T he pentose phosphate (PP) pathway interconverts phosphosugars having 3 to 7 carbon atoms principally by the action of the reversible enzymes transketolase and transaldolase. During the consumption of hexoses such as D-glucose or D-fructose, entry of carbon into this pathway provides many microorganisms, including Escherichia coli, the means to generate the reduced cofactor NADPH and to synthesize specific building-block compounds derived from intermediates of this pathway (e.g., phenylalanine, histidine, and ribose). For microorganisms having the requisite kinases and sugar transport mechanisms, the PP pathway also provides convenient entry points for the catabolism of many other sugars, including D-xylose and L-arabinose (49).We have previously studied D-xylose and L-arabinose metabolism in E. coli that lacks the ability to metabolize D-glucose due to knockouts in the ptsG, manZ, and glk genes (1-3). Recently, small but consistent amounts (about 50 mg/liter) of D-glucose were observed as the accumulated end product when E. coli ptsG manZ glk was grown on 5 g/liter of either pentose in a defined medium (unpublished data). How might D-glucose be derived from these pentoses?Both D-xylose and L-arabinose are converted to the common intermediate D-xylulose-5-phosphate (D-xylulose-5P) (Fig. 1), which via the PP pathway partitions to 67% D-fructose-6P and 33% D-glyceraldehyde-3P without the involvement of ATP:(1) During growth of cells having a complete glycolytic pathway, the 2 mol of D-fructose-6P formed via equation 1 readily generates 4 mol of D-glyceraldehyde-3P. For D-glucose to accumulate from pentoses in cells prevented from metabolizing D-glucose, we reasoned that some D-fructose-6P generated from these pentoses (i.e., by equation 1) is converted "back" to D-glucose and that once formed, the D-glucose is unable to reenter metabolism in the triple knockout strain. We furthermore hypothesized that even more D-glucose would accumulate from pentoses in cells that were further constrained from metabolizing D-fructose-6P or D-glucose-6P.Because D-fructose-6P conversion to D-glyceraldehyde-3P is ubiquitous in wild-type organisms, D-glucose is not typically considered a product of D-xylose or L-arabinose metabolism, and the conversion of these pentoses to readily available D-glucose would in itself not seem to be an economically viable process. However, if the yields and rates were sufficiently large, the accumulation of hexoses directly from pentoses might advance the use of lignocellulosic hydrolysates with organisms, such as Saccharomyces cerevisiae, which metabolize D-glucose readily but are natively unable to consume pentoses. Moreover, conversion of 5-carbon saccharides into 6-carbon saccharides derived from D-fructose-6P offers a unique platform both to build carbon length and potentially to generate compounds in industrially relevant organisms such as E. coli that might not be possible under typical conditions in which products of D-fructose-6P do not accumulate.The objectives of this study were to examine ...