Psilocybin is the principal natural product of Psilocybe and other fungal genera, collectively referred to as “magic mushooms”. Therapeutic evaluation of psilocybin has revealed a remarkable potential to treat a variety of psychological conditions, including major depressive disorder, substance dependence, and end-of-life anxiety. As a result, psilocybin has received “breakthrough therapy” status by the US Food and Drug Administration and is currently the subject of approximately 200 clinical trials. The rising interest in potential medical applications has prompted efforts to produce psilocybin biotechnologically, as well as to explore the properties of novel analogs. Four enzyme-encoding genes required for the biosynthetic pathway from L-tryptophan to psilocybin were recently identified in Psilocybe cubensis (P. cubensis) and other species, paving the way for large-scale heterologous production in microorganisms. Although heterologous production of psilocybin has been achieved, scalable production of psilocybin analogs requires detailed characterization of the enzymes involved to optimize their activity on non-native substrates. The final reaction in the psilocybin biosynthetic pathway is catalyzed by PsiM, an S-adenosyl-L-methionine (SAM) dependent methyltransferase. This enzyme carries out two successive methylations on its native substrate, norbaeocystin, to produce the tertiary amine, psilocybin. PsiM exhibits strict substrate specificity toward norbaeocystin as indicated by low to no turnover of analogous substrates. However, the pertinent structural features and subsequent mechanistic implications for the specificity of PsiM have not been identified. To address these fundamental gaps in our understanding of the last step in the biosynthesis of psilocybin, we structurally and biochemically characterized PsiM from P. cubensis. Using X-ray crystallography, we achieved sub-angstrom resolution of the first crystal structures of PsiM in all stages of its reaction cycle revealing geometric restraints that dictate activity and selectivity. We also include a full kinetic characterization of PsiM toward norbaeocystin and baeocystin. Inspection of the secondary sphere of the active site in conjunction with sequence alignments to homologous methyltransferases and phylogenetic analysis strongly suggest PsiM evolved from regulatory monomethylating RNA methyltransferases. Mutagenesis studies support our evolutionary hypothesis through the production of PsiM mutants that mimic ancestral activity through the inability to produce psilocybin. Ultimately, our findings suggest PsiM is not an ideal methyltransferase for the biotechnological production of psilocybin analogs due to the delicate nature of substrate binding resulting from its evolutionary history.
