Potential conflict of interest: Nothing to report.
This work was supported by research grant AI068769 from the National Institutes of Health (Bethesda, MD).
Our previous studies demonstrated that hepatitis C virus (HCV) envelope glycoproteins 1 and 2 (E1 and E2) display distinct reactivity to different cell-surface molecules. In this study, we characterized the interaction of E1 and E2 with apolipoproteins in facilitating virus entry. The results suggested a higher neutralization of vesicular stomatitis virus (VSV)/HCV E1-G pseudotype infectivity by antibodies to apolipoprotein E (ApoE) than apolipoprotein B (ApoB), with VSV/HCV E2-G pseudotype infectivity remaining largely unaffected. Neutralization of cell-culture–grown HCV infectivity by antiserum to ApoE and, to a lesser extent, by ApoB further verified their involvement in virus entry. HCV E1, but not E2, displayed binding with ApoE and ApoB by enzyme-linked immunosorbent assay. Binding of E1 with apolipoproteins were further supported by coimmunoprecipitation from human hepatocytes expressing E1. Rabbit antiserum to a selected E1 ectodomain-derived peptide displayed ∼50% neutralization of E1-G pseudotype infectivity. Furthermore, E1 ectodomain-derived synthetic peptides significantly inhibited the interaction of E1 with both the apolipoproteins. Investigation on the role of low-density lipoprotein receptor (LDL-R) as a hepatocyte surface receptor for virus entry suggested a significant reduction in E1-G pseudotype plaque numbers (∼70%) by inhibiting LDL-R ligand-binding activity using human proprotein convertase subtilisin/kexin type 9 and platelet factor-4, whereas they had a minimal inhibitory effect on the E2-G pseudotype. Conclusion: Together, the results suggested an association between HCV E1 and apolipoproteins, which may facilitate virus entry through LDL-R into mammalian cells. (HEPATOLOGY 2011;)
Hepatitis C virus (HCV) envelope glycoproteins 1 and 2 (E1 and E2) are anchored onto the envelope lipid bilayer of HCV and facilitate virus entry by interaction with host cell-surface molecules. We were the first to generate vesicular stomatitis virus (VSV) pseudotypes from HCV E1 and/or E2 chimeric gene constructs to study HCV entry using a single-cycle infection system.1-5 Subsequently, pseudotype derived from murine leukemia virus (MuLV) and human immunodeficiency virus systems with unmodified E1 and E26, 7 and baculovirus-derived HCV virus-like particles8 were generated and used as models to study the function of HCV envelope glycoproteins. These models suggested important contributions by the two HCV envelope glycoproteins in HCV entry mediated through interactions with sulfated polysaccharides, cluster of differentiation 81 (CD81), low-density lipoprotein receptor (LDL-R), and scavenger-receptor class B type 1 (SR-B1).3 These observations suggested that the two different forms of recombinant HCV envelope glycoproteins (chimeric E1-G/E2-G or unmodified E1-E2) display many similar functional profiles, and that more than one of these cellular proteins bind with the virus envelope glycoproteins. Claudin-1 has been shown to act late in the HCV entry process.9
Although HCV entry may occur through a hetero-oligomeric complex of the envelope glycoproteins, questions remain as to the specific role of E1 and/or E2 in this complex.4 The role of the individual HCV glycoproteins in VSV-derived pseudotype infectivity could be the result of recognition of distinct cell-surface receptors3, 4 and the presence of class I and/or class II fusion peptide motifs on each envelope protein.10-13 Recently, a new functional segment in the stem region of E2 (residues 703-715) has been suggested to play an active role in the reorganization of the envelope glycoproteins during the fusion process.14
HCV associates with very low-density lipoprotein (VLDL)-like or low-density lipoprotein (LDL)-like lipoproteins in human sera15 and circulate as virions packaged as lipoviroparticles.16 Both apolipoprotein B (ApoB) and apolipoprotein E (ApoE) were detected in low-density fractions of the HCV RNA-containing particles, and the virions can be precipitated by ApoB- and ApoE-specific antibodies.16 The production of ApoB containing VLDL in cell culture may help in HCV assembly and maturation, and the presence of ApoE seems to be more crucial in the production of infectious viral particles.17-20 As a result of the association between HCV and lipoproteins, LDL-R has been proposed as another potential entry factor for HCV.21-23
A detailed understanding of the mechanism of HCV entry is critical for preventive or therapeutic intervention. The present study was focused on defining the individual role of HCV envelope glycoproteins in LDL-R-mediated entry into hepatocytes. The results using the VSV/HCV pseudotype system suggested a preferential association of HCV E1 glycoprotein mediated by its N-terminal ectodomain with ApoE and ApoB. This association directs VSV/HCVE1-G pseudotype entry by LDL-R into mammalian cells.
Generation of VSV/HCV Pseudotypes and Plaque Assay.
VSV-derived pseudotypes were generated using the ectodomains of E1 and/or E2 glycoproteins from HCV genotype 1a (GenBank accession no.: M62321).1 A sulfated sialyl lipid (NMSO3) was used in the VSV/HCV pseudotype plaque assay to inhibit any potential residual uptake of parental G glycoprotein to the VSVts045 backbone.3
Antibodies to Apolipoproteins and Virus Neutralization.
Goat antiserum to human Apo B or Apo E (Calbiochem, La Jolla, CA), known to cross-react with many other species for considerable homology,24 was used for the neutralization of VSV-derived pseudotype or cell-culture–grown HCV.25
Antibodies to E1 Linear Epitopes.
Peptide (P1) with the sequence NH2–Cys-Ser-Ser-Ile–Val-Tyr-Glu-Ala-Ala-Asp-Met-Ile-Met-His-Thr-COOH representing an identified B cell epitope26 was synthesized (AnaSpec, Fremont, CA). Rabbit antiserum was generated to keyhole limpet hemocyanin conjugated at the N-terminus of the peptide (Cocalico Biologicals, Reamstown, PA) and purified by affinity chromatography against P1 peptide using a Sulfolink Immobilization Kit (Pierce Biotechnology, Rockford, IL).
Synthetic Peptides for ELISA.
Four other peptides, P2 (Gly-His-Arg-Met-Ala-Trp-Asp-Met-Met-Met-Asn-Trp-Ser-Pro), P3 (Val-Thr-Asn-Asp-Cys-Pro-Asn-Ser-Ser-Ile-Val-Tyr-Glu), P4 (Ile-Leu-His-Thr-Pro-Gly-Cys-Val-Pro-Cys-Val-Arg), and P5 (Ser-Arg-Cys-Trp-Val-Ala-Val-Thr-Pro-Thr), were selected from the hydrophilic region of E1 of the HCV genotype 1a sequence (accession no.: M62321). The peptides were synthesized (AnaSpec), dissolved in phosphate-buffered saline (PBS; pH 7.4), and used in enzyme-linked immunosorbent assay (ELISA). An E2 hypervariable region 1 (HVR1) peptide (Glu-Thr-His-Val-Thr-Gly-Gly-Ser-Ala-Gly-His-Thr-Val-Ser-Gly-Phe-Val-Ser-Leu-Leu-Ala-Pro-Gly-Ala-Lys-Gln-Asn) was used as a negative control in discriminating E1 binding specificity.
Highly purified recombinant HCV E1191-326 (Innogenetics Biologicals, Gent, Belgium)27 and E2384-715 (kindly provided by Michael Houghton; Chiron Corporation, Emeryville, CA) were examined for binding with ApoB and ApoE (Calbiochem). Briefly, an ELISA plate (Maxisorp; Nunc Corporation, Rochester, NY) was coated with ApoB or ApoE (100 ng/well) in PBS (pH 7.4) overnight at 4°C. The plate was blocked with 3% bovine serum albumin (BSA) in PBS. Purified E1 or E2 protein was added at serial dilutions in triplicate wells. After incubation overnight at 4°C, plates were washed, and bound protein was detected by the addition of anti-E1 or anti-E2 monoclonal antibody (AUSTRAL Biologicals, San Ramon, CA), followed by antimouse immunoglobulin G/horseradish peroxidase (IgG-HRP) and peroxidase substrate (Sigma-Aldrich, St. Louis, MO). Absorbance values were calculated after subtraction from BSA binding values used as the negative control. Results are presented as the mean, together with the standard deviations, from three independent experiments.
Four other peptides, P2, P3, P4, and P5, derived from HCV E1, were tested for binding to ApoB or ApoE by ELISA. For this, the peptides were biotinylated (Pierce), and the binding to immobilized ApoB or ApoE was quantitated by ELISA, as described above. Biotinylated E2 HVR1 peptide was included as an unrelated negative control. Inhibition of E1 binding to apolipoproteins by the same set of peptides was verified, following a similar procedure.
Fluorescence-Activated Cell-Sorting Analysis.
Cell-surface expression of LDL-R or CD81 on Huh-7 cells was quantified by flow cytometry (BD, Franklin Lakes, NJ) using a fluorochrome-tagged specific antibody (BD Pharmingen, San Diego, CA).3
Interaction between HCV envelope glycoproteins with apolipoproteins, without appearance of coprecipitating E2 with E1, was examined. For this, Huh-7 cells were infected separately with recombinant vaccinia virus (vvC-E1-E2-NS21-967 for detection of E1 or vvE2-NS2347-906 for detection of E2 only) at a multiplicity of infection of ∼2. After 48 hours, cells were lysed with TNTG buffer (30 mM of Tris [pH 8.0], 150 mM of NaCl, 0.5% Triton X-100, 10% glycerol, and a cocktail of protease inhibitors). Clarified cell lysates were mixed with goat antiserum to ApoB (Calbiochem), mouse monoclonal antibody (mAb) to E1 (AUSTRAL), ApoE (Santa Cruz Biotechnology, Santa Cruz, CA), or E2 (Chiron). The immunoprecipitates were immobilized on protein-G Sepharose beads, washed, and subjected to western blotting analysis for the detection of E1, E2, or ApoE using specific antibodies.
Proprotein Convertase Subtilisin/Kexin Type 9 Treatment.
Huh-7 cells were transfected with proprotein convertase subtilisin/kexin type 9 (PCSK9) plasmid DNA, its gain of function mutant, D374Y, or with vector control at varying concentrations using lipofectamine and incubated for 48 hours at 37°C. VSV/HCV E1-G and/or E2-G pseudotype was added to the cell monolayer for plaque assay.3
Platelet Factor-4 Treatment.
Huh-7 cells were rinsed with prechilled media and incubated at 4°C in the presence of varying concentrations of platelet factor-4 (PF-4; Hematologic Technologies, Essex Junction, VT) for 15 minutes before the addition of pseudotype virus. Plates were incubated for an additional 90 minutes at 4°C, rinsed with cold medium, and analyzed for plaque development.3
Association of HCV E1 Glycoprotein With Apolipoproteins.
We examined the role of the individual HCV envelope glycoproteins for interactions with the apolipoproteins. For this, VSV-derived pseudotypes reconstituted with chimeric E1-G and/or E2-G were utilized as surrogate models. Inhibition study for pseudotype entry was examined using goat antiserum to ApoB or ApoE (Fig. 1A,B). Antiserum to ApoB displayed a maximum of ∼55% reduction in E1-G/E2-G as well as E1-G pseudotype infectvity. On the other hand, <25% reduction of E2-G pseudotype and ∼5% reduction of VSVG control pseudotype infectivity was observed (Fig. 1A). Unrelated goat antiserum, when used as a negative control, did not display a significant inhibition of VSV/HCV or VSV-G pseudotype infection at similar dilutions. Antiserum to ApoE displayed a maximum of ∼80% reduction in E1-G/E2-G as well as E1-G infectivity (Fig. 1B). In contrast, the antiserum to ApoE displayed a much lower level (<30%) of infectivity of E2-G pseudotype or VSV-G control pseudotype. Neutralization of cell-culture–grown HCV by antiserum to apolipoproteins was also tested. Goat antiserum to ApoB or ApoE displayed ∼50%-70% reduction of cell-culture–grown HCV genotype 2a infectivity at a 1:20 dilution (Fig. 1C). Similar results were obtained with the cell-culture–grown HCV genotype 1a (data not shown). Neutralization of VSV/HCV E1-G pseudotype infectivity by antisera to apolipoprotein indicates an important role played by E1 glycoprotein in HCV infection.
Interaction of HCV E1 Glycoprotein With Apolipoproteins.
We also examined the potential for a direct interaction between HCV envelope glycoproteins and ApoB or ApoE. Binding studies by ELISA revealed an interaction of HCV E1 glycoprotein with both ApoB and ApoE (Fig. 2A,B). Interestingly, a very weak binding for E2 with both ApoB and ApoE were observed. Coimmunoprecipitation experiments were also performed from Huh-7 cells infected with a recombinant vaccinia virus expressing HCV proteins (VVC-E1-E2-NS21-967 or VVE2-NS2347-906) to determine an association of HCV E1 glycoprotein with ApoB or ApoE. Precipitation of ApoB and ApoE by specific antibody from cell lysates brought down HCV E1, as detected by western blotting using anti-E1 antibody (Fig. 2). In a reciprocal experiment, immunoprecipitation of cell lysates with E1 antibody coprecipitated ApoE. However, immunoprecipitation with antibody to ApoB/ApoE or E2 did not coprecipitate E2 and ApoE, respectively (figure not shown). Together, the results suggested an association of HCV E1 with both ApoB and ApoE.
Synthetic Peptides Derived From the E1 Ectodomain Bind With Apolipoproteins.
We examined the potential regions in E1 that may be responsible for interaction with apolipoproteins. Two hydrophobic regions, internal (H1) and C-terminal (H2), are localized in the E1 protein28 (Fig. 3A). For binding with apolipoproteins by ELISA (Fig. 2A,B), we used a C-terminally truncated variant of E11-326 (Innogenetics), lacking the H2 domain. To investigate the role of N-terminal ectodomain in the binding of E1 with apolipoproteins, three peptides (P2, P3, and P4) were selected from adjacent regions of the N-terminal ectodomain bearing high homology across HCV genotypes (Fig. 3A). A different peptide (P5) was selected from a highly conserved, centrally located sequence, GHRMAWDMMMNWSP, flanked between two hydrophobic domains (Fig. 3A).
For the ELISA-based quantitative detection of interaction with E1, the peptides were biotinylated at their N-terminus and the binding was measured based on detection by avidin-HRP conjugate. Biotinylated HVR1 peptide from HCV E2 (residues 384-411) was used as a negative control. The results suggested that the peptides, P2, P3, and P4, bound to both ApoB and ApoE (Fig. 3B,C). Peptide P5, designed from the internal conserved sequence of E1, showed very little binding ability with both the apolipoproteins. The results further indicated the involvement of the E1 N-terminal ectodomain in this interaction.
Antibodies were generated against the synthetic peptide, P1 (Fig. 3A), representing a linear epitope from the N-terminal ectodomain of E1.26 Neutralization of VSV/HCV E1-G pseudotype infectivity reached a maximal value of 50% with the antibody to peptide P1 (Fig. 4A). We did not observe a detectable neutralization of VSV/HCV E2-G pseudotype infectivity with the antibody to peptide P1, as expected. This observation suggests a possible involvement for the E1 ectodomain in mediating VSV/HCV E1-G pseudotype infectivity.
E1-Derived Peptides Inhibit Interaction With Apolipoproteins.
Next, peptides P2-P5 were used for ELISA-based inhibition of binding of E1 with the apolipoproteins. The results (Fig. 4B) indicated that the peptides, P2 and P4, exhibit ∼40% inhibition of interaction of E1 with ApoB, whereas peptide P3 exhibits ∼ 20% inhibition. As expected, the nonectodomain-derived peptide, P5, did not inhibit the binding of E1 with ApoB. Similar results were observed with ApoE by these peptides. The extent of inhibition was not improved by combining all three of the peptides in equimolar amounts.
PCSK9 Inhibits HCV E1 Pseudotype Infectivity.
PCSK9 is a proprotein convertase belonging to the proteinase K subfamily of the secretory subtilase family. It binds to the epidermal growth factor-like repeat A domain of LDL-R, inducing LDL-R degradation.29 We verified the down-regulation of LDL-R expression in Huh-7 cells by PCSK9 from fluorescence-activated cell-sorting (FACS) analysis (Fig. 5A); whereas no significant effect on the expression of HCV E2 envelope glycoprotein binding molecule, CD81 was observed (Fig. 5B).
We observed a 40%-70% reduction of VSV/HCV E1-G/E2-G pseudotype infectivity in the presence of PCSK9 or its gain-of-function mutant (Fig. 5C). Furthermore, a 50%-70% reduction in plaque numbers of pseudotype bearing the E1-G glycoprotein was observed (Fig. 5D). On the other hand, infection with pseudotype virus bearing E2-G did not lead to a detectable reduction in plaque numbers. Thus, HCV association with lipoproteins and the ability of LDL-R to act as an HCV E1-G pseudotype receptor may occur through the preferential association of E1 with apolipoproteins.
PF-4 is a chemokine released from activated platelets and binds to the ligand-binding domain of recombinant soluble LDL-R and partially inhibits the binding of LDL,30 disrupting the normal endocytic trafficking of the LDL-bound receptor. Varying amounts of PF-4 were added to Huh-7 cells at 4°C for 15 minutes before the addition of pseudotype virus for infection. The plaque assay mirrored the results obtained with PCSK9, with a 60%-80% reduction in plaque numbers when E1-G pseudotype virus was used to infect cells (Fig. 5E). A similar reduction was observed with pseudotype reconstituted with E1-G/E2-G. On the other hand, ∼5% reduction in plaque numbers was observed with VSV/HCV E2-G pseudotype.
We have used VSV/HCV pseudotypes expressing individual HCV envelope glycoproteins to dissect their functional interaction with apolipoproteins in mediating entry into hepatocytes. Earlier studies using pseudotypes expressing either E1-G or E2-G chimeric glycoprotein have suggested a distinct role for each glycoprotein in interaction with the host cell surface.3, 31 We observed the ability of antisera against both ApoB and ApoE to significantly neutralize E1-G pseudotype infectivity with little or no change in E2-G pseudotype infectivity. Much like HCV, VSV associates with VLDL, which leads to its endocytosis by LDL-R.21 However, there is no association of VSV-G with either of the apolipoproteins, as indicated by minimal inhibition of VSV infectivity by both antisera. This result raises the possibility of a preferential attachment of apolipoproteins to E1 glycoprotein. Indeed, our results suggested that only HCV E1 glycoprotein specifically interacts with ApoB and ApoE.
To determine the region of E1 responsible for interaction with apolipoproteins, we generated synthetic peptides, based on the homologous regions from the E1 ectodomain, and performed binding experiments with apolipoproteins by ELISA. We also included a peptide based on the highly conserved hydrophilic 14-amino-acid region located between the two hydrophobic domains. The peptides (P2, P3, and P4) derived from the ectodomain displayed appreciable binding with the apolipoproteins. These results indicate that the N-terminal ectodomain of HCV E1 is responsible for interaction with both apolipoproteins. To establish this observation, we employed ectodomain-derived peptides to inhibit the binding of E1 to apolipoproteins. We observed a significant inhibition by all three peptides, with a maximum of ∼40% exhibited by P2 and P4. However, there was no increase in extent of inhibition when the peptides were used in combination, suggesting that the apolipoprotein binding determinants present in the ectodomain may be distributed in these synthetic peptides and, therefore, need to be oriented in a proper conformation to exhibit maximal binding. Future site-directed mutagenesis studies may identify individual amino acids responsible for this protein-protein interaction. It is likely that this interaction may help in packaging HCV particles with VLDL to promote entry through LDL-R23 or SR-B1.32
The observation that only E1 is responsible for interaction with apolipoproteins signifies, for the first time, a distinct role for this envelope glycoprotein in HCV entry into target cells. VSV pseudotype viruses were employed for this specific purpose, as our several studies suggest that the properties of individual envelope glycoproteins could be studied as VSV/HCV E1-G or VSV/HCV E2-G pseudoparticles. Studies using these pseudotypes further revealed that inhibiting ligand binding activity of LDL-R by PCSK9 or PF-4 reduced E1-G pseudotype infectivity, whereas E2-G pseudotype infectivity was virtually unchanged. This implies that HCV E1-G pseudotype infectivity in hepatocytes is mediated by LDL-R. Interestingly, the interaction of E1 with apolipoproteins provides a possible insight into the LDL-R-mediated mechanism of HCV E1-G pseudotype entry. It is likely that an affinity of E1 for apolipoproteins directs the HCV E1-G pseudotype virus toward LDL-R, driven, of course, by the binding of apolipoproteins with LDL-R mediated by oppositely charged residues on the interacting proteins.33, 34 The fact that E2 lacks this affinity may explain that the HCV E2-G pseudotype infectivity is LDL-R independent. Interestingly, gain or loss of PCSK9 has been observed to correlate with liver cholesterol status.35 On the other hand, PF-4 status is associated with liver disease progression.36, 37 Further work should determine the structural insights of the E1-apolipoprotein interaction and the contribution of PCSK9 and PF-4 in LDL-R-mediated virus entry regulation and liver disease progression.
The authors are grateful to Jay Horton, Michael Houghton, Steven Foung, and George Luo for providing reagents, Ratna B. Ray and Sandip Bose for helping with data analysis and presentation, and Lin Cowick for preparation of the manuscript for this article.