Résumé
Human myxovirus resistance 2 (MX2/MXB) is an interferon-stimulated gene (ISG) and was recently identified as a late postentry suppressor of human immunodeficiency virus type 1 (HIV-1) infection, inhibiting the nuclear accumulation of viral cDNAs. Although the HIV-1 capsid (CA) protein is believed to be the viral determinant of MX2-mediated inhibition, the precise mechanism of antiviral action remains unclear. The MX family of dynamin-like GTPases also includes MX1/MXA, a well-studied inhibitor of a range of RNA and DNA viruses, including influenza A virus (FLUAV) and hepatitis B virus but not retroviruses. MX1 and MX2 are closely related and share similar domain architectures and structures. However, MX2 possesses an extended N terminus that is essential for antiviral function and confers anti-HIV-1 activity on MX1 [ MX1(NMX2)]. Higher-order oligomerization is required for the antiviral activity of MX1 against FLUAV, with current models proposing that MX1 forms ring structures that constrict around viral nucleoprotein complexes. Here, we performed structure-function studies to investigate the requirements for oligomerization of both MX2 and chimeric MX1(NMX2) for the inhibition of HIV-1 infection. The oligomerization state of mutated proteins with amino acid substitutions at multiple putative oligomerization interfaces was assessed using a combination of covalent cross-linking and coimmunoprecipitation. We show that while monomeric MX2 and MX1(NMX2) mutants are not antiviral, higher-order oligomerization does not appear to be required for full antiviral activity of either protein. We propose that lower-order oligomerization of MX2 is sufficient for the effective inhibition of HIV-1.