If you're a latent virus and lack a centromere, you're in a bind: how do you ensure that your genome is replicated in step with the host's and is passed on to the daughter cells at mitosis? One way would be to integrate into the host DNA so that you become an integral part of the cell's replicon. But several viruses, including Kaposi sarcoma herpesvirus (KSHV), use a different method: they hitch a ride on the host cell's chromosomes.
When KSHV infects Kaposi sarcoma cells, it adopts a circular episomal form. The virus encodes a protein required for replication termed latency-associated nuclear antigen (LANA). LANA was believed to function as a tether, since it could interact with both viral episomes and mitotic chromosomes. The N-terminal 22 amino acid residues of LANA are essential for chromosome attachment, but the target protein bound by LANA was unknown; several candidates (including the methyl-CpG binding protein MeCP2 and histone H1) had been proposed.
Barbera et al. (Science 311, 856–861, 2006) recently clarified what the target of LANA is. By alanine-scanning mutagenesis, residues 5–15 of LANA were identified as being important for chromosome binding. In the presence of viral episomes, LANA forms discrete foci (yellow) on chromosomes (red). Replacing residues 14 and 15 with alanine reduces the affinity of LANA for the chromosome (panel labeled TG). Barbera et al. isolated LANA-interacting proteins by fusing residues 1–32 of LANA to green fluorescent protein and performing a pull-down experiment. The interacting proteins constituted mostly core histone components. Full-length LANA also precipitated core histones in KSHV-infected cells. To define which specific protein(s) of the core histone was the target, the fusion protein was incubated with H2A-H2B dimers and H3-H4 tetramers. Only the H2A-H2B dimers were bound. This interaction was independent of the H2A-H2B tails, suggesting that it involved the folded histone domain.
To confirm further that H2A-H2B is the relevant LANA target, binding was compared between Xenopus laevis sperm chromatin, which lacks H2A and H2B, and sperm chromatin in which egg H2A and H2B were introduced. Only chromatin on which H2A and H2B had been deposited was able to bind the LANA fusion. To gain insight about this interaction, the authors solved the structure of the LANA1–23 peptide with the nucleosomal core particle. They found that LANA has high structural complementarity with an acidic patch on the surface of H2A-H2B, a region that overlaps the binding site for the N-terminal tail of H4. Binding of the LANA peptide buries a large surface area, consistent with the tight binding between LANA and H2A-H2B.
Other viruses that piggyback on host chromosomes use different binding targets; for example, Epstein-Barr virus uses a nucleolar protein, EBP2, and bovine papillomavirus interacts with a bromodomain protein, Brd4. But the KSHV LANA protein is the first shown to bind the histones directly. As LANA has been shown to affect transcription, these results may offer insight into how that type of regulation occurs as well.