Linking genome content to biofuel production yields: a meta-analysis of major catabolic pathways among select H(2) and ethanol-producing bacteria

BACKGROUND: Fermentative bacteria offer the potential to convert lignocellulosic waste-streams into biofuels such as hydrogen (H(2)) and ethanol. Current fermentative H(2) and ethanol yields, however, are below theoretical maxima, vary greatly among organisms, and depend on the extent of metabolic pathways utilized. For fermentative H(2) and/or ethanol production to become practical, biofuel yields must be increased. We performed a comparative meta-analysis of (i) reported end-product yields, and (ii) genes encoding pyruvate metabolism and end-product synthesis pathways to identify suitable biomarkers for screening a microorganism’s potential of H(2) and/or ethanol production, and to identify targets for metabolic engineering to improve biofuel yields. Our interest in H(2) and/or ethanol optimization restricted our meta-analysis to organisms with sequenced genomes and limited branched end-product pathways. These included members of the Firmicutes, Euryarchaeota, and Thermotogae. RESULTS: Bioinformatic analysis revealed that the absence of genes encoding acetaldehyde dehydrogenase and bifunctional acetaldehyde/alcohol dehydrogenase (AdhE) in Caldicellulosiruptor, Thermococcus, Pyrococcus, and Thermotoga species coincide with high H(2) yields and low ethanol production. Organisms containing genes (or activities) for both ethanol and H(2) synthesis pathways (i.e. Caldanaerobacter subterraneus subsp. tengcongensis, Ethanoligenens harbinense, and Clostridium species) had relatively uniform mixed product patterns. The absence of hydrogenases in Geobacillus and Bacillus species did not confer high ethanol production, but rather high lactate production. Only Thermoanaerobacter pseudethanolicus produced relatively high ethanol and low H(2) yields. This may be attributed to the presence of genes encoding proteins that promote NADH production. Lactate dehydrogenase and pyruvate:formate lyase are not conducive for ethanol and/or H(2) production. While the type(s) of encoded hydrogenases appear to have little impact on H(2) production in organisms that do not encode ethanol producing pathways, they do influence reduced end-product yields in those that do. CONCLUSIONS: Here we show that composition of genes encoding pathways involved in pyruvate catabolism and end-product synthesis pathways can be used to approximate potential end-product distribution patterns. We have identified a number of genetic biomarkers for streamlining ethanol and H(2) producing capabilities. By linking genome content, reaction thermodynamics, and end-product yields, we offer potential targets for optimization of either ethanol or H(2) yields through metabolic engineering.