Unstructured Modeling of a Synthetic Microbial Consortium for Consolidated Production of Ethanol

Thursday, October 20, 2011: 1:30 PM
L100 J (Minneapolis Convention Center)
Timothy Hanly, Chemical Engineering, University of Massachusetts, Amherst, MA and Michael H. Henson, Department of Chemical Engineering, University of Massachusetts Amherst, Amherst, MA

The conversion of lignocellulosic biomass to liquid fuels such as ethanol is required for the commercialization of second generation biofuels.  Reducing operating costs by combining the saccharification and fermentation steps of this process into one reactor has long been a goal of biofuels research.  The typical approach to consolidated bioprocessing, engineering a single, omnipotent microbe to perform both these tasks, can result in undue metabolic burden and bottlenecks.  Using a defined mixed culture of specialized microbes is a promising alternative that exploits the native capabilities of each member species.  One possible consortium is composed of the filamentous fungus Trichoderma reesei, and two fermentative yeasts, Saccharomyces cerevisiae and Pichia stipitisT. reesei is able produce high concentrations of cellulolytic enzymes.  S. cerevisiae and P. stipitis efficiently produce ethanol from glucose and xylose, the predominant monomeric sugars that result from biomass hydrolysis, and have similar growth requirements as T. reesei.

We have developed a simple dynamic model of the proposed consortium with unstructured descriptions of enzyme synthesis, lignocellulose degradation, sugar uptake, cell growth and ethanol production. The batch growth model contained 10 ordinary differential equations with parameters obtained from the literature to the extent possible. The dynamic model equations were solved over a range of initial feedstock concentrations with 2:1 cellulose-hemicellulose ratios to identify the inoculum concentration that maximized the ethanol productivity.  The model was further used to examine how the ratio of cellulose to hemicellulose in the feedstock affected ethanol productivity and the optimal inoculum.  A sensitivity analysis of model parameters identified several promising targets for the improvement of ethanol production through genetic engineering.  The five highest ranked targets were the hydrolysis rate constant, the basal and induced production rates of the T. reesei hemicellulase, the saturation constant for S. cerevisiae growth on glucose, and the ethanol inhibition constant for P. stipitis. Simultaneous implementation of 25% improvements in these parameters was predicted to yield a 15% increase in ethanol productivity compared to the original consortium.  A series of batch mixed-culture experiments was performed to validate the predictive power of the model and to identify unmodeled interactions between the three microbes.


Extended Abstract: File Not Uploaded