Enhancing the methanol tolerance of Candida antarctica lipase B by saturation mutagenesis for biodiesel preparation

Methanol tolerance of lipase is one of the important factors affecting its esterification ability in biodiesel preparation. By B factor indicated prediction of Candida antarctica lipase B (CalB) surface amino acids, eight sites (Val139, Ala146, Leu147, Pro218, Val286, Ala287, Val306, and Gly307) with high B value indicating more flexibility were chosen to perform saturation mutagenesis. High-methanol-tolerant variants, CalB-P218W and -V306N, created larger haloes on emulsified tributyrin solid plate including 15% (v/v) methanol and showed 19% and 31% higher activity over wild-type CalB (CalB-WT), respectively. By modeling, a newly formed hydrogen bond in CalB-V306N and hydrophobic force in CalB-P218W contributing more stability in protein may have resulted in increased methanol tolerance. CalB-P218W and -V306N transesterified the soybean oil into biodiesel at 30 °C by 85% and 89% yield, respectively, over 82% by CalB-WT for 24 h reactions. These results may provide a basis for molecular engineering of CalB and expand its applications in fuel industries. The as-developed semi-rational method could be utilized to enhance the stabilities of many other industrial enzymes.


Introduction
Dramatically needing renewable and greener fuels was triggered by depletion of fossil fuels and rising climate warming (Unglert et al. 2020). As a biodegradable, renewable, and non-toxic fuel, biodiesel (fatty acid methyl esters, FAMEs) acquired from transesterification of vegetable or microalgae oils and fats has received increasing interest in the last few decades (Srivastava et al. 2018). Compared with harsh alkali or acid conversion at high temperatures, enzymatic transesterification has emerged as an eco-friendly approach for biodiesel production with high outcomes (Panichikkal et al. 2018).
Candida antarctica lipase B (CalB) is one of the most researched lipases for biodiesel preparation (Bhan and Singh 2020). Mature CalB is composed of 317 amino acid residues, which constitute an α/β-hydrolase fold consisting of 10 α-helixes and 9 β-strands in moisture surrounding with high stability (Benjamin et al. 2015). A universal lipase always shows interfacial activation due to the existence of a lid, either a loop or a helix, shielding the active site (Ser-His-Asp catalytic triad, Ser 105 -His 224 -Asp 187 in CalB) from the solvent and thus regulating the substrate accessibility to the catalytic cavity (Heidi et al. 2018). CalB does not show any significant interfacial activation due to the absence of a traditional lid, but its α-helix 5 (works like a lipase lid) and α-helix 10 are shown at atomic resolution to be responsible for active site closing (Benjamin et al. 2015). The reaction mixture for transesterification is hydrophobic and has a high concentration of methanol or ethanol to get a high conversion rate of oil to FAMEs. In such an environment, the biocatalyst's structure tends to denature (Lotti et al. 2015). For this reason, searching for high methanol tolerant lipases from the natural world (Malekabadi et al. 2018;Park et al. 2019) or improving the existing lower-methanol-tolerant lipases into hyper-ones has become meaningful to achieve 1 3 22 Page 2 of 11 competitive biodiesel and cost-effective production technology (Syal et al. 2017).
There are a few works conducted on improving lipase methanol stability through protein engineering. Directed evolution method using error-prone polymerase chain reaction (PCR) was performed to enhance methanol tolerance property of Proteus vulgaris lipase for biodiesel synthesis in harsh conditions (Gupta et al. 2020). By directed evolution using iterative saturation mutagenesis (ISM), methanol tolerance of Thermomyces lanuginosus lipase has been enhanced for efficient biodiesel production from waste grease (Tian et al. 2017). Saturation mutagenesis is a semirational approach to change the enzyme characteristics, such as activity (Zheng et al. 2016), stability (Liu et al. 2021), specificity (Li et al. 2020b), and enantioconvergency (Hu et al. 2018).
As an important industrial enzyme, CalB was engineered with a few outstanding enzymatic properties (Brito e Cunha et al. 2019). CalB was divided into two regions, substratebinding and solvent-affecting regions. Its variant (V139E, A92E) was yielded by integrating the activity and stability by multiple-site mutagenesis (Yagonia et al. 2015). Following a structure-based engineering strategy, CalB active site was redesigned and constructed by saturation mutagenesis to improve the catalytic efficiency of poly(ε-caprolactone) synthesis and size of polymer products (Montanier et al. 2017). To date, different 3-D structures of CalB in different conformations (Benjamin et al. 2015) and at different resolutions [0.91 Å PDB code 5A71 (Benjamin et al. 2015), 1.5 Å PDB code 4K6G (Xie et al. 2014), 2.0 Å PDB code 4ZV7 (Strzelczyk et al. 2016), 2.5 Å PDB code 1TCC (Uppenberg et al. 1994)] have been resolved. These should provide a unique opportunity to improve its capacity by in silico design. In this work, the methanol tolerance of CalB was enhanced using a semi-rational approach of B-factor indicated prediction, saturation mutagenesis, and screening on tributyrin solid plate containing methanol. Application of high-methanol-tolerant variants of CalB for biodiesel preparation and molecular basis for methanol tolerance rising was also investigated.

Strains, culture media and chemicals
The recombinant vector, pET28a-CalB, containing CalB gene (GenBank accession no. Z30645) was constructed and preserved in Jiangsu Provincial Engineering Laboratory for Biomass Conversion and Process Integration, Huaiyin Institute of Technology, China. The vector pET28a (Novagen, Madison, WI, USA) was used to express gene CalB and its mutants in Escherichia. coli BL21(DE3), which was cultured in Luria-Bertani (LB) medium (Hassan et al. 2021). The Pfu DNA Polymerase and T4 DNA ligase were purchased from Sangon Biotech (Shanghai, China). All other chemicals of high grade were obtained from commercial sources.

Design of saturation mutagenesis
The crystal structure of lipase B from Candida antarctica (PDB code 5A71, chain B) was used as the original CalB 3-D structure (Benjamin et al. 2015) and visualized using the PyMOL software (http:// pymol. org). The B factor values, namely atomic displacement parameters, of amino acids were calculated using the B-FITTER program (Watanabe et al. 2021). The B factor values indicate the degree of flexibility of residue. The high B factor value indicates being more flexible; conversely, it means being less flexible.

Saturation mutagenesis
By B factor indicated prediction of CalB surface amino acids, eight sites (Val 139 , Ala 146 , Leu 147 , Pro 218 , Val 286 , Ala 287 , Val 306 , and Gly 307 ) with high B value indicating being more flexible were chosen to perform saturation mutagenesis. Using pET28a-CalB as a template, a fulllength expression plasmid with a mutant CalB gene was amplified by long-distance inverse PCR (Fig. 1

Screening of variants
The full length PCR amplification DNA fragments with two blunt ends fused automatically (cyclized) from the fore-end to the tail-end by catalysing T4 DNA ligase ( Fig. 1) and transformed into E. coli BL21(DE3). The transformants growing on LB solid plate were picked into liquid LB, cultivated, and induced by isopropyl-βd-thiogalactopyranoside (IPTG) to produce recombinant wild-type CalB (CalB-WT) or CalB variants. The crude lipases were obtained by ultrasonic treatment of the cultivated cells following centrifugation. The crude CalB mutants were screened preliminarily on a tributyrin (Sangon, Shanghai, China)-methanol solid plate for acquiring high-methanol-tolerant variants. The screening plate was prepared according to the method reported previously with slight modifications as follows: 3.75-g agar, 7.5-mL tributyrin, 19.0-mL 3% (w/v) polyvinyl alcohol, and 221-mL 50 mM KH 2 PO 4 -Na 2 HPO 4 buffer (pH 7.0) were added into a 500-mL flask and blended under microwave heating for homogeneity (Tan et al. 2011). The mixture was cooled to ~ 50 °C, then added 15% (v/v) methanol, immediately poured into φ12-cm plates and cooled down at room temperature. The φ0.6-cm filter paper rounds were put upon the tributyrin-methanol solid plate, and each 5.0 μL CalB sample was added on each filter paper round, followed by incubating at 35 °C for 24 h. The heat-inactivated CalB sample was used as the control. The methanol tolerance changing of the CalB variant was estimated by the difference value of the halo diameter (clear zone encircled the filter paper round) minus CalB-WT. There is a linear relationship between the logarithm of enzymatic activity and the halo diameter (Wu et al. 2000). The linear relationship in this research can be expressed by the equation of lgA = 0.1216 × D + 0.0415 (R 2 = 0.998), in which A is enzyme activity (U/mL), while D is the diameter difference value of halo and filter paper round (mm).

Purification of CalB-WT and CalB variants
After inducing with 0.25 mM IPTG at 25 °C for 8 h, the transformant E. coli BL21(DE3) cells were harvested and suspended in the same volume of 50 mM KH 2 PO 4 -Na 2 HPO 4 buffer (pH 7.0) and followed by ultrasonic processing. The resulting supernatant was concentrated by ultrafiltration and purified using a Ni 2+ -chelating agent (Novagen, Madison, WI, USA) (Li et al. 2019). Imidazole at low concentration (20 mM in this research) was used in the binding and wash buffer to minimize the binding of unwanted host cell proteins. Imidazole at optimized high concentration (100 mM in this research) was applied to elute the enzyme (target CalB protein). The purified CalB was concentrated in an aqueous solution and quantified with a BCA protein assay kit (Real-Times, Beijing, China).

Partial enzymatic characteristics analysis
The activities of high-methanol-tolerant CalB variants were quantified according to the method described previously with slight modifications (Tan et al. 2014). Briefly, 1 vol of tributyrin and 7 vol of 1.0% (w/v) polyvinyl alcohol in 50 mM KH 2 PO 4 -Na 2 HPO 4 buffer (pH 6.5) were blended (10,000 rpm) in a high-speed homogenizer till homogeneity was achieved under strong mechanical shear force. Then, 9 mL of the emulsified tributyrin and 1 mL of the suitably diluted enzyme were mixed, incubated at 30 °C for 10 min, and terminated by adding 15 mL of ethanol. The heat-inactivated CalB instead of the active one was used as the control. The amount of free butyric acid released from tributyrin was measured by titration using 50 mM NaOH. One unit (U) of lipase activity corresponds to the amount of enzyme able to release 1 μmol/min of fatty acid under the standard assay conditions (at pH 6.5 and 30 °C for 10 min). Corresponding to U/mg, it was defined as the active lipase amount (U) per milligram of purified CalB protein. All data of enzyme activities or relative ones were expressed as the mean ± standard deviation from three independent experiments or parallel measurements. Statistical comparison was made using Student's t-test. The level of statistical significance was defined as P < 0.05 or P < 0.01.
Determination of temperature and pH optimum and stability of CalB was performed according to the method described previously with modifications (Tan et al. 2014). The temperature optimum of CalB was determined at a temperature range of 20-50 °C under the standard conditions for lipase activity assay. To measure its temperature stability, CalB aliquots were incubated at different temperatures of 20-50 °C for 1.0 h before evaluating the retained activity of CalB under the standard assay conditions. In this work, the thermostability was defined as the temperature at or below which the residual activity of CalB was over 85% of its initial activity. The pH optimum of CalB was estimated using the emulsified tributyrin prepared with 50 mM KH 2 PO 4 -Na 2 HPO 4 buffer (pH 5.0-8.0), at the optimum temperature for 10 min. For its pH stability estimating, CalB aliquots were incubated at 30 °C and different pH values of 50 mM KH 2 PO 4 -Na 2 HPO 4 buffer (pH 4.5-8.5) for 1.0 h before assessing the retained activity of CalB under the standard assay conditions. The pH range over which the residual activity of CalB was more than 85% of its initial activity was defined as the range of its pH stability.

Methanol tolerant assay
Aliquots of purified CalB were diluted in different concentrations of methanol buffer (50 mM KH 2 PO 4 -Na 2 HPO 4 , pH 6.5) solution (0-35% of methanol), and incubated at 30 °C for 24 h on a rotary incubator (100 rpm). The residual activity was measured under the standard assay conditions. CalB diluted in 50 mM KH 2 PO 4 -Na 2 HPO 4 (pH 6.5) without methanol was used as a control.

Biodiesel preparation
CalB-WT and high-methanol-tolerant variants (CalB-P218W and -V306N) were used for biodiesel preparation. To prepare biodiesel, 20.0-g soybean oil, 3.7 mL methanol (kept 4:1 molar ratio of methanol to oil), 1-mL water (5% of oil weight), and 100 mg CalB were added to a 250-mL stoppered flask. The transesterification reactions were catalyzed by CalB at 30 °C for 24 h on a rotary incubator (200 rpm). Each 100 μL sample was withdrawn from the reaction mixture at specified time points and centrifuged to collect the upper phase of FAMEs for quantitative analysis (Quayson et al. 2020).

Biodiesel analysis by gas chromatography
GC-2030 gas chromatograph (Shimadzu Corp., Kyoto, Japan) equipped with a capillary column (30 m × 0.25 mm × 0.25 μm, INNOWAX, Angellym, USA) and a flame ionization detector was used for quantitative analysis of FAMEs. Each 1.0 μL sample was injected into the GC using a split mode with a split ratio of 1:30 for analysis. High purity nitrogen was used as carrier gas at a flow rate of 1.0 mL/min. The column temperature was raised from 200 to 240 °C at a rate of 4 °C/min, and then kept at 240 °C for 5 min (Li et al. 2020a). The injector and detector

Analysis of methanol stable variants
The crystal structure of lipase B from Candida antarctica (PDB code 5A71) was used as the modeling template for 3-D structure predictions of the CalB-P218W and -V306N. The 3-D structures of CalB-P218W and -V306N were homologically modeled using the MODELLER 9.9 program (http:// salil ab. org/ model ler/) and visualized by using the PyMOL software. The alignment of 3-D structures was carried out using the PyMOL software.

Saturation mutagenesis
The average B factor value of all residues in CalB was 18.1. The amino acid residues of CalB with top-16 high B factor values are shown in Table 2. The candidate sites for saturation mutagenesis should have high B factor values and be on the molecular surface, not at the strand ends or in the catalytic cavity. Amino acid residues at the strand end usually have high B factor values for their easily and freely wiggling, though the residue distributions around the end limit the end's moving (Duan et al. 2010). Changing amino acid residues in or adjacent to the catalytic cavity may decrease enzyme activity (Long et al. 2016

Screening of high-methanol-tolerant variants
For screening of target variants after saturation mutagenesis for each position, 94 colonies, in theory, should be picked for 95% coverage (Reetz 2006). After saturation mutagenesis, mutants on the plate were picked, cultivated, induced by IPTG, treated by ultrasonic, and finally screened by adding crude CalB variants to filter paper rounds on emulsified tributyrin solid plate including 15% (v/v) methanol. Each library has gotten more than 100 variants for screening on an agar plate; most of them showed lacklustre methanol stability performance compared to CalB-WT. Among nearly 900 variants of 8 libraries, 11 variants showed remarkable methanol stabilities. Two outstanding variants that finally confirmed CalB-P218W and -V306N created larger haloes of φ7.72 mm and φ8.06 mm, and showed 19% and 31% higher activity over CalB-WT, respectively. One-site saturation mutagenesis at one amino acid site yields only one site mutant and usually a limited low level changing of methanol tolerance. It is worthy of integrating different low-level positive mutants to varying positions into one CalB molecule to obtain high-methanol-tolerant variants by site-directed mutagenesis or iterative saturation mutagenesis (ISM). As a result of applying ISM, a pronounced increase in thermostability of maize endosperm ADP-glucose pyrophosphorylase has been reported (Boehlein et al. 2015).

Partial enzymatic properties
The methanol-tolerant variants (CalB-P218W and -V306N) and original lipase (CalB-WT) were purified by Ni 2+ -chelating affinity chromatography for analyzing their enzymatic properties. By SDS-PAGE analysis, the purified CalB-P218W and -V306N showed protein bands of about 39 kDa, similar to the theoretical M.W. (38.7 kDa) of His-tag added CalB. The specific activity of purified CalB-P218W and -V306N towards tributyrin under the standard assay conditions were 81.5 ± 2.6 and 86.7 ± 3.2 U/mg, respectively. They were like the specific activity of CalB-WT (82.1 ± 2.3 U/mg) and slightly more than the specific activity of CalB reported previously (~ 30 U/mg, Larsen et al. 2008). The temperature optima (100% relative activity) of CalB-P218W and -V306N at pH 6.5 were all 30 °C same as CalB-WT (Fig. 3a). After incubating at different temperatures for 1.0 h, the CalB-WT, -P218W, and -V306N retained over 85% of their original activities at 35 °C or below but declined rapidly over 35 °C (Fig. 3b). The purified CalB-WT, -P218W, and -V306N exhibited higher relative activities over a pH range of 5.5-8.0 (measured at 30 °C). Their pH optima (100% relative activity) were at pH 6.5 (Fig. 4a). Relative activities or relative residual activities of CalB-WT and its variants at the same temperature are not statistically different (P > 0.05). Incubated at different pH values (4.5-8.5) for 1.0 h, CalB-WT, -P218W, and -V306N were highly stable (more than 85% of their original activities) over a wide pH range of 5.0-8.0 (Fig. 4b). Relative activities or relative residual activities of CalB-WT and its variants at the same pH are not statistically different (P > 0.05), except relative activities at pH 8 are statistically different (P < 0.01). CalBs expressed in different host cells, including Escherichia coli (Tassel et al. 2020), Pichia pastoris (Larsen et al. 2008), and Aspergillus oryzae (CalB on the resin of Novozym 435) (Endo et al. 2018) exhibit similar effects by temperature and pH. As an important biocatalyst for transesterification, CalB could act in almost neutral pH and room temperature Fig. 3 Effects of temperature on CalB-WT and its variants; a activities of CalB-WT and its variants at different temperatures, and b thermostabilities of CalB-WT and its variants. Reaction conditions: for a 9 mL of the emulsified tributyrin (pH 6.5) and 1 mL of suitably diluted CalB, at 20-50 °C for 10 min, and for b 9 mL of the emulsified tributyrin (pH 6.5) and 1 mL of suitably diluted CalB, at 20-50 °C for 1 h, followed by at 30 °C for 10 min. (closed square) CalB-WT; (closed diamond) CalB-P218W; (closed triangle)  conditions that indicate more eco-friendly and cost-saving biodiesel production.

Methanol tolerance analysis
With an increase of methanol concentration, the stabilities of CalB-WT, -P218W, and -V306N were decreased (Fig. 5). After 24-h incubation at different methanol concentrations, the CalB-V306N retained over 85% of its original activity in 20% of methanol or below but declined rapidly over 25% of methanol (Fig. 5). Relative residual activities of CalB-WT and its variants at the same methanol concentration are not statistically different (P > 0.05) when methanol concentration is below 5%, and are statistically different (P < 0.01) when methanol concentration is between 10 and 35%. The CalB-P218W retained over 85% of its original activity in 15% of methanol or below but declined rapidly over 20% of methanol. However, CalB-WT was stable in 10% of methanol or below, quickly declined over 15% of methanol. CalB-V306N showed more stability than CalB-WT and -P218W in the presence of methanol. For methanol concentrations above 10% (vol/vol), the catalytic activity of CalB in the methanol-toluene mixture was at 30% of the maximum activity (Kulschewski et al. 2013).

Biodiesel preparation
By a quantitative analysis of FAMEs collected at specified reaction time points, biodiesel yields increased along with the extension of reaction time (Fig. 6). FAME yields of CalB-WT and -P218 at the same reaction time are not statistically different (P > 0.05). FAME yields of CalB-WT and -V306N at the same reaction time are not statistically different (P > 0.05) when reaction time is below 4 h and are statistically different (P < 0.05) when reaction time is between 8 and 24 h. CalB-P218W and -V306N transesterified the soybean oil into biodiesel by 85% and 89% yield, respectively, over 82% by CalB-WT for 24 h reactions. Biodiesel preparations by CalBs at different statuses have been reported. CalB in an Aspergillus oryzae whole-cell biocatalyst afforded a methyl ester content of more than 90% after 6 h with the addition of 1.5 M equivalents of methanol Fig. 4 Effects of pH on CalB-WT and its variants; a activities of CalB-WT and its variants in different pH buffer, and b pH stabilities of CalB-WT and its variants. Reaction conditions: for a 9 mL of the emulsified tributyrin (pH 5-8) and 1 mL of suitably diluted CalB, at 30 °C for 10 min, and for b 1 mL of suitably diluted CalB (pH 4.5-8.5), at 30 °C for 1 h, followed by adding to 9 mL of the emulsified tributyrin (pH 6.5), at 30 °C for 10 min. (closed square) CalB-WT; (closed diamond) CalB-P218W; (closed triangle) CalB-V306N

Fig. 5
Methanol tolerance of CalB-WT and its variants. Reaction conditions: CalB diluted in methanol-buffer (pH 6.5) solution (0-35% of methanol), at 30 °C for 24 h, followed by 9 mL of the emulsified tributyrin (pH 6.5) and 1 mL of above-mentioned diluted CalB, at 30 °C for 10 min. (closed square) CalB-WT; (closed diamond) CalB-P218W; (closed triangle) CalB-V306N (Adachi et al. 2013). Covalently immobilized CalB on carboxylated single-walled carbon nanotubes using 3.7 wt% of the enzyme to sunflower oil converted the oil in 83.4% yield after 4 h at 35 °C (Bencze et al. 2016). Immobilized CalB (Novozym 435) using 4 wt% of the enzyme to soybean oil and methanol, preincubated in methyl oleate for 0.5 h, could transesterify soybean oil into biodiesel in 97% yield within 3.5 h by stepwise addition of 0.33 molar equivalent of methanol at 0.25-0.4 h intervals (Samukawa et al. 2000). To avoid the negative effect of a high concentration of methanol on lipase, adding methanol to the reaction mixture at intervals is a valuable strategy for biodiesel production (Lee et al. 2009). Dissolving methanol and soybean oil in the appropriate amount of solvent can also reduce the harmfulness of methanol to an enzyme (Tan et al. 2021). CalB co-immobilized with Thermomyces lanuginose lipase (TLL) (CalB: TLL ratio 2.1) converted palm oil to FAMEs with 94% yield within 24 h, in which 0.15% weight of enzyme to oil emerged synergistic effect between CalB and TLL (Shahedi et al. 2021). There is a high expectancy of CalB-P218W and -V306N to convert oil into biodiesel in a high yield by an integrated strategy of immobilizing lipase, pretreating lipase, adding methanol at intervals though the lipases are high-methanol-tolerant, varying molar ratio of methanol to oil, and synergistic effect with other lipases (Shahedi et al. 2019).

Molecular basis for methanol tolerance
The open and closed conformations of CalB variants were modeled using chain A (open conformation) and chain B (closed conformation) of CalB-WT crystal structure (PDB code 5A71) as templates, respectively. By 3-D structure alignment, the difference between the open and closed conformations of CalB-WT, and these conformation differences in modeled CalB variant (CalB-P218W or -V306N) mainly existed in a state of α-helix 5 and 10. Either for open or closed conformations between CalB variant and CalB-WT, the difference mainly existed in the site of the mutant amino acid (P218W or V306N), and its surrounding. In modeled CalB-P218W, the side chain of Trp 218 oriented to Val 194 , leading to drastically shrinking of the shortest distance between site 218 and Val 194 from 7.75 to 3.42 Å, where the final distance was still larger than the maximum one for hydrogen bond formation (Fig. 7a, b). Hydrogen bonds make favorable contribution to protein stability, and their lengths are usually 2.5-3.3 Å (Pace et al. 2014). Whereas the indole ring of Trp 218 brought an enhanced hydrophobic force and may result in a CalB-P218W methanol tolerance increase. In modeled CalB-V306N, the side chain of Asn 306 was much longer and oriented to Thr 316 , which made the distance between site 306 and 316 shorter to 2.88 Å and more beneficial for hydrogen bond formation (Fig. 7c, d). It also promoted β-sheet folding in strands consisting of Gly 307 through to Val 315 , making the local molecular surface more rigid. The introduced polar interactions or hydrophobic forces should be responsible for the stability elevation of lipases (Dror et al. 2015) and CalB variant (CalB-V306N or -P218W) in methanol solution.
It is necessary and valuable to consider the dielectric constant in the discussion of hydrogen bonds formation. The dielectric constant of water, methanol, and soybean oil are 80.3, 33.0, and 2.9, respectively, at room temperature (Shcherbakov and Artemkina 2013;Gabriel et al. 1998;Muley and Boldor 2013). The dielectric constant of a mixture is complicated and related to each component, e.g., water-in-oil-emulsion is 24.2, which is intermediate between water and oil (Vlachou et al. 2020). The methanol-containing media are emulsified tributyrin solid plate including 15% (v/v) methanol for high-methanol-tolerant variant screening. The medium's dielectric constant is theoretically slightly lower than the water-in-oil-emulsion cause of added 15% (v/v) methanol (Khaled et al. 2016). A water molecule layer tends to overlay the solute protein even in the blend containing a small amount of water for activating enzymes for transesterification (Shiraga et al. 2016). However, the decreased dielectric constant has a negative impact on intermolecular hydrogen-bond formation on protein surface between protein and solvents, including water and methanol, as well as on intramolecular hydrogen-bond. The presumable additional hydrogen bond in CalB-V306N apparently led to protein stabilization even in a methanol-containing reaction medium of lower dielectric constant (compared with water). Thus, CalB variant (CalB-P218W or -V306N) has improved methanol tolerance in 15% (v/v) methanol-containing media, but its transesterification rate has not increased sharply.

Conclusions
In this work, eight amino acid sites on the molecular surface of Candida antarctica lipase B (CalB) with high B values indicating being more flexible were chosen to perform saturation mutagenesis. High-methanol-tolerant variants, CalB-P218W and -V306N, created larger haloes on emulsified tributyrin solid plate including 15% (v/v) methanol and showed 19% and 31% higher activity over CalB-WT, respectively. A newly formed hydrogen bond in CalB-V306N and hydrophobic force in CalB-P218W might have enhanced their methanol tolerance. CalB-P218W and -V306N transesterified the soybean oil into biodiesel by 85% and 89% yield, respectively, over 82% by CalB-WT for 24-h reactions.
Author contributions ZT performed the experiments, wrote the original manuscript, and reviewed the manuscript. XL checked the original data. HS performed the data analysis. XY conceived and designed the experiments. XZ contributed analysis tools. MB reviewed the manuscript. MMO reviewed the manuscript. All authors read and approved the final manuscript.  In modeled CalB-P218W, the side chain of Trp 218 oriented to Val 194 , leading to drastically shrinking of the shortest distance between site 218 and Val 194 from 7.75 to 3.42 Å where it enhanced hydrophobic force. While in modeled CalB-V306N, the side chain of Asn 306 was much longer and oriented to Thr 316 , which made the distance between site 306 and 316 shorten to 2.88 Å in which it became more beneficial for hydrogen bond formation and also promoted β-sheet folding in strand consisted of Gly 307 through to Val 315