Z proteins of New World arenaviruses bind RIG-I and interfere with type I interferon induction

The retinoic acid-inducible gene I product (RIG-I) is a cellular sensor of RNA virus infection that regulates the cellular beta interferon (IFN-beta) response. The nucleoproteins (NP) of arenaviruses are reported to antagonize the IFN response by inhibiting interferon regulatory factor 3 (IRF-3). Here, we demonstrate that the Z proteins of four New World (NW) arenaviruses, Guanarito virus (GTOV), Junin virus (JUNV), Machupo virus (MAVC), and Sabia virus (SABV), bind to RIG-I, resulting in downregulation of the IFN-beta response. We show that expression of the four NW arenavirus Z proteins inhibits IFN-beta mRNA induction in A549 cells in response to RNA bearing 5' phosphates (5'pppRNA). NW arenavirus Z proteins interact with RIG-I in coimmunoprecipitation studies and colocalize with RIG-I. Furthermore, expression of Z proteins interferes with the interaction between RIG-I and MAVS. Z expression also impedes the nuclear factor kappa light chain enhancer of activated B cells (NF-kappaB) and IRF-3 activation. Our results indicate that NW arenavirus Z proteins, but not Z protein of the Old World (OW) arenavirus lymphocytic choriomeningitis virus (LCMV) or Lassa virus, bind to RIG-I and inhibit downstream activation of the RIG-I signaling pathway, preventing the transcriptional induction of IFN-beta.

Dual host-virus arms races shape an essential housekeeping protein

Relentless selective pressures exerted by viruses trigger arms race dynamics that shape the evolution of even critical host genes like those involved in iron homeostasis.

Transferrin Receptor (TfR1) is the cell-surface receptor that regulates iron uptake into cells, a process that is fundamental to life. However, TfR1 also facilitates the cellular entry of multiple mammalian viruses. We use evolutionary and functional analyses of TfR1 in the rodent clade, where two families of viruses bind this receptor, to mechanistically dissect how essential housekeeping genes like TFR1 successfully balance the opposing selective pressures exerted by host and virus. We find that while the sequence of rodent TfR1 is generally conserved, a small set of TfR1 residue positions has evolved rapidly over the speciation of rodents. Remarkably, all of these residues correspond to the two virus binding surfaces of TfR1. We show that naturally occurring mutations at these positions block virus entry while simultaneously preserving iron-uptake functionalities, both in rodent and human TfR1. Thus, by constantly replacing the amino acids encoded at just a few residue positions, TFR1 divorces adaptation to ever-changing viruses from preservation of key cellular functions. These dynamics have driven genetic divergence at the TFR1 locus that now enforces species-specific barriers to virus transmission, limiting both the cross-species and zoonotic transmission of these viruses.

Genetic differences between mammalian species dictate the patterns of viral infection observed in nature. They also define how viruses must evolve in order to infect new mammalian hosts, giving rise to new and sometimes pandemic diseases. Because viruses must enter cells before they can replicate, new diseases often emerge when existing viruses evolve the ability to bind to the cell-surface receptor of a new species. At the same time, host cell receptors also evolve to counteract virus attacks. This back-and-forth evolution between virus and host can lead to an arms race that shapes the sequences of the proteins involved. In wild rodent populations, the retrovirus MMTV and New World arenaviruses both exploit Transferrin Receptor 1 (TfR1) to enter the cells of their hosts. Here we show that the physical interactions between these viruses and TfR1 have triggered evolutionary arms race dynamics that have directly modified the sequence of TfR1 and at least one of the viruses involved. Computational evolutionary analysis allowed us to identify specific residues in TfR1 that define patterns of viral infection in nature. The approach presented here can theoretically be applied to the study of any virus, through analysis of host genes known to be key to controlling viral infection. As such, this approach can expand our understanding of how viruses emerge from wildlife reservoirs, and how they drive the evolution of host genes.

Host-species transferrin receptor 1 orthologs are cellular receptors for nonpathogenic new world clade B arenaviruses

The ability of a New World (NW) clade B arenavirus to enter cells using human transferrin receptor 1 (TfR1) strictly correlates with its ability to cause hemorrhagic fever. Amapari (AMAV) and Tacaribe (TCRV), two nonpathogenic NW clade B arenaviruses that do not use human TfR1, are closely related to the NW arenaviruses that cause hemorrhagic fevers. Here we show that pseudotyped viruses bearing the surface glycoprotein (GP) of AMAV or TCRV can infect cells using the TfR1 orthologs of several mammalian species, including those of their respective natural hosts, the small rodent Neacomys spinosus and the fruit bat Artibeus jamaicensis. Mutation of one residue in human TfR1 makes it a functional receptor for TCRV, and mutation of four residues makes it a functional receptor for AMAV. Our data support an in vivo role for TfR1 in the replication of most, if not all, NW clade B arenaviruses, and suggest that with modest changes in their GPs the nonpathogenic arenaviruses could use human TfR1 and emerge as human pathogens.

Pathogenesis of a pichinde virus strain adapted to produce lethal infections in guinea pigs

A model for studying the pathogenesis of virulent arenavirus infection was developed by adapting Pichinde virus to produce lethal infections of inbred guinea pigs. This adapted Pichinde virus retained low virulence for primates, thus potentially reducing the biohazard to investigators. Whereas all inbred (strain 13) guinea pigs were infected and killed by 3 plaque-forming units or more of adapted Pichinde virus injected subcutaneously, outbred (Hartley strain) guinea pigs were relatively resistant. All infected, inbred guinea pigs died at 13 to 19 days after inoculation, with viremias in excess of 5 log(10) plaque-forming units/ml, severe lymphopenia (<1,000/mm(3)), and elevated serum glutamic oxaloacetic acid transaminase levels. Immunofluorescent antibody examination of tissues and infectivity titrations of tissue homogenates obtained at 3- to 4-day intervals demonstrated significant viral replication in all visceral tissues examined, but not in brain. Livers of all moribund guinea pigs contained moderate to severe hepatocellular necrosis and diffuse fatty change. Splenic red pulp and adrenal cortical tissues were engorged with blood and contained necrotic foci. Pancreatic acinar tissues were atrophied and vacuolated; lung sections typically contained areas of moderate to severe interstitial pneumonia. Inflammatory cells were conspicuously absent from all lesions. The virological and pathological features of adapted Pichinde infection in guinea pigs are remarkably similar to those described for Lassa virus infections in rhesus monkeys and humans, suggesting that this model might provide insight into the pathogenesis and treatment of Lassa fever in humans.

Role of the endothelium in viral hemorrhagic fevers

OBJECTIVE

To describe endothelial participation in the pathogenesis of viral hemorrhagic fevers and certain other acute infectious diseases.

DATA EXTRACTION AND SYNTHESIS

Survey of published literature on viral hemorrhagic fevers interpreted in light of observations in patients and research on those diseases.

CONCLUSIONS

Endothelial involvement is an extremely important factor in the clinical syndrome termed viral hemorrhagic fever. Endothelial dysfunction is important in the genesis of bleeding, which is not universal and is commonly seen only in the presence of thrombocytopenia or severe platelet dysfunction. The pathogenesis of endothelial dysfunction varies in the different diseases. In some situations, direct endothelial infection is important in increased vascular permeability, changes in the procoagulant vs. anticoagulant balance, or cytokine production. In all the viral hemorrhagic fevers studied to date, cytokine induction is an important factor and also acts on the endothelium. Poor myocardial contractility is a very important issue in viral hemorrhagic fever and is not caused by direct viral infection of the heart; it is increasingly being recognized that these patients present with low cardiac output and high peripheral resistance and that they respond poorly to fluid infusion. The clinical findings in viral hemorrhagic fever differ from those in the sepsis syndrome and should be studied and interpreted separately; this approach will sharpen therapeutic approaches and could shed light on the problems of sepsis in general.

Description of Guanarito virus (Arenaviridae: Arenavirus), the etiologic agent of Venezuelan hemorrhagic fever

This paper characterizes Guanarito virus, the etiologic agent of Venezuelan hemorrhagic fever. Based on its morphology and antigenic properties, Guanarito virus appears to be a new member of the Tacaribe complex of the genus Arenavirus, family Arenaviridae. Complement fixation and indirect fluorescent antibody tests showed that Guanarito virus and its antiserum are broadly cross-reactive with other members of the Tacaribe complex, but it can be differentiated from other members of the complex by neutralization test. Guanarito virus causes mortality in suckling mice and adult guinea pigs, but not in adult mice. Inoculated rhesus monkeys developed viremia and became ill; however, they subsequently recovered and responded with production of antibody. To date, all isolates of Guanarito virus have come from sick persons or wild rodents living within a single geographic focus in the central plains of Venezuela.

CpG oligodeoxynucleotides protect newborn mice from a lethal challenge with the neurotropic Tacaribe arenavirus

The innate immune system is key to limiting the early spread of most pathogens and directing the development of Ag-specific immunity. Recently, a number of synthetic molecules that activate the innate immune system by stimulating TLRs have been identified. Among them, synthetic oligodeoxynucleotides (ODNs) containing unmethylated CpG motifs (CpG ODNs) were shown to activate TLR9-bearing B cells, macrophages, and dendritic cells to induce a strong proinflammatory milieu and a type 1-biased immune response that protects mice from a variety of parasitic, bacterial, and viral infections. Although the protective effect of CpG ODN in adult mice was well established, its effectiveness in neonates, which have lower numbers of dendritic, B, and T cells and tend to favor Th2 responses, was unclear. This study uses the New World arenavirus Tacaribe, a neurotropic pathogen that is lethal in newborn mice, to explore the effectiveness of TLR-mediated innate immune responses. Neonatal BALB/c mice treated with CpG ODN at the time of infection had reduced viral load (p < 0.01) and increased survival (52%, p < 0.001 i.p.; 36%, p < 0.05 intranasally). Protection was achieved in mice treated no later than 3 days postchallenge and appears to be mediated by an increase in Ag-specific Abs (IgG and IgM) and to require inducible NO synthase expression and NO production. To our knowledge, this is the first study assessing the mechanisms by which CpG ODN can protect mice from a neurotropic viral infection.

Comparative analysis of disease pathogenesis and molecular mechanisms of New World and Old World arenavirus infections

Arenaviruses can cause fatal human haemorrhagic fever (HF) diseases for which vaccines and therapies are extremely limited. Both the New World (NW) and Old World (OW) groups of arenaviruses contain HF-causing pathogens. Although these two groups share many similarities, important differences with regard to pathogenicity and molecular mechanisms of virus infection exist. These closely related pathogens share many characteristics, including genome structure, viral assembly, natural host selection and the ability to interfere with innate immune signalling. However, members of the NW and OW viruses appear to use different receptors for cellular entry, as well as different mechanisms of virus internalization. General differences in disease signs and symptoms and pathological lesions in patients infected with either NW or OW arenaviruses are also noted and discussed herein. Whilst both the OW Lassa virus (LASV) and the NW Junin virus (JUNV) can cause disruption of the vascular endothelium, which is an important pathological feature of HF, the immune responses to these related pathogens seem to be quite distinct. Whereas LASV infection results in an overall generalized immune suppression, patients infected with JUNV seem to develop a cytokine storm. Additionally, the type of immune response required for recovery and clearance of the virus is different between NW and OW infections. These differences may be important to allow the viruses to evade host immune detection. Understanding these differences will aid the development of new vaccines and treatment strategies against deadly HF viral infections.

Epidemiology and pathogenesis of Bolivian hemorrhagic fever

The etiologic agent of Bolivian hemorrhagic fever (BHF), Machupo virus (MACV) is reported to have a mortality rate of 25-35%. First identified in 1959, BHF was the cause of a localized outbreak in San Joaquin until rodent population controls were implemented in 1964. The rodent Calomys collosus was identified as the primary vector and reservoir for the virus. Multiple animal models were considered during the 1970s with the most human-like disease identified in Rhesus macaques but minimal characterization of the pathogenesis has been published since. A reemergence of reported BHF cases has been reported in recent years, which necessitates the further study and development of a vaccine to prevent future outbreaks.

Candid No. 1 Argentine hemorrhagic fever vaccine protects against lethal Junin virus challenge in rhesus macaques

The protective efficacy of Candid No. 1, a live-attenuated vaccine against Argentine hemorrhagic fever (AHF), was evaluated in non-human primates. Twenty rhesus macaques immunized 3 months previously with graded doses of Candid No. 1 (16-127, 000 PFU), as well as 4 placebo-inoculated controls, were challenged with 4.41 log10 PFU of virulent P3790 strain Junin virus. All controls developed severe clinical disease; 3 of 4 died. In contrast, all vaccinated animals were fully protected; none developed any signs of AHF during a 105-day follow-up period. Viremia and virus shedding were readily detected in all placebo-vaccinated controls, while virus could be recovered only once (by amplification) from throat swabs of 2 Candid No. 1 vaccinees on day 21. Vigorous secondary-type neutralizing and immunofluorescent antibody responses were seen in most vaccinees that had received 3 log10 PFU Candid No. 1 or fewer; all others, including those receiving 127,200 PFU, maintained relatively stable titers during follow-up. Candid No. 1 was highly immunogenic and fully protective against lethal Junin virus challenge in rhesus macaques, even at extremely low (16 PFU) vaccine doses.

Experimental infection of the cane mouse Zygodontomys brevicauda (family Muridae) with guanarito virus (Arenaviridae), the etiologic agent of Venezuelan hemorrhagic fever

Chronic infections in specific rodents appear to be crucial to the long-term persistence of arenaviruses in nature. The cane mouse, Zygodontomys brevicauda, is a natural host of Guanarito virus (family Arenaviridae), the etiologic agent of Venezuelan hemorrhagic fever. The purpose of this study was to elucidate the natural history of Guanarito virus infection in Z. brevicauda. Thirty-nine laboratory-reared cane mice each were inoculated subcutaneously with 3.0 log10 plaque-forming units of the Guanarito virus prototype strain INH-95551. No lethality was associated with infection in any animal, regardless of age at inoculation. The 13 newborn, 14 weanling, and 8 of the 12 adult animals developed chronic viremic infections characterized by persistent shedding of infectious virus in oropharyngeal secretions and urine. These findings indicate that Guanarito virus infection in Z. brevicauda can be chronic and thus support the concept that this rodent species is the natural reservoir of Guanarito virus.

Junin virus infection of guinea pigs: immunohistochemical and ultrastructural studies of hemopoietic tissue

An association between viral antigens, cytopathic effect (CPE) and viral titers in blood and lymphoid tissues suggests a direct CPE of Junin virus on the lymphopoietic organs of guinea pigs infected with 10(3) 50% lethal doses of the XJ prototype strain. After seven days of infection, all lymphoreticular organs had infectivity titers higher than those for blood. Virus was recovered from bone marrow and lymph nodes at day 5 after infection; peak titers were obtained from bone marrow, spleen, and lymph nodes after day 10. Granular specific fluorescence was detected in the cytoplasm of reticular monocytes after day 7; megakaryocytes showed positive fluorescence, but specific staining of other lymphoid cells was not observed. Necrosis of bone marrow, lymph nodes, and spleen was observed after day 9. CPE consisted of overdevelopment of reticuloendoplasmic cisterne of reticulomonocytes and myeloblasts. Typical Junin virus particles were observed. Reticular cells were gradually destroyed, and simultaneous necrosis of surrounding lymphoid cells was observed.

Variation between strains of hamsters in the lethality of Pichinde virus infections

Infection by Pichinde virus, a member of the arenavirus group, was studied in Golden Syrian hamsters (Mesocricetus auratus) with regard to possible mechanisms of resistance to virus infection in adult hamsters. Two hamster strains were found to differ in their susceptibility to lethal Pichinde virus infection. LVG/Lak randomly bred hamsters were found to be 100% susceptible to low doses of Pichinde virus during the first 6 days of life, but after 8 days of life, mortality was uncommon. Peak virus titers in the serum of animals infected at 3 days of life were 4 logs greater than in animals infected at 12 days. MHA/Lak inbred hamsters, in contrast, were found to be susceptible to lethal virus infection both as newborns and as adults. Peak virus titers of greater than 10(8) plaque-forming units/ml were observed in serum 8 days after infection of adult MHA hamsters as compared with less than 10(3) plaque-forming units/ml in the serum of adult LVG hamsters. Cultured primary kidney cells and peritoneal macrophages from either hamster strain supported Pichinde virus replication equally well in vitro. Antibodies to the complement-fixing antigens and to antigens at the surface of virus-infected cells were produced by both strains of hamsters. Cyclophosphamide immunosuppression rendered adult LVG animals susceptible to lethal infections, and virus grew to high titers in the treated animals. These findings suggest that immunological factors that appear early in life in LVG hamsters and are deficient in MHA hamsters limit Pichinde virus infection. Unlike previously reported arenavirus diseases, the observations suggest that death is produced by a direct viral effect and not through immunopathological mechanisms.

Viral hemorrhagic fevers and hantavirus infections in the Americas

Several arenaviruses and hantaviruses have been isolated in the Americas during the last 4 decades. These are rodent-borne viruses responsible for the South American hemorrhagic fevers (SAHF) and hantavirus pulmonary syndrome (HPS). Although rare, SAHF and HPS are serious illnesses with high mortality rates. Most viral isolates found in the Americas represent New World lineages of their respective viral families. Their presence in the Western hemisphere is likely ancient, their relationship with their rodent hosts is likely coevolutionary, and their recent detection forebodes the likelihood of detecting additional arena- and hantaviral species in the Americas.

Safety and immunogenicity of a live-attenuated Junin (Argentine hemorrhagic fever) vaccine in rhesus macaques

The safety and immunogenicity of Candid #1, a live-attenuated Junin virus vaccine, were evaluated in rhesus macaques. Candid #1 was inoculated subcutaneously in graded doses ranging from 16 to 127,200 plaque-forming units (PFU) into four groups of five animals each; four controls received saline. There was no significant effect of the immunization on any physical, hematologic, or biochemical parameter measured. Junin virus was recovered by cocultivation from peripheral blood mononuclear cells (PBMC) of 14 (70%) of 20 animals from 1 to 21 days after immunization; 27 (12%) of 223 PBMC samples that represented animals in all four dose groups were positive. In contrast, virus was recovered from the plasma of only two of 20 macaques (two of 225 samples [0.9%]), and only once (by amplification) from throat swabs. No evidence of reversion was detected in any blood isolate. All animals developed a detectable neutralizing antibody response following vaccination. These results indicate that Candid #1 is safe and immunogenic in nonhuman primates.

Experimental Argentine hemorrhagic fever in rhesus macaques: virus-specific variations in pathology

Two isolates of Junin virus (Espindola and Ledesma) inoculated into rhesus macaques produced distinct lesions which were strain-constant and similar to reported human cases of Argentine hemorrhagic fever. The Espindola isolate was associated with hemorrhagia, necrosis of bone marrow, and mild hepatocellular necrosis. Ledesma isolate was associated with pronounced polioencephalomyelitis and autonomic ganglioneuritis, but very mild or absent hepatocellular necrosis, bone marrow necrosis, and hemorrhagia. Deaths of Espindola-infected macaques were usually attributed to hemorrhagia with severe secondary bacterial infections, while in Ledesma-infected macaques, death was associated either with early severe secondary bacterial infections or slowly progressive polioencephalomyelitis. These two Junin virus isolates demonstrated hemorrhagic viscerotropism or neurotropism in macaques, suggesting that the variety of Argentine hemorrhagic fever syndromes in man may be virus-isolate determined.

Antigenic relationships between attenuated and pathogenic strains of Junin virus

Antigenic relationships between attenuated and pathogenic strains of Junin virus (JV) were investigated. Five strains of either human or rodent origin were tested by cross-neutralization assay with hyperimmune antisera, raised in rabbits, against each strain. Polyclonal antisera could be used to distinguish among these JV strains, as the titer values differed significantly with ratios of homologous to heterologous titers, which ranged from 1.3 to 22.3. This demonstrates, independent of their virulence, a heterogeneity among the JV strains tested. The relatedness among JV strains was expressed quantitatively through a dendrogram based on taxonomic distance coefficients. The field strains of JV were grouped into two clusters, according to their geographic origin.

Pirital virus (Arenaviridae) infection in the syrian golden hamster, Mesocricetus auratus: a new animal model for arenaviral hemorrhagic fever

Adult Syrian golden hamsters inoculated intraperitoneally with Pirital virus, a recently discovered member of the Tacaribe complex of New World arenaviruses, developed a progressively severe, fatal illness with many of the pathologic features observed in fatal human cases of Lassa fever and other arenaviral hemorrhagic fevers. Most of the animals became moribund by Day 5 and were dead by Day 7 after inoculation. The most consistent histopathologic changes included interstitial pneumonitis, splenic lymphoid depletion and necrosis, and multifocal hepatic necrosis without significant inflammatory cell infiltration. The liver changes ranged from single cell death by apoptosis to coagulative necrosis of clusters of hepatocytes. Immunohistochemical studies of the liver demonstrated the presence and accumulation ot Pirital virus antigen within hepatocytes as well as Kupffer cells. An in situ terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) assay showed progressively increasing apoptotic activity in the liver of infected hamsters. A human hepatoblastoma cell line (Hep G2/C3A) inoculated with Pirital virus also developed progressive cell destruction and accumulation of viral antigen, as demonstrated by immunofluorescence. Results of this pilot study suggest that the Pirital virus-hamster model is a very promising new small animal model for studying the pathogenesis of arenavirus infections, particularly, the mechanism of direct virus-induced hepatic injury. It may also be useful for testingantiviral agents for treatment of arenaviral hemorrhagic fevers.

Viral strain dependent differences in experimental Argentine hemorrhagic fever (Junin virus) infection of guinea pigs

Guinea pigs infected with low-passage Junin virus of human origin showed viral strain dependent differences in mortality, LD50, time to death, and in viral spread and distribution. Different Junin strains appeared to cause at least two broad patterns of Argentine hemorrhagic fever in guinea pigs. A number of strains of Junin virus caused a viscerotropic type of illness in which virus replicated predominantly in lymph nodes, spleen, and bone marrow. With the most severe visceral forms of Argentine hemorrhagic fever, the guinea pigs became viremic, developed necrosis of spleen, lymph nodes, and bone marrow, showed gastric hemorrhages, and all animals died within 13-15 days. Other Junin strains induced a neurological type of illness with transient viral replication in and lymphocyte depletion of spleen and lymph nodes, with no detectable viremia or viral replication in bone marrow. Subsequently, virus was found in the brain with varying severities of polioencephalitis, and the guinea pigs frequently showed rear leg paralysis before death occurred 28-34 days after inoculation. Not all animals infected with a neurotropic strain developed all these signs. One viral strain induced some signs characteristic of both patterns of illness. Although the disease forms in the guinea pig model did not strictly correlate with those observed in the humans from which these strains were obtained, the different strains of Junin virus consistently caused very different patterns of illness in infected guinea pigs.

Differences in tropism and pH dependence for glycoproteins from the Clade B1 arenaviruses: implications for receptor usage and pathogenicity

The Clade B lineage of the New World arenaviruses contains four viruses capable of causing severe hemorrhagic fevers in humans. Within this group, the B1 sub-lineage contains the pathogenic viruses Junin (JUNV) and Machupo (MACV), as well as the non-pathogenic Tacaribe virus (TCRV). In order to elucidate differences that may determine pathogenicity, we studied the entry pathways directed by the glycoproteins (GPs) from these related B1 viruses, using pseudotyped retroviral vectors and GP1 immunoadhesin constructs. Our data revealed variations in the efficiency with which different cell types could be transduced by B1 vectors, and this correlated with the ability of the immunoadhesins to bind to those cells. Interestingly, the tropism directed by the TCRV GP proved to be distinct from that of JUNV and MACV, in particular on lymphocyte cell lines. In addition, the GPs showed variations in their sensitivity to an inhibitor of endosome acidification, with the TCRV GP again being the outlier. Together these data suggest that more than one entry pathway can be used by these closely related viruses and that the ability to cause human disease may be highly dependent on receptor usage.

Protection of guinea pigs inoculated with Tacaribe virus against lethal doses of Junin virus

Guinea pigs were protected against lethal doses of Junin virus by a previous inoculation with Tacaribe virus. Fourteen guinea pigs were infected with 10(6) 50% lethal doses (LD50) of Tacaribe virus and superinfected 45 days later with 10(3) LD50 of Junin virus. Appropriate control groups for both infections were also studied. The replication of Junin virus was impaired, as no virus was isolated from blood and organs of animals killed on days 9 and 11 after infection. High levels of neutralizing antibodies to Tacaribe virus were present by day 45, and no cross-reacting neutralizing antibodies to Junin virus were detected. However, three days after challenge with Junin virus, a typical secondary immune response to this virus was established. From these results it was concluded that the protection observed was mainly due to a specific immune response that was triggered by the Junin virus but primed by Tacaribe virus, which is antigenically similar to Junin virus.