ω-transaminase from Ochrobactrum anthropi Is Devoid of Substrate and Product Inhibitions

AEM Accepts, published online ahead of print on 12 April 2013 Appl. Environ. Microbiol. doi:10.1128/aem.03811-12 Copyright 2013, American Society for Microbiology. All Rights Reserved. (Enzymology and Protein Engineering) ω-transaminase from Ochrobactrum anthropi Is Devoid of Substrate and Product Inhibitions Eul-Soo Park and Jong-Shik Shin* Department of Biotechnology, Yonsei University, Seoul 120-749, South Korea *To whom correspondence should be addressed. Engineering Building 2 (Rm 507), Yonsei University, Shinchon-Dong 134, Seodaemun-Gu, Seoul 120-749, South Korea (E-mail: enzymo@yonsei.ac.kr, Tel: (+82) 2-2123-5884, Fax: (+82) 2-362-7265) Running Title: Inhibition-free ω-transaminase 1

Abstract: ω-transaminases display complicated inhibitions by ketone products and both enantiomers of amine substrates. Here, we report the first example of ω-transaminase devoid of such inhibitions. Owing to the lack of enzyme inhibitions, the ω-transaminase from Ochrobactrum anthropi enabled efficient kinetic resolution of α-methylbenzylamine (500 mm) even without product removal. ω-transaminase (ω-ta) catalyzes reversible transfer of an amino group between primary amines and carbonyl compounds, which is mediated by pyridoxal 5 -phosphate (PLP) bound to the enzyme as a prosthetic group (6, 7). During a last decade, ω-ta has gained increasing attention for chiral amine production owing to a unique enzyme property enabling both oxidative deamination of an amine involving conversion of enzyme-bound PLP to pyridoxamine 5 -phosphate (PMP) and reductive amination of a carbonyl compound accompanied by regeneration of the PLP form of enzyme (E-PLP) from the PMP form of enzyme (E-PMP) as shown in Fig. 1 (2, 9, 11, 12). ω-tas exhibit broad substrate specificity, high turnover rate, stringent stereoselectivity, high enzyme stability and no requirement for an external cofactor, which renders the enzyme suitable for industrial process development (7, 21). In the biocatalytic process design, complicated enzyme properties often limit efficient process operation (10, 23). Despite the enzyme properties of ω-tas beneficial for industrial applications, severe product inhibition causing drastic reductions in the enzyme activity has been a major challenge to exploit the ω-ta reactions for industrial processes (2, 9, 11, 18-20). For example, ω-ta from Bacillus thuringiensis JS64 (BT ω-ta) showed only 5 % residual 2

activity at 20 mm acetophenone in transamination between α-methylbenzylamine (α-mba) and pyruvate (18). Therefore, several reaction engineering approaches were developed to alleviate the inhibition by removing the ketone product using solvent extraction (18, 20). The product inhibition has been regarded as a general property of ω-tas and there have been no reports on the discovery of a ω-ta lacking the product inhibition (2, 9, 11, 12, 16). Yun et al. reported that product inhibition of ω-ta from Vibrio fluvialis JS17 (VF ω-ta) by aliphatic ketones could be attenuated by directed evolution (24). However, it remains challenging to completely eliminate the product inhibition behaviors although such an inhibition-free enzyme is highly demanded for cost-effective processes. In the previous studies, we cloned, overexpressed and purified (S)-selective ω-tas from Ochrobactrum anthropi (OA ω-ta)(13) and Paracoccus denitrificans PD1222 (PD ω-ta)(14). Briefly, Escherichia coli BL21(DE3) cells carrying a pet28a(+) expression vector harboring the OA ω-ta or PD ω-ta gene were cultivated in LB medium (typically 1 L) and the Histagged ω-ta was purified using a HisTrap HP column (GE Healthcare) followed by a HiTrap Desalting column (GE Healthcare) as described elsewhere (13, 14). When necessary, the purified enzyme solution was concentrated by an ultrafiltration kit (Ultracel-30) from Millipore Co. (Billerica, MA). Substrate specificities of the two ω-tas were remarkably similar (13, 14), leading us to presume that both enzymes were prone to product inhibition. As previously observed with VF ω-ta (16), PD ω-ta exhibited strong product inhibition (only 10 % residual activity at 20 mm acetophenone) (Fig. 2). Assuming that the product inhibition followed a hyperbolic decay (see detailed procedures in the Supplemental Material), product inhibition constant (K PI ) of acetophenone in the transamination between (S)-α-MBA and pyruvate (both 3

20 mm) was determined to be 2.4 ± 0.3 mm. However, to our surprise, OA ω-ta did not show such inhibition at all up to 20 mm acetophenone. These contrasting inhibition properties of the two ω-tas were not affected by the choice of an amino acceptor (see entry 1 in Table S1 where 2-oxobutyrate was used instead of pyruvate). Based on the kinetic model we built previously to describe kinetic properties of BT and VF ω-tas (17, 19), the product inhibition results from strong binding of acetophenone to the active site of E-PMP that hinders an amino acceptor from binding to E-PMP. Consistent with this, acetophenone strongly inhibited PD ω-ta even in the reaction between L-alanine and 2-oxobutyrate where acetophenone was not a cognate product (entry 2 in Table S1). In contrast, OA ω-ta did not show such inhibition, suggesting that the lack of product inhibition observed with OA ω-ta results from binding affinity of E-PMP toward acetophenone much lower than that toward pyruvate and 2-oxobutyrate. We further examined whether OA ω-ta was also devoid of product inhibitions by other ketones, e.g. two arylaliphatic ketones (propiophenone and 1-indanone) and two aliphatic ketones (2-butanone and cyclopropyl methyl ketone). None of the reactions between chiral amines (α-ethylbenzylamine, 1-aminoindan, sec-butylamine and cyclopropylethylamine) and pyruvate were inhibited by the ketone product (entries 3-6 in Table S1), indicating that the lack of the product inhibition is a general property of OA ω-ta. In contrast, PD ω-ta showed strong product inhibitions by all those ketones (i.e. K PI values ranging from 4.3 to 28.5 mm; Table S1). Besides the product inhibition, substrate inhibitions by both enantiomers of chiral amines, as previously observed with VF ω-ta, render kinetic properties of ω-tas much more complicated (19). Because OA and PD ω-tas displayed contrasting product inhibitions, we 4

examined whether inhibition properties of the two ω-tas were different toward amine substrates. Indeed, the ω-tas showed completely different substrate inhibitions by α-mba (Fig. 3). PD ω-ta showed significant substrate inhibition by (S)-α-MBA (i.e. a reacting enantiomer) above 100 mm, whereas OA ω-ta was not inhibited at all up to 500 mm (Fig. 3A). Moreover, in contrast to strong inhibition of PD ω-ta by (R)-α-MBA in the transamination between (S)-α-MBA and pyruvate (both 20 mm), OA ω-ta is devoid of such inhibition by the non-reacting enantiomer of α-mba (Fig. 3B). The strong substrate inhibition of PD ω-ta was also observed with achiral amine substrates such as benzylamine (Fig. S1). From the initial rate measurements shown in Fig. 3, we determined kinetic parameters of the two enzymes (Table 1). For the kinetic analysis, we used a pseudo-one-substrate model as previously described (15). In the case of PD ω-ta, initial rate data obtained in the absence of the substrate inhibitions were used to determine the Michaelis constant (K M ) and the maximum reaction rate (V max ). Similarly to the product inhibitions, substrate inhibitions (K SI ) were assumed to follow hyperbolic decay (see detailed procedures in the Supplemental Material). The K (S)-α-MBA M of PD ω-ta (i.e. 31 mm) was much lower than that of OA ω-ta (i.e. 126 mm), indicating that the active site of OA ω-ta forms a relatively weak Michaelis complex between the E-PLP form and (S)-α-MBA. In the previous study, the K M values of four reactive (S)- arylalkylamines measured with VF ω-ta were between 0.29 and 33 (19). Therefore, the K M α-mba of OA ω-ta is unusually larger than typical K M values of reactive arylalkylamines. We previously proposed that substrate inhibitions by (S)- and (R)-enantiomer of chiral amines resulted from non-productive binding to E-PMP and E-PLP, respectively (19). It is likely that such a low binding affinity of the E-PLP form of OA ω-ta toward (S)-α-MBA leads to 5

negligible binding of (S)- and (R)-enantiomer of α-mba to E-PMP and E-PLP, respectively, which may explain the lack of the substrate inhibitions of OA ω-ta. Similarly, the lack of product inhibition by acetophenone can be explained by the high K (S)-α-MBA M because the same active site residues are involved in the Michaelis complex formation between E-PMP and acetophenone. Despite the weaker binding affinity toward (S)-α-MBA, catalytic turnover by OA ω-ta was found to be even 4-fold faster than that by PD ω-ta (Table 1). The dual substrate inhibitions by both enantiomers of chiral amines are highly detrimental to kinetic resolution of racemic amines where use of high concentrations of racemic amine substrates is preferred (4, 5, 22). The lack of enzyme inhibition by amine substrates turned out to be a general property of OA ω-ta because none of the racemic amines tested (i.e. α-mba, α-ethylbenzylamine, p-fluoro-α-mba, 1-methyl-3-phenylpropylamine and sec-butylamine) caused substrate inhibition (Fig. S2A). In contrast, PD ω-ta showed strong substrate inhibitions by all these amines (Fig. S2B). Compared with more than twenty ω-tas identified so far (9), OA ω-ta displayed an unprecedented property devoid of enzyme inhibitions by ketone products as well as amine substrates. To assess how beneficial this unique property is to chiral amine production, we carried out kinetic resolution of 500 mm α-mba at 300 mm pyruvate and 75 U/mL ω-ta (Fig. 4). When using OA ω-ta, enantiomeric excess (ee) of the resulting (R)-α-MBA reached 95.3 % at 3 h and exceeded 99.9 % (i.e. absolutely enantiopure) at 10 h even without acetophenone removal. In contrast, kinetic resolution using PD ω-ta did not show any substantial reaction progress after 1 h due to strong enzyme inhibition by produced 6

acetophenone and led to only 22 % ee at 10 h, which illustrates why ketone product removal was indispensable to the previous studies using inhibition-susceptible ω-tas (18, 20). This example clearly indicates that lack of the inhibition behaviors renders OA ω-ta ideal for kinetic resolution of chiral amines. In addition, we expect that the inhibition-free OA ω-ta may benefit asymmetric synthesis of chiral amines from prochiral ketones (1, 3, 8). Supplemental Material Materials and Methods, Supplemental Table S1 and Supplemental Figures S1 S2. Acknowledgements This work was supported by the Advanced Biomass R&D Center (ABC-2010-0029737) and the Basic Science Research Program (2010-0024448) through the National Research Foundation of Korea funded by the Ministry of Education, Science and Technology. We are grateful to the technical assistance of Dr. Chao Li in the kinetic resolution using OA ω-ta. References 1. Fuchs, M., D. Koszelewski, K. Tauber, W. Kroutil, and K. Faber. 2010. Chemoenzymatic asymmetric total synthesis of (S)-Rivastigmine using ω-transaminases. Chem. Commun. 46:5500-5502. 7

2. Höhne, M., and U. T. Bornscheuer. 2009. Biocatalytic Routes to Optically Active Amines. ChemCatChem 1:42-51. 3. Höhne, M., S. Kühl, K. Robins, and U. T. Bornscheuer. 2008. Efficient asymmetric synthesis of Chiral amines by combining transaminase and pyruvate decarboxylase. ChemBioChem 9:363-365. 4. Höhne, M., K. Robins, and U. T. Bornscheuer. 2008. A protection strategy substantially enhances rate and enantioselectivity in ω-transaminase-catalyzed kinetic resolutions. Adv. Synth. Catal. 350:807-812. 5. Hanson, R. L., B. L. Davis, Y. Chen, S. L. Goldberg, W. L. Parker, T. P. Tully, M. A. Montana, and R. N. Patel. 2008. Preparation of (R)-amines from racemic amines with an (S)-amine transaminase from Bacillus megaterium. Adv. Synth. Catal. 350:1367-1375. 6. Hirotsu, K., M. Goto, A. Okamoto, and I. Miyahara. 2005. Dual substrate recognition of aminotransferases. Chemical Record 5:160-172. 7. Hwang, B. Y., B. K. Cho, H. Yun, K. Koteshwar, and B. G. Kim. 2005. Revisit of aminotransferase in the genomic era and its application to biocatalysis. J. Mol. Catal. B: Enzym. 37:47-55. 8. Koszelewski, D., I. Lavandera, D. Clay, G. M. Guebitz, D. Rozzell, and W. Kroutil. 2008. Formal asymmetric biocatalytic reductive amination. Angew. Chemie Int. Ed. 47:9337-9340. 9. Koszelewski, D., K. Tauber, K. Faber, and W. Kroutil. 2010. ω-transaminases for the synthesis of non-racemic α-chiral primary amines. Trends Biotechnol. 28:324-332. 10. Liese, A., K. Seelbach, and C. Wandrey (ed.). 2006. Industrial Biotransformation, Second ed. Wiley-VCH Verlag GmbH & Co. KGaA. 8

11. Malik, M. S., E. S. Park, and J. S. Shin. 2012. Features and technical applications of ω-transaminases. Appl. Microbiol. Biot. 94:1163-1171. 12. Mathew, S., and H. Yun. 2012. ω-transaminases for the Production of Optically Pure Amines and Unnatural Amino Acids. ACS Catalysis:993-1001. 13. Park, E.-S., M. Kim, and J.-S. Shin. 2012. Molecular determinants for substrate selectivity of ω-transaminases. Appl. Microbiol. Biotechnol. 93:2425-2435. 14. Park, E., M. Kim, and J. S. Shin. 2010. One-pot conversion of L -threonine into L- homoalanine: Biocatalytic production of an unnatural amino acid from a natural one. Adv. Synth. Catal. 352:3391-3398. 15. Park, E. S., and J. S. Shin. 2011. Free energy analysis of ω-transaminase reactions to dissect how the enzyme controls the substrate selectivity. Enzyme Microb. Technol. 49:380-387. 16. Shin, J. S., and B. G. Kim. 2001. Comparison of the ω-transaminases from different microorganisms and application to production of chiral amines. Biosci. Biotech. Biochem. 65:1782-1788. 17. Shin, J. S., and B. G. Kim. 1998. Kinetic modeling of ω-transamination for enzymatic kinetic resolution of α-methylbenzylamine. Biotechnol. Bioeng. 60:534-540. 18. Shin, J. S., and B. G. Kim. 1997. Kinetic resolution of α-methylbenzylamine with ω- transaminase screened from soil microorganisms: Application of a biphasic system to overcome product inhibition. Biotechnol. Bioeng. 55:348-358. 19. Shin, J. S., and B. G. Kim. 2002. Substrate inhibition mode of ω-transaminase from Vibrio fluvialis JS17 is dependent on the chirality of substrate. Biotechnol. Bioeng. 77:832-837. 9

20. Shin, J. S., B. G. Kim, A. Liese, and C. Wandrey. 2001. Kinetic resolution of chiral amines with ω-transaminase using an enzyme-membrane reactor. Biotechnol. Bioeng. 73:179-187. 21. Taylor, P. P., D. P. Pantaleone, R. F. Senkpeil, and I. G. Fotheringham. 1998. Novel biosynthetic approaches to the production of unnatural amino acids using transaminases. Trends Biotechnol. 16:412-418. 22. Truppo, M. D., N. J. Turner, and J. D. Rozzell. 2009. Efficient kinetic resolution of racemic amines using a transaminase in combination with an amino acid oxidase. Chem. Commun. 28:2127-2129. 23. Wenda, S., S. Illner, A. Mell, and U. Kragl. 2011. Industrial biotechnology-the future of green chemistry? Green Chemistry 13: 3007-3047. 24. Yun, H., B. Y. Hwang, J. H. Lee, and B. G. Kim. 2005. Use of enrichment culture for directed evolution of the Vibrio fluvialis JS17 ω-transaminase, which is resistant to product inhibition by aliphatic ketones. Appl. Environ. Microb. 71:4220-4224. 10

Table 1. Kinetic parameters a of PD and OA ω-tas PD ω-ta OA ω-ta V max (mm/min/(u/ml)) b 2.2 ± 0.1 9.4 ± 2.7 K (S)-α-MBA M (mm) 31 ± 3 126 ± 33 K (S)-α-MBA SI (mm) 294 ± 13 n.a. c K SI (R)-α-MBA (mm) 39 ± 6 n.a. c a Kinetic parameters represent the apparent rate constants determined at a fixed concentration of pyruvate (20 mm). b V max represents a maximum reaction rate normalized by an enzyme concentration. c not applicable: enzyme inhibition was not observed. 11

Figure legends. Fig.1 Reaction scheme of ω-ta-catalyzed transamination between an amino donor (shown in a box) and an amino acceptor (in a circle). Fig.2 Product inhibition of OA and PD ω-tas by acetophenone in transamination between (S)-α-MBA and pyruvate. Fig. 3 Substrate inhibitions of OA and PD ω-tas by α-mba. (A) Enzyme inhibition at high concentrations of (S)-α-MBA in the reaction with pyruvate. Enzyme concentrations were 0.05 U/mL. (B) Enzyme inhibition by (R)-α-MBA in the reaction between (S)-α- MBA and pyruvate. Fig. 4 Comparison of the kinetic resolutions of α-mba using OA and PD ω-tas. 12

NH 2 oxidative deamination O R 1 R 2 R 1 R 2 NH 2 E-PLP E-PMP O R 3 R 4 reductive amination R 3 R 4 Fig. 1 13

1.2 1.0 relative initial rate 0.8 0.6 0.4 0.2 OA ω-ta PD ω-ta 0.0 0 5 10 15 20 acetophenone (mm) Fig. 2 14

A 0.30 initial rate (mm/min) 0.25 0.20 0.15 0.10 0.05 OA ω-ta PD ω-ta B 0.00 0 100 200 300 400 500 1.2 (S)-α-MBA (mm) 1.0 relative initial rate 0.8 0.6 0.4 0.2 OA ω-ta PD ω-ta 0.0 0 20 40 60 80 100 (R)-α-MBA (mm) Fig. 3 15

100 ee of (R)-α-MBA (%) 80 60 40 20 OA ω-ta PD ω-ta 0 0 2 4 6 8 10 Reaction time (h) Fig. 4 16