Ethanol from lignocellulosic biomass is a promising alternative to rapidly depleting oil reserves. However, natural recalcitrance of lignocelluloses to biological and chemical treatments presents major engineering challenges in designing an ethanol conversion process. Current methods for pretreatment and hydrolysis of lignocelluloses generate a mixture of pentose (C5) and hexose (C6) sugars, and several microbial growth inhibitors such as acetic acid and phenolic compounds. Hence, an efficient ethanol production process requires a fermenting microorganism not only capable of converting mixed sugars to ethanol with high yield and productivity, but also having high tolerance to inhibitors. Although recombinant bacteria and yeast strains have been developed, ethanol yield and productivity from C5 sugars in the presence of inhibitors remain low and need to be further improved for a commercial ethanol production. The overarching objective of this work is to transform Zymomonas mobilis into an efficient whole-cell biocatalyst for ethanol production from lignocelluloses. Z. mobilis, a natural ethanologen, is ideal for this application but xylose (a C5 sugar) is not its 'natural' substrate. Back in 1995, researches at National Renewable Energy Laboratory (NREL) had managed to overcome this obstacle by metabolically engineering Z. mobilis to utilize xylose. However, even after more than a decade of research, xylose fermentation by Z. mobilis is still inefficient compared to that of glucose. For example, volumetric productivity of ethanol from xylose fermentation is 3- to 4- fold lower than that from glucose fermentation. Further reduction or complete inhibition of xylose fermentation occurs under adverse conditions. Also, high concentrations of xylose do not get metabolized completely. Thus, improvement in xylose fermentation by Z. mobilis is required. In this work, xylose fermentation in a metabolically engineered Z. mobilis was markedly improved by applying the technique of adaptive mutation. The adapted strain was able to grow on 10% (w/v) xylose and rapidly ferment xylose to ethanol within 2 days and retained high ethanol yield. Similarly, in mixed glucose-xylose fermentation, the strain produced a total of 9% (w/v) ethanol from two doses of 5% glucose and 5% xylose (or a total of 10% glucose and 10% xylose). Investigation was done to identify the molecular basis for efficient biocatalysis. An altered xylitol metabolism with reduced xylitol formation, increased xylitol tolerance and higher xylose isomerase activity were found to contribute towards improvement in xylose fermentation. Lower xylitol production in adapted strain was due to a single mutation in ZMO0976 gene, which drastically lowered the reductase activity of ZMO0976 protein. ZMO0976 was characterized as a novel aldo-keto reductase capable of reducing xylose, xylulose, benzaldehyde, furfural, 5-hydroxymethyl furfural, and acetaldehyde, but not glucose or fructose. It exhibited nearly 150-times higher affinity with benzaldehyde than xylose. Knockout of ZMO0976 was found to facilitate the establishment of xylose fermentation in Z. mobilis ZM4. Equipped with molecular level understanding of the biocatalytic process and insight into Z. mobilis central carbon metabolism, further genetic engineering of Z. mobilis was undertaken to improve the fermentation of sugars and lignocellulosic hydrolysates. These efforts culminated in construction of a strain capable of fermenting glucose-xylose mixture in presence of high concentration of acetic acid and another strain with a partially operational EMP pathway.