We demonstrated the successful optimization of a recombinant multi-subunit malaria vaccine candidate protein for production in the methylotrophic yeast Pichia pastoris by the identification and subsequent removal of two protease cleavage sites. After observing protein degradation in the culture supernatant of a fed-batch fermentation, the predominant proteolytic fragment of the secreted recombinant protein was analyzed by mass spectrometry. The MS data indicated the cleavage of an amino acid sequence matching the yeast KEX2-protease consensus motif EKRE. The cleavage in this region was completely abolished by the deletion of the EKRE motif in a modified variant. This modified variant was produced, purified, and used for immunization of rabbits, inducing high antigen specific antibody titers (2 × 10(6) ). Total IgG from rabbit immune sera recognized different stages of Plasmodium falciparum parasites in immunofluorescence assays, indicating native folding of the vaccine candidate. However, the modified variant was still degraded, albeit into different fragments. Further analysis by mass spectrometry and N-terminal sequencing revealed a second cleavage site downstream of the motif PEVK. We therefore removed a 17-amino-acid stretch including the PEVK motif, resulting in the subsequent production of the full-length recombinant vaccine candidate protein without significant degradation, with a yield of 53 mg per liter culture volume. We clearly demonstrate that the proteolytic degradation of recombinant proteins by endogenous P. pastoris proteases can be prevented by the identification and removal of such cleavage sites. This strategy is particularly relevant for the production of recombinant subunit vaccines, where product yield and stability play a more important role than for the production of a stringently-defined native sequence which is necessary for most therapeutic molecules.
Keywords: Pichia pastoris; Plasmodium falciparum; protein engineering; proteolytic degradation.
© 2014 Wiley Periodicals, Inc.