The NF- (cid:1) B Activation in Lymphotoxin (cid:2) Receptor Signaling Depends on the Phosphorylation of p65 at Serine 536*

NF- (cid:1) B-inducing kinase (NIK) has been shown to play an essential role in the NF- (cid:1) B activation cascade elicited by lymphotoxin (cid:2) receptor (LT (cid:2) R) signaling. How-ever, the molecular mechanism of this pathway remains unclear. In this report we demonstrate that both NIK and I (cid:1) B kinase (cid:3) (IKK (cid:3) ) are involved in LT (cid:2) R signaling and that the phosphorylation of the p65 subunit at serine 536 in its transactivation domain 1 (TA1) plays an essential role. We also found that NF- (cid:1) B could be activated in the LT (cid:2) R pathway without altering the level of the phosphorylation of I (cid:1) B and nuclear localization of p65. By using a heterologous transactivation system in which Gal4-dependent reporter gene is activated by the Gal4 DNA-binding domain in fusion with various portions of p65, we found that TA1 serves as a direct target in the NIK-IKK (cid:3) pathway. In addition, mutation studies have revealed the essential role of Ser-536 within TA1 of p65 in transcriptional control mediated by NIK-IKK (cid:3) . Furthermore, we found that Ser-536

NF-B-inducing kinase (NIK) has been shown to play an essential role in the NF-B activation cascade elicited by lymphotoxin ␤ receptor (LT␤R) signaling. However, the molecular mechanism of this pathway remains unclear. In this report we demonstrate that both NIK and IB kinase ␣ (IKK␣) are involved in LT␤R signaling and that the phosphorylation of the p65 subunit at serine 536 in its transactivation domain 1 (TA1) plays an essential role. We also found that NF-B could be activated in the LT␤R pathway without altering the level of the phosphorylation of IB and nuclear localization of p65. By using a heterologous transactivation system in which Gal4-dependent reporter gene is activated by the Gal4 DNA-binding domain in fusion with various portions of p65, we found that TA1 serves as a direct target in the NIK-IKK␣ pathway. In addition, mutation studies have revealed the essential role of Ser-536 within TA1 of p65 in transcriptional control mediated by NIK-IKK␣. Furthermore, we found that Ser-536 was phosphorylated following the stimulation of LT␤R, and this phosphorylation was inhibited by the kinase-dead dominant-negative mutant of either NIK or IKK␣. These observations provide evidence for a crucial role of the NIK-IKK␣ cascade for NF-B activation in LT␤R signaling.
The lymphotoxin (LT) 1 system plays crucial roles in the embryonic development of lymphoid organ, the maintenance of lymphoid architecture, and the formation of ectopic lymphoid tissue adjacent to chronic inflammatory sites (1)(2)(3). LT is a heterotrimer complex consisting of ␣ (LT␣) and ␤ (LT␤) subunits (as LT␣1␤2), which bind to its specific receptor (LT␤R) (1). LT␤R signaling involves NF-B-inducing kinase (NIK), which eventually activates nuclear factor B (NF-B) (4,5). The involvement of NIK in the LT␤R signaling has been suggested by the shared phenotypes of gene knock-out mice of LT␣, LT␤, and LT␤R (1,6,7) and alymphoplasia mice (aly/aly) (8) in which spontaneous mutations of NIK were found responsible (9 -11). Moreover, LT␤R signaling was shown to involve both NIK and IKK␣ for NF-B activation (4). It has also been shown that NIK is indispensable for LT␤R signaling but not for tumor necrosis factor (TNF) signaling (5). However, the molecular mechanism by which NF-B is activated by LT␤R signaling has not been clarified.
In cells, NF-B is largely cytoplasmic and therefore remains transcriptionally inactive until a cell receives an appropriate stimulus. In response to proinflammatory cytokines such as TNF and interleukin-1␤ (IL-1␤), the IB proteins become phosphorylated on two serine residues located in the N-terminal region (21). Phosphorylation of IB proteins results in rapid ubiquitination and subsequent proteolysis by the 26 S proteasome (15,22,23), which allows the liberated NF-B to translocate to the nucleus and participate in target gene transactivation (12)(13)(14)(15). The large molecular weight complex consisting of two catalytic subunits, IB kinases ␣ and ␤ (IKK␣ and IKK␤), and a regulatory subunit IKK␥ was identified and shown to be responsible for phosphorylating IB proteins (24 -29). It has recently been shown that IKK␣ is not required for IB degradation or induction of NF-B DNA binding but essential for the generation of transcriptionally competent NF-B (30). The kinase activity of IKKs is induced by a wide variety of NF-B inducers such as TNF or IL-1␤, and mediated by the upstream kinases including NIK and the extracellular signalregulated kinase kinase kinase 1/3 (31)(32)(33)(34). NIK was originally identified as a protein interacting with the TNF receptor-associated factor 2 component of the TNF receptor complex (35). NIK physically interacts via its C-terminal region with IKK␣ and IKK␤ and stimulates their catalytic activity as an upstream effector kinase (32, 36 -39).
The NF-B p65 subunit contains at least two independent transactivation (TA) domains (TA1 and TA2) within its C-terminal 120 amino acids and is responsible for binding to the basal transcription factor TFIIB and CBP/p300 coactivators (19,20). The TNF-mediated signaling was shown to involve phosphorylation of Ser-529 within TA1 by casein kinase II (CKII) (40,41). Similarly, overexpression of IKK␤ induced phosphorylation of p65 at Ser-536 (42). These two serine residues within p65 TA1 were also shown to be essential for Rasmediated NF-B activation involving phosphatidylinositol 3-ki-nase and Akt serine/threonine kinase (43). These signalinduced p65 phosphorylation events appear to induce NF-Bdependent gene expression by augmenting the transcriptional activity of NF-B (p65) rather than by inducing IB phosphorylation and promoting its nuclear translocation. Interestingly, recent studies revealed the presence of NF-B and IB in the nucleus even in the resting unstimulated cells (44 -47), thus making NF-B susceptible for regulatory phosphorylation in the nucleus.
In this study, we have attempted to clarify the molecular mechanism by which NIK activates NF-B and found that the TA1 domain of p65 subunit is indispensable for NF-B transcriptional activity. We demonstrate that the phosphorylation of p65 at Ser-536, mediated by NIK-IKK␣, is crucial for LT␤R signaling.
Cell Culture and Transfection-293 cells were grown at 37°C in Dulbecco's modified Eagle's medium (Sigma) with 10% heat-inactivated fetal bovine serum (IBL, Maebashi, Japan). Cells were transfected using FuGENE TM 6 transfection reagent (Roche Molecular Biochemicals) according to the manufacturer's recommendations. HT29 cells were grown at 37°C in McCoy's 5A Medium Modified (Sigma) with 10% heat-inactivated fetal bovine serum (IBL), and cells were transfected using LipofectAMINE TM reagent (Invitrogen) according to the manufacturer's recommendations. At 48-h post-transfection, the cells were harvested, and the cell extracts were prepared for the luciferase assay. Luciferase activity was measured using the luciferase assay system (Promega) as described previously (52). Transfection efficiency was monitored by Renilla luciferase activity using the pRL-TK plasmid (Promega) as an internal control, and the luciferase activity was normalized by the Renilla luciferase activity. For each transfection, 50 ng of the luc reporter plasmid and 25 ng of internal control plasmid pRL-TK were used. pUC19 was used to adjust the total amount of DNA (500 ng) transfected. Cells without the stimulation of TNF were lysed 48 h after transfection, and the luciferase activity was measured. Other cells, as indicated, were stimulated with 10 ng/ml of TNF after 24 h of transfection and lysed after an additional incubation for 24 h or stimulated with 2 g/ml of agonistic anti-LT␤R monoclonal antibody (mAb) (AC.H6) (53) 10 h before cells were harvested. The data are presented as the fold increase in luciferase activity (mean Ϯ S.D.) relative to the control of three independent transfections.
Immunostaining-The intracellular localization of p65 in HT29 cells was examined by immunostaining as described previously (50). Briefly, HT29 cells were cultured in 2-well chamber slides and after stimulating with 10 ng/ml of TNF for 15 min or 2 g/ml of agonistic anti-LT␤R mAb for 40 min, cells were fixed in 4% (w/v) paraformaldehyde/PBS at room temperature for 20 min and then permeabilized by 0.5% Triton X-100/ PBS for 20 min at room temperature. They were then incubated with rabbit polyclonal antibody against p65 (Santa Cruz Biotechnology) for 1 h at 37°C, rinsed three timed with 0.05% Triton X-100/PBS, and incubated with secondary antibody, fluorescein-conjugated goat antirabbit IgG (CAPPEL; ICN Pharmaceuticals), for 1 h at 37°C. The slides were rinsed three times with PBS and mounted with buffered glycerol for fluorescent microscopic examination. Primary and secondary antibodies were diluted at 1:100 and 1:200 in PBS containing 3% bovine serum albumin, respectively.
Western Blotting-In order to monitor the phosphorylation of IB, HT29 cells were stimulated with TNF or agonist anti-LT␤R mAb, and the cells were lysed in 350 l of ice-cold lysis buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, 0.2% Nonidet P-40, 10 mM sodium fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, and 1 g/ml pepstatin A). To evaluate the protein level of p65, 293 cells were harvested 48 h after transfection. Whole cell extracts were lysed in 350 l of ice-cold lysis buffer. In order to evaluate the protein level of p65, NIK(KM), IKK␣(KM), and IB(S32A/S36A), 48 h after transfection, cells were stimulated by 2 g/ml of agonist anti-LT␤R mAb for 40 min, lysed, cleared by centrifugation, and determined for the protein concentration using Bio-Rad DC protein assay kit (Bio-Rad). The cell lysate was resolved by SDS-PAGE and transferred on polyvinylidene difluoride membranes (Millipore). The membranes were incubated with antibodies to anti-IB␣ (New England Biolabs), anti-phospho-IB␣ (New England Biolabs), anti-␤-tubulin (MONOSAN), anti-Gal4 (Santa Cruz Biotechnology), anti-FLAG epitope (M2 antibody; Sigma), or anti-Myc epitope (Santa Cruz Biotechnology). The immunoreactive proteins were visualized by enhanced chemilluminescence (ECL) (Amersham Biosciences).
Immunoprecipitation-To detect the phosphorylated p65 at Ser-536, HT29 cells were transfected with the indicated plasmids including pcDNA3.1(Ϫ)-FLAG-p65, treated with 2 g/ml of anti-LT␤R mAb for 40 min, and lysed by incubation at 4°C for 30 min in 2 ml of ice-cold lysis buffer (50 mM Tris, pH 7.8, 300 mM KCl, 1 mM EDTA, 10% of glycerol, 0.3% Nonidet P-40, 1ϫ Complete (Roche Molecular Biochemicals), 5 mM sodium fluoride, 0.4 mM sodium orthovanadate). The lysates were cleared by centrifugation, and the supernatants were incubated with anti-FLAG M2 affinity gel (Kodak) for 1 h at 4°C. The beads were washed five times with 1 ml of lysis buffer, and the bound proteins were eluted with an equal volume of 2ϫ SDS loading buffer and resolved on 7.5% SDS-PAGE. Western blot was conducted by using anti-phospho-p65 NF-B (Ser-536) antibody (Cell Signaling). Fig. 1A, the effects of TNF and LT␤R signaling on NF-B-mediated gene expression were compared. TNF stimulated NF-B-dependent gene expression in both 293 and HT29 cells. However, the agonistic LT␤R mAb stimulated gene expression only in the LT␤R-expressing HT29 cells as reported (54). Overexpression of NIK stimulated gene expression in both cells, indicating that the differences in these cells depend on LT␤R. In addition, overexpression of IKK␣ alone did not significantly activate the NF-B-dependent gene expression in both cells whereas that of IKK␤ activated the gene expression by 3-and 1.8-fold in 293 and HT29 cells, respectively. In fact, whereas IKK␤ overexpression induced IB␣ degradation, IKK␣ overexpression did not (data not shown). These findings suggested that the upstream signal is required for optimal activation as previously reported (24,26).

Activation of NF-B-mediated Gene Expression by TNF and LT␤R-In
As demonstrated in Fig. 1B, TNF stimulation induced the phosphorylation of IB␣ and subsequent degradation in HT29 cells. However, stimulation of LT␤R did not induce either IB␣ phosphorylation or its degradation, yet induced NF-B-dependent gene expression. Moreover, whereas TNF induced the nuclear translocation of NF-B in HT29 cells, LT␤R did not (Fig.  1C). The presence of NF-B in the nucleus even in the resting cells has been demonstrated recently by Birbach et al. (44) (see also Ref. 55 for review). These findings suggest that the LT␤R signaling stimulates NF-B activity without inducing IB␣ degradation or NF-B nuclear translocation as reported by Yin et al. (5) and that the activation of NF-B by the LT␤R signaling is not through alteration of intracellular localization of NF-B but presumably by augmenting its transcriptional activity.
NIK-IKK␣ Plays a Key Role in Regulating the Transcriptional Activity of NF-B during LT␤R Signaling-In a series of experiments using transient B luciferase assays, we have explored the kinase responsible for NF-B activation in the LT␤R cascade by using the dominant-negative kinase mutants of NIK (NIK(KM)), IKK␣ (IKK␣(KM)), and IKK␤ (IKK␤(KM)). As demonstrated in Fig. 2A, when NIK(KM), IKK␣(KM), or IKK␤(KM) were overexpressed in 293 cells, the TNF-induced NF-B activation was greatly inhibited by either of these mutants, more remarkable by IKK␤(KM). A similar observation was obtained with HT29 cells (data not shown). These results are consistent with the fact that the TNF-mediated NF-B activation was abolished in the IKK␤ gene knock-out mice but could not entirely be abolished in IKK␣ and NIK knock-out mice (5, 56 -60). Interestingly, the induction of NF-B-dependent gene expression by agonist anti-LT␤R mAb was strongly inhibited by NIK(KM) or IKK␣(KM) in HT29 cells but not by IKK␤(KM). In Fig. 2B, synergistic activation of gene expression was investigated. When wild-type NIK, IKK␣, or IKK␤ were overexpressed together with TNF signaling in 293 cells, there was no significant augmentation by IKK␣ as compared from TNF alone. However, either NIK or IKK␤ augmented the effect of TNF in inducing NF-B-dependent gene expression, which were statistically significant (p Ͻ 0.01 and p Ͻ 0.05, respectively). On the other hand, in HT29 cells, the gene expression elicited by anti-LT␤R mAb was augmented significantly by IKK␣ and NIK (p Ͻ 0.05 and p Ͻ 0.01, respectively) but not at Since 293 cells do not have LT␤R, we used HT29 cells to assess the effect of LT␤R signaling as reported previously (49). The luciferase activity was normalized by Renilla luciferase activity that was co-transfected as an internal control. The data are presented as the fold increase in luciferase activities (mean Ϯ S.D.) relative to control transfection of three independent experiments. B, phosphorylation and degradation of IB␣ induced by TNF but not by agonistic anti-LT␤R mAb. HT29 cells were stimulated with 10 ng/ml of TNF or 2 g/ml of anti-LT␤R mAb, the levels of IB␣ protein and its phosphorylated form were detected by Western blotting using specific antibodies. The Western blotting using anti-␤-tubulin antibody indicated that equivalent amounts of protein prepared from each fraction were resolved on each loading. C, nuclear localization of p65 in HT29 cells induced by TNF but not by anti-LT␤R mAb. After treating with 10 ng/ml of TNF for 15 min or 2 g of agonistic anti-LT␤R mAb for 40 min, HT29 cells were immunostained using primary (rabbit polyclonal antibody against p65) and secondary (fluorescein-conjugated goat anti-rabbit IgG) antibodies, and the intracelluar location of p65 was examined by fluorescence microscopy. all by IKK␤. These data collectively indicated that the TNFinduced NF-B activation is mainly through IKK␤ but the NF-B activation in LT␤R pathway is mediated by NIK and IKK␣, which was consistent with the previous study (4,5). The results of Fig. 2C demonstrated that the synergism between NIK and IKK␣ or IKK␤ was observed irrespective of the presence or absence of LT␤R in cells. Moreover, the abolishment of the effect of NIK by IKK␣(KM), not by IKK␤(KM), was observed equally in both cells. These observations suggest that activation of NIK is mainly coupled with IKK␣ but not IKK␤.
Involvement of the p65 C-terminal TA Domain in Signaling Mediated by NIK-IKK␣-Since NIK-IKK␣ was shown to activate the NF-B-dependent gene expression by augmenting the transcriptional activity of NF-B independently of the IB degradation pathway, we examined whether the p65 subunit is directly involved. In Fig. 3, we adopted a heterologous luciferase reporter system with Gal4-luc from which gene expression is under the control of Gal4. As shown in Fig. 3B, pM-p65, expressing the Gal4-p65 (full-length) fusion protein, aug-mented the gene expression from the Gal4-dependent promoter when NIK was overexpressed.

FIG. 2. The essential roles of NIK and IKK␣ in the transcriptional activation of NF-B induced by LT␤R signaling.
A, distinct inhibition profiles by kinase-defective mutants of NIK and IKK in the LT␤R and TNF signaling. Their effects on the NF-B-dependent luc gene expression were evaluated. Expression plasmids for dominant-negative NIK (pcDNA3-NIK(KM)), IKK␣ (pCR2FL-IKK␣(KM)), or IKK␤ (pCR2FL-IKK␤ (KM)) were cotransfected with 4ϫ Bwluc, and the extent of stimulation was compared when NF-B was activated either by TNF or agonistic anti-LT␤R mAb. After transfection with the indicated plasmids, 293 cells and HT29 cells were stimulated by TNF (10 ng/ml) for 24 h and anti-LT␤R mAb (2 g/ml) for 10 h, respectively. Note that the LT␤R signaling was blocked by NIK(KM) or IKK␣(KM) but not by IKK␤(KM). B, synergism between the signaling effectors in NF-B activation. In 293 cells, the synergistic activation was examined between TNF and wild-type NIK, IKK␣, or IKK␤. Similarly, in HT29 cells, the synergistic activation was examined between anti-LT␤R mAb and IKK␣ or IKK␤. C, synergism between NIK and the downstream kinases in the NF-B activation. NIK was overexpressed together with wild-type IKK␣, IKK␤, or their kinase-defective mutants, and the effect on NF-B-dependent gene expression was determined. Note that IKK␣ and IKK␤ augmented the effect of NIK, yet only IKK␣(KM) inhibited the effect of NIK. There was no effect of IKK␣, IKK␤, or NIK overexpression on the levels of endogenous p65 (data not shown). The data are presented as the fold increase in luciferase activities (mean Ϯ S.D.) relative to the control of three independent transfections. only the TA1 domain had the highest susceptibility for the NIK-mediated transactivation (19.7-fold, comparing lanes 15 and 16) although the basal transcription level was relatively low. pM-p65-(1-286) containing only RHD and pM-p65-(286 -430) lacking both TA2 and TA1 supported no effect of NIK. These results suggested that the effect of NIK was mainly mediated by TA1.
As it was demonstrated that the activation of NIK was coupled with IKK␣ (4, 61) (Fig. 2C), we next examined whether the dominant-negative IKK␣ mutant could block these effects of NIK. As shown in Fig. 3C, NIK-mediated activation of the transcriptional activity of Gal4-p65 fusion proteins was inhibited by the overexpression of IKK␣(KM). There was no significant effect with IKK␤(KM) (data not shown). Similar results were obtained with HT29 cells (data not shown). These data collectively demonstrated that NF-B transcriptional activation elicited by NIK-IKK␣ was mediated through the C-terminal TA1 domain of p65.
Serine 536 in the p65 TA1 Domain Is Responsible for the Effect of NIK-Since the effect of NIK-IKK␣ on p65 was primarily mediated by the TA1 domain of p65, we further examined the effect of mutation in Ser-536 within TA1. We also addressed whether IB␣ could block the effect of NIK since it was recently demonstrated that IB␣ is present in the nucleus and exhibits the inhibition of NF-B transcriptional activity (41, 42, 44 -47). In Fig. 4, the Gal4-luc reporter plasmid was co-transfected with pM-p65-(521-551), pM-p65-(521-551: S529A), and pM-p65-(521-551:S536A) with or without pcDNA3-NIK (expressing the wild-type NIK). When IB␣⌬N (a superactive mutant of IB␣) was expressed, the effect of NIK was inhibited. Although NIK stimulated the transcriptional activities of pM-p65-(521-551) and pM-p65-(521-551:S529A) similarly as in Fig. 3, B and C, the extent of stimulation was significantly reduced with pM-p65-(521-551:S536A), indicating that Ser-536 is indispensable for the transcriptional activity of p65 in response to the LT␤R signaling mediated by NIK-IKK␣. These findings indicated that the effect of NIK on the p65 TA1 domain might depend on the phosphorylation of p65 at serine 536, and this action of NIK could be inhibited by IB␣.
Phosphorylation of p65 at Ser-536 in LT␤R Pathway-In Fig.  5, we further examined whether Ser-536 in p65 is essential for the LT␤R signaling involving NIK and IKK␣,and could be phosphorylated in HT29 cells when stimulated with the agonist anti-LT␤R mAb. We first addressed whether the mutant p65, in which Ser-536 is substituted by Ala, is still responsive to LT␤R signaling. As demonstrated in Fig. 5A, although anti-LT␤R mAb stimulated the transcriptional activity of pM-p65-(521-551), its action was completely abolished when Ser-536 was substituted by Ala. In addition, when dominant-negative mutants of NIK and IKK␣ were expressed, this action of LT␤R signaling was blocked, suggesting that the effect of LT␤R is mediated by NIK and IKK␣ leading to the phosphorylation at Ser-536 in p65. We examined more directly whether Ser-536 is phosphorylated in response to the LT␤R signaling (Fig. 5B). When full-length p65 (FLAG-tagged) was expressed, Ser-536 phosphorylation was detected by the specific antibody (antiphospho-p65 NF-B (Ser-536)), and this phosphorylation was blocked upon coexpression of dominant-negative mutants IKK␣(KM)) inhibited the effect of NIK on the transcriptional activity of Gal4 fusion p65 mutants. 293 cells were transfected with Gal4-luc, pcDNA3-NIK in the presence or absence of pCR2FL-IKK␣(KM) (50 ng) together with various pM-p65 constructs. The data are presented as the relative folds relative to control transfection of three independent experiments. Similar observation was obtained in HT29 cells (data not shown). of NIK or IKK␣, or phosphodefective IB␣ (Myc-tagged IB␣(S32A/S36A)). These data clearly indicate that LT␤R signaling eventually leads to the phosphorylation of p65 at Ser-536. Both NIK and IKK␣ are involved, and this process can be blocked by IB␣.

DISCUSSION
In this study we have explored a mechanism by which NF-B is activated by LT␤R signaling. We found that both NIK and IKK␣ are critically involved in this pathway since either the kinase-deficient mutant of NIK or IKK␣, but not IKK␤, could block LT␤R-mediated NF-B activation and that the LT␤Rmediated NF-B activation did not induce phosphorylation and subsequent degradation of IB␣. We also found that serine 536 phosphorylation in its TA1 transactivation domain is essential. These findings collectively demonstrate the presence of a novel signaling mechanism in the NF-B activation, which is unique to the LT␤R signaling.
Matsushima et al. (4) found that NIK and IKK␣ were indispensable for NF-B activation in the LT␤R signaling using aly/aly mice in which formation of the secondary lymphoid tissues was affected although the TNF-and IL-1-mediated NF-B activation signaling remained intact. Similar observations were reported with IKK␣-deficient mice (57) and NIKdeficient mice (5). In addition, Cao et al. (62) has recently reported another interesting feature of the IKK␣ pathway with the mutant IKK␣ knock-in mice that the signal transduction of receptor activator of NF-B (RANK) leading to NF-B activation was abolished and thus the inducible expression of target cyclin D1 gene was affected. Whereas the responsiveness to proinflammatory stimuli including TNF, IL-1, and LPS are largely dependent on the IKK␤, other stimuli such as LT␤, RANK ligand, and Blys/BAFF depend on IKK␣-mediated signaling (55). These biological features of IKK␣ explain the characteristic developmental defect of the secondary lymphophoid tissues in the IKK␣ knock-out mice. Thus, although IKK␣ and IKK␤ are cross-related and both serve as the catalytic subunits of IKK complex, these findings illustrate the functional heterogeneity of IKK␣ and IKK␤.
Although a major step that regulates NF-B activity is the removal of IB from the NF-B/IB complex, the capacity of nuclear NF-B to drive transcription is also a regulated process. A number of studies support the possibility that p65 phosphorylation regulates the transcriptional competence of nuclear NF-B (41,(63)(64)(65)(66)(67). Although the role of PKA in phosphorylating p65 is still controversial (68 -70), regulation of the transcriptional competence of p65 by phosphorylation has been widely accepted. Protein kinases such as CKII, PKC, and IKK have been implicated in this process (40 -42, 67). For example, Wang and Baldwin (40) reported that the phosphorylation of Ser-529 at the TA1 domain of p65 is associated with the TNF-induced NF-B activation. They later found that CKII interacts with p65 and directly phosphorylates p65 at Ser-529 (41). In addition, Sakurai et al. (42) reported that TNF induced phosphorylation of p65 at Ser-536 in the cell and showed that the p65 could be phosphorylated at Ser-536 by IKK␤ at least in vitro. Moreover, Madrid et al. (43) have recently demonstrated that phosphatidylinositol 3-kinase activates Akt, which subsequently activates IKK␣ and leads to p65 phosphorylation at Ser-536.
One of the possible mechanisms of p65 phosphorylation at its TA domain in controlling its transcriptional competency is to recruit coactivator proteins such as histone acetyl transferases (71,72) and TLS (73) to NF-B when it binds to the target promoter sequence. Alternatively, p65 phosphorylation may preclude the recruitment of corepressor proteins such as Groucho family proteins that is known to interact with the p65 TA domain (48) and histone deacetylases (HDACs) (74 -77). For example, it was reported that cAMP-dependent kinase (PKA)mediated phosphorylation of p65 caused the p65 association with CBP in vitro (72). The same group has recently demon- Effects of IB⌬N were also examined. The constructs of these plasmids are described in the legend to Fig. 3. Lower panels show the results of Western blotting using anti-Gal4 antibody indicating that equivalent amounts of pM-p65 TA1 and its mutants were expressed in each transfection irrespective of cotransfection with pcDNA3-NIK or pcDNA3-IB⌬N. strated with cultured cells that p65 was associated with HDAC-1 in unstimulated cells, and it was dissociated from HDAC1 but associated with CBP upon cotransfection with the PKA catalytic subunit (77). Thus, it is possible that p65 phosphorylation may act as a determinant for selecting the interacting partner of NF-B.
The results in this study revealed that LT␤R signaling induced the p65 phosphorylation at Ser-536 by using phosphorylation-specific antibody. This finding was confirmed with the Ser-536 mutant of p65 TA1, which could not mediate the effect of LT␤R signaling. Then, where is NF-B (p65) phosphorylated in the cell? In fact, a number of studies have revealed that NF-B and IB shuttle in and out of the nucleus (44 -47). Therefore, NF-B is present in the nucleus even in the unstimulated cells, although to a lesser amount than that in the cytoplasm. More importantly, Birbach et al. (44) found that the treatment of cells with leptomycin B, an inhibitor of CRM1 and a blocking agent of nuclear export, resulted in the nuclear accumulation of NIK and IKK␣, but not IKK␤, indicating that these kinases also shuttle between the cytoplasm and the nucleus. IKK␣ has been initially identified as NIK-interacting protein in yeast two-hybrid screens (36). Thus, IKK␣ appears to preferentially associate with NIK where the large IKK complex is not found, such as in the nucleus. Interestingly, when IKK␣ was mutated at lysine 44, the shuttle of IKK␣ between cytoplasm and nucleus was prevented because it is known that the lysine residue at position 44 was also essential for the kinase activity (44), which is consistent with our observation that either dominant-negative IKK␣ (IKK␣(KM)) or phosphorylation-defective mutant IB␣ (IBs⌬N) efficiently blocked LT␤R signaling. Together with our findings, it is likely that the p65 subunit of NF-B is phosphorylated by the NIK/IKK␣ cascade in the nucleus.