Investigating metabolic burden and stress responses in prokaryotic model organisms through an integrated computational and synthetic biology approach

Abstract

Expression of heterologous proteins in bacteria is commonplace in research and industry across many disciplines including medicine, agriculture, and civil engineering. Overexpression of proteins may result in greater yields; however, it can induce a stress response from metabolic burden which can reduce yields by mechanisms like cell death or sporulation. This thesis investigated the stress responses in prokaryotic model organisms and employed a combined approach of computational methods and synthetic biology to develop systems to detect these stresses. Furthermore, this project aimed to work toward portable systems - the ability to transfer a system from one species to another with minimal refactoring. This work comprised of three major components: 1. A condition specific biomarker selection algorithm that takes an input of regulation data under different conditions and generates sets of biomarkers that best classify the selected condition versus all other conditions. This thesis thoroughly examines this algorithm through testing on diverse datasets including tiling array and RNA-seq regulation data. 2. A genetic logic gate was designed to use stress-specific biomarkers as inputs to produce a measurable output to detect the stress response. A testing system was developed to work between two species and early characterisation of this system was completed. 3. in vivo testing of a cross-species codon optimisation algorithm. A framework for testing coding sequences generated by a cross-species codon optimisation algorithm has been developed for this study. In summary, this research explored approaches to detect stress responses in prokaryotic model organisms and worked toward making systems portable without the need to refactor between species. The approaches developed formed a foundation for combined computational and synthetic biology approaches.