A phage display study of interacting peptide binding partners of malarial S-Adenosylmethionine decarboxylase/Ornithine decarboxylase

Abstract

Due to the increasing resistance against the currently used antimalarial drugs, novel chemotherapeutic agents that target new metabolic pathways for the treatment of malarial infections are urgently needed. One approach to the drug discovery process is to use interaction analysis to find proteins that are involved in a specific metabolic pathway that has been identified as a drug target. Protein-protein interactions in such a pathway can be preferential targets since a) there is often greater structural variability in protein-protein interfaces, which can lead to more effective differentiation between the parasite and host proteins; and b) the important amino acids in a protein-protein interface are often conserved and even one amino acid mutation can lead to the dissociation of the complex, implying that resistance should be slower to appear. Since polyamines and their biosynthetic enzymes occur in increased concentrations in rapidly proliferating cells, the inhibition of polyamine metabolism is a rational approach for the development of antiparasitic drugs. Polyamine synthesis in P. falciparum is uniquely facilitated by a single open reading frame that encodes both rate-limiting enzymes in the pathway, namely ornithine decarboxylase (ODC) and S-adenosylmethionine decarboxylase (AdoMetDC). The AdoMetDC/ODC domains are assembled in a heterotetrameric bifunctional protein complex of ~330 kDa. Inhibition of both decarboxylase activities is curative of murine malaria and indicates the viability of such strategies in malaria control. It was hypothesized that protein ligands to this enzyme can be utilized in targeting the polyamine biosynthetic pathway in a novel approach. The bifunctional PfAdoMetDC/ODC was recombinantly expressed with a C-terminal Strep-tag-II to allow affinity purification. Subsequent gel electrophoresis analysis showed the presence of 3 contaminating proteins (~60 kDa, ~70 kDa and ~112 kDa) that co-elute with the ~330 kDa AdoMetDC/ODC. Efforts to purify the bifunctional protein to homogeneity included subcloning into a double-tagged vector for tandem affinity purification as well as size-exclusion HPLC. SDS-PAGE analysis of these indicated that separation of the four proteins was not successful, implicating the presence of strong protein-protein interactions. Western blot analysis showed that the ~112 kDa and ~70 kDa peptides were recombinantly produced with a C-terminal Strep-tag, indicating their heterologous origin. The ~60 kDa fragment was however not recognised by the tag-specific antibodies. This implies that this fragment is of E. coli origin. MS-analysis of the contaminating bands showed that the ~112 kDa peptide is an N-terminally truncated form of the full-length protein, the ~70 kDa peptide is a mixture of N-terminally truncated recombinant protein and E. coli DnaK and the ~60 kDa peptide is E. coli GroEL. A P. falciparum cDNA phage display library was used to identify peptide ligands to PfAdoMetDC/ODC. Of the peptides isolated through the biopanning process, only one was shown to occur in vivo. It could however not be conclusively shown that the isolated peptides bind to PfAdoMetDC/ODC and not to the co-eluting E. coli proteins. It is thought that while it is extremely likely that interacting protein partners to PfAdoMetDC/DOC exist, the available technologies are not sufficient to lead to the identification of such partners.