Arenaviruses and hantaviruses are human pathogens that cause significant morbidity and mortality. The current lack of vaccines and treatment options for these viruses is a global concern. Despite producing only 4 proteins, these viruses are able to maintain a persistent and asymptomatic infection in wild rodents while being continuously shed into the environment. In humans, these viruses cause a spectrum of diseases ranging from aseptic meningitis to severe hemorrhagic fever syndromes. Little is known about how arenavirus and hantavirus proteins engage and interact with the human proteome during the complex process of viral biogenesis, or how the interactions with human proteins contribute to viral propagation as well as the onset and progression of disease. This dissertation provides a road map of the protein interactions formed between a prototypic envelope glycoprotein encoded by either an arenavirus or hantavirus, and the human proteome. The viral envelope glycoprotein (GP) decorates the surface of the virion. The primary function of the GP is to mediate attachment of the virus to specific cellular receptors, and after internalization of the virion, fuse the viral membrane with an internal endosomal membrane. In order to carry out these specific tasks, the viral GPs must first co-opt the extensive machinery found within the cellular secretory pathway to coordinate the proper glycosylation, folding, proteolytic maturation, and targeting of the GP during its biosynthesis. We identified a human protein with a conserved interaction amongst these two groups of viral GPs termed the Endoplasmic Reticulum (ER)-Golgi Intermediate Compartment Protein of 53 kiloDaltons (ERGIC-53). ERGIC-53 is an intracellular cargo receptor that normally cycles within the early secretory pathway of cells, where it is responsible for ferrying a small subset of cellular glycoproteins, most notably the coagulation factors FV and FVIII, from the ER to the Golgi apparatus. Herein we describe a novel role for ERGIC-53 in the propagation of not only arenaviruses, but also coronaviruses and filoviruses. Following infection with an arenavirus, ERGIC-53 leaves the early secretory pathway and becomes incorporated into the virus as it pinches off from the cell surface. Newly formed viruses lacking ERGIC-53 are no longer infectious due, in part, to a defect in their ability to attach to host cells. We suggest that ERGIC-53 represents a promising broad-spectrum antiviral target because of its association with the GPs from many families of pathogenic viruses, as well as its ability to exert control over their infectivity; and finally, because ERGIC-53 itself is not required for human health. The discovery of ERGIC-53 outside of its normal location inside of cells suggests that it may have additional unknown functions. Lastly, by revealing the importance of the cellular protein in controlling viral infectivity, we provide insight into the ongoing co-evolution of virus and host.

La famille des Arenaviridae contient des virus transmis par les rongeurs responsables notamment de graves fièvres hémorragiques dont la fièvre de Lassa qui infecte chaque année 300 000 personnes. L’absence de solutions thérapeutiques efficaces (vaccins et/ou antiviraux) en fait une priorité d’étude. Il est donc urgent d’identifier de nouvelles cibles thérapeutiques et des antiviraux spécifiques associés. Le génome des Arénavirus code pour deux nucléases possédant des fonctions cruciales pour la survie du virus. La protéine L possède un domaine endonucléase indispensable à la réplication virale réalisant le mécanisme de vol de coiffe. La protéine NP comporte un domaine exonucléase impliqué dans l’échappement au système immunitaire de l’hôte infecté. Ces deux enzymes possèdent un mécanisme d’action commun métal-dépendant de clivage de leur ARN substrat. Ce projet :1) se propose d’apporter la preuve de concept que ces deux enzymes sont de nouvelles cibles thérapeutiques pour combattre les Arénavirus.2) se base sur le criblage préalable d’une chimiothèque de 1ère génération sur l’endonucléase d’un Arénavirus et la découverte d’une famille de dicéto-­‐acides (DCAs) inhibitrice de son activité. Au cours de ce travail de thèse, une chimiothèque de 2nde génération de 22 composés a été synthétisée. Des voies de synthèses nouvelles et originales ont été développées. Les composés ont été évalués par des méthodes biophysiques pour leur capacité à se lier aux enzymes cibles, ont été évalués comme potentiels inhibiteurs dans des tests d’activité in vitro, et in cellula dans un test en mini-­‐génome. Certains composés se sont montrés de bons ligands, actifs in vitro et in cellula. / The Arenaviridae family comprises viruses transmitted by rodents responsible for several diseases such as hemorragic fevers among which Lassa fever infects 300,000 people each year. The current lack of effective vaccines and/or antivirals makes the development of therapeutic solutions of high priority. The genome of Arenaviruses encodes for two nucleases bearing crucial functions for virus survival. The L protein bears an endonuclease domain essential for viral replication through the cap-­‐snatching mechanism. The NP protein has an exonuclease domain involved in anti-­‐immune activity. Both enzymes share a common mode of action as they cleave their RNA substrate via a metal-­‐dependant (Mg2+) strategy. This work intends to :1) consider these two enzymes as novel Arenavirus therapeutic targets.2) take over a preliminary screening of a 1st generation of ligands on an Arenavirus endonuclease. Diketo-­‐acids (DKAs) have been discovered as potential inhibitors of its activity. To this end, a library of 2nd generation of 22 compounds was synthetized thanks to the development of an optimized synthesis pathway. Compounds were further evaluated by biophysical methods for their ability to bind therapeutic targets, as potential inhibitors in an in vitro assay and in cellula in a mini-­‐genome assay. Some of them are active in vitro and in cellula

Serologically related arenaviruses have been isolated from West Africa, Mozambique, Zimbabwe and the Central African Republic. Human disease is only associated with the West African isolates. The virulence of Mozambique, Zimbabwe and Central African Republic isolates in humans is not known. This Thesis is an account of work carried out by the author to compare the biological characteristics of isolates from West Africa, Mozambique and Zimbabwe. It describes the successful isolation and identification of the aetiological agents, their physicochemical and antigenic characteristics and describes in vivo studies using mice, guinea pigs and Rhesus monkeys. A direct comparison was made with a patient diagnosed as having Lassa fever. The disease in man and monkeys following infection with Lassa virus was similar. The Rhesus monkey and guinea pig proved suitable experimental models in which to study and compare the pathogenic responses and also to evaluate various aspects of protection. These animal models when immunised with the viruses from Mozambique and Zimbabwe were protected when subsequently challenged with Lassa virus. The Mozambique and Zimbabwe isolates proved to have morphological and physicochemical characteristics not dissimilar from West African Lassa viruses and those members of the arenavirus family from South America. Serological and immunochemical investigations suggest the existence of both common and unique antigenic determinants on the viruses from Mozambique, -Zimbabwe and West Africa. This grouping also coincides with the geographic separation of the viruses, i.e. Lassa - West Africa and Mopeia -southeast Africa. Similar differences in host susceptibility have also been demonstrated. Lassa virus produces a fatal haemorrhagic disease while Mopeia isolates produce only an asymptomatic infection. The combined data suggests the possibility of two virus groups within the 'Old World' arenavirus classification. The proposed name, 'Mopeia', forms one group and includes the viruses from Mozambique and Zimbabwe. The Lassa strains from West Africa form the second group. It is suggested that the Mopeia viruses are minor antigenic variants of Lassa and should be included within the arenavirus family.

Viruses are infectious agents that must infect the cells of living organisms in order to reproduce. They have relatively simple genomes which encode few proteins but can compensate for their simplicity by hijacking components of their cellular hosts. Arenaviruses, a family of zoonotic viruses carried by rodents, encode only 4 proteins. One of these proteins, Z, is responsible for several functions during the virus life cycle including driving the formation and release of new virus particles at the plasma membrane of infected cells. Relatively little is known about how this viral protein is regulated or the complement of host proteins it engages in order to produce new virus particles or augment Z's other functions. To address this gap in knowledge, mass spectrometry was used to identify phosphorylation sites in the Old World arenavirus, lymphocytic choriomeningitis virus (LCMV) Z protein. Phosphorylation sites were identified at serine 41 (S41) and tyrosine 88 (Y88). Functional studies using recombinant (r)LCMV containing mutations at these phosphorylation sites revealed that both were important for the production of defective interfering (DI) particles. DI particles are replication-incompetent virus particles that interfere with the production of infectious virus and mitigate its cytopathic effect. While a mutation that mimics phosphorylation at S41 reduced LCMV's ability to produce both infectious and DI particles, this mutation had a much stronger impact on DI particles. Production of DI particles in Y88-mutant rLCMV was drastically reduced while the impact on infectious virus was minimal. Y88 lies within a type of viral late domain (PPXY) also found in matrix proteins of several disparate virus families where it has been shown to drive infectious virus release by recruiting the membrane scission machinery of the cellular endosomal sorting complex required for transport (ESCRT). Inhibition of the ESCRT pathway drastically reduced LCMV DI particle but not infectious virus release indicating that Z's PPXY late domain and the cellular ESCRT complex are required specifically for the production of DI particles. Mass spectrometry was also used to identify host protein partners of Z as well as the host proteins recruited into virus particles for the New World arenavirus, Junin (JUNV). ESCRT complex proteins were enriched in JUNV virus-like particles (VLPs) and bona fide virions. In contrast to LCMV, inhibition of the ESCRT complex resulted in significantly less infectious JUNV release. This indicates that the ultimate role of ESCRT engagement by the Old World arenavirus, LCMV, differs from that of New World, JUNV. This work represents the first demonstration that a viral protein motif and the host machinery it engages selectively drive DI particle production independently of infectious virus. It also suggests that host cell kinases can dynamically regulate the production of DI particles through phosphorylation of Z. Finally, the late domain-mutant rLCMV generated in these studies represents the first LCMV strain known to produce undetectable levels of DI particles which provides the opportunity to assess the impact that a loss of DI particles has on the ability of LCMV to establish or maintain a persistent infection.

MY-24 is an aristeromycin derivative previously shown to protect AG129 type I and II interferon receptor knockout mice from lethal challenge with Tacaribe virus (TCRV). TCRV is nonpathogenic to humans, but is closely related to the highly pathogenic New World arenaviruses that cause often-fatal viral hemorrhagic fever syndromes. Remarkably, MY-24 prevented mortality without reducing TCRV burden in the circulation or tissues. To investigate the mechanism by which MY-24 protects AG129 mice against TCRV infection, we first characterized the natural history of disease in the model with an emphasis on cytokine responses and vascular integrity to establish the best times to evaluate the effects of MY-24 treatment on host responses believed to contribute to pathogenesis and fatal outcome. We found that viral replication in the blood and in various tissues precedes a hyperproduction of proinflammatory mediators that may lead to the destabilization of the endothelial barrier and increased vascular leakage believed to contribute to terminal shock associated with severe cases of hemorrhagic fever. We also found slightly reduced virus titers in certain tissues from MY-24-treated mice, suggesting that there may be a weak antiviral effect; however, TCRV was not cleared from lung, spleen, brain or kidney in recovering animals out to 40 days post-infection, indicative of the establishment of chronic infection in mice that are able to survive the initial challenge. Neutralizing antibodies do not appear to play a major role in the antiviral effect of MY-24, whereas reductions in several key proinflammatory cytokines in mice treated with MY-24 may serve to reduce vascular leakage caused by TCRV infection.

Viruses are the only organisms known to store their genetic information solely in the form of RNA, and have thus evolved unique machinery to replicate an RNA genome and initiate viral gene expression in the infected cell. The large polymerase protein (L) of negative-strand (NS) RNA viruses is a particularly intriguing model for viral replication, where all of the enzymatic activities required for mRNA transcription, RNA modification, and genomic RNA replication are contained within a single polypeptide. Whereas the host cell requires a suite of enzymes to accomplish these tasks, L alone is the catalytic engine driving NS RNA viral replication. Here we demonstrate purification of functional L protein from Machupo virus (MACV) and reconstitute arenavirus RNA synthesis initiation and gene expression regulation in vitro using purified recombinant components. Through single-molecule electron microscopy analysis of MACV L, we provide the first structural information of viral L proteins. Comparative analysis with nonsegmented NS RNA viral L proteins reveals how the various enzymatic domains are arranged into a conserved architecture shared by both polymerases. Our in vitro RNA synthesis data defines the basis of arenavirus sequence-specific polymerase recruitment and how inter-termini interactions regulate template recognition. Moreover, we discover a new role for the arenaviral matrix protein in regulating viral RNA synthesis by locking a polymerase-template complex. The inhibitory matrix-L-RNA assembly functionally links transcription regulation and polymerase packaging, and reveals a mechanism for NS RNA viruses to ensure polymerase incorporation during virion maturation. Reconstitution of RNA synthesis in vitro establishes a new framework to understand the arenaviral polymerase complex, and our structural and biochemical experiments provide a basis for mechanistic analysis of the NS RNA viral replication machinery.

Background: Parasite evolution is hypothesized to select for levels of parasite virulence that maximise transmission success. When host population densities fluctuate, low levels of virulence with limited impact on the host are expected, as this should increase the likelihood of surviving periods of low host density. We examined the effects of Morogoro arenavirus on the survival and recapture probability of multimammate mice (Mastomys natalensis) using a seven-year capture-mark-recapture time series. Mastomys natalensis is the natural host of Morogoro virus and is known for its strong seasonal density fluctuations. Results: Antibody presence was negatively correlated with survival probability (effect size: 5-8% per month depending on season) but positively with recapture probability (effect size: 8%). Conclusions: The small negative correlation between host survival probability and antibody presence suggests that either the virus has a negative effect on host condition, or that hosts with lower survival probability are more likely to obtain Morogoro virus infection, for example due to particular behavioural or immunological traits. The latter hypothesis is supported by the positive correlation between antibody status and recapture probability which suggests that risky behaviour might increase the probability of becoming infected.

Background: In order to optimize net transmission success, parasites are hypothesized to evolve towards causing minimal damage to their reservoir host while obtaining high shedding rates. For many parasite species however this paradigm has not been tested, and conflicting results have been found regarding the effect of arenaviruses on their rodent host species. The rodent Mastomys natalensis is the natural reservoir host of several arenaviruses, including Lassa virus that is known to cause Lassa haemorrhagic fever in humans. Here, we examined the effect of three arenaviruses (Gairo, Morogoro and Lassa virus) on four parameters of wild-caught Mastomys natalensis: body mass, head-body length, sexual maturity and fertility. After correcting for the effect of age, we compared these parameters between arenavirus-positive (arenavirus RNA or antibody) and negative animals using data from different field studies in Guinea (Lassa virus) and Tanzania (Morogoro and Gairo viruses). Results: Although the sample sizes of our studies (1297, 749 and 259 animals respectively) were large enough to statistically detect small differences in body conditions, we did not observe any adverse effects of these viruses on Mastomys natalensis. We did find that sexual maturity was significantly positively related with Lassa virus antibody presence until a certain age, and with Gairo virus antibody presence in general. Gairo virus antibody-positive animals were also significantly heavier and larger than antibody-free animals. Conclusion: Together, these results suggest that the pathogenicity of arenaviruses is not severe in M. natalensis, which is likely to be an adaptation of these viruses to optimize transmission success. They also suggest that sexual behaviour might increase the probability of M. natalensis to become infected with arenaviruses.

Arenaviruses cause an estimated 300,000 – 500,000 infections annually. Currently there is no arenavirus-specific antiviral drug available to treat these infections. This study sought to use the non-pathogenic New World arenavirus Tacaribe virus (TCRV) as a model for the pathogenic Junin virus (JUNV) and Machupo virus (MACV) that cause haemorrhagic fevers in South America. TCRV was used to explore three different approaches in the search for an antiviral drug against arenavirus infection targeted specifically against the viral nucleoprotein (NP). Of the four expressed arenaviral proteins, NP is the most abundant and is thought to be of multifunctional nature involved in viral replication, suppression of the innate immune system and viral egress. The approaches to find targets for broad-spectrum anti-arenaviral drugs were high throughput screens (HTS) with purified NP using thermal shift assays, exploring the virus interactions with the innate immune system and identifying virus- host protein-protein interactions. HTS resulted in the identification of two small- molecule compounds, [5-(2-Furyl)thien-2-yl]methanol and cyclosporine A (CsA), showing broad-spectrum activity against arenaviruses. Interferon-stimulated genes (ISGs), such as IFIT3, were identified to reduce viral titres and potential 202 protein- interactions between NP and host cell proteins were identified, of which the interaction with apoptosis-inducing factor 1 (AIF1) was described further. To characterise the importance of these interactions as potential drug targets further, a TCRV reverse genetics system was constructed.

Arenaviruses are enveloped negative-strand RNA viruses that cause significant human disease. Despite decades of research, it is still unclear how these viruses establish a lifelong, asymptomatic infection in their rodent hosts while infection of humans often results in severe disease. Unable to enter a state of bona fide latency, the transcription and replication of the viral genomic RNA is likely highly regulated in time and subcellular space. Moreover, we hypothesize that the viral nucleoprotein (NP), responsible for the encapsidation of the viral RNA and the most highly expressed viral gene product, plays a key role in the regulation of the viral gene expression program. Further, exploring host-virus interactions may elucidate the basic aspects of arenavirus biology and how they cause such severe disease in humans. To explore these questions in greater detail, this dissertation has pursued three main avenues. First, to better understand lymphocytic choriomeningitis mammarenavirus (LCMV) genome replication and transcription at the single-cell level, we established a high-throughput, single-molecule (sm)FISH image acquisition and analysis pipeline and followed viral RNA species from viral entry through the late stages of persistent infection in vitro. This work provided support for a cyclical model of persistence where individual cells are initially transiently infected, clear active infection, and become re-infected from neighboring reservoir cells within the population. Second, we used FISH to visualize viral genomic RNA to describe the subcellular sites where LCMV RNAs localize during infection. We observed that, viral RNA concentrates in large subcellular structures located near the cellular microtubule organizing center and colocalizes with the early endosomal marker Rab5c and the viral glycoprotein in a proportion of infected cells. We propose that the virus is using the surface of a cellular membrane bound organelle as a site for the pre-assembly of viral components including genomic RNA and viral glycoprotein prior to their transport to the plasma membrane where new particles will bud. Last, we used mass spectrometry to identify human proteins that interact with the NPs of LCMV and Junín mammareanavirus (JUNV) strain Candid #1. We provided a detailed map of the host machinery engaged by arenavirus NPs, and in particular, showed that NP associates with the double-stranded RNA (dsRNA)-activated protein kinase (PKR), a well-characterized antiviral protein that inhibits cap-dependent protein translation initiation via phosphorylation of eIF2α. We demonstrated that JUNV antagonizes the antiviral activity of PKR completely, effectively abrogating the antiviral activity of this surveillance pathway. In sum, the work composing this dissertation has given us fresh insight into how arenaviruses establish and maintain persistence; the nature of the subcellular site where viral genomic RNA is transcribed, replicated, and assembled with other viral components; and a global view of the cellular machinery hijacked by the viral nucleoprotein. This work improves our basic understanding of the arenavirus life cycle and may suggest novel antiviral therapeutic targets that could be exploited in the future.