Engineering the Active Site of the Amine Transaminase from Vibrio fluvialis for the Asymmetric Synthesis of Aryl– Alkyl Amines and Amino Alcohols Alberto Nobili, [a] Fabian Steffen-Munsberg, [a, b] Hannes Kohls, [a] Ivan Trentin, [a] Carola Schulzke, [a] Matthias Hçhne, [a] and Uwe T. Bornscheuer* [a] Although the amine transaminase from Vibrio fluvialis has often been applied as a catalyst for the biocatalytic prepara- tion of various chiral primary amines, it is not suitable for the transamination of a-hydroxy ketones and aryl-alkyl ketones bearing an alkyl substituent larger than a methyl group. We addressed this problem through a systematic mutagenesis study of active site residues to expand its substrate scope to- wards two bulky ketones. We identified two mutants (F85L/ V153A and Y150F/V153A) showing 30-fold increased activity in the conversion of (S)-phenylbutylamine and (R)-phenylglycinol, respectively. Notably, they facilitated asymmetric synthesis of these amines with excellent enantiomeric purities of 98 % ee. Enantiomerically pure amines and amino alcohols play a funda- mental role in the pharmaceutical industry. One in four of the 200 top-sold drugs contains a chiral amine moiety and these drugs had a total market value of more than 88 billion USD in 2013 according to Weber and Sedelmeier. [1] When it comes to the choice of the synthetic strategy for the preparation of the amine building blocks, amine transaminases (ATAs) are increas- ingly recognized as an attractive option as they facilitate a one-step asymmetric synthesis starting from the correspond- ing prochiral ketone. [2] A very impressive example is the appli- cation of an engineered (R)-selective ATA from Arthrobacter sp. (ATA117-mut), which is currently being used for the production of sitagliptin, the active ingredient of the drugs Januvia and Ja- numet. [3] This example demonstrates the importance of protein engineering of wild-type amine transaminases to expand their limited substrate scope. Known wild-type ATAs are not able to convert bulky compounds demanded by the pharmaceutical industry. Compared to the success story of engineered (R)-se- lective transaminases with relaxed substrate specificity, (S)-se- lective ATAs that convert a range of bulky ketones with similar efficiency as the engineered ATA117-mut are still not available, despite the progress of first engineering studies. [4] The crystal structures of several (S)-selective ATAs were solved recently, en- abling a detailed understanding of the mechanism of substrate binding. [5] Both (R)- and (S)-selective ATAs that were found in nature possess a large and a small pocket in their active sites (Fig- ure 1 a). [5a, 6] Although the large pocket can accommodate sub- stituents with a rather broad size distribution, such as small alkyl to naphthyl groups, the small pocket creates a strict steric constraint: if the size of the small substituent exceeds that of a methyl group, activity drops significantly. [6] For instance, ke- tones with a hydroxymethyl group as small substituent are hardly accepted. [7] This active site architecture limits the sub- strate scope, but at the same time contributes to the usually high enantioselectivity of these ATAs. Midelfort et al. [4a] and Park et al. [4b] recently reported the first attempts of rational engineering: they identified key residues via bioinformatic methods or structural inspection and investi- gated up to two substitutions per position by site-directed mutagenesis to achieve the transamination of their bulky target ketones. By combining eight mutations in Vibrio fluvialis ATA, a b-keto ester bearing a long (6 carbon) alkyl chain could be converted employing 1-phenylethylamine 1b as amino donor, affording the amine imagabalin at 28% yield via asym- metric synthesis. A single mutant in Paracoccus denitrificans ATA [4b] showed increased activity in the deamination of 1-alkyl substituted benzyl amines and the amination of 2-oxo-octa- noate. Interestingly, this study showed that larger n-alkyl sub- stituents are accepted in the small binding pocket if the sub- strate bears an a-carboxylate functional group instead of a large hydrophobic substituent such as a phenyl group. Despite these first successes, further efforts are needed to create an (S)-selective ATA that is useful for asymmetric synthe- sis of bulky amines. In the present study, we systematically ad- dress this problem by a (partial) saturation mutagenesis of all amino acids that form the small binding pocket of the ATA of Vibrio fluvialis. We employed 1-phenylbutane-1-one 2a and the hydroxy ketone 2-hydroxyacetophenone 3a as model substrates (Table 1). The amine product (R)-phenylglycinol 3b is a building block for many important pharmaceuticals, such as an inhibitor of the 3-phosphoinositide-dependent protein kinase-1 (PDK1), which was identified as a target enzyme for cancer therapy. [8] Additionally, 3b is applied as a chiral auxiliary in the synthesis of some of the top selling drugs, saxagliptin [9] (treatment of type 2 diabetes), femoxetine and paroxetine [10] (antidepres- [a] A. Nobili, F. Steffen-Munsberg, H. Kohls, I. Trentin, Prof. Dr. C. Schulzke, Prof. Dr. M. Hçhne, Prof. Dr. U. T. Bornscheuer Institute of Biochemistry, University of Greifswald Felix-Hausdorff Str. 4, 17487 Greifswald (Germany) Fax: (+ 49)3834-86-794367 E-mail : uwe.bornscheuer@uni-greifswald.de [b] F. Steffen-Munsberg KTH Royal Institute of Technology, School of Biotechnology Division of Industrial Biotechnology AlbaNova University Center, SE-106 91 Stockholm (Sweden) Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cctc.201403010. ChemCatChem 0000, 00,0–0  0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 1 & These are not the final page numbers! ÞÞ These are not the final page numbers! ÞÞ Communications DOI: 10.1002/cctc.201403010