The Ketosynthase Domain Constrains the Design of Polyketide Synthases

• Published in: bioRxiv
• Main Authors: Klaus, Maja, Buyachuihan, Lynn, Grininger, Martin
• Format: Paper
• Language: English
• Published: Cold Spring Harbor Cold Spring Harbor Laboratory Press 24.01.2020
• Subjects: Bioactive compounds; Biological activity; Chimeras; Interfaces; Metabolites; Mutants; Protein interaction; Proteins; Secondary metabolites; Substrate specificity
• DOI: 10.1101/2020.01.23.916668

Modular polyketide synthases (PKSs) produce complex, bioactive secondary metabolites in assembly line-like multistep reactions. The sequence of reaction steps determines the identity of the compound and is defined by the order of modules within the assembly line. This type of chemical programming makes modular PKSs ideal templates for engineering multistep C-C bond forming reactions. However, engineering of modular PKSs by recombining intact modules to produce novel, biologically active compounds has often resulted in chimeras exhibiting poor turnover rates and decreased product yields. Recent findings demonstrate that the low efficiencies of modular chimeric PKSs also result from rate limitations in the translocation step; i.e., the transfer of the growing polyketide from one module to the other and further processing of the polyketide substrate by the ketosynthase (KS) domain. In this study, we systematically vary protein:protein interfaces, substrate identity and substrate specificity of the KS involved reactions of bimodular chimeric PKSs. First, we broadened the substrate specificity of the C-C bond forming KS domains with the aim to increase the turnover rates of kinetically impaired chimeric PKSs. Diverse multipoint mutants were calculated with a Rosetta program suite to result in stable active site environments. Some mutations efficiently released the kinetic penalties imposed by the non-native substrate and especially one mutant exhibited unexpected higher turnover numbers than the reference system. Second, we changed the identity of the donor module in the chimeric constructs to also vary protein-protein interactions and substrate identity. We observed that the KS domain, specifically the translocation reaction, is sensitive towards both substrate identity and protein-protein interactions in a complex and often non-intuitive interplay of these constraints. Overall, our systematic study can explain why early oversimplified engineering strategies based on the plain shuffling of modules mostly failed and why more recent approaches show improved success rates.