A structured model for penicillin production on mixed substrates

A structured kinetic model previously developed to describe the growth, differentiation, and penicillin production of Penicillium chrysogenum has been enhanced and extended in order to apply it to a mixed carbon source fermentation. The filamentous hyphae are divided into four distinct regions on the basis of their activities and the physiological structure ( i.e., vacuolation) of the hyphal compartments: viz., actively growing (mainly apical) regions, non-growing or penicillin producing regions, vacuoles, and degenerated or metabolically inactive regions. A simple approach is taken to give quantitative descriptions of hyphal extension, branch formation, vacuolation and differentiation. The fermentation medium contained glucose and lactose monohydrate as the main carbon sources. The source of the lactose was whey powder used in excess in the inoculum medium, whilst glucose was fed continuously throughout the fermentation. Lactose, a disaccharide, is hydrolysed to two monosaccharides, glucose and galactose, when the residual glucose concentration in the medium drops to a very low level. The utilisation of glucose and that of galactose following the hydrolysis of lactose were observed to occur simultaneously. This allowed the assumption of simple lactose utilisation kinetics in which lactose hydrolysis could be considered as producing an equivalent amount of glucose. The model has been used for successful predictions of fed-batch penicillin fermentations using an industrial P. chrysogenum strain under different glucose feed rates. Quantitative information on proportions of the hyphal regions was obtained from image analysis measurements and the parameters of the model were identified. When the glucose feed rate to the production culture was switched between a high and a low value, the model successfully predicted the dynamic changes of differentiation and the resulting penicillin production caused by the variations in the nutrient conditions. The use of image analysis to characterise differentiation as a basis for structured modelling of the penicillin fermentation appears to be very powerful, and such models have great potential for use in process simulation and control of antibiotic fermentations. (C) 1998 Elsevier Science S.A. All rights reserved.