Efficient synthesis of bepotastine and cloperastine intermediates using engineered alcohol dehydrogenase with a hydrophobic pocket.

(S)-4-Chlorophenylpyridylmethanol and (R)-4-chlorobenzhydrol are key pharmaceutical intermediates for the synthesis of bepotastine and cloperastine, respectively. However, the biocatalytic approach to prepare these bulky diaryl ketones remains challenging because of the low activity of naturally occurring alcohol dehydrogenases (ADH). In the present study, ADH seq5, which has an adequate binding pocket volume and accepts bulky diaryl ketones, was further engineered with a binding pocket of increased hydrophobicity. Based on molecular simulation and binding free energy analyses, a small mutation library was constructed, and mutant seq5-D150I with a threefold increase in k and a low K was obtained successfully. The comparison of kinetic parameters, binding free energy, docking conformation, and critical catalytic distances calculated by molecular dynamic simulations revealed the source of increased activity. To develop a practical approach with seq5-D150I, reaction conditions including pH, temperature, buffer, and metal ions were optimised and applied to synthesise (S)-4-chlorophenylpyridylmethanol and (R)-4-chlorobenzhydrol with high enantiomeric excess. The space-time yields for (S)-4-chlorophenylpyridylmethanol and (R)-4-chlorobenzhydrol increased dramatically to as high as 263.4 g∙L day and 150 g∙L day, respectively, which, to our knowledge, is the highest reported yield to date. These results show that the biocatalytic approach with seq5-D150I may be practical for future industrial applications.Key points An alcohol dehydrogenase was engineered based on binding free energy analysis. The mutant seq5-D150I obtained a threefold increase in k and a low K. Two important pharmaceutical intermediates were obtained with high space-time yield.