The purpose of this research was to obtain the mutant of Bacillus licheniformis alpha amylase (BLA) with an improved acid stability and elucidate the difference in catalytic mechanism under acidic conditions between wild-type and mutant BLAs.
The purpose of this research was to obtain the mutant of Bacillus licheniformis alpha amylase (BLA) with an improved acid stability and elucidate the difference in catalytic mechanism under acidic conditions between wild-type and mutant BLAs.
The stability of BLA under acid condition was enhanced through direct evolution using error-prone polymerase chain reaction. Two mutation sites, T353I and H400R, were obtained in BLA. To identify the mutation of amino acids in Thr353Ile/His400Arg related to its acid stability, single mutants Thr353Ile and His400Arg were obtained via site-directed mutagenesis. Among the resulting mutant enzymes, the kcat/Km values of the mutants Thr353Ile, His400Arg and Thr353Ile/His400Arg under pH 4·5 were 3·5-, 6·0- and 11·3-fold higher, respectively, than that of the wild-type. Thr353Ile/His400Arg exhibited stronger tolerance towards a lower pH without obvious difference in thermostability when compared with wild-type.
The results combined with three-dimensional structure analysis of mutant BLAs demonstrated that Thr353Ile/His400Arg showed an improved acid stability under low pH condition as a result of the interactions of hydrogen bonding, hydrophobicity, helix propensity and electrostatic fields.
It provides theoretical basis and background data for the improvement of acid stability in BLA by protein engineering.
Alpha amylase isolated from the mesophile Bacillus licheniformis (BLA) is more thermostable than that isolated from Bacillus stearothermophilus (BSA) and Bacillus amyloliquefaciens (BAA) despite the strong sequence similarity among the three proteins (Hmidet et al. 2008). BLA is widely used in industrial processes of starch hydrolysis for its remarkably thermostable characteristics (Machius et al. 2003). However, the application of BLA is limited by its loss of hydrolysis ability in acidic environments (Lee et al. 2006). The pH value in several industrial processes is lower than the stable pH of BLA. As a result, significant raw material and process operating costs are needed for large-scale pH adjustments (Crabb and Shetty 1999). Therefore, the development of acid stability in BLA that can operate at low pH is imperative. So far, the alpha amylase with increasing activity towards acid and alkaline pH has been engineered by protein engineering. A variety of variants of BLA was constructed by site-directed mutagenesis to investigate the relationship between the activity and pH (Nielsen et al. 1999, 2001), and the activity of alpha amylases from Bacillus sp. TS-25 and B. amyloliquefaciens was improved under low and high pH conditions, respectively, by directed evolution using error-prone PCR and DNA shuffling (Bessler et al. 2003; Jones et al. 2008). Comparatively little work has been carried out on enhancing the stability of BLA in acid environment and has focused instead on the improvement of activity at low and high pH values in BLA and other alpha amylases. The studies of acid stability in BLA thus far are still limited, resulting in little theoretical basis and background data of the structure–function relationship on BLA were obtained towards acidic pH.
Bacillus subtilis is an efficient expression host for the expression of heterologous genes (Doi et al. 1998; Liu et al. 2010, 2012). However, the B. subtilis expression–secretion system for the construction of a random mutant library and for the screening of mutant genes still has some limitations. The greatest limitation is that the transformation efficiency of B. subtilis is generally lower than that of Escherichia coli. Furthermore, B. subtilis can produce high levels of extracellular protease, which affects the stability of secreted foreign proteins.
In our previous studies, the mutation of amino acids in L134R/S320A was established via site-directed mutagenesis in BLA for acid-resistant capability (Cai et al. 2005). Further work confirmed that the electrostatic effects play a significant role in determining the stability of BLA through the three-dimensional structure analysis (Liu et al. 2008a,b). In the current study, the mutant BLAs with improved acid stability were achieved by the construction of a random mutant library based on error-prone PCR. Under the control of the pelB signal peptide of pET-22b(+) in Escherichia coli BL21 (DE3), the mutant BLAs could be directly secreted into the culture medium via isopropyl-β-d-thiogalactoside (IPTG) induction, which was convenient for the screening of the mutant genes. Subsequently, the acid-resistant Thr353Ile/His400Arg was obtained, and its kinetics and properties were studied. Meanwhile, the three-dimensional structure of the mutant was analysed to elucidate the mechanism of its acid stability under low pH conditions.
Escherichia coli DH5α, E. coli BL21 (DE3) and plasmid pET-22b(+) were preserved in our laboratory. Escherichia coli DH5α was used as an intermediate host for various plasmid constructions. Escherichia coli BL21 (DE3) was used as a host for the expression of recombinant protein. The plasmid pUAM, which contains the wild-type (WT) alpha amylase gene (amy) from B. licheniformis CICC 10181 with mature peptide to pUC19, was constructed in our laboratory. All bacteria were cultivated at 37°C in Luria–Bertani (LB) medium (Bactotryptone 10 g l−1, yeast extract 5 g l−1, and NaCl 10 g l−1). Ampicillin (30 μg ml−1) was added to the growth medium when necessary.
Error-prone PCR was performed using the recombinant plasmid pUAM containing the BLA gene as a template. Two primers, upstream (5′-cgcggatccggcaaatcttaatgggacgct-3′) and downstream (5′-cccaagctttctttgaacataaattgaaacc-3′), were used to incorporate BamHI and HindIII restriction sites (underlined), respectively. The optimal reaction mixture (100 μl) contained 0·2 mmol l−1 dATP and dGTP, 1 mmol l−1 dCTP and dTTP, 3 mmol l−1 MgCL2, 0·05 mmol l−1 MnCl2, 10 ng of template, 10 μl 10× buffer (Tiangen, Hai Dian District, Beijing, China) and 5 units (U) of Taq polymerase (Tiangen). PCR was carried out at a single cycle of 95°C for 5 min, 30 cycles of 94°C for 30 s, 55°C for 30 s and 72°C for 1 min and a single cycle of 72°C for 10 min. The PCR products digested by BamHI–HindIII were inserted into the BamHI–HindIII site of pET-22b(+) with the 6-His tag at the C terminus for purification, and the resulting plasmids were designated pET-amyM. The plasmids were then transformed into competent E. coli DH5α through electroporation. Subsequently, the plasmids extracted from the colonies of E. coli DH5α were transformed into the expression host, E. coli BL21 (DE3).
A double-layer plate was developed for the initial screening of the mutant BLAs. The lower layer of the plate was used as the nutrient culture medium for the growth of colonies, which contained LB/agar plate (pH 6·5), 1 mmol l−1 IPTG and ampicillin (30 μg ml−1). The upper layer was used as the detection medium (pH 4·5), which contained 1% (w/v) soluble starch and 0·02% (w/v) trypan blue. The standard for initial screening is the diameter ratio of transparent halo and colony. The selected mutant BLAs showing higher activity than WT at pH 4·5 were further identified through rescreening in 96-well plates. Single colony of the transformants was transferred into 96-well plates containing LB medium supplemented with 30 μg ml−1 ampicillin. The 96-well plates were incubated at 200 rev min−1 for 18 h at 37°C until the OD600 value reached 0·6, followed by the addition of isopropyl-β-d-1-thiogalactopyranoside (IPTG) at a final concentration of 1 mmol l−1 to induce the expression of recombinant enzymes. After incubation at 18°C for 24 h, cell debris was removed by centrifugation (4000 rev min−1 at 4°C for 10 min), and the supernatant was inoculated into replica 96-well plates, followed by dilution in phosphate buffer solution (0·2 mol l−1, pH 4·5) containing 1% (w/v) soluble starch. These plates will be used for initial activity assay.
The mutant BLAs, with residues T353I and H400R, were obtained by site-directed mutagenesis using an overlap extension-PCR method (Ho et al. 1989). All PCR protocols were performed using the primers listed in Table 1. Briefly, in the first step, two simultaneous PCRs were performed. One reaction contained the primer pair AMY-F and T353I-R, whereas the other reaction contained the primer pair T353I-F and AMY-R with plasmid pUAM as the template. Primers T353I-F and T353I-R are mutant primers that substitute Ile residues for Thr 353 residues. To obtain full-length mutant fragments, the two PCR products from the first step were mixed in equimolar concentrations and used as templates during the second PCR reaction with the AMY-F and AMY-R primers. The mutant gene digested by BamHI–HindIII was inserted into the same endonuclease sites of pET-22b(+) to generate the recombinant plasmid pET-amy1. In addition, H400R-F and H400R-R are mutation primers that substituted Arg residues for His400 residues. The recombinant plasmid pET-amy2 was generated through the method described earlier. Subsequently, the recombinant plasmids pET-amy1 and pET-amy2 were transformed into E. coli DH5α competent, analysed by the digestion with restriction enzymes and by DNA sequencing, and then transformed into E. coli BL21 (DE3) for protein expression.
One colony of E. coli BL21 (DE3) recombinants was inoculated into 2 ml of LB medium containing ampicillin (30 μg ml−1) and grown overnight at 37°C in a rotary shaker (200 rev min−1). One millilitre of the overnight culture was transferred into 100 ml of LB medium containing ampicillin (30 μg ml−1) at 37°C in a rotary shaker (200 rev min−1) to an OD600 of 0·6–0·8. Protein production was induced by the addition of IPTG to a final concentration of 1 mmol l−1, and cultivation was continued at 18°C in a rotary shaker (80 rev min−1) for 24 h. To purify the His-tagged WT and mutant BLAs, the culture supernatants were collected by the centrifugation at 4000 rev min−1 for 20 min and applied to Ni-NTA agarose gel column equilibrated with PBS buffer (50 mmol l−1 sodium phosphate, 0·5 mol l−1 NaCl, pH 7·4). Subsequently, the enzymes were eluted with a linear imidazole gradient of 10–400 mmol l−1 in the same PBS buffer. The purity and apparent molecular mass of the recombinant BLAs were monitored by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS-PAGE). The protein concentration was determined by the Bradford method (1976), using bovine serum albumin as a standard.
The definition of alpha amylase is described below. One unit of enzyme is the amount of amylase needed to complete the liquefaction of starch into dextrin per minute at 70°C and pH 6·0. The measurement was performed according to Chinese Industrial Standard (QB/T 2306-97). The enzyme activity was calculated as follows: X = c × n × 16·67, where X is the enzyme activity of the sample (U ml−1), c is the concentration of the control enzyme (U ml−1) corresponding to the absorbance and n is the dilution fold.
The kinetic parameters of WT and mutant BLAs were determined at different pH conditions (pH 4·5 and 6·5) using the synthetic chromogenic substrate [4, 6-ethylidene(G7)-p-nitrophenyl(G1)-α, d-maltoheptaoside (Et-G7-PNP)] from a Sigma diagnostic amylase kit (Sigma) in MOPS buffer (50 mmol l−1, pH 7·0, 1 mmol l−1 CaCl2 and 50 mmol l−1 NaCl). The samples were incubated in a suspension of Et-G7-PNP at concentrations ranging from 0 mmol l−1–5·0 mmol l−1 in the reaction mixture at 45°C, and then the ΔOD per minute was read at 405 nm. The Michaelis constants were calculated by building Lineweaver–Burk type plots.
For determining the thermostability, each purified enzyme was pre-incubated in 50 mmol l−1 sodium phosphate buffer (pH 6·0) for various times at 90°C in the absence of the substrate. To determine the pH stability, the enzyme was pre-incubated in the 50 mmol l−1 sodium phosphate buffer (pH 4·5) without substrate for various times at 90°C. After cooling the treated enzyme samples on ice for 10 min, the residual activity was measured according to the standard assay. The activity of the enzyme without pre-incubation was defined as 100%.
The three-dimensional structures of WT and mutant BLAs were yielded through homology modelling based on the structure of BLA (1VJS pdb entry) using the Swiss-Model server (Peitsch et al. 1995, 1996; Peitsch 1996; Guex and Peitsch 1997). The structure of BLA (1VJS pdb entry) was obtained from the Protein Databank (PDB; Bernstein et al. 1977).
Approximately, 5500 clones were screened to identify mutants with increased acid stability. Screening for alpha amylase-active colonies was performed using double-layer plate. Because of growth dormancy on LB/agar plate (pH 4·5), single colony of E. coli. BL21 (DE3) harbouring the pET-amyM was inoculated on the lower nutrient layer, which contained LB/agar plate (pH 6·5), 1 mmol l−1 IPTG and ampicillin (30 μg ml−1). Meanwhile, hydrolytic activity in acidic condition was detected by the upper layer which contained 1% (w/v) soluble starch and 0·02% (w/v) trypan blue. Under the control of the pelB signal peptide of pET-22b(+) in E. coli BL21 (DE3), the mutant BLAs could be directly secreted into the culture medium through IPTG induction, which was convenient for the screening of the mutant genes. Therefore, the presence of alpha amylase was indicated by the decolorization of the blue complex, resulting in a transparent halo around the colonies growing on the lower nutrient layer. The E. coli BL21 (DE3) cells harbouring pET-amy (1–2), pET-22b (3) and pET-amyM (4–21) grown on the double-layer plate were shown in Fig. 1. The amylase activity of the mutant BLAs was obvious compared with that of the transformants harbouring pET-amy and pET-22b(+). At the initial screening, 1100 mutants are exhibiting higher amylase activity than WT at pH 4·5. Further screening was carried out using 96-well plates, and the enzyme activity was assayed. Subsequently, the mutant exhibiting the highest amylase activity at pH 4·5 was selected, and nucleotide sequence alignment of the selected mutant BLA gene (amyh) with the WT (amy) revealed the three mutation sites resulting in substitution of T353I and H400R in amino acid sequence.
Confirmed by nucleotide sequencing analysis, the mutagenesis of two sites with T353I and H400R were obtained by site-directed mutation. The amplified amy1 (T353I) and amy2 (H400R) by PCR were cloned into BamHI–HindIII restriction sites of the pET-22b(+) vector to construct recombinant plasmids pET-amy1 and pET-amy2. Competent E. coli DH5α cells were transformed with the recombinant plasmids and screened on a LB/agar plate containing 30 μg ml−1 ampicillin. The plasmids that extracted from the colonies of E. coli DH5α were confirmed by double digestion and sequencing (data not shown) and then transformed into E. coli BL21 (DE3) for protein expression.
After induction with 1 mmol l−1 IPTG at 18°C for 24 h, the amylase activity of WT, Thr353Ile, His400Arg and Thr353Ile/His400Arg was detected in the supernatant of the culture medium by the expression with pET-amy, pET-amy1, pET-amy2 and pET-amyh in E. coli BL21 (DE3). The mutant BLAs revealed a band of c.53 kDa on SDS-PAGE, which was consistent with the calculated molecular weight of WT. After Ni2+ resin purification, the purified recombinant proteins formed a single band with purity of at least 95% respectively, which is convenient for functional analysis.
The kinetic parameters for the WT and mutant BLAs were determined under different pH conditions (4·5 and 6·5) using Et-G7-PNP as the substrate at 45°C (Table 2). Under a pH value of 6·5, the kcat/Km values of WT were c.1·3 times that of Thr353Ile/His400Arg. Meanwhile, the kcat/Km values at pH 4·5 of Thr353Ile, His400Arg and Thr353Ile/His400Arg were 3·5-, 6·0- and 11·3-fold higher, respectively, than those of WT. Among the recombinant BLAs, the Thr353Ile/His400Arg exhibited a highest kcat/Km value at pH 4·5 as a result of a decreased Km and an increased kcat. The kcat/Km values corresponding to the single mutants Thr353Ile and His400Arg were between those of WT and Thr353Ile/His400Arg.
|kcat(s−1)b||Km (μmol l−1)b||kcat/Km (106/s mol l−1)||kcat (s−1)b||Km(μmol l−1)b||kcat/Km (106/s mol l−1)|
|WT||41 ± 1·9||273 ± 6·9||0·15 ± 0·012||204 ± 5·1||130 ± 3·9||1·57 ± 0·08|
|Thr353Ile||123 ± 5·0||237 ± 4·5||0·52 ± 0·02||200 ± 4·4||135 ± 5·4||1·48 ± 0·06|
|His400Arg||156 ± 4·1||173 ± 6·4||0·90 ± 0·03||167 ± 4·5||201 ± 5·3||0·83 ± 0·04|
|Thr353Ile/His400Arg||239 ± 5·3||141 ± 3·6||1·70 ± 0·05||190 ± 4·6||160 ± 4·8||1·19 ± 0·05|
The purified mutant BLAs were operated optimally at 95°C. At pH 6·0, the purified enzyme solutions of WT and mutant BLAs were incubated in water bath for various times (i.e. 20, 40, 60, 80, 100 and 120 min) at 90°C. Around 80% of the enzyme activity was still detectable in WT and mutant BLAs after incubating in the water bath for 60 min (Fig. 2). The temperature stability of the mutant BLAs was similar to that of WT (Fig. 2), which demonstrated that there is no obvious difference in thermostability between the WT and mutant BLAs after mutation.
Meanwhile, Thr353Ile/His400Arg had a pH optimum of 5·5 compared with pH 6·5 for WT. The activity of WT rapidly declined, whereas that of the mutant BLAs was strongly maintained at pH 4·5 and 90°C. Compared with the WT, which retained only 27% of its initial activity after incubation at pH 4·5 and 90°C for 10 min, c.40% residual amylase activity could be still detectable in Thr353Ile/His400Arg after incubation at pH 4·5 for 60 min (Fig. 3). Moreover, the curves corresponding to the single mutants Thr353Ile and His400Arg were at an intermediate position between those for WT and Thr353Ile/His400Arg demonstrated that Thr353Ile/His400Arg was more effective in increasing enzyme stability against acidic pH than Thr353Ile and His400Arg.
The acid stability of BLA was enhanced through direct evolution using error-prone PCR. Two mutation sites, T353I and H400R, were obtained in BLA. To identify the mutation of amino acids in Thr353Ile/His400Arg related to its acid stability, single mutants Thr353Ile and His400Arg were obtained via site-directed mutagenesis. The mutant BLAs were obtained from the supernatant of the culture medium by the expression with the mutant gene in E. coli BL21 (DE3). Compared with WT, which had a rapid decline with the activity, the mutant BLAs could maintain its activity strongly in low pH, and the temperature stability of the mutant BLAs was similar to that of WT, indicating that the mutant BLAs were more acid resistant than WT without obvious difference in thermostability between the recombinant BLAs.
To elucidate the differences in catalytic mechanism under acidic pH conditions between WT and mutant BLAs (Thr353Ile, His400Arg and Thr353Ile/His400Arg), three-dimensional structures of the WT and mutant BLAs were built using the public website Swiss-Model (http://www.expasy.ch/swissmod/SWISS-MODEL.html). The overall three-dimensional structure of BLA (Fig. 4) has shown that alpha amylases consist of three domains, called A, B and C. Domain A is the central region containing a (α/β)8-barrel forming the core of the enzyme and contains the three active site residues (Fig. 4) Asp231, Glu261 and Asp328 (BLA numbering). Many studies have presented the evidence that Asp231 is the nucleophile, and Glu261 has been suggested to be the hydrogen donor to glycosidic oxygen in the reaction mechanism (Uitdehaag et al. 1999; Rydberg et al. 2002). The third conserved acid (Asp328) is believed to play an important role in this catalytic process by hydrogen bonding to the substrate and by elevating the pKa of Glu261 (Nielsen and Mccammon 2003). Domain B, containing a six-stranded β sheet, is a protrusion between the third β-strand and the third α-helix of the TIM barrel (MacGregor 1993). Domain C, which forms a Greek key motif, is located at opposite sides of this TIM barrel to domain B and contains the C terminus (Nielsen et al. 2001).
Extensive studies thus far established a fact that the chemistry of the catalytic mechanism demands on that both the proton donor and the catalytic nucleophile should be in a particular protonation state (Nielsen and Mccammon 2003), as the active site residues must be in a catalytically competent protonation state to be active for the enzyme (Nielsen and Borchert 2000). Thus, a protonated state of Glu261 is necessary, while Asp231 must remain deprotonation state. It is generally assumed that the acidic limb of the pH–activity profile is determined by the titration of the catalytic nucleophile and that the basic limb reflects the titration of the catalytic proton donor (Fig. 5) by the bell-shaped pH–activity profiles exhibited by the alpha amylases (Takagi et al. 1971; Qian et al. 1994; Strokopytov et al. 1995).
Two residues corresponding to Thr353 and His400 in BLA were replaced by Ile and Arg, respectively. To identify the determinants of protein stability at the mutation sites, the details of the BLA three-dimensional structure were analysed at two crucial positions 353 and 400.
Position 353 Thr353 is the last residue of α-helix 7 in the interior of domain A. Compared with Thr, which has a low propensity with forming α-helix, Ile having high helix propensity is statistically favoured in the secondary structure. It is known that protein denaturation is a complex process, and most energy is consumed by breaking the contacts to neighbouring residues in an α-helix or a β-sheet. The replacement of Thr with a low α-helix propensity to Ile having high helix propensity is beneficial to stabilize the domain A which forms the core of the molecule, resulting in the improvement of protein overall stability at extreme pH values.
Moreover, it is generally assumed that in proteins hydrophilic residues are not favourable at interior sites for protein stability. Although two hydrogen bonds (Thr353 ↔ Ala349 and Thr353 ↔ Trp411) were undermined (Fig. 6), the Ile to Thr substitution that replaced a hydrophilic side chain at this interior site with an apolar group is favourable to increase the hydrophobic nature in the interior, resulting in the improvement of protein stability.
Position 400 It is assumed that alpha amylase catalysis is limited at low pH conditions by the protonation of the nucleophile (D231) and at high pH conditions by the deprotonation of the hydrogen donor (E261); moreover, the relationship between the activity of enzyme and pH is determined by the pKa values of these two active site groups (Kyte 1995). The pKa value of a residue can be changed by altering the electrostatic field as a result of both desolvation effects and other charged groups and dipoles in the protein. Mutations that aim at changing the pKa value of a titratable group by changing desolvation effects and dipole interaction should be placed in the vicinity of the titratable group because of the short-range of these energies. Because the interaction energy between a titratable group and a unit charge is less dependent on distance, mutations that introduce or remove unit charges can be placed adjacent and distant the titratable group. However, the mutations close to the catalytic residues most likely have a negative effect on enzyme activity (Wind et al. 1998; Nielsen et al. 1999). So the mutations aiming at perturbing the pKa values of the residue by introducing or removing unit charges can be placed further away from the residue. Furthermore, placing a titratable group in a positive environment would decrease its pKa (Nielsen et al. 2001). At most physiological pH values, Arg with a guanidyl can be expected to bear more positive charges than the residue of His. Although the positive charges are far from nucleophiles (D231), they can still stabilize the negative charge on this aspartate residue to stabilize its deprotonated form and thereby reduced its pKa. Therefore, the mutation site of H400R would produce a shift of the acidic limb to more acidic values. Likewise, the mutation site of H400R could also stabilize the negative charge on the glutamate residue (E261) to a decrease of the pKa of Glu261, resulting in a acidic shift of the pH–activity profile.
Generally, although the hydrogen bond between positions 400 and 414 was undermined (Fig. 7), the substitution of His400 to Arg, replacing a residue at solvent-exposed sites by a more hydrophilic residue would increase the surface hydrophilicity, causing the improvement of the protein stability.
The analysis of enzymatic properties combined with three-dimensional structure of mutant BLAs showed that the mutation site of H400R was more effective for increasing BLA stability against acidic pH than the mutation site of T353I, and the combination of these two effects obtained for mutation sites of H400R and T353I would result in a cumulative effect observed in the double mutant T353I/H400R. The results of the present study could provide theoretical basis and background data for the improvement of acid stability in BLA by protein engineering.
This work was supported by the National Natural Science Fund (31101219) and the programme for Changjiang Scholars and Innovative Research Team in University (IRT1166) and National High-Tech Research and Development Plan (‘863’ Plan) (no. 2011AA100905-4).