Advertisement

Structural basis of the strict specificity of a bacterial GH31 α-1,3-glucosidase for nigerooligosaccharides

Open AccessPublished:March 12, 2022DOI:https://doi.org/10.1016/j.jbc.2022.101827
      Carbohydrate-active enzymes are involved in the degradation, biosynthesis, and modification of carbohydrates and vary with the diversity of carbohydrates. The glycoside hydrolase (GH) family 31 is one of the most diverse families of carbohydrate-active enzymes, containing various enzymes that act on α-glycosides. However, the function of some GH31 groups remains unknown, as their enzymatic activity is difficult to estimate due to the low amino acid sequence similarity between characterized and uncharacterized members. Here, we performed a phylogenetic analysis and discovered a protein cluster (GH31_u1) sharing low sequence similarity with the reported GH31 enzymes. Within this cluster, we showed that a GH31_u1 protein from Lactococcus lactis (LlGH31_u1) and its fungal homolog demonstrated hydrolytic activities against nigerose [α-D-Glcp-(1→3)-D-Glc]. The kcat/Km values of LlGH31_u1 against kojibiose and maltose were 13% and 2.1% of that against nigerose, indicating that LlGH31_u1 has a higher specificity to the α-1,3 linkage of nigerose than other characterized GH31 enzymes, including eukaryotic enzymes. Furthermore, the three-dimensional structures of LlGH31_u1 determined using X-ray crystallography and cryogenic electron microscopy revealed that LlGH31_u1 forms a hexamer and has a C-terminal domain comprising four α-helices, suggesting that it contributes to hexamerization. Finally, crystal structures in complex with nigerooligosaccharides and kojibiose along with mutational analysis revealed the active site residues involved in substrate recognition in this enzyme. This study reports the first structure of a bacterial GH31 α-1,3-glucosidase and provides new insight into the substrate specificity of GH31 enzymes and the physiological functions of bacterial and fungal GH31_u1 members.

      Keywords

      Abbreviations:

      CADE (GH31 cycloalternan-specific α-1,3-glucosidase), cryo-EM (cryogenic electron microscopy), CTF (contrast transfer function), CtGII (α-glucosidase II from Chaetomium thermophilum var. thermophilum), ER (endoplasmic reticulum), GHs (glycoside hydrolases), MGAM (mammalian intestinal maltase-glucoamylase), NtMGAM (N-terminal maltase domain of human maltase-glucoamylase), PDB (Protein Data Bank), pNP-α-Glc (p-nitrophenyl α-D-glucopyranoside)
      Carbohydrates are widely distributed in nature, and their structures are diverse as per their physiological functions. The biosynthesis and degradation of carbohydrates are catalyzed by carbohydrate-active enzymes, including glycoside hydrolases (GHs), glycosyltransferases, and polysaccharide lyases, and are classified into many families based on sequence similarity in the CAZy database (http://www.cazy.org/) (
      • Lombard V.
      • Golaconda Ramulu H.
      • Drula E.
      • Coutinho P.M.
      • Henrissat B.
      The carbohydrate-active enzymes database (CAZy) in 2013.
      ,
      • Drula E.
      • Garron M.L.
      • Dogan S.
      • Lombard V.
      • Henrissat B.
      • Terrapon N.
      The carbohydrate-active enzyme database: Functions and literature.
      ). GHs are involved in carbohydrate degradation and glycoside formation through transglycosylation reactions (
      • Light S.H.
      • Cahoon L.A.
      • Mahasenan K.V.
      • Lee M.
      • Boggess B.
      • Halavaty A.S.
      • Mobashery S.
      • Freitag N.E.
      • Anderson W.F.
      Transferase versus hydrolase: The role of conformational flexibility in reaction specificity.
      ,
      • Tagami T.
      • Miyano E.
      • Sadahiro J.
      • Okuyama M.
      • Iwasaki T.
      • Kimura A.
      Two novel glycoside hydrolases responsible for the catabolism of cyclobis-(1→6)-α-nigerosyl.
      ). There are >160 GH families, and recent studies have discovered new enzymes in various microorganisms (
      • Garron M.L.
      • Henrissat B.
      The continuing expansion of CAZymes and their families.
      ,
      • Helbert W.
      • Poulet L.
      • Drouillard S.
      • Mathieu S.
      • Loiodice M.
      • Couturier M.
      • Lombard V.
      • Terrapon N.
      • Turchetto J.
      • Vincentelli R.
      • Henrissat B.
      Discovery of novel carbohydrate-active enzymes through the rational exploration of the protein sequences space.
      ). However, there are many putative GHs with unknown functions among the known GH families.
      α-1,4-Glucosidases (EC 3.2.1.20) hydrolyze α-1,4-glucosidic linkages at the nonreducing end of substrates and release α-glucose via the anomer-retaining reaction mechanism. The enzymes are present in numerous organisms and primarily belong to the GH13 and GH31 families (
      • Okuyama M.
      • Saburi W.
      • Mori H.
      • Kimura A.
      α-Glucosidases and α-1,4-glucan lyases: Structures, functions, and physiological actions.
      ). GH31 is one of the most diversified GH families and harbors not only α-1,4-glucosidases but also α-1,3-glucosidases (
      • Kang M.S.
      • Okuyama M.
      • Mori H.
      • Kimura A.
      The first α-1,3-glucosidase from bacterial origin belonging to glycoside hydrolase family 31.
      ), endoplasmic reticulum (ER) α-glucosidase II (
      • Caputo A.T.
      • Alonzi D.S.
      • Marti L.
      • Reca I.B.
      • Kiappes J.L.
      • Struwe W.B.
      • Cross A.
      • Basu S.
      • Lowe E.D.
      • Darlot B.
      • Santino A.
      • Roversi P.
      • Zitzmann N.
      Structures of mammalian ER α-glucosidase II capture the binding modes of broad-spectrum iminosugar antivirals.
      ,
      • Satoh T.
      • Toshimori T.
      • Yan G.
      • Yamaguchi T.
      • Kato K.
      Structural basis for two-step glucose trimming by glucosidase II involved in ER glycoprotein quality control.
      ), α-xylosidases (
      • Lovering A.L.
      • Lee S.S.
      • Kim Y.W.
      • Withers S.G.
      • Strynadka N.C.
      Mechanistic and structural analysis of a family 31 α-glycosidase and its glycosyl-enzyme intermediate.
      ), α-galactosidases (
      • Miyazaki T.
      • Ishizaki Y.
      • Ichikawa M.
      • Nishikawa A.
      • Tonozuka T.
      Structural and biochemical characterization of novel bacterial α-galactosidases belonging to glycoside hydrolase family 31.
      ), sulfoquinovosidases (
      • Speciale G.
      • Jin Y.
      • Davies G.J.
      • Williams S.J.
      • Goddard-Borger E.D.
      YihQ is a sulfoquinovosidase that cleaves sulfoquinovosyl diacylglyceride sulfolipids.
      ), α-N-acetylgalactosaminidases (
      • Rahfeld P.
      • Wardman J.F.
      • Mehr K.
      • Huff D.
      • Morgan-Lang C.
      • Chen H.M.
      • Hallam S.J.
      • Withers S.G.
      Prospecting for microbial α-N-acetylgalactosaminidases yields a new class of GH31 O-glycanase.
      ,
      • Miyazaki T.
      • Park E.Y.
      Crystal structure of the Enterococcus faecalis α-N-acetylgalactosaminidase, a member of the glycoside hydrolase family 31.
      ,
      • Ikegaya M.
      • Miyazaki T.
      • Park E.Y.
      Biochemical characterization of Bombyx mori α-N-acetylgalactosaminidase belonging to the glycoside hydrolase family 31.
      ), and α-glucan lyases (
      • Rozeboom H.J.
      • Yu S.
      • Madrid S.
      • Kalk K.H.
      • Zhang R.
      • Dijkstra B.W.
      Crystal structure of α-1,4-glucan lyase, a unique glycoside hydrolase family member with a novel catalytic mechanism.
      ). GH31 α-glucoside hydrolases also vary in their function and substrate specificity; these include mammalian intestinal maltase-glucoamylase (MGAM) (
      • Sim L.
      • Quezada-Calvillo R.
      • Sterchi E.E.
      • Nichols B.L.
      • Rose D.R.
      Human intestinal maltase-glucoamylase: Crystal structure of the N-terminal catalytic subunit and basis of inhibition and substrate specificity.
      ,
      • Ren L.
      • Qin X.
      • Cao X.
      • Wang L.
      • Bai F.
      • Bai G.
      • Shen Y.
      Structural insight into substrate specificity of human intestinal maltase-glucoamylase.
      ), sucrase-isomaltase (
      • Gray G.M.
      • Lally B.C.
      • Conklin K.A.
      Action of intestinal sucrase-isomaltase and its free monomers on an α-limit dextrin.
      ), and lysosomal acid α-glucosidase, which catalyzes glycogen breakdown (
      • Kishnani P.S.
      • Howell R.R.
      Pompe disease in infants and children.
      ). Some GH31 enzymes can efficiently catalyze transglycosylation and produce oligosaccharides containing α-1,3- and α-1,6-glucosidic bonds (
      • Ma M.
      • Okuyama M.
      • Tagami T.
      • Kikuchi A.
      • Klahan P.
      • Kimura A.
      Novel α-1,3/α-1,4-glucosidase from Aspergillus niger exhibits unique transglucosylation to generate high levels of nigerose and kojibiose.
      ,
      • Ma M.
      • Okuyama M.
      • Sato M.
      • Tagami T.
      • Klahan P.
      • Kumagai Y.
      • Mori H.
      • Kimura A.
      Effects of mutation of Asn694 in Aspergillus Niger α-glucosidase on hydrolysis and transglucosylation.
      ). Furthermore, some bacterial GH31 enzymes catalyze unique reactions. While most known GH31 enzymes are exo-acting, a GH31 dextranase derived from Flavobacterium johnsoniae (FjDex31A) is endo-acting and releases isomaltooligosaccharides from dextran (
      • Tsutsumi K.
      • Gozu Y.
      • Nishikawa A.
      • Tonozuka T.
      Structural insights into polysaccharide recognition by Flavobacterium johnsoniae dextranase, a member of glycoside hydrolase family 31.
      • Gozu Y.
      • Ishizaki Y.
      • Hosoyama Y.
      • Miyazaki T.
      • Nishikawa A.
      • Tonozuka T.
      A glycoside hydrolase family 31 dextranase with high transglucosylation activity from Flavobacterium johnsoniae.
      ). An α-transglucosylase derived from Cellvibrio japonicus (Agd31B) transfers single glucosyl units from α-1,4-glucans to the 4-OH group of glucose (
      • Larsbrink J.
      • Izumi A.
      • Hemsworth G.R.
      • Davies G.J.
      • Brumer H.
      Structural enzymology of Cellvibrio japonicus Agd31B protein reveals α-transglucosylase activity in glycoside hydrolase family 31.
      ). Cycloalternan [cyclic α-nigerosyl-(1→6)-nigerose] is formed by two GH31 transglucosidases, namely, 6-α-glucosyltransferase and cycloalternan-forming enzyme (3-α-isomaltosyltransferase) and degraded by GH31 cycloalternan-specific α-1,3-glucosidase (CADE) (
      • Light S.H.
      • Cahoon L.A.
      • Mahasenan K.V.
      • Lee M.
      • Boggess B.
      • Halavaty A.S.
      • Mobashery S.
      • Freitag N.E.
      • Anderson W.F.
      Transferase versus hydrolase: The role of conformational flexibility in reaction specificity.
      ). Despite differences in their substrate specificities and reactions, the characterized GH31 enzymes share two aspartic acid residues as catalytic residues in the anomer-retaining catalytic mechanism (
      • Okuyama M.
      Function and structure studies of GH family 31 and 97 α-glycosidases.
      ).
      Although many GH31 enzymes have been characterized to date, there are still many proteins in microbial genomes whose activities are unclear. The diversity of the GH31 family indicates the existence of novel enzymes. Phylogenetic analysis and the subclassification of the enzyme family is a powerful strategy for discovering new enzymes from a large sequence space (
      • Okuyama M.
      Function and structure studies of GH family 31 and 97 α-glycosidases.
      ,
      • Jones D.R.
      • Thomas D.
      • Alger N.
      • Ghavidel A.
      • Inglis G.D.
      • Abbott D.W.
      Saccharis: An automated pipeline to streamline discovery of carbohydrate active enzyme activities within polyspecific families and de novo sequence datasets.
      ). Recently, some open-source software and servers for protein sequence analysis have been developed for handling a large dataset (
      • Jones D.R.
      • Thomas D.
      • Alger N.
      • Ghavidel A.
      • Inglis G.D.
      • Abbott D.W.
      Saccharis: An automated pipeline to streamline discovery of carbohydrate active enzyme activities within polyspecific families and de novo sequence datasets.
      ,
      • Xu J.
      • Zhang H.
      • Zheng J.
      • Dovoedo P.
      • Yin Y.
      eCAMI: simultaneous classification and motif identification for enzyme annotation.
      ,
      • Barrett K.
      • Lange L.
      Peptide-based functional annotation of carbohydrate-active enzymes by conserved unique peptide patterns (CUPP).
      ). In the present study, we performed a phylogenetic analysis of GH31 proteins and identified new microbial members of the GH31 family, which showed strict α-1,3-glucosidase activities on nigerooligosaccharides. Furthermore, using X-ray crystallography and cryogenic electron microscopy (cryo-EM), the three-dimensional structure of the bacterial members was determined to discuss their substrate recognition mechanism.

      Results and discussion

      GH31 uncharacterized protein cluster

      We performed a phylogenetic analysis using protein sequences classified as the GH31 family on the CAZy database. Protein sequences sharing <70% identity extracted from all GH31 sequences and characterized sequences (1194 sequences) were subjected to the phylogenetic analysis (Fig. 1). A clade of uncharacterized proteins derived from bacteria and fungi was detected (Fig. S1). These proteins are classified as GH31_u1 family in National Center for Biotechnology Information Conserved Domain Database and showed <25% amino acid sequence identity with the previously characterized GH31 enzymes, indicating that the activities and substrate specificities of GH31_u1 proteins differed from those of characterized GH31 enzymes. In this study, GH31_u1 proteins from Lactococcus lactis subsp. cremoris MG1363 (named LlGH31_u1) and Cordyceps militaris (named CmGH31_u1) were targeted to clarify the functions and structures of GH31_u1 proteins. LlGH31_u1 is the sole GH31 protein encoded in the genome of L. lactis subsp. cremoris, whereas the C. militaris has three other GH31 proteins (not GH31_u1) compared with CmGH31_u1. LlGH31_u1 and CmGH31_u1 share 37.5% amino acid sequence identity, have no signal peptide, and have no domain annotated in the Conserved Domains Database in addition to the conserved GH31 region, indicating that they possibly function intracellularly.
      Figure thumbnail gr1
      Figure 1Phylogenetic tree of GH31 proteins. A total of 1194 GH31 sequences were used to generate the phylogenetic tree. GH31 catalytic domains were aligned using the MUSCLE algorithm, and the phylogenetic tree was built using the maximum likelihood method with MEGA X (
      • Kumar S.
      • Stecher G.
      • Li M.
      • Knyaz C.
      • Tamura K.
      Mega X: Molecular evolutionary genetics analysis across computing platforms.
      ). Symbols: solid circle, the enzymatically characterized proteins; solid star, the enzymatically characterized and structure-determined proteins; open star, the protein whose structure has been deposited on PDB. Symbol colors are listed below the tree. The clade of GH31_u1 proteins is highlighted with gray background. GH, glycoside hydrolase.

      Enzymatic features of recombinant GH31_u1 proteins

      N-terminally His-tagged LlGH31_u1 and CmGH31_u1 were expressed in Escherichia coli and purified to homogeneity. The yields of the purified LlGH31_u1 and CmGH31_u1 were 2 and 0.1 mg from 1 L of culture, respectively. These enzymes were active against only p-nitrophenyl α-D-glucopyranoside (pNP-α-Glc) among the p-nitrophenyl glycosides tested, suggesting that they are α-glucoside hydrolases. The effects of pH and temperature on the hydrolytic activity of LlGH31_u1 were evaluated using pNP-α-Glc as a substrate. LlGH31_u1 showed the highest activity at pH 7.0 for 5-min incubation at 30 °C and maintained >79% of its maximal activity in the pH range of 4.5 to 8.0 after a 20-h incubation at 4 °C. The enzyme showed the highest activity at 35 °C for 5-min incubation at pH 7.0, but it was unstable at temperatures >35 °C (<63%) and approximately 5% of its activity remained after a 30-min incubation at 40 °C (Fig. S2). CmGH31_u1 exhibited the highest activity at pH 6.0 and 40 °C for 5-min incubation in the condition described in Experimental procedures (Fig. S3).
      Next, we analyzed hydrolytic activities toward disaccharides, trehalose [α-D-Glcp-(1↔1)-α-D-Glcp], kojibiose [α-D-Glcp-(1→2)-D-Glc], nigerose [α-D-Glcp-(1→3)-D-Glc], maltose [α-D-Glcp-(1→4)-D-Glc], isomaltose [α-D-Glcp-(1→6)-D-Glc], and sucrose [β-D-Fruf-(2↔1)-α-D-Glcp] using thin-layer chromatography (TLC). After a 1-h incubation with 0.1 mg/ml LlGH31_u1, only nigerose was completely hydrolyzed; the hydrolysis of kojibiose and maltose was also observed (Fig. 2). Both LlGH31_u1 and CmGH31_u1 exhibited the highest hydrolytic activity toward nigerose, followed by kojibiose, maltose, and isomaltose in that order (Table 1). We detected no hydrolytic activity when either trehalose or sucrose was used as a substrate. The Km and kcat values of nigerose were 9.2 ± 0.6 mM and 9.7 ± 0.3 s−1, respectively (Table 2). By contrast, the Km value of kojibiose was almost the same as that of nigerose, and the kcat value was 8.6 times lower than that of nigerose. The higher Km and lower kcat values were observed when maltose was employed as a substrate. We then analyzed the substrate specificity of LlGH31_u1 for longer oligosaccharides. The Km values of nigerotetraose and nigerotriose were approximately three times lower than that of nigerose. The Km values of LlGH31_u1 for nigerose are in the same range (millimolar order) as those of other GH31 α-glucosidases for their preferred substrates (Table S1). However, maltotriose and maltotetraose have higher Km values than nigerose. LlGH31_u1 did not release glucose from nigeran, an insoluble linear α-glucan with alternating α-1,3- and α-1,4-glucosidic linkages derived from Aspergillus niger.
      Figure thumbnail gr2
      Figure 2TLC analysis of LlGH31_u1 hydrolysis toward disaccharides. LlGH31_u1 (0.1 mg/ml) was incubated with 10 mM disaccharides (trehalose, kojibiose, nigerose, maltose, isomaltose, and sucrose) for 1 h at 30 °C. Glucose standard and the reaction mixtures were developed on a TLC plate with 1-butanol/ethanol/water (10:5:2, vol/vol).
      Table 1Activities of LlGH31_u1 WT, Y99F, and CmGH31_u1 against disaccharides
      SubstrateLlGH31_u1 WTLlGH31_u1 Y99FCmGH31_u1
      Activity (μmol/mg/min)Relative activity
      Each activity for nigerose is taken as 100%.
      (%)
      Activity (μmol/mg/min)Relative activity (%)Activity (μmol/mg/min)Relative activity (%)
      TrehaloseND
      ND, not detected.
      NDNDNDNDND
      Kojibiose0.34 ± 0.025.2(49 ± 0.7) × 10−31.3(2.4 ± 0.1) × 10−24.2
      Nigerose6.3 ± 0.51003.9 ± 0.2100(5.6 ± 0.7) × 10−2100
      Maltose(10 ± 0.8) × 10−30.16(20 ± 11) × 10−30.53(8.2 ± 3.8) × 10−31.5
      Isomaltose(2.9 ± 3.5) × 10−30.045(5.4 ± 0.2) × 10−30.15(8.4 ± 0.3) × 10−31.5
      SucroseNDNDNDNDNDND
      a Each activity for nigerose is taken as 100%.
      b ND, not detected.
      Table 2Kinetic parameters of LlGH31_u1 and CmGH31_u1
      EnzymeSubstrateKm (mM)kcat (s −1)kcat/Km (s−1 mM−1)Relative kcat/Km (%)
      kcat/Km value of LlGH31_u1 WT for nigerose is taken as 100%.
      LlGH31_u1 WTpNP-α-Glc20 ± 0.831 ± 0.71.5 ± 0.08145
      Nigerose9.2 ± 0.69.7 ± 0.31.1 ± 0.2100
      Nigerotriose3.1 ± 0.48.1 ± 0.42.6 ± 0.3246
      Nigerotetraose2.9 ± 0.25.6 ± 0.21.9 ± 0.2182
      Kojibiose8.2 ± 0.41.1 ± 0.02(1.4 ± 0.05) × 10−113
      Maltose36 ± 1.9(7.9 ± 0.2) × 10−1(2.2 ± 0.03) × 10−22.1
      Maltotriose39 ± 2.7(6.2 ± 0.2) × 10−1(1.6 ± 0.09) × 10−21.5
      Maltotetraose82 ± 17(1.5 ± 0.2) × 10−1(1.8 ± 0.4) × 10−30.2
      LlGH31_u1 Y99FpNP-α-Glc28 ± 0.859 ± 0.82.1 ± 0.02204
      Nigerose12 ± 2.05.2 ± 0.4(4.3 ± 0.2) × 10−138
      Nigerotriose2.0 ± 0.24.9 ± 0.42.3 ± 0.9221
      CmGH31_u1pNP-α-Glc4.7 ± 0.84.1 ± 0.3(8.9 ± 2) × 10−1
      Nigerose6.3 ± 0.6(4.5 ± 0.3) × 10−1(7.2 ± 1) × 10−2
      Nigerotriose2.0 ± 0.6(4.7 ± 0.3) × 10−1(2.3 ± 0.4) × 10−1
      a kcat/Km value of LlGH31_u1 WT for nigerose is taken as 100%.
      Based on the CAZy classification, while α-1,4-glucosidases (EC 3.2.1.20) are present in the GH families 4, 13, 31, and 122 (
      • Lombard V.
      • Golaconda Ramulu H.
      • Drula E.
      • Coutinho P.M.
      • Henrissat B.
      The carbohydrate-active enzymes database (CAZy) in 2013.
      ,
      • Okuyama M.
      Function and structure studies of GH family 31 and 97 α-glycosidases.
      ,
      • Chang S.T.
      • Parker K.N.
      • Bauer M.W.
      • Kelly R.M.
      α-Glucosidase from Pyrococcus furiosus.
      ,
      • Lodge J.A.
      • Maier T.
      • Liebl W.
      • Hoffmann V.
      • Sträter N.
      Crystal structure of Thermotoga maritima α-glucosidase AglA defines a new clan of NAD+-dependent glycosidases.
      ), the GH31 and GH63 families contain nigerose-active α-glucoside hydrolases (
      • Okuyama M.
      • Saburi W.
      • Mori H.
      • Kimura A.
      α-Glucosidases and α-1,4-glucan lyases: Structures, functions, and physiological actions.
      ,
      • Kurakata Y.
      • Uechi A.
      • Yoshida H.
      • Kamitori S.
      • Sakano Y.
      • Nishikawa A.
      • Tonozuka T.
      Structural insights into the substrate specificity and function of Escherichia coli K12 YgjK, a glucosidase belonging to the glycoside hydrolase family 63.
      ). However, YgjK, a GH63 enzyme, from E. coli had a high Km value (approximately 230 mM) for nigerose (
      • Kurakata Y.
      • Uechi A.
      • Yoshida H.
      • Kamitori S.
      • Sakano Y.
      • Nishikawa A.
      • Tonozuka T.
      Structural insights into the substrate specificity and function of Escherichia coli K12 YgjK, a glucosidase belonging to the glycoside hydrolase family 63.
      ), and its natural substrate was assumed to be another α-glucoside (
      • Miyazaki T.
      • Ichikawa M.
      • Yokoi G.
      • Kitaoka M.
      • Mori H.
      • Kitano Y.
      • Nishikawa A.
      • Tonozuka T.
      Structure of a bacterial glycoside hydrolase family 63 enzyme in complex with its glycosynthase product, and insights into the substrate specificity.
      ,
      • Miyazaki T.
      • Nishikawa A.
      • Tonozuka T.
      Crystal structure of the enzyme-product complex reveals sugar ring distortion during catalysis by family 63 inverting α-glycosidase.
      ). Other than GHs, a nigerose phosphorylase from Lachnoclostridium phytofermentans (formerly Clostridium phytofermentans) (Cphy1874) belonging to the GH65 family displays high specificity against nigerose (Table S1), but it does not catalyze hydrolysis (
      • Nihira T.
      • Nakai H.
      • Chiku K.
      • Kitaoka M.
      Discovery of nigerose phosphorylase from Clostridium phytofermentans.
      ). At present, the catalytic α subunit of ER α-glucosidase II, CADE (
      • Light S.H.
      • Cahoon L.A.
      • Mahasenan K.V.
      • Lee M.
      • Boggess B.
      • Halavaty A.S.
      • Mobashery S.
      • Freitag N.E.
      • Anderson W.F.
      Transferase versus hydrolase: The role of conformational flexibility in reaction specificity.
      ,
      • Tagami T.
      • Miyano E.
      • Sadahiro J.
      • Okuyama M.
      • Iwasaki T.
      • Kimura A.
      Two novel glycoside hydrolases responsible for the catabolism of cyclobis-(1→6)-α-nigerosyl.
      ), and Lactobacillus johnsonii α-1,3-glucosidase (LJAG31) (
      • Kang M.S.
      • Okuyama M.
      • Mori H.
      • Kimura A.
      The first α-1,3-glucosidase from bacterial origin belonging to glycoside hydrolase family 31.
      ) have been identified as GH31 enzymes that display specificity against α-1,3-glucosidic linkage. ER α-glucosidase II hydrolyzes α-1,3 linkages in the α-Glc-(1→3)-Glc and α-Glc-(1→3)-mannose moieties of N-glycans. Despite its biological role in cleaving α-1,3-linked glucose, ER α-glucosidase II has a significant kcat/Km value on maltose (27%–50% of that for nigerose, Table S1) (
      • Okuyama M.
      • Miyamoto M.
      • Matsuo I.
      • Iwamoto S.
      • Serizawa R.
      • Tanuma M.
      • Ma M.
      • Klahan P.
      • Kumagai Y.
      • Tagami T.
      • Kimura A.
      Substrate recognition of the catalytic α-subunit of glucosidase II from Schizosaccharomyces pombe.
      ). As shown in Figure 1, mammalian α-glucosidases, i.e., ER α-glucosidase II, MGAM, sucrase-isomaltase, and lysosomal acid α-glucosidase, are phylogenetically close, and this agrees with the enzymatic feature acting on both the α-1,3 and α-1,4 bonds. LJAG31 displays the highest nigerose hydrolytic activity, although it exhibits a broad substrate specificity; the kcat/Km values for maltulose [α-D-Glcp-(1→4)-D-Fru] and kojibiose were 73% and 61%, respectively, of those for nigerose (
      • Kang M.S.
      • Okuyama M.
      • Mori H.
      • Kimura A.
      The first α-1,3-glucosidase from bacterial origin belonging to glycoside hydrolase family 31.
      ). CADE is cycloalternan-specific, and its hydrolytic activity against nigerose is <0.1% of that for cycloalternan (
      • Tagami T.
      • Miyano E.
      • Sadahiro J.
      • Okuyama M.
      • Iwasaki T.
      • Kimura A.
      Two novel glycoside hydrolases responsible for the catabolism of cyclobis-(1→6)-α-nigerosyl.
      ). However, the kcat/Km value of LlGH31_u1 against other glucobiose is <13% of that against nigerose, and this enzyme exhibits the highest substrate specificity for nigerose among the characterized GH31 enzymes.

      Crystal structure of LlGH31_u1

      WT LlGH31_u1 was crystallized in two different crystal forms (named, crystal forms 1 and 2, see Experimental procedures). Crystal forms 1 and 2 belong to the space groups P21 and P6322, respectively, and comprise six and one molecule of LlGH31_u1, respectively, in an asymmetric unit. The P21 and P6322 structures of unliganded WT (named WT_P21 and WT_P6322, respectively) were determined at 1.75- and 1.85-Å resolutions (Fig. 3A and Table S2). Despite the space group difference, hexamer formation could be generated via the symmetry operations in the P6322 space group, and the hexamer assembly is isomorphous with six monomers in WT_P21. LlGH31_u1 has a theoretical mass of 85.7 kDa and a molecular weight of 453 kDa as determined using gel filtration chromatography. With molecular assembly in the crystal structures, LlGH31_u1 probably forms a hexamer. Two GH31 enzymes, α-xylosidase YicI from E. coli and α-glucosidase MalA from Sulfolobus solfataricus, reportedly formed a hexamer (
      • Lovering A.L.
      • Lee S.S.
      • Kim Y.W.
      • Withers S.G.
      • Strynadka N.C.
      Mechanistic and structural analysis of a family 31 α-glycosidase and its glycosyl-enzyme intermediate.
      ,
      • Ernst H.A.
      • Lo Leggio L.
      • Willemoës M.
      • Leonard G.
      • Blum P.
      • Larsen S.
      Structure of the Sulfolobus solfataricus α-glucosidase: Implications for domain conservation and substrate recognition in GH31.
      ); however, the hexameric arrangement of LlGH31_u1 differs from theirs (Fig. S4).
      Figure thumbnail gr3
      Figure 3Overall structure of LlGH31_u1 determined using X-ray crystallography. A, hexameric structure of LlGH31_u1 determined using the crystal form 1 in which the space group is P21. Each monomer is shown in different colors. B, ribbon model of LlGH31_u1 monomer. Left, view from the front of the active site cleft; right, view rotated to adjust to the same orientation as the monomer colored in red in the right panel of (A). Individual domains are colored as follows: N-domain, yellow; A-domain, red; insert 1, pink; insert 2, blue; proximal C-domain, cyan; distal C-domain, orange; linker, gray; and C-terminal α-helix domain, green.
      The monomeric structure of LlGH31_u1 contains four domains that are generally conserved in GH31 enzymes: N-terminal β-sandwich domain (N-domain, residues 1–172); (β/α)8-barrel catalytic domain (A-domain, residues 173–502) with two inserted components insert 1 (residues 232–249) and insert 2 (residues 286–308) located between β3 and α4 and between β4 and α5, respectively; proximal C-domain (residues 503–586); and distal C-domain (residues 587–669). An extra domain consisted of four α-helices (residues 680–739) at the C terminus connected by a linker polypeptide (residues 670–679) (Fig. 3B). Interactions between insertions in the A-domain and a β-sheet of the N-domain from an adjacent protomer as well as the C-terminal α-helix domain contribute to hexamer formation (Figs. 3A and S5). We performed a structural similarity search on the DALI server (
      • Holm L.
      • Rosenström P.
      Dali server: Conservation mapping in 3D.
      ) using WT_P6322 as a query. An α-xylosidase from Bacteroides ovatus [BoGH31; Protein Data Bank (PDB) 5JOV; Z = 33.1; amino acid sequence identity = 23%] had the highest Z-score, followed by a cycloalternan-degrading enzyme from Trueperella pyogenes (TpCADE; PDB 5I0G; Z = 32.1; identity = 19%) and other structurally identified GH31 enzymes. The DALI results also included ER α-glucosidase II from Chaetomium thermophilum var. thermophilum (CtGII, 5DKY, Z = 28.1; identity = 19%), similarly as other GH31 enzymes. A DALI search using the C-terminal α-helix domain revealed that it has the highest structural similarity with harmonin homology domain 2 of whirlin (PDB 6FDD, Z = 5.8), which is expressed in hair and photoreceptor cells and is essential for sound and light perception, although their amino acid sequence identity was 12%. The harmonin homology domain of cerebral cavernous malformations 2 (PDB 4FQN) was reported to be involved in dimer formation (
      • Fisher O.S.
      • Zhang R.
      • Li X.
      • Murphy J.W.
      • Demeler B.
      • Boggon T.J.
      Structural studies of cerebral cavernous malformations 2 (CCM2) reveal a folded helical domain at its C-terminus.
      ). Although these proteins and LlGH31_u1 have no physiological function in common, this α-helix fold may be relevant in oligomer formation.

      Comparison of cryo-EM and crystal structures of LlGH31_u1

      We determined the cryo-EM structure of LlGH31_u1 at 2.73 Å resolution using single-particle analysis (Fig. 4). Table S3 summarizes the statistics of data collection, image processing, and 3D reconstruction steps. The cryo-EM structure revealed the hexameric state of LlGH31_u1, and the overall cryo-EM structure was similar to the crystal structure (Figs. 4, 5 A and B). The Cα root-mean-square deviation between the cryo-EM and crystal structures is 1.340 Å. In both the cryo-EM and crystal structures, most amino acid residues at the hexamer interface were identical (Figs. 5C and S5). The cryo-EM maps for the side chains of the amino acid residues in the C-terminal α-helix domain were well resolved, and most of their conformations were almost identical to those in the crystal structure (Fig. 5, C and D). At the C-terminal α-helix domain, Val716 and Phe732 hydrophobically interact, and Gln728 forms a hydrogen bond with Lys720 (Fig. S6). By contrast, the local resolutions of insert 1 and insert 2 on A-domain and a loop containing Tyr99 on N-domain (discussed below) are lower than those of the other regions, and the cryo-EM map of these side chains is unclear (Figs. 4C and 5E). Furthermore, slight structural differences between the crystal and cryo-EM structures were observed (Fig. 5F). Although the insertion regions are determined with low B factors in the crystal structure (the average B-factor values of insert 1 and insert 2 were 30.2 Å2 and 33.5 Å2, respectively, and those of the other structural components was 33.7 Å2), the structural difference may suggest insert 1 and insert 2 flexibility without the crystal packing effect.
      Figure thumbnail gr4
      Figure 4Cryo-EM single-particle analysis of LlGH31_u1. A and B, representative micrograph (A) and two-dimensional class averages of the cryo-EM dataset (B). C, local resolution of final cryo-EM structure. Unit of color bar label is in Å. The diagram is viewed from the same orientation as the ribbon models in A and A. cryo-EM, cryogenic electron microscopy.
      Figure thumbnail gr5
      Figure 5Comparison with cryo-EM and crystal structures of LlGH31_u1. A and B, superimposition of hexamer (A) or monomer (B) of the cryo-EM structure (cyan) and crystal structure (red). C, cryo-EM map (contoured at 5 σ) (right) and 2Fo − Fc map (contoured at 1 σ) (left) of the C-terminal α-helix domain. The main and side chains of the domain are shown as stick models. D, superimposition of the side chains (stick model) of the C-terminal α-helix domain in the cryo-EM structure (magenta) and crystal structure (slate blue). The residues located in the interface are labeled. E, cryo-EM map (contoured at 5 σ) (right) and 2Fo − Fc map (contoured at 1 σ) (left) of the loop in which Y99 is present. F, superimposition of the loop where Y99 exists in the cryo-EM structure (magenta) and crystal structure (slate blue). cryo-EM, cryogenic electron microscopy.

      Crystal structures of LlGH31_u1 complexed with ligands

      First, we determined the crystal structure of WT in complex with glucose (WT-Glc). Asp341 and Asp394, which are located near the C1 atom of the glucose, are assumed to be catalytic nucleophile and acid/base, respectively, similar to other GH31 enzymes (
      • Okuyama M.
      • Saburi W.
      • Mori H.
      • Kimura A.
      α-Glucosidases and α-1,4-glucan lyases: Structures, functions, and physiological actions.
      ). In accordance, the structures of inactive mutant D394A, in which the catalytic acid/base D394 was substituted with alanine, and the structure of complexes with nigerose (D394A-Nig2), nigerotriose (D394A-Nig3), nigerotetraose (D394A-Nig4), and kojibiose (D394A-Koj2) were also determined. At the catalytic site of LlGH31_u1, we observed the electron density maps for all ligands. In D394A-Nig3 and WT-Glc, both the α- and β-anomers of the ligand were modeled, while an electron density map for α-nigerose, α-nigerotetraose, and α-kojibiose was observed in D394A-Nig2, D394A-Nig4, and D394A-Koji2, respectively (Fig. 6A). Hereafter, the glucose residues bound to subsite −1, subsite +1, subsite +2, and subsite +3 (subsite nomenclature is according to literature (
      • Davies G.J.
      • Wilson K.S.
      • Henrissat B.
      Nomenclature for sugar-binding subsites in glycosyl hydrolases.
      )) are referred to as Glc−1, Glc+1, Glc+2, and Glc+3, respectively.
      Figure thumbnail gr6
      Figure 6Ligand–complex structures of LlGH31_u1. A, Fo − Fc electron density map for ligands (contoured at 2 σ). B, active sites of D394A-Nig3. Side chains of amino acid residues that interact with nigerotriose are shown as stick models. Residues forming hydrogen bonds are colored cyan, residues forming a hydrophobic environment are yellow, and nigerotriose is green. A water molecule is shown as the red sphere model and hydrogen bonds are shown as the dashed line. C, surface model of D394A-Nig4. Nigerotetraose is shown as the green stick model, and an adjacent protomer is shown in slate blue. D, superposition of D394A-Nig2 (green) and D394A-Koj2 (pink). Side chain of the general acid/base Asp394 (yellow) in WT_P6322 is also overlaid. Ligands and amino acid residues forming hydrogen bonds with Glc+1 are shown as thick and thin stick models, respectively. The distance (1.2 Å) between the OD2 of Asp394 and Glc+1 O1 of kojibiose is indicated as the double arrow.
      In the −1 subsite, the Glc−1 of all ligands (glucose, nigerose, nigerotriose, nigerotetraose, and kojibiose) are accommodated in the active site pocket in almost the same manner (Fig. S7) and recognized by surrounding residues via hydrogen bonds and hydrophobic interaction (Fig. 6B). In the subsite +1, in addition to three hydrogen bonds with the catalytic site of A-domain, Tyr99 on the N-domain of an adjacent protomer forms a hydrogen bond with the 4-OH group of Glc+1 of nigerooligosaccharides (Fig. 6B). The electron densities of two distinct rotamers of Ser455 are observed in some monomers of WT_P21 and ligand–complex structures, including D394A-Nig3. One rotamer does not interact with the ligand, but others form hydrogen bonds with the 2-OH of Glc+1 and 4-OH of Glc+2 (Fig. 6B). Ser455 may contribute to substrate recognition at subsite +2, and no other amino acid residue forms a hydrogen bond with Glc+2. In D394-Nig4, the Glc+3 of nigerotetraose protrudes from the catalytic pocket (Fig. 6C). These findings imply that the substrate recognition of LlGH31_u1 in subsites +2 and +3 is less strict than that in subsites −1 and +1.
      In the D394A-Koj2 structure, the Glc+1 adopts the same chair 4C1 conformation as in the D394A-Nig2, but its sugar ring is overturned (Fig. 6D). His427 interacts with the 3-OH group of kojibiose Glc+1. Gln343 interacts with the axial 1-OH group of α-kojibiose Glc+1, but Tyr99 on an adjacent subunit does not interact with kojibiose. However, the superimposition of D394A-Koj2 with WT_P6322 shows that the estimated distance between the Asp394 and 1-OH of kojibiose was 1.2 Å, indicating a steric hindrance. Although it is unclear how kojibiose binds to the WT catalytic site, β-kojibiose may bind to WT enzyme and/or the glycosidic bond of kojibiose may not be in a suitable position for proton acceptance, resulting in a low turnover.

      Difference in substrate recognition mechanisms between LlGH31_u1 and GH31 α-glucoside hydrolases

      The active site of D394A-Nig2 was compared with CtGII (PDB code 5DKZ). At subsite −1, the glucose recognition mechanisms of LlGH31_u1 and CtGII are similar. At the subsite −1 of CtGII structure, in addition to a pair of catalytic residues, residues that identified Glc−1, Asp443, Trp554, Asp662, and His691 are conserved in LlGH31_u1 (Fig. 7, A and B). Although Trp630 in CtGII is substituted with Phe391 in LlGH31_u1, the aromatic rings of the residues are located in a similar position and are involved in a hydrophobic environment formation. While the 6-OH group of Glc−1 interacts with Asp482 and Trp517 in CtGII via a water molecule, it establishes a direct hydrogen bond with His280 in LlGH31_u1 (Fig. 7A). At subsite +1, the amino acid residues involved in glucose recognition differ. Asp303 and Arg617 in CtGII recognize the 4-OH of Glc+1, whereas it is recognized by the Gln343 and Tyr99 of the neighbor protomer in LlGH31_u1. Unlike CtGII, no hydrogen bond with the 2-OH of Glc+1 is observed because the His427 of LlGH31_u1 is substituted with Phe666 in CtGII (Fig. 7C).
      Figure thumbnail gr7
      Figure 7Comparison of the active sites of LlGH31_u1 with CtGII and NtMGAM. A, superimposition of CtGII complexed with nigerose (PDB 5DKZ, orange) and D394A-Nig2 (cyan). Side chains of amino acid residues at subsite −1 and nigerose are shown as thick and thin stick models, respectively. Water molecules in CtGII and LlGH31_u1 are shown as red and violet sphere models, respectively. Hydrogen bonds with the O6 of Glc−1 are shown as dashed lines. Asterisks indicate an artificially mutated residue (Asp394→Ala). B, the amino acid sequence alignment of LlGH31_u1 and eukaryotic α-glucosidases in which structure is determined, i.e., CtGII (GenBank ID, EGS17181.1); MmGII, ER α-glucosidase II from Mus musculus (AAC53182.1); hGAA, human lysosomal acid α-glucosidase (CAA68763.1); C-terminal (Ct), and N-terminal (Nt) domains of human MGAM (AAC39568.2); hNtSI, N-terminal sucrase domain of human sucrase-isomaltase (AAT18166.1); SBG, α-glucosidase from sugar beet Beta vulgaris (BAM74081.1). Red triangles indicate the catalytic residues; blue and light blue triangles indicate the residues that interact with Glc−1 and Glc+1, respectively. C, comparison of subsite +2 in CtGII complexed with nigerose (orange) and D394A-Nig2 (cyan). Hydrogen bonds between nigerose and the subsite +1 residues are shown as dashed lines. Tyr99 side chain of an adjacent protomer in D394A-Nig2 is shown as a slate blue stick. D, comparison of subsite +1 in NtMGAM complexed with acarbose (yellow) and D394A-Nig2 (cyan). Acarviosine moiety of acarbose is shown for clarity. Ligands are shown in thin sticks. Residues of NtMGAM interacting with α-D-6-deoxy-glucopyranose and those of LlGH31_u1 interacting with nigerose are compared. Hydrogen bonds between substrate and protein are shown as yellow dashed lines in NtMGAM and cyan dashed line in LlGH31_u1.CtGII, α-glucosidase II from Chaetomium thermophilum var. thermophilum; NtMGAM, N-terminal maltase domain of human maltase-glucoamylase.
      Because no crystal structure of GH31 maltase complexed with maltooligosaccharide substrates is available, the structure of the N-terminal maltase domain of human maltase-glucoamylase (NtMGAM) complexed with inhibitor acarbose (PDB code 2QMJ) (
      • Sim L.
      • Quezada-Calvillo R.
      • Sterchi E.E.
      • Nichols B.L.
      • Rose D.R.
      Human intestinal maltase-glucoamylase: Crystal structure of the N-terminal catalytic subunit and basis of inhibition and substrate specificity.
      ) was used for comparison to elucidate why LlGH31_u1 has a low activity for maltooligosaccharides. The residues that interact with Glc−1 and Glc+1 of the substrate are conserved in eukaryotic GH31 α-glucosidases, including ER α-glucosidase II and NtMGAM, whose structures have been determined (
      • Satoh T.
      • Toshimori T.
      • Yan G.
      • Yamaguchi T.
      • Kato K.
      Structural basis for two-step glucose trimming by glucosidase II involved in ER glycoprotein quality control.
      ,
      • Sim L.
      • Quezada-Calvillo R.
      • Sterchi E.E.
      • Nichols B.L.
      • Rose D.R.
      Human intestinal maltase-glucoamylase: Crystal structure of the N-terminal catalytic subunit and basis of inhibition and substrate specificity.
      ) (Fig. 7D). In the NtMGAM–acarbose complex, the α-D-6-deoxy-glucopyranose residue adopts the same 4C1 chair conformation as the Glc+1 of nigerose in D394A-Nig2, but the pyranose ring is overturned. If maltose interacts with acarbose, Gln343, and Tyr99 in a similar orientation, residues that form hydrogen bonds with a 4-OH group of Glc+1 of nigerose may not prevent maltose from binding. A space exists surrounding the methyl group of acarbose in the NtMGAM–acarbose complex structure, which is probably capable of accepting the 6-OH group of Glc+1 of maltooligosaccharides (Fig. 8A). However, because LlGH31_u1 lacks additional space surrounding the 2-OH group of Glc+1 of nigerooligosaccharides (Fig. 8B), the 6-OH group of maltose may have difficulty fitting into the catalytic pocket. Moreover, compared with the NtMGAM–acarbose complex, Met122 and Asn429 in LlGH31_u1 may inhibit acarbose binding (Fig. 8, A and B). A similar steric hindrance for acarbose is also observed in the α-subunit of ER α-glucosidase II (
      • Satoh T.
      • Toshimori T.
      • Yan G.
      • Yamaguchi T.
      • Kato K.
      Structural basis for two-step glucose trimming by glucosidase II involved in ER glycoprotein quality control.
      ). These findings support the result that LlGH31_u1 is substantially less active for maltooligosaccharides.
      Figure thumbnail gr8
      Figure 8Surface model of the active sites of NtMGAM and LlGH31_u1. A, surface model of NtMGAM in complex with acarbose. Acarbose is shown as a white stick model. B, surface model of D394A-Nig4. Nigerotetraose is shown as green stick models, and adjacent protomer is shown in slate blue. Dotted circles indicate the space around the methyl group of acarbose and the space around the 2-OH group of Glc+1 of nigerooligosaccharides. NtMGAM, N-terminal maltase domain of human maltase-glucoamylase.

      Function of LlGH31_u1 Tyr99

      Structural analysis suggested that the Tyr99 on N-domain is involved in the formation of subsite +1 of an adjacent protomer. However, the cryo-EM structure suggested the flexibility of this residue (Fig. 5E), and this tyrosine residue is not strictly conserved among LlGH31_u1 homologs, including CmGH31_u1 (Fig. S8). Therefore, we constructed a valiant Y99F in which Tyr99 is substituted with phenylalanine to determine whether Tyr99 affects LlGH31_u1 activity. Y99F was expressed in E. coli and purified in the same manner as the WT enzyme, and its activities for various substrates were determined. Contrary to our expectations, Y99F exhibits the same substrate preference as the WT enzyme (nigerose >> kojibiose > maltose > isomaltose) but is inactive against trehalose and sucrose (Table 1). The Km values of the LlGH31_u1_Y99F mutant for nigerose and nigerotriose were 12 ± 2 mM and 2.0 ± 0.2 mM, respectively, which were nearly identical to those of WT. The kcat values for nigerose and nigerotriose were 5.2 ± 0.4 s−1 and 4.9 ± 0.4 s−1, respectively, and are approximately two times lower than that of WT. By contrast, the kcat value for pNP-Glc, 59 ± 0.8 s−1, was two times higher than that of WT (Table 2). Thus, it was proposed that Tyr99 affects the turnover of the hydrolytic reaction rather than the substrate specificity. However, further analysis is required to elucidate the role of Tyr99.

      Estimation of the physiological role of LlGH31_u1 and its homologs

      Although some bacterial species degrade eukaryotic N-glycans, mature proteins lack nigerose units on the nonreducing ends of N-glycans because the α-1,3-glucosidic moiety is cleaved in ER (
      • Okuyama M.
      • Miyamoto M.
      • Matsuo I.
      • Iwamoto S.
      • Serizawa R.
      • Tanuma M.
      • Ma M.
      • Klahan P.
      • Kumagai Y.
      • Tagami T.
      • Kimura A.
      Substrate recognition of the catalytic α-subunit of glucosidase II from Schizosaccharomyces pombe.
      ,
      • Higgins M.A.
      • Tegl G.
      • MacDonald S.S.
      • Arnal G.
      • Brumer H.
      • Withers S.G.
      • Ryan K.S.
      N-glycan degradation pathways in gut- and soil-dwelling actinobacteria share common core genes.
      ). Therefore, eukaryotic N-glycan may not be a natural substrate of LlGH31_u1, unlike ER α-glucosidase II. One possible source of the substrate of LlGH31_u1 is mutan, which is an exopolysaccharide that contains α-1,6 and α-1,3 bonds and is synthesized by three different glucansucrases belonging to the GH70 family in Streptococcus mutans (
      • Bowen W.H.
      • Koo H.
      Biology of Streptococcus mutans-derived glucosyltransferases: Role in extracellular matrix formation of cariogenic biofilms.
      ). The genome of S. mutans encodes GH31_u1 and may be involved in mutan metabolism. Some lactic acid bacteria also synthesize exopolysaccharides, which contain a major α-1,6 linkage and a minor α-1,3 linkage (
      • Bounaix M.S.
      • Gabriel V.
      • Morel S.
      • Robert H.
      • Rabier P.
      • Remaud-Siméon M.
      • Gabriel B.
      • Fontagné-Faucher C.
      Biodiversity of exopolysaccharides produced from sucrose by sourdough lactic acid bacteria.
      ,
      • Côté G.L.
      • Skory C.D.
      Cloning, expression, and characterization of an insoluble glucan-producing glucansucrase from Leuconostoc mesenteroides NRRL B-1118.
      ,
      • Malik A.
      • Radji M.
      • Kralj S.
      • Dijkhuizen L.
      Screening of lactic acid bacteria from Indonesia reveals glucansucrase and fructansucrase genes in two different Weissella confusa strains from soya.
      ). However, there has been no report on the synthesis of glucans with α-1,3 linkage from Lactococcus species, and there is no gene estimated as GH70 glucansucrase in L. lactis genome.
      α-1,3-Glucan produced by fungi is another potential natural source of GH31_u1 substrate. The cell wall of filamentous fungi contains α-1,3-glucan, which has an α-1,3-glucosidic linkage and partial α-1,4 linkage as one of the principal components (
      • Yoshimi A.
      • Miyazawa K.
      • Abe K.
      Cell wall structure and biogenesis in Aspergillus species.
      ). Based on the result that CmGH31_u1 exhibited high substrate specificity to nigerose, fungal GH31_u1 homologs may be involved in the metabolism of α-1,3-glucans. Because the hydrolytic activity of LlGH31_u1 for nigeran was not detected, GH31_u1 may require cooperation with an endo-type enzyme to degrade polysaccharides. GH71 proteins are known as fungal endo-α-1,3-glucanases, and C. millitalis also has a gene encoding a GH71 protein. In addition, Talaromyces verruculosus has a gene for GH31_u1 fused with the GH71 domain (GenBank, KUL90319.1). Therefore, fungal GH31_u1 enzymes may be involved in the degradation of cell wall α-1,3-glucan in cooperation with GH71 enzymes.
      Nigerose phosphorylase from L. phytofermentans (Cphy1874), belonging to GH65, has high substrate specificity for nigerose. In the genome of L. phytofermentans, a putative GH87 endo-α-1,3-glucanase is located near the GH65 nigerose phosphorylase, and its participation in nigerose degradation is proposed (
      • Nihira T.
      • Nakai H.
      • Chiku K.
      • Kitaoka M.
      Discovery of nigerose phosphorylase from Clostridium phytofermentans.
      ). In addition, GH31_u1 (locus tag Cphy_1877) is located in the same gene cluster as nigerose phosphorylase, along with genes for ABC transporter proteins and LacI family transcription regulators (Fig. 9). In the genome of L. lactis, LlGH31_u1 is also surrounded by the genes for ABC transporter proteins and LacI family transcription regulator. S. mutans has a gene cluster similar to that of L. lactis; however, it has the genes for phosphotransferase system instead of the ABC transporter system (
      • Deutscher J.
      • Francke C.
      • Postma P.W.
      How phosphotransferase system-related protein phosphorylation regulates carbohydrate metabolism in bacteria.
      ,
      • Koropatkin N.M.
      • Cameron E.A.
      • Martens E.C.
      How glycan metabolism shapes the human gut microbiota.
      ). Other bacterial species harboring GH31_u1 possess various GHs and sugar utilization proteins genes in the vicinity of GH31_u1; therefore, GH31_u1 may be involved in various degradation systems for saccharides in widespread species. L. lactis is isolated from various sources, including dairy products, fermented foods, plants, and soil (
      • Siezen R.J.
      • Bayjanov J.R.
      • Felis G.E.
      • van der Sijde M.R.
      • Starrenburg M.
      • Molenaar D.
      • Wels M.
      • van Hijum S.A.
      • van Hylckama Vlieg J.E.
      Genome-scale diversity and niche adaptation analysis of Lactococcus lactis by comparative genome hybridization using multi-strain arrays.
      ). L. lactis may absorb and use oligosaccharides, which might be derived from polysaccharides produced by other organisms, using LlGH31_u1. However, LlGH31_u1 may require cooperation with other enzymes, at least with an endo-type glycoside degradation enzyme, to degrade polysaccharides with α-1,3-glucosidic linkage; therefore, further study is required to clarify the entire pathway in which LlGH31_u1 is involved.
      Figure thumbnail gr9
      Figure 9Gene clusters, including GH31_u1, in bacteria. Gene clusters containing genes for GH31_u1 proteins from Lactococcus lactis subsp. cremoris MG1363 (GenBank ID, CAL98407.1), Streptococcus mutans UA159 (AAN57886.1), Roseburia intestinalis L1-82 (VCV21248.1), Lachnoclostridium phytofermentans ISDg (ABX42246.1), and Bacteroides cellulosilyticus WH2 (CCP32040.1) are shown. Open reading frames are shown as arrows, and colors are listed below the gene clusters. LacI indicates the LacI-like transcription regulator.

      Conclusion

      We found that microbial GH31 enzymes displayed more strict specificity to α-1,3-glucosides than other characterized GH31 enzymes. X-ray crystallographic and cryo-EM structural analyses revealed that LlGH31_u1 forms a hexamer, and the C-terminal α-helix domain, which has not been observed in other GH31 enzymes, is involved in hexamer formation. Moreover, the residues forming subsite +1, including the Tyr99 residue on an adjacent protomer, of LlGH31_u1 are different from those of the other GH31 α-glucoside hydrolases and are important for the mechanism of the strict recognition of α-1,3-glucosidic linkages. This study suggests that LlGH31_u1 and its homologs are involved in the degradation of nigerooligosaccharides in various microbial species, and further studies are needed to clarify the entire pathway of polysaccharide degradation in which GH31_u1 enzymes are involved.

      Experimental procedures

      Chemicals and strains

      p-Nitrophenyl α-D-galactopyranoside and GlcNAc-β-1,3-GalNAcα-pNP were purchased from Tokyo Chemical Industry Co, Ltd p-Nitrophenyl α-L-fucopyranoside, pNP-α-Glc, and p-nitrophenyl α-D-mannopyranoside were obtained from Merck. p-Nitrophenyl N-acetyl-α-D-glucopyranoside, p-nitrophenyl α-L-rhamnopyranoside, p-nitrophenyl α-D-xylopyranoside, nigerose, and nigeran were purchased from Carbosynth. GalNAcα-pNP was obtained from Cayman Chemical, and nigerotriose and nigerotetraose were purchased from Megazyme. Maltose, maltotriose, and maltotetraose were obtained from Hayashibara Co E. coli DH5α and BL21 (DE3) were used for DNA manipulation and protein expression, respectively.

      Phylogenetic analysis

      The extraction of accession numbers registered on the CAZy database, download of the corresponding amino acid sequences from National Center for Biotechnology Information, and pruning of extra domains was performed using the SACCHARIS program (
      • Jones D.R.
      • Thomas D.
      • Alger N.
      • Ghavidel A.
      • Inglis G.D.
      • Abbott D.W.
      Saccharis: An automated pipeline to streamline discovery of carbohydrate active enzyme activities within polyspecific families and de novo sequence datasets.
      ). These sequences were then clustered at 70% sequence similarity using CD-HIT to remove redundancy (
      • Fu L.
      • Niu B.
      • Zhu Z.
      • Wu S.
      • Li W.
      CD-HIT: Accelerated for clustering the next-generation sequencing data.
      ,
      • Li W.
      • Godzik A.
      Cd-hit: A fast program for clustering and comparing large sets of protein or nucleotide sequences.
      ). The sequence alignment of GH31 catalytic domains was performed using the MUSCLE algorithm, and the phylogenetic tree was built via the maximum likelihood method using MEGA (
      • Kumar S.
      • Stecher G.
      • Li M.
      • Knyaz C.
      • Tamura K.
      Mega X: Molecular evolutionary genetics analysis across computing platforms.
      ). The phylogenetic tree was visualized using iTOL (
      • Letunic I.
      • Bork P.
      Interactive tree of life (iTOL) v3: An online tool for the display and annotation of phylogenetic and other trees.
      ).

      Plasmid construction and site-directed mutagenesis

      The absence of a signal peptide of LlGH31_u1 and CmGH31_u1 was evaluated using SignalP 5.0 (http://www.cbs.dtu.dk/services/SignalP/). The DNA fragment encoding LlGH31_u1 (llmg_1836, GenBank CAL98407.1) was amplified from L. lactis subsp. cremoris MG1363 using colony-directed polymerase chain reaction (PCR) with Ex Taq DNA polymerase (TaKaRa). The DNA fragment encoding CmGH31_u1 was amplified from the cDNA of C. militaris NBRC 103752 (NITE BioResource Center). The resultant DNA fragments were digested with NdeI and NotI (for LlGH31_u1) or NdeI and HindIII (for CmGH31_u1) and then ligated into a pET28a vector (Merck) digested using the same restriction enzymes. Site-directed mutagenesis was performed via inverse PCR with the desired primers using pET28a expression plasmid harboring DNA encoding LlGH31_u1 as a template. All single amino acid substitution mutants were generated according to literature (
      • Zheng L.
      • Baumann U.
      • Reymond J.L.
      An efficient one-step site-directed and site-saturation mutagenesis protocol.
      ). All primers used are listed in Table S4. All the constructed plasmids were confirmed via DNA sequencing.

      Recombinant expression and purification

      E. coli BL21 (DE3) harboring the desired plasmid was cultured at 37 °C to an absorbance (600 nm) of 0.8 in 1 L Luria–Bertani medium containing 50 μg/ml kanamycin. After cooling the medium to 20 °C on ice, protein expression was induced by adding 0.05 and 0.5 mM IPTG at 20 °C for 20 h for WT LlGH31_u1 and the other expression constructs, respectively. The cells were harvested via centrifugation for 10 min (4 °C 5000g) and stored at −20 °C. The cell pellet was resuspended in 50 mM sodium phosphate buffer (pH 8.0) containing 300 mM NaCl and 20 mM imidazole. After resuspension, the cells were disrupted via ultrasonication and then centrifuged to remove insoluble materials. Cell lysate supernatant was loaded onto a Ni-NTA Agarose (Qiagen) column, and the unbound proteins were washed with the same buffer. Proteins were eluted with the same buffer containing 250 mM imidazole and then concentrated via ultrafiltration using 30K Amicon Ultracentrifugal units. Gel filtration chromatography was performed using the ÄKTA explorer 10S system (GE Healthcare). LlGH31_u1 was applied onto Superdex 200 Increase 10/300 column for protein purification or HiLoad 16/60 Superdex 200 prep grade column for molecular weight determination and eluted by 1.5 column volumes of 20 mM sodium phosphate buffer (pH 7.0) containing 300 mM NaCl. Protein purity was confirmed using sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

      Enzyme assays

      The hydrolytic activity toward various p-nitrophenyl (pNP) glycosides was measured in 50 μl reaction mixtures containing 40 μg/ml of LlGH31_u1 or CmGH31_u1, 0.5 mM of a substrate, and 20 mM sodium phosphate buffer (pH 7.0) at 30 °C. To examine the effect of pH on hydrolytic activity, reaction mixtures containing LlGH31_u1 (30 μg/ml) or CmGH31_u1 (34 μg/ml) and 0.5 mM pNP-α-Glc were prepared with McIlvaine (sodium citrate–phosphate) buffer at pH 3.5 to 8.0 or in glycine–HCl buffer at pH 9.0 to 10. The mixtures were incubated for 5 min at 30 °C. The effect of temperature on the hydrolytic activity was examined using 50 mM sodium phosphate buffer (pH 7.0) containing 0.5 mM pNP-α-Glc. The mixtures were incubated at temperatures ranging from 25 °C to 55 °C. The pH stability was tested by incubating 500 μg/ml of LlGH31_u1 for 24 h in McIlvaine buffer at pH 3.5 to 8.0 or in glycine–HCl buffer at pH 9.0 to 10. The thermostability was tested by incubating 300 μg/ml of LlGH31_u1 at 4 °C to 45 °C for 30 min. The residual activities for pH stability and thermostability were measured using 30 μg/ml of LlGH31_u1 and 50 mM phosphate buffer (pH 7.0) containing 0.5 mM pNP-α-Glc at 30 °C. All reactions above were terminated by adding 100 μl of 1 M Na2CO3, and the amount of released pNP was quantified by measuring absorption at 405 nm.
      To examine the activity toward disaccharides, the reaction mixture containing 20 mM sodium phosphate buffer, 10 mM disaccharides (trehalose, kojibiose, nigerose, maltose, isomaltose, or sucrose), and 0.1 mg/ml purified LlGH31_u1 was incubated at 30 °C for 1 h and analyzed by TLC using Silica Gel 60 F254 TLC plates (Merck). TLC plates were developed using a solvent with 1-butanol/ethanol/water (10:5:2, vol/vol) and sprayed with 10% sulfuric acid in methanol and then baked. For the quantitative analysis of GH31_u1 enzyme hydrolytic activity against α-glucosides, the reaction was terminated by adding the same amount of 0.5 M sodium carbonate (pH 10.0), and the liberated glucose was quantified by the glucose oxidase–peroxidase method using Glucose CII-test Wako (Wako Pure Chemicals). To determine kinetic parameters, initial velocities at five or six concentrations for each substrate were measured and fitted to the Michaelis−Menten equation using Kaleida Graph software (Synergy Software). The reaction was performed using various concentrations of LlGH31_u1 (62.0 nM for nigerose and nigerotetraose, 32.0 nM for nigerotriose, 611 nM for maltose, 1.55 × 103 nM for maltotriose, 3.10 × 103 nM for maltotetraose, 310 nM for kojibiose, and 49.0 nM for pNP-Glc; all concentrations were calculated as a monomer) or CmGH31_u1 (620 nM for nigerose and nigerotriose and 124 nM for pNP-α-Glc) in 50 mM sodium phosphate buffer (pH 7.0 and pH 6.0 for LlGH31_u1 and CmGH31_u1, respectively) containing 50 mM sodium chloride. Substrate concentrations were as follows: 1 to 16 mM for pNP-α-Glc, 1 to 20 mM for nigerose, 0.5 to 10 mM for nigerose and nigerotetraose, 1 to 30 mM for kojibiose, 1 to 50 mM for maltose, and 5 to 100 mM for maltotriose and maltotetraose.

      X-ray crystallography

      The purified protein was concentrated to 15 mg/ml in 10 mM Hepes (pH 7.0) via ultrafiltration and crystallized at 20 °C using the hanging-drop vapor diffusion method, in which 1.0 μl protein solution was mixed with an equal volume of a crystallization reservoir solution. Initial crystallization screening was performed using Crystal Screen, Crystal Screen 2, PEG/Ion Screen, and PEG/Ion 2 Screen kits (Hampton Research). Crystallization trials using PEG/Ion Screen and PEG/Ion 2 Screen yielded snowflake-shaped crystals in several conditions. Additive screening was performed using an Additive Screen kit (Hampton Research), and crystal form 1 was obtained using a crystallization solution containing 16% (w/v) polyethylene glycol (PEG) 3350, 200 mM sodium citrate buffer (pH 7.0), and 3% (w/v) xylitol. The crystal form 2 of native LlGH31_u1, D394A, and selenomethionine (SeMet)-substituted LlGH31_u1 were obtained using a crystallization solution containing 13% to 7% (w/v) PEG 3350 and 400 mM sodium citrate buffer (pH 7.0). The crystal was soaked for 45 min in the reservoir solution containing 100 mM glucose to determine LlGH31_u1 in complex with glucose. To determine the structures of LlGH31_u1 in complex with the other ligands, enzymes were cocrystallized using a reservoir solution containing 10 mM ligands. Crystals were cryoprotected with the reservoir solution supplemented with ethylene glycol at a final concentration of 22% (v/v) and then flash-frozen in liquid nitrogen or flash cooled to 100 K in a nitrogen gas stream. Diffraction data were collected at the BL5A and AR-NW12A beamlines (Photon Factory).
      Data were processed using XDS (
      • Kabsch W.
      Xds.
      ). The initial phase was calculated using the single-wavelength anomalous dispersion dataset of SeMet-substituted LlGH31_u1 via the Crank2 program in the CCP4 suite (
      • Winn M.D.
      • Ballard C.C.
      • Cowtan K.D.
      • Dodson E.J.
      • Emsley P.
      • Evans P.R.
      • Keegan R.M.
      • Krissinel E.B.
      • Leslie A.G.
      • McCoy A.
      • McNicholas S.J.
      • Murshudov G.N.
      • Pannu N.S.
      • Potterton E.A.
      • Powell H.R.
      • et al.
      Overview of the CCP4 suite and current developments.
      ). The structures of WT_P21 and WT_P6322 were solved by the molecular replacement method using MOLREP (
      • Vagin A.
      • Teplyakov A.
      Molecular replacement with MOLREP.
      ) with the coordinates of the refined SeMet-LlGH31_u1 as the search models. Refinement and manual model building were performed using REFMAC5 (
      • Murshudov G.N.
      • Skubák P.
      • Lebedev A.A.
      • Pannu N.S.
      • Steiner R.A.
      • Nicholls R.A.
      • Winn M.D.
      • Long F.
      • Vagin A.A.
      REFMAC5 for the refinement of macromolecular crystal structures.
      ) and COOT (
      • Emsley P.
      • Lohkamp B.
      • Scott W.G.
      • Cowtan K.
      Features and development of coot.
      ), respectively, and the models were validated using MolProbity (
      • Williams C.J.
      • Headd J.J.
      • Moriarty N.W.
      • Prisant M.G.
      • Videau L.L.
      • Deis L.N.
      • Verma V.
      • Keedy D.A.
      • Hintze B.J.
      • Chen V.B.
      • Jain S.
      • Lewis S.M.
      • Arendall 3rd, W.B.
      • Snoeyink J.
      • Adams P.D.
      • et al.
      MolProbity: More and better reference data for improved all-atom structure validation.
      ).

      Cryo-EM sample preparation and data acquisition

      A 3 μl sample of LlGH31_u1 (1.13 μM, calculated as a hexamer) in 20 mM sodium phosphate buffer (pH 7.0) containing 300 mM NaCl was applied to a holey carbon grid (Quantifoil, Cu, R1.2/1.3, 300 mesh) rendered hydrophilic by a 30-s glow-discharge in air at 11 mA current with PIB-10. The grid was blotted for 5 s (blot force 15) at 18 °C and 100% humidity and flash-frozen in liquid ethane using Vitrobot Mark IV (FEI).
      A total of 995 micrographs were acquired by two cryo-EM sessions of Talos Arctica (FEI) microscope operating at 200 kV in the nanoprobe mode using EPU software for automated data collection. For both sessions, the movie frames were collected by 4 k × 4 k Falcon 3 direct electron detector in electron counting mode at a nominal magnification of 1,200,00 × , which yielded a pixel size of 0.88 Å/pixel. Fifty movie frames were recorded at an exposure of 1.00 electrons per Å2 per frame, corresponding to a total exposure of 50 e2. The defocus steps used were –0.8, –1.2, –1.6, and –2.0 μm.

      Cryo-EM data processing

      Movie frames were aligned, dose-weighted, and averaged using MotionCor2 (
      • Zheng S.Q.
      • Palovcak E.
      • Armache J.P.
      • Verba K.A.
      • Cheng Y.
      • Agard D.A.
      MotionCor2: Anisotropic correction of beam-induced motion for improved cryo-electron microscopy.
      ) (the version implemented in RELION3.1 (
      • Zivanov J.
      • Nakane T.
      • Forsberg B.O.
      • Kimanius D.
      • Hagen W.J.
      • Lindahl E.
      • Scheres S.H.
      New tools for automated high-resolution cryo-EM structure determination in RELION-3.
      )) on 5 × 5 tiled frames with a B-factor of 200 applied to correct for beam-induced specimen motion and account for radiation damage using an exposure-dependent filter. The micrographs whose total accumulated motions were >60 Å were discarded. The nonweighted movie sums were used for contrast transfer function (CTF) estimation with the program Gctf (
      • Zhang K.
      Gctf: Real-time CTF determination and correction.
      ) (512-pixel box size, 30 Å minimum and 4 Å maximum resolution, and 0.10 amplitude contrast), whereas the dose-weighted sums were used for all subsequent steps of image processing. First, the images whose CTF max resolutions were better than 5.5 Å were selected. The particles were collected using SPHIRE crYOLO with a generalized model (
      • Wagner T.
      • Merino F.
      • Stabrin M.
      • Moriya T.
      • Antoni C.
      • Apelbaum A.
      • Hagel P.
      • Sitsel O.
      • Raisch T.
      • Prumbaum D.
      • Quentin D.
      • Roderer D.
      • Tacke S.
      • Siebolds B.
      • Schubert E.
      • et al.
      SPHIRE-crYOLO is a fast and accurate fully automated particle picker for cryo-EM.
      ,
      • Moriya T.
      • Saur M.
      • Stabrin M.
      • Merino F.
      • Voicu H.
      • Huang Z.
      • Penczek P.A.
      • Raunser S.
      • Gatsogiannis C.
      High-resolution single particle analysis from electron cryo-microscopy images using SPHIRE.
      ) using a selection threshold of 0.1. The subsequent processes of 2D classification, ab initio reconstruction, 3D classification, 3D refinement, CTF refinement, Bayesian polishing, and local resolution estimation were conducted using RELION3.1. The details of the processes are described in the Supporting information (Supplementary Methods, Figs. S9, and S10).
      For calculating the global resolution estimation after each 3D refinement, the gold standard FSC resolution with 0.143 criterion (
      • Rosenthal P.B.
      • Henderson R.
      Optimal determination of particle orientation, absolute hand, and contrast loss in single-particle electron cryomicroscopy.
      ) was used, including the phase randomization to account for the possible artifactual resolution enhancement caused by solvent mask (
      • Chen S.
      • McMullan G.
      • Faruqi A.R.
      • Murshudov G.N.
      • Short J.M.
      • Scheres S.H.
      • Henderson R.
      High-resolution noise substitution to measure overfitting and validate resolution in 3D structure determination by single particle electron cryomicroscopy.
      ). The model-to-map FSC resolution with 0.5 criterion was calculated using phenix.mtriage (
      • Afonine P.V.
      • Klaholz B.P.
      • Moriarty N.W.
      • Poon B.K.
      • Sobolev O.V.
      • Terwilliger T.C.
      • Adams P.D.
      • Urzhumtsev A.
      New tools for the analysis and validation of cryo-EM maps and atomic models.
      ). To visualize the output 2D/3D images, UCSF Chimera and e2display.py of EMAN2 (
      • Tang G.
      • Peng L.
      • Baldwin P.R.
      • Mann D.S.
      • Jiang W.
      • Rees I.
      • Ludtke S.J.
      EMAN2: An extensible image processing suite for electron microscopy.
      ) were used. The ctflimit function (
      • Penczek P.A.
      • Fang J.
      • Li X.
      • Cheng Y.
      • Loerke J.
      • Spahn C.M.
      CTER-rapid estimation of CTF parameters with error assessment.
      ) implemented in SPARX/SPHIRE (
      • Moriya T.
      • Saur M.
      • Stabrin M.
      • Merino F.
      • Voicu H.
      • Huang Z.
      • Penczek P.A.
      • Raunser S.
      • Gatsogiannis C.
      High-resolution single particle analysis from electron cryo-microscopy images using SPHIRE.
      ,
      • Hohn M.
      • Tang G.
      • Goodyear G.
      • Baldwin P.R.
      • Huang Z.
      • Penczek P.A.
      • Yang C.
      • Glaeser R.M.
      • Adams P.D.
      • Ludtke S.J.
      SPARX, a new environment for Cryo-EM image processing.
      ) was used to calculate the smallest box size that ensures no CTF aliasing in the reciprocal space up to an expected resolution for the maximum defocus value of the dataset.

      Data availability

      The nucleotide sequence of CmGH31_u1 was submitted to the DDBJ/EMBL/GenBank databases under the accession number LC660181. The atomic coordinates and structure factors of WT_P21, WT_P6322, WT-Glc, D394A-Nig2, D394A-Nig3, D394A-Nig4, and D394A-Koj2 have been deposited in the Worldwide Protein Data Bank (wwPDB, http://wwpdb.org/) under the accession codes 7WJ9, 7WJA, 7WJB, 7WJC, 7WJD, 7WJE, and 7WJF, respectively. The cryo-EM map and the atomic coordinate of LlGH31_u1 determined using cryo-EM have been deposited in the Electron Microscopy Data Bank (https://www.ebi.ac.uk/pdbe/emdb/) and wwPDB, respectively, with the accession codes EMD-32571 and 7WLG.

      Supporting information

      This article contains supporting information (
      • Kang M.S.
      • Okuyama M.
      • Mori H.
      • Kimura A.
      The first α-1,3-glucosidase from bacterial origin belonging to glycoside hydrolase family 31.
      ,
      • Barrett K.
      • Lange L.
      Peptide-based functional annotation of carbohydrate-active enzymes by conserved unique peptide patterns (CUPP).
      ,
      • Nihira T.
      • Nakai H.
      • Chiku K.
      • Kitaoka M.
      Discovery of nigerose phosphorylase from Clostridium phytofermentans.
      ,
      • Okuyama M.
      • Miyamoto M.
      • Matsuo I.
      • Iwamoto S.
      • Serizawa R.
      • Tanuma M.
      • Ma M.
      • Klahan P.
      • Kumagai Y.
      • Tagami T.
      • Kimura A.
      Substrate recognition of the catalytic α-subunit of glucosidase II from Schizosaccharomyces pombe.
      ,
      • Kumar S.
      • Stecher G.
      • Li M.
      • Knyaz C.
      • Tamura K.
      Mega X: Molecular evolutionary genetics analysis across computing platforms.
      ,
      • Letunic I.
      • Bork P.
      Interactive tree of life (iTOL) v3: An online tool for the display and annotation of phylogenetic and other trees.
      ,
      • Zivanov J.
      • Nakane T.
      • Scheres S.H.W.
      Estimation of high-order aberrations and anisotropic magnification from cryo-EM data sets in RELION-3.1.
      ,
      • Zivanov J.
      • Nakane T.
      • Scheres S.H.W.
      A Bayesian approach to beam-induced motion correction in cryo-EM single-particle analysis.
      ,
      • Kita A.
      • Matsui H.
      • Somoto A.
      • Kimura A.
      • Takata M.
      • Chiba S.
      Substrate specificity and subsite affinities of crystalline α-glucosidase from Aspergillus niger.
      ,
      • Matsui H.
      • Sasaki M.
      • Takemasa E.
      • Kaneta T.
      • Chiba S.
      Kinetic studies on the substrate specificity and active site of rabbit muscle acid α-glucosidase.
      ,
      • Lee B.H.
      • Rose D.R.
      • Lin A.H.
      • Quezada-Calvillo R.
      • Nichols B.L.
      • Hamaker B.R.
      Contribution of the individual small intestinal α-glucosidases to digestion of unusual α-linked glycemic disaccharides.
      ,
      • Landau M.
      • Mayrose I.
      • Rosenberg Y.
      • Glaser F.
      • Martz E.
      • Pupko T.
      • Ben-Tal N.
      ConSurf 2005: The projection of evolutionary conservation scores of residues on protein structures.
      ,
      • Celniker G.
      • Nimrod G.
      • Ashkenazy H.
      • Glaser F.
      • Martz E.
      • Mayrose I.
      • Pupko T.
      • Ben-Tal N.
      ConSurf: Using evolutionary data to raise testable hypotheses about protein function.
      ,
      • Ashkenazy H.
      • Erez E.
      • Martz E.
      • Pupko T.
      • Ben-Tal N.
      ConSurf 2010: Calculating evolutionary conservation in sequence and structure of proteins and nucleic acids.
      ,
      • Tan Y.Z.
      • Baldwin P.R.
      • Davis J.H.
      • Williamson J.R.
      • Potter C.S.
      • Carragher B.
      • Lyumkis D.
      Addressing preferred specimen orientation in single-particle cryo-EM through tilting.
      ).

      Conflict of interest

      The authors have no conflicts of interest to declare.

      Acknowledgments

      We thank the staff of the Photon Factory for their assistance with X-ray data collection as well as the staff of the cryo-EM facility in KEK for their assistance with cryo-EM data collection. We also thank Dr Ahmad Suparmin for preparing the cDNA of C. militaris, Enago (www.enago.jp) for the English language review and Prof. Toshiya Senda for managing cryo-EM data collections. The X-ray crystallographic study was approved by the Photon Factory Program Advisory Committee (proposals 2019G097 and 2021G013). This research was supported by Platform Project for Supporting Drug Discovery and Life Science Research [Basis of Supporting Innovative Drug Discovery and Life Science Research (BINDS)] from AMED under Grant Number JP21am0101071 (support number 2932).

      Author contributions

      T. Miyazaki conceptualization; M. I., T. Moriya, N. A., M. K., E. Y. P., and T. Miyazaki formal analysis; M. I., T. Moriya, N. A., M. K., and T. Miyazaki investigation; T. Moriya, N. A., M. K., and T. Miyazaki methodology; T. Miyazaki project administration; T. Miyazaki supervision; T. Miyazaki funding acquisition; M. I., T. Moriya, and T. Miyazaki visualization; M. I., T. Moriya, N. A., and T. Miyazaki writing-original draft; M. I., T. Moriya, N. A., E. Y. P. and T. Miyazaki writing-review & editing.

      Funding and additional information

      This work was supported by the Japan Society for the Promotion of Science KAKENHI (Grant No. 19K15748) to T. Miyazaki.

      Supporting information

      References

        • Lombard V.
        • Golaconda Ramulu H.
        • Drula E.
        • Coutinho P.M.
        • Henrissat B.
        The carbohydrate-active enzymes database (CAZy) in 2013.
        Nucleic Acids Res. 2014; 42: D490-D495
        • Drula E.
        • Garron M.L.
        • Dogan S.
        • Lombard V.
        • Henrissat B.
        • Terrapon N.
        The carbohydrate-active enzyme database: Functions and literature.
        Nucleic Acids Res. 2021; 50 (in press): D571-D577
        • Light S.H.
        • Cahoon L.A.
        • Mahasenan K.V.
        • Lee M.
        • Boggess B.
        • Halavaty A.S.
        • Mobashery S.
        • Freitag N.E.
        • Anderson W.F.
        Transferase versus hydrolase: The role of conformational flexibility in reaction specificity.
        Structure. 2017; 25: 295-304
        • Tagami T.
        • Miyano E.
        • Sadahiro J.
        • Okuyama M.
        • Iwasaki T.
        • Kimura A.
        Two novel glycoside hydrolases responsible for the catabolism of cyclobis-(1→6)-α-nigerosyl.
        J. Biol. Chem. 2016; 291: 16438-16447
        • Garron M.L.
        • Henrissat B.
        The continuing expansion of CAZymes and their families.
        Curr. Opin. Chem. Biol. 2019; 53: 82-87
        • Helbert W.
        • Poulet L.
        • Drouillard S.
        • Mathieu S.
        • Loiodice M.
        • Couturier M.
        • Lombard V.
        • Terrapon N.
        • Turchetto J.
        • Vincentelli R.
        • Henrissat B.
        Discovery of novel carbohydrate-active enzymes through the rational exploration of the protein sequences space.
        Proc. Natl. Acad. Sci. U. S. A. 2019; 116: 6063-6068
        • Okuyama M.
        • Saburi W.
        • Mori H.
        • Kimura A.
        α-Glucosidases and α-1,4-glucan lyases: Structures, functions, and physiological actions.
        Cell Mol. Life Sci. 2016; 73: 2727-2751
        • Kang M.S.
        • Okuyama M.
        • Mori H.
        • Kimura A.
        The first α-1,3-glucosidase from bacterial origin belonging to glycoside hydrolase family 31.
        Biochimie. 2009; 91: 1434-1442
        • Caputo A.T.
        • Alonzi D.S.
        • Marti L.
        • Reca I.B.
        • Kiappes J.L.
        • Struwe W.B.
        • Cross A.
        • Basu S.
        • Lowe E.D.
        • Darlot B.
        • Santino A.
        • Roversi P.
        • Zitzmann N.
        Structures of mammalian ER α-glucosidase II capture the binding modes of broad-spectrum iminosugar antivirals.
        Proc. Natl. Acad. Sci. U. S. A. 2016; 113: E4630-E4638
        • Satoh T.
        • Toshimori T.
        • Yan G.
        • Yamaguchi T.
        • Kato K.
        Structural basis for two-step glucose trimming by glucosidase II involved in ER glycoprotein quality control.
        Sci. Rep. 2016; 6: 20575
        • Lovering A.L.
        • Lee S.S.
        • Kim Y.W.
        • Withers S.G.
        • Strynadka N.C.
        Mechanistic and structural analysis of a family 31 α-glycosidase and its glycosyl-enzyme intermediate.
        J. Biol. Chem. 2005; 280: 2105-2115
        • Miyazaki T.
        • Ishizaki Y.
        • Ichikawa M.
        • Nishikawa A.
        • Tonozuka T.
        Structural and biochemical characterization of novel bacterial α-galactosidases belonging to glycoside hydrolase family 31.
        Biochem. J. 2015; 469: 145-158
        • Speciale G.
        • Jin Y.
        • Davies G.J.
        • Williams S.J.
        • Goddard-Borger E.D.
        YihQ is a sulfoquinovosidase that cleaves sulfoquinovosyl diacylglyceride sulfolipids.
        Nat. Chem. Biol. 2016; 12: 215-217
        • Rahfeld P.
        • Wardman J.F.
        • Mehr K.
        • Huff D.
        • Morgan-Lang C.
        • Chen H.M.
        • Hallam S.J.
        • Withers S.G.
        Prospecting for microbial α-N-acetylgalactosaminidases yields a new class of GH31 O-glycanase.
        J. Biol. Chem. 2019; 294: 16400-16415
        • Miyazaki T.
        • Park E.Y.
        Crystal structure of the Enterococcus faecalis α-N-acetylgalactosaminidase, a member of the glycoside hydrolase family 31.
        FEBS Lett. 2020; 594: 2282-2293
        • Ikegaya M.
        • Miyazaki T.
        • Park E.Y.
        Biochemical characterization of Bombyx mori α-N-acetylgalactosaminidase belonging to the glycoside hydrolase family 31.
        Insect Mol. Biol. 2021; 30: 367-378
        • Rozeboom H.J.
        • Yu S.
        • Madrid S.
        • Kalk K.H.
        • Zhang R.
        • Dijkstra B.W.
        Crystal structure of α-1,4-glucan lyase, a unique glycoside hydrolase family member with a novel catalytic mechanism.
        J. Biol. Chem. 2013; 288: 26764-26774
        • Sim L.
        • Quezada-Calvillo R.
        • Sterchi E.E.
        • Nichols B.L.
        • Rose D.R.
        Human intestinal maltase-glucoamylase: Crystal structure of the N-terminal catalytic subunit and basis of inhibition and substrate specificity.
        J. Mol. Biol. 2008; 375: 782-792
        • Ren L.
        • Qin X.
        • Cao X.
        • Wang L.
        • Bai F.
        • Bai G.
        • Shen Y.
        Structural insight into substrate specificity of human intestinal maltase-glucoamylase.
        Protein Cell. 2011; 2: 827-836
        • Gray G.M.
        • Lally B.C.
        • Conklin K.A.
        Action of intestinal sucrase-isomaltase and its free monomers on an α-limit dextrin.
        J. Biol. Chem. 1979; 254: 6038-6043
        • Kishnani P.S.
        • Howell R.R.
        Pompe disease in infants and children.
        J. Pediatr. 2004; 144: S35-S43
        • Ma M.
        • Okuyama M.
        • Tagami T.
        • Kikuchi A.
        • Klahan P.
        • Kimura A.
        Novel α-1,3/α-1,4-glucosidase from Aspergillus niger exhibits unique transglucosylation to generate high levels of nigerose and kojibiose.
        J. Agric. Food Chem. 2019; 67: 3380-3388
        • Ma M.
        • Okuyama M.
        • Sato M.
        • Tagami T.
        • Klahan P.
        • Kumagai Y.
        • Mori H.
        • Kimura A.
        Effects of mutation of Asn694 in Aspergillus Niger α-glucosidase on hydrolysis and transglucosylation.
        Appl. Microbiol. Biotechnol. 2017; 101: 6399-6408
        • Tsutsumi K.
        • Gozu Y.
        • Nishikawa A.
        • Tonozuka T.
        Structural insights into polysaccharide recognition by Flavobacterium johnsoniae dextranase, a member of glycoside hydrolase family 31.
        FEBJ J. 2020; 287: 1195-1207
        • Gozu Y.
        • Ishizaki Y.
        • Hosoyama Y.
        • Miyazaki T.
        • Nishikawa A.
        • Tonozuka T.
        A glycoside hydrolase family 31 dextranase with high transglucosylation activity from Flavobacterium johnsoniae.
        Biosci. Biotechnol. Biochem. 2016; 80: 1562-1567
        • Larsbrink J.
        • Izumi A.
        • Hemsworth G.R.
        • Davies G.J.
        • Brumer H.
        Structural enzymology of Cellvibrio japonicus Agd31B protein reveals α-transglucosylase activity in glycoside hydrolase family 31.
        J. Biol. Chem. 2012; 287: 43288-43299
        • Okuyama M.
        Function and structure studies of GH family 31 and 97 α-glycosidases.
        Biosci. Biotechnol. Biochem. 2011; 75: 2269-2277
        • Jones D.R.
        • Thomas D.
        • Alger N.
        • Ghavidel A.
        • Inglis G.D.
        • Abbott D.W.
        Saccharis: An automated pipeline to streamline discovery of carbohydrate active enzyme activities within polyspecific families and de novo sequence datasets.
        Biotechnol. Biofuels. 2018; 11: 27
        • Xu J.
        • Zhang H.
        • Zheng J.
        • Dovoedo P.
        • Yin Y.
        eCAMI: simultaneous classification and motif identification for enzyme annotation.
        Bioinformatics. 2020; 36: 2068-2075
        • Barrett K.
        • Lange L.
        Peptide-based functional annotation of carbohydrate-active enzymes by conserved unique peptide patterns (CUPP).
        Biotechnol. Biofuels. 2019; 12: 102
        • Chang S.T.
        • Parker K.N.
        • Bauer M.W.
        • Kelly R.M.
        α-Glucosidase from Pyrococcus furiosus.
        Methods Enzymol. 2001; 330: 260-269
        • Lodge J.A.
        • Maier T.
        • Liebl W.
        • Hoffmann V.
        • Sträter N.
        Crystal structure of Thermotoga maritima α-glucosidase AglA defines a new clan of NAD+-dependent glycosidases.
        J. Biol. Chem. 2003; 278: 19151-19158
        • Kurakata Y.
        • Uechi A.
        • Yoshida H.
        • Kamitori S.
        • Sakano Y.
        • Nishikawa A.
        • Tonozuka T.
        Structural insights into the substrate specificity and function of Escherichia coli K12 YgjK, a glucosidase belonging to the glycoside hydrolase family 63.
        J. Mol. Biol. 2008; 381: 116-128
        • Miyazaki T.
        • Ichikawa M.
        • Yokoi G.
        • Kitaoka M.
        • Mori H.
        • Kitano Y.
        • Nishikawa A.
        • Tonozuka T.
        Structure of a bacterial glycoside hydrolase family 63 enzyme in complex with its glycosynthase product, and insights into the substrate specificity.
        FEBJ J. 2013; 280: 4560-4571
        • Miyazaki T.
        • Nishikawa A.
        • Tonozuka T.
        Crystal structure of the enzyme-product complex reveals sugar ring distortion during catalysis by family 63 inverting α-glycosidase.
        J. Struct. Biol. 2016; 196: 479-486
        • Nihira T.
        • Nakai H.
        • Chiku K.
        • Kitaoka M.
        Discovery of nigerose phosphorylase from Clostridium phytofermentans.
        Appl. Microbiol. Biotechnol. 2012; 93: 1513-1522
        • Okuyama M.
        • Miyamoto M.
        • Matsuo I.
        • Iwamoto S.
        • Serizawa R.
        • Tanuma M.
        • Ma M.
        • Klahan P.
        • Kumagai Y.
        • Tagami T.
        • Kimura A.
        Substrate recognition of the catalytic α-subunit of glucosidase II from Schizosaccharomyces pombe.
        Biosci. Biotechnol. Biochem. 2017; 81: 1503-1511
        • Ernst H.A.
        • Lo Leggio L.
        • Willemoës M.
        • Leonard G.
        • Blum P.
        • Larsen S.
        Structure of the Sulfolobus solfataricus α-glucosidase: Implications for domain conservation and substrate recognition in GH31.
        J. Mol. Biol. 2006; 358: 1106-1124
        • Holm L.
        • Rosenström P.
        Dali server: Conservation mapping in 3D.
        Nucleic Acids Res. 2010; 38: W545-W549
        • Fisher O.S.
        • Zhang R.
        • Li X.
        • Murphy J.W.
        • Demeler B.
        • Boggon T.J.
        Structural studies of cerebral cavernous malformations 2 (CCM2) reveal a folded helical domain at its C-terminus.
        FEBS Lett. 2013; 587: 272-277
        • Davies G.J.
        • Wilson K.S.
        • Henrissat B.
        Nomenclature for sugar-binding subsites in glycosyl hydrolases.
        Biochem. J. 1997; 321: 557-559
        • Higgins M.A.
        • Tegl G.
        • MacDonald S.S.
        • Arnal G.
        • Brumer H.
        • Withers S.G.
        • Ryan K.S.
        N-glycan degradation pathways in gut- and soil-dwelling actinobacteria share common core genes.
        ACS Chem. Biol. 2021; 16: 701-711
        • Bowen W.H.
        • Koo H.
        Biology of Streptococcus mutans-derived glucosyltransferases: Role in extracellular matrix formation of cariogenic biofilms.
        Caries Res. 2011; 45: 69-86
        • Bounaix M.S.
        • Gabriel V.
        • Morel S.
        • Robert H.
        • Rabier P.
        • Remaud-Siméon M.
        • Gabriel B.
        • Fontagné-Faucher C.
        Biodiversity of exopolysaccharides produced from sucrose by sourdough lactic acid bacteria.
        J. Agric. Food Chem. 2009; 57: 10889-10897
        • Côté G.L.
        • Skory C.D.
        Cloning, expression, and characterization of an insoluble glucan-producing glucansucrase from Leuconostoc mesenteroides NRRL B-1118.
        Appl. Microbiol. Biotechnol. 2012; 93: 2387-2394
        • Malik A.
        • Radji M.
        • Kralj S.
        • Dijkhuizen L.
        Screening of lactic acid bacteria from Indonesia reveals glucansucrase and fructansucrase genes in two different Weissella confusa strains from soya.
        FEMS Microbiol. Lett. 2009; 300: 131-138
        • Yoshimi A.
        • Miyazawa K.
        • Abe K.
        Cell wall structure and biogenesis in Aspergillus species.
        Biosci. Biotechnol. Biochem. 2016; 80: 1700-1711
        • Deutscher J.
        • Francke C.
        • Postma P.W.
        How phosphotransferase system-related protein phosphorylation regulates carbohydrate metabolism in bacteria.
        Microbiol. Mol. Biol. Rev. 2006; 70: 939-1031
        • Koropatkin N.M.
        • Cameron E.A.
        • Martens E.C.
        How glycan metabolism shapes the human gut microbiota.
        Nat. Rev. Microbiol. 2012; 10: 323-335
        • Siezen R.J.
        • Bayjanov J.R.
        • Felis G.E.
        • van der Sijde M.R.
        • Starrenburg M.
        • Molenaar D.
        • Wels M.
        • van Hijum S.A.
        • van Hylckama Vlieg J.E.
        Genome-scale diversity and niche adaptation analysis of Lactococcus lactis by comparative genome hybridization using multi-strain arrays.
        Microb. Biotechnol. 2011; 4: 383-402
        • Fu L.
        • Niu B.
        • Zhu Z.
        • Wu S.
        • Li W.
        CD-HIT: Accelerated for clustering the next-generation sequencing data.
        Bioinformatics. 2012; 28: 3150-3152
        • Li W.
        • Godzik A.
        Cd-hit: A fast program for clustering and comparing large sets of protein or nucleotide sequences.
        Bioinformatics. 2006; 22: 1658-1659
        • Kumar S.
        • Stecher G.
        • Li M.
        • Knyaz C.
        • Tamura K.
        Mega X: Molecular evolutionary genetics analysis across computing platforms.
        Mol. Biol. Evol. 2018; 35: 1547-1549
        • Letunic I.
        • Bork P.
        Interactive tree of life (iTOL) v3: An online tool for the display and annotation of phylogenetic and other trees.
        Nucleic Acids Res. 2016; 44: W242-W245
        • Zheng L.
        • Baumann U.
        • Reymond J.L.
        An efficient one-step site-directed and site-saturation mutagenesis protocol.
        Nucleic Acids Res. 2004; 32: e115
        • Kabsch W.
        Xds.
        Acta Crystallogr. D Biol. Crystallogr. 2010; 66: 125-132
        • Winn M.D.
        • Ballard C.C.
        • Cowtan K.D.
        • Dodson E.J.
        • Emsley P.
        • Evans P.R.
        • Keegan R.M.
        • Krissinel E.B.
        • Leslie A.G.
        • McCoy A.
        • McNicholas S.J.
        • Murshudov G.N.
        • Pannu N.S.
        • Potterton E.A.
        • Powell H.R.
        • et al.
        Overview of the CCP4 suite and current developments.
        Acta Crystallogr. D Biol. Crystallogr. 2011; 67: 235-242
        • Vagin A.
        • Teplyakov A.
        Molecular replacement with MOLREP.
        Acta Crystallogr. D Biol. Crystallogr. 2010; 66: 22-25
        • Murshudov G.N.
        • Skubák P.
        • Lebedev A.A.
        • Pannu N.S.
        • Steiner R.A.
        • Nicholls R.A.
        • Winn M.D.
        • Long F.
        • Vagin A.A.
        REFMAC5 for the refinement of macromolecular crystal structures.
        Acta Crystallogr. D Biol. Crystallogr. 2011; 67: 355-367
        • Emsley P.
        • Lohkamp B.
        • Scott W.G.
        • Cowtan K.
        Features and development of coot.
        Acta Crystallogr. D Biol. Crystallogr. 2010; 66: 486-501
        • Williams C.J.
        • Headd J.J.
        • Moriarty N.W.
        • Prisant M.G.
        • Videau L.L.
        • Deis L.N.
        • Verma V.
        • Keedy D.A.
        • Hintze B.J.
        • Chen V.B.
        • Jain S.
        • Lewis S.M.
        • Arendall 3rd, W.B.
        • Snoeyink J.
        • Adams P.D.
        • et al.
        MolProbity: More and better reference data for improved all-atom structure validation.
        Protein Sci. 2018; 27: 293-315
        • Zheng S.Q.
        • Palovcak E.
        • Armache J.P.
        • Verba K.A.
        • Cheng Y.
        • Agard D.A.
        MotionCor2: Anisotropic correction of beam-induced motion for improved cryo-electron microscopy.
        Nat. Methods. 2017; 14: 331-332
        • Zivanov J.
        • Nakane T.
        • Forsberg B.O.
        • Kimanius D.
        • Hagen W.J.
        • Lindahl E.
        • Scheres S.H.
        New tools for automated high-resolution cryo-EM structure determination in RELION-3.
        ELife. 2018; 7e42166
        • Zhang K.
        Gctf: Real-time CTF determination and correction.
        J. Struct. Biol. 2016; 193: 1-12
        • Wagner T.
        • Merino F.
        • Stabrin M.
        • Moriya T.
        • Antoni C.
        • Apelbaum A.
        • Hagel P.
        • Sitsel O.
        • Raisch T.
        • Prumbaum D.
        • Quentin D.
        • Roderer D.
        • Tacke S.
        • Siebolds B.
        • Schubert E.
        • et al.
        SPHIRE-crYOLO is a fast and accurate fully automated particle picker for cryo-EM.
        Commun. Biol. 2019; 2: 218
        • Moriya T.
        • Saur M.
        • Stabrin M.
        • Merino F.
        • Voicu H.
        • Huang Z.
        • Penczek P.A.
        • Raunser S.
        • Gatsogiannis C.
        High-resolution single particle analysis from electron cryo-microscopy images using SPHIRE.
        J. Vis. Exp. 2017; https://doi.org/10.3791/55448
        • Rosenthal P.B.
        • Henderson R.
        Optimal determination of particle orientation, absolute hand, and contrast loss in single-particle electron cryomicroscopy.
        J. Mol. Biol. 2003; 333: 721-745
        • Chen S.
        • McMullan G.
        • Faruqi A.R.
        • Murshudov G.N.
        • Short J.M.
        • Scheres S.H.
        • Henderson R.
        High-resolution noise substitution to measure overfitting and validate resolution in 3D structure determination by single particle electron cryomicroscopy.
        Ultramicroscopy. 2013; 135: 24-35
        • Afonine P.V.
        • Klaholz B.P.
        • Moriarty N.W.
        • Poon B.K.
        • Sobolev O.V.
        • Terwilliger T.C.
        • Adams P.D.
        • Urzhumtsev A.
        New tools for the analysis and validation of cryo-EM maps and atomic models.
        Acta Crystallogr. D Struct. Biol. 2018; 74: 814-840
        • Tang G.
        • Peng L.
        • Baldwin P.R.
        • Mann D.S.
        • Jiang W.
        • Rees I.
        • Ludtke S.J.
        EMAN2: An extensible image processing suite for electron microscopy.
        J. Struct. Biol. 2007; 157: 38-46
        • Penczek P.A.
        • Fang J.
        • Li X.
        • Cheng Y.
        • Loerke J.
        • Spahn C.M.
        CTER-rapid estimation of CTF parameters with error assessment.
        Ultramicroscopy. 2014; 140: 9-19
        • Hohn M.
        • Tang G.
        • Goodyear G.
        • Baldwin P.R.
        • Huang Z.
        • Penczek P.A.
        • Yang C.
        • Glaeser R.M.
        • Adams P.D.
        • Ludtke S.J.
        SPARX, a new environment for Cryo-EM image processing.
        J. Struct. Biol. 2007; 157: 47-55
        • Zivanov J.
        • Nakane T.
        • Scheres S.H.W.
        Estimation of high-order aberrations and anisotropic magnification from cryo-EM data sets in RELION-3.1.
        IUCrJ. 2020; 7: 253-267
        • Zivanov J.
        • Nakane T.
        • Scheres S.H.W.
        A Bayesian approach to beam-induced motion correction in cryo-EM single-particle analysis.
        IUCrJ. 2019; 6: 5-17
        • Kita A.
        • Matsui H.
        • Somoto A.
        • Kimura A.
        • Takata M.
        • Chiba S.
        Substrate specificity and subsite affinities of crystalline α-glucosidase from Aspergillus niger.
        Agric. Biol. Chem. 1991; 55: 2327-2335
        • Matsui H.
        • Sasaki M.
        • Takemasa E.
        • Kaneta T.
        • Chiba S.
        Kinetic studies on the substrate specificity and active site of rabbit muscle acid α-glucosidase.
        J. Biochem. 1984; 96: 993-1004
        • Lee B.H.
        • Rose D.R.
        • Lin A.H.
        • Quezada-Calvillo R.
        • Nichols B.L.
        • Hamaker B.R.
        Contribution of the individual small intestinal α-glucosidases to digestion of unusual α-linked glycemic disaccharides.
        J. Agric. Food Chem. 2016; 64: 6487-6494
        • Landau M.
        • Mayrose I.
        • Rosenberg Y.
        • Glaser F.
        • Martz E.
        • Pupko T.
        • Ben-Tal N.
        ConSurf 2005: The projection of evolutionary conservation scores of residues on protein structures.
        Nucleic Acids Res. 2005; 33: W299-W302
        • Celniker G.
        • Nimrod G.
        • Ashkenazy H.
        • Glaser F.
        • Martz E.
        • Mayrose I.
        • Pupko T.
        • Ben-Tal N.
        ConSurf: Using evolutionary data to raise testable hypotheses about protein function.
        Isr. J. Chem. 2013; 53: 199-206
        • Ashkenazy H.
        • Erez E.
        • Martz E.
        • Pupko T.
        • Ben-Tal N.
        ConSurf 2010: Calculating evolutionary conservation in sequence and structure of proteins and nucleic acids.
        Nucleic Acids Res. 2010; 38: W529-W533
        • Tan Y.Z.
        • Baldwin P.R.
        • Davis J.H.
        • Williamson J.R.
        • Potter C.S.
        • Carragher B.
        • Lyumkis D.
        Addressing preferred specimen orientation in single-particle cryo-EM through tilting.
        Nat. Methods. 2017; 14: 793-796