SSDesign: Computational metabolic pathway design based on flux variability using elementary flux modes

Metabolic pathway modification based on the stoichiometric model has been an effective approach for enhancing microbial bio-production. The network of optimal pathways for "growth-associated" and "non-growth-associated" production can be designed from the flux variability (solution space). The present study introduces a new computational method (solution space design [SSDesign]) that visually designs the gene knockout solution space. The smallest reaction sets that satisfy the mass balances of intermediates are called elementary flux nodes (EFMs). Because some of the EFMs necessarily occupy the outer boundary nodes of the flux solution space, the proposed SSDesign determines the area over which EFMs should be removed from the solution space of the parent strain, and explores the gene knockouts that will eliminate these undesirable EFMs. To evaluate the performance of SSDesign, the model was applied to growth-associated and non-growth-associated succinate production in Escherichia coli. In the growth-associated case, the deletion mutants that promoted succinate production at maximum biomass yield were predicted, and a candidate of ΔptsG ΔpykA,F ΔpflA has been experimentally confirmed as a succinate producer. Simply by changing the parameters, the gene knockout combinations yielding high growth yield were successfully predicted by SSDesign. In the non-growth-associated case, strong candidates for succinate production were the deletion mutants ΔpntAB ΔsfcA ΔpykA,F and ΔsfcA ΔmaeB ΔpykA,F Δzwf. According to the solution spaces, these strains allow high growth yield and inevitably produce succinate at zero biomass yield, since their metabolic pathways cannot sustain steady-state without discarding succinate from the cell.