Report

Regulation of NF-κB Activation in T Cells via Association of the Adapter Proteins ADAP and CARMA1

See allHide authors and affiliations

Science  04 May 2007:
Vol. 316, Issue 5825, pp. 754-758
DOI: 10.1126/science.1137895

Abstract

The adapter protein ADAP regulates T lymphocyte adhesion and activation. We present evidence for a previously unrecognized function for ADAP in regulating T cell receptor (TCR)–mediated activation of the transcription factor NF-κB. Stimulation of ADAP-deficient mouse T cells with antibodies to CD3 and CD28 resulted in impaired nuclear translocation of NF-κB, a reduced DNA binding, and delayed degradation and decreased phosphorylation of IκB (inhibitor of NF-κB). TCR-stimulated assembly of the CARMA1–BCL-10–MALT1 complex was substantially impaired in the absence of ADAP. We further identified a region of ADAP that is required for association with the CARMA1 adapter and NF-κB activation but is not required for ADAP-dependent regulation of adhesion. These findings provide new insights into ADAP function and the mechanism by which CARMA1 regulates NF-κB activation in T cells.

Adapter proteins nucleate multimolecular complexes that are essential for effective transmission of intracellular signals during an adaptive immune response (1). In T lymphocytes, the adhesion- and degranulation-promoting adapter protein (ADAP) regulates T cell receptor (TCR)–dependent changes in the function of integrin adhesion receptors (2, 3). ADAP-deficient (ADAP–/–) T cells also exhibit impaired proliferation and cytokine production after stimulation of the TCR and the CD28 costimulatory receptor (2, 3). Stimulation of these receptors leads to activation of the NF-κB family of transcription factors, which are critical for T cell activation and survival (4). A multiprotein complex consisting of the membrane-associated adapter protein CARMA1 (5, 6), the caspase-like protein MALT1 (7, 8), and the adapter protein BCL-10 (9) is critical for TCR-dependent activation of the IκB kinase complex and subsequent NF-κBnuclear translocation (10).

Like ADAP-deficient T cells, protein kinase Cθ (PKCθ)–deficient T cells exhibit defective TCR-mediated proliferation, even though proximal TCR signaling events, such as extracellular signal–regulated kinase (ERK) activation, are normal (11). Therefore, we examined PKCθ-dependent signaling in ADAP–/– T cells (12). Membrane localization of PKCθ was similar in ADAP+/– and ADAP–/– T cells upon stimulation with antibodies to CD3 and CD28 (Fig. 1, A and B). Stimulated ADAP+/– and ADAP–/– T cells also showed similar levels of PKCθ phosphorylation (Fig. 1C). Thus, ADAP is not required for TCR signaling events leading to and including PKCθ activation. Because PKCθ regulates NF-κB activation downstream of the TCR (11, 13), we next examined NF-κB signaling in ADAP–/– T cells. Image scanning flow cytometry (14, 15) (fig. S1) revealed a striking defect in p65 nuclear translocation after stimulation of ADAP–/– lymph node T cells (Fig. 2, A and B) or CD4 T cells (fig. S2) by CD3 and CD28 (CD3/CD28). In contrast, no impairment in NF-κB activation was detected after stimulation with tumor necrosis factor–α (TNF-α), which activates NF-κB independently of the TCR. These results were confirmed with electrophoretic mobility shift assays (Fig. 2C). ADAP–/– Tcells also displayed defective NF-κB translocation after treatment with phorbol 12-myristate 13-acetate (PMA), which activates PKC (Fig. 2, A and B). CD3/CD28 stimulation of ADAP–/– T cells also resulted in reduced induction of intercellular adhesion molecule–1 (ICAM-1), which is encoded by a NF-κB–regulated gene (16) (Fig. 2D).

Fig. 1.

TCR-dependent membrane localization and activation of PKCθ in ADAP –/– T cells. (A) Localization of PKCθ (bottom) in ADAP+/– and ADAP–/– T cells to the contact site with beads coated with antibodies to CD3 and CD28. Differential interference contrast (DIC) images are shown in top panels. (B) Quantification of PKCθ localization. T cell–bead conjugates (minimum 90 per group) were scored for PKCθ polarization from two independent experiments. Graph shows the average percent of T cell–bead conjugates with polarized PKCθ (±SD). (C) Phosphorylation of PKCθ after CD3/CD28 stimulation of ADAP+/– and ADAP–/– T cells for the indicated time points was assessed by Western blotting of whole-cell lysates with antibody to phosphorylated PKCθ (Thr538) (top panels). Blots were also probed with antibody to β-actin (bottom panels).

Fig. 2.

Activation of the NF-κB pathway is impaired in ADAP–/– T cells. (A) ADAP+/– and ADAP–/– T cells were unstimulated (unstim.) or stimulated for 10 min with antibodies to CD3 and CD28, PMA, or TNF-α, stained with 7-AAD and fluorescein isothiocyanate–conjugated antibody to p65, and analyzed on a multispectral imaging flow cytometer (15). Histograms show the NF-κB–7-AAD similarity index, which reflects colocalization of the 7-AAD and p65 signals in the population of T cells analyzed. Marker values indicate the percentage of T cells in each sample with translocated p65. (B) Nuclear localization of p65 in unstimulated T cells was set to 1; results show the average increase (±SD) in p65 nuclear translocation in stimulated T cells relative to unstimulated T cells for four independent experiments. (C) Electrophoretic mobility shift assay (EMSA) of NF-κB in ADAP–/– T cells. ADAP+/– or ADAP–/– lymph node T cells were stimulated with antibodies to CD3 and CD28 for the indicated time periods (in minutes), or for 10 min with PMA and ionomycin (P/I) or with TNF-α. Nuclear extracts were prepared and EMSA was performed with a biotin-labeled NF-κBprobe. (D) ADAP+/– and ADAP–/– lymph node T cells were stimulated in vitro with antibodies to CD3 and CD28 for 24 hours. Cells were harvested and ICAM-1 expression on CD4 T cells was determined by flow cytometry. (E) IκBα degradation and phosphorylation. ADAP+/– and ADAP–/– T cells were either unstimulated or stimulated as in (C) before lysis. Lysates were analyzed by Western blotting with antibodies specific for IκBα, phosphorylated IκBα, phosphorylated ERK, ERK2, or β-actin.

We also examined signaling events proximal to nuclear translocation of the p65 NF-κB subunit in ADAP–/– T cells. In resting T cells, NF-κB subunits are sequestered in the cytoplasm via interactions with IκBα (4). Relative to ADAP+/– T cells, CD3/CD28 stimulation of ADAP–/– T cells led to a delay in IκBα degradation and decreased IκB phosphorylation (Fig. 2E and fig. S3). Consistent with previous results (2, 3), the kinetics of ERK phosphorylation after CD3/CD28 stimulation were not affected by loss of ADAP (Fig. 2E). Activation of the IκB kinase (IKK) complex, which phosphorylates IκBα, was also impaired after CD3/CD28 stimulation of ADAP–/– T cells (fig. S4). Thus, ADAP acts as a positive regulator of TCR-dependent NF-κB activation, downstream of PKCθ yet upstream of IKK activation, IκBα degradation, and NF-κB subunit translocation.

We next examined the role of ADAP in inducible membrane localization of BCL-10, MALT1, and CARMA1 (17, 18). In unstimulated ADAP+/– or ADAP–/– T cells, only low levels of BCL-10 and MALT1 were detected in membrane fractions (Fig. 3A). CD3/CD28 or PMA stimulation of ADAP+/– T cells, but not ADAP–/– T cells, resulted in enhanced membrane localization of both BCL-10 and MALT1 (Fig. 3A). ADAP also localized to the membrane in stimulated ADAP+/– T cells (Fig. 3A). The levels of CARMA1 in membrane fractions were comparable between ADAP+/– and ADAP–/– T cells both before and after stimulation (Fig. 3A), consistent with previous results (17). Thus, ADAP is critical for the activation-dependent membrane localization of BCL-10 and MALT1.

Fig. 3.

ADAP is critical for the assembly of the CARMA1–BCL-10–MALT1 complex and associates directly with the C-terminal end of CARMA1. (A) ADAP+/– and ADAP–/– T cells were either unstimulated (U) or stimulated with antibodies to CD3 and CD28 (3/28) or with PMA (P). Cytosolic (Cyto) and membrane (Memb) fractions were prepared from each cell sample and analyzed by Western blotting with the indicated antibodies. (B) Membrane fractions were prepared from unstimulated and stimulated ADAP+/– and ADAP–/– T cells as in (A). ADAP or CARMA1 was immunoprecipitated and the immunoprecipitates analyzed as in (A). (C) T cells isolated from ADAP+/– and ADAP–/– transgenic mice expressing the hCAR adenovirus receptor were transduced with control adenovirus encoding Thy1.1 (ctrl) or adenovirus expressing wild-type mouse ADAP and Thy1.1 (ADAPwt). T cells were then left unstimulated or were stimulated with antibodies to CD3 and CD28 or with PMA before lysis and immunoprecipitation with a monoclonal antibody (mAb) to BCL-10. Immunoprecipitates were analyzed by Western blotting as in (A). (D) Interaction of ADAP with CARMA1 in vitro. GST only or GST-ADAP fusion proteins were incubated with in vitro transcribed/translated BCL-10, MALT1, CARMA1, or truncated FLAG epitope–tagged CARMA1/651-1147. GST pull-downs were analyzed by Western blotting for the presence of the indicated proteins. An equivalent amount of the purified proteins used in the pull-down assays, along with a sample from in vitro transcription and translation reactions using a control vector (vec), were analyzed in separate gels (input). (E) Diagram of CARMA1 and CARMA1/651-1147 constructs used in (D).

To define potential interactions between ADAP and the CARMA1–BCL-10–MALT1 signaling complex, we performed coimmunoprecipitation experiments. ADAP could be immunoprecipitated only from membrane fractions isolated from activated T cells (Fig. 3B). Coimmunoprecipitation of CARMA1, BCL-10, and MALT1 with ADAP was observed from stimulated membrane fractions of either lymph node T cells (Fig. 3B) or purified CD4 T cells (fig. S5). Similarly, MALT1, BCL-10, and ADAP coimmunoprecipitated with CARMA1 only after T cell stimulation (Fig. 3B). However, in activated ADAP–/– T cells, MALT-1 and BCL-10 did not coimmunoprecipitate with CARMA1 (Fig. 3B), and CARMA1 and MALT-1 did not coimmunoprecipitate with BCL-10 (Fig. 3C and fig. S6). To confirm that ADAP is required for inducible complex assembly, we used resting T cells expressing the hCAR adenovirus receptor to permit adenoviral-mediated expression of ADAP (19). ADAP reexpression in ADAP–/– T cells restored coimmunoprecipitation of CARMA1 and MALT1 with BCL-10 after CD3/CD28 or PMA stimulation (Fig. 3C). Thus, ADAP is a component of the CARMA1–BCL-10–MALT1 complex and is required for normal complex formation.

We next examined interactions between ADAP and purified BCL-10, MALT1, and CARMA1. A glutathione S-transferase (GST)–ADAP fusion protein interacted in vitro with purified CARMA1, but not with purified BCL-10 or MALT1 (Fig. 3D). A truncated form of CARMA1 (CARMA1/651-1147) containing just the C-terminal PDZ, SH3, and GUK-like domains typical of membrane-associated guanylate kinase (MAGUK)–family proteins (20) (Fig. 3E) also interacted with GST-ADAP in vitro (Fig. 3D). Thus, the interaction of ADAP with CARMA1 is not dependent on the caspase-recruiting domain that mediates the interaction of CARMA1 with BCL-10. Colocalization of ADAP with CARMA1 was also observed at the contact site between wild-type T cells and beads coated with antibodies to CD3 and CD28 (fig. S7).

Truncation and deletion mutants of ADAP were used to define sites within ADAP critical for its interaction with CARMA1 (Fig. 4A). Wild-type and mutant forms of hemagglutinin (HA) epitope–tagged ADAP were immunoprecipitated from transiently transfected Jurkat T cells after PMA stimulation, and these immunoprecipitates were analyzed for the presence of coimmunoprecipitating BCL-10, MALT1, and CARMA1 proteins (Fig. 4B). Wild-type ADAP and an ADAP mutant lacking the N-terminal 327 amino acids (ADAPΔ1-327), but not an ADAP mutant containing only the N-terminal 426 amino acids (ADAP-426), coimmunoprecipitated the CARMA1–BCL-10–MALT1 complex (Fig. 4B). The CARMA1 binding site in ADAP was mapped to a region of ADAP between amino acids 426 and 541 (Fig. 4A), because a deletion mutant of ADAP lacking this region (ADAPΔ426-541) was completely unable to coimmunoprecipitate the CARMA1–BCL-10–MALT1 complex (Fig. 4B). This region of ADAP contains the N-terminal helical SH3 (hSH3) domain (2123) between amino acids 482 and 541, and an adjacent region rich in Glu and Lys residues (E/K-rich region) between amino acids 426 and 481. ADAP deletion mutants lacking either the N-terminal hSH3 domain (ADAPΔ482-541) or the E/K-rich region (ADAPΔ426-481) (Fig. 4A) showed weak coimmunoprecipitation of the CARMA1–BCL-10–MALT1 complex relative to wild-type ADAP (Fig. 4B). A GST-ADAP fusion protein expressing both the hSH3 domain and the E/K-rich region of ADAP was able to associate with the truncated CARMA1/651-1147 protein in vitro (fig. S8), which suggests that this region of ADAP is sufficient for CARMA1 association.

Fig. 4.

The region of ADAP containing the N-terminal helical SH3 domain and an E/K-rich region is critical for ADAP-CARMA1 association and TCR-dependent activation of NF-κB, but is dispensable for ADAP-dependent regulation of antigen-dependent conjugate formation. (A) Diagram of the HA-tagged ADAP truncation and deletion mutants used in this study. Numbers indicate amino acid position in mouse ADAP. Asterisks indicate Tyr residues (amino acids 547/549, 584, 615, and 687) implicated in ADAP binding to the SLP-76 adapter protein. (B) Jurkat T cells were transiently transfected with the indicated HA-ADAP constructs and then stimulated with PMA before immunoprecipitation with BCL-10 mAb, followed by Western blotting with antibodies specific for the indicated proteins. (C) T cells isolated from ADAP+/– and ADAP–/– hCAR transgenic mice were transduced with control adenovirus encoding Thy1.1 (ctrl) or adenovirus encoding wild-type ADAP and Thy1.1 (ADAPwt). Cells were either unstimulated or stimulated with PMA or TNF-α before analysis of NF-κB p65 nuclear translocation as in Fig. 2. (D and E) T cells isolated from ADAP+/+ and ADAP–/– hCAR transgenic mice were transduced as in (C) with either a control adenovirus or adenovirus encoding the indicated ADAP constructs. Cells were either unstimulated or stimulated with antibodies to CD3 and CD28 before analysis of NF-κBp65 nuclear translocation. Graphs show the average increase (±SD) in p65 nuclear translocation in stimulated relative to unstimulated T cells for three [(C) and (E)] or two (D) independent experiments. (F) T cells isolated from ADAP+/+ and ADAP–/– DO11.10/hCAR transgenic mice were transduced as in (E) with control adenovirus or the indicated ADAP constructs. T cells were then analyzed by flow cytometry for their ability to form conjugates with unpulsed (unstim.) or ovalbumin-pulsed (OVA) splenocytes.

Expression of wild-type ADAP in resting ADAP–/– T cells restored the ability of PMA (Fig. 4C) or CD3/CD28 stimulation (Fig. 4D) to induce nuclear translocation of p65. ADAP expression in transduced T cells was verified by intracellular flow cytometry (fig. S9). Expression of the ADAPΔ1-327 mutant, but not the ADAP-426 mutant, in ADAP–/– T cells was also able to fully restore CD3/CD28-mediated NF-κBtranslocation (Fig. 4D). In contrast, the ADAPΔ426-541 deletion mutant did not restore CD3/CD28-mediated NF-κB translocation after expression in ADAP–/– T cells (Fig. 4E). Expression of either the ADAPΔ426-481 or ADAPΔ482-541 deletion mutants partially restored NF-κB p65 nuclear translocation (Fig. 4E).

ADAP also regulates TCR-mediated integrin activation (2, 3) and thus ADAP–/– T cells exhibit impaired integrin-dependent conjugate formation with antigen-pulsed antigen-presenting cells (24) (Fig. 4F). Expression of either wild-type ADAP or the ADAPΔ426-541 mutant in ADAP–/– T cells restored TCR-induced conjugate formation to levels observed in control T cells (Fig. 4F). Thus, the region of ADAP between amino acids 426 and 541 is critical for NF-κB activation but is not required for the regulation of integrin-dependent conjugate formation.

We have identified a novel function for ADAP in the regulation of NF-κB activation in T cells. We propose that the association between the C-terminal end of CARMA1 and the region of ADAP between amino acids 426 and 541 is critical for assembly of the CARMA1–BCL-10–MALT1 complex at the membrane. ADAP may provide mechanisms for membrane localization and stabilization of the CARMA1–BCL-10–MALT1 complex, as the C-terminal ADAP hSH3 domain can associate with membrane phospholipids (22). The interaction of ADAP with the MAGUK region of CARMA1 may also alter intramolecular interactions within CARMA1 (25, 26), thereby promoting recruitment of BCL-10 and MALT1. The region of ADAP critical for association with CARMA1 is not required for ADAP-dependent regulation of integrins (2, 3), which involves the association of ADAP with the SKAP-55 and SLP-76 adapters (2729). Two biochemically distinct pools of ADAP can be identified in CD3/CD28-stimulated T cells: one that interacts with the CARMA1–BCL-10–MALT1 complex, and one that interacts with SLP-76 (fig. S10). In contrast, CARMA1 is required for NF-κB activation (2, 3, 30, 31) but is not required for conjugate formation (31). Thus, ADAP serves distinct roles downstream of the TCR that promote functions critical to T cell immune responses.

Supporting Online Material

www.sciencemag.org/cgi/content/full/316/5825/754/DC1

Materials and Methods

Figs. S1 to S10

References

References and Notes

View Abstract

Navigate This Article