Molecular biology and replication of hepatitis E virus

Non-structural proteins (NSPs)

The HEV NSPs are encoded by ORF1. This coding region begins immediately after the 5′ NCR and extends 5082 bp (for the genotype 1 HEV Sar-55 strain). ORF1 encodes a 1693-amino acid (aa) polypeptide that is involved in viral replication and protein processing.⁴⁰ ORF1 contains several putative functional domains^1,42 (Figure 2), including a methyltransferase (Met) domain for capping the 5′ end of the viral genomic RNA,⁴³ a ‘Y' domain of unknown function, a papain-like cysteine protease (PCP) domain, a proline-rich region that contains a hypervariable region (HVR), an ‘X' domain of unknown function that is next to the PCP domain, a helicase (Hel) domain, and an RNA-dependent RNA polymerase (RdRp) domain that is responsible for viral replication.⁴⁴

The Met domain is the first functional domain at the 5′ end of ORF1.⁴³ Both guaine-7-Met and guanyltransferase activities were detected in a baculovirus-expressed 110-kDa ORF1 polyprotein. Capping of the HEV genomic RNA was verified using a 5′ RNA ligase-mediated rapid amplification of cDNA ends method that can selectively amplify capped RNAs.⁴⁵ The cap is critical for viral infectivity, as only capped genomic RNA is infectious in non-human primate models.⁴⁰ A purified recombinant HEV Hel protein was shown to possess γ-phosphatase activity that may catalyze the first step in the formation of the cap,⁴⁶ suggesting that the Hel domain of the HEV protein is active and may be involved in the synthesis of the 5′ capped HEV RNA.

When expressed in Escherichia coli (E. Coli), the Hel domain of HEV possesses both nucleoside triphosphatase and 5′–3′ RNA duplex unwinding activities.⁴⁷ Mutations in the first walker A motif, including substitution of the K residue of the nucleotide-binding motif with an A residue, can lead to a 70% loss in ATPase activity. Similarly, mutating the DE residues in the Mg²⁺-binding motif present in the second walker B motif to aa residues caused a 50% reduction in ATPase activity. Mutations in both motifs abolished the Hel activity of the protein.

A putative PCP domain has been predicted in HEV ORF1;⁴² however, definitive evidence for the translational and post-translational processing of the ORF1 polyprotein is still lacking. HEV ORF1 has been expressed as a 185-kDa polyprotein in E. coli, insect cells and mammalian cells.^48,49,50 When ORF1 was expressed in mammalian cells using a vaccinia virus-based expression system, two potential post-translational processing products of 107 and 78 kDa were observed only after an extended incubation time.⁴⁹ However, disruption of the predicted protease catalytic site in the ORF1 PCP domain did not affect the processing of the protein products. This suggests that the observed cleavage products are not generated by a viral protease and that they are likely produced by the expression system that was used in the study. Additionally, when HEV ORF1 was expressed in a baculovirus expression system, multiple smaller proteins were seen, and the presence of the putative processing products was abolished after the addition of the cell-permeable cysteine protease inhibitor E-64d.⁵¹ Therefore, conclusive evidence detailing the processing of the ORF1 protein in the baculovirus system is still lacking.

More recently, the deconjugation activity of a purified recombinant protein that contained the Met and the papain-like cysteine protease domains (MetT-PCP) from ORF1 was analyzed using the following fluorogenic substrates: ubiquitin-7-amino-4-methylcoumarin (Ub-AMC), interferon-stimulated gene 15 (ISG15)-AMC, neural precursor cells expressed developmentally downregulated 8 (Nedd8)-AMC and small ubiquitin-like modifier (SUMO)-AMC. The MetT-PCP recombinant protein hydrolyzed all four substrates, and the hydrolysis of ISG15-AMC was significantly greater compared with the other substrates, suggesting that the HEV MetT-PCP recombinant protein possesses deubiquitinating activity.⁵² However, processing of the Hel and the RdRp domains that contained the conserved LXGG site (position 1205) was not observed. Thus, it remains unclear whether the ORF1 polyprotein functions as a single protein with multiple functional domains or if the cleaved smaller proteins function independently.

The viral RdRp contains eight conserved motifs (motif I–VIII) that are similar to the RdRps from other positive-sense RNA viruses.⁵³ The GDD sequence is essential for HEV replicase activity, as demonstrated by mutation of an enhanced green fluorescent protein (EGFP)-containing HEV replicon.⁵⁴ The HEV RdRp can bind to the 3′ NCR of the HEV genome, and when expressed in E. coli, recombinant HEV RdRp can synthesize RNA in vitro using the 3′ polyadenylated HEV RNA as a template in a primer-independent manner.⁴⁴ Furthermore, the RdRp is localized to the endoplasmic reticulum (ER) of the cells that express RdRp recombinant protein,⁵⁵ suggesting that the ER may be involved in HEV replication.

An HVR was identified in ORF1of HEV. The HVR overlaps the proline-rich sequence that is located between the N terminus of the X domain and the C-terminal portion of the putative PCP domain.⁴² The HVR varies both in length and sequence among different HEV strains.⁵⁶ It has been demonstrated that HEV can tolerate small deletions in the HVR and that the aa residues in this region are dispensable for viral infectivity.⁵⁶ However, the replication of HVR deletion mutants was markedly reduced in Huh7 cells, and a similar reduction in HEV replication was observed when HVR deletion mutants from avian HEV were tested in LMH chicken liver cells.⁵⁷ The HVR sequences are functionally interchangeable between different HEV genotypes with respect to viral replication and infectivity in vitro; however, genotype-specific HVR differences in replication efficiency have also been observed.⁵⁷ This suggests that while small deletions in the HVR do not have a significant impact on viral infectivity, the HVR may influence the efficiency of HEV replication by interacting with viral and/or host factors. Interestingly, a rare, recombinant HEV that contained a 174-base insertion from a human ribosomal protein gene was isolated when the Kernow-C1 strain of HEV, which was originally isolated from a human immunodeficiency virus-infected individual, was adapted to grow in human hepatocellular carcinoma cells.¹⁵ The insertion in the HVR region was stable during serial passages in cell culture, and using an infectious cDNA clone, this insertion was found to be a key factor in the adaptation of the virus to growth in cell culture. Both the aa composition and the size of the insertion are important in viral replication.⁵⁸

Recently, it has been demonstrated that chimeric viruses that contain the ORF2 capsid gene either alone or in combination with the adjacent 5′ JR and 3′ NCR from a genotype 4 human HEV in the backbone of a genotype 3 swine HEV are replication competent in Huh7 cells, infectious in HepG2/C3A cells, and infectious in pigs. However, chimeric viruses that contain the JR, ORF2 and 3′ NCR from zoonotic genotypes 3 or 4 HEV that infect humans and pigs in the backbone of genotype 1 human HEV failed to infect pigs. This suggests that the 5′ NCR and ORF1 may be involved in cross-species HEV infection or may be related to host tropism.⁵⁹

The capsid protein

The HEV capsid protein is encoded by ORF2 and is 660 aa in length with a molecular weight of approximately 72 kDa. The capsid protein has a typical arginine-rich signal peptide sequence and three potential N-linked glycosylation sites.^1,9 When expressed in vitro, capsid proteins of various sizes have been observed.^60,61 In cells that are infected with recombinant Semliki Forest virus that contains ORF2, two larger ORF2-specific proteins were identified that were likely due to N-linked glycosylation.⁶² However, when the HEV ORF2 protein was expressed in a baculovirus system, four capsid protein products of 72, 63, 56 and 53 kDa were observed. Sequence analyses revealed that the 53-, 56- and 63-kDa proteins were cleaved products from the full-length 72-kDa protein.⁶³

The ORF2 capsid protein is involved in the assembly of the HEV particle and its interaction with host cells. The capsid protein contains a potential ER localization signal at its N terminus, and while the protein does enters the ER, a fraction of the protein retrotranslocates to the cytoplasm.⁶⁴ The majority of the capsid protein expressed in mammalian cells is glycosylated, and the glycosylation is important for the formation of infectious virus particles.²⁰ The capsid also binds to HEV genomic RNA⁶⁵ and may play a role in viral assembly. The capsid interacts with host cells and binds to heat shock protein 90 (HSP90),⁶⁶ glucose-regulated protein 78 (Grp78)⁶⁷ and heparin sulfate proteoglycans (HSPGs).⁶⁸ It has been demonstrated that HSPGs may serve as an attachment receptor to facilitate HEV entry into the host cells, whereas Grp78 or HSP90 may be involved in intracellular transport. Recently, it has been shown that three aa mutations (F51L, T59A and S390L) in the HEV ORF2 capsid protein²¹ result in viral attenuation. These three mutations significantly reduced viremia, delayed the onset time of viremia, shortened the duration of fecal virus shedding and viremia, and reduced the viral loads in the liver, bile and intestinal contents.⁶⁹

The HEV ORF2 capsid protein is immunogenic, and neutralizing antibodies have been shown to target it. Both linear and conformational epitopes for neutralizing antibodies have been mapped to aa residues 578–607,⁷⁰ 452–617,⁷¹ and 458–607⁷² in the C-terminal portion of ORF2. Recent studies demonstrated that only antibodies recognizing conformational epitopes are neutralizing, and the aa residues Leu477 and Leu613 in the capsid protein are important in forming a neutralization-sensitive epitope.⁷³ Recently, the epitope of the 8C11 neutralizing antibody was identified using the complex structure of the HEV E2s domain and the 8C11 antibody. The neutralizing epitope was found to be conformational and located at aa residues 496–499, 510–514 and 573–578.⁷⁴ Western blot analysis showed that the recombinant avian HEV capsid protein reacted with both antisera against genotype 1 human HEV and convalescent antisera against genotype 3 swine and human HEVs. Similarly, convalescent sera from chickens that were experimentally infected with avian HEV reacted with the recombinant capsid proteins from the genotype 3 swine HEV and genotype 1 human HEV.⁷⁵ The convalescent serum from animals that were infected with any of the four mammalian genotypes of HEV all neutralized genotype 1 HEV.⁷⁶ Despite the broad antigenic cross-reactivity between the known HEV genotypes, antigenic variations have been observed among HEV strains using genotype- and strain-specific monoclonal antibodies.⁷⁷ Therefore, both pan-genotypic and strain-specific epitopes exist. It has been reported that aa residue 606 in the capsid is important in maintaining the antigenicity of the HEV ORF2 protein.⁷⁷ The E2 domain of the capsid protein, including aa residues 455–602, forms a homodimer. A strong hydrophobic cluster at the dimer interface is maintained by the side chains of residues Val503, Trp548, Thr552, Ala555, Tyr557, Tyr561, Val598 and Val600 in both subunits of the dimer. The C-terminal residues Ala597, Val598, Ala599, Leu601 and Ala602 are also involved in the formation of the dimmer.²⁶

The cytoskeleton-associated multifunctional phosphoprotein

HEV ORF3 encodes a small, 114-aa, cytoskeleton-associated phosphoprotein that is translated from the bicistronic SG RNA. At its 3′ end, ORF3 overlaps ORF2 by approximately 300 nt, but it does not overlap ORF1.^19,39 The protein encoded by ORF3 contains two N-terminal hydrophobic domains (D1 and D2) and two C-terminal proline-rich regions (P1 and P2).¹⁷ The phosphorylated ORF3 protein interacts with the non-glycosylated capsid protein via a 25-aa region in the ORF3 protein. This ORF2–ORF3 interaction depends on the phosphorylation of the ORF3 protein at the Ser71 residue and may play a regulatory role in the assembly of HEV virions.⁷⁸ The C-terminal region of the ORF3 protein is multifunctional and appears to be involved in virion morphogenesis and pathogenesis.^78,79,80

When expressed in mammalian cells, the ORF3 protein is approximately 13 kDa in size, it is phosphorylated at the Ser71 residue, and it is associated with the cytoskeletal and membrane fractions.⁸¹ Similarly, ORF3 proteins were also detected in cells that were transfected with the HEV genomic RNA.^18,41,82 The expression of the ORF3 protein displays both filamentous and punctate distribution patterns. The filamentous distribution reflects the interaction between the ORF3 protein and microtubules,⁸³ whereas the punctate distribution is associated with both early and recycling endosomes.⁸⁴ Using an in vitro replicon system, it was shown that the ORF3 protein is not required for HEV replication in Huh7 cells and other cell lines,¹⁸ but the ORF3 protein is essential for both infection of rhesus macaques⁴¹ and the release of virions from infected cells.^16,85 Monoclonal antibodies against the ORF3 protein can capture HEV particles from cell culture supernatants and serum from HEV-infected patients but cannot capture fecal virus. This suggests that the ORF3 protein is present on the surface of nascent HEV virions.⁸⁶

It has been shown that the ORF3 protein colocalizes with the cytoskeleton via interactions through the ORF3 D1 hydrophobic domain,⁸¹ and it binds to mitogen-activated protein kinase (MAPK) phosphatase⁸⁷ and microtubules.⁸³ The D2 hydrophobic domain interacts with hemopexin and affects cellular iron homeostasis.⁸⁸ The P2 region contains a proline-rich PxxP motif and binds to sarcoma(src)-homology 3 (SH3) domain-containing proteins,⁸⁹ which have important functions in signal transduction pathways that promote cell survival.⁹⁰ It has been reported that the ORF3 protein activates extracellular signal-regulated kinases (Erks)⁸⁹ via the binding and inactivation of an Erk-specific MAPK phosphatase, Pyst1. The ORF3 protein binds Pyst1 through its D1 domain and blocks conformational changes in Pyst1 that are required for Erk-mediated activation.⁸⁷ It has also been reported that the ORF3 protein upregulates the expression of enzymes in the glycolytic pathway via the stabilization of hypoxia-inducible factor 1 (HIF-1),⁹¹ and it interacts with the fibrinogen (FBG) Bβ chain.⁹²

The ORF3 protein inhibits the mitochondrial apoptosis pathway⁹³ and delays the transport of phosphorylated signal transducer and activator of transcription 3 (pSTAT3) into the cell nucleus.⁸⁴ The ORF3 protein also delays the trafficking and degradation of the activated hepatocyte growth factor receptor (c-Met). Because of its effects on growth factor receptor trafficking, it is believed that the ORF3 protein promotes cell survival.⁹⁴ The ORF3 protein has also been shown to interact with α1-microglobulin and bikunin.^79,80 In cells that express the ORF3 protein, there is an increase in the secretion of α1-microglobulin,⁹⁵ which is mediated by tumor susceptibility gene 101 (Tsg101). Recent studies have revealed that the ORF3 protein binds to Tsg101 via its PSAP motif,^95,96 and that the ORF3 protein interacts with Tsg101 and α1-microglobulin to facilitate the assembly of the endosomal sorting complex required for transport (ESCRT) complexes.⁹⁵ It has also been reported that Tsg101 and the vacuolar protein-sorting (Vps) proteins Vps4A and Vps4B are involved in release of HEV particles. This suggests that the multivesicular body (MVB) pathway is involved in viral release.⁹⁷ It appears that the ORF3 PSAP motif is a late domain involved in the release of membrane-associated HEV particles from infected cells. The interactions between the ORF3 protein and host proteins may create a favorable environment for HEV replication and pathogenesis. In Huh7 liver cells, it has been demonstrated that ORF3 increases the phosphorylation of hepatocyte nuclear factor 4 (HNF4). Phosphorylation impairs the nuclear translocation of HNF4 and downregulates HNF4-dependent gene expression.⁹⁸ Taken together, these results suggest that the ORF3 protein is a multifunctional protein that appears to play an important role in HEV replication and pathogenicity.