ABSTRACT
Hepatitis C virus (HCV) is highly dependent on cellular factors for viral propagation. Using high-throughput next-generation sequencing, we analyzed the host transcriptomic changes and identified 30 candidate genes which were upregulated in cell culture-grown HCV (HCVcc)-infected cells. Of these candidates, we selected Rab32 for further investigation. Rab32 is a small GTPase that regulates a variety of intracellular membrane-trafficking events in various cell types. In this study, we demonstrated that both mRNA and protein levels of Rab32 were increased in HCV-infected cells. Furthermore, we showed that HCV infection converted the predominantly expressed GTP-bound Rab32 to GDP-bound Rab32, contributing to the aggregation of Rab32 and thus making it less sensitive to cellular degradation machinery. In addition, GDP-bound Rab32 selectively interacted with HCV core protein and deposited core protein into the endoplasmic reticulum (ER)-associated Rab32-derived aggregated structures in the perinuclear region, which were likely to be viral assembly sites. Using RNA interference technology, we demonstrated that Rab32 was required for the assembly step but not for other stages of the HCV life cycle. Taken together, these data suggest that HCV may modulate Rab32 activity to facilitate virion assembly.
IMPORTANCE Rab32, a member of the Ras superfamily of small GTPases, regulates various intracellular membrane-trafficking events in many cell types. In this study, we showed that HCV infection concomitantly increased Rab32 expression at the transcriptional level and altered the balance between GDP- and GTP-bound Rab32 toward production of Rab32-GDP. GDP-bound Rab32 selectively interacted with HCV core protein and enriched core in the ER-associated Rab32-derived aggregated structures that were probably necessary for viral assembly. Indeed, we showed that Rab32 was specifically required for the assembly of HCV. Collectively, our study identifies that Rab32 is a novel host factor essential for HCV particle assembly.
INTRODUCTION
Hepatitis C virus (HCV) is the major causative agent of chronic liver diseases, including liver cirrhosis and hepatocellular carcinoma (1). Recent reports have shown that total global viremic HCV infections were estimated at 80 million cases, contributing to 350,000 deaths annually (2, 3). HCV has a positive, single-stranded RNA genome that is approximately 9.6 kb in length and encodes a large polyprotein of about 3,000 amino acids. The polyprotein is posttranslationally cleaved by both host and viral proteases into 10 different polypeptides including three structural (core, E1, and E2) and seven nonstructural (p7, NS2, NS3, NS4A, NS4B, NS5A, and NS5B) proteins (4). Previously, the combination of pegylated alpha interferon (IFN-α) and ribavirin has been a standard regimen to treat chronic hepatitis C. Currently, numerous direct-acting antivirals (DAAs) are available to treat HCV patients, with high cure rates. However, sustained virologic responses of new DAAs vary depending on HCV genotypes (5), and the sky-high prices of DAAs are too burdensome for the majority of HCV patients worldwide (6). Moreover, long-term treatment with DAAs may lead to the selection of drug escape mutants due to their low genetic barriers to drug resistance (7). Since HCV propagation and pathogenesis largely depend on host cellular machinery, host-targeting antivirals may have some advantages in terms of high genetic barriers to resistance and a potential for pan-genotypic antiviral activity (8).
Rab32 belongs to the Ras superfamily of small GTPases which contains at least 60 Rab genes in the human genome which play important roles in regulating membrane traffic by cycling between a GDP-bound form and GTP-bound form (9). Murine Rab32 and its homologue Rab38 were reported to be involved in the regulation of skin melanocyte pigmentation and organization of trans-Golgi networks (TGNs) as Rab32 silencing resulted in the mistargeting and subsequent degradation of key melanogenic enzymes, including tyrosinase and tyrosinase-related protein 1 (10). In Xenopus melanophores, Rab32 controls melanosome transport in a cyclic AMP (cAMP)-dependent protein kinase A (PKA)-dependent manner (11). Despite the ubiquitous expression of Rab32 in most human tissues (12, 13), the precise functions of Rab32 in nonmelanogenic cells and tissues are poorly characterized. In cell types other than melanocytes, such as COS7 and WI-38 fibroblasts, Rab32 was found to colocalize with mitochondria. In addition, Rab32 modulates targeting of PKA to mitochondrial and endoplasmic reticulum (ER) membranes and determines mitochondrial dynamics and apoptosis onset (13, 14). Furthermore, Rab32 has been demonstrated to be essential for the autophagic response in HeLa and COS7 cells (15). Recently, it has been reported that Rab32 increases lipid biosynthesis and autophagosome formation during the reprogramming process (16). Rab32 has also been involved in acute brain inflammation in mice (17). Moreover, Rab32 interacts with leucine-rich repeat kinase 2 (LRRK2) and regulates LRRK2 transport, implicated in Parkinson's disease (18). To date, the functional involvement of Rab32 in the HCV life cycle or HCV-induced pathogenesis has not been demonstrated.
In the present study, we demonstrate that HCV concomitantly upregulated Rab32 expression and induced conversion of the predominantly expressed GTP-bound Rab32 to GDP-bound Rab32, which resulted in the aggregation of Rab32 protein and thus made it less susceptible to cellular degradation machinery. We further show that GDP-bound Rab32 selectively interacts with HCV core protein and deposits core in ER-associated Rab32-derived aggregated structures in the perinuclear region that are likely to be viral assembly sites. Moreover, we demonstrate that Rab32 is specifically required for HCV particle assembly. Collectively, these data suggest that HCV may modulate Rab32 activity to generate the core protein-containing structures necessary for HCV virion assembly.
RESULTS
Rab32 level is increased in the context of HCV infection.In an attempt to identify host factors that play essential roles in HCV propagation, we previously employed high-throughput RNA sequencing (RNA-Seq) technology to characterize the genome-wide transcriptomic changes in cell culture-grown (HCVcc)-infected cells. By performing quantitative real-time PCR (qRT-PCR analysis), we ultimately verified that 30 host genes were markedly increased in the context of HCV infection (19). In the present study, we selected Rab32 for more elaborate characterization in order to delineate its possible functional involvement in regulating HCV propagation. To confirm the increase in Rab32 expression in HCVcc-infected cells, we measured Rab32 mRNA levels in Jc1-infected Huh7.5 cells at different time points. As expected, Rab32 mRNA was noticeably increased at day 2, and its level was doubled at day 6 in HCV-infected cells compared with the level in mock-infected cells (Fig. 1A). To investigate if the transcriptional level of Rab32 was also regulated by HCV infection, Huh7.5 cells were either mock infected or infected with Jc1. At 4 h postinfection, cells were further transfected with a luciferase (Luc) reporter plasmid consisting of nucleotides (nt) −643 to +260 of the Rab32 promoter, and then luciferase activity was analyzed at 2 days postinfection. Figure 1B shows that Rab32 promoter activity was significantly increased in HCV-infected cells. Consistently, the protein level of Rab32 was proportionally elevated during the course of HCV infection (Fig. 1C). We further verified that the Rab32 mRNA level in HCV-replicating primary human hepatocytes significantly increased compared with the level in the replication-defective control (Fig. 1D). Additionally, we also examined the Rab32 level in HCV subgenomic replicon cells derived from genotype 1b. We showed that both the mRNA level (Fig. 1E) and the protein expression level (Fig. 1F) of Rab32 in the HCV replicon cells were markedly higher than those in parental Huh7 or IFN-cured cells. These data suggest that an HCV nonstructural protein may be responsible for the upregulation of Rab32 in HCV-infected cells. Indeed, overexpression of HCV NS3 increased the mRNA level of Rab32 (data not shown). Together, these data clearly indicate that HCV upregulates Rab32 gene expression.
Rab32 level is increased in the context of HCV infection. (A) Huh7.5 cells were either mock infected or infected with HCV Jc1 for 4 h, and then mRNA levels of Rab32 were analyzed by qRT-PCR at the indicated time points. dpi, days postinfection. (B) Rab32 promoter activity is upregulated by HCV infection. A schematic diagram of Rab32 promoter construct is shown at the top of the panel. Huh7.5 cells were either mock infected or infected with Jc1. At 4 h postinfection, cells were further transfected with a Rab32-Luc promoter reporter plasmid. Luciferase activity was determined at 48 h postinfection (bottom). (C) Huh7.5 cells were either mock infected or infected with HCV Jc1 for 4 h. Total cell lysates were immunoblotted with the indicated antibodies at various time points. (D) Primary human hepatocytes (PHHs) were electroporated with either in vitro-transcribed HCV Jc1 RNA or a replication-deficient GNN mutant. At 4 days after electroporation, mRNA levels of Rab32 were analyzed by qRT-PCR. (E) Total cellular RNAs were extracted from parental Huh7, HCV replicon derived from genotype 1b, and IFN-cured cells, and the Rab32 mRNA level was analyzed by qRT-PCR. (F) Total cellular lysates harvested from Huh7, HCV replicon, and IFN-cured cells were immunoblotted with the indicated antibodies. Data represent averages from at least three independent experiments for panels A, B, D, and E. The asterisks indicate significant differences (*, P < 0.05; **, P < 0.01) from the value for the control. β-Gal, β-galactosidase.
Rab32 is predominantly present in the GTP-bound form in naive Huh7 cells.It has been shown that Rab32 in nonmelanogenic cell types partially localizes to the ER-mitochondrion contact sites and regulates certain key cellular events (13–15, 20). In this study, a subpopulation of Rab32 in Huh7 cells was also detected in mitochondrion- and ER-containing fractions (Fig. 2A). Because Rab GTPases cycle between the GDP- and GTP-bound states and are regulated by guanine nucleotide exchange factors (GEFs) and GTPase activating proteins (GAPs) that influence their subcellular localization and activities (21), it is important to study the nucleotide-dependent function of Rab32 protein in both naive hepatoma cells and in virus-infected cells. The nucleotide status of Rab32 in murine melanocytes has been frequently measured by employing vacuolar protein sorting 9 (VPS9) ankyrin repeat protein (VARP) as its first ankyrin repeat domain (ANKRD1) (Fig. 2B) specifically interacts with Rab32 in a GTP-dependent manner (22). However, in other cell types that do not produce melanosome and exhibit different Rab32 subcellular distributions (14, 15), the nucleotide status of Rab32 has not been characterized so far. For this purpose, we constructed a cDNA clone encoding the human ANKRD1 domain of VARP (VARP-ANKRD1) that had been amplified from Huh7 cells and attempted to take advantage of this construct to examine the nucleotide status of Rab32. First, we investigated the subcellular distribution of VARP-ANKRD1 in Huh7 cells by immunofluorescence assay. Notably, overexpression of VARP-ANKRD1 produced cytoplasmic speckle-like staining patterns which resembled membrane-bound structures (Fig. 2C). We further showed a significant overlap of these speckles and both the ER and mitochondria with high levels of colocalization coefficients, suggesting that VARP-ANKRD1 is preferentially present at the ER-mitochondrion contact sites. We further examined the staining pattern of VARP-ANKRD1 together with either the GDP-locked (T39N) or GTP-locked (Q85L) form of Rab32 in Huh7 cells. Rab32-T39N preferentially binds GDP because of the mutation in the conserved GTP-binding site, whereas Rab32-Q85L is permanently trapped in its GTP form since this mutation abolishes its intrinsic GTPase activity (13). It has been previously reported that Rab32-L188P is unable to anchor PKA due to the disruption of its secondary structure (13). We showed that Flag-tagged VARP-ANKRD1 colocalized with GDP-locked Rab32 (T39N) in the speckle-like structures (Fig. 2D). This also implied that GDP-bound Rab32 displayed a preferential association with the ER and mitochondria, as reported previously (14). In contrast, GTP-locked Rab32 (Q85L) was widely distributed throughout the cells and exhibited less overlapped signal with VARP-ANKRD1 (Fig. 2D). As a result, we postulated that human VARP-ANKRD1 specifically recognized Rab32-GDP in Huh7 cells. We further verified the nucleotide-dependent binding pattern of Rab32 with VARP-ANKRD1 by a glutathione (GSH) S-transferase (GST) pulldown assay. For this purpose, a GST-tagged VARP-ANKRD1 fusion protein was induced in Escherichia coli, and purified protein was analyzed on 10% SDS-PAGE gels (Fig. 2E, left panel). Huh7 cell lysates containing Flag-tagged Rab32 were treated with either GTPγS or GDP, and cell lysates were further precipitated with GST-tagged VARP-ANKRD1. We demonstrated that VARP-ANKRD1 specifically interacted with GDP-bound but not with GTPγS-bound Rab32 (Fig. 2E, right panel). Consistently, we showed that GST-tagged VARP-ANKRD1 specifically captured Rab32-T39N but not Rab32-Q85L from both HEK293T and Huh7 cells (Fig. 2F). We therefore concluded that VARP-ANKRD1 selectively interacted with GDP-bound Rab32. It was noteworthy that wild-type Rab32 protein overexpressed in Huh7 cells exhibited a barely detectable interaction with VARP-ANKRD1 in contrast to GDP-bound Rab32 (T39N) (Fig. 2F, right panel). Using Flag-tagged VARP-ANKRD1, we confirmed that VARP-ANKRD1 was coimmunoprecipitated with wild-type Rab32 (Rab32-WT), Rab32-T39N, and Rab32-L188P in HEK293 cells (Fig. 2G, left panel). However, the interaction between wild-type Rab32 and VARP-ANKRD1 was undetectable in Huh7 cells (Fig. 2G, right panel). These data suggest that Rab32 may exist predominantly in the GTP-bound form in naive Huh7 cells.
Rab32 is predominantly present in the GTP-bound form in naive Huh7 cells. (A) Naive Huh7 cells were fractionated into cytosol, mitochondrial, and microsomal fractions. The voltage-dependent anion channel (VDAC) protein was used as the mitochondrion-specific marker. Calnexin and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) represent ER and cytosol fractions, respectively. Mito, mitochondria; Micro, microsomes; Lysate, whole-cell lysate. (B) Schematic illustration of both human VARP and Rab32 protein domain structures. (C) VARP-ANKRD1 localizes to the ER and mitochondrial structures. Huh7 cells seeded on glass coverslips were transfected with a V5-tagged VARP-ANKRD1 expression plasmid. In order to label the ER, cells were incubated with CellLight ER-GFP reagent for 16 h at 10 h after transfection. For mitochondrial staining, cells were incubated with 100 nM MitoTracker for 40 min at 30 h after transfection. Cells were then fixed in 4% paraformaldehyde, and immunofluorescence staining was performed by using an anti-V5 monoclonal antibody. Cells were also counterstained with DAPI to label nuclei (blue). Higher magnifications of the boxed areas are shown in the cropped images. Quantification of colocalization between VARP-ANKRD1 and both the ER and mitochondrial marker was determined using the Pearson's correlation coefficient and the Manders' overlap coefficient (right panel). R, red; G, green; M1 and M2, Manders' coefficients. (D) Subcellular distribution of GDP- and GTP-locked forms of Rab32 and VARP-ANKRD1 in Huh7 cells. Huh7 cells seeded on glass coverslips were cotransfected with Flag-tagged VARP-ANKRD1 and either a V5-tagged Rab32-T39N or Q85L expression plasmid. At 36 h after transfection, cells were fixed in 4% paraformaldehyde, and immunofluorescence staining was performed by using an anti-V5 and an anti-Flag monoclonal antibody. Dual staining showed colocalization of Rab32-T39N and VARP-ANKRD1 as yellow fluorescence in the merged image. Cells were counterstained with DAPI to label nuclei (blue). Quantification of colocalization between VARP-ANKRD1 and both Rab32-T39N and Rab32-Q85L was determined using the Pearson's correlation coefficient and the Manders' overlap coefficient. R, red; G: green. (E) The left panel shows purification of GST-VARP-ANKRD1 protein. GST-VARP-ANKRD1 protein was expressed in E. coli BL21(DE3), purified on glutathione-Sepharose 4B beads, and eluted in buffer containing 30 mM reduced glutathione. Purified protein was resolved on 10% SDS-PAGE and stained with Coomassie brilliant blue. BSA was run in parallel for reference. The right panel shows that VARP-ANKRD1 specifically interacts with GDP-bound Rab32. Huh7 cells were transfected with a Flag-tagged Rab32 expression plasmid. At 36 h after transfection, 80 μg of harvested cell lysates was treated with either 0.5 mM GTPγS or 1 mM GDP for 30 min at 37°C. Cell lysates were further precipitated for 45 min by 2 μg of GST-VARP-ANKRD1 that had been immobilized on 40 μl of 50% (vol/vol) glutathione-Sepharose slurry for 3 h at 4°C. Bound protein was detected by immunoblot (IB) analysis using an anti-Flag monoclonal antibody. (F) Either HEK293T cells or Huh7 cells were transiently transfected with an empty vector or Flag-tagged Rab32-WT, Rab32-T39N, or Rab32-Q85L expression plasmid, as indicated. At 36 h after transfection, total cell lysates were incubated with either GST or GST-VARP-ANKRD1 protein. Protein complexes were precipitated with glutathione beads for 45 min at 4°C. Bound protein was detected by immunoblot analysis using an anti-Flag monoclonal antibody. (G) Either HEK293T cells or Huh7 cells were transiently cotransfected with Flag-tagged VARP-ANKRD1 and an empty vector or V5-tagged Rab32-WT, Rab32-T39N, Rab32-Q85L, or Rab32-L188P expression plasmid, as indicated. At 36 h after transfection, total cell lysates were immunoprecipitated with an anti-Flag monoclonal antibody. Bound proteins were detected by immunoblot analysis using an anti-V5 monoclonal antibody.
HCV infection shifts the balance between the GDP- and GTP-bound forms of Rab32.To investigate the nucleotide status of Rab32 in HCV-infected cells, Huh7 cells were either mock infected or infected with Jc1 for 2 and 4 days, and the endogenous Rab32-GDP level was analyzed by using purified GST-tagged VARP-ANKRD1. The experiment shown in Fig. 3A indicates that endogenous Rab32 in mock-infected cells was not pulled down by GST-tagged VARP-ANKRD1 (Fig. 3A, lanes 3). However, endogenous Rab32 in HCV-infected cells was coprecipitated with GST-tagged VARP-ANKRD1 (Fig. 3A, lanes 4). To further corroborate our result, endogenous Rab32 protein in both mock- and Jc1-infected cells was immunoprecipitated, purified under high-stringency conditions, and then incubated with GST-VARP-ANKRD1. Consistently, we showed that Rab32 derived from HCV-infected cells displayed higher affinity with GST-VARP-ANKRD1 (Fig. 3B, lane 4). Given the exclusive interaction between VARP-ANKRD1 and Rab32 in HCV-infected cells, we demonstrated that HCV not only upregulated the total protein expression level of Rab32 but also switched the balance between the GDP- and GTP-bound forms toward more production of GDP-bound Rab32. Nevertheless, the mechanism underlying the shift in nucleotide status of Rab32 in HCV-infected cells remains to be elucidated.
HCV infection converts the predominantly expressed GTP-bound Rab32 to GDP-bound Rab32. (A) Huh7 cells were either mock infected or infected with Jc1 for 4 h. Cells were harvested at the indicated days postinfection, and then cell lysates were precipitated with GST-VARP-ANKRD1 that had been immobilized on glutathione-Sepharose 4B beads for 3 h at 4°C. Bound Rab32 protein was detected by immunoblot analysis using an anti-Rab32 polyclonal antibody. (B) Huh7.5 cells were either mock infected or infected with HCV Jc1 for 4 h. At 4 days postinfection, total cell lysates were immunoprecipitated (IP) with an anti-Rab32 polyclonal antibody and incubated with protein A beads for 45 min. Bound Rab32 proteins were washed under high-stringency conditions and were further incubated with purified GST-VARP-ANKRD1. Bound GST-VARP-ANKRD1 proteins were analyzed by immunoblotting with an anti-GST monoclonal antibody.
Rab32 forms aggregates in HCV-infected cells.The nucleotide-dependent function of Rab32 has been extensively studied in melanocytes as it plays a key role in melanosome trafficking at TGNs (10, 22). However, in nonmelanogenic cell types, membrane-bound endogenous Rab32 exhibits distinct functions (13–15). In the present study, we also observed a partial distribution of endogenous Rab32 in the ER and mitochondrial fractions in Huh7 cells (Fig. 2A). Here, we sought to understand why HCV infection caused a shift from the predominant GTP-bound Rab32 to GDP-bound Rab32. Because it has been reported that overexpression of Rab32-T39N in COS-1 cells led to the formation of aggresome-like structures consisting of dot-like and/or chain-like aggregated Rab32 and that this was not due to the alteration of the secondary structure of the mutant (15), we also speculated that HCV infection might cause the aggregation of Rab32 protein. To test this hypothesis, we first examined the solubility of various forms of Rab32 protein in 1% Triton X-100 as the aggregated proteins are largely detergent resistant. Calnexin was used as the marker for the soluble fraction as reported previously (23, 24). As shown in Fig. 4A, both wild-type and GTP-locked Rab32 proteins (Q85L) were detected mainly in soluble fractions, whereas both GDP-locked T39N and PKA-binding-defective L188P mutants of Rab32 were relocalized into insoluble fractions. These data indicate that the solubility of Rab32 decreases as it shifts from the GTP- to GDP-bound state. Consistent with a previous report (15), we further observed that both Rab32-T39N and Rab32-L188P proteins were polyubiquitinated in Huh7 cells (data not shown). To further characterize the protein half-lives of both soluble and insoluble fractions of GDP-bound Rab32, Huh7 cells overexpressing Flag-tagged Rab32-T39N plasmid were treated with cycloheximide (CHX) and harvested at the time points indicated on the figures. Cells were fractionated, and equivalent amounts of proteins among different soluble and insoluble fractions were further analyzed by immunoblot analysis. As shown in Fig. 4B, the T39N protein in the soluble fraction was degraded gradually, whereas the insoluble fraction of the T39N protein was highly resistant to cellular degradation under the CHX treatment condition. This result demonstrated that the GDP-bound form of Rab32 predominantly localized in the insoluble fraction, making it less susceptible to the enzymatic attacks from the host cells. To further investigate the subcellular distribution of wild-type and mutant forms of Rab32 protein, Huh7 cells transfected with the V5-tagged wild-type (WT), T39N, and Q85L forms of the Rab32 expression plasmid were analyzed by an immunofluorescence assay using MitoTracker and anti-V5 monoclonal antibody. In accord with a previous report (15), we also showed that cells expressing Rab32-T39N but not Rab32-Q85L formed concentrated foci (Fig. 4C, white arrow) that may represent aggresome-like structures. Of note, we also observed that mitochondria were frequently and highly condensed in the perinuclear region in cells overexpressing Rab32-T39N protein as reported by Bui et al. (14). We subsequently investigated the solubility of endogenous Rab32 protein in both mock-infected and Jc1-infected Huh7 cells. As expected, Rab32 was not detected in the detergent-insoluble fraction in mock-infected cells (Fig. 4D, lane 2). In contrast, Rab32 was found in not only the detergent-soluble fraction but also the detergent-insoluble fraction at a certain level (Fig. 4D, lane 4) in Jc1-infected cells. These results indirectly suggested that HCV might promote GDP-bound Rab32 production which contributed to the altered subcellular distribution of Rab32 in HCV-infected cells. We then asked if HCV-induced Rab32 enrichment of the detergent-insoluble fraction could extend its half-life. For this purpose, we investigated the stability of endogenous Rab32 protein in mock- and Jc1-infected CHX-treated cells. As expected, the Rab32 protein level in Jc1-infected cells remained steady over time, whereas its level in mock-infected cells decreased gradually over the course of CHX treatment (Fig. 4E). Finally, we examined the level of Rab32 aggregation in both mock- and HCV-infected cells by confocal microscopic analysis. As shown in Fig. 4F, aggregation of Rab32 was hardly detectable in mock-infected cells, whereas 22% of those infected with HCV exhibited perinuclear aggregated Rab32 protein which was highly localized to the ER (Fig. 4F). Since certain viruses utilize intracellular aggregated protein structures for viral replication and/or assembly (25, 26), we speculated that HCV might induce an aggregated form of Rab32 to create an intracellular milieu that is favorable for viral propagation.
Rab32 forms aggregates in HCV-infected cells. (A) Solubility of WT and mutants of Rab32. Huh7 cells were transiently transfected with an empty vector or Flag-tagged Rab32-WT, Rab32-T39N, Rab32-Q85L, or Rab32-L188P expression plasmid, as indicated. At 36 h after transfection, total cell lysates were separated into detergent-soluble (S) and detergent-insoluble (I) fractions as described in Materials and Methods. Both fractions were analyzed by immunoblot analysis with an anti-Flag monoclonal antibody. Calnexin was used as the marker for the soluble fraction. (B) Insoluble Rab32 is highly resistant to cellular degradation machinery. Huh7 cells were transiently transfected with a Flag-tagged Rab32-T39N expression plasmid. At 36 h after transfection, cells were then treated with 10 μg/ml of CHX and harvested at the indicated time points. Total cell lysates were separated into detergent-soluble (S) and detergent-insoluble (I) fractions. Rab32 protein was detected by immunoblot analysis using an anti-Flag monoclonal antibody. Calnexin was used as the marker for the soluble fraction. (C) Aggresome-like structure formation in cells expressing Rab32-T39N protein. Huh7 cells seeded on glass coverslips were transfected with a V5-tagged Rab32-WT, Rab32-T39N, or Rab32-Q85L expression plasmid, as indicated. At 30 h after transfection, cells were incubated with 100 nM MitoTracker for 40 min to stain mitochondria. Cells were further fixed in 4% paraformaldehyde, and immunofluorescence staining was performed by using an anti-V5 monoclonal antibody. Cells were counterstained with DAPI to label nuclei (blue). The arrow indicates perinuclear aggregated structures of Rab32. Higher magnifications of the boxed areas are shown in the cropped image. (D) Deposition of endogenous Rab32 into the insoluble fraction in HCV-infected cells. Huh7 cells were either mock infected or infected with Jc1 for 4 h. At 4 days postinfection, total cell lysates were separated into detergent-soluble (S) and detergent-insoluble (I) fractions. Equivalent amounts of proteins were analyzed by immunoblot analysis using the indicated antibodies. Calnexin was used as the marker for the soluble fraction. (E) Huh7.5 cells were either mock infected or infected with HCV Jc1 for 4 h. At 4 days postinfection, cells were then treated with 10 μg/ml of CHX and harvested at the indicated time points. Total cell lysates were immunoblotted with the indicated antibodies. Band intensities of normalized core protein compared to the initial time point were quantified by ImageJ. (F) Perinuclear ER-associated Rab32 aggregates in HCV-infected cells. Huh7.5 cells were either mock infected or infected with HCV Jc1 for 4 h. At 4 days postinfection, cells were transfected with V5-tagged Rab32-WT. At 30 h after transfection, cells were further incubated with CellLight ER-GFP reagent for 16 h. Cells were then fixed in 4% paraformaldehyde, and immunofluorescence staining was performed by using an anti-V5 monoclonal antibody. Cells were also counterstained with DAPI to label nuclei (blue). The arrows indicate perinuclear ER-associated aggregated Rab32. Higher magnifications of the boxed areas are shown in the cropped images. At least 100 transfected cells were assessed for determining the percentage of aggregated Rab32-WT in mock- and Jc1-infected cells.
HCV core protein expression level is elevated by a T39N mutant of Rab32.To investigate the functional involvement of Rab32 activity in HCV propagation, Huh7.5 cells were either mock infected or infected with Jc1 and then further transfected with an empty vector, WT, T39N, Q85L, or L188P form of the Rab32 expression plasmid. At 2 days after transfection, HCV protein levels were analyzed by immunoblot analysis using the antibodies indicated on the figures. As shown in Fig. 5A, viral protein levels, including NS3 and NS5A in cells expressing various constructs of Rab32, were not changed compared with levels in the vector transfected control. Strikingly, core protein levels in cells expressing either wild-type or T39N Rab32 were dramatically higher than those in cells expressing the Q85L or L188P form of Rab32 protein. Of note, extracellular HCV RNA levels released into the culture supernatants were proportional to intracellular core levels (Fig. 5B), raising the possibility that Rab32-T39N was involved in HCV propagation by increasing the core protein level. When naive Huh7.5 cells were infected with virus-containing culture supernatants harvested from Fig. 5A, HCV protein levels in the second infection were also higher in cells expressing either the wild type or T39N than in cells expressing Q85L or L188P (Fig. 5C, lanes 2 and 3 versus lanes 4 and 5). These data suggest that GDP-bound Rab32 displayed a positive effect on HCV production. In addition, we showed that overexpression of L188P, which harbors an aberrant PKA-anchoring signal, exhibited a marked reduction in HCV propagation (Fig. 5C, lane 5). This was confirmed by the titration of released extracellular viruses (Fig. 5D). Collectively, these data suggest that HCV may upregulate GDP-bound Rab32 to potentiate core protein expression and thus facilitate viral propagation, and PKA-anchoring activity of Rab32 appears to have a pivotal role in this regulation. We further confirmed that ectopic expression levels of core (left panel) but not NS5A (right panel) protein were enhanced in the presence of Rab32-T39N (Fig. 5E). We also verified that core protein but not other HCV proteins was accumulated by Rab32-T39N (data not shown).
Rab32-GDP positively modulates HCV core protein expression and facilitates viral propagation. (A) Huh7.5 cells were either mock infected or infected with Jc1 for 4 h. At 24 h postinfection, cells were further transiently transfected with the indicated Rab32 expression plasmid. At 48 h after transfection, cells were harvested, and total cell lysates were immunoblotted with the indicated antibodies. Band intensities of normalized core protein were quantified by ImageJ. (B) Extracellular HCV RNA harvested from cell culture supernatants of the experiment shown in panel A were quantified by qRT-PCR. (C) Naive Huh7.5 cells were infected with virus-containing supernatants harvested from panel A. At 48 h postinfection, cells were harvested, and total cell lysates were immunoblotted with the indicated antibodies. Band intensities of normalized HCV proteins were quantified by ImageJ. (D) A TCID50 assay was used to determine the titer of viruses released in cell culture supernatants of the experiment shown in panel A. (E) Huh7.5 cells were cotransfected with either Myc-tagged HCV core protein (left panel) or Myc-tagged HCV NS5A (right panel) and various constructs of the indicated Flag-tagged Rab32 expression plasmid. At 36 h after transfection, cells were harvested, and total cell lysates were immunoblotted with the indicated antibodies. Band intensities of normalized core protein were quantified by ImageJ.
Rab32 selectively interacts with HCV core protein and enriches core protein in the detergent-insoluble fraction.Given that the expression level and activity of Rab32 lead to a profound alteration in HCV propagation and core protein expression, we sought to determine the possible interaction between Rab32 and HCV proteins. For this purpose, a GST pulldown assay was performed using GST-tagged Rab32 protein produced in E. coli. As demonstrated in Fig. 6A, Rab32 selectively interacted with core protein but not with other HCV proteins. This interaction was confirmed by an immunoprecipitation assay (Fig. 6B). We further verified that core protein interacted with endogenous Rab32 in HCV-infected cells (Fig. 6C). These data suggest that Rab32 may colocalize with core protein in HCV-infected cells. To investigate this possibility, Huh7 cells were either mock infected or infected with Jc1 and then subjected to an immunofluorescence assay. We showed that both endogenous Rab32 and HCV core protein were colocalized in the cytoplasm, as shown by yellow fluorescence (Fig. 6D). We further confirmed the colocalization of Rab32 and HCV core protein by determining both Pearson's and Manders' coefficients. It was noteworthy that yellow fluorescent signals were found in regions showing dotted structures of Rab32 protein. These structures may represent intracellular Rab32 inclusions that were absent in mock-infected cells (Fig. 6D). These data suggest that HCV infection promoted the formation of small cytoplasmic Rab32 aggregates, as we showed previously (Fig. 4F), and that Rab32 specifically interacted with HCV core protein. We next dissected the nucleotide dependency of Rab32 in protein interaction with core protein. For this purpose, either HEK293T cells (upper panel) or Huh7 cells (lower panel) were cotransfected with Myc-tagged HCV core protein and various constructs of a Flag-tagged Rab32 expression plasmid. Immunoprecipitation data showed that Rab32-T39N but not the Q85L protein was coimmunoprecipitated with core protein (Fig. 6E), implying that HCV core protein specifically interacted with GDP-bound Rab32. Of note, only trace amounts of wild-type Rab32 interacted with core protein in Huh7 cells (Fig. 6E, lower panel), further confirming that the majority of Rab32 proteins in naive Huh7 cells are in the GTP-bound state. It has been previously reported that expression of Rab32-T39N but not Q85L led to the redistribution of ubiquitinated proteins, including Hsc70 and α-synuclein, from the cytoplasm to aggresome-like structures in the microtubule-organizing center (15). In addition, HCV core protein has been shown to be ubiquitinated by ubiquitin ligase E6AP and to be degraded by the proteasome-dependent pathway (27, 28). Recently, Afzal et al. reported that core protein is likely to interfere with its own degradation by undergoing a posttranslational modification that makes the core protein partition into detergent-resistant microdomains of the ER membranes (29). We therefore speculated that HCV-induced GDP-bound Rab32 production might play an important role in core stabilization by enriching core protein in Rab32-derived aggregated structures. For this purpose, Huh7 cells transfected with various constructs of Rab32 were infected with Jc1, and then cell lysates were separated into detergent-soluble and -insoluble fractions. Figure 6F shows that a large portion of wild-type Rab32 protein was redistributed from the detergent-soluble to -insoluble fraction upon HCV infection (lane 4), further confirming that HCV infection shifted the nucleotide status of Rab32 from a predominantly GTP-bound form to a GDP-bound form. Importantly, major changes in protein solubility occurred in core protein but not in other HCV proteins. As expected, ectopic expression of both Rab32-WT and T39N increased the insoluble core protein level 1.64 and 2.29 times, respectively, compared with the level of the vector control (Fig. 6F, lanes 4 and 6). This result indicated that GDP-bound Rab32 interacted with core protein and enriched core protein in the detergent-insoluble fraction. Notably, despite its interaction with core protein, expression of an L188P mutant of Rab32 resulted in the lowest level of detergent-insoluble core protein, demonstrating the functional importance of the PKA-anchoring activity of Rab32 as well as PKA signaling. By immunofluorescence assay, we further showed that core proteins in cells overexpressing Rab32-T39N but not Q85L accumulated as yellow granular dots in cytoplasmic juxtanuclear structures (Fig. 6G) that we observed previously (Fig. 6D). Colocalization of Rab32-T39N and core protein was further verified by both Pearson's and Manders' overlap coefficients. These data indicate that GDP-bound Rab32 selectively interacted with HCV core protein and enriched core in Rab32 aggregates. To validate the impact of Rab32 protein on the detergent resistance of core protein, Huh7 cells infected with Jc1 were further transfected with either an empty vector or V5-tagged wild-type Rab32. At 30 h after transfection, cells were briefly dipped in 1% Triton X-100 solution and then further processed for immunofluorescence assay. As reported previously (23), some dots and patches of core protein were resistant to detergent treatment (Fig. 6H, arrows in the left panel). Importantly, core proteins appeared to localize onto detergent-resistant Rab32 proteins and accumulated at a high level in cells expressing Rab32 protein (Fig. 6H, arrows in the right panel), whereas only small dots of core proteins were detected in neighboring untransfected cells. These data show that insoluble Rab32 protected core protein from detergent extraction and thus that more core proteins were accumulated in the detergent-insoluble fraction.
Rab32-GDP selectively interacts with HCV core protein and enriches core in the detergent-insoluble fraction. (A) Rab32 selectively interacts with HCV core protein but not with other HCV proteins. HEK293T cells were transiently transfected with an Myc-tagged HCV core, NS3, NS4B, NS5A, or NS5B expression plasmid, as indicated. At 36 h after transfection, total cell lysates were incubated with either GST or GST-Rab32 protein. Protein complexes were further precipitated with glutathione beads for 45 min at 4°C. Bound proteins were detected by immunoblot analysis using an anti-Myc monoclonal antibody. (B) HEK293T cells were cotransfected with Flag-tagged Rab32 and either an empty vector, Myc-tagged HCV core protein, or HCV NS5B expression plasmid, as indicated. At 36 h after transfection, total cell lysates were then immunoprecipitated with an anti-Myc monoclonal antibody. Bound proteins were detected by immunoblot analysis using an anti-Flag monoclonal antibody. The arrowheads indicate IgG. (C) Huh7.5 cells were infected with Jc1 for 4 h. At 4 days postinfection, total cell lysates were immunoprecipitated with either IgG or an anti-Rab32 polyclonal antibody. Bound proteins were analyzed by immunoblotting with an anti-HCV core protein or an anti-NS5A polyclonal antibody. The arrowhead indicates IgG. (D) Huh7.5 cells were either mock infected or infected with Jc1 for 4 h. At 48 h postinfection, cells were seeded on glass coverslips and were further cultured for 36 h. Cells were then fixed in ice-cold methanol, and immunofluorescence staining was performed by using anti-Rab32 and anti-HCV core protein polyclonal antibodies. Cells were counterstained with DAPI to label nuclei (blue). The arrow indicates the dotted structures of Rab32. Higher magnifications of the boxed areas are shown in the cropped images. Quantification of colocalization between HCV core protein and Rab32 was determined using the Pearson's and Manders' coefficients. R, red; G, green. (E) Either HEK293T cells or Huh7 cells were transiently cotransfected with Myc-tagged HCV core protein and an empty vector, Flag-tagged Rab32-WT, Rab-T39N, Rab32-Q85L, or Rab32-L188P expression plasmid, as indicated. At 36 h after transfection, total cell lysates were then immunoprecipitated with an anti-Myc monoclonal antibody. Bound proteins were detected by immunoblot analysis using an anti-Flag monoclonal antibody. The arrowheads indicate IgG. (F) GDP-bound Rab32 enriches core protein in the detergent-insoluble fraction. Huh7 cells were transiently transfected with either an empty vector or a Flag-tagged Rab32-WT, Rab32-T39N, Rab32-Q85L, or Rab32-L188P expression plasmid, as indicated. At 24 h after transfection, cells were further infected with Jc1 for 4 h. At 36 h postinfection, total cells lysates were separated into the 1% Triton X-100 soluble (S) and insoluble (I) fractions. Equivalent amounts of proteins in each fraction were analyzed by immunoblot analysis using the indicated antibodies. Calnexin was used as the marker for the soluble fraction. Band intensities of core protein in the detergent-insoluble fraction were quantified by ImageJ. (G) GDP-bound but not GTP-bound Rab32 colocalizes with HCV core protein. Huh7 cells were infected with Jc1 for 4 h. At 48 h postinfection, cells were further transfected with V5-tagged T39N or Q85L of the Rab32 expression plasmid. At 36 h after transfection, cells were fixed in 4% paraformaldehyde, and immunofluorescence staining was performed by using an anti-HCV core protein polyclonal antibody and an anti-V5 monoclonal antibody. Dual staining showed colocalization of HCV core protein and Rab32-T39N in the perinuclear aggregated structures as yellow fluorescence in the cropped images. Cells were counterstained with DAPI to label nuclei (blue). Quantification of colocalization between HCV core protein and both Rab32-T39N and Rab32-Q85L was determined using the Pearson's and Manders' coefficients. R, red; G, green. (H) Huh7 cells were infected with Jc1 for 4 h. At 48 h postinfection, cells were further transfected with either an empty vector (upper panel) or V5-tagged Rab32-WT (lower panel). At 30 h after transfection, cells were dipped in 1% Triton X-100 solution for 5 s four times. Cells were further fixed in 4% paraformaldehyde, and immunofluorescence staining was performed by using an anti-HCV core protein polyclonal antibody and an anti-V5 monoclonal antibody. Yellow fluorescence in the merged image indicates the colocalization of detergent-resistant core protein and Rab32. Cells were counterstained with DAPI to label nuclei (blue). +, Rab32 transfected cells. The arrows indicate detergent-resistant core protein.
Rab32 is involved in the assembly step of the HCV life cycle.HCV core protein has been shown to be localized in the detergent-resistant membranes which are putative platforms for HCV particle assembly (23, 24). In addition, HCV core protein is localized on the surface of lipid droplets (LDs) in virus-infected cells, and core protein-coated LDs aggregate in the perinuclear region which has been proven to be essential for virion assembly (30–32). Since the stabilization of these core protein-coated LDs could be mediated through the formation of aggresome-like structures (25, 26), we hypothesized that HCV might usurp aggregated Rab32 for virion assembly. To dissect the functional role of Rab32 in the HCV life cycle, we initially investigated the effect of Rab32 silencing on HCV pseudoparticle (HCVpp) entry and HCV internal ribosome entry site (IRES)-mediated translation. We observed that both HCVpp entry and IRES-mediated translation were not affected by the interference of Rab32 protein (data not shown). We then asked if Rab32 was required for HCV RNA replication. To address this question, HCV subgenomic replicon cells derived from either genotype 1b or genotype 2a were treated with the small interfering RNAs (siRNAs) indicated on the figure. We demonstrated that silencing of Rab32 displayed no discernible effect on intracellular HCV RNA and protein levels in HCV subgenomic replicon cells derived from both genotype 1b (Fig. 7A, left panel) and genotype 2a (Fig. 7A, right panel). These data suggest that Rab32 may not be involved in early stages of the HCV life cycle. To further investigate if Rab32 is required for later steps of the HCV life cycle, Huh7.5 cells transfected with two different siRNAs targeting Rab32 were infected with Jc1. At 48 h postinfection, both HCV RNA and protein levels were determined. As shown in Fig. 7B, knockdown of Rab32 significantly suppressed the HCV RNA level. Consistently, HCV protein expression was also impaired in Rab32 knockdown cells (Fig. 7C). We further showed that extracellular HCV RNAs (Fig. 7D, upper panel) as well as core protein (Fig. 7D, lower panel) levels in the culture supernatants were markedly reduced upon silencing of Rab32. Extracellular virus-containing supernatants were quantified by titration and were used to infect naive Huh7.5 cells. As shown in Fig. 7E and F, extracellular HCV infectivity as well as the HCV protein level in the second-round infection in Rab32 knockdown cells was dramatically suppressed compared to the level in the negative control, resulting in a significant decrease in the specific infectivity of extracellular viruses (Fig. 7G). We further investigated the specific infectivity of intracellular viruses in Rab32 knockdown cells. For this purpose, Huh7.5 cells transfected with the siRNAs indicated on the figure were further infected with Jc1 and then disrupted by multiple cycles of freezing and thawing. Subsequently, a 50% tissue culture infectious dose (TCID50) assay was used to determine the titer of intracellular HCV virus. As shown in Fig. 7H, intracellular HCV infectivity was significantly decreased in Rab32 knockdown cells. The parallel decrease in the intracellular (Fig. 7I) and extracellular (Fig. 7G) specific infectivities of HCV in Rab32 knockdown cells argued for a specific impairment in HCV assembly. We further showed that the treatment of siRNAs displayed no effect on cell viability, confirming that the silencing effect was specific to Rab32 (Fig. 7J). Since HCV capsid assembly and envelopment occur in association with LD-associated ER membranes (33–35) and since core protein binds onto ER-associated Rab32 aggregates (Fig. 4 and 6), we hypothesized that the ER-associated Rab32-derived aggregated structures could provide the platforms for the capsid assembly process and thus might be important for the recruitment of core protein from LDs to assembly sites. To verify our assumption, we investigated the silencing effect of Rab32 on core protein and LD distribution by confocal microscopic analysis. Consistent with previous reports (35, 36), Jc1 core protein exhibited a reticular staining pattern with a low level of core protein-LD association in negative siRNA-treated cells (Fig. 7K, left panel). Of note, knockdown of Rab32 resulted in the retention of core protein on the surface of clustered LDs, which implied an impairment of core movement to assembly sites (Fig. 7K, right panel). Taking these observations together, we provide compelling evidence that Rab32 is a host factor required for HCV particle assembly.
Rab32 is involved in the assembly step of the HCV life cycle. (A) HCV replicon cells derived from genotype 1b or genotype 2a were transfected with a 40 nM concentration of the indicated siRNAs. At 72 h after transfection, both intracellular RNA and protein levels were analyzed. (B) Huh7.5 cells were transfected with a 20 nM concentration of the indicated siRNAs for 48 h and were further infected with Jc1 for 4 h. At 2 days postinfection, intracellular RNA levels were determined by qRT-PCR. (C) Total cell lysates from the experiment shown in panel B were immunoblotted with the indicated antibodies. Band intensities of normalized HCV proteins were quantified by ImageJ. (D) Extracellular HCV RNA in cell culture supernatants from the experiment shown in panel were quantified by qRT-PCR (top). Extracellular HCV core protein in cell culture supernatants from the experiment shown in panel B were concentrated and detected by immunoblot analysis (bottom). Band intensities of core proteins were quantified by ImageJ. (E) Infectivity of viruses released in cell culture supernatants from the experiment shown in panel B was determined by using 50% tissue culture infective doses (TCID50s). (F) Naive Huh7.5 cells were infected with Jc1 harvested from culture supernatants of the experiment shown in panel B. At 48 h postinfection, total cell lysates were immunoblotted with the indicated antibodies. Band intensities of normalized HCV proteins were quantified by ImageJ. (G) Extracellular specific infectivity was calculated as a ratio of the viral titers to the corresponding HCV RNA genomes in the culture supernatants. (H) Huh7.5 cells were treated as described for panel B. At 2 days postinfection, cells were trypsinized and disrupted by multiple cycles of freezing and thawing. A TCID50 assay was used to determine the titer of viruses in postnuclear supernatants. (I) Intracellular specific infectivity was calculated as a ratio of the intracellular viral titers to the corresponding intracellular HCV RNA genomes. (J) Huh7.5 cells were transfected with 40 nM concentrations of the indicated siRNAs. At 72 h after transfection, cell viability was determined by WST assay. (K) Rab32 silencing results in the retention of core protein on the surface of lipid droplets. Huh7 cells were infected with HCV Jc1 for 4 h. At 48 h postinfection, cells were then treated with a 20 nM concentration of either a negative or Rab32-specific siRNA. At 48 h after transfection, cells were fixed in 4% paraformaldehyde, and immunofluorescence staining was performed by using an anti-HCV core protein polyclonal antibody. Lipid droplets were visualized by staining with BODIPY (439/503). Cells were also counterstained with DAPI to label nuclei (blue). Data represent averages from at least three independent experiments for panels A, B, D, E, G, H, I, and J. The asterisk indicates significant difference (*, P < 0.05; **, P < 0.01) from the value for the control.
DISCUSSION
The life cycle of HCV is inextricably linked to the host cell's machinery, and thus targeting host factors may provide a potential approach for the development of antiviral agents. Here, we show that Rab32 is a novel host factor playing a critical role in viral propagation, in particular, the assembly step of the HCV life cycle. By employing RNA-Seq technology, we identified that Rab32 expression was upregulated in HCVcc-infected cells. It has been previously reported that some members of the Rab GTPase family were involved in various steps of the HCV life cycle (37–39). Rab32 is ubiquitously expressed in human tissues and has diverse functions, including PKA-mediated calcium signaling, apoptosis and mitochondrial dynamics, lipid biosynthesis, and autophagosome formation (13–16). Nonetheless, its functional involvement in HCV propagation has not been clearly characterized so far. In the present study, we showed that HCV infection positively regulated Rab32 expression at the transcriptional level and that HCV NS3 was mainly responsible for this upregulation (data not shown). We further showed that knockdown of Rab32 significantly impaired viral infectivity. Indeed, Mankouri et al. have investigated the functional significance of several TGN-associated Rab GTPases and suggested that Rab32 might be important for HCV assembly/secretion (40). Melanocyte-specific murine Rab32 predominantly localizes to the TGNs, where it regulates the transport of key melanogenic enzymes (10). In contrast, membrane-bound Rab32 is present primarily at mitochondrion-associated ER membranes in nonmelanogenic murine liver cells (20). Consistently, as shown in our study, endogenous membrane-bound Rab32 in Huh7 cells is largely present in the perinuclear ER and mitochondria and thus probably has a distinctive functional role in regulating HCV production.
Because the Rab GTPase switches between GDP- and GTP-bound forms that determine its specific activities, it is important to know whether there is a difference in the nucleotide status of Rab32 between naive and HCV-infected cells. Since the first ankyrin repeat domain of murine VPS9-ankyrin repeat protein functions as a GTP-dependent Rab32-binding protein, it was used to quantify the amount of GTP-bound Rab32 in melanocytes (22). Paradoxically, we showed that human GDP-bound Rab32 interacted with VARP-ANKRD1 and colocalized in perinuclear ER and mitochondria. Indeed, Bui et al. reported that GDP-locked Rab32-T39N in nonmelanogenic cell types is preferentially distributed to the ER-mitochondrion contact sites (14). By employing a VARP-ANKRD1 construct, we demonstrated that endogenous Rab32 was maintained in the GTP-bound form in naive Huh7 cells, as in the case of Rab3D and Rab27 (41, 42). Interestingly, Rab32 switched its status from the GTP- to GDP-bound form in response to HCV infection. Although the mechanisms underlying this conversion are still uncertain, HCV core protein may be involved in the Rab32 GDP-GTP cycle because core protein selectively downregulated Rab32-GTP but not Rab32-GDP in a lysosome-dependent manner (data not shown). Another scenario could be due to the physical binding of HCV core protein on GDP-bound Rab32 that probably prevents the release of GDP during the process of GDP-GTP exchange.
The GDP-bound forms of certain Rab GTPases, such as Rab8a and Rab27a, exert positive effects on specific biological functions (43, 44). HCV-induced GDP-bound Rab32 production led us to dissect its functional significance in HCV propagation. Overexpression of GDP-locked Rab32 in COS-1 cells results in the formation of aggresome-like structures at the microtubule-organizing center without affecting proteasome activity (15). It has also been reported that GDP-bound Rab32 causes mitochondria to cluster in the perinuclear region in HeLa cells (13, 14). In the present study, we also observed that overexpression of Rab32-T39N in Huh7 cells resulted in a more juxtanuclear condensation of mitochondria and deposition of aggregated Rab32 protein near the nuclei. Importantly, HCV infection led to the formation of small cytoplasmic ER-associated Rab32 aggregates. In addition, HCV-infected Huh7 cells also displayed the condensation of mitochondria around the cell nuclei (data not shown) as reported previously (45). A variety of viruses has been shown to trigger cellular remodeling that develops intracellular aggregates containing viral capsid proteins in order to support viral propagation. Poxviruses, African swine fever virus, iridoviruses, and phycodnaviruses hijack the aggresome pathway while human cytomegalovirus and herpes simplex virus utilize the aggresome-like structures to generate viral assembly compartments that recruit capsid proteins, mitochondria, and heat shock proteins to facilitate capsid assembly (25, 26, 46). It therefore raised the possibility that perinuclear aggregates derived from HCV-induced Rab32 GDP production might contribute to putative assembly compartments required for HCV particle assembly. Indeed, HCV particle production was more efficient in cells transfected with GDP-locked Rab32-T39N than in cells transfected with an empty vector. Notably, viral proteins in functional complexes could counter the host defense mechanism by sheltering in aggregated structures (46–48). Interestingly, in the present study, we showed that the GDP- but not GTP-bound form of Rab32 specifically interacted with HCV core protein and enriched core in Rab32-derived perinuclear aggregated structures and resulted in a high level of core protein expression. Our hypothesis also fits well with the study of Shanmugam et al. showing that HCV core protein-associated detergent-resistant membranes may serve as platforms for virion assembly (24). Importantly, it has been previously reported that Rab32 localizes at the mitochondrion-associated membrane (MAM) and regulates MAM properties (14). Since MAMs enrich for viral assembly complexes (49), it is plausible that Rab32 might localize at the viral assembly sites. On the basis of our results, we would predict that ER-associated Rab32-derived aggregated structures in HCV-infected cells might constitute viral assembly compartments that not only concentrate core proteins required for viral capsid morphogenesis and protect core from the host cellular degradation machinery but also alter mitochondrial networks to provide energy for this process. However, further detailed studies are needed to clarify our assumption. Indeed, silencing of Rab32 significantly impaired HCV particle production and HCV intracellular infectivity. Moreover, knockdown of Rab32 resulted in the retention of core protein on the surface of clustered LDs, arguing for a defect in the transport of core protein from LDs to viral assembly sites. Consistent with a previous report (36), we showed that the impairment of viral particle assembly was correlated with the accumulation of core protein on LDs. All these data indicate that Rab32 is required for HCV assembly.
Several studies have shown that core protein has different putative PKA phosphorylation sites (50, 51). In the present study, both Rab32-T39N and Rab32-L188P (a PKA-binding-defective mutant) interacted with core protein, whereas overexpression of Rab32-L188P uprooted core protein from the detergent-insoluble fraction and thus impaired viral propagation. These results suggest that proper PKA anchoring as well as PKA signaling may play a critical role in HCV propagation via modulating core protein distribution and stability. However, the precise role of the putative PKA anchoring activity of Rab32 on HCV propagation may merit further investigation. In summary, this study provides some compelling evidence showing that Rab32 is an essential host factor required for HCV particle assembly and thus that Rab32 could be a potential therapeutic target to interrupt HCV propagation.
MATERIALS AND METHODS
Plasmid constructions.Total cellular RNAs were isolated from Huh7 cells by using RiboEx (GeneAll), followed by cDNA synthesis with a reverse transcription kit (Toyobo). These cDNAs were used to amplify full-length human Rab32 and the first ankyrin repeat domain (ANKRD1) of the human VPS9-ankyrin repeat protein (VARP) (amino acids [aa] 451 to 650). PCR products were inserted into a p3XFlag-CMV-10 (Sigma-Aldrich) or pGEX-4T-1 expression vector (Amersham Biosciences) to generate Flag-tagged Rab32 and GST-tagged VARP-ANKRD1 constructs. Rab32 was further subcloned into pGEX-4T-1 vector for protein expression. T39N, Q85L, and L188P mutants of Rab32 were constructed by PCR-based mutagenesis, and each construct was inserted into the corresponding enzyme sites of the plasmid p3XFlag-CMV-10. Both the wild type and mutants of Rab32 were subcloned into the pEF6-V5-HisB expression vector (Invitrogen) to generate V5-tagged constructs. A Rab32 promoter construct (nt −643 to +260) was amplified from Huh7 genomic DNA and inserted into the pGL3-basic vector (Promega). Myc-tagged HCV core protein, NS3, NS4B, NS5A, and NS5B plasmids were described previously (52). Primers are listed in Table 1.
List of primers used in this studya
Cell culture.All cell lines were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum and 100 units/ml of penicillin-streptomycin in 5% CO2 at 37°C. Huh7 cells harboring HCV subgenomic replicon derived from genotype 1b or 2a and IFN-cured cells were grown as reported previously (52). Primary human hepatocytes were cultured as we described elsewhere (53).
Antibodies and reagents.Antibodies were purchased from the following sources: mouse polyclonal anti-Rab32 antibody was from Abcam, mouse anti-Flag antibody was from Sigma-Aldrich, mouse anti-c-Myc was from Santa Cruz, mouse anti-β-actin antibody was from Sigma-Aldrich, and mouse anti-V5 antibody was from Invitrogen. HCV core protein, NS3, and NS5A antibodies have been described elsewhere (53). Either horseradish peroxidase-conjugated goat anti-rabbit antibody or goat anti-mouse antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) was used as a secondary antibody. GTPγS and GDP were purchased from Sigma-Aldrich.
Immunoprecipitation.HEK293T or Huh7 cells were transfected with the indicated expression plasmids. Total amounts of DNA were adjusted by adding an empty vector. At 36 h after transfection, cells were harvested and lysed in cell lysis buffer. Cell lysates were centrifuged at 13,500 rpm for 15 min. The supernatant was then incubated overnight at 4°C with the antibodies indicated on the figures. Protein complexes were further precipitated with 40 μl of protein A beads (Sigma-Aldrich) for 45 min at 4°C. The beads were subsequently washed three times in washing buffer, and then bound proteins were detected by immunoblot analysis.
GDP- and GTP-binding assay.Huh7.5 cells were transfected with a Flag-tagged wild-type Rab32 expression plasmid. At 36 h after transfection, cells were lysed in lysis buffer (50 mM Tris, [pH 7.5], 150 mM NaCl, 5 mM MgCl2, 1% NP-40, 0.25% sodium deoxycholate, and protease inhibitors). A total of 80 μg of cell lysates supplemented with 10 mM EDTA (pH 8.0) was then treated with either 0.5 mM GTPγS or 1 mM GDP for 30 min at 37°C. Reactions were terminated by placing samples on ice and adding 60 mM MgCl2. Cell lysates were further precipitated for 45 min with 2 μg of GST-VARP-ANKRD1 that had been immobilized on 40 μl of a 50% (vol/vol) glutathione-Sepharose 4B slurry for 3 h at 4°C. Bound protein was detected by immunoblot analysis using an anti-Flag monoclonal antibody.
Purification of GST fusion protein.The glutathione S-transferase (GST)-Rab32 and GST-VARP-ANKRD1 fusion proteins were expressed in E. coli BL21(DE3) (Novagen). Cells were lysed in lysis buffer (10 mM Tris [pH 7.5], and 150 mM NaCl) supplemented with 5 mM dithiothreitol (DTT), 5 mM MgCl2, 1 mM MnCl2, 10 μg/ml DNase, 10 μg/ml RNase, and 1 mM phenylmethylsulfonyl fluoride (PMSF). GST fusion proteins were then purified with glutathione-Sepharose 4B beads (Amersham Biosciences) and eluted in buffer containing 20 mM Tris (pH 8.5), 500 mM NaCl, 20 mM DTT, and 40 mM reduced GSH.
RNA interference.The universal negative-control siRNA and following siRNAs targeting Rab32 were purchased from Bioneer (South Korea): Rab32 siRNA 1, 5′-CUG AGA AUG GGU UAC AGA U (dTdT)-3′ (sense) and 5′-AUC UGU AAC CCA UUC UCA G (dTdT)-3′ (antisense); Rab32 siRNA 2, 5′-CUA GUG GCU CUG UAA CUU A (dTdT)-3′ (sense) and 5′-UAA GUU ACA GAG CCA CUA G (dTdT)-3′ (antisense). The siRNA targeting the 5′ untranslated region (UTR) of the Jc1 virus (5′-CCU CAA AGA AAA ACC AAA CUU-3′) was used as a positive control (54). siRNA transfection was performed using Lipofectamine RNAiMax reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's protocol.
Cell fractionation.Isolation of Triton X-100-soluble and -insoluble fractions was performed as described previously (55) with few modifications. Briefly, Huh7 cells grown on 60-mm dishes were washed twice with phosphate-buffered saline (PBS) and lysed in 200 μl of lysis buffer (1% Triton X-100, 120 mM KCl, 30 mM NaCl, 5 mM MgCl2 and 10% glycerol supplemented with protease inhibitors). After incubation on ice for 15 min, the supernatant was collected by centrifugation at 500 × g for 5 min at 4°C (soluble fraction). The pellet was then washed twice with the lysis buffer and solubilized in 50 μl of SDS buffer (10 mM Tris [pH 7.5], 1% SDS, and protease inhibitors) for 5 min at room temperature. The dissolved fraction was further added with 150 μl of radioimmunoprecipitation assay (RIPA) buffer and sonicated for 30 s (insoluble fraction). Equivalent amounts of proteins from the soluble and insoluble fractions were analyzed by immunoblot assay. The pure mitochondrial fraction and ER-containing microsome were isolated as we described previously (56).
Immunofluorescence assay.Huh7 cells seeded on glass coverslips were transfected with the plasmids indicated on the figures. At the time points given in the figure legends, cells were washed twice with PBS and fixed with either 4% paraformaldehyde in PBS for 10 min at room temperature or with ice-cold methanol for 20 min at −20°C and then permeabilized with 0.1% Triton X-100 in PBS for 10 min at 37°C. After three washes with PBS, fixed cells were blocked with 3% bovine serum albumin (BSA) in PBS for 1 h at room temperature. Cells were then incubated with the primary antibodies indicated in the figure legends at 4°C overnight. After three washes with PBS, cells were incubated with tetramethylrhodamine isothiocyanate (TRITC)-conjugated donkey anti-mouse/rabbit IgG and/or fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse/rabbit IgG for 60 min at room temperature. Cells were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) to label nuclei. Mitochondria were stained with 100 nM MitoTracker (CMXRos; Molecular Probes) for 40 min at 37°C. The ER was labeled with CellLight ER-green fluorescent protein (GFP) (BacMam 2.0; Molecular Probes). Lipid droplets were visualized by staining with BODIPY (4,4-difluoro-3a,4a-diaza-s-indacene) (439/503 nm; Invitrogen) for 30 min at room temperature. After three washes with PBS, cells were analyzed using a Zeiss LSM 700 laser confocal microscopy system (Carl Zeiss, Inc., Thornwood, NY). Pearson's correlation coefficient and the Manders' overlap coefficient were used to quantify the level of colocalization as described previously (57).
TCID50.A 50% tissue culture infectious dose (TCID50) assay was used to determine the titer of cell culture-produced HCV as described previously (58), with few modifications. Briefly, 6.4 × 103 Huh7.5 cells/well were seeded on collagen-coated 96-well plates and incubated overnight. Cells were infected with 5-fold serial dilutions of the virus-containing supernatants and were further cultured for 96 h. Infected cells were then fixed in ice-cold methanol at −20°C for 20 min. Endogenous peroxidase was quenched with 3% H2O2 in PBS for 5 min. Cells were treated with blocking buffer, followed by incubation with an anti-NS5A polyclonal antibody (1:1,000) at 4°C overnight. Next, cells were incubated with the secondary horseradish peroxidase-conjugated goat anti-rabbit antibody (1:200) for 30 min at room temperature. Bound peroxidase was developed with diaminobenzidine (DAB) substrate (DAB+; Dako, USA) for 10 min. HCV NS5A-positive wells were counted under a light microscope to determine the TCID50 using the method of Reed and Muench (59).
WST assay.Approximately 1.5 × 104 cells/well seeded on 24-well plates were transfected with either a negative, positive, or Rab32-specific siRNA. Cell viability was measured by using water-soluble tetrazolium salt (WST) (Dail Lab) as we reported previously (57).
Quantification of RNA.Quantitative real-time PCR (qRT-PCR) experiments were performed as described previously (56) using Rab32 primers: 5′-TGG GAC AGC AGG ACT CTG G-3′ (sense) and 5′-ACT CGG GTC ATG TTG CCA A-3′ (antisense).
Statistical analysis.Data are presented as means ± standard deviations (SD). A Student t test was used for statistical analysis. The asterisks on the figures indicate significant differences, as noted in the figure legends.
ACKNOWLEDGMENTS
We thank Ralf Bartenschlager (University of Heidelberg) for the Jc1 construct and Charles Rice (The Rockefeller University) for Huh7.5 cells. We also thank Hee-Jun Kim for helpful discussions.
We have no conflicts of interest to report.
This work, including the efforts of S.B.H, was funded by the Ministry of Science, ICT and Future Planning (MSIP) (2016R010886) and the Ministry of Health and Welfare (HI13C1746).
FOOTNOTES
- Received 20 August 2016.
- Accepted 9 November 2016.
- Accepted manuscript posted online 16 November 2016.
- Copyright © 2017 American Society for Microbiology.
