Crystal structure of Mokola virus glycoprotein in its post-fusion conformation.

Mokola virus (MOKV) belongs to the lyssavirus genus. As other genus members-including rabies virus (RABV)-it causes deadly encephalitis in mammals. MOKV entry into host cells is mediated by its transmembrane glycoprotein G. First, G binds cellular receptors, triggering virion endocytosis. Then, in the acidic endosomal environment, G undergoes a conformational change from its pre- toward its post-fusion state that catalyzes the merger of the viral and endosomal membranes. Here, we have determined the crystal structure of a soluble MOKV G ectodomain in which the hydrophobic fusion loops have been replaced by more hydrophilic sequences. The crystal structure corresponds to a monomer that is similar to the protomer of the trimeric post-fusion state of vesicular stomatitis virus (VSV) G. However, by electron microscopy, we show that, at low pH, at the surface of pseudotyped VSV, MOKV spikes adopt the trimeric post-fusion conformation and have a tendency to reorganize into regular arrays. Sequence alignment between MOKV G and RABV G allows a precise location of RABV G antigenic sites. Repositioning MOKV G domains on VSV G pre-fusion structure reveals that antigenic sites are located in the most exposed part of the molecule in its pre-fusion conformation and are therefore very accessible to antibodies. Furthermore, the structure allows the identification of pH-sensitive molecular switches. Specifically, the long helix, which constitutes the core of the post-fusion trimer for class III fusion glycoproteins, contains many acidic residues located at the trimeric interface. Several of them, aligned along the helix, point toward the trimer axis. They have to be protonated for the post-fusion trimer to be stable. At high pH, when they are negatively charged, they destabilize the interface, which explains the conformational change reversibility. Finally, the present structure will be of great help to perform rational mutagenesis on lyssavirus glycoproteins.

Structural comparison of the C-terminal domain of functionally divergent lyssavirus P proteins.

Lyssavirus P protein is a multifunctional protein that interacts with numerous host-cell proteins. The C-terminal domain (CTD) of P is important for inhibition of JAK-STAT signaling enabling the virus to evade host immunity. Several regions on the surface of rabies virus P are reported to interact with host factors. Among them, an extended, discrete hydrophobic patch of P CTD is notable. Although structures of P CTD of two strains of rabies virus, and of mokola virus have been solved, the structure of P CTD for Duvenhage virus, which is functionally divergent from these species for immune evasion function, is not known. Here, we analyze the structures of P CTD of Duvenhage and of a distinct rabies virus strain to gain further insight on the nature and potential function of the hydrophobic surface. Molecular contacts in crystals suggest that the hydrophobic patch is important to intermolecular interactions with other proteins, which differ between the lyssavirus species.

Quantifying antigenic relationships among the lyssaviruses.

All lyssaviruses cause fatal encephalitis in mammals. There is sufficient antigenic variation within the genus to cause variable vaccine efficacy, but this variation is difficult to characterize quantitatively: sequence analysis cannot yet provide detailed antigenic information, and antigenic neutralization data have been refractory to high-resolution robust interpretation. Here, we address these issues by using state-of-the-art antigenic analyses to generate a high-resolution antigenic map of a global panel of 25 lyssaviruses. We compared the calculated antigenic distances with viral glycoprotein ectodomain sequence data. Although 67% of antigenic variation was predictable from the glycoprotein amino acid sequence, there are in some cases substantial differences between genetic and antigenic distances, thus highlighting the risk of inferring antigenic relationships solely from sequence data at this time. These differences included epidemiologically important antigenic differences between vaccine strains and wild-type rabies viruses. Further, we quantitatively assessed the antigenic relationships measured by using rabbit, mouse, and human sera, validating the use of nonhuman experimental animals as a model for determining antigenic variation in humans. The use of passive immune globulin is a crucial component of rabies postexposure prophylaxis, and here we also show that it is possible to predict the reactivity of immune globulin against divergent lyssaviruses.

Glycoproteins of Predicted Amphibian and Reptile Lyssaviruses Can Mediate Infection of Mammalian and Reptile Cells.

Lyssaviruses are neurotropic rhabdoviruses thought to be restricted to mammalian hosts, and to originate from bats. The identification of lyssavirus sequences from amphibians and reptiles by metatranscriptomics thus comes as a surprise and challenges the mammalian origin of lyssaviruses. The novel sequences of the proposed American tree frog lyssavirus (ATFLV) and anole lizard lyssavirus (ALLV) reveal substantial phylogenetic distances from each other and from bat lyssaviruses, with ATFLV being the most distant. As virus isolation has not been successful yet, we have here studied the functionality of the authentic ATFLV- and ALLV-encoded glycoproteins in the context of rabies virus pseudotype particles. Cryogenic electron microscopy uncovered the incorporation of the plasmid-encoded G proteins in viral envelopes. Infection experiments revealed the infectivity of ATFLV and ALLV G-coated RABV pp for a broad spectrum of cell lines from humans, bats, and reptiles, demonstrating membrane fusion activities. As presumed, ATFLV and ALLV G RABV pp escaped neutralization by human rabies immune sera. The present findings support the existence of contagious lyssaviruses in poikilothermic animals, and reveal a broad cell tropism in vitro, similar to that of the rabies virus.

Structural Elucidation of Viral Antagonism of Innate Immunity at the STAT1 Interface.

To evade immunity, many viruses express interferon antagonists that target STAT transcription factors as a major component of pathogenesis. Because of a lack of direct structural data, these interfaces are poorly understood. We report the structural analysis of full-length STAT1 binding to an interferon antagonist of a human pathogenic virus. The interface revealed by transferred cross-saturation NMR is complex, involving multiple regions in both the viral and cellular proteins. Molecular mapping analysis, combined with biophysical characterization and in vitro/in vivo functional assays, indicates that the interface is significant in disease caused by a pathogenic field-strain lyssavirus, with critical roles for contacts between the STAT1 coiled-coil/DNA-binding domains and specific regions within the viral protein. These data elucidate the potentially complex nature of IFN antagonist/STAT interactions, and the spatial relationship of protein interfaces that mediate immune evasion and replication, providing insight into how viruses can regulate these essential functions via single multifunctional proteins.

Expression, purification and crystallization of a lyssavirus matrix (M) protein.

The matrix (M) proteins of lyssaviruses (family Rhabdoviridae) are crucial to viral morphogenesis as well as in modulating replication and transcription of the viral genome. To date, no high-resolution structural information has been obtained for full-length rhabdovirus M. Here, the cloning, expression and purification of the matrix proteins from three lyssaviruses, Lagos bat virus (LAG), Mokola virus and Thailand dog virus, are described. Crystals have been obtained for the full-length M protein from Lagos bat virus (LAG M). Successful crystallization depended on a number of factors, in particular the addition of an N-terminal SUMO fusion tag to increase protein solubility. Diffraction data have been recorded from crystals of native and selenomethionine-labelled LAG M to 2.75 and 3.0 A resolution, respectively. Preliminary analysis indicates that these crystals belong to space group P6(1)22 or P6(5)22, with unit-cell parameters a = b = 56.9-57.2, c = 187.9-188.6 A, consistent with the presence of one molecule per asymmetric unit, and structure determination is currently in progress.

Isolation of a European bat lyssavirus type 2 from a Daubenton's bat in the United Kingdom.

European bat lyssavirus type 2 (EBLV-2) has been isolated once previously from a bat in the UK in June 1996. In September 2002, a Daubenton's bat (Myotis daubentonii) found in Lancashire developed abnormal behaviour, including unprovoked aggression, while it was in captivity. Brain samples from the bat were tested for virus of the Lyssavirus genus, which includes EBLV-2 (genotype 6), and classical rabies virus (genotype 1). A positive fluorescent antibody test confirmed that it was infected with a lyssavirus, and PCR and genomic sequencing identified the virus as an EBLV-2a. Phylogenetic comparisons with all the published sequences from genotype 6 showed that it was closely related to the previous isolate of EBLV-2 in the UK and suggested links to isolates from bats in The Netherlands. The isolation of EBLV-2 from a bat found on the west coast of England provides evidence that this virus may be present within the UK Daubenton's bat population at a low prevalence level.

Lyssavirus P gene characterisation provides insights into the phylogeny of the genus and identifies structural similarities and diversity within the encoded phosphoprotein.

A comprehensive phylogenetic analysis of the Lyssavirus genus, employing P gene sequences from 128 isolates recovered globally, is presented. While confirming prior suggestions of the genetic distinctness of the Australian bat lyssaviruses, these data also revealed a clear division within the rabies virus clade (Genotype 1) between globally distributed viruses circulating predominantly in canid species (subgroup 1-1), and American strains harbored by both chiropteran and terrestrial hosts (subgroup 1-2). Nucleotide substitution patterns within the P locus suggested differential selection operating along the length of the open reading frame (ORF) between rabies viruses of these two subgroups. Comparison of the deduced primary sequences of the encoded phosphoproteins of all isolates provided insights into the product's structural organization. Two conserved (CD1,2) and two variable (VD1,2) domains were evident; high variation in the VD2 region was reflected by differences in hydropathic profiles. Only two of five serine residues reported to function as phosphate acceptors in the P protein of the rabies challenge virus standard (CVS) strain were absolutely conserved; similarly, out of four internal methionines reported to direct internal translation initiation of the CVS strain to produce N-terminally truncated P proteins, only Met(20) was universally retained. In contrast, two sequence motifs, one believed to confer binding to the cytoplasmic dynein light chain LC8, and a lysine-rich sequence probably contributing to N protein binding, were conserved throughout the genus. Most rabies viruses of the carnivora (1-1) contain a potential C ORF in alternate frame to that of P, a feature limited or absent in most other isolates of the genus, an observation interpreted with consideration to the predicted course of lyssavirus evolution.

Mokola virus glycoprotein and chimeric proteins can replace rabies virus glycoprotein in the rescue of infectious defective rabies virus particles.

A reverse genetics approach which allows the generation of infectious defective rabies virus (RV) particles entirely from plasmid-encoded genomes and proteins (K.-K. Conzelmann and M. Schnell, J. Virol. 68:713-719, 1994) was used to investigate the ability of a heterologous lyssavirus glycoprotein (G) and chimeric G constructs to function in the formation of infectious RV-like particles. Virions containing a chloramphenicol acetyltransferase (CAT) reporter gene (SDI-CAT) were generated in cells simultaneously expressing the genomic RNA analog, the RV N, P, M, and L proteins, and engineered G constructs from transfected plasmids. The infectivity of particles was determined by a CAT assay after passage to helper virus-infected cells. The heterologous G protein from Eth-16 virus (Mokola virus, lyssavirus serotype 3) as well as a construct in which the ectodomain of RV G was fused to the cytoplasmic and transmembrane domains of the Eth-16 virus G rescued infectious SDI-CAT particles. In contrast, a chimeric protein composed of the amino-terminal half of the Eth-16 virus G and the carboxy-terminal half of RV G failed to produce infectious particles. Site-directed mutagenesis was used to convert the antigenic site III of RV G to the corresponding sequence of Eth-16 G. This chimeric protein rescued infectious SDI-CAT particles as efficiently as RV G. Virions containing the chimeric protein were specifically neutralized by an anti-Eth-16 virus serum and escaped neutralization by a monoclonal antibody directed against RV antigenic site III. The results show that entire structural domains as well as short surface epitopes of lyssavirus G proteins may be exchanged without affecting the structure required to mediate infection of cells.

Functional interaction map of lyssavirus phosphoprotein: identification of the minimal transcription domains.

Lyssaviruses, the causative agents of rabies encephalitis, are distributed in seven genotypes. The phylogenetically distant rabies virus (PV strain, genotype 1) and Mokola virus (genotype 3) were used to develop a strategy to identify functional homologous interactive domains from two proteins (P and N) which participate in the viral ribonucleoprotein (RNP) transcription-replication complex. This strategy combined two-hybrid and green fluorescent protein-reverse two-hybrid assays in Saccharomyces cerevisiae to analyze protein-protein interactions and a reverse genetic assay in mammalian cells to study the transcriptional activity of the reconstituted RNP complex. Lyssavirus P proteins contain two N-binding domains (N-BDs), a strong one encompassing amino acid (aa) 176 to the C terminus and a weak one in the 189 N-terminal aa. The N-terminal portion of P (aa 52 to 189) also contains a homomultimerization site. Here we demonstrate that N-P interactions, although weaker, are maintained between proteins of the different genotypes. A minimal transcriptional module of the P protein was obtained by fusing the first 60 N-terminal aa containing the L protein binding site to the C-terminal strong N-BD. Random mutation of the strong N-BD on P protein identified three highly conserved K residues crucial for N-P interaction. Their mutagenesis in full-length P induced a transcriptionally defective RNP. The analysis of homologous interactive domains presented here and previously reported dissections of the P protein allowed us to propose a model of the functional interaction network of the lyssavirus P protein. This model underscores the central role of P at the interface between L protein and N-RNA template.