Translation of human immunodeficiency virus type 1 (HIV-1) mRNAs is entirely dependent on the host translation machinery. There are two prevailing hypotheses regarding the translation initiation mechanism in HIV-1; conventional cap-dependent ribosomal scanning mechanism (CDRSM) and cap-independent entry of the ribosome, usually through an internal ribosome entry site (IRES). The first mechanism makes use of the Kozak consensus sequence in locating the translation initiation codon, similar to the mechanism observed in human mRNAs. Therefore, a thorough understanding of the Kozak consensus and translation initiation in human would also shed light on the mechanism of translation initiation in HIV-1. The role of Kozak +4G site in translation initiation has been controversial, with the alternative hypothesis explaining the prevalence of +4G by invoking the observation that small amino acids, coded by G-starting codons, which are efficient for N-terminal methionine excision (NME), are preferred at the penultimate (second) position. Using two bioinformatics approaches we provide strong support for this alternative hypothesis and provide evidence contradicting the involvement of +4G in translation initiation. One of the predictions of the CDRSM hypothesis is a high conservation of Kozak consensus sequence in different HIV-1 sequences. Our results presented here validate this prediction. The CDRSM hypothesis also predicts that there should be a selective pressure against ATG usage in optimal context in the HIV-1 5'-UTR to avoid their erroneous detection by the scanning ribosome, whereas the IRES-dependent mechanism in the presence of stable secondary structures, predicts no such selective pressure because these ATGs would be embedded in the secondary structures. Here we demonstrate this selective pressure in the HIV-1 5'-UTR which further supports the CDRSM hypothesis. Finally, we present evidence for strong site conservation in the 5'-UTR of HIV-1 sequences, which not only point to as yet unknown mechanisms of translation initiation, but also provide a mean to separate HIV-1 and human mRNAs. This implies that it is theoretically possible to design HIV-1-specific translation inhibition drugs.

<p> Animal models are invaluable tools to further the understanding of human disease transmission or pathogenesis. In the context of human immunodeficiency virus (HIV), non-human primate models are frequently used to answer questions too costly, impractical or unethical to address in human clinical science. We sought to develop the model of simian immunodeficiency virus (SIV) as an alternative to simian/human immunodeficiency virus (SHIV), specifically focusing on passive immunization, by identifying and characterizing a panel of monoclonal antibodies (mAbs) targeting the viral envelope. The most promising of these mAbs were then advanced to in vivo NHP studies to characterize delivery platforms, specifically vectored immunoprophylaxis (VIP) via adenoassociated virus (AAV). Macaques were passively immunized therapeutically and prophylactically with anti-SIV Envelope mAbs. The goal of these studies was to determine if this model of passive immunization could predict outcomes of anti-HIV broadly neutralizing antibody (bNAb) based interventions against transmission or pathogenesis. The resulting protection observed in these studies highlights the discrepancies between different NHP models (SHIV vs SIV). Knowing which model to choose for a particular clinical candidate will be crucial for conserving resources, effort and time spent conducting human studies. These studies have laid the ground work for predicting the in vivo utility of NHP models for a myriad of passive vaccine concepts, including antibody infusion, VIP, chimeric antigen receptor T cells, and bi/tri-specific antibody technologies. Additionally, the characterization of VIP here has clinical implications for optimizing delivery of antibodies. Lastly, the characterization of these antibodies in vitro has revealed virology behaviors not yet observed in HIV-1. Together, this work expands the SIV model to meet the needs of the rapidly advancing field of antibody research which is poised to translate many concepts to the clinic.</p><p>

<p> The influenza virus continues to cause significant morbidity and mortality in humans, resulting in up to 50,000 deaths per year in the United States. Annual vaccination remains the recommended prophylaxis for influenza. However, vaccines must be reformulated to account for antigenic drift and, even when vaccines contain strains that antigenically match circulating strains, they display suboptimal efficacies. Two glycoproteins coat the surface of the influenza virus &ndash; the more abundant and immunodominant hemagglutinin (HA), which serves as the receptor-binding protein, and the neuraminidase (NA), an enzyme that functions to free budding viruses from infected cells. Current influenza virus vaccine strategies aim to elicit neutralizing antibodies against the HA, but past studies have demonstrated that neuraminidase inhibition titers are correlated with reduced illness and viral shedding in humans. Despite the accumulated evidence that an anti-NA immune response is beneficial, the NA content in vaccines is not standardized. </p><p> Here, the potential breadth of protection afforded by NA antibodies was investigated by studying the use of NA as a vaccine antigen and by characterizing broadly cross-reactive murine monoclonal antibodies against the NA. Using baculovirus-expressed, purified protein, it was demonstrated that vaccination with adjuvanted NA was sufficient to induce protection against lethal influenza virus challenge in mice. In the same study, the N1 NA content of inactivated influenza virus vaccines from different companies was found to be highly variable. Furthermore, in humans vaccinated with standard inactivated influenza virus vaccine, the induction of serum NA titers was significantly lower than that of HA titers. In the second part of this dissertation, panels of monoclonal antibodies were generated against the N8 NA of an emerging H10N8 influenza virus strain and against the NA of influenza B virus. Monoclonal antibodies against the influenza B virus NA displayed <i>in vivo</i> prophylactic and therapeutic protection in mice, robustly activated antibody-dependent cellular cytotoxicity (ADCC) <i>in vitro,</i> and displayed neuraminidase inhibition against an oseltamivir-resistant influenza B virus. As a whole, our data strongly suggest that targeting the influenza virus NA may be beneficial when designing novel influenza virus vaccines or antibody-based therapeutics. </p><p>

<p> Lassa virus (LASV) is the causative agent of Lassa Fever (LF), an acute and occasionally fatal disease with hemorrhagic features. Current treatments for LF are limited to ribavirin and supportive care. The objective of this dissertation project was to characterize a human monoclonal antibody (mAb) with the potential to treat LASV infections. This antibody, mAb 8.9F, was isolated from the serum of a convalescent patient and has been shown to protect guinea pigs from lethal LASV challenge. Our approach focused on confirming that mAb 8.9F bound a quaternary neutralizing epitope (QNE), determining which residues comprised the epitope, and examining how mutations to critical residues would impact viral fitness. Our results showed that mAb 8.9F shared a number of features similar to previously characterized QNEs including broad neutralization of viral subtypes and high sensitivity to epitope disruption by detergents. We found that mAb 8.9F recognizes sites on both the GP1 and GP2 subunits of the LASV glycoprotein. Both subunits dissociate from the antibody at similar rates under chaotropic conditions, suggesting that these subunits have equal binding affinity with mAb 8.9F. To identify amino acids important to the epitope, we used a method of knockout site-directed mutagenesis that replaced sequences of LASV glycoprotein with homologous sequences from lymphocytic choriomeningitis virus. We created single mutations for those regions unreactive with mAb 8.9F and identified H124F, P145R, and F147N as mutations capable of abrogating mAb 8.9F reactivity. Viral fitness of these mutants was evaluated by measuring glycoprotein processing and viral infectivity. Our results showed that mutations abrogating mAb 8.9F binding did not significantly inhibit glycoprotein production, processing, or surface transport; neither was pseudovirus formation affected. However, these mutations did significantly reduce pseudovirus infectivity suggesting that mutants escaping mAb 8.9F neutralization may be less fit. The role of N-linked glycosylation in mAb 8.9F recognition was also examined but results proved inconclusive. In summary, this work provides a detailed analysis of the first documented QNE found on the Lassa virus glycoprotein. This work will help direct rational vaccine design and post-exposure antibody therapy.</p><p>

CD4 T cells play an important and central role in the immune system, coordinating the arms of the adaptive immune system to shape an effective response while simultaneously regulating non-essential or deleterious activities. Their critical necessity is demonstrated most strikingly during acquired immunodeficiency syndrome (AIDS), when depletion of CD4 T cells by human immunodeficiency virus (HIV) type 1 ultimately results in a host of immune dysfunctions and susceptibility to opportunistic pathogens. Although virus-specific CD4 T cell responses are generally vital for the control of viral infections, HIV-specific CD4 T cells have long been recognized to be preferentially targeted and depleted by the virus—raising questions about their utility as immune effectors both during HIV-1 infection and in the context of a prospective HIV-1 vaccine. However, more recent research has challenged the notion that HIV-specific CD4 T cells are only relevant as targets. There is a growing appreciation for the crucial role that these cells may play in mediating anti-HIV immunity through a diverse array of effector functions—including direct anti-viral cytotoxicity. Here we show that HIV-specific CD4 T cell responses are evident throughout the course of HIV disease, including acute infection. In particular, an expansion of HIV-specific CD4 T cells with cytolytic potential early after acute HIV infection is associated with lower viral set point and better clinical progression. This expansion is evident both as an increase in HIV-specific CD4 T cells able to degranulate upon antigen recognition and as cells with a unique granzyme and perforin expression pattern. Further, HIV-specific cytolytic CD4 T cell responses are functionally enhanced in the setting of durable HIV control during chronic infection, where they exhibit a profile reminiscent of HIV-specific CD8 T cells and are associated with T-bet and Eomesodermin expression. We also show that in HIV controllers, the function of HIV-specific CD4 T cells is very tightly correlated with clinical status. Together these results strongly support the concept that CD4 T cells are critical players in the cellular immune response to HIV, and point to specific CD4 T cell functions—including direct cytolysis—which may be most important for anti-HIV immunity.

Merkel cell polyomavirus (MCPyV) small T antigen (ST) is expressed in the majority of Merkel cell carcinomas (MCC), a highly lethal and aggressive cancer of the skin. Since the discovery of MCPyV in 2008, the role of ST in the context of the virus and MCC has been under intense investigation. Much of our knowledge of polyomavirus ST comes from research on other polyomaviruses, including mouse polyomavirus (MPyV) and simian virus 40 (SV40). Both MPyV and SV40 ST contribute to transformation in part by binding to and inhibiting the cellular phosphatase PP2A. Likewise, MCPyV ST interacts with PP2A, although mutants that are reported to abolish this interaction still transform cells, suggesting that MCPyV ST has PP2A-independent functions. Understanding the unique cellular perturbations induced by MCPyV ST will thus be important for understanding the tumorigenesis of MCC. In this dissertation, we sought to understand the manipulation of cellular functions by MCPyV ST. We began by characterizing the MCPyV ST protein itself, starting with structural and functional comparisons with other well characterized polyomaviruses and identifying the interaction of MCPyV ST with cellular proteins. We observed that MCPyV ST uniquely interacts with the TIP60 cellular complex, which contains an ATPase and an acetyltransferase and is involved in histone modifications and DNA damage repair. Through predictions of the structure, we identified a surface-exposed region of ST, loop 4, and observed that regions in this loop were important for regulating the binding of ST to the TIP60 complex. Functionally, we investigated the role of MCPyV ST in the DNA damage response because of its interaction with TIP60 and because DNA damaging agents are used to treat MCCs. In addition, overcoming checkpoint regulation in the p53 pathway is an open question in MCPyV infection. We determined that ST increased sensitivity to DNA damage by γ-irradiation and etoposide and that expression of ST caused persistence of double strand DNA breaks (DSB) after damage, suggesting that DSB repair was delayed in ST expressing cells. Specifically, we observed that ST expression inhibits repair of breaks by nonhomologous end joining (NHEJ) but does not inhibit repair by homologous recombination (HR). These effects on the DNA damage response are explained in part by a less robust phosphorylation of DNA-PKcs at serine residue 2056, which is important for regulating end processing and repair by NHEJ. Taken together, these results indicate that MCPyV ST disrupts the cellular DNA damage response, which has implications on the viral life cycle and the initiation and treatment of MCC. / Medical Sciences

The formation of viral particles requires the coordinated assembly of both nucleic acids and proteins. In the case of Rhabdoviruses, such as vesicular stomatitis virus (VSV), the particles display a characteristic bullet-shape. VSV virions consist of the matrix protein (M), glycoprotein (G), and viral ribonucleoprotein (RNP), which contains the nucleocapsid protein (N) coated RNA bound to the large polymerase protein (L) through the phosphoprotein (P). During assembly, these components are recruited to the plasma membrane where the viral RNP undergoes condensation by M and envelopment with G containing membranes. To address whether formation of the bullet-shape requires a consistent packaging of the viral proteins, the composition of single virions was measured with fluorescence microscopy. We generated autonomously replicating VSV bearing up to 3 fluorescent protein fusions in the disordered N-terminal region of M and N-terminus of P and G. Quantification of single particles reveals that VSV assembles with a range of M, P, and G molecules, suggesting a flexible packaging mechanism. The maintenance of the bullet-shape with significantly less M proposes that condensation does not require the particle to be saturated with M. Our fluorescent VSV clones permit the tracking of viral components in live cells. We observed that assembly of M into particles requires ~2 min and can be broken into 4 stages. First, M forms a small preassembly complex. Second, M rapidly assembles into particles where its incorporation initiates before P, although they are packaged concurrently. This is followed by a delay before final release of particles into the supernatant. Late domains in M were thought to only recruit the endosomal sorting complexes required for transport (ESCRT) pathway to mediate fission. However, using our M fusions we demonstrate that these motifs are required for efficient competition into released particles and a step in assembly prior to pinching off. These constructs have permitted the study of viral assembly at the single particle level and are useful tools for studying viral entry and egress. Specifically, VSV containing M-eGFP and the lassa virus glycoprotein instead of G was used to demonstrate the requirement of a host factor for lassa virus fusion. / Medical Sciences

Epigenetic regulation plays a major role in whether the herpes simplex virus 1 (HSV-1) will initiate viral gene expression and lytic infection or instead suppress its gene expression and establish a latent infection. Prior to this study, it was known that cells respond to naked DNA by assembling chromatin to silence foreign genetic material. However, during lytic infection of epithelial cells, viral proteins VP16 and ICP0 have been implicated in limiting chromatin association and promoting euchromatic histone modifications on the HSV-1 genome. We hypothesized that the viral genome would also be subject to silencing by heterochromatin modification during lytic infection. To test this we examined the association of chromatin and heterochromatic modifications during lytic infection with WT viruses and ICP0-null mutant viruses. We found that heterochromatin modifications H3K9me3 and H3K27me3 associate initially with all viruses, but were removed rapidly during infection with WT HSV-1. ICP0-null viruses were not able to remove histones or heterochromatin, indicating a role for ICP0 in reversing epigenetic silencing. In latent infection, HSV-1 undergoes epigenetic silencing as a means to suppress gene expression and persist in neurons. Surprisingly, in this study, we find that ICP0-null viruses accumulate less heterochromatin on lytic gene promoters relative to WT viruses. This suggests that ICP0 may function to promote infection of neurons, or assist in the establishment or maintenance of latent infection. Additionally, during latency the viral genome maintains active expression from the latency-associated transcript (LAT) region, and this region retains markers of euchromatin that are excluded from the lytic viral genes. The insulator protein, CTCF, binds to a site downstream of this region between the LAT and ICP0 promoters. We find that during latent infection, deletion of this site promoted accumulation of H3K27me3 at the LAT promoter and reduced reactivation competence of the virus, but surprisingly enhanced LAT expression. This suggests that CTCF balances epigenetic repression to promote latency and maintain reactivation competence. In summary, this dissertation suggests that during lytic infection HSV reverses cell-mediated epigenetic repression and promotes viral gene expression, while during latency, the virus co-opts epigenetic mechanisms to maintain a silenced but poised genome. / Medical Sciences

The innate immune sensor MDA5, a RIG-I-like receptor (RLR), is critical for the detection of viral nucleic acid, eliciting an antiviral immune response. Aberrant immune activation can be detrimental to the host; therefore, RLR activity is strictly regulated. In the uninfected cell, MDA5 is constitutively phosphorylated at S88, preventing antiviral signaling. Upon sensing of viral RNA, MDA5 is activated via dephosphorylation by the phosphatases PP1a/c. Dephosphorylation of MDA5 allows interaction with the mitochondrial adaptor protein MAVS, inducing downstream signaling, leading to the production of antiviral cytokines including type-I interferons (IFNs). Many viruses have evolved sophisticated mechanisms to avoid detection by RLRs. Here, we present a novel evasion mechanism of paramyxoviruses to escape the MDA5-induced innate immune response: inhibition of its key regulators, PP1a and PP1c. The V proteins of measles virus (MV) and multiple other paramyxoviruses interact with PP1a/c, preventing MDA5 S88 dephosphorylation and subsequent innate immune signaling. In Chapter 2, we identify a conventional PP1-binding motif in the unique C-terminal region of the MV V protein which mediates this interaction. Mutation of this motif abrogates PP1 binding and MDA5 antagonism without effecting other known activities of the V protein such as STAT inhibition. To determine the physiological relevance of the V-PP1 interaction for MDA5 antagonism, we generated a recombinant MV carrying a PP1-binding deficient V protein. This mutant virus no longer suppressed MDA5 dephosphorylation by PP1, resulting in increased expression of IFN and IFN-stimulated genes (ISGs) and impaired replication in lung epithelial and dendritic cells compared to the parental virus. In Chapter 3, we expand our understanding of paramyxovirus antagonism of MDA5 by examining the virus-specificity of inhibition of S88 dephosphorylation. We found that the ability to interact with PP1 and inhibit MDA5 S88 dephosphorylation is shared by multiple viruses, including mumps (MuV), Nipah, and Hendra viruses. We mapped the PP1 interaction to a minimal binding region in the MuV-V C-terminal domain. This region contains a putative PP1-binding motif which is not conserved between viruses. The importance of this putative motif requires additional investigation. Together, our findings reveal PP1 antagonism as a novel immune evasion strategy of paramyxoviruses. / Medical Sciences