Computational studies of industrial hosts for improved production of the 2nd generation of biofuels

In an effort of overcoming the limited availability of fossil energy resources and moving toward a sustainable economy, the focus of the research and development in the area of biofuels has shifted towards developing the 2nd generation of fuels that should be produced via microbial fermentation. The 2nd generation biofuels should satisfy several criteria such as lower emission, higher energy density and should be less corrosive to engines. Although for many of these molecules, natural producers are known, they are not produced in the appreciated quantities. Heterologous expression of biosynthetic pathways taken from natural producers or expression of de novo synthetic pathways into microbial workhorses such as E. coli allows for production of a wide spectra of biofuels. Recently, P. putida has emerged as an amenable production hosts with a number of advantages over natural producers. P. putida is a non-pathogenic soil bacterium known for its versatile metabolism. This highly adaptive bacterium has been found to survive and grow on a wide range of substrates from pure caffeine to toxic industrial waste. Moreover, P. putida is tolerant to high toxicity compounds such as 2nd generation biofuel butanol. Counterintuitively, P. putida was seldom used as a host for the production of biofuels. In this thesis, we performed a computational analysis of this organism to evaluate its metabolic capacities to serve as a potential 2nd generation biofuels production host. Its capacity was compared against heavily used host E. coli on the test example of production of one of the most prominent fuel candidate Methyl Ethyl Ketone (MEK). To this end, we first performed a thermodynamic curation of the genome-scale iJN1411 model of P. putida, and we then used redGEM and lumpGEM algorithms to derive a consistently reduced large-scale stoichiometric model of P. putida. We integrated different omics data into resulting models and we proposed a novel way of constraining concentrations of the same species across different compartments while maintaining the consistency with the experimental measurements. To assess its capability to serve as a host, we evaluated and analyzed more than 3.6 millions biosynthetic pathways for production of 5 MEK precursors, in both heavily used industrial workhorse E. coli and rising P. putida. We compared their capability and performance with respect to thermodynamic feasibility and yield and we identified the most promising pathways for MEK production. Beside the discovered and evaluated pathways, we present a new way of clustering of feasible pathways and pathway precursors that allows us to classify and evaluate alternative ways for production and to better understand chemistry that leads towards the target molecule. Identification of metabolic engineering targets for the improved biofuel production requires kinetic models. We used the ORACLE framework to generate a population of large-scale kinetic models of P. putida, and we employed these models in two studies. In the first study, for a wild-type strain of P. putida grown under aerobic conditions using glucose as a carbon source, we evaluated and validated the predictions of the generated kinetic models against a collection of experimental single-gene knockouts. In the second study, we analyzed the capacity of P. putida to adapt to increased energy demand, and we identified potential metabolic engineering targets for improved resistance of this organism to stress conditions.