| Production of biofuels and bio-based chemicals is being increasingly pursued by chemical industry to reduce its dependence on petroleum. Lignocellulosic biomass (LCB) is an abundant source of sugars that can be used for producing biofuels and bio-based chemicals using fermentation. Hydrolysis of LCB results in a mixture of sugars mainly composed of glucose and xylose. Fermentation of such a sugar mixture presents multiple technical challenges at industrial scale. Most industrial microorganisms utilize sugars in a sequential manner due to the regulatory phenomenon of carbon catabolite repression (CCR). Due to sequential utilization of sugars, the LCB-based fermentation processes suffer low productivities and complicated operation. Performance of fermentation processes can be improved by metabolic engineering of microorganisms to obtain superior characteristics such as high product yield. With increased computational power and availability of complete genomes of microorganisms, use of model-based metabolic engineering is now a common practice. The problem of sequential sugar utilization, however, is a regulatory problem, and metabolic models have never been used to solve such regulatory problems. The focus of this thesis is to use model-guided metabolic engineering to construct industrial strains capable of co-utilizing sugars. First, we develop a novel bilevel optimization algorithm SimUp, that uses metabolic models to identify reaction deletion strategies to force co-utilization of two sugars. We then use SimUp to identify reaction deletion strategies to force glucose-xylose co-utilization in Escherichia coli. To validate SimUp predictions, we construct three mutants with multiple gene knockouts and test them for glucose-xylose utilization characteristics. Two mutants, designated as LMSE2 and LMSE5, are shown to co-utilize glucose and xylose in agreement with SimUp predictions. To understand the molecular mechanism involved in glucose-xylose co-utilization of the mutant LMSE2, the mutant is subjected to targeted and whole genome sequencing. Finally, we use the mutant LMSE2 to produce D-ribose from a mixture of glucose and xylose by overexpressing an endogenous phosphatase. The methods developed in this thesis are anticipated to provide a novel approach to solve sugar co-utilization problem in industrial microorganisms, and provide insights into microbial response to forced co-utilization of sugars. |
