S113: Computational design of cellulase cocktails that improve sugar release from real-world feedstocks

Fuels based on plant biomass show great potential as a sustainable source of energy.  However, many technical problems must be overcome before cellulosic fuels can compete with traditional fuels.  Enzyme costs currently contribute significantly to the price of cellulosic fuel; we estimate that a 4-fold decrease in enzyme loading coupled with a 2-fold increase in sugar-release rate could yield a 30% reduction in the total production cost of cellulosic ethanol.  Thermostabilized cellulases might foment these improvements by enabling saccharification to continue longer at standard or elevated temperatures.  Towards this goal, we applied a variety of computational strategies to improve the thermostability of cellulases, including core repacking, consensus design, proline introduction, glycine elimination, and disulfide design.  The designed variants were evaluated by measuring glucose release from AFEX-pretreated corn stover in a minimal 3-component cellulase cocktail.  By screening small libraries of these designs, we discovered several cellulase variants with enhanced thermostability and expression yield; one variant was thermostabilized by 9 degrees.  By substituting this variant for its wild type in the 3-component cocktail, improved glucose release compared to the original cocktail was achieved, even when only a third as much total enzyme was applied.  This encouraging result validates the hypothesis that thermostabilized cellulases might significantly reduce production costs for cellulosic fuels.  It also demonstrates that economically relevant enzyme improvements may be achieved with low screening capacity when computational modeling is employed as an in silico prescreen.