In this work, the stoichiometric metabolic network of Escherichia coli
has been formulated as a comprehensive mathematical programming model
, with a view to identifying the optimal redirection of metabolic flux
es so that the yield of particular metabolites is maximized. Computati
on and analysis has shown that the over-production of a given metaboli
te at various cell growth rates is only possible for a finite ordered
set of metabolic structures which, in addition, are metabolite-specifi
c. Each regime has distinct topological features, although the actual
flux values differ. Application of the model to the production of 20 a
mino acids on four carbon sources (glucose, glycerol, lactate, and cit
rate) has also indicated that, for fixed cell composition, the maximum
amino acid yield decreases linearly with increasing cell growth rate.
However, when the cell composition varies with cell growth rate, the
amino-acid yield varies in a nonlinear manner. Medium optimization stu
dies have also demonstrated that, of the above substrates, glucose and
glycerol are the most efficient from the energetic viewpoint. Finally
, model predictions are analyzed in the light of experimental data.