Deciphering the evolutionary origin of the enantioselectivity of short-chain dehydrogenases from plants toward 1-borneol.

Enzyme engineering has produced numerous methods to optimize enzymes for biotechnological processes; however, less is known about how natural evolution creates new functionalities. We investigate the evolutionary emergence of enantioselectivity in plant borneol dehydrogenases, which feature hydrophobic active-sites and are enantioselective towards dibornane-type monoterpenols. Ancestral sequence reconstruction provided a trajectory from the oldest unselective ancestor N30 (E = 12) toward a more recent selective ancestor N32, involving 19 mutations: 18 mutations are peripheral, one (I111L) occurs in the active-site. The mutation L111I in the hydrophobic pocket increased the selectivity of N30, while the back-mutation I111L decreased the selectivity of N32. Additional peripheral mutations (V136L/G169A/V183I) were required for high selectivity. Crystal structures suggested that protein dynamics rather than structural changes shape these catalytic properties; this was confirmed by Machine Learning/Molecular Mechanics simulations of ligand binding. Funnel-metadynamics simulations revealed a correlation between the active-site's solvent-accessible surface area and selectivity. This potential evolutionary pathway shapes enantioselectivity, and guides future enzyme engineering campaigns.