By analysis of the genome sequence of T. thermosaccharolyticum DSM 571, a protein (Tthe_1813), defined as β-galactosidase in Genbank, consists of a 1,329-bp fragment encoding 443 amino acids, which belonged to family 1 of the glycoside hydrolases. It shares the highest sequence similarity of 66% with the β-glucosidses from Thermoanaerobacter mathranii (Genbank No. YP_003676178.1) and Thermoanaerobacter pseudethanolicus ATCC 33223 (Genbank No. YP_001665894.1), which were revealed by whole-genome sequencing but has not been biochemically characterized. Alignment of the BGL cluster with several representative members of GH1 indicated that they share similar blocks. The catalytic proton donor, Glu¹³⁵ and Glu³⁵¹ in BGL are well conserved among all GH1 proteins (Figure 1). The sequence around Glu³⁵¹ in BGL is [LYT-NGAA], which is consistent with the consensus pattern of PS00572. The results indicated that the protein (Tthe_1813) could be a novel β-glucoside. Then the DNA fragment of a protein (Tthe_1813) gene was amplified from genomic DNA of T. thermosaccharolyticum DSM 571, and ligated to pET-20b at Nde I and Xho I sites to generate plasmid pET-20-BGL.

In order to increase the expression level of BGL in E. coli, site-directed mutagenesis were designed and performed to optimize condons of BGL for E. coli expression system. pET-20-BGLII was obtained from pET-20-BGL in which the rare condons for the N-terminal amino acid residues were replaced by optimal codons in E. coli without and change of amino acid sequence (Figure 2), so pET-20-BGLII encodes the same β-glucosidase as that encoded by the wild-type gene. The β-glucosidase activity expression from pET-20-BGLII was 7.5 U/mL (11.2 U/mg total of cell protein) and was estimated to be about 30% of the total protein, which was about 1.7 times higher than the expressed from pET-20-BGL (Figure 3, lane 2 and 3).

The protein in the cell-free extract was purified to gel electro homogeneity after a heat treatment and a Ni-NTA affinity. The final preparation gave a single band on SDS-PAGE gel and the molecular mass of the enzyme was estimated to be 52 kDa (Figure 3, lane 4).

The biochemical properties of BGL were investigated by using the purified recombinant BGL. The optimal pH of the BGL was determined to be 6.4 (Figure 4a), while the β-glucosidase activity was higher than 50% of the maximum activity at the pH range from 5.6 to 7.2. The enzyme was stable for about 1 h at pH 5.6 to 8.0 at 60°C in the absence of the substrate (Figure 4c). The optimal temperature of the BGL was 70°C, which the β-glucosidase activity was higher than 40% of the maximum activity at the temperature range from 45 to 75°C (Figure 4b). Thermostability assays indicated that its residual activity was more than 80% after being incubated at 60°C for 2 h (pH 6.4, Figure 4d).

The effects of metal ions and some chemicals on the enzyme activity were shown in (Table 1). In various assays, the enzyme activity was significantly enhanced by Fe²⁺, or Mn²⁺, and completely inactivated by Zn²⁺, Cu²⁺, Ag²⁺, or Hg²⁺. The effects of Mg²⁺, Ca²⁺, K⁺, Li², or EDTA (10 mM) on the enzyme activity were not so significant.

^a Final concentration, 1 mM or as indicated. Values shown are the mean of duplicate experiments, and the variation about the mean was below 5%.

The enzyme was able to hydrolyze p-nitrophenyl-β-D-glucopyranoside, cellobiose, and p-nitrophenyl-β-D-galactopyranoside, while no activity was detected upon p-nitrophenyl-α-L-arabinofuranoside, p-nitrophenyl-β-D-xylopyranoside, maltose, CMC, and sucrose. p-nitrophenyl-β-D-Galactopyranoside was hydrolyzed at 40% of that of p-nitrophenyl-β-D-glucopyranoside. The dependence of the rate of the enzymatic reaction on the substrates concentration followed Michaelis-Menten kinetics, with K_m and V_max values of 0.62 mM and 64 U/mg for p-nitrophenyl-β-D-glucopyranoside, and for cellobiose 7.9 mM and 120 U/mg under optimal conditions. The effects of the substrate, cellobiose (290 mM), on the enzyme activity were not significant. The K_cat/K_m value for cellobiose 13.3 mM^-1 s^-1 was less than the β-glucosidase from A. oryzae, but the activity of β-glucosidase from A. oryzae was inhibited by cellobiose, and rapidly decreased above 50°C (Table 2). Furthermore, the enzyme activity was enhanced by the concentrations of glucose below 200 mM, and the enzyme activity was increased 110% when adding 100 mM glucose into reaction mixtures (Figure 5). When glucose was increased, the enzyme activity of BGL was gradually inhibited, with a K_i of 600 mM glucose (Figure 5). The properties of the glucose-tolerant β-glucosidase from other microorganisms are summarized in Table 2. As Table 2 shows, these enzymes have many distinct features, especially in their catalytic properties 121318192021.

^a pNPG: p-nitrophenyl-β-D-glucopyranoside.

^b ND: not determined.

^c It was calculated by the data based on the reference.

Production of glucose from 290 mM cellobiose (10%) by the purified BGL was examined. Even if the final concentration of glucose in reaction reached about 580 mM, cellobiose (290 mM) was found to be degraded completely (Figure 6a, b). At the beginning of the reaction, the K_cat was 67.7 s^-1 within one hour at 60°C which was identical to the theoretical value. During the whole degradation process, the K_cat was 28.2 s^-1.

To gain insights into the evolutionary relationship among β-glucosidases, we constructed the phylogenetic trees of 40 candidate sequences using he NJ method and the MP method respectively, both supporting almost the same topology. The results revealed the presences of five well-supported clades: Clade II was GH1 β-glucosidases from fungi, and Clade III was the GH3 β-glucosidases from bacteria, and Clade IV was the GH3 β-glucosidases from fungi. The GH1 β-glucosidases from bacteria was divided into two clades: Clade I mainly contained mesophilic bacteria; Clade V mainly contained thermophile, which is formed by further divided into two subclades, of which one contains all thermophile, and the other Bacillus GH1 β-glucosidases. Clade II and clade III had a relatively close relationship, and the GH1 β-glucosidases from thermophile were distant from the other clades (Figure 7).

Thermoanaerobacterium thermosaccharolyticum DSM 571 was purchased from DSMZ (http://www.dsmz.de). It was grown anaerobically at 60°C as described previously 17. Escherichia coli JM109 and JM109(DE3) was grown at 37°C in Luria-Bertani medium (LB) and supplemented with ampicillin when required. The expression vectors pET-20b (Novagen) were employed as cloning vector and expression vector.

DNA was manipulated by standard procedures 25. QIAGEN Plasmid Kit and QIAGEN MinElute Gel Extraction Kit (Qiagen, USA) were employed for the purification of plasmids and PCR products. DNA restriction and modification enzymes were purchased form TaKaRa (Dalian, China). DNA transformation was performed by electroporation using GenePulser (Bio-Rad, USA). Site-directed mutagenesis of genes and the modification of the plasmids were performed by inverse-PCR followed by phosporylation and self-ligation using T4 polynucleotide kinase and T4 DNA ligase.

The β-glucosidase gene bgl was amplified from T. thermosaccharolyticum DSM 571 genomic DNA by PCR using primers bgl-1 and bgl-2 (Table 3), the PCR products were digested with Nde I and Xho I and inserted into pET-20b at Nde I and Xho I sites, yielding the plasmid pET-20-BGL.

The boldface italic nucleotides represented mutations for optimizing codons.

In order to improve the expression level of recombinant BGL, the internal region from 1^st to 19^th amino acids in open reading frame of bgl was mutated in situ by inverse-PCR to replace the rare codons with the optimal codons of E. coli; the primers for the inverse-PCR were designated as bgl-3 and bgl-4 (Table 3). Inverse-PCR with primers was carried out using Pyrobest with pET-20-BGL as template, generating the plasmid pET-20-BGLII.

Plasmids pET-20-BG and pET-20-BGLII were transformed into E. coli JM109(DE3), and induced to expressed recombinant BGL by adding isopropyl-β-D-thiogalactopyranoside (IPTG) to final concentration of 0.8 mM at OD₆₀₀ about 0.7, and incubated further at 30°C for about 6 h.

One liters of the recombinant cells carrying pET-20-BGLII were harvested by centrifugation at 5,000 g for 10 min at 4°C, and washed twice with distilled water, resuspended in 50 mL of 5 mM imidazole, 0.5 mM NaCl, and 20 mM Tris–HCl buffer (pH 7.9), and French-pressured for three times. The cell extracts were heat treated (60°C, 30 min), and then cooled in an ice bath, and centrifuged (20,000 g, 4°C, 30 min). The resulting supernatants were loaded on to an immobilized metal affinity column (Novagen, USA), and eluded with 1 M imidazole, 0.5 M NaCl, and 20 mM Tris–HCl buffer (pH 7.9). Protein was examined by SDS-PAGE 26, and the protein bands were analyzed by density scanning with an image analysis system (Bio-Rad, USA). Protein concentration was determined by the Bradford method using BSA as a standard.

The reaction mixture, containing 50 mM imidole-potassium buffer (pH 6.4), 1 mM p-nitrophenyl-β-D-glucopyranoside, and certain amount of β-glucosidase in 0.2 mL, was incubated for 5 min at 70°C. The reaction was stopped by adding 1 mL of 1 M Na₂CO₃. The absorbance of the mixture was measured at 405 nm. One unit of enzyme activity was defined as the amount of enzyme necessary to liberate 1 μmol of pNP per min under the assay conditions.

The optimum pH for activity β-glucosidase was determined by incubation at 70°C for 5 min in the 50 mM imidole-potassium buffer from pH 4.8 to 8.4. The optimum temperature for the enzyme activity was determined by standard assay ranging from 45 to 85°C in the 50 mM imidole-potassium buffer, pH 6.0. The results were expressed as percentages of the activity obtained at either the optimum pH or the optimum temperature.

The pH stability of the enzyme was determined by measuring the remaining activity after incubating the enzyme (0.1 μg) at 50°C for 1 h in the 50 mM imidole-potassium buffer from pH 5.2 to 8.0. To determine the effect of temperature on the stability of BGL, the enzyme (0.1 μg) in the 50 mM imidole-potassium buffer (pH 6.4) was pre-incubated for various times at 50°C, 65°C, 68°C and 70°C in the absence of the substrate. The activity of the enzyme without pre-incubation was defined as 100%.

The effects of metals and chemical agents on β-glucosidase activity of purified enzyme (0.1 μg) were determined. Fe²⁺, Mg²⁺, Zn²⁺, Mn²⁺, Ca²⁺, K⁺, Al³⁺, Li²⁺, Cu²⁺, Co²⁺, and Hg²⁺ were assayed at concentrations of 1 mM in the reaction mixture. The chemical agents EDTA (10 mM) were assayed. The enzyme was incubated with each reagent for 10 min at 50°C before addition of p-nitrophenyl-β-D-glucopyranoside to initiate the enzyme reaction. Activity was determined as described above and was expressed as a percentage of the activity obtained in the absence of the chemical agents and metal cations.

The substrate specificity of the enzyme (0.1 μg) was tested by using following p-nitrophenyl-β-D-glucopyranoside, p-nitrophenyl-β-D-xylopyranoside, p-nitrophenyl-α-L-arabinofuranoside, maltose, sucrose, and cellobiose. Kinetic constant of BGL was determined by measuring the initial rates at various p-nitrophenyl-β-D-glucopyranoside concentrations (0.2, 0.4, 0.6, 0.8, 1, 2, and 4.0 mM) or various cellobiose concentration (2, 4, 6, 8, 10, 12, 14, and 16 mM) under standard reaction conditions. The K_i value of glucose was defined as amount of glucose required for inhibiting 50% of the β-glucosidase activity and was given as the averages of three separate experiments performed in duplicate.

The condon usage preference of E. coli in translation initiation region of pET-20-BGL was analyzed by using codon usage tool (http://gcua.schoedl.de/). The potential ORF of bgl was searched using the ORF search tool provided by the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov). Database searching was performed with Blast at NCBI and against CAZy (http://www.cazy.org). The active site of the enzyme was analyzed with the prosite tool (http://prosite.expasy.org/scanprosite). The multiple sequence alignment tool Clustal X2.0 was used for multiple protein sequence alignment 27. Sequences were further edited and aligned manually, when necessary, using the Mega 5 for editing. For phylogenetic analyses of conserved domains, sequences were trimmed so that only the relevant protein domains remained in the alignment 28. Phylogenetic relationships were inferred using the Neighbor-Joining (NJ) and Maximum-Parsimony (MP) method as implemented in Paup 4.0 for the NJ and MP trees, the results were evaluated with 1000 bootstrap replicates 29. The generated trees were displayed using TREEVIEW 1.6.6 (http://taxonomy.zoology.gla.ac.uk/rod/treeview.html).

The cellobiose was treated with purified BGL, and the degradation was subjected to analysis of thin-layer chromatography (TLC) and HPLC. The reaction mixture (20 μL) contained 290 mM cellobiose, and 1 μg of BGL in 50 mM imidole-potassium buffer (pH 6.4). The reaction was performed for various times at 60°C, and stopped by heating for 5 min in a boiling water bath. After centrifuged for 10 min at 10,000 g, supernatants of the reaction mixtures were applied on silica gel TLC plates (60F254, Merck Co.). Sugars on the plates were partitioned with a solvent system consisting of n-butanol, acetic acid, and water (2:1:1, by vol/vol), and detected using the orcinol reagent 30. The concentration of glucose was examined by HPLC on a carbohydrate analysis column (Waters Sugarpak1, USA) with water as a mobile phase.