Replacement of the Hepatitis E Virus ORF3 Protein PxxP Motif with Heterologous Late Domain Motifs Affects Virus Release Via Interaction with TSG101

Abstract

The ORF3 protein of hepatitis E virus (HEV) contains a “PSAP” amino acid late domain motif, which allows for interaction with the endosomal sorting complexes required for transport (ESCRT) pathway aiding virion release. Late domain motifs are interchangeable with other viral late domain motifs in several enveloped viruses, however, it remains unknown whether HEV shares this functional interchangeability and what implications this might have on viral replication. In this study, by substituting heterologous late domain motifs (PPPY, YPDL, and PSAA) for the HEV ORF3 late domain (PSAP), we demonstrated that deviation from the PSAP motif reduces virus release as measured by viral RNA in culture media. Virus release could not be restored by insertion of a heterologous late domain motif or by supplying wild-type ORF3 in trans, suggesting that the HEV PSAP motif is required for viral exit which cannot be bypassed by the use of alternative heterologous late domains.

Introduction

Hepatitis E virus (HEV) is a non-enveloped, single-stranded, positive-sense RNA virus belonging to the family Hepeviridae (1). HEV is transmitted via a fecal-oral route through contaminated water and food (2, 3) and also as an emerging zoonotic threat (4–6) causing acute hepatitis in immunocompetent and chronic hepatitis in immunocompromised patients (7). For these reasons, hepatitis E is considered a public health concern especially in many developing nations. Currently there are no FDA-approved diagnostics, antivirals, or vaccines on the market. In order to develop effective antivirals and vaccines, it is imperative that different aspects of the HEV life cycle be studied so that they can be explored for potential therapeutic targets.

The HEV genome consists of three open reading frames (ORFs): ORF1 encodes the nonstructural proteins containing methyltransferase (8), cysteine protease (9, 10), helicase (11, 12), and RNA-dependent RNA polymerase (13, 14) domains. The ORF1 also plays a role in interaction with the host causing immunomodulation through inhibition of type I interferon induction (15) and aiding in viral adaptation to the host through the hypervariable polyproline region (16, 17). ORF2 encodes the viral capsid protein (18, 19), and ORF3 encodes a small, 13.5 kDa, multifunctional phosphoprotein (20, 21). Although not critical for virus infection in vitro (22), the ORF3 protein plays essential roles within the HEV lifecycle, and is critical for viral infection in macaques and pigs (23, 24). The ORF3 protein interacts with a myriad of host cellular proteins including SH3 domain containing proteins involved in the MAPK pathway (25, 26), hemopoxin (27), hypoxia-inducible factor 1 (28), fibrinogen Bβ chain (29), retinoic acid-inducible protein I (30), α1-microglobulin, bikunin, and hepsin (31–34) aiding in prolonging cell survival, manipulating the liver microenvironment, and contributing to overall HEV pathogenesis.

Recently, HEV virions were shown to be associated with lipids and this association was linked to a PSAP late domain motif within the ORF3 (35–38). HEV is similar to the hepatitis A virus in that both viruses appear to be released from cells into the bloodstream as enveloped viruses which enables the virus to be protected from the immune system before being excreted through the intestinal tract (39). Once reaching the intestinal lumen, these viruses then lose their envelopes prior to infecting the next susceptible host cell.

Late or “L” domains, so named for their function late in the viral assembly process, play a critical role in the formation of enveloped viral particles from the plasma membrane of infected cells (40, 41). Most enveloped viruses including retroviruses (40, 42), rhabdoviruses (43), and filoviruses (44) encode late domains in their viral genomes. Late domains are characterized by a highly conserved amino acid motif known to facilitate protein-protein interactions between viral and host cellular proteins. Three canonical L domains have been studied to great extent to date: P[TS]AP, PPXY, and YPDL. The P[TS]AP domain binds to the TSG101 (ESCRT-I) gene which also binds to ubiquitin (45) facilitating interaction with the ESCRT machinery. The YPDL “L domain” binds to ALIX/AIP1 which also interacts through TSG101 via its own PTAP motif (46). PPXY domain binds Nedd4-like HECT-domain possessing ubiquitin ligases which allows participation in the ESCRT pathway without the necessity for TSG101 interaction. This pathway has been shown to induce cellular membrane scission events through a process believed to involve ubiquitination, resulting in the formation of viral particles which are released from the cell (47). It is believed that the interactions between components of the ESCRT pathway such as Alix, TSG101, and HECT ubiquitin ligases are facilitated by arrestin-related trafficking (ART) proteins which have been found to be recruited to cellular sites of viral budding (48).

It has been shown that there is a functional interchangeability of these late domains for other virus systems. For example, the human immunodeficiency virus (HIV) PSAP motif or the YPDL motif of equine infectious anemia virus (EIAV) could be used to induce the budding of Rous sarcoma virus which uses a PPPY late-domain to initiate virus release (49). Other studies have shown that the need to recruit ESCRT machinery through the use of other co-factors can be bypassed by experimentally-driven direct recruitment of the ESCRT machinery to allow for viral exit, and ubiquitination of the viral structural protein itself enhances virus release (47). Taken together, the available data suggest that viruses can make use of multiple mechanisms to access the ESCRT pathway, regardless of the late domain that is contained in their genome. Currently, it is unknown if there is a functional interchangeability of the PSAP domain in HEV with late domains used by other viruses.

Results

The PSAP late domain motif is highly conserved among ORF3 proteins from different HEV strains but lysine residues are absent. A BLAST search against the genotype 3 HEV Kernow strain ORF3 protein sequence yielded 258 novel ORF3 protein sequences deposited within the GenBank database. Multiple sequence alignment of the ORF3 sequences showed that there is a highly conserved PSAP motif occurring at ORF3 amino acids 106–109 in HEV genotypes 1–4 (Table S1). Additionally, in 46 out of 258 ORF3 sequences, the PSAP motif was duplicated at amino acid residues 96–99. Only two HEV strains, AFB71119 (a genotype 1 strain isolated from India), and ACU12606 (a genotype 3 strain), lacked a PSAP motif at amino acids 106–109. Strain AFB71119 had no PSAP motif within the entire ORF3 protein whereas strain ACU12606 had the second PSAP motif located at amino acids 96–99 (Table S2). In addition to the conserved PSAP motif, there is a notable lack of lysine residues with only a single lysine residue found in a single genotype 4 HEV strain Q91V27 (Table S1).

Replacement of the HEV ORF3 PSAP motif with heterologous late domain motifs. Overlap extension PCR was utilized to engineer a PPPY, YPDL, and PSAA amino acid motif in place of the wild-type HEV ORF3 PSAP motif (Fig. 1). Due to the specificity and complexity of the inserted amino acid sequences, we were not able to generate the mutations silently within ORF2 because ORF2 overlaps ORF3. Insertion of the PPPY motif within ORF3 caused a Q92P mutation in ORF2 at amino acid position 92, an R93P mutation, and a P94L mutation in ORF2. Similarly, generation of the ORF3 YPDL insertion led to a S91L mutation, a Q92S mutation, and an R93G mutation in ORF2. The insertion of the PSAA motif resulted in a P94R mutation and an A96T mutation in ORF2 (Fig. 1).

Figure 1.

A schematic diagram of substitutions introduced in the PxxP late domain motif within the ORF2/3 reading frames of the Kernow-C1 P6 strain of HEV. Diagram represents the full-length genome of HEV. Black line represents the viral genomic RNA with 5′ cap and 3′ polyadenylation sequences. Boxes represent open reading frames (ORFs) within the genome as labeled. Blowout shows the sequence between nucleotides 5615–5650 with the ORF3 amino acid sequence showing below, followed by the ORF2 reading frame. Mutations inserted in place of the wild-type PSAP motif (underlined) are shown in bold in the nucleotide, ORF3, and ORF2 reading frames.

HEV Kernow C-1 P6 mutant viruses containing substitutions in the ORF3 PxxP motif are replication competent and infectious in vitro. In vitro-transcribed capped RNA from full-length clones of wild-type, PSAA, YPDL, and PPPY HEV mutants were capable of producing HEV ORF2 capsid protein when transfected into the Huh7 S10-3 cells. Immunofluorescent assay (IFA) using antibodies against the HEV ORF2 capsid protein followed by fluorescently-labeled secondary antibodies produced specific signals for each wild-type and mutant HEV tested but not for the untransfected mock sample (Fig. 2A). When quantified via flow cytometry, approximately 28–35% of the transfected cells including the PSAA, YPDL, and PPPY HEV mutants produced ORF2 protein. When the freeze-thawed cell lysates from these transfected Huh7 S10-3 cells were used to infect HepG2/C3A cells, we again observed HEV-specific fluorescent signals after 5 days post-infection for wild-type and mutant HEVs, indicating that these HEV mutants are infectious although to a lower degree than the wild-type HEV (Fig. 2B). When quantified, wild-type HEV infected 4.2% of the cells, whereas PSAA mutant HEV infected 0.5% of the cells, PPPY mutant infected 1.75% of the cells, and YPDL mutant infected 0.35% of the cells (Fig. 2C) while all mock infected cells consistently generated zero or negative values.

Figure 2.

Wild-type and ORF3 HEV mutants retain replication competence and infectivity in vitro. (A). Fluorescent microscopy images of Huh7 S10-3 cells transfected with in vitro synthesized capped RNA transcripts from wild-type or ORF3 mutant HEV probed at 5 days post-transfection with anti-ORF2 rabbit antibodies followed by anti-rabbit fluorescein isothiocyanate (FITC) secondary antibody. (B). Freeze-thawed cellular lysates from Huh7 cells transiently transfected with in vitro synthesized capped RNA transcripts were used to infect HepG2/C3A cells. At seven days post-infection, cells were probed with anti-ORF2 rabbit antibodies followed by anti-rabbit FITC secondary antibody. (C). Flow cytometry quantification of transfected Huh7 S10-3 (T) or infected HepG2/C3A cells (I) from each mutant virus at 5–7 days post-transfection or infection probed with anti-ORF2 rabbit antibodies followed by anti-rabbit phycoerythrin (PE) secondary antibodies.

Wild-type and ORF3 PxxP HEV mutants localize to cellular endosomes within Huh7 cells. In order to understand why the late domain mutants were less infectious than the wild-type virus, we constructed fluorescent expression vectors with the wild-type and mutant HEV ORF3s tagged with fluorescent molecules allowing visualization of ORF3 localization within the cells. HEV ORF3 localized to distinct regions within the cellular cytoplasm and was not found within the nuclei (Fig. 3, and Fig. S1). By co-transfecting the Huh7 cells with expression vectors encoding endosomal markers, we were able to show that the wild-type HEV ORF3 localized predominantly with Rab11, to a much lesser extent Rab7, but not with TGN38 (Fig. 3), showing that the localization of ORF3 is predominantly within recycling endosomes and to a lesser extent the late endosomes reconfirming previous studies (36, 50). The ORF3 from YPDL, PPPY, and PSAA HEV mutants colocalized with wild-type ORF3 when expressed within the same cell (Fig. S1).

Figure 3.

Wild-type and mutant HEV ORF3 proteins localize to cellular late and recycling endosomes. Huh7 cells were transfected with either wild-type HEV-eCFP or mutant (PPPY, YPDL, PSAA) ORF3 HEV-eCFP constructs and were co-transfected with either Rab7-eYFP, a marker for late endosome (A), Rab-11-eYFP a marker for recycling endosomes (B), or TGN38-eYFP a marker for the trans-golgi network (C). The eYFP fluorescence is pseudo-colored green while the eCFP fluorescence is pseudo-colored red. Colocalization is seen as orange in the merged image.

As ORF3 is known to self-associate (51), it was possible that wild-type ORF3 could function to properly redistribute mutant ORF3. Therefore, we expressed each mutant strain alone within cells expressing Rab7, Rab11, and TGN38 to determine the subcellular localization of each individual wild-type and mutant ORF3. The ORF3 YPDL, PPPY, and PSAA mutants each showed a similar localization as the wild-type ORF3, predominantly recycling endosomal (Fig. 3), although the PSAA mutant had slightly more late endosomal localization than the other ORF3 proteins (Fig 3A, PSAA). The fact that the PSAA mutant also localized to the endosomes demonstrates that the localization of ORF3 to the cellular endosomes is not dependent on the interaction between the PSAP late domain and the ESCRT pathway proteins.

Wild-type HEV, YPDL and PPPY mutant ORF3 proteins colocalize with the TSG101 protein and have positive Förster resonance energy transfer (FRET) efficiencies suggesting interaction between these late domains and TSG101. Co-expressing wild-type and mutant ORF3 eCFP with TSG101 eYFP in the same cell demonstrated colocalization of TSG101 and each ORF3 within the cell (Fig. 4A). Though colocalization occurred with all of the wild-type and mutant ORF3 proteins and TSG101, this colocalization occurred less with ORF3 PSAA eCFP than anticipated since we did not expect PSAA to interact with either TSG101 or the rest of the ESCRT machinery.

Figure 4.

Wild-type and mutant ORF3s colocalize and interact with the host protein cellular protein tumor suppressor gene 101 (TSG101). (A). TSG101 tagged with eYFP colocalizes with HEV ORF3 proteins tagged with eCFP. (B). Förster resonance energy transfer (FRET) demonstrates that TSG101 is within a biologically relevant distance to be interacting with wild-type, PPPY, YPDL, but not PSAA-containing ORF3 proteins tagged with eCFP. Asterisks (*) indicate significantly greater FRET efficiencies than eCFP-eYFP control (P< 0.05).

To further determine whether ORF3 interacts with TSG101, we performed acceptor photobleaching FRET which is a biophysical non-radiative transfer of energy between fluorescent molecules with overlapping excitation and emission spectra that are brought within a distance of 1–10nm of each other consistent with a biologically relevant physical interaction (52–55). The FRET efficiencies for the negative control of eCFP and eYFP was 6.18%, whereas wild-type ORF3 and TSG101 produced a FRET efficiency significantly greater than eCFP and eYFP (21.49%, P = 0.0007). The FRET efficiency found for the PSAA mutant was not significantly different than the FRET efficiency of the negative control (9.05%, P = 0.265). The FRET efficiencies for ORF3 PPPY (20.89%, P < 0.0001) and YPDL (13.25%, P = 0.013) were significantly greater than eCFP and eYFP alone, suggesting that they were within 10 nm of each other, thus indicative of an interaction between these ORF3 proteins and TSG101 (Fig. 4B). The fact that these late domains were found in such close proximity to the TSG101 motif implies that they are likely interacting with each other. This was expected with the wild-type HEV ORF3 as the PSAP motif is thought to interact directly with TSG101 but is unexpected in the cases of PPPY and YPDL, as these domains are thought to interact with other components of the ESCRT machinery. PSAA was not found in close proximity to TSG101 as expected from previous studies showing that this mutation had release defects (36, 56).

PPPY, YPDL, and PSAA HEV mutants have significantly decreased viral release compared to wild-type HEV. Since we demonstrated that each PSAP ORF3 substitution mutation was still capable of interacting with TSG101 and hence the ESCRT pathway, we next determined if they were capable of allowing virus release from the cell. In order to determine virus release, we performed HEV-specific qRT-PCR on RNA isolated from cellular media and lysates separately, and then compared the amount of viral RNA released into the media to the total level of RNA expression in media and lysate collected from separate samples on days 5, 6, and 7 then combining the results taken over these days. Wild-type HEV had the most viral RNA in the media with 8.25% of the total RNA being released into the cell culture media, and as expected the PSAAORF3 mutant had the least amount of RNA released into the media (0.76%) (Fig. 5A). The HEV mutants containing the PPPY ORF3 and YPDL substitutions released significantly less viral RNA than that of the wild-type virus (1.30%, P = 0.039 and 1.32%, P= 0.031, respectively).

Figure 5.

Substitutions within the HEV ORF3 PSAP motif do not restore virus release and cannot be complemented by providing ORF3 protein in trans. (A). Quantitative reverse transcription real time PCR (qRT-PCR) using primers specific for HEV RNA from samples taken from infected cell culture supernatant compared to RNA within the cellular lysate demonstrated that disruption of the PSAP motif decreases the ability of HEV to be released into the media (PSAA compared to wild-type). Substitution PPPY and YPDL were not capable of restoring release to wild-type levels. (B). Silent mutation of the ORF3 reading frame to insert a Q92L and R93C amino acids in ORF2 while retaining the PSAP motif within ORF3 reduce levels of virus release. (C). Cells infected with either wild-type or PSAA ORF3 mutants subsequently transfected with plasmid DNA expressing wild-type ORF3 GFP or ORF3PSAAGFP were not capable of restoring or augmenting release of wild-type or PSAA mutant virus as assessed by titering the amount of virus released over a period of 24 hours on naïve HepG2/C3A cells. Asterisks (*) indicate significantly lower than wild-type (P < 0.05). Octothorpes (#) indicate significantly greater than PSAA mutant (P > 0.05).

Density of cell culture-derived HEV virions. Cell culture supernatants harvested from cells transfected with wild-type capped HEV RNA at 5 days post-transfection predominantly banded at densities of 1.15–1.16 g/ml (21%) and 1.25g/ml (20%) (Fig. 6, wild-type), consistent with the findings of Takahashi et al (57) in which our sample contained both lipid associated (1.16g/ml) and non-lipid associated (1.25g/ml) HEV virions. Disruption of the PSAP ORF3 late domain motif led to a dramatic shift of the HEV bands to the 1.25g/ml fraction (44%) compared to only 1% remaining within the 1.16g/ml fraction (Fig 6, PSAA). Insertion of the YPDL and PPPY motifs led to a higher percentage of HEV RNA at the 1.16g/ml peak than PSAA (6.7% and 4.8%, respectively) and retained wild-type percentages of HEV RNA at the 1.25g/ml peaks (18.2% and 29.3%, respectively). The YPDL mutant contained an intermediate peak (24.6%) at ~1.19 g/ml, and PPPY mutant contained an intermediate peak (23.6%) at 1.21g/ml, suggesting a certain level of association with lipids was present but not to the extent of wild-type HEV (Fig. 6 YPDL and PPPY).

Figure 6.

Sucrose density centrifugation of HEV particles from the supernatants of Huh7 cells transfected with wild-type and HEV ORF3 mutants. Twenty-one fractions from a 10–60% sucrose gradient overlaid with cell culture media from Huh7 cells transfected with the indicated full-length viral RNA were collected. RNA was isolated, quantified via qRT-PCR and plotted against the density of the fraction determined via refractometry. RNA is plotted as percentage of the total RNA measured from all fractions. Dotted lines were inserted at 1.16g/ml and 1.25g/ml as this was previously published to coincide with lipid-associated virions and unenveloped virions, respectively (57).

Mutations within the ORF2 contribute to the reduction in virus release from the cell. As we were not able to restore viral release to wild-type levels with either the PPPY or the YPDL substitutions, we wanted to determine whether viral factors outside of ORF3 could be contributing factors to viral RNA release from the cell. Since both the YPDL and PPPY substitutions also altered amino acids Q92 and R93 within the ORF2 reading frame (Fig 1), we wanted to assess whether these amino acid substitutions within ORF2 could be contributing to a decrease in virus release. To test this, we generated Q92L and R93C mutations within ORF2 without altering the amino acid sequences within ORF3 (Fig. 1 Q92L and R93C). When these mutants were assayed for viral RNA release, both appeared to be numerically reduced from wild-type, ORF2 Q92L released 3.36% of its viral RNA into the media and ORF2 R93C released 1.85% of its viral RNA into the culture media (Fig. 5 B), however neither was significantly different from wild-type (WT:Q92L, P = 0.26; WT:R93C, P = 0.13). Additionally, the amount of viral RNA released by ORF2 Q92L and R93C were significantly greater than the release of the PSAA mutation (P = 0.025 and 0.0272, respectively), suggesting that the impact of these ORF2 mutations was minor but, as we were not capable of silently recapitulating the exact mutations found in the PPPY and YPDL substitutions, we cannot rule out more adverse effects on virus release from the specific amino acids found within our ORF3 substitution mutant’s capsids.

Addition of ORF3 protein in trans was not sufficient to restore release of virus containing the PSAA ORF3 mutation. In order to further rule out mutations within the ORF2 protein affecting viral release, we attempted to use a gain of function assay by providing wild-type ORF3 protein in trans by transient transfection of cells previously transfected with either wild-type or PSAA mutant HEV viral RNA and then assessed the amount of infectious virus released from the cells in a 24 hour period. Wild-type virus produced enough particles to infect 3.51% of HepG2/C3A cells, wild-type plus ORF3GFP produced particles which infected 1.89% of the HepG2/C3A cells, and wild-type with PSAAORF3GFP produced enough particles to infect 1.32% of the HepG2/C3A cells, none of which was significantly different from one another (Fig. 5C). PSAA mutant alone produced particles capable of infecting 0.59% of the HepG2C3a cells, significantly less than wild-type alone (P = 0.048). PSAA with wild-type ORF3GFP produced particles capable of infecting 1.045% of the HepG2/C3A cells and PSAA plus PSAAGFP produced particles capable of infecting 0.94% of the HepG2/C3A cells, not significantly greater than the PSAA mutant alone (Fig. 5C). Additionally, there was no significant difference between both wild-type plus ORF3GFP and wild-type plus PSAAGFP compared to PSAA alone suggesting that overexpression of a C-terminally tagged ORF3 was detrimental to virus release. When observing viral RNA release, there was also no significant difference between wild-type and wild-type plus ORF3GFP or wild-type plus PSAAORF3GFP, and there were no significant difference in viral RNA release between PSAA, PSAA plus ORF3GFP, and PSAA plus PSAAORF3GFP (data not shown).

Discussion

Much of the assembly stage during HEV virion formation remains unexplored. Previous studies have shown that other viruses are able to use alternative amino acid domains in order to bypass specific TSG101 interactions and allow for viral exit (47, 49). In this study, we investigated the mechanisms of envelopment and release phases of HEV assembly through substitutions of heterologous PPPY and YPDL late domains in place of the HEV ORF3 wild-type PSAP late domain. In an effort to determine if HEV shares this functional interchangeability with other heterologous late domains, multiple approaches were used to create different virus mutants and then examine how they replicated, how their proteins interacted with the host ESCRT pathway, and how these mutations affected their ability to exit cells.

The results from this study showed that, when the HEV ORF3 PSAP motif is substituted with heterologous late domain motifs, the resulting substitution mutants of HEV were replication competent, producing viral capsid protein, and the ORF3 protein of the HEV mutants was capable of localizing to the same cellular organelles as the wild-type HEV ORF3 protein. However, the HEV ORF3-specific PSAP motif was required for the virus to efficiently exit host cells. Without this interaction, viral exit was significantly impaired as seen with the PPPY and YPDL ORF3 substitution mutants. These results also imply that this specific interaction between PSAP and the ESCRT pathway is crucial for viral exit and spread to other host cells. Unlike retroviral structural proteins which can be directly ubiquitinated by host factors to aid in particle release (47), direct ubiquitination of HEV ORF2 (which previous studies have shown that it does not appear to be ubiquitinated (58) or direct ubiquitination of HEV ORF3 (as there are almost no mammalian HEV strains possessing lysine) likely does not play a role in HEV envelopment and release. Therefore, future experiments could explore blocking this specific interaction as a potential antiviral target.

We demonstrated that the PSAP late domain is not required for cellular localization of ORF3 to endosomes, as evidenced by the fact that all the HEV mutants were able to localize to endosomal vesicles. This implies that interaction with the ESCRT machinery is not required for localization to endosomes and suggests that the ESCRT machinery may more specifically be required solely for HEV viral budding rather than intracellular transport of the viral ORF3 protein. The results from our FRET results yield interesting findings in that we did not expect the PPPY substitution mutant to interact with TSG101 as this motif is known to interact with the NEDD4 family of ubiquitin ligases rather than TSG101. It may be possible that TSG101 localization to the same endosomal compartment as the PPPY ORF3 mutant is causing an artificial FRET signal due to macromolecular crowding. The small increase that was not significantly different than eCFP/eYFP alone in the PSAAORF3-TSG101 would also argue that this could be occurring or the interaction of the PPPY motif with the NEDD4 family may create some types of higher ordered complex which places PPPY ORF3 in very close proximity to TSG101. Further research is required to determine whether these PPPY and YPDL motifs are indeed directly interacting with TSG101 and if so, why this interaction is not capable of wild-type levels of virus particle release.

Though the HEV mutants in this study were able to infect, replicate and localize in vitro, the use of real time qPCR demonstrated that the late domain is required for viral exit from the host cell. By determining the ratio of viral RNA that was found in the cell lysates compared to cellular media in the HEV mutants, we found that mutant HEV viral exit from Huh7 cells was significantly reduced when compared to wild-type HEV. This was unexpected, as it was thought that HEV would show functional interchangeability in late domains similar to other viruses.

It is intriguing that we were not capable of restoring PSAA HEV virus release to wild-type levels by simply providing ORF3 protein in trans. Several explanations for this phenomenon could be proposed. The simplest explanation is that the ORF3 protein requires an unhindered C-terminus to contribute to the release stage of the virus assembly process, and that by using an ORF3 protein with a fluorescent tag rendered it unable to complement the PSAA mutation. This is evident in Figure 5C when wild-type virus is expressed with either ORF3GFP or PSAAGFP, where wild-type alone becomes statistically insignificant from PSAA virus alone. A second possibility is that the PSAA mutations act as a dominant negative mutation and if any PSAA ORF3 is present, the budding efficiency is altered. We do see a numerical decrease in the number of infectious wild-type particles released when cotransfected with PSAA GFP, although it was not statistically different. It could also be that supplying an overabundance of ORF3 interferes with other aspects of the virus assembly pathway as we saw a decrease, although not statistically significant, in infectious particles simply by co-expressing wild-type ORF3GFP protein with wild-type virus. More experiments will be necessary in the future to define why we could not complement the ORF3 mutations with wild type ORF3 protein supplied in trans.

Due to its ability to cause acute hepatitis in the general population, fulminant hepatitis in pregnant women, and chronic hepatitis in immunosuppressed individuals, HEV is an important but extremely understudied pathogen. It is imperative that we understand key aspects of HEV infection and viral spreading in order to devise effective preventive and control strategies against HEV. By demonstrating that the PSAP motif in the late domain of HEV ORF3 is required for viral exit and cannot be bypassed by the use of alternative heterologous late domains, it highly suggests that the PSAP-TSG101 interaction is specifically required for virion release. In this way, attempting to antagonize or block this interaction may block virion exit and the spread of virus from cell to cell, which could have a clinical impact on the course or severity of HEV infection. By identifying this specific interaction required for viral spreading, we have identified a potential antiviral target. Previous studies have been successful at blocking the Gag-TSG101 interaction in HIV with the use of small molecule therapeutics (59), and inhibiting the PPxY-Nedd4 interaction (60) in other viruses. Such targeted inhibitions of the ESCRT machinery and its components should be explored for HEV in future studies.

Materials and Methods

Expression vectors, plasmids, and cells. The pBlueScript SK (+) vector containing the Kernow-C1 passage 6 (P6) genotype 3 HEV infectious cDNA clone has been previously described (61). Fluorescent vectors used for the ORF3 expression in this study were enhanced cyan fluorescent protein (eCFP) and enhanced yellow fluorescent protein (eYFP)-N1 vectors (Clontech, Mountain View, CA). The Huh7 S10-3 human hepatoma cell line (a gift of Suzanne U. Emerson, NIAID, NIH, Bethesda, MD) and HepG2/C3A hepatocellular cell line (ATCC: CRL10741) were used for localization and infectivity assays.

Construction of mutant HEV viruses containing substitutions in the HEV ORF3 PxxP motif. Using two rounds of overlap extension PCR (OEPCR) with pBlueScript Kernow-C1 passage 6 HEV infectious clone as the template, HEV mutants with PPPY and YPDL motifs substituted for the wild-type PSAP motif were created. Additionally, a PSAA mutant was created to act as a negative control. For generation of the PPPY mutant, primer pairs JLW10/JLW2 and JLW3/JLW11 (Table 1) were used for the first round OEPCR, and primer pair JLW10/JLW11 was used for the 2^nd round PCR. After OEPCR, a PPPS mutation was verified by sequencing. Primer pair JLW1/JLW4 was used to correct PPPS to PPPY in an additional, third round of PCR-directed mutagenesis. For generation of YPDL mutant, primer pairs JLW10/SPK348 and SPK349/JLW11 were used for the first round OEPCR and primer pair JLW10/JLW11 was used for 2^nd round PCR (Table 1). For generation of PSAA mutant, primer pairs JLW10/JLW8 and JLW9/JLW11 were used for the first round OEPCR and primer pair JLW10/JLW11 was used for 2^nd round OEPCR (Table 1). For generation of the ORF2 Q92L mutation, primer pairs JLW10/SPK575 and SPK576/JLW11 were used for the first round OEPCR followed by JLW10/JLW11 for the second round PCR amplification (Table 1). Generation of the ORF2 R93C mutation was done using primer pairs JLW10/SPK577 and SPK578/JLW11 for the first round PCR followed by JLW10/JLW11 for the second round PCR amplification (Table 1). Following the OEPCR mutagenesis, the PCR products were ligated into the Kernow-C1 P6 viral vector and verified by sequencing.

Table 1.

Oligonucleotide sequences utilized in OEPCR mutagenesis, qRT-PCR, and for cloning into expression vectors.

Name	Sequence (5′–3′)	Function
SPK337	CATCATCTCGAGCCATGGGATCACCATGCG	Used to amplify ORF3 for insertion into fluorescent vector using Xho1 restriction site.
SPK341	CATCATAAGCTTGTACAGGGGCTGTGTCAGG	Used to amplify ORF3 for insertion into fluorescent vector, HindIII restriction site.
JLW10	CAGTGTTTGCTGCCGCTG	External OEPCR for ORF3 mutations.
JLW11	GGAGGTAGCCTCCTCCTC	External OEPCR for ORF3 mutations.
JLW8	GCAGTGGAGCGGCGC	Creation of PSAA mutant, round 1 OEPCR.
JLW9	GCGCCGCTCCACTGC	Creation of PSAA mutant, round 1 OEPCR.
SPK348	AAGGTCCGGATAACTGGTCACGCCAAGCGG	Creation of YPDL mutant, round 1 OEPCR.
SPK349	TATCCGGACCTTCCGCTGCCCCCCG	Creation of YPDL mutant, round 1 OEPCR.
JLW2	ATAGGGGGGGGGACTGGTCACGCC	Creation of PPPY mutant, round 1 OEPCR.
JLW3	GGCGTGACCAGTCCCCCCCCCTAT	Creation of PPPY mutant, round 1 OEPCR.
JLW1	CTGATGTTGTGCGTTTCGC	PPPS to PPPY in 3rd round OEPCR.
JLW4	CATACATAGGGTTGGTGGACG	PPPS to PPPY in 3rd round OEPCR.
SPK575	GACCAGTCCCTCCGCC	Creation of Q92L mutant, round 1 OEPCR.
SPK576	GGCGGAGGGACTGGTC	Creation of Q92L mutant, round 1 OEPCR.
SPK577	GACCAGTCCCAGTGCCC	Creation of R93C mutant, round 1 OEPCR.
SPK578	GGGCACTGGGACTGGTC	Creation of R93C mutant, round 1 OEPCR.
SPK373	GGTGGTTTCTGGGGTGAC	qRT-PCR forward ORF2 5311
SPK374	AGGGGTTGGTTGGATGAA	qRT-PCR forward ORF2 5380
SPK375	56-FAM/TGATTCTCA/ZEN/GCCCTTCGC/3IABKFQ	qRT-PCR ORF2 probe
SPK592	GCGATGCTTATGAGGAGTCA	qRT-PCR forward ORF1 4549
SPK593	GGCCGAGAGAGAAGTTGTTC	qRT-PCR reverse ORF1 4646
SPK594	56-FAM/TGCTTGCTG/ZEN/CCGCTGTGTCA/3IABKFQ	qRT-PCR ORF1 probe

In vitro transcription of full-length capped viral genomic RNA, transfection of Huh7 S10-3 cells, and infection of HepG2/C3A cells. The pBluescript Kernow-C1 P6 HEV infectious clone and its derived subsequent mutants were linearized via MIuI restriction enzyme digest, proteinase K treated, then purified via phenol chloroform extraction and ethanol precipitation. Linearized DNA was used as the template for in vitro transcription via a T7 Ribomax RNA transcription kit (Promega) following the manufacturer’s protocol for capped RNA synthesis using anti-reverse cap analogue (ARCA) RNA (Tri Link Biotechnologies). RNA transcripts were visualized via gel electrophoresis on 0.8% agarose gels, and subsequently purified via Qiagen RNA columns and treated with turbo DNase (Ambion). Reactions were subjected to qRT-PCR and normalized based on CT values prior to transfection. Huh7 subclone S10-3 cells noted for their ability to support HEV replication were used for initial transfection. Normalized in vitro-transcribed capped RNA was transfected into Huh7 S10-3 cells using a Trans-IT-mRNA transfection kit (Mirus). Cells at day 2 post transfection were passaged 1:2 onto two new T25 flasks. Five to six days post transfection flasks of cells were frozen at −80°C and thawed 3 times. Cellular debris was spun out at 1000 × g for 10 minutes with the resultant supernatant being used to infect naïve HepG2/C3A cells. Equivalent volumes of cleared cellular lysate was incubated with naïve HepG2/C3A cells for 4 hours at 37°C before changing to fresh primary growth media and allowing the cells to grow for an additional 4–5 days.

Immunofluorescence assays (IFA) and flow cytometry. Four to five days post-transfection or infection cells were trypsinized and either re-seeded onto chamber slides and fixed one day after re-seeding using 80% methanol, or cells were spun down at 525 × g and resuspended in 100% methanol at 4°C for 15 minutes then frozen at −80°C until processed for flow cytometry. For confocal imaging, cells were probed using anti-HEV ORF2 rabbit antibodies followed by goat anti-rabbit fluorescein isothiocyanin (FITC) secondary antibodies (KPL). Cells were stained with 20 mg/ml DAPI (Sigma Aldrich) diluted 1:10,000 in phosphate buffered saline (PBS) for 5 minutes at room temperature, washed twice more in PBS, and mounted on slides using aqua polymount (Polysciences Inc). Fluorescence microscopy was performed using a Nikon TE2000 sweptfield confocal system using Elements software (Nikon, USA). To visualize the FITC and DAPI signals, an X-cite series 120 light source was used for excitation in conjunction with a DAPI-FITC-TRITC emission filter cube and Photometrics CoolSNAP HQ² camera. For flow cytometry, cells were pelleted out of methanol and resuspended in 300ul of blocking solution (1.0% nonfat dried milk, 1.0% Odyssey Block (Licor Biosciences), 1.0% normal goat serum, and 0.1% Triton X-100) at 37°C for 30 minutes. Cells were probed using anti-HEV ORF2 rabbit antibodies for 1 hour at 37°C, followed by 2 washes with PBS, incubation with goat anti-rabbit IgG (H+L) phycoerythrin conjugated secondary antibody (Life Technologies) at a 1:1000 dilution for 30 minutes at 37°C, followed by 2 more washes with PBS. Cells were analyzed with a FACS Aria II (BD Biosciences) and Flow Jo software. A minimum of 20,000 events were recorded with gates being set based on mock samples returning the lowest amount of background signal with positive signal being subtracted from all test samples. Randomly tested mock samples returned no positive signals.

Construction of recombinant vectors expressing fluorescently-tagged ORF3 fusion proteins and determination of sub-cellular localization of wild-type and mutant ORF3. HEV Kernow-C1 P6 plasmids containing wild-type ORF3 or ORF3 mutants containing PPPY, YPDL, and PSAA motifs in place of the wild-type PSAP motif were amplified using primers SPK337 and SPK341 (Table 1). These primers were used to create XhoI and HindIII restriction sites to clone ORF3 in frame with eYFP or eCFP. All mutations were confirmed via sequencing. Huh7 S10-3 cells were co-transfected with eYFP expression vectors containing intra-cellular organelle markers for recycling endosomes (Rab11-eYFP) (62), late endosomes (Rab7-eYFP) (62), and trans Golgi network (TGN38-eYFP) (63) using lipofectamine LTX reagent (Invitrogen) and TSG101-eYFP (64). Cells were fixed 17–24 hours post-transfection using 4% paraformaldehyde and mounted to slides using aqua polymount. Confocal microscopy was performed using a Zeiss LSM 510 confocal microscope with Zen 2009 software (Zeiss, USA) to look for overlap between eCFP and eYFP fluorescence to determine the subcellular localization of the wild-type HEV ORF3 and mutant ORF3. eYFP signals were acquired using the 514 nm laser setting and eCFP signal was obtained using the 458 nm laser settings using a Plan-Apo 63X oil immersion lens.

Confocal microscopy and Förster resonance energy transfer (FRET). Huh7 S10-3 cells grown on coverslips were transfected using lipofectamine LTX. At 18–24 hours post-transfection, cells were fixed in 4% paraformaldehyde and mounted using aquapolymount (Polysciences Inc). Cells were imaged on a Zeis LSM510 confocal microscope using a plan-apo 63X oil immersion objective. eCFP fluorophores were excited using 458 nm light and eYFP was excited using 514 nm.

For FRET assay, cells were initially imaged, bleached for 11–16 iterations of 100% 514 nm laser intensity. Cells were bleached to 30% or less of initial eYFP intensity. FRET intensity was calculated as (Donor Post corrected for non-specific bleaching-Donor Pre)/Donor Pre *100. A minimum of 10 total cells were analyzed for each construct from at least two separate transfections.

RT-PCR detection of HEV RNA released into cell culture media. T25 flasks containing Huh7 S10-3 cells were transfected with capped HEV RNA transcripts using the Trans-IT mRNA transfection kit (Mirus) and washed twice with PBS 24 hours post-transfection. At two days post-transfection, cells were removed from the flask using 2 ml of 0.25% trypsin and 4 ml of Dulbecco’s modified eagle’s media (DMEM) with 10% fetal bovine serum (FBS). Approximately 500 μl of cells were replated onto 12 well dishes. Media was harvested from each well at 5, 6, and 7 days post-transfection. Cellular debris was spun out of the media (10,000 x G for 1 minute). Cells were removed from the plate using 500 μl of trypsin and 500 μl of DMEM/10% FBS. Samples were frozen at −80°C until processed. Viral RNA was extracted from media and lysates using a viral RNA isolation kit (Zymo). RNA was DNase-treated using a turbo DNA-free kit (Life Technologies). Quantitative real time qRT-PCR was performed using a SensiFAST real-time PCR kit (Bioline, USA) using primers SPK373 and SPK374 with probe SPK375 (Table 1) (65) to determine the amount of RNA released into the media compared to the total RNA expression in each well. Purified in vitro transcribed full-length viral RNA transcripts from either genotype 1 or genotype 3 HEV were quantified via spectrophotometry, serially diluted from 0.01ng to 0.00000001ng, and used in the RT-PCR reaction to generate a standard curve to which the unknown samples were compared to generate absolute values for the quantity of HEV RNA in each tested sample. Once the absolute quantity of RNA was generated from each sample, the amount of viral RNA in the media was divided by the total amount of RNA in both the media and lysate giving us a ratio of total viral RNA released from the cell compared to the amount of total RNA in both media and lysate.

Complementation of HEV RNA release with ORF3 provided in trans. T25 flasks containing Huh7 S10-3 cells were seeded and transfected with wild-type or PSAA HEV RNA as described for RT-PCR detection of HEV RNA release. On day 5 post-transfection, cells were transfected with either eYFP vector alone, wild-type ORF3 eYFP, or PSAA ORF3 eYFP. Transfected cells were washed three times at 4–6 h post-transfection, and 0.2 ml of DMEM/10%FBS was added to each well at 12 h post-transfection. Transfected cells were allowed to grow for another 24 h before media and lysates were harvested separately and stored at −80°C until processed. Equivalent volumes of media were used to infect 1.5 × 10⁴ HepG2/C3A cells plated on 48 well dishes in quadruplicate. Cells at two days post-infection were split 1:2 and allowed to grow for an additional 4 days before being removed from the plate and processed for flow cytometry.

Sucrose density gradient equilibrium centrifugation. A density gradient was formed in an SW50Ti tube (Beckman Coulter Inc., Fullerton CA) containing 0.8 ml 60% sucrose [wt/wt], 0.6 ml of 50%, 40%, 30%, and 20% [wt/wt] sucrose, and 0.3 ml of 10% sucrose in PBS (66). One hundred microliters of each sample (wild-type HEV Kernow P6, PPPY, YPDL, and PSAA) was layered onto the top of each gradient and overlaid with 0.4 ml of PBS. Samples were centrifuged at 179,200 x g at 4°C for 19 h, and 200 μl aliquots were removed from the top of the gradient. The density of each fraction was measured via refractometry. RNA was isolated from 100 μl of each fraction using a ZR viral RNA kit (Zymo Research) and the amount of vRNA in each fraction was determined using quantitative real-time qRT-PCR using the SPK373, SPK374, and SPK375 primer/probe set (Table 1).

Highlights.

• HEV viral egress is reliant on an intact PSAP motif in ORF3.

• HEV ORF3 subcellular localization is not dependent on the PSAP motif.

• Direct ubiquitination of ORF3 does not play a role in HEV lipid association.

• Interactions between ORF3 and TSG101 can be exploited as a potential antiviral target.