Positive regulation of IkappaB kinase signaling by protein serine/threonine phosphatase 2A.

Transcription factor NF-kappaB plays a key regulatory role in the cellular response to pro-inflammatory cytokines such as tumor necrosis factor-alpha (TNF). In the absence of TNF, NF-kappaB is sequestered in the cytoplasm by inhibitory IkappaB proteins. Phosphorylation of IkappaBby the beta-catalytic subunit of IKK, a multicomponent IkappaB kinase, targets the inhibitor for proteolytic destruction and facilitates nuclear translocation of NF-kappaB. This pathway is initiated by TNF-dependent phosphorylation of T loop serines in IKKbeta, which greatly stimulates IkappaB kinase activity. Prior in vitro mixing experiments indicate that protein serine/threonine phosphatase 2A (PP2A) can dephosphorylate these T loop serines and inactivate IKK, suggesting a negative regulatory role for PP2A in IKK signaling. Here we provided several in vivo lines of evidence indicating that PP2A plays a positive rather than a negative role in the regulation of IKK. First, TNF-induced degradation of IkappaB is attenuated in cells treated with okadaic acid or fostriecin, two potent inhibitors of PP2A. Second, PP2A forms stable complexes with IKK in untransfected mammalian cells. This interaction is critically dependent on amino acid residues 121-179 of the IKKgamma regulatory subunit. Third, deletion of the PP2A-binding site in IKKgamma attenuates T loop phosphorylation and catalytic activation of IKKbeta in cells treated with TNF. Taken together, these data provide strong evidence that the formation of IKK.PP2A complexes is required for the proper induction of IkappaB kinase activity in vivo.

Insults to the immune system and various forms of cellular stress activate the transcription factor NF-B 3 and other dimeric members of the Rel polypeptide family (reviewed in Refs. 1 and 2). NF-B regulates the expression of multiple genes involved in the control of cell growth, division, and survival. A variety of stimuli can activate NF-B-mediated gene transcription, including tumor necrosis factor-␣ (TNF), interleukin-1, T and B cell mitogens, bacterial products, viral proteins, doublestranded RNA, as well as physical and chemical stress. Most of these agonists converge on a latent, cytoplasmic form of NF-B that associates with IB␣ or other members of the inhibitory IB family. Following cellular stimulation, IB␣ is phosphorylated, ubiquitinated, and degraded by the 26 S proteasome. In turn, NF-B is free to translocate to the nuclear compartment where it activates transcription units containing the B-binding site (1,2).
Phosphorylation of IB is catalyzed by a multicomponent protein kinase termed the IKK signalsome (3). Within the prototypic complex are two highly homologous IB kinase subunits, termed IKK␣ and IKK␤, which form homo-or heterodimers via their leucine zippers (4 -6). In response to various cell stimuli, IKK␣ and IKK␤ are activated by a mechanism involving phosphorylation of their respective T loops at Ser-176/Ser-180 and Ser-177/Ser-181 (7). T loop phosphorylation occurs via IKK subunit trans-autophosphorylation and/or the action of upstream kinases (3). In addition to phosphorylation of T loop serines, activated IKK␣ and IKK␤ also undergo autophosphorylation at a C-terminal serine cluster, resulting in decreased kinase activity (7). Despite extensive study of these kinases, the exact molecular mechanisms involved in the activation and post-inductive repression of IKK are poorly understood.
Signal-induced activation of IKK␤ is dependent on its interaction with the regulatory subunit IKK␥, which lacks a kinase domain (8 -10). The N-terminal half of IKK␥ is required for its assembly into a functional IB kinase complex containing IKK␣ and IKK␤. In contrast, the C-terminal half of this regulatory subunit is required for the induction of IB kinase activity, presumably by serving as an interaction surface for upstream signal transducers. For example, IKK␥ serves as a molecular adaptor for the Tax oncoprotein of human T cell leukemia virus, type 1, which stimulates persistent activation of IKK (11). However, the precise molecular composition of the IKK signalsome remains unclear in terms of the cellular proteins that modulate its transient activity following cellular stimulation with physiologic agonists of NF-B (12,13).
The multilevel regulation of IKK by phosphorylation indicates that protein phosphatases may play an important role in IKK signaling. Consistent with this possibility, prior studies with purified proteins have indicated that protein serine/threonine phosphatase 2A (PP2A) inhibits IB kinase activity in vitro (14). The predominant form of PP2A in cells is a heterotrimeric holoenzyme consisting of a catalytic C subunit (PP2A C ), a structural A subunit, and a variable B subunit (reviewed in Refs. 15 and 16). PP2A associates with many cellular proteins, including cytoskeletal components (17,18), receptors (19,20), monoamine transporters (21), metabolic enzymes (22), transcription factors (23,24), and several viral proteins (reviewed in Ref. 25). More recent studies indicate that PP2A forms stable complexes with a growing set of cellular protein kinases, presumably to facilitate substrate recognition and rapid signal-dependent changes in their phosphorylation status (reviewed in Ref. 26).
Here we provide several in vivo lines of evidence indicating that PP2A binds to IKK and modulates its catalytic activity. Specifically, we show that TNF-induced degradation of IB is attenuated in cells treated with okadaic acid, an inhibitor of PP2A. PP2A and IKK derived from mammalian cell extracts co-immunoprecipitate and fractionate together on both kinase and phosphatase affinity resins. This phosphatase/kinase interaction is mediated by the IKK␥ regulatory subunit. Deletion of the PP2A-binding site in IKK␥ (amino acids 121-179) attenuates T loop phosphorylation and catalytic activation of IKK␤ induced by enforced cellular expression of either the Tax oncoprotein or the type 1 receptor for TNF. Furthermore, in sharp contrast to wild type IKK␥, IKK␥ lacking the PP2A-binding site is unable to restore TNF-induced phosphorylation of IKK␤, degradation of IB␣, and nuclear translocation of NF-B when the mutant is stably introduced into IKK␥-deficient fibroblasts. Taken together, these data provide strong evidence that the formation of IKK⅐PP2A complexes is required for TNF-dependent phosphorylation and activation of IKK in vivo.

MATERIALS AND METHODS
Reagents-Polyclonal IKK subunit antibodies (H744, H470, FL-419, and M280) and the p65 antibody were purchased from Santa Cruz Biotechnology, and the monoclonal IKK subunit (␣, ␤, and ␥) antibodies were from Pharmingen. Monoclonal antibodies recognizing the T7 epitope and agarose conjugates of the T7 antibody were obtained from Novagen. The PP2A C monoclonal antibodies were from BD Transduction Laboratories or Upstate Biotechnology, Inc. A phospho-specific antibody recognizing IKK␣ and IKK␤ phosphorylated at Ser-180 and Ser-181, respectively, was purchased from Cell Signaling Technology. Anti-FLAG M2 monoclonal antibodies were purchased from Sigma. Rabbit anti-Tax antibodies were provided by Dr. Bryan Cullen (Duke University). Normal rabbit IgG, normal mouse IgG, and secondary antibodies for alkaline phosphatase colorimetric detection were obtained from The Jackson Laboratory; secondary antibodies for fluorescence detection were obtained from Rockland or Molecular Probes. Horseradish peroxidase-conjugated secondary antibodies were obtained from Pierce. Mammalian expression plasmids for Tax, TNF-R1, and IKK subunits have been described previously (27,28). Expression plasmids for IKK␥ used in Fig. 4 were kindly provided by Dr. M. Horwitz (Albert Einstein College of Medicine). pBABE-Puro, pCL-Ampho, and pHSCMV-VSVg were gifts from Dr. Christopher Aiken (Vanderbilt University). PolyFect transfection reagent and plasmid purification kits were purchased from Qiagen. The nuclear and cytoplasmic extraction kit was purchased from Active Motif. Okadaic acid, fostriecin, and microcystin were obtained from Alexis Biochemicals, and microcystin-Sepharose was obtained from Upstate Biotechnology, Inc. ATP-Sepharose was a gift from Dr. Timothy Haystead (Duke University) or purchased from Upstate Biotechnology, Inc. [␥-32 P]ATP was obtained from ICN. Western blot Blocking Buffer and protein G-Sepharose were obtained from Zymed Laboratories Inc. Phenyl-Sepharose, MonoQ, and Superdex-200 columns and gel filtration standards were from Amersham Biosciences. Centricon-30 filters were obtained from Millipore.
B Lymphocytes, Cell Culture, and Transfections-Primary B lymphocytes were isolated from spleens of C57Bl6 mice and purified by AutoMACS (Miltenyi Biotec) by a negative selection protocol as described previously (29,30). All mice that were used as the source of splenocytes were treated humanely in accordance with federal and state government guidelines, and their use was approved by the Institutional Animal Care and Use Committee (Vanderbilt University). Jurkat T cells were maintained in Roswell Park Memorial Institute (RPMI) media supplemented with 10% fetal bovine serum, 2 mM glutamine, and antibiotics. Murine embryonic fibroblasts (MEFs) derived from mice lacking IKK␥ were a gift from Dr. Joseph DiDonato (Cleveland Clinic) and have been described previously (31). HeLa cells, human embryonic kidney-293T cells (HEK-293T), and MEFs were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 2 mM glutamine, and antibiotics. HEK-293T and HeLa cells were transfected using PolyFect according to the manufacturer's recommended protocol or by the calcium phosphate method as described previously (27).
Generation of T7-tagged IKK␥ Internal Deletion Mutant of IKK␥ Lacking Amino Acids 121-179-Amino acids 121-179 were deleted from IKK␥ by PCR amplification of the flanking regions, with an engineered overhang, followed by a second round of PCR in which the two flanking regions were allowed to anneal, and the resulting sequence was amplified. PCR amplification of the sequence encoding the T7 tag and first 120 amino acids of IKK␥ (with a 3Ј overhang) was accomplished by using the sense primer 5Ј-ATGGGTCGCATGGATCCG-3Ј, the antisense primer 5Ј-GCCGCGCCTCCTTCTGCCTC-3Ј, and T7-tagged full-length IKK␥ (T7-IKK␥)/pcDNA3 (28) as a template; the second region of IKK␥ encoding amino acid residues 180 -419 (with a 5Ј overhang) was amplified in another PCR by using the sense primer 5Ј-GAAGGAGGCGCGGCAGCTGG-3Ј, the antisense primer 5Ј-TC-CGCCTTGTAGATATCCG-3Ј, and T7-IKK␥/pcDNA3 as a template. The PCR products were allowed to anneal and then amplified by using the outside primers 5Ј-ATGGGTCGCATGGATCCG-3Ј and 5Ј-TC-CGCCTTGTAGATATCCG-3Ј. The resulting PCR product of IKK␥, lacking the sequence encoding amino acids 121-179, was digested with BamHI and EcoRV and ligated back into BamHI/EcoRV-digested T7-IKK␥/pcDNA3. Proper construction of the plasmid encoding the deletion mutant was verified by automated DNA sequencing (Vanderbilt University DNA Core Facility).
Generation of MEF Cell Lines Stably Expressing T7-tagged IKK␥ Constructs-Viral stocks were produced by transfecting HEK-293T cells (35-mm dish) with 3 g of pBABE-Puro encoding the IKK construct, 3 g of pCL-Ampho, and 1.2 g of pHSCMV-VSVg (32,33). Supernatants containing fully packaged retrovirus were recovered 48 h after transfection and were immediately used for transduction. Transduction was achieved by incubating IKK␥-deficient MEFs with the appropriate viral stock in the presence of Polybrene (0.8 g/ml). Supernatants were removed 8 h post-infection, and the cells were cultured an additional 48 h in complete DMEM. Transduced cells were selected in DMEM containing 2 g/ml puromycin.
Immunoblotting-Protein samples were separated on SDS-polyacrylamide gels (10%) and electrophoretically transferred to nitrocellulose in 10 mM CAPS, pH 11, containing 10% methanol (1 h at 1 A). Proteins on the membrane were visualized with Ponceau S, followed by washing in TTBS (25 mM Tris-HCl, pH 7.4, 137 mM NaCl, 3 mM KCl, and 0.2% Tween 20). The membrane was blocked in TTBS containing 0.5% bovine serum albumin or in Blocking Buffer (Zymed Laboratories Inc.), followed by incubation with the diluted primary antibody. After washing, the membranes were then incubated with alkaline phosphatase-or fluor-conjugated secondary antibodies; bound antibodies were visualized by colorimetric detection or via the Odyssey Infrared Imaging system (LiCor). Alternatively, immunoblotting was performed using horseradish peroxidase-conjugated secondary antibodies and an enhanced chemiluminescence detection system from Amersham Biosciences.
Fast Performance Liquid Chromatography-Jurkat T cells (100 ml of cell suspension) were stimulated with PMA (50 ng/ml) and ionomycin (1 M) for 15 min. Cells were pelleted and lysed in 12 ml of Buffer A (20 mM Tris-HCl, pH 7.5, 20 mM ␤-glycerol phosphate, 5 mM Na 4 P 2 O 7 , 10 mM NaF, 0.5 mM Na 3 VO 4 , 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 10% glycerol, and 0.25% Triton X-100) containing 0.1 mM phenylmethylsulfonyl fluoride, 2 g/ml leupeptin, 1 g/ml pepstatin, 2 g/ml aprotinin, and 2 mM benzamidine. The clarified cell lysates obtained following centrifugation at 12,000 ϫ g were adjusted to 1 M ammonium sulfate and applied to a phenyl-Sepharose column (20 ml) equilibrated in Buffer A containing 1 M ammonium sulfate and reduced protease inhibitor concentrations (400 ng/ml pepstatin, 200 ng/ml aprotinin, and 2 mM benzamidine). The column was developed (2 ml/min) with a linear gradient (1 to 0 M ammonium sulfate in the same buffer), and 4-ml fractions were collected. The peak fractions of IKK activity (fractions 34 -45) were pooled and applied to a MonoQ column (1 ml). The MonoQ column was developed (1 ml/min) using a linear gradient (0 to 0.5 M NaCl in Buffer A), and 1-ml fractions were collected. The peak fractions of IKK, which eluted from the MonoQ column between 370 and 400 mM NaCl (fractions 20 -22), were pooled and concentrated to 200 l using a Centricon-30. The concentrated sample was fractionated (0.5 ml/min) on an analytical Superdex-200 gel filtration column (20 ml) equilibrated in Buffer A containing 150 mM NaCl, and 0.5-ml fractions were collected. Aliquots of each column fraction were subjected to Western analysis or assayed for IKK activity.
ATP-Sepharose and Microcystin-Sepharose Affinity Purifications-Jurkat T cell lysates (3.5 mg of protein) in 1 ml of Buffer C (50 mM ␤-glycerol phosphate, pH 7.4, 1.5 mM EGTA, 0.15 mM Na 3 VO 4 , 1 mM DTT, 50 mM MgCl 2 , and 1% Nonidet P-40) were incubated with 50 l of a 50% slurry of ATP-Sepharose overnight at 4°C. ATP-Sepharose-bound proteins were washed five times with Buffer C and once with "wash buffer" (25 mM Tris-HCl, pH 7.4, 1 mM DTT, and 1% Nonidet P-40). After washing, bound proteins were eluted three times with 75 l of elution buffer containing 25 mM Tris-HCl, pH 7.4, 100 mM ATP, 1 mM DTT, 5 mM EGTA, and 1% Nonidet P-40 (ATP). The eluted material was then incubated with 50 l of a 50% slurry of microcystin-Sepharose overnight at 4°C. Microcystin-Sepharose-bound proteins were washed five times with wash buffer and eluted two times with 50 l of SDS sample buffer (ATP3 MC). Twenty-l aliquots of the ATP and ATP3 MC samples and a 20-l aliquot of Jurkat T cell extracts were subjected to SDS-PAGE and analyzed by silver staining or immunoblotting by using IKK␣-, IKK␤-, and PP2A C -specific antibodies.
Immunoprecipitation of epitope-tagged IKK subunits was accomplished by incubating cellular extracts (50 -60 g of protein) prepared in Buffer B with 10 -15 l of a 50% slurry of anti-T7-agarose or anti-FLAG M2-agarose beads for 2 h at 4°C. For immunoblot analyses, bound proteins were washed four times with ELB buffer (50 mM HEPES, pH 7.5, 5 mM EDTA, and 250 mM NaCl), eluted with SDS sample buffer, and subjected to SDS-PAGE. For kinase assays, bound proteins were washed sequentially with ELB buffer and kinase buffer (10 mM HEPES, pH 7.4, 5 mM MgCl 2 , 1 mM MnCl 2 , 2 mM NaF, and 50 M Na 3 VO 4 ) and assayed as described below.
For immunoprecipitation studies with primary B lymphocytes, lysates were prepared from 10 7 cells in RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 g/ml aprotinin, 1 mM Na 3 VO 4 , 1 mM NaF, 1% Triton X-100, and 0.1% SDS). Extracts were pre-cleared by incubating with 12 l of either protein G-Sepharose (PP2A C immunoprecipitations) or protein A-Sepharose (IKK immunoprecipitations) for 1 h at 4°C. The precleared cell extracts were incubated with 2 g of the PP2A C monoclonal antibody (BD Transduction Laboratories) and 12 l of protein G-Sepharose or with 2 g of the IKK␥ polyclonal antibody (FL-419, Santa Cruz Biotechnology) and 12 l of protein A-Sepharose for 12-14 h at 4°C. As controls, precleared extracts were also incubated with equivalent amounts of normal mouse IgG and protein G-Sepharose or normal rabbit IgG and protein A-Sepharose. Bound proteins were washed three times with ELB buffer, eluted with SDS sample buffer, and subjected to immunoblot analysis.
In Vitro Kinase Assays-IB kinase activity was measured in reaction mixtures containing recombinant glutathione S-transferase protein fused to amino acids 1-54 of IB␣(GST-IB␣), 10 M ATP, and 5 Ci of [␥-32 P]ATP, as described previously (14,34). A substrate containing alanine replacements for Ser-32 and Ser-36 of IB␣(GST-IB␣ MU) was used to assess kinase specificity. The kinase reactions were incubated at 30°C for 30 min and then terminated by heat denaturation in the presence of 1% SDS. Radiolabeled phosphoproteins were resolved by SDS-PAGE and visualized by autoradiography.
Microcystin-Sepharose Affinity Isolations-Soluble protein extracts (0.5 mg) of HEK-293T or HeLa cells expressing T7-tagged IKK␥ proteins were incubated with 30 l of a 50% slurry of microcystin-Sepharose overnight at 4°C. The beads were washed four times with Buffer A, and bound proteins were eluted with SDS sample buffer and subjected to immunoblot analysis. Soluble protein extracts of MEF cell lines (0.1 mg) were incubated with 20 l of a 50% slurry of microcystin-Sepharose in ELB buffer for 2 h at 4°C. The beads were washed four times with ELB buffer, and bound proteins were eluted with SDS sample buffer and subjected to immunoblot analysis.

RESULTS AND DISCUSSION
Effects of PP2A Inhibitors on IB Degradation-Several studies have implicated a negative regulatory role for PP2A in signal-dependent activation of IKK and NF-B (7,14,23,(35)(36)(37)(38)(39)(40)(41). For example, Delhase et al. (7) showed that PP2A inactivates purified IKK in vitro. Furthermore, several studies (37)(38)(39)(40)(41) have revealed that prolonged cell treatment (Ͼ45 min) with the PP2A inhibitor okadaic acid leads to NF-B activation in some cell types. However, cellular induction of NF-B activity in response to prolonged incubation with okadaic acid also coincides with cell death (42), making it difficult to distinguish whether the effects of this phosphatase inhibitor are the direct result of inhibition of phosphatase(s) regulating IKK/NF-B activities or an indirect result of cellular apoptosis. Although these earlier studies have linked PP2A to the control of NF-B activation, the precise in vivo role of PP2A in IKK/ NF-B signaling remains unclear.
To investigate this question further, we monitored the fate of IB␣ in HeLa cells following acute treatment with okadaic acid, the cytokine TNF, or both. As shown in control experiments, TNF rapidly induced the degradation of IB␣ followed by de novo synthesis of the inhibitor as described previously (Fig. 1A, top, lanes 1-6) (43). Of special interest, this TNF response was partially suppressed rather than enhanced following exposure to okadaic acid (Fig. 1A, top, lanes 7-11). Moreover, treatment with the PP2A inhibitor in the absence of TNF led to an increase rather than a decrease in IB␣ levels (Fig. 1A, top, lanes 12-16). The observed patterns of IB␣ protein expression could not be attributed to okadaic acid-induced changes in the steady-state level of IKK␣ and IKK␤ (Fig. 1A, bottom panels). These findings suggested that PP2A plays a positive regulatory role in the mechanism underlying TNF-induced degradation of IB␣.
To extend these findings with okadaic acid to a more specific inhibitor of PP2A, we performed similar experiments with fostriecin (44). In keeping with Fig. 1A, fostriecin failed to induce the degradation of IB in the absence of TNF (Fig. 1B, compare lanes 8-9 with lane 7) and partially suppressed TNF-induced IB degradation (Fig. 1B, compare lane 5 with  lane 2). The observed patterns of IB␣ protein expression could not be attributed to fostriecin-induced changes in the steady-state level of IKK␣ and IKK␤ (Fig. 1B, bottom panel). Thus, despite the capacity of PP2A to inactivate IKK in vitro, exposure of TNF-treated cells to an antagonist of PP2A impairs rather than enhances the degradation of IB␣. These results raised the intriguing possibility that PP2A functions as a positive regulator of the NF-B signaling pathway, perhaps at the level of IKK.   6 and 7, respectively). The resulting immune complexes were washed, eluted with SDS buffer, and subjected to immunoblot analysis using antibodies recognizing IKK␤ or PP2A C . B, immunoprecipitations were performed from extracts of primary B lymphocytes using an IKK␥ polyclonal antibody (FL-419, Santa Cruz Biotechnology) and protein A-Sepharose (lane 1) or a PP2A monoclonal antibody (BD Transduction Laboratories) and protein G-Sepharose (lane 3). As controls, the cell extracts were also incubated with equivalent amounts of normal rabbit IgG and protein A-Sepharose (lane 2) or normal mouse IgG and protein G-Sepharose (lane 4). The resulting immune complexes were washed, eluted with SDS buffer, and subjected to immunoblot analysis using antibodies recognizing IKK␣ or PP2A C . Immunoblot analyses of the corresponding cell lysates (10 -20 g of protein) are also shown.
PP2A Associates with IKK Complexes-To explore the molecular basis for these results with PP2A inhibitors, cellular extracts were fractionated and monitored for the presence of IKK⅐PP2A complexes. For these studies, Jurkat T cells were stimulated with combinations of PMA and ionomycin, which together activate IKK and NF-B (45). Stimulated cells were lysed in a buffer containing a mixture of phosphatase inhibitors to preserve kinase activity. Sequential fractionation of the IKK signalsome on phenyl-Sepharose, MonoQ, and Superdex-200 gel filtration columns was monitored by using an in vitro IB kinase assay (12). Immunoblot analysis of the gel filtration profile revealed a portion of the cellular PP2A pool in high molecular weight fractions containing IKK␣, IKK␤, and IKK␥ (ϳ700 -900 kDa; Fig. 2A, bottom panels). These fractions also contained peak levels of IB kinase activity ( Fig. 2A, top  panel). SDS-PAGE and silver stain analysis of the gel filtration fractions revealed a significantly enriched protein preparation as compared with earlier chromatographic steps (data not shown). Similar results were obtained using extracts from unstimulated Jurkat T cells and rat brain (data not shown), suggesting that cellular stimulation is not a prerequisite for co-fractionation of PP2A and IKK.
The finding that PP2A and IKK co-fractionate during multiple chromatographic steps was consistent with the presence of a stable protein kinase-phosphatase complex. In this regard, prior studies have estab-lished that protein kinases can be affinity-purified from crude extracts via their signature interaction with ATP (46). Indeed, DiDonato et al. (14) reported that IKK is efficiently captured on ATP-Sepharose. To test whether PP2A interacts with ATP-Sepharose, we incubated the resin with extracts of unstimulated Jurkat T cells, eluted bound polypeptides with 100 mM ATP, and then analyzed eluates via SDS-PAGE and immunoblotting. All three of the prototypical IKK subunits were readily detected in the ATP eluates (Fig. 2B, top panel, and data not shown), consistent with the multisubunit composition and function of the IB kinase signalsome. More importantly, in keeping with the proposed interaction of PP2A with IKK, the PP2A catalytic subunit (PP2A C ) coeluted with these IKK subunits in the presence of ATP (Fig. 2B, bottom  panel).
In reciprocal experiments, we investigated whether the IKK complexes isolated from cell extracts via ATP-Sepharose interact with an affinity resin for protein phosphatases. We have established previously that kinase-phosphatase complexes containing PP2A can be isolated from crude extracts using Sepharose beads bound to microcystin, a potent inhibitor of protein phosphatases (47)(48)(49)(50)(51)(52). Because microcystin-Sepharose engages the substrate-binding site of the PP2A catalytic subunit (53), PP2A catalytic activity does not appear to be required for its association with protein kinases. To test whether IKK subunits bind to  A, schematic representation of IKK␥ wild type (wt), IKK␥ ND200, and IKK␥(⌬121-179) constructs. B, the T7 tagged IKK␥ constructs were transiently expressed in HeLa cells, and protein expression was confirmed by immunoblotting the cell lysates (ϳ20 g of protein) with antibodies recognizing the T7 tag. Cell lysates were incubated with microcystin-Sepharose, and bound proteins (MC-Sepharose isolations) were analyzed by immunoblotting using the T7 antibody.
this phosphatase affinity resin, we incubated the ATP eluates with microcystin-Sepharose, fractionated absorbed polypeptides by SDS-PAGE, and then probed them for IKK subunit content on immunoblots. As shown in Fig. 2B (top panel), the IKK subunits co-eluted with PP2A C from the phosphatase affinity resin. These sequential chromatography experiments reinforce the model that IKK and PP2A form a stable macromolecular complex that can be captured using conventional fractionation methods for either protein kinases or phosphatases.
To confirm this model, we performed immunoprecipitation experiments using lysates from unstimulated Jurkat T cells and a monoclonal antibody that recognizes the C terminus of the PP2A catalytic subunit. PP2A immunocomplexes were fractionated by SDS-PAGE and monitored for protein content by immunoblotting with antibodies for PP2A and IKK. As shown in Fig. 3A (lane 2), IKK␤ was readily detected in PP2A immune complexes. Immunoprecipitation of these IKK␤⅐PP2A complexes was blocked by preincubation of PP2A-specific antibodies with a peptide corresponding to the C-terminal epitope in PP2A, confirming specificity (Fig. 3A, lane 3). Furthermore, stimulation of the cells with TNF prior to lysis had no significant effect on the amount of IKK␤ that co-immunoprecipitated with PP2A (Fig. 3A, lanes 6 and 7).
To determine whether PP2A forms a stable complex with IKK in nontransformed cells, we performed immunoprecipitations from extracts of primary murine B cells. As shown in Fig. 3B, IKK␣ and PP2A co-purify in reciprocal immunoprecipitations (Fig. 3B, lanes 1 and 3). Very little if any IKK␣ or PP2A was detected in the corresponding control immunoprecipitations (Fig. 3B, lanes 2 and 4). Taken together with the results shown in Fig. 2, A and B, these data provide multiple independent lines of evidence indicating that PP2A forms stable complexes with IKK in untransfected mammalian cells.
Identification of a Putative PP2A-binding Site in IKK␥-Given the evidence for a constitutive interaction between PP2A and IKK in mammalian cells (Figs. 2 and 3), we next performed studies to dissect the relevant binding mechanism. To determine whether PP2A interfaces with the regulatory subunit of IKK in the absence of cellular stimulation, we programmed HEK-293T cells with expression vectors for either fulllength IKK␥ or the deletion mutants shown in Fig. 4A. PP2A complexes were then purified from recipient cell extracts on microcystin-Sepharose (see Fig. 2B), fractionated by SDS-PAGE, and probed for IKK␥ content by immunoblotting. As shown in Fig. 4B, wild type IKK␥ was readily detected in these PP2A complexes (lane b). Deletion of the entire C-terminal half of IKK␥ (amino acids 180 -419), which is required for signal-induced activation of IKK␤, failed to prevent its association with the phosphatase affinity resin (Fig. 4B, lane i). Similar results were obtained with an IKK␥ construct lacking amino acids 1-100, which are required for the formation of IKK␥⅐IKK␤ complexes (Fig. 4B, lane c). However, further deletion from the N-terminal end of IKK␥ to amino acid 179 completely disrupted this interaction (Fig. 4B, lane d), indicating that the putative PP2A-binding site lies between amino acids 101 and 179. These results could not be attributed to effects of the mutations on steady-state levels of IKK␥ because all of the mutants were efficiently expressed in recipient cells as demonstrated by immunoblotting (Fig. 4B, right panel).
Role of PP2A in IKK Signaling-Previous studies have established that amino acids 1-120 of IKK␥ are responsible for the association of this regulatory subunit with IKK␣ and IKK␤ (28). In light of these studies and our mapping data (Fig. 4), we removed amino acids 121-179 from IKK␥ and introduced the corresponding deletion mutant in HeLa cells (IKK␥ ⌬121-179, Fig. 5A). Control transfections were performed with an expression vector for IKK␥ lacking amino acids 1-200 (ND200, Fig.  5A). In keeping with the results shown in Fig. 4, IKK␥(⌬121-179) was unable to bind microcystin-Sepharose, indicating a significant defect in its capacity to engage PP2A (Fig. 5B, lane 3).
The human T cell leukemia virus, type 1, Tax oncoprotein binds to IKK␥ and chronically activates IKK␤ (reviewed in Ref. 11). To confirm the structural integrity of IKK␥(⌬121-179), we transfected HEK-293T cells with combinations of expression plasmids for Tax, IKK␥, and IKK␤. IKK␤ complexes were then isolated from recipient cell extracts by immunoprecipitation and tested for the formation of IKK⅐Tax complexes by immunoblotting. As shown in Fig. 6A, the content of Tax was negligible in IKK␤ immune complexes derived from control cells lacking co-transfected IKK␥ (Fig. 6A, top panel, lane 2). However, IKK⅐Tax complexes were readily detected in cells programmed with either fulllength IKK␥ or IKK␥(⌬121-179) (Fig. 6A, lanes 4 and 6). We conclude that deletion of the PP2A-binding site in IKK␥ has no effect on its ability to recruit Tax into a higher order IB kinase complex.
Whereas Tax activates IKK via direct interactions, TNF receptormediated activation of IKK proceeds through several signaling intermediates. To investigate the role of PP2A in these two pathways, HEK-293T cells were co-transfected with expression plasmids for IKK subunits and either Tax or TNF receptor type 1 (TNF-R1). We then monitored recipient cell extracts for T loop phosphorylation of IKK␤ and IB kinase activity. As shown in Fig. 6B, enforced expression of either Tax or TNF-R1 in cells expressing wild type IKK␥ led to a significant increase in IKK␤ T loop phosphorylation (lanes 4 -6). Both of these signal-dependent responses were attenuated in the IKK␥(⌬121-179) cellular background, albeit to different degrees (Fig. 6B, lanes 7-9). Consistent with the observed defect in IKK signaling, IB␣ kinase activity stimulated by either Tax or TNF-R1 was significantly impaired in cells expressing the PP2A-binding site mutant of IKK␥ (Fig. 6C, top panel, lanes [7][8][9]. All of the ectopic proteins were efficiently expressed in these mammalian cell transfectants (Fig. 6, B and C, bottom panels). Thus, deletion of the PP2A-binding site in IKK␥ impairs Tax-induced activation of IKK and leads to a profound signaling defect in the TNF-R1/IKK axis.
Rescue Studies of IKK␥-deficient Cells-To confirm these findings with transiently transfected cells in a more physiologically relevant setting, wild type and mutant IKK␥ were stably introduced into IKK␥deficient MEFs for subsequent rescue experiments. Prior studies have established that TNF-induced activation of IKK␤ is blocked in IKK␥deficient MEFs, consistent with the essential role that IKK␥ plays in this process (54,55). As shown in Fig. 7A, both full-length IKK␥ and IKK␥(⌬121-179) were comparably expressed in MEFs (lanes 2 and 3). Consistent with the loss of PP2A binding, the IKK␥(⌬121-179) mutant was defective for binding to microcystin-Sepharose, whereas full-length IKK␥ interacted readily with the phosphatase affinity resin (Fig. 7B,  compare lanes 2 and 3). More importantly, both constructs retained the capacity to associate with endogenous IKK␣ and IKK␤ (Fig. 7C, lanes 2  and 3). Formation of higher order IKK complexes (ϳ400 -800 kDa) in MEFs stably expressing either wild type IKK␥ or IKK␥(⌬121-179) was also confirmed by gel filtration experiments (data not shown). Together, these data demonstrate that the wild type and PP2A-binding defective IKK␥ constructs were efficiently integrated into endogenous IKK complexes containing the IKK␣ and IKK␤ catalytic subunits.
We next monitored IKK signaling in reconstituted MEFs following treatment with the pro-inflammatory cytokine TNF. As shown in immunoblotting experiments with T loop-specific antibodies (Fig. 8A), IKK␤ was rapidly phosphorylated within 10 min following TNF stimulation (top panel, lane 5). In sharp contrast, signal-induced phosphorylation of IKK␤ was not apparent in cells expressing IKK␥(⌬121-179), even after stimulation for 1 h (Fig. 8A, lanes 7-9). These phosphorylation data correlated strongly with the TNF-inducible pattern of IB kinase activity in MEFs expressing wild type versus mutant IKK␥ (Fig.  8B, top panel, lanes 4 -9). Comparable levels of IKK␥ and IKK␣/␤ subunits were detected in these experiments (Fig. 8B, middle and bottom panels), thus ensuring that the observed differences in kinase activities were not because of TNF-induced changes in IKK subunit expression. These data with reconstituted MEFs provide strong evidence that PP2A plays a positive regulatory role in IKK signaling following cellular stimulation with the pro-inflammatory cytokine TNF.
To determine whether disruption of the IKK/PP2A interaction affects downstream steps in NF-B signaling, we next monitored the degradation of IB␣ and nuclear translocation of NF-B in TNF-treated MEFs. In keeping with the in vitro kinase experiments using a recombinant IB␣ substrate (Fig. 8B), steady-state levels of the endogenous IB␣ protein decreased following TNF stimulation and returned to near basal levels by 60 min in cells expressing wild type IKK␥ (Fig. 8C, top  panel, lanes 4 -6). In contrast, TNF-induced degradation of IB␣ was  3) were subjected to SDS-PAGE and immunoblotting with an IKK␥-specific antibody. B, microcystin-Sepharose isolations were performed from the lysates of these cells. Cell lysates and microcystin-Sepharose bound proteins were subjected to SDS-PAGE and immunoblotting with monoclonal antibodies recognizing T7 and PP2A C . C, T7-immunoprecipitations (T7 IPs) were performed on lysates from these cells. Cell lysates and T7 immune complexes were analyzed by SDS-PAGE and immunoblotting with a T7 monoclonal antibody (bottom panels) and IKK␣-and IKK␤-specific antibodies (top panels). mu, mutant. MEFs stably expressing empty vector (ev), wild type IKK␥ (wt), or IKK␥(⌬121-179) (mu) were treated with TNF␣ (20 ng/ml) for the indicated times. T7 immune complexes were isolated from extracts of these cells and analyzed by SDS-PAGE and immunoblotting with the specified antibodies. B, T7 immune complexes isolated from cells treated as described in A were assayed for kinase activity using the GST-IB␣ as the substrate. The kinase reaction mixtures were subjected to SDS-PAGE; the bottom portion of the gel was exposed to film (top panel), and the top portion of the gel was analyzed by immunoblotting with IKK␣ and IKK␤ antibodies (middle panel) and an IKK␥ antibody (bottom panel). C, cell lysates were prepared from cells treated as described in A and analyzed by SDS-PAGE and immunoblotting with the specified antibodies. D, IKK␥-null MEFs stably expressing wild type IKK␥ (wt) or IKK␥(⌬121-179) (mu) were treated with (ϩ) or without (Ϫ) TNF␣ (20 ng/ml) for 15 min. Nuclear and cytoplasmic extracts were obtained from these cells using with the active motif nuclear extract kit and analyzed by SDS-PAGE and immunoblotting with the p65 antibody. impaired in cells expressing IKK␥(⌬121-179) (Fig. 8C, top panel, lanes  7-9). As shown in Fig. 8D, IB␣ breakdown in MEFs expressing wild type IKK␥ led to mobilization of the p65/RelA subunit of NF-B to the nuclear compartment, whereas this translocation step was blocked in cells expressing the PP2A-binding mutant of IKK␥ (Fig. 8D, lanes 7 and  8). These functional results, together with data shown in Fig. 8, A and B, further reinforce the model that PP2A binds to IKK, facilitating the induction of IB kinase activity, targeted degradation of IB, and release of NF-B to its nuclear site of action.
Our studies indicate that amino acid residues 121-179 of IKK␥ mediate its interaction with PP2A either directly or via an intermediate(s). The PP2A-binding domain lies within the first coiled-coil domain of IKK␥ (CC1; amino acid residues 98 -196) (56), which has not been extensively characterized. The C-terminal coiled-coil domain (CC2; amino acid residues 254 -298) of IKK␥ is required for multimerization of this regulatory subunit (56,57), whereas the N terminus of IKK␥ appears to be crucial for its interaction with IKK␣ and IKK␤ (8,28,58,59). Consistent with these earlier reports, our studies indicate that amino acid residues 121-179 are not required for IKK␥ multimerization or for its interaction with IKK␣ and IKK␤ (Figs. 6, 7C, and 8, A and B). Instead, our studies support the idea that this region of IKK␥ is a novel protein/protein interaction domain, which facilitates the interaction of PP2A with the IKK complex, allowing this phosphatase to function as a positive regulator of IKK signaling.
In summary, we report evidence indicating that PP2A is required for signal-dependent activation rather than repression of IKK. We find that PP2A engages IKK via its interaction with IKK␥, a regulatory subunit that is essential for IKK␤ phosphorylation and activation. Removal of the putative PP2A-binding site in IKK␥ impairs TNF-induced activation of IKK, indicating that proper induction of IB kinase activity is contingent upon the formation of IKK⅐PP2A complexes. The precise mechanism of PP2A action on IKK remains unclear, especially with respect to the relevant phosphoacceptor sites under PP2A control. In this regard, Delhase et al. (7) have reported that autophosphorylation of multiple serines positioned in the C-terminal region of IKK␤ leads to down-regulation of the IKK complex. Further investigation will be required to determine whether PP2A positively regulates IKK by dephosphorylating these C-terminal residues in IKK␤. Notwithstanding, given the key role that IKK plays in inflammation, the kinase/phosphatase interaction described in this report represents an attractive therapeutic target for small molecule inhibitors.