Large scale transient 5-HT3 receptor production with the Semliki Forest Virus Expression System.

The expression of recombinant proteins with the Semliki Forest Virus (SFV) system has been scaled up to bioreactor scale. As a model protein for this study the human 5-HT(3) receptor was chosen. The gene for the receptor was subcloned into the SFV expression plasmid pSFV1. Virus production by in vivo packaging and production of the recombinant protein was scaled up, the latter to a reactor volume of 11.5 l. A Vibromix(TM) agitation system was chosen to overcome aggregation problems of BHK cells in suspension. In the process, cells were first grown to a density of 10(6) cells/ml, the medium was then exchanged with fresh medium and the culture was infected with the recombinant virus at an estimated multiplicity of infection of 30. 24 h post infection we measured an expression level of 3 million functional 5-HT(3) receptors per cell. For harvesting, the cells were pelleted by centrifugation. The receptor protein was purified in a single step (Hovius et al., 1998) by exploiting the hexa-His tag at minimal protein loss (51% yield). Experiments to optimise expression resulted in yields up to 8 million receptors per cell, when the pH of a suspension culture was controlled at pH 7.3. Rapid virus generation and protein production, high protein yields as well as successful large scale application have made the SFV expression system attractive to produce large quantities of recombinant protein in a very short time. After optimisation of the expression conditions (in particular by setting the pH at 7.3), yields were increased twofold.

Recombinant protein production using the Semliki Forest Virus expression system.

We report here the successful scale up of transient recombinant protein expression to litre scale using Semliki Forest Virus System. The expression of bacterial ß-galactosidase was initially compared in BHK and CHO cells and the conditions for optimal infection of BHK cells were identified. 10% FCS in a medium at pH 6.9 and infection in small volumes were found to be optimal. A high MOI results in an increased recombinant protein yield. Stirring does not affect the infection process. Finally we applied these optimal conditions to the production of a microsomal enzyme, human cyclooxygenase-2 in suspension spinners. Five independant productions at the 1 litre scale yielded reproducible substantial amounts of recombinant protein (16 mg microsomal protein 10(9) cells(-1)) with an average specific activity of 3942 ± 765 pg PGE(2) µg(-1) microsomal protein 5 min(-1).

Large scale transient expression with COS cells.

We demonstrate here that transient expression with COS cells can be performed at the one litre scale for a period of more than 10 days. Cells grown in T225 flasks were transfected by electroporation, transferred into spinners, and then grown either in suspension or on microcarriers. A daily medium change significantly extented culture life and production time, compared with standard protocols.Concentrations of the product, the secreted fusion protein CD40-Fc, were comparable in microcarrier and suspension culture. Cultures were started in fetal calf serum containing medium and the subsequent production process was performed in a low protein serum free medium which allowed easy downstream processing. 10 litres of supernatant, collected from one transfected batch of cells, yielded 30 mg of purified and biologically active protein.In addition to developing a simplified protocol for generation of cells we also reduced the material (DNA, cuvettes) required for electroporation. Our results show that scale up of transient expression to the litre scale can be successfully acieved. This provides a new tool to generate milligram quantities of protein within weeks of gene cloning.

Repeated hybridoma batch culture with cell recycle.

At the end of a hybridoma batch culture, the cells are usually discarded after separation from the culture broth. If, however, they are aseptically recycled into the reactor, the production process can be resumed simply by the addition of fresh medium. This cycle can then be repeated several times consecutively. In a test case, with a mouse hybridoma, we found antibody yields for each cycle in the same range as for a standard batch. In a 15 1 stirred tank reactor we could, within 6 days, produce 2.8 g of monoclonal antibody (MAb). This type of reactor operation allowed a doubling in the reactor volumetric productivity (mg/l/day).