(285e) Multiscale Multiobjective Systems Analysis (MIMOSA): An Advanced Metabolic Modeling Framework for Complex Systems

Photosynthetic microorganisms have the potential to be a renewable source of a wide variety of fuels and chemicals. Unfortunately, their potential has not yet been realized due to the complexity of these organisms and a general lack of tools when compared to the premiere model organisms such as E. coli and yeast. One tool that has proven extremely effective in production strain design is metabolic modeling but unfortunately, current metabolic modeling approaches are not able to accurately predict carbon fluxes in photosynthetic microorganisms because they are overly simplistic, rely on a single objective function (such as maximize biomass), and do not account for environmental effects such as light limitations or nutrient diffusion. To address these limitations, we have developed a new modeling approach, MIMOSA. MIMOSA is a multi-scale, multi-paradigm modeling approach that can track individual cells in a population, account for differences in the environment (such as nutrient or light availability) and allows for multi-objective optimization. We have used MIMOSA to investigate nutrient cycling in the marine nitrogen fixing cyanobacterium, Trichodesmium erythraeum. The ocean is an extremely nutrient deplete environment and T. erythraeum has a suite of metabolic reactions that enable nitrogen and carbon fixation and storage as well as phosphorus storage and utilization, which provides it a distinct advantage in the marine environment. Our simulations have illustrated some of the mechanisms through which cells control their environment, including the capacity to increase local concentrations of dissolved carbon up to 100% over a 12-hour light period. Integration of diffusion and cell physiology also justify the role of metabolic oxygen sequestration methods, like Mehler’s reactions, to reduce intracellular oxygen concentration and halt inhibition of nitrogenase through poisoning. It was effective in predicting when nitrogenase was limited when more energetically favorable nitrogen compounds like nitrate were available, yielding 0% to 55% inhibition of nitrogenase between 0 and 10 M NO₃^-, consistent with values reported through experimentation. Overall, the model is effective in predicting the confluence of metabolism and physiology in governing whole-population behaviors.