Persistent Activation of NF-κB by the Tax Transforming Protein Involves Chronic Phosphorylation of IκB Kinase Subunits IKKβ and IKKγ*

The Tax transforming protein encoded by human T-cell leukemia virus type 1 (HTLV1) persistently activates transcription factor NF-κB and deregulates the expression of downstream genes that mediate cell cycle entry. We recently found that Tax binds to and chronically stimulates the catalytic function of IκB kinase (IKK), a cellular enzyme complex that phosphorylates and inactivates the IκB inhibitory subunit of NF-κB. We now demonstrate that the IKKβ catalytic subunit and IKKγ regulatory subunit of IKK are chronically phosphorylated in HTLV1-infected and Tax-transfected cells. Alanine substitutions at Ser-177 and Ser-181 in the T loop of IKKβ protect both of these IKK subunits from Tax-directed phosphorylation and prevent the induction of IκB kinase activity. Each of these inhibitory effects is recapitulated in Tax transfectants expressing the bacterial protein YopJ, a potent in vivo agonist of T loop phosphorylation. Moreover, ectopically expressed forms of IKKβ that contain glutamic acid substitutions at Ser-177 and Ser-181 have the capacity to phosphorylate a recombinant IKKγ substrate in vitro. We conclude that Tax-induced phosphorylation of IKKβ is required for IKKβ activation, phosphoryl group transfer to IKKγ, and acquisition of the deregulated IKK phenotype.


The Tax transforming protein encoded by human Tcell leukemia virus type 1 (HTLV1) persistently activates transcription factor NF-B and deregulates the expression of downstream genes that mediate cell cycle entry. We recently found that Tax binds to and chronically stimulates the catalytic function of IB kinase (IKK), a cellular enzyme complex that phosphorylates and inactivates the IB inhibitory subunit of NF-B. We now demonstrate that the IKK␤ catalytic subunit and IKK␥ regulatory subunit of IKK are chronically phosphorylated in HTLV1-infected and Tax-transfected cells. Alanine substitutions at Ser-177 and Ser-181 in the T loop of IKK␤ protect both of these IKK subunits from
Tax-directed phosphorylation and prevent the induction of IB kinase activity. Each of these inhibitory effects is recapitulated in Tax transfectants expressing the bacterial protein YopJ, a potent in vivo agonist of T loop phosphorylation. Moreover, ectopically expressed forms of IKK␤ that contain glutamic acid substitutions at Ser-177 and Ser-181 have the capacity to phosphorylate a recombinant IKK␥ substrate in vitro. We conclude that Tax-induced phosphorylation of IKK␤ is required for IKK␤ activation, phosphoryl group transfer to IKK␥, and acquisition of the deregulated IKK phenotype.
To initiate an adaptive immune response, T lymphocytes execute a signal transduction program that triggers the transient induction of transcription factor NF-B, activation of downstream growth-related genes, and cell cycle entry (1). Part of this program is regulated from the cytoplasm by IB␣, an inhibitory subunit of NF-B, and an inducible IB kinase called IKK 1 (2). In response to immune system cues such as the cytokine tumor necrosis factor-␣ (TNF), IKK phosphorylates IB␣ at Ser-32 and Ser-36 (2). In turn, phosphorylated IB␣ is degraded and NF-B translocates to the nucleus (2).
The most well-characterized form of IKK contains two catalytic subunits, termed IKK␣ and IKK␤, and a regulatory subunit called IKK␥ (NEMO) (2). In response to TNF, IKK␤ is rapidly phosphorylated, activated, and down-regulated within 30 min (3). The relevant phosphoacceptors in IKK␤ have been mapped to a region in its catalytic domain that shares strong homology with "T loop" regulatory sequences found in members of the mitogen-activated protein kinase kinase (MAP2K) family of enzymes (3). Consistent with this structural link, members of the MAP2K kinase (MAP3K) family of enzymes have been implicated in TNF-induced activation of IKK (4).
In contrast to their transient action in TNF-treated cells, IKK and NF-B are constitutively activated in T lymphocytes infected with human T-cell leukemia virus type 1 (HTLV1) (5). This process is mediated by the Tax transforming protein of HTLV1 and appears to play an essential role in the pathogenesis of HTLV1-associated disease (5). Chronic stimulation of IKK catalytic activity by Tax is dependent on IKK␥, which directs the assembly of Tax⅐IKK complexes (5). Tax also binds to and activates MEKK1, a MAP3K that phosphorylates IKK in vitro (6). The dual specificity of Tax for these two enzymes may promote chronic phosphorylation of IKK and acquisition of the deregulated IKK phenotype. However, the phosphorylation status of IKK in Tax-expressing cells has not been examined.
We now demonstrate that IKK␤ and IKK␥ are chronically phosphorylated in Tax-expressing cells and HTLV1-infected T lymphocytes. Alanine replacements at Ser-177 and Ser-181 in the T loop of IKK␤ inhibit Tax-directed phosphorylation of both subunits and block the chronic stimulatory effects of Tax on IB kinase activity. Moreover, Tax-induced phosphorylation of IKK␤ and IKK␥ is antagonized by the bacterial protein YopJ, a potent in vivo inhibitor of T loop phosphorylation and IKK␤ catalytic activity (7). The finding that Tax-induced phosphorylation of IKK␥ is contingent upon IB kinase activity may reflect IKK␤-mediated phosphorylation of IKK␥ within individual Tax⅐IKK signaling complexes. Consistent with this finding, constitutively active forms of IKK␤ have the capacity to phosphorylate a recombinant IKK␥ substrate in vitro.
Reporter gene activity was determined using a Promega Luciferase Assay Kit and a Turner Designs luminometer. All data were normalized to the activity of a cotransfected ␤-galactosidase expression vector.
Metabolic Radiolabeling-Following transfection (18 h), HeLa and 293T cells were labeled for 8 h with [ 32 P]orthophosphate (1 mCi/ml; ICN) in phosphate-free DMEM (Mediatech). Immunocomplexes were washed with ELB buffer containing 0.5 M NaCl and 1 M urea, followed by RIPA buffer. Jurkat and MT-2 T cells were labeled in phosphate-free DMEM for 8 h with [ 32 P]orthophosphate (2 mCi/ml). Cytoplasmic extracts were precleared with anti-FLAG M2 antibodies prior to immunoprecipitation with anti-IKK␥ antibodies. Resultant complexes were washed with ELB buffer containing 2 M urea followed by RIPA buffer. Phosphoproteins were resolved by SDS-PAGE, transferred to PVDF membranes, and analyzed by autoradiography and immunoblotting.

RESULTS AND DISCUSSION
Tax-dependent Phosphorylation of IKK in Vitro-Prior studies have established that IKK␥ directs the assembly of Tax⅐IKK complexes (5). However, the biochemical mechanism by which Tax stimulates IKK␤ catalytic activity in these complexes remains unknown. To explore the role of phosphorylation in Tax⅐IKK signaling, HeLa cells were transfected with combinations of expression vectors for IKK␤, IKK␥, and wild-type Tax. Parallel transfections were performed with mutants of Tax that are selectively defective for either CREB/ATF (Tax-M47) or NF-B (Tax-M22) activation (8). We then isolated IKK␤ complexes from recipient cell extracts by immunoprecipitation and subjected them to in vitro kinase assays in the presence of [␥-32 P]ATP. As shown in Fig. 1A (top panel), the phosphorylation status of IKK␤ was essentially unaffected by Tax in the absence of ectopic IKK␥ (lanes 1 and 2). Programming cells with this Tax docking subunit resulted in significant phospho-rylation of both IKK␤ and IKK␥ (lane 4). Similar results were obtained in experiments with Tax-M47 (lane 6), but not Tax-M22 (lane 5), consistent with their differing capacities to engage and activate IKK (18). This pattern of phosphorylation was not attributable to inefficient or variable protein expression (lower panels).
Phosphorylation of Ser-177 and Ser-181 in the T loop of IKK␤ is required for its activation by TNF (3). To explore the role of Ser-177 and Ser-181 in Tax-dependent phosphorylation of IKK␤, HeLa cells were transfected with vectors for Tax, IKK␥, and either wild-type IKK␤ or a mutant containing alanine replacements at these two sites (IKK␤.SA). We then prepared IKK␤ (Fig. 1B) or Tax (Fig. 1C) immunoprecipitates for in vitro kinase assays. As shown in Fig. 1B, top panel, mutations affecting Ser-177 and Ser-181 in IKK␤ completely blocked its phosphorylation in the presence of Tax (lanes 3 and 6). These mutations also prevented Tax-induced phosphorylation of IKK␥. Similar results were obtained with a kinase-dead mutant of IKK␤ that is defective for ATP binding (IKK␤.KM, lanes 7-9) (10). These mutations had no significant effect on IKK␤ protein levels (middle panel). However, we detected a significant shift in the electrophoretic mobility of IKK␥ when coexpressed with Tax and wild-type IKK␤, consistent with a change in its phosphorylation status (bottom panel). All of these findings were recapitulated with Tax immunoprecipitates (Fig. 1C), indicating that the kinase activity responsible for IKK␤ and IKK␥ subunit phosphorylation is stably associated with Tax. We conclude that Ser-177 and/or Ser-181 in the T loop of IKK␤ are required for Tax-directed phosphorylation of IKK␤ and IKK␥ in vitro.
Tax-dependent Phosphorylation of IKK in Vivo-We next used [ 32 P]orthophosphate to metabolically label endogenous IKK in MT-2 cells. This transformed T lymphocyte line is chronically infected with HTLV1 and expresses high constitutive levels of IB kinase activity (18). Parallel experiments were conducted with Jurkat T lymphocytes, which are transformed by an HTLV1-independent mechanism. Endogenous IKK was immunoprecipitated from the corresponding extracts, fractionated by SDS-PAGE, and analyzed by autoradiography. As shown in Fig could not be attributed to cell type-specific differences in the steady-state level of IKK protein expression (lower panels, lanes 3 and 4). These findings demonstrate that endogenous IKK␤ and IKK␥ are chronically phosphorylated in the physiologically relevant setting of HTLV1-infected T cells.
To extend these findings, HeLa cells were transfected with various combinations of Tax and IKK expression vectors and then metabolically labeled with [ 32 P]orthophosphate. Ectopic IKK␤ was isolated by immunoprecipitation, fractionated by SDS-PAGE, and analyzed by autoradiography. As shown in Fig. 2B, top panel, Tax induced significant phosphorylation of IKK␤ and IKK␥ in cells programmed with the wild-type cata-lytic subunit (lanes 1 and 2). In contrast, Tax-directed phosphorylation of both subunits was blocked in cells expressing either IKK␤.SA (lanes 3 and 4), which has alanine replacements at Ser-177/Ser-181, or the kinase-dead mutant IKK␤.KM (lanes 5 and 6). Comparable amounts of IKK␤ and IKK␥ protein were detected in all of the samples analyzed (Fig.  2B, middle and lower panels). These in vivo results correlated strongly with the in vitro phosphorylation data shown in Fig. 1B.
To explore the functional consequences of IKK␤ phosphorylation, IKK␤ complexes were immunopurified from HeLa cell transfectants expressing Tax, IKK␥, and either wild-type IKK␤

FIG. 2. Tax induces phosphorylation of IKK␤ and IKK␥ in vivo.
A, Jurkat and MT-2 T cells were radiolabeled with [ 32 P]orthophosphate for 8 h. Cytoplasmic extracts (400 g) were subjected to immunoprecipitation with the indicated monoclonal antibodies. Resultant complexes were fractionated by SDS-PAGE and analyzed for 32 P incorporation (top panel). IKK protein content was determined by immunoblotting with subunit-specific antibodies (middle and lower panels). B, HeLa cells (1 ϫ 10 6 ) were transfected with vectors for Tax (100 ng), murine IKK␥ (Myc epitope-tagged, 100 ng), and FLAG-tagged forms of either wild-type IKK␤ (WT), IKK␤.SA, or IKK␤.KM (50 ng each). Cells were radiolabeled with [ 32 P]orthophosphate for 8 h. Ectopic IKK␤ complexes were isolated from cytoplasmic extracts using anti-FLAG M2-agarose beads and fractionated by SDS-PAGE. Resolved proteins were subjected to autoradiography (top panel) and immunoblotting with IKK subunit-specific antibodies (middle and bottom panels). C, HeLa cells were transfected as described in B. Ectopic IKK␤ was isolated from cytoplasmic extracts with anti-FLAG M2-agarose beads and assayed for IB kinase activity in the presence of GST-IB␣ (1 g) and [␥-32 P]ATP. Phosphoproteins were resolved by SDS-PAGE and visualized by autoradiography (top panel). Relative levels of IKK␤ protein were determined by immunoblotting with IKK␤-specific antibodies (bottom panel).

FIG. 3. YopJ prevents Tax-induced activation of IKK and NF-B.
A, 293T cells (5 ϫ 10 5 ) were transfected with a Tax expression vector (1 g), the indicated amounts of an effector plasmid encoding FLAG-tagged YopJ, and either the NF-B-Luc or HTLV1 LTR-Luc reporter plasmid (100 ng). Whole cell extracts were prepared after 24 h and assayed for luciferase activity. Average values obtained from four replicates are reported as the mean -fold induction (ϮS.E.) of luciferase activity by Tax relative to basal expression of the reporter gene in Tax-deficient cells. In the absence of Tax, YopJ-induced changes in the basal activity of either reporter plasmid was essentially negligible (Ͻ2-fold). B, 293T cells (1 ϫ 10 6 ) were transfected with expression plasmids for Tax (2 g) and FLAG-tagged YopJ (0.5 g) as indicated. Endogenous IKK complexes were immunoprecipitated with monoclonal anti-IKK␥ antibodies (PharMingen) and assayed for IB kinase activity as described in the Fig. 2 legend (top  panel). IKK␤ protein levels were monitored by immunoblotting with polyclonal anti-IKK␤ antibodies (second panel). Levels of Tax and YopJ protein expression were determined by immunoblotting cytoplasmic extracts with anti-Tax or anti-FLAG antibodies (lower two panels). C, 293T cells (1 ϫ 10 6 ) were transfected with vectors for HA-tagged IKK (25 ng), Tax (0.5 g), human IKK␥ (T7-tagged, 25 ng), and YopJ (100 ng). Cells were radiolabeled with [ 32 P]orthophosphate for 8 h. Ectopic IKK␤ complexes were isolated using anti-HA antibodies, washed at high stringency, and fractionated by SDS-PAGE. Resolved proteins were subjected to sequential autoradiography (top panel) and immunoblotting (middle and bottom panels).
or IKK␤.SA. These complexes were then monitored for IB kinase activity in vitro using a recombinant IB␣ substrate (GST-IB␣). As shown in Fig. 2C, upper panel, Tax potently induced the catalytic activity of wild-type IKK␤ via an IKK␥dependent mechanism (lanes 1-3). In contrast, we were unable to detect IB kinase activity in IKK␤.SA immunoprecipitates (lanes 4 -6). These differences in Tax responsiveness were significant, because wild-type IKK␤ and IKK␤.SA were comparably expressed at the protein level (lower panel). Given that IKK␤.SA escapes from phosphorylation in Tax-expressing cells (Fig. 2B), we conclude that Tax-induced phosphorylation of the IKK␤ catalytic subunit is required for acquisition of constitutive IB kinase activity.
YopJ Interferes with Tax⅐IKK Signaling-The bacterial virulence factor YopJ binds to multiple MAP2K proteins and interferes with T loop phosphorylation (7). In keeping with the structural link between MAP2K and IKK proteins, YopJ also forms complexes with IKK␤ (7). To determine whether YopJ affects the Tax/NF-B signaling axis, 293T cells were transfected with an NF-B reporter plasmid (NF-B-Luc) along with expression vectors for Tax and YopJ. As shown in Fig. 3A, left panel, Tax potently stimulated NF-B-directed transcription in the absence of YopJ. However, coexpression with YopJ blocked this Tax response in a dose-dependent fashion. In contrast, YopJ failed to inhibit Tax-induced activation of the HTLV1 LTR (right panel), which involves the transcriptional action of CREB/ATF rather than NF-B (8).
To examine the mechanism of YopJ action on the Tax/NF-B axis, 293T cells were transfected with vectors for Tax and YopJ, either alone or in combination. We then purified endogenous IKK complexes from recipient cells and monitored them for IB kinase activity. As shown in Fig. 3B, top panel, Tax potently induced the catalytic activity of IKK in the absence of YopJ (lanes 1 and 2). In contrast, endogenous IKK complexes isolated from cells coexpressing Tax and YopJ failed to affect GST⅐IB␣ phosphorylation (lane 4). Immunoblotting experiments confirmed that comparable amounts of endogenous IKK␤ were co-immunoprecipitated under each condition and that YopJ and Tax were both expressed efficiently (lower two panels).
To determine whether YopJ interferes with Tax-induced phosphorylation of IKK, 293T cells expressing ectopic IKK, Tax, and YopJ were radiolabeled with [ 32 P]orthophosphate. We then isolated IKK␤ complexes and analyzed their phosphoprotein content. As shown in Fig. 3C, top panel, phosphorylation of IKK␤ and IKK␥ was significantly increased in the presence of Tax relative to the level of subunit radiolabeling detected in Tax-deficient cells (lanes 3 and 4). Coexpression with YopJ completely blocked this Tax-dependent increase in IKK phosphorylation (lane 5). Given the capacity of YopJ to prevent T loop phosphorylation (7), these findings provide further evidence indicating that Ser-177 and Ser-181 in the T loop of IKK␤ function as Tax-responsive phosphoacceptors.
Phosphorylation of IKK␥ by IKK␤-The finding that IKK␥ phosphorylation induced by Tax is dependent on the catalytic activity of IKK␤ (Figs. 1-3) led us to hypothesize that IKK␥ is a substrate of IKK␤. To test this possibility, 293T cells were transfected with an expression vector encoding a constitutively active mutant of IKK␤ that contained glutamic acid substitutions at Ser-177 and Ser-181 (IKK␤.SE) (10). IKK␤.SE immunoprecipitates derived from these transfectants were incubated with [␥-32 P]ATP and a substrate containing GST fused to fulllength IKK␥ (GST⅐IKK␥). As shown in Fig. 4 (top, lane 2), GST⅐IKK␥ phosphorylating activity was readily detected in IKK␤.SE immunocomplexes. Removal of the IKK␥ sequences from the GST⅐IKK␥ fusion protein eliminated phosphoryl group transfer, thus confirming specificity (middle, lane 2). IKK␥ kinase activity was not detected in immunoprecipitates derived from cells expressing a kinase-deficient mutant of IKK␤ (IKK␤.SA) (top, lane 3), indicating that phosphorylation of IKK␥ is dependent on the catalytic function of IKK␤. Moreover, the IKK␥ kinase activity detected in IKK␤.SE immunocomplexes was retained after high stringency washing with 3 M urea (top, lane 5), excluding the involvement of a kinase that associates loosely with IKK␤.
In summary, our data indicate that the assembly of Tax⅐IKK complexes leads to chronic phosphorylation of IKK␤ and IKK␥. Activation of IKK␤ by Tax appears to involve a kinase that phosphorylates the T loop of IKK␤ at Ser-177 and Ser-181. Given that kinase-dead mutants of IKK␤ are defective for Tax-induced phosphorylation, this kinase may be IKK␤. Point mutations in IKK␤ that disrupt its catalytic function also prevent Tax-induced phosphorylation of IKK␥, suggesting that IKK␥ is phosphorylated by IKK␤ within the same complex. In keeping with this proposal, IKK␤ has the capacity to phosphorylate a recombinant IKK␥ substrate in vitro. As such, further studies are warranted to define whether the phosphorylation status of IKK␥ affects the temporal regulation of IB kinase activity.