Molecular basis for the stereoselective ammoniolysis of N-alkyl aziridine-2-carboxylates catalyzed by Candida antarctica lipase B.

Candida antarctica lipase B catalyzed the stereoselective ammoniolysis of N-alkyl aziridine-2-carboxylates in tBuOH saturated with ammonia and yielded the (2S)-aziridine-2-carboxamide and unreacted (2R)-aziridine-2-carboxylate. Varying the N-1 substituent on the aziridine ring changed the rate and stereoselectivity of the reaction. Substrates with a benzyl substituent or a (1'R)-1-phenylethyl substituent reacted approximately ten times faster than substrates with a (1'S)-1-phenylethyl substituent. Substrates with a benzyl substituent showed little stereoselectivity (E=5-7) while substrates with either a (1'R)- or (1'S)-1-phenylethyl substituent showed high stereoselectivity (D>50). Molecular modeling by using the current paradigm for enantioselectivity-binding of the slow enantiomer by an exchange-of-substituents orientation-could not account for the experimental results. However, modeling an umbrella-like-inversion orientation for the slow enantiomer could account for the experimental results. Steric hindrance between the methyl in the (1'S)-1-phenylethyl substituent and Thr138 and Ile189 in the acyl-binding site likely accounts for the slow reaction. Enantioselectivity likely stems from an unfavorable interaction of the methine hydrogen with Thr40 for the slow enantiomer and from subtle differences in the orientations of the other three substituents. This success in rationalizing the enantioselectivity supports the notion that an umbrella-like-inversion orientation can contribute to enantioselectivity in lipases.