A comprehensive metabolic network that considers 147 reaction fluxes and 105 metabolites is used in a mass-Aux-balance-based stoichiometric model for Bacillus licheniformis for serine alkaline protease (SAP) overproduction. The theoretical capacity analysis leading to optimized SAP overproduction was carried out by using a linear constrained optimization technique for several specific growth rates and the variation of the fluxes were calculated by fixing the sole carbon source citrate＇s uptake rate at 10 mmol/gDW/h. The theoretical data-based capacity analysis was conducted by using the model in combination with the off-line extracellular analyses of the dry cell and the metabolites that were citrate, organic acids, amino acids, and SAP; and the variations in the intracellular fluxes were obtained for the three periods of the batch bioprocess. The flux distribution maps of the analyses showed that the tricarboxylic acid cycle was active and the cells utilized the gluconeogenesis pathway, the pentose phosphate pathway, and the anaplerotic reactions; nevertheless, the glyoxylate shunt and the glycolysis pathway were inactive. The theoretical capacity analysis showed that SAP synthesis flux increased with the decrease in the specific growth rate, and was the highest at mu = 0 h(-1) as 0.0260 mmol/gDW/h. Both in the theoretical capacity and the theoretical data-based capacity analyses, among the fluxes towards the amino acid groups, aspartic acid group had the highest value and aromatic acid group had the lowest flux value; the flux distributions are similar. The Aux values towards SAP was maximum in Period II, whereas it was minimum in Period I. In Period II of the theoretical data-based capacity analysis, the fluxes of alanine and valine are higher than the other amino acid fluxes; and the pyruvate branch point seems to be the potential metabolic engineering site. The results reveal that SAP production can theoretically be increased 1.09, 16.68, and 7.21 folds, respectively, in Periods I, II, and III. The diversions in the pathways and certain metabolic reactions depending on the bioprocess periods and potential strategies for improving SAP production are also discussed. (C) 2000 Elsevier Science Inc. All rights reserved.