Abstract:
A system-level analysis of Saccharomyces cerevisiae metabolism was performed through integration of stoichiometric modeling and high-throughput ‘omics’ data. A bridge between metabolic networks and transcriptomics was built by employing the reactions involved in central carbon metabolism of the baker’s yeast. The fold changes in controleffective fluxes (CEF), the weighted sum of calculated elementary modes passing through the reactions, were used for the prediction of the fold changes in mRNA transcripts of metabolic genes on different growth media (glucose-ethanol and galactoseethanol). An acceptable correlation was obtained between the theoretical CEF-based flux ratios and experimental mRNA level ratios of 38 genes. Applicability of the approach to mammalian cell metabolism through analysis of red blood cell enzymopathies was also demonstrated. CEF approach was then employed to investigate the transcriptional regulation of fluxes in yeast metabolism for carbon shifts from fermentative (glucose) to nonfermentative (ethanol, acetate, lactate) substrates. An acceptable correlation was obtained for the analysis of such perturbation experiments, indicating that fluxes of yeast central metabolism are mainly transcriptionally regulated when there is a shift in carbon source. An algorithm was developed to integrate metabolome data with metabolic network topology. The approach enables identification of reporter reactions, around which there are significant coordinated changes following a perturbation. Applicability of the algorithm was demonstrated for S. cerevisiae. Further combination of the results with transcriptome data enabled to infer whether the reactions are hierarchically or metabolically regulated. Model-based structural robustness of yeast metabolism was analyzed to guide the research on phenomics. In silico lethality information of gene deletions on different carbon substrates indicated a more robust metabolism for S. cerevisiae than for E. coli bacterium.
