Apolipoprotein E interacts with hepatitis C virus nonstructural protein 5A and determines assembly of infectious particles

Authors

  • Wagane J. A. Benga,

    1. Institut National de la Santé et de la Recherche Médicale (INSERM) Unite 748, Strasbourg, France
    2. Institut de Virologie, Université de Strasbourg, Strasbourg, France
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Sophie E. Krieger,

    1. Institut National de la Santé et de la Recherche Médicale (INSERM) Unite 748, Strasbourg, France
    2. Institut de Virologie, Université de Strasbourg, Strasbourg, France
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Maria Dimitrova,

    1. Institut National de la Santé et de la Recherche Médicale (INSERM) Unite 748, Strasbourg, France
    2. Institut de Virologie, Université de Strasbourg, Strasbourg, France
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Mirjam B. Zeisel,

    1. Institut National de la Santé et de la Recherche Médicale (INSERM) Unite 748, Strasbourg, France
    2. Institut de Virologie, Université de Strasbourg, Strasbourg, France
    Search for more papers by this author
  • Marie Parnot,

    1. Institut National de la Santé et de la Recherche Médicale (INSERM) Unite 748, Strasbourg, France
    2. Institut de Virologie, Université de Strasbourg, Strasbourg, France
    Search for more papers by this author
  • Joachim Lupberger,

    1. Institut National de la Santé et de la Recherche Médicale (INSERM) Unite 748, Strasbourg, France
    2. Institut de Virologie, Université de Strasbourg, Strasbourg, France
    Search for more papers by this author
  • Eberhard Hildt,

    1. Institut für Infektionsmedizin, Molekulare Medizinische Virologie, Kiel, Germany
    Search for more papers by this author
  • Guangxiang Luo,

    1. Department of Microbiology, Immunology and Molecular Genetics, University of Kentucky College of Medicine, Lexington, KY
    Search for more papers by this author
  • John McLauchlan,

    1. MRC Virology Unit, Glasgow, UK
    Search for more papers by this author
  • Thomas F. Baumert,

    Corresponding author
    1. Institut National de la Santé et de la Recherche Médicale (INSERM) Unite 748, Strasbourg, France
    2. Institut de Virologie, Université de Strasbourg, Strasbourg, France
    3. Pôle Hépato-digestif, Nouvel Hôpital Civil, Hôpitaux Universitaires de Strasbourg, Strasbourg, France
    • Inserm U748, 3 rue Koeberlé, F-67000 Strasbourg, France
    Search for more papers by this author
    • fax: (33)-3-68-85-37-50.

  • Catherine Schuster

    Corresponding author
    1. Institut National de la Santé et de la Recherche Médicale (INSERM) Unite 748, Strasbourg, France
    2. Institut de Virologie, Université de Strasbourg, Strasbourg, France
    • Inserm U748, 3 rue Koeberlé, F-67000 Strasbourg, France
    Search for more papers by this author
    • fax: (33)-3-68-85-37-24.


  • Potential conflict of interest: Nothing to report.

Abstract

Chronic hepatitis C virus (HCV) infection is a major cause of liver disease worldwide. Restriction of HCV infection to human hepatocytes suggests that liver-specific host factors play a role in the viral life cycle. Using a yeast-two-hybrid system, we identified apolipoprotein E (apoE) as a liver-derived host factor specifically interacting with HCV nonstructural protein 5A (NS5A) but not with other viral proteins. The relevance of apoE–NS5A interaction for viral infection was confirmed by co-immunoprecipitation and co-localization studies of apoE and NS5A in an infectious HCV cell culture model system. Silencing apoE expression resulted in marked inhibition of infectious particle production without affecting viral entry and replication. Analysis of particle production in liver-derived cells with silenced apoE expression showed impairment of infectious particle assembly and release. The functional relevance of the apoE–NS5A interaction for production of viral particles was supported by loss or decrease of apoE–NS5A binding in assembly-defective viral mutants. Conclusion: These results suggest that recruitment of apoE by NS5A is important for viral assembly and release of infectious viral particles. These findings have important implications for understanding the HCV life cycle and the development of novel antiviral strategies targeting HCV–lipoprotein interaction. (HEPATOLOGY 2010)

Hepatitis C virus (HCV) is a major cause of liver disease, including liver cirrhosis and hepatocellular carcinoma.1 Current treatment by interferon-alpha and ribavirin is limited by resistance, toxicity, and high costs.1, 2 Novel treatment approaches are therefore urgently needed. HCV is an enveloped single-stranded RNA virus of positive polarity that is a member of the genus Hepacivirus within the family Flaviviridae.3, 4 The HCV RNA genome encodes a unique polyprotein of approximately 3000 amino acids and is flanked at its 5′ and 3′ ends by two highly conserved untranslated regions involved in the translation and replication processes of the virus, respectively. The virus enters the cell through interaction of the viral glycoproteins with cellular co-factors.3, 4 After viral entry, viral translation and replication occurs in a cell compartment termed the “membranous web,” which is followed by viral assembly and particle egress.3, 4

An important feature of the hepatocyte is its key role in lipid metabolism. Increasing evidence suggests that the HCV life cycle and hepatocyte lipid metabolism pathways are closely linked. Indeed, HCV replication takes place, as noted previously, at specialized rearranged intracellular membranes termed “membranous web.”4 Membrane vesicles containing the HCV replication complex have been shown to be highly enriched in proteins required for very-low-density lipoprotein (VLDL) assembly, including apolipoprotein B (apoB), apolipoprotein E (apoE), and microsomal triglyceride transfer protein.5 Furthermore, the VLDL pathway has been shown to play a role in the assembly and maturation of infectious viral particles.5, 6 Moreover, a hepatocyte organelle storing lipids—the lipid droplet—has been shown to be important for the production of infectious virus particles.7, 8 Miyanari and colleagues8 demonstrated that the viral capsid protein core recruits viral nonstructural proteins and replication complexes to lipid droplet–associated membranes, and that this recruitment is critical for producing infectious viruses. Furthermore, HCV core protein has been shown to induce lipid droplet redistribution,9 and different HCV genotype core sequences have been shown to induce morphological changes in intracellular lipid droplets.10 Finally, Chang et al.11 have shown that HCV infectious virions purified from low-density fractions of cell culture supernatants are assembled as apoE-enriched lipoprotein particles and can be specifically precipitated by anti-apoE and anti-E2 monoclonal antibodies.11 These findings indicate that apoE is required for HCV virion infectivity and production. However, how apoE is recruited to the infectious particle and which viral factors are implicated in that interaction process remain unclear.

Recent evidence suggests that the viral nonstructural protein 5A (NS5A) plays an important role in HCV virion production. NS5A was found to be recruited by core-associated lipid droplets in replicating Huh7.5 cells for production of infectious particles.8, 12 The introduction of specific mutations into the NS5A showed that mutations in NS5A C-terminal domain III abolished core-NS5A colocalization in the HCV-replicating cells and hampered virion production.8, 13 These data identify NS5A as a viral factor for assembly of infectious viral particles.8, 13 However, the functional link between NS5A and virus production is unknown.

Abbreviations

apoE, apolipoprotein E; HCV, hepatitis C virus; HCVcc, cell culture-derived HCV; IgG, immunoglobulin G; MOI, multiplicity of infection; NS5A, nonstructural protein 5A; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; siApoE, small interfering apolipoprotein ERNA; siRNA, small interfering RNA; VLDL, very-low-density lipoprotein; TCID50, 50% tissue culture infectious dose.

Materials and Methods

Plasmids.

Yeast two-hybrid system expression plasmids have been described.14 The full-length apoE complementary DNA cloned in pOTB7 vector was purchased from Open Biosystems and contains an apoE consensus sequence. After digestion of the plasmid by EcoRI-XhoI apoE, complementary DNA was inserted in the EcoRI-XhoI sites of a yeast plasmid expressing the activation domain of Gal4 (pGADT7). Plasmids pFK-Jc1 (Jc1), pFK-Luc-Jc1 (Luc-Jc1) and J6/JFH1 constructs have been described.15–19 JFH1 or H77 NS5A encoding regions were polymerase chain reaction (PCR) amplified using Jc1 or p90/HCV FL-long pU20 as templates and inserted in the pGBKT7 EcoRI-SalI sites.14 Alanine triplet substitutions at positions 99 to 101 and 102 to 104 in the JFH1 NS5A coding region as described by Miyanari et al.8 and “domain III” deletion mutant described by Appel et al.13 were obtained as described recently.21

Yeast Strains.

The pGADT7 plasmid expressing a fusion between the activation domain of Gal4 and ApoE and the pGBKT7 plasmids expressing fusions between the DNA binding domain of Gal4 and the HCV nonstructural proteins were transformed in the S. cerevisiae strain AH109 (MATa, trp1-901, leu2-3, 112, ura3-52, his3-200, gal4Δ, gal80Δ, LYS2:GAL1UAS-GAL1TATA-HIS3, GAL2UAS-GAL2TATA-ADE2, URA3:MEL1UAS-MEL1TATA-lacZ).

Antibodies.

Monoclonal mouse anti-NS5A was obtained from Virostat. Rabbit polyclonal anti-NS5A,22 and monoclonal anti-apoE (Ab23, Ab33) have been described,11 polyclonal rabbit anti-apoE were from Epitomics and Abgent, and polyclonal rabbit anti-apoA1 was from Santa Cruz.

Yeast Two-Hybrid Assays.

Yeast two-hybrid assays were performed as described by Dimitrova et al.14 The pGBKT7-derived and pGADT7-derived constructs encoding apoE or HCV nonstructural proteins were co-transformed into AH109 yeast cells. Co-transformants were grown for approximately 24 hours in –LEU – TRP medium and then spotted onto – LEU –TRP culture plates to select for co-transformants and onto –TRP – LEU – HIS - ADE culture plates to allow for selection of interactants.

RNA Interference Assay.

Commercially available small interfering RNAs (siRNA) pools targeting apoE and apoA1 as well as control nontargeting siRNAs were purchased from Dharmacon and transfected into Huh7.5.1 cells using DharmaFect solution following the manufacturer's protocol. Silencing of protein expression was assessed by immunoblotting19 for apoE or immunofluorescence14 for apoA1.

Production of HCV Pseudotype Particles and Infection of Huh7.5.1 Cells.

Production and infection of Huh7.5.1 cells using murine leukemia virus MLV-based HCV pseudotype particles (HCVpp) (strain H77) were performed as described.19, 23, 24

Analysis of HCV Infection and Replication.

For the study of the viral life cycle in cells with silenced apoE expression, siRNAs were either co-electroporated with HCV Luc-Jc1 RNA16, 19 (protocol A) or transfected 3 days before electroporation of HCV RNA (protocol B). HCV replication was analyzed by luciferase activity as described.15, 19 For infection assays, Huh7.5.1 cells were incubated with Luc-Jc1 cell culture-derived HCV (HCVcc) (50% tissue culture infectious dose (TCID50) of 103, 104, and 105/mL, corresponding to a multiplicity of infection [MOI] of 0.1, 1, and 10). Forty-eight hours later, infection was analyzed by quantitation of protein expression using luciferase activity.15, 16, 19

Purification of Intracellular or Extracellular HCV Virions Using Sucrose Gradient Ultracentrifugation.

HCV virions were partially purified by pelleting lysates or culture supernatants through a sucrose cushion (20% sucrose in TN buffer) using a SW55Ti rotor (100 000 g for 4 hours at 4°C) and a Beckman L8-80M preparative ultracentrifuge. In addition, HCV RNA of cell lysates and supernatants were purified on iodixanol gradients as described by Gastaminza et al.25 Viral RNA content was determined by quantitative reverse transcription PCR, and infectivity was assessed by incubation of pelleted HCV or iodixanol gradient fractions with naïve Huh7.5.1 cells and subsequent detection of infection (72 hours after incubation) by luciferase activity.

Co-immunoprecipitation of apoE and NS5A in HCV Replicating Cells.

Co-immunoprecipitation experiments were performed in Huh7.5.1 cells containing replicating Jc1 as described,14 using monoclonal anti-NS5A (Virostat) or an anti-c-myc-tag isotype control antibody (Sigma). Immunoprecipitated proteins were subjected to immunoblotting using monoclonal anti-apoE.11 In a second approach, co-immunoprecipitation was performed in lysates of Huh7.5.1 cells transfected with HCV J6/JFH1 RNA17 or Huh9-13 cells containing the HCV Con-1 replicon,26 using rabbit anti-NS5A,22 anti-apoE (Abgent), or an anti-HBc control antibody27 and a previously described co-immunoprecipitation protocol.28 Immunoprecipitated proteins were analyzed by immunoblotting using peroxidase-conjugated apoE and anti-NS5A (Peroxidase conjugation kit, Pierce).

Co-localization of apoE and NS5A in HCV Replicating and Infected Cells.

Huh7.5.1 cells containingJc1 HCV RNA were fixed with 3% paraformaldehyde/phosphate-buffered saline (PBS) and permeabilized with 0.1% Triton X-100 in PBS. Proteins were stained using anti-apoE (Epitomics), anti-NS5A (Virostat), and Alexa Fluor 488-conjugated goat anti-rabbit immunoglobulin G (IgG) and Alexa Fluor 568-conjugated goat anti-mouse IgG (Molecular Probes) and a Zeiss Axiovert microscope (Zeiss). Immunofluorescence and confocal laser scanning microscopy of Huh7.5.1 cells infected with HCV JFH-1 or Jc1 was performed as described.29

Results

ApoE Interacts with the HCV NS5A Protein in a Yeast Two-Hybrid Assay.

To investigate apoE-HCV interactions, we first studied the interaction of apoE with HCV proteins using a yeast two-hybrid system. A yeast two-hybrid assay was constructed using HCV proteins fused to Gal4 DNA binding domain as bait and a full-length apoE fused to the Gal4 activation domain as a prey. After viral protein–apoE interaction, reporter gene His3 is activated, allowing selection on defined media (Fig. 1A). As shown in Fig. 1, only yeast co-transformed with apoE and HCV NS5A protein was able to grow on selective media (Fig. 1B). The specificity of this interaction for NS5A was confirmed by the absence of growth in co-transformation experiments using the HCV core and nonstructural proteins NS2, NS3, NS4A, NS4B, and NS5B and nonrelated nonviral proteins (Fig. 1B). Interestingly, an apoE interaction was not observed with C-terminally truncated envelope glycoproteins E1 (aa 170–311) and E2 (aa 371–661) (data not shown). The binding of apoE to NS5A appeared to be conserved among different genotypes because both NS5A of strain H77 as well as NS5A derived from strain JFH1 bound to apoE in the yeast two-hybrid system (Fig. 8A). The validity of the yeast two-hybrid system for the study of viral protein–protein interactions was further confirmed by core–NS5A interaction of both strains H77 and JFH1 (Fig. 1C), which has been shown to be required for viral assembly.8, 13 These results demonstrate that HCV NS5A interacts specifically with full-length apoE in a yeast two-hybrid assay.

Figure 1.

Interaction of apoE with HCV NS5A protein in a yeast two-hybrid assay. (A) Principle of the yeast two-hybrid screen: In this system, interaction between two proteins is indicated by the activation of the reporter gene HIS3, which allowed growth on histidine-depleted (HIS−) plates. (B) AH109 yeasts (matchmaker Clontech) were co-transformed with a plasmid encoding a fusion of the apoE coding region and the activation domain of Gal 4 and plasmids encoding the fusion of the DNA binding domain of Gal4 and HCV H77 strain proteins (core, NS2, NS3, NS4B, NS5A, and NS5B) or control protein laminin (Lam) or the Gal4 DNA activation domain only (pGBKT7). After 2 weeks at 30°C on selective His (−) media, yeast growth was analyzed. (C) AH109 yeasts (matchmaker Clontech) were co-transformed with a plasmid encoding a fusion of HCV core protein and the activation domain of Gal4 and plasmids encoding the fusion of DNA binding domain of Gal4 and NS5A from H77 or JFH1 strains or control protein Lam.

Figure 8.

ApoE-NS5A interaction is absent or decreased in viral mutants with a defect in viral assembly. AH109 yeasts (matchmaker Clontech) were co-transformed with a plasmid encoding fusion of apoE protein and the activation domain of Gal4 and plasmids encoding fusions of the DNA binding domain of Gal4 and wild-type JFH1 NS5A or mutant JFH1 NS5A containing (A) an alanine triplet substitution at position 99 (NS5AAAA99 JFH1) or 102 (NS5AAAA102 JFH1)8 or (B) a deletion in domain III of NS5A resulting in abolishment of particle production (“domain III deletion”).13 Lam, pGBKT7, pGAD expressing a fusion between Gal 4 DNA binding domain and control protein laminin, DNA binding domain of Gal4, and activation domain of Gal4, respectively, were used as controls.

ApoE and NS5A Co-immunoprecipitate and Co-localize in Huh7.5.1 Cells Containing Replicating HCV.

To confirm binding of host factor apoE and viral protein NS5A during the HCV life cycle, we performed co-immunoprecipitation experiments in Huh7.5.1 cells containing replicating HCV. Anti-NS5A antibody specifically co-immunoprecipitated viral protein NS5A together with apoE (Fig. 2). Detection of apoE–NS5A binding was dependent on the concentration of anti-NS5A antibody, because low concentrations of anti-NS5A antibody did not allow detection of apoE–NS5A binding. The validity of the results was further confirmed by absence of immunoprecipitation of apoE using anti-NS5A antibody in non–HCV-replicating Huh7.5.1 cells (Fig. 2A). ApoE–NS5A interaction was further confirmed by reciprocal co-immunoprecipitation using anti-apoE (Fig. 2B, C). Furthermore, apoE–NS5A interaction appeared to be genotype-independent, because it was also easily observed for NS5A of the HCV genotype 1b Con-1 replicon (Fig. 2B, C). These results clearly demonstrate that HCV protein NS5A binds to host cell apoE in target cells with replicating HCV. The detection of apoE–NS5A binding by co-immunoprecipitation did not require overexpression of apoE or NS5A using complementary DNA expression constructs but was easily detectable with endogenous apoE present in physiological levels in hepatoma cells and NS5A in expression levels present during viral replication.

Figure 2.

Co-immunoprecipitation of apoE and HCV NS5A protein in HCV replicating Huh7.5.1 cells. (A) Full-length replication-competent HCV RNA derived from isolate HCV Luc-Jc1 was transfected into the hepatoma cell line Huh7.5.1. Seventy-two hours after transfection, Huh7.5.1 cells containing replicating HCV were lysed and subjected to immunoprecipitation using a monoclonal anti-NS5A or an unrelated control monoclonal anti-myc antibody as described in Materials and Methods. As a positive control (PC), apoE expression was analyzed in Huh7.5.1 lysates shown in lane 1. Proteins immunoprecipitated by anti-NS5A or anti-myc control antibody (“++” corresponds to an antibody concentration of 15 μg/mL, “+” to 7.5 μg/mL) were analyzed by immunoblot using anti-apoE antibody 33 (1/500) and horseradish peroxidase–conjugated anti-mouse secondary antibody. (B, C) Reciprocal co-immunoprecitation of NS5A of genotype 1 and 2 in HCV replicating cells. Huh7 cells containing replicating HCV Con1 (genotype 1b) or JFH1/J6 (genotype 2a) were lysed and subjected to immunoprecipitation using anti-NS5A, anti-apoE, or anti-HBc as an unrelated control antibody. Immunoprecipitated proteins were analyzed by immunoblot using peroxidase-conjugated anti-NS5A and anti-apoE antibodies. Peroxidase conjugation of anti-apoE and anti-NS5A antibodies in immunoblot analysis shown in panels (B) and (C) eliminated detection of cross-reacting light and heavy chains of the immunoprecipitating antibodies. PC, positive control.

To further confirm apoE–NS5A binding during viral infection, we performed co-localization studies in HCV replicating and infected cells. As shown in Fig. 3A, Huh7.5.1 expressed apoE at various levels, including cells with high-level expression and cells in which expression of apoE was virtually absent. Confirming the interaction of NS5A with apoE in living cells, NS5A partially co-localized with apoE in Huh7.5.1 cells containing replicating HCV (Fig. 3A). This partial co-localization between apoE and NS5A was also observed at distinct “dot”-like structures in HCV-infected cells (Fig. 3B).

Figure 3.

Co-localization of apoE with HCV NS5A in Huh7.5.1 cells containing replicating HCV. (A) Co-localization of apoE and NS5A in Jc1 replicating cells. Full-length replication-competent HCV RNA derived from isolate Jc1 was transfected into the hepatoma cell line Huh7.5.1 cells as described in Materials and Methods. Seventy-two hours later, cells were fixed and stained for apoE and NS5A expression using rabbit anti-apoE antibody and Alexa Fluor 488–conjugated goat anti-rabbit IgG (green fluorescence) and anti-NS5A antibody and Alexa Fluor 568-conjugated goat anti-mouse IgG (red fluorescence) as described in Materials and Methods. Nuclei were stained in blue (DAPI; left upper panel). Co-localization of stained proteins is shown in the right lower panel. (B) Confocal laser scanning microscopy of Huh7.5.1 cells infected with HCV JFH-1 (TCID50 104/mL). Infection and staining of apoE and NS5A was performed as described in Materials and Methods.

ApoE Is Required for Production of Infectious HCV Particles.

To investigate the role of apoE and the impact of apoE–NS5A interaction on the HCV life cycle, we silenced apoE expression in Huh7.5.1 cells using apoE-specific siRNAs (siApoE). As shown in immunoblot analyses, transfection of siApoE reproducibly and specifically silenced apoE expression (Fig. 4A). Using this protocol, we investigated whether silencing apoE had any impact on production of infectious viral particles. As shown in Fig. 4B, supernatants obtained from donor cells containing replicating HCV Jc1 with silenced apoE expression (siApoE) showed markedly reduced or total lack of infectivity. To confirm these results, we investigated the impact of apoE silencing in single-cycle infection experiments in a protocol similar to previous studies analyzing the impact of apoB on HCV production.6 Again, apoE silencing resulted in marked inhibition of infectivity of supernatants of infected cells (Fig. 4C). Inhibition of virus production was observed for different MOIs (0.1, 1, and 10), with the most pronounced inhibition occurring at the lowest MOI (Fig. 4C). These single-cycle experiments using infectious recombinant HCV at different MOIs confirm that apoE is required for production of infectious virions.

Figure 4.

Silencing of apoE expression in Huh7.5.1 cells results in inhibition of HCV particle production. (A) Silencing of apoE expression in Huh7.5.1 cells. Huh7.5.1 cells were transfected with siRNAs as described in Materials and Methods. Lysates of control naïve Huh7.5.1 (Mock), PBS, siApoE or control siRNA (siCTRL) transfected cells (siCTRL) were subjected to immunoblotting using rabbit anti-apoE and anti-beta-actin monoclonal antibodies and horseradish peroxidase–conjugated secondary antibodies. ApoE and beta-actin are indicated on the left, and molecular weight (MW) markers (kDa) are indicated on the right. (B) Inhibition of HCV particle production in Huh7.5.1 cells with silenced apoE expression. Huh7.5.1 cells were transfected with siRNAs and HCV Luc-Jc1 RNA. Cell culture supernatants of nontransfected cells (Mock), PBS, siCTRL, siApoE, siApoA1, and HCV Luc-Jc1 co-transfected Huh7.5.1 cells were concentrated 50-fold 72 hours after transfection of HCV Luc-Jc1 RNA. Concentrated supernatants (150 μL) were then used to infect 6 × 104 naive Huh7.5.1 cells. Infectivity of supernatants from mock, PBS, siCTRL, or siApoE treated cells was quantified by measuring of luciferase activity in Huh7.5.1 lysates 72 hours after infection (mean ± SD; n = 4). (C) Impact of apoE silencing on HCV production in single-cycle infection experiments. Huh7.5.1 cells were transfected with control (CTRL) siRNA or apoE siRNA. Transfected cells were then infected with recombinant Luc-Jc1 HCV at different multiplicities of infection (MOI) using virus stocks with TCID50 of 103/mL, 104/mL, and 105/mL, corresponding to MOIs of 0.1, 1, and 10, respectively. Seventy-two hours after infection, cell supernatants containing infectious virions were used to infect new naïve Huh7.5.1 cells. Infectivity of supernatants was assessed 48 hours later as described previously. The results are expressed as a percentage of infection of cells treated with CTRL siRNA. A representative experiment of two experiments performed in triplicate is shown.

A previous report has described an interaction of NS5A with apoA1 using co-immunoprecipitation experiments.30 As shown in Fig. 4B, silencing of apoA1 by a pool of validated siRNAs did not result in detectable decrease of virus production.

ApoE Is a Host Factor Required for a Late Stage in the Viral Life Cycle.

Next, we mapped the stage of the viral life cycle requiring apoE as a host cell factor. It is conceivable that apoE could be involved in the early infection stages such as entry or replication as well as late stages of the viral life cycle, such as assembly or release of viral particles. To exclude an effect of apoE silencing on viral entry, we infected Huh7.5.1 cells with HCVpp and HCVcc. As shown in Fig. 5, apoE silencing had no effect on HCVpp entry or HCV infection. In contrast, silencing the HCV entry factor CD81 resulted in marked inhibition of HCVpp entry (Fig. 5A) and HCVcc infection (Fig. 5B). These data demonstrate that host cell apoE is not essential for viral entry.

Figure 5.

Impact of apoE silencing on entry and replication of HCV Luc-Jc1. Huh7.5.1 cells were transfected with siRNA targeting apoE (siApoE) and CD81 (siCD81) expression or irrelevant control RNAs (siCTRL). Forty-eight hours later, Huh7.5.1 cells were infected with HCVpp (A) or HCVcc Luc-Jc1 (B). Seventy-two hours later, HCV entry was assessed by quantitation of luciferase activity (mean ± SD; n = 4). (C) Huh7.5.1 were co-electroporated with HCV Luc-Jc1 and PBS (dark gray curve), control siRNA (light gray curve), siApoE (black curve), or siRNA targeting HCV translation (dashed curve). Twenty-two to 96 hours later, replication was assessed by quantitation of luciferase reporter activity. The results are expressed as a percentage of entry (A) or infection (B) or Relative light units (RLU) per microgram protein (C).

To investigate a potential effect of apoE silencing on viral replication, Huh7.5.1 cells were co-electroporated with HCV Luc-Jc1 RNA and apoE siRNA. As shown in Fig. 5C, silencing of apoE expression did not significantly alter HCV replication. In contrast, an antiviral siRNA (siHCV331) targeting the 5′ nontranslated region of the viral genome21 markedly inhibited viral replication (Fig. 5C). These results demonstrate that silencing of apoE expression did not modulate HCV replication. This finding is further supported by the results of co-localization studies (Fig. 3) in which HCV NS5A protein was also detected in cells with low or absent apoE expression.

ApoE Is Involved in Assembly and Release of Infectious HCV Virions.

Because apoE was required to produce infectious viral particles but did not interfere with viral replication, it is likely that apoE is involved in a postreplication stage of the viral life cycle. Postreplication virus–host interaction stages include viral assembly and release of viral particles from the infected hepatocytes. To address whether apoE was required for assembly and release, we partially purified intracellular infectious viral particles from cellular lysates or supernatants from Luc-Jc1 replicating cells. Infectivity of the particles present in cell culture lysates or supernatants was analyzed after incubation with naïve Huh7.5.1 cells. Two experimental conditions were used for apoE silencing: protocol A, in which siRNAs targeting apoE or control siRNA were transfected 72 hours before electroporation of cells with HCV RNA, and protocol B, in which siRNAs were co-electroporated with HCV RNA (Fig. 6A). ApoE silencing was analyzed by immunoblot 72 hours after HCV RNA electroporation (Fig. 6B). Protocol A resulted in less efficient apoE silencing than protocol B at the time of HCVcc harvesting (Fig. 6B). In both protocols, apoE silencing resulted in a marked decrease in the infectivity of extracellular virions, confirming a functional role for apoE in virion production (Fig. 6C). When siRNAs and HCV RNA were co-transfected (protocol B, Fig. 6A), silencing of apoE resulted in a marked decrease in the infectivity of virions purified from intracellular lysates as well as virions present in supernatants from transfected cells (Fig. 6C, right panel). A dual effect on both intracellular and released infectious particles suggests that apoE is a co-factor for viral assembly. When siRNAs were transfected before electroporation of HCV RNA (protocol A, Fig. 6A), a similar decrease in the infectivity of cell culture supernatants was observed. In contrast, the infectivity of particles in cellular lysates was almost unchanged 72 hours after HCV RNA electroporation (Fig. 6C). The marked inhibition of released infectious particles (present in cell culture supernatants) without a concomitant decrease of intracellular particles (present in cellular lysates) suggests an additional effect of apoE silencing on release of viral particles. The minor effect of apoE silencing on assembly in this experiment may be attributable to the fact that low-level apoE silencing may still allow particle assembly. Furthermore, the impairment in release may have resulted in an accumulation of intracellular virions.

Figure 6.

Silencing of apoE expression results in an impairment of viral assembly and release. (A) Schematic outline of apoE silencing protocols. In protocol A, apoE siRNAs or CTRL siRNAs were transfected 72 hours before electroporation (EP) of cells with HCV Luc-Jc1 RNA, whereas in protocol B apoE siRNAs or CTRL siRNAs were co-electroporated with HCV Luc-Jc1 RNA. (B) ApoE expression after RNAi silencing. Seventy-two hours after electroporation of HCV RNA, apoE expression was analyzed by immunoblot as described in Fig. 4. (C) Effect of apoE silencing on the quantity of infectious virions purified from the intracellular and extracellular fractions of HCV replicating Huh7.5.1 cells. Supernatants (SN) or lysates (LYS) from cells treated with CTRL siRNAs (black bars) or apoE siRNAs (gray bars) as shown in A were subjected to sucrose cushion ultracentrifugation, and pelleted particles were used to infect naïve Huh7.5.1 cells. Infectivity was assessed as described in Fig. 4. The results are expressed as a percentage of infection of cells treated with CTRL siRNA (mean ± SD; n = 6). (D) Biophysical characterization of released HCV virions in cells with silenced apoE expression. Cells were transfected with siCTRL or siApoE as described in protocol A. Concentrated (50×) cell culture supernatants were subjected to 10% to 50% iodixanol step gradients and subjected to equilibrium ultracentrifugation as described in Materials and Methods. Fractions obtained from gradients with siCTRL (in black) or siApoE (in gray) transfected cells were measured for RNA content by quantitative reverse transcription PCR quantification (lines) and tested for infectivity on naïve Huh7.5.1 cells (bars). Infectivity was measured 72 hours later by quantitative reverse transcription PCR quantification. RNA cell extraction was normalized with glyceraldehyde 3-phosphate dehydrogenase quantification. Mean and standard deviation of three independent experiments are shown. The dotted line represents the density (in g/mL) of each fraction.

To further address the impact of apoE silencing on viral assembly and egress, we performed time course experiments after apoE silencing in HCV replicating cells. ApoE silencing resulted in an early effect on the release of infectious particles and a late effect on the assembly of intracellular infectious virions (Fig. 7). These findings suggest that apoE is a co-factor for two distinct steps in the assembly–egress process. The delayed effect of apoE silencing on assembly may be attributable to the impairment of viral release resulting in an accumulation of intracellular virions as well as the presence of a pool of pre-assembled particles that are not affected by apoE silencing during early time points.

Figure 7.

Time-course of apoE silencing and impact on production of infectious virions. (A) Schematic outline of the time course of the experimental protocol. Cells were electroporated with HCV Luc-Jc1 RNA 3 days (day −3) before transfection with apoE and control (CTRL) siRNAs (day 0). On day 0, supernatants were removed and replaced by fresh tissue culture medium to deplete HCVcc in the extracellular medium at the time of siRNA transfection. On days 1, 2, and 3 after siRNA transfection, supernatants or lysates from cells treated with CTRL siRNAs or apoE siRNAs were subjected to sucrose cushion ultracentrifugation, and pelleted particles were used to infect naïve Huh7.5.1 cells. Infectivity was assessed as described in Fig. 4. (B) Analysis of silenced apoE expression in cell lysates using immunoblotting as described in Fig. 4A. (C, D) Infectivity of supernatants (SN, shown in panel C) or lysates (LYS, shown in panel D) from cells treated with CTRL siRNAs (black bars) or apoE siRNAs (gray bars). The results are expressed as a percentage of infection of cells treated with CTRL siRNA (mean ± SD; n = 6).

As shown in Fig. 6D, apoE silencing did not markedly modify the biophysical properties of released infectious viral particles (Fig. 6D).

ApoE-NS5A Interaction Is Lost in Viral Mutants with a Defect in Viral Assembly.

To further confirm the functional impact of apoE–NS5A interaction for viral assembly, we studied apoE–NS5A interaction using mutant NS5A from viral variants with a defect in virus production. Using site-directed mutagenesis, we introduced an alanine triplet substitution at positions 99 to 101 or 102 to 104 of the JFH1 NS5A protein as described by Miyanari et al.8 These mutations have been shown to result in a defect of virion production in an infectious cell culture model.1, 8 As shown in Fig. 8A, NS5A mutants containing these mutations lost their ability to interact with apoE in a yeast two-hybrid system. A marked reduction in apoE–NS5A interaction (Fig. 8B) was also observed for NS5A of a viral variant containing a deletion of the NS5A domain III and a marked impairment of viral assembly.1, 8 This functional correlation between apoE–NS5A binding (Fig. 8) and assembly/production phenotype1, 8 further confirms the functional relevance of apoE–NS5A interaction in the production of infectious viral particles.

Discussion

Using an infectious HCV cell culture system, we demonstrate that apoE binds to HCV nonstructural protein NS5A. This conclusion is supported by a specific and easily detectable apoE–NS5A interaction in co-immunoprecipitation and co-localization studies in human hepatoma cells containing replicating infectious HCV as well as a yeast two-hybrid assay (Figs. 1–3).

The HCV NS5A protein is anchored in the endoplasmic reticulum (ER) via its N-terminal part (31) and forms ‘claw-like’ dimers presumably facing away from the membrane where it could accommodate either single-or double-stranded RNA and interact with viral and cellular proteins and membranes (32, 33). Having identified an interaction between apoE and NS5A in HCV replicating cells, it is conceivable that the apoE-NS5A interaction takes place in the vicinity or within the replication complex and that this interaction recruits apoE to the viral assembly modules which results in the production of viral particles released from the infected hepatocytes. The hypothesis of apoE-NS5A interaction within the membranous web is supported by the vesicular membranous structures harbouring replicating virus being enriched with lipoproteins including apoE (5, 6). This hypothesis is also supported by recent studies demonstrating an important role of NS5A in HCV virion production: NS5A has been shown to be recruited by core-associated lipid droplets in replicating Huh7.5 cells to allow robust infectious particle production (8, 12). Using reverse genetics, the NS5A domain III has been identified as an important viral determinant for virion production (1, 8). The relevance of apoE-NS5A interaction for virus production is further supported by the loss of apoE-NS5A binding from viral variants (8, 13) with a defect in virus production in the yeast-two-hybrid system (Fig. 8). Thus, our results identify apoE-NS5A binding as a novel virus-host interaction and point to apoE-NS5A binding having a crucial function in the production of infectious particles. Further studies are underway to fine-map the structural determinants of NS5A-apoE interactions and its impact on the viral phenotype.

The functional characterization in an infectious cell culture system suggests that apoE is involved in assembly and release of viral infectious particles. This conclusion is supported by the finding that apoE silencing results in a marked decrease in viral particles in cell lysates and supernatants of Huh7.5.1 cells electroporated with HCV RNA (Protocol B; Fig. 6C). Interestingly, when apoE silencing was less pronounced at the time of HCVcc harvest (protocol A; Fig. 6B), apoE silencing still resulted in a decrease in released infectious particles with concomitant retention of infectious viral particles in the host cell (Fig. 6C). It seems likely, therefore, that efficient silencing of apoE at the time of HCVcc harvest results in impairment of viral assembly and release, whereas low-level silencing may allow assembly of viral particles but results in impairment of viral release. Time course experiments (Fig. 7) showed an early effect of apoE silencing on released infectious particles and a late effect on the assembly of intracellular infectious virions. These data further support a model for apoE as a co-factor for two distinct steps of the assembly-egress process of the viral life cycle.

Recent studies have provided evidence that the very low density lipoprotein (VLDL) pathway appears to play a role in HCV assembly, maturation and egress. Indeed, the membrane vesicles in which HCV replicates are highly enriched in proteins required for VLDL assembly, including apoB, apoE, and microsomal triglyceride transfer protein (5, 6). Moreover, Gastaminza et al. (6) demonstrated that HCV assembly and maturation occur in the endoplasmic reticulum and post-endoplasmic reticulum compartments, respectively, in a manner that parallels the formation of VLDL. In addition, it was demonstrated that only low-density particles are efficiently secreted and that immature particles are actively degraded.6 Gastaminza and colleagues6 postulated that by coopting the VLDL assembly, maturation, degradation, and secretory machinery of the cell, HCV acquires its hepatocyte tropism. Efficient HCV spread has been shown to be impaired in cells expressing reduced apoB levels.6 Furthermore, reduced apoB secretion by short hairpin RNA has been shown to reduce viral particle assembly and secretion without interfering with infection efficiency or HCV RNA replication.6 Complementing these studies, Chang et al.11 demonstrated that HCV infectious virions purified from low-density fractions from cell culture supernatants are assembled as apoE-enriched lipoprotein particles and could be precipitated specifically by anti-apoE and anti-E2 antibodies. Collectively, our results identify NS5A as a viral factor providing a link between virus production and the VLDL pathway.

Several studies in HCV-infected patients have shown a statistically significant association between the presence of allelic isoforms of apoE and the severity of HCV-induced liver disease.34, 35 The apoE-epsilon4 allele was associated with a poor sustained viral response to interferon-alpha–based treatment.36 Taken together, these observations underline the impact of apoE as a host factor playing an important role in HCV–host interactions and the pathogenesis of disease.

Because our results demonstrate that silencing of intracellular apoE results in impaired particle assembly and release, it is conceivable that the interaction of HCV with components of the VLDL export pathway is a novel target for antiviral therapies. This concept is supported by a recent study demonstrating that stimulation of infected hepatocytes with the flavonoid naringenin significantly inhibits HCV secretion.37 Thus, drug-induced decrease of apoE in the liver of the HCV-infected patient may represent a novel approach for control of viral infection.

Acknowledgements

The authors thank C. Thumann for support in HCVcc purification (Inserm U748), R. Bartenschlager (University of Heidelberg), T. Wakita (National Institutes of Health, Tokyo), C. M. Rice (Rockefeller University, New York), F. V. Chisari (The Scripps Research Institute, La Jolla), and F.-L. Cosset (Inserm U758, ENS, Lyon) for providing HCV strains and human hepatoma cell lines, and K. Shimotohno, J. Dubuisson, and R. Bartenschlager for helpful discussions.

Ancillary