Computational design of myoglobin-based carbene transferases for monoterpene derivatization

Abstract

Carbene transfer reactions have emerged as pivotal methodologies for the synthesis of complex molecular architectures. Heme protein-catalyzed carbene transfer reactions have shown promising results on model compounds. However, their limited substrate scope has hindered their application in natural product functionalization. Building upon the foundation of previously published work on a carbene transferase-myoglobin variant, this study employs computer-aided protein engineering to design myoglobin variants, using either docking or the deep learning-based LigandMPNN method. These variants were utilized as catalysts in carbene transfer reactions with a selection of monoterpene substrates featuring C–C double bonds, leading to seven target products. This cost-effective methodology broadens the substrate scope for heme protein-catalyzed reactions, thereby opening novel pathways for research in heme protein functionalities and offering fresh perspectives in the synthesis of bioactive molecules.