p62dok belongs to a newly identified family of adaptor proteins. In T cells, the two members that are predominantly expressed, p56dok and p62dok, are tyrosine phosphorylated upon CD2 or CD28 stimulation, but not upon CD3 ligation. Little is known about the biological role of Dok proteins in T cells. In this study, to evaluate the importance of p62dok in T cell function, we generated Jurkat clones overexpressing p62dok. Our results demonstrate that overexpression of p62dok in Jurkat cells has a dramatic negative effect on CD2-mediated signaling. The p62dok-mediated inhibition affects several biochemical events initiated by CD2 ligation, such as the increase of intracellular Ca2+, phospholipase Cγ1 activation, and extracellular signal-regulated kinase 1/2 activation. Importantly, these cellular events are not affected in the signaling cascade induced by engagement of the CD3/TCR complex. However, both CD3- and CD2-induced NF-AT activation and IL-2 secretion are impaired in p62dok-overexpressing cells. In addition, we show that CD2 but not CD3 stimulation induces p62dok and Ras GTPase-activating protein recruitment to the plasma membrane. These results suggest that p62dok plays a negative role at multiple steps in the CD2 signaling pathway. We propose that p62dok may represent an important negative regulator in the modulation of the response mediated by the TCR.
Dok proteins belong to a newly identified family. Their overall structure suggests that they might serve as docking proteins for signaling molecules. They contain an amino-terminal pleckstrin homology domain (PH),3 a central phosphotyrosine binding domain (PTB), and a carboxyl-terminal domain rich in proline and tyrosine residues. The lowest sequence similarity between the members of the family resides in the carboxyl terminus, suggesting that this region is involved in the recruitment of different sets of downstream signaling molecules. The first member identified, p62dok, was originally shown to be a target of activated protein tyrosine kinases and, when phosphorylated, to associate with Ras GTPase-activating protein (RasGAP) (1). p62dok was subsequently cloned from a p210bcr-abl-transformed myeloid cell line (2) and a v-Abl-transformed precursor B cell line (3). Two other members, p56dok (4, 5, 6, 7) and dok-3 (8, 9), have been characterized.
In T cells, p62dok and p56dok are expressed (10), whereas dok-3 is absent in thymus and in most T cell lines examined (9). Little information is available on the biological function of these proteins in T cells. Recent evidence suggests that they play a specific role in signal transduction pathways initiated by costimulatory receptors. The phosphorylation of p62dok has been reported to occur following CD2 (10) or CD28 stimulation (11), whereas CD3 stimulation does not induce p62dok phosphorylation. Phosphorylation of the other Dok member, p56dok, seems to be regulated in the same way since we have shown that it is phosphorylated upon CD2 stimulation and not upon CD3 stimulation (10). Several lines of evidence support the conclusion that Src family kinases phosphorylate Dok proteins. Cell adhesion-dependent tyrosine phosphorylation of Dok is mediated by Src tyrosine kinases (12). We have shown that Lck is required for CD2-mediated phosphorylation of Dok proteins (10). In transient transfection assays, Dok proteins are good substrates for Src family kinases (9). Phosphorylation of p56dok by Lyn generates binding sites for RasGAP and Nck (7). Depending on the transduction pathway examined, Dok proteins can act as a positive or negative regulator. In B cells, phosphorylation of p62dok is involved in the FcγRIIB-mediated inhibition of B cell receptor signaling (13, 14) most likely by negatively regulating the activity of Ras (15), thereby inhibiting the Ras-dependent activation of extracellular signal-regulated kinase 1/2 (Erk1/2) (13, 14). By contrast, transient overexpression of p62dok in 293 cells (8) and in Chinese hamster ovary cells expressing insulin receptors (12) has been reported to have no effect on v-abl-dependent and insulin-dependent mitogen-activated protein kinase activation, respectively. Moreover, overexpression of p62dok enhances cell migration in response to insulin. p56dok has been reported to be a negative regulator of the Ras activation pathway. Transient overexpression of p56dok diminishes the IL-2-induced activation of mitogen-activated protein kinase and AP-1 (5). The lowered p56dok expression in hr/hr T cells induces an increase of T cell proliferation in response to cytokine and TCR stimulation (5).
In this study, to gain insight into the mechanisms involved in Dok function in T cells, we examined the role of p62dok in regulating T cell signaling. We studied the effect of p62dok overexpression on CD2- and CD3-mediated signal transduction events in Jurkat cells. We demonstrate that overexpression of p62dok specifically inhibits CD2-mediated phospholipase Cγ1 (PLCγ1) phosphorylation, Erk1/2 activation, and Ca2+ mobilization, and that these events remain unaffected when initiated via the CD3/TCR complex.
Materials and Methods
Cell lines and Abs
Jurkat cells, clone 77-6, were grown in RPMI 1640 supplemented with 10% FCS, 2 mM l-glutamine, penicillin, and streptomycin. Puromycin at 1 μg/ml was added to the medium when required.
mAbs used included: anti-CD3ε UCHT1 (IgG2a; kindly provided by A. Alcover, Institut Pasteur, Paris, France); anti-CD2 (anti-T11-2 and T11-3, kindly provided by E. Reinherz, Harvard Medical School, Boston, MA); anti-RasGAP (B4F8; Santa Cruz Biotechnology, Santa Cruz, CA); anti-phosphotyrosine (4G10; Upstate Biotechnology, Lake Placid, NY); and anti-PLCγ1 (a mixture of mAbs; Upstate Biotechnology). Polyclonal Abs used included: anti-p62dok PTB directed against p62dok PTB domain (produced by immunizing rabbits with a GST fusion protein bearing residues 152–259); anti-p62dok directed against aa residues 425–439 of p62dok (kindly provided by B. Stillman, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY); phospho-specific anti-Erk1/2 (Promega, Madison, WI); and total anti-Erk1/2 (New England Biolabs, Beverly, MA).
Plasmids and transfections
The plasmid pSRα-dok was generated by cloning an EcoRI fragment corresponding to the human p62dok cDNA (kindly provided by B. Stillman, Cold Spring Harbor, NY) into the plasmid pSRα puromycin. The hemagglutinin (HA)-p62dok construct (pSRα-HA-dok) was generated as follows. The first methionine residue of p62dok was deleted by PCR and cloned into the plasmid MT073 (kindly provided by M. Thome, Institut de Biochimie, Epalinges, Swizerland) in frame with the sequence encoding the HA epitope. The HA-p62dok was then subcloned into the plasmid pSRα puromycin. J.77-6 cell line was transfected with 20 μg of pSRα-HA-dok or pSRα-dok by electroporation using a Gene Pulser (Bio-Rad, Hercules, CA) set at 250 mV and 960 μF. Drug-resistant cells were cloned by limited dilution in puromycin-containing medium. Expression levels of p62dok were evaluated by immunoblotting of cell extracts with anti-p62dok. Expression levels of CD2 and CD3 were evaluated by flow cytometric analysis with an EPICS XL (Coulter Electronics, Hialeah, FL). Clones expressing similar levels of CD3 and CD2 compared with the parental Jurkat cells were kept for further studies. To quantify the amounts of p62dok expressed in the transfectants, we performed serial dilution of p62dok immunoprecipitates. After Western blotting with p62dok Abs, the ECL signal was quantified using Kodak Image Station 440cf to acquire the image and the 1D Image Analysis software.
Measurement of intracellular Ca2+
Cells were washed twice with HBSS and incubated at 107 cells/ml with 3 μM Indo-1 (Molecular Probes, Eugene, OR) and 0.4 mg/ml Pluronic acid F-127 (Molecular Probes) for 25 min at room temperature. Cells were washed in HBSS and resuspended at 106 cells/ml, and Ca2+ mobilization studies were conducted on an EPICS ELITE ESP cell sorter (Coulter Electronics). Cells were stimulated with anti-CD3ε mAbs (UCHT1, 1/1000 dilution of ascites) or the anti-CD2 mAb pair T11-2 + T11-3 (1/1000 dilution of ascites). Successful loading with Indo-1 was confirmed by subsequently treating the cells with 1 μM ionomycin. Violet/blue ratio signals were analyzed using the MultiTime software (Phoenix Flow Systems, San Diego, CA).
Immunoprecipitations and immunoblotting
Cells were washed twice in RPMI 1640 and resuspended at 5 × 107 cells/ml in RPMI 1640. Cells were left unstimulated or stimulated with anti-CD3ε (UCHT1 at 1/500 dilution of ascites) or anti-CD2 (a combination of T11-2 and T11-3 at 1/1000 dilution of ascites) for the indicated times. Cells were harvested and solubilized for 30 min at 4°C in 1% Nonidet P-40 containing 20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM MgCl2, and 1 mM EGTA in the presence of inhibitors of proteases and phosphatases (10 μg/ml leupeptin and aprotinin, 1 mM Pefabloc-sc, 50 mM NaF, 10 mM Na4P2O7, and 1 mM NaVO4). Immunoprecipitations and immunoblotting were performed as described previously (10).
Cells were washed twice in RPMI 1640 and resuspended at 5 × 107 cells/ml in RPMI 1640. Cells were left unstimulated or stimulated with anti-CD3ε (UCHT1 at 1/500 dilution of ascites) or anti-CD2 (a combination of T11-2 and T11-3 at 1/1000 dilution of ascites) for the indicated times. To determine p62dok and RasGAP cellular redistribution, cells were fractionated into cytosolic and membrane fractions, as described (16). Both fractions were immunoprecipitated with either anti-p62dok or anti-RasGAP Abs.
Jurkat cells (106 cells) were transfected with 2.5 μg NF-AT-firefly luciferase (kindly provided by O. Acuto, Institut Pasteur) and 0.5 μg thymidine kinase (TK)-Renilla luciferase constructs (Promega, Madison, WI) using Fugene transfection assays (Roche Diagnostics, Laval, Quebec, Canada). For transient overexpression of p62dok, 1 μg of pSRα-dok or empty vector was used in combination with 1.5 μg NF-AT-firefly luciferase and 0.5 μg TK-Renilla luciferase constructs. After 24 h, 3 × 105 cells were stimulated with plate-bound anti-CD3 or soluble anti-CD2 Abs for 6 h in a 24-well plate. Maximal stimulation was obtained by a combination of PMA (10 ng/ml) and iononycin (1 μM). Cells were then lysed and assayed for luciferase activity using the dual luciferase reporter assay system (Promega, Madison, WI) and a luminometer (Berthold, LUMAT LB 9507). The NF-AT-firefly luciferase values were normalized based on the constitutive Renilla luciferase activity.
A total of 105 Jurkat cells was stimulated by plate-bound anti-CD3ε mAb (50 μl of UCHT1 at 1/100 dilution of ascites in PBS), soluble anti-CD2 (a combination of T11-2 and T11-3 at 1/1000 dilution of ascites), or PMA (10 ng/ml) and ionomycin (1 μM) for 20 h in 96-well plates. Fifty microliters of supernatant were assayed for IL-2 production using the IL-2-dependent cell line CTLL-2, as described previously (17). 3H incorporation was assessed on a liquid scintillation counter (MicroBeta Trilux).
Overexpression of p62dok does not affect CD2- or CD3-induced tyrosine phosphorylation
To evaluate the importance of p62dok in T cell function, we stably transfected Jurkat cells with an expression vector containing an unmodified or an HA-tagged version of p62dok. Drug-resistant clones overexpressing p62dok were identified by immunoblotting of total cell lysates with anti-p62dok Abs (data not shown). Clones were also tested for expression of CD2 and CD3 cell surface molecules, and those expressing levels comparable with that of the Jurkat parental cell line were kept for further study (data not shown). Three representative clones (clones 2, 74, and HA-1), shown in Fig. 1, that express at least 2.5 times more p62dok than the parental Jurkat cells were subsequently characterized for their signaling capacity.
We investigated whether overexpression of p62dok altered the patterns of tyrosine phosphorylation of cellular proteins following either CD3 or CD2 cross-linking (Fig. 1,A). Phosphorylation patterns remained unaffected in CD3-stimulated cells in clones overexpressing p62dok, indicating that p62dok overexpression did not noticeably prevent the activation of CD3-mediated tyrosine phosphorylation pathways. CD2 stimulation led to the phosphorylation of a similar set of proteins in all the clones, with the exception of a 62-kDa protein, which is hyperphosphorylated in clones overexpressing p62dok. We have previously shown that CD2 stimulation of Jurkat cells induces tyrosine phosphorylation of p62dok (10). To ascertain that the 62-kDa band corresponded to p62dok, p62dok immunoprecipitations were performed from the different Jurkat clones that were left unstimulated, CD2 stimulated, or CD3 stimulated, and were analyzed by anti-phosphotyrosine immunoblotting (Fig. 1 B). As expected, CD2 stimulation of the parental Jurkat cell line and clones that overexpress p62dok led to an increased phosphorylation of p62dok. In contrast, CD3 stimulation did not induce p62dok phosphorylation. It should be noted that there is a higher basal level of p62dok phosphorylation in unstimulated clones overexpressing p62dok compared with the parental Jurkat cells. The increase in constitutively phosphorylated p62dok correlates with the levels of p62dok expressed in these clones. Therefore, overexpression of p62dok does not seem to affect the constitutive phosphorylation of p62dok.
Effect of p62dok overexpression on Ras signaling
To assess the role of p62dok in the Ras signaling pathway, we examined whether p62dok overexpression interferes with the activation of Erk1/2 following CD2 or CD3 engagement (Fig. 2,A). To monitor the activation of Erk1/2, we used Abs specific for the phosphorylated form of Erk1/2. As shown in Fig. 2, phospho-Erk1/2 induction after CD3 stimulation was similar in all clones, indicating that CD3-induced Erk1/2 activation was not inhibited by p62dok overexpression. In contrast, there was a decrease in CD2-mediated Erk1/2 activation in p62dok-overexpressing clones. The intensity of this inhibitory effect correlated with the levels of p62dok overexpression, as evidenced by a greater decrease in CD2-induced Erk1/2 phosphorylation in clone 74 and HA-1 when compared with clone 2. Moreover, the inhibition of CD2-induced Erk1/2 phosphorylation could be rescued by PMA (Fig. 2 B). These results show that p62dok overexpression selectively inhibits CD2- but not CD3-mediated activation of Erk1/2.
Overexpression of p62dok inhibits CD2-induced Ca2+ mobilization
We evaluated the effect of p62dok overexpression on Ca2+ mobilization after CD2 and CD3 stimulation. In Fig. 3, we show that Ca2+ mobilization after CD3 stimulation was unaltered in all cell lines, regardless of p62dok expression levels. In contrast, the CD2-induced Ca2+ response was abolished or greatly diminished in clones overexpressing p62dok. Inositol 3,4,5-triphosphate-mediated mobilization of Ca2+ requires the activation of PLCγ1. Since tyrosine phosphorylation of PLCγ1 contributes to the activation of the enzymatic activity of PLCγ1, we next investigated whether PLCγ1 tyrosine phosphorylation is affected by p62dok overexpression after CD2 or CD3 stimulation (Fig. 4). After CD3 stimulation, PLCγ1 was highly phosphorylated in all cell lines tested. However, there was a slight decrease in the phosphorylation levels of PLCγ1 after CD2 stimulation when we compared clones 2 and 74 with the control J.77-6 cells. More importantly, PLCγ1 phosphorylation was abolished after CD2 stimulation of the HA-1 cell line. The decreased or absence of phosphorylation of PLCγ1 in p62dok-overexpressing cells might be, at least in part, responsible for the diminished Ca2+ influx following CD2 stimulation.
p62dok overexpression interferes with IL-2 gene expression
The activation of signaling pathways initiated by engagement of the CD3/TCR or the CD2 receptor leads to IL-2 production in Jurkat cells. We examined whether p62dok overexpression would influence IL-2 production. Jurkat clones were left unstimulated or stimulated by anti-CD3, anti-CD2 Abs, or by PMA and ionomycin. After 24 h, the culture supernatants were assayed for IL-2 production (Fig. 5,A). Activation of the different clones by PMA and ionomycin resulted in equivalent levels of IL-2 production, indicating that p62dok overexpression did not alter the capacity of the cells to produce IL-2. CD2 stimulation of p62dok-overexpressing clones led to significantly reduced levels of IL-2 production compared with the parental cell line. This result was expected given that IL-2 gene expression is dependent on Ca2+ mobilization and Erk1/2 activation. The inhibition of CD2-induced IL-2 secretion seems to correlate with the amount of p62dok present in the cells. Surprisingly, PMA restored CD2-induced IL-2 secretion (Fig. 5 A). In the clones 74 and HA-1, CD3-induced IL-2 secretion was decreased, whereas CD3-induced IL-2 production in the clone 2 was comparable with that of the parental cell line.
To study in greater detail the mechanisms involved in the inhibition of CD3- and CD2-induced IL-2 secretion in clones overexpressing p62dok, we examined the effect of p62dok overexpression on the activation of transcription factors known to regulate IL-2 gene expression. The control Jurkat cell line as well as clones overexpressing p62dok were transiently transfected with luciferase reporter constructs under the control of NF-AT (Fig. 5,B). Cells were left unstimulated, or stimulated by anti-CD3 or anti-CD2 Abs or PMA and ionomycin for 6 h. The activation of NF-AT was evaluated by a luciferase activity assay. p62dok overexpression inhibited CD2-induced activation of NF-AT (Fig. 5,B). Moreover, the intensity of the inhibition correlates with the levels of p62dok expressed in the cells. We observed a 5- and 15-fold decrease in NF-AT activation for clones 2 and clones 74 and HA-1, respectively. These results were confirmed by performing transient transfection assays with plasmids containing p62dok cDNA and the NF-AT luciferase reporter construct. Transient overexpression of wild-type p62dok or HA-tagged p62dok inhibited (2- to 3-fold) CD2-induced NF-AT activation (Fig. 5 C).
As shown in Fig. 5,D, PMA increases NF-AT activity upon CD2 stimulation in both wild-type Jurkat cells and cells overexpressing p62dok. However, overexpression of p62dok considerably inhibited NF-AT activation in response to CD2 and PMA when compared with the parental cell line. The NF-AT activity in clone HA-1 following treatment with anti-CD2 and PMA is similar to the CD3-induced NF-AT activity in wild-type cells. This result might, at least partially, explain why CD2-induced IL-2 secretion is restored in presence of PMA in clones overexpressing p62dok. NF-AT activation requires the cooperative binding between NF-AT and AP-1 to the IL-2 promoter NF-AT site. NF-AT nuclear translocation is mediated by the Ca2+-regulated phosphatase calcineurin, whereas the synthesis and activation of Fos and Jun, components of the AP-1 family of transcription complexes, are mediated by the protein kinase C (PKC)/Ras pathway. Therefore, the inability of PMA treatment to rescue CD2-induced NF-AT activity is most likely due to the absence of Ca2+ mobilization in these clones. As expected, in cells overexpressing p62dok, restoration of Ca2+ flux by treatment with ionomycin is not sufficient to rescue the CD2-induced NF-AT activation, which is partly dependent on Erk1/2 activation. In contrast, treatment with PMA and ionomycin restored CD2-mediated activation of NF-AT in these cells (Fig. 5 E). This indicates that p62dok overexpression does not lead to nonspecific CD2-mediated inhibition of NF-AT activity.
We were surprised to find that p62dok overexpression interfered with CD3-induced NF-AT activation given that p62dok overexpression does not seem to affect Ca2+ response and Erk1/2 activation (Fig. 5,B). The CD3-induced activation of NF-AT was only slightly reduced in clone 2, whereas there was a marked decrease in CD3-mediated NF-AT activation in clones 74 and HA-1 (2-, 6-, and 28-fold decrease in NF-AT activation for clones 2, 74, and HA-1, respectively). Therefore, as shown for CD2 stimulation, the p62dok-mediated inhibition of CD3-induced NF-AT activation is dependent on the level of Dok overexpression. The inhibition of CD3-induced NF-AT activation can be bypassed by treatment with PMA and calcium ionophore used in combination, but not alone. We have shown that the CD3-mediated Ca2+ response is not affected in clones overexpressing p62dok (Fig. 3). These data clearly suggest that there is a difference between anti-CD3 Abs and Ca2+ ionophore with respect to Ca2+ mobilization. Moreover, CD3-mediated activation of Erk1/2 activation seems normal in clones overexpressing p62dok (Fig. 2). Therefore, in these clones, PMA treatment most likely compensates for a defective component that is required in the NF-AT activation pathway and is downstream or independent of Erk1/2 activation. In addition, in wild-type Jurkat cells, CD3 stimulation or CD2 stimulation further increased NF-AT activity induced by the combination of PMA and ionomycin. Therefore, compared with antireceptor stimuli, the pharmacological stimuli PMA and ionomycin may utilize an additional Ras-independent pathway, which leads to NF-AT activation. These results underline the complexity of the regulation of NF-AT activation. Additional experiments will be required to elucidate the mechanisms by which p62dok mediates inhibition of CD3-induced activation of NF-AT.
Membrane localization and interaction of p62dok with RasGAP following CD2 stimulation
p62dok binding to the Src homology 2 (SH2) domain of RasGAP requires p62dok phosphorylation, and RasGAP is known to be an important negative regulator of Ras. Therefore, Dok-mediated inhibition of Ras signaling might occur through its interaction with RasGAP. To test this hypothesis, we compared the amount of RasGAP associated with p62dok in clones ovexpressing p62dok and in the parental cell line. Lysates from cells that were left unstimulated or stimulated with anti-CD2 mAbs for 3 and 10 min were immunoprecipitated with anti-p62dok Abs. CD2 stimulation resulted in an increase in the amount of RasGAP associated with Dok in cells overexpressing Dok and in the parental cell line (Fig. 6). Importantly, the amount of RasGAP associated with p62dok correlates with the amount of phosphorylated p62dok in the cells (Fig. 6).
Several pieces of evidence suggest that p62dok phosphorylation depends on its subcellular localization and occurs at the plasma membrane (12, 14). We therefore tested whether p62dok overexpression would increase the amount of membrane-bound p62dok and, consequently, would alter the subcellular localization of RasGAP. We performed subcellular fractionation on the parentalJurkat cells and cells overexpressing p62dok that were left unstimulated, CD2 stimulated, or CD3 stimulated (Fig. 7). In all the cell lines, the majority of p62dok was found in the soluble fraction (Fig. 7,A–C). The amount of membrane-bound p62dok in clones overexpressing p62dok was significantly higher than in the parental cell line. However, the percentage of total p62dok present in the membrane fraction in unstimulated cells did not change significantly with varying levels of p62dok. There was a slight increase in the total amount of p62dok in the particulate fraction after CD2 stimulation, whereas CD3 stimulation did not alter the localization of p62dok. The relative proportion of phosphorylated p62dok at each stimulation condition and cytofraction was calculated by normalizing the phosphotyrosine signal to the amount of precipitated protein, and the ratio of these two signals is presented in Fig. 7. These data clearly show that in all the clones, the percentage of phosphorylated p62dok was higher in the membrane fraction than in the cytosolic fraction at each stimulation condition. These results indicate that the phosphorylation of p62dok is mainly occurring at the membrane or that once phosphorylated, p62dok is preferentially recruited to the membrane.
To analyze the effect of CD2 stimulation on RasGAP subcellular distribution, we performed subcellular fractionation. As reported previously, we found that the majority of RasGAP was present in the cytosol (Fig. 7 D). As shown for p62dok, CD2 stimulation induced the translocation of RasGAP to the plasma membrane, whereas CD3 stimulation had no effect on RasGAP subcellular localization. Altogether, these results suggest that recruitment of RasGAP to the membrane following CD2 stimulation is mediated by tyrosine-phosphorylated Dok present at the membrane.
In this study, we analyzed the effect of p62dok overexpression on the signal transduction pathways initiated by CD2 or CD3 cross-linking. We showed that p62dok overexpression interferes with CD2- and CD3-induced IL-2 expression in Jurkat cells. In contrast, overexpression of p62dok specifically inhibits CD2-mediated PLCγ1 phosphorylation, Erk1/2 activation, and Ca2+ mobilization, whereas these events remain unaffected when initiated via the CD3/TCR complex.
Our finding that overexpression of Dok affects PLCγ1 phosphorylation upon CD2 stimulation indicates that p62dok functions early in the signal transduction cascade initiated via CD2. Activation of PLCγ1 requires several targeting signals for its recruitment to the plasma membrane, where its substrate, phosphatidylinositol 4,5-biphosphate, is found. SH2-mediated interaction of PLCγ1 with linker for activation of T cells (18) and binding of its PH domain to the phosphoinositide 3-kinase products, phosphatidylinositol 3,4,5-triphosphate molecules in the membrane, represent two important steps for PLCγ1 activation (19, 20). In addition, tyrosine phosphorylation of PLCγ1 is required for complete stimulation of its enzymatic activity and is most likely mediated by the tyrosine kinase Itk (21, 22). The Gads/SLP-76 complex recruited by linker for activation of T cells also plays an important role for PLCγ1 activation (23, 24). p62dok overexpression may interfere with the formation of any of these signaling complexes. Additional experiments are required to define the molecular basis of p62dok involvement in PLCγ1 activation.
Importantly, we found that Dok-mediated inhibition of PLCγ1 phosphorylation, Erk1/2 activation, and Ca2+ mobilization correlates with phosphorylation of specific tyrosine residues that are phosphorylated upon CD2 cross-linking. Indeed, although the amount of tyrosine-phosphorylated p62dok is higher in unstimulated or CD3-stimulated clones 74 and HA-1 than in CD2-stimulated clone 2, there is no inhibition of CD3-mediated PLCγ-1 phosphorylation, Erk1/2 activation, and Ca2+ mobilization in these clones. Therefore, if phosphotyrosine residues are involved in CD2-mediated inhibition of PLCγ1, Erk1/2, and/or Ca2+ mobilization, these tyrosine residues are specifically phosphorylated upon CD2 stimulation and are different from those that are constitutively phosphorylated.
Since PLCγ1 is a critical regulator of Ca2+ mobilization, defect in PLCγ1 activation is a likely cause of the impaired Ca2+ response in clones overexpressing p62dok. However, we cannot exclude the possibility that overexpression of Dok may directly interfere with Ca2+ mobilization downstream of inositol 3,4,5-triphosphate production.
The inhibition of Erk1/2 activation following CD2 stimulation in clones overexpressing p62dok may also be secondary to defective induction of PLCγ1 activation and diacylglycerol (DAG) production. DAG, a product of PLCγ1, activates PKC, which in turn stimulates Erk1/2 via the Ras/Raf pathways (25). Consistent with a role of PKC-mediated Erk1/2 activation, PMA rescued CD2-induced IL-2 secretion and Erk1/2 activation in clones overexpressing Dok (Figs. 5 A and 2A). Since the contribution of PKC activation in the CD2 regulation of Erk1/2 activity has not been studied, it is not possible to discriminate whether p62dok acts in CD2 signaling upstream of PKC or acts on a CD2-induced Erk1/2-activating pathway independent of PKC.
The recently described Ras-guanine nucleotide exchange factor, RasGRP, has been shown to be involved in CD3-induced Ras activation (26, 27). DAG binding to RasGRP can recruit RasGRP to the membrane and thereby promote Ras signaling. Although the importance of RasGRP in the regulation of CD2-induced Ras activation is unknown, impaired production of DAG in clones overexpressing p62dok might be in part responsible for the decrease in Ras-mediated Erk1/2 activation.
Ras function is negatively regulated by RasGAP, which stimulates the GTPase activity of Ras (28). In addition, RasGAP binds to tyrosine-phosphorylated p62dok via an SH2-mediated interaction (2, 3). Therefore, RasGAP was a good candidate for mediating the negative effect conducted by p62dok in the Ras/Erk1/2 signaling pathway. We have shown that there is a specific recruitment of RasGAP to the plasma membrane following CD2 stimulation and not CD3 stimulation. Moreover, the relocalization of RasGAP is mediated by phosphorylated p62dok. Although it was recently reported using an in vitro assay that p62dok binding to RasGAP diminished its catalytic activity (29), recruitment of RasGAP in the proximity of Ras is likely to have a global negative effect on Ras function. In addition, binding of p62dok to RasGAP might induce a conformational change that exposes domains of RasGAP (such as SH3) to binding partners. This would allow RasGAP to undergo additional protein interactions that might be important in Dok-mediated functions.
Surprisingly, we find that IL-2 secretion and NF-AT activation induced by CD3 cross-linking are inhibited by p62dok overexpression. The level of inhibition correlates with the amount of p62dok overexpression. The inhibition of CD3-induced IL-2 expression might be related to the increased amount of tyrosine-phosphorylated p62dok present in clones overexpressing p62dok. Alternatively, but not exclusively, other structural elements of p62dok might be responsible for Dok-mediated effects on CD3-induced IL-2 secretion. Activation of Erk1/2 is required, but not sufficient for NF-AT induction (30). Since we have shown that CD3-induced Erk1/2 activation is not affected by overexpression of p62dok, p62dok may inhibit the Rac-regulated pathway involved in the regulation of NF-AT induction (31). In unstimulated cells, the amount of membrane-bound p62dok in clones overexpressing p62dok is significantly higher than in the parental cell line. Our current data do not allow us to distinguish whether CD3-mediated inhibition of NF-AT activation takes place in the cytosol or/and in the membrane. In any case, it is important to point out that CD2 stimulation, and not CD3 stimulation, specifically increases tyrosine phosphorylation and induces membrane translocation of p62dok. Therefore, the involvement of p62dok in the signaling cascade initiated by CD2 is likely to be different from that initiated by CD3.
Our data support a model in which phosphorylation of p62dok following CD2 engagement enables its translocation from the cytosolic to the membrane compartment. Activation of phosphoinositide 3-kinase upon CD2 stimulation (32, 33) is likely to be involved in the PH domain-dependent recruitment of p62dok to the membrane (12). Recently identified CD2-binding proteins (34, 35, 36) might also regulate CD2-mediated p62dok membrane localization. Once at the membrane, p62dok is in the proximity of various protein tyrosine kinases, such as Lck, thereby allowing phosphorylation of tyrosine sites that act as SH2 acceptor sites for downstream signaling molecules such as RasGAP. p62dok-mediated RasGAP recruitment to the proximity of Ras leads to the accumulation of RasGDP and consequently to the down-regulation of CD2-mediated Erk1/2 activation. Although this model takes into account the p62dok-mediated Ras inhibition, other molecules are likely to be recruited by p62dok and involved in Dok-mediated inhibition of PLCγ1. To identify such molecules, we have tested the co-association of p62dok with inhibitory molecules such as Csk, SH2-containing inositol 5′-phosphatase-1, or -2, but were not able to detect any significant interactions (data not shown).
Maximal phosphorylation of p62dok occurred 10 min following CD2 stimulation and remained high for 30 min (data not shown). This slow time course supports a model in which p62dok could be considered a hinge molecule that comes into play to down-regulate CD2-mediated activation signals. Although originally described as a receptor delivering stimulatory signals to the cell upon ligand binding (37), CD2 has also been described as a molecule capable of transducing inhibitory signals (38, 39, 40). Since p62dok is clearly involved in the negative regulation of B cell signaling mediated by FcγRIIB (13, 14), p62dok might be considered as a key molecule involved in the modulation of the response mediated by the TCR and the B cell receptor.
In conclusion, p62dok is a multifunctional adaptor protein that most likely plays a negative role at multiple steps in the CD2 signaling pathway. Better insight into the molecules recruited by p62dok will help to identify the molecular mechanisms involved in Dok-mediated inhibition of T cell signaling.
We thank Drs. O. Acuto, A. Alcover, B. Stillman, M. Thome, and E. Reinherz for providing reagents. We also thank Dr. K. Gehring for critical reading of this manuscript.
This work was supported by a grant from the Medical Research Council of Canada.
Abbreviations used in this paper: PH, pleckstrin homology; DAG, diacylglycerol; Erk, extracellular signal-regulated kinase; HA, hemagglutinin; PKC, protein kinase C; PLC, phospholipase C; PTB, phosphotyrosine binding domain; RasGAP, RasGTPase-activating protein; SH, Src homology; TK, thymidine kinase.