A detailed genome-scale metabolic model of Clostridium thermocellum investigates sources of pyrophosphate for driving glycolysis.

Lignocellulosic biomass is an abundant and renewable source of carbon for chemical manufacturing, yet it is cumbersome in conventional processes. A promising, and increasingly studied, candidate for lignocellulose bioprocessing is the thermophilic anaerobe Clostridium thermocellum given its potential to produce ethanol, organic acids, and hydrogen gas from lignocellulosic biomass under high substrate loading. Possessing an atypical glycolytic pathway which substitutes GTP or pyrophosphate (PP) for ATP in some steps, including in the energy-investment phase, identification, and manipulation of PP sources are key to engineering its metabolism. Previous efforts to identify the primary pyrophosphate have been unsuccessful. Here, we explore pyrophosphate metabolism through reconstructing, updating, and analyzing a new genome-scale stoichiometric model for C. thermocellum, iCTH669. Hundreds of changes to the former GEM, iCBI655, including correcting cofactor usages, addressing charge and elemental balance, standardizing biomass composition, and incorporating the latest experimental evidence led to a MEMOTE score improvement to 94%. We found agreement of iCTH669 model predictions across all available fermentation and biomass yield datasets. The feasibility of hundreds of PP synthesis routes, newly identified and previously proposed, were assessed through the lens of the iCTH669 model including biomass synthesis, tRNA synthesis, newly identified sources, and previously proposed PP-generating cycles. In all cases, the metabolic cost of PP synthesis is at best equivalent to investment of one ATP suggesting no direct energetic advantage for the cofactor substitution in C. thermocellum. Even though no unique source of PP could be gleaned by the model, by combining with gene expression data two most likely scenarios emerge. First, previously investigated PP sources likely account for most PP production in wild-type strains. Second, alternate metabolic routes as encoded by iCTH669 can collectively maintain PP levels even when previously investigated synthesis cycles are disrupted. Model iCTH669 is available at github.com/maranasgroup/iCTH669.