Noncatalytic Docking Domains of Cellulosomes of Anaerobic Fungi

4 downloads 2146 Views 1MB Size Report
domain, mostly two copies of a 40-amino-acid cysteine-rich, noncatalytic ... enzymes act individually and are free in solution, whereas ...... animal hosts, p.
JOURNAL OF BACTERIOLOGY, Sept. 2001, p. 5325–5333 0021-9193/01/$04.00⫹0 DOI: 10.1128/JB.183.18.5325–5333.2001 Copyright © 2001, American Society for Microbiology. All Rights Reserved.

Vol. 183, No. 18

Noncatalytic Docking Domains of Cellulosomes of Anaerobic Fungi PETER J. M. STEENBAKKERS,1 XIN-LIANG LI,2 EDUARDO A. XIMENES,3 JORIK G. ARTS,1 HUIZHONG CHEN,2 LARS G. LJUNGDAHL,2 AND HUUB J. M. OP DEN CAMP1* Department of Microbiology, Faculty of Science, University of Nijmegen, NL-6525 ED Nijmegen, The Netherlands1; Department of Biochemistry and Molecular Biology and Center for Biological Resource Recovery, University of Georgia, Athens, Georgia 30602-72292; and Laboratorio De Enzimologia, Departamento De Biologia Celular, Universidade De Brasilia, Asa Norte, Brasilia-DF, Brazil 70910-9003 Received 10 April 2001/Accepted 22 June 2001

A method is presented for the specific isolation of genes encoding cellulosome components from anaerobic fungi. The catalytic components of the cellulosome of anaerobic fungi typically contain, besides the catalytic domain, mostly two copies of a 40-amino-acid cysteine-rich, noncatalytic docking domain (NCDD) interspaced by short linkers. Degenerate primers were designed to anneal to the highly conserved region within the NCDDs of the monocentric fungus Piromyces sp. strain E2 and the polycentric fungus Orpinomyces sp. strain PC-2. Through PCR using cDNA from Orpinomyces sp. and genomic DNA from Piromyces sp. as templates, respectively, 9 and 19 PCR products were isolated encoding novel NCDD linker sequences. Screening of an Orpinomyces sp. cDNA library with four of these PCR products resulted in the isolation of new genes encoding cellulosome components. An alignment of the partial NCDD sequence information obtained and an alignment of database-accessible NCDD sequences, focusing on the number and position of cysteine residues, indicated the presence of three structural subfamilies within fungal NCDDs. Furthermore, evidence is presented that the NCDDs in CelC from the polycentric fungus Orpinomyces sp. strain PC-2 specifically recognize four proteins in a cellulosome preparation, indicating the presence of multiple scaffoldins. essential for attachment (24). The scaffoldin includes nine repeated cohesins, a cellulose binding domain, and a special dockerin type II domain that, through a special polypeptide, binds the cellulosome to the cell surface (17, 20). The threedimensional structures of two cohesins of CipA are known (32, 34), and the cohesin-dockerin interaction has been extensively studied (25). Cellulosomes from anaerobic fungi are similar in size and contain about as many polypeptides as the C. thermocellum cellulosome. Molecular biological evidence is accumulating that enzymes associated with the fungal cellulosomes from the genera Neocallimastix, Orpinomyces, and Piromyces, like those of anaerobic bacteria, are modular. In addition to the catalytic domain they contain, one, but mostly two or, in a few cases, three copies of a conserved 40-amino-acid cysteinerich, noncatalytic docking domain (NCDD), which shows no sequence homology to bacterial dockerins (1, 16, 18, 22, 26, 28, 41). NCDDs are interspaced by short unique linkers and are separated from the catalytic domain(s) by a serine-threoninerich linker(s). The general opinion is that enzymes with NCDDs are subunits of fungal cellulosomes. Recently, it was shown that a glutathione S-transferase (GST) reporter protein fused to one, two, or three NCDDs from Piromyces equi were able to specifically recognize a 97-kDa protein present in a cellulosome preparation purified by a cellulose affinity procedure (15). These results strongly indicate the presence of a scaffoldin analogue in the fungal cellulosome and also show that, in contrast to bacterial dockerins, a single NCDD may serve as the interacting unit. Thus far, estA from P. equi (15) and celB2 from Orpinomyces joyonii (28) are the only examples of genes encoding a cellulosome component containing only one NCDD. The majority of the genes encoding components contain two NCDDs. The genes encoding scaffoldins from several

Obligately anaerobic fungi, first described by Orpin (27), are part of the intestinal flora of herbivorous animals. They are involved in the solubilization and fermentation of plant cell wall material (31, 35, 36). For that purpose, they secrete a variety of hydrolytic enzymes, including cellulases, xylanases, mannanases, esterases, glucosidases, and glucanases, which effectively hydrolyze cellulose and hemicellulose. Some of these enzymes act individually and are free in solution, whereas others are constituents of large (hemi)cellulase multienzyme complexes remarkably similar to the cellulosomes of several species of clostridia and other anaerobic bacteria (11, 14, 23, 33, 37). Among anaerobic bacteria, the cellulosome of Clostridium thermocellum was the first described and is the most investigated (33). It has a molecular mass of about 3,000 kDa and consists of at least 26 different polypeptides with masses ranging from 38 to 210 kDa (19). Nineteen of the encoding genes have been sequenced. All cellulosomal components have modular structures. The enzymatically active components are composed of a catalytic domain and a conserved domain called the dockerin or the protein docking domain. A serine-threoninerich linker separates the domains. The dockerin consists of a 22-amino-acid tandem repeat interspaced by a short linker. The function of dockerins is to attach the enzymatically active subunits to cohesins of a noncatalytic polypeptide called scaffoldin or cellulosome-integrating protein (CipA) (17). It has been shown that both amino acid stretches of the dockerin are * Corresponding author. Mailing address: Department of Microbiology, Faculty of Science, University of Nijmegen, Toernooiveld 1, NL-6525 ED Nijmegen, The Netherlands. Phone: 31-(0)243652657. Fax: 31-(0)243652830. E-mail: [email protected]. 5325

5326

STEENBAKKERS ET AL.

anaerobic bacteria have been sequenced (2, 12), but no scaffoldin-encoding gene has been isolated from an anaerobic fungus. For the isolation of genes encoding (hemi)cellulases from anaerobic fungi, the general approach has been to screen expression libraries in Escherichia coli with appropriate substrates. Unfortunately, this approach does not discriminate between cellulosome-associated and freely occurring (hemi) cellulases and selects against those enzymes, which require eukaryotic transcription and translation machinery for activity. In this paper, we describe a strategy by which to obtain most of the genes specifically encoding (hemi)cellulases with NCDDs and therefore constituents of anaerobic fungal cellulosomes. The strategy is based on obtaining novel NCDD sequences by PCR. Degenerate primers were designed to anneal to the highly conserved region of NCDDs. PCR with cDNA or genomic DNA from the polycentric fungus Orpinomyces sp. strain PC-2 or the monocentric fungus Piromyces sp. strain E2, respectively, as the template was used to isolate 16 and 19 PCR products, respectively, encoding distinctive but incomplete NCDDs and their linker sequences. Screening of an Orpinomyces sp. cDNA library with four of these PCR products resulted in the isolation of new full-length genes encoding cellulosome components. Further, a three-subfamily division for the fungal NCDD is proposed, based on alignments of novel and published NCDD sequences. In addition, it is shown that the NCDDs in cellulase CelC specifically bind potential scaffoldins of the cellulosome of Orpinomyces sp. strain PC-2. MATERIALS AND METHODS Organisms and DNA isolation. Piromyces sp. strain E2 (ATCC 76762) was cultured in M2 medium at 39°C as described before (35), using 0.5% (wt/vol) fructose as the carbon source. Media were inoculated from exponentially growing precultures (1%, vol/vol). After 48 h of growth, the fungal biomass was harvested by filtration. Genomic DNA was isolated as described by Brownlee (5). Orpinomyces sp. strain PC-2 isolated from the rumen of a cow (4) was cultured with Avicel or oat spelt xylan (1.0%, wt/vol) as the carbon source, and a cDNA library was constructed as previously described (7). E. coli strain INV⬘⬘F⬘ and plasmid pCRII, purchased from Invitrogen (Carlsbad, Calif.) as the TA cloning kit, were used for PCR cloning. E. coli strains JM109 and BL21(DE3), from Stratagene (La Jolla, Calif.), and plasmid pRSETB, from Invitrogen, were used for heterologous protein production. Isolation of the cellulase-hemicellulase complex of Orpinomyces. Orpinomyces sp. strain PC-2 was grown on a medium containing Avicel as the sole carbon source at 39°C for 3 days. The association between Avicel and the fungal mycelium was disrupted, and the mycelium was removed by passing the culture liquid through three layers of Miracloth (Calbiochem, Anaheim, Calif.). The Avicel with the cellulase-hemicellulase complex attached was collected by centrifugation (4,000 ⫻ g, 20 min). Five grams of wet Avicel was washed three times with 40 ml of 50 mM sodium citrate buffer, pH 5.8, and then the complex was eluted with three 40-ml washes with distilled water. The extractions were done at 4°C by shaking the suspension at 150 rpm on a BellyDancer shaker (Strovall Life Science, Inc., Greensboro, N.C.) for 15 min, followed by centrifugation (4,000 ⫻ g, 15 min). Concentrated sodium citrate buffer was added to the water wash fractions to a final concentration of 50 mM. The suspension was incubated overnight at 37°C. Residual insoluble material was removed by centrifugation (8,000 ⫻ g, 30 min). The supernatant was concentrated with an Amicon cell (50 ml) equipped with a PM10 membrane. The final concentrate constituted the cellulosomal preparation. Design of NCDD-specific primers. Degenerate primers were designed to anneal to the highly conserved region within the 40 amino acids of NCDDs (Fig. 1). Primers directed toward NCDDs of Piromyces sp. strain E2 were constructed based on genes from P. equi available in databases (Fig. 2). The following genes were used (accession numbers are in parentheses): manA (X91857), manB (X97408), manC (X97520), and xynA (X91858) (14, 26). The Piromyces sp. strain E2 primer set consisted of NCDD forward, 5⬘-AAAAYRRHGAHTGGTGTG

J. BACTERIOL.

FIG. 1. Schematic representation of modular (hemi)cellulases isolated from anaerobic fungi, explaining the concept of the PCR method applied. PCR results in amplification of the unique linker region between the NCDD repeats. Because the majority of NCDDs from anaerobic fungi contain two repeats, every PCR product will represent the partial sequence of one NCDD repeat-containing gene. SP, signal peptide.

G-3⬘, and NCDD reverse, 5⬘TTCAAYACCCCADTCACC-3⬘. Y represents C or T, R represents A or G; H represents A, T, or C; and D represents G, A, or T. Primers used to amplify NCDD-specific sequences of Orpinomyces sp. strain PC-2 were designed based on conserved NCDD sequences within genes previously identified by activity screening. The genes include those coding for cellulases A (U63837), B (U57818), C (U63838), and E (U97153) and xylanase A (U57819) (6, 21, 22). The Orpinomyces primer set contained NCDD forward, 5⬘-GGIAAITGGGGIGTIGARAA-3⬘, and NCDD reverse, 5⬘-TTYTCIACICC CCAITTICC-3⬘, coding for GK(N)WGVEN and GK(N)WGVEN (reverse and complementary), respectively. R represents A or G; Y represents C or T; and I represents inosine. PCR conditions and analysis of PCR products. PCR with genomic DNA from Piromyces sp. strain E2 as the template was performed with a Perkin-Elmer (Norwalk, Conn.) DNA thermal cycler. It involved a hot start, 40 cycles at 94°C for 30 s, and annealing at 50°C for 30 s. The PCR was performed in a volume of 50 ␮l containing 1 ␮g of genomic DNA, 2.5 mM MgCl2, 0.5 U of Taq polymerase (Gibco, Bio-Rad Laboratories, Richmond, Calif.), and 200 pmol of each primer. The PCR with the primers from Orpinomyces sp. was performed with the cDNA library as the template and involved 35 cycles at 95°C for 1 min, 40°C for 1 min, and 72°C for 1.5 min. The PCR products from Orpinomyces and Piromyces spp. were analyzed on a 2 or 3% (wt/vol) agarose gel and cloned in pCRII (Invitrogen) and pGEM-T Easy (Promega, Madison, Wis.). Transformants were grown in Luria-Bertani medium containing ampicillin (50 ␮g/ml). Vectors with inserts of 110 to 200 bp were sequenced. Both strands of Piromyces genomic PCR clones were sequenced with the dRhodamine DNA sequencing kit (Perkin-Elmer) on a Perkin-Elmer automated fluorescent ABI Prism 310 sequencer. Both strands of the Orpinomyces cDNA products were sequenced with an automated PCR sequencer (Ap-

FIG. 2. Alignment of the highly conserved nucleotide sequence within NCDDs from P. equi showing the relative positions and directions of the primers designed for the Piromyces sp. strain E2 genomic DNA PCR. The following sequences were used for the alignment (accession numbers are in parentheses): manA (X91857), manB (X97408), manC (X97520), and xynA (X91858) (14, 26). Forw., forward; Rev., reverse.

VOL. 183, 2001

NONCATALYTIC DOCKING DOMAINS OF ANAEROBIC FUNGI

plied Biosystems, Foster City, Calif.). Alignment of translated clones was performed by hand in the Seaview alignment editor after Phylip alignment. Data were also analyzed with the Genetics Computer Group program (version 8; University of Wisconsin Biotechnology Center, Madison) on the VAX/VMs system of the Bioscience Computing Resource at the University of Georgia. Hybridization specificity of the isolated NCDD-PCR products. To investigate the specificity of the isolated clones for use as a probe in the screening of a genomic or cDNA library, we tested standard hybridization conditions by using the largest, NCDD-PCR19, and smallest, NCDD-PCR8, clones isolated. All 19 novel Piromyces sp. clones were reamplified by using vector primers, separated on a 2% agarose gel, and scanned with a Gel Doc 1000 scanner and the multi analyst software (Bio-Rad Laboratories, Richmond, Calif.). DNA was transferred to a Nytran supercharge membrane by downward blotting in accordance with the manufacturer’s (Schleicher & Schuell) protocol. The membrane was dried on Whatman 3MM paper and wrapped in Cling Rap. The denatured DNA was covalently bound to the membrane by UV cross-linking. Clones NCDDPCR8 and NCDD-PCR19 were labeled by PCR using the degenerate NCDD primers and incorporation of radioactively labeled dATP. Radioactively labeled PCR products were purified by using a Sephadex G-100 minicolumn. Hybridization conditions were performed at 65°C as described in the protocol of Amersham (Pharmacia, Biotech Inc., Uppsala, Sweden). Final washing steps were done under stringent (0.5⫻ SSC[1⫻ SSC is 0.15 M NaCl plus 0.015 M sodium citrate]–0.1% sodium dodecyl sulfate [SDS]) or nonstringent (2⫻ SSC–0.1% SDS) conditions, both at 65°C. Overexpression and purification of CelC. CelC of Orpinomyces sp. strain PC-2 contains an NCDD repeat at the N terminus that is separated from the catalytic domain by a linker (22). To produce the enzyme with [CelC(⫹)] and without [CelC(⫺)] the NCDDs, two forward oligonucleotide primers, CLF (5⬘-TTTCT GCAGCTAGATGTCATCCAAGTTACCC-3⬘) and CLSF (5⬘-TTTCTGCAGCT AGTGATAACTTCTTTGAAAAT-3⬘), and one reverse primer, ACR (5⬘-GGA ATTCTTAGAATGGTGGGTTAGCGTTTT-3⬘), were designed and synthesized (Applied Biosystems). PstI and EcoRI restriction sites are in boldface. Primers CLF and CLSF correspond to amino acid residues RCHPSYP (positions 20 to 26) and SDNFFEN (positions 129 to 135), respectively, whereas ACR corresponded to residues ENANPPF (positions 443 to 449) of CelC (22). PCR was performed by using CLF and CLSF separately in combination with ACR and using the cDNA library as the template. Reagents for PCR were purchased from Roche Molecular Biochemicals, Indianapolis, Ind., except that Pfu polymerase was a product of Stratagene. A 480 Thermal Cycler (Perkin-Elmer) was used with 30 cycles of 95°C for 1 min, 50°C for 1 min, and 72°C for 1.5 min. PCR products were analyzed by 1.0% agarose gels and visualized by ethidium bromide staining. Bands with the correct sizes were excised from the gels, and DNA was purified by using the Geneclean kit (Bio 101). Plasmid pRSET B and the purified celC PCR products were digested by PstI and EcoRI. The digested samples were again purified by the Geneclean kit and ligated by using the Rapid Ligation kit (Roche Molecular Biochemicals). Transformation of E. coli JM109 and BL21(DE3) was done as described by Sambrook et al. (30), and transformants were grown on solid Luria-Bertani medium containing ampicillin at 100 ␮g/ml. Plasmids were purified by using the Qiaprep Spin Miniprep kit (Qiagen, Inc., Valencia, Calif.) and analyzed for the presence of the appropriate inserts by restriction digestion and DNA sequencing (Applied Biosystems). Growth of transformants, induction of celC constructs, disruption of E. coli cells, and purification of CelC(⫹) and CelC(⫺) were performed in accordance with the instructions supplied by Invitrogen, except that two additional purification steps were used to achieve purity of the proteins. Biotin labeling of the CelC expression products and binding analysis. Purified CelC(⫹) and CelC(⫺) proteins (100 ␮g) were labeled with biotin by using the Biotin Protein Labeling kit (Roche Molecular Biochemicals). Free biotin was removed with the gel filtration column provided with the kit. The polypeptides of the Orpinomyces (hemi)cellulase complex preparations were separated by SDSpolyacrylamide gel electrophoresis (PAGE) and transferred to nitrocellulose membranes. The membranes were blocked with 0.1% (wt/vol) bovine serum albumin in 50 mM sodium citrate, pH 6.0, for 4 h. Binding between the cellulosomal polypeptides and the biotinylated CelC protein was studied by incubating the membranes in biotinylated CelC preparations (10 ␮g/ml in 10 mM CaCl2–50 mM sodium citrate buffer, pH 6.0). Detection of bound biotinylated CelC was done with the Biotin Detection kit (Roche). Nucleotide sequence accession numbers. The nucleotide sequences of the cellulolytic enzymes determined in this study have been deposited in the GenBank/EMBL database under the following accession numbers: manA, AF177206; celH, AF177204; celI AF177205; celJ, AF177207.

5327

FIG. 3. (A) Agarose (3.0%, wt/vol) gel analysis of PCR products amplified by using the Orpinomyces sp. NCDD forward and reverse primers and an Orpinomyces cDNA library as the template. (B) Orpinomyces PCR control, without template. (C) Agarose (2.0%, wt/vol) gel analysis of PCR products amplified by using the Piromyces sp. NCDD forward and reverse primers and Piromyces sp. strain E2 genomic DNA as the template. Molecular sizes are given in base pairs on the right.

RESULTS AND DISCUSSION Evidence has been published showing that anaerobic fungi produce high-molecular-mass (hemi)cellulase complexes similar to cellulosomes of anaerobic bacteria. It has been suggested that the fungal cellulosomes have scaffoldins that bind enzymatic subunits through interactions between cohesins of the scaffoldins and NCDDs of the catalytic subunits (14, 15). Most of the genes encoding catalytic cellulosome components of anaerobic fungi consist of two NCDDs interspaced by a rather specific linker sequence, making each NCDD repeat unique. Furthermore, by using specific antibodies against NCDDs, it has been shown that many additional proteins of Orpinomyces and Neocallimastix spp. contain them (22). These observations encouraged us to use the conserved sequences of the repeated NCDDs to design degenerate primers for amplification of the unique sequences interspacing them. This resulted in the isolation of unique NCDD-specific DNA probes for use in the isolation of genes encoding NCDD-containing enzymes. Analysis and isolation of NCDD-based PCR products. PCR was performed as outlined in Materials and Methods by using the degenerate primers based on NCDD sequences from P. equi and Orpinomyces sp. and with genomic DNA and cDNA, respectively, as templates. The products were analyzed on agarose gels and were visible for both fungi as smears ranging from 110 to 200 bp (Fig. 3). The PCR products were directly ligated and transformed into E. coli to obtain specific NCDDPCR clones. Sixty-nine and 45 clones were sequenced for Piromyces and Orpinomyces, respectively. According to the expectations of this PCR-based strategy, among these sequences were those that completely matched already known NCDDs of previously isolated genes. NCDD-PCR products were regarded as novel when they fulfilled the following criteria. (i) PCR products should contain both primers and encode the typical vicinal cysteine residues of NCDD sequences (i.e., LGYPCC or QGYKCC). (ii) PCR products were expected to

5328

STEENBAKKERS ET AL.

be a coding sequence. In the case of the Piromyces manA gene, this would predict a PCR product of 119 bp. NCDD-PCR products should have a size of 119 bp or sizes that differ by multiples of 3 bp. (iii) For sequence comparison, a general indication for sequence variation was deduced from an alignment of published NCDDs from P. equi. This alignment (not shown) indicates that the two theoretical PCR products from the manA gene from P. equi are equal in size and have a nucleotide and amino acid sequence identity of approximately 90%. Consequently, in the case of identical sizes, NCDD-PCR clones were regarded as different when they showed a sequence diversity of more than 10%. (iv) Because of the presence of a highly conserved stretch within the NCDD-PCR products, the sequences were checked for the formation of chimerical products. All clones fulfilled the above criteria and showed sequence homology with NCDDs of known enzymes of anaerobic fungi. In addition, the clones had GC contents ranging from 40 to 45%, indicating that they represent coding regions. Noncoding regions of DNA of anaerobic fungi have a remarkably low GC content of approximately 5%. Introns, which are rare in genes from anaerobic fungi (23), were not found. Among the clones, several showed small sequence variations (⬍10%). These variations may have occurred during amplification, but they may also be regarded as NCDD representatives of recently duplicated genes (6, 22, 26) or as the result of allelic variation. Distinctive or novel NCDD sequences. A comparison of NCDD-PCR sequences of 69 clones from Piromyces sp. strain E2 and 45 clones from Orpinomyces sp. strain PC-2 revealed that 19 and 16 distinctive NCDD-PCR clones were obtained, respectively. Of these, one from Piromyces (NCDD-PCR10, a family 6 cellulase [unpublished data]) and seven from Orpinomyces (celA to celE and xynA) (7, 21, 22, and unpublished data) were previously identified. The recognition of previously identified sequences clearly demonstrated that the cloning strategy was successful and served as a positive control. When the PCR with gDNA as a template was performed at an annealing temperature of 60°C instead of 50°C, a strong selection could be observed of those PCR products that contained the primers with the highest GC content. The NCDD-PCR product of a family 6 cellulase, which served as a positive control, could not be isolated under these conditions. The large number of novel NCDD-PCR clones isolated indicates that the region within the NCDDs is indeed highly conserved among NCDD-containing genes of anaerobic fungi and also that a large number of NCDDs in enzymes of anaerobic fungi contain at least two NCDDs. Only estA from P. equi (15) and celB2 from O. joyonii (28) are examples of genes encoding catalytic components of the cellulosome containing only one NCDD. Also two other Piromyces sp. genes, both encoding family 5 catalytic domains, contain a single N-terminal NCDD but also an NCDD repeat at the C terminus (13; unpublished data; EMBL protein database accession no. AAD43818). Sequence alignment of translated NCDD-PCR products. The NCDDs of anaerobic fungi have an exceptionally high cysteine content and typically contain 3 to 6 cysteine residues and as many as 6 to 12 in an NCDD repeat. Despite the similar functions of bacterial dockerins and fungal NCDDs, they share no sequence homology. C. thermocellum cellulosomal dockerins, for instance, contain no, one, or a few cysteine residues

J. BACTERIOL.

(8). Cysteine residues are generally known to be involved in disulfide bridges, which significantly influence the three-dimensional structure of proteins. Cysteine residues are also involved in the proper folding and secretion of assembled extracellular oligomeric proteins (29). Because of the relatively high number and the general importance of cysteine residues, we aligned the deduced amino acid sequences of the NCDDPCR clones from both Piromyces sp. strain E2 and Orpinomyces sp. strain PC-2 (Fig. 4). All of the sequences start with the C-terminal end of the first NCDD, followed by the interspacing linker and the second NCDD. The alignments reveal that the NCDD-PCR clones, encoding partial NCDD-linker sequences from Piromyces sp. and Orpinomyces sp. can be divided into three sequence groups, NCDD-PCR groups A to C, although the alignment was less consistent for the sequences of the latter fungus. The coding sequences of the Orpinomyces NCDD-PCR clones were larger because of the complementary primers used, which annealed to a smaller part of the NCDD repeatcontaining gene. All of the Orpinomyces sp. sequences encoded WCG immediately after the primer encoded sequence (except for NCDD-PCR12), indicating that the conserved stretch within NCDDs probably extends beyond the sequence that was used to design the Orpinomyces sp. primers. The sequence downstream of WCG was used to identify NCDD-PCR sequence groups A to C. All of the NCDD-PCR sequences encoded three to five cysteine residues, of which two were vicinal cysteines, except for NCDD-PCR12 (Fig. 4). NCDD-PCR group A has a total of five cysteines (except for NCDD-PCR29), as does NCDD-PCR group C, and typically starts with CG. Group C begins with a conserved isoleucine residue followed by three other amino acids before the first cysteine. NCDD-PCR group B contains sequences with a total of three cysteines and also starts with a conserved isoleucine. The first cysteine of the group B sequences is followed by a highly conserved tryptophan residue, and the vicinal cysteines are preceded by a highly conserved stretch, LGYP. A tyrosine is also conserved in the other two groups, but the context for group A sequences is typically QGYKCC, whereas group C sequences show a much lower degree of conservation. Groups A and C have the fifth cysteine residue only two to four amino acids beyond the cysteine pair. The final part extending to the reverse primer sequence is conserved for all three NCDD-PCR groups. It consists of 8 to 12 amino acids and starts with one or two valines, followed by a conserved tyrosine residue and two conserved aspartates at the C-terminal end. The conserved aspartate residues were previously reported to resemble a part of the consensus for a calcium-binding motif that is present in clostridial dockerins (3). The clostridial dockerin and cohesin interaction requires calcium (8, 9, 40). This could indicate a similar calcium dependency of interactions between NCDDs and scaffoldins of the fungal cellulosomes, but this has yet to be investigated. Sequence alignment and classification of fungal NCDDs. Because of the recent publications of estA containing a functional N-terminal NCDD (15), cel5A with a single N-terminal NCDD and a C-terminal NCDD repeat from P. equi and Piromyces rhizinflata (13; unpublished data; EMBL protein database accession no. AAB69348), and celB2 with a single Cterminal NCDD (28), there are strong indications that one

VOL. 183, 2001

NONCATALYTIC DOCKING DOMAINS OF ANAEROBIC FUNGI

5329

FIG. 4. Translated NCDD-PCR clones isolated from Piromyces sp. strain E2 (roman) and Orpinomyces sp. strain PC-2 (italic). The alignment with fixed cysteine residues reveals a division into sequence groups A to C. The first and last amino acids are encoded by the degenerate primers and are underlined. The presumed linker region is boxed. Cysteine residues are in boldface, and conserved residues are in shaded boxes. Spaces are indicated by dashes. The sequences that served as a positive control are in boldface.

NCDD is capable of interacting with a hypothetical scaffoldin protein. This implies that the classification of the isolated NCDD-PCR products reflects the combination of functional NCDDs. For this reason, an alignment was made of all accessible NCDDs from anaerobic fungi, again based on the num-

ber, position, and context of cysteine residues (Fig. 5). The alignment revealed that all NCDD sequences can be divided into three different types. Type 2 NCDDs contain four cysteines, and types 1 and 3 contain six cysteines. The three different NCDD types can be used to fully explain NCDD-

FIG. 5. Alignment of all accessible NCDD sequences from anaerobic fungi (except manB [⬇manC] from P. equi and celE [⬇celB] from Orpinomyces sp. strain PC-2), indicating a three-subfamily division. From left to right: organism (Pir, Piromyces; Neo, Neocallimastix; Orp, Orpinomyces; eq., equi; rh., rhizinflata; pa., patriciarum; fr., frontalis; jo., joyonii; pc, strain PC-2), database accession number, name of gene, NCDD number (N terminus to C terminus). Cysteine residues are in boldface and boxed. Conserved residues are in boxes shaded dark grey for residues present in more than 90% of the sequences and light grey for residues present in more than 50% of the sequences. Spaces are indicated by dashes. An asterisk indicates that the data were taken from a partial cDNA sequence. 5330

VOL. 183, 2001

NONCATALYTIC DOCKING DOMAINS OF ANAEROBIC FUNGI

FIG. 6. SDS-PAGE and Western blot analysis of Orpinomyces CelC(⫹) and CelC(⫺). Purified forms of CelC were denatured in SDS buffer and analyzed by using 10% (wt/vol) acrylamide gels. Gels were stained with Coomassie brilliant blue (A) and subjected to Western blotting, followed by incubation with anti-NCDD polyclonal antibodies (B) (21). Lanes 1 and 2 were loaded with CelC(⫹) and CelC(⫺), respectively. The values on the left are molecular sizes in kilodaltons.

PCR sequence groups A to C, indicating the validity of the division of NCDD sequences into three different types. It seems that NCDD-PCR group A reflects the PCR products of an NCDD type 1 repeat, NCDD-PCR group B reflects the products of an NCDD type 2 repeat, and NCDD-PCR group C reflects the products of an NCDD type 3 repeat. The NCDDPCR products that do not exactly fit into the group classification with regard to their cysteine residues can also be explained. PCR product NCDD-PCR29 seems to be a combination of NCDD type 1 with NCDD type 2, and NCDDPCR2 seems to be a combination of types 2 and 3. The NCDDPCR products from celB and celE are a combination of NCDD types 3 and 2. The type 1 NCDDs seem to be confined to hemicellulolytic enzymes. The possible implications of the existence of these three structural NCDD types remain to be determined. The NCDDs of Orpinomyces CelC bind to multiple polypeptides of the cellulosomal complex of anaerobic fungi. The only previous evidence of a scaffoldin in anaerobic fungi and the possible involvement of NCDD-like sequences was obtained by Fanutti et al. (14) and Fillingham et al. (15). They showed that one, two, or three NCDDs coupled to a reporter protein bind to a secreted protein in the culture fluids of the monocentric fungi P. equi and Neocallimastix sp., as well as in a cellulosomal preparation. To investigate if the NCDDs from the polycentric fungus Orpinomyces sp. share the same function and to detect hypothetical scaffoldin proteins from this fungus, the binding of two CelC constructs to immobilized cellulosomal protein was investigated. The CelC constructs were expressed in E. coli and purified to homogeneity, one containing the original NCDD repeat, CelC(⫹), and one without the repeat, CelC(⫺) (22). The molecular weights of both purified polypeptides matched the calculated values (Fig. 6A). CelC(⫹), but not the CelC(⫺) protein, reacted to antibodies raised against XynA NCDDs (Fig. 6B), indicating that the CelC(⫹) protein indeed contained NCDDs and that the NCDDs of XynA and CelC

5331

FIG. 7. Analysis of CelC NCDD binding to the Orpinomyces (hemi)cellulase complex. The purified complex sample was boiled in SDS buffer, and the denatured polypeptides were separated by SDSPAGE using 10% (wt/vol) acrylamide. Gels were stained with Coomassie brilliant blue (A) or labeled by using biotinylated full-length CelC (B) and CelC minus the NCDD repeat (C). The values on the left are molecular sizes in kilodaltons. The arrowheads indicate the scaffoldin polypeptides.

have epitopes in common. Both forms of CelC labeled with biotin were used to detect polypeptides in the cellulosomal complex. Only the CelC(⫹) protein containing the NCDDs (Fig. 7) bound to the immobilized cellulosomal proteins. The isolated complex visualized after SDS-PAGE contains perhaps as many as 20 different polypeptides (Fig. 7, lane A); of these, four bound CelC(⫹) whereas none bound CelC(⫺) (Fig. 7, lanes B and C, respectively). The molecular masses of the NCDD-binding polypeptides, as indicated by the SDS-PAGE, were 64, 66, 95, and 130 kDa. Our finding of four NCDDbinding polypeptides contrasts with the observations of Fanutti et al. (14) and Fillingham et al. (15), who detected only one 97-kDa NCDD-binding polypeptide. The lower bands may represent proteolytic products of the 130-kDa band. However, due to the fact that the intensities of the bands did not change after 6 months of storage of the complexes at 4°C, this seems unlikely. However, the NCDD sequences from P. equi, which were part of the GST fusion proteins used to detect scaffoldins, according to the proposed classification, are classified as type 1 (xynA) and 2 (estA) NCDDs. The sequence encoding a protein of unknown function used for the GST fusion protein containing three NCDDs was not reported (15). The NCDDs from CelC are classified as NCDD type 3. A possible explanation would be that the type 3 NCDDs recognize multiple or other scaffoldins. Therefore, the possibility exists that cellulosomes from anaerobic fungi have several different scaffoldin polypeptides. Hybridization of NCDD clones to genes encoding (hemi)cellulases of cellulosomes of anaerobic fungi. The PCR directed against the NCDD conserved region, followed by specific hybridization, makes it possible to develop a new strategy by which to specifically clone genes encoding (hemi)cellulases present in the cellulosomal complexes from anaerobic fungi. Besides the intrinsic speed of the PCR, the use of this ap-

5332

STEENBAKKERS ET AL.

proach will give definite advantages over conventional strategies, since this strategy is specifically suited to the isolation of genes containing the NCDD conserved region. Roughly 70% of the extracellular (hemi)cellulolytic activities of Piromyces sp. strain E2, after growth on filter paper, are present as individual enzymes (10, 11) that, in analogy to their clostridial counterparts, probably do not contain the NCDD sequences. A standard activity screening of a genomic or cDNA library will not discriminate between these individual and complexed (hemi) cellulases. Furthermore, the PCR of the unique linker region between NCDDs results in one PCR product for every (hemi) cellulase containing an NCDD repeat. In addition, there are indications that anaerobic fungi contain hydrolase families with duplicate genes. The examples of duplicated genes are manB and manC of P. equi (26) and celA and celC (22), as well as celB and celE (6, 21), of Orpinomyces sp. strain PC-2. By using the above-mentioned criteria, the isolation of members of the same multigene family or isolation of allelic counterparts can be excluded. Also, because most activity screening is performed with E. coli as an expression host, this DNA approach will prevent selection against those enzymes produced by the anaerobic fungus that depend on eukaryotic transcription and translation machinery for proper folding. To see if the isolated, novel NCDD-PCR clones could be used as specific probes to screen genomic or cDNA libraries, we tested hybridization and washing conditions. The 19 Piromyces sp. strain E2 clones were reamplified and separated on a 2% agarose gel and finally transferred to a nitrocellulose membrane. The smallest and largest clones isolated were radioactively labeled and used to investigate hybridization (NCDD-PCR8 and NCDD-PCR19, respectively, in Fig. 4). Clone NCDD-PCR19 already hybridized specifically after nonstringent washing conditions (2⫻ SSC–0.1% SDS, 65°C). Clone NCDD-PCR8, however, showed strong cross-reactivity with NCRPD-PCR14, even under the most stringent conditions tested (0.5⫻ SSC–0.1% SDS, 65°C), indicating that the hybridization temperature (65°C) should be optimized (data not shown). Four different NCDD-PCR probes were used successfully in a DNA hybridization screening of the cDNA library of the anaerobic fungus Orpinomyces sp. strain PC-2 constructed in E. coli (39). Four new cDNA sequences corresponding to hydrolytic enzymes not previously reported were isolated by this procedure. The first cDNA sequence isolated was identified as encoding a family 5 mannanase, based on the homology between its deduced amino acid sequence and a mannanase from aerobic fungi. The manA cDNA was 1,924-bp long. The second (celH) and third (celI) cDNA sequences isolated were classified, based on homology of the deduced amino acid sequence, as encoding cellulases belonging to family 6 of glycosyl hydrolases. The total length of the celH cDNA was 1,712 bp, and that of the celI cDNA was 1,784 bp. CelH and CelI shared 84% amino acid identity. The fourth cDNA sequence isolated (celJ) appeared to be homologous to endoglucanases from bacteria (family 5 of glycosyl hydrolases). This was only a partial sequence, lacking the region corresponding to the N terminus. Because of the large amount of sequence data collected, this strategy can be supplemented to directly correlate PCR products with the specific proteins visible in a cellulosome preparation. Wu et al. (38) described a method of chemically cleav-

J. BACTERIOL.

ing proteins at the N-terminal peptide bond of cysteine residues to determine which cysteine residues within a polypeptide are involved in disulfide bridges. The chemically produced peptides are analyzed by matrix-assisted laser desorption ionization–time of flight mass spectrometry, and the molecular masses are compared to calculated values and used to identify the cysteines involved. This same method can be applied to identify which PCR products partially encode the various proteins visible in a cellulosome preparation. Cellulosomal proteins can be excised from an SDS-PAGE gel and subjected to chemical cysteine cleavage, followed by matrixassisted laser desorption ionization–time of flight mass spectrometry analysis. The molecular masses can be compared to the calculated values predicted by the NCDD-PCR sequences. The corresponding PCR product can be identified, which can be used to isolate the gene encoding the cellulosome component. Although the method presented does not make a direct correlation between genes and cellulosomal proteins, the strategy described has proven to be a very swift method by which to isolate a large number of previously unknown sequences, partially encoding components of the fungal cellulosome, which can be used to isolate full-length genes. This method has been developed for fungal cellulosomes, but the same strategy can be applied for the isolation of genes encoding components of bacterial cellulosomes. The finding of 16 and 19 PCR products encoding novel NCDD repeats indicates at least as many polypeptides in the cellulosomes of the polycentric fungus Orpinomyces sp. strain PC-2 and the monocentric fungus Piromyces sp. strain E2. These results seem to agree with the number of identified components in the cellulosome of Clostridium species. Apparently, convergent evolution has led to similar functions of the cellulosomes of anaerobic fungi and anaerobic bacteria, despite differences in structure. ACKNOWLEDGMENTS This investigation was done at The University of Georgia and partly supported by contract DE-FG05-93ERZ0217 from the U.S. Department of Energy and partly by a Technology Development Partnership between the Georgia Research Alliance and Aureozyme, Inc., Atlanta, Ga. We acknowledge Jelle Eygenstein for analyzing sequencing samples and Jan Keltjens for valuable discussion on cysteine residues. REFERENCES 1. Aylward, J. H., K. S. Gobius, G.-P. Xue, G. D. Simpson, and B. P. Dalrymple. 1999. The Neocallimastix patriciarum cellulase, CelD, contains three almost identical domains with high specific activities on Avicel. Enzyme Microb. Technol. 24:609–614. 2. Bayer, E. A., S.-Y. Ding, Y. Shoham, and R. Lamed. 1999. New perspectives in the structure of cellulosome-related domains from different species, p. 428–436. In K. Ohmiya, K. Hayashi, K. Sakka, Y. Kobayashi, S. Karita, and T. Kimura (ed.), Genetics, biochemistry and ecology of cellulose degradation. Uni Publishers Co., Ltd., Tokyo, Japan. 3. Bayer, E. A., E. Morag, R. Lamed, S. Yaron, and Y. Shoham. 1998. Cellulosome structure: four-pronged attack using biochemistry, molecular biology, crystallography and bioinformatics, p. 39–65.In M. Claeyssens, W. Nerinckx, and K. Piens (ed.), Carbohydrolases from Trichoderma reesei and other microorganisms. Royal Society of Chemistry, London, England. 4. Borneman, W. S., D. E. Akin, and L. G. Ljungdahl. 1989. Fermentation products and plant cell wall-degrading enzymes produced by monocentric and polycentric anaerobic ruminal fungi. Appl. Environ. Microbiol. 55:1066– 1073. 5. Brownlee, A. G. 1994. The nucleic acids of anaerobic fungi, p. 241–256. In D. O. Mountford and C. G. Orpin (ed.), Anaerobic fungi. Marcel Dekker, Inc., New York, N.Y. 6. Chen, H., X. L. Li, D. L. Blum, and L. G. Ljungdahl. 1998. Two genes of the

VOL. 183, 2001

7.

8.

9. 10.

11.

12. 13.

14.

15.

16.

17.

18. 19. 20. 21. 22.

23.

NONCATALYTIC DOCKING DOMAINS OF ANAEROBIC FUNGI

anaerobic fungus Orpinomyces sp. strain PC-2 encoding cellulases with endoglucanase activities may have arisen by gene duplication. FEMS Microbiol. Lett. 159:63–68. Chen, H., X. L. Li, and L. G. Ljungdahl. 1995. A cyclophilin from the polycentric anaerobic rumen fungus Orpinomyces sp. strain P-2 is highly homologous to vertebrate cyclophilin B. Proc. Natl. Acad. Sci. USA 92:2587– 2591. Choi, S. K., and L. G. Ljungdahl. 1996. Dissociation of the cellulosome of Clostridium thermocellum in the presence of ethylenediaminetetraacetic acid occurs with the formation of truncated polypeptides. Biochemistry 35:4897– 4905. Choi, S. K., and L. G. Ljungdahl. 1996. Structural role of calcium for the organization of the cellulosome of Clostridium thermocellum. Biochemistry 35:4906–4910. Dijkerman, R., M. B. F. Vervuren, H. J. M. Op den Camp, and C. van der Drift. 1996. Adsorption characteristics of cellulolytic enzymes from the anaerobic fungus Piromyces sp. strain E2 on microcrystalline cellulose. Appl. Environ. Microbiol. 62:20–25. Dijkerman, R., H. J. M Op den Camp, C. Van der Drift, and G. D. Vogels. 1997. The role of the cellulolytic high molecular mass (HMM) complex of the anaerobic fungus Piromyces sp. strain E2 in the hydrolysis of microcrystalline cellulose. Arch. Microbiol. 167:137–142. Ding, S.-Y., E. A. Bayer, D. Steiner, Y. Shoham, and R. Lamed. 2000. A scaffoldin of the Bacteroides cellulosolvens cellulosome that contains 11 type II cohesins. J. Bacteriol. 182:4915–4925. Eberhardt, R. Y., H. J. Gilbert, and G. P. Hazlewood. 2000. Primary sequence and enzymic properties of two modular endoglucanases, Cel5A and Cel45A, from the anaerobic fungus Piromyces equi. Microbiology 146:1999– 2008. Fanutti, C., T. Ponyi, G. W. Black, G. P. Hazlewood, and H. J. Gilbert. 1995. The conserved non-catalytic 40-residue sequence in cellulases and hemicellulases from anaerobic rumen fungi functions as a docking domain. J. Biol. Chem. 270:29314–29322. Fillingham, I. J., P. A. Kroon, G. Williamson, H. J. Gilbert, and G. P. Hazlewood. 1999. A modular cinnamoyl ester hydrolase from the anaerobic fungus Piromyces equi acts synergistically with xylanase and is part of a multiprotein cellulose-binding cellulase-hemicellulase complex. Biochem. J. 343:215–224. Fujino, Y., K. Ogata, T. Nagamine, and K. Ushida. 1998. Cloning, sequencing, and expression of an endoglucanase gene from the rumen anaerobic fungus Neocallimastix frontalis MCH3. Biosci. Biotechnol. Biochem. 62:1795–1798. Gerngross, U. T., M. P. M. Romaniec, T. Kobayashi, N. S. Huskisson, and A. L. Demain. 1993. Sequencing of a Clostridium thermocellum gene (cipA) encoding the cellulosomal SL-protein reveals an unusual degree of internal homology. Mol. Microbiol. 8:325–334. Gilbert, H. J., G. P. Hazlewood, J. I. Laurie, C. G. Orpin, and G. P. Xue. 1992. Homologous catalytic domains in a rumen fungal xylanase: evidence for gene duplication and prokaryotic origin. Mol. Microbiol. 6:2065–2072. Kohring, S., J. Wiegel, and F. Mayer. 1990. Subunit composition and glycosidic activities of the cellulase complex from Clostridium thermocellum JW20. Appl. Environ. Microbiol. 56:3798–3804. Leibovitz, E., and P. Be´guin. 1996. A new type of cohesin domain that specifically binds the dockerin domain of the Clostridium thermocellum cellulosome-integrating protein CipA. J. Bacteriol. 178:3077–3084. Li, X.-L., H. Chen, and L. G. Ljungdahl. 1997. Monocentric and polycentric anaerobic fungi produce structurally related cellulases and xylanases. Appl. Environ. Microbiol. 63:628–635. Li, X.-L., H. Chen, and L. G. Ljungdahl. 1997. Two cellulases, CelA and CelC, from the polycentric anaerobic fungus Orpinomyces strain PC-2 contain N-terminal docking domains for a cellulase-hemicellulase complex. Appl. Environ. Microbiol. 63:4721–4728. Ljungdahl, L. G., X.-L. Li, and H. Chen. 1998. Evidence in anaerobic fungi

24. 25. 26.

27. 28. 29. 30. 31. 32.

33. 34. 35.

36.

37.

38.

39.

40. 41.

5333

of transfer of genes between them and from aerobic fungi, bacteria and animal hosts, p. 187–197. In J. Wiegel and W. W. Adams (ed.), Thermophiles: the keys to molecular evolution and the origins of life. Taylor and Frances Inc., Philadelphia, Pa. Lytle, B., and J. H. D Wu. 1998. Involvement of both dockerin subdomains in assembly of Clostridium thermocellum cellulosome. J. Bacteriol. 180:6581– 6585. Mechaly, A., S. Yaron, R. Lamed, H. P. Fierobe, A. Belaich, J. P. Belaich, Y. Shoham, and E. A. Bayer. 2000. Cohesin-dockerin recognition in cellulosome assembly: experiment versus hypothesis. Proteins 39:170–177. Millward-Sadler, S. J., J. Hall, G. W. Black, G. P. Hazlewood, and H. J. Gilbert. 1996. Evidence that the Piromyces gene family encoding endo-1,4mannases arose through gene duplication. FEMS Microbiol. Lett. 141:183– 188. Orpin, C. G. 1975. Studies on the rumen flagellate Neocallimastix frontalis. J. Gen. Microbiol. 91:249–262. Qiu, X., B. Selinger, L.-J. Yanke, and K.-J. Cheng. 2000. Isolation and analysis of two cellulase cDNAs from Orpinomyces joyonii. Gene 245:119– 126. Reddy, P. S., and R. B. Corley. 1998. Assembly, sorting, and exit of oligomeric proteins from the endoplasmic reticulum. Bioessays 20:546–554. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. Selinger, L. B., C. W. Forsberg, and K.-J. Cheng. 1996. The rumen: a unique source of enzymes for enhancing livestock production. Anaerobe 2:263–284. Shimon, L. J. W., F. Frolow, S. Yaron, E. A. Bayer, R. Lamed, E. Morag, and Y. Shoham. 1997. Crystallization and preliminary X-ray analysis of a cohesin domain of the cellulosome from Clostridium thermocellum. Acta Crystallogr. Sect. D 53:114–115. Shoham, Y., R. Lamed, and E. A. Bayer. 1999. The cellulosome concept as an efficient microbial strategy for the degradation of insoluble polysaccharides. Trends Microbiol. 7:275–281. Tavares, G. A., P. Be´guin, and P. M. Alzari. 1997. The crystal structure of a type I cohesin domain at 1.7 Å resolution. J. Mol. Biol. 273:701–713. Teunissen, M. J., H. J. M. Op den Camp, C. G. Orpin, J. H. J. Huis in’t Veld, and G. D. Vogels. 1991. Comparison of growth characteristics of anaerobic fungi isolated from ruminant and non-ruminant herbivores during cultivation in a defined medium. J. Gen. Microbiol. 137:1401–1408. Teunissen, M. J., A. A. M. Smits, H. J. M. Op den Camp, J. H. J. Huis in’t Veld, and G. D. Vogels. 1991. Fermentation of cellulose and production of cellulolytic and xylanolytic enzymes by anaerobic fungi from ruminant and non-ruminant herbivores. Arch. Microbiol. 156:290–296. Wilson, C. A., and T. M. Wood. 1992. The anaerobic fungus Neocallimastix frontalis: isolation and properties of a cellulosome-type enzyme fraction with the capacity to solubilize hydrogen-bond-ordered cellulose. Appl. Microbiol. Biotechnol. 37:125–129. Wu, J., D. A. Cage, and J. T. Watson. 1996. A strategy to locate cysteine residues in proteins by specific chemical cleavage followed by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Anal. Biochem. 235:161–174. Ximenes, E. A. 1999. Cloning and sequencing of cDNAs encoding one ␤-glucosidase, three cellulases and one mannanase of the fungus Orpinomyces sp. strain PC-2, and over-expression of the ␤-glucosidase and mannanase in Saccharomyces cerevisiae. Ph.D. thesis. University of Brasilia, Brasilia, Brazil. Yaron, S., E. Morag, E. A. Bayer, R. Lame, and Y. Shoham. 1995. Expression, purification and subunit-binding properties of cohesins 2 and 3 of the Clostridium thermocellum cellulosome. FEBS Lett. 360:121–124. Zhou, L., G.-P. Xue, C. G. Orpin, G. W. Black, H. J. Gilbert, and G. P. Hazlewood. 1994. Intronless celB from the anaerobic fungus Neocallimastix patriciarum encodes a modular family A endoglucanase. Biochem. J. 297: 359–364.