Rational Engineering of a Cold-Adapted α-Amylase from the Antarctic Ciliate Euplotes focardii for Simultaneous Improvement of Thermostability and Catalytic Activity

ABSTRACT The α-amylases are endo-acting enzymes that hydrolyze starch by randomly cleaving the 1,4-α-d-glucosidic linkages between the adjacent glucose units in a linear amylose chain. They have significant advantages in a wide range of applications, particularly in the food industry. The eukaryotic α-amylase isolated from the Antarctic ciliated protozoon Euplotes focardii (EfAmy) is an alkaline enzyme, different from most of the α-amylases characterized so far. Furthermore, EfAmy has the characteristics of a psychrophilic α-amylase, such as the highest hydrolytic activity at a low temperature and high thermolability, which is the major drawback of cold-active enzymes in industrial applications. In this work, we applied site-directed mutagenesis combined with rational design to generate a cold-active EfAmy with improved thermostability and catalytic efficiency at low temperatures. We engineered two EfAmy mutants. In one mutant, we introduced Pro residues on the A and B domains in surface loops. In the second mutant, we changed Val residues to Thr close to the catalytic site. The aim of these substitutions was to rigidify the molecular structure of the enzyme. Furthermore, we also analyzed mutants containing these combined substitutions. Biochemical enzymatic assays of engineered versions of EfAmy revealed that the combination of mutations at the surface loops increased the thermostability and catalytic efficiency of the enzyme. The possible mechanisms responsible for the changes in the biochemical properties are discussed by analyzing the three-dimensional structural model. IMPORTANCE Cold-adapted enzymes have high specific activity at low and moderate temperatures, a property that can be extremely useful in various applications as it implies a reduction in energy consumption during the catalyzed reaction. However, the concurrent high thermolability of cold-adapted enzymes often limits their applications in industrial processes. The α-amylase from the psychrophilic Antarctic ciliate Euplotes focardii (named EfAmy) is a cold-adapted enzyme with optimal catalytic activity in an alkaline environment. These unique features distinguish it from most α-amylases characterized so far. In this work, we engineered a novel EfAmy with improved thermostability, substrate binding affinity, and catalytic efficiency to various extents, without impacting its pH preference. These characteristics can be considered important properties for use in the food, detergent, and textile industries and in other industrial applications. The enzyme engineering strategy developed in this study may also provide useful knowledge for future optimization of molecules to be used in particular industrial applications.

T he ␣-amylases (Enzyme Commission [EC] no. 3.2.1.1) are endo-acting enzymes that hydrolyze starch by randomly cleaving the 1,4-␣-D-glucosidic linkages between the adjacent glucose units in a linear amylose chain (1,2). On the basis of sequence similarity, ␣-amylases have been classified into the glycoside hydrolase (GH) families (3), mostly incorporated into the GH13 family in the carbohydrate-active enzymes (CAZy) database (4). However, with the rapidly increasing number of ␣-amylases recently discovered, they have also been identified as representatives of other GH families (3). The evolutionary relatedness among members of the GH13 family has been recently extensively revised, and the family is presently subdivided into 42 subfamilies on the basis of similarities in their sequences (5). The tertiary structure of each of these enzymes is characterized by a (␤/␣) 8 barrel containing a conserved catalytic triad (Asp, Glu, and Asp) that forms the active site (6).
␣-Amylases have been utilized in a wide range of industrial processes, such as in food, detergent, textile, and paper industries (7,8). They represent approximately 30% of the world's enzyme market (2). ␣-Amylases can be obtained from several sources, including plants, animals, and microorganisms. Microbial ␣-amylases are generally attractive for biotechnological and industrial applications (1,9), particularly those produced by extremophiles, as they can withstand harsh conditions (10)(11)(12)(13)(14)(15). In general, cold-adapted enzymes produced by psychrophiles possess high biotechnological value and provide economic benefit, being more productive at low temperatures than mesophilic or thermophilic homologs. This implies energy savings in industrial processes (16,17). However, although a few number of cold-adapted ␣-amylases have been studied to understand the molecular adaptation of cold-adapted enzymes (18)(19)(20)(21)(22), they are very rarely developed for immediate use in industrial applications (23). As a matter of fact, there has been a high demand for the discovery of a novel cold-adapted ␣-amylase for use in various industrial processes (24).
Cold-adapted enzymes possess a range of structural features that confer higher structural flexibility than thermostable homologs, which usually translates to high specific activity at low temperatures (25). However, an increased flexibility may also represent a two-edged sword for psychrophilic organisms, since it also increases the likelihood that the proteins may undergo denaturation in response to small changes in temperature (26). Therefore, the high thermolability of cold-adapted enzymes has been a major drawback for their use in industrial applications (17,27). One way to solve this problem is to modify existing cold-adapted enzymes by site-directed mutagenesis combined with a rational design approach (28), which generally is based on structureguided consensus sequence alignments with sequences that have moderate or high amino acid identity to reduce the number of possible target residues to be mutated. Previously, rational design principles were successfully utilized to enhance the stability and/or catalytic efficiency of enzymes such as maltogenic amylase and patatin-like phospholipase (29)(30)(31). These changes reduced the molecular flexibility of the polypeptide by (i) introducing hydrophobic residues in the protein core, (ii) introducing disulfide and salt bridges to increase electrostatic interactions in the polypeptide, (iii) stabilizing ␣-helix dipoles, (iv) introducing Pro residues, and (v) loop shortening to decrease loop entropy (28,32).
Recently, the cold-adapted ␣-amylase from the Antarctic ciliate Euplotes focardii, named EfAmy, was heterologously expressed in Escherichia coli and biochemically characterized (33). The ciliated protozoon E. focardii shows a strictly psychrophilic phenotype (34,35). The amylolytic activity of EfAmy was compared with that of the homologous enzyme, EcAmy, from the mesophilic congeneric species E. crassus (33). The results showed that EfAmy represents a classical psychrophilic enzyme with high hydrolytic activity at low temperatures (5 to 25°C) and high thermolability. Furthermore, EcAmy displayed a 2-fold increased thermostability at 50°C compared with that of EfAmy. In this study, we applied site-directed mutagenesis to generate a cold-active EfAmy with improved thermostability and catalytic efficiency at low temperatures. We applied a rational design approach to identify residues that contribute to the structural stability and catalytic efficiency of the cold-active EfAmy ␣-amylase. Biochemical assays of engineered versions of EfAmy revealed that the combination of mutations at the surface loops and catalytic core resulted in significant increases in thermostability and catalytic efficiency of the enzyme. The novel ␣-amylases generated in our study could have important applications in the detergent and food and beverage industries, where the cold-adapted enzymes have already been widely applied and have even led to the evolution of the conventional processes (36). Furthermore, the success of our rational design strategy suggests that it may be universally applied to other enzymes.

RESULTS AND DISCUSSION
EfAmy sequence analysis. In a previous study (33), we biochemically characterized two homologous ␣-amylases from two closely related Euplotes species, the psychrophilic E. focardii (named EfAmy) and the mesophilic E. crassus (named EcAmy). We showed that EfAmy and EcAmy share a relatively high sequence similarity (68% identity, 82% similarity). To identify their evolutionary relations with other members of the GH13 subfamilies, EfAmy and EcAmy were aligned with ␣-amylases from representative taxa of Bacteria, Archaea, and Eukarya. The alignment was then converted into a phylogenetic tree (Fig. 1). The tree clearly shows that EfAmy (named in the tree as Euplotes focardii) and EcAmy (named Euplotes crassus) together with homologous sequences from other ciliates, such as Euplotes harpa, Oxytricha trifallax, Tetrahymena thermophila, and Paramecium tetraurelia, cluster within the GH13_1 subfamily clade that is represented by several well-studied fungus and yeast ␣-amylases, including Aspergillus niger acid ␣-amylase (37), Aspergillus oryzae TAKA ␣-amylase (38), and Saccharomycopsis fibuligera ␣-amylase (39).
Rational design of EfAmy mutations. Previous findings from the biochemical characterization of EfAmy and EcAmy indicate significant differences in the catalytic properties of the two enzymes, such as in their thermostability and temperature-dependent catalytic efficiency (33). In this study, we combined rational design and site-directed mutagenesis approaches to modify EfAmy to develop novel ␣-amylases that are both thermostable and highly active at low temperatures. As the first step of our strategy, we performed a detailed sequence analysis of the cold-adapted EfAmy with the homologous mesophilic EcAmy (Fig. 2). This included a multiple-sequence alignment (MSA) and protein FIG 2 Euplotes ␣-amylase sequence analysis. Alignment of the predicted amino acid sequences of EfAmy and EcAmy with the homologous A. oryzae TAKA ␣-amylase A (TAA) that was included due to available molecular structures used in the homology model shown in Fig. 3. Conserved residues are lightly shaded. The ␣-amylase catalytic residues of the Asp, Glu, and Asp triad are marked in yellow. The five mutation sites are labeled with red arrows. The seven conserved ␣-amylase sequence regions are boxed. The amino acids of putative signal peptides are in italics. Secondary structure elements are presented below the sequences, which correspond to the TAA ␣/␤ barrel, as described by Machius et al. (82). TAA residues involved in Ca 2ϩ (*) and Cl Ϫ (•) binding (82) are indicated below the sequences. Sequence regions belonging to domains A (-· -), B (---), and C (-· · -) are labeled above the sequences. secondary structure analysis. In the MSA, the TAKA ␣-amylase A (TAA) from A. oryzae was also included, since the three-dimensional protein structure of this enzyme was used as the template for homology modeling of Euplotes ␣-amylases in the subsequent analysis. The overall low conservation of the protein sequences among GH13 family enzymes prevented the inclusion of additional ␣-amylases in the alignment (40). The purpose of this analysis was to compare the sequences of two homologous enzymes from closely related species to identify the residues possibly responsible for their distinguished catalytic properties revealed by the previous biochemical characterization (33). In general terms, the conservation between EfAmy and EcAmy was not uniform along the entire sequence. The highest amino acid conservation level was found at the sequences corresponding to the A, B, and C domains (Fig. 2), which showed 71, 81, and 53% identity, respectively. Subsequently, for the selection of the mutation sites, we focused our attention on the A and B domains, since changes in the C domain appear to have no effects on ␣-amylase thermostability and activity (41).
Through an intensive analysis of the sequences, we detected three substitutions in the A and B domains of EfAmy and EcAmy, i.e., E166P, S185P, and T350P (where each number refers to the sequence position that is involved in the substitution and the letters that precede and follow the number stand for amino acids present in E. focardii and E. crassus, respectively). It is unsurprising that a higher number of Pro residues is observed in the mesophilic ␣-amylase than in the psychrophilic EfAmy. In fact, Pro avoidance has been generally assumed as the molecular mechanism of cold adaptation of the psychrophilic enzymes, since Pro residues in loops are supposed to infer increased rigidity to the polypeptide and a smaller number of Pro residues has been noted in loops connecting secondary structures in several cold-adapted enzymes investigated so far (42). Very intriguingly, we observed through structural modeling analysis that the three Pro substitutions between EfAmy and EcAmy were all located in the surface loops ( Fig. 3) (i.e., two of them were found in the loop of the B domain and another was found in the loop of the A domain), which have been reported to be of major importance to the overall stability of ␣-amylases (7). Therefore, we primarily chose to mutate the three residues in the loops of EfAmy (Glu166, Ser185, and Thr350) to Pro to achieve an increased thermostability of the cold-adapted enzyme.
Through sequence alignment and modeling analysis, we also observed that Val212 and Val232 in EfAmy were replaced by Thr in EcAmy. As shown in Fig. 3, Val212 is located in close proximity to the catalytic residues, Asp210 (at a distance of approximately 4.1 Å) and Glu234 (at a distance of approximately 3.6 Å), while Val232 is located in proximity to residue Arg208 (at a distance of approximately 3.5 Å), which according to Brzozowski and Davies (43) plays a role in the stabilization of ligand (e.g., with ligand ABC as shown in Fig. 3A). Val212 and Val232 interact with the catalytic residues by nonpolar interactions as shown in Fig. 3A. Previous studies have indicated that the modification of amino acids surrounding the catalytic sites of enzymes will significantly affect the catalytic activity and stability (31,44). In our case, the replacement of these residues with polar amino acids, such as Thr (as occurs in EcAmy), most likely will produce the establishment of H bonds between the two Thr residues and Asp and Arg residues of the catalytic site as predicted and shown in Fig. 3B. The predicted shorter length and, consequently, the higher strength of these hydrogen bonds with respect to the nonpolar interactions (which, moreover, become weaker under cold temperature conditions [45]) may confer higher stability versus the ligand and better catalytic activity of the enzyme. Therefore, Val212 and Val232, located around the active sites of EfAmy, were chosen for mutagenesis to probe their effects on the catalytic efficiency and thermostability of EfAmy. In addition to the single mutations, variants containing these substitutions combined (V212T/V232T, E166P/S185P/T350P, and E166P/S185P/ T350P/V212T/V232T) were also generated and studied.
pH dependency of EfAmy and site-directed variants. The pH dependence of wild-type and mutant ␣-amylase activities were examined in various buffers at pH values that ranged from 5.0 to 11.0 under standard assay conditions. The results show that the optimal pH of all site-directed variants was 9, which was similar to that of the wild-type enzyme (see Fig. S2 in the supplemental material). While ␣-amylases from most bacteria and fungi have an optimal pH in the acidic to neutral range (1), EfAmy represents an alkaline enzyme similar to the previously characterized alkaline ␣-amylases from alkalophilic Bacillus spp. (46). This can be considered an important property for utilizing the enzyme in detergent and textile industries and in other industrial applications.
Effects of temperature on activity and stability of site-directed variants. The effect of temperature on wild-type and mutant ␣-amylase activities was determined at various temperatures ranging from 5 to 60°C (Fig. 4). Under the standard conditions used, the wild-type EfAmy showed the highest activity at approximately 25°C. Fifty percent maximal activity was observed at 5°C and approximately 10% residual activity remained at 45°C. All these characteristics indicate that the wild type behaves as a classic psychrophilic enzyme. To optimize the enzyme activity and stability at high temperatures and without a dramatic compromise of its activity at low temperatures for a broader application in industrial uses, we introduced mutations in both surface loops and in the catalytic domain of the EfAmy based on our structural analysis. As shown in Fig. 4A and B, the mutant enzymes with single mutations (E166P, S185P, T350P, V212T, and V232T) showed optimal activity at 25°C, with only slightly decreased activity at low temperatures (5 to 20°C) and increased activity at high temperatures (30 to 55°C). These data may indicate that single mutagenesis is not sufficient to drastically modify the catalytic parameters of the cold-adapted enzyme to convert it to a true mesophilic ␣-amylase. Similar phenomena were also observed with a number of other microbial ␣-amylases (47)(48)(49). When the mutations were combined, the generated variant enzymes (E166P/S185P/T350P and E166P/S185P/T350P/V212T/V232T) displayed a shift in optimal temperature from 25 to 30°C that is typical of the mesophilic counterpart, EcAmy (Fig. 4C). These results clearly demonstrate that the combination of mutations shows an effect of superimposing the activation that alters the behavior of the variant toward that of a mesophilic enzyme (50). As mentioned above, Pro has a more rigid conformation. Therefore, point mutations introducing extra Pro residues in the surface loops could thus increase the enzyme's surface rigidity. Recently, Brandsdal and coworkers disclosed the correlation between enzyme surface rigidity and temperature-dependent activity (51). Through computer simulation and mutation study, they determined that increasing the restraints in protein surface loops can lead to an unambiguous effect of turning a cold-adapted enzyme into a variant with mesophilic characteristics. On the other hand, our structural analysis indicates that the two mutations in the catalytic core region of the enzyme (V212T and V232T) may result in the formation of extra H bonds between the substituted residues and the catalytic sites (Fig. 3). It is known that the hydrogen bonding network around the catalytic domain of ␣-amylase can be crucial for the enzyme to maintain its stability and catalytic efficiency at a high temperature (52,53), which is also supported by our results from this study.
To analyze the thermostability of the mutant ␣-amylases, we incubated the enzymes at 40 and 50°C for 2 to 20 min before measuring the residual activity at their optimal FIG 4 (A to C) Effect of temperature on the amylolytic activity of EfAmy, EcAmy, and mutants. For each enzyme, the total activity at the optimal temperature was set at 100%. This represented 1.27 U/ml for EfAmy, 1.86 U/ml for EcAmy, 1.31 U/ml for the E166P mutant, 1.34 U/ml for S185P, 1.31 U/ml for T350P, 1.53 U/ml for V212T, 1.51 U/ml for V232T, 1.91 U/ml for V212T/V232T, 1.32 U/ml for E166P/S185P/T350P, and 2.03 U/ml for E166P/S185P/T350P/V212T/V232T. temperatures (Fig. 5). As shown in Table 1, all single mutant enzymes (E166P, S185P, T350P, V212T, and V232T) displayed increased stability at the same temperatures with respect to the wild-type EfAmy, which showed 4.1-and 1.8-min half-lives at 40 and 50°C, respectively. In addition to single mutations, our study also examined the possible synergistic effects of combining mutations. Notably, the E166P/S185P/T350P/V212T/ V232T variant, combining all the single mutations, was the most stable; half-life was more than 1.8-fold that of the wild-type enzyme at 50°C. Although single-site mutagenesis of the rationally selected residues was not able to raise the thermostability of the cold-adapted EfAmy to the same level as EcAmy, which represents the mesophilic counterpart, the half-lives are comparable when the combined mutations are taken into account (Table 1). These data reveal the effect of beneficial amino acid mutations that are synergistic or additive for the thermostability of the enzyme, which is consistent with the findings from other studies (54, 55). The generated variant   (30,(57)(58)(59). In this study, by adopting sequence alignment and structural modeling analyses, we engineered EfAmy with substituted Pro residues in the surface loops that promoted new hydrogen bonds in the catalytic domain. This most likely confers a reduced flexibility of the loops and catalytic core of EfAmy and thus an increased thermostability of the enzyme. Kinetic parameters. Kinetic studies were performed on the wild-type and mutated enzymes at 5, 25, and 35°C using starch as the substrate. As indicated by the results presented in Table 2, in comparison with the wild-type EfAmy, the E166P, S185P, and T350P mutants showed slightly decreased turnover rates (k cat ) and increased substrate binding affinities (decreased K m ) at all temperatures. However, the overall catalytic efficiency was increased for these single-site mutants. This effect was greater when the three mutations were combined (E166P/S185P/T350P), with an increased k cat /K m value of 56% at 35°C. On the other hand, although slight decreases in k cat values were observed at 5 and 25°C when the single-site variants with amino acid substitutions in the catalytic core of the enzyme (V212T or V232T) were tested, significant increases of catalytic efficiency were detected at all temperatures measured. This can be explained by the dramatic decrease in K m values for both single-site mutants. Interestingly, when the engineered enzyme that carried the combination of all mutations (E166P/S185P/ T350P/V212T/V232T) was analyzed, a significant increase of turnover rate was observed at a high temperature (35°C). Indeed, this enzyme represents the most efficient catalyst among all the enzymes tested in this study when the reactions were carried out at 35°C. It is worth noting that the engineered enzyme that carried the five mutations combined showed an even better catalytic efficiency than the mesophilic EcAmy (33). There is agreement that a trade-off between thermostability and catalytic activity has taken place during the natural evolution of enzymes to suit different temperature niches in the environment (60). This is supported by the fact that cold-adapted enzymes with high catalytic efficiency at a low temperature often show high thermolability at high temperatures due to the loss of native structures (61,62). However, thermostability and catalytic activity are not mutually exclusive in a cold-adapted enzyme. It has been reported that directed-evolution methods can confer enzymes with both high thermostability and high catalytic activity (63,64). D. Kern and coworkers have recently applied ancestral sequence reconstruction (ASR) approaches to create a "superenzyme" that displays high thermostability and catalytic activity at low temperatures (65). In this study, we further proved that rational design-based protein engineering can also achieve the same effect.
Conclusion. Protein engineering strategies are often used to optimize enzyme traits. Both rational design and directed evolution have been used to tailor enzyme properties (66,67). Rational design requires a thorough understanding of parental structures, consensus sequences, and interactions between amino acid residues to identify mutations that will lead to desired enzyme properties, especially the high level of thermostability and catalytic efficiency required for industrial applications (68). In this work, we applied a rational design method to engineer a novel cold-adapted ␣-amylase with improved thermostability, substrate binding affinity, and catalytic efficiency to various extents. In one variant (E166P/S185P/T350P), we introduced Pro residues in the A and B domains of surface loops. In the second variant (V212T/V232T), we facilitated extra molecular interactions in the catalytic core of the protein. The aim of these substitutions was to rigidify the molecular structure of the enzyme. Indeed, the E166P/S185P/T350P site-directed variant exhibited a strong increase in thermostability in comparison to that of the wild type, whereas the V212T/V232T site-directed mutation had a high impact on the catalytic efficiency compared with those of the enzymes with surface mutations. All our mutation sites were selected based on structural and consensus analyses of the cold-adapted enzyme and its mesophilic counterpart. Unlike the site-directed mutagenesis solely based on protein structural analysis that often shows adverse impacts on the kinetic parameters, the structureguided consensus approach assumes that nature has efficiently optimized the protein sequence space. Consequently, the targeting of a single mutational point of interest at positions which match the desired consensus cutoff has proven to be a successful approach for engineering enzymes with improved kinetic parameters (30, 69, 70). The combination of sequence analysis and homology-based model comparison appears to be a good approach for the rational design of mutations to improve the thermal stability and catalytic efficiency of this enzyme (71,72). Our study supports this multidisciplinary approach for engendering structural diversities in catalysis and for optimizing particular industrial applications. The E. focardii ␣-amylase that we have engineered is promising for industrial applications not only for its efficiency in the cold but also because it represents an alkaline enzyme, similar to the ␣-amylase from alkalophilic Bacillus spp. (11,46) and different from most of the ␣-amylases characterized so far. These characteristics can be considered important properties for use in detergent and textile industries and in other industrial applications.
Bacterial strains, plasmids, and culture conditions. E. coli DH5␣ was used as the host for cloning, whereas E. coli BL21(DE3)/pLysS harbored the wild-type and recombinant plasmids for gene expression. The pET22b(ϩ) vector (Novagen), containing the inducible T7 promoter, was used as the expression vector. The plasmid pET22b-EfAmy, containing the gene encoding EfAmy (33), was used for production of the wild-type EfAmy protein. Different E. coli strains harboring wild-type and mutated genes were routinely grown in Luria-Bertani (LB) medium at 37°C. When required, antibiotics and chromogenic substrates were added at the following concentrations: 100 g/ml ampicillin, 34 g/ml chloramphenicol, and 30 g/ml X-Gal.
Construction, expression, and purification of the mutant plasmids. The gene encoding the wild-type EfAmy was previously cloned and overexpressed in E. coli BL21(DE3)/pLysS (28), while the mutated genes encoding the mutated EfAmy were constructed by PCR-based site-directed mutagenesis (73). For each single mutation, PCRs were carried out using the plasmid pET-EfAmy as the template, Pfu DNA polymerase, and two complementary mutagenic primers for each site-specific mutation. For each combined mutation, PCR was performed using the former mutated plasmid as the template with the corresponding primers for each reaction. The sequences of the oligonucleotides are shown in Table S1 in the supplemental material. The final amplified products were verified by bidirectional DNA sequencing. As a result, PCR mutagenesis yielded eight expression plasmids: pET-EfAmy-E166P, pET-EfAmy-S185P, pET-EfAmy-T350P, pET-EfAmy-V212T, pET-EfAmy-V232T, pET-EfAmy-V212T/V232T, pET-EfAmy-E166P/ S185P/T350P, and pET-EfAmy-E166P/S185P/T350P/V212T/V232T, where the numbers indicate the position in the polypeptide of the mutated residue and the letters before and after the numbers indicate the original and the substituted amino acid, respectively.
E. coli BL21(DE3)/pLysS cells carrying wild-type and mutant EfAmy plasmids were grown overnight at 37°C in LB medium supplemented with 100 g/ml ampicillin and 34 g/ml chloramphenicol. The overnight cultures were diluted to an optical density at 600 nm (OD 600 ) of approximately 0.08 in 600 ml of LB medium supplemented with 100 g/ml ampicillin and 34 g/ml chloramphenicol using a 2-liter 60°C in 0.1 M Tris-HCl buffer (pH 9). Half-lives of thermal inactivation were determined for purified ␣-amylases by incubating the enzymes in 0.1 M Tris-HCl buffer (pH 9) at 40°C and 50°C for 0 to 20 min at regular time intervals. Initial and residual activities were measured under the standard assay conditions as described above. The first-order rate constant, K d , of irreversible thermal denaturation was obtained from the slope of the plots of ln (initial activity/residual activity) versus time, and the half-lives (t 1/2 ) were calculated as ln 2/K d .
Influence of mutations on kinetic parameters. The kinetic parameters K m and k cat of wild-type and mutant ␣-amylases were measured using soluble starch as the substrate at 5, 25, and 35°C. The initial velocities of substrate hydrolysis were monitored for a minimum of substrate concentration. The substrate concentrations used ranged from 0.5 to 8 g/liter. All kinetic data were analyzed by nonlinear regression using Origin 8.0. The standard errors for each parameter were estimated from the curve fitting. Assays were performed in duplicate, and results for kinetic data were the means from two independent experiments.
Sequence analysis, phylogenetic tree construction, and comparative modeling. Sequence similarity and analysis for conserved sequence regions (CSR) were performed using BLAST programs on the National Center for Biotechnology Information (NCBI) website (http://www.ncbi.nlm.nih.gov). Sequence alignment was performed using Clustal Omega (http://www.ebi.ac.uk/Tools/msa/clustalo/), and the related phylogenetic tree was calculated using the neighbor-joining method and displayed using the R package APE (version 3.5) (76). The comparative homology models of E. focardii and E. crassus ␣-amylases were obtained by the Modeler software (77) using TAKA ␣-amylase A (TAA) from Aspergillus oryzae (PDB no. 2TAA and 2GVY) (78,79) and the ␣-amylase from Malbranchea cinnamomea (PDB no. 3VM7) (80) as specific templates. The sequence identities shown by these three templates against EfAmy were 33.3%, 33.8% and 32.9%, respectively, and against EcAmy the identities were 37.4%, 37.6%, and 35.9%, respectively. It is relevant to note that each template had a coverage percentage higher than 95%. To visualize residues involved in ligand interaction, the EfAmy homology model was superimposed on the structure of TAA from A. oryzae in complex with ABC, an acarbose-derived hexasaccharide, as the ligand (PDB no. 7TAA) (43). The Deep View Swiss-PDB viewer software and PyMOL v1.5 were used to visualize and analyze the three-dimensional models.

SUPPLEMENTAL MATERIAL
Supplemental material for this article may be found at https://doi.org/10.1128/AEM .00449-17.