An optimized repeated-fed-batch fermentation process for the synthesis of dihydroxyacetone (DHA) from glycerol utilizing Gluconobacter oxydans is presented. Cleaning, sterilization, and inoculation procedures could be reduced significantly compared to the conventional fed-batch process. A stringent requirement was that the product concentration was kept below a critical threshold level at all times in order to avoid irreversible product inhibition of the cells. On the basis of experimentally validated model calculations, a threshold value of about 60 kg x m(-3) DHA was obtained. The innovative bioreactor system consisted of a stirred tank reactor combined with a packed trickle-bed column. In the packed column, active cells could be retained by in situ immobilization on a hydrophilized Ralu-ring carrier material. Within 17 days, the productivity of the process could be increased by 75% to about 2.8 kg x m(-3) h(-1). However, it was observed that the maximum achievable productivity had not been reached yet.

No abstract

The influence of the product inhibition by dihydroxyacetone (DHA) on Gluconobacter oxydans for a novel semi-continuous two-stage repeated-fed-batch process was examined quantitatively. It was shown that the culture was able to grow up to a DHA concentration of 80 kg m(-3) without any influence of product inhibition. The regeneration capability of the reversibly product inhibited culture from a laboratory-scale bioreactor system was observed up to a DHA concentration of about 160 kg m(-3). At higher DHA concentrations, the culture was irreversibly product inhibited. However, due to the robust membrane-bound glycerol dehydrogenase of G. oxydans, product formation was still active for a prolonged period of time. The reachable maximum final DHA concentration was as high as 220 kg m(-3). The lag phases for growth increased exponentially with increasing DHA threshold values of the first reactor stage. These results correlated well with fluorescence in situ hybridization (FISH) measurements confirming that the number of active cells decreased exponentially with increasing DHA concentrations.

For the mathematical description of the semicontinuous two-stage repeated-fed-batch fermentation of dihydroxyacetone (DHA), a novel segregated model incorporating transient growth rates was developed. The fermentation process was carried out in two stages. A viable, not irreversibly product-inhibited culture was maintained in the first reactor stage until a predetermined DHA threshold value was reached. In the second reactor stage, high final product concentrations of up to 220 g L(-1) were reached while the culture was irreversibly product-inhibited. The experimentally observed changes of the physiological state of the culture due to product inhibition were taken into account by introducing a segregation into the mathematical model. It was shown that the state of the cells was dependent on the current environment and on the previous history. This phenomenon was considered in the model by utilizing delay time equations for the specific rates of growth on the primary and the secondary substrate. A comparison with reproducible measurements gave a good correlation between computation and experiment. The mathematical model was validated using independent own experimental data. A comparison with a stationary and nonsegregated model demonstrated the essential improvements of the novel model. It was deduced from the model calculations that high product formation rates of 3.3-3.5 g L(-1) h(-1) as well as high final DHA concentrations of 196-215 g L(-1) can be obtained with a residual broth volume in the first reactor stage of 2% and a DHA threshold value in the range of 100-110 g L(-1).

No abstract