The thermophilic route to succinic acid

Pseudoclostridium thermosuccinogenes is the only known thermophile that produces succinic acid as one of its main fermentations products, together with acetic acid and formic acid, as well as ethanol, lactic acid, and hydrogen in smaller quantities. Thermophilic cell-factories, and succinic acid are both of interest for industrial biotechnology, and it is therefore that we set out to try to understand the metabolism of P. thermosuccinogenes.In Chapter 2 we published the genome sequences of the four available P. thermosuccinogenes strains (at that point still called Clostridium thermosuccinogenes). The genome of the type strain (DSM 5807) was fully closed. Using these annotated genomes, we were able to reconstruct its central metabolism. The genes for the pathways towards all the fermentation products were identified, as well as the complete Embden-Meyerhof-Parnas pathway. All genes for the pentose phosphate pathway (including those for xylose assimilation) were identified except for a transaldolase. Transcriptomics during growth on xylose versus glucose did not provide any leads to potential transaldolase genes or alternatives. Enzyme assays with cell-free extract were conducted to study cofactor usage of various glycolytic reactions. We were able to show that glucokinase was GTP-dependent and that 6-phosphofructokinase was PPi-dependent, verifying what was previously shown in Hungateiclostridium thermocellum, a close relative of P. thermosuccinogenes. Furthermore, xylulokinase was shown to use GTP as well.In Chapter 3 we looked further at the cofactor usage of P. thermosuccinogenes. Thirteen genes were cloned and heterologously expressed in Escherichia coli (encoding ribokinase, galactokinase, acetate kinase, isocitrate dehydrogenase, three 6-phosphofructokinase (PFK) orthologs, three glyceraldehyde 3-phosphate dehydrogenase (GAPDH) orthologs, and three genes encoding either pyruvate kinase, pyruvate, phosphate dikinase, or phosphoenolpyruvate synthase). Via enzyme assays we showed that besides glucokinase and xylulokinase, galactokinase and ribokinase are also GTP-dependent, suggesting that sugar phosphorylation by and large is GTP-dependent in P. thermosuccinogenes. Of the three PFKs, we confirmed which was PPi-dependent, and we found that another was active with both ATP and GTP; no activity was found for the third. Two GAPDHs were found to be NAD+-dependent. Further, the use of PPi and GTP as phosphoryl carriers was extensively discussed; we hypothesize that the use of GTP allows for different (or more flexible) reaction thermodynamics compared to reactions relying on ATP.In Chapter 4 the pathway to succinic acid in P. thermosuccinogenes was investigated in more detail. The fumarate hydratase and fumarate reductase (FRD) genes reside in an operon together with the genes for a large electron bifurcating NADH-reductase-heterodisulfide reductase complex (Flx-Hdr) that takes electrons from NADH, reducing ferredoxin and a disulfide bond simultaneously. The FRD differs significantly from studied isoforms, but is closest related to methanogenic FRDs that use thiols to reduce fumarate. Based on this genomic context and comparative genomics, we propose two hypothetical mechanisms through which the FRD associates with the electron bifurcating Flx-Hdr complex: (1) A disulfide bond from a hitherto unknown cofactor is reduced by the Flx-Hdr complex, using NADH to generate two thiol groups, while facilitating the unfavourable reduction of ferredoxin by NADH. The disulfide bond is subsequently regenerated via the reduction of fumarate to succinate by the FRD using the previously formed thiol groups. Or, (2) the FRD forms an integral part of the FlxABCD-HdrABC complex, and NADH is used to reduce ferredoxin and fumarate directly, without an intermediate disulfide-forming cofactor. Either way enables the conservation of additional energy (in the form of reduced ferredoxin) by a soluble FRD. Some preliminary, inconclusive experimental data are presented as well.In Chapter 5 the effect of CO2 limitation on succinate yield and on the metabolism of P. thermosuccinogenes in general was studied. Succinate production is connected to net fixation of CO2 (by PEP carboxykinase) and was, therefore, expected to be impacted significantly. Batch cultivations in bioreactors sparged with 1% and 20% CO2 were conducted that allowed us to carefully study the effect of CO2 limitation. Formate yield was greatly reduced at low CO2 concentrations, signifying a switch from pyruvate formate lyase (PFL) to pyruvate:ferredoxin oxidoreductase (PFOR) for acetyl-CoA formation. The corresponding increase in endogenous CO2 production (by PFOR) enabled succinic acid production to be largely maintained as its yield was reduced by only 26%. Acetate yield was slightly reduced as well, while that of lactate was slightly increased. CO2 limitation also prompted the formation of significant amounts of ethanol, which is only marginally produced during CO2 excess. Altogether, the changes in fermentation product yields result in increased ferredoxin and NAD+ reduction, and increased NADPH oxidation during CO2 limitation, which must be linked to reshuffled (trans)hydrogenation mechanisms of those cofactors, to keep them balanced. RNA sequencing, to investigate transcriptional effects of CO2 limitation, yielded only ambiguous results regarding the known (trans)hydrogenation mechanisms, hinting at a decreased NAD+/NADH ratio, which could ultimately be responsible for the stress observed during CO2 limitation. Clear overexpression of an alcohol dehydrogenase (adhE) was observed, which explains the increased ethanol production, while no changes were seen for PFL and PFOR expression that could explain the anticipated switch based on the fermentation results.In Chapter 6 the pentose phosphate pathway of Hungateiclostridiaceae was investigated to find out how they are able to interconvert C5 and C3/C6 metabolites in the absence of a transaldolase. We were able to confirm that Hungateiclostridiaceae rely on the sedoheptulose 1,7-bisphosphate (SBP) pathway, using pyrophosphate-dependent phosphofructokinase (PPi-PFK) instead of transaldolase. In the SBP pathway, sedoheptulose 7-phosphate is converted to SBP by PPi-PFK after which fructose bisphosphate aldolase cleaves SBP into dihydroxyacetone phosphate and erythrose 4-phosphate. We showed that PPi-PFK of P. thermosuccinogenes and of H. thermocellum indeed can convert S7P to SBP, and that they have similar affinities for S7P and fructose 6-phosphate (F6P), the canonical substrate. By contrast, (ATP-dependent) PfkA of Escherichia coli (which does rely on transaldolase) has a very poor affinity for S7P, indicative of the fact that the PPi-PFK of the Hungateiclostridiaceae has evolved for the use of S7P. We further show that P. thermosuccinogenes contains a significant SBP pool, an otherwise unusual metabolite, which is elevated during growth on xylose, demonstrating its relevance for pentose assimilation.