The HIV capsid is a conical structure that is formed by a single protein, the capsid protein. Capsid monomers are arranged into a symmetrical hexamer around a central axis and the authors hypothesized that a pore might exist along this axis. However, there was no evidence for such a pore from previous hexamer structures. In fact, previous structures showed that the amino-terminal β-hairpins of each capsid monomer blocked the opening of such a potential pore. When the authors examined this region more closely, including in structures of capsid monomers, they noticed that the β-hairpin is flexible and can assume different conformations by tilting up to 15 Å away from the axis of symmetry. Indeed, when reconstructing a capsid hexamer based on monomer structures with this 'open' β-hairpin conformation, a central pore is formed. The conformational change is likely to depend on the protonation status of a histidine residue at the base of the β-hairpin. This is supported by the observation that structures that were obtained at a high crystallisation pH adopted a closed conformation, whereas structures that were obtained at a low pH adopted an open conformation. Interestingly, at physiological pH, the β-hairpin assumes an intermediate position, indicating high flexibility. Thus, the β-hairpin functions as a 'molecular iris' that controls entry to a central pore in the capsid.
Fluorescence anisotropy confirmed that dNTPs bind to capsid hexamers with very high affinity. This is likely to be coordinated by a highly electropositive ring of arginine side chains inside the pore that attracts the negatively charged dNTPs and is just large enough to let them pass. Dissociation experiments showed that the off-rate of dNTPs from the capsid hexamers is very fast as well, which suggests that the pore efficiently recruits dNTPs for reverse transcription inside the capsid. Indeed, the addition of ribonuclease (RNase) did not block reverse transcription of purified capsids in vitro, which suggests that the reaction occurs inside the capsid and that the RNA genome is protected inside this structure. When the negative charge inside the pore was abrogated by mutating the arginine to glycine, the resulting capsid hexamers no longer bound dNTPs. Furthermore, the infectivity of HIV that contained a mutant capsid sharply decreased and less HIV DNA, the product of reverse transcription, was produced compared with wild-type virus. Importantly, hexacarboxybenzene, a positively charged small molecule, efficiently inhibited reverse transcription, which led the authors to conclude that the capsid pore is a potential drug target. Interestingly, alignment of capsid sequences suggests that such a negatively charged pore is conserved among retrovirus families, and several residues in the β-hairpin are invariant or highly conserved in HIV; mutation of these residues abrogates or markedly decreases infectivity.
This is a preview of subscription content, access via your institution