Active Site Engineering of co-Transaminase Guided by Docking Orientation Analysis and Virtual Activity Screening

Creation of enzyme variants displaying desirable catalytic performance usually necessitates tedious and time-consuming procedures for library generation and selection, which may be circumvented by a computational method based on a precise understanding of the reaction mechanism in the context of active site environment. Despite the great potential of omega-transaminases (omega-TAs) for asymmetric synthesis of chiral amines from ketones, it remains elusive why omega-TAs exhibit marginal activities for most ketones in contrast to their high activities for alpha-keto acids and aldehydes. To address this puzzling question, crystal structure determination and molecular modeling of co-TAs were carried out to analyze docking orientations of the amino acceptors in the Michaelis complex. We found that ketones, unlike the reactive substrates, led to nonproductive binding complexes where the bound substrate was hardly accessible to a nucleophilic attack by the pyridoxamine cofactor to initiate reductive amination of the amino acceptor. This finding led us to perform in silico mutagenesis of the S-selective omega-TA from Ochrobactrum anthropi to ameliorate the unfavorable nucleophilic attack trajectory to structurally demanding ketones. The resulting variant, carrying L57A/W58A mutations, was predicted to allow an unprecedented re-face attack on butyrophenone, leading to 10s-fold activity improvement with no loss in stereoselectivity. This study is expected to provide an efficient computational strategy for creation of high-turnover omega-TA variants tailored for a target ketone by affording in silico assessment of the effect of active site mutation on an enzyme activity.