Journal of Molecular Graphics and Modelling 105 (2021) 107897 Contents lists available at ScienceDirect Journal of Molecular Graphics and Modelling journal homepage: www.elsevier.com/locate/JMGM Enhancing the stability of Geobacillus zalihae T1 lipase in organic solvents and insights into the structural stability of its variants Jonathan Maiangwa a, b, d, Siti Hajar Hamdan c, Mohd Shukuri Mohamad Ali c, Abu Bakar Salleh d, Raja Noor Zaliha Raja Abd Rahman e, Fairolniza Mohd Shariff f, Thean Chor Leow a, d, f, * a Department of Cell and Molecular Biology, Enzyme Microbial Technology Research Center, Faculty of Biotechnology and Biomolecular Science, Universiti Putra Malaysia Serdang, 43400, UPM Serdang, Selangor, Malaysia Department of Microbiology Kaduna State University, Nigeria c Department of Biochemistry, Enzyme Microbial Technology Research Centre, Faculty of Biotechnology and Biomolecular Science, Universiti Putra Malaysia Serdang, 43400, UPM Serdang, Selangor, Malaysia d Enzyme Microbial Technology Research Centre, Faculty of Biotechnology and Biomolecular Science, Universiti Putra Malaysia Serdang, 43400, UPM Serdang, Selangor, Malaysia e Department of Microbiology, Enzyme Microbial Technology Research Centre, Faculty of Biotechnology and Biomolecular Science, Universiti Putra Malaysia, 43400, UPM Serdang, Selangor, Malaysia f Institute of Bioscience, 43400, UPM Serdang, Universiti Putra Malaysia Serdang, Selangor, Malaysia b a r t i c l e i n f o a b s t r a c t Article history: Received 24 August 2020 Received in revised form 4 March 2021 Accepted 4 March 2021 Available online 10 March 2021 Critical to the applications of proteins in non-aqueous enzymatic processes is their structural dynamics in relation to solvent polarity. A pool of mutants derived from Geobacillus zalihae T1 lipase was screened in organic solvents (methanol, ethanol, propanol, butanol and pentanol) resulting in the selection of six mutants at initial screening (A83D/K251E, R21C, G35D/S195 N, K84R/R103C/M121I/T272 M and R106H/ G327S). Site-directed mutagenesis further yielded quadruple mutants A83D/M121I/K251E/G327S and A83D/M121I/S195 N/T272 M, both of which had improved activity after incubation in methanol. The km and kcat values of these mutants vary marginally with the wild-type enzyme in the methanol/substrate mixture. Thermally induced unfolding of mutants was accompanied with some loss of secondary structure content. The root mean square deviations (RMSD) and B-factors revealed that changes in the structural organization are intertwined with an interplay of the protein backbone with organic solvents. Spatially exposed charged residues showed correlations between the solvation dynamics of the methanol solvent and the hydrophobicity of the residues. The short distances of the radial distribution function provided the required distances for hydrogen bond formation and hydrophobic interactions. These dynamic changes demonstrate newly formed structural interactions could be targeted and incorporated experimentally on the basis of solvent mobility and mutant residues. © 2021 Elsevier Inc. All rights reserved. Keywords: Lipases Organic solvent Mutagenesis Stability Molecular simulations 1. Introduction The most important feature of proteins is the stability conferred on them by the delicate balance of network of intramolecular interactions. The folded structure of the protein is maintained in an * Corresponding author. Department of Cell and Molecular Biology, Enzyme Microbial Technology Research Center, Faculty of Biotechnology and Biomolecular Science, Universiti Putra Malaysia Serdang, 43400, UPM Serdang, Selangor, Malaysia. E-mail address: adamleow@upm.edu.my (T.C. Leow). https://doi.org/10.1016/j.jmgm.2021.107897 1093-3263/© 2021 Elsevier Inc. All rights reserved. active state for efficient catalytic function [1]. Non-aqueous media have been exploited in recent years for several enzymatic industrial processes, given the fact that enzymes in organic solvents have long been shown to behave in a strikingly activated and enhanced manner distinct from that in an aqueous environment [2,3]. The processing conditions and the solvent environments are the major bottlenecks causing loss of the native protein conformation [4,5]. The catalysis of reactions in organic solvents in combination with stability at high temperatures are attractive attributes of numerous industrial lipases [6e8]. For this purpose, lipases are the preferred group of enzymes employed as biocatalysts. J. Maiangwa, S.H. Hamdan, M.S. Mohamad Ali et al. Journal of Molecular Graphics and Modelling 105 (2021) 107897 conformational switch, and the protein solvent interaction strength. Therefore, further fine-tuning of the mutants’ network of interactions will ultimately yield the desired biocatalysts with robust stability in hydrophilic solvents. The use of these enzymes in an organic solvent environment comes at a cost to the network of interactions that holds the constituent amino acids in the overall structure. The solvent-induced deleterious effects have been validated both experimentally and computationally, providing insights into the mechanisms underlying protein stability [9]. Notably, these organic solvents are capable of permeating the protein surface and causing damage to the secondary structure [10e12]. The “stripping off” effect and solvent polarity contribute in part to the unfolding and denaturation of most enzymes in organic solvents [13,14]. Despite these limitations, the biocatalytic turnover of enzymes could still be enhanced by using various techniques of protein engineering [15e17]. Increases in substrate and product solubility, a shift in the kinetic and thermodynamic constraints of hydrolysis reactions, alterations in substrate specificity, control of water-mediated side reactions, elimination of product contamination and increased organic solvent stability are properties of interest during the engineering of enzymes [18e20]. Several novel thermostable and organic solvent-tolerant lipases derived from different microbial sources have been investigated among which is the Geobacillus zalihae T1 lipase [21e26]. Geobacillus zalihae T1 lipase (a thermoalkaliphic lipase) is a 43-kDa protein that catalyses the hydrolysis of long-chain triglycerides into fatty acids at a high temperature of ~70  C [27]. Structural resolution of T1 lipase has been determined at 1.5 Å in a closed conformation [28]. The lipase shares the common canonical alpha/ beta hydrolase fold [29e31]. The catalytic triad of Ser113, His358 and Asp317 are conserved in an active site shielded by a lid domain, which modulates the phenomenal interfacial activation by the movement of helix a6 of the lid [32]. Geobacillus zalihae T1 lipase is a suitable candidate for use in various organic solvent processes due to its hydrolytic activity at elevated temperatures. Studies have revealed that T1 lipase shares the same distinctive properties with other thermoalkalophilic lipases from the same I.5 family. The ability of this group of lipases to catalyse a reaction under extreme conditions with high expression levels makes them a suitable choice for most biocatalytic processes [33e35]. However, molecular dynamic simulations have revealed the effects of some conformational variations in T1 lipase on its overall stability in organic solvents at high concentrations [36]. Although some microbial lipases are reported to be poor targets for directed evolution, thermostable enzymes are thought to have a high capacity to accommodate mutations that allow for broad-spectrum sampling of catalytically favourable mutations [37]. In this study, we successfully engineered T1 lipase to enhance its stability in organic solvents. Insights into how the motion of protein atoms and their interactions with organic solvents are utilized in modelling effective strategies that compensate for the deleterious effects of the organic medium guided the design of efficient mutants. Using high-throughput screening for improved solvent stability, mutants were identified and characterized in the presence of high concentrations of hydrophilic organic solvents (50%e70% (v/ v)). There have been a few reports of evolutionary engineering of thermostable lipases that have shown enhanced stability in high concentrations of methanol [26,38]. However, variants generated in this study did not show remarkable improvement in their activity in organic solvents as expected but proved useful in reactions in non-aqueous media. At the molecular level, the events that characterize the structural properties and stabilization of the T1 lipase variants in an organic solvent cannot be fully understood with the current experimental techniques [39]. Hence, molecular dynamic (MD) simulations were relied upon to evaluate related changes and the core stability properties of the mutants as characterized by the local effects of structural fluctuations, the compactness of secondary structure, hydrophobic surface area, secondary structure and 2. Results 2.1. Screening and activity assay of ePCR mutants of T1 lipase in organic solvents Following error-prone PCR of the T1 lipase gene, a large number of mutant libraries of ~20,000 variants were chosen at random. During the initial screening (Supplementary Fig. S1) mutants A83D/ K251E, R21C, G35D/S195 N, K84R/R103C/M121I/T272 M and R106H/G327S, which showed improved activity upon further screening, were selected (Fig. 1). The relative activity of the mutants upon further incubation resulted in complete inactivation and total loss of activity in butanol and pentanol solvent mixtures. Mutants were inactivated with longer incubation (Supplementary Fig. S2). All mutant activity decreased relative to the wild type T1 lipase (Supplementary Fig. S3). Ethanol and propanol had intrinsically variable stabilizing effects on the mutants, which suggested an inverse relationship between solvent polarity and enzyme stability. 2.2. In silico assessment of mutants and site-directed mutagenesis Not all amino acids substitutions contribute equally to protein stability [40,41]. Therefore, based on the free energy (Supplementary Tables S1&S2) of each mutant amino acid residue spot obtained in Fig. 1, the single mutants G35D, A83D, M121I, S195 N, K251E, T272 M and G327S were constructed via sitedirected mutagenesis and further screened for organic solvent stability [42,43]. Most of the substitutions were far from the active site region (Fig. 2) [44]. Investigations of activity assays in methanol, ethanol and propanol showed that all mutants were completely inactivated after 1-h incubation (Supplementary Fig. S4). Further activity assays were conducted for mutants in 50e60% methanol as solvent of choice with mutants M121I, S195 N and T272 M considered to be more stable variants in 50% (v/v) methanol, as shown in Fig. 3. Site-directed quadruple and double mutants generated from pairwise combinations of methanol tolerant single mutants in Fig. 3, revealed improved activity after incubation in methanol for mutants A83D/S195 N/K251E/G327S and A83D/M121I/S195 N/T272 M, and A83D/K251E (Fig. 4). In principle, the amino acid substitutions and the effects of single mutations in these mutants may provide independent contributions that leads to the overall improvement of their given protein property [45]. 2.3. Kinetic parameters of the mutants An increase in substrate solubility in methanol would lead to a higher substrate concentration, which would ultimately increase hydrolysis and enzyme rate kinetics. It has been shown that a high methanol concentration in a reaction medium exerts a greater inhibitory effect and lowers the reaction velocity (Vo) [46,47]. The apparent kinetic constants are summarized in Table 1 (Supplementary Fig. S5) and were found to favour faster reaction rates for mutants A83D/K251E and A83D/M121I/K251E/G327S in pnitrophenyl palmitate substrates only. The results showed that methanol increased the km for all mutants except mutant A83D/ M121I/S192 N/T272 M. However, while variations in the reaction rates of the mutants were observed for all mutants, wildtype T1 lipase had a steady state reaction rates in the two different substrate mixtures. Similarly, the catalytic efficiency in substrate 2 J. Maiangwa, S.H. Hamdan, M.S. Mohamad Ali et al. Journal of Molecular Graphics and Modelling 105 (2021) 107897 Fig. 1. The relative activity of T1 lipase mutants in methanol. Relative activity was defined as the specific activity relative to organic solvent untreated T1 lipase held at 100%. The control (TC), T1 lipase in 50% methanol, (Mmc), Methanol with substrate without enzyme, and (C), Assay buffer with substrate without enzyme and methanol. thermal denaturation measurements in 50% (v/v) methanol depict an increase in the Tm of all mutants from 60 to 80  C compared with a decrease for native T1 lipase by < 10  C. The secondary structure in the far-UV (200e250 nm) spectra in Figs. 5 and 6 showed the proteins adopted the features of a mixed a/ß fold, which is typical of most lipases. The presence of negative minima at 208 and 220 nm and a positive maximum at 190 nm provided evidence that methanol influences protein secondary structure. 2.5. General features of simulation dynamics of selected mutants Assessments of the possible local effects of the in silico modifications on the developed mutants are given in Supplementary Table S3, which describes the quality and stability of the mutant models. As observed from the FoldX force field energy of unfolding protein stability, the overall effects induced by the mutations appeared to restore the kinetic and thermodynamic properties of the mutant models. The complex effects and interaction networks that govern protein stability are considered to be well established, as revealed in the packing densities of all the residues, which were within the specified proximity of <1.0 Å separating the surfaces of atoms [48e50]. The cavity volume associated with some mutants will, on the average, not have large solvation effects on the interfaces of the protein structure. This is because relative to the cavity volume of the T1 (2DSN), the packing of the residues that are involved in protein-protein interactions could still accommodate buried water molecules which are integral to the stabilizing increase in hydrophobic and van der Waals interactions [51]. The information obtained from the dynamic simulation properties in Fig. 7 are in a way relevant to the conformational stability transitions of proteins [10,52]. These conformational deviations revealed an equilibrated stable RMSD over the time course of 100ns simulations in just the double mutants. However, the observed steep RMSD of the quadruple mutants and T1 lipase could generally be correlated with a higher number of hydrogen bonds, and the converse could also hold true. Fig. 2. Structure of the closed conformation of T1 lipase with the mutation points generated randomly across the structural elements of the protein. Mutants are highlighted in various colours and the active site Ser113 (red), Asp317 (yellow) and His358 (gray) are shown. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) transformation was found to be more favoured in the presence of methanol solvents and p-nitrophenyl palmitate mixtures. 2.4. Biophysical properties and protein folding of mutants in organic solvents The thermodynamic properties of each protein mutant in the presence and absence of methanol varied as shown in Table 2. The 3 J. Maiangwa, S.H. Hamdan, M.S. Mohamad Ali et al. Journal of Molecular Graphics and Modelling 105 (2021) 107897 Fig. 3. The relative activity of single mutants in 50e60%(v/v) methanol solvent mixtures. Representative panels are (A) 50% methanol (B) 60% methanol. Fig. 4. The relative activity of double, and quadruple mutants in 50e70% methanol, solvent mixtures. Representative panels are (A) 50% methanol (B) 60% methanol (C) 70% methanol. All results are replicates of independent experiment of 1 mg/mL enzyme solution incubated in various concentrations of methanol and assay at 60  C. Table 1 P-nitrophenol palmitate and methanol kinetics of selected mutants as compared to wild-type T1 lipase. pNPP Mutants km (mol/L) (10 A83D/K251E A83D/M121I/S192N/T272M A83D/M121I/K251E/G327S T1 5.9 9.8 5.3 9.4 MtOH/pNPP mixtures 1 ) kcat(10 4.6 4.7 4.6 6.9 2 ) km (mol/L) (10 8.7 4.3 9.4 9.0 1 ) kcat(10 2 ) 7.9 7.2 4.7 7.7 through the protein residues, especially those at the C-terminal extremities. They also showed that the surrounding network of the solvent molecules could drive the protein fluctuations in a dynamic character with less steric stress. Conversely, the rigidity of the protein was also associated with increased binding to solvent molecules, whereby the regions with a rigid distribution were considered to be areas where binding hot spots were localized [53]. 2.6. Functional flexibility The positional fluctuations in terms of the flexibility given by the B-factors are presented in Fig. 8. The flexibility of the residues enabled proteins to change conformation from closed to open in response to the solvent state of the environment. The patterns of atomic fluctuations in all mutants simultaneously followed all 4 J. Maiangwa, S.H. Hamdan, M.S. Mohamad Ali et al. Journal of Molecular Graphics and Modelling 105 (2021) 107897 modified by the greater relative mobility of the a-helices and bstrands, allowing for many possibilities for the stabilization of catalytically active residues. However, in some cases these structural conservations may differ greatly, with functionality and dynamics often influencing the robustness of these proteins to mutations [54]. The secondary structural elements in Fig. 9 did not show any significant perturbation as observed with the equilibrated peaks over the simulated timescale. All mutants exhibited secondary structure persistence, which would exclusively provide effective local stabilization interactions that would keep the structure in a desired flexible fold. The mechanism of organic solvent tolerance of proteins had been described for a long time [58], but it is still uncertain how methanol modulates protein structural contents. However, solvent size or volume and the dielectric constant (log P), which are ultimately dependent on the hydrophobicity of the organic solvent, are long considered to be some of the parameters involved [59,60]. The character and positional orientation of substituted amino acids can provide increased stability by promoting electrostatic interactions, especially when in a buried position [61,62]. When the side chains of the charged amino acids are oriented toward the exterior of the protein surface as shown in Fig. 10, they provide a hydration shell with the solvent molecules preventing both protein aggregation and electrostatic repulsions, which is vital to stability in such a solvent environment [55]. When the side chains of the charged amino acids are oriented toward the exterior of the protein surface as shown in Fig. 10, they provide a hydration shell with the solvent molecules preventing both protein aggregation and electrostatic repulsions, which is vital to stability in such a solvent environment [55]. Although penetration of solvents into the protein core has been previously Table 2 Thermal denaturation of mutants in buffer and methanol at 30e100  C as compared with wild-type T1 lipase. Mutants Tm (Buffer) Tm (Methanol) A83D/K251E A83D/M121I/K251E/G327S A83D/M121I/S192N/T272M T1 69 60 63 70 81 88 79 61 In this situation, B-factors showed high deviations as methanol molecules strip off water molecules in the quadruple mutants around residues Asn141-Gly150. The higher flexibility. Indicates that both the hydrophobic and hydrophilic side chains reorient themselves on the protein surface in organic solvents (particularly for Asn141-Gly150). Similar observations have been reported for other organic solvent media with variations of the solvent composition (Zhu et al., 2012). Usually, solvent-buried residues would have the lower RMSF/B-factors and solventexposed residues would have the higher RMSF/B-factors (Maiangwa et al., 2017). Higher B-factors deviations could reflect changes in the solvent water interface around the protein leading to changes in the dynamical properties of the protein. In a corresponding manner, the solvent accessible surface area (SASA) (Supplementary Fig. S6), indicates conformational changes also did not have an effect on the compactness of the protein mutant models. 2.7. Secondary structural variability and conformational switch The intrinsic dynamics of the structurally rigid folded state is Fig. 5. The Far-UV circular dichroism spectra after incubation of selected best mutants in the absence of methanol (A&B), and in 50% methanol solvent mixture (C&D) showing progressive loss in ellipticity. 5 J. Maiangwa, S.H. Hamdan, M.S. Mohamad Ali et al. Journal of Molecular Graphics and Modelling 105 (2021) 107897 Fig. 6. The Far-UV circular dichroism spectra after incubation of selected best mutants in the in 60% methanol solvent mixture for 1e4 h (A&B) showing secondary structure induced changes. Fig. 7. Evolution of the root-mean-square deviation (RMSD) of T1 lipase mutant models simulated for 1000 ns. Equilibrated Molecular Dynamic simulations were done in replicates runs. Title legends are designation of different mutant models. functionality: hydrogen bonding and hydrophobic interactions [1,56]. The best set of mutations exerted a local effect that modified the interplay between the protein dynamics of hydrogen bonding and hydrophobic interactions. The results showed that mutations elicited the formation of electrostatic interactions and intramolecular hydrogen bonds, particularly in the quadruple mutants A83D/M121I/K251E/G327S and A83D/M121I/S195 N/T272 M (Figs. 11 and 12 and Spplementary Table S4). This mediated a network extension of electrostatic interactions with a more demonstrated by MD simulations to cause conformational changes [5,36], it can be assumed that such penetration into the core will most likely not cause interactions with the hydrophobic residues, hence the hydration shell and stability could be preserved. 2.8. Local effects of mutations on structural interactions Two types of interactions are traditionally considered to hold the balance between the stability of folded proteins and their 6 J. Maiangwa, S.H. Hamdan, M.S. Mohamad Ali et al. Journal of Molecular Graphics and Modelling 105 (2021) 107897 Fig. 8. The average B-factors estimated from the root-mean-square fluctuations (rmsf) of entire protein residues from the equilibrated length of 100 ns simulations of mutant models as compared to T1 lipase in methanol. Simulations are average of replicates of three independent runs represented. Title legends are designation of different mutant models. Fig. 9. The Percentage of preserved secondary structural elements of T1 lipase mutants simulated in methanol solvent. All structural elements are replicates of three independent simulations of the last 40ns of 100ns simulation runs. favourable aromatic stacking environment via cation teractions [57]. affected by the spatial distance of surrounding solute atoms [63]. The orientation of the oxygen, hydrogen, hydroxyl and methyl groups of water and methanol are presented in Fig. 13. There is significantly higher contact of the methanol oxygen and methyl groups with the protein, as revealed by the contact peak. The contact peak of water (Fig. 13), moved towards shorter peak distances suggesting oxygen groups are closer to the hydrophobic surface than the hydrogen group in the water molecules. The p in- 2.9. Local structural orientation of solvents for protein-solvent interactions The radial distribution function (RDF) describes how the structural system, particularly that of the surrounding solvents, is 7 J. Maiangwa, S.H. Hamdan, M.S. Mohamad Ali et al. Journal of Molecular Graphics and Modelling 105 (2021) 107897 conformational state [65]. The loss of stability of mutants in butanol and pentanol clearly demonstrates the complex phenomenon underlying the mechanisms of inactivation by these solvents. It is proposed that when amino acid residue hot spots are substituted in a combinative manner, their combined effect could further enhanced certain protein properties [66]. The correlation between thermostability and solvent tolerance is in line with the observations of Hamamatsu et al. [67]. In principle, recombination of finely tuned single mutations can have profound benefits for the local fitness landscapes of proteins [83]. However, this may not be applicable to other combinatorial mutants, in which a combination of single mutations may decrease, rather than increase, the thermal stability of a lipase [68]. The catalytic efficiency of the mutants correlated with their miscibility in the organic solvents used in this study. Moreover, the unfolding and exposure of the hydrophobic core of proteins is instantaneous in a non-aqueous system [69]. The deactivation profiles of these mutants were mediated by first-order kinetics, with the stability of mutants A83D/K251E, M121I/S195 N, A83D/ M121I/K215E/G327S and A83D/M121I/S195 N/T272 M being more improved. The first-order kinetics for lipases in organic solvents are known to strip off enzyme-bound water molecules more rapidly [70,71]. Consequently, activation by methanol mediates solubility and substrate diffusion towards active sites and allows the proteins to remain kinetically entrapped in a folded and thus active state. As demonstrated by CALB lipase and other enzymes such as phytase, low methanol concentrations do not have significant effects on enzyme kinetics, and higher enzyme concentrations can suppress the inhibitory effects of methanol [47,72]. The thermal behaviour of the mutants could be attributable to their ability to either maintain rigidity (in the case of increased thermostability) or gradually lose their rigidity at temperatures above 70  C. The thermal denaturation (Tm) and melting temperatures (Tm) in methanol appear dependent on an increase in the thermal stability of mutants with a Tm  80  C. This highlights the cooperativity in unfolding and the favourable interaction of the protein with solvents at no cost to protein flexibility. The Circular Dicrhoism spectral properties of mutant and wildtype T1 lipase revealed that the electrostatic interactions between the proteins and the solvent molecules can alter their structures, particularly in the case of methanol. The spectral properties of methanol-treated Candida albicans lipase also suggested that the formation of bsheets is methanol induced [47]. The loss of hydrogen bond interactions between solvent molecules and surface residues allows proteins to have more stability and structural rigidity in organic solvents due to b-sheet formation, as has been previously described in a Pseudomonas cepacia lipase [73]. The contents of the secondary structure did not change significantly in the absence of methanol and were kept constant despite variations in the incubation time. The propensity of proteins to remain in an ordered state is ascribed to induced changes in the b-structural contents, since the response to methanol inactivation is highly protein specific [72,73]. The evolution of higher functional flexibility in the quadruple mutants could be best described as a “reciprocal influence” exerted by the combinatorial mutation of the functionally flexible single mutants A83D, M121I and S192 N [74]. In aqueous environments, the largest contribution to stabilization of the native conformation of proteins is made by hydrophobic interactions [2,75]. It was, however, observed in this study that the non-differential changes in the secondary structural conformational switch implied that destabilization or denaturation by polar solvents would occur in a sequential manner with increasing concentrations and hydrophilicity of the polar solvents. The preservation of protein secondary structure in organic solvents is a function of the hydration attributes of particularly exposed electrically charged amino acids. Fig. 10. Exposed charged amino acids (red spheres) of mutants with electrically charged amino acids more exposed during 100 ns simulations in methanol solvent. Mutants are; (A) T1 native structure (B) A83D/K251E (C) A83D/M121I/K251E/G327S (D) A83D/M121I/S195 N/T272 M. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) Fig. 11. The structural cluster of bond formation of single mutant residues with neighbouring residues (A) M121I/S192 N, compared with (B) Native T1 lipase. And (C) A83D/K251E, compared with (D) Native T1 lipase. Hydrogen bonds are in red between atoms of mutant residues (green) and neighbouring residues (blue). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) methyl and oxygen group of the methanol are both closer to the hydrophobic patch of the protein. This describes the preferential orientation of the solvent molecules, binding and consequently stripping of water molecules within the protein surface. Considering the role of water in protein dynamics, it was also possible for water structures in the simulation studies to be disrupted and devoid of positional correlation [64]. The results revealed no observed differences between the hydroxyl groups of methanol at the air-water interface in relation to the protein surface of the quadruple mutants (Fig. 14). There was no possible contact between the surface residues and the hydroxyl groups which could be attributable to the lower dielectric constant of the protein relative to the selected solvent. 3. Discussion Enzymes sometimes achieve good reaction synthesis rates in organic solvents media but could however, be inactivated [22]. Comparisons of the organic solvent stability of the T1 lipase from Geobacillus zalihae with that of evolved mutants provided evidence that the behaviour of enzymes in organic solvents varies with their 8 J. Maiangwa, S.H. Hamdan, M.S. Mohamad Ali et al. Journal of Molecular Graphics and Modelling 105 (2021) 107897 Fig. 12. The structural cluster of bond formation of quadruple mutants (A) A83D/M121I/K251E/G327S, compared to the native T1 (B). The quadruple mutant (C) A83D/M121I/ S195 N/T272 M, as compared to native T1 lipase (D). Hydrogen bonds are in red between atoms of mutant residues (green) and neighbouring residues (blue). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) Fig. 13. 100 ns simulations of solute to solvent radial distribution functions of water, oxygen and metanol atoms. Legends are designation of each mutant models as watermethanol-oxygen interactions (red peaks), methanol-oxygen interactions (black peaks) and, water-oxygen interactions (blue peaks). Simulations are replicates of independent runs. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) Fig. 14. 100 ns simulations of solute to solvent radial distribution featuring spatial correlation between solute to solvent functions of the hydroxyl groups (OH) of methanol only for mutants (A) A83D/M121I/K251E/G327S and (B) A83D/M121I/S195 N/T272 M. 9 J. Maiangwa, S.H. Hamdan, M.S. Mohamad Ali et al. Journal of Molecular Graphics and Modelling 105 (2021) 107897 Importantly, the irreversible unfolding of protein secondary structure by polar solvents, imposes more structural constraints to stability [10]. The cooperative hydration and network interactions of the polar and charged residues clustered away from the protein’s interior would promote effective clustering of the hydrophobic groups to form a hydrophobic core [56]. Proteins can effectively modulate the structuring of their solvent environment in the presence of water, which is somewhat dependent on the monolayer of water covering the enzyme surface [76]. Depending on the orientation distribution of the OH group of alcohols, a linear orientation could favour strong hydrogen bond interactions when the OH acts as a hydrogen bond donor when interacting with the solute atom in the radial distribution [77]. Generally, in an analogous way, the organic solvent molecules can induce positional fluctuations of the atom groups towards the charged sites of the proteins far more rapidly. When freely mobile, the oxygen atom of water could experience repulsion by the organic solvent atoms, hence orienting them into the solution away from the protein surface. However, when in contact with the oxygen atom of methanol, the contact distance was increased, thereby favouring further interaction with the surface due to the hydrophilicity of the methanol solvent molecules. The low dielectric constant of the protein surface influenced the repellent behaviour of the organic solvents toward the water molecules. Simulation studies had demonstrated this behaviour to be analogous to that at the air-water interface, whereby organic solvent molecules achieved a higher frequency of interaction with the protein surface than water [78]. 5. Materials and methods 5.1. Bacteria strains and construction of T1 lipase mutants using error-prone PCR The Geobacillus zalihae T1 lipase gene was synthesized with NcoI and Bpu1102I restriction endonuclease sites incorporated upstream and downstream of the N- and C- terminus, respectively, and cloned into a pET28b(þ) vector (Novagen, USA). The recombinant plasmid (pET28b/lipT1) habouring T1 lipase gene was used as a template for the first round of random mutagenesis using a of forward (GT1L-F:50 TATACCATGGGCCATCATCATCATCATCAC30 ) and reverse (GT1L-R:50 CAAGGGGTTATGCTAGTTATTGCTCAGCTTA30 ) primers. Error-prone PCR (ePCR) was conducted in a 50-mL reaction mixture using the Genemorph II random mutagenesis Kit (Stratagene, Agilent). The reaction mix contained 0.1 mg DNA template, 200 mM dNTP mix, 200 ng/mL of each primer, 10X Mutazyme II Buffer and 2.5 U of Mutazyme II polymerase. Using an S1000™ Thermal cycler (Bio-Rad, USA), PCR was conducted with an initial denaturation at 95  C for 2 min, followed by 35 cycles of denaturation at 95  C for 30 s, annealing at 58  C for 30 s and extension at 72  C for 1 min. The final  extension was conducted at 72 C for 10 min. The purified PCR product and pET28b(þ) vector were digested with NcoI and Bpu1102I (NEB) restriction enzymes, ligated, and transformed into E. coli BL21 (DE3) chemically competent cells. The transformed cells were grown on tributyrin Luria-Bertini agar supplemented with 50 mg/mL kanamycin and incubated at 37  C for 24 h. 5.2. Screening of organic solvent tolerant mutants of T1 lipase Over 20,000 colonies were selected for the initial screening €ttcher & Bornscheuer [79], with minor modificaaccording to Bo tions. In the first step of screening, the mutant libraries were randomly selected by picking colonies with degraded halos that were grown on tributyrin Luria-Bertini agar plates. The colonies were cultured in Luria-Bertini broth supplemented with kanamycin (50 mg/mL) in 96-well microtiter plates as the master plate. Replica plates were prepared from the master plate and each well was induced with 0.05 mM isopropyl b-D-1-thiogalactopyranoside (IPTG), and the expressed mutants were assayed for lipase activity after preincubation in methanol. Variants that showed lipase activity were selected from the corresponding master plate for further validation. Further screening was done by the colony lift approach using filter paper as described by Reiter et al. [80], and Korman et al., [26]. Overnight cultures of mutant clones were lifted onto sterile filter paper (Advantec 5C) and placed colony side up on a Luria-Bertini-agar plate containing 50 mg/mL kanamycin and 0.4 mM IPTG, which was then incubated for 12 h at 37  C to induce lipase expression. The colony filters were soaked in lysis solution and immersed in 50% methanol solution and incubated for 30 min at 60  C. The filters were overlaid with developing solution [3 mM fatty acid naphthyl palmitate, 1 mM fast blue B (Sigma), 100 mM Tris-HCl buffer, pH 8.0, dispersed in 1% (w/v) agar]. Corresponding spots with dark precipitates were selected from the master plates and preserved for further validation (see Supplementary Figure S1). 4. Conclusion Random mutagenesis and the combination of beneficial mutations generated improved mutants with improved organic-solvent stability A83D/K251E, A83D/M121I/K251E/G327S and A83D/ M121I/S195 N/T272 M. Mutations in the set of mutants established a hydrophobic cluster with more rigidifying reinforcement of the stability of the structure according to the MD simulations. The additive stabilizing effects of the individual mutations benefited the evolved quadruple mutants. Considerable variations in kinetics among mutants highlights the concentration dependence of methanol-induced kinetics. The catalytically active state of the mutants was improved in the half-life retention profile. The thermostability of the mutants could suggest the immutable state of the salt bridges to the mutations in silico. Further engineering of salt bridges at target hot spots determinant of solvent stability could possibly improve the catalytic activity of the mutants. The proximal locations of some mutations in the quadruple mutants in the study showed a higher persistence of hydrophobic residues with particular regards to electrostatic interactions. Clustered interactions provided molecular details related to the observed variations in the cooperative network of interactions and the general energy minimum of fluctuations and flexibility. The mutations were capable of eliciting effects in relation to the restoration or strengthening of the target interactions of the residues with the solvent in such a dynamic environment. Thus, there was no denaturing effect on the secondary structural elements, which was thought to be due to the gain in intermolecular hydrogen and hydrophobic interactions. Integrating different properties related to the protein dynamics of the mutants with the correlated motions and electrostatic interaction network could be useful to predict and further fine-tune the important residues of these mutants with the aim of improving upon their protein activity and stability. 5.3. Expression, purification and validation of methanol-tolerant mutants Overnight cultures of screened mutants were induced in 40 mL Luria-Bertini broth with 0.025 mM isopropyl-b-D-thiogalactopyranoside (IPTG) and cell cultures were harvested and the pellet 10 J. Maiangwa, S.H. Hamdan, M.S. Mohamad Ali et al. Journal of Molecular Graphics and Modelling 105 (2021) 107897 suspended in 50 mM phosphate buffer, pH 8. The suspended cells were sonicated with an amplitude of 35% and pulse length of 30 s for 3 min, and clarified by centrifugation (10,000g), and the crude cell supernatants were filtered through a 0.42-mm pore filter. The supernatants were loaded onto a 5-mL Ni2þ Sepharose 6 Fast Flow prepacked column (GE Healthcare, USA) mounted on an AKTA Prime Plus (GE Healthcare, USA). The column was equilibrated with 5 CV of binding buffer (20 mM phosphate buffer, 20 mM imidazole, 500 mM NaCl, pH 7.4). The purified mutants were dialyzed against 50 mM phosphate buffer (pH 8.0) for 24 h at 4  C. The enzymatic assay was performed at 60  C for 5 min using 0.5 mM p-nitrophenol palmitate (pNPP) as a substrate and methanol at a concentration of 50% (v/v). The residual lipase activity was determined and one unit of lipase activity was defined as the activity required to produce 1 mmol p-nitrophenol per minute under standard assay conditions. 5.7. Molecular dynamic simulations of mutants of T1 lipase Models of mutants shown to give better methanol stability (A83D/K251E, A83D/M121I/K251E/G327S and A83D/M121I/S192 N/ T272 M) were constructed using the empirical potential function of FoldX 4 (http://foldxsuite.crg.eu/). The mutants were generated by using the reference crystal structure of T1 lipase (PDB code: 2DSN). The stability of each corresponding mutant model was estimated by calculating the free energy change (DDG) required to fold the proteins from their unfolded state based on the following linear combination of empirical terms used to calculate free energy (in kcal mol 1) by FoldX as described in detail by Schymkowitz et al. [81]. All models were validated by performing a number of quality checks with YASARA at 500ps simulations to determine the Model Quality, Bond angles, Bond length, Dihedral angles and the 3DPacking of residues of models. Further structural assessment was performed using the RAMPAGE server http://mordred.bioc.cam.ac. uk/. 5.4. In silico modeling of mutations and site-directed mutagenesis All mutants generated via random mutagenesis were identified and single amino acid mutant models designed in silico from each mutant. The empirical potential function of FoldX [81], the Medusa force field of Eris [42] and YASARA 16.1.0 computational package [82] were used for predicting the folding free-energy for model quality checks and protein stability prediction. The single mutant models with better quality were constructed by site-directed mutagenesis with different primers (Supplementary Table S5). The constructed mutants were expressed, purified and assayed for enzyme activity. One unit of lipase activity was defined as the activity required to produce 1 mmol p-nitrophenol per minute under standard conditions. DG ¼ a:DGvwd þ b:DGsolvH þ c:DGsolvP þ d:DGwb þe:DGhbond þ f :DGel þ g:DGkon þ h:T DS me þk:T DS sc þ l:DGclash The stabilizing effects of the substitutions in the packing of protein atoms, which served as an indicator for their stability and functionality, were further investigated on the basis of atomic-scale packing data for protein 3D structures using a Voronoi Cell algorithm [87,88]. Parameterization steps prior to simulations, as well as the analysis of dynamic properties of the protein, were carried out as described in Maiangwa et al. [36] with minor modifications. Simulations were done using methanol as solvent molecules and the mutant models A83D/K251E, A83D/M121I/K251E/G327S and A83D/M121I/S192 N/T272 M. The atoms of the protein structure in the different solvent mixtures in the simulation box were constrained (fixed). This was to equilibrate the solvent mixtures in the ensemble box and allow the relaxation of the solvent molecules and prevent large distortions to the protein. To remove bumps, delete unwanted water molecules and correct the covalent geometry, the atoms were unconstrained and the system setup was energy-minimized with the AMBER14 force field. The geometric parameters (dihedral angles, bond angles, and length) including Van der Waals and Coulomb interactions of the force field were truncated at a 7.86 Å force cut-off. The Particle Mesh Ewald algorithm was used to treat long-range electrostatic interactions with a grid spacing of <1 Å. After removal of conformational stress by a short steepest descent minimization of 200 steps, the procedure continued by simulated annealing (time step 2 fs, atom velocities scaled down by 0.9 every 10th step) until convergence was reached. This final system setup was used for the production runs. The Amber14 force field adapted in the YASARA macros protocol described in detail in Krieger et al. [82], was used throughout the entire simulation process. All simulations were run with periodic boundary conditions and in replicates based on different initial velocities as modified by the RandomSeed number based on the temperature control (343K) assigned using a Maxwell-Boltzmann distribution to achieve different initial velocities for each replicate simulation over a 100 ns timescale. The velocities of the atoms were rescaled according to the Berendsen barostat, and pH 9 was used in all solvents. The manometer pressure mode was used to control the pressure from the kinetic energy/virial and slowly rescaled the simulation cell along the X, Y, and Z-plane to smoothly approach 1 bar. Simulation snapshot coordinates were saved every 30 ps. 5.5. Determination of kinetic parameters Initial rate measurements with 50 mg/mL of purified mutants were performed in 50 mM phosphate buffer, pH 8.0 at 70  C for 25 min with an increasing substrate concentration of p-nitrophenol palmitate (0.2e3 mM) in methanol. The kinetic constants of the mutants were measured by fitting the experimental data to a nonlinear regression to obtain the Michaelis constant (km) and maximum velocity (Vmax) using the GraphPad Prism 8. 5.6. Biophysical analysis of mutant T1 lipase The thermal transitions of lipase mutants were monitored by visible absorption using circular dichroism (CD). Purified mutants and native T1 lipase (final concentration of 4 mM in 10 mM phosphate buffer) were respectively incubated at 60  C in phosphate buffer (10 mM, pH 8) and 50% methanol for 1e4 h. The CD spectra were recorded by using a Jasco J-815 circular dichroism spectropolarimeter equipped with Peltier temperature control and cell stirring and an N2 purge with a 10-mm path length quartz cuvette, 0.1-nm data pitch, 1.00-nm bandwidth and scan speed of 100 nm/ min. The melting temperature (Tm) curves were obtained by heating the sample at 30e100  C/min and recording the ellipticity value at 222 nm. The Tm values were extracted from the sigmoidal part of the plot for all experiments. Spectra in the far-UV region (190e260 nm) were recorded as the mean of three consecutive scans [83e85]. Appropriate baseline subtraction and noise reduction analysis were performed for methanol and the buffer. The mean residue molecular ellipticity [q] was expressed as deg. cm2.dmoL 1 using a mean residue weight of 110. The percentages of secondary structures were calculated and cross-referenced against other databases as described by Raussens, Ruysschaert & Goormaghtigh, [86]. 11 J. Maiangwa, S.H. Hamdan, M.S. Mohamad Ali et al. Journal of Molecular Graphics and Modelling 105 (2021) 107897 J. Chem. Inf. Model. 52 (2012) 465e473, https://doi.org/10.1021/ci200455z. [11] L. Yang, J.S. Dordick, S. Garde, Hydration of enzyme in nonaqueous media is consistent with solvent dependence of its activity, Biophys. J. 87 (2004) 812e821, https://doi.org/10.1529/biophysj.104.041269. [12] M. Silberstein, S. Dennis, L. Brown, T. Kortvelyesi, K. Clodfelter, S. Vajda, Identification of substrate binding sites in enzymes by computational solvent mapping, J. Mol. Biol. 332 (2003) 1095e1113, https://doi.org/10.1016/ j.jmb.2003.08.019. [13] D. Chakravorty, S. Parameswaran, V.K. Dubey, S. Patra, Unraveling the rationale behind organic solvent stability of lipases, Appl. Biochem. Biotechnol. 167 (2012) 439e461, https://doi.org/10.1007/s12010-012-9669-9. [14] F.H. Arnold, Engineering enzymes for non-aqueous solvents, Trends Biotechnol. 8 (1990) 244e249, https://doi.org/10.1016/0167-7799(90)90186-2. [15] F. Hasan, A.A. Shah, A. Hameed, Industrial applications of microbial lipases, Enzym. Microb. Technol. 39 (2006) 235e251, https://doi.org/10.1016/ j.enzmictec.2005.10.016. [16] S.B. Rubin-Pitel, H. Zhao, Recent advances in biocatalysis by directed enzyme evolution, Comb. Chem. High Throughput Screen. 9 (2006) 247e257. http:// www.ncbi.nlm.nih.gov/pubmed/16724916. [17] H. Yu, H. Huang, Engineering proteins for thermostability through rigidifying flexible sites (RFS), Biotechnol. Adv. (2013), https://doi.org/10.1016/ j.biotechadv.2013.10.012. [18] M.T. Reetz, J.D. Carballeira, Iterative saturation mutagenesis (ISM) for rapid directed evolution of functional enzymes, Nat. Protoc. 2 (2007) 891e903, https://doi.org/10.1038/nprot.2007.72. [19] S. Cesarini, C. Bofill, F.I.J. Pastor, M.T. Reetz, P. Diaz, A thermostable variant of P. aeruginosa cold-adapted LipC obtained by rational design and saturation mutagenesis, Process Biochem. 47 (2012) 2064e2071, https://doi.org/ 10.1016/j.procbio.2012.07.023. [20] Z. Han, S. Han, S. Zheng, Y. Lin, Enhancing thermostability of a Rhizomucor miehei lipase by engineering a disulfide bond and displaying on the yeast cell surface, Appl. Microbiol. Biotechnol. 85 (2009) 117e126, https://doi.org/ 10.1007/s00253-009-2067-8. [21] J. Yang, D. Guo, Y. Yan, Cloning, expression and characterization of a novel thermal stable and short-chain alcohol tolerant lipase from Burkholderia cepacia strain G63, J. Mol. Catal. B Enzym. 45 (2007) 91e96. http://www. sciencedirect.com/science/article/pii/S1381117707000033. (Accessed 14 November 2013). [22] E. Vazquez-Figueroa, V. Yeh, J.M. Broering, J.F. Chaparro-Riggers, A.S. Bommarius, Thermostable variants constructed via the structure-guided consensus method also show increased stability in salts solutions and homogeneous aqueous-organic media, Protein Eng. Des. Sel. 21 (2008) 673e680, https://doi.org/10.1093/protein/gzn048. [23] S.-L. Wang, Y.-T. Lin, T.-W. Liang, S.-H. Chio, L.-J. Ming, P.-C. Wu, Purification and characterization of extracellular lipases from Pseudomonas monteilii TKU009 by the use of soybeans as the substrate, J. Ind. Microbiol. Biotechnol. 36 (2009) 65e73, https://doi.org/10.1007/s10295-008-0473-z. [24] M. Gatti-Lafranconi, P. Natalello, A. Rehm, S. Doglia, S.M. Pleiss, J. Lotti, Evolution of stability in a cold-active enzyme elicits specificity relaxation and highlights substrate-related effects on temperature adaptation, Biol. J. Mol. 1 (2010) 155e166. [25] C.M. Romero, L.M. Pera, F. Loto, C. Vallejos, G. Castro, M.D. Baigori, Purification of an organic solvent-tolerant lipase from Aspergillus Niger MYA 135 and its application in ester synthesis, Biocatal. Agric. Biotechnol. 1 (2012) 25e31, https://doi.org/10.1016/j.bcab.2011.08.013. [26] T.P. Korman, B. Sahachartsiri, D.M. Charbonneau, G.L. Huang, M. Beauregard, J.U. Bowie, Dieselzymes: development of a stable and methanol tolerant lipase for biodiesel production by directed evolution, Biotechnol. Biofuels 6 (2013) 70, https://doi.org/10.1186/1754-6834-6-70. [27] T.C. Leow, R.N.Z.R.A. Rahman, M. Basri, A.B. Salleh, A thermoalkaliphilic lipase of Geobacillus sp. T1, Extremophiles 11 (2007) 527e535, https://doi.org/ 10.1007/s00792-007-0069-y. [28] H. Matsumura, T. Yamamoto, T.C. Leow, T. Mori, A.B. Salleh, M. Basri, T. Inoue, Y. Kai, R.N.Z.R.A. Rahman, Novel cation- p interaction revealed by crystal structure of thermoalkalophilic lipase, Proteins (2007) 592e598, https:// doi.org/10.1002/prot. [29] D.L. Ollis, E. Cheah, M. Cygler, B. Dijkstra, F. Frolow, S.M. Franken, M. Harel, S.J. Remington, I. Silman, J. Schrag, The alpha/beta hydrolase fold, Protein Eng. 5 (1992) 197e211. http://www.ncbi.nlm.nih.gov/pubmed/1409539. (Accessed 10 November 2013). [30] M. Nardini, B.W. Dijkstra, a/b Hydrolase fold enzymes: the family keeps growing, Curr. Opin. Struct. Biol. 9 (1999) 732e737, https://doi.org/10.1016/ S0959-440X(99)00037-8. [31] J.L. Arpigny, K.E. Jaeger, Bacterial lipolytic enzymes: classification and properties, Biochem. J. 343 (1999) 177e183. http://www.pubmedcentral.nih.gov/ articlerender.fcgi?artid¼1220539&tool¼pmcentrez&rendertype¼abstract. [32] Y. Wang, D.Q. Wei, J.F. Wang, Molecular dynamics studies on T1 lipase: insight into a double-flap mechanism, J. Chem. Inf. Model. 50 (2010) 875e878, https://doi.org/10.1021/ci900458u. [33] J. Kristjansson, Thermophilic organisms as sources of thermostable enzymes, Trends Biotechnol. 7 (1989) 349e353, https://doi.org/10.1016/0167-7799(89) 90035-8. [34] M. Guncheva, D. Zhiryakova, Catalytic properties and potential applications of Bacillus lipases, J. Mol. Catal. B Enzym. 68 (2011) 1e21, https://doi.org/ 10.1016/j.molcatb.2010.09.002. 5.8. Analysis of trajectories Analysis of dynamic properties of the temperature factors (Bfactors), root mean square deviations (RMSD) from the alpha carbon (Ca) and structural and hydrogen interactions were carried out as described previously in Maiangwa et al. [36]. The radial distribution function was used to measure the distance correlation between water and methanol atoms and sorted into 40 bins, each 0.25 A wide (covering the range up to 10 A). The RDF was obtained with the following formula: RDF i ¼ CountsInBini Atoms1* 43 p*BinWidth3 *ðði þ 1Þ3 i3 Þ Where Atoms1 is the number of atoms in the first selection of the BinDistance command, and the rest specifies the volume of the spherical shell whose counts are collected in bin i. 4. The secondary structure analysis was carried out using the Kabsch & Sander (1983) algorithm incorporated in their Dictionary of Secondary Structure for Proteins (DSSP) program to analyse the variation of protein secondary structure changes. The secondary structure assignment algorithm program installed onto YASARA was used to analyse the change in secondary structure contents. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments This research was funded by Universiti Putra Malaysia through the High-Performance Individual Research Grants Scheme (UPM/ 700-1/2/GPPI/2017/9532200). Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.jmgm.2021.107897. References [1] R.E. Hubbard, M. Kamran Haider, Hydrogen bonds in proteins: role and strength, in: Encycl. Life Sci., John Wiley & Sons, Ltd, Chichester, UK, 2010, pp. 1e6, https://doi.org/10.1002/9780470015902.a0003011.pub2. [2] A.L. Serdakowski, J.S. Dordick, Enzyme activation for organic solvents made easy, Trends Biotechnol. 26 (2008) 48e54, https://doi.org/10.1016/ j.tibtech.2007.10.007. [3] A. Klibanov, Why are enzymes less active in organic solvents than in water? Trends Biotechnol. 15 (1997) 97e101, https://doi.org/10.1016/S01677799(97)01013-5. [4] V. Stepankova, J. Damborsky, R. Chaloupkova, Organic co-solvents affect activity, stability and enantioselectivity of haloalkane dehalogenases, Biotechnol. J. 8 (2013) 719e729, https://doi.org/10.1002/biot.201200378. [5] Z. Kamal, P. Yedavalli, M.V. Deshmukh, N.M. Rao, Lipase in aqueous-polar organic solvents: activity, structure, and stability, Protein Sci. 22 (2013) 904e915, https://doi.org/10.1002/pro.2271. [6] A.M. Klibanov, Improving enzymes by using them in organic solvents, Nature 409 (2001) 241e246, https://doi.org/10.1038/35051719. [7] N. Doukyu, H. Ogino, Organic solvent-tolerant enzymes, Biochem. Eng. J. 48 (2010) 270e282, https://doi.org/10.1016/j.bej.2009.09.009. [8] K. De Godoy Daiha, R. Angeli, S.D. De Oliveira, R.V. Almeida, Are lipases still important biocatalysts? A study of scientific publications and patents for technological forecasting, PLoS One 10 (2015) 1e20, https://doi.org/10.1371/ journal.pone.0131624. [9] E.P. Hudson, R.K. Eppler, D.S. Clark, Biocatalysis in semi-aqueous and nearly anhydrous conditions, Curr. Opin. Biotechnol. 16 (2005) 637e643, https:// doi.org/10.1016/j.copbio.2005.10.004. [10] D. Lousa, A.M. Baptista, C.M. Soares, Analyzing the molecular basis of enzyme stability in ethanol/water mixtures using molecular dynamics simulations, 12 J. Maiangwa, S.H. Hamdan, M.S. Mohamad Ali et al. Journal of Molecular Graphics and Modelling 105 (2021) 107897 [35] X.-L. Qin, D.-M. Lan, J.-F. Zhong, L. Liu, Y.-H. Wang, B. Yang, Fatty acid specificity of T1 lipase and its potential in acylglycerol synthesis, J. Sci. Food Agric. 94 (2014) 1614e1621, https://doi.org/10.1002/jsfa.6468. [36] J. Maiangwa, M.S. Mohamad Ali, A.B. Salleh, R.N.Z.R.A. Rahman, Y.M. Normi, F. Mohd Shariff, T.C. Leow, Lid opening and conformational stability of T1 Lipase is mediated by increasing chain length polar solvents, PeerJ 5 (2017), e3341, https://doi.org/10.7717/peerj.3341. [37] H. Leemhuis, R.M. Kelly, L. Dijkhuizen, Directed evolution of enzymes: library screening strategies, IUBMB Life 61 (2009) 222e228, https://doi.org/10.1002/ iub.165. [38] A. Dror, E. Shemesh, N. Dayan, A. Fishman, Protein engineering by random mutagenesis and structure-guided consensus of Geobacillus stearothermophilus lipase T6 for enhanced stability in methanol, Appl. Environ. Microbiol. 80 (2014) 1515e1527, https://doi.org/10.1128/AEM.03371-13. [39] H. Monhemi, M.R. Housaindokht, A.A. Moosavi-Movahedi, M.R. Bozorgmehr, How a protein can remain stable in a solvent with high content of urea: insights from molecular dynamics simulation of Candida Antarctica lipase B in urea : choline chloride deep eutectic solvent, Phys. Chem. Chem. Phys. 16 (2014) 14882, https://doi.org/10.1039/c4cp00503a. [40] A. Dixit, A. Torkamani, N.J. Schork, G. Verkhivker, Computational modeling of structurally conserved cancer mutations in the RET and MET kinases: the impact on protein structure, dynamics, and stability, Biophys. J. 96 (2009) 858e874, https://doi.org/10.1016/j.bpj.2008.10.041. [41] S.E. Acuner Ozbabacan, A. Gursoy, O. Keskin, R. Nussinov, Conformational ensembles, signal transduction and residue hot spots: application to drug discovery, Curr. Opin. Drug Discov. Dev 13 (2010) 527e537. [42] S. Yin, F. Ding, N. V Dokholyan, Eris, An automated estimator of protein stability, Nat. Methods 4 (2007) 466e467, https://doi.org/10.1038/nmeth0607466. [43] S. Khan, M. Vihinen, Performance of protein stability predictors, Hum. Mutat. 31 (2010) 675e684, https://doi.org/10.1002/humu.21242. [44] G.P. Horsman, A.M.F. Liu, E. Henke, U.T. Bornscheuer, R.J. Kazlauskas, Mutations in distant residues moderately increase the enantioselectivity of Pseudomonas fluorescens esterase towards methyl 3bromo-2-methylpropanoate and ethyl 3phenylbutyrate, Chemistry 9 (2003), https://doi.org/10.1002/ chem.200204551, 1933e9. [45] D.M. McCandlish, E. Rajon, P. Shah, Y. Ding, J.B. Plotkin, The role of epistasis in protein evolution, Nature 497 (2013) E1eE2, https://doi.org/10.1038/ nature12219. [46] F.S. Tobias Kulschewski, Francesco Sasso, J.P. Marina Lotti, Molecular mechanism of deactivation of C.Antarctica lipase B by methanol, J. Biotechnol. 168 (2013) 462e469. [47] X. Fang, Y. Zhan, J. Yang, D. Yu, A concentration-dependent effect of methanol on Candida Antarctica lipase B in aqueous phase, J. Mol. Catal. B Enzym. 104 (2014) 1e7, https://doi.org/10.1016/j.molcatb.2014.03.002. [48] P.J. Fleming, F.M. Richards, Protein packing: dependence on protein size, secondary structure and amino acid composition, J. Mol. Biol. 299 (2000) 487e498, https://doi.org/10.1006/jmbi.2000.3750. [49] M. Gerstein, E.L. Sonnhammer, C. Chothia, Volume changes in protein evolution, J. Mol. Biol. 236 (1994) 1067e1078, https://doi.org/10.1016/00222836(94)90012-4. [50] F.M. Richards, The interpretation of protein structures: total volume, group volume distributions and packing density, J. Mol. Biol. 82 (1974) 1e14, https://doi.org/10.1016/0022-2836(74)90570-1. [51] K. Takano, Y. Yamagata, K. Yutani, Buried water molecules contribute to the conformational stability of a protein, Protein Eng. Des. Sel. 16 (2003) 5e9, https://doi.org/10.1093/proeng/gzg001. [52] M.R. Ganjalikhany, B. Ranjbar, A.H. Taghavi, T. Tohidi Moghadam, Functional motions of Candida Antarctica lipase B: a survey through open-close conforhttps://doi.org/10.1371/ mations, PLoS One 7 (2012), e40327, journal.pone.0040327. [53] D. Alvarez-garcia, X. Barril, Relationship between protein flexibility and binding: lessons for structure-based drug design, J. Chem. Theor. Comput. 10 (2014) 2608e2614, https://doi.org/10.1021/ct500182z. [54] S.P. Tiwari, N. Reuter, Similarity in shape dictates signature intrinsic dynamics despite No functional conservation in TIM barrel enzymes, PLoS Comput. Biol. 12 (2016), e1004834, https://doi.org/10.1371/journal.pcbi.1004834. [55] D. Jain, I. Pancha, S.K. Mishra, A. Shrivastav, S. Mishra, Purification and characterization of haloalkaline thermoactive, solvent stable and SDS-induced protease from Bacillus sp.: a potential additive for laundry detergents, Bioresour. Technol. 115 (2012) 228e236, https://doi.org/10.1016/ j.biortech.2011.10.081. [56] D.J. Huggins, Studying the role of cooperative hydration in stabilizing folded protein states, J. Struct. Biol. 196 (2016) 394e406, https://doi.org/10.1016/ j.jsb.2016.09.003. [57] E. Papaleo, M. Pasi, M. Tiberti, L. De Gioia, Molecular dynamics of mesophiliclike mutants of a cold-adapted enzyme: insights into distal effects induced by the mutations, PLoS One 6 (2011), e24214, https://doi.org/10.1371/ journal.pone.0024214. [58] H. Ogino, T. Uchiho, J. Yokoo, R. Kobayashi, R. Ichise, H. Ishikawa, Role of intermolecular disulfide bonds of the organic solvent-stable PST-01 protease in its organic solvent stability, Appl. Environ. Microbiol. 67 (2001) 942e947, https://doi.org/10.1128/AEM.67.2.942-947.2001. [59] J. Ottosson, L. Fransson, J.W. King, K. Hult, Size as a parameter for solvent effects on Candida Antarctica lipase B enantioselectivity, Biochim. Biophys. [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] 13 Acta Protein Struct. Mol. Enzymol. 1594 (2002) 325e334, https://doi.org/ 10.1016/S0167-4838(01)00324-7. R. Affleck, Z. Xut, V. Suzawa, K. Fochtt, D.S. Clark, J.S. Dordicktt, Enzymatic Catalysis and Dynamics in Low-Water Environments E20-, vol. 89, 1992, pp. 1100e1104. Z.S.H.A.B. TIDOR, Do salt bridges stabilize proteins - a continuum electrostatic analysis, Protein Sci. 3 (1994) 211e226, https://doi.org/10.1002/ pro.5560030206. R. Sinha, S.K. Khare, Effect of organic solvents on the structure and activity of moderately halophilic Bacillus sp. EMB9 protease, Extremophiles 18 (2014) 1057e1066, https://doi.org/10.1007/s00792-014-0683-4. R.M. Sabuti, M.R. Bozorgmehr, A. Morsali, Molecular dynamics simulations on the heterocyclic cyclodecapeptide and its linear analogous in water and octanol solvents, J. Mol. Liq. 229 (2017) 583e590, https://doi.org/10.1016/ j.molliq.2016.11.118. H. Kovacs, A.E. Mark, W.F. van Gunsteren, Solvent structure at a hydrophoic protein surface, Proteins 27 (1997) 395e404. X. Tang, M.J. Pikal, Design of freeze-drying processes for pharmaceuticals: practical advice, Pharm. Res. (N. Y.) 21 (2004) 191e200, https://doi.org/ 10.1023/B:PHAM.0000016234.73023.75. O. Kuchner, F.H. Arnold, Directed evolution of enzyme catalysts, Trends Biotechnol. 15 (1997) 523e530, https://doi.org/10.1016/S0167-7799(97)011384. T.A. Norio Hamamatsu, Y.H. Yukiko Nomiya, Motowo Nakajima1, Y. Shibanaka, Directed evolution by accumulating tailored mutations : thermostabilization of lactate oxidase with less trade-off with catalytic activity, Protein Eng. Des. Sel. 19 (2006) 483e489, https://doi.org/10.1093/protein/ gzl034. P. Yedavalli, N. Madhusudhana Rao, Engineering the loops in a lipase for stability in DMSO, Protein Eng. Des. Sel. 26 (2013) 317e324, https://doi.org/ 10.1093/protein/gzt002. A.M. Azevedo, D.M. Prazeres, J.M. Cabral, L.P. Fonseca, Stability of free and immobilised peroxidase in aqueous?organic solvents mixtures, J. Mol. Catal. B Enzym. 15 (2001) 147e153, https://doi.org/10.1016/S1381-1177(01)00017-0. T. Kawata, H. Ogino, Enhancement of the organic solvent-stability of the LST03 lipase by directed evolution, Biotechnol. Prog. 25 (2009), https://doi.org/ 10.1002/btpr.264. NA-NA. H.J. Park, J.C. Joo, K. Park, Y.J. Yoo, Stabilization of Candida Antarctica lipase B in hydrophilic organic solvent by rational design of hydrogen bond, Biotechnol. Bioproc. Eng. 17 (2012) 722e728, https://doi.org/10.1007/s12257012-0092-4. M.M. Abdullaeva, Effect of methanol on enzymatic synthesis of phosphatidylinositol, Chem. Nat. Coumpounds. 40 (2004) 139e140. Y. Liu, X. Zhang, H. Tan, Y. Yan, B.H. Hameed, Effect of pretreatment by different organic solvents on esterification activity and conformation of immobilized Pseudomonas cepacia lipase, Process Biochem. 45 (2010) 1176e1180, https://doi.org/10.1016/j.procbio.2010.03.023. H. Xu, X. Li, Z. Zhang, J. Song, Identifying coevolution between amino acid residues in protein families: advances in the improvement and evaluation of correlated mutation algorithms, Curr. Bioinf. 8 (2013) 148e160, https:// doi.org/10.2174/1574893611308020003. N.K.P. Kumari, M. V Jagannadham, Organic solvent induced refolding of acid denatured HEYNEIN : evidence of domains in the molecular, J. Protein Proteonomics 2 (2011) 11e21. doux, Analysis of G. Bellavia, L. Paccou, S. Achir, Y. Guinet, J. Siepmann, A. He bulk and hydration water during thermal lysozyme denaturation using Raman scattering, Food Biophys. 8 (2013) 170e176, https://doi.org/10.1007/ s11483-013-9294-3. K. Smail, N. Tchouar, M. Barj, B. Marekha, A. Idrissi, Luteolin organic solvent interactions. A molecular dynamics simulation analysis, J. Mol. Liq. 212 (2015) 503e508, https://doi.org/10.1016/j.molliq.2015.09.043. M. Khabiri, B. Minofar, J. Brezovský, J. Damborský, R. Ettrich, Interaction of organic solvents with protein structures at protein-solvent interface, J. Mol. Model. 19 (2013) 4701e4711, https://doi.org/10.1007/s00894-012-1507-z. €ttcher, U.T. Bornscheuer, High-throughput screening of activity and D. Bo enantioselectivity of esterases, Nat. Protoc. 1 (2006) 2340e2343, https:// doi.org/10.1038/nprot.2006.391. B. Reiter, a. Glieder, D. Talker, H. Schwab, Cloning and characterization of EstC from Burkholderia gladioli , a novel-type esterase related to plant enzymes, Appl. Microbiol. Biotechnol. 54 (2000) 778e785, https://doi.org/10.1007/ s002530000468. J. Schymkowitz, J. Borg, F. Stricher, R. Nys, F. Rousseau, L. Serrano, The FoldX web server: an online force field, Nucleic Acids Res. 33 (2005) W382eW388, https://doi.org/10.1093/nar/gki387. E. Krieger, G. Vriend, New ways to boost molecular dynamics simulations, J. Comput. Chem. 36 (2015) 996e1007, https://doi.org/10.1002/jcc.23899. A. Micsonai, F. Wien, L. Kernya, Y.H. Lee, Y. Goto, M. Refregiers, J. Kardos, Accurate secondary structure prediction and fold recognition for circular dichroism spectroscopy, Proc. Natl. Acad. Sci. U. S. A. 112 (2015) E3095eE3103, https://doi.org/10.1073/pnas.1500851112. N.J. Greenfield, Using circular dichroism spectra to estimate protein secondary structure, Nat. Protoc. 1 (2006) 2876e2890, https://doi.org/10.1038/ nprot.2006.202.Using. S.M. Kelly, N.C. Price, The application of circular dichroism to studies of protein folding and unfolding, Biochim. Biophys. Acta 1338 (1997) 161e185, J. Maiangwa, S.H. Hamdan, M.S. Mohamad Ali et al. Journal of Molecular Graphics and Modelling 105 (2021) 107897 analyzing packing in protein structures, Nucleic Acids Res. 37 (2009) D393eD395, https://doi.org/10.1093/nar/gkn769. €mmel, Voronoi cell: new method for allocation of [88] A. Goede, R. Preissner, C. Fro space among atoms: elimination of avoidable errors in calculation of atomic volume and density, J. Comput. Chem. 18 (1997) 1113e1123, https://doi.org/ 10.1002/(SICI)1096-987X(19970715)18:9<1113::AID-JCC1>3.0.CO;2-U. https://doi.org/10.1016/S0167-4838(96)00190-2. [86] V. Raussens, J.M. Ruysschaert, E. Goormaghtigh, Protein concentration is not an absolute prerequisite for the determination of secondary structure from circular dichroism spectra: a new scaling method, Anal. Biochem. 319 (2003) 114e121, https://doi.org/10.1016/S0003-2697(03)00285-9. [87] K. Rother, P.W. Hildebrand, A. Goede, B. Gruening, R. Preissner, Voronoia: 14