Modeling of Biochemical System Theory with Applications to Recombinant Escherichia coli Fermentation Processes

• Main Authors: Chih-lung Ko, 柯誌龍
• Other Authors: Feng-Sheng Wang
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/04779389393904287397

博士 === 國立中正大學 === 化學工程所 === 96 === In this study, we introduced Biochemical Systems Theory (BST) modeling technique. Structured and unstructured modeling methods are discussed respectively. First of all, we discussed how to establish a suitable model to describe the dynamic characteristic of microbial growth in the unstructured model. Then, the optimal models are applied to the development and optimization of fed-batch bioprocesses, and design on-line state estimator for the system monitoring. In addition, we proposed structured techniques which make parameters of model not only be able to describe the dynamic characteristics, but also be able to meet the system sensitivity. First, we analyze the growth dynamics and production of an industrially interesting amino acid in a recombinant E. coli strain, using in parallel two alternative modeling frameworks, namely S-systems and lin-log models. Other, a traditional Monod model is applied to describe the dynamic behaviors and it compared to with them. Among the S-system models, several alternatives representing slightly different pathway structures, were tested. All were found to be capable of capturing the dynamics of all variables in the test systems and also of predicting the dynamic responses under new conditions. The lin-log models also captured the dynamics, but not as well as the S-system models. A probable reason for the inferior performance of lin-log models is their intrinsic property of not representing situations well where variable concentrations are moderately small. However, the Monod’s model could only predict the growth characteristic of the recombinant E. coli, qualitatively. Then, we take S-system model to design operation condition of fed-batch bioprocesses. A two-phase design approach is introduced to determine the optimal feed rate, fed glucose concentration and fermentation time for maximizing protein productivity using recombinant Escherichia coli BL21 (pBAW2) strain. The first phase is applied to determine a raw S-system kinetic model with using batch time-series data. Two runs were carried out in the second phase to achieve the maximum protein productivity for the fed-batch fermentation process. The computational results using the second S-system kinetic model are more satisfactory to the experiments than those by using the kinetic model obtained from batch time-series data. To cross validation, two extra fed-batch experiments with different feed strategies were carried out to compare with the optimal fed-batch result. This approach could improve at least 3% productivity from the experimental comparison. And then, we take models to design on-line state estimator of fermentation processes. One of the difficulties encountered in the control of bioprocesses is the lack of reliable on-line sensors for their key state variables. On-line measurements of the recombinant biomass and intracellular protein are the key issues necessary to achieve an effective process operation. This section investigates the suitability of using on-line estimation to predict concentrations of biomass and protein for recombinant E. coli cultivated in batch and fed-batch modes. Various pairs of on-line accessible measurements of glucose, lactate, and acetate are fed back into the on-line estimator to train the tuning parameters in order to predict concentrations of biomass and intracellular protein. Such predictions are consistent with experimental data using off-line measurements. Because the life scientific analysis technology is high-developed, it enters the microscopic viewpoint time of the reacted research for metabolism. However it is complex relationship between metabolic pathways inside of the cell, best researching strategy is to establish mathematical model. To ensure suitable structured model becomes quite important. In this study, we established metabolic reacting model by using biochemical systems theory. It was structured by power law way, however the best parameter of that model is uncertain to reflect the physical property of metabolic pathway. If using this model to carry on the control analysis or the sensitivity analysis, the result might not be quite exactly. Therefore, we introduced the metabolism control analysis (metabolic control analysis, MCA) connectivity theory, it is mainly for the connection between the description metabolite concentration, the enzyme concentration and the flux relates. The connectivity theory was set as constraint condition in the parameter estimated process. In that progress, it doesn’t only obtain the model value with the experimental value consistent result, but also guarantee the best parameter being truly represented for the performance system physical characteristic.