Enhancing Paenibacillus sp. Cold-Active Acetyl Xylan Esterase Activity through Semi-Rational Protein Engineering

Abstract

Interest in protein engineering for the enzymatic production of valuable products, such as pharmaceutical compounds and biofuels, is growing rapidly. The cold-active acetyl xylan esterase from Paenibacillus sp. (PbAcE) presents unusually broad substrate specificity. Here, we engineered a hydrophobic substrate-binding pocket to enable the accommodation of relatively large alcohol substrates, such as linalyl acetate and α-terpinyl acetate. To identify candidate residues for engineering, we performed covalent docking of substrates to the Ser185 active site using the HCovDock program. Functional hotspots were analyzed using HotSpot Wizard 3.1. Lys91, His93, and Tyr182 were selected for site-saturation mutagenesis (SSM). After generating the SSM mutant library, a qualitative colorimetric assay was conducted to identify positive mutants. Three, two, and five single mutants were selected for Lys91, His93, and Tyr182, respectively. The best single mutants were then sequentially combined to generate double and triple mutants. Single mutants exhibited a 10–30% increase in activity compared to that of wild-type PbAcE, while no significant synergistic improvements were observed in the double and triple mutants. The increase in activity against both linalyl acetate and α-terpinyl acetate was similar. Mutation did not affect the acetyl binding and catalysis. Further research on the acetyl binding pocket will provide insights into substrate specificity and aid in efficient biocatalyst development for industrial applications.

1. Introduction

The enzymatic synthesis of pharmaceutical compounds has garnered increasing attention in recent years, offering numerous advantages over traditional chemical synthesis methods [1]. This approach harnesses the catalytic capabilities of enzymes to facilitate specific chemical transformations for producing valuable drugs. Enzymes have high selectivity, enabling the synthesis of complex pharmaceutical molecules with precise stereo- and regioselectivity [2]. The mild reaction conditions and aqueous environments used in enzymatic synthesis reduce the need for toxic solvents, enhancing the environmental sustainability of the process [3,4]. Additionally, enzymes can be engineered and optimized to enhance their catalytic activity, stability, and substrate specificity [5,6,7]. As such, the production of a wide range of pharmaceutical compounds is viable, including antibiotics, anticancer agents, and atorvastatin [8,9]. Enzymatic synthesis also enables enantiopure compound production, which are crucial for drug efficacy and safety [2]. Enzymatic synthesis can be seamlessly integrated into existing pharmaceutical manufacturing processes, lowering costs and improving efficiency [1]. Ongoing research and development efforts in enzymatic synthesis hold substantial potential for the synthesis of novel and intricate pharmaceutical compounds, contributing to filling the biocatalytic toolbox [1]. In this regard, psychrophilic organisms present strong candidates for novel enzyme identification [10,11].

Carbohydrate esterases (CE) catalyze the deacetylation of acetylated saccharide residues in hemicellulose and pectin [12,13]. As such, CE facilitates the breakdown of plant biomass into simpler sugars for biofuels or value-added product production. Hemicellulose, a heterogeneous polysaccharide, constitutes a substantial portion of plant cell walls, typically including short chains of xylan, mannan, galactan, and arabinan. Among these polymeric xylans, the major component (20~35%) in hemicellulose, are acetylated at the C2 or C3 positions of xylose residues and require deacetylation for xylose production. Among CE enzymes, CE1–7, 12, and 16 are acetyl xylan esterases (AXEs). These AXEs have broad industrial applications, such as biobleaching, semi-synthetic β-lactam antibiotic production, and animal feed production [12].

Recently, the cold-active Paenibacillus sp. AXE (PbAcE) was isolated and characterized, along with its X-ray crystal structure [14]. PbAcE, displaying higher activity at 4 °C than that at 25 °C, belongs to the CE7 family, which catalyzes the hydrolysis of ester linkages between xylose and acetic acid, and is able to deacetylate cephalosporin antibiotics [15]. Similarly to other AXEs [12,16,17,18], PbAcE exhibits considerable potential for industrial applications. Moreover, its cold-active nature offers distinct advantages, enabling bioconversion processes at low temperatures. This characteristic reduces energy costs and minimizes contamination risk [19]. PbAcE shares common structural and catalytic characteristics with CE7 family members. The PbAcE X-ray crystal structure revealed a donut-shaped hexameric structure comprising a trimer of dimers. Because of this shape, the substrates can access active sites via the tunnel formed in the central ring of the hexameric structure, and the active site of each subunit faces toward the center of the ring [14]. Each PbAcE subunit has typical α/β hydrolase folds with a catalytic triad of Ser185-His303-Asp274. PbAcE, similarly to other members of the CE7 family, exhibits broad substrate specificity. The substrates are mostly acetate esters, including glucose penta-acetate and xylan acetate, tertiary alcohol esters, and antibiotics (such as cephalosporin C [CPC], cefotaxime and, 7-amino cephalosporanic acid [7-ACA]) [14]. PbAcE demonstrates diverse activities across different substrates, likely attributable to its substrate-binding pockets (Figure 1a and Figure S1). PbAcE accommodates the diverse alcohols moieties of substrates in a substrate binding pocket, designated as Figure S1 (Figure 1a); however, it can only fit short acyl moieties in the other substrate-binding pocket, designated as S0. Among the three tertiary alcohol esters examined in the previous study, namely tert-butyl, linalyl, and α-terpinyl acetates, PbAcE exhibited higher hydrolytic activity to tert-butyl acetate than the other two [14]. The enzyme also showed higher activity to 7-ACA and cefotaxime than to CPC. Enzymatic CPC deacetylation by PbAcE to deacetyl-CPC is an initial production step for deacetyl 7-ACA, which is a key pharmaceutical intermediate for semi-synthetic β-lactam antibiotic production [20]. Here, we engineered the S1 substrate-binding pocket of PbAcE to improve its activity toward tertiary alcohols, such as linalyl and α-terpinyl acetates. We covalently docked four ligands, namely tert-butyl acetate, linalyl acetate, α-terpinyl acetate, and glyceryl tribuyrate, to the Ser 185 of PbAcE using HCovDock [21,22] to map out the residues interacting with substrates. Furthermore, HotSpot Wizard 3.1 was used to find functional residues in AXE with high mutability [23]. Lys91, His93, and Tyr182 were selected for site-saturation mutagenesis (SSM). Positive mutants for each residue were screened from single mutant libraries using a colorimetric assay method: Ala, His, and Ser for Lys91 residue; Pro and Asp for His93; and Glu, Thr, Val, Met, and Trp for Tyr182. The activity of the mutant enzymes towards linalyl acetate and α-terpinyl acetates increased by 10–30% compared to that of the wild-type PbAcE. By combining the best single mutants from each residue, we sequentially generated double and triple mutants. However, no synergistic improvements were observed in the double and triple mutants compared to the wild-type and single mutants. Steady-state kinetics of mutant PbAcEs against the pNP-C₂ substrate revealed that acetyl binding and catalysis were not significantly affected by the mutations. Our findings may provide insight into engineering the substrate binding pockets of biocatalysts for industrial applications.

2. Materials and Methods

2.1. Chemicals and Reagents

Primers were synthesized from Macrogen (Seoul, Republic of Korea). p-nitrophenyl acetate (pNP-C₂), linalyl acetate, α-terpinyl acetate, tert-butyl acetate, and glyceryl tributyrate were purchased from Sigma-Aldrich (St. Louis, MO, USA). B-PER™ Bacterial Protein Extraction Reagent was from ThermoFisher Scientific (Waltham, MA, USA). Ni²⁺ affinity resin was obtained from Qiagen (Hilden, Germany). A QuikChange II kit was purchased from Agilent (Santa Clara, CA, USA). The 96-well PCR plate (Code 781368) was purchased from Axygen (Corning, NY, USA). All other reagents were of analytical grade, unless otherwise stated.

2.2. Molecular Docking, Functional Residue Identification, and In Silico Mutant Analysis

To identify putative residues involved in substrate selectivity, covalent docking of substrates with the protein was performed using the HCovDock server [21,22]. Linalyl, α-terpinyl, tert-butyl acetates, and glyceryl tributyrate were used as substrates. Input PDB files for the substrates were generated using MarvinSketch software. The substrates were imported into MarvinSketch, and the γ-oxygen and ß-carbon of the active Ser 185 of PbAcE were covalently attached to the carbonyl carbon of the substrates that underwent nucleophilic substitution during hydrolysis. Covalent docking was performed using the HCovDock protocol [21,22]. The ten best binding models were investigated to find important residues for substrate selectivity.

Hotspot Wizard 3.1 (http://loschmidt.chemi.muni.cz/hotspotwizard/; accessed on 23 March 2021) was used to predict functional hotspots [23]. The X-ray crystal structures of PbAcE, Thermotoga maritima AcE, and Bacillus subtilis AcE were submitted as analysis input. After comparison of the results from AcEs, residues detected in the access tunnels with high or moderate mutability in all AcEs were selected for site-saturation mutagenesis. The DynaMut web server [24] was used to analyze the stability of substitutions with the default parameters, and the resulting structures were visualized using PyMOL [25]. The substrate-binding pockets of mutants were calculated using the CavityPlus server (http://www.pkumdl.cn/cavityplus; accessed on 20 September 2021) [26].

2.3. Construction and Screening of Site-Saturation Mutagenesis Library

All molecular procedures were performed as previously described [27]. The pET28a vector harboring the PbAcE gene was used as a template to generate an SSM library for each residue selected to improve substrate specificity. Polymerase chain reaction (PCR) was performed with NNK (N = G or A or T or C, K = G or T) degenerate codon primers (Table S1) using a standard protocol: 5 min denaturation at 95 °C, 18 cycles of 95 °C for 30 s denaturation, 55 °C for 1 min annealing, 68 °C for 6 min extension, and a final 7 min extension at 68 °C. The PCR product was then treated with DpnI for 1 h at 37 °C and transformed into Escherichia coli (strain DH5α). The plasmids were harvested and transformed into E. coli BL21(DE3) for expression.

Two hundred colonies were initially screened for each SSM. Based on the change in color of phenol red as result of esterase activity, positive PbAcE colonies were screened [14,28,29,30]. Linalyl and terpinyl acetates were mainly used as substrates. Two hundred successful transformant colonies for each SSM were inoculated into a 96-well microplate containing 500 μL Luria–Bertani (LB) broth supplemented with 50 μg/mL substrate and grown at 37 °C overnight. Each colony was re-inoculated into the same medium, grown at 37 °C until the optical density (OD) at 600 nm reached 0.5. The culture was induced by 0.5 mM isopropyl-1-thio-β-D-galactopyranoside (IPTG) and incubated at 25 °C overnight. The cell lysates containing PbAcE mutant proteins were then prepared as follows: The cells were pelleted and suspended in 100 μL B-PER™ Bacterial Protein Extraction Reagent. Protein extract (100 μL) was transferred to a 96-well microplate (Corning, Costar, NY, USA), 100 μL 100 mM linalyl acetate or α-terpinyl acetate in 10 mM Tris-HCl was added, the pH was adjusted to 8.0, and 0.04 mg/mL phenol-red was added with thorough mixing. The reaction mixture was incubated at 25 °C until wild type wells indicated a color change. The change was photographed every 5 min. All screening was repeated twice. The top 12 positive mutants for each mutant library were selected using the intensity of color change.

The nucleotide sequence of the positive mutants was verified using DNA sequencing. The second substitutions of His93 to Pro or Asp were introduced in the Lys91Ala, Lys91His, and Lys91Ser mutants, respectively, using site-directed mutagenesis. In total, six double mutants were constructed. The third mutation of Tyr182 to Val, Thr, Glu, Met, and Trp was combined with Lys91Ala/His93Asp or Lys91A/His93Pro, generating 10 triple mutants. The activity of double and triple mutants was screened as described above. Finally, mutants exhibiting higher activity than that of the wild type were further examined.

2.4. Overexpression, Purification, and Kinetic Characterization of Mutant PbAcEs

To investigate substrate specificity and kinetic parameters, mutant PbAcEs were prepared as described previously [14]. The pET28a vectors containing PbAcE mutants were transformed into E. coli BL21(DE3) and grown in 100 mL of LB media. Protein expression was induced using 0.1 mM IPTG. The cell pellets were resuspended in buffer A (50 mM sodium phosphate, pH 8.0, 300 mM NaCl, 5 mM imidazole) with 0.2 mg/mL lysozyme, and lysed by sonication on ice for 5 min (5 s pulse and 10 s pause). The cell lysate was clarified by centrifugation at 12,000 rpm for 30 min at 4 °C. The supernatant was incubated with Ni²⁺ affinity resin (Qiagen, Hilden, Germany) for 30 min with gentle agitation. The protein-bound resin solution was loaded onto the column and settled. The resin was then washed three times with buffer A containing 20 mM imidazole. The protein was eluted in buffer A containing 250 mM imidazole. The fraction was visualized using SDS-PAGE. The fractions containing PbAcEs were concentrated and exchanged with buffer B (50 mM Tris-HCl, pH 8.0, and 150 mM NaCl) using Amicon ultracentrifuge filters (Ultracel-3K; Millipore, Darmstadt, Germany). Protein concentration was determined using the Bradford assay with bovine serum albumin (BSA) as a protein standard [28]. The activity of wild-type and mutant PbAcEs was measured in 10 mM Tris-HCl, pH 8.0 in the presence of 0.04 mg/mL phenol-red using a microplate reader (FilterMax F5, Molecular Devices LLC, Sunnyvale, CA, USA) as previously described [14]. Briefly, 200 μL reaction mixture containing 10 mM substrate and 5 μg enzyme was incubated for 80 min at 25 °C in a 96-well microplate. Linalyl and α-terpinyl acetates were used as substrates. Proton release owing to enzymatic cleavage of esters of substrates led to a reduction in the absorbance of phenol red at 550 nm, which was measured using a photometer. The enzyme activity was determined by converting ΔA_{550 nm}·min⁻¹ using the extinction coefficient of phenol red at 550 nm (8450 M^–1 cm^–1). One unit (U) of enzyme activity was defined as the amount of enzyme required to transform 1 μmol of substrate in 1 min under the assay conditions. The activity of wild-type PbAcE against each substrate was defined as 100%. All tests were conducted in triplicate. The results were expressed as the mean ± standard deviation (SD) with a sample size of three (n = 3).

The kinetic parameters of mutants were investigated with pNP-C₂ as a substrate. The reaction mixture contained 5 μg of wild-type or mutant PbAcEs and varying concentrations of pNP-C₂ in 1 mL of buffer B at 25 °C. The absorbance was recorded at 405 nm for 10 min, and the initial velocity was calculated from the initial linear slope of the measurement. The k_cat and K_M were determined using the Michalis–Menten equation.

3. Results and Discussion

3.1. Identification of Functional Residues for SSM

To increase the substrate specificity of PbAcE towards tertiary alcohol esters, we adopted a semi-rational protein engineering approach. Covalent docking is widely used to find residues that interact with ligands or substrates [21,29,30,31]. The X-ray crystal structures of CE7 family AcEs demonstrated that they have two pockets for substrate binding: one for acetyl groups and the other for alcohol groups (Figure 1a and Figure S1) [14,32,33,34,35]. The catalytic triad of Ser815-His303-Asp274 is located near the smaller acetyl binding pocket (designated as S0). The alcohol groups bind to the larger pocket (designated as S1), which are mostly composed of hydrophobic amino acids. Previously, the putative hydrophobic substrate binding site was suggested to consist of Lys91, Glu104, Trp108, Tyr182, Phe208, Leu306, His309, Glu310, and Met313 [14]. In this study, we covalently docked the substrates to the active site Ser185 using HCovDock [21]. The substrates docked included tert-butyl acetate, linalyl acetate, α-terpinyl acetate, glyceryl tributyrate, and glyceryl trihexanoate. To find common residues that interacted with most substrates and ones that interacted with longer aliphatic substrates, the substrates to be docked were not limited to tertiary alcohol esters. The interaction of PbAcE with each substrate is shown in Figure 1b–f. Figure 1f shows the top 10 glyceryl tributyrate models with lowest energy. The residues interacting with substrates in the top 10 models were analyzed and are presented in

Table 1. List of residues involved in substrate binding based on covalent docking model.

Residues	Substrates 1	Location/Function	Mutability 2
Lys91	GTB, GTH	S1	7
Tyr92	GTB, GTH	S1	-
His93	GTB, GTH	S1	5
Gly94	GTB, GTH	Oxyanion hole	-
Tyr95	All	Oxyanion hole	-
Ser96	LA, GTB, GTH	Oxyanion hole	-
Gly97	GTH	Gate keeper	6
Asn98	GTH	Gate keeper	6
Glu104	GTH	Gate keeper	-
Tyr182	GTH	S1	9
Gly184	All	Active site	-
Gln186	All	Oxyanion hole	-
Phe210	All	S0	-
Pro225	All	S0	5
Thr276	All	S0	5
Cys277	All	S0	-
His303	TA, LA, GTB	Catalytic triad	-

¹ Abbreviations: TBA, tert-butyl acetate; TA, α-terpinyl acetate; LA, linalyl acetate; GTB, glyceryl tributyrate; GTH, glyceryl trihexanoate. ² Data from HotSpot Wizard analysis. Values greater than six imply that the residues are located in a highly mutable position.

. In this step, amino acids involved in the active site or S0 were eliminated. To further sort the residues for SSM, functional hotspots in PbAcE were investigated using the HotSpot Wizard program. Hotspot Wizard performs in silico identification of functional hotspots of the protein, which are highly mutable residues in the pockets or tunnels. This helps to reduce the size of mutant libraries [36,37]. The prediction proposed 35 hotspot residues with mutability that exceeded six (Table S2). Hotspots were located evenly in the protein sequence, but only two residues, Lys91 and Tyr182 (in the S1 pocket), were predicted as hotspots. Taken together, we selected Lys91, Tyr182, and His93 for SSM, to improve substrate specificity for linalyl and α-terpinyl acetates; however, Tyr92 was excluded, since it is not exclusively conserved in CE7 AcEs, but its side chain OH group faces the interior and forms hydrogen bonds with Ile162 NH.

3.2. Screening of Mutants with Improved Enzme Activity

Initially, single mutant libraries for each residue in the S1 pocket were generated. From each SSM library, approximately 200 colonies were randomly selected to establish 95% coverage of the desired amino acids substitutions and screened [30,36]. The qualitative colorimetric screening of mutant library was conducted against linalyl and α-terpinyl acetates, since PbAcE showed lower activity towards two substrates compared to tert-butyl acetate. The reaction lasted until the wild type showed a color change from red to yellow, during which the changes of the 96-well plates were photographed every 5 min (Figure 2). We attempted to narrow down the positive mutants to only those that showed improved activity in both substrates. In all libraries, the top twelve positive colonies for each residue were selected and sequenced. The sequencing results for each residue were as follows, with the substituted amino acid and time of appearance in parentheses: Lys91 substituted with Ala (6), His (5), Ser (1); His93 with Pro (4), Asp (2), Ala (2), Gln (1), Val (1), Cys (1), Gly (1); and Tyr182 with Val (3), Thr (2), Glu (3), Met (2), Trp (2). Notably, most replacements were frequently occurring amino acids in AcEs of CE7 (Figure S2). The His and Ala at position 91 occurred more frequently than Lys, while His was the most commonly occurring amino acids at residue 93, followed by Pro, Ile, and Leu. Position 182 was occupied by smaller and more polar residues compared to that of Tyr.

The single positive mutants exhibiting relatively higher activity compared to that of the wild-type were selected as follows: Ala, His, and Ser at Lys91; Pro and Asp at His93; and Glu, Val, Met, Thr, and Trp at Tyr182. To efficiently harness the potential synergistic effects of the mutants, we sequentially combined mutations from each residue. The combination of individual positive mutants often results in an increase in enzyme activity [30,38]. In this study, we introduced Pro and Asp mutations at His93 into three K91 mutants: K91A, K91H, and K91S. Colorimetric screening was performed as described above; however, no distinct changes were detected in the double mutants (Figure 3a). Of the double mutants, K91A/H93P and K91H/H93P showed slightly better activity compared to that of the other double mutants. Not surprisingly, Proline was the second most frequently occurring amino acid at position 93 (Figure S2). Subsequently, triple mutants were generated by introducing an additional mutation—Glu, Val, Met, Thr, and Trp at Tyr182—into the double mutant templates of K91A/H93P and K91H/H93P. The activity of the triple mutants was screened, revealing that the Y182E, Y182V, Y182M, Y182T, and Y182W variants in the K91A/H93P double mutant exhibited slightly higher or comparable activity to that of the wild type (Figure 3b). Contrary to other reports, which indicated that combining individual mutations into the wild type increased substrate selectivity of lipase [30,38] and thermal tolerance of phytase [30,38], our results did not show this cumulative effect.

3.3. Enzyme Activity and Structural Analysis

To further characterize the mutants, we expressed and purified eight single mutants (K91A, K91H, K91S, H93P, Y182E, Y182V, Y182M, and Y182W), two double mutants (K91A/H93P and K91H/H93P), and five triple mutants (K91A/H93P/Y182E, K91A/H93P/Y182M, K91A/H93P/Y182T, K91A/H93P/Y182V, and K91A/H93P/Y182W) (Figure S3). The enzymatic properties of these mutants were evaluated by measuring their relative activity against tertiary alcohols, such as linalyl acetate and α-terpinyl acetate. The relative activities are presented in Figure 4a, with the corresponding colorimetric assay data displayed in Figure 4b. For linalyl acetate as the substrate, all mutants, except K91A/H93P/Y182W, exhibited higher activity than the wild type. Single mutants at the three different residues showed increased activities ranging between 10 and 30% relative to the wild-type PbAcE. The His93 mutations exhibited the smallest increase (10%) compared to mutations at Lys91 and Y182, which showed a 20–30% increase. Despite the introduction of multiple positive mutations, double and triple mutants only increased activity by approximately 10% relative to the wild type. For α-terpinyl acetate, the relative activity enhancement pattern of mutant PbAcEs was similar to that observed with linalyl acetate, although the overall increment was generally lower. Single mutants demonstrated a greater increase in activity compared to that in double and triple mutants.

Steady-state kinetics of PbAcE mutants were conducted using pNP-C₂ as a substrate to understand the impact of mutations on substrate affinity and catalysis. As shown in Figure 5a and

Table 2. The steady-state kinetic parameters of PbAcEs using pNP-C₂ as a substrate.

Enzyme	kcat (s−1)	KM (mM)	kcat/KM (mM−1s−1)
Wild type	52.5 ± 0.9	0.19 ± 0.01	282.3
K91A	53.5± 1.1	0.17 ± 0.01	308.8
K91H	55 ± 0.8	0.16 ± 0.02	343.8
H93P	52.3 ± 0.9	0.18 ± 0.01	288.9
Y182E	55 ± 0.9	0.17 ± 0.01	323.1
Y182V	54 ± 0.9	0.16 ± 0.01	337.1
Y182M	52.5 ± 1.2	0.16 ± 0.01	328.1
Y182T	52.5 ± 1.0	0.16 ± 0.01	328.1
Y182W	55 ± 1.5	0.16 ± 0.01	343.8
K91A/H93P	50.5 ± 0.9	0.17 ± 0.02	297.1
K91H/H93P	51.4 ± 1.0	0.17 ± 0.01	302.4
K91A/H93P/Y182E	50.3 ± 0.7	0.16 ± 0.01	312.1
K91A/H93P/Y182V	52.5 ± 0.7	0.17 ± 0.01	308.1
K91A/H93P/Y182M	52.5 ± 0.9	0.17 ± 0.01	308.7
K91A/H93P/Y182T	54.3 ± 1.2	0.17 ± 0.01	319.3
K91A/H93P/Y182W	51.2 ± 1.1	0.17 ± 0.01	301.2

, most of the mutants showed slightly lower K_M values for pNP-C₂ compared to that of the wild type, but the differences were not significant. This suggests that the mutations had little effect on kinetic parameters related to pNP-C₂ substrate. This is not surprising, as the mutations at three residues only enlarge the alcohol moiety binding pocket that pNP-C₂ does not occupy (Figure 5b). pNP-C₂ only binds the S0 pocket in the PbAcE. The p-nitrophenyl group has no interaction with the hydrophobic substrate binding pocket (S1), thus the mutation did not affect catalysis.

Furthermore, we sought to understand the effects of the introduced mutations on the substrate-binding pocket size, substrate binding, protein stability, and conformation. To achieve this, we used the CavityPlus [26] and Dynamut [24] programs. The CavityPlus analysis demonstrated that the mutation improved the substrate-binding pocket (Figure 6). The pocket volume was 818.38 ų for the wild type, 858.38 ų for the K91A mutant, 884.88 ų for Y182E mutant, 855.38 ų for K91A/H93P mutant, and 974.62 ų for K91A/H93P/Y182E mutant. Notably, mutations in Lys91 and Tyr182 to smaller amino acids resulted in an enlarged binding pocket, potentially easing the accommodation of a large moiety of substrates.

The impact of mutation on protein stability and conformation was predicted using the DynaMut web server [24]. The protein stability and interatomic interaction were calculated and visualized (Figure S5). The mutation at Lys91 to Ala caused a reduced number of interactions because of the loss of positive charge and two methylene groups, which led to a slight destabilization (ΔΔG −0.334 kcal/mol) but increased the flexibility (Figure S5a and Figure 5b). The second mutation His93 to Pro slightly increased the stability and flexibility (Figure S5c and Figure 5d). The triple mutant (K91A/H93P/Y182E) had a lower stability than the double mutant (K91A/H93P), owing to less hydrophobic interactions (Figure S5e and Figure 5f). However, the mutations in this study did not seem to impact protein stability, since ΔΔG was not significant [24].

We evaluated the residues of cold-active PbAcE that can prevent the proper binding of substrates using docking simulation and HotSpot Wizard software. PbAcE is an intriguing enzyme, since it presents broad substrate specificity and high activity at low temperatures. In particular, its characteristic of deacetylating CPC and related antibiotics render this enzyme a strong protein engineering candidate for enzyme-based pharmaceutical compound synthesis [1,7,14,39]. Of its substrates, tertiary alcohol esters have low binding affinity to the substrate-binding pockets. PbAcE has a small acetyl binding pocket and large hydrophobic pocket, where the alcohol group of tertiary alcohol esters binds. To improve the suitability of the suRbstrate-binding of the large pocket, we performed structure-guided engineering. Presently, there is growing number of publications supporting structure-guided protein engineering [7,30,35,36,37,40,41]. Based on docking simulation and hotspot analysis, we determined three suitable amino acids to be engineered. Structure-guided protein engineering combined with SSM or iterative saturation mutagenesis (ISM) has been widely used [30,36,37,42,43]. Usually in an ISM procedure, an improved mutant from the first-generation mutagenesis is chosen for the second-generation mutagenesis at other selected sites. After the first generation of SSM, we successfully harvested single mutants with improved activity. Instead of taking the second generation of SSM, we mainly combined the single best mutants using site-directed mutagenesis [36,42,43]. Contrary to other data showing the synergistic effect of introducing multiple mutations, our double and triple mutants did not exhibit higher activity than that of the corresponding single mutants. The pNP-C₂ kinetic analysis, which did not face or bind to the engineered pocket, revealed that the introduced mutations did not affect the kinetic parameters for the substrate. This approach will provide insights into the substrate specificity and the interaction of substrate and pockets. Further investigation to expand the acetyl binding pocket to accept large acyl groups will fuel the development of efficient biocatalysts for industrial applications.