Analyses

Comparison of the enzymatic redundancy between several models

In a first analysis, we sought to evaluate the extent of the functional redundancy (pathway redundancy plus genetic redundancy) providing robustness within the metabolic network of R. solanacearum. Indeed, this prior analysis is required to further deconvoluate if there is some specificity of the metabolic network of Ralstonia solanacearum in term of robustness before analysis of the virulence regulatory network contribution. Hence, we quantified the genetic redundancy of enzymes, i.e. the proportion of the reactions that can be catalysed by alternative enzymes, including alternative protein complexes harbouring at least one different subunit. We found that 26% of the reactions are catalysed by redundant enzymes. This level of enzyme redundancy is similar to the level reported in other bacteria like E. coli and Bacillus subtilis, 31% and 30% respectively, or P. aeruguinosa, 22 %. Next, we predicted the robustness R(proliferation,πi) of the biomass production, later referred as the proliferation phenotype, in front of internal perturbation (πi). We thereafter refer as internal perturbation any failure of internal network components leading to loss of function, like gene facing deleterious mutations, stochastic expression, or even enzymes facing loss of activity due to miss folding or inhibitions by chemicals. The simulations were carried out using the genome-scale metabolic models of R. solanacearum, E. coli and P. aeruginosa. To apply a similar genetic perturbation to each organism we determined groups of orthologs shared by the three species and present in each metabolic model using the INPARANOID algorithm. Hence, the robustness of the proliferation phenotype of the three bacteria growing with D-glucose and L-glutamate as sole source of carbon was calculated. To do so, we run a BECO analysis and collected predictions for the 421 single gene knock-out perturbations corresponding to each of the 421 orthologous families identified. The R(proliferation,πi) of R. solanacearum in D-glucose was found to be 0.58 whereas a value of 0.65 and 0.66 was found for E. coli and P. aeruginosa, respectively. The higher robustness observed for E. coli and P. aeruginosa was tracked and founded to be mainly due to differences in the biomass equation. Hence, the analysis showed that the contribution of the genetic redundancy to sustain the proliferation phenotype is almost similar in R. solanacearum and E. coli (0.10 and 0.12, respectively) whereas the contribution of pathway redundancy is only slightly lower (0.20 versus 0.25 in E. coli). A similar pattern was observed after simulation of growth in L-glutamate.