Chapter 1: Communication with the Mononuclear Molybdoenzymes: Emerging Opportunities and Applications in Redox Enzyme Biosensors

The mononuclear Mo-enzymes are remarkable in their coherence of active site structure and function yet equally interesting in the diversity of substrates that they are capable of oxidising or reducing. Apart from the well-studied enzyme nitrogenase, where the Mo ion is found within a S-bridged cluster of metals including Fe, all other enzymes containing Mo bear a single metal at the active site.¹  A recent exception to this may be the novel Mo enzyme CO dehydrogenase, where a Cu ion shares a sulfide bridging ligand with the Mo ion at the active site.

All known enzymes from this family bear either one or two bidentate pterin-dithiolene (molybdopterin, MPT) ligands bound to the Mo ion at the active site (Figure 1.1). Hille proposed²  a classification of this group of enzymes into three families based on the coordination environment of the metal as shown in Figure 1.1.

At this time, enzymes from the DMSO reductase family (the most diverse of all) have only been found in bacteria and archea whilst enzymes from the other two families are found in all forms of life.

Although subtle differences exist in the mechanism of the mononuclear Mo enzymes, as a starting point, the reactions catalysed by this enzyme superfamily can be generalised by eqn (1) written in either the forward (reductase) or reverse (oxidase/dehydrogenase) direction. The substrates represented by the generic symbols Z and ZO are apparent from the names of the respective oxidases/dehydrogenases (Z) or reductases (ZO) shown in Figure 1.1.

The use of Mo enzymes in electrochemically driven (amperometric) biosensors relies on connecting a working electrode with the redox active species involved in the catalytic cycle i.e. the enzyme and/or its substrates and products. The electrons required to sustain catalysis are provided or accepted by the electrode rather than the enzyme's natural cosubstrate. The various ways in which this can be done are summarised in the following section, but the most relevant point is that the Mo ion at the active site always cycles between its Mo^VI and Mo^IV oxidation states during catalysis and an O-donor ligand (oxo or hydroxo) is exchanged with the substrate during turnover. The Mo^VI form is the catalytically active form of the oxidases/dehydrogenases whilst the reductases must be reduced to Mo^IV before turnover can commence. The Mo^V form is a thermodynamically stable intermediate in most, but not all, cases, but it is incapable of turning over substrates in either direction. However, it may become an important rate-limiting intermediate in some cases.

Ligands bonded to the Mo ion are activated by coordination to perform some remarkable bond-breaking and formation reactions that otherwise do not occur in the absence of the enzyme. In the well-studied xanthine oxidoreductases, a hydroxo ligand participates in a base-assisted nucleophilic attack at C-8 of xanthine coupled with a hydride abstraction by the sulfido ligand (Figure 1.2A).³  This mechanism is significantly different from that seen in enzymes from the sulfite oxidase (Figure 1.2B) and DMSO reductase (Figure 1.2C) families where an oxo ligand is exchanged directly between the Mo ion and substrate during turnover.

The development of enzyme-based biosensors in general has evolved over recent times as methods for addressing the active sites of enzymes have become better understood. Initially, enzyme electrochemistry relied upon the voltammetric detection of either the product or cosubstrate (so-called first-generation biosensors, Figure 1.3). The most common analyte that has been detected in this way is hydrogen peroxide, a typical product of oxidase enzyme turnover where the cosubstrate dioxygen is reduced in a two-electron proton-coupled reaction by the enzyme after substrate turnover. Alternatively the product itself may be electroactive but this is exceptional. A complementary approach is to monitor the depletion of cosubstrate, e.g. dioxygen, during turnover. However, this approach has limitations as variations in dioxygen concentrations may result from changes to the solution during analysis, e.g. temperature, stirring etc., and thus give false readings.

Second generation biosensors removed the natural cosubstrate from the system altogether, replacing it with a small molecule electron transfer mediator e.g. a redox active coordination compound or organic molecule. This approach has its roots in traditional enzyme solution assays where a mediator is oxidised or reduced chemically rather than electrochemically to drive the catalytic reaction. In electrochemistry, the mediator serves the dual purpose of (i) undergoing homogeneous electron transfer with the enzyme to restore it to its active form following turnover and (ii) undergoing heterogeneous electron transfer with the working electrode to provide the current that quantifies the enzyme-substrate reaction. This approach underpins most commercial enzyme biosensors to date including the glucose oxidase biosensor,⁴  which utilised a ferrocenium mediator (rather than dioxygen) as its artificial cosubstrate. The use of redox active polymers adsorbed on the electrode also comes under this classification. In all cases it should be emphasised that the currents observed are due to the mediator and they appear at the formal potential of the mediator and not of the enzyme. Ideally the redox potential of the mediator is in the vicinity of that of the active site. This avoids excessively large overpotentials which may lead to non-specific redox reactions with species in the sample other than the substrate.

The final approach (third generation) is to remove all cosubstrates from the system (natural or artificial) and to achieve direct electron exchange between the enzyme and the electrode. Although this has yet to be applied in a commercial biosensor, it offers significant advantages over (ternary) mediated systems, which each require a certain artificial electron relay tailor-made for the enzyme in question. The challenges in achieving direct enzyme electrochemistry are many.^5–7  These include avoiding enzyme denaturation of the necessarily electrode-adsorbed enzyme whilst ensuring efficient electronic communication between the active site (or other redox cofactors within the enzyme) and the electrode. However, significant progress has been made in the last 15 years or so and a number of robust enzyme electrode systems have been reported which require no mediators. In the absence of mediators, which mask electron transfer events with the enzyme, some interesting mechanistic studies have been possible which provide new insight to electron transfer pathways in complex enzyme systems.

The following section will review the evolution of Mo enzyme based electrochemical biosensors over recent times with examples of all three types of biosensor being covered. The ordering of sections follows that of the three enzyme sub-families.

Xanthine oxidase from bovine milk is the most studied of all mononuclear Mo enzymes. The volume of literature on this enzyme alone and its employment in enzyme electrode biosensors far outweighs all other Mo enzymes accordingly. Xanthine oxidase primarily catalyses the oxidation of xanthine to uric acid as part of the process of purine metabolism, but it also is capable of oxidising hypoxanthine to xanthine (Figure 1.4).

Regardless of their origin, all xanthine oxidase/dehydrogenase enzymes contain an active site comprising a single bidentate molybdopterin chelate, an equatorial terminal sulfido, an axial oxo and an equatorial hydroxo/aqua ligand depending on pH (Figure 1.1). The enzymes from this family bear three additional redox cofactors comprising two [2Fe-2S] clusters and an FAD cofactor. Crystal structures of various xanthine oxidoreductases^8,9  have shown that electrons flow along the pathway Mo→[2Fe-2S]→[2Fe-2S]→FAD. The FAD cofactor is oxidised by either NADP⁺ or dioxygen depending on whether it is present in its dehydrogenase or oxidase form.

Like many other oxidase enzymes, hydrogen peroxide is a product of substrate turnover in xanthine oxidase when dioxygen is the cosubstrate. The electroactivity of H₂O₂ enables its voltammetric detection and provides a method for monitoring turnover without requiring direct or mediated electron transfer with the enzyme itself. A wide variety of electrode systems have been described that utilise immobilised xanthine oxidase to produce H₂O₂ as an electrochemically detectable product or alternatively to monitor the depletion of cosubstrate dioxygen.

Hypoxanthine in particular is an analyte of interest as its presence is an indicator of spoilage in otherwise fresh fish^10,11  and beef products.¹²  Sol-gel methods were used to produce an electrode coated with a silica-graphite matrix in which xanthine oxidase is entrapped.¹³  Following substrate (hypoxanthine) turnover, the H₂O₂ produced is detected electrochemically by poising the electrode at 0.58 V vs. Ag/AgCl. Alternatively the complementary consumption of dioxygen may be determined at low potential voltammetrically (mediated by viologens) or with an oxygen electrode. In H₂O₂ production mode the sensor maintained a linear current/concentration response up to 500 μM hypoxanthine. Above this substrate concentration the response saturated following Michaelis–Menten kinetics (K_M,app 450 μM). A lower detection limit of 1.3 μM hypoxanthine was reported.¹³  The high surface area and biocompatibility of the silica matrix was found to be ideal for encapsulating the enzyme/graphite composite. Furthermore, the mild synthetic sol-gel techniques enabled the enzyme to be incorporated during the synthesis of the matrix.

Screen-printed xanthine oxidase electrodes have also been reported using a vast array of mediators including metal oxides (RuO₂, Pd-IrO₂) within the matrix to lower the overpotential for H₂O₂ oxidation.^14,15  Even more elaborate bi-enzyme systems have been developed comprising both xanthine oxidase and peroxidase (an enzyme that reduces H₂O₂ to water) where the H₂O₂ produced by xanthine oxidase turnover is quantified by the current produced through its ferrocyanide-mediated reduction by peroxidase¹⁶  thus enabling the bi-enzyme system to function at a low potential (ca. −100 mV vs. Ag/AgCl) and minimising possible interferences from oxidation of other analytes at higher potentials.

A miniaturised xanthine oxidase electrode has been developed for monitoring the concentrations of hypoxanthine in myocardial cell culture media.¹⁷  Purines are also associated with signalling in the nervous system and multienzyme electrodes (including xanthine oxidase) have been developed to monitor the local changes in purine concentrations in vivo.¹⁸  The enzyme purine nucleoside phosphorylase (NP) catalyses the phosphorylation of inosine (by phosphate) to release hypoxanthine and ribose-1-phosphate (Scheme 1.1). The stoichiometry of the overall reaction coupled with xanthine oxidase activity means that one equivalent of H₂O₂ is produced for every (hydrogen)phosphate anion present.¹⁹ 

A number of groups have developed amperometric bi-enzyme systems of this type.^20–22  Haemmerli et al.²¹  reported a bi-enzyme (NP/xanthine oxidase) system which exhibited a linear response up to 250 μM phosphate. Various ratios of the two enzymes were investigated and the most ideal was found to be 10 : 1 NP : xanthine oxidase. This novel approach provides a viable alternative to otherwise tedious wet chemical (colorimetric) methods for phosphate determination.

Purine analysis can also be achieved with this system. For example, the concentration of inosine (the cosubstrate, with phosphate, in eqn (2a)) can be determined.^23,24  The so-called hypoxanthine ratio ([hypoxanthine]/([hypoxanthine]+[inosine]+[inosine-monophosphate])²⁵  is an important parameter in the analysis of fresh fish as hypoxanthine is a product of nucleotide degradation. Like hydrogen peroxide, uric acid (the product of enzymatic xanthine oxidation, Figure 1.4) is electroactive and it can also be detected electrochemically thus providing a method for quantifying xanthine oxidase turnover.²⁶ 

Electron transfer mediators can be used to great effect in providing a link between the electrode and the enzyme cofactors. There are many approaches that may be taken. Conducting polymers such as poly-p-benzoquinone and poly(mercapto-p-benzoquinone) have been used to provide a redox active matrix in which xanthine oxidase can be both immobilised and addressed electrochemically.²⁷  The well-studied Os-pyridine-based hydrogel polymers developed by the Heller group²⁸  effectively mediate electron transfer in horseradish peroxidase (HRP). This polymer has been cast on a glassy carbon electrode and coupled with xanthine oxidase which enables the production of H₂O₂ from (hypo)xanthine turnover to be monitored electrochemically via the Os-mediated reduction of HRP within the polymeric hydrogel.²⁹  This is illustrated in Scheme 1.2.

The benefit of using Os-mediated reduction of HRP is that the biosensor functions at a low potential (ca. 0 mV vs. Ag/AgCl) relative to that required for the direct oxidation of H₂O₂ (> 600 mV), where other species may be oxidised non-specifically as well. The sensor responds to hypoxanthine in a continuous flow system (linear response up to 80 μM with a detection limit of 0.2 μM) and also xanthine but is unaffected by potential interference from glutamate, lactate, glucose and glutathione. Ascorbate interference (significant at a working potential of 0 mV) was negligible when the working electrode was poised at −200 mV vs. Ag/AgCl. Otherwise redox inert polymers such as Nafion® can be made electroactive by the inclusion of small cationic mediators such as methyl viologen, which ion-exchanges within the anionic polymer. Xanthine oxidase adsorbed on such a viologen-modified polymer (cast on a glassy carbon electrode) produces a hypoxanthine biosensor³⁰  that functions in O₂-consumption mode. In this case the viologen mediates the reduction of O₂ and lowers the overpotential for reduction by about 100 mV relative to direct oxygen reduction at the electrode.

High potential redox mediators, such as ferrocenium derivatives, have been very effective in acting as artificial electron acceptors for a range of oxidase enzymes; glucose oxidase being the most famous example. Similarly, hydroxymethylferrocenium can accept electrons from xanthine oxidase (replacing dioxygen) to produce a mediated amperometric hypoxanthine biosensor. In this case the enzyme was entrapped within a membrane covering the carbon-paste working electrode.³¹  A wide linear range (up to 700 μM hypoxanthine with a detection limit to 0.6 μM) was reported.

Ironically, despite its intensive investigation by enzymologists and spectroscopists for almost 50 years, it is only recently that a direct electrochemical study of a xanthine oxidoreductase was reported, namely the bacterial xanthine dehydrogenase from R. capsulatus.³²  Non-turnover signals from all cofactors were seen and EPR potentiometry was also undertaken to resolve the potentials of the cofactors. Pronounced catalytic voltammetry was seen in the presence of xanthine. The bell-shaped pH profile of the catalytic wave mirrored that seen in solution. A high pH deprotonation of xanthine (pK_a 7.7) and a low potential protonation of a glutamate residue (pK_a 6) essential for base catalysis each switch off catalysis. An unusual feature of this study was that the potential at which catalysis was observed was ca. 600 mV more positive than that of the highest potential cofactor (FAD). Essentially the same potential “delay” in catalysis has been seen for bovine milk xanthine oxidase when immobilised on a pyrolytic graphite electrode.³³ 

A recent addition to the Mo enzyme family is the novel CO dehydrogenase from the bacterium Oligotropha carboxidovorans, which catalyses the oxidation of CO to CO₂.³⁴  Its crystal structure first appeared to reveal an unusual Mo-S-Se group at the active site³⁵  but this interpretation was later revisited. Coupled with spectroscopic measurements, an equally novel Mo-S-Cu moiety,³⁶  was identified instead of the terminal Mo=S group typically found in enzymes from the xanthine oxidase family (Figure 1.1). Given its dinuclear active site CO dehydrogenase, strictly speaking, does not belong in the mononuclear enzyme family at all. Notwithstanding, CO dehydrogenase possesses a single molybdopterin ligand bound to the Mo ion and two [2Fe-2S] clusters in addition to an FAD cofactor like all other members of the xanthine oxidase family and on this basis it appears to belong within this grouping.

The importance of quantitatively detecting CO, a potentially lethal gas, in the home and in the field is a significant driving force for the development of a CO dehydrogenase biosensor. Very little electrochemical work has been performed on this enzyme. The most significant appeared more than 25 years ago, well before its novel structure was known, when ferricyanide and ferrocenium-mediated catalytic electrochemistry of CO dehydrogenase was reported.³⁷  More studies with this remarkable addition to the Mo enzyme family are eagerly awaited.

The sulfite oxidases and dehydrogenases dominate this Mo-enzyme family;²  a plant nitrate reductase being the only other member of this sub-grouping that has been characterised to date.³⁸  Sulfite oxidoreductases are found in most forms of life and there now exist structurally characterised examples from animals,^39,40  plants⁴¹  and most recently bacteria.⁴²  In humans they catalyse the last stage in the degradation of S-containing amino acids where sulfite is oxidised to sulfate. The eukaryotic sulfite oxidases can donate electrons to dioxygen (eqn (2a)) or cytochrome c (eqn (2b)) whilst the dehydrogenases only use cytochrome c as their electron acceptor.

In animal and bacterial sulfite oxidoreductases, the Mo active site is coupled with a heme cofactor that accepts electrons in sequence from the Mo^IV ion after turnover and passes them on to a c-type cytochrome in solution. In plants it appears that the heme cofactor is absent⁴¹  and the Mo ion donates electrons directly to dioxygen.

The first crystal structure of a sulfite oxidoreductase to appear was that of the enzyme from chicken liver.³⁹  This structure presented a mechanistic problem in that the heme and Mo cofactors (supposedly electron partners) were separated by 33 Å; well in excess of the distance that electrons are known to tunnel between redox centres.⁴³  Subsequently it has emerged from spectroscopic studies of electron transfer^44,45  that the crystallographically characterised conformation of chicken liver sulfite oxidase is not that of the active enzyme. Instead, a polypeptide “hinge” enables the heme cofactor to “swing” around from this remote location to be in proximity with the Mo active site such that intramolecular electron transfer is then possible. The initial uncertainty surrounding the active conformation of chicken liver sulfite oxidase has been resolved by the crystal structure of a bacterial sulfite dehydrogenase⁴²  where the Mo and heme cofactors are adjacent as expected (Figure 1.5). This structure serves as a good model for the chicken liver and human enzyme in their active conformations, which so far have eluded identification by X-ray crystallography.^39,40 

The substrate sulfite is of great importance to the food and beverage industry as it is used extensively as a preservative due to its antioxidant and antimicrobial properties.⁴⁶  The concentration of added sulfites is regulated as allergies to this chemical in some people can present a serious health risk.⁴⁷  However, the analytical determination of sulfite using wet chemical methods is not a simple procedure and the development of alternative procedures that can directly determine sulfite in foods and beverages is an important goal. Although the direct electrochemical oxidation of sulfite to sulfate can in principle be performed, the high potential that is required becomes problematic due to the simultaneous oxidation of species such as ascorbate and polyphenols, which are also often present in food and beverage samples. The development of enzyme-based (sulfite oxidase) biosensors offers a great improvement in selectivity.

The commercial availability of chicken liver sulfite oxidase has led to its dominance of the amperometric sulfite biosensor literature. In principle, the oxidase activity of the enzyme (in producing electroactive H₂O₂) can be exploited to develop a first generation biosensor where hydrogen peroxide is detected amperometrically. However, a major problem to be overcome with this approach is that H₂O₂ and sulfite are each oxidised at similar potentials on bare electrodes (ca. 400 mV vs. Ag/AgCl at neutral pH). The use of electropolymerised polytyramine to entrap sulfite oxidase has proven a successful solution to this problem and a first-generation sulfite biosensor that detects H₂O₂ produced as a result of enzyme turnover (eqn (2a)) has been reported.⁴⁸  Direct oxidation of sulfite (Figure 1.6, curve a) is seen on a bare GC electrode but this response is completely suppressed when the polytyramine film is cast on the same electrode (Figure 1.6, curve b). Introduction of sulfite oxidase and sulfite (Figure 1.6, curve d) generates a pronounced wave due to electrochemical H₂O₂ oxidation produced from enzyme turnover. This biosensor gave a linear response up to 300 μM sulfite.

An Hg-film coated glassy carbon electrode was used in oxygen depletion mode to monitor sulfite oxidase activity.⁴⁹  The Hg film significantly lowered the overpotential for O₂ reduction at the electrode. Entrapment of sulfite oxidase within electropolymerised polypyrrole films has also been reported.⁵⁰  Additives such as dextran⁵¹  were shown to improve mechanical stability of the enzyme-polymer film.

The problems associated with accurately measuring oxygen consumption or detecting H₂O₂ amperometrically in the presence of sulfite are avoided if mediators are used to accept electrons from sulfite oxidase instead of dioxygen. The electron transfer protein cytochrome c is an ideal choice (Scheme 1.3) as it is known to be a physiological electron partner of sulfite oxidase and its direct electrochemistry at a number of different electrode surfaces has been intensively studied. Under anaerobic conditions, no H₂O₂ is produced from enzyme turnover and the anodic current is merely that of the cytochrome.

A comprehensive study by Ferapontova and coworkers⁵²  showed that horse heart cytochrome c at a 1 : 1 mercaptoundecanoic acid and mercaptoundecanol modified Au electrode is capable of mediating sulfite oxidase electron transfer (Figure 1.7). The electrochemistry of an almost equimolar mixture of cytochrome c and sulfite oxidase (Figure 1.7, curves 1 and 3 in inset) is characteristic of diffusion-controlled voltammetry of the cytochrome, indicating only weak association with the thiol-modified Au electrode. Loss of the proteins from the electrode surface into the bulk was prevented by their entrapment beneath a dialysis membrane ensuring all proteins remained in proximity to the electrode. In the presence of sulfite the reversible peak-shaped response of the cytochrome transforms into a sigmoidal wave (curve 2) where the amplification of current is due to recycling of reduced cytochrome c, which accepts electrons from sulfite oxidase following turnover. A very similar study using human sulfite oxidase has appeared more recently and utilising the same thiol-modified Au procedure as Ferapontova et al. and with cytochrome c acting as a mediator.⁵³ 

Screen-printed carbon electrode composites containing both sulfite oxidase and cytochrome c have been reported for the purpose of analysing gaseous SO₂ (dissolved in water).⁵⁴  The sensor operated at 300 mV vs. Ag/AgCl and exhibited a linear response in the range 4–50 ppm sulfite.

Artificial mediators have also been used to accept electrons from sulfite oxidase during catalysis. Murray and coworkers reported a series of papers where a number of different high-potential mediators were used.^55–58  Given the nature of the experiment, the intrinsic catalytic constants were hidden by the use of mediators, but information regarding the relative rates of bimolecular (mediator-enzyme) electron transfer, modelled with Marcus theory,⁵⁶  and intramolecular (Mo-heme) electron transfer⁵⁸  were elucidated. Interestingly when the heme domain was cleaved from the enzyme, the mediator was unable to oxidise the reduced Mo active site thus indicating that the heme moiety was the site at which intermolecular electron transfer took place.⁵⁶ 

Towards third-generation biosensors, the direct electrochemistry of sulfite oxidoreductases has been a more recent development. Direct electrochemistry of the bacterial sulfite dehydrogenase from S. novella adsorbed on an edge plane pyrolytic graphite electrode has been reported⁵⁹  where non-turnover responses from both the Mo and heme cofactors were identified. A sigmoidal catalytic oxidation wave was seen upon addition of sulfite to the cell (Figure 1.8) at a potential corresponding to the heme potential (and ca. 200 mV lower than that due to direct sulfite oxidation).

This represents an ideal example of direct enzyme electrochemistry where the potential at which a catalytic current emerges is that of the enzyme rather than of a mediator. The apparent Michaelis constant derived from the concentration dependence of the steady state limiting current (K_M,app = 26 μM) was similar to that obtained from solution studies. It is notable that the absence of mediators and the fact that the enzyme is adsorbed on an electrode significantly increases the sensitivity whilst narrowing the linear response of the enzyme electrode to around 10 μM sulfite. Elliot et al. reported direct electrochemistry of chicken liver sulfite oxidase adsorbed on a pyrolytic graphite electrode. The only non-turnover signal in the absence of substrate was assigned to the heme which was replaced by sigmoidal catalytic oxidation wave in the presence of sulfite.⁶⁰  Ferapontova et al. reported a similar activity with chicken sulfite oxidase but used self-assembled monolayers of long-chain alkanethiols bound to an Au working electrode.⁵²  Interestingly, the electrocatalytic activity was very sensitive to both the terminal group on the alkanethiol and the length of the chains comprising the self-assembled monolayer.

Unlike the xanthine oxidase and sulfite oxidase families, which contain very few members, the DMSO reductase family boasts an impressive array of enzymes that act upon a wide variety of inorganic and organic substrates. This family perhaps offers the most opportunities for the development of novel biosensors for challenging (and otherwise chemically inert) substrates. This is a rapidly developing field and new members of the DMSO reductase family are continuing to be identified each year.

The parent enzyme from this family has been isolated from three different bacteria (E. coli, R. capsulatus and R. sphaeroides). The E. coli DMSO reductase is a complex membrane-bound enzyme which bears five Fe-S clusters in addition to its Mo active site. It obtains its electrons from the quinone pool. The two Rhodobacter enzymes are periplasmic, very similar to each other and contain no other redox active cofactors. They are reduced by a multi-heme containing electron partner.

Regardless of these differences, all catalyse the reduction of DMSO to dimethyl sulfide (DMS) where the substrate is the terminal electron acceptor for anaerobic respiration. As well as being present in seawater (and an important component of the global sulfur cycle) DMSO is found in food and beverages. Its reduction to highly volatile DMS is responsible for the unpleasant odour of otherwise pure DMSO. The analytical determination of DMSO is complicated by its chemical inertness and its high miscibility with most solvents, making separation and isolation by extraction very difficult.

The highly selective reduction of DMSO by DMSO reductase offers a direct way of determining this compound analytically in the presence of other species. DMSO reductase may accept electrons from the reduced methyl viologen radical cation (MV^+.) in solution assays and this has been adapted for the electrochemically driven catalysis of the enzyme from R. sphaeroides (Scheme 1.4) in the presence of the viologen as the electrochemically detectable species. Apart from DMSO itself, enantioselective reduction of various chiral sulfoxides was demonstrated with very high enantiomeric excess.⁶¹  This has been further developed into a biosensor for DMSO in solution.⁶²  A linear current response was found up to 6 mM DMSO using methyl or benzyl viologen as the mediator. Figure 1.9 (left) shows the transformation of the first electron reduction of methyl viologen (MV^2+/+.) from a peak-shaped reversible response to a sigmoidal steady state wave where the dication is regenerated. The second wave at lower potential is due to the MV^+./0 couple.

The first direct (unmediated) electrochemical investigation of a mononuclear Mo enzyme comprised a voltammetric study of DMSO reductase from E. coli (DmsABC).⁶³  No response from any of the redox active cofactors was seen in the absence of substrate but, upon addition of DMSO, a cathodic catalytic wave emerged. Curiously, the catalytic voltammogram at pH 8.9 (Figure 1.10) is peak shaped instead of the ideally sigmoidal profile of a steady state electrochemical system more apparent at pH 7.0.

This behaviour was rationalised by a model whereby protonation steps at the active site accompanying reduction (Mo^VI → Mo^V → Mo^IV) become rate limiting in the catalytic cycle. Central to this hypothesis is the assumption that the Mo^V form undergoes protonation much more rapidly than Mo^IV. At low potentials (high driving force) and high pH the Mo^IV form is generated rapidly but catalysis is limited by a slow protonation reaction on the Mo ion. Conversely at intermediate potentials, the intermediate Mo^V form accumulates and undergoes rapid (non-rate-limiting) protonation before further reduction to the active Mo^IV state and the overall reaction proceeds more quickly. Non-physiological oxidation of PMe₃ to OPMe₃ was also seen.

The periplasmic DMSO reductase from Rhodobacter capsulatus (DorA) is structurally unrelated to membrane-bound DmsABC except that they are believed to share the same active site. This enzyme (and the highly homologous enzyme from R. sphaeroides) possesses no other cofactors and is thus an attractive choice for spectroscopists where competing optical and EPR signals from Fe-S clusters and heme moieties do not mask those of the Mo cofactor. In fact this simplicity enabled clear resolution, for the first time, of non-turnover Mo^VI/V and Mo^V/IV signals from a Mo enzyme active site (in the absence of substrate).⁶⁴  Optical spectroelectrochemistry was also employed to observe the enzyme in its fully oxidised and reduced forms. Upon addition of DMSO, a catalytic wave was seen.

Like DMSO reductase, nitrate reductase can be found in distinctly different forms; a soluble periplasmic form (Nap) and a membrane-bound form (Nar) to name but two. Crystal structures of both NapAB^65,66  and NarGH(I)^67–69  have appeared recently. The NarGHI system is a membrane-bound quinol nitrate oxidoreductase complex comprising a number of redox active cofactors in each subunit. The NarI sub-unit is membrane bound and contains two hemes which receive electrons from the quinol pool, but this sub-unit may be separated to leave a catalytically competent and soluble NarGH dimer. The NarG component contains the Mo cofactor and a recently identified Fe-S cluster, while the NarH sub-unit bears a number of Fe-S clusters. Butt and coworkers have reported studies on the soluble NarGH enzyme from Paracoccus pantotrophus^70,71  where catalytic reduction (NO₃⁻ → NO₂⁻) waves were seen in the presence of substrate. At low substrate concentrations, a peak appears in the catalytic wave. The authors suggest that rate-limiting substrate binding to the Mo^V form is more rapid than to Mo^IV (a similar argument to the protonation model proposed by Heffron et al. for DMSO reductase (Figure 1.10),⁶³  except that no substrate concentration-dependence of the catalytic waveform was seen in that case. However, given the number of other redox centres in the enzyme it is also possible that a different redox event is responsible for the attenuation in catalytic current at lower potentials. The NarGH enzyme from E. coli has more recently been investigated⁷²  and again complex behaviour is seen comprising low- and high-potential catalytic waves. At micromolar nitrate concentrations a peak is seen in the higher potential component of the catalytic wave, which is then dominated by a more intense lower potential wave at millimolar concentrations of substrate. The high-potential component is pH dependent while the lower component is not. The conclusions of this study again implicate the Mo^V form as a more effective Lewis acid in being able to bind nitrate more rapidly and more tightly than the Mo^IV form. It should be reiterated that the catalytic cycle cannot finish until the tetravalent oxidation state is reached but substrate binding emerges as a rate-limiting event in this system.

The periplasmic (Nap) nitrate reductases have received comparatively little attention from electrochemists. They are quite different from the NarGHI nitrate reductases both in the active site structure (a cysteine is coordinated to the Mo instead of an aspartate) and their cofactors. The NapA sub-unit contains the Mo active site and a [4Fe-4S] cluster while the NapB sub-unit has two heme cofactors. Frangioni et al. have communicated their results from the direct protein film voltammetry of NapAB from R. sphaeroides.⁷³  Once again, despite the significant differences in active site structure and cofactors, a peak in the ideally sigmoidal catalytic rotating disk voltammogram was seen across a wide range of nitrate concentrations. The authors refer to more complicated behaviour at high nitrate concentrations but no further analysis or data were given.⁷³ 

Reports of mediated nitrate reductase voltammetry appeared earlier than the above-mentioned direct electrochemistry. An early paper used a methyl-viologen-appended poly-pyrrole matrix within which nitrate reductase (E. coli) was entrapped.⁷⁴  The enzyme/monomer mixture was pre-adsorbed on the electrode to ensure a significant amount of enzyme remained within the poly-pyrrole matrix. A later study from the same group utilised a clay-modified electrode as a template for electropolymerisation which improved the current response.⁷⁵  The need for low-potential mediators was established in separate investigations with nitrate reductase enzymes from different organisms. Low-potential viologens and bromophenol dyes (red and blue) are particularly effective whilst higher-potential dyes (e.g. indigo sulfonates, with potentials closer to the Mo^VI/V and Mo^V/IV redox couples) are ineffective.^76,77 

Arsenic poisoning from anthropogenic sources is a major health problem in some parts of the world. The most toxic form is trivalent As (which exists naturally as arsenite (As(OH)₃). A number of bacteria can defend themselves from arsenite poisoning by using the Mo enzyme arsenite oxidase which oxidises arsenite to the less toxic arsenate (AsO₄³⁻). One of these (Rhizobium sp. str. NT-26) is unique in being able to draw energy from arsenite oxidation through its own arsenate oxidase. Direct electrochemistry of this enzyme adsorbed on a pyrolytic graphite electrode has been reported, where a catalytic current in the presence of arsenite is seen (Figure 1.11).⁷⁸  The pH optimum and kinetic constants mirrored those seen in solution and the enzyme exhibits stability on an electrode for periods of days. This study laid the foundation for the use of a carbon-nanotube-arsenite-oxidase modified electrode that was used to detect arsenite in a number of water samples.⁷⁹  A similar but distinct arsenite oxidase (from A. faecalis) is also electroactive and appears to be oxidised in a cooperative 2-electron reaction (Mo^IV to Mo^VI)⁸⁰  in the absence of substrate making it unusual compared with most Mo enzymes.

There are bacteria that are capable of respiring on oxoanions such as perchlorate (ClO₄⁻) or chlorate (ClO₃⁻). The overall picture is quite complex.^81,82  Some bacteria may reduce either anion for respiration by utilising distinctly different (but related) Mo enzymes perchlorate reductase⁸³  and chlorate reductase.⁸⁴  The analytical detection of perchlorate is difficult as the anion, although in a very high oxidation state and a potentially strong oxidant, is rather inert in solution. Immobilisation of perchlorate reductase in a Nafion® film cast on a glassy carbon electrode has recently enabled the construction of a functional perchlorate biosensor.⁸⁵  Although chlorate reductase electrochemistry has not been investigated to date, chlorate can be catalytically reduced by the Mo enzyme nitrate reductase (NarGH)⁷¹  in a non-physiological but important reaction.

Studies of mononuclear Mo enzymes received a tremendous boost with the publication of the first crystal structures of these enzymes a little more than 10 years ago,⁸⁶  which provided a clearer view of the organisation of electron transfer within these mostly complex metalloenzymes than was ever possible before. Since then the combined efforts of microbiologists, biochemists, geneticists, structural biologists, spectroscopists and, most recently, electrochemists have unearthed an ever-expanding family of enzymes whose substrates span a remarkably wide range of species, both organic and inorganic, yet who share essentially the reaction stoichiometry (eqn (1)). The observation of some unusual electrochemical behaviour, particularly the attenuation of activity at high overpotential, is remarkable but has now been observed in a number of Mo enzymes (mostly from the DMSO reductase family). Whether this non-classical electrochemical behaviour has a physiological significance or instead it is an experimental artefact is yet to be established.

New, as-yet biochemically uncharacterised, mononuclear molybdenum enzymes are now appearing in genomes of various organisms at an overwhelming rate and there is evidently much work to be done in elucidating the natural substrates of these enzymes and ultimately being able to express these proteins in sufficient quantities to enable full characterisation and then utilisation in novel biosensors.

Despite the many opportunities for determining the solution concentrations of substrates that present significant challenges for wet chemical methods, Mo enzymes biosensors are yet to reach commercial development, but this is a familiar tale. The phenomenal success of the glucose oxidase biosensor was driven by a need in the marketplace for such a device.⁸⁷  The technological issues for development of Mo-enzyme-based biosensors are no more challenging than those presented by the glucose biosensor but economics and market forces will decide which systems warrant intensive development into commercially viable biosensors.⁸⁸  At the moment the Mo-enzyme systems that probably warrant the most attention are those for the determination of sulfite (sulfite oxidase/dehydrogenase) and arsenite (arsenite oxidase) due to the potentially adverse health effects of each to humans and the difficulties surrounding their determination by wet chemical methods. This chapter has hopefully provided a perspective of the systems that are available for future development and an impetus for their further study.