Advertisement

Regulation of Nuclear Factor κB Transactivation

IMPLICATION OF PHOSPHATIDYLINOSITOL 3-KINASE AND PROTEIN KINASE C ζ IN c-Rel ACTIVATION BY TUMOR NECROSIS FACTOR α*
  • Angel G. Martin
    Affiliations
    Centro de Biologı́a Molecular “Severo Ochoa,” Universidad Autónoma de Madrid, Cantoblanco, Madrid 28049, Spain
    Search for articles by this author
  • Belén San-Antonio
    Affiliations
    Centro de Biologı́a Molecular “Severo Ochoa,” Universidad Autónoma de Madrid, Cantoblanco, Madrid 28049, Spain
    Search for articles by this author
  • Manuel Fresno
    Correspondence
    To whom correspondence should be addressed.
    Affiliations
    Centro de Biologı́a Molecular “Severo Ochoa,” Universidad Autónoma de Madrid, Cantoblanco, Madrid 28049, Spain
    Search for articles by this author
  • Author Footnotes
    * This work was supported by grants from Dirección General de Investigación Cientı́fica y Técnica, Fondo de Investigaciones Sanitarias, Comunidad Autónoma de Madrid, and Fundación Ramón Areces.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Open AccessPublished:May 11, 2001DOI:https://doi.org/10.1074/jbc.M011313200
      Transactivation by c-Rel (nuclear factor κB) was dependent on phosphorylation of several serines in the transactivation domain, indicating that it is a phosphorylation-dependent Ser-rich domain. By Ser → Ala mutational and deletion analysis, we have identified two regions in this domain: 1) a C-terminal region (amino acids 540–588), which is required for basal activity; and 2) the 422–540 region, which responds to external stimuli as tumor necrosis factor (TNF) α or phorbol myristate acetate plus ionomycin. Ser from 454 to 473 were shown to be required for TNFα-induced activation, whereas Ser between 492 and 519 were required for phorbol myristate acetate plus ionomycin activation. Phosphatidylinositol 3-kinase (PI3K) and protein kinase C (PKC) ζ were identified as downstream signaling molecules of TNFα-activation of c-Rel transactivating activity. Interestingly, dominant negative forms of PI3K inhibited PKCζ activation and dominant negative PKCζ inhibited PI3K-mediated activation of c-Rel transactivating activity, indicating a cross-talk between both enzymes. We have identified the critical role of different Ser for PKCζ- and PI3K-mediated responses. Interestingly, those c-Rel mutants not only did not respond to TNFα but also acted as dominant negative forms of nuclear factor κB activation.
      NF-κB
      nuclear factor κB
      DBD
      DNA binding domain, FBS, fetal bovine serum
      GST
      glutathione S-transferase
      IKK
      IκB kinase
      PAGE
      polyacrylamide gel electrophoresis
      PCR
      polymerase chain reaction
      PI3K
      phosphatidylinositol 3-kinase
      PMA
      phorbol myristate acetate
      TNF
      tumor necrosis factor
      TPCK
      l-1-chloro-3-[4-tosylamido]-4-phenyl-2-butanone
      WCE
      whole cell extract
      PKC
      protein kinase C
      HA
      hemagglutinin
      PIP3
      phosphatidylinositol 3,4,5-trisphosphate
      Transcription factors belonging to the nuclear factor κB (NF-κB)1 family regulate several of the most important genes induced during T cell activation (for review, see Ref.
      • Baldwin Jr., A.S.
      ). The NF-κB family of transcription factors is composed of homo- and heterodimers of a family of proteins, which include the Dorsal gene ofDrosophila and the mammalian genesnfκb1, nfκb2, c-rel, relA (p65), and relB (for review, see Ref.
      • Ghosh S.
      • May M.J.
      • Kopp E.B.
      ). All members share a conserved 300-amino acid region in their N terminus that includes the dimerization, nuclear localization, and DNA binding regions. c-Rel, RelB, and RelA also have C-terminal transactivation domains, which strongly activate transcription from NF-κB sites. NF-κB is rapidly activated by the T cell receptor complex, but, at later phases of T cell activation, autocrine or paracrine secreted TNFα takes control of NF-κB activation (
      • Pimentel-Muiños F.X.
      • Mazana J.
      • Fresno M.
      ). Tumor necrosis factor (TNF) α is a pleiotropic cytokine with biological effects ranging from promoting growth and differentiation to induction of apoptosis. Those effects rely, at least in part, in the activation of the transcription factor NF-κB (for review, see Ref.
      • Wallach D.
      • Varfolomeev E.E.
      • Malinin N.L.
      • Goltsev Y.V.
      • Kovalenko A.V.
      • Boldin M.P.
      ). In T cells, the initial phase of NF-κB activation after T cell receptor triggering mainly relies on p65 translocation, whereas the later phase is controlled by c-Rel. We have previously shown that autocrine or paracrine TNFα secretion controls the c-Rel levels in T cells (
      • Pimentel-Muiños F.X.
      • Mazana J.
      • Fresno M.
      ). Thus, c-Rel activation emerges as a key point for the later phase of T lymphocyte activation, a fact that is supported by the functional unresponsiveness of T lymphocytes from the c-Rel knock out mice (
      • Gerondakis S.
      • Strasser A.
      • Metcalf D.
      • Grigoriadis G.
      • Scheerlinck J.Y.
      • Grumont R.J.
      ,
      • Liou H.C.
      • Jin Z.
      • Tumang J.
      • Andjelic S.
      • Smith K.A.
      • Liuo M.L.
      ).
      NF-κB activity is regulated, at least in part, by its subcellular localization. Thus, functional NF-κB complexes are held in the cytoplasm of resting T cells in an inactive state complexed with members of the IκB family. In response to different activators, which include T cell receptor and TNFα, IκB is phosphorylated by IκB kinases (IKKs), and subsequently degraded, liberating the active NF-κB complex, which translocates to the nucleus and activates transcription (for review, see Ref.
      • May M.J.
      • Ghosh S.
      ). Recently, a second level of regulation of NF-κB activity independent of IκB, which relies in the activation of the transcriptional activity of p65, has been described (
      • Schmitz M.L.
      • dos Santos Silva M.A.
      • Baeuerle P.A.
      ,
      • Schmitz M.L.
      • Baeuerle P.A.
      ,
      • Wang D.
      • Baldwin Jr., A.S.
      ). Thus, the catalytic subunit of protein kinase A was shown to be bound to inactive NF-κB complexes, and upon IκB degradation this catalytic subunit phosphorylated p65, resulting in an enhanced transcription promoting activity (
      • Zhong H.
      • Voll R.E.
      • Ghosh S.
      ). Moreover, TNFα treatment of cells results in phosphorylation of Ser529 in the transactivation domain of p65, resulting in the activation of the transcriptional activity of the protein (
      • Wang D.
      • Baldwin Jr., A.S.
      ). The small GTP-binding protein Ras enhanced p65/RelA transcriptional activity through a pathway that required the stress-activated protein kinase p38 or a related kinase (
      • Norris J.L.
      • Baldwin Jr., A.S.
      ), although it was not demonstrated whether this kinase was directly involved in activating NF-κB or instead a transcriptional co-activator. The activity of Ras as well as the atypical protein kinase C ζ (PKCζ) has been also shown to be essential for the transcriptional activity of p65/RelA in endothelial cells (
      • Anrather J.
      • Csizmadia V.
      • Soares M.P.
      • Winkler H.
      ). This activation relies in the phosphorylation of the N-terminal Rel homology domain and not on the C-terminal transactivation domain. PKCζ was able to phosphorylate and activate IKK2 (
      • Lallena M.J.
      • Dı́az-Meco M.T.
      • Bren G.
      • Pay C.V.
      • Moscat J.
      ), thus demonstrating its direct implication in the NF-κB activation process by participating in IκB degradation. A recently identified 62-kDa protein (named p62) might function as a bridge between PKCζ and the TNF receptor-associated protein RIP (
      • Sanz L.
      • Sánchez P.
      • Lallena M.J.
      • Dı́az-Meco M.T.
      • Moscat J.
      ). On the other hand, PI3K activity seems to be required for interleukin-1- and TNFα-induced NF-κB activity (
      • Sizemore N.
      • Leung S.
      • Stark G.R.
      ,
      • Reddy S.A.
      • Huang J.H.
      • Liao W.S.
      ). The PI3Ks are a family of lipid kinases that catalyze the addition of a phosphate group to the 3′-OH position of the inositol ring of phosphoinositides. The 3-phosphoinositides are second messengers that exert specific regulatory functions inside the cells (
      • Toker A.
      • Cantley L.C.
      ). PI3K is composed of two different subunits, a regulatory subunit (p85) and a catalytic subunit, termed p110. Upon stimulation, p85 becomes associated to the cytosolic portion of tyrosine-phosphorylated receptors via its SH2 domains, which in turn promotes its association with the catalytic subunit p110 and its subsequent activation. The activation of PI3K triggers a signaling cascade that leads to the specific phosphorylation of p65/RelA subunit. This phosphorylation enhances p65-mediated transcription without affecting IκB degradation, nuclear translocation of NF-κB, or the ability of NF-κB to bind to DNA (
      • Sizemore N.
      • Leung S.
      • Stark G.R.
      ).
      Analysis of the transactivation domain of p65 by CD and NMR spectroscopy revealed no defined structure (
      • Schmitz M.L.
      • dos Santos Silva M.A.
      • Altmann H.
      • Czisch M.
      • Holak T.A.
      • Baeuerle P.A.
      ). Two differentiated acidic regions (termed TA1 and TA2) were identified as essential for its transcription promoting activity. Only TA2, however, was responsible for the activation by phorbol ester stimulation by a mechanism that involved phosphorylation of Ser residues (
      • Schmitz M.L.
      • dos Santos Silva M.A.
      • Baeuerle P.A.
      ). Moreover, the high Ser content of the transactivation domain of the avian c-Rel-related oncogene v-rel has been demonstrated to be essential for its transforming capabilities (
      • Chen C.
      • Agnës F.
      • Gèlinas C.
      ). Furthermore, the mutation of the Ser residue 471 in the human c-Rel transactivation domain abrogated TNFα-induced NF-κB activity in a Jurkat T cell clone (
      • Martı́n A.G.
      • Fresno M.
      ). Taken together, all these works point to a key functional role of the regulation of the transactivation domain of c-Rel family proteins for NF-κB function.
      In this work we have characterized the regulation of c-Rel transactivation domain. This domain seems to belong to the family of the phosphorylation-dependent Ser-rich acidic transactivation domains. We have revealed the critical role of several Ser residues for TNFα-dependent activation. Interestingly, mutations of those Ser residues not only abrogated c-Rel transactivating activity, but also acted as dominant negative forms of NF-κB activation, further stressing the importance of this regulation in the activity of NF-κB. Additionally, we have identified PI3K and PKCζ as enzymes participating in the signaling route that leads to c-Rel activation by TNFα.

      RESULTS

      Mapping of the c-Rel Transactivation Domain

      Mouse c-Rel transactivation domain has been previously localized in the C-terminal region of the protein, between positions 403 and 568 (
      • Bull P.
      • Morley K.L.
      • Hoekstra M.F.
      • Hunter T.
      • Verma I.M.
      ). In order to delineate the transcriptionally active region of human c-Rel we constructed several fusion plasmids between the Gal4 DBD and c-Rel, providing a system where transcriptional activity of this protein could be assayed without interference of IκB association and/or degradation. The fragment of c-Rel from position 309 to 588 was fused to Gal4 DBD, and several deletion mutants of this region were generated (Fig. 1). These constructs were transfected into Jurkat T cells along with a 5xGal4 Luc reporter plasmid and luciferase activity was recorded. As a negative control, a construction which covered only the c-Rel fragment from position 309 to 318 fused to Gal4 DBD was used.
      Figure thumbnail gr1
      Figure 1Deletion mapping of the C terminus of c-Rel. Left panel, structure of c-Rel transactivation domain deletion mutants. RHD, Rel homology domain; NLS, nuclear localization signal. Right panel, effect of deletion mutants on the transcriptional activity of c-Rel fusions to Gal4 DBD. Jurkat T cells were transfected with each of the mutants described and with a reporter plasmid containing five tandem repeats of Gal4 site upstream from the luciferase gene. Results are expressed as percentage of activity compared with the wild-type construct. Transfection efficiency was normalized using the Dual Luciferase assay (Promega). Additionally, similar amounts of the different construction were expressed in transfected cells as detected by EMSA assays (data not shown). The results shown are the mean ± S.D. of three independent experiments.
      Constructions spanning the c-Rel region from 309 to 421 had no basal transcriptional activity, as the control construction Gal4-(309–318) (Fig. 1, right). A minimal transcriptional activation was observed when Gal4-(309–455) was transfected. By contrast, Gal4-(422–588) induced an activity of the Gal4 reporter that was 236% of the wild-type fusion, indicating that this region possessed all the transcriptionally active sequences of c-Rel and even behaved as a better autonomous transactivation domain that the whole c-Rel C-terminal region (from position 309 to 588). Progressive deletions toward the C terminus were introduced in this region (Gal4-(456–588), Gal4-(498–588), and Gal4-(541–588)), which increasingly reduced the transcriptional activity. However, the smallest construction Gal4-(541–588) still evidenced a significant transcription promoting activity. That opened the possibility of the co-existence of several subdomains within region 422–588. To test this, smaller fragments covering this region were fused to Gal4 DBD (Gal4-(498–540), Gal4-(456–497), and Gal4-(422–455)). As shown in Fig. 1, none of them were transcriptionally active, indicating that region 422–588 behaved as a single transcription activation domain.

      Identification of the Regions in c-Rel Transactivation Domain Activated by TNFα and PMA + Ionophore

      We used the Gal4 DBD fusion plasmids described in the preceding section to map the region responsible for the transcriptional activation of c-Rel by TNFα and compared it with the activation produced by PMA + ionophore. Table Ilists the TNFα and PMA + ionophore inducibility of the different constructions. The region responsive to activation mapped between positions 422 and 540. Although region 540–588 showed a strong basal transcription activity (Fig. 1), it did not respond to stimulation, suggesting that this C-terminal region was necessary for the basal transcriptional activity of c-Rel but was not involved in TNFα or PMA + ionophore stimulation.
      Table IStimulation of the transactivating activity of Gal4 DBD-c-Rel wild-type and deletion mutant fusions
      c-Rel fragmentTNFα
      Fold activation above corresponding unstimulated construction: ++, >3; +, 2–3; +/−, >2; −, >1.
      PMA + ionophore
      Fold activation above corresponding unstimulated construction: ++, >3; +, 2–3; +/−, >2; −, >1.
      Wild-type+++
      309–318
      309–372
      309–421
      309–455+/−
      309–497++++
      309–540+++
      422–588++
      456–588++++
      498–588++
      541–588+/−
      498–540
      456–497
      422–455
      1-a Fold activation above corresponding unstimulated construction: ++, >3; +, 2–3; +/−, >2; −, >1.

      Analysis of c-Rel Region 422–588 by Ser → Ala Substitutions

      The above results indicate the region 422–588 includes the transcription promoting region of c-Rel, whereas the region 422–540 was responsible for integrating signals derived from activation by TNFα and PMA + ionophore. This region contains 33 Ser residues (20%), 20 acidic residues (12%), and 8 Pro residues (5%), suggesting that it could be an acidic transcription activation domain, despite not having any significant homology to conventional acidic and Pro-rich transactivation domains (). The existence of 20% Ser residues could confer it with the properties of a transcription activation domain regulated by phosphorylation. Those Ser residues are strongly conserved between the human and murine proteins, suggesting the relevance of those residues for function. In order to study their functional relevance, we introduced Ser → Ala mutations into the Gal4-c-Rel-(422–588) (Fig.2 A) fusion. We subsequently assayed for the basal and the TNFα- or PMA/ionophore-induced transcriptional activity of the substitution mutants transfected into Jurkat T cells. However, we did not observe significant differences in the basal activity in any of the Ser → Ala substitutions (Fig.2 B), indicating that none of them is absolutely necessary for the basal transcriptional activity of c-Rel. Interestingly, when we assayed for activation by TNFα or PMA + ionophore, we observed that mutants A3, A4, A5, and A6 failed to respond to TNFα stimulation by increasing its transactivating activity, whereas mutants A8, A9, and A10 showed a reduced response to either PMA + ionophore or TNFα. Mutants A13-A16, included within the constitutively active region 541–588, showed a reduced response, as well. This results identified one region essential for the activation of c-Rel transcriptional capabilities by TNFα, which includes the Ser residues 454, 460, 463, 470, 471, and 473. Substitution of those residues abrogated the activation of c-Rel by TNFα. A second region including Ser residues 491, 494, 508, 509, 510, 511, 513, and 518 was identified to be involved in c-Rel activation, although it was not essential (TableII).
      Figure thumbnail gr2
      Figure 2Activity of Ser → Ala substitutions of c-Rel-(422–588). A, structure of c-Rel transactivation domain with the indicated 16 Ser → Ala substitution mutants (A1–A16). B, effect of substitution mutants on the basal transcriptional activity of c-Rel fusions to Gal4 DBD. Jurkat T cells were transfected with each of the Ser → Ala mutants and co-transfected with a reporter plasmid, which contains five tandem repeats of the Gal4 site upstream the luciferase gene. Results are expressed as percentage of activity compared with the wild-type construct. Transfection efficiency was normalized using the Dual Luciferase assay (Promega). Additionally, similar amounts of the different constructions were expressed in transfected cells as detected by EMSA assays (data not shown). The results represents the mean ± S.D. of three independent experiments.
      Table IIStimulation of the transactivating activity of Gal4 DBD-c-Rel wild-type and Ser → Ala mutants
      c-Rel mutantTNFα
      Fold activation above corresponding unstimulated construction: ++, >3; +, 2–3; +/−, >2; −, >1.
      PMA + ionophore
      Fold activation above corresponding unstimulated construction: ++, >3; +, 2–3; +/−, >2; −, >1.
      Wild-type+++
      A1+++
      A2+++
      A3++
      A4++
      A5++
      A6++
      A7+++
      A8+/−+/−
      A9+/−+/−
      A10+/−+/−
      A11+++
      A12++
      A13+/−+
      A14+/−+
      A15+/−+
      A16+/−+/−
      2-a Fold activation above corresponding unstimulated construction: ++, >3; +, 2–3; +/−, >2; −, >1.

      Identification of the Signaling Route Involved in TNFα Activation of c-Rel Transactivating Activity

      Implication of Phosphatidylinositol 3-Kinase (PI3K)

      We used several commercial inhibitors of putative signaling enzymes to define the route leading to c-Rel activation by TNFα. We first transfected Jurkat T cells with a NF-κB reporter plasmid and tested the effect of different inhibitors in TNFα-stimulated NF-κB activity. As a control, the proteasome inhibitor TPCK was used as a generic inhibitor of NF-κB activity since it interferes with the degradation of IκB. Of the different inhibitors used, only the PI3K inhibitor wortmannin significantly inhibited TNFα stimulation of NF-κB activity (Fig.3 A). To corroborate the effect of wortmannin, another inhibitor of PI3K, LY294002, structurally unrelated to wortmannin, produced similar inhibition of NF-κB activity (Fig. 3 B). Neither LY294002 was wortmannin affected cell viability (data not shown). However, NF-κB activity results from the combined effect of IκB degradation and c-Rel and p65 activation. Thus, in order to study exclusively the implication of PI3K on c-Rel transactivating activity, we transfected Jurkat cells with the Gal4 c-Rel-(309–588) construct and tested the effect of the PI3K inhibitors. Both PI3K inhibitors prevented the transcription activity of c-Rel stimulated by TNFα (Fig. 3 C).
      Figure thumbnail gr3
      Figure 3PI3K is involved in the TNF α-induced activation of the c-Rel transactivation domain. A and B, Jurkat (106) cells were transfected with 500 ng of the reporter NF-κB plasmid NF3TKLuc and cultures were treated with medium alone or with TNFα (10 ng/ml). The effect of the indicated inhibitors is shown. C, Jurkat cells were cotransfected with 0.5 μg of Gal4 c-Rel-(309–588) fusion construct and 10 ng of 5xGal4 luciferase reporter. The effect of the PI3K inhibitors wortmannin and LY294002 was assayed in B and C. Mean luciferase activity ± S.D. of three different experiments is shown.

      PKCζ Involvement in the Activation of the c-Rel Transactivation Domain

      PKCζ has been recently found to be involved in the process of NF-κB activation by TNFα stimulation (
      • Domı́nguez I.
      • Sanz L.
      • Arenzana-Seisdedos F.
      • Dı́az-Meco M.T.
      • Virelizier J.L.
      • Moscat J.
      ), so we studied the effect of this kinase on c-Rel activation. Jurkat T cells were co-transfected with the NF-κB reporter along with a plasmid that expressed either a wild-type or a dominant-negative mutant form of PKCζ. Co-transfection of wild-type PKCζ induced a strong increase in NF-κB-driven reporter activity, compared with cells co-transfected with an empty plasmid (Fig. 4). Furthermore, co-transfection of the PKCζ dominant-negative mutant inhibited the activation of NF-κB transcriptional activity by TNFα (Fig. 4 A). Parallel experiments were carried out assaying the activity of the Gal4 reporter driven by Gal4 c-Rel-(309–588) construct. Co-transfection of PKCζ induced a strong activity of the Gal4 reporter, in a similar way as it did with the NF-κB reporter, and TNFα stimulation induced a still higher activation of the reporter (Fig. 4 B). Interestingly, co-transfection of a PKCζ dominant-negative mutant completely inhibited TNFα activation. Taken together, those results suggested the involvement of PKCζ in the activation of the c-Rel transactivation domain by TNFα and subsequently in NF-κB activation.
      Figure thumbnail gr4
      Figure 4PKC ζ activates the transactivation domain of c-Rel. Jurkat cells (106) were transfected with 500 ng of the reporter NF-κB plasmid NF3TKLuc (A) or with 100 ng of 5xGal4 luciferase reporter along with 10 ng of Gal4 c-Rel-(309–588) fusion construct (B). Cultures were co-transfected with 500 ng of expression plasmids for PKCζ wild-type, PKCζ dominant negative mutant (mut ), PI3K dominant negative mutant (Δp85), or empty vector (pcDNA3) and stimulated with TNFα (10 ng/ml) for 6 h. Luciferase activity is expressed normalized for transfection efficiency with the Dual Luciferase assay (Promega). Expression of wild-type and dominant negative PKCζ was similar as detected by Western blot (data not shown). The results shown are the mean ± S.D. of three independent experiments.

      Cross-talk between PI3K and PKCζ

      The above results indicated that PI3K and PKCζ were involved in the activation of the c-Rel transactivation domain by TNFα. On the other hand, the products of PI3K activity, PIP2 and PIP3, have been described to activate PKCζ (
      • Standaert M.L.
      • Galloway L.
      • Karnam P.
      • Bandyopadhyay G.
      • Moscat J.
      • Farese R.V.
      ). Thus, these enzymes could be participating in the same signaling pathway leading to c-Rel activation by TNFα. To investigate this, Jurkat cells were transfected with a dominant-negative mutant of PI3K (termed Δp85), resulting in the inhibition of both the NF-κB- and Gal4 c-Rel-driven activity induced by co-transfection of active PKCζ (Fig. 4). Additionally, the PI3K inhibitors wortmannin and LY294002 inhibited the activation of NF-κB-dependent reporter activity (Fig.5 A), as well as activation of Gal4 c-Rel-driven activity (Fig. 5 B) induced by transfection of wild-type PKCζ. On the other hand, PIP3, the product of PI3K activity, was able to induce the activity of Gal4 reporter driven by Gal4 c-Rel when added exogenously. Addition of PIP3 to cultures of Jurkat cells co-transfected with wild-type PKCζ, along with the Gal4 reporter and Gal4 c-Rel, resulted in a reporter activity similar to that observed after TNFα stimulation (Fig. 5 C), suggesting that PIP3 is able to activate PKCζ activity. However, PIP3 could not revert the inhibition produced by co-transfection of the dominant-negative mutant of PKCζ (Fig. 5 C). These data indicate that both PI3K and PKCζ are necessary for the activation of c-Rel transactivation domain but do not allow establishment of the relative position of each other in the signaling pathway.
      Figure thumbnail gr5
      Figure 5Cross-talk between PKC ζ and PI3K in the activation of the c-Rel transactivation domain. Jurkat cells (106) were transfected with 500 ng of the reporter NF-κB plasmid NF3TKLuc (A) or with 100 ng of 5xGal4 luciferase reporter along with 10 ng of Gal4 c-Rel-(309–588) fusion construct (B and C). Cultures were co-transfected with 500 ng of expression plasmids for PKCζ wild-type, empty vector (pcDNA3) or PKCζ dominant negative mutant (mut , only inC). The effect of the PI3K inhibitors wortmannin and LY294002 in PKCζ activation of the c-Rel transactivation domain was assayed (A and B). Effect of exogenously added PIP3 (20 ng/ml) was tested (C). Luciferase activity is expressed normalized for transfection efficiency with the Dual Luciferase assay (Promega). Expression of wild-type or mutant PKCζ was similar in transfected cells as detected by Western blot (data not shown). The results shown are the mean ± S.D. of three independent experiments.

      Mapping of the Target Sites of PKCζ and PI3K on the c-Rel Transactivation Domain

      Results shown above indicate that TNFα-dependent activation of c-Rel transactivation domain relies on the Ser residues defined by the Ser → Ala substitution mutants A3, A4, A5, A6, A8, A9, and A10. In order to identify the exact residues that were affected by TNFα-induced PKCζ activity, these mutants were co-transfected into Jurkat cells with the wild-type form of PKCζ. As shown in Fig.6 A, co-transfection of PKCζ along with wild-type Gal4 c-Rel-(422–588) induced an average of 2.5-fold induction of reporter activity. Co-transfection of Ser → Ala substitution mutants A3, A5, A6, A9, and A10 with PKCζ produced a similar activation of the reporter. However, mutants A4 (Ser460) and A8 (Ser491, Ser494) were not activated by PKCζ co-transfection. These results indicate that those sites were necessary for PKCζ activation of c-Rel transactivation domain. Strikingly, both sites are palindromes of the sequence SNCS, not found in other part of the c-Rel transactivation domain.
      Figure thumbnail gr6
      Figure 6Mapping of the target sites of PKC ζ and PI3K in the activation of the c-Rel transactivation domain. Jurkat cells (106) were transfected with 100 ng of 5xGal4 luciferase reporter along with 100 ng of the different TNFα-sensitive Ser → Ala substitution mutants together with 500 ng of expression plasmid for PKCζ wild-type (A) or tested for the effect of exogenously added PIP3 (B). Luciferase activity was normalized for transfection efficiency with the Dual Luciferase assay (Promega). The results represent the increase above the control culture transfected with empty vector (A) or the non-stimulated control (B). A fragment of the sequence of the c-Rel transactivation domain is shown, and the relevant sites are included inboxes. Similar amounts of the different construction were expressed in transfected cells as detected by EMSA assays (data not shown). The results represent the mean ± S.D. of three independent experiments.
      A similar approach was used to map the sites relevant in PI3K activation of c-Rel. Exogenous PIP3 was used to mimic PI3K activation. Wild-type Gal4 c-Rel-(422–588) was successfully activated by PIP3 treatment, as well as the Ser → Ala substitution mutants A4, A5, A9, and A10 (Fig. 6 B). Substitution mutant A8 (Ser491, Ser494), which was not activated by PKCζ, was not activated either by PIP3 treatment, while substitution mutants A3 (Ser454) and A6 (Ser470, Ser471, Ser473) displayed only a reduced activation. These results suggest that the positions defined by substitution mutant A8 may be the point of coincidence of PKCζ and PI3K activation on c-Rel transactivation domain. Noteworthy, both substitution mutants A3 and A6 show a Ser residue close to an Asp residue, pointing out to a possible unique kinase dependent on PI3K activity.

      c-Rel Transactivation Domain Mutants Act as Dominant Negative Forms in NF-κB Activation by TNFα

      In order to test the functional significance of several of those mutations in c-Rel functioning as well as in NF-κB activation in T cells, we transfected expression plasmid of c-Rel mutants into Jurkat cells together with a NF-κB-luc reporter gene. Basal reporter activity was not altered by overexpression of any c-Rel protein (data not shown). As shown in Fig. 7, transfection of wild-type c-Rel slightly, although significantly, increased TNFα-induced stimulation. However, mutants A3, A4, A5, and A8 could not support TNFα-induced activation of NF-κB activity in contrast to wild-type c-Rel. More interestingly, those results indicate that overexpression of these mutant proteins acted as dominant negative forms for NF-κB activation (that includes both endogenous p65 and c-Rel), further pointing out to the importance of these pathways in NF-κB activity.
      Figure thumbnail gr7
      Figure 7Functional activity of c-Rel mutants.Jurkat cells (106) were transfected with 500 ng of the indicated c-Rel cytomegalovirus-expressing plasmids together with 500 ng of the NF-κB reporter plasmid NF-3TK-luc and stimulated with TNFα (10 ng/ml) for 6 h. Luciferase activity was normalized with the Dual Luciferase assay (Promega) and represented as -fold induction over basal activity. The results shown are the mean ± S.D. of three independent experiments. Expression of all c-Rel proteins was similar as detected by Western blot data (data not shown).

      Phosphorylation of c-Rel Region 422–588

      The above results suggested that Ser residues in region 422–588 may be activated by TNFα through phosphorylation. Previous studies have shown that activity of c-Rel and other NF-κB proteins are indeed regulated by phosphorylation (
      • Naumann M.
      • Scheidereit C.
      ). In order to study the stimulation-dependent phosphorylation of this region, we made a recombinant GST fusion protein comprising region 422–588 of c-Rel transactivation domain. Solid phase phosphorylation assays using this recombinant protein revealed that extracts from non-stimulated cells already had strong c-Rel basal phosphorylation activity. Nonetheless, addition of extracts made from Jurkat cells stimulated either with TNFα or PMA + ionophore gave rise to a significant increase (about 2-fold) in the level of phosphorylation state of the recombinant protein (Fig.8 A). When a fusion construct lacking the region 541–588 was used, a great reduction in phosphorylation by unstimulated extracts was observed (Fig.8 B). Interestingly, this construct, comprising region 422–540, was more heavily phosphorylated by extracts from TNFα-stimulated Jurkat cells (about 5-fold increase). Those results suggest that region 541–588 retains most of the basal phosphorylation of c-Rel, but the phosphorylation dependent on TNFα activation resides in the region 422–540, thus corroborating the results obtained for transcriptional stimulation of the deletion mutants. In addition, PI3K inhibitors prevented the increase of the in vitrophosphorylation of c-Rel transactivation domain by cell extracts from TNFα-stimulated cells (Fig. 8 C). Although both TNFα and PMA + ionophore stimulation induced the specific phosphorylation of the c-Rel transactivation domain, we could not detect any kinase activity in TNFα- or PMA/ionophore-treated cell extracts on this domain using “in-gel” kinase phosphorylation assays (data not shown).
      Figure thumbnail gr8
      Figure 8TNF α-induced phosphorylation of c-Rel transactivation domain. A, recombinant GST c-Rel-(422–588) protein was used as substrate forin vitro solid-phase phosphorylation assays using WCE from cells treated with medium alone, TNFα (10 ng/ml), or PMA (10 ng/ml) + ionophore (1 μm) for 30 min, as indicated. B, recombinant GST c-Rel-(422–588) protein or GST c-Rel-(422–540) were used as substrates for in vitro solid-phase phosphorylation assays (left panel) using WCE from the same number of cells treated with medium alone or TNFα (10 ng/ml). The same amount of recombinant protein was run in a 10% SDS-PAGE and stained with Coomassie Blue to reveal the recombinant proteins (right panel). Intensity of the bands was quantified by densitometric scanning of the films and expressed as optical density (O.D.) inA or -fold increase above extracts from non-stimulated control in B. This set of results represents a representative experiment of the three performed. C, effect of PI3K inhibitors wortmannin and LY294002 on TNFα-induced phosphorylation of c-Rel. Recombinant GST-c-Rel-(422–588) was used as substrate for in vitro solid phase phosphorylation assays using WCE from non-stimulated cells or cells stimulated with TNFα or PMA + ionophore for 30 min, as indicated. The intensity of the bands was determined by densitometry of the film and the optical density (O.D.) units plotted. Results in panels A–C are the mean ± S.D. of three independent experiments. In C a representative experiment is shown.

      DISCUSSION

      The intermediate and late phase of T cell activation is controlled by the autocrine or paracrine effect of the cytokine TNFα, which in turn modulates the transcription factor NF-κB through the sustained activation of c-Rel (
      • Pimentel-Muiños F.X.
      • Mazana J.
      • Fresno M.
      ). On the other hand, it is well established that IKK activation by TNFα stimulation, and subsequent IκB degradation, is a pre-requisite for NF-κB activation (
      • Mercurio F.
      • Zhu H.
      • Murray B.W.
      • Shevchenko A.
      • Bennett B.L.
      • Li J.
      • Young D.B.
      • Barbosa M.
      • Mann M.
      • Manning A.
      • Rao A.
      ,
      • Mercurio F.
      • Manning A.M.
      ,
      • Zandi E.
      • Rothwarf D.M.
      • Delhase M.
      • Hayakawa M.
      • Karin M.
      ). However, a second level of regulation of NF-κB activity that is independent of IKK degradation has been recently described. This second level involves the signal-dependent phosphorylation and activation of the transactivation domain of p65 (
      • Schmitz M.L.
      • dos Santos Silva M.A.
      • Baeuerle P.A.
      ,
      • Wang D.
      • Baldwin Jr., A.S.
      ,
      • Sizemore N.
      • Leung S.
      • Stark G.R.
      ,
      • Zhong H.
      • SuYang H.
      • Erdjument-Bromage H.
      • Tempst P.
      • Ghosh S.
      ), although the exact mechanism has not yet been defined.
      The transactivation domain of c-Rel has been previously defined as the region downstream the Rel homology domain to the C terminus of the protein (positions 309–588), as in this region reside all the transcription promoting capabilities of c-Rel (
      • Bull P.
      • Morley K.L.
      • Hoekstra M.F.
      • Hunter T.
      • Verma I.M.
      ,
      • Ishikawa H.
      • Asano M.
      • Kanda T.
      • Kumar S.
      • Gélinas C.
      • Ito Y.
      ). The deletion analysis described in this work has indicated that positions 309–421 in the C terminus are dispensable for c-Rel-mediated transcription, further delimiting the location of this transactivation domain between positions 422 and 588 of the C terminus. However, a role of the region 309–421 in the regulation of c-Rel stability through degradation by the proteasome has been proposed (
      • Chen E.
      • Hrdlickova R.
      • Nehyba J.
      • Longo D.L.
      • Bose Jr., H.R.
      • Li C.C.
      ). Furthermore, as the region 422–588 fused to the Gal4 DBD had a stronger transcriptional activity than the complete C-terminal region (309), this may indicate the existence of repressor sequences within the 309–421 region. Alternatively, region 422–588 could acquire a configuration in the Gal4 fusion that is more competent for transcriptional activation. Sequential deletions from position 422 to 588 (C terminus) were associated with parallel decrease in transcriptional activity. The region from position 541 to 588 still retained significant transcriptional activity, whereas smaller fragments (422–455, 456–497, and 498–540) showed no transcriptional activity, indicating that region 422–588 is a functional transactivation domain non divisible in smaller units. This is not the case of the Rel family member p65, where two different transcriptional activating regions were found in the transactivation domain, named TA1 and TA2 (
      • Schmitz M.L.
      • dos Santos Silva M.A.
      • Baeuerle P.A.
      ,
      • Schmitz M.L.
      • dos Santos Silva M.A.
      • Altmann H.
      • Czisch M.
      • Holak T.A.
      • Baeuerle P.A.
      ).
      Stimulation of the cells transfected with the Gal4 c-Rel fusion constructs revealed that only region 422–540 was activated by TNFα, as well as PMA + ionophore, stimulation. Region 541–588 was not further activated by these stimuli, even though it showed a strong basal transcriptional activity. Thus, those analyses indicate that this region is necessary for the function of c-Rel as a transcription factor but dispensable for inducible activation of this factor. In vitro solid phase phosphorylation assays demonstrated that this region (541) was strongly phosphorylated by cell extracts obtained from non-stimulated cells, thus supporting its role in basal transcriptional activity of c-Rel. In contrast, region 422–540 was weakly phosphorylated by cell extracts from unstimulated cells, whereas extracts from TNFα- or PMA/ionophore-stimulated Jurkat cells strongly phosphorylated it. This indicate that it is within this region where stimulus-induced activation occurs. These data suggested that the activation of this region requires extracellular signal-dependent phosphorylation.
      This region (422) is negatively charged (15 acidic residues, 3 basic) and is rich in Pro and Ser. These characteristics resemble to acidic transactivation domains (). Although this region is not conserved among different Rel family members, the acidic and Ser residues are well conserved between human and mouse c-Rel, suggesting a critical role for c-Rel functioning. This region does not show homology to other described acidic transactivation factors. However, its high Ser content is typical of other transcription factors like CREB, TCF/Elk-1, or STAT that are regulated by phosphorylation (
      • Karin M.
      ). Thus, phosphorylation may provide the additional negative charges necessary to constitute an acidic transactivation domain. The effect observed in Ser → Ala mutants of the transactivation domain of c-Rel corroborates this hypothesis. Although none of the mutants had any effect on the basal transcriptional activity of c-Rel, the substitution of the Ser residues at position 455 (mutant A3), 461 (mutant A4), 464 (mutant A5), or 470, 471, and 473 (mutant A6) completely prevented the activation of this domain by TNFα. Any of these positions is absolutely required and must be activated, as the substitution of any of them was enough to abrogate the activation of this domain by TNFα. Additionally, the substitution of the Ser residues at position 492 and 494 (mutant A8), 509–512 and 514 (mutant A9), or 519 (mutant A10) had a less pronounced inhibitory effect on both TNFα- and PMA/ionophore-induced activation of c-Rel transactivation domain, indicating that they may participate in c-Rel activation although they are probably not essential.
      Transcription factors like NF-κB activate gene transcription through the interaction with basal transcriptional machinery, or with co-factors that modulate that machinery (
      • Greenblatt J.
      ). In this regard, both c-Rel and v-Rel have been found to interact with the basal transcription factors TBP and TFIID directly through their transactivation domains (
      • Xu X.
      • Prorock C.
      • Ishikawa H.
      • Maldonado E.
      • Ito Y.
      • Gelinas C.
      ). However, other reports indicate that only the first 50 amino acids were necessary for TBP interaction (
      • Kerr L.D.
      • Ransone L.J.
      • Wamsley P.
      • Schmitt M.J.
      • Boyer T.G.
      • Zhou Q.
      • Berk A.J.
      • Verma I.M.
      ). p65 could interact with TBP and TFIID through its transactivation domain (
      • Schmitz M.L.
      • dos Santos Silva M.A.
      • Baeuerle P.A.
      ,
      • Schmitz M.L.
      • Stelzer G.
      • Altmann H.
      • Meisterernst M.
      • Baeuerle P.A.
      ). Furthermore, other proteins that are able to interact with NF-κB have been described: the HMG(I)Y nuclear protein (
      • Thanos D.
      • Maniatis T.
      ), the SP1 factor (
      • Hirano F.
      • Tanaka H.
      • Hirano Y.
      • Hiramoto M.
      • Handa H.
      • Makino I.
      • Scheidereit C.
      ), or a 40-kDa protein, which acts as a target for the quinone derivative E3330 in the inhibition of NF-κB activity (
      • Hiramoto M.
      • Shimizu N.
      • Sugimoto K.
      • Tang J.
      • Kawakami Y.
      • Ito M.
      • Aizawa S.
      • Tanaka H.
      • Makino I.
      • Handa H.
      ). The mutations described in this work may be very useful for the study of the interactions of c-Rel with the transcriptional machinery and to define protein-protein interactions that regulates the process of transcription activation.
      We were unable to identify a kinase activity, capable of phosphorylating c-Rel transactivation domain, by in-gel phosphorylation assays. However, only a fraction of cellular kinases can remain active after the renaturalization process required in those assays. Furthermore, no kinase that requires a co-factor or association with other proteins to form an active complex will remain active in those assays. On the other hand, the use of commercially available inhibitors revealed the dependence of NF-κB activation by TNFα on the activity of PI3K. PI3K activity was previously described as necessary for NF-κB activity in several cell types but not in lymphocytes (
      • Kaliman P.
      • Canicio J.
      • Testar X.
      • Palacı́n M.
      • Zorzano A.
      ). Furthermore, PI3K-dependent phosphorylation and activation of p65 has been described as a requisite for interleukin-1 activation of p65 transactivation domain, without affecting IκB degradation or DNA binding activity of p65 complexes (
      • Sizemore N.
      • Leung S.
      • Stark G.R.
      ). In a similar manner, TNFα-dependent activation of NF-κB-induced transcription in HepG2 cells has been shown to depend on PI3K activity, which does not affect NF-κB binding to DNA or IκB degradation induced by the cytokine (
      • Reddy S.A.
      • Huang J.H.
      • Liao W.S.
      ). Our results indicate that a similar mechanism is taking place for c-Rel activation by TNFα. Thus, PI3K inhibitors wortmannin and LY294002 inhibited not only NF-κB dependent activity, but also c-Rel activation and phosphorylation of its transactivation domain, supporting a critical role of PI3K activity for TNFα activation of c-Rel. However, whether PI3K can associate with TNF receptors or any of its associated factors is not known yet. Furthermore, the PI3K metabolite, PIP3, was capable of activating Gal4 c-Rel when transfected into Jurkat cells but not NF-κB reporter activity (data not shown). Thus, TNFα might activate several signaling routes, leading to the direct modulation of transcriptional abilities of c-Rel, as well as to activation of IKKs for IκB degradation and subsequent nuclear translocation of active NF-κB heterodimers. PI3K may therefore be implicated in the activation of c-Rel transactivation domain but not in the activation of IKK by TNFα. Nevertheless, one of the well known targets of PI3K activity, the protein kinase Akt, has been recently demonstrated to interact with IKK upon TNFα stimulation (
      • Romashkova J.A.
      • Makarov S.S.
      ,
      • Ozes O.N.
      • Mayo L.D.
      • Gustin J.A.
      • Pfeffer S.R.
      • Pfeffer L.M.
      • Donner D.B.
      ). However, our results with PIP3 stimulation, as well as other recent reports looking directly at NF-κB binding activity (
      • Sizemore N.
      • Leung S.
      • Stark G.R.
      ,
      • Reddy S.A.
      • Huang J.H.
      • Liao W.S.
      ), suggest that PI3K may not be essential for IKK activation and subsequent nuclear translocation of NF-κB. Although those discrepancies cannot be explained yet, PI3K might differentially affect NF-κB activation depending on the cell type.
      On the other hand, PKCζ has been implicated in NF-κB activation, through mechanisms that involve IKK activation (
      • Lallena M.J.
      • Dı́az-Meco M.T.
      • Bren G.
      • Pay C.V.
      • Moscat J.
      ,
      • Folgueira L.
      • McElhinny J.A.
      • Bren G.D.
      • MacMorran W.S.
      • Dı́az-Meco M.T.
      • Moscat J.
      • Paya C.V.
      ,
      • Muller G.
      • Ayoub M.
      • Storz P.
      • Rennecke J.
      • Fabbro D.
      • Pfizenmaier K.
      ) or directly through the phosphorylation and activation of p65 (
      • Anrather J.
      • Csizmadia V.
      • Soares M.P.
      • Winkler H.
      ). Thus, PKCζ may play a dual role in the activation of NF-κB, activating IKK and also participating in the activation the transactivation domain of members of the Rel family. Our results clearly support this hypothesis, indicating that PKCζ, besides participating in IKK activation, is also involved in the activation of c-Rel transactivation domain. Thus, co-transfection experiments into Jurkat cells of PKCζ wild-type with Gal4 c-Rel showed a strong potentiation of c-Rel transcriptional promoting capabilities. Furthermore, a dominant-negative mutant of PKCζ abrogated TNFα-induced c-Rel activation. Surprisingly, PKCζ activation of c-Rel transactivation domain was inhibited in the presence of wortmannin and LY294002, both inhibitors of PI3K activity, as well as the co-transfection of a dominant-negative mutant of PI3K, suggesting that PI3K might act downstream of PKCζ . In addition, the activation of c-Rel transactivation domain by PIP3, a product of PI3K activity, was inhibited by a dominant-negative mutant of PKCζ. However, PKCζ has been shown to be activated by the protein kinase PDK-1, an effector of PI3K activity (
      • Chou M.M.
      • Hou W.
      • Johnson J.
      • Graham L.K.
      • Lee M.H.
      • Chen C.S.
      • Newton A.C.
      • Schaffhausen B.S.
      • Toker A.
      ), suggesting that PKCζ would be downstream of PI3K in the route of c-Rel activation. A possible explanation for this apparent different positioning in the signaling pathway of PI3K and PKCζ is that both pathways are parallel and required for activity.
      The fact that PIP3 and active PKCζ seem to act on different Ser residues of the c-Rel transactivation domain would be in agreement with the above hypothesis. The Ser → Ala substitution analysis showed that positions defined by the mutants A3, A4, A5, and A6 are essential and in lesser extent by the mutants A8, A9, and A10 for c-Rel activation by TNFα. Thus, the failure to activate one of them resulted in the inability of c-Rel to be activated by TNFα. Mapping of the sites activated by PKCζ and PIP3 revealed that the only Ser residues substituted in mutant A8 are the point of convergence for PKCζ and PI3K activation. However, PKCζ also required the Ser residue defined by mutant A4, whereas PIP3required Ser residues defined by mutants A3 and A6. Thus, the inhibition of either PKCζ or PI3K would result in an inhibition of c-Rel transactivation domain. The sites defined by mutants A4 and A8 showed a striking palindromic similarity (SNCS for A4 and SCNS for A8). Mutant A5, however, which substituted the second Ser residue downstream mutant A4, was not essential for PKCζ activation. Thus the Ser residue close to Asn (Ser460 and Ser494) may be the actual target of PKCζ activation. Furthermore, A8, which defined the point of convergence between PI3K and PKCζ, mutants A3 (Ser454) and A6 (Ser471, Ser473, and Ser474), failed to be activated by PIP3. These mutants involve Ser residues that are in close proximity of an Asp residue, suggesting that the same kinase might be activating both sites. Furthermore, a natural mutant of Ser471 to Asn produced a form of c-Rel that could not be activated by TNFα stimulation (
      • Martı́n A.G.
      • Fresno M.
      ), suggesting that PI3K may be an essential part of the signaling mechanisms activated by TNFα, resulting in the activation of c-Rel transactivation domain. Furthermore, the different responses of several Ser mutants to the different stimuli clearly discard nonspecific effects of the pathways or activators used.
      More interestingly, those mutants not only are defective in transactivating activity, but they prevent NF-κB-dependent reporter activity. Those results indicate that they act as dominant negative forms of NF-κB activation, either by recruiting the kinases required for phosphorylation of transactivation domains of c-Rel and/or p65 or by binding to NF-κB sites on DNA and then prevent active NF-κB complexes (either p65 or endogenous c-Rel) for binding. The fact that the spontaneous Ser471 mutant failed to activate at all NF-κB (
      • Martı́n A.G.
      • Fresno M.
      ) tends to support the first hypothesis.
      In summary, our results have revealed an important level of TNFα-induced NF-κB activation mediated through PI3K and PKCζ, which are absolutely required for c-Rel transactivating activity (Fig.9).
      Figure thumbnail gr9
      Figure 9Model of activation of c-Rel transactivation domain by TNF α.

      Acknowledgments

      We thank Marı́a Chorro and Lucı́a Horrillo for excellent technical assistance.

      REFERENCES

        • Baldwin Jr., A.S.
        Annu. Rev. Immunol. 1996; 14: 649-683
        • Ghosh S.
        • May M.J.
        • Kopp E.B.
        Annu. Rev. Immunol. 1998; 16: 225-260
        • Pimentel-Muiños F.X.
        • Mazana J.
        • Fresno M.
        Eur. J. Immunol. 1995; 25: 179-186
        • Wallach D.
        • Varfolomeev E.E.
        • Malinin N.L.
        • Goltsev Y.V.
        • Kovalenko A.V.
        • Boldin M.P.
        Annu. Rev. Immunol. 1999; 17: 331-367
        • Gerondakis S.
        • Strasser A.
        • Metcalf D.
        • Grigoriadis G.
        • Scheerlinck J.Y.
        • Grumont R.J.
        Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 3405-3409
        • Liou H.C.
        • Jin Z.
        • Tumang J.
        • Andjelic S.
        • Smith K.A.
        • Liuo M.L.
        Int. Immunol. 1999; 11: 361-371
        • May M.J.
        • Ghosh S.
        Immunol. Today. 1998; 19: 80-88
        • Schmitz M.L.
        • dos Santos Silva M.A.
        • Baeuerle P.A.
        J. Biol. Chem. 1995; 270: 15576-15584
        • Schmitz M.L.
        • Baeuerle P.A.
        Immunobiology. 1995; 193: 116-127
        • Wang D.
        • Baldwin Jr., A.S.
        J. Biol. Chem. 1998; 273: 29411-29416
        • Zhong H.
        • Voll R.E.
        • Ghosh S.
        Mol. Cell. 1998; 1: 661-671
        • Norris J.L.
        • Baldwin Jr., A.S.
        J. Biol. Chem. 1999; 274: 13841-16846
        • Anrather J.
        • Csizmadia V.
        • Soares M.P.
        • Winkler H.
        J. Biol. Chem. 1999; 274: 13594-13603
        • Lallena M.J.
        • Dı́az-Meco M.T.
        • Bren G.
        • Pay C.V.
        • Moscat J.
        Mol. Cell. Biol. 1999; 19: 2180-2188
        • Sanz L.
        • Sánchez P.
        • Lallena M.J.
        • Dı́az-Meco M.T.
        • Moscat J.
        EMBO J. 1999; 18: 3044-3053
        • Sizemore N.
        • Leung S.
        • Stark G.R.
        Mol. Cell. Biol. 1999; 19: 4798-4805
        • Reddy S.A.
        • Huang J.H.
        • Liao W.S.
        J. Immunol. 2000; 164: 1355-1363
        • Toker A.
        • Cantley L.C.
        Nature. 1997; 387: 673-676
        • Schmitz M.L.
        • dos Santos Silva M.A.
        • Altmann H.
        • Czisch M.
        • Holak T.A.
        • Baeuerle P.A.
        J. Biol. Chem. 1994; 269: 25613-25620
        • Chen C.
        • Agnës F.
        • Gèlinas C.
        Mol. Cell. Biol. 1999; 19: 307-316
        • Martı́n A.G.
        • Fresno M.
        J. Biol. Chem. 2000; 275: 24383-24391
        • Yano O.
        • Kanellopoulos J.
        • Kieran M.
        • Le Bail O.
        • Israel A.
        • Kourilsky P.
        EMBO J. 1987; 6: 3317-3324
        • Bull P.
        • Morley K.L.
        • Hoekstra M.F.
        • Hunter T.
        • Verma I.M.
        Mol. Cell. Biol. 1990; 10: 5473-5485
        • Hahn S.
        Cell. 1993; 72: 481-483
        • Domı́nguez I.
        • Sanz L.
        • Arenzana-Seisdedos F.
        • Dı́az-Meco M.T.
        • Virelizier J.L.
        • Moscat J.
        Mol. Cell. Biol. 1993; 13: 1290-1295
        • Standaert M.L.
        • Galloway L.
        • Karnam P.
        • Bandyopadhyay G.
        • Moscat J.
        • Farese R.V.
        J. Biol. Chem. 1997; 272: 30075-30082
        • Naumann M.
        • Scheidereit C.
        EMBO J. 1994; 13: 4597-4607
        • Mercurio F.
        • Zhu H.
        • Murray B.W.
        • Shevchenko A.
        • Bennett B.L.
        • Li J.
        • Young D.B.
        • Barbosa M.
        • Mann M.
        • Manning A.
        • Rao A.
        Science. 1997; 278: 860-866
        • Mercurio F.
        • Manning A.M.
        Curr. Opin. Cell Biol. 1999; 11: 226-232
        • Zandi E.
        • Rothwarf D.M.
        • Delhase M.
        • Hayakawa M.
        • Karin M.
        Cell. 1997; 91: 243-252
        • Zhong H.
        • SuYang H.
        • Erdjument-Bromage H.
        • Tempst P.
        • Ghosh S.
        Cell. 1997; 89: 413-424
        • Ishikawa H.
        • Asano M.
        • Kanda T.
        • Kumar S.
        • Gélinas C.
        • Ito Y.
        Oncogene. 1993; 8: 2889-2896
        • Chen E.
        • Hrdlickova R.
        • Nehyba J.
        • Longo D.L.
        • Bose Jr., H.R.
        • Li C.C.
        J. Biol. Chem. 1998; 273: 35201-35207
        • Karin M.
        Curr. Opin. Cell Biol. 1994; 6: 415-424
        • Greenblatt J.
        Nature. 1992; 360: 16-17
        • Xu X.
        • Prorock C.
        • Ishikawa H.
        • Maldonado E.
        • Ito Y.
        • Gelinas C.
        Mol. Cell. Biol. 1993; 13: 6733-6741
        • Kerr L.D.
        • Ransone L.J.
        • Wamsley P.
        • Schmitt M.J.
        • Boyer T.G.
        • Zhou Q.
        • Berk A.J.
        • Verma I.M.
        Nature. 1993; 365: 412-419
        • Schmitz M.L.
        • Stelzer G.
        • Altmann H.
        • Meisterernst M.
        • Baeuerle P.A.
        J. Biol. Chem. 1995; 270: 7219-7226
        • Thanos D.
        • Maniatis T.
        Cell. 1992; 71: 777-789
        • Hirano F.
        • Tanaka H.
        • Hirano Y.
        • Hiramoto M.
        • Handa H.
        • Makino I.
        • Scheidereit C.
        Mol. Cell. Biol. 1998; 18: 1266-1274
        • Hiramoto M.
        • Shimizu N.
        • Sugimoto K.
        • Tang J.
        • Kawakami Y.
        • Ito M.
        • Aizawa S.
        • Tanaka H.
        • Makino I.
        • Handa H.
        J. Immunol. 1998; 160: 810-819
        • Kaliman P.
        • Canicio J.
        • Testar X.
        • Palacı́n M.
        • Zorzano A.
        J. Biol. Chem. 1999; 274: 17437-17444
        • Romashkova J.A.
        • Makarov S.S.
        Nature. 1999; 401: 86-90
        • Ozes O.N.
        • Mayo L.D.
        • Gustin J.A.
        • Pfeffer S.R.
        • Pfeffer L.M.
        • Donner D.B.
        Nature. 1999; 401: 82-85
        • Folgueira L.
        • McElhinny J.A.
        • Bren G.D.
        • MacMorran W.S.
        • Dı́az-Meco M.T.
        • Moscat J.
        • Paya C.V.
        J. Virol. 1996; 70: 223-231
        • Muller G.
        • Ayoub M.
        • Storz P.
        • Rennecke J.
        • Fabbro D.
        • Pfizenmaier K.
        EMBO J. 1995; 14: 1961-1969
        • Chou M.M.
        • Hou W.
        • Johnson J.
        • Graham L.K.
        • Lee M.H.
        • Chen C.S.
        • Newton A.C.
        • Schaffhausen B.S.
        • Toker A.
        Curr. Biol. 1998; 8: 1069-1077