Structural basis for the removal of ubiquitin and interferon-stimulated gene 15 by a viral ovarian tumor domain-containing protease.

The attachment of ubiquitin (Ub) and the Ub-like (Ubl) molecule interferon-stimulated gene 15 (ISG15) to cellular proteins mediates important innate antiviral responses. Ovarian tumor (OTU) domain proteases from nairoviruses and arteriviruses were recently found to remove these molecules from host proteins, which inhibits Ub and ISG15-dependent antiviral pathways. This contrasts with the Ub-specific activity of known eukaryotic OTU-domain proteases. Here we describe crystal structures of a viral OTU domain from the highly pathogenic Crimean-Congo haemorrhagic fever virus (CCHFV) bound to Ub and to ISG15 at 2.5-Å and 2.3-Å resolution, respectively. The complexes provide a unique structural example of ISG15 bound to another protein and reveal the molecular mechanism of an ISG15 cross-reactive deubiquitinase. To accommodate structural differences between Ub and ISG15, the viral protease binds the ß-grasp folds of Ub and C-terminal Ub-like domain of ISG15 in an orientation that is rotated nearly 75° with respect to that observed for Ub bound to a representative eukaryotic OTU domain from yeast. Distinct structural determinants necessary for binding either substrate were identified and allowed the reengineering of the viral OTU protease into enzymes with increased substrate specificity, either for Ub or for ISG15. Our findings now provide the basis to determine in vivo the relative contributions of deubiquitination and deISGylation to viral immune evasion tactics, and a structural template of a promiscuous deubiquitinase from a haemorrhagic fever virus that can be targeted for inhibition using small-molecule-based strategies.

Identification of two functional domains within the arenavirus nucleoprotein.

Tacaribe virus (TCRV) belongs to the Arenaviridae family. Its bisegmented negative-stranded RNA genome encodes the nucleoprotein (N), the precursor of the envelope glycoproteins, the polymerase (L), and a RING finger matrix (Z) protein. The 570-amino-acid N protein binds to viral RNA, forming nucleocapsids, which are the template for transcription and replication by the viral polymerase. We have previously shown that the interaction between N and Z is required for assembly of infectious virus-like particles (VLPs) (J. C. Casabona et al., J. Virol. 83:7029-7039, 2009). Here, we examine the functional organization of TCRV N protein. A series of deletions and point mutations were introduced into the N-coding sequence, and the ability of the mutants to sustain heterotypic (N-Z) or homotypic (N-N) interactions was analyzed. We found that N protein displays two functional domains. By using coimmunoprecipitation studies, VLP incorporation assays, and double immunofluorescence staining, the carboxy-terminal region of N was found to be required for N-Z interaction and also necessary for incorporation of N protein into VLPs. Moreover, further analysis of this region showed that the integrity of a putative zinc-finger motif, as well as its amino-flanking sequence (residues 461 to 489), are critical for Z binding and N incorporation into VLPs. In addition, we provide evidence of an essential role of the amino-terminal region of N protein for N-N interaction. In this regard, using reciprocal coimmunoprecipitation analysis, we identified a 28-residue region predicted to form a coiled-coil domain (residues 92 to 119) as a newly recognized molecular determinant of N homotypic interactions.

Molecular determinants of arenavirus Z protein homo-oligomerization and L polymerase binding.

The arenavirus Z is a zinc-binding RING protein that has been implicated in multiple functions during the viral life cycle. These roles of Z involve interactions with viral and cellular proteins that remain incompletely understood. In this regard, Z inhibits viral RNA transcription and replication through direct interaction with the viral L polymerase. Here, we defined the L-binding domain of Tacaribe virus (TCRV) Z protein and the structural requirements mediating Z homo-oligomerization. By using site-directed mutagenesis, coimmunoprecipitation, and functional assays, we showed that residues R37, N39, W44, L50, and Y57, located around the zinc coordination site I, play a critical role in the Z-L interaction. We also found that Z protein from either TCRV or the pathogenic Junin virus (JUNV) self-associates into oligomeric forms in mammalian cells. Importantly, mutation of the myristoylation site, the strictly conserved residue G at position 2, severely impaired the ability of both TCRV Z and JUNV Z to self-interact as well as their capacity to accumulate at the plasma membrane, strongly suggesting that Z homo-oligomerization is associated with its myristoylation and cell membrane targeting. In contrast, disruption of the RING structure or substitution of W44 or N39, which are critical for L protein recognition, did not affect Z self-binding. Overall, the data presented here indicate that homo-oligomerization is not a requirement for Z-L interaction or Z-mediated polymerase activity inhibition.

The RING domain and the L79 residue of Z protein are involved in both the rescue of nucleocapsids and the incorporation of glycoproteins into infectious chimeric arenavirus-like particles.

Arenaviruses, such as Tacaribe virus (TacV) and its closely related pathogenic Junin virus (JunV), are enveloped viruses with a bipartite negative-sense RNA genome that encodes the nucleocapsid protein (N), the precursor of the envelope glycoprotein complex (GP), the polymerase (L), and a RING finger protein (Z), which is the driving force of arenavirus budding. We have established a plasmid-based system which allowed the successful packaging of TacV-like nucleocapsids along with Z and GP of JunV into infectious virus-like particles (VLPs). By coexpressing different combinations of the system components, followed by biochemical analysis of the VLPs, the requirements for the assembly of both N and GP into particles were defined. We found that coexpression of N with Z protein in the absence of minigenome and other viral proteins was sufficient to recruit N within lipid-enveloped Z-containing VLPs. In addition, whereas GP was not required for the incorporation of N, coexpression of N substantially enhanced the ratio of GP to Z into VLPs. Disruption of the RING structure or mutation of residue L79 to alanine within Z protein, although it had no effect on Z self-budding, severely impaired VLP infectivity. These mutations drastically altered intracellular Z-N interactions and the incorporation of both N and GP into VLPs. Our results support the conclusion that the interaction between Z and N is required for assembly of both the nucleocapsids and the glycoproteins into infectious arenavirus budding particles.

Targeting of arenavirus RNA synthesis by a carboxamide-derivatized aromatic disulfide with virucidal activity.

Several arenaviruses can cause severe hemorrhagic fever (HF) in humans, representing a public health threat in endemic areas of Africa and South America. The present study characterizes the potent virucidal activity of the carboxamide-derivatized aromatic disulfide NSC4492, an antiretroviral zinc finger-reactive compound, against Junín virus (JUNV), the causative agent of Argentine HF. The compound was able to inactivate JUNV in a time and temperature-dependent manner, producing more than 99 % reduction in virus titer upon incubation with virions at 37 °C for 90 min. The ability of NSC4492-treated JUNV to go through different steps of the multiplication cycle was then evaluated. Inactivated virions were able to bind and enter into the host cell with similar efficiency as control infectious particles. In contrast, treatment with NSC4492 impaired the capacity of JUNV to drive viral RNA synthesis, as measured by quantitative RT-PCR, and blocked viral protein expression, as determined by indirect immunofluorescence. These results suggest that the disulfide NSC4492 targets on the arenavirus replication complex leading to impairment in viral RNA synthesis. Additionally, analysis of VLP produced in NSC4492-treated cells expressing JUNV matrix Z protein revealed that the compound may interact with Z resulting in an altered aggregation behavior of this protein, but without affecting its intrinsic self-budding properties. The potential perspectives of NSC4492 as an inactivating vaccinal compound for pathogenic arenaviruses are discussed.

Uncovering viral protein-protein interactions and their role in arenavirus life cycle.

The Arenaviridae family includes widely distributed pathogens that cause severe hemorrhagic fever in humans. Replication and packaging of their single-stranded RNA genome involve RNA recognition by viral proteins and a number of key protein-protein interactions. Viral RNA synthesis is directed by the virus-encoded RNA dependent-RNA polymerase (L protein) and requires viral RNA encapsidation by the Nucleoprotein. In addition to the role that the interaction between L and the Nucleoprotein may have in the replication process, polymerase activity appears to be modulated by the association between L and the small multifunctional Z protein. Z is also a structural component of the virions that plays an essential role in viral morphogenesis. Indeed, interaction of the Z protein with the Nucleoprotein is critical for genome packaging. Furthermore, current evidence suggests that binding between Z and the viral envelope glycoprotein complex is required for virion infectivity, and that Z homo-oligomerization is an essential step for particle assembly and budding. Efforts to understand the molecular basis of arenavirus life cycle have revealed important details on these viral protein-protein interactions that will be reviewed in this article.