|Home | About | Journals | Submit | Contact Us | Français|
Hepatitis C virus (HCV) nonstructural protein 5A (NS5A) regulates viral replication through its interaction with host and other viral proteins. We have previously shown that FK506-binding protein 8 (FKBP8) binds to NS5A and recruits Hsp90 to form a complex that participates in the replication of HCV. In this study, we examined the biochemical characteristics of the interaction and the intracellular localization of NS5A and FKBP8. Surface plasmon resonance analysis revealed that the dissociation constant of the interaction between the purified FKBP8 and NS5A expressed in bacteria was 82 nM. Mutational analyses of NS5A revealed that a single amino acid residue of Val or Ile at position 121, which is well conserved among all genotypes of HCV, is critical for the specific interaction with FKBP8. Substitution of the Val121 to Ala drastically impaired the replication of HCV replicon cells, and the drug-resistant replicon cells emerging after drug selection were shown to have reverted to the original arrangement by replacing Ala121 with Val. Examination of individual fields of the replicon cells by both fluorescence microscopy and electron microscopy (the correlative fluorescence microscopy-electron microscopy technique) revealed that FKBP8 is partially colocalized with NS5A in the cytoplasmic structure known as the membranous web. These results suggest that specific interaction of NS5A with FKBP8 in the cytoplasmic compartment plays a crucial role in the replication of HCV.
Hepatitis C virus (HCV) infects more than 170 million people worldwide, a large percentage of whom suffer from persistent infection and severe chronic liver diseases, culminating in cirrhosis and hepatocellular carcinoma (51). Combination therapy with pegylated interferon (IFN) and ribavirin achieves a 40 to 50% sustained virological response in patients infected with genotype 1 HCV (30). Recently, therapeutics have been developed to target the protease and polymerase of HCV, as well as the host factors required for the viral replication (24, 42).
HCV belongs to the Flaviviridae family and has a single-stranded positive-sense RNA genome with a nucleotide length of 9.6 kb. The viral genome, translation of which depends on its own internal ribosomal entry site found within the 5′ nontranslated region, encodes a large precursor protein composed of about 3,000 amino acids. The polyprotein is cleaved by host and viral proteases, resulting in viral structural proteins (core, E1, and E2), a putative viropore protein (p7), and nonstructural proteins (NS2, NS3, NS4A, NS4B, NS5A, and NS5B) (33). In the last decade, the mechanism by which HCV replicates in the hepatoma cell line Huh-7 has been partially revealed in studies using a cell culture system. The HCV replicon system, which encompasses the autonomously replicable genome of HCV in the Huh-7 cell line or other cell lines derived from it, has been established to accumulate information about the mechanism of HCV replication and to be utilized for screening antiviral drug candidates (27). In addition, the cell culture system for the propagation of infectious HCV particles was developed by using a full-length genome of HCV genotype 2a, JFH1 virus, which was isolated from a fulminant hepatitis C patient (25, 49, 57). However, a robust cell culture system for HCV of genotypes 1a and 1b, the most prevalent genotypes in the world, has not yet been successfully developed, with the exception of the cell culture systems for strains H77 and H77-S of the 1a genotype (21, 56). Furthermore, it is currently impossible to obtain a sufficient amount of HCV particles for biological and physiochemical studies due to the low viral load in the sera of hepatitis C patients and the low yield of HCV particles in the present cell culture system.
HCV NS5A is a membrane-anchored phosphoprotein that appears to possess multiple and diverse functions in viral replication, as well as in the establishment and maintenance of persistent infection (29, 38). Structural analyses suggest that NS5A forms a dimer and has a zinc-binding motif required for replication in the N-terminal domain (45, 46). NS5A has the IFN sensitivity-determining and MyD88-binding regions in the central domain (1, 10), and the SH3-binding region and nuclear localization signal in the C-terminal domain (28, 29). Adaptive mutations of NS5A have frequently been found in the replicon cells exhibiting efficient replication (4, 55). Several host proteins and lipids have been reported to interact with NS5A to upregulate the viral replication. For example, HCV replication was inhibited by treatment with lovastatin, an inhibitor of 3-hydroxy-3-methylglutaryl coenzyme A reductase, and this inhibition was restored by the addition of geranylgeraniol, suggesting that HCV replication requires geranylgeranylated proteins (22, 54). In addition, the F-box and leucine-rich repeat protein 2 (FBL2) was identified as a binding partner of NS5A, and geranylgeranylation of FBL2 was shown to be required for replication of HCV RNA (50). Vesicle-associated membrane protein (VAMP)-associated protein (VAP) subtype A (VAP-A) and subtype B (VAP-B) were also shown to interact with NS5A and NS5B through the coiled-coil domain and the N-terminal major sperm protein domain, respectively (11, 16, 39).
Immunophilins are known to share the peptidyl prolyl cis/trans isomerase activity, thereby basically conserving the ability to interact with immunosuppressive drugs such as cyclosporine and tacrolimus (FK506). Cyclophilin B, one of the cyclosporine-binding immunophilins, can bind to NS5B and upregulate the replication of HCV (53). We have previously reported that NS5A specifically interacts with FK506-binding protein 8 (FKBP8) and recruits heat shock protein 90 (Hsp90) to the viral RNA replication complex through the interaction of the carboxylate clump structure of FKBP8 with the C-terminal MEEVD motif of Hsp90 (37). Knockdown of FKBP8 reduced the replication efficiency of the HCV genome in the replicon cells and the cells infected with JFH1 virus (37), suggesting that FKBP8 is required for the replication of HCV via formation of the replication complex. In the present study we identified an amino acid residue in NS5A responsible for specific interaction with FKBP8 and examined the biochemical interaction and intracellular localization of NS5A and FKBP8.
Human embryo kidney 293T cells, and human hepatoma cell line Huh-7 and its derivatives were maintained in Dulbecco modified Eagle medium (DMEM; Sigma, St. Louis, MO) containing 10% fetal calf serum (FCS) and nonessential amino acid (NEAA). The Huh-7 9-13 cell line, which harbors an HCV subgenomic replicon (4, 27), was cultured in DMEM supplemented with 10% FCS and 1 mg of G418 and NEAA/ml. The Huh-7 9-13 cell line was treated with IFN-α to deplete the HCV RNA replicon. A cell line exhibiting the highest efficiency of propagation of JFH1 virus was selected by limited dilution and designated Huh-7OK1. The Huh-7OK1 cell line retained the ability to produce type I IFNs through the RIG-I-dependent signaling pathway upon infection with RNA viruses and exhibited a cell surface expression level of human CD81 comparable to that of the parental cell line. Detailed characteristics of this cell line are described elsewhere.
Rabbit antibody to NS5A was prepared by immunization with the NS5A peptide as described previously (16). Mouse monoclonal antibody to NS5A was purchased from Austral Biologicals (San Ramon, CA). Mouse monoclonal antibody to FKBP8 (KDM11) was described previously (37).
cDNA encoding NS5A was amplified from the HCV genotype 1b Con1 strain, kindly provided by R. Bartenschlager, by PCR using Pfu Turbo DNA polymerase (Stratagene, La Jolla, CA). The DNA fragment was cloned into pCAGGs-PUR/N-HA (36, 37). Human FKBP8 cDNA was amplified from the total cDNA of Huh-7 cells by PCR, and the fragment was introduced into pcDNA3.1 N-Flag, in which a Flag tag is introduced in the 5′ terminus of the cloning site of pcDNA3.1(+) (Invitrogen, Carlsbad, CA). The point mutations of NS5A were generated by the method of splicing by overlap extension (17, 18) and introduced into pCAGGs-PUR/N-HA. The mutant NS5A cDNAs were amplified by PCR, digested with MluI and XhoI, and introduced into the replicon plasmid pFKI389/neo/NS3-3′/5.1 (23), provided by R. Bartenschlager, or pFKI389/hRL/NS3-3′/5.1 (37). The cDNA encoding NS3 to NS5A was excised from pFKI389/neo/NS3-3′/5.1 and cloned into pCAGGs-PUR (36, 37). pET-UbCHis-del32-NS5A encoding an NS5A lacking the membrane-anchoring region (amino acid residues 1 to 32) and Escherichia coli strain BL21(DE3)/pCG1 was kindly provided by C. E. Cameron (19). The DNA fragment encoding the regions spanning from amino acid residues 2 to 389 of FKBP8 lacking the transmembrane region was amplified by PCR and replaced with the NS5A coding region of pET-UbNHis-del32-NS5A. The resulting plasmid encoding the amino acid residues from 2 to 389 of FKBP8 was designated pET-UbNHis-FKBP8(dTM) in this report. The DNA fragment encoding FKBP52 was amplified from the human fetal brain library (Clontech, Palo Alto, CA) by PCR and then was introduced into pET30a (Novagen, San Diego, CA) to be expressed in E. coli. The resulting plasmid was designated pET30a-FKBP52. The sequences of the plasmids were confirmed by using an ABI Prism 3130 genetic analyzer (Applied Biosystems, Tokyo, Japan).
The procedure used for protein purification was basically that of Huang et al. (19), with minor modifications that have been described previously (37). Briefly, overnight culture of E. coli strains transformed with pET-UbCHis-del32-NS5A, pET-UbNHis-FKBP8(dTM), or pET30a-FKBP52 were added at 1/100 volume into 250 ml of 2xYT medium and incubated at 37°C with shaking at 200 rpm. IPTG(isopropyl β-d-thiogalactoside) was added at a final concentration of 0.5 mM when the absorbance of the culture reached an optical density at 600 nm of 0.6 to 0.8, and then the culture solution was incubated at 20°C for 4 h with shaking at 200 rpm. After centrifugation of the culture at 3,000 × g for 5 min, the pellets were washed once with phosphate-buffered saline (PBS); suspended in 5 ml of 100 mM Tris-HCl (pH.8.0)-200 mM NaCl-10 mM 2-mercaptoethanol (lysis buffer) containing 0.5% Nonidet P-40, EDTA-free complete protease inhibitor (Roche, Indianapolis, IN), and 0.2 μg of lysozyme/ml; incubated at 4°C for 1 h; and subjected to freezing-thawing once. The resulting mixture was sonicated at 4°C for 5 min and was treated with 0.02 mg of DNase per ml at room temperature for 5 min. The suspension was centrifuged at 4°C at 30,000 rpm for 1 h in a Beckman SW50.1 (Beckman Coulter, Fullerton, CA), and the resulting supernatant was mixed with 0.5 ml of nickel agarose (Sigma) and gently rotated at 4°C for 60 min. The nickel resins were washed twice by spinning down with lysis buffer containing 10 mM imidazole. The recombinant protein was eluted from the nickel resin with lysis buffer containing 0.25 M imidazole and then dialyzed in 20 mM Tris-HCl (pH 8.0) containing 100 mM NaCl. The dialyzed eluates were applied to a Resource Q Sepharose column (GE Healthcare, Tokyo, Japan), washed with a ten-column volume of 20 mM Tris-HCl buffer (pH 8.0) containing 100 mM NaCl, and eluted under a linear gradient of 100 to 1,000 mM NaCl in 20 mM Tris-HCl buffer (pH 8.0). The peak fractions were pooled into a tube and concentrated by using Amicon Ultra-4 (Millipore, Bedford, MA). A half volume of the concentrated fraction was dialyzed against 10 mM HEPES (pH 7.4) containing 150 mM NaCl and 3 mM EDTA (HBS-E buffer) for analysis of the binding kinetics, while the remaining half was dialyzed in PBS for the immobilization on the sensor chip and pull-down assay. The protein concentration was measured using a Coomassie protein assay kit (Pierce, Rockford, IL).
Surface plasmon resonance (SPR) measurements were made at 25°C by using a Bicore 2000 biosensor (GE Healthcare) in accordance with the manufacturer's instructions to determine the affinity between NS5A and FKBP8. Briefly, The NS5A-His was immobilized as ligand on a carboxymethyl-dextran (CM5) sensor chip with an amine coupling kit (Biacore). His-FKBP8 and His-FKBP52 were diluted with HBS-E buffer containing 0.0005% surfactant P20 (HBS-EP buffer) at the concentrations indicated in Fig. Fig.1.1. The diluted sample was applied to the sensor chip at a flow rate of 20 μl/min in HBS-EP. The raw data were analyzed with a Biaevaluation software package (version 3.0; GE Healthcare).
Huh-7 9-13 replicon cells cultured on glass slides overnight were fixed with 4% paraformaldehyde in PBS at room temperature for 20 min. After two washes with PBS, cells were permeabilized for 15 min at room temperature with PBS containing 0.25% saponin and blocked with PBS containing 1% bovine serum albumin (PBS-BSA) for 30 min at room temperature. The cells were then incubated with PBS-BSA containing mouse anti-FKBP8 antibody (KDM11) and/or rabbit anti-NS5A antibody at 37°C for 60 min, washed three times with PBS-BSA, and incubated with PBS-BSA containing Alexa Fluor 488 (AF488)-conjugated anti-mouse immunoglobulin G (IgG), AF488-conjugated anti-rabbit IgG, AF594-conjugated anti-rabbit IgG, and/or AF546-conjugated anti-mouse IgG antibody (Molecular Probes, Eugene, OR) at 37°C for 60 min. Finally, the cells were washed three times with PBS-BSA and observed on a FluoView FV1000 laser scanning confocal microscope (Olympus, Tokyo, Japan). For mitochondria and lipid droplet staining, cells were incubated with culture medium containing Mitotracker Deep-Red (200 nM; Molecular Probes) and Bodipy 558/568 C12 (20 μg/ml; Molecular Probes), respectively, for 20 min at 37°C. After staining, cells were washed once with fresh and prewarmed culture medium and incubated at 37°C for 20 min.
Correlative fluorescence microscopy-electron microscopy (FM-EM) allows individual cells to be examined both in an overview with FM and in a detailed subcellular structure view with EM (40). For the observation by FM-EM, the Huh-7 9-13 replicon or Huh-7OK1 cells were cultured on gridded, 35-mm glass-bottom dishes (Mat Tek, Ashland, MA) in 1 ml of DMEM containing 10% FCS at 37°C overnight. Cells on the grid were fixed and stained with the specific antibodies as described above and then examined by using a confocal laser scanning microscope. The same specimens were then further incubated with 2.5% glutaraldehyde and 2% formaldehyde in PBS at 4°C overnight. After three washings with PBS, the samples were postfixed with 1% osmium tetroxide and 0.5% potassium ferrocyanide in PBS for 1 h, washed with distilled water three times, dehydrated in ethanol, and embedded in Epon812 (Structure Probe, West Chester, PA). Ultrathin sections of the cell (70-nm thick) were stained with saturated uranyl acetate and Reynolds lead citrate solution. The electron micrographs were taken with a JEOL JEM-1011 transmission electron microscope (JEOL, Ltd., Tokyo, Japan).
The transfection and immunoprecipitation tests were carried out as described previously (37). The immunoprecipitated samples were subjected to sodium dodecyl sulfate (SDS)-12.5 or 10% polyacrylamide gel electrophoresis. The proteins were transferred to polyvinylidene difluoride membranes (Millipore) and reacted with the appropriate antibodies. The immune complexes were visualized with Super Signal West Femto substrate (Pierce) and were detected by using an LAS-3000 image analyzer system (Fujifilm, Tokyo, Japan).
The HCV replicon plasmid, pFK-I389 hRL/NS3-3′/5.1 (37), was cleaved with ScaI and transcribed in vitro by using a MEGAscript T7 kit (Ambion, Austin, TX). Then, 10 μg of the transcribed RNA was electroporated at 270 V and 960 μF by a Gene Pulser (Bio-Rad, Hercules, CA) into 10 million cells of Huh-7OK1 of cell line per ml, suspended in 25 ml of culture medium, and then seeded at 1 ml per well on 12-well culture plates. Luciferase activity was measured at 4 and 48 h posttransfection using a Renilla luciferase assay system (Promega, Madison, WI) according to the manufacturer's protocol. The relative luciferase activity was presented as the ratio of the luciferase activity measured at 48 h posttransfection to that at 4 h.
The plasmid pFK-I389 neo/NS3-3′/NK5.1 (23) was digested with ScaI, and 10 μg of the in vitro-transcribed RNA was electroporated onto 4 million Huh-7 cells per 0.4 ml and suspended in 10 ml of the culture medium as described above. A 3-ml aliquot of the resulting cell suspension was mixed with 7 ml of the culture medium and inoculated into a culture dish 10 cm in diameter. The culture medium was replaced with fresh DMEM containing 10% FCS and 1 mg of G418 (Nakarai Tesque, Tokyo, Japan)/ml at 24 h posttransfection. The medium was exchanged once a week with fresh DMEM containing 10% FCS and 1 mg of G418/ml, and the remaining colonies were fixed with 4% paraformaldehyde at 28 days posttransfection and stained with crystal violet.
Total RNA was prepared from G418-resistant colonies by using an RNeasy minikit (Qiagen, Valencia, CA), and first-strand cDNA was synthesized with random primers by using a first-strand cDNA synthesis kit (GE Healthcare). The NS5A genes were amplified with the primer pair 5′-GACGGCATCATGCAAACCAC-3′ and 5′-CGTGGAGGTGGTATCGGAGG-3′. The PCR products were applied to agarose gel electrophoresis and purified by using a gel extraction kit (Qiagen). The purified PCR products were sequenced with the inside primer 5′-ATTAACGCGTACACCACGGG-3′ by using an ABI Prism 3130 genetic analyzer (Applied Biosystems).
We have previously reported that the thioredoxin-tagged domain I of NS5A (Trx-NS5A) binds directly to His6-tagged FKBP8 (37), although we could not obtain sufficient amounts of the recombinant FKBP8 for further biochemical analysis. Huang et al. reported that C-terminally His6-tagged NS5A lacking the N-terminal 32 amino acid residues of the membrane anchoring region (NS5A-His) could be purified by using a pET-ubiquitin expression system, in which the NS5A-His fused with ubiquitin at the C terminus was cleaved off by a ubiquitin-specific protease, Ubp1, in E. coli and then purified by using nickel-charged resin (19). By using the pET-ubiquitin expression system, we could obtain 1 mg of the purified His-FKBP8 protein from 1 liter of a culture of E. coli harboring a pET-UbCHis-FKBP8-dTM encoding an N-terminally His6-tagged FKBP8 lacking the transmembrane region (His-FKBP8), which is five times greater production than that achieved by the previous method we used (37). His-FKBP8, NS5A-His, and His-FKBP52 (10 μg) were purified with nickel-charged resin (Fig. (Fig.1A)1A) and subjected to the pull-down assay. Immunoblotting, instead of protein staining, was used for detection of the precipitates due to the similar molecular sizes of NS5A-His and His-FKBP8 (Fig. (Fig.1A).1A). To confirm the specificity of the antibodies to NS5A and FKBP8, NS5A-His and His-FKBP8 were immunoprecipitated and then subjected to immunoblotting by the antibodies. The antibodies to NS5A and FKBP8 specifically recognize NS5A and FKBP8, respectively (Fig. (Fig.1B).1B). The antibody to NS5A, but not nonspecific rabbit IgG, precipitated NS5A together with FKBP8 (Fig. (Fig.1C).1C). The reverse experiment was also successful in demonstrating that antibody to FKBP8, but not nonspecific mouse IgG, precipitated FKBP8 with NS5A (Fig. (Fig.1D).1D). The binding kinetics was analyzed on the basis of the SPR technology to examine the specificity of interaction between FKBP8 and NS5A. His-FKBP8 or His-FKBP52 was applied to a flow line at various concentrations on the sensor chip on which NS5A-His was immobilized. Each background signal was determined by flowing the FKBPs over a blank chip. The SPR signal of His-FKBP8 or His-FKBP52 was determined after subtraction by the background signals. The SPR was increased corresponding to the amount of His-FKBP8, but no response was observed with His-FKBP52 (Fig. (Fig.1D).1D). The values of the dissociation constant, Kd (10−3 s−1), and Ka (103 M−1 s−1) were calculated to be 1.86 and 22.8, respectively. Therefore, the equilibrium dissociation constant (KD) of the interaction between His-FKBP8 and NS5A-His was determined as 82 nM, suggesting a specific binding of the proteins. These results indicate that FKBP8 directly and specifically interacts with NS5A.
The domain I of NS5A (amino acid residues 1 to 213) was shown to interact with FKBP8 (37). However, further analyses on the specific interaction of NS5A with FKBP8 have not yet been carried out. To determine the amino acid residues in NS5A responsible for specific interaction with FKBP8, Flag-FKBP8 was coexpressed with C-terminal deletion mutants of the hemagglutinin (HA)-tagged NS5A domain I in 293T cells and immunoprecipitated with appropriate antibodies (Fig. (Fig.2A).2A). Although the C-terminal deletions up to the residue 141 in HA-NS5A exhibited no effect on the coimmunoprecipitation with Flag-FKBP8, further deletion beyond the amino acid residue 121 of HA-NS5A abrogated the coprecipitation with Flag-FKBP8 (Fig. (Fig.2B,2B, upper panel), suggesting that residues from 121 to 140 in NS5A are responsible for the interaction with FKBP8. Further deletion mutants of HA-NS5A revealed that the amino acid residues from 121 to 125 are required for the interaction with Flag-FKBP8 (Fig. (Fig.2B,2B, lower panel). To identify a specific amino acid residue critical for interaction with FKBP8, we generated substitution mutants of HA-NS5A(1-125) in which each of the amino acid residues from 121 to 125 were replaced with Ala. The mutant in which Val121 was replaced with Ala completely abrogated the interaction of HA-NS5A(1-125) with Flag-FKBP8, but the other substitution mutants did not (Fig. (Fig.2C).2C). However, we could not obtain a clear reduction in the interaction of FKBP8 with a full-length of NS5A mutant substituted Val121 with Ala by immunoprecipitation analysis (data not shown). To examine the interaction of NS5A with FKBP8 in more functional setting, we examined the colocalization of the wild-type or mutant NS5A with an endogenous FKBP8 in Huh-7OK1 cells by transfection of the expression plasmids encoding HCV nonstructural proteins carrying a wild-type or mutant NS5A substituted Val121 with Ala. As shown in Fig. Fig.3,3, colocalization of an endogenous FKBP8 with NS5A was reduced by the introduction of substitution of Val121 to Ala. These results suggest that Val121 of NS5A plays a critical role in the specific interaction with FKBP8.
The amino acid alignment of NS5A derived from several genotypes on the basis of the European HCV database (http://euhcvdb.ibcp.fr/euHCVdb/jsp/index.jsp) revealed that the amino acid residue Val121 is well conserved among NS5A of various genotypes of HCV, with the exception of the genotype 1a strains, which have Ile in place of Val (Fig. (Fig.4A).4A). We have previously shown that NS5A of genotype 1a (H77C strain), which has an amino acid residue Ile121, was able to interact with FKBP8 (37). To determine the role of Ile121 on the binding of NS5A to FKBP8, HA-NS5A(1-125) of the genotype 1b Con1 strain in which Ile was substituted for Val121 was coexpressed with Flag-FKBP8 and immunoprecipitated with specific antibodies (Fig. (Fig.4B).4B). The HA-NS5A mutant possessing the substitution of Val121 to Ile interacted with Flag-FKBP8 at the same level as the wild-type NS5A. Next, to determine the role of Val121 or Ile121 in the replication of HCV, we generated replicon RNAs in which Val121 of NS5A was replaced with either Ala or Ile. In vitro-transcribed RNAs from the pFK-I389 hRL/NS3-3′/5.1 carrying the mutation were introduced into the Huh-7 cell line by electroporation. The Renilla luciferase activity was measured at 4 and 48 h posttransfection and is represented as the ratio of luciferase activity measured at 48 h posttransfection to that measured at 4 h. The replacement of Val121 with Ala severely impaired the RNA replication, whereas substitution of Val121 to Ile, as seen in genotype 1a strains, had no apparent effect on the replication (Fig. (Fig.4C).4C). These results suggest that Val121 and Ile121 of NS5A play crucial roles in the interaction with FKBP8 and the transient replication of HCV replicons. We have previously reported that NS5A interacts with an endogenous FKBP8 in replicon cells harboring the subgenomic viral RNA (37). Therefore, we tried to demonstrate the lack of interaction of FKBP8 with the mutant NS5A substituted Val121 to Ala. However, we could not detect a sufficient amount of HCV proteins due to a low level of replication of the subgenomic replicon carrying the mutation in NS5A (Fig. (Fig.4C4C and and5A5A).
To further confirm the importance of Val121 and Ile121 in NS5A on the replication of HCV RNA, a colony formation assay was carried out. The replicon RNA carrying a neomycin resistance gene transcribed from pFKI389/neo/NS3-3′/5.1 (23) was introduced into Huh-7 cells and cultivated under the pressure of G418. The number of remaining cell colonies was determined at 4 weeks posttransfection. There were more than 1,000 colonies on the plate of Huh-7 cells into which the parent replicon RNA or an RNA carrying the substitution of Val121 of NS5A to Ile was introduced, but only a few colonies appeared on the plate of cells into which RNA carrying the mutation of Val121 to Ala was introduced (Fig. (Fig.5A).5A). To characterize the colonies emerging on the plate of Huh-7 cells into which the mutant Ala121 replicon RNA was introduced, the total RNAs were purified from the seven resistant colonies. The NS5A cDNAs were amplified after reverse transcription but not in the absence of reverse transcription (Fig. (Fig.5B),5B), suggesting that the amplified cDNAs were derived from RNA but not from the remaining transfected plasmid DNA. The NS5A genes were subjected to direct sequencing and revealed that the transfected mutant replicon RNA had GCG corresponding to the Ala121, in contrast to the parental replicon, which had GTT corresponding to the Val121. On the other hand, all of the RNAs prepared from the individual resistant colonies had GTG encoding Val (Fig. (Fig.5C),5C), indicating that the resistant colonies were not derived from the contamination of the wild-type replicon RNA but emerged by the mutation after replication. These results further support the notion that Val121 in NS5A is an indispensable amino acid and plays an important role in the replication of HCV though interaction with FKBP8.
Previous reports suggest that FKBP8 is mainly localized on mitochondria (7, 44), whereas NS5A is mainly localized on the endoplasmic reticulum (ER) and Golgi apparatus (2, 6, 16). HCV is reported to replicate in a raft-like intracellular compartment or the folded membranous compartment known as a membranous web in the replicon cells (8, 13, 15). In the present work, intracellular localization of FKBP8 was examined by immunofluorescence staining of the replicon cell line, Huh-7 9-13, which harbored an HCV subgenomic replicon, with the antibodies to NS5A and to FKBP8. Endogenous FKBP8 was mainly found in mitochondria and was partially colocalized with NS5A in a few compartments sharing a dot-like structure (Fig. (Fig.6A).6A). Lipid droplets were required for production of infectious HCV (5) and were colocalized with NS5A and core protein (43), although NS5A formed as dot-like structures but was not found in lipid droplets stained with Bodipy 558/568 C12 in the replicon cell line (Fig. (Fig.6B).6B). On the other hand, FKBP8 was mainly localized on mitochondria and partially together with NS5A on dot-like structures that were distinct from the mitochondria (Fig. (Fig.6C6C).
To further analyze the subcellular compartments where FKBP8 and NS5A were colocalized, the same fields of Huh-7 9-13 replicon cells were observed with FM and EM by using the correlative FM-EM technique described above. This method allowed us to examine the colocalization of the molecules by both FM and EM in the same samples, yielding two different but complementary data sets. The replicon cells were stained with antibodies to FKBP8 and NS5A and examined under FM (Fig. (Fig.7A,7A, left panels), and the same fields were observed under EM (Fig. (Fig.7A,7A, right panels). The compartments colocalizing FKBP8 and NS5A (arrows) exhibited a high electron density and a folded membranous structure that was similar to a membranous web (15, 32). In contrast, the replicon cells cured by IFN-α treatment did not have the electron-dense structure (Fig. (Fig.7B).7B). These results suggest that FKBP8 interacts with NS5A on the membranous web in cells replicating HCV RNA.
HCV NS5A is a multifunctional protein involved in viral replication and pathogenesis (29). In a previous study, we have shown that NS5A specifically interacts with FKBP8 and recruits Hsp90 to the viral RNA replication complex through the interaction of the carboxylate clump structure of FKBP8 with the C-terminal MEEVD motif of Hsp90 (37). Although we demonstrated that a TPR domain other than the carboxylate clump region of FKBP8 was responsible for the specific interaction with NS5A (37), the precise binding amino acid residue of the interaction was not determined. In the present study, FKBP8 exhibited a specific interaction with the immobilized NS5A in a dose-dependent manner with an equilibrium dissociation constant (Kd) of 82 nM as determined by the SPR, but no interaction with FKBP52 was detected. Furthermore, mutational analysis suggested that Val or Ile at the amino acid residue 121 of NS5A was responsible for the specific interaction with FKBP8. The subgenomic HCV replicon RNA harboring the mutation of Val121 to Ala within NS5A leads to severe impairment of RNA replication, and reversion from Ala121 to Val was detected, suggesting that interaction of FKBP8 with NS5A through the Val121 is crucial for HCV replication. The crystal structure of NS5A domain 1 revealed that Val121 is located on one of the β-sheet structures in the 1B subdomain and the side chain of the residue is located within the hydrophobic core (46); therefore, the Val121 may be involved in the maintenance of the β-sheet structure in the subdomain rather than the direct interaction with FKBP8. However, it remains feasible to speculate that unidentified host factors may be involved in the conformational change of region, including Val121 for direct interaction with FKBP8. Further studies, including a structural analysis of FKBP8, are needed to clarify the mechanisms by which HCV is replicated through the interaction of NS5A, FKBP8, and Hsp90.
The current combination therapy with pegylated IFN-α and ribavirin achieves a sustained virological response in half of the patients infected with a high viral load of HCV of genotype 1b (30). However, it is difficult to achieve the complete removal of viruses by antiviral drugs targeted to the viral enzymes, including proteases and polymerases, from patients persistently infected with RNA viruses that exhibit a quasispecies nature, such as human immunodeficiency virus and HCV. Viral quasispecies are not a simple collection of diverse mutants but a group of interactive variants capable of adapting to new environments (48). Cyclosporine treatment has been shown to be effective for patients infected with HCV of genotype 1b (20) and suppresses HCV RNA replication in vitro (52). In addition, cyclosporine has been shown to disrupt the interaction between NS5B and cyclophilin B, which is required for an efficient RNA-binding of NS5B (53). Cyclophilins and FKBPs are classified as immunophilins capable of binding to the immunosuppressants cyclosporine and FK506, respectively (26). The family members do not share a homologous domain other than drug-binding and enzymatically active domains, based on their amino acid sequences, substrate specificities, and inhibitor sensitivities. However, cyclosporine-resistant RNA replicon was shown to exhibit mutations not only in NS5B but also in NS5A (12, 41), suggesting that cyclosporine might affect the viral replication through the nucleotide-binding ability of NS5B, as well as the function of NS5A. Recently, geldanamycin, an inhibitor of Hsp90, was shown to drastically impair the replication of poliovirus without any emergence of escape mutants (14). Therefore, the elucidation of host proteins, including immunophilins and chaperones, participating in the HCV replication complex may lead to the development of new therapeutics for chronic hepatitis C with a broad spectrum and a low possibility of emergence of revertant viruses. In particular, disruption of the specific interaction of Val121 of NS5A with the TPR domain of FKBP8 might be an ideal target for a novel therapeutic measure.
Egger et al. reported that NS4B alters the intracellular membrane to form a membranous web structure consisting of a membrane-associated multiprotein complex localized in the cytoplasmic compartments distinct from the mitochondria in vitro and in the liver of an HCV-infected chimpanzee, suggesting that the membranous web forms the viral replication complex (8). An N-terminal amphipathic helix of NS4B plays an important role in the viral replication, as well as in the correct localization of other NS proteins including NS5A (9). Furthermore, VAP-B was reported to interact with Nir2 protein through the FFAT (named for two phenylalanines [i.e., FF] in the acidic tract) motif and to remodel the ER structure to form a convoluted membrane structure resembling a membranous web (3). In addition, VAP-A and B interact with not only NS5A but also NS5B (13, 16, 47), suggesting that the complex of NS5A with FKBP8 might be recruited on the membranous web by NS4B and/or VAPs and participate in the HCV replication.
FKBP8 has been shown to be localized mainly on the mitochondria and to interact with Bcl-2 to sequester Bcl-2 on the mitochondria (7, 44). However, HCV RNA was suggested to be replicated in the membranous web structure in replicon cells (8, 13, 15), and NS5A was reported to localize on the ER, Golgi apparatus (2, 6, 16), and lipid droplets (43). Figures Figures6C6C and and7A7A clearly indicate that the intracellular compartment including NS5A and FKBP8 is distinct from mitochondria. The HCV core protein was shown to upregulate genes related to fatty acid biosynthesis through the interaction with proteasome activator PA28γ/REGγ in the nucleus (34) and to induce accumulation of cytoplasmic lipid droplets in the mouse liver (35). Recently, it was shown that the HCV core protein of the genotype 2a JFH1 strain recruits the replication complex to the lipid droplet-associated membranes, and HCV particles were detected in close proximity to the lipid droplets, suggesting that lipid droplets induced by core protein participate in the assembly of HCV particles (31). In addition, the lipid droplets including the core protein were surrounded by the nonstructural proteins was also detected in cells expressing the chimeric HCV genomes encoding core to a part of NS2 proteins of genotype 1b or 1a strain and the nonstructural proteins of JFH1 strain (31). In the present study, FKBP8 was shown to be colocalized with NS5A in a highly electron-dense intracellular compartment indistinguishable from the membranous web. Although the total amount of FKBP8 was not changed by the treatment of the replicon cells by IFN-α (data not shown), the membranous web structure where FKBP8 and NS5A had accumulated was removed by the treatment (Fig. (Fig.7B).7B). These results suggest that the replication of the subgenomic HCV RNA induces the formation of a membranous web structure in which NS5A and FKBP8 are colocalized but has no effect on the expression level of FKBP8. Furthermore, we could not detect any colocalization of FKBP8 and NS5A with the lipid droplets in the replicon cells harboring a full-length genome of the genotype 1b Con1 strain (data not shown). Although the relationships between the membranous web and lipid droplets remain unknown, these discrepancies might be attributable to the difference in HCV genotypes of the nonstructural proteins that consist of the major components of the replication complex determining the efficiency of HCV replication.
In conclusion, our data indicate that NS5A directly binds to FKBP8 through the Val121 and colocalizes in the convoluted membrane structure known as the membranous web. Future studies on the role of FKBP8 in the replication of HCV might contribute to the development of a new type of anti-HCV drugs with a low frequency of emergence of drug-resistant breakthrough viruses.
We thank H. Murase for secretarial work. We also thank R. Bartenschlager and C. E. Cameron for providing the plasmids. We are grateful to the staff of the Center for Research and Education for their help in using the Biacore 2000.
This work was supported in part by grants-in-aid from the Ministry of Health, Labor, and Welfare; the Ministry of Education, Culture, Sports, Science, and Technology; the 21st Century Center of Excellence Program; and the Foundation for Biomedical Research and Innovation.
Published ahead of print on 23 January 2008.