Functional Evolution of Influenza Virus NS1 Protein in Currently Circulating Human 2009 Pandemic H1N1 Viruses

ABSTRACT

In 2009, a novel H1N1 influenza virus emerged in humans, causing a global pandemic. It was previously shown that the NS1 protein from this human 2009 pandemic H1N1 (pH1N1) virus was an effective interferon (IFN) antagonist but could not inhibit general host gene expression, unlike other NS1 proteins from seasonal human H1N1 and H3N2 viruses. Here we show that the NS1 protein from currently circulating pH1N1 viruses has evolved to encode 6 amino acid changes (E55K, L90I, I123V, E125D, K131E, and N205S) with respect to the original protein. Notably, these 6 residue changes restore the ability of pH1N1 NS1 to inhibit general host gene expression, mainly by their ability to restore binding to the cellular factor CPSF30. This is the first report describing the ability of the pH1N1 NS1 protein to naturally acquire mutations that restore this function. Importantly, a recombinant pH1N1 virus containing these 6 amino acid changes in the NS1 protein (pH1N1/NSs-6mut) inhibited host IFN and proinflammatory responses to a greater extent than that with the parental virus (pH1N1/NS1-wt), yet virus titers were not significantly increased in cell cultures or in mouse lungs, and the disease was partially attenuated. The pH1N1/NSs-6mut virus grew similarly to pH1N1/NSs-wt in mouse lungs, but infection with pH1N1/NSs-6mut induced lower levels of proinflammatory cytokines, likely due to a general inhibition of gene expression mediated by the mutated NS1 protein. This lower level of inflammation induced by the pH1N1/NSs-6mut virus likely accounts for the attenuated disease phenotype and may represent a host-virus adaptation affecting influenza virus pathogenesis.

IMPORTANCE Seasonal influenza A viruses (IAVs) are among the most common causes of respiratory infections in humans. In addition, occasional pandemics are caused when IAVs circulating in other species emerge in the human population. In 2009, a swine-origin H1N1 IAV (pH1N1) was transmitted to humans, infecting people then and up to the present. It was previously shown that the NS1 protein from the 2009 pandemic H1N1 (pH1N1) virus is not able to inhibit general gene expression. However, currently circulating pH1N1 viruses have evolved to encode 6 amino acid changes (E55K, L90I, I123V, E125D, K131E, and N205S) that allow the NS1 protein of contemporary pH1N1 strains to inhibit host gene expression, which correlates with its ability to interact with CPSF30. Infection with a recombinant pH1N1 virus encoding these 6 amino acid changes (pH1N1/NSs-6mut) induced lower levels of proinflammatory cytokines, resulting in viral attenuation in vivo. This might represent an adaptation of pH1N1 virus to humans.

INTRODUCTION

Influenza A (IAV) and B (IBV) viruses are members of the Orthomyxoviridae family and contain a segmented genome containing eight single-stranded RNA molecules with negative polarity (1). Despite comprehensive vaccination programs, the World Health Organization (WHO) estimates that the global disease burden from seasonal influenza results in 1 billion infections, 3 to 5 million cases of severe disease, and between 300,000 and 500,000 deaths annually (2). Whereas IBVs are principally human pathogens, diverse IAVs are maintained in a wide range of animal species, including wild aquatic birds, domestic poultry, and mammals, such as swine, dogs, cats, and horses, and can cause zoonotic infections in humans. Actually, in April 2009, a quadruple-reassortant swine-origin H1N1 IAV was transmitted from swine to humans (3, 4) and was spread rapidly around the world, leading to the declaration of a global pandemic by the WHO on 11 June 2009 (5). Since then, the pandemic H1N1 (pH1N1) virus has remained in the human population, and in the 2015-2016 season, it was the predominant virus infecting humans (https://www.cdc.gov/flu/about/season/flu-season-2015-2016.htm).

Host innate immune responses restrict influenza virus replication (6). Pathogen-associated molecular patterns (PAMPs) are recognized in infected cells by pattern recognition receptors (PRRs), which initiate signaling pathways leading to the production of type I and III interferons (IFNs) and proinflammatory cytokines (6). Influenza virus double-stranded RNA (dsRNA) is recognized by the membrane-associated PRR Toll-like receptor 3 (TLR-3) (7), influenza virus single-stranded RNA (ssRNA) is recognized by TLR-7 (8), and virus-specific RNAs are recognized by the cytoplasmic PRR retinoic acid-inducible gene I (RIG-I) (9) and the NOD-like receptor family member LRR and Pyrin domain containing 3 (NLRP3) (6, 10, 11). TLR recognition activates transcription factors, such as IFN regulatory factor 3 (IRF3), NF-κB, and activating transcription factor 2 (ATF-2)/c-Jun, that are responsible for the transcription of type I (IFN-α and IFN-β) and type III (IFN-λ) IFNs and proinflammatory cytokines (6, 12, 13). Secreted IFNs act in a paracrine and/or autocrine fashion to induce the expression of hundreds of IFN-stimulated genes (ISGs), many of which possess antiviral activity (6, 13, 14).

Influenza virus nonstructural protein 1 (NS1) is a multifunctional protein mainly involved in limiting IFN and proinflammatory responses, allowing the virus to efficiently replicate in infected cells (15–19). Influenza virus NS1 counteracts innate immune responses by different mechanisms, including inhibiting cellular transcription elongation (20); blocking posttranscriptional RNA processing and nuclear export (21); decreasing RIG-I activation through sequestration of its RNA helicase and its activating ligand (22–24) or through inhibition of tripartite motif family 25 (TRIM25)-mediated RIG-I ubiquitination (25); interfering with IFN signaling and directly inhibiting specific ISGs, such as protein kinase R (PKR) (26) and RNase L (27); and binding to dsRNA and PKR directly (reviewed in reference 18). Furthermore, NS1 proteins block pre-mRNA processing and the nuclear export of mRNAs (21), therefore leading to a general inhibition of host gene expression, including that of IFN genes, ISGs, and proinflammatory genes. To this end, the NS1 proteins of some human IAV strains bind to the 30-kDa subunit of the cleavage and polyadenylation specificity factor (CPSF30), blocking the processing of mRNAs (28–31), and to poly(A)-binding protein II (PABPII), inhibiting the ability of PABPII to stimulate the synthesis of long poly(A) tails (32). The NS1 protein binding site for CPSF30 is centered around amino acid 186, and mutations in NS1 amino acids 184 to 188 (GLEWN) lead to a virus with highly impaired growth (30). In addition, amino acid residues 103 and 106 (29), 108, 125, and 189 (33), 106 (34), 103 and 106 (35), and 64, 189, and 194 (36, 37) are important for NS1 binding to CPSF30 for the A/PuertoRico/8/34 (PR8), pH1N1, H7N9, H5N1, and seasonal H3N2 IAVs, respectively.

The 2009 pH1N1 virus is slightly more pathogenic than seasonal H1N1 viruses in different animal systems, replicating to higher titers and deeper in the lungs than the case for seasonal influenza virus and causing more pathological lesions in the lungs of mice, ferrets, and nonhuman primates than those seen with seasonal IAVs (38–41). Besides this, the pH1N1 virus induced higher levels of proinflammatory cytokines in mouse and cynomolgus macaque lungs, probably contributing to the higher level of pathogenesis induced by the virus (39).

In this work, we present, for the first time, data showing that unlike the original pH1N1 NS1 protein, the NS1 protein from currently circulating (now seasonal) pH1N1 viruses has evolved to contain six amino acid mutations (E55K, L90I, I123V, E125D, K131E, and N205S) that restore NS1's ability to bind to CPSF30, therefore allowing inhibition of host gene expression. Remarkably, a recombinant pH1N1 virus encoding the 6 amino acid changes in the NS1 protein (pH1N1/NSs-6mut) grew similarly to a virus encoding the original NS1 protein (pH1N1/NSs-wt) in cultured cells and in vivo but induced lower levels of IFN and proinflammatory cytokines, most likely due to a general inhibition of gene expression mediated by binding to CPSF30. This lower level of inflammation induced by pH1N1/NSs-6mut correlates with the partially attenuated phenotype observed for this virus in mice and most likely provides a novel mechanism for the adaptation of pandemic IAVs to humans.

RESULTS

The NS1 proteins from currently circulating pH1N1 viruses carry 6 amino acid changes with respect to the original A/California/04/2009 pH1N1 virus.

To analyze whether the NS1 protein of currently circulating pH1N1 viruses is identical to the protein encoded by the A/California/04/2009 virus circulating at the origin of the 2009 pandemic, we analyzed the NS1 sequences of 9 clinical samples collected in the Rochester, NY, area during the 2015-2016 season (subjects ACU001, ACU004, ACU005, ACU007, ACU008, ACU009, ACU012, ACU017, and ACU022). The sequences from the 9 clinical isolates were identical at the amino acid level. Interestingly, 6 amino acid changes (E55K, L90I, I123V, E125D, K131E, and N205S) (Table 1) were identified in the NS1 protein (NS1-6mut) from these isolates compared to that from pH1N1 A/California/04/2009 circulating at the origin of the 2009 pandemic (NS1-wt).

TABLE 1.

Amino acid changes in NS1 proteins from influenza pH1N1 viruses circulating in the Rochester, NY, area and from globally circulating viruses

Virus	Amino acid at NS1 position:
	55	90	123	125	131	205
A/California/04/2009	E	L	I	E	K	N
Viruses from patients:
001	K	I	V	D	E	S
004	K	I	V	D	E	S
005	K	I	V	D	E	S
007	K	I	V	D	E	S
008	K	I	V	D	E	S
009	K	I	V	D	E	S
012	K	I	V	D	E	S
017	K	I	V	D	E	S
022	K	I	V	D	E	S
Viruses circulating globally	K	I	V	D	E	S
% of sequences with amino acid changea	100	100	99.72	93.96	98.3	99.73

^aPercentages of sequences encoding the indicated amino acid changes in viruses circulating globally during the 2015–2016 season.

To analyze whether the NS1 amino acid changes found in viruses isolated in the Rochester, NY, area are representative of viruses circulating globally, these NS1 sequences were compared to the publicly available NS1 sequences of human pH1N1 viruses isolated worldwide from October 2015 to September 2016, which were downloaded from the Influenza Research Database (1,477 sequences) (www.fludb.org). Of these 1,477 sequences, 100% contained the E55K mutation, 100% contained the L90I mutation, 99.72% contained the I123V mutation, 93.96% contained the E125D mutation, 98.3% contained the K131E mutation, and 99.73% contained the N205S mutation (Table 1), indicating that the viruses circulating in Rochester, NY, are representative of the viruses circulating globally. To assess the timeline over which these mutations were incorporated into NS1, the percentages of strains isolated each season whose NS1 sequences contained the different amino acids at a given position were plotted (n = 7,027) (Fig. 1). The timelines of amino acid changes were different for the different residues. The I123V mutation was the first to appear, and as early as the period of October 2009 to September 2010, more than 90% of the sequences contained this amino acid change (Fig. 1A). The L90I and N205S mutations appeared after that, with around 90% of the sequences containing these amino acid changes by the period of October 2012 to September 2013 (Fig. 1B). The E55K and K131E mutations appeared next, with more than 95% of the sequences containing these amino acid changes during the period of October 2014 to September 2015 (Fig. 1C). Lastly, the E125D amino acid change appeared, with more than 95% of the sequences containing that mutation in the period of October 2015 to September 2016 (Fig. 1D). Interestingly, 100% of the sequences from currently circulating pH1N1 viruses contain these 6 amino acid changes (Fig. 1). Taken together, the timeline results suggest that the mutations present in the NS1 proteins of currently circulating pH1N1 viruses were gradually selected (I123V, L90I-N205S, E55K-K131E, and E125D) during adaptation of the virus to humans.

FIG 1.

Frequencies of identified NS1 mutations in currently circulating pH1N1 viruses over time. Publicly available sequences in the Influenza Research Database were downloaded, and the frequencies of sequences containing the original amino acid (black) and the mutated amino acid (white) at positions 123 (A), 90 and 205 (B), 55 and 131 (C), and 125 (D) are presented according to the month and season of virus isolation.

The NS1 protein of currently circulating pH1N1 viruses is able to inhibit general host gene expression.

One of the mechanisms by which the IAV NS1 protein counteracts innate immune responses is through binding to CPSF30, leading to host gene expression suppression, including that of IFN genes and ISGs with antiviral activities (28, 30). However, it was previously shown that the NS1 protein from pH1N1 A/California/04/2009 (NS1-wt) does not inhibit general gene expression and does not bind to CPSF30 (33). To analyze whether the 6 amino acid changes in the NS1 protein (NS1-6mut) from currently circulating pH1N1 viruses restore NS1's ability to inhibit general gene expression, human 293T cells were transiently cotransfected with plasmids expressing green fluorescent protein (GFP) and Gaussia luciferase (Gluc) and with plasmids encoding the different NS1 variants under the control of a polymerase II promoter (42), or with an empty plasmid as an internal control (Fig. 2). Plasmids expressing the NS1 proteins from the influenza A/PuertoRico/8/34 (PR8) and influenza A/BrevigMission/01/1918 (1918) H1N1 viruses were used as controls that do not and do block host gene expression, respectively (29). In addition, a plasmid encoding the NS1 protein of a pH1N1 virus with the previously described amino acid changes R108K, E125D, and G189D (NS1-3mut) was used as an additional control (33). GFP expression was evaluated by fluorescence microscopy (Fig. 2A), and Gluc expression was quantitated in a Lumicount luminometer (Fig. 2B) at 30 h posttransfection (hpt). As expected (29, 33), expression of the NS1-1918 and NS1-3mut proteins inhibited GFP and Gluc expression, whereas that of the NS1-PR8 protein did not. Expression of the pH1N1 NS1-wt protein did not efficiently inhibit reporter gene expression, as previously described (33). Interestingly, GFP and Gluc expression was inhibited in cells transfected with the plasmid expressing the NS1-6mut protein from currently circulating pH1N1 viruses, to levels similar to those for the NS1-1918- or NS1-3mut-transfected cells (Fig. 2A and B). These results correlated with the levels of NS1 expression detected by Western blotting (Fig. 2C), showing that NS1 proteins that efficiently blocked GFP and Gluc expression were barely detectable by Western blotting (Fig. 2C). These results indicate that at least 1 of the 6 amino acid changes selected in the recently circulating pH1N1 viruses restored the ability of NS1 to inhibit general gene expression.

FIG 2.

Effects of identified pH1N1 NS1 mutations on inhibition of host gene expression. Human 293T cells were transiently cotransfected with plasmids expressing the indicated HA epitope-tagged NS1 proteins and with GFP- and Gluc-expressing plasmids by use of DNA-In. Empty and 1918 NS1- and PR8 NS1-expressing plasmids were included as controls. At 30 hpt, GFP (A), Gluc (B), and NS1 (C) expression levels were analyzed. (A) GFP was visualized using a fluorescence microscope, and representative images obtained with a 20× objective are shown. Bars, 50 μm. (B) Gluc expression was quantified in a Lumicount luminometer. Error bars represent the standard deviations for triplicates. P values obtained using Student's t test are indicated for comparisons between NS1-wt and NS1-6mut and between NS1-wt and NS1-3mut. (C) pH1N1 NS1 protein and cellular actin expression levels were analyzed by Western blotting using cell extracts and antibodies specific to the NS1 protein and actin (loading control). Western blots were quantified by densitometry using the software ImageJ (v1.46), and the amounts of NS1 proteins were normalized to the amounts of actin. Protein expression in cells transfected with PR8 NS1 was considered 100% for comparison with the levels of expression of the other NS1 variants (numbers below the NS1 blot; ND, not done). Molecular mass markers (in kilodaltons) are indicated on the right. The experiments were repeated 3 times, with similar results.

To analyze whether the 6 amino acid changes improve the ability of the NS1 protein to counteract IFN responses, human 293T cells were transiently cotransfected with plasmids expressing the different NS1 proteins and a plasmid expressing firefly luciferase (Fluc) under the control of an IFN-stimulated response element (ISRE) promoter (Fig. 3A). At 24 hpt, cells were mock infected or infected (multiplicity of infection [MOI] = 3) with Sendai virus (SeV), and at 16 h postinfection (hpi), activation of the ISRE promoter was determined by measuring Fluc activity (Fig. 3A). As expected, SeV infection induced a robust activation of the ISRE promoter in cells transfected with the empty plasmid. In cells transfected with the different NS1 variants, levels of Fluc were induced to different extents, but always to smaller amounts than those in cells transfected with the empty plasmid (Fig. 3A), consistent with results showing that influenza virus NS1 counteracts host innate immune responses (18). The NS1-PR8 and NS1-wt proteins counteracted ISRE promoter activation to a lesser extent than that with the NS1-1918, NS1-6mut, and NS1-3mut proteins, most probably because the NS1-1918, NS1-6mut, and NS1-3mut proteins not only counteract IFN responses but also block general host gene expression (29) (Fig. 3A). To further analyze the effects of the different NS1 proteins on innate immune responses, tissue culture supernatants (TCS) from cells transfected with the different NS1 variants and infected with SeV were collected at 16 hpi, UV inactivated, and used to treat fresh A549 cells. As a control, cells were treated with 250 U/ml of universal IFN-α. The cells were then infected with recombinant vesicular stomatitis virus expressing GFP (rVSV-GFP), a virus highly sensitive to previous antiviral states, and the levels of GFP expression were analyzed using a microplate reader (Fig. 3B) as an indirect measure of the levels of IFN present in the UV-inactivated TCS. As expected, levels of GFP expression were higher in cells treated with TCS collected from mock-infected cells than in cells treated with TCS collected from SeV-infected cells. Likewise, levels of GFP in IFN-treated cells were very low (Fig. 3B). Importantly, the results for TCS obtained from NS1-expressing cells correlated with the results shown in Fig. 3A. Levels of GFP expression were >2-fold higher for cells treated with TCS from SeV-infected, NS1-wt-transfected cells than for cells treated with TCS from SeV-infected, NS1-6mut- or NS1-3mut-transfected cells (Fig. 3B), indicating that the NS1-6mut and NS1-3mut proteins, which inhibit general host gene expression, decrease the amounts of IFN induced after SeV infection to a larger extent than that in cells expressing the NS1-wt protein.

FIG 3.

Effects of pH1N1 NS1 mutations on IFN responses induced by SeV infection. Human 293T cells were transiently cotransfected, using calcium phosphate, with the indicated NS1-expressing plasmids and a plasmid expressing Fluc under the control of an ISRE promoter. Empty and 1918 NS1- and PR8 NS1-expressing plasmids were included as controls. At 24 hpt, cells were mock infected or infected (MOI = 3) with the SeV Cantell strain to induce activation of the ISRE promoter. At 16 hpi, cell lysates were prepared for reporter gene expression. (A) Fluc expression was measured by luminescence assay. Data show the means and standard deviations of the results determined for triplicate wells. Experiments were repeated 3 times in triplicate, with similar results. P values obtained using Student's t test are indicated for comparisons between NS1-wt and NS1-6mut or NS1-3mut. (B) Sixteen hours after SeV infection, TCS were collected and, after UV inactivation, were used to treat fresh A549 cells. Alternatively, A549 cells were treated with 250 U/ml of universal IFN-α as a control. At 24 h posttreatment, cells were infected (MOI = 0.1) with rVSV-GFP. At 16 hpi, rVSV-GFP expression was quantified in a microplate reader. P values obtained using Student's t test are indicated for comparisons of GFP expression levels in cells treated with TCS from SeV-infected cells previously transfected with plasmids encoding the NS1-wt, NS1-6mut, and NS1-3mut proteins.

The NS1 protein from currently circulating pH1N1 viruses binds to CPSF30.

To assess whether the 6 amino acid mutations in currently circulating pH1N1 viruses that allow inhibition of general host gene expression (Fig. 2 and 3) restore the binding of NS1 to CPSF30, extracts of human 293T cells transfected with a plasmid expressing CPSF30 fused to a FLAG tag or transfected with an empty plasmid were incubated with in vitro-transcribed and -translated NS1-wt or NS1-6mut protein and with agarose beads conjugated to an anti-FLAG polyclonal antibody (pAb). The NS1-wt and NS1-6mut proteins were expressed at similar levels (Fig. 4A, anti-NS1 blot). After coimmunoprecipitation, the CPSF30 protein was detected by Western blotting, at similar levels for all cases (Fig. 4B, CPSF30 panel, anti-FLAG blot). As expected, NS1-wt was not coimmunoprecipitated with CPSF30 (Fig. 4B, CPSF30 panel, anti-NS1 blot), consistent with previous results showing that NS1-wt does not interact with CPSF30 (33). Interestingly, the NS1-6mut protein was coimmunoprecipitated with CPSF30 (Fig. 4B, CPSF30 panel, anti-NS1 blot) and was not detected when the CPSF30 protein was not expressed (Fig. 4B, empty panel, anti-NS1 blot). These data indicate that the NS1-6mut protein and CPSF30 interact, providing a mechanism for NS1-6mut-mediated inhibition of host gene expression.

FIG 4.

The NS1 protein from currently circulating pH1N1 viruses binds to CPSF30 and shows a subcellular localization similar to that of the NS1-wt protein. FLAG-tagged CPSF30 plasmid- or empty plasmid-transfected 293T cells were mixed with in vitro-synthesized HA-tagged NS1 variants and immunoprecipitated using an anti-FLAG resin. Following SDS-PAGE, input proteins (A) and immunoprecipitated (IP) proteins from CPSF30 plasmid- or empty plasmid-transfected cells (B) were detected by Western blotting using antibodies specific for the NS1 protein or the FLAG tag (CPSF30 protein). Molecular mass markers are indicated on the right. The experiments were repeated twice, with similar results. (C) 293T cells were transiently transfected with pcDNA plasmids expressing HA-tagged NS1-wt (upper panels) and NS1-6mut (lower panels) proteins. Confocal microscopy images showing anti-HA staining (NS1 protein; green), DAPI staining (blue), and merged images are shown. Bars, 10 μm.

Since the ability of the NS1 protein to interact with CPSF30 could be affected by its subcellular localization, we evaluated the subcellular localization of the NS1-wt and NS1-6mut proteins in transfected 293T cells (Fig. 4C). The specific signal observed for NS1-wt was more intense than the signal observed for NS1-6mut (Fig. 4C), consistent with previous data showing that unlike NS1-wt, NS1-6mut inhibits its own gene expression (Fig. 2). Regardless, both NS1-wt and NS1-6mut localized to the nuclei of transfected cells (Fig. 4C), as previously described (43), suggesting that the 6 mutations in the NS1 protein from currently circulating pH1N1 viruses do not change the subcellular localization of the protein, and therefore that the restored binding to CPSF30 is not due to a change in the subcellular localization of the NS1-6mut protein compared to that of NS1-wt.

The 6 amino acid changes in the NS1 protein of currently circulating pH1N1 viruses contribute to the restored ability of NS1 to inhibit general gene expression.

To discern whether the 6 amino acid changes in the NS1 protein of currently circulating pH1N1 viruses contribute to the restored inhibition of general gene expression, plasmids encoding single amino acid changes separately in the backbone of NS1-wt (Fig. 5A and C) and in the backbone of NS1-6mut (Fig. 5B and D) were generated. Human 293T cells were transiently cotransfected with the different NS1-expressing plasmids and with plasmids expressing the reporter proteins GFP and Gluc (Fig. 5). The levels of Gluc expression were around 10-fold lower in cells transfected with the NS1-6mut plasmid than in cells transfected with the NS1-wt plasmid (Fig. 5C and D). Compared to that in cells expressing NS1-wt, Gluc expression was about 2-fold lower in the cells expressing the NS1-wt proteins containing the L90I, I123V, E125D, K131E, and N205S single mutations and was similar in cells expressing NS1-wt-E55K (Fig. 5C). On the other hand, Gluc expression was around 2- to 5-fold higher in cells expressing the NS1-6mut proteins containing the K55E, I90L, V123I, D125E, E131K, and S205N single mutations than in cells encoding the NS1-6mut protein, with the K55E mutation showing the least effect (Fig. 5D). The levels of GFP expression correlated well with the levels of Gluc expression, as expected (Fig. 5A and B). These data indicate that each of the 6 mutations partially contribute to restoring the ability of the NS1-6mut protein to inhibit general gene expression, with the mutation at residue 55 showing the smallest effect.

FIG 5.

Effects of individual mutations of residues 55, 90, 123, 125, 131, and 205 on inhibition of host gene expression by the pH1N1 NS1 protein. Human 293T cells were transiently cotransfected with GFP- and Gluc-expressing plasmids along with plasmids expressing HA epitope-tagged NS1-wt proteins containing single amino acid changes (A and C) or plasmids expressing HA epitope-tagged NS1-6mut proteins containing single amino acid changes (B and D). At 30 hpt, GFP (A and B) and Gluc (C and D) protein expression levels were analyzed. (A and B) GFP was visualized using a fluorescence microscope, and representative images obtained with a 20× objective are shown. Bars, 100 μm. (C and D) Gluc expression was quantified in a Lumicount luminometer. Error bars represent the standard deviations for triplicates. P values obtained using Student's t test are indicated for comparisons between NS1-wt and NS1-wt mutants with single amino acid changes (C) and between NS1-6mut and NS1-6mut mutants with single amino acid changes (D).

Effects of NS1 mutations on pH1N1 virus growth and induction of IFN-β in cell cultures.

IAV NS1 has previously been shown to play a significant role in viral pathogenesis in vivo (15, 44, 45). Generally, mutations decreasing the ability of NS1 to counteract host innate immune responses have been shown to reduce the virulence of IAV in vivo (46, 47). However, in the case of pH1N1 viruses, the virus encoding the three mutations restoring the ability of NS1 to interact with CPSF30 and to inhibit host gene expression (R108K, E125D, and G189D) was slightly attenuated compared to the pH1N1-wt virus (33). To compare the pathogenicities of pH1N1 viruses encoding the NS1-wt, NS1-6mut, and NS1-3mut proteins, we generated recombinant viruses encoding these NS1 proteins (pH1N1/NSs-wt, pH1N1/NSs-6mut, and pH1N1/NSs-3mut, respectively) (Fig. 6). To this end, a pH1N1 virus encoding a split NS segment (NSs) was used. In this virus, the NS1 and NEP open reading frames (ORFs) are separated by the porcine teschovirus (PTV) 2A autoproteolytic cleavage site (Fig. 6A), as previously described for the A/Puerto Rico/8/34 strain (48). This virus allowed the introduction of mutations in the NS1 protein without affecting the amino acid sequence of NEP, since otherwise the mutation at position 205 would have generated an amino acid change in the NEP protein (Fig. 6A) (48). The three recombinant NSs viruses, as well as a wt virus encoding a nonsplit NS segment and wt NS1 (pH1N1-wt), were rescued using previously described plasmid-based reverse genetics approaches (49). Growth of these viruses was determined in canine MDCK (Fig. 6B) and human A549 (Fig. 6C) cells infected at low MOI (0.001 and 0.1, respectively). Compared to those of the NSs viruses, the pH1N1-wt virus titers were slightly higher (around 5-fold) at 24 hpi in MDCK cells (Fig. 6B) and at 24, 48, and 72 hpi in A549 cells (Fig. 6C). However, the titers of the NS split viruses encoding the NS1-wt, NS1-3mut, and NS1-6mut proteins were similar in both cell lines, suggesting that the 3 (R108K, E125D, and G189D) and 6 (K55E, I90L, V123I, D125E, E131K, and S205N) mutations which restore NS1-mediated general gene expression inhibition did not significantly affect virus growth in cell cultures. To analyze whether the pH1N1/NSs-wt, pH1N1/NSs-6mut, and pH1N1/NSs-3mut viruses show differences in the ability to mount an effective antiviral IFN response, MDCK cells constitutively expressing GFP and Fluc reporter genes under the control of the IFN-β promoter (MDCK IFNβ GFP-CAT/IFNβ Fluc) (50) were mock infected or infected (MOI = 5) with the pH1N1/NSs-wt, pH1N1/NSs-6mut, or pH1N1/NSs-3mut virus (Fig. 7). Importantly, comparable levels of infection were verified by immunofluorescence assay using an anti-nucleoprotein (anti-NP) antibody (Fig. 7A). At 12 hpi, activation of the IFN-β promoter was measured by evaluating GFP expression (data not shown) and quantifying the levels of Fluc expression (Fig. 7B). Correlating with the levels of GFP expression, the levels of Fluc expressed after infection were 2- to 2.5-fold higher in cells infected with the pH1N1/NSs-wt virus than in cells infected with the pH1N1/NSs-6mut or pH1N1/NSs-3mut virus (Fig. 7B), similar to the results obtained for plasmid-transfected cells infected with SeV (Fig. 3). Newcastle disease virus (NDV) infection is highly reduced by the antiviral state induced in cells (51). Taking this into account, as a second approach to analyze the presence of IFN in TCS from pH1N1-infected cells, we evaluated the inhibition of rNDV-GFP infection in MDCK cells treated with UV-inactivated TCS. Inhibition of rNDV-GFP was evaluated by GFP expression and was dependent on the amounts of IFN present in the TCS of the cells infected with the different pH1N1 viruses. As expected, rNDV-GFP replicated efficiently in cells pretreated with TCS from mock-infected cells, as evidenced by high levels of GFP expression (Fig. 7C). However, rNDV-GFP replication was reduced in cells treated with TCS from pH1N1/NSs-wt, pH1N1/NSs-6mut, and pH1N1/NSs-3mut virus-infected cells, consistent with the data showing that pH1N1 infection induces IFN production (Fig. 7B). Interestingly, rNDV-GFP replication was significantly reduced in cells treated with TCS from pH1N1/NSs-wt-infected cells compared to that in cells treated with TCS from pH1N1/NSs-6mut- or pH1N1/NSs-3mut-infected cells (Fig. 7C), confirming that infection with pH1N1/NSs-6mut or pH1N1/NSs-3mut inhibited the cellular antiviral state to a larger extent, as previously described for virus-infected cells (Fig. 7B) and plasmid-transfected, SeV-infected cells (Fig. 3).

FIG 6.

Recombinant pH1N1 mutant virus growth kinetics in vitro. (A) Schematic of the wt virus, carrying overlapping NS1 and NEP genes, which are translated by alternative splicing (67) (left), and the virus containing the NS split (NSs) segment, carrying the NS1 and NEP genes separated by the porcine teschovirus 1 (PTV-1) 2A autoproteolytic cleavage site (48) (right). Numbers in the left panel indicate the amino acid residues changed in currently circulating pH1N1 viruses. Viral 3′ and 5′ noncoding regions are indicated with black boxes at the ends of the viral segments. Viral products from the NS segment (NS1 and NEP) are indicated with white boxes. The region up to the splicing donor in the viral segments is indicated with light gray boxes. The 2A autoproteolytic cleavage site is indicated with a dark gray box. Canine MDCK (B) and human A549 (C) cells were infected (MOI of 0.001 and 0.1, respectively) in triplicate with the recombinant wt and NS split viruses encoding NS1-wt, NS1-6mut, and NS1-3mut (pH1N1-wt, pH1N1/NSs-wt, pH1N1/NSs-6mut, and pH1N1/NSs-3mut). Virus titers in TCS from infected cells were determined at the indicated times by immunofocus assay. P values obtained using Student's t test for comparisons of the pH1N1-wt virus and the pH1N1-NSs viruses are indicated. The dotted line indicates the limit of detection (20 FFU/ml).

FIG 7.

Induction of IFN by pH1N1/NSs-wt, pH1N1/NSs-6mut, and pH1N1/NSs-3mut viruses in canine MDCK cells. (A to C) MDCK cells constitutively expressing GFP and Fluc reporter genes under the control of the IFN-β promoter (MDCK IFN-β GFP-CAT/IFNβ Fluc) were infected (MOI = 5) with the pH1N1/NSs-wt, pH1N1/NSs-6mut, or pH1N1/NSs-3mut virus. (A and B) At 12 hpi, the levels of viral infection were determined by immunofluorescence assay using an antibody specific for viral NP (A), and the activation of the IFN-β promoter was determined by assessing Fluc activity by use of a microplate reader (B). (C) TCS of previously infected MDCK cells were collected and, after UV inactivation, were used to treat fresh MDCK cells. After 24 h of incubation, cells were infected (MOI = 5) with the IFN-sensitive virus rNDV-GFP. At 24 hpi, rNDV-GFP expression was quantified in a microplate reader. Experiments were repeated 3 times in triplicate wells, with similar results. Data represent the means and standard deviations for one representative experiment. P values obtained using Student's t test are indicated.

pH1N1/NSs-6mut and pH1N1/NSs-3mut pH1N1 viruses further reduce induction of IFN and proinflammatory cytokines.

To further analyze the ability of the pH1N1 NS1 variants to inhibit IFN responses in a more relevant system, human A549 cells were infected (MOI = 5) with the pH1N1/NSs-wt, pH1N1/NSs-6mut, or pH1N1/NSs-3mut virus for 12 h (Fig. 8A to C). After confirming similar levels of infection with an anti-NP antibody (data not shown), expression levels of IFN-β and IFN-induced protein with tetratricopeptide repeats 2 (IFIT2) were measured at the mRNA level by quantitative reverse transcription-PCR (qRT-PCR) (Fig. 8A and B, respectively). Consistent with the results shown in MDCK cells for IFN-β (Fig. 7B), the mRNA levels of IFN-β and IFIT2 were higher (around 2.5-fold) in pH1N1/NSs-wt-infected A549 cells than in pH1N1/NSs-6mut- or pH1N1/NSs-3mut-infected A549 cells (Fig. 8A and B, respectively). To indirectly measure the levels of IFN secreted into the medium after infection, TCS collected from mock-infected and infected A549 cells were UV treated and used to treat fresh A549 cells, which were then infected with the IFN-sensitive virus rVSV-GFP (Fig. 8C). Compared to that in the cells treated with TCS from mock-infected cells, expression of GFP was slightly decreased in cells pretreated with TCS from pH1N1/NSs-6mut- or pH1N1/NSs-3mut-infected cells (Fig. 8C). However, GFP expression was significantly reduced in the cells treated with the supernatants from pH1N1/NSs-wt-infected cells (Fig. 8C), indicating that the amino acid changes in the NS1-6mut and NS1-3mut proteins further decrease the ability of the proteins to inhibit IFN responses compared to that of the NS1-wt protein, most probably because the NS1 proteins from these viruses inhibit IFN responses as well as general gene expression.

FIG 8.

Induction of IFN and proinflammatory cytokine responses by pH1N1/NSs-wt, pH1N1/NSs-6mut, and pH1N1/NSs-3mut viruses. Human A549 cells were infected (MOI = 5) with the pH1N1/NSs-wt, pH1N1/NSs-6mut, or pH1N1/NSs-3mut virus. (A, B, D, and E) At 12 hpi, the levels of mRNAs for IFN-β (A), IFIT2 (B), CXCL10 (D), and TNF (E) were determined by qRT-PCR, using specific TaqMan assays. Data represent the means and standard deviations for one representative experiment of three performed in triplicate. P values obtained using Student's t test are indicated. (C) TCS of previously infected A549 cells were collected and, after UV inactivation, were used to treat fresh A549 cells. After 24 h of incubation, cells were infected (MOI = 0.1) with the IFN-sensitive virus rVSV-GFP. At 16 hpi, GFP expression in rVSV-GFP-infected cells was quantified in a microplate reader. Experiments were repeated 3 times in triplicate wells, with similar results.

To analyze whether the pH1N1/NSs-6mut and pH1N1/NSs-3mut viruses induce lower proinflammatory cytokine levels as well, human lung A549 cells were infected (MOI = 5) with the pH1N1/NSs-6mut, pH1N1/NSs-3mut, and pH1N1/NSs-wt viruses, and the levels of expression of C-X-C motif chemokine 10 (CXCL10) and tumor necrosis factor (TNF) were evaluated by qRT-PCR at 12 hpi (Fig. 8D and E). Interestingly, the levels of CXCL10 and TNF induced by viral infection were significantly higher in pH1N1/NSs-wt-infected cells than in pH1N1/NSs-3mut-infected and, particularly, pH1N1/NSs-6mut-infected cells, indicating that NS1-3mut and, particularly, NS1-6mut further decrease the induction of proinflammatory cytokines compared to that with the NS1-wt protein.

Effects of NS1 mutations on virus virulence and replication in mice.

We next evaluated the effects of NS1 mutations on virus pathogenesis in vivo (Fig. 9). To that end, mice (n = 5) were infected intranasally (i.n.) with 100 (Fig. 9A) or 1,000 (Fig. 9B) focus-forming units (FFU) of the pH1N1/NSs-wt, pH1N1/NSs-6mut, pH1N1/NSs-3mut, or pH1N1-wt virus. Morbidity (body weight loss) and mortality (% survival) were monitored for 14 days. Maximum weight losses were detected between days 9 and 10 for all the infected mice. Interestingly, at the 100-FFU dose (Fig. 9A), whereas the mice infected with the pH1N1-wt virus lost a maximum of 22% of body weight, mice infected with the pH1N1/NSs-wt virus lost a maximum of 13% of body weight, and mice infected with the pH1N1/NSs-3mut and pH1N1/NSs-6mut viruses lost a maximum of 5% and 2.5% of body weight, respectively (Fig. 9A). With the higher dose of 1,000 FFU (Fig. 9B), weight losses were higher, as expected. Similar to the results observed with the lower dose, mice infected with the pH1N1-wt virus lost the most weight (25%), followed by the mice infected with the pH1N1/NSs-wt virus (20%), the mice infected with the pH1N1/NSs-3mut virus (18%), and pH1N1/NSs-6mut-infected mice (7%) (Fig. 9B). With the lower dose, only two mice infected with the pH1N1-wt virus died, whereas all mice infected with the pH1N1/NSs-wt, pH1N1/NSs-3mut, and pH1N1/NSs-6mut viruses survived infection (Fig. 9A). However, at the higher dose, 20%, 20%, and 100% of the mice infected with the pH1N1/NSs-3mut, pH1N1/NSs-wt, and pH1N1-wt viruses, respectively, succumbed to the infection, whereas 0% of the mice infected with the pH1N1/NSs-6mut virus died (Fig. 9B). These results suggested that compared to the pH1N1-wt virus, the pH1N1/NSs-wt virus was slightly attenuated, causing less morbidity and mortality. In addition, the virus encoding the NS1-3mut protein was partially attenuated, and the virus encoding the NS1-6mut protein was further attenuated. To analyze whether virus attenuation correlates with virus replication, viral titers in the lungs of mice infected with the pH1N1/NSs-wt, pH1N1/NSs-3mut and pH1N1/NSs-6mut viruses were evaluated at 2, 5, and 9 days postinfection (dpi) (Fig. 9C). Virus titers in mice infected with the three viruses were similar at 2 and 5 dpi. Only slightly lower (5-fold) titers were observed at 2 dpi for mice infected with the pH1N1/NSs-3mut virus compared to those for mice infected with the pH1N1/NSs-6mut and pH1N1/NSs-wt viruses, which were similar (Fig. 9C). At 9 dpi, we could detect viral titers for only 1 of 3 mice infected with the pH1N1/NSs-6mut or pH1N1/NSs-wt virus (Fig. 9C). These data suggest that the pH1N1/NSs-6mut and pH1N1/NSs-wt viruses replicate similarly, and therefore the attenuation observed for the pH1N1/NSs-6mut virus is likely not due to differences in virus replication.

FIG 9.

Virulence and growth of recombinant pH1N1 viruses containing mutations restoring the ability of NS1 to inhibit general gene expression. Groups of 7- to 8-week-old C57BL/6 female mice (n = 5) were infected with 100 (A) and 1,000 (B) FFU per mouse of the pH1N1-wt, pH1N1/NSs-wt, pH1N1/NSs-6mut, or pH1N1/NSs-3mut virus. Weight loss (left panels) and survival (right panels) were evaluated daily for 2 weeks. (C) Groups of 7- to 8-week-old C57BL/6 female mice (n = 3) were infected (1,000 FFU/mouse) with the pH1N1/NSs-wt, pH1N1/NSs-6mut, or pH1N1/NSs-3mut virus. Mice were sacrificed at 2, 5, and 9 dpi, and lungs were harvested, homogenized, and used to quantify viral titers by immunofocus assay (giving FFU per milliliter). #, infectious virus was detected in only 1 of 3 mice. The dotted line indicates the limit of detection (20 FFU/ml).

Evaluation of innate immune responses in mice.

To determine whether restoring NS1's ability to inhibit host gene expression results in reduced expression levels of IFN, ISGs, and proinflammatory cytokines in vivo, the levels of mRNAs encoding IFN-β, IFIT2, chemokine (C-C) motif ligand 2 (CCL2), and TNF were evaluated in mouse lungs at 2, 5, and 9 dpi by qRT-PCR (Fig. 10). The induction of these host genes was maximal at 5 dpi (Fig. 10A and B). Interestingly, the pH1N1/NSs-wt virus induced higher expression levels of IFN-β and IFIT2 at 2 and 5 dpi than those with the pH1N1/NSs-6mut and pH1N1/NSs-3mut viruses, consistent with our previous results for cell cultures (Fig. 7) and with the idea that NS1 proteins with a restored ability to block general gene expression induce smaller innate immune responses in vitro and in vivo. The levels of IFIT2, CCL2, and TNF mRNAs at 9 dpi were higher for pH1N1/NSs-3mut-infected mice than for pH1N1/NSs-6mut-infected mice (Fig. 10A and B), suggesting that the NS1-6mut protein found in currently circulating viruses may be slightly better at inhibiting innate immune responses than the NS1-3mut protein, which was artificially designed in a laboratory. Interestingly, it has been shown that highly pathogenic influenza viruses (i.e., H5N1) induce high levels of proinflammatory genes, contributing to the virus-induced pathogenesis (52, 53). Moreover, the NS1 protein inhibits the expression of proinflammatory cytokines, partly due to its CPSF30-binding function (54). Remarkably, in mouse lungs evaluated at 2 and 5 dpi, the levels of proinflammatory cytokine mRNAs, such as those for TNF and CCL2, induced after infections with the pH1N1/NSs-6mut and pH1N1/NSs-3mut viruses were significantly lower than those induced by pH1N1/NSs-wt viral infection (Fig. 10B). However, whereas at day 9 the levels of CCL2 and TNF mRNAs were lower in pH1N1/NSs-6mut-infected mice, the levels in pH1N1/NSs-3mut- and pH1N1/NSs-wt-infected mice were similar. These effects were most probably due to the restoration of NS1's CPSF30-binding function and provide a likely explanation for the lower morbidity observed for the pH1N1/NSs-3mut virus and, particularly, the pH1N1/NSs-6mut virus (Fig. 9A and B). With human cells, the levels of proinflammatory cytokines induced by viral infection were significantly higher in pH1N1/NSs-wt-infected cells than in pH1N1/NSs-3mut- and, particularly, pH1N1/NSs-6mut-infected cells (Fig. 8D and E), indicating that the effect of the NS1 mutations on the induction of proinflammatory cytokine expression may also apply to human systems.

FIG 10.

In vivo IFN-β, ISG, and proinflammatory responses induced by pH1N1 mutant viruses containing NS1 mutations that restore the ability to inhibit general gene expression. Groups of 7- to 8-week-old C57BL/6 female mice (n = 3) were infected with 1,000 FFU of the pH1N1/NSs-wt, pH1N1/NSs-6mut, or pH1N1/NSs-3mut virus. Mice were sacrificed at 2, 5, and 9 dpi, and total RNA was extracted from lungs to quantify the levels of IFN-β and IFIT mRNAs (A) and CCL2 and TNF mRNAs (B) by qRT-PCR. P values obtained using Student's t test are indicated. ns, not significant.

DISCUSSION

Interaction of the IAV NS1 protein with CPSF30 blocks the processing of cellular mRNAs, leading to inhibition of general gene expression, including that of antiviral activity genes and proinflammatory cytokine genes (28–31). However, this CPSF30-NS1 binding property is not conserved in all IAV NS1 proteins. For example, the NS1 proteins of the 1918 (H1N1) strain and currently circulating H3N2 IAVs bind CPSF30 and inhibit general host gene expression (29, 36, 37), whereas the NS1 proteins of PR8, pH1N1, and H7N9 IAVs do not (29, 33, 34). Interestingly, starting with pH1N1 viruses circulating in the 2015-2016 season, six mutations (E55K, L90I, I123V, E125D, K131E, and N205S) that restore pH1N1 NS1's ability to inhibit general gene expression were identified (Table 1). These mutations were selected in viruses circulating worldwide, with almost 100% of the currently circulating pH1N1 viruses encoding these 6 mutations (Fig. 1). Similarly, it was shown that the NS1 protein of H5N1 viruses presented a defect in inhibiting general gene expression when the transmission from birds to humans occurred in 1997; however, the viruses isolated since 1998 have gained this NS1 function (35), suggesting a selective advantage for viruses encoding an NS1 protein able to inhibit general gene expression in the human host.

The NS1 binding site for CPSF30 is centered around amino acid 186 (30). In addition, amino acid residues 103 and 106 (29), 108, 125, and 189 (33), 106 (34), 103 and 106 (35), and 64, 189, and 194 (36, 37) are important for NS1 binding to CPSF30 for the PR8, pH1N1, H7N9, H5N1, and seasonal H3N2 IAVs, respectively. In this work, we found that the naturally occurring mutations E55K, L90I, I123V, E125D, K131E, and N205S each partially contribute to the restored ability of pH1N1 NS1 to inhibit general gene expression, although the mutation at residue 55 has an effect only in the backbone of the NS1-6mut protein, not in the NS1-wt backbone (Fig. 5). It was previously shown that the E125D mutation partially restored the ability of the NS1 protein of the pH1N1 2009 virus to inhibit host gene expression, reinforcing the data presented in this work (33). Similar to the results of our work, a pH1N1 NS1 protein containing three artificially introduced mutations (R108K, E125D, and G189D; the E125D mutation is also described here, but the other two amino acid changes are different) also binds CPSF30 and inhibits general gene expression (33), showing that there are multiple ways that NS1 can regain the ability to bind CPSF30 and inhibit general gene expression.

While it was thought that only the NS1 protein effector domain contributes to the inhibition of general host gene expression through binding to CPSF30, we recently described that an amino acid in the NS1 dsRNA-binding domain (residue 64) affects CPSF30 binding and is important for NS1's inhibition of host gene expression (36), providing a previous example of a residue in the dsRNA-binding domain (like residue 55 from this study) modulating this function. The three-dimensional structure of NS1 bound to the F2/F3 domain of CPSF30 (28) shows that NS1 residues 123 and 125 are close to the CPSF30 protein (Fig. 11A), providing a likely explanation for the effects of their amino acid changes in the binding to CPSF30. Residues 55 and 205 are not in the region of the crystallized NS1, which includes residues 85 to 203 (28), and residues 90 and 131 are not close to the CPSF30-binding sites (Fig. 11A). Similarly, using the NS1-3mut protein, containing the R108K, E125D, and G189D mutations restoring NS1-CPSF30 binding, two residues (residues 125 and 189) are close to CPSF30, whereas residue 108 is not as close to the CPSF30-binding sites (Fig. 11B). However, we cannot rule out that these residues are important for maintaining the structure of other residues important for CPSF30 binding. Moreover, given that our model is the crystal structure of the A/Udorn/72 H3N2 NS1 protein (28), since the crystal structure for the pH1N1 NS1 protein is not currently available, we cannot discard the possibility of differences between both viral proteins and different effects on CPSF30 binding.

FIG 11.

Three-dimensional structure of the IAV NS1 effector domain bound to the F2/F3 domain of CPSF30. The NS1 effector domain of IAV NS1 (A/Udorn/72 H3N2) coupled to the F2/F3 fragment of human CPSF30 was previously crystalized (28) (PDB entry 2RHK). The structure was colored using the MacPyMOL molecular graphics system. The monomers of the NS1 effector domain are presented in light and dark green. Monomers of the F2/F3 fragment are presented in dark and light blue. Residues 90, 123, 125, and 131, found in currently circulating pH1N1 viruses (A), and residues 108, 125, and 189, artificially introduced into the pH1N1 virus NS1 protein to restore NS1-CPSF30 binding (33) (B), are indicated in red.

Depending on the influenza virus strain, NS1-mediated inhibition of host gene expression may lead to increased virus virulence or attenuation. Increasing the ability of the NS1 protein to inhibit host gene expression in H7N9 and H5N1 viruses by introducing the L103F and I106M mutations led to decreased IFN responses after infection, increased virus titers in vivo, and higher virulence levels in mice (34, 47). Similarly, blocking NS1-mediated inhibition of host gene expression in an H3N2 seasonal virus led to virus attenuation (36). In contrast, the pH1N1 viruses containing 6 mutations (E55K, L90I, I123V, E125D, K131E, and N205S) and the virus encoding 3 mutations (R108K, E125D, and G189D), all of which have a restored ability of NS1 to block host gene expression, were attenuated in mice compared to the virulence of the original virus (Fig. 9) (33). Unexpectedly, although the induction of IFN and antiviral genes was reduced after infections with the pH1N1/NSs-6mut and pH1N1/NSs-3mut viruses in cell cultures and in vivo, we did not observe an increase in virus replication, similar to the previous results with the pH1N1/NSs-3mut virus (33). However, we cannot rule out that the cell cultures and the mouse strain used for these experiments are not sensitive enough to detect small differences, and the situation in humans may be different. In addition, the absence of increased virus titers may be due to a negative effect of these mutations on other NS1 functions (18). Although mice are a valid animal model for studying influenza virus pathogenicity, we cannot discard the possibility that mutations in NS1 affect virus transmission in humans. Therefore, future experiments to analyze virus transmission in more relevant animal transmission models, such as guinea pigs or ferrets (55, 56), will be required to evaluate the roles of NS1 binding to CPSF30 and inhibition of host general gene expression in viral transmission.

The pH1N1/NSs-6mut virus grew to titers similar to those of the pH1N1/NSs-wt virus in mouse lungs (Fig. 9). Interestingly, the induction of proinflammatory cytokines was higher in mice infected with the pH1N1/NSs-wt virus than in mice infected with the pH1N1/NSs-3mut virus and, particularly, the pH1N1/NSs-6mut virus (Fig. 10 and 8D and E), most probably because the NS1-6mut and NS1-3mut proteins are capable of inhibiting general gene expression (including that of proinflammatory cytokine genes), whereas the wt NS1 protein is not. This correlates with the virulence observed for these viruses as determined by weight loss and survival (Fig. 9). It was shown previously that highly pathogenic influenza viruses (i.e., H5N1 viruses) induce high levels of proinflammatory genes, contributing to the virus-induced pathogenesis (52, 53, 57). Furthermore, mice and cynomolgus macaques infected with pH1N1 virus showed higher levels of several proinflammatory cytokines and chemokines than animals infected with seasonal H1N1 viruses, resulting in higher pathogenicity (39). In a pig model, swine-origin pH1N1 viruses (one derived from a human patient and the other derived from swine) were more virulent than a swine-origin 1918-like classical IAV strain (58). In addition, these viruses upregulated the expression of proinflammatory genes to a higher extent than that with the swine-origin 1918-like classical IAV, indicating that both pH1N1 isolates were more virulent due in part to differences in the host transcriptional response during acute infection (58). Similarly, for other respiratory viruses, such as severe acute respiratory syndrome coronavirus (SARS-CoV), pathogenicity is significantly due to the cytokine storm induced during infection (59). Therefore, the lower levels of induction of proinflammatory cytokine expression after infection with the pH1N1/NSs-6mut and pH1N1/NSs-3mut viruses are a likely explanation for the attenuation observed for these viruses. Notably, since the pH1N1 virus has naturally acquired the 6 reported mutations, it is feasible to speculate that they confer an advantage to the virus and represent an adaptation to the new host.

Whereas the effect of NS1-mediated inhibition of gene expression on IFN responses has been studied for many IAV strains, the effect of this NS1 function on the induction of proinflammatory cytokine/chemokine expression has remained elusive. Thus, this work introduces novel effects of NS1 inhibition of gene expression on the induction of inflammation and viral pathogenesis in vivo, which may be related to virus adaptation to replicate in the human host.

MATERIALS AND METHODS

Human subjects.

Human subjects were enrolled as part of an “acute influenza” surveillance protocol (IRB number 09-0034). Individuals reporting influenza-like illness (fever, cough, and rhinitis) were asked to visit the Vaccine Research Unit (VRU) at the University of Rochester for nasal and nasopharyngeal swabbing and blood sampling. The study was approved by the University of Rochester Human Research Subjects Review Board. Informed written individual or parental consent was obtained for each participant.

NS1 sequencing.

RNA was obtained from 300 μl of patient nasal wash/swab by use of a QIAamp viral RNA extraction kit (Qiagen) according to the manufacturer's instructions. Reverse transcription (RT) reactions were performed for 2 h at 37°C by use of a High Capacity cDNA reverse transcription kit (Applied Biosystems) and the primers NS1-NCR-5′-VS (5′-AGCAAAAGCAGGGTGACAAAGACATAATGG-3′), complementary to the 5′ noncoding region (NCR) and the first 4 nucleotides (nt) of the pandemic A/California/04/2009 H1N1 (pH1N1) influenza virus NS1 gene (GenBank accession no. FJ969514.1), and NEP-3′-RS (5′-GAGATAAGAGCTTTCTCGTTTCAGCTTATTTAATG-3′), complementary to the pH1N1 NEP ORF. The cDNAs were amplified by PCR using Platinum Pfx polymerase (Life Technologies) and the same VS and RS primers. Amplified PCR products were used for Sanger sequencing (Genewiz).

Cells.

Human embryonic kidney 293T cells (ATCC CRL-11268), human lung epithelial carcinoma A549 cells (ATCC CCL-185), and canine kidney epithelial MDCK cells (ATCC CCL-34) were grown at 37°C in air enriched with 5% CO₂, using Dulbecco's modified Eagle's medium (DMEM; Gibco) supplemented with 10% fetal bovine serum (Gibco), 100 U/ml penicillin, and 50 μg/ml gentamicin (Gibco).

Plasmids.

pcDNA3 polymerase II expression plasmids expressing NS1 sequences fused to a hemagglutinin (HA) epitope tag (YPYDVPDYA) at the N terminus (23) were obtained by overlapping PCR using the first 4 primers described in Table 2. To abolish NS mRNA splicing, two silent mutations were introduced, at nt 501 and 504 of the NS1 gene (shown in bold in Table 2). RNAs extracted from nasal washes/swabs from patient ACU001 and from MDCK cells infected with pH1N1 were used for RT reactions, performed at 37°C for 2 h by use of a High Capacity cDNA reverse transcription kit (Applied Biosystems) and an oligo(dT) primer, to obtain the cDNAs. Overlapping PCRs using Platinum Pfx polymerase (Life Technologies) and the primers and template cDNAs described in Table 2 were used to amplify the viral NS1 protein. PCR products were cloned into the pcDNA3 expression plasmid by digestion with EcoRI plus NheI (PCR products) and EcoRI plus XbaI (pcDNA3 plasmid). Plasmids encoding the NS1 proteins of the influenza A/PuertoRico/8/34 (PR8) and influenza A/BrevigMission/01/1918 (1918) H1N1 viruses were described previously (23, 29). Plasmid pcDNA-3mut, encoding an HA-tagged NS1 protein containing the R108K, E125D, and G189D mutations (33) (pcDNA-HA NH₂-NS1-3mut), was generated by site-directed mutagenesis, using a plasmid encoding the wt NS1 protein as the template. To generate the pDZ-NSs-6mut and -3mut plasmids, we cloned the NS1-6mut and NS1-3mut ORFs into a pDZ plasmid carrying nonoverlapping NS1 and NEP genes (pDZ-NSs) separated by the porcine teschovirus 2A autoproteolytic site (48). NS1 ORFs were amplified by using the pcDNA3-HA NH₂-NS1-6mut and -NS1-3mut plasmids as templates and the primers NS1-EcoRI-NS1cal-VS and NS1cal-AgeI-RS (Table 2). PCR products were digested with the EcoRI and AgeI restriction enzymes and cloned into the pDZ-NSs plasmid.

TABLE 2.

Primers and templates used to clone NS1 proteins into pcDNA3 (first 4 primers) and into pDZ (last 2 primers)

Template	PCR no.	Primer	Sequence (5′–3′)
RNA from subject 001 or RNA from A/California/04/2009	1	NS1cal-NS1-EcoRI-VS	CGAGGGAATTCATGTACCCTTATGATGTGCCAGATTATGACTCCAACACCATGTCAAGCTTTCa
		NS1cal-mutsplic-RS	CCATTACCTTCTCTTCCCGGGCATACTTATGAGGATGTCb
	2	NS1cal-mutsplic-VS	GACATCCTCATAAGTATGCCCGGGAAGAGAAGGTAATGGb
		NS1cal-NheI -RS	GGACCTACCGCTAGCTCATTTCTGCTCTGGAGGTAGTGAAGGc
pcDNA-NS1-6mut or pcDNA3-NS1-3mut	1	NS1-EcoRI-NS1cal-VS	GGCTGAATTCGAGCTCGGTACCCGGGGATCCTCTAGACCGGAGTACTGGTCGACCTCCGAAGTTGGGGGGGAGCAAAAGCAGGGTGACAAAAACATAatggactccaacaccatgtcaagcd
		NS1cal-AgeI-RS	gagaccttcactacctccagagcagaaaGGCACCGGTGCTAGCGGe

^aLetters in bold italics represent the EcoRI restriction site used for cloning. Underlined letters encode the HA tag.

^bLetters in bold represent the silent mutations introduced to abolish NS mRNA splicing.

^cLetters in bold italics represent the NheI restriction site used for cloning.

^dUnderlined letters represent the EcoRI restriction site, italic letters represent the 5′ NCR, and lowercase letters represent the first 24 nt of the NS1 ORF.

^eLowercase letters represent the last 28 nt of the NS1 ORF. Underlined letters represent the AgeI restriction site.

Virus rescues.

Cocultures (1:1) of 293T and MDCK cells in 6-well plates were transiently cotransfected, in suspension, with 1 μg each of the seven ambisense wt plasmids (pDZ-PB2, -PB1, -PA, -HA, -NP, -NA, and -M) of pH1N1 (kindly provided by A. Garcia-Sastre, Icahn School of Medicine at Mount Sinai, NY) plus the pDZ-NSs plasmid encoding the pH1N1/NS1-wt, NS1-6mut, or NS1-3mut protein by using DNA-In (MTI-GlobalStem) (49). Recombinant viruses encoded the HA protein from an egg-adapted pH1N1 strain (A/California/04/2009/E3) which is highly pathogenic in mice (61). At 12 hpt, the medium was replaced with DMEM containing 0.3% bovine serum albumin (BSA), antibiotics, and 0.5 μg/ml tosylsulfonyl phenylalanyl chloromethyl ketone (TPCK)-treated trypsin (Sigma). At 48 hpt, tissue culture supernatants (TCS) were collected and used to infect fresh MDCK cells. At 3 dpi, recombinant viruses were plaque purified and scaled up in MDCK cells. Virus stocks were generated by infecting confluent 10-cm dishes of MDCK cells at a low MOI (0.001). Stocks were titrated by an immunofocus assay (giving FFU per milliliter) on MDCK cells as described above. The identities of the NS1 ORFs in the rescued viruses were confirmed by restriction analysis and sequencing (Genewiz).

Virus growth kinetics.

To determine virus growth rates in vitro, confluent monolayers of MDCK or A549 cells (24-well-plate format, 2 × 10⁵ cells/well, triplicates) were infected at low MOI (0.001 and 0.1, respectively). After 1 h of virus adsorption at room temperature, cells were washed and overlaid with DMEM containing 0.3% BSA, antibiotics, and TPCK-treated trypsin (1 μg/ml for MDCK cells and 0.25 μg/ml for A549 cells). At the indicated hpi, TCS were collected and titrated by immunofocus assay (giving FFU per milliliter) on MDCK cells as described above.

Virus titrations.

Confluent plates of MDCK cells (96-well-plate format, 5 × 10⁴ cells/well) were infected with 10-fold serial dilutions of TCS and incubated at 37°C. At 8 hpi, cells were fixed and permeabilized using 4% formaldehyde and 0.5% Triton X-100 in phosphate-buffered saline (PBS) for 20 min at room temperature. After washing with PBS, cells were incubated in blocking solution (2.5% BSA in PBS) for 1 h at room temperature and then incubated with an IAV nucleoprotein (NP) monoclonal antibody (MAb) purified from hybridoma cells (HB-65; ATCC), diluted in 1% BSA, for 2 h at 37°C. After washing with PBS, cells were incubated with a fluorescein isothiocyanate (FITC)-conjugated rabbit anti-mouse IgG secondary antibody (62-6511; Invitrogen), diluted 1:1,000 in 1% BSA, for 1 h at 37°C. IAV NP-positive cells were visualized and enumerated to determine the virus titer (in FFU per milliliter) by use of a fluorescence microscope. Viral infections were performed in the presence of 1 μg/ml TPCK-treated trypsin (Sigma).

Inhibition of host protein expression.

To evaluate the effects of IAV NS1 proteins on inhibition of host protein synthesis, confluent monolayers of human 293T cells (96-well-plate format, 5 × 10⁴ cells/well, triplicates) were transiently cotransfected, by use of DNA-In, with 200 ng/well of pcDNA3-HA NH₂-NS1 expression plasmids, or an empty plasmid as a control, together with 50 ng/well of pCAGGS plasmids expressing GFP (29) and Gaussia luciferase (Gluc) (62). GFP expression levels were measured at 30 hpt by fluorescence microscopy. Gluc expression levels were measured by mixing TCS with equal volumes of Biolux Gaussia luciferase reagent (New England BioLabs) and using a Lumicount luminometer.

Western blots.

Cells were lysed in passive lysis buffer (Promega) and clarified. Laemmli sample buffer (Bio-Rad) and β-mercaptoethanol (2.5%) were added, and samples were heated at 90°C for 5 min before running in 10% SDS-PAGE gels. Proteins were transferred to nitrocellulose membranes (Bio-Rad) and detected by Western blotting. Primary polyclonal antibodies (pAbs) against NS1 (rabbit pAb against the first 73 amino acids of NS1 A/swine/Texas/4199-2/98; 1:1,000 dilution) (63) and the FLAG tag (to detect CPSF30; diluted 1:1,000) (F7425; Sigma-Aldrich) and an anti-actin MAb (diluted 1:2,000) (A1978; Sigma-Aldrich) were used, followed by incubation with a 1:1,000 dilution of goat anti-rabbit or goat anti-mouse IgG antibodies conjugated to horseradish peroxidase (RNP4301 and NA931V, respectively; GE Healthcare). Membranes were revealed by chemiluminescence, using SuperSignal West Femto maximum sensitivity substrate (Thermo Scientific) according to the manufacturer's recommendations.

Reporter assays and bioassays to assess induction of innate immune responses.

To evaluate the effect of NS1 protein expression on the induction of IFN mediated by Sendai virus (SeV) infection, triplicate wells (96-well-plate format, 5 × 10⁴ cells/well) of human 293T cells were transiently cotransfected (using calcium phosphate) with 100 ng/well of the pcDNA3-HA NH₂-NS1 plasmids, or the empty plasmid as a control, and 50 ng/well of pCAGGS plasmid expressing firefly luciferase (Fluc) under the control of the ISRE (pISRE-Fluc) promoter (29). At 24 hpt, cells were infected with the SeV Cantell strain (29), and at 16 hpi, Fluc expression in cell lysates was quantified by use of a luciferase reporter assay (Promega) and a Lumicount luminometer. In addition, 16 h after SeV infection, TCS were collected, and SeV was inactivated by exposure to shortwave (254 nm) UV radiation for 40 min at a distance of 6 cm (64). Fresh human A549 cells (96-well-plate format, 5 × 10⁴ cells/well, triplicates) were treated with the UV-inactivated TCS for 24 h and then infected at an MOI of 0.1 with a recombinant vesicular stomatitis virus expressing GFP (rVSV-GFP) for 16 h (64). GFP expression was quantified in a microplate reader (Beckman Coulter). GFP expression of mock-treated, rVSV-GFP-infected cells was considered to represent 100% infection. Mean values and standard deviations (SD) were calculated using Microsoft Excel software.

Coimmunoprecipitation of NS1 and CPSF30 proteins.

Human 293T cells (10-cm plates, 1 × 10⁷ cells/plate) were transiently transfected with 30 μg of a pCAGGS plasmid expressing a FLAG-tagged version of human CPSF30 (29) or with the empty plasmid, as a control. At 30 hpt, cells were lysed with 50 mM Tris-HCl (pH 7.5), 500 mM NaCl, 5 mM EDTA, 5% glycerol, and 0.5% Igepal supplemented with Complete Mini protease inhibitor cocktail (Pierce). In parallel, HA-tagged NS1 variants were synthesized in vitro by using pcDNA3 plasmids and a TNT7 transcription/translation kit (Promega) following the manufacturer's recommendations. Cleared cell lysates expressing FLAG-CPSF30 were incubated overnight at 4°C with the in vitro-synthesized NS1 proteins and 30 μl of an anti-FLAG affinity resin (Sigma-Aldrich). After washing three times in Tris-buffered saline (TBS) containing 0.1% Tween 20, precipitated proteins were dissociated from the resin by use of disruption buffer and high temperature (95°C) and then analyzed by Western blotting as described above, using rabbit anti-NS1 and anti-FLAG (CPSF30) pAbs.

Subcellular localization of NS1 proteins.

Subconfluent plates of 293T cells seeded on coverslips (24-well-plate format, 1.5 × 10⁵ cells/well) were transiently transfected with 1.25-μg aliquots of the pcDNA3 plasmids expressing the NS1 proteins fused to an HA tag or with the empty plasmid, as a control, using DNA-In (MTI-GlobalStem). At 24 h posttransfection, the cells were fixed and permeabilized using 4% formaldehyde and 0.5% Triton X-100 in PBS for 20 min at room temperature. After washing with PBS, cells were incubated in blocking solution (2.5% BSA in PBS) for 1 h at room temperature and then incubated with an anti-HA pAb (H6908; Sigma-Aldrich), diluted 1:1,000 in 1% BSA, for 2 h at room temperature. After washing with PBS, cells were incubated with an Alexa Fluor 488-conjugated mouse anti-rabbit IgG secondary antibody (A-11008; Thermo Fisher Scientific), diluted 1:500 in 1% BSA, for 1 h at room temperature. Coverslips were mounted using ProLong Gold antifade reagent with DAPI (4′,6-diamidino-2-phenylindole; Invitrogen). Representative images were taken using a confocal microscope (FV1000; Olympus).

Assays to detect IFN and proinflammatory cytokines in infected cells.

To measure the levels of IFN produced by infected cells, confluent monolayers of MDCK cells (24-well-plate format, 5 × 10⁴ cells/well, triplicates) constitutively expressing GFP and Fluc reporter genes under the control of the IFN-β promoter (MDCK IFNβ-GFP/IFNβ-Fluc) (50) were mock infected or infected (MOI = 5) with pH1N1 viruses encoding the NS1-wt, NS1-6mut, and NS1-3mut proteins (pH1N1/NSs-wt, pH1N1/NSs-6mut, and pH1N1/NSs-3mut, respectively). At 12 hpi, activation of the IFN-β promoter was determined by assessing GFP expression under a fluorescence microscope or by quantifying Fluc expression in cell lysates by using a luciferase reporter assay (Promega) and a Lumicount luminometer. Influenza virus infection levels were evaluated by immunofluorescence assay using an anti-NP MAb (HB-65) as described above. In addition, TCS of infected MDCK cells were collected, and viruses were UV inactivated as previously described (64). Fresh MDCK cells seeded in 96-well plates were treated (triplicate wells) with the UV-inactivated TCS for 24 h and then infected (MOI = 5) with a recombinant Newcastle disease virus expressing GFP (rNDV-GFP) (64). The GFP intensity was measured at 24 hpi by use of a microplate reader (DTX880; Beckman Coulter). GFP expression of mock-treated cells was considered 100%. Mean values and SD were calculated using Microsoft Excel software. In addition, human A549 cells (24-well-plate format, 5 × 10⁴ cells/well, triplicates) were mock infected or infected (MOI = 5) with the pH1N1/NSs-wt, pH1N1/NSs-6mut, and pH1N1/NSs-3mut viruses. Then, at 12 hpi, levels of IFN-β, IFIT2, CCL2/monocyte chemotactic protein 1 (MCP1), CXCL10, and TNF induction were analyzed. To that end, total RNA was extracted using an RNeasy minikit (Qiagen). RT reactions were performed at 37°C for 2 h by use of a High Capacity cDNA reverse transcription kit (Applied Biosystems) and oligo(dT) primers to amplify mRNAs, starting from 300 ng of total RNA. qPCRs were performed using TaqMan gene expression assays (Applied Biosystems) specific for the IFN-β (Hs01077958_s1), IFIT2 (Hs00533665_m1), CCL2 (Hs00234140_m1), CXCL10 (Hs00171042_m1), and TNF (Hs00174128_m1) human genes. Quantification was achieved using the 2^−ΔΔCT method (65). In addition, 12 h after IAV infection, TCS were collected, and IAV was inactivated by exposure to shortwave (254 nm) UV radiation for 40 min at a distance of 6 cm (64). Fresh human A549 cells (96-well-plate format, 5 × 10⁴ cells/well, triplicates) were treated with the UV-inactivated TCS for 24 h and then infected (MOI = 0.1) with rVSV-GFP for 16 h (64). GFP expression was analyzed as described above.

Mouse experiments.

Female 6-week-old C57BL/6 mice were purchased from the Jackson Laboratory and maintained in a pathogen-free environment in the animal care facility at the University of Rochester. All animal protocols were approved by the University of Rochester Committee of Animal Resources and complied with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Research Council (66). Mice were anesthetized intraperitoneally (i.p.) with 2,2,2-tribromoethanol (Avertin) and then inoculated intranasally (i.n.) with 30 μl of the indicated pH1N1/NSs-wt, pH1N1/NSs-6mut, pH1N1/NSs-3mut, or pH1N1-wt (encoding a nonsplit NS segment) virus and monitored daily for morbidity (body weight loss; n = 5) and mortality (survival; n = 5). Mice showing a >25% loss of body weight were considered to have reached the experimental endpoint and were humanely euthanized. Virus replication was evaluated by determination of viral titers in the lungs at 2, 5, and 9 dpi. To that end, mice (n = 3) were sacrificed, and lungs were extracted and homogenized. Virus titers were determined by an immunofocus assay (giving FFU per milliliter) with MDCK cells, as described above.

Levels of IFN-β, IFIT2, CCL2, and TNF induction were analyzed in mouse lungs at 2, 5, and 9 dpi. To that end, mice (n = 3) were sacrificed, and their lungs were extracted and incubated in RNAlater (Ambion) at 4°C for 24 h prior to freezing at −80°C. Lungs were homogenized in lysis buffer by use of a gentleMACS dissociator (Miltenyi Biotec), and total RNA was extracted by use of an RNeasy minikit (Qiagen). RT reactions were performed at 37°C for 2 h by use of a High Capacity cDNA reverse transcription kit (Applied Biosystems) and a dT primer to amplify mRNAs, starting from 300 ng of total RNA. qPCRs were performed using TaqMan gene expression assays (Applied Biosystems) specific for the IFN-β (Mm00439552_s1), IFIT2 (Mm00492606_m1), CCL2 (Mm00441242_m1), and TNF (Mm00443258_m1) murine genes. Quantification was achieved using the 2^−ΔΔCT method (65).

Accession number(s).

The NS1 sequences obtained from subjects ACU001, ACU004, ACU005, ACU007, ACU008, ACU009, ACU012, ACU017, and ACU022 are available in GenBank under accession numbers MF150392, MF150397, MF150393, MF150395, MF150400, MF150399, MF150398, MF150394, and MF150396, respectively.