Files in this item



application/pdfPETERSON-DISSERTATION-2017.pdf (65MB)
(no description provided)PDF


Title:A tale of two microbes: Computational investigations of biological processes in escherichia and methanosarcina at multiple scales
Author(s):Peterson, Joseph Ryan
Director of Research:Luthey-Schulten, Zaida
Doctoral Committee Chair(s):Luthey-Schulten, Zaida
Doctoral Committee Member(s):Chemla, Yann; Gruebele, Martin; Metcalf, William W.
Department / Program:Chemistry
Degree Granting Institution:University of Illinois at Urbana-Champaign
Computational biology
Stochastic physics
Genome scale modelling
Systems biology
Abstract:Biological phenomena, while grounded in the laws of physics and chemistry, often exhibit behaviours too complex to be ascribed to a single electron density, bond vibration or chemical reaction. Rather, the processes inherent to living organisms---adaptation, growth, homeostasis, metabolism, replication, and response to stimuli---result from concerted function of the physical and chemical interactions of thousands to trillions of molecules interacting with one another. In principle, the governing equations for a biological system could be written in terms of the fundamental physical and chemical laws; however, such an approach would suffer from the intractability in its solution and the incomprehensibility of the result; it would be difficult to see the forest from the trees. This leaves the possibility of study using either analytical studies of toy models or computation. Adopting the latter approach in this work, a subset of cellular processes including metabolism, growth, and response to stimuli are examined. A variety of modeling approaches are employed to capture phenomena at different scales. Using as model systems Methanosarcina acetivorans, an anaerobic methane producing archaeum adapted to niche environments, and Escherichia coli, a faculative anaerobic bacteria with diverse capabilities, the influence of extrinsic (environmental) and intrinsic (inherent) factors on the organisms' behaviours is examined. In Part I of the thesis, studies of Methanosarcina species are presented. A kinetic model of methanogenesis---the metabolic pathways unique to Archaea that produce methane---in M. acetivorans is developed and used to examine the sensitivity of methane production rates to abundances of methanogenesis proteins. Subsequently, the metabolism of M. acetivorans when grown on several substrates is examined using genome-scale metabolic modeling. Metabolic phenotypes, wherein the methanogens utilize metabolic pathways to different extents, were predicted by integrating RNA expression and half-life data with the models. Strikingly, it was shown that the organism adjusts RNA half-lives of nearly half of metabolic genes to optimize metabolic flux for different growth substrates. This discovery was the first to show such a global role for half-life in defining metabolic phenotype. Concomitantly, the metabolic model was corrected and expanded, especially in the context of the cell's compositional requirements, by adding new terms to the model's biomass equation. Two comparative genomic studies were subsequently undertaken, enabled primarily by the availability of this and other highly curated metabolic models. First, the genomes of all fully sequenced Archaea were mapped across the available metabolic models to examine conservation of metabolic function. This revealed that amino acid metabolic pathways relatively more highly conserved than coenzyme, lipid, nitrogen, and transport metabolism. Second, the metabolic models of several Methanosarcina species were mapped across the genomes of 30 Methanosarcina species, enabling a pan-reactome study of these metabolically diverse methanogens. By examining the resulting core-reactome in the context the conserved genome, knowledge gaps in the metabolism could be filled. Importantly, by examining the pan- and core-genome of the Methanosarcina, a biosynthetic pathway for methanophenazine, a methanogenesis cofactor, was hypothesized. In Part II of the thesis, causes of stochasticity and heterogeneity were examined in the model organism E. coli. Adopting a simulation technique designed to sample the chemical master equation noise in gene expression was examined. Inspired by the inability of traditional models (which neglected genome replication) to fit the distributions obtained using single-molecular fluorescence in situ hybridization experiments, the effect of genome replication on the noise observed in genes placed at different locations around the circular genome was examined. Simulation results indicated that relaxation of the RNA count from a pre- to a post-replication steady-state significantly affected the shape of the resulting distributions. Analytical results showed that the noise of a constitutively expressed gene could be completely defined by three variables: the location of the gene on the chromosome, the RNA half-life, and the cell doubling time. Overall, this showed that previous studies that neglected to handle genome replication explicitly could both qualitatively and quantitatively misinterpret experimental data. Finally, adopting a completely different modeling framework, metabolic cooperativity in E. coli colonies growing on agar surfaces were examined. Building upon previous work that identified acetate cross-feeding using reaction-diffusion partial-differential equations, the effects of strain specific differences in the metabolic capacities and geometrical confinement were examined. The behavior of five different E. coli strains were examined; it was found that the extent and timing of metabolic cross-feeding were significantly different, even for closely related strains. Finally, cross-feeding was found to only vary when the growth substrate had abrupt changes in geometry (e.g. a wall or pit), and that smooth changes caused imperceptible changes in growth.
Issue Date:2017-11-28
Rights Information:Copyright 2017 Joseph Ryan Peterson
Date Available in IDEALS:2018-03-13
Date Deposited:2017-12

This item appears in the following Collection(s)

Item Statistics