal potential of the cofactor in the negative potential direction due to the stabilization of the oxidized state of the cofactor and vice versa. The formal potentials of redox enzymes are also greatly interfered with by conformational changes caused by mutated amino acid residues.

FDH is among the accepted model enzymes to investigate the effects of protein engineering on DET-type bioelectrocatalysis. The enzyme is a heterotrimeric enzyme composed of a FAD-containing large catalytic subunit, a three-heme c (called hemes 1c, 2c, and 3c) from the N-terminus-containing cytochrome subunit, and a small subunit and proceeds with a DET-type reaction by transferring electrons from FAD, heme 3c, and heme 2c to an electrode in this order [102,103,104]. The sixth axial ligand (methionine 450) of heme 2c was then replaced by glutamine with electron-donating characteristics to shift the formal potential of heme 2c in the direction of the negative potential. The FDH variant (M450Q_FDH) provided DET-type bioelectrocatalytic waves for the oxidation of fructose at a half-wave potential of approximately 50 mV, more negative than that of the recombinant native enzyme, as shown in Figure 4A [103,104]. A drastic change in the limiting catalytic current was not observed. This suggests that the rate constant of the intramolecular electron transfer from FAD to heme 2c in the enzyme is sufficiently large compared with the catalytic reaction rate constant at the catalytic center (FAD); the latter predominantly determines k_c,DET in Equation (5).

On the other hand, such replacement is also effective for the axial ligand of the electron-accepting site (type I blue copper site) to negatively shift the formal potential. Replacement of the axial ligand methionine 467 in the type I copper site of BOD with glutamine (M467Q_BOD) caused a large negative shift (approximately 0.23 V) in the half-wave potential of the DET-type bioelectrocatalytic waves. This means an increase in the overpotential in the intramolecular electron transfer in the enzyme. Fortunately, in this case, the catalytic limiting current density increased compared with that of the recombinant native BOD, as shown in Figure 4B (note here that the catalytic wave in Figure 4B is illustrated for substrate oxidation) [105]. Most probably, the catalytic rate constant of the dioxygen reduction at the type II/III catalytic site is sufficiently large compared with the rate constant of the intramolecular electron transfer from the electron-accepting type I site to the dioxygen-reducing type II/III site; the intramolecular electron transfer is the rate-determining step to determine k_c,DET in Equation (5). In addition, the electron transfer kinetics seems to obey the linear free energy relationship (LFER); the increase in the formal potential difference between type I and II/III sites (that is, the driving force of the reaction) results in an increase in the intramolecular electron transfer rate constant.

Even in MET-type bioelectrocatalysis, the redox potential shift of the electron-donating site (for substrate oxidation) can tune the overpotential in the intramolecular electron transfer process. In addition, the intermolecular electron transfer kinetics between the electron-donating site in the enzyme and the mediator seem to be improved in theory by the potential shift mutation of the electron-donating site (for substrate oxidation) on the basis of the concept of LFER, though there is no report on this matter. Therefore, the mutational tuning of the redox potential of the electron-donating site (for substrate oxidation) may also expand the strategy of the mediator selection for MET-type bioelectrocatalysis.

2.2.2. Downsizing

As described by Equation (5), an increase in Γ_E,eff is essential to getting a large bioelectrocatalytic current density. Γ_E,eff can be then increased by downsizing enzymes from which the regions not deeply involved in bioelectrocatalysis are deleted, as shown in Figure 4C.

Downsizing effects were also investigated in FDH. Heme 1c of FDH is suggested to be uninvolved in the DET-type reaction, and the downsized FDH without the heme 1c region (Δ1c_FDH) showed a larger DET-type bioelectrocatalytic current density than the recombinant native FDH (r_FDH) [104,106]. In addition, a further downsized FDH without the heme 1c and 2c regions (Δ1c2c_FDH) also showed DET-type bioelectrocatalytic activity with reduced overpotential due to direct electrical communication between an electrode and heme 3c with the most negative formal potential [104,107]. However, whereas Γ_E,eff was suggested to be increased, j_cat of Δ1c2c_FDH was as large as that of r_FDH (smaller than that of Δ1c_FDH), which seemed to be ascribed to a decrease in enzymatic activity (≈ k_c,DET) of Δ1c2c_FDH by excessive deletion [104,107]. Thus, it is important to avoid conformational changes due to deletion and retain enzymatic activity as high as possible. Furthermore, a double mutant of downsizing and the potential shift (M450QΔ1c_FDH) accomplished both an increase in j_cat and an overpotential reduction [104,108,109].

2.2.3. Surface Amino Acid Mutation

It is desirable to tightly immobilize enzymes on electrodes in both DET- and MET-type bioelectrocatalytic reactions. Cross-linkers such as glutaraldehyde [110], carbodiimide [29], and maleimide [71] are often used for the covalent immobilization of enzymes [111]. The mutation of amino acid residue(s) located on the enzyme surface also enhances the cross-coupling reactions and can control the orientation of the enzyme on a suitable electrode surface. For example, cysteine introduction onto the enzyme surface close to the electrode-active site can increase the enzyme orientations suitable for DET-type reactions at thiol- and maleimide-functionalized electrodes by forming (di)sulfide bonds. Holland et al. reported DET-type bioelectrocatalysis of cysteine-introduced GOD conjugated with maleimide-modified gold nanoparticles on a gold electrode [71]. In addition, Ferapontova et al. revealed that cysteine mutation of HRP was effective for its orientation on gold electrodes to improve DET-type bioelectrocatalytic properties [112].

2.2.4. Fusion Protein

DET-type bioelectrocatalysis is sometimes achieved by introducing an electrode-active domain into a native enzyme. Cellobiose dehydrogenase (CDH) is often used as a model of DET-type fusion enzymes. CDH has two domains: a larger catalytic dehydrogenase domain and a smaller electrode-active cytochrome domain. The domains are linked by a flexible polypeptide [113]. The fused cytochrome domain mediates the electron transfer from the catalytic domain to electrodes, such as a “built-in mediator.” Utilizing the fusion protein-engineering methods, DET-type reactions by FAD-GDH [114,115], pyrroloquinoline quinone (PQQ)-dependent GDH [116], and flavodoxin [117] were reported.

3. Novel Bioelectrochemical System

Major applications in enzymatic bioelectrocatalysis are biosensors and biofuel cells. Biosensors are utilized as analytical devices for food analyses and clinical diagnoses [6,7,8,9,10,11,12,13,14], and biofuel cells are ecofriendly energy conversion devices in which the chemical energy of fuels is electrically extracted [11,15,16,17,18,19,20,21]. The characteristics, properties, and progress of these devices were previously summarized in the corresponding reviews cited above. Recently, in addition, novel methods for bioelectrocatalytic applications are reported as described below.

3.1. Biosupercapacitor

Biosupercapacitors are self-powered energy storage devices using bioelectrocatalysis to charge capacitors [22]. In the biosupercapacitor, electric power is generated by a biocathode and/or a bioanode. The fundamental difference between biosupercapacitors and biofuel cells is the separation of the output of the current and bioelectrocatalytic reaction. Expressed as an equivalent circuit, a biosupercapacitor is a biofuel cell and a capacitor connected in parallel. Regardless of the connection to the external circuit, the bioelectrocatalytic reaction injects the electrons into the supercapacitor. When the external circuit requires the current, the biosupercapacitor provides the current. An advantage of biosupercapacitors is that a large current can flow even if the kinetics of the bioelectrocatalysis is inadequate. Additionally, biosupercapacitors are compatible with the self-powered biosensors [118,119].

The supercapacitor is classified into two types: the electrical double layer capacitor (EDLC) and the electrochemical pseudocapacitor (EPC). EDLC is the electrode with a large surface area, the electric charge is accumulated in the electrical double layer at the interface between the electrode and the electrolyte solution. EPC is constructed with the electrode and the reversible electrode-active redox species. The externally applied voltage shifts the electrode potential at the anode and a cathode in the negative and positive directions, respectively, as expressed in the Nernst equation.

Generally, redox-active species are immobilized at the electrode surface. Since an electrochemical capacitor contains two electrodes and an electrolyte solution, electrochemical capacitors are also classified into three types: EDLC-EDLC, EPC-EPC, and EDLC-EPC [120]. Therefore, biosupercapacitors are basically classified into the three types. The EDLC is compatible with DET-type bioelectrocatalysis and MET-type bioelectrocatalysis is adapted to the EPC. In order to avoid the mixing of mediators, EPCs in the biosupercapacitor are frequently employed redox polymers as mediators to immobilize at the electrode surface.

Electrochemically inactive porous electrodes in an electrolyte solution work as the EDLC. Additionally, DET-type bioelectrocatalytic activity is required for the EDLC-type biosupercapacitor. The reported materials suitable for the EDLC and the bioelectrode are carbon nanotubes [121,122,123], porous gold [124], gold nanoparticles [125], and indium tin oxide nanoparticles [126]. Porous electrodes often show high activity for DET-type bioelectrocatalysis. Therefore, it is considered that the investigation of bioelectrodes for DET-type biofuel cells will be directly useful for the biosupercapacitor.

The separation of oxidant and reductant is important in EPCs, since the mixing of the two species causes the cross-reaction discharge. Therefore, in the case of biosupercapacitors using the MET-type bioelectrocatalysis, careful attention is required to head off the outflow of mediators. As mentioned above, redox polymers are employed as mediators in biosupercapacitors [42,127,128,129]. The immobilization of redox-active proteins at the electrode surface has been investigated for the improvement of the capacitance of EPC-type biosupercapacitors [130,131].

3.2. Bioelectrosynthesis

Bioelectrosynthesis is a coined term for the generation of renewable energy resources and useful organic compounds [132]. Particularly, stereoselective enzymatic bioelectrocatalysis is desirable for pharmaceutical use because pharmaceutical precursors are required to have high enantiomeric purity [133,134].

Some useful compounds are reduced forms of redox couples, and reductive conversion is often the opposite reaction in metabolisms in vivo. In bioelectrocatalysis, on the other hand, several enzymes can proceed bidirectional electroenzymatic reactions (both the oxidation and reduction of substrate redox couples), in which the driving forces are reversed depending on pH. For example, H₂ase [135,136], formate dehydrogenase (FoDH) [75,136,137,138,139], carbon monoxide dehydrogenase (CODH) [136], diaphorase (DI) [136,140], ferredoxine-NADP⁺ reductase (FNR) [28], and some NAD(P)⁺-dependent dehydrogenases [28,140,141,142,143] show bidirectional bioelectrocatalytic activities. The characteristic of these enzymes is that the formal potentials of the substrate redox couples and the cofactors in the enzymes are relatively close to each other, and thus, the small uphill intramolecular electron transfer proceeds in acceptable velocity [34]. Bidirectional bioelectrocatalysis is a key reaction for constructing bioelectrochemical energy/compound conversion systems.

There are mainly two types of bioelectrosynthetic systems: a fuel-cell-type [142,144,145,146,147,148,149] and an electrolysis-type [28,53,141,143,150,151,152,153,154]. The former realizes a spontaneous production of compounds without any external power supplies, whereas the latter proceeds relatively rapid reactions due to an optimally controlled electrode potential to realize diffusion-controlled conditions. We will show examples of bioelectrosyntheses in the following sections.

3.2.1. Dihydrogen Production

Dihydrogen (H₂) is a typical clean energy source that emits no harmful products in combustion. H₂ase, a common anodic bioelectrocatalyst for H₂ oxidation, can also catalyze proton (H⁺) reduction [135,136]. The reversible reaction is biased to favor H₂ oxidation, and it is difficult to realize a large cathodic current density of H⁺ reduction in DET-type systems. H⁺ reduction with a large current density was achieved in a H₂ase MET-type system using a viologen polymer as a redox mediator [53]. On the other hand, a dual DET-type bioelectrocatalytic water–gas shift reaction using CODH and H₂ase (H₂O + CO → H₂ + CO₂) was reported [146]. The system requires no external power supplies.

3.2.2. Formate Production

Formate/formic acid (HCOO⁻/HCOOH) is a liquid energy resource with a high energy density due to its property of being completely oxidized to CO₂. FoDH interconverts HCOO⁻ and CO₂ both in DET- and MET-type systems [75,136,137,138,139]. Sakai et al. reported efficient bioelectrocatalytic CO₂ reduction using FoDH from Methylobacterium extorquens AM1 and a synthesized redox mediator with the low formal potential on a gas-diffusion electrode [150]. In addition, spontaneous interconversion between HCOO⁻ and H₂ without any external power supplies was achieved by coupling two bioelectrocatalyses by FoDH and H₂ase (H₂ + CO₂ → HCOO⁻ + H⁺) [147]. The direction of the reaction is controlled by pH- and concentration-dependent equilibrium potentials of the 2H⁺/H₂ and CO₂/HCOO⁻ couples.

3.2.3. Ammonia Production

Ammonia (NH₃) is an industrially beneficial compound used as a chemical raw material, fuel, hydrogen storage, and so on [155,156]. Nitrogenase (N₂ase) is an enzyme that catalyzes a nitrogen fixation from dinitrogen (N₂) to NH₃ using adenosine triphosphate (ATP) hydrolysis energy in microorganisms under an anaerobic condition [155,156]. Particularly, molybdenum-dependent N₂ase, which comprises a catalytic MoFe protein and a homodimeric MgATP-binding Fe protein, has been mainly investigated in view of bioelectrocatalysis [156,157]. N₂ase reduces not only N₂ but also nitrite (NO₂⁻) and azide (N₃⁻) to NH₃ on electrodes, and the MoFe protein disassociated with the Fe protein shows the DET-type bioelectrocatalytic activity without ATP [154]. Furthermore, ATP-independent NH₃ bioelectrosynthesis was improved in the MET-type system using N₂ase and a cobaltocene-functionalized polymeric mediator [51]. On the other hand, Milton et al. reported a NH₃-producing H₂/N₂ biofuel cell using MET-type reactions of N₂ase and H₂ase [148]. If the fatal weakness of N₂ase, a lack of oxygen tolerance, is improved, N₂ase will be expected to have further industrial applications.

3.2.4. NAD(P)⁺-Dependent Bioelectrosynthesis

Nicotinamide coenzymes (NAD(P)⁺/NAD(P)H) play essential roles in the function of many NAD(P)⁺-dependent dehydrogenases, but the dehydrogenases need suitable redox mediators in single-step enzymatic bioelectrocatalysis because the NAD(P)⁺/NAD(P)H interconversion requires high overpotentials at conventional electrodes due to its hydride ion-transfer-type characteristics that are completely different from two-step single-electron transfer-type electrode reactions. On the other hand, the introduction of dipaphorase (DI) and FNR, which are bioelectrocatalysts for the interconversion of the NAD⁺/NADH and NADP⁺/NADPH redox couples, respectively, can realize various NAD(P)⁺-dependent bioelectrocatalyses with relatively low overpotentials [28,140].

The bienzymatic MET-type system of NAD⁺-dependent dehydrogenase and DI is kinetically and thermodynamically investigated and often applied to biosensors [40,140,158,159]. In addition, the coulometric electrooxidation of organic substances was achieved in the same multienzymatic system mimicking the tricarboxylic acid cycle [151]. On the other hand, reductive production of chiral compounds is useful in many cases. Minteer’s group reported various NAD⁺-dependent multienzymatic cascade reactions for generating chiral compounds, as shown in Figure 5 [149,152,153]. They also demonstrated the bioelectrosynthesis of polyhydroxybutarate with NADH regeneration by DI [143].

In an NADP⁺-dependent system, on the other hand, FNR showed DET-type bioelectrocatalytic activity [28]. Particularly, Armstrong’s group investigated NADP⁺-dependent bienzymatic bioelectrocatalysis using FNR. They reported the reductive carboxylation of pyruvate to malate by malate dehydrogenase [142], reductive amination of 2-oxoglutamate to l-glutamate by glutamate dehydrogenase [28,144], and enantioselective interconversion of the secondary alcohol/ketone couple by engineered alcohol dehydrogenase [141,145].

3.3. Photo-Bioelectrocatalysis

Photo-bioelectrocatalysis enables reducing a substrate with a lower formal potential by oxidizing a sacrificial reagent with a higher formal potential, which cannot spontaneously occur without solar energy. Electrons that photosensitizers accept from sacrificial reagents are excited by solar energy, and the electric potential is shifted in the negative direction, corresponding to the wavelength of the adsorbed light. The electrons are then donated to substrates via enzymes and mediators, as shown in Figure 6. Photosensitizers such as TiO₂ [80,81,160,161,162], PbS quantum dots [162], silver nanoclusters [80,81], and organic dyes [160,161] are incorporated in anodes of transparent electrode bases (ITO in general) with or without other catalysts. Particularly, in addition, the photosystem II (PSII) complex in the thylakoid membrane of cyanobacteria and higher plants is often used as a water-splitting anodic photo-bioelectrocatalyst [129,160,161,162,163,164,165,166,167,168,169,170,171,172,173,174,175,176,177,178,179,180,181]. Biosolar cells [163,164,165,166,167,168,169,170,171,172,173] and solar biosupercapacitors [129,164,168,174], using PSII/I, thylakoid membranes, or cyanobacteria in anodes, and BOD or laccase in cathodes, realized the conversion from solar to electric energy without any sacrificial reagents in total. On the other hand, photo-bioelectrosyntheses, also called artificial photosyntheses, are reported. H₂ is generated by H₂ase [80,160], and CO₂ is fixed by FoDH and CODH [81,161]. These cathodic reactions proceed at quite low potentials. Thus, in order to improve the Faradaic efficiency of these photo-bioelectrosyntheses, it is essential to reduce electron leakage to dissolved oxygen as much as possible, especially when oxygen-generating photosynthetic proteins or organisms are used as anodic photo-bioelectrocatalysts.

4. Conclusions

Enzymatic bioelectrocatalysis is advancing day by day due to the finding of novel enzymes and the development of electrode nanomaterials. However, there are still issues to be solved such as thermostability, oxygen tolerance, and physical fragility, compared to inorganic electrocatalysts. Further understanding and consideration of the electrode interface and the interaction between enzymes and electrodes from kinetics and thermodynamics are required. Other perspectives, such as the improvement of redox enzymes by protein-engineering approaches, the selection of electrode materials and mediators (including redox hydrogels), the immobilization of enzymes, and the layout of electrodes will be also required for the improvement of present bioelectrochemical devices.