Identification of spliceosome SNRNP200 required for SeV-induced IFNB1 transcription

A genome-wide gene silencing screen that assessed the transcriptional activity of the IFNB1 promoter following SeV infection was previously performed to identify novel regulators of innate immunity [8]. Six genes that encode spliceosome components that reduced IFNB1 transcription upon gene silencing were identified (Fig 1A). Among these genes, one encodes an RNA helicase (SNRNP200) and two (SF3A1 and SRSF1) were shown to regulate innate immune responses by the alternative splicing of either Myd88 or IRF3 [10,11]. To further explore RNA helicases that play central roles in splicing and that often function in proofreading events in pre-mRNA splicing [21], an RNAi mini-screen using five independent lentiviruses that express short hairpin RNA (shRNA) targeting most spliceosomal RNA helicases was performed (S1A Fig). SNRNP200 was the only RNA helicase assigned to the Ski2-like helicase subfamily that showed a significant reduction in IFNB1 promoter-driven reporter activity. The KD of SNRNP200 was validated through the marked decrease in the mRNA and the protein levels while its specificity was validated by the absence of any off-target effects on other spliceosome gene hits (S1B and S1C Fig). The depletion of SNRNP200 reduced IFN-β production at 8 hours post-infection reaching levels comparable to those obtained in DDX58 KD cells at 48 hours post-infection (Fig 1B). Interestingly, while the depletion of SNRNP200 completely inhibited IFIT1 (also known as ISG56) induction (S1C Fig), its overexpression could not increase IFIT1 levels in neither the non-infected nor the SeV-infected cells. To further investigate SNRNP200’s contribution to the antiviral response, the viral susceptibility of SNRNP200 KD cells was monitored in a time-course experiment by following IFIT1 induction along with the production of infectious particles and the viral protein levels (Fig 1C and 1D). In control HEK 293T cells transduced with a non-target sequence shRNA (shNT)-expressing lentivirus, SeV protein was only detectable at 24 hours post-infection which coincided with IFIT1 induction (Fig 1C). In contrast, in SNRNP200 KD cells, SeV protein was readily detectable at 8 hours post-infection becoming more significant at 24 hours post-infection. However, IFIT1 induction was only detected at 48 hours post-infection. Of notable importance, SNRNP200 KD cells were observed to yield up to a 2-log increase in viral titers when compared to the control (Fig 1D). Correlating with a reduced early IFNB1 induction, it was confirmed that silencing SNRNP200 also increased the replicative potential of influenza A virus (FLUA) and hepatitis C virus (HCV) in HEK 293T and Huh7 cells, respectively (S2 Fig). Based on the epistasis analysis, the transcriptional activity of the IFNB1 promoter is slightly affected by ectopic expression of constitutively active IRF3(5D) [22], while it was completely blocked in SeV infected or MAVS overexpressing cells transduced at a high multiplicity of infection (MOI of 20) with lentiviral-expressing shRNA (Fig 1E). Similar results were obtained in A549 cells (S3A and S3B Fig). Interestingly, upon SeV infection and in contrast to IRF3 overexpression, the ectopic expression of IRF3(5D) could rescue the induction of IFIT1 (Fig 1F), suggesting a role for SNRNP200 in IRF3 activation. The effect of SNRNP200 KD in NF-κB-dependent transcription was then investigated using a reporter assay (p2xNF-κB_fLUC) in HEK 293T cells. It was found that SNRNP200 KD cells display no attenuation of poly (I:C)-, MAVS-, TBK1-, or p65-mediated activation of the NF-κB promoter (S4A Fig). In contrast, in these cells, there was a significant inhibition of SeV-, poly (I:C)-, TBK1-, and IFN-α-mediated activation of the ISG56 promoter (S4B Fig). It was confirmed that SNRNP200 silencing does not affect NF-κB-dependent transcription in SeV-infected A549 cells through the quantification of TNF, NFKBIA, and TNFAIP3 mRNA levels using qRT-PCR (S3C Fig). Interestingly, neither the TRIF nor the cGAS/STING pathways are affected in SNRNP200 KD cells whereas the RLR pathway, which converge to the TBK1-mediated phosphorylation of their respective adaptors (TRIF, STING, and MAVS) to recruit IRF3 and to license IRF3 for phosphorylation [23]. These data suggest that, upon RNA virus infection, SNRNP200 may function between MAVS signal transduction and TBK1-mediated IRF3 licensing. These observations led to exploring a specific regulatory role of SNRNP200, a core component of U4/U6-U5 small nuclear RNA (snRNA) [24], in the downstream activation of IRF3, of production of IFNB1 and ultimately of an optimal antiviral response.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. SNRNP200 spliceosome protein is required for virus-induced IFNB1 production to control viral replication.

(A) HEK 293T cells stably expressing an IFNB1promoter-driven luciferase gene (HEK 293T pIFNB1-Luc) are transduced with different lentiviral-expressing shRNA targeting SNRNP200, SFRS1, SNRNP35, SF3A1, PHF5A and NHP2L1 genes. Left panel—Heat map (log2 scale) indicating the modulation of IFNB1 promoter activity following silencing of spliceosome genes and infection with SeV or transfection of poly I:C, MAVS or IRF3(5D) expression plasmids for 16 hours. Right panel—qRT-PCR validation data of the endogenous IFNB1 mRNA levels and target gene KD efficiency of cells transduced with shRNA. (B) HEK 293T are transduced with lentiviral-expressing shRNA control (shNT) or targeting SNRNP200 (shSNRNP200) or DDX58 (shDDX58) for three days and infected cells with SeV for 8, 24 or 48 hours. Supernatants are harvested and IFN-β secretion levels are measured by ELISA. (C) Immunoblot analysis of HEK 293T cells infected with SeV for 8, 24 or 48 hours following treatment with shNT or shSNRNP200 for three days. SeV, IFIT1 and actin proteins are resolved by immunobloting at the indicated time. (D) Infectivity titers of SeV particles produced as indicated in (C) are determined by harvesting supernatants at the indicated time and infecting VERO cells in virus plaque assays. (E) HEK 293T pIFNB1-Luc cells are transduced with four different shSNRNP200 at a multiplicity of infection (MOI) of 5 and 20 for three days. Relative IFNB1 promoter activity are reported as percentage of the control shRNA NT after infection with SeV or transfection of poly I:C, MAVS or IRF3(5D) expression plasmids for 16 hours (left). Simplified schematic of RLR signaling pathway leading to IFN-β promoter induction (right). Deduced points of action of SNRNP200 are marked with asterisks (blue and green for MOI = 5 and 20, respectively). Knockdown efficiencies at the various MOI are determined by immunobloting analysis of SNRNP200 protein levels. (F) Immunoblot analysis of HEK 293T cells transduced with shNT or shSNRNP200 for three days and subjected to SeV infection for 16 hours. Plasmids encoding eYFP, IRF3 and IRF3(5D) are transfected for 48 hours. Following cell harvesting, IRF3 and IFIT1 protein levels are resolved by immunobloting analysis of cell lysates.

https://doi.org/10.1371/journal.ppat.1005772.g001

SNRNP200 specifically regulates IRF3 signaling upon RNA virus infection

To understand the manner in which SNRNP200, upon viral infection, contributes to IRF3-mediated IFNB1 production, the effect of SNRNP200 silencing on the expression of established members of the RLR pathway was evaluated using a western blot analysis (S5A Fig). First, a decreased protein expression of IRF3 in SNRNP200 KD cells was observed. This correlated with the blockage of the SeV-mediated induction of IFIT1, DDX58, and IFIH1 proteins. The decreased IRF3 protein levels were further confirmed at the mRNA level by qRT-PCR, paralleling the reduced mRNA levels of SNRNP200 and its effector genes (IFNB1, IFIT1, DDX58, and IFIH1) (S5B Fig). While the residual IRF3 protein levels of KD cells are sufficient for the activation of the cGAS/STING pathway (S4B Fig), a complete inhibition of IRF3 phosphorylation at serine 386 following SeV infection (Fig 2A, see IRF3-p386) was observed. This suggested a specific contribution of SNRNP200 during IRF3-activation-mediated IFNB1 production. A weak decrease of the basal protein expression levels of DDX58 in SNRNP200 KD cells was also observed (S5A Fig). The mRNA levels, however, were comparable to the control shNT cells (S5B Fig), suggesting that SNRNP200 enhances the RLR-mediated antiviral signaling potential of DDX58 at the basal level. In contrast, the protein expression of MAVS, TBK1, IKBKE, RELA (p65), and TRAF3, which all contribute to the signal propagation for IFNB1 induction, remained unchanged in all conditions (S5A Fig). Similar observations were made for the protein expression of the housekeeping genes ACTIN, TUBULIN, and GAPDH (S5A Fig). To better evaluate the outcome of the reduced basal protein levels of DDX58 and IRF3 on IFNB1 production, ectopic expression of DDX58 and IRF3 in SNRNP200 KD cells was used in an attempt to restore antiviral response. Surprisingly, overexpression of DDX58 and IRF3 neither alone nor in combination could restore SeV-mediated IFIT1 induction (Fig 2B and S6A and S6B Fig). Furthermore, ectopic expression of neither DDX58 nor IRF3 restored IFNB1 promoter reporter activity or IFN-β cytokine production upon SeV infection (Fig 2D and 2E). This was in stark contrast to the almost complete rescue achieved by the ectopic expression of IRF3(5D) (Fig 2D and 2E). The phosphorylation of IRF3 at serine 386 (IRF3-p386) was also noted as a key event for IRF3 activation and, as such, the proportion of IRF3-p386 in relation to total IRF3 was investigated. The quantitative ratios of IRF3-p386 to total IRF3 in context of endogenous or overexpressed DDX58, IRF3, IRF3(5D), and cGAS/STING were determined (Fig 2B). When comparing control shNT-treated with SNRNP200 KD cells in the context of SeV-mediated infection, regardless of whether or not DDX58 or IRF3 was overexpressed, a significant reduction in the IRF3-p386/IRF3 ratios (from 0.6–0.9 to 0.1–0.2) was observed (Fig 2B). This establishes the requirement of SNRNP200 for downstream IRF3 activation independent of its effects on protein expression. Furthermore, ectopic expression of IRF3(5D) in SNRNP200 KD cells yielded IRF3-p386/IRF3 ratios comparable to those of the control shNT-treated cells (0.8 vs 0.6–1.3) correlating with the almost complete restoration of SeV-mediated IFNB1 production. The effect of SNRNP200 KD on the DNA sensing arm of the antiviral response downstream of cGAS/STING was also investigated. It was shown that SNRNP200 is dispensable for cGAS/STING-mediated IFIT1 induction, IFN-β production, and IFNB1 promoter activity (Fig 2B, 2D and 2E), further implying a specific role for SNRNP200 in the RLR-mediated IRF3 signaling pathway upon RNA virus infection. Interestingly, the higher IRF3-p386/IRF3 ratios (4.5) in SNRNP200 KD cGAS/STING expressing cells versus control cells (2.3) largely reflects a significant increase in IRF3 activation and hence in IFIT1 induction (Fig 2B). This suggests that SNRNP200 potentially competes with the STING adaptor during TBK1-mediated IRF3 phosphorylation. Although IRF3 expression slightly increased IFN-β secretion and the IFNB1 promoter activity when activation of the cGAS/STING pathway and SeV infection were combined in SNRNP200 KD cells (Fig 2D and 2E), similar IFNB1 induction was observed in uninfected cells (Fig 2C), demonstrating that IRF3 protein levels in SNRNP200 KD cells have little functional consequence on the cytosolic DNA sensing pathway. Finally, the IRF3 mRNA splice junctions and the presence of exons were investigated to explain the reduced mRNA and protein levels of IRF3. No splicing variants were observed excluding an alternative splicing regulation of IRF3 and supporting a reduction in the efficiency of pre-mRNAs splicing to explain the phenotype of SNRNP200 KD cells (S7 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. SNRNP200 KD restricts SeV- and type I IFN-mediated induction of antiviral response.

(A) HEK 293T cells are transduced with shNT or shSNRNP200 and infected with SeV or stimulated with IFN-α for 24 hours. Selected genes are resolved by immunobloting and compared to shNT control cells. (B) HEK 293T are transduced with shNT or shSNRNP200 for 3 days and transfected with DDX58, IRF3, IRF3(5D) and cGAS-STING expression plasmids for the last 48 hours, and subsequently infected with SeV for 24 hours. Selected genes are resolved by immunobloting and compared with cells transduced with shNT. (C) HEK 293T pIFNB1-Luc cells are transduced with shNT or shSNRNP200 for 3 days and transfected with DDX58, IRF3, IRF3(5D) and cGAS-STING expression plasmids for the last 48 hours. Luciferase levels are resolved and compared to shNT control cells. (D) HEK 293T pIFNB1-Luc cells are transduced with shNT or shSNRNP200 for 3 days and transfected with DDX58, IRF3, IRF3(5D) and cGAS-STING expression plasmids for the last 48 hours, and subsequently infected with SeV for 24 hrs. Luciferase levels are resolved and compared to shNT cells. (E) HEK 293T cells are treated as indicated in D. At 24 hrs post-infection, supernatants are harvested and IFN-β secretion levels are measured by ELISA.

https://doi.org/10.1371/journal.ppat.1005772.g002

Interestingly, the induction of IFIT1, DDX58, and IFIH1 was also inhibited upon treatment of SNRNP200 KD cells with IFN-α (Fig 2A). Furthermore, similar levels of IFNα/β receptor alpha chain (IFNAR1), of STAT1, and of phosphorylation at tyrosine 701 (STAT1pY701) were observed between control and SNRNP200 KD cells following stimulation with IFN-α. This suggests the involvement of SNRNP200 at a later stage of type I IFN signaling downstream of STAT1 phosphorylation. The negative effect of SNRNP200 KD on IFN-α signaling was also demonstrated by the reduced mRNA levels of IFN inducible genes IFIT1, DDX58, and IFIH1 in A549 cells (S3C Fig).

To comprehensively understand the effect of SNRNP200 KD on type I IFN production (early) and signaling (late), expression profiling studies of non-stimulated (NS), SeV-infected and IFN-α-treated SNRNP200 KD cells versus control shNT HEK 293T cells were performed to assess differential gene expression (Fig 3 and S8 Fig). The effect of SNRNP200 KD on basal gene expression of NS cells was characterized; 2,880 altered transcripts (cutoff of 1,5 log2 fold induction) that are primarily associated with immune system functions and cell cycle regulation were found based on a Reactome pathway enrichment analysis (S8A Fig). A list of transcriptionally altered genes by SeV infection or by IFN-α stimulation in control shNT cells was then established. Using a stringent significance cutoff (p ≤0,001), 52 genes altered by SeV infection and 55 genes altered by IFN-α stimulation were found to be transcriptionally affected by SNRNP200 (Fig 3A). Within these subsets of genes, 13 were in common reflecting the expected overlap of the early (SeV-mediated IFNB1 production) and the late (type I IFN signaling) arms of the antiviral response. Within the subset of commonly affected genes, all showed altered expression upon SNRNP200 silencing with a mean difference of 3,8 log2 fold change when compared to the shNT control cells (Fig 3B). On the other hand, the two subsets of 39 SeV-specific and 42 IFN-α specific genes were hindered by SNRNP200 silencing by a mean difference of 2,6 and 2,5 log2 fold change, respectively, when compared to the shNT control cells (Fig 3C and 3D). This demonstrated that SNRNP200 plays a distinct role in the early and the late antiviral response pathways. The relationship between the affected gene subsets was assessed by resolving their interaction and functional alignment networks. The assessment showed that the top GO term for genes exclusively affected by SeV is “response to virus”, and the top GO term for IFN-α is “response to type I interferon”. Furthermore, it showed that SeV-specific genes affected by silencing SNRNP200 are IRF3-dependent (IFNB1, IL29, and BIRC6) and that several IFN-α-specific genes are JAK-STAT1-dependent (ACVR1C, CIQA, IFIT5, and OAS1). The latter might be explained by the reduction of IFN-induced IRF9 and STAT1 mRNA levels encoding key transcription factors of ISGF3 that mediates signaling of type I and III IFNs (S2 Table). The molecular signature of differential gene expression strengthens the observation that SNRNP200 silencing hinders IRF3-dependent gene induction, leading to a general weakening of the RLR signaling pathway, and further suggests that SNRNP200 plays a distinct regulatory role in type I IFN signaling. The results suggest that SNRNP200 specifically regulates IRF3 activation upon RNA virus infection to promote IFNB1 induction and IFN effector responses, and thus demonstrates its importance in controlling viral infections.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Transcriptional profiles of SNRNP200 KD cells reveal altered expression of genes induced by SeV infection and IFN-α treatment.

A) Summary diagram of the transcriptional analysis of shNT control cells to illustrate that out of the 525 SeV altered transcripts and 957 IFN-α altered transcripts, 52 and 55 genes have a differential gene expression (p ≥ 0.001) upon SeV infection or IFN-α stimulation, respectively. (B) Differential gene expression of the 13 transcripts affected by SeV infection and IFN-α stimulation. Expression Fold Change are shown for shSNRNP200 cells (200_NS), shSNRNP200 cells + SeV (200 SEV) and shSNRNP200 cells + IFN-α (200_IFNa) and for shNT control cells (NT_SEV, NT_IFNa). Numerical values are log2 fold change. (C) Differential gene expression of the 39 transcripts affected by SeV infection. Expression Fold Change are shown for shSNRNP200 cells (200_NS), shSNRNP200 cells + SeV (200 SEV) and shSNRNP200 cells + IFN-α (200_IFNa) and for shNT control cells (NT_SEV, NT_IFNa). Numerical values are log2 fold change. Top 15 genes are displayed. Complete list and gene network are available in supporting information (S1 Fig and S1 Table). (D) Differential gene expression of the 42 transcripts affected by IFN-α stimulation. Expression Fold Change are shown for shSNRNP200 cells (200_NS), shSNRNP200 cells + SeV (200 SEV) and shSNRNP200 cells + IFN-α (200_IFNa) and for shNT control cells (NT_SEV, NT_IFNa). Numerical values are log2 fold change. Top 15 genes are displayed. Complete list and gene network are available in supporting information. (E) Interaction network of the 13 common genes shown in (a) and affected by SNRNP200 silencing. The colors inside the dot represent their biological processes (see legend on left). Lines represent physical interactions (protein-protein interactions), pathways (blue), and co-localization (purple) attributes. Shaded nodes represent input data; Black nodes represent most likely first degree interactor. Higher magnification and input genes of this network can be found in supporting documentation. The network on the right corner represents the connections to IRF3.

https://doi.org/10.1371/journal.ppat.1005772.g003

Sec63-1 domain of SNRNP200 is required for virus-mediated IFNB1 production

To examine the manner in which SNRNP200 directly contributes to IRF3-mediated IFNB1 activation upon SeV infection, a series of recombinant SNRNP200-truncated mutants were tested for their ability to rescue IFNB1 reporter activity and ISG expression in SNRNP200 KD cells (Fig 4 and S9 Fig). None of the truncated mutants could induce an antiviral response in SNRNP200 KD cells (with the exception of a weak IFNB1 activation by expression of a D1-D3 construct). Indeed, the deletion of the C-terminal Sec63 domain (Sec63-2) alone, which was reported in yeast (Sec63-2 deleted Brr2 protein) to reduce ATPase/helicase activity and splicing [25], completely abolished the activation of IFNB1 promoter-driven reporter activity and the induction of IFIT1 and DDX58 upon SeV infection. To further explore a dual regulatory role in splicing and RNA-mediated antiviral responses, the described SNRNP200 heterozygous mutations associated with the autosomal dominant RP33 disease were considered [26–28]. In particular, the SNRNP200 S1087L and R681C variants located within the Sec63-1 homology domain and the N-terminal RecA-like ATPase/helicase domains, respectively, were investigated (Fig 4A). It was first demonstrated that the ectopic expression of RNAi-resistant WT SNRNP200 rescues the SeV-mediated IFN-β secretion and the IFNB1-driven reporter activity in KD cells, further validating the specificity and minimal off-target effects of shSNRNP200 and its associated immunoregulatory phenotype (Fig 4). Surprisingly, expression of the SNRNP200 S1087L mutant completely eliminated the ability to rescue IFNB1 activation (Fig 4B and 4C). Similar results were obtained using qRT-PCR as the rescue of endogenous IFNB1 mRNA levels was achieved only by the expression of WT SNRNP200 (S10 Fig). Concordantly, WT SNRNP200, but not the S1087L mutant, restores IRF3 protein levels, and more importantly, restores the phosphorylation of IRF3 at serine 386 as well as the inducible levels of DDX58 and IFIT1 upon SeV infection (Fig 4D). WT SNRNP200, but not the S1087L mutant, also restores IFN-α-dependent DDX58 and IFIT1 induction. It was also determined that expression of R681C variant only slightly rescues IFNB1 promoter-driven reporter activity and IFN-β secretion (Fig 4B and 4C). Interestingly, while investigating a mutation within the ATP binding motif, it was found that the ectopic expression of a SNRNP200 C502A variant elicited an IFNB1 response independent of viral infection (S11 Fig) in line with the recently reported natural gain-of-function of DDX58 and IFIH1 ATPase-deficient variants [29]. The constitutive induction of IFNB1 with expression of SNRNP200 C502A is further enhanced upon SeV infection to levels similar to the WT enzyme (Fig 4B and 4C), thereby suggesting the requirement of a functional SNRNP200 ATPase in conferring specificity to viral RNA and in preventing signaling through the recognition of self-RNA. Thus, the data firmly establishes a critical role of the Sec63-1 domain to promote virus-mediated IFNB1 production and further suggests a contribution of the N-terminal ATPase/helicase domain in sensing viral RNA.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. SNRNP200, but not Sec63-containing S1087L mutant, rescues SeV- and IFN-α-mediated induction of antiviral response in SNRNP200 KD cells.

(A) Schematic representation of SNRNP200 protein, C-terminal truncation and clinically relevant mutants. (B) HEK 293T pIFNB1-Luc cells are transduced with shSNRNP200 and transfected with RNAi-resistant SNRNP200 WT and variants expression plasmids bearing the indicated mutation or eYFP as a control. Following 24 hours of SeV infection, total luciferase levels are measured and compared with control shNT cells. (C) HEK 293T are treated as indicated in B. IFN-β secretion levels are measured by ELISA and compared with shNT cells. (D) HEK 293T cells are transduced with shSNRNP200 and transfected with RNAi-resistant SNRNP200 WT and variants expression plasmids bearing the indicated mutation or eYFP as a control. At 24 hours, cells are harvested and DDX58, IFIT1, IRF3 and IRF3pS386 levels are resolved by immunobloting analysis of cell lysates.

https://doi.org/10.1371/journal.ppat.1005772.g004

Sec63-1 domain of SNRNP200 is a major determinant of viral RNA recognition

DExD/H-box helicases, such as RIG-I, are engaged in antiviral innate immunity because they detect viral nucleic acids and prevent the recognition of self-RNA through ATP hydrolysis (29). As the Sec63-1 containing S1087L mutation was reported to diminish binding to RNA duplex and to reduce RNA-stimulated ATPase/helicase activity without any discernible effect on the folding of SNRNP200 (27), it was hypothesized that this natural loss-of-function mutation abolishes the recognition of viral RNA for IFNB1 induction. To determine whether or not the S1087L variant impaired the binding of the immunostimulatory RNA in SeV-infected cells, the in vitro ability of exogenously expressed SNRNP200 to bind biotinylated polyinosinic-polycytidylic acid (poly (I:C)) was measured using an RNA pull-down and western blot analysis of bead-bound protein fractions (Fig 5). It was shown that FLAG-WT SNRNP200 binds poly (I:C), which is used as a viral double-stranded RNA (dsRNA) surrogate, only in SeV-infected cell extracts (Fig 5A). Furthermore, a complete loss of poly (I:C) binding by the FLAG-SNRNP200 S1087L variant was observed. Interestingly, the FLAG-Sec63-1 domain, but not the FLAG-Sec63-2, is sufficient to bind poly (I:C) (Fig 5B). These observations were confirmed with biotinylated HCV genomic RNA. As expected, WT SNRNP200 and the Sec63-1 domain, but neither the S1087L variant nor the Sec63-2 domain, were able to successfully pull-down HCV RNA (Fig 5C). To provide insight into the interaction of SNRNP200 and immunostimulatory RNA molecules, SNRNP200’s ability to bind a synthetic 5’-triphosphate (5’ppp) and a double-stranded stretch of RNA using the full-length HCV genome, produced in vitro by transcription with a T7 polymerase, was investigated (Fig 5D). A comparable binding of FLAG-WT SNRNP200 to the untreated and to the calf-intestine alkaline phosphatase (CIAP)-treated blunt-ended HCV RNA was observed. This comparable binding suggests that the 5’ppp moiety is not essential for the recognition of viral dsRNA. FLAG-SNRNP200 does not bind dsDNA molecules unlike the FLAG-cGAS control which does [30]. This is reflected by the lack of pull-down by FLAG-SNRNP200 with biotinylated polydeoxyadenylic acid-polythymidylic acid (poly (dA:dT)) and polydeoxyguanylic acid-polydeoxycytidylic acid (poly (dG:dC)) homopolymer molecules (Fig 5D). To assess the requirement of a protein complex with DDX58, MAVS, or TBK1 for binding HCV RNA, expressions of these proteins both individually and together were silenced, and RNA pull-down assays were performed to detect SNRNP200 (Fig 5E). It was determined that SNRNP200 binds HCV RNA regardless, ruling out a contribution of these proteins in its ability to recognize viral RNA. Finally, FLAG-tagged WT SNRNP200 and S1087L variant were immunoprecipitated upon SeV infection, and the co-purified RNA molecules were analyzed using qRT-PCR in SNRNP200 KD and in control shNT cells (Fig 5F). Increased amounts of actin mRNA for both immunoprecipitated proteins were found compared to the eYFP control (normalized to RNA levels of cell lysates). A significant enrichment of SeV RNA, which is more important in SNRNP200 KD cells than in shNT control cells that express the endogenous untagged protein, was observed upon immunoprecipitation of FLAG-WT SNRNP200, demonstrating a direct binding to viral genomes. The amount of SeV RNA recovered with the WT was almost 10- to 20-fold higher than with the S1087L variant in KD cells (and 3-fold in shNT cells), reflecting an altered RNA binding ability of the mutant. Despite the weak binding of SeV RNA by S1087L which is possibly due to its N-terminal RecA domains, the loss-of-function in IFNB1 induction reveals that this interaction is not biologically active. The data demonstrate that the Sec63-1 domain of SNRNP200 is a major determinant of viral RNA recognition and consequently of SNRNP200’s ability to activate antiviral innate immune responses.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. SNRNP200, but not S1087L mutant, binds viral RNA in vitro.

(A) HEK 293T cells are transfected with FLAG-eYFP (control), FLAG-SNRNP200 or FLAG-SNRNP200 S1087L mutant expression plasmids for 48 hours and infected with SeV for 16 hours. RNA pull-down assays are performed with cell lysates using biotinylated poly (I:C). Cell lysates and bead-bound complexes are analyzed by Western blotting and compared to uninfected control cells. (B) HEK 293T cells are transfected with FLAG-Sec63-1 and FLAG-Sec63-2. RNA pull-down assays are performed and analyzed as indicated in (A). (C) HEK 293T are treated as indicated in (B) and RNA pull-down assays are performed on cell lysates using biotinylated HCV RNA and analyzed as indicated in (A). (D) HEK 293T cells are transfected with FLAG-SNRNP200 (top panel) or FLAG-cGAS (bottom) expression plasmids for 48 hours. RNA pull-down assays are performed with cell lysates using biotinylated HCV RNA (5’ppp) that is either left untreated or treated with CIAP or heat inactivated (h/i) CIAP (control), and with biotinylated poly (dA:dT) and poly (dG:dC) DNA molecules. Pull-down complexes are resolved by immunobloting and compared to protein input and uncoated beads as negative control. (E) Pull-down assays with biotinylated HCV RNA and transfected FLAG-SNRNP200 are performed with lysates of cells transduced with shRNA targeting MAVS, RIG-I or TBK1, either alone or in combination. (F) HEK 293T cells are transduced with shNT or shSNRNP200 and transfected with FLAG-eYFP, RNAi-resistant FLAG-SNRNP200 WT or S1087L variant expression plasmids and both (comb) for 48 hours. At 16 hours post-infection with SeV, cell lysates are subjected to an anti-FLAG immunoprecipitation, and RNA molecules are extracted from the immune complexes and analyzed by qRT-PCR. SeV and actin RNA levels are determined and normalized to RNA levels of cell lysates.

https://doi.org/10.1371/journal.ppat.1005772.g005

Sec63-1 domain of SNRNP200 interacts with TBK1

To better define a specific immunoregulatory role of SNRNP200, binding partners were identified by screening proteins of the antiviral signaling pathways upon immunoprecipitation of FLAG-tagged SNRNP200. This method successfully allowed the detection of a constitutive interaction between SNRNP200 and the ubiquitously expressed kinase TBK1. This interaction was also detected when using the SNRNP200 S1087L mutant (Fig 6A). The ability of various SNRNP200-truncated mutants (Fig 4A) to bind TBK1 (Fig 6B) was then assessed. A mutagenesis analysis showed that the Sec63-1 domain of SNRNP200 is required and sufficient for TBK1 interaction (Fig 6C), which is similar to the observation for RNA binding (Fig 5B and 5C). Both Sec63 homology domains of SNRNP200 contain a helical bundle (HB) and immunoglobulin-like (IG) sub-domains separated by a helix loop helix (HLH) motif. To more accurately map the TBK1 binding domain, the sub-domains of Sec63-1 were expressed separately. A weak interaction with the HLH-IG sub-domain was observed, suggesting its contribution to the binding of TBK1 (Fig 6C). It was also demonstrated that the C-terminal Sec63-2 domain could not bind TBK1, which corroborates the detected interaction of the N-terminal truncated D1-3, D1-4, and D1-5 mutants with TBK1 (Fig 6B and 6C). In reciprocal experiments, immunoprecipitation of FLAG-tagged TBK1 confirmed the interaction with ectopically expressed SNRNP200 in uninfected and in SeV-infected cells (Fig 6D). In addition, the kinase-dead TBK1 mutant (K38A) was still shown to interact with SNRNP200, demonstrating that this interaction is not dependent on TBK1 activity (Fig 6D). Finally, the interaction was further confirmed in A549 cells by the co-immunoprecipitation of endogenous SNRNP200 and TBK1 proteins (S12 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. SNRNP200 Sec63-1 domain interacts with TBK1.

(A) HEK 293T cells are transfected with FLAG-eYFP (control), FLAG-SNRNP200 or FLAG-SNRNP200 S1087L mutant expressing plasmids for 48 hours. Cell lysates are prepared following 16 hours of SeV infection and subjected to immunoprecipitation with anti-FLAG antibodies. Cell lysates and immune complexes are resolved by immunobloting analysis using anti-FLAG and anti-TBK1 antibodies. (B) Immunoprecipitation of FLAG-SNRNP200 C-terminal deletion mutants are performed and analyzed as indicated in (A). C) Immunoprecipitation of FLAG-SNRNP200 Sec63-1, HB-HLH or HLH-IG subdomains of Sec63-1 and Sec63-2 are performed and analyzed as indicated in (A). (D) Reciprocal immunoprecipitation of FLAG-eYFP (control), FLAG-TBK1 or FLAG-TBK1 K38A mutant following ectopic expression of SNRNP200 are performed as indicated in (A) and analyzed as indicated in (A). Cell lysates and immune complexes are resolved by immunobloting analysis using anti-FLAG and anti-SNRNP200 antibodies.

https://doi.org/10.1371/journal.ppat.1005772.g006

To further assess the interaction of SNRNP200 and TBK1, their intracellular localization was investigated by examining confocal fluorescence microscope images of HEK 293T cells and of HeLa cells in response to SeV infection (Fig 7 and S13 Fig). It was observed that FLAG-SNRNP200 (HEK 293T cells) and endogenous SNRNP200 (Hela cells) are localized to the nucleus and cytoplasm with a diffuse staining prior to stimulation. Upon viral infection, a subcellular fraction of SNRNP200 relocalizes with TBK1 into perinuclear cytoplasmic speckles (Fig 7A and S13B Fig). SNRNP200 and TBK1 colocalization can be easily observed in the 3D-stack and lateral view of infected cells (Fig 7B and 7C). Unlike WT SNRNP200, the staining of the FLAG-SNRNP200 S1087L mutant shows neither relocalization of the protein nor colocalization with TBK1 into these cytoplasmic speckles upon infection (Fig 7A and S13A Fig), correlating with its lack of RNA binding (Fig 5). Thus, the data suggest that viral RNA recognition by the Sec63-1 domain is responsible for the relocalization of SNRNP200 to perinuclear cytoplasmic speckles, and that SNRNP200 possibly functions as a novel adaptor via its interaction with TBK1 to promote IRF3 phosphorylation and antiviral innate responses.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Re-localization of SNRNP200 in perinuclear cytoplasmic speckles and colocalization with TBK1 in response to SeV infection.

(A) Confocal analysis of HEK 293T co-transfected with FLAG-SNRNP200 or FLAG-SNRNP200-S1087L and MYC-TBK1 using Hoechst, anti-FLAG and anti-MYC antibodies without virus infection or following a 16-hour infection with SeV. Imaging was done using a 63x/1.40 Oil DIC objective. Intensity analysis showed that 19/19 cells have cytoplasmic colocalization between SNRNP200 and TBK1 in SeV-infected cells. (B) Z-stack and lateral view of SNRNP200 and TBK1 in SeV-infected HEK 293T cells treated as indicated in (A). (C) Z-stacks reconstitution of a complete cell with colocalization plot and cut view, showing an exclusive cytoplasmic colocalization of SNRNP200 and TBK1.

https://doi.org/10.1371/journal.ppat.1005772.g007

SNRNP200 regulates innate immune responses of SeV-infected human MDM

The regulation of antiviral responses by SNRNP200 was further investigated in immune cells using primary cultures of purified human monocyte-derived macrophages (MDM). It was found that SeV infection leads to an increase in the immunodetection of SNRNP200 without affecting mRNA levels (Fig 8A), as observed in SeV-infected and in IFN-α-treated HEK 293T cells (S14 Fig). More importantly, the silencing of SNRNP200 in MDM decreases the induction of IFIH1 and IFIT1, and completely blocks IRF3 Ser386 phosphorylation within 3 hours post-infection (Fig 8B). Kinetic studies on IFN-β production have further demonstrated a complete blockage of its secretion at 3 hours post-infection (Fig 8C). In contrast to the unchanged TNFα mRNA levels, a decrease in IFNB1 mRNA was observed, which correlates with the reduced SNRNP200 mRNA at 1 hour post-infection in MDM (Fig 8D, 8E and 8F). Interestingly, the duration of SNRNP200 gene silencing is not sufficient to affect the steady-state levels of IRF3 protein, though it completely inhibits its phosphorylation. In addition, SNRNP200 KD increased SeV protein levels, as observed in HEK 293T cells (Figs 1C and 8B). These results confirm a regulatory role of SNRNP200 in the IRF3-mediated antiviral response in human macrophages.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. SNRNP200 KD restricts SeV-mediated antiviral response of human MDM.

(A) MDM are infected with SeV for 1, 3 or 5 hours. SNRNP200 and IFIT1 protein levels are resolved by immunobloting of cell lysates. (B) MDM are transfected with a pool of siRNA targeting SNRNP200 for 48 hours and infected with SeV. At 3 hours post-infection, cells are harvested and selected proteins (SNRNP200, IFIT1, IFIH1, IRF3, IFR3-386, SeV and actin) are resolved by immunobloting of cells lysates and compared to control cells treated with scrambled siRNA. (C) MDM are transfected with a pool of siRNA targeting SNRNP200 for 48 hours and infected with SeV for 1, 3 or 5 hours. Supernatants are harvested and IFN-β secretion levels are measured by ELISA and compared to control cells treated with scrambled siRNA (siNT). (D-F) MDM are transfected with a pool of siRNA targeting SNRNP200 (siSNRNP200) or scrambled siRNA (siNT) and infected with SeV for 1 hour. Cells are harvested and relative gene expression of IFNB1 (D), TNFα (E) and SNRNP200 (F) are measured by qRT-PCR and compared with scrambled control cells. mRNA RQ are normalized versus ACTIN and HPRT1 mRNA. P values <0.0001 (****) are indicated. Data are pooled results from two experiments of two biological replicates.

https://doi.org/10.1371/journal.ppat.1005772.g008

Impaired antiviral response of PBMCs from RP33 patients

RP is an inherited degenerative eye disease that causes severe vision impairment and blindness due to mutations in several core spliceosomal proteins. The antiviral responses of peripheral blood mononuclear cells (PBMCs) of RP33 patients that are genotyped for a particular monoallelic mutation in SNRNP200 were characterized: p.S1087L- c.3260C>T in the Sec63-1 domain and p.R681C c.2122G>A in the N-terminal helicase domain (see S1 Table for donor information). Interestingly, all RP33 patients showed a complete blockage of IFN-β cytokine production at 3 hours post-infection with a significant two-fold reduction in IFN-β secretion at 7 hours (Fig 9A and 9B). The decreased IFN-β production was corroborated by the reduction of virus-induced IRF3-dependent IFNB1 and IFIT1 mRNA. On the other hand, NF-κB-dependent TNF mRNA levels were not significantly affected (Fig 9C, 9D and 9E). The IRF3 mRNA levels determined by qRT-PCR showed no difference between healthy donors (HD) and RP33 patients (Fig 9F). Finally, a cytokine 41-plex assay performed on supernatants of infected PBMCs from HD and RP33 patients showed a significant decrease in IFN-α2, but showed similar cytokine/chemokine levels of RANTES, IL6, CXCL10, and IL1B (S15 Fig). The defective antiviral response of PBMCs from RP33 patients demonstrates that SNRNP200 plays a crucial role in regulating the IRF3-dependent pathway of IFNB1 production and does so without altering the NF-κB-dependent inflammatory pathway.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. PBMCs of RP33 patients bearing monoallelic point mutation in SNRNP200 show hindered antiviral innate immune response.

(A) PBMCs of RP33 patients are infected for 3, 5 and 7 hours with SeV. Supernatants are harvested and IFN-β secretion levels are measured by ELISA and compared to PBMCs of three healthy donors (HD). (B) Alternative representation of the 7-hours SeV infection of individual RP33 patients and HD as in (A), where the horizontal bar represents the mean of each group. P value <0.01(**) is indicated. (C-F) PBMCs of RP33 patients are infected for 1 hour with SeV. Cells are harvested and relative gene expression of IFIT1(C), IFNB1 (D), TNFα (E) and IRF3 (F) are measured by qRT-PCR and compared with PBMCs of HD. mRNA RQ are normalized versus ACTIN and HPRT1 mRNA. P values <0.0001 (****) are indicated.

https://doi.org/10.1371/journal.ppat.1005772.g009

Ethics statement

This study was approved by Institutional Review Board (IRB) of the participating institution (McGill Children’s Hospital, McGill University Health Centre and Centre Hospitalier Universitaire de l’Université de Montréal (CHUM)) and written informed consent was obtained from all participants before participation.

Expression vectors

SF3A1, NHP2L1 and PHF5A cDNAs were purchased from GE Dharmacon/Open Biosystems. Following PCR-amplification, PCR products were cloned into pcDNA3.1-Hygro-MCS using EcoRV/HindIII[6]. SNRNP200 was subcloned from the pBluescriptSK-hBrr2 obtained from R. Lührmann [50] into pcDNA3.1-Hygro(+) (Life Technologies) using NotI and XhoI restriction sites. SNRNP200 deletion mutants and S1087L point mutation were generated by PCR. All constructs were verified by Sanger sequencing and subsequent western blot analysis. If necessary, validated constructs where subcloned into pcDNA3.1-MCS-FLAG. pIFNB1-LUC and p2xNF-κB-LUC luciferase reporter constructs were previously described [51–53]. Generation of stable HEK 293T cells harboring the pIFNB1-LUC and pEF1α-LUC promoters was previously described [8].

Cells lines and culture

Human embryonic kidney HEK 293T (ATCC), human epithelial adenocarcinoma HELA (ATCC) and human hepatoma cell lines Huh7 / Huh7.5 (ATCC) were cultured in Dulbecco's modified Eagle's medium (DMEM, Wisent). Human lung adenocarcinoma epithelial A549 (ATCC) were cultured in Ham’s F-12 medium (Life Technologies). Both media were supplemented with 10% fetal bovine serum, 100 units/ml penicillin, 100 μg/ml streptomycin and 2 mM glutamine (all from Wisent) at 37°C in an atmosphere of 5% CO₂. Transient transfections were performed with lipofectamine 2000 (Life Technologies) according to manufacturer’s protocol. Peripheral blood mononuclear cells (PBMCs) were isolated from fresh heparinized peripheral blood samples by Ficoll-Histopaque gradient centrifugation (Sigma-Aldrich). Unfrozen PBMCs were washed twice in 10 ml of sterile RPMI 1640 and re-suspended in RPMI 1640 supplemented with 10% FBS. PBMCs were counted using a haemocytometer and counts were adjusted using trypan blue exclusion to plate 1x10⁶ PBMCs in 100 μl RPMI 1640 supplemented with 10% FBS in 96-well plate. For monocyte-derived macrophage (MDM), PBMCs were harvested has described above and monocyte were isolated using MACS Monocyte Isolation Kit II human (Miltenyi Biotec) as per manufacturer’s protocol before differentiation into MDM for five days in the presence of 10 ng/mL granulocyte-monocyte colony stimulating factor (M-CSF, R&D).

shRNA and siRNA gene silencing

shRNAs from MISSION TRC shRNA lentiviral library (Sigma-Aldrich) were used as followed: shRNA targeting SNRNP200 (TRCN0000051831), SF3A1 (TRCN0000006597), PHF5A (TRCN0000074878), NHP2L1 (TRCN0000074799), or shRNA non-target (NT). shRNA were transfected in combination with a standard packaging mix (1.5 μg pMDLg/pRRE, 1.5 μg pRSV-REV and 3 μg pVSVg) as previously described[54]. siRNA ON-TARGETplus SMARTpool, Human SNRNP200 and siRNA non-targeting #1 Human, ON-TARGETplus (GE Healthcare, Dharmacon), Santa Cruz HELIC2 siRNA (h) (sc-75243) were transfected with lipofectamine RNAi Max (Life Technologies) for 48 hours according to manufacturer’s protocol.

Firefly luminescence assay

For assays in 96-well plates, cells were seeded in white 96-well plates at a density of 5,000 HEK 293T pIFNB1_LUC and 1,250 293T pEF1α-LUC in 100 μl of complete phenol-red free DMEM containing 4 μg/ml polybrene. Infection with lentivirus encoding shRNA were carried out immediately after cell seeding at a MOI of 10 (except when specified otherwise) and incubated for three days at 37°C in an atmosphere of 5% CO₂. Cells were infected with 100 HAU/ml of SeV (Cantell Strain, Charles River Labs) for 16 hours before cell lysis and firefly luminescence reading in a 100 mM Tris acetate, 20 mM Mg acetate, 2 mM EGTA, 3.6 mM ATP, 1% Brij 58, 0.7% β-mercaptoethanol and 45 μg/ml luciferine pH 7.9 buffer. All infections were performed in an enclosed in a class II cabinet.

Assays used in western blot or qRT-PCR experiments where scaled up accordingly and carried out with the maternal HEK 293T or appropriate cell line.

Influenza A/Gaussia luminescence assay

For influenza infection, 3x10⁵ HEK 293T pIFNB1_LUC cells were seeded in 6-well plates. The next day, cells were transfected with an influenza vRNA reporter plasmid [55] and infected with 0.1 ul of purified influenza virus (A/PR/8/34, from Charles River). Five days later, 20 ul of cell supernatant was used to quantify the Gaussia luciferase using a Gaussia Luciferase Assay HTS (Nanolight Technology). Cell lysates were used to quantify the IFNB1 induction according to the Firefly luminescence assay described above.

HCV/Renilla luminescence assay

J6/JFH(p7-Rluc2a) virus production was conducted as previously described [56]. Briefly, HCV DNA template used for in vitro transcription was linearized using XbaI and subsequently transcribed using TranscriptAid T7 High Yield Transcription Kit according to manufacturer protocol (Life Technologies). The resulting HCV RNA was then electroporated into Huh7.5 and virus-containing culture medium was collected, filtered (0.45 μm) and kept at -80°C. For infection, 100μl of virus was added to 5,000 Huh7 cells that had been plated in 96-well white opaque plates one day before. Culture medium was replaced six hours later and Huh7 cells were transfected with pIFNB1-LUC (50 ng/well) the next day. Three days later Huh7 cells were washed twice with PBS, before Rluc and Fluc quantification using the Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer protocol.

Western immunoblot analysis

Cells were washed twice with ice-cold phosphate-buffered saline (PBS; Wisent), harvested and lysed in 10mM Tris-HCl, 100mM NaCl, 0.5% Triton X-100, pH7.6 with EDTA-free Protease Inhibitor Cocktail (Roche). Cell lysates were clarified by centrifugation at 13,000 g for 20 min at 4°C and subjected to sodium dodecyl sulfate-polyacrylamide gel (SDS-PAGE). Western blot analysis was performed using mouse anti-PHF5A (Abnova), anti-IRF3 (Santa Cruz), anti-TRAF3 (Santa Cruz), anti-RIG-I (Alexis Biochemicals), anti-ACTIN (Chemicon International), anti-TBK1 (Imgenex and Santa Cruz), anti-IKBKE (Santa Cruz), anti-TUBULIN (ICN), anti-GAPDH (RDI) and rabbit anti-SNRNP200 (Sigma-Aldrich), anti-SF3A1 (Santa Cruz), anti-RELA (Santa Cruz), anti-NHP2L1 (Abcam), anti-DDX3X (Bethyl), anti-DDX60 (Abcam), TRIF (Cell Signaling), anti-ISG56 (Novus Biologicals), anti-MDA5 (Alexis Biochemicals), anti-MAVS (Alexis Biochemicals), anti-IKBKE (eBioscience), STAT1 (ABCAM), STAT1 tyr701 (ABCAM), IFNAR1 (Santa Cruz) and anti-IRF3-P-ser386 (Abcam). HRP-conjugated secondary antibodies were from Bio-Rad. The chemiluminescence reaction was performed using the Western Lighting Chemiluminescence Reagent Plus (PerkinElmer).

Co-immunoprecipitation

For co-immunoprecipitation, FLAG-tagged protein expressing cells were harvested and lysed as described above. Resulting cell extracts were adjusted to 1 mg/ml and subjected to IP as follows: preclearing of the lysates was done by incubating lysates with 40 μl of a 50:50 slurry of immunoglobulin G-Sepharose (GE Healthcare) prepared in the lysis buffer with IgG beads for 1 hour. Pre-cleared lysate were immunoprecipitated by adding 20 μl of M2 anti-FLAG affinity gel (Sigma-Aldrich) prepared in TBS buffer (50 mM Tris-HCl, 150 mM NaCl, pH 7.4) overnight as described by the manufacturer. Immunoprecipitates were washed five times in lysis buffer. For interaction analysis, elution was performed using 250 ng/μl purified FLAG peptide for 45 min at 4°C (Sigma-Aldrich). Eluates were analyzed by western immunoblotting.

Microarray analysis

The microarray studies were performed with HEK 293T cells transduced with lentiviral-expressing shNT (control) or shSNRNP200 RNA targeting SNRNP200 gene for three days following 16 hours infection with SeV (100 U/ml) or 16 hours of treatment with a mixture of IFN-α from human leukocytes (400 U/ml; Sigma-Aldrich). A total of 10 μg of RNA was reverse transcribed using oligo(dT) 16–18 primers and SuperScript II Reverse Transcriptase (Life Technologies) according to the manufacturer's instructions. Following purification using QIAquick PCR Purification kit (Qiagen), up to 1 μg of purified cDNA was mixed with 5'-Cy3 labeled random nonamers (Trilink Biotechnology) and heated at 95°C for 10 minutes and transferred on ice for 10 minutes. Samples were mixed with 1 mM dNTP and 2 μl of 3’-5’ exo-Klenow fragment (New England Biolabs) and incubated at 37°C for 2 hours. The labeling reaction was stopped using 50 μM EDTA and the DNA precipitated using 0.5 M NaCl and 1 volume isopropanol, washed with 80% ethanol and resuspended in water. Hybridizations were carried out using the Human GE 4x44K v2 Microarrays (Agilent Technologies) containing probes targeting 27,958 Entrez Gene RNAs. Arrays were scanned at 5 μm resolution using a GenePix4000B scanner (Molecular Devices). Data from scanned images were extracted using GenePix 6.1 (Axon) and processed and normalized using ArrayPipe (v2.0). Processed data was used as input for linear modeling using Bioconductor's limma package, which estimates the fold-change between predefined groups by fitting a linear model and using an empirical Bayes method to moderate standard errors of the estimated log-fold changes in expression values from each probe set. P values from the resulting comparison are adjusted for multiple testing according to the method of Benjamini and Hochberg. Subsequently, gene enrichment analysis are conducted using DAVID [57,58], STRING [59,60] and Gene networks were constructed using GENEMANIA [60].

Biotin-RNA/Biotin-DNA pull-down

RNA pull-down assay was performed using Dynabeads M270 Streptavidin (Life Technologies). Dynabeads were incubated with biotin-labeled RNA (poly I:C (InvivoGen) and full-length Jc1 HCV) for 1 hours according to manufacturer’s protocol. Biotin-HCV RNA was obtained by subjecting linearized HCV DNA to T7 reverse transcription (TranscriptAid T7 High Yield, Life Technologies) and biotin-dUTP (Enzo Life Sciences). Saturated beads were added to whole 100 μg cell lysate and incubated, in a cold room, on a rotating wheel. Beads were washed three times and RNA-bound proteins were eluted after boiling in 0.1% SDS and analyzed by western blot. Poly (dA:dT) and Poly(dG:dC) were purchased from Sigma and labeled using Label IT Nucleic Acid Labeling Kit (Mirus Bio) and biotin-DNA pull-down assays were performed as described above.

RNA extraction and qRT-PCR assays

Total cellular RNA was extracted with the RNeasy Mini kit (Qiagen). Reverse transcription was performed on 1 μg total cellular RNA using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems). In order to amplify only the cDNA, primers were located in the splicing junction between two exons. PCR reactions were performed using 1.5 μl of cDNA samples (15 ng), 5 μl of the Fast TaqMan PCR Master Mix (Applied Biosystems), 10 pmol of each primer (IDT) and 5 pmol of the UPL probe (Roche) in a total volume of 10 μl. The ABI PRISM 7900HT Sequence Detection System (Applied Biosystems) was used to detect the amplification level and was programmed to an initial step of 3 minutes at 95°C, followed by 40 cycles of 5 seconds at 95°C, 30 seconds at 60°C and 1 second at 72°C. All reactions were run in duplicate on biological duplicate and the average values were used for quantification. ACTIN (β-actin) or GAPDH (Glyceraldehyde 3-phosphate dehydrogenase) and HPRT1 (hypoxanthine phosphoribosyltransferase 1) were used as endogenous controls. The relative quantification (RQ) of target genes was determined by using the ΔΔCt method with the Sequence Detection System (SDS) 2.2.2 software (Applied Biosystems).

Virus plaque assays

Plaque assays were conducted in VERO cells and MDCK.2 cells (ATCC) using a method described elsewhere [61]. Briefly, supernatants were harvested from infected cells and used to inoculate in serial dilutions VERO (SeV) and MDCK.2 cells (FLUA) for 45 minutes and 1 hour, respectively. After infection, cells were wash with PBS and an overlay of 0,6% agarose was superimposed to 2X DMEM medium. At 72 hours post-infection, cells were colored with crystal violet, washed with PBS and colonies (lysed cells) were counted to compute viral titers.

ELISA assays

ELISA assays were carried out with 50 μl of cell culture supernatants using the VeriKine Human Interferon Beta Elisa Kit (PBL Assay Science) according to the manufacturer’s protocol. Samples were run as technical duplicates on biological triplicates.

Immunofluorescence analysis

HEK 293T were seeded in cover slip-containing 24-well plates and co-transfected with FLAG-SNRNP200 WT or S1087L mutant and MYC-TBK1 24 hours later. The following day, cells were infected or not with SeV for 16 hours before being washed twice with PBS, fixed with 4% paraformaldehyde-containing PBS during 20 minutes at room temperature and then permeabilized in 0.2% Triton X-100/PBS during 15 minutes. Blocking was made in PBS with 10% normal goat serum, 5% bovine serum albumin (BSA) and 0.02% sodium azide during 45 minutes at room temperature. Following three rapid washes, cells were labelled with rabbit anti-FLAG (Sigma-Aldrich) and mouse anti-MYC (Santa Cruz) primary antibodies diluted in 5% BSA/0.02% sodium azide/PBS for 2 hours. Slides were washed three times in PBS and then labeled with anti-rabbit or anti-mouse AlexaFluor 488, 594 or 647 secondary antibodies (Life Technologies) diluted in 5% BSA/0.02% sodium azide/PBS for 1 hour. Cells were extensively washed and incubated with Prolong Gold with DAPI (Life Technologies). Alternatively, nuclei were labeled with Syox Green (Life Technologies). Labelled cells were then examined by laser scanning microscopy using a TCS SP5 (Leica).