GH3 (glycoside hydrolase family 3) BGLs (β-glucosidases) from filamentous fungi have been widely and commercially used for the supplementation of cellulases. AaBGL1 (Aspergillus aculeatus BGL1) belongs to the GH3 and shows high activity towards cellooligosaccharides up to high degree of polymerization. In the present study we determined the crystal structure of AaBGL1. In addition to the substrate-free structure, the structures of complexes with glucose and various inhibitors were determined. The structure of AaBGL1 is highly glycosylated with 88 monosaccharides (18 N-glycan chains) in the dimer. The largest N-glycan chain comprises ten monosaccharides and is one of the largest glycans ever observed in protein crystal structures. A prominent insertion region exists in a fibronectin type III domain, and this region extends to cover a wide surface area of the enzyme. The subsite +1 of AaBGL1 is highly hydrophobic. Three aromatic residues are present at subsite +1 and are located in short loop regions that are uniquely present in this enzyme. There is a long cleft extending from subsite +1, which appears to be suitable for binding long cellooligosaccharides. The crystal structures of AaBGL1 from the present study provide an important structural basis for the technical improvement of enzymatic cellulosic biomass conversion.

INTRODUCTION

The production of biofuels and other chemicals from biomass has attracted significant attention with the creation of ‘biorefineries’ [1,2]. Cellulosic biomass is one of the most attractive renewable resources because it is the most abundant polysaccharide on Earth and, for the most part, does not compete for raw materials with food and animal feed production. Cellulosic biomass is, however, naturally resistant to microbial and enzymatic degradation [3]. Development of a cost-competitive process for converting cellulosic biomass into fermentable sugar is required for the widespread utilization of this process. The filamentous fungi Trichoderma reesei produces large amounts of extracellular glycosidases and is widely used for industrial cellulase production [4]. T. reesei produces a complete set of cellulases, including endo-β-1,4-glucanases (EC 3.2.1.4), cellobiohydrolases (EC 3.2.1.91) and BGLs (β-glucosidases; EC 3.2.1.21) [5]. However, BGL activity in the cellulase mixture is generally quite low and limits the total hydrolysis of cellulose [6]. This problem has been alleviated by the addition of exogenous BGL [79] or by construction of recombinant T. reesei strains that overproduce endogenous [10,11] or exogenous [12] BGL. BGLs from Aspergillus niger and Aspergillus aculeatus have been often used for the supplementation of T. reesei cellulases [13,14].

AaBGL1 [A. aculeatus (number F-50) BGL1] is potently active not only on soluble cellooligosaccharides, but also on insoluble cellooligosaccharides [the average DP (degree of polymerization)=20] [1517]. AaBGL1 shows a higher specific activity against cellobiose compared with that of T. reesei BGLs, and a recombinant T. reesei strain expressing AaBGL1 yielded enzyme preparations that had high saccharification activity for various pre-treated biomasses at a low enzyme dose [12,18]. Moreover, AaBGL1 has been repeatedly used for biotechnological developments and research studies such as the construction of ‘arming yeast’, which display enzymes on their cell surface [1926], lactic acid and amino acid fermentation by bacteria [27,28], and glucose production from cellulose with hyperthermophilic endoglucanase [29].

BGLs have been found in the GH (glycoside hydrolase) families GH1, GH3, GH5, GH9 and GH30 of the CAZy database [30]. Many fungal BGLs, including AaBGL1 [31], extracellular BGLs from A. niger and BGL1 from T. reesei (Bgl1 or Cel3A) are classified as part of the GH3 family. The GH3 family is a very large family, which comprises more than 4000 members. The GH3 family includes BGLs, NagZs (N-acetyl-β-D-glucosaminidases), α-L-arabinofuranosidases and β-D-xylopyranosidases. The position and the identity of the acid/base catalytic residues of the GH3 family members are highly diverse and not identical [32]. Crystal structures of GH3 family BGLs have been reported for only four enzymes: β-D-glucan glucohydrolase ExoI from barley Hordeum vulgare [33], Bgl3B from the hyperthermophilic bacterium Thermotoga neapolitana [34], BglI from the yeast Kluyveromyces marxianus [35] and exo-1,3-1,4-β-glucanase ExoP from the marine bacterium Pseudoalteromonas sp. BB1 [36]. Among these four BGLs, T. neapolitana Bgl3B and K. marxianus BglI are most similar to AaBGL1, but they have only a 29% amino acid sequence identity (Supplementary Table S1 at http://www.biochemj.org/bj/452/bj4520211add.htm). Fungal BGLs are distantly related to the structurally characterized BGLs and are classified into a distinct subfamily of the GH3 family. In the system described by Harvey et al. [37], in which six major phylogenetic branches are defined, H. vulgare ExoI, T. neapolitana Bgl3B, K. marxianus BglI and fungal BGLs are classified into branches 1, 5, 5 and 4 respectively. The branch including Pseudoalteromonas sp. ExoP is not defined in this system, but is located between branches 1 and 2. In the system described by Cournoyer and Faure [38], H. vulgare ExoI, Pseudoalteromonas sp. ExoP, T. neapolitana Bgl3B, K. marxianus BglI and fungal BGLs are classified into subclusters D, D, C3, C3 and C2 respectively. However, no crystal structure of GH3 family BGLs from filamentous fungi have been available despite their importance in industrial applications. Owing to the low sequence homology and numerous insertions and deletions, homology modelling of the fungal GH3 family BGLs based on the BGL structures from plant, bacteria and yeast and quite unreliable. Very recently, the crystal structures of the GH3 family Bgl1 from T. reesei have been deposited (PDB codes 3ZYZ, 3ZZ1 and 4I8D) in the PDB, but the accompanying research article has not yet been published. In the present study, we report the crystal structure of AaBGL1, providing a three-dimensional view of a fungal GH3 family BGL. To clarify the interactions at the active site, structures of complexes with D-glucose and various inhibitors were determined.

MATERIALS AND METHODS

Protein expression and purification

Aspergillus oryzae niaD300 and the high-level expression shuttle vector plasmid pNAN8142 were used as the host vector system for expression of the protein [39]. The pH3K plasmid, pUC118 carrying the entire bgl1 gene [31] between the XhoI and SphI restriction sites, was used as the source of the gene. The XhoI/SphI fragment was inserted into the same sites of pNAN8142, and the constructed plasmid was named pNAN-BGL1-A (12.8 kb). A. oryzae was transformed as described previously [40,41]. For enzyme expression, the transformant was grown in 200 ml of minimal medium [(w/v) 5% glucose, 1% NaNO3, 0.13% KCl, 0.13% MgSO4·7H2O, 0.38% KH2PO4, 0.00011% (NH4)6Mo7O24·4H2O, 0.00011% H3BO3, 0.00016% CoCl2·6H2O, 0.00016% CuSO4·5H2O, 0.005% EDTA, 0.0005% FeSO4·7H2O, 0.0005% MnCl2·4H2O and 0.0022% ZnSO4·7H2O (pH 6.5)] in a 500 ml baffle flask with rotary shaking at 160 rev./min. After incubation for 3 days at 30°C, mycelia were harvested by filtering using a Büchner funnel with filter paper, rinsed with 150 ml of 20 mM sodium acetate (pH 5.0) and were suspended in 200 ml of enzyme-releasing solution consisting of 20 μg/ml cycloheximide, 1 mM PMSF and 20 mM sodium acetate (pH 5.0). After incubation at 30°C for 4 days with shaking at 160 rev./min, the recombinant protein was released into the solution. The mycelia were removed by filtering using a Büchner funnel with filter paper and centrifugation at 10000 g for 30 min at 4°C. From the supernatant the AaBGL1 protein was purified using a column of DEAE–Toyopearl® 650 M (Tosoh) with a linear gradient of 0–0.3 M NaCl in 20 mM sodium acetate (pH 5.0). The enzyme was then further purified on a column of Butyl–Toyopearl® 650 M (Tosoh) with a decreasing linear gradient of 30–0% saturation of ammonium sulfate in 20 mM sodium acetate (pH 5.0). The sample was then subjected to ammonium sulfate precipitation (80% saturation). The pellet was resuspended in 20 mM sodium acetate (pH 5.0) and dialysed against the same buffer. The protein sample (88 mg) was treated with 50 units of endoglycosidase Hf (fusion with maltose-binding protein; New England Biolabs) in 50 mM sodium citrate buffer (pH 5.5) overnight at 37°C. After the deglycosylation reaction, endoglycosidase H was removed using a column of amylose resin (New England Biolabs). The sample was finally subjected to ammonium sulfate precipitation (80% saturation), resuspension and dialysis against 20 mM sodium acetate (pH 5.0).

Inhibition kinetics

The enzyme activity towards pNP-Glc (p-nitrophenyl-β-D-glucopyranoside) was measured at various substrate (0.1–1.2 mM) and inhibitor concentrations in 50 mM sodium citrate (pH 5.0) at 37°C. The following inhibitor concentrations were used: 10 and 50 μM for IFG (isofagomine); 2 and 5 μM for DNJ (1-deoxynojirimycin) and CTS (castanospermine); and 4 mM for (+)-CgB2 (calystegine B2). The reaction was initiated by the addition of 2.94 μg/ml enzyme (final concentration) and the release of p-nitrophenol was continuously monitored using a V-550 spectrophotometer (Jasco). The molar absorption coefficient of p-nitrophenol at 405 nm (pH 5.0) was 174 M−1·cm−1. In the absence of inhibitors, the Km and Vmax values for pNP-Glc were determined to be 0.42±0.05 mM and 294±14 units/mg respectively. Each inhibition pattern, i.e. competitive or mixed, was judged using a 1/[S]−1/v plot in the presence and absence of each inhibitor. The inhibition parameters (Ki and Ki′) were calculated by using SigmaPlot (Systat Software). The equations used for calculating competitive and mixed inhibition constants were v=Vmax[S]/{Km(1+[I]/Ki)+[S]} and v=Vmax[S]/{Km(1+[I]/Ki)+[S](1+[I]/Ki′)} respectively.

Crystallography

Crystallization was performed at 20°C for 3 weeks, using the hanging-drop vapour-diffusion method. Crystals were obtained by mixing 1 μl of protein solution, consisting of 21 mg/ml AaBGL1 and 20 mM sodium acetate (pH 5.0), and 1 μl of reservoir solution, consisting of 35% MPD (2-methyl-2,4-pentanediol) and 0.1 M sodium acetate (pH 4.8). X-ray diffraction data were collected at the NE3A and NW12A stations [λ=1.0 Å (1 Å=0.1 nm)] at the Photon Factory AR, High Energy Accelerator Research Organization, Tsukuba, Japan. Complex crystals were prepared by soaking the crystals in the reservoir solution supplemented with each ligand. IFG and TCB (thiocellobiose) were purchased from Toronto Research Chemicals. CgB2 was purchased from Alexis Biochemicals. DNJ and CTS were purchased from Wako Pure Chemical Industries. The soaking conditions for crystals complexed with IFG, DNJ, CgB2, CTS, Glc and TCB were 10 mM for 60 min, 10 mM for 90 min, 10 mM for 1 min, 1 mM for 1 min, 100 mM for 1 min and 10 mM for 1 min respectively. Crystals were flash-cooled at 95 K in a stream of nitrogen gas. Data were processed using the HKL-2000 program [42]. Molecular replacement was performed using MOLREP [43]. Automated model building was performed using ARP/wARP [44]. Manual model rebuilding and refinement were performed using Coot [45] and Refmac5 [46]. Data collection and refinement statistics are shown in Supplementary Table S2 (at http://www.biochemj.org/bj/452/bj4520211add.htm). The molecular interfaces were analysed using the PDBe PISA server [47].

RESULTS AND DISCUSSION

Purification, inhibition kinetics, crystallization and structure determination

Recombinant AaBGL1 protein was overexpressed in the A. oryzae niaD mutant strain under the control of the promoter P-No8142 [39]. The recombinant protein was efficiently secreted by the original secretion signal sequence for A. aculeatus and then purified by several steps of column chromatography. Using the purified enzyme, we measured inhibition by four azasugar (or iminosugar)-type glucosidase inhibitors (IFG, DNJ, CgB2 and CTS) using pNP-Glc as a substrate. The inhibition modes and constants are summarized in Table 1. IFG, DNJ and CTS were potent competitive inhibitors and their Ki values were less than 15 μM. CgB2 was a weak mixed-type inhibitor with inhibition constants larger than 4 mM.

Table 1
Inhibition modes and inhibition constants by azasugars

The inhibition parameters, Ki and Ki′, correspond to the inhibitors against the free enzyme and the enzyme–substrate complex respectively.

InhibitorModeKi (μM)Ki′ (μM)
IFG Competitive 14±2 – 
DNJ Competitive 2.4±0.3 – 
CgB2 Mixed 4600±1000 10000±5000 
CTS Competitive 6.6±0.6 – 
InhibitorModeKi (μM)Ki′ (μM)
IFG Competitive 14±2 – 
DNJ Competitive 2.4±0.3 – 
CgB2 Mixed 4600±1000 10000±5000 
CTS Competitive 6.6±0.6 – 

The theoretical molecular mass of the AaBGL1 protein deduced from its amino acid sequence is 91151 (841 aa without the 19 aa signal sequence) [31]. During SDS/PAGE analysis the recombinant AaBGL1 protein appeared as a smeared band at approximately 130–140 kDa. Endoglycosidase H treatment under non-denaturing conditions reduced the smearing and the size of the resulting band shifted to approximately 115 kDa, indicating that approximately half of the N-glycans were removed. The endoglycosidase H treatment significantly improved the reproducibility of the crystallization and the resolution of the X-ray diffractions. Crystals of the deglycosylated AaBGL1 diffracted beyond a 1.8 Å resolution (Supplementary Table S2). The initial phases were determined by molecular replacement using the structure of T. neapolitana Bgl3B (PDB code 2X40) as a search model. The crystal structure (substrate-free structure) was determined at a 1.80 Å resolution and refined to R/Rfree=13.4/16.8%. The crystal structure contains two molecules in the asymmetric unit (Figure 1A), and the final model of the substrate-free structure contains residues from Leu22 to Ala669 and Val675 to Gln860 of chain A and from Leu22 to Asn668 and Ala676 to Gln860 of chain B. The structure also contains nine N-linked glycan chains, three MPD molecules, one Na+ ion and one acetate ion in each polypeptide.

Overall structure of AaBGL1

Figure 1
Overall structure of AaBGL1

Dimer (A) and monomer (B) structures. The catalytic TIM barrel-like domain (blue), α/β sandwich domain (green), FnIII domain (yellow), insertion region (red) and linker regions (black) are shown. The N-glycans and ligands in the active site (acetate and MPD) are shown as cyan sticks and magenta spheres respectively. In (A), one monomer is shown in grey and (B) is a stereo figure. (C) The rainbow-coloured FnIII domain. The insertion region is shown as magenta.

Figure 1
Overall structure of AaBGL1

Dimer (A) and monomer (B) structures. The catalytic TIM barrel-like domain (blue), α/β sandwich domain (green), FnIII domain (yellow), insertion region (red) and linker regions (black) are shown. The N-glycans and ligands in the active site (acetate and MPD) are shown as cyan sticks and magenta spheres respectively. In (A), one monomer is shown in grey and (B) is a stereo figure. (C) The rainbow-coloured FnIII domain. The insertion region is shown as magenta.

Overall structure

The molecular mass of AaBGL1 estimated by calibrated gel-filtration chromatography in a solution containing 150 mM NaCl and 50 mM Tris/HCl (pH 5.0) was 320 kDa (results not shown), suggesting that AaBGL1 is a dimer in solution. When the dimer interface in the crystallographic asymmetric unit was analysed using protein (polypeptide) atoms only, the buried surface area of the interface (1450 Å2) is only approximately 5% of the entire surface area of the monomer (27780 Å2), however, 25 hydrogen bonds are involved. The dimer interface is expanded by N-glycans (Figure 1A), which increases the buried surface area by 470 Å with three additional hydrogen bonds. The AaBGL1 monomer consists of three domains; a catalytic TIM (triosephosphateisomerase) barrel-like domain (Leu22–Ser356; Figure 1B, blue), an α/β sandwich domain (Gln385–Gly588; Figure 1B, green) and a FnIII (fibronectin type III) domain (Tyr654–Gln860; Figure 1B, yellow). These domains are connected with two linker regions (residues 357–384 and 589–653; Figure 1B, black). Within the FnIII domain there is a prominent insertion region (Val675–Asn755; Figure 1A, red). The FnIII domain is located at the ‘back’ side of the molecule, but the insertion loop extends to the ‘front’ side and covers a wide area of the molecular surface. A short region between the insertion loop and the FnIII domain (residues 670–674 in chain A and 669–675 in chain B) was not included in the model owing to disorder. There are three disulfide bonds per monomer (Cys73–Cys89 and Cys246–Cys257 in the barrel domain and Cys435–Cys440 α/β sandwich domain).

The TIM barrel-like domain actually adopts a partially broken ββ(β/α)6-barrel fold, in which the first and the second helices are deleted and the second β-strand is antiparallel. The partially broken ββ(β/α)6-barrel is also found in T. neapolitana Bgl3B and K. marxianus BglI [34,35], whereas H. vulgare ExoI and Pseudoalteromonas sp. ExoP have a canonical (β/α)8-barrel fold [33,36]. This structural feature is consistent with the subfamily classification system described by Cournoyer and Faure [38], in which (β/α)8-barrel and ββ(β/α)6-barrel fold-containing enzymes are classified into clusters D and C respectively. In the α/β sandwich domain, a six-stranded β-sheet is sandwiched between two layers both consisting of three helices. The α/β sandwich domain of AaBGL1 is similar to that of K. marxianus BglI, whereas that of T. neapolitana Bgl3B lacks one strand (five-stranded). The FnIII domain consists of two β-sheets facing each other (Figure 1C). One of the β-sheets is three-stranded (A1+A2, B and E), and the other is four-stranded (C1+C2, D1+D2, F and G). The strands A1–A2 and C1–C2 are divided by kinks formed by Gly661 and Pro781 respectively. Although there are no glycine or proline residues in the kink between strands D1 and D2, Arg802 is located in this region and appears to be pulled by Thr115 in the barrel domain, Asp398 in the α/β sandwich domain and Asp778 in the FnIII domain.

The structure of AaBGL1 was compared with other GH3 family BGLs. As shown in Supplementary Figure S1, H. vulgare ExoI, T. neapolitana Bgl3B and K. marxianus BglI have two-domain (TIM barrel+α/β sandwich), three-domain (TIM barrel+α/β sandwich+FnIII) and four-domain (TIM barrel+α/β sandwich+FnIII+PA14) structures respectively. Although the domain construction of AaBGL1 is similar to that of T. neapolitana Bgl3B, the insertion region in the FnIII domain is uniquely present in AaBGL1. Supplementary Table S1 shows sequence and structural comparisons between the GH3 family BGLs. Using BLAST sequence similarity search, AaBGL1 was observed to be most similar to T. neapolitana Bgl3B (E-value=7e−73), reflecting the similar domain construction. From a Dali structural similarity search [48], K. marxianus BglI shows the highest similarity [Z score=42.3 and RMSD (root mean square deviation)=1.9 Å for 635 Cα atoms]. We also performed a structural comparison with the T. reesei Bgl1 structure, which is available from the PDB. The domain structure of T. reesei Bgl1 is basically similar to that of AaBGL1, consisting of TIM barrel, α/β sandwich and FnIII domains (Supplementary Figure S1E). A pairwise structural comparison between AaBGL1 and T. reesei Bgl1 (PDB code 3ZYZ) using the Dali server shows that they have high similarity (Z score=49.5 and RMSD=1.5 Å for 701 Cα atoms). However, T. reesei Bgl1 lacks the insertion loop within the FnIII domain.

N-glycans

The amino acid sequence of AaBGL1 includes 16 potential glycosylation sites. Before crystallization, the protein sample was treated with endoglycosidase H, which cleaves the glycosidic bond of (GlcNAc)2 and leaves one asparagine residue-linked GlcNAc. Surprisingly, the AaBGL1 structure is still highly glycosylated with high-mannose type N-glycans (Table 2). The final model contains nine N-glycan chains per polypeptide chain, and only two of them (Asn211 and Asn712) consist of a single GlcNAc (Supplementary Figures S2a and S2b). These two sites are located at the protein surface and appear to be susceptible to the endoglycosidase. The other seven N-glycans are partially or fully buried and interact with the protein. For Asn252, Asn315 and Asn564 the N-glycans in the chain B are larger than those in the chain A by one to three sugar residues. The N-glycan at Asn322 is the largest one, consisting of eight Man and two GlcNAc, with a clear electron density (Supplementary Figure S3 at http://www.biochemj.org/bj/452/bj4520211add.htm). The large N-glycan at Asn322 forms many hydrogen bonds with protein residues and is buried in a pocket formed by a part of the insertion loop (Asp722–His732; Figure 2A). The N-glycans at Asn422 and Asn523 are located at the dimer interface (Figure 1A) and form inter- and intra-subunit hydrogen bonds (Figures 2B and 2C). Other N-glycans at Asn61, Asn252, Asn315 and Asn564 also form multiple hydrogen bonds with the protein (Supplementary Figures S2c–S2f). In particular, the root GlcNAc residue of the N-glycan at Asn564 is involved in the co-ordination of a metal ion that was assigned as Na+ according to B-factor and difference-Fourier map analyses. These high-mannose N-glycan structures are consistent with those observed in secreted proteins from Aspergillus species [4951].

Table 2
N-glycan modifications in the final model of the substrate-free structure
ResidueStructure
Chain A  
 Asn61 Man-α1,6–(Man-α1,2–Man-α1,2–Man-α1,3)Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn211 GlcNAc 
 Asn252 Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn315 GlcNAc-β1,4–GlcNAc 
 Asn322 Man-α1,2–Man-α1,6–(Man-α1,2–Man-α1,3-)Man-α1,6–(Man-α1,2–Man-α1,3)Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn442 Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn523 Man-α1,2–Man-α1,6-Man-α1,6–(Man-α1,3-)Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn564 Man-α1,6–(Man-α1,2-Man-α1,3-)Man-α1,6–Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn712 GlcNAc 
Chain B  
 Asn61 Man-α1,6–(Man-α1,2-Man–α1,2-Man-α1,3)Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn211 GlcNAc 
 Asn252 Man-α1,3–Man-α1,6–(Man-α1,3)Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn315 Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn322 Man-α1,2-Man-α1,6-(Man-α1,2–Man-α1,3)Man-α1,6–(Man-α1,2–Man-α1,3)Man-β1,4-GlcNAc-β1,4–GlcNAc 
 Asn442 Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn523 Man-α1,2–Man-α1,6-Man-α1,6–(Man-α1,3)Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn564 Man-α1,6–(Man-α1,2–Man-α1,3)Man-α1,6–(Man-α1,3)Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn712 GlcNAc 
ResidueStructure
Chain A  
 Asn61 Man-α1,6–(Man-α1,2–Man-α1,2–Man-α1,3)Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn211 GlcNAc 
 Asn252 Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn315 GlcNAc-β1,4–GlcNAc 
 Asn322 Man-α1,2–Man-α1,6–(Man-α1,2–Man-α1,3-)Man-α1,6–(Man-α1,2–Man-α1,3)Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn442 Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn523 Man-α1,2–Man-α1,6-Man-α1,6–(Man-α1,3-)Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn564 Man-α1,6–(Man-α1,2-Man-α1,3-)Man-α1,6–Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn712 GlcNAc 
Chain B  
 Asn61 Man-α1,6–(Man-α1,2-Man–α1,2-Man-α1,3)Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn211 GlcNAc 
 Asn252 Man-α1,3–Man-α1,6–(Man-α1,3)Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn315 Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn322 Man-α1,2-Man-α1,6-(Man-α1,2–Man-α1,3)Man-α1,6–(Man-α1,2–Man-α1,3)Man-β1,4-GlcNAc-β1,4–GlcNAc 
 Asn442 Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn523 Man-α1,2–Man-α1,6-Man-α1,6–(Man-α1,3)Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn564 Man-α1,6–(Man-α1,2–Man-α1,3)Man-α1,6–(Man-α1,3)Man-β1,4–GlcNAc-β1,4–GlcNAc 
 Asn712 GlcNAc 

N-glycan structures

Figure 2
N-glycan structures

N-glycan at the Asn322 (stereo view) (A), Asn422 (B) and Asn523 (C) sites in the chain B are shown. GlcNAc, β-linked Man and α-linked Man are coloured as green, orange and yellow respectively. The N-glycans and polypeptides in the chain A are coloured as cyan.

Figure 2
N-glycan structures

N-glycan at the Asn322 (stereo view) (A), Asn422 (B) and Asn523 (C) sites in the chain B are shown. GlcNAc, β-linked Man and α-linked Man are coloured as green, orange and yellow respectively. The N-glycans and polypeptides in the chain A are coloured as cyan.

With regard to the GH3 family BGL structures, two members contain N-glycans. In the crystal structure of H. vulgare ExoI, N-glycosylations were observed at three sites (Asn221, Asn498 and Asn600), and a six sugar N-glycan structure derived from the plant (one Fuc, two Man and three GlcNAc) was modelled at Asn498 (Supplementary Figure S1B) [33]. The crystal structure of T. reesei Bgl1 contains two single GlcNAc residues at two sites (Asn208 and Asn310), which were probably truncated by endoglycosidase H treatment (Supplementary Figure S1E). In the crystal structure of AaBGL1, 32 GlcNAc and 56 Man moieties, which comprise 18 N-glycans, were included in the dimeric final model. This corresponds to 7.64 kDa per monomer and approximately 32% of the carbohydrate moiety of the partially deglycosylated sample as estimated by SDS/PAGE. To the best of our knowledge, this is one of the largest N-glycan structures observed in protein crystal structures. The crystal structure of glucoamylase from Aspergillus awamori var. X100 contains two high-mannose N-glycan sites (Asn171 and Asn395) [52]. Asn171 has five sugar residues (three Man and two GlcNAc), whereas Asn395 has eight sugar residues (six Man and two GlcNAc). In the crystal structure of RGase A [A. aculeatus (strain KSM) rhamnogalacturonase A], which is heterologously overexpressed in A. oryzae A1560, two N-linked and 18 O-linked glycosylation sites were observed [53]. In the O-linked sites, single Man-α1 sugars are linked to threonine residues. Asn32 has five sugar residues (three Man and two GlcNAc) and Asn299 has three sugar residues (one Man and two GlcNAc). Interestingly, in all of the three highly glycosylated enzymes from the Aspergillus species, distribution of the glycosylation sites are spatially biased to one side of the molecules. In the AaBGL1 and A. awamori glucoamylase structures, the N-glycosylation sites are located at the ‘front’ side, surrounding the substrate-binding pocket (Figures 1A and 1B) [52]. In contrast, the glycosylation sites of RGase A are mainly located at the ‘back’ side of the molecule [53].

The crystal packing of AaBGL1 shows that the carbohydrate moiety is involved in the packing interfaces (Supplementary Figure S4a at http://www.biochemj.org/bj/452/bj4520211add.htm). The crystal-packing interfaces can be divided into protein–protein, protein–carbohydrate and carbohydrate–carbohydrate types. The third type of the packing is only present as a hydrogen bond (3.2 Å) between the GlcNAc of Asn221 site in the chain B and one of the three terminal Man residues of Asn322 site in the chain A (Supplementary Figure S4b). Within the solvent channel of the crystal packing, there is sufficient room for the disordered carbohydrate moieties (14 kDa and 65% of total carbohydrates). When the total molecular mass (115 kDa) of the sample is assumed to be packed in the AaBGL1 crystal, the Matthews coefficient (VM) and solvent content (Vs) are estimated to be 2.44 Å3/Da and 49.7% respectively. In the crystal structure of RGase A, ‘carbohydrate interfaces’ by O-linked single Man residues mainly dictate the crystal packing, and only a few protein–protein interactions were found [53]. Although highly glycosylated proteins are generally difficult to crystallize, there are at least three exceptions to this rule (A. awamori glucoamylase, RGase A and AaBGL1). In some cases, such as AaBGL1 and RGase A, glycosylated moieties can be involved in crystal packing.

Active site and ligand complex structures

The active site and the catalytic residues of AaBGL1 are located at the domain interface between the barrel and the α/β sandwich domains. From the crystal structure, the nucleophile and the acid/base residues were assigned to be Asp280 in the barrel domain and Glu509 in the α/β sandwich domain respectively. The distance between the two catalytic residues is 5.6 Å, which is a typical distance for retaining GHs [54]. The position of the catalytic residues in the crystal structure is consistent with the previous results of biochemical determination of the catalytic residues using BGLs from A. niger B1 (nucleophile) [55] and from A. niger ASKU28 (acid/base) [56]. In the substrate-binding pocket of the substrate-free structure, the molecules derived from the crystallization buffer (an acetate ion at subsite −1 and an MPD molecule at subsite +1) were bound (Supplementary Figure S5a at http://www.biochemj.org/bj/452/bj4520211add.htm). To clarify the interactions at the active site of AaBGL1, complex crystals were prepared by soaking the crystals in the crystallization buffer supplemented with various ligands: four azasugar-type glucosidase inhibitors, Glc and a S-glycosyl analogue TCB. The complex structures were determined at 1.90–2.45 Å resolutions (Supplementary Table S2 and Supplementary Figure S5b–S5g). The ligand complex structures are shown in Figure 3.

Active-site structures of AaBGL1

Figure 3
Active-site structures of AaBGL1

IFG (A), DNJ (B), CgB2 (C), CTS (D), Glc (E) and TCB (F) complexes. The ligands, phenylalanine, tryptophan and tyrosine are shown as green, yellow, olive and orange sticks respectively. Asp280 (red) and Glu509 (blue) are the nucleophile and the acid/base catalyst residues respectively.

Figure 3
Active-site structures of AaBGL1

IFG (A), DNJ (B), CgB2 (C), CTS (D), Glc (E) and TCB (F) complexes. The ligands, phenylalanine, tryptophan and tyrosine are shown as green, yellow, olive and orange sticks respectively. Asp280 (red) and Glu509 (blue) are the nucleophile and the acid/base catalyst residues respectively.

All of the azasugars and one Glc molecule are bound to subsite −1 (Figures 3A–3E). The position and recognition of these ligands at subsite −1 are very similar. Superimposition of these structures illustrates that subsite −1 is precisely recognized (Figure 4). The two-carbon bridge of CgB2, which is a relatively weak inhibitor compared with other azasugars, appears to cause a slight steric hindrance with Asp92. Asp92, Arg156, Lys189 and His190 form hydrogen bonds with the sugar hydroxyls. These residues correspond to Asp95, Arg158, Lys206 and His207 of H. vulgare ExoI and are highly conserved in GH3 family BGLs. As shown in Supplementary Figure S6 (at http://www.biochemj.org/bj/452/bj4520211add.htm), the glucose-recognizing interactions at subsite −1 in AaBGL1 (green) and T. neapolitana Bgl3B (cyan) are almost identical. Trp281 of AaBGL1 forms an aromatic stacking interaction with the sugar ring at subsite −1. The conformation of Trp2811=−92.5°) is similar to those of equivalent tryptophan residues in T. neapolitana Bgl3B (Trp243, ϕ1=−92.5°) and K. marxianus BglI (Trp226, ϕ1=−90°) [35]. In contrast, the equivalent tryptophan residue (Trp286) of H. vulgare ExoI has a quite different conformation (ϕ1=−169°) and constitutes subsite +1 [34]. The Trp281 residue equivalent (Trp262) of A. niger BGL, a close homologue of AaBGL1, is a key residue for determining the ratio of hydrolytic and transglycosidic activities [57].

Superimposition of complex structures

Figure 4
Superimposition of complex structures

Stereo views of the superimposition of IFG (green), DNJ (cyan), CgB2 (yellow), CTS (magenta), Glc (grey) and TCB (slate blue) complex structures are shown. Protein residue side chains of the Glc complex are shown as sticks.

Figure 4
Superimposition of complex structures

Stereo views of the superimposition of IFG (green), DNJ (cyan), CgB2 (yellow), CTS (magenta), Glc (grey) and TCB (slate blue) complex structures are shown. Protein residue side chains of the Glc complex are shown as sticks.

In the Glc complex structure, the other Glc molecule is bound to subsite +1 (Figure 3E). The Glc molecule at subsite +1 is clamped by Tyr511 (upper side) and Trp68 and Phe305 (lower side). The Glc at subsite +1 points its C6 hydroxyl to C1 hydroxyl (β-anomer) of Glc at subsite −1 (O6–O1 distance=2.4 Å). This is consistent with the report that the main transglycosylation product of A. niger BGL is gentiobiose (Glc-β1,6-Glc) [58,59]. The substrate specificity of another close homologue of AaBGL1, BGL from A. oryzae, against various pyranose substrates and Glc–Glc disaccharides has been studied [60]. Consistent with the structural feature of AaBGL1, the subsite −1 of A. oryzae BGL has a strict stereochemical requirement to β-D-glucopyranoside, but its subsite +1 has a relatively loose stereochemical specificity as it shows similarly high activities against cellobiose, gentiobiose, laminaribiose (Glc-β1,3-Glc), sophorose (Glc-β1,2–Glc) and pNP-Glc. The kcat/Km values of A. oryzae BGL for all of the five substrates were within the range 3.3×105–2.6×106 M−1·s−1.

In the TCB complex, the non-reducing end Glc moiety at subsite −1 is significantly displaced toward subsite +1 because the hydrogen-bond interactions with Arg169, Lys190 and His190 are absent (Figure 3F). Superimposition of the TCB complex with other structures revealed that the reducing-end Glc moiety at subsite +1 is also displaced slightly away from subsite −1 (Figure 4). Therefore the TCB molecule in the complex structure does not reflect the exact subsite positions. Because subsite +1 is exclusively formed by hydrophobic interactions, this subsite appears to have a strong affinity; however, there may be some positional looseness. The glycosidic S atom is significantly located away from the Oϵ1 atom of Glu509 (3.6 Å). In the case of H. vulgare ExoI complexed with TCB the corresponding distance is 2.75 Å, clearly indicating that the glutamic acid residue (Glu491 in H. vulgare ExoI) is the acid/base residue [61].

The MPD molecule, which is used as a cryoprotectant, is bound at subsite +1 in the IFG, CgB2 and CTS complex structures (Figures 3A, 3C and 3D), as well as in the substrate-free (acetate-bound) structure. In all of the seven structures, an MPD molecule is also bound at a position corresponding to subsite +4. As shown in Figure 5(A) (the molecular surface of AaBGL1), there is a long cleft extending from subsite +1 to subsite >5, and Trp358 is present at subsite +4. Consistent with this feature, AaBGL1 exhibits high activity against cellooligosaccharides with a DP greater than 4 [16].

Molecular surface of the active-site cleft of GH3 family BGLs (Glc complex)

Figure 5
Molecular surface of the active-site cleft of GH3 family BGLs (Glc complex)

AaBGL1 (A), H. vulgare ExoI (B), T. neapolitana Bgl3B (C) and K. marxianus BglI (D). The acid/base catalyst, phenylalanine, tryptophan and tyrosine are coloured as blue, yellow, olive and orange respectively. In (D) the PA14 domain is shown as ribbon and a stick model.

Figure 5
Molecular surface of the active-site cleft of GH3 family BGLs (Glc complex)

AaBGL1 (A), H. vulgare ExoI (B), T. neapolitana Bgl3B (C) and K. marxianus BglI (D). The acid/base catalyst, phenylalanine, tryptophan and tyrosine are coloured as blue, yellow, olive and orange respectively. In (D) the PA14 domain is shown as ribbon and a stick model.

Comparison of the active site with other GH3 family BGLs

The active sites of the four representative GH3 family BGLs (AaBGL1, H. vulgare ExoI, T. neapolitana Bgl3B and K. marxianus Bgl1) were compared, mainly focusing on the aromatic residues at subsite +1 (Figures 5 and 6). As discussed above, subsite +1 of AaBGL1 is formed by Tyr511, Phe305 and Trp68 (Figure 6A), and there is a long cleft along with Trp358 (Figure 5A). H. vulgare ExoI has a narrower coin-slot-like subsite +1, which is formed by Trp434 and Trp286 (Figure 5B). The side chain of Trp286 is flipped to form the subsite +1, but is not involved in formation of subsite −1 (Figure 6B) [62]. This structural feature of H. vulgare ExoI constitutes a binding pocket that is suitable for various disaccharides [61]. In T. neapolitana Bgl3B, there is no aromatic residue at subsite +1 (Figure 5C). The substrate-binding pocket of T. neapolitana Bgl3B is widely open to the solvent, and it can accommodate various alcohol acceptors to synthesize alkyl glucosides by a transglycosylation reaction [34]. The subsite +1 of K. marxianus Bgl1 is surrounded by many aromatic residues (Figure 6D). However, a PA14 domain covers the active site of K. marxianus Bgl1 to prevent binding of longer cellooligosaccharides (DP>4) [35]. In comparison with the other three GH3 family BGLs, AaBGL1 appears to have the most suitable binding pocket for cellooligosaccharides. As shown in Figure 6, Trp68, Phe305 and Tyr511 of AaBGL1 are located in short loop regions that are not present in other GH3 family BGLs.

The active site and aromatic residue at subsite +1 of GH3 family BGLs (Glc complex)

Figure 6
The active site and aromatic residue at subsite +1 of GH3 family BGLs (Glc complex)

AaBGL1 (A), H. vulgare ExoI (B), T. neapolitana Bgl3B (C) and K. marxianus BglI (D). The nucleophile, acid/base catalyst, phenylalanine, tryptophan and tyrosine are coloured as red, blue, yellow, olive and orange respectively.

Figure 6
The active site and aromatic residue at subsite +1 of GH3 family BGLs (Glc complex)

AaBGL1 (A), H. vulgare ExoI (B), T. neapolitana Bgl3B (C) and K. marxianus BglI (D). The nucleophile, acid/base catalyst, phenylalanine, tryptophan and tyrosine are coloured as red, blue, yellow, olive and orange respectively.

In comparison with T. reesei Bgl1, the structures of the two fungal GH3 family BGLs are very similar at subsite −1 (Supplementary Figure S7A at http://www.biochemj.org/bj/452/bj4520211add.htm). However, there is a significant difference at subsite +1. Two aromatic residues of T. reesei Bgl1 (Trp37 and Phe260) have different side-chain conformations from those of corresponding residues in AaBGL1 (Trp68 and Phe305). T. reesei Bgl1 lacks a tryptophan residue and a loop corresponding to Trp358 in AaBGL1, but Tyr68 is present at the other side of the pocket. These three aromatic residues (Trp37, Tyr68 and Phe260) form the entrance of the active-site pocket of T. reesei Bgl1 (Supplementary Figure S7B). Therefore the diverse structures of the loops forming subsite +1 of GH3 family BGLs result in the diversified substrate specificity of these enzymes.

Ambiguity of the catalytic residue positions of the GH3 family enzymes

The GH3 family is one of the most divergent families in the CAZy database. The catalytic nucleophile of GH3 family enzymes is always present just after the β7 strand in the TIM barrel domain. However, the positions and identities of the acid/base catalyst are not completely conserved. In GH3 family BGLs, the acid/base catalyst (Glu590 in AaBGL1 and Glu491 in H. vulgare ExoI) is located in the α/β sandwich domain (Figure 7). Surprisingly, Litzinger et al. [32] reported that the acid/base catalyst of a two-domain NagZ from Bacillus subtilis is a unique aspartic acid–histidine residue dyad (Asp232 and His234) located in the TIM barrel domain (Figure 7C), and the histidine residue is proposed to function as a proton donor. A DSH motif containing the aspartic acid–histidine residue dyad is highly conserved in both two- and one-domain NagZ enzymes. In the structure of a one-domain NagZ from Vibrio cholerae [63], however, a flexible loop carrying the aspartic acid–histidine residue dyad (Asp171 and His173) is flipped outward, and the nucleophile (Asp242) is distorted (Figure 7D). In the case of GH3 family BGLs, the catalytic acid/base glutamic acid residue is located in a less-conserved region of the α/β sandwich domain. As shown in Figure 7, this region has distinct main chain structures between AaBGL1 (loop region) and H. vulgare ExoI (a short helix). The ambiguity of the catalytic residue is one of the most interesting aspects of GH3 family enzymes, and further structural analysis will clarify the whole picture of this diverse family.

Locations and identities of the catalytic residues of the GH3 family enzymes

Figure 7
Locations and identities of the catalytic residues of the GH3 family enzymes

AaBGL1 (Glc complex) (A), H. vulgare ExoI (gluco-phenylimidazole complex, PDB code 1X38) (B), B. subtilis two-domain NagZ (PUGNAc complex, PDB code 3NVD) (C) and (D) V. cholerae one-domain NagZ (PUGNAc complex, PDB code 2OXN). The α/β-sandwich domain, the FnIII domain and the insertion region are coloured as green, yellow and red respectively. The ligand in the active site, nucleophile, acid/base catalyst and acid/base-supporting aspartic acid are coloured as yellow, red, blue and orange respectively.

Figure 7
Locations and identities of the catalytic residues of the GH3 family enzymes

AaBGL1 (Glc complex) (A), H. vulgare ExoI (gluco-phenylimidazole complex, PDB code 1X38) (B), B. subtilis two-domain NagZ (PUGNAc complex, PDB code 3NVD) (C) and (D) V. cholerae one-domain NagZ (PUGNAc complex, PDB code 2OXN). The α/β-sandwich domain, the FnIII domain and the insertion region are coloured as green, yellow and red respectively. The ligand in the active site, nucleophile, acid/base catalyst and acid/base-supporting aspartic acid are coloured as yellow, red, blue and orange respectively.

Conclusions

The crystal structure of a fungal GH3 family BGL, AaBGL1, has been determined. The AaBGL1 crystal structure contains many large N-glycan chains that are resistant to endoglycosidase treatment. In addition to the TIM barrel, α/β sandwich and FnIII domains, a prominent insertion loop is found. The insertion loop has not been found in other structures of GH3 family BGLs and covers a wide surface area of the enzyme molecule. Although the biological function of these structural elements (N-glycans and insertion loop) remains to be studied, these elements most likely contribute to the solubility and stability of the enzyme in aqueous solution (e.g. protection from proteolytic cleavage). The active site of AaBGL1 was investigated via the complex structures with various inhibitors and ligands. Subsite −1 was determined exactly using azasugar inhibitors and Glc, and the interactions are similar to those of other GH3 family BGLs. Subsite +1 is formed by three aromatic residues located in short loop regions that are uniquely present in AaBGL1. A long cleft extending from subsite +1 is present, and it appears to be suitable for binding long cellooligosaccharides. Because fungal GH3 family BGLs are widely and commercially used as an effective cellulase supplement, the crystal structure of AaBGL1 will provide an important platform for enzyme engineering to improve the conversion of plant cellulosic biomass.

Abbreviations

     
  • AaBGL1

    Aspergillus aculeatus BGL1

  •  
  • BGL

    β-glucosidase

  •  
  • DP

    degree of polymerization

  •  
  • CgB2

    calystegine B2

  •  
  • CTS

    castanospermine

  •  
  • DNJ

    1-deoxynojirimycin

  •  
  • FnIII

    fibronectin type III

  •  
  • GH

    glycoside hydrolase

  •  
  • IFG

    isofagomine

  •  
  • MPD

    2-methyl-2,4-pentanediol

  •  
  • NagZ

    N-acetyl-β-D-glucosaminidase

  •  
  • pNP-Glc

    p-nitrophenyl-β-D-glucopyranoside

  •  
  • RGase

    A, A. aculeatus rhamnogalacturonase A

  •  
  • RMSD

    root mean square deviation

  •  
  • TCB

    thiocellobiose

  •  
  • TIM

    triosephosphateisomerase

AUTHOR CONTRIBUTION

Kentaro Suzuki and Shinya Fushinobu performed crystallization, X-ray data collection and structure determination. Jun-ichi Sumitani purified the protein. Young-Woo Nam measured the enzyme activity. Toru Nishimaki and Shuji Tani constructed the A. oryzae heterologous overexpression system. Shinya Fushinobu, Takayoshi Wakagi and Takashi Kawaguchi conceived and supervised the project. Shinya Fushinobu analysed the results and wrote the paper.

We thank the staff of the Photon Factory and SPring-8 for the X-ray data collection. Synchrotron beam time at SPring-8 was supported by the Priority Program for Disaster-Affected Quantum Beam Facilities (2011G080 for KEK-PF and 2011A1908 for SPring-8).

FUNDING

This work was supported by The New Energy and Industrial Technology Development Organization and by JSPS (Japan Society for the Promotion of Science) KAKENHI [grant number 24380053 (to S.F.)].

References

References
1
Ragauskas
 
A. J.
Williams
 
C. K.
Davison
 
B. H.
Britovsek
 
G.
Cairney
 
J.
Eckert
 
C. A.
Frederick
 
W. J.
Hallett
 
J. P.
Leak
 
D. J.
Liotta
 
C. L.
, et al 
The path forward for biofuels and biomaterials
Science
2006
, vol. 
311
 (pg. 
484
-
489
)
2
Kamm
 
B.
Kamm
 
M.
 
Biorefineries: multi product processes
Adv. Biochem. Eng. Biotechnol.
2007
, vol. 
105
 (pg. 
175
-
204
)
3
Himmel
 
M. E.
Ding
 
S. Y.
Johnson
 
D. K.
Adney
 
W. S.
Nimlos
 
M. R.
Brady
 
J. W.
Foust
 
T. D.
 
Biomass recalcitrance: engineering plants and enzymes for biofuels production
Science
2007
, vol. 
315
 (pg. 
804
-
807
)
4
Nakari-Setala
 
T.
Paloheimo
 
M.
Kallio
 
J.
Vehmaanpera
 
J.
Penttila
 
M.
Saloheimo
 
M.
 
Genetic modification of carbon catabolite repression in Trichoderma reesei for improved protein production
Appl. Environ. Microbiol.
2009
, vol. 
75
 (pg. 
4853
-
4860
)
5
van den Brink
 
J.
de Vries
 
R. P.
 
Fungal enzyme sets for plant polysaccharide degradation
Appl. Microbiol. Biotechnol.
2011
, vol. 
91
 (pg. 
1477
-
1492
)
6
Kubicek
 
C. P.
 
Release of carboxymethylcellulose and β-glucosidase from cell walls of Trichoderma reesei
Eur. J. Appl. Miciobiol. Biotechnol.
1981
, vol. 
13
 (pg. 
226
-
231
)
7
Xin
 
Z.
Yinbo
 
Q.
Peiji
 
G.
 
Acceleration of ethanol production from paper mill waste fiber by supplementation with beta-glucosidase
Enzyme Microb. Technol.
1993
, vol. 
15
 (pg. 
62
-
65
)
8
Berlin
 
A.
Maximenko
 
V.
Gilkes
 
N.
Saddler
 
J.
 
Optimization of enzyme complexes for lignocellulose hydrolysis
Biotechnol. Bioeng.
2007
, vol. 
97
 (pg. 
287
-
296
)
9
Chen
 
M.
Zhao
 
J.
Xia
 
L.
 
Enzymatic hydrolysis of maize straw polysaccharides for the production of reducing sugars
Carbohydr. Polym.
2008
, vol. 
71
 (pg. 
411
-
415
)
10
Barnett
 
C. C.
Berka
 
R. M.
Fowler
 
T.
 
Cloning and amplification of the gene encoding an extracellular beta-glucosidase from Trichoderma reesei: evidence for improved rates of saccharification of cellulosic substrates
Biotechnology
1991
, vol. 
9
 (pg. 
562
-
567
)
11
Rahman
 
Z.
Shida
 
Y.
Furukawa
 
T.
Suzuki
 
Y.
Okada
 
H.
Ogasawara
 
W.
Morikawa
 
Y.
 
Application of Trichoderma reesei cellulase and xylanase promoters through homologous recombination for enhanced production of extracellular β-glucosidase I
Biosci. Biotechnol. Biochem.
2009
, vol. 
73
 (pg. 
1083
-
1089
)
12
Nakazawa
 
H.
Kawai
 
T.
Ida
 
N.
Shida
 
Y.
Kobayashi
 
Y.
Okada
 
H.
Tani
 
S.
Sumitani
 
J.
Kawaguchi
 
T.
Morikawa
 
Y.
Ogasawara
 
W.
 
Construction of a recombinant Trichoderma reesei strain expressing Aspergillus aculeatus β-glucosidase 1 for efficient biomass conversion
Biotechnol. Bioeng.
2012
, vol. 
109
 (pg. 
92
-
99
)
13
Sternberg
 
D.
Vijayakumar
 
P.
Reese
 
E. T.
 
β-Glucosidase: microbial production and effect on enzymatic hydrolysis of cellulose
Can. J. Microbiol.
1977
, vol. 
23
 (pg. 
139
-
147
)
14
Chauve
 
M.
Mathis
 
H.
Huc
 
D.
Casanave
 
D.
Monot
 
F.
Lopes Ferreira
 
N.
 
Comparative kinetic analysis of two fungal β-glucosidases
Biotechnol. Biofuels.
2010
, vol. 
3
 pg. 
3
 
15
Sakamoto
 
R.
Kanamoto
 
J.
Arai
 
M.
Murao
 
S.
 
Purification and physicochemical properties of three β-glucosidases from Aspergillus aculeatus No. F-50
Agric. Biol. Chem.
1985
, vol. 
49
 (pg. 
1275
-
1281
)
16
Sakamoto
 
R.
Arai
 
M.
Murao
 
S.
 
Enzymic properties of three β-glucosidases from Aspergillus aculatus No. F-50
Agric. Biol. Chem.
1985
, vol. 
49
 (pg. 
1283
-
1290
)
17
Murao
 
S.
Sakamoto
 
R.
Arai
 
M.
 
Cellulases of Aspergillus aculeatus
Methods Enzymol.
1988
, vol. 
160C
 (pg. 
274
-
299
)
18
Kawai
 
T.
Nakazawa
 
H.
Ida
 
N.
Okada
 
H.
Tani
 
S.
Sumitani
 
J.
Kawaguchi
 
T.
Ogasawara
 
W.
Morikawa
 
Y.
Kobayashi
 
Y.
 
Analysis of the saccharification capability of high-functional cellulase JN11 for various pretreated biomasses through a comparison with commercially available counterparts
J. Ind. Microbiol. Biotechnol.
2012
, vol. 
39
 (pg. 
1741
-
1749
)
19
Murai
 
T.
Ueda
 
M.
Kawaguchi
 
T.
Arai
 
M.
Tanaka
 
A.
 
Assimilation of cellooligosaccharides by a cell surface-engineered yeast expressing β-glucosidase and carboxymethylcellulase from Aspergillus aculeatus
Appl. Environ. Microbiol.
1998
, vol. 
64
 (pg. 
4857
-
4861
)
20
Fujita
 
Y.
Takahashi
 
S.
Ueda
 
M.
Tanaka
 
A.
Okada
 
H.
Morikawa
 
Y.
Kawaguchi
 
T.
Arai
 
M.
Fukuda
 
H.
Kondo
 
A.
 
Direct and efficient production of ethanol from cellulosic material with a yeast strain displaying cellulolytic enzymes
Appl. Environ. Microbiol.
2002
, vol. 
68
 (pg. 
5136
-
5141
)
21
Fujita
 
Y.
Ito
 
J.
Ueda
 
M.
Fukuda
 
H.
Kondo
 
A.
 
Synergistic saccharification, and direct fermentation to ethanol, of amorphous cellulose by use of an engineered yeast strain codisplaying three types of cellulolytic enzyme
Appl. Environ. Microbiol.
2004
, vol. 
70
 (pg. 
1207
-
1212
)
22
Tokuhiro
 
K.
Ishida
 
N.
Kondo
 
A.
Takahashi
 
H.
 
Lactic fermentation of cellobiose by a yeast strain displaying β-glucosidase on the cell surface
Appl. Microbiol. Biotechnol.
2008
, vol. 
79
 (pg. 
481
-
488
)
23
Ito
 
J.
Kosugi
 
A.
Tanaka
 
T.
Kuroda
 
K.
Shibasaki
 
S.
Ogino
 
C.
Ueda
 
M.
Fukuda
 
H.
Doi
 
R. H.
Kondo
 
A.
 
Regulation of the display ratio of enzymes on the Saccharomyces cerevisiae cell surface by the immunoglobulin G and cellulosomal enzyme binding domains
Appl. Environ. Microbiol.
2009
, vol. 
75
 (pg. 
4149
-
4154
)
24
Yanase
 
S.
Yamada
 
R.
Kaneko
 
S.
Noda
 
H.
Hasunuma
 
T.
Tanaka
 
T.
Ogino
 
C.
Fukuda
 
H.
Kondo
 
A.
 
Ethanol production from cellulosic materials using cellulase-expressing yeast
Biotechnol. J.
2010
, vol. 
5
 (pg. 
449
-
455
)
25
Yanase
 
S.
Hasunuma
 
T.
Yamada
 
R.
Tanaka
 
T.
Ogino
 
C.
Fukuda
 
H.
Kondo
 
A.
 
Direct ethanol production from cellulosic materials at high temperature using the thermotolerant yeast Kluyveromyces marxianus displaying cellulolytic enzymes
Appl. Microbiol. Biotechnol.
2010
, vol. 
88
 (pg. 
381
-
388
)
26
Sakamoto
 
T.
Hasunuma
 
T.
Hori
 
Y.
Yamada
 
R.
Kondo
 
A.
 
Direct ethanol production from hemicellulosic materials of rice straw by use of an engineered yeast strain codisplaying three types of hemicellulolytic enzymes on the surface of xylose-utilizing Saccharomyces cerevisiae cells
J. Biotechnol.
2012
, vol. 
158
 (pg. 
203
-
210
)
27
Okano
 
K.
Zhang
 
Q.
Yoshida
 
S.
Tanaka
 
T.
Ogino
 
C.
Fukuda
 
H.
Kondo
 
A.
 
D-Lactic acid production from cellooligosaccharides and beta-glucan using L-LDH gene-deficient and endoglucanase-secreting Lactobacillus plantarum
Appl. Microbiol. Biotechnol.
2010
, vol. 
85
 (pg. 
643
-
650
)
28
Tsuchidate
 
T.
Tateno
 
T.
Okai
 
N.
Tanaka
 
T.
Ogino
 
C.
Kondo
 
A.
 
Glutamate production from β-glucan using endoglucanase-secreting Corynebacterium glutamicum
Appl. Microbiol. Biotechnol.
2011
, vol. 
90
 (pg. 
895
-
901
)
29
Kim
 
H. W.
Ishikawa
 
K.
 
Complete saccharification of cellulose at high temperature using endocellulase and β-glucosidase from Pyrococcus sp
J. Microbiol. Biotechnol.
2010
, vol. 
20
 (pg. 
889
-
892
)
30
Cantarel
 
B. L.
Coutinho
 
P. M.
Rancurel
 
C.
Bernard
 
T.
Lombard
 
V.
Henrissat
 
B.
 
The Carbohydrate-Active EnZymes database (CAZy): an expert resource for glycogenomics
Nucleic Acids Res.
2009
, vol. 
37
 (pg. 
D233
-
D238
)
31
Kawaguchi
 
T.
Enoki
 
T.
Tsurumaki
 
S.
Sumitani
 
J.
Ueda
 
M.
Ooi
 
T.
Arai
 
M.
 
Cloning and sequencing of the cDNA encoding β-glucosidase 1 from Aspergillus aculeatus
Gene
1996
, vol. 
173
 (pg. 
287
-
288
)
32
Litzinger
 
S.
Fischer
 
S.
Polzer
 
P.
Diederichs
 
K.
Welte
 
W.
Mayer
 
C.
 
Structural and kinetic analysis of Bacillus subtilis N-acetylglucosaminidase reveals a unique Asp–His dyad mechanism
J. Biol. Chem.
2010
, vol. 
285
 (pg. 
35675
-
35684
)
33
Varghese
 
J. N.
Hrmova
 
M.
Fincher
 
G. B.
 
Three-dimensional structure of a barley β-D-glucan exohydrolase, a family 3 glycosyl hydrolase
Structure
1999
, vol. 
7
 (pg. 
179
-
190
)
34
Pozzo
 
T.
Pasten
 
J. L.
Karlsson
 
E. N.
Logan
 
D. T.
 
Structural and functional analyses of β-glucosidase 3B from Thermotoga neapolitana: a thermostable three-domain representative of glycoside hydrolase 3
J. Mol. Biol.
2010
, vol. 
397
 (pg. 
724
-
739
)
35
Yoshida
 
E.
Hidaka
 
M.
Fushinobu
 
S.
Koyanagi
 
T.
Minami
 
H.
Tamaki
 
H.
Kitaoka
 
M.
Katayama
 
T.
Kumagai
 
H.
 
Role of a PA14 domain in determining substrate specificity of a glycoside hydrolase family 3 β-glucosidase from Kluyveromyces marxianus
Biochem. J.
2010
, vol. 
431
 (pg. 
39
-
49
)
36
Nakatani
 
Y.
Cutfield
 
S. M.
Cowieson
 
N. P.
Cutfield
 
J. F.
 
Structure and activity of exo-1,3/1,4-β-glucanase from marine bacterium Pseudoalteromonas sp. BB1 showing a novel C-terminal domain
FEBS J.
2012
, vol. 
279
 (pg. 
464
-
478
)
37
Harvey
 
A. J.
Hrmova
 
M.
De Gori
 
R.
Varghese
 
J. N.
Fincher
 
G. B.
 
Comparative modeling of the three-dimensional structures of family 3 glycoside hydrolases
Proteins
2000
, vol. 
41
 (pg. 
257
-
269
)
38
Cournoyer
 
B.
Faure
 
D.
 
Radiation and functional specialization of the family-3 glycoside hydrolases
J. Mol. Microbiol. Biotechnol.
2003
, vol. 
5
 (pg. 
190
-
198
)
39
Minetoki
 
T.
Tsuboi
 
H.
Koda
 
A.
Ozeki
 
K.
 
Development of high expression system with the improved promoter using the cis-acting element in Aspergillus species
J. Biol. Macromol.
2003
, vol. 
3
 (pg. 
89
-
96
)
40
Kanamasa
 
S.
Takada
 
G.
Kawaguchi
 
T.
Sumitani
 
J.
Arai
 
M.
 
Overexpression and purification of Aspergillus aculeatus β-mannosidase and analysis of the integrated gene in Aspergillus oryzae
J. Biosci. Bioeng.
2001
, vol. 
92
 (pg. 
131
-
137
)
41
Rashid
 
M. H.
Javed
 
M. R.
Kawaguchi
 
T.
Sumitani
 
J.
Arai
 
M.
 
Improvement of Aspergillus oryzae for hyperproduction of endoglucanase: expression cloning of cmc-1 gene of Aspergillus aculeatus
Biotechnol. Lett.
2008
, vol. 
30
 (pg. 
2165
-
2172
)
42
Otwinowski
 
Z.
Minor
 
W.
 
Processing of X-ray diffraction data collected in oscillation mode
Methods Enzymol.
1997
, vol. 
276
 (pg. 
307
-
326
)
43
Vagin
 
A.
Teplyakov
 
A.
 
MOLREP: an automated program for molecular replacement
J. Appl. Cryst.
1997
, vol. 
30
 (pg. 
1022
-
1025
)
44
Perrakis
 
A.
Morris
 
R.
Lamzin
 
V. S.
 
Automated protein model building combined with iterative structure refinement
Nat. Struct. Biol.
1999
, vol. 
6
 (pg. 
458
-
463
)
45
Emsley
 
P.
Lohkamp
 
B.
Scott
 
W. G.
Cowtan
 
K.
 
Features and development of Coot
Acta Crystallogr., Sect. D: Biol. Crystallogr.
2010
, vol. 
66
 (pg. 
486
-
501
)
46
Murshudov
 
G. N.
Vagin
 
A. A.
Dodson
 
E. J.
 
Refinement of macromolecular structures by the maximum-likelihood method
Acta Crystallogr., Sect. D: Biol. Crystallogr.
1997
, vol. 
53
 (pg. 
240
-
255
)
47
Krissinel
 
E.
Henrick
 
K.
 
Inference of macromolecular assemblies from crystalline state
J. Mol. Biol.
2007
, vol. 
372
 (pg. 
774
-
797
)
48
Holm
 
L.
Sander
 
C.
 
Dali: a network tool for protein structure comparison
Trends Biochem. Sci.
1995
, vol. 
20
 (pg. 
478
-
480
)
49
Chiba
 
Y.
Yamagata
 
Y.
Nakajima
 
T.
Ichishima
 
E.
 
A new high-mannose type N-linked oligosaccharide from Aspergillus carboxypeptidase
Biosci. Biotechnol. Biochem.
1992
, vol. 
56
 (pg. 
1371
-
1372
)
50
Archer
 
D. B.
Peberdy
 
J. F.
 
The molecular biology of secreted enzyme production by fungi
Crit. Rev. Biotechnol.
1997
, vol. 
17
 (pg. 
273
-
306
)
51
Kainz
 
E.
Gallmetzer
 
A.
Hatzl
 
C.
Nett
 
J. H.
Li
 
H.
Schinko
 
T.
Pachlinger
 
R.
Berger
 
H.
Reyes-Dominguez
 
Y.
Bernreiter
 
A.
, et al 
N-glycan modification in Aspergillus species
Appl. Environ. Microbiol.
2008
, vol. 
74
 (pg. 
1076
-
1086
)
52
Aleshin
 
A. E.
Hoffman
 
C.
Firsov
 
L. M.
Honzatko
 
R. B.
 
Refined crystal structures of glucoamylase from Aspergillus awamori var. X100
J. Mol. Biol.
1994
, vol. 
238
 (pg. 
575
-
591
)
53
Petersen
 
T. N.
Kauppinen
 
S.
Larsen
 
S.
 
The crystal structure of rhamnogalacturonase A from Aspergillus aculeatus: a right-handed parallel β-helix
Structure
1997
, vol. 
5
 (pg. 
533
-
544
)
54
Davies
 
G.
Henrissat
 
B.
 
Structures and mechanisms of glycosyl hydrolases
Structure
1995
, vol. 
3
 (pg. 
853
-
859
)
55
Dan
 
S.
Marton
 
I.
Dekel
 
M.
Bravdo
 
B. A.
He
 
S.
Withers
 
S. G.
Shoseyov
 
O.
 
Cloning, expression, characterization, and nucleophile identification of family 3, Aspergillus niger β-glucosidase
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
4973
-
4980
)
56
Thongpoo
 
P.
McKee
 
L. S.
Araujo
 
A. C.
Kongsaeree
 
P. T.
Brumer
 
H.
 
Identification of the acid/base catalyst of a glycoside hydrolase family 3 (GH3) β-glucosidase from Aspergillus niger ASKU28
Biochim. Biophys. Acta
2012
, vol. 
1830
 (pg. 
2739
-
2749
)
57
Seidle
 
H. F.
McKenzie
 
K.
Marten
 
I.
Shoseyov
 
O.
Huber
 
R. E.
 
Trp262 is a key residue for the hydrolytic and transglucosidic reactivity of the Aspergillus niger family 3 beta-glucosidase: substitution results in enzymes with mainly transglucosidic activity
Arch. Biochem. Biophys.
2005
, vol. 
444
 (pg. 
66
-
75
)
58
Seidle
 
H. F.
Marten
 
I.
Shoseyov
 
O.
Huber
 
R. E.
 
Physical and kinetic properties of the family 3 beta-glucosidase from Aspergillus niger which is important for cellulose breakdown
Protein J.
2004
, vol. 
23
 (pg. 
11
-
23
)
59
Seidle
 
H. F.
Huber
 
R. E.
 
Transglucosidic reactions of the Aspergillus niger family 3 β-glucosidase: qualitative and quantitative analyses and evidence that the transglucosidic rate is independent of pH
Arch. Biochem. Biophys.
2005
, vol. 
436
 (pg. 
254
-
264
)
60
Langston
 
J.
Sheehy
 
N.
Xu
 
F.
 
Substrate specificity of Aspergillus oryzae family 3 β-glucosidase
Biochim. Biophys. Acta
2006
, vol. 
1764
 (pg. 
972
-
978
)
61
Hrmova
 
M.
De Gori
 
R.
Smith
 
B. J.
Fairweather
 
J. K.
Driguez
 
H.
Varghese
 
J. N.
Fincher
 
G. B.
 
Structural basis for broad substrate specificity in higher plant β-D-glucan glucohydrolases
Plant Cell
2002
, vol. 
14
 (pg. 
1033
-
1052
)
62
Hrmova
 
M.
Varghese
 
J. N.
De Gori
 
R.
Smith
 
B. J.
Driguez
 
H.
Fincher
 
G. B.
 
Catalytic mechanisms and reaction intermediates along the hydrolytic pathway of a plant β-D-glucan glucohydrolase
Structure
2001
, vol. 
9
 (pg. 
1005
-
1016
)
63
Stubbs
 
K. A.
Balcewich
 
M.
Mark
 
B. L.
Vocadlo
 
D. J.
 
Small molecule inhibitors of a glycoside hydrolase attenuate inducible AmpC-mediated β-lactam resistance
J. Biol. Chem.
2007
, vol. 
282
 (pg. 
21382
-
21391
)

Author notes

The co-ordinates and structure factors have been deposited in the PDB under accession codes 4IIB, 4IIC, 4IID, 4IIE, 4IIF, 4IIG and 4IIH.

Supplementary data