Journal list menu

Volume 283, Issue 12 p. 2340-2353
Original Article
Free Access

Crystal structure and identification of a key amino acid for glucose tolerance, substrate specificity, and transglycosylation activity of metagenomic β-glucosidase Td2F2

Tomohiko Matsuzawa

Tomohiko Matsuzawa

Bioproduction Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki, Japan

Contributed equally.Search for more papers by this author
Toshinori Jo

Toshinori Jo

Department of Biotechnology, The University of Tokyo, Japan

Contributed equally.Search for more papers by this author
Taku Uchiyama

Taku Uchiyama

Department of Biomaterial Sciences, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Japan

Search for more papers by this author
Jenny A. Manninen

Jenny A. Manninen

Department of Biotechnology, The University of Tokyo, Japan

Search for more papers by this author
Takatoshi Arakawa

Takatoshi Arakawa

Department of Biotechnology, The University of Tokyo, Japan

Search for more papers by this author
Kentaro Miyazaki

Kentaro Miyazaki

Bioproduction Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki, Japan

Department of Computational Biology and Medical Sciences, Graduate School of Frontier Sciences, The University of Tokyo, Chiba, Japan

Search for more papers by this author
Shinya Fushinobu

Shinya Fushinobu

Department of Biotechnology, The University of Tokyo, Japan

Search for more papers by this author
Katsuro Yaoi

Corresponding Author

Katsuro Yaoi

Bioproduction Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki, Japan

Correspondence

K. Yaoi, Bioproduction Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba Central 6, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8566, Japan

Fax: +81 29 861 6226

Tel: +81 29 861 7867

E-mail: [email protected]

Search for more papers by this author
First published: 19 April 2016
Citations: 46

Abstract

β-Glucosidase Td2F2 isolated from a compost metagenome has high glucose tolerance and transglycosylation activity. In this study, we determined the high-resolution crystal structure of Td2F2. It has a unique structure at the −1 subsite that is important for substrate specificity but not for glucose tolerance. To elucidate the mechanism(s) of glucose tolerance, we isolated a glucose-sensitive Td2F2 mutant using random mutagenesis. In this mutant, Asn223 residue located between subsites +1 and +2 was mutated. The Asn223 mutation resulted in reduced glucose tolerance and transglycosylation activity, and drastically changed substrate specificity. These results indicate that the structure between subsites +1 and +2 is critical for the glucose tolerance and substrate specificity of Td2F2. Our findings shed light on the glucose tolerance and transglycosylation activity mechanisms of glycoside hydrolase family 1 β-glucosidases.

Database

The atomic coordinates and structure factors (codes 3WH5, 3WH6, 3WH8, 3WH7, 5AYB, and 5AYI) have been deposited in the Protein Data Bank (http://wwpdb.org/).

Abbreviations

  • GH
  • glycoside hydrolase
  • LB
  • Luria–Bertani
  • PCR
  • polymerase chain reaction
  • PDB
  • Protein Data Bank
  • pNP
  • p-nitrophenol
  • pNP-β-d-Fuc
  • p-nitrophenyl-β-d-fucopyranoside
  • pNP-β-d-Gal
  • p-nitrophenyl-β-d-galactopyranoside
  • pNP-β-d-Glc
  • p-nitrophenyl-β-d-glucopyranoside
  • SsBGL
  • Sulfolobus solfataricus β-glycosidase
  • TmBGL
  • Thermotoga maritima β-glucosidase
  • Introduction

    Many β-glucosidases (β-d-glucoside glucohydrolases, EC 3.2.1.21) belonging to glycoside hydrolase (GH) families 1, 3, 5, 9, 30, and 116 have been isolated from archaea [1, 2], bacteria [3, 4], fungi [5, 6], plants [7-10], and environmental metagenomes [11-13]. β-Glucosidases play vital roles in several biological processes, such as biomass conversion and saccharification. For example, they are important for the degradation of cello-oligosaccharides, including cellobiose, as well as the saccharification and refining of lignocellulosic biomass [14]. However, most β-glucosidases are significantly inhibited by glucose, a primary product of a β-glucosidase reaction [15]. This product inhibition is a critical bottleneck in the saccharification of biomass [16]. To overcome this problem, several glucose-tolerant β-glucosidases have been isolated and characterized [13, 17-19]. We also isolated a glucose-tolerant β-glucosidase, termed Td2F2, belonging to a glycoside hydrolase family 1 (GH1) from a compost metagenome [12]. It not only shows β-glucosidase activity on sophorose, laminaribiose, cellobiose, gentiobiose (in this order of catalytic efficiency), and the chromogenic substrate p-nitrophenyl (pNP)-β-d-glucopyranoside (pNP-β-d-Glc) but also shows β-d-fucosidase and β-d-galactosidase activities. Td2F2 has high glucose tolerance and its activity is enhanced by various monosaccharides, including glucose, galactose, fucose, and xylitol. The high glucose tolerance and enhancement of activity induced by monosaccharides are related to the high transglycosylation activity of Td2F2.

    In this study, we determined the crystal structure of Td2F2 to clarify the molecular mechanism of the high glucose tolerance and transglycosylation activity. In addition, we isolated glucose-sensitive Td2F2 mutants by random and saturation mutagenesis. The structure of positive subsites was important for the glucose tolerance, transglycosylation activity, and substrate specificity of Td2F2.

    Results and discussion

    Crystal structures of wild-type Td2F2

    The crystal structure of Td2F2 was determined by molecular replacement using the structure of Thermus thermophilus β-glucosidase (PDB ID 1UG6, amino acid sequence identity = 53%) as the search model. Wild-type Td2F2 structures in ligand-free form, and in complex form with d-glucose, isofagomine (a potent glucosidase inhibitor), and d-fucose were determined at resolutions up to 1.10 Å (Table 1). The overall structure of Td2F2 was similar to known GH1 enzymes, which have typical (β/α)8 barrel folds. The active site was located at the bottom of a deep pocket in the shape of a coin slot, characteristic of most GH1 enzymes. The asymmetric unit contained one molecule, which corresponded to the monomer structure of this enzyme in solution [12].

    Table 1. Data collection and refinement statistics
    Ligand-free Glucose Isofagomine Fucose Asn223Gly Asn223Gln + glucose
    Data collection
    PDB entry 3WH5 3WH6 3WH8 3WH7 5AYB 5AYI
    Beamline NW12A NW12A BL5A BL5A NW12A NW12A
    Space group P212121 P212121 P212121 P212121 P212121 P212121
    Unit cell (Å)

    a = 68.8

    b = 69.6

    c = 96.1

    a = 68.1

    b = 68.5

    c = 95.9

    a = 68.4

    b = 69.2

    c = 96.0

    a = 69.1

    b = 69.7

    c = 96.2

    a = 68.9

    b = 69.8

    c = 96.1

    a = 69.1

    b = 69.8

    c = 96.3

    Resolution (Å)a 50.00–1.60 (1.66–1.60) 50.00–1.60 (1.63–1.60) 50.00–1.90 (1.93–1.90) 50.00–1.10 (1.12–1.10) 50.00–1.80 (1.83–1.80) 50.00–1.85 (1.88–1.85)
    Total reflections 441 100 411 504 258 389 1 290 836 318 219 197 226
    Unique reflections 61 578 59 793 36 608 185 141 43 580 40 537
    Completeness (%)a 99.8 (99.5) 99.5 (98.8) 99.9 (100.0) 98.4 (98.6) 100.0 (100.0) 100.0 (100.0)
    Rmerge (%)a 11.3 (49.8) 6.1 (45.1) 13.3 (47.8) 6.1 (34.5) 11.0 (70.2) 13.1 (45.4)
    Mean I/σ (I)a 22.0 (3.2) 33.0 (3.0) 16.4 (3.3) 34.0 (4.7) 20.0 (2.9) 13.4 (3.3)
    Redundancya 7.2 (7.0) 6.9 (6.5) 7.1 (7.4) 7.0 (6.5) 7.3 (7.2) 4.9 (4.9)
    Refinement
    Resolution (Å) 26.08–1.60 32.23–1.60 29.04–1.90 24.05–1.10 50.00–1.80 56.62–1.85
    No. of reflections 58 392 56 664 34 721 175 615 41 327 38 446
    R/Rfree (%) 13.7/16.7 15.7/19.0 13.2/16.9 11.3/13.4 13.8/16.4 14.0/17.5
    No. of protein atoms 3458 3475 3436 3475 3459 3475
    No. of solvents 403 (water), 1 (CHES), 1 (Na), 1 (glycerol) 255 (water), 1 (CHES), 1 (Na), 3 (β-d-glucose), 1 (α-d-glucose)b 346 (water), 1 (CHES), 1 (Na), 3 (glycerol), 1 (isofagomine) 649 (water), 1 (CHES), 1 (Na), 1 (β-d-fucose) 283 (water), 1 (CHES), 1 (Na), 2 (ethyleneglycol) 378 (water), 1 (CHES), 1 (Na), 2 (glycerol), 1 (β-d-glucose)
    RMSD from ideal values
    Bond lengths (Å) 0.025 0.023 0.021 0.024 0.018 0.020
    Bond angles (°) 2.292 2.276 1.927 2.226 1.671 1.864
    Ramachandran plot (%)
    Favored 97.1 96.6 96.8 96.6 97.0 97.5
    Allowed 2.7 3.4 3.2 3.4 3.0 2.5
    Outlier 0.2 0.0 0.0 0.0 0.0 0.0
    • a Values in parentheses correspond to the highest resolution shell.
    • b One β-d-glucose was bound to the active site, and others (two β-d-glucose and one α-d-glucose) were bound to nonspecific sites on the surface of the protein.

    Structure at subsite −1

    In ligand-free form, a glycerol molecule, which was used as a cryoprotectant, was bound at subsite −1, which is frequently observed in GHs (Fig. 1A). All three ligands used in this study (d-glucose, isofagomine, and d-fucose) were also bound at subsite −1 (Fig. 1B–D). Superposition of the four structures showed no significant differences in the protein structure, but interactions with the ligands differ due to the differences in the chemical structures of the ligand sugars. The glucose molecule was in a β-anomer, hydrogen bonded with the catalytic base residue Glu166 (Fig. 1B). The N1 atom of isofagomine forms tight electrostatic interactions with the catalytic residues (Glu166 and Glu352) as observed in Thermotoga maritima β-glucosidase (TmBGL) [20] (Fig. 1C). The O1 hydroxyl of d-fucose was not clearly observed despite the very high resolution (Fig. 1D). The O1 atom is located in an intermediate position between the α- and β-anomers, and appears to form hydrogen bonds with both Glu166 and the nucleophile residue Glu352. The O2 hydroxyl is hydrogen bonded with His121, Asn165, and Glu352 in the glucose and fucose complex, whereas isofagomine lacks O2. The O3 hydroxyl is hydrogen bonded with Gln20, His121, and Trp407 in all complex structures. In the complexes with glucose and isofagomine, the equatorial O4 hydroxyl is hydrogen bonded with Gln20 and Glu406. However, in the fucose complex structure, the axial O4 forms hydrogen bonds with Glu406, Trp407, and a water molecule. In the glucose and isofagomine complexes, the O6 hydroxyl forms hydrogen bonds with Arg325 and Glu406, whereas a water molecule substitutes the O6 atom in the fucose (6-deoxy sugar) complex. As a stacking platform, Trp399 is located beneath the sugar ring in all complex structures.

    Details are in the caption following the image
    Structure of the Td2F2 active site. (A) Ligand-free, (B) d-glucose complex, (C) isofagomine complex, and (D) d-fucose complex. Ligands, catalytic residues, and residues forming hydrogen bonds and stacking interactions are shown as sticks in green, magenta, yellow, and orange, respectively. The |Fo| – |Fc| omit electron density maps (4.0 σ) of the ligands are shown as blue mesh.

    Figure 2 shows the Td2F2 active site structure compared to two of the best-characterized GH1 β-glycosidases from a thermophilic bacterium (TmBGL) and a thermophilic archaeon, Sulfolobus solfataricus β-glycosidase (SsBGL) [21]. Both SsBGL and TmBGL not only have β-glucosidase activity but also β-galactosidase and β-fucosidase activities [2, 22-24]). Almost all of the residues that recognize the sugar at subsite −1 except for Arg325 are conserved in these enzymes (Fig. 2A). Arg325 in Td2F2 (green) forms an electrostatic interaction with the neighboring residue Asp324. In SsBGL (orange) and TmBGL (cyan), Arg325 is replaced with Trp, Asp324 is replaced with a hydrophobic residue (Phe in SsBGL and Met in TmBGL), and one Gly residue is inserted between them (Fig. 2B). Because Arg325 is involved in Td2F2 recognition of O6, the structural difference at this region seemed to be responsible for the substrate specificity of GH1 enzymes.

    Details are in the caption following the image
    Structural comparison of GH1 β-glycosidases at subsite −1 (A) and the C6 hydroxyl (B). Td2F2 complexed with glucose (green), TmBGL (orange), and SsBGL (cyan) are superimposed.

    Mutational analysis at subsite −1

    To examine the importance of the structure at the −1 subsite, we constructed three Td2F2 mutants. The DGR mutant was created by inserting a Gly residue between Asp324 and Arg325, and the FGW and MGW mutants were created by mimicking the sequences of SsBGL and TmBGL, respectively. Td2F2 showed broad substrate specificity toward pNP-β-d-Glc, pNP-β-d-galactopyranoside (pNP-β-d-Gal), and pNP-β-d-fucopyranoside (pNP-β-d-Fuc) [12]. DGR and FGW mutants exhibited lower specific activities and kcat values for pNP-β-d-Glc, pNP-β-d-Gal, and pNP-β-d-Fuc compared to the wild-type enzyme. However, the MGW mutant exhibited about 2.5-fold and 1.4-fold higher specific activities toward pNP-β-d-Glc and pNP-β-d-Fuc, respectively (Fig. 3A,B, Table 2). Although specific activities toward pNP-β-d-Glc, pNP-β-d-Gal, and pNP-β-d-Fuc were reduced in the DGR and FGW mutants (Fig. 3A,B), Km values of these mutants for pNP-β-d-Glc, pNP-β-d-Gal, and pNP-β-d-Fuc were much lower than that of the wild-type enzyme (Table 2). The kcat/Km values of wild-type Td2F2 for pNP-β-d-Glc, pNP-β-d-Gal, and pNP-β-d-Fuc were 32.3, 8.52, and 115 mm−1s−1, respectively. Although the substrate specificities of the wild-type Td2F2 resembles that of SsBGL (kcat/Km values for pNP-β-d-Glc, pNP-β-d-Gal, and pNP-β-d-Fuc were 29.8, 12.7, and 42.6 mm−1·s−1, respectively [25]), the kcat/Km value of the wild-type Td2F2 for pNP-β-d-Fuc was more than double that of SsBGL. In the MGW mutant, the specific activity for pNP-β-d-Fuc was further increased (Fig. 3B and Table 2). With regard to glucose tolerance in the presence of 500 mm glucose, both Km and kcat values of the wild-type and mutated enzymes for pNP-β-d-Glc were increased. In the presence of 500 mm glucose, the specific β-glucosidase activities of DGR, FGW, and MGW mutants for 5 mm pNP-β-d-Glc were approximately 5–6 times higher than in the absence of glucose, similar to the activity levels of the wild-type enzyme (Fig. 3C). These results indicate that the unique structure of the Td2F2 −1 subsite is important for substrate specificity and affinity for substrates, but not for glucose tolerance.

    Details are in the caption following the image
    Glucose tolerance and substrate specificity of −1 subsite mutants. (A) β-Glucosidase activities for pNP-β-d-Glc with or without 500 mm glucose. (B) Specific activities of −1 subsite mutants for pNP-β-d-Gal and pNP-β-d-Fuc. (C) Activation ratio in the presence of 500 mm glucose. The β-glucosidase activities of each enzyme in the absence of glucose were valued at 1. Error bars, SD (n = 3).
    Table 2. Steady-state kinetic constants of the subsite −1 and Asn223 mutants
    Substrates and additive Km (mm) kcat (s−1) kcat/Km (mm−1·s−1)
    Wt pNP-β-d-Glc 0.21 ± 0.02 6.94 ± 0.16 32.3
    pNP-β-d-Glc + 500 mm Glc 2.65 ± 0.23 69.0 ± 1.9 26.1
    pNP-β-d-Gal 3.61 ± 0.50 30.8 ± 1.5 8.52
    pNP-β-d-Fuc 0.14 ± 0.03 16.0 ± 0.6 115
    DGR pNP-β-d-Glc 0.037 ± 0.004 1.82 ± 0.04 48.9
    pNP-β-d-Glc + 500 mm Glc 0.95 ± 0.06 17.5 ± 0.3 18.4
    pNP-β-d-Gal 2.20 ± 0.29 7.62 ± 0.29 3.46
    pNP-β-d-Fuc 0.080 ± 0.015 7.44 ± 0.42 92.8
    FGW pNP-β-d-Glc 0.089 ± 0.018 1.16 ± 0.08 13.0
    pNP-β-d-Glc + 500 mm Glc 1.12 ± 0.10 17.9 ± 0.4 16.0
    pNP-β-d-Gal 1.39 ± 0.13 3.11 ± 0.07 2.23
    pNP-β-d-Fuc 0.034 ± 0.007 3.74 ± 0.16 109
    MGW pNP-β-d-Glc 0.092 ± 0.007 5.87 ± 0.16 63.7
    pNP-β-d-Glc + 500 mm Glc 3.50 ± 0.41 124 ± 5 35.5
    pNP-β-d-Gal 0.94 ± 0.15 9.35 ± 0.32 9.96
    pNP-β-d-Fuc 0.13 ± 0.02 19.0 ± 1.3 151
    Asn223Tyr pNP-β-d-Glc 0.85 ± 0.07 36.7 ± 1.1 43.2
    pNP-β-d-Glc + 500 mm Glc 3.99 ± 0.26 145 ± 3.5 36.4
    pNP-β-d-Gal 0.71 ± 0.06 17.6 ± 0.3 24.7
    pNP-β-d-Fuc 0.39 ± 0.04 38.7 ± 1.0 98.3
    Asn223Gln pNP-β-d-Glc 7.69 ± 0.84 42.5 ± 1.3 5.53
    pNP-β-d-Glc + 500 mm Glc 34.3 ± 5.8 82.8 ± 6.4 2.41
    pNP-β-d-Gal 17.6 ± 1.5 32.3 ± 1.1 1.83
    pNP-β-d-Fuc 5.00 ± 0.70 80.4 ± 5.2 16.1

    Screening for glucose-sensitive Td2F2 mutants

    To elucidate the mechanism involved in Td2F2 glucose tolerance, glucose-sensitive Td2F2 mutants lacking enhanced β-glucosidase activity in the presence of glucose were screened from a random mutagenesis library (see Materials and methods section). Screening involved comparing the extent of pNP release in the presence and absence of 500 mm glucose. Note that pNP release at the high glucose concentration reflects both hydrolysis and transglycosylation (discussed below). Among approximately 200 mutants, only one exhibited a reduced activity ratio (with/without 500 mm glucose) compared to the wild-type enzyme. In this mutant, an asparagine residue (Asn223) was substituted to an aspartic acid residue. Considering the shape of the pocket and the distance of the pocket from subsite −1, the asparagine residue at site 223 was located between subsites +1 and +2 (Fig. 4A). The Asn223Asp mutant enzyme was purified using an Ni-column and then characterized. In the wild-type enzyme, the hydrolysis rate of pNP-β-d-Glc was approximately 5.5 times higher in the presence of 500 mm glucose than in the absence of glucose (Fig. 5A,B). In contrast, the hydrolysis activity of the Asn223Asp mutant for pNP-β-d-Glc was only about double in the presence of 500 mm glucose than in the absence of glucose (Fig. 5A,B). In addition, in the Asn223Asp mutant, β-galactosidase and β-fucosidase activities were 15% and 66% of wild-type enzyme activity, respectively (Fig. 5C). These results indicate that Asn223 is important not only for Td2F2 glucose tolerance but also substrate specificity.

    Details are in the caption following the image
    Structures of Td2F2 and its Asn223 mutants at positive subsites. Molecular surfaces of the wild-type Td2F2 complexed with glucose (A, green), Asn223Gly (B, cyan), and Asn223Gln complexed with glucose (C, magenta), and superimposition of the three structures (D) are shown.
    Details are in the caption following the image
    Glucose tolerance and substrate specificity of Asn223 variants. (A) Hydrolysis activities for pNP-β-d-Glc with or without 500 mm glucose. (B) Activation ratio in the presence of 500 mm glucose. β-glucosidase activities of each enzyme in the absence of glucose were valued at 1. (C) Specific activities of Asn223 variants for pNP-β-d-Gal and pNP-β-d-Fuc. Error bars, SD (n = 3).

    Saturation mutagenesis of the 223rd asparagine residue

    The importance of the structure at the Asn223 site was examined by saturation mutagenesis. Asn223 residue was replaced with all other amino acid residues by site-directed mutagenesis. Asn223 variants were expressed in Escherichia coli cells and purified in an Ni-column. The β-glucosidase activities of the purified Asn223 variants were measured in the presence or absence of 500 mm glucose (Fig. 5A,B). β-glucosidase activities toward pNP-β-d-Glc of the Asn223Thr and Asn223Arg mutants was drastically reduced. Interestingly, β-glucosidase activities toward pNP-β-d-Glc increased by approximately 2.5-, 1.8-, 1.5-, 3.2-, 2.5-, 5.4-, and 3.0-fold in the Asn223Leu, Asn223Ile, Asn223Met, Asn223Phe, Asn223Trp, Asn223Tyr, and Asn223Gln mutants, respectively (Fig. 5A). This indicates that the presence of a hydrophobic residue increases activity toward the pNP substrate, possibly by increasing the affinity between subsite +1 and the pNP group. Aside from the Asn223Gly, Asn223Lys, and Asn223Gln mutants, most mutants had higher β-glucosidase activities in the presence of 500 mm glucose than in its absence (Fig. 5B). However, the glucose activation ratios for all Asn223 variants were lower than that of the wild-type enzyme. Both the increased (wild-type and Asn223Asp) and decreased (Asn223Gln and Asn223Gly) activities were modulated by glucose concentration (Fig. 6A). The Asn223His mutant exhibited comparable activities in the presence and absence of 500 mm glucose (Fig. 5B), but the fact that a dependence on glucose concentration remained evident indicated that the result reflected a mixture of both slight activation and mild inhibition (Fig. 6A).

    Details are in the caption following the image
    Effects of monosaccharides and sugar alcohol. (A) The β-glucosidase activities of Asn223 mutants for pNP-β-d-Glc were measured in the presence of 25–750 mm glucose. The activities of each enzyme for pNP-β-d-Glc without glucose were valued at 1. Wild-type, Asn223Asp, Asn223Gly, Asn223His, and Asn223Gln are shown. (B) The β-glucosidase activities of Asn223 mutants for pNP-β-d-Glc in the presence of 100 mm d-galactose, d-fucose, and xylitol. Hydrolysis activities toward pNP-β-d-Glc without additives were valued at 1. Error bars, SD (n = 3).

    The hydrolysis activity of Td2F2 is also activated by various monosaccharides and sugar alcohols, including d-galactose, d-fucose, and xylitol [12]. In this study, in the wild-type enzyme, β-glucosidase activity for pNP-β-d-Glc was activated by 100 mm d-galactose, d-fucose, and xylitol by a 2.9-, 2.6-, and 1.6-fold increase, respectively (Fig. 6B). In contrast, the activation of β-glucosidase activity by d-galactose, d-fucose, and xylitol was dramatically reduced in the Asn223Asp mutant, and was almost completely inhibited in the Asn223Gly and Asn223Gln mutants. These results suggest that the structure between subsites +1 and +2 relates to Td2F2 activation by not only glucose but also other monosaccharides and sugar alcohols.

    Substrate specificity of Asn223 variants

    Except for Asn223Tyr, the β-galactosidase activities of Asn223 variants decreased, particularly for the Asn223Asp, Asn223Gly, Asn223Ala, Asn223Pro, Asn223Ser, Asn223Thr, Asn223Lys, and Asn223Arg mutants (Fig. 5C). The β-fucosidase activities of Asn223Leu, Asn223Phe, Asn223Trp, Asn223His, Asn223Tyr, and Asn223Gln were approximately 1.5, 1.9, 1.5, 1.6, 2.8, and 3.3 times that of the wild-type Td2F2. In contrast, the activities decreased in the following mutants: Asn223Asp, Asn223Gly, Asn223Ala, Asn223Val, Asn223Met, Asn223Pro, Asn223Ser, Asn223Thr, Asn223Lys, Asn223Arg, and Asn223Cys. Next, we investigated the specificity of the Asn223Asp, Asn223Gly, Asn223Tyr, and Asn223Gln mutants for disaccharides, including cellobiose, laminaribiose, sophorose, gentiobiose, and lactose (Table 3). β-Glucosidase activities toward cellobiose, laminaribiose, and sophorose decreased in these Asn223 variants. Although the Asn223Tyr mutant exhibited approximately a 5.4 times higher activity toward pNP-β-d-Glc than the wild-type enzyme, the hydrolysis activities of this mutant for cellobiose, laminaribiose, and sophorose were only 1.7%, 11.3%, and 5.7% of the wild-type enzyme activity, respectively (Table 3). Asn223Gly and Asn223Gln mutants, which were glucose-sensitive mutants, had higher β-glucosidase activities for sophorose than Asn223Asp and Asn223Tyr mutants. Although the wild-type Td2F2 had marginal β-glucosidase activity for gentiobiose (0.42 μmol·min−1·mg protein−1), surprisingly, the Asn223Tyr and Asn223Gln mutants had remarkably higher β-glucosidase activity for gentiobiose (1.33 and 2.08 μmol·min−1·mg protein−1, respectively).

    Table 3. Disaccharides degradation using Td2F2 mutants
    Activity (μmol·min−1·mg−1)
    Wt Asn223Asp Asn223Gly Asn223Tyr Asn223Gln
    Cellobiose 4.62 ± 0.43 0.54 ± 0.02 0.62 ± 0.08 0.08 ± 0.01 0.50 ± 0.07
    Laminaribiose 6.13 ± 0.78 0.55 ± 0.05 0.95 ± 0.09 0.69 ± 0.03 0.84 ± 0.10
    Sophorose 10.6 ± 0.4 1.32 ± 0.12 6.16 ± 0.40 0.60 ± 0.07 6.48 ± 0.98
    Gentiobiose 0.42 ± 0.04 0.03 ± 0.01 0.13 ± 0.02 1.33 ± 0.18 2.08 ± 0.25
    Lactose 5.73 ± 0.63 0.49 ± 0.03 1.27 ± 0.11 0.05 ± 0.01 0.65 ± 0.05

    Kinetic constants of Asn223 variants

    As described above, Asn223 mutants showed various changes in glucose tolerance and substrate specificity. For example, the specific activity for pNP-β-d-Glc was dramatically enhanced by substitution of asparagine to a tyrosine or glutamine residue (Fig. 5). Therefore, we examined the kinetic parameters of Asn223Tyr and Asn223Gln against various substrates. The Km and kcat values of Asn223Tyr for pNP-β-d-Glc were approximately 4.0-fold and 5.3-fold higher than that of the wild-type enzyme, respectively (Table 2). Similarly, the Km and kcat values for pNP-β-d-Fuc increased in the Asn223Tyr mutant (Table 2). In contrast, the Km and kcat values of Asn223Tyr mutant for pNP-β-d-Gal were lower than that of the wild-type enzyme by factors of 0.2 and 0.6, respectively (Table 2). Although the kcat value of the Asn223Gln mutant for pNP-β-d-Glc was approximately 6.1-fold greater than that of the wild-type enzyme, the Km value was dramatically increased in the Asn223Gln mutant. In Asn223Gln mutants, Km values for pNP-β-d-Gal and pNP-β-d-Fuc also increased, and as a result, kcat/Km values decreased.

    Similar to the wild-type enzyme, in Asn223Tyr and Asn223Gln mutants, Km and kcat values increased with the addition of glucose. Specifically, the Km value for pNP-β-d-Glc was dramatically increased in the Asn223Gln mutant in the presence of 500 mm glucose (Table 2). These results indicate that the glucose sensitivity of the Asn223Gln mutant was caused by the decrease in affinity for pNP-β-d-Glc in the presence of glucose.

    Transglycosylation activity of Td2F2 mutants

    Td2F2 has strong transglycosylation activity to produce disaccharides, including laminaribiose and sophorose, in the presence of high glucose concentrations [12]. In the presence of 125 mm glucose and the substrate pNP-d-Glc, the wild-type Td2F2 enzyme generated laminaribiose and sophorose (Table 4). In contrast, sophorose production was dramatically decreased and laminaribiose production was completely suppressed in the Asn223Asp, Asn223Gln, Asn223Gly, and Asn223His mutants. Interestingly, the Asn223Leu and Asn223Tyr mutants, which exhibited higher activity toward pNP-d-Glc than did the wild-type enzyme (Fig. 5A), generated significant amounts of gentiobiose, while they generated lower amounts of sophorose and laminaribiose. When the glucose concentration was elevated to 500 mm, the extent of sophorose production by the wild-type; and the Asn223Asp, Asn223Gln, Asn223His, Asn223Leu, and Asn223Tyr mutants, was two- to fourfold greater than that in the presence of 125 mm glucose. However, the sophorose production rates (sophorose/pNP ratios) differed greatly between the wild-type and mutant enzymes. The sophorose production rates of the wild-type enzyme in the presence of 125 and 500 mm glucose were 0.40 (9.04/22.4) and 0.49 (20.2/41.1), respectively. In contrast, the production rates of the Asn223 mutants were <0.14 and 0.21, respectively. Our results indicate that site 223 is vital for transglycosylation activity that results in sophorose production, and that this activity is strongly related to glucose tolerance/activation and substrate specificity.

    Table 4. Transglycosylation activity of the Td2F2 Asn223 variants
    Glucose concentration (mm)
    0 125 500
    Production activity (μmol·min−1·mg−1)
    Wt pNPa 11.2 ± 0.2 22.4 ± 0.6 41.1 ± 1.4
    Sophoroseb n.d.c 9.04 ± 1.37 20.2 ± 1.7
    Laminaribioseb n.d. 1.39 ± 0.33 4.11 ± 0.78
    Gentiobioseb n.d. n.d. n.d.
    Asn223Asp pNP 3.66 ± 0.12 5.79 ± 0.16 9.44 ± 0.71
    Sophorose n.d. 0.79 ± 0.05 1.97 ± 0.29
    Laminaribiose n.d. n.d. n.d.
    Gentiobiose n.d. n.d. n.d.
    Asn223Gln pNP 21.1 ± 1.6 18.4 ± 0.3 15.1 ± 0.5
    Sophorose n.d. 0.55 ± 0.05 1.33 ± 0.20
    Laminaribiose n.d. n.d. n.d.
    Gentiobiose n.d. n.d. n.d.
    Asn223Gly pNP 2.97 ± 0.13 2.77 ± 0.07 2.33 ± 0.09
    Sophorose n.d. 0.22 ± 0.04 n.d.
    Laminaribiose n.d. n.d. n.d.
    Gentiobiose n.d. n.d. n.d.
    Asn223His pNP 9.97 ± 0.13 11.7 ± 0.6 13.0 ± 0.3
    Sophorose n.d. 0.42 ± 0.03 1.25 ± 0.15
    Laminaribiose n.d. n.d. n.d.
    Gentiobiose n.d. n.d. n.d.
    Asn223Leu pNP 31.2 ± 0.5 48.2 ± 1.4 68.8 ± 3.6
    Sophorose n.d. 0.93 ± 0.11 3.84 ± 0.77
    Laminaribiose n.d. n.d. n.d.
    Gentiobiose n.d. 14.8 ± 1.7 45.0 ± 4.1
    Asn223Tyr pNP 45.1 ± 1.0 62.8 ± 1.7 81.1 ± 1.7
    Sophorose n.d. 1.29 ± 0.27 4.23 ± 0.45
    Laminaribiose n.d. 1.02 ± 0.39 3.06 ± 0.18
    Gentiobiose n.d. 11.8 ± 0.8 27.1 ± 2.6
    • a The concentration of pNP was determined by measuring the solution absorbance at 405 nm.
    • b The concentration of sophorose, laminaribiose, and gentiobiose was determined by high-performance liquid chromatography.
    • c Not detected.

    Structures of the Asn223 mutants

    We determined the crystal structures of the two mutants with the highest glucose sensitivity, Asn223Gly and Asn223Gln. The structures of these mutants were determined at 1.80 Å and 1.85 Å resolutions, respectively (Table 1). The Asn223Gln mutant structure was obtained as a glucose complex using a soaking method at a concentration of 10 mm. A β-d-glucose molecule was bound at subsite −1, similar to the wild-type glucose complex structure. The molecular surfaces of these mutants were compared to that of the wild-type enzyme (Fig. 4A–C). As shown in Fig. 4A, the side chain of Asn223 in the wild-type enzyme forms a platform that is likely suitable for binding a sugar at subsite +1. In Asn223Gly, there is a big hole between subsites +1 and +2 because of the lack of a side chain at this residue (Fig. 4B). In Asn223Gln, there is a wall-blocking subsite +1 (Fig. 4C). When the wild-type and mutant structures are superimposed at this site, the side chain of Gln223 is displaced and has a different conformation (Fig. 4D). Therefore, subsite +1 of the mutants does not appear to have a suitable surface for binding the sugar acceptor molecule necessary for the transglycosylation reaction, and thus we conclude that this site is crucial for Td2F2 glucose tolerance.

    In conclusion, the Asn223 residue of Td2F2 is critical for the structure of the positive subsites, and is closely related to glucose tolerance, transglycosylation activity, and substrate specificity. An asparagine residue corresponding to Asn223 of Td2F2 is highly conserved in other GH1 β-glucosidases (Fig. 1 of [12]). In previous studies on a Thermus thermophilus β-glycosidase of the GH1 family, it was found that the transglycosylation/hydrolysis ratio was increased by mutations around subsite −1 [26, 27]. In addition, the structure of the entrance to the active site, including subsite +2, plays a key role in the glucose tolerance of fungal GH1 β-glucosidases [28-30]. Our findings also suggest that not only subsite −1 but also positive subsites are essential for the glucose tolerance and transglycosylation activity of GH1 glycosidases. The transglycosylation reactions of retaining GHs require appropriate binding of acceptor molecules after the formation of covalent enzyme-glycosyl intermediates (Fig. 7). Subsite +1, supported by Asn223, provides a structural platform for the acceptor glucose involved in transglycosylation. We assume that the high-level glucose tolerance of Td2F2 is related to the high transglycosylation activity of the enzyme [12]. The aromatic side chains of Trp168, Phe243, and Phe246 form a partially hydrophobic pocket located near subsite +2 (Fig. 4A); this may provide an additional binding site for glucose, even when the concentrations of both substrate and glucose are high. In addition, certain Asn223 mutants with hydrophobic side chains did not lose glucose tolerance (e.g., Asn223Tyr and Asn223Leu; Fig. 5), suggesting that the presence of a hydrophobic pocket at this site could support glucose tolerance and transglycosylation to produce gentiobiose (Table 4). Our findings contribute to a better understanding of the mechanisms of glucose tolerance, transglycosylation activity, and substrate specificity, and could aid protein engineering of GH1 β-glucosidases.

    Details are in the caption following the image
    The possible transglycosylation reaction mechanism of Td2F2 at high concentrations of pNP-β-d-Glc (the donor) and glucose (the acceptor). A reaction forming a sophorose (Glc-β1,2-Glc) is shown.

    Materials and methods

    Materials

    Cellobiose, pNP-β-d-Gal, and d-fucose were purchased from Sigma-Aldrich (St. Louis, MO, USA). pNP-β-d-Glc, pNP-β-d-Fuc, and gentiobiose were purchased from Tokyo Chemical Industry Co. (Tokyo, Japan). Xylitol, lactose monohydrate, and glucose were purchased from Wako Pure Chemical Industries (Osaka, Japan). d-Galactose was purchased from Nacalai Tesque (Kyoto, Japan). Laminaribiose and sophorose were purchased from Seikagaku Biobusiness (Tokyo, Japan) and Extrasynthese (Lyon, France), respectively.

    Protein expression and purification

    N-terminally His6-tagged Td2F2 was expressed and purified as described previously [12], with some modifications. For crystallographic analysis, His6-tagged Td2F2 and its mutants were purified using an Ni-column (HisTrap HP; GE Healthcare, Buckinghamshire, England) and a gel-filtration column (HiLoad 16/60 Superdex 200 prep-grade; GE Healthcare).

    Crystallography

    Crystals from the wild-type Td2F2 and its Asn223 mutants were obtained using the sitting-drop vapor diffusion method at 20 °C. We mixed 1.0 μL protein solution (7.6 mg·mL−1) with an equal volume of a reservoir solution containing 1.0–1.2 m K/Na tartrate, 0.1 m CHES (pH 8.5), and 0.2 m Li2SO4. As a cryoprotectant, ethylene glycol was added to 20% (v/v) for the Asn223Gln crystals, and glycerol to 20–30% (v/v) for all other crystals. Wild-type Td2F2 crystals complexed with glucose were obtained by cocrystallization in the presence of 500 mm glucose. Wild-type Td2F2 crystals complexed with d-fucose and isofagomine, and Asn223Gln crystals complexed with glucose were obtained using a soaking method; the cryoprotectant solution contained 20–30% (v/v) glycerol and 10 mm of each ligand. Crystals were flash-cooled in a nitrogen stream at 100 K. Diffraction data were collected at the Photon Factory of the High Energy Accelerator Research Organization (KEK; Tsukuba, Japan) at a wavelength of 1.000 Å. Diffraction data were processed using hkl2000 software [31]. Molecular replacement was performed with the aid of molrep [32]. Manual model building and refinement were conducted using coot [33] and refmac5 [34], respectively. Molecular graphical images were prepared using pymol (Schrödinger, LLC, New York, NY, USA).

    Random mutagenesis of Td2F2

    Random mutagenesis employed error-prone PCR using a GeneMorph II Random Mutagenesis Kit (Agilent Technologies, Santa Clara, CA, USA) and the ‘megaprimer PCR of whole plasmid’ (MEGAWHOP) method [35]. Escherichia coli BL21-CodonPlus(DE3)-RP (Agilent Technologies) cells harboring the Td2F2 random mutagenesis library were selected from Luria–Bertani broth (LB) agar plates containing 100 μg·mL−1 ampicillin and 30 μg·mL−1 chloramphenicol (LB + Amp + Cm). Selected cultures were inoculated in 96-well plates with 900 μL Overnight Express Instant LB medium (OELB; Novagen, Madison, WI, USA) containing 100 μg·mL−1 ampicillin and 30 μg·mL−1 chloramphenicol (OELB + Amp + Cm). After overnight incubation at 37 °C, cells were collected by centrifugation (2465 g, 5 min) and resuspended in 100 μL BugBuster (Novagen). Cell suspensions were incubated at room temperature for 40 min with agitation. After adding 1 mL 50 mm sodium acetate buffer (pH 5.5), soluble fractions (2465 g, 15 min) of the cell lysates were collected (crude extract). The activities of crude extracts of β-glucosidase were measured with or without 500 mm glucose as mentioned below. Each 20 μL reaction mixture contained 3 μL of crude enzyme solution, 50 mm sodium acetate (pH 5.5), 5 mm pNP-β-d-Glc, and 0 or 500 mm glucose. The reaction was performed at 70 °C for 5 min. To stop the reaction, 50 μL 1 m NaHCO3 was added. The concentration of released pNP was determined by the absorbance at 405 nm. A mutant with a lower glucose activation ratio (β-glucosidase activity in the presence of 500 mm glucose/in the absence of glucose) compared to the wild-type Td2F2 enzyme was selected, and the mutation site was identified by sequencing.

    Introduction of mutations into Td2F2

    The primers used for both site-directed mutagenesis of the −1 subsite and saturation mutagenesis of the asparagine residue at site 223 are listed in Table S1 in the Supporting information. We amplified 21 PCR fragments using the following primer combinations: Nos. 1 and 2 (DGW mutant), Nos. 3 and 4 (FGW mutant), Nos. 5 and 6 (MGW), and Nos. 7 and 8–25 (saturation mutagenesis of the 223rd asparagine residue). The Td2F2 expression vector served as a template and each PCR fragment was circularized using an In-Fusion PCR cloning kit (Takara Bio Inc., Shiga, Japan).

    β-Glucosidase, β-fucosidase, and β-galactosidase activity assays using Td2F2 mutants

    Activity assays for pNP substrates were performed at 75 °C for 5 min. The total reaction volume was 20 μL. The reaction mixture contained 100 mm sodium acetate (pH 5.5), 5 mm of a pNP substrate (pNP-β-d-Glc, pNP-β-d-Fuc, or pNP-β-d-Gal), 0 or 500 mm glucose, and 0.05 μg of a purified enzyme. To stop the reaction, 50 μL 1 m NaHCO3 was added and the concentration of pNP was determined by measuring the absorbance at 405 nm. Activity assays for disaccharides were performed at 75 °C for 10 min. The 20 μL reaction mixture contained 100 mm sodium acetate (pH 5.5), 10 mm disaccharides (cellobiose, laminaribiose, sophorose, gentiobiose, and lactose), and 0.4 μg purified enzymes. To inactivate the enzymes, reaction mixtures were incubated at 98 °C for 10 min. The product, glucose, was measured using a LabAssay Glucose Kit (Wako Pure Chemical Industries).

    Kinetic analysis

    The kinetic parameters of Td2F2 and variants thereof were determined using pNP-β-d-Glc, pNP-β-d-Gal, and pNP-β-d-Fuc upon incubation at 75 °C for 5 min in 100 mm sodium acetate buffer (pH 5.5), as described below. For wild-type Td2F2, the −1 subsite mutants (DGR, FGW, and MGW), and the Asn223Tyr mutant, the kinetic parameters were determined using 0.05 μg enzyme/20 μL, and pNP-β-d-Glc, pNP-β-d-Gal, and pNP-β-d-Fuc concentrations of 0.125–6 mm, 0.3125–20 mm, and 0.125–6 mm, respectively. In the presence of 500 mm glucose, the kinetic parameters were determined at pNP-β-d-Glc concentrations of 0.3125–20 mm. In the case of the Asn223Gln mutant, the kinetic parameters were determined using 0.05 μg enzyme/20 μL and pNP-β-d-Glc, pNP-β-d-Gal, and pNP-β-d-Fuc concentrations of 1.5–75 mm, 1–60 mm, and 0.3–10 mm, respectively. In the presence of 500 mm glucose, the kinetic parameters were determined at pNP-β-d-Glc concentrations of 1.5–75 mm. Kinetic constants (Km and kcat) were calculated using nonlinear regression of the Michaelis–Menten equation with graphpad prism version 5.0 (GraphPad Software, La Jolla, CA, USA).

    Analysis of transglycosylation activity

    Transglycosylation products were analyzed as described previously [12], with some modifications. Each reaction mixture (200 μL) containing 1 μg of Td2F2 or a variant thereof, 100 mm sodium acetate buffer (pH 5.5), 10 mm pNP-β-d-Glc, and 0, 125 or 500 mm glucose, was incubated at 70 °C for 15 min in the case of the wild-type and Asn223Leu and Asn223Tyr mutants; or for 30 min in the case of the Asn223Asp, Asn223Gln, Asn223Gly, and Asn223His mutants. To stop the reaction, all samples were heated to 95 °C for 10 min. The reactants were pretreated at 37 °C for 16 h with glucose oxidase from Aspergillus niger (Wako Pure Chemical Industries) to degrade any excess glucose prior to analysis. Transglycosylation products were analyzed by high-performance liquid chromatography, as described previously [12].

    Acknowledgements

    We thank the staff of the Photon Factory and SPring-8 for collecting X-ray data. This study was supported by JSPS KAKENHI Grant Number 26850067.

      Conflict of interest

      The authors declare that they have no conflicts of interests.

      Author contributions

      T.M.: experimental design, performance of experiments, data analysis, and manuscript preparation. T.J.: performance of experiments. T.U.: performance of experiments, data analysis, and manuscript preparation. J.A.M.: performance of experiments. T.A.: performance of experiments. K.M.: experimental design and manuscript preparation. S.F.: experimental design, performance of experiments, data analysis, and manuscript preparation. K.Y.: experimental design, data analysis, and manuscript preparation.