Ebola Virus Protein VP35 Impairs the Function of Interferon Regulatory Factor-Activating Kinases IKKɛ and TBK-1^▿

Antibodies.

Monoclonal antibodies against the Zaire EBOV VP35 (6C5) and Zaire EBOV nucleoprotein (NP) were generated in collaboration with the Mount Sinai Hybridoma Center. The monoclonal anti-hemagglutinin (HA) and anti-FLAG (M2) and polyclonal anti-HA and anti-FLAG antibodies were purchased from Sigma (St. Louis, MO).

Cell lines and viruses.

293T cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum at 37°C and 5% CO₂. Sendai virus (SeV) strain Cantell was grown in 10-day-old embryonated chicken eggs for 2 days at 37°C.

Plasmids.

Zaire EBOV VP35, FLAG-RIG-I, and IPS-1 were cloned into pCAGGS as described elsewhere previously (3, 7, 37). Plasmids encoding human cDNAs for wild-type TBK-1 and ΙΚΚɛ were kindly provided by John Hiscott (McGill University). FLAG-tagged versions of these cDNAs were amplified by PCR and inserted into expression plasmid pCAGGS (37). A kinase-inactive ΙΚΚɛ (IKKɛKN) was generated by introducing the previously described K38A mutation into IKKɛ (39, 40, 47). A kinase-inactive K38M mutant of TBK-1 (TBK-1KN) was kindly provided by Benjamin tenOever (Mount Sinai School of Medicine). Plasmids encoding human IRF-3 were previously described (2). IRF-3 amino acids 375 to 427 were amplified by PCR for expression as glutathione S-transferase (GST) fusions in Escherichia coli. The pCAGGS-FLAG-IRF-7 construct was kindly provided by Adolfo García-Sastre (Mount Sinai School of Medicine).

Bacterial expression and purification of GST and the GST-IRF-3 C terminus.

GST and the GST-IRF-3 C terminus, residues 375 to 427 (IRF-3-C), were expressed in E. coli Origami B BL21(DE3)pLysS host strains (Stratagene). Cultures were grown at 37°C to an optical density at 600 nm of 0.57, and IRF-3 expression was induced by the addition of 0.1 mM IPTG (isopropyl-β-d-thiogalactopyranoside). Induced cells were grown further at 18°C for 24 h. Lysates were prepared by sonication for 10 s five times in lysis buffer (25 mM Tris-HCl [pH 7.5], 200 mM NaCl, 0.1% NP-40, 1 mM EDTA, 0.1 mM dithiothreitol, and a cocktail of protease inhibitors [Roche]). Bacterially produced protein was purified from cell lysates on a glutathione-Sepharose (Amersham Biosciences) column. After loading, the column was washed with a solution containing 25 mM Tris-HCl, 1 M NaCl, 0.1% NP-40, and 1 mM EDTA and eluted with a solution containing 5 mM glutathione, 25 mM Tris, 200 mM NaCl, 1 mM EDTA, 1 mM Tris(2-carboxyethyl)phosphine, 5% glycerol, and 0.2% CHAPS {3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate}. Dialysis was then performed overnight in 1 liter of kinase buffer (20 mM HEPES, 1 mM beta-glycerophosphate, 50 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, and 0.1 mM NaVO₃).

Transfections.

HEK 293T cells were transfected with a 1:1 ratio of Lipofectamine 2000 to plasmid DNA in OptiMEM medium (Gibco) at 37°C for 8 h. For subsequent infection of cells, the transfection medium was removed, and SeV was added, at a multiplicity of infection of 10, in a solution containing phosphate-buffered saline and 0.3% bovine serum albumin for 1 h. Infection medium was then replaced with Dulbecco's modified Eagle's medium with 10% fetal bovine serum, and cells were incubated at 37°C overnight. Following overnight incubation, cells were lysed in lysis buffer (50 mM Tris [pH 8], 1% NP-40, 280 mM NaCl, 0.2 mM EDTA, 2 mM EGTA, and 10% glycerol).

Immunoprecipitations.

Lysates were incubated with 1 μg of the indicated antibody for 4 h at 4°C, followed by a 1-h incubation with protein A-Sepharose beads (Roche). Beads were washed five times with lysis buffer. After washing, beads were resuspended in sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) sample loading buffer, separated by 10% SDS-PAGE, and analyzed by Western blotting as indicated.

Purification of FLAG-tagged proteins.

HEK 293T cells were transfected with 2 μg of expression plasmids for FLAG-tagged ΙΚΚɛ, ΙΚΚɛKN, TBK-1, TBK-1KN, or VP35. The transfected cell lysates were immunoprecipitated with M2 anti-FLAG affinity gel (Sigma). The FLAG-tagged proteins were eluted from the affinity gel by two sequential incubations with FLAG peptide at 100 μg/ml. The eluate was concentrated 20-fold, and buffer was exchanged to kinase buffer with a Microcon centrifugal filter device (Millipore). Proteins were stored at −80°C in kinase buffer (without dithiothreitol or NaVO₃) supplemented with 4% glycerol.

In vitro kinase assays.

Purified FLAG-tagged ΙΚΚɛ, TBK-1, TBK-1KN, or ΙΚΚɛKN was incubated with a solution containing 5 μCi [γ-³²P]ATP (Perkin-Elmer); 0.1 mM unlabeled ATP; and either GST, GST-IRF-3 (residues 375 to 427), or FLAG-tagged VP35 in 30 μl kinase buffer (20 mM HEPES, 1 mM beta-glycerophosphate, 50 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, and 0.1 mM NaVO₃) as described previously (27). Reaction mixtures were incubated at 30°C for 30 min and terminated by the addition of SDS sample loading buffer. Proteins were separated by 12% SDS-PAGE, and phosphorylation was visualized by autoradiography.

To examine IRF-3 phosphorylation by cell lysates, in vitro kinase assays were performed as described above using GST-IRF-3-C as a substrate and lysates from 12% of 2 × 10⁶ cells (2.4 × 10⁵ cell equivalents) cotransfected with FLAG-ΙΚΚɛ expression plasmid and either empty vector as a source of kinase, increasing concentrations of VP35 expression plasmid (1 μg, 2 μg, and 4 μg), or EBOV NP plasmid. Kinase reactions were terminated after 1 h by the addition of glutathione-Sepharose (Amersham Biosciences) in lysis buffer for affinity purification of GST-IRF-3-C. Following incubation on a Nutator mixer for 1 h, beads were washed five times with lysis buffer. SDS-PAGE sample loading buffer was added, and proteins were separated by 10% or 12% SDS-PAGE. Phosphorylation was visualized by autoradiography and quantified by ImageJ software. Transfected proteins were visualized by Western blotting of lysates from 1 × 10⁵ cell equivalents with the indicated antibodies.

Reporter assay.

293T cells were transfected with the indicated amount of expression plasmid DNA together with 400 ng of the IFN-α4-chloramphenicol acetyltransferase (CAT) reporter plasmid (2) and 200 ng of the constitutive firefly luciferase reporter plasmid. Twelve hours posttransfection, cells were infected with SeV or mock infected for 1 h. Twelve hours postinfection, cells were lysed with reporter lysis buffer (Promega), and CAT activities were measured (43). Firefly luciferase activity was determined as recommended by the manufacturer (Promega) and was used to normalize CAT activity. IFN-α4 reporter gene activation is expressed as induction over an empty vector mock-infected control.

VP35 interacts with the IRF-3 kinases IKKɛ and TBK-1.

VP35 was previously reported to block the phosphorylation of IRF-3 and to inhibit the IRF-3-dependent gene expression induced by the overexpression of ΙΚΚɛ and TBK-1 (2, 7). To test whether ΙΚΚɛ and TBK-1 are targeted by VP35, coimmunoprecipitation (co-IP) experiments were performed. 293T cells were transfected with expression plasmids for FLAG-tagged kinase-inactive forms of IKKɛ (IKKɛKN) (Fig. 1, lanes 2, 5, 7, 8, and 10) or TBK-1 (TBK-1KN) (Fig. 1, lanes 1, 4, and 6) (39, 40, 47) alone or with VP35 (Fig. 1, lanes 3 to 7, 9, and 10). Inactive kinases were used for co-IP experiments because the interaction of the kinases with their substrate IRF-3 is more readily detected by co-IP when inactive rather than functional kinases are used (data not shown). This presumably reflects the fact that active kinases, upon overexpression, rapidly phosphorylate IRF-3, resulting in its nuclear accumulation. The VP35-kinase interactions, however, can be detected by co-IP with equal efficiency using either the kinase-active or -inactive forms (data not shown). Twelve hours posttransfection, cells were either mock infected (Fig. 1, lanes 1 to 5) or infected with SeV (lanes 6 to 10). Twenty-four hours posttransfection, cells were lysed as described in Materials and Methods. Immunoprecipitations were then performed on lysates by using an anti-VP35 monoclonal antibody (Fig. 1, lanes 1 to 7). IKKɛKN and TBK-1KN each coprecipitated with VP35 (lanes 4 to 7). The reciprocal co-IP was also performed with FLAG-IKKɛKN using anti-FLAG monoclonal antibodies (lanes 8 to 10). When an anti-FLAG immunoprecipitation was performed, VP35 coprecipitated with FLAG-IKKɛKN (Fig. 1, lane 10). Therefore, VP35 interacts with the IRF kinases.

IKKɛKN and TBK-1 can phosphorylate VP35 in vitro.

Having demonstrated a physical interaction between VP35 and both ΙΚΚɛ and TBK-1, we sought to determine whether IKKɛ or TBK-1 can phosphorylate VP35 in an in vitro kinase assay. FLAG-tagged IKKɛ, FLAG-tagged TBK-1, and FLAG-tagged VP35 were each purified from separate, transiently transfected 293T cell cultures. Increasing amounts of FLAG-IKKɛ (Fig. 2A and B) were incubated with constant amounts of GST (Fig. 2A and B, lanes 1 to 3), GST fused to the C-terminal region of IRF-3 (amino acids 375 to 427) (IRF-3-C) (Fig. 2A and B, lanes 4 to 6), or FLAG-VP35 (Fig. 2A and B, lanes 7 to 9). In vitro kinase assays were performed as described in Materials and Methods. The products were separated by SDS-PAGE and developed by autoradiography (Fig. 2A) and by Coomassie blue staining (Fig. 2B). Alternatively, FLAG-TBK-1 was used in in vitro kinase assays (Fig. 2C and D, lanes 4 to 7) with GST (lanes 1 and 5), GST-IRF-3-C (lanes 2 and 6), or FLAG-VP35 (lanes 3 and 7).

As previously reported, both IKKɛ and TBK-1 undergo apparent autophosphorylation (Fig. 2A and C) (47). Neither kinase phosphorylated the negative control protein GST (Fig. 2A, lanes 1 to 3, and C, lane 5), nor was VP35 or GST-IRF-3-C phosphorylated when the kinase-negative forms of the kinases were used (Fig. 2A, lanes 10 to 11, and C, lane 8). The kinase-competent IKKɛ phosphorylated GST-IRF-3-C as well as VP35, shown in Fig. 2A, lanes 4 to 6 and 7 to 9, respectively. TBK-1 also phosphorylated GST-IRF-3-C and VP35, as shown in Fig. 2C, lanes 6 and 7, respectively. The examination of the Coomassie blue-stained gels demonstrated that the purified protein preparations did not contain visible amounts of contaminating cellular proteins and that comparable amounts of GST, GST-IRF-3-C, and VP35 were present in the kinase reactions (Fig. 2B and D). These data demonstrate that IKKɛ and TBK-1 can phosphorylate VP35.

VP35 disrupts IKKɛKN-IRF-3 and IKKɛKN-IRF-7 interactions.

The interaction of VP35 with IKKɛ and TBK-1, coupled with the phosphorylation of VP35 by these kinases, suggested that VP35 might act as an alternative substrate which blocks the interaction between IKKɛ or TBK-1 and their IRF substrates. To address this question, co-IP experiments were performed, focusing on IKKɛ as a representative IRF-3 kinase. Cells were cotransfected with full-length FLAG-IKKɛKN (Fig. 3A, lanes 1 and 4 to 7) and HA-IRF-3 (lanes 2 and 4 to 7) plasmids in the absence (lane 4) or the presence (lanes 5 to 7) of increasing amounts of VP35 plasmid. IRF-3 was immunoprecipitated using monoclonal anti-HA antibody, and coprecipitated FLAG-IKKɛKN was analyzed by Western blotting with anti-FLAG polyclonal antibody. HA-IRF-3 pulled down FLAG-IKKɛKN, and the amount of IKKɛKN that coprecipitated with IRF-3 decreased as the amount of VP35 increased (Fig. 3A, top).

VP35 was previously reported to block IRF-3-dependent gene expression induced by SeV infection (2, 7). To determine whether VP35 could block IRF-7-dependent gene expression as well, reporter assays were performed utilizing an IRF-7-dependent promoter, IFN-α4. 293T cells were transfected with an IFN-α4-CAT reporter plasmid, a constitutive firefly luciferase plasmid, and either empty vector (Fig. 3B, samples 1 and 4), FLAG-IRF-7 alone (samples 2 and 5), or FLAG-IRF-7 with VP35 expression plasmids (samples 3 and 6). Cells were subsequently mock infected or infected with SeV at a multiplicity of infection of 10. The expression of IRF-7 alone was sufficient to weakly induce IFN-α4 promoter activation relative to empty vector-transfected cells (Fig. 3B, samples 1 and 2). This IRF-7-induced activation of the promoter was decreased upon coexpression with 3 μg of VP35 (sample 3). SeV infection of IRF-7-expressing cells resulted in a dramatic induction of the IFN-α4 reporter (sample 5) compared with that seen for mock-infected cells (samples 1 and 2) and SeV-infected empty vector-expressing cells (sample 4), highlighting the role of IRF-7 in this reporter activation. The coexpression of VP35 with IRF-7 drastically decreased SeV-induced IFN-α4 reporter activation (sample 6). Therefore, VP35 is able to block the SeV-mediated activation of an IRF-7-dependent promoter.

Consistent with this functional assay, a biochemical assay similar to that shown in Fig. 3A was performed. 293T cells were cotransfected with full-length HA-IKKɛKN (Fig. 3C, lanes 1 and 3 to 6) and FLAG-IRF-7 (lanes 2 to 6) plasmids in the absence (lane 3) or the presence (lanes 4 to 6) of increasing amounts of VP35 plasmid. IKKɛKN was immunoprecipitated using a monoclonal anti-HA antibody, and coprecipitated FLAG-IRF-7 was analyzed by Western blotting with anti-FLAG polyclonal antibody. HA-IKKɛKN pulled down FLAG-IRF-7 (Fig. 3C, lane 3), and the amount of IRF-7 that coprecipitated with IKKɛKN decreased when increasing amounts of VP35 were present. The loss of the IRF-7-kinase interaction was most dramatic when the highest concentration of VP35 was present (Fig. 3C, lane 6). Notably, in both the IRF-3 (Fig. 3A) and IRF-7 (Fig. 3C) experiments, the presence of VP35 results in increased levels of expression of the cotransfected IRF and, to a variable extent, the cotransfected IKKɛKN. The molecular basis of this effect is unclear, but this may influence the efficiency with which VP35 appears to affect kinase-IRF interactions in these different experiments. As previously reported by Cárdenas et al. (7), the amounts of VP35 produced in transfected cells is comparable to what is seen in EBOV-infected cells, suggesting that results obtained using transfected cells are likely to be biologically relevant. Cumulatively, these data demonstrate that VP35 can disrupt the physical interaction between full-length IKKɛKN and either IRF-3 or IRF-7, and this physical disruption contributes to the IFN antagonist function of VP35.

The IKKɛ amino-terminal kinase domain can interact with IRF-3 and with VP35.

The amino-terminal kinase domains of IKKɛ and TBK-1 are quite homologous, with the two proteins exhibiting approximately 70% amino acid identity over their first 350 amino acids (data not shown). To determine if VP35 can interact with the kinase domain of IKKɛ, a FLAG-tagged IKKɛKN deletion mutant consisting of the amino-terminal 315 amino acids (N315) was created (39, 40, 47, 50). This mutant was tested for interactions with either HA-tagged IRF-3 or untagged VP35 by co-IP assay (Fig. 4). Cells were transfected with either full-length IKKɛKN (Fig. 4, lanes 1, 3, and 5) or the N315 truncation mutant (lanes 1, 4, and 6) and either HA-tagged IRF-3 (lanes 3 and 4) or VP35 (lanes 5 and 6). HA-IRF-3 and VP35 were expressed in the absence of kinase (Fig. 4, lane 2). Anti-HA and anti-VP35 monoclonal antibodies were then added to the transfected cell lysates for the immunoprecipitation of IRF-3 and VP35, respectively. As shown in Fig. 4, VP35 and IRF-3 each physically interacted with the N315 kinase domain (Fig. 4, lanes 4 and 6) as assessed by Western blotting with anti-FLAG polyclonal antibody, further suggesting that VP35 might physically block the IRF-3-IKKɛ interaction.

Overexpression of VP35 disrupts IKKɛKN N315-IRF-3 interactions.

To determine whether VP35 can disrupt binding between the ΙΚΚɛ kinase domain and IRF-3, co-IP experiments were performed using the IKKɛKN kinase domain (N315). FLAG-tagged N315 was expressed in 293T cells alone (Fig. 5, lane 1) or with HA-IRF-3 (lanes 3 to 9) in the absence (lane 4) or presence (lanes 6 to 9) of increasing amounts of VP35. VP35 was expressed in the absence of a kinase domain or HA-IRF-3 (Fig. 5, lane 2). IRF-3 was immunoprecipitated with anti-HA antibody, and coprecipitation of the N315 kinase domain was assessed by Western blotting using polyclonal anti-FLAG antibody. IRF-3 interacted with N315, but the amount of N315 coprecipitated decreased as levels of VP35 increased (Fig. 5, top). Therefore, the presence of VP35 physically disrupts IKKɛ kinase domain-IRF-3 interaction in a concentration-dependent manner.

VP35 overexpression disrupts the IKKɛ-IPS-1 interaction.

IKKɛ also interacts with IPS-1 (33). To determine whether VP35 might influence this interaction as well, we performed co-IP experiments similar to those shown in Fig. 5, using IPS-1 as an IKKɛ binding partner. Cells were transfected with HA-IPS-1 (Fig. 6, lanes 3 to 9) and FLAG-IKKɛ (lanes 2, 4, and 6 to 9) expression plasmids in the absence or presence of increasing amounts of VP35 (lanes 5 to 9). HA-IPS-1 was immunoprecipitated by adding anti-HA monoclonal antibody to the cell lysates, and co-immunoprecipitation of FLAG-IKKɛ was assessed by Western blotting using anti-FLAG polyclonal antibody. The IKKɛKN-IPS-1 interaction was impaired, as shown in Fig. 6 (lane 9, top). However, the inhibition did not show a linear dose response to VP35 and was seen only in the samples where the maximum amount of VP35 plasmid was transfected. Cumulatively, these data suggest that the presence of VP35 can disrupt interactions between IKKɛ and IRF-3, IRF-7 (Fig. 3A and C), and IPS-1 (Fig. 6).

VP35 decreases IRF-3 kinase activity in IKKɛ-expressing cells.

Previous data showed that the presence of VP35 decreased levels of phosphorylated IRF-3 present in cells following SeV infection (2). The data described above suggest that VP35 inhibits the ability of IKKɛ to interact with IRF-3. To determine whether VP35 expression decreases the ability of IKKɛ to phosphorylate IRF-3, an in vitro kinase assay was performed using as a source of enzyme lysates from cells transfected with IKKɛ plasmid in the presence or absence of VP35 plasmid. Cells were transfected with either empty vector (Fig. 7, lane 1) or FLAG-IKKɛ (lanes 2 to 6) and increasing amounts of VP35 (1, 2, and 4 μg) (lanes 3 to 5) or EBOV NP (4 μg) (lane 6) as an irrelevant protein control. Transfected cell lysates were added to kinase assays where GST-IRF-3-C served as a substrate (Fig. 7, lanes 1 to 6) and were subsequently purified on glutathione beads. The precipitated products were then separated by SDS-PAGE and developed by Coomassie blue staining (Fig. 7, middle) and by autoradiography (top). As shown in Fig. 7, equal amounts of IRF-3 phosphorylation (top), as determined by densitometry, were detectable in samples with IKKɛ alone or IKKɛ plus EBOV NP. However, the phosphorylation of IRF-3 decreased in the presence of VP35 to 89%, 79%, and 57% of the control. Levels of GST-IRF-3 (Fig. 7, middle), IKKɛ, NP, and VP35 are provided for comparison (Western blotting using anti-FLAG, anti-NP, and anti-VP35 monoclonal antibodies are shown in the bottom three panels). Therefore, the presence of VP35 reduces IRF-3 phosphorylation by IKKɛ.