Abstract: Gamma-aminobutyric acid (GABA) is a natural non-protein functional amino acid, which has potential for fermentation industrial production by Lactobacillus brevis. This work investigated the batch fermentation process and developed a kinetic model based on substrate restrictive model established by experimental data from L₂₅(5⁶) orthogonal experiments. In this study, the OD₆₀₀ value of fermentation broth was fixed to constant after reaching its maximum because the microorganism death showed no effect on the enzyme activity of glutamate decarboxylase (GAD). As pH is one of the key parameters in fermentation process, a pH-dependent kinetic model based on radial basis function was developed to enhance the practicality of the model. Furthermore, as to decrease the deviations between the simulated curves and the experimental data, the rolling correction strategy with OD₆₀₀ values that was measured in real-time was introduced into this work to modify the model. Finally, the accuracy of the rolling corrected and pH-dependent model was validated by good fitness between the simulated curves and data of the initial batch fermentation (pH 5.2). As a result, this pH-dependent kinetic model revealed that the optimal pH for biomass growth is 5.6-5.7 and for GABA production is about 5, respectively. Therefore, the developed model is practical and convenient for the instruction of GABA fermentation production, and it has instructive significance for the industrial scale.

Key words: Kinetic modeling, Fermentation, Gamma-aminobutyric, Lactobacillus brevis, pH control, Rolling correction

摘要： Gamma-aminobutyric acid (GABA) is a natural non-protein functional amino acid, which has potential for fermentation industrial production by Lactobacillus brevis. This work investigated the batch fermentation process and developed a kinetic model based on substrate restrictive model established by experimental data from L₂₅(5⁶) orthogonal experiments. In this study, the OD₆₀₀ value of fermentation broth was fixed to constant after reaching its maximum because the microorganism death showed no effect on the enzyme activity of glutamate decarboxylase (GAD). As pH is one of the key parameters in fermentation process, a pH-dependent kinetic model based on radial basis function was developed to enhance the practicality of the model. Furthermore, as to decrease the deviations between the simulated curves and the experimental data, the rolling correction strategy with OD₆₀₀ values that was measured in real-time was introduced into this work to modify the model. Finally, the accuracy of the rolling corrected and pH-dependent model was validated by good fitness between the simulated curves and data of the initial batch fermentation (pH 5.2). As a result, this pH-dependent kinetic model revealed that the optimal pH for biomass growth is 5.6-5.7 and for GABA production is about 5, respectively. Therefore, the developed model is practical and convenient for the instruction of GABA fermentation production, and it has instructive significance for the industrial scale.

关键词: Kinetic modeling, Fermentation, Gamma-aminobutyric, Lactobacillus brevis, pH control, Rolling correction