Natural products have outstanding pharmaceutical pedigree - approximately 50% of FDA approved small molecule drugs are derived from plant, bacterial or fungal sources. To discover new natural products, Warp Drive Bio has assembled an enormous proprietary and searchable database of microbial genomes and biosynthetic gene clusters. To date, our database comprises ~135,000 genomes sequenced encoding the biosynthesis of ~3.5 million natural compounds. Over the course of our genome mining campaign, we identified WDB-002, a cyclic polyketide synthase (PKS) natural product, by a search for biosynthetic clusters related to FK506 and Rapamycin. Both FK506 and Rapamycin target important signaling pathways with high potency and selectivity, and these compounds have entered the clinic with annual sales of $3 billion. We propose to force the evolution of this clinically-relevant family of natural products by applying combinatorial biosynthesis techniques to create a library of new WDB-002 analogs. The innovation described here represents an integrated platform for the combinatorial biosynthesis of polyketides on an industrial scale. We built a prediction algorithm for the engineering of novel PKS clusters by analyzing our vast genomic inventory of PKS natural products. Second, we developed a robust genetic assembly methodology to construct of a chemically diverse PKS library of WDB-002 analogs, and we expressed the library in an optimized Streptomyces host. Third, we have developed two mass spectrometry assays based on FKBP12 binding that allow the precise identification of WDB-002 analogs and a FKBP12 affinity proteomics-LC/MS2 assay to identify the protein targets of WDB-002 analogs in human lysates. The goal of this proposal is to assess the feasibility of large-scale evolvalog generation by performing a detailed analysis of a 650-member evolvalog library.
In Aim 1, we propose to analyze the library by mass spectrometry to determine if the engineered clusters produce novel evolvalogs. In parallel, we will assess the genetic identity of each evolvalog expression construct by next-generation sequencing. By linking the identity of module-module combinations to successful polyketide production, we can refine our module swapping algorithm to increase the success rate for new library designs based on high-value PKS clusters.
In Aim 2, we will map the human protein targets of 50 evolvalogs by a target-ID assay to understand the rules by which polyketide engineering can reprogram target engagement in human proteome. These data will accelerate the rational design of polyketide combinatorial libraries with novel target profiles, thus yielding a direct path to diverse polyketide natural product libraries for drug discovery.
Using multiple bioinformatic methods and a robust genetic assembly methodology, we have assembled a 650- member natural products library to experimentally assess and computationally optimize polyketide combinatorial biosynthesis, in hopes of identifying novel therapeutics to treat infections and other diseases.