| In this dissertation, aiming at improving porcine interferon-α(pIFN-α) fermentation performance by a recombinant Pichia pastoris, with the aids of using artificial intelligent, on-line process monitoring, multi-variables clustering, and metabolic/enzymatic analysis methods or techniques, pIFN-αefficient fermentation were carried out in 5 L and 10 L bioreactors. The major contents of this dissertation were focused on process optimization/control, fault diagnosis and metabolic regulation of pIFN-αfermentation, and the main results were summarized as below:(1) The previously proposed ANNPR-Ctrl (artificial neural network pattern recognition based control) system was successfully applied to the high cell density fed-batch cultivation processes of the two recombinant P.pastoris strains, producing pIFN-αand phytase respectively, and its effectiveness and universal ability were testified. The results indicated that, comparing with the traditional DO-Stat substrate feeding strategy, the ANNPR-Ctrl system could not only maintain the substrates (glycerol and glucose) concentration at low level to prevent the possible accumulation of the protein expression inhibitory by-metabolites, but also obtain a much higher specific growth rate in growth phase, cell concentrations of the two recombinant P.pastoris strains could increase 81% and 44% within the same cultivation period. The ANNPR-Ctrl system supplies a universal platform for realizing high cell density cultivation of recombinant P.pastoris strains.(2) The methanol concentration was on-line monitored and controlled by a methanol electrode, while simultaneously monitoring other important state variables reflecting cell metabolic activities, to seek the optimal induction conditions for pIFN-αproduction. The results indicated that the formation of high and stable OUR (oxygen uptate rate) in induction phase was the key factor for effective pIFN-αproduction. Achievement of high and stable OUR relied on simultaneously maintaining a moderate methanol concentration around 10 g/L in induction phase and controlling specific growth rate in glycerol transition phase at certain suitable level. Under the high and stable OUR environment (200-300 mmol/L/h), the highest pIFN-αantiviral activity could reach a high level of 6.6×106 IU/mL.(3) The effects of induction at low temperature on cell metabolic activity, pIFN-αproduction, transcriptional levels of the key enzymes in methanol metabolism were investigated. The results indicated that when shifting induction temperature from 30°C to 20°C: 1) the cells adaptation period in response to methanol induction environment was shortened for about 5 hours; 2) the methanol metabolism ability of cells was significantly enhanced, the specific AOX activity, specific methanol and oxygen consumption rates were 2-3 folds of those obtained at 30°C; 3) pIFN-αproduction was largely improved and the highest pIFN-αantiviral activity was raised about 10 folds in a 5 L fermentor; 4) the transcriptional levels of the key enzymes including the detoxifying enzyme in methanol metabolism increased 1.36-16.47 folds.(4) The foreign protein production by recombinant P.pastoris is a high oxygen consumption process. The underlying reasons of the unavailability of pIFN-αcontinuous accumulation under low induction temperature were analyzed and interpreted from the standpoint of methanol metabolism. Based on the analysis results, a combinational control strategy of low induction temperature and high dissolved oxygen concentration was proposed. With the combinational control strategy, the limitation in oxygen supply could be relieved and the oxidative phosphorylation be activated, resulting that a more efficient NADH consumption rate and ATP regeneration rate directing an enhanced carbon flow towards to pIFN-αsynthesis. As a result, the pIFN-αantiviral activity could be increased continuously, and pIFN-αactivity reached to a highest level of 3.62×107 IU/mL.(5) The availability and stability of the pIFN-αfermentation process could not be simply ensured by the control and optimization technique. A simple, straightforward and effective multivariable clustering analysis method and an early fault alarming/diagnosis system were proposed, and the feasibility of this system was testified by experiments. With pIFN-αantiviral activity as the evaluation index, this method classified fermentation runs into two groups:“normal”and“abnormal”, by and plotting and visualizing the multivariable data collected during induction phase in two dimensional (2D) planes in any possible and random combination marked with the specified symbols. The 2D-plane figures with good clustering characteristics were then selected to interpret the physiological features of“normal”and“abnormal”fermentations in growth phase backward, and to determine the prerequisite of the formation of“normal”fermentation and how the dynamic behaviors in this phase affect the control performance in the subsequent induction phase. Based on a resulting optimal“normal/abnormal boundary trajectory”, a simple fault detection/ alarming system was proposed, and its effectiveness was verified using two testing data sets (one“normal”and one“abnormal”).(6) The scaled-up pIFN-αproduction was carried out in a 10 L fermentor. The results indicated that the pIFN-αactivities obtained in 10 L fermentor were obviously lower than those obtained in 5 L fermentor under the same control conditions (methanol conc., temperature, etc.). However, the pIFN-αantiviral activity in 10 L fermentor could be improved by adopting low induction temperature strategy at 20°C or the methanol/sorbitol co-feeding strategy at 30°C. The enzymatic analysis indicated that the sorbitol/methanol co-feeding strategy changed energy metabolism route, and energy metabolism was shifted from formaldehyde dissimilatory route into TCA cycle, which is beneficial for relieving the toxic effect of formaldehyde. The energy system could run normally and the highest pIFN-αactivity obtained with the methanol/sorbitol co-feeding strategy at 30°C (1.8×107 IU/mL) reached or even exceeded the highest level obtained with 20°C induction and the combinational induction strategy (20°C + high DO) in 5 L fermentor. As a result, fermentation cost reduced significantly and the entire fermentation performance was improved. |
PDF Full Text Request