Signaling And Metabolic Coupling Model For2-KGA Mixed Fermentation

China is the most important country in vitamin C production with thecontribution of over80%of the world supply. The two-step fermentationprocess is commonly used. L-sorbose is converted into2-keto-L-gulonic acid(2-KGA) in the second step, via mixed culture of Bacillus megaterium andKetogulonicigenium vulgare (also called big bacterium and small bacterium).Due to the nonlinear characteristic of fermentation process and complexinteraction in mixed culture system, the modeling of the second-step becomesvery difficult and only little references may be found in the literature.In this paper, the interaction between big bacterium and small bacteriumis summarized based on the most recent literatures:(1) big bacterium mayadjust its growth behavior by quorum sensing;(2) metabolites and autolysissubstances of big bacterium are believed beneficial to overcome themetabolic defects of the small bacterium and therefore accelerate its growth;(3) the small bacterium releases lysozyme to promote the big bacterium’sautolysis;(4) the big bacterium releases specific protease substances toenhance the sorbitol dehydrogenase (SDH) which may increase the synthesisrate of2-KGA. After that, a kinetic model of2-KGA mixed fermentation isestablished. Model validation is carried out with four sets of experimentaldata under different cultivation conditions. Results demonstrate that theproposed model is able to well describe the growth of two bacteria and the2-KGA production.With the development of genetic and metabolic engineering, themodeling of fermentation process showes some micro-and multi-scale tendency. On the basis of previous work, a simplified carbon sourcemetabolic network of small bacterium is constructed. With the help of fluxbalance analysis method (FBA), a metabolic network model of smallbacterium is established. The model describes small bacterium’s growthprocess on the metabolic level. Although we only have some substrate andproduct data, on the reference of other’s experiment description and previousliterature knowledge, the scope of the model parameters can be roughlydetermined. The small bacteria’s specific substrate absorption rate and theflux distribution of each metabolic branch can be obtained by this model. Themodeling system can be expanded: on the one hand, the metabolic networkmodel of big bacteria can be combined in; on another hand, if somemetabolic level data was available, we can build more complex metabolicmodel to study some micro-phenomenon mechanism.Mechanism modeling of fermentation process can be used for predictivecontrol, process optimization, fault batches of early diagnosis and so on. Thepreviously established kinetic model for2-KGA mixed culture is tested withthe data of80industrial batches. Based on sensitivity analysis, it is found thatsome insensitive parameters may take fixed values to minimize thecomputing time. Then, the model is used to predict the most important statevariables, i.e., the substrate and the product concentration. Moving datawindow technique and rolling parameter identification are used in theprediction process.4h and8h ahead prediction errors for2-KGAconcentration can all be limitied in5%. The results showes that the model hasits potential application value.