S65: Protein engineering of polyketide synthase from fredericamycin biosynthetic pathway using machine learning and synthetic genes

Polyketide synthases (PKSs) are a family of enzymes that catalyze the biosynthesis of structurally diverse and pharmacologically important natural products. Two major types of PKSs have been identified in microorganisms. Type I PKSs have multiple active sites on each polypeptide, in contrast to type II PKSs where each polypeptide has a single active site. In type II PKSs, a minimal PKS is composed of the ketosynthase (KS), the chain length factor (CLF), the acyl carrier protein, and the malonyl-CoA:ACP transacylase. The heterodimeric KS-CLF catalyzes polyketide chain elongation and CLF controls the polyketide chain length. The chain length of their polyketide backbones is the most critical property that influences product structure. Successful engineering of the polyketide chain length will set the stage for a major expansion of polyketide natural product libraries. Fredericamycin (fdm) A, a pentadecaketide, is the longest type II polyketide natural product characterized to date. Previous structural and mutagenic studies of actinorhodin KS-CLF suggested that the polyketide chain length was gated by seven residues. Based on sequence analysis of all known KS-CLF enzymes using machine learning algorithms, we hypothesized that additional thirty-five residues play potentially important roles in chain length control. Accordingly, forty-eight quadruple or quintuple fdm KS-CLF mutants were made by high throughput gene synthesis. These forty-eight mutants were introduced into Streptomyces coelicolor and characterized. Interestingly, some of them showed an altered product profile compared to that of the wild-type enzyme. Taken together, the design and characterization of the fredericamycin KS-CLF mutants will be presented in this talk.