Firstly, we performed a sensitivity analysis, i.e. how biomass production rate changed as the flux over a specific reaction of interest varied in magnitude. The target reactions to perform this analysis were those involving the exchange of essential and additional growth sources used in the FBA simulations described in the previous section. We also analyzed the effect of oxygen uptake since the metabolic inference from the two cockroach endosymbiont genomes CDK inhibitor indicates the presence of a complete electron transport chain terminated with a high-affinity cbb 3-type cytochrome oxidase [1, 2]. Furthermore, the cockroach fat body, the tissue where

endosymbionts are located, exhibits the characteristics of an active aerobic environment (e.g. peroxisome

abundance and urate catabolism, [23, 1] and references therein). Both the iCG238 and the iCG230 models, showed a strict dependence on the import of L-Asn, Gly and L-Pro, in accordance with the metabolic inference selleck from the genomes [1, 2]. Our simulations using Bge model show that there is a range of metabolic flux values for oxygen and L-Gln exchange reactions over which it is possible to produce an optimum phenotype in terms of biomass (Fig. 5). A similar result was observed for the growth dependence on L-Gln with the Pam model (data not shown). Figure 5 Effect of oxygen and L-Gln uptake on metabolic network performance. Biomass production rates (mmol g DW-1 h-1) by the Bge strain model were measured at different uptake rates of oxygen (left) and L-Gln (right). We also evaluated the sensitivity of the Bge metabolic network to variations in the

three first reactions of the TCA cycle, absent in the metabolic network of the strain Pam ([2]; see Fig. 1). We simulated the minimal conditions and those considering the additional uptake of some intermediates of the cycle as well as the anaplerotic amino acids L-Glu and L-Asp, precursors of 2-oxoglutarate and oxalacetate, respectively. As shown in Figure 6, a viable phenotype is produced even when the flux values Phospholipase D1 through the three aforementioned reactions are null. Moreover, the biomass production reaches a maximum value when the flux across such reactions is zero and 2-oxoglutarate or L-Glu is added. Figure 6 Sensitivity analysis for the first three reactions of the TCA cycle. Biomass production rates (mmol g DW-1 h-1) by the Bge strain model were measured under different metabolic environments (minimal conditions or the uptake of the indicated metabolites, see inset) and diverse reaction flux through the first enzymatic steps of the TCA cycle: citrate synthase, aconitase and isocitrate dehydrogenase. Finally, we also explored the robustness of both metabolic networks by randomly removing genes.