Nội dung

Eur. J. Biochem. 271, 1266–1276 (2004)  FEBS 2004 doi:10.1111/j.1432-1033.2004.04031.x Heterogeneity of homologously expressed Hypocrea jecorina (Trichoderma reesei ) Cel7B catalytic module Torny Eriksson1,*, Ingeborg Stals2,*, Anna Collén1, Folke Tjerneld1, Marc Claeyssens2, Henrik Stålbrand1 and Harry Brumer3 1 Department of Biochemistry, Center for Chemistry and Chemical Engineering, Lund University, Sweden; 2Laboratory for Biochemistry, Department of Biochemistry, Physiology and Microbiology, Ghent University, Belgium; 3Department of Biotechnology, Royal Institute of Technology (KTH), AlbaNova University Centre, Stockholm, Sweden The catalytic module of Hypocrea jecorina (previously Trichoderma reesei) Cel7B was homologously expressed by transformation of strain QM9414. Post-translational modifications in purified Cel7B preparations were analysed by enzymatic digestions, high performance chromatography, mass spectrometry and site-directed mutagenesis. Of the five potential sites found in the wild-type enzyme, only Asn56 and Asn182 were found to be N-glycosylated. GlcNAc2Man5 was identified as the predominant N-glycan, although lesser amounts of GlcNAc2Man7 and glycans carrying a mannophosphodiester bond were also detected. Repartition of neutral and charged glycan structures over the two glycosylation sites mainly accounts for the observed microheterogeneity of the protein. However, partial deamidation of Asn259 and a partially occupied O-glycosylation site give rise to further complexity in enzyme preparations. The filamentous fungus Hypocrea jecorina (previously Trichoderma reesei [1]) produces several extracellular cellulases, which cooperate in the degradation of paracrystalline cellulose. The five known endoglucanases, including Cel7B, generally act by hydrolysing the b(1fi4) glucan chains internally [2,3], whereas the two cellobiohydrolases (Cel6A and Cel7A) release cellobiose from the nonreducing and reducing chain ends, respectively [4,5]. b-Glucosidase eventually hydrolyses this cellobiose to glucose which is taken up by the fungal hyphae. All but one of the Hypocrea jecorina cellulases share a similar modular structure comprised of a catalytic module connected to a carbohydrate-binding module (CBM) by a flexible linker peptide. The 3D structures of the catalytic modules of Cel7A (formerly cellobiohydrolase I, CBH I [6]) and Cel7B (formerly endoglucanase I, EG I) both exhibit a similar overall fold but are different in their active site topologies; the former has a tunnel-shaped active site whereas the latter possesses an open cleft [7,8]. This reflects their different specificity, i.e. exo vs. endo activity [8]. The catalytic modules of fungal glycoside hydrolases are often glycosylated on asparagine residues in the consensus sequence Asn-Xaa-(Ser/Thr), where Xaa is not Pro [9]. This post-translational modification is thought to affect protein secretion and enzyme stability [10,11]. Of the structures studied so far, most fungal N-glycans contain the mammalian-type core structure (Man3GlcNAc2) [12]. However, the occurrence of a single N-acetyl glucosamine on Cel7A from H. jecorina strains ALKO2877 and QM9414 [13,14], indicates glycosylation may be processed differently in some cases. The N-glycosylation of both Cel7A and Cel7B, isolated from different strains and grown under different conditions, has been studied by several groups, but disparate and inconclusive results have been published [8,13–18]. The Hypocrea jecorina Cel7B catalytic module (Swiss-Prot number P07981) possesses five potential N-glycosylation sites. Single N-acetyl glucosamine (GlcNAc) residues have been observed by X-ray crystallography on Asn56 and Asn182 of Cel7B produced in the H. jecorina strain QM9414 [8]. In a later study, this enzyme was suggested to carry only one high mannose N-glycan, some forms of which carried mannophosphodiester linkages [16]. O-Mannosylation was also indicated by this study, but the sites of attachment of this and the N-glycan were not determined. Recently, Cel7B from the H. jecorina strain Rut-C30 was shown to bear a single GlcNAc on Asn56, while Asn182 was occupied with higher-order glycans, primarily GlcNAc2Hex8 [18]. In the present study, we describe the glycoform analysis of the catalytic module of H. jecorina Cel7B homologously expressed in a QM9414 decendent strain using a range of experimental techniques. Detailed analysis using mass spectrometry and high-performance chromatography indicated that two of the five potential N-glycosylation sites of Correspondence to H. Brumer, Department of Biotechnology, Royal Institute of Technology (KTH), AlbaNova University Centre, S-106 91 Stockholm, Sweden. Fax: + 46 85537 8468, Tel.: + 46 85537 8367, E-mail: harry@biotech.kth.se Abbreviations: Endo H, Streptomyces plicatus endoglycosidase H; CID MS/MS, collision-induced dissociation tandem mass spectrometry; HPAEC-PAD, high-performance anion-exchange chromatography with pulsed amperometric detection; PAG-IEF, polyacrylamide gel isoelectric focusing. Enzyme: endoglycosidase H (EC 3.2.1.96). *Note: These authors contributed equally to this work. Note: A website is available at http://www.biotech.kth.se/ woodbiotechnology/ (Received 18 November 2003, revised 16 January 2004, accepted 6 February 2004) Keywords: protein glycosylation; O-glycan; N-glycan; Trichoderma reesei; cellulase.  FEBS 2004 Heterogeneity of H. jecorina (T. reesei) Cel7B (Eur. J. Biochem. 271) 1267 the enzyme were glycosylated with high-mannose structures, predominantly GlcNAc2Man5. Additional heterogeneity in the purified protein arises from a partially occupied O-glycosylation site, as well as from partial deamidation of asparagine. Materials and methods Enzyme production The gene sequence encoding the catalytic module (Glu1– Thr371) of H. jecorina Cel7B (Cel7Bcat) was expressed under the control of the gpdA promotor from Aspergillus nidulans as described previously [19] by transforming the vector pAC1 into H. jecorina (Trichoderma reesei) QM9414, to yield strain QM9414-Cel7Bcat. H. jecorina strain QM9414-Cel7BcatN182Q, expressing Cel7Bcat(Asn182Gln) under the regulation of the A. nidulans gpdA promotor, was constructed as follows. Site directed mutagenesis was carried out using the PCR, according to the QuickChange method (Stratagene, La Jolla, CA, USA) using native Pfu polymerase and vector pAC1 as the template. The following oligonucleotide primer was used: 5¢-CGTCCAGACATGGAGGcaaGGtACCCTCAACAC TAGC-3¢. Mismatches are indicated in lower case and the introduced KpnI restriction site used for screening of transformants is shown in bold. Amplified and purified plasmid preparations were screened using KpnI and two positives from 10 were found. The open reading frame of the construct was sequenced prior to transformation into H. jecorina QM9414 (gift from M. Penttilä, VTT Biotechnology, Espoo, Finland) to yield strain QM9414Cel7BcatN182Q. Transformation and selection was performed as described by Collen et al. [19], based upon the method described by Penttilä et al. [20]. The strains H. jecorina QM9414-Cel7Bcat and H. jecorina QM9414-Cel7BcatN182Q were cultivated in minimal medium with glucose as the sole carbon source according to Collén et al. [21], which is a modification of the medium used by Nakari-Setälä et al. [22] and Penttilä et al. [20]. The medium contained 30 gÆL)1 K2HPO4, 8 gÆL)1 KH2PO4, 4 gÆL)1 (NH4)2SO4, 0.6 gÆL)1 CaCl2, 0.6 gÆL)1 MgSO4, 5 mgÆL)1 FeSO47H2O, 1.6 mg L)1 MnSO4H2O, 1.4 mgÆL)1 ZnSO47H2O, 2 mgÆL)1 CoCl2 and 4% (w/v) glucose. The pH was adjusted to 6.0. The fermentation was performed in 1 L baffled shake-flasks with 200 mL medium at 28 C and 180 r.p.m. The glucose concentration was monitored daily as described previously [21] and was kept above 1% (w/v). After 7 days of cultivation, the mycelia were removed and the buffer was exchanged to 20 mM NH4OAc, pH 4.5 (Buffer A) by ultrafiltration. The proteins were purified by anion-exchange chromatography (Source Q; Amersham Pharmacia Biotech) using a linear gradient generated by mixing Buffer A with Buffer B (1 M NH4OAc, pH 4.5). All fractions containing significant activity toward p-nitrophenyl-b-cellobioside (measured as described in [23]) were pooled and used in further analyses. dry precast homogeneous polyacrylamide gel (3.8 cm · 3.3 cm). The gel was rehydrated with 120 lL PharmalyteTM pH 2.5–5 (Amersham Biosciences, Uppsala Sweden), 20 lL ServalytTM pH 3–7 (Serva Electrophoresis GmbH, Heidelberg, Germany) and 1860 lL bidistilled water for 2 h. In a prefocusing step, the pH gradient was generated (75 Vh, 2000 V, 2.5 mA) and 1 lL samples (10 mg proteinÆmL)1) were subsequently applied at the cathode end. Electrophoresis was started at low voltage (15 Vh) and run to a final 450 Vh (2.5 mA, 2000 V). At the end of the run the locations of Cel7B activity were revealed by immersing the gel in 2 mM 4-methylumbelliferyl b-lactoside (NaOAc buffer, pH 5). Staining with Coomassie Blue R-350 was performed according to the manufacturer’s instructions (Pharmacia Biotech). High-performance anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD) A HPAEC-PAD system (DionexTM, Sunnyvale, CA, USA), equipped with an ED40 electrochemical detector, a GP40 gradient pump and a LC30 chromatography oven (40 C) was used. Chromatographic data were analysed using DIONEX PEAKNET software (release 5.1). Monosaccharide mixtures resulting from total acid hydrolysis were analysed on a CarboPac PA-10 column using isocratic elution (16 mM NaOH, 1 mLÆmin)1). Enzymatically released N-glycans were separated on a CarboPac PA-100 column; neutral oligosaccharides were first resolved using a 0–60 mM NaOAc gradient in 100 mM NaOH for 35 min (1 mLÆmin)1). A 60–500 mM NaOAc gradient in 100 mM NaOH was subsequently applied to elute carbohydrates carrying negatively charged substituents. Mass spectrometry Mass spectrometric analysis was carried out on a Q-TofTM II mass spectrometer fitted with a nano Z spray source (Waters Corporation, Micromass MS Technologies, Manchester, UK), essentially as described previously [24]. Endoglycosidase H digestion Enzymatic N-deglycosylation was performed by adding 0.02 U Endo H (Sigma-Aldrich, Bornem, Belgium) per microgram of Cel7Bcat or Cel7BcatN182Q in 10 mM NaOAc buffer, pH 4.5, for 12 h at 37 C. N-Deglycosylated proteins were subsequently precipitated with three volumes of ethanol and were redissolved in bidistilled water prior to PAG-IEF analysis. For carbohydrate analysis, the supernatant was desalted [25] on a Carbograph column (Alltech Associates Inc., Lokeren, Belgium). After extensive washing with bidistilled water, N-glycans were eluted with 2 mL 25% (v/v) CH3CN(aq) containing 0.05% (v/v) trifluoroacetic acid. Following evaporation of the solvent, the N-glycans were redissolved in bidistilled water for further analysis. Alkaline phosphatase treatment Polyacrylamide gel isoelectric focusing (PAG-IEF) PAG-IEF experiments were performed with a PhastSystemTM (Amersham Biosciences, Uppsala, Sweden) using a Enzymatic dephosphorylation of released N-glycans was attempted on both untreated and mild acid-hydrolysed samples (0.01 M HCl, 100 C, 30 min) as follows. One unit 1268 T. Eriksson et al. (Eur. J. Biochem. 271)  FEBS 2004 of calf intestine alkaline phosphatase (Roche Diagnostics, Vilvoorde, Belgium) dissolved in 20 lL 100 mM Tris/HCl, pH 8.8 containing 10 mM ZnCl2, was added to 20 lL oligosaccharides (25 lgÆmL)1). Reactions were allowed to proceed overnight at room temperature prior to product analysis by HPAEC-PAD. a-Mannosidase treatment Jack bean mannosidase (1 unit; Sigma-Aldrich, Bornem, Belgium) was added to oligosaccharide mixtures (20 lL, 25 lgÆmL)1) obtained from Cel7Bcat or the reference protein RNAse B in 20 mM NaOAc buffer, pH 5, containing 2 mM ZnCl2. The products of the overnight reaction at room temperature were analysed by HPAEC-PAD. Total acid hydrolysis Oligosaccharide samples (25 lgÆmL)1) were hydrolysed in 4 M trifluoroacetic acid. After heating at 100 C for 4 h in Teflon capped tubes, the acid was removed by evaporation and the sugars were identified by HPAEC-PAD. Fig. 1. Primary amino acid sequence of H. jecorina Cel7Bcat. Labels T1–T16 and S1–S11 denote predicted peptides from trypsin or S. aureus V8 protease digestion, respectively. Predicted N-glycosylation sites are shown in bold italic type, with Asn underlined. Q, Glnderived pyroglutamate; Cys residues are highlighted in bold. Protease digestions Prior to protease digestion, proteins were denatured and reduced by incubating 0.8 mgÆmL)1 Cel7Bcat or 0.05 mg mL)1 Cel7BcatN182Q in 0.1 M NH4HCO3, containing 6 M urea and 5 mM dithiothreitol, for 30 min at 60 C. Iodoacetamide (25 mM final concentration) was added and the samples were incubated in the dark for 30 min at 25 C. Subsequent dialysis was performed against 1 M urea either in 10 mM NH4HCO3 (for trypsin digestions) or in 10 mM sodium phosphate buffer, pH 7.5 (for V8 protease digestions). Modified trypsin (Promega, Madison, WI, USA) was added in a protease/cellulase ratio of 1/20 (w/w; 37 C, 12 h). V8 protease digestions were performed by incubating Staphylococcus aureus V8 protease (V8 endoproteinase Glu-C; Sigma, St Louis, MO, USA) in a 1/20 (w/w) protease/cellulase ratio (37 C, 12 h). Results Protein expression and purification The gene sequence encoding Glu1–Thr371 of Hypocrea jecorina Cel7B (Cel7Bcat, Fig. 1) was homologously expressed under the regulation of the gpdA promotor of Aspergillus nidulans [19]. After cultivation (7 days) the endoglucanase activity was 0.63 nkatÆmL)1, corresponding to an extracellular expression level of 27 mgÆL)1. The pH was 5.5 at the start of the cultivation and decreased to approximately 3 at days six and seven of growth. Purification by anion-exchange chromatography yielded several endoglucanase-active peaks, which suggested the presence of protein isoforms (data not shown). These combined fractions were used in further experiments to ensure that all produced isoforms of the protein were analysed. SDS/ PAGE analysis indicated a major protein band with an apparent molecular mass of 44 ± 1 kDa (data not shown), which is higher than that calculated for Cel7Bcat (39.1 kDa). The results from the cultivation of H. jecorina expressing Cel7BcatN182Q were similar to that of Cel7Bcat, except that the detected endoglucanase activity in the cultivation broth was lower (0.18 nkatÆmL)1). PAG-IEF analysis of intact and Endo H digested Cel7Bcat After PAG-IEF over a narrow pH gradient, at least five enzymatically active isoforms were observed. Analysis of a sample treated with endoglycosidase H (Endo H), which cleaves the core GlcNAcb(1fi4)GlcNAc bond in high mannose-type N-glycans, yielded one predominant and one minor isoform (Fig. 2). ESI-MS analyses of intact and Endo H-digested Cel7Bcat X-ray crystallography has previously revealed that the N-terminal residue in Cel7B is pyroglutamate and that the protein contains eight disulphide bonds [8]. TOF MS analysis of tryptic digests confirmed the presence of the N-terminal pyroglutamate in the protein produced in this study (Table 1). After correction for these post-translational modifications, the calculated molecular mass of Cel7Bcat is 39133.8 Da. Figure 3A shows the reconstructed zerocharge spectrum of purified Cel7Bcat, in which a range of peaks is observed. The mass of the major component corresponds well with the calculated molecular mass for Cel7Bcat substituted with Man5GlcNAc2 on two Asn residues (calculated molecular mass, 41 567.6 Da; observed molecular mass 41 566.6 Da). The observation of species of increasing mass spaced by 162 Da reflects the presence of glycoforms with an increasing number of hexose (probably mannose) units. Phosphorylation is indicated by the presence of +80 Da species interspersed within the hexose ladder. This was further confirmed by carbohydrate analysis (see below). The peak at 40 552.4 Da may result from a glycoform on which one of the high-mannose glycans has been trimmed to a single GlcNAc. MS analysis of  FEBS 2004 Heterogeneity of H. jecorina (T. reesei) Cel7B (Eur. J. Biochem. 271) 1269 additional peak observed at 39 700.6 Da (39 538+ 162 Da) is probably due to O-linked glycosylation of the protein (see below). The results are summarized in Table 2. Identification of N-glycosylation sites Fig. 2. PAG-IEF of Cel7Bcat (PyrGlu1–Thr371) expressed in H. jecorina QM9414-Cel7Bcat over the pH range 2.5–7. Lane 1, markers (amyloglucosidase, pI 3.50; methyl red dye, pI 3.75; soybean trypsin inhibitor, pI 4.55; b-lactoglobulin A, pI 5.20; bovine carbonic anhydrase, pI 5.85). Lane 2, purified Cel7Bcat; lane 3: purified Cel7Bcat after Endo H treatment. Endo H-treated Cel7Bcat shows a major peak with a molecular mass of 39 538.0 Da (Fig. 3B), which corresponds to the Cel7Bcat polypeptide plus two N-acetyl glucosamine residues (calculated molecular mass, 39 540.2 Da). The A series of detailed TOF MS and CID MS/MS experiments were performed to identify the N-glycosylation sites of wildtype Cel7Bcat. After deglycosylation by Endo H, and prior to digestion with either trypsin or V8 protease, the protein was denatured, the disulphide bonds reduced and the free Cys residues converted to carboxyamidomethyl derivatives. The predicted cleavage sites for trypsin and V8 protease under the conditions used for each digestion are indicated in Fig. 1. Peptides were infused directly into the MS without prior separation, and the observed m/z values were matched against those calculated for various protonated forms of the peptides (Table 1). Matching values were found for the T8, S7 and T12 peptides, thus indicating absence of glycosylation at Asn142 and Asn259. Due to the presence of Pro in the second position of the Asn-Xaa-(Ser/Thr) consensus sequence, glycosylation at Asn344 in the T15 peptide is not expected [9]. Indeed, no evidence for glycosylation at this site was observed in the TOF MS data. No signals were observed for the remaining two potential tryptic glycopeptides, T5 (which contains Asn56) and T9 (which contains both Asn182 and Asn186). However, peaks arising from these two peptides, each with an appendant GlcNAc residue, were observed (Table 1), thus indicating that these two peptides are N-glycosylated in the intact glycoprotein. The identities of peptides containing all five predicted N-glycosylation sites were further confirmed by CID MS/MS experiments. Fragmentation of ions corresponding to peptides T5+GlcNAc (Table 1, m/z 1159.2, [M+3H]3+) and T9+GlcNAc (Table 1, m/z 986.4, Table 1. Selected proteolytic fragments of Cel7Bcore. Peptides represent sequential fragment numbering from the N-terminus, T, tryptic peptides (cleavage after K and R except when preceeding P); S, S. aureus V8 proteolytic fragments (cleavage after E except when preceeding E or P). Potential N-linked glycosylation sites are shown in bold; C, carboxyamidomethyl cysteine; Q, Gln-derived pyroglutamic acid. All masses are monoisotopic. Ions indicated in bold were selected for CID MS/MS experiments. N.O., not observed. [M+H]+ Peptide Residues T1 T5 1–13 40–68 T5 + GlcNAc T8 40–68 123–181 [M+2H]2+ Sequence Calculated Observed Calculated Observed Calculated Observed QQPGTSTPEVHPK WMHDANYNSCTVNGGVNTTLCP DEATCGK WMHDANYNSCTVNGGVNTTLCP DEATCGK + GlcNAc (203.079) LNGQELSFDVDLSALPCGENGS 1388.68 3272.35 1388.65 N.O. 694.84 1636.68 694.84 N.O. 463.57 1091.46 N.O. N.O. 3475.43 N.O. 1738.22 1738.29 1159.15 1159.17a 6466.94 N.O. 1616.46 [M+4H]4+ 1616.48 1293.37 [M+5H]5+ 1293.59 1256.55 2754.17 2957.25 1256.61 N.O. N.O. 628.78 1377.59 1479.13 628.81 N.O. 1479.16 2583.31 1524.96 [M+4H]4+ N.O. 1524.99 1292.16 1220.17 [M+5H]5+ 1292.18 1220.21+ LYLSQMDENGGANQYNTAGANY S7 T9 T9 + GlcNAc T12 T15 142–152 182–205 40–68 GSGYCDAQCPVQTWR NGSLYLSQMDE NGTLNTSHQGFCCNEMDILEGNSR NGTLNTSHQGFCCNEMDILEGNSR 419.52 918.73 986.42 N.O. N.O. 986.43 + GlcNAc (203.079) 248–271 308–363 TFTIITQFNTDNGSPSGNLVSITR ALSSGMVLVFSIWNDNSQYMNWLD SGNAGPCSSTEGNPSNILANNPNT HVVFSNIR a [M+3H]3+ [M+4H]4+ ion also observed at m/z 869.62. 861.77 1016.98 [M+6H]6+ 861.79 1017.00  FEBS 2004 1270 T. Eriksson et al. (Eur. J. Biochem. 271) Fig. 3. Reconstructed zero-charge spectra of Cel7Bcat (A) and Endo H-treated Cel7Bcat (B). Table 2. Potential glycan structures correlated with observed glycoprotein masses. Proposed occupation of glycosylation sites Spectrum Fig. 3A Fig. 3B Observed mass (Da) Asn56 Asn182 40 41 41 41 41 42 42 42 42 GlcNAc Man5GlcNAc2 Man5GlcNAc2 Man5GlcNAc2 Man5GlcNAc2 Man5GlcNAc2 Man7GlcNAc2 (ManP)Man6GlcNAc2 Man7GlcNAc2 Man5GlcNAc2 Man5GlcNAc2 Man5GlcNAc2 Man7GlcNAc2 (ManP)Man6GlcNAc2 (ManP)Man6GlcNAc2 (ManP)Man6GlcNAc2 (ManP)Man6GlcNAc2 (ManP)Man6GlcNAc2 GlcNAc GlcNAc GlcNAc GlcNAc 552 566 729 890 970 133 295 374 457 39 537 39 700 [M+3H]3+) produced the peptide sequence tags NTTLCPDEATC and NGTLNTSHQGFCCNEMDL LEGNSR, respectively. Isobaric Leu/Ile is denoted by L, while C denotes carboxylamidomethyl Cys. In both cases, a neutral loss of 203 Da, the mass of GlcNAc, is observed from the [M + H]+ ion in deconvoluted, singlecharge spectra. In contrast, the T8 peptide is too large to generate useful CID MS/MS information. In this case, fragmentation of the much smaller S7 peptide ion (Table 1, m/z 628.8, [M + 2H]2+), which also contained the potential glycosylation site Asn142, produced a confirmatory peptide sequence tag SQMD. Thr/Ser Hex Hex Hex Hex CID MS/MS of the nonglycosylated T12 peptide ion (Table 1, m/z 1292.2, [M + 2H]2+) gave rise to two series of daughter ions, which correspond to the sequences LTQFNTDNGSPSGNLVSITR and LTQFNTDDGSPS GNLVSITR (Fig. 4). The data indicate that deamidation of Asn259 has occurred to produce an aspartic acid residue at this location (bold). The resulting 1 Da increase in the peptide mass results in the production of an [M + 2H]2+ ion for the Asn259Asp variant (m/z 1292.7) which was not resolved from the parent peptide ion in the quadrupole stage of the MS; simultaneous CID of both ions generates the overlapping series of daughter ions.  FEBS 2004 Heterogeneity of H. jecorina (T. reesei) Cel7B (Eur. J. Biochem. 271) 1271 Fig. 4. CID MS/MS analysis of the double-charged Cel7Bcat T12 peptide ion at m/z 1292.2. (A) Complete MaxEnt3 spectrum. (B) Expansion of spectrum A. (C) Assignment of the second series of daughter ions with aspartic acid as residue 259. 1272 T. Eriksson et al. (Eur. J. Biochem. 271)  FEBS 2004 Fig. 5. Reconstructed zero-charge spectra of Cel7BcatN182Q mutant (A) and Endo H treated Cel7BcatN182Q mutant (B). Analysis of the Cel7BcatN182Q mutant MS/MS analysis of the T9 peptide produced from Cel7Bcat failed to provide information about the site of N-glycan attachment (either Asn182 or Asn186), as the GlcNAc moiety is readily lost during CID. To resolve this ambiguity, a glycosylation site mutant was constructed in which Asn182 was replaced by glutamine. This mutant was expressed, purified and analysed in the same way as the wild-type Cel7Bcat protein. Figure 5 shows the reconstructed zero-charge spectra of the Cel7BcatN182Q protein before and after Endo H treatment. Accounting for an N-terminal pyroglutamate and eight cystine bridges, the calculated average molecular mass for the Cel7BcatN182Q polypeptide chain is 39 148.9 Da. The difference of 1216.6 Da between this value and that observed for the major isoform of the intact Cel7BcatN182Q protein (40 365.5 Da) corresponds to a modification of the polypeptide chain with a single GlcNAc2Man5 N-glycan (calculated molecular mass, 1217.1 Da). As with the wild-type Cel7Bcat, a range of isoforms was observed by MS; those separated by 162 Da reflect variable mannosylation of the N-glycan or O-glycosylation, while those separated from this series by 80 Da indicate glycan phosphorylation. Endo H treatment significantly reduced the number of glycoforms of the Cel7BcatN182Q observed by MS. The observed mass of the major species corresponds well with that expected for the Cel7BcatN182Q polypeptide bearing one asparagine-linked GlcNAc residue (calculated molecular mass, 39 352.0; observed molecular mass, 39 351.0). Similar to wild-type Cel7Bcat (Fig. 3B), the zero-charge spectrum of Endo H-treated Cel7BcatN182Q exhibits a minor peak 162 Da larger than the main glycoform, possibly reflecting an O-linked hexose modification. To confirm that Asn182 had indeed been mutated to Gln and that the remaining glycosylation site was identical to that of the wild-type protein, Endo H-treated Cel7BcatN182Q was similarly subjected to trypsin digestion and TOF MS analysis. As in Cel7Bcat (Table 1), peptide T5 of Cel7BcatN182Q was observed bearing a single GlcNAc residue (Table 3), and thus carried the N-glycan. The T9 peptide ionized as a triple-charged ion, m/z 923.4 (Table 3), and yielded the complete sequence QGTLNTSHQGF CCNEMDLLEGNSR upon CID MS/MS analysis, which verified the mutation and the absence of N-glycosylation. The remaining peptides bearing potential N-glycosylation sites (T8, T12 and T15) were only observed as their unmodified forms (Table 3). Characterization of Endo H-released N-glycans Total acid hydrolysis indicates that N-glycans released by Endo H treatment contain only mannose and N-acetyl glucosamine in a ratio of 6 : 1. Oligosaccharide analysis by HPAEC-PAD showed the presence of GlcNAcMan5, GlcNAcMan7 and, eluting at high salt concentrations, small amounts of negatively charged glycans (Fig. 6B).  FEBS 2004 Heterogeneity of H. jecorina (T. reesei) Cel7B (Eur. J. Biochem. 271) 1273 Table 3. Selected proteolytic fragments of Cel7BcoreN182Q. Peptides represent sequential fragment numbering from the N-terminus; T, tryptic peptides (cleavage after K and R except when preceeding P). Potential N-linked glycosylation sites are shown in bold; C, carboxyamidomethyl cysteine; Q, Gln-derived pyroglutamic acid. All masses are monoisotopic. Ions indicated in bold were selected for CID MS/MS experiments. N.O., not observed. [M + H]+ Peptide T1 T5 Residues Sequence 1–13 40–68 T5 + 40–68 GlcNAc T8 123–181 QQPGTSTPEVHPK WMHDANYNSCTVNGGVNTTLCP DEATCGK WMHDANYNSCTVNGGVNTTLCP DEATCGK + GlcNAc (203.079) LNGQELSFDVDLSALPCGENGS [M + 2H]2+ Calculated Observed Calculated Observed Calculated Observed 1388.68 3272.35 N.O. N.O. 694.84 1636.68 694.84 N.O. 463.57 1091.46 N.O. N.O. 3475.42 N.O. 1738.22 N.O. 1159.15 1159.17a 6466.94 N.O. 1616.46 1616.46 [M+4H]4+ 2768.18 2971.26 N.O. N.O. 1384.60 1486.13 N.O. N.O. 923.40 991.09 923.40 N.O. 1292.16 1220.17 [M+5H]5 N.O. 1220.18 1016.98+ 861.77 861.76 1016.96 [M+6H]6+ LYLSQMDENGGANQYNTAGANY T9 182–205 T9 + 40–68 GlcNAc T12 248–271 T15 308–363 GSGYCDAQCPVQTWR QGTLNTSHQGFCCNEMDILEGNSR QGTLNTSHQGFCCNEMDILEGNSR 1293.37 1293.56 [M+5H]5+ + GlcNAc (203.079) TFTIITQFNTDNGSPSGNLVSITR ALSSGMVLVFSIWNDNSQYMNWLD SGNAGPCSSTEGNPSNILANNPNT HVVFSNIR a [M + 3H]3+ 2583.31 N.O. 1524.96 1524.97 [M+4H]4+ [M+4H]4+ ion also observed at m/z 869.60. a-Mannosidase hydrolyses only the neutral oligosaccharides to mannose and GlcNAcMan (Fig. 6C). Acid treatment of the oligosaccharides (0.01 M HCl, 100 C) leads to the formation of mannose (Fig. 6D). The concomittant formation of a terminal phosphate could not be demonstrated with HPAEC-PAD, but PAGE analysis of fluorescent labelled oligosaccharides clearly indicates a shift in electrophoretic mobility (data not shown). Alkaline phosphatase digestion alone did not change the HPAEC-PAD elution profile. Conversion of charged N-glycans to their uncharged counterparts (GlcNAcMan5 and GlcNAcMan7) could only be achieved by combined mild acid hydrolysis and alkaline phosphatase treatment (Fig. 6E), thus indicating the presence of a phosphodiester linkage between sugar residues [26]. O-glycosylation The presence of O-glycosylation at serine or threonine residues was indicated by the observation of a second minor species with a mass 162 Da greater than the major glycoform in both Endo H-digested Cel7Bcat and Cel7BcatN182Q (Figs 3B and 5B). Analysis of the TOF MS data obtained for the tryptic digest of wild-type enzyme indicates that, in addition to peaks arising from the unmodifed T6 peptide (Fig. 1) at m/z 1384.64 ([M + 3H]3+) and m/z 1038.72 ([M + 4H]4+), peaks corresponding to T6 + Hex (T6 +162 Da) are present at m/z 1438.68 ([M +3H]3+) and m/z 1079.26 ([M + 4H]4+). CID MS/MS analysis of the m/z 1384.64 ion yielded near-complete sequence coverage for the T6 peptide: LEXXDYAASGVTTSGSSLTMNQY MPSSSGGYSSVSPR. CID MS/MS analysis of the putative glycopeptide ion at m/z 1438.68 generated less complete sequence tag information, and failed to pinpoint the site of glycosylation due to facile cleavage of the carbohydratepeptide linkage (spectrum not shown). Discussion Understanding the factors that contribute to the glycoform heterogeneity of proteins expressed in fungal sources is a fundamental concern in a wide range of applications from basic biochemical characterization to crystallographic studies and medicinal applications [12]. Previous studies on the catalytic module of Cel7B from H. jecorina have generated conflicting results about the extent and localization of glycosylation of this protein [8,16,18]. The present detailed structural study of Cel7Bcat expressed in an H. jecorina QM9414 derivative strain aims to elucidate discrepancies found in previous publications. N-glycan heterogeneity in Cel7Bcat PAG-IEF analysis clearly shows that the purified wildtype Cel7Bcat consists of several isoforms and that this heterogeneity is considerably reduced after enzymatic N-deglycosylation (Fig. 2). Comparative MS analysis of the intact protein before and after Endo H treatment confirms that it consists of a mixture of glycoforms, which vary in the extent of both mannosylation and phosphorylation. Moreover, the data on the intact protein alone are sufficient to show that the protein carries two N-linked high-mannose glycans. The results of detailed carbohydrate analysis (Fig. 6) and peptide MS experiments (Fig. 3A) define the glycoform structure and the sites of glycan attachment in the present enzyme preparation. TOF MS and CID MS/MS analyses of Endo H-treated Cel7Bcat and Cel7BcatN182Q proteolytic digests unequivocally show that Asn56 and Asn182 of the Cel7B catalytic module are glycosylated. These two residues are located on exposed loops of the protein (Protein Data Bank code 1EG1 [8]), as is often observed for N-glycosylation sites [27]. The multiplicity of glycoforms in the intact  FEBS 2004 1274 T. Eriksson et al. (Eur. J. Biochem. 271) that H. jecorina produces one or more enzymes responsible for processing glycan structures during cultivation. Indeed, recent results indicate that H. jecorina produces an array of extracellular glycan processing enzymes, including an enzyme with Endo H-like activity [32a]. Most interestingly, these activities exhibit widely different pH profiles, which implies that the spectrum of enzyme activities changes with pH variations in the medium during fermentation. Thus, the incongruous results could be explained by postsecretorial effects due to the growth conditions used in the different studies. The minimal medium used here results in a low pH at the final culture stages, where the glycan hydrolases become inactive. As such, the high-mannose N-glycans decorated with phosphodiester modifications represent the initial complexity of secreted fungal proteins. Reports of Cel7B carrying single GlcNAc residues undoubtly result from cultivations in rich medium, which buffers the pH near the optima for the extracellular hydrolases [32a]. O-glycan heterogeneity in Cel7Bcat Fig. 6. HPAEC-PAD analysis of N-glycans released from Cel7Bcat. (A) Reference N-glycans. (B) N-glycans released from Cel7B. (C) N-glycans released from Cel7B treated with Jack bean a-mannosidase. (D) N-glycans released from Cel7B treated with mild acid. (E) N-glycans released from Cel7B treated with mild acid followed by alkaline phosphatase. protein (Fig. 3A and Table 2) can be explained by the presence of Man5GlcNAc2 and Man7GlcNAc2, their derived phosphodiesters and/or one extra O-linked mannose residue. The predominant Man5GlcNAc2 N-glycan reflects normal processing in the endoplasmatic reticulum by a(1fi3) glucosidases and further trimming by a nonspecific a(1fi2) mannosidase present in H. jecorina [28,29]. As suggested previously, the presence of minor amounts of GlcNAc2Man6)7 could be due to reduced catalytic activity of the latter enzyme. Previous studies on both Cel7A and Cel7B isolated from strain Rut-C30 report glucosylated N-glycans (GlcMan7GlcNAc2) [15,18]. Thus, although the glycosylation sites are invariant, the structures of the attached glycans can be markedly different between strains of different mutational lineage [30,31]. N-glycan phosphorylation has previously been demonstrated in proteins from both strain QM9414 and RUT-C30 (e.g [15,16,32]). In the present study, the presence of Man-PiMan phosphodiester structures in Cel7B is confirmed by mild acid hydrolysis (release of Man) and conversion of the free glycans to uncharged counterparts by subsequent phosphatase treatment (Fig. 6). The observed differences in N-glycan structures in this study compared to the other studies of Hypocrea jecorina Cel7B [8,16,18] may further be explained, in part, by differences in cultivation and/or purification conditions. Previous studies have shown N-glycan structures on Cel7B with only a single GlcNAc [16,18], which may indicate In addition to the observed N-glycan heterogeneity, Cel7Bcat has one partially occupied O-glycosylation site located in peptide T6 Asn69–Arg109. Whereas N-glycosylation has been observed only in the catalytic module of fungal cellulases, O-linked glycosylation has been shown to be a typical feature of the linker peptide. Heterogeneous phosphorylation or sulfation of mono- di- or tri-hexose (predominantly mannose) units on serine or threonine residues has been reported [14,17,32]. The function of this O-glycosylation is not fully understood but it has been suggested that it contributes to enzyme stability [11] and helps define the conformation of the linker [33]. In contrast, the O-glycosylation of core modules of H. jecorina cellulases has not been widely documented. To our knowledge, the present work is only the second report of this phenomenon; the first is the observation of O-linked a-mannosyl units on multiple serine and threonine residues in Cel6A (formerly cellobiohydrolase II, CBHII) [34]. CID MS/MS analysis did not allow the determination of the exact site of hexose attachment within the T6 peptide. Analysis of the Cel7Bcat structure (PDB code 1EG1) shows that this peptide is rich in surface-exposed serine and threonine residues found in tandem or triple repeats. Although further chemical analysis will be required, it is tempting to speculate that the sequence 97-SSS-99, which is at the tip of a loop region, may be a likely O-glycosylation site. Heterogeneity in Cel7Bcat due to protein deamidation The spontaneous deamidation of asparagines to yield aspartic acid residues is well documented for both proteins and peptides [35]. In particular, the sequence Asn–Gly has been shown to be especially susceptible [36]. Our results indicate that partial deamidation at Asn259–Gly260 has occurred in Cel7Bcat. Deamidation has recently been observed in the protein hydrophobin HFBI expressed in H. jecorina [37], and it is likely that this process contributes to a further increase in the heterogeneity of H. jecorina proteins in general.  FEBS 2004 Heterogeneity of H. jecorina (T. reesei) Cel7B (Eur. J. Biochem. 271) 1275 The three types of protein heterogeneity in Cel7Bcat observed by MS analysis all contribute to the complex PAGIEF patterns observed with purified samples of the enzyme. Native Cel7Bcat exhibits a pattern resulting from at least five different forms of the enzyme. Because Endo H treatment reduces the number of isoforms to two, the three bands observed with more acidic isoelectric points probably correspond to phosphorylated forms (substitution of charged N-glycans on one or both glycosylation sites, Table 2). Heterogenous mannosylation of neutral N-glycans at one or both glycosylation sites is only expected to cause a minor shift in the pI value. The major component resulting from Endo H treatment is Cel7Bcat bearing two single GlcNAc substitutions (calculated pI 4.44), as observed by MS. The minor, more acidic isoform could either result from the deamidation of Asn259 (calculated pI 4.39) or from the presence of a species carrying charged O-glycosylation. Indeed, a species corresponding to the Cel7Bcat polypeptide bearing a hexosephosphate (or sulfate) moiety was observed by MS analysis of the intact protein from some preparations (not shown). However, the very low abundance of this peak and the lack of a confirmatory observation in peptide digests makes such an assignment uncertain. In conclusion, the heterogeity observed in the Cel7B preparation by electrophoresis and mass spectrometry can be attributed to N-glycosylation. Moreover, the results indicate both O-glycosylation and deamidation in the catalytic domain are further complicating factors. Acknowledgements I. S. and M. C. thank the Ghent University Research Council for support (B/03197/04 IV 1). H.S. thanks the Swedish Research Council (Vetenskapsrådet) and the Carl Trygger Foundation (CTS) for funding. The purchase of mass spectrometry equipment at the Royal Institute of Technology was funded by the Wallenberg Consortium North. 8. 9. 10. 11. 12. 13. 14. 15. 16. References 1. Kuhls, K., Lieckfeldt, E., Samuels, G.J., Kovacs, W., Meyer, W., Petrini, O., Gams, W., Borner, T. & Kubicek, C.P. (1996) Molecular evidence that the asexual industrial fungus Trichoderma reesei is a clonal derivative of the ascomycete Hypocrea jecorina. Proc. Natl Acad. Sci. USA 93, 7755–7760. 2. Srisodsuk, M., Kleman-Leyer, K., Keränen, S., Kirk, K. & Teeri, T. (1998) Modes of action on cotton and bacterial cellulose of a homologous endoglucanase-exoglucanase pair from Trichoderma reesei. Eur. J. Biochem. 251, 885–892. 3. Kleman-Leyer, K.M., Siika-aho, M., Teeri, T.T. & Kirk, T.K. (1996) The cellulases endoglucanase I and cellobiohydrolase II of Trichoderma reesei act synergistically to solubilize native cotton cellulose but not to decrease its molecular size. Appl. Env. Microb. 62, 2883–2887. 4. Barr, B.K., Hsieh, Y.L., Ganem, B. & Wilson, D.B. (1996) Identification of two functionally different classes of exocellulases. Biochemistry 35, 586–592. 5. Nutt, A., Sild, V., Pettersson, G. & Johansson, G. (1998) Progress curves. A mean for functional classification of cellulases. Eur. J. Biochem. 258, 200–206. 6. Henrissat, B., Teeri, T.T. & Warren, R.A.J. (1998) A scheme for designating enzymes that hydrolyse the polysaccharides in the cell walls of plants. FEBS Lett. 425, 352–354. 7. Divne, C., Ståhlberg, J., Reinikainen, T., Ruohonen, L., Pettersson, G., Knowles, J.K.C., Teeri, T.T. & Jones, T.A. (1994) The 17. 18. 19. 20. 21. 22. three-dimensional crystal structure of the catalytic core of cellobiohydrolase I from Trichoderma reesei. Science 265, 524–528. Kleywegt, G., Zou, J.-Y., Divne, C., Davies, G.J., Sinning, I., Ståhlberg, J., Reinikainen, T., Srisodsuk, M., Teeri, T. & Jones, T.A. (1997) The crystal structure of the catalytic core domain of endoglucanase I from Trichoderma reesei at 3.6 Å resolution, and a comparision with related enzymes. J. Mol. Biol. 272, 383–397. Gavel, Y. & von Heijne, G. (1990) Sequence differences between glycosylated and non-glycosylated Asn-X-Thr/Ser acceptor sites: implications for protein engineering. Protein Eng. 3, 433–442. Chen, H.-M., Ford, C. & Reilly, P.J. (1994) Substitution of asparagine residues in Aspergillus awamori glucoamylase by site-directed mutagenesis to eliminate N-glycosylation and inactivation by deamidation. Biochem. J. 301, 275–281. Neustroev, K.N., Golubev, A.M., Firsov, L.M., Ibatullin, F.M., Protasevich, I.I. & Makarov, A.A. (1993) Effect of modification of carbohydrate component on properties of glucoamylase. FEBS Lett. 316, 157–160. Maras, M., van Die, I., Contreras, R. & van den Hondel, C. (1999) Filamentous fungi as production organisms for glycoproteins of bio-medical interest. Glycoconjugate J. 16, 99–107. Klarskov, K., Piens, K., Ståhlberg, J., Høj, P.B., van Beeumen, J. & Claeyssens, M. (1997) Cellobiohydrolase I from Trichoderma reesei: Identification of an active-site nucleophile and additional information on sequence including the glycosylation pattern of the core protein. Carbohydr. Res. 304, 143–154. Harrison, M., Nouvens, A., Jardine, D., Zachara, N., Gooley, A., Nevalainen, H. & Packer, N. (1998) Modified glycosylation of cellobiohydrolase I from a high cellulase-producing mutant strain of Trichoderma reesei. Eur. J. Biochem. 256, 119–127. Maras, M., De Bruyn, A., Schraml, J., Herdewijn, P., Claeyssens, M., Fiers, W. & Contreras, R. (1997) Structural characterization of N-linked oligosaccharides from cellobiohydrolase I secreted by the filamentous fungus Trichoderma reesei Rut-C30. Eur. J. Biochem. 245, 617–625. Garcı́a, R., Cremata, J.A., Quintero, O., Montesino, R., Benkestock, K. & Ståhlberg, J. (2001) Characterisation of protein glycoforms with N-linked neutral and phosphorylated oligosaccharides: studies on the glycosylation of endoglucanase I from Trichoderma reesei. Biotechnol. Appl. Biochem. 33, 141–152. Hui, J.P.M., Lanthier, P., White, T.C., McHugh, S.G., Yaguchi, M., Roy, R. & Thibault, P. (2001) Characterization of cellobiohydrolase I (Cel7A) glycoforms from extracts of Trichoderma reesei using capillary isoelectric focusing and electrospray mass spectrometry. J. Chromatogr. B 752, 349–368. Hui, J.P.M., White, T.C. & Thibault, P. (2002) Identification of glycan structures and glycosylation sites in cellobiohydrolase II and endoglucanase I and II from Trichoderma reesei. Glycobiology 12, 837–849. Collén, A., Ward, M., Tjerneld, F. & Stålbrand, H. (2001) Genetic engineering of the Trichoderma reesei endoglucanase I (Cel7B) for enhanced partitioning in aqueous two-phase systems containing thermoseparating ethylene oxide-propylene oxide copolymers. J. Biotechnol. 87, 179–191. Penttilä,M.,Nevalainen,H.,Rättö,M.,Salminen,E.&Knowles,J. (1987) A versatile transformation system for the cellulolytic filamentous fungus Trichoderma reesei. Gene 61, 155–164. Collén, A., Ward, M., Tjerneld, F. & Stålbrand, H. (2001) Genetically engineered peptide fusions for improved protein partitioning in aqueous two-phase systems. Effect of fusion localization on endoglucanase I of Trichoderma reesei. J. Chromatogr. A 910, 275–284. Nakari-Setälä, T. & Penttilä, M. (1995) Production of Trichoderma reesei cellulases on glucose-containing media. Appl. Environ. Microb. 61, 3650–3655.