Distinct Roles of the IκB Kinase α and β Subunits in Liberating Nuclear Factor κB (NF-κB) from IκB and in Phosphorylating the p65 Subunit of NF-κB

Phosphatidylinositol 3′-kinase (PI3K) and the serine/threonine kinase AKT have critical roles in phosphorylating and transactivating the p65 subunit of nuclear factor κB (NF-κB) in response to the pro-inflammatory cytokines interleukin-1 (IL-1) and tumor necrosis factor (TNF). Mouse embryo fibroblasts (MEFs) lacking either the α or β subunit of IκB kinase (IKK) were deficient in NF-κB-dependent transcription following treatment with IL-1 or TNF. However, in contrast to IKKβ-null MEFs, IKKα-null MEFs were not substantially defective in the cytokine-stimulated degradation of Iκβα or in the nuclear translocation of NF-κB. The IKK complexes from IKKα- or IKKβ-null MEFs were both deficient in PI3K-mediated phosphorylation of the transactivation domain of the p65 subunit of NF-κB in response to IL-1 and TNF, and constitutively activated forms of PI3K or AKT did not potentiate cytokine-stimulated activation of NF-κB in either IKKα- or IKKβ-null MEFs. Collectively, these data indicate that, in contrast to IKKβ, which is required for both NF-κB liberation and p65 phosphorylation, IKKα is required solely for the cytokine-induced phosphorylation and activation of the p65 subunit of NF-κB that are mediated by the PI3K/AKT pathway.

The NF-B 1 family of transcription factors consists of binary complexes of subunits with related promoter-binding and transactivation properties. The p65/RelA, RelB, and c-Rel subunits stimulate transcription, whereas the p50 and p52 subunits serve primarily to bind to DNA (1). The prototypical NF-B complex is the p65-p50 heterodimer (2). NF-B is sequestered in a latent form in the cytoplasm through its interaction with the inhibitory IB proteins. In response to signals, IB kinase is activated, and IB is phosphorylated and degraded, releasing NF-B, which enters the nucleus and binds to DNA (2)(3)(4)(5). However, the phosphorylation and degradation of IB and the consequent liberation of NF-B are not sufficient to activate NF-B-dependent transcription, which also relies on a second pathway, which leads to the stimulus-induced phosphorylation of the p65/RelA, RelB, and c-Rel subunits of NF-B (6 -15).
Our laboratory (13) and others (7,14) have shown that the pro-inflammatory cytokines IL-1 and TNF induce the phosphorylation and activation of the p65 subunit of NF-B, a pathway distinct from the one leading to IB degradation and NF-B nuclear translocation. Additionally, phosphatidylinositol 3Ј-kinase (PI3K) and the serine/threonine kinase AKT play critical roles in this pathway (13,16,17). Recently, an additional function for IL-1-stimulated PI3K/AKT activation has been reported: phosphorylation of the NF-B p50 subunit in response to these kinases increases the DNA-binding capacity of the NF-B complex (18).
Targeted gene disruptions have demonstrated that IKK␤ (but not IKK␣) is largely responsible for cytokine-induced IB degradation and NF-B nuclear translocation (19 -24). However, IKK␣-null mouse embryo fibroblasts (MEFs) are deficient in inducing several NF-B-dependent mRNAs in response to IL-1 and TNF (21). Activated AKT interacts with IKK␣ upon cytokine stimulation and induces the phosphorylation of threonine 23 (25). These findings raise the interesting possibility that, although IKK␣ is dispensable for IB␣ degradation and NF-B nuclear translocation, it may be required in the PI3K/ AKT pathway that leads to the phosphorylation and activation of NF-B. Therefore, we have investigated the roles of the IKK ␣ and ␤ subunits in the IL-1-and TNF-mediated phosphorylation and activation of the p65 subunit of NF-B.

EXPERIMENTAL PROCEDURES
Biological Reagents and Cell Culture-Recombinant human IL-1␤ was from NCI, National Institutes of Health. Recombinant human TNF was from Preprotech (Rocky Hill, NJ). LY 294002 was from Sigma. Polyclonal anti-IKK␣, anti-IKK␤, anti-IKK␥, anti-p65/RelA, and anti-IB␣ antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA). Polyclonal anti-phospho-p38 MAPK, anti-p38 MAPK, anti-phospho-AKT, anti-AKT, anti-phospho-ERK1/2, anti-ERK1/2, anti-phospho-p90 RSK , anti-p90 RSK , anti-phospho-JNK1, and anti-JNK antibodies were from Cell Signaling Technologies (Beverly, MA). Protein A-Sepharose and glutathione-agarose beads were from Amersham Biosciences, Inc. (Buckinghamshire, United Kingdom). Wild-type and IKK␣-and IKK␤-null MEFs, kindly provided by Dr. Inder Verma (21), were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 100 g/ml penicillin G, and 100 g/ml streptomycin. For all experiments, unless otherwise indicated, cells at 80% confluence on 100-mm dishes were preincubated with the PI3K inhibitor LY 294002 (20 M) for 30 min at 37°C prior to stimulation with IL-1 (2 ng/ml) or TNF (25 ng/ml) at 37°C for the indicated time periods. All results shown are typical of at least three independent experiments. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Northern Analyses-The cells were stimulated with either IL-1 or  TNF for 4 h or, where indicated, preincubated with LY 294002 for 30  min prior to stimulation with IL-1 or TNF for 4 h. Total RNA was  isolated using the TRIzol reagent (Invitrogen). RNA was fractionated by  electrophoresis on a formaldehyde gel and transferred to Hybond-N, a  positively charged nylon membrane, according to the procedures provided by Amersham Biosciences, Inc. cDNA probes for murine IL-6,  murine interferon-inducible protein 10 (IP10), and human glyceraldehyde-3-phosphate dehydrogenase mRNAs were made using the random priming kit from Amersham Biosciences, Inc. Probe hybridization and washing were performed according to procedures provided by Amersham Biosciences, Inc., and signals were visualized by autoradiography. The murine probes for IL-6 and IP10 were kindly provided by Dr. Thomas Hamilton (Cleveland Clinic Foundation).
Transfection and Reporter Assay-The NF-B-dependent reporter plasmid p5XIP10B, a kind gift of Dr. Bryan Williams (Cleveland Clinic Foundation), contains five tandem copies of the NF-B site from the IP10 gene. For reporter assays, wild-type and IKK␣-and IKK␤-null MEFs were stably transfected using Lipofectin (Invitrogen) with 10 g of p5XIP10B and 1 g of pBABEPuro. Pools of stably transfected cells were selected with and maintained in puromycin. In separate experiments, the NF-B reporter cells were transiently transfected using Lipofectin with 0.5 g of pSV2-␤gal and 5 g of either vector or construct expressing wild-type or lipid phosphatase-deficient mutants of PTEN, kindly provided by Dr. Kenneth Yamada (26); constitutively activated p110 PI3K , kindly provided by Drs. Doreen Cantrell and Karin Reif (27); constitutively activated AKT, kindly provided by Dr. Julian Downward (28); wild-type p65/RelA, kindly provided by Dr. Dean Ballard (29); wild-type IKK␣, kindly provided by Dr. David Donner (25); or wild-type IKK␤, kindly provided by Dr. Zhaodan Cao (30). Cells were divided into the appropriate number of plates for treatment 8 h following transfection. After 24 h, the cells were harvested. The cells were stimulated with either IL-1 or TNF for 4 h or, where indicated, preincubated with LY 294002 for 30 min prior to stimulation with IL-1 or TNF for 4 h. Luciferase or galactosidase activity was determined with the luciferase assay system or chemiluminescent reagents (both from Promega, Madison, WI). Luciferase activity was normalized to ␤-galactosidase activity to control for transfection efficiency. The viability of each transfected cell population was measured at the time of harvesting by trypan blue exclusion.
Gel Electrophoretic Mobility Shift Assays-For electrophoretic mobility shift assays (EMSAs), where indicated, the cells were preincubated with LY 294002 prior to stimulation with IL-1 or TNF for 20 min. The NF-B-binding site (5Ј-GAGCAGAGGGAAATTCCGTAACTT-3Ј) from the IP10 gene was used as a probe. Briefly, complementary oligonucleotides, end-labeled with polynucleotide kinase and [␥-32 P]ATP, were annealed by slow cooling. Approximately 20,000 cpm of probe were used per reaction mixture. Nuclear and cytoplasmic extracts were prepared in binding reaction buffer as described previously (31). The binding reaction was carried out with nuclear extracts containing equal amounts of protein at room temperature for 30 min in a total volume of 20 l. The DNA⅐NF-B complexes were separated on 5% polyacrylamide gels by electrophoresis in low ionic strength Tris borate/EDTA buffer. The gels were dried, and the labeled complexes were visualized by autoradiography.
Immunoblotting and Immunoprecipitation-Cells were washed once with phosphate-buffered saline and lysed for 30 min at 4°C in 1 ml of 0.5% Nonidet P-40 lysis buffer as described previously (32). Cellular debris was removed by centrifugation at 16,000 ϫ g for 15 min. For immunoblotting, cell extracts were fractionated directly by SDS-PAGE and transferred to nitrocellulose membranes. Immunoblot analysis was performed with the indicated primary antibodies, which were visualized with horseradish peroxidase-coupled goat anti-rabbit or antimouse immunoglobulins using the ECL Western blotting detection system (PerkinElmer Life Sciences). For immunoprecipitations, cell extracts were incubated with 3 g of primary antibody for 4 h, followed by incubation for 1 h with 50 l of protein A-Sepharose beads (50% suspension). The beads were washed three times with lysis buffer, and samples were analyzed by SDS-PAGE and autoradiography.
Analysis of the Phosphorylation of the p65 Transactivation Domain and IB␣-(1-54)-A fragment of p65 (residues 354 -551) representing the transactivation domain (TAD) was used (12). The IB␣ fragment (residues 1-54) was kindly provided by Dr. Joseph DiDonato (Cleveland Clinic Foundation). These proteins were expressed as glutathione Stransferase (GST) fusions in bacteria and purified using glutathioneagarose beads after sonication at 4°C in 0.5% Nonidet P-40 lysis buffer (33). The GST fusion proteins were eluted from the beads in kinase buffer (20 mM HEPES (pH 7.6), 20 mM MgCl 2 , 2 mM dithiothreitol, 20 M ATP, 20 mM ␤-glycerophosphate, 20 mM disodium p-nitrophenyl phosphate, 0.1 mM sodium orthovanadate, 3 Ci of [␥-32 P]ATP, and 10 mM reduced glutathione). The cells were stimulated with either IL-1 or TNF for 20 min or, where indicated, preincubated with LY 294002 for 30 min prior to stimulation with IL-1 or TNF for 20 min. Nuclear and cytoplasmic extracts were prepared as described previously (31); or the cells were lysed, and anti-IKK␥ antibody was used to immunoprecipitate the IKK complex from each sample. In vitro phosphorylation was performed using ϳ1 g of the p65 TAD/GST or IB␣-(1-54)/GST fusion protein as a substrate with either cytoplasmic or nuclear extracts (3 g of protein) or the immunoprecipitated IKK complex as the kinase in kinase buffer at 30°C for 30 min (12). Following the kinase reaction, phosphorylation of either substrate was analyzed by SDS-PAGE, followed by autoradiography. In separate experiments, wild-type MEFs were transiently transfected using Lipofectin with 5 g of either vector or construct expressing constitutively activated AKT (28). The transfected cells were divided into three plates for treatment 8 h following transfection. After 24 h, the cells were stimulated with either IL-1 or TNF for 20 min. The cells were lysed, and anti-IKK␥ antibody was used to immunoprecipitate the IKK complex from each sample. In vitro phosphorylation of the p65 TAD/GST fusion protein was performed as described above with the immunoprecipitated IKK complex.

IL-1-and TNF-induced NF-B-dependent Transcription Is Deficient in Both IKK␣-and IKK␤-null MEFs, but Only IKK␤null MEFs Are Deficient in NF-B Liberation and Nuclear
Translocation-The complete activation of NF-B requires two pathways leading to the liberation of NF-B and to the activation of the transcription function of NF-B. Studies from our laboratory (13) and by Madrid et al. (16,34) have demonstrated a critical role of the PI3K/AKT pathway in activating NF-B in response to the pro-inflammatory cytokines IL-1 and TNF by phosphorylation of the carboxyl-terminal transactivation do-

FIG. 1. IKK␣-and IKK␤-null MEFs are both deficient in IL-1and TNF-stimulated induction of NF-B-dependent transcription and promoter activation.
A, transcription. Cells were stimulated with IL-1 (IL) or TNF (T) for 4 h, and total RNA was isolated. Equal amounts of RNA were electrophoresed and subjected to Northern analysis with probes for murine IL-6 and IP10 and human glyceraldehyde-3-phosphate dehydrogenase (GAPDH). B, reporter assay. Wildtype (WT) and IKK␣and IKK␤-null MEFs, stably transfected with an NF-B-dependent luciferase reporter construct containing five copies of the NF-B consensus site from the IP10 gene, were either unstimulated or stimulated with IL-1 or TNF for 4 h and lysed. Equal amounts of protein were assayed for luciferase activity. C, control. main of p65. IKK␣-null MEFs were reported to be deficient in NF-B-dependent transcription following stimulation with IL-1 and TNF (21), even though IKK␣ seemed to be dispensable for cytokine-induced IB degradation and NF-B nuclear translocation (19 -24). We have now investigated the roles of both IKK␣ and IKK␤ in the activation of NF-B through the PI3K/AKT pathway. Both IKK␣-and IKK␤-null MEFs were deficient in IL-1-and TNF-stimulated induction of the NF-Bdependent endogenous genes IL-6 and IP10 (Fig. 1A). Both were also deficient in activating an NF-B-dependent reporter (Fig. 1B), confirming that the defects are at the level of NF-B function. To explore why IL-1 and TNF fail to activate NF-B in IKK␣-and IKK␤-null MEFs, we investigated known pathways leading to NF-B activation and the signaling deficiencies responsible for causing the null MEFs to be unresponsive to cytokine stimulation. Wild-type and IKK␣-and IKK␤-null MEFs were exposed to either IL-1 or TNF, and total cell extracts were analyzed for phosphorylated AKT, ERK, JNK, p90 RSK , and p38. No defects in the induction of these kinase pathways by IL-1 and TNF were detected (Fig. 2). Interestingly, the down-regulation of these kinase activities appeared to be more sustained in IKK␣-and IKK␤-null MEFs compared with wild-type MEFs (Fig. 2). The Western blots were stripped and reprobed to determine total levels of each of the above kinases, which were the same in the different cell lines (data not shown). The data of Fig. 2 demonstrate that the induction of these kinase pathways was completely intact in response to IL-1 and TNF in both IKK␣-and IKK␤-null MEFs. However, IKK␤-null MEFs were deficient in cytokine-stimulated IB␣ degradation. In contrast, IKK␣-null MEFs were as efficient as wild-type MEFs in degrading IB␣ in response to IL-1 and TNF (Fig. 3A). There were no substantial defects in the ability of NF-B to translocate to the nucleus and to bind to DNA following IL-1 or TNF stimulation in IKK␣-null MEFs, but IKK␤-null MEFs were deficient in cytokine-stimulated NF-B nuclear translocation and DNA binding (Fig. 3B).

IKK␣-and IKK␤-null MEFs Are Both Deficient in IL-1-and TNF-stimulated Phosphorylation of the Transactivation Domain of the p65 Subunit of NF-B, but Only IKK␤-null MEFs
Are Deficient in IL-1-and TNF-stimulated Phosphorylation of IB␣-Although there was no substantial defect in the degradation of IB␣ or in the nuclear translocation and DNA binding of NF-B in IKK␣-null MEFs (Fig. 3), these cells were deficient in IL-1-and TNF-stimulated activation of NF-B-dependent endogenous genes and an NF-B-dependent reporter construct (Fig. 1). We analyzed of the ability of cytoplasmic and nuclear extracts from wild-type and IKK␣-and IKK␤-null MEFs exposed to either IL-1 or TNF to phosphorylate a GST fusion protein containing the transactivation domain of p65 (p65 TAD/GST). The results of this experiment indicated that the majority of the IL-1-and TNF-induced kinase activity for this substrate resided in the cytoplasmic fraction of wild-type MEFs and was absent in both IKK␣-and IKK␤-null MEFs (data not shown). In view of these results, we assayed the ability of immunoprecipitated IKK complexes from wild-type and IKK␣and IKK␤-null MEFs to phosphorylate p65 TAD/GST. The cells were treated with IL-1 or TNF and, where indicated, incubated with LY 294002 before treatment to inhibit the cytokine-stimulated PI3K/AKT pathway. The composition of the immunoprecipitated IKK complex from each cell type is shown in Fig. 4A. The IKK complex from IKK␣-null MEFs was unable to phosphorylate p65 TAD/GST efficiently (Fig. 4B, upper panel). Phosphorylation of p65 TAD/GST by the wild-type IKK complex depends on activation of the PI3K/AKT pathway by IL-1 and TNF, as preincubation with LY 294002 almost completely blocked phosphorylation in wild-type MEFs (Fig. 4B, upper  panel). Interestingly, IKK␤-null MEFs were also deficient in phosphorylating p65 TAD/GST (Fig. 4B, upper panel). However, the IKK complex from IKK␣-null MEFs was not defective FIG. 2. IL-1-and TNF-stimulated phosphorylation and activation of AKT, ERK, JNK, p90 RSK , and p38 are intact in both IKK␣-and IKK␤-null MEFs. Cells were exposed to either IL-1 or TNF for the times indicated. Total cell extracts were prepared, and equal amounts of protein were analyzed by SDS-PAGE and Western blotting for activated, phosphorylated AKT (pAKT), ERK (pERK), JNK (pJNK), p90 RSK (pRSK), and p38 (pp38). No defects were detected in any of these pathways in IKK␣-or IKK␤-null MEFs. WT, wild-type MEFs.
FIG. 3. IKK␤-null MEFs (but not IKK␣-null MEFs) are substantially deficient in IL-1-and TNF-stimulated IB␣ degradation, p65 nuclear translocation, and NF-B DNA binding. A, IB␣ degradation. Cells were exposed to either IL-1 or TNF for the times indicated, and total cell extracts were prepared. Equal amounts of protein were analyzed by SDS-PAGE and Western blotting. B, EMSAs. Cells were exposed to either IL-1 (IL) or TNF (T) for 20 min, and nuclear extracts were prepared. Equal amounts of each extract were analyzed by EMSA for the ability of NF-B to bind to a labeled NF-B consensus site from the IP10 gene. WT, wild-type MEFs; C, control.
in the IL-1-and TNF-stimulated phosphorylation of IB␣-(1-54)/GST, as was the IKK complex from IKK␤-null MEFs (Fig.  4B, lower panel) (21). Phosphorylation of IB␣-(1-54)/GST does not depend on activation of the PI3K/AKT pathway by IL-1 and TNF, as preincubation with LY 294002 had no effect on the phosphorylation of this substrate by the IKK complex (Fig. 4B,  lower panel). To test whether IKK is capable of phosphorylating p65 in AKT-activated cells, the effect of overexpressing constitutively activated AKT on IKK phosphorylation of p65 TAD/GST was investigated. Wild-type MEFs were transiently transfected with either vector alone or activated AKT. 8 h after transfection, the cells were divided into three plates each. After 24 h, the cells were left unstimulated or were stimulated with IL-1 or TNF for 20 min, and the immunoprecipitated IKK complex from each sample was assayed for p65 TAD kinase activity by the in vitro kinase assay. Phosphorylation of p65 TAD/GST by the wild-type IKK complex was highly induced by activated AKT compared with the vector-transfected control and could not be significantly further induced by either IL-1 or TNF (Fig. 4C). Therefore, one reason that IKK␣-null MEFs are unable to activate NF-B in response to IL-1 and TNF is because they fail to phosphorylate and activate the p65 subunit of NF-B in response to cytokine-stimulated PI3K and AKT. Our data also indicate a role for IKK␤ in this pathway in addition to its role in phosphorylating IB. Only reconstitution with IKK␣ (and not IKK␤) can restore cytokine-stimulated NF-B-dependent promoter and endogenous gene activation in IKK␣-null MEFs (data not shown). All together, these data indicate a separate, essential function of IKK␣, distinct from that of IKK␤, in the activation of NF-B.

IL-1-and TNF-dependent Activation of NF-B through the PI3K/AKT Pathway Requires IKK␣ Independently of IB␣
Degradation-We tested the ability of constitutively activated PI3K and AKT to enhance IL-1-and TNF-stimulated NF-Bdependent promoter activation in wild-type and IKK␣-and IKK␤-null MEFs. These cells, stably transfected with the NF-B-dependent luciferase reporter construct p5XIP10B, were transiently transfected with vector alone, activated p110, activated AKT, or wild-type p65. 8 h after transfection, the cells were divided into three plates each. After 24 h, the cells were left unstimulated or were stimulated with IL-1 (Fig. 5A) or TNF (Fig. 5B) for 4 h and assayed for luciferase activity. Transfection of wild-type p65 weakly increased both IL-1-induced (IKK␣ ϭ 5.4-fold and IKK␤ ϭ 3.8-fold) and TNF-induced (IKK␣ ϭ 3.9-fold and IKK␤ ϭ 3.0-fold) NF-B-dependent promoter activation in IKK␣-and IKK␤-null MEFs, but not nearly as well as in wild-type MEFs (IL-1 ϭ 23.9-fold and TNF ϭ 17.9-fold), over their respective untreated p65-transfected controls (Fig. 5, A and B). Transfection of p65 alone with no cytokine treatment resulted in an 8-fold increase in wild-type MEFs, a 3-fold increase in IKK␣-null MEFs, and a 2-fold increase in IKK␤-null MEFs over their respective vector-transfected controls (data not shown), indicating that both basal and cytokine-induced p65-dependent transactivation is diminished in both IKK␣-and IKK␤-null MEFs compared with wild-type MEFs. However, constitutively activated PI3K and AKT both

FIG. 4. The IKK complex from IKK␣-and IKK␤-null MEFs is deficient in both the IL-1-and TNF-stimulated phosphorylation of the transactivation domain of the p65 subunit of NF-B, but only the IKK complex from IKK␤-null MEFs is defective in phosphorylating IB␣.
Cells were stimulated with IL-1 (IL) or TNF (T) for 20 min or, where indicated, preincubated with LY 294002 (LY) for 30 min prior to stimulation with IL-1 or TNF for 20 min. Cells were lysed, and anti-IKK␥ antibody was used to immunoprecipitate the IKK complex from each sample. A, Western analysis of the composition of the IKK complex from each cell type. B, in vitro phosphorylation. Approximately 1 g of p65 TAD/GST (upper panel) or IB␣-(1-54)-GST (lower panel) was used as substrate, and the immunoprecipitated IKK complex was the kinase. Following the kinase reaction, phosphorylation of p65 TAD/GST was analyzed by SDS-PAGE, followed by autoradiography. C, in vitro phosphorylation of p65 TAD/GST. Wild-type MEFs (WT) were transiently transfected with 5 g of either vector or activated AKT construct. After 8 h, transfected cells were divided into three plates each. After 24 h, each transfected sample was left unstimulated or was stimulated with either IL-1 or TNF for 20 min. Cells were lysed; anti-IKK␥ antibody was used to immunoprecipitate the IKK complex from each sample; and the p65 TAD in vitro kinase assay was performed with the IKK complexes as described for B. The phosphorylation of p65 TAD/GST was analyzed by SDS-PAGE, followed by autoradiography. C, control. failed to enhance IL-1-or TNF-stimulated NF-B-dependent promoter activation in IKK␣-and IKK␤-null MEFs in contrast to wild-type MEFs (Fig. 5, A and B). As we reported previously (13), neither activated PI3K nor AKT can induce NF-B activation alone, presumably because a second signal is needed for IB␣ degradation. As there was no defect in the cytokinestimulated phosphorylation and degradation of IB␣ in IKK␣null MEFs, a major role of IKK␣ in cytokine-dependent signaling must be in the phosphorylation and activation of the p65 subunit of NF-B. Blockade of IL-1-and TNF-induced PI3K activity with LY 294002 inhibited IL-1-and TNF-stimulated induction of IL-6 and IP10 mRNAs (Fig. 6A) as well as NF-Bdependent promoter activation (Fig. 6D). Despite dramatically inhibiting NF-B-dependent endogenous gene transcription and promoter activation (Fig. 6, A and D), neither IL-1-or TNF-induced IB␣ degradation (Fig. 6B, upper panel) nor NF-B DNA binding (Fig. 6C) was affected by inhibiting PI3K. However, the phosphorylation and activation of AKT in response to IL-1 and TNF were dramatically inhibited by PI3K blockade (Fig. 6B, lower panel). The IL-1-and TNF-stimulated phosphorylation of ERK, JNK, p90 RSK , and p38 was unaffected by pretreatment with LY 294002 (data not shown). The production of the phospholipid second messenger phosphatidylinositol 3,4,5-trisphosphate by PI3K could be inhibited directly by its endogenous antagonist, the tumor suppressor PTEN, a dual-specificity phosphatase that dephosphorylates phosphatidylinositol 3,4,5-trisphosphate. We explored the effects of inhibiting IL-1-and TNF-stimulated PI3K activity by LY 294002 as well as by expressing PTEN on NF-B-dependent promoter activation in wild-type MEFs stably transfected with the NF-B-dependent luciferase reporter p5XIP10B. The cells were transiently transfected with either vector or wild-type PTEN. After 48 h, cells were preincubated with either vehicle or LY 294002 for 30 min prior to stimulation with IL-1 or TNF, and luciferase activity was measured (Fig. 6D). Blockade of IL-1and TNF-induced PI3K activity with either wild-type PTEN or LY 294002 inhibited both IL-1 and TNF stimulation of the NF-B-dependent promoter. We also compared the effects of wild-type PTEN with those of three different mutants that lack lipid phosphatase activity. Only wild-type PTEN suppressed NF-B activation (data not shown). Therefore, the ability of PTEN to suppress NF-B-dependent transcription driven by IL-1 and TNF depends on its lipid phosphatase activity. These data confirm our previous results showing that the PI3K/AKT pathway functions to stimulate p65 activation independently of the degradation of IB␣ and activation of the ability of NF-B to bind to DNA (13). DISCUSSION

IKK␣ and IKK␤ Are Both Required for PI3K-and AKTmediated Activation of NF-B in Response to IL-1 and TNF-
We have investigated the roles of IKK␣ and IKK␤ in the activation of NF-B through the PI3K/AKT pathway in response to IL-1 and TNF. IKK␣ is required for the PI3K/AKT pathway to phosphorylate and transactivate the p65 subunit of NF-B. In addition to the role of IKK␤ as the IB␣ kinase, it also plays a role in the PI3K/AKT-mediated phosphorylation and activation of p65. Our previous findings indicated that activation of the PI3K/AKT pathway is not sufficient to activate NF-B-dependent transcription, but is necessary for IL-1 and TNF to activate NF-B together with the pathway that leads to IB␣ degradation and nuclear translocation of NF-B (13). Our present results (Figs. 4B, lower panel; and 6) agree with these findings and with the studies of others, indicating that the IL-1-and TNF-stimulated PI3K/AKT pathway functions to stimulate the transactivation potential of NF-B and does not participate in the pathway leading to IB␣ phosphorylation, IB␣ degradation, and NF-B liberation (13,16,34). We also demonstrated that the endogenous antagonist of the PI3K/AKT pathway, the tumor suppressor PTEN, inhibits the IL-1-and TNF-stimulated activation of NF-B in a manner that depends on its lipid phosphatase function ( Fig. 6 and data not shown). Our finding that wild-type PTEN is able to inhibit NF-B-dependent transcriptional activation agrees with two previous reports. In one, PTEN suppressed both the activation of IKK and the ability of NF-B to bind to DNA (35), whereas in the second, PTEN affected the ability of NF-B to bind to DNA by inhibiting the phosphorylation of the p50 subunit of NF-B independently of IB␣ degradation (18). We found no substantial difference in IL-1-and TNF-stimulated NF-B DNA binding between wild-type and IKK␣-null MEFs (Fig. 3B). However, the total levels of induced NF-B DNA binding in IKK␣null MEFs are ϳ2-fold lower than those in wild-type MEFs (Fig. 3B). This may be due to reduced phosphorylation of p50 in these cells and remains to be investigated. Neither constitutively activated PI3K nor AKT is able to enhance the activation of NF-B by IL-1 or TNF in either IKK␣-or IKK␤-null MEFs (Fig. 5). These data indicate that both IKK␣ and IKK␤ are required for the PI3K/AKT pathway to activate NF-B.
IKK␣, Unlike IKK␤, Is Dispensable for NF-B Liberation, but Both Are Required for the PI3K/AKT-mediated Phosphorylation of p65-Targeted gene disruption studies have demonstrated that IKK␤ (but not IKK␣) is largely responsible for cytokine-induced IB degradation and translocation of NF-B to the nucleus in response to IL-1 and TNF. In this respect, the failure of PI3K and AKT to activate NF-B in IKK␤-null MEFs may be due in part to the defective liberation of NF-B in these FIG. 5. Constitutively activated PI3K and AKT are unable to enhance IL-1-or TNF-stimulated activation of a reporter gene in the absence of either IKK␣ or IKK␤. Wild-type (WT) and IKK␣and IKK␤-null MEFs, stably transfected with an NF-B-dependent luciferase reporter construct controlled by five copies of the NF-B consensus site from the IP10 gene, were transiently transfected with 5 g of either vector (Vec) or construct encoding activated p110, activated AKT, or the p65 subunit of NF-B plus 0.5 g of pSV2-␤gal. After 8 h, transfected cells were divided into three plates each. After 24 h, each transfected sample was left unstimulated or was stimulated with either IL-1 (A) or TNF (B) for 4 h and lysed. Equal amounts of protein were assayed for luciferase and ␤-galactosidase activities. Luciferase activity was normalized to ␤-galactosidase activity to control for transfection efficiency. Data are expressed as the fold induction, the ratio of the luciferase activity of each IL-1-or TNF-treated transfected sample to the same activity of the untreated transfected control. The data shown are representative of at least three separate independent experiments. cells (Fig. 3). However, the diminished capacity of overexpressed p65 to drive both basal and cytokine-induced NF-B promoter activation in IKK␤-null MEFs also indicates a role of IKK␤ in the phosphorylation and activation of NF-B p65 (Fig.  5). Both IKK␣ and IKK␤ are required for IL-1 and TNF to activate NF-B-dependent transcription (Fig. 1). However, the IKK complex from IKK␣-null MEFs is not deficient in IB␣ phosphorylation (Fig. 4B, lower panel), IB␣ degradation, or NF-B liberation, but is deficient in activating NF-B-dependent transcription in response to IL-1 and TNF ( Figs. 1 and 3, A  and B). There are no defects in the IL-1-or TNF-dependent phosphorylation and activation of AKT, ERK, JNK, p90 RSK , or p38 in either IKK␣-or IKK␤-null MEFs (Fig. 2). Our demonstration that the phosphorylation and activation of p38 in response to IL-1 are intact in IKK␣-null MEFs differs from a recent report indicating that IL-1 activates NF-B by inducing p38 activity in a manner dependent on AKT and IKK␣ (34). We have no explanation for this discrepancy other than that the IKK␣-null MEFs utilized in the study of Madrid et al. (34) were obtained from a different source (19). However, our demonstration that the absence of IKK␣ (Fig. 2) or inhibition of PI3K (data not shown) does not affect p38 phosphorylation and activation in response to IL-1 and TNF, but does inhibit AKT activation and NF-B-dependent transcription (Fig. 6), indicates that cytokine-mediated p38 activation is independent of the IKK␣ and PI3K/AKT pathways. However, we cannot exclude a separate role of p38 in the activation of NF-B as suggested by Madrid et al. (34) and corroborated by previous reports of NF-B modulation by p38 (36,37). The sustained activation of p38 and the other kinases observed in IKK␣-and IKK␤-null MEFs compared with wild-type MEFs is an interesting observation (Fig. 2). A recent report suggests that both IKK␤-and p65-null MEFs demonstrate a sustained JNK activation in response to TNF compared with wild-type MEFs, contributing to TNF-induced apoptosis, which can be reversed by overexpression of NF-B-induced XIAP (X-chromosomelinked inhibitor of apoptosis) (38). A second report demonstrated that inhibition of NF-B activation by the super-repressor IB␣(A32/36) mutant elicits sustained JNK activation without inhibiting MAPK phosphatase-1, a JNK phosphatase (39). These reports suggest that NF-B target genes may downregulate JNK activation. Our results indicate that inhibition of NF-B and NF-B target genes may have a broader impact on the down-regulation of the IL-1-and TNF-stimulated MAPKs and remain to be further investigated. Our studies demonstrate a novel role of IKK␣ in the phosphorylation and activation of p65 by the PI3K/AKT pathway. We have demonstrated that the IKK complex from IKK␣-null MEFs is unable to phosphorylate the transactivation domain of p65 efficiently (Fig.  4B, upper panel). The phosphorylation of the p65 transactivation domain by the IKK complex depends on the activation of the PI3K/AKT pathway by IL-1 and TNF, as preincubation with LY 294002 almost completely blocked this phosphorylation in wild-type MEFs (Fig. 4B, upper panel). Our study demonstrates that the IKK complex from IKK␤-null MEFs is also deficient in phosphorylating the transactivation domain of p65 (Fig. 4B, upper panel). The deficiency of IKK␤-null MEFs in phosphorylating p65 is in agreement with the study of Madrid (IL) or TNF (T) for 4 h, and total RNA was isolated. Equal amounts of RNA were electrophoresed and subjected to Northern analysis with probes for murine IL-6 and IP10 and human glyceraldehyde-3-phosphate dehydrogenase (GAPDH). B, analysis of IB␣ and phosphorylated AKT (pAkt). Cells were preincubated with either vehicle or 20 M LY 294002 for 30 min before stimulation with either IL-1 or TNF for 20 min. Total cell extracts were prepared, and equal amounts of protein were analyzed by SDS-PAGE and Western blotting. C, EMSAs. Cells were preincubated with either vehicle or 20 M LY 294002 for 30 min prior to stimulation with either IL-1 or TNF for 20 min, and nuclear extracts were prepared. Equal amounts of each extract were analyzed by EMSA for the binding of NF-B to an NF-B consensus probe. D, analysis of the activation of an NF-B-dependent reporter. Wild-type MEFs stably transfected with an NF-B-dependent luciferase reporter construct were transiently transfected with 5 g of either vector (VEC) or construct encoding wild-type PTEN plus 0.5 g of pSV2-␤gal. After 24 h, each cell population was split into three plates. After an additional 24 h, cells were preincubated with either vehicle or 20 M LY 294002 for 30 min before being stimulated with IL-1 or TNF for 4 h. After lysis, equal amounts of protein were assayed for luciferase activity, which was normalized to ␤-galactosidase activity to control for transfection efficiency. C, control. et al. (34), demonstrating that AKT targets the transactivation function of p65 in a manner that is dependent on IKK␤, and studies demonstrating the direct phosphorylation of p65 by IKK␤ (12). Therefore, both IKK␣ and IKK␤ are required for the IL-1-and TNF-stimulated PI3K/AKT pathway to induce the phosphorylation and transactivation potential of p65. However, the inability of IKK␤ to substitute for IKK␣ in restoring NF-B activation to IKK␣-null MEFs (data not shown) indicates an essential function of IKK␣ separate from that of IKK␤ in mediating the cytokine-stimulated activation of p65.
Is the IB Kinase the PI3K-and AKT-activated p65 Kinase?-Our data that the IKK complex from wild-type MEFs expressing activated AKT displays significantly increased kinase activity for the p65 TAD (Fig. 4C) indicate that the IKK complex phosphorylates the p65 TAD in response to activated AKT. We plan to investigate whether IKK␣ and IKK␤ require their kinase activities to stimulate the pathway leading to p65 activation. Reconstitution of IKK␣/IKK␤ double knockout MEFs with combinations of wild-type and kinase-dead mutants of IKK␣ and IKK␤ will help to answer this interesting question. This line of experimentation may also lead to resolution of the controversy over the order of IKK activation and the possible interdependence of these two highly homologous kinases (40 -42). In conclusion, IL-1 and TNF activate two separate but interrelated signal transduction pathways, and both are necessary for the full activation of NF-B. One set of signals activates the IKK complex, primarily through IKK␤, to phosphorylate IB␣, inducing IB␣ degradation and liberating NF-B, which can then enter the nucleus and bind to DNA. The second set of signals utilizes PI3K and AKT to activate the IKK complex, requiring both IKK␣ and IKK␤ to phosphorylate and activate the transactivation domain of the p65 subunit of NF-B.