Hepatitis E virus (HEV), the causative agent of hepatitis E, is an important public health concern in many developing countries. It mainly infects young adults and has a mortality of up to 25% in pregnant women. Although hepatitis E is only sporadic in industrialized countries including the United States, a relative high seroprevalence rate has been reported in healthy individuals. Evidence suggests that there exist animal reservoirs for HEV and HEV transmission is zoonotic. Animal strains of HEV, swine HEV and avian HEV have been identified from a pig and a chicken, respectively, in the United States. Studies showed that swine HEV and avian HEV are genetically and antigenically related to human HEV, and that pigs and chickens are useful animal models to study HEV replication, pathogenesis and cross-species infection. The objectives of this dissertation were to genetically characterize both avian HEV and swine HEV, to determine their serological and molecular epidemiology in the United States, to assess the ability of avian HEV cross-species infection in non-human primates, to determine the full-length genomic sequence and genome organization, and to construct an infectious cDNA clone of avian HEV. The prevalence of swine HEV infections in US swine herds and the heterogeneity of swine HEV isolates from different geographic regions of the United States were determined. We found that 35% pigs and 54% swine herds were positive for swine HEV RNA. Partial capsid gene region of twenty-seven US swine HEV isolates was sequenced and was showed to share 88%-100% nucleotide sequence identity to each other and 89-98% identity with the prototype US swine HEV, but only <79% identity with Taiwanese swine HEV isolates and most known human strains of HEV worldwide. All US swine HEV isolates belong to the same genotype 3 with the prototype US swine HEV and the two US strains of human HEV. Similarly, the prevalence of avian HEV infections in US chicken flocks and the heterogeneity of avian HEV isolates were also determined. Helicase gene region of eleven field isolates of avian HEV from chickens with hepatitis-splenomegaly (HS) syndrome was sequenced and was found to share 78-100% nucleotide sequence identities with each other, 79-88% identities with the prototype avian HEV, 76-80% identities with Australian chicken big liver and spleen disease virus (BLSV), and 56-61% identities with other known strains of mammalian HEV. A relative high prevalence of anti-avian HEV antibodies was found in apparently healthy chicken flocks in 5 states. Like swine HEV, the seropositivity of avian HEV in adult chickens was higher than that in young chickens. To genetically characterize the avian HEV genome, we determined the full-length genomic sequence of avian HEV, which is 6,654 bp in length excluding the poly (A) tail, and 600 bp shorter than that of mammalian HEVs. Avian HEV has similar genomic organization with human and swine HEVs, but shared only about 50% nucleotide sequence identity with mammalian HEVs in the complete genome. Significant genetic variations such as deletions and insertions, particularly in the ORF1 of avian HEV, were observed, but motifs in the putative functional domains of the ORF1 were relatively conserved between avian HEV and mammalian HEVs. Phylogenetic analyses based on the full-length genomic sequence revealed that avian HEV represents a branch distinct from human and swine HEVs. Since swine HEV infects non-human primates and possibly humans, the ability of avian HEV cross-species infection in non-human primates was also assessed. However, unlike swine HEV, avian HEV failed to infect two rhesus monkeys under experimental conditions. With the availability of the complete genome sequence of avian HEV, we constructed three full-length cDNA clones of avian HEV and tested their infectivity by in vitro transfection of the LMH chicken liver cells and by in vivo intrahepatic inoculation of specific-pathogen-free (SPF) chickens. The results showed that all 3 cDNA clones of avian HEV were infectious both in vitro and in vivo, as the capped RNA transcripts from each of the clones were replication-competent in transfected LMH cells and developed active infection in inoculated SPF chickens. In summary, avian HEV and swine HEV infections are enzootic in chicken flocks and in swine herds in the United States, respectively. Like human HEV, swine HEV and avian HEV isolates from different geographic regions are also genetically heterogenic. Complete genomic sequence analyses showed that avian HEV is related to, but distinct from, human and swine HEVs. Unlike swine HEV, avian HEV is probably not transmissible to non-human primates. Infectious cDNA clones of avian HEV have been successfully constructed. The availability of the infectious clones for a chicken strain of HEV now affords us an opportunity to study the mechanisms of HEV replication, pathogenesis and cross-species infection. / Ph. D.

As novel or variant strains of HEV continue to evolve rapidly both in humans and other animals, it is important to develop a rapid pre-sequencing screening method to select field isolates for further molecular characterization. Two heteroduplex mobility assays (HMA) were developed to genetically differentiate field strains of swine HEV and avian HEV from known reference strains. It was shown that the HMA profiles generally correlate well with nucleotide sequence identities and with phylogenetic clustering between field strains and the reference swine HEV or avian HEV strains. Therefore, by using different HEV isolates as references, the HMA developed in this study can be used as a pre-sequencing screening tool to identify variant HEV isolates for further molecular epidemiological studies. Our previous study showed that avian HEV antibody is prevalent in apparently healthy chickens. A prospective study was conducted on a known seropositive but healthy chicken farm. Avian HEV was identified from the healthy chicken flock. Avian HEV isolates recovered from the healthy chicken share 70-97% nucleotide sequence identities with those isolates which cause hepatitis-splenomegaly (HS) syndrome based on partial helicase and capsid gene regions. Recovery of identical viruses from the experimentally inoculated chickens in the subsequent transmission study further confirmed our field results. The capsid gene of avian HEV isolates from chickens with HS syndrome were also characterized and found to be heterogeneic, with 76-100% nucleotide sequence identities to each other. The study indicates that avian HEV is enzootic in chicken flocks and spread subclinically among chicken populations, and that the virus is heterogeneic. As HEV can not be propagated <i>in vitro</i>, in order to further characterize avian HEV, an infectious viral stock with a known infectious titer must be generated. Bile and feces collected from specific-pathogen-free (SPF) chickens experimentally infected with avian HEV were used to prepare an avian HEV infectious stock. The infectivity titer of this infectious stock was determined, by intravenously inoculating one-week old SPF chickens, to be 5 x 10^4.5 50% chicken infectious doses (CID₅₀) per ml. Seroconversion, viremia as well as fecal virus shedding were observed in the inoculated chickens. Contact control chickens also became infected via direct contact with inoculated ones. Avian HEV infection in chickens was found to be dose-dependent. To determine if avian HEV can infect across species, one-week old SPF turkeys were intravenously inoculated each with 10^4.5(CID₅₀) of avian HEV. The inoculated turkeys seroconverted to avian HEV antibodies at 4-8 weeks postinoculation (WPI). Viremia was detected at 2-6 WPI, and fecal virus shedding at 4-7 WPI in inoculated turkeys. This is the first demonstration of cross-species infection by avian HEV. Little is known regarding the characteristics of the small ORF3 protein largely due to the lack of a cell culture system for HEV. To characterize the small protein, the ORF3 proteins of avian HEV and swine HEV were expressed in <i>Escherchia coli</i>, and purified by BugBuster His-Bind Purification System. Western blot analysis showed that avian HEV ORF3 protein is unique and does not share common antigenic epitopes with those of swine HEV and human HEV. However, swine HEV (genotype 3) and human HEV (genotype 1) ORF3 proteins cross-react with each other antigenically. To determine if the ORF3 protein is a virion protein, infectious stocks of avian HEV and swine HEV were first generated in SPF chickens and pigs, respectively. Virions were subsequently purified by sucrose density gradient centrifugation and virion proteins were characterized by SDS-PAGE and Western blot analysis. Two major forms of ORF2 proteins of avian HEV were identified: a 56 kDa and an 80 kDa proteins. Multiple immunoreactive forms of ORF2 proteins of swine HEV were also observed: 40 kDa, 53 kDa, 56 kDa and 72 kDa. However, the ORF3 protein was not detected from the native virions of avian HEV or swine HEV. These findings provide direct evidence that ORF2 indeed encodes a structural protein of HEV, whereas ORF3 does not. To search for other potential animal reservoirs for HEV, the prevalence of IgG anti-HEV antibody was determined in field mice caught in chicken farms to assess the possibility of mice as a potential reservoir for HEV infection in chickens. Three different recombinant HEV antigens derived from avian HEV, swine HEV, and human HEV were used in the ELISA assays. The anti-HEV seropositive rates in wild field mice (<i>Mus musculus</i>), depending upon the antigen used, are 15/76 (20%), 39/74 (53%), and 43/74 (58%), respectively. HEV RNA was also detected from 29 fecal and/or serum samples of mice. The HEV sequences recovered from field mice shared 72-100% nucleotide sequence identities with each other, 73-99% sequence identities with avian HEV isolates, and 51-60% sequence identities with representative strains of swine and human HEVs. However, attempts to experimentally infect laboratory mice (Mus musculus) with the PCR-positive fecal materials recovered from the wild field mice were unsuccessful. We also attempted to experimentally infect 10 Wistar rats each with avian HEV, swine HEV, and an US-2 strain of human HEV, respectively. However, the inoculated rats did not become infected as evidenced by the lack of viremia, virus shedding in feces or seroconversion. These data suggest that mice caught in chicken farms are infected by a HEV-like virus, but additional work is needed to determine the origin of the mouse virus as well as the potential role of rodents in HEV transmission. In summary, we developed two HMAs which are useful for differentiation and identification of variant strains of swine and avian HEVs. We genetically identified and characterized an avian HEV strain from apparently healthy chickens in seropositive flocks. We showed that avian HEV can cross species barriers and infect turkeys. Our data indicated that avian and swine HEV ORF2 genes encode structural proteins, whereas ORF3 genes do not. Evidence in this study also showed that HEV or HEV-like agent exists in field mice on a chicken farm. / Ph. D.

Hepatitis E virus (HEV) is a major cause of enterically transmitted acute viral hepatitis in developing countries that lack proper hygienic infrastructure. Hepatitis E is globally distributed and has emerged as an important public health disease in both developing and industrialized countries. HEV is a non-enveloped virus carrying a single-stranded positive-sense RNA genome of approximately 7.200 bp in length. The life cycle of HEV is poorly understood due to the lack of an efficient cell culture system. Animal model systems, including non-human primates, swine, and chickens are being used to study some fundamental aspects of the HEV biology. Recently, novel animal strains of rat and rabbit HEV have been discovered, and whose usage as animal model systems needs to be established. HEV infections in pigs and chickens provide excellent model systems to study the replication and pathogenesis aspects of HEV. Recently, we identified a hypervariable region (HVR) in the open reading frame 1 (ORF1) of HEV. The objectives of this dissertation were to utilize chicken and swine model systems to study the role of HVR in HEV infection in vivo, to determine the effects of HVR on replication of HEV in vitro, and to analyze the effect of exchange of HVR among different genotypes on the replication-competency and virion production in vitro. Extensive sequence variability in the HVR among HEV strains of different genotypes prompted us to study the dispensability of this region. Initially we constructed two partial deletion mutants of genotype 1 human HEV, hHVRd1 and hHVRd2, with in-frame deletion of amino acids (aa) 711 to 777 and 747 to 761 in the HVR of a sub-genomic GFP HEV replicon. Expression of enhanced green fluorescent protein by the mutant hHVRd2 confirmed the dispensability of amino acid residues 747-761 of the HVR. To confirm our in vitro results, specific-pathogen-free (SPF) chickens were intra-hepatically inoculated with capped RNA transcripts from three avian HEV HVR-deletion mutants: mutants aHVRd1 (Δ557-585), aHVRd2 (Δ612-641), and aHVRd3 (Δ557-641). Chickens intra-hepatically inoculated with the mutants, aHVRd1 and aHVRd2, developed active viral infection as evidenced by seroconversion, viremia, and fecal virus shedding. Mutant aHVRd3, with a larger HVR deletion, was apparently attenuated in chickens. Additionally, we used the swine model system to further verify our results from the chicken study. The infectivity of four genotype 3 swine HEV HVR-deletion mutants, sHVRd1 (Δ712-790), sHVRd2 (Δ722-781), sHVRd3 (Δ735-765), and sHVRd4 (Δ712-765) constructed using the genotype 3 swine HEV as the backbone was determined in SPF pigs. Pigs intra-hepatically inoculated with capped RNA transcripts from the mutants sHVRd2, sHVRd3, and sHVRd4 developed active viral infection, whereas mutant sHVRd1 (Δ712-790), with a nearly complete HVR deletion, exhibited an attenuation phenotype. The data from these studies indicate that deletions in HVR do not abolish HEV infectivity in vitro or in vivo, although evidence for attenuation was observed for HEV mutants with a larger or nearly complete HVR deletion. To further elucidate the role of HVR in HEV replication, we investigated the effects of serial amino acid deletions in HVR on the replication of HEV. We first constructed a genotype 1 human HEV luciferase replicon by replacing the ORF2 gene that encodes for the capsid protein with the fire fly luciferase reporter gene. Using the backbone of human HEV genotype 1 luciferase replicon, we constructed a series of HVR-deletion mutants with deletions of variable lengths in the HVR. Amino acid deletions Δ711-725, 711-740 and Δ711-750 were engineered at the N-terminus, deletions Δ729-754, Δ721-766, and Δ716-771 were engineered in the central region, and deletions Δ761-775, Δ746-775, and Δ736-775 were engineered at C-terminus of the HVR. The effects of these serial deletions on HEV RNA replication in the human liver carcinoma cell line, Huh7, were examined. Replication levels of mutants carrying these deletions were compared with that of the wild-type HEV in Huh7 cells. We observed that deletions in the HVR did not abolish viral RNA synthesis but substantially reduced the replication levels of viral RNA, as measured by the reporter luciferase activity. To further verify the effects of HVR deletions on viral RNA replication as observed with the genotype 1 human HEV replicon, we subsequently used a genetically-distinct strain of HEV, avian HEV, and constructed an avian HEV sub-genomic luciferase replicon by substituting the ORF2 gene of avian HEV with the fire fly luciferase gene. Avian HEV HVR-deletion mutants Δ557-603, Δ566-595, and Δ573-587 were then engineered using the backbone of avian HEV luciferase replicon. The replication efficiency of the three deletion mutants of avian HEV in chicken liver hepatoma cell line, LMH, was evaluated. Compared with the wild-type avian HEV, the viral RNA synthesis of the avian HEV HVR-deletion mutants was considerably reduced by the HVR deletions. To analyze the impact of the complete HVR deletion on avian HEV infectivity, we constructed an avian HEV mutant with a deletion of the entire HVR region (aaΔ557-603) using the avian HEV infectious cDNA clone as the backbone. After confirming the viability of the complete HVR-deletion mutant in LMH cells, SPF chickens were intrahepatically inoculated with capped RNA transcripts generated from the mutant. None of the chickens inoculated with the complete HVR-deletion mutant showed evidence of HEV infection, indicating that drastic reduction in replication levels due to complete HVR deletion has resulted in the loss of virus infectivity. The results indicated that HVR may have critical residues that may interact with viral/and or host factors and modulate the replication efficiency of HEV. In the final part of the dissertation research, we sought to determine if the variable sequences of HVR are genotype-specific for in vitro virus replication. By using the genotype 1 human HEV as the backbone, we swapped the HVR of genotype 1 human HEV with the HVRs of the genotype 3 swine HEV and the distantly-related avian HEV to construct two inter-genotypic chimeras, pSKHEV2-Sw and pSKHEV2-Av. Similarly, by using the genotype 3 swine HEV as the backbone, the HVR of genotype 3 swine HEV was swapped with the HVR of genotype 1 human HEV to construct the chimera, pSHEV3-Hu. The viability of these chimeras was tested in Huh7 cells that are permissive for HEV replication. Immunofluorescence assay (IFA) with anti-HEV antibodies revealed that all the three chimeras were replication-competent in Huh7 cells. The infectivity of these chimeras was subsequently evaluated in HepG2 cells. The results showed that exchange of the HVR between different genotypes of mammalian HEVs does not abolish the replication competency and infectivity of HEV. This finding suggests that HVR is not genotype-specific with respect to viral replication and infectivity. The absence of detectable viral antigen in HepG2 cells infected with chimera pSKHEV2-Av suggested a functional incompatibility of the HVR of avian HEV in the mammalian HEV genome. In summary, we identified a highly variable sequence, HVR, in the ORF1 of the HEV genome, and the sequences of the HVR vary significantly among HEV strains of different genotypes. We found that the HVR contain sequences that are dispensable for virus infectivity both in vitro and in vivo. Deletion analysis of HVR revealed that the region may play a role in modulating the replication efficiency of HEV RNA by interacting with viral and/or host factors. Finally, we demonstrated that HVR is not genotype-specific for virus replication and the region can be functionally replaced between mammalian HEV genotypes for virus replication and virion production in vitro. The results from this dissertation research have important implications for better understanding the biology and mechanism of HEV replication and may aid in our efforts to eventually develop a modified live-attenuated vaccine against HEV. / Ph. D.