Ifit1 Inhibits Japanese Encephalitis Virus Replication through Binding to 5′ Capped 2′-O Unmethylated RNA

Abstract

The interferon-inducible protein with tetratricopeptide (IFIT) family proteins inhibit replication of some viruses by recognizing several types of RNAs, including 5′-triphosphate RNA and 5′ capped 2′-O unmethylated mRNA. However, it remains unclear how IFITs inhibit replication of some viruses through recognition of RNA. Here, we analyzed the mechanisms by which Ifit1 exerts antiviral responses. Replication of a Japanese encephalitis virus (JEV) 2′-O methyltransferase (MTase) mutant was markedly enhanced in mouse embryonic fibroblasts and macrophages lacking Ifit1. Ifit1 bound 5′-triphosphate RNA but more preferentially associated with 5′ capped 2′-O unmethylated mRNA. Ifit1 inhibited the translation of mRNA and thereby restricted the replication of JEV mutated in 2′-O MTase. Thus, Ifit1 inhibits replication of MTase-defective JEV by inhibiting mRNA translation through direct binding to mRNA 5′ structures.

INTRODUCTION

mRNA has a 5′ cap structure, in which the N-7 position of the guanosine residue is methylated. The 5′ cap structure is known to be responsible for the stability and efficient translation of mRNA (1, 2). In higher eukaryotes, the first one or two 5′ nucleotides are additionally methylated at the ribose 2′-O position by distinct host nuclear 2′-O methyltransferases (MTases) (3, 4). However, the functional role of 2′-O methylation (2′-O Me) remains poorly understood. Several viruses that replicate in the cytoplasm possess their own mRNA capping machineries (5–10). For positive-stranded flaviviruses, nonstructural protein 3 (NS3) acts as an RNA 5′-triphosphatase and NS5 possesses both N-7 and 2′-O MTase activities (8, 11, 12). Recent studies have revealed that 2′-O methylation of the mRNA 5′ cap in these viruses is important for evasion from the host innate immune responses (13–15). However, the 2′-O MTase activity has been shown to be absent from several paramyxoviruses, such as Newcastle disease virus (NDV) and respiratory syncytial virus (RSV) (16, 17).

Type I interferons (IFNs) induce the expression of a large number of antiviral genes through a Janus kinase/signal transducer and activator of transcription (JAK/STAT) pathway (18, 19). Among the IFN-inducible genes, the IFN-inducible protein with tetratricopeptide (IFIT) genes comprise a large family with three (Ifit1, Ifit2, and Ifit3) and four (IFIT1, IFIT2, IFIT3, and IFIT5) members in mice and humans, respectively. The murine and human genes are clustered in loci on chromosomes 19C1 and 10q23, respectively (20). IFIT family proteins reportedly associate with several host proteins to exert various cellular functions (21, 22). For example, human IFIT1/IFIT2 and murine Ifit1/Ifit2 bind to eukaryotic translational initiation factor 3 (eIF3) subunits to inhibit translation (23–26). IFIT1 has been suggested to interact with STING/MITA to negatively regulate IRF3 activation (27), whereas IFIT3 may bind TBK1 to enhance type I IFN production and with JAB1 to inhibit leukemia cell growth (28, 29).

In addition to binding host factors, IFIT proteins have functional effects by interacting directly with products of viruses. Human IFIT1 interacts with the human papillomavirus E1 protein and human IFIT2 interacts with the AU-rich RNA of NDV to exert antiviral effects (30, 31). Direct binding of IFIT proteins to virus RNA has also been demonstrated in several recent studies. IFIT1 and IFIT5 bind to the 5′-triphosphate (5′-PPP) RNA that is present in the genomes of viruses (32, 33). Structural studies of human IFIT2 and human IFIT5 identified an RNA-binding site and defined the structural basis of a complex with 5′-PPP RNA (31, 33). However, these structural studies did not explain how IFIT binds to or restricts virus RNA that has a 5′ cap but lacks methylation at the 2′-O position (13–15). Thus, it remains unclear how IFITs mediate antiviral activities against viruses that have a 5′ cap but lack 2′-O MTase activity.

In this study, we analyzed the mechanisms by which murine Ifit1 exerts the host defense against a flavivirus lacking 2′-O MTase activity. Ifit1 was found to preferentially interact with 5′ capped mRNA without 2′-O methylation and inhibit its translation. Thus, Ifit1 participates in antiviral responses targeting 5′ capped mRNA without 2′-O methylation.

MATERIALS AND METHODS

Mice.

All animal experiments were conducted in accordance with the guidelines of the Animal Care and Use Committee of the Graduate School of Medicine, Osaka University. The gene-targeting strategies for generating Ifit1-knockout (Ifit1^−/−) mice were described previously (34). The Ifit1-targeting vector was designed to replace a 1.8-kb fragment encoding the exon of Ifit1 with a neomycin resistance gene cassette (Neo). A short arm and a long arm of the homology region from the v6.5 embryonic stem (ES) cell genome were amplified by PCR. A herpes simplex virus (HSV) thymidine kinase (tk) gene was inserted into the 3′ end of the vector. After the Ifit1-targeting vector was electroporated into ES cells, G418 and ganciclovir doubly resistant clones were selected and screened by PCR and Southern blot analysis. An ES cell clone correctly targeting Ifit1 was microinjected into C57BL/6 mouse blastocysts. Chimeric mice were mated with female C57BL/6 mice, and heterozygous F1 progenies were intercrossed to obtain Ifit1^−/− mice.

Cells.

HEK293T cells, Vero cells, and mouse embryonic fibroblasts (MEFs) were maintained in Dulbecco's modified Eagle's medium (Nakalai Tesque) supplemented with 10% fetal bovine serum (JRH Bioscience), 100 U/ml penicillin, and 100 μg/ml streptomycin (Gibco). MEFs were prepared from wild-type (WT) and Ifit1^−/− day 14.5 embryos and immortalized by introduction of a plasmid encoding the simian virus 40 large T antigen. MEFs stably expressing Ifit1 were established by the previously described method with some modifications (34). In short, full-length cDNA of Ifit1 was cloned into pMRX-puro (pMRX/Ifit1). Retrovirus was produced by introduction of pMRX/Ifit1 into Plat-E packaging cells (35). MEFs were infected with the retrovirus, cultured in the presence of 1 μg/ml of puromycin (Sigma) for 5 days, and harvested for subsequent studies. To isolate peritoneal macrophages, mice were intraperitoneally injected with 5 ml of 4% thioglycolate medium (Sigma), and peritoneal exudative cells were isolated from the peritoneal cavity at 3 days postinjection. The cells were incubated for 2 h and then washed three times with Hanks' balanced salt solution. The remaining adherent cells were used as peritoneal macrophages in the experiments.

Viruses.

Japanese encephalitis virus (JEV) strain AT31 (36) was used for the experiments. An NS5 K61A mutation of JEV was introduced into pMWATG1 (37) by PCR-based mutagenesis with the primers 5′-GCGAGGCTCAGCAGCTCTCCGTTGGCTCG-3′ and 5′-CGAGCCAACGGAGAGCTGCTGAGCCTCGC-3′ (the mutagenesis site is underlined) and verified by DNA sequencing. A recombinant virus, the JEV K61A mutant, was generated from pMWJEATG1/JEV K61A as previously described (36). MEFs or macrophages were infected with JEV at specified multiplicities of infection (MOIs). The virus yields in the culture supernatants were titrated by focus-forming assays on Vero cells and expressed as the number of focus-forming units (FFU), as previously described (38). The virus RNA accumulations in the JEV-infected cells were determined by real-time reverse transcription-PCR (RT-PCR) with primers targeting JEV NS5, normalized to the level of host GAPDH (glyceraldehyde-3-phosphate dehydrogenase), and expressed as the fold change in Ifit1^−/− cells versus wild-type cells (value for wild type = 1).

Preparation of RNA.

The 5′-terminal 200 nucleotides of the JEV genome were amplified by PCR using pMWATG1 (37) with the primers 5′-TAATACGACTCACTATTAGAAGTTTATCT-3′ (the T7 class II promoter sequence is underlined) and 5′-CATTACTACCCTCTTCACTCCCACTAGTGG-3′, and the luciferase reporter gene (luc2) was amplified using pGL4.14 (Promega) with the primers 5′-TAATACGACTCACTATAGGCCACCATGGAAGATGCCAAAAA-3′ (the T7 class III promoter sequence is underlined) and 5′-TACCACATTTGTAGAGGTTTTACTTGCTTT-3′. Subsequently, the PCR products were in vitro transcribed under the control of the T7 promoter with MEGAScript (Ambion). Biotin-labeled RNA was prepared by in vitro transcription in the presence of biotin-labeled UTP (PerkinElmer). Capped RNA substrates were produced with a ScriptCap 7-methylguanosine (m7G) capping system (Epicentre) in the presence (5′ cap positive [5′ cap⁺]/2′-O Me positive [2′-O Me⁺]) or absence (5′ cap⁺/2′-O Me negative [2′-O Me⁻]) of a ScriptCap vaccinia virus 2′-O MTase (Epicentre). ³²P-labeled m7GpppA-RNA substrate was prepared with a ScriptCap m7G capping system in the presence of ³²P-labeled GTP. A 5′ OH-RNA substrate was produced by incubating in vitro-transcribed RNA with calf intestinal alkaline phosphatase (CIAP) for 3 h at 37°C. All RNA substrates were purified with an RNeasy minikit (Qiagen) and stored at −80°C until use.

Real-time RT-PCR.

Total RNA was isolated with the TRIzol reagent (Invitrogen), and 1 to 2 μg of RNA was reverse transcribed using Moloney murine leukemia virus reverse transcriptase (Promega) and random primers (Toyobo) after treatment with RQ1 DNase I (Promega). Real-time RT-PCR was performed in an ABI 7300 apparatus (Applied Biosystems) using a GoTaq real-time PCR system (Promega). All values were normalized by the expression of the GAPDH gene. The following primer sets were used: for the JEV NS5 gene, 5′-AACGCACATTACGCGTCCTAGAGATGA-3′ and 5′-CTAACCCAATACATCTCGTGATTGGAGTT-3′; for Ifnb, 5′-GGAGATGACGGAGAAGATGC-3′ and 5′-CCCAGTGCTGGAGAAATTGT-3′; for luc2, 5′-CCATTCTACCCACTCGAAGACG-3′ and 5′-CGTAGGTAATGTCCACCTCGA-3′; and for the GAPDH gene, 5′-CCTCGTCCCGTAGACAAAATG-3′ and 5′-TCTCCACTTTGCCACTGCAA-3′.

Recombinant proteins.

Wild-type and K61A mutant JEV N-terminal NS5 (MTase domain) cDNAs were obtained by PCR using pMWATG1 with the primers 5′-GGATCCGGAAGGCCTGGGGGCAGGACGCTA-3′ and 5′-CTCGAGATGCTCAGGGTCTTTGTGCCACGT-3′. Full-length murine Ifit1 cDNA and JEV MTase cDNA were inserted into pET-15b and pGEX-6P, respectively. pET/Ifit1 and pGEX/JEV MTases were transformed into the Escherichia coli BL21(DE3) strain. Expression of the Ifit1 and JEV NS5 proteins was induced by addition of 0.5 mM isopropyl-1-thio-β-d-galactopyranoside (IPTG), and the expressed Ifit1 and JEV MTase proteins were purified using Ni²⁺-affinity chromatography (Novagen) and glutathione-Sepharose 4B (Amersham Biosciences), respectively, according to each manufacturer's instructions. The purified protein was desalted and concentrated using an Amicon Ultra centrifugal filter unit (Millipore) and stored at −80°C until use.

In vitro MTase activity assay.

The MTase reaction was performed in a 20-μl reaction mixture of 50 mM Tris-HCl (pH 8.0), 6 mM KCl, 1.25 mM MgCl₂, and 0.5 mM S-adenosylmethionine (AdoMet) containing 10 nmol of ³²P-labeled m7GpppA-RNA substrate (JEV 5′-terminal 200 nucleotides) and 30 pmol of JEV MTase or 80 units of vaccinia virus 2′-O MTase (Epicentre) for 3 h at 37°C. The RNA was purified by passage through a postreaction cleanup column (Sigma) and digested with 10 U of nuclease P1 (Wako) in 50 mM sodium acetate overnight at 37°C. The samples were analyzed on thin-layer chromatography polyethyleneimine (PEI)-cellulose plates developed with 0.3 M ammonium sulfate.

RNA EMSAs.

RNA electrophoretic mobility shift assays (EMSAs) were performed using a LightShift chemiluminescent RNA EMSA kit (Thermo Scientific) according to the manufacturer's instructions. Briefly, 0 to 20 pmol of recombinant murine Ifit1 and 2.5 pmol of in vitro-transcribed and biotin-labeled RNA were coincubated for 30 min at room temperature in RNA EMSA binding buffer (10 mM HEPES, pH 7.3, 20 mM KCl, 1 mM MgCl₂, 1 mM dithiothreitol, 0.1 μg/μl of yeast tRNA, 2% glycerol). The resulting Ifit1/RNA complexes were electrophoresed in a 7.5% native polyacrylamide gel. The separated RNAs were transferred to a positively charged nylon membrane and cross-linked at 120 mJ/cm² and an absorbance of 254 nm. The membrane was incubated with stabilized streptavidin-horseradish peroxidase conjugate (1:300 dilution; a component of the EMSA kit), and the bound stable peroxide was detected with luminol/enhancer solution (another component of the EMSA kit). The gel-shift band intensities were quantified using ImageJ software (National Institutes of Health).

RNA pulldown assay.

For RNA pulldown assays, an expression vector for hemagglutinin (HA)-tagged murine full-length Ifit1 was transfected into HEK293T cells using Lipofectamine 2000 (Invitrogen). The Ifit1-transfected cells were lysed in RNA-binding buffer (10 mM HEPES, pH 7.3, 500 mM KCl, 1 mM EDTA, 2 mM MgCl₂, 0.1% NP-40, 0.1 μg/μl of yeast tRNA (Ambion), 1 U/ml of RNase inhibitor [Toyobo]), and the lysate (200 μg) was coincubated with 25 pmol of biotin-labeled RNA and streptavidin-agarose (Invitrogen) in RNA-binding buffer for 30 min at room temperature. The binding complexes were washed five times with RNA-binding buffer, followed by SDS-PAGE and immunoblotting with an anti-HA probe (F-7) antibody (Santa Cruz Biotechnology). The intensity of the detected Ifit1 band was quantified using ImageJ software (National Institutes of Health).

RNA immunoprecipitation.

RNA immunoprecipitation was performed as described previously (38) with slight modifications. MEFs (2 × 10⁵) stably expressing Flag-tagged Ifit1 were infected with JEV at an MOI of 1.0 and cultured for 24 h. The cells were then lysed in 500 μl of RNA buffer (10 mM HEPES, pH 7.3, 500 mM KCl, 1 mM EDTA, 2 mM MgCl₂, 0.1% NP-40, 0.1 μg/μl of yeast tRNA (Ambion), 1 U/ml of RNase inhibitor [Toyobo], 1 tablet/10 ml of Complete mini-protease inhibitor cocktail [Roche]). After centrifugation at 15,000 rpm for 20 min at 4°C, 50-μl aliquots of the supernatants were recovered as input samples and the remaining supernatants were precleared with 30 μl of 50% protein G-conjugated Sepharose and 1 μg of mouse normal IgG for 1 h. After centrifugation of the beads, the supernatants were immunoprecipitated with 1 μg of mouse normal IgG or anti-Flag M2 antibody (Santa Cruz Biotechnology) and 30 μl of 50% protein G-conjugated Sepharose. The beads were washed five times with RNA buffer without yeast tRNA, and RNA was isolated from the precipitants and input samples with the TRIzol reagent. The RNA was reverse transcribed as described above and subjected to the first round of PCR with JEV NS1-specific primers 5′-TCTGTCACTAGACTGGAGCA-3′ and 5′-CCAGAAACATCACCAGAAGG-3′. The PCR products were then analyzed by quantitative PCR with nested primers 5′-GAGCACTGACGAGTGTGATG-3′ and 5′-AGCGACTCTCAATCCAGTAC-3′. All values were normalized by the values for the input samples (indicated as percent input).

Cellular translational reporter assay.

MEFs (2 × 10⁵) were pretreated with 1,000 U/ml of universal type I interferon (PBL Biomedical Laboratories) for 6 h. Three types of 5′ modified luciferase mRNAs (2 μg of 5′-PPP, 5′ cap⁺/2′-O Me⁻, and 5′ cap⁺/2′-O Me⁺) were transiently transfected into MEFs using the Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer's instructions. At 6 h after the transfection, RNA was isolated and analyzed for the quantity of the luciferase mRNAs (luc2). The luciferase activities of whole-cell lysates were measured using a dual-luciferase reporter assay system (Promega). The numbers of relative light units (RLU) were normalized by the concentrations of proteins determined by use of a bicinchoninic acid protein assay kit (Thermo Scientific).

Statistical analysis.

Statistical analyses were conducted on each independent data set. An unpaired Student's t test was used to determine the statistical significance of differences in the experimental data. P values of <0.05 were considered to indicate statistical significance.

RESULTS

Ifit1^−/− cells fail to restrict the replication of a mutant JEV lacking 2′-O MTase activity.

Previous analysis of the flavivirus West Nile virus (WNV) 2′-O MTase revealed residues in NS5 (K61, D146, K182, and E218) that were essential for its biochemical activity (8). A WNV mutant (E218A) lacking 2′-O MTase activity was attenuated in mouse MEFs and macrophages but showed enhanced replication in Ifit1^−/− cells (13, 15). As NS5 is a highly conserved protein in flaviviruses, the four residues integral to the 2′-O MTase activity are identical in WNV and JEV (Fig. 1A). Replacement of lysine 61 by alanine in the JEV NS5 MTase domain (JEV K61A) abolished the JEV 2′-O MTase activity in vitro (Fig. 1B). We generated Ifit1^−/− mice (Fig. 2A to C) and infected MEFs with JEV WT and JEV K61A strains (Fig. 3A). The JEV WT replicated equivalently in wild-type and Ifit1^−/− MEFs. In comparison, the production of the JEV K61A mutant was decreased in wild-type MEFs, suggesting that 2′-O MTase activity is required for JEV replication. Consistent with this and analogous studies with an WNV E218A strain (13), replication of the JEV K61A strain was enhanced (approximately 173-fold increased at 4 days postinfection; P < 0.05) in Ifit1^−/− MEFs compared with wild-type MEFs. We also infected peritoneal macrophages with JEV WT and JEV K61A strains (Fig. 3B). Similar to the results obtained with MEFs, replication of the JEV WT was similarly observed in wild-type and Ifit1^−/− macrophages. However, replication of the JEV K61A mutant was severely decreased in wild-type but not Ifit1^−/− macrophages, and the virus was not detected at 4 days postinfection in wild-type cells. For further confirmation, we analyzed virus RNA accumulation at 4 days postinfection (Fig. 3C and D). Whereas RNA levels of JEV WT were similar in wild-type and Ifit1^−/− MEFs, those of the JEV K61A mutant were markedly higher (approximately 13-fold; P < 0.05) in Ifit1^−/− MEFs than in wild-type MEFs. To further corroborate these findings, we reintroduced the Ifit1 gene into Ifit1^−/− MEFs using a retrovirus vector. Replication of the JEV K61A mutant was considerably suppressed (approximately 4-fold; P < 0.05) by ectopic Ifit1 expression in Ifit1^−/− MEFs (Fig. 3E). Ifnb was similarly induced in wild-type and Ifit1^−/− MEFs after infection with the JEV K61A mutant, excluding the possibility that defective type I IFN production is responsible for the high sensitivity to infection with the JEV K61A mutant in Ifit1^−/− cells (Fig. 3F). Thus, consistent with the findings of previous studies (13, 15), Ifit1 inhibits replication and infection of flavivirus mutants that lack 2′-O MTase activity.

Fig 1.

Generation of an MTase-defective JEV mutant. (A) Sequence homology between NS5 proteins of JEV (AT31 strain, GenBank accession number AB196926) and WNV (00-3356 strain, GenBank accession number EF530047). Arrowheads, MTase catalytic K-D-K-E tetrad; *, consensus sequences between the two proteins. (B) 2′-O MTase activity of JEV WT and JEV K61A mutant recombinant NS5 proteins by thin-layer chromatography assays. The substrate m7GpppA-RNA (³²P-labeled JEV 5′-terminal 200 nucleotides) was methylated in vitro with the respective recombinant NS5 proteins or vaccinia virus 2′-O MTase, digested with P1 nuclease, and developed on PEI-cellulose plates. The positions of 2′-O methylated (m7GpppAm) and unmethylated (m7GpppA) RNA are indicated. Data are representative of four independent experiments.

Fig 2.

Generation of Ifit1^−/− mice. (A) Schematic representation of the Ifit1 gene-targeting strategies. Solid boxes, coding regions of the Ifit1 gene; open boxes, untranslated regions; Neo and HSV tk, a neomycin-resistance gene cassette and a herpes simplex virus thymidine kinase gene, respectively. The positions of the probe and restriction enzyme site for Southern blotting are shown. (B) Genomic DNA was isolated from the tails of wild-type (+/+), heterozygous (+/−), and homozygous (−/−) Ifit1 mutant mice. A Southern blot analysis performed after digestion of the genomic DNA with BamHI shows the correct targeting of the locus. (C) Peritoneal exudative macrophages were harvested from wild-type (+/+) or Ifit1-deficient (−/−) mice. Total RNA (10 μg) was blotted onto a nylon membrane, and Ifit1 and β-actin mRNA expression was detected by Northern blot analysis with the respective cDNA probes. LPS lanes, cells stimulated with 100 ng/ml of lipopolysaccharide for 4 h to induce endogenous Ifit1 expression; Med lanes, cells treated with medium alone.

Fig 3.

Ifit1^−/− MEFs and macrophages fail to restrict the replication of the 2′-O MTase mutant JEV. (A, B) Culture supernatants of wild-type and Ifit1^−/− MEFs (A) and macrophages (B) infected with JEV WT and the JEV K61A mutant (MOIs, 0.1 for MEFs and 0.5 for macrophages) were harvested at the indicated days postinfection. The virus titers in 1-ml supernatant aliquots were determined by focus-forming assays on Vero cells and expressed as the log₁₀ number of FFU/ml. Data are shown as means ± SDs of quadruplicate samples generated from four independent experiments with statistical significance. ND, not detected. *, P < 0.05. (C, D) Accumulation of JEV WT (C) and the JEV K61A mutant (D) RNA in wild-type and Ifit1^−/− MEFs at 4 days postinfection determined by quantitative real-time RT-PCR. JEV NS5 RNA levels were normalized to the level of host GAPDH and are expressed as the fold change in Ifit1^−/− cells versus wild-type cells (value for wild type = 1). Data are representative of three independent experiments with statistical significance. *, P < 0.05. (E) Culture supernatants of vector-transduced (+vector) and Flag-tagged Ifit1 gene-transduced (+Ifit1) Ifit1^−/− MEFs infected with the JEV K61A mutant (MOI, 0.1) were harvested at 3 days postinfection. The virus titers in 1-ml supernatant aliquots were determined by focus-forming assays on Vero cells and expressed as the log₁₀ number of FFU/ml. Expression of Ifit1 and β-actin determined by immunoblotting with anti-Flag or anti-β-actin antibodies is shown at the bottom. Data are representative of three independent experiments. *, P < 0.05. (F) Wild-type and Ifit1^−/− MEFs were infected with the JEV K61A mutant (MOI, 0.1). At 4 days postinfection, cells were harvested and analyzed for Ifnb expression by quantitative RT-PCR. Ifnb RNA levels were expressed relative to those of GAPDH. ND, not detected. Data are shown as means ± SDs and are representative of data from three independent experiments.

Ifit1 preferentially binds to virus RNA lacking 2′-O methylation.

Next, we analyzed how Ifit1 recognizes 2′-O MTase mutant viruses. While recombinant IFIT1 reportedly binds to 5′-PPP RNA (32), the mRNA of the JEV K61A mutant has a 5′ m7G cap but lacks 2′-O methylation (5′ cap⁺/2′-O Me⁻). We examined whether Ifit1 can also interact directly with 5′ cap⁺/2′-O Me⁻ RNA using electrophoretic mobility shift assays. Consistent with a previous report (32), bands of 5′-PPP RNA but not RNA lacking phosphate at the 5′ end (5′ OH) were diminished after addition of recombinant Ifit1 (Fig. 4A). Furthermore, Ifit1 blocked the electrophoretic mobility of the 5′ cap⁺/2′-O Me⁻ RNA. However, this effect was rescued by exogenous addition in vitro of 2′-O methylation (5′ cap⁺/2′-O Me⁺). The efficient binding of Ifit1 to 5′ cap⁺/2′-O Me⁻ RNA was corroborated by RNA pulldown assays (Fig. 4B). HA-tagged Ifit1 was expressed in HEK293T cells, and cell lysates were incubated with biotin-labeled in vitro-transcribed RNA and streptavidin-agarose. Then, binding complexes of Ifit1/RNA were analyzed by Western blotting. While Ifit1 was not pulled down with 5′ OH RNA, modest binding of Ifit1 to 5′-PPP RNA and 5′ cap⁺/2′-O Me⁺ RNA was observed. In comparison, the strongest Ifit1 protein signal was observed with 5′ cap⁺/2′-O Me⁻ RNA. These findings suggest that Ifit1 preferentially binds to 5′ capped RNA lacking 2′-O methylation.

Fig 4.

Ifit1 preferentially binds to virus RNA lacking 2′-O methylation. (A) Electrophoretic mobility shift of biotin-labeled RNA (JEV 5′-terminal 200 nucleotides) with recombinant Ifit1. The presence or absence of a 5′ cap and 2′-O Me of the JEV 5′-terminal 200 nucleotides is indicated. Unlabeled 5′-PPP RNA was used as a competitor. The loss of the band indicates binding of RNA and Ifit1 (top). The band intensities (in percent) calculated by ImageJ are shown at the bottom. Data are representative (top) and means ± SDs (bottom) of five independent experiments. *, P < 0.05. (B) Lysates from HEK293T cells transfected with HA-tagged Ifit1 were incubated with 2.5 pmol of biotin-labeled RNA. The presence or absence of a 5′ cap and 2′-O Me of the JEV 5′-terminal 200 nucleotides is indicated. 5′ OH RNA was produced by incubating in vitro-transcribed RNA with CIAP. RNA was incubated with streptavidin beads, and the precipitates were separated by SDS-PAGE and immunoblotted with an anti-HA antibody (top). Med and Input, samples from whole-cell lysates of empty vector- and Ifit1-transfected 293T cells, respectively. The percent band intensities calculated by ImageJ are shown at the bottom. Data are representative (top) and means ± SDs (bottom) of three independent experiments. *, P < 0.05. (C) MEFs stably expressing Ifit1 (Ifit1^+/+ + Ifit1) or Ifit1^−/− MEFs were infected with JEV WT or the JEV K61A mutant at an MOI of 1.0. The cells were harvested after 24 h, and JEV RNA/Ifit1-binding complexes were immunoprecipitated with a mouse anti-Flag antibody or mouse IgG. The immunoprecipitated RNA was analyzed by nested RT-PCR using primers that detect the JEV NS1 gene. Each value was normalized by the value for the input (indicated in percent). Data are means ± SDs of three independent experiments. *, P < 0.05.

To confirm independently that Ifit1 interacts with 5′ capped RNA lacking 2′-O methylation, we performed RNA immunoprecipitation assays using cell lysates from JEV-infected MEFs that ectopically expressed a Flag-tagged Ifit1. After immunoprecipitation with an anti-Flag antibody, the JEV mRNA was measured by nested RT-PCR analysis (Fig. 4C). JEV RNA was only marginally detected in lysates precipitated with control IgG and lysates of Ifit1^−/− MEFs infected with the JEV K61A mutant, indicating the specificity of Ifit1 binding in the assay. Virus RNA in JEV K61A mutant-infected MEFs was detected at a level over 37-fold higher than that in JEV WT-infected MEFs. Taken together, these findings suggest that Ifit1 directly interacts with virus mRNA lacking 2′-O methylation.

Ifit1 selectively inhibits translation of 5′ capped 2′-O unmethylated mRNA.

To examine the mechanism by which Ifit1 exerts an antiviral effect by associating with mRNA lacking 2′-O methylation, we used a luciferase translational reporter assay. Luciferase RNAs with different 5′ structures were transfected into type I IFN-primed MEFs, and total RNA and cell lysates were harvested 6 h later. Importantly, the levels of luciferase RNAs in wild-type and Ifit1^−/− cells were unaffected by any of the 5′ modifications (Fig. 5A). We then analyzed the translational efficiency of the transfected RNAs by measuring the luciferase activity (Fig. 5B). As expected (1), uncapped 5′-PPP luciferase mRNA was not translated in either wild-type or Ifit1^−/− MEFs. Capping of the mRNA (5′ cap⁺/2′-O Me⁻) increased translation in wild-type cells, although the levels were profoundly lower (P < 0.05) than those in Ifit1^−/− cells. In comparison, addition of 2′-O methylation to the 5′ cap (5′ cap⁺/2′-O Me⁺) in vitro resulted in similar levels of translation in wild-type and Ifit1^−/− MEFs. Even in MEFs that were not treated with type I IFN, similar patterns of luciferase activity were observed (Fig. 5C), indicating that slightly expressed Ifit1 might contribute to the inhibition. Taken together, our data establish that Ifit1 preferentially binds to 5′ capped mRNA lacking 2′-O methylation and inhibits its translation.

Fig 5.

Ifit1 selectively inhibits the translation of mRNA lacking 2′-O methylation. (A) The luciferase RNA amounts at 6 h after RNA transfection were determined by quantitative real-time RT-PCR. The relative luciferase mRNA amounts, calculated as the amount of each transfected mRNA (luc2) divided by the level of GAPDH mRNA expression, are shown. The presence or absence of a 5′ cap and 2′-O Me of the introduced luciferase RNA is indicated. Data are shown as means ± SDs and are representative of three independent experiments. (B, C) Wild-type and Ifit1^−/− MEFs pretreated with type I IFN (B) or untreated (C) were transiently transfected with luciferase mRNA. Luciferase activities were measured at 6 h after the transfection and are shown as relative light units (RLU). The presence or absence of a 5′ cap and 2′-O Me of the introduced luciferase RNA is indicated. Data are shown as means ± SDs of triplicate samples of the representative results. Similar results were obtained in three independent experiments. *, P < 0.05.

DISCUSSION

In this study, we investigated the mechanisms by which Ifit1 recognizes RNA of JEV lacking 2′-O MTase activity. Ifit1 inhibited the translation of mRNA through association with mRNA lacking 2′-O methylation.

To analyze the role of Ifit1 in 5′ cap structure-dependent antiviral responses, we generated a JEV MTase mutant. The K61, D146, K182, and E218 residues have all been shown to be essential for the MTase activity of the NS5 protein and replication of WNV (8, 11). While a WNV E218A mutant was previously used for analysis of Ifit1-mediated antiviral responses (13), in our assays, the corresponding JEV E218A mutant was severely impaired in replication in Vero cells and rapidly reverted to the wild type during cell culture, preventing its use (data not shown). A similar phenotype was observed with the WNV D146A 2′-O methylation mutant (11). However, unlike our results, it has recently been reported that a JEV E218A mutant is stable in Vero cells (39). This would be due to the different strains used in the two studies. Thus, mutation of residues that are essential for the 2′-O MTase activity of a flavivirus NS5 protein can differentially impact replication of JEV and WNV even in cells lacking type I IFN responses and IFIT1 expression.

Previous in vitro studies indicated that IFIT family proteins bind to several types of RNA, including 5′-PPP RNA and AU-rich double-stranded RNA (31, 32). Indeed, an analysis of the IFIT2 crystal structure indicated the presence of a positively charged RNA-binding channel (31), findings which were supported by the X-ray crystallographic structure of complexes of 5′-PPP RNA with human IFIT5 (33, 40). We also observed that Ifit1 could bind to 5′-PPP RNA. However, our biochemical analysis showed that Ifit1 bound strongly to 5′ capped RNA lacking 2′-O methylation and addition of 2′-O methylation weakened the binding of Ifit1 to the RNA. Since mRNAs of virtually all higher eukaryotes are believed to be methylated at the ribose 2′-O position (41), this modification likely serves as a molecular pattern for discriminating self from nonself.

Although it remains unclear how 2′-O methylation reduces Ifit1 binding to RNA, structural changes to the RNA at the 5′ terminus after 2′-O methylation could sterically hamper Ifit1 binding. The crystal structure of the 5′-PPP RNA/IFIT5 complex has indicated that the RNA-binding site on human IFIT5 is located in a narrow pocket, thus raising the possibility that 5′ capped and 2′-O methylated RNA cannot fit within an analogous pocket of Ifit1 due to a size limitation (33). Future structural analyses of the binding complex of 5′ capped RNA with Ifit1 will be required to reveal the precise mechanisms by which Ifit1 recognizes 5′ capped RNA lacking 2′-O methylation. Additional studies must also test whether other IFITs preferentially associate with 5′ capped RNA lacking 2′-O methylation.

Ifit1 also has an antiviral activity against several negative-stranded viruses, such as vesicular stomatitis virus (VSV) and parainfluenza virus type 5 (PIV5) (32, 42), whose mRNAs are 2′-O methylated (6, 42). In this regard, Ifit1 is supposed to have an antiviral effect independent of 2′-O methylation. Indeed, IFIT1 is able to bind 5′-PPP genomic RNA (32).

Given the previous and present findings that Ifit1 inhibits mRNA translation (23–26), our data are most consistent with a model in which Ifit1 restricts replication of viruses with 5′ capped RNA lacking 2′-O methylation through direct RNA binding and subsequent inhibition of translation. Human IFIT1 and murine Ifit1 were previously reported to interact with eIF3 to interfere with translation (23–26), and replication of hepatitis C virus, whose RNA lacks a 5′ cap, was also impaired by IFIT1 through binding to eIF3 (43). Thus, Ifit1 may associate with both eIF3 and virus mRNA to inhibit translation and infection.

The Ifit family proteins consist of several conserved members. However, Ifit1 and Ifit2 appear to have distinct antiviral activities (44). Thus, the nonredundant and redundant roles of the Ifit family proteins remain to be elucidated. Generation of mice lacking the other members or all of the Ifit family proteins will be useful to reveal the physiological functions.