Proper regulation of cellular proliferation is critical for normal development and cancer prevention. Most, if not all, cancers contain mutations in the Rb/E2F pathway, which controls cellular proliferation. Inactivation of the retinoblastoma protein (Rb) can occur through Rb loss, mutation, or inactivation by cellular or viral oncoproteins leading to unrestrained proliferation. This occurs primarily by de-repression and activation of the E2F transcription factors, which promote the transition of cells from the G1to S phase of the cell cycle. In order to protect against loss of growth control, the p53 tumor suppressor is able to induce programmed cell death, or apoptosis, in response to loss of proper Rb cell cycle regulation. E2F1 serves as the primary link between the Rb growth control pathway and the p53 apoptosis pathway. While the pathway(s) linking E2F1 to p53 activation and apoptosis are unclear, it has been proposed that E2F1 activates p53-dependent apoptosis by transactivation of p19ARF leading to inhibition of Mdm2-promoted degradation of p53. We tested this hypothesis, and found that p19ARFis not required for E2F1-induced apoptosis. Instead, we find that expression of E2F1 leads to covalent modifications of p53 that correlate with p53 activation and are required for apoptosis. The observation that E2F1 induces covalent modification of p53 is consistent with the p53 modifications observed following DNA damage. We therefore hypothesized that E2F1 may be activating components of the DNA damage response to activate p53 and kill cells. Consistent with the DNA damage response, we find that E2F1-induced apoptosis is compromised in cells from patients with the related disorders ataxia telangiectasia and Nijmegen breakage syndrome, lacking functional Atm and Nbs1 gene products, respectively. E2F1-induced apoptosis and p53 modification also requires the human checkpoint kinase Chk2, another component of the DNA damage response. We find that the commitment step in E2F1-induced apoptosis is the induction of Chk2. Having found that E2F1 requires DNA damage kinases to induce apoptosis, we next examined events upstream of kinase activation. To this end, we observe relocalization of the DNA damage repair MRN complex (composed of Mre11, Nbs1, and Rad50) to nuclear foci specifically following expression of E2F1. Expression of E2F1 also induces relocalization of the DNA damage recognition proteins γH2AX and 53BP1 to nuclear foci, consistent with the location of these complexes observed following DNA double strand breaks. As a consequence of activating some or all of these DNA damage signaling proteins, expression of E2F1 blocks cell cycle progression in diploid human fibroblasts. The observed block in cell cycle progression is found to be, in part, due to activation of a p21-dependent cell cycle checkpoint. The E7 protein from the oncogenic human papillomavirus (HPV) is able to bind to and inactivate members of the Rb family. HPV infects quiescent, non-cycling cells that lack expression of DNA replication machinery that is essential for replication of the viral genome. By expression of the E7 protein, HPV is able to bypass normal Rb-mediated growth control and induce quiescent cells to enter S phase where the host cell DNA replication enzymes are present for viral replication. We find that expression of E7 can also result in apoptosis that is dependent specifically on E2F1. Additionally, E7-induced apoptosis, like E2F1-induced apoptosis, requires Atm, Nbs1, and Chk2. Expression of E7, like that of E2F1, induces E2F1-dependent covalent modification of p53 that correlates with apoptosis induction. These findings demonstrate that deregulation of the Rb/E2F growth control pathway leads to activation of an apoptosis program with some similarity to the pathways activated by DNA damage. Our observations suggest that E2F1 not only functions as a sensor for deregulation of Rb, but may also play an important role in regulating cellular growth control in response to other oncogenic stimuli.

Generation of the primitive neuroectoderm into specialized brain subdivisions, such as the hindbrain primordium, involves the regulated coordination of complex morphogenetic and molecular mechanisms. These processes are evident in the segregation of the zebrafish hindbrain into seven distinct lineage-restricted compartments, termed rhombomeres (r), which are established by the interplay of several spatially-restricted expressed genes. These include transcription factors, members of specific signaling pathways and specialized molecules that mediate cell adhesion and identity. Despite their extensive characterization, it is evident that other genes are involved to mediate the proper specification and segregation of individual rhombomeres. One candidate that likely fits this role is related to the no ocelli/l(2)35Ba gene in Drosophila, termed nlz (nocA-like zinc-finger). Nlz-related proteins behave as transcriptional repressors and are related to the vertebrate Sp1-like family of transcription factors. nlz is dynamically expressed in the zebrafish hindbrain, residing in the caudal hindbrain at gastrula stages and rostrally expanding from presumptive r3/r4 boundary to encompass r3 and r2 at segmentation stages. Nlz localizes to the nucleus and associates with the co-repressors Groucho and histone deacetylases, suggesting that Nlz acts as a repressor. Consistent with this, misexpression of nlz into zebrafish embryos results in a loss of gene expression in the rostral hindbrain (rl-r3). Taken together, the findings in this thesis suggest that Nlz functions as a transcriptional repressor to control segmental gene expression in the rostral hindbrain.

Influenza A virus infections are a major cause of morbidity and mortality in the United States and throughout the world. The current vaccine elicits primarily a humoral response that is specific for the external glycoproteins hemagglutinin (HA) and neuraminidase (NA). However, these are the viral proteins that are most susceptible to antigenic shift and drift, and can evade the humoral response. Cytotoxic T lymphocytes (CTL) recognize and lyse virus-infected cells and are important in clearing influenza A virus infections. CTL can recognize epitopes on both the external glycoproteins and the more conserved internal viral proteins. This thesis investigates the hypothesis that there is a broad CTL memory response in humans, and, if boosted by vaccines, these CTL may help clear influenza A virus strains of different subtypes. The CTL repertoire specific for influenza A viruses reported in inbred mice is extremely limited and has focused on a few immunodominant epitopes. We perfonned preliminary bulk culture chromium release assays using human peripheral blood mononuclear cells (PBMC) stimulated with influenza virus strain A/PR/8/34 (H1N1) in vitro. CTL activity was observed against autologous B-lymphoblastoid cell lines (B-LCL) infected with vaccinia constructs that expressed several influenza A viral proteins, including nucleoprotein (NP), matrix (M1), nonstructural 1 (NS1) and polymerase (PB1). This was more diverse than the limited response reported in inbred mice. To further characterize the CTL repertoire in humans, PBMC from healthy adult donors were stimulated and CTL were cloned by limiting dilution. Isolated cell lines were further characterized by their CD4/CD8 surface expression, histocompatibility leukocyte antigen (HLA) restriction, cross-reactive or subtype-specific influenza A subtype recognition, and epitope recognition. CTL lines isolated from three donors recognized epitopes on many different influenza virus proteins. The ELISPOT assay was used to identify the number of IFN-γ- secreting cells and determine the precursor frequency of the CTL specific for the epitopes that were mapped. The precursor frequency of IFN-γ producing CTL ranged from 1 in 4,156 PBMC to 1 in 31,250 PBMC. The precursor frequency for one epitope was below the level of detection of this assay, but most of the memory CTL were readily detected. The cross-reactive or subtype-specific recognition of various human influenza A subtypes by these T cell lines was determined by chromium release assays. Most of the CTL lines recognized B-LCL infected with any of the three influenza A subtypes that have caused epidemics in the last century (H1N1, H2N2, and H3N2) and recognized epitopes on conserved internal influenza viral proteins. Most of the subtype-specific cell lines recognized the surface HA or NA glycoproteins, which are not well conserved between influenza subtypes. Although most of the T cell lines that were characterized were cross-reactive with influenza viruses of human origin, infection of humans with a divergent swine or avian derived strain could cause a global pandemic. To study the human CTL responses to non-human influenza viruses, B-LCL were infected with an Hsw1N1 influenza A virus of swine origin, and cell lines were tested for recognition of these targets in a chromium release assay. Most cell lines lysed the targets infected With the Hsw1N1 subtype to the same degree as targets infected with the human H1N1 strain. Two influenza viruses of duck origin were also tested and were recognized by many of the cell lines. The subtypes of these duck strains were Hav1N1 and H5N2. The isolates of influenza A virus from the Hong Kong outbreak of 1997 were also used to infect targets and analyze recognition by these CTL. We found that approximately 50% of the human T cell lines tested recognized both of the Hong Kong isolates, 25% recognized at least one isolate, and 25% recognized neither isolate to the same degree as the A/PR/8/34 (H1N1) virus. We analyzed the amino acid (aa) changes in the epitopes of the T cells lines from the 25% of cell lines that did not recognize either Hong Kong virus isolate. Non-conservative mutations were found in all of the epitopes that lost recognition by the human CTL lines. Bulk cultures of PBMC from three donors that were stimulated with A/PR/8/34 (H1N1) influenza A virus of human origin recognized all of the non-human virus strains tested. Thus, humans have memory CTL that recognize influenza viruses of avian and swine species. This may provide a second line of defense against influenza infection in case of exposure to a novel influenza A virus derived from these species. These results made it clear that humans have broad CTL memory to influenza A virus. In order to determine whether these T cells could be boosted in a vaccine, immune-stimulatory complexes (Iscom) incorporating inactivated influenza particles were tested in vitro. Iscoms containing inactivated influenza A vaccine (Flu-Iscom) were used to pulse autologous B-LCL overnight that were then used as targets in chromium release assays with human CTL lines as effectors. A CD8+ HA-specific CTL line lysed these targets, but not targets pulsed with Iscoms alone or with inactivated influenza A vaccine alone. An NS1-specific cell line recognized targets pulsed with NS1 protein and Iscoms, but not targets pulsed with Iscoms or NS1 protein alone. Therefore, CTL could recognize in vitrotarget cells that were exposed to the Iscom vaccines containing their specific epitope. Flu-Iscom and Iscom mixed with inactivated influenza virus particles (Flu-Iscomatrix) were then used as vaccines in a clinical trial to test CTL and neutralizing antibody induction against influenza. Fifty-five donors were bled pre-vaccination, and on days 14 and day 56 post-vaccination. Bulk culture chromium release assays were performed using targets infected with live vaccine strain viruses. There were significantly more increases in the influenza A specific CTL activity in the PBMC of donors that were vaccinated with the Flu-Iscom and Flu-Iscomatrix vaccines than in recipients of the standard vaccine. In order to determine whether these increases in cytotoxicity were due to an increase in the precursor frequency of influenza specific CTL, the PBMC were used in ELISPOT assays to assess the changes pre-and post-vaccination. When there was an increase in the level of cytotoxicity detected in bulk culture CTL, there was often also an increase in the precursor frequency of influenza-specific CTL. Peptide-specific increases in the number of CTL that recognize epitopes such as M1 aa 58-66 were detected in several donors confirming the increase in influenza-specific CTL post-vaccination. Another type of T cell that may be involved in defense against viruses is the γδ T cell. T cells expressing the γδ T cell receptor (TCR) have been found extensively in mucosal tissues in mice and humans. Influenza A viruses enter via the airway tract, infecting the epithelial cells at the mucosal surface. These epithelial cells have been shown in vitro to be targets for influenza-specific cytolytic recognition of αβ T cells. To analyze whether γδ T cells can respond to influenza A-infected APCs, PBMC were stimulated with influenza A virus. Intracellular IFN-γ staining was used to determine whether γ/δ T cells can secrete IFN-γ in response to the influenza A virus infection. We observed an increase in the percentage of γ/δ T cells secreting IFN-γ post-influenza A virus infection of PBMC compared to uninfected or allantoic fluid-stimulated cultures. These T cells also upregulated CD25 and CD69 in response to live influenza A virus. We focused on the responses in the CD8- population of γδ T cells, which are the majority of γδ T lymphocytes. Furthermore, the increases in IFN-γ production and activation marker expression were much more clear in the CD8- γδ+ T cells. The level of CD8- γδ T cell activation with inactivated influenza A virus was much less, and in some cases no higher than uninfected PBMC. The CD8+ αβ and γδ responses could be partially blocked by anti-class I antibodies, but the CD8- γδ responses could not. Vaccinia virus infection did not activate the CD8- γδ T cells to the same degree as influenza virus infection. γδ T cells are thought to have a regulatory role that includes the secretion of cytokines and epithelial growth factors to help restore tissue back to health. Humans have broad multi-specific T lymphocyte responses by αβ T cells to influenza A viruses and those responses are cross-reactive with human, avian, and swine virus strains. These CTL can be activated in vitro and boosted in number in vivo by Iscom incorporating vaccines. There is also a population of γδ+ T lymphocytes in humans that responds to influenza virus infection by producing cytokines and becoming activated. Increasing memory CTL as a second line of defense against influenza A viruses may be important in future vaccine development.

The sterol biosynthetic pathway is essential for growth of all eukaryotic cells and the main target of antifungal agents. The emergence of resistance to these antifungals in an already ill patient population indicates a need to develop drugs that have a broad spectrum of activity among pathogenic fungi and have minimal patient toxicity. Squalene synthase is the first committed step in the sterol pathway and has been studied intensively for development of antifungal agents. While the overall architecture of this enzyme is identical throughout eukaryotes, it was shown that plant and animal genes cannot complement a squalene synthase knockout mutation in yeast unless the carboxy-terminal domain is swapped for one of fungal origin. This implies that there is a component of the fungal carboxy-terminal domain that is responsible for the complementation phenotype and that is unique to the fungal kingdom of life. To determine the role of the carboxy-terminal domain of squalene synthase in the sterol pathway, we used the yeast Saccharomyces cerevisiae with a squalene synthase knockout mutation and expressed squalene synthases originating from fungi, plants, and animals. In contrast to previous observations, all enzymes tested could partially complement the knockout mutation when the genes were weakly expressed. When induced, non-fungal squalene synthases could not complement the knockout mutation and instead led to the accumulation of carboxysterol intermediates, suggesting an interaction between squalene synthase and the downstream sterol C4-decarboxylase. Overexpression of a sterol C4-decarboxylase from any kingdom of life both decreased the accumulation of carboxysterol intermediates and allowed non-fungal squalene synthases to complement the squalene synthase knockout mutation. Using chimeric squalene synthases from each kingdom of life, the motif in the C-terminal domain responsible for preventing this toxicity was mapped to a kingdom-specific 26-amino acid hinge motif adjacent to the catalytic domain. Furthermore, over-expression of the carboxy-terminal domain alone containing a hinge motif from fungi, not from animals or plants, led to growth inhibition of wild-type yeast. Since this hinge region is unique to and highly conserved within each kingdom of life, this data provides evidence for the development of an antifungal therapeutic as well as for tools to develop an understanding of triterpene catalytic activity and identify similar motifs in other biosynthetic pathways.

Development of functional neural circuits involves a series of complicated steps, including neurogenesis and neuronal morphogenesis. To understand the molecular mechasnims of neurite complexity, especially neurite branching/arborization, the Drosophila brain, especially MBNs (mushroom body neurons) and PNs (projection neurons) in olfactory circuitry, was used in this dissertation work as the model system to study how two molecules, Dscam and Kr-h1 affect neurite complexity in the Drosophilabrain. For the Drosophila Dscam, through alternative splicing it could encode up to 152,064 distinct immunoglobulin/fibronectin type cell adhesion molecules. Each Dscam isoform is derived from one of the 19,008 ectodomain variants connected with one of the two alternative transmembrane segments and one of the four possible endodomain portions. Recent studies revealed that Dscam was widely required for neurite branching/arborizaiton. However, due to the technical difficulty, the functional roles of Dscam transmembrane variants and ectodomain variants remain unclear. In this thesis work, a microRNA based RNA interference was used to knock down distinct subsets of Dscam isoform. First, loss of Dscam[TM1] versus Dscam[TM2], two distinct Dscam transmembrane variants, disrupted the dendritic versus axonal morphogenesis, respectively. Furthermore, structural analysis suggested that the juxtamembrane portion of transmembrane segment was required for the Dscam protein targeting in dendrites/axons and this differential protein targeting might account for the functional distinction between Dscam[TM1] and Dscam[TM2]. Second, to further address the functional significance of having two Dscam transmembrane variants in axons versus dendrites, the possibility that there might be different usage of Dscam repertoire between axons and dendrites that lead to different levels of morphological complexity between axons and dendrites in the same neuron was examined. To this end, end-in targeting approaches were used to exchange Dscam populations between axons and dendrites. Though the genetic data suggested that Dscam populations were exchanged between axons and dendrites, the phenotypic analysis in various neuronal types revealed that depending on the neuronal types, exchange of Dscam populations between axons and dendrites might primarily affect either axonal or dendritic morphology, suggesting that different usage of Dscam population between axons and dendrites might regulate complex patterns of neurite morphology. Finally, the functions of Dscam exon 4 variants had been addressed in different model neurons in the Drosophilabrain. First, 12 Dscam exon 4 variants were divided into three groups based on their phylogenetic distance. Then, three miRNA constructs were engineered to knock down one group at a time. The genetic data suggested that different Dscam exon 4 variants are differentially required in different neurons to support their proper neuronal morphogenesis. In summary, this part of my thesis work identified and characterized previously unrecognized functions of all these distinct Dscam variants and provided novel insights into how diverse Dscam isoforms regulate the different aspects of neuronal morphogenesis. In the honey bee brain, Kr-h1 is upregulated during the behavioral shift from nursing to foraging when there is increased neurite branching in the brain. To directly examine the hypothesis that altered Kr-h1 expression might regulate morphological complexity of neurites, this research work involved the MARCM (mosaic analysis with a repressible cell marker) and TARGET (temporal and regional gene expression targeting) techniques to analyze the roles of Kr-h1 in Drosophila neuronal morphogenesis. Interestingly, increased expression of Kr-h1 blocked the axon branching and further disrupted the lobe formation in the mushroom body whereas the loss-of-Kr-h1 did not show any apparent neuronal morphogenetic defects. In addition, it was observed that Kr-h1 was expressed when MB (mushroom body) did not undergo active morphogenesis, suggesting its potential anti-morphogenetic activity. Indeed, loss of Kr-h1 (Kruppel homolog 1) enhanced the neuronal morphogenesis that was otherwise delayed due to the defective TGF-beta signaling. Furthermore, Kr-h1 expression was closely linked to ecdysone dependent signaling: Kr-h1 was first regulated by usp (ultraspiracle), which dimerized with various ecdysone receptors and then Kr-h1 expression was essential for proper ecdysone patterning in the larval CNS (central nervous system). Together, though Kr-h1could potentially regulate the neurite complexity, it seems primarily involved in the coordinating ecdysone signaling. In conclusion, the powerful genetic toolkit available in the Drosophila has allowed the investigation in the molecular mechanisms of neuronal morphogenesis and understanding of these mechanisms will enhance our understanding of how the complex nervous system is wired to perform the delicate behaviors.

Argonaute proteins are the core component of an RNA silencing complex. The human genome encodes four Argonaute paralogs –Ago1, Ago2, Ago3 and Ago4– proteins that are guided to target mRNAs by microRNAs. More than 500 miRNAs are conserved between mammals, and each microRNA can repress hundreds of genes, regulating almost every cellular process. We still do not fully understand the molecular mechanisms by which miRNAs regulate gene expression. Although we understand many aspects of microRNA biogenesis and formation of the RNA-induced silencing complex, much less is known about the subsequent steps leading to target mRNA regulation. Mammalian microRNAs rarely have complete complementarity to their target mRNAs so, instead of endonucleolytic cleavage by Ago2, microRNAs destabilize or repress translation of target mRNAs. Here I explored the functional limits of Argonaute proteins bound to their targets directly and indirectly through microRNAs in mammalian cells. I revealed the different abilities for Argonaute proteins bound at multiple sites in a target to generate cooperativity in silencing based on the extent of pairing between the microRNA and target mRNA. Further, I harnessed the endogenous microRNA silencing mechanism to repress an mRNA that is not a direct target of the microRNA by tethering the RNA-induced silencing complex to the 3´ UTR of an mRNA. This strategy allows tissue-specific gene silencing due to the limited endogenous expression profile of the recruited microRNA. Efforts made herein further our mechanistic knowledge of microRNA-induced gene silencing in mammalian cells and advance microRNA-based strategies toward treating human disease.

The first step of viral infection requires the binding of the viral attachment protein to cell surface receptors. Following binding, viruses penetrate the cellular membrane to deliver their genome into the host cell. For enveloped viruses, which have a lipid bilayer that surrounds their nucleocapsids, entry into the host cell requires the fusion of viral and cellular membranes. This process is mediated by viral glycoproteins located on the surface of the virus. For many enveloped viruses, such as influenza, Ebola, and human immunodeficiency virus, the fusion protein is responsible for mediating both attachment to cellular receptors and membrane fusion. However, paramyxoviruses are unique among fusion promoting viruses because their receptor binding and fusion activities reside on two separate proteins. This unique distribution of functions necessitates a mechanism by which the two proteins can transmit the juxtaposition of the viral and host cell membranes, mediated by the attachment protein (HN/H), into membrane fusion, mediated by the fusion (F) protein. This mechanism allows for paramyxoviruses to gain entry into and spread between cells, and therefore, is an important aspect of virus infection and disease progression. Despite the conservation of receptor binding activity among members of the Paramyxovirinaesubfamily, for most of these viruses, including Newcastle disease virus (NDV), heterologous HN proteins cannot complement F in the promotion of fusion; both the HN and F proteins must originate from the same virus. This is consistent with the existence of a virus-specific interaction between the two glycoproteins. Thus, one or more domains on the HN and F proteins is thought to mediate a specific interaction between them that is an integral part of the fusion process. Therefore, the primary focus of this thesis is the identification of the site(s) on HN that directly contacts F in the HN-F interaction. The ectodomain of the HN protein consists of a stalk and a terminal globular head. Analysis of the fusion activity of chimeric paramyxovirus HN proteins indicates that the stalk region of HN determines its F protein specificity. The first goal of this research was to address the question of whether the stalk not only determines F-specificity, but does so by directly mediating the interaction with F. To establish a correlation between the amount of fusion and the extent of the HN-F interaction, a specific and quantitative co-immunoprecipitation assay was used that detects the HN-F complex at the cell surface. As an initial probe of the role of the HN stalk in mediating the interaction with F, N-glycans were individually added at several positions in the region. N-glycan addition at positions 69 and 77 in the stalk specifically and completely block both fusion and the HN-F interaction without affecting either HN structure or its other activities. However, though they also prevent fusion, N-glycans added at other positions in the stalk also modulate activities that reside in the globular head of HN. This correlates with an alteration of the tetrameric structure of the protein as indicated by sucrose gradient sedimentation analyses. These additional N-glycans likely indirectly affect fusion, perhaps by interfering with changes in the conformation of HN that link receptor binding to the fusion activation of F. To address the issue of whether N-glycan addition at any position in HN would abolish fusion, an N-glycan was added in another region at the base of the globular head of HN (residues 124-152), which was previously predicted by a peptide-based analysis to mediate the interaction with F. HN carrying this additional N-glycan exhibits significant fusion promoting activity, arguing against this site being part of the F-interactive domain in HN. These data support the idea that the F-interactive site on HN is defined by the stalk region of the protein. Site-directed mutagenesis was used to begin to explore the role of individual residues in the stalk in the interaction with F. The characteristics of the F-interactive domain in the stalk of HN are that it is a conserved motif with enough sequence heterogeneity to account for the specificity of the interaction. One such region that meets these requirements is the intervening region (IR) (residues 89-95); a non-helical domain situated between two conserved heptad repeats. Several amino acid substitutions for a completely conserved proline residue in this region impair not only fusion and the HN-F interaction, but also decrease neuraminidase activity in the globular domain and alter the structure of the protein, suggesting that the substitutions indirectly affect the HN-F interaction. Substitutions for L94 also interfere with fusion, but have no significant effect on any other HN function or its structure. Amino acid substitutions at two other positions in the IR (A89 and L90) also modulate only fusion. In all cases, diminished fusion correlates with a decreased ability of the mutated HN protein to interact with F at the cell surface. These findings indicate that the IR is critical to the role of HN in the promotion of fusion and are consistent with its direct involvement in the interaction with the homologous F protein. These are the first point mutations in the HN protein for which a correlation has been demonstrated between the extent of the HN-F interaction and the amount of fusion. This argues strongly that the co-IP assay is an accurate reflection of the HN-F interaction. The second goal of this research was to address the HN-F interaction from the perspective of the F protein by investigating the relationship between receptor binding, the HN-F interaction, and fusion using a highly fusogenic form of the F protein. It has previously been shown that an L289A substitution in NDV F eliminates the requirement for HN in the promotion of fusion and enhances HN-dependent fusion above wild-type (wt) levels. Here, it was shown that the HN-independent fusion exhibited by L289A-F in Cos-7 cells cannot be duplicated in BHK cells. However, when L289A-F is co-expressed with wt HN, enhanced fusion above wt levels is observed in BHK cells. Additionally, when L289A-F is co-expressed with IR-mutated HN proteins previously shown to promote low levels of fusion with wt F, a 2.5-fold increase in fusion was observed. However, similar to wt F, an interaction between L289A-F and the IR-mutated HN proteins was not detected. These results imply that the attachment function of HN, as well as the conformational change in L289A-F, are necessary for the enhanced level of fusion exhibited by HN proteins co-expressed with L289A-F. Indeed, two MAbs detected a conformational difference between L289A-F and the wt F protein. These findings support the idea that the L289A substitution converts F to a form that is less dependent on an interaction with HN for conversion to the fusion-active form. The last goal of this research was to address the cellular site of the HN-F interaction, still a controversial issue based on conflicting data from studies of different paramyxoviruses, using various approaches. This is a particular point of interest, as it speaks to the mechanism by which the HN-F interaction regulates fusion. Thus, NDV HN and F were successfully retained intracellularly with a multiple arginine or KK motif, respectively. The results of Endoglycosidase H resistance and F cleavage studies indicate that the mutated proteins, HN-ER and F-ER, are retained in a compartment prior to the medial-Golgi apparatus and that they are unable to interact with a high enough affinity to co-retain or even cause reduced transport of their wt partner glycoproteins. This is consistent with the HN-F interaction occurring at the cell surface, possibly triggered by receptor binding. In conclusion, this thesis presents evidence to argue that the IR in the stalk of the NDV HN protein directly mediates the interaction with the F protein that is necessary for fusion. Overall, the data presented in this thesis extend the current knowledge of the mechanism by which the paramyxovirus attachment protein can trigger the F protein to initiate membrane fusion. A clear understanding of this process has the potential to identify new anti-viral strategies, such as small molecule inhibitors, aimed at controlling paramyxovirus infection by interfering with early steps in the virus infection cycle.

The Tec family tyrosine kinase, Itk, plays an important role in signal transduction following T cell receptor engagement. Several prior studies have established the importance of Itk in immune system processes, such as T cell development and T cell activation. Additional biochemical studies have found that Itk specifically functions within a multi-molecular signalosome complex, which ultimately functions to provide a platform by which Itk can phosphorylate and activate PLC-γ1, a crucial step in T cell activation. To further study how Itk regulates distinct immune outcomes via T cell effector processes within the peripheral immune system, and to further understand how Itk functions in T cells in response to a physiological ligand-receptor interaction, I crossed Itk-deficient mice to mice transgenic for a TCR specific for a moth cytochrome C peptide. My studies have established a unique role for Itk in several important aspects of T cell function. Following T cell activation, I identified an imperative role for Itk in activation-induced cell death via FasL, a mechanism of immune homeostasis. Furthermore, I found Itk plays a unique role in the process of T cell differentiation, where Itk positively regulates the induction of cytokine genes, such as IL-4, while negatively regulating the induction of T-bet, a transcription factor important for Th1 differentiation. Lastly, following T cell differentiation, I found that Itk mRNA and protein are up-regulated during Th2 differentiation, while Rlk, a related Tec kinase, disappears rapidly from Th2 cells, indicating a critical role for Itk in Th2 cell function. Collectively, my thesis work has more clearly defined an important function for Itk not only in TCR signaling, but also in immune processes such as T cell differentiation and activation-induced cell death that are required for proper immune function.

Many intricate and highly conserved mechanisms have evolved to safeguard organisms against errors in gene expression. The nonsense-mediated mRNA decay pathway (NMD) exemplifies one such mechanism, specifically by eliminating mRNAs containing premature translation termination codons within their protein coding regions, thereby limiting the synthesis of potentially deleterious truncated polypeptides. Studies in Saccharomyces Cerevisiae have found that the activity of at least three trans-acting factors, known as UPF1, UPF2/NMD2, and UPF3is necessary for the proper function of the NMD pathway. Further research conducted in yeast indicates that the degradation of substrates of the NMD pathway is dependent on their translation, and that the sub-cellular site of their degradation in the cytoplasm. Although most evidence in yeast suggests that substrates of the NMD pathway are degraded in the cytoplasm while in association with the translation apparatus, some mammalian studies have found several mRNAs whose decay appears to occur within the nucleus or before their transport to the cytoplasm has been completed. In addition, study of the mammalian TPI mRNA found that this transcript was unavailable as a substrate for the NMD pathway once it had been successfully exported to the cytoplasm, further supporting the notion that the degradation of mammalian substrates of the NMD pathway occurs in association with the nucleus, or during export from the nucleus to the cytoplasm. To determine if yeast cytoplasmic nonsense-containing mRNA can become immune to the NMD pathway we examined the decay kinetics of two NMDS substrate mRNAs in response to repressing or activating the NMD pathway. Both the ade2-1 and pgk1-UAG-2nonsense-containing mRNAs were stabilized by repressing this pathway, while activation of NMD resulted in the rapid and immediate degradation of each transcripts. These findings demonstrate that nonsense-containing mRNAs residing in the nucleus are potentially susceptible to NMD at each round of translation. The remainder of this thesis utilizes protein overexpression studies to gain understanding into the function of factors related to the processes of nonsense-mediated mRNA decay and translation in Saccharomyces cerevisiae. Overexpression of a C-terminal truncated form of Nmd3p was found to be dominant-negative for cell viability, translation and the normal course of rRNA biogenesis. Overexpression studies conducted with mutant forms of the nonsense-mediated mRNA decay protein Upf1p, found that overexpression of mutants in the ATP binding and ATP hydrolysis region ofUpflp were dominant-negative for growth in an otherwise wild-type yeast strain. Furthermore, overexpression of the ATP hydrolysis mutant of Upf1p (DE572AA), resulted in the partial inhibition of NMD and a general perturbation of the translation apparatus. These results support previous studies suggesting a general role for Upf1p function in translation.

In mammals, newly synthesized proteins destined for secretion are translocated cotranslationally into the lumen of the Endoplasmic Reticulum (ER). Once inside, these nascent polypeptide chains are bound by a lumenal ER protein called BiP (Immunoglobulin Binding Protein) or Grp 78 (Glucose Regulated Protein 78). It is hypothesized that this binding is necessary to protect the nascent chains until they are properly folded or assembled with other subunits. When the proteins are folded and assembled, they are released from BiP by a process that is dependent on ATP hydrolysis. Since ATP is synthesized mainly in the mitochondria, we hypothesized that there must be an ATP transporter in the ER which would allow the protein mediated transport of ATP from the cytosol into the ER lumen. We studied the transport of ATP in vitro and found that ATP enters the lumen of the ER in a saturable manner with a Kmapp~3μM. ATP transport is dependent on time, protein, and vesicle integrity, it is also inhibited by the general anion transport inhibitor, 4,4' diisothiocyano-2,2'-disulfonic acid stilbene (DIDS). We also found that the transport was inhibited by membrane impermeable protein modifying agents such as N-ethlymaleamide (NEM) and Pronase when added to intact ER vesicles. These results suggest that the transport is mediated by a protein with an active cytoplasmic face. Using monoclonal and polyclonal antibodies to BiP and Grp94 (another resident ER protein) and U.V. crosslinking, we demonstrated that after transport of ATPα32P into intact vesicles, radiolabeled BiP and Grp94 could be immunoprecipitated. We also found that labeling of lumenal proteins with ATP is dependent on the transport of ATP. Finally using ATP labeled with 35S, we concluded that BiP was able to bind intact ATP and we confirmed earlier work that BiP was thiophosphorylated while Grp94 is not. The second area of study involves processes that occur further along the secretory pathway in the Golgi apparatus. It was known from previous work that the nucleotide sugar substrates necessary for the synthesis of the linkage region, UDP-xylose (UDP-Xyl), UDP-galactose (UDP-Gal) and UDP-glucuronic acid (UDP-GlcA) were transported into the Golgi apparatus from the cytosol via protein mediated transporters. In order to eventually purify one of these transporter proteins, we wanted to reconstitute their activities. We were able to reconstitute the activities that exhibited kinetic parameters and inhibitor sensitivities very similar to those seen in intact Golgi vesicles. In the case of UDP-xylose it was necessary to prepare the liposomes using endogenous Golgi lipids in order to get transport activity similar to that seen in the intact Golgi vesicles. This suggested a specific lipid requirement for the UDP-xylose transporter. These transporters seem to be antiporters, whereby the nucleotide sugar enters the lumen of the Golgi coupled to the equimolar exit of the corresponding nucleoside monophosphate (Hirschberg, C.B. and Snider, M.D. 1987). We also showed that we could reproduce the hypothesized antiporter system in the reconstituted proteoliposomes by preloading the proteoliposomes with the putative antiporter molecule UMP. The rationale for developing the reconstituted system is eventually to use this system to purify one of these nucleotide sugar translocators. In the last set of studies, I have shown that this reconstituted system can be used to monitor the purification of the UDP-galactose translocator. Using column chromatography we were able to purify this membrane translocator protein 45,000 fold from a rat liver homogenate.