A systems biology understanding of protein constraints in the metabolism of budding yeasts

Abstract: Fermentation technologies, such as bread making and production of alcoholic beverages, have been crucial for development of humanity throughout history. Saccharomyces cerevisiae provides a natural platform for this, due to its capability to transform sugars into ethanol. This, and other yeasts, are now used for production of pharmaceuticals, including insulin and artemisinic acid, flavors, fragrances, nutraceuticals, and fuel precursors. In this thesis, different systems biology methods were developed to study interactions between metabolism, enzymatic capabilities, and regulation of gene expression in budding yeasts. In paper I, a study of three different yeast species (S. cerevisiae, Yarrowia lipolytica and Kluyveromyces marxianus), exposed to multiple conditions, was carried out to understand their adaptation to environmental stress. Paper II revises the use of genome-scale metabolic models (GEMs) for the study and directed engineering of diverse yeast species. Additionally, 45 GEMs for different yeasts were collected, analyzed, and tested. In paper III, GECKO 2.0, a toolbox for integration of enzymatic constraints and proteomics data into GEMs, was developed and used for reconstruction of enzyme-constrained models (ecGEMs) for three yeast species and model organisms. Proteomics data and ecGEMs were used to further characterize the impact of environmental stress over metabolism of budding yeasts. On paper IV, gene engineering targets for increased accumulation of heme in S. cerevisiae cells were predicted with an ecGEM. Predictions were experimentally validated, yielding a 70-fold increase in intracellular heme. The prediction method was systematized and applied to the production of 102 chemicals in S. cerevisiae (Paper V). Results highlighted general principles for systems metabolic engineering and enabled understanding of the role of protein limitations in bio-based chemical production. Paper VI presents a hybrid model integrating an enzyme-constrained metabolic network, coupled to a gene regulatory model of nutrient-sensing mechanisms in S. cerevisiae. This model improves prediction of protein expression patterns while providing a rational connection between metabolism and the use of nutrients from the environment. This thesis demonstrates that integration of multiple systems biology approaches is valuable for understanding the connection of cell physiology at different levels, and provides tools for directed engineering of cells for the benefit of society.