Modular polyketide synthases (PKSs) produce complex, bioactive secondary metabolites in assembly line-like multistep reactions. Longstanding efforts to produce novel, biologically active compounds by recombining intact modules to new modular PKSs have mostly resulted in poorly active chimeras and decreased product yields. Recent findings demonstrate that the low efficiencies of modular chimeric PKSs also result from rate limitations in the transfer of the growing polyketide chain across the non-cognate module:module interface and further processing of the non-native polyketide substrate by the ketosynthase (KS) domain. In this study, we aim at disclosing and understanding the low efficiency of chimeric modular PKSs and at establishing guidelines for modular PKSs engineering. To do so, we work with a bimodular PKS testbed and systematically vary substrate specificity, substrate identity, and domain:domain interfaces of the KS involved reactions. We observe that KS domains employed in our chimeric bimodular PKSs are bottlenecks with regards to both substrate specificity as well as interaction with the ACP. Overall, our systematic study can explain in quantitative terms why early oversimplified engineering strategies based on the plain shuffling of modules mostly failed and why more recent approaches show improved success rates. We moreover identify two mutations of the KS domain that significantly increased turnover rates in chimeric systems and interpret this finding in mechanistic detail..

Modular polyketide synthases (PKSs) are attractive targets for the directed, biosynthetic production of platform chemicals and pharmaceuticals by protein engineering. In this study, we analyze docking domains from the 6-deoxyerythronolide B synthase, SYNZIP domains, and the SpyCatcher:SpyTag complex as engineering tools to couple the polypeptides VemG and VemH to functional venemycin synthases. Our data show that the high-affinity interaction or covalent connection of modules, enabled by SYNZIP domains and the SpyCatcher:SpyTag complex, can be advantageous, e.g., in synthesis at low protein concentrations, but their rigidity and steric demand decrease synthesis rates. However, we also show that efficiency can be recovered when inserting a hinge region distant from the rigid interface. This study demonstrates that engineering approaches should take the conformational properties of modular PKSs into account and establishes a three-polypeptide split venemycin synthase as an exquisite in vitro platform for the analysis and engineering of modular PKSs.

Modular polyketide synthases (PKSs) are attractive targets for the directed, biosynthetic production of platform chemicals and pharmaceuticals by protein engineering. In this study, we analyze docking domains from the 6-deoxyerythronolide B synthase, SYNZIP domains, and the SpyCatcher:SpyTag complex as engineering tools to couple the polypeptides VemG and VemH to functional venemycin synthases. Our data show that the high-affinity interaction or covalent connection of modules, possible by SYNZIP domains and the SpyCatcher:SpyTag complex, can be advantageous, e.g., in synthesis at low protein concentrations, but their rigidity and steric demand decrease synthesis rates. However, we also show that efficiency can be recovered when inserting a hinge region distant from the rigid interface. This study demonstrates that engineering approaches should take the conformational properties of modular PKSs into account and establishes a three-polypeptide split-venemycin synthase as an exquisite in vitro platform for the analysis and engineering of modular PKSs.

Modular polyketide synthases (PKSs) produce complex, bioactive secondary metabolites in assembly line-like multistep reactions. Longstanding efforts to produce novel, biologically active compounds by recombining intact modules to new modular PKSs have mostly resulted in poorly active chimeras and decreased product yields. Recent findings demonstrate that the low efficiencies of modular chimeric PKSs also result from rate limitations in the transfer of the growing polyketide chain across the non-cognate module:module interface and further processing of the non-native polyketide substrate by the ketosynthase (KS) domain. In this study, we aim at disclosing and understanding the low efficiency of chimeric modular PKSs and at establishing guidelines for modular PKSs engineering. To do so, we work with a bimodular PKS testbed and systematically vary substrate specificity, substrate identity, and domain:domain interfaces of the KS involved reactions. We observe that KS domains employed in our chimeric bimodular PKSs are bottlenecks with regards to both substrate specificity as well as interaction with the ACP. Overall, our systematic study can explain in quantitative terms why early oversimplified engineering strategies based on the plain shuffling of modules mostly failed and why more recent approaches show improved success rates. We moreover identify two mutations of the KS domain that significantly increased turnover rates in chimeric systems and interpret this finding in mechanistic detail.