Identification, characterization and site-saturation mutagenesis of a thermostable ω-transaminase from Chloroflexi bacterium

In present study, we have mined a ω-transaminase (ω-TA) from Chloroflexi bacterium from genome database by using an ω-TA sequence ATA117 Arrmut11 from Arthrobacter sp . KNK168 and an amine transaminase from Aspergillus terreus as templates in a BLASTP search and motif sequence alignment. The protein sequence of the ω-TA from C. bacterium shows 38% sequence identity to ATA117 Arrmut11. The gene sequence of the ω-TA was inserted into pRSF-Duet1 and functionally expressed in E. coli BL21(DE3). Results showed that the recombinant ω-TA has a specific activity of 1.19 U/mg at pH 8.5, 40 °C. The substrate acceptability test showed that ω-TA has significant reactivity to aromatic amino donors and amino receptors. More importantly, the ω-TA also exhibited a good affinity towards some cyclic substrates such as 1-Boc-3-piperidone. The homology model of the ω-TA was built by Discovery Studio and docking was performed to describe the relative activity towards some substrates. The ω-TA was evolved by site-saturation mutagenesis and found that the Q192G mutant increased the activity to the (R)-α-methylbenzylamine (MBA) by around seven-fold. The Q192G mutant was then used to convert two cyclic ketones, N -Boc-3-pyrrolidinone and N -Boc-3-aminopiperidine, and the conversions were both improved compared to the parental ω-TA. Graphical Abstract


Introduction
Chiral amines are very important intermediates, which have a wide range of applications in medicine and fine chemical industries [1]. The production of chiral amines include chemical and biochemical methods. As far as chemical methods are concerned, chiral amines are usually produced by asymmetric catalysis of prochiral molecules, such as hydrogenation of imine or enamine, alkylation of imine, amino hydroxylation and reductive amination [2]. In order to overcome the drawbacks of low enantioselectivity, use of noble metal catalyst, harsh reaction condition, and environmental concern in asymmetric synthesis, several biochemical methods using efficient biocatalysts have been developed, and manage to replace chemical methods. These methods mainly involve enzymes such as amine oxidases, ammonia lyases, amine dehydrogenase and transaminases [3][4][5]. Among them, ω-transaminases (ω-TAs), which catalyze the transfer of ketone and amino groups, have received extensive attention [6][7][8][9]. The well-known example of the synthesis of chiral amines by ω-TAs is the manufacture of sitagliptin, an oral antihyperglycemic drug [10].
Interestingly, ω-TAs originating from various sources have showed distinct selectivity, that is (R)-selective or (S)-selective. Hence, both enantiomers of specific amines can be produced by ω-TAs with different enantioselectivity. Although the ω-TA Ata-117 Arrmut11 from Arthrobacter sp. KNK168 is commercially available now, only a few (R)-selective ω-TAs have been identified [11,12]. Therefore, the discovery of novel R-selective ω-TAs was still the focus issue in the past few years. Bornscheuer et al. [13,14] developed a method for predicting and screening enzyme function on the basis of key motif in sequence, which has been proved to be effective and successfully used to search and identify 17 (R)-selective ω-TAs by database mining. The method has also been applied to discover five other different sources of (R)-selective ω-TAs [15]. In recent years, more and more (R)-selective ω-TAs have been discerned and investigated, but their wide applications are still limited by the poor properties [16][17][18]. In order to improve the performance of enzymes, protein engineering has usually been adopted to conduct enzyme design and optimization. A (R)-selective ω-TA has been created by bioinformatic analysis combined with computational redesign of the D-amino acid aminotransferase, exhibiting a specific activity close to those of natural (R)-selective ω-TAs [19].On the basis of a fluorescence-based screening system, a KnowVolution campaign has been carried out to optimize a (R)-selective ω-TA from Mycobacterium vanbaalenii and the best resulting mutant showed specific activity to acetonaphthone more than 100 times higher than parental enzyme [20]. In another study, a (R)-selective ω-TA from A. cumminsii ZJUT212 has been modified using a semi-rational protein design, and a mutant has been screened to produce sitagliptin intermediate on a kilogram scale with >99 % e.e. and approximately 80% conversion [21].
In the current study, BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi) was performed against thermophilic bacteria in NCBI to identify a putative ω-TA sequence from Chloroflexi bacterium using Ata-117 Arrmut11 from Arthrobacter sp. KNK168 and amine transaminase from Aspergillus terreus as templates. Subsequently, the putative ω-TA gene was cloned into E. coli for expression and its kinetic parameters and substrate spectra were characterized. Thereafter, the ω-TA was improved by site-saturation mutagenesis on the basis of reported results and modeling analysis.

Reagents
All the chemicals used were of analytical grade and purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. or Sigma-Aldrich, Inc. unless otherwise noted.

Sequence analysis
The Ata-117 Arrmut11 sequence from Arthrobacter sp. KNK168 and amine transaminase from A. terreus, were used as templates for BLAST search in NCBI to find new ω-TA sequence. After removing incomplete and redundant sequences, MEGA 7.0 was used to align all unknown sequences, and the target sequences which conform to the functional characteristics of (R)-selective ω-TA were screened out according to the motif sequence confirmed in reference [9].
Among them, a putative (R)-selective ω-TA from C. bacterium was chosen for further study, which shows 38% identity with Ata-117 Arrmut11. A phylogenetic tree was also generated with MEGA 7.0 [22] by using the neighbor-joining algorithm [23].

Plasmid and strain Construction
The gene sequence of the putative (R)-selective ω-TA from C. bacterium was further codon-optimized and synthesized by GenScript (Nanjing, China). The encoding region was inserted between the BamH I and Xho I, with a 6-histidine tag at the N-terminal, resulting in plasmid pRSF-CbTA. The recombinant plasmid was transformed into E. coli BL21(DE3) to generate the recombinant strain of E. coli BL21(DE3)/pRSF-CbTA.
When the OD600 of the culture reaches about 0.6, 0.02 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) was added, and the culture was cultivated at 20°C for another 24 h. Thereafter the cells were harvested and re-suspended in 20 mL of glycine-NaOH buffer (50 mM, pH 8.5), and sonicated in an ice bath for 30 min. The cell debris was removed by centrifugation at 8000 × g at 4°C for 40 min. The supernatant was passed through a 5 mL Ni-NTA column to purify the recombinant ω-TA, and then the purified ω-TA solution was then stored at 4°C for further experiments. The molecular weight of the purified ω-TA were determined by 12% (w/v) SDS-PAGE [24]. A control experiment was performed under the same conditions using E. coli transformant carrying empty pRSFDuet-1.

Assay of enzyme activity
Enzyme activity analysis was performed in 500 µL glycine-NaOH buffer (50 mM, pH 8.5) containing pyridoxal 5'-phosphate (PLP) (20 μM), (R)-α-methylbenzylamine (MBA, 10 mM), sodium pyruvate (10 mM). The reaction started at 40°C by adding 0.05 mg of the purified ω-TA to the mixture. After 30 min the mixture was heated and boiled for 20 min to stop the reaction. All experiments were performed in three replicates. The resulting acetophenone was analyzed by HPLC. One unit of enzyme activity is defined as the amount of ω-TA catalyzing the release of 1 μmol acetophenone per minute.

Characterization of biochemical properties
By using the following buffers with different pH 6.0-10.5 (50 mM phosphate buffer, pH 6.0-8.0; 50 mM glycine-NaOH buffer, pH 8.5-10.5), the optimal pH of the ω-TA was determined at 40°C.
The optimal temperature for the ω-TA was investigated from 25°C to 65°C at pH 8.5. Relative activity (%) was calculated using the maximal activity as a control (100%). The temperature stability was measured by incubating the ω-TA for a period in the temperature range of 40-65°C .
The tolerance of the ω-TA to organic solvents is determined by incubating 100 µL of the ω-TA in glycine-NaOH buffer (50 mM, pH 8.5) for 5 h at room temperature. The buffer contains 0%, 5%, 10%, 20%, 40%, 60%, 80% (v/v) organic co-solvents (acetonitrile or DMSO or methanol or ethanol). After that, residual activities were evaluated using the assay method of enzyme activity described above.

Determination of kinetic parameters
The kinetic parameters of the ω-TA to (R)-MBA or pyruvate were measured at 45°C and pH 8.5. According to the method described above, the initial reaction rate was determined under the conditions of different concentrations of (R)-MBA or pyruvate. The production of acetophenone and the consumption of pyruvate were analyzed by HPLC. In order to determine the kinetic parameters of the ω-TA to (R)-MBA, the reaction solution was 500 µL glycine-NaOH buffer (50

Characterization of the substrate specificity of the ω-TA
Substrate specificity of the ω-TA was tested by reaction between different groups of amino donors and amino receptors. All experiments were performed in triplicate. When the specificity of the amino donors were studied, the amino donors were listed in table 2,3. The activity assay for each reaction was carried out in a 50 mM glycine-NaOH buffer (pH 8.5) containing 25 mM amino donor, 25 mM sodium pyruvate, 0.02 mM PLP and a suitable amount of purified enzyme, with a final volume of 500 µL at 40°C for 30 min. When the specificity of the amino receptors were studied, the amino receptors were listed in table 4,5. The activity assay for each reaction was carried out in a 50 mM glycine-NaOH buffer (pH 8.

Site-saturation mutagenesis
The amino acid residues Y190, Q192, and G288 were selected for site-saturation mutagenesis on the basis of previous results and modeling analysis [19]. The site saturation mutagenesis was designed by NDT codon design. All primers are listed in Table 1 Thereafter, the activity of the mutant was tested as described above.

Sequence analysis
Nowadays, more and more protein sequences are stored in the database, which provides a huge resource for the mining of biocatalysts. Two independent BLAST searches were performed using two crystallized (R)-selective ω-TAs, Ata-117 Arrmut11 (PDB: 5FR9) from Arthrobacter sp.
KNK168 and amine transaminase (PDB: 4CE5) from A. terreus as templates. Then a rational analysis of the aligned sequences was performed use motif sequence alignment following the criteria previously established by Höhne and coworkers [13]. Four putative sequences derived from thermophilic bacteria were obtained from the NCBI non-redundant protein sequence library.
According to preliminary activity test, the ω-TA from C. bacterium with the highest activity was selected as the target.
The ω-TA from C. bacterium showed identity with amine transaminases from the following analysis was performed to verify the taxonomic and evolutionary relationship of the ω-TA to previously reported transaminases, and the results are presented in Fig. 1. Ten amino acid sequences with different degrees of sequence identity to the ω-TA were collected and compared.
The ω-TA was found to have a close evolutionary relationship with other sequences, except for thermophilic archaea G. acetivoras.

Purification and identification of recombinant ω-TA
The recombinant E. coli BL21(DE3)/pRSF-CbTA cells were cultured, collected and sonicated. The expressed recombinant protein with a theoretical molecular weight of 44.3kda accounted for more than 30% of the total protein in the cells, and was detectable in both the soluble and insoluble parts of E. coli BL21 (DE3)/pRSF-CbTA cells (Fig. 2). The recombinant protein was purified by immobilized metal ion affinity chromatography with the assistance of His-tag at the N-terminus of the sequence, and the purification results are shown in Fig. 2. The followed activity test indicated that the specific activity of the ω-TA was 1.19U/mg at pH 8.5, 40°C.

Enzymatic properties of the ω-TA
As is showed in Fig. 3a, the effect of pH on the ω-TA activity was assessed over the range of pH 6.0 to 10.5, and the ω-TA has the highest activity at pH 8.5, which is similar to the known ω-TAs [27,28]. For example, a typical (R)-selective ω-TA Arthrobacter sp. KNK168 has showed an optimal pH of 8.0 to 9.0 [29]. Besides, a branched-chain amino acid aminotransferase TUZN1299 identified from the genome of the hyperthermophilic archaeon Thermoproteus uzoniensis has exhibited the optimal pH of 8.0 [30]. The effect of temperature on the ω-TA activity was evaluated over the range of 25-65°C, and the highest activity was detected at 45°C (Fig. 3b). Fig. 3c presents the temperature profile of the ω-TA, and results the ω-TA has good thermal stability.
Interestingly, with the increase of incubation time at 45°C, the highest activity of the ω-TA can reach twice of the original activity, which was similar to an (S)-selective thermophilic transaminase from Geobacillus thermodelinificans [31]. This increase in activity may be due to the elevated stable interaction of temperature dependent folding [32]. The temperature dependence also showed that the ω-TA remain more than 50% activity after 12 h of incubation at 50-55°C, but the ω-TA was significantly inactivated at above 60°C. Similarly, most of the thermophilic TAs have showed significant thermal stability at about 55°C [33]. For instance, it has been found that a transaminase identified from Thermomicrobium roseum retained 50% activity after 5 h of incubation at 70°C [34]. A β-amino acid transaminase Ms-TA2 has also been discovered from a Meiothermus strain isolated in an Icelandic hot spring, which kept around 60% activity after incubation at 50°C for 3 h [35]. However, few thermophilic TAs remain active after incubation at temperatures above 65°C. A novel amine transaminase has demonstrated spectacular thermostability, and it activity can be maintained at 85% or around 40% after being incubated at 80°C for 5 d or 14 d [36].
Organic cosolvent is very necessary to increase the solubility of substrates. The influence of organic cosolvent on the ω-TA activity was determined by adding methanol, ethanol, DMSO and acetonitrile in reaction solutions (Fig. 3d). Results showed that the ω-TA activity increased significantly in all four 20-40% organic co-solvents, while decreased in 60% organic co-solvents.
This indicates that ω-TA is well tolerated by organic solvents and the results are superior to that of a solvent-tolerant haloarchaeal (R)-selective transaminase isolated from a Triassic Period salt mine [37].

Kinetic Parameters of the ω-TA
The kinetic parameters are helpful to evaluate the catalytic ability of enzymes. Given that transaminases operate via a typical dual substrate recognition, the Km and kcat to either substrate of the ω-TA were determined by changing the concentration of (R)-MBA or pyruvate from 2.5 mM to 40 mM respectively. With 20 mM pyruvate as co-substrate, the Km and kcat of the ω-TA to (R)-MBA are 14.68 mM and 3397.63 s -1 , while the Km and kcat of the ω-TA for pyruvate are 2.834 mM and 1617.5 s -1 , respectively. Compared with the ATA117 from Arthrobacter sp. KNK168, the ω-TA showed similar affinity for pyruvate, but its Km value for (R)-MBA was significantly higher than that of ATA117, which indicated that the ω-TA had slightly lower affinity for amine. However, the ω-TA showed lower Km to pyruvate and higher kcat than a transaminase from Fusarium oxysporum [38], which means higher reactivity on pyruvate (Fig. 4).

Fig. 4
The nonlinear regression fitting of Michaelis-Menten equation

Substrate specificity of the ω-TA
Previous studies on substrate range and stereoselectivity have shown that the active center of transaminase is composed of two active pockets. The large pocket can embrace the large structure and the small pocket is usually limited to methyl-sized substituents [39]. Considering the binding of the substrate to the ω-TA active center, we resolved the structure of the substrate on the basis of the ω-TA active site model, ω-TA's large pocket uses R1, while ω-TA's small pocket contains R2.
All substrates were divided into four groups: aromatic and aliphatic amines, amino acids, ketones and aldehydes to allow us to compare reactivity chemically.   Table 2 and Table 3: 0.1 mg mL −1 purified enzyme, 25 mM pyruvate and either 25 mM amino donor, 1 mL glycine-NaOH buffer (50 mM, pH 8.5) and 40°C. One unit of the enzyme activity was defined as the conversion of 1 μmol pyruvate per minute. The relative activity for (R)-MBA was designated as 100%.   Table 4 and Table 5: 0.1 mg mL −1 purified enzyme, 25 mM (R)-MBA and either 25 mM amino acceptor, 1 mL glycine-NaOH bufer (50 mM, pH 8.5) and 40°C. One unit of the enzyme activity was defined as the conversion of 1 μmol acetophenone per minute. The relative activity for pyruvate was designated as 100%.
As is showed in Table 2, Table 3, Table 4 and Table 5, we can find that the ω-TA can utilize a wide range of amino donors and acceptors for the substrate range, which is consistent with the literature [25], indicating their potential for synthetic applications. It can be seen from Table 2 that the ω-TA has good activity for most aliphatic and aromatic amines, but there are a marked difference among various aliphatic or aromatic amines, which is similar to the results reported by Jiang et al [15]. A comparison of the relative activity of the ω-TA to aliphatic amines revealed that it showed higher activity for the long chain amines (A5-A8), while it exhibited a very low reactivity against the short chain amines isopropylamine (A3) and 2-butylamine A4) except for propylamine (A2).
Interestingly, the ω-TA displayed higher reactivity to propylamine but extremely low to isopropylamine unlike other amine transaminases. Furthermore, the ω-TA is quite reactive to cyclic amines, and its relative activity is normally better with the increase of the ring (A9-A13).
Unexpectedly, when amino acids are used as the amino donors (B1-B4), the activity given by ω-TA is not very high, or even very low, e.g. 2-aminobutyric acid (B5).
The amino receptor specificity of the ω-TA was studied and the results are presented in Table 4 and Table 5. The ω-TA appears to have similar amino receptor specificity for Capronia semiimmersa [40]. The amino receptor spectrum shows that nearly all aliphatic, cyclic, and aromatic ketones show low reactivity except for 2'-nitroacetophenone (C7), 4'-methoxyacetophenone (C8), benzaldehyde (D5) and 5-nitrosalicylaldehyde (D6). As the results given in tables, the amino acceptors with aryl group have comparatively good reactivity, and the aryl ketones and aromatic aldehydes presents similar results. However, the low reactivity is observed in the case where the selected aliphatic ketones or aldehydes were used as amino acceptors. In comparison, the reactivity of the ω-TA towards aliphatic ketones and aldehydes was different from previously reported results [41]. It should be noted that in most cases the benzene ring with stronger electron withdrawing group or the methyl with substituent group in aryl ketones would dramatically inhibit the reactivity of the amino acceptors, and the existence of methylene between aryl group and aldehyde group also significantly reduced the activity of the ω-TA to aryl aldehydes. Additionally, the ω-TA appears to be reactive to two selected cyclic substrates with ketone groups. Discovery studio proposed an active site model of the ω-TA binding to PLP. The most favorable binding model contains several reported key residues. Previous study has reported that residue K252 shows significant catalytic reactivity with coenzymes and substrates [42]. As is shown in Fig. 5, K252 which binds to the center of the pocket has a strong hydrogen bond force and attractive charge with PLP. At the same time, strong hydrogen bond force was also found between PLP and residues G288 and S346. The residue E285 is located in the substrate-cofactor binding pocket and has conventional hydrogen bonding with the amino group of (R)-MBA. In addition, it is observed that the force contribution of the oxygen atom of phosphate group in PLP was provided by K252, G288 and E285, which is similar to other (R)-selective transaminase [43].

Fig. 6
Docking results of the ω-TA with selected amino donors listed in Table 2 and Table 3 To manage to explain the possible cause of substrate specificity, some selected amino donors and acceptors were docked with ω-TA and PLP, and the results were visualized by Discovery Studio.
The most favorable docking configuration was selected to analyze the forces. From the docking results of aliphatic amines (A2-A8), it can be seen that the residues in the binding sphere have Pi-alkyl and attractive charge on the carbon chain, but there is no such force on isopropylamine (A3). Meanwhile, the residue Q259 also has unfavorable bump on A3, which explains the low reactivity of A3. According to the docking results of cyclic amines, PLP has strong acting forces such as Pi-cation interaction and conventional hydrogen bond on the amino groups of cyclic substrates cyclobutanamine (A10) and cyclopentylamine (A11), while the acting force on amino groups of cyclopropylamine (A9) is only C-H bond. This indicates that with the increase of the size of the cyclic substrate, the L-pocket becomes more suitable for binding ligands. When aromatic substrates were involved, there were many intermolecular forces between the residue in the binding sphere and the benzene ring, including T-shaped Pi-Pi stack and Pi-alkyl interaction.
Taking (R)-1-(3-chlorophenyl) ethylamine (A17) as an example. we can see that residues R287 and Q259 have strong conventional hydrogen bond with amino group of A17, and chloro group in benzene ring is an electron-withdrawing group forming Pi-alkyl force with residue H260, which may lead to high reactivity of A17. In addition, the low activity of benzylamine (A14) may be due to the fact that coenzyme PLP has no interaction with it, and residue E259 also has an Unfavorable Donor-Donor to its amino group.  Tables 4 and Table 5 The docking results of the selected amino acceptors with the active site model of the ω-TA binding to PLP. are shown in Fig. 7. We can observed that pyridoxamine 5'-phosphate (PMP) has a strong hydrogen bond to the aldehyde group in the typical acceptor pyruvate (C1), and residues T242, N253, and H258 also bear corresponding hydrogen bond to the aldehyde group and carboxyl group, indicating that the ω-TA has a strong affinity for pyruvic acid. As for α-ketoglutarate (C6), the interaction between PMP and substrate was not found, which may be the reason why the ω-TA has no activity on C6. From the docking results of ketones, it could be perceived that the active site of PMP and the enzyme tended to bind to stable structures containing benzene rings or rings, and only exhibited weak van der waals for straight chain substrates, which might be the reason why most of the aliphatic ketones were low or even inactive, e.g. pentanone (C3). For aromatic ketones, the electron-withdrawing groups of 2'-nitroacetophenone (C7) with good reactivity shows the attractive charge and salt bridge with PMP. In the case of 4'-methoxyacetophenone (C8), PMP has Pi-alkyl interaction to its benzene ring, and residues L242 and N241 have Pi-Pi T-shaped interaction and hydrogen bond to the benzene ring and ketone groups, with its methoxy group having additional C-H interaction binding to the ω-TA. In contrast to other acceptors, the substituted R2 groups show little interaction with the ω-TA and PMP, suggesting that the small pocket of the ω-TA are not suitable for substrates with R2 more than one methyl group, which may be responsible for the lack of activity of 2-hydroxyacetophenone (C13) and valerophenone (C14). For aldehydes, residues in both PMP and the binding sphere have a conventional hydrogen bond to the aldehyde group except for propionaldehyde (D1) and 2-ethylhexanal (D4), which explains the low reactivity of D1 and D4. With regard to the most reactive amino acceptors benzaldehyde (D5) and 5-nitrosalicylaldehyde (D6), residue G259 exerted a strong force on the aldehyde groups of D5 and D6, which was not available for other aldehydes from the docking results. In addition, the nitro group of D6 has an additional conventional hydrogen bond with the active site as an electron withdrawing group.

Site-saturation mutagenesis of the ω-TA
Enzymatic transamination of ketones to amines is an important method to obtain various chiral building blocks or pharmaceuticals. In this study, a new ω-TA was obtained and its enzymatic properties and substrate spectrum were determined, showing the potential of the ω-TA for the production of chiral amines and unnatural amino acids. However, the unsatisfactory activity on many ketones of the ω-TA is a vital obstacle for its application in chiral amine production. In order to improve its activity, site-saturation mutagenesis was conducted on the basis of previous results and homology modeling.
As confirmed by Voss et al, residues F130, Y190, Q190, G288, and A348 showed significant influence on the activity of amine transaminase. Docking result of the ω-TA with PLP and (R)-MBA indicated that residues K252, E285, G288, Q192 and Y190 are all located in the small pocket of the active site. Hereby, residues Y190, Q192, and G288 were selected for site saturation mutation, and the results are presented in Fig. 9. The results showed that the changes in residues Y190 and Q192 did have a significant effect on the activity of the ω-TA to (R)-MBA. Among them, mutant Q192G was the most beneficial mutation (Fig. 10). However, there was no significant change in the activity of mutants G288 for the ω-TA. The ω-TA and mutant Q192G was used to convert acetophenone to produce (R)-MBA, and results showed that after 16 h of reaction, the conversion of (R)-MBA for mutant Q192G was around 55% with e.e. ≥99%, which was more than seven times higher than that of the ω-TA. Fig. 9 The results of site saturation mutation of the ω-TA. Subsequently, the ω-TA and mutant Q192G were further applied to evaluate the conversion of cyclic substrates with ketone group, and results are shown in Fig. 10. After 36 h of reaction, the conversion of N-Boc-pyrrolidinone and N-Boc-piperidone by the ω-TA was about 15% and 12%, and the conversion of N-Boc-pyrrolidinone and N-Boc-piperidone by mutant Q192G was increased to 23% and 30%, respectively. Compared with the ω-TA, the activity of mutant Q192G was improved evidently. However, further study is necessary to continually enhance the activity of the ω-TA.

Conclusion
(R) -Selective ω-TAs has important potential for industrial applications. In this work, a hypothetical (R)-selective ω-TA from C. bacterium was found by motif sequence blast from genome mining. On the basis of the amino acid sequences, a phylogenetic tree was constructed to verify the taxonomic and evolutionary relationships of the ω-TA with other amine transaminase family homologues. Ten amino acid sequences with different degrees of sequence identity to the ω-TA were selected and analyzed, and the ω-TA showed a close evolutionary relationship to pyruvate transferase from M. vanbaalenii PYR-1 and Ata-117 Arrmut11. By characterizing the enzymatic properties of the recombinant ω-TA, results showed that the ω-TA has good thermal stability, organic solvent tolerance and broad substrate specificity. Thereafter, the site-saturation mutation of the ω-TA was conducted and the mutant Q192G with higher activity was screened, which was applied in a conversion of N-Boc-pyrrolidinone and N-Boc-piperidone. Results showed that the ω-TA can be used as a valuable catalyst for the asymmetric synthesis of chiral amines from the corresponding aldehydes or ketones.

Availability of data and materials
The data and the materials are all available in this article.