Modulation of HCV replication and translation by ErbB3 binding protein1 isoforms

Abstract

We recently identified a cell-factor, ErbB3 binding protein 1 (Ebp-1), which specifically interacts with the viral RNA genome and modulates HCV replication and translation. Ebp1 has two isoforms, p48, and p42, that result from differential splicing. We found that both isoforms interact with HCV proteins NS5A and NS5B, as well as cell-factor PKR. The p48 isoform, which localizes in the cytoplasm and nuclei, promoted HCV replication, whereas the shorter p42 isoform, which resides exclusively in the cytoplasm, strongly inhibited HCV replication. Transient expression of individual isoforms in Ebp1-knockdown MH14 cells confirmed that the p48 isoform promotes HCV replication, while the p42 isoform inhibits it. We found that Ebp1-p42 significantly enhanced autophosphorylation of PKR, while Ebp1-p48 isoform strongly inhibited it. We propose that modulation of autophosphorylation of PKR by p48 isoform is an important mechanism whereby the HCV virus escapes innate antiviral immune responses by circumventing p42-mediated inhibition of its replication.

INTRODUCTION

Hepatitis C has long been regarded as the silent global epidemic. Worldwide, the number of people infected with HCV is greater than 185 million (Mohd Hanafiah et al., 2013). HCV infection is mostly asymptomatic; the general public has poor knowledge of the disease, and many individuals are unaware of their infection (Denniston et al., 2012). The situation has changed overwhelmingly in the last few years as a result of increased publicity about chronic hepatitis C (CHC) and the availability of improved treatment options.

About 15%–25% of infected individuals clear the virus without treatment. However, the majority of infections persist, leading to CHC, which is closely linked with the risk of liver cirrhosis (LC) and hepatocellular carcinoma (HCC) (Chien et al., 1992). Although enormous progress has been made in understanding the molecular and cell biology of HCV, we have not yet clearly defined the roles of specific cell factors in providing innate immunity to HCV or establishing chronic HCV infection. The HCV genome is a positive-stranded RNA with conserved and highly structured untranslated 5’ and 3’ terminal regions. These regions have multiple regulatory elements that are essential to viral replication and translation. Some cell factors interact with 5’NTR and 3’NTR (Ali and Siddiqui, 1995, 1997; Anwar et al., 2000; Isken et al., 2007; Paek et al., 2008; Randall et al., 2007). We have identified many cell factors associated with HCV 3’NTR and confirmed the effect of some of these factors on HCV replication (Harris et al., 2006).

Recently, using a strategy we designed to capture the replicating HCV RNA genome in situ, we have identified many cell factors associated with viral genome (Upadhyay et al., 2013). One protein we identified was Ebp1, a double-stranded RNA-binding protein (DRBP), which strongly inhibits HCV replication (Upadhyay et al., 2013). In humans, two differently spliced transcripts of Ebp1 mRNA, 2.2 kb, and 1.7 kb, have been found, respectively, to yield translation products p48 and p42(Liu et al., 2006). The sequence alignment of mRNAs of Ebp1 isoforms indicated that both the 5’ and 3’ NTRs of Ebp1-p42 mRNA is shorter than the Ebp1-p48 mRNA. Ebp1-p48 is involved in regulating cell survival; its interaction with Akt kinase suppresses apoptosis (Ahn et al., 2006; Liu et al., 2006). This interaction between Akt kinase and Ebp1-p48 depends on the phosphorylation of Ebp1 on Ser 360 (Lessor and Hamburger, 2001). Ebp1 can promote the initiation of translation by interacting with PKR and inhibiting phosphorylation of the eIF2α subunit of eIF2 (Squatrito et al., 2006). The overexpression of Ebp1-p42 inhibits proliferation of human fibroblasts (Squatrito et al., 2004). Ebp1 also co-represses many proliferation-associated genes, including cyclinD1 and E2F1 (Zhang et al., 2003).

The expression level of Ebp1 significantly decreases in prostate cancer; restoring its level has an anti-tumor effect (Zhang et al., 2008a). The p42 isoform of Ebp1 displays antiproliferative activity (Oh et al., 2010). Replication and production of the influenza virus are significantly reduced by overexpression of Ebp1 (Honda et al., 2007). But although this protein has been studied in many different cancer cell lines, nothing has been reported concerning its role in the HCV life cycle and associated pathogenesis. Also, the molecular mechanism whereby HCV modulates the function of Ebp1 to facilitate its persistent replication and associated pathogenesis is not known.

MATERIALS AND METHODS

Plasmids, oligonucleotides and antibodies: Plasmids carrying the HCV subgenomic replicon (pMH14) (Miyanari et al., 2003) were obtained from Dr. Makoto Hijikata (Kyoto University, Japan). Huh7.5 and pFL-J6/JFH were gifts from Dr. Charles Rice (Lindenbach et al., 2006). Plasmid pEGFP-p42 and pEGFP-p48 were gifts from Dr. Anne Hamburger (University of Maryland, Washington). Plasmids pET28a-Ebp1p42 and pET28a-Ebp1p48 were constructed by cloning the PCR-amplified Ebp1 isoform coding region between BamH1 and Nde1 sites. A bicistronic reporter plasmid, pGEM-REN-HCV IRES-Luc, containing renilla and firefly luciferase, was obtained from Dr. Fanxiu Zhu (Florida) (Kuang et al., 2011). Plasmid pPET- PKR/λPP was purchased from Addgene (Cambridge, MA) for the expression of unphosphorylated PKR. Plasmids expressing GFP fused to the N-terminus of human CD81 (pTRIP-GFP-hCD81), expressing miR-122 (pTRIP-Puro-miR122), as well as their negative controls were generous gifts from Dr. Matthews J Evans (Narbus et al., 2011).

Ebp1 shRNA lentiviral particles were purchased from Santa Cruz Biotechnology (sc-77220-V). Ebp1 siRNA (sense 5’-CUG AAU UUG AGG UAC AUG Att-3’, antisense 5’-UCA UGU ACC UCA AAU UCA Gtt-3’) was purchased from Sigma. Ebp1p48 siRNA sense, 5’-GAU GGG GGG CGA CAU CGC Ctt-3’, antisense, 5’-GGC GAU GUC GCC CCC CAU Ctt-3’ was ordered from Sigma-Aldrich (St. Louis, MO). Control siRNA with scramble sequence (5’- UUC UCC GAA CGU GUC ACG Utt-3’, antisense 5’-ACG UGA CAC GUU CGG AGA Att-3’) was obtained from Santa Cruz. The primers for RT-PCR and real-time RT-PCR of HCV 5’NTR (up 5’- CGG GAG AGC CAT AGT GG-3’), HCV 5’NTR (dn 5’- AGT ACC ACA AGG CCT TTC G −3’), GAPDH mRNA (up: 5’-CTC TGC TCC TCC TGT TCG AC-3’; down: 5’-ATG GGT GGA ATC ATA TTG GAAC-3’); actin mRNA (up: 5’-CAG GCA CCA GGG CGT GAT GG-3’; down: 5’-AGG CGT ACA GGG ATA GCA CA-3’) were purchased from Sigma. Primers for cloning p42 and p48 were purchased from Sigma-Aldrich. Primary antibodies against HCV NS5B, NS5A, Ebp1, and protein kinase PKR were obtained from Santa Cruz Biotechnology

Cell culture: MH14 cells with stable replicating HCV subgenomic replicons and cured MH14 cells devoid of HCV replicons were gifts from Dr. Makoto Hijikata (Kyoto University, Japan) (Murata et al., 2005). MH14 cells were maintained in complete DMEM supplemented with 300 µg/ml of G418 (Calbiochem). Cured MH14 and Huh7.5 cells were cultured in the same medium without G418. Cells were grown at 37°C with 5% CO₂.

Infection of HepG2 cells with JFH1 HCV RNA. We used two sets of HepG2 cells (10⁶ cells) cotransfected with plasmids expressing human CD81 (pTRIP-GFP-hCD81) and miR-122 (pTRIP-Puro-miR122). After 24 h post transfection, cells were layered with HCV-JFH1 virions in the presence of 5-µg/ml polybrene (Dixit et al., 2015). After 6 h, cells were washed two times with PBS, supplemented with fresh medium. Seventy-two hours later, total RNA was isolated from the first set and the level of HCV RNA was quantitated by q-RT-PCR. Cell lysates from the other set of cells were prepared, and Western blotted for Ebp1 and actin.

Preparation of cell extract and Western blotting: Cells washed with phosphate-buffered saline (PBS) were lysed in lysis buffer (1% SDS, 10% glycerol, 1 mM DTT, and 10 mM Tris HCl, pH 6.8), boiled for 5 min, and centrifuged at 13,000 rpm for 10 min at 4°C. The clear supernatant, equivalent to 15 µg of protein, was resolved on 8% SDS-PAGE and transferred to a nitrocellulose membrane. The membrane was processed for Western blotting using the primary and secondary antibody as described earlier (Zhang et al., 2008b).

Preparation of nuclear and cytoplasmic extracts: We prepared cytoplasmic and nuclear extracts from MH14 cells using NE-PER Nuclear and Cytoplasmic Extract Reagent (Pierce, Appleton WI) according to the manufacturer’s protocol. An aliquot of each extract was subjected to SDS-PAGE, and Western blotted for Ebp1. The samples were also Western blotted for poly(ADP-ribose) polymerase (PARP) and α-tubulin as respective markers for nuclei and cytosol.

Transfection of Ebp1 siRNA, p48 siRNA, pEGFP-p48, and pEGFP-p42 expression plasmid: Two sets of MH14 cells were grown in 6-well plates (2 × 10⁵/well) for 24 h, then transfected with siRNA (final concentration, 20 nM per well). For overexpression experiments, pEGFP-p48 or pEGFP-p42 (4 µg per well) was transfected according to the manufacturer’s protocol, using Lipofectamine 2000 (Invitrogen) as the transfection reagent. The transfected cells were grown for 48 h. One set of cells was processed for the isolation of total protein for Western blotting; the other set was processed for the isolation of total RNA (TRIzol; Invitrogen) for RT-PCR analysis of HCV RNA and actin mRNA. These experiments were done in triplicate.

Isolation of total RNA and quantitative real-time RT-PCR: We isolated total RNA from cells using TRIzol^® reagent (Invitrogen) according to the manufacturer’s protocol. One microgram of total RNA was used to synthesize cDNA corresponding to the mRNA of Ebp1, HCV 5’NTR, and GADPH by reverse transcription, as described earlier (Dixit et al., 2014). We used 50 ng of cDNA in Fast SYBR® Green Master Mix (Applied Biosystems) with primers directed to specific mRNA to do quantitative real-time PCR using the Fast Real PCR System (Applied Biosystems) as described (Krieger et al., 2001; Yokota et al., 2003) or by using ΔΔCT method (Livak and Schmittgen, 2001). For each point of QPCR data statistically significant p-value was obtained by using two-tailed unpaired t-test with Graph Pad software. Relative fold changes in specific mRNA copies were calculated by normalizing the amount of GADPH mRNA in each sample. All experiments were done in triplicate for each data point.

Immunoprecipitation. Washed MH14 cells (5 × 10⁶ cells per assay) were lysed in cold lysis buffer (50 mM Tris-HCl, pH 7.4; 150 mM NaCl; 1% Triton X-100) supplemented with 1x protease inhibitor cocktail (Roche Applied Science). The lysates were treated with 50 units of benzonase (Sigma) to avoid nonspecific RNA binding proteins that might be captured as a consequence of RNA bridging effect. We incubated the cell lysates with antibody against the target protein for 1 h at 4°C, then captured the IP-complex on protein A/G Plus agarose beads as described earlier (Zhang et al., 2008b). The immunoprecipitates released from the beads by boiling them in 1 × Laemmli gel loading buffer were subjected to SDS-PAGE, and Western blotted for the target proteins.

Construction of stably transduced Ebp1-knockdown MH14 cells. Following the manufacturer’s protocol, we generated stably transduced MH14 cells with lentivirus vectors encoding shRNA and targeting both isoforms of Ebp1 (sc-77220-V; Santa Cruz, CA). As a control, MH14 cells were also transduced with control lentiviral particles with empty vector (sc-108080, Santa Cruz, CA). Stable knockdown clones were chosen after several passages via puromycin selection in DMEM medium containing 3 µg/mL of puromycin (Santa Cruz, CA). Western blot analysis confirmed the knockdown of Ebp1 expression as compared to cells transformed with vector alone.

Transient expression of p42 and p48 isoform of Ebp1 in Ebp1-kd cells. Lentivirus-based Ebp1 shRNAs target two different regions of Ebp1 transcripts that are present in both isoforms. The first target of shRNA spans nucleotide positions +150–168 in the p48 transcript and −13 to +6 in the p42 transcript from the start codon. The second target is at position +825 to 843 in the p48 transcript and +662 to 681 in the p42 transcript from the start codon. We constructed shRNA-resistant Ebp1 expression clones (pEGFP-p42_SHR and pEGFP-p48_SHR) by point mutations in the degenerate codons without altering the amino acid sequence. Resistance to shRNA was confirmed by transient expression of Ebp1 isoforms in Ebp1-kd cells.

Downregulation of PKR in Ebp1-kd Huh7.5 cells. Two sets of Huh7.5 cells grown overnight in 6-well culture plates were transfected with PKR siRNA. Twenty-four hours later, cells were infected with HCV-JFH1 virions in the presence of 5-µg/ml polybrene (Dixit et al., 2016). At 6 h, cells after infection, cells were washed two times with PBS and supplemented with fresh medium. At 72h after infection, total RNA was isolated from one first set to quantify the level of HCV RNA by qRT-PCR; cell lysates were prepared from the other set to determine the expression levels of EBP1, PKR, and actin by Western blot.

Preparation of cell-free replication lysate: We prepared the replicative cytoplasmic fractions from MH14 cells following with minor modification, the protocol described previously (Ali et al., 2002; Waris et al., 2004; Zhang et al., 2008b). In brief, MH14 cells were grown in 10 cm Petri dishes, then washed with ice-cold buffer containing 150 mM sucrose, 30 mM HEPES (pH 7.4), 33 mM ammonium chloride, 7 mM KCl, and 4.5 mM magnesium acetate. The washed cells were first treated with lysolecithin solution (250 µg/ml) in the washing buffer for 1 min; this was followed by washing with 3 ml of wash buffer. The cells were scraped from the plate after the addition of 200 µl of replication buffer containing 100 mM HEPES (pH 7.4); 50 mM ammonium chloride; 7 mM KCl; 1 mM spermidine; 0.5 mM each of ATP, GTP, UTP, and CTP; 1 mM DTT; and 10% glycerol. The cells were gently lysed by pipetting up and down several times, then centrifuged at 1,600 rpm for 5 min at 4°C. The supernatant fraction (replicative lysate) was stored at −80°C until use.

Endogenous HCV replication assays in cell-free replication lysate. For this assay, an aliquot of normalized cell-free replication lysate (equivalent to 100 µg protein) containing 0.5 mM each of four rNTPs was incubated for one hour at 30°C. The reaction was terminated by adding stop buffer containing 0.5% SDS in 10 mM Tris-HCl, pH 7.5; 1 mM EDTA; and 150 mM NaCl. After extracting the reaction mixture twice with phenol-chloroform-isoamyl alcohol (25:24:1) and twice with water-saturated ether, we precipitated the RNA and analyzed it by RT-PCR for HCV RNA and actin mRNA.

RNA-dependent RNA polymerase (RdRp) assay. The RdRp assay was done in a final volume of 100µl. The polymerase reaction in a final volume of 100µl contained 20 mM Tris HCl (pH 7.8), 40 mM NaCl, 40 mM sodium glutamate, 0.5 mM DTT, 0.01% BSA, 0.01% tween-20, 5% glycerol, 20 U/ml RNase Out, 50 µM cold UTP, 1 µCi ³H-UTP per assay, 500 nM PolyrA/dT₁₈, 0.5 mM MnCl₂, and 200 nM of NS5B. After 30 min of incubation at 37°C, reactions were terminated by the addition of 5% ice-cold TCA. The acid-precipitable nucleic acid material was filtered on glass fiber filters (GF/B), washed successively with 5% TCA, water, and ethanol. Filters were air dried and placed in a vial containing 5 ml EcoLite scintillation fluid, then counted for radioactivity using a Packard 2200-CA Tri-Carb scintillation counter.

Expression and purification of recombinant PKR: A recombinant clone of pPET-PKR/λPP was expressed in E. coli Rosetta (DE3) and purified by Hi-Trap heparin columns (Pharmacia). In brief, transformed E. coli Rosetta (DE3) cells were grown at 37°C in Luria broth (LB) containing 100 µg/ml of ampicillin until an OD₅₉₅ of 0.4 was achieved. The medium was then cooled to 25°C, supplemented with 1 mM IPTG, and incubated at 25°C for 16 h with vigorous shaking. The cells were harvested, washed, and resuspended in lysis buffer (20 mM Tris-HCl, pH 7.4; 200 mM NaCl; 1 mM β-mercaptoethanol; 10% glycerol; and 1% triton-X 100) containing 1× ProteoBlock protease inhibitor cocktail (Fermentas), and 2 mg/ml lysozyme. The suspension was sonicated and centrifuged. The clear supernatant was applied to an Hi-Trap Heparin FPLC column preequilibrated with binding buffer (20 mM Tris-HCl, pH 7.4; 200 mM NaCl; 10% glycerol). After washing the column extensively with binding buffer, PKR was eluted with a linear 0%–80% of 1 M KCl in the same buffer for 20 min (1 ml/min). Eluted fractions showing more than 95% purity on SDS-PAGE (8%) were pooled and dialyzed against a buffer containing 50 mM Tris-HCl (pH 7.5), 2 mM DTT, 100 mM NaCl, and 50% glycerol.

Expression and purification of recombinant Ebp1 isoforms: Recombinant clones of His-tagged Ebp1 isoforms (pET28a-Ebp1p42 and pET28a-Ebp1p48) were expressed in E. coli Rosetta (DE3) and purified by affinity chromatography using Ni-NTA and Hi-Trap heparin columns (Pharmacia). In brief, transformed E. coli Rosetta (DE3) cells were grown at 37°C in Luria broth (LB) containing 30 µg/ml of kanamycin until an OD₅₉₅ of 0.4 was achieved. The medium was cooled to 26°C, supplemented with 1.0 mM IPTG, and incubated at 26°C for 16 h with vigorous shaking. The cells were harvested, washed, and resuspended in a lysis buffer containing 20 mM Tris-HCl, pH 7.4; 200 mM NaCl; 1 mM β-mercaptoethanol; 10% glycerol; 1% triton-X 100; 5 mM imidazole; and 1× ProteoBlock protease inhibitor cocktail (Fermentas) containing 2 mg/ml lysozyme. The suspension was sonicated and centrifuged, after which the clear supernatant was applied to a Ni-NTA column pre-equilibrated with binding buffer (20 mM Tris-HCl, pH 7.4; 200 mM NaCl; 10% glycerol; and 5 mM imidazole). After extensively washing the column with the binding buffer, the column was then successively washed with 20 mM and 50 mM imidazole in the binding buffer. His-tagged Ebp1 isoform was then eluted with 200 mM imidazole in the same buffer. Fractions showing greater than 90% purity on SDS-PAGE were pooled, then diluted 2-fold by adding an equal volume of a buffer containing 20mM Tris-HCl (pH 7.5), 5% glycerol, 0.5% NP-40, and one mM β-mercaptoethanol. The diluted fractions were then applied to the Hi-Trap heparin column. The column was extensively washed. Ebp1 isoform was eluted from it with a linear gradient (0% to 80%) of 1 M KCl in the same buffer for 20 min (1 ml/min). Eluted fractions showing more than 95% purity on SDS-PAGE (8%) were pooled and dialyzed against buffer containing 50 mM Tris-HCl (pH 7.5), 2 mM DTT, 100 mM NaCl, and 50% glycerol.

Photoaffinity Crosslinking of 5’-NTR with Ebp1 isoforms. We incubated 500 ng of purified recombinant Ebp1-p42 or Ebp1-p48 isoform with internally Cy5-labeled HCV 5’-NTR. The incubation buffer contained 50mM Tris HCl (pH7.8), 1 mM DTT, 1 mM MgCl₂, and 0.01% bovine serum albumin in a final volume of 30 µl. After 20 min of incubation on ice, we UV irradiated the mixture at 360 mJ/cm² in a Spectrolinker XL-1000 (Spectronics). We treated the irradiated samples with RNase A (0.1 µg/µl; Qiagen, Valencia CA) for 15 min at 37°C and then resolved the crosslinked RNA-protein complexes by SDS-PAGE. The labeled RNA-protein complex was visualized using a Typhoon scanner (Amersham).

PKR autophosphorylation The PKR autophosphorylation assay was done under standard conditions as described earlier (Dey et al., 2005; Vyas et al., 2003). The reactions in a final volume of 20 µl contained 20mM Tris HCl (pH7.5), 5mM MgCl₂, 5 mM MnCl₂, 100mM KCl, 0.1 mM ATP, 0.1 mM EDTA, 0.1 µg poly (I).poly (C), 30% (v/v) glycerol, and variable or fixed concentrations of unphosphorylated recombinant PKR. The reaction, done for 30 min at 30°C, was terminated by the addition of 2× SDS gel loading buffer. The phosphorylated products were resolved by SDS PAGE and Western blotted with an antibody against phospho serine-tyrosine.

Luciferase assay. MH14 cells in 6-well plates were transfected with Ebp1 siRNA (final concentration, 20 nM/well), p48 siRNA, pEGFPp48, or pEGFP-p42 overexpressing plasmid (4 µg per well), using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s protocol. At 24 h posttransfection, cells were transfected with HCV replicon RNA carrying luciferase reporter (0.5 µg/well) and 100 ng of pGEM-REN-HCV IRES-Luc, a reporter plasmid expressing firefly luciferase under the control of HCV IRES and renilla luciferase under the control SV40 promoter. Cells were grown for another 48 h, lysed, and assayed for luciferase using a dual luciferase reporter assay kit (Promega). Assays were done in four parallel sets.

Immunofluorescence and co-localization: Cells (1 × 10⁵ cells/well) were grown on BD Falcon 8-chamber tissue-culture slides at 37°C in a 5% CO₂ atmosphere for 24 h. Cultures were washed, fixed in 4% paraformaldehyde, permeabilized with 0.1% Triton, and stained with first primary monoclonal antibody as described earlier (Zhang et al., 2008b). Cells were washed with PBS containing 0.3% Triton-X 100, then incubated with anti-mouse IgG antibody conjugated with Alexa Fluor 568 (Invitrogen) at a 1:500 dilution. After 1 h of incubation at room temperature, cells were washed, treated with rabbit monoclonal antibody against another target protein (Santa Cruz Biotechnology), and incubated in the same chambers for 1 h at room temperature with anti-rabbit IgG antibody conjugated with Alexa Fluor488 (Invitrogen). Cells were then washed and stained with DAPI (4’, 6’-Diamidino-2-phenyindole; Sigma Chemical) for 10 min, washed four times with PBS, and air-dried. The slides were mounted with mounting medium (ProLong® Kit; Molecular Probes, OR) and visualized under a Multiphoton Confocal Microscope System (Nikon A1R). Images were processed using NIS software. To determine the extent of co-localization, we used Nikon NIS viewer analysis software to calculate the overlap coefficient between 0 to 1 range (0 is minimum, and 1 is maximum co-localization).

RESULTS

HCV replication and translation are negatively affected by selective downregulation of p48 isoform of Ebp1 but significantly enhanced by downregulation of total Ebp1. Two differently spliced transcripts of Ebp1 mRNA, 2.2 kb, and 1.7 kb have been detected in humans (Liu et al., 2006). The spliced transcript Ebp1-p42 isoform contains a deletion of 29 nucleotides at positions 135–164 and translated from 3rd ATG codon starting at 237 position (Fig 1A). The two isoforms regulate cell survival and differentiation differently. The longer isoform, p48, induces cell survival; the shorter p42, inhibits cell growth (Liu et al., 2006). We found that higher levels of p42 isoform than p48 isoform were expressed in the cytoplasm, although, in HCV infected cells, the two isoforms were equally distributed in both the cytoplasm and nuclei (Fig. 1B).

Figure 1.

Ebp1 isoforms have opposite effects on HCV replication and translation. (A) The N-terminal sequence of transcripts of p42 and p48 showing identical sequence except for the 29 nucleotides missing from the p42 transcripts. The smaller isoform skips 29 nucleotides and starts from the 3rd ATG, beginning at position 237 (Liu et al., 2006). (B) Ebp1-p42 is predominantly localized in the cytoplasm, while p48 is localized in the nuclei of cells. Nuclear (lane 1) and cytoplasmic (lane 2) fractions of MH14 cells were Western blotted for Ebp1. These fractions were also blotted for PARP (lane 3) and tubulin (lane 4), which, respectively, are nuclear and cytoplasmic markers. (C) The effect of siRNA-mediated selective downregulation of p48 isoforms and total Ebp1 on HCV replication and translation. MH14 cells carrying HCV replicons were transfected with siRNA targeting total Ebp1 transcript (lane 5) or selectively targeting the p48 isoform at nucleotide positions 132–150 and 148–166 (lane 4). Transfected cells were grown for 48h, harvested, lysed, and Western blotted for Ebp1, HCV NS5A, and actin. Total RNA was isolated from a parallel set of cells and RT-PCR was done on HCV RNA and GAPDH mRNA. Lane 1, untransfected control MH14 cells; lane 2, cells transfected with reagents only; lane 3, cells transfected with control siRNA; lane 4, cells transfected with Ebp1p48-siRNA; lane 5, cells transfected with Ebp1 siRNA targeting both isoforms. (D) Effect of downregulation of total Ebp1 or p48 isoform on luciferase reporter activity. Cured MH 14 cells devoid of HCV replicons were transfected with siRNA targeting either total Ebp1 transcript (lane 3) or p48 isoform at nucleotide positions 132–150 and 148–166 (lane 4). At 24 h posttransfection, cells were transfected with a bicistronic reporter plasmid, pGEM-REN-HCV IRES-Luc or empty vector alone. After 48 h, cells lysates were assayed for reporter activity. Lane 1, control; lane 2, vector control; lane 3, downregulation of total Ebp1; lane 4, downregulation of the p48-isoform of Ebp1.

We have used MH14 cells (derived from Huh7) carrying replicating HCV subgenomic replicons to downregulate total Ebp1 expression by transfecting with Ebp1-siRNA. The p42 isoform cannot be selectively downregulated by siRNA because both p42 and p48 isoform contain identical transcripts except that 29 nucleotides (from 135–164) that are present in p48 transcripts are missing in the spliced transcript of p42. Therefore, we selectively downregulated p48 isoform by targeting the 29 nucleotides in it that are missing from p42. We used two sets of siRNAs to target the p48 transcript at the nucleotide position from 132–150 and 148–166. We transfected siRNAs into MH14 cells carrying stable replicating HCV subgenomic replicons. After 48 h of transfection, we examined the cell lysate for expression of Ebp1 and viral proteins by Western blotting, and for HCV RNA by RT-PCR. We found that selective downregulation of p48 isoform reduced the level of both HCV RNA and viral protein NS5A (Fig.1C, lane 4) while downregulation of total Ebp1 significantly enhanced both viral replication and translation of viral protein (Fig1C, lane 5).

These results suggest that HCV replication and translation are promoted by p48 isoform but significantly inhibited by p42-isoform. Since siRNA against the longer isoform of Ebp1 may specifically target the p48 mRNA, it is likely that it may also target pre-spliced Ebp1 transcript, thus reducing the level of the p42 transcript. This may explain the low level of expression of Ebp1–42 isoform in cells transfected with siRNA against p48 isoform. However, in the absence of p48-isoform, the residual expression of p42 in cells treated with p48-siRNA may have an inhibitory effect on HCV replication and translation. An approximately 1.5-fold increase in the expression of viral protein NS5A was observed in total Ebp1-downregulated cells as compared to controls, while the HCV RNA level was also increased 2-fold in Ebp1-siRNA treated cells (Fig. 1C, lane 5, right panel). Since siRNA against the p48 isoform of Ebp1 may specifically target the longer mRNA transcript, the residual expression of p42 isoform may inhibit HCV replication and translation. Alternatively, p48 isoform may have a positive effect on viral replication and translation.

We further examined whether downregulation of either total Ebp1 or p48 isoform influences translation directed by HCV IRES. We first downregulated the p48 isoform of total Ebp1 by siRNA. We then transfected it with bicistronic reporter plasmid pGEM-REN-HCV IRES-Luc expressing firefly luciferase under the control of HCV IRES and renilla luciferase under the control SV40 promoter. When we determined relative luciferase activity at 48h after transfection, there was a 2-fold increase in the reporter activity when total Ebp1 was downregulated (Fig. 1D, lane 3). In contrast, an approximately, a 2-fold decrease in reporter activity was observed in cells selectively downregulated for p48 expression (Fig. 1 D, lane 4).

Ebp1-IP co-immunoprecipitates HCV-NS5A and HCV NS5B. Since Ebp1 seems to inhibit HCV replication, we examined whether it interacts with HCV proteins. When we used Ebp1-IP on RNase A-treated cell lysates from MH14 cells and Western blotted for NS5A, NS5B and NS3, only NS5A and NS5B co-precipitated in the Ebp1-IP fraction (Fig 2A, lane 3). This suggests that the interaction of viral proteins NS5A and NS5B with Ebp1 may modulate the inhibitory effect of Ebp1 on HCV replication or translation. Using recombinant p42 and p48 isoforms of Ebp1-purified proteins and viral proteins, we confirmed that both isoforms of Ebp1 directly interact with NS5B and NS5A. Ebp1-IP, on a mixture of Ebp1p42+NS5A and Ebp1p42+NS5B, coprecipitated NS5A and NS5B (Fig. 2B). Similarly, Ebp1-IP on p48 isoform coprecipitated both NS5A and NS5B (Fig. 2C).

Figure 2.

Ebp1 physically interacts with NS5A, NS5B, and PKR. (A)Ebp1-IP co-immunoprecipitates NS5A and NS5B. Ebp-1-IP was done on cell lysates from MH14 cells using Ebp1 antibody. The IP complex was captured on protein A/G agarose beads, extensively washed, and resolved by SDS-PAGE as well as Western blotted for Ebp1, NS5A, NS5B, and NS3. Lane1, IgG control; lane 2, beads control; lane 3, Ebp1-IP; lane 4, cell lysates. (B, C) Reciprocal IP on a mixture of recombinant Ebp1 isoforms and HCV proteins. The purified recombinant proteins of Ebp1-p42 or p48 isoforms were incubated with NS5A or NS5B, then subjected to reciprocal IP and immunoblotted for the targeted protein partner. Lane 1, IgG control; lane 2, beads control, lane 3, IP-sample; lane 4, protein control. (D) Reciprocal Ebp1-IP and PKR-IP on MH14 cell lysates coimmunoprecipitate each other. Lane 1, IgG control; lane 2, beads control; lane 3, IP-sample; lane 4, cell lysate. (E, F) Reciprocal IP on a mixture of recombinant Ebp1 isoforms PKR. The purified recombinant proteins of Ebp1-p42 or Ebp1-p48 isoforms were incubated with PKR, then subjected to reciprocal IP and immunoblotted for the targeted protein partner. Lane 1, IgG control; lane 2, beads control, lane 3, IP-sample; lane 4, protein control.

Ebp1 isoforms physically interact with PKR. In the cytoplasm, Ebp1, which is associated with ribosomes, inhibits the phosphorylation of eukaryotic initiation factor 2 alpha (eIF2α) (Squatrito et al., 2006). We, therefore, examined whether Ebp1 isoforms interact with PKR. We did reciprocal Ebp1-IP and PKR-IP on RNase A-treated cell lysate from MH14 cells; then Western blotted for PKR and Ebp1. The results (Fig. 2D) indicate that Ebp1 and PKR physically interact with each other. To confirm this, we did IP on purified recombinant PKR and Ebp1 isoforms p42 and p48. We found that Ebp1-IP on a mixture of purified Ebp1p42 + PKR effectively coprecipitated PKR (Fig. 2E, lane 3). Similarly, Ebp1-IP on a mixture of Ebp1-p48 + PKR coimmunoprecipitated PKR (Fig. 2F, lane 3). These results confirmed that both isoforms of Ebp1 physically interact with PKR.

Ebp1 isoforms have opposite effects on activation and autophosphorylation of PKR. Since both isoforms of Ebp1 interact with PKR, we examined whether Ebp1 isoforms have any effect on the activation and autophosphorylation of PKR. Autophosphorylation of PKR is an essential step for its activation and dimerization. We incubated increasing concentrations of purified PKR (100–400 ng) under standard autophosphorylation assay conditions (Dey et al., 2005; Vyas et al., 2003) and detected autophosphorylation by Western blotting using an antibody against phospho-serine-threonine. We found that PKR is moderately autophosphorylated in these conditions (Fig. 3A, lanes 1–4). We then used 0.4µg of PKR (equivalent to lane 4) and incubated it in the autophosphorylation assay mix with increasing concentrations of the p42 or p48 isoform of Ebp1. We found that autophosphorylation of PKR was significantly enhanced by Ebp1-p42 (Fig 3B, lanes 2–4), but strongly inhibited by the Ebp1-p48 isoform (Fig. 3C, lanes 2–4). These results indicate that the p42 isoform of Ebp1 positively regulates PKR activation, which may explain its inhibitory effect on HCV replication. In contrast, the p48 isoform negatively regulates PKR activation, which may facilitate viral replication and translation.

Figure 3.

PKR autophosphorylation is promoted by p42 isoform but inhibited by p48 isoform. (A) Autophosphorylation of PKR as a function of increasing PKR concentration. Lanes 1–4, increasing concentrations of PKR (100–400 ng) were incubated in autophosphorylation assay mix. The reaction was done at room temperature for 30 min. An aliquot of phosphorylated PKR was resolved on SDS-PAGE, and Western blotted with an antibody against phospho-serine-threonine. (B) Ebp1-p42 isoform promotes PKR autophosphorylation. A fixed concentration of PKR (400 ng) was incubated in autophosphorylation assay mix in the presence of increasing concentrations of Ebp1-p42 (lanes 2–4). (C) Ebp1-p48 isoform inhibits PKR autophosphorylation. A fixed concentration of PKR (400 ng) was incubated in autophosphorylation assay mix in the presence of increasing concentrations of Ebp-p48 (lanes 2–4). (D) Ebp1-p48 and NS5A together have a synergistic effect on PKR autophosphorylation, while Ebp1-p42 antagonizes the effect of NS5A. We incubated 100 ng of purified recombinant PKR in autophosphorylation assay mix in the absence (lane1) or presence of 200 ng of NS5A (lane 2), Ebp1-p48 (lane 3), Ebp1-p42 ( lane 5), or both NS5A+p48 (lane 4) and NS5A+p42 (lane 6). Reactions were done for 30 min at room temperature. Phosphorylated PKR was immunoblotted with antibody against phospho-serine-threonine. Bottom panel shows the quantitation of Western blots included in the upper panel.

The NS5A-mediated inhibitory effect on PKR activation is blocked by Ebp1-p42 but synergistically promoted by Ebp1-p48. HCV NS5A is known to have an inhibitory effect on the activation and autophosphorylation of PKR. Since both Ebp1 isoforms interact with NS5A and have opposite effects on PKR activation, we examined their effect on PKR autophosphorylation in the absence and presence of NS5A. We carried out autophosphorylation of PKR in the presence and absence (control) of NS5A, Ebp1-p48, or Ebp1-p42. We also did autophosphorylation of PKR in the presence of both NS5A and the individual Ebp1 isoforms. The level of phosphorylation of PKR was then determined by Western blotting, using an antibody against phospho-serine-threonine. The results, shown in Figure 3D, indicated that in the presence of either NS5A alone (lane 2) or Ebp1-p48 alone (lane 3), autophosphorylation of PKR was moderately inhibited, whereas 75% inhibition was observed in the presence of both NS5A and the p48 isoform of Ebp1 (lane 4). This demonstrated the synergistic inhibitory effect of NS5A and p48 isoform on the inhibition of PKR activation. In contrast, the Ebp1-p42 isoform promoted PKR activation (lane 5) and antagonized the inhibitory effect of NS5A (lane 6).

The p42 isoform strongly binds to HCV −5’NTR, whereas p48 binds to it poorly. Many viruses have developed mechanisms to inhibit PKR to prevent the inhibition of cellular protein synthesis, which would be detrimental to their replication (Katze, 1995). HCV-IRES and NS5A both circumvent this inhibition by directly interacting with PKR and inhibiting its autophosphorylation and activation (Gale et al., 1998; Vyas et al., 2003). The p48 isoform, which is phosphorylated by PKR, protects eIF-2a from phosphorylation (Squatrito et al., 2006). We found that the p42 isoform promotes activation of PKR, while the p48 isoform has an opposite effect on the activation of PKR. We examined whether Ebp1 isoforms have any affinity for HCV IRES, which also interacts with PKR and inhibits its activation. We incubated in-vitro transcribed Cy5-labeled full-length HCV 5’NTR (Fig. 4A) with increasing concentration of His-tagged Ebp1-p42 or Ebp1-p48 isoform. We UV-irradiated the mixture, then treated with RNase A and resolved by SDS-PAGE. We found that the p42 isoform efficiently binds to 5’NTR RNA (Fig. 4B, lanes 1–3), while the p48 isoform binds 5’NTR less efficiently (lanes 4–7). These results suggest that selective interaction of HCV IRES with p42 isoform may be a viral strategy to facilitate its replication by attenuating p42-mediated activation of PKR.

Figure 4.

Ebp1-p42 efficiently binds HCV IRES, whereas p48 isoform binds poorly it. (A) Predicted secondary and tertiary structure of 5’NTR containing HCV IRES (Honda et al., 1999). (B) Binding of Ebp1 isoforms to full-length 5’NTR. Internally Cy5- labeled in-vitro that transcribed full-length 5’NTR RNA was incubated with increasing concentrations (200 ng to 1 µg) of His-tagged p42 or p48 isoforms of Ebp1. The complex was photo-crosslinked by UV exposure. The crosslinked complex was treated with RNase A and resolved by SDS-PAGE. Lanes 1–3, Cy5-5’NTR crosslinked to His-tagged p42 isoform; lanes 4–6, Cy5-5’NTR crosslinked to His-tagged p48 isoform. Lane 7 is the negative control without Ebp1 isoform.

Effect of overexpression of Ebp1 isoforms on HCV replication and translation. We have shown that siRNA-mediated downregulation of total Ebp1 enhances HCV replication and translation, whereas selective downregulation of the p48 isoform of Ebp1 reduces both viral replication and translation. Thus, the p42 and p48 isoforms of Ebp1 have opposite effects on HCV replication.

The next obvious issue was to examine the effect of overexpression of each isoform on viral replication. We overexpressed GFP-tagged Ebp1 isoforms by transfecting MH14 cells with pEGFP-p48 or pEGFP-p42, then examined the expression level of GFP-tagged Ebp1 isoform in the cell lysate, using an antibody against GFP (Fig. 5A). While cells transfected with vector alone expressed GFP, which migrated as 32kDa protein, GFP-tagged p42 and p48 Ebp1 isoforms migrated, respectively, as 72kDa and 78kDa proteins. We then examined the expression of viral protein NS5A and actin by Western blotting, as well as the HCV RNA level by RT-PCR (Fig. 5B) and quantitative RT- PCR (Fig. 5C). We found that , as compared to untransfected control and vector control, overexpression of the Ebp1-p42 isoform significantly reduced the HCV RNA level and expression level of viral protein NS5A (Fig. 5B, lane 4). A two-fold increase in the level of NS5A and HCR RNA was observed when we overexpressed the longer isoform p48 (Fig. 5B, lane 3). Quantitative RT-PCR confirmed the two-fold increase in HCV RNA level in cells overexpressing the p48 isoform (Fig. 5C, lane 3) and the twofold decrease that occurred with overexpression of the p42 isoform (Fig. 5C, lane 4). These results further demonstrate that p42 and p48 isoforms of Ebp1 are, respectively, negative and positive regulators of HCV. To establish persistent replication in the cell, HCV may use its protein NS5A and the p48 isoform to overcome the negative effect of p42 by synergistically inhibiting PKR activation in HCV-infected cells.

Figure 5.

Effect of overexpression of Ebp1 isoforms and knockdown of total Ebp1 on HCV replication and translation. Overexpression of GFP-tagged Ebp1 isoforms in MH14 cells. MH14 cells were transfected with pEGFP-p42, pEGFP-p48, or empty vector. At 48 h posttransfection, cells were analyzed by immunoblotting with anti-GFP antibody to confirm the expression of GFP-p42 and GFP-p48, which migrate, respectively, as 72kDa and 78kDa fusion protein. (B, C) Overexpression of Ebp1-p42 isoform inhibits HCV replication; Ebp1-p48 overexpression promotes viral replication. Cell lysates containing overexpressed GFP-tagged p42 or p48 isoforms of Ebp1 were immunoblotted for NS5A and actin. From a parallel set, total RNA was isolated and analyzed for HCV RNA and actin mRNA by RT-PCR and (C) quantitated by real-time qPCR. Lane 1, MH14 control; lane 2, MH14 cells transfected with empty vector only; lane 3, MH 14 cells transfected with pEGFP-p48; lane 4, MH14 cells transfected with pEGFP-p42 overexpression plasmid. (D) Effect of overexpression of p48 and p42 isoform of Ebp1 on HCV translation. Cured MH14 cells devoid of HCV subgenomic replicons were transfected with overexpression plasmid, pEGFP-p42, or pEGFP-p48. At 24h, cells were transfected with bicistronic reporter plasmid pGEM-REN-HCV IRES-Luc or empty vector alone. After 48, cell lysates were assayed for reporters activity. Lane 1, control; Lane 2, vector control; lane 3, overexpression of p48; lane 4, overexpression of p42. (E) Stable knockdown of Ebp1 in MH14 cells by shRNA and its effect on the expression of HCV proteins: MH14 cells were transfected with a lentivirus vector encoding Ebp1 targeting shRNA or with empty vector alone. The stable clone was selected after several passages via puromycin and confirmed by Western blot analysis for stable knockdown of Ebp1 expression (lane 2) as compared to cells transformed with vector alone (lane 1). We also Western blotted for expression of NS5A and NS5B in control and Ebp1-kd cells. The right panel shows quantitation of the Western blot bands of Ebp1, NS5A, NS5B, and actin in control and Ebp1-kd MH14 cells.

We next examined the effect of overexpression of p42 and p48 isoform on reporter activity under the control of HCV IRES. Cured MH14 cells overexpressing p42 or p48 isoform were transfected with the bicistronic vector pGEM-REN-HCV IRES-Luc or empty vector alone. At 48 h posttransfection, cell lysates were assayed for reporter activity (Fig. 5D). We found a 2-fold increase in relative luciferase reporter activity when the p48 isoform was overexpressed (Fig. 5D, lane 3) while a 3-fold decrease in the reporter activity in cells overexpressing p42-isoform (Fig. 5D, lane 4). These results confirm that p42 and p48 isoforms have opposite effects on HCV translation.

Effect of transient expression of p42 and p48 isoform on HCV replication in Ebp1-kd cells. We generated stable Ebp1 knockdown (Ebp1-kd) MH14 cells using lentivirus shRNA targeting Ebp1 transcripts (Santa Cruz, CA). After several passages, stable clones were selected by puromycin selection. We also generated shRNA control cells by transducing with an empty lentiviral vector alone. Stable knockdown was confirmed using Western blot for Ebp1 (Fig. 5E, lane 2, left panel). We also Western blotted HCV NS5A and NS5B in the cell lysate and found significantly increased expression of the viral protein in Ebp1-kd cells (Fig. 5E, lane 2, left panel). Quantitation of Western blot images confirmed a 3-fold increase in the expression of the viral proteins, NS5A and NS5B in Ebp1-kd cells (Fig. 5E, right panel).

For transient expression of either the p42 or p48 isoform in Ebp1 knockdown cells, we constructed shRNA-resistant expression clones of p42 and p48 isoforms by point mutations in the Ebp1 encoding region, targeted with shRNA without altering the amino acid sequence (Fig. 6A). The lentivirus-based Ebp1 shRNAs target two different regions of Ebp1 transcripts that are present in both of the isoforms. The first target of shRNA spans nucleotide position +150–168 in the p48 transcript and −13 to +6 from their start codon in the p42 transcript. The second target is at position +825 to 843 in the p48 transcript and +662 to 681 from their start codon in the p42 transcript. We carried out point mutations in these targets in p48 and p42 expression plasmids without altering codon use. We transfected shRNA-resistant expression clones of p42 and p48 isoforms of Ebp1 separately or with an empty vector in Ebp1-kd MH14 cells. After 48 h of posttransfection, we Western blotted cell lysates for the expression level of transiently expressed GFP-tagged Ebp1 isoforms by anti-GFP antibody (Fig. 6B) or anti-Ebp1 antibody, actin, and viral protein NS5A (Fig. 6C). We determined levels of HCV RNA by RT-PCR (Fig. 6C, left panel). Using quantitative RT-PCR on total RNA, we also quantitated the relative fold-change in HCV RNA levels with respect to actin mRNA (Fig. 6C, right panel).

Figure 6.

Transient expression of Ebp1-p42 in Ebp1-kd MH14 cells reverses enhanced HCV replication and restores HCV to control level; Ebp1-p48 isoform promotes HCV replication. (A) Construction of shRNA-resistant expression clones of Ebp1 isoforms. Ebp1-shRNA targets two different regions of Ebp1 transcripts that are present in both p48 and p42 isoforms. The first target spans nucleotide positions +150–168 in the p48 transcript and −13 to +6 in the p42 transcript from their start codons. The second target is at position +825 to 843 in the p48 transcript and +662 to 681 in the p42 transcript from their start codons. We introduced two point mutations in each target without altering codon use. (B) Transient expression of shRNA-resistant GFP-tagged Ebp1 isoforms in Ebp1-kd MH14 cells. shRNA-resistant pEGFP-p42_SHR and pEGFP-p48_SHR expression clones were transfected into Ebp1-kd MH14 cells; 48 h later, cells lysates were analyzed for expression of GFP-p48, GFP-42, or vector alone by immunoblotting with the anti-GFP antibody. (C) Effect of transient expression of Ebp1 isoforms on replication of HCV replicons in MH14. Cell lysates containing transiently expressed GFP-tagged p42 or p48 isoforms in Ebp1-kd MH14 cells were immunoblotted for Ebp1, NS5A and actin. From a parallel set, total RNA was isolated and analyzed for HCV RNA and actin mRNA by RT-PCR and quantitated by real-time qPCR (right panel). Lane 1, MH14 control; Lane 2, untransfected Ebp1-kd MH14 cells; lanes 3–5, Ebp1-kd MH14 cells transfected, respectively, with pEGFP-p42_SHR, pEGFP-p48_SHR and empty vector. (D) Effect of transient expression of Ebp1 isoforms on replication of JFH1 HCV in HepG2 cells. Ebp1-kd HepG2 cells were cotransfected with plasmids expressing human CD81 (pTRIP-GFP-hCD81) and miR-122 (pTRIP-Puro-miR122) with or without shRNA resistant expression clone of Ebp1-isoform. At 24 h posttransfection, cells were infected with HCV-JFH1 virions (Dixit et al., 2015). At 72h, total RNA was isolated to determine the level of HCV RNA, which was quantitated by q-RT-PCR; cell lysates were Western blotted for Ebp1 and actin. Lane 1, HepG2/CD81/miR122 control; Lane 2, Ebp1-kd HepG2/CD81/miR122 cells control; lanes 3–5, Ebp1-kd HepG2/CD81/miR122 cells transfected, respectively, with pEGFP-p42_SHR, pEGFP-p48_SHR and empty vector.

We found increased HCV replication and translation in Ebp1-kd control cells (Fig. 6C, lane 2). In Ebp1-kd cells, transient expression of the p42 isoform significantly suppressed both HCV replication and translation (Fig. 6C, lane 3) as compared to Ebp1-kd control cells (Fig. 6C, lane 2) or control cells transfected with empty vector alone (Fig. 6C, lane 5). In contrast, transient expression of the p48 isoform in Ebp1-kd cells had an effect opposite to that of p42, moderately enhancing HCV replication (Fig. 6C, lane 4) as compared to that in Ebp1-kd cells (Fig. 6C, lane 2). There was a 2-fold increase in HCV RNA level (Fig. 6C, lane 2) as compared MH14 controls (Fig. 6C. lane 1). These results indicate that p42 isoform of Ebp1 inhibits HCV replication and translation, whereas the p48 form promotes them. The experiments shown in Figure 6C were also performed with HepG2/CD81/miR122 cells infected with JFH1 HCV virions (Fig. 6D). We found that, as compared to Ebp1-kd cell (Fig 6D, lane 2, right panel), a 4-fold reduction in HCV replication was observed upon transient expression of p42 isoform (Fig. 6D, right panel, lane 3) while transient expression of p48 isoform resulted in an approximately 1.5- to 2-fold increase in the level of HCV RNA (Fig. 6D, right panel, lane 4).

Downregulation of PKR in Ebp1-kd cells transiently expressing the p42 or p48 isoform of Ebp1. Ebp1-kd Huh7.5 cells were cotransfected with PKR-siRNA and shRNA-resistant expression plasmid of Ebp1-p42 or Ebp1-p48 isoform. At 24 h posttransfection, cells were infected with JFH1 HCV virions. After 72 h, levels of HCV RNA in the cells were determined by qRT-PCR, while expression levels of PKR, Ebp1, and actin were established by Western blotting. We found a 2-fold increase in HCV replication in Ebp1-Kd cells (Fig. 7, lane 2) as compared to control Huh7.5 cell (lane 1). Downregulation of PKR in Ebp1-kd cells further increased HCV replication (lane 3) as compared to the level in control (lane 1) and Ebp1-kd cells (lane 2). We also found that transient expression of p42 in Ebp1-kd cells reduced HCV RNA levels 4-fold (lane 4) as compared to that in Ebp1-kd control (lane 2). This inhibition was reversed by downregulation of PKR (lane 5). In contrast, transient expression of the Ebp1-p48 isoform in Ebp1-kd cells resulted 2.5 fold increase in HCV RNA levels (lane 6) as compared to control (lane 1). However, downregulation of PKR in Ebp1-kd cells transiently expressing the p48 isoform displayed no significant increases in the levels of HCV RNA (lane 7). Thus, the p42 isoform may inhibit HCV replication via activation of PKR, while the p48 isoform have opposite effect on HCV replication by inhibiting PKR.

Figure 7.

Effect of downregulation of PKR on JFH1 HCV replication in Ebp1-kd Huh7.5 cells transiently expressing the p42 or p48 isoform of Ebp1. Ebp1-kd Huh7.5 cells were transfected with PKR-siRNA with or without shRNA-resistant expression plasmid of Ebp1-p42 or Ebp1-p48 isoforms. At 24 h posttransfection, cells were infected with JFH1 HCV virions. At 72 h, the levels of HCV RNA in cells were determined by qRT-PCR. Expression levels of PKR, Ebp1, and actin were determined by Western blotting. Lane 1, Huh7.5 control; lane 2,Ebp1-kd control; lanes 3 to 8 respresents, respectively, Fbp1-kd cells transfected with PKR-siRNA alone (lane 3), transfected with transient expression clone of p42 alone (lane 4), cotransfected with PKR-siRNA with p42 expression plasmid (lane 5), transfected with transiently expression clone of p48 (lane 6), cotransfected with PKR-siRNA with p48 expression plasmid (lane 7), and empty vector alone (lane 8).

Endogenous HCV replication in cell-free replication lysate is inhibited by exogenously added Ebp1-p42 but enhanced by the Ebp1-p48 isoform. Since both isoforms of Ebp1 co-immunoprecipitate viral proteins NS5B and NS5A, which are essential components of the HCV replication complex, we examined whether the addition of purified Ebp1 isoforms affects in-vitro endogenous HCV replication in cell-free replication lysate. We prepared cell-free replication lysate from MH14 cells (Ali et al., 2002; Zhang et al., 2008b) and applied it to a cell-free HCV replication assay in the absence or presence of either exogenous Ebp1-p42 (Fig. 8A) or Ebp1-p48 isoform (Fig. 8B). After 3 h of incubation, we isolated total RNA and examined the level of HCV RNA by RT-PCR and qPCR, which indicated the activity of endogenous HCV replicative complexes in the replicative lysate. After incubation, as a consequence of newly synthesized HCV RNA, the level of HCV RNA in controls was approximately 2-fold higher (Fig. 8A, B, lane 2) than the basal level at zero time (Fig. 8A, B, lane 1). In the presence of exogenous Ebp1 isoforms, the level of newly synthesized HCV RNA was 2.5-fold lower, at 0.25 pmol of Ebp1-p42 (Fig. 8A, lane 3). At 0.5 pmol and 1 pmol, this level further decreased to 3.5 and 4-fold, respectively (Fig. 8A, lanes 4 and 5), then to the basal level of replication activity. In contrast to the inhibition observed with Ebp1-p42, we found that HCV replication in cell-free replication lysate was moderately enhanced in the presence of Ebp1-p48 isoform (Fig. 8B, lanes 3–5). The inhibition of HCV replication by the Ebp1-p42 isoform may be a consequence of its binding to and sequestering of NS5B and NS5A components from the HCV replication complex. In contrast, the p48 isoform, which also binds to NS5A and NS5B, may facilitate their association with the replication complex. These results indicate that the Ebp1 isoforms function as positive and negative regulators of HCV replication.

Figure 8.

Endogenous in-vitro HCV replication in cell-free replication lysate in the absence or presence of exogenously added Ebp1-p42 (A) or Ebp1-p48 (B). An aliquot of a cell-free replicative lysate of MH14 cells was incubated in the absence and presence of 0.25, 0.5 pmol, and 1 pmol of Ebp1-p42 or Ebp1-p48 isoform. RT-PCR for HCV RNA and GAPDH mRNA was then done. Quantification of HCV RNA by qPCR is shown in the bottom panel of A and B.

Effects of Ebp1-isoforms on in-vitro RdRp activity of NS5B in the absence and presence of activated PKR. Since exogenously added Ebp1-p42 and Ebp1-p48 isoforms respectively inhibit and promote endogenous HCV replication in cell-free replication lysate, we examined whether or not they have any effect on the in-vitro RdRp activity of NS5B. We incubated purified recombinant NS5B in the polymerase assay reaction in the absence or presence of Ebp1 isoform and activated PKR. In the reaction, 0.1 mM ATP was also supplemented in the absence or presence of activated PKR. The reactions were carried out at 37°C for 30 min. The acid precipitable polymerase reaction products were collected and counted for radioactivity.

The results shown in Figure 9 indicate that both the Ebp1 isoforms alone had no effect on the RdRp activity of HCV replicase (Fig. 9, lanes 2 and 6). The addition of activated PKR alone drastically inhibited polymerase activity of NS5B (lanes 3 and 7). The addition of Ebp1 isoform in the presence of activated PKR did not affect the PKR-mediated inhibitory effect on the polymerase activity of NS5B (lanes 4 and 8). These results suggested that Ebp1-isoforms do not directly affect the polymerase activity of NS5B; their influence on the RdRp activity of the enzyme may be by regulating the activation of PKR. The inhibition of RdRp activity by activated PKR may be due to PKR-mediated phosphorylation of NS5B at a regulatory site that may interfere with its catalytic activity of the enzyme.

Figure 9.

In-vitro RdRp activity of NS5B in the absence or presence of Ebp1 isoforms and activated PKR. We used poly rA/dT₁₈ as the template-primer and ³H-UTP as the rNTP substrate. Reaction mixture was also supplemented with 0.1 mM ATP. Reactions were carried out at 37°C for 30 min and products were analyzed by counting TCA precipitable radioactive materials in a scintillation counter. Lanes 1 and 5, control; lanes 2 and 6, in the presence of indicated Ebp1 isoform alone; lanes 3 and 7, in the presence of activated PKR alone; lanes 4 and 8, in the presence of activated PKR and indicated Ebp1 isoform.

Co-localization of Ebp1 with NS5A, NS5B, and PKR. Ebp1-p42 isoform, being a cytoplasmic dsRNA binding protein, is abundantly expressed and mainly localized in the cytosol; on the other hand, expression of the p48 isoform, which is primarily localized in the nucleus is low. Since Ebp1-IP co-immunoprecipitates NS5A and PKR from the cell lysate of HCV-infected cells, we examined whether Ebp1 is co-localized with these proteins. We did immunofluorescence staining of Ebp1 and NS5A, as well as Ebp1 and PKR, then examined their images by confocal microscope. We found that Ebp1 is colocalized with NS5A (Fig. 10A), NS5B (Fig. 10B) and PKR (Fig. 10C) in the cytoplasm of MH14 cells. Cured MH14 cells devoid of HCV replicon were used as a negative control (Fig. 10D). We found that Ebp1 was colocalized with NS5A, NS5B and PKR with an overlap coefficient of 0.65, 0.70 and 0.75, respectively.

Figure 10.

Co-localization of Ebp1 with HCV NS5A, NS5B and PKR in MH14 cells. Cells grown on 8-chamber tissue-culture slides for 24 h were fixed, treated with the primary anti-Ebp1 antibody (rabbit), then treated with anti-rabbit Alexa-594-labeled secondary antibody (red). The Ebp1-antibody treated cells were then separately treated with mouse anti-NS5A antibody (A), goat anti-NS5B (B), or mouse anti-PKR antibody (C), then treated with their respective Alexa 488-labeled secondary antibodies (Dixit et al., 2015). DAPI was used to stain nuclei. Cells were observed by confocal microscope. Pictures were taken individually and merged using NIS viewer’s software (Nikon). Colocalization of Ebp1 with NS5A (A), NS5B (B) and PKR (C). Ebp1 colocalized with NS5A, NS5B and PKR with an overlap coefficient of 0.65, 0.70 and 0.75 respectively. Cured MH14 cells devoid of HCV replicon were used as a negative control (D).

DISCUSSION

Earlier, we identified Ebp1 as a cell factor that binds to the HCV RNA genome and modulates its replication and translation (Upadhyay et al., 2013). There are two alternatively spliced isoforms of Ebp1, the longer isoform p48, and the shorter isoform p42 (Liu et al., 2006). The Ebp1 transcript contains three in-frame ATG codons in the N-terminus. The first ATG codon is used to initiate translation of the longer p48 isoform, while the third ATG codon is used to translate the shorter p42 isoform (Fig. 1A) (Xia et al., 2001b). The longer and the shorter isoforms have opposite biological effects (Liu et al., 2006). The longer p48 isoform has an oncogenic character that induces cell survival (Kim et al., 2010), whereas the shorter p42 isoform inhibits cell growth and acts as a tumor suppressor. The Ebp1-p42 is abundantly expressed and exclusively localized in the cytosol while the expression level of p48 isoform is very low and distributed mainly in nucleus (Figure 1B). However, Western blot analysis of total cell lysate showed only one band of Ebp1 corresponding to the most abundantly expressed p42 isoform (Figure 1C, lanes 1–3). This could be caused by the masking effect of abundantly expressed cytosolic p42 on the less abundant nuclear-localized p48 isoform.

We found that downregulation of total Ebp1 significantly enhanced HCV replication and translation indicating its inhibitory effect on HCV replication. Our effort at selective downregulation of each of the isoforms was inconclusive, probably for the following reasons. Since mRNAs of both isoforms are derived from the same pre-spliced mRNA by differential splicing, targeting the mRNA of Ebp1–42 isoform may also target mRNA levels of the larger p48 isoform. Downregulation of Ebp1-p48 isoform by targeting the 29-nucleotide-long region, which is absent from p42-mRNA, may also target the pre-spliced mRNA and thus reducing the mRNA level of both the isoforms. This may be reason of reduced expression level of p42 isoform when Ebp1-p48 was selectively downregulated by p48-siRNA.

Since the siRNA experiment was not conclusive, we resorted to overexpression of each isoform ectopically in cell and examined their effect on HCV replication and translation. However, the effect of overexpression of individual isoforms on HCV replication and translation was complicated by the presence of endogenously expressing isoforms. Since the isoforms have opposite effects on HCV replication and translation, the net effect of overexpression of each isoform may also include the antagonizing effect of the endogenously expressing isoform. We therefore, resorted to transient expression of individual Ebp1 isoforms in the absence of endogenously expressed Ebp1. For this purpose, we first constructed stable Ebp1-kd cells using shRNA against Ebp1. We then constructed shRNA-resistant expression clones of p42 and p48 isoforms for transient expression in Ebp1-kd cells. We found that transient expression of the Ebp1-p42 isoform in Ebp1-kd cells significantly inhibited HCV replication and translation, whereas transient expression of the longer Ebp1-p48 isoform had no inhibitory effect on HCV replication; its selective expression in Ebp1-kd cells marginally enhanced HCV replication and translation (Fig. 5).

Interestingly, only the p42 isoform interacted with and bound to the HCV IRES. The p48 isoform bound poorly (Fig. 4). The most probable explanation for the differential binding affinity of the isoforms to HCV 5’NTR is based on their structures. The N-terminal domain comprising 54 amino acids in the longer p48- isoform is missing from the N-terminal region of the shorter p42 isoform (Monie et al., 2007) suggesting that the absence of this domain in the shorter isoform may be essential for its binding to HCV 5’NTR. In the p48 isoform, the N-terminal 54-residues long motif is comprised of α1 and half of the α2 structure located at the entrance of the central cavity where RNA may bind (Figure 11). This explains why, unlike p42, the p48 isoform does not efficiently bind HCV IRES.

Figure 11.

The structure of p48 and p42 isoforms of Ebp1. This structure, a ribbon representation with the beta barrel forming the central cavity, was taken from protein data bank (PDB: 2Q8K). The α1 and part of α2 at the N-terminal domain positioned at the entrance of the central cavity of the p48 isoform are missing in the p42 isoform.

We also found that both isoforms of Ebp1 interact with two viral nonstructural proteins, NS5A, and NS5B, and cell factor PKR. Although both isoforms interact with the viral proteins, only the shorter isoform inhibited HCV replication. The endogenous HCV replication in cell-free replication lysate was also significantly inhibited by purified recombinant p42 isoform, while the p48 isoform promoted endogenous viral replication in the cell-free replication lysate (Figure 8). Since both isoforms interact with viral proteins NS5A and NS5B, it was intriguing that they had opposite effects on endogenous HCV replication in cell-free replication lysate. It is possible that p42 isoform may bind and sequester away NS5B and NS5A from the HCV replication complex while p48-isoform may promote their association with the replication complex. Alternatively, p42 may directly inhibit the polymerase activity of NS5B.

The in-vitro RdRp activity of recombinant HCV replicase is drastically inhibited in the presence of activated PKR but is not affected by either of the isoforms alone. Interestingly, the inhibited RdRp activity in the presence of activated PKR is not affected by either of the Ebp1 isoforms. These results suggest that PKR may negatively influence the RdRp activity by phosphorylating NS5B at a regulatory site that may be detrimental to its catalytic activity. Since p42 isoform promotes activation and autophosphorylation of PKR, it may promote PKR-mediated phosphorylation of NS5B. Although the p48 isoform inhibits autophosphorylation and activation of PKR, it may not block phosphorylation of NS5B by activated PKR. It has been shown that Ser 29 and Ser 42, located on the Δ1 finger loop region of NS5B, are the sites of phosphorylation by protein kinase C that are required for efficient HCV replication (Han et al., 2014). The Δ1 finger loop structure caps the polymerase cleft by interacting with the thumb subdomain and thus confers closed conformation to the enzyme required for de-novo synthesis of viral RNA genome (Hong et al., 2001). It is possible that PKR may kinase NS5B at a different regulatory site(s) that may not be favorable for the catalytic function of the enzyme. It will be interesting to phosphorylate NS5B with PKR in the presence and absence of Ebp1 isoforms and identify the site of phosphorylation by LC/MS/MS, and then validate the significance of these sites by site-directed mutagenesis.

Based on our results, a model of the possible interaction of Ebp1 isoform was constructed to explain their effect on HCV replication and translation (Figure 12). Both of the isoforms interact with HCV NS5B, NS5A and host cell factor PKR, but have opposite effects. The p42 isoform antagonizes NS5A and NS5B and activates PKR, whereas the p48 isoform, together with NS5A, synergistically inhibits PKR activation but promotes the RdRp activity of NS5B.

Figure 12.

Schematic model of Ebp1 interaction and its implication with regard to HCV replication and translation. Ebp1 isoforms are the translation product of differentially spliced mRNAs from the same pre-sliced transcript. Ebp1-p42 isoform interacts with and activates PKR, antagonizes NS5A and NS5B, and inhibits HCV replication and translation. Ebp1-p48 isoform together with NS5A blocks PKR activation promotes the RdRp activity of NS5B and facilitates HCV replication and translation.

Ebp1 is also known to interact with retinoblastoma and affects its transcriptional regulation (Xia et al., 2001a). Its interaction with nucleophosmin is essential for regulating cell proliferation and suppressing apoptosis (Okada et al., 2007). Nucleophosmin (NPM/B23) a multifunctional oncoprotein that is involved in regulating cell growth and proliferation (Okuwaki et al., 2002; Olanich et al., 2011), also functions as a positive and negative regulator of p53 and HDM2/MDM2 (Kurki et al., 2004). Earlier we have identified another cell factor, FUSE-binding protein, which represses NPM (Olanich et al., 2011) physically interacts with and antagonizes p53 function, and promotes persistent HCV replication (Dixit et al., 2015). The Ebp1-p48 isoform also negatively regulates p53 and thus may support HCV replication by antagonizing p53 function (Kim et al., 2010). In contrast, the Ebp1-p42 isoform upregulates p53 (Zhang et al., 2014) which, in turn, may facilitate p53-mediated inhibition of HCV replication. The Ebp1-p42 isoform also interacts with corepressor histone deacetylase (HDAC) complex and inhibits cell proliferation by blocking the activation of transcription factor E2F1 (Zhang et al., 2005; Zhang et al., 2003). Activated E2F1 is required to promote HCV replication (Hassan et al., 2004), which is strongly inhibited by Ebp1 (Okada et al., 2007).

In the cytoplasm, Ebp1 is associated with ribosomes and inhibits the phosphorylation of eukaryotic initiation factor 2 alpha (eIF2α) (Squatrito et al., 2006). The overexpression of Ebp1-p48 protects eIF2α from phosphorylation in response to cellular stress (Squatrito et al., 2006). In this context, our finding that the longer p48 isoform of Ebp1 interacts with PKR and inhibits its autophosphorylation (Figure 3C) may explain the mechanism by which the p48-isoform facilitates HCV replication. This possibility was supported by the fact that the p48 isoform together with HCV NS5A, had synergistic inhibitory effect on PKR autophosphorylation, while the p42 isoform antagonized the effect of NS5A and promoted PKR activation and phosphorylation (Figure 3D). The activated PKR phosphorylates eIF2α resulting in the shutdown of the translation of specific cell factors that may be required for persistent viral replication. Thus, Ebp1-p42 isoform may offer cellular innate immunity to HCV infection by promoting activation of PKR, which, in turn, may phosphorylate eIF2α to shut off the cellular translation, and NS5B to block HCV replication. This effect of Ebp1-p42 on PKR activation is effectively antagonized by HCV NS5A together with the p48 isoform of Ebp1.

Highlights.

• The p42 and p48 isoforms of Ebp1 affect HCV replication by modulation of PKR activation.

• PKR activation is promoted by p42 isoform, and inhibited by p48 isoform.

• Activated PKR strongly inhibits RdRp activity of HCV replicase.