First Glycoside Hydrolase Family 2 Enzymes from Thermus antranikianii and Thermus brockianus with β-Glucosidase Activity

Two glycoside hydrolase encoding genes (tagh2 and tbgh2) were identified from different Thermus species using functional screening. Based on amino acid similarities, the enzymes were predicted to belong to glycoside hydrolase (GH) family 2. Surprisingly, both enzymes (TaGH2 and TbGH2) showed twofold higher activities for the hydrolysis of nitrophenol-linked β-D-glucopyranoside than of -galactopyranoside. Specific activities of 3,966 U/mg for TaGH2 and 660 U/mg for TbGH2 were observed. In accordance, Km values for both enzymes were significantly lower when β-D-glucopyranoside was used as substrate. Furthermore, TaGH2 was able to hydrolyze cellobiose. TaGH2 and TbGH2 exhibited highest activity at 95 and 90°C at pH 6.5. Both enzymes were extremely thermostable and showed thermal activation up to 250% relative activity at temperatures of 50 and 60°C. Especially, TaGH2 displayed high tolerance toward numerous metal ions (Cu2+, Co2+, Zn2+), which are known as glycoside hydrolase inhibitors. In this study, the first thermoactive GH family 2 enzymes with β-glucosidase activity have been identified and characterized. The hydrolysis of cellobiose is a unique property of TaGH2 when compared to other enzymes of GH family 2. Our work contributes to a broader knowledge of substrate specificities in GH family 2.

Glycoside hydrolases find a wide range of industrial applications. For instance, β-glucosidases are applied in biorefineries to reduce cellobiose-mediated product inhibition of endoglucanases. Thus, the addition of β-glucosidases leads to increased glucose concentrations during the enzymatic hydrolysis of cellulose, thereby increasing ethanol yields (Viikari et al., 2007). Furthermore, β-glucosidases and β-galactosidases are applied in the food industry to improve the aroma of juices and wine (Bhatia et al., 2002). Many industrial processes run at harsh conditions, e. g., elevated temperatures and extremes of pH. Hence, the demand for novel enzymes which function at high temperatures or acidic or alkaline pH values is growing (Antranikian and Egorova, 2007).
Bacteria belonging to the genus Thermus grow at temperatures between 53 and 86°C, and at pH values between 6.0 and 10.5 (Dworkin et al., 2006). Since the discovery of the type strain of this genus, Thermus aquaticus in 1969, numerous species have been isolated from hot environments (Brock and Freeze, 1969;Oshima and Imahori, 1974). Especially, the DNA polymerase from T. aquaticus (Taq DNA polymerase) became one of the key enzymes in molecular biology (Chien et al., 1976). Furthermore, thermostable DNA-processing enzymes such as ligases, helicases, or endonucleases have been identified (Pantazaki et al., 2002). Although Thermus spp. produce different hydrolytic enzymes like proteases or lipases, only few glycoside hydrolases have been characterized so far (Dion et al., 1999;Fridjonsson et al., 1999;Pantazaki et al., 2002;Kim et al., 2006;Nam et al., 2010;Blank et al., 2014).
In the current report, two genes coding for GHs (TaGH2 and TbGH2) were identified from Thermus antranikianii and Thermus brockianus. Conserved domains of GH2 were detected with incomplete motifs. The recombinant proteins showed highest activity toward 4-NP-β--glucopyranoside. The activity of TaGH2 toward cellobiose makes this enzyme unique when compared to the GH family 2.
Isolated genomic DNA from Thermus strains was partially digested with BamHI to generate fragments of 5-10 kb. DNA fragments of the desired size were separated by agarose gel electrophoresis. To prepare an agarose gel of 1%, 1 g agarose was dissolved in 100 mL TAE buffer (40 mM Tris, 1 mM EDTA, 40 mM acetic acid, pH 8.5). The DNA was extracted from the gel by using the "GeneJet Gel Extraction Kit" (Fermentas, St. Leon-Rot, Germany). The ligation into the "ZAP Express Vector, " as well as further steps for gene library construction were carried out according to the manufacturer's instructions.
To detect β-glucosidase-encoding genes, the gene libraries were plated on LB agar supplemented with 50 µg/mL kanamycin, and after growth over night the colonies were replicated on plates containing additionally 1 mM IPTG. The clones were overlayed with buffer (25 mM sodium acetate, 2.5 mM CaCl 2 × 2 H 2 O, 170 mM NaCl, 1% agarose, pH 6.5) containing 2.5 mM esculin and 0.4 mM ammonium-iron(III)-citrate. After incubation at 70°C for 1-16 h, β-glucosidase positive clones were observed by the formation of a brown halo around the colonies. E. coli XLOLR pBK-CMV clones displaying β-glucosidase activity were conserved as cryostocks containing 25% glycerol.

Sequencing and Sequence Analysis
Plasmids were isolated ("GeneJET TM Plasmid Miniprep Kit", Fermentas, St. Leon-Rot, Germany) from selected clones, which conferred β-glucosidase activity. To determine the DNA-sequence of integrated DNA-fragments, the plasmids were sent to Eurofins MWG Operon (Ebersberg, Germany) for sequencing.
The PCRs were carried out in a thermocycler according to the manufacturer's instructions. A mixture of 20 µL contained 0.4 U "Phusion High-Fidelity DNA-Polymerase" (Fermentas, St. Leon-Rot, Germany), 0.2 mM dATP, dCTP, dGTP, dTTP, 0.5 µM of the forward and reverse primer, 10-300 ng template, reaction buffer, and distilled autoclaved water. The obtained PCR products were cloned into the pJet1.2/blunt vector (Fermentas, St. Leon-Rot, Germany), which was thereafter used to transform competent E. coli Nova Blue Single TM cells. Positive transformants were identified by colony PCR. Isolated pJet1.2/blunt plasmids were double digested with PaeI/SalI or BamHI/SalI to excise the β-glucosidase-encoding genes. The purified genes were ligated into the PaeI/SalI and BamHI/SalI digested expression vector pQE-80L, which was used to transform competent E. coli BL21 Star TM (DE3) cells. Positive transformants were identified by colony PCR and conserved as cryostocks containing 25% glycerol.

Production and Purification of the Enzymes
E. coli BL21 Star TM (DE3) harboring the recombinant pQE-80L plasmids were cultivated in 5 mL LB medium containing 100 µg/mL ampicillin at 37°C and 160 rpm for 16 h. The preculture was subcultivated in 500 mL LB medium containing 100 µg/mL ampicillin until the culture reached an optical density of A 600 nm = 0.5-0.7. To induce protein production, 0.5 M IPTG was added. The cultivation was continued for 6 h, afterwards the cells were harvested by centrifugation at 13,000 rpm for 20 min at 4°C. The cell pellet was resuspended in lysis buffer (50 mM NaH 2 PO 4 , 300 mM NaCl, 10 mM imidazole, pH 8) in the ratio 5 mL buffer/1 g pellet. Cells were disrupted by French press (French R Pressure Cell Press, SLM-Aminco, MD, USA).
Crude extracts with protein contents of 1.5 mg/mL (TaGH2) and 1.07 mg/mL (TbGH2) were obtained by centrifugation at 13,000 rpm at 4°C for 20 min.
For heat precipitation, the crude extract was incubated at 70°C for 15 min, and subsequently centrifuged at 13,000 rpm at 4°C for 20 min.
In the case of TbGH2, a gel filtration using the "ÄKTA TM Fast Protein Liquid Chromatography" (FPLC)-system (GE Healthcare, München, Germany) was carried out. One milliliter of the protein solution was load on a "HiLoad 16/60 Superdex 200 prep grade"column (GE Healthcare, München, Germany) with a flow rate of 1 mL/min. The protein was eluted with a 1.5-fold column volume of 50 mM sodium phosphate buffer, pH 7.2, which contained 150 mM NaCl. Enzyme fractions were pooled.
The purified proteins were dialyzed against 20 mM citrate buffer (TaGH2) or 20 mM maleate buffer (TbGH2) and stored at 4°C.
Protein concentrations were determined using the method described by Bradford (1976).

Determination of Enzyme Activity
Enzyme activity was determined by the released amounts of 2-nitrophenol (2-Np) and 4-nitrophenol (4-Np) from several nitrophenol-linked substrates (4-Np-β--glucopyranoside, The reaction mixture of 1 mL contained 2 mM Np-substrate and 20 mM citrate buffer pH 6.5 (TaGH2), or 20 mM maleate buffer pH 6.5 (TbGH2). After preincubation of the reaction mixture for 5 min at 95°C (TaGH2) or 90°C (TbGH2), 10 µL of diluted enzyme solution were added. The hydrolysis was stopped after 10 min by adding 100 µL of 100 mM Na 2 CO 3 . The absorbance was measured at 410 nm. All measurements were determined in triplicates. About 1 U of enzyme activity was defined as the amount of enzyme needed for the release of 1 µmol 4-nitrophenol per minute. Kinetic parameters were determined according to Michaelis and Menten (1913).

Effect of pH and Temperature
Relative activities against 4-Np-β--glucopyranoside (2 mM) were measured in the range of pH 4.0-10.0 using 20 mM Britton-Robinson buffer (Britton and Robinson, 1931). To determine the pH stability, both enzymes were preincubated in 20 mM Britton-Robinson buffer, pH 3.0-10.0 at 4°C for 24 h. To stabilize the enzymes, the concentration of the solution was adjusted to 0.1 mg/mL with BSA. After preincubation, the relative activity toward 4-Np-β--glucopyranoside (2 mM) was determined in 20 mM Britton-Robinson buffer pH 6.5 and 95°C (TaGH2), or in 20 mM maleate buffer at 90°C (TbGH2).

Determination of Hydrolysis Products
Hydrolysis products of 1% (w/v) cellobiose and lactose were examined by HPLC (Agilent Technology 1260 Infinity Quarternary LC system with 1260 ALS sampler, 1260 Quat pump and 1260 R i detector). The enzymes were incubated with the substrates at 90°C for 1 h in 20 mM citrate buffer (TaGH2) or 20 mM maleate buffer (TbGH2), pH 6.5. After hydrolysis, 20 µL of the centrifuged and filtered solution was applied to a Hi-Plex H column (Agilent Technologies, Waldbronn, Germany). Water was used as solvent with a flowrate of 0.6 mL/min. To identify the hydrolysis products, the retention times (min) were compared to the standards cellobiose (9.433), lactose (9.789), and glucose (11.247).

Identification of Novel GH-Encoding Genes
Gene libraries were constructed from pure cultures of T. antranikianii and T. brockianus, and screened for genes coding for enzymes active toward esculin. One activity-conferring E. coli clone was detected from each library. The respective inserts of the phagemids (pBK-CMV:Ta and pBK-CMV:Tb) harbored tagh2 and tbgh2. The nucleotide sequences of tagh2 and tbgh2 were 80% identical. GC contents of 68.4 and 66.7% with GC proportions in the third codon position (GC 3 ) of 86.3 and 82.3% were observed.

Expression of tagh2 and tbgh2 and Protein Purification
The genes tagh2 and tbgh2 were amplified and expressed in E. coli BL21 Star TM (DE3) using pQE-80L (Figure 3). A high proportion of TbGH2 was produced in insoluble form. TaGH2 and TbGH2 were purified from soluble crude extracts. After heat precipitation and Ni-NTA affinity chromatography, TaGH2 was homogenous with a yield of 17.3%, whereas TbGH2 was subsequently subjected to size exclusion chromatography with a yield of 6.7%. The calculated molecular masses of 79.0 kDa (TaGH2) and 78.1 kDa (TbGH2) were confirmed by SDS gel analysis (Figure 3).

FIGURE 4 | HPLC analyses of cellobiose degradation.
The degradation of cellobiose (1% w/v) by TaGH2 and by TbGH2 is depicted. The hydrolysis was carried out at 90°C for 1 h in 20 mM citrate or maleate buffer (pH 6.5).
Subsequently, the samples were subjected to HPLC analyses using an Hi-Plex H column and water with a flow rate of 0.6 mL/min. Standards are shown in gray. nRIU, nano Refractive Index Units.

FIGURE 5 | Effect of temperature on activity of TaGH2 and TbGH2.
Activity of TaGH2 and TbGH2 was measured at different temperatures (50-115°C) for 10 min in 20 mM citrate or maleate buffer (pH 6.5) with 2 mM 4-Np-β-D-glucopyranoside. Temperatures above 95°C were measured in heated oil. The presence of 5 mM additives, such as β-mercaptoethanol, DTT, EDTA, urea, iodoacetic acid, sodium azide, and Tween 80 did not decrease the activity of both enzymes. Moreover, Triton X-100 and Tween 20 did not influence the performance of TbGH2, whereas TaGH2 exhibited 49 ± 2 and 76 ± 1% residual activity. Complete loss of activity for both glycoside hydrolases was detected when SDS or CTAB (5 mM) were added.

Discussion
Screening of gene libraries from T. antranikianii and T. brockianus on esculin resulted in the identification of two novel glycoside hydrolases (TaGH2 and TbGH2). Identities of amino acid sequences higher than 80% were observed to proteins from other Thermus species. The glycoside hydrolase family 2 domains (sugar binding domain and TIM-barrel domain) appeared to be incomplete. The domains are either truncated but functional or they are not recognized as domains due to large differences to other representatives of GH2. Comparison with the GH2 LacZ from E. coli (1HN1_D) showed low similarities within the corresponding domains. Additionally, conserved amino acid signatures typical for members of GH2 were partially identified (Figure 2). Similar hypothetical proteins with typical conserved domains from different genera were detected with identities of 39% (TaGH2) and 42% (TbGH2) or below. Thus, a differentiation of proteins produced from Thermus spp. would be suggested. Low sequence similarities within one protein superfamily could be the result of adaptation to different environmental conditions. Different enzymatic activities may have developed from a common ancestor (Galperin and Koonin, 2012). Structural diversification that preserved the active site residues may result in catalytically active enzymes with altered substrate specificity. Evolutionary pressure promotes functional and effective enzymes, which may result in reduction of conserved domains with remaining functionality (Juers et al., 1999). However, substrate specificity can be neglected for classification of glycoside hydrolases, since the structural homologies especially the motif forming the catalytic center is highly conserved (Henrissat and Davies, 1997). Homologies of up to 98 and 82%, respectively, were observed by comparison with annotated β-galactosidases or β-mannosidases from Thermus species. No similar characterized enzyme to TaGH2 or TbGH2 was detected in the NCBI database; thus, classification of similar proteins was achieved by domain prediction rather than by functional analyses. Highest catalytic efficiencies for TaGH2 and TbGH2 were observed toward 4-Np-β--glucopyranoside with residual activities toward 4-Np-β--galactopyranoside. Cellobiose conversion distinguishes TaGH2 from TbGH2. The hydrolysis characteristics may vary between artificial and natural substrates. This may be due to differences in size, charge, and binding properties of the artificial compound. Higher activity toward artificial substrates was also described as common phenomenon for α-galactosidases (Wang et al., 2014). Hydrolysis of cellobiose or lactose by TbGH2 could not be detected, although high β-glucosidase activity was observed when the artificial substrate was used. β-Glucosidases are reported to often exhibit a broad substrate spectrum. A β-glucosidase from Thermotoga neapolitana hydrolyzed among others 4-Np-β--glucopyranoside, cellobiose, and lactose (Park et al., 2005). Likewise, a GH1-β-glucosidase derived from a hot-spring metagenome exhibited activity toward 4-Np-β--glucopyranoside, 4-Np-β--galactopyranoside, cellobiose, and lactose . However, the newly discovered GH2 β-glucosidases TaGH2 and TbGH2 showed a narrow substrate range.
Protein production in the mesophilic host E. coli resulted in notable amount of inclusion bodies due to differences in genomic GC contents of the genus Thermus, as previously reported (Ishida and Oshima, 1994;Fridjonsson et al., 1999). The β-glucosidase TaGH2 and the β-glucosidase/galactosidase TbGH2 showed high activities at elevated temperatures. Other β-glycosidases from Thermus thermophilus also show activity at 88-90°C and pH 5.4-7.0 (Dion et al., 1999;Nam et al., 2010). A thermal activation was detected at 50 and 60°C, especially for TbGH2 with 246 and 202% activity after 24 h, respectively. It was reported in the literature that enzymes from thermophilic organisms produced in mesophilic hosts may require thermal activation as demonstrated for a β-glycosidase from T. thermophilus at 70°C (Gerard et al., 2002). Specific activities were considered high with 3,966 and 660 U/mg, when compared to the majority of previously reported β-glucosidases and β-galactosidases with <1-100 U/mg 7 . Similar to GH family 1 β-glycosidases from T. thermophilus, significantly lower Km values were observed with 4-Np-β-glucopyranoside when compared to -galactopyranoside (Dion et al., 1999;Nam et al., 2010). Highest substrate affinities for -glucopyranoside were also reported for other β-glucosidases from an uncultured soil bacterium and Cellulomonas flavigena with Km-values of 0.16 and 7.1 mM, respectively (Barrera-Islas et al., 2007;Kim et al., 2007). Hence, the enzymes described here exhibit Km-values in the range of other described GH1 βglucosidases.
Ag + , Cu 2+ , and Fe 3+ are most frequently described as glycoside hydrolase inhibitor (Cairns and Esen, 2010). By contrast, Fe 3+ (1 mM) had no influence on the activity of both enzymes. Zinc ions inhibited TbGH2 completely but had, interestingly, no effect on activity of TaGH2. The surfactant CTAB (5 mM) inhibited both enzymes. The negatively charged catalytic amino acids in the active center might be covered by the positively charged additive. Moreover, negatively charged surface residues might be influenced resulting in structural destabilization of the enzyme. Although three cysteins are present in the amino acid sequence of both enzymes, reducing agents such as DTT and βmercaptoethanol did not show a negative effect on the activity. Either no disulfide bond is formed or it is not affected by the prevalent conditions. It is a frequently described phenomenon that reducing agents can also have stabilizing effects on enzymes (Park et al., 2005). The influence of additives on β-glucosidases and βgalactosidases does not follow a certain pattern, and appears to be specific for each individual enzyme.
TaGH2 and TbGH2 were classified as members of glycoside hydrolase family 2 based on amino acid sequence similarities. Although GH family 2 comprises enzymes with different substrate specificities, β-glucosidase activity has not been reported in this family, yet (Henrissat, 1991). This finding proves the necessity of function-based screening to identify genes coding for proteins with unusual domain structures or unexpected activities. The substrate specificity was reported to be less relevant than structural homologies for classification of members of glycoside hydrolase families (Henrissat and Davies, 1997).
In conclusion, our study demonstrates that enzymes structurally related to GH family 2 can exhibit more diverse substrate specificities than previously predicted. Therefore, we recommend to incorporate β-glucosidases into GH family 2 and consequently to evaluate activity on glucopyranosides including cellobiose for characterization of enzymes from this family.