UC Davis

The economic contributions from yeast (Saccharomyces cerevisiae) alcoholic fermentation are vital to the Californian economy by contributing to over $43.6 billion in sales in 2019. Alcoholic fermentation is a dynamic and intricate process where yeast cells are exposed to many changing stress conditions such as hyperosmotic shock, nutrient limitation, temperature variations, and ethanol toxicity. Environmental conditions and genetic traits dictate yeast's metabolic activities, thus significantly impacting the overall fermentation performance. Two crucial metabolic activities relevant to wine fermentations, nutrient utilization efficiency (NUE) and volatile organic compound (VOC) (i.e., aroma compounds) formation, are poorly understood. Therefore, it is essential for fermentation industries as well as beneficial for the broader scientific community to investigate the pathways responsible for regulating NUE and production of VOCs. Overall, this dissertation provides insights for improving the quality and efficiency of VOC production by improving the understanding of yeast cellular metabolism.To gain necessary insight about metabolic mechanisms, several methods rooted in systems biology were used in this work. This dissertation focuses on curating and expanding current yeast genome-scale metabolic models (GSMMs) and employing the GSMM to understand the underlying metabolic differences regarding aroma production under two different growth conditions. Furthermore, the advances and remaining limitations of this expanded GSMM to simulate the metabolism and production of aroma impact molecules under enological conditions were explored. In a subsequent study in the dissertation, quantitative analysis of key VOCs and nitrogenous compounds from various commercial yeast strains over multiple time points at different phases of fermentation was performed. This analysis provided a unique insight into these strains' specific aroma production profiles and how they change over time and between strains. Multivariate statistics coupled with genome-scale modeling were used to identify several essential amino acids as well as their metabolic routes known to be involved in forming essential VOCs that are desired in wines. Consequently, it was determined that glycine, tyrosine, leucine, and lysine, were positively correlated with fusel alcohols and acetate esters concentrations during wine fermentation. In addition, these nitrogen utilization routes govern strain-specific variation in aroma profiles. In the following work, an experimentally- validated genome-scale dynamic flux balance analysis model was developed that can be used to predict fermentation kinetic behavior under enological conditions. The model was validated using experimental data for fermentation profiles and final concentrations of fermentation products of Enoferm T306 strain yeast grown under low nitrogen conditions (123 mg/L) in synthetic minimal must media (MMM). Furthermore, it was observed that not only are the biomass coefficients within the biomass equation crucial for accurate model predictions using dFBA, but so are the kinetic constraints, especially the maximum sugar uptake rate (υ_smax) and production yield of ethanol (f_ethanol). Finally, a metabolic modeling approach (random flux sampling) was applied to analyze and compare the various intracellular metabolic flux states of commercial yeast strains during enological fermentation. The intracellular predictions show qualitative agreement with the specific variation found from performing principal component analysis. These results indicated which gene-associated reactions are responsible for key strain variations. From the fluctuating degrees of variation among the strains from gene-associated reactions, a comparison of differences in GPR activity was evaluated. Thus, probable gene similarity could be inferred among the strains, and targets could be explored for genetic modification.