Springer 2005
Biotechnology Letters (2005) 27: 555–560 DOI 10.1007/s10529-005-2881-8
Electrochemical oxidation of water by a cellobiose dehydrogenase from Phanerochaete chrysosporium Jun Feng, Michael E. Himmel & Stephen R. Decker* National Bioenergy Center, National Renewable Energy Laboratory, Golden, Colorado, CO 80401, USA *Author for correspondence (Fax: +303-384-7752; E-mail:
[email protected]) Received 21 December 2004; Revisions requested 17 January 2005; Revisions Received 15 February 2005; Accepted 15 February 2005
Key words: cellobiose dehydrogenase, cyclic voltammetry, enzyme electrode, pyrolytic graphite
Abstract Cellobiose dehydrogenase (CDH) is a redox protein containing two electron transfer centers; a flavin coenzyme performing a two-electron transfer reaction and an iron-heme coenzyme facilitating singleelectron transfer. Purified CDH from Phanerochaete chrysosporium was immobilized on a pyrolytic graphite electrode and electron transfer from cellobiose to the electrode was generated. With cellobiose present during cyclic voltammetry, this novel enzyme/electrode system exhibited sharp, stable oxidation peaks with slower, though equivalent, reduction peaks. During cyclic voltammetry without substrate, the enzyme was rapidly oxidized during the initial scan, with no corresponding enzyme reduction during the reducing half of the cycle. After resting for several hours in aqueous buffer, the full oxidation current appeared again. These results suggest that the CDH is reduced by water splitting, albeit at a slow rate.
Introduction Cellobiose dehydrogenase is produced by certain fungi and bacteria, generally in conjunction with cellulose and lignin degradation. It is a twodomain hemoflavo-oxidoreductase carrying two prosthetic groups: a catalytically active flavin adenine dinucleotide (FAD) and a cytochrome b-type heme domain. Though the biological role of CDH is still under debate (Goodell 2003, Mason et al. 2003), it readily accepts electrons from cellobiose, other disaccharides, and cellooligomers and donates electrons to acceptors such as metal ions, certain redox dyes, and possibly lignin and polysaccharides found in plant cell walls (Henriksson et al. 2000). The heme coenzyme, located in the smaller domain, is tethered to the larger FAD-containing domain by a short linker peptide (Stoica et al. 2004). The FAD-containing domain is responsible for catalysis as well as cellulose binding and subsequently transfers
single electrons within the CDH protein from flavin to heme, an orientation potentially contributing to unidirectional electron transfer (Henriksson et al. 1993, Larsson et al. 2000). Many studies have reported the potential uses of CDH-immobilized electrodes, usually as some form of biosensor to measure various CDH substrate levels (Vilkanauskyte et al. 2002). When CDH is immobilized on to a graphite electrode, a complex electron transfer system is established with electrons flowing in pairs from cellobiose to FAD, as single electrons from FAD to the heme through the linker peptide, and finally from the heme to electrode directly (Lindgren et al. 2000). This system has three electron transfer barriers: substrate to FAD, FAD to heme, and heme to electrode. Proteins bind strongly with pyrolytic graphite electrode (PGE) surfaces, being held in close association with the graphite through interactions along the ÔedgesÕ of carbon nano-islands (Leger
556 et al. 2003). Carbon nano-islands are slight irregularities on the graphite surface that provide protein interaction sites through available hydroxyl and carboxyl groups (Azamian et al. 2002). We reasoned that this interaction could lead to the formation of CDH Ôpseudo-crystalsÕ originating from these nano-islands and thus investigated these surfaces using atomic force microscopy and electrochemical measurements.
ing and washing in water, the CDH electrode was stored wet for up to several months, with little or no change in effectiveness. Cyclic voltammetry was carried out over a scan range of )0.5 to +0.3 V. The scan rate was typically 1 V min)1. The buffer used was 50 mM Na3PO4/ H3PO4, 100 mM NaCl, pH 6.5. Cellobiose was ACS reagent grade (Sigma-Aldrich) and cellulose was Sigmacell-50 (Sigma-Aldrich).
Results and discussion Materials and methods Cellobiose dehydrogenase was purified from the white-rot fungus Phanerochaete chrysosporium after growth on cellulose in 10 l submerged fermentation. Briefly, the enzyme was isolated by sequential FPLC. Culture broth was loaded on to a 20 ml Q-Sepharose anion exchange column (Amersham Biosciences) and eluted with a 0–1 M NaCl gradient in 20 mM Tris/HCl pH 8.0. Active fractions were pooled, brought to 1 M (NH4)2SO4, loaded on to a 20 ml Phenyl Sepharose hydrophobic interaction column (Amersham Biosciences) and eluted with a descending 1–0 M (NH4)2SO4 gradient, also in 20 mM Tris/HCl pH 8.0. Pooled active fractions were concentrated with a 10 kDa MWCO membrane (Millipore) and buffer exchanged into 100 mM NaCl, 20 mM sodium acetate/HCl buffer pH 5.0 by size exclusion chromatography on a 26/60 SuperDex200 column (Amersham Biosciences). The purified protein was concentrated using a 10 kDa MWCO membrane (Millipore) to 310 lg ml)1 (Pierce BCA assay) and stored at 4 C. Pyrolytic graphite electrodes were fabricated by tightly fitting a 2-mm length of 3-mm diameter graphite rod into an acrylic cylinder, wiring the electrode, and sealing the system with epoxy. After construction, the face of the electrode was polished using a commercial polishing kit and set up in a 3-electrode cell connected to a computercontrolled CHI618B electrochemical analyzer (CH Instruments). A platinum wire counter electrode and Ag/AgCl reference electrode were used in all experiments and all potentials are reported against this reference. The working electrode was derivatized with CDH either by immersion in CDH solution (310) for a few minutes, or by pipetting 2 ll CDH onto the surface. After dry-
Atomic force microscopy (AFM) using the tapping probe mode showed a topographically intriguing surface profile for the CDH treated pyrolytic graphite electrodes (see Figure 1). Three essential features were observed using AFM. First, the graphite surface appears to be extensively covered with a fine granular material that extends about 1.5 nm from the baseline grid (i.e. z axis). Throughout this region of the film, there is also a high density of cylindrical structures about 5 nm high, suggesting specific growth zones with a distinct capping material. Lastly, these cylinders appear to coalesce into complex but regular structures of larger size and height (about 15 nm in z axis). Figure 2 shows the electrochemical behavior of the CDH PGE in the presence of cellobiose, where the electrode displays a repeatable cyclic voltammetry scan. The oxidation peak appears at 0.06 V vs. Ag–AgCl and the reduction peak appears at )0.06 V vs. Ag–AgCl. Compared to the oxidation spike, the reduction peak is much reduced in amplitude, but longer in duration, resulting in a balanced system. This behavior can be explained by the rapid transfer of electrons from FAD fi heme fi electrode and the relatively slow transfer of electrons from cellobiose fi FAD, perhaps due to diffusional constraints of cellobiose entering and cellobionolactone leaving the CDH active site. It has been well established that cellobiose is the preferred substrate for CDH, with cellulose acting as a weak electron donor (Kajisa et al. 2004). With cellulose, the CDH film electrode behaved identically as with cellobiose, except for the decreased peak current (Figure 2). This behavior may be explained by the high mobility of soluble cellobiose relative to insoluble cellulose, as well as the com-
557
Fig. 1. Atomic force microscopy of CDH film on the surface of pyrolytic graphite electrode. (a) Region of surface height profile about 1.5 nm above baseline which extends to about 5 nm above baseline at tops of growing ÔcylindersÕ. (b) Region with height profiles from baseline to about 15 nm for the ÔbutterflyÕ pattern.
paratively few reducing ends available in cellulose. In this enzyme/electrode system, there are two pairs of electron donor/acceptors, the flavin–heme pair and the heme–electrode pair. The protein linker connecting the FAD and heme provides an effective kinetic barrier in the former case, while direct contact is likely for the latter case. There are at least two factors important to consider relative to these barriers: one is the van der Waals distance and the other is the actual distance separating the donor and acceptor centers, which contributes to the barrier kinetics and the reorganization energy (Onuchic et al. 1986), and then generates a gate effect in the electron transfer reaction (Jeuken et al. 2002). Such gate effects are indicated by the continuously cyclic voltammetry of CDH/electrode system in buffer solution (Figure 3). When we examined the behavior of the CDH enzyme/electrode system in buffer solution, several effects were noted. First, the spike peak
appeared in the oxidation scan with less than 15 mV in half width peak (HWP). Because of the limitation of a one-electron transfer reaction at room temperature, this value of 15 mV is much smaller than the 59 mV predicted by the Nernst equation. This may be an example of inner molecule attraction as a consequence of the electron transferring through these molecules. Second, the oxidation peak shifted +0.02 V when substrate was absent. Third, after donating the electrons to the electrode, the enzyme could not be reduced by the electrode during the reverse cycle. Finally, the electrochemical behavior of CDH/electrode system in substrate-free buffer is not only irreversible but is also not continuous. The most probable explanation for the oxidation potential shift between substrate and buffer is that after being oxidized, the FAD exerts an electronegative force on the donated electrons and impedes their transfer. The presence of cellobiose would allow a rapid reduction of the FAD, reducing the electronegativity of the FAD and
558
Fig. 2. Voltammetric behavior of CDH film on pyrolytic graphite electrode (10 mV s)1) in 50 mM phosphate buffer with (a) cellobiose (blue), (b) cellulose (green), and (c) buffer only (red).
Fig. 3. Cyclic voltammetry of CDH on pyrolytic graphite electrode in 50 mM phosphate buffer displaying diminishing current with successive scans. The peak current dropped about 88% from the first scan to the last scan, which suggests that the CDH enzyme performs as an electron sink, quickly losing electrons to the electrode, and slowly filling via solution electrochemistry.
lowering the needed potential. A conformational change in the protein induced by cellobiose binding could also affect the transfer by altering the interaction of the heme and electrode surface or the linker peptide. We propose that in the presence of cellobiose, the protein is induced to a conformation that enables rapid electron transfer from the FAD fi heme fi electrode. In the absence of cellobiose, electron transfer still occurs rapidly, but requires a higher potential and is essentially irreversible by electrode-mediated reduction. This could indicate that the ready replacement of electrons in the oxidized FAD group enhances the electron transfer, lowering the required potential. The lack of a reduction peak in the substratefree system indicates that the system is a one way transfer, i.e. FAD fi heme fi electrode. If the FAD is distal to the electrode, direct electron transfer would be prohibited. The terminal oxidation species, i.e. the heme group, is presumably close to and interacts with the electrode. It does not interact directly with substrate, but transfers electrons rapidly to the electrode, causing the current density to quickly drop, forming the spike peak. On the other end of the CDH molecule, the electric potential of the FAD coenzyme is increased after losing electrons. In the absence of cellobiose, there is no substrate present to perform the fast electrochemical reaction on the interface of enzyme/solution and replace the electrons to the FAD sink. The increased potential of FAD domain is the direct reason for rapid decrease in the current density. The lack of a reduction peak in the absence of substrate demonstrates that the heme domain either moves away from the electrode in the absence of substrate, or is connected to a one-way electron path, i.e. an enzyme-diode, which increases the electric barrier of heme/electrode. If cellobiose is present, the fast electrochemical reaction on the enzyme/solution interface causes FAD reduction followed by rapid electron transfer to the heme group, allowing the oxidation and reduction peaks to repeatedly cycle. In a redox enzyme, the electric potential within the enzyme may dominate the electron transfer (Adams 1990). For CDH, the electron transfer from the flavin domain to the heme domain is dependent on the Ômolecular inner recognition’; the efficiency of the protein to channel electrons from FAD fi heme (Larsson et al. 2000).
559 Biologically, the mechanisms of Ômolecular inner recognitionÕ are still unclear, with at least two parallel interpretations for this recognition (Henriksson et al. 1993). Our experimental results support heme proximity to the electrode as the determining factor in whether the circuit is open or closed, with the redox state of the FAD dictating the heme conformation. In solution, this has little effect on the ability of the heme to transfer electrons to a soluble acceptor, as the entire CDH is free to move, so heme orientation is somewhat moot. When the CDH is immobilized on a PGE, however, the freedom of movement is restricted and the proximity of the heme to the electrode surface can control the electron transfer. We propose that reduced FAD orients the heme proximal to the electrode, allowing electron transfer. Oxidized FAD re-orients the heme group distal to the electrode, preventing reduction during cyclic voltammetry. The CDH enzyme/electrode system in buffer cycles with a continuous reduction in current density from the first scan to the last scan dropping about 88% (Figure 3). However, the full spike peak reappears after several hours of ÔrestÕ and this electron releasing and refilling again continued cyclically. This finding may indicate that the oxidized CDH is recovering electrons from something in the substrate-free solution. In the absence of traditional substrates, the most likely candidate is water, with the low potential of the oxidized FAD group necessitating a long time to obtain electrons. The oxidized flavin prosthetic group appears to be slowly reduced in substrate-free buffer, implying that the CDH/ electrode system can split water (analysis of bubbles formed during the reaction is underway). In the study of the relationship of peak current and scan rate, we find that peak current increases a little with scan rate increase if the interval time between each measurement is 5 min, and even decreases if the interval time between each measurement is less than 1 min (Figure 4). Only with enough ‘rest’ time within each measurement, does the enzyme film display the linear relationship of the peak current with the scan rate (Figure 4), which is in agreement with the diffusion free principle of the enzyme film electrode process. In conclusion, the CDH/electrode system performs charge separation in buffer via normal dynamic electrochemistry at the heme electro-
Fig. 4. (e) Linear relationship between peak current and scan rate. Each measurement was performed after one day. (h and n) Exhaustive discharge of the enzyme film on PGE electrode. The recovery time for (h) was 5 min and for (n) less than 1 min.
chemical potential, leaving the FAD in an oxidized state. In the absence of substrate, slow reduction of the FAD through water splitting is one possible explanation of the ‘rest’ required to allow additional electrode oxidation during subsequent oxidative scans. This raises the concept of splitting water via enzyme/electrode systems, potentially a convenient way to generate bioelectricity.
References Adams MWW (1990) The structure and mechanism of ironhydrogenases. Biochim. Biophys. Acta 1020: 115–145. Azamian BR, Davis JJ, Coleman KS, Bagshaw CB, Green MLH (2002) Bioelectrochemical single-walled carbon nanotubes. J. Am. Chem. Soc. 124: 12664–12665. Goodell B (2003) Brown-rot fungal degradation of wood: our evolving view. Wood Deter. Preserv. 845: 97–118. Henriksson G, Johansson G, Pettersson G (2000) A critical review of cellobiose dehydrogenases. J. Biotechnol. 78: 93–113. Henriksson G, Johansson G, Pettersson G (1993) Is cellobiose oxidase from phanerochaete chrysosporium a one-electron reductase? Biochim. Biophys. Acta 1144: 184–190. Jeuken LJC, McEvoy JP, Armstrong FA (2002) Insights into gated electron-transfer kinetics at the electrode-protein interface: a square wave voltammetry study of the blue copper protein azurin. J. Phys. Chem. B 106: 2304–2313. Kajisa T, Yoshida M, Igarashi K, Katayama A, Nishino T, Samejima M (2004) Characterization and molecular cloning of cellobiose dehydrogenase from the brown-rot fungus coniophora puteana. J. Biosci. Bioeng. 98: 57–63. Larsson T, Lindgren A, Ruzgas T, Lindquist SE, Gorton L (2000) Bioelectrochemical characterisation of cellobiose dehydrogenase modified graphite electrodes: ionic strength and ph dependences. J. Electroanal. Chem. 482: 1–10.
560 Leger C, Elliott SJ, Hoke KR, Jeuken LJC, Jones AK, Armstrong FA (2003) Enzyme electrokinetics: using protein film voltammetry to investigate redox enzymes and their mechanisms. Biochemistry 42: 8653–8662. Lindgren A, Larsson T, Ruzgas T, Gorton L (2000) Direct electron transfer between the heme of cellobiose dehydrogenase and thiol modified gold electrodes. J. Electroanal. Chem. 494: 105–113. Mason MG, Nicholls P, Wilson MT (2003) Rotting by radicals – the role of cellobiose oxidoreductase?. Biochem. Soc. Trans. 31: 1335–1336.
Onuchic JN, Beratan DN, Hopfield JJ (1986) Some aspects of electron-transfer reaction dynamics. J. Phys. Chem. 90: 3707–3721. Stoica L, Lindgren-Sjolander A, Ruzgas T, Gorton L (2004) Biosensor based on cellobiose dehydrogenase for detection of catecholamines. Anal. Chem. 76: 4690–4696. Vilkanauskyte A, Erichsen T, Marcinkeviciene L, Laurinavicius V, Schuhmann W (2002) Reagentless biosensors based on co-entrapment of a soluble redox polymer and an enzyme within an electrochemically deposited polymer film. Biosens. Bioelectron. 17: 1025–1031.
