Computational design of a Novel dehydratase for renewable fuels and chemicals

The reaction catalyzed by dihydroxyacid dehydratase is a key bottleneck in the biosynthesis of isobutanol from glucose, caused in part by extremely inefficient maturation of the catalytically essential Fe-S cofactor (~1%) when the dehydratase is recombinantly expressed. One strategy to relieve this bottleneck is to replace this complicated, energetically expensive enzyme with a
simpler one that uses Mg2+ instead of an Fe-S cluster as cofactor. To accomplish this, computational protein design will be used to engineer dihydroxyacid dehydratase activity into a sugar acid dehydratase. These enzymes express well in recombinant organisms, and although the wild-type enzymes do not act on the target substrate, they catalyze analogous dehydration reactions using simple metal ions. Starting from the E. coli sugar acid dehydratase EcYfaW, the active site will be redesigned to accommodate the target substrate, generating a novel dihydroxyacid dehydratase that will be amenable to expression in yeast; this should increase active enzyme concentration and enhance the overall reaction rate, leading to improved isobutanol yields and decreased production costs. Phase I objectives are: (a) apply state-of-the- art computational protein design to design combinatorial libraries of EcYfaW variants with dihydroxyacid dehydratase activity; (b) use high-throughput screening and/or selection to identify active EcYfaW variants; (c) apply directed evolution to further improve catalytic activity. This research will improve the efficiency of biosynthetic production of isobutanol, a product that holds substantial value as a fuel additive and in the production of hydrocarbon transportation fuels, plastics, rubber, and other polymers. An Fe-S-independent dehydratase could also be valuable in several other transformations leading to key industrial chemicals such as isobutylene, polypropylene, and polystyrene. By facilitating the biosynthesis of isobutanol and other chemicals from a biomass-derived sustainable resource (glucose), this project can help reduce U.S. dependence on foreign oil and spur domestic manufacturing, investment, and job creation. Furthermore, it may ultimately provide a carbon-neutral route to transportation fuel, resulting in myriad environmental benefits. Our understanding of enzyme-substrate-cofactor interactions and how computational tools can be applied to tailor enzyme activity and specificity will be enhanced in the execution of this project. The principles learned may facilitate the use of computer-aided design methods and encourage the design of other novel enzymes to replace the many whose dependence on complex cofactors have hindered their use in industry. This work will also demonstrate the utility of combining computational protein design and directed evolution on a specific problem of commercial significance. The wider adoption of hybrid computational- and evolution-based protein engineering methods could significantly reduce research time and costs by replacing the bulk of expensive laboratory-evolution efforts with in silico screening. A hybrid approach may enable the use of custom engineered proteins in previously intractable applications such as patient-specific medical treatments and facilitate the development of novel, highly active catalysts for industrial applications, broadening the scope of safe, clean enzymatic catalysis to reactions never addressed by Nature.