Abstract
Pristinamycin biosynthesis using Streptomyces pristinaespiralis and date sirup (DS) as substrates was optimized before scale-up. DS was filter sterilized as heat sterilization primes Maillard reactions having negative effects on antibiotic production. Multilinear regression modeling (MLR) predicted optimum medium composition, specifying components with positive and negative effects on production. The MLR showed that to maximize bacterial growth, DS, arginine, CaCl
2
, and KH
2
PO
4
must be fixed at the highest concentration, but to maximize antibiotic production, these factors have to be fixed at a low level. A noticeable difference in productivity was observed in a shake flask experiments with 50.4 and 43.1 mg/L pristinamycin final concentration for the DS and the glucose substrates, respectively. In the 2 L bioreactor, the DS medium resulted in a 66.6 mg/L antibiotic, while the scale-up in the 100 L resulted in 39.0 mg/L. The low yield in the 100 L bioreactor could be attributed to the relatively high stirring rate applied which was the minimum possible in the bioreactor used. This high stirring rate prevented pellet formation by the cells, which is described as necessary for antibiotic formation by the bacterium. Hence, a successful scale-up to pilot-scale should consider the effect of stirring rate.