UC Berkeley

Metabolic engineering has been applied to a variety of microbial hosts to enhance production of compounds spanning from biofuels to pharmaceuticals. In all cases metabolic engineering requires the identification and optimization of the production pathway; however, engineering of optimal host strains capable of reliable production is multi-layered and complex. Strain engineering beyond the primary conversion pathway can lead to improved production levels, more robust strains, and better compatibility with downstream recovery methods. A common area where this facet of strain engineering becomes important is in alleviating the encumbrance on the cell caused by high product concentrations. In particular, these concerns are most relevant where productivity has to be maximized to achieve economic value, such as in commodity chemical and biofuel production. Many of the compounds valuable on an industrial scale are solvent-like hydrocarbons, where the accumulation of the compound is toxic to the cell. Furthermore separation of the somewhat hydrophobic compound becomes difficult as the large scale of the product formation leads to dissolved cell contaminants in the final product.

One potential engineering platform – efflux pumps – addresses many of the difficulties that come about from increased product yields. Engineered efflux transporters can secrete the product from the host thereby eliminating any potentially unwanted interactions between the product and the entirety of cellular machinery. Furthermore, secretion serves as a preliminary purification step for those compounds that are easily miscible with the rest of the cellular environment. This work discusses advances in the use of native and engineered efflux transporters to increase biochemical productivity. To this end, our first approach was to utilize directed evolution to mutate the native Escherichia coli efflux transporter AcrB to secrete the biofuel butanol. While AcrB’s native function is mainly in stabilizing E. coli tolerance to bile salts and antibiotics, we were able to alter its specificity towards n-butanol and other short, straight-chain alcohols. We managed to increase tolerance of the cells expressing these mutant pumps to butanol as well as increase titers of butanol when the production pathways were added into the strains.

In addition to demonstrating that directed evolution approaches can be used to enhance secretion of toxic products from the cell, we developed a continuum kinetics model of transporter activity and molecular diffusion of small molecules across the membrane. An interesting conclusion of this model shows that efflux transporters are likely to have the most benefit in the secretion of more hydrophilic molecules. However, when the model is expanded to include phase formation and separation of highly hydrophobic molecules, a new function for transporters is found. Transporters acting on very hydrophobic molecules are theoretically capable of creating a gradient strong enough to drive phase formation outside of the cell while keeping cellular concentration of the molecule at or below saturation concentration. Using this insight we searched for and identified native pumps in Saccharomyces cerevisiae capable of secreting the biodiesel molecule farnesene. Inactivation or deletion of these pumps results in intracellular accumulation of farnesene during production or from exogenous addition of the molecule.

This thesis serves as a foundation for developing the tools and for study and application of efflux pumps in biochemical production platforms. Our model, supported by empirical findings, can be used to help in the search of native transporters with unknown substrate specificities, and the tools we developed in engineering AcrB can aid in expanding the metabolic engineer’s toolbox for high yield chemical production.