Modelling the structure of full-length Epstein-Barr virus nuclear antigen 1

Epstein-Barr virus (EBV) is a clinically important human virus associated with several cancers and is the etiologic agent of infectious mononucleosis. The viral nuclear antigen-1 (EBNA1) is central to the replication and propagation of the viral genome and likely contributes to tumourigenesis. We have compared EBNA1 homologues from other primate lymphocryptoviruses (LCV) and found that the central glycine/alanine repeat (GAr) domain, as well as predicted cellular protein (USP7 and CK2) binding sites are present in homologues in the Old World primates, but not the marmoset; suggesting that these motifs may have co-evolved. Using the resolved structure of the C-terminal one third of EBNA1 (homodimerisation and DNA binding domain), we have gone on to develop monomeric and dimeric models in silico of the full length protein. The C-terminal domain is predicted to be structurally highly similar between homologues, indicating conserved function. Zinc could be stably incorporated into the model, bonding with two N-terminal cysteines predicted to facilitate multimerisation. The GAr contains secondary structural elements in the models, while the protein binding regions are unstructured, irrespective of the prediction approach used and sequence origin. These intrinsically disordered regions may facilitate the diversity observed in partner interactions. We hypothsise that the structured GAr could mask the disordered regions, thereby protecting the protein from default degradation. In the dimer conformation, the C-terminal tails of each monomer wrap around a proline-rich protruding loop of the partner monomer, providing dimer stability, a feature which could be exploited in therapeutic design