- Email: [email protected]
Development of a catalytically stable and efficient lipase through an increase in hydrophobicity of the oxyanion residue
Journal of Molecular Catalysis B: Enzymatic 122 (2015) 282–288
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic journal homepage: www.elsevier.com/locate/molcatb
Development of a catalytically stable and efficient lipase through an increase in hydrophobicity of the oxyanion residue Roswanira Abdul Wahab a,∗ , Mahiran Basri b,c,e,∗∗ , Raja Noor Zaliha Raja Abdul Rahman c,d,e , Abu Bakar Salleh c,d , Mohd Basyaruddin Abdul Rahman b,e , Thean Chor Leow c,d,e a
Department of Chemistry, Faculty of Science, Universiti Teknologi Malaysia, Skudai, Johor 81310, Malaysia Department of Chemistry, Faculty of Science, Universiti Putra Malaysia, Serdang, Selangor 43400, Malaysia c Institute of Bioscience, Universiti Putra Malaysia, UPM, Serdang, Selangor 43400, Malaysia d Department of Cell and Molecular Biology, Faculty of Biotechnology and Biomolecular Sciences, Universiti Putra Malaysia, Serdang, Selangor 43400, Malaysia e Enzyme and Microbial Technology Research Centre, Universiti Putra Malaysia, Serdang, Selangor 43400, Malaysia b
a r t i c l e
i n f o
Article history: Received 8 January 2015 Received in revised form 1 October 2015 Accepted 6 October 2015 Keywords: Geobacillus Lipase Oxyanion In-silico Site-directed mutagenesis
a b s t r a c t In-silico and empirical site-directed substitutions of oxyanion Q114 of the wildtype T1 lipase with that of hydrophobic Leu and Met residues were carried out to afford the Q114L and Q114M lipases, respectively. Using the esterification production of menthyl butyrate as a reaction model, evaluation on the catalytic efficiency of the three lipases was performed. It was found that Leu evidently improved the catalytic activity of the Q114L lipase, achieving the highest conversions of menthyl butyrate under varying experimental conditions that may be attributable to its lower total energy when compared with both the T1 and Q114M lipases. The diminishing catalytic activity of T1 lipase observable following substitution with Met (Q114M lipase) may be due to formation of three additional cavities within the vicinity of the mutation, lesser density of the protein core and susceptibility to high temperature, particularly under prolonged reaction time. Therefore, it can be concluded that the substitution of Leu into the oxyanion-114 had rendered favorable structural changes that enhanced the catalytic activity of the T1 lipase, envisaging that such approach may also be of applied value for improving the catalytic activity of other bacterial lipases within the I.5 family. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Hydrophobicity as well as high density of residues within the interior protein have been suggested as major factors affecting stabilization of protein structure [1,2]. A compromise between close packing and conformational strain to form a dense protein core packing is often required to overcome energetically unfavorable perturbations [3]. Such a compromise would lead to formation of cavities especially among large proteins [4,5], an inevitable
∗ Corresponding author. Department of Chemistry, Faculty of Science, Universiti Teknologi Malaysia, Skudai, Johor 81310, Malaysia. Fax: +60 7 556 6162. ∗∗ Corresponding author at: Department of Chemistry, Faculty of Science, Universiti Putra Malaysia, UPM, Serdang, Selangor 43400, Malaysia. Fax: +60 3 8943 2508. E-mail addresses: [email protected], [email protected] (R. Abdul Wahab), [email protected] (M. Basri). http://dx.doi.org/10.1016/j.molcatb.2015.10.003 1381-1177/© 2015 Elsevier B.V. All rights reserved.
aspect in protein architecture [1]. This is particularly true for folded proteins such as enzymes which hydrophobic amino acids are generally located deep within the protein for efficient packing [6]. It has been suggested that it is possible that he stability of the protein interior against various denaturants with a single amino acid substitution [7] that increases the hydrophobicity of the protein core. However, possibility of expansion in the size and increased number of cavity within the hydrophobic core (i.e., defects in atomic packing) may counteract the benefits of such modifications and destabilize the protein molecule [8]. Lipases (triacylglycerol hydrolases E.C.3.1.1.3) are industrially important enzymes that catalyze a broad range of reactions such as esterification, trans-esterification, organic synthesis under waterrestricted environment, and stereospecific hydrolysis of racemic esters. Many efforts have been made to further optimize their enzymatic properties through advances in site-directed mutagenesis
R. Abdul Wahab et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 282–288
and directed evolution [9]. In view of elucidating the possible role of amino acid sequence in stability of protein in lipases, molecular approaches that explore the effect of amino acid substitution on conformational stability of mutant proteins have been suggested [10]. Molecular approaches such as site-directed mutagenesis that occurs within the 11 Å distance from active site may alter enantioselectivity of the molecule more strongly than that of distant mutations [10]. Since such mutation may substantially change the topology of catalytic pocket and subsequently affect the stability as well as efficacy of lipase-catalyzed reactions; improved tolerance of lipases towards the different solvents, high temperatures, and denaturants as well as enantioselectivity may possibly be obtained [10–13]. Therefore, developing potentially more expedient approaches for creating new catalytically more stable and efficient lipases through protein mutation may prove useful. Previously, we reported on a lipase called T1 produced by the thermoalkalophilic Geobacillus zalihae. The lipase gene encoding 388 amino acids residues using vector pGEX 4-T1 was highly expressed in recombinant Escherichia coli BL21 (DE3) pLysS. The catalytic machinery is of T1 lipase consists of a triad of amino acids formed by Ser-114, His-359 and Asp-314 [14]. The oxyanion of T1 lipase was deduced to be Q114 and F16, based on another wellknown thermoalkalophilic lipase called BTL2 lipase of the Bacillus thermocanulatus. The BTL2 lipase has been successfully crystallized in open conformation which also revealed that residues Q115 and F17 formed the oxyanion of the enzyme. The F17 residue in BTL2 lipase was reportedly to be highly conserved [15], hence, it was as expected of the F16 in T1 lipase. More importantly, the BTL2 lipase shares a high sequence similarity to the T1 lipase, thereby, serves as an excellent reference template for the in-silico mutational studies. In this paper, site-directed mutagenesis on the oxyanion Q114 in the protein core of the wildtype T1 was performed whereby the hydrophilic Gln was substituted with hydrophobic, Leu and Met for producing variants Q114L and Q114M lipases, respectively. During in-silico investigations, FoldX and Voronoia 1.0 were used for predicting the conformational changes on the Q114L and Q114M structures and compared with that of wildtype T1 lipase. In addition, an empirical study for comparing the catalytic efficacy of lipases Q114L and Q114M with that of the wildtype T1 lipase was carried out using solvent-free esterification production of menthol butyrate as a reaction model. The efficiency of lipases Q114L, Q114M and T1 were systematically investigated under a variety of experimental conditions that included incubation time, temperature, enzyme amount, substrate molar ratio and agitation speed.
2. Materials and methods 2.1. Materials The components for the growth media for the lipases were purchased from Difco Laboratories (Detroit, USA). Menthol and butyric anhydride was purchased from Sigma–Aldrich (St. Louis, USA), and HPLC grade isooctane was obtained from Merck (Darmstadt, Germany). Antibiotics, chloramphenicol, and ampicillin were acquired from Amresco (Ohio, USA). Glutathione-Sepharose HP, Sephadex G25 were from GE Healthcare (Buckinghamshire, United Kingdom). Dithioreitol (DTT), phosphate buffer (pH 7.0 and pH 7.4), sodium hydroxide (NaOH), buffer Tris–HCl (pH 8.0), NaCl and CaCl2 were purchased from Sigma–Aldrich (St. Louis, USA). Amresco assay reagent and bovine serum albumin (BSA) were also procured from Sigma–Aldrich (St. Louis, USA). Amicon Ultra-15 centrifugal and 0.45-m membrane filter (Sartorius) filter were purchased from Millipore (Bedford, USA) and Sartorius (USA),
283
respectively. Distilled and deionized water were produced in our laboratory. 2.2. Computational studies: in-silico mutation, calculation of structural stability and protein compactness In-silico protein mutation and estimation of protein stability were performed using FoldX Version Beta 3.0. The crystal structure of T1 lipase (2DSN) of the thermoalkalophilic lipase from G. zalihae was retrieved from PDB file and the structure was repaired using FoldX for correction of residues with bad torsion angles. For accuracy, the presence of crystal water and metal within the protein structure were also taken into consideration. The three dimensional model of each lipase variant was constructed by the software and possible conformations of each residue following mutation were analyzed. Each lipase variant was subjected to simple energy minimization to remove high energy local minima. Finally, calculation of individual protein stability was performed and compared to the wild-type T1 lipase. The standalone program downloaded from the website http://bioinformatics.charite.de/voronoia [16] was used to calculate the packing density and total cavities of each lipase mutant. The Protein Data Bank files of the wildtype T1, Q114L and Q114M lipase, previously constructed in FoldX were submitted to the Voronoia 1.0 standalone program which calculated for the packing density and total cavities of each lipase. 2.3. Preparation of working culture and purification of enzyme The wildtype T1 lipase is from a thermoalkalophilic lipase called G. zalihae (2DSN) previously isolated from palm effluent. For the working culture, the Q114L, Q114M and wildtype T1 strains were revived from stock culture and grown in Luria Bertani (LB) broth supplemented with antibiotics ampicillin (50 g/mL) and chloramphenicol (35 g/mL). The cultures were centrifuged (10,000 rpm, 10 min), the liquid decanted and the pellet was re-suspended in distilled water prior to sonication (Branson 250 sonifier: output 2, duty cycle 30 and min 2). The cell lysate was cleared by centrifugation (12,000 rpm, 20 min) and the supernatant collected. The pH of supernatant was adjusted to pH 9.0 using NaOH (2M), lyophilized and stored in −20 ◦ C until further use. For the purification process, the cultures containing Q114L, Q114M and wildtype T1 lipase (1000 mL) were harvested by centrifugation, re-suspended in 40 mL phosphate-buffered saline (PBS; pH 7.4) containing 5 mM DTT, and sonicated (Branson 250 sonifier: output 2, duty cycle 30 and min 2). The cell lysate was cleared by centrifugation at 12,000 × g for 30 min and filtered with a 0.45-m membrane filter. Glutathione-Sepharose HP (10 mL) was packed into an XK 16/20 column and was equilibrated with 10 column volumes of PBS. The cleared cell lysate was loaded on the glutathione-Sepharose HP column at a flow rate of 0.25 mL/min. The column was washed with phosphate buffer solution (pH 7.0) until no protein was detected. The bound lipase was eluted with a buffer containing 100 mM Tris–HCl, 100 mM NaCl, and 0.33 mM CaCl2 , pH 8.0. The enzyme-containing fractions were determined by SDS-PAGE, pooled, and concentrated using Amicon Ultra-15 centrifugal filter and was tested for lipase activity. The concentrated solution was subjected to gel filtration on Sephadex G25 in an XK16/20 column. The fractions containing lipase activity were collected and concentrated with an Amicon Ultra-15 centrifugal unit. The homogeneity of the partially purified protein was confirmed by SDS-PAGE. The protein was lyophilized and stored at –20 ◦ C. 2.4. Standard lipase activity assay and protein concentration Lipase activity was determined according to a previously described method by Leow et al. [14] using an emulsion of olive oil
284
R. Abdul Wahab et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 282–288
and 50 mM glycine-NaOH (pH 9.0) as substrate. The protein content was determined by the method of Bradford (Bradford, 1976), utilizing the commercial Amresco assay reagent and bovine serum albumin (BSA) were used as standards. 2.5. Lipase-catalyzed esterification of menthol with butyric anhydride A typical esterification reaction was performed in a 30 mL mixture which consisted of menthol and butyric anhydride (1:1–4:1;butyric anhydride:menthol) in 100 mL screw-capped bottles. The lyophilized Q114L, Q114M and T1 lipase (1–5 mg/mL) were weighed and added to the reaction mixture. The reaction temperature investigated ranged from 40 to 90 ◦ C and the reaction was initiated by the addition of butyric anhydride to the reaction mixture and the mixture refluxed with stirring (100–400 rpm) in a paraffin oil bath. Aliquots (0.5 mL) of the reaction mixture were withdrawn periodically for analysis. A control without enzyme was also run in parallel under the same conditions. 2.6. Analytical methods The reaction was terminated by dilution in 5 mL isooctane and the solution was centrifuged (10, 000 rpm, 5 min) to remove the enzyme. An aliquot (1 mL) of the solution was removed and 1decanol (40 L) was added as internal standard. The sample was analyzed using gas chromatography (Perkin Elmer (model Clarus 600, USA) equipped with flame-ionization detector and Ultra 1 capillary column (25 m × 0.25 mm i.d. × 25 m film thickness). The temperature program was as follows: 110 to 150 ◦ C (4 ◦ C/min) and 200 ◦ C (5 ◦ C/min). The injector and detector temperatures were 230 and 250 ◦ C, respectively. The flow rate of the N2 carrier gas was 1 mL/min and the split ratio was 50:1. The composition of each reaction mixture was calculated from the concentration of the menthyl butyrate produced. 3. Results and discussion 3.1. Rationale for site-directed mutagenesis of the oxyanion Q114 Appreciation of the rules governing the protein core evolution and the restrictions in folding requirements are essential for designing proteins with superior folding properties. The study focused on the oxyanion hole, Q114 and F16, to redesign the protein core of the wildtype T1 for improved stability towards various factors that contribute to premature denaturation. It is pertinent to highlight that the oxyanion F17 in BTL2 is highly conserved within the bacterial lipases of the 1.5 family. Apart from maintaining proper substrate orientation for the binding of substrate during catalysis, the side-chain of F17 also shapes the oxyanion binding pocket [15]. Considering the fact that mutation on both Q114 and F16 may prove too extreme and cause impending diminished catalytic activity of T1, mutation on the F16 in T1 was not an option, which leaves only the Q114. Since the side-chain of Q114 has no role in shaping the oxyanion binding pocket, replacing Q114 with hydrophobic amino acids, Leu and Met may be more tolerable and less probable to cause adverse effects. It is anticipated that mutation on the Q114 may alter the shape and create new architectural landscape of the T1lipase binding pocket. Review of the literature reveals that the side-chain of a substituted hydrophobic amino acid residue can be packed more efficiently within the core of folded proteins (i.e., enzymes) than a hydrophilic one [6]. Furthermore, it is fascinating to see how changing the hydrophobicity the oxyanion Q114 in T1 can affect the catalytic activity and, potentially create new enzyme variants with novel substrate specificities [17]. The position of the oxyanion Q114 for site-directed the mutation study
Fig. 1. The three-dimensional depiction of the wildtype T1 lipase protein illustrated in ribbons showing (a) position of the Gln114 oxyanion relative to the the catalytic triad residues (Ser113, Asp317, His358) and (b) positions of neigbouring cavities (represented as green blobs) in the immediate vicinity of the mutation site. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
relative to the catalytic triad (S113, D317, H358) is illustrated in Fig. 1a. 3.2. Computational studies: in-silico protein mutation, calculation of structural stability, total number of cavity and protein compactness Site-directed mutagenesis experiments have supported the general view that the hydrophobic core of a protein represents a delicate balance between high packing density and the conformational strain needed to achieve it [18]. In-silico site-directed mutation on Q114 was executed to evaluate the structural quality of the Q114L and Q114M lipase protein relative to the wildtype T1 lipase. Factors evaluated include protein density, total number of cavities and total energy of each lipase molecule. Table 1 details the Goldman–Engelman–Steitz hydrophobicity index and the structural profile of Q114L, Q11M and wildtype T1 calculated using FoldX and Voronoia 1.0. It was revealed that the total energy for Q114L and Q114M were lower than that of the wildtype T1. FoldX calculated a decrease of 3.37 and 4.41 kcal/ mole, respectively, from the estimated total energy of T1. The reduced total energies of Q114L and Q114M signified enhanced stability of both lipases over the wildtype T1,
R. Abdul Wahab et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 282–288
285
Table 1 Summary of the hydrophobicity of oxyanion-114 and the structural profile of lipases Q114L, Q11M and T1 according to FoldX and Voronoia 1.0, respectively. Lipase
Q114L Q114M T1 lipase (Q114)
FoldX
Voronoia 1.0
Goldman–Engelman– Steitz hydropathy scale
Total energy (kcal/mol)
Energy difference (kcal/mol)
Total cavities
Protein compactness (van der Waals/total volume)
2.8 (hydrophobic) 3.4 (hydrophobic) 1.0 (hydrophilic)
−146.71 −147.75 −143.34
−3.37 −4.41 –
75 76 73
0.704 0.704 0.706
suggesting each lipase has adapted a new folded structure that improved intermolecular interactions and was energetically more favorable. Therefore, the decrease in total energy of Q114L and Q114M can be attributed to the replacement of the hydrophilic Gln (1.0) with the more hydrophobic Leu (2.8) and Met (3.4) [6] according to the Goldman–Engelman–Steitz hydropathy scale [19], respectively. This is possibly due to creation of new hydrophobic and van der Waals interactions between the side chain of Leu and Met with adjacent hydrophobic amino acids to result in a more rigid interior and stabilizing the protein packing. Eijsink et al. indicated that the packing within the hydrophobic core of a protein can be adjusted to accommodate a great variety of mutations as long as hydrophobicity is preserved [1]. In this context it is pertinent to point out that the larger decline in total energy of Q114M (−4.47 kcal/mol) as compared to Q114L (−3.37 kcal/mol) is consistent with the substitution the more hydrophobic Met (3.4) as opposed to Leu (2.8). In this study, results of FoldX afforded significant clues on the structural packing Q114L and Q114M lipase following substitution with Leu and Met, indicating a more efficiently packed hydrophobic core of both lipases than in T1 (Table 1). Apart from the estimation of total energy of the lipase, the quality of the protein packing can also be inferred from total number of cavity and compactness/density. The tabulated total number of cavity and compactness/density of Q114L, Q114M and wildtype T1 calculated by Voronoia 1.0 is illustrated in Table 1. Interestingly, the results indicated that substitution with Leu and Met increased the number of total cavity and reduced the compactness in Q114L and Q114M, respectively. The compactness of Q114L (0.704 vdW/total volume) and Q114M (0.704 vdW/total volume) were marginally lower than the wild type T1 (0.706 vdW/total volume), but remarkably, the latter had developed three new cavities as compared to the former which formed only two new ones. In is vital to point out that, the increased number of cavities and reduced compactness are features associated with less desirable enzyme structure. This can be attributed to the less dense enzyme core that is prone to premature unfolding when subjected to various denaturing conditions i.e., extreme temperature and pH. The higher number of total cavities and lower protein compactness in Q114L and Q114M confirmed that the three-dimensional structures of both lipase proteins were different to that of T1 but were less stable. Fig. 1b depicts the location of the cavities in the immediate vicinity of the substrate binding pocket of the wildtype T1. In this perspective, features such as a lower total cavity number and a more compact structure are regarded as advantageous, inferring less susceptibility of the mutant lipases towards denaturation. These structural changes are the result of a more efficient packing due to increased hydrophobic interaction, hence, the enzyme molecule is less likely to prematurely unwind when the reaction temperature is elevated or when subjected to extreme pH. Hence, the resultant mutant lipase is anticipated to be more thermostable. Creighton indicated that a compromise between close packing and conformational strain is often required as energetically unfavorable perturbations of protein chemistry will most certainly occur following a mutation, particularly when the mutation is within the tight core of a folded protein [3]. In view of the contradictive results acquired using FoldX and Voronoia 1.0, empirical
Table 2 The parametric values that gave maximal yield of menthyl butyrate for reactions catalyzed by lipases wildtype T1, Q114L and Q114M. Reaction condition
Wildtype T1
Q114L
Q114M
Incubation time (h) Temperature (◦ C) Amount of enzyme (mg/mL) Molar ratio (butyric anhydride/menthol) Agitation speed (rpm)
16 70 3 2:1 200
16 80 3 2:1 300
16 70 3 2:1 300
assessment merits consideration using a solvent-free esterification of menthol with butyric anhydride as a model reaction. The empirical approach was necessary to elucidate the role the hydrophobic amino acids Leu and Met on the catalytic efficacy of the wildtype T1, as well as to prove, once and for all, which residue to confer improved/diminished activity in T1. 3.3. Effect of reaction parameters on the lipase-catalyzed esterification of menthol with butyric anhydride 3.3.1. Effect of incubation time Time course study is a good indicator of enzyme performance as well as product yield. The effect of incubation time was monitored up to 30 h with sampling intervals of 0, 2, 4, 8, 16, 16, 24 and 30 h, respectively. Constant parameters utilized were temperature (60 ◦ C), enzyme amount (2 mg/mL), substrate molar ratio butyric anhydride/menthol (1:1) and stirring speed (200 rpm). Prior to the investigation, the lipase activity of Q114L, Q114 M and wild type T1 were determined as 126.3, 82.5 and 103 U/mg, respectively. Fig. 2a represents the reaction time profile for the esterification of butyric anhydride and menthol under constant parameters; temperature (60 ◦ C), enzyme amount (2 mg/mL), molar ratio butyric anhydride/menthol (1:1) and stirred at 200 rpm catalyzed by lipases Q114L and Q114M and T1 monitored over for a 30 h period. It can be seen that the conversion of menthyl butyrate increased with the increase in reaction time up to 16 h and declined thereafter. Remarkably, the Q114L was found to show improved conversion of menthyl butyrate, reaching the highest percentage conversion (92%) as compared to the wildtype T1 (84.6%) and Q114M lipase (80.3%) (Table 2). The higher percentage of menthyl butyrate in reactions catalyzed by Q114L can be attributed to an improved protein packing as a result of the smaller Leu side chain [20], implying that the oxyanion is easily adapted to substitution with Leu to result in the reduced total energy of Q114L. The methylene side of Leu in Q114L possibly promotes formation of hydrophobic interaction [21] with adjacent amino acids and consequently, reinforces protein rigidity within the catalytic pocket. Therefore, the active site of Q114L is made less flexible and can better retain its catalytically active shape, hence, less susceptible to denaturation under long incubation time, particularly under elevated reaction temperature. Conversely, the low percentage conversion of menthyl butyrate in reactions catalyzed by Q114M infers that Met could not be well accommodated within the Q114 site. The observation is in accordance with results obtained in Voronoia 1.0 that presented the Q114M having a higher number of total cavities and low protein
286
R. Abdul Wahab et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 282–288
Fig. 2. The esterification of menthol and butyric anhydride catalyzed by lipases T1, Q114L and Q114M. (a) Effect of incubation time at constant parameters: temperature (60 ◦ C), enzyme amount (2 mg), molar ratio butyric anhydride/menthol (1:1) and agitation speed (150 rpm), and (b) effect of temperature at constant parameters: time = 8 h, enzyme amount = 2 mg/mL, molar ratio butyric anhydride/menthol = 1:1 and stirring speed = 200 rpm.
compactness as compared to the T1 and Q114L. These changes can be inferred as the consequence of inefficient protein folding that resulted from protein perturbations, causing the inner core of Q114M to expand, hence, the observed reduction in compactness of Q114M. According to Leow et al., cavities are energetically unfavorable architecture in the structure of folded proteins as the surrounding amino acids contain fewer van der Waals interactions that could destabilize the protein [22]. This statement is evidently true for Q114M as the lipase consistently recorded the lowest percentage conversions of menthyl butyrate throughout the course of the reaction. The observation suggests that changes within the binding pocket following mutation to Q114 may have been more extensive that it was thought to be. It was highly probable that replacement of Q114 with Met had altered the binding affinity of Q114M to the substrate, thereby, adversely affecting catalytic efficiency and slower production of menthyl butyrate. Meanwhile, the apparent decrease in the yield of menthyl butyrate after 16 h was perhaps due to the counterproductive liberation of water in the reaction media that hydrolyzes the ester product [23], a common occurrence in many esterification reactions. The production of surplus water is a natural phenomenon and an unwanted byproduct of a prolonged esterification reaction [24], as when in excess, the enzyme-catalyzed hydrolysis process is favored and the equilibrium shifts towards reformation of the substrates. 3.3.2. Effect of temperature Since temperature can have an effect on the equilibrium of reaction as well as the activity and stability of the enzymes [25], such factor was also evaluated in this present study. The effects of temperature on the Q114L, Q114M and T1 catalyzed esterification was investigated by increasing the reaction temperature from 40 to 90 ◦ C (Fig. 2b), under constant parameters; incubation time (8 h), enzyme amount (2 mg/mL), substrate molar ratio butyric anhydride/menthol (1:1) and stirring speed (200 rpm). Interestingly, results revealed high conversions of menthyl butyrate for lipases T1 (83.1%) and Q114M (±83.3%), as well as the Q114L lipase (80 ◦ C) were approximately at 70 ◦ C (T1 and Q114M) and 80 ◦ C (Q114L), respectively. Production of the ester was at the lowest for Q114L, Q114M and T1 at 40 ◦ C, yielding only 28, 35.1 and 38.5% of menthyl butyrate, respectively (Table 2). The high yield of menthyl butyrate achieved by Q114L at a considerably higher temperature (80 ◦ C) over lipases Q114M and T1 suggests that substitution with Leu has positively improved the activity and stability of Q114L under elevated temperature. The shift in the optimum temperature may be due to the hydrophobicity of Leu (−2.8) that altered the molecular interactions between amino acids close to the mutation point. In fact, Leu has been described as an acceptable substitute to the buried Gln due to its
similarity in terms of hydrophobicity [20,21] with other hydrophobic amino acids that predominates the protein core of T1. The increased interactions between the adjacent amino acids and new methylene side chain of Leu114 can effectively facilitate local structure folding [26] and reduce susceptibility of Q114L to high temperature deactivation. However, despite the presence of the more hydrophobic Met, the optimum temperature of Q114M remained unchanged. This may largely be due to the formation of three additional cavities that counteracts the benefit of increased molecular interactions between adjacent amino acids in the vicinity of the mutation. Alternatively, the overall decline in the yield of menthyl butyrate at conditions exceeding the optimal temperature of the lipases is illustrative of increased deactivation rate, signifying partial lipase inactivation, usually associated with excessive protein unfolding that causes enzymes to partially lose their catalytically competent form [27,28]. Alternatively, the low yield of menthyl butyrate at low reaction temperatures (40–50 ◦ C) is attributable to the fact that the protein structures of Q114M and Q114L, also variants of the thermophilic T1, were still rigid and have yet to achieve their catalytically active shape [29]. A compromise between high reaction temperature and solubility of reactants is often necessary to attain efficient biocatalysis, as at the right temperature, solubility of reactants is increased and catalytic activity of thermophiles is at their optimum, thereby accelerating the conversion of products [30]. 3.3.3. Effect of enzyme amount Fig. 3a details the reaction profile for the Q114L, Q114M and wildtype T1 in catalyzing the conversion of menthyl butyrate following the use of increasing enzyme loading (1–5 mg/mL) under fixed parameters; incubation time (8 h), temperature (60 ◦ C), substrate molar ratio butyric anhydride/menthol (1:1) and stirring speed (200 rpm). It was observed that the highest conversions of menthyl butyrate for the reactions by the Q114L, Q114M and wildtype T1 were all attained at 3 mg/mL, yielding up to 68.5, 42.1 and 68.9% of the ester, respectively (Table 2). The percent conversion of menthyl butyrate dropped in all reactions when the enzyme amount exceeded 3 mg/mL, indicating ineffective use of the biocatalysts at high enzyme loading. Beyond this concentration, substantial decrease in conversion of menthyl butyrate prevailed, usually associated with the high number of enzyme molecules in the reaction medium, apart from inefficient mixing that resulted in reduced effective enzyme-substrate collisions and rate of reaction. Review of the literature also revealed a similar trend whenever a high enzyme amount is used [29,31,32]. Thus, similar trend in Q114L, Q114M and T1 catalyzed synthesis may too prove common; suggesting further additions of the enzyme into the reaction is futile and will pointlessly boosts the production cost of menthyl
R. Abdul Wahab et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 282–288
287
Fig. 3. The esterification of menthol and butyric anhydride catalyzed by lipases wildtype T1, Q114L and Q114M. (a) Effect of enzyme amount at constant parameters: incubation time = 8 h, temperature = 60 ◦ C, molar ratio butyric anhydride/menthol = 1:1 and stirring speed = 200 rpm, (b) effect of substrate molar ratio of butyric anhydride/menthol at constant parameters: incubation time = 8 h, temperature = 60 ◦ C, enzyme amount = 2 mg/mL and stirring speed = 200rpm.
butyrate. Correspondingly, the low yield is also associated with the phenomenon of agglomeration whereby lipase molecules cluster and aggregate [33] to form inhomogeneous enzyme distribution which impedes accessibility of substrates to the enzymes, hence, the decrease in reaction yields. 3.3.4. Effect of substrate molar ratio Considering that molar ratio of anhydride:menthol can influence the rates of reactions [34] as well as the activity of the lipases [35], such factor too was examined in this study. The influence of substrate molar ratio in the production of menthyl butyrate catalyzed by lipases Q114L, Q114M and T1 was determined at ratios of butyric anhydride:menthol that ranged between 1:1–4:1 (Fig. 3b). Remarkably, substitution of the hydrophilic oxyanion Q114 with Leu and Met did not affect the optimum substrate molar ratio of substrates but effectively improve activity of Q114L as compared to T1 and Q114M. The highest conversion of menthyl butyrate for Q114L, Q114M and wildtype T1 were achieved at anhydride:menthol molar ratio of 2:1, yielding up to a 71, 52 and 64% of the ester, correspondingly (Table 2). Percent conversion of menthyl butyrate was the lowest for Q114L, Q114M and wildtype T1 at an anhydride/menthol molar ratio of 4:1, affording only 51, 37 and 55%, respectively. The noteworthy higher percentage of menthyl butyrate in Q114L-catalyzed reactions substantiates the enhanced catalytic activity of Q114L as compared to the Q114M and wildtype T1. The reduction in yield of menthyl butyrate at molar ratio exceeding 2:1 in all reactions is due to the inhibitory effect of excessive butyric anhydride present in the reaction medium or by formation of butyric acid [36]. A rather low yield of menthyl butyrate was also recorded at equimolar of anhydride:alcohol, is usually due to the increased dehydrating effects of the menthol molecules on the structures [37] whereby in most cases, the reaction substrates are inherently disruptive towards the structurally sensitive active form of the lipases, reducing the rate of reaction and yield of ester. Apart from the dehydrating effects, the excess menthol molecules also compete for the same active site [34]. 3.3.5. Effect of stirring speed The rate of stirring is important in many enzyme catalyzed processes as proper integration of reaction components is needed to obtain a homogeneous mixture of substrates and good contact between the enzyme molecules and substrates. The reaction profile based on increasing stirring rate speed (100–400 rpm) is detailed in Fig. 4. The results indicated that the percentage conversion of menthyl butyrate in reactions catalyzed by Q114L, Q114M and T1 positively correlated with the increase in the stirring speed, with lipases Q114L and Q114M attaining the highest conversion (72.2
Fig. 4. The esterification of menthol and butyric anhydride catalyzed by lipases wildtype T1, Q114L and Q114M. Effect of agitation speed at constant parameters: incubation time = 16 h, temperature = 60 ◦ C, enzyme amount = 2 mg/mL and molar ratio butyric anhydride/menthol = 1:1.
and 63.1%, respectively) of menthyl butyrate at 300 rpm, and the wildtype T1 (65.3%) at a lower stirring speed of 200 rpm. In a solvent-free biotransformation process, factors such as mixing and mass transfer require due consideration in the absence of solvent as the suspending medium. Mass and heat transfer play an important role under such condition as these factors affect the substrate availability as well as removal of inhibitory cellular metabolites from the biocatalyst [38]. The higher percentage of menthyl butyrate at increased stirring speed is attributable to a more efficient transfer of mass and heat throughout the reaction medium [38]. This promotes effective enzyme-substrate collisions that accelerate reaction rates [30] of the Q114L, Q114M and T1 lipases, and reduces viscosity of the substrates to result in higher reaction yields. 4. Conclusions This present study revealed that substituting the hydrophilic oxyanion Gln in the wildtype T1 with Leu (Q114L lipase) would considerably improve the catalytic activity of this lipase enzyme that can be attributable to lower total molecular energy predicted by FoldX. In contrast, such improved in the catalytic activity was not observed when the hydrophilic oxyanion Gln in the wildtype T1 was substituted with Met (Q114M lipase). In addition, increased in
288
R. Abdul Wahab et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 282–288
localized hydrophobicity may potentially enhance the hydrophobicity as well as van der Waals interactions, allowing the catalytic pocket to retain its active shape for longer duration. Contrasting to Leu, substituting Gln with Met at the oxyanion-114 would create additional cavities in the molecule that subsequently alter catalytic pocket of Q114M lipase due to the larger side chain of Met when compared with Leu; a problematic approach for enhancing the catalytic activity of the enzyme. Hence, it can be construed that such substitution as Leu into the wildtype T1 at oxyanion-114 had conferred favorable structural changes that rendered better catalytic activity; an approach that may prove suitable for improving the catalytic activity of other bacterial lipases within the I.5 family. Acknowledgements The authors would like to thank Dr Naji Arafat for proof reading this manuscript. This project was funded by a grant from Ministry of Science and Technology and Innovation, grant no. 5487729 (Green Chemistry), Malaysia. References [1] V.G.H. Eijsink, B.W. Djikstra, G. Vriend, J.R. van der Zee, O.R. Veltman, B. van der Vinne, G. van den Burg, S. Kempe, G. Venema, Protein Eng. 5 (1992) 421–426. [2] C.N. Pace, B.A. Shirley, M. McNutt, K. Gajiwala, FASEB J. 10 (1996) 75–83. [3] T.E. Creighton, Curr. Opin. Struc. Biol. 1 (1991) 5–6. [4] M. Eilers, S.C. Shekar, T. Shieh, S.O. Smith, P.J. Fleming, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 5796–5801. [5] T. Ohmura, T. Ueda, K. Ootsuka, M. Saito, T. Imoto, Protein Sci. 10 (2001) 313–320. [6] J. Liang, K.A. Dill, Biophys. J. 81 (2001) 751–766. [7] K. Yutani, K. Ogasahara, T. Tsujuta, Y. Sugino, Proc. Natl. Acad. Sci. U. S. A. 84 (1987) 4441–4444. [8] S. Sonavane, P. Chakrabarti, PLoS Comput. Biol. 4 (2008) e1000188, http://dx. doi.org/10.1371/journal.pcbi.1000188. [9] R.A. Wahab, M. Basri, M.B. Abdul Rahman, R.N.Z. Raja Abdul Rahman, A.B. Salleh, T.C. Leow, Int. J. Mol. Sci. 13 (2012) 11666–11680. [10] K.L. Morley, R.J. Kazlauskas, Trends Biotechnol. 23 (2005) 231–237. [11] A. Nobili, M.G. Gall, I.V. Pavlidis, M.L. Thompson, M. Schmidt, U.T. Bornscheur, FEBS J. (2013), http://dx.doi.org/10.1111/febs.12137. [12] J. Gu, L. Ye, F.L.X. Guo, W. Lu, H. Yu, Appl. Microbiol. Biotechnol. (2014), http:// dx.doi.org/10.1007/s00253-014-5995-x.
[13] V. Stepankova, T. Bidmanova, T. Kudelakova, Z. Prokop, R. Chaloupkova, J. Damborsky, ACS Catal. 3 (2013) 2823–2836. [14] T.C. Leow, R.N.Z. Rahman, M. Basri, A.B. Salleh, Biosci. Biotechnol. Biochem. 68 (2004) 96–203. [15] C. Carrasco-Lopez, C. Godoy, B. de las Rivas, G. Fernandez-Lafuente, J.M. Palomo, J.M. Guisan, R. Fernandez-Lafuente, M. Martinez-Ripoll, J.A. Hermoso, J. Biol. Chem. 284 (2009) 4365–4372. [16] http://bioinformatics.charite.de/voronoia. [17] A. Nayeem, S.J. Chiang, S.W. Liu, Y. Sun, L. You, J. Basch, Protein Eng. Des. Sel. 22 (2009) 257–266. [18] W.C. Sandberg, T.C. Terwilliger, Proc. Natl. Acad. Sci. U. S. A. 74 (1991) 643–657. [19] D.M. Engelman, T.A. Steitz, A. Goldman, Ann. Rev. Biophys. Biophys. Chem. 15 (1986) 321–353. [20] R.A. Wahab, M.B. Abdul Rahman, U.H. Zaidan, M. Basri, R.N.Z. Raja Abdul Rahman, A.B. Salleh, T.C. Leow, Adv. Biosci. Bioeng. 3 (2012) 158–167. [21] K. Ishikawa, H. Nakamura, K. Morikawa, S. Kanaya, Biochemistry 32 (1993) 6171–6178. [22] T.C. Leow, R.N.Z. Raja Abdul Rahman, M. Basri, A.B. Salleh, Extreme 11 (2005) 527–535. [23] A.L. Ong, A.H. Kamaruddin, W.S. Long, S.T. Lim, R. Kumari, Enzyme Microb. Technol. 39 (2006) 924–929. [24] Q.X. Song, D.Z. Wei, J. Mol. Catal. B: Enzym. 18 (2002) 261–266. [25] D. Zhao, E. Xun, J. Wang, R. Wang, X. We, L. Wang, Z. Wang, Biotechnol. Bioprocess Eng. 6 (2011) 638–644. [26] E. Eyal, R. Najmanovich, M. Edelman, V. Sobolev, Protein Struct. Funct. Bioinform. 50 (2003) 272–282. [27] M.L. Foresti, M. Galle, M.L. Ferreira, L.E. Briand, J. Chem. Technol. Biotechnol. 84 (2009) 1461–1473. [28] S. Sun, L. Shan, Y. Liu, Q. Jin, Y. Son, X. Wang, J. Mol. Catal. B: Enzym. 57 (2009) 104–108. [29] R.A. Wahab, M. Basri, R.N.Z. Abdul Rahman, M.B. Abdul Rahman, A.B. Ariff, A.B. Salleh, N. Chaibakhsh, T.C. Leow, Biotechnol. Biotechnol. Equip. 28 (2014) 1065–1072. [30] M.B. Abdul Rahman, N. Chaibakhsh, M. Basri, R.N.Z. Abdul Rahman, A.B. Salleh, S. Md Radzi, J. Chem. Technol. Biotechnol. 83 (2008) 1534–1540. [31] C. Dianoczki, R. Recseg, K. Kovari, P. Poppe Laszlo, J. Mol. Catal. B: Enzym. 45 (2007) 45–49. [32] K.M.W. Syamsul, M.R. Salina, S.O. Siti, M.N. Hanina, M.A.R. Basyaruddin, K. Jusoff, World Appl. Sci. J. 11 (2010) 401–407. [33] G. Trubiano, D. Borio, A. Errazu, Enzyme Microb. Technol. 40 (2007) 716–722. [34] P. Pires-Cabral, M.M.R. da, S. Ferreira-Dias, Biochem. Eng. J. 43 (2009) 327–332. [35] D. Li, W. Xiaojing, N. Kaili, W. Fang, L. Junfeng, W. Pu, T. Tianwei, Chin. J. Chem. Eng. 19 (2011) 978–982. [36] E.T. Liaw, K.J. Liu, Bioresour. Technol. 101 (2010) 3320–3324. [37] M.L. Foresti, A. Errazu, M.L. Ferreira, Biochem. Eng. J. 25 (2005) 69–77. [38] V.S. Meena, U.C. Banerjee, Ind. J. Biotechnol. 10 (2011) 452–457.
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic journal homepage: www.elsevier.com/locate/molcatb
Development of a catalytically stable and efficient lipase through an increase in hydrophobicity of the oxyanion residue Roswanira Abdul Wahab a,∗ , Mahiran Basri b,c,e,∗∗ , Raja Noor Zaliha Raja Abdul Rahman c,d,e , Abu Bakar Salleh c,d , Mohd Basyaruddin Abdul Rahman b,e , Thean Chor Leow c,d,e a
Department of Chemistry, Faculty of Science, Universiti Teknologi Malaysia, Skudai, Johor 81310, Malaysia Department of Chemistry, Faculty of Science, Universiti Putra Malaysia, Serdang, Selangor 43400, Malaysia c Institute of Bioscience, Universiti Putra Malaysia, UPM, Serdang, Selangor 43400, Malaysia d Department of Cell and Molecular Biology, Faculty of Biotechnology and Biomolecular Sciences, Universiti Putra Malaysia, Serdang, Selangor 43400, Malaysia e Enzyme and Microbial Technology Research Centre, Universiti Putra Malaysia, Serdang, Selangor 43400, Malaysia b
a r t i c l e
i n f o
Article history: Received 8 January 2015 Received in revised form 1 October 2015 Accepted 6 October 2015 Keywords: Geobacillus Lipase Oxyanion In-silico Site-directed mutagenesis
a b s t r a c t In-silico and empirical site-directed substitutions of oxyanion Q114 of the wildtype T1 lipase with that of hydrophobic Leu and Met residues were carried out to afford the Q114L and Q114M lipases, respectively. Using the esterification production of menthyl butyrate as a reaction model, evaluation on the catalytic efficiency of the three lipases was performed. It was found that Leu evidently improved the catalytic activity of the Q114L lipase, achieving the highest conversions of menthyl butyrate under varying experimental conditions that may be attributable to its lower total energy when compared with both the T1 and Q114M lipases. The diminishing catalytic activity of T1 lipase observable following substitution with Met (Q114M lipase) may be due to formation of three additional cavities within the vicinity of the mutation, lesser density of the protein core and susceptibility to high temperature, particularly under prolonged reaction time. Therefore, it can be concluded that the substitution of Leu into the oxyanion-114 had rendered favorable structural changes that enhanced the catalytic activity of the T1 lipase, envisaging that such approach may also be of applied value for improving the catalytic activity of other bacterial lipases within the I.5 family. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Hydrophobicity as well as high density of residues within the interior protein have been suggested as major factors affecting stabilization of protein structure [1,2]. A compromise between close packing and conformational strain to form a dense protein core packing is often required to overcome energetically unfavorable perturbations [3]. Such a compromise would lead to formation of cavities especially among large proteins [4,5], an inevitable
∗ Corresponding author. Department of Chemistry, Faculty of Science, Universiti Teknologi Malaysia, Skudai, Johor 81310, Malaysia. Fax: +60 7 556 6162. ∗∗ Corresponding author at: Department of Chemistry, Faculty of Science, Universiti Putra Malaysia, UPM, Serdang, Selangor 43400, Malaysia. Fax: +60 3 8943 2508. E-mail addresses: [email protected], [email protected] (R. Abdul Wahab), [email protected] (M. Basri). http://dx.doi.org/10.1016/j.molcatb.2015.10.003 1381-1177/© 2015 Elsevier B.V. All rights reserved.
aspect in protein architecture [1]. This is particularly true for folded proteins such as enzymes which hydrophobic amino acids are generally located deep within the protein for efficient packing [6]. It has been suggested that it is possible that he stability of the protein interior against various denaturants with a single amino acid substitution [7] that increases the hydrophobicity of the protein core. However, possibility of expansion in the size and increased number of cavity within the hydrophobic core (i.e., defects in atomic packing) may counteract the benefits of such modifications and destabilize the protein molecule [8]. Lipases (triacylglycerol hydrolases E.C.3.1.1.3) are industrially important enzymes that catalyze a broad range of reactions such as esterification, trans-esterification, organic synthesis under waterrestricted environment, and stereospecific hydrolysis of racemic esters. Many efforts have been made to further optimize their enzymatic properties through advances in site-directed mutagenesis
R. Abdul Wahab et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 282–288
and directed evolution [9]. In view of elucidating the possible role of amino acid sequence in stability of protein in lipases, molecular approaches that explore the effect of amino acid substitution on conformational stability of mutant proteins have been suggested [10]. Molecular approaches such as site-directed mutagenesis that occurs within the 11 Å distance from active site may alter enantioselectivity of the molecule more strongly than that of distant mutations [10]. Since such mutation may substantially change the topology of catalytic pocket and subsequently affect the stability as well as efficacy of lipase-catalyzed reactions; improved tolerance of lipases towards the different solvents, high temperatures, and denaturants as well as enantioselectivity may possibly be obtained [10–13]. Therefore, developing potentially more expedient approaches for creating new catalytically more stable and efficient lipases through protein mutation may prove useful. Previously, we reported on a lipase called T1 produced by the thermoalkalophilic Geobacillus zalihae. The lipase gene encoding 388 amino acids residues using vector pGEX 4-T1 was highly expressed in recombinant Escherichia coli BL21 (DE3) pLysS. The catalytic machinery is of T1 lipase consists of a triad of amino acids formed by Ser-114, His-359 and Asp-314 [14]. The oxyanion of T1 lipase was deduced to be Q114 and F16, based on another wellknown thermoalkalophilic lipase called BTL2 lipase of the Bacillus thermocanulatus. The BTL2 lipase has been successfully crystallized in open conformation which also revealed that residues Q115 and F17 formed the oxyanion of the enzyme. The F17 residue in BTL2 lipase was reportedly to be highly conserved [15], hence, it was as expected of the F16 in T1 lipase. More importantly, the BTL2 lipase shares a high sequence similarity to the T1 lipase, thereby, serves as an excellent reference template for the in-silico mutational studies. In this paper, site-directed mutagenesis on the oxyanion Q114 in the protein core of the wildtype T1 was performed whereby the hydrophilic Gln was substituted with hydrophobic, Leu and Met for producing variants Q114L and Q114M lipases, respectively. During in-silico investigations, FoldX and Voronoia 1.0 were used for predicting the conformational changes on the Q114L and Q114M structures and compared with that of wildtype T1 lipase. In addition, an empirical study for comparing the catalytic efficacy of lipases Q114L and Q114M with that of the wildtype T1 lipase was carried out using solvent-free esterification production of menthol butyrate as a reaction model. The efficiency of lipases Q114L, Q114M and T1 were systematically investigated under a variety of experimental conditions that included incubation time, temperature, enzyme amount, substrate molar ratio and agitation speed.
2. Materials and methods 2.1. Materials The components for the growth media for the lipases were purchased from Difco Laboratories (Detroit, USA). Menthol and butyric anhydride was purchased from Sigma–Aldrich (St. Louis, USA), and HPLC grade isooctane was obtained from Merck (Darmstadt, Germany). Antibiotics, chloramphenicol, and ampicillin were acquired from Amresco (Ohio, USA). Glutathione-Sepharose HP, Sephadex G25 were from GE Healthcare (Buckinghamshire, United Kingdom). Dithioreitol (DTT), phosphate buffer (pH 7.0 and pH 7.4), sodium hydroxide (NaOH), buffer Tris–HCl (pH 8.0), NaCl and CaCl2 were purchased from Sigma–Aldrich (St. Louis, USA). Amresco assay reagent and bovine serum albumin (BSA) were also procured from Sigma–Aldrich (St. Louis, USA). Amicon Ultra-15 centrifugal and 0.45-m membrane filter (Sartorius) filter were purchased from Millipore (Bedford, USA) and Sartorius (USA),
283
respectively. Distilled and deionized water were produced in our laboratory. 2.2. Computational studies: in-silico mutation, calculation of structural stability and protein compactness In-silico protein mutation and estimation of protein stability were performed using FoldX Version Beta 3.0. The crystal structure of T1 lipase (2DSN) of the thermoalkalophilic lipase from G. zalihae was retrieved from PDB file and the structure was repaired using FoldX for correction of residues with bad torsion angles. For accuracy, the presence of crystal water and metal within the protein structure were also taken into consideration. The three dimensional model of each lipase variant was constructed by the software and possible conformations of each residue following mutation were analyzed. Each lipase variant was subjected to simple energy minimization to remove high energy local minima. Finally, calculation of individual protein stability was performed and compared to the wild-type T1 lipase. The standalone program downloaded from the website http://bioinformatics.charite.de/voronoia [16] was used to calculate the packing density and total cavities of each lipase mutant. The Protein Data Bank files of the wildtype T1, Q114L and Q114M lipase, previously constructed in FoldX were submitted to the Voronoia 1.0 standalone program which calculated for the packing density and total cavities of each lipase. 2.3. Preparation of working culture and purification of enzyme The wildtype T1 lipase is from a thermoalkalophilic lipase called G. zalihae (2DSN) previously isolated from palm effluent. For the working culture, the Q114L, Q114M and wildtype T1 strains were revived from stock culture and grown in Luria Bertani (LB) broth supplemented with antibiotics ampicillin (50 g/mL) and chloramphenicol (35 g/mL). The cultures were centrifuged (10,000 rpm, 10 min), the liquid decanted and the pellet was re-suspended in distilled water prior to sonication (Branson 250 sonifier: output 2, duty cycle 30 and min 2). The cell lysate was cleared by centrifugation (12,000 rpm, 20 min) and the supernatant collected. The pH of supernatant was adjusted to pH 9.0 using NaOH (2M), lyophilized and stored in −20 ◦ C until further use. For the purification process, the cultures containing Q114L, Q114M and wildtype T1 lipase (1000 mL) were harvested by centrifugation, re-suspended in 40 mL phosphate-buffered saline (PBS; pH 7.4) containing 5 mM DTT, and sonicated (Branson 250 sonifier: output 2, duty cycle 30 and min 2). The cell lysate was cleared by centrifugation at 12,000 × g for 30 min and filtered with a 0.45-m membrane filter. Glutathione-Sepharose HP (10 mL) was packed into an XK 16/20 column and was equilibrated with 10 column volumes of PBS. The cleared cell lysate was loaded on the glutathione-Sepharose HP column at a flow rate of 0.25 mL/min. The column was washed with phosphate buffer solution (pH 7.0) until no protein was detected. The bound lipase was eluted with a buffer containing 100 mM Tris–HCl, 100 mM NaCl, and 0.33 mM CaCl2 , pH 8.0. The enzyme-containing fractions were determined by SDS-PAGE, pooled, and concentrated using Amicon Ultra-15 centrifugal filter and was tested for lipase activity. The concentrated solution was subjected to gel filtration on Sephadex G25 in an XK16/20 column. The fractions containing lipase activity were collected and concentrated with an Amicon Ultra-15 centrifugal unit. The homogeneity of the partially purified protein was confirmed by SDS-PAGE. The protein was lyophilized and stored at –20 ◦ C. 2.4. Standard lipase activity assay and protein concentration Lipase activity was determined according to a previously described method by Leow et al. [14] using an emulsion of olive oil
284
R. Abdul Wahab et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 282–288
and 50 mM glycine-NaOH (pH 9.0) as substrate. The protein content was determined by the method of Bradford (Bradford, 1976), utilizing the commercial Amresco assay reagent and bovine serum albumin (BSA) were used as standards. 2.5. Lipase-catalyzed esterification of menthol with butyric anhydride A typical esterification reaction was performed in a 30 mL mixture which consisted of menthol and butyric anhydride (1:1–4:1;butyric anhydride:menthol) in 100 mL screw-capped bottles. The lyophilized Q114L, Q114M and T1 lipase (1–5 mg/mL) were weighed and added to the reaction mixture. The reaction temperature investigated ranged from 40 to 90 ◦ C and the reaction was initiated by the addition of butyric anhydride to the reaction mixture and the mixture refluxed with stirring (100–400 rpm) in a paraffin oil bath. Aliquots (0.5 mL) of the reaction mixture were withdrawn periodically for analysis. A control without enzyme was also run in parallel under the same conditions. 2.6. Analytical methods The reaction was terminated by dilution in 5 mL isooctane and the solution was centrifuged (10, 000 rpm, 5 min) to remove the enzyme. An aliquot (1 mL) of the solution was removed and 1decanol (40 L) was added as internal standard. The sample was analyzed using gas chromatography (Perkin Elmer (model Clarus 600, USA) equipped with flame-ionization detector and Ultra 1 capillary column (25 m × 0.25 mm i.d. × 25 m film thickness). The temperature program was as follows: 110 to 150 ◦ C (4 ◦ C/min) and 200 ◦ C (5 ◦ C/min). The injector and detector temperatures were 230 and 250 ◦ C, respectively. The flow rate of the N2 carrier gas was 1 mL/min and the split ratio was 50:1. The composition of each reaction mixture was calculated from the concentration of the menthyl butyrate produced. 3. Results and discussion 3.1. Rationale for site-directed mutagenesis of the oxyanion Q114 Appreciation of the rules governing the protein core evolution and the restrictions in folding requirements are essential for designing proteins with superior folding properties. The study focused on the oxyanion hole, Q114 and F16, to redesign the protein core of the wildtype T1 for improved stability towards various factors that contribute to premature denaturation. It is pertinent to highlight that the oxyanion F17 in BTL2 is highly conserved within the bacterial lipases of the 1.5 family. Apart from maintaining proper substrate orientation for the binding of substrate during catalysis, the side-chain of F17 also shapes the oxyanion binding pocket [15]. Considering the fact that mutation on both Q114 and F16 may prove too extreme and cause impending diminished catalytic activity of T1, mutation on the F16 in T1 was not an option, which leaves only the Q114. Since the side-chain of Q114 has no role in shaping the oxyanion binding pocket, replacing Q114 with hydrophobic amino acids, Leu and Met may be more tolerable and less probable to cause adverse effects. It is anticipated that mutation on the Q114 may alter the shape and create new architectural landscape of the T1lipase binding pocket. Review of the literature reveals that the side-chain of a substituted hydrophobic amino acid residue can be packed more efficiently within the core of folded proteins (i.e., enzymes) than a hydrophilic one [6]. Furthermore, it is fascinating to see how changing the hydrophobicity the oxyanion Q114 in T1 can affect the catalytic activity and, potentially create new enzyme variants with novel substrate specificities [17]. The position of the oxyanion Q114 for site-directed the mutation study
Fig. 1. The three-dimensional depiction of the wildtype T1 lipase protein illustrated in ribbons showing (a) position of the Gln114 oxyanion relative to the the catalytic triad residues (Ser113, Asp317, His358) and (b) positions of neigbouring cavities (represented as green blobs) in the immediate vicinity of the mutation site. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
relative to the catalytic triad (S113, D317, H358) is illustrated in Fig. 1a. 3.2. Computational studies: in-silico protein mutation, calculation of structural stability, total number of cavity and protein compactness Site-directed mutagenesis experiments have supported the general view that the hydrophobic core of a protein represents a delicate balance between high packing density and the conformational strain needed to achieve it [18]. In-silico site-directed mutation on Q114 was executed to evaluate the structural quality of the Q114L and Q114M lipase protein relative to the wildtype T1 lipase. Factors evaluated include protein density, total number of cavities and total energy of each lipase molecule. Table 1 details the Goldman–Engelman–Steitz hydrophobicity index and the structural profile of Q114L, Q11M and wildtype T1 calculated using FoldX and Voronoia 1.0. It was revealed that the total energy for Q114L and Q114M were lower than that of the wildtype T1. FoldX calculated a decrease of 3.37 and 4.41 kcal/ mole, respectively, from the estimated total energy of T1. The reduced total energies of Q114L and Q114M signified enhanced stability of both lipases over the wildtype T1,
R. Abdul Wahab et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 282–288
285
Table 1 Summary of the hydrophobicity of oxyanion-114 and the structural profile of lipases Q114L, Q11M and T1 according to FoldX and Voronoia 1.0, respectively. Lipase
Q114L Q114M T1 lipase (Q114)
FoldX
Voronoia 1.0
Goldman–Engelman– Steitz hydropathy scale
Total energy (kcal/mol)
Energy difference (kcal/mol)
Total cavities
Protein compactness (van der Waals/total volume)
2.8 (hydrophobic) 3.4 (hydrophobic) 1.0 (hydrophilic)
−146.71 −147.75 −143.34
−3.37 −4.41 –
75 76 73
0.704 0.704 0.706
suggesting each lipase has adapted a new folded structure that improved intermolecular interactions and was energetically more favorable. Therefore, the decrease in total energy of Q114L and Q114M can be attributed to the replacement of the hydrophilic Gln (1.0) with the more hydrophobic Leu (2.8) and Met (3.4) [6] according to the Goldman–Engelman–Steitz hydropathy scale [19], respectively. This is possibly due to creation of new hydrophobic and van der Waals interactions between the side chain of Leu and Met with adjacent hydrophobic amino acids to result in a more rigid interior and stabilizing the protein packing. Eijsink et al. indicated that the packing within the hydrophobic core of a protein can be adjusted to accommodate a great variety of mutations as long as hydrophobicity is preserved [1]. In this context it is pertinent to point out that the larger decline in total energy of Q114M (−4.47 kcal/mol) as compared to Q114L (−3.37 kcal/mol) is consistent with the substitution the more hydrophobic Met (3.4) as opposed to Leu (2.8). In this study, results of FoldX afforded significant clues on the structural packing Q114L and Q114M lipase following substitution with Leu and Met, indicating a more efficiently packed hydrophobic core of both lipases than in T1 (Table 1). Apart from the estimation of total energy of the lipase, the quality of the protein packing can also be inferred from total number of cavity and compactness/density. The tabulated total number of cavity and compactness/density of Q114L, Q114M and wildtype T1 calculated by Voronoia 1.0 is illustrated in Table 1. Interestingly, the results indicated that substitution with Leu and Met increased the number of total cavity and reduced the compactness in Q114L and Q114M, respectively. The compactness of Q114L (0.704 vdW/total volume) and Q114M (0.704 vdW/total volume) were marginally lower than the wild type T1 (0.706 vdW/total volume), but remarkably, the latter had developed three new cavities as compared to the former which formed only two new ones. In is vital to point out that, the increased number of cavities and reduced compactness are features associated with less desirable enzyme structure. This can be attributed to the less dense enzyme core that is prone to premature unfolding when subjected to various denaturing conditions i.e., extreme temperature and pH. The higher number of total cavities and lower protein compactness in Q114L and Q114M confirmed that the three-dimensional structures of both lipase proteins were different to that of T1 but were less stable. Fig. 1b depicts the location of the cavities in the immediate vicinity of the substrate binding pocket of the wildtype T1. In this perspective, features such as a lower total cavity number and a more compact structure are regarded as advantageous, inferring less susceptibility of the mutant lipases towards denaturation. These structural changes are the result of a more efficient packing due to increased hydrophobic interaction, hence, the enzyme molecule is less likely to prematurely unwind when the reaction temperature is elevated or when subjected to extreme pH. Hence, the resultant mutant lipase is anticipated to be more thermostable. Creighton indicated that a compromise between close packing and conformational strain is often required as energetically unfavorable perturbations of protein chemistry will most certainly occur following a mutation, particularly when the mutation is within the tight core of a folded protein [3]. In view of the contradictive results acquired using FoldX and Voronoia 1.0, empirical
Table 2 The parametric values that gave maximal yield of menthyl butyrate for reactions catalyzed by lipases wildtype T1, Q114L and Q114M. Reaction condition
Wildtype T1
Q114L
Q114M
Incubation time (h) Temperature (◦ C) Amount of enzyme (mg/mL) Molar ratio (butyric anhydride/menthol) Agitation speed (rpm)
16 70 3 2:1 200
16 80 3 2:1 300
16 70 3 2:1 300
assessment merits consideration using a solvent-free esterification of menthol with butyric anhydride as a model reaction. The empirical approach was necessary to elucidate the role the hydrophobic amino acids Leu and Met on the catalytic efficacy of the wildtype T1, as well as to prove, once and for all, which residue to confer improved/diminished activity in T1. 3.3. Effect of reaction parameters on the lipase-catalyzed esterification of menthol with butyric anhydride 3.3.1. Effect of incubation time Time course study is a good indicator of enzyme performance as well as product yield. The effect of incubation time was monitored up to 30 h with sampling intervals of 0, 2, 4, 8, 16, 16, 24 and 30 h, respectively. Constant parameters utilized were temperature (60 ◦ C), enzyme amount (2 mg/mL), substrate molar ratio butyric anhydride/menthol (1:1) and stirring speed (200 rpm). Prior to the investigation, the lipase activity of Q114L, Q114 M and wild type T1 were determined as 126.3, 82.5 and 103 U/mg, respectively. Fig. 2a represents the reaction time profile for the esterification of butyric anhydride and menthol under constant parameters; temperature (60 ◦ C), enzyme amount (2 mg/mL), molar ratio butyric anhydride/menthol (1:1) and stirred at 200 rpm catalyzed by lipases Q114L and Q114M and T1 monitored over for a 30 h period. It can be seen that the conversion of menthyl butyrate increased with the increase in reaction time up to 16 h and declined thereafter. Remarkably, the Q114L was found to show improved conversion of menthyl butyrate, reaching the highest percentage conversion (92%) as compared to the wildtype T1 (84.6%) and Q114M lipase (80.3%) (Table 2). The higher percentage of menthyl butyrate in reactions catalyzed by Q114L can be attributed to an improved protein packing as a result of the smaller Leu side chain [20], implying that the oxyanion is easily adapted to substitution with Leu to result in the reduced total energy of Q114L. The methylene side of Leu in Q114L possibly promotes formation of hydrophobic interaction [21] with adjacent amino acids and consequently, reinforces protein rigidity within the catalytic pocket. Therefore, the active site of Q114L is made less flexible and can better retain its catalytically active shape, hence, less susceptible to denaturation under long incubation time, particularly under elevated reaction temperature. Conversely, the low percentage conversion of menthyl butyrate in reactions catalyzed by Q114M infers that Met could not be well accommodated within the Q114 site. The observation is in accordance with results obtained in Voronoia 1.0 that presented the Q114M having a higher number of total cavities and low protein
286
R. Abdul Wahab et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 282–288
Fig. 2. The esterification of menthol and butyric anhydride catalyzed by lipases T1, Q114L and Q114M. (a) Effect of incubation time at constant parameters: temperature (60 ◦ C), enzyme amount (2 mg), molar ratio butyric anhydride/menthol (1:1) and agitation speed (150 rpm), and (b) effect of temperature at constant parameters: time = 8 h, enzyme amount = 2 mg/mL, molar ratio butyric anhydride/menthol = 1:1 and stirring speed = 200 rpm.
compactness as compared to the T1 and Q114L. These changes can be inferred as the consequence of inefficient protein folding that resulted from protein perturbations, causing the inner core of Q114M to expand, hence, the observed reduction in compactness of Q114M. According to Leow et al., cavities are energetically unfavorable architecture in the structure of folded proteins as the surrounding amino acids contain fewer van der Waals interactions that could destabilize the protein [22]. This statement is evidently true for Q114M as the lipase consistently recorded the lowest percentage conversions of menthyl butyrate throughout the course of the reaction. The observation suggests that changes within the binding pocket following mutation to Q114 may have been more extensive that it was thought to be. It was highly probable that replacement of Q114 with Met had altered the binding affinity of Q114M to the substrate, thereby, adversely affecting catalytic efficiency and slower production of menthyl butyrate. Meanwhile, the apparent decrease in the yield of menthyl butyrate after 16 h was perhaps due to the counterproductive liberation of water in the reaction media that hydrolyzes the ester product [23], a common occurrence in many esterification reactions. The production of surplus water is a natural phenomenon and an unwanted byproduct of a prolonged esterification reaction [24], as when in excess, the enzyme-catalyzed hydrolysis process is favored and the equilibrium shifts towards reformation of the substrates. 3.3.2. Effect of temperature Since temperature can have an effect on the equilibrium of reaction as well as the activity and stability of the enzymes [25], such factor was also evaluated in this present study. The effects of temperature on the Q114L, Q114M and T1 catalyzed esterification was investigated by increasing the reaction temperature from 40 to 90 ◦ C (Fig. 2b), under constant parameters; incubation time (8 h), enzyme amount (2 mg/mL), substrate molar ratio butyric anhydride/menthol (1:1) and stirring speed (200 rpm). Interestingly, results revealed high conversions of menthyl butyrate for lipases T1 (83.1%) and Q114M (±83.3%), as well as the Q114L lipase (80 ◦ C) were approximately at 70 ◦ C (T1 and Q114M) and 80 ◦ C (Q114L), respectively. Production of the ester was at the lowest for Q114L, Q114M and T1 at 40 ◦ C, yielding only 28, 35.1 and 38.5% of menthyl butyrate, respectively (Table 2). The high yield of menthyl butyrate achieved by Q114L at a considerably higher temperature (80 ◦ C) over lipases Q114M and T1 suggests that substitution with Leu has positively improved the activity and stability of Q114L under elevated temperature. The shift in the optimum temperature may be due to the hydrophobicity of Leu (−2.8) that altered the molecular interactions between amino acids close to the mutation point. In fact, Leu has been described as an acceptable substitute to the buried Gln due to its
similarity in terms of hydrophobicity [20,21] with other hydrophobic amino acids that predominates the protein core of T1. The increased interactions between the adjacent amino acids and new methylene side chain of Leu114 can effectively facilitate local structure folding [26] and reduce susceptibility of Q114L to high temperature deactivation. However, despite the presence of the more hydrophobic Met, the optimum temperature of Q114M remained unchanged. This may largely be due to the formation of three additional cavities that counteracts the benefit of increased molecular interactions between adjacent amino acids in the vicinity of the mutation. Alternatively, the overall decline in the yield of menthyl butyrate at conditions exceeding the optimal temperature of the lipases is illustrative of increased deactivation rate, signifying partial lipase inactivation, usually associated with excessive protein unfolding that causes enzymes to partially lose their catalytically competent form [27,28]. Alternatively, the low yield of menthyl butyrate at low reaction temperatures (40–50 ◦ C) is attributable to the fact that the protein structures of Q114M and Q114L, also variants of the thermophilic T1, were still rigid and have yet to achieve their catalytically active shape [29]. A compromise between high reaction temperature and solubility of reactants is often necessary to attain efficient biocatalysis, as at the right temperature, solubility of reactants is increased and catalytic activity of thermophiles is at their optimum, thereby accelerating the conversion of products [30]. 3.3.3. Effect of enzyme amount Fig. 3a details the reaction profile for the Q114L, Q114M and wildtype T1 in catalyzing the conversion of menthyl butyrate following the use of increasing enzyme loading (1–5 mg/mL) under fixed parameters; incubation time (8 h), temperature (60 ◦ C), substrate molar ratio butyric anhydride/menthol (1:1) and stirring speed (200 rpm). It was observed that the highest conversions of menthyl butyrate for the reactions by the Q114L, Q114M and wildtype T1 were all attained at 3 mg/mL, yielding up to 68.5, 42.1 and 68.9% of the ester, respectively (Table 2). The percent conversion of menthyl butyrate dropped in all reactions when the enzyme amount exceeded 3 mg/mL, indicating ineffective use of the biocatalysts at high enzyme loading. Beyond this concentration, substantial decrease in conversion of menthyl butyrate prevailed, usually associated with the high number of enzyme molecules in the reaction medium, apart from inefficient mixing that resulted in reduced effective enzyme-substrate collisions and rate of reaction. Review of the literature also revealed a similar trend whenever a high enzyme amount is used [29,31,32]. Thus, similar trend in Q114L, Q114M and T1 catalyzed synthesis may too prove common; suggesting further additions of the enzyme into the reaction is futile and will pointlessly boosts the production cost of menthyl
R. Abdul Wahab et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 282–288
287
Fig. 3. The esterification of menthol and butyric anhydride catalyzed by lipases wildtype T1, Q114L and Q114M. (a) Effect of enzyme amount at constant parameters: incubation time = 8 h, temperature = 60 ◦ C, molar ratio butyric anhydride/menthol = 1:1 and stirring speed = 200 rpm, (b) effect of substrate molar ratio of butyric anhydride/menthol at constant parameters: incubation time = 8 h, temperature = 60 ◦ C, enzyme amount = 2 mg/mL and stirring speed = 200rpm.
butyrate. Correspondingly, the low yield is also associated with the phenomenon of agglomeration whereby lipase molecules cluster and aggregate [33] to form inhomogeneous enzyme distribution which impedes accessibility of substrates to the enzymes, hence, the decrease in reaction yields. 3.3.4. Effect of substrate molar ratio Considering that molar ratio of anhydride:menthol can influence the rates of reactions [34] as well as the activity of the lipases [35], such factor too was examined in this study. The influence of substrate molar ratio in the production of menthyl butyrate catalyzed by lipases Q114L, Q114M and T1 was determined at ratios of butyric anhydride:menthol that ranged between 1:1–4:1 (Fig. 3b). Remarkably, substitution of the hydrophilic oxyanion Q114 with Leu and Met did not affect the optimum substrate molar ratio of substrates but effectively improve activity of Q114L as compared to T1 and Q114M. The highest conversion of menthyl butyrate for Q114L, Q114M and wildtype T1 were achieved at anhydride:menthol molar ratio of 2:1, yielding up to a 71, 52 and 64% of the ester, correspondingly (Table 2). Percent conversion of menthyl butyrate was the lowest for Q114L, Q114M and wildtype T1 at an anhydride/menthol molar ratio of 4:1, affording only 51, 37 and 55%, respectively. The noteworthy higher percentage of menthyl butyrate in Q114L-catalyzed reactions substantiates the enhanced catalytic activity of Q114L as compared to the Q114M and wildtype T1. The reduction in yield of menthyl butyrate at molar ratio exceeding 2:1 in all reactions is due to the inhibitory effect of excessive butyric anhydride present in the reaction medium or by formation of butyric acid [36]. A rather low yield of menthyl butyrate was also recorded at equimolar of anhydride:alcohol, is usually due to the increased dehydrating effects of the menthol molecules on the structures [37] whereby in most cases, the reaction substrates are inherently disruptive towards the structurally sensitive active form of the lipases, reducing the rate of reaction and yield of ester. Apart from the dehydrating effects, the excess menthol molecules also compete for the same active site [34]. 3.3.5. Effect of stirring speed The rate of stirring is important in many enzyme catalyzed processes as proper integration of reaction components is needed to obtain a homogeneous mixture of substrates and good contact between the enzyme molecules and substrates. The reaction profile based on increasing stirring rate speed (100–400 rpm) is detailed in Fig. 4. The results indicated that the percentage conversion of menthyl butyrate in reactions catalyzed by Q114L, Q114M and T1 positively correlated with the increase in the stirring speed, with lipases Q114L and Q114M attaining the highest conversion (72.2
Fig. 4. The esterification of menthol and butyric anhydride catalyzed by lipases wildtype T1, Q114L and Q114M. Effect of agitation speed at constant parameters: incubation time = 16 h, temperature = 60 ◦ C, enzyme amount = 2 mg/mL and molar ratio butyric anhydride/menthol = 1:1.
and 63.1%, respectively) of menthyl butyrate at 300 rpm, and the wildtype T1 (65.3%) at a lower stirring speed of 200 rpm. In a solvent-free biotransformation process, factors such as mixing and mass transfer require due consideration in the absence of solvent as the suspending medium. Mass and heat transfer play an important role under such condition as these factors affect the substrate availability as well as removal of inhibitory cellular metabolites from the biocatalyst [38]. The higher percentage of menthyl butyrate at increased stirring speed is attributable to a more efficient transfer of mass and heat throughout the reaction medium [38]. This promotes effective enzyme-substrate collisions that accelerate reaction rates [30] of the Q114L, Q114M and T1 lipases, and reduces viscosity of the substrates to result in higher reaction yields. 4. Conclusions This present study revealed that substituting the hydrophilic oxyanion Gln in the wildtype T1 with Leu (Q114L lipase) would considerably improve the catalytic activity of this lipase enzyme that can be attributable to lower total molecular energy predicted by FoldX. In contrast, such improved in the catalytic activity was not observed when the hydrophilic oxyanion Gln in the wildtype T1 was substituted with Met (Q114M lipase). In addition, increased in
288
R. Abdul Wahab et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 282–288
localized hydrophobicity may potentially enhance the hydrophobicity as well as van der Waals interactions, allowing the catalytic pocket to retain its active shape for longer duration. Contrasting to Leu, substituting Gln with Met at the oxyanion-114 would create additional cavities in the molecule that subsequently alter catalytic pocket of Q114M lipase due to the larger side chain of Met when compared with Leu; a problematic approach for enhancing the catalytic activity of the enzyme. Hence, it can be construed that such substitution as Leu into the wildtype T1 at oxyanion-114 had conferred favorable structural changes that rendered better catalytic activity; an approach that may prove suitable for improving the catalytic activity of other bacterial lipases within the I.5 family. Acknowledgements The authors would like to thank Dr Naji Arafat for proof reading this manuscript. This project was funded by a grant from Ministry of Science and Technology and Innovation, grant no. 5487729 (Green Chemistry), Malaysia. References [1] V.G.H. Eijsink, B.W. Djikstra, G. Vriend, J.R. van der Zee, O.R. Veltman, B. van der Vinne, G. van den Burg, S. Kempe, G. Venema, Protein Eng. 5 (1992) 421–426. [2] C.N. Pace, B.A. Shirley, M. McNutt, K. Gajiwala, FASEB J. 10 (1996) 75–83. [3] T.E. Creighton, Curr. Opin. Struc. Biol. 1 (1991) 5–6. [4] M. Eilers, S.C. Shekar, T. Shieh, S.O. Smith, P.J. Fleming, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 5796–5801. [5] T. Ohmura, T. Ueda, K. Ootsuka, M. Saito, T. Imoto, Protein Sci. 10 (2001) 313–320. [6] J. Liang, K.A. Dill, Biophys. J. 81 (2001) 751–766. [7] K. Yutani, K. Ogasahara, T. Tsujuta, Y. Sugino, Proc. Natl. Acad. Sci. U. S. A. 84 (1987) 4441–4444. [8] S. Sonavane, P. Chakrabarti, PLoS Comput. Biol. 4 (2008) e1000188, http://dx. doi.org/10.1371/journal.pcbi.1000188. [9] R.A. Wahab, M. Basri, M.B. Abdul Rahman, R.N.Z. Raja Abdul Rahman, A.B. Salleh, T.C. Leow, Int. J. Mol. Sci. 13 (2012) 11666–11680. [10] K.L. Morley, R.J. Kazlauskas, Trends Biotechnol. 23 (2005) 231–237. [11] A. Nobili, M.G. Gall, I.V. Pavlidis, M.L. Thompson, M. Schmidt, U.T. Bornscheur, FEBS J. (2013), http://dx.doi.org/10.1111/febs.12137. [12] J. Gu, L. Ye, F.L.X. Guo, W. Lu, H. Yu, Appl. Microbiol. Biotechnol. (2014), http:// dx.doi.org/10.1007/s00253-014-5995-x.
[13] V. Stepankova, T. Bidmanova, T. Kudelakova, Z. Prokop, R. Chaloupkova, J. Damborsky, ACS Catal. 3 (2013) 2823–2836. [14] T.C. Leow, R.N.Z. Rahman, M. Basri, A.B. Salleh, Biosci. Biotechnol. Biochem. 68 (2004) 96–203. [15] C. Carrasco-Lopez, C. Godoy, B. de las Rivas, G. Fernandez-Lafuente, J.M. Palomo, J.M. Guisan, R. Fernandez-Lafuente, M. Martinez-Ripoll, J.A. Hermoso, J. Biol. Chem. 284 (2009) 4365–4372. [16] http://bioinformatics.charite.de/voronoia. [17] A. Nayeem, S.J. Chiang, S.W. Liu, Y. Sun, L. You, J. Basch, Protein Eng. Des. Sel. 22 (2009) 257–266. [18] W.C. Sandberg, T.C. Terwilliger, Proc. Natl. Acad. Sci. U. S. A. 74 (1991) 643–657. [19] D.M. Engelman, T.A. Steitz, A. Goldman, Ann. Rev. Biophys. Biophys. Chem. 15 (1986) 321–353. [20] R.A. Wahab, M.B. Abdul Rahman, U.H. Zaidan, M. Basri, R.N.Z. Raja Abdul Rahman, A.B. Salleh, T.C. Leow, Adv. Biosci. Bioeng. 3 (2012) 158–167. [21] K. Ishikawa, H. Nakamura, K. Morikawa, S. Kanaya, Biochemistry 32 (1993) 6171–6178. [22] T.C. Leow, R.N.Z. Raja Abdul Rahman, M. Basri, A.B. Salleh, Extreme 11 (2005) 527–535. [23] A.L. Ong, A.H. Kamaruddin, W.S. Long, S.T. Lim, R. Kumari, Enzyme Microb. Technol. 39 (2006) 924–929. [24] Q.X. Song, D.Z. Wei, J. Mol. Catal. B: Enzym. 18 (2002) 261–266. [25] D. Zhao, E. Xun, J. Wang, R. Wang, X. We, L. Wang, Z. Wang, Biotechnol. Bioprocess Eng. 6 (2011) 638–644. [26] E. Eyal, R. Najmanovich, M. Edelman, V. Sobolev, Protein Struct. Funct. Bioinform. 50 (2003) 272–282. [27] M.L. Foresti, M. Galle, M.L. Ferreira, L.E. Briand, J. Chem. Technol. Biotechnol. 84 (2009) 1461–1473. [28] S. Sun, L. Shan, Y. Liu, Q. Jin, Y. Son, X. Wang, J. Mol. Catal. B: Enzym. 57 (2009) 104–108. [29] R.A. Wahab, M. Basri, R.N.Z. Abdul Rahman, M.B. Abdul Rahman, A.B. Ariff, A.B. Salleh, N. Chaibakhsh, T.C. Leow, Biotechnol. Biotechnol. Equip. 28 (2014) 1065–1072. [30] M.B. Abdul Rahman, N. Chaibakhsh, M. Basri, R.N.Z. Abdul Rahman, A.B. Salleh, S. Md Radzi, J. Chem. Technol. Biotechnol. 83 (2008) 1534–1540. [31] C. Dianoczki, R. Recseg, K. Kovari, P. Poppe Laszlo, J. Mol. Catal. B: Enzym. 45 (2007) 45–49. [32] K.M.W. Syamsul, M.R. Salina, S.O. Siti, M.N. Hanina, M.A.R. Basyaruddin, K. Jusoff, World Appl. Sci. J. 11 (2010) 401–407. [33] G. Trubiano, D. Borio, A. Errazu, Enzyme Microb. Technol. 40 (2007) 716–722. [34] P. Pires-Cabral, M.M.R. da, S. Ferreira-Dias, Biochem. Eng. J. 43 (2009) 327–332. [35] D. Li, W. Xiaojing, N. Kaili, W. Fang, L. Junfeng, W. Pu, T. Tianwei, Chin. J. Chem. Eng. 19 (2011) 978–982. [36] E.T. Liaw, K.J. Liu, Bioresour. Technol. 101 (2010) 3320–3324. [37] M.L. Foresti, A. Errazu, M.L. Ferreira, Biochem. Eng. J. 25 (2005) 69–77. [38] V.S. Meena, U.C. Banerjee, Ind. J. Biotechnol. 10 (2011) 452–457.