Novel ( S )-Selective Hydrolase from Arthrobacter sp. K5 for Kinetic Resolution of Cyclic Amines

: Chiral 2-methylpiperidine (2-MPI) is an important building block that has potential for applications in pharmaceuticals and pesticides. In this study, we observed that the hydrolase in Arthrobacter sp. K5 exhibits high ( S )-selectivity toward rac - N -pivaloyl-2-MPI to yield ( S )-2-MPI with 80.2% enantiomeric excess ( ee ) in a 38.2% conversion. The hydrolase, which was identiﬁed by analyses of partial amino acid sequences of the puriﬁed enzyme and genome sequence of Arthrobacter sp. K5, exhibited moderate homology with amidohydrolases up to 67% (molinate hydrolase from Gulosibacter molinativorax ). The hydrolase gene was overexpressed in Rhodococcus erythropolis . The recombinant cells produced ( S )-2-MPI with 83.5% ee in a 48.4% conversion ( E = 26.3) from 100 mM rac - N -pivaloyl-2-MPI. These results suggest the possibility of an efﬁcient preparation of chiral 2-MPI in kinetic resolution.

In this study, we screened the stereoselective hydrolases acting on rac-N-acyl-2-MPI from soil microorganisms for one-step preparation of chiral 2-MPI (Scheme 2). We then purified and characterized the hydrolase of Arthrobacter sp. K5. To ensure efficient chiral piperidine preparation, we further constructed recombinant cells expressing the hydrolase gene and investigated the optically pure 2-MPI production via whole cells reaction.

Screening of Microorganisms
To obtain microorganisms exhibiting hydrolase activity toward N-acyl-2-MPI, we isolated 294 strains from soil samples using a medium containing N-acetyl-2-MPI as the

Screening of Microorganisms
To obtain microorganisms exhibiting hydrolase activity toward N-acyl-2-MPI, we isolated 294 strains from soil samples using a medium containing N-acetyl-2-MPI as the sole carbon source. Among 60 bacteria that hydrolyzed N-acetyl-2-MPI, the majority of them did not display high stereoselectivity. Only the strain K5 revealed moderate (S)selectivity kinetically, yielding racemic 2-MPI ( Figure 2a). We predicted that the acyl group affected the enantiomeric recognition of the hydrolase in strain K5. To improve the enantioselectivity of the hydrolysis, we substituted the acetyl group with a bulky pivaloyl group for appropriate chiral recognition. The reaction using whole cells of strain K5 exhibited good (S)-stereoselectivity toward N-pivaloyl-2-MPI to yield (S)-2-MPI with 22% conversion and 88% ee in 72 h (Figure 2b). The strain K5 was identified as Arthrobacter sp. based on 16S rDNA sequence analysis.

Optimization of Culture Conditions and Kinetic Resolution Using Optimized Whole Cells
To enhance the hydrolase activity in Arthrobacter sp. K5 cells, cyclic amine derivatives were added to the culture medium, and the hydrolase activity was examined. The addition of N-acetyl-2-MPI, N-acetylpiperidine, or N-acetyl-2-methylpyrrolidine resulted in a high induction of the hydrolase activity, with the best compound being N-acetylpiperidine, which also enhanced Arthrobacter sp. K5 growth (Table S1). The highest activity was found in the cells cultivated for one day with 0.4% (v/v) N-acetylpiperidine (Tables S2 and  S3). Prolonged cultivation (≥2 days) decreased enzyme activity. The culture conditions were optimized as follows: time, 24 h and culture medium, 0.4% (v/v) N-acetylpiperidine, 5 g L −1 of polypeptone, 5 g L −1 of meat extract, 2 g L −1 NaCl, and 0.5 g L −1 yeast extract in tap water (pH = 7). Using the cells cultivated under optimal conditions, we performed kinetic resolution of N-pivaloyl-2-MPI. In the reaction with a 100 mM concentration of Npivaloyl-2-MPI, Arthrobacter sp. K5 cells achieved 38.2% conversion to (S)-2-MPI at 80.2%

Optimization of Culture Conditions and Kinetic Resolution Using Optimized Whole Cells
To enhance the hydrolase activity in Arthrobacter sp. K5 cells, cyclic amine derivatives were added to the culture medium, and the hydrolase activity was examined. The addition of N-acetyl-2-MPI, N-acetylpiperidine, or N-acetyl-2-methylpyrrolidine resulted in a high induction of the hydrolase activity, with the best compound being N-acetylpiperidine, which also enhanced Arthrobacter sp. K5 growth (Table S1). The highest activity was found in the cells cultivated for one day with 0.4% (v/v) N-acetylpiperidine (Tables S2 and S3). Prolonged cultivation (≥2 days) decreased enzyme activity. The culture conditions were optimized as follows: time, 24 h and culture medium, 0.4% (v/v) N-acetylpiperidine, 5 g L −1 of polypeptone, 5 g L −1 of meat extract, 2 g L −1 NaCl, and 0.5 g L −1 yeast extract in tap water (pH = 7). Using the cells cultivated under optimal conditions, we performed kinetic resolution of N-pivaloyl-2-MPI. In the reaction with a 100 mM concentration of N-pivaloyl-2-MPI, Arthrobacter sp. K5 cells achieved 38.2% conversion to (S)-2-MPI at 80.2% ee (E = 14.8) after 115 h. However, the whole cells reaction stopped after 75 h, unable to achieve 50% conversion of the substrate. This result suggested the possibility of the enzyme inactivation caused by a long-time reaction or insufficient stability of the hydrolase.

Properties of the Hydrolase
To obtain a homogeneous enzyme, the hydrolase was purified from the cell-free extract of Arthrobacter sp. K5 through five steps, including ammonium sulfate fractionation, ion-exchange chromatography on DEAE-Sephacel, and hydrophobic interaction chromatography on phenyl-Sepharose and butyl-Toyopearl (Table S4). The purified enzyme was obtained with a specific activity of 35.5 µmol min −1 mg −1 . The overall purification was 10.8-fold (yield = 22%). The molecular mass of the hydrolase was estimated to be 50 kDa by SDS-PAGE and 238 kDa by gel permeation high performance liquid chromatography (HPLC), which suggested it to be a homo-tetrameric hydrolase. The optimum reaction temperature and pH of the hydrolase were 45 • C and pH = 8.0 (Tris-HCl), respectively ( Figures S2 and S3). The enzyme retained 75% of its maximum activity <40 • C; however, its activity decreased at a temperature >45 • C ( Figure S4). Conversely, it retained 80% of its maximum activity at pH = 6.0-7.5 and more than 80% inhibition of the hydrolase activity was observed at other pH values ( Figure S5). According to amino acid sequence analysis, the N-terminal and internal amino acid sequences of the hydrolase were obtained as ATQTVITNGTLIDGTGNQPQ and GGVTTVFDTWNA, respectively. Based on the amino acid sequence and genome DNA sequence of Arthrobacter sp. K5, we identified (S)-selective hydrolase (SHA). The enzyme gene was composed of 1384 bp, and coded for a protein of 481 amino acids with the molecular mass of 49,725 Da. This value is in agreement with molecular mass determined on SDS-PAGE. A BLAST search with full-length amino acid sequence of SHA revealed moderate sequence identity with the amidohydrolase protein family, including phenylurea hydrolases (<66%), and the highest sequence identity (67%) with the molinate hydrolase from Gulosibacter molinativorax. These results suggested SHA as a novel enzyme.

Substrate Specificity
We examined the substrate specificity of purified SHA using various N-acyl cyclic amines and comparing the activity toward them with the activity toward N-benzoyl-2-MPI ( Table 1). The hydrolase exhibited almost the same activity toward N-benzoyl-2-MPI and N-pivaloyl-2-MPI; however, the latter was more hydrolyzed with higher (S)-selectivity. N-Acetyl-2-MPI was hydrolyzed with 24.8-fold higher activity than Nbenzoyl-2-MPI. SHA displayed high activity on N-acyl 2-unsubstituted cyclic amines, such as N-benzoylpiperidine, N-benzoyl-3-MPI, N-pivaloyl-3-MPI, N-benzoyl-4-MPI, and N-benzoylpyrrolidine but showed no stereoselectivity toward N-acyl-3-MPI. N-benzoyl-2methylpyrrolidine displayed four-fold higher reactivity than N-benzoyl-2-MPI, whereas enantioselectivity was low in slight favor of the (R)-enantiomer (11% ee). N-benzoyl-2methylindoline was also a preferable substrate, reacting with medium enantioselectively (Sor R-enantiomers not determined); however, N-pivaloyl-2-methylindoline and N-benzoyl-1,2,3,4-tetrahydroquinaldine were not hydrolyzed. The reactivity and stereoselectivity of SHA for tested compounds depended on the acyl groups and the distance between the acyl group and chiral center. SHA exhibited no activity toward N-acetyl D-or L-amino acid (data not shown).

(S)-2-MPI Synthesis Using Recombinant Cells
Since the SHA gene sequence has GC-content, we overexpressed the gene in Rhodococcus erythropolis L88, which are high GC-content bacteria that can express high GC-content genes [34]. Rhodococcus sp. cells are robust and show resistance to various stress conditions, expecting tolerance to organic solvent, high concentration of substrates, and long-time reaction [35]. The reaction with 100 mM of N-pivaloyl-2-MPI using the recombinant cells reached 48.4% conversion to (S)-2-MPI with 83.5% ee (E = 26.3) in 74 h (Figure 3, closed circles). Compared with the reaction using Arthrobacter sp. K5 cells that almost stopped after 60 h with the conversion level of <40% (Figure 3, closed triangles), R. erythropolis transformant retained the hydrolase activity and the reaction proceeded to 50% conversion after 72 h. Compared to the two-step preparation using the Aspergillus sp. protease and acylation of the racemic cyclic amine by the lipase with 3-methoxyphenyl allyl carbonate (Scheme 1B,C), the kinetic resolution by Arthrobacter sp. K5 hydrolase achieved one-step preparation of (S)-2-MPI requiring only inexpensive reagents. However, optical purity of 2-MPI was decreased due to gradual hydrolysis of the (R)-enantiomer. For the kinetic resolution of chiral 2-MPI production, it is essential to transfer the higher stereoselectivity toward N-pivaloyl-2-MPI to SHA using protein engineering.

(S)-2-MPI Synthesis Using Recombinant Cells
Since the SHA gene sequence has GC-content, we overexpressed the gene in Rhodococcus erythropolis L88, which are high GC-content bacteria that can express high GC-content genes [34]. Rhodococcus sp. cells are robust and show resistance to various stress conditions, expecting tolerance to organic solvent, high concentration of substrates, and longtime reaction [35]. The reaction with 100 mM of N-pivaloyl-2-MPI using the recombinant cells reached 48.4% conversion to (S)-2-MPI with 83.5% ee (E = 26.3) in 74 h (Figure 3, closed circles). Compared with the reaction using Arthrobacter sp. K5 cells that almost stopped after 60 h with the conversion level of <40% (Figure 3, closed triangles), R. erythropolis transformant retained the hydrolase activity and the reaction proceeded to 50% conversion after 72 h. Compared to the two-step preparation using the Aspergillus sp. protease and acylation of the racemic cyclic amine by the lipase with 3-methoxyphenyl allyl carbonate (Scheme 1B,C), the kinetic resolution by Arthrobacter sp. K5 hydrolase achieved one-step preparation of (S)-2-MPI requiring only inexpensive reagents. However, optical purity of 2-MPI was decreased due to gradual hydrolysis of the (R)-enantiomer. For the kinetic resolution of chiral 2-MPI production, it is essential to transfer the higher stereoselectivity toward N-pivaloyl-2-MPI to SHA using protein engineering.

General Information
Commercially available reagents were used without purification and purchased

N-Acyl Cyclic Amines Synthesis
Acyl chloride (50 mmol) was added to a solution of cyclic amine (50 mmol) and pyridine (50 mmol) in dichloromethane (200 mL) at 0 • C. The reaction was performed overnight at room temperature with stirring. The reaction mixture was concentrated under reduced pressure. Ethyl acetate was added to the residue, and pyridine hydrochloride was removed by filtration. The supernatant was concentrated under reduced pressure and purified by column chromatography (silica gel, n-hexane:ethyl acetate = 6:1) to obtain N-acyl cyclic amines in moderate to good yield.

Hydrolysis of N-Pivaloyl-2-Methylpiperidine Using Whole Cells of Arthrobacter sp. K5
The reaction was performed at 30 • C with shaking (120 rpm) in 25 mL of 100 of mM potassium phosphate buffer (pH = 7.0) containing 100 mM of N-pivaloyl-2-MPI and whole cells derived from 125 mL of culture broth. Samples were collected multiple times and analyzed by HPLC after derivatization of samples with GITC.
3.6. Genome Sequence of Arthrobacter sp. K5 Arthrobacter sp. K5 cells were lysed at room temperature overnight with 0.5 mg mL −1 of achromopeptidase (crude) in 10 mM of Tris-HCl (pH = 8.0). To this solution, 0.02 mg mL −1 of proteinase K, 10 mM of CaCl 2 , and 0.5% (w/v) SDS were added, and the mixture was incubated overnight at 37 • C. To an equal volume of the lysed cells, 2×CTAB solutions were added and incubated at 60 • C for 1 h. The 2×CTAB solution contained 20 g L −1 of cetyltrimethylammonium bromide, 50 mM of Tris-HCl, 20 mM of EDTA, 111 g L −1 of NaCl, and 10 g L −1 of polyvinylpyrrolidone in distilled H 2 O. To an equal volume of the treatment solution, a mixture of phenol, chloroform, and isoamyl alcohol (25:24:1, v/v/v) was added, mixed gently, and centrifuged at 4 • C and 8000 rpm for 30 min. The supernatant was washed with chloroform and one-tenth the volume of 3 M of sodium acetate (pH = 5.2) was added and mixed with isopropanol until genome DNA was thoroughly precipitated. The genome DNA obtained was washed twice with 70% (v/v) ethanol and diluted with distilled H 2 O. Genome analysis was commissioned to Gifu University NGS service.

Hydrolase Overexpression in Rhodococcus Erythropolis L88
The SHA gene was identified based on the partial amino acid sequence of SHA and genome sequence analysis of Arthrobacter sp. K5. The gene sequence was deposited in the DDBJ database under the accession number LC633519. The gene amplification was performed via PCR using the primers 5 -CTATCCATGGCAACGCAGACAGTG-3 and 5 -TAATCTCGAGTCAGACGTTGTCGTCGAGG-3 with PrimeSTAR ® Max DNA Polymerase (Takara Bio). The amplified DNA fragments and pTipQC1 vector (Hokkaido System Science, Hokkaido, Japan) were digested with Nco I and Xho I and ligated using a DNA Ligation Kit (Takara Bio). The resulting plasmid was transformed into Rhodococcus erythropolis L88 cells (Hokkaido System Science) by electroporation with Eporator ® (Eppendorf, Hamburg, Germany). The transformed cells were cultivated with 30 µg mL −1 of chloramphenicol at 120 rpm and 28 • C in 5 mL of the nutrient medium containing 10 g L −1 of tryptone, 5 g L −1 of yeast extract, 4 g L −1 of Na 2 HPO 4 , 1 g L −1 of KH 2 PO 4 , 1 g L −1 of NaCl, 0.2 g L −1 of MgSO 4 ·7H 2 O, and 0.01 g L −1 of CaCl 2 ·7H 2 O in tap water. The preculture was inoculated into 90 mL of the nutrient medium and 0.2 µg ml −1 of thiostrepton was added and incubated at 20 • C and 120 rpm for 24 h. The cells were harvested by centrifugation, washed twice with 0.85% (w/v) NaCl, and suspended in the same solution.

(S)-2-MPI Synthesis Using Recombinant Cells
The reaction was performed at 30 • C with shaking (120 rpm) in 2 mL of 100 mM of potassium phosphate buffer (pH = 7.0) containing 37 mg (100 mM) of N-pivaloyl-2-MPI and whole cells obtained from 10 mL of culture broth. Samples were collected multiple times and analyzed by HPLC after derivatization with GITC.

Conclusions
SHA exhibited high (S)-selectivity toward N-pivaloyl-2-MPI to produce (S)-2-MPI with 80.2% ee. We successfully overexpressed the SHA gene in R. erythropolis L88 and improved the hydrolase activity. The biocatalytic process achieved a kinetic resolution of 100 mM of N-pivaloyl-2-MPI in one step using inexpensive synthetic substrates, forming (S)-2-MPI with 48.4% conversion and 83.5% ee. As a potential enzyme for practical applications, SHA may allow highly enantioenriched (S)-MPI production by improving its stereoselectivity in the future.