|Home | About | Journals | Submit | Contact Us | Français|
Like almost all of the positive-strand RNA viruses, hepatitis C virus (HCV) induces host intracellular membrane modification to form the membrane-bound viral replication complex (RC), within which viral replicases amplify the viral RNA genome. Despite accumulated information about how HCV co-opts host factors for viral replication, our knowledge of the molecular mechanisms by which viral proteins hijack host factors for replicase assembly has only begun to emerge. Purification of the viral replicase and identification of the replicase-associated host factors to dissect their roles in RC biogenesis will shed light on the molecular mechanisms of RC assembly. To purify the viral replicase in the context of genuine viral replication, we developed an HCV subgenomic replicon system in which two different affinity tags were simultaneously inserted in frame into HCV NS5A and NS5B. After solubilizing the replicon cells, we purified the viral replicase by two-step affinity purification and identified the associated host factors by mass spectrometry. We identified valosin-containing protein (VCP), a member of the ATPases associated with diverse cellular activities (AAA+ATPase) family, as an active viral replication modulator whose ATPase activity is required for viral replication. A transient replication assay indicated that VCP is involved mainly in viral genome amplification. VCP associated with viral replicase and colocalized with a viral RC marker. Further, in an HCV replicase formation surrogate system, abolishing VCP function resulted in aberrant distribution of HCV NS5A. We propose that HCV may co-opt a host AAA+ATPase for its replicase assembly.
IMPORTANCE Almost all of the positive-strand RNA viruses share a replication strategy in which viral proteins modify host membranes to form the membrane-associated viral replicase. Viruses hijack host factors to facilitate this energy-unfavorable process. Understanding of this fundamental process is hampered by the challenges of purifying the replicase because of the technical difficulties involved. In this study, we developed an HCV subgenomic replicon system in which two different affinity tags were simultaneously inserted in frame into two replicase components. Using this dual-affinity-tagged replicon system, we purified the viral replicase and identified valosin-containing protein (VCP) AAA+ATPase as a pivotal viral replicase-associated host factor that is required for viral genome replication. Abolishing VCP function resulted in aberrant viral protein distribution. We propose that HCV hijacks a host AAA+ATPase for its replicase assembly. Understanding the molecular mechanism of VCP regulates viral replicase assembly may lead to novel antiviral strategies targeting the most conserved viral replication step.
The positive-strand RNA viruses are the largest class of viruses and include many medically and economically important pathogens, including hepatitis C virus (HCV); picornaviruses, which cause hand, foot, and mouth disease; and flaviviruses, such as the West Nile virus and the Zika virus. Positive-strand RNA viruses share a conserved replication mechanism in which viral proteins induce host membrane modification to assemble membrane-associated viral replication complexes (RCs) (1). Viruses hijack host factors to facilitate this energy-unfavorable process (2).
HCV, a member of the Flaviviridae family, chronically infects approximately 160 million people worldwide and causes hepatocellular carcinoma in a significant proportion of the chronically infected population (3). Its 9.6-kb positive-sense RNA genome encodes a single polyprotein that is cleaved into at least 10 individual polypeptides by the host and viral proteinase in the following protein order: 5′-C-E1-E2-p7-NS2-NS3-NS4A-NS4B-NS5A-NS5B-3′ (reviewed in reference 4). The single open reading frame (ORF) is flanked by the highly conserved 5′ and 3′ untranslated regions (UTRs). The 5′ UTR contains an internal ribosome entry site (IRES) to initiate cap-independent translation. The 3′ UTR is required for RNA replication and is composed of three sequential elements: a nonconserved variable region (30 to 50 nucleotides), a poly(U/C) stretch (20 to 200 nucleotides), and a conserved 98-nucleotide sequence, termed the 3′X region, which contains three stem-loop structures (reviewed in reference 4). Upon polyprotein processing, HCV nonstructural proteins residing in the endoplasmic reticulum (ER) induce the formation of double-membrane vesicles (DMVs), which are protrusions from the ER membranes toward the cytosol (5). Nonstructural proteins NS3, NS4A, NS4B, NS5A, and NS5B constitute the viral replicase. NS3 is a bifunctional enzyme with protease and helicase activities, with NS4A as a protease cofactor. NS4B is a multispanning integral membrane protein. NS5A is a multifunctional viral protein with no enzymatic activity. NS5B is an RNA-dependent RNA polymerase (reviewed in reference 6).
Overexpression of the HCV polyprotein encompassing NS3 to NS5B can induce DMVs, as observed in virus-infected cells. Overexpression of NS5A alone, although it is less efficient, can also induce DMVs, which argues for the central role of NS5A in RC assembly (5). HCV NS5A recruits host factors and, combined with other viral proteins, locally modifies the membrane lipid composition to facilitate RC assembly (7). Despite great advances in the studies of HCV RC biogenesis, insights into the underlying molecular mechanism of how viral proteins hijack host factors for replicase assembly have only begun to emerge.
Identification of replicase-associated host factors and dissection of their roles in replicase assembly will shed light on the molecular mechanisms of RC biogenesis. Purification of the viral replicase during authentic viral replication is hampered by the membrane association of the replicase, the low expression levels of the replicase components, and the complexity of the replicase modules. The replicase-enriched intracellular membrane fractions and the associated host factors have been studied by selective fractionation of the RC-associated detergent-resistant membranes (8). Anti-NS4B antibody (9) and tagged-NS4B replicons (10) were used to affinity capture the replicase-containing membrane fractions. Chung et al. reported purification of the tagged-NS5A-interacting proteins during viral infection (11). There is no report regarding purification of the solubilized HCV replicase as a protein complex.
In this study, we developed an HCV subgenomic replicon system in which two different affinity tags were simultaneously inserted in frame into HCV NS5A and NS5B without obviously attenuating viral replication, which allows us to purify the NS5A- and NS5B-containing replicase by sequential affinity purification. Affinity purification of the viral replicase and identification of the associated host factors uncovered valosin-containing protein (VCP), a member of the ATPases associated with diverse cellular activities (AAA+ATPase) family, as a pivotal host factor required for viral replication.
The plasmid encoding HCV Jc1FLAG2 (p7-nsGluc2A, named Jc1G here) was reported previously (12). To generate Jc1G with the duplicated NS5B stem-loop SL3, first a PacI site was introduced into the Jc1G downstream stop codon of NS5B and then a PCR-amplified NS5B SL3 fragment flanked by PacI recognition sites was ligated into the PacI site in Jc1G to get Jc1G-DU. To generate Jc1G-NS5B.HA, a nucleotide sequence encoding a hemagglutinin (HA) peptide (TAC CCA TAC GAT GTT CCA GAT TAC GCT) was inserted into the NS5B region by fusion PCR. HCV subgenomic replicon construct sgJFH1 was generated by replacing the fragment of Jc1G that encompasses the sequence extending from the 5′ UTR to the structural protein region with a fragment from the BB7 subgenomic replicon encompassing the 5′ UTR-BSD-EMCV IRES region (13). To generate sgJFH1-DU-NS5B.HA1, sgJFH1-DU-NS5B.HA2, and sgJFH1-DU-NS5B.His constructs, first an RsrII/XbaI fragment was transferred from Jc1G-DU into sgJFH1 to get the sgJFH1-DU construct and then the nucleotides encoding the HA peptide, the HA peptide with flanking amino acids (GAC GCT TAC CCA TAC GAT GTT CCA GAT TAC GCT AGC CTA), and the His6 (CAT CAT CAT CAT CAT CAC) peptide were inserted into the NS5B region by fusion PCR to get sgJFH1-DU-NS5B.HA1, sgJFH1-DU-NS5B.HA2, and sgJFH1-DU-NS5B.His, respectively. To generate the sgJFH1-NS5A.HA construct, the nucleotide encoding the HA peptide (TAC CCA TAC GAT GTT CCA GAT TAC GCT) was inserted into the NS5A region of sgJFH1 by fusion PCR as previously reported (14). The sgJFH1-NS4B.HA construct was generated essentially as previously reported, with the additional adaptive mutations (10). To generate the sgJFH1-NS5A.HA.NS5B.His construct, a BsrG1/SanD1 fragment was transferred from the sgJFH1-NS5A.HA construct into the sgJFH1-DU-NS5B.His construct. Plasmid sgJFH1-NS5A.ypet was generated by transferring an SpeI/RsrII fragment from the Jc1G-378-1 construct that includes an in-frame fusion of NS5A.Ypet (Charles Rice, unpublished data) into the sgJFH1 plasmid. To generate the phCMV-3-5B.ypet construct, first plasmid phCMV-3-5B was generated by inserting the PCR-amplified HCV NS3-5B region from Jc1G into the BglII/EcoRI sites in the phCMV vector (Genlantis) and then a SanDI/HindIII fragment was transferred from the Jc1-378-1 plasmid into phCMV-3-5B.
The lentivirus-based short hairpin RNA (shRNA) plasmids were generated with a pLKO.1 (Sigma) backbone. The oligonucleotides targeting the VCP coding region (GCA TTA CTG GTA ATC TCT TCG) and an irrelevant target (IRR) (TCT CGC TTG GGC GAG AGT AAG) were synthesized and ligated into pLKO.1.
For CRISPR-mediated knockdown (Kd), the oligonucleotides targeting VCP sg1 (GAT GAA TTG CAG TTG TTC CG), sg2 (GAC ACA GTG ATC CAC TGC GA), sg3 (AAA CTC ACC TCT CGT TTG AT), and sg4 (GCG TAT CGA CCC ATC CGG AA) and a scrambled oligonucleotide (IRR) (ATA GCG ACT AAA CAC ATC AA) were synthesized and ligated into LentiCRISPR-V2 (Addgene plasmid no. 52961).
To generate the retroviral vectors Lenti-RFP and Lenti-VCP, PCR amplicons were digested with XbaI/BamH1 and ligated into the Lenticas9-BLAST plasmid (Addgene plasmid no. 52962) to replace the Cas9 ORF. The VCP gene fragment was amplified from Huh7 cDNA, which was generated by reverse transcription by superscript III (Invitrogen). The red fluorescent protein (RFP) fragment was amplified from plasmid pTRIP-RFP (15). To generate Lenti-RFP-VCP, in which RFP is inserted in frame upstream of the VCP coding region, the amplified RFP-encoding gene was digested with BamHI and ligated into similarly digested Lenti-VCP. The retroviral vector pTRIP, which was reported previously (15), was modified to contain a blasticidin resistance gene (BSD) and a multicloning site and renamed pTRIP-BSD. PCR-amplified green fluorescent protein (GFP) and VCP-es, which contains shRNA-resistant mutations (GG ATA ACA GGA AAC CTG TTT G), were digested with XbaI and BamHI and ligated into similarly digested pTRIP-BSD to get plasmids pTRIP-BSD-GFP and pTRIP-BSD-VCP-es. The dominant negative mutant construct pTRIP-VCP-es-DN (E305Q/E578Q) was generated by site-directed mutagenesis. All of the constructs were proofed by DNA sequencing. Detailed information is available upon request.
Human hepatoma cell lines Huh7 (Cell Bank of the Chinese Academy of Sciences, Shanghai, China, http://grbio.org/institution/cell-bank-type-culture-collection-chinese-academy-sciences) and Huh7.5 (kindly provided by Charles Rice) were routinely maintained in Dulbecco modified medium supplemented with 10% fetal bovine serum (Biological Industries catalog no. 04-001-1), 25 mM HEPES (Gibco), and nonessential amino acids (Gibco). Replicon cells were generated by electroporation of in vitro-transcribed RNAs into Huh7.5 cells and selected in conditioned medium supplemented with 5 μg/ml blasticidin (Invitrogen). The resulted replicon cells were pooled and maintained with the conditioned medium supplemented with 0.5 μg/ml blasticidin. Huh7.5 cells harboring pTRIP-BSD-GFP, pTRIP-BSD-VCP-es, and pTRIP-BSD-VCP-es-DN were generated by transducing Huh7.5 cells with vesicular stomatitis virus G protein (VSV-G)-pseudotyped lentiviral particles and growing cells in conditioned medium supplemented with 5 μg/ml blasticidin. The surviving cells were pooled and maintained in conditioned medium with 5 μg/ml blasticidin. For HCVcc infection, these three stable cell lines with exactly the same number of passages were used to avoid the effect of cell passage variation on HCV replication (16).
Anti-β-actin antibody from Sigma (A1978) was used in Western blotting analyses at a 1:5,000 dilution. Anti-c-Myc antibody from Santa Cruz Biotechnology (sc-40) was utilized at a 1:1,000 dilution for Western blotting. Anti-NS5A antibody (9E10) was used at a 1:2,000 dilution for Western blotting analyses. Anti-GFP antibody (sc-9996; Santa Cruz) was used for Western blotting at a 1:2,000 dilution. Rabbit anti-VCP antibody (ABclonal; A2795) was used at a 1:2,000 dilution for Western blotting. A mouse monoclonal anti-VCP antibody (Abcam; ab11433) was used at a 1:50 dilution for immunostaining. An anti-His antibody (Abcam; ab18184) was used at a 1:2,000 dilution for Western blotting. An anti-HA antibody (Abcam; ab130275) was used at a 1:1,000 dilution for Western blotting. An anti-HCV NS3 antibody (Virogen; 217-A) was used at a 1:1,000 dilution for Western blotting. Horseradish peroxidase (HRP)-conjugated anti-mouse (Santa Cruz; sc-2005) and anti-rabbit IgG secondary antibodies (Santa Cruz; sc-2004) were used at a 1:2,000 dilution for Western blotting. Cy3-conjugated anti-mouse IgG (Jackson, 115-165-062) was used at a 1:200 dilution for immunostaining. EerI was purchased from Tocris Bioscience (catalog no. 3922) and dissolved in dimethyl sulfoxide to make a 10 mM stock.
HCVcc was generated as described previously, by electroporation of in vitro-transcribed Jc1G RNA into Huh7.5 cells, and the virus titer was determined in Huh7.5 cells by limiting dilution (17). VSV-G-pseudotyped lentiviral particles were generated by cotransfection of HEK293T cells with plasmids encoding VSV-G and HIV gag-pol and with the lentiviral provirus plasmids. The medium overlying the cells was harvested at 48 to 72 h after transfection, filtered through a 0.45-μm filter, and stored at −80°C. Cells were transduced with the pseudoparticles in the presence of 8 μg/ml Polybrene. The pseudoparticle dosage was determined by limiting dilution to produce about 60% transduced-cell survival in conditioned medium supplemented with 5 μg/ml puromycin (for pLKO.1 and LentiCRISPR_v2) or 5 μg/ml blasticidin (for pTRIP-BSD).
For small-scale experiments, HCV replicon cells harboring sgJFH1 and sgJFH1-NS5A.HA.NS5B.His were seeded onto a 10-cm tissue culture dish. When subconfluent, the cells were scraped into Dulbecco's phosphate-buffered saline (PBS) (DPBS; Biological Industries, catalog no. 02-023-1) and centrifuged at 500 × g for 5 min. Cell pellets were resuspended in 500 μl of ice-cold lysis buffer (50 mM Tris Cl [pH 7.5], 1 mM EDTA, 15 mM MgCl2, 10 mM KCl, 1% Triton X-100, proteinase inhibitor cocktail [Roche]). After disruption by being passed through a 27-gauge needle 20 times and clarification by centrifugation at 15,000 × g for 10 min at 4°C, the soluble fraction was incubated with 30 μl of anti-HA antibody-coated beads (Sigma, A 2095) for 4 h at 4°C. The beads were collected by centrifugation and then washed four times with 1.2 ml of washing buffer (50 mM Tris Cl [pH 7.5], 1 mM EDTA, 15 mM MgCl2, 10 mM KCl, 1% Triton X-100, 20 mM imidazole). After being washed, the bound proteins were eluted with 50 μl of elution buffer (50 mM Tris Cl [pH 7.5], 1 mM EDTA, 15 mM MgCl2, 10 mM KCl, 1% Triton X-100, 250 μg/ml HA peptide [Sigma, I2149]) by incubation in a Thermomixer R (Eppendorf) at room temperature for 15 min (1,000 rpm). After centrifugation at 3,000 × g for 5 min, the supernatants were collected and diluted to 300 μl by lysis buffer supplemented with 20 mM imidazole. A 20-μl volume of Ni-Sepharose (GE Healthcare, 17-5268-01) was then added to the supernatants. After incubation at 4°C for 1 h and clarification, the Sepharose was washed four times with 600 μl of washing buffer containing 20 mM imidazole. The captured proteins were eluted in 30 μl of washing buffer containing 240 mM imidazole and mixed with 30 μl of 2×SDS loading buffer (100 mM Tris Cl [pH 6.8], 4% [wt/vol] SDS, 0.2% [wt/vol] bromophenol blue, 20% [vol/vol] glycerol, 10% 2-mercaptoethanol). After being boiled for 5 min, the proteins samples were separated by SDS-PAGE and transferred to a nitrocellulose membrane for immunoblotting.
For large-scale experiments, the cells from 10 150-mm tissue culture dishes were harvested and pooled. After centrifugation, the cells were resuspended in 3 ml of lysis buffer and broken by 20 strokes in a Dounce tissue grinder. After clarification, the soluble fraction was incubated by head-to-tail rotation with 200 μl of anti-HA antibody-coated beads (Sigma, A2095) for 4 h at 4°C. The beads were collected by centrifugation and then washed four times with 8 ml of washing buffer. The proteins were eluted in 300 μl of washing buffer supplemented with HA peptide (250 μg/ml), followed by a second elution with 100 μl of washing buffer supplemented with HA peptide (250 μg/ml). The eluted solutions were combined and diluted into 1.2 ml of lysis buffer containing 20 mM imidazole, and 40 μl of Ni Sepharose was added. After incubation at 4°C for 1 h, the beads were washed four times with 1.2 ml of washing buffer. The captured proteins were eluted in 40 μl of washing buffer containing 240 mM imidazole and mixed with 10 μl of 5× SDS loading buffer (250 mM Tris Cl [pH 6.8], 30% glycerol, 10% SDS, 0.02% bromophenol blue). After the addition of 2-mercaptoethanol to a final concentration of 5%, the eluted solution was boiled for 5 min. Proteins were resolved with NuPAGE Novex Bis-Tris Mini Gels (Invitrogen), and protein bands were visualized by silver staining.
RNAs were purified by TRIzol extraction and reverse transcribed with the PrimeScript RT reagent kit with gDNA Eraser (Perfect Real Time) from TaKaRa according to the manufacturer's instructions. The cDNA samples were subjected to real-time PCR (SYBR Premix Ex Taq Tli RNase H Plus) with the following primers for specific genes: GAPDH, (s [sense]) GGT ATC GTG GAA GGA CTC ATG A and (as [antisense]) ATG CCA GTG GCT TCC CGT TCA GC; HCV, (s) CCC TGT GAG GAA CTA CTG TCT TCA CGC and (as) GCT CAT GGT GCA CGG TCTA CGA GAC CT; ANKFY1, (s) CAG AGA AGC GCTG CGC TC and (as) CTC CAG CTG TCA CTG CGG; SHROOM3, (s) GGC GCA GTG AGC CTG C and (as) TGG TCA GGG CTC CGC C; SEC16A, (s) GAC ACA GCC CAG AGC ACA and (as) CTG CAG ACC AGG CTG AGG; VCP, (s) CCC AGC CCA AGA TGG ATG and (as) ATC CCC TAG GCG TAC ACG; SEC24C, (s) CCC CGC CAT TCC CTA TGG and (as) CTG AGA CCC AGG CAG CC; MOV10, (s) GAC AGA TCG CGA GCG GC and (as) CTC GACC GCT GGG CCG A; UPF1, (s) GCC AAG ACC AGC CAG TTG and (as) GCA GGG TCA CCT CTT TGC. The relative RNA levels were calculated by the 2−ΔΔCT method (18). The gene for glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as a housekeeping gene for loading normalization.
Before transfection, 1.25 × 105 Huh7.5 cells were seeded into a 24-well plate. When the cells had grown to nearly 80 to 90% confluence, 0.5 μg of RNA per well was transfected with the TransIT-mRNA Transfection kit (Mirus) at a ratio of 0.5:1:1 (RNA to boost to transfection reagent) according to the manufacturer's protocol.
Supernatants were taken from each well and mixed with an equal volume of 2× passive lysis buffer (Promega). Luciferase activity was measured with Renilla luciferase substrate (Promega) according to the manufacturer's protocol, or cultured cells, after being washed with DPBS, were lysed with 1× passive lysis buffer and luciferase activity was measured with Renilla luciferase substrate.
Cells were directly lysed with 2× SDS loading buffer (100 mM Tris-Cl [pH 6.8], 4% [wt/vol] SDS, 0.2% [wt/vol] bromophenol blue, 20% [vol/vol] glycerol, 10% 2-mercaptoethanol) and boiled for 5 min. Proteins were separated by SDS-PAGE and transferred to a nitrocellulose membrane (GE Healthcare Life Sciences, catalog no.10401396). The membrane was incubated in blocking buffer (PBS, 0.05% Tween 20, 5% dried milk) and then incubated with primary antibody diluted in blocking buffer. The membrane was washed three times in PBS supplemented with 0.05% Tween 20 and incubated with an HRP-conjugated secondary antibody. After three washes, the membrane was visualized with Western Lightning Plus-ECL substrate (PerkinElmer, catalog no. NEL10500). Blots were reprobed with another antibody after being stripped with restore stripping buffer (Thermo, 21059), if necessary. Protein bands were quantified by densitometry with ImageJ, if necessary.
For small interfering RNA (siRNA) screening, Huh7.5 cells were seeded into 24-well plates at a density of 1 × 105 cells/ml. After overnight growth, the cells were transfected with Silencer siRNAs targeting ANKFY1 (Ambion, s28200), SHROOM3 (Ambion, s33455), SEC16A (Ambion, s19236), VCP (Ambion, s14765), SEC24C (Ambion, s18515), MOV10 (Ambion, s8922), and UPF1 (Ambion, s11927) at a final concentration of 30 nM by RNAiMAX (Invitrogen) according to the manufacturer's protocol. At 24 h posttransfection, the cells were retransfected essentially as described above. After another 24 h, cell viability was determined with Cell Counting Kit-8 (Dojinodo Laboratories, CK04) according to the manufacturer's protocol. The cells were then washed once with DPBS and infected with Jc1G at a multiplicity of infection (MOI) of 0.1. At 2 days postinfection, the supernatants were collected and mixed with an equal volume of 2× passive lysis buffer (Promega) and luciferase activity was measured with Renilla luciferase substrate (Promega, E2820) according to the manufacturer's protocol. Cells from triplicate wells were washed with DPBS and lysed in TRIzol (Ambion) for RNA analysis.
For shRNA-mediated gene Kd, Huh7.5 or Huh7.5 cells harboring pTRIP-BSD were seeded into six-well plates at a density of 1 × 105/ml. After overnight growth, the cells were transduced with pLKo.1 pseudoparticles. When near confluence, the cells were split 1:3 with conditioned medium supplemented with 5 μg/ml puromycin. Two days later, cells were infected with Jc1G or transfected with Jc1G RNA. The supernatants were collected, and the cells were harvested at the time points indicated.
For CRISPR single guide RNA (sgRNA)-mediated gene Kd, similarly, Huh7.5 cells were first seeded into six-well plates and transduced with LentiCRISPR_v2 pseudoparticles. When near confluence, the cells were split 1:3 with conditioned medium supplemented with 5 μg/ml puromycin. Two days later, the cells were infected with Jc1G, the supernatants were collected, and the cells were harvested at the time points indicated.
Cells were fixed with 4% formaldehyde in PBS at room temperature for 15 min. After being washed with DPBS, the coverslips were immunostained as described previously (15) or directly mounted with Mowiol mounting medium (0.1 M Tris HCl [pH 8.5], 25% glycerol, 10% Mowiol 4-88 [Calbiochem 475904]) and observed with a Leica TCS SP8 confocal laser microscope with a 63× numerical aperture 1.3 oil immersion objective. Images were captured with the LAS software and processed with Image J. For colocalization assessment, Pearson's correlation coefficients were quantified for the selected regions with the ImageJ plugin as previously described (19).
Statistical analysis was performed with the GraphPad Prism 6 software. Specific tests are described in the figure legends.
Ammonium bicarbonate, dithiothreitol, iodoacetamide, and SDS were obtained from Sigma-Aldrich (St. Louis, MO). Acetonitrile (ACN, 99.9%), trifluoroacetic acid (TFA, 99.8%), and formic acid (FA) were purchased from Merck (Darmstadt, Germany). Sequencing grade modified trypsin was from Promega (Madison, WI, USA). The deionized water used for all experiments was obtained from a Milli-Q system (Millipore, Bedford, MA, USA). All other chemicals and reagents were of analytical grade and were obtained from the Shanghai Chemical Reagent Company.
The gels were subjected to silver staining as described previously (20) and then scanned with a GS-800 calibrated densitometer (Bio-Rad). Overall, 10 bands were selectively sliced from the silver-stained gels as indicated in the figures. Each band was then cut into slices, and in-gel digestion was performed as previously described (21). Gel pieces were washed twice with a buffer including 15 mM ferricyanide potassium and 50 mM sodium hyposulfite and then washed with water until the gel pieces were destained completely. The gel pieces were reduced with dithiothreitol and alkylated with iodoacetamide. Afterwards, gel pieces were dehydrated with 100% ACN and dried at 37°C for 10 min. The dried gel pieces were subjected to trypsin digestion by the addition of a solution containing 25 mM ammonium bicarbonate and 12.5 ng/μl trypsin and then incubated at 37°C overnight for digestion. After digestion, 0.1% TFA in 50% ACN was added to extract peptides from gels. The extracted peptides were dried completely in a SpeedVac.
Samples were resuspended with buffer A (5% ACN containing 0.1% FA) and separated on a reverse-phase analytical column (Eksigent, C18, 3 μm, 150 mm by 75 μm) by nano-high-performance liquid chromatography (Eksigent Technologies). Peptides were subsequently eluted under the following gradient conditions with phase B (98% ACN with 0.1% FA) from 5 to 45% B (5 to 100 min), and the total flow rate was maintained at 300 nl/min. An electrospray voltage of 2.5 kV versus the inlet of the mass spectrometer was used. The TripleTOF 4600 mass spectrometer used was operated in information-dependent data acquisition mode to switch automatically between the mass spectrometry (MS) and tandem MS (MS/MS) acquisition modes. MS spectra were acquired across a mass range of 350 to 1,250 m/z with a 250-ms accumulation time per spectrum. MS/MS spectra were scanned from 100 to 1,250 m/z in high-sensitivity mode with rolling collision energy. The 25 most intense precursors were selected for fragmentation per cycle with a dynamic exclusion time of 25 s.
MS database searching was performed as previously described, with some modifications (22). Peak lists were first generated from raw files with ProteoWizard v3.0 at default settings. Searching of a custom database that contains 20,264 human protein entries and 130,224 HCV protein entries was performed with Mascot v2.3.2 (Matrix Science, United Kingdom). These protein sequences were extracted from the UniProtKB/Swiss-Prot database (release 2014_03) by using the species names as keyword constraints with an in-house perl script. The search parameters were set as follows. The digestion enzyme selected was trypsin (fully enzymatic). A maximum of two missed cleavages was allowed. The mass value was set as monoisotopic, and peptide charge states of 2+ and 3+ were taken into account. Carbamidomethyl (C) was configured as a fixed modification, and oxidation (M) was configured as a variable modification. Mass tolerances for precursor and product ions were 25 ppm and 0.2 Da, respectively. The searching results from Mascot were parsed into Scaffold QS (Proteome Software, v4.3.4). All of the search results were combined automatically by Scaffold, and data evaluation was performed. The Scaffold local false-discovery rate algorithm assigned probabilities to Mascot identifications. Proteins were grouped by the legacy-independent sample protein grouping method. Protein identifications with a threshold of >99.0% probability assigned by the ProteinProphet algorithm (23), at least one unique peptide, and a peptide threshold of >95% probability assigned by the PeptideProphet algorithm (24, 25) were accepted.
To uncover the host factors that are intimately associated with HCV replication, we attempted to purify the viral replicase formed during authentic viral replication and identify the associated host factors. We aimed to affinity tag two replicase components with different tags to facilitate tandem affinity purification. Two viral replicase components, NS5A and NS4B, were successfully tagged. Tagging of NS5A does not affect viral replication, whereas tagging of NS4B attenuates viral replication (10, 14). In addition to NS5A, we aimed to tag a replicase component other than NS4B without obviously attenuating viral replication. We selected NS5B because of its constitutive role in replicase assembly rather than for its enzymatic activity (26, 27). It is tolerant to replacement with exogenous sequences and to insertion mutations (28, 29). We utilized the HCV replicon system where the self-replicating bicistronic viral RNA contains an antibiotic resistance gene and encodes the viral polypeptide encompassing NS3 to NS5B (30). In the first attempt with the BB7 subgenomic replicon (genotype con1b) (13), we selected a tolerant site for 1D4 peptide insertion with attenuated viral replication (data not shown).
We next sought to modify this tolerant site to increase viral replication. First, the HA-encoding sequence was inserted at the equivalent site in the backbone of the Jc1G genotype 2a HCV infectious clone containing a secreted Gaussia luciferase reporter (12) (Fig. 1A). The recombinant infectious clone, named Jc1G-NS5B.HA, could not replicate (Fig. 1B), possibly because the insertion may disrupt the RNA kissing loop between 5BSL3 and 3′SL2, which is essential for RNA replication (31, 32). Introduction of a duplicated NS5B stem-loop into an artificially introduced PacI site downstream of the stop codon of NS5B to restore the kissing loop (Fig. 1A) resulted in efficient replication of Jc1G-DU-NS5B.HA (Fig. 1B). It was noted that insertion of the duplicated region (Jc1G-DU) attenuated viral replication compared with that of the wild-type virus (Fig. 1B). Further insertion of HA into the NS5B region in Jc1G-DU enhanced viral replication (Fig. 1B). Although attenuated, Jc1G-DU-NS5B.HA replicated to a level similar to that of wild-type Jc1G at late time points (5 days posttransfection) (Fig. 1B). We then generated a subgenomic replicon construct of Jc1G-DU-NS5B.HA and subgenomic replicon constructs with HA insertion in NS5A (14) and NS4B (10) as reported, respectively. The resulting replicon, sgJFH1-DU-NS5B.HA1 (Fig. 1C), along with sgJFH1-NS5A.HA, replicated similarly to the nontagged sgJFH1 replicon, whereas sgJFH1-NS4B.HA (Fig. 1C), as reported, replicated to a lesser extent, as judged by the NS3 protein levels (Fig. 1D). The addition of amino acids to the ends of the HA peptide for epitope accessibility optimization (sgJFH1-DU-NS5B.HA2) also attenuated viral replication, as evidenced by the lower NS3 protein levels (Fig. 1D). There was a lower level of NS5B-HA than NS5A-HA, as judged by anti-HA antibody blotting (Fig. 1D), which is most likely due to the lower intrinsic stability of NS5B protein (33). We could not detect NS4B-HA in sgJFH1-NS4B.HA replicon cells (data not shown), most likely because of attenuated viral replication (10). Thus, we used sgJFH1-DU-NS5B-HA1 in the subsequent experiments. We then replaced the HA peptide in NS5B in this construct with a His6 peptide (Fig. 1C), and the resulting sgJFH1-NS5B.His replicon also replicated similarly to sgJFH1 (Fig. 1E). Insertion of the His peptide enhanced viral replication, as higher NS3 levels were detected (Fig. 1E), although this enhancement disappeared after passaging (data not shown). Finally, we introduced HA and His peptides into NS5A and NS5B simultaneously to obtain double-affinity-tagged replicon sgJFH1-NS5A.HA.NS5B.His. The expression of NS5A-HA and NS5B-His was confirmed via Western blotting with the anti-HA and anti-His antibodies, respectively (Fig. 1F).
We optimized the conditions for solubilizing the replicase from cells and meanwhile maintaining the interactions between the replicase components. After optimization, we tried to affinity capture NS5A-HA and NS5B-HA in sgJFH1-NS5A.HA and sgJFH1-NS5B.HA1 replicon cells and examined if another replicase component, NS3, could be captured. In both cell types, we detected NS3 but not the ER membrane protein calnexin in the captured protein complexes (Fig. 2A), indicating that the interactions between the viral replicase components were maintained and not due to a membrane-dependent nonspecific association. Detection of only a weak NS3 protein band in anti-NS5B-HA captured protein complexes was probably due to the low level of NS5B-HA protein (Fig. 2A and and1D).1D). We then used sgJFH1-NS5A.HA.NS5B.His replicon cells and performed two-step affinity purification. Solubilized cell lysates were sequentially purified by using the HA and His tags (Fig. 3A). At each step, NS3 could be detected in the captured complexes (Fig. 2B). In addition, after two-step affinity purification, NS3, NS5A, and NS5B were detected in the captured complexes (Fig. 2B). Thus, by affinity tagging HCV NS5A and NS5B in a replicon system, we could purify the solubilized viral replicase that contains at least NS3, NS5A, and NS5B.
We then scaled up the affinity purification experiment by using untagged sgJFH1 as a negative control. After two-step affinity purification, the eluted protein complexes were resolved by SDS-PAGE and the protein bands were visualized by silver staining (Fig. 3A). Overall, 10 sgJFH1-NS5A.HA.NS5B.His-specific or -enriched bands were sliced from the sgJFH1-NS5A.HA.NS5B.His lane and the proteins they contained were identified by MS (Table 1). The identified host proteins included proteins that were associated with intracellular transport, cytoskeleton components, RNP complex components, and chaperones. Among these proteins, numerous known HCV-interacting proteins like PI4KA, HnRNPK, PTB, HSP90, and nucleolin were unambiguously identified (Fig. 3B and Table 1).
We then selected host factors that have not been reported, on the basis of a PubMed search, to be involved in HCV replication and performed siRNA screening to determine the effect of each factor on viral replication. We selected COPII vesicle coat subunit Sec24C (34) and regulating protein Sec16A (35, 36); Shroom3, which is required for apical constriction and epithelial invagination (37) and is probably involved in cytoskeleton organization (38); ANKFY1 (also known as rabankyrin-5), which is a Rab5 effector (39); AAA+ATPase VCP, which is well known for its role in ER-associated degradation (ERAD) (40); RNA helicase MOV10, which is a putative RNA helicase and a component of the RNA-induced silencing complex (41); and UPF1, which is involved in the nonsense-mediated decay (NMD) pathway (42). Huh7.5 cells were transfected with siRNAs targeting the genes of interest (GOIs) and then infected with HCV Jc1G. The supernatants were collected to measure reporter luciferase activity (Fig. 3C). Intracellular viral RNA levels were examined by quantitative RT-PCR (Fig. 3D). The Kd efficiency of selected genes was confirmed by quantitative RT-PCR with gene-specific primers (Fig. 3E), and cell viability upon gene Kd was measured (Fig. 3F). Kd of Sec16A, Shroom3, ANKFY1, UPF1, and VCP significantly reduced viral replication, as determined by viral replication reporter and viral RNA quantification (Fig. 3C and andD).D). Among the seven selected genes, that for VCP was found among the top hits affecting viral replication (Fig. 3C and andDD).
We further investigated the role of VCP in HCV replication. To complement siRNA-mediated silencing, we performed shRNA-mediated Kd in Huh7.5 cells and infected the shRNA-transduced cells with Jc1G. Kd of VCP dramatically reduced HCV replication, as assessed via luciferase reporter activity, viral RNA quantification, and viral protein expression (Fig. 4A to toC).C). We also performed CRISPR (clustered regularly interspaced short palindromic repeats)-mediated gene Kd. All sgRNAs targeting VCP in Huh7.5 cells reduced VCP protein expression (Fig. 4F) and significantly reduced Jc1G viral replication, as assessed via the luciferase reporter activity (Fig. 4D), viral RNA level (Fig. 4E), and viral protein level (Fig. 4F).
To determine the viral life cycle in which VCP was involved, we overcame viral entry by transfecting viral RNAs into VCP-Kd cells. Incoming viral RNA translation was comparable in shIRR control cells and shVCP cells at 4 and 8 h posttransfection, suggesting that VCP was not involved in viral translation (Fig. 4G). In contrast, luciferase activity and viral protein abundance were dramatically reduced in shVCP cells at later time points (Fig. 4G and andH),H), suggesting that VCP was mainly involved in viral genome amplification.
We then examined the association between VCP and viral replicase. First, HA affinity purification of NS5A in sgJFH1-NS5A.HA.NS5B.His replicon cells showed that VCP was enriched in the pulled replicase complexes, similarly to the replicase component NS3, whereas the ER marker calnexin was not (Fig. 5A). We then examined the colocalization of RFP-VCP with the replicase marker NS5A in sgJFH1-NS5A.ypet replicon cells in which ypet is fused in frame to NS5A. We observed RFP-VCP (Pearson's colocalization coefficient [rp] = 0.54 and 0.62) but not RFP (rp = 0.003 and 0.113) colocalized with NS5A.yet clustering sites in these cells (Fig. 5B). We also examined the colocalization of endogenous VCP with NS5A in sgJFH1-NS5A.ypet replicon cells. Endogenous VCP exhibited cytoplasmic and nuclear distribution and colocalized with NS5A.yet in the cytoplasm (rp = 0.54 and 0.64) (Fig. 5C). Further, we used the HCV RC surrogate system in which the HCV polypeptide NS3-5B is expressed and RC assembly occurs independently of viral replication (5) to examine the association of VCP with the viral replicase marker. As shown in Fig. 5D, RFP-VCP changed its localization and relocalized to the NS5A.ypet clustering sites in the NS3-5B.ypet-expressing cells in which ypet is fused in frame to NS5A (rp = 0.67 and 0.82). Thus, VCP associates with the HCV RC and is required for viral replication.
VCP contains two ATP-binding (AAA) domains (Fig. 6A). Double mutation (E305Q/E578Q) of these two domains completely abolishes ATPase activity (43). To determine whether the ATPase activity of VCP is required for HCV replication, we performed a rescue experiment by expressing the shRNA-resistant version of VCP and dominant negative mutant VCP (E305Q/E578Q) in VCP-Kd cells. We first generated cell lines stably expressing shRNA-resistant, myc-tagged VCP and VCP-DN and then knocked down the endogenous VCP in these cells by using shRNA. Expression of exogenous myc-tagged VCP was confirmed by Western blotting with anti-myc antibody (Fig. 6C). Although exogenously expressed VCP did not completely restore VCP expression in VCP-Kd cells, most likely because of the high abundance of endogenous VCP, it partially restored HCV replication (Fig. 6B and andC).C). In contrast, VCP-DN did not rescue HCV replication but synergistically reduced viral replication in VCP-Kd cells (Fig. 6B), demonstrating that it acted as a dominant negative mutant protein. We then used a VCP inhibitor to block its ATPase activity and examined the effect on HCV replication. EerI inhibits VCP-dependent ERAD, most likely by irreversibly binding to VCP (44). EerI treatment at 2 μM dramatically reduced HCV replication and viral protein levels (Fig. 6D and andF)F) with only about 10% cell viability reduction (Fig. 6E). EerI treatment at 4 μM completely blocked viral replication (Fig. 6D and andF)F) with only about 15% cell viability reduction (Fig. 6E). Taken together, these data indicate that the ATPase activity of VCP is required for HCV replication.
We took advantage of the fact that HCV replication was dramatically reduced in VCP-DN-expressing cells in which endogenous VCP was further knocked down by shRNA. We attempted to examine if there was a change in NS5A distribution in the context of NS3-5B in these cells, which may indicate alteration of viral replicase assembly. We cotransfected NS3-5B.ypet with VCP-DN into the shRNA-transduced cells and monitored NS5A-ypet distribution by confocal microscopy. Compared with shIRR cells with NS5A-ypet, shVCP cells with NS5A-ypet exhibited a “coalesced” phenotype in which NS5A proteins were clustered and condensed (Fig. 7A). We also examined the NS5A-ypet distribution when cells were treated with EerI. We treated NS3-5B.ypet-expressing cells with 2 μM EerI for 24 h. At this dosage, EerI dramatically reduced viral replication (Fig. 6D). As in shVCP cells, EerI treatment altered the distribution of NS5A-ypet and resulted in a “coalesced” phenotype (Fig. 7B). Thus, abolishing VCP function resulted in aberrant HCV NS5A distribution in the replicase assembly surrogate system.
Almost all positive-strand RNA viruses use a replication strategy by which viral proteins modify the host membrane to assemble the membrane-associated viral replicase. Viruses hijack host factors to facilitate this energy-unfavorable process. Identification of the replicase-associated host factors and dissection of their roles in replicase assembly will help us understand the molecular mechanisms of RC biogenesis.
Purification of membrane-associated viral replicase remains technically challenging. By dual-affinity tagging two HCV replicase components, we tried to purify the solubilized HCV replicase. We detected the replicase components NS5A, NS5B, and NS3 in the purified complexes after sequential affinity purification (Fig. 2B and and3B).3B). We found specifically enriched host proteins associated with viral replicase, including numerous known HCV-interacting proteins (Fig. 3B and Table 1).
Of the genes selected for further study, Kd of those for Shroom3, ANKFY1, UPF1, and VCP significantly reduced viral replication (Fig. 3C and andD),D), suggesting potential roles for these proteins in viral replication. Shroom3, a cytoskeleton organization regulator (38), is required for apical constriction and epithelial invagination (37). HCV replication factories are protrusions from the ER membranes toward the cytosol (5) that may employ cytoskeleton reorganization via Shroom3. ANKFY1 (also known as rabankyrin-5), a Rab5 effector, is required for macropinocytosis (39). Rab proteins are required for HCV RC formation (9). Rab5 associated with HCV RC, and Kd of Rab5 reduced HCV replication (45, 46). As ANKFY1 was to be copurified with HCV replicase and Kd of ANFY1 significantly reduced viral replication (Fig. 3B to toD),D), Rab5-ANKFY1 pairs may play a role in RC formation. UPF1 is well known for its involvement in the NMD pathway (42) and was recently reported to restrict Semliki Forest virus infection. In contrast to restriction of viral replication, UPF1 is required for HIV1 virion production (47). In this study, we found that UPF1 is required for HCV replication (Fig. 3C and andDD).
Among the seven genes selected, that for VCP was the top hit affecting viral replication (Fig. 3C and andD)D) and is involved in HCV genome amplification (Fig. 4G). The ATPase activity of VCP was required for HCV replication (Fig. 6). VCP associated with viral replicase (Fig. 5A) and colocalized with viral RC components (Fig. 5B and andC).C). In addition, in an RC surrogate system, VCP relocalized to the RC factories (Fig. 5D). Further, when VCP function was abolished by overexpression of VCP-DN and knocking down by shRNA or EerI, HCV NS5A in the RC surrogate system exhibited a “coalesced” phenotype (Fig. 7), indicating aberrant replicase assembly. These data indicate that VCP is hijacked to the RC and associates with the viral replicase. It may participate in replicase assembly via its AAA+ATPase activity.
The AAA+ proteins play an important role in diverse cellular functions that are dependent on their ATPase activity to induce conformational changes in diverse substrate proteins (48). AAA+ATPase proteins have been reported in the positive-strand RNA viral life cycle. VPS4 is involved in brome mosaic virus and tomato bushy stunt virus RC formation by cooperating with ESCRT complexes (49, 50) and in HCV virion production (51). VCP was identified in two RNA interference screening experiments to be involved in poliovirus replication (52) and in the regulation of Sindbis virus entry (53). VCP colocalizes with poliovirus nonstructural proteins in virus-infected cells (52). It is likely that hijacking host AAA+ATPases to modulate RC assembly or the viral life cycle is conserved among positive-strand RNA viruses.
In conclusion, we developed an HCV subgenomic replicon system by dual-affinity tagging two replicase components. Using this dual-affinity-tagged replicon system, we purified the viral replicase and identified VCP as a pivotal replicase-associated host factor that is required for viral genome replication. We propose that HCV hijacks host AAA+ATPase for its replicase assembly. Dissection of the molecular mechanism by which VCP regulates viral replicase assembly may lead to new insights into viral RC biogenesis and novel antiviral strategies.
We are grateful to Yumei Wen for her warm encouragement, to Charles Rice for enlightening discussion and kindly providing the HCV reagents, and to Margaret R. MacDonald for her critical reading of the manuscript.
This work was supported in part by the National Basic Research Program of China (2015CB554301 to Z. Yi), Fudan University intramural funds (Zhuoxue fellowship to Z. Yi), the National Key Basic Research Program of China (2012CB519005 to Z. Yuan), and the Sino-German Transregional Collaborative Research and National Natural Science Foundation of China for International Collaboration (81461130019 to Z. Yuan).