vCLAP, a Caspase-recruitment Domain-containing Protein of Equine Herpesvirus-2, Persistently Activates the I k B Kinases through Oligomerization of IKK g *

vCLAP, the E10 gene product of equine herpesvirus-2, is a caspase-recruitment domain (CARD)-containing protein that has been shown to induce both apoptosis and NF- k B activation in mammalian cells. vCLAP has a cellular counterpart, Bcl10/cCLAP, which is also an ac-tivator of apoptosis and NF- k B. Recent studies demonstrated that vCLAP activates NF- k B through an I k B kinase (IKK)-dependent pathway, but the underlying mechanism remains unknown. In this report, we demonstrate that vCLAP associates stably with the IKK complex through direct binding to the C-terminal region of IKK g . Consistent with this finding, IKK g was found to be essential for vCLAP-induced NF- k B activation, and the association between vCLAP and the IKK complex induced persistent activation of the IKKs. Moreover, enforced oligomerization of the isolated C-terminal region of vCLAP, which interacts with IKK g , can trigger NF- k B activation. Finally, substitution of the C-terminal region of IKK g , which interacts with vCLAP, with the CARD of vCLAP or Bcl10 produced a molecule that was able to activate NF- k B when ectopically expressed in IKK g -de-ficient cells. These data suggest that vCLAP-induced (GFP) (pEGFP-C1) and the red fluorescent protein (RFP) (pD- sRed1-N1) were from CLONTECH. FLAG-M5 antibody was from Sigma. T7-horseradish peroxidase conjugate antibody was from Nova- gen. IKK a and IKK g polyclonal antibodies were from Santa Cruz. Biochemical Analysis— Immunoprecipitations were performed as de- scribed previously (20), and the precipitated proteins were analyzed by

Equine herpesvirus-2 (EHV-2) 1 is a gammaherpesvirus related to other lymphotropic herpesviruses such as herpesvirus saimiri and Epstein-Barr virus. EHV-2 contains 79 reading frames that encode 77 distinct molecules, several of which show significant similarity to cellular genes. One of these molecules, vCLAP (also called vCIPER/E10/vCARMEN) (1-4), a CARDcontaining apoptotic protein, was recently found to induce both apoptosis and activation of the transcription factor NF-B in mammalian cells. vCLAP, like its cellular counterpart Bcl10, contains two domains, an N-terminal CARD that can oligomerize via homotypic interactions and a C-terminal domain that probably functions as the NF-B activation domain. Because NF-B activation is considered to be a survival signal, virally encoded proteins, such as vCLAP, may be utilized by viruses as a strategic tool to initiate self-replication or to suppress apoptosis in infected cells (5).
In most resting cells, NF-B is sequestered in the cytoplasm through interaction with the IB inhibitory proteins. IBs mask the NF-B nuclear localization signal, thereby preventing its nuclear uptake. Exposure of cells to a wide variety of stimuli, such as viral or bacterial infection, inflammatory cytokines, or UV irradiation leads to the rapid phosphorylation, ubiquitination, and ultimately proteolitic degradation of the IBs (6 -9). This allows the activated NF-B to translocate to the nucleus and activate the transcription of several NF-B target genes.
The kinase activity responsible for phosphorylation of IBs is present in a large (700 -900 kDa) cytoplasmic complex composed of two catalytic subunits, IKK␣ and IKK␤ (10 -14), and a noncatalytic subunit termed IKK␥ (also called NEMO, IKKAP1, or FIP-3) (15)(16)(17)(18). We and others have recently demonstrated that activation of the IKK complex could be achieved through IKK␥-mediated oligomerization of the IKK kinases, indicating that IKK␥ functions as an adaptor to link the IKKs with the upstream regulators of NF-B (19,20). Here we show that vCLAP associates directly and specifically with IKK␥ through its C-terminal glycine-rich domain and may regulate the activity of the IKK complex through CARD-mediated oligomerization of IKK␥.
Biochemical Analysis-Immunoprecipitations were performed as described previously (20), and the precipitated proteins were analyzed by SDS polyacrylamide gel electrophoresis followed by immunoblotting. GST pull-down assays, luciferase reporter gene assays, and IKK kinase assays were performed as described (20,22).
Confocal Microscopy-293T cells were grown on coverslips and then transfected with the GFP-tagged IKK␥ or RFP-tagged vCLAP separately or together, with the indicated vectors. 24 h after transfection, cells were fixed with 4% paraformaldehyde in phosphate-buffered saline for 30 min. The coverslips were mounted on a glass slide, and the fluorescence was detected by confocal microscopy using an excitation wavelength of 488 nm and a detection wavelength of 522 nm (GFP) or an excitation wavelength of 568 nm and a detection wavelength of 585 nm (RFP). Images were Kalman-averaged with a Kalman filter to increase the signal/noise ratio.

RESULTS
vCLAP Interacts with the IKK Complex and Persistently Activates the IKK Kinases-Whereas in resting cells the IKK kinases are inactive, potent activators, such as TNF-␣, interleukin-1, or lipopolysaccharide, induce a very rapid IKK activation, detectable within minutes. However, numerous studies have shown that this activation is only transient and after ϳ30 min decreases to about 25% of its peak value (12,14,23). Using a luciferase reporter assay, we and others demonstrated that expression of vCLAP results in a robust activation of NF-B (1)(2)(3)(4). Because of the very high level of this activation, we asked whether vCLAP could persistently activate the IKKs, resulting in a sustained rather than transient activation of NF-B. To answer this question, endogenous IKK␣ was isolated by immunoprecipitation from extracts prepared from vCLAP-transfected or TNF-␣-treated HeLa cells and assayed for IKK catalytic activity using a GST-IB␣ fusion protein as a substrate. Consistent with previous observations, TNF-␣ stimulation of HeLa cells induced high but transient IKK␣ kinase activity (Fig. 1A); the activity, which was maximum after 10 min of stimulation, declined sharply with time and was barely detectable after 90 min of stimulation. However, compared with TNF-␣ stimulation, overexpression of vCLAP in HeLa cells induced a robust and sustained IKK␣ kinase activity in the absence of any external stimulation (Fig. 1B). IKK␣ protein expression was comparable in vCLAP-transfected and nontransfected cells, indicating that vCLAP activates endogenous IKK␣ by a post-translational mechanism. To determine whether vCLAP associates stably with the IKK complex, immunoprecipitates obtained using anti-IKK␣ or anti-FLAG antibodies were assayed for the presence of vCLAP or the IKK components, respectively. As shown in Fig. 1B, vCLAP was readily detected after precipitation of endogenous IKK␣. Moreover, IKK␣ and IKK␥ were also detected in immunocomplexes obtained after precipitation of vCLAP (Fig. 1C). The vCLAP immunoprecipitates also possessed IKK kinase activity (Fig.  1C). These results provide direct biochemical evidence that vCLAP associates stably with the IKK complex and is able to persistently activate the IKK kinases.
vCLAP Interacts Directly with and Requires IKK␥ for Activation of NF-B-To determine which component of the IKK complex interacts with vCLAP, 293T cells were transfected with expression vectors for FLAG-tagged vCLAP and T7-IKK␤ with or without T7-IKK␥. As shown in Fig. 2A, a small amount of IKK␤ was coimmunoprecipitated with vCLAP in the absence of ectopic IKK␥. However, a remarkably higher amount of IKK␤ was coimmunoprecipitated with vCLAP in the presence of coexpressed IKK␥ ( Fig. 2A). The ectopic T7-IKK␥ was also detected in these complexes ( Fig. 2A). No IKK␤ or IKK␥ were precipitated with the FLAG antibody in the absence of FLAG-vCLAP ( Fig. 2A). This result shows that IKK␥ mediates the interaction of vCLAP with the IKK complex.
To rule out the possibility that other proteins were necessary for the vCLAP-IKK␥ interaction, we analyzed the ability of a GST-IKK␥ fusion protein to associate with an in vitro-trans-lated 35 S-labeled vCLAP. In agreement with a direct interaction between vCLAP and IKK␥, 35 S-labeled vCLAP bound to the GST-IKK␥ fusion protein but not the GST control (Fig. 2B).
To address the physiological relevance of this finding, we transiently expressed vCLAP in wild type or IKK␥-deficient Rat-1 cells (15). In contrast to wild type Rat-1 cells, no NF-B activation was elicited in the IKK␥-deficient 5R cells after transfection with the vCLAP construct or treatment with TNF-␣ (Fig. 2C). This result provides genetic proof for the requirement of IKK␥ in vCLAP-induced activation of NF-B, confirming its role as a molecular adaptor in the assembly of the vCLAP-IKK complexes. The inability of vCLAP to induce NF-B activation in 5R cells cannot be attributed to defects in the NF-B pathway downstream of IKK␥, because transfection of these cells with IKK␥ can restore NF-B activation by Tax, which is expressed stably in this cell line (Ref. 15 and data not shown).
In contrast to vCLAP, Bcl10 was unable to interact with IKK␥ in vitro (data not shown). However, like vCLAP, Bcl10 was able to induce NF-B activation in Rat-1, but not in 5R cells (Fig. 2C), suggesting that Bcl10 could relay its signal to IKK␥ indirectly.
Mapping of the Interaction Domains of vCLAP and IKK␥-We next mapped the regions of vCLAP and IKK␥ that are required for their interaction. FLAG-tagged IKK␥ was expressed in 293T cells with T7-tagged full-length domain, CARD (residues 1-107), or CTD (residues 108 -311) of vCLAP. Extracts prepared from the transfected cells were immunoprecipitated with an anti-FLAG antibody, and the resulting immune complexes were analyzed by Western blotting with an anti-T7 antibody that recognizes the T7-vCLAP variants. Both the full-length domain and the CTD of vCLAP were able to bind to IKK␥ (Fig. 3A). In contrast, the CARD did not interact with IKK␥ (Fig. 3A). The same results were obtained using a GST-IKK␥ pull-down assay, which showed that the recombinant GST-IKK␥ fusion protein is able to bind the in vitro-translated 35 S-labeled full-length domain or the CTD of vCLAP, but not the CARD of vCLAP (not shown). Combined, these results show that the CTD of vCLAP mediates its interaction with IKK␥.
To extend the characterization of the vCLAP-IKK␥ interaction, we expressed T7-tagged vCLAP in 293T cells together with several FLAG-tagged full-length or truncated IKK␥. vCLAP was found to specifically associate with full-length IKK␥ but not with the C-terminally truncated IKK␥ (1-200) or IKK␥ (1-251) (Fig. 3B). Removal of the last 119 amino acids of IKK␥ strongly reduced its interaction with the vCLAP (Fig.  3B). Taken together, these data indicate that the interaction between vCLAP and IKK␥ involves sequences in the C-terminal region of IKK␥.
To confirm that vCLAP associates intercellularly with IKK␥ when the two proteins are coexpressed in the same cell, we cotransfected 293T cells with constructs encoding GFP-IKK␥ and RFP-vCLAP fusion proteins and then monitored the subcellular localization of these proteins by confocal microscopy. As shown in Fig. 3C, the two proteins exhibited different pat-terns of cellular localization when expressed alone. Whereas vCLAP exhibited a clear pattern of discrete and interconnecting cytoplasmic filaments, IKK␥ displayed a somewhat punctuate cytoplasmic or whole-cell distribution. However, coexpression of the two proteins resulted in redistribution of IKK␥ to the vCLAP filaments. This observation is consistent with a direct interaction between vCLAP and IKK␥.
vCLAP Mediates Activation of the IKKs through Oligomerization of IKK␥-Our data demonstrate that IKK␥ mediates the association of vCLAP with the IKK kinases and is required for vCLAP-induced activation of NF-B. This suggest that IKK␥ is an adaptor molecule with a C-terminal half that binds to NF-B activators like vCLAP (see above), RIP (18 -20, 24), or Tax (25)(26)(27) and an N-terminal half that is required for interaction with the IKK kinases (20, 28). vCLAP has a bipartite structure consisting of an N-terminal CARD and a C-terminal

FIG. 2. IKK␥ mediates the assembly of the vCLAP-IKK complexes and is required for vCLAP-induced NF-B activation. A,
293T cells were transfected with expression constructs for T7-IKK␤, T7-IKK␥, and FLAG-vCLAP as indicated. 24 h after transfection, cells were lysed, and the lysates immunoprecipitated with anti-FLAG antibody. The immunoprecipitates were immunoblotted (IB) with anti-T7 antibody. Expression of T7-IKK␤, T7-IKK␥, or FLAG-vCLAP was determined by immunoblotting with anti-T7 or anti-FLAG antibodies, respectively. B, in vitro interaction of vCLAP with GST-IKK␥. 35 S-Labeled vCLAP (lane 1, 10% input) was incubated with an equal amount of GST (lane 2) or GST-IKK␥ (lane 3) proteins bound to glutathione-Sepharose beads. Bound proteins were then eluted and analyzed by SDS-PAGE and autoradiography. C, Rat-1 cells or the IKK␥-deficient Rat-1 derivative (5R) cells were transfected with 5ϫ B-luciferase reporter together with either empty vector or vCLAP or Bcl10 expression construct. 24 h after transfection, cells were either left untreated or incubated with TNF-␣ for 5 h. Cells were then collected and lysed, and the luciferase activity in the cell lysates was determined. pRSC-LacZ was included in all transfection reactions to normalize the transfection efficiency. Mean values Ϯ S.E. are shown from three independent experiments performed in duplicate.

FIG. 3. Interaction domains of IKK␥ and vCLAP.
The CTD of vCLAP mediates its interaction with IKK␥. vCLAP, IKK␥, and their deletion mutants are represented on the top. A, 293T cells were transfected with FLAG-IKK␥ together with either empty vector or T7-tagged full-length (FL) vCLAP, vCLAP-CARD, or vCLAP-CTD expression constructs. 24 h after transfection, cells were lysed, and the lysates were immunoprecipitated (IP) with anti-FLAG antibody. The immunoprecipitates were immunoblotted (IB) with anti-T7 antibody. The expression of FLAG-IKK␥ and the different T7-CLAP chimeras was determined by immunoblotting with anti-FLAG or anti-T7 antibodies, respectively. B, IKK␥ interacts with vCLAP via its C terminus. 293T cells were transfected with T7-vCLAP and different FLAG-tagged fulllength (FL) or C-terminally truncated IKK␥ expression constructs. 24 h after transfection, cells were lysed, and the lysates were immunoprecipitated (IP) with anti-FLAG antibody. The immunoprecipitates were immunoblotted (IB) with anti-T7 antibody. The expression of T7-vCLAP and the different FLAG-IKK␥ truncated mutants was determined by immunoblotting with anti-T7 or anti-FLAG antibodies, respectively. LZ, leucine zipper. C, recruitment of IKK␥ into vCLAP filaments. RFP-tagged vCLAP and GFP-tagged IKK␥ were expressed either alone or together in 293T cells. 24 h after transfection, cells were then examined and photographed under a fluorescent microscope. domain that interacts with IKK␥ (see above). Interestingly, we and others have shown that the CARDs of cellular and viral CLAP proteins are important for homo-and heterotypic interactions (1)(2)(3)(4). It is therefore likely that CARD-mediated selfassociation of vCLAP could induce oligomerization of IKK␥ resulting in activation of the IKK complex. To test this hypothesis, we generated a fusion protein (CARD-IKK␥-⌬C) composed of the CARD of vCLAP linked to the N-terminal part (residues 1-200) of IKK␥ (IKK␥-⌬C) and determined its ability to induce NF-B activation. To rule out the possibility that the CARD-IKK␥-⌬C chimera functions through interaction with the endogenous IKK␥ protein, we examined its ability to activate NF-B in the IKK␥-deficient 5R cells. As shown in Fig. 4A, transient transfection of the CARD-IKK␥-⌬C chimera resulted in a large increase of NF-B activity in a dose-dependent man-ner. In contrast, neither the separate CARD of vCLAP nor IKK␥-⌬C were able to activate the NF-B when transfected at either low or high doses (Fig. 4A). Moreover, a single point mutation of a conserved residue in the CARD of vCLAP that abrogates homodimerization (L49R) (1) prevented NF-B activation by the chimeric protein (Fig. 4A). Similar results were obtained when the CARD of Bcl10 was used instead of vCLAP-CARD in the above experiments (data not shown).
We then tested whether enforced oligomerization of the CTD of vCLAP could induce NF-B activation. For this purpose, the CTD of vCLAP was fused to a 3-fold repeat of the FKBP12 polypeptide, which oligomerizes when it binds to the cell-permeable synthetic organic ligand AP1510 (29). As shown in Fig.  4B, transient tranfection of this construct into Rat-1 cells, but not in the IKK␥-deficient 5R cells, induced a large NF-B activation in a ligand-dependent manner. No NF-B activation was detected when the FKBP12-CTD construct was cotransfected with kinase-inactive IKK␤ (not shown) or after treatment of empty vector-transfected 293T cells with AP1510 (Fig.  4B). Taken together, these results demonstrate that vCLAPinduced oligomerization of IKK␥ is the triggering event leading to activation of the IKK complex.

DISCUSSION
Although previous studies have shown that the EHV-2-encoded vCLAP protein activates NF-B in mammalian cells (1,4), the mechanism by which vCLAP interfaces with the cell's NF-B-activating machinery remains unclear. In this paper, we report several observations that, when combined, provide a potential mechanism for vCLAP-induced NF-B activation. First, we show that the IKK kinases are persistently activated in vCLAP-expressing cells, which might explain the robust NF-B activity observed in vCLAP-transfected cells (1,4). Second, we demonstrate that vCLAP, via its C-terminal domain, interacts physically with the IKK complex through direct binding to the C-terminal part of IKK␥. Consistent with this observation, IKK␥ was found to be essential for vCLAP-induced activation of NF-B. Third, we demonstrate that vCLAP activates the IKK complex through oligomerization of IKK␥. Indeed, CARD-dependent clustering of the N-terminal part of IKK␥, which we have previously shown to interact with the IKK kinases (20), was able to activate NF-B. Moreover, enforced oligomerization of the CTD of vCLAP was able to induce a large increase in NF-B activity in Rat-1 cells but not in the IKK␥-deficient 5R cells. This mechanism of NF-B activation by vCLAP, namely oligomerization-induced activation of the IKK kinases via IKK␥, is reminiscent of the model we proposed for RIP-induced activation of NF-B after ligation of TNF-R1 (20). Based on this model, TNF-␣ stimulation induces binding of RIP to, and concomitant oligomerization of IKK␥, which in turn passes the oligomerization signal to the effector kinases, resulting in their activation through autophosphorylation of their T-loop serines (20). However, in contrast to vCLAP, which binds stably to the IKK complex, RIP releases the activated IKK complex after its oligomerization, resulting in transient rather than persistent activation. Therefore, our results illustrate the ubiquitous role of oligomerization in IKK activation.
Activation of NF-B by expression of a single viral protein has been described in several studies, indicating that infection with an intact virus is not always required for NF-B activation. One well documented example of this is the human T-cell leukemia virus-Tax protein (30 -33). Tax has been shown to interact with the IKK complex through direct interaction with IKK␥ (25)(26)(27). Moreover, Tax mutants defective in IKK␥ binding failed to activate NF-B (34). Interestingly, Tax has been shown to function as a dimerizer that stabilizes dimer formation in other proteins (35). It is therefore possible that Tax- induced oligomerization of IKK␥ is, as for vCLAP, the triggering event in the activation of the IKK kinases by Tax.
Recently, several independent groups have demonstrated that the cellular homologue of vCLAP, Bcl10 (also called cCLAP/CIPER/hE10/CARMEN), was also able to activate NF-B when expressed in cells, although at a lesser degree than vCLAP. Bcl10 requires IKK␥ for activation of NF-B (Fig.  2C). Our preliminary results suggest that cCLAP interacts with the IKK complex, as immunoprecipitation of the endogenous cCLAP results in isolation of the IKK components. However, in a GST-IKK pull-down assay, in vitro-translated 35 Slabeled cCLAP was not able to bind to recombinant GST-IKK␥ or GST-IKK␣ (data not shown), indicating that the cCLAP-IKK complex interaction could be indirect or regulated by a posttranslational modification of cCLAP, such as phosphorylation. Future studies will reveal the precise physiological function of cCLAP and how it activates the IKK complex.