Host specificity of the anti-interferon and anti-apoptosis activities of parainfluenza virus P/C gene products

Abstract

Human parainfluenza virus type 1 (HPIV-1) and Sendai virus (SeV) are highly homologous in structure and sequence, whilst maintaining distinct host ranges. These viruses express accessory proteins from their P/C gene that are known to have activities against innate immunity. The accessory proteins expressed from the P/C gene of these viruses are different. In addition to the nested set of C proteins, SeV expresses V protein from edited P mRNA, which is not expressed by HPIV-1. This study evaluated the host specificity and role of the P/C gene products in anti-interferon (IFN) and anti-apoptosis activity by characterizing a recombinant SeV, rSeVhP, in which the SeV P/C gene was replaced with that of HPIV-1. Unlike SeV, rSeVhP infection strongly activated IFN regulatory transcription factor (IRF)-3 and nuclear factor-κB, resulting in an increased level of IFN-β induction compared with SeV in murine cells. In contrast, activation of IRF-3 was not observed in rSeVhP-infected human A549 cells. rSeVhPSV, which expressed SeV V protein from an inserted gene in rSeVhP, induced less IFN-β than rSeVhP, suggesting that V contributes to the suppression of IFN production in murine cells. Furthermore, rSeVhP induced apoptotic cell death in murine but not in A549 cells. These data indicate the functional difference in P/C gene products from SeV and HPIV-1 in antagonizing IFN induction and apoptosis, which is likely to be one of the major factors for pathogenicity in specific hosts.

INTRODUCTION

The family Paramyxoviridae comprises negative-strand RNA viruses including many important human and animal pathogens that share a high degree of genetic and structural homology. Human parainfluenza virus type 1 (HPIV-1), a human respiratory virus, and Sendai virus (SeV), a pneumotropic murine virus, are one example, sharing sequence identities between 53 and 87 % (Takimoto et al., 2000), whilst maintaining a strict host range (Sangster et al., 1995). There have been few studies on the viral genes that are critical for efficient growth in specific hosts. One of the genes that may perform a significant role in host-range restriction in parainfluenza viruses is the P/C gene, as the P/C genes of SeV and HPIV-1 express different proteins. The SeV P/C gene expresses P, C, C′, Y1, Y2 and V (Curran et al., 1991). The C′, C, Y1 and Y2 proteins are a nested set of proteins expressed from intact P and V mRNAs. V protein is expressed from an edited P gene transcript, which includes a single non-templated G residue. In contrast, only the P and C open reading frames (ORFs) have been detected in HPIV-1 P/C gene analysis; there is no evidence of an HPIV-1 ORF capable of encoding a V protein (Matsuoka et al., 1991; Power et al., 1992). These accessory proteins are non-essential for viral replication in vitro, but are required for interferon (IFN) antagonism and virus virulence (Kato et al., 1997; Komatsu et al., 2004).

Viral methods of counteracting innate immunity include inhibition of the Jak/STAT pathway and suppression of IFN-β induction (Conzelmann, 2005; Haller et al., 2006; Randall & Goodbourn, 2008). Many studies have identified the V and C proteins of various paramyxoviruses as significant in counteracting IFN (Gotoh et al., 2001; Randall & Goodbourn, 2008). Parainfluenza virus 5 (PIV-5) V protein blocks both types I (IFN-α/β) and II (IFN-γ) IFN signalling by initiating the ubiquitination and subsequent degradation of STAT1 (Didcock et al., 1999). SeV and HPIV-3 C proteins inhibit IFN signalling by precluding the appropriate phosphorylation of STAT1 (Komatsu et al., 2002; Young et al., 2000). HPIV-1 C protein also prevents the nuclear translocation of STAT1 and STAT2 in human cells (Bousse et al., 2006).

Viruses also suppress IFN-β induction to circumvent the host immune response. In order for IFN-β to be induced, viral infection or dsRNA is detected by retinoic acid inducible gene I (RIG-I), melanoma differentiation-associated gene 5 (MDA-5) or dsRNA-activated protein kinase stimulating activation of transcription factors nuclear factor (NF)-κB, interferon regulatory transcription factor (IRF)-3 and ATF-2/c-Jun (Andrejeva et al., 2004; Kato et al., 2006; Strahle et al., 2007). Activated NF-κB, IRF-3 and ATF-2/c-Jun are translocated to the nucleus where they form a transcription complex on the IFN-β promoter along with CBP/p300 (Randall & Goodbourn, 2008). Antagonism of IFN-β induction has been attributed to the V and C proteins of SeV, which block activation of IRF-3 and NF-κB (Kiyotani et al., 2007; Komatsu et al., 2004; Strahle et al., 2007). Additionally, the C protein of HPIV-1 has been shown to inhibit IFN-β production and IRF-3 activation in infected human cells (Bartlett et al., 2008b; Van Cleve et al., 2006). Regulation of virally induced apoptosis is another viral mechanism of evading host protection. The V and C proteins of SeV, as well as HPIV-1 C protein, have been shown to be important elements in eliminating apoptotic cell death (Bartlett et al., 2008a; Kato et al., 1997; Koyama et al., 2003).

Although regulation of IFN-β induction and antagonism of the IFN pathway by P/C gene products has been well demonstrated, the specificity of the activities against innate immunity in different host cells is less apparent. It is also unclear why SeV expresses V protein, which is not encoded by the HPIV-1 gene. To analyse the functional specificity of the P/C gene products of these viruses, we characterized a recombinant SeV, rSeVhP, in which the P/C gene of SeV was replaced with that of HPIV-1. Our analysis demonstrated that rSeVhP infection blocks the Jak/STAT pathway in murine cells. However, unlike SeV, rSeVhP infection strongly induced IFN-β production and apoptosis in the murine cells we tested. Further analysis suggested a contribution by SeV V protein to activity against innate immunity and to anti-apoptosis in infected murine cells. Our data indicate a functional difference in the ability of HPIV-1 and SeV P/C gene products to counteract innate immunity.

METHODS

Cells and viruses.

LLC-MK₂, MLg2908, MM14.Lu, 293T, HeLa T4⁺ and A549 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) with 10 % fetal calf serum. SeV strain Z, HPIV-1 strain C35 and recombinant SeVs (rSeVhP, rSeVhPSV and rSeVhPSC) were grown in LLC-MK₂ cells in DMEM supplemented with acetylated trypsin (2 μg ml⁻¹). Recombinant vaccinia virus vTF7.3 (Fuerst et al., 1986) was grown in HeLa T4⁺ cells.

cDNA synthesis and cloning.

The full-genome cDNA of rSeVhP (pSeVhP) was constructed as follows. We used pSeV(+)AN (Alymova et al., 2004), which includes NotI and AscI sites upstream and downstream, respectively, of the translation initiation codon of the haemagglutinin–neuraminidase (HN) gene of pSeV(+)(Kato et al., 1996). First, we created SbfI and PacI sites upstream and downstream of the P gene ORF, respectively, using a QuikChange Site-Directed Mutagenesis kit (Stratagene), and designated this construct pSeV(+)ANPS. Next, the P gene of pSeV(+)ANPS was replaced with that of HPIV-1, which was amplified by PCR from pTF1HPIV-1P (Takimoto et al., 2000) using Pfu DNA polymerase (Stratagene). Primers used for PCR included the SbfI and PacI sites.

pSeVhPSC, used to rescue rSeVhPSC, was constructed as follows. First, the translation initiation codon of the P protein in the SeV P gene was mutated from ATG to CTG by site-directed mutagenesis as described above. The coding region for the C proteins was then amplified by PCR using primers that contained NotI sites in the forward and reverse primers. The reverse primer also included transcription termination, intergenic and transcription initiation signals for expression of the following HN gene. The PCR product was cleaved with NotI and ligated to the NotI site of pSeVhP. pSeVhPSV, used to rescue rSeVhPSV, was constructed as follows. The editing site of the SeV P/C gene in pTF1SeVP for expression of V was mutated from AAAAGGG to AGAAGGGG, so that the V protein would be expressed without mRNA editing. A translation termination codon was inserted in-frame in the C protein codon (TCG→TAG at nt 231–233 of the P/C gene) to stop expression of full-sized C proteins. The region including the V coding region was amplified by PCR using primers that included NotI sites in the forward and reverse primers, and the PCR product was inserted at the same NotI site in pSeVhP.

Rescue of recombinant SeVs.

The recombinant SeVs were rescued by a reverse genetics system from pSeVhP, pSeVhPSC or pSeVhPSV as described previously (Bousse et al., 2002). Briefly, 293T cells infected with UV-inactivated vTF7.3 were transfected with full-genome rSeV cDNAs together with the supporting plasmids pTF1SVNP, pTF1SVP and pTF1SVL. The cells were treated with trypsin and lysed by freezing and thawing. The viruses in the lysates were passaged in LLC-MK₂ cells. Rescued viruses were plaque-purified and stock virus was grown in LLC-MK₂ cells.

Virus growth in LLC-MK₂ cells.

Cells in a six-well plate were infected with the viruses at an m.o.i. of 0.01 and incubated at 34 °C in 2 ml DMEM containing 0.15 % BSA and acetylated trypsin (2 μg ml⁻¹). Culture medium (200 μl) was harvested every 12 h after infection. Virus titres in the supernatant were determined by tissue culture infection in LLC-MK₂.

Western blot analysis.

To detect nucleoprotein (NP) synthesis after IFN treatment, infected cells in a 12-well plate were lysed with 200 μl TNE buffer [10 mM Tris/HCl (pH 7.4), 150 mM NaCl, 0.5 % NP-40, 1 mM EDTA]. After clarification by centrifugation, the protein concentrations of the cell lysates were ascertained by using a BCA assay (Pierce Chemicals). Proteins were separated by SDS-PAGE, electrotransferred onto PVDF membranes and probed with specific monoclonal antibodies (mAbs) against SeV or HPIV-1 NP, and against actin as an internal control. The NP and actin band intensities were quantified by using Bio-Rad Quantity One 1-D analysis software. For STAT1 levels in infected cells, lysates were prepared as above and the blot was probed with STAT1 polyclonal antibody (pAb) (Santa Cruz Biotechnology). To detect V or C proteins in infected cells, we used anti-SeV V or C rabbit serum (Kato et al., 1997).

Analysis of IFN-β.

MM14.Lu cells were infected with SeV, rSeVhP, rSeVhPSV or rSeVhPSC at an m.o.i. of 1.0 for 1 h and cultured for 24 h. Supernatants were collected and IFN-β levels were detected using a mouse IFN-β ELISA kit (PBL Biomedical Laboratories). Total cellular RNAs were extracted from infected cells using an illustra RNAspin Mini Extraction kit (GE Healthcare). One microgram of extracted RNA was used for RT-PCR using AffinityScript Multiple Temperature Reverse Transcriptase (Stratagene) and TaKaRa Ex Taq (TaKaRa Bio) with primers specific for murine IFN-β (5′-CCTGCTGTGCTTCTCCAC-3′ and 5′-CAACAATAGTCTCATTCC-3′) and murine actin (5′-CACACCTTCTACAATGAG-3′ and 5′-CCGGCCAGGTCCAGACGC-3′). Samples were collected after 25 cycles for actin and after 35 cycles for IFN-β. PCR products were electrophoresed on a 1 % agarose gel and stained with ethidium bromide.

IRF-3 dimerization assay.

Oligomer formation of IRF-3 was determined as described previously (Dillon & Parks, 2007; Iwamura et al., 2001). Briefly, infected cells were lysed in buffer containing 50 mM Tris/HCl (pH 8.0), 150 mM NaCl, 1 % NP-40, 1 mM EDTA, 0.1 mg leupeptin ml⁻¹, 1 mM PMSF and 1 mM sodium orthovanadate. After centrifugation, supernatants containing 80 μg protein were analysed on a 7.5 % non-denaturing acrylamide gel (Ready Gel; Bio-Rad Laboratories) with 0.2 % deoxycholate added to the cathode buffer, and transferred to a PVDF membrane. IRF-3 was detected by Western blot analysis using rabbit anti-IRF3 antibody (Zymed Laboratories). Band intensities were quantified and the ratio of dimer to monomer was calculated for each blot.

Immunofluorescence assays.

For STAT1 localization, MLg2908 cells were infected with the viruses at an m.o.i. of 0.5. After 16 h, murine IFN-β (1000 U ml⁻¹) was added and the cells were incubated for 2 h at 34 °C. Fixed and permeabilized cells were reacted with anti-STAT1 rabbit pAb (diluted 1 : 200; Santa Cruz Biotechnology) and anti-HPIV-1 P or anti-SeV P mAb (diluted 1 : 2000), followed by Texas Red-conjugated goat anti-mouse (Molecular Probes) and fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit (Molecular Probes). For NF-κB localization, cells infected with the viruses at an m.o.i. of 0.8 were incubated for 24 h and reacted with anti-p65 antibody (diluted 1 : 100; Santa Cruz Biotechnology) and anti-SeV P or anti-HPIV-1 P mAb, followed by FITC-conjugated goat anti-rabbit or Texas Red-conjugated goat anti-mouse. For IRF-3 localization, cells were infected with the viruses for 12 h and reacted with rabbit anti-mouse IRF-3 antibody (Zymed Laboratories) or anti-human IRF-3 antibody (Santa Cruz) at a 1 : 100 dilution, along with anti-P (SeV or HPIV-1) or anti-NP (SeV or influenza strain PR8) at a 1 : 2000 dilution.

Apoptosis assay.

MLg2908 cells in six-well plates were infected with the viruses at an m.o.i. of 3 and incubated with or without 0.5 μM geldanamycin for 40 h. Apoptotic cells were detected using a Vybrant Apoptosis Assay kit #2 (Invitrogen) and analysed on a Becton Dickinson FACSCalibur.

RESULTS

Rescue and characterization of rSeVhP, rSeVhPSC and rSeVhPSV

SeV C and V proteins have anti-IFN activities, are crucial for multiplication in mice in vivo and are known to block the Jak/STAT pathway and IFN production in human and murine cells (Garcin et al., 2003; Kato et al., 2007; Komatsu et al., 2004; Strahle et al., 2007). The HPIV-1 P/C gene expresses C but not V protein; however, the C protein blocks the Jak/STAT pathway in normal human lung fibroblast MRC-5 cells (Bousse et al., 2006) and inhibits IRF-3 activation and subsequent IFN production in human respiratory epithelial A549 cells and primary human airway epithelial (HAE) cells (Bartlett et al., 2008b; Van Cleve et al., 2006). The specificity of the anti-IFN activities of HPIV-1 C protein has not been characterized in detail. To determine the functional specificity of the HPIV-1 and SeV P/C gene products, we created a recombinant SeV, rSeVhP, whose P/C gene was replaced with that of HPIV-1, using reverse genetics. rSeVhP was successfully rescued from transfected 293T cells. To confirm that rSeVhP contained the HPIV-1 P/C gene and expressed HPIV-1 but not SeV P protein, we purified the virus and analysed the viral proteins by SDS-PAGE. The P protein of rSeVhP was detected at the same position as HPIV-1 P, which migrates slower than SeV P (Fig. 1a). Western blot analysis using mAbs specific to SeV or HPIV-1 P confirmed that rSeVhP does not express SeV P protein (Fig. 1b).

Fig. 1.

Rescue and characterization of rSeVhP, rSeVhPSC and rSeVhPSV. (a) Purified SeV, HPIV-1 and rSeVhP virions were analysed by SDS-PAGE. A graph of the retention factor (Rf) versus band intensity was prepared from the gel image. Rf was calculated as migration distance of substance/migration of tracking dye front. Note that rSeVhP P protein is located at the same position as HPIV-1 P protein, which migrates slower than SeV P. The rSeVhP HN protein migrates the same distance as SeV HN, which migrates slower than HPIV-1 HN. SP, SeV P; HP, HPIV-1 P; SHN, SeV HN; HHN, HPIV-1 HN. (b) SDS-PAGE and Western blot analysis of the P protein of purified SeV, HPIV-1 and rSeVhP. Proteins of the purified virions (1 μg) were detected by Coomassie staining (upper panel) or by Western blotting using anti-SeV- (SP) or HPIV-1 P (HP)-specific mAb (lower panels). (c) Schematic representation of rescued rSeVhP containing the sequence for the SeV V or C gene. The SeV V or C gene was inserted between the F and HN genes. Black and grey boxes represent transcription initiation and termination sequences, respectively. Numbers represent translation initiation sites, except for 1053, which denotes the position of the non-template C residue insertion for V protein expression. IG, intergenic sequence. Nt 231–233 of the V gene and nt 104–106 of the C gene were mutated to terminate expression of the C proteins and the N-terminal region of SeV P or V protein, respectively.

We also rescued additional rSeVhP viruses, rSeVhPSC and rSeVhPSV, which express SeV C or V protein, respectively (Fig. 1c), to determine whether SeV C or V protein is required for activity against innate immunity in murine cells. These two viruses had additional genes that were designed to express either the nested set of C proteins (C′, C, Y1 and Y2) or the V protein of SeV. Both viruses were rescued successfully. Expression of the inserted C or V proteins was confirmed by Western blot analysis using specific antibodies against C or V proteins (see Fig. 3d).

Fig. 3.

Translocation of NF-κB and IRF-3 in infected murine or human cells. (a, b) Immunofluorescence assay of infected murine MLg2908 cells using anti-p65 (a) or anti-IRF3 (b) rabbit pAb and anti-rabbit IgG–FITC. Cells were also reacted with anti-P mAb and anti-mouse IgG–Texas Red. Arrows and arrowheads indicate localization or lack of localization in the nucleus, respectively. (c) Human A549 cells were infected with the viruses and localization of IRF-3 in infected cells was determined as in (b). (d) Western blot analysis of viral proteins expressed in infected MLg2908 cells. Equal amounts of infected cell lysate were reacted with antibodies against HPIV-1 P, SeV NP, SeV V or SeV C.

We determined the growth characteristics of the viruses in LLC-MK₂ cells, which supported multi-step replication of both SeV and HPIV-1. Cells were infected with the viruses at an m.o.i. of 0.01 and the virus growth kinetics were determined. rSeVhP grew at similar levels to SeV and HPIV-1 in the cell line (Fig. 2), supporting the previous finding that SeV P protein can be replaced with HPIV-1 P and still remain functional with SeV L for polymerase activity. Growth of rSeVhPSV was as efficient as rSeVhP, but rSeVhPSC replicated to a lower level than rSeVhP (Fig. 2), possibly due to the additional expression of SeV C, which is known to inhibit transcription/replication of the viral genome (Kato et al., 2004).

Fig. 2.

Growth kinetics of the viruses. The amount of infectious virus in the culture medium at various times after infection was quantified by plaque assay in LLC-MK₂ cells. The data represent the mean±sem of three different experiments. ◊, SeV; □, rSeVhP; ▵, HPIV-1; ○, rSeVhPSV; ×, rSeVhPSC.

Specificity of the P/C gene products in suppressing IFN-β induction in murine cells

We next analysed IFN induction in murine cells infected with rSeVs. To determine the effect of virus infection on activation of NF-κB and IRF-3 and therefore induction of IFN-β, localization of NF-κB or IRF-3 in infected murine MLg2908 cells was determined by immunofluorescence (Fig. 3). Approximately 90 % of cells infected with rSeVhP showed NF-κB and IRF-3 nuclear localization, which was in sharp contrast to less than 10 % found in SeV-infected cells (Table 1). Infection with rSeVhPSC resulted in extensive nuclear localization of NF-κB and IRF-3 as observed with rSeVhP, indicating that the additional expression of SeV C protein did not reduce the level of NF-κB and IRF-3 activation. However, in rSeVhPSV-infected cells, the nuclear localization of NF-κB and IRF-3 was suppressed approximately twofold and fivefold, respectively, compared with rSeVhP- and rSeVhPSC-infected cells (Table 1). The level of HPIV-1 C protein expressed by rSeVhP and rSeVhPSV in MLg2908 cells was almost equal, suggesting that the expressed V protein suppressed NF-κB and IRF-3 activation (Fig. 3d).

Table 1.

Quantification of the percentage of cells with NF-κB and IRF-3 nuclear localization

Infections were carried out as described in Methods. Infected cells (500 per sample) were counted in three separate experiments. The percentage of infected cells with NF-κB or IRF-3 nuclear localization out of the total infected cell number was calculated as mean±sem (95 % confidence limits: P≤0.05). nd, not detected.

Virus	Nuclear localization (%)
	NF-κB	IRF-3
Mock	nd	nd
SeV	9.33±6.43	6.03±2.08
rSeVhP	92.7±3.06	84.8±6.54
rSeVhPSV	42.0±7.21	16.0±14.8
rSeVhPSC	89.3±4.62	84.9±2.68

We next determined IRF-3 activation of these viruses in human lung A549 cells. In contrast to the results in murine MLg2908 cells, nuclear translocation of IRF-3 was not evident in any of the recombinant SeV-infected A549 cells (Fig. 3c). These results are consistent with previous reports showing that HPIV-1 C protein suppresses IFN induction in A549 and HAE cells (Bartlett et al., 2008b; Van Cleve et al., 2006). Nuclear localization of IRF-3 was clearly detected in A549 cells infected with influenza strain PR8 ΔNS1 virus, which is known to be a strong inducer of IFN (Fig. 3c) (Talon et al., 2000). These results suggested that strong activation of NF-κB and IRF-3 in MLg2908 cells is not due to the presence of defective interfering particles in our stock viruses.

IRF-3 activation by virus infection was also analysed by determining IRF-3 dimerization in infected cells. Consistent with the nuclear localization data, IRF-3 dimerization in rSeVhP-infected cells was significantly higher than that of SeV (Fig. 4a). The level of IRF-3 dimer in rSeVhPSC was approximately equivalent to that of rSeVhP. In contrast, in rSeVhPSV-infected cells, the level of IRF-3 dimerization was reduced by 50 % compared with cells infected with rSeVhP or rSeVhPSC.

Fig. 4.

IRF-3 activation and IFN-β induction in infected murine cells. (a) IRF-3 dimers in cell lysates were analysed by native PAGE followed by Western blotting using rabbit anti-IRF3 antibody. The bar graph represents the percentage of IRF-3 dimer out of the total IRF-3. Error bars represent sem at 95 % confidence limits. (b) IFN-β mRNA transcript levels in infected cells. RT-PCR products of IFN-β and actin from cells mock infected or infected with SeV, rSeVhP, rSeVhPSV or rSeVhPSC, or transfected with poly(I : C).

To confirm the results of IRF-3 and NF-κB activation in murine cells, we determined the level of IFN-β transcripts by RT-PCR. As shown in Fig. 4(b), the level of IFN-β transcripts was much higher in rSeVhP- or rSeVhPSC-infected murine cells than that detected in wild-type SeV- or rSeVhPSV-infected cells, consistent with the results from NF-κB and IRF-3 activation. In addition, we measured the amount of IFN-β released into the medium by ELISA. SeV-infected cells produced low levels of IFN-β (0.3 ng ml⁻¹), whilst rSeVhP and rSeVhPSC produced high levels of IFN-β (20.1 and 14.4 ng ml⁻¹, respectively). rSeVhPSV produced 4.8 ng IFN-β ml⁻¹, which was less than either rSeVhP or rSeVhPSC. These results indicated that HPIV-1 P/C gene products are unable to block NF-κB and IRF-3 activations and subsequent IFN release in MLg2908 cells infected with rSeVhP, and that SeV V contributes to the suppression of IFN induction.

Both SeV and rSeVhP prevent IFN-induced nuclear translocation of STAT1 in murine cells

To evaluate the effect of rSeVhP infection on the Jak/STAT pathway in murine cells, we infected MLg2908 cells with virus, and STAT1 levels in cell lysates were analysed by Western blotting (Fig. 5a). Degradation of STAT1, as seen with PIV-5 (Precious et al., 2005), was not observed. Next, nuclear translocation of STAT1 was assessed after infection with virus. Cells infected with the viruses at an m.o.i. of 0.5 were treated with exogenous IFN-β, and localization of STAT1 was visualized by immunofluorescence. As shown in Fig. 5(b), all of the viruses blocked nuclear translocation of STAT1 induced by exogenously added IFN-β. These data indicated that, although the HPIV-1 P/C gene products were unable to block NF-κB and IRF-3 activation, they were capable of blocking the Jak/STAT pathway in murine cells, as in human MRC-5 cells (Bousse et al., 2006).

Fig. 5.

Effect of virus infections on STAT1 protein levels and nuclear translocation. (a) STAT1 levels in cells mock infected (lane 1) or infected with SeV (lane 2), rSeVhP (lane 3), rSeVhPSC (lane 4) or rSeVhPSV (lane 5) were determined by Western blotting using an anti-STAT1 antibody. (b) Localization of STAT1 in infected murine cells. Exogenous murine IFN-β (1000 U ml⁻¹) was added to infected cells and processed for immunofluorescence using anti-STAT1 rabbit serum and anti-SeV or HPIV-1 P mAb. Images of STAT1 staining alone (left panels) and merged images of STAT1 and P proteins (right panels) are shown.

SeV, but not rSeVhP, replicates in murine cells pre-treated with IFN-β

We reported previously that SeV is able to replicate in murine cells pre-treated with IFN-β, but not in human cells that are in an IFN-induced antiviral state (Bousse et al., 2006). We next determined the ability of the recombinant viruses to replicate in murine cells in an antiviral state induced by IFN-β pre-treatment. First, murine cells were treated with IFN-β (500 U ml⁻¹) for 24 h and then infected with virus. At 16, 24 and 48 h post-infection (p.i.), NP levels in cell lysates were detected by Western blot analysis and normalized against the actin control (Fig. 6a). In SeV-infected cells, NP production was restored by 48 h p.i. The amount of NP at 48 h p.i. was eightfold higher than that at 16 h. In contrast, the levels of NP detected in rSeVhP- and rSeVhPSC-infected cells were limited, even at 48 h p.i., only increasing 1.6-fold from 16 to 48 h p.i. The increase in NP levels in rSeVhPSV infection was fourfold, suggesting that SeV V protein contributes to viral replication in cells pre-treated with IFN.

Fig. 6.

SeV and rSeVhPSV, but not rSeVhP or rSeVhPSC, restore replication in IFN-pre-treated murine cells. (a) MM14.Lu cells pre-treated with murine IFN-β were infected with SeV, rSeVhP, rSeVhPSC or rSeVhPSV, and levels of viral NP protein were analysed by Western blotting using anti-SeV NP mAb. Cellular actin was also detected as a loading control and used to normalize the NP levels. (b) Exogenous IFN-β (500 U ml⁻¹) was added at 16 h p.i. Cell lysates at 8 or 32 h after the addition of IFN (24 or 48 h p.i.) were analysed for the levels of NP and actin as a loading control.

Even though the Jak/STAT pathway was blocked in cells infected with rSeVhP and rSeVhPSC (Fig. 5b), it was possible that continuous stimulation by IFN-β produced in response to infection (Fig. 4b) maintained the antiviral state and prevented genome replication. To determine whether exogenous IFN-β added after infection suppressed SeV replication, we treated cells with murine IFN-β (500 U ml⁻¹) for 24 h and then infected them with SeV or rSeVhPSV. At 16 h p.i., an additional 500 U IFN-β ml⁻¹ was added. The cells were cultured for a further 8 and 32 h, and the level of NP was determined by Western blotting (Fig. 6b). Addition of exogenous IFN-β after infection did not significantly affect the replication of SeV or rSeVhPSV, suggesting that continuous stimulation of IFN from infected cells is not likely to be the factor that suppresses replication of rSeVhP or rSeVhPSC.

rSeVhP infection induces significant cytopathogenicity due to apoptosis in murine cells compared with SeV

Several studies have suggested that the accessory proteins of the P/C/V genes of various paramyxoviruses are able to antagonize cell death and apoptotic activity (Bartlett et al., 2008a; He et al., 2002; Kato et al., 1997). To determine the effect and specificity of the P/C gene products in cell death, we infected MLg2908 and A549 cells with the viruses and observed the cytopathic effect (CPE). rSeVhP infection induced a significant level of cell death in MLg2908 cells compared with infected A549 cells or SeV-infected MLg2908 or A549 cells (Fig. 7a). Quantification of the number of live cells at 48 h after infection showed a threefold increase in cell death in rSeVhP-infected murine cells compared with SeV-infected cells. Infection with rSeVhPSC also induced high levels of CPE in murine cells, whilst rSeVhPSV infection was much less cytopathic (Fig. 7a).

Fig. 7.

Induction of apoptosis in MLg2908 cells infected with rSeVhP and rSeVhPSC. (a) Morphological changes in A549 and MLg2908 cells infected with various viruses. Infected cells were cultured for 48 h and analysed by phase-contrast microscopy. Magnification ×100. (b) Levels of apoptosis in infected cells. Infected MLg2908 cells were cultured in the presence or absence of 0.5 μM geldanamycin. Apoptotic cells were detected and analysed by flow cytometry. The percentage of apoptotic cells in the absence (black bars) or presence (grey bars) of geldanamycin is shown as the mean±sem of three separate experiments (P≤0.05).

In order to understand the mechanism of cytopathogenicity induced by rSeVhP and rSeVhPSC infections, we determined apoptosis levels in infected cells by flow cytometry. rSeVhP- and rSeVhPSC-infected murine cells exhibited a higher percentage of apoptotic cells compared with SeV- or rSeVhPSV-infected cells (Fig. 7b). These data suggested that the proteins produced by the HPIV-1 P/C gene do not protect infected cells from apoptosis, whilst SeV V protein provides a protective effect.

To determine the role of NF-κB and IRF-3 activation in apoptotic induction in rSeVhP- and rSeVhPSC-infected cells, we treated infected cells with geldanamycin, which has been shown to block NF-κB and IRF-3 activation. MLg2908 cells infected with the viruses were cultured in the presence of 0.5 μM geldanamycin, which blocked approximately 60 % of NF-κB nuclear translocation in rSeVhP-and rSeVhPSC-infected cells (data not shown). As shown in Fig. 7(b), apoptotic death was strongly inhibited by geldanamycin, indicating that NF-κB and/or IRF-3 activation plays a role in apoptosis induction in rSeVhP- or rSeVhPSC-infected cells.

DISCUSSION

It is well established that paramyxovirus P/C gene products have anti-IFN activities, which play a major role in viral pathogenicity towards their specific hosts. We showed previously that HPIV-1 infection or expression of HPIV-1 C protein blocks the Jak/STAT pathway in human lung MRC-5 cells (Bousse et al., 2006). Another study with rescued HPIV-1 containing mutations in the C protein showed that HPIV-1 C protein is responsible for blocking IRF-3 activation and IFN induction in human A549 and HAE cells (Bartlett et al., 2008b; Van Cleve et al., 2006). These studies suggested that HPIV-1 C protein is capable of blocking both IFN induction and signalling in its natural host, human cells. In this study, we replaced the P/C gene of SeV with that of HPIV-1 and determined the innate immune response in murine cells. We found that, in the murine cell lines tested, the HPIV-1 P/C gene products were unable to block NF-κB and IRF-3 activation and IFN induction induced by rSeVhP infection (Figs 3 and 4, Table 1), could not overcome the antiviral state induced by IFN pre-treatment (Fig. 6) and were unable to prevent apoptosis (Fig. 7), but they were able to block the Jak/STAT pathway (Fig. 5).

Unlike wild-type SeV, rSeVhP was not able to block NF-κB and IRF-3 activation or IFN induction in infected murine cells (Figs 3 and 4), suggesting that the HPIV-1 C protein is unable to block the IFN induction pathway induced by SeV infection in at least some murine cells, in contrast to its ability to block IFN activation in human respiratory A549 (Fig. 3c) and primary HAE cells (Bartlett et al., 2008b; Van Cleve et al., 2006). To determine whether amino acid differences between SeV and HPIV-1 C proteins were responsible for this difference, we rescued rSeVhPSC, which expresses the nested set of SeV C proteins from additional genomic transcripts (Fig. 1c). Unexpectedly, rSeVhPSC did not inhibit activation of NF-κB and IRF-3 or IFN induction (Figs 3 and 4, Table 1), suggesting that a difference in the C protein sequence is not the reason for the deficiency of rSeVhP in blocking IFN induction, although it is possible that additional expression of SeV C protein causes unbalanced genome replication and activation of IFN production. Interestingly, however, rSeVhPSV, which expresses SeV V protein in addition to HPIV-1 C, reduced the activation of NF-κB and IRF-3, as well as IFN induction, although it was less efficient than that of wild-type SeV. These results indicated that, in the case of some murine cell lines, SeV V protein is necessary for the suppression of IRF-3 and NF-κB activation induced by SeV infection. It is likely that both V and C proteins are required for inhibition of IFN induction, as SeV deletion mutants that do not express either the V protein or the nested set of C proteins are unable to block IRF-3 activation or IFN induction in HeLa cells (Komatsu et al., 2004). A possible explanation for the necessity for both SeV C and V proteins in inhibition of IFN production in murine cells begins with the ability of these proteins to block signalling of the RNA helicases MDA-5 and RIG-I. MDA-5 recognizes dsRNA, and RIG-I recognizes 5′-triphosphate RNA in infected cells, which leads to activation of IRF-3 and NF-κB and consequently IFN induction. It has been shown previously that infection with SeV activates both RIG-I and MDA-5 signalling, and that SeV C protein eliminates RIG-I signalling (Strahle et al., 2007), whilst V protein is required to block MDA-5 signalling (Andrejeva et al., 2004; Komatsu et al., 2004). It is possible that HPIV-1 infection in human cells is recognized only by RIG-I, and that expression of the C protein is sufficient for blocking IFN production.

In contrast to the difference observed in the ability to block IFN induction, rSeVhP was able to block IFN-induced nuclear translocation of STAT1 protein in murine cells (Fig. 5). Previously, we reported that both HPIV-1 and SeV C proteins were able to block nuclear translocation of STAT1 and STAT2 in human cells (Bousse et al., 2006). Together, these results suggest that the C protein sequences conserved among SeV and HPIV-1 are responsible for inhibition of the Jak/STAT pathway in both murine and human cells. Overall sequence identity between the SeV and HPIV-1 C proteins is 70 %; however, the C-terminal 131 residues (aa 74–204) are 88.5 % identical. A study using site-directed mutagenesis of SeV C indicated that residues 151, 153 and 154 were important for anti-IFN function (Kato et al., 2004), and these residues are shared with HPIV-1 C protein.

In addition to analysing viral capacity to block IFN induction and signalling, we also determined the ability of the recombinant viruses to overcome the antiviral state induced by pre-treatment with IFN-β. rSeVhP was unable to counteract the antiviral activity induced by IFN-β in murine cells, whilst SeV replicated in the pre-treated cells at a later stage of infection (Fig. 6a). This is similar to our previous findings that, in human cells, HPIV-1 was able to replicate in cells pre-treated with IFN at a later stage of infection, whilst SeV could not (Bousse et al., 2006). As rSeVhP could block the Jak/STAT pathway in murine cells and the addition of exogenous IFN did not suppress SeV or rSeVhPSV replication (Fig. 6b), the inhibition of viral replication could be due to antiviral immunity induced directly by activated NF-κB and IRF-3, without the involvement of the Jak/STAT pathway. A previous study showed that the SeV V protein counteracts antiviral activity induced by IRF-3 without the involvement of IFN (Kiyotani et al., 2007). Consistent with their results, rSeVhPSV replicated better than rSeVhP or rSeVhPSC in IFN-pre-treated cells, further supporting the idea that SeV V protein, possibly together with C protein, is required to antagonize antiviral immunity induced upon viral infection.

In sharp contrast to SeV, infection of murine cells with rSeVhP exhibited significant CPE due to apoptosis in murine MLg2908 cells, but not in human A549 cells (Fig. 7), showing that the P/C gene products of HPIV-1 are unable to elicit a protective effect against apoptosis in infected murine cells. A reduction in apoptosis was observed in rSeVhPSV- but not rSeVhPSC-infected cells, suggesting that the SeV V protein, possibly in conjunction with the C protein, functions to suppress apoptosis in infected murine cells. When we infected murine cells with rSeVhP or rSeVhPSC in the presence of geldanamycin, an inhibitor of IRF-3 and NF-κB, apoptosis was markedly reduced (Fig. 7b). This result supports the idea that the pathway that leads to IRF-3 and/or NF-κB activation is involved in regulation of apoptosis (Heylbroeck et al., 2000). A recent study also showed that induction of IRF-3, but not NF-κB or the Jak/STAT pathway, was required for cell death (Peters et al., 2008). Apoptosis is a significant host defence factor limiting virus replication in infected cells. Expression of both C and V proteins from SeV is likely to contribute to the extended production of progeny viruses from infected murine cells and to pathogenicity in murine hosts.