The major targets for improvement of recombinant expression systems in microbial cells are gene dosage, transcriptional control machinery and, to some extent, translation. Here we show that optimization of fermentation conditions by applying statistically designed, multifactorial experiments offers an additional method for potential enhancement of gene expression systems. A chromosomally encoded fusion between the Bacillus subtilis aprE regulatory region and the E. coli lacZ gene carried by the B. subtilis host cells was used. The 2 x SG sporulation medium was used as a basal medium. Among the 11 fermentation factors we examined, the most significant variables influencing beta -galactosidase expression were statistically elucidated for optimization and included peptone, MgSO4. 7H(2)O, and KCI. The optimum concentrations of these variables were predicted by using a second-order polynomial model fitted to the results obtained by applying the Box-Behnken design, a response surface method. Calculated optimum concentrations were predicted to confer a maximum yield of 2,423.5 beta -galactosidase specific activity units. A verification experiment performed under optimal conditions yielded 96% of the predicted specific activity value with an increase by a factor of almost 5 compared with the results obtained under basal conditions.