Gene Mining,Design And Engineering Of Stereocomplementary Ketoreductases

With excellent selectivity and environment compatibility,ketoreductases have gained extensive attention and have been used commonly in organic synthesis and green pharmaceutical in recent years.Natural ketoreductases/alcohol dehydrogenase usually display Prelog preference,which would produce Prelog products.However,both Prelog alcohols and anti-Prelog alcohols could be used as building blocks in pharmaceutical and fine chemical synthesis,thus,ketoreductases following Prelog's or anti-Prelog's rules are equally desired.In this paper,ketoreductase 1 obtained from Candida glabrata(CgKR1)catalyze the reduction of various keto esters and substituted acetophenones following Prelog's rule to the corresponding chiral alcohols,such as(R)-o-chloromandelate(R-CMM).To reverse its stereoselectivity,we first solved the crystal structure of apo-state CgKR1.And then,Phe92 and Tyr208 were selected for engineering based on the rational design,which was guided by structure and molecular docking.To minimize effort in mutant library screening,point mutagenesis to 6 smaller residues(Ala,Val,Ile,Ser,Thr,and Leu)were constructed for Phe92 and Tyr208 of CgKR1.Fortunately,five variants with higher stereoselectivity were obtained,among which,three variants catalyzed the asymmetric reduction of a-halogen-substituted acetophenones following Anti-Prelog's rule,and the remaining two variants catalyzed the asymmetric reduction of 2-hydroxy-1-phenylethanone and propiophenone following Prelog's rule.Furthermore,the whole cell biocatalytic processes for reduction of a-bromoacetophenone with high substrate loading were achieved by coexpression of CgKR1-F92V-Y208I and BsGDH for NADPH regeneration,the space-time conversion up to 600 g L-1 day-1.Although the variants of CgKR1 with Anti-Prelog preference were gained,it is difficult to transfer the knowledge from engineered CgKR1 to the others.With the challenge in mind,we cloned several short-chain dehydrogenases/reductases from bacterias by gene mining.Through homology modeling,molecular docking,carefully analyses the structural information and multiple-sequence alignment of homology SDRs,the key residues that might control their stereopreferences were identified using LfSDR1 as the starting enzyme.Protein engineering at these positions of LfSDR1 could improve its anti-Prelog stereoselectivity or switch its stereopreference to Prelog,stereocomplementary variants were got.Moreover,the knowledge obtained from LfSDR1 could be further transferred to other five SDRs that have 21-48%sequence identities with LfSDR1.The stereopreferences of these SDRs were able to be switched either from anti-Prelog to Prelog or from Prelog to anti-Prelog by mutagenesis at related positions.