Improving CBCA synthase activity through rational protein design

Abstract

Global interest for the minor cannabinoid cannabichromene (CBC) is growing steadily, as potential pharmaceutical applications continue to emerge. Due to low-yielding and unspecific extraction processes from its plant host Cannabis sativa, a biotechnological production is desirable. The complete heterologous biosynthesis of several other cannabinoids has recently been demonstrated as an accessible platform. However, the enzyme involved in the biosynthesis of CBC precursor cannabichromenic acid (CBCA) suffers from comparatively low catalytic efficiency, has not been crystallized, and remains poorly characterized. This study contributes to overcoming these challenges in three unique aspects. A deep‑learning‑assisted prediction of the CBCA synthase crystal structure using DeepMinds AlphaFold is performed and evaluated. The predicted CBCA synthase structure scored considerably higher in various quality assessments than the alternative template‑based homology modeling approach. A robust and practical understanding of crucial structure-function relationships for CBCA synthase is provided and a new binding mode for the substrate uncovered. Rational design approaches and computational analyses to suggest CBCAS variants with facilitated activity are applied. Through subsequent screening the substrate conversion of those variants is compared to the native enzyme. The best variant presented in this study increases CBCA production from crude lysate 22-fold and is one of five positions where substitutions had a significantly favorable impact on product formation.