STUDIES ON THE KINETICS AND MODELING OF L-GLUTAMIC ACID FERMENTATION

Abstract

L-glutamic acid (LGA) is commercially oneof the most important amino acids. Industrially, it is manufactured by batch/fed-batch fermentation processes using various strains of Corynebacterium glutamicum. India does not have any production unit for LGA, and its requirements are met through imports. The present thesis aims at studies on the kinetics and modeling of batch fermentation ofLGA in shake flasks by making the use of experimental observations. Three types of C. glutamicum, namely - MTCC 2745, MTCC 2679 and MTCC 2807 obtained from IMTECH, Chfedigarh, India were used for LGA fermentation at a constant temperature of30 °C and initial pH 7. Using the reported production medium, the efficacy of various nitrogen sources were assessed and it was found that urea is the best source of nitrogen for LGA production irrespective of the type of strains used. Other studies were made only with C.glutamicum MTCC 2745. The concentrations of biotin ( a growth factor) and urea (a nitrogen source) were optimized by classical method and were found to be 1 ug/1 and 8 g/1, respectively. In order to optimize the concentration of glucose ( a carbon source), batch fermentation studies were carried out with substrate (glucose) concentration ranging from 10 g/1 to 300 g/1 for a period of 16 h. The cell growth and LGA production were found to be maximum at an initial glucose concentration of 50 g/1. The fermentation was found to be substrate (glucose) limited below 50 g/1 of its concentration, whereas, growth inhibition by the substrate was observed above 50 g/1 of its concentration. The growth and product formation were found to be completely inhibited at the initial substrate concentration of 300 g/1. For studying the kinetics of fermentation, a glucose concentration of 50 g/1 along with the optimized concentrations ofbiotin and urea were used for a fermentation period of36 hwith free cells. Specific growth rate-, substrate and product concentration plots showed that the LGA fermentation is inhibited by substrate as well as by the product, and that it does not follow Monod kinetics. The growth behaviour of the cells could not be modeled by Monod equation, therefore, modified form of Monod equation with product inhibition term was used to simulate the experimental data. It was found that the linear product inhibition model of Levenspiel (1980) can describe the LGA fermentation In a satisfactory manner. in Abstract The laboratory data for the growth of C. glutamicum (Bona and Moser, 1997 c; Zhang et al. 1998) were also explained by the same model with better correlation. The growth data obtained in this work were also modeled satisfactorily by the logistic equation (Shuler and Kargi, 2002), modified form of logistic equation (Bona and Moser, 1997 a) and the modified form of Monod equation (Yamashita et al., 1969). The model developed with complete set of differential equations for cell growth, substrate consumption and product formation showed good agreement with the experimental data. The equations of this model were used for the presentation of a graphical performance between kinetically relevant variables along with simulation. In order to study the diffusional mass transfer effect of glucose, attempts were made to develop a model with incorporation of diffusion coefficient. Good agreement between experimental and simulated results was obtained. Experiments were also carried out with C. glutamicum entrapped in Ca-alginate matrix. Gel beads of various sizes were formed with the help of a syringe fitted with hypodermic needles of different sizes. The diameter of individual beads was calculated microscopically and the size was expressed as Sauter Mean Diameter (SMD) (Allen, 1992). The synthetic medium and other conditions (pH, Temperature) as used for fermentation of free cell system were also used for immobilized whole cells. The experimental data for cell loading, free cell, substrate and product concentrations were collected with respect to 120 h of fermentation. Over all biomass yield coefficient and product yield coefficient based on substrate were not affected by initial free cell concentration, initial substrate concentration, initial cell loading and alginate concentration in the gel matrix. Whereas, the yield coefficients in the free space and in the immobilization matrix; and the yield factor were found to vary. With the increasing initial free cell concentration, product formation was found to increase with continuous depletion of substrate. The increase in substrate concentration, resulted in the increase of the cell loading and product formation. Over a substrate concentration ranging from 25 - 100 g/1, no substrate inhibition was observed. This is in contrast to the free cell condition. Cell leaching was found to be more and the leaching occurred earlier with increasing cell loading. Rates of product formation and substrate consumption were observed to increase. Similar observation was reported by Nampoothiri and Pandey (1998) with regard to the initial cell loading, product formation and substrate depletion. iv Abstract Increasing alginate concentration in the gel matrix was found to have adverse effect on product formation and cell leaching. Cell leaching was noticed to be delayed and less due to increased alginate concentration. Nampoothiri and Pandey (1998) also reported similar findings. Amodel was developed for immobilized whole cell system which incorporates the effectiveness factor concepts along with bioreaction contribution in liquid space as well as immobilization matrix. The resultant equation for immobilized whole cell systems was reduced to the equation derived by Prasad and Mishra (1995) after elimination of the product inhibition, and lag terms. The equation was further reduced to the equation derived by Gates and Marlar (1968) for free cell systems. Time course of fermentation data for batch fermentation were simulated with the model developed for batch biosystem using immobilized whole cells with the assumption that there is no substrate or product inhibition. Other assumptions were same as those made for free cell system. Apart from the free cell concentration (X), other process variables (S, P and CimX') showed a closer correlation. Free cell concentration showed deviation initially but agreement later on. Initial free cell concentrations showed no effect on the cell loading, maximum specific growth rate, Monod saturation constant and effectiveness factor. Maximum specific growth rate and Monod saturation constant were invariant with initial substrate concentration, whereas the effectiveness factor was found to increase. Prasad (1991) also reported similar observation. With the increasing initial cell loading on the immobilization matrix, effectiveness factor and maximum specific growth rate were found to decrease, whereas the Monod saturation constant showed an increasing trend. Buchholz (1982) and Hiemstra et al. (1983) also made the same observations. The decreasing trend of maximum specific growth rate was also reported by Vorlop and Klein (1983); whereas the increasing nature ofMonod saturation constant was supported by Klein and Sachara (1980), and Van Ginekel (1983). Monod saturation constant increased, whereas the maximum specific growth rate and effectiveness factor decreased with the increasing alginate concentration. On the basis of the above results, it was concluded that the product inhibition model developed for free cell condition in batch cultures, the model developed for Abstract studying the diffusional mass transfer of glucose (free cells), and for immobilized whole cell used in batch bioreactor could be used to explain the kinetics of batch fermentation of L-glutamic acid using Corynebacterium glutamicum.