ISGylation of Hepatitis C Virus NS5A Protein Promotes Viral RNA Replication via Recruitment of Cyclophilin A

Host cells have evolved host defense machinery (such as innate immunity) to eliminate viral infections. Viruses have evolved several counteracting strategies for achieving an immune escape from host defense machinery, including type I interferons (IFNs) and inflammatory cytokines. ISG15 is an IFN-inducible ubiquitin-like protein that is covalently conjugated to the viral protein via specific Lys residues and suppresses viral functions and viral propagation. Here, we demonstrate that HCV NS5A protein accepts ISG15 conjugation at specific Lys residues and that the HERC5 E3 ligase specifically promotes NS5A ISGylation. We obtained evidence suggesting that NS5A ISGylation facilitates the recruitment of CypA, which is the critical host factor for HCV replication, thereby promoting HCV replication. These findings indicate that E3 ligase HERC5 is a potential therapeutic target for HCV infection. We propose that HCV hijacks an intracellular ISG15 function to escape the host defense machinery in order to establish a persistent infection.

ABSTRACT

Interferon-stimulated gene 15 (ISG15) is a ubiquitin-like protein that is covalently conjugated to many substrate proteins in order to modulate their functions; this conjugation is called ISGylation. Several groups reported that the ISGylation of hepatitis C virus (HCV) NS5A protein affects HCV replication. However, the ISG15 conjugation sites on NS5A are not well determined, and it is unclear whether the role of NS5A ISGylation in HCV replication is proviral or antiviral. Here, we investigated the role of NS5A ISGylation in HCV replication by using HCV RNA replicons that encode a mutation at each lysine (Lys) residue of the NS5A protein. Immunoblot analyses revealed that 5 Lys residues (K44, K68, K166, K215, and K308) of the 14 Lys residues within NS5A (genotype 1b, Con1) have the potential to accept ISGylation. We tested the NS5A ISGylation among different HCV genotypes and observed that the NS5A proteins of all of the HCV genotypes accept ISGylation at multiple Lys residues. Using an HCV luciferase reporter replicon assay revealed that residue K308 of NS5A is important for HCV (1b, Con1) RNA replication. We observed that K308, one of the Lys residues for NS5A ISGylation, is located within the binding region of cyclophilin A (CypA), which is the critical host factor for HCV replication. We obtained evidence derived from all of the HCV genotypes suggesting that NS5A ISGylation enhances the interaction between NS5A and CypA. Taken together, these results suggest that NS5A ISGylation functions as a proviral factor and promotes HCV replication via the recruitment of CypA.

IMPORTANCE Host cells have evolved host defense machinery (such as innate immunity) to eliminate viral infections. Viruses have evolved several counteracting strategies for achieving an immune escape from host defense machinery, including type I interferons (IFNs) and inflammatory cytokines. ISG15 is an IFN-inducible ubiquitin-like protein that is covalently conjugated to the viral protein via specific Lys residues and suppresses viral functions and viral propagation. Here, we demonstrate that HCV NS5A protein accepts ISG15 conjugation at specific Lys residues and that the HERC5 E3 ligase specifically promotes NS5A ISGylation. We obtained evidence suggesting that NS5A ISGylation facilitates the recruitment of CypA, which is the critical host factor for HCV replication, thereby promoting HCV replication. These findings indicate that E3 ligase HERC5 is a potential therapeutic target for HCV infection. We propose that HCV hijacks an intracellular ISG15 function to escape the host defense machinery in order to establish a persistent infection.

INTRODUCTION

Hepatitis C virus (HCV) infection is a leading cause of chronic hepatitis, liver cirrhosis, and hepatocellular carcinoma and remains a public health burden worldwide. HCV is an enveloped, single-stranded positive-sense RNA virus classified in the Hepacivirus genus of the Flaviviridae family. The HCV genome consists of 9.6 kb of RNA encoding a single polyprotein that is processed by viral proteases and cellular signalases to produce three structural proteins (core, envelope 1 [E1], and E2) and seven nonstructural proteins (p7, nonstructural protein 2 [NS2], NS3, NS4A, NS4B, NS5A, and NS5B) (1). Replication of the viral RNA occurs on endoplasmic reticulum (ER)-derived double membrane vesicles and requires the participation of NS2 to NS5B, in addition to cellular host factors (2–4).

NS5A is a membrane-associated RNA-binding protein involved in HCV RNA replication and viral assembly through the recruitment of various host factors (e.g., VAP-A/B, cyclophilin A [CypA], phosphatidylinositol 4-kinase type IIIα [PI4KIIIα], and others) (5). The N-terminal membrane anchor domain of NS5A forms an amphipathic α-helix (AH). NS5A has been divided into three distinct domains (I, II, and III) connected by two solvent-exposed, low-complexity sequences (LCS-I and LCS-II). Domain I (DI) has been crystallized only in the dimer formation (6–8), and the structures of domains II (DII) and III (DIII) have not been established. NS5A is a multifunctional protein, and its precise functions and modes of action during the HCV life cycle are not fully understood.

Interferon (IFN)-stimulated gene 15 (ISG15) is a ubiquitin-like protein that is induced by IFN production after viral or bacterial infection (9). ISG15 is covalently conjugated to the target substrate protein via specific lysine (Lys) residues by three enzymes: an E1 activating enzyme (UBE1L), an E2 conjugating enzyme (UbcH8), and E3 ligase (HERC5). This process is called “ISGylation” and is a posttranslational protein modification that is similar to ubiquitylation. The covalently conjugated ISG15 is removed from target substrates by USP18 (also referred to as UBP43), an ISG15-deconjugating enzyme. USP18 cleaves isopeptide bonds between ISG15 and the substrate protein (10). In addition to its involvement in the regulation of signal transduction and tumorigenesis, ISG15 plays important roles in several virus infections (9).

Using ISG15-deficient mice, it was demonstrated that ISG15 may function as an antiviral factor against influenza A and B viruses, herpes simplex virus 1, murine gammaherpesvirus 68, Sindbis virus (11), and chikungunya virus (12). However, no effects of ISG15 on vesicular stomatitis virus and lymphocytic choriomeningitis virus infection were observed (13). Investigations of gain or loss of function of ISG15 in cultured cells revealed that ISG15 has an antiviral function against Japanese encephalitis virus (14), Dengue virus, and West Nile virus (15), human immunodeficiency virus type 1 (16–18), Ebola virus (19), vaccinia virus lacking its coding for E3 protein (VV-delta E3) (20), and murine norovirus (21). The function of ISG15 has thus been suggested to be an antiviral function to several viruses. However, there were conflicting observations suggesting that ISG15 has a proviral function in HCV infection (22–25).

IFN-induced ISG15 also exists in the free, unconjugated form and functions intra- or extracellularly (26). Evidence accumulated that the role of free ISG15 may be to modulate viral infections without ISGylation (17, 19, 27, 28). ISG15-defective fibroblast cells derived from a patient with inherited ISG15 deficiency showed enhanced resistance against several viral infections due to increased IFN signaling activation (29). Zhang and colleagues mechanistically demonstrated that free ISG15 negatively regulates the IFN response via USP18 stabilization (30). These results indicate that there are distinct roles of free ISG15 in humans and rodents. Most recently, we demonstrated that HCV NS5A protein can interact with free ISG15 without ISGylation (31). However, the role of NS5A ISGylation and its impact on HCV infection remain to be elucidated.

In the present study, we sought to elucidate the role of NS5A ISGylation in HCV RNA replication. We determined the ISG15 conjugation sites on the specific Lys residues of NS5A protein and examined the effects of mutations on viral replication. We obtained evidence suggesting that NS5A ISGylation is involved in the positive regulation of HCV RNA replication through an enhanced association of NS5A with cyclophilin A (CypA), which is the critical host factor for HCV RNA replication (32–35).

RESULTS

HCV NS5A is a target substrate of ISGylation.

First, to determine whether NS5A (1b, Con1) is a target substrate for ISGylation, we coexpressed hemagglutinin (HA)-NS5A and FLAG-ISG15 together with E1 (UBE1L), E2 (UbcH8), and E3 (HERC5) in 293T cells, followed by immunoprecipitation with anti-FLAG antibody and detection with anti-HA antibody (Fig. 1A). In the immunoblot analysis with anti-FLAG or anti-ISG15 antibody, FLAG-ISG15 and ISG15 conjugates were detected only in the E1/E2/E3-transduced cells (Fig. 1A to D and F, lanes 2). The immunoprecipitation analysis coupled with immunoblotting revealed that HA-NS5A was precipitated with FLAG-ISG15, and slowly migrating forms of HA-NS5A were clearly detected in the cotransduced cells compared to the cells transduced with HA-NS5A alone (Fig. 1A, left and middle, lanes 2, asterisks).

FIG 1.

HCV NS5A is a target substrate of ISGylation. (A to D) Expression vector encoding HA-NS5A (1b, Con1) was coexpressed with FLAG-ISG15 together with E1 (UBE1L), E2 (UbcH8), and E3 ligase (HERC5) in 293T cells, followed by immunoprecipitation with anti-FLAG mouse MAb or anti-HA rabbit PAb and detection with anti-HA rabbit PAb (A), anti-FLAG mouse MAb (B), and specific antibodies against NS5A (C) and ISG15 (D), respectively. (E) Schematic diagrams of ISG15 and ISG15 mutant ISG15-AA. ISG15 was mutated by replacing the amino acid glycine (G) with alanine (A) within the C-terminal LRLRGG motif sequence. UBL1, ubiquitin-like domain 1; LRLRGG, Leu-Arg-Leu-Arg-Gly-Gly. (F) Expression vector encoding HA-NS5A was coexpressed with FLAG-ISG15 or FLAG-ISG15-AA together with E1, E2, and E3 ligase in 293T cells, followed by immunoprecipitation with anti-HA rabbit PAb and detection with anti-ISG15 mouse MAb or anti-HA rabbit PAb. Input samples (Lysate) were detected by anti-HA rabbit PAb, anti-FLAG mouse MAb, anti-NS5A mouse MAb, or anti-ISG15 mouse MAb. The asterisks indicate the various ISG15-conjugated NS5A proteins. IP, immunoprecipitation; IB, immunoblotting; HC, immunoglobulin heavy chain.

These slowly migrating forms of HA-NS5A were also observed in the cotransduced cells when anti-HA antibody was used for immunoprecipitation, followed by detection with anti-FLAG antibody (Fig. 1B, left and middle, lanes 2, asterisks). To determine whether these bands represent the NS5A ISGylation, we performed an immunoprecipitation analysis of samples with anti-HA antibody and detection with anti-NS5A or anti-ISG15 antibody, respectively. As shown by the results in Fig. 1C and D, these bands were also detected with their specific antibodies, indicating that the bands represent the ISG15-conjugated NS5A protein.

Next, to verify the specificity of the ISGylation of NS5A protein, we performed an immunoprecipitation analysis using the conjugation-defective ISG15 mutant that possesses the Gly-Gly to Ala-Ala substitution within the C-terminal LRLRGG motif sequence (Fig. 1E). The expression levels of FLAG-ISG15 and FLAG-ISG15-AA (approximately 15 kDa) were comparable (Fig. 1F, fourth panel, lanes 2 and 3). The transduction of FLAG-ISG15, but not of FLAG-ISG15-AA, resulted in NS5A ISGylation in both immunoprecipitates and lysates (Fig. 1F, first, second, and third panels, compare lanes 2 with lanes 3).

To seek evidence suggesting that NS5A protein is ISGylated in the HCV-replicating cells, we electroporated the in vitro-transcribed chimeric JFH1/5A-Con1 RNA that encodes the NS5A gene from the Con1 strain together with ISGylation components into the Huh7.5 cells. The immunoprecipitation analysis coupled with immunoblotting using NS5A-specific antibody indicated that the NS5A ISGylation with 100-kDa conjugates was clearly detected in the HCV-replicating cells (Fig. 2, lane 2, asterisk). These results suggest that NS5A protein is a target substrate of ISGylation in the HCV-replicating cells.

FIG 2.

Detection of NS5A ISGylation in the HCV-replicating cells. Huh-7.5 cells were electroporated with in vitro-transcribed chimeric JFH1/5A-Con1 RNA together with E1 (UBE1L), E2 (UbcH8), and E3 ligase (FLAG-tagged HERC5), followed by immunoprecipitation analysis coupled with immunoblotting using NS5A-specific antibody. Input samples (Lysate) were detected with anti-NS5A mouse MAb, anti-FLAG mouse MAb, or anti-ISG15 mouse MAb. The asterisk indicates the 100-kDa conjugates of NS5A ISGylation. IP, immunoprecipitation; IB, immunoblotting.

HERC5 is a specific E3 ligase for NS5A ISGylation.

HERC5, TRIM25 (also referred to as EFP), and HHARI have been shown to act as E3 ligases for ISGylation to different substrates (36–38). To investigate the specificity of E3 ligases in NS5A ISGylation, we coexpressed 293T cells with NS5A-Myc-His₆ and each of the HA-tagged E3 ligases, including HA-HERC5, HA-TRIM25, or HA-HHARI, together with E1 (UBE1L) and E2 (UbcH8), followed by detection with anti-NS5A antibody. The immunoblot analysis revealed that the transduction of HERC5, but not of TRIM25 and HHARI, strongly induced NS5A ISGylation (Fig. 3A, lane 2, top). These results suggest that HERC5 is a specific E3 ligase for NS5A ISGylation.

FIG 3.

HERC5 is an E3 ligase that is specific for NS5A ISGylation. (A) Expression vector encoding NS5A-Myc-His₆ was coexpressed with FLAG-ISG15 and each of three HA-tagged E3 ligases, including (a) HERC5, (b) TRIM25, and (c) HHARI, together with E1 (UBE1L) and E2 (UbcH8) in 293T cells, followed by detection with anti-NS5A antibody. (B) Expression vector encoding HA-NS5A (1b, Con1) was coexpressed with FLAG-ISG15 together with E1, E2, and E3 ligase (HERC5) or without each of the enzymes, as indicated, in 293T cells, followed by immunoprecipitation with anti-HA rabbit PAb and detection with anti-ISG15 mouse MAb or anti-HA rabbit PAb. (C) Expression vector encoding HA-NS5A was coexpressed with FLAG-ISG15 together with E1, E2, and E3 ligase (HERC5) or an inactive C994A HERC5 mutant in 293T cells, followed by immunoprecipitation with anti-HA rabbit PAb and detection with anti-ISG15 mouse MAb and anti-HA rabbit PAb. Input samples (Lysate) were detected with anti-HA rabbit PAb, anti-FLAG mouse MAb, or anti-HERC5 rabbit PAb. The asterisks indicate the various ISG15-conjugated NS5A proteins.

To further investigate whether HERC5 acts as a specific E3 ligase to promote NS5A ISGylation, we coexpressed HA-NS5A and FLAG-ISG15 together with E1 (UBE1L), E2 (UbcH8), and E3 (HERC5) in 293T cells, followed by immunoprecipitation with anti-HA antibody and detection with anti-ISG15 antibody. The immunoprecipitation analysis coupled with immunoblotting clearly detected NS5A ISGylation when all components were expressed (Fig. 3B, lane 3, first, second, and third panels, asterisks), whereas NS5A ISGylation was markedly diminished in the absence of HERC5 (Fig. 3B, lane 6, top, middle, and bottom). In contrast, when the cells were depleted of E1 or E2, both NS5A ISGylation and ISG15 conjugates had completely disappeared (Fig. 3B, lanes 4 and 5, top, middle, and bottom).

Next, to determine whether HERC5 ligase activity is involved in NS5A ISGylation, we performed an ISGylation assay of NS5A protein using the catalytically inactive C994A mutant of HERC5 (bearing a change of C to A at position 994). The transduction of wild-type (WT) HERC5 induced NS5A ISGylation, whereas the transduction of C994A HERC5 showed a marked reduction in NS5A ISGylation (Fig. 3C, lanes 2 and 3, first, second, and third panels), indicating the requirement of HERC5 ligase activity for NS5A ISGylation. These results indicate that HERC5 acts as a specific E3 ligase for NS5A ISGylation.

Identification of ISGylation sites on NS5A protein.

ISG15 is conjugated to the substrate protein via a Lys residue as a monomer, but not as a polymer (39). Because some slowly migrating forms of NS5A were detected in the NS5A ISGylation (Fig. 1), we hypothesized that NS5A protein accepts ISGylation at multiple Lys residues. To map the ISGylation sites on NS5A protein (1b, Con1), we constructed a series of NS5A mutants in which all except 1 of the 14 Lys residues were mutated to Ala residues (K mutant series). We also constructed a series of NS5A mutants containing point mutations of Lys to Arg at the corresponding Lys residues (K/R mutant series) (Fig. 4A). The immunoprecipitation analysis coupled with immunoblotting revealed that HA-NS5A (WT) exhibited the NS5A ISGylation but HA-NS5A with all Lys residues mutated to Ala (K-Null) did not (Fig. 4B and D, lanes 2 and 8, top and middle), indicating that these slowly migrating bands are dependent upon the Lys residues of NS5A protein.

FIG 4.

Identification of ISGylation sites on NS5A protein. (A) Schematic diagram of HCV NS5A (1b, Con1). NS5A possesses 14 Lys residues within the coding sequence. (B, D) Expression vector encoding HA-NS5A or HA-NS5A mutants in which all Lys residues except one (K44, K68, K166, K215, and K308) are mutated to Ala and HA-NS5A mutants containing a point mutation of Lys to Arg at corresponding Lys residues (K44R, K68R, K166R, K215R, and K308R) was coexpressed with FLAG-ISG15 together with E1 (UBE1L), E2 (UbcH8), and E3 ligase (HERC5) in 293T cells, followed by immunoprecipitation with anti-FLAG mouse MAb and detection with anti-HA rabbit PAb. Null indicates a mutant with mutation of all of the Lys residues on NS5A. (C) Expression vector encoding either HA-NS5A (WT) or HA-NS5A mutants (K44, K68, K166, K215 and K308) was coexpressed with FLAG-ISG15 or FLAG-ISG15-AA together with E1, E2, and E3 ligase (HERC5) in 293T cells, followed by immunoprecipitation with anti-HA rabbit PAb and detection with anti-FLAG mouse MAb. (E) Expression vector encoding either HA-NS5A (WT) or NS5A Lys double mutants (K44/68R) was coexpressed with FLAG-ISG15 together with E1, E2, and E3 ligase (HERC5) in 293T cells, followed by immunoprecipitation with anti-FLAG mouse MAb and detection with anti-HA rabbit PAb. Input samples (Lysate) were detected with anti-HA rabbit PAb and anti-FLAG mouse MAb. The asterisks indicate the various ISG15-conjugated NS5A proteins.

The immunoblot analysis using the K mutant series of HA-NS5A revealed that five Lys residues (K44, K68, K166, K215, and K308) on NS5A are responsible for ISGylation with different molecular masses (Fig. 4B, top and middle, red boxes). In detail, 75-kDa conjugates were mediated by both the K44 and K68 residues. The 100-kDa conjugates were mediated by the K166, K215, and K308 residues. The slowly migrating bands with the 100-kDa conjugates were mediated by the K166 and K215 residues. The clearest band was mediated by the K308 residue, suggesting that three different Lys residues (i.e., K166, K215, and K308) are involved in the 100-kDa band of ISGylated NS5A. On the other hand, there were no NS5A ISGylation bands from the remaining NS5A mutants with all lysine residues except one mutated to an Ala residue (referred to as K20, K26, K139, K240, K277, K330, K348, K358, and K378 mutants) (data not shown). Moreover, we observed that the NS5A ISGylation via the K44, K68, K166, K215, or K308 Lys residue was completely abolished in cells coexpressed with the FLAG-ISG15-AA mutant (Fig. 4C, top, lanes 7 to 12).

Using a series of K/R mutants of HA-NS5A, NS5A ISGylation via the K308 residue was abolished by the mutation of K308R, whereas the slowly migrating bands with the 100-kDa conjugates via the K166 and K215 residues were still detected (Fig. 4D, top and middle, lane 7, red boxes). The 130-kDa conjugates of NS5A ISGylation (Fig. 4D, top and middle, three asterisks) were also abolished when the cells were transduced with the K308R mutant (Fig. 4D, top, lane 7), suggesting the possibility that the K308 residue may also be involved in the 130-kDa conjugates of NS5A ISGylation. On the other hand, the NS5A ISGylation with 75-kDa conjugates via the K44 and K68 residues was reduced by the K44R and K68R mutations (Fig. 4D, top and middle, lanes 3 and 4, red boxes). Further immunoblot analysis using the K44/68R double mutation revealed that the NS5A ISGylation with 75-kDa conjugates was completely abolished (Fig. 4E, top and middle, lanes 2 and 4, asterisks), suggesting that the K44 and K68 residues are both involved in the 75-kDa band of ISGylated NS5A. Taken together, these results suggest that ISG15 is conjugated at the K44, K68, K166, K215, and K308 residues on NS5A protein.

NS5A ISGylation via K308 participates in the positive regulation of HCV RNA replication.

To determine the role of the Lys residues on NS5A in HCV RNA replication, we constructed plasmids expressing a series of HCV (1b, Con1) luciferase reporter subgenomic RNA replicons (SGRs) (HCV-SGR-Luc RNA), each with a point mutation of NS5A that replaces one of the 14 Lys residues with Arg. We then electroporated in vitro-transcribed RNAs into Huh-7.5 cells to evaluate the viral replication by measuring the luciferase activity (Fig. 5A). The transduction of HCV-SGR-Luc RNA derived from the NS5A WT exhibited high luciferase activities at 48, 72, and 96 h, but transductions of the NS5A K-Null mutant lacking all of the Lys residues and the polymerase-inactive mutant (GND), which contains a point mutation at the GDD motif in NS5B, did not (Fig. 5B). These results suggest an essential role of Lys residues on NS5A for HCV RNA replication.

FIG 5.

NS5A ISGylation via K308 participates in the positive regulation of HCV RNA replication. (A) Experimental procedures for the HCV replication assay. Huh-7.5 cells were electroporated with either HCV subgenomic replicon (SGR) (genotype 1b, Con1) RNA possessing the firefly luciferase gene (HCV-SGR-Luc RNA) or the replicon possessing an NS5A Lys mutant in which a Lys is replaced with Arg. The luciferase activity was measured at the indicated time points after electroporation. The luciferase activity measured at 4 h after electroporation was used to normalize for the input RNA. IRES, internal ribosomal entry site; UTR, untranslated region; EMCV, encephalomyocarditis virus. (B to D) HCV-SGR-Luc RNAs derived from Con1 WT, a replication-defective mutant (GND), or a series of NS5A Lys mutants were electroporated into Huh-7.5 cells. At 48, 72, and 96 h postelectroporation of RNA, cells were harvested, lysed, and subjected to a luciferase assay. Results shown are the mean normalized relative light unit (RLU) values from triplicates ± standard errors (SE) at the indicated time points. *, P < 0.05, and **, P < 0.01, versus the results for cells transduced with SGR-Luc Con1 WT RNAs.

To investigate the importance of ISGylation sites on NS5A protein, we analyzed HCV replication using HCV-SGR-Luc RNA with a substitution of K44R, K68R, K166R, K215R, or K308R. As shown by the results in Fig. 5C, the transduction of HCV-SGR-Luc RNA with a K308R mutation exhibited a marked reduction of luciferase activity. The luciferase activities induced by the other nine mutants were comparable to that of the WT, except for the K348R mutation (Fig. 5D). This result is consistent with a report that proposed the positive regulation of HCV RNA replication through NS5A SUMOylation at residue K348 (40). Taken together, these results suggest that residue K308 on NS5A participates in HCV RNA replication through NS5A ISGylation.

The involvement of NS5A ISGylation in HCV genotypes.

We next investigated the role of NS5A ISGylation in several HCV genotypes: 1a (H77c), 2a (JFH1), 3a (S52), 4a (ED43), and 5a (SA1). The Lys residues on NS5A are well conserved among all the HCV genotypes, with the exception of genotype 2a (JFH1) (Fig. 6A). Three Lys residues (K20, K26, and K139) were well conserved among all the HCV genotypes. In contrast, residue K308, which is important for HCV RNA replication through the NS5A ISGylation of genotype 1b (Con1), was conserved only in genotype 3a (S52).

FIG 6.

NS5A ISGylation in HCV genotypes. (A) Schematic diagrams of the locations of Lys (K) residues on HCV NS5A of different genotypes (1b, Con1; 1a, H77c; 2a, JFH1; 3a, S52; 4a, ED43; 5a, SA1). (B) Expression vector encoding HA-NS5A of different genotypes (1a [H77c], 2a [JFH1], 3a [S52], 4a [ED43], and 5a [SA1]) was coexpressed with FLAG-ISG15 together with E1 (UBE1L), E2 (UbcH8), and E3 ligase (HERC5) in 293T cells, followed by immunoprecipitation with anti-HA rabbit PAb and detection with anti-ISG15 mouse MAb and anti-HA antibody. Input samples (Lysate) were detected with anti-HA rabbit PAb and anti-FLAG mouse MAb. The slowly migrating bands indicate various ISG15-conjugated NS5A proteins.

To examine whether NS5A genotypes accept ISGylation, we coexpressed 293T cells with each of the HA-NS5A genotypes and FLAG-ISG15 together with E1 (UBE1L), E2 (UbcH8), and E3 (HERC5), followed by immunoprecipitation with anti-HA antibody and detection with anti-ISG15 or anti-HA antibody, respectively. The immunoprecipitation analysis coupled with immunoblotting revealed that the several slowly migrating forms of NS5A ISGylation were detected in cells cotransduced with each of the NS5A genotypes (Fig. 6B, first, second, and third panels, lanes 2, 4, 6, 8, and 10), although less ISG15 conjugation of NS5A from genotype 5a (SA1) was detected. These results indicate that the NS5A genotypes investigated in this study have the potential to accept ISGylation in distinct conjugation manners.

The role of ISG15 in HCV propagation.

To further examine the role of ISG15 in HCV propagation, we established Huh-7/scr cells stably expressing small hairpin RNA (shRNA) targeted to ISG15 (sh-ISG15), and we evaluated the effect of ISG15 knockdown on cell culture-adapted HCV (HCVcc) (J6/JFH1) propagation. The core protein levels were markedly decreased in the cells expressing sh-ISG15 (ISG15) at 2 to 3 days postinfection compared to the core protein levels in the cells expressing scrambled sh-ISG15 (sc-ISG15) (Fig. 7A, top, lanes 2, 4, 6, and 8). We also observed significant reductions of the extracellular virus titers (Fig. 7B), extracellular core protein levels (Fig. 7C), and intracellular HCV RNA (Fig. 7D) in the cells stably expressing sh-ISG15. In addition, the luciferase activity of transduced J6/JFH1-FGR-Luc RNA in the cells stably expressing sh-ISG15 was significantly reduced compared to that in the sc-ISG15-expressing cells (Fig. 7E, left). Here, we confirmed that the NS3 protein levels were also decreased in the sh-ISG15-expressing cells (Fig. 7E, right, top, lanes 1 and 3). Moreover, to exclude possible off-target effects in the sh-ISG15-expressing cells infected with HCV, we transduced the cells with a FLAG-tagged ISG15 wobble mutant (FLAG-ISG15-shR), which is resistant to the shRNA targeted to ISG15. In accordance with the rescue of ISG15 expression, the expression of core protein was shown to be recovered in the sh-ISG15 cells transduced with FLAG-ISG15-shR compared to its expression in the untransduced cells (Fig. 7F, top and middle, lanes 6 and 8).

FIG 7.

The silencing of ISG15 suppresses HCV propagation in human hepatoma Huh-7/scr cells. (A) Stable knockdown cell lines based on the Huh-7/scr cells expressing shRNA targeted to ISG15 or its scrambled sequence (sc-ISG15) were infected with HCVcc (J6/JFH1) at a multiplicity of infection (MOI) of 1. At 2 or 3 days postinoculation, cell lysates were harvested and subjected to immunoblot analysis with the indicated antibodies. (B to D) The extracellular HCV titers (B), the amounts of extracellular core protein (C), and the viral RNA expression levels (D) were determined at 3 days postinoculation by a focus-forming assay (FFA), core ELISA, and real-time PCR, respectively. Results are the mean values from triplicates ± SE. *, P < 0.05 versus the results for Huh-7/scr cells (sc-ISG15). Data from the real-time PCR were normalized to the amount of GAPDH mRNA. (E) HCV full-genomic replicon (FGR) RNAs possessing the firefly luciferase gene (HCV-FGR-Luc RNA) and derived from J6/JFH1 WT or a replication-defective mutant (GND) were electroporated into Huh-7.5 cells. At 72 h postelectroporation of RNA, cells were harvested, lysed, and subjected to a luciferase assay (left) or IB analysis with the indicated antibodies (right). Results shown are the mean normalized RLU values ± SE from triplicates at 72 h after electroporation. (F) Huh7/scr cells expressing shRNA targeted to ISG15 and sc-ISG15 were infected with HCVcc (J6/JFH1) at an MOI of 1. At 2 days postinoculation, the cells were transfected with pCAG-FLAG-ISG15 (WT) or pCAG-FLAG-ISG15 (shR) and subjected to immunoblot analysis with the indicated antibodies at 72 h after transfection.

Next, to examine a role of endogenous ISG15 in NS5A ISGylation in the HCV-replicating cells, we transfected the JFH1/5A-Con1 RNA together with ISGylation components into the sc-ISG15 and sh-ISG15 cells by electroporation. The immunoprecipitation analysis coupled with immunoblotting using NS5A-specific antibody indicated that the NS5A ISGylation with 100-kDa conjugates was substantially diminished in the sh-ISG15 cells compared to that in the sc-ISG15 cells (Fig. 8, first and second panels, lanes 2 and 4, asterisks). Taken together, these results suggest that ISG15 plays an important role in efficient HCV propagation through NS5A ISGylation.

FIG 8.

Reduced expression of NS5A ISGylation in the ISG15 knockdown cells. Huh7/scr cells expressing shRNA targeted to ISG15 and sc-ISG15 were electroporated with the in vitro-transcribed chimeric JFH1/5A-Con1 RNA together with E1 (UBE1L), E2 (UbcH8), and E3 ligase (HA-tagged HERC5), followed by immunoprecipitation analysis coupled with immunoblotting using NS5A-specific antibody. Input samples (Lysate) were detected with anti-NS5A mouse MAb, anti-HA rabbit PAb, anti-GAPDH mouse MAb, or anti-ISG15 mouse MAb. The asterisk indicates the 100-kDa conjugates of NS5A ISGylation. NS, nonspecific bands.

NS5A ISGylation promotes the association of NS5A with CypA.

Finally, we investigated how NS5A ISGylation is involved in the positive regulation of HCV RNA replication. Based on the report that cyclophilin A (CypA) may be involved in the positive regulation of HCV RNA replication through an association with NS5A domain II (DII) (41–49), we noticed that the CypA-binding sequence (³⁰⁸KFPRAMPIWARPDYNPP³²⁴) (45) within the NS5A (1b, Con1) DII includes the K308 residue (Fig. 9A). This CypA-binding sequence also overlapped the RNA- and NS5B-binding sequences. We therefore hypothesized that NS5A ISGylation via the K308 residue within the CypA-binding sequence may be involved in the regulation of HCV RNA replication. To address this hypothesis, we coexpressed 293T cells with HA-NS5A and CypA-Myc-His₆ in the presence or absence of the ISGylation components, followed by immunoprecipitation with anti-Myc- or anti-HA antibody and detection with the indicated antibodies. Interestingly, the induction of ISGylation resulted in a marked enhancement of the association of HA-NS5A with CypA-Myc-His₆ (Fig. 9B, top, left and right, lanes 3). Further immunoprecipitation analysis revealed that CypA did not interact with ISG15 in the cells cotransduced with or without ISGylation components, whereas CypA interacted with NS5A in the cells transduced with ISGylation components (Fig. 10). These results indicate that NS5A ISGylation may function to support the association of NS5A and CypA. Additionally, HA-NS5A from several HCV genotypes also showed the enhanced association with CypA-Myc-His₆ by the induction of ISGylation (Fig. 11), suggesting a common regulatory role in HCV genotypes.

FIG 9.

NS5A ISGylation promotes the association of NS5A with cyclophilin A (CypA). (A) The CypA-binding region on NS5A (1b, Con1) domain II. The amino acid regions from 305 to 311, 308 to 324, and 308 to 329 comprising the RNA-binding motif, the CypA-binding region (highlighted in pink), and the NS5B-binding region, respectively, are indicated. W316 is one of the key residues on NS5A for the protein-protein interactions with CypA and NS5B. (B) Expression vector encoding HA-NS5A (1b, Con1) was coexpressed with C-terminal Myc-His₆-CypA in the presence or absence of FLAG-ISG15 together with E1 (UBE1L), E2 (UbcH8), and E3 ligase (HERC5) in 293T cells, followed by immunoprecipitation with anti-Myc mouse MAb or anti-HA rabbit PAb and detection with anti-HA rabbit PAb and anti-Myc mouse MAb, respectively. (C) HA-NS5A was coexpressed with CypA-Myc-His₆ in the presence or absence of FLAG-ISG15 together with E1, E2, and E3 ligase in 293T cells. At 24 h posttransfection, the cells were treated with 5 μg/ml of CsA for 24 h, followed by immunoprecipitation with anti-Myc mouse MAb and detection with anti-HA rabbit PAb and anti-Myc mouse MAb. Input samples (Lysate) were detected with anti-Myc mouse MAb, anti-HA rabbit PAb, and anti-FLAG mouse MAb. NS, nonspecific bands. The binding ratio of HA-NS5A and CypA-Myc-His₆ was determined to measure the band intensities (HA-NS5A [first panel]/CypA-Myc-His₆ [second panel], lanes 3 and 6), using ImageQuant TL software (version 7). Results are the mean values from triplicates ± SE. *, P < 0.05 versus the results for the cells treated with dimethyl sulfoxide (DMSO). HC, immunoglobulin heavy chain. (D) The proximity ligation assay (PLA). Huh-7.5 cells were coexpressed with HA-NS5A and CypA-Myc-His₆ together with pEGFP or pEGFP-ISG15 plus E1, E2, and E3 ligase, respectively. Cells were fixed, permeabilized, and incubated with anti-HA rabbit PAb and anti-Myc mouse MAb, followed by staining with PLA probes. Red fluorescent spots indicate an interaction with NS5A and CypA. Nuclei were stained with Hoechst 33342. Scale bars, 10 μm.

FIG 10.

NS5A, but not CypA, associates with ISG15. Expression vector encoding either CypA-Myc-His₆ or NS5A-Myc-His₆ was coexpressed in the presence or absence of FLAG-ISG15 together with or without E1 (UBE1L), E2 (UbcH8), and E3 ligase (HERC5) in 293T cells, followed by immunoprecipitation with anti-FLAG mouse MAb and detection with anti-Myc mouse MAb or anti-FLAG mouse MAb. Input samples (Lysate) were detected by anti-Myc mouse MAb and anti-FLAG mouse MAb. HC, immunoglobulin heavy chain; LC, immunoglobulin light chain.

FIG 11.

Enhancement of the interaction between NS5A and CypA by ISGylation in an HCV genotype-dependent manner. Expression vector encoding each of the HA-NS5A genotypes (1a [H77c], 2a [JFH1], 3a [S52], 4a [ED43], and 5a [SA1]) was coexpressed with CypA-Myc-His₆ in the presence or absence of FLAG-ISG15 together with E1 (UBE1L), E2 (UbcH8), and E3 ligase (HERC5) in 293T cells, followed by immunoprecipitation with anti-Myc mouse MAb and detection with anti-HA rabbit PAb and anti-Myc mouse MAb. Input samples (Lysate) were detected with anti-Myc mouse MAb, anti-HA rabbit PAb, and anti-ISG15 mouse MAb.

To confirm the association of NS5A with CypA, we treated the cells with cyclosporine (CsA). The association of NS5A with CypA was significantly inhibited by the treatment with CsA (Fig. 9C, first panel, lane 6). Increasing amounts of CsA resulted in reductions of the NS5A-CypA association (Fig. 12, first panel, lanes 8 to 10). These results suggest that ISGylation promotes the association of NS5A with CypA.

FIG 12.

CsA interferes with the ISGylation-promoted association of NS5A with CypA in a dose-dependent manner. Expression vector encoding HA-NS5A was coexpressed with CypA-Myc-His₆ in the presence or absence of FLAG-ISG15 together with E1 (UBE1L), E2 (UbcH8), and E3 ligase (HERC5) in 293T cells. At 24 h posttransfection, the cells were treated with 5, 0.5, or 0.05 μg/ml of CsA for 24 h, followed by immunoprecipitation with anti-Myc mouse MAb and detection with anti-HA rabbit PAb and anti-Myc mouse MAb. Input samples (Lysate) were detected with anti-Myc mouse MAb, anti-HA rabbit PAb, and anti-FLAG mouse MAb. NS, nonspecific bands.

To further investigate the association of NS5A with CypA in the cells, we performed a proximity ligation assay (PLA) to detect protein-protein interactions as fluorescence signal spots in cells. We cotransduced cells with enhanced green fluorescent protein (EGFP) or EGFP-ISG15 plus ISGylation components together with HA-NS5A and CypA-Myc-His₆. A strong PLA signal (Fig. 9D, red spots) was observed predominantly in the cytoplasm in the cells cotransduced with EGFP-ISG15 compared to those cotransduced with EGFP.

To determine whether the enhanced association of NS5A with CypA is dependent on the ISGylation of NS5A, we performed an immunoprecipitation analysis using the mutants of ISG15 and NS5A. The transduction of FLAG-ISG15, but not the FLAG-ISG15-AA mutant, resulted in an enhanced association of HA-NS5A with CypA-Myc-His₆ (Fig. 13A, first panel, lanes 3 and 4), suggesting that NS5A ISGylation enhances the association of NS5A with CypA. A further immunoprecipitation analysis revealed that WT NS5A, but not the K-null mutant of HA-NS5A, resulted in an enhanced association of NS5A with CypA-Myc-His₆ in the cells cotransduced with ISGylation components (Fig. 13B, first panel, lanes 3 and 6), suggesting the involvement of NS5A ISGylation in the enhanced association of NS5A with CypA.

FIG 13.

Enhancement of the interaction between NS5A and CypA in an ISGylation-dependent manner. (A) Expression vector encoding HA-NS5A (1b, Con1) was coexpressed with CypA-Myc-His₆ in the presence or absence of FLAG-ISG15 or its conjugation-defective mutant (FLAG-ISG15-AA) together with E1 (UBE1L), E2 (UbcH8), and E3 ligase (HERC5) in 293T cells, followed by immunoprecipitation with anti-Myc mouse MAb and detection with anti-HA rabbit PAb and anti-Myc mouse MAb. (B, C) Expression vector encoding either HA-NS5A (1b, Con1) or NS5A Lys mutant K-Null (B) or K308R (C) was coexpressed with CypA-Myc-His₆ in the presence or absence of FLAG-ISG15 together with E1, E2, and E3 ligase in 293T cells, followed by immunoprecipitation with anti-Myc mouse MAb or anti-HA rabbit PAb and detection with anti-HA rabbit PAb or anti-Myc mouse MAb, respectively. Input samples (Lysate) were detected by anti-Myc mouse MAb, anti-HA rabbit PAb, and anti-FLAG mouse MAb. NS, nonspecific bands. The binding ratios of HA-NS5A and CypA-Myc-His₆ were determined to measure the band intensities of HA-NS5A (first panel)/CypA-Myc-His₆ (second panel), lanes 3 and 4 (A), HA-NS5A (first panel)/CypA-Myc-His₆ (second panel), lanes 3 and 6 (B), and CypA-Myc-His₆ (first panel)/HA-NS5A (second panel), lanes 3 and 4 (C), using ImageQuant TL software (version 7). Results are the mean values from triplicates ± SE. *, P < 0.05, and **, P < 0.01, versus the results for the cells transduced with WT NS5A.

We also observed that the transduction of the HA-NS5A K308R mutant resulted in a significant reduction of the association with CypA-Myc-His₆ (Fig. 13C, first panel, lane 4), whereas the transduction of either the HA-NS5A K44R mutant or the HA-NS5A K166R mutant had no effect on the interaction between CypA-Myc-His₆ and HA-NS5A (Fig. 14, first panel, lanes 3 and 4). These results indicate that the K308 residue on NS5A is responsible for not only the acceptance of ISGylation but also the association with CypA.

FIG 14.

The K44 and K166 residues on NS5A are not involved in the enhanced association with CypA. Expression vector encoding HA-NS5A (1b, Con1) or an NS5A Lys mutant (K44R or K166R) was coexpressed with CypA-Myc-His₆ in the presence or absence of FLAG-ISG15 together with E1 (UBE1L), E2 (UbcH8), and E3 ligase (HERC5) in the 293T cells, followed by immunoprecipitation with anti-HA rabbit PAb and detection with anti-Myc mouse MAb and anti-HA rabbit PAb. Input samples (Lysate) were detected by anti-Myc mouse MAb, anti-HA rabbit PAb, and anti-FLAG mouse MAb.

Collectively, these results suggest that the HCV NS5A protein has the potential to accept ISG15 conjugation via a specific Lys (K) residue, leading to enhanced association of NS5A with CypA for efficient HCV RNA replication (Fig. 15).

FIG 15.

A model of the functional role of NS5A ISGylation in efficient HCV RNA replication. An intracellular ISG15 is covalently conjugated to HCV NS5A protein by the action of three enzymes, such as E1 (UBE1L), E2 (UbcH8) and E3 ligase (HERC5). The K308 residue within domain II (DII) of NS5A (genotype 1b, Con1) is the acceptor site for ISGylation. The ISGylated NS5A DII may allow its association with CypA, facilitating efficient HCV RNA replication.

DISCUSSION

The IFN-inducible protein ISG15 and ISGylation are believed to play important roles in antiviral activity against many types of RNA and DNA viruses (9, 26). Our present findings demonstrated that an intracellular ISG15 is involved in the positive regulation of HCV RNA replication through NS5A ISGylation at multiple specific Lys residues. Our analyses further demonstrated that the NS5A ISGylation was specific to HCV genotypes, although the details of their individual contributions to the viral replication remain to be elucidated. Further mechanistic insights indicated that the induction of ISGylation may lead to the enhancement of the association of NS5A with CypA (Fig. 15), which is the critical host factor for HCV RNA replication (3, 4). To our knowledge, the present study is the first to demonstrate the proviral ISG15 function through the selective recruitment of CypA for efficient HCV RNA replication.

Domain II (DII) and domain III (DIII) of NS5A have been implicated as the region of a potential binding site for CypA (41–50). Here, we observed that the K308 residue in DII of NS5A (1b, Con1), which is one of the ISG15 conjugation sites on the NS5A protein, was located within the CypA-binding region (³⁰⁸KFPRAMPIWARPDYNPP³²⁴). In fact, our analyses demonstrated that mutation of K308 on NS5A is correlated with both attenuation of the interaction with CypA and significant reduction of HCV RNA replication in cells transduced with HCV reporter RNA (Fig. 5C and Fig. 13C). However, the exact role of the CypA recruitment by NS5A ISGylation via DII of NS5A remains to be elucidated, due to the lack of evidence of the K308 residue within DII of NS5A from other HCV genotypes, except for genotype 3a (S52) (Fig. 6A). We speculate that the nearby Lys residues (e.g., K304/305) on NS5A from other HCV genotypes may compensate for the role of the NS5A ISGylation on the K308 residue, except for genotype 2a (JFH1), because of the lack of Lys residues within DII of NS5A (Fig. 6A). The recent report by Shanmugam and colleagues (51) suggests that NS5A dimer interface residues within DI have an important role in the association of NS5A with CypA via DII. We thus speculate that the Lys residues within DI of NS5A from genotype 2a (JFH1) may be involved in the NS5A ISGylation and promote an association with CypA. Further investigation of ISGylation sites on NS5A from genotype 2a (JFH1) is needed to understand the genotype-specific regulation in NS5A-CypA interaction. Crystal structures of NS5A from genotypes 1a (H77c strain) and 1b (Con1 strain) have been reported (6–8), whereas the crystal structure of NS5A from genotype 2a (JFH1 strain) has not been reported yet. Detailed information about the crystal structure of NS5A from genotype 2a may give us new insights into the role of the NS5A ISGylation and recruitment of CypA. Further investigation of the intermolecular interactions of CypA and DII of NS5A among the HCV genotypes may contribute to our understanding of the functional link between CypA recruitment and NS5A ISGylation.

It was reported that CsA, the prototype Cyp inhibitor, may suppress HCV RNA replication by interfering with the peptidyl-prolyl cis-trans-isomerase (PPIase) activity of Cyp (52). On the other hand, Daito et al. (53) reported that SCY-635, which is one of the CsA derivatives, may suppress double-stranded RNA (dsRNA)-mediated protein kinase R (PKR) activation through a dissociation of the PKR-CypA complex, resulting in augmentation of the expression of several ISGs to suppress HCV replication. Intriguingly, the treatment of the cells with CsA suppressed not only the association of NS5A with CypA but also the NS5A ISGylation (Fig. 9C, third panel, lane 6). Further studies are needed to clarify the alternative mechanisms of Cyp inhibitors as anti-HCV agents.

Our present findings demonstrated that NS5A protein may accept ISG15 conjugation at five different Lys residues, although only the K308 residue exhibited a significant impact for HCV RNA replication. The virological impact of the other four Lys residues involved in the NS5A ISGylation remains to be elucidated. Immunoblot analyses indicated that the mono-ISG15-conjugated form is approximately 75 kDa via K44 and K68 on NS5A, which equals the molecular mass of one ISG15 (15 kDa) plus one NS5A (56 to 58 kDa). ISG15 is believed to be covalently conjugated to a target substrate protein via a Lys residue as a monomer but not as a polymer (39). However, the conjugated form via K166, K215, or K308 is approximately 100 kDa, suggesting the possibility that there might be poly-ISGylated NS5A. It is interesting to determine how ISGylation confers ISGylated NS5A at 100 kDa. Another possibility is that an additional posttranslational modification, such as ubiquitin and another ubiquitin-like modifier, may be involved in NS5A ISGylation. Interestingly, it was reported that ubiquitin is a substrate of ISG15 conjugation in a degradation signal-independent manner (54). We thus speculate that poly-ISGylation or ISG15-ubiquitin mixed chains might be involved in NS5A ISGylation. The role of the other protein modifications to ISGylated NS5A protein merits further investigation.

We attempted to detect the endogenous ISGylated NS5A in HCV-infected cells and HCV replicon cells. However, it was difficult to detect the endogenous ISGylated NS5A in the HCV-replicating cells, probably due to a transient ISG15 conjugation to substrate protein. In general, the covalent ISG15 conjugation is known to be immediately removed from target substrates by USP18, an ISG15-deconjugating enzyme.

It was reported that NS5A protein may accept ISG15 conjugation at the K378 residue, which leads to the proteasomal degradation of NS5A protein (55). Those authors proposed that this may be antiviral activity against HCV infection by ISG15. However, we obtained evidence herein that suggests that (i) NS5A protein accepts ISG15 conjugation at multiple Lys residues other than K378 and (ii) the NS5A ISGylation acts as a proviral function for HCV RNA replication.

There are some recent reports that are in agreement with our notion that ISG15 functions as a proviral factor for HCV infection. For example, overexpression of UBE1L showed the promotion of HCV particle production in HCV J6/JFH1-infected cells, whereas the HCV particle production was decreased in cells in which UBE1L was silenced (23). It was also reported that ISG15 may contribute to the HCV life cycle through the suppression of IFN responsiveness in patients (24). A high expression of hepatic ISG15 in patients was correlated with a high viral load and low responsiveness to interferon alpha (IFN-α) treatment. These results suggest that ISG15 acts as a proviral factor that promotes HCV propagation and counteracts the IFN-signaling pathway.

In summary, the results of the present study demonstrated that NS5A ISGylation may participate in the positive regulation of HCV RNA replication through an enhancement of NS5A-CypA interaction. Targeting the machinery of ISGylation may lead to the development of the next generation of therapeutics for the treatment of chronic HCV infection.

MATERIALS AND METHODS

Cell culture and viruses.

Huh-7.5 cells (kindly provided by C. M. Rice, The Rockefeller University, NY) and Huh-7/scr cells (the differentiated Huh-7 cells) (kindly provided by F. V. Chisari, The Scripps Research Institute, CA) are highly permissive for propagation of the cell culture-adapted HCV of the genotype 2a JFH1 strain (HCVcc) (56, 57). Huh-7.5, Huh-7/scr, and 293T cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) (high glucose) with l-glutamine (Wako, Osaka, Japan) supplemented with 50 IU/ml penicillin, 50 μg/ml streptomycin (Gibco, Grand Island, NY), and 10% heat-inactivated fetal bovine serum (Biowest, Nuaillé, France) at 37°C in a 5% CO₂ incubator.

Cells were transfected with plasmid DNA using FuGene 6 transfection reagents (Promega, Madison, WI). The pFL-J6/JFH1 plasmid that encodes the entire viral genome of a chimeric strain of HCV-2a, JFH1 (58), was kindly provided by C. M. Rice. The HCV genome RNA was synthesized in vitro using pFL-J6/JFH1 as the template and was transfected into Huh-7.5 cells by electroporation. The virus produced in the culture supernatant was used for the infection experiments.

Plasmids.

The cDNA fragments of NS5A (1b, Con1, and 2a, JFH1) were inserted into the NotI site of pCAG-HA using the In-Fusion HD cloning kit (Clontech, Mountain View, CA). The cDNA fragments of NS5A (genotype 1b, Con1, and genotype 2a, JFH1) with all of the Lys residues mutated to Ala residues were synthesized at Eurofins Genomics (Tokyo) and then cloned into the NotI site of pCAG-HA. We designated the resultant plasmid pCAG-HA-NS5A (K-Null). The construction of NS5A Lys mutants with all of the Lys residues except one mutated to Ala residues was generated by the mutagenesis PCR method using pCAG-HA-NS5A (K-Null) as a template.

HCV subgenomic replicon plasmids (SG-Feo) of various genotypes, including H77c (L+8), Con1 (I), JFH1, S52 (AII), ED43 (VYG), and SA1 (SKIP), and possessing a chimeric gene encoding the firefly luciferase and neomycin resistance were kindly provided by C. M. Rice (59, 60). The cDNA fragments of NS5A genotypes were cloned into the XhoI/NotI site of pCAG-HA using the In-Fusion HD cloning kit. The HCV plasmid encoding JFH1 cDNA by replacing the NS5A gene from the Con1 strain (pJFH1/5A-Con1) was kindly provided by Takaji Wakita, NIID, Japan (61). The cDNA fragment encoding human CypA was amplified by reverse transcription-PCR from the total RNA of Huh-7.5 cells and cloned into pEF1-Myc-His₆ (Invitrogen, Eugene, OR) using the In-Fusion HD cloning kit. The expression plasmids for pEF1-NS5A-Myc-His₆, pCAG-FLAG-ISG15, and its conjugation-defective mutant (pCAG-FL-ISG15-AA) were previously described (31). The FLAG-tagged ISG15 wobble mutant (FLAG-ISG15-shR) was generated by the mutagenesis PCR method using pCAG-FLAG-ISG15 as a template. The cDNA fragments encoding UBE1L, UbcH8, and HERC5 were cloned into the NotI/BglII or SmaI/KpnI sites, respectively, of pCAG-MCS2 using the In-Fusion HD cloning kit.

The C994A HERC5 point mutant was generated by the mutagenesis PCR method using pCAG-MCS-HERC5 as a template. The cDNA fragments of ISG15 were cloned into pEGFP-C3 (Clontech) using the In-Fusion HD cloning kit, and the resultant plasmid was designated pEGFP-ISG15. The construction of the HCV reporter subgenomic replicon (SGR) plasmid encoding the firefly luciferase gene (pSGR-Luc) was kindly provided by R. Bartenschlager (Heidelberg University, Germany). The sequences of the inserts were extensively confirmed by sequencing (Eurofins Genomics).

Antibodies and reagents.

The mouse monoclonal antibodies (MAbs) used in this study were anti-FLAG (M2) MAb (F-3165; Sigma-Aldrich, St. Louis, MO), anti-NS5A MAb (MAB8694; Millipore, Billerica, MA), anti-NS3 MAb (MAB8691; Millipore), anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) MAb (MAB374; Millipore), anti-c-Myc MAb (9E10; Santa Cruz Biotechnology, Santa Cruz, CA), anti-ISG15 MAb (F-9; Santa Cruz Biotechnology), and anti-Core MAb (2H9 clone, kindly provided by T. Wakita, NIID, Japan).

The rabbit polyclonal antibodies (PAbs) used in this study were anti-HA PAb (H-6908; Sigma-Aldrich) and anti-Herc5 PAb (BML-PW0920; Enzo Life Sciences, New York, NY). Horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG and goat anti-rabbit IgG (Molecular Probes, Eugene, OR) and HRP-conjugated donkey anti-goat IgG (Santa Cruz Biotechnology) were used as secondary antibodies. CsA was purchased from Sigma-Aldrich.

Immunoprecipitation and immunoblot analysis.

Cells were transfected with the plasmids using FuGene 6 (Promega), harvested at 48 h posttransfection, and suspended in 0.5 ml of radioimmunoprecipitation assay (RIPA) buffer containing 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.1% SDS, 1% NP-40, 0.5% deoxycholate (DOC), and protease inhibitor cocktail tablets (Roche Molecular Biochemicals, Mannheim, Germany). Cell lysates were incubated for 2 h at 4°C and centrifuged at 20,400 × g for 30 min at 4°C (TOMY centrifuge MX-307 with rotor rack AR015-SC24; TOMY, Tokyo). The supernatant was immunoprecipitated with protein G Sepharose 4 fast flow (GE Healthcare, Buckinghamshire, UK) and incubated with appropriate antibodies at 4°C overnight.

After being washed with the RIPA buffer five times, the samples were boiled in 15 μl of sodium dodecyl sulfate (SDS) sample buffer and then subjected to SDS–10% polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride membranes (PVDF) (Millipore). The membranes were blocked with Tris-buffered saline containing 20 mM Tris-HCl (pH 7.6), 135 mM NaCl, and 0.05% Tween 20 (TBST) containing 5% skim milk at room temperature for 2 h and incubated with corresponding antibodies. The membranes were then incubated with HRP-conjugated secondary antibody at room temperature for 2 h. The immune complexes and cell lysates were visualized with ECL Western blotting detection reagents (GE Healthcare, Buckinghamshire, UK) and detected by an LAS-4000 image analyzer system (GE Healthcare). The band intensities were quantified using ImageQuant TL software (version 7.0).

RNA interference and stable ISG15 knockdown cell clones.

The short hairpin RNA (shRNA) sequences of the sense strands targeted to ISG15 (5′-UGAGCACCGUGUUCAUGAA-3′) and its scrambled sequence (5′-GGACAUCGACGGCUUUAUA-3′) were inserted into the pSilencer 2.1 U6 puro vector (Ambion, Austin, TX). To establish the stable ISG15 knockdown cell clones, Huh-7/scr cells were transfected with the plasmids and drug-resistant clones were selected by puromycin treatment (Sigma-Aldrich) at a final concentration of 1 μg/ml.

Real-time PCR.

Total RNA was prepared from each of the cells using an RNeasy minikit (Qiagen, Valencia, CA). First-strand cDNA was synthesized using the GoScript reverse transcription system (Promega). The real-time PCR was performed using SYBR Premix Ex Taq II (Tli RNaseH plus) (TaKaRa Bio, Shiga, Japan) according to the manufacturer’s protocol. Fluorescence signals were analyzed by a StepOnePlus real-time PCR system (Applied Biosystems, Foster City, CA).

The HCV, ISG15, and GAPDH genes were amplified using the primer pairs 5′-GAGTGTCGTGCAGCCTCCA-3′ and 5′-CACTCGCAAGCACCCTATCA-3′, 5′-AGCGAACTCATCTTTGCCAGTACA-3′ and 5′-CAGCTCTGACACCGACATGGA-3′, and 5′-GCCATCAATGACCCCTTCATT-3′ and 5′-TCTCGCTCCTGGAAGATGG-3′, respectively. The expression of each of the genes was normalized to that of the GAPDH gene.

Quantification of extracellular core protein.

The HCV core protein in the culture supernatants was quantified by highly sensitive enzyme immunoassay using the Ortho HCV core antigen enzyme-linked immunosorbent assay (ELISA) (Ortho-Clinical Diagnostics, Raritan, NJ).

HCV infectivity assay.

Briefly, culture supernatants were serially diluted 10-fold and used to infect duplicate 24-well cultures of Huh-7.5 cells. At 24 h postinoculation, the cultures were overlaid with complete DMEM containing a final concentration of 0.25% methylcellulose (Sigma-Aldrich). At 72 h of incubation, the cells were fixed in 4% paraformaldehyde (PFA) and immunohistochemically stained using anti-Core MAb (2H9 clone). The infectious HCV titers were determined based on focus-forming units (FFU)/ml (62).

In vitro transcription of HCV RNA, electroporation into the cells, and reporter analysis.

The HCV replicon pSGR (1b, Con1)-Luc, pJFH1/5A-Con1, and J6/JFH1 reporter replicon plasmids (FGR-luc) were digested with ScaI or XbaI, respectively, and transcribed in vitro by using a MEGAscript T7 kit (Ambion, Austin, TX). Then, 10 μg of in vitro-transcribed HCV RNA was electroporated at 270 V and 960 μF by using a GenePulser Xcell (Bio-Rad, Hercules, CA) into 4 × 10⁶ Huh-7.5 cells treated with BTXpress buffer (BTX, Holliston, MA). The electroporated cells were seeded into 24-well plates and harvested at the time points indicated in Fig. 5B to D and Fig. 7E, and luciferase activity was determined in triplicate using a GloMax 96 microplate luminometer (Promega).The luciferase activity at 4 h after electroporation was used for normalization to account for the differing transduction efficiencies of HCV RNAs.

PLA.

Cells seeded on glass coverslips in 24-well plates were cotransfected with pCAG-HA-NS5A and pEF1-CypA-Myc-His₆ together with pEGFP or pEGFP-ISG15 plus E1 (UBE1L), E2 (UbcH8), and E3 (HERC5), respectively. At 48 h posttransfection, the cells were fixed in 4% PFA at room temperature (RT) for 30 min. After being washed with phosphate-buffered saline (PBS), the cells were permeabilized for 10 min at RT with PBS containing 0.5% Triton X-100. The proximity ligation assay (PLA) was performed on the fixed cells using the Duolink in situ PLA (Sigma-Aldrich) according to the manufacturer’s instructions. Anti-HA rabbit PAb (H-6908) and anti-c-Myc mouse MAb (9E10) were diluted in an antibody diluent supplied with the kit. Positive interactions were produced with the probes for the PLA in combination with the anti-rabbit Minus and anti-mouse Plus. The cells were counterstained with Hoechst 33342 solution, mounted on glass slides, and observed under a confocal laser scanning microscope (LSM700, Carl Zeiss, Oberkochen, Germany).

Statistics.

Results are expressed as the mean values ± standard errors. Statistical significance was determined by Student’s t test for the results shown in Fig. 7B to E, Fig. 9C, and Fig. 13A to C and by one-way analysis of ANOVA for the results shown in Fig. 5B to D. P values of <0.05 and <0.01 were considered significant.