Invited speaker: Combinatorial biosynthesis of redesigned polyketide synthase genes

HUGO MENZELLA

Baltimore, MD, USA

Conferencia; Society of Industrial Microbiology Annual meeting; 2006

Type I polyketide synthase (PKS) genes consist of modules ~3-6 kb that encode the structures of 2-carbon units in theirpolyketide products. Alteration or replacement of PKS modules can lead to novel products, but techniques to effect changes have been extremely slow due to the lack of convenient tools to rapidly engineer the natural hosts or manipulate the wild type genes for heterologous expression. To overcome these limitations we have created a generic design of synthetic PKS genes where the DNA sequences are codon-optimized for expression in an E. coli strain engineered for polyketide production. In the redesigned sequences all the modules and domains are flanked by a repeated set of unique restriction sites, allowing facile cassette assembly and interchange.To test feasibility, we synthesized 17 modules from nine different PKS clusters and associated them in 264 bimodular combinations flanked by the loading module and the thioesterase domain of the 6-dEB synthase cluster. Remarkably, nearly half the combinations made the predicted polyketide in E. coli. All individual modules participated in at least one productive bimodular combination showing that all are catalytically competent to accept and extend an appropriate substrate.Next, we investigated the causes of the failures of the unproductive bimodular PKSs. Unnatural combinations of modules often fail to make a polyketide product because of a failure of the ketosynthase (KS) domain to extend the ketide donated to it by the module upstream of it. The causes of such problems are likely complex and are not yet amenable to rational correction. We therefore addressed the problem by exchanging KS domains of the redesigned acceptor modules in a combinatorial fashion, and co-expressing these chimeric modules with ketide-donor modules that naturally interact with the transplanted KS. This approach was remarkably successful in activating previously unproductive bimodular combinations.Overall, more than 3 million bp of novel PKS gene sequences were assembled to validate our approach to identify novel module-module interfaces. In these experiments the results augur well for the ongoing development of molecular tools to design and produce novel polyketides.