Hepatitis C Virus Egress and Release Depend on Endosomal Trafficking of Core Protein

Hepatitis C virus (HCV) assembly is known to occur in juxtaposition to lipid droplets, but the mechanisms of nascent virion transport and release remain poorly understood. Here we demonstrate that HCV core protein targets to early and late endosomes but not to mitochondria or peroxisomes. Further, by employing inhibitors of early and late endosome motility in HCV-infected cells, we demonstrate that the movement of core protein to the early and late endosomes and virus production require an endosome-based secretory pathway. We also observed that this way is independent of that of the internalization of endocytosed virus particles during virus entry.Hepatitis C virus (HCV) is a major causative agent of chronic hepatitis, liver cirrhosis, and hepatocellular carcinoma. HCV usually infects host cells via receptor-mediated endocytosis (6, 21), followed by the release of genomic RNA after uncoating of the nucleocapsid in the endosome. HCV core protein constitutes the viral nucleocapsid and may possess multiple functions. Intracellular HCV core protein is localized mainly in lipid droplets (LDs) (23, 29). Recent studies have indicated that core protein promotes the accumulation of LDs to facilitate virus assembly (1, 10) and recruits viral replication complexes to LD-associated membranes, where virus assembly takes place (23). However, the precise mechanisms of HCV assembly, budding, and release remain largely unclear. Most recently, HCV virion release has been shown to require the functional endosomal sorting complex required for transport III (ESCRT-III) and Vps4 (an AAA ATPase) (13), which are required for the biogenesis of the multivesicular body (MVB), a late endosomal compartment (12). Late endosomes have been implicated in the budding of several other viruses, including retroviruses (8, 17, 24, 25, 27), rhabdoviruses (14), filoviruses (18, 20), arenaviruses (26, 32), and hepatitis B virus (35). However, little is known about the roles of late endosomes in the HCV life cycle.Since LDs are associated with the endoplasmic reticulum membrane, endosomes, peroxisomes, and mitochondria (16, 37), we investigated what subcellular compartments may be involved in HCV assembly and release. We first compared the intracellular distribution of HCV core protein with that of early endosome markers Rab5a and early endosome antigen 1 (EEA1), as well as the late endosome marker CD63 in the HCV Jc1-infected Huh7.5 cells at day 10 postinfection (p.i.). In immunofluorescence studies, we demonstrated that the core protein partially colocalized with Rab5a (Fig. ​(Fig.1A,1A, left panel) or EEA1 (Fig. ​(Fig.1A,1A, right panel). This finding was confirmed by the expression of enhanced green fluorescent protein (EGFP)-tagged Rab5a (Fig. ​(Fig.1A,1A, middle panel). Similarly, core protein also showed partial colocalization with CD63 (Fig. ​(Fig.1B).1B). In particular, core protein showed numerous vesicle-like structures of homogeneous size that partially colocalized with CD63 at the cell periphery (Fig. ​(Fig.1B,1B, right panel inset and drawing). This result contrasts with that of Ai et al. (2), who observed, by confocal microscopy, that core protein did not interact with markers of early and late endosomes. Ai et al. did find, however, that multimeric core complexes cofractionated with ER/late endosomal membranes in HCV-infected cells.Open in a separate windowFIG. 1.HCV core protein colocalized with early and late endosomes but not mitochondria and peroxisomes. HCV-infected cells were costained with anti-core protein (red) and anti-Rab5a (A, left panel), -EEA1 (A, right panel), or -CD63 antibodies (green) (B) or transfected with plasmids expressing enhanced green fluorescent protein (EGFP)-Rab5a (A, middle panel), enhanced yellow fluorescent protein (EYFP) (C, left panel), EYFP-mitochondria (C, middle panel), or EYFP-peroxisome (C, right panel). Cellular DNA was labeled with DAPI (4′,6-diamidine-2-phenylindole) (blue). Images shown were collected sequentially with a confocal laser scanning microscope and merged to demonstrate colocalization (yellow merge fluorescence). Enlarged views of parts of every image are shown (insets). The cartoon in panel B illustrates the core protein-containing vesicle-like structures (depicted as red circles), which partially colocalized with CD63 at the cell periphery in HCV-infected cells. PM, plasma membrane.To demonstrate the specificity of the association of core protein with endosomes, we transfected HCV-infected cells (at day 10 p.i.) with pEYFP, pEYFP-mito, and pEYFP-peroxi (Clontech) (Fig. ​(Fig.1C),1C), which label the cytoplasm/nucleus, mitochondria, and peroxisomes, respectively. The results showed that HCV core protein did not colocalize with mitochondria or peroxisomes. Taken together, these results indicate that core protein is partially associated with early and late endosomes.To investigate the functional involvement of the endosomes in HCV release, we employed HCV-infected cells (at day 10 p.i.). In our observation, at day 10 p.i., not all the cells were infected with HCV, as revealed by immunofluorescence staining against core protein (data not shown), suggesting that these cells are a mixture of infected and noninfected cells. We examined the effects of inhibitors of endosome movement, including 10 μM nocodazole (which induces microtubule depolymerization), 100 nM wortmannin (which inhibits early endosomes), 20 nM Baf-A1 (which blocks early endosomes from fusing with late endosomes) (Sigma), and 10 μg/ml U18666A (which arrests late endosome movement) (Biomol), on the release of HCV in the HCV-infected cells. We first determined the possible cytotoxicity of these drugs. We found that within 20 h of the drug application, no significant effect on cell viability, as revealed by 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium salt (MTS) assay, was observed (Fig. ​(Fig.2C).2C). We therefore treated the cells with the various drugs for 20 h. This protocol focuses only on virus release, not on virus entry, as the reinfection of Huh7.5 cells could not account for the effects on virus release, as one round of HCV replication requires about 24 h (11).Open in a separate windowFIG. 2.HCV virion release requires endosome motility. HCV-infected cells (at day 10 p.i.) were treated with DMSO, nocodazole (10 μM), U18666A (10 μg/ml), Baf-A1 (20 nM), or wortmannin (100 nM) for 20 h, and then the levels of extracellular (A) and intracellular (B) HCV RNA and HCV core proteins (F) in cells were analyzed by RT-qPCR and Western blotting, respectively. Results of Western blotting were quantified by PhosphorImager counting. (C) Analysis of cellular proliferation and survival by MTS assay. (D) Assay of extracellular viral infectivity. The culture supernatants from the cells treated with the various drugs as indicated were used to infect naïve Huh7.5 cells. The cells were stained with anti-core protein antibody (green) and DAPI (blue). The images were analyzed by using Metamorph, and the proportion of cells (of 5,000 counted) expressing core protein was counted (E). Results are presented as percentages and are averages and standard deviations from results of triplicate experiments. (G) In parallel, the HCV-infected cells were costained with anti-Rab5a (green) and -NS5A (red) antibodies. Cellular DNA was labeled with DAPI (blue). Enlarged views of parts of every image are shown (insets). (H) Colocalization efficiency between NS5A and early endosomes was analyzed by using Zeiss LSM Zen software. Error bars represent standard deviations of the mean result from 20 cells of two experiments. (I) Assay of intracellular HCV titers. Intracellular HCV particles were prepared from the HCV-infected cells (at day 10 p.i.) treated with U18666A at concentrations varying from 2.5 to 10 μg/ml for 20 h. The titers of intracellular HCV particles were determined by immunofluorescence staining for core-positive cell foci and are reported in focus-forming units (FFU)/ml. (J) HCV-infected cells (at day 10 p.i.) were treated with U18666A at concentrations varying from 2.5 to 20 μg/ml for 8 h, and then the levels of extracellular HCV RNA were analyzed by RT-qPCR. (K) Effects of nocodazole (10 μM), U18666A (10 μg/ml), Baf-A1 (20 nM), or wortmannin (100 nM) on HCV production in single-cycle HCV growth assays. Huh7.5 cells were infected with HCV at an MOI of 1 and then incubated with DMEM containing the various drugs. At 24 h p.i., the cells and their culture supernatants were collected and used to determine the levels of intracellular and extracellular HCV RNA, which were converted to percentages of the control levels (DMSO) as 100%. Noc, nocodazole; U18, U18666A; Baf-A1, Bafilomycin A1; Wortman, wortmannin.After treatments for 20 h, the cells and their culture supernatants were collected. Intracellular RNA was isolated from cell lysates using a High Pure RNA isolation kit (Roche), and viral RNA was isolated from cell culture supernatants using a QIAamp viral RNA kit (Qiagen). Equivalent RNA volumes were subsequently analyzed on a LightCycler 1.5 real-time PCR system (Roche) for quantitative PCR (qPCR), with a primer-probe set specific for the 5′ untranslated region (UTR) sequence of HCV Jc1 and a second set specific for the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene to quantitate the RNA amount. To calculate the percentage of HCV RNA remaining after the inhibitor treatments, the mean HCV RNA levels from triplicate wells of each sample type were standardized to the mean GAPDH RNA level in the dimethyl sulfoxide (DMSO) control wells. The relative levels of HCV RNA (percentage of control) were then analyzed with LightCycler software (version 3.53) and calculated using the relative quantification method as described in Roche Applied Science, Relative Quantification, Technical Note no. LC 13/2001 (http://www.gene-quantification.de/roche-rel-quant.pdf).As indicated in Fig. ​Fig.2A,2A, the extracellular HCV RNA levels were reduced to 32.1%, 21.7%, 76.2%, and 53.8% of the DMSO control after treatments with nocodazole, U18666A, Baf-A1, and wortmannin, respectively. We further determined the intracellular levels of HCV RNA to test whether these drugs have an effect on HCV RNA replication. Intracellular HCV RNA was reduced to 64% by nocodazole treatment (Fig. ​(Fig.2B),2B), confirming previous reports that microtubules are required for HCV RNA synthesis (9, 28). Treatments with Baf-A1 and wortmannin, however, did not affect HCV RNA replication. Interestingly, the intracellular HCV RNA level was increased to 167% by U18666A treatment (Fig. ​(Fig.2B),2B), suggesting that U18666A blocks HCV particle release (Fig. ​(Fig.2A),2A), thereby causing accumulation of HCV RNA in the late endosomes. We further determined the extracellular viral infectivity after treatments with these drugs by using the culture supernatant to infect naïve Huh7.5 cells. The infectivity was checked by counting core protein-expressing cells. Raw acquired 8-bit images (Fig. ​(Fig.2D)2D) were converted to 16-bit and analyzed with the Multi Wavelength Cell Scoring application module in Metamorph (Molecular Device). All of these treatments significantly inhibited the production of infectious HCV particles, as shown by the reduced infectivity in the supernatant of the infected cells (Fig. 2D and E). Determination of the amounts of core protein in the lysates showed that the levels of core protein in cells were not significantly affected by the various drug treatments (Fig. ​(Fig.2F).2F). Taken together, these results suggest that the inhibitors of endosome movement, including U18666A, Baf-A1, and wortmannin, reduced the secretion of HCV particles but not the HCV RNA replication. On the other hand, nocodazole-induced microtubule depolymerization reduced both HCV RNA synthesis (Fig. ​(Fig.2B)2B) (9, 28) and virus release (Fig. 2A and E). Since nocodazole also blocks the movement of endosomes between pericentriolar regions and the cell periphery (3, 7), it therefore should perturb particle release. Our present results show that nocodazole had a greater effect on the reduction in extracellular HCV RNA levels (to 32%) than intracellular HCV RNA levels (to 64%) (Fig. 2A and B), and this effect may be due to blocking of endosome movement and reduced HCV RNA replication. These results can be explained on the premise that nocodazole treatments may affect both HCV RNA replication and virus release; it led us to conclude that microtubules may simultaneously play a key role in both HCV RNA replication and virus egress. Thus, both microtubules and the movement of endosomes are required for HCV particle egress.Earlier reports indicated that Rab5, an early endosomal protein, interacts with NS4B (31) and is required for HCV RNA replication. This effect was demonstrated in Rab5 small interfering RNA (siRNA)-transfected replicon cells (5, 31). However, in our current studies (Fig. ​(Fig.2B)2B) treatments with endosome inhibitors did not reduce the levels of intracellular HCV RNA. In order to investigate the discrepancy between our results and the earlier studies, we examined the effects of endosome inhibitors on the colocalization of NS5A with Rab5a in the HCV-infected cells (at day 10 p.i.) by calculating the percentage of colocalization in the cells. Colocalization scatter diagrams were generated using the colocalization function of the Zeiss LSM Zen software. The weighted colocalization coefficient, defined as the sum of intensities of colocalizing pixels for NS5A with Rab5a in comparison to the overall sum of pixel intensities (above threshold) for NS5A, was determined. Under control conditions (DMSO), NS5A was colocalized with Rab5a throughout the cytoplasm and perinuclear region (Fig. ​(Fig.2G,2G, upper left panel). The proportion of NS5A that colocalized with Rab5a was 41% (Fig. ​(Fig.2H).2H). After treatment with U18666A, the dispersed Rab5a compartments were found only in the perinuclear region (Fig. ​(Fig.2G,2G, upper right panel). Importantly, the colocalization of NS5A with Rab5a was increased to 57% by U18666A treatment. Treatments with Baf-A1 and wortmannin, however, had no significant effect. In contrast, nocodazole treatment reduced the colocalization of NS5A with Rab5a to 12% (Fig. ​(Fig.2G2G and ​and2H).2H). Previous studies have reported that the expression levels of Rab5 and its colocalization with HCV NS4B (or NS5A) play a functional role in HCV RNA replication (31). In addition, Rab5 remained in an early endosome fraction and the expression levels of Rab5 showed no significant difference between wortmannin (100 nM)-treated or untreated cells (22). More importantly, the total levels of early endosome proteins, including EEA1 (Fig. ​(Fig.3F)3F) and Rab5a (data not shown), are not altered by the endosome inhibitors. The colocalization efficiency of NS5A with Rab5a was not reduced by U18666A, Baf-A1, or wortmannin treatments, suggesting that these drugs cannot decrease HCV RNA replication. This finding is consistent with the previous results (Fig. ​(Fig.2B).2B). These findings suggest that the observed discrepancy between our results in Fig. ​Fig.2B2B and those of the other studies (5, 31) is most likely due to differences in the expression levels of Rab5a.Open in a separate windowFIG. 3.HCV particles formed are transported from early to late endosomes. HCV-infected cells (at day 10 p.i.) were treated with the various drugs and 14 h later labeled with antibodies specific for core protein (red) and CD63 (green) (A) or EEA1 (green) (D). At the right is an enlarged area from the merged image. Nuclei were stained with DAPI (blue). (B) (E) Colocalization efficiency between core protein and early or late endosomes was analyzed by using Zeiss LSM Zen software. Error bars represent standard deviations of the mean result from 20 cells of two experiments. In parallel, the cell lysates were collected and then immunoblotted with antibodies against CD63 (C) and EEA1 (F). Results were quantified by PhosphorImager counting. The HCV-infected cells (at day 10 p.i.) were fixed either for immunofluorescence microscopy (G) or for thin-section electron microscopy (H and I). Cells were costained with anti-Rab5a (green) (G, left panel), -CD63 (green) (G, right panel) and -core protein (red). Lipid droplets (LDs) and nuclei were stained with BODYPI 493/503 (blue) and DAPI (white), respectively (G). Enlarged views of parts of every image are shown (insets). (H, left panel) Early endosome (EE) (white arrow) containing particles resembling HCV adjacent to the LDs. (I, left panel) MVB/late endosome (LE) containing particles resembling HCV and internal vesicles (white arrow). High-magnification images of the early endosome (H, right panel) and MVB/late endosome (I, right panel) harboring particles resembling HCV (black arrow).To further confirm that U18666A suppresses HCV release and/or virus assembly, we determined the titer of the accumulated infectious virus particles inside the cells. The HCV-infected cells (at day 10 p.i.) were treated with U18666A at various concentrations between 2.5 and 10 μg/ml for 20 h, and then intracellular HCV particles were isolated from the cells by repeated freezing and thawing (15). The infectivity was assayed on naïve Huh7.5 cells. The results showed that the titers of infectious intracellular HCV particles were increased in a dose-dependent manner by the U18666A treatments (Fig. ​(Fig.2I).2I). These data, combined with our previous results (Fig. 2A, B, D, and E), indicate that U18666A blocks HCV particle release. Further, we determined the 50% effective dose (ED₅₀) and 50% cytotoxic concentration (CC₅₀), which are defined as the concentration of U18666A that reduced the levels of extracellular HCV RNA by 50% and the concentration of U18666A that produced 50% cytotoxicity in an MTS assay, respectively. We observed that the HCV-infected cells (at day 10 p.i.) treated with U18666A at concentrations varying from 2.5 to 20 μg/ml for 8 h showed a dose-dependent reduction in extracellular HCV RNA levels (Fig. ​(Fig.2J).2J). The ED₅₀ and CC₅₀ were calculated by polynomial regression analysis. An ED₅₀ of 8.18 μg/ml (19.2 μM) and a CC₅₀ of 40.26 μg/ml (94.9 μM) were observed for U18666A for the reduction of extracellular HCV RNA and the cytotoxicity of HCV-infected cells, respectively. These results indicate that U18666A acts as a specific inhibitor of HCV release.In our previous results in Fig. 2A and D, we used a multiple-cycle virus growth assay, which could not discriminate the role of endosome movement in infection from that in virus assembly or egress. We therefore used a single-cycle HCV growth assay to further confirm that these endosome inhibitors could suppress HCV release. Previous studies have suggested that one round of HCV replication requires about 24 h (11). Therefore, Huh7.5 cells were infected with HCV JC1 at a multiplicity of infection (MOI) of 1 for 3 h. The HCV-infected cells were washed with phosphate-buffered saline (PBS) and then incubated with Dulbecco modified Eagle medium (DMEM) containing 10% fetal bovine serum (FBS) and nocodazole (10 μM), U18666A (10 μg/ml), Baf-A1 (20 nM), or wortmannin (100 nM). Therefore, the experiment will focus on virus egress and RNA replication instead of virus entry because the cells were first inoculated with HCV and then treated with the various drugs. At 24 h p.i., the cells and their culture supernatants were collected. Intracellular RNA and viral RNA were isolated from cell lysates and cell culture supernatants, respectively. The percentage of HCV RNA remaining after the inhibitor treatments was determined in the same way as for Fig. 2A and B. As indicated in Fig. ​Fig.2K,2K, the levels of extracellular HCV RNA (or viral particles) were reduced to 61%, 48.8%, 59.7%, and 60.4% after treatments with nocodazole, U18666A, Baf-A1, and wortmannin, respectively, compared to the control DMSO treatment. We further determined the intracellular levels of HCV RNA to test whether these drugs have an effect on HCV RNA replication. Intracellular HCV RNA was reduced to 69% by nocodazole treatment (Fig. ​(Fig.2K),2K), but not by U18666A, Baf-A1, and wortmannin. Overall, these results of single-cycle HCV growth assay are similar to those of multiple-cycle virus growth assay (Fig. 2A and B), again suggesting that the endosome movement inhibitors reduced the secretion of HCV particles.To further understand the roles of the early and late endosomes in the HCV life cycle, we next examined the effects of endosome inhibitors on the colocalization of core protein with CD63 or EEA1 in the HCV-infected cells (at day 10 p.i.) by calculating the percentage of their colocalization in the cells. The percentage of colocalization of core protein with CD63 or EEA1 was determined in the same way as for Fig. ​Fig.2H.2H. Under control conditions (DMSO), core protein was colocalized with CD63 throughout the cytoplasm, perinuclear region, and cell periphery (Fig. ​(Fig.3A,3A, top row). The proportion of core protein that colocalized with CD63 was 17% (Fig. ​(Fig.3B).3B). After treatment with U18666A, a characteristic collapse of dispersed CD63 compartments to the perinuclear region of the cells was revealed (Fig. ​(Fig.3A,3A, second row). Importantly, the colocalization of core protein with CD63 increased to 30% when the movement of late endosome was arrested by U18666A, whereas it was reduced to 2% and 7% by Baf-A1 and wortmannin treatments (Fig. 3A and B), respectively, suggesting that the movement of core protein was blocked by U18666A and accumulated in the juxtanuclear region. Additionally, EEA1-labeled distinct puncta, which are dispersed throughout the cytoplasm and partially colocalized with core protein in DMSO treatment (Fig. ​(Fig.3D,3D, top row), were found clustered in the perinuclear region following treatments with U18666A and Baf-A1 (Fig. ​(Fig.3D,3D, second and third rows). Correspondingly, the colocalization of core protein and EEA1 was increased after these treatments. In contrast, very little colocalization between core protein and EEA1 was seen after treatment with wortmannin (Fig. ​(Fig.3D,3D, bottom row). The colocalization of core protein and EEA1 was 10% in the DMSO control, in contrast with 35% and 20% after treatments with U18666A and Baf-A1, respectively, and 4% after treatment with wortmannin (Fig. ​(Fig.3E).3E). These data suggest that core protein was blocked by U18666A and accumulated in the juxtanuclear region. In parallel, the levels of CD63 and EEA1 protein in the lysates were determined by Western blotting. The results indicated that the total levels of CD63 and EEA1 were not altered by the various drug treatments (Fig. 3C and F, respectively). These results collectively indicate that the colocalization of core protein with CD63 or EEA1 and the release of virus particles depend on the motility of endosomes. Thus, we suggest that HCV core and/or the viral particles formed are transported from early to late endosomes.The above results prompted us to characterize the location of the early and late endosomes in relation to the site of core-LD colocalization, where HCV assembly takes place (23). In HCV-infected cells (at day 10 p.i.), early endosomes (Fig. ​(Fig.3G,3G, left panel) were colocalized with core protein and were located in juxtaposition to LDs, whereas the late endosomes were located far away from LDs (Fig. ​(Fig.3G,3G, right panel). These data suggest that following the assembly of viral particles in juxtaposition to LDs, the HCV particles are transported through early to late endosomes. To gain further insight into the trafficking patterns of HCV particles in Huh7.5 cells, we performed electron microscopy of the HCV-infected cells (at day 10 p.i.). Particles resembling HCV were present in both the early endosomes adjacent to the LDs (Fig. ​(Fig.3H)3H) and MVBs (late endosomes) (Fig. ​(Fig.3I).3I). The morphology of these endosomal compartments is similar to that reported previously (34, 36). These results again suggest that the HCV particles formed are transported from early to late endosomes.In order to rule out the possibility that the reduction in HCV particle secretion by endosome inhibitors (Fig. ​(Fig.2)2) was caused by the inhibition of endocytosis-mediated virus entry, we determined the percentage of cells that could be infected by HCV in the presence of endosome inhibitors. Cells were first treated with inhibitors of endosome movement, followed by HCV inoculation for 3 days. This analysis revealed that the same percentage of cells was infected and produced core protein following DMSO and U18666A treatments, 52% and 54%, respectively (Fig. 4A and B), demonstrating that U18666A did not affect HCV entry, and HCV could proceed normally to RNA replication. In contrast, Baf-A1 and wortmannin treatments yielded 7% and 20% (Fig. 4A and B), respectively, of infected cells, indicating that they blocked HCV entry, as previously reported (6, 21). Overall, these findings indicate that late endosome motility is dispensable for HCV entry and subsequent RNA replication and translation but is required for viral egress.Open in a separate windowFIG. 4.HCV entry and RNA replication are not affected by inhibiting late endosome movement, and endosomal localization of core protein is not affected by inhibiting endocytosis. Huh7.5 cells were treated with the various drugs and 14 h later washed and inoculated with HCV Jc1. At 3 days p.i., cells were stained with anti-core protein antibody (green) and DAPI (blue) (A). The images were analyzed by using Metamorph and the proportion of cells (of 5,000 counted) expressing core protein was counted (B). (C) Colocalization of HCV core protein and late endosomes is not affected by DN mutants of Esp15 or Rab5a. HCV-infected cells (at day 5 p.i.) were transfected with a control plasmid pCMV-IE (C, left panel) or with dominant negative mutants of Eps15 (pEGFP-Eps15-DN) (C, middle panel) or Rab5a (pEGFP-Rab5a-DN) (C, right panel). At day 2 posttransfection, cells were labeled with antibodies specific for core protein (blue) and CD63 (red). Cells expressed EGFP-Eps15-DN or -Rab5-DN proteins (green). Nuclei were stained with DAPI (white). Enlarged views of parts of every image (insets) are shown. Colocalization of core protein and CD63 is depicted in magenta. (D) Results from colocalization analysis are shown using Zeiss LSM Zen software. Error bars represent standard deviations of the mean result from 20 cells of two experiments.Moreover, we observed an almost complete block in virus infection after expression of either an Eps15 dominant negative (DN) mutant, EGFP-Eps15D95/295 (EGFP-Eps15-DN) (4), or a Rab5a dominant negative mutant, EGFP-Rab5a-S34N (EGFP-Rab5a-DN) (30), in naïve Huh7.5 cells (data not shown). This finding is consistent with a previous report (21). We further studied the effects of the DN mutants of Eps15 and Rab5a on the colocalization of core protein and the late endosomes to confirm that this colocalization was not due to the process of virus entry. Since the DN mutants of endocytosis will block HCV infection, we first performed virus infection followed by expression of the DN mutants. This strategy has been used to demonstrate that the trafficking of HIV-1 RNA and Gag protein to late endosome is independent of the endocytosed virus (19). EGFP-Eps15-DN and EGFP-Rab5-DN were transfected into HCV-infected cells (at day 5 p.i.). As shown in Fig. ​Fig.4C,4C, the localization of core protein with CD63 in these cells was not affected by the expression of these DN mutants; core protein remained well colocalized with CD63 in both the cell periphery and juxtanuclear positions in the multiple-cycle HCV growth assays. The calculated efficiency of colocalization of core protein with CD63 (19 to ∼20%; Fig. ​Fig.4D)4D) was not altered. These results indicate that core protein localization with late endosomes is not the result of the accumulation of endocytosed viruses but rather represents the trafficking intermediates of the core protein during the late viral replication stages. Thus, we conclude that the late endosome-based secretory pathways are not involved in virus entry but rather deliver the assembled virions to the extracellular milieu.Taken together, our results indicate that HCV egress requires the motility of early to late endosomes, which is microtubule dependent, and that this pathway is independent of the one required for virus entry. Thus, we postulate that following the assembly of virus particles in juxtaposition to LDs, the HCV particles are transported through early to late endosomes to the plasma membrane, where the membrane of late endosomes is fused with plasma membrane to release virions into the extracellular milieu. This transport appears to be important for HCV egress, but it is not clear how endosomes adapt in these processes. It is possible that HCV particles are transported into endosomes after their synthesis near LD. The endosome may facilitate transport of the virus particles to the plasma membrane or even to specialized cell surface domains, such as cell junctions. Notably, the tight junction protein claudin 1 has been reported to be required for HCV cell-to-cell transmission (33), which may help viruses to sequester away from the immune system.     

All Three Domains of the Hepatitis C Virus Nonstructural NS5A Protein Contribute to RNA Binding

The hepatitis C virus (HCV) nonstructural protein NS5A is critical for viral genome replication and is thought to interact directly with both the RNA-dependent RNA polymerase, NS5B, and viral RNA. NS5A consists of three domains which have, as yet, undefined roles in viral replication and assembly. In order to define the regions that mediate the interaction with RNA, specifically the HCV 3′ untranslated region (UTR) positive-strand RNA, constructs of different domain combinations were cloned, bacterially expressed, and purified to homogeneity. Each of these purified proteins was probed for its ability to interact with the 3′ UTR RNA using filter binding and gel electrophoretic mobility shift assays, revealing differences in their RNA binding efficiencies and affinities. A specific interaction between domains I and II of NS5A and the 3′ UTR RNA was identified, suggesting that these are the RNA binding domains of NS5A. Domain III showed low in vitro RNA binding capacity. Filter binding and competition analyses identified differences between NS5A and NS5B in their specificities for defined regions of the 3′ UTR. The preference of NS5A, in contrast to NS5B, for the polypyrimidine tract highlights an aspect of 3′ UTR RNA recognition by NS5A which may play a role in the control or enhancement of HCV genome replication.Hepatitis C virus (HCV) is a human pathogen which chronically infects nearly 3% of the world''s population (36, 37). Persistent infection, in 80% of cases, leads to chronic hepatitis which can progress to liver cirrhosis and, in the worst cases, hepatocellular carcinoma (37). Current therapies lack specificity and efficacy due largely to an incomplete understanding of the complex molecular mechanisms of virus infectivity, RNA replication, and assembly (4, 36). HCV is a member of the Flaviviridae family of enveloped viruses (30), with a positive-sense RNA genome of ∼9.6 kb consisting of a single open reading frame (ORF) that encodes 10 structural and nonstructural viral proteins (3, 16, 25). Cap-independent translation of the ORF (29) yields a large polyprotein of approximately 3,000 amino acid residues that is cleaved co- and posttranslationally by host and viral proteases into 10 mature virus proteins; these cleavage products are ordered from the amino to the carboxy terminus as follows: core (C), envelope proteins 1 and 2 (E1 and E2), p7, nonstructural protein 2 (NS2), NS3, NS4A, NS4B, NS5A, and NS5B (3, 16, 25). At the flanking ends of the genome are two highly conserved untranslated regions (UTRs). The 5′ UTR is highly structured and consists of the internal ribosome entry site (IRES), which is important for the initiation of cap-independent translation of the polyprotein (29). The 3′ UTR consists of a short genotype-specific variable region, a tract of variable length comprising solely pyrimidine residues (predominantly U), and a conserved 98-nucleotide sequence, known as the X region, containing three stem-loops (13, 23) (Fig. ​(Fig.1A).1A). The 3′ UTR is the initiation site for the synthesis of the negative-strand RNA during viral replication (13) and is involved in translational regulation.Open in a separate windowFIG. 1.The HCV 3′ UTR RNA. (A) The positive-strand 3′ UTR consists of three distinct regions, i.e., a short genotype-specific variable region, a polypyrimidine tract [poly(U/UC)] of variable length, and a conserved 98-nucleotide sequence known as the X region containing three stable stem-loops. The predicted structure of the genotype 1b 3′ UTR is shown. (B) Left panel, the integrities of in vitro-transcribed radiolabeled full-length 3′ UTR RNAs of genotypes 1b (nucleotides 9375 to 9595) and 2a (nucleotides 9443 to 9678) and the poly(U/UC) (nucleotides 9406 to 9497) and X region (nucleotides 9498 to 9595) of genotype 1b are shown on denaturing polyacrylamide gels. Right panel, the integrities of in vitro-transcribed radiolabeled RNAs comprising the 3′-terminal NS5B-coding region plus the 3′ UTR RNAs of genotypes 1b (nucleotides 9136 to 9595) and 2a (nucleotides 9204 to 9678) (KL-3′ UTR) are shown on denaturing polyacrylamide gels.HCV RNA replication occurs on membranous structures derived from the endoplasmic reticulum (ER) in a complex that includes host cell factors as well as viral nonstructural proteins, including NS5B, the RNA-dependent RNA polymerase (RdRp) which replicates the viral genome in vivo and in vitro (2, 25, 30). Initiation of the synthesis of the negative-strand RNA is thought to occur upon recognition and specific binding of the NS5B polymerase to the 3′ UTR of the genomic RNA (2, 16, 26). This replication activity and template specificity of NS5B in vivo are dependent, however, on the presence of the other nonstructural proteins, such as the proteases NS2 and NS3, which are required for polyprotein processing and helicase activity, and the multifunctional protein NS5A (16).NS5A is a proline-rich phosphoprotein that is absolutely required for viral replication and is also involved in virus particle assembly (9, 10, 20, 22, 35). Its specific function in the latter process is, however, still unknown. NS5A is membrane associated due to the presence of an N-terminal amphipathic helix that serves as a membrane anchor allowing association with ER-derived membranes (Fig. ​(Fig.2)2) (24, 27). The cytoplasmic portion of NS5A is organized into three domains that are separated by low-complexity sequences (Fig. ​(Fig.2A)2A) (20). The X-ray crystal structure of domain I has revealed that it is a zinc binding domain which forms a homodimer with contacts at the N-terminal ends of the molecules; the resultant large, basic groove at the dimeric interface has been proposed to be involved in RNA binding during viral replication (17, 33). NS5A has also been shown to interact with uridylate and guanylate-rich RNA and to bind to the 3′ ends of the HCV positive- and negative-strand RNAs (8). These observations suggest that NS5A may specifically interact with the large U/G stretches in the IRES of the 5′ UTR, implying a role in HCV translation and genome multiplication, while its interactions with the polypyrimidine tract of the 3′ UTR suggest that NS5A may affect the efficiency of RNA synthesis by NS5B (8, 28, 32). The reported interactions with both flanking regions of the HCV genome imply that NS5A may play a role in the switch between translation and replication that must occur during the viral life cycle (8).Open in a separate windowFIG. 2.Domain structure and expression of HCV NS5A. (A) Schematic diagram of the functional domains of NS5A and design of the constructs used in the study (genotype 1b NS5A protein numbering). The N-terminal amphipathic helix of NS5A (black box) is responsible for the interaction of NS5A with membranes. NS5A is organized into three domains that are separated by low-complexity sequences, indicated by black boxes. The NS5A constructs used all lacked the N-terminal amphipathic helix and were designed to include an N-terminal Strep tag and a C-terminal hexahistidine tag. (B and C) SDS-PAGE and Western blot analysis of the NS5A(ΔAH) and NS5A domain constructs purified by nickel affinity and Streptactin tag affinity chromatography. Coomassie brilliant blue-stained gels and Western blots (WB) using anti-NS5A antibodies for NS5A proteins of genotype 1b strain J4 (B) and genotype 2a strain JFH-1 (C) are shown.Among HCV genotypes, domains II and III are less well conserved than domain I (34). By mutational analysis, domain II, along with domain I, has been attributed to the replicase activity of NS5A (12). Contrastingly, domain III has been shown to be dispensable for RNA replication, and large heterologous insertions and deletions in this region can be tolerated, maintaining RNA replication (34). It has been shown, however, that these insertions and deletions within domain III do have an impact on virus particle assembly, highlighting the critical role of domain III NS5A in the viral life cycle (1, 10). Recent nuclear magnetic resonance (NMR) studies of domains II and III of NS5A revealed that they both adopt a natively unfolded state (6, 14, 15). The high degree of disorder and flexibility observed in these domains may contribute to the promiscuity of NS5A, which has been shown to interact with a variety of biological partners essential for NS5A function and virus persistence (11, 18, 19, 21, 31). In addition, regions within domains I and II of NS5A interact with NS5B, stimulating the in vitro activity of the polymerase and supporting the hypothesis that NS5A has a role in the modulation of RNA replication (28, 32).In this study, we have investigated in detail the RNA binding properties of NS5A. We have mapped the RNA binding regions of NS5A using bacterially expressed deletion constructs of NS5A and have assayed their binding affinity for HCV positive-strand 3′ UTR RNA. In addition, we provide evidence that the RNA binding activity of NS5A is specific and that NS5A interacts preferentially with the polypyrimidine region of the 3′ UTR.     

Detection of Nonstructural Protein 6 in Murine Coronavirus-Infected Cells and Analysis of the Transmembrane Topology by Using Bioinformatics and Molecular Approaches

Coronaviruses encode large replicase polyproteins which are proteolytically processed by viral proteases to generate mature nonstructural proteins (nsps) that form the viral replication complex. Mouse hepatitis virus (MHV) replicase products nsp3, nsp4, and nsp6 are predicted to act as membrane anchors during assembly of the viral replication complexes. We report the first antibody-mediated Western blot detection of nsp6 from MHV-infected cells. The nsp6-specific peptide antiserum detected the replicase intermediate p150 (nsp4 to nsp11) and two nsp6 products of approximately 23 and 25 kDa. Analysis of nsp6 transmembrane topology revealed six membrane-spanning segments and a conserved hydrophobic domain in the C-terminal cytosolic tail.Coronaviruses are enveloped, positive-stranded RNA viruses that sequester host cell membranes to assemble viral replication complexes in the cytoplasm of infected cells (2, 21). In the case of murine coronavirus mouse hepatitis virus (MHV), three viral proteases process the replicase polyproteins (3, 4, 5, 9, 12, 13, 14, 16, 18, 19, 24, 26) into intermediates and 16 mature nonstructural protein (nsp) products (Fig. ​(Fig.1A).1A). It is unclear whether the intermediate forms or the mature nsps are responsible for assembly of the viral replication complex. The replicase proteins nsp3, nsp4, and nsp6 contain transmembrane (TM)-spanning sequences that are proposed to be important for sequestering endoplasmic reticulum (ER) membranes to form the double-membrane vesicles which are the site of viral RNA synthesis (11, 17). However, the mechanism used by the nsps to generate double-membrane vesicles is not yet understood. Recent reports (8, 15, 22, 23, 28) and the study presented here have unraveled the membrane topology of these nsps. nsp4 is a glycoprotein with four TM domains (8, 22, 23, 28). nsp3 anchors its 213-kDa multidomain protein to ER membranes, likely using two TM domains (15, 22). Recently, nsp6 was shown to contain six TM domains (22); however, the authors were unable to resolve which of two C-terminal hydrophobic domains can act as the final membrane-spanning region.Open in a separate windowFIG. 1.Schematic diagram of MHV RNA genome, indicating the proteolytic processing scheme of the replicase polyprotein and Western blot detection of MHV nsp6. (A) MHV-A59 linear RNA genome with the canonical representation of replicase, structural, and accessory genes. The replicase polyprotein intermediates and mature nsps generated during processing are depicted. The mature nsp6 replicase protein (hatched box) and the antibodies used to detect nsp6 and nsp8 (solid black boxes) are indicated. aa''s, amino acids. (B) Western blot analysis of nsp6. Whole-cell lysates were prepared from mock-infected (M) and MHV-infected (I) HeLa-MHVR cells, and the lysates were separated by 12.5% SDS-PAGE. Products were detected by probing with nsp6- or nsp8-specific antibodies.In this report, we show the first antibody-mediated detection of MHV-A59 nsp6 in virally infected cells. We also report the TM topology of nsp6, as determined by glycosylation tagging and N-linked glycosylation sequence insertion mutagenesis approaches, providing evidence that nsp6 contains six membrane-spanning segments with a large C-terminal tail exposed to the cytosol. Multiple alignment of the nsp6 amino acid sequences from each coronavirus group revealed a high level of conservation at the C-terminal end, suggesting an evolutionarily conserved function.To detect nsp6 replicase protein in MHV-A59-infected cells, we used a polyclonal rabbit antiserum directed against a peptide (PLGVYNYKISVQEL) from the C-terminal region of nsp6. We detected the replicase intermediate p150 (nsp4 to nsp11) and two nsp6-specific products of 23 and 25 kDa (Fig. ​(Fig.1B,1B, lane 2) in MHV-infected HeLa-MHVR (25) cells by Western blot analysis. We found similar mature products of nsp6 in MHV-infected murine cell lines 17Cl-1 and DBT (data not shown). The same MHV-infected cell lysate was used to detect nsp8 replicase protein with a specific antibody that also recognizes p150 (Fig. ​(Fig.1B,1B, lane 4). The reason for the existence of multiple forms of nsp6 is currently unknown, although posttranslational modification or alternative processing of nsp6 cannot be ruled out at this point. Future experiments will be directed at purification and analysis of the two forms of nsp6 detected here.To develop a framework for understanding the membrane topology of nsp6, we first performed nsp6 bioinformatics analysis. Five out of the seven bioinformatics tools predicted that nsp6 would encode seven TM domains, whereas two programs predicted that it would encode eight TM domains (Fig. ​(Fig.2).2). However, because both the N and C termini of nsp6 must be processed in the cytosol by the viral 3C-like protease (3CLpro), we expected nsp6 to encode an even number of TM domains and established a consensus TM domain prediction for nsp6 (Fig. ​(Fig.2,2, bottom row). The consensus provided a working model for generating plasmid DNA constructs for evaluating the membrane topology of MHV nsp6. First, we employed enhanced green fluorescent protein (EGFP) glycosylation tagging (EGFPglyc) experiments as previously used for determining the membrane topology of other viral replicase TM proteins (20, 22). This approach allowed us to determine the localization of the tagged protein based on the differences in the mobility of the endoglycosidase H (endo H)-treated protein versus that of the untreated protein by the use of sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis. Based on the consensus topology model (Fig. ​(Fig.3A)3A) suggesting a maximum of eight TM domains, we generated a series of plasmid DNA constructs starting with the N-terminal putative TM1 domain, and successively larger constructs were fused at their carboxyl terminus in frame with EGFPglycV5. The plasmid DNAs were individually transfected into BsrT7 cells (6), and the newly synthesized fusion proteins were radiolabeled with 100 μCi of [³⁵S]methionine-cysteine per ml from 20 to 22 h posttransfection. Chimeric proteins expressed from the cell lysates were immunoprecipitated with V5 antibody, either endo H treated or mock treated, separated using 12.5% SDS-PAGE, and analyzed by autoradiography as described previously (15).Open in a separate windowFIG. 2.Summary of TM predictions for MHV nsp6 obtained from membrane topology bioinformatics tools. The nsp6 amino acid sequence (amino acids 3637 to 3923 in the MHV A59 genome are numbered 1 to 287 for simplicity) was analyzed for TM-spanning domains by the use of various bioinformatics tools, and the residue numbers with predicted TM domains are displayed. The consensus TM topology of MHV nsp6 used as a basis for the topology experiments is depicted at the bottom row (shaded in gray).Open in a separate windowFIG. 3.Determining the topology of nsp6 by the use of EGFPglyc and insertion of glycosylation consensus sites. (A) Schematics of a working topology model of MHV nsp6 (obtained from our consensus experiments) and nsp6-EGFPglycV5 fusion constructs generated for endo H assay. (B) Metabolic labeling and endo H treatment of nsp6-EGFPglycV5 fusion proteins. The nsp6-EGFPglycV5 fusion proteins expressed in transfected BsrT7 cells were radiolabeled from 20 to 22 h posttransfection, and then cell lysates were subjected to immunoprecipitation with V5 antibody, treated with endo H or left untreated, separated by 12.5% SDS-PAGE, and analyzed by autoradiography. (C) Map of plasmid DNA construct showing the sites of inserted glycosylation acceptor consensus sequences (NXS). The locations of glycosylation insertion in the nsp6-V5 construct are represented, with the amino acid number at the site of insertion. (D) Metabolic labeling and endo H analysis of glycosylation sequence insertion expression constructs of nsp6-V5. The plasmid DNAs (iNsp6-V5 constructs) were transfected and analyzed as described for panel B. (E) MHV-A59 nsp6 topology model, summarizing the results of EGFPglycV5 and glycosylation sequence insertion experiments. Amino acid positions indicated by the symbol “Y” were glycosylated and were positive by endo H assay, whereas those positions tested but found not glycosylated and negative by endo H assay are depicted by solid black horizontal lines. The inserted glycosylation acceptor sequence positions precede the letter i. Selected charged residues are shown in white characters on a black background. K, lysine residues; R, arginine residues; E, glutamic acid residues.We found that fusion protein products expressed from the reporter constructs (nsp6-35glycV5, nsp6-86glycV5, and nsp-165glycV5) were glycosylated, as shown by sensitivity to endo H treatment, indicating that the C-terminal end of these chimeric proteins must extend into the ER lumen (Fig. ​(Fig.3B,3B, lanes 4, 8, and 10). In contrast, the remaining reporter constructs were not sensitive to endo H treatment; therefore, the C-terminal end of the chimeric constructs must extend into the cytoplasm (Fig. ​(Fig.3B,3B, lanes 6, 12, 14, 16, and 18). Thus, these results indicate the presence of three luminal loops in nsp6. Identical results were obtained when we used PNGaseF (data not shown), which indicates that the lack of endo H sensitivity was not attributable to the protein transiting through the Golgi body, thereby rendering the protein insensitive to endo H treatment.To further investigate nsp6 topology in detail, we exploited a glycosylation sequence insertion mutagenesis approach (7) to create acceptor sequences in the region between amino acids 86 and 200 of nsp6 by the use of site-directed mutagenesis as described in reference 32 in order to independently investigate the topology, since bioinformatics predictions of the TM domains within this region differ (Fig. ​(Fig.2).2). Consensus glycosylation acceptor sites (NXS) were generated at four sites in the nsp6-V5 plasmid backbone by introducing single-codon insertions as depicted in Fig. ​Fig.3C.3C. All the glycosylation insertion constructs were expressed and analyzed by use of the endo H assay as described above. As expected, the parental nsp6-V5 protein is not glycosylated and did not show a mobility shift after endo H treatment (Fig. ​(Fig.3D,3D, lanes 1 and 2). In contrast, expression of 99iNsp6-V5 revealed evidence of endo H sensitivity (Fig. ​(Fig.3D,3D, lanes 3 and 4), indicating the ER luminal localization of the N99 introduced into MHV nsp6. This result is in agreement with those obtained with the nsp6-86glycV5 construct that is also endo H sensitive (Fig. ​(Fig.3B,3B, lanes 7 and 8). The insertion of glycosylation acceptor sequences at other sites yielded endo H-negative results (lanes 6, 8, and 10), indicating the possibility that the introduced NX(S/T) motifs (i) are localized in the cytosol, (ii) are localized within the membrane, or (iii) are not used, as the glycosylation site is not at least 12 amino acids away from the end of the preceding TM and 14 amino acids away from the beginning of the following TM (12 + 14 rule), thus rendering it inaccessible for glycosylation (1, 7, 30). Our results confirm and extend the results of a recent study (22) in which authors were unable to resolve whether TM6 or TM7 acted as the final TM domain. Our results indicate that TM6 is the final TM domain for MHV nsp6. We propose a topology model of MHV-A59 nsp6 in Fig. ​Fig.3E3E which is in accordance with the distribution of positively charged residues (positive inside rule; reviewed in reference 31), depicting the higher number of lysine and arginine residues facing the cytosolic side of the membrane and the majority of charged residues excluded from the TM domain. Taken together, the results presented above are consistent with a six-TM domain model of MHV nsp6. This report provides new information on the membrane topology of nsp6 and provides potential clues with respect to the assembly of the coronavirus replication complex.To determine whether the experimentally determined six-TM-spanning domain topology of MHV-A59 is conserved among coronaviruses, we performed MUSCLE (10) and ClustalW (29) multiple sequence alignment of nsp6 amino acid sequences representing group 1, group 2, and group 3 coronaviruses obtained from PATRIC (http://patric.vbi.vt.edu/) (27). The most striking observations were the amino acid sequence conservation in the C terminus of all nsp6 proteins and the conservation in the hydrophobicities within the putative TM domains (Fig. ​(Fig.4).4). This analysis revealed several conserved sites that may be important for the function of nsp6. We designated the conserved region between TM2 and TM3 the “KH loop” because of the invariant lysine and histidine residues that are present in the cytosolic loop (Fig. ​(Fig.3E),3E), although the function of these amino acids is not yet known. We also designated the hydrophobic region in the C-terminal tail the “conserved hydrophobic domain” (Fig. ​(Fig.4).4). We speculate that cysteine residue(s) within the region we designated the “conserved G(X)C(X)G motif” may be modified by palmitoylation, indicating that this region of nsp6 may have important functions in establishing protein-protein or protein-membrane interactions during the assembly of the viral replication complex. Additionally, for the MHV-A59 nsp6 protein, the NetPhosK 1.0 server (http://www.cbs.dtu.dk/services/NetPhosK/) predicted serine and tyrosine residues (serine 244 and tyrosine 250; see Fig. ​Fig.4)4) at the C-terminal region as sites of possible phosphorylation by epidermal growth factor receptor kinase and protein kinase C, respectively. Both predicted sites are highly conserved in all coronavirus nsp6 proteins (Fig. ​(Fig.4).4). Overall, our analysis revealed conserved features in the nsp6 C-terminal region whose importance in viral replication can be investigated using a coronavirus reverse genetics system.Open in a separate windowFIG. 4.Multiple sequence alignment (MSA) and percent sequence identity of coronavirus nsp6. The nsp6 amino acid sequences of 18 different coronaviruses were obtained from PATRIC (http://patric.vbi.vt.edu/) and aligned using MUSCLE and ClustalW software. The experimentally determined TM domains of MHV-A59 nsp6 were used as a reference for alignment. Unshaded boxes indicate the conserved TM domains that aligned with other coronavirus nsp6 sequences; the conserved hydrophobic domain (CHD) predicted by all the topology programs is indicated by gray shading. The residues of the peptide against which the nsp6 antibody was raised are boxed, with residue designations shown in boldface. Putative sites for palmitoylation (cysteine residue[s]) within the GXCXG motif) and phosphorylation (serine 244 and tyrosine 250 in MHV-A59 nsp6) are indicated. Percent identity (% ID) values are indicated. In MSA, the following notations were used: asterisk indicate invariant amino acids, colons indicate highly similar amino acids, and dots indicate similar amino acids. HCoV, human coronavirus; PHEV, porcine hemagglutinating encephalomyelitis virus; BCoV, bovine coronavirus; BatSARS, bat severe acute respiratory syndrome coronavirus; BatCoV, bat coronavirus; SARSCoV, severe acute respiratory syndrome coronavirus; FIPV, feline infectious peritonitis virus; PRCoV, porcine respiratory coronavirus; TGEV, transmissible gastroenteritis virus; PEDV, porcine epidemic diarrhea virus; IBV, infectious bronchitis virus.     

Association of Rice Gall Dwarf Virus with Microtubules Is Necessary for Viral Release from Cultured Insect Vector Cells

Vector insect cells infected with Rice gall dwarf virus, a member of the family Reoviridae, contained the virus-associated microtubules adjacent to the viroplasms, as revealed by transmission electron, electron tomographic, and confocal microscopy. The viroplasms, putative sites of viral replication, contained the nonstructural viral proteins Pns7 and Pns12, as well as core protein P5, of the virus. Microtubule-depolymerizing drugs suppressed the association of viral particles with microtubules and prevented the release of viruses from cells without significantly affecting viral multiplication. Thus, microtubules appear to mediate viral transport within and release of viruses from infected vector cells.Rice gall dwarf virus (RGDV), Rice dwarf virus (RDV), and Wound tumor virus, members of the genus Phytoreovirus in the family Reoviridae, multiply both in plants and in invertebrate insect vectors. Each virus exists as icosahedral particles of approximately 65 to 70 nm in diameter, with two concentric layers (shells) of proteins that enclose a core (1, 13). The viral genome of RGDV consists of 12 segmented double-stranded RNAs that encode six structural (P1, P2, P3, P5, P6, and P8) and six nonstructural (Pns4, Pns7, Pns9, Pns10, Pns11, and Pns12) proteins (reference 21 and references therein). The core capsid is composed of P3, the major protein, which encloses P1, P5, and P6 (12). The outer layer consists of two proteins, namely, P2 and P8 (10, 12).Cytoplasmic inclusion bodies, known as viroplasms or viral factories, are assumed to be the sites of replication of viruses in the family Reoviridae. After infecting insect vector cell monolayers (VCMs) in culture with RDV, Wei et al. (19) examined the generation of RDV particles in and at the periphery of such viroplasms. VCMs are also useful for studies of RGDV, allowing detailed analysis of the synchronous replication and multiplication of this virus (14). In order to identify the viroplasms in RGDV-infected VCMs, we examined the subcellular localization of Pns7, Pns12, P5, and RGDV particles by confocal immunofluorescence microscopy. Pns7 and Pns12 of RGDV correspond to Pns6 and Pns11, respectively, which are components of the viroplasm of RDV (12, 19). RGDV P5 is a counterpart of RDV P5, a core protein that locates inside the viroplasm in RDV-infected cells. We inoculated VCMs with RGDV, purified by the method reported in reference 15, at a multiplicity of infection (MOI) of 1; fixed them 48 h postinfection (p.i.); probed the cells with Pns7-, Pns12-, P5-, and viral-antigen-specific antibodies (11, 12) that had been conjugated to fluorescein isothiocyanate (FITC) (Sigma, St. Louis, MO) or rhodamine (Sigma); and examined them by confocal microscopy, as described previously (19). In RGDV-infected cells, Pns7, Pns12, and P5 were detected as punctate inclusions (Fig. ​(Fig.1).1). Immunostained viral antigens formed ringlike structures around the punctate inclusions. When the images were merged, Pns7, Pns12, and P5 were colocalized in the punctate inclusions, indicating that these proteins were constituents of the viral inclusions (Fig. ​(Fig.1).1). Our observations revealed the similar respective localizations of the corresponding nonstructural proteins, core proteins, and viral particles of two phytoreoviruses, RGDV and RDV, in infected cells. Thus, Pns7 and Pns12 of RGDV had attributes common to their functional counterparts—Pns6 and Pns11, respectively—of RDV (19). The core protein P5 was located inside the viroplasms, and the viral antigens were distributed at the periphery of the viroplasms. The results, together, suggest that RGDV and RDV exploit similar replication strategies. Specific fluorescence was not detected in noninfected cells after incubation with Pns7-, Pns12-, P5-, and viral-antigen-specific antibodies (data not shown).Open in a separate windowFIG. 1.Subcellular localization of Pns7, Pns12, and P5 of RGDV and viral antigens in RGDV-infected VCMs 48 h p.i. Arrowheads show ringlike profiles of viral antigens that surround viral inclusions, which have been immunostained with the Pns12-specific antibodies. Arrows show the fibrillar profiles of immunostained viral antigens. Bars, 5 μm.In addition to the viral location at the periphery of the viral inclusions visualized as immunostained Pns12 (Fig. ​(Fig.1),1), the antigens were distributed as bundles of fibrillar structures, a form not observed in RDV-infected cells. To analyze the entity of the bundles of fibrillar structures, VCMs on coverslips were inoculated with RGDV at an MOI of 1, fixed at 48 h p.i., and examined by electron microscopy (EM), as described previously (19). We observed viral particles of approximately 70 nm in diameter in close association with the free ends, as well as along the edges, of tubules of approximately 25 nm in diameter (Fig. 2A to D). The abundant bundles of tubules with closely associated viral particles were clearly in contact with the periphery of granular, electron-dense inclusions of 800 to 1,200 nm in diameter (Fig. ​(Fig.2B),2B), namely, viroplasms. The dimensions and appearance of the tubular structures resembled those of microtubules (Fig. ​(Fig.2C)2C) (17). Transverse sections of tubules revealed arrays of closed circles of approximately 25 nm in diameter, with viral particles attached directly or via a filament to the circumference (Fig. ​(Fig.2D2D).Open in a separate windowFIG. 2.Association of RGDV particles with microtubules. (A) Electron micrograph showing RGDV particles associated with microtubules in virus-infected VCMs 48 h p.i. Bar, 300 nm. VP, electron-dense inclusion. (B) Virus-associated microtubules in contact with the periphery of the electron-dense inclusion indicated by a white rectangle in panel A. Bar, 300 nm. (C) Viral particles along the edges of tubules of approximately 25 nm in diameter. Bar, 150 nm. (D) Transverse sections of arrays of closed circles of approximately 25 nm in diameter with viral particles attached to their circumference directly (arrow) or via a filament (arrowhead). Bar, 150 nm. (E) Confocal micrograph showing the association of viral particles with microtubules in virus-infected VCMs 48 h p.i. Microtubules were stained with α-tubulin-specific antibodies conjugated to FITC; viral particles were stained with viral-antigen-specific antibodies conjugated to rhodamine. Arrowheads indicate the ringlike organization of viral antigens. Arrows show the colocalization of fibrillar profiles of viral antigens with microtubules. The insets show ringlike and fibrillar profiles of immunostained viral antigens. The circular areas inside the ringlike structures are viroplasms. Bar, 5 μm.Our observations suggested that RGDV particles might attach to microtubules in infected cells. To examine this possibility, we inoculated VCMs with RGDV at an MOI of 1, fixed the cells 48 h p.i., immunostained them with α-tubulin-specific antibodies conjugated to FITC and with viral-antigen-specific antibodies conjugated to rhodamine, and analyzed them by confocal microscopy, as described previously (19). Viral antigens were visualized as ringlike and fibrillar structures (Fig. ​(Fig.2E).2E). Double immunostaining of the infected cells revealed that a network of microtubule-based filaments colocalized with most of the fibrillar structures that represented viral antigens, confirming the association of viral particles with the microtubule-like inclusions visualized by EM (Fig. ​(Fig.2A).2A). Nonspecific reactions were not detected with either of the stainings (data not shown). Our results suggested that RGDV particles, which assembled at the periphery of viroplasms, might be transported along microtubules. Due to the lack of RGDV infectious clones fused with green fluorescent protein and the effective gene transfection system for VCMs, we could not observe the trafficking of RGDV particles along microtubules in living cells.We then used three-dimensional (3-D) electron tomographic microscopy (ET) to reveal a new level of morphological detail about the association of RGDV with microtubules. To produce 3-D reconstructions of RGDV-infected cells, we fixed, embedded, and sectioned infected leafhopper cells as described previously (5). We chose a representative region that showed numerous RGDV particles close to bundles of microtubules for this novel tomographic analysis. A single-axis tilt series was collected manually from −60° to 60° with 2° increments using an H9500SD EM (Hitachi, Tokyo) operated at 200 kV. These tomographic data were recorded at a defocus of 3.6 μm on the TVIPS 2k × 2k charge-coupled-device camera (TVIPS, Gauting, Germany). Microscopic magnification of ×15,000, providing 1.28 nm/pixel, was enough to view the microtubules and virus particles following tomographic reconstruction of the tilt series using IMOD (7). As shown in the 3-D tomogram in Fig. ​Fig.3,3, most of the RGDV particles were bound to the edges of bundles of microtubules. The RGDV particles along the edges of microtubules were arrayed in an orderly but uncrowded manner (Fig. ​(Fig.3).3). Our ET analysis also revealed that some viral particles were linked to filaments of approximately 10 nm in diameter (Fig. ​(Fig.3B).3B). Morphologically, these filaments resembled vimentin intermediate filaments (4). In many lines of cultured cells, vimentin intermediate filaments partially overlap the microtubules, and there is evidence that the two filament systems interact (3, 9, 20). Unfortunately, vimentin-specific monoclonal antibodies from mouse and rabbit did not react specifically with our leafhopper cells (data not shown), but the nature of the intermediate filaments was apparent from their dimensions, intracellular location, and organization. Thus, our ET analysis indicated that RGDV particles were able to associate directly and/or via intermediate filaments with microtubules.Open in a separate windowFIG. 3.ET analysis showing the association of RGDV particles with microtubules either directly or via intermediate filaments. (A) Translucent representation of the reconstructed viruses lining up with microtubules. (B) Slice of the reconstructed volumes from the inset of A to show the association of RGDV particles with intermediate filaments (arrows). Bars, 150 nm.To examine the role of the microtubules for RGDV activity, we added a microtubule-disrupting agent, either nocodazole (Sigma) or colchicine (Sigma), 2 hours after inoculation of VCMs with RGDV at an MOI of 1 and then continued the incubation for a further 46 h. Cells were fixed 48 h p.i. and stained with α-tubulin-specific antibodies conjugated to FITC (Sigma) and viral-particle-specific antibodies conjugated to rhodamine, with subsequent confocal fluorescence microscopy, as described previously (19). We tested a range of drug concentrations in preliminary experiments (data not shown) and determined optimal concentrations. Treatment of infected cells with 10 μM nocodazole or 5 μg/ml colchicine resulted in the complete disassembly of microtubules, with the accumulation of ringlike structures exclusively and no fibrillar structures representative of viral antigens in the cytoplasm (Fig. ​(Fig.4A).4A). These ringlike aggregates of viral antigens were confirmed to surround viroplasms when the latter were immunostained for Pns12, as described above and shown in Fig. ​Fig.1.1. Nonspecific reactions were not detected with either staining (data not shown). These results suggest that RGDV particles multiply around the viroplasm but are unable to distribute along the microtubules in the presence of the chemicals.Open in a separate windowFIG. 4.(A) Effects of microtubule-disrupting agents on the formation of microtubules and fibrillar profiles of immunostained viral antigens. Bars, 5 μm. The insets show the ringlike profiles of immunostained viral antigens after treatment with inhibitors, suggesting that viral replication occurs in the presence of each inhibitor. (B) Effects of drugs on the production of cell-associated (gray bars) and extracellular (black bars) viruses in VCMs infected with RGDV. The error bars indicate standard deviations.During the process of infection, microtubules play important roles in viral entry, intracellular trafficking, and extracellular release (2, 8, 16). We next investigated the effects of the microtubule-disrupting agents on the production in and release of viruses from virus-infected cells by the method described previously (18). Nocodazole or colchicine was added 2 h after inoculation of VCMs with RGDV at an MOI of 1, and incubation was continued for a further 46 h. The extracellular medium and the cells were collected separately. The medium was centrifuged for 30 min at 15,000 × g, and the supernatant was collected. The cells were subjected to three cycles of freezing and thawing to release viral particles. The viral titer of each sample was determined, in duplicate, by the fluorescent focus assay as described previously (6), with VCMs and a magnification of ×10. As shown in Fig. ​Fig.4B,4B, nocodazole (20 μM) and colchicine (10 μg/ml) caused a fivefold reduction in the number of released viruses, compared to that from untreated control infected cells. In contrast, each inhibitor at the selected dose failed to significantly reduce the titer of cell-associated viruses (less then 5% compared to that from untreated control). These results suggest that the inhibitors impeded the release of viruses into the medium without affecting viral production in infected cells. We do not yet understand why the viral titer was not elevated in drug-treated cells from which viral release was inhibited. However, our data show clearly that disruption of microtubules directly inhibited the release of mature viral particles from infected cells.In conclusion, EM, ET, immunofluorescence staining, and experiments with two inhibitors support the hypothesis that the transport of RGDV from viroplasms to the plasma membrane and into the medium is dependent on microtubules. In the case of RDV, vesicular compartment-containing viral particles that locate adjacent to the viroplasms were considered to play an important role in the transport and release of the virus from the viroplasm to the culture medium in infected VCMs (18). On the other hand, a fibrillar structure (Fig. ​(Fig.11 and ​and2),2), not observed in RDV-infected cells, was considered to function in the trafficking of RGDV from viroplasm into the culture medium (Fig. ​(Fig.4)4) in the present study. RGDV and RDV, both members of the Phytoreovirus genus, have some common biological and biochemical properties but are distinct from each other (13). For example, viruses are restricted to phloem-related cells in RGDV-infected plants but distributed in many types of cells in RDV-infected plants, and a P2 protein with a function to adsorb to and/or penetrate into insect vector cells is present in RGDV and absent in RDV in particles purified using carbon tetrachloride. The present molecular cytopathological study revealed one more difference between the viruses: they have different means for transporting and releasing infectious particles to the cell exterior. The presence of such a molecular mechanism may accelerate the secondary infections by the viruses in infected vector insects, and the high propagation speed would allow the viruses to complete infection cycles through insects and plants.     

Molecular Chaperone BiP Interacts with Borna Disease Virus Glycoprotein at the Cell Surface

Borna disease virus (BDV) is characterized by highly neurotropic infection. BDV enters its target cells using virus surface glycoprotein (G), but the cellular molecules mediating this process remain to be elucidated. We demonstrate here that the N-terminal product of G, GP1, interacts with the 78-kDa chaperone protein BiP. BiP was found at the surface of BDV-permissive cells, and anti-BiP antibody reduced BDV infection as well as GP1 binding to the cell surface. We also reveal that BiP localizes at the synapse of neurons. These results indicate that BiP may participate in the cell surface association of BDV.Borna disease virus (BDV) belongs to the Bornaviridae family of nonsegmented, negative-strand RNA viruses and is characterized by highly neurotropic and noncytopathic infection (18, 33). BDV infects a wide variety of host species and causes central nervous system (CNS) diseases in animals, which are frequently associated with behavioral disorders (14, 19, 29, 31). BDV cell entry is mediated by endocytosis, following the attachment of viral envelope glycoprotein (G) to the cellular receptor (2, 7, 8). BDV G is translated as a precursor protein, GP, which is posttranslationally cleaved by the cellular protease furin to generate two functional subunits of the N (GP1) and C (GP2) termini (28). Recent studies revealed that GP1 is involved in virus interaction with as-yet-unidentified cell surface receptor(s) and that GP2 mediates a pH-dependent fusion event between viral and cell membranes (2, 7, 27). In addition, a previous work using a hippocampal culture system suggested that BDV G is required for viral dissemination in neurons (2); however, cellular factors involved in BDV cell entry, especially cell surface association, remain to be elucidated.To extend our understanding of the role of BDV G in the interaction with the cell plasma membrane, we transfected GP1 fused with hemagglutinin-tobacco etch virus protease cleavage site-FLAG tags (GP1-TAP) into human oligodendroglioma OL cells. GP1-TAP was purified using anti-FLAG M2 affinity gel (Sigma). To verify that GP1-TAP binds to OL cells, the cells were incubated with 4 μg/ml GP1-TAP, and binding was detected by anti-FLAG M2 antibody (Sigma). A flow cytometric analysis indicated that GP1-TAP binds to OL cells (Fig. ​(Fig.1A).1A). To further validate the binding of GP1-TAP, we tested whether GP1-TAP inhibits BDV infection. OL cells were pretreated with 4 μg/ml GP1-TAP for 30 min. Proteins purified from mock-transfected cells using an anti-FLAG M2 affinity gel served as a control. The cells were then mixed with cell-free BDV. After 1 h of absorption, the supernatants were removed and fresh medium was added. At 3 days postinfection, the viral antigens were stained with anti-nucleoprotein (N) monoclonal and anti-matrix (M) polyclonal antibodies. As shown in Fig. ​Fig.1B,1B, GP1-TAP reduced BDV infection by 40% compared to levels for mock-treated cells. This result was consistent with earlier reports showing that recombinant GP1 protein binds to the cell surface and inhibits BDV infection (6, 20).Open in a separate windowFIG. 1.BDV GP1 binds to the cell surface. (A) Binding of BDV GP1 to OL cells. OL cells were incubated with GP1-TAP (solid line), and its binding was detected using anti-FLAG M2 antibody and flow cytometry. As a control, cells incubated with proteins purified from mock-transfected cells were detected by an anti-FLAG M2 antibody (dotted line). (B) Inhibition of BDV infection by GP1. OL cells pretreated with GP1-TAP were inoculated with the BDV huP2br strain. Values are the means + standard deviations (SD) from three independent experiments. **, P < 0.01.To investigate the host factor(s) that mediates the interaction of GP1 with the cell surface, a combination of tandem affinity purification (TAP) and liquid chromatography tandem mass spectrometry analyses was designed (13). We transfected GP1-TAP into OL cells and then purified GP1 from cell homogenates using a TAP strategy. We compared the purified proteins from the whole-cell and cytosol fractions (Fig. ​(Fig.2A),2A), and the bands detected only in the whole-cell fraction were determined as GP1-binding proteins in the membrane and/or nuclear fractions. In addition to GP1 protein (Fig. ​(Fig.2A,2A, arrow), we identified a specific band around 80 kDa in the whole-cell homogenate, but not in the cytosol fraction (Fig. ​(Fig.2A,2A, arrowhead), and determined that the band corresponded to the BiP (immunoglobulin heavy chain-binding protein) molecular chaperone, also called glucose-regulated protein 78 (GRP78), by mass spectrometry analysis. We confirmed the specific interaction between endogenous BiP and BDV G in infected cells by immunoprecipitation analysis (Fig. ​(Fig.2B).2B). To map the binding domain on BiP to GP1, we constructed a series of deletion mutants of the green fluorescent protein (GFP)-tagged BiP plasmid (Fig. ​(Fig.2C).2C). We transfected the mutant plasmids into BDV-infected OL cells and then performed an immunoprecipitation assay using anti-GFP antibody (Invitrogen). As shown in Fig. ​Fig.2D,2D, BDV G was coimmunoprecipitated with truncated BiP mutants, except for BiPΔN-GFP, which lacks the ATP-binding domain of BiP (lane 3), suggesting that BiP interacts with GP1 via its N-terminal region.Open in a separate windowFIG. 2.BDV GP1 interacts with BiP molecular chaperone. (A) TAP analysis of BDV GP1. Proteins coimmunoprecipitated with GP1-TAP in OL cells were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and visualized by silver staining. Cyt, cytosol fraction; Wc, whole-cell homogenate. Arrow, GP1-TAP; arrowhead, BiP. (B) Coimmunoprecipitation (IP) of BDV G and endogenous BiP. BDV G was immunoprecipitated from BDV-infected OL cells by anti-BDV G polyclonal antibody. Endogenous BiP was then detected by anti-BiP monoclonal antibody (Becton Dickinson). IgG, immunoglobulin G. (C) Schematic representation of deletion mutants of recombinant BiP-GFP. The known functional regions are indicated. (D) Immunoprecipitation analysis of BiP-GFP mutants in BDV-infected OL cells. The deletion plasmids were transfected and immunoprecipitated by anti-GFP antibody. Specific binding was detected using anti-BDV G antibody. Lane 1, GFP; lane 2, BiP-GFP; lane 3, BiPΔN-GFP; lane 4, BiPΔPB-GFP; lane 5, BiPΔC-GFP.BiP is known to be resident primarily in the endoplasmic reticulum and functions as a molecular chaperone involved in the folding process of nascent proteins, mostly through interaction with its peptide-binding domain (12, 17, 21). On the other hand, BiP has been reported to serve as a coreceptor of certain viruses at the plasma membrane (15, 34). Recent studies also revealed that cell surface BiP mediates the internalization of its ligands into cells (1, 10). We first investigated whether BiP is expressed on the cell surface of BDV-permissive OL and 293T cells using an anti-BiP polyclonal antibody (H-129; Santa Cruz Biotechnology, Inc.). As shown in Fig. ​Fig.3A,3A, BiP expression is detected on the surface of both cell lines. This result is in agreement with recent observations that BiP is expressed on the surface of various types of cells (9, 10, 15, 23, 24, 34). We also investigated whether BiP is expressed on the cell surface of BDV-nonpermissive cell lines, such as HeLa and CHO cells. As shown in Fig. ​Fig.3A,3A, we detected BiP expression on the surface of HeLa, but not CHO, cells. These observations were confirmed by immunofluorescence analysis (Fig. ​(Fig.3B).3B). Note that BiP is clearly detected at the endoplasmic reticulum in the permeabilized CHO cells by the antibody (see Fig. S1 in the supplemental material), suggesting that BiP is expressed at a very low level, if at all, on the surface of CHO cells. We next examined whether cell surface BiP serves as a binding molecule of BDV GP1. To test this, we performed an inhibition assay using an anti-BiP polyclonal antibody (N-20; Santa Cruz Biotechnology, Inc.) which recognizes the N terminus of BiP. As shown in Fig. ​Fig.3C,3C, the antibody inhibited GP1 binding to the cell surface by 40%. Furthermore, BDV infection was found to decrease by 70% when cells were treated with the antibody (Fig. ​(Fig.3D3D).Open in a separate windowFIG. 3.Cell surface BiP mediates cell association of BDV. (A) Flow cytometric analysis was performed with anti-BiP antibody (H-129) in BDV-permissive (OL and 293T) and -nonpermissive (HeLa and CHO) cells (solid lines). Cells stained with normal rabbit immunoglobulin G were used as a control (dotted lines). (B) Immunofluorescence analysis was performed by using anti-BiP antibody (H-129) with BDV-permissive and -nonpermissive cells. Arrows indicate BiP staining at the membrane. Scale bars, 10 μm. (C) Inhibition of GP1 binding by anti-BiP antibody (N-20). OL cells were pretreated with anti-BiP antibody, followed by labeling with GP1. GP1 binding on the cell surface was detected using flow cytometry. Values are the means + SD from three independent experiments. *, P < 0.05. (D) Inhibition of BDV infection by anti-BiP antibody. OL cells were incubated with 10 μg/ml anti-BiP antibody or normal goat immunoglobulin G and then the cells were mixed with cell-free BDV. After 1 h absorption, the supernatants were replaced with fresh medium. Virus infection was measured by immunofluorescence analysis using anti-N and -M antibodies at 3 days postinfection. Values are the means + SD from three independent experiments. *, P < 0.05. IgG, immunoglobulin G.To investigate the role of cell surface BiP in the infection of BDV, the BiP expression was inhibited by short interfering RNA (siRNA) in OL cells (see Fig. S2A in the supplemental material). We selected an siRNA (Hs_HSPA5_4; Qiagen, Inc.) which could partially downregulate the cell surface expression of BiP (see Fig. S2B in the supplemental material). However, siRNA treatment of BiP did not influence the infectivity of BDV in OL cells (see Fig. S2C in the supplemental material). This may be due to an incomplete reduction of BiP expression on the cell surface. Alternatively, while BiP mediates at least in part the cell surface association of BDV particles, this result may exhibit the presence of another, as-yet-unidentified BDV G-binding protein that is involved in the binding and subsequent cell entry of BDV.Previous studies demonstrated that BDV can be traced centripetally and transsynaptically after olfactory, ophthalmic, or intraperitoneal inoculation (3, 25). Migration of BDV to the CNS after footpad infection can be prevented by sciatic nerve transection (3). These observations suggest that BDV may disseminate primarily via neural networks. Recently, it has been demonstrated that BDV G was expressed at the termini of neurites or at contact sites of neurites (2), suggesting that local assembly of BDV may take place at the presynaptic terminals of synapses, similar to assembly of other neurotropic viruses (22, 26, 32). If BiP localizes at synapse sites, BiP may efficiently participate in the transmission of BDV particles at the synapses. To evaluate this hypothesis, we examined BiP localization in primary culture of mouse hippocampal neurons. After in vitro culture for 17 days, BiP localization was determined by an immunofluorescence assay without permeabilization. As shown in Fig. ​Fig.4A,4A, BiP signals were clearly detected at neurites, including the contact sites between dendrites and axons, as punctate staining (arrows), suggesting that BiP is expressed at the neuronal surface, most likely at the synapses. We next examined the localization of BiP with postsynaptic density 95 (PSD-95), a marker of postsynaptic density (5). Although BiP signals were detected mainly in the perinuclear area of the hippocampal neurons, punctate staining was also found at neurites colocalized with PSD-95 (Fig. ​(Fig.4B,4B, arrows). Taken together, these observations suggested that BiP is distributed at the synaptic surface, including the postsynaptic membrane, of neurons, a possible site for BDV budding and entry (2).Open in a separate windowFIG. 4.BiP localizes at the synaptic surface of hippocampus neurons. (A) Localization of BiP at synaptic surface. Hippocampal neurons were immunostained with anti-BiP antibody (N-20) without permeabilization. A differential interference contrast (DIC) image is shown. Dotted lines in the Merge panel indicate the dendrite outline. Arrows indicate BiP staining at the contact sites between axons and dendrites. (B) Colocalization between BiP and a postsynaptic protein. Hippocampal neurons were immunostained with anti-BiP (N-20) and anti-PSD-95 (Millipore) antibodies. Arrows indicate colocalized signals of BiP and PSD-95 at neurites. Scale bars, 10 μm.In summary, this study demonstrates that BiP is a GP1-binding protein at the synaptic surface. This is the first report showing the BDV G-binding factor on the cell surface. The first step of BDV entry might be mediated by the interaction of GP1 with as-yet-unidentified cell surface receptors, which may form a complex with other molecules, such as BiP. We showed that treatment with anti-BiP antibody affects BDV infection as well as GP1 binding to the cell surface (Fig. ​(Fig.3).3). Furthermore, synaptic distribution of BiP was found in hippocampal primary neurons (Fig. ​(Fig.4).4). These findings strongly suggest that BiP plays critical roles in BDV association with the neuronal surface via interaction with GP1. On the other hand, a BDV-nonpermissive cell line, HeLa, appeared to express BiP on the cell surface, suggesting that the cell surface BiP may not be necessarily involved in the infectivity of BDV. A recent study by Clemente et al. (6) revealed that following initial attachment to the cell surface, BDV is recruited to the plasma membrane lipid raft (LR) prior to internalization of the particles. The study suggested that BDV may use the LR as a platform to interact with additional host cell factor(s) required for efficient BDV internalization. Because BiP does not contain transmembrane regions, BiP needs another host protein(s) with transmembrane regions on the cell surface. It has been reported that cell surface BiP interacts with diverse proteins, such as major histocompatibility complex class I molecules (34), the voltage-dependent anion channel (9), and the DnaJ-like protein MTJ-1 (4), all of which associate with LR in the plasma membrane (16, 24, 35). Once BDV has attached to the cell surface, it might utilize such BiP-associated LR proteins for efficient cell surface attachment or internalization. Previously, it has been proposed that kainate 1 (KA-1) receptor might represent the BDV receptor within the CNS (11). Because some glutamate receptors are shown to bind to BiP (30), KA-1 receptors might interact with BiP and serve as a receptor complex for BDV. Further studies are required for a full understanding of the cell association processes, especially receptor binding, of BDV.      

Enhanced Gene Silencing in Cells Cured of Persistent Virus Infection by RNA Interference

We compared HEp-2-derived cells cured of persistent poliovirus infection by RNA interference (RNAi) with parental cells, to investigate possible changes in the efficiency of RNAi. Lower levels of poliovirus replication were observed in cured cells, possibly facilitating virus silencing by antiviral small interfering RNAs (siRNAs). However, green fluorescent protein (GFP) produced from a measles virus vector and also GFP and luciferase produced from plasmids that do not replicate in human cells were more effectively silenced by specific siRNAs in cured than in control cells. Thus, cells displaying enhanced silencing were selected during curing by RNAi. Our results strongly suggest that the RNAi machinery of cured cells is more efficient than that of parental cells.Small interfering RNAs (siRNAs) mediate RNA interference (RNAi), a natural biological phenomenon regulating a wide range of cellular pathways (8, 20). RNAi-based therapies with siRNAs or small hairpin RNAs (shRNAs) have been developed against several viral infections, and a reduction of the viral yield by several orders of magnitude has frequently been obtained (4, 9). However, virus clearance from cells and the complete cure of persistent virus infections have only rarely been reported (24, 25). We have developed several models of persistent virus infection by using poliovirus (PV), a positive-strand RNA virus of the Picornaviridae family (5, 7, 16, 21). We previously studied the effects of antiviral siRNAs applied months after the infection of HEp-2 cells with a persistent PV mutant (7, 25). We used a mixture (“the Mix”) of two synthetic siRNAs targeting the viral RNA genome in the 5′ noncoding (NC) region and the 3D polymerase (3D^pol) (siRNA-5′NC and siRNA-3D^pol, respectively; synthesized by Sigma-Proligo). When repeated transfections with the Mix were performed in persistently PV-infected cultures, most cultures stopped producing virus (25). Here, we investigate the important issue of changes in RNAi efficacy following siRNA treatment, 2 to 5 months after the cure. The efficiency of gene silencing in cells was stable during this period.We used the HEp-Q4 and -Q5 cell lines, which were cured of persistent PV infection after transfections with the Mix (25). The cured cells and their parental cell line, HEp-2, had similar growth rates (data not shown). To compare PV silencing efficiencies in the three cell lines, they were transfected either with the Mix or with an irrelevant siRNA (siRNA-IRR) in the presence of Lipofectamine 2000 (Invitrogen) in 24-well plates as previously described (25). Treated and mock-treated cells were infected 16 h posttransfection with PV strain Sabin 3, at a multiplicity of infection (MOI) of 1 50% infectious dose (ID₅₀) per cell. The viral progeny was titrated 24 h postinfection, as previously described (16). HEp-Q4 and HEp-Q5 were permissive to PV infection, although viral yields were about 1 log lower in these cells than in HEp-2 cells (Fig. ​(Fig.1A).1A). Virus silencing was observed in all three cell lines treated with the Mix; however, silencing was significantly more efficient in HEp-Q4 (≈2.2 times more efficient; P = 0.013, Student''s t test) and HEp-Q5 (≈5.6 times more efficient; P = 0.015) than in HEp-2 cells (Fig. 1A and B). Similar results were obtained with an shRNA (Thermo Scientific) targeting the same region as the siRNA-5′NC (data not shown).Open in a separate windowFIG. 1.Efficiency of enterovirus silencing in HEp-2, HEp-Q4, and HEp-Q5 cells after transfection with specific siRNAs. (A) Yield of progeny virus produced by cells infected at an MOI of 1 ID₅₀, 16 h posttransfection with the antiviral Mix containing two anti-PV siRNAs (20 pmol), the irrelevant siRNA-IRR (20 pmol), or no siRNA. Samples were harvested 24 h postinfection. Each bar represents the mean value ± SEM of six infected cultures from three independent experiments. (B to E) For each cell line, silencing efficiency is expressed as the ratio of infectious virus yield (titer in ID₅₀/ml) in the presence of the irrelevant siRNA-IRR to infectious virus yield (titer in ID₅₀/ml) in the presence of the antiviral siRNAs in cured cells, normalized with respect to the silencing efficiency in HEp-2 cells. S2, PV strain Sabin 2. (F) GFP silencing efficiency for each cell line is expressed as a ratio [1 − (mean GFP levels in the presence of siRNA-eGFP)/(mean GFP levels in the presence of siRNA-IRR)] in cured cells, normalized with respect to the efficiency of silencing in HEp-2 cells. Each bar represents the mean value ± SEM of at least four cultures from two independent experiments. *, P < 0.05 based on Student''s t test comparing HEp-Q4 and HEp-Q5 with HEp-2 cells.We investigated whether the differences in silencing efficacies between the three cell lines were due to differences in siRNA transfection efficiency by transfecting HEp-2, HEp-Q4, and HEp-Q5 cells with fluorescein isothiocyanate-conjugated siRNA (siRNA-FITC; 20 pmol/well; Cell Signaling) and testing them between 4 and 48 h posttransfection. The fluorescence of transfected cells was measured with a FACScan flow cytometer (Becton Dickinson), and data were analyzed with CellQuest software (Becton Dickinson). The percentages of siRNA-FITC-positive cells were similar for all cell types (Fig. ​(Fig.2A).2A). The mean fluorescence per positive cell and the percentage of cells displaying fluorescence peaked 16 and 24 h posttransfection, respectively, and decreased thereafter (Fig. ​(Fig.2).2). These findings suggest both that the presence of siRNAs in cells was similarly transient in the three cell types, as previously reported (27), and that the high silencing efficiencies in cured cells were not a consequence of higher transfection efficiencies. All subsequent experiments were performed between 16 and 40 h posttransfection.Open in a separate windowFIG. 2.Transfection efficiencies of fluorescein-conjugated siRNAs in HEp-2, HEp-Q4, and HEp-Q5 cells. A fluorescent siRNA-FITC (20 pmol) was used to transfect each of the three cell lines in the presence of Lipofectamine 2000. Fluorescent cells were analyzed 4 to 48 h posttransfection by using a FACScan flow cytometer (Becton Dickinson). The percentage of fluorescent cells (A) and the mean fluorescence per positive cell, in arbitrary units (B), are shown. Each bar represents the mean value ± SEM. (C) Representative FACS plots (cell granularity versus cell size), showing the similarities between the three cell populations.Fluorescence-activated cell sorting (FACS) plots for granularity versus cell size were very similar for the three cell lines (Fig. ​(Fig.2C),2C), as were those for cell numbers versus fluorescence (not shown), suggesting highly related cell populations. Although highly probable, it remains to be confirmed that the cured cells originated from a subpopulation of HEp-2 cells.Virus silencing was also investigated in cured cells infected with Sabin 2 or coxsackievirus A17 (CAV17) strain 67591 (22) or in cells transfected with Sabin 2 RNA. The experimental conditions used for Sabin 2 and CAV17 were identical to those for Sabin 3, except that only the 3D polymerase was targeted by siRNAs. Sabin 2 RNA (1 μg) was prepared as previously described (12) and used with siRNA-3D^pol (20 pmol/well) for the cotransfection of cells in the presence of Lipofectamine 2000. Virus yields were determined 7.5 h after transfection. In all cases, virus silencing was more effective in HEp-Q4 and -Q5 cells than in HEp-2 cells (Fig. 1C to E). Additional experiments were performed with a PV replicon encoding the green fluorescent protein (GFP), PV-eGFP (28) (2 μg/well), which was used with siRNA-eGFP (20 pmol/well; Ambion) for cotransfection. GFP fluorescence was measured by flow cytometry, 16 h after transfection. As for PV, a higher silencing efficiency was observed in cured cells than in HEp-2 cells (Fig. ​(Fig.1F1F).We then investigated whether the lower level of viral multiplication in HEp-Q4 and -Q5 cells in the absence of siRNAs involved an entry or postentry step. We quantified the expression of the PV receptor (CD155) at the surface of cells. We used flow cytometry after indirect immunofluorescence labeling with anti-CD155 antibodies, as previously described (16). More than 98.4% ± 2% (mean ± standard error of the mean [SEM]) of cured cells, like HEp-2 cells, tested positive for CD155 (data not shown). In the absence of siRNAs, a decrease in viral replication was also observed in HEp-Q4 and -Q5 cells infected with the Sabin 2 PV strain in cells, in which the early stages of the viral cycle were bypassed by transfection with Sabin 2 RNA, and in cells infected with the CAV17 virus, which uses a cell receptor other than CD155 (12) (data not shown). Together, these results suggest that PV multiplication is reduced at a postentry step, probably at replication, in cured cells.We investigated whether PV silencing was also enhanced in other HEp-derived cells in which Sabin 3 PV multiplication was reduced by using HEp-S31 (cl18) cells that had been cured of persistent PV infection by growth at a supraoptimal temperature rather than by RNAi (2). PV yield was ≈1.6 logs lower in HEp-S31 (cl18) cells than in HEp-2 cells (data not shown). Sabin 3 PV silencing in HEp-S31 (cl18) cells was 1.7 ± 0.9 times more effective (mean of six experiments) than that in HEp-2 cells (relative efficacy of 1) (data not shown), but this difference was not significant. However, these results do not exclude the possibility that reduced PV replication facilitates PV silencing by the Mix in cured cells. We therefore pursued our work with a different virus.We investigated whether the high silencing efficiency in HEp-Q4 and -Q5 cells was specific to enteroviruses by using a measles virus expressing GFP, MV-eGFP (26), and siRNA-eGFP to silence GFP expression. Cells were transfected with either siRNA-eGFP or siRNA-IRR, infected with MV-eGFP (1 ID₅₀ per cell, 16 h posttransfection), and the GFP silencing efficiency was determined 40 h posttransfection by flow cytometry. For each cell line, silencing efficiency was expressed as a percentage {[1 − (percentage of siRNA-eGFP-transfected cells expressing GFP)/(percentage of siRNA-IRR-transfected cells expressing GFP)] × 100}. GFP silencing was significantly stronger in HEp-Q4 cells (≈14%; P = 0.048) and HEp-Q5 cells (≈17%; P = 0.010) than in HEp-2 cells (Fig. ​(Fig.3A).3A). There was no significant difference in the silencing efficiency of GFP between HEp-Q4 and -Q5 cells (Fig. ​(Fig.3A).3A). The anti-PV Mix did not silence GFP expression (data not shown), indicating that the silencing of GFP was not due to anti-PV siRNAs persisting in cured cells months after the initial treatment.Open in a separate windowFIG. 3.Efficiency of GFP and luciferase silencing in HEp-2, HEp-Q4, and HEp-Q5 cells after transfection with specific siRNAs. (A and B) GFP silencing, expressed as a percentage calculated for each cell line as follows: {[1 − (GFP expression in the presence of siRNA-eGFP)/(GFP expression in the presence of the irrelevant siRNA-IRR)] × 100}. (A) Cells were infected 16 h posttransfection with a measles virus encoding eGFP (MV-eGFP [26]) at an MOI of 1 ID₅₀/cell, and fluorescent cells were analyzed 24 h after infection (40 h posttransfection). Each bar represents the mean value ± SEM of three independent experiments. (B) Cells were cotransfected with pEGFP-C1 and siRNA-eGFP or siRNA-IRR and analyzed 40 h later. Each bar represents the mean value ± SEM of four independent experiments. (C) Luciferase silencing efficiency for each cell line, expressed as the ratio of luciferase activity in the presence of the irrelevant siRNA-IRR to luciferase activity in the presence of the specific siRNAs in cured cells, normalized with respect to silencing efficiency in HEp-2 cells. Relative efficiencies are shown as in Fig. ​Fig.11 for luciferase, because the enzymatic reaction amplified the signal. Each bar represents the mean value ± SEM of triplicates from three independent experiments. *, P < 0.05 based on Student''s t test comparing HEp-Q4 and HEp-Q5 with HEp-2 cells.To test whether the high silencing efficiency in HEp-Q4 and -Q5 cells was dependent on viral infection, plasmid vectors pEGFP-C1 (Clontech Laboratories) and pRL-CMV (Promega) were used to generate GFP (6) and Renilla luciferase (18), respectively. These plasmids do not replicate in human cells. Cells (10⁶) were cotransfected with pEGFP-C1 (1 μg) and siRNAs (20 pmol) in the presence of Lipofectamine 2000, as recommended by the manufacturer. GFP fluorescence was analyzed by flow cytometry 40 h posttransfection. Silencing efficiencies were expressed as a percentage {[1 − (mean GFP levels in the presence of siRNA-eGFP)/(mean GFP levels in the presence of siRNA-IRR)] × 100)}. Mean silencing efficiency was significantly higher in HEp-Q4 (≈15%; P = 0.003) and HEp-Q5 (≈15%; P = 0.002) cells than in HEp-2 cells (Fig. ​(Fig.3B).3B). The efficiency with which the GFP encoded by pEGFP-C1 was silenced was similar in HEp-Q4 and -Q5 cells.The efficacy of siRNAs was then assessed with pRL-CMV, which encodes the Renilla luciferase and Silencer Renilla luciferase (AM4630; Ambion). Cells (10⁶) were cotransfected with the plasmid (100 ng) and either specific or irrelevant siRNA (7 pmol) in the presence of Lipofectamine 2000. Luciferase assays were performed with a Dual-Glo luciferase assay system (Promega), as recommended by the manufacturer at 40 h posttransfection, and luminescence was measured with a luminometer (Centro LB960; Berthold). The results of the sensitive luciferase assays confirmed that the relative efficiency of silencing was significantly higher in cured than in parental cells (Fig. ​(Fig.3C).3C). By contrast, results obtained in HEp-S31 (cl18) cells, cured without siRNAs, were not significantly different from those obtained in control HEp-2 cells (data not shown), strongly suggesting that the treatment of HEp-Q4 and -Q5 cells with specific siRNAs selected cells in which siRNAs mediated silencing more efficiently than in parental cells.The difference in silencing efficiency between cured and HEp-2 cells may be due to differences in the abundance and/or efficacy of cellular factors involved in gene silencing. Some major actors of the RNAi pathway, particularly those associated with the RNA-induced silencing complex (RISC), have been identified (3, 10, 13, 19). The active endonucleolytic core of the RISC includes the guide strand of the siRNA and a slicer protein called Argonaute 2 (Ago2) (17). We used Western blotting to study Ago-2 and other factors contributing to the function of RISC (3, 10, 11, 14, 19, 23): the endonuclease Dicer, the transactivation response RNA binding protein (TRBP), the protein activator of double-stranded RNA-dependent protein kinase (PACT), and the RNA helicase A (RHA) (Fig. ​(Fig.4).4). Exportin 5, which plays a role upstream from the dicing process in the export of small RNA precursors (29), was included as a control.Open in a separate windowFIG. 4.Comparative analysis of proteins involved in RNAi in HEp-2, HEp-Q4, and HEp-Q5 cell lines. Whole-cell lysates were tested for Exportin 5 (A), Dicer (B), Ago-2 (C), the helicase RHA (D), TRBP (E to H) and PACT (I) by Western blotting with the corresponding specific antibodies. Blots were subsequently stripped and reprobed with antiactin antibodies to confirm equal protein loading. (E and F) TRBP levels in HEp-Q4 and HEp-Q5 cells were determined by densitometry and are plotted in arbitrary units, as ratios relative to the level of actin and to the level of TRBP in HEp-2 cells. In panel F the symbols correspond to TRBP levels determined in nine different experiments. (G) TRBP levels in HEp-2 cells transfected with pcDNA-TRBP (14) and in cells cotransfected with pcDNA-TRBP and siRNA-TRBP. (H) TRBP levels were compared in human IMR5 cells, HEpS31 (cl18) cells previously cured of persistent PV infection by growth at a supraoptimal temperature, and the control HEp-2 cell line. TRBP/actin densitometry and PACT/actin densitometry results are indicated in arbitrary units in the histograms below the corresponding Western blot results shown in panels H and I.Proteins (30 to 50 μg) from each cell line were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (10 to 20% Tricine gels; Invitrogen) and transferred to nitrocellulose membranes (Amersham Biosciences) as previously described (1). The membranes were incubated with one of the following primary antibodies (1): anti-Ago2 monoclonal antibody (MAb; Abcam), anti-RHA MAb (Abcam), and anti-TRBP2 MAb (Santa Cruz Biotechnology); rabbit antibodies against Dicer (Santa Cruz Biotechnology); anti-PACT MAb (Santa Cruz Biotechnology), and anti-Exportin 5 MAb (Abcam). The antiactin MAb (AC-40; Sigma-Aldrich) was used to check for equal protein loading. Membranes were then washed and treated with appropriate horseradish peroxidase-conjugated secondary antibodies (Amersham Biosciences) for 2 h at room temperature. Protein bands were detected with an enhanced chemiluminescence detection kit (ECL⁺; Amersham Biosciences) and a G:box (Syngene).Exportin 5, Dicer, Ago-2, and RHA were similarly abundant in all three cell lines (Fig. 4A to D), suggesting that quantitative differences in protein levels were unlikely to be responsible for the enhanced silencing in HEp-Q4 and -Q5 cells. There was significantly more TRBP in HEp-Q4 (≈21%; P = 0.026) and HEp-Q5 (≈28%; P = 0.016) cells than in HEp-2 cells, as indicated by the results of nine experiments (Fig. 4E and F). The specificity of the anti-TRBP antibody was checked on extracts of HEp-2 cells transfected with a plasmid encoding TRBP, pcDNA-TRBP (14), with and without silencing by siRNA-TRBP (Fig. ​(Fig.4G).4G). GFP silencing was not enhanced in HEp-2 cells overproducing TRBP, and it was not decreased by downregulating TRBP gene expression with siRNA-TRBP (data not shown). These results suggest that the high levels of TRBP in the cured cell lines are not the cause of the enhanced silencing in these cells.There was less TRBP protein in HEp-S31 (cl18) cells (2) than in HEp-2 and other control cells (IMR5) (Fig. ​(Fig.4H),4H), indicating that high levels of TRBP are not necessarily selected in cells persistently infected with PV. PACT was slightly downregulated in the cured cells (Fig. ​(Fig.4I).4I). Moreover, PACT is unlikely to be involved in the enhanced silencing in cured cells, because we used synthetic siRNAs and PACT functions principally during siRNA production by Dicer (14). We did not investigate the activities or subcellular distributions of the various factors involved in RNAi in the three cell lines, and they may differ. It is also possible that other factors, not tested here, contribute to the efficacy of siRNAs in cured cells. The molecular details of the mechanism involved remain to be determined.Overall, our results suggest that both a decrease in viral replication and the enhancement of gene silencing contributed to the mechanism by which cells persistently infected with poliovirus were cured by RNAi. Our results also indicate that cells displaying enhanced silencing may be selected during treatment with siRNAs. This may result in profound changes to cell phenotype, because RNAi plays an essential role in the regulation of cellular gene expression (15).