Use of de novo Protein Design Methods to Rationally Redesign P450 Enzymes for New Cofactors and Substrates - PROJECT SUMMARY The synthesis and modification of biologically active molecules is critical to the develop- ment of new pharmaceuticals. Carbene and nitrene group-transfer reactions are particularly valuable for these purposes because they enable the rapid construction of complex scaffolds and the introduction of functional groups into complex molecules. However, selective, intermolecular, group-transfer reactions are challenging to achieve. Biocatalysis with artificial metalloenzymes is an increasingly promising approach to meet this challenge, and the development of artificial metalloenzymes has enabled a wide variety of C–H functionalization and olefin cyclization reactions by group transfers. P450 enzymes have been used for these transformations, due to the presence of a highly evolved, enclosed active site that provides greater activity and selectivity than the rigid and/or shallow binding pockets common in other metalloenzymes. As a result, unnatural intermolecular reactions are usually accomplished with this class of enzyme. However, accomplishing new transformations with P450 enzymes is typically limited to directed evolution of a natural P450 enzyme or further evolution of an evolved variant. Because current P450 enzymes with porphyrinoid cofactors do not catalyze group-transfers between many synthetically valuable combinations of carbenes or nitrenes and alkenes or C–H bonds, this limitation impedes the application of P450 enzymes to synthetic problems. Thus, new strategies are needed to redesign P450 enzymes for binding unnatural cofactors and new classes of substrates. One possible strategy is the de novo design of new enzymes. Methods for de novo design are increasingly powerful, and it is now possible to generate complex tertiary structures that bind to any desired reactive complex. However, it remains challenging to design de novo active sites with the high activity and selectivity of natural enzymes. We propose to use de novo protein design strategies to redesign cytochrome P450 enzymes to incorporate new cofactors and substrates, while retaining the natural P450 active site. This redesign will allow us to develop group-transfer reactions that are difficult to achieve with current P450 enzymes. We will follow two approaches to test this hypothesis. First, we will use a diffusion model to design a new cofactor binding motif for incorporation of an unnatural piano-stool iridium cofactor. Because this cofactor is more active than porphyrin cofactors for acyl nitrene transfers, this approach will allow us to evolve highly active en- zymes for olefin aziridination and C–H amidation. Second, we will use sequence redesign to stabilize an enzyme that possesses an active site that can accommodate large molecules but that has been too unstable to evolve for unnatural reactivity. Together, our approach for the development of artificial metalloenzymes will combine the flexibility of de novo design methods with the highly evolved active site of natural P450 enzymes. In doing so, we will accomplish the first use of a non-porphyrinoid cofactor within a P450 enzyme and enable the late-stage diversification of molecules by selective nitrene and carbene group-transfer reactions.
