|Home | About | Journals | Submit | Contact Us | Français|
Hepatitis B virus (HBV) and hepatitis delta virus (HDV) share the HBV envelope proteins. When woodchucks chronically infected with woodchuck hepatitis virus (WHV) are superinfected with HDV, they produce HDV with a WHV envelope, wHDV. Several lines of evidence are provided that wHDV infects not only cultured primary woodchuck hepatocytes (PWH) but also primary human hepatocytes (PHH). Surprisingly, HBV-enveloped HDV (hHDV) and wHDV infected PHH with comparable efficiencies; however, hHDV did not infect PWH. The basis for these host range specificities was investigated using as inhibitors peptides bearing species-specific pre-S (where S is the small envelope protein) sequences. It was found that pre-S1 contributed to the ability of wHDV to infect both PHH and PWH. In addition, the inability of hHDV to infect PWH was not overcome using a chimeric form of hHDV containing WHV S protein, again supporting the essential role of pre-S1 in infection of target cells. One interpretation of these data is that host range specificity of HDV is determined entirely by pre-S1 and that the WHV and HBV pre-S1 proteins recognize different receptors on PHH.
Hepatitis B virus (HBV) still remains a major health problem, with about 400 million infected people worldwide and approximately 1 million deaths a year caused by HBV-associated liver cancer (35). In natural situations people can be coinfected with hepatitis delta virus (HDV) and HBV, or HBV carriers can be superinfected with HDV (40). HDV is a subviral agent that uses the envelope proteins of HBV for virion formation and infectivity (42). The HDV genome is a negative-sense single-stranded circular RNA, which is known to fold into a rod-like structure with 74% self-complementarity. The genome is replicated via RNA-directed RNA synthesis, apparently catalyzed by host RNA polymerase II (41). Three processed HDV RNAs accumulate in infected cells: (i) the genome, (ii) its exact complement, the antigenome, and (iii) a 5′-capped and 3′-polyadenylated mRNA for the only viral protein, the delta antigen (δAg), which is essential for replication. An RNA-editing event, which occurs on antigenomic RNA, leads to the production of a longer form of δAg that does not support RNA replication but is required for HDV assembly (6, 7).
Hepadnaviruses are considered to be highly species specific due to recognition of species-specific receptors on hepatocytes, the target of infection. HDV is presumed to use the same means for attachment and entry as HBV (36) and, therefore, to share this species specificity. A number of candidates have been proposed as the HBV and HDV receptors, but none of them has been shown to be sufficient or even necessary for infection (35). In addition, the number of the cell surface receptors used by HBV, HDV, and other hepadnaviruses is unknown. Woodchuck hepatitis virus (WHV) is a hepadnavirus with many similarities to HBV. Like HBV, the envelope proteins of WHV can be used to assemble HDV particles (31), referred to here as wHDV, to distinguish them from hHDV, HDV with HBV envelope.
It is known that chimpanzees can be infected with HBV obtained from patient serum or transfected cells (34) and also with hHDV collected from infected patients (28-30). Furthermore, hHDV from either an infected chimpanzee or transfected cells is able to infect primary chimpanzee and human hepatocytes (4, 37, 38). hHDV passaged in chimps can infect WHV-chronic carrier woodchucks (21, 26), and yet, as considered further in the Discussion, this probably involves the spread of wHDV. Serum-derived hHDV cannot infect primary woodchuck hepatocytes (PWH) (39). Also there are no data supporting the possibility of productive infection of woodchucks with HBV (26).
The HDV from an infected woodchuck, wHDV, is able to infect PWH regardless of the presence of WHV (39) and is also able to reinfect chimpanzees (20, 29). In contrast, there is no evidence that WHV itself can infect humans or other primates (25, 35, 43). Similarly, primary tupaia hepatocytes which are susceptible to HBV infection are resistant to WHV (17). It is also known that woodchucks are susceptible to infection with either WHV (32) or wHDV that has been assembled in cultured cells (22, 23, 31).
In the present study, we have asked if host range susceptibilities of HBV, WHV, and the homologous forms of HDV are really at the level of receptor recognition. Our results, albeit with HDV, are consistent with the following interpretations: (i) WHV and HBV and the related HDV recognize different host receptors, (ii) recognition is via pre-S1 (where S is the small envelope protein), and (iii) WHV is able to recognize a receptor on human cells but not vice versa (that is, HBV cannot recognize a receptor on woodchuck cells). The failure of WHV to infect human hepatocytes occurs, by inference, at a step after receptor recognition.
One source of wHDV was serum from woodchucks chronically infected with WHV and then superinfected with HDV (23). Alternatively wHDV or hHDV was assembled from Huh7 cells cotransfected with pSVLD3 to initiate HDV replication, together with plasmids expressing WHV or HBV envelope proteins, using procedures as previously described (16).
For the experiments shown in Fig. Fig.7,7, assembly was performed using combinations of HBV and WHV envelope-expressing constructs. HBV S envelope protein was expressed from pSVBX24H (11). WHV S envelope protein was expressed from pSV24W (11). The HBV large envelope protein (L) protein was expressed from a construct, pSVL, on which the initiation codons for pre-S2 and S were mutated to threonine (10). In order to express WHV L protein, the relevant sequence from pUC119CMVWHV (33) was transferred to vector (also named) pSVL (Pharmacia) to create pSG322. This construct was not mutated to change the initiation codons for pre-S2 and S. However, L was driven from a strong simian virus 40 late promoter, and consistent with the ineffective expression of M (middle envelope protein) and S proteins, we observed that this construct by itself was insufficient to achieve detectable assembly of HDV RNA-containing particles (see Fig. 7A1 and A2). Presumably the WHV promoter for the M and S mRNA was much less efficient in the Huh7 cells that the simian virus 40 late promoter driving expression of the L mRNA.
Primary human hepatocytes (PHH) plated on rat tail collagen (type I) in a 48-well format were obtained commercially (CellzDirect, Lonza, or BD Gentest) and infected the day after arrival. PWH were prepared from the liver of a WHV-positive woodchuck and cryo-preserved (1). This prior infection is not a concern since in a previous study we showed that wHDV infection was equally efficient in hepatocytes taken from either naive or WHV-infected animals (39). After thawing, PWH were resuspended in prewarmed complete hepatocyte medium (CellzDirect), seeded on 48-well collagen-coated plates (BD BioScience) at 2 × 105 cells/well, and infected the next day.
The features and preparation of the six peptide inhibitors tested in the experiment shown in Fig. Fig.66 are summarized in Table Table1.1. The immunoadhesins (5) bearing either WHV pre-S1 or the whole pre-S were constructed using the WHV sequence of plasmid pUC119CMVWHV.
Primary hepatocyte infections were performed as described previously for PHH (16). Unless otherwise stated, these were performed at a multiplicity of infection (MOI) of about 50 HDV genome equivalents (GE) per cell in the presence of 5% polyethylene glycol (PEG). In order to examine the ability of peptides and immunoadhesins to block infection, these potential inhibitors were present along with the virus throughout the incubation with hepatocytes, that is, for 6 to 16 h.
HDV RNA titers were determined by quantitative real-time PCR (qPCR), as previously described (16). For hepatocyte infections, total RNA was extracted at 6 days after infection and assayed by qPCR to determine the number of HDV RNA GE per average cell (5, 16).
As an alternative assay of hepatocyte infections, at day 6 cells were fixed, permeabilized, and immunostained for δAg, as previously described (16). In addition, PHH were counterstained with anti-albumin antibody (16), and PWH were counterstained with anti-α-tubulin (a gift from Elena Pugacheva).
It is known that HDV can be passaged in woodchucks infected with WHV to produce wHDV, that is, HDV with a WHV envelope (26, 27). Both WHV and HBV encode three co-C-terminal envelope proteins (35). Figure Figure1A1A shows the sequence alignment of the L proteins of HBV and WHV. The two sequences share very little similarity in most of the pre-S1 domain. However, beginning at the C terminus of pre-S1, sequence homology increases significantly, and clearly the majority of identical amino acids are located in the S domains.
For both envelope proteins TMHMM2.0 software (19) was used to predict folding, including transmembrane domains (TMD) (Fig. (Fig.2).2). It was assumed that 22 ± 2 amino acids constitute the average size for each TMD (2). Also taken into consideration was that the high level of sequence conservation between S domains of HBV and WHV will lead to significant structural-functional conservations between these two viral envelopes. The modeling took account of the fact that TMD III does not possess clear signals for breaking the transmembrane helix and that the well-conserved patch of amino acids IWM(M/I)W(Y/F)W should be exposed in the cytosolic compartment, consistent with its critical role in interacting with cytoplasmic HDV RNP (36). In the structures shown here, the entire pre-S1 domain is exposed on the luminal side of the endoplasmic reticulum or the outer side of the viral membrane. However, previous studies of the HBV envelope have shown that, for a fraction of the envelope proteins, an alternative conformation can exist with this region located inside the virion (36).
In summary, WHV and HBV L proteins are considered to share the following characteristic elements: (i) four TMD, I to IV; (ii) two cytosolic loops, a large one located between TMD I and II and a small one between TMD III and TMD IV; (iii) an N-terminal pre-S region on the outer side; and (iv) an external large loop between TMD II and TMD III. The structures presented in Fig. Fig.22 are in a good correlation with experimental data placing a receptor-interacting site(s) for HBV near the N terminus of pre-S1 (3, 14) and an HDV RNP-binding region in the small cytosolic loop (18).
Next, we compared the ability of wHDV obtained from the serum of an infected woodchuck to infect PHH and PWH. Infections were performed in the presence of 5% PEG, a strategy known to enhance HBV and HDV infections of PHH (3, 15, 16). At 6 days after infection, HDV replication was assayed by immunostaining to detect hepatocytes positive for δAg. As shown in Fig. Fig.3A,3A, both PWH and PHH were infected with wHDV. hHDV infected PHH but not PWH, which is consistent with a previous report (39). In the three situations where infection was achieved, the subcellular distributions of δAg were similar, typically with nucleoplasmic localization and sometimes with a distribution throughout the cell.
Quantitation of such infections performed at the same MOIs showed that wHDV infected 1.5 and 2.4% of PHH and PWH, respectively; that is, the infection levels were comparable. In contrast, hHDV infected 5.3% of PHH but <0.0003% of PWH. The fraction of PHH infected with wHDV increased with the MOI and reached 14% when 5,000 GE of wHDV per hepatocyte was used. This is comparable to a report that HBV infection of primary tupaia hepatocytes at an MOI of 10,000 resulted in infection in 25% of cells (12).
As an independent assay of infection, we used real-time qPCR, as summarized in Fig. Fig.3B.3B. These data confirmed the immunostaining results. Furthermore, given the increased sensitivity of the qPCR, we also quantitated infections carried out in the absence of 5% PEG. Again, wHDV was able to infect both PHH and PWH, while hHDV infected PHH but not PWH. (Similarly to HBV and hHDV, the infectivity of wHDV was enhanced in the presence of 5% of PEG.)
Several other isolates of wHDV obtained from different woodchucks were tested and found to be infectious for both PWH and PHH. Also, HDV RNA genome replication was confirmed in both kinds of wHDV-infected hepatocytes by Northern analysis (data not shown).
Next, we examined the effect of different MOIs using qPCR, and the results are summarized in Fig. Fig.4.4. In a range of MOIs up to 1,000 GE/cell, wHDV readily infected PWH, while the same cells were resistant to hHDV. However, for PHH we observed comparable efficiencies of infection with both viruses.
As another parameter of infection, we performed a time course analysis, again using qPCR, and the results are shown in Fig. Fig.5.5. PWH or PHH were infected with hHDV or wHDV at an MOI of 500. For PHH infections both viruses revealed time-dependent accumulation of HDV RNA GE/cell in amounts, at late times, greatly exceeding input MOI, clearly demonstrating efficient HDV RNA replication (Fig. 5A and B). wHDV productively infected PWH (Fig. (Fig.5C).5C). With hHDV there was signal detectable on PWH cultures at the earliest times, but it decreased progressively (Fig. (Fig.5D),5D), consistent with the interpretation that no significant replication took place.
The above studies using different assay procedures demonstrated that wHDV can infect both PWH and PHH while hHDV infects PHH but not PWH. While it might be agreed that these findings are consistent with the idea that the infection is largely controlled by the presentation of envelope proteins to receptor(s) at the cell surface, the results do not address how this presentation can differ and whether or not a different host receptor(s) is involved.
As an approach to understanding wHDV attachment and entry into primary hepatocytes, we made use of a panel of six peptides bearing pre-S sequences, with three derived from HBV L and three from WHV L, as summarized in Table Table1.1. The rationale was that these peptides could potentially compete with the virus for binding of the receptor(s) on the hepatocyte surfaces and thus block the infection. Note that two of the peptides were chemically synthesized and then myristoylated. The other four were created as immunoadhesins (5). These potential inhibitors were tested at a concentration of 50 nM present during the time cells were exposed to virus, and the results are summarized in Fig. Fig.6.6. Previous studies have shown that at 50 nM, the three HBV peptides inhibit infection of PHH by HBV and hHDV (5, 9, 14). Consistent with this, peptides 1 to 3 inhibited hHDV infection of PHH (Fig. (Fig.6B,6B, lanes 1 to 3). However, under the same conditions these peptides had little effect on wHDV infections of PHH (Fig. (Fig.6A)6A) or PWH (Fig. (Fig.6C).6C). Next, we tested WHV peptides 4 to 6. Of these, only the synthetic peptide (Fig. (Fig.6,6, lanes 4), inhibited infection by wHDV of PHH (panel A) and PWH (panel C) but had no effect on infection by hHDV of PHH (panel B). The other two WHV peptides (lanes 5 and 6) that were presented as immunoadhesins failed to inhibit any of the infections. One possible reason for this is that the sequences added to the N terminus could no longer interact with the receptor because of intramolecular folding. Such a phenomenon has been reported for synthetic HBV peptides that tend to loose their inhibitory potential after exceeding a certain length (14).
Overall, these studies support the interpretation that wHDV, like hHDV, needs pre-S1 sequences to achieve infection. More importantly, and consistent with the fact that the WHV and HBV pre-S1 regions share very little sequence homology (Fig. (Fig.1),1), we interpret these inhibition studies as evidence that wHDV and hHDV interact with PHH via different receptors. Furthermore, the receptor used by wHDV on PHH might be closely related to that used on PWH.
In all of the above studies the sources of wHDV and hHDV differed not only in terms of the envelope proteins used but also in the way the particles were assembled. wHDV was obtained from infected woodchucks, and hHDV was assembled, as previously described, using cells transfected to express HBV L, M, and S envelope proteins and also replicating HDV RNA (16). Therefore, it was important to prepare wHDV by a transfection strategy similar to that used for the hHDV. To do this, we expressed in Huh7 cells the WHV L and S envelope proteins in various combinations, along with a plasmid to initiate HDV genome replication. As summarized in Fig. 7A1, HDV RNA-containing particles were released with WHV S in the absence of WHV L, and as the percentage of WHV L plasmid transfected increased to 100%, the amount of released particles dropped to undetectable levels. In this respect these results were similar to studies of assembly using HBV L and S (16). Next, we tested aliquots of medium collected from the transfected cells for the ability to infect PHH and PWH. After 6 days, the total cell RNAs were extracted, and HDV replication was quantitated by qPCR. The number of GE produced per average cell was normalized relative to the input MOI, in GE/per average cell, to determine what we refer to as the specific infectivity of the virus on the susceptible cells. The specific infectivities of the assembled wHDV on PHH and PWH are shown in Fig. 7B1 and C1, respectively. Note that the particles assembled with WHV S alone were not infectious, consistent with the interpretation that pre-S regions are needed for infectivity. Also, with both cell types the specific infectivities demonstrated a peak at the same percentage of WHV L, and these peak values were not significantly different.
These results not only demonstrate that infectious wHDV can be assembled in transfected cells but also confirm that such virus can infect both PHH and PWH. That is, the results obtained with wHDV assembled in animals were extended to wHDV assembled from transfected cells.
We next extended the study to look for the assembly of particles with intermolecular combinations of WHV and HBV envelope proteins. The aim was to determine whether such particles could be assembled and, if so, whether they would be infectious on PHH and/or PWH.
First, we considered combinations of WHV L with HBV S and of HBV L with WHV S and tested for the assembly of HDV RNA-containing particles. As shown in Fig. 7A2 and A3, respectively, assembly took place, and as before (Fig. 7A1), when the proportion of the L protein was increased to 100%, the assembly dropped to undetectable levels.
Next, we determined the specific infectivities for these particles on PHH and PWH. Virus assembled with WHV L plus HBV S infected both cell types, and the peak of specific infectivity was at about the same percentage of WHV L (Fig. 7B2 and C2). Clearly, the presence of HBV S in the particles did not interfere with the ability to infect PWH (Fig. 7C2), and, if anything, it enhanced the ability to infect PHH (Fig. 7B2). Virus with combinations of HBV L with WHV S infected PHH (Fig. 7B3) but gave no detectable infectivity on PWH (Fig. 7C3). The presence of WHV S in the particles was not sufficient for infection of PWH.
In summary, the use of combinations of WHV and HBV envelope proteins did not interfere with the assembly of HDV RNA-containing particles. Further, these particles could be infectious on primary hepatocytes. And in all cases it was the origin of the L protein, whether WHV or HBV L, that determined the specificity of infection on PWH and PHH. Furthermore, we can deduce that it was the origin of the pre-S sequences and not that of the S protein that controlled the infectivity and the species specificity. Incidentally, it could be noted that in most cases increasing the relative amount of L in an HDV particle initially achieved greater infectivity but then led to suppression (Fig. 7B1, B3, C1, and C2).
A prior study reported that wHDV could infect PHH (4). We have confirmed and extended this result, using immunostaining (Fig. (Fig.3A)3A) and RNA analyses by Northern blotting and real-time qPCR (Fig. (Fig.33 to to7).7). We used wHDV isolates from infected animals and also wHDV assembled in vitro from transfected cultured cells. Different sources of wHDV did have different titers and probably had different ratios of virions to subviral particles and also different relative amounts of the L/M/S envelope proteins. Except for the study shown in Fig. Fig.7,7, we did not attempt to control or determine the relative amounts of L/M/S in the infectious particles. However (again, except for Fig. Fig.7),7), the sources of wHDV and hHDV that we used were able to infect PWH and PHH, respectively, with comparable efficiencies.
Independent of potential variations between the sources of virus, we observed that for a given source of wHDV, the infections of PHH and PWH were of comparable extent. In contrast, the sources of hHDV infected PHH but gave no detectable infection of PWH.
There is no reported evidence for the productive infection of PHH by WHV, and this might be considered to be due to a block at attachment and entry. However, the results presented here for infection by wHDV favor the possibility that WHV can enter PHH but is blocked at some postentry step. For example, WHV enhancers and promoters may not function correctly in PHH (8).
The virtual inability of hHDV to infect PWH might seem to be in contradiction to the in vivo observation that hHDV can be transmitted to a woodchuck in the presence of WHV (26). However, in our studies we are detecting only a primary infection without subsequent spread. In contrast, for the in vivo studies, a rare infection event into a hepatocyte already infected with WHV can lead to the assembly and release of new HDV. This will be wHDV rather than hHDV, which will be able to amplify and spread throughout the susceptible hepatocytes of the woodchuck liver.
We have demonstrated that for wHDV, as for hHDV (13, 35), the ability to infect a susceptible cell depended upon sequences within the pre-S1 domain. Evidence for this was obtained using peptides related to the pre-S1 region of the envelope proteins of both WHV and HBV (Fig. (Fig.6).6). However, a comparison of the pre-S1 sequences of WHV and HBV showed little sequence conservation (Fig. (Fig.1).1). And while sequences from the pre-S1 of HBV could block infection of PHH by hHDV, they did not block the infection by wHDV.
Furthermore, we have exploited the simpler assembly requirements of HDV relative to hepadnaviruses (16, 36) to achieve the assembly of infectious HDV RNA-containing particles containing known combinations of WHV and HBV envelopes and thus demonstrated that the ability to infect PHH is provided by HBV or WHV L but not by HBV or WHV S (Fig. (Fig.77).
These findings lead us to suggest that hHDV and wHDV might use pre-S1 domains to recognize different receptors on the surface of PHH to achieve infection. Such an interpretation may be presumptive in that we still do not know the identity of the receptor(s) used by HBV and HDV for infection of PHH (13). Nevertheless, we trust that our studies will ultimately contribute to a more complete picture of how hHDV and wHDV, as well as HBV and WHV, attach to and enter susceptible hepatocytes.
J.M.T. was supported by grants AI-058269 and CA-06927 from the NIH and by an appropriation from the Commonwealth of Pennsylvania. N.C. was supported in part by the Elizabeth Knight Patterson Fellowship.
Constructive comments on the manuscript were given by Glenn Rall and Richard Katz. We acknowledge assistance from Emmanuelle Nicolas and the Biochemistry and Biotechnology Facility, Roland Dunbrack and the Molecular Modeling Facility, Sandra Jablonsky and the Cell Imaging Facility, and Carol Aldrich.
Published ahead of print on 21 May 2008.