2014 Zanariah Binti Hashim


Metabolomics-based analysis of gene-metabolite correlations in yeast transcription factor knockouts

Abstract of Thesis

Cellular functions are determined by integrative interactions between various constituents, i.e. genes, transcripts, proteins, and metabolites. Thus, it is important to study these interactions to understand the whole biological system. Genetic perturbations are often used to investigate the contribution of individual components. One of such components is a transcription factor. Transcription factors (TFs) are the regulatory proteins that interact with DNA to either promote or suppress gene expression. Due to the importance of TFs in gene regulation, they have been extensively studied and several databases dedicated to TFs were established. However, our understanding is still lacking since many regulatory events that link gene expression to the final phenotypic changes remain poorly characterized. In particular, systematic analyses of global metabolic alteration following a TF perturbation have been largely unexplored.

This thesis regards the effects of transcription factor-related single gene deletion towards metabolic alteration, using yeast Saccharomyces cerevisiae as a model system. Specifically, the correlations between TF gene and metabolites were investigated. In Chapter 1, general introduction and research background are presented. In particular, yeast transcription factors and metabolomics techniques are discussed. In Chapter 2, metabolic profiling of two representative transcription factors Rtg1 and Rtg3 yeast knockout is demonstrated as a proof-of-principle of the utility of metabolomics approach in finding novel TF-metabolite correlations. These two proteins are positive regulators of the mitochondrial retrograde response (RTG), a signaling pathway activated under repressed mitochondrial function. Using a widely-targeted metabolomics approach, polyamine biosynthesis and other amino acid metabolism were found to be significantly altered in RTG-deficient strains, apart from the expected TCA and glyoxylate cycles. A characteristic decrease of 2-oxoglutarate preceding the decreases of other TCA cycle intermediates in RTG disruptants suggests that 2-oxoglutarate may play a pivotal role in controlling the flow and balance of TCA/glyoxylate cycles under RTG response.

In Chapter 3, a global metabolome analysis was performed for 154 TF deletion strains. Characterization using hierarchical clustering analysis (HCA) and differential analysis showed that the strains can be categorized according to their metabolic phenotype, and both known and unknown correlations were demonstrated. Differential strains (strains with large differences in metabolite levels compared with wild-type) and strain clusters that share a highly similar metabolic profile were identified. About 30% of the strains were classified as differential, while 27 individual clusters consisting of differential and non-differential strains were observed. The comprehensive metabolome dataset presented here may serve as an input for deeper investigations into transcription factors. Also discussed are issues regarding data normalization and correction of batch-to-batch variation, a prevalent problem in mid- to large-scale metabolomics studies. Finally, in Chapter 4, general conclusion and future perspective are presented. It is expected that metabolomics will be routinely performed, whether as a primary or complementary means in many gene regulation studies.