Active Ca2/Calmodulin-Dependent Protein Kinase IIB Impairs Positive Selection of T Cells by Modulating TCR Signaling1
http://www.100md.com
免疫学杂志 2005年第14期
Abstract
T cell development is regulated at two critical checkpoints that involve signaling events through the TCR. These signals are propagated by kinases of the Src and Syk families, which activate several adaptor molecules to trigger Ca2 release and, in turn, Ca2/calmodulin-dependent protein kinase II (CaMKII) activation. In this study, we show that a constitutively active form of CaMKII antagonizes TCR signaling and impairs positive selection of thymocytes in mice. Following TCR engagement, active CaMKII decreases TCR-mediated CD3 chain phosphorylation and ZAP70 recruitment, preventing further downstream events. Therefore, we propose that CaMKII belongs to a negative-feedback loop that modulates the strength of the TCR signal through the tyrosine phosphatase Src homology 2 domain-containing phosphatase 2 (SHP-2).
Introduction
The development of T cells is an ordered process that is tightly regulated by signals generated through the TCR. The pre-TCR governs the transition of immature thymocytes from the double-negative (DN)3 to the double-positive (DP) stage. During the DP stage, after rearrangement of the TCR chain is complete, thymocytes must receive a signal of the appropriate strength through the TCR to proceed through development. The absence of a signal results in death by neglect, an intermediate signal promotes positive selection and development into CD4 or CD8 single-positive (SP) cells, and a strong signal induces apoptosis and negative selection (1).
Signals through the TCR are propagated by several molecules. Following TCR engagement, ITAMs within the cytoplasmic tails of the CD3 subunits become phosphorylated by LCK and, to a lesser extent, by FYN (2). ZAP70 binds to phosphorylated ITAMs and undergoes full catalytic activation by LCK (3). Two identified substrates of activated ZAP70 are the adapter proteins Src homology 2 (SH2) domain-containing leukocyte protein of 76 kDa (SLP-76) and linker for activation of T cells (LAT) (4, 5). Phosphorylated LAT induces the formation of a multimolecular complex including VAV, NCK, GADS, phospholipase C (PLC)-, PI3K, CBL, SLAP-130, and GRB2/SOS/RAS (6). The formation of this complex initiates an increase of intracellular Ca2 and inositol phosphates, which ultimately activate members of the MAPK family and transcription factors of the NF-AT and NF-Bfamilies (7).
Among these various signaling molecules, intracellular Ca2 plays an important role in translating the signal through the TCR into a particular cellular outcome (8). For example, high-affinity peptide-TCR interactions induce high-amplitude Ca2 elevations and negative selection, whereas low-affinity peptide-TCR interactions induce low-amplitude Ca2 elevations and positive selection (9). Moreover, TCR stimulation initiates either amplitude- or frequency-encoded Ca2 changes depending on the maturation state of thymocytes, the avidity of TCR interaction, and the coreceptor involved in the engagement (9). Therefore, the transduction of many cellular stimuli during T cell development results in oscillations in the intracellular concentration of Ca2 and influences the fate of thymocytes.
Members of the Ca2/calmodulin-dependent protein kinase II (CaMKII) family are biochemical decoders of intracellular Ca2 oscillations. CAMKII is encoded by four genes (, , , and ) that have distinct expression patterns (10). The and isoforms are expressed in T cells and are activated by binding with the Ca2/calmodulin complex after Ca2 influx. Following activation, CaMKII undergoes an important autophosphorylation of Thr287 in the inhibitory domain (11). This autophosphorylation disrupts the kinase regulatory domain, thereby converting the enzyme into a Ca2-independent (autonomous) form (12). The level of autonomous CaMKII activity is dependent on the initial frequency of Ca2 oscillations produced by various stimuli (13), and thus, CaMKII acts as a biochemical Ca2 decoder. Substitution of Thr287 by an aspartic acid mimics autophosphorylation and generates a kinase that has up to 80% autonomous activity (14, 15). Previously, we generated transgenic mice expressing the CAMKII-T287D mutation (CaMKIIB*) in the T cell lineage, and demonstrated that this active form of CAMKIIB increased the life span of DP thymocytes and the number of mature T cells with a memory phenotype (16). In this study, we provide evidence for another role played by CAMKII. Expression of CaMKIIB* impaired positive selection due to an antagonistic effect on TCR signaling. Our results suggest that CaMKII plays a previously unrecognized, regulatory, feedback role in TCR signaling of thymocytes by sustaining SH2 domain-containing phosphatase 2 (SHP-2) interaction with the TCR complex.
Materials and Methods
Animals
All mice were bred and maintained in the animal facility at the University of California, San Diego, in sterile isolator cages on irradiated food and autoclaved water. AND, H-Y TCR transgenic mice, and CaMKIIB* transgenic mice (founders 50, 66, and 32) were previously described (16, 17, 18). These mice have been backcrossed to C57BL/6J for 11 generations. MHC double-knockout mice were a generous gift from C. Surh (The Scripps Research Institute, La Jolla, CA), and were generated through the mating of 2M–/– and I-Ab–/– deficient mice (19) on C57BL/6J genetic background.
Flow cytometry
Abs to CD4 (CT-CD4) and CD8 (CT-CD8) were from Caltag. The following Abs from BD Pharmingen were used: TCR (H57), CD3 (145-2C11), CD4 (GK1.5), CD5 (53-7.3), CD24 (M1/69), CD25 (7D.4), CD69 (H1.2F3), and CD44 (IM7). T3.70 Ab was produced in our laboratory and revealed with goat anti-mouse IgG (Caltag).
Western blots
A total of 4 x 107 thymocytes was lysed in 200 μl of 1% Triton X-100 buffer (100 mM NaCl, 50 mM HEPES (pH 7.4), 5 mM MgCl2, 1 mM EDTA, 10% glycerol) supplemented with protease inhibitor mixture set I and phosphatase (PTPase) inhibitor mixture set II (Calbiochem) at 4°C for 15 min, followed by a 20-min high-speed spin to remove debris. Following electrophoresis, samples were transferred to a polyvinylidene difluoride membrane (Immobilon-P; Millipore) and blocked with PBS-0.1% Tween 20 supplemented with 5% BSA (Sigma-Aldrich). Blots were developed using ECL (Amersham Biosciences) and analyzed on a Versadoc 3000 (Bio-Rad) or a PhosphorImager Storm 960 (Molecular Dynamics). Anti-phospho-src family Tyr416 was the anti-Tyr394-LCK Ab (Cell Signaling Technology). Anti-phospho-Tyr (clone 4G10), anti-CD4 (H-370), and anti-CD3 (M-20) were from Santa Cruz Biotechnology.
Kinase assay
LCK and FYN activity were monitored by an in vitro kinase assay as follows. Thymocytes were homogenized in lysis buffer (50 mM HEPES (pH 7.4), 100 mM NaCl, 1 mM EGTA, 5 mM MgCl2, 0.5 μg/ml leupeptin, 200 μM sodium vanadate, 1% Triton X-100, 10% glycerol), and the lysate was incubated with an anti-LCK or an anti-FYN Ab or a control Ig at 4°C for 90 min. The immunocomplex was precipitated with protein A-Sepharose. Kinase reactions were conducted with the immunoprecipitates at 37°C for 20 min in 50 μl of kinase buffer (20 mM PIPES (pH 7.6), 10 mM MgCl2, 1 mM EGTA) containing 5 μCi of [32P]ATP (Amersham Biosciences) supplemented by an exogenous substrate, enolase (1 μg; Sigma-Aldrich). The reaction was stopped by adding 10 μl of Laemmli sample loading buffer and was analyzed by SDS-PAGE under reducing conditions. To visualize tyrosine phosphorylation, the gel was subjected to film autoradiography. 32P incorporation into enolase (41 kDa) was measured by densitometry using ImageQuant (Amersham Biosciences).
Retrovirus infection
Retrovirus was produced by transfection of 293T cells by using a Ca2 phosphate transfection kit (Invitrogen Life Technologies). A total of 2.5 x 106 T cells was infected by spinning at 32°C for 120 min. After 8 h at 32°C, cells were transferred to 37°C and analyzed 24–48 h later. Thy1.1 or hu-CD25 were used as a cell surface marker of infection. Thy1.1 or hu-CD25 cells were positively selected using magnetic beads (Miltenyi Biotec).
Calcium flux
A total of 107 thymocytes were labeled in DMEM containing 10 mM HEPES, 2 μM Fura Red (Molecular Probes), and 1 μM Fluo 4 (Molecular Probes), in the presence of 0.2% Pluronic (Molecular Probes) for 25 min at room temperature. Cells were washed in DMEM, 1% FCS with 10 mM HEPES, twice. After adding stimulating Abs (CD3-biotin and/or CD4-biotin), the cells were incubated on ice for 20 min, washed twice, and resuspended in DMEM/1% FCS with 10 mM HEPES. The cells were collected on a FACSCalibur (BD Biosciences) for 30 s to establish a baseline FL1/FL3 ratio (reflective of intracellular free Ca2 levels) for unstimulated cells. Prewarmed streptavidin (4 μg/ml; Pierce) was added to the cells, and the changes in the FL1/FL3 ratio was monitored for the following 6 min. Changes in intracellular Ca2 levels were analyzed by using FlowJo software (Tree Star) (20).
Results
Altered thymocyte populations in CaMKIIB* mice
Although the human CaMKIIB and C isoforms were cloned from T cells (15), the expression of these isoforms in thymocyte subsets was unknown. Western blot analysis indicated that CaMKIIB and C are expressed in all thymocyte subsets (Fig. 1A). To evaluate the role of CaMKIIB in T cell development, we generated transgenic mice expressing CaMKIIB* under the control of the human CD2 promoter or the mouse CD4 promoter (16). Several founders were obtained that expressed the transgene at different levels; founder CaMKIIB*-50 (F50), founder CaMKIIB*-66 (F66), and founder CaMKIIB*-32 (F32) expressed the transgene in the thymus at about five times, four times, and two times the level of endogenous protein, respectively (Fig. 1B). In F32, the CD2 promoter drives expression of the transgene, whereas in F50 and F66, the transgenes contain the mouse CD4 promoter.
To test whether expression of active CAMKIIB altered T cell development, we analyzed the thymocyte cellularity and the developmental subsets from wild-type (WT) and CaMKIIB* mice. The fraction of DP thymocytes was increased in CaMKIIB* mice (percentage of DPs in WT, 82.5 ± 3%; F50, 92 ± 1.4%; F66, 85.5 ± 3%; F32, 90 ± 1%), whereas the percentage of SP cells was decreased in CaMKIIB* mice compared with WT mice (percentage of SPs in WT, 13.4 ± 2.6%; F50, 3 ± 0.6%; F66, 9.9 ± 2.3%; F32, 6.4 ± 1%; Fig. 2A). All founders had a decrease in the number of both CD4 and CD8 SP thymocytes (Fig. 2B). Interestingly, the decrease of SP thymocytes was more severe in the founder with the highest transgene expression level. No major differences were observed in the number of DP and DN thymocytes in mice between 6 and 8 wk of age. In mice older than 12 wk, the number of DP and DN thymocytes was increased up to 2-fold in F32 as described in Bui et al. (16), suggesting that there could be an increase in T cell precursors in mice expressing the transgene under the control of the CD2 promoter (data not shown). Therefore, CaMKIIB* plays a role in T cell development and alters the selection of SP thymocytes.
CaMKIIB* inhibits TCR-proximal tyrosine phosphorylation
Ca2 flux that is initiated by TCR signaling occurs as a result of a series of molecular events that include tyrosine phosphorylation of CD3, recruitment of ZAP70, phosphorylation of PLC, inositol 1,4,5-triphosphate production, and release of Ca2 stores from the endoplasmic reticulum. To investigate the mechanism by which CaMKIIB* decreased Ca2 flux, we next looked at PLC activation. In MHC–/– thymocytes, PLC was phosphorylated after TCR activation; however, the amount of phosphorylated PLC was dramatically reduced in MHC–/–;F50 thymocytes (Fig. 4C). Subsequently, we examined signaling events more proximal to the TCR, and found that the amount of ZAP70 associated with the TCR after cross-linking was significantly reduced in MHC–/–;F50 thymocytes compared with MHC–/– thymocytes (Fig. 4D). This was evident in the amount of phosphorylated and nonphosphorylated ZAP70 associated with the TCR complex. In addition, MHC–/–;F50 thymocytes exhibited decreased CD3 phosphorylation after TCR cross-linking (Fig. 4D). Finally, there was less Erk phosphorylation after TCR cross-linking in CaMKIIB* thymocytes, although it was similar in response to PMA/ionomycin (data not shown). Similar results were obtained in ANDR compared with ANDR50 mice (data not shown). There was no difference in the level of expression of LCK, ZAP70, LAT, SLP-76, PLC, and CD3 in CaMKIIB* thymocytes (data not shown), which rules out the possibility that CaMKIIB* regulates the transcription or the translation of any of these signaling molecules. In summary, CD3 phosphorylation, ZAP70 activation and association with CD3, and PLC activation were all impaired in the presence of constitutively active CaMKIIB*. This suggests that CaMKIIB* inhibits ITAM phosphorylation and, in turn, blocks activation of PLC and prevents optimal release of intracellular Ca2 stores.
Impaired mobilization of intracellular Ca2 in a cell line expressing CaMKIIB*
The simplest interpretation of the data is that CaMKIIB* antagonizes TCR signaling in DP thymocytes. However, this inhibition could be due to abnormal T cell development because we observed a partial block at the DN3 stage (Fig. 2C). Thus, we confirmed our results by expressing the constitutively active CaMKIIB* mutant in a thymocyte cell line. 166 is a p53-deficient thymoma line expressing CD4, CD8, and a low level of CD3 (26). We infected 166 cells with a retrovirus expressing Thy1.1 as a reporter gene, with or without CaMKIIB*. The expression level of CD3, CD4, and Thy1.1 was comparable on the infected cell lines (Fig. 5A). Based on expression of Thy1.1, 98% of the cells were infected. However, the expression level of Thy1.1 was slightly lower in the cells cotransduced with CaMKIIB*, which suggests a toxic effect of high CaMKIIB* expression, as previously observed for transduced fibroblasts (data not shown). In response to CD3/CD4 cross-linking, the cell line expressing CaMKIIB* triggered less Ca2 release than the cell line with the empty vector, whereas the response to ionomycin was similar for both lines (Fig. 5B). This is consistent with the result observed in the CaMKIIB* transgenic thymocytes. Therefore, inhibition of TCR signaling by CaMKIIB* is not due to a developmental defect, but rather it is due to an inhibitory loop involving CaMKII, which can reduce the strength of TCR signaling.
CaMKIIB* inhibitory effect is dominant to the LCKF505 mutant
CD3-ITAMs are phosphorylated mainly by LCK and, to a lesser extent, by FYN (27). The decrease in phosphorylated CD3 and ZAP70 association after TCR engagement in CaMKIIB* mice suggests an inhibition of LCK activity. Thus, we tested whether a constitutively active form of LCK would rescue the defect in positive selection caused by the CaMKIIB* transgene. LCK activation occurs by dephosphorylation of the inhibitory Tyr505 by the transmembrane PTPase, CD45 (28); hence substituting tyrosine for a phenylalanine at this position (Y505F mutation) results in a constitutively active form of LCK (29). Transgenic mice expressing the LCKF505 mutant (PLGF mice) have a block in T cell development at the DP stage because the cells progress from the DN3 stage to the DP stage without TCR gene rearrangement (30). This developmental problem can be partially rescued by expressing a TCR transgene (30). To determine whether a constitutively active LCK could rescue the signaling defect caused by active CaMKIIB*, we bred CaMKIIB* mice with AND;LCKF505 mice. In the latter, expression of CaMKIIB* still blocked positive selection of CD4 cells (Fig. 6, A and B). Therefore, these results suggest that CaMKIIB* impairs positive selection, even when LCK is constitutively activated.
LCK mutations do not restore Ca2 mobilization
Although a constitutively active LCK did not rescue the defect in selection caused by active CaMKII, it was still possible that CaMKII was altering TCR signaling through LCK. In fact, analysis of the LCK amino acid sequence revealed two motifs containing the CaMKII consensus phosphorylation site: R-X-X-S/T (31). These sites are located at residues 39–42 and 207–210 in LCK. We tested whether CAMKII could interfere with LCK activity through these two sites by mutating the serine and threonine residues of the consensus sites to produce: S42A-LCK, T210A-LCK, and the double mutant, dbl-LCK. These mutants and WT LCK were transduced by retrovirus infection into an LCK-deficient cell line, JCaM. This particular JCaM cell line stably expressed mouse CD4 to provide costimulation to CD3. The LCK constructs also contained Thy1.1 to serve as a reporter for transduction. Stable clones were obtained after Thy1.1 positive selection and tested for the ability to flux Ca2 after CD3 and CD4 stimulation. There was no Ca2 mobilization after TCR activation in JCaM cells in the absence of LCK (data not shown); however, it was restored following expression of WT LCK, or either mutant form of LCK (Fig. 6C). This suggests that LCK activity is not affected by mutations at Ser42, Thr210, or both positions. To determine whether CaMKIIB* would differentially affect signaling in these mutants, CaMKIIB* was also transduced into each of the mutants using coexpressed Tac (huCD25) as a reporter gene, and the transduced cells were again tested for the release of Ca2. The presence of CaMKIIB reduced the amount of Ca2 fluxed in response to anti-CD3 stimulation (Fig. 6C). Interestingly, constitutively active CaMKIIB also reduced the amount of Ca2 flux regardless of whether LCK was mutated at one or both putative phosphorylation sites (Fig. 6C). Therefore, CaMKII is not reducing Ca2 flux by directly phosphorylating LCK on either Ser42 or Thr210.
LCK activity is not reduced in CaMKIIB* thymocytes
Finally, to confirm that CaMKII is not inhibiting TCR signaling through LCK activity, we performed an in vitro LCK kinase assay. LCK was immunoprecipitated from AND and ANDR50 thymocytes before and after activation and incubated with the substrate enolase. After activation of thymocytes, LCK activity was not reduced in ANDR50 vs ANDR thymocytes (Fig. 6D). We also did not observe a difference in FYN kinase activity (data not shown). LCK is regulated by phosphorylation on multiple residues including Tyr505 and Tyr394. Phosphorylation at Tyr505 creates an intramolecular association with the SH2 domain, rendering LCK biologically inactive. Phosphorylation of Tyr394 overrides inhibition by phosphorylated Tyr505, resulting in kinase activation (32). To confirm that LCK activity was not modified in CaMKIIB* mice, we looked specifically at LCK Tyr394 phosphorylation and found no difference in the amount of active LCK before and after activation in CaMKIIB* and WT mice (Fig. 6E). Therefore, CaMKIIB* does not affect TCR signaling by inhibiting LCK activity, which suggests that CaMKII targets a different molecule, which is responsible for reduced CD3 phosphorylation.
Inhibition of PTPase restored CD3 Tyr phosphorylation in CaMKIIB* mice
Little is known about the mechanism of TCR inactivation, but several PTPases have been implicated in this process (33). It is possible that CaMKII inhibits TCR signaling by activating a PTPase. If this is true, then inhibition of PTPases should restore a normal CD3 phosphorylation profile and ZAP70 binding in the presence of CaMKIIB*. To test this, we stimulated MHC–/– and MHC–/–;F50 thymocytes in the presence of pervanadate, a general PTPase inhibitor, and then immunoprecipitated the CD3/CD4 complex to analyze CD3 phosphorylation and ZAP70 association. To avoid activation of the T cells before TCR stimulation, we did not pretreat the cells with pervanadate; it was only added in with the streptavidin. As described previously, after TCR stimulation, CD3 phosphorylation and ZAP70 association was decreased in MHC–/–;F50 thymocytes compared with MHC–/– thymocytes. However, in the presence of pervanadate, CD3 phosphorylation, ZAP70 association, and PLC phosphorylation were restored in MHC–/–;F50 thymocytes (Fig. 7A). These data are consistent with a CaMKII-mediated activation of a PTPase, which, in turn, alters the proximal TCR signaling events. Furthermore, this PTPase activity is dominant to LCK activity.
To identify this PTPase, we used a program that predicts substrates for kinases based on the primary sequence of a protein kinase catalytic domain (34). This search predicted a consensus phosphorylation site of RQXSFE, and a search of protein databases revealed four PTPases that contained this site: SHP-1; SHP-2; protein phosphatase 2A (PP2A); and protein tyrosine phosphatase, nonreceptor type 2 (PTPN2). Therefore, we investigated the association of these PTPases with the TCR complex following activation in MHC–/–;F50 thymocytes compared with MHC–/– thymocytes. After stimulation, the CD3/CD4 complex was immunoprecipitated and analyzed for SHP-1, SHP-2, and PP2A association by Western blot (we could not obtain an Ab that recognized PTPN-2, in mouse thymocytes). There was no difference in SHP-1 or PP2A association with the TCR in MHC–/–;F50 and MHC–/– thymocytes, before or after activation (Fig. 7B). In WT thymocytes, SHP-2 was associated with the TCR before activation to maintain TCR in an unphosphorylated state. Upon stimulation, SHP-2 disassociated from the complex and TCR became phosphorylated. However, in MHC–/–;F50 thymocytes, SHP-2 remained associated with the TCR complex even after stimulation (Fig. 7B). This suggests that CaMKII may be phosphorylating and activating SHP-2, maintaining SHP-2 association with the TCR complex and, in turn, halting TCR signaling. This is an additional, novel mechanism in which T cell signaling is regulated.
Discussion
The elevation of intracellular free Ca2 following TCR engagement is essential for T cell activation. The nature of intracellular Ca2 that is released, whether it is a transient elevation, a repetitive oscillation, or a sustained elevation, depends on the avidity of the TCR for the ligand. Ca2 signals are not a binary switch but contain waves of information that need to be decoded (8, 35). CaMKII, a serine/threonine kinase, is an important biochemical decoder of intracellular Ca2 oscillations (13). In this study, we examined the role of CaMKII in T cell development and found that a constitutively active, Ca2-independent form of CaMKIIB reduced positive selection of T cells by maintaining association of SHP-2 with the TCR complex. The persistence of SHP-2 with the TCR complex after activation correlated with a reduction in CD3 phosphorylation and ZAP70 association. In turn, PLC activation, Ca2 flux, and Erk activation were also reduced following TCR stimulation in cells that express active CaMKII. The decrease in Ca2 mobilization was not due to negative regulation of CRAC channels by CaMKII as it has been reported in a neuron cell line (36), because thymocytes from CaMKIIB* mice responded to ionomycin with the same intensity as WT thymocytes, whether or not the medium contained extracellular free Ca2. This work describes a previously unrecognized role of CaMKII in proximal TCR signaling.
CaMKII could be affecting TCR signaling in a number of ways. One possibility is that CaMKII directly inhibits LCK. LCK is responsible for the majority of CD3 phosphorylation and contains two CaMKII consensus phosphorylation sites (31). Furthermore, LCK was previously reported to be a substrate of CaMKII (37). Therefore, we were surprised to find that CaMKII did not inhibit TCR signaling by acting directly upon LCK. This was demonstrated by the fact that the presence of a constitutively active LCK did not rescue the defect in positive selection observed in the presence of CaMKIIB* (Fig. 6, A and B). In addition, mutation of the two potential CaMKII phosphorylation sites on LCK did not abrogate the defect in Ca2 flux that occurs in the presence of CaMKIIB* (Fig. 6C). Finally, there was no difference in LCK phosphorylation or kinase activity in CaMKIIB* and WT thymocytes (Fig. 6, D and E). Together, these data demonstrate that CaMKII does not alter TCR signaling by phosphorylating LCK.
Another possibility is that CaMKII could antagonize TCR signaling by activating and/or recruiting a PTPase to CD3. The fact that the presence of a PTPase inhibitor rescues the signaling defect caused by CaMKIIB* (Fig. 7A) suggests that CaMKII functions upstream of a PTPase. Furthermore, a search of proteins containing the CaMKIIB* phosphorylation site revealed that four PTPases contained this site. We demonstrated that in the presence of CaMKIIB*, more SHP-2 was associated with the TCR complex after activation compared with WT thymocytes (Fig. 7B). We did not find a difference in association of the other PTPases, SHP-1 or PP2A. SHP-2 is a PTPase containing a tandem SH2 domain that is expressed ubiquitously. Compared with SHP-1, much less is known about its physiological function in lymphocyte signaling (38). Previous biochemical data suggest that SHP-2 participates in signaling events downstream of Ag receptors and IL-2, physically interacting with a number of cell surface and cytoplasmic signaling proteins in lymphocytes (39). Notably, the interaction of TCR and CTLA-4 results in SHP-2 recruitment and dephosphorylation of CD3 (40). Together, these data suggest that CaMKII alters TCR signaling through SHP-2, which dephosphorylates CD3 and prevents further signaling through the TCR. Because CaMKII activation depends on the release of Ca2, this suggests that CaMKII belongs to a regulatory loop where initial signaling events will activate CaMKII, leading to thymocyte activation; however, persistent activation of CaMKII will recruit SHP-2 to the TCR complex and, as a result, modulate TCR signaling. We speculate that the level of SHP-2 recruitment to the TCR will depend on the level of CaMKII-autonomous activity, thus providing a mechanism by which TCR signaling can be fine-tuned by Ca2 oscillations. If there are frequent Ca2 oscillations, more CaMKII subunits will be phosphorylated, as in F50 thymocytes, and more SHP-2 will be recruited, blocking positive selection. In contrast, if a low-affinity ligand induces low or infrequent oscillations, then all of the CaMKII subunits will not be phosphorylated, and SHP-2 will not be maintained at the TCR, thereby allowing positive selection to occur.
To our surprise, we were unable to demonstrate phosphorylation of SHP-2 by CAMKII in an in vitro kinase assay, with immunoprecipitated SHP-2 or rSHP-2 as a substrate (data not shown). However, this may be due to the fact that SHP-2 needs to bind to CD3, or another unidentified protein to be phosphorylated by CaMKII. Moreover, SHP-2 may need to be altered by activation to be phosphorylated by CaMKII. It is also possible that the immunoprecipitated SHP-2 was not phosphorylated, because both the kinase and the substrate were immobilized on beads and they could not interact freely. Alternatively, SHP-2 may not be a direct substrate of CAMKII, and constitutively active CAMKII may increase the recruitment of SHP-2 to the TCR via phosphorylation of another protein. Additional experiments need to be done to address this question. Regardless, the presence of constitutively active CAMKII increases SHP-2 association with the TCR and inhibits positive selection.
It is interesting to note that altered CD3 and the lack of ZAP70 recruitment, as well as hyporesponsiveness in IL-2 secretion (41), were also observed in anergic T cells (42). Moreover, a block in inducible IL-2 reporter gene activity was initiated by transfection of CaMKII* in Jurkat T cells (43). Our data suggest the possibility that CaMKII could control IL-2 secretion upstream and downstream of Ca2 signaling by acting on TCR signaling.
We previously reported that, in the presence of the CaMKIIB* transgene, DP thymocytes have a longer life span in the thymus. From the data presented here, we conclude that this is due to the defect in positive selection. Thymocytes remain at the DP stage until they receive a positive selection signal; thus, in a nonselecting background, the DP thymocytes accumulate until they die by neglect (44, 45). Likewise, the CaMKIIB* transgene inhibits positive selection and results in an accumulation of DP thymocytes.
It seems unlikely that the previously unsuspected role of CaMKII in a negative-feedback loop, as described here, would be unique to the TCR, given that CaMKII is ubiquitously expressed. CaMKII is highly concentrated at the CNS synapse (46). Moreover, it associates with the N-methyl-D-asparate (NMDA) receptor and is necessary for normal synaptic plasticity in pyramidal neurons (47). Binding of CaMKII to the NMDA receptor is potentiated by autophosphorylation at Thr286 (48) and may be part of a mechanism modulating the signal strength. Our data suggest that CaMKII modulates the strength of TCR signaling, while regulating a negative feedback loop that could be a more common phenomenon than is currently recognized.
Acknowledgments
We thank Camil Sayegh for the 166 cell line, and Irene Chen for critical reading of the manuscript.
Disclosures
The authors have no financial conflict of interest.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grants AI21372-21 (to S.M.H.)
2 Address correspondence and reprint requests to Dr. Sébastien Calbo at the current address: Institut National de la Santé et de la Recherche Médicale, Unité 519, Faculté de Médecine, 22 boulevard Gambetta, 76183 Rouen Cedex, France. E-mail address: scalbo{at}biomail.ucsd.edu
3 Abbreviations used in this paper: DN, double negative; DP, double positive; SP, single positive; SLP-76, SH2 domain-containing leukocyte protein of 76 kDa; LAT, linker for activation of T cells; PLC, phospholipase C; CaMKII, Ca2/calmodulin-dependent protein kinase II; CaMKIIB*, Ca2-independent mutant of CaMKIIB; SH, Src homology; SHP, SH2 domain-containing phosphatase; WT, wild type; PTPase, phosphatase; PP2A, protein phosphatase 2A; PTPN-2, protein tyrosine phosphatase, nonreceptor type 2; NMDA, N-methyl-D-asparate.
Received for publication January 25, 2005. Accepted for publication April 21, 2005.
References
Starr, T. K., S. C. Jameson, K. A. Hogquist. 2003. Positive and negative selection of T cells. Annu. Rev. Immunol. 21: 139-176.
van Oers, N. S., N. Killeen, A. Weiss. 1996. Lck regulates the tyrosine phosphorylation of the T cell receptor subunits and ZAP-70 in murine thymocytes. J. Exp. Med. 183: 1053-1062.
Chan, A. C., M. Dalton, R. Johnson, G. H. Kong, T. Wang, R. Thoma, T. Kurosaki. 1995. Activation of ZAP-70 kinase activity by phosphorylation of tyrosine 493 is required for lymphocyte antigen receptor function. EMBO J. 14: 2499-2508.
Zhang, W., J. Sloan-Lancaster, J. Kitchen, R. P. Trible, L. E. Samelson. 1998. LAT: the ZAP-70 tyrosine kinase substrate that links T cell receptor to cellular activation. Cell 92: 83-92.
Wardenburg, J. B., C. Fu, J. K. Jackman, H. Flotow, S. E. Wilkinson, D. H. Williams, R. Johnson, G. Kong, A. C. Chan, P. R. Findell. 1996. Phosphorylation of SLP-76 by the ZAP-70 protein-tyrosine kinase is required for T-cell receptor function. J. Biol. Chem. 271: 19641-19644.
Pivniouk, V. I., R. S. Geha. 2000. The role of SLP-76 and LAT in lymphocyte development. Curr. Opin. Immunol. 12: 173-178.
Crabtree, G. R., N. A. Clipstone. 1994. Signal transmission between the plasma membrane and nucleus of T lymphocytes. Annu. Rev. Biochem. 63: 1045-1083.
Dolmetsch, R. E., K. Xu, R. S. Lewis. 1998. Calcium oscillations increase the efficiency and specificity of gene expression. Nature 392: 933-936.
Freedman, B. D., Q. H. Liu, S. Somersan, M. I. Kotlikoff, J. A. Punt. 1999. Receptor avidity and costimulation specify the intracellular Ca2 signaling pattern in CD4CD8 thymocytes. J. Exp. Med. 190: 943-952.
Tobimatsu, T., H. Fujisawa. 1989. Tissue-specific expression of four types of rat calmodulin-dependent protein kinase II mRNAs. J. Biol. Chem. 264: 17907-17912.(Maureen A. McGargill*, Le)
T cell development is regulated at two critical checkpoints that involve signaling events through the TCR. These signals are propagated by kinases of the Src and Syk families, which activate several adaptor molecules to trigger Ca2 release and, in turn, Ca2/calmodulin-dependent protein kinase II (CaMKII) activation. In this study, we show that a constitutively active form of CaMKII antagonizes TCR signaling and impairs positive selection of thymocytes in mice. Following TCR engagement, active CaMKII decreases TCR-mediated CD3 chain phosphorylation and ZAP70 recruitment, preventing further downstream events. Therefore, we propose that CaMKII belongs to a negative-feedback loop that modulates the strength of the TCR signal through the tyrosine phosphatase Src homology 2 domain-containing phosphatase 2 (SHP-2).
Introduction
The development of T cells is an ordered process that is tightly regulated by signals generated through the TCR. The pre-TCR governs the transition of immature thymocytes from the double-negative (DN)3 to the double-positive (DP) stage. During the DP stage, after rearrangement of the TCR chain is complete, thymocytes must receive a signal of the appropriate strength through the TCR to proceed through development. The absence of a signal results in death by neglect, an intermediate signal promotes positive selection and development into CD4 or CD8 single-positive (SP) cells, and a strong signal induces apoptosis and negative selection (1).
Signals through the TCR are propagated by several molecules. Following TCR engagement, ITAMs within the cytoplasmic tails of the CD3 subunits become phosphorylated by LCK and, to a lesser extent, by FYN (2). ZAP70 binds to phosphorylated ITAMs and undergoes full catalytic activation by LCK (3). Two identified substrates of activated ZAP70 are the adapter proteins Src homology 2 (SH2) domain-containing leukocyte protein of 76 kDa (SLP-76) and linker for activation of T cells (LAT) (4, 5). Phosphorylated LAT induces the formation of a multimolecular complex including VAV, NCK, GADS, phospholipase C (PLC)-, PI3K, CBL, SLAP-130, and GRB2/SOS/RAS (6). The formation of this complex initiates an increase of intracellular Ca2 and inositol phosphates, which ultimately activate members of the MAPK family and transcription factors of the NF-AT and NF-Bfamilies (7).
Among these various signaling molecules, intracellular Ca2 plays an important role in translating the signal through the TCR into a particular cellular outcome (8). For example, high-affinity peptide-TCR interactions induce high-amplitude Ca2 elevations and negative selection, whereas low-affinity peptide-TCR interactions induce low-amplitude Ca2 elevations and positive selection (9). Moreover, TCR stimulation initiates either amplitude- or frequency-encoded Ca2 changes depending on the maturation state of thymocytes, the avidity of TCR interaction, and the coreceptor involved in the engagement (9). Therefore, the transduction of many cellular stimuli during T cell development results in oscillations in the intracellular concentration of Ca2 and influences the fate of thymocytes.
Members of the Ca2/calmodulin-dependent protein kinase II (CaMKII) family are biochemical decoders of intracellular Ca2 oscillations. CAMKII is encoded by four genes (, , , and ) that have distinct expression patterns (10). The and isoforms are expressed in T cells and are activated by binding with the Ca2/calmodulin complex after Ca2 influx. Following activation, CaMKII undergoes an important autophosphorylation of Thr287 in the inhibitory domain (11). This autophosphorylation disrupts the kinase regulatory domain, thereby converting the enzyme into a Ca2-independent (autonomous) form (12). The level of autonomous CaMKII activity is dependent on the initial frequency of Ca2 oscillations produced by various stimuli (13), and thus, CaMKII acts as a biochemical Ca2 decoder. Substitution of Thr287 by an aspartic acid mimics autophosphorylation and generates a kinase that has up to 80% autonomous activity (14, 15). Previously, we generated transgenic mice expressing the CAMKII-T287D mutation (CaMKIIB*) in the T cell lineage, and demonstrated that this active form of CAMKIIB increased the life span of DP thymocytes and the number of mature T cells with a memory phenotype (16). In this study, we provide evidence for another role played by CAMKII. Expression of CaMKIIB* impaired positive selection due to an antagonistic effect on TCR signaling. Our results suggest that CaMKII plays a previously unrecognized, regulatory, feedback role in TCR signaling of thymocytes by sustaining SH2 domain-containing phosphatase 2 (SHP-2) interaction with the TCR complex.
Materials and Methods
Animals
All mice were bred and maintained in the animal facility at the University of California, San Diego, in sterile isolator cages on irradiated food and autoclaved water. AND, H-Y TCR transgenic mice, and CaMKIIB* transgenic mice (founders 50, 66, and 32) were previously described (16, 17, 18). These mice have been backcrossed to C57BL/6J for 11 generations. MHC double-knockout mice were a generous gift from C. Surh (The Scripps Research Institute, La Jolla, CA), and were generated through the mating of 2M–/– and I-Ab–/– deficient mice (19) on C57BL/6J genetic background.
Flow cytometry
Abs to CD4 (CT-CD4) and CD8 (CT-CD8) were from Caltag. The following Abs from BD Pharmingen were used: TCR (H57), CD3 (145-2C11), CD4 (GK1.5), CD5 (53-7.3), CD24 (M1/69), CD25 (7D.4), CD69 (H1.2F3), and CD44 (IM7). T3.70 Ab was produced in our laboratory and revealed with goat anti-mouse IgG (Caltag).
Western blots
A total of 4 x 107 thymocytes was lysed in 200 μl of 1% Triton X-100 buffer (100 mM NaCl, 50 mM HEPES (pH 7.4), 5 mM MgCl2, 1 mM EDTA, 10% glycerol) supplemented with protease inhibitor mixture set I and phosphatase (PTPase) inhibitor mixture set II (Calbiochem) at 4°C for 15 min, followed by a 20-min high-speed spin to remove debris. Following electrophoresis, samples were transferred to a polyvinylidene difluoride membrane (Immobilon-P; Millipore) and blocked with PBS-0.1% Tween 20 supplemented with 5% BSA (Sigma-Aldrich). Blots were developed using ECL (Amersham Biosciences) and analyzed on a Versadoc 3000 (Bio-Rad) or a PhosphorImager Storm 960 (Molecular Dynamics). Anti-phospho-src family Tyr416 was the anti-Tyr394-LCK Ab (Cell Signaling Technology). Anti-phospho-Tyr (clone 4G10), anti-CD4 (H-370), and anti-CD3 (M-20) were from Santa Cruz Biotechnology.
Kinase assay
LCK and FYN activity were monitored by an in vitro kinase assay as follows. Thymocytes were homogenized in lysis buffer (50 mM HEPES (pH 7.4), 100 mM NaCl, 1 mM EGTA, 5 mM MgCl2, 0.5 μg/ml leupeptin, 200 μM sodium vanadate, 1% Triton X-100, 10% glycerol), and the lysate was incubated with an anti-LCK or an anti-FYN Ab or a control Ig at 4°C for 90 min. The immunocomplex was precipitated with protein A-Sepharose. Kinase reactions were conducted with the immunoprecipitates at 37°C for 20 min in 50 μl of kinase buffer (20 mM PIPES (pH 7.6), 10 mM MgCl2, 1 mM EGTA) containing 5 μCi of [32P]ATP (Amersham Biosciences) supplemented by an exogenous substrate, enolase (1 μg; Sigma-Aldrich). The reaction was stopped by adding 10 μl of Laemmli sample loading buffer and was analyzed by SDS-PAGE under reducing conditions. To visualize tyrosine phosphorylation, the gel was subjected to film autoradiography. 32P incorporation into enolase (41 kDa) was measured by densitometry using ImageQuant (Amersham Biosciences).
Retrovirus infection
Retrovirus was produced by transfection of 293T cells by using a Ca2 phosphate transfection kit (Invitrogen Life Technologies). A total of 2.5 x 106 T cells was infected by spinning at 32°C for 120 min. After 8 h at 32°C, cells were transferred to 37°C and analyzed 24–48 h later. Thy1.1 or hu-CD25 were used as a cell surface marker of infection. Thy1.1 or hu-CD25 cells were positively selected using magnetic beads (Miltenyi Biotec).
Calcium flux
A total of 107 thymocytes were labeled in DMEM containing 10 mM HEPES, 2 μM Fura Red (Molecular Probes), and 1 μM Fluo 4 (Molecular Probes), in the presence of 0.2% Pluronic (Molecular Probes) for 25 min at room temperature. Cells were washed in DMEM, 1% FCS with 10 mM HEPES, twice. After adding stimulating Abs (CD3-biotin and/or CD4-biotin), the cells were incubated on ice for 20 min, washed twice, and resuspended in DMEM/1% FCS with 10 mM HEPES. The cells were collected on a FACSCalibur (BD Biosciences) for 30 s to establish a baseline FL1/FL3 ratio (reflective of intracellular free Ca2 levels) for unstimulated cells. Prewarmed streptavidin (4 μg/ml; Pierce) was added to the cells, and the changes in the FL1/FL3 ratio was monitored for the following 6 min. Changes in intracellular Ca2 levels were analyzed by using FlowJo software (Tree Star) (20).
Results
Altered thymocyte populations in CaMKIIB* mice
Although the human CaMKIIB and C isoforms were cloned from T cells (15), the expression of these isoforms in thymocyte subsets was unknown. Western blot analysis indicated that CaMKIIB and C are expressed in all thymocyte subsets (Fig. 1A). To evaluate the role of CaMKIIB in T cell development, we generated transgenic mice expressing CaMKIIB* under the control of the human CD2 promoter or the mouse CD4 promoter (16). Several founders were obtained that expressed the transgene at different levels; founder CaMKIIB*-50 (F50), founder CaMKIIB*-66 (F66), and founder CaMKIIB*-32 (F32) expressed the transgene in the thymus at about five times, four times, and two times the level of endogenous protein, respectively (Fig. 1B). In F32, the CD2 promoter drives expression of the transgene, whereas in F50 and F66, the transgenes contain the mouse CD4 promoter.
To test whether expression of active CAMKIIB altered T cell development, we analyzed the thymocyte cellularity and the developmental subsets from wild-type (WT) and CaMKIIB* mice. The fraction of DP thymocytes was increased in CaMKIIB* mice (percentage of DPs in WT, 82.5 ± 3%; F50, 92 ± 1.4%; F66, 85.5 ± 3%; F32, 90 ± 1%), whereas the percentage of SP cells was decreased in CaMKIIB* mice compared with WT mice (percentage of SPs in WT, 13.4 ± 2.6%; F50, 3 ± 0.6%; F66, 9.9 ± 2.3%; F32, 6.4 ± 1%; Fig. 2A). All founders had a decrease in the number of both CD4 and CD8 SP thymocytes (Fig. 2B). Interestingly, the decrease of SP thymocytes was more severe in the founder with the highest transgene expression level. No major differences were observed in the number of DP and DN thymocytes in mice between 6 and 8 wk of age. In mice older than 12 wk, the number of DP and DN thymocytes was increased up to 2-fold in F32 as described in Bui et al. (16), suggesting that there could be an increase in T cell precursors in mice expressing the transgene under the control of the CD2 promoter (data not shown). Therefore, CaMKIIB* plays a role in T cell development and alters the selection of SP thymocytes.
CaMKIIB* inhibits TCR-proximal tyrosine phosphorylation
Ca2 flux that is initiated by TCR signaling occurs as a result of a series of molecular events that include tyrosine phosphorylation of CD3, recruitment of ZAP70, phosphorylation of PLC, inositol 1,4,5-triphosphate production, and release of Ca2 stores from the endoplasmic reticulum. To investigate the mechanism by which CaMKIIB* decreased Ca2 flux, we next looked at PLC activation. In MHC–/– thymocytes, PLC was phosphorylated after TCR activation; however, the amount of phosphorylated PLC was dramatically reduced in MHC–/–;F50 thymocytes (Fig. 4C). Subsequently, we examined signaling events more proximal to the TCR, and found that the amount of ZAP70 associated with the TCR after cross-linking was significantly reduced in MHC–/–;F50 thymocytes compared with MHC–/– thymocytes (Fig. 4D). This was evident in the amount of phosphorylated and nonphosphorylated ZAP70 associated with the TCR complex. In addition, MHC–/–;F50 thymocytes exhibited decreased CD3 phosphorylation after TCR cross-linking (Fig. 4D). Finally, there was less Erk phosphorylation after TCR cross-linking in CaMKIIB* thymocytes, although it was similar in response to PMA/ionomycin (data not shown). Similar results were obtained in ANDR compared with ANDR50 mice (data not shown). There was no difference in the level of expression of LCK, ZAP70, LAT, SLP-76, PLC, and CD3 in CaMKIIB* thymocytes (data not shown), which rules out the possibility that CaMKIIB* regulates the transcription or the translation of any of these signaling molecules. In summary, CD3 phosphorylation, ZAP70 activation and association with CD3, and PLC activation were all impaired in the presence of constitutively active CaMKIIB*. This suggests that CaMKIIB* inhibits ITAM phosphorylation and, in turn, blocks activation of PLC and prevents optimal release of intracellular Ca2 stores.
Impaired mobilization of intracellular Ca2 in a cell line expressing CaMKIIB*
The simplest interpretation of the data is that CaMKIIB* antagonizes TCR signaling in DP thymocytes. However, this inhibition could be due to abnormal T cell development because we observed a partial block at the DN3 stage (Fig. 2C). Thus, we confirmed our results by expressing the constitutively active CaMKIIB* mutant in a thymocyte cell line. 166 is a p53-deficient thymoma line expressing CD4, CD8, and a low level of CD3 (26). We infected 166 cells with a retrovirus expressing Thy1.1 as a reporter gene, with or without CaMKIIB*. The expression level of CD3, CD4, and Thy1.1 was comparable on the infected cell lines (Fig. 5A). Based on expression of Thy1.1, 98% of the cells were infected. However, the expression level of Thy1.1 was slightly lower in the cells cotransduced with CaMKIIB*, which suggests a toxic effect of high CaMKIIB* expression, as previously observed for transduced fibroblasts (data not shown). In response to CD3/CD4 cross-linking, the cell line expressing CaMKIIB* triggered less Ca2 release than the cell line with the empty vector, whereas the response to ionomycin was similar for both lines (Fig. 5B). This is consistent with the result observed in the CaMKIIB* transgenic thymocytes. Therefore, inhibition of TCR signaling by CaMKIIB* is not due to a developmental defect, but rather it is due to an inhibitory loop involving CaMKII, which can reduce the strength of TCR signaling.
CaMKIIB* inhibitory effect is dominant to the LCKF505 mutant
CD3-ITAMs are phosphorylated mainly by LCK and, to a lesser extent, by FYN (27). The decrease in phosphorylated CD3 and ZAP70 association after TCR engagement in CaMKIIB* mice suggests an inhibition of LCK activity. Thus, we tested whether a constitutively active form of LCK would rescue the defect in positive selection caused by the CaMKIIB* transgene. LCK activation occurs by dephosphorylation of the inhibitory Tyr505 by the transmembrane PTPase, CD45 (28); hence substituting tyrosine for a phenylalanine at this position (Y505F mutation) results in a constitutively active form of LCK (29). Transgenic mice expressing the LCKF505 mutant (PLGF mice) have a block in T cell development at the DP stage because the cells progress from the DN3 stage to the DP stage without TCR gene rearrangement (30). This developmental problem can be partially rescued by expressing a TCR transgene (30). To determine whether a constitutively active LCK could rescue the signaling defect caused by active CaMKIIB*, we bred CaMKIIB* mice with AND;LCKF505 mice. In the latter, expression of CaMKIIB* still blocked positive selection of CD4 cells (Fig. 6, A and B). Therefore, these results suggest that CaMKIIB* impairs positive selection, even when LCK is constitutively activated.
LCK mutations do not restore Ca2 mobilization
Although a constitutively active LCK did not rescue the defect in selection caused by active CaMKII, it was still possible that CaMKII was altering TCR signaling through LCK. In fact, analysis of the LCK amino acid sequence revealed two motifs containing the CaMKII consensus phosphorylation site: R-X-X-S/T (31). These sites are located at residues 39–42 and 207–210 in LCK. We tested whether CAMKII could interfere with LCK activity through these two sites by mutating the serine and threonine residues of the consensus sites to produce: S42A-LCK, T210A-LCK, and the double mutant, dbl-LCK. These mutants and WT LCK were transduced by retrovirus infection into an LCK-deficient cell line, JCaM. This particular JCaM cell line stably expressed mouse CD4 to provide costimulation to CD3. The LCK constructs also contained Thy1.1 to serve as a reporter for transduction. Stable clones were obtained after Thy1.1 positive selection and tested for the ability to flux Ca2 after CD3 and CD4 stimulation. There was no Ca2 mobilization after TCR activation in JCaM cells in the absence of LCK (data not shown); however, it was restored following expression of WT LCK, or either mutant form of LCK (Fig. 6C). This suggests that LCK activity is not affected by mutations at Ser42, Thr210, or both positions. To determine whether CaMKIIB* would differentially affect signaling in these mutants, CaMKIIB* was also transduced into each of the mutants using coexpressed Tac (huCD25) as a reporter gene, and the transduced cells were again tested for the release of Ca2. The presence of CaMKIIB reduced the amount of Ca2 fluxed in response to anti-CD3 stimulation (Fig. 6C). Interestingly, constitutively active CaMKIIB also reduced the amount of Ca2 flux regardless of whether LCK was mutated at one or both putative phosphorylation sites (Fig. 6C). Therefore, CaMKII is not reducing Ca2 flux by directly phosphorylating LCK on either Ser42 or Thr210.
LCK activity is not reduced in CaMKIIB* thymocytes
Finally, to confirm that CaMKII is not inhibiting TCR signaling through LCK activity, we performed an in vitro LCK kinase assay. LCK was immunoprecipitated from AND and ANDR50 thymocytes before and after activation and incubated with the substrate enolase. After activation of thymocytes, LCK activity was not reduced in ANDR50 vs ANDR thymocytes (Fig. 6D). We also did not observe a difference in FYN kinase activity (data not shown). LCK is regulated by phosphorylation on multiple residues including Tyr505 and Tyr394. Phosphorylation at Tyr505 creates an intramolecular association with the SH2 domain, rendering LCK biologically inactive. Phosphorylation of Tyr394 overrides inhibition by phosphorylated Tyr505, resulting in kinase activation (32). To confirm that LCK activity was not modified in CaMKIIB* mice, we looked specifically at LCK Tyr394 phosphorylation and found no difference in the amount of active LCK before and after activation in CaMKIIB* and WT mice (Fig. 6E). Therefore, CaMKIIB* does not affect TCR signaling by inhibiting LCK activity, which suggests that CaMKII targets a different molecule, which is responsible for reduced CD3 phosphorylation.
Inhibition of PTPase restored CD3 Tyr phosphorylation in CaMKIIB* mice
Little is known about the mechanism of TCR inactivation, but several PTPases have been implicated in this process (33). It is possible that CaMKII inhibits TCR signaling by activating a PTPase. If this is true, then inhibition of PTPases should restore a normal CD3 phosphorylation profile and ZAP70 binding in the presence of CaMKIIB*. To test this, we stimulated MHC–/– and MHC–/–;F50 thymocytes in the presence of pervanadate, a general PTPase inhibitor, and then immunoprecipitated the CD3/CD4 complex to analyze CD3 phosphorylation and ZAP70 association. To avoid activation of the T cells before TCR stimulation, we did not pretreat the cells with pervanadate; it was only added in with the streptavidin. As described previously, after TCR stimulation, CD3 phosphorylation and ZAP70 association was decreased in MHC–/–;F50 thymocytes compared with MHC–/– thymocytes. However, in the presence of pervanadate, CD3 phosphorylation, ZAP70 association, and PLC phosphorylation were restored in MHC–/–;F50 thymocytes (Fig. 7A). These data are consistent with a CaMKII-mediated activation of a PTPase, which, in turn, alters the proximal TCR signaling events. Furthermore, this PTPase activity is dominant to LCK activity.
To identify this PTPase, we used a program that predicts substrates for kinases based on the primary sequence of a protein kinase catalytic domain (34). This search predicted a consensus phosphorylation site of RQXSFE, and a search of protein databases revealed four PTPases that contained this site: SHP-1; SHP-2; protein phosphatase 2A (PP2A); and protein tyrosine phosphatase, nonreceptor type 2 (PTPN2). Therefore, we investigated the association of these PTPases with the TCR complex following activation in MHC–/–;F50 thymocytes compared with MHC–/– thymocytes. After stimulation, the CD3/CD4 complex was immunoprecipitated and analyzed for SHP-1, SHP-2, and PP2A association by Western blot (we could not obtain an Ab that recognized PTPN-2, in mouse thymocytes). There was no difference in SHP-1 or PP2A association with the TCR in MHC–/–;F50 and MHC–/– thymocytes, before or after activation (Fig. 7B). In WT thymocytes, SHP-2 was associated with the TCR before activation to maintain TCR in an unphosphorylated state. Upon stimulation, SHP-2 disassociated from the complex and TCR became phosphorylated. However, in MHC–/–;F50 thymocytes, SHP-2 remained associated with the TCR complex even after stimulation (Fig. 7B). This suggests that CaMKII may be phosphorylating and activating SHP-2, maintaining SHP-2 association with the TCR complex and, in turn, halting TCR signaling. This is an additional, novel mechanism in which T cell signaling is regulated.
Discussion
The elevation of intracellular free Ca2 following TCR engagement is essential for T cell activation. The nature of intracellular Ca2 that is released, whether it is a transient elevation, a repetitive oscillation, or a sustained elevation, depends on the avidity of the TCR for the ligand. Ca2 signals are not a binary switch but contain waves of information that need to be decoded (8, 35). CaMKII, a serine/threonine kinase, is an important biochemical decoder of intracellular Ca2 oscillations (13). In this study, we examined the role of CaMKII in T cell development and found that a constitutively active, Ca2-independent form of CaMKIIB reduced positive selection of T cells by maintaining association of SHP-2 with the TCR complex. The persistence of SHP-2 with the TCR complex after activation correlated with a reduction in CD3 phosphorylation and ZAP70 association. In turn, PLC activation, Ca2 flux, and Erk activation were also reduced following TCR stimulation in cells that express active CaMKII. The decrease in Ca2 mobilization was not due to negative regulation of CRAC channels by CaMKII as it has been reported in a neuron cell line (36), because thymocytes from CaMKIIB* mice responded to ionomycin with the same intensity as WT thymocytes, whether or not the medium contained extracellular free Ca2. This work describes a previously unrecognized role of CaMKII in proximal TCR signaling.
CaMKII could be affecting TCR signaling in a number of ways. One possibility is that CaMKII directly inhibits LCK. LCK is responsible for the majority of CD3 phosphorylation and contains two CaMKII consensus phosphorylation sites (31). Furthermore, LCK was previously reported to be a substrate of CaMKII (37). Therefore, we were surprised to find that CaMKII did not inhibit TCR signaling by acting directly upon LCK. This was demonstrated by the fact that the presence of a constitutively active LCK did not rescue the defect in positive selection observed in the presence of CaMKIIB* (Fig. 6, A and B). In addition, mutation of the two potential CaMKII phosphorylation sites on LCK did not abrogate the defect in Ca2 flux that occurs in the presence of CaMKIIB* (Fig. 6C). Finally, there was no difference in LCK phosphorylation or kinase activity in CaMKIIB* and WT thymocytes (Fig. 6, D and E). Together, these data demonstrate that CaMKII does not alter TCR signaling by phosphorylating LCK.
Another possibility is that CaMKII could antagonize TCR signaling by activating and/or recruiting a PTPase to CD3. The fact that the presence of a PTPase inhibitor rescues the signaling defect caused by CaMKIIB* (Fig. 7A) suggests that CaMKII functions upstream of a PTPase. Furthermore, a search of proteins containing the CaMKIIB* phosphorylation site revealed that four PTPases contained this site. We demonstrated that in the presence of CaMKIIB*, more SHP-2 was associated with the TCR complex after activation compared with WT thymocytes (Fig. 7B). We did not find a difference in association of the other PTPases, SHP-1 or PP2A. SHP-2 is a PTPase containing a tandem SH2 domain that is expressed ubiquitously. Compared with SHP-1, much less is known about its physiological function in lymphocyte signaling (38). Previous biochemical data suggest that SHP-2 participates in signaling events downstream of Ag receptors and IL-2, physically interacting with a number of cell surface and cytoplasmic signaling proteins in lymphocytes (39). Notably, the interaction of TCR and CTLA-4 results in SHP-2 recruitment and dephosphorylation of CD3 (40). Together, these data suggest that CaMKII alters TCR signaling through SHP-2, which dephosphorylates CD3 and prevents further signaling through the TCR. Because CaMKII activation depends on the release of Ca2, this suggests that CaMKII belongs to a regulatory loop where initial signaling events will activate CaMKII, leading to thymocyte activation; however, persistent activation of CaMKII will recruit SHP-2 to the TCR complex and, as a result, modulate TCR signaling. We speculate that the level of SHP-2 recruitment to the TCR will depend on the level of CaMKII-autonomous activity, thus providing a mechanism by which TCR signaling can be fine-tuned by Ca2 oscillations. If there are frequent Ca2 oscillations, more CaMKII subunits will be phosphorylated, as in F50 thymocytes, and more SHP-2 will be recruited, blocking positive selection. In contrast, if a low-affinity ligand induces low or infrequent oscillations, then all of the CaMKII subunits will not be phosphorylated, and SHP-2 will not be maintained at the TCR, thereby allowing positive selection to occur.
To our surprise, we were unable to demonstrate phosphorylation of SHP-2 by CAMKII in an in vitro kinase assay, with immunoprecipitated SHP-2 or rSHP-2 as a substrate (data not shown). However, this may be due to the fact that SHP-2 needs to bind to CD3, or another unidentified protein to be phosphorylated by CaMKII. Moreover, SHP-2 may need to be altered by activation to be phosphorylated by CaMKII. It is also possible that the immunoprecipitated SHP-2 was not phosphorylated, because both the kinase and the substrate were immobilized on beads and they could not interact freely. Alternatively, SHP-2 may not be a direct substrate of CAMKII, and constitutively active CAMKII may increase the recruitment of SHP-2 to the TCR via phosphorylation of another protein. Additional experiments need to be done to address this question. Regardless, the presence of constitutively active CAMKII increases SHP-2 association with the TCR and inhibits positive selection.
It is interesting to note that altered CD3 and the lack of ZAP70 recruitment, as well as hyporesponsiveness in IL-2 secretion (41), were also observed in anergic T cells (42). Moreover, a block in inducible IL-2 reporter gene activity was initiated by transfection of CaMKII* in Jurkat T cells (43). Our data suggest the possibility that CaMKII could control IL-2 secretion upstream and downstream of Ca2 signaling by acting on TCR signaling.
We previously reported that, in the presence of the CaMKIIB* transgene, DP thymocytes have a longer life span in the thymus. From the data presented here, we conclude that this is due to the defect in positive selection. Thymocytes remain at the DP stage until they receive a positive selection signal; thus, in a nonselecting background, the DP thymocytes accumulate until they die by neglect (44, 45). Likewise, the CaMKIIB* transgene inhibits positive selection and results in an accumulation of DP thymocytes.
It seems unlikely that the previously unsuspected role of CaMKII in a negative-feedback loop, as described here, would be unique to the TCR, given that CaMKII is ubiquitously expressed. CaMKII is highly concentrated at the CNS synapse (46). Moreover, it associates with the N-methyl-D-asparate (NMDA) receptor and is necessary for normal synaptic plasticity in pyramidal neurons (47). Binding of CaMKII to the NMDA receptor is potentiated by autophosphorylation at Thr286 (48) and may be part of a mechanism modulating the signal strength. Our data suggest that CaMKII modulates the strength of TCR signaling, while regulating a negative feedback loop that could be a more common phenomenon than is currently recognized.
Acknowledgments
We thank Camil Sayegh for the 166 cell line, and Irene Chen for critical reading of the manuscript.
Disclosures
The authors have no financial conflict of interest.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grants AI21372-21 (to S.M.H.)
2 Address correspondence and reprint requests to Dr. Sébastien Calbo at the current address: Institut National de la Santé et de la Recherche Médicale, Unité 519, Faculté de Médecine, 22 boulevard Gambetta, 76183 Rouen Cedex, France. E-mail address: scalbo{at}biomail.ucsd.edu
3 Abbreviations used in this paper: DN, double negative; DP, double positive; SP, single positive; SLP-76, SH2 domain-containing leukocyte protein of 76 kDa; LAT, linker for activation of T cells; PLC, phospholipase C; CaMKII, Ca2/calmodulin-dependent protein kinase II; CaMKIIB*, Ca2-independent mutant of CaMKIIB; SH, Src homology; SHP, SH2 domain-containing phosphatase; WT, wild type; PTPase, phosphatase; PP2A, protein phosphatase 2A; PTPN-2, protein tyrosine phosphatase, nonreceptor type 2; NMDA, N-methyl-D-asparate.
Received for publication January 25, 2005. Accepted for publication April 21, 2005.
References
Starr, T. K., S. C. Jameson, K. A. Hogquist. 2003. Positive and negative selection of T cells. Annu. Rev. Immunol. 21: 139-176.
van Oers, N. S., N. Killeen, A. Weiss. 1996. Lck regulates the tyrosine phosphorylation of the T cell receptor subunits and ZAP-70 in murine thymocytes. J. Exp. Med. 183: 1053-1062.
Chan, A. C., M. Dalton, R. Johnson, G. H. Kong, T. Wang, R. Thoma, T. Kurosaki. 1995. Activation of ZAP-70 kinase activity by phosphorylation of tyrosine 493 is required for lymphocyte antigen receptor function. EMBO J. 14: 2499-2508.
Zhang, W., J. Sloan-Lancaster, J. Kitchen, R. P. Trible, L. E. Samelson. 1998. LAT: the ZAP-70 tyrosine kinase substrate that links T cell receptor to cellular activation. Cell 92: 83-92.
Wardenburg, J. B., C. Fu, J. K. Jackman, H. Flotow, S. E. Wilkinson, D. H. Williams, R. Johnson, G. Kong, A. C. Chan, P. R. Findell. 1996. Phosphorylation of SLP-76 by the ZAP-70 protein-tyrosine kinase is required for T-cell receptor function. J. Biol. Chem. 271: 19641-19644.
Pivniouk, V. I., R. S. Geha. 2000. The role of SLP-76 and LAT in lymphocyte development. Curr. Opin. Immunol. 12: 173-178.
Crabtree, G. R., N. A. Clipstone. 1994. Signal transmission between the plasma membrane and nucleus of T lymphocytes. Annu. Rev. Biochem. 63: 1045-1083.
Dolmetsch, R. E., K. Xu, R. S. Lewis. 1998. Calcium oscillations increase the efficiency and specificity of gene expression. Nature 392: 933-936.
Freedman, B. D., Q. H. Liu, S. Somersan, M. I. Kotlikoff, J. A. Punt. 1999. Receptor avidity and costimulation specify the intracellular Ca2 signaling pattern in CD4CD8 thymocytes. J. Exp. Med. 190: 943-952.
Tobimatsu, T., H. Fujisawa. 1989. Tissue-specific expression of four types of rat calmodulin-dependent protein kinase II mRNAs. J. Biol. Chem. 264: 17907-17912.(Maureen A. McGargill*, Le)