(194ae) Multi-Paradigm Multi-Scale Metabolic Modeling of a Nitrogen Fixing Cyanobacterium with Two Distinct Metabolic Modes

Trichodesmium erythraeum is a marine dwelling nitrogen fixing cyanobacterium which is responsible for up to 42% of the annual biological nitrogen fixation. Nitrogenase, the enzyme which converts ambient nitrogen to ammonia, is poisoned by oxygen; cyanobacteria, which produce oxygen as a product of the water splitting reaction in photosynthesis, must segregate the two reactions. One approach is to create specialized cells (called heterocysts) that keep the microenvironment anaerobic and another is to fix nitrogen at night, when metabolism is respiratory. T. erythraeum is unique in how it handles this challenge: cells of similar structure to the rest of the filament carry out nitrogen fixation with a specialized metabolism (no photosystem II, carbon source is glycogen). Thus, the filament grows with two unique metabolic modes simultaneously. In order to model this organism more accurately the traditional flux balance analysis, we have developed a multi-scale multi-paradigm metabolic model which accounts for the diffusion of metabolites, spatial distribution and cell cycle.

The basis of this model is a genome-scale metabolic reconstruction of T. erythraeum which includes 971 reactions, 986 metabolites, and 647 unique genes and a biomass formation equation was based on experimental evidence collected in our laboratory. Constraints were developed for each specific cell type based on the differences in metabolism reported in literature. This general model was then used to predict fluxes for exponential growth and also used in dFBA to predict equilibrium compositions. Without any additional constraints, the dFBA model predicts an equilibrium composition of 15.5% diazotrophs which agrees well with published in situ data which report filaments are 10-20% diazotrophs. Moreover, the model predicts that nitrogen leakage is an essential condition of optimality for T. erythraeum; cells leak approximately 29.4% total fixed nitrogen when growing at the optimal growth rate, which agrees with values observed in situ. These results will be compared to predictions from the multi-scale multi-paradigm metabolic model we have created in an agent based framework. We will describe how placement in the filament and distance from the nitrogen fixing cells affects the metabolism of each cell. How this modeling framework can then be applied to more complex consortia will also be discussed.