A. Kimura, Research Faculty of Agriculture, Hokkaido University, Kita-9 Nishi-9, Kita-ku, Sapporo 060-8589, Japan Fax: +81 11 706 2808 Tel: +81 11 706 2808 E-mail: firstname.lastname@example.org
Bacteroides thetaiotaomicron VPI-5482 harbors a gene encoding a putative cycloisomaltooligosaccharide glucanotransferase (BT3087) belonging to glycoside hydrolase family 66. The goal of the present study was to characterize the catalytic properties of this enzyme. Therefore, we expressed BT3087 (recombinant endo-dextranase from Bacteroides thetaiotaomicron VPI-5482) in Escherichia coli and determined that recombinant endo-dextranase from Bacteroides thetaiotaomicron VPI-5482 preferentially synthesized isomaltotetraose and isomaltooligosaccharides (degree of polymerization > 4) from dextran. The enzyme also generated large cyclic isomaltooligosaccharides early in the reaction. We conclude that members of the glycoside hydrolase 66 family may be classified into three types: (a) endo-dextranases, (b) dextranases possessing weak cycloisomaltooligosaccharide glucanotransferase activity, and (c) cycloisomaltooligosaccharide glucanotransferases.
endo-dextranase from Bacteroides thetaiotaomicron VPI-5482
carbohydrate-binding module 35
degree of polymerization
fast-atom bombardment/mass spectrometry
glycoside hydrolase family
Dex from Paenibacillus sp.
Endo-dextranases (Dex; EC 126.96.36.199) randomly hydrolyze the α-1,6-linkages of dextran. Endo-dextranases are classified into the glycoside hydrolase (GH) families 49 and 66 on the basis of amino acid sequence similarity . Cycloisomaltooligosaccharide (CI) glucanotransferases (CITases; EC 188.8.131.52)  are also classified into GH66 and catalyze conversion of dextran to CIs with degrees of polymerization of 7–14 (i.e. CI-n, where n is the degree of polymerization) by intramolecular transglycosylation [3,4]. The 3D structure of a GH66 Dex [5,6] indicated that the amino acid sequences of this family of enzymes comprise seven regions [7,8] (Fig. 1): (a) an N-terminal variable region, (b) an N-terminal carbohydrate-binding region, (c) conserved region I, (d) a CITase-specific region, (e) conserved region II, (f) a C-terminal carbohydrate-binding region, and (g) a C-terminal variable region. However, most GH66 endo-dextranases except Dex from Paenibacillus sp. (PsDex) lack the CITase-specific region that is responsible for the generation of CIs [7,9]. The 3D structure of Dex from Streptococcus mutans UA159 reveals that the conserved regions I and II of GH66 members form a catalytic (β/α)8 barrel that includes two catalytic amino acids [5,6]. The two catalytic residues of PsDex (Asp189 and Glu412) serve as the nucleophile and proton donor, respectively, in chemical rescue experiments .
PsDex exhibits strong dextranolytic activity with weak cyclization activity, generating CI-7 to CI-14 . We therefore propose that GH66 enzymes may be categorized into three types: (a) type I, pure Dex lacking detectable cyclization activity, (b) type II, mainly Dex with weak CITase activity, and (c) type III, CITase cyclization activity with little dextranolytic activity. However, to date, there is only one example of a type II enzyme , which encouraged us to identify more type II Dex.
The Bacteroides thetaiotaomicron VPI-5482 genome  harbors a putative CITase gene (BT3087), carrying a 1779 bp open reading frame encoding 592 amino acids. However, the function of this protein is unknown. In the present study, we cloned the putative CITase gene, and expressed its product in Escherichiacoli. Characterization of its properties showed that the recombinant enzyme (rBtDex) was a typical Dex. Furthermore, we focused on its dextranolytic activity, together with CI formation ability, and found that rBtDex possessed both CITase and Dex activities. To the best of our knowledge, BtDex is the second example of a type II Dex member of GH66.
Results and Discussion
The BtDex amino acid sequence exhibited 30–44% identity with five thermostable Paenibacillus sp. endo-dextranases [11,12], 29% identity with Bacilluscirculans T-3040 CITase , 24% identity with PsDex , and 24% with Streptococcus sp. endo-dextranases [13–15] (Fig. 1). BtDex is the shortest and lacks a C-terminal variable region. Similar to Streptococcus and the thermostable Paenibacillus endo-dextranases, BtDex also lacked a CITase-specific region that contributes to CI formation. BtDex possesses two catalytic amino acid residues (Asp297 in conserved region I and Glu360 in conserved region II), which correspond to Asp340 or Glu412 of Paenibacillus sp. Dex , respectively. Computer-based analyses using BLAST and SignalP suggested that BtDex is a putative extracellular CITase. The signal peptide is predicted to be the N-terminal 23 amino acids. However, there is a possibility that BtDex is a Dex, due to the lack of CITase-specific region in its sequence.
Production and characterization of rBtDex
We purified rBtDex using a four-step procedure, and determined its biochemical properties. The N-terminal amino acid sequence (PQNGGAS) and ESI-MS (63 964 Da) indicated that rBtDex comprises the sequence Pro24–Glu592, lacking the 23 amino acid residues as a signal peptide. Its molecular mass was estimated at 64 000 Da by SDS/PAGE (Fig. 2A).
The optimum pH of rBtDex was 6.2, with broad pH stability of 5.1–10.6, and its thermostable range was < 45 °C. We determined the substrate specificity of the enzyme for 2 mmα-glucosides [panose, methyl α-isomaltooligosides (MIGn), maltose, maltotriose, kojibiose, nigerose, sucrose and trehalose] and 0.1% w/v α-glucans [Dextran T2000 (mean molecular weight 2 × 106), pullulan and soluble starch]. The enzyme was most active with Dextran T2000 (100%), followed by MIG6 (82%), MIG5 (20%) and MIG4 (6.6%). In contrast, no activity was detected in the presence of the other saccharides. The major hydrolysis product in reactions including MIGn substrates (DP 4–6) was isomaltotetraose, indicating that BtDex hydrolyzes substrates from their non-reducing side. This evidence may be useful to understand the hydrolysis pattern towards Dextran T2000. The kinetic parameters (Km and kcat) for the hydrolysis of MIGn (DP = 4–6) and Dextran T2000 are summarized in Table 1. MIGn substrates were used instead of isomaltooligosaccharides with high reducing values. The values of 1/Km and kcat/Km for MIGn increased as a function of DP. The kcat/Km for Dextran T2000 was 26 times greater than that of MIG6, suggesting that Dextran T2000 is the most favorable substrate for rBtDex. Taken together, these findings indicate that the recombinant enzyme possessed catalytic properties typical of those of an endo-dextranase.
Table 1. Kinetic parameters for the hydrolysis of MIGn and Dextran T2000 by rBtDex.
a Mean molecular mass 200 000 Da.
CI formation by rBtDex
We next investigated rBtDex-catalyzed production of CI from Dextran T10 (mean molecular weight 1 × 104) (Fig. 2B). Long CIs (DP 10–14) were formed during the first 5 min of the reaction. CI-10 was the major product present after 10 min. The CIs were degraded into linear oligosaccharides after 1 h. HPLC analysis of a 1 h reaction revealed a product consistent with CI-10. The molecular mass of an isolated compound was determined as 1643 Da (10 glucosyl residues × 162 + 23, M + Na+) by fast-atom bombardment/mass spectrometry (FAB-MS). 1H-NMR analysis did not detect a signal for the anomeric proton of a non-reducing glucose residue. The non-reducing sugar exhibited 13C-NMR chemical shifts (C1, 97.6 p.p.m.; C2, 71.4 p.p.m.; C3, 73.4 p.p.m.; C4, 70.2 p.p.m.; C5, 69.6 p.p.m.; C6, 65.8 p.p.m.) that were almost identical to those of CI-8  (C1, 97.9 p.p.m.; C2, 71.8 p.p.m.; C3, 73.4 p.p.m.; C4, 70.1 p.p.m.; C5, 70.3 p.p.m.; C6, 65.7 p.p.m.). These results indicate that rBtDex produced CI-8 to CI-14, with CI-10 predominant early in the reaction.
Although computer analyses indicated that rBtDex is a GH66 CITase, our analysis revealed it to be a Dex with minor CITase activity such as that exhibited by PsDex. These results indicate that rBtDex is a second example of a Dex/CITase, and supports our published classification of GH66 enzymes into three types .
We performed multiple sequence comparisons to construct a phylogenetic tree (Fig. 3), the structure of which shows that GH66 enzymes fall into two groups. Group 1 includes endo-dextranases from streptococcal bacteria, which possess only dextranolytic activity. Group 2 includes CITases, PsDex and BtDex, all of which display CITase activity. The genes encoding group 1 and 2 enzymes may share the same ancestor. Our results raise a point regarding the CITase-specific region: BtDex lacks a CITase-specific region, whereas it is present in CITase and PsDex. The CITase-specific region includes a carbohydrate-binding module 35 (CBM 35) that contributes to CI formation [7,9]. Thus it remains to be determined whether the CITase-specific region is essential for CITase activity.
A CITase truncated at its CITase-specific region including CBM 35 has been generated to identify the function of CBM 35 . Surprisingly, the CITase-specific region-truncated CITase still exhibits weak CI formation activity, suggesting that another structural element is required for the cyclization activity. Molecular analysis of BtDex, which lacks the CITase-specific region, may elucidate new structural element(s). Research is continuing to identify further candidates for CITase activity.
Bacterial strains, plasmids, reagents, and bioinformatics analysis
B. thetaiotaomicron VPI-5482 was obtained from the American Culture Type Collection (Manassas, VA, USA). E. coli DH5α and E. coli BL21 CodonPlus (DE3)-RIL (Stratagene, La Jolla, CA, USA) were used as hosts for gene cloning and protein expression, respectively. The plasmids pBluescript II SK(+) (Stratagene) and pET23d (Novagen, Darmstadt, Germany) were used as the cloning and expression vectors, respectively. E. coli cells were grown at 37 °C in Lenox broth (1% w/v tryptone, 0.5% w/v yeast extract, 0.5% w/v NaCl). Dextran sucrase from Leuconostocmesenteroides B-512 FMCM  and α-glucosidase from Streptococcusmutans ATCC 25175  were used for synthesis and enzymatic hydrolysis of MIGn of various DP.
Cloning of the BtDex gene and construction of an expression vector
B. thetaiotaomicron VPI-5482 was grown in chopped meat medium (Becton-Dickinson, Sparks, MD, USA) for 2 days under anaerobic conditions at 37 °C. Cellular genomic DNA was prepared by alkaline cell lysis, followed by extraction with phenol/chloroform, and ethanol precipitation. Two PCR primers were prepared according to the DNA sequence of B. thetaiotaomicron VPI-5482 : 5′-GAAACGATTTCAATTGTAAAACAATAAAC-3′ (sense, corresponding to bp −40 to −12 upstream of the translation initiation codon) and 5′-CCTATCGTTTTTAGTTATTCCGTTGC-3′ (antisense, corresponding to bp +33 to +8 upstream of the translation termination codon). Amplified fragments were inserted into the EcoRV site of pBluescript II SK(+) to construct the plasmid pSK-BtDex. Its sequence, determined twice using an ABI PRISM 310 genetic analyzer (Applied Biosystems, Foster City, CA, USA), was identical to that of the published sequence .
The gene encoding BtDex was amplified from the pSK-BtDex plasmid using the primers 5′-AACCATGGAGAAGATAATATATTTGGTG-3′ (sense, NcoI site underlined) and 5′-GCACTCGAGTTATTCAGCTACAATCATTG-3′ (antisense, XhoI site underlined). Insertion of NcoI site at the 5′-terminus of the open reading frame changed Glu to Lys (G2K mutation). The resulting PCR product and pET23d were digested with NcoI and XhoI, and were then ligated using a DNA ligation kit version 2 (Takara, Otsu, Japan). We confirmed that the inserted BtDex sequence was a perfect match to the sequence deposited in GenBank .
BLAST analysis was performed using the Swiss Institute of Bioinformatics website . The secretory signal peptide was predicted using the SignalP server . Multiple alignments of the amino acid sequence were performed using MUSCLE , and the phylogenetic tree was constructed by the neighbor-joining and unweighted pair group method with arithmetic mean (UPGMA) methods at http://toolkit.tuebingen.mpg.de/phylip.
Expression of the gene and rBtDex purification
E. coli BL21 CodonPlus (DE3)-RIL strain was transformed with pET23d-BtDex DNA. The transformants were grown in 600 mL Lenox broth containing 50 μg·mL−1 ampicillin and 30 μg·mL−1 chloramphenicol at 37 °C until the culture reached an attenuance of 0.5 at 600 nm. Cultures were agitated (180 rpm) for 30 min at 37 °C, followed by induction of rBtDex expression using 0.2 mm isopropyl β-d-thiogalactopyranoside overnight at 18 °C. Cells were harvested by centrifugation at 8000 g for 20 min at 4 °C, suspended in 50 mL 20 mm potassium phosphate buffer (pH 6.8) containing 0.5 m NaCl and 1.0 mm phenylmethansulfonyl fluoride, and then disrupted by sonication using a Sonifier-250 sonicator (Branson, Danbury, CT, USA). The soluble fraction (764 U, 0.59 U·mg−1) obtained by centrifugation (14 000 g for 20 min at 4 °C), was applied to a DEAE-TOYOPEARL 650M column (2.0 × 15 cm; Tosoh, Tokyo, Japan) equilibrated with 20 mm potassium phosphate buffer (pH 6.5, buffer A), and eluted with a linear gradient of 0–1.0 m NaCl. The pooled active fractions (434 U, 5.40 U·mg−1) were adjusted to 1.5 m ammonium sulfate, and loaded onto a butyl-TOYOPEARL 650M column (1.6 × 12.1 cm; Tosoh) equilibrated with buffer A containing 1.5 m ammonium sulfate, and eluted using a linear gradient of 1.5–0 m ammonium sulfate. The active fractions (214 U, 16.3 U·mg−1) were concentrated and loaded onto a Sephacryl S-100 HR column (3.0 × 80 cm; GE Healthcare, Piscataway, NJ, USA) equilibrated and eluted with buffer A containing 50 mm NaCl. The pool of active fractions (224 U, 20.4 U·mg−1) was dialyzed against buffer A, applied to a DEAE-TOYOPEARL 650M column (2.0 × 15 cm) equilibrated with buffer A, and eluted using a linear gradient of 0–1.0 m NaCl. The pool of active fractions (186 U, 22.7 U·mg−1) was then dialyzed against 20 mm potassium phosphate buffer (pH 6.5) and concentrated using a CentriPrep YM-30 centrifugal filter unit (Millipore, Bedford, MA, USA). All purification steps were performed at 4 °C. Enzyme concentrations were estimated from the amino acid content of 500 pmol protein hydrolysates (6 N HCl for 24 h at 110 °C) using an Amino Tac JLC-500/V amino acid analyzer (JEOL, Tokyo, Japan). One unit (U) of dextranolytic activity was defined as the amount of enzyme that released 1 μmol reducing power per min as determined using the copper bicinchoninate method , with glucose as the standard. The reducing power was measured using 0.4% w/v Dextran T2000 in 20 mm MES/NaOH (pH 6.2) at 35 °C for 10 min.
Gel electrophoresis and protein analysis
Enzyme purity was evaluated using 10% SDS/PAGE. Protein bands were visualized using Coomassie Brilliant Blue . The relative mobility of the enzyme band was compared to that of size markers (25–200 kDa; Invitrogen, Carlsbad, CA, USA). The proteins were transferred electrophoretically to a ProSorb membrane (Applied Biosystems), and a piece of the membrane containing protein was subjected to protein sequencing. The amino acid sequence of the protein bound to the membrane was determined using a Procise model 490 automated protein sequencer (Applied Biosystems). The molecular mass of the protein was determined by ESI-MS using a JMS-700TZ mass spectrometer (JEOL).
Characterization of rBtDex
To determine the enzyme’s pH optimum, 22.5 nm rBtDex was incubated at 35 °C in 32 mm Britton–Robinson buffer (pH 3.7–11.5) with 0.1% w/v Dextran T2000 for 10 min. For pH stability assays, 22.5 nm rBtDex was kept at 4 °C for 24 h in 32 mm Britton–Robinson buffer (pH 3.7–12.0), and the remaining enzyme activity was examined at 35 °C in 180 mm MES/NaOH buffer (pH 6.2) with 0.1% w/v Dextran T2000. The thermal stability of the enzyme was determined by incubation of 22.5 nm rBtDex in 40 mm sodium citrate buffer (pH 6.2) at 25–65 °C for 15 min. Enzyme activity was measured in terms of the reducing power using the copper bicinchoninate method  at 35 °C in 40 mm sodium citrate buffer (pH 6.2) with 0.1% w/v Dextran T2000.
The kinetic parameters (Km and kcat) for rBtDex were determined in reactions using MIGn and Dextran T2000. The initial velocities (v) were measured in the presence of varying concentrations of MIG4 (0.25–2.50 mm), MIG5 (0.25–2.50 mm), MIG6 (0.05–0.97 mm) and Dextran T2000 (1–25 μm) in 40 mm MES/NaOH buffer (pH 6.2) at 35 °C using 562, 56.2, 56.2 and 11.2 nm rBtDex, respectively. The kinetic parameters were calculated from initial rates by fitting to the Michaelis–Menten equation by non-linear regression using the Curve Expert 1.3 computer program (http://www.curveexpert.net/).
The 20 μL reaction mixture containing 2 mm MIG5, 8.36 μg rBtDex and 40 mm MES/NaOH buffer (pH 6.2) was incubated at 35 °C for 10 h. Reaction products were determined using TLC on a Kiesel Gel silica gel 60 F254 plate (Merck, Darmstadt, Germany) using a solvent system comprising 85 : 70 : 20 : 50 v/v/v/v acetonitrile/H2O/ethyl/acetate/1-propanol and standard isomaltooligosaccharides (DP = 2–7) and MIG4–MIG6. Sugars on the TLC plate were visualized by dipping into 0.03%N-(1-naphthyl) ethylenediamine, 5% v/v H2SO4 in methanol, followed by heating at 120 °C for 5 min .
Synthesis of MIGn and CI
MIG4–MIG6 were prepared by means of an acceptor reaction catalyzed by dextran sucrose, using sucrose as the glucosyl donor and methyl α-glucoside as the acceptor. The 10 mL reaction mixture containing 0.35 m sucrose, 0.15 m methyl α-glucoside, 40 mm sodium acetate buffer (pH 5.1), 10 mm CaCl2 and 50 μL dextran sucrase (67.4 U·mL−1) was incubated at 37 °C for 26.5 h. The reaction products were separated by TLC using acetonitrile/water (85 : 15 v/v). The resultant dextran was precipitated by adding 80% v/v ethanol, and then desalted using Amberlite MB-3 (Ograno, Tokyo, Japan). The concentrated MIGn preparations were analyzed using a PU-2089 HPLC system (Jasco, Tokyo, Japan) equipped with a refractive index detector (model D-2000; Hitachi, Tokyo, Japan) and a Shodex RS Pak DC-613 polymer-based column (Showa Denko, Tokyo, Japan). The mobile phase was 61% v/v acetonitrile at 50 °C.
To analyze the synthesis of CI, the reaction mixture (25 mL) containing 0.139 U·mL−1 rBtDex, 2% w/v Dextran T10 and 40 mm MES/NaOH buffer (pH 6.2) was incubated at 35 °C for 1 h, followed by termination of the reaction in boiling water for 10 min. After discarding the excess dextran by adding 80% v/v ethanol and desalting the sample using Amberlite MB-3, the remaining linear isomaltooligosaccharides were hydrolyzed using buckwheat α-glucosidase (7.6 U·mL−1) and Aspergillus nigerα-glucosidase (12 U·mL−1) for 12 h at 37 °C. The CIs were isolated by applying the solution to a SepPack18 cartridge (Waters, Milford, CT, USA), followed by elution with 20% v/v ethanol, concentrated by vacuum evaporation, and subjected to HPLC through a Shodex RS Pak DC-613 column at 60 °C using a mobile phase of 57% v/v acetonitrile.
The structures of MIGn and CI (2 mg) were determined using FAB-MS on a JEOL JMS-SX102A mass spectrometer. NMR was performed using an AMX-500 spectrometer (Bruker, Rheinstetten, Germany) in 99.98% D2O (Sigma-Aldrich, St Louis, MO, USA) at 500 MHz for 1H nuclei and at 125 MHz for 13C nuclei using tetramethylsilane (external standard). The FAB-MS molecular masses with sodium ion were 865 Da (MIG4), 1027 Da (MIG5) and 1189 Da (MIG6). Isolated MIGn were completely hydrolyzed to glucose by using α-glucosidase from Streptococcusmutans ATCC 25175 , indicating an α-1,6-glucosidic linkage. Sugar concentrations were determined using the phenol/sulfuric acid method .
This work was partially supported by Grant-in-Aids for Scientific Research from the Japan Society for the Promotion of Science, and from the Program for Promotion of Basic and Applied Researches for Innovations in Bio-oriented Industry (BRAIN).