Enantioselective Synthesis of 1-Aryl-Substituted Tetrahydroisoquinolines Employing Imine Reductase

Tetrahydroisoquinolines (THIQs) with a C1-aryl-substituted groups are common in many natural and synthetic compounds of biological importance. Currently, their enantioselective synthesis are primarily reliant on chemical catalysis. Enzymatic synthesis using imine reductase is very attractive, because of the cost-effectiveness, high catalytic efficiency, and enantioselectivity. However, the steric hindrance of the 1-aryl substituents make this conversion very challenging, and current successful examples are mostly restricted to the simple alkyl-THIQs. In this report, through extensive evaluation of a large collection of IREDs (including 88 enzymes), we successfully identified a panel of steric-hindrance tolerated IREDs. These enzymes are able to convert meta- and para-substituted chloro-, methyl-, and methoxyl-benzyl dihydroisoquinolines (DHIQs) into corresponding R- or S- THIQs with very high enantioselectivity and conversion. Among them, the two most hindrance-tolerated enzymes (with different stereospecificity) are also able to convert ortho-substituted chloro-, methyl-, and methoxyl-benzyl DHIQs and dimethoxyl 1-chlorobenzyl-DHIQs with good enantiometric excess. Furthermore, using in silico modeling, a highly conserved tryptophan residue (W191) was identified to be critical for substrate accommodation in the binding cavity of the S-selective IRED (IR45). Replacing W191 with alanine can dramatically increase the catalytic performance by decreasing the Km value by 2 orders of magnitude. Our results provide an effective route to synthesize these important classes of THIQs. Moreover, the disclosed sequences and substrate binding model set a solid basis to generate more-efficient and broad-selective enzymes via protein engineering.