Rab5 and Class III Phosphoinositide 3-Kinase Vps34 Are Involved in Hepatitis C Virus NS4B-Induced Autophagy

Abstract

Autophagy has been shown to facilitate replication or production of hepatitis C virus (HCV); nevertheless, how HCV induces autophagy remains unclear. Here, we demonstrate that HCV nonstructural protein 4B (NS4B) alone can induce autophagy signaling; amino acid residues 1 to 190 of NS4B are sufficient for this induction. Further studies showed that the phosphorylation levels of S6K and 4E-BP1 were not altered, suggesting that the mTOR/S6 kinase pathway and mTOR/4E-BP1 pathway did not contribute to NS4B- or HCV-induced autophagy. Inhibition of Rab5 function by silencing Rab5 or overexpressing dominant-negative Rab5 mutant (S34N) resulted in significant reduction of NS4B- or HCV-induced autophagic vesicle formation. Moreover, the autophagy induction was impaired by inhibition of class III phosphoinositide 3-kinase (PI 3-kinase) Vps34 function. Finally, the coimmunoprecipitation assay indicated that NS4B formed a complex with Rab5 and Vps34, supporting the notion that Rab5 and Vps34 are involved in NS4B-induced autophagy. Taken together, these results not only reveal a novel role of NS4B in autophagy but also offer a clue to the mechanism of HCV-induced autophagy.

INTRODUCTION

Hepatitis C virus (HCV) infections are a growing public health burden, with more than 180 million people infected worldwide. A striking feature of HCV infection is its tendency toward chronicity, often of significant liver disease, including chronic hepatitis, cirrhosis, and hepatocellular carcinoma (48). HCV is a positive-stranded RNA virus and classified into six genotypes (20). Its 9.6-kb genome encodes a single polyprotein, which is proteolytically processed into structural proteins (core, E1, E2, and p7), primarily forming the viral nucleocapsid and envelope, as well as nonstructural proteins (nonstructural protein 2 [NS2], NS3, NS4A, NS4B, NS5A, and NS5B) (35). Nonstructural proteins NS3 to NS5B are components of the membrane-associated HCV replication complex (16). NS3 is a bifunctional protein containing protease and helicase/nucleoside triphosphatase (NTPase) activities, and NS4A serves as a cofactor for NS3 protease. NS4B protein is known to induce formation of the membranous web that serves as the site for viral RNA replication. NS5A is required for RNA replication; phosphorylation of NS5A plays an important role in the HCV life cycle. NS5B is the RNA-dependent RNA polymerase (39). Although the roles of HCV proteins have been investigated, there is a great need for more understanding of the virus-host interaction and critical cellular players in the HCV life cycle that could be harnessed for anti-HCV therapy.

Autophagy is a cellular response to a variety of stimuli, including nutrient depletion, hormone treatment, and viral or bacterial infection in eukaryotic cells (31). The autophagy pathway comprises stepwise processes, including initiation as a phagophore, elongation to form an autophagosome, and ultimately the formation of autolysosomes to degrade long-lived proteins and organelles (9, 12). The autophagic vesicle refers to an autophagosome, amphisome, or autolysosome (11).

Two functional groups, mammalian target of rapamycin (mTOR) complex and Beclin1-hVps34 (human Vps34 [vacuolar sorting protein 34]) complex, are primarily involved in the initiation step of autophagosome formation (40). mTOR is a serine/threonine protein kinase that regulates cell growth, protein synthesis, translation, and autophagy. mTOR is recognized as the negative regulator of autophagy induction by inhibiting the expression and/or activity of several crucial Atg proteins (24). Beclin1 and Vps34, a class III phosphoinositide 3-kinase (PI 3-kinase), are known to be positive regulators of autophagy, through initial nucleation and assembly of the primary autophagosome membrane (42). Recent studies have revealed that Beclin1 and Vps34 play roles in the central control of autophagic activity and other trafficking events through the formation of distinct protein complexes (13).

The small GTPase Rab5, a member of the Ras superfamily, cycles between active (GTP-bound) and inactive (GDP-bound) forms (49) to regulate intracellular membrane trafficking such as endocytosis and early endosome fusion (19). Rab5 has been demonstrated to play an important role in autophagy by inhibiting mTOR kinase activity (32) or forming the complex with Beclin1 and Vps34 (41). Previous studies have demonstrated that Rab5 is an NS4B-interacting protein and crucial for HCV replication (36, 46). However, how Rab5 affects HCV replication remains unclear.

HCV genotypes 1a, 1b, and 2a have been reported to induce autophagy, probably to facilitate viral replication or virus production or suppress innate immunity (1, 10, 26, 43, 47). It has been proposed that HCV induces autophagy via the unfolded protein response (UPR) (43). Although HCV core, E1/E2, NS4B, and NS5A are capable of inducing the UPR (3, 5, 15, 33, 50), which protein of HCV is responsible for induction of autophagy remains unknown. In this study, we found that HCV NS4B protein alone is sufficient to induce autophagy. Furthermore, the results of mechanistic studies showed that Rab5 and Vps34 participated in the induction of autophagy by NS4B or HCV, whereas the common mTOR pathway did not. Thus, we propose that NS4B may modify the membrane environment via Rab5 and Vps34, thereby contributing to the induction of autophagy by HCV.

MATERIALS AND METHODS

Plasmids, small interfering RNA (siRNA), short hairpin RNA (shRNA), and viruses.

The open reading frame of mammalian Atg8 homolog microtubule-associated protein light chain 3 (LC3) was obtained from Gendiscovery (clone identification [ID] 7189855) and amplified by PCR; the sequence was verified, and it was subcloned into pLKO_AS3W.GFP.bsd lentivirus-based vector (provided by National RNAi Core Facility [RNAi stands for RNA interference], Academia Sinica, Taiwan) to generate the green fluorescent protein-light chain 3 (GFP-LC3) expression construct pLKO_AS3W.GFP-LC3.bsd. mCherry-LC3 plasmid was kindly provided by Yueh-Hsin Ping (Institute of Pharmacology, Yang-Ming University). The open reading frames of HCV proteins were amplified by PCR with appropriate hemagglutinin (HA) tag-containing primers and cloned to produce the following expressing plasmids: pUI-HA-core, pCAG2-HA-E1E2, pCI-HA-NS3/4A, pUI-NS4B-HA, pCI-NS5A-HA, pCAG2-HA-NS5B. pCI-HA-GST, and JC1 plasmid encoding a chimeric genome of HCV J6CF/JFH1 were constructed as previously described (6). The truncation mutants of NS4B were amplified by PCR with appropriate HA tag-containing primers, fused with mCherry, and subsequently cloned to pUI vector to obtain pUI-T1-mCherry, pUI-T2-mCherry, and pUI-T3-mCherry.

GFP-Rab5 (wild type [WT]) and GFP-Rab5 (S34N) were kindly provided as previously reported (29). pCMV-Vps34 (CMV stands for cytomegalovirus) and pCMV-Beclin1 were purchased from Open Biosystems. The on-target plus smart pool Atg7 siRNA and scramble siRNA were purchased from Dharmacon. shRNAs targeting LacZ, Rab5, and Vps34 were obtained from the National RNAi Core Facility (Academia Sinica, Taiwan). The ID numbers of the effective clones are as follows: TRCN0000072229 for clone shLacZ, TRCN0000007974 for clone shRab5#1, TRCN0000273640 for clone shRab5#2, TRCN0000037794 for clone shVps34#1, and TRC0000296151 for clone shVps34#2. Lentivirus expressing GFP-LC3 or harboring shRNAs was produced by following the established protocol (http://rnai.genmed.sinica.edu.tw:88/en/Protocols.asp). HCV JC1 virus was produced and titrated based on the described method (51). The JC1 virus was used at an multiplicity of infection (MOI) of 0.1 in the experiment.

Antibodies and reagents.

Rabbit anti-HA and anti-ULK1 (ULK1 stands for UNC-51-like kinase) antibodies were purchased from Santa Cruz Biotechnology (sc-805 and sc-33182). Mouse antiactin antibody was obtained from Millipore (MAB1501). Rabbit anti-LC3 and anti-phospho-LC3 antibodies were purchased from Abgent (AP1802a and AP3301a). Mouse anticore antibody was obtained from Thermo Scientific (MA1-080). Rabbit anti-Rab5 (GTX108605) and anti-Beclin1 (GTX113039) antibodies were obtained from GeneTex, Inc. S6 kinase (S6K) (catalog no. 9202), phosphorylated S6K (p-S6K) (Thr389) (catalog no. 9205), 4E-BP1 (catalog no. 9452), and phosphorylated 4E-BP1 (p-4E-BP1) (Thr37/46) (catalog no. 9459) were from Cell Signaling Technology. Rabbit anti-Vps34 antibody was from Invitrogen (38-2100). The LysoTracker red and all Alexa Fluor-conjugated secondary antibodies used for immunofluorescence were procured from Molecular Probes (Invitrogen). CelLytic reagent, 3-methyladenine (3-MA), 4′,6′-diamidino-2-phenylindole dihydrochloride (DAPI), and thapsigargin (TG) were purchased from Sigma-Aldrich. M-PER mammalian protein extraction reagent was obtained from Thermo Scientific, and anti-HA affinity matrix was from Roche Diagnostics.

Cells and transfection.

Huh7.5 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 1% nonessential amino acids, 100 U/ml penicillin, and 100 μg/ml streptomycin. GFP-LC3/Huh7.5 cells that stably expressed GFP-LC3 were produced by transduction of cells with lentiviruses expressing GFP-LC3. Positive clones were selected in the presence of 10 μg/ml blasticidin and eventually maintained with 4 μg/ml blasticidin. For transfection of siRNA, DMRIE-C (liposome formulation of the cationic lipid DMRIE [1,2-dimyristyloxypropyl-3-dimethyl-hydroxy ethyl ammonium bromide] and cholesterol) (Invitrogen) was used according to the manufacturer's instructions. For exogenous expression of protein, two kinds of transfection reagents were used for different purposes. For lipidated LC3 detection by Western blotting, Huh7.5 cells were transfected with plasmid DNA using Lipofectamine 2000 (Invitrogen) to achieve a higher transfection efficiency. For observation of autophagic vesicle by indirect immunofluorescence staining, TransIT-LT1 (Mirus) was used to reduce cytotoxicity.

Immunofluorescence microscopy and quantification of cells showing LC3-positive vesicles.

For indirect immunofluorescence staining, cells were fixed with 4% paraformaldehyde for 20 min and incubated with the blocking buffer (containing 3% bovine serum albumin and 0.01% saponin) for 45 min and then probed with the indicated primary antibody overnight at 4°C. Secondary antibody incubation was done for 1 h with the appropriate Alexa Fluor-conjugated antibody. To visualize DNA, cells were counterstained with DAPI. Images were acquired using an LSM 510 or LSM 780 confocal laser scanning microscope (Zeiss). For quantification of cells showing LC3-positive vesicles, approximately 50 cells were counted, and the cells with more than 20 GFP-LC3-labeled puncta were labeled as having formed an autophagosome.

Western blot analysis and coimmunoprecipitation.

Cell lysates were prepared using CelLytic reagent, subjected to electrophoresis on an SDS-polyacrylamide gel, and transferred onto a Hybond-P membrane (Amersham Biosciences). The membrane was probed with the indicated primary antibodies and appropriate secondary antibodies, detected using an enhanced chemiluminescence detection kit (Thermo Scientific), and then exposed to X-ray film. For the coimmunoprecipitation assay, cell lysates were prepared using M-PER reagent and then incubated with anti-HA affinity matrix overnight at 4°C. The immunoprecipitated proteins were subjected to Western blot analysis.

Electron microscopy.

All electron microscopy procedures were done at 4°C or on ice. Cells were cultured on ACLAR embedding film (Electron Microscopy Sciences) and then rinsed with 0.1 M cacodylate buffer (0.1 M sodium cacodylate, 3.4% sucrose, pH 7.4), fixed with 2.5% glutaraldehyde-0.1 M cacodylate, washed with 0.1 M cacodylate buffer, postfixed in 1% osmium tetroxide, prestained with 1% uranyl acetate, dehydrated in a graded series of ethanol concentrations, and embedded in Spurr resin. Then, 100-nm sections were cut, stained with 5% uranyl acetate in 50% methanol, and viewed on a Tecnai G2 Spirit TWIN transmission electron microscope (FEI Company) with Gatan 794 MultiScan charge-coupled-device (CCD) camera.

RNA preparation and reverse transcription-PCR (RT-PCR).

Total cellular RNA was extracted by using High Pure RNA Isolation kit (Roche Diagnostics) according to the manufacturer's protocol. cDNA was synthesized with oligo(dT)₂₀ by using SuperScript III first-strand synthesis system (Invitrogen). To amplify XBP1 and actin mRNA, the following primers were used: for X-box binding protein 1 (XBP1), 5′-CCTTGTAGTTGAGAACCAGG-3′ (sense) and 5′-GGGGCTTGGTATATATGTGG-3′ (antisense); for actin, 5′-TCACCCACACTGTGCCCATCTACG-3′ (sense) and 5′-CAGCGGAACCGCTCATTGCCAATG-3′ (antisense).

RESULTS

NS4B alone is sufficient to induce autophagic vesicle formation.

The mammalian Atg8 homolog microtubule-associated protein light chain 3 (LC3) is a specific marker protein for monitoring autophagic vesicle formation by its vesicle formation and lipidation reaction (25). To facilitate the observation of autophagic vesicles by fluorescence microscopy, a line of Huh7.5 cells that stably expressed green fluorescent protein-tagged LC3 (GFP-LC3) was first established. The patterns of GFP-LC3 in Huh7.5 cells were diffuse; in contrast, the patterns in the cells treated with thapsigargin (TG), a known autophagic inducer (21), showed punctate or dot-like structures, characteristic of autophagic vesicles (Fig. 1A). Furthermore, upon hepatitis C virus (HCV) infection, the green fluorescent pattern also showed punctate morphology, resembling the pattern of autophagic vesicles induced by TG, indicating that GFP-LC3/Huh7.5 cells were suitable for studying autophagy.

Fig. 1.

Evaluation of autophagy induction by HCV proteins. (A) Indirect immunofluorescence analysis was performed on GFP-LC3/Huh7.5 cells, which were left untreated or treated with 150 nM thapsigargin (TG) for 40 h or infected with HCV JC1 virus for 3 days. GFP-LC3 signal (green), DAPI staining of nuclei (blue), and staining of HCV core protein (red) are shown. (B) GFP-LC3/Huh7.5 cells were infected with HCV JC1 or transfected individually with various plasmids expressing HA-tagged HCV proteins or GST (as a control). After 40 h, cells were processed for immunofluorescence staining with HA antibody or anti-HCV core antibody (red) and GFP-LC3 (green). (C) Quantitative presentation of the percentage of GFP-LC3 punctum-forming cells in the HCV core or HA-tagged protein-expressing cells. GFP-LC3 punctum-positive cells were defined as described in Materials and Methods. The data shown represent the means plus standard deviations (error bars) from two independent experiments. (D) Immunoblot analysis of HA-tagged protein, LC3, phosphorylated LC3 (P-LC3), and actin in the cell extracts of Huh7.5 cells that were treated with 150 nM TG or transfected with various plasmids expressing HA-tagged HCV proteins. The band intensities of LC3-II, P-LC3, and actin were quantified, and the relative LC3-II/actin and P-LC3/actin ratios are shown below the blots. The positions of molecular mass markers (in kilodaltons) are shown to the right of the gel.

To assess which viral protein could trigger autophagy, plasmids expressing various HA-tagged HCV proteins were transfected individually into GFP-LC3/Huh7.5 cells. As shown in Fig. 1B, the punctate GFP pattern was clearly observed in the cells expressing NS4B. By calculating the percentage of the overexpressed cells containing GFP-LC3 puncta, we found that approximately 91% of NS4B-expressing cells showed autophagosome formation (Fig. 1C). None of the other HCV proteins induced significant formation of puncta. Upon autophagosome formation, LC3 is known to be converted from the cytosolic form (LC3-I) to the lipidated, autophagosome-associated form (LC3-II), which has faster mobility on SDS-polyacrylamide gels. Thus, to further verify that NS4B triggers autophagy, Western blot analysis was used to monitor the LC3-II level. Since LC3-II tends to be much more reactive than LC3-I with most antibodies, we measured the amount of LC3-II relative to actin to reflect the degree of autophagosome formation in each sample (38). The amount of LC3-II was most abundant in TG-treated cells (Fig. 1D). Among the HCV proteins, NS4B induced the highest level of LC3-II (Fig. 1D). The status of LC3 phosphorylation, which is known to be reduced when autophagy is induced (7), was also examined. The decreased phosphorylation of LC3 was noticeable in TG-treated and NS4B-transfected cells (Fig. 1D).

To rule out the possibility that induction of autophagy by NS4B is due to endoplasmic reticulum (ER) stress which is elicited by overloaded protein expression, various amounts of NS4B plasmids were transfected into GFP-LC3/Huh7.5 cells for observation of autophagic vesicle formation. Compared to control cells expressing glutathione S-transferase (GST), cells expressing NS4B showed autophagic vesicle formation (Fig. 2A). Notably, the punctate GFP-LC3 pattern was observed even at the lowest expression level of NS4B (Fig. 2A and B). The effects of various amounts of NS4B proteins on autophagic vesicle formation were further demonstrated by the formation of lipidated LC3. As shown in Fig. 2C, higher levels of LC3-II were detected in NS4B-expressing cells than in the GST-expressing cells.

Fig. 2.

Induction of autophagic vesicle formation by NS4B. (A) GFP-LC3/Huh7.5 cells were transfected with various amounts of HA-GST or HA-NS4B plasmids as indicated. After 40 h, the cells were processed for indirect immunofluorescence analysis. GFP-LC3 signal (green) and staining of HA-tagged protein (red) are shown. (B) Quantitative presentation of the percentage of GFP-LC3 punctum-forming cells in the HA-tagged protein-expressing cells. The data shown represent the means plus standard deviations from two independent experiments. (C) Huh7.5 cells were transfected with various amounts of HA-GST or HA-NS4B plasmids as indicated (−, none). After 40 h, the cell extracts were harvested and analyzed by Western blotting with the indicated antibodies. The densitometric LC3-II/actin ratios are shown below the blot. (D) GFP-LC3/Huh7.5 cells were transfected with 100 μM siAtg7 or scramble siRNA (as a control). After 30 h, the knockdown cells were reseeded, transfected with NS4B or infected with JC1 virus at an MOI of 0.5 for an additional 48 h, and processed for indirect immunofluorescence analysis. GFP-LC3 signal (green) and staining of HA-NS4B or HCV Core protein (red) are shown. The white arrows indicate the GFP-LC3-harboring cells that simultaneously expressed NS4B or HCV core. The quantitative presentation of the percentage of GFP-LC3 punctum-forming cells was shown in the graph to the right of the micrographs. (E) GFP-LC3/Huh7.5 cells were treated with 150 nM TG or infected with JC1 virus or transfected with HA-NS4B. After 40 h, indirect immunofluorescence staining was performed. GFP-LC3 signal (green), LysoTracker red (red), DAPI staining of nuclei (blue), and HA-tagged NS4B expression (white) are shown. (F) Huh7.5 cells were transfected with NS4B-GFP for 2 days, and NS4B-GFP-expressing cells were sorted using fluorescence-activated cell sorting (FACS) and processed for transmission electron microscopy (TEM). For HCV-infected cells, Huh7.5 cells were infected with JC1 virus for 6 days and processed for TEM. Representative images of naïve (left), NS4B-transfected, and JC1 virus-infected Huh7.5 cells are shown. The black arrows indicate autophagic vesicles. N, nucleus; M, mitochondria; LD, lipid droplet. Bars, 1 μm.

To verify that the phenomenon of LC3 punctum formation was caused by autophagic signaling instead of merely reflecting NS4B-induced membrane alteration, the knockdown effect of Atg7, one of the crucial factors for autophagosome formation by lipidation of LC3 (27), on induced autophagic vesicle formation was assessed. Figure 2D showed that the induction of GFP-LC3 puncta by NS4B overexpression or HCV infection was significantly inhibited upon silencing Atg7. Moreover, by staining GFP-LC3 cells with LysoTracker red, which can be used for staining lysosomes, we found that most of the NS4B- and HCV-induced autophagic vesicles could not colocalize with lysosomes, consistent with published findings (43); in contrast, TG treatment led to the colocalization of the induced autophagic vesicles with lysosomes as predicted (Fig. 2E). To further confirm that NS4B induces autophagy, we observed the formation of autophagic vesicles by electron microscopy (EM). As shown in Fig. 2F, the closed double-membraned vesicles with a diameter of 300 to 900 nm, resembling autophagic vesicles, were detected in NS4B-transfected and HCV-infected Huh7.5 cells but rarely in naïve Huh7.5 cells. Taken together, these results support the conclusion that NS4B alone can induce autophagy signaling.

Amino acids 1 to 190 of NS4B are responsible for induction of autophagy.

To determine which region of NS4B is crucial for inducing autophagy, various NS4B deletion mutants were used for assays. On the basis of the secondary structure of NS4B as predicted by the TMHMM software (Fig. 3A), we constructed NS4B mutants, including T1 (amino acids [aa] 1 to 63), covering the N-terminal α helix domain; T2 (aa 64 to 190), composed of four transmembrane domains; and T3 (aa 191 to 261), covering the C-terminal α helix domain. All the mutants were fused to the red fluorescent protein mCherry to facilitate fluorescence microscope observation. As shown in Fig. 3B, T1-mCherry and T2-mCherry led to autophagic vesicle formation. Notably, T2-mCherry-expressing cells induced more puncta than T1-mCherry-expressing cells; this result is consistent with the quantification data, showing more cells containing GFP-LC3 puncta in T2-mCherry (∼80%) than in T1-mCherry (∼55%) (Fig. 3B and C). For comparison, mCherry and T3-mCherry did not induce autophagy (Fig. 3B and C). Additionally, the results of immunoblotting analysis indicated that, consistent with the fluorescence microscopic evidence, expression of T1-mCherry and T2-mCherry enhanced lipidated LC3-II formation (Fig. 3D). In conclusion, the first 190 amino acids of NS4B are accountable for its induction of autophagy.

Fig. 3.

Amino acids 1 to 190 of NS4B mediate autophagosome formation. (A) Schematic representation of NS4B truncation mutants. The numbers in parentheses after T1 to T3 are the amino acids (e.g., T1 consists of amino acids [aa] 1 to 63, while T2 consists of aa 64 to 190). (B) GFP-LC3/Huh7.5 cells were transfected individually with mCherry, T1-mCherry, T2-mCherry, or T3-mCherry. After 40 h, the cells were fixed, stained with DAPI, and observed by fluorescence microscopy. The left panels show GFP-LC3 signals, the middle panels show mCherry signals, and the right panels show the merged images of the two. (C) Quantitative presentation of the percentage of GFP-LC3 punctum-forming cells in the protein-expressing cells. GFP-LC3 punctum-positive cells were defined as described in Materials and Methods. The data shown represent the means plus standard deviations from two independent experiments. (D) Huh7.5 cells were transfected with mCherry or NS4B truncation mutants as indicated. After 24 h, the cell extracts were harvested, subjected to 15% SDS-PAGE, and processed for Western blot analysis with the indicated antibodies. The positions of molecular mass markers (MM) (in kilodaltons) are shown to the right of the top gel. The predicted molecular masses of the mutants follow: T1-mCherry, 37 kDa; T2-mCherry, 42 kDa; and T3-mCherry, 38 kDa.

NS4B- or HCV-induced autophagy is not via the mTOR/S6K and mTOR/4E-BP1 pathways.

To dissect the mechanism of induction of autophagy by HCV or NS4B, we first examined whether HCV or NS4B could induce the unfolded protein response (UPR). As shown in Fig. 4A, NS4B and HCV activated the IRE1-XBP1 pathway, which is one of the UPR pathways, as indicated by splicing of Xbp1 mRNA. Since the stress response can trigger autophagy via an mTOR-dependent pathway (24), we next addressed whether the autophagy induced by NS4B or HCV is mediated via a common mTOR-dependent pathway by examining the phosphorylation levels of its substrates, ribosomal S6 protein kinase (S6K1, also known as p70S6K) and eukaryotic initiation factor 4E-binding protein 1 (4E-BP1) at Thr³⁸⁹ and Thr^37/46, respectively. By treating Huh7.5 cells with TG, which is known to induce mTOR-dependent autophagy (21), the phosphorylation levels of S6K and 4E-BP1 were decreased as predicted (Fig. 4B and C). In contrast, upon exogenous expression of NS4B or infection with HCV, phosphorylation levels of S6K and 4E-BP1 were not significantly changed in comparison to the mock-treated control (Fig. 4B and C). To rule out the possibility that the phosphorylated S6K and 4E-BP1 from untransfected cells might have compensated for the reducing signal, we transfected cells with NS4B-GFP and sorted the NS4B-GFP-expressing cells using fluorescence-activated cell sorting (FACS) for Western blot analysis. As shown in Fig. 4D, exogenous expression of NS4B-GFP did not cause a discernible effect on phosphorylation levels of S6K and 4E-BP1 compared to those in the GFP-expressing control. These results together suggest that NS4B- or HCV-induced autophagy is not mediated by mTOR/S6K and mTOR/4E-BP1 pathways.

Fig. 4.

NS4B- or HCV-induced autophagy is independent of mTOR/S6K and mTOR/4E-BP1 pathways. (A) Activation of IRE1-XBP1 signaling. Huh7.5 cells were transfected with NS4B or infected with HCV JC1 virus or treated with 150 nM TG. After 40 h, total cellular RNA was prepared and analyzed for XBP1 RNA by RT-PCR. The PCR products were separated by 2% agarose gel electrophoresis. Unspliced and spliced forms of XBP1 transcript are indicated. Actin RNA was also analyzed as an internal control. (B) Huh7.5 cells were transfected with NS4B or infected with HCV JC1 virus or treated with 150 nM TG. After 40 h, the cell extracts were harvested and processed for Western blot analysis with the indicated antibodies. (B) Immunoblotting for phosphorylated S6K (P-S6K) and total S6K. (C) Immunoblotting for phosphorylated 4E-BP1 (P-4E-BP1) and total 4E-BP1. (D) Huh7.5 cells were transfected with GFP or NS4B-GFP for 2 days, and the GFP-expressing cells were sorted by FACS and processed for Western blot analysis with the indicated antibodies. The control was TG-treated cells as mentioned above.

Rab5 is crucial for NS4B- and HCV-induced autophagy signaling.

Rab5, known to be involved in endocytosis and autophagy, has been identified as an NS4B-interacting protein and is important for HCV replication (36, 46). Accordingly, we examined whether Rab5 participated in NS4B- or HCV-induced autophagy signaling. Huh7.5 cells were transduced by lentiviral short hairpin RNAs (shRNAs) targeting Rab5, selected for knocked-down cells with puromycin, and then transfected with NS4B or infected with HCV. As shown in Fig. 5A, compared to the shLacZ control, knockdown of Rab5 significantly reduced LC3-II formation upon exogenous expression of NS4B. Moreover, silencing of Rab5 also decreased formation of LC3-II upon JC1 virus infection (Fig. 5B). Notably, a certain degree of LC3-II formation, which might be caused by lentivirus transduction was found in the mock-treated cells (Fig. 5A and B). Besides, the levels of NS4B and HCV, as detected by anti-hemagglutinin (anti-HA) and anti-core antibodies, respectively, were also diminished by Rab5 silencing, probably due to inhibition of endocytosis as a result of the knockdown of Rab5. To rule out the possibility that decrease of autophagy signaling was caused by the reduction of expression of NS4B or the inhibition of HCV entry, a dominant-negative (DN) Rab5 mutant, which is defective in GTP binding because of an S34N point mutation, was cotransfected with NS4B or cotreated with HCV infection. Overexpression of wild-type (WT) GFP-Rab5, which did not cause autophagic vesicle formation by itself, did not alter the NS4B-induced autophagic vesicle formation (Fig. 5C). In contrast, the DN GFP-Rab5 (S34N) altered the distribution pattern of mCherry-LC3 from punctate to homogeneous, suggesting that impairment of Rab5 function could prevent the induction of autophagy by NS4B (Fig. 5C). The quantification data showed that DN Rab5 (S34N) drastically abolished the formation of autophagic vesicle induced by NS4B (Fig. 5C). Upon HCV infection, the DN Rab5 (S34N), compared to the WT, also significantly inhibited autophagic vesicle formation caused by HCV (Fig. 5D). In sum, abolishing Rab5 function by RNA interference (RNAi) or DN Rab5 effectively impairs NS4B- or HCV-induced autophagic vesicle formation, suggesting that Rab5 is involved in NS4B- or HCV-induced autophagy signaling.

Fig. 5.

Rab5 is involved in NS4B- or HCV-induced autophagy. (A and B) Huh7.5 cells were transduced with lentiviruses expressing shRNA against Rab5 (two different clones denoted as shRab5#1 [lanes 2 and 5] and shRab5#2 [lanes 3 and 6]) or shLacZ (as control [lanes 1 and 4]) and then treated with puromycin for selection of shRNA-carrying cells. At 4 days posttransduction, 1 × 10⁵ of the transduced cells were reseeded onto a 12-well plate for further experiments. The knockdown cells were transfected with 0.5 μg of NS4B for 40 h (A) or infected with HCV JC1 virus at an MOI of 0.1 for 72 h (B). The cells were harvested, and the lysates were used for Western blot analysis with the indicated antibodies. The relative densitometric LC3-II/actin ratios are shown below the blots. Mock vector-transfected and noninfected cells were used as a control. (C and D) Huh7.5 cells were transfected with the combination of mCherry-LC3, HA-NS4B, GFP-Rab5 (WT), or GFP-Rab5 (S34N) as indicated. After 5 h, the culture medium was replaced with fresh medium (C) or HCV JC1-containing medium (D). After 2 days, indirect immunofluorescence staining with anti-HA antibody (blue) or anticore antibody (white) was performed. mCherry-LC3 (red) and GFP-Rab5 (green) are shown.

Class III PI 3-kinase Vps34 is crucial for NS4B- and HCV-induced autophagy signaling.

Class III phosphoinositide 3-kinase (PI 3-kinase) Vps34 is one of regulators that mediate autophagy and has been reported as a Rab5 effector (8); thus, we next assessed whether induction of autophagy by NS4B or HCV is dependent on Vps34. 3-Methyladenine (3-MA), an inhibitor of Vps34, effectively reduced punctate staining of GFP-LC3 in a dose-dependent manner, suggesting that PI 3-kinase (PI3K) signaling plays a role in NS4B- and HCV-induced autophagy (Fig. 6A and B). Since 3-MA is a broad-spectrum PI 3-kinase inhibitor, the effect of Vps34 knockdown on autophagy was then investigated to confirm that Vps34 mediates NS4B- or HCV-induced autophagy. Silencing Vps34 significantly reduced NS4B- and HCV-triggered LC3-II formation (Fig. 6C and D). Interestingly, in mock-treated controls, silencing Vps34 also slightly reduced LC3-II formation, which was probably caused by transduction of lentiviral shRNA. Taken together, these data showed that inhibition of Vps34 function by 3-MA or RNA interference prevented NS4B- or HCV-induced autophagosome formation, implying that Vps34 is important to the induction of autophagy by NS4B or HCV.

Fig. 6.

Vps34 participates in NS4B- and HCV-induced autophagy. (A and B) GFP-LC3 cells were transfected with HA-NS4B (A) or infected with HCV JC1 virus (B). After 5 h, cells were treated with 3-MA at various concentrations for an additional 40 h of incubation and then fixed for indirect immunofluorescence staining with anti-HA or anticore antibody. GFP-LC3 (green) and HA-tagged NS4B or HCV core (red) are shown. (C and D) Huh7.5 cells were transduced with lentiviruses expressing shRNA against Vps34 (two different clones denoted as shVps34#1 [lanes 2 and 5] and shVps34#2 [lanes 3 and 6]) or shLacZ (as control [lanes 1 and 4]) and then treated with puromycin for selection of shRNA-carrying cells. At 4 days posttransduction, 1 × 10⁵ of the transduced cells were reseeded into a 12-well plate for further experiments. The knockdown cells were transfected with 0.5 μg NS4B for 40 h (C) or infected with HCV JC1 virus at an MOI of 0.1 for 72 h (D). The cells were harvested, and the lysates were used for Western blot analysis with the indicated antibodies. The relative densitometric LC3-II/actin ratios are shown below the blots. Mock vector-transfected and noninfected cells were used as a control.

It has been documented that Rab5, Vps34, and Beclin1 could form a macromolecular complex (41). Thus, we examined whether NS4B associates with Rab5, Vps34, and Beclin1 by performing a coimmunoprecipitation assay. As shown in Fig. 7A, by using anti-HA antibody-conjugated agarose beads, HA-NS4B, but not HA-GST, coprecipitated with Rab5, Vps34, and Beclin1. In contrast, UNC-51-like kinase 1 (ULK1), one of the mammalian Atg1 homologs, could not coprecipitate with HA-NS4B, revealing the specificity of this interaction. Moreover, a reciprocal coimmunoprecipitation assay was also performed to confirm the formation of a macromolecular complex. As shown in Fig. 7B, GFP-Rab5 coprecipitated with HA-NS4B, Vps34, and Beclin1, compared to the GFP control. Taken together, these results suggested that NS4B may induce autophagy by forming a macromolecular complex containing Rab5, Vps34, and Beclin1.

Fig. 7.

NS4B associates with Rab5, Vps34, and Beclin1. (A) Huh7.5 cells were cotransfected with Vps34, Beclin1, and HA-GST or HA-NS4B. After 2 days, cell lysates were prepared using M-PER reagent and then incubated with anti-HA affinity matrix overnight at 4°C. The immunoprecipitated (IP) proteins were subjected to Western blot analysis with the indicated antibodies (α HA, anti-HA). (B) Huh7.5 cells were cotransfected with HA-NS4B, Vps34, Beclin1, and GFP or GFP-Rab5. After 2 days, cell lysates were prepared using M-PER reagent and then incubated with mouse anti-GFP antibody overnight at 4°C. The GFP-containing complex was coprecipitated with protein G beads and then subjected to Western blot analysis with the indicated antibodies (αGFP, anti-GFP).

DISCUSSION

Several viruses have been shown to utilize the autophagic machinery for their own replication and survival advantage (28, 30, 34). Recent studies have shown that some viruses can induce autophagy by their proteins. For example, the NSP4 protein of rotavirus (4), 2BC and 3A proteins of poliovirus (22), matrix protein 2 of influenza A virus (14), and HBx of hepatitis B virus (44) are responsible for the induction of autophagy. We have found that the HCV genotype 1b (HCV1b) subgenomic replicon, composed of the NS3-NS5B-coding region, could induce autophagy (data not shown), consistent with the previous report (43), implying that viral nonstructural proteins have the potential to trigger autophagy. In the present study, we clearly demonstrated that NS4B by itself was able to induce autophagy (Fig. 1 and 2). None of the other viral proteins, though capable of inducing the unfolded protein response (UPR), could induce autophagy. It is notable that autophagic vesicles of various sizes were observed in NS4B-overexpressed and HCV-infected cells by using immunofluorescence microscopy and electron microscopy (Fig. 1A and B and 2F). The size variation probably reflects autophagic vesicles encompassing different cytoplasmic macromolecules, including organelles and proteins. In addition, HCV-infected cells showed only modest GFP-LC3 puncta in subcellular distribution compared to NS4B-transfected cells, which showed more robust punctum formation (Fig. 1B and C). This difference was probably due to the overexpression scenario of NS4B. Therefore, though NS4B induced stronger autophagic signaling than the other HCV proteins, it should be noted that this individual expression may not accurately reflect the physiological context of HCV infection.

NS4B is known to be responsible for the formation of a membranous web, where the membrane-associated HCV RNA replication complex is located (11). Considering the characteristics of the membranous web and autophagosome, both of which are derived by membrane alterations and pivotal to HCV replication, NS4B likely plays a role in both functions. By using truncation mutants of NS4B, we found that T1 (aa 1 to 63) and T2 (aa 64 to 190), comprising the cytosolic N-terminal part and four transmembrane domains, could induce autophagy (Fig. 3). The transmembrane domain is frequently involved in triggering endoplasmic reticulum (ER) stress; therefore, it is conceivable that the transmembrane domain is sufficient for induction of autophagy. Both the N- and C-terminal domains of NS4B have been reported to mediate membrane association and are required for HCV replication (17, 18). However, the C-terminal domain of NS4B is insufficient for membranous web formation (2), consistent with our results that the C-terminal domain could not induce autophagy. Interestingly, though T3 (aa 191 to 261) has the highest expression level (Fig. 3D), it could not induce autophagy, strengthening the conclusion that autophagy induction by NS4B is not caused by overloading protein. Our findings that NS4B triggers autophagosome formation may contribute to the study of the novel functions of NS4B.

Though we demonstrated that NS4B is sufficient for induction of autophagy, it is difficult to assess whether NS4B is necessary for HCV-induced autophagosome formation by performing studies based on the depletion of NS4B, since HCV cannot replicate without NS4B. Instead, we addressed the role of NS4B in HCV-induced autophagy by delineating the molecular pathways by which NS4B and HCV induce autophagy. It has been recognized that the signaling pathways responsible for induction of autophagy vary according to cell type and stimulus. mTOR is a negative regulator of autophagy induction; the inhibition of mTOR complex 1 is sufficient to induce autophagy, indicating that it serves as a major gatekeeper of autophagy induction under normal conditions (37). Although we observed that NS4B and HCV activated the UPR (Fig. 4A), we unexpectedly found that neither NS4B nor HCV induced autophagy through the common mTOR-dependent pathway (Fig. 4).

The Rab family, which is involved in regulating vesicle budding, transport, and fusion with target membranes (45), includes more than 50 members. It has been reported that Rab1, -2, -5, -6, and -7 are enriched in an isolated NS4B-bound subcellular fraction, which is competent for HCV RNA synthesis; nevertheless, only Rab5 and Rab7 are important for HCV replication (36). Rab5 plays a role in the early endocytic pathway and initiation of autophagy, while Rab7 is involved in the late endocytic pathway and the final maturation of late autophagic vacuole (23). HCV is known to induce autophagosome but not autolysosome, the late autophagic vacuole, formation (43). We have investigated whether the NS4B-induced autophagic vesicle could complete its maturation step by fusing with lysosome to form an autolysosome and found that the autophagy induced by NS4B did not lead to autolysosome formation (shown in Fig. 2E). Accordingly, we focused on addressing the role of Rab5 in NS4B- and HCV-induced autophagy. On the basis of our findings that Rab5 is crucial for HCV-induced autophagy and other reports that autophagy is important for HCV replication (10, 43), we speculate that Rab5 may influence HCV replication by affecting autophagy signaling.

Furthermore, similar to Rab5 inhibition, inhibition of Vps34 function by 3-methyladenine (3-MA) treatment or short hairpin RNA (shRNA) knockdown blocked NS4B- or HCV-induced autophagosome formation. Notably, inhibition of Rab5 or Vps34 could not block autophagosome formation completely (Fig. 5 and 6), implying that other signal pathways may also be involved in NS4B- or HCV-induced autophagy. Besides, the amounts of HCV were reduced upon silencing of Rab5 and Vps34. There are two possibilities for this finding. One possibility is that suppressing Rab5 and Vps34 blocked autophagy and thereby inhibited HCV replication; the other possibility is that suppressing Rab5 and Vps34 blocked endocytosis and thereby reduced HCV uptake. The current results do not allow us to distinguish between these two possibilities. Nevertheless, it has been reported that knockdown of Beclin1, which forms a complex with Vps34 and is crucial for autophagy, drastically abolished HCV replication (10); this finding favors the first possibility. Finally, we demonstrated that NS4B could associate with Rab5 and the Vps34-Beclin1 complex (Fig. 7), strengthening the finding that both Rab5 and Vps34 are involved in NS4B-induced autophagy. The results of our studies showed that the molecular pathways involved in NS4B- or HCV-induced autophagy are similar; thus, we propose that NS4B can recruit the Rab5 and Vps34 complex to induce autophagy and contribute to induction of autophagy by HCV.