Search

Tuberculosis (TB) is a critical infectious disease world-wide, and the increasing development of antibiotic resistance drives the search for effective host-directed therapies. One molecular target of potential host-directed therapy is Type 1 Interferon, (IFN-I or IFNβ), an excess of which correlates with TB progression. The mechanisms underlying IFNβ overproduction are still unclear. In this dissertation we review cellular mechanisms, including mitochondrial function and metabolism, oxidative stress, and the Integrated Stress Response, which are involved in IFNβ production and macrophage function. We also describe an experimental model of human-like TB, the B6J.C3-Sst1C3HeB/Fej Krmn (B6.Sst1S) mouse, which provides a unique and convenient system for studying mechanisms of necrosis in TB granulomas. We use primary macrophages from the B6.Sst1S mouse to establish a mechanism that links the B6.Sst1S genotype to a cascade of dysregulation that drives IFNβ superinduction and susceptibility to TB infection. TNF is necessary for granuloma formation in vivo, but in the context of transcriptional dysregulation and excess free iron, it drives oxidative stress, which amplifies IFNβ induction to pathologic levels. This induction is maintained by positive feedback through the double stranded RNA-dependent Protein Kinase (PKR). We demonstrate that interruption of this cascade by iron chelation or inhibition of lipid peroxidation attenuates IFNβ induction and improves subsequent infection outcomes. We conclude by comparing the in vitro model system to an in vivo necrotic TB granuloma, describing similarities between our system and human TB, and discussing the connections between IFN-I and autoimmune and degenerative disease and the broader application of the B6.Sst1S model system to studies of human immunity.

Myxoma virus (MYXV) is a rabbit-specific poxvirus. It causes an innocuous localized cutaneous fibroma in brush rabbits, its nature host. In contrast, the same virus exhibits up to 99.8 % fatality rates in European rabbits, causing the lethal disease myxomatosis. MYXV was deliberately released as a biocontrol agent against European rabbits in Australia and European. The subsequent results of coevolution are both attenuation of MYXV and increased resistance in rabbits. This provided us with one of the best documented examples of host-pathogen evolution and the opportunity to investigate the molecular mechanism underlying virulence changes after encountering with new hosts. In response to virus infection, host protein kinase R (PKR) can sense double-stranded RNA, which is generated during virus replication. Activated PKR phosphorylates the translation initiation factor eIF2a, resulting in the globally inhibition of protein synthesis. PKR has also been reported to involve in the regulation of other signaling pathways, including eIF2a/ATF4 responses and NF-kB activities. Poxviruses encodes two inhibitors, E3 and K3, to inhibit PKR mediated eIF2a phosphorylation and repress the inhibition of protein synthesis, facilitating virus replication. The molecular mechanisms behind PKR associated NF-kB activation and regulation of PKR associated NF-kB transcriptional activities by K3 orthologs are still largely unknown. The main goal of this study is to elucidate molecular mechanisms for the crosstalk between PKR-eIF2a antiviral pathway and NF-kB dependent pro-inflammatory response during MYXV infection. We examined the antiviral function of PKRs from diverse mammalian species during MYXV infection. Activation of PKR induced ATF4 responses and NF-kB transcriptional activities. Of note, rabbit PKRs from nature host and new host of MYXV showed different sensitivity to M156 inhibition. Higher inhibition of PKR inhibition exerted by M156 or other K3 orthologs from different poxviruses resulted in low level of NF-kB activation. In conclusion, our research reveals the interaction between PKR and K3 orthologs affect the outcome of NF-kB dependent pro-inflammatory response, possibly contributing to viral replication and pathogenesis in vivo.

Most virus families produce double-stranded (ds) RNA during their lifecycles, which can be sensed by multiple dsRNA-dependent innate immune sensors, including two potent host restriction factors OAS/RNase L and PKR. PKR mainly restricts replication of viruses that are dependent on eIF2, and RNase L inhibits viral replication via degradation of host and viral RNAs. However, while PKR is directly activated by dsRNA, RNase L activation depends on the OAS to both bind dsRNA and synthesize 2’-5’ oligoadenylate (2-5A). We hypothesized that engineering proteins to combine the dsRNA sensor domain of PKR with the effector domain of RNase L would bypass the need for OAS activation, making them less susceptible to inhibition by viral molecules, and preserve their potent antiviral activity. To test this, we generated Recombinant Enhanced Antiviral Restrictors (REAVRs) by combining dsRNA-binding domains of PKR from different species with the effector domain of human RNase L. We show that REAVRs led to RNA degradation and decreased the activity of a luciferase reporter, suggesting the REAVRs are functionally active. To investigate whether REAVRs could restrict viral replication, we generated T-REx 293 cells containing a single REAVR copy under the control of a doxycycline-inducible promoter. Some REAVRs exerted potent antiviral activities against five tested viruses: vaccinia virus, dengue virus, Zika virus, SARS-CoV-2, and vesicular stomatitis virus. Importantly, these REAVRs were also effective against viruses that are resistant to PKR activation, for example, SARS CoV-2 and flaviviruses. This study provides proof-of-concept that REAVRs are active in vitro and exhibit antiviral effects on various families of viruses. We envision that REAVRs can be used to generate transgenic organisms with increased viral resistance and represent promising candidates as therapeutics for viral infections.

Virus infection leads to activation of the interferon (IFN)-induced endoribonuclease RNase L, which results in degradation of viral and cellular RNAs. Both cellular and viral RNA cleavage products of RNase L bind pattern recognition receptors (PRRs), like retinoic acid-inducible I (RIG-I) and melanoma differentiation-associated protein 5 (MDA5), to further amplify IFN production and antiviral response. Although much is known about the mechanics of ligand binding and PRR activation, how cells coordinate RNA sensing with signaling response and interferon production remains unclear. We show that RNA cleavage products of RNase L activity induce the formation of antiviral stress granules (avSGs) by regulating activation of double-stranded RNA (dsRNA)-dependent protein kinase R (PKR) and recruit the antiviral proteins RIG-I, PKR, OAS, and RNase L to avSGs. Biochemical analysis of purified avSGs showed interaction of a key stress granule protein, G3BP1, with only PKR and RIG-I and not with OAS or RNase L. AvSG assembly during RNase L activation is required for IRF3-mediated IFN production, but not IFN signaling or proinflammatory cytokine induction. Consequently, cells lacking avSG formation or RNase L signaling produced less IFN and showed higher susceptibility during Sendai virus infection, demonstrating the importance of avSGs in RNase L-mediated host defense. We propose a role during viral infection for RNase L-cleaved RNAs in inducing avSGs containing antiviral proteins to provide a platform for efficient interaction of RNA ligands with pattern recognition receptors to enhance IFN production to mount an effective antiviral response.

Eukaryotes have evolved stress response pathways to adapt to a range of environmental assaults. One conserved pathway is the integrated stress response, which allows cells to cope with endoplasmic reticulum stress, heat shock, oxidative stress, and viral infection. A consequence of the integrated stress response is the formation of stress granules, cytoplasmic foci of translationally stalled mRNAs and proteins including ribosomes, translation initiation factors, and RNA-binding proteins. Stress granules have a diverse proteome, and stress granules formed during viral infection are unique in that they are enriched in host antiviral proteins and regulate the immune response. These stress responses are well characterized in mammalian systems, but studies in non-mammalian vertebrates are poorly understood. Viral Hemorrhagic Septicemia Virus (VHSV) is an aquatic rhabdovirus found in over 80 fish species worldwide. Viruses must subvert host immune responses, and previous studies from our lab and others have identified the VHSV matrix (M) and non-virion (NV) proteins as inhibitors of the immune response. Whether there is any interplay between VHSV and aspects of the integrated stress response remain uncharacterized. In order to explore this, we first established that rainbow trout and fathead minnow cells express proteins homologous to mammalian proteins needed to form stress granules. We showed that in response to heat shock and oxidative stress, two stressors known to induce stress granule formation in mammalian cells, both fish cell types formed stress granules. Furthermore, we observed that infection by two different strains of VHSV, VHSV Ia and IVb, induces the phosphorylation of eIF2alpha and stress granule formation, albeit with different kinetics and to different extents, in their respective host species. Phosphorylation of eIF2α and the subsequent formation of stress granules is regulated by the stress-activated kinases HRI, GCN2, PERK and PKR. After establishing VHSV infection induces eIF2alpha phosphorylation and stress granule formation, we sought to determine the kinases responsible for inducing this response. By utilizing pharmacological inhibitors of PKR and PERK, two kinases responsible for activating the integrated stress response, we found that PERK is required for eIF2α phosphorylation during VHSV Ia or IVb infection. With inhibition of either kinase, we observed reduced stress granule formation during VHSV Ia infection. Our results indicate that both PKR and PERK are involved in stress granule induction during VHSV Ia infection, with PERK having a larger impact. Interestingly, inhibition of either of these kinases significantly reduced VHSV Ia and IVb-induced cytopathicity. We have shown in other studies that the VHSV IVb NV protein utilizes the PERK-eIF2α pathway for virus-mediated host translation and IFN signaling shutoff to regulate the host response. Here, we determined that the VHSV IVb NV protein modulates stress granule formation, however, the role of PKR and or PERK in NV-induced SG formation requires further experimentation. These results suggest that VHSV induces the integrated stress response, and benefits from its induction, and the molecular mechanism underlying these processes will be the subject of further studies.

In Arabidopsis thaliana, Caenorhabditis elegans and Drosophila, long dsRNA (ldsRNA) accumulation during virus infection, transposon activation or transgene expression, elicits a potent RNA interference (RNAi) response: ldsRNA is processed into small interfering RNAs by Dicer proteins, before loading into Argonaute (Ago) proteins which assemble into RNA-induced silencing complex (RISC) to slice complementary RNA or methylate DNA. The response can also be amplified and spread systemically throughout the organism. In mammals, specific RNAi can be observed in specific cell lines, but this effect is often mitigated by activation type I interferon (IFN) dsRNA binding protein (dsRBPs) which are responsible for cytotoxicity. In this PhD thesis, we aim to unravel these processes using state-of-the-art biochemical-, genetic- and molecular approaches. Here, we first show, that Dicer interactors protein activator of protein kinase R (PKR) (PACT) and TAR-RNA binding protein (TRBP) rapidly congregate in ldsRNA-rich granules (DRGs) upon ldsRNA stress. Mass spectrometry analysis reveals the recruitment of a compendium of additional RNA silencing- and type I IFN factors to these DRGs, thus providing evidence for an integrated cellular platform for recognition and processing of ldsRNA. Second, we confirm that modulating a key innate immunity receptor (PKR) in 293T cells dampens the ldsRNA-induced cytotoxic effects and uncovers a specific RNAi response on transgene- and endogenous targets in an Ago2- and a Dicer-dependent manner. Interestingly, our in-depth analysis reveals two modes of Dicer processing depending on the exogenous ldsRNA template used, suggesting ldsRNA sequence or structure specifically affects these processes. Furthermore, we also reveal that TRBP is a master regulator of Dicer-processing and Ago-loading during RNAi, most notably of ldsRNA templates which are fully processed by Dicer. Finally, we perform several genome-wide forward genetic screens to dissect the mechanisms leading to the uptake and sensing of exogenous ldsRNA. These experiments allowed us to identify a network of proteins involved in these processes, from cell surface binding to nuclear ldsRNA responses. By extensive validation of knock-outs from the type I IFN screen we managed to (i) molecularly dissect ldsRNA uptake by lipofection and (ii) describe a novel role of signal transducer and activator of transcription 1 (STAT1) in human RNAi. All together the work of this PhD thesis sheds new light on the cellular- and molecular impact of ldsRNA in mammalian cells, an important challenge in the fields of RNA therapeutics and virology.

Hyperosmotic stress is relevant in both physiological and pathological conditions. The response to hyperosmotic stress includes a vast body of research that with each passing year becomes more diverse. Aspects of this multifaceted response include the mechanism (or mechanisms) by which cells sense extracellular hyperosmolarity, the capacity with which various cell types respond, the relationship between hyperosmolarity and inflammation, and mechanisms by which hyperosmolarity contributes to the pathology of disease. We used the response to hyperosmotic stress in cell culture as a tool to study potential mechanisms of disease pathology in tissues affected by hyperosmolarity.In response to an increased concentration of extracellular osmolytes, cells will undergo transcriptional, translational, and signaling responses that either enable adaptation and survival, or initiate apoptosis. This decision will depend on the intensity of stress. The Rel family of transcription factors, including the Tonicity Enhancer Binding Protein (TonEBP) and Nuclear Factor-κB (NF-κB), play critical roles in the switch from osmoadaptive homeostasis to inflammation. Lower intensity hyperosmotic conditions will promote TonEBP and NF-κB family member c-Rel-dependent transcriptional induction of osmoadaptive genes SNAT2 and GADD34. As stress intensity increases, however, the inability of cells to adapt to higher environmental tonicity can lead to inflammatory gene expression and pathogenesis. High extracellular osmolarity results in activation of protein kinase R (PKR) via interaction with the PKR-Activating protein (PACT). Here we identified PACT-mediated PKR activation as a marker of the termination of adaptation and initiation of inflammation. We found that high stress-induced PACT-mediated PKR activation inhibits the interaction between NF-κB c-Rel and TonEBP essential for the increased expression of TonEBP-dependent osmoprotective genes; this resulted in enhanced formation of TonEBP/NF-κB p65 complexes and enhanced transcription of proinflammatory genes, including inducible nitric oxide synthase (iNOS) and IL6. PKR mediated-hyperinduction of iNOS decreased cell survival in mouse embryonic fibroblasts via mechanisms involving nitric oxide (NO) synthesis and post-translational modification of proteins. Moreover, we demonstrate that treatment with the PKR inhibitor C16 dampens both iNOS hyperamplification and phenotypic breakdown of the intestinal epithelial barrier caused by an increase in extracellular osmolarity induced by dextran sodium sulfate (DSS) in vivo.In the following pages, we describe a unique signaling pathway that is activated upon hyperosmotic stimulation that mediates the switch between adaptation and inflammation. Collectively, these findings indicate that PKR functions as a signal transducer in the sensing of extracellular hyperosmolarity. These data demonstrate a novel role of c-Rel in the adaptive response to hyperosmotic stress, which is inhibited via a PACT/PKR-dependent reorganization of Rel family transcription factors. Our results suggest that inhibiting PACT-PKR signaling may prove a novel target for alleviating stress-induced inflammatory diseases.

During viral infection, a major innate host defense mechanism is to reduce global protein synthesis by phosphorylation of eukaryotic translation inhibition factor 2a (eIF2a). eIF2a is phosphorylated by four different cellular kinases, protein kinase R (PKR), general control nonderepressible 2 (GCN2), PKR-like endoplasmic reticulum kinase, and heme-regulated inhibitor, each responding to a different cellular stress. GCN2, which phosphorylates eIF2a in response to amino acid starvation, UV irradiation, and oxidative stress, can also phosphorylate eIF2a in response to viral infection. PKR senses double-stranded RNA produced by virus infection. Many viruses have methods of inhibiting PKR activation or its downstream effects, thus circumventing protein synthesis shutdown. These methods include sequestering double-stranded RNA or producing proteins that bind to and inhibit PKR activation. Here we describe our finding that PKR was antiviral in mouse adenovirus type 1 (MAV 1) infection. We also showed that in multiple cell types, PKR was depleted during MAV 1 infection. Inhibiting the proteasome reduced the PKR depletion seen in MAV-1-infected cells, indicating that proteasomal degradation is responsible for PKR degradation during MAV-1 infection. Time course experiments showed that the degradation occurred early after infection. Infecting cells with UV-inactivated virus prevented PKR degradation, whereas inhibiting viral DNA replication did not. Together these results suggest that an early viral gene is responsible for the degradation. Degradation of PKR is a rare mechanism to oppose PKR activity, and it has only been described in six RNA viruses. To our knowledge, this is the first example of a DNA virus counteracting PKR by degrading it. In addition to PKR being degraded by MAV-1 during infection, we tested whether GCN2 plays an antiviral role, as seen for several other viruses. Here we describe that GCN2-/- deficient (atc) mice and peritoneal macrophages were significantly more susceptible to infection by MAV 1. However, atc mouse embryonic fibroblasts had similar viral yields compared to wild type mouse embryonic fibroblasts, suggesting that there is a cell-type specific antiviral effect of GCN2. There were no differences in viral yields in infected organs of atc mice at 8 days post infection compared to wild type organs, except for the cecum, indicating that the difference in survival between atc and wild type mice is not likely a result of increased viremia in atc mice. There was also no significant difference in histology of organs from MAV-1-infected atc and wild type mice. However, cytokine analysis showed that MAV-1-infected atc mice had significantly higher levels of interleukin 1a, interleukin 1b, and interferon y in the brain compared to infected wild type mouse brains at 7 days post infection, suggesting that a difference in inflammatory response could be responsible for the decreased survival of atc mice in response to MAV-1 infection. Determining how GCN2 affects the immune response to MAV 1 infection could provide insight into how GCN2 is playing an antiviral role in DNA virus infection.

PKR is well characterized for its function in antiviral immunity. Using Toxoplasma gondii, we examined if PKR promotes resistance to disease caused by a non-viral pathogen. PKR-/- mice exhibited higher parasite load and worsened histopathology in the brain and eye. Macrophages/microglia with defective PKR signaling were unable to kill the parasite in response to CD40 stimulation. The TRAF6 but not the TRAF2,3 binding sites of CD40 was required for PKR phosphorylation through CD40 ligation. TRAF6 co-immunoprecipitated with PKR upon CD40 ligation. TRAF6-PKR interaction appeared to be indirect since knock-down of TRAF2 inhibited TRAF6-PKR co-immunoprecipitation. PKR was required for accumulation LC3 around the parasite, vacuole-lysosomal fusion and killing of T. gondii in CD40-activated cells. Thus, we identified PKR as a promoter of protection against disease caused by a non-viral pathogen. PKR is activated by CD40 via TRAF6 and TRAF2 and PKR links CD40-TRAF signaling and stimulation of the autophagy pathway.

The dsRNA binding protein PACT/RAX activates the dsRNA dependent protein kinase (PKR) in response to stress. Several biochemical functions of PACT/RAX are well established, however the physiological role of this protein is unclear. As such the goal of this thesis is to investigate the physiological and developmental functions of PACT/RAX. Two RAX mutant mice with very similar phenotypes resulting from developmental defects were used to investigate the physiological and developmental functions of PACT/RAX. One mouse contained a missense mutation in PACT/RAX which was found to reduce the steady state levels of RAX and disrupt its ability to bind dsRNA. We were unable to distinguish the relative contributions of reduced protein expression and impaired dsRNA binding to the phenotype of this mutant. A second mouse with a targetted disruption in Exon 8 develops an anterior pituitary hypoplasia due to reduced postnatal proliferation. A cell culture model was developed to investigated the role of RAX in pituitary development. Ectopic RAX expression is not sufficient to restore proliferation of anterior pituitary cells lacking RAX; however over the course of these experiments three novel splice variants of RAX were identified which may be required for anterior pituitary proliferation. PACT/RAX is also a component of the RNA silencing (siRNA/miRNA) pathway, although current knowledge of its function in this pathway is limited. Two approaches aimed at investigating the function of PACT/RAX in the RNA silencing pathway were employed. The model organism D. melanogaster was used to determine whether PACT is capable of substituting for the Drosophila dsRNA binding proteins LOQS and/or R2D2, both of which have well defined roles in RNA silencing. PACT is unable to substitute for LOQS or R2D2 at the whole organism level. Expression levels of miRNAs in RAX knockdown pituitary cells were measured to determine whether RAX is required for miRNA biogenesis. Removing RAX from pituitary cells does not affect miRNA expression profile. Results presented here characterize a RAX mutation involved in development, establish a cell culture model to investigate the developmental function of PACT/RAX, identify three novel splice variants of RAX and investigate the function of PACT/RAX in RNA silencing.

Two cellular innate responses, the dsRNA protein kinase (PKR) pathway and the 2'-5' oligoadenylate synthetase (OAS)/RNase L pathway, are activated by dsRNAs produced by viruses and reduce translation of host and viral mRNAs. PKR activation results in eIF2a phosphorylation. As a consequence of eIF2a phosphorylation, stress granules (SGs) are formed by the aggregation of stalled SG proteins with pre-initiation complexes and mRNA. West Nile virus (WNV) infections do not induce eIF2a phosphorylation despite upregulation of PKR mRNA and protein suggesting an active suppression of PKR activation. Assessment of the mechanism of suppression of PKR activation in WNV-infected cells indicated that WNV infections do not induce PKR phosphorylation so that active suppression is not required.             In contrast to infections with "natural" strains of WNV, infections with the chimeric W956 infectious clone (IC) virus efficiently induce SGs in infected cells. After two serial passages, the IC virus generated a mutant (IC-P) that does not induce SGs efficiently but does induce the formation of NS3 granules that persist throughout the infection. This mutant was characterized.             2'-5' oligoadenylate synthetases (OAS) are activated by viral dsRNA to produce 2-5A oligos that activate RNase L to digest viral and cellular RNAs. Resistance to flavivirus-induced disease in mice is conferred by the full-length 2'-5' oligoadenylate synthetase 1b (Oas1b) protein. Oas1b is an inactive synthetase that is able to suppress the in vitro synthetase activity of the active synthetase Oas1a. The ability of Oas1b to inhibit Oas1a synthetase activity in vivo and to form a heteromeric complex with Oas1a was investigated. Oas1b suppressed 2-5A production in vivo. Oas1a and Oas1b overexpressed in mammalian cells co-immunoprecipitated indicating the formation of heteromeric complexes by these proteins.             Unlike mice, humans encode a single OAS1 gene that generates alternatively spliced transcripts encoding different isoforms. Synthetase activity has previously been reported for only three of the isoforms.  The in vitro synthetase activity of additional OAS1 isoforms was analyzed. All tested isoforms synthesized higher order 2-5A oligos. However, p44A only produced 2-5A dimers which inhibit RNase L.

Viral infection induces expression and activation of several genes involved directly or indirectly in antiviral defense. Protein kinase R (PKR), a Ser/Thr kinase induced by interferons and activated by double-stranded RNA (dsRNA) is a crucial component in mediating cellular antiviral action. Following activation by dsRNA, PKR initiates antiviral defense by inhibiting translation and inducing apoptosis. Apart from its role in translational inhibition, PKR has been implicated as a key integrator of many other signaling pathways stimulated by lipopolysaccharide, cytokines, growth factors and different types of cellular stresses. In order to define PKR dependent genes to better elucidate its physiological function, gene expression profiling following dsRNA treatment was performed on wild-type (wt) and PKR knockout (pkr-ko) spleen derived macrophage cell lines (SM) cells). The microarray results identified several novel PKR-dependent genes whose expression is induced by dsRNA treatment of mouse macrophages. The expression of early growth response-1 (Egr-1), an immediate early gene was highly induced in a PKR dependent manner on dsRNA treatment. NF-kB was determined to be a key component in regulation of the dsRNA mediated PKR-dependent, Egr-1 induction.Earlier reports have implied a role for PKR in induction and activation of AP-1 family members including c-Fos and ATF2 in response to different stimuli. A novel role for PKR in up-regulation of the AP-1 transcription factor c-Jun in response to dsRNA is documented in this thesis. Depending on the cell type, dsRNA mediated c-Jun induction is shown to be dependent on c-Jun NH2-terminal kinase JNK or extracellular signal related kinase ERK. Additionally, a role for c-AMP response element binding protein CREB is implicated in dsRNA mediated c-Jun induction. Importantly, activation of CREB, in response to dsRNA was impaired in the absence of PKR in SM cells.A role for PKR in the regulation of different pro-inflammatory cytokines has been established from earlier reports. However a role for PKR in regulation of anti-inflammatory cytokine production is not well documented. This study shows for the first time PKR dependent induction of of IL-10 in response to dsRNA. This dissertation demonstrates that the PKR mediated IL-10 induction in response to dsRNA is dependent on JNK and NF-kB. In addition, by bioinformatic and chromatin immunoprecipitation assay, a novel functional NF-kB binding site was identified at the -861/-851 region of the mouse IL-10 promoter. These results suggest that PKR plays a significant role in innate immunity by maintaining a equilibrium between pro- and anti-inflammatory cytokines in response to viral infection.Overall this study provided us with a better understanding of the global regulation of gene expression by PKR following dsRNA treatment. In addition by identifying the signaling pathways involved in the regulation of three novel dsRNA induced PKR dependent genes, new insights have been provided into the role of this important kinase in regulating innate immunity.

PACT contains three modular domains. While PACT domains 1 and 2 can mediate strong interactions with PKR and dsRNA, the C-terminal domain 3 binds to PKR weakly but is necessary and sufficient for activating PKR in vitro. In vivo, PACT-mediated PKR activation requires either domain 1 or domain 2 for anchoring domain 3 to PKR following cellular stress. The mechanism of PACT-mediated PKR activation was not well understood. The studies on PACT function in this dissertation led to the following findings: 1. We showed that PACT suppressed eIF2α phosphorylation without stress through domains 1 and 2, but enhanced eIF2α phosphorylation with stress through domain 3. We also showed that PACT was phosphorylated by stress, and there might be two phosphorylation sites in PACT, as shown by two dimensional gel analysis. 2. Latent PKR is thought to adopt an inactive confirmation by virtue of an intramolecular inhibitory interaction between the second of the two N-terminal dsRNA-binding motifs (dsRBM2) and the C-terminal kinase domain. Using combined structural and functional approaches, we characterized that this autoinhibitory interaction is mediated by residues 328-336 within the eIF2α kinase insertion region of PKR and a surface on dsRBM2 overlapping with putative dsRNA-binding sites. Residues 328-336 were also found to comprise the region to which PACT domain 3 bound. These results suggested that dsRNA and PACT domain 3 may competitively bind to either of the intramolecular interaction components of PKR to cause PKR activation. This model was supported by our mutational analysis, which showed that mutations within residues 328-335 could activate PKR. 3. We showed that PKR activation by PACT or dsRNA could be blocked by the RNA-binding NS1 protein of influenza A virus both in vitro and in vivo. PKR-binding, but neither dsRNA-binding nor PACT-binding, was essential for this inhibitory effect of NS1. Furthermore, we mapped the PKR-binding region in NS1 and showed that recombinant influenza A viruses encoding PKR-binding deficient NS1 point mutants could not block PKR activation in virus-infected cells. The results presented here provide models to study dsRNA- and PACT-mediated PKR activation pathways.

Viral myocarditis is an important human disease associated with many viruses. Mechanistically, cardiac damage associated with viral myocarditis can be immune mediated and/or the result of direct cytopathic effect. Reovirus induced myocarditis is not immune mediated, and thus, provides an excellent model for the study of direct cytopathic effect in the heart. Previous work has demonstrated that reovirus myocarditic potential reflects induction and/or sensitivity to interferon (IFN). Specifically, nonmyocarditic reoviruses induce greater IFN-β and/or are more sensitive to the antiviral effects IFN than are myocarditic reoviruses. Importantly, IFN mediates its antiviral effects through the induction of interferon-stimulated genes (ISGs); ISGs function as the antiviral effector proteins that block replication in the host cell. The work presented in this dissertation further defines the cardiac response to reovirus infection. In chapter 2, we examined the double-stranded activated protein kinase PKR: the role of PKR in cardiac IFN induction and protection against reovirus-induced myocarditis. We demonstrate that PKR is critical to the robust induction of IFN-β in primary cardiac myocyte cultures (PCMCs). Additionally, we show that nonmyocarditic reoviruses become myocarditic in PKR-null mice, even though reovirus growth in PKR-null hearts is similar to that in wild type mouse hearts. Finally, we demonstrate that relative to wild type mice, reoviruses induce significantly greater morbidity in PKR-null mice. In the following chapter, we compare the IFN response in PCMCs: a non-replenished, critical cell type, to that in primary cardiac fibroblast cultures (PCFCs): a readily replenished cardiac cell type. By quantitative real-time PCR, we find that PCMCs express IFN-β and the ISGs IRF-7 and 561 constitutively at higher levels than PCFCs. We also identify constitutive IFN as a primary means of constitutive ISG expression. Additionally, we demonstrate that PCFCs, more so than PCMCs are more dependent on IFN-mediated protection against reovirus infection. The regulation of IFN-β expression is achieved primarily through the actions of Interferon Regulatory Factors (IRFs). In Appendix 1, we show contributions to a publication demonstrating that IRF-1 is dispensable for IFN-β induction in PCMCs, and yet critical for defense against reovirus-induced myocarditis in the mouse. Finally, a summary of this work is provided.