Engineering of protein purification from Escherichia coli homogenates

Abstract

The term Quality by Design defines a concept for processing biologics in pharmaceutical biotechnology. In short it means that quality is built into the product rather than prove it afterwards. Knowing the product and even more a fully understanding of the process is mandatory. This work helps to gain process knowledgment of industrial downstream processing of E.coli homogenates. Escherichsa coli (E. coli) is a well-known microorganism, easy to manipulate and popular in biotechnology processes. The fermentation process is fast, cheap and simple. Due to the lack of secretion properties isolation and purification of recombinant proteins is challenging. To characterize the unit operations and their mutual influence, the E. coli strain HMS 174 (DE3) with the plasmid pET11a-GFPmut3.1 was used to establish a downstream model process. gfp was chosen because of its easy quantification. The recently established BIO INDUSTRIAL PILOT PLANT provided the possibility to work on a pilot scale level. First part of the investigation was the homogenization and its effectiveness in dependence of cell density, number of passages and operating pressure. Over the whole range of operating pressure (20MPa to 90MPa) cell density had a negligible effect on the breakage of the cells. Above 60MPa over 90% protein release was achieved with one passage. The disruption efficiency was highly dependent on operating pressure. It was also possibly to establish a correlation between solid content in the solution and its viscosity. It provides a useful estimate of solid content with a short measurement. This is particularly useful for design of centrifuge separations. Furthermore, the influence of the homogenization on the centrifugation behavior was studied. It could be shown that, the higher the operating pressure and the more passages had been performed the clearance efficiency was dramatically decreasing. Another important point is the reduction factor of the pellet volume. At high cell density the pellet volume reduction was lower, leading to more discharges when the homogenate was clarified in a disc stack centrifuge. In turn, this led to significant product losses because 0.4 L of liquid is discharged during every ejection. The chromatographic capture step using anion exchange chromatography was investigated with the resins CaptoQ and Q-Sepharose FF in detail for crude and diafiltrated homogenate. Adsorption isotherms revealed a displacement effect of gfp by other components at high feed concentrations. The displacement can be modeled by the extended Langmuir isotherm, which takes into account the variable qi. It represents the available binding sites for each component and it is suggested to be in correlation with their molecular size. Additionally, was investigated the adsorption kinetics of the diafiltrated and crude homogenate. Kinetic measurements showed that an overshoot above equilibrium capacity took place. The higher the feed concentration the more pronounced was the overshoot. According to these data, it is suggested that a low residence time and a high feed concentration is required to obtain the maximum binding capacity. In general, CaptoQ had a higher binding capacity than Q-Sepharose FF. Also diafiltrating the homogenate enhanced binding capacity. Furthermore break through curves (BTC) were performed with both resins. Remarkably, changes in the residence time had no effect on the dynamic binding capacity (DBC) for the crude homogenate. Additionally, a higher DBC was exhibited at a higher feed concentration. However, the recovery in the elution step was quite low due to the fact that a high amount of product was lost during the wash out phase. BTC's with the diafiltrated revealed a higher binding capacity which was positively influenced by enhancing the residence time and lowering the feed concentration.