Orthologue inference-based enzyme mining for diversification of the anti-cancer evodiamine scaffold

Selective functionalization is central to the structural and functional diversification of pharmacologically important alkaloids, yet the complexity of these scaffolds often hinders chemical modifications. Enzymes such as cytochrome P450 monooxygenases can overcome these challenges by catalyzing stereo- and regioselective transformations at typically inert C-H sites. Here, we present a targeted workflow for mining such enzymes by integrating orthologue-inference bioinformatics to leverage the increasingly available biological data. Comparative analysis of 15 alkaloid-producing plant species generated a focused library of 15 P450s, five of which exhibited activities toward alkaloid scaffolds. Four enzymes from Camptotheca acuminata and Tabernaemontana elegans selectively oxygenated two positions of the anticancer scaffold evodiamine, a compound not reported in either lineage. Structural modelling and mutagenesis revealed that hydrophobic bulk in the active site governs evodiamine positioning and catalytic selectivity. Overall, we demonstrated that combining biocatalysts and substrates from distantly related plants with orthology-guided discovery enables selective functionalization of pharmacologically active alkaloids.