Potential conflict of interest: Nothing to report.
Entry and intracellular transport of hepatitis B viruses have several unusual, largely unknown aspects. In this study, we explored the mode of virus entry using the duck hepatitis B virus (DHBV) and the primary hepatocyte infection model. Upon internalization, viral particles were enriched in an endosomal compartment, as revealed by biochemical and ultrastructural analysis. Virus-containing vesicles harbored early endosome markers. Kinetic analysis revealed time-dependent partial translocation of viral DNA from endosomes into the cytosol. This was strongly reduced by inhibition of vacuolar ATPase; (vATPase) activity with bafilomycin A1 and resulted in abortive infection and prevention of cccDNA formation. Inactivation of vATPase induced accumulation and stabilization of incoming viral particles in endosomes, presumably by blocking endosomal carrier vesicle–mediated cargo transport and sorting. Although neutralization of the endomembrane organelles alone led to stabilization of incoming viral particles, it did not inhibit virus infection. In line with this, a pH-dependent ectopic virus fusion at the plasma membrane could not be artificially induced. This provided further evidence for a pH-neutral translocation mechanism. Endosomal membrane potential was required for viral infection because cotreatment of cells with monensin partially overcame the inhibitory effect of bafilomycin A1. In conclusion, hepatitis B viral infection is mediated by a novel cellular entry mechanism with features different from that of all other known viruses. (HEPATOLOGY 2006;44:685–693.)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Enveloped animal viruses must overcome membrane barriers posed by host cells in order to deliver their genomes to their cellular replication site. Some viruses achieve this directly at the plasma membrane. For most viruses, fusion or penetration processes occur in the cell interior such as in endosomes, the endoplasmic reticulum (ER), or occasionally the trans-Golgi network (TGN). Entry steps exploited by the hepatitis B virus to enter host cells remain largely unknown. Several major issues remain unresolved, including receptor binding, mode of virus entry, membrane fusion, and capsid release as well as delivery of the viral genome to the nuclear replication site. Using DHBV and primary duck hepatocytes (PDHs), evidence has been provided: virus binding is species- and cell type–specific, the preS region of the large viral envelope protein governs virus entry, a small number of virions bind and enter host cells, viral uptake requires energy and proceeds with a remarkably slow kinetic, and virus trafficking requires dynamic microtubules.1–4
On binding to host cells, enveloped viruses are frequently endocytosed and transported in a coordinated and directional manner along the endocytic pathway.5 The transport and sorting processes involved are highly regulated and rely on adaptor proteins like Rab5 as well as facilitation by microtubules.6 Beyond early endosomes, transport is associated with an increasingly acidic environment that often triggers fusion of the viral envelope with endosomal membranes. Early endosomes are characterized by a mildly acidic environment, with a pH of approximately 6.2, whereas late endosomes have a lower pH, typically about 5.5.7 Endosomes are acidified by vacuolar proton-ATPase (vATPase).8 In addition to its numerous advantages such as regulated transport and protection from cytoplasmic enzymes, endocytic uptake also bears a certain risk for viruses in that transport to lysosomes and other cellular degradation centers can occur, which are a dead-end for most viruses.9
In the present study, we used the DHBV and PDHs to investigate the entry mechanism of hepatitis B viruses.
Primary Hepatocyte Cultures, Viruses, Antibodies, and Drugs.
Fetal PDHs were prepared and cultivated as previously described.10 PDHs were seeded in 6- or 12-well plates at a density of about 106 or 5 × 105 liver cells per well, respectively. PDHs were infected as previously described.10 If not otherwise indicated, infection was performed using 220 genome equivalents (GE)/cell. Infections with Herpes simplex virus (HSV) and Semliki Forest virus (SFV) were performed at MOI 3.
Bafilomycin A1 (BafA1) and concanamycin A were purchased from Calbiochem (Schwalbach, Germany) and Fluka (Buchs, Switzerland), respectively. Ammonium chloride (NH4Cl) was obtained from Merck (Darmstadt, Germany), reconstituted in water, and freshly prepared for each experiment. Monensin, nocodazole, wortmannin, and brefeldin A (BFA) were all from Sigma (Taufkirchen, Germany). For immunoblot and -fluorescence analysis, we used DHBV-preS- and core-specific antiserum and secondary antibodies as previously described.10 The following antisera were used: caveolin-1 (BD Transduction Laboratories), Rab5B (Santa Cruz Biotechnology), beta-actin (Sigma), Hsp70/Hsc70 (Stressgen Biotechnologies), early endosomal antigen 1 (EEA1; Affinity BioReagents). Antibody to vimentin, SFV core, and HSV were kindly provided by G. Tolstonog (Heinrich-Pette-Institute, Germany), J. Pavlovic (University of Zurich, Switzerland), and J. Kühn (University of Münster, Germany), respectively.
Infection Assays in Treated Cultures.
Cells were preincubated with the indicated concentrations of BafA1, 150 mmol/L NH4Cl, 50 μmol/L monensin, 5 μg/mL BFA, 5 μmol/L concanamycin A, or 300 nmol/L wortmannin for 1 hour, after which inoculum was added. Infection was allowed to proceed for about 16 hours in the presence of drugs. Subsequently, cells were washed with medium to remove unbound external virus as well as drugs. Cells were further cultivated in new medium for 3 days and harvested for analyses of viral L and core protein. cccDNA isolation and detection as well as immunoblotting and -fluorescence staining, and PCR was done as described.4
Attachment and Entry Assay.
This assay was performed as previously described.4
Low pH Treatment After Attachment.
To investigate if low pH treatment can induce viral fusion at the cell surface leading to infection, viral particles were attached to cultures. Cells were then washed and incubated with 1 mL of citrate buffer of a specified pH11 for 2 minutes. Finally, cultures were washed twice with PBS and warmed to 37°C to allow infection to proceed. Three days after treatment, the cultures were harvested. As controls, cells were first treated with the buffers, washed, and then infected.
To investigate uptake of DHBV into subcellular compartments, DHBV was attached to treated or untreated cultures in 6-well plates. Cells were then washed and either directly fractionated or trypsin-digested and then fractionated. Alternatively, cultures were shifted without washing to 37°C to allow viral entry for the indicated periods. These cultures were harvested with trypsin and fractionated.
Subcellular fractionation after DHBV inoculation was performed as previously described.12 Purity of the fractions was controlled by immunoblot analysis of aliquots for endosomal and cytoplasmic marker proteins as well as by electron microscopy. The first subcellular fraction contained the marker proteins caveolin-1 and Rab5, indicating enrichment of caveosomal and endosomal compartments, respectively (data not shown). The supernatant of the last ultracentrifugation step corresponded to the cytosolic fraction as shown by enrichment of the cytosolic proteins Hsc70 and vimentin as well as the exclusion of caveolin-1 (data not shown).
Immunoprecipitation of Endosomal Compartments.
Cultures in a 6-well plate were inoculated with viremic serum, followed by incubation for 2 hours at 4°C and then incubated for 6 hours at 37°C. The cells were harvested with trypsin and pelleted. The pellet was dissolved in a homogenization buffer (0.25 mol/L sucrose, 1 mmol/L EDTA, 10 mmol/L HEPES [pH 7.4] and protease inhibitor cocktail) and pottered 15 times. Half the homogenate was used for immunoprecipitation. For immunoblot, 2% input and 20% IP pellet were loaded.
Cell Viability Assays.
Cytotoxicity of the drugs was assayed as previously described.4
Results and Discussion
DHBV Infection of Primary Hepatocytes Does Not Rely on Acidic pH.
For the vast majority of enveloped viruses entering host cells by endocytosis, pH-dependent maturation and sorting of endocytic vesicles is essential. Consequently, pharmacological prevention of this by, for example, NH4Cl or monensin results in abortive infection by pH-dependent viruses such as SFV or influenza virus, but not by HSV fusing at a neutral pH. To test whether DHBV entry into hepatocytes is pH dependent, which would be suggestive of endocytic uptake mechanism, cells were treated with NH4Cl and monensin for 1 hour before inoculation. After 16 hours of incubation, medium containing drugs and virus was removed and replaced with medium without inhibitors. Inhibitors were omitted from the medium to study early viral entry steps. Three days later, cells were harvested and examined by immunoblot and -fluorescence staining for viral envelope protein L. Although both treatments resulted in an efficient increase in the endo-/lysosomal pH as judged by acridine orange counterstaining (data not shown), they did not reduce DHBV infection (Fig. 1A, data only shown for NH4Cl). The infection rate in treated cultures was comparable to that in untreated control cultures. The pH-dissipating effect of the drugs ammonium chloride and bafilomycin A1 in primary duck hepatocytes was additionally controlled by infection experiments using the prototype viruses HSV and SFV. As expected, both drugs efficiently reduced SFV infection, whereas they did not significantly influence HSV infection (Fig. 1A, lower panel).
To obtain additional evidence that DHBV infection is a pH-independent event, we tested whether fusion of prebound virions at the cell surface could be artificially induced by exposure to a low pH, as shown for low pH-dependent viruses.13 DHBV was allowed to bind to hepatocytes at 4°C and was then exposed to buffers in a pH ranging from 2.2 to 7 for 2 minutes. Successful infection occurred at a pH of 5, but exposure to a lower pH resulted in abortive infection (Fig. 1B). Thus, the pH sufficient for virus inactivation was higher than that normally encountered in endosomes. These results have several possible implications; after exposure to a low pH, one of the following occurred: DHBV nonproductively fused at the plasma membrane, DHBV fused but treatment resulted in irreversible rearrangements in virus particles required for a postentry step, or the DHBV viral particles simply did not fuse. As a control for adverse effects of the treatment on the cells, cultures were treated with the buffers, washed, and subsequently infected. This did not influence infection efficiency (data not shown).
Together, the data indicate that DHBV infection of PDHs does not strictly depend on an acidic pH step, a requirement that has been disputed in contradictory reports.1–3 The behavior of DHBV more resembles that of pH-neutral viruses. Unlike with pH-dependent viruses, exposure of prebound DHBV to a low pH did not induce productive ectopic fusion at the cell surface (Fig. 1B). A pH-neutral membrane fusion upon endocytic uptake is rather unusual for enveloped viruses. Usually fusion is either pH independent and occurs at the cytoplasmic membrane (HSV in certain cell types14) or is induced by acidification following endocytosis (influenza virus, vesicular stomatitis virus.15, 16). The slow internalization of DHBV appears to parallel more the fusion kinetics of the avian sarcoma/leukosis virus, which fuses at a neutral pH17 and contrasts with the rapid uptake and fusion of low pH-dependent viruses.
Incoming Viral Particles Reside in Cellular Subfractions Containing Endosomes.
A pH-independent membrane fusion that requires endocytosis is rather uncommon for enveloped viruses, and to our knowledge, there are only a few such examples, like Epstein-Barr virus or HSV into certain cell types.18, 19 Therefore, we decided to test whether DHBV entry into hepatocytes is mediated by endocytosis. If true, incoming viral particles should be found in endosomal subcellular fractions. However, if the virus directly fuses with the plasma membrane, then capsid-coated viral DNA should be found mainly in the cytosol. To discriminate between these two principal entry modes, we used a biochemical approach and subfractionated inoculated hepatocytes by differential ultracentrifugation. PDHs were incubated with DHBV at multiplicity of genome equivalents (MGE) 880 for 2 hours at 4°C in order to allow virus attachment. Virus entry was initiated by warming cultures to 37°C for the indicated durations. An aliquot of the subfractions was analyzed for the presence of viral markers biochemically and by electron microscopy. As a control, noninoculated cultures were used. Three hours after viral uptake into cells, viral DNA could be exclusively detected in the endosomal fraction (Fig. 2A, upper panel). Six hours after entry, the largest amount of viral DNA was still found in the endosomal fraction, but a faint signal could be detected in the cytosolic fraction. This fraction slightly increased 9 hours after viral entry. Analysis of the total cell-associated viral DNA showed that the signals were proportional to the fractions and increased with incubation time (Fig. 2A, lower panel). Viral envelope protein L was exclusively found in endosomal fractions, regardless of when analyzed (Fig. 2B). In contrast to viral DNA, the amount of L protein increased between 3 and 6 hours, before decreasing at 9 hours despite continuous inoculation. This indicates that L protein first accumulated before being further transported along the endocytic route, which probably led to its destabilization and degradation after viral entry. The viral DNA in contrast did not show degradation in this time frame.
Thin-section electron microscopy of endosomal fractions 9 hours postentry showed spherical vesicles representing different endosomal compartments. Within these vesicles were intact viral particles with a diameter of 40–50 nm (Fig. 2C, upper panel, arrows). Smaller structures without a visible envelope, presumably naked capsids, were also observed within the lumen. The surface of endosomes containing viral particles was rather smooth and appeared devoid of a clathrin coat. Readily identifiable organelles such as mitochondria, Golgi stacks, and plasma membrane sheets were absent or infrequently detected. The corresponding fractions of noninoculated PDHs contained similar vesicular structures but lacked structures resembling viral particles (Fig. 2C, lower panels).
We next determined the nature of virus-containing cellular vesicles biochemically using the immunoadsorption protocol described in the Experimental Procedures section. Early endocytic vesicles were enriched by immunocapture from homogenates of inoculated PDHs using an anti-Rab5B antibody. To stabilize the viral proteins, vATPase inhibitor concanamycin A and microtubule (MT)-depolymerizing agent nocodazole were used. Immunoblot and PCR analysis of the precipitates revealed enrichment of L protein and DNA, respectively (Fig. 3, DNA data not shown). In addition to Rab5, the isolated compartments also contained EEA1 (data not shown). Nonspecific binding of membranes to the beads was controlled by using rabbit antimouse serum for immunoprecipitation.
Both formation and sorting of endosomal carrier vehicles (ECVs) and multivesicular bodies are regulated by phosphatidylinositol-3-phosphate kinase (PtIns3PK) signaling.20 Pharmacological inactivation of PtIns3PK with wortmannin results in inhibition of membrane sorting and trafficking from late endosomes to lysosomes and TGN as well as recycling between compartments.21 To test the relevance of this late-endosomal trafficking for DHBV infection, PDHs were pretreated with 300 nmol/L wortmannin for 1 hour before the addition of virus for 16 hours in the presence of the drug. Immunoblot and -fluorescence analysis of cells 3 days later revealed no interference with DHBV infection (Supplementary Fig. 1A; Supplementary material for this article can be found on the Hepatology website (http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html). The concentration used was sufficient to induce endosomal vacuolation as reported (data not shown).23 From these experiments we concluded that DHBV entry into PDHs does not rely on a PtIns3PK-regulated late-endosomal trafficking step.
To test whether DHBV entry employs the Golgi apparatus, we treated cells with 5 μg/mL BFA, which disrupts the Golgi apparatus and interferes with the traffic of cargo between endosomes and lysosomes.22 The PDHs were pretreated for 1 hour prior to the addition of the virus, and cultures were incubated overnight. The following day, both virus and drug were removed, and the cultures were further incubated for 3 days before being analyzed for infection. The drug was removed because we only wanted to study the influence of Golgi disruption during virus entry. BFA treatment resulted in disruption of the Golgi compartment, as judged by the distribution of the Golgi marker protein beta-galactosyltransferase, and was reversible about 3 hours after removal of the drug (data not shown). Both immunoblot and -fluorescence studies showed no inhibitory effect of BFA on infection (Supplementary Fig. 1B). This indicates that the Golgi apparatus is not the site of viral translocation and that retrograde viral transport across the Golgi apparatus to the ER is also not required to initiate infection.
Taken together, the data indicate that DHBV is taken up in cells via an endocytic pathway involving early endosomes defined by Rab5B and further suggest that once fusion has occurred, capsids reside in the lumen of endosomal vesicles. This uptake route presumably leads to degradation of at least some of the viral surface proteins.
Infectious DHBV Entry in PDHs Requires vATPase Activity.
The unexpected findings that viral particles reside in endosomal compartments but that a low pH is dispensable for DHBV infection prompted us to investigate the possible role of acidification and trafficking of earlier compartments, those that precede late endosomes or lysosomes. We thus made use of bafilomycin A1 (BafA1) and concanamycin A, strong and very specific inhibitors of vATPase.23 Treatment of cells with BafA1 leads to neutralization of the luminal pH within about 20 minutes24 and interferes with formation and maturation of ECVs from early endosomes.25 This results in a transport block and stabilization of endocytosed material further along the endo-/lysosomal pathway.24
If DHBV needs an ECV-mediated endosomal exit from and acidification of early endosomes, then BafA1 treatment of hepatocytes should strongly reduce DHBV infection. To test this, we pretreated PDHs with BafA1 for 1 hour at different concentrations prior to infection and analyzed cultures 3 days later. Dose-response escalation, ranging from 60 to 600 nmol/L, revealed a dose-dependent reduction of infection in BafA1-treated cultures as determined by L-immunoblot analysis (data not shown). The strongest antiviral effect was seen at 600 nmol/L of BafA1. Consequently, in further experiments 600 nmol/L was used. The effect of concanamycin A, a compound closely related to BafA1, on DHBV infection was also tested and led to similar results but at an approximately 10-fold higher concentration (data not shown). Immunofluorescence staining for L protein showed a strong reduction in the number of infected cells in treated cultures compared to the number of infected controls (Fig. 4A). Consistent with these observations, cccDNA was barely detectable in treated cultures compared to in untreated controls (Fig. 4B). The strongly reduced efficiency of infection was not a result of diminished viral entry into PDHs, which could have been caused by disturbed receptor trafficking after treatment (Supplementary Fig. 2). These findings indicate that the infection block occurred before nuclear delivery of incoming viral DNA but after viral entry.
In summary, these results show that DHBV uptake is mediated by an endocytic pathway. Moreover, they show that DHBV depends on cellular vATPase activity after entry. Because vATPases are found in highly specialized endomembrane organelles including endosomes, lysosomes, and the Golgi apparatus, these observations indicate that DHBV is transported through intracellular membrane-bound compartments. If this trafficking pathway is disturbed, incoming viral DNA does not arrive in the nucleus and is therefore not converted to cccDNA, a prerequisite for infection.
Arrest of Endosomal Trafficking and Cargo Sorting Results in Stabilization of Incoming Viral Particles and Abortive Infection.
Assuming BafA1 can inhibit infection by preventing viral escape from early endosomes or ECVs, treatment should result in accumulation of viral particles in these compartments. To test this hypothesis, we pretreated PDHs with BafA1 prior to inoculation with DHBV. Cells were harvested and fractionated after different entry periods and analyzed by immunoblot for L protein and PCR. We found a strong accumulation of L protein in the endosomal fraction of BafA1-treated cells (Fig. 5A, upper panel). The amount of L strongly increased over the incubation period, whereas untreated cells did not show such an accumulation (Fig. 5A, lower panel). The bands in the lower part of the gel indicate proteolytic processing of L protein. Similarly, viral DNA was also found to accumulate over time in endosomal compartments when the vATPase was inhibited (Fig. 5B).
To determine whether BafA1-induced sequestration and stabilization of viral particles resulted from its interference with endosomal pH and/or its inhibition of membrane trafficking, we evaluated the impact of the arrest of endosomal transport on the stability of incoming DHBV particles. Transport and sorting of cargo from early endosomes requires microtubules. Consequently, disruption of microtubules with nocodazole results in accumulation of endocytosed tracers beyond early endosomes, but before late endosomes in ECVs.26 Nocodazole treatment prior to inoculation indeed resulted in accumulation and stabilization of endocytosed viral particles, albeit to a lesser extent than that seen with substances that dissipate endosomal pH (Supplementary Fig. 3). Thus, vesicle traffic along the endocytic pathway appears to be involved in DHBV entry, as already previously shown, because disruption of microtubules (MTs) results in abortive infection.4 Fluorescence imaging of the pH in cellular vesicles using the pH-sensitive probe acridine orange after MT depolymerization showed no difference in endosomal acidification compared to that in untreated cultures (data not shown).
Taken together, the data indicate that viral particles accumulate in endosomal compartments after endocytosis in the absence of functional vATPase because of disruption of proper trafficking. Because DHBV entry does not rely on a low pH; the data demonstrate that the inhibition of DHBV infection by BafA1 treatment is a result of prevention of certain endosomal transport/sorting steps and probably inactivation of the endosomal proton gradient rather than prevention of luminal acidification. One explanation for this observed phenomenon is that treatment with BafA1 results in inhibition of vesicle budding and a carrier vesicle-mediated transport block along the endocytic route without a concomitant block of incoming traffic.
Endosomal Membrane Potential Is Critical for Viral Infection.
Neutralization of vesicular pH by monensin or NH4Cl does not result in membrane depolarization, whereas treatment with BafA1 prevents creation of the membrane potential that is normally achieved by the proton pump. Because a low pH is not crucial for DHBV infection, we asked whether the endosomal transmembrane potential plays a critical role. vATPase generates not only a low vesicular pH but also an internal positive electrical potential across the membrane.27 This potential is lost after treatment with BafA1 because protons are no longer transported into the vesicle. To test whether vATPase-dependent transmembrane potential is required for DHBV infection, we treated cells with BafA1 alone and in combination with monensin. In the presence of the ionophore monensin and BafA1, K+ is exchanged for Na+ inside the vesicle, which reactivates the Na+/K+-ATPase and results in regeneration of a membrane potential even in the presence of BafA1.28 As shown in Fig. 6, combined treatment of cells with BafA1 and monensin resulted in partial rescue of DHBV infection compared to the results with cultures treated only with BafA1. However, this recovery was only partial because comparison of infection efficacy estimated by the amount of L protein was less than that in untreated cultures. Interestingly, PDHs that received a single treatment of monensin showed slightly more efficient infection compared to that in the untreated cells. Therefore, it appears that activation of Na+/K+-ATPase by monensin generates a membrane potential that can partially rescue proper viral translocation or sorting when the proton pump is blocked.
In summary, the data suggest that the transmembrane potential is at least partially necessary for the establishment of DHBV infection in PDHs.
It has been extremely difficult to characterize both the viral and cellular components mediating hepatocyte-virus interaction during entry of hepadnaviruses. Using DHBV and a combination of biochemical, ultrastructural, and pharmacological approaches, we have shown the first evidence of an entry mechanism with unusual and novel features. On basis of the presented data, we propose the following model summarized in Fig. 7. After binding to hepatocellular membranes, viral particles are taken up by endocytosis. Within the early endosome, a membrane potential-dependent but pH-independent translocation and subsequent sorting and transport of the virus from the endosome occurs, and infection is established. It remains unclear at which step of the endocytic membrane sorting, translocation and release of capsids occur. However, proteolytic processing of L protein and activation of a translocation motif (TLM) is a critical step during translocation because mutational impairment of the TLM results in strongly reduced infection most likely, because of a block of endosmal release.29
We thank J. Kühn and J. Pavlovic for kindly providing HSV as well as HSV-1 antibody and SFV as well as polyclonal C protein antibody, respectively; and E. Hildt for sharing the protocol for the subcellular fractionation. We are indebted to L. Cova for providing the DHBV core antiserum. We thank E. Grgacic for sharing unpublished data on the effect of bafilomycin on DHBV infection. We acknowledge the excellent technical support from N. Schmalstieg and C. Schneider.