|Home | About | Journals | Submit | Contact Us | Français|
Studies involving limited numbers of rotavirus (RV) strains have shown that the viral gene 5 product, NSP1, can antagonize beta interferon (IFN-β) expression by inducing the degradation of IFN-regulatory factors (IRFs) (IRF3, IRF5, and IRF7) or a component of the E3 ubiquitin ligase complex responsible for activating NF-κB (β-transducin repeat-containing protein [β-TrCP]). To gain a broader perspective of NSP1 activities, we examined various RV strains for the ability to inhibit IFN-β expression in human cells. We found that all strains encoding wild-type NSP1 impeded IFN-β expression but not always through IRF3 degradation. To identify other degradation targets involved in suppressing IFN-β expression, we used transient expression vectors to test the abilities of a diverse collection of NSP1 proteins to target IRF3, IRF5, IRF7, and β-TrCP for degradation. The results indicated that human RVs rely predominantly on the NSP1-induced degradation of IRF5 and IRF7 to suppress IFN signaling, whereas NSP1 proteins of animal RVs tended to target IRF3, IRF5, and IRF7, allowing the animal viruses a broader attack on the IFN-β signaling pathway. The results also suggested that the NSP1-induced degradation of β-TrCP is an uncommon mechanism of subverting IFN-β signaling but is one that can be shared with NSP1 proteins that induce IRF degradation. Our analysis reveals that the activities of NSP1 proteins are diverse, with no obvious correlations between degradations of pairs of target proteins. Thus, RVs have evolved functionally distinct approaches for subverting the host antiviral response, a property consistent with the immense sequence variation noted for NSP1 proteins.
The production of type I interferons (IFNs), including IFN-α and IFN-β, is critical to an effective innate immune response to viral infection. Unless prevented by the activity of an antagonist, infection leads to IFN-β expression, which in turn induces the expression of IFN-α (22). RNA viruses trigger the IFN response through interactions of viral components with cellular transmembrane Toll-like receptors (TLRs) or cytoplasmic retinoic acid-inducible gene I (RIG-I)-like receptors (RLRs). The stimulation of these receptors leads to the activation of numerous transcription factors, including those belonging to the IFN-regulatory factor (IRF) and nuclear factor κB (NF-κB) families (17, 20).
Two IRF proteins, IRF3 and IRF7, have important roles in the expression of type I IFN (3, 22). IRF3 is constitutively expressed in cells, where it accumulates at elevated levels in the cytoplasm. In contrast, IRF7 is present in most cells at low levels, and its expression is amplified by type I IFN (17, 38). Both IRF3 and IRF7 reside in inactive forms in the cytoplasm and undergo activation through the phosphorylation of C-terminal serine residues. These active forms of IRF3 and IRF7 then undergo dimerization and translocation to the nucleus, where they interact with specific promoters to enhance the transcription of IFN-stimulated genes (ISGs) (33, 37). IRF3, working in cooperation with other transcription factors, promotes the expression of IFN-β, a few IFN-α subtypes, and IRF7. IRF7 has broader effects than IRF3, promoting the expression of even higher levels of IFN-β, a greater number of IFN-α subtypes, and additional IRF7. A maximal level of expression of type I IFN is achieved through the combined actions of IRF3 and IRF7 (12, 21, 30, 31). Other members of the IRF family, such as IRF5, are known to support the development of antiviral responses. However, their contribution to effective IFN expression is not well understood and may vary depending on cell type and host species (4, 26).
The expression of IFN-β requires the coordinated activation and assembly of IRF3/IRF7, along with ATF2/c-Jun and NF-κB, into an enhanceosome complex on the IFN-β promoter. NF-κB is typically held as part of an inactive cytoplasmic complex through interactions with proteins of the inhibitors of NF-κB (IκB) family. NF-κB activation occurs when signals stimulate the phosphorylation of IκB, which is subsequently recognized by the β-transducin repeat-containing protein (β-TrCP) component of the Skp1-Cul1-F-box E3 ubiquitin ligase complex (SCFβ-TrCP). The recognition of IκB by SCFβ-TrCP leads to its ubiquitination and degradation. Consequently, NF-κB is free to move to the nucleus, where, in the presence of IRF3/IRF7, the factor induces the expression of IFN-β (7, 16).
Type I IFNs activate the expression of hundreds of ISGs, which play critical roles not only in innate immune responses but also in adaptive immune responses, apoptosis, and cell growth regulation (9, 35). Collectively, the ISGs have the potential to target various stages of the virus replicative cycle; thus, the unrestricted expression of IFNs in infected cells can have profoundly negative consequences for virus viability. To subvert the antiviral effects of IFN, many viruses have evolved mechanisms for antagonizing the activities of the IRF proteins or NF-κB (28, 29, 35).
Rotavirus (RV), the primary cause of severe dehydrating gastroenteritis in infants and young children, is sensitive to the antiviral effects of IFN (8, 19, 24). The RV gene 5 product, NSP1, antagonizes the innate immune response by inducing the degradation of one or more components required for the expression of type I IFN. Studies of the NSP1 proteins of a limited number of RV strains have shown that the degradation targets can include IRFs or β-TrCP. For example, NSP1 of simian SA11-4F RV has been shown to induce the degradation of IRF3, IRF5, and IRF7, while NSP1 of porcine OSU induces the degradation of β-TrCP (5, 6, 13-15). However, the NSP1 activities of many RV strains, including those relevant to human disease, have not been analyzed.
NSP1 exhibits the greatest sequence variability of any of the RV proteins, with nucleotide identities falling below 40% between some mammalian strains. Sequence variability is particularly evident among virus strains that infect different animal species and much less so for virus strains that share common natural hosts. This pattern suggests that the NSP1 proteins of RV strains infecting different animal species may have genetically diverged as the proteins evolved to better counter the antiviral pathways inherent to each host (10).
The poor sequence conservation of NSP1, especially in the C-terminal half of the protein that includes the IRF-interactive domain (15), may affect its capacity to recognize and induce the degradation of any one IRF. Because the activities of so few NSP1 proteins have been defined, our overall understanding of what constitutes typical IFN antagonist activities for the protein is quite limited. To gain greater insight into the activities of RV NSP1, we sought to examine the ability of a variety of strains to inhibit the expression of IFN-β and how it correlates with the NSP1-induced degradation of the target proteins IRF3, IRF5, IRF7, and β-TrCP. Our analysis revealed that all RV strains encoding wild-type NSP1 suppressed IFN-β expression although not exclusively by IRF3 degradation. Moreover, we found that the most common targets of NSP1-mediated degradation are IRF5 and IRF7, indicating that these may be the preferred target of the protein. In fact, several human RVs that, although capable of targeting IRF5 and IRF7, showed little effect on IRF3 were identified.
African green monkey kidney MA104 cells were maintained in medium 199 (Invitrogen) supplemented with 5% fetal bovine serum (FBS) (Invitrogen). Human colon HT29 cells (ATCC) were maintained in Dulbecco's modified Eagle's medium (DMEM) (Quality Biological, Inc.) supplemented with 10% FBS. Human 293T cells were maintained in DMEM supplemented with 10% FBS and 1% nonessential amino acids (Lonza). All viruses were propagated and titers were determined with MA104 cells. Viruses were activated by incubation with 5 to 10 μg trypsin per ml for 30 to 60 min prior to infection.
NSP1 polyclonal antiserum was produced by Pacific Immunology Corporation (Ramona, CA). Peptides corresponding to amino acids 257 to 273 of simian SA11-5S NSP1 (C-RDELELYSDLKNDKNKL), amino acids 479 to 494 of simian RRV NSP1 (C-LSEEYELLISDSEDDD), amino acids 477 to 491 of bovine UK NSP1 (C-NELIDEYDLELSDVE), and amino acids 479 to 491 of lapine 30-96 NSP1 (C-EEYELLISDSEDD) were conjugated to the carrier protein keyhole limpet hemocyanin. Each peptide was then used to immunize individual New Zealand White rabbits. NSP1-specific antisera were collected and affinity purified by using the immunizing peptide. The affinity-purified antibodies tested negative for cross-reactivity with other RV proteins. Each NSP1 antibody was used at a 1:1,000 dilution. Polyclonal VP6 antiserum was produced in guinea pigs by immunization with VP6 stripped from DxRRV double-layered particles by treatment with 1 M CaCl2 (25). The VP6 antibody was used at a 1:2,500 dilution. Rabbit polyclonal antibodies to IRF3, IRF7, and PCNA (used at 1:2,000 dilutions) and mouse monoclonal antibody to IRF5 (1:500 dilution) were purchased from Santa Cruz Biotechnology. Monoclonal anti-FLAG M2 antibody (1:1,000 dilution) was purchased from Sigma-Aldrich.
The NSP1 open reading frames (ORFs) for RV strains AU-1 (GenBank accession number D45244), DS-1 (accession number EF672578), WI61 (accession number EF672620), Gottfried (accession number U08431), and 30-96 (accession number DQ205225) were synthesized de novo and inserted into the entry vector pENTR221 by Blue Heron Biotechnology (Bothell, WA). The NSP1 ORF cDNAs were then transferred into the destination vector pcDNA-DEST40 (Invitrogen) by recombination with LR Clonase II (Invitrogen).
To prepare vectors containing the NSP1 ORFs of RV strains RRV (GenBank accession number AY117048), UK (accession number HQ186289), Wa (accession number L18943), K9 (accession number EU708929), KU (accession number AB022769), NCDV (accession number HQ186290), and OSU (accession number D38153), viral RNA was extracted from infected cell lysates by using TRIzol LS reagent (Invitrogen). Gene 5 cDNAs were prepared from the RNA by using a SuperScript II Platinum Taq HiFi one-step reverse transcription (RT)-PCR kit (Invitrogen) and forward and reverse primers that annealed to the untranslated regions of the RNA. The PCR product was ligated into the TOPO TA vector pCR2.1 (Invitrogen). The NSP1 ORFs in these vectors were then amplified by using forward and reverse primers that contained unique restriction sites (Table (Table1).1). The RRV, UK, Wa, K9, KU, and NCDV PCR products were digested with applicable restriction enzymes and inserted into the entry vector pENTR-1A (Invitrogen). The inserts were then transferred into pcDNA-DEST40 by recombination with LR Clonase II. The OSU PCR product was digested with the applicable restriction enzymes and inserted into the expression vector pCI (Promega).
A vector containing the NSP1 ORF of RV strain ETD (GenBank accession number GQ479951) was prepared by the PCR amplification of a plasmid that contained an ETD gene 5 cDNA (kindly provided by Harry Greenberg) (Table (Table1).1). The PCR product was digested with the applicable restriction enzymes and inserted into the entry vector pENTR-1A (Invitrogen). Afterwards, the insert was transferred into pcDNA-DEST40 by recombination with LR Clonase II.
Vectors containing the NSP1 ORFs of RV strains SA11-4F (GenBank accession number AF290881) and SA11-5S were prepared by the PCR amplification of a plasmid containing a SA11-4F gene 5 cDNA (kindly provided by Zenobia Taraporewala) with the primer pairs indicated in Table Table1.1. The reverse primer used for the SA11-5S NSP1 vector contained a stop codon that introduced a 17-amino-acid truncation into the protein product of the ORF. Otherwise, the amino acid sequences of the protein products of the SA11-4F and SA11-5S ORFs were identical. The PCR products were digested with the applicable restriction enzymes and inserted into the pCI vector.
To construct the plasmid expressing FLAG-tagged β-TrCP, the sequence encoding β-TrCP (GenBank accession number NM_033637) was amplified from pCMV6-XL5-β-TrCP (OriGene) by using forward and reverse primers containing unique restriction sites (Table (Table1).1). The PCR product was digested with applicable restriction enzymes and inserted into the pCI vector, producing pCI-β-TrCP. The FLAG tag was added by annealing forward and reverse primers containing the epitope flanked by restriction sites (Table (Table1).1). The annealed primers were ligated into the NheI and XhoI sites of pCI-β-TrCP, generating pCI-FLAG-β-TrCP. Plasmid pCMVSport-IRF7H expresses human IRF7 (accession number AF076494) (2). Plasmid pCMV6-IRF5, which expresses human IRF5 (accession number NM_032643), was purchased from OriGene. Restriction enzymes were purchased from New England BioLabs.
Unless indicated otherwise, transfection conditions were as follows. Approximately 2.0 × 106 293T cells in 6-well plates were transfected with vector DNA using 8 μl of Lipofectamine 2000 (Invitrogen) per well. To analyze IRF protein levels, 4 μg of the NSP1 expression vector was either transfected alone (to analyze IRF3) or cotransfected with 0.25 μg of plasmid expressing IRF5 or IRF7. To analyze β-TrCP protein levels, 2 μg of the NSP1 expression vector was cotransfected with 2 μg of plasmid expressing FLAG-β-TrCP. At 24 h posttransfection (p.t.), cells were washed with phosphate-buffered saline and then lysed in 200 μl radioimmunoprecipitation assay (RIPA) buffer (150 mM NaCl, 150 mM Tris-HCl [pH 8.0], 1.0% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate [SDS], 1× Complete protease inhibitor cocktail [Roche]) and assayed for protein content by immunoblot analysis. For experiments examining the effect of the proteasome inhibitor MG132 (Boston Biochem), medium was removed at 8 h p.t. and replaced with fresh medium containing 20 μM MG132. Cells were incubated for an additional 16 h and harvested at 24 h p.t. by lysis in 200 μl RIPA buffer. Approximately 1.5 × 106 HT29 cells in 6-well plates were infected at a multiplicity of infection (MOI) of 5 with plaque-determined titers of virus. At 10 h postinfection (p.i.), the cells were lysed in 200 μl RIPA buffer and assayed for protein content by immunoblot analysis.
Whole-cell lysates were briefly sonicated and then diluted 1:1 in 2× Tris-glycine SDS sample buffer (Invitrogen). Proteins were resolved by electrophoresis in 10% Tris-glycine gels (Invitrogen) and transferred onto nitrocellulose membranes. The membranes were blocked with Odyssey blocking buffer (Li-Cor) and then incubated with primary antibody in Odyssey blocking buffer containing 0.1% Tween 20. Following primary incubation, blots were washed with Tris-buffered saline (50 mM Tris-HCl [pH 7.5], 150 mM NaCl) containing 0.1% Tween 20 (TBS-T) and incubated with IRDye680- or IRDye800-conjugated secondary antibodies in TBS-T and 1% milk. Following incubation with the secondary antibody, blots were washed with TBS-T. Blots were imaged and quantified with the Odyssey infrared imaging system (Li-Cor). Bands of interest were quantified and normalized to the PCNA levels to control for loading. One-way analysis of variance (ANOVA) was performed on endogenous IRF3 immunoblots using GraphPad Prism 5 software.
At 10 h p.i. total RNA was harvested from infected cells by using the RNeasy minikit (Qiagen, Inc.), and contaminating genomic DNA was removed by using RNase-free DNase I (Qiagen, Inc.). RNA (500 ng) was reverse transcribed by using a high-capacity cDNA reverse transcription kit (Applied Biosystems). cDNA was amplified by quantitative TaqMan PCR for the quantification of IFN-β mRNA (Applied Biosystems). Samples were run on a 7900HT Fast real-time PCR system (Applied Biosystems). Quantitative PCR results were analyzed by using the comparative threshold cycle (ΔΔCT) method, normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA. Results were expressed as a fold difference relative to mock-infected cells.
Phylogenetic trees were generated with MacVector (10.6.0) from NSP1 amino acid alignments using the neighbor-joining method, a systematic best tree with gaps distributed proportionally, and the Poisson correction parameter. Bootstrap values were based on 2,000 repetitions.
We initiated our experiments by comparing the abilities of a diverse collection of RVs to inhibit IFN-β gene activation. This was accomplished by infecting HT29 cells with several common laboratory strains of RV and a panel of monoreassortant viruses containing gene 5 RNAs derived from different RVs on the genetic background of the SA11-L2 virus (1). The monoreassortants allowed us to examine the properties of NSP1 proteins from RVs of human origin, which do not replicate well in HT29 cells. Total RNA recovered from the infected cells at 10 h p.i. was analyzed for levels of IFN-β mRNA by quantitative RT-PCR. The results showed that the SA11-5S strain of RV, which encodes a defective NSP1 protein and therefore is unable to inhibit the IFN response (5), induced high levels of IFN-β mRNA (Fig. (Fig.1).1). In contrast, the other strains of RV examined did not trigger similarly high levels of IFN-β mRNA synthesis but instead induced levels of IFN-β mRNA only slightly above those of the mock-infected control. Notably, viruses encoding full-length NSP1 proteins of strains SA11, RRV (SRF), NCDV, and K9 (SNF) effectively suppressed IFN-β mRNA production by more than 100-fold compared to the level induced by SA11-5S. RV strains encoding OSU (SOF), KU (SKF), and DS-1 (SDF) full-length NSP1 proteins were slightly less efficient at suppressing the production of IFN-β mRNAs (20- to 70-fold compared to SA11-5S) but were nonetheless strong inhibitors of IFN-β mRNA expression. Thus, the ability to antagonize IFN-β gene activation was a shared feature of RV strains that grew well in HT29 cells and that encoded full-length NSP1 proteins.
Previous studies have shown that some RV NSP1 proteins inhibit IFN-β expression by inducing the degradation of IRF3 (5, 13, 32). To examine whether the inhibition of IFN-β expression always correlates with IRF3 degradation, HT29 cells were infected with the same collection of viruses used in the IFN-β expression assays described above (Fig. (Fig.1),1), and total cellular protein was harvested at 10 h p.i. Quantitative immunoblot analysis with IRF3 antiserum showed that infection with SA11-4F, RRV, NCDV, SA11-L2, and reassortants encoding K9 (SNF) or RRV (SRF) NSP1 proteins reduced endogenous IRF3 levels to less than 10% of that of mock-infected cells, indicating that these strains were all strong inducers of IRF3 degradation (Fig. (Fig.2).2). The degradation of IRF3 by these RV strains correlated with the inhibition of IFN-β transcription (Fig. (Fig.1).1). Infection with SA11-5S did not result in IRF3 degradation, consistent with the ability of this virus to stimulate high levels of IFN-β expression.
In contrast, cells infected with OSU or with reassortants encoding DS-1 (SDF), KU (SKF), or OSU (SOF) NSP1 proteins had IRF3 levels similar to those of mock-infected cells. Since each of these viruses was capable of inhibiting IFN-β expression (Fig. (Fig.1),1), this finding indicates that some strains of RV inhibit IFN-β via a mechanism other than IRF3 degradation. Furthermore, the NSP1 proteins of human RV prototype strains KU and DS-1 failed to induce IRF3 degradation under the same conditions and in the same genetic background in which NSP1 proteins of several other strains (e.g., SA11, RRV, and K9) successfully triggered IRF3 degradation (Fig. (Fig.2).2). The analysis using SA11 monoreassortants definitively showed that the ability of RV strains to induce IRF3 degradation genetically maps to gene 5 RNA, similar to data from previous studies carried out with a panel of RRV × UK reassortant viruses (11).
To examine whether the activities of some RV NSP1 proteins went undetected because the protein failed to accumulate efficiently in infected cells, due to either poor expression or a rapid rate of turnover, we produced antisera by immunizing rabbits with peptides representing portions of the SA11-5S, RRV, UK, and 30-96 NSP1 proteins. Because of the inherent sequence variation among NSP1 proteins, each antiserum reacted with the NSP1 proteins of only a limited number of RV strains. Nonetheless, immunoblot analyses performed with the NSP1 antisera showed that NSP1 was present in cells infected with each of the isolates tested irrespective of IRF3 degradation (Fig. (Fig.2).2). This finding established that the failure of the OSU, SOF, SDF, and SKF viruses to cause IRF3 degradation was not due to a lack of NSP1 expression. Notably, the quantities of NSP1 detected in SA11-4F- and SA11-5S-infected cells were similar, suggesting that the truncated NSP1 species of SA11-5S was no less stable than the wild-type NSP1 species of SA11-4F. Immunoblot analysis of the RV protein VP6 showed similar expression levels for all viruses, suggesting similar levels of infection in HT29 cells (Fig. (Fig.2).2). The viral yield at 10 h p.i. ranged from 4 × 106 to 1 × 108 PFU per ml, indicating that all strains tested replicated productively.
Having identified several RV strains that did not target IRF3 for degradation in infected cells despite the ability to suppress IFN-β expression, we sought to gain a more comprehensive understanding of the activities of NSP1 proteins. To accomplish this, we created a set of 15 gene 5 cDNAs that represented seven phylogenetically distinct genotypes of the gene 5 RNA, that were derived from viruses that infect different animal species, and that represented prototype strains of human genogroups 1, 2, and 3. In addition, the set of gene 5 cDNAs included those originating from viruses that have been used commonly in studies of rotavirus biology and that have been used for vaccine development (23). The cDNAs were introduced into expression vectors under the control of a cytomegalovirus (CMV) promoter. The origins of the gene 5 cDNAs (and their genotypes) are as follows: murine ETD (A7); simian SA11-4F (A5), SA11-5S (A5), and RRV (A9); canine K9 (A9); lapine 30-96 (A9); human Wa (A1), DS-1 (A2), AU-1 (A3), and WI61 (A1); bovine UK (A3) and NCDV (A3); and porcine OSU (A1) and Gottfried (A8).
To assess the effect of each NSP1 protein on human IRF3, 293T cells were transfected with individual NSP1 expression vectors, and at 24 h p.t., the levels of endogenous IRF3 were analyzed by a quantitative immunoblot assay. Consistent with IRF3 results for virus-infected cells (Fig. (Fig.2),2), the transient expressions of the SA11-5S, OSU, KU, and DS-1 NSP1 proteins were found not to trigger a statistically significant decrease (95% confidence) in IRF3 levels (Fig. (Fig.3).3). Likewise, the AU-1 and Wa NSP1 proteins failed to degrade IRF3. In contrast, the NSP1 proteins of virus strains ETD, SA11-4F, K9, 30-96, RRV, UK, NCDV, WI61, and Gottfried significantly reduced IRF3 levels relative to control (pCI)-transfected cells. Thus, IRF3 appears to be frequently targeted for degradation by the NSP1 proteins of animal RVs but less so by the NSP1 proteins of human RVs.
To analyze the effect of NSP1 proteins on IRF7, 293T cells were cotransfected with vectors encoding human IRF7 and an NSP1 protein, and at 24 h p.t., IRF7 levels were assessed by a quantitative immunoblot assay. The results indicated that the NSP1 proteins of virus strains ETD, SA11-4F, 30-96, RRV, UK, NCDV, WI61, KU, Wa, Gottfried, and DS-1 reduced IRF7 levels by 50% or more (Fig. (Fig.4A).4A). Notably, the NSP1 proteins of all the human viruses tested in our analysis targeted IRF7 for degradation, although some failed to target IRF3 (Fig. (Fig.22 and and3).3). Thus, the degradation of IRF7, and not IRF3, may be a principal mechanism by which human RVs antagonize IFN-β signaling. The only wild-type NSP1 proteins with no activity on the IRF7 target in our assay were those of virus strains K9, AU-1, and OSU. The AU-1 and OSU NSP1 proteins also showed little or no activity on the IRF3 target.
To determine if the degradation results were affected by changes in the time between cell transfection and harvest, we cotransfected 293T cells with NSP1 and IRF7 expression vectors (as described above) and harvested cells at 12, 24, and 36 h p.t. Analysis of cell lysates from 24 and 36 h p.t. by a quantitative immunoblot assay revealed little difference in IRF7 levels (Fig. (Fig.4B).4B). In contrast, IRF7 levels in cell lysates containing some types of NSP1 proteins (30-96, WI61, KU, Wa, and DS-1) were markedly higher at 12 h p.t. than at 24 and 36 h p.t. This change from early to later times suggests that in these transfected cells, there may be an initial delay in the interaction of some NSP1 and IRF7 proteins or in the degradation of the IRF target. Regardless, the data suggest that sample preparation at 24 h p.t. suitably reflects the activity of NSP1 proteins on IRF targets, whereas sample preparation at earlier times may give misleading results.
In our assays, we noted that the IRF7 accumulation in 293T cells cotransfected with OSU NSP1 and IRF7 expression vectors was 2- to 3-fold higher than that of control (pCI)-transfected cells (Fig. (Fig.4A).4A). This phenomenon was not seen in experiments examining the effects of transiently expressed NSP1 proteins on endogenous levels of IRF3 (Fig. (Fig.3),3), suggesting that the enhanced IRF7 level might result from the effect of OSU NSP1 on the activity of the IRF7 transcription vector, the export of the IRF7 mRNA to the cytoplasm, or the translational efficiency of the IRF7 mRNA. By quantitative RT-PCR experiments, we were able to rule out the possibility that the enhanced IRF7 level in 293T cells expressing OSU NSP1 stemmed from increased levels of IRF7 mRNA accumulation (data not shown). Importantly, we observed enhanced levels of IRF7 accumulation even when the 293T cells were transfected with various amounts of OSU NSP1 and IRF7 vectors (Fig. 3C and D). Similarly high levels of IRF7 accumulation were not obtained when 293T cells were cotransfected with an expression vector encoding SA11-4F NSP1 instead of OSU NSP1. Together, these data suggest that the enhanced accumulation of IRF7 does not stem from a technical artifact but rather from some specific but undefined effect that OSU NSP1 has on the posttranscriptional events of the IRF7 mRNA.
To examine the effect of NSP1 on IRF5, 293T cells were cotransfected with vectors encoding human IRF5 and an NSP1 protein, and IRF5 levels were analyzed at 24 h p.t. by a quantitative immunoblot assay. The results indicated that the NSP1 proteins of viral strains ETD, SA11-4F, 30-96, RRV, WI61, Gottfried, and DS-1 were the most effective in inducing the degradation of IRF5, while the NSP1 proteins of KU and Wa displayed more moderate levels of activity on the IRF5 target (Fig. (Fig.5).5). In contrast, the UK and NCDV NSP1 proteins showed little or no activity on this target. Notably, our results indicated that the NSP1 proteins of the human viruses WI61, KU, Wa, and DS-1 all targeted IRF5 and IRF7 for degradation, but only one of these strains (WI61) also targeted IRF3. Thus, human RVs may rely predominantly on the degradation of IRF5 and IRF7 to subvert IFN-β signaling. On the other hand, the NSP1 proteins of those animal RVs that targeted IRF5 and/or IRF7 for degradation also targeted IRF3, a feature that may allow these viruses to more broadly attack the IFN-β signaling pathway.
Just as the coexpression of OSU NSP1 and IRF7 resulted in the enhanced accumulation of IRF7 (Fig. (Fig.4A),4A), the coexpression of OSU NSP1 and IRF5 led to the enhanced accumulation of IRF5 (Fig. (Fig.5).5). In both cases, OSU NSP1 expression correlated with increases of the IRF target to levels that were 2- to 3-fold higher than that of control (pCI)-transfected cells. The SA11-5S and K9 NSP1 proteins also showed evidence of an ability to stimulate IRF5 levels beyond that of control (pCI)-transfected 293T cells. The result with SA11-5S NSP1 is particularly interesting since its C-terminal truncation causes it to lack the necessary activity required to induce IRF3, IRF5, or IRF7 degradation.
A previous study investigating the inability of OSU NSP1 to induce IRF3 degradation led to the discovery that OSU NSP1 targeted β-TrCP for degradation. The loss of β-TrCP was suggested previously to inhibit the activation of NF-κB and thus prevent the expression of IFN-β by an IRF-independent mechanism (13). To determine whether the induction of β-TrCP degradation is a common activity among NSP1 proteins, 293T cells were cotransfected with vectors encoding FLAG-tagged β-TrCP and an individual NSP1 protein. Subsequently, β-TrCP levels in the cells were assayed by quantitative immunoblotting using an antibody against the FLAG tag. The results were surprising in that the expressions of several NSP1 proteins were found to be associated with β-TrCP accumulation to levels that were 2-fold or more above that of control (pCI)-transfected cells (Fig. (Fig.6).6). These NSP1 proteins were derived predominantly from animal RVs or animal-like human RVs (ETD, SA11-4F, SA11-5S, K9, 30-96, RRV, AU-1, and UK), raising the possibility that animal NSP1 proteins may share a unique activity that causes the enhanced accumulation of some proteins. Because the defective NSP1 protein of SA11-5S was one of the proteins that triggered enhanced β-TrCP accumulation, it is unlikely that the accumulation phenomenon is connected directly to the protein's impact on IRF degradation.
The coexpression of β-TrCP with other NSP1 proteins (NCDV, OSU, WI61, KU, Wa, Gottfried, and DS-1) resulted in β-TrCP accumulation to levels at or below that of control (pCI)-transfected cells (Fig. (Fig.6).6). Included in this set were NSP1 proteins representing human RV strains that are common causes of infant diarrhea: KU (G1P), Wa (G1P), DS-1 (G2P), and WI61 (G9P). Thus, human RVs may encode NSP1 proteins that by and large lack the activity found for the NSP1 proteins of several animal RVs that leads to the enhanced accumulation of β-TrCP in our assay system. In agreement with previous results (13), coexpression with OSU NSP1 resulted in the near-complete loss of the β-TrCP target (Fig. (Fig.6).6). Only one additional NSP1, that of the closely related human isolate WI61, triggered a reduction in β-TrCP levels to less than that of control (pCI)-transfected cells. Thus, of the 14 wild-type NSP1 proteins tested, only two exhibited activity on the β-TrCP target, suggesting that this is an uncommon mechanism of subverting IFN-β expression. However, we cannot rule out the possibility that other NSP1 proteins analyzed in our assays can target β-TrCP for degradation but that this activity has been obscured by the ability of some NSP1 proteins to trigger an enhanced accumulation of the target.
Previous reports indicated that the actions of OSU NSP1 on β-TrCP and of SA11-4F NSP1 on IRF3, IRF5, and IRF7 are mediated by proteasomal degradation (5, 6, 13). To determine whether this is a common pathway by which NSP1 proteins direct the degradation of target proteins, 293T cells were transfected with vectors encoding NSP1 proteins and treated with the proteasome inhibitor MG132. Analysis of endogenous IRF3 levels in the transfected cells at 24 h p.t. by an immunoblot assay showed that all the treated cells, regardless of the type of NSP1 expressed, contained IRF3 levels similar to that of control transfected cells (Fig. (Fig.7A).7A). Thus, MG132 prevented the NSP1 proteins of ETD, SA11-4F, K9, 30-96, RRV, UK, NCDV, WI61, and Gottfried from inducing IRF3 degradation (Fig. (Fig.3).3). This observation indicates that the mechanisms of IRF3 degradation are similar for all the NSP1 proteins tested and involve the proteasome.
Similarly, 293T cells were cotransfected with vectors encoding FLAG-β-TrCP and an NSP1 protein and then treated with MG132. Analysis of the transfected cells at 24 h p.t. by an immunoblot assay showed that β-TrCP levels were close to or below those detected in control transfected cells (Fig. (Fig.7B).7B). Remarkably, the enhanced level of β-TrCP accumulation observed for transfected cells expressing the ETD, SA11-4F, SA11-5S, K9, 30-96, RRV AU-1, or UK NSP1 protein (Fig. (Fig.6)6) was not seen when identically transfected cells were treated with MG132 (Fig. (Fig.7B).7B). Also, the levels of β-TrCP accumulation in those MG132-treated cells coexpressing some NSP1 proteins, including those of human RV strains WI61, KU, Wa, and DS-1, were substantially lower that those of control (pCI)-transfected cells. This result led to the conclusion that, in some cases, the levels of β-TrCP accumulation may be affected by more than its use as a target of NSP1-mediated proteasome degradation.
NSP1 subverts the host innate immune response by inducing the degradation of factors necessary for IFN-β signaling. Descriptions of NSP1 activities have come from previous studies examining a limited number of RVs, all representing culture-adapted animal virus strains. For instance, insight into the activity of NSP1 on IRF3 came from analyses with SA11-4F, RRV, OSU, NCDV, UK, and EW or gene 5 clones of these viruses. Moreover, the activity of NSP1 on all three IRF targets (IRF3, IRF5, and IRF7) has been examined only for SA11-4F, and the activity on β-TrCP has been examined only for OSU and NCDV. To gain a broader perspective of the activities of NSP1, we analyzed the activities of NSP1 proteins originating from RV strains recovered from a range of mammalian species, including human, on IRF and β-TrCP targets. We found that all tested RV strains encoding full-length NSP1 proteins shared the ability to inhibit IFN-β gene expression in a human cell line but that this activity was not due strictly to the degradation of IRF3. Furthermore, we found that the ability of an NSP1 protein to degrade one target protein generally does not indicate whether the protein will have activity on a second target. Our results indicate that different strains of RV can take distinct approaches for subverting the host antiviral response.
This study extensively probed the link between the capacity of RV NSP1 proteins to induce the degradation of targets of the IFN signaling pathway and the capacity to suppress IFN-β gene expression. We found that infection with SA11-5S, which expresses a C-terminally truncated and therefore defective form of NSP1, stimulated the expression of IFN-β and was unable to induce the degradation of IRF3 or any other known target. In contrast, the degradation of IRF3 by numerous RV strains correlated with an inhibition of IFN-β expression (Fig. (Fig.11 and and2).2). These strains, including SA11-4F, RRV, NCDV, SA11-L2, SNF, and SRF, resulted in the near-complete degradation of endogenous IRF3 (Fig. (Fig.2).2). These results were expected, as previous studies correlated the inhibition of IFN-β with IRF3 degradation, suggesting that this mechanism is commonly used among RV strains to inhibit host antiviral responses (5, 13, 32).
Interestingly, infection with SKF and SDF, which express the human KU and DS-1 NSP1 proteins, respectively, inhibited IFN-β expression (Fig. (Fig.1),1), even though these viruses were unable to induce a significant degradation of IRF3 (Fig. (Fig.2).2). This finding indicates that SKF and SDF interfere with the production of IFN-β via a mechanism other than through IRF3 degradation. Transient expression assays showed that both of these NSP1 proteins were able to target IRF7 and IRF5 for degradation, suggesting that these activities represent a primary mechanism by which human RVs inhibit IFN-β expression. In support of this idea, transient expression assays showed that the NSP1 proteins of human RV strains Wa and WI61 also targeted IRF7 and IRF5 for degradation. Because IRF7 and IRF5 are expressed at high levels in dendritic cells (2, 4), the ability of human RVs to induce the degradation of these targets may enable these viruses to spread beyond intestinal epithelial cells, the initial sites of infection. While the NSP1 proteins of most of the tested human RVs triggered IRF7 and IRF5 but not IRF3 degradation (KU, Wa, and DS-1), most of the animal RVs caused the degradation of all three (SA11-4F, 30-96, RRV, UK, and NCDV). Thus, animal RVs may mount a broader attack on the IFN-β signaling pathway, which may also explain why animal RVs tend to grow more efficiently in cell cultures.
OSU infection also inhibited IFN-β transcription (Fig. (Fig.1),1), and in agreement with data from previous reports, the low level of IFN-β mRNA did not correlate with IRF3 degradation (Fig. (Fig.2)2) (13). The β-TrCP component of the cellular E3 ubiquitin ligase complex SCFβ-TrCP, when expressed as a FLAG-tagged protein, was confirmed to be targeted for degradation by OSU NSP1 (Fig. (Fig.6),6), indicating that this is indeed a possible mechanism of inhibiting IFN-β expression. WI61 was the only other virus in our assays that appeared to induce β-TrCP degradation. Although OSU NSP1 caused the degradation of β-TrCP but not the IRF proteins, WI61 NSP1 was novel in that it exhibited some activity against both IRF proteins and β-TrCP. The ability of WI61 NSP1 to target both IRF family members and β-TrCP is surprising given that they are structurally distinct targets (27, 36). The fact that only 2 of the 15 species of NSP1 tested in this study could direct β-TrCP degradation implies that this does not represent a common mechanism by which RV subverts IFN-β expression.
An unexpected finding of our analysis was that the transient expression of OSU NSP1 with IRF5 or IRF7 stimulated the accumulation of the IRF targets to levels 2- to 3-fold higher than those in the absence of NSP1 expression (Fig. (Fig.44 and and5).5). Similarly, the expression of any of several NSP1 proteins with β-TrCP stimulated the accumulation of this target at levels well above that detected for cells expressing β-TrCP only (Fig. (Fig.6).6). Quantitative RT-PCR indicated that the enhanced accumulation of the IRF and β-TrCP targets did not stem from changes in the amounts of IRF and β-TrCP transcripts produced in the transfected cells. The enhanced accumulation of β-TrCP occurred even upon coexpression with a mutant form of NSP1 (SA11-5S), indicating that the enhancement phenomenon is not tied directly to the process by which NSP1 proteins induce target degradation. Perhaps, the enhancement phenomena are connected to one of the other features described previously for NSP1, including the capacity to bind RNA or localize to the nucleus when truncated (18). Through these activities, NSP1 proteins may impact the export of target transcripts to the cytoplasm or the efficiency of their translation.
Our results show that the pattern of IRF and β-TrCP degradation varies greatly among NSP1 proteins. Similarly, the amino acid sequences of NSP1 proteins are known to display extensive variation. Despite such sequence variation, phylogenetic analysis reveals that NSP1 proteins resolve within two classes, which, for the sake of discussion, we term classes I and II (Fig. (Fig.8).8). All of the class I NSP1 proteins that we tested in our assays were found to have activities. For example, SA11-4F NSP1 induced the degradation of IRF3, IRF5, and IRF7, although it did not affect β-TrCP. Like SA11-4F NSP1, the ETD, RRV, and 30-96 NSP1 proteins induced the degradation of IRF3, IRF5, and IRF7 but not of β-TrCP. Slightly different results were obtained for canine virus strain K9. Its NSP1 protein induced the degradation of IRF3 but caused little or no degradation of IRF5 and IRF7 (nor did it cause the degradation of β-TrCP). Thus, the class I NSP1 proteins are characterized by the ability to induce the degradation of IRF proteins, which in all cases included IRF3. None of the class I proteins targeted β-TrCP for degradation.
In comparison to the class I proteins, the class II proteins showed more diverse activities. For example, WI61 NSP1 directed the degradation of not only IRF proteins (IRF5 and IRF7) but also β-TrCP. In contrast, OSU NSP1 induced β-TrCP degradation but did not affect IRF proteins. The class II UK and NCDV NSP1 proteins induced the efficient degradation of IRF3 and IRF7 but had more limited effects on IRF5 or β-TrCP. In all cases where the class II proteins had activity on an IRF target, one of these targets included IRF7. Collectively, our analysis indicates that class I NSP1 proteins antagonize IFN signaling by inducing the degradation of IRF proteins, while the class II NSP1 proteins use more diverse approaches, including in some cases targeting IRF proteins and/or β-TrCP.
The prototypic human RV strains Wa and DS-1 appeared to preferentially target IRF5 and IRF7 for degradation. Only AU-1 NSP1 failed to exhibit any activity at all in our assays. AU-1 represents an uncommon human strain distinct from the predominant Wa and DS-1 genogroups, likely originating from the interspecies transmission of a feline RV to a human (34). The basis for the inability of AU-1 NSP1 to target any of the IRF proteins or β-TrCP is not known but may stem from a failure to interact with the target protein or from an intrinsic defect in some associated catalytic function (e.g., E3 ubiquitin ligase-like activity). However, the related bovine UK NSP1 was able to target IRF3 and IRF7 for degradation, raising the possibility that the differences between these two proteins could be useful for mapping the specific sites in the NSP1 protein necessary for target activity.
The results of this study indicate that RV NSP1 has evolved numerous ways to antagonize the IFN-β signaling pathway. The recognition that different functional classes of NSP1 proteins exist suggests that multifaceted experimental approaches should be taken in characterizing the activities of the protein and that any one assay system is not likely to reveal the complete nature of its IFN-β antagonist activity. In addition to activities against IRF proteins and β-TrCP, NSP1 may subvert innate immune responses via other mechanisms, such as through its RNA-binding activity. The identification of NSP1 sequence variations that influence target specificity and activity may help to define virulence factors for RVs as well as provide clues as to the basis of host range restriction. Further insights into an understanding of the mechanisms of NSP1 activity may lead to the establishment of alternative live-attenuated vaccine candidates through reverse vaccinology.
We are grateful to Nobumichi Kobayashi for providing gene 5 reassortant viruses, Zenobia Taraporewala and Harry Greenberg for sharing gene 5 plasmids, Andrew Rolle and Christine Rippinger for technical support, and Kristen Guglielmi, Sarah McDonald, Aitor Navarro, and Shane Trask for critical review of the manuscript and helpful discussions.
This work was supported by the Intramural Research Program of the National Institute of Allergy and Infectious Diseases, National Institutes of Health.
Published ahead of print on 22 December 2010.