Reconstruction And Analysis Of Genome-scale MetabolicNetwork Of Brady Rhizobium Japonicum USDA110

Symbiotic nitrogen fixation is an important part of the nitrogen biogeochemical cycles and the main source of nitrogen in the biosphere. Rhizobium-host plant symbiotic nitrogen fixation system is used as the model system to study symbiotic relationship.Bradyrhizobium japonicum USDA110has strong symbiotic nitrogen fixation ability and iswidely used in the fields of agriculture and environmental protection. It is the representative strain of Bradyrhizobium and one of the model organisms in symbiotic nitrogen fixation research.Todeepenthe understandingof symbioticnitrogen fixationfrom thesystem leveland metabolicperspective, it is helpful and meaningful to reconstruct, simulate and analyze B. japonicum USDA110genome-scale metabolic network.Based on the genome-scale metabolic network model of Sinorhizobium meliloti1020, iHZ565, combined with metabolic data set of B. japonicum USDA110in KEGG database, thedraft of B. japonicum USDA110genome-scale metabolic network was obtained. Then the draft was refined by the integration of the information from database, online tools and softwareprediction and literatures, and the genome-scale metabolic network of B. japonicum USDA110was reconstructed. This model, named as iYY1065, includes1053reactions and663metabolites, involving1065genesthat account for12.8%of total genes of B. japonicum USDA110. By adding condition-dependant objective function and simplest substrate input set, specific metabolic network models that reflect three specific physiological conditions-free living (FL), symbioticnitrogen fixation with branched-chain amino acid transport (SNF1) and symbioticnitrogen fixation without branched-chain amino acid transport (SNF2)—were established, respectively.Flux balance analysis (FBA) was performed to simulate physiological states of free living and symbioticnitrogen fixation. Using iYY1065model, utilization ability of42carbon sources and9nitrogen sources waspredictedand compared with experimental results. Between these two kinds of result,60.8%were consistent,23.5%were weakly consistent,15.7%were conflicting.Overlap between reaction-associatedgene set of three specific metabolic network models and transcriptomedata were75%(FL),42%(SNF1), 45%(SNF2), respectively.Basic metabolicpathwayanalysis, KEGG pathwayanalysis, metabolic pathwayflux analysis, analysis ofsymbiotic nitrogen fixationphysiologicalindices, robustness analysis andgene knockoutanalysis were performed for the threespecific metabolic network models.233and20essential genes were detected for FL and SNF statesrespectively by single gene knockout.123and21essential genes pairs were detected for FL and SNF statesrespectively by double gene knockout. Integrating all this analysis, we obtained twokey assumptions: branched-chain amino acid transport from host plant to bacteroid which is essential to fast-growing rhizobia is non-essential to Bradyrhizobium; SNF metabolic network is more robust than FL metabolic network.