Fingerprint-guided engineering of selective P450 catalysts for sp³ C—H functionalization in complex molecules

Cytochrome P450s constitute attractive catalysts for the controlled oxidation of unactivated aliphatic C–H bonds in organic molecules. However, fine-tuning the regio- and stereoselectivity of P450 enzymes in the context of non-native transformations currently poses a major hurdle toward exploiting these biocatalysts for chemical synthesis. Here, we describe a systematic, rationally-driven strategy to expedite the development of P450 oxidation catalysts with refined regio- and stereoselectivity for the hydroxylation of remote, unactivated C?H sites in a complex scaffold. Using artemisinin as model substrate, we demonstrate how a three-tier strategy involving first-sphere active site mutagenesis, high-throughput P450 fingerprinting, and fingerprint-driven P450 reactivity predictions enabled the rapid evolution of three efficient biocatalysts for the selective hydroxylation of a primary and a secondary C?H site (with both S and R stereoselectivity) in a relevant yet previously inaccessible region of this complex natural product. The evolved P450 variants could be applied to provide direct access to the desired hydroxylated derivatives at preparative scales and in high yields, thereby enabling further elaboration of this molecule. As an example, enantiopure C7-fluorinated derivatives of the clinical antimalarial drugs artesunate and artemether, in which a major metabolically sensitive site is protected by means of a C?H to C?F substitution, could be prepared for the first time via P450-mediated chemoenzymatic synthesis.