Scale-down models to optimize a filter train for the downstream purification of recombinant pharmaceutical proteins produced in tobacco leaves.

The extraction of biopharmaceutical proteins from intact leaves involves the release of abundant particulate contaminants that must be removed economically from the process stream before chromatography, for example, using disposable filters that comply with good manufacturing practice. We therefore scaled down an existing 200-kg process for the purification of two target proteins from tobacco leaves (the monoclonal antibody 2G12 and the fluorescent protein DsRed, as monitored by surface plasmon resonance spectroscopy and fluorescence imaging, respectively) and screened different materials on the 2-kg scale to reduce the number of depth filtration steps from three to one. We assessed filter cost and capacity, filtrate turbidity, and protein recovery when the filter materials were challenged with extracts from different tobacco varieties and related species grown in soil or rockwool. PDF4 was consistently the most suitable depth filter because it was the least expensive, it did not interact significantly with the target proteins, and it had the greatest overall capacity. The filter capacity was generally reduced when plants were grown in rockwool, but this substrate has a low bioburden, thus improving process safety. Our data concerning the clarification of plant extracts will help in the design of more cost-effective downstream processes and accelerate their development.