Apolipoprotein E Likely Contributes to a Maturation Step of Infectious Hepatitis C Virus Particles and Interacts with Viral Envelope Glycoproteins

ABSTRACT

The assembly of infectious hepatitis C virus (HCV) particles is tightly linked to components of the very-low-density lipoprotein (VLDL) pathway. We and others have shown that apolipoprotein E (ApoE) plays a major role in production of infectious HCV particles. However, the mechanism by which ApoE contributes to virion assembly/release and how it gets associated with the HCV particle is poorly understood. We found that knockdown of ApoE reduces titers of infectious intra- and extracellular HCV but not of the related dengue virus. ApoE depletion also reduced amounts of extracellular HCV core protein without affecting intracellular core amounts. Moreover, we found that ApoE depletion affected neither formation of nucleocapsids nor their envelopment, suggesting that ApoE acts at a late step of assembly, such as particle maturation and infectivity. Importantly, we demonstrate that ApoE interacts with the HCV envelope glycoproteins, most notably E2. This interaction did not require any other viral proteins and depended on the transmembrane domain of E2 that also was required for recruitment of HCV envelope glycoproteins to detergent-resistant membrane fractions. These results suggest that ApoE plays an important role in HCV particle maturation, presumably by direct interaction with viral envelope glycoproteins.

IMPORTANCE The HCV replication cycle is tightly linked to host cell lipid pathways and components. This is best illustrated by the dependency of HCV assembly on lipid droplets and the VLDL component ApoE. Although the role of ApoE for production of infectious HCV particles is well established, it is still poorly understood how ApoE contributes to virion formation and how it gets associated with HCV particles. Here, we provide experimental evidence that ApoE likely is required for an intracellular maturation step of HCV particles. Moreover, we demonstrate that ApoE associates with the viral envelope glycoproteins. This interaction appears to be dispensable for envelopment of virus particles but likely contributes to the quality control of secreted infectious virions. These results shed new light on the exploitation of host cell lipid pathways by HCV and the link of viral particle assembly to the VLDL component ApoE.

INTRODUCTION

Hepatitis C virus (HCV) is a major cause of chronic liver disease, leading to liver cirrhosis and hepatocellular carcinoma. Currently, ∼170 million people are thought to be persistently infected with HCV. In spite of the availability of highly active antiviral drugs, it is expected that the number of patients with serious liver diseases, including hepatocellular carcinoma, will increase further in the next 5 to 10 years (1).

Several studies suggest that the HCV life cycle is closely linked to host cell lipid metabolism and that the virus exploits lipid synthesis pathways for its replication and virus particle formation (2). Indeed, HCV replication is abolished by treatment with inhibitors of cholesterol and fatty acid biosynthesis, and blockage of very-low-density lipoprotein (VLDL) formation also affects virion assembly and release (3, 4). Furthermore, lipid droplets (LDs), which are the source for VLDL production (5), play an important role in HCV assembly (6), and several host factors involved in VLDL synthesis participate in HCV particle production (3, 4, 7, 8). The tight link between VLDL and HCV assembly would be in line with the notion that HCV particles are secreted as lipoviroparticles (LVPs). These hybrid particles are enriched in triglycerides and cholesterol esters and are composed of the structural proteins and human apolipoproteins, including ApoB, ApoE, ApoA-I, and ApoC-I (7, 9–12). Of these, ApoE appears to have a dual function for HCV. First, as an integral part of HCV particles, ApoE contributes to virus entry into the hepatocyte by mediating high-affinity interactions with cell surface molecules, such as LDL receptor (LDLR), scavenger receptor class B type I, and heparan sulfate proteoglycan (13). Second, ApoE is required for the production of infectious HCV particles (7, 8, 14, 15).

ApoE is an exchangeable apolipoprotein that plays an important role in VLDL assembly and cellular lipid transport by high-affinity binding to the LDLR and the LDLR-related protein (16). In the lipid-free state, ApoE has two independently folded structural domains: an N-terminal domain containing the LDLR-binding region and a C-terminal domain containing the major lipoprotein-binding elements. However, in the absence of lipid, ApoE has limited structural stability but undergoes large conformational changes upon lipid binding (16).

HCV particle assembly, i.e., formation of infectious virions, can be divided into three distinct steps: formation of the nucleocapsid by packaging of viral RNA into the capsid shell, envelopment of the nucleocapsid, which is the process of acquisition of the lipid envelope surrounding the nucleocapsid, and maturation of virions, which can be regarded as the process by which assembled virus particles acquire full infectivity. Whether these processes occur in a sequential manner or are coupled is not known. It is assumed that HCV assembly occurs in specialized lipid-rich microdomains at the endoplasmic reticulum (ER) membrane located in close proximity to cytosolic LDs. Indeed, it has been reported that the structural proteins core, E1, and E2 localize to intracellular lipid rafts (17, 18) and that viral replication complexes are enriched in these specialized membrane microdomains (19). Furthermore, it was shown that HCV particles contain large amounts of cholesterol and sphingolipids, the main components of lipid rafts, and that these lipids are essential for virus maturation and infectivity (12, 17, 20). These studies strongly suggest that these lipids play an important role in HCV assembly, but their precise contribution to virus formation and release is unclear. Moreover, it is unknown how the envelope glycoproteins are targeted to the assembly sites and what the exact role of ApoE in the assembly and release process is. Therefore, in the present study we determined the mechanism by which ApoE is associated with and contributes to assembly and release of infectious HCV particles.

MATERIALS AND METHODS

Cell culture.

The highly permissive Huh7-derived cell line overexpressing CD81, Huh7/LunetCD81H (21), was maintained in Dulbecco's modified Eagle medium (DMEM) (Invitrogen, Karlsruhe, Germany) supplemented with 2 mM l-glutamine, nonessential amino acids, 100 U/ml penicillin, 100 μg/ml streptomycin, 10% fetal calf serum (DMEMcplt), and 500 μg/ml G418 (Geneticin; Invitrogen). Puromycin (Sigma) was added at a final concentration of 5 μg/ml in the case of short hairpin RNA (shRNA)-transduced stable cell lines (sh-NT and sh-ApoE). sh-ApoE cells stably overexpressing ApoE or ApoE-HA, which possesses a C-terminal hemagglutinin (HA) affinity tag (ApoE^WT and ApoE^HA cell lines [for wild-type or C-terminally HA-tagged ApoE]) were generated by lentiviral transduction and cultured in the presence of 500 μg/ml G418 (Geneticin; Invitrogen), 5 μg/ml puromycin, and 10 μg/ml blasticidin (Invitrogen). sh-NT cells stably overexpressing E1-E2 (spE1E2) or spE1E2_DelE2TMD were generated by lentiviral transduction and selection in the presence of 500 μg/ml zeocin. For maintenance, cells were cultured in the presence of 500 μg/ml G418, 5 μg/ml puromycin, and 50 μg/ml zeocin (Invitrogen).

Plasmids.

Plasmids encoding a subgenomic reporter replicon (pFK_I₃₈₉LucNS3-3′_JFH_dg), the nonreplicative mutant pFK_I₃₈₉LucNS3-3′_NS5BΔGDD_JFH_dg (abbreviated as sgNS3-5B/Fluc-WT and sgNS3-5B/Fluc-ΔGDD, respectively), full-length genomes derived from the HCV chimera Jc1 (pFK_Jc1_dg, pFK_Jc1_p7KRQQ_dg, and pFK_i389RLuc2a_Core-3′_Jc1; abbreviated as Jc1, Jc1_p7KRQQ, and JcR2a, respectively), and the full-length dengue virus (DENV) reporter genome (pFK-DVs-R2A) have been described previously (22–26). The plasmids containing the coding sequences of human ApoE3 and C-terminally HA-tagged ApoE3 (pWPI_ApoE3 and pWPI_ApoE3-HA) were described elsewhere (27). For the expression of the structural region of the HCV polyprotein, plasmids derived from pWPI_zeo were constructed. In brief, DNA fragments generated by PCR and containing flanking recognition sequences for BamHI and SpeI were inserted into the vector, giving rise to plasmids pWPI_Core-NS2, pWPI_spE1-NS2, pWPI_spE1E2, pWPI_Core-ΔE1E2-NS2, pWPI_Core-p7, and pWPI_spE1-p7. Note that plasmids pWPI_spE1-NS2 and pWPI_spE1E2 contain the 3′ terminal coding sequence of core, which is the signal sequence of E1 (28). The sequence encoding the HCV envelope glycoproteins of the H77 strain (genotype 1a; GenBank accession no. AF011751) was amplified by PCR and inserted into pWPI_zeo or pcDNA3.1(+) (giving rise to plasmids pWPI_spE1E2[H77], pcDNA_spE1E2[H77], pcDNA_spE1[H77], and pcDNA_spE2[H77]). In the case of plasmids pWPI_spE2_DelE2TMD[H77], pcDNA_spE1E2_DelE2TMD[H77], and pcDNA_spE2_DelE2TMD[H77], the C-terminal transmembrane domain of E2 (amino acid residues 716 to 746; numbers refer to the position of the H77 polyprotein) was deleted. The integrity of all constructs was verified by nucleotide sequence analysis.

Production of lentiviruses.

Human immunodeficiency virus (HIV)-based particles that were pseudotyped with the vesicular stomatitis virus glycoprotein (VSVg) were generated by transfection of 293T cells as described previously, with slight modifications (29). For the production of lentiviruses encoding shRNA that were used to establish cell lines sh-NT and sh-ApoE, 293T cells were cotransfected with the following vectors: first, transfer vector pAPM, encoding the puromycin resistance gene and an shRNA sequence targeting the 3′-untranslated region of ApoE (shmirAPOE1141) or control shRNA (sh-NT); second, the HIV-1 packaging plasmid (psPAX2); third, a VSVg expression vector (pMD2.G). Cells were transfected by using the CalPhos mammalian transfection kit as recommended by the manufacturer (Becton Dickinson). The following shRNA targeting sequences were used: shmirAPOE1141, 5′-TGCTGTTGACAGTGAGCGCGGACCCTAGTTTAATAAAGATTAGTGAAGCCACAGATGTAATCTTTATTAAACTAGGGTCCATGCCTACTGCCTCGGA-3′; sh-NT, 5′-TGCTGTTGACAGTGAGCGCTCTCGCTTGGGCGAGAGTAAGTAGTGAAGCCACAGATGTACTTACTCTCGCCCAAGCGAGATAGTGAAGCCACAGATGTA-3′ (underlined sequences indicate the core sequence of the shRNA). For the production of lentiviral particles suitable for overexpression, 293T cells were cotransfected with the transfer vector encoding the gene of interest and a blasticidin or zeocin resistance gene (pWPI_BLR or pWPI_zeo, respectively), the HIV-1 packaging plasmid (pCMV), and a VSVg expression vector (pMD.G) by using the same transfection method as that described above.

Transient transfection of DNA expression constructs.

Cells were transfected with pcDNA3.1(+)-based constructs by using Lipofectamine 2000 (Invitrogen). Cells were seeded into each well of a 6-well plate 1 day prior to transfection. For transfection, 4 μg plasmid was mixed with 10 μl Lipofectamine 2000 and applied to cells as recommended by the manufacturer. After 2 days, cells were harvested for coimmunoprecipitation (co-IP) assay.

ELISA.

Intracellular and extracellular apolipoproteins (ApoCI and ApoE) were measured by using the human ApoCI enzyme-linked immunosorbent assay (ELISA) kit (Cell Biolabs, USA) and the human ApoE ELISA kit (Cell Biolabs, USA) according to the instructions of the manufacturer.

Antibodies.

For the detection of viral antigens, the following antibodies were used: mouse monoclonal antibody reacting with NS5A domain III of the HCV isolate Con1 and cross-reacting with NS5A of JFH1 (9E10; a kind gift from C. M. Rice, NY, USA), mouse monoclonal antibody reacting with NS3 of the JFH1 isolate (2E3) (30), rabbit polyclonal antibody reacting with NS3 (31), mouse monoclonal antibody reacting with NS2 (6H6; a kind gift from C. M. Rice, NY, USA) (32), mouse monoclonal antibody reacting with E2 (Ap33; a kind gift from Genentech Inc., CA, USA) (33), rat monoclonal antibody reacting with E2 (3/11; a kind gift from J. A. McKeating, Birmingham, United Kingdom) (34), mouse monoclonal antibody reacting with E1 (A4; a kind gift from J. Dubuisson, Lille, France) (35), rabbit polyclonal antibody reacting with core (C830) (36), and mouse monoclonal antibody reacting with core (C7-50) (sc57800; Santa Cruz, TX, USA).

For detection of cellular proteins, the following antibodies were used: rabbit polyclonal antibody reacting with caveolin-1 (sc894; Santa Cruz), goat polyclonal antibody reacting with ApoE (AB947; Millipore, MA, USA), rabbit polyclonal antibody reacting with protein disulfide isomerase (PDI; Sigma), and mouse monoclonal antibody against a cis-Golgi marker (GM130; BD Bioscience). Primary antibodies reacting with β-actin (A5441), HA-specific agarose beads (A2095), and secondary horseradish peroxidase-conjugated antibodies were purchased from Sigma-Aldrich.

In vitro transcription and RNA transfection.

In vitro transcripts generated with pFK-based plasmids were transfected into cells by electroporation as described previously (37). Methods for in vitro transcription and capping of DENV RNA, as well as transfection of DENV RNA into Huh7 cells, have been reported previously (26).

Determination of virus titers and HCV core protein amounts.

Cells (4 × 10⁶) of the stable cell lines sh-NT and sh-ApoE were transfected with 10 μg of Jc1 RNA by electroporation and resuspended in 15 ml of DMEMcplt. Cells were seeded into 6-well plates (2 ml per well) and harvested 4, 24, 48, and 72 h later. To determine the amounts of extracellular infectivity, at each time point supernatants were harvested, filtered through a 0.45-μm-pore-size filter, and stored at 4°C. For determination of intracellular infectivity titers, cells from two wells of a 6-well plate were rinsed three times with phosphate-buffered saline (PBS) and scraped into 0.6 ml of PBS. Cells were pelleted by centrifugation for 5 min at 500 × g, resuspended in 0.6 ml of DMEMcplt, and subjected to three freeze-thaw cycles, and cell debris was removed by centrifugation for 10 min at 20,000 × g. Virus titers were determined by limiting-dilution assay using Huh7.5 target cells and staining of the NS3 protein with the 2E3 antibody as described elsewhere (38). Titers of infectious virus were expressed as 50% tissue culture infective doses (TCID₅₀) per ml according to methods described previously (39, 40).

For quantification of core protein amounts, suspensions of transfected cells were seeded into a 24-well plate (0.5 ml/well). At different time points, cell culture supernatants were prepared as described above. To determine intracellular core amounts, cells were washed twice with PBS and lysed by addition of 0.2 ml PBS containing 1% Triton X-100 and protease inhibitor (Roche). Lysates were cleared by centrifugation at 20,000 × g for 10 min at 4°C. HCV core protein was quantified using a commercial chemiluminescent microparticle immunoassay (CMIA) (6L47; Architect HCV Ag reagent kit; Abbott Diagnostics, Abbott Park, USA) according to the manufacturer's instructions. If required, samples were diluted with PBS containing 0.5% Triton X-100.

Amounts of infectious DENV particles were determined by plaque formation unit (PFU) assay. Briefly, cell culture supernatants harvested at different time points were serially diluted and used to inoculate VeroE6 cells that had been seeded into 24-well plates. After 1 h of incubation, inocula were removed and cells were covered with Eagle's minimum essential medium (MEM) (Invitrogen) containing 1.5% carboxy-methylcellulose (Sigma-Aldrich). Cultures were incubated for 7 days at 37°C, fixed, and stained with a 1% crystal violet solution.

Transient replication assay.

Cells transfected with subgenomic replicon RNAs encoding the firefly luciferase (sgNS3-5B/Fluc-WT and sgNS3-5B/Fluc-ΔGDD) or full-length HCV RNA encoding Renilla luciferase (JcR2a) were seeded into 12-well plates. Replication was determined by measuring luciferase activity in cells 4, 24, 48, and 72 h after transfection as described previously (37). Each lysate was measured in triplicate. Luciferase activities, measured 4 h after transfection, reflect transfection efficiency and were used for normalization. In the case of DV-R2A-infected cultures, luciferase activity was normalized to values measured 8 h after infection.

Coimmunoprecipitation.

Huh7/LunetCD81H-based stable cell lines, ApoE^WT and ApoE^HA, were transfected with the constructs specified in the results section. After 72 h, cell lysates were prepared by scraping cells off the plate with IP buffer 1 (20 mM Tris-HCl [pH 7.4], 0.1% NP-40, 50 mM NaF, 5 mM sodium orthovanadate, 100 mM NaCl) or IP buffer 2 (20 mM Tris-HCl [pH 7.4], 1% Triton X-100, 100 mM NaCl). In the case of experiments shown in Fig. 7C and 8, cells were lysed by using IP buffer 2. In the case of experiments shown in Fig. 8A, IP buffer 2 was supplemented with 0.5% (wt/vol) saponin or 0.1% (wt/vol) SDS. After 1 h of incubation on ice, cell debris was removed by 30 min of centrifugation at 20,000 × g. For HA affinity capture, samples were incubated with HA-specific agarose beads (Sigma-Aldrich) for 5 h by continuously inverting the tubes at 4°C. Beads were washed three times with the respective IP buffer, samples were eluted with 5% SDS in PBS, 5× Laemmli buffer (250 mM Tris-HCl [pH 6.8], 10% SDS, 30% glycerol, 0.02% bromophenol blue, 10% β-mercaptoethanol) was added, and samples were incubated for 5 min at 95°C. Eluted proteins were analyzed further by Western blotting.

FIG 7.

Interaction of ApoE with HCV envelope glycoproteins. (A) ApoE^WT and ApoE^HA cells were electroporated with Jc1 and lysed 72 h later by using lysis buffers with different levels of stringency (see Materials and Methods). After affinity capture with HA-specific antibody, captured immunocomplexes were analyzed by Western blotting using antibodies specified on the left. To determine capture efficiency, 0.5% of the total cell lysates used for the immunoprecipitation was analyzed in parallel (input). Note that in the case of ApoE Western blotting, only 0.5% of immunocaptured protein was loaded onto the gel. (B) Schematic representation of used expression constructs that are derived from the Jc1 chimera. The dark-shaded area refers to the sequence of the HCV J6 isolate (63), whereas the light-shaded area refers to the JFH-1 sequence (64). (C) ApoE^WT and ApoE^HA cells were transduced with lentiviruses encoding HCV proteins specified on the left and 72 h later were lysed using buffer 2. After HA-specific immunoprecipitation, protein complexes were analyzed by Western blotting. Relative co-IP efficiencies of E2 are given below the respective lanes of the HA-IP E2-specific blot. To determine capture efficiency, 2.5% of the total cell lysates used for the immunoprecipitation was analyzed in parallel (input). In the case of ApoE Western blot, only 2.5% of immunocaptured protein was loaded onto the gel. Numbers on the right of panels A and C refer to molecular mass standards in kDa.

FIG 8.

Critical role of the E2 transmembrane domain for ApoE interaction with E1/E2 envelope glycoproteins. (A) ApoE^WT and ApoE^HA cells were transfected with the H77-derived construct spE1E2 (sp, signal peptide). After 48 h, cells were lysed with buffer 2 alone, with saponin, or with 0.1% SDS as specified at the top. ApoE-HA was captured by HA-specific immunoprecipitation. (B) ApoE^WT and ApoE^HA cells were transfected with H77-derived expression constructs specified at the top. After 48 h, cells were lysed with buffer 2, and ApoE-HA was captured by HA-specific immunoprecipitation. Immunocomplexes and 2.5% of the input were analyzed by Western blotting using antibodies specified on the left. Note that in the case of the ApoE Western blotting, only 2.5% of immunocaptured protein was loaded onto the gel. Co-IP efficiencies of E2 and E1 are given below the respective lanes in the HA IP blots. Numbers on the right of panels A and B refer to molecular mass standards in kDa.

Western blot and immuno-dot blot analysis.

Samples for both Western blotting and immuno-dot blotting were denatured in 5× Laemmli buffer by incubation at 95°C for 5 min. For Western blotting, proteins were resolved by electrophoresis into SDS-polyacrylamide gels and transferred onto a polyvinylidene difluoride (PVDF) membrane. For immuno-dot blot analysis, samples were spotted onto a nitrocellulose membrane by vacuum transfer using a dot blot apparatus. Membranes were blocked by incubation with 5% nonfat milk, and proteins were detected by using primary antibodies specified in Results and secondary antibodies conjugated with horseradish peroxidase. Membranes were developed by using the Western Lightning Plus-ECL reagent (PerkinElmer, Waltham, MA).

Iodixanol density gradient analysis.

For iodixanol density gradient analysis, cells were electroporated with Jc1 RNA. After 48 h of transfection, supernatants were concentrated using a Centricon instrument (Amicon Ultra; Millipore), while intracellular lysates were prepared by repetitive cycles of freezing and thawing. Samples were layered on top of a PBS-based 20 to 80% OptiPrep (Axis-Shield) step gradient and subjected to isopycnic gradient centrifugation overnight at 120,000 × g at 4°C using an SW60 rotor (Beckman Coulter, Inc.). Eleven fractions were harvested from the top of the gradient and used for the determination of density, infectivity titer, and HCV core protein amounts.

Protease protection and rate zonal centrifugation assay.

The protease protection and rate zonal centrifugation assays were performed as previously described, with minor modifications (41). Cells were transfected with HCV RNA by electroporation and seeded into 10-cm-diameter dishes. After 2 days, cells were washed three times with PBS, scraped into 0.5 ml PBS, and subjected to five cycles of freezing and thawing. Cell lysates were cleared by centrifugation at 20,000 × g for 10 min at 4°C, and supernatants were collected. For the protease protection assay, samples were divided into three aliquots that were subjected to different experimental conditions: treatment with 15 μg/ml proteinase K (P2308; Sigma) for 40 min on ice, treatment with 15 μg/ml proteinase K and 1% Triton X-100 on ice, and mock treatment. Protease digestion was terminated by addition of 30 mM phenylmethylsulfonyl fluoride (PMSF; AppliChem) and 5× Laemmli buffer. Samples were heated for 5 min at 95°C and analyzed by Western blotting. For rate zonal centrifugation, cell lysates were layered on top of a continuous 0 to 30% sucrose-PBS gradient and centrifuged at 270,000 × g for 1 h at 4°C in an SW60 Ti rotor (Beckman Coulter, Inc.). Ten fractions were collected from the top, and their densities were determined by measuring refractive indexes.

Immunofluorescence analysis and confocal microscopy.

Cells seeded onto coverslips were fixed with 4% paraformaldehyde (PFA) for 10 min at room temperature (RT) and permeabilized with 0.1% Triton X-100 in PBS for 5 min at RT. After blocking with 3% bovine serum albumin (BSA) in PBS for 30 min at RT, cells were incubated with primary antibodies, diluted with 1% BSA-PBS, for 1 h at RT. Cells were incubated with secondary antibodies conjugated with Alexa fluorophore (Molecular Probes). For counterstaining of the nuclei, cells were incubated with 4′,6′-diamidino-2-phenylindole dihydrochloride (DAPI; Molecular Probes). Coverslips were mounted in Fluoromount-G mounting medium (Electron Microscopy Sciences, Fort Washington, PA, USA). Images were acquired by using a PerkinElmer spinning-disk confocal microscope. For selective permeabilization assay, cells were fixed with 4% PFA and permeabilized by treatment with 0.1% Triton X-100 for 5 min at RT or 5 μg/ml digitonin dissolved in 20 mM HEPES (pH 6.9), 0.3 M sucrose, 0.1 M KCl, 2.5 mM MgCl₂, and 1 mM EDTA for 15 min at 4°C. To quantify the degree of colocalization, images were analyzed by using the intensity correlation analysis plugin of the Fiji software package (NIH).

Membrane floatation assay.

The membrane floatation assay was performed as described previously, with minor modifications (42). Cells were scraped into TNE buffer (150 mM NaCl, 2 mM EDTA, 50 mM Tris-HCl [pH 7.4]) supplemented with protease inhibitor cocktail (Roche) and passed 25 times through a 25-gauge needle. Triton X-100 was added to the lysate at a final concentration of 1% (vol/vol). After incubation for 30 min on ice, cell lysates were cleared by centrifugation at 10,000 × g for 5 min at 4°C and supernatants were harvested. Samples were adjusted to a final concentration of 40% sucrose in TNE, and 1.2 ml was transferred to the bottom of a centrifuge tube and overlaid with 2.8 ml of 35% sucrose, followed by 0.4 ml of 5% sucrose dissolved in TNE. After 18 h of centrifugation at 271,000 × g at 4°C in an SW60 Ti rotor (Beckman Coulter, Inc.), 10 fractions of 400 μl were collected from the top of the gradient, mixed with 5× Laemmli buffer, and analyzed by immuno-dot blotting or Western blotting.

Statistical analysis.

Differences between sample populations were assessed using the two-tailed, unpaired Student's t test available in the GraphPad Prism 5 software package (GraphPad Software, Inc., La Jolla, CA). P values of less than 0.05 (indicated by asterisks) were considered statistically significant. The following categories were used: ***, P ≤ 0.001; **, P ≤ 0.01; *, P ≤ 0.05.

RESULTS

Impact of ApoE silencing on the production of infectious HCV particles.

Previous studies have shown that reducing ApoE expression impairs the production of infectious HCV particles; however, the precise step has not been identified (7, 8, 14, 15). To better define the role of ApoE in the HCV life cycle, we generated a highly permissive Huh7-derived cell pool with stable knockdown of ApoE expression (sh-ApoE). This shRNA led to a profound reduction of ApoE abundance both at the level of mRNA and protein amounts compared to control cells transduced with a nontargeting shRNA (sh-NT) (Fig. 1A and B). In agreement with previous reports (7, 8, 14, 15), silencing of ApoE did not affect replication of either a JFH1-derived luciferase reporter replicon (sgNS3-5B/Fluc-WT; not shown) or a full-length reporter virus genome (JcR2a) (Fig. 1C).

FIG 1.

Knockdown of ApoE expression does not affect HCV replication. (A) Huh7/LunetCD81H cells were transduced with shRNAs targeting the 3′-UTR of ApoE (sh-ApoE) or an irrelevant nontargeting control (sh-NT), and cell pools with stable shRNA expression were established. Amounts of ApoE mRNA in either cell pool were quantified by RT-qPCR and normalized to GAPDH mRNA amounts. The graph represents the mean values from three independent experiments and standard deviations of the means. (B) ApoE-specific knockdown was determined at the single-cell level by immunofluorescence. Scale bars represent 10 μm. (C) sh-ApoE and sh-NT cells were electroporated with fully functional Renilla luciferase-containing reporter virus genome (JcR2a) RNA. Cells were lysed at given time points after transfection, and luciferase activity was determined. Values are expressed relative to reporter activity measured 4 h posttransfection and reflect transfection efficiency. Data shown in panel C represent mean values and standard deviations of the means from three independent experiments.

To determine the impact of ApoE silencing on HCV particle production, sh-ApoE and sh-NT cells were transfected with the reporter-free HCV chimera Jc1, and at different time points, amounts of infectious intra- and extracellular virus were determined. Knockdown of ApoE reduced virus titers within cells and in culture supernatants ∼10-fold (Fig. 2A). However, the relative amounts of intra- and extracellular infectivity were not altered (Fig. 2B), arguing that silencing of ApoE expression affects assembly but not release of HCV particles. The quantification of core protein amounts from intra- and extracellular fractions revealed a significant reduction of core protein secretion from sh-ApoE cells (Fig. 2C, middle). Although intracellular core abundance was not affected, the total core protein amount produced by sh-ApoE cells was reduced (Fig. 2C, left and right, respectively).

FIG 2.

Knockdown of ApoE expression reduces amounts of infectious HCV but not DENV particles. (A) sh-ApoE and sh-NT cell lines were electroporated with Jc1 RNA. After 24, 48, and 72 h, supernatants were harvested, whereas cells were lysed by repetitive cycles of freezing and thawing. Intra- and extracellular infectivity was determined by limiting-dilution assay (TCID₅₀). (B) Ratio of intra- and extracellular infectivity relative to the total amount of infectivity. (C) Amount of HCV core protein contained within transfected cells or released into the supernatant of Jc1-transfected cells (left and middle, respectively). Total core protein amounts are shown on the right. (D) sh-ApoE cell pools stably expressing a wild-type ApoE (ApoE^WT) or a C-terminally HA-tagged ApoE (ApoE^HA) were generated by lentiviral transduction. sh-ApoE (ApoE-KD), sh-NT, ApoE^WT, and ApoE^HA cell lines were electroporated with Jc1. After 72 h, supernatants were harvested and extracellular infectivity was determined by limiting-dilution assay (TCID₅₀). (Lower) ApoE protein amounts were determined by Western blotting of cell lysates. Beta-actin served as the loading control. Relative ApoE amounts, normalized to sh-NT cells, are given below the respective lanes. (E) sh-ApoE and sh-NT cells were infected with the dengue reporter virus DV-R2A (1 PFU/cell). Cells were lysed at given time points, and supernatants were harvested to quantify virus production by plaque assay (right). Replication kinetics was determined by Renilla luciferase assay, and values were normalized to those measured 8 h postinfection. Data shown in panels A to E represent mean values and standard deviations of the means from three independent experiments. ***, P ≤ 0.001; **, P ≤ 0.01; *, P ≤ 0.05.

To exclude off-target effects of the knockdown, we reconstituted ApoE expression in sh-ApoE cells by stable transduction of ApoE^WT and ApoE^HA. As shown in Fig. 2D, in both cases virus titers were restored to those obtained in sh-NT cells, excluding off-target effects of the ApoE-specific shRNA and confirming the functionality of both ApoE proteins.

With the aim of determining the possible role of ApoE for particle assembly of other members of the Flaviviridae family, we analyzed the kinetics of DENV replication and virus production in sh-ApoE and sh-NT cell lines. In contrast to HCV, the silencing of ApoE had no effect on replication or production of infectious DENV particles (Fig. 2E). Taken together, these results suggest that ApoE plays a crucial role for HCV particle assembly and maturation but not virus release, and this role is specific to HCV but does not apply to DENV.

Evidence that ApoE is required for proper lipidation of infectious HCV particles.

To determine possible differences between intra- and extracellular HCV particles produced from sh-NT and sh-ApoE cells, we used iodixanol density gradient centrifugation. For virtually all fractions, amounts of intra- and extracellular infectivity were lower in the case of sh-ApoE cells (Fig. 3A). Importantly, although the density distribution of intra- and extracellular virus particles produced from sh-ApoE and sh-NT cells appeared rather similar, we observed a reproducible shift toward higher density (>1.1 g/ml) of virus particles released from sh-ApoE cells. This was best visible when the infectivity of each fraction was normalized to the total infectivity (Fig. 3A, lower right). In contrast, density distributions of core protein were comparable overall between the two cell lines (Fig. 3B). Most likely this is because an ∼100- to 1,000-fold excess of noninfectious virus particles is released from HCV-containing cells (43 and references cited therein); thus, the rather moderate density shift observed by infectivity assay would not be detectable when using core-based measurements. Importantly, the density shift observed with sh-ApoE-derived HCV particles was specific, because ApoE knockdown had no effect on the release of other apolipoproteins, as deduced from the unaffected production and release of ApoCI (Fig. 3C). Moreover, distribution patterns along the density gradient of apolipoproteins (ApoCI and ApoE) that were released from cells after transfection with full-length HCV RNA were comparable to those of mock-transfected control cells (data not shown).

FIG 3.

Biophysical properties of intra- and extracellular HCV particles isolated from sh-ApoE and sh-NT cells. sh-ApoE (red) and sh-NT (black) cells were electroporated with Jc1 RNA. After 48 h, supernatants were concentrated, intracellular lysates were prepared, and samples were fractionated by density gradient centrifugation. Eleven fractions were harvested from the top of the gradient and used for the determination of density, infectivity titer, and HCV core protein amounts. (A) Upper panels display absolute amounts of intra- and extracellular infectivity (TCID₅₀) and the density for each fraction. Lower panels display relative amounts of infectivity normalized to total infectivity detected in all fractions. (B) Analogous to panel A, but displaying amounts of HCV core protein. Data shown in panels A and B represent mean values and standard deviations of the means from three independent experiments. (C) Amounts of ApoCI and ApoE in gradient fractions as determined by ELISA. Mean values from three independent experiments are plotted. *, P ≤ 0.05. n.s, not significant.

No major role of ApoE for envelopment of HCV particles.

We next aimed to determine the step in virus assembly affected by ApoE depletion. First, we probed protection of nucleocapsids/core protein against proteolytic digestion in vitro as described recently (41). In this assay, the protease sensitivity of core protein indicates improper envelopment, whereas in an intact virus particle, core protein is protected by the lipid envelope. As a reference, we included a Jc1-derived p7 mutant (Jc1_p7KRQQ), which was shown to have a defect in core envelopment (41). Cell lysates from Jc1- or Jc1_p7KRQQ-transfected sh-ApoE or sh-NT cells were treated with proteinase K (PK) in the absence or presence of Triton X-100 (Fig. 4A). In the absence of detergent, a fraction of core protein was resistant to the digestion, indicating protection by the membranous envelope, whereas solubilization with the detergent led to complete core protein digestion, showing that the PK amount was not a limiting factor in the assay. The quantification of signal intensities of Western blots demonstrated that in control sh-NT cells, ∼80% of intracellular core protein is PK resistant, whereas this percentage is significantly lower in the case of the p7 mutant (Fig. 4A). However, there was no significant difference between ApoE-silenced and control cells both for the wild type and the mutant (Fig. 4A), suggesting that capsid envelopment is not compromised by ApoE silencing.

FIG 4.

Role of ApoE in HCV particle formation. (A) sh-ApoE and sh-NT cells were transfected with Jc1 or the assembly-deficient mutant Jc1_p7KRQQ. After 48 h, cell lysates were prepared by repetitive cycles of freezing and thawing, and envelopment of particles was probed by protease protection assay. Aliquots of each cell lysate were either mock treated, treated with 15 μg/ml proteinase K (PK) for 30 min on ice, or treated in the same way in the presence of 1% Triton X-100. The amount of PK-resistant core protein was determined by Western blotting (WB). (Upper) Representative core-specific blots are shown. (Lower) Core-specific bands detected in three independent experiments were quantified by densitometry, and values obtained with PK-treated samples were normalized to those of untreated samples. *, P ≤ 0.05. (B) Sedimentation profiles of core protein complexes produced in cells transfected with Jc1 or Jc1_p7KRQQ. Cell lysates were subjected to rate zonal centrifugation for 1 h at 270,000 × g. Twelve fractions were harvested from the top of the gradient and used for determination of density as well as core protein amount by using core ELISA. Mean values from two independent experiments are plotted. (C) Fractions of lysates from cells transfected with Jc1 (solid lines) or mock transfected (dashed lines) were used to determine amounts of ApoE and ApoCI by ELISA. Mean values from two independent experiments are plotted.

We next assessed the efficiency of nucleocapsid assembly by using rate zonal centrifugation of cell lysates, allowing separation of high-molecular-weight core protein-containing structures according to size and molecular mass. A single peak of fast-sedimenting core, likely corresponding to nucleocapsids or fully assembled virions, was detected in Jc1-transfected sh-NT cells, while no distinct peak could be detected in lysates of the p7 mutant, as described earlier (41) (Fig. 4B). Consistent with the previous results (Fig. 3C), this rate zonal centrifugation of the distribution of ApoCI across the gradient was not affected in the case of ApoE-depleted cells (Fig. 4C). Moreover, the sedimentation profiles of apolipoproteins (ApoCI and ApoE) were not affected by replicating HCV (Fig. 4C; compare Jc1 transfection to mock transfection). Taking these findings together, we conclude that ApoE is not required for the envelopment of HCV.

Several reports have indicated that defects in HCV assembly lead to an accumulation of core protein at the surface of LDs (6, 44). However, we did not detect an alteration in the subcellular localization pattern of core protein upon ApoE silencing (data not shown), consistent with our conclusion that ApoE acts at a stage postenvelopment. Therefore, we determined the colocalization between core and E2, which are integral components of HCV particles. In the case of the assembly-deficient p7 mutant, a clustering of core protein, most likely around LDs, was observed with only limited E2 colocalization with core (Fig. 5A, lower half; B, left, shows quantification). In the case of Jc1-transfected cells, a strong colocalization between core and E2 protein was found that was not affected by ApoE silencing (Fig. 5A, upper half; B, left, shows quantification). This result suggested that the colocalization between core and E2 does not depend on ApoE. Furthermore, the unaffected colocalization of ApoE and E2 in the case of the p7 mutant transfected into sh-NT cells suggested that the close proximity of these two proteins is independent of assembly (Fig. 5B, right).

FIG 5.

Colocalization of core and E2 proteins is not affected by ApoE knockdown. (A) Stable cell lines were electroporated with Jc1 or Jc1_p7KRQQ, and 48 h later immunofluorescence was performed by using antibodies specified at the top. The inserts in each panel correspond to magnifications of the boxed areas, with white arrows indicating the lines selected for the RGB line profiles shown in the right panel of each image. Scale bars represent 10 μm or 5 μm in regular and magnified images, respectively. (B) The degrees of colocalization between E2 and core proteins under all conditions and ApoE and E2 in sh-NT cells were quantified and are shown in the left and right panels, respectively. Each dot in the graph corresponds to one cell. ***, P ≤ 0.001.

Predominant colocalization of ApoE and E2 in the ER lumen.

Several previous reports have highlighted the importance of the interaction of ApoE with NS5A for the assembly and release of infectious HCV particles (8, 14, 45). This result is perplexing due to the membrane topologies of these two proteins. While NS5A resides on the cytoplasmic side of the ER membrane, ApoE localizes to the ER lumen, making a direct protein-protein interaction difficult to envisage. Therefore, we investigated whether viral proteins other than NS5A also interact with ApoE. As a first step, we determined the colocalization of viral proteins in HCV-infected cells by using immunofluorescence. Although we cannot exclude possible confounding effects caused by the use of different antibodies to detect the viral proteins, the highest degrees of colocalization of ApoE were found with E2, NS2, and reticular (most likely ER-resident) core proteins (Fig. 6A and B). Of note, ApoE-NS5A colocalization was significantly lower than that between ApoE and E2 (Fig. 6B). Moreover, consistent with earlier reports (46, 47), we observed that ApoE and E2 colocalized with an ER marker (PDI) and a Golgi marker (GM130), in agreement with the assumed envelopment site of virus particles and virion transport via the Golgi apparatus, respectively (Fig. 6C).

FIG 6.

Subcellular distribution and colocalization of ApoE with viral proteins. (A) Huh7/LunetCD81H cells were infected with 30 TCID₅₀/cell of Jc1. Colocalization of ApoE with viral proteins specified in the top of each panel was determined by immunofluorescence using monospecific antibodies. (B) The degree of colocalization of ApoE with given viral proteins was quantified. Each dot in the graph corresponds to one cell. Core_ring, core protein with ring-like staining pattern corresponding to lipid droplet localization; core-retic, core protein with ER-like staining pattern. (C) Huh7/LunetCD81H cells were electroporated with Jc1 RNA and fixed 48 h later. Immunofluorescence was performed using antibodies against E2, ApoE, and cellular organelle markers (PDI or the cis-Golgi marker GM130). (D) Huh7/LunetCD81H cells were transfected with Jc1 RNA and fixed 48 h later. Cells were permeabilized with digitonin or Triton X-100, and immunofluorescence was performed using antibodies specified at the top. Insets in the lower left of merged images show magnifications of boxed areas. White arrows indicate the selected lines in the RGB line profiles shown at the right of each row. Scale bars represent 10 μm or 5 μm in regular and magnified images, respectively.

To further validate ApoE-E2 colocalization, we performed immunofluorescence-based assays under permeabilization conditions that leave the ER membrane intact (low concentration of digitonin) or disrupt it (Triton X-100), with the latter allowing antibody detection of proteins residing in the ER lumen. As expected, in Jc1-transfected cells, ApoE, E2, and NS5A were visualized in cells permeabilized with Triton X-100, while only NS5A was detected in digitonin-permeabilized cells (Fig. 6D), arguing that an association between ApoE and E2 occurs only at the luminal side of the ER and Golgi structure.

Interaction between ApoE and E2 envelope glycoprotein.

We addressed a possible ApoE-E2 interaction more directly by using coimmunoprecipitation. To this end, ApoE^WT and ApoE^HA cell lines (the latter expressing a fully functional HA-tagged ApoE; Fig. 2D) were transfected with Jc1 RNA, and 72 h later cells were lysed and subjected to HA-specific immunoprecipitation. Under low-stringency conditions, ApoE coprecipitated with E2, NS2, and NS5A, the latter being consistent with earlier reports (8) (Fig. 7A, buffer 1). However, when we increased stringency, the coprecipitation of ApoE with NS5A or NS2 was no longer detectable (Fig. 7A, buffer 2). In contrast, interaction between ApoE and E2 was only slightly reduced, arguing for tight association between these two proteins.

We and others have shown earlier that E2 is part of a large protein complex orchestrated by NS2 and likely is required for envelopment of HCV particles (48–50). Thus, ApoE-E2 interaction might be mediated by other viral proteins. To address this possibility, we determined the minimum region of the HCV assembly module (core to NS2) required for interaction with ApoE by using various polyprotein fragments or individual proteins that were expressed in ApoE^WT and ApoE^HA cells (Fig. 7B). Among all tested conditions, E2 was consistently coprecipitated with ApoE (Fig. 7C). Of note, we found that in the absence of p7, E2-ApoE^HA coprecipitation efficiency was lowest (Fig. 7C, IP efficiency). This might be due to defects in p7-dependent protection from degradation (51) or to a possible impact of p7 on the folding of E2 affecting interaction efficiency with ApoE^HA. Importantly, neither core protein nor p7 and NS2 were necessary for coprecipitation of ApoE with E2.

To discriminate whether the ApoE-envelope glycoprotein interaction is mediated by protein-lipid or protein-protein interaction, we performed ApoE-specific immunocapture by using lysis buffers containing saponin to dissociate proteins from cholesterol-rich lipid microdomains (52) or SDS to disrupt protein-protein interactions (53). For this and the subsequent analyses we used the H77 isolate, because we had antibodies allowing the detection of E1 only for this strain. We found that the ApoE-E1/E2 interaction was resistant to saponin but not to SDS treatment (Fig. 8A), arguing for a direct protein-protein, rather than lipid-mediated, interaction.

We next determined whether ApoE associates with E1 or E2, each expressed on its own, or only with the complex by using cell lines expressing E1, E2, or E1/E2. Consistent with an earlier report (54), we found that E1 expressed on its own is aberrantly glycosylated, indicative of misfolding (not shown). Although these E1 species weakly coprecipitated with ApoE, this result was ambiguous and did not allow us to draw firm conclusions (not shown). In the case of the cell line expressing only E2 (spE2), no apparent effect on glycosylation was detected, as deduced from the electrophoretic mobility of the protein compared to the one expressed as part of the E1/E2 complex (Fig. 8B). Importantly, the efficiency of coprecipitation of singly expressed E2 with ApoE was only slightly lower than that of the E1/E2 heterodimer (Fig. 8B).

Dependency of ApoE interaction with E2 on the transmembrane domain of the viral envelope glycoprotein.

In an attempt to map the region of E2 responsible for interaction with ApoE, we designed several internal deletions, removing various parts of the E2 ectodomain as well as a deletion spanning the transmembrane domain (TMD) of E2. While the internal E2 deletions did not affect interaction with ApoE to a measurable extent (not shown), we found that removal of the E2 TMD in the context of the E1/E2 complex reduced E1-ApoE coprecipitation close to the background and completely abrogated E2-ApoE association in the case of sole E2 expression (Fig. 8B). These results suggest that the primary interaction partner of ApoE in the E1/E2 complex is E2.

It has been reported that the TMDs of the HCV envelope glycoproteins are multifunctional and serve different purposes, such as ER retention, membrane translocation of the ectodomains, and E1/E2 heterodimerization (reviewed in reference 55). Moreover, for several viruses it has been reported that the TMD of the envelope protein is required for targeting to cholesterol-rich lipid microdomains, which often are exploited as assembly platforms (56). Importantly, the structural proteins of HCV (core and the E1/E2 complex) were reported to associate with lipid rafts (17, 18). Using a membrane floatation assay, we found that in the case of the TMD deletion mutant (spE1E2_DelE2TMD), E2 remained at the bottom of the gradient, and its abundance was drastically reduced in the detergent-resistant membranes (DRMs), which represent the lipid raft fraction (Fig. 9A; B shows quantification). The analysis of selected gradient fractions by Western blotting corroborated that the presence of E2 in the DRM fraction depended on the E2 TMD (Fig. 9C). In addition, detection of caveolin-1, a marker for detergent-resistant membranes, in the top fraction and the detergent-soluble protein PDI in the bottom fraction validated the effectiveness of the separation procedure.

FIG 9.

Presence of E2 glycoprotein in detergent-resistant membrane fractions depends on the E2 transmembrane domain. (A) Huh7/LunetCD81H cells stably expressing spE1E2 or spE1E2_DelE2TMD were lysed by treatment with 1% Triton X-100 for 30 min on ice to keep cholesterol-rich microdomains intact and fractionated by discontinuous sucrose gradient centrifugation. Fractions were harvested from the top and analyzed by E2-specific immuno-dot blotting. (B) Quantification of the immuno-dot blot. The intensity of each dot was quantified by densitometry and normalized to total E2 detected in all fractions. (C) Gradient fractions of panel A and the input were analyzed by Western blotting using antibodies specified on the left of each panel. Note that twice the amount of fraction 1 compared to fraction 10 was loaded onto the gel. Numbers on the right refer to molecular mass standards in kDa. PDI, protein disulfide isomerase; Cav-1, caveolin-1, a marker for detergent-resistant membranes. (D) sh-NT cells were transfected with Jc1 RNA, and 72 h later, cells were lysed by treatment with 0.5% Triton X-100 on ice. Samples were fractionated by discontinuous sucrose gradient centrifugation and analyzed as described for panel A. Fractions were harvested from the top and analyzed by E2-specific immuno-dot blotting, core ELISA, and ApoE ELISA. The graph represents relative quantities of E2, core, and ApoE along the gradient. ***, P ≤ 0.001; **, P ≤ 0.01; *, P ≤ 0.05.

We note that the distribution of ApoE along the gradient did not mirror the ones of E2 and core protein (Fig. 9D). This most likely is because only a very small fraction of ApoE associates with E2, as deduced from the efficiency of coprecipitation (∼1.5%) (Fig. 8B). In conclusion, these results argue for an ApoE-E2 interaction required for recruitment of the envelope glycoprotein into the low-density fraction. These may correspond to detergent-resistant lipid microdomains or to lipoprotein particles to which E1/E2 complexes are bound.

DISCUSSION

Accumulating evidence shows that ApoE plays a major role in the formation of infectious HCV particles (7, 8, 14, 15). This is best exemplified by the fact that the ectopic expression of ApoE in cells unable to produce this protein renders them competent for production of infectious HCV (27, 57). Nevertheless, the molecular mechanisms by which ApoE contributes to HCV assembly remain largely unexplored.

Taking advantage of a cell line with efficient knockdown of ApoE expression, we observed a proportional reduction of intra- and extracellular virus titers without overt effect on viral replication, arguing for a role of ApoE in assembly, rather than virus release. However, silencing of ApoE expression reduced only extracellular core amounts, whereas the abundance of core protein within the cell was not elevated as might be expected. This seemingly paradoxical observation might be explained by the observation that even in sh-ApoE cells, enveloped virus particles are formed. Consistent with an earlier report (4), we speculate that owing to the lack of ApoE, they are not released but are instead degraded. In this respect, ApoE might act as a quality control, as only particles containing sufficient ApoE amounts would be secreted. The observation that infectious virus particles produced in sh-ApoE cells have an increased density further suggests that ApoE is required for the acquisition of lipids, with poorly lipidated particles having a low specific infectivity, which is consistent with what we found for particles in sh-ApoE cells. In this respect, HCV particles might undergo a postenvelopment maturation step to ascertain that only low-density (fully lipidated) particles are efficiently secreted, as proposed in an earlier report (4).

Recently it has been shown that the ability of cells to produce VLDL is not a prerequisite for the association with ApoB, suggesting that HCV assembly is not strictly dependent on VLDL formation (58). Moreover, ApoE participates in the secondary lipidation of VLDL precursors via fusion with luminal LDs (59). Therefore, ApoE might play an important role in the intracellular fusion between nascent VLDL and viral precursors, which lead immature virions to acquire proper lipidation.

Results obtained with protease protection assays and rate zonal centrifugation suggest that nucleocapsid formation and envelopment are unimpaired by ApoE silencing. This conclusion is supported by our immunofluorescence analysis showing that ApoE knockdown did not affect the high degree of core-E2 colocalization. However, it is unclear whether sites of high core-E2 colocalization correspond to assembly sites, since different core protein species exist: ER-associated core immediately after translation, ER- and LD-associated core that is not yet involved in assembly of virus particles, and core within productive assembly sites or within assembled intracellular particles. Nevertheless, we conclude that ApoE is not involved in virion assembly per se but rather is required for a step after envelopment and prior to release. These results are in perfect agreement with those of a very recent study also showing that ApoE is required for a postenvelopment step of HCV assembly (57).

Several previous reports suggested that the interaction of ApoE with NS5A is essential for the assembly of infectious HCV particles (8, 45). Although we also detected this interaction when using low-stringency conditions for coimmunoprecipitation, it is unknown how these proteins interact given their different membrane topologies. Since ApoE and E1/E2 complexes are components of infectious HCV particles and localize to the luminal side of the ER, and since NS5A appears to be absent from HCV particles (12), we speculated that E1/E2 interacts with ApoE. Indeed, several lines of evidence support this assumption. First, immunofluorescence analysis of infected cells demonstrated ApoE colocalization with NS2, core, E2, and NS5A. The latter showed the weakest colocalization with ApoE, and it occurred only upon solubilization of the ER membrane. Second, coimmunoprecipitation assays revealed tight interaction between ApoE and HCV envelope proteins that resisted even high stringency, whereas coprecipitation of ApoE with NS5A or NS2 was lost under these conditions. Consistent with these results, it has been reported that the sole expression of HCV envelope glycoproteins is sufficient to produce empty lipoviroparticles composed of E1/E2 and triglyceride-rich lipoproteins, indicating an intrinsic capacity of these viral proteins to associate with lipoproteins (60). In support of this assumption, we observed an intracellular interaction of ApoE with E1 and E2. We note that an interaction between ApoB and ApoE with E1, but not E2, has been reported (61). However, the authors of this study used primarily an ELISA-based read-out to compare ApoB and ApoE binding to E1 and E2, whereas we used comparative coprecipitation analysis and HCV-infected cells. Thus, the use of very different experimental conditions might account for these apparent discrepancies.

While the manuscript was in preparation, Boyer and colleagues also reported an association of HCV glycoproteins with ApoB and ApoE (62). Although our results are in agreement with that report, neither the determinants in E1/E2 nor the kind of interaction with ApoE has been studied. Moreover, mechanistic aspects of how ApoE is associated with HCV particles have not been addressed. In the present study, we show that the E2 TMD is required for ApoE-E2 interaction. We note that ApoE is able to interact with phospholipid head groups and is known to undergo large conformational changes in the presence of lipids (16). Moreover, it was shown that HCV envelope proteins reside in lipid rafts, and replacing the TMD of E1 and E2 by the C-terminal domain of vesicular stomatitis virus glycoprotein disturbs this localization (17). Thus, one attractive possibility is that localization of HCV envelope glycoproteins in lipid rafts facilitates the interaction with ApoE. It is tempting to speculate that HCV assembly occurs at lipid rafts of the ER membrane and fulfills a dual role: as specialized membrane compartments, allowing the concentration of viral proteins at a distinct site, and as a source of cholesterol and sphingolipids that are part of infectious HCV particles (12, 17).

In conclusion, we show that ApoE interacts with HCV envelope glycoproteins, presumably by protein-protein interaction. This association is dispensable for the envelopment of virions but appears to be required for lipidation of virus particles, contributing to their maturation and, ultimately, secretion. These results provide novel and important mechanistic insights into how ApoE contributes to the assembly of HCV particles and underscore the remarkable link between the life cycle of this virus and host cell lipid metabolism.