The primary objective of this research is to develop microbial production strains which will serve to host NAD(P)H-dependent, heterologous reactions with improved efficiency in the generation and subsequent utilization of reduced cofactors derived from glucose or other renewable energy sources. E. coli is the microbial host chosen for this work, and we are initially studying the reduction of xylose to xylitol (using NADPH-dependent xylose reductase, XR) to serve as an experimental platform that allows us to systematically characterize the influences of select genetic modifications on strain performance, measured as the yield on xylitol produced per glucose consumed as co-substrate. Maximizing this yield translates to uncoupling carbon metabolism from respiration or fermentation. Stoichiometric network analysis and a bilevel optimization framework (“OptKnock”) are being used to understand the influence of potentially critical enzymes on theoretical yields and to suggest knockout strategies that will constrain the network such that cell growth or ATP production are coupled to xylitol production.

In order to accurately reflect conditions of low growth, we are using metabolically active but non-growing “resting cells” to evaluate strain performance. Use of resting cells additionally eliminates growth as a variable that alters partitioning of carbon and reducing equivalents, improving xylitol yield and allowing for more accurate comparisons between different engineered strains with different growth characteristics. Yield measurements from several knockout strains (e.g., atpA, pgi, zwf, sthA, pntA) under different production conditions are helping to elucidate how various metabolic pathways contribute to NADPH-dependent xylose reduction, and to characterize the effects of overexpressing an NADPH sink (a common scenario in whole-cell biocatalysis) on partitioning of co-substrate carbon and reducing equivalents. Our results indicate that native transhydrogenase activity is insufficient for providing XR with reducing equivalents derived from NADH, and NADPH is primarily derived from the pentose phosphate pathway and TCA cycle. We will report results for the case of transhydrogenase overexpression. Finally, approaches to overcoming xylose transport limitations will be addressed.