Introduction

Enterodiol (END) and enterolactone (ENL), known as mammalian lignans, are phytoestrogens that may be used for the prevention of menopausal syndrome, cardiovascular diseases, and cancers of the colon, breast, and prostate (Kitts et al. 1999; Meagher and Beecher 2000; Wang 2002). The human intestines contain bacteria that are capable of converting plant lignans, such as secoisolariciresinol diglucoside (SDG), to END and ENL (Clavel et al. 2006; Wang et al. 2000; 2010).

Secoisolariciresinol (SECO) was found to be the first intermediate in the biotransformation pathways from SDG to END and ENL (Raffaelli et al. 2002). To date, several bacterial strains capable of converting SDG to SECO have been reported, including Bacteroides distasonis DSM 20701T, Bacteroides fragilis DIfE-05, Clostridium cocleatum DSM 1551T, Clostridium ramosum DSM 1402T (Clavel et al. 2006), Clostridium saccharogumia (Clavel et al. 2007), and Klebsiella sp. (Wang et al. 2010). In our previous work, we isolated an active bacterial strain, Bacteroides uniformis ZL1, from human intestinal flora, which could effectively transform SDG in flaxseed to SECO, with a transformation efficiency of more than 80 % (Li et al. 2012). However, several factors limit the application of bacterial biotransformation in large scale, such as the low content of functional enzyme in the bacteria, which leads to the need of a huge system to meet the desired production. The anaerobic culture condition was also a serious challenge as it will significantly increase the cost of production. Furthermore, the repeated subculture processes may reduce the activity of the bacteria and may even lead to their genetic changes (Iguchi et al. 2002).

To solve these problems, one effective way would be to identify, clone and express at high level the functional gene(s) involved in the hydrolysis of SDG in a bacterial host. Unlike the wild-type bacterial strain that can carry out the biotransformation but tend to lose the capability following repeated sub-culturing, the enzymes expressed and isolated through the gene engineering process would be able to convert the SDG to SECO more stably and effectively (Imanaka and Aiba 1981; Kirk et al. 2002). In addition, biotransformation with the cloned enzymes will be more efficient due to much higher substrate concentration tolerance, greater catalytic rate, and easier post-processing (de Carvalho 2011; Faber 2011).

Although there are reports about the isolation and characterization of SDG-β-d-glucosidases, such as that purified from Aspergillus oryzae (Zhang et al. 2009), the sequence of the enzyme has not been determined. The purpose of this study is to identify the β-glucosidase gene of B. uniformis ZL1 that is involved in SDG hydrolysis. We cloned and sequenced the SDG-hydrolyzing β-glucosidase gene and characterized it after expressing it in a prokaryotic host.

Materials and methods

Materials

The pEASY-E1 Expression Vector and PCR-related reagents were purchased from TransGen Biotech CO., Ltd (Beijing, China). TIANAmp bacteria DNA kit, TIANprep Mini Plasmid Kit and isopropyl β-d-1-thiogalactopyranoside (IPTG) were purchased from TianGen Biotech CO., Ltd (Beijing, China). Primers for PCR reaction were synthesized by Sangon Biotech Co., Ltd (Shanghai, China). 4-Nitrophenyl β-d-glucopyranoside (pNPG) and p-nitrophenol were purchased from Sigma-Aldrich (St. Louis, MO, USA). 5-bromo-4-chloro-3-indolyl glucopyranoside (X-Glu) was purchased from BioDee BioTech Co., Ltd (Beijing, China). SDG was prepared in our laboratory from defatted flaxseed with the purity of 96 % by HPLC analysis (Li et al. 2012). Crude SDG (50 %) was purchased from Baoji FangSheng Biotechnology Co., Ltd (Shanxi, China). The HPLC-grade acetonitrile and methanol were purchased from Merck KGaA Co., Ltd (Darmstadt, German), and the purified water was purchased from Hangzhou Wahaha Co., Ltd (Zhejiang, China).

Bacterial strains and culture condition

B. uniformis ZL1 (CGMCC No.4897) was isolated from the human intestine in our previous study (Li et al. 2012). Escherichia coli strains Trans1-T1 and BL21(DE3)pLysS and ArtMedia Protein Expression medium were purchased from TransGen Biotech Co., Ltd (Beijing, China). E. coli strain Rosetta(DE3) was purchased from CoWin Bioscience Co., Ltd (Beijing, China). Anaerobic Broth medium base, Luria-Bertani (LB) nutrient agar and LB Broth medium base were purchased from Beijing Land Bridge Technology Co., Ltd (Beijing, China). Anaerobic Broth (g/L) consisted of peptone (15.0), yeast (5.0), soybean peptone (5.0), beef powder (5.0), glucose (5.0), sodium chloride (5.0), soluble starch (3.0), cysteine (0.5), potassium dihydrogen phosphate (2.5), hemin (0.005), and vitamin K1 (0.001).

B. uniformis ZL1 was cultured in Anaerobic Broth medium at 37 °C. E. coli BL21(DE3)pLysS cells were cultured in LB Broth media supplemented with 100 μg/ml ampicillin and 34 μg/ml chloramphenicol. E. coli Trans1-T1 and E. coli Rosetta(DE3) cells were cultured in LB Broth media containing 100 μg/ml ampicillin.

Cloning of the β-glucosidase gene from B. uniformis ZL1

The genome DNA of B. uniformis ZL1 was extracted using the TianAmp bacteria DNA kit according to the manufacturer’s protocol. Primers were designed based on the sequences of 24 putative β-glucosidase genes screened in the genome of B. uniformis ATCC 8492. The PCR reaction mixture contained 12.5 μl 2× TransTaq TM High Fidelity (HiFi) PCR SuperMix, 1 μl genome DNA as template, 1 μl of each PCR primer and 9.5 μl deionized water. Thermal cycles included initial denaturation at 94 °C for 5 min, 30 cycles of 94 °C for 30 s, appropriate annealing temperature for 30 s and 72 °C for 2 min, and a final extension for 10 min at 72 °C. The amplified PCR products of gene bgl1–bgl15 were cloned into the pEASY-E1 Expression vector according to the manufacturer’s protocol, and then were used to transform E. coli Trans1-T1. The positive colonies were verified by colony PCR, by using the T7 promoter primer and each of the reverse primers of bgl1–bgl15. The recombinant plasmids (pEASY-BuBGL1–pEASY-BuBGL15) derived from these positive colonies were extracted using the TIANprep Mini Plasmid Kit and then sequenced by Beijing Genomics Institute (China).

Screening for the SDG-hydrolyzing β-glucosidase

The recombinant plasmids pEASY-BuBGL1–pEASY-BuBGL15 were used to transform E. coli BL21(DE3)pLysS according to the manufacturer’s protocol. The selection of the positive recombinant colonies was performed by plate assays with X-Glu as indicator substrates (Bhatia et al. 2002; Love and Streiff 1987). The blue colonies were cultured in 3 ml ArtMedia Protein Expression medium (a kind of commercial automatic induction medium) at 37 °C overnight (Studier 2005). The induced cells were lysed according to established protocol of molecular cloning (Sambrook and Russell 2001). Then, the lysate was centrifuged, and 10 μl supernatant (crude enzyme extract) was incubated with 190 μl SDG solution (1.0 mg crude SDG/ml) in 20 mM sodium phosphate buffer (pH 7.0) at 37 °C for 1 h, then the reaction mixture was analyzed by HPLC.

Sequence analysis of recombinant β-glucosidase BuBGL8

The BLAST program at NCBI server (http://www.ncbi.nlm.nih.gov/BLAST/) was used to perform DNA/Protein sequence similarity search (Altschul et al. 1990). Motif was identified using PROSITE program (http://prosite.expasy.org/) (Sigrist et al. 2010), and conserved domain analysis was retrieved form SMART server (http://smart.embl-heidelberg.de/) (Letunic et al. 2012). Multiple sequence alignment was performed using ClustalW2 (http://www.ebi.ac.uk/Tools/msa/clustalw2/). ExPASy ProtParam tool was used to calculate the molecular weight and isoelectric point (pI) (http://web.expasy.org/protparam/) (Gasteiger et al. 2011). Signal sequence prediction was performed using SignalP 4.0 server (http://www.cbs.dtu.dk/services/SignalP/) (Petersen et al. 2011).

Homology modeling and substrate docking analysis of recombinant β-glucosidase BuBGL8

The 3-D structure model of BuBGL8 was generated using Swiss-Model server (http://swissmodel.expasy.org/) (Schwede et al. 2003). Then, the predicted structure model was optimized by using NAMD 2.6 software. According to the NAMD tutorials, the model was subjected to 5,000 steps of conjugate gradient minimization, followed by 100 ps of constrained molecular dynamics simulation with temperature at 310 K with a time step of 1.0 fs. Afterwards, the model was minimized again for 5,000 steps and equilibrated for 100 ps again with the protein moving freely. Finally, the model was simulated and equilibrated via 1 ns with a time step of 2.0 fs (Graciano et al. 2012).

The minimized and equilibrated model was used for the docking study. pNPG and SDG were docked into the active site of BuBGL8 model by using AutoDock 4.2 program (Morris et al. 1998; Morris et al. 2009).

Expression and Purification of the recombinant β-glucosidase BuBGL8

Plasmid pEASY-BuBGL8 was used to transform E. coli Rosetta(DE3) to construct the recombinant bacteria pEASY-BuBGL8/Rosetta(DE3) according to the manufacturer’s protocol. pEASY-BuBGL8/Rosetta(DE3) cells were grown in 4 L LB medium until the optical density of the culture at 600 nm (OD600) reached 0.5–0.6 (Jabbour et al. 2012). Afterwards, the expression of recombinant BuBGL8 was performed by adding IPTG (0.48 g) to a final concentration of 0.5 mM and subsequent incubation at 20 °C overnight (Ferrer et al. 2004; Graslund et al. 2008; Rabhi-Essafi et al. 2007). The induced cells were harvested by centrifugation (5,000×g, 4 °C, 30 min) and resuspended in 100 ml PBS (NaCl 137 mM, KCl 2.7 mM, Na2HPO4 10 mM, KH2PO4 2 mM, pH 7.4). Then, the cells were lysed by sonication till sample was no longer viscous, and the lysate was centrifuged (10,000×g, 4 °C, 30 min) and the supernatant (crude enzyme extract) was collected. BuBGL8 was purified from crude enzyme extract with His-tag Protein Purification Kit (CoWin Bioscience Co., Ltd, Beijing, China). The purified BuBGL8 was analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) followed by Coomassie blue R-250 staining.

Enzyme assay

The activity of recombinant β-glucosidase BuBGL8 was determined by using 1 mM pNPG as substrate. For each assay, 100 μl enzyme solution was incubated with 1.9 ml sample solution in 20 mM sodium phosphate buffer (pH 6.0) at 40 °C for 5 min. The reaction was stopped by adding 2 ml Na2CO3 (0.2M), the amount of released pNPG aglycone (p-nitrophenol) was measured at OD400 by TU-1810 spectrophotometer (Purkinje General Instrument, Beijing, China). The activity of BuBGL8 (in units per milliliter) was calculated as (C × V) ′/(V × T), where C (in micromole per liter) represents the concentration of p-nitrophenol, V ′ (in liter) represents the total volume of the reaction system, V (in milliliter) represents the volume of BuBGL8 added, and T (in minute) represents the reaction time. The amount of enzyme required to release 1 μmol of p-nitrophenol per minute was defined as one unit (U) (Yan et al. 2008).

Effects of pH and temperature on the activity of recombinant β-glucosidase BuBGL8

To investigate the effect of pH on the activity of recombinant β-glucosidase BuBGL8, it was incubated at 37 °C for 5 min in varying buffers with pH range of 3.0–9.0 (20 mM citric acid–sodium citrate, pH 3.0–5.0; 20 mM sodium phosphate, pH 6.0–7.0; 20 mM Tris–HCl, pH 8.0–9.0), respectively. The effect of temperature on the activity of BuBGL8 was determined by incubating the enzyme at temperatures ranging from 20 to 70 °C for 5 min in 20 mM sodium phosphate buffer (pH 6.0) (Quan et al. 2012).

Effects of pH and temperature on the stability of recombinant β-glucosidase BuBGL8

The effect of pH on the stability of recombinant β-glucosidase BuBGL8 was studied by pre-incubating BuBGL8 solution in the buffers described above (pH 3.0–9.0) at 20 °C for 60 min, and then the activity of BuBGL8 was measured by the standard assay method. The thermal stability of recombinant BuBGL8 was investigated by pre-incubating the enzyme in 20 mM phosphate buffer (pH 6.0) at different temperatures (30, 37, 40, and 50 °C) for 0–120 min, respectively. The activity of BuBGL8 in different time point was determined as described above.

Comparison of SDG-hydrolyzing effects of purified recombinant β-glucosidase BuBGL8 and crude enzyme extract

One milliliter BuBGL8 solution (0.035 mg/ml) and 1 ml crude enzyme extracts (0.14, 0.18, and 0.23 mg crude protein/ml) of induced pEASY-BuBGL8/Rosetta(DE3) cells in 20 mM sodium phosphate buffer (pH 6.0) were reacted with equal volume of SDG solution (2.0 mg/ml) at 30 °C, respectively. The concentration of SDG-hydrolyzing product SECO in the reaction mixture at different time points was determined by HPLC.

The effect of substrate concentration on the activity of crude enzyme extract

One milliliter SDG solutions, each at different concentrations of 1.0, 2.0, and 3.0 mg/ml, was incubated with 1-ml crude enzyme extract (0.2 mg crude protein/ml) in 20 mM sodium phosphate buffer (pH 6.0) at 30 °C, respectively. The concentration of SECO in the reaction mixture was determined by HPLC.

Kinetic analysis of recombinant β-glucosidase BuBGL8 acting on pNPG

The initial velocities of BuBGL8 using pNPG as substrate with various concentrations of 32–320 μM were measured at 40 °C and pH 6.0 for 5 min. The maximum reaction rate V max and Michaelis constant K m were determined using Lineweaver–Burk plotting method based on the Michaelis–Menten equation : 1/v = K m/V max[S] + 1/V max, where v is the initial velocity and [S] is the substrate concentration (Chen et al. 2010; Michaelis et al. 2011).

HPLC analysis

The above SDG-hydrolyzing reaction was stopped by adding an equal volume of methanol. After centrifugation at 10,000×g for 10 min, the supernatant was used for HPLC analysis. The HPLC system consisted of Agilent 1260 series HPLC apparatus (Agilent Technologies, USA), including binary pump, standard autosampler, thermostatted column compartment, diode array detector and Agilent Chemstation (D. 04. 03). The BDS HYPERSIL C18 (4.6 mm × 250 mm, 5 μm) (Thermo Fisher Scientific Inc., USA) was used to analyze the samples. The temperature of the column oven was 25 °C. The mobile phase consisted of water (A) and acetonitrile (B) in a linear gradient change from 100 % A and 0 % B to 50 % A and 50 % B in 30 min, with flow rate of 1.0 ml/min. The samples were detected at UV 280 nm.

Preparation of the SECO standard curve

A SECO stock solution of 2.36 mg/ml was prepared by exactly weighting dissolving SECO (4.71 mg) in methanol. The working solutions (0.03–2.36 mg/ml) were prepared by diluting the stock solutions with methanol. Twenty microliters of each working solution was analyzed by HPLC. The SECO standard curve was obtained by plotting HPLC peak areas on the y-axis versus the concentration (in milligram per milliliter) of SECO on x-axis, which was resulted as follows: y = 14798x − 10.74(R 2 = 0.9999), with a linearity over the concentration range of 0.03 to 2.36 mg/ml.

Nucleotide sequence accession numbers

The nucleotide sequences of bgl1–bgl15 from B. uniformis ZL1 have been deposited in GenBank nucleotide sequence database. The accession numbers of bgl1–bgl15 are from KF110993 to KF111007, respectively.

Results

Cloning of putative β-glucosidase genes and screening of the SDG-hydrolyzing β-glucosidase

Although the genome of B. uniformis ZL1 was unknown, the genome sequence data of B. uniformis ATCC 8492 was available in NCBI database, and 24 putative β-glucosidases (BuBGL1-BuBGL24), as well as their encoding genes, were annotated in the genome of strain ATCC 8492 (Table 1). Based on the fact that the strains belong to same genus may have high genetic similarities (Kuhnert and Korczak 2006), the primers for the target genes in strain ZL1 were designed by using the 24 encoding genes in strain ATCC 8492 as templates. And 15 bgl genes (bgl1-bgl15) from B. uniformis ZL1 were amplified by using the 24 pairs of primers successfully. We cloned them into the pEASY-E1 expression vector and then transformed the E. coli expression strain BL21(DE3)pLysS with the constructs. As shown in Fig. 1, the colonies of recombinant bacteria harboring pEASY-BuBGL8 displayed blue color on LB medium plate with X-Glu as indicator, while others (without pEASY-BuBGL8) formed white transparent colonies, indicating that the recombinant BuBGL8 played the roles of β-glucosidase. To further validate the activity of BuBGL8, we cultured the blue colonies in automatic induction medium and extracted the enzyme. When incubated with the crude enzyme for 60 min, SDG was completely transformed into its aglycone SECO (Fig. 2), indicating that BuBGL8 was the SDG-hydrolyzing β-glucosidase.

Table 1 Putative β-glucosidases retrieved in genome of Bacteroides uniformis ATCC 8492 in NCBI database and primers for the target genes in strain ZL1
Full size table
Fig. 1
figure 1

Plate assays with X-Glu as indicator. The colonies harboring pEASY-BuBGL8 (a) and the colonies without pEASY-BuBGL8 (b) were grown on LB agar plate with X-Glu at 37 °C overnight

Full size image
Fig. 2
figure 2

HPLC profile of crude enzyme transformation products of SDG after incubating at 37 °C for 0 min and 60 min

Full size image

Sequence analysis of the recombinant β-glucosidase BuBGL8

BuBGL8 is composed of three domains as revealed by analysis through the SMART server. Domains 1 (residues 37 to 326) and 2 (residues 359 to 638) are conserved domains of glucoside hydrolase family 3, and domain 3 (residues 669 to740) has a fibronectin type III fold with unknown function. When multiple sequence alignments of BuBGL8 with six similar known proteins were made by ClustalW2 (Fig. 3), a conserved region (S/TDW, residues 254–256) was found; the central aspartate (ASP255) was thought to be catalytic nucleophile residue (Dan et al. 2000; Li et al. 2001; Pozzo et al. 2010) and GLU462 was predicted to be catalytic acid/base residue (Bhatia et al. 2002; Pozzo et al. 2010). The molecular weight of BuBGL8 was 81.5 kDa and pI was 5.89 calculated by ProtParam tool.

Fig. 3
figure 3

Identification of the catalytic nucleophile (ASP255) and the acid/base catalyst (GLU462) of recombinant β-glucosidase BuBGL8. Multiple sequence alignment of BuBGL8 was performed with other six known proteins using ClustalW2. These proteins include: β-glucosidase from Prevotella bergensis DSM 17361 (accession number ZP_06005092.1, identity 74 %), glucan 1,4-β-glucosidase from Prevotella ruminicola 23 (accession number YP_003575940.1, identity 69 %), glycosyl hydrolase family 3 protein from Bacteroides fluxus YIT 12057 (accession number ZP_08300624.1, identity 61 %), β-glucosidase from Flavobacterium frigoris PS1 (accession number ZP_09895978.1, identity 57 %), β-glucosidase from Thermotoga neapolitana DSM 4359 (accession number YP_002534212.1, identity 50 %), glycoside hydrolase family 3 protein from Fervidobacterium nodosum Rt17-B1 (accession number YP_001409810.1, identity 49 %). The catalytic nucleophile and acid/base catalyst sites were indicated by filled triangles

Full size image

A signal peptide was found in BuBGL8 with a cleavage site between residues 19 and 20 when analyzed by SignalP 4.0 server. To heterogeneously express BuBGL8 in E. coli, the signal peptide had to be cut-off because the signal peptidase from B. uniformis did not exist in E. coli. Thus, the design of primers was based on the partial sequence of the encoding gene bgl8 without signal peptide sequence.

Homology modeling and substrate docking analysis of recombinant β-glucosidase BuBGL8

The sequence similarity search for BuBGL8 was performed in PDB archive by BLAST program, and the highest identity of 48.7 % was found with the β-glucosidase 3B from Thermotoga neapolitana (TnBgl3B, PDB ID 2X40). Thus, the 3-D structure model of the recombinant β-glucosidase BuBGL8 was built by using the crystal structure of TnBgl3B as template by Swiss-Model server, and energetically optimized by the NAMD software. The refined structure was validated using the PROCHECK and ERRAT programs. The Ramachandran plot generated by PROCHECK showed that more than 99 % of the residues were at the core region (89.0 %) or allowed region (10.8 %), indicating that the model has good stereochemical quality (Kalyani et al. 2012; Almeida et al. 2002). The overall quality factor obtained by ERRAT was 88.853, which also confirmed a good quality of the structure (accepted range for a high quality model is higher than 50) (Banerjee et al. 2009; Colovos and Yeates 1993). Therefore, the structure model of BuBGL8 was reliable for further docking studies.

Two substrates, pNPG and SDG, were taken to proceed the molecular docking with the 3-D structure model of BuBGL8. By using AutoDock 4.2 program, the structure of pNPG was docked with that of BuBGL8 (Fig. 4a). H3 and H1 of pNPG could form hydrogen bonds with O atom in carboxyl groups of D255 (2.18 Å) and E462 (2.00 and 2.15 Å), respectively (Fig. 4b). Similar interactions formed between BuBGL8 and SDG, as shown in Fig. 4c, d. For example, the O atoms in carboxyl groups of D255 made hydrogen bonding interactions with H2 (2.23 Å) and H3 (1.65 Å and 2.23 Å) of one glucose moiety in SDG, and O atom in E462 made hydrogen bonds with H1 of the glucose moiety (2.05 and 2.26 Å). Moreover, the carboxyl groups of D255 and E462 pointed to the C1 atoms of pNPG and SDG (Fig. 4a–c). Based on these findings, we predicted that D255 made nucleophilic attack on the C1 atom and E462 played the role of catalytic proton donor (Ketudat Cairns and Esen 2010). Besides, another six residues of BuBGL8 including ASP71 (only with pNPG), ARG143, LYS176, ARG187 (only with SDG), TRP256, and SER377 also participated in the formation of H-bond interactions with pNPG and SDG, which were also identified as active sites to take part in the catalysis process of BuBGL8.

Fig. 4
figure 4

Substrate docking study of recombinant β-glucosidase BuBGL8 with pNPG or SDG as ligand. a Stereo view of active-site pocket of BuBGL8 docked with pNPG, hydrogen bonds between pNPG and the contact residues are shown as dashed lines (green color), D255 and E462 are located closely with C1 atom and O1 atom of pNPG (3.93 and 4.66 Å). b Schematic representation of hydrogen bond network between pNPG and active site of BuBGL8, C and H atoms of glucose moiety of pNPG were numbered. H-bond interactions were shown as dashed line (distances are in angstrom), the intermolecular H-bonds of BuBGL8 itself are not shown but the residues are represented on an arc. c Stereo view of active-site pocket of BuBGL8 bonded with SDG. d Schematic representation of hydrogen bond network between SDG and active site of BuBGL8. The bonding interactions were represented using Discovery Studio 2.5 and ChemBioDraw Ultra 11.0

Full size image

Expression and purification of the recombinant β-glucosidase BuBGL8

The plasmid pEASY-BuBGL8 containing the encoding gene of BuBGL8 was propagated in E. coli Rosetta(DE3), with the recombinant strain being designated pEASY-BuBGL8/Rosetta(DE3). The soluble protein and insoluble protein extracts from pEASY-BuBGL8/Rosetta(DE3) cells induced by IPTG, as well as that from uninduced pEASY-BuBGL8/Rosetta(DE3) cells, were analyzed by SDS-PAGE. As shown in Fig. 5a, there was a band of about 80 kDa in the soluble protein extract from the induced cells (lane 2), which was in accordance with the calculated weight of BuBGL8 (81.5 kDa). There was no similar band in the insoluble fraction of the induced cells (lane 1). The content of BuBGL8 in the total soluble protein extract was about 20 % calculated by Quantity One v4.62. Thus, the recombinant β-glucosidase BuBGL8 was expressed in a soluble form in pEASY-BuBGL8/Rosetta(DE3) cells induced by IPTG. A volume of 1 ml BuBGL8 (2.8 mg/ml) was purified from the soluble protein extracts of induced pEASY-BuBGL8/Rosetta(DE3) cells from 4 L culture medium by affinity chromatography on Ni-Agarose Resin, which was detected as a single band (81.5 kDa) by SDS-PAGE analysis (Fig. 5b).

Fig. 5
figure 5

SDS-PAGE analysis of recombinant β-glucosidase BuBGL8 expression (a) and purification (b). a lane M molecular weight standard; lanes 1,2 insoluble and soluble extracts of pEASY-BuBGL8/Rosetta(DE3) cells induced by IPTG, respectively; lanes 3, 4 insoluble and soluble extracts of pEASY-BuBGL8/Rosetta(DE3) cells without induction. b lane M molecular weight standard, lane 1 purified BuBGL8

Full size image

Effects of pH and temperature on the activity of recombinant β-glucosidase BuBGL8

It is well known that pH and temperature are the most two important factors that may affect the activity of an enzyme. Thus, the effects of pH and temperature on the activity of the recombinant β-glucosidase BuBGL8 were investigated with pNPG (1 mM) as the substrate. As shown in Fig. 6a, BuBGL8 could transform pNPG to p-nitrophenol at the pH range of 5.0–9.0, with the best activity at pH 6.0. The effect of temperature was shown in Fig. 6b: the activity of BuBGL8 increased gradually with temperature varying from 20 °C to 40 °C under the condition of pH 6.0; when the temperature was above 40 °C, the activity of BuBGL8 begun to decrease significantly, and there was almost no activity detected when the temperature was higher than 60 °C.

Fig. 6
figure 6

Effects of pH (a) and temperature (b) on the activity of recombinant β-glucosidase BuBGL8. Values are means calculating from triplicate determinations ± SD. Relative activity (%) was calculated by assuming that activities of BuBGL8 at pH 6.0 (a) or 40 °C (b) was 100 %. The analysis of data acquired was performed using Origin Pro 8.5

Full size image

Effects of pH and temperature on the stability of recombinant β-glucosidase BuBGL8

To make sure that the enzymatic reaction was successfully developed, BuBGL8 should remain stable during the reaction period. Although BuBGL8 had optimal activity at pH 6.0 and 40 °C, the stability of BuBGL8 under this condition was unknown. Thus, the effects of pH and temperature on the stability of BuBGL8 were investigated. After pre-incubating BuBGL8 at pH range of 3.0-9.0 and 20 °C for 60 min, we measured its activity for transforming pNPG (1 mM) to p-nitrophenol at 40 °C for 5 min. As shown in Fig. 7a, BuBGL8 showed best stability at pH 6.0, with 93 % of initial activity left after incubating for 60 min. In other conditions BuBGL8 retained 86 % (pH 5.0), 60 % (pH 7.0), and 67 % (pH 8.0) of initial activity, and almost no activity was detected at pH 4.0. Similarly, BuBGL8 was pre-incubated at temperatures of 30–50 °C and pH 6.0 for 0–120 min, and then its activity was measured at 40 °C for 5 min. As shown in Fig. 7b, the temperature of 30 °C was most suitable for the stability of BuBGL8, more than 90 % initial activity was retained after 120 min. When the temperature was above 37 °C, the activity of BuBGL8 decreased quickly, with only 50 and 31 % of the initial activity remaining at 37 °C and 40 °C in 120 min, respectively; and almost no activity was detected when pre-incubated at 50 °C for just 20 min.

Fig. 7
figure 7

Effects of pH (a) and temperature (b) on the stability of the recombinant β-glucosidase BuBGL8. a. BuBGL8 was pre-incubated at various pH (3.0–9.0) at 20 °C for 60 min, then the remaining enzyme activity was detected as the standard assay method. b. Thermal stability of BuBGL8 at temperatures of 30 °C (open square), 37 °C (open circle), 40 °C (open upright triangle), and 50 °C (open inverse triangle). The enzyme activity was measured as the standard assay method. Relative activity (%) was calculated by assuming that the initial activity of the BuBGL8 was 100 %

Full size image

Comparison of the SDG-hydrolyzing effects of purified recombinant β-glucosidase BuBGL8 and crude enzyme extract

In large-scale production technology, the crude enzyme extract would be more economical than the purified enzyme for avoiding the expensive and complex protein purification process (Hou et al. 2012; Zhang et al. 2009). Therefore, the SDG-hydrolyzing activities of purified BuBGL8 protein (0.035 mg/ml) and crude enzyme extracts at three different concentration of crude protein at 0.14 mg (I), 0.18 mg (II), 0.23 mg (III), were compared, all of them were incubated with 4.0 mg/ml SDG at pH 6.0 and 30 °C. As shown in Fig. 8, among the three crude enzyme extracts, III showed similar activity to that of BuBGL8, with the SDG transformation efficiency of about 90 % after 90 min, while the activities of I and II were much lower. Hence, we could replace purified glucosidase BuBGL8 with the crude enzyme extract of induced pEASY-BuBGL8/Rosetta(DE3) cells to complete the large-scale enzymatic catalysis, the ratio of protein concentration in BuBGL8 versus that in crude enzyme extract was 1:7.

Fig. 8
figure 8

Conversion efficiency-time courses of SDG to SECO by purified BuBGL8 and crude enzyme extract of induced pEASY-BuBGL8/Rosetta(DE3) cells. SDG solutions (2.0 mg/ml) were incubated with 0.035 mg/ml BuBGL8 (open square) and crude enzyme extracts (0.23 mg crude protein/ml, open circle; 0.18 mg crude protein/ml, open upright triangle; 0.14 mg crude protein/ml, open inverse triangle). The conversion efficiency of SDG to SECO was calculated as C SECO/C ′SECO  × 100%: C SECO was the concentration of SECO in the reaction solution, C ′SECO was the concentration of SECO which was transformed completely by the original pure SDG

Full size image

The effect of substrate concentration on the activity of crude enzyme extract

Theoretically, if the amount of an enzyme remains constant, a lower concentration of substrate means fewer molecules will get attached to the enzyme active sites and leads to a slower reaction rate. On the contrary, the substrate may be an inhibitor at improper high level (Hou et al. 2012; Reed et al. 2010). Thus, the SDG-hydrolyzing activities of crude enzyme extract by taking three substrate concentrations of SDG at 1.0, 2.0, and 3.0 mg/ml were investigated. As shown in Fig. 9, the optimal reaction rate for SECO production was observed at 2.0 mg/ml SDG within a 90-min period, which was two times more than that by taking 1.0 mg/ml SDG as substrate. And when SDG concentration increased to 3.0 mg/ml, the reaction rate was slightly reduced, which means that excess substrate will show inhibitory action on the SDG-hydrolyzing activity. Thus, 2.0 mg/ml was opted as the substrate concentration for SDG in the present system, i.e., the amount ratio of crude enzyme extract (0.2 mg) versus substrate SDG (2.0 mg) was appropriately 1:10.

Fig. 9
figure 9

Content-time courses of SECO produced from SDG (1.0, 2.0 and 3.0 mg/ml) by crude enzyme extract (0.2 mg crude protein/ml) of induced pEASY-BuBGL8/Rosetta(DE3) cells

Full size image

Kinetic parameters of recombinant β-glucosidase BuBGL8 acting on pNPG

The kinetic parameters of BuBGL8 by using pNPG as substrate were determined by using the Lineweaver–Burk plotting method. The equation was created by plotting the inverse of initial velocity as a function of the inverse of the substrate concentration, which resulted as y = 1.3905x + 0.0317(R 2 = 0.9985). The substrate concentration for half-maximum reaction rate (K m) was calculated to be 44 μM and V max was 31.5 μmol min−1 mg-protein−1.

Discussion

In this study, the β-glucosidase BuBGL8 involved in SDG hydrolysis was identified from B. uniformis ZL1, its encoding gene was cloned into the pEASY-E1 expression vector and heterogeneously expressed in E. coli Rosetta(DE3). To the best of our knowledge, the sequence of SDG-hydrolyzing β-glucosidase was first identified and this is the first report using the genetic recombination method to produce the SDG-hydrolyzing β-glucosidase.

To date, several methods have commonly been used to identify functional gene or protein from microorganism, such as DNA/cDNA library screening, protein extraction and purification, bioinformatics analysis, and so on (Bhatia et al. 2002; Bork et al. 1998; Kaul and Asano 2012). Library screening may be the first choice when there was little information about the functional gene, but the library need contain a huge number of clones for being screened to find the functional gene. And even if the high-throughput screening method was developed, the procedure was still very time-consuming (Wahler and Reymond 2001). With the development of chromatography technology and protein purification equipment, the purification strategy was adopted to obtain the pure active protein from microorganism, its sequence will be determined on the basis of mass spectroscopy analysis (Domon and Aebersold 2006; Smith 2005). However, the functional protein is usually produced in low-level in natural host strains, which leads to the requirement for large abundant sources of strain cells to enrich the proteins. Furthermore, after several protein purification steps, the activity of the target protein will be lost in the procedure if it had low stability (Linn 2009). With the increasing biological data and developed bioinformatics software in recent years, bioinformatics analysis was used to identify functional genes/proteins more and more, the computational method of bioinformatics could predict the protein function prior to the screening experiments, which will significantly shorten and streamline the functional gene/protein selection progress (Lee et al. 2007; von der Lieth et al. 2006). Nonetheless, the known genome sequence of microorganism as well as rich research on the related functional gene/protein was prerequisites. In this study, both the whole genome sequencing of B. uniformis and the sequence information of known β-glucosidases are available (Bhatia et al. 2002; Opassiri et. al. 2006; Rojas and Romeu 1996), the identification of the sequence of β-glucosidase in B. uniformis was smoothly identified via bioinformatics analysis.

The stability of an enzyme is one of the influencing factors for an enzymatic biocatalysis, β-glucosidases are usually unstable at high temperature above 40 °C (Ketudat Cairns and Esen 2010; Nakajima et al. 2012; Woodley 2013). In our research, the activity and stability of purified recombinant β-glucosidase BuBGL8 were investigated simultaneously, it was found that BuBGL8 had the highest activity at the condition of pH 6.0 and 40 °C, while being unstable at 40 °C with 60 % of the initial activity left after 60 min and 30 % left after 120 min. The thermal stability study of BuBGL8 indicated that the temperature being lower than 30 °C could keep it highly active for at least 120 min, with 90 % of initial activity remained. As the enzymatic conversion of SDG to SECO in our present work need to take about 90 min, it is suggested that the enzymatic reaction could be performed at the condition of pH 6.0 and 30 °C in which both the enzyme activity and stability are better.

In this study, the enzymatic reaction could be finished within 90 min with transformation efficiency of SDG to SECO being more than 90 %, which was at least 48-fold faster than that by bacteria transformation of B. uniformis (72–120 h). And in our previous work, the reaction time of 7 days was needed to reach the maximum transformation efficiency of about 80 % for producing SECO from defatted flaxseed (containing 6.7 mg SDG/g) (Li et al. 2012). It is obvious that the enzymatic reaction rate and transformation efficiency were both improved in comparison to that of bacteria transformation. Besides that, the recombinant bacteria E. coli could grow rapidly and easily, which solved the difficulties for culturing B. uniformis in scale-up, such as the complex anaerobic culture condition and slower grown rate.

It has been confirmed that the recombinant β-glucosidase BuBGL8 also exhibited the deglycosylation activities for other natural glucosides, such as astragalin, hyperoside, isoquercetin, and rutin (data not shown). Therefore, BuBGL8 could be applied to obtain active aglycones from corresponding glucosides which are ubiquity in plant resources.