Biosynthetic engineering and fermentation media development leads to gram-scale production of spliceostatin natural products in Burkholderia sp.

A key challenge in natural products drug discovery is compound supply. Hundreds of grams of purified material are needed to advance a natural product lead through preclinical development. Spliceostatins are polyketide-nonribosomal peptide natural products that bind to the spliceosome, an emerging target in cancer therapy. The wild-type bacterium Burkholderia sp. FERM BP-3421 produces a suite of spliceostatin congeners with varying biological activities and physiological stabilities. Hemiketal compounds such as FR901464 were the first to be described. Due to its improved properties, we were particularly interested in a carboxylic acid precursor analog that was first reported from Burkholderia sp. MSMB 43 and termed thailanstatin A. Inactivation of the iron/α-ketoglutarate-dependent dioxygenase gene fr9P had been shown to block hemiketal biosynthesis. However, a 4-deoxy congener of thailanstatin A was the main product seen in the dioxygenase mutant. We show here that expression of the cytochrome P450 gene fr9R is a metabolic bottle neck, as use of an l-arabinose inducible system led to nearly complete conversion of the 4-deoxy analog to the target molecule. By integrating fermentation media development approaches with biosynthetic engineering, we were able to improve production titers of the target compound >40-fold, going from the starting ~60 mg/L to 2.5 g/L, and to achieve what is predominantly a single component production profile. These improvements were instrumental in enabling preclinical development of spliceostatin analogs as chemotherapy.