Virus‐induced autophagic degradation of STAT2 as a mechanism for interferon signaling blockade

Abstract

The mammalian interferon (IFN) signaling pathway is a primary component of the innate antiviral response, and viral pathogens have evolved multiple mechanisms to antagonize this pathway and to facilitate infection. Bluetongue virus (BTV), an orbivirus of the Reoviridae family, is transmitted by midges to ruminants and causes a disease that produces important economic losses and restriction to animal trade and is of compulsory notification to the World Organization for Animal Health (OIE). Here, we show that BTV interferes with IFN‐I and IFN‐II responses in two ways, by blocking STAT1 phosphorylation and by degrading STAT2. BTV‐NS3 protein, which is involved in virion egress, interacts with STAT2, and induces its degradation by an autophagy‐dependent mechanism. This STAT2 degradative process requires the recruitment of an E3‐Ub‐ligase to NS3 as well as NS3 K63 polyubiquitination. Taken together, our study identifies a new mechanism by which a virus degrades STAT2 for IFN signaling blockade, highlighting the diversity of mechanisms employed by viruses to subvert the IFN response.

Introduction

Interferons (IFNs) are early critical mediators of the innate antiviral host response 1. IFN production is initiated upon recognition of specific pathogen‐associated molecular patterns (PAMPs) by host pattern recognition receptors (PRRs) 2. PRR sensing triggers the activation of a signaling cascade that leads to type I IFN (IFN‐I) secretion 3. New synthesized IFNs bind to their receptors activating phosphorylation of the Janus kinases JAK1 and TYK2, and consequently the phosphorylation of the signal transducer and activators of transcription 1 (STAT1) and 2 (STAT2) 4, 5. Phosphorylated STAT1 and STAT2 heterodimerize and bind to interferon regulatory factor 9 (IRF9), forming the IFN‐stimulated gene factor 3 (ISGF3). This complex translocates to the nucleus and activates the interferon‐stimulated response element (ISRE) inducing a broad range of IFN‐stimulated genes (ISGs) 6, 7. Type II IFN (IFN‐γ) is produced by activated immune cells; it binds to IFN‐γ receptors leading to STAT1 phosphorylation, which then homodimerizes to form the IFN‐γ factor (GAF). GAF translocates to the nucleus and binds γ‐activated sequence (GAS) elements that induce ISGs 8. Regulation of the IFN pathway is tightly controlled and involves post‐translational mechanisms, such as phosphorylation, acetylation, and ubiquitination 9, 10.

The ubiquitin (Ub) system was first described as a regulator of protein expression levels that labeled misfolded proteins for their removal 11, 12. Ubiquitination can also be used to target proteins to specific cellular locations and/or allow new protein–protein interactions, thus regulating signaling events. Ub is a 76‐amino acid polypeptide with 7 lysine residues, which can form a bond with the C‐terminal glycine residue of the following Ub giving rise to a poly‐Ub chain 13. This poly‐Ub chain is often bound covalently to target proteins through specific lysine residues and directs polyubiquitinated proteins for proteasome (K48‐linked) or lysosome (K63‐linked) degradation 14. The fundamental contributors to this cascade are the E1‐activating, E2‐conjugating, and E3 protein ligase enzymes that attach Ub to the substrate. The host‐Ub system is one of the several cellular components hijacked by viruses to evade the host immune response (reviewed in 15).

The large group of arthropod‐borne viruses (arboviruses) uses arthropod vectors for transmission between vertebrate hosts and includes the agents responsible for most emergent and re‐emergent viral infections in the past decade (reviewed in 16, 17). Among them, bluetongue virus (BTV) is the causative agent of one of the most relevant diseases in ruminants. BTV belongs to the Orbivirus genus of the Reoviridae family and is transmitted by Culicoides biting midges 18. The genome is composed of 10 segments of dsRNA enclosed by a complex capsid structure and encoding for 7 structural and 5 non‐structural (NS) proteins 19, 20. BTV infection in sheep results in acute disease with high morbidity and mortality 21, while in goats, cattle, and wild ruminants, the infection is asymptomatic although a prolonged viremia is developed 22, 23. This differential outcome of BTV infection in ruminants may be determined by the interplay between the host immune response and the virus. It is therefore important to identify the mechanisms that drive disease development.

Members of the Reoviridae family, including rotavirus, reovirus, and BTV, although susceptible to IFN, have developed mechanisms to avoid or suppress IFN effects 24, 25, 26, 27. BTV is an inducer of type I IFN responses through RIG‐I and MDA5 helicase sensors in non‐hematopoietic cells 28, while in plasmacytoid dendritic cells, it is recognized through a MyD88‐dependent TLR‐independent pathway 29. BTV is nonetheless capable of modulating the IFN response. Among BTV proteins, NS3 and NS4 have also been identified as antagonist of the IFN‐I system 30, 31.

The present study explores in more detail the disruption of IFN signaling by BTV. We demonstrate that BTV infection inhibits type I IFN and IFN‐γ signaling. This inhibition is mediated by NS3 that reduces STAT1 phosphorylation and degrades STAT2. More importantly, we show that NS3 uses ubiquitination to target STAT2 for degradation through an autophagy–lysosome‐dependent mechanism.

Results

BTV infection suppresses type I and II IFN signaling and blocks IFN‐β induction

BTV infection impairs ISRE promoter activation after IFN‐I stimulation in A549 cells 32. Using a dual‐luciferase reporter gene assay with 293T cells, we confirmed that BTV infection also interfered with this signaling pathway in our model system (Fig 1A). The effect of BTV infection on IFN‐γ‐stimulated gene expression was also studied using an IFN‐γ‐triggered GAS activation luciferase reporter assay in Vero cells (Fig 1B). GAS promoter activation by IFN‐γ was significantly reduced at 16 hpi, indicating that BTV inhibits IFN‐γ signaling. BTV infection also interferes with IFN‐I induction in our system 30 (Fig EV1A). These inhibitory effects on IFN signaling and induction pathways are triggered before viral shutoff occurs as assessed by control CMV promoter activity (Fig EV1B and C).

Figure 1. BTV impairs IFN‐I and IFN‐II signaling.

• A, B
  (A) ISRE‐luciferase reporter assay in 293T cells or (B) GAS‐luciferase reporter assay in Vero cells infected with BTV‐8 or mock‐infected (control) and treated with (A) 1,000 U/ml of universal IFN, or (B) 5 ng/ml of IFN‐γ. Luminescence fold induction was measured at different time points p.i. Mean luciferase activity fold inductions in control or BTV‐infected in three experiments are presented. *P < 0.05; ***P < 0.001, one‐way ANOVA.
• C, D
  qRT–PCR analysis of ISG induction in mock‐ or BTV‐8‐infected (C) 293T or (D) Vero cells after (C) type I or (D) type II IFN treatment for 8 h. Mean mRNA fold inductions for three experiments are presented. *P < 0.05; ***P < 0.001, one‐way ANOVA.

Figure EV1. BTV‐8 infection interferes with IFN‐I induction, starts producing viral shutoff after 24 h, and also degrades STAT2 in sheep cells.

• A
  Vero cells were co‐transfected with pRL‐null‐Renilla luciferase plasmid and IFN‐β promoter‐driven luciferase plasmid (p125‐luc). 24 h post‐transfection, cells were infected with BTV‐8 at MOI 1 PFU/cell or mock‐infected (control). For IFN‐I induction, cells were then infected with SeV and luminescence was measured at different time points post‐BTV infection. Mean fold inductions of luciferase activity in three experiments are presented. BTV‐8 infection interferes with IFN‐β promoter activity. ***P < 0.001, one‐way ANOVA.
• B, C
  (B) 293T cells or (C) Vero cells were co‐transfected with pGL3‐luciferase plasmid and pCMV‐Renilla plasmid. 24 h post‐transfection, cells were infected with BTV‐8 at MOI 1 or mock‐infected (control). Cells were lysed at the indicated time points (6, 16, or 24 h) and luminescence measured. BTV‐8 infection produces viral shutoff after 24 h, indicating that it interferes with IFN‐I induction and IFN‐I and IFN‐II signaling before shutoff occurs. Mean % CMV inductions in three experiments are shown. **P < 0.01; ***P < 0.001, one‐way ANOVA.
• D
  Immunoblot of SSC cells infected with BTV‐8 for 16 h at indicated MOI and probed for STAT1, STAT2, and GAPDH expression. BTV‐8 infection reduces STAT2 expression in cells from the natural host.

To functionally confirm these findings, ISG induction was assessed by qRT–PCR in BTV‐8‐infected cells (Fig 1C and D). BTV‐infected 293T cells expressed 45‐fold lower MxA, 18‐fold lower ISG15, and 17‐fold lower OAS than mock‐infected after IFN‐I stimulus, further confirming the inhibition of IFN‐I signaling by BTV (Fig 1C). The expression of IRFI and IP10 as representative IFN‐γ‐induced genes, in BTV‐infected Vero cells (Fig 1D), showed that IRFI was expressed eightfold lower and IP‐10 sevenfold lower in BTV‐infected cells than in mock‐infected cells after IFN‐γ stimulus, which confirmed IFN‐II signaling inhibition by BTV. Thus, BTV specifically blocks IFN‐I induction and IFN‐I and IFN‐II signaling pathways before viral shutoff occurs.

BTV inhibits STAT1 and STAT2 phosphorylation and degrades STAT2 through lysosome/autophagy pathway

Since BTV inhibits IFN‐I and IFN‐II signaling pathways, the ability of the virus to affect STAT1/2 levels and phosphorylation after IFN‐I treatment was studied. Vero cells were infected at different MOI, treated with IFN‐I, and cell lysates analyzed by Western blotting (Fig 2A). Flow cytometry staining for BTV‐VP7 (Fig 2B) showed increased VP7 expression with increasing MOI, which confirmed BTV‐8 cell infection. Western blots showed that IFN‐I induced STAT1 and STAT2 phosphorylation in mock‐infected cells (Fig 2A). By contrast, STAT1 phosphorylation was inhibited in BTV‐infected cells after IFN‐I stimulation but total STAT1 levels remain constant. Flow cytometry analyses confirmed this significant reduction of pSTAT1 in BTV‐infected cells (Fig 2C). Immunofluorescence assays in IFN‐treated mock‐ or BTV‐infected cells (Fig 2D) showed translocated pSTAT1 expression in the nucleus of mock‐infected cells but no pSTAT1 expression in BTV‐infected cells. BTV therefore inhibits STAT1 phosphorylation to impair IFN‐I and IFN‐II signaling.

Figure 2. BTV prevents STAT1 phosphorylation and nuclear translocation and reduces STAT2 levels.

• A
  Immunoblot of Vero cells infected with BTV‐8 for 16 h and probed for total and phosphorylated STAT1/2 levels.
• B
  Vero cells infected with BTV‐8 were fixed at 16 hpi and stained with anti‐BTV‐VP7 antibody.
• C, D
  STAT1 phosphorylation in IFN‐I‐treated Vero cells infected with BTV‐8 detected by (C) flow cytometry and (D) confocal microscopy (Scale bar = 30 μm). Means ± SD of P‐STAT1 mean fluorescence intensity (MFI) in three experiments are shown. *P < 0.05, one‐way ANOVA (mock + IFN versus infection).
• E, F
  STAT2 expression in BTV‐8‐infected Vero cells detected by (E) flow cytometry and (F) confocal microscopy (Scale bar = 30 μm). Means ± SD of STAT2 MFI in three experiments are shown. **P < 0.01; ***P < 0.001, one‐way ANOVA (mock versus infection).

Source data are available online for this figure.

We also examined BTV effects on STAT2 expression and phosphorylation. BTV infection not only reduced STAT2 phosphorylation but also decreased STAT2 expression (Fig 2A). BTV‐induced STAT2 degradation was also detected in an ovine cell line, indicating that this phenomenon occurs in cells derived from a natural host (Fig EV1D). STAT2 degradation by BTV was confirmed by flow cytometry (Fig 2E) and immunofluorescence (Fig 2F).

To determine which cellular protein degradation pathway BTV employs to reduce STAT2 expression, BTV‐infected Vero cells were treated with proteasome inhibitors MG132 or lactacystin, lysosome acidification inhibitor chloroquine (CQ) 33, autophagosome formation inhibitor 3‐MA 34, or autophagic maturation vacuole inhibitor bafilomycin A1 (BAF‐A1) 35 (Fig 3A–C). Flow cytometry showed that BTV infection was not affected by drug treatment in these experiments (Fig 3A). The reduction in LC3‐cleaved form in BTV‐8‐infected cells confirmed that CQ, 3‐MA, and BAF‐A1 inhibited autophagy (Fig 3B and C). STAT2 expression in infected cells was only rescued with CQ, 3‐MA (Fig 3B), or BAF‐A1 (Fig 3C) treatment indicating that lysosomal activity and autophagosome formation and maturation, but not proteasome activity, were necessary for BTV‐induced STAT2 degradation. Immunofluorescence also showed STAT2 colocalized with the autophagosome marker SQSTM‐1/p62 36, confirming the involvement of this pathway in STAT2 degradation in BTV‐infected cells (Fig 3D). Taken together, these results strongly indicate that BTV degrades STAT2 through autophagy/lysosome pathway as a mean to evade innate immune responses.

Figure 3. BTV infection induces STAT2 degradation through the lysosome/autophagy pathway.

• A–C
  Vero cells were infected at MOI 1 (or mock‐infected) and treated with proteasome inhibitors MG132 (MG 20 μM), lactacystin (LC 20 μM), chloroquine (CQ 50 μM), 3‐MA (5 mM), or bafilomycin A1 (BAF‐A1 0.2 μM) for the final 6 h. (A) Representative flow cytometry histogram of BTV‐VP7 expression in inhibitor‐treated cells. (B and C) Immunoblots probed for STAT1, STAT2, LC3, and GAPDH expression.
• D
  Immunofluorescence of BTV‐8‐ or mock‐infected Vero cells (treated for the last 6 h with BAF‐A1) stained for BTV proteins, STAT2 and p62. Inset shows detail of STAT2 and p62 colocalization. Nuclei were counterstained with DAPI. Arrowheads in inset indicate p62 and STAT2 colocalization. Scale bar = 20 μm. Inset scale bar = 3 μm.

Expression of BTV‐NS3 and BTV‐NS4 proteins suppresses IFN‐I and IFN‐II signaling

We next set out to identify which BTV proteins are involved in the inhibition of IFN‐I and IFN‐II signaling. We examined the effects of VP7, VP6, NS4, NS3, NS2, or NS1 BTV protein expression on these pathways in a dual‐luciferase reporter gene assays for ISRE (Fig 4A) or GAS (Fig 4B) activity after IFN‐I or IFN‐γ treatment, respectively. The expression of FLAG‐tagged BTV proteins driven by the CMV promoter in 293T or Vero‐transfected cells was confirmed by Western blot analysis and immunofluorescence (Fig EV2A and B). Our data showed that cells expressing NS4 or NS3 significantly reduced luciferase activity driven by the ISRE or the GAS promoter compared to the control or cells expressing VP7, VP6, NS1, or NS2 (Fig 4A and B). Unaltered CMV‐driven Renilla expression excluded the possibility that the reduction in ISRE‐, or GAS promoter‐driven luciferase activity was caused by a general inhibition of gene expression by NS3 or NS4 (Fig EV2C–E). In addition, expression of increasing amounts of NS3 or NS4 proteins (Fig EV2G) inhibited the tested reporter activity in a dose‐dependent manner (Fig EV2G and F), while CMV promoter‐driven Renilla expression was not affected even at the higher dose of NS3 or NS4 (Fig EV2H and I). Finally, we confirmed the potential of BTV‐NS3 and NS4 proteins to modulate IFN‐I promoter activation using SeV infection as stimulus in p125‐Luc and pRL‐null co‐transfected cells (Fig EV3A and B). Taken together, these data indicate that NS3 and NS4 proteins contribute to the viral antagonism of IFN‐I/IFN‐γ signaling and IFN‐I production.

Figure 4. BTV‐NS3 and NS4 contribute to IFN signaling antagonism.

• A, B
  (A) ISRE‐luciferase reporter assays in 293T cells stimulated with IFN‐U and (B) GAS‐luciferase reporter assays in Vero cells stimulated with IFN‐γ co‐transfected with BTV‐VP6, BTV‐VP7, BTV‐NS1, BTV‐NS2, BTV‐NS3, or BTV‐NS4‐expressing plasmids or empty expression plasmid (control). Representative mean fold inductions of luciferase activity of three independent experiments are shown. ***P < 0.001, one‐way ANOVA.
• C
  Immunoblot of 293T cells transfected with BTV‐NS2, BTV‐NS3, BTV‐NS4 (FLAG) plasmids and NiV V (HA) and WNV‐NS5 (HA) used as control for STAT1 phosphorylation inhibition and DENV‐NS5 (HA) used as control for STAT2 degradation.
• D
  Three independent immunoblot experiments as in (C) were quantified for STAT1, P‐STAT1, STAT2, and P‐STAT2. Mean ± SD normalized to GAPDH and non‐transfected cells. **P < 0.01; ***P < 0.001, one‐way ANOVA (mock versus transfected cells).
• E
  Immunoprecipitation (IP) for STAT1/2 binding using 293T transfected with BTV‐NS2, BTV‐NS3, or BTV‐NS4 plasmids. NiV V used as positive control for STAT1 and STAT2 binding, and DENV‐NS5 used as positive control for STAT2 binding. Cells were treated (or left untreated) with 1,000 U/ml universal IFN for the final 16 h. Immunoblots of WCE and IP were probed (IB) for STAT1, STAT2, HA (NiV V, DENV NS5), FLAG (NS2, NS3, NS4), and GAPDH.
• F, G
  (F) STAT1 phosphorylation and (G) STAT2 nuclear translocation in BTV‐NS3‐transfected Vero cells detected by immunofluorescence and confocal microscopy after 30‐min IFN treatment (Scale bar = 20 μm). Arrowheads in merge images indicate examples of transfected cells.

Figure EV2. BTV protein expression and BTV‐NS3 and BTV‐NS4 effects on IFN signaling.

• A
  293T cells were transfected with expression vectors for FLAG‐tagged BTV‐NS1, BTV‐NS2, BTV‐NS3, BTV‐NS4, BTV‐VP6, or BTV‐VP7 proteins, or empty vector (pIRES), and lysates were separated by SDS–PAGE (8‐15% polyacrylamide). BTV proteins were visualized by immunoblotting with antibodies to the FLAG epitope tag. GAPDH was used as loading control. Western blot analysis shows the expression of BTV proteins, with the predicted molecular weight.
• B
  293T cells were transfected with OFP‐containing expression vectors containing FLAG‐tagged BTV‐NS1, BTV‐NS2, BTV‐NS3, BTV‐NS4, BTV‐VP6, or BTV‐VP7. Cells were fixed, permeabilized, and stained with antibodies to FLAG, as described in Materials and Methods. OFP was used as control for transfected cells, and nuclear DNA was stained with DAPI. Images were obtained using confocal microscopy (Scale bar = 10 μm). VP7 and VP6 were expressed into the cytoplasm (Roy et al [37]); NS4 localized in both cytoplasm and nuclei, and accumulated into the nucleoli (Ratinier et al [20]); NS2 was found in the cytoplasm, where it formed inclusion bodies (Kar et al [38]); NS1 localized into cytoplasmic tubules (Owens et al [39]), and NS3 localized to the plasma membrane and Golgi apparatus (Han & Harty [40]).
• C, D
  (C) 293T cells or (D) Vero cells were co‐transfected with pGL3‐luciferase plasmid and pCMV‐Renilla plasmid and either with BTV‐VP6, BTV‐VP7, BTV‐NS1, BTV‐NS2, BTV‐NS3, or BTV‐NS4‐expressing plasmids or empty expression plasmid (control). 24 h post‐transfection, cells were lysed and luminescence was measured. Mean % CMV stimulation in three experiments is shown.
• E
  Immunoblots of 293T cells transfected with 50, 300, or 600 ng expression plasmids and probed for FLAG‐tag and GAPDH. Arrows indicate the expression of NS3 or NS4 at the predicted molecular weight.
• F, G
  (F) ISRE or (G) GAS reporter assays performed in (F) 293T or (G) Vero cells transfected with increasing amounts of NS3‐ and NS4‐expressing plasmids (50, 300, and 600 ng) and measured 16 h after stimulation. Mean luciferase activity fold inductions in 3 experiments are shown. **P < 0.01; ***P < 0.001, one‐way ANOVA.
• H, I
  (H) 293T cells or (I) Vero cells were co‐transfected with pGL3‐luciferase plasmid and pCMV‐Renilla plasmid and with increasing amounts of NS3‐ or NS4‐expressing plasmids (50, 300, and 600 ng). At 24 h post‐transfection, cells were lysed and luminescence was measured. Mean % CMV stimulation in three experiments is shown. *P < 0.05, one‐way ANOVA.

Figure EV3. BTV‐NS3 and BTV‐NS4 inhibit IFN‐I induction in a dose‐dependent manner.

• A
  Vero cells were co‐transfected with BTV‐VP6, BTV‐VP7, BTV‐NS1, BTV‐NS2, BTV‐NS3, or BTV‐NS4‐expressing plasmids or the empty expression plasmid (control), a plasmid encoding a 125‐driven luciferase reporter and a pRL‐null‐Renilla. At 24 h post‐transfection, cells were infected with SeV for IFN‐I induction and luminescence measured 16 h later. Mean luciferase activity fold inductions in three experiments are presented. ***P < 0.001, one‐way ANOVA.
• B
  125 reporter assays performed in Vero cells transfected with increasing amounts of NS3‐ and NS4‐expressing plasmids (50, 300, and 600 ng) and measured 16 h after stimulation by SeV infection. Mean luciferase activity fold inductions in three experiments are presented. ***P < 0.001, one‐way ANOVA.

NS3 interferes with STAT1 phosphorylation and is responsible for STAT2 degradation

In order to investigate the mechanisms by which NS3 or NS4 interferes with IFN‐I/IFN‐γ signaling, we monitored STAT1/2 expression levels and phosphorylation before and after IFN treatment in transfected 293T cells. In these experiments, NS5 protein from West Nile virus (WNV) that inhibits STAT1 phosphorylation after IFN treatment 41; NS5 protein from Dengue virus (DENV) that binds to STAT2 independently of IFN treatment and promotes its proteasomal degradation 42; and V protein from Nipah virus (NiV) that binds STAT1 and STAT2 and prevents STAT1 and STAT2 nuclear translocation in response to IFN 43, 44, 45, were used as controls. Western blotting experiments showed that transient expression of BTV‐NS3 protein inhibited STAT1 phosphorylation after IFN‐I treatment (Fig 4C and D), consistent with the mechanism employed by BTV to prevent ISRE/GAS activation after IFN‐I/IFN‐II treatment. BTV‐NS3 did not however appear to directly interact with STAT1 to inhibit phosphorylation, as STAT1 was not detected in NS3 pulled down fractions (Fig 4E). NS3 inhibition of STAT1 phosphorylation was more pronounced in Vero cells, which are defective in the IFN‐β gene (Fig EV4A), than in 293T cells which retain the IFN‐I‐producing capacity. Immunofluorescence assays in cells expressing BTV‐NS3 protein showed inhibition of STAT1 phosphorylation and its subsequent translocation to the nucleus (Fig 4F). These data strongly suggest that BTV‐NS3 protein inhibits STAT1 phosphorylation.

Figure EV4. BTV‐NS3 antagonizes IFN signaling in 293T, Vero, and SSC cells.

• A
  Immunoblot of Vero cells transfected with BTV‐NS3 plasmid and sorted by flow cytometry and probed for STAT1/2 phosphorylation, STAT1/2 total levels, BTV‐NS3 (FLAG), and GAPDH.
• B
  Immunoblot of 293T cells transfected with increasing amounts of DENV‐NS5 or BTV‐NS3 plasmids (0.5 and 1 μg) showing STAT1, STAT2, DENV‐NS5 (HA), BTV‐NS3 (FLAG), and GAPDH expression. Densitometry result presenting mean ± SD STAT2 band intensity normalized to GAPDH and non‐transfected cells in three independent experiments. **P < 0.01; ***P < 0.001, one‐way ANOVA (mock versus transfected cells).
• C
  Immunoblot of SSC cells transfected with BTV‐NS3 plasmid, sorted by flow cytometry, and probed for STAT1, STAT2, NS3 (FLAG), and GAPDH expression.

Western blot experiments showed that BTV‐NS3 expression in 293T cells decreased STAT2 total levels independently of IFN treatment, suggesting that this protein could play a role in STAT2 degradation (Fig 4C and D). Similarly to NiV V protein and DENV‐NS5 protein, BTV‐NS3 protein interacted with STAT2 independently of IFN treatment in immunoprecipitation (IP) assay (Fig 4E). None of the other BTV proteins were found to interact either with STAT1 or STAT2. The expression of NS3 in 293T cells reduced STAT2 levels in a dose‐dependent manner (Fig EV4B). Western blot analysis also showed that NS3 expression in ovine cells reduced STAT2 levels, indicating that NS3 induced STAT2 degradation in the natural host (Fig EV4C). Immunofluorescence assays confirmed STAT2 degradation by NS3 (Fig 4G), and this was independent of IFN‐I stimulation (Appendix Fig S1). These data imply that BTV‐NS3 protein interacts specifically with STAT2 and induces its degradation.

BTV‐NS3 mediates STAT2 degradation through a lysosome/autophagy pathway

To determine whether BTV‐NS3 was implicated in the autophagic/lysosomal degradation of STAT2 detected during BTV infections, BTV‐NS3‐transfected 293T cells were treated with proteasome, lysosome, or autophagy inhibitors and STAT2 expression assessed in the resulting cell lysates by Western blotting (Fig 5A–D). Treatment with proteasome inhibitors MG‐132 or lactacystin did not rescue STAT2 expression in BTV‐NS3 transfected cells, but restored STAT2 levels in DENV‐NS5‐transfected cells that degrade STAT2 through the proteasome 42 (Fig 5E). Inhibition of lysosome acidification with chloroquine (Fig 5F) or NH₄Cl (Fig 5G) 33 restored STAT2 expression in a dose‐dependent manner. Similarly, STAT2 expression was rescued in a dose‐dependent manner with the autophagosome formation inhibitor 3‐MA 34. To further confirm the involvement of autophagic pathways in NS3‐mediated STAT2 degradation, autophagy key protein Atg7 and Beclin‐1 46 expression was silenced with specific siRNA. Partial inhibition of Atg7 and Beclin‐1 expression was achieved by siRNA transfection in control cells (empty plasmid) and in NS3‐transfected cells. This reduction was sufficient to rescue STAT2 expression in NS3 expressing cells. These data strongly indicate that BTV‐NS3 is the viral protein that mediates STAT2 autophagic/lysosomal degradation during BTV infections.

Figure 5. BTV‐NS3 mediates STAT2 degradation through an autophagy/lysosome pathway.

• A
  Immunoblots of BTV‐NS3‐transfected 293T cells treated with proteasome inhibitors MG or LC. Membranes were probed for STAT1, STAT2, BTV‐NS3 (FLAG), and GAPDH.
• B–D
  Immunoblots of BTV‐NS3‐transfected 293T cells treated with increasing concentration of lysosome acidification inhibitors (B) CQ or (C) NH₄Cl or with increasing concentration of the (D) autophagosome formation inhibitor 3‐MA, and probed (IB) for STAT1, STAT2, BTV‐NS3 (FLAG), and GAPDH expression.
• E, F
  293T cells were co‐transfected with NS3 expression plasmid (or empty plasmid as control) and siRNA for (E) Atg7 or (F) Beclin‐1 (or siRNA control or no RNA as controls). Lysates were obtained from FACS‐sorted transfected cells. Immunoblots of sorted cells were probed for STAT1; STAT2; Atg7/Beclin‐1; FLAG‐tagged NS3; LC3; and GAPDH.
• G
  Representative flow cytometry dot plots of Vero cells infected with rgBTV‐8 or BTV‐8 ΔNS3 and stained for BTV‐VP7. Gating on VP7⁺ cells was used for sorting to obtain lysates from infected cells.
• H
  Immunoblot of FACS‐sorted Vero cells infected with rgBTV‐8 or BTV‐8 ΔNS3 and treated with IFN‐U for 30 min prior to fixation, staining, and sorting. Membranes were probed for STAT1, P‐STAT1, STAT2, and GAPDH.
• I, J
  Immunofluorescence confocal images of Vero cells infected with rgBTV‐8, BTV‐8 ΔNS3, or mock‐infected and treated with IFN‐U for 30 min prior to fixation and staining. Cells were stained for BTV‐VP7 and (I) P‐STAT1 or (J) STAT2. Nuclei were counterstained with DAPI (Scale bar = 30 μm). (I) Arrowhead in merge image indicates partial inhibition of STAT1 phosphorylation in BTV‐8 ΔNS3‐infected cells. (J) Arrowhead in merge image shows STAT2 translocation to nucleus in BTV‐8 ΔNS3‐infected cells.

NS3‐defective BTV partially inhibits STAT1 phosphorylation and cannot degrade STAT2

To confirm NS3 role in STAT1 phosphorylation inhibition and STAT2 degradation during an infection, Vero cells were infected with reverse‐engineered BTV‐8 (rgBTV‐8) or reverse‐engineered BTV‐8 without the NS3 segment (BTV‐8 ΔNS3). Flow cytometry staining for VP7 confirmed infection (Fig 5G). To assess STAT1 phosphorylation and STAT2 expression, rgBTV‐8‐ or BTV‐8 ΔNS3‐infected cells were treated with IFN‐U prior to fixation, staining, and sorting for VP7⁺ cells. These sorted cells were lysed and used for Western blot analysis (Fig 5H). Infection with rgBTV‐8 inhibited STAT1 phosphorylation, whereas infection with BTV‐8 ΔNS3 only partially inhibited this phosphorylation (Fig 5H). Immunofluorescence studies showed that STAT1 phosphorylation and nuclear translocation were completely inhibited in rgBTV‐8‐infected cells, but in BTV‐8 ΔNS3 infections, some infected cells still responded to the IFN stimulus, while in others, STAT1 phosphorylation and nuclear translocation were inhibited (arrowheads in Fig 5I). This indicated that NS3 contributes to STAT1 phosphorylation inhibition during BTV infections but is not solely responsible for this effect. STAT2 expression was also assessed by immunoblot in rgBTV‐8‐ and BTV‐8 ΔNS3‐infected cells (Fig 5H). STAT2 degradation was detected in rgBTV‐8‐infected cells but not in BTV‐8 ΔNS3‐infected cells. In immunofluorescence studies of Vero cells stimulated with IFN‐U (Fig 5J), rgBTV‐8 infection inhibited STAT2 translocation to the nucleus, whereas BTV‐8 ΔNS3‐infected cells could still translocate STAT2 to the nucleus (arrowhead in Fig 5J). These data confirm in an infection model that NS3 participates in the IFN antagonism by partially inhibiting STAT1 phosphorylation and by degrading STAT2.

BTV‐NS3 protein is ubiquitinated on lysine 13 and 15 and recruits an E3 ligase for STAT2 degradation

Viruses can use host ubiquitin (Ub) ligases to target host immune proteins for degradation and thus evade immunity (reviewed in 47). Ubiquitin is a globular protein that is covalently attached to target proteins through a modification known as ubiquitination (reviewed in 48). Given NS3 banding pattern observed in Western blotting and the presence of two putative ubiquitination sites on NS3 49, we tested by Co‐IP whether NS3 bound ubiquitin. To this end, 293T cells were co‐transfected with increasing amounts of HA‐tag Ub‐expressing plasmid and a FLAG‐tag NS3 expression plasmid; and Ub presence was assessed in NS3 pull downs by immune blotting (IB) (Fig 6A). Ub could be detected in NS3‐FLAG‐pulled down lysates indicating that Ub specifically binds to NS3. Reverse IP assays in which increasing amounts of BTV‐NS3 were pulled down by anti‐HA (Ub) antibodies confirmed this result (Fig EV5A). Importantly, poly‐Ub chains did not co‐IP with NS4 protein (Fig EV5B), indicating that the ubiquitination is specific to NS3 protein. Supporting the IP data, immunofluorescence in 293T cells transfected with plasmids expressing FLAG‐NS3 and HA‐Ub showed colocalization of NS3 and Ub fluorescent signals (Figs 6B and C, and EV5C).

Figure 6. BTV‐NS3 protein ubiquitination by K63 Ub chains is critical to its IFN‐I antagonistic effects.

• A
  WCE and FLAG‐tag NS3 IP assay immunoblots of 293T transfected with BTV‐NS3‐FLAG plasmids and increasing amounts of ubiquitin‐HA plasmids and probed (IB) for ubiquitin (HA), BTV‐NS3 (FLAG), or GAPDH expression.
• B
  Immunofluorescence staining of Vero cells co‐transfected with BTV‐NS3 and ubiquitin and acquired with Airyscan super‐resolution. Scale bar = 10 μm. Inset shows detailed area, and arrowheads indicate NS3 and ubiquitin signal co‐incidence (scale bar = 2 μm).
• C
  Fluorescence intensity profile of cell section indicated by the white line (blue: DAPI, green: NS3; red: Ub). Arrowheads on image and arrows on intensity profile indicate matching fluorescence signal for NS3 and Ub.
• D
  FLAG‐tag IP assay of BTV‐NS3 and WCE immunoblot of 293T co‐transfected with BTV‐FLAG‐NS3‐, BTV‐FLAG‐NS3‐K13R, BTV‐FLAG‐NS3‐K15R, BTV‐FLAG‐NS3‐K13/15R, or BTV‐FLAG‐NS3‐PPRY/AARH plasmid and ubiquitin‐HA plasmid and probed (IB) for ubiquitin (HA), BTV‐NS3 (FLAG), STAT2, and GAPDH expression.
• E
  WCE and HA‐tag ubiquitin IP assay immunoblots of 293T transfected with ubiquitin‐HA plasmids and increasing amounts of BTV‐FLAG‐NS3‐K13/15R, and probed (IB) for ubiquitin (HA), BTV‐NS3‐K13/15R (FLAG), or GAPDH expression.
• F
  ISRE‐driven luciferase reporter assay in 293T cells transfected with empty expression plasmid (control), BTV‐NS2 (300 ng), BTV‐NS3 (50, 300, and 600 ng), or BTV‐NS3‐K13/15R (50, 300, and 600 ng)‐expressing plasmids. Mean % inductions of luciferase activity of three independent experiments are shown. ***P < 0.001, one‐way ANOVA.
• G
  FLAG‐tag IP assay of BTV‐NS3 and WCE immunoblot of 293T co‐transfected with BTV‐FLAG‐NS3 plasmid and HA‐UB‐WT or HA‐UB lysine (K to R) mutants. IP was probed (IB) for ubiquitin (HA) and BTV‐NS3 (FLAG) expression. WCE was probed for STAT1, STAT2, ubiquitin (HA), BTV‐NS3 (FLAG), and GAPDH expression.

Figure EV5. BTV‐NS3, but not BTV‐NS4, immunoprecipitates and colocalizes with ubiquitin.

• A, B
  Whole cell extract (WCE) and HA‐tag ubiquitin IP assay immunoblots of 293T transfected with ubiquitin‐HA plasmids and increasing amounts of (A) BTV‐NS3‐FLAG or (B) BTV‐NS4‐FLAG, and probed (IB) for ubiquitin (HA), BTV‐NS3 or BTV‐NS4 (FLAG), or GAPDH expression.
• C
  Alternative colocalization analysis of cell presented in Fig 6B and C. Immunofluorescence staining of Vero cells co‐transfected with BTV‐NS3 and ubiquitin and acquired with Airyscan super‐resolution (120 nm in xy). Thresholded fluorescence signal co‐incidence shown as pixel co‐incidence mask and LUT 16 colors for NS3 and Ub channels determined using ImageJ software and Image calculator tool.

Source data are available online for this figure.

NS3 has two putative lysine residues at positions 13 and 15 susceptible to ubiquitination 49. We mutated these lysine residues on NS3 to arginine to produce 3 mutants: NS3‐K13R, NS3‐K15R, and double‐mutant NS3‐K13/15R (Fig 6D). BTV‐NS3 protein also contains a highly conserved PPXY motif known to recruit the NEDD4 family of HECT E3‐Ub ligases 50, 51. We also mutated the motif PPRY to AARH to determine whether the E3 ligase‐recruiting domain participated in NS3 ubiquitination. To assess Ub binding to these NS3 mutants, the NS3 IP fraction of co‐transfected cells was probed with anti‐HA to detect ubiquitin presence. Ubiquitin binding was detected in all mutants except for the double‐mutant NS3‐K13/15R (Fig 6D). A co‐IP assay in which cells were co‐transfected with increasing amounts of the NS3 double‐mutant K13/15R‐expressing plasmid and Ub, and pulled down with anti‐HA (Ub) antibody confirmed the absence of NS3‐K13/15R in Ub pull downs, showing that lysine 13 and 15 are the main ubiquitination sites on NS3 (Fig 6E). These data also indicate that the PPRY motif does not participate in NS3 ubiquitination. STAT2 was detected in the pull down of all mutants, thus indicating that STAT2 binding to NS3 is independent of its ubiquitination status and the presence of a PPRY motif (Fig 6D). When STAT2 levels were assessed in WCE, only wild‐type NS3 could reduce STAT2 expression. It thus appears that ubiquitination on both lysine residues on NS3 is necessary for the correct targeting of STAT2 for degradation. Moreover, the rescued expression of STAT2 in NS3 PPRY/AARH‐transfected cells implicates that, in addition to NS3 ubiquitination, the PPRY motif is required for STAT2 degradation.

To determine the functional relevance of NS3 ubiquitination, an ISRE‐luciferase reporter assay was performed in cells expressing the NS3‐K13/15R double mutant (Fig 6F). NS3‐K13/15R expression did not interfere with luciferase activity in these assays, while wild‐type NS3 expression reduced ISRE promoter activity. These data indicate that NS3 ubiquitination on both lysine residues (13 and 15) and E3 ligase recruitment is required for its IFN‐I signaling interference.

NS3 is ubiquitinated through Ub K63 chains

Ubiquitin has seven lysines (K6, K11, K27, K29, K33, K48, and K63) that form polyubiquitin chains (reviewed in 52). Modification with these polyubiquitin chains can determine the intracellular fate of a protein. We studied the type of NS3 ubiquitination. This was done by detecting Ub presence in NS3‐immunoprecipitated extracts from cells co‐expressing NS3 and Ub mutants with lysine residues mutated to arginine that prevent Ub modifications (Fig 6G). These Ub mutants include single lysine residue mutants in which only one lysine residue has been modified to arginine (K6/11/27/29/33/48/63R mutants), and two mutants that only maintain one functional lysine residue, while all other lysines are mutated to arginine (K48 only and K63 only mutants). NS3 co‐IP with Ub was reduced with the Ub‐K63R mutant that only prevents K63‐linked polyubiquitin chain formation (Fig 6G), while the mutant Ub‐K63 only that form poly‐Ub chains exclusively through K63 linkage (as all other lysine residues on Ub are mutated to arginine) co‐IP more efficiently with NS3 as compared to UB‐WT. The poly‐Ub forms on NS3 were therefore predominantly K63‐linked. The importance of NS3 polyubiquitination through K63 chain was further supported by the observation that when K63 was mutated to R (in Ub K63R and UbK48 only mutants), NS3 failed to degrade STAT2 (WCE in Fig 6E). Therefore, NS3 linkage to K63 polyubiquitin appears essential for STAT2 degradation.

Discussion

Viruses have evolved a variety of mechanisms to evade the IFN response (reviewed in 15). Among them, degradation of essential factors for the IFN response through labeling proteins for proteasomal degradation by attachment of K48 polyubiquitin chains is triggered by some viruses 42, 53, 54. Here, we show that BTV inhibits type I and II IFN signaling and promotes STAT2 degradation through an autophagy/lysosome pathway rather than proteasomal activity. Moreover, we identified the BTV‐NS3 protein as the IFN antagonist that binds STAT2, leading to its degradation. Our findings also indicate that NS3 is polyubiquitinated by Ub K63 chains which drive STAT2 degradation through the lysosome/autophagosome pathway. Lysosomal degradation of a host protein involved in the IFN system has been documented 55. To our knowledge, our report is nonetheless the first description of a viral protein that uses its ubiquitination to degrade a host protein through autophagy.

It has been previously shown that BTV evolved mechanisms to subvert the IFN‐I response 32. In our study, BTV counteracts the IFN‐I response and interferes with IFN‐I induction and IFN‐II signaling. We found that BTV inhibits STAT1 phosphorylation and degrades STAT2 to prevent the IFN‐I signaling activation. Lack of STAT1 phosphorylation will also inhibit the homodimerization of pSTAT1 that activates IFN‐II pathway. These events take place early in the infection and before shutoff occurs in BTV‐infected cells (24 hpi), indicating that STAT2 degradation was not due to viral shutoff. STAT2 degradation is a mechanism used by flaviviruses such as Dengue 42, 54 and Zika 53, paramyxoviruses 56, 57, and respiratory syncytial virus 58. BTV also promotes STAT2 degradation to evade IFN responses, although in this case it uses an autophagy/lysosome pathway to degrade this transcription factor.

Our data showed that NS4 and NS3 proteins antagonize IFN responses in luciferase reporter assays and that NS3 protein expression reduces STAT2 expression and inhibits STAT1 phosphorylation. By contrast, NS4 does not affect STAT1 and STAT2 expression levels or their phosphorylation. It has been previously described that NS4 facilitates BTV replication in IFN‐I‐treated cells 20. In line with this, cells infected with an NS4 deletion mutant showed increased IFN‐I synthesis compared to cells infected with wild‐type BTV 31. Due to the nucleolar localization of NS4 in infected cells 20, it is plausible that NS4 is mediating IFN antagonism by a mechanism that involves interaction with other nucleolar proteins to inhibit IRF3 binding to its promoter. We are currently investigating the possible mechanism used by NS4 to interfere with IFN signaling and induction.

BTV‐NS3 protein interferes with IFN‐I induction by inhibiting the RLR‐dependent signaling pathway 30. Here, we show that NS3 antagonizes type I and II IFN signaling by targeting STAT2 for autophagic/lysosomal degradation and by inhibiting STAT1 phosphorylation. Infection with BTV‐8 ΔNS3 confirmed the role of NS3 in STAT2 degradation and STAT1 phosphorylation inhibition. BTV‐8 ΔNS3 infection only partially inhibited STAT1 phosphorylation indicating that BTV can block STAT1 phosphorylation through other pathways. Further work will be required to better characterize the mechanisms through which NS3 impairs STAT1 phosphorylation.

BTV‐NS3 protein is glycosylated and is the only membrane protein encoded by orbiviruses 59, associated with smooth intracellular membranes and plasma membrane 60. NS3 contains a long N‐terminal domain and a shorter C‐terminal cytoplasmic domain that are connected by two transmembrane domains and a short extracellular domain. In here, we demonstrate that NS3 interacts with STAT2, an interaction that is probably necessary but not sufficient to reduce STAT2 levels. We show that NS3‐mediated STAT2 degradation is dependent on lysosome activity, autophagosome formation and maturation, and autophagy key components Atg7 and Beclin‐1. NS3 therefore appears to be the viral protein responsible for the autophagic/lysosomal degradation of STAT2 observed during BTV infections.

The mechanism through which NS3 mediates STAT2 degradation requires NS3 ubiquitination anchored to K13 and K15 on NS3 and assembled through isopeptidic bonds involving K63 on ubiquitin. Ubiquitination on both NS3 residues is required for STAT2 degradation, and NS3 ubiquitination is critical to NS3 antagonistic effects on IFN‐I signaling since cells expressing the double‐mutant K13/15R could still signal through this pathway. STAT2 degradation is therefore a central mechanism employed by NS3 to interfere with IFN‐I signaling. Importantly, STAT2 was still immunoprecipitated with the NS3 single mutants, K13R or K15R, and the double‐mutant K13/15R, indicating that STAT2 interaction with NS3 occurs through a region that does not involve K13 or K15 or NS3 ubiquitination.

Co‐transfection of NS3 with ubiquitin mutants demonstrated that Ub linkage through K63 promoted NS3‐mediated STAT2 degradation. In mammals, K‐63‐linked ubiquitin chain is a signal that appears to be critical for endocytosis and efficient sorting to MVB 61, 62. Pharmacological disruption of MVB reduces BTV titers, probably by impairing virion release, a mechanism that involves NS3 activity 50, 51. NS3 can interact through its PSAP domain with Tsg101, a component of the ESCRT‐I complex involved in virion release 50 and in MVB protein sorting 62. NS3 also possesses PSAP and PPXY late‐domain motifs, which are highly conserved protein–protein interaction motifs 59. PSAP has been shown to play a role in BTV release, binding specifically the cellular protein Tsg101, a component of the ESCRT‐I complex 50. The PPRY domain of BTV‐NS3 is known to bind the WW domains of the NEDD4 family of HECT ubiquitin E3 ligases 46, 60. When the PPRY domain from NS3 was mutated to AARH, we found that NS3 was still ubiquitinated but that STAT2 expression was restored. We conclude that the E3 ligase responsible for NS3 ubiquitination is not recruited by the PPRY motif of NS3 but STAT2 degradation requires this intact PPRY domain on NS3 to recruit an E3 ligase to drive the degradation.

NS3 labeling with Ub K63 chains and E3 ligase recruitment on NS3 appears thus to promote STAT2 degradation through autophagy/lysosomal pathway. Based on these data, we propose a model in which the viral protein NS3 degrades STAT2 via an autophagy/lysosome pathway (Fig 7). In this model, NS3 act as an adaptor/scaffolding protein that interacts with STAT2, while NS3 ubiquitination with Ub K63 chains through the action of the host E1‐E2‐E3 ubiquitination machinery targets the complex to the autophagosome for lysosomal degradation. Although this model is consistent with the known activities of NS3 and our experimental results, we cannot exclude the possibility that STAT2 is targeted to lysosomal degradation through other mechanisms. Untangling the complex NS3 sorting mechanisms that are likely to take place in the MVB would allow for a better understanding not only of BTV biology but also of NS3 capacity to target cellular proteins for lysosomal degradation.

Figure 7. Model for BTV‐NS3 mediated autophagic/lysosomal STAT2 degradation.

BTV‐NS3 protein contains two conserved late domains PPRY and PSAP, which bind to ubiquitin E3 ligases and ubiquitin E2, respectively. BTV‐NS3 is likely to interact directly with STAT2 and act as a scaffolding/adaptor protein that facilitates the transport of STAT2 for degradation. NS3 contains two lysine residues at positions 13 and 15, which are K63 polyubiquitinated that target the NS3/STAT2 complex to the endosomes. STAT2 is targeted to autophagosome vesicles whose content is degraded after fusion with lysosomes. NS3 is probably deubiquitinated and allowed to recycle to the plasma membrane (as NS3 degradation was never observed).

In summary, our data reveal for the first time the use of autophagy by a viral protein to degrade an IFN signaling factor. Although further work will be necessary to fully characterize how this viral protein achieves STAT2 degradation, this study highlights the diversity and the complexity of the cellular mechanisms that viruses can hijack in order to favor their replication.

Materials and Methods

Cells, virus, and inhibitors

293T (ATCC CRL‐1573), BHK‐21 (ATCC CRL‐6281C), SSC 65, and Vero cells (ATCC CCL‐81) were grown in Dulbecco's modified Eagle medium (Invitrogen) supplemented with 10% fetal bovine serum (FBS), 2 mM l‐glutamine, 1% 100× non‐essential amino acids, 1 mM sodium pyruvate, and 100 U/ml penicillin/100 μg/ml streptomycin. All cells were kept at 37°C in the presence of 5% CO₂. Reverse genetic BTV‐8 and reverse genetic BTV‐8 without the NS3 encoding segment were prepared as described by Feenstra et al 66. rgBTV‐8, BTV‐8 ΔNS3, and BTV serotype 8 (Belgium/06) stocks were prepared from infection of BHK‐21 cells with multiplicity of infection (MOI) of 0.1. For inhibitor experiments, cells were treated 18 hpi and for 6 h. Inhibitors were used at the following concentration: proteasome inhibitors MG132 (20 μM) and lactacystin (20 μM); lysosome acidification inhibitors chloroquine (1, 10, 20, 50 μM) (Sigma) and NH₄Cl (0.2, 2, 10 mM); autophagosome formation inhibitor 3‐MA (0.5, 5, and 10 mM) (all from Sigma); autophagosome maturation inhibitor, bafilomycin A1 (0.2 μM) (InvivoGen).

Plasmids

BTV‐8 VP6, VP7, NS1, NS2, NS3, and NS4 open reading frame (ORF) were amplified by RT–PCR of RNA extracted from BTV‐8‐infected BHK‐21 cells. PCR amplicons were cloned into pIRES‐cOFP‐COOH‐FLAG (Promega) expression vector. All primer sequences are available upon request. Intact NS3 gene recombinant construct (NS3‐FLAG‐pIRES‐cOFP) was used as template to make the different point mutants from NS3 gene by substituting K amino acid residues or PPRY region. Four FLAG‐tagged expression vectors were generated and from here on are referred to as NS3‐FLAG‐K13R, NS3‐FLAG‐K15R, NS3‐FLAG‐K13/15R, and NS3‐FLAG‐PPRY/AARH.

The NiV V‐HA, DenV‐NS5‐HA, WNV‐NS5‐HA, HA‐UB‐WT, HA‐UB‐K48 only, HA‐UB‐K63 only, HA‐UB‐K63R, HA‐UB‐K48R, HA‐UB‐K33R, HA‐UB‐K29R, HA‐UB‐K27R, HA‐UB‐K11R, and HA‐UB‐K6R expression plasmid were previously described 42, 67.

Reporter assays

Cells (293T or Vero) were co‐transfected with 0.3 μg of pISRE‐firefly‐luc, pGAS‐firefly‐luc (Promega), or pIFN‐β‐firefly‐luc, and 0.1 μg of a constitutively expressing Renilla luciferase reporter construct (pRL‐null). For BTV‐8 infection, at 24 h post‐transfection cells were infected at MOI1 in serum‐free medium. After 1 h, cells were incubated in culture medium supplemented with 1,000 U/ml universal type I IFN (PBL) or IFN‐γ (5 ng/ml) (PBL) or infected with Sendai virus (SeV) at MOI of 1 PFUs/cell. At 6, 16, and 24 h pi, luciferase activity was analyzed using the Dual‐Glo luciferase assay system (Promega).

To determine BTV protein effects on IFN signaling or induction, cells were co‐transfected with reporter plasmids as before and 0.3 μg (or the indicated amounts) of the BTV protein expression plasmids. Cells were treated 24 h later with 1,000 U/ml universal type I IFN or IFN‐γ (5 ng/ml) or infected with SeV (MOI of 1). Luciferase activity was measured 16 h post‐stimulation. Luciferase activities were normalized to Renilla activity in all assays. To assess endogenous transcriptional activity in infected or transfected cells and control for cellular shutoff, cells (293T or Vero) were co‐transfected with 0.3 μg pGL3‐firefly‐luc and 0.3 μg pCMV‐ren‐luc and luciferase activity measured 24 h post‐transfection. Luciferase activities were normalized to firefly activity in all assays.

RNA extraction and RT–PCR

293T or Vero cells were infected with BTV‐8 at a MOI of 1, washed, and then treated 16 h later with 1,000 U/ml universal IFN or 5 ng/ml IFN‐γ for an additional 12 h. Total cellular RNA was then extracted using TRIzol Reagent Solution (Thermo Scientific) and cDNA generated by reverse transcription using oligo dT primers with SuperScript III reverse transcriptase (Invitrogen). Transcription levels of MxA, ISG15, OAS, IP10, and IRF1 genes were evaluated by real‐time PCRs performed in a LightCycler 480 System instrument (Roche) using the LightCycler 480 SYBR Green I Master Reagents (Roche). Gene expression was normalized to 18S rRNA gene expression, and relative expression levels were calculated using the 2^−ΔCt method. A melting curve for each PCR fluorescence reading, every degree between 60 and 95°C, was determined to ensure that only a single product had been amplified.

Western blots

For cell extract preparation, 293T, Vero, or SSC cells were cultured in six‐well plates and transfected with our experimental and control plasmids. After 48 h, cells were treated (or mock treated) with 1,000 U/ml universal type I IFN for 30 min at 37°C, washed with PBS, and subsequently lysed on ice for 20 min with 250 μl of RIPA buffer (Sigma) supplemented with protease inhibitor cocktail (Roche). Lysates were centrifuged for 20 min at 20,000 ×g at 4°C, and protein content was determined using the Micro BCA Protein Assay kit (Thermo Scientific). Cell lysates in SDS sample buffer were boiled for 10 min at 95°C and equal protein amounts resolved on sodium dodecyl sulfate (SDS) 4–15% polyacrylamide gels. Proteins were then transferred to PVDF membranes and probed with anti‐STAT1 (Sigma #SAB1105169), anti‐STAT2 (Santa Cruz, #sc‐476), anti‐phospho‐STAT1 (Tyr701) (Cell Signaling Technologies, #58D6), anti‐phospho‐STAT2 (Tyr689) (Millipore, #07‐224), anti‐GAPDH (Sigma #G8795), anti‐Atg7 (Cell Signaling, D12B11), anti‐Beclin‐1 (Cell Signaling, D40C5), anti‐HA‐tag epitope (Sigma, #H3663) or anti‐FLAG (Sigma, #085M4774V) antibodies. Secondary HRP‐conjugated anti‐mouse or anti‐rabbit IgG (both from GE Healthcare) antibodies were used and protein bands visualized by chemiluminescence (ECL Plus, Thermo Scientific).

Immunoprecipitation assays

Cells lysates were obtained as previously described in Western blot section using non‐denaturing co‐immunoprecipitation buffer (50 mM Tris pH 8.0, 280 mM NaCl, 0.5% IGEPAL, 0.2 mM EDTA, 2 mM EGTA, 10% glycerol, 1 mM dithiothreitol, supplemented with protease inhibitor cocktail). A portion of total lysates was kept as whole cell extract (WCE). Immunoprecipitation (IP) was performed on the remaining lysate with red anti‐HA or anti‐FLAG affinity gel antibody beads (Sigma, #E6779) incubated overnight at 4°C under rolling agitation. HA or FLAG beads were washed, and bound protein was eluted by boiling the beads in Laemmli sample buffer. SDS–PAGE followed by immunoblot analysis was performed on the IP and whole cell extracts with the indicated antibodies.

Immunofluorescence assays

Vero cells were grown to 70% confluence on coverslips in 12‐well plates transfected with the indicated expression plasmids and 24 h later treated with 1,000 U/ml universal type I IFN for 30 min. Cells were fixed with 4% PFA for 15 min and permeabilized with PBS containing 0.1% NP40 for 10 min at room temperature (RT). Nonspecific binding sites were blocked with Dako antibody diluent (Dako, #S3022) for 1 h at RT. Primary antibodies were incubated overnight at 4°C; coverslips were washed in PBS and incubated for 1 h at RT with secondary antibodies. Anti‐mouse IgG or anti‐rabbit IgG secondary antibodies conjugated to Alexa Fluor 647 or 488 (Invitrogen) were used to visualize phospho‐STAT1, STAT2, BTV‐VP7 protein (VMRD), HGS (Abcam, ab72053), LAMP1‐Alexa647 (BioLegend, 328612), p62 (Santa Cruz, D‐3) HA‐ or FLAG‐tagged proteins. Nuclear chromatin staining was performed by incubation in PBS containing 0.5 mg/ml 4′,6‐diamidino‐2‐phenylindole DAPI (Sigma). Coverslips were mounted using Prolong Gold antifade reagent (Invitrogen). Images were captured using a LSM 880 confocal microscope (Zeiss). Image analysis was performed with the ImageJ software (http://rsbweb.nih.gov/ij/ US National Institutes of Health).

Cell sorting

293T, SSC, or Vero OFP‐positive cells transfected with OFP expressing plasmids pIRES‐FLAG‐tagged BTV‐NS3 or pIRES‐empty were sorted using a BD FACSAria III cell sorter. Recovered cells were used to prepare cell extracts as described in Western blot section.

To sort BTV‐infected cells, Vero cells were infected with BTV8‐ΔNS3 or rgBTV‐8 (MOI of 1) and at 16hpi cells were fixed with 4% PFA for 15 min and intracellularly staining with anti‐VP7‐FITC antibody (VMRD). VP7⁺ cells were sorted using a BD FACSAria III cell sorter. Recovered cells were treated for protein extraction according to Sadick et al 68.

Flow cytometry

For flow cytometry analysis of STAT2 expression, Vero cells were infected with BTV‐8 at the indicated MOI and fixed after 24 h with 4% PFA for 15 min. Cells were then permeabilized and intracellular staining performed as described in 69. Cells were stained with anti‐STAT2 antibody and secondary Alexa Fluor 647‐conjugated anti‐rabbit IgG antibody (Thermo Fisher). For STAT‐1 phosphorylation assays, Vero cells were infected with BTV‐8 at the specified MOI for 24 h and stimulated thereafter for 30 min with 1,000 U/ml universal type I IFN. Cells were then fixed and permeabilized with 100% methanol at −20°C for 10 min and after washing stained with anti‐phospho‐STAT1 antibody. Samples were acquired on a BD FACScalibur flow cytometer and analyzed using the FlowJo software (Tree Star Inc, USA).

Atg7 and Beclin‐1 siRNA transfection

Atg7 (SignalSilence Atg7 siRNA I #6604), Beclin‐1 (SignalSilence Beclin‐1 siRNA I #6222), and control siRNA (SignalSilence control siRNA #6568) were purchased from Cell Signaling Technology. For Atg7 silencing, 293T cells were co‐transfected in 6‐well plates with TransIT‐X2 (Mirus Bio) with siRNA (25 nM) and NS3‐ or empty‐OFP expression plasmid (1 μg). For Beclin‐1 silencing, 293T cells were transfected in 6‐well plates with 25 nM siRNA using Lipofectamine 3000 (Invitrogen) and co‐transfected 24 h later with NS3‐ or empty‐OFP expression plasmid (1 μg) using TransIT‐LT1 (Mirus Bio). OFP⁺ cells were sorted by flow cytometry and lysate prepared 24 h after NS3‐plasmid transfection.

Densitometry analysis

Densitometric image analysis was performed with the ImageJ software (http://rsbweb.nih.gov/ij/ US National Institutes of Health). The GAPDH protein band intensity was used as internal control to normalize protein expression.

Statistical analysis

Statistical analysis was performed using Prism 5.0 software (GraphPad Software Inc, USA).

Author contributions

MA, JMR, LM, and EP performed the experiments. MA, JMR, LM, VM, AG‐S, and NS analyzed the data. PAVR provide the rgBTV‐8 and BTV‐8 ΔNS3 viruses. NS and VM designed the study. NS and JMR wrote the manuscript.

Conflict of interest

The authors declare that they have no conflict of interest.

Supporting information

Appendix

Expanded View Figures PDF

Source Data for Expanded View and Appendix

Review Process File

Source Data for Figure 2

EMBO Reports (2019) 20: e48766