The Putative Endoglucanase PcGH61D from Phanerochaete chrysosporium Is a Metal-Dependent Oxidative Enzyme that Cleaves Cellulose

1Traditionally, enzyme systems capable of degrading recalcitrant polysaccharides, such as cellulose and chitin, are thought to consist of endo-acting enzymes that are cutting randomly in the polysaccharide chain, and processive exo-acting enzymes (chito- or cellobiohydrolases) that degrade the polymers from chain ends. These enzymes often contain one or more carbohydrate binding modules that may contribute to their activity by promoting substrate binding and perhaps even some sort of substrate disrupting action,. Still, it remains difficult to understand how these glycoside hydrolases could gain access to a polysaccharide chain in its crystalline environment, and one may wonder about the existence of additional factors that would make the substrate more accessible. Recent studies of bacterial proteins currently classified as family 33 Carbohydrate Binding Modules (CBM33) have shown that the classical endo/exo scheme indeed may be too simple,. CBM33 proteins have flat substrate-binding surfaces and are capable of cleaving polysaccharide chains in their crystalline contexts using an oxidative mechanism that depends on the presence of bivalent metal ions and an electron donor. Most importantly, CBM33 activity can boost the activity of classical cellulases or chitinases towards certain recalcitrant forms of cellulose and chitin, respectively.

2Biomass-degrading fungi produce a plethora of enzymes when growing on cellulose,,. Among proteins upregulated during growth on cellulose are proteins that are currently classified as Glycoside Hydrolases family 61 (GH61),. It has been claimed that GH61 proteins exhibit weak endoglucanase activity, but structural and recent functional studies indicate that they are not classical glycoside hydrolases. Since GH61 proteins structurally resemble CBM33s and since some of them are known to act synergistically with cellulases, it has been suggested that GH61 proteins may be the fungal analogues of CBM33's. A recent study, showing that a combination of a GH61 from Thermoascus aurantiacus (TaGH61A) and cellobiose dehydrogenase (CDH) acts synergistically with cellulases and yields oxidized products, may be taken to support this suggestion. So far, very little is known about these GH61 proteins and more experimental work is needed to unravel their functions.

3Genes coding for products classified as GH61 are abundant in the genomes of biomass-degrading fungi. One of the most extreme examples is Coprinopsis cinerea the genome of which contains as many as 33 putative GH61-encoding genes. According to current publicly available annotations,, the genome of the wood-degrading white-rot fungus Phanerochaete chrysosporium has at least 11 genes encoding a GH61 single domain protein or a GH61 fused to a carbohydrate-binding module (CBM). In a recent study on the P. chrysosporium secretome, it was shown that CDH and GH61 proteins are upregulated by xylan. In this study, we have identified a GH61 protein, hereafter referred to as PcGH61D, the expression of which is also upregulated when the fungus is grown on xylan. We have cloned and overexpressed PcGH61D in Pichia pastoris and we have used purified correctly processed recombinant protein to characterize its function. We show that PcGH61D cleaves cellulose using an oxidative mechanism and, by doing so, acts synergistically with cellulases.

4Phanerochaete chrysosporium strain K-3 was used as the source for the target gene. Escherichia coli strain JM109 (Takara Bio, Otsu, Japan) was used as subcloning host and Pichia pastoris strain KM71H (Invitrogen, Carlsbad, CA) for heterologous production of recombinant PcGH61D. The primers for subcloning and construction of expression vector were purchased from Proligo (Boulder, CO), and the primers for Quikchange were purchased from Invitrogen (Carlsbad, CA, USA). The deglycosylation enzyme, EndoH (endo-β-N-acetyl glucosaminidase, EC 3.2.1.96, from Streptomyces plicatus), was a kind gift by Genencor, a Danisco division (Palo Alto, CA, USA).

5The identification of PcGH61D protein was carried out as described in. P. chrysosporium strain K-3 was cultivated at 26.5°C using a Kremer and Wood's minimal medium with 2% cellulose as carbon source (CF11; Whatman, Fairfield, NJ, USA), and with or without an addition of 0.2% (w/v) beech xylan (Sigma-Aldrich, St. Louis, MO, USA) as described in. After 3 days of cultivation the mycelium was collected by filtrating the culture medium, and collected mycelium was frozen with liquid nitrogen. Proteins in the culture medium were analyzed using two-dimensional electrophoresis as described in. The N-terminal sequences of interesting proteins was determined by sequencing the proteins on a protein sequencer from Applied Biosystems (model 491 cLC; Applied Biosystems, Foster City, CA, USA) using the protocol described in. This procedure led to the identification of PcGH61D. The synthesis of cDNA from P. chrysosporium culture was carried out using the same protocol as described in. Total RNA was extracted from approximately 100 mg of mycelial powder of P. chrysosporium culture grown on cellulose media, using an E.Z.N.A Fungal RNA Kit (Omega Biotech, GA, USA). The first strand of the cDNA was synthesized using a Ready-To-Go™ T-primed First-strand cDNA synthesis Kit (GE Healthcare UK Ltd., Buckinghamshire, UK). Based on the annotated gene model ID 4691 (scaffold_7:1591640-1592893 in the P. chrysosporium v. 2.0 genome database at the JGI genome portal (http://genome.jgi-psf.org/Phchr1/Phchr1.home.html)), the nucleotide primers F1: 5′-CACTACACCTTCCCCGACTTCATTGAGCCTAGC-3′ and R1: 5′-TCCTTGCCAGACAGCGGGGCCAG-3′ were designed for PCR amplification of the cDNA sequence. The amplified fragment was ligated into pGEM-T Easy Vectors (Promega, Madison, WI, USA) and transformed into E. coli strain JM109. The fragment was sequenced using a Thermo Sequenase primer cycle sequencing kit (GE Healthcare) using a DNA sequencer model SQ5500E (Hitachi High Technologies, Tokyo, Japan), as described previously.

6The nucleotide sequence of the cDNA encoding PcGH61D has been submitted to DDBJ/EMBL/GeneBank databases with accession number: AB670125.

7The oligonucleotide primers F2: 5′-TTTGAATTCCACTACACCTTCCCCGACTTC-3′ and R2: 5′-TTTGCGGCCGCTATCCTTGCCAGACAGCG-3′, introducing EcoRI and NotI cleavage sites, respectively (underlined sequence) were designed to amplify a fragment of PcGH61D cDNA sequence encoding the predicted mature enzyme (i.e. without secretion signal in the cDNA sequence). The PCR product was ligated into the P. pastoris pPICZα-A vector (Invitrogen, Carlsbad, CA) using the same restriction sites as in the PCR amplified PcGH61D cDNA product. The resulting plasmid has an EcoRI restriction site that introduces additional amino acid residues, Glu-Phe, at the N-terminal of the expressed protein, as well as a STE13 protease site that may cause heterogeneous N-terminus of the protein. To substitute the STE13 site and the EcoRI restriction site in the plasmid with an enterokinase cleavage site, the pPICZα vector containing the PcGH61D cDNA sequence was modified by means of the Quikchange® method (Agilent Technologies, Santa Clara, CA, USA), using the primers 5′-CTCTCGAGAAAAGAGATGATGACGACAGACACTACACCTTCCC -3′ and 5′-GGGAAGGTGTAGTGTCTGTCGTCATCATCTCTTTTCTCGAGAG-3′. In the resulting final construct, the sequence near the start of the mature product looks as follows (His1 in bold face; enterokinase recognition motif in italics, enterokinase cleavage site indicated by a slash): LEKRDDDDR / HYTF.

8Transformation of P. pastoris was conducted as described in. Prior to transformation of P. pastoris, approximately 5 µg of the plasmid was linearized using the restriction enzyme Bpu1102I (Takara Bio, Otsu, Japan). Electroporation and selection of transformants were carried out according to the EasySelect™ P. pastoris expression kit standard protocol (Invitrogen, Carlsbad, CA). The expression of recombinant protein was performed as described in. After induction with 1% (w/v) methanol for 3 days, P. pastoris cells were removed from the culture medium by centrifugation of the culture for 15 min at 1500 x g.

9One mL of the culture supernatant, from expression in P. pastoris, was taken for checking the effect of enzymatic processing mentioned below. After concentration using VIVAspin 500 (GE Healthcare UK Ltd., Buckinghamshire, UK), 2 µL of concentrated culture medium, estimated at approximately 1 mg/mL protein concentration (BioRad Protein assay with BSA as standard), was incubated with 10 ng of EndoH for 30 min in 20 mM citrate buffer pH 5.5 at 35°C. One EndoH treated protein sample was diluted to 20 µL and 50 mM Tris-HCl buffer pH 8.0 with 2 mM calcium chloride and 50 mM sodium chloride, and then further incubated with 1 ng of enterokinase, light-chain (New England Biolabs) at 25°C for 16 hours. The protein samples treated by only EndoH and both EndoH and enterokinase were subjected to SDS-PAGE followed by protein transfer to a polyvinylidene difluoride membrane (Millipore) using a Trans-Blot SD cell (Biorad). N-terminal amino acid sequences of the protein was determined using the same method as described above.

10Ammonium sulphate was added to 150 mL of P. pastoris culture supernatant to a final concentration of 1 M, and the protein solution was applied to a Phenyl-Sepharose column (ϕ16 x 100 mm, GE Healthcare UK Ltd., Buckinghamshire, UK). Protein bound to the column was eluted by applying a linear reverse gradient of 1 to 0 M ammonium sulphate in 20 mM sodium acetate buffer pH 5.0. Fractions containing PcGH61D, as monitored by SDS-PAGE analysis, were pooled and incubated with 0.5 µg of EndoH for 24 hours at 35°C. After buffer exchange to 20 mM Tris-HCl buffer, pH 8.0, containing 50 mM sodium chloride and 2 mM calcium chloride, using VIVAspin 20 (GE Healthcare UK Ltd., Buckinghamshire, UK). Approximately 10 mL of 1 mg/mL protein solution was incubated with 10 ng of enterokinase, light chain, at 30°C for 24 hours. The enterokinase treated protein solution was applied to a source30Q column (ϕ16 x 90 mm) equilibrated with 20 mM Tris-HCl buffer pH 8.0. Proteins were eluted from the column with a linear gradient from 0 to 0.25 M sodium chloride in the same buffer. This procedure yielded two peaks containing a 27 kDa and a 25 kDa protein, respectively. Fractions belonging to the same peak were diluted to 20 mM sodium acetate buffer pH 5.0 and concentrated to 1 mg/mL using VIVAspin 20 tubes (GE Healthcare UK Ltd., Buckinghamshire, UK).

11The Expresso program at the T-COFFEE Multiple Sequence Alignment Server and ESPript were used to prepare the structure based sequence alignment of PcGH61D (residue 1 to 217 of the mature sequence starting from His 1) with the two GH61s with known crystal structures; GH61E from Thielavia terrestris (TtGH61E;; PDB code 3EII; (residue 1 to 208)) and Cel61B from Hypocrea jecorina (HjCel61B;; PDB code 2VTC; residue 1 to 230). For comparison, the TaGH61A sequence was included in the alignment.

12A homology model of PcGH61D was built using the SWISS-MODEL workspace, using the crystal structure of TtGH61E (PDB code: 3EII) as template. PcGH61D has 41% amino acid sequence identity with TtGH61E. Structural comparisons were carried out with MacPyMOL (The PyMOL Molecular Graphics System, Version 1.3, Schrödinger, LLC, USA).

13Avicel PH-101 was obtained from Fluka analytical (Sigma Aldrich, St. Louis, USA). Cellulose nanofibrils were produced as described in, and were a kind gift from Prof. Dimitris Argyropoulos, NC State university. Phosphoric acid swollen cellulose (PASC) was prepared from Avicel using the method described by Wood. Reduced PASC cellulose was prepared as follows: 2 mL of a 2% PASC suspension were centrifuged for 3 minutes at 14800 rpm, the supernatant was removed and the pellet was resuspended in 1 mL MilliQ H2O. After centrifugation (3 min, 21000 x g), the pellet was re-suspended in 4 mL 12.5 mM NaOH after which 25 mg NaBH4 was added and the tube was left at ambient temperature over night with occasional stirring. The reaction was quenched by neutralizing with 100 µL glacial acetic acid, followed by centrifugation as described above. The pellet was washed 4 times with MilliQ H2O and finally re-suspended in MilliQ H2O to obtain a 2% solution of reduced PASC.

14Various types of functional assays were conducted as described in the Results section, using chromatography and mass spectrometry for product identification as described in detail by Vaaje-Kolstad et al. and Forsberg et al.,. Depending on the type of reactions, reaction products were detected by MALDI-TOF MS (Matrix assisted laser desorption ionisation time-of-flight mass spectrometry) as described by Vaaje-Kolstad et al., and/or by High-Performance Anion-Exchange Chromatography (HPAEC) to detect native and oxidized cello-oligosaccharides, as described by Forsberg et al., and/or by measuring release of cellotriose, cellobiose and glucose, as described in Forsberg et al.. The PcGH61D-containing chromatographic fraction used contained purified correctly processed PcGH61D. The cellulase mixture used for synergy experiments was Celluclast 1.5 L (Novozymes, Copenhagen, Denmark).

15Standard reaction mixtures contained the cellulose substrate (Avicel or filter paper 10 mg/mL; phosphoric acid swollen cellulose (PASC) or nano fibrillated cellulose 1mg/mL), 4–40 µg/mL PcGH61D, 50 mM MES buffer pH 6.6, and 1.7 mM reduced glutathion, and reactions were incubated at 50°C with 900 rpm vertical shaking in an Eppendorf Thermo mixer. For the synergy experiments the enzyme concentrations were 56 µg/mL PcGH61D and 1.3 µg/mL Celluclast 1.5 L (Novozymes, Copenhagen, Denmark). In experiments were MALDI-TOF analyses were applied for product analysis, the buffer was 25 mM Tris-HCl pH 6.5 and the reductant was 1 mM ascorbic acid (instead of reduced glutathione). Samples were taken at different time points. To obtain reproducible sampling condensed water from the upper part of tube was by spun down, and the sample was then collected after first mixing the contents of the tube with the pipette. The samples were centrifuged at 21000 x g for 3 minutes. Supernatants containing soluble oxidized oligosaccharides were collected and immediately applied to HPLC or MALDI-TOF analyses. The auto-sampler of the HPLC was kept at 4°C. Synergy experiments were run in triplicates. MALDI-TOF/MS spectra (see Vaaje-Kolstad et al. for conditions) were acquired from averaging 250 arbitrary shots on each spot.

16In the two-dimensional gel electrophoresis analysis of P. chrysosporium culture filtrates, one protein spot at approximately 25 kDa molecular weight and pI of 4.8 showed much higher intensity when xylan had been added to the cellulose-containing medium. The N-terminal amino acid sequence of this protein was determined to be XYTFPDFIEPS. A BLAST search against the P. chrysosporium genome database revealed that this sequence did match with a P. chrysosporium gene model with protein ID 4691 (scaffold_7:1591640-1592893 in P. chrysosporium v. 2.0 genome database (http://genome.jgi-psf.org/Phchr1/Phchr1.home.html). Based on the annotated sequence, the corresponding cDNA was cloned and sequenced, and the correct prediction of the gene model with 10 introns was confirmed. The cloned cDNA consist of 708 bp, including an open reading frame encoding a 18-amino acid signal peptide at the N-terminal region followed by a 217-amino acid mature protein with a predicted mass of 23.5 kDa, and a theoretic pI value of 4.54. The protein coded for by the cDNA has two putative N-glycosylation sites and it consists of a single GH61 catalytic module that shows significant sequence similarity with previously characterized GH61 enzymes (22–41% identity). It has no CBM1 in contrast to the previously reported PcCel61A and B from P. chrysosporium that do, but in similarity with PcCel61C that does not,,. This novel P. chrysosporium protein was designated PcGH61D.

17PcGH61D was heterologously expressed in P. pastoris using a strategy where the key element was to generate an enterokinase cleavage site that would allow correct processing of the secreted protein, so that the catalytically important N-terminus of the mature recombinant protein would be the same as in the natural protein (i.e. starting with His1; see Materials and Methods for details). Figure 1 shows that the P. pastoris expressed PcGH61D protein appeared as a smeary 40 to 50 kDa protein band on a SDS-PAGE gel (lane 2). However, after deglycosylation with EndoH a clear major band appeared at approximately 27 kDa (lane 3). Treatment with enterokinase led to formation of a 25 kDa protein (lane 4), which was shown by amino acid sequencing to have the correct N-terminal sequence, XYTFPDFI. Note that this is the same N-terminal sequence as found for the protein produced from natural sources, and that the first X must be a histidine (Fig. 2A). The 25 kDa and 27 kDa proteins could be separated into pure fractions of each by ion exchange chromatography and their purity was confirmed by SDS.PAGE (lanes 6 & 7). The protein shown in lane 7 is not active (data not shown), most likely due to incomplete processing of the N-terminus.

18Figure 2 shows a structure based sequence alignment of PcGH61D with the two GH61 proteins for which the structure is known (TtGH61E and HjCel61B), as well as a structural comparison based on the two crystal structures and a homology model of PcGH61D. Figure 2C shows a close-up of the catalytic center highlighting four conserved residues His1, Tyr75, His76, and Tyr160 of PcGH61D. Three of these four residues are highly conserved in family GH61 proteins but Tyr75 is less conserved. This residue is substituted by a proline in HjCel61B and also in TaGH61A recently described by Langston et al., the sequence of which has been included in Figure 2A.

19Figure 3 shows analyses of products released from cellulose upon incubation with PcGH61D in the presence of a reducing agent. The data show that PcGH61D releases both native and oxidized cello-oligosaccharides, similar to what has previously been observed for the CBM33 CelS2. The release of minor fractions of native cello-oligosaccharides is most likely due to the fact that polysaccharide chains have been cleaved close to their original reducing ends, as discussed in more detail below. Mass spectrometry analysis of the products revealed masses and adduct clusters that are typical for aldonic acids (Figure 3B-D). Furthermore, MS/MS fragmentation analysis of major ions (Figure 3E) showed that the end opposite to the non-reducing end carries the extra mass introduced by the oxidation. Minor peaks with masses possibly corresponding to the lactone form were also detected in the mass spectra (Figure 3D).

20Several control experiments were done. Firstly, we were not able to detect any oxidized products when incubating PcGH61D with soluble oligosaccharides (DP2-5), which is in accordance with the previous observation that CBM33s are not active on soluble chito- or cello-oligosaccharides,. Secondly, incubation of 0.1% PASC in the presence of 1mM ascorbic acid and copper (II) sulfate at pH 5, which could lead to cellulose cleavage by enzyme-independent oxidative reactions, did not yield detectable levels of the products found in the reactions with PcGH61D.

21By analogy to observations made for chitin-active CBM33s, the presence of reducing agents (i.e. external electron donors) was expected to have a beneficial effect on the cellulose-degrading activity of PcGH61D. This is confirmed by Figure 4, which shows a clear dose-response effect of ascorbic acid on product formation. Other reductants such as reduced glutathione and gallic acid used at concentrations in the 1 mM range also stimulated PcGH61D activity, approximately to the same level as ascorbic acid. Without addition of reductants the products of PcGH61D was hardly detectable (Figure 4).

22Figure 5 shows that similar product profiles were obtained when incubating PcGH61D with different types of cellulosic substrates, including nanofibrillated cellulose Avicel and PASC. For comparison, a product profile obtained with the cellulolytic CBM33 CelS2 is also shown. While the product profiles generally are quite similar, they do show subtle differences with respect to the length distributions of the released products. Notably, CelS2 yields products that are predominantly even-numbered, whereas PcGH61D does not (see below for further discussion).

23GH family 61 proteins have a metal binding site and it is known that metals are necessary for activity. In the experiments described above it was not necessary to add extra metals to the reaction mix (addition did not lead to higher activity), indicating that sufficient amount of metals were already present in the purified protein and/or in the substrates. In an attempt to identify which is the preferred metal for GH61, a stock solution of enzyme was treated by ethylenediaminetetraacetic acid (EDTA), and different metals were tested to reactivate the protein. Figure 6 clearly shows that under the conditions used in this experiment, including use of very low metal concentrations, only copper and manganese were able to reactivate PcGH61D (at a 10 µM metal concentration in the presence of 20 µM EDTA). At 5 µM metal concentration only copper could activate the enzyme (data not shown).

24GH family 61 proteins have till now been claimed to possess a weak endoglucanase activity, based on the release of low amounts of cello-oligosaccharides from certain substrates. In order to investigate whether native cello-oligosaccharides released from cellulose by GH61 have reducing ends that were already present in the original substrate and thus not a result of GH61 hydrolytic activity, a reduced cellulose substrate was prepared in which the reducing ends are converted to glucitols. Figure 7 shows that the major products observed upon degradation of this substrate by PcGH61D were reduced and oxidized cello-oligosaccharides, whereas the amounts of released native cello-oligosaccharides were much lower than observed with normal cellulose. Interestingly, the amounts of oxidized oligosaccharides compared to glucitol oligosaccharides, as estimated from the mass spectra, increased over time (3:1 after 4 h; 13:1 after 20 h). Furthermore, the native:reduced ratio increased over time (not shown). The fact that reduced products became less prominent in the later phase of the reaction with reduced PASC may be taken to indicate that the enzyme penetrates the substrate where the number of existing chain ends and/or the degree of reduction of these chain ends are likely to be lower. Taking into account that the reduction of the substrate was incomplete (Figure 7A), these data clearly show that at least the large majority and, possibly, all of the native products observed when incubating PASC with PcGH61D have reducing ends that were already present in the substrate and that have thus not been generated by the enzyme.

25To check for synergistic effects with cellulases, cellulose was degraded by Celluclast in the presence or absence of PcGH61D. Figure 8 shows a synergistic effect of adding GH61 to the reaction mixture. Figure 6B shows that the main products produced by Celluclast are glucose and cellobiose. However, the presence of GH61 also leads to accumulation of oxidized dimers and trimers. These oxidized products are not included as glucose equivalents in Panel A, meaning that the synergistic effect shown by this figure is somewhat underestimated.

26Characterization of GH61 proteins was described for the first time in 1992, but their function has remained enigmatic until very recently (,, and this paper). They were originally thought to exhibit weak endoglucanase activity based on the detection of low levels of newly formed reducing end upon the incubation of the protein with cellulose. The data described here show that PcGH61D does cleave cellulose but that this generally does not lead to generation of new reducing ends. PcGH61D-catalyzed oxidative cleavage of cellulose close to existing reducing ends will release soluble native cello-oligosaccharides. This was indeed observed and might explain why low endoglucanase activities have been detected in some previous studies. It should be noted though that although the experiments with reduced cellulose show that the large majority of released native cello-oligosaccharides have reducing-ends that were already present in the substrate, we cannot exclude that PcGH61D has a side-activity that entails hydrolysis without oxidation.

27A potential boosting effect of GH61 proteins on cellulose activity was reported in 2007 and underpinned in a later study by Harris et al.. The discovery of the chitin-cleaving activity of the structurally homologous CBM33 proteins provided a possible explanation for this GH61 activity. Indeed, we show here that PcGH61D has a similar effect on cellulose as the chitin-active CBM 33 CBP21 has on chitin and as the cellulose-active CBM33 CelS2 has on cellulose. PcGH61D cleaves cellulose at the glycosidic bond, leaving one of the new chain ends oxidized to a lactone, which subsequently is spontaneously converted to an aldonic acid in solution. The enzyme is not active on soluble cello-oligosaccharides and its flat active site-containing surface, devoid of potential substrate-binding grooves and pockets,, suggests that it is optimized for interacting with ordered substrate surfaces such as they occur in crystalline and otherwise well-ordered substrates. This is analogous to what has been claimed, and to a large extent proven, for CBM33s.

28The activity of both CBM33s and GH61s is increased in the presence of external electron donors, be it chemical reductants such as ascorbic acid (see above) or electron generating enzyme systems such as CDH. It should be noted though that the few studies now available in the literature show differences between the systems studied. The original work on chitin-active CBP21 showed dramatic effects both in terms of the effect of adding reductants and the boosting effect on chitinase activity. For the cellulolytic CBM33 and GH61 enzymes studied so far, these effects are less dramatic and generally in the same range as the effects described here for PcGH61D,,. The product profiles in Figure 5 reveal differences between cellulolytic CBM33s and GH61s and between substrates. For example, whereas the products generated from Avicel by CelS2 (a CBM33) show a dominance of even-numbered products, as previously seen for CBP21 acting an chitin (see Vaaje-Kolstad et al., 2010 for further discussion), the products released by PcGH61D from the same substrate do not show this periodicity. Figure 5 also shows that chromatographic product profiles generated by PcGH61D differ slightly for the different substrates. It is conceivable that these differences reflect different binding modes of the enzymes; for example, the enzymes may vary in terms of their ability to act on substrate regions of varying crystallinity or they may bind to different faces of the crystalline material. Interestingly, the recent study by Langston et al. indicated that TaGH61A, which shows some interesting sequence differences with PcGH61D (Figure 2A) may in fact have a slightly different oxidative mechanism since several products were detected in addition to aldonic acids. Further work is needed to unravel the causes and implications of the possible differences between members of the GH61 and CBM33 families.

29Another issue that so far has remained partly unresolved is the dependency of both CBM33s and GH61s on bivalent metals. Initially, it was reported for both types of enzymes that they can work with a wide variety for bivalent metal ions, including several that are not redox-active. It is likely that these remarkable observations were partly misinterpreted due to the fact that the concentrations of added metals were too high, thus releasing already present metals from binding sites through displacement. It would in fact be quite remarkable if the activity of these redox enzymes did not depend on the presence of a redox metal. Our present data indicate quite clearly that PcGH61D is a copper oxidase, and that it can use manganese in the absence of copper. Interestingly, our own (partly unpublished) observations on the metal-dependency of CBM33s have so far not yielded equally clear results.

30When this paper was about ready for submission, a report by Quinlan et al. appeared online that describes an in-depth study of TaGH61A and demonstrates that this is an oxidative enzyme that cleaves cellulose. TaGH61A resembles PcGH61D (and cellulose-active CBM33s; ) in that the enzyme produces aldonic acids and works better in the presence of redox-active cofactors such as ascorbate. The study by Quinlan et al. also provides evidence for TaGH61A being a copper oxidase. Most interestingly, this very recent study also confirmed differences between PcGH61D and TaGH61A, since the latter enzyme was found to produce additional oxidized species that we have not observed for PcGH61D. As noted above and in Figure 2, the sequences of the two enzymes show some interesting differences that could underlie functional differences.

31The recent findings on CBM33 and GH61 proteins add a completely new dimension to the classical concept of cellulose degradation by endo- and exo-acting cellulases. Clearly, nature has developed additional oxidative enzyme systems for tackling this recalcitrant substrate that may be specifically tailored for acting on the least accessible regions. It is important to note the abundance of GH61 in fungal genomes and the (not quite as large) abundance of CBM33s in some bacterial genomes. It seems quite likely that additional substrate specificities will be discovered for these proteins. For example, GH61s acting on chitin are likely to exist, as are GH61s acting on certain xylans, galactomannans, or heteropolymeric polysaccharide materials. The data presented in Figure 5 and discussed above suggest functional differences between cellulolytic GH61s/CBM33s and it is thus conceivable that synergistic effects may be observed when combining several of these proteins working on the same substrate. We envisage that more knowledge on these GH61 and CBM33 enzymes will contribute considerably to our general understanding of enzymatic biomass conversion and open up new avenues towards efficient industrial biomass processing.