Résumé
Controlling the amounts of redox cofactors to manipulate metabolic fluxes is emerging as a useful approach to optimizing byproduct yields in yeast biotechnological processes. Redox cofactors are extensively interconnected metabolites, so predicting metabolite patterns is challenging and requires in-depth knowledge of how the metabolic network responds to a redox perturbation. Our aim was to analyze comprehensively the metabolic consequences of increased cytosolic NADPH oxidation during yeast fermentation. Using a genetic device based on the overexpression of a modified 2,3-butanediol dehydrogenase catalyzing the NADPH-dependent reduction of acetoin into 2,3-butanediol, we increased the NADPH demand to between 8 and 40-fold the anabolic demand. We developed (i) a dedicated constraint-based model of yeast fermentation and (ii) a constraint-based modeling method based on the dynamical analysis of mass distribution to quantify the in vivo contribution of pathways producing NADPH to the maintenance of redox homeostasis. We report that yeast responds to NADPH oxidation through a gradual increase in the flux through the PP and acetate pathways, providing 80% and 20% of the NADPH demand, respectively. However, for the highest NADPH demand, the model reveals a saturation of the PP pathway and predicts an exchange between NADH and NADPH in the cytosol that may be mediated by the glycerol–DHA futile cycle. We also reveal the contribution of mitochondrial shuttles, resulting in a net production of NADH in the cytosol, to fine-tune the NADH/NAD+ balance. This systems level study helps elucidate the physiological adaptation of yeast to NADPH perturbation. Our findings emphasize the robustness of yeast to alterations in NADPH metabolism and highlight the role of the glycerol–DHA cycle as a redox valve, providing additional NADPH from NADH under conditions of very high demand.
► A constraint-based model of yeast fermentation was developed and used to study NADPH homeostasis. ► The NADPH demand was increased to between 8 and 40-fold the anabolic demand. ► At moderate NADPH demand the PP and acetate pathways provide 80% and 20% of the NADPH demand. ► At high NADPH demand, the model predicts additional NADPH production from NADH by the glycerol–DHA futile cycle.