Role of N Terminus-Truncated NS1 Proteins of Influenza A Virus in Inhibiting IRF3 Activation

ABSTRACT

The NS1 protein encoded by influenza A virus antagonizes the interferon response through various mechanisms, including blocking cellular mRNA maturation by binding the cellular CPSF30 3′ end processing factor and/or suppressing the activation of interferon regulatory factor 3 (IRF3). In the present study, we identified two truncated NS1 proteins that are translated from internal AUGs at positions 235 and 241 of the NS1 open reading frame. We analyzed the cellular localization and function of the N-truncated NS1 proteins encoded by two influenza A virus strains, Udorn/72/H3N2 (Ud) and Puerto Rico/8/34/H1N1 (PR8). The NS1 protein of PR8, but not Ud, inhibits the activation of IRF3, whereas the NS1 protein of Ud, but not PR8, binds CPSF30. The truncated PR8 NS1 proteins are localized in the cytoplasm, whereas the full-length PR8 NS1 protein is localized in the nucleus. The infection of cells with a PR8 virus expressing an NS1 protein containing mutations of the two in-frame AUGs results in both the absence of truncated NS1 proteins and the reduced inhibition of activation of IRF3 and beta interferon (IFN-β) transcription. The expression of the truncated PR8 NS1 protein by itself enhances the inhibition of the activation of IRF3 and IFN-β transcription in Ud virus-infected cells. These results demonstrate that truncated PR8 NS1 proteins contribute to the inhibition of activation of this innate immune response. In contrast, the N-truncated NS1 proteins of the Ud strain, like the full-length NS1 protein, are localized in the nucleus, and mutation of the two in-frame AUGs has no effect on the activation of IRF3 and IFN-β transcription.

IMPORTANCE Influenza A virus causes pandemics and annual epidemics in the human population. The viral NS1 protein plays a critical role in suppressing type I interferon expression. In the present study, we identified two novel truncated NS1 proteins that are translated from the second and third in-frame AUG codons in the NS1 open reading frame. The N-terminally truncated NS1 encoded by the H1N1 PR8 strain of influenza virus that suppresses IRF3 activation is localized primarily in the cytoplasm. We demonstrate that this truncated NS1 protein by itself enhances this suppression, demonstrating that some strains of influenza A virus express truncated forms of the NS1 protein that function in the inhibition of cytoplasmic antiviral events.

INTRODUCTION

Infections from influenza A virus can result in severe respiratory complications and death in humans. The viral RNA genome is comprised of eight negative-sense RNA segments that encode PB1, PB2, PA, NP, HA, NA, M1, and NS1 viral proteins. In addition to these eight proteins, two viral proteins, M2 and NS2, are encoded by alternatively spliced transcripts of M1 and NS1 mRNAs, respectively.

Additional viral proteins are produced from polymerase protein mRNAs by the initiation of translation at internal AUGs or by frameshifting. The PB1-F2 protein is translated from the +1 open reading frame (ORF) of the PB1 mRNA of several strains of influenza A virus (1), and the PB1-N40 protein is an N terminus-truncated PB1 protein (2). In addition, the PA segment encodes three novel PA-related proteins, denoted PA-X, PA-N155, and PA-N182 (3, 4). The PA-X protein, which is generated by frameshifting at codons 190 to 193, is comprised of the endonuclease domain of the PA protein and 61 amino acids from the +1 reading frame at its C terminus and is involved in regulating the host response to influenza A virus infection (3). The PA-N155 and PA-N182 proteins are N terminus-truncated PA proteins initiated from the 11th and 13th AUG codons, respectively, in PA mRNA. These truncated PA proteins were predicted to have a role in influenza A virus replication and viral pathogenicity, although the detailed mechanisms have yet to be elucidated (4).

An additional protein also is encoded by the M1 mRNA, specifically, a protein denoted as M42 is encoded by an alternatively splicing transcript of M1 mRNA, mRNA4 (5, 6). This splicing event is enhanced by the presence of a U148A mutation in the M1 mRNA. The M42 protein is translated from a +2 reading frame of mRNA4, has the same amino acid sequence from positions 12 to 99 as the M2 ion channel protein, and compensates for a null-M2 mutation (6).

The NS segment of influenza A virus typically encodes two proteins, NS1 and the nuclear export protein NS2. In addition, Selman et al. discovered a novel alternatively spliced product of NS1 mRNA that encodes a deleted NS1 protein, designated NS3, in certain influenza A virus strains that contain the A374G nucleotide substitution in their NS1 ORFs (7). The NS1 protein, which is abundantly expressed in influenza A virus-infected cells, is a small (219 to 237 amino acids long) multifunctional protein. It is divided into two functional domains, an N-terminal double-stranded RNA (dsRNA)-binding domain and a C-terminal effector domain (8). The dsRNA binding domain is comprised of 1 to 73 amino acids and forms a symmetric homodimer. The major function of this RNA-binding activity is the suppression of the activation of the 2′-5′-oligo(A) synthetase/RNase L pathway (9). The effector domain (comprising amino acid 85 to the end of the protein) has been shown to bind numerous cellular proteins (10). The effector domains of NS1 proteins encoded by H3N2, H2N2, and seasonal H1N1 viruses strongly bind to the 30-kDa subunit of the cleavage and polyadenylation specificity factor (CPSF30), thereby inhibiting the 3′ end processing of host pre-mRNAs and the production of mature mRNAs, including beta interferon (IFN-β) and other antiviral mRNAs (11–13). However, the NS1 proteins encoded by some strains of influenza A viruses, such as Puerto Rico/8/34/H1N1 (PR8) and the 2009 pandemic H1N1 viruses, do not efficiently bind CPSF30 (13–15). Consequently, these viruses depend solely on the ability of their NS1 proteins to efficiently suppress the activation of interferon regulatory factor 3 (IRF3) to inhibit the expression of type I IFN. The NS1 proteins of some virus strains bind CPSF30 and also inhibit IRF3 activation. However, the NS1 proteins of circulating H3N2, H2N2, and a subset of H1N1 viruses encode NS1 proteins that do not inhibit IRF3 activation (16). The lack of inhibition is correlated with the presence of K (rather than E) at position 196 in the NS1 protein, and changing the K to E leads to a virus that partially inhibits IRF3 activation (16). It is not known why one of these two mechanisms for inhibiting the production of IFN suffices for some viruses, whereas other viruses employ both mechanisms. All of these NS1 proteins, whether or not they inhibit IRF3 activation, bind TRIM25 (16, 17), demonstrating that TRIM25 binding does not necessarily lead to the inhibition of IRF3 activation. Indeed, it has not been definitively established how some NS1 proteins inhibit IRF3 activation.

Here, we determined whether NS1 mRNA, like several other influenza viral mRNAs, encodes additional proteins by utilizing one or more downstream AUGs to initiate translation. We identified such N-terminally truncated proteins translated from the second and/or third in-frame AUGs of the NS1 ORFs of two influenza A virus strains, Udorn/72/H3N2 (Ud) and PR8. The N-terminally truncated Ud NS1 proteins function like the full-length NS1 protein: they are localized primarily in the nucleus and bind CPSF30. In contrast, the N-terminally truncated PR8 NS1 proteins, unlike the full-length protein, are localized primarily in the cytoplasm. Here, we demonstrate that the truncated PR8 NS1 protein plays an important role in infected cells by contributing to the inhibition of IRF3 activation.

MATERIALS AND METHODS

Cells, viruses, and viral infections.

The A549, MDCK, HEL-299, HeLa, and 293T cell lines were cultivated in Dulbecco's modified Eagle medium (DMEM) (Life Technologies, USA) with 10% fetal bovine serum. Influenza A virus stocks of Ud and PR8 strains were amplified using 10-day-old embryonic chicken eggs and then titrated using a plaque formation assay with MDCK cells. The mutant viruses were generated using the reverse genetics technique, as described previously (18, 19). Influenza A virus replication was detected using A549 cells at the indicated MOI by employing a previously described method (13). The HEL-299 and A549 cells were applied in assaying the IFN-β response after influenza A virus infection at the indicated multiplicity of infection (MOI).

Immunoprecipitation.

The 293T cells were cotransfected with the indicated NS1-expressing plasmid and a 3×FLAG CPSF30-expressing plasmid, a FLAG-tagged RIG-I-expressing plasmid, or an empty vector as the control. All of the expressing vectors were in the pcDNA3 plasmid backbone. At 24 h posttransfection, the cells were lysed using a buffer containing 100 mM Tris-HCl (pH 7.5), 250 mM NaCl, 0.5% sodium deoxycholate, 0.5% NP-40, and 1 mM phenylmethylsulfonyl fluoride (PMSF). The cell extracts then were subjected to immunoprecipitation using an anti-FLAG M2 affinity gel (Sigma-Aldrich, USA) in accordance with the manufacturer's instructions. After washing the affinity gel with a buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 0.5% NP-40 4 times, the proteins that remained on the affinity gel were eluted using an SDS-PAGE sample buffer or 3×FLAG peptide for use in further analysis.

Immunoblot assay and antibodies.

The protein extracts of transfected or influenza A virus-infected cells were prepared, separated using SDS-PAGE, and transferred to PVDF membranes, as described previously (20). The antibodies against influenza A virus NS1 and NP proteins were generated by rabbits that were immunized with purified WSN NS1 and NP proteins, respectively. Other antibodies were anti-total IRF3 (Santa Cruz Biotechnology, USA), anti-phospho-IRF3 (Ser396; Cell Signaling, USA), anti-β-actin (Sigma-Aldrich, USA), and horseradish peroxidase (HRP)-conjugated secondary antibodies (GE Healthcare, USA).

In vitro translation assay.

The wild-type (WT) or mutated NS1 ORFs from Ud and PR8 viruses in the pcDNA3 vector backbone were linearized using a restriction enzyme, purified using the QIAquick gel extraction kit (Qiagen, Germany), and then extracted using a mixture of phenol, chloroform, and isoamyl alcohol. In total, 1 μg of purified linear plasmid DNA was mixed into a TNT coupled reticulocyte lysate system (Promega, USA) with T7 RNA polymerase and [³⁵S]methionine/[³⁵S]cysteine in accordance with the manufacturer's protocol. The mixture then was incubated at 30°C for 90 min. After the product of the transcription/translation couple reaction was analyzed using 13% SDS-PAGE, the gel was dried and exposed to X-ray film.

Immunofluorescence and confocal microscopy.

The HeLa cells were grown on coverslips for 24 h and then transfected with the full-length NS1-expressing plasmids, plasmids encoded with N-truncated NS1 proteins (amino acids 79 to 237 or 230 for the Ud virus; amino acids 79 to 230 for the PR8 virus), or an empty vector. The transfected cells then were fixed with 4% paraformaldehyde at 24 h posttransfection. After being washed with phosphate-buffered saline (PBS), the cells were treated with 0.3% Triton X-100 and then blocked with 5% bovine serum albumin for 1 h. The rabbit anti-NS1 polyclonal antibody and anti-rabbit IgG conjugated with Alexa Fluor 594 or Alexa Fluor 488 dye (Life Technologies, USA) was applied to the treated cells to detect NS1 proteins. After 4′,6-diamidino-2-phenylindole (DAPI) nuclear counterstaining, the localization of the NS1 protein was observed under a confocal microscope.

Quantitative reverse transcription-PCR (RT-PCR) for detecting IFN-β pre-mRNA and mRNA.

At 9 h after influenza virus infection, the total RNA of infected cells was isolated using TRIzol reagent (Invitrogen, USA). To determine the relative amounts of IFN-β pre-mRNA and mRNA, the total RNA was subjected to a reverse transcription reaction with an oligonucleotide complementary to sequence downstream of the poly(A) addition site of IFN-β pre-mRNA and an oligo(dT) primer using SuperScript III reverse transcriptase (Life Technologies, USA) (15, 20), and quantitative PCR (qPCR) was conducted subsequently to detect the pre-mRNA of IFN-β and the mRNAs of IFN-β and β-actin using Kapa SYBR fast qPCR kits (Kapa Biosystems, USA) with primers described previously (15, 20) and a Roche LightCycler 480II instrument.

IFN-β reporter assay.

The 293T cells were cotransfected with a firefly luciferase reporter plasmid containing an IFN-β promoter (provided by Michael Gale, Jr.), a pRL-TK renilla luciferase control plasmid (for normalizing the transfection efficiency), a Myc-tagged RIG-I-CARD-expressing plasmid (provided by Helene M. Liu), and a plasmid that expressed WT or N-terminally truncated PR8 NS1 (amino acids 79 to 230). At 24 h posttransfection, the firefly and renilla luciferase activities of transfected cells were determined using a dual-luciferase reporter assay (Promega, USA) according to the manufacturer's instructions. The normalized firefly luciferase activity represented the promoter activity of IFN-β. 293T cells were transfected with the reporters and the NS1-expressing plasmids for 18 h and then infected with Sendai virus (provided by Lih-Hwa Huang). At 24 h postinfection, the luciferase activity assay was performed as described above.

Establishment of A549 cells that constitutionally express the truncated PR8 NS1 protein.

The 293T cells were cotransfected with the plasmids pCMV-deltaR8.91 and pMD.G and a lentiviral transfer vector that expresses the PR8 truncated NS1 protein in the pLAS2w.Ppuro backbone to generate a recombinant lentivirus. The lentivirus-based vectors for gene expression were obtained from the RNA interference (RNAi) core laboratory, Academia Sinica, Taiwan. At 48 h posttransfection, the supernatants of the transfected 293T cells were collected and then added to A549 cells. After transduction for 24 h, the A549 cells were treated with puromycin at a concentration of 5 μg/ml. The A549 cells that survived from puromycin treatment were maintained in DMEM with 10% fetal bovine serum (FBS) for further studies.

Mouse experiments.

C57BL/6 mice were purchased from BioLasco (Taiwan) and maintained in the specific-pathogen-free animal facilities in Chang Gung University by following the guidelines of the Animal Care and Use of Laboratory Animals of the Taiwanese Council of Agriculture. The mice were infected with 200 PFU of the PR8 WT or mutant virus intranasally as described previously (21).

RESULTS

N-terminally truncated NS1 proteins are translated from the second and third in-frame AUGs in the NS1 ORF.

We previously forecasted two downstream in-frame AUGs that might initiate the synthesis of N-truncated NS1 proteins (21). The third in-frame AUG, located at bases 241 to 243 (codon 81), is predicted to be an efficient translation initiation site according to Kozak's rules (22) (Fig. 1A). There is also an in-frame AUG located at nucleotides 235 to 237 (the second start codon; codon 79), which is predicted to be a less efficient translation initiation site (22). Translation initiated at these downstream in-frame AUGs would generate two polypeptides similar in length, comprised of 159 and 157 amino acids (based on the Ud NS1 protein sequence) and a molecular mass of approximately 17 kDa.

FIG 1.

N terminus-truncated NS1 protein is translated from the second or third in-frame AUG codon of NS1 ORF of influenza A virus. (A) A diagram of the NS1 ORF of the Ud strain of influenza A virus. (B) 293T cells were transfected with a plasmid expressing WT Ud NS1, a Ud NS1 plasmid that is mutated at codons 79 and 81 (ATG to ATT), Ud NS1 plasmids expressing ATG-to-ATT mutation at either codon 79 or 81, or a plasmid expressing Ud NS1 amino acids 79 to 237 (as a length control). (C) Linearized plasmids that encode WT Ud NS1, Ud NS1 with mutations at codons 79 and 81, WT PR8 NS1, or mutated PR8 NS1 were applied to in vitro transcription/translation-coupled reactions with [³⁵S]methionine/[³⁵S]cysteine. The translated products were analyzed with SDS-PAGE and autoradiography. tNS1, truncated NS1 protein.

To determine whether these downstream AUGs generate the predicted truncated NS1 proteins, 293T cells were transfected with a plasmid expressing a WT Ud NS1 protein or a plasmid expressing a Ud NS1 protein that was mutated in the second and third ATG, i.e., with a replacement of ATG with ATT (denoted M79.81I). At 48 h posttransfection, cell extracts were analyzed by immunoblots probed with an anti-NS1 polyclonal antibody. A 17-kDa protein band was found in the WT-transfected cells but not in the M79.81I-transfected cells (Fig. 1B, lanes 2 and 3), demonstrating that a 17-kDa NS1-related polypeptide was produced by the initiation of translation at the second and/or third in-frame AUG. To determine whether translation was initiated at one or both of these downstream AUGs, 293T cells were transfected with a plasmid expressing a Ud NS1 containing an ATT replacement at either the second or third ATG, denoted M79I and M81I, respectively. An immunoblot detected a 17-kDa polypeptide in cells transfected with either plasmid, demonstrating that the initiation of translation can occur at both the second and third in-frame AUG.

As verification, we carried out coupled transcription-translation experiments in vitro using [³⁵S]methionine and [³⁵S]cysteine to radiolabel the synthesized proteins (Fig. 1C). The WT Ud mRNA directed the synthesis of both full-length NS1 protein and a 17-kDa protein, whereas the M79.81I mutant mRNA directed the synthesis of only the full-length NS1 protein (lanes 2 and 3), confirming that the synthesis of the17-kDa protein is initiated at the internal AUGs of the NS1 mRNA. The same results were obtained with the NS1 mRNA of another virus, PR8 virus (lanes 4 and 5), demonstrating that synthesis of a 17-kDa protein also is initiated at the internal AUGs of PR8 NS1 mRNA.

N-terminal truncated NS1 proteins initiated at the second and third downstream AUGs are synthesized in virus-infected cells.

To investigate whether the truncated NS1-related proteins are translated using the downstream AUGs of the NS1 ORF in cells infected with influenza A virus, we generated Ud viruses that express NS1 proteins containing mutations in the second and/or third AUG. A549 cells were infected with each of these viruses and WT virus. As shown in Fig. 2A, the truncated 17-kDa NS1 protein was produced at 6 and 9 h after infection with WT Ud virus (lanes 7 and 12), whereas the17-kDa virus was not detected in cells infected with the Ud virus whose NS1 contained mutations at both the second and third AUGs (M79.81I virus) (lanes 10 and 15). These results demonstrate that the 17-kDa protein is synthesized in infected cells utilizing the second and third downstream AUG initiation sites. Although mutating the second or third in-frame start codon of the NS1 ORF (M79I or M81I virus, respectively) reduced the production of the truncated NS1 protein (lanes 8, 9, 13, and 14), the mutation in the third AUG (M81I virus) led to a greater reduction in the synthesis of the truncated NS1 protein (compare lanes 13 and 14).

FIG 2.

Mutations at the second and third in-frame AUG codons abolish the expression of the truncated NS1 protein in infected cells. A549 cells were infected with WT or mutated Ud viruses that express NS1 with codon 79 or 81 single mutations or 79 and 81 double mutations (A) or with WT or mutated PR8 viruses that are mutated at codons 79 and 81 in the NS1 protein (B) at an MOI of 2. At 3, 6, and 9 h postinfection, cell extracts were collected and analyzed by immunoblotting with anti-NS1 and anti-β actin antibodies.

We also generated a PR8 virus which expresses an NS1 protein containing mutations at the second and third AUGs (M79.81I virus). Whereas the 17-kDa truncated NS1 protein was produced in WT virus-infected cells, it was not detected in cells infected with the PR8 mutant virus (Fig. 2B, lanes 6 and 9), verifying that the 17-kDa protein is synthesized in infected cells utilizing the second and/or third downstream AUG initiation sites of NS1 mRNA.

The localization of the N-terminally truncated NS1 Ud and PR8 proteins.

To explore the biological function of 17-kDa N terminus-truncated NS1 proteins, we determined their sites of cellular localization. For this purpose, we constructed plasmids that express either amino acids 79 to 237 of the NS1 protein from the Ud strain or amino acids 79 to 230 of NS1 protein from the PR8 strain and determined the cellular localization of these two N terminus-truncated NS1 proteins in transfected HeLa cells using immunofluorescence and confocal microscopy. Like the localization of the WT full-length NS1 protein of the Ud virus, the truncated Ud NS1 protein containing the amino acids 79 to 237 was localized principally in the nucleus (Fig. 3F and I). This result is consistent with the presence of a second nuclear localization sequence in the C terminus (amino acids 219 to 237) of the Ud NS1 protein (22). A truncated Ud protein lacking the 7 C-terminal amino acids (231 to 237) also was located primarily, though not entirely, in the nucleus (Fig. 4A), consistent with a previous study (23). Consequently, because the NS1 proteins of H3N2 viruses isolated since 1987 lack the 231 to 237 sequence, the truncated NS1 proteins produced by these viruses would be expected to be located primarily in the nucleus. The N terminus-truncated Ud NS1 protein binds CPSF30 (Fig. 4B), indicating that it inhibits the 3′-end processing of cellular pre-mRNAs, like the full-length Ud NS1 protein, whereas the PR8 NS1 does not bind CPSF30 (Fig. 4C).

FIG 3.

Localization of N-truncated NS1 proteins. HeLa cells were transfected with plasmids expressing WT or truncated (containing amino acids 79 to 237) Ud NS1 or plasmids expressing WT or truncated (containing amino acids 79 to 237) PR8 NS1. At 24 h posttransfection, cells were fixed and stained with anti-NS1 antibody and DAPI. The localizations of NS1 proteins were observed under confocal microscopy. (Upper) Schematic representation of the domains and nuclear localization signals of Ud NS1 protein.

FIG 4.

N-truncated NS1 proteins of Ud virus can be localized in nucleus and bind to CPSF30. (A) HeLa cells were transfected with plasmids expressing truncated (containing amino acids 79 to 237 or 79 to 230) Ud NS1 proteins. At 24 h posttransfection, cells were fixed and stained with anti-NS1 antibody and DAPI. The localizations of tNS1 proteins were observed under confocal microscopy. (Upper) Schematic representation of the Ud N-truncated NS1 proteins and the predicted nuclear localization signal. (B) 293T cells were cotransfected with plasmid expressing either truncated NS1 protein (containing amino acids 79 to 237 derived from Ud virus) or full-length NS1 protein and 3× FLAG-tagged CPSF30-expressing plasmid. At 24 h posttransfection, cell extracts were collected and subjected to immunoprecipitation (IP) by anti-FLAG resin. The cell extracts (lanes 1 to 6) and the precipitated products (lanes 7 to 12) were analyzed by immunoblotting with anti-FLAG and anti-NS1 antibodies (Ab). (C) 293T cells were cotransfected with plasmid expressing PR8 NS1 protein and 3× FLAG-tagged CPSF30-expressing plasmid for 24 h. As previously described, the FLAG-tagged CPSF30 and the NS1 proteins in the cell extracts and anti-FLAG immunoprecipitates were analyzed by immunoblotting.

In contrast, the N terminus-truncated NS1 proteins containing the amino acids 79 to 230 of the PR8 virus were localized primarily in the cytoplasm (Fig. 3L and O), whereas the full-length PR8 NS1 protein was localized mainly in the nucleus, consistent with the absence of a second nuclear localization sequence in the PR8 NS1 protein (23).

Elimination of the N-truncated PR8 NS1 proteins reduces the inhibition of the activation of IRF3 and IFN transcription.

Because the N-terminally truncated PR8 NS1 protein is localized primarily in the cytoplasm, it might enhance cytoplasmic functions of the full-length PR8 NS1, particularly its inhibition of IRF3 activation and, as a consequence, inhibition of the activation of IFN-β transcription. To test this hypothesis, A549 cells were infected with either the WT or M79.81I mutant PR8 virus, and cell extracts collected at 9 h postinfection were assayed for the presence of activated (phosphorylated) IRF3 (Fig. 5A). The suppression of IRF3 activation in cells infected with WT PR8 virus was relieved in cells infected with the mutant PR8 virus that does not produce truncated NS1 proteins (Fig. 5A, compare lanes 2 and 3), although the activation of IRF3 was not as strong as that in Ud virus-infected cells. The level of IFN-β transcription was measured by determining the amount of IFN-β pre-mRNA by quantitative RT-PCR as previously described (Fig. 5B). The IFN-β transcription was activated by 2.5-fold in mutant PR8 virus-infected cells, resulting in the increased production of IFN-β mRNA. Similar results were obtained in another cell line, HEL-299 cells (Fig. 5C and D).

FIG 5.

Mutations at the second and third in-frame AUG codons in NS1 ORF of PR8 virus increase IRF3 activation and IFN-β production in infected cells. A549 cells were infected with WT or mutated influenza A virus that expresses NS1 with mutations at the second and third AUG (codons 79 and 81, respectively) in the PR8 backbone (A and B) or Ud backbone (E and F) at an MOI of 2. (A and E) At 9 h postinfection, cell extracts were collected and analyzed by immunoblotting with anti-NS1, anti-phospho-IRF3 (serine 396), anti-total IRF3, and anti-β actin antibodies. (B and F) Relative amount of IFN-β mRNA of infected cells was quantitated by real-time RT-PCR. HEL-299 cells were infected with the WT or the PR8 virus mutated at codons 79 and 81 at an MOI of 2. (C) At 9 h postinfection, cell extracts were collected and analyzed for the proteins of NS1, phosphorylated IRF3, and total IRF3. (D) Relative amount of IFN-β pre-mRNA and mRNA of infected cells was quantitated by real-time RT-PCR. tNS1, truncated NS1 protein. *, P < 0.05; ***, P < 0.005. n.s., no significant difference.

Since it has been proposed that binding of TRIM25 is responsible for suppression of IRF3 activation by the NS1 protein (17), we determined whether the mutations of M79I and M81I affect the TRIM25 binding ability of the NS1 protein. 293T cells were transfected with a plasmid that expresses HA-tagged TRIM25 for 24 h, followed by infection of PR8 WT or M79.81I mutant at an MOI of 2. At 12 h postinfection, cell lysates were subjected to immunoprecipitation by anti-NS1 antibody. As shown in Fig. 6, the mutations M79I and M81I did not change the interaction of TRIM25 and the NS1 protein.

FIG 6.

Determination of the interaction of TRIM25 and the NS1 protein containing M79I and M81I mutations encoded by a mutant PR8 virus. 293T cells were transfected with plasmids expressing HA-tagged TRIM25 protein. At 24 h posttransfection, the transfected cells were infected with either the WT or NS1-mutated PR8 virus at an MOI of 2. At 12 h postinfection, cell extracts were collected and subjected to immunoprecipitation by anti-NS1 antibody. The HA-tagged TRIM25 and NS1 protein were detected in the precipitates by immunoblotting with anti-HA and anti-NS1 antibodies.

The N-truncated PR8 NS1 protein by itself inhibits activation of IRF3 and IFN-β transcription.

We evaluated the ability of the truncated NS1 protein to suppress the RIG-I-activated IFN-β promoter. 293T cells were cotransfected with an IFN-β promoter fused to a firefly luciferase gene plasmid, a renilla luciferase control plasmid, a plasmid expressing Myc-tagged RIG-I CARD domains (RIG-I-CARDs), and a plasmid expressing N-truncated PR8 NS1 or an empty vector (as the control). At 24 h posttransfection, the transfected cells were lysed and analyzed by using a dual-luciferase activity assay. As shown in Fig. 7A, the N terminus-truncated NS1 protein of the PR8 virus inhibited the RIG-I-CARD-mediated activation of the IFN-β promoter as efficiently as the WT NS1 protein (Fig. 7A, lane 4). In addition, the truncated PR8 NS1 protein inhibited the IFN-β promoter activation stimulated by Sendai virus infection almost as efficiently as the WT NS1 protein (Fig. 7B, lane 4).

FIG 7.

N-terminally truncated NS1 protein of PR8 virus inhibits IFN-β promoter activation. (A) 293T cells were cotransfected with a reporter plasmid containing IFN-β promoter fused to the firefly luciferase gene, a renilla luciferase control plasmid, a plasmid expressing WT or truncated NS1 protein (containing amino acids 79 to 230 derived from PR8 virus), and either the plasmid expressing Myc-tagged RIG-I CARD domains or an empty vector. At 24 h posttransfection, cell extracts were collected and analyzed by dual-luciferase activity assay. (B) 293T cells were cotransfected with the aforementioned luciferase reporters and the plasmid expressing PR8 WT or truncated NS1 for 18 h, followed by Sendai virus (SeV) infection for another 24 h. IFN-β promoter activity was determined by dual-luciferase activity assay. A549 cells that constitutionally express RFP or PR8 truncated NS1 were infected with Sendai virus at a final concentration of 13 HAU/ml (C) or with PR/M79.81I (D) or Ud/M79.81I (E) mutant virus at an MOI of 0.2. At 9 h postinfection, the activation of IRF3 was examined by anti-phosphorylated IRF3 antibody. The relative levels of IFN-β mRNA or pre-mRNA in the infected cells were analyzed by real-time RT-PCR. Protein expression was monitored by immunoblotting with anti-NS1, β-actin, c-Myc, or total IRF3 antibodies. *, P < 0.05; **, P < 0.01; ***, P < 0.005.

To verify the role of N-truncated PR8 NS1 in inhibiting IRF3 activation, A549 cells that constitutionally express the 79- to 230-amino-acid-long fragment of PR8 NS1 (denoted A549-tNS1 cells) were constructed and infected with Sendai virus or with PR8/M79.81I or Ud/M79.81I mutant virus, which does not express the N-terminally truncated NS1 proteins. At 9 h postinfection, cell extracts were collected to analyze the activation of IRF3. Compared to the infection in control A549 cells that express the red fluorescent protein (RFP), virus infection induces less activated (phosphorylated) IRF3 in the A549-tNS1 cells (Fig. 7C to E). Consequently, the transcription of IFN-β, determined by measuring the production of IFN-β mRNA or pre-mRNA, was reduced in infected A549-tNS1 cells compared to that of the infection in A549-RFP cells (Fig. 7C to E). These results show that the N terminus-truncated NS1 protein of PR8 virus enhances the inhibition of the activation of IRF3 and IFN-β transcription in infected cells. In addition, the elimination of the production of the truncated PR8 NS1 proteins decreased virus replication compared to that of WT virus (Fig. 8A). The lower replication of the mutant PR8 virus also was observed in vivo (Fig. 8C).

FIG 8.

Effect of abolishment of the second and third in-frame AUGs in NS1 ORF on influenza A virus replication. A549 cells were infected with a WT or mutated virus that does not express the N-truncated NS1 (mutated at the second and third AUG codons of NS1 ORF) in the PR8 (A) or Ud backbone (B) at an MOI of 0.001. At 16, 24, 36, 48, 60, and 72 h postinfection, supernatants of infected cells were collected and titrated by plaque formation assay. (C) C57BL/6 mice were inoculated with 200 PFU of the PR8 WT or mutant virus intranasally. Lungs were harvested for determining virus titers at 7 days postinfection. *, P < 0.05; ***, P < 0.005.

The results with the Ud virus were different. As shown previously (16), Ud virus encoding its WT NS1 protein does not inhibit the activation of IRF3 and IFN-β transcription, and as shown here, the elimination of the production of the N-terminally truncated NS1 proteins did not have any effect on these activations (Fig. 5E and F) or on virus replication (Fig. 8B).

DISCUSSION

Previous studies have identified 14 to 17 proteins that are produced in infected cells by various influenza A virus strains (24). In the present study, we identified two additional viral proteins, namely, truncated NS1 proteins that are translated from the second and third in-frame AUG codons located at nucleotides 235 to 237 (codon 79) and 241 to 243 (codon 81), respectively, in the NS1 ORF of influenza A virus. The third in-frame AUG codon was predicted to be more efficient at translation initiation than the second in-frame AUG, a prediction that was supported by our results (Fig. 2). Almost all (95 to 99%) of H2N2 and H3N2 strains that have circulated in humans have both of these in-frame AUG codons, which is also the case for H1N1 strains circulating prior to 2009. Consequently, the two truncated NS1 proteins would be expected to be produced in cells infected by these human influenza A viruses. In contrast, the pandemic H1N1 (pH1N1) strains that have been circulating in humans since 2009 have only the second in-frame AUG codon, indicating that only one of the truncated NS1 proteins is expected to be produced in cells infected by these viruses. In addition, we analyzed the other start codons before the initiation sites of the truncated NS1 and identified 4 AUGs in the +2 frame of the NS1 gene. These AUGs are not in the best context for translation initiation according to Kozak's rules. Translation initiated from these four AUGs meets downstream in-frame stop codons in the 7th, 15th, 6th, and 6th codons, respectively. This indicates that initiation at these AUGs yields very small peptides. It is not known whether such small peptides are synthesized in infected cells.

Mutation of the two in-frame AUGs of the PR8 NS1 protein resulted in both the absence of the truncated NS1 proteins and reduced inhibition of the activation of IRF3 and IFN-β transcription. Because these mutations also result in methionine (M)-to-isoleucine (I) substitution at amino acid positions 79 and 81 in the full-length NS1 protein, it was possible that the reduced activation of IRF3 resulted from the inactivation of the full-length NS1 protein due to these amino acid changes. We tested the binding of TRIM25 (17) to the mutated PR8 NS1 and found that the mutations M79I and M81I did not change this interaction (Fig. 6), demonstrating that increased IRF3 activation and IFN-β expression in the mutant PR8 virus-infected cells was not caused by the loss of the binding of the full-length NS1 to TRIM25. More importantly, we demonstrated that the truncated PR8 NS1 protein by itself inhibits the activation of IRF3 and IFN-β transcription. In one approach, we used a reporter assay to demonstrate that the N terminus-truncated PR8 NS1 protein efficiently inhibited RIG-I-CARD-mediated activation of the IFN-β promoter. In the second approach, we generated a cell line that expresses the truncated PR8 NS1 protein and showed that the truncated PR8 NS1 protein expressed in these cells inhibits the activation of IRF3 and IFN-β transcription in cells infected by PR8 and Ud viruses with the M79I and M81I mutations. Based on these results, we concluded that the truncated PR8 NS1 protein, which is localized in the cytoplasm, contributes to the inhibition of the activation of this innate immune response that is initiated in the cytoplasm. The resulting activation of IRF3 and IFN-β transcription leads to a decrease in the replication of the PR8 virus that does not express the truncated NS1, most likely because the PR8 NS1 protein does not bind CPSF30 (Fig. 4C) and does not block the processing of IFN-β pre-mRNA to form mature IFN-β mRNA (25).

The truncated PR8 NS1 protein enhances the inhibition of IRF3 activation even though it lacks the N-terminal RNA-binding domain, indicating that the effector domain is largely responsible for inhibiting IRF3 activation. Interestingly, we did not detect binding of the truncated PR8 NS1 protein to RIG-I under conditions in which the full-length PR8 NS1 protein binds to RIG-I (data not shown), indicating that the interaction of the NS1 protein with RIG-I is mediated by the NS1 RNA-binding domain; hence, such an interaction of the NS1 protein with RIG-I is not absolutely required for the inhibition of IRF3 activation. A PR8 virus expressing an NS1 protein lacking most of the effector domain (99- to 230-amino-acid deletion) is severely attenuated (data not shown), which is similar to the results described in a previous publication (26), and the ability to suppress IRF3 activation was reduced dramatically (data not shown). These results are consistent with our previous finding that the identity of the amino acid (K or E) at position 196 of the NS1 protein plays a significant role in the ability of the NS1 protein to inhibit IRF3 activation (16).

In contrast, we did not detect any role of the truncated NS1 proteins of the Ud virus in the RIG-I pathway. The full-length Ud NS1 protein, which contains K at position 196, does not block IRF3 activation, and mutation of the two in-frame AUGs had no detectable effect on the activation of IRF3 and IFN-β transcription. It is possible that the truncated Ud NS1 proteins have a role in other NS1-mediated activities.