Cellobiose dehydrogenase in biofuel cells

Abstract

Enzymatic biofuel cells utilize oxidoreductases as highly specific and highly active electrocatalysts to convert a fuel and an oxidant even in complex biological matrices like hydrolysates or physiological fluids into electric energy. The hemoflavoenzyme cellobiose dehydrogenase is investigated as a versatile bioelectrocatalyst for the anode reaction of biofuel cells, because it is robust, converts a range of different carbohydrates, and can transfer electrons to the anode by direct electron transfer or via redox mediators. The versatility of cellobiose dehydrogenase has led to the development of various electrode modifications to create biofuel cells and biosupercapacitors that are capable to power small electronic devices like biosensors and connect them wireless to a receiver.

Introduction

Fuel cells generate electrical energy by converting externally delivered fuel and oxidant into electric energy through electrocatalysts performing separated electrode reactions. Biofuel cells use either enzymes, cell organelles or microorganisms as biocatalysts to greatly accelerate anodic fuel oxidation and cathodic oxidant reduction [1]. In enzymatic biofuel cells enzymes are used as bioelectrocatalysts with their catalytic reaction coupled to the electrode either via redox mediators or by direct electron transfer (DET). The supply of fuel and oxidant for a biofuel cell can come from hydrolytic processes releasing mono- or disaccharides, hydrogen from anaerobic wastewater treatment as well as blood, sweat, or tear fluid for miniaturized, implantable or wearable biofuel cells powering biosensors or other devices [1,2]. Based on the electric discharge pattern and their capacity to store electrons they can be subdivided into conventional biofuel cells operating under constant discharge and charge-storing biofuel cells (or self-charging biosupercapacitors) which accumulate electrons on their electrodes until the power retrieved in a discharge.

Great advances have been made in the past decade in regard to the employed bioelectrocatalysts and the electrode materials. The benefit of an enzymatic biofuel cell is its operation at a low, ambient temperature because of its very efficient electrode catalysts, which are extremely well evolved to convert fuels of biological origin. They can therefore be employed in situations where regular fuel cells cannot be employed for reasons of fuel composition, low temperature, size, or biocompatibility. The prerequisites for enzymes used in biofuel cell anodes are an as low as possible redox potential to maximize the cell voltage, a reasonably high turnover number (>1 s^-1), and an electron transfer rate that does not limit the enzymatic activity. Another benefit are non-reactive reaction products. Dehydrogenases which have a very low side-reactivity with oxygen generate less deactivating H₂O₂ when not efficiently contacted to the electrode than oxidases.

Cellobiose dehydrogenase

Cellobiose dehydrogenase (CDH, EC 1.1.99.18, CAZy AA3_1) is an extracellular fungal hemoflavoenzyme involved in the degradation of lignocellulose. The physiological role of this auxiliary enzyme is to transfer electrons via its cytochrome domain to lytic polysaccharide monooxygenase (LPMO) to reduce its active site copper and thereby to initiate oxidative depolymerization of cellulose, hemicelluloses and other polysaccharides. CDH has been evolved by fungi to transfer electrons efficiently from its catalytically active dehydrogenase domain via its cytochrome domain to macromolecular electron acceptors. It also has been evolved to survive harsh extracellular conditions of fungal environments including the binding to macromolecular surfaces and even drying, and still maintaining its active conformation in low water activity, which are important prerequisites for immobilization processes on electrodes.

CDH catalyzes the two-electron oxidation of aldopyranoses at the free, anomeric C1 into aldonic acids via the corresponding lactones at its FAD cofactor in the dehydrogenase domain. The natural substrates are cellobiose or cello-oligosaccharides, but also activity on other mono-, di- and oligosaccharides has been reported (Figure 1). Being an auxiliary enzyme, the more important part of the reaction is connected to the obtained electrons at the FAD. Numerous electron acceptors such as quinones, redox dyes or chelated metal centers have been found to efficiently reoxidize the FAD with turnover numbers ranging from 1–50 s^-1, which is the basis of CDH’s excellent performance in mediated electron transfer-based electrodes. The most efficient electron acceptors/redox mediators are converted several hundred times faster than O₂, an important feature to reduce the amount of formed reactive oxygen species at the electrode. The most important electron acceptor, however, is CDH’s own N-terminal cytochrome domain. The interdomain electron transfer (IET) from the FAD to the heme b in the cytochrome domain occurs subsequently and depends on the close vicinity of both redox centers. The closest distance of ~8 Å between the edges of the cofactors is achieved when both domains form contact at their domain interface in the IET-competent closed state conformation [3]. Efficient IET depends on the complementary of the domain interfaces, favorable electrostatic interactions as well as on the optimal length of the linker region connecting the dehydrogenase and the cytochrome domain [4].

Figure 1. Kinetic constants and pH ranges of CDHs.

(a) Box plots (top) depicting ranges of Michaelis-Menten constants of basidiomycete Class I (blue) and ascomycete Class II (red) CDHs for carbohydrates (bottom). (b) pH ranges for the dehydrogenase domain assessed by the two-electron acceptor 2,6-dichloroindophenol (DCIP, top) and the cytochrome domain assessed by the one-electron acceptor cytochrome c of Class I (blue) and Class II (red) CDHs from: Tv, Trametes villosa; Tp, Trametes pubescens; Ps, Phanerochaete sordida; Pc, Phanerochaete chrysosporium; Gs, Gelatoporia subvermispora; Ar, Athelia rolfsii; Sb, Stachybotrys bisbyi; Nc, Neurospora crassa; Hi, Humicola insolens; Ds, Dichomera saubinetii; Ch, Crassicarpon hotsonii; Aa, Amesia athrobrunnea; Hh, Hypoxylon haematostroma; Ct, Crassicarpon thermophilum.

The interaction of CDH’s domains is the rate-limiting step of the electron transfer. Typically, basidiomycete Class I CDHs have a faster interdomain electron transfer (IET, 5–50 s^-1) than ascomycete Class II CDHs (0.1–10 s^-1). The interaction of both domains depends on the pH. It has been shown that the deprotonation of Asp and Glu residues, especially around the domain interface lead to electrostatic repulsion that prevents the closed state for most Class I CDHs above pH 6, but is less pronounced for Class II CDHs with efficient IET rates observed even around pH 7–8 (Figure 1).

Structurally, the flavodehydrogenase domain of CDH belongs to the glucose-methanol-choline (GMC)-oxidoreductase superfamily, which all share a common βαβ-mononucleotide binding motif (Rossmann fold) [5]. The substrate binding subdomain contains the substrate active site and a substrate channel, which also is part of the domain interface with the cytochrome domain. Both domains are connected via a flexible linker (Figure 2). Electrons from the FAD can be delivered via IET to the cytochrome domain and further on to an external sink such as the physiological electron acceptor LPMO or an artificial one – an anode. The cytochrome domain thus acts as a built-in redox mediator for CDH and enables DET to electrodes [6]. Using the multi-domain-architecture of CDH as a blueprint, engineered chimeric enzymes based on glucose dehydrogenases and cytochrome domains have been reported recently that are capable of DET [7–9]. A homologous cytochrome domain to that found in CDHs was found in fungal pyrroloquinoline quinone (PQQ)-dependent glucose dehydrogenases, which also allows DET for this this class of enzymes [10,11]. Additionally, this enzyme was also shown to use PQQ as redox mediator to shuttle electrons to an electrode even in the absence of its cytochrome domain [12,13].

Figure 2. CDH structure and electron transfer pathways.

The crystal structure of CDH (PDB: 4QI6, 4QI7) in its open state and closed state conformation are shown as ribbons. The flavodehydrogenase domain (DH) comprises a substrate binding (S, yellow) and an FAD binding (F, cyan) subdomain. The carbohydrate oxidizing isoalloxazine ring of the FAD molecule (orange spheres) and the substrate binding pocket are located in the interface of the S- and F-domain. The heme b carrying cytochrome domain (CYT) consists of two β-sheets (inner - red; outer - green) forming an ellipsoidal antiparallel β-sandwich. The inner sheet forms a concave pocket for the heme b (red spheres). After substrate oxidation at the FAD electrons are shuttled to the cytochrome domain of CDH in the closed state. In the open state direct electron transfer (DET) to electrode is possible. Alternatively, electrons can be transferred via mediated electron transfer (MET) directly from FAD via a suitable redox mediator or a redox polymer to the electrode.

Electrode materials and modifications

Direct electrochemistry or DET between the bioelectrocatalyst and an electrode depends on the distance between the redox center and the electrode and should not exceed 15 Å to still occur at a relevant rate. In the case of CDH the mobile cytochrome domain can easily bridge the distance and transfer electrons to the electrode. It also takes care of a second factor: the orientation of the enzyme on the electrode. CDH has been found to transfer electrons in any bound position with the DET being only one third of the maximum in the most unfavorable orientation which restricts the movement of the cytochrome domain [14]. Electrodes that have been used as support and collector for electrons accumulated in CDH are either carbon-based such as spectroscopic graphite, pyrolytic graphite, glassy carbon, carbon nanotubes or carbon nanofibers as well as gold-based using either plain gold electrodes or gold nanoparticles (Figure 3).

Figure 3. Electrode materials used to optimize DET of CDH in bioanodes.

a) Scanning electron microscopy (A) and atomic force microscopy (B) image of the nanostructured surface of nanoimprint lithography (NIL) gold electrodes. Reprinted from [15], with permission from Elsevier, Copyright (2015). b) SEM images of (a), (b) a carbon cloth and (c), (d) a CNTs-modified carbon cloth at different magnification. Reprinted from [16], with permission from John Wiley and Sons, Copyright (2012). c) SEM images of 3D hierarchically structured CNT/CMF/GE. a) Graphite rod with hierarchically structured carbon surfaces, b) CMFs covered with CNTs, c) magnification of one CMF with CNTs, and d) close up of the CNTs. Reprinted from [17], with permission from John Wiley and Sons, Copyright (2014). d) gold electrode etched in concentrated aqua regia. Reprinted from [18], with permission from the American Chemical Society, Copyright (1992). e) AFM image of thiol-modified gold electrode before (a) and after cytochrome c immobilization (b) and at higher resolution (c). Reprinted from [19], with permission from the American Chemical Society, Copyright (2007). f) TEM images of MWCNTs (a), PtNPs–MWCNTs (b) and PdNPs–MWCNTs (c). Reprinted from [20], with permission from the Royal Society of Chemistry, Copyright (2015). g) SEM images of drop-coated (left) and spray-coated (right) ITO NP modified electrodes with different magnification. From top to bottom: 200×, 12,000×, 100,000×. Reprinted from [21], with permission from Elsevier, Copyright (2017).

Several studies focused on the enhancement of current densities via combining CDH with a variety of redox polymers [22,23]. In DET-based biofuel cells electrode materials were modified to improve the electron transfer rate, e.g. by aryl diazonium modified single walled carbon nanotubes [24], gold nanoparticles [25] the application of polymer-coated gold nanoparticles [26] of which especially the positively charged polyethyleneimine polymer promotes immobilization of CDH and high currents [27].

Biofuel cells

CDH-based biofuel cells have to be separated in DET-based and MET-based architectures, both with benefits and limitations. The enzymes applied in combination with CDH are either bilirubin oxidase or laccase. The advantage of bilirubin oxidase as cathode biocatalyst is its good activity at neutral pH, whereas fungal laccases an acidic pH optimum. Otherwise, fungal laccases can achieve higher redox potentials (650–800 mV vs. SHE, [28,29]) than bilirubin oxidase (500–650 mV vs. SHE, [28,30]).

DET-based CDH biofuel cells and their application to power diagnostic devices has been intensively researched. During the first years establishing solid contact between enzyme and electrode was paramount and a variety of chemical modifications, e.g. thiol-based self-assembled monolayers (SAM) like N-(6-mercapto)hexyl-pyridinium on gold nanoparticles were established [31]. After this the adaption of the biofuel cells to naturally occurring operating conditions became more prominent. An interesting example are CDH-based biofuel cells developed to operate in human sweat, saliva or tear-liquid as well as cheese whey [15,21,32–34]. Another example is the application of CDH-based biofuel cells in culture supernatants to prove the applicability in this environment together with the use of unpurified CDH and laccase which resulted in a still respectable power output of 6.2 μW cm^-2 [35], or the application in cell culture with a power output of 25.5 μW cm^-2 at 0.5 V [36]. The concept of having a cheap source of enzymatic activity was taken a step further by using yeast surface display of CDH which goes in the direction of a microbial fuel cell, but is still mediatorless with a respectable power output of 3.3 μW cm^-2 at 0.35 V with a purified laccase as biocathode [37].

In regard to MET-based CDH biofuel cells a variety of polymers with different redox mediator groups were tested to achieve two aims: a fast and efficient electron transfer process from the enzyme to the electrode and a small as possible loss of voltage due to the mediators own redox potential. Low-potential biofuel cell anodes were based on toluidine blue-modified polymers [38] which led to a high open circuit potential of 720 mV when coupled to a bilirubin oxidase cathode or in a subsequent approach the usage of phenothiazine derivates in combination with cross-linkers to enhance the film stability together with an open circuit potential of 700 mV in the same setup [22]. Further studies focused on Os-polymers by optimizing the polymer backbone to Os-complex ratio [23], the properties of the hydrogel [39], or the combination of carbon microfibers, carbon nanotubes and an Os-redox polymer in a hierarchically structured electrode [17] to improve the specific power output, or a combination of carbon cloth and carbon nanotubes with the Os-polymer to increase the surface area [16]. CDH-biofuel cells based on MET have been used to power immunosensors for the determination of the sulfonamide antibiotic residue sulfapyridine in milk [40].

While a power output of 350 μW cm^-2 at 350 mV of a CDH-based bioanode was achieved with an Os-polymer/carbon nanotube embedded CDH under optimal conditions against a platinum black cathode, the highest reported power output of a full biofuel cell using also Os-polymer CDH-based bioanode and a bilirubin oxidase-based biocathode is 54 μW cm^-2 [17]. A compilation of the highest achieved power outputs in different setups and cell voltages for DET and MET-based biofuel cells based on CDH is given in Table 1.

Table 1. Highest published OCVs and power outputs of CDH-based biofuel cells.

		highest OCV [V]	Power output [μW cm-2]	at cell voltage [V]	Ref	highest Power output [μW cm-2]	at cell voltage [V]	OCV [V]	Ref
Lac	DET	0.86	3.3	0.45	[37]	15	0.55	0.73	[41]
	MET	0.55	1.9	0.275	[42]	5.87	0.2	0.5	[16]
BOX	DET	0.7	15	0.525	[25]	25.5	0.5	0.65	[36]
	MET	0.72	6.1	0.5	[38]	54	0.35	0.6	[17]

Lac, Laccase; BOX, bilirubin oxidase

An approach to increase the efficiency of biofuel cells to improve the ratio of generated electrons per substrate molecule, that is the optimization of the coulombic efficiency, was tested by employing a reaction cascade formed by a combination of CDH and pyranose dehydrogenase [43,44] by which up to six electrons could be harvested per glucose molecule. The stability of CDH-based biofuel cell anodes is affected by several factors, but most importantly by substances that inhibit or deactivate the bioelectrocatalyst. In the case of CDH, plasma components, especially uric acid, may reduce the operational stability of CDH embedded in an Os-polymer [45], but also the formation of H₂O₂ by CDH itself, although only residual side activity, limits CDH’s operational stability and was tried to be improved by protein engineering [46–48]. A compilation of all enzymatic biofuel cells employing CDH as anode biocatalyst is given in Table S1.

Biosupercapacitors

The initial work in this area started with an enzymatic fuel cell with a bioanode and a platinum modified carbon cathode in combination with a capacitor to measure glucose via monitoring the charge/discharge cycles of this “BioCapacitor” [49]. The full potential of this approach became clear soon as a way to deliver short pulses of power on demand by charge-storing biofuel cells and various enzymes were investigated for their suitability as anodic biocatalysts including glucose oxidase [50], glucose dehydrogenase [51], fructose dehydrogenase [52] as well as CDH [53]. The first CDH-based self-charging biosupercapacitor was mediator-less and because of the electrode architecture based on conducting polymer, carbon nanotubes, and gold nanoparticles also needed no external capacitor, but worked as single module [53]. This design was soon expanded to power a biodevice with the potential to measure the glucose and O₂ concentration in blood and transmit the results as a radio signal powered by the biosupercapacitor [54], and in the application of non-invasive bioelectronics working in tear-liquid [55]. The development of a polymer-based biosupercapacitors does not depend on DET and therefore also other GMC-oxidoreductases [50] or PQQ-dependent glucose dehydrogenase [56] could be employed, the latter one to fabricate a transparent and flexible supercapacitor based on indium tin oxide with a power output of 0.030 mW cm^-2 in 50 μM glucose containing phosphate buffered saline.

Conclusions

Initial research on CDH based bioanodes in biofuel cells focused on the establishment of either DET or MET of the bioelectrocatalyst with the electrode by investigating surface modifications to improve the binding or orientation of CDH. The first generated prototypes with CDH were invaluable to elucidate the performance of DET and MET based biofuel cells but also to evaluate factors needing improvement such as the operational stability, how to deal with interfering electroactive compounds in complex matrices, and how to improve power output while maintaining a high cell voltage. Current research focuses on enzyme engineering to improve the electron transfer between the biocatalyst and the electrode or redox mediator as well as designing new redox mediators and redox polymers for efficient electron transfer and minimal voltage drop in the cell. Current CDH-based biofuel cells work efficiently in complex matrices and are applicable in miniaturized sensor-transmitter devices, but still lack electrode architectures that are scalable and could promise a power output above one Watt.

Figure 4. Scheme of a MET-based biofuel cell and a biosupercapacitor.

a) Principle of a biofuel cell. On the anode, CDH (PDB: 4QI6) oxidizes cellobiose. Electrons are mediated via a low potential redoxpolymer to the electrode and via an external circuit to the cathode where electrons are donated via a high potential redoxpolymer to laccase (PDB: 3SQR), which reduces O₂ to H₂O. No separator between the electrodes is necessary. b) Principle of a self-charging biofuel cell (biosupercapacitor). When the circuit is open, the redox polymers at the electrodes serve as storage for charges originating from the catalytic reactions.