Background
Human respiratory syncytial virus (RSV) is the most common cause of pediatric viral bronchiolitis and pneumonia in infants and young children worldwide, and also causes severe respiratory infection in immunocompromised adults and the elderly [
1,
2]. Despite its world-wide importance, and several decades of research, there is still no vaccine or specific antiviral therapy for RSV disease [
3]. RSV has a single-stranded negative-sense RNA genome, and belongs to the genus
Pneumovirus of the family Paramyxoviridae [
1]. The RSV genome encodes 11 proteins, including attachment and fusion proteins G and F, nucleocapsid-associated proteins N, P and L, transcription and RNA replication factors M2-1 and M2-2, the matrix M protein, small hydrophobic SH protein, and two non-structural proteins NS1 and NS2. The NS1 and NS2 proteins are dispensable for viral replication
in vitro. However, ablation of either NS protein, or both, significantly attenuates the growth of RSV
in vitro and
in vivo[
4‐
7].
Most viruses encode proteins that inhibit the innate immune response to viral infection and promote virus replication [
8,
9]. NS1 and NS2 of both bovine and human RSV are type I Interferon (IFN α/β) antagonists and target type I IFN induction and signaling [
7,
10‐
13]. Deletion of NS1, more so than NS2, from human recombinant (r) RSV (rRSVΔNS1) attenuates replication and results in an increase in the expression of type I IFN-α/β and type III IFN-λ, compared to wild-type (wt) rRSV [
7]. However, deletion of both NS proteins (rRSVΔNS1/2) results in a greater induction of type I and type III IFN expression and attenuates rRSV to a greater extent than deletion of either single NS protein. Deletion of NS1 and/or NS2 also attenuates rRSV in Vero cells, which do not express type I IFN [
6,
7]. This suggests that NS1 and NS2 have additional functions, independent of the type I IFN response, that affect RSV replication. One such function is the suppression of early apoptosis (<18 h) in RSV-infected cells [
14]. RSV induces both pro- and anti-apoptotic factors in A549 and primary epithelial cells [
15]. The NS proteins, both individually and together, delay apoptosis and promote viral replication via an IFN-independent pathway [
14]. RSV NS1 and NS1/2 deletion mutants enhance maturation of infected human dendritic cells, also suggesting that NS1, and to a lesser extent NS2, suppress DC maturation leading to a weakened immune response to infection [
16].
The mechanisms by which NS1 and NS2 suppress the antiviral response are proving to be complex. RSV is known to degrade STAT2, which is required for the transcription of genes encoding a range of antiviral cellular factors [
17‐
20]. Recently, a mechanism by which NS1 targets STAT2 for ubiquitination and proteasome-mediated degradation has been proposed. Elliot
et al., (2007), have identified consensus elongin C and cullin 2 binding sequences within NS1. They have described the potential of NS1 to bind directly to elongin C and act as an E3 ligase to target STAT2 to the proteasome for degradation.
NS1/2 deletion mutants are being developed as live-attenuated vaccine candidates. Preclinical studies in chimpanzees demonstrated that both ΔNS1 and ΔNS2 deletion viruses were substantially attenuated in the upper and lower respiratory tracts and induced significant resistance to challenge with wild-type virus [
5,
21,
22]. Combination of the NS2 deletion with cold-passaged (cp) and temperature-sensitive (ts) mutations, resulting in the vaccine candidates rA2
cpts 248/404ΔNS2 and rA2
cpts 530/1009ΔNS2, proved to be overattenuating in seronegative children [
3]. Evaluation of a ΔNS1 vaccine candidate is planned. However, virus that lacks NS1 replicates less efficiently
in vitro. It would be advantageous to identify residues in NS1 that are involved in antagonising the IFN response, and if possible, to ablate these activities. To this end we sought to develop and characterize a live rRSV containing modification of one of the putative functional regions of NS1, namely the proposed elongin C binding domain [
19].
This domain (VxxLxxxCxxxK) in NS1 does not conform exactly with the consensus sequence (VxxLxxxCxxx(A/I/L/V) of other elongin C-interacting proteins such as VHL and SOCS 1-3 [
23,
24]. For this reason we chose to modify only the 3 consensus residues (V, L and C) by substitution with alanine (A). Here we demonstrate that this region of NS1 is critical for the survival of the NS1 protein, and that rapid degradation of NS1 as a result of this mutation correlates with viral attenuation in both type I IFN-competent (A549) and -incompetent (Vero) cells.
Materials and methods
Cells and virus stocks
Vero (African green monkey), HEp-2a (human epithelial) and A549 (human type II alveolar epithelial) cells were grown in Opti-MEM (Invitrogen) containing 5% FBS (Sigma-Aldrich). Cells were incubated at 37°C in 5% CO
2. Viral stocks were generated in Vero cells infected at a MOI of 0.1 PFU/cell and harvested 7 days later. The titre of viral stocks was determined by plaque assay. Infected cells were incubated at 37°C for 7 days under 0.8% methyl cellulose (Sigma) in OptiMEM with 2% FBS, and plaques visualized by immunostaining using a goat anti-RSV polyclonal antibody (Virostat) followed by horse-radish peroxidase-coupled anti-goat IgG antibodies and DAB peroxidase substrate (Sigma) as previously described [
25].
Construction of NS1F, NS1F-ELCmut, ΔNS1 cDNA clones
Construction of a cDNA copy of the RSV genome under the control of a T7 promotor has been described previously [
26]. The full-length RSV cDNA used in the present study was D46/6120, which is based on strain A2 and contains a stabilizing 112 nt deletion of the downstream non-coding region of SH as well as several silent mutations in the last codons of the SH ORF [
27,
28]. The full-length cDNA clone was not mutated directly, rather a pGEM 7Z(+) plasmid containing the
Aat II-
Xba I fragment (leader, NS1, NS2 and part of the N genes) was generated and used for site directed mutagenesis (Stratagene Quikchange kit). A FLAG (Sigma) tag and diagnostic
Kpn I site were introduced to the N-terminus of the NS1 ORF directly following the start codon using the oligonucleotide: 5'- GG TTA GAG ATG
GAC TAC AAG GAC GAC GAC GAC AAG GGT ACC GGC AGC AAT TC-3' (FLAG tag in italics,
Kpn I site underlined). The putative elongin C binding domain of NS1, also was mutated by site directed mutagenesis. The 3 consensus amino acids of the NS1 elongin C binding motif VxxLxxxC were mutated to alanine using the oligonucleotide: 5'-G TTT GAC AAT GAT GAA
GCA GCA TTG
GCA AAA ATA ACA
GCC TAT ACT G-3' (alanine underlined).
Aat II-
Xba I mutated fragments, containing either the introduced FLAG tag within NS1, or the FLAG tag and the alanine substitution mutations, were introduced into the full-length antigenome cDNA D46/6120 using standard restriction enzyme digestion and T4 ligase (New England Biolabs) according to the manufacturer's instructions. The mutated cDNA clones were designated NS1F (FLAG tag) and NS1F-ELCmut (FLAG tag and alanine substitutions within the elongin C binding domain), and sequenced to ensure the presence of the mutations. A ΔNS1 cDNA clone, containing deletion of the NS1 open reading frame from D46/6120, was also engineered, by excising the entire NS1 gene, including the gene start and gene end signals, following the introduction of flanking
Pst I sites (Stratagene).
Recombinant RSV (rRSV) recovery
Mutated rRSVs were recovered using a modified transfection protocol described previously [
26,
29]. Briefly, 75% subconfluent BSR T7/5 cells were transfected in 6-well dishes simultaneously with 5 μg of antigenome plasmid (NS1F, NS1F-ELCmut or ΔNS1), 2 μg each of the support plasmids pTM1-N and pTM1-P, and 1 ug each of pTM1-L and pTM1-M2-1. Transfections were performed with Lipofectamine 2000 (Invitrogen) in OptiMEM without serum at 37°C. After 1 day the transfection medium was replaced with Glasgow's minimal essential medium (supplemented with 5% FBS and 1% L-glutamine) and the cells were expanded. After a further 4 days of incubation at 37°C, cell-medium mixtures were passaged onto fresh HEp-2a cells. Once cytopathic effects indicative of RSV infection were observed 5-7 days later, cell-medium mixtures were harvested and used to infect HEp-2a monolayers for three rounds of plaque purification. A stock of each recovered virus was then generated in Vero cells. The presence of the expected mutations and the absence of other spurious mutations were confirmed by RT-PCR and nucleotide sequencing of viral genomic RNA for all rRSVs generated.
Growth of rRSV in vitro
Quadruplicate cell culture monolayers of A549 and Vero cells were infected with either wt RSV (D46/6120), NS1F, NS1F-ELCmut, or ΔNS1 at a MOI of 0.01 PFU/cell. Cell culture supernatants were collected daily for 5 days and the virus titres determined by plaque assay. Statistically significant differences in growth between rRSVs was detected by ANOVA
Western Blot analysis
Triplicate cell monolayers were infected with either wt RSV, NS1F or NS1F-ELCmut at a MOI of 1 PFU/cell, or mock infected with media, and incubated at 37°C. At 24 h and 48 h p.i., cells were washed in PBS and lysed with RIPA buffer (25 mM Tris-HCl pH 7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS) containing protease inhibitors (Pierce). 20 μg total protein (BCA assay; Pierce) from each lysate was separated through pre-cast 12% polyacrylamide Bis-Tris gels (Invitrogen). Proteins were then transferred to nitrocellulose membranes and blocked in either 5% skim milk (Diploma)/PBS or 1% BSA (Sigma)/PBS containing 0.1% Tween 20. The blots were analysed using a goat anti-RSV antiserum (Virostat), or a rabbit antiserum raised against the C-terminal 13 amino acids of NS2, which reacts with both NS1 and NS2, and FLAG was detected using a mouse anti-FLAG IgG (Sigma). As a marker for autophagy, LC3 was detected using an antibody against LC3B (Cell Signaling Technology). β-actin was used as a loading standard and detected using a mouse anti-β-actin IgG (Sigma). Bound antibodies were visualized with species-specific IgG conjugated to either IR800 (Rockland Inc), or HyLyte680 (AnaSpec) and a Li-Cor Odyssey scanner. RSV protein expression was quantified using Odyssey™ densitometry software and statistical differences identified using Students t-test.
RNA extraction and PCR
Cells were infected at a MOI of 1 with either NS1F, NS1F-ELCmut, ΔNS1, or mock infected with media, and harvested 24 h p.i., Total RNA was isolated using TRIzol (Invitrogen) and the RNeasy total RNA isolation kit (QIAGEN). RT-PCR was performed using RSV-specific primers and the PCR products sequenced also using RSV-specific primers. For quantitiative (q) PCR, RT was performed with oligo (dT)
20 primers (Invitrogen) to select mRNA from the total RNA preparations. qPCR was performed using dual-labeled probe and primer sets as previously described [
7] to quantify β-actin and IFN-β. PCR primers and dual-labeled probes to quantify NS1 and N mRNA were designed using Primer 3
http://frodo.wi.mit.edu/primer3/ software. qPCR was performed using QuantiTect™ reagents (QIAGEN) and the Rotogene 3000 (Roche). The fold increase in target expression compared to mock-infected was calculated using the 2
-ΔΔCt formula.
Immunofluorescence
Confluent monolayers of Vero cells were cultured on coverslips and infected with either NS1F, NS1F-ELCmut or ΔNS1 at a MOI of 1 PFU/cell, or mock infected with media. At specific times, cells were fixed with 4% formaldehyde (ProSciTech) for 30 min and permeabilized with 0.1% Triton X-100 for 10 min. Primary and secondary antibodies were diluted in 1% BSA/PBS. NS1 and other RSV proteins were detected using antiserum also used for western blot anlaysis (above). Cellular organelles were detected using monoclonal antibodies to EEA, GM130, LAMP1 (BD Transduction Labs), and PDI (Invitrogen). Secondary antibodies, anti-mouse and anti-rabbit specific IgG were conjugated to Alexa Fluor 488 or 594 (Invitrogen). Following antibody incubation and washes, coverslips were mounted on glass microscope slides using ProLong Gold mounting medium (Invitrogen) and visualized using a Nikon Eclipse E3600 fluorescence microscope. Images were collated using adobe Photoshop software. Statistical differences in protein detection were identified using Student's t test.
Discussion
It is well documented that both of the non-structural proteins of RSV, NS1 and NS2, are type I IFN antagonists and target both the induction and signaling pathways. RSV mediates type I IFN signaling by degrading STAT2. Both NS1 and NS2 have been documented to play a role in this response [
17‐
20]. Elliot
et al. (2007) demonstrated that NS1 contains a consensus elongin C binding domain, and can form an elongin C-cullin-SOCS box-type E3 ubiquitin ligase. This may provide a mechanism by which RSV targets STAT2 for proteosomal degradation. The putative elongin C binding site within NS1, as described by Elliot
et al, (2007), is VxxLxxxCxxxK, which differs in the fourth conserved residue from the consensus motif of VxxLxxxCxxxA/I/L/V [
23,
24]. Due to the variability of the final amino acid, we chose to replace the 3 residues with full homology, V, L and C, with alanines (A), as they are likely to be the most critical for function.
The resulting live rRSV NS1F-ELCmut was attenuated in both type I IFN competent (A549) and -incompetent (Vero) cells. This attenuation was similar to that observed when NS1 was ablated entirely (ΔNS1) in A549 cells, suggesting that mutation of the elongin C binding domain affected viral growth nearly as much as complete deletion of NS1. In fact, in A549 cells, mutation of the elongin C binding region affected NS1 expression such that it was never detected by western blot analysis or immunofluorescence. This suggests that the attenuation of NS1F-ELCmut and ΔNS1' were similarly due to the significant reduction or absence of NS1 protein expression respectively. Both NS1F-ELCmut and ΔNS1 also induced similarly increased levels of IFN-β mRNA, which correlated to viral attenuation and the absence or significant reduction of NS1 protein. The effect of deletion of NS1 on the replication of live rRSV is well documented [
6,
7,
21,
22], as is the correlating increase in the type I IFN response [
7,
17,
18].
More insight into the effect of mutations within the elongin C binding domain of NS1 was gained using Vero cells. In these cells, the mutations in the elongin C binding motif of NS1 did lead to NS1 degradation, but the effect was delayed. Thus, NS1 was detected by immunofluorescence up to 48 h p.i. The presence of detectable NS1 early in infection and the slower rate of NS1 degradation in Vero cells, correlated to moderate attenuation of NS1F-ELCmut to levels between those of NS1F and ΔNS1. It has been observed previously that deletion of the NS proteins results in attenuation of RSV in type I IFN-incompetent Vero cells [
6,
7].
In Vero cells infected with NS1F-ELCmut, NS1 was degraded via a lysosomal-directed pathway. From 15 h pi to 48 h pi, there was a shift in the location of NS1 from being both nuclear and diffusely cytoplasmic, to being completely located within LAMP1-positive cytoplasmic bodies. The co-localisation of NS1 and LAMP1 suggested that NS1 was associated with lysosomes/autolysosomes and that degradation of NS1 may occur via autophagy. Autophagy involves the sequestration of cytoplasmic cellular material into double-membrane bound autophagosomes, which then fuse with lysosomes to form autolysosomes for degradation of intracellular materials [
31]. Interactions between autophagy and viral infection have been documented for many viruses [
31]. However, little is documented concerning autophagy during RSV infection. RSV-induced autophagy in dendritic cells has been shown to promote antiviral cytokine responses [
32]. During autophagy LC3 is activated and recruited to the surface of autophagosomes. Activation of LC3, as demonstrated here in NS1F-infected cells, shows that RSV does induce autophagy in epithelial cells. The accumulation of LAMP1-positive bodies in the NS1F-ELCmut-infected cells suggested that autophagy may have been induced. However, the lack of LC3 activation by NS1F-ELCmut by 48 h p.i. indicates that the association of NS1 with lysosomes is not a consequence of increased autophagy. The lack of activated LC3 in NS1F-ELCmut-infected cells compared to NS1F-infected cells is most likely correlated with the reduced replication of the NS1F-ELCmut virus.
It is possible that the mutations introduced into the elongin C binding motif of NS1 caused misfolding of the protein leading to degradation. The tertiary structure is not known and therefore we were not able to model these mutations to assist with protein engineering. There is one report of a method for purifying rNS1-HIS
6 from
E. coli and some information concerning secondary structure of NS1 [
33]. However, functional regions have not yet been mapped and the consequences of changes in secondary structure on NS1 function are not known. Protein structure prediction using Protein Predict [
34] indicated that this motif may be a beta strand (weak prediction), and that the alanine substitutions induce a predicted change to an alpha helix (results not shown). Given the increased accuracy of secondary predictions [
35] it is reasonable to suspect that the introduced mutations indeed changed the local character of the secondary structure. However, whether this resulted in global misfolding leading to degradation is unknown.
Proteasomes and lysosomes represent the main proteolytic pathways in mammalian cells. Misfolded proteins are usually targeted for degradation via proteosomes [
36], as the proteosomal pathway requires substrates to be unfolded to enter the narrow catalytic core [
37]. Lysosomes, in contrast, fuse to vesicles containing proteins for degradation that are not necessarily misfolded. Thus, the finding that degradation of NS1 occurred via a lysosome-driven pathway rather than a proteosomal pathway in Vero cells suggests that the mutations in NS1 may not have resulted in global unfolding. Instead, it may be that the mutated NS1 protein became targeted for degradation as a consequence of its failure to associate with the E3 ligase complex. Although we attempted to investigate binding of NS1 to elongin C, by using the FLAG tag to immunoprecipitate NS1-associated proteins, the degradation of NS1 made this difficult. Hence we cannot directly confirm or deny binding of NS1 to elongin C.
The initial purpose of engineering mutations within the elongin C biding domain of NS1 was to investigate if NS1 did form an E3 ligase complex to degrade STAT2 in a live virus infection, and to alter this function, such that STAT2 degradation was reduced during RSV infection. Due to the rapid degradation of NS1 by these mutations we were not able to demonstrate the effect of these mutants on STAT2 within whole cell populations. We did perform infection experiments in A549 cells and identified isolated infected cells by immunofluorescence 24 h after infection. We found that STAT2 was degraded in NS1F-infected cells, as expected, and not degraded in NS1F-ELCmut- and ΔNS1-infected cells (additional file
1). This, however, was most likely the result of a reduction or lack of NS1 expression, not the result of mutations within NS1.
There has been considerable effort over the last 30 years to develop a vaccine for human RSV and currently live attenuated RSVs may offer the best candidates [
3]. As deletion of NS1 or NS2 is attenuating and results in elevated type I IFN-mediated antiviral responses each deletion is being considered for inclusion in live-attenuated vaccine candidates. Modification or deletion of specific functional regions that inhibit the antiviral response may provide improved mutants in which the interferon antagonist activities have been ablated while providing improved growth. Alanine replacement of the conserved residues within the elongin C binding domain of NS1, however, resulted in attenuation similar to full deletion of NS1 in A549 cells. It is therefore unlikely that this present set of mutations will be useful in vaccine development. The data presented here do suggest, however, that this region is critical to the survival of NS1 and may be a potential target for antiviral agents. NS1 has proven to be a fruitful target for anti-RSV therapy. RNAi using nanoparticle delivery of siRNA targeting NS1 was used successfully to protect against infection and to clear an existing RSV infection in Fischer 344 rats [
38]. A specific antiviral molecule that disabled elongin C binding and degraded NS1 may prove effective for clearing RSV infection.
Authors' contributions
KMS, JJG and PLC conceived of the study. KMS, CPS and WHL made the rRSVs and performed most experiments. FMP performed to qPCR experiments. MJH and JJG advised the mutations and MJH performed the protein prediction. KMS and CPS analysed the data and wrote the manuscript with assistance from JJG, MJH and PLC. All authors read and approved the final manuscript
Competing interests
The authors declare that they have no competing interests.