Full transcriptional activity of the nuclear, DNA-bound form of NF-κB requires additional posttranslational modifications. In this study, we systematically mapped the T cell costimulation-induced phosphorylation sites within the C-terminal half of the strongly trans-activating NF-κB p65 subunit and identified serine 536 as the main phosphorylation site. The transient kinetics of serine 536 phosphorylation paralleled the kinetics of IκBα and IκB kinase (IKK) phosphorylation and also mirrored the principle of T cell costimulation. The TCR-induced pathway leading to serine 536 phosphorylation is regulated by the kinases Cot (Tpl2), receptor interacting protein, protein kinase Cθ, and NF-κB-inducing kinase, but is independent from the phosphatidylinositol 3-kinase/Akt signaling pathway. Loss-of-function and gain-of-function experiments showed phosphorylation of p65 serine 536 by IKKβ, but not by IKKα. Phosphorylation occurs within the cytoplasmic and intact NF-κB/IκBα complex and requires prior phosphorylation of IκBα at serines 32 and 36. Reconstitution of p65−/− cells either with wild-type p65 or a p65 mutant containing a serine to alanine mutation revealed the importance of this phosphorylation site for cytosolic IκBα localization and the kinetics of p65 nuclear import.
Full activation of T lymphocytes depends on stimulation of the TCR and the occupancy of auxiliary receptors such as CD28 (1). At the biochemical level, costimulation is typically exemplified at the level of NF-κB activation, which needs input from both receptors for full activation (2). The NF-κB/Rel transcription factor family is essential for the maturation of T cells and regulation of the survival and activation of mature T cells (3, 4). Thus, NF-κB-regulated target genes essentially contribute to the innate and adaptive immune response, as revealed by mouse models and the association of dysregulated NF-κB with human immunodeficiency syndromes (5). The signaling pathway required for NF-κB activation induced by T cell costimulation depends on the activation of membrane-proximal tyrosine kinases of the Src and Syk families, which allow the inducible formation of multiprotein complexes (6). These complexes inducibly translocate to sphingolipid- and cholesterol-rich membrane microdomains, termed lipid rafts. The multiprotein complexes contain several proteins required for NF-κB activation: the adapter protein Src homology 2 domain-containing leukocyte phosphoprotein 76 (SLP-76) 3 and proteins with enzymatic activity such as phospholipase Cγ and the exchange factor Vav1 (7, 8, 9). Vav1 contributes to lipid raft recruitment of protein kinase Cθ (PKCθ), an essential component of the NF-κB activation pathway that needs raft translocation for its enzymatic activity (10). Downstream from PKCθ, the caspase recruitment domain (CARD) proteins CARD11/CARMA1, MALT1, and Bcl10 relay TCR-derived signals to the IκB kinase (IKK) complex (11, 12, 13, 14, 15).
The IKK complex is composed of three subunits: the two kinases IKKα and IKKβ and the regulatory subunit IKKγ/NF-κB essential modulator (NEMO). IKKα also has other functions and contributes to stimulation of gene expression upon phosphorylation of histone H3 (16, 17), while IKKβ and IKKγ/NEMO are critical for induced phosphorylation of IκB (18). Phosphorylation tags IκB for ubiquitination and proteasomal degradation, thus resulting in the accumulation of NF-κB in the nucleus and induced transcription of target genes. There is recent evidence for an IκB-independent, so-called noncanonical pathway that is induced in response to a subset of stimuli, including lymphotoxin β (LTβ), LPS, and B cell activating factor belonging to the TNF family. These stimuli lead to NF-κB-inducing kinase (NIK)- and IKKα-dependent processing of the p100 precursor protein, thus generating DNA-binding p52 subunits (18, 19).
Once activated, inducible posttranslational modifications including acetylations and phosphorylations allow the regulation of NF-κB transcriptional activity (20, 21). The kinases implicated in the regulation of NF-κB trans activation include NIK, as NIK−/− mice show normal induction of DNA-binding, but a defective induction of NF-κB target genes in response to LTβ (22). Also, phosphatidylinositol 3-kinase (PI3K) and protein kinase B (PKB)/Akt were suggested to regulate the transcriptional activity of p65 in response to IL-1 stimulation (23). The strongly trans-activating p65 subunit is inducibly phosphorylated at serines 276 and 311 within the Rel homology domain. Serine 276 can be phosphorylated by protein kinase A catalytic subunit (24) or mitogen- and stress-activated protein kinase 1 (25). Phosphorylation of this site enhances p65 DNA binding and interaction with the transcriptional coactivator CBP/p300 (26). Serine 311 can be inducibly phosphorylated by PKCζ, which is important for coactivator binding and recruitment to their target promoters (27). The two C-terminal trans activation domains contain three known inducible phosphorylation sites. p14ARF-mediated repression of p65 trans activation requires threonine 502 phosphorylation (28) within p65 trans activation domain 2 (TA2). TNF-α stimulation induces phosphorylation of serines 529 and 536 contained within the C-terminal TA1 (29, 30). Phosphorylation of serine 536 contributes to NF-κB trans activation, and mutation of this serine to alanine impairs LPS- or LTβ-induced stimulation of NF-κB p65 trans activation (31). Serine 536 can be phosphorylated by the IKKs (29), but there is also evidence for phosphorylation by calmodulin-dependent protein kinase IV (32) or the NF-κB-activating kinase/NF-κB-activating kinase-associated protein 1 complex (33).
In this study, we investigated inducible phosphorylation of NF-κB p65 trans activation domains and the involved signaling pathways in costimulated T cells. These experiments identified serine 536 as the main inducible phosphorylation site. Phosphorylation of this site is mediated by IKKβ and occurs within the cytosolic p50/p65/IκBα complex. Reconstitution of p65−/− cells revealed the contribution of this phosphorylation site for the kinetics of p65 nuclear import and cytosolic IκBα retention.
Materials and Methods
Cell culture, transfections, and stimulations
Jurkat T leukemia cells and their derivatives deficient and repleted for SLP-76 (J14-v-29 and J14-76-11, respectively) (34) or IKKγ (35) or Jurkat cells overexpressing dominant-negative forms of IκBα or receptor interacting protein (RIP) (36) were grown at 37°C and 5% CO2 in supplemented RPMI 1640 medium; retransfected cells were cultured in the presence of G418 (1 mg/ml). A total of 1.5 × 107 Jurkat cells was transfected by electroporation using a gene pulser (Bio-Rad, Hercules, CA) at 250 V/950 μF with constant amounts of DNA. Costimulation of Jurkat cells was performed in a final volume of 500 μl (2 × 107 cells) by adding αCD3 (clone OKT3) and αCD28 (clone 15E8) at a final concentration of 1.5 μg together with 1 μg of protein A from Staphylococcus aureus for cross-linking. Mouse embryonic fibroblasts stably expressing wild-type p65 or p65 with a serine 536 to alanine mutation (37) were cultivated in DMEM supplemented with 10% (v/v) FCS and 1% (v/v) penicillin/streptomycin containing 1 μg/ml puromycin.
Antisera, plasmids, and reagents
38) and the wild-type and kinase-inactive forms of NIK, IKKα, IKKβ, and PKCθ (39), PKB/Akt DD, PKB/Akt AAA (40), Cot, and Cot K/M (41) were described. Methyl-β-cyclodextrin, PMA, bisindolylmaleimide, ionomycin, and parthenolide were from Sigma-Aldrich; nystatin (Fluka, Buchs, Switzerland) and MG132 (Calbiochem, La Jolla, CA) were purchased from the indicated sources.
Cell extracts and Western blotting
Stimulation was terminated upon the addition of ice-cold PBS to the cells. After centrifugation, the cell pellet was resuspended in Nonidet P-40 lysis buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM PMSF, 10 mM NaF, 0.5 mM sodium vanadate, leupeptine (10 μg/ml), aprotinin (10 μg/ml), 1% (v/v) Nonidet P-40, and 10% (v/v) glycerol) and incubated on ice for 30 min. The cell debris was pelleted upon centrifugation with 13,000 rpm at 4°C for 10 min. Equal amounts of protein contained in the supernatant were further analyzed by reducing SDS-PAGE and Western blotting onto a polyvinylidene difluoride membrane (Millipore, Bedford, MA). After incubation with primary Abs and extensive washing, HRP-coupled secondary Abs were added. Immunoreactive bands were visualized by ECL, according to the instructions of the manufacturer (Amersham Pharmacia Biotech, Piscataway, NJ).
Purification of GST fusion proteins and kinase assays
The GST-p65 (285–470) vector was cloned upon ligation of the EcoRI/SalI fragment from a p65 expression vector into pGEX-4T-1. The vectors encoding GST-p65 (354–551), GST-p65 (354–551) serine 529/536 Ala, and GST-GST-p65 (354–551) serine 536 Ala (29) were described. GST fusion proteins were expressed in Escherichia coli BL21 cells and purified by affinity chromatography on glutathione-Sepharose 4B, according to standard protocols. For total cell extract kinase assays, cells were lysed in kinase/lysis buffer (20 mM HEPES/KOH, pH 7.4, 0.5% (v/v) Nonidet P-40, 20 mM MgCl2, 2 mM DTT, 10 mM NaF, 1 mM sodium vanadate, leupeptine (10 μg/ml), aprotinin (10 μg/ml), 10 mM β-glycerophosphate, and 1 mM PMSF), and after 20-min incubation cell debris was pelleted. The supernatant was supplemented with 5 μCi of [γ-32P]ATP, 40 μM ATP, and 4 μg of substrate protein still attached to reduced glutathione-coupled Sepharose beads, and the kinase reaction was allowed to proceed for 2 h at 30°C. After extensive washing with radioimmunoprecipitation buffer, substrate proteins were eluted from glutathione-Sepharose by boiling in 1× SDS sample buffer and analyzed by SDS-PAGE and autoradiography. The immune complex IKK kinase assay was done by immunoprecipitation of the IKK complex using αIKKα and αIKKγ Abs. The precipitate was washed three times in lysis buffer and twice in kinase buffer (20 mM HEPES/KOH, pH 7.4, 25 mM β-glycerophosphate, 2 mM DTT, 20 mM MgCl2). The kinase assay was performed in a final volume of 20 μl of kinase buffer containing 5 μCi of [γ-32P]ATP, 40 μM ATP, and the purified substrate proteins. After incubation for 20 min at 30°C, the reaction was stopped, separated by SDS-PAGE, and analyzed by autoradiography.
Nuclear and cytosolic proteins were separated upon resuspending pelleted cells (1.5 × 107) in 200 μl of cold buffer A (10 mM HEPES/KOH, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT, and 0.5 mM PMSF) by gentle pipetting. After incubation for 20 min on ice, 5 μl of 10% Nonidet P-40 was added and cells were lysed by vortexing. The homogenate was centrifuged for 30 s in a microfuge. The supernatant representing the cytosolic fraction was collected, and the pellet containing the cell nuclei was dissolved in 40 μl of buffer C (20 mM HEPES/KOH, pH 7.9, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 0.5 mM PMSF). The Eppendorf tubes were incubated for 15 min on ice and briefly vortexed every 3 min. After centrifugation for 10 min with 13,000 rpm at 4°C, the supernatant was modified by Western blotting.
T cell costimulation induces NF-κB p65 C-terminal phosphorylation selectively at serine 536
Although the T cell costimulation-induced pathways leading to IKK activation and NF-κB activation start to unfold, information on costimulation-induced p65 phosphorylation is still lacking. As a starting point, we mapped inducible p65 phosphorylation within the trans-activating C terminus. Jurkat T leukemia cells were either left untreated or stimulated with agonistic αCD3 (TCR) and αCD28 Abs. Total cell lysates were incubated with recombinant substrate proteins covering the entire C-terminal portion of p65 in the presence of [γ-32P]ATP and analyzed for substrate protein phosphorylation (Fig. 1⇓A). These experiments showed CD3/CD28-inducible phosphorylation of the GST-p65 (354–551) protein, but not of GST-p65 (285–470), thus mapping the phosphorylation site to the region between aa 470 and 551. The involvement of the known phosphorylation sites serine 529 and 536 within this domain was tested by performing similar in vitro kinase assays with the substrate GST-p65 (354–551) protein mutated either in serine 536 or serines 536 and 529. These experiments revealed that mutation of serine 536 largely impaired CD3/CD28-inducible p65 phosphorylation, thus identifying this residue within TA1 as the main phosphorylation site. We then investigated whether the p65 serine 536 kinase is a component of the IKK complex or another kinase such as the reported calmodulin-dependent protein kinase IV (32). To address this question, the IKK complex was immunoprecipitated from extracts of either unstimulated or costimulated T cells and assayed for its p65-phosphorylating activity (Fig. 1⇓B). CD3/CD28 costimulation allowed inducible phosphorylation of GST-p65 (354–551) by a kinase activity associated with the IKK complex. This kinase activity was strongly diminished upon mutation of serine 536 within the substrate protein. Full NF-κB activation in T cells needs input signals from the TCR and also from costimulatory receptors (2). To test whether the principle of costimulation is also reflected at the level of p65 serine 536 phosphorylation, T cells were treated with αCD3 and αCD28 Abs either alone or in combination. Stimulation of either CD28 or CD3 alone induced phosphorylation of p65 only faintly, as detected by a phospho-specific Ab that allows monitoring of phosphorylation of the endogenous p65 protein at serine 536 (Fig. 1⇓C). T cell costimulation upon simultaneous triggering of both receptors strongly enhanced phosphorylation of the endogenous p65 protein. This costimulation was also mirrored at the level of phosphorylation of the endogenous IκBα and IKK proteins (Fig. 1⇓C). Identical results were obtained with primary T cells isolated from human blood (data not shown). A kinetic analysis revealed maximal CD3/CD28-induced serine 536 phosphorylation at 15 min after costimulation and no detectable signal after 1 h (Fig. 1⇓D), showing that this posttranslational modification is early and transient. Phosphorylation steps occurred in a staggered fashion, as phosphorylation of the IKK activation loop was maximal at 5 min, while phosphorylation of IκBα and p65 peaked at 10 and 15 min, respectively.
CD3/CD28-induced p65 TA1 phosphorylation depends on SLP-76, RIP, and IκBα phosphorylation
To investigate the p65-phosphorylating signaling pathways, we used Jurkat T cells either lacking or expressing nonfunctional signaling proteins. Jurkat cells deficient in the adapter protein SLP-76 failed to induce p65 phosphorylation in response to T cell costimulation (Fig. 2⇓). Retransfection of SLP-76 restored CD3/CD28-induced p65 phosphorylation, demonstrating the importance of SLP-76 for this signaling pathway. As T cells from RIP2−/− mice show strongly impaired phosphorylation of IκBα and NF-κB DNA binding after TCR engagement (42), we investigated p65 phosphorylation in RIP−/− Jurkat cells stably retransfected with a dominant-negative form of RIP lacking aa 391–422. These cells did not show elevated p65 phosphorylation, implying the RIP protein as an important signaling mediator. Jurkat cells stably expressing a mutant form of IκBα not phosphorylatable at serines 32 and 36 also failed to display p65 phosphorylation, showing that IκBα phosphorylation is a prerequisite for p65 phosphorylation. The role of IKK activation for p65 TA1 phosphorylation was further investigated in Jurkat cells lacking the IKKγ/NEMO protein. As these cells lack a functional TCR (43), they were stimulated with the phorbol ester PMA and αCD28 Abs and analyzed for p65 phosphorylation. Although the absence of this regulatory protein precluded p65 phosphorylation, IKKγ/NEMO-retransfected cells showed inducible serine 536 phosphorylation, thus revealing the importance of the IKK complex for costimulation-induced p65 C-terminal modification. This notion is supported by the parallel analysis of inducible IκBα phosphorylation in the various cell lines (Fig. 2⇓), as these experiments revealed a perfect coincidence between CD3/CD28-induced phosphorylation of p65 and IκBα.
CD3/CD28-induced p65 serine 536 phosphorylation requires IKKβ activation
To collect further evidence for the role of IKKs by an alternative experimental approach, cells were pretreated with the IKK inhibitor parthenolide (44) and analyzed for costimulation-induced p65 and IκBα phosphorylation (Fig. 3⇓A). Parthenolide efficiently blocked phosphorylation of both proteins, arguing for the importance of IKK activation. Induced lipid raft compartmentation of signaling proteins is required for efficient T cell activation (45) and also for costimulation-mediated NF-κB activation. The role of lipid raft formation for CD3/CD28-induced p65 serine 536 phosphorylation was investigated upon pretreatment of T cells with either the cholesterol synthesis inhibitor nystatin or the cholesterol-depleting drug methyl-β-cyclodextrin, followed by the analysis of costimulation-induced p65 phosphorylation. Interference with lipid raft recruitment averted p65 serine 536 phosphorylation (data not shown), thus pointing to the relevance of lipid rafts for this signaling pathway. To investigate the relative role of IKKα or IKKβ for p65 phosphorylation, Jurkat T cells were transfected with expression vectors encoding dominant-negative forms of the IKKs together with very low amounts of the GFP-tagged p65 protein, allowing expression of this fusion protein at physiological levels. This experimental approach was taken, as the low transfection efficiency of Jurkat cells does not allow the analysis of the endogenous p65 protein, thus enabling the detection of the slower migrating GFP-p65 fusion protein, which is fully functional and regulated (38). CD3/CD28-induced p65 TA1 phosphorylation was retained in the presence of kinase-inactive IKKα K/A, but expression of kinase-dead IKKβ K/A alone was sufficient to block this phosphorylation completely (Fig. 3⇓B). Although these experiments show the necessity of IKKβ for p65 C-terminal phosphorylation, we next asked whether IKKβ is also sufficient to induce this modification. To address this question, a similar experimental approach was used by expressing IKKα and IKKβ together with GFP-p65 or Flag-tagged IκBα as a control (Fig. 3⇓C). Phosphorylation of p65 occurred after expression of IKKα and IKKβ, showing that these kinases are sufficient to trigger C-terminal p65 phosphorylation.
Costimulation-induced p65 serine 536 phosphorylation is affected by Cot, NIK, and PKC θ , but does not use the PI3K/Akt pathway
We then assessed the role of various kinases implicated in the TCR-dependent NF-κB activation (7) for their importance for NF-κB p65 phosphorylation. GFP-p65 was expressed in T cells along with active or inactive variants of the kinases. Expression of NIK, which plays a role in TCR-mediated NF-κB activation (46), resulted in serine 536 phosphorylation and further enhanced CD3/CD28-induced p65 phosphorylation (Fig. 4⇓A). In contrast, a dominant-negative form of NIK failed to interfere with CD3/CD28-induced p65 phosphorylation, indicating that NIK is not essential in this study. Similarly, expression of Cot, which contributes to CD3/CD28-induced NF-κB activation by physically assembling with NIK and IKKα (41), induced p65 phosphorylation (Fig. 4⇓B). Expression of a point-mutated, kinase-inactive form of Cot strongly impaired CD3/CD28-induced p65 phosphorylation. Because PKB/Akt is one of the activators of Cot (47) and the PI3K/Akt pathway has been implicated in the stimulation of phosphorylation and trans activation by p65 (23, 48), we tested the impact of PKB/Akt on p65 serine 536 phosphorylation. A constitutive active form of PKB/Akt failed to induce p65 phosphorylation, while a dominant-negative variant reduced p65 serine 536 phosphorylation only barely (Fig. 4⇓C). To investigate the importance of the PI3K/Akt pathway by a different experimental approach, costimulation-induced p65 phosphorylation was tested in the presence of wortmannin, a potent and specific PI3K inhibitor (49). Wortmannin failed to interfere with CD3/CD28-induced p65 phosphorylation (data not shown), suggesting that the PI3K/Akt pathway is not operational in CD3/CD28-induced p65 serine 536 phosphorylation. Because cooperation between PKB/Akt and PKCθ regulates NF-κB-dependent trans activation (50), we investigated the effect of the PKC inhibitor bisindolylmaleimide on T cell costimulation-induced p65 phosphorylation (Fig. 4⇓D). This PKC inhibitor completely prevented CD3/CD28-induced serine 536 phosphorylation, thus supporting the role of PKCs. To directly address the role of PKCθ, constitutively active (PKCθ A/E) or dominant-negative (PKCθ K/R) forms of PKCθ were coexpressed with GFP-p65 in unstimulated or costimulated T cells (Fig. 4⇓E). PKCθ A/E resulted in constitutive serine 536 phosphorylation that was further augmented by costimulation. In contrast, the expression of kinase-inactive PKCθ failed to completely prevent CD3/CD28-induced p65 phosphorylation, suggesting the involvement of further PKC isoforms such as PKCα (51). Similarly, the analysis of IκBα phosphorylation revealed incomplete inhibition of costimulation-induced phosphorylation in cells containing PKCθ K/R and induced phosphorylation in the presence of PKCθ A/E. Also, PMA alone was able to trigger p65 serine 536 phosphorylation, corroborating the relevance of PKCs for this process (data not shown).
T cell costimulation-induced NF-κB p65 serine 536 phosphorylation occurs within the intact cytosolic NF-κB/IκB complex
Given the importance of IKKβ for the phosphorylation of IκB and p65, we addressed the question as to whether p65 phosphorylation occurs within the intact NF-κB/IκB complex or after liberation of the DNA-binding subunits from IκBα. Jurkat T cells were either left untreated or preincubated with MG132, an inhibitor that prevents proteasomal IκBα degradation and thus stabilizes the phosphorylated form of IκBα in complex with the DNA-binding subunits. Cells were then stimulated either with αCD3/αCD28 Abs or PMA in combination with ionomycin. Western blotting revealed that induced phosphorylation of endogenous p65 was even enhanced in the presence of MG132, indicating that the phosphorylation takes place within the intact complex (Fig. 5⇓A, upper). Control experiments ensured the inhibition of IκBα degradation and enhancement of IκBα phosphorylation in the presence of MG132 (Fig. 5⇓A, lower). The cytosolic phosphorylation of p65 and the transient nature of this modification raise the questions as to whether phosphorylated p65 is also found in the nucleus and whether the nuclear import kinetics of phosphorylated p65 is altered upon stabilization of the phosphorylated NF-κB/IκBα complex by MG132.
To address this question, cells were either left untreated or pretreated with MG132 and stimulated for various time periods with αCD3/αCD28 Abs, followed by subcellular fractionation into cytosolic and nuclear extracts (Fig. 5⇑B). The main fraction of phosphorylated p65 occurred in the cytosol and only a minor fraction was found in the nucleus. Nuclear serine 536 phosphorylated p65 was detectable already 10 min after stimulation. Thirty minutes after stimulation, phosphorylated p65 was not detectable in the cytosol and only occurred in the nucleus. Nuclear immigration of p65 was delayed in the presence of MG132. Control Western blots showed MG132-mediated IκBα stabilization and the purity of cytosolic and nuclear fractions, as revealed by the marker proteins HDAC-1 and p105, respectively. To directly compare the differential distribution of phosphorylated and nonphosphorylated p65 between cytosol and nucleus by a semiquantitative approach, a constant volume of nuclear extract and increasing volumes of cytosolic extract were separated on the same SDS gel and analyzed by immunoblotting (Fig. 5⇑C). An αp65 Ab revealed comparable amounts of p65 in 20 μl of nuclear and 2.5 μl of cytosolic extracts. Analysis of p65 phosphorylation using the phosphospecific Ab showed resembling serine 536 phosphorylation in 20 μl of nuclear extracts and only 0.5 μl of cytosolic extracts. Thus, the number of p65 molecules phosphorylated in the cytosol is ∼5 times higher in the cytosol, when compared with the nucleus.
Serine 536 phosphorylation controls the kinetics of NF-κB p65 nuclear import
The predominant cytosolic occurrence and short duration of p65 phosphorylation argue against a prominent role of this posttranslational modification for NF-κB-dependent gene expression, which requires longer time periods (52). To investigate the importance of p65 serine 536 phosphorylation for nuclear import, we used p65−/− cells stably retransfected with wild-type p65 or with a p65 mutant in which serine 536 was changed to an alanine. As the embryonic lethality of p65−/− mice precludes the use of p65−/− T cells (53), these experiments were performed with TNF-α-stimulated fibroblasts. Cells expressing either the wild-type or point-mutated form of p65 were stimulated for various periods and cells were fractionated into nuclear and cytosolic extracts (Fig. 6⇓). A kinetic analysis of p65 nuclear import by immunoblotting revealed faint amounts of mutant p65 in unstimulated cells and a faster and increased nuclear import of point-mutated p65 at the early time points 7.5 and 15 min after stimulation, when compared with the wild-type protein. These differences disappeared after stimulation periods exceeding 30 min, showing that p65 phosphorylation negatively regulates the kinetics of p65 import. Concomitantly, cells expressing the mutant form of p65 showed a significantly higher amount of IκBα in the nucleus of unstimulated cells, suggesting a contribution of serine 536 for cytosolic retention of IκBα. Thus, we suggest that CD3/CD28-induced phosphorylation of p65 at serine 536 contributes to control the kinetics of nuclear import and allows a fine tuning of the NF-κB-mediated transcriptional response.
In this study, we identified NF-κB p65 serine 536 as the main phosphorylation site in the trans-activating C terminus. However, it cannot be ruled out that at different time points, other amino acids are also phosphorylated. Only serine 536, but not the neighboring serine 529, is evolutionary conserved between mammals and amphibians (Fig. 7⇓A), arguing for the functional importance of this site.
This study reveals that T cell costimulation-induced p65 TA1 phosphorylation depends on the kinases RIP and Cot and is regulated by NIK and PKCθ. The importance of RIP, NIK, and PKCθ for TCR-mediated NF-κB activation was revealed in animal models. Accordingly, RIP2−/− mice show strongly impaired phosphorylation of IκBα and NF-κB DNA binding after TCR engagement (42). T cell costimulation results in the association of RIP2 with Bcl10 and induces the phosphorylation of this CARD domain protein (54). Because RIP2 is also important for signals triggered by the Toll-like receptor 2 and 4, this kinase operates in several pathways. A role for NIK was revealed by the analysis of T cells from aly mice, a strain with mutant NIK. TCR stimulation resulted in an impaired NF-κB DNA binding in immature and mature T cells, while responses to costimulatory signals were not changed (46). Support for a role of NIK in serine 536 phosphorylation comes from mapping experiments using various portions of p65 fused to Gal4, which identified TA1 as the NIK-responsive region (55). The relevance of PKCθ was revealed by the analysis of knockout mice, which fail to induce IκB degradation after costimulation of mature T cells, while NF-κB activation induced by IL-1 or TNF-α remains unaffected (56). A dominant-negative form of Cot selectively prevents CD3/CD28-induced NF-κB activation, but does not interfere with TNF-α-derived signals (41, 55). Cot−/− mice show no defects in NF-κB activation in response to LPS treatment (57) or CD40 ligation (58), but the role of this kinase in TCR-induced NF-κB activation remains to be tested. A schematic summary of the signaling pathway mediating p65 serine 536 phosphorylation in T cells is shown in Fig. 7⇑B. Our data suggest that the Cot-activating PI3K/Akt does not play a role for CD3/CD28-induced p65 phosphorylation. The involvement of the PI3K/Akt pathway for p65-mediated trans activation has been proposed for IL-1-stimulated HepG2 cells (23) and Ras-expressing NIH 3T3 cells (48). In contrast, other groups found no involvement of PI3K/Akt signaling for p65-dependent trans activation (20, 59) or p65 phosphorylation (60). Further evidence for a lacking relevance of the PI3K/Akt signaling pathway for p65 phosphorylation in T cells comes from the finding that Jurkat T cells show no elevated basal p65 serine 536 phosphorylation, although these cells display a constitutively active PI3K/Akt signaling pathway (61).
Our data suggest that serine 536 is directly phosphorylated by IKKβ, but not by IKKα. The observed inhibition of TNF-α-induced p65 phosphorylation by kinase-inactive IKKα in HeLa cells (62) argues for a cell- or stimulus-dependent role of IKKα. T cell costimulation-induced serine 536 phosphorylation is mediated by IKKβ or alternatively by a kinase that is controlled by IKKβ. We cannot rule out that the relevance of IKKβ for p65 phosphorylation is attributable to its IκBα-phosphorylating activity, which is a prerequisite for p65 phosphorylation (compare Fig. 2⇑). In such a case, IKKβ-mediated phosphorylation of IκBα would enable the subsequent p65 serine 536 phosphorylation by a different kinase. This model would be compatible with all the experiments, showing that interference with IKKβ activity would also prohibit p65 phosphorylation. In support of this, TNF-α still induces p65 serine 536 phosphorylation in IKKβ−/− cells (60). However, because the p65-phosphorylating kinase activity can be immunoprecipitated with the IKK complex and p65 phosphorylation occurs in the cytosolic IκB/NF-κB complex, it is also plausible that IKKβ directly phosphorylates p65 in stimulated T cells. Accordingly, in vitro experiments showed that p65 and IκBα are comparable substrates with similar Km and Kcat values for rIKKβ (62, 63). Thus, we suggest IKKβ as the T cell costimulation-induced p65 kinase, although the existence of further kinases acting on the same site is well possible.
Phosphorylation of p65 occurs within the cytosolic NF-κB/IκB complex, and only a minor fraction of phosphorylated p65 is found in the nucleus. The cytoplasmic phosphorylation and also the very fast phosphorylation kinetics suggest that p65 phosphorylation exerts its effects at the very early steps of NF-κB activation. Using reconstituted fibroblasts, this study unexpectedly reveals that a nonphosphorylatable form of p65 is recruited faster to the nucleus, implicating that TA1 phosphorylation serves to prolong the residence time of p65 in the cytoplasm. In addition, p65 serine 536 alanine-expressing cells showed elevated levels of nuclear IκB, which suggests that this TA1 modification contributes to the interaction of p65 with IκBα. IκBα occurs in the cytosol and the nucleus, while its ubiquitin ligase β-transducing repeats-containing protein is found exclusively in the nucleus (64), raising the possibility that the nucleus is actually the location of IκBα proteolysis (65). Does NF-κB p65 serine 536 phosphorylation affect transcription of NF-κB target genes positively or negatively? We could not address this question using Gal4-p65 one-hybrid proteins, as these constitutively nuclear fusion proteins were only marginally phosphorylated in response to T cell costimulation (data not shown). Although transcription of some NF-κB target genes such as IL-6 is elevated in p65 serine 536 alanine-expressing cells (37), kinetic studies suggest that these cells display differences in the amplitude and duration of the transcriptional response when compared with cells expressing wild-type p65 (C. Bucher and M. Lienhard Schmitz, unpublished results). Thus, it is conceivable that the outcome of serine 536 phosphorylation depends on each individual target gene.
We thank Drs. H. Sakurai, B. Seed, R. de Martin, and D. Krappmann for kindly providing plasmids or cell lines.
↵1 This work was supported by grants from the Deutsche Forschungsgemeinschaft (Schm 1417/3-1), Fonds der chemischen Industrie, European Union project (QLK3-CT-2000-00463) sponsored by the Schweizerisches Bundesamt für Bildung und Wissenschaft, Oncosuisse, Schweizerischer Nationalfonds, Association for International Cancer Research, the “Stiftung zur Förderung der wissenschaftlichen Forschung an der Universität Bern,” and the Roche Foundation. H.N. is supported by a grant from Human Frontier Science Program.
↵2 Address correspondence and reprint requests to Dr. M. Lienhard Schmitz, University of Bern, Department for Chemistry and Biochemistry, Freiestr. 3, CH-3012 Bern (Switzerland). E-mail address:
↵3 Abbreviations used in this paper: SLP-76, protein Src homology 2 domain-containing leukocyte phosphoprotein 76; CARD, caspase recruitment domain; GFP, green fluorescent protein; IKK, IκB kinase; LTβ, lymphotoxin β; NEMO, NF-κB essential modulator; NIK, NF-κB-inducing kinase; PI3K, phosphatidylinositol 3-kinase; PKB, protein kinase B; PKC, protein kinase C; RIP, receptor interacting protein; TA, trans activation domain.
- Received November 21, 2003.
- Accepted March 1, 2004.
- Copyright © 2004 by The American Association of Immunologists