A multi-phase multi-objective dynamic genome-scale model shows different redox balancing among yeast species in fermentation

Abstract

Yeasts constitute over 1500 species with great potential for biotechnology. Still, the yeast Saccharomyces cerevisiae dominates industrial applications and many alternative physiological capabilities of lesser-known yeasts arenot being fully exploited. While comparative genomics receives substantial attention, little is known about yeasts' metabolic specificity in batch cultures. Here we propose a multi-phase multi-objective dynamic genome-scale model of yeast batch cultures that describes the uptake of carbon and nitrogen sources and the production of primary and secondary metabolites. The model integrates a specific metabolic reconstruction, based on the consensus Yeast8, and a kinetic model describing the time-varying culture environment. Besides, we proposed a multi-phase multi-objective flux balance analysis to compute the dynamics of intracellular fluxes. We then compared the metabolism of S. cerevisiae and S. uvarum strains in a rich medium fermentation. The model successfully explained the experimental data and brought novel insights into how cryotolerant strains achieve redox balance. The proposed model (along with the corresponding code) provides a comprehensive picture of the main steps occurring inside the cell during batch cultures and offers a systematic approach to prospect or metabolically engineering novel yeast cell factories.

Dynamic genome-scale modeling ofSaccharomyces cerevisiaeunravels mechanisms for ester formation during alcoholic fermentation

“…This key set of fluxes includes presenting the role of NAD-dependent acetaldehyde dehydrogenase (r_2115) to restore redox balance and permit the anaerobic growth at differing levels among the strains as shown previously (Vargas et al, 2011, Scott et al, 2021b). Moreover, the simulations pointed to many of the same reaction fluxes such as aspartate -semialdehyde dehydrogenase (6.33 × 10 −4 , 1.05 × 10 −3 , 1.22 × 10 −4 , and 9.36 × 10 −4 mmol/mmolH for Opale, R2, Elixir, and Uvaferm respectively), homoserine dehydrogenase (6.33 × 10 −4 , 1.05 × 10 −3 , 1.22 × 10 −4 , and 9.36 × 10 −4 mmol/mmolH for Opale, R2, Elixir, and Uvaferm respectively), and glycerol-3-phosphate dehydrogenase (7.10 × 10 −2 , 8.00 × 10 −2 , 6.31 × 10 −2 , and 7.95 × 10 −2 mmol/mmolH for Opale, R2, Elixir, and Uvaferm, respectively) (where mmolH is millimoles of consumed hexose x 100) responsible for strain-dependent behavior as demonstrated in recent studies (Henriques et al, 2021b, Scott et al, 2021b). This result underscores the tight link between glycolysis, the TCA cycle, and amino acid metabolism under nitrogen-limited conditions.…”

“…In this period, the cellular objective was the maximization of biomass. In contrast with the previous model, during carbohydrate accumulation (previously growth under nitrogen limitation), the cellular objective of protein maximization and the procedure to simulate protein turnover (described in (Henriques et al, 2021b)) were activated. Also, to simulate carbohydrate accumulation, during this period, an exchange flux for this compound (s_3717[c]) was added to the stoichiometric network determined by the equation: where X carb (g/L) is the carbohydrate quantity present in the biomass, X is the biomass (g/L), θ carb is the final carbohydrate content and τ carb is the parameter controlling the convergence speed towards θ carb (see complete equations in supplementary code example (dynamic genome-scale modeling of yeast fermentation): https://sites.google.com/site/amigo2toolbox/examples).…”

“…These higher alcohols and their respective acetate esters are produced predominantly via the Ehrlich pathway (Ehrlich, 1907, Hazelwood et al, 2008) or de novo synthesis routes (Gonzalez and Morales, 2017). Higher alcohols have been shown to be produced through this route to regenerate NADH and preserve the redox balance in yeast cells (Lambrechts and Pretorius, 2000, Jain et al, 2012, Henriques et al, 2021b). The Ehrlich pathway consists of three steps: a transamination step, a decarboxylation step, and a reduction step.…”

“…Yeast8 has been combined with constraint-based modeling approaches to understand yeast metabolism in a few studies. However, the use of this yeast GEM has thus far been confined to glucose-limited, aerobic conditions (Moreno-Paz et al, 2022), nutrient-rich cases (Henriques et al, 2021b) and/or under non-transient model scenarios (Scott et al, 2021b), thus limiting its scope and applicability to accurately predict enological conditions.…”

“…In this study, the Yeast8.5.0 GEM (Lu et al, 2019), along with the dFBA framework (Henriques et al, 2021b), were employed to predict the metabolic behavior of Saccharomyces cerevisiae under enological conditions. The model was mainly used to model nitrogen-limited, anaerobic growth of S. cerevisiae with appropriate kinetic constraints, and utilizing a biomass equation tailored for enological conditions.…”