Identification of Subfamily Specific Residues within Highly Active and Promiscuous Alcohol Dehydrogenases

Enzyme selection is an essential process in the biobased production of chemicals. It is essential to develop a method to extract yet unknown useful enzymes from protein databases. Enzymes that exhibit substrate promiscuity and high activity hold the potential to access unknown reactions and mediate known reactions with a higher performance. Herein, we propose and validate a principal component analysis (PCA)-based classification method, termed MUSASHI (MUltiple-Sequence Alignment-based protein Selection via clustering using HIgh-dimensional analysis), to identify subfamily-specific residues that are highly conserved among promiscuous alcohol dehydrogenase (ADH). Specifically, zinc-dependent ADH homologues retrieved from the protein database were classified into 9 groups, and according to PCA-based clustering, the activities of 18 ADHs, with representative enzymes from each group, were characterized. As a result, we identified two promiscuous ADH groups: Group 1 ADH, efficient with short-chain and aromatic aldehydes, and Group 3 ADH, efficient with aliphatic and aromatic ketones. Sequence feature analysis then revealed subfamily-specific residues, which are highly conserved only in promiscuous ADH Groups 1 and 3, with the potential to biosynthesize a wide spectrum of target compounds. Tatumella ptyseos ADH, identified from Group 1 of this study, showed higher isobutanol and 2-phenylethanol bioconversions than that of a conventional ADH (Ahr). These results indicate that the MUSASHI method for subfamily-specific residue identification can enable optimal enzyme selection from protein databases.