Main content area

Dynamic modeling of Escherichia coli metabolic and regulatory systems for amino-acid production

Usuda, Yoshihiro, Nishio, Yosuke, Iwatani, Shintaro, Van Dien, Stephen J., Imaizumi, Akira, Shimbo, Kazutaka, Kageyama, Naoko, Iwahata, Daigo, Miyano, Hiroshi, Matsui, Kazuhiko
Journal of biotechnology 2010 v.147 no.1 pp. 17-30
DNA-directed RNA polymerase, Escherichia coli, acetates, batch fermentation, biotechnology, carbon, cyclic AMP, dynamic models, enzyme activity, fructose, glucose, glutamic acid, glycolysis, metabolites, operon, primary productivity, pyruvate dehydrogenase (lipoamide), ribosomes, simulation models, specific growth rate, transcription (genetics), transcription factors, tricarboxylic acid cycle
Our aim is to construct a practical dynamic-simulation system that can model the metabolic and regulatory processes involved in the production of primary metabolites, such as amino acids. We have simulated the production of glutamate by transient batch-cultivation using a model of Escherichia coli central metabolism. Kinetic data were used to produce both the metabolic parts of the model, including the phosphotransferase system, glycolysis, the pentose-phosphate pathway, the tricarboxylic acid cycle, the glyoxylate shunt, and the anaplerotic pathways, and the regulatory parts of the model, including regulation by transcription factors, cyclic AMP receptor protein (CRP), making large colonies protein (Mlc), catabolite repressor/activator (Cra), pyruvate dehydrogenase complex repressor (PdhR), and acetate operon repressor (IclR). RNA polymerase and ribosome concentrations were expressed as a function of the specific growth rate, μ, corresponding to the changes in the growth rate during batch cultivation. Parameter fitting was performed using both extracellular concentration measurements and in vivo enzyme activities determined by ¹³C flux analysis. By manual adjustment of the parameters, we simulated the batch fermentation of glucose or fructose by a wild-type strain (MG1655) and a glutamate-producing strain (MG1655 ΔsucA). The differences caused by the carbon source, and by wild-type and glutamate-producing strains, were clearly shown by the simulation. A sensitivity analysis revealed the factors that could be altered to improve the production process. Furthermore, an in silico deletion experiments could suggested the existence of uncharacterized regulation. We concluded that our simulation model could function as a new tool for the rational improvement and design of metabolic and regulatory networks.