Polyketides form many clinically valuable compounds. However, manipulation of their biosynthesis remains highly challenging. An understanding of gene cluster evolution provides a rationale for reprogramming of the biosynthetic machinery. Herein, we report characterization of giant modular polyketide synthases (PKSs) responsible for the production of aminopolyol polyketides. Heterologous expression of over 150 kbp polyketide gene clusters successfully afforded their products, whose stereochemistry was established by taking advantage of bioinformatic analysis. Furthermore, phylogenetic analysis of highly homologous but functionally diverse domains from the giant PKSs demonstrated the evolutionary mechanism for structural diversification of polyketides. The gene clusters characterized herein, together with their evolutionary insights, are promising genetic building blocks for de novo production of unnatural polyketides.

Versipelostatin (VST) is an unusual 17-membered macrocyclic polyketide product that contains a spirotetronate skeleton. In this study, the entire VST biosynthetic gene cluster (vst) spanning 108 kb from Streptomyces versipellis 4083-SVS6 was identified by heterologous expression using a bacterial artificial chromosome vector. Here, we demonstrate that an enzyme, VstJ, catalyzes the stereoselective [4+2]-cycloaddition between the conjugated diene and the exocyclic olefin of a newly identified tetronate-containing intermediate to form the spirotetronate skeleton during VST biosynthesis.

Quinolidomicin A 1 is the largest macrolide compound from terrestrial sources, consisting of a 60membered ring, and its biosynthetic gene cluster was revealed to be over 200 kb. The gene cluster for quinolidomicin was cloned and heterologously expressed. Confirmation of the product led to a structural revision, in which the hydroxy group in the chromophore moiety of the reported structure was replaced by an amine group.

One major bottleneck in natural product drug development is derivatization, which is pivotal for fine tuning lead compounds. A promising solution is modifying the biosynthetic machineries of middle molecules such as macrolides. Although intense studies have established various methodologies for protein engineering of type I modular polyketide synthase(s) (PKSs), the accurate targeting of desired regions in the PKS gene is still challenging due to the high sequence similarity between its modules. Here, we report an innovative technique that adapts in vitro Cas9 reaction and Gibson assembly to edit a target region of the type I modular PKS gene. Proof-of-concept experiments using rapamycin PKS as a template show that heterologous expression of edited biosynthetic gene clusters produced almost all the desired derivatives. Our results are consistent with the promiscuity of modular PKS and thus, our technique will provide a platform to generate rationally designed natural product derivatives for future drug development.

Natural enzymes that catalyze Diels-Alder reactions have long been sought after, yet few enzymes have been experimentally confirmed to perform this reaction. In the past five years, several stand-alone enzymes that can catalyze the Diels-Alder reaction had been identified and characterized. Among which, the crystal structures of SpnF, PyrI4 and AbyU have been determined. The structures of PyrI4 and AbyU, which are involved in spirotetronate/spirotetramate biosynthesis, are particularly informative since they shed light on how a natural catalyst captures the flexible substrate and facilitates the intramolecular Diels-Alder reaction through stabilization of the transition state in catalysis. These pioneering studies will inspire the design of artificial catalysts for Diels-Alder reactions.