Barros, Georgia (2012-07). Downstream Processing of Recombinant Proteins from Transgenic Plant Systems: Phenolic Compounds Removal from Monoclonal Antibody Expressing Lemna minor and Purification of Recombinant Bovine Lysozyme from Sugarcane. Doctoral Dissertation.
Thesis
Transgenic plant systems have been proposed as bioreactors in the production of pharmaceutical and industrial proteins. The economic benefits of inexpensive plant production systems could be erased if the downstream processing ends up being expensive. To avoid monoclonal antibody (mAb) modification or fouling of chromatography resins, removal of phenolics from plant extracts is desirable. Removal of major phenolics in Lemna extracts was evaluated by adsorption to PVPP, XAD-4, IRA-402 and Q-Sepharose resins. Analysis of phenolics adsorption to XAD-4, IRA-402 and Q-Sepharose showed superior dynamic binding capacities at pH 4.5 than at 7.5. The economic analysis using SuperPro Designer 7.0 indicated that addition of a phenolics adsorption step would increase mAb production cost only 20% by using IRA-402 compared to 35% for XAD-4 resin. The overall mAb processing cost can be reduced by implementing a phenolics removal step. To understand phenolics-resin interactions, adsorption isotherms of phenolic compounds (chlorogenic acid, ferulic acid, rutin, syringic acid and vitexin-2-O-rhamnoside) from different phenolic classes on three resins (IRA-402, PVPP, XAD-4) at pH 4.5 and 7.5 were determined. Differences in adsorption with the type of phenolics were observed, and PVPP was not efficient for phenolics removal. Screening of sugarcane lines for bovine lysozyme (BvLz) accumulation indicated that expression levels are still inadequate for commercial development. To maximize BvLz extraction, pH and ionic strength were evaluated; five conditions resulted in equivalent BvLz/TSP ratio. Membrane filtration process using BvLz extracts attained partial removal of native proteins by the 100 kDa membrane step, but also BvLz loss (21-29%). Regardless of the extraction condition, at least 47% of the starting BvLz was lost during the membrane processing. None of the evaluated extraction conditions caused a substantial recovery of BvLz in the concentrate. Alternative purification options for the IEX+HIC process, which achieved 95% BvLz purity, were tested. Direct loading of sugarcane extract concentrate on HIC and XAD-4 pretreatment of juice did not recovered BvLz as effectively as the IEX chromatography. Pure BvLz was obtained by the XAD+HIC process, but higher purification fold and HIC yield were achieved by the IEX+HIC process, due to the complete separation of BvLz and 18-kDa protein.
