The heat shock cognate protein 70 is associated with hepatitis C virus particles and modulates virus infectivity


  • Potential conflict of interest: Nothing to report.


There is growing evidence that virus particles contain host cell proteins. These proteins may provide viruses with means to evade the immune system or with mechanisms for cell entry and release. A proteomic analysis performed on highly purified hepatitis C virus (HCV) J6/JFH virions identified the heat shock cognate protein 70 (HSC70) as part of the viral particles. These results were further validated via immunogold electron microscopy. The HSC70 interaction HPD motif was found present on the E2 envelope of the J6/JFH strain, as well as in over 50% of genotype 2 clinical HCV isolates. In addition, HSC70 was found associated with viral particles from an HCV genotype 2a–infected patient. Preincubation of HCV particles with anti-HSC70 antibodies decreased viral infectivity. Within infected cells, colocalization of HSC70 with the HCV core and E2 proteins was observed around lipid droplets. Reduction of HSC70 expression using an RNA interference approach decreased the volume of lipid droplets as well as viral release without affecting HCV replication levels. Conclusion: These results suggest that HSC70 modulates HCV infectivity and lipid droplet–dependent virus release. (HEPATOLOGY 2009.)

Hepatitis C virus (HCV) is an enveloped, positive-strand RNA virus and is the only member of the Hepacivirus genus within the family Flaviviridae.1 HCV often causes persistent infection in humans, a serious condition that is associated with chronic liver disease, cirrhosis, and hepatocellular carcinoma, and therefore represents a major issue of public health worldwide.2 Until recently, HCV research has been hampered by the lack of a robust cell culture system that recapitulates the complete viral life cycle. These limitations have been recently overcome through the use of the JFH genotype 2a strain of HCV, which permits propagation of infectious HCV particles in cell culture at high yields.3–5 The particles produced in this system have been shown to be infectious in vivo.6 Therefore, this cell system allows for the study of the viral life cycle and for the characterization of the composition of the HCV virion. In this study, we show that the heat shock cognate protein 70 (HSC70), also named HSPA8, is associated with HCV particles and is likely to play an important role in the virus life cycle.


BSA, bovine serum albumin; HCV, hepatitis C virus; HSC70, heat shock cognate protein; IgG, immunoglobulin G; MS, mass spectrometry; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; SEM, standard error of the mean; siRNA, small interfering RNA.

Materials and Methods

Cell Culture and HCV Infection.

Huh7.5 cells7 (kindly provided by Dr. Charles Rice, Rockefeller University) were cultured in Dulbecco's minimal essential medium (Invitrogen, Carlsbad, CA) supplemented with 10% dialyzed fetal bovine serum and 1% penicillin-streptomycin (Invitrogen) to minimize cytopathic effects. Viral stocks were generated via transfection of in vitro transcripts of the HCV J6/JFH genotype 2a strain (kindly provided by Dr. Charles Rice) using DMRIE-C (Invitrogen). Six hundred thousand cells were then infected with HCV at a dose of 5.6 × 103 TCID50/mL. All subsequent experiments were performed between 7 and 14 days postinfection.

Purification of HCV Virions.

Thirty-six million of infected (day 10 postinfection) or uninfected Huh7.5 cells were seeded at a density of 0.02 million/cm2 and grown for 4 days as described above. Supernatants were centrifuged at 8,000g for 15 minutes at 4°C, filtered through 0.45-μm membranes, layered onto a 20% sucrose cushion in TNE (10 mM Tris, 150 mM NaCl, 2 mM ethylene diamine tetraacetic acid) and ultracentrifuged at 27,000 rpm for 4 hours at 4°C. Pellets were then resuspended in 1 mL of TNE, layered at the top of a 10% to 60% sucrose linear gradient, and submitted to isopycnic ultracentrifugation for 16 hours at 31,200 rpm at 4°C. Sixteen fractions (750 μL each) were then harvested from the top of the gradient. Fractions containing the highest HCV RNA signals (typically fractions 10 to 13) as evaluated via real-time polymerase chain reaction (PCR) and their uninfected counterparts were pooled and dialyzed against TNE overnight at 4°C. Fractions were then concentrated 10- to 20-fold in YM-3 concentration devices (Centricon; Millipore, Billerica, MA) and further processed for electron microscopy applications. For mass spectrometry (MS)-based identification of virus bound proteins, a second sucrose gradient purification was performed.

Mass Spectrometry.

HCV-infected or uninfected samples prepared as described above were separated via one-dimensional sodium dodecyl sulfate–polyacrylamide gel electrophoresis and in-gel digested with trypsin (Promega, Madison, WI) in 200 mM sodium bicarbonate at 37°C overnight. Peptides were extracted twice with 10% acetonitrile and 10% formic acid and analyzed using nanoflow capillary LC-ESI/OrbiTrap MS/MS. The MS/MS spectra were searched against the human International Protein Index database (version 3.14) using the database search program X!Tandem. Peptide and protein identifications were then analyzed using PeptideProphet8 and ProteinProphet9 programs, respectively, to estimate statistical confidence of the protein identifications.

Electron Microscopy and Immunogold Labeling.

Fractions enriched in HCV particles were filtered through 1-μm microspin columns (Falcon; BD Biosciences, San Jose, CA) and deposited onto glow-discharged 200-mesh nickel grids (Ted Pella, Redding, CA) using an Airfuge centrifugation device (Beckman Coulter, Fullerton, CA) at 26 lbs/inch2 for 30 minutes. Preparations were negatively stained using either 1% Nanovan (Electron Microscopy Sciences, Hatfield, PA) in phosphate-buffered saline (PBS). For immunogold labeling, virus-containing grids were fixed with 2% paraformaldehyde in PBS for 5 minutes, washed, and incubated with the primary antibodies anti-HSC70 (Santa Cruz Biotechnology, Santa Cruz, CA, catalog number sc-1059) and normal goat immunoglobulin G (IgG) (Santa Cruz Biotechnology, catalog number sc-2028), both at a final concentration of 40 μg/mL for 16 hours, and anti-HCV E2 (kindly provided by Dr. Steven Foung, Stanford University) or an irrelevant isotype-matched antibody, both at a final concentration of 100 μg/mL for 16 hours at 4°C. Grids were washed and incubated with gold-conjugated rabbit anti-human (diameter of gold particles: 6 nm) or mouse anti-goat (diameter of gold particles: 15 nm) secondary antibodies for 1 hour at room temperature before being negatively stained with 1% Nanovan (Electron Microscopy Sciences) in PBS for immuno-electron microscopy. Samples were visualized under a Jeol 1020 electron microscope equipped with a Gatan digital camera.

Patient serum-derived HCV particles (genotype 1a/2a)10 were mounted onto electron microscope nickel grids by adsorption for 2 minutes of purified HCV pellet (120 μg of protein, corresponding to 2.2 × 105 copies of HCV RNA). Nonspecific reactive sites on the grids were blocked with 0.5% bovine serum albumin (BSA) for 20 minutes at room temperature. The grids were then incubated overnight at 4°C with the primary antibodies (normal goat IgG or anti-HSC70 antibody), both diluted 1/5 in 0.05 M Tris-HCl buffer (pH 7.4) containing 0.5% BSA, and washed once with 0.05 M Tris-HCl buffer (pH 7.4, then pH 8.2). The grids were then incubated with 0.5% BSA in Tris-HCl buffer (pH 8.2) for 5 minutes at room temperature, and incubated with 15-nm gold-conjugated rabbit anti-goat secondary antibodies diluted 1/25 in Tris-HCl buffer (pH 8.2) for 30 minutes at room temperature. The grids were washed in Tris-HCl buffer (pH 8.2, then pH 7.4), fixed with 2% glutaraldehyde in Tris-HCl buffer (pH 7.4), and finally negatively stained with 4% phosphotungstic acid for 30 seconds and examined in a Jeol 1400 electron microscope equipped with a Gatan digital camera (Orius type).

Neutralization Assays.

Supernatants harvested from infected cultures 12 days postinfection (infectivity of ≈5 × 103 TCID50/mL) were incubated overnight with anti-HSC70 (Santa Cruz Biotechnology, catalog number sc-1059), normal goat IgG (Santa Cruz Biotechnology, catalog number sc-2028), anti-HCV E2, or the irrelevant isotype-matched RO4 antibody (kindly provided by Steven Foung, Stanford University) at a final concentration of 5 μg/mL. Infection of naïve Huh7.5 cells seeded the day before at a density of 0.01 million/cm2 was performed for 3 hours. Cells were washed three times with complete medium and grown for 3 days before total RNA extraction using Trizol reagent (Invitrogen). In a separate set of experiments, cells seeded at 0.02 million/cm2 and cultured overnight were incubated with 10 μg/mL anti-HSC70 (Santa Cruz Biotechnology, catalog number sc-1059), anti-CD81 (BD Biosciences, catalog number 555675), or control IgG for 1 hour at 37°C. Cells were then incubated for 3 hours at 37°C with viral supernatants harvested from infected cultures 12 days postinfection. Following the infection, cells were washed three times with complete medium and grown for 24 hours before total RNA extraction using the RNeasy Mini Kit (Qiagen, Valencia, CA).


Cells seeded at a density of 0.01 million/cm2 in 8-well chamber slides (Falcon; BD Biosciences) or coverslips for 3 days were washed twice in ice-cold PBS, fixed in −20°C methanol for 1 minute and in 2% paraformaldehyde for 10 minutes. Cells were then permeabilized for 30 minutes in 0.1% Tween-20 in PBS and blocked for 30 minutes in 3% BSA/0.1% Tween-20 in PBS. Incubation with primary antibodies was performed for 1 to 2 hours in 3% BSA/0.1% Tween-20 in PBS at room temperature. The following antibodies were used: anti-HSC70 (clone 1B5, Santa Cruz Biotechnology, catalog number sc-59560); anti-HCV core (clone 1851; ViroStat, Portland, ME), and anti-HCV E2 (kindly provided by Steven Foung, Stanford University), all at a concentration of 2 μg/mL. After three washes in 0.1% Tween-20 in PBS, bound primary antibodies were probed with goat anti-rat Alexa-680, goat anti-mouse Alexa-594, or goat anti-human Alexa-488 secondary antibodies (Molecular Probes, Invitrogen) at a concentration of 2 μg/mL for 1 hour at room temperature. After three washes, nuclei were counterstained with 0.25 μg/mL diamidinophenylindole. For lipid droplet analysis, cells were incubated after the secondary antibody staining with Bodipy 493/503 (Molecular Probes, Invitrogen) at a concentration of 10 μg/mL in the first wash for 10 minutes. Slides were sealed and visualized under a Zeiss Axiovert or an Olympus IX70 deconvolution microscope, with Z-sections taken every 200 nm. Deconvolution was performed using softWoRx software (Applied Precision, Issaquah, WA). Images obtained by selective projection of the most intense pixel values for each channel, thereby allowing the calculation of the largest diameter and the volume of each droplet, were generated using softWoRx software. Tridimensional images shown in the Supplementary Movie were obtained using Volocity software (Improvision, Waltham, MA). Quantification of fluorescence intensities was performed using ImageJ software (National Institutes of Health,

Quantification of HCV RNA.

Supernatants were centrifuged at 8,000g for 15 minutes at 4°C to remove cellular debris. Viral RNA contained in supernatants or fractions harvested from sucrose gradients was extracted and purified using the Qiamp Viral RNA kit (Qiagen, Valencia, CA). Total cellular RNA was extracted and purified using the Trizol reagent (Invitrogen). Samples were then submitted to DNAse digestion, reverse-transcription using random hexamers, and real-time PCR as described,11 using the following primer sequences: HCV forward, 5′-CGGGAGAGCCATAGTGGTCTGCG-3′, HCV reverse, 5′-CTCGCAAGCACCCTATCAGGCAGTA-3′; actin forward, 5′-TGGACTTCGAGCAAGAGATGG-3′, and actin reverse, 5′-GGAAGGAAGGCTGGAAGAGTG-3′.

HCV Infectivity Assay.

Cells (6,400 per well) were seeded the day before infection in 96-well plates precoated with poly-L-lysine (Sigma-Aldrich, St. Louis, MO), infected with serial 10-fold dilutions of the supernatants of interest for 96 hours, washed with ice-cold PBS, fixed for 10 minutes in methanol, treated with 3% hydrogen peroxide to quench endogenous peroxidase activity, permeabilized for 30 minutes in Tris-buffered saline/0.1% Tween-20, and blocked for 30 minutes in 3% BSA/Tris-buffered saline/0.1% Tween-20. Cells were then probed with primary anti-HCV capsid monoclonal antibody (clone 1851, ViroStat, Portland, ME) at a concentration of 2.5 μg/mL in 3% BSA/Tris-buffered saline/0.1% Tween-20 for 2 hours at room temperature. After three washes in 0.1% Tween-20 in PBS, bound primary antibodies were probed with rabbit anti-mouse secondary antibodies conjugated to horseradish peroxidase (Dako North America, Carpinteria, CA) at a concentration of 10 μg/mL for 1 at room temperature. After three additional washes in 0.1% Tween-20 in PBS and then in PBS alone, cells were stained with diamidinobenzidine using the DAB+ kit (Dako North America). Titers were calculated using the method of Reed and Muench12 as modified by Lindenbach et al.3 without further standardization as RNA quantification and cell observation by light microscopy demonstrated that HSC70 small interfering RNA (siRNA) treatment did not affect cell proliferation.

siRNA Experiments.

Infected Huh7.5 cells (10 days postinfection) were seeded the day before transfection of siRNAs at a density of 0.01 million/cm2. siRNAs targeting HSC70 or irrelevant sequences (Dharmacon; Thermo Fisher Scientific, Waltham, MA) were transfected at a final concentration of 25 nM using lipofectamine 2000 (Invitrogen). Quantification of HSC70 knockdown was performed using HSC70 primers as follows: HSC70 forward, 5′-TCTTGTGTGGGTGTTTTCCAG-3′; HSC70 reverse, 5′-GACATAGCTTGGAGTGGTTCG-3′, leading to an 81-bp amplicon. Primer sequences and PCR conditions were obtained from the Primer Bank Web site (Harvard University, Specificity of the primers was verified by post-run melting curve analysis and agarose gel electrophoresis. The medium was renewed 2 days after transfection. Three days after transfection, supernatant and intracellular HCV RNA and HSC70 mRNA levels, as well as HCV infectivity levels, were determined as described above.


Identification of HSC70 as an HCV Virion-Associated Protein.

An MS analysis of HCV particles was performed in order to identify host cell factors that associate with HCV virions. Performance in MS-based identification of proteins is inversely correlated with the complexity of the analyzed sample. In addition to increasing sample complexity, the cytopathic effect associated with replication of the J6/JFH strain may result in unspecific binding of irrelevant proteins to HCV particles. We observed that dialyzed serum minimized HCV-induced cytopathic effect as determined by the absence of microcellular debris and floating necrotic cells 15 days postinfection. Following sedimentation of HCV particles on sucrose gradient via isopycnic centrifugation, HCV RNA was monitored in all 15 collected fractions. A peak of viral RNA was observed at 1.18 g/mL sucrose (Fig. 1A). After dialysis and concentration of the selected fractions, particles were visualized with transmission electron microscopy. The great majority of the particles was negatively stained and did not uptake any electron microscopy stain demonstrating the integrity of the particles at the end of the purification protocol (Fig. 1B,C). Particles were spherical, although some distortion was occasionally observed probably due to the conditions of sample processing.

Figure 1.

Association of HSC70 with J6/JFH HCV particles produced in vitro. (A) Concentrated supernatants of infected Huh7.5 cells were submitted to isopycnic fractionation using a 10% to 60% linear sucrose gradient. HCV RNA levels were determined in each fraction. Density of each fraction was determined by weighing. Graph is representative of three independent experiments performed on three independent infections. (B,C) Particles harvested from fractions 11 and 12 after dialysis and concentration were deposited onto grids using the Airfuge centrifugation system and processed for immunogold labeling using (B) the anti-E2 monoclonal antibody CBH539 or (C) the anti-HSC70 K19 polyclonal antibody. Antibodies bound to HCV E2 and HSC70 were detected using anti-human (B) or anti-goat (C) secondary antibodies conjugated to 6-nm and 15-nm gold particles, respectively. Prior to visualization, grids were negatively stained using 1% Nanovan. Pictures are representative of two independent experiments performed on two independent infections.

To enhance confidence on protein identifications to be obtained via MS on purified HCV preparations, a second sucrose gradient–based purification was performed on sucrose gradient–purified HCV-infected as well as uninfected supernatants processed in parallel. Following separation via sodium dodecyl sulfate–polyacrylamide gel electrophoresis and in-gel tryptic digestion, MS analysis identified several host proteins. Identified proteins include apolipoprotein E, previously reported associated with HCV virions,14 thereby validating our experimental approach; claudin 6, a coreceptor for HCV15, 16 and HSC70. To provide further evidence for the presence of HSC70 in association with HCV virions, immunogold electron microscopy was performed. As a positive control for the nature of these particles, grids were stained with a human anti-E2 monoclonal antibody developed from an HCV-infected patient. Although the labeling yield was low, virions were decorated at their periphery with gold particles (Fig. 1B). No staining was observed using a human isotype-matched monoclonal antibody on the same preparation, although rare background labeling could be observed. Neither gold particles nor virus-like shapes were observed in uninfected control preparations processed in parallel (data not shown). Immunogold staining was performed using an anti-HSC70 primary antibody. The presence of this protein in association with the virus particles was confirmed, while no staining was observed with purified normal total IgG (Fig. 1C). The lower number of gold particles found per virion in the case of the anti-HSC70 antibody is likely due to the stronger steric hindrance induced by the 15-nm gold beads in comparison with the 6-nm gold beads used with the anti-E2 antibody. As for E2 labeling, not all virions were found to be HSC70-positive. Whether this is linked to an intrinsic heterogeneity of HSC70 expression on the viral particle or the low labeling yield frequently observed by immuno-electron microscopy is unknown. Taken together, these data demonstrate the presence of HSC70 in association with the HCV particle.

HSC70 Is Associated with Serum-Derived HCV Genotype 1a/2a Particles.

HSC70 is composed of an ATPase domain and a substrate-binding domain. The substrate-binding domain of HSC70 is known to interact with a large number of targets through the highly conserved HPD motif in the J-domain of these proteins.17, 18 A search performed against this motif in the protein sequence of the J6/JFH genome used in this study revealed the presence of the HPD motif at the position 593 of the HCV polyprotein, within the E2 region (Fig. 2). To determine if the presence of this tripeptide sequence was restricted to the J6 strain or was a more general feature present across HCV genomes and genotypes, alignments of these HPD motifs with the contents of the Los Alamos National Laboratory HCV database were performed, using the Epilign function (Fig. 2B). Relevant results were limited to genotypes 1a, 1b, 2a, and 2b because of the low number of viral sequences available from other genotypes. The HPD motif was found in half of the 2a and 2b sequences. This motif was also present in a few genotype 1, 3, and 5 sequences. These results suggest that HCV/HSC70 interactions are not limited to the particular strain used in this study and that HSC70 is likely to play a role in the biology of a significant subset of HCV clinical isolates.

Figure 2.

Presence of the HSC70-binding motif HPD in the E2 sequence of the J6/JFH strain, and prevalence among HCV clinical isolates. (A) Amino acid sequence of the HCV E2 glycoprotein (J6 strain; GenBank accession number AF177036). The HPD motif, present at positions 593-595 of the polyprotein, is underlined. (B) Prevalence of the HPD motif in HCV clinical isolates. Sequences were retrieved from the Los Alamos National Laboratory HCV Database (

We investigated the presence of HSC70 on virions purified from a genotype 1a/2a-coinfected patient.10 Immunogold labeling experiments confirmed the presence of HSC70 on the surface of the serum-derived HCV particles. None or only one gold particle per virion on the viral surface or at a distance could be observed with the control immunoglobulins (Fig. 3A). In contrast, an average of three gold particles (up to six) were observed per virion when incubated with the anti-HSC70 antibody (Fig. 3B). Omitting the use of primary antibodies abolished the detection of colloidal gold on the virus surface, suggesting that no unspecific binding of secondary gold-labeled antibody occurred (data not shown). These data demonstrate the presence of HSC70 in association with particles harvested from HCV-infected patients.

Figure 3.

Association of HSC70 with serum-derived HCV 1a/2a particles. Particles harvested from a genotype 1a/2a–infected patient were deposited onto grids by adsorption and processed for immunogold labeling. Electron micrographs using (A) normal goat IgG or (B) the anti-HSC70 K19 polyclonal antibody. Bound antibodies were detected using anti-goat secondary antibodies conjugated to 15-nm gold particles. Pictures are representative of three labeling experiments.

Neutralization of Viral Infectivity by Anti-HSC70 Antibodies.

In order to gain some insight into the role of HSC70 on the particles, we investigated if the infectivity of the particles could be inhibited by their preincubation with anti HSC70 antibodies. The human anti-E2 monoclonal antibody developed from an HCV-infected patient and used in the electron microscopy experiments described above reduced infectivity by 69 ± 6.7% (mean ± SEM) (P = 0.009) (Fig. 4A). Preincubation of the particles with a polyclonal antibody raised against the C-terminus (residues 600-646) of HSC70 significantly inhibited HCV infectivity by 49.7 ± 8.4% (mean ± SEM) (P = 0.02) (Fig. 4A) while preincubation of the cells with anti-HSC70 antibodies did not significantly modify HCV infectivity (P = 0.24) (Fig. 4B). Preincubation of the cells with anti-CD81 antibodies reduced infectivity by 72.7 ± 4.9% (mean ± SEM) (P <0.001) as expected (Fig. 4B). These results suggest that HSC70, present on HCV virions in a topology accessible to antibody binding, plays a role in early interactions of HCV with its target cells.

Figure 4.

Neutralization of viral infectivity by anti-HSC70 antibodies. (A) Naïve Huh7.5 cells were infected with virus previously incubated with the anti-E2 CBH5 antibody39 or an antibody raised against residues 600-646 of the HSC70 protein and their respective controls, washed and grown for 3 days before lysis. HCV RNA quantification is shown as the mean ± SEM of three independent experiments performed using virus produced from three independent infections. (B) Huh7.5 cells were preincubated with an anti-HSC70 antibody or an anti-CD81 antibody and subsequently infected with HCV virus. HCV RNA quantification was performed 24 hours after infection and is shown as the mean ± SEM of three independent experiments.

HSC70, HCV Core, and E2 Proteins Colocalize Around Lipid Droplets.

We wished to investigate which area of the cell is the main site of association of HSC70 with HCV particles. Lipid droplets are vesicles frequently observed in hepatocytes and are known to be important sites of viral assembly and budding in HCV-infected Huh7 cells. Almost all Huh7.5 cells contained lipid droplets as evidenced by Bodipy 493/503 staining, a neutral lipid-specific dye. A fraction of the HSC70 pool localized around lipid droplets in naïve Huh7.5 cells (Fig. 5A). We performed a triple immunolocalization of HCV core, E2, and HSC70 in infected Huh7.5 cells (Fig. 5B). No changes in HSC70 expression levels were observed upon HCV replication from day 0 to day 14 postinfection, neither at the transcriptional nor at the protein level (data not shown). HSC70 was distributed between the cytosol and the immediate periphery of lipid droplets. As expected, the HCV core protein predominantly localized around lipid droplets. As observed for HSC70, the E2 protein was equally distributed between some, but not all, lipid droplets and the cytosol. Most importantly, HSC70 colocalized with the core and E2 proteins or with the core protein alone around lipid droplets. These results were confirmed by a plot profile showing the fluorescence intensity of each protein analyzed plotted against a line delineating the equator of the colocalization site, the lipid droplet (Fig. 5C). Correlation coefficients were calculated for each couple of intensity values along the equator of the vesicle (R = 0.95 for core/E2; R = 0.92 for core/HSC70; R = 0.91 for E2/HSC70). These high correlation coefficients further demonstrated the colocalization of the HCV core, E2, and HSC70 proteins. To further evaluate the percentage of overlap of the three colocalizations around the entire lipid droplet, we performed a three-dimensional reconstitution of the droplet by stacking Z-axis deconvolution images (Supplementary Movie). This analysis revealed that all three proteins remained confined in the same areas of the immediate periphery of the droplet. The colocalization of core, E2, and HSC70 proteins at the immediate periphery of lipid droplets, an important site for viral assembly and release, suggests that this area of the cell is the main site of association of HSC70 with HCV particles undergoing morphogenesis and secretion.

Figure 5.

HSC70 colocalizes with the HCV core and E2 proteins on lipid droplets. (A) HSC70 localizes in the immediate vicinity of lipid droplets. Huh7.5 cells were stained with an anti-HSC70 monoclonal antibody, then with an Alexa-680–conjugated secondary antibody. The neutral lipid-specific dye Bodipy 493/503 was used to stain lipid droplets, and cells were counterstained with diamidinophenylindole. Merged images were obtained using the softWoRx software. Bar = 10 μm. (B) Triple colocalization of core, E2, and HSC70. HCV-infected cells (day 14 postinfection) were fixed, permeabilized, and stained with the anti-HCV core antibody clone 1851, the anti-E2 CBH5 antibody,39 and the anti HSC70 antibody raised against residues 540-646. Bound primary antibodies were then detected using Alexa-546 (red fluorescence)–, Alexa-488 (green fluorescence)–, and Alexa-680 (far-red fluorescence converted into false blue signal)–conjugated secondary antibodies for HCV core, HCV E2, and HSC70, respectively. Deconvoluted and merged images were obtained using softWoRx software. Images are representative of two independent experiments performed on two independent infections. Bar = 10 μm. (C) Statistical assessment of the colocalization of core, E2, and HSC70. Red (HCV core), green (HCV E2), and blue (HSC70) fluorescence intensities were plotted for each pixel across the lipid droplet shown (white dashed line) using the Plot Profile function of ImageJ software. Correlation coefficients for each couple of intensity values (core/E2, core/HSC70, E2/HSC70) are shown.

We also observed some aggregation of HSC70 specific to HCV-infected cells. A heavy HSC70 punctuate pattern was found restricted to approximately 50% of core positive cells (Fig. 6A). This result was quantitatively confirmed by measuring the size of the 20 most intense HSC70 cytosolic aggregates in infected as well as noninfected cells in a Z-section taken in the middle of the cell thickness. Whereas the 20 most intense punctuates in uninfected cells had an average size of 129 ± 45 nm, their counterparts in HCV-positive cells had a significantly higher size of 253 ± 62 nm (P < 0.001) (Fig. 6B). These large HSC70 aggregates did not colocalize with the core protein (Fig. 6A) or the E2 protein (data not shown). Remarkably, the localization of these large HSC70 aggregates and of the lipid droplets appeared to be mutually exclusive.

Figure 6.

Viral replication induces intracellular aggregation of HSC70. HCV-infected cells (day 14 postinfection) were fixed, permeabilized, and stained with the anti-HCV core antibody clone 1851 and the anti-HSC70 antibody raised against residues 540-646. Bound primary antibodies were then detected using Alexa-488 (green fluorescence)– or Alexa-546 (red fluorescence)–conjugated secondary antibodies for HCV core and HSC70, respectively. Before mounting, nuclei were counterstained with diamidinophenylindole. Deconvoluted and merged images were obtained using SoftWorx software. Images are representative of three independent experiments performed on three independent infections. Bar = 10 μm.

Intracellular HSC70 Expression Levels Affect the Volume of Lipid Droplets.

We investigated whether the lipid droplet compartment was affected by HSC70 expression levels. HSC70-specific siRNA-treated Huh7.5 cells were quantified for the number and the volume of lipid droplets at the individual cell level via quantitative immunofluorescence using deconvolution microscopy. Not all siRNA-treated cells are identically responsive to knockdown in a given monolayer, thereby enabling statistical evaluation of the influence of the knockdown at a unicellular level. HSC70 siRNA-treated cells were labeled with an anti-HSC70 antibody and stained with the neutral lipid-specific dye Bodipy493/503. A representative picture of HSC70high and HSC70low cell populations is shown in Fig. 7A. HSC70 expression levels in selected HSC70high and HSC70low cell populations (n = 30) define two statistically different cell populations (P < 0.001) (Fig. 7B). The volume and the number of lipid droplets per cell in these two cell populations were quantified using deconvolution microscopy. Although no statistically significant difference was observed in the number of lipid droplets per cell between HSC70high and HSC70low cell populations (data not shown), the average volume of lipid droplets per cell was 1.55 ± 0.14 μm3 (mean ± SEM) in HSC70high cells and 0.62 ± 0.05 μm3 (mean ± SEM) in HSC70low cells (P < 0.001) (Fig. 7C). These results suggest that HSC70 is involved in the biogenesis or the stability of the lipid droplets.

Figure 7.

Association between HSC70 expression levels and volume of lipid droplets. (A) Representative picture of HSC70 siRNA-treated Huh7.5 cells. Cells were stained with the neutral lipid-specific dye Bodipy 493/503, an anti-HSC70 monoclonal antibody, and counterstained with diamidinophenylindole. The merged image was obtained using softWoRx software. Bar = 10 μm. (B) HSC70 siRNA-treated cells from a randomly chosen area of the slide using the “panel” function of the microscope were grouped into HSC70high (n = 30) and HSC70low (n = 30) expression levels. The HSC70 fluorescence intensity values for the two groups were quantified at the individual cell level using ImageJ software. (C) The largest diameter of each lipid droplet for each cell of the two groups was measured using softWoRx and ImageJ softwares. Averages of droplet diameters were calculated for each cell. Values were plotted for the HSC70high and the HSC70low cell populations. The statistical significance of the results was evaluated using a paired Student's t test. Horizontal bars indicate average values.

Modulation of Viral Release by HSC70.

We then determined the effect of reducing HSC70 expression on HCV replication and release. Three days after siRNA transfection, HSC70 mRNA level was reduced by 63 ± 9.8% by HSC70-specific siRNAs in comparison with nontargeting sequences, as quantified by real-time PCR (P = 0.01) (Fig. 8A). As expected, supernatant and intracellular HCV RNA levels were strongly decreased by HCV-specific siRNA sequences (90.8 ± 1.8% and 85.2 ± 1.9%, respectively). Whereas no effect of HSC70 reduction was observed on intracellular HCV RNA levels, HCV RNA release was strongly inhibited (65.2 ± 5.3%) in HSC70 siRNA-treated cells in comparison with cells treated with nontargeting sequences (P < 0.001) (Fig. 8B). A concomitant decrease in infectivity (78 ± 9.6%) was observed (P = 0.015) (Fig. 8C). These data suggest that HSC70 does not modulate the intracellular replication levels of HCV, but instead is involved in the assembly or budding steps of the viral life cycle.

Figure 8.

HSC70 is involved in viral secretion. (A) Knockdown of HSC70 expression in HCV-infected Huh7.5 cells by RNA interference. Specificity of HSC70 siRNA sequences was verified using the BLASTN program ( (B) Secreted and intracellular HCV RNA levels of irrelevant or HSC70-specific, siRNA-treated, HCV-infected Huh7.5 cells. (C) Infectivity of supernatants. Results shown are the mean ± SEM of four independent experiments performed on four independent infections.


The optimized cell culture system used in this study allowed for the recovery of a single viral RNA peak at a sucrose density of 1.18 g/mL, in agreement with other studies using a closely related system4 or clinical specimens.10, 19 A virus may contain host proteins for the following reasons: the host protein is present at the site of assembly; the protein interacts with a viral protein and is swept up into the virion during budding, or its incorporation is needed to perform a specific function for the virus.20 In this study, we demonstrate that the constitutively expressed HSC70 associates to cell culture grown HCV virions and regulates at least two crucial steps of the viral cycle: budding/egress and entry. This observation is of clinical relevance, because we also observed the association of HSC70 with serum-purified HCV particles.

Members of the HSP70 family of chaperones are involved in numerous biological processes. Among these are the folding of newly synthesized proteins, the refolding of stress-denatured proteins, the disaggregation of protein aggregates, the translocation of organellar and secretory proteins across membranes, the assembly and disassembly of oligomeric structures, and the modulation of the biological activity and stability of regulatory proteins. HSC70 is a cytosolic member of the HSP70 family. Unlike other members of this family, HSC70 is constitutively expressed. Chaperone proteins in general and HSC70 in particular have been found to play roles in the life cycle of a variety of RNA and DNA viruses.21 HSC70 binds to rotaviruses and contributes to the virus entry,22, 23 associates with and promotes in vitro and in vivo assembly of polyomavirus capsid proteins,24, 25 is involved in the disassembly of the hexon capsid protein of adenoviruses after entry,26 activates the reverse transcriptase of the duck hepatitis B virus,27 and participates in hepatitis B virus morphogenesis through specific interaction with the L protein.28 Finally, HSC70 as well as several other heat shock proteins have been found incorporated into primate and murine lentiviral virions.29

The intact HSC70 protein was present on HCV particles. Indeed, the protein was identified from a sodium dodecyl sulfate–polyacrylamide gel electrophoresis gel band of molecular weight compatible with the molecular weight of the full-length protein. In addition, the peptide identified via MS corresponds to the extreme N-terminal region of the protein, and HSC70 was recognized via electron microscopy on the particles by an antibody raised against the C-terminus part of its sequence. The substrate-binding domain of HSC70 is known to interact with its targets via the highly conserved and essential HPD triad.17, 18 This motif was found in the E2 glycoprotein sequence of the J6/JFH chimeric strain used in this study. Localization of this triad within the E2 scaffold using the tridimensional model developed by Yagnik et al.30 showed its presence at one of the highest peaks of accessibility and hydrophily (positions 593-595) of the E2 sequence. The J6/JFH strain is known for its atypical replication potential and genetic characteristics. The fact that this motif was found not only in this strain but also in over 50% of the genotype 2 sequences, as well as in some genotype 1 sequences, argues in favor of the association of HSC70 with a significant percentage of clinical isolates, as well as a potential role of HSC70 in HCV pathogenesis.

The results of the neutralization experiments suggest that the C-terminal domain of HSC70 may be involved in HCV entry, although the steric hindrance caused by the anti-HSC70 antibody could mask another cluster of residues necessary for HCV–target cell interactions. HSC70 has been initially identified as an uncoating ATPase, an enzyme that releases clathrin from coated vesicles.31–33 Interestingly, the heavy chain of the clathrin protein was also identified in our proteomic analysis of the HCV particles. In addition, HCV entry is mediated by a clathrin-dependent endocytosis.34, 35 The association of HSC70 with HCV particles may therefore contribute to the virus release from endocytic vesicles to the cytosol of the target cell.

Lipid droplets are to date the only known sites for HCV morphogenesis and secretion.36 RNA interference–mediated depletion of HSC70 led to a decrease in the volume of lipid droplets suggesting that HSC70 is necessary for lipid droplet biogenesis or stability. Remarkably, HSC70 has been shown to be involved in autophagy, a function by which the cell catabolizes its own energetic sources37 and the lipid droplet is, with glycogen, the most abundant reserve of energy of the hepatocyte. RNA interference–mediated depletion of HSC70 also led to a decrease in HCV release without any effect on intracellular HCV RNA levels. Involvement in HCV release could constitute a general function for HCV particle–bound proteins. Indeed, it has been shown that apolipoprotein E not only associates with HCV, a result we confirmed in our study, but also modulates viral secretion.14 The same result has been found for apolipoprotein B,38 a protein we did not detect in our study, perhaps because the HCV-ApoB complex is particularly prone to dissociate upon centrifugation. The presence of HSC70 around lipid droplets and the presence of large HSC70 cytosolic aggregates appear to be mutually exclusive. The HCV-associated cytosolic aggregation of HSC70 may therefore represent aggregates resulting from misfolding of HCV proteins. It would be interesting to investigate whether the presence of these large cytosolic aggregates of HSC70 is indicative of an impairment of viral release at the unicellular level, while in contrast, the presence of areas of HSC70, core, and E2 colocalization around lipid droplets correlates with efficient HCV secretion.

Future experiments are needed to further evaluate the role of HSC70 in the life cycle of HCV of different genotypes.


We thank Dr. Charles M. Rice (Rockefeller University, New York, NY) for the gift of the HCV J6/JFH strain and Huh7.5 cells and Dr. Steven Foung (Stanford University, Stanford, CA) for the gift of CBH5 and RO4 antibodies. We are grateful to Bobbie Schneider (Electron Microscopy Core Facility, Fred Hutchinson Cancer Research Center, Seattle, WA), Simone Peyrol (Centre Commun d'Imagerie Laënnec (CeCIL), Université Lyon 1, IFR62 Lyon-Est, Lyon, France), and Dr. Julio Vazquez (Scientific Imaging Core Facility, Fred Hutchinson Cancer Research Center, Seattle, WA) for excellent technical assistance.