(430f) Combinatorial Pathway Optimization and Metabolic Engineering for Increased Production of the Anti-Malarial FR900098

It has been estimated that malaria kills over half a million people annually.  As resistance to current anti-malarial therapeutics spreads, new compounds to treat malaria are increasingly needed. One promising compound is FR900098, a naturally occurring phosphonate. This molecule mimics 1-deoxy-D-xylulose 5-phosphate (DOXP) of the non-mevalonate pathway, blocking essential isoprenoid synthesis in the malaria causing protozoan Plasmodium falciparum while not affecting the alternate isoprenoid pathway in humans. Due to limitations in both chemical synthesis and biosynthetic methods for FR900098 production, this potential therapeutic has yet to be considered for widespread use. Fermentation in the natural producer, Streptomyces rubellomurinus, yielded titers of 22.5 mg/L, and initial expression of the gene cluster in Escherichia coli led to an even lower production of 6 mg/L.  Several factors lead to less than optimal bioproduction of FR900098, including unfavorable thermodynamics, toxic intermediates, and diversion of substrates to dead-end byproducts. Due to these complications and the difficulty of detecting the pathway intermediates, to optimize the pathway a combinatorial library construction and screening was applied. To achieve high assembly efficiency for the combinatorial library, the construction was broken down into two rounds of Golden Gate assembly. In this library, expression of each of the eight biosynthetic genes and the FR900098 immunity gene was modulated with promoters of four strengths, leading to a library size of 4⁹ or 262,144. The library was screened by feeding the supernatants to an E. coli strain susceptible to FR900098 and measuring the resulting cell density.  Since each screening can only accommodate less than 1% of the theoretical library, a searching strategy was implemented by building subsequent libraries enriched in promoter assignments found in the previous library’s optimal strain. After two rounds of screening, a strain was found to produce over 50 mg/L FR900098 with a third round of screening yielding no significantly improved strains. In addition to pathway engineering, strain engineering methods are underway. Since each molecule of FR900098 requires one phosphoenol pyruvate and two acetyl-CoA molecules as substrates, gene knockout and knockdown of enzymes associated with these two metabolite pools are being implemented with promising results.