Mouse Hepatitis virus (MHV) is a member of the coronavirus family in the order Nidovirales. The 32 kb genome contains cis-acting sequences necessary for replication of the viral genome. Those cis-acting sequences have been shown to bind host proteins, and binding of those proteins is necessary for virus replication. One of the cis-acting elements is the 3' terminal 42 nucleotide host protein binding element. Previous work has demonstrated that mitochondrial aconitase, mitochondrial heat shock protein 70, heat shock protein 60 and heat shock protein 40 bind to the 3' terminal 42 nucleotide host protein binding element. We demonstrated that RNA secondary structure of the 3' terminal 42 nucleotide host protein binding element is necessary for host protein binding in vitro. We also demonstrate that primary structure of the 3' terminal 42 nucleotide host protein binding element is necessary for viral replication by targeted recombination. DI replication assays infer that the 3' terminal 42 nucleotide host protein binding element plays a role in positive strand synthesis from the negative strand template. Current studies involve the infectious cDNA clone, which will provide definitive answers on the role of the 3' terminal 42 nucleotide host protein binding element in MHV replication.

The non-structural protein 15 (NSP15) locus in Lineage A β-coronaviruses has two important functions during replication. The encoded endoribonuclease is conserved among coronaviruses. The function of the nsp15 protein is still not fully understood, but recent evidence suggests it may be involved in both replication and inhibiting viral sensing of double stranded RNA. In Lineage A β-coronaviruses, the RNA locus contains an inserted packaging signal (P/S). The P/S is essential for selectively packaging viral genomic RNA. While the P/S is required for selective packaging, it is not required for nsp15 protein function or viral replication. Utilizing this region, I studied the interactions of nsp15 protein during infection. Additionally, I studied the effect of selective packaging on virulence. Coronaviruses encode 16 nonstructural proteins in two open reading frames. These proteins are responsible for forming the replication/ transcription complex (RTC) and creating an environment conducive to viral replication. The RTC is an intricate complex of viral and host proteins with a largely unknown composition. While almost all nsps studied to date localize to sites of replication, the interactions between these proteins are not fully understood. In Chapter II, I describe studies of the interactions and localization of Nsp15 by creating an in situ hemagglutinin epitope tag. I found that mouse hepatitis virus nsp15 could tolerate an in situ tag when placed into the P/S (MHVNsp15-HA). MHVNsp15-HA had wild-type like replication in vitro. Nsp15 was localized to sites of replication throughout infection, with no localization to sites of assembly. Finally, nsp15 interacted with the RNA dependent RNA polymerase and putative primase during infection. These data demonstrate that nsp15 is a member of the RTC. During coronavirus replication two species of viral RNA are present, genomic RNA (gRNA) and sub-genomic RNA (sgRNA). These RNAs are co-terminal on both their 5’ and 3’ ends, containing the leader sequence and 3’ UTR/ polyA respectively. Even with these similarities, coronaviruses are adept at selectively packaging gRNA over sgRNA. This selective packaging is determined by the P/S, a 95 base pair stem-loop structure in the nsp15 locus. This RNA motif is sufficient for packaging of nonviral RNAs and has been shown to interact with the M protein from MHV. Moreover, when this RNA motif is deleted from MHV, (MHVPS-) selective packaging is lost during infection as sgRNAs become a large percentage of packaged viral RNA. In chapter IV I determined the effect of selective packaging on pathogenicity in vivo. Immunocompetent mice infected with MHVPS- had significantly better outcomes compared to MHV wild-type (MHVWT) infected mice. Peak viral loads were decreased in MHVPS- compared to MHVWT. Strikingly I found MHVPS- infected bone marrow derived macrophages had significant increases in type-I interferons (IFNs) and pathogenesis of MHVPS- was restored in mice deficient in IFN signaling. These data indicate that the P/S of MHV is an uncharacterized MHV virulence factor, which acts by preventing an increased IFN response during infection. In MHV, the nsp15 locus is translated into a functional protein and contains functional cis acting RNA elements both of which play a role in MHV replication. This work provides understanding of nsp15 localization and interactions which educate our understanding of the function of this conserved endoribonuclease. Additionally, this work demonstrates a unique function for the P/S not previously described. This work informs future studies of nsp15 protein function and the function of selective packaging during coronavirus infection.

Mice infected with neurotropic variants of the murine coronavirus, mouse hepatitis virus, (strains JHMV or J2.2–V–1) develop acute and chronic CNS infections, and provide a model system to study the pathogenesis of virus–induced neuroinflammation, mechanisms of virus persistence, and anti–viral immune responses in the CNS. Using the J2.2–V–1 model of CNS infection, we addressed the role of sustained CCL2 production during viral infection using mice in which CCL2 was expressed transgenically in oligodendrocytes. Tonic CCL2 expression in the CNS resulted in delayed kinetics of virus clearance, and converted what is typically a mild, nonlethal disease to acutely lethal encephalitis, with the majority of mice succumbing to the infection. CCL2 induced a rapid and dysregulated inflammatory response that was no longer protective and was unable to efficiently clear virus from the CNS. Infected CCL2 Tg mice had increased numbers of Foxp3–expressing CD4 T cells (Tregs) and of macrophages and microglia expressing elevated levels of YM–1, a marker for alternatively activated macrophages, and nitric oxide. Our results showed that CCL2 has effects beyond serving as a chemoattractant for leukocytes, and has effects on the composition and function of inflammatory cells at sites of infection. In a separate set of experiments, I identified and characterized two additional heteroclitic variants of the JHMV epitope S598 that induced CD8 T cells with greater antigen sensitivity to the native S598 determinant relative to the cells primed by the native epitope. One of these heteroclitic epitopes elicited a T cell response with nearly complete cross–reactivity towards the native peptide. The structural data show that these heteroclitic epitopes induced modest conformational changes in the local environment of the peptide–MHCI complex. I also provide data to support the notion that heteroclitic determinants augment functional avidity by increasing surface epitope density. Collectively, these data will help guide the design of heteroclitic epitopes in the setting of vaccine development. Lastly, I examined the consequences of oxidative stress induced by viral infection on antigen presentation. The brains of JHMV–infected mice were found to have signs of oxidative stress, with significantly decreased ratios of reduced (GSH) to oxidized (GSSG) glutathione, suggesting that there is an environment that is conducive for cysteine modification with oxidized glutathione. We found that virus–induced oxidative stress resulted in the presentation of both native and S–glutathionylated forms of the JHMV epitope S510 by infected cells. A subset of the S510–specific CD8 T cells failed to recognize the modified form of the epitope, suggesting that GSH–modification of a cysteine–containing viral epitope might interfere with T cell recognition. Further, GSH-modified peptides were identified in stressed human cells, including herpes virus–transformed B cells, suggesting that the modification is not limited to mouse cells. Collectively these findings have implications for both anti–viral immunity and anti–tumor immunity, where oxidative stress has been shown to play a role during infection and tumorgenesis.

RNA sequences and higher-order structures in the 5’ and 3’ untranslated regions (UTRs) of positive-strand RNA viruses are known to function as cis-acting elements for translation, replication, and transcription. In coronaviruses, these are best characterized in the group 2a bovine coronavirus (BCoV) and mouse hepatitis virus (MHV), yet their precise mechanistic features are largely undefined. Here, we use a reverse genetics system in MHV to exploit the ~30% nt sequence divergence between BCoV and MHV to establish structure/function relationships of 5’ UTR cis-replication elements. It had been previously shown that a precise replacement of the 391-nt MHV 3’ UTR with the 288-nt BCoV 3’ UTR yields wt-like MHV. Our attempts to replace the 209-nt MHV 5’ UTR with the 210-nt BCoV 5’ UTR, however, yielded a non-viable chimera. Therefore, a systematic analysis of individual 5’-terminal structures was made to identify compatible elements. By placing each of four putative cis-acting domains from the BCoV 5’ UTR into the MHV genome, we learned that (i) stem-loops (SLs) I & II and SLIII are functionally compatible, (ii) SLIV is compatible if it spans parts of the 5’ UTR and the nonstructural protein 1 (nsp1) cistron, thus identifying this part of ORF 1 as a component of the cis-replication signal, (iii) a relatively unstructured 32-nt region mapping between SLIII and SLIV defines a novel virus species-specific cis-replication element, (iv) spontaneous suppressor mutations within MHV SLI and nsp1 cistron compensated for growth defects arising from the BCoV 32-nt element in the MHV genome, (v) cross talk between the 32-nt element, SLI, and the nsp1 cistron appears essential for virus replication, (vi) the BCoV 5’ UTR and nsp1 cistron function together in the MHV genome to generate a wt-like MHV phenotype, and (vii) a functional 5’ UTR-nsp1 domain in group 2a coronaviruses cannot be substituted by the corresponding genomic element from the group 2b SARS-CoV. We postulate that the interaction between the 5’ UTR and nsp1 cistron (or possibly nsp1 protein) functions as a molecular switch between genome translation and ignition of negative-strand RNA synthesis.