Acetogens such as Clostridium ljungdahlii can play a crucial role reducing the human CO 2 footprint by converting industrial emissions containing CO 2 , CO and H 2 into valuable products such as organic acids or alcohols. The quantitative understanding of cellular metabolism is a prerequisite to exploit the bacterial endowments and to fine-tune the cells by applying metabolic engineering tools. Studying the three gas mixtures CO 2 + H 2 , CO and CO + CO 2 + H 2 (syngas) by continuously gassed batch cultivation experiments and applying flux balance analysis, we identified CO as the preferred carbon and electron source for growth and producing alcohols. However, the total yield of moles of carbon (mol-C) per electrons consumed was almost identical in all setups which underlines electron availability as the main factor influencing product formation. The Wood-Ljungdahl pathway (WLP) showed high flexibility by serving as the key NAD + provider for CO 2 + H 2, whereas this function was strongly compensated by the transhydrogenase-like Nfn complex when CO was metabolized. Availability of reduced ferredoxin (Fd red) can be considered as a key determinant of metabolic control. Oxidation of CO via carbon monoxide dehydrogenase (CODH) is the main route of Fd red formation when CO is used as substrate, whereas Fd red is mainly regenerated via the methyl branch of WLP and the Nfn complex utilizing CO 2 + H 2. Consequently, doubled growth rates, highest ATP formation rates and highest amounts of reduced products (ethanol, 2,3-butanediol) were observed when CO was the sole carbon and electron source.

Clostridium ljungdahlii (C. ljungdahlii, CLJU) is natively endowed producing acetic acid, 2,3-butandiol, and ethanol consuming gas mixtures of CO2, CO, and H2 (syngas). Here, we present the syngas-based isobutanol formation using C. ljungdahlii harboring the recombinant amplification of the “Ehrlich” pathway that converts intracellular KIV to isobutanol. Autotrophic isobutanol production was studied analyzing two different strains in 3-L gassed and stirred bioreactors. Physiological characterization was thoroughly applied together with metabolic profiling and flux balance analysis. Thereof, KIV and pyruvate supply were identified as key “bottlenecking” precursors limiting preliminary isobutanol formation in CLJU[KAIA] to 0.02 g L–1. Additional blocking of valine synthesis in CLJU[KAIA]:ilvE increased isobutanol production by factor 6.5 finally reaching 0.13 g L–1. Future metabolic engineering should focus on debottlenecking NADPH availability, whereas NADH supply is already equilibrated in the current generation of strains.

Two different isobutanol synthesis pathways were cloned into and expressed in the two model acetogenic bacteria Acetobacterium woodii and Clostridium ljungdahlii. A. woodii is specialized on using CO2 + H2 gas mixtures for growth and depends on sodium ions for ATP generation by a respective ATPase and Rnf system. On the other hand, C. ljungdahlii grows well on syngas (CO + H2 + CO2 mixture) and depends on protons for energy conservation. The first pathway consisted of ketoisovalerate ferredoxin oxidoreductase (Kor) from Clostridium thermocellum and bifunctional aldehyde/alcohol dehydrogenase (AdhE2) from C. acetobutylicum. Three different kor gene clusters are annotated in C. thermocellum and were all tested. Only in recombinant A. woodii strains, traces of isobutanol could be detected. Additional feeding of ketoisovalerate increased isobutanol production to 2.9 mM under heterotrophic conditions using kor3 and to 1.8 mM under autotrophic conditions using kor2. In C. ljungdahlii, isobutanol could only be detected upon additional ketoisovalerate feeding under autotrophic conditions. kor3 proved to be the best suited gene cluster. The second pathway consisted of ketoisovalerate decarboxylase from Lactococcus lactis and alcohol dehydrogenase from Corynebacterium glutamicum. For increasing the carbon flux to ketoisovalerate, genes encoding ketol-acid reductoisomerase, dihydroxy-acid dehydratase, and acetolactate synthase from C. ljungdahlii were subcloned downstream of adhA. Under heterotrophic conditions, A. woodii produced 0.2 mM isobutanol and 0.4 mM upon additional ketoisovalerate feeding. Under autotrophic conditions, no isobutanol formation could be detected. Only upon additional ketoisovalerate feeding, recombinant A. woodii produced 1.5 mM isobutanol. With C. ljungdahlii, no isobutanol was formed under heterotrophic conditions and only 0.1 mM under autotrophic conditions. Additional feeding of ketoisovalerate increased these values to 1.5 mM and 0.6 mM, respectively. A further increase to 2.4 mM and 1 mM, respectively, could be achieved upon inactivation of the ilvE gene in the recombinant C. ljungdahlii strain. Engineering the coenzyme specificity of IlvC of C. ljungdahlii from NADPH to NADH did not result in improved isobutanol production.