Computationally-guided exchange of substrate selectivity motifs in a modular polyketide synthase acyltransferase.

Polyketides, one of the largest classes of natural products, are often clinically relevant. The ability to engineer polyketide biosynthesis to produce analogs is critically important. Acyltransferases (ATs) of modular polyketide synthases (PKSs) catalyze the installation of malonyl-CoA extenders into polyketide scaffolds. ATs have been targeted extensively to site-selectively introduce various extenders into polyketides. Yet, a complete inventory of AT residues responsible for substrate selection has not been established, limiting the scope of AT engineering. Here, molecular dynamics simulations are used to prioritize ~50 mutations within the active site of EryAT6 from erythromycin biosynthesis, leading to identification of two previously unexplored structural motifs. Exchanging both motifs with those from ATs with alternative extender specificities provides chimeric PKS modules with expanded and inverted substrate specificity. Our enhanced understanding of AT substrate selectivity and application of this motif-swapping strategy are expected to advance our ability to engineer PKSs towards designer polyketides.

Development of a Genetically Encoded Biosensor for Detection of Polyketide Synthase Extender Units in Escherichia coli.

The scaffolds of polyketides are constructed via assembly of extender units based on malonyl-CoA and its derivatives that are substituted at the C2-position with diverse chemical functionality. Subsequently, a transcription-factor-based biosensor for malonyl-CoA has proven to be a powerful tool for detecting malonyl-CoA, facilitating the dynamic regulation of malonyl-CoA biosynthesis and guiding high-throughput engineering of malonyl-CoA-dependent processes. Yet, a biosensor for the detection of malonyl-CoA derivatives has yet to be reported, severely restricting the application of high-throughput synthetic biology approaches to engineering extender unit biosynthesis and limiting the ability to dynamically regulate the biosynthesis of polyketide products that are dependent on such α-carboxyacyl-CoAs. Herein, the FapR biosensor was re-engineered and optimized for a range of mCoA concentrations across a panel of E. coli strains. The effector specificity of FapR was probed by cell-free transcription-translation, revealing that a variety of non-native and non-natural acyl-thioesters are FapR effectors. This FapR promiscuity proved sufficient for the detection of the polyketide extender unit methylmalonyl-CoA in E. coli, providing the first reported genetically encoded biosensor for this important metabolite. As such, the previously unknown broad effector promiscuity of FapR provides a platform to develop new tools and approaches that can be leveraged to overcome limitations of pathways that construct diverse α-carboxyacyl-CoAs and those that are dependent on them, including biofuels, antibiotics, anticancer drugs, and other value-added products.