In T lymphocytes, the CD2 and CD5 glycoproteins are believed to be involved in the regulation of signals elicited by the TCR/CD3 complex. Here we show that CD2 and CD3 independently associate with CD5 in human PBMC and Jurkat cells. CD5 coprecipitates with CD2 in CD3-deficient cells and, conversely, coprecipitates with CD3 in cells devoid of CD2. In unstimulated CD2+ CD3+ Jurkat cells, CD5 associates equivalently with CD2 and CD3 and is as efficiently phosphorylated in CD2 as in CD3 immune complexes. However, upon activation the involvement of CD5 is the opposite in the CD2 and CD3 pathways. CD5 becomes rapidly tyrosine phosphorylated after CD3 stimulation, but is dephosphorylated upon CD2 cross-linking. These opposing effects correlate with the decrease in the activity of the SH2 domain-containing protein phosphatase 1 (SHP-1) following CD3 activation vs an enhanced activity of the phosphatase after CD2 triggering. The failure of CD5 to become phosphorylated on tyrosine residues in the CD2 pathway has no parallel with the lack of use of ζ-chains in CD2 signaling; contrasting with comparable levels of association of CD2 or CD3 with CD5, ζ associates with CD2 only residually and is nevertheless slightly phosphorylated after CD2 stimulation. The modulation of CD5 phosphorylation may thus represent a level of regulation controlled by CD2 in signal transduction mechanisms in human T lymphocytes.
Major histocompatibility complex/peptide recognition by the appropriate TCR is central to the process of T lymphocyte activation. The simultaneous binding of TCR/CD3 and the coreceptors CD4 and CD8 to the same MHC/peptide complexes present on APC induces the approximation of the CD4/CD8-associated tyrosine kinase Lck to the CD3 chains. This results in the phosphorylation of the two conserved tyrosine residues within immune receptor tyrosine-based activation motifs (ITAM)3 contained in CD3 chains and TCR-ζ, and recruits through SH2 domains the ζ-associated protein ZAP-70, a kinase of the Syk family (1, 2, 3, 4). Concomitantly with the binding, ZAP-70 becomes phosphorylated on tyrosine residues, and newly formed phosphotyrosine residues in ZAP-70 then become docking sites for other SH2-containing enzymes (2).
CD2 is a 45- to 58-kDa type I integral protein expressed in human T and NK cells (5). Binding of CD2 on T cells to its counter-receptor CD58 contributes not only to the stabilization of interactions between lymphocytes and APC, but also to the transduction of activation signals, as CD58 in combination with CD2 mAbs can induce T cell activation and proliferation (6, 7). Signaling through CD2 depends on the integrity of its 116-aa-long cytoplasmic tail (8, 9), which is responsible for the association with the tyrosine kinases Lck and Fyn through proline sequence-SH3 domain interactions (10, 11, 12). Indeed, activation of CD2 with mAb, like that of TCR/CD3, has been shown to induce the activation of Lck (13). However, it was reported that signal transduction via CD2 fails to phosphorylate ζ-chains and consequently does not use ZAP-70 (14) despite the fact that many features of the CD2 pathway are similar to those of the pathway originated by the TCR (15, 16, 17).
In T lymphocytes, CD2 is embodied in a loosely associated membrane complex that additionally comprises the TCR/CD3 chains, CD4 or CD8, Lck and Fyn, and CD5, a membrane Ag expressed mainly on T cells (18). CD5 is a 67-kDa type I transmembrane glycoprotein whose cytoplasmic domain contains multiple potential sites for the phosphorylation of threonine, serine, and tyrosine residues (5). CD5 is rapidly phosphorylated after stimulation of the TCR/CD3 complex (19, 20), and this may allow Lck to bind through its SH2 domain and, as a result, to increase its catalytic activity (21). This can have some consequences for the phosphorylation of Lck substrates such as the CD3 chains, because CD5 is closely associated with the TCR/CD3 complex (22), and, in fact, tyrosine phosphorylation of CD5 precedes that of ζ-chains (19). Given the failure of the CD2 pathway to progress through the ITAMs present on CD3 chains, and as CD2 constitutively associates with Lck, which is an effector of CD5, we investigated the role of CD5 in signal transduction via CD2.
Materials and Methods
Human PBMC were obtained from buffy coats of normal healthy donors after centrifugation over Lymphoprep (Nycomed, Oslo, Norway). The Jurkat E6.1 and JRT3-T3.5 cell lines (23) were obtained from A. Weiss (University of California, San Francisco, CA). The Jurkat CD2+ CD3+ JKHM cell line was donated by D. A. Cantrell (Imperial Cancer Research Fund, London, U.K.). Cell lines were maintained in RPMI with 10% FCS, 1 mM sodium pyruvate, 2 mM l-glutamine, penicillin G (50 U/ml), and streptomycin (50 μg/ml).
mAbs against cell surface Ags were: CD2-RFT11 (24), given by G. Jánossy (Royal Free Hospital, London, U.K.), OKT11 (25), obtained from European Cell Culture Collection (ECACC, Porton Down, U.K.), GT2 (26), donated by D. A. Cantrell, and CD2–300 (27) (a polyclonal Ab recognizing the C-terminal-conserved end of CD2); CD3-OKT3 (28), and anti-CD3 polyclonal (29), gifts from M. H. Brown (Medical Research Council Cellular Immunology Unit); CD4-OKT4 (28), obtained from ECACC; CD5-Y2/178 (11), and a polyclonal anti-CD5 raised against a peptide sequence of 451–471 aa of human CD5 (18), gifts from D. Y. Mason (John Radcliffe Hospital, University of Oxford, Oxford, U.K); CD45-BMAC-1 (30), donated by J. Fabre (Institute of Child Health, University of London, London, U.K.); C3bi-OX21 (31); polyclonal anti-Lck, raised against a peptide consisting of 39–64 aa of murine Lck (18), a gift from J. Borst (The Netherlands Cancer Institute, Amsterdam, The Netherlands); anti-Fyn rabbit polyclonal Ab, donated by P. Burn (Hoffmann-LaRoche, Basel, Switzerland); polyclonal anti-CD3 ζ (32), a gift from D. A. Cantrell; anti-protein tyrosine phosphatase 1C, a polyclonal Ab recognizing 576–595 aa at the C terminus, from Santa Cruz Biotechnology (Santa Cruz, CA); anti-phosphotyrosine PY-20, purchased from Transduction Laboratories (Lexington, KY); goat anti-mouse peroxidase conjugate, from Transduction Laboratories; rabbit anti-mouse Ig (RAM), from Serotec (Kidlington, U.K.); and RAM conjugated with fluorescein (RAM-FITC), donated by S. Simmonds (Medical Research Council Cellular Immunology Unit).
Between 1–5 × 106 cells were resuspended in 50 μl of PBS containing 0.25% (w/v) BSA (PBS/BSA) and incubated with mAb (50 μl of hybridoma tissue culture supernatant) for 30 min on ice. Cells were washed twice at 4°C with 1 ml of PBS/BSA and 10 mM NaN3 (PBS/BSA/NaN3) and incubated with 50 μl of RAM-FITC (10 μg/ml) for 30 min on ice. Cells were then resuspended in 1 ml of PBS/BSA/NaN3 and analyzed on a FACScan (Becton Dickinson, Mountain View, CA).
Cell surface biotinylation
Cell surface biotinylation was performed as previously described (33). Briefly, cells were washed three times with ice-cold PBS and incubated for 10 min at room temperature with PBS containing 0.5 mg/ml of EZ-Link sulfo-NHS-LC-biotin (Pierce, Rockford, IL). Cells were then washed for an additional three rounds with PBS, divided into aliquots of 3.5 × 107 cells, and lysed for 30 min in ice-cold 1% Brij 96 lysis buffer (10 mM Tris-Cl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1 mM PMSF, and 1% (v/v) Brij 96 or Nonidet P-40).
Aliquots of 3.5 × 107 cells were lysed for 30 min on ice in lysis buffer (10 mM Tris-Cl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1 mg/ml BSA (not used in cell surface biotinylation), 1 mM PMSF, and 1% (v/v) Brij 96 or Nonidet P-40), the nuclear pellet was removed by centrifugation at 12,000 × g for 10 min at 4°C, and the supernatants were precleared by end-over-end rotation with protein A-Sepharose CL-4B (Pharmacia, Aylesbury, U.K.) for 30 min at 4°C. Abs (10 μg) or antisera (1–3 μl) and 100 μl of 10% protein A-Sepharose beads were added to the samples and rotated for 90 min at 4°C. The beads containing the immune complexes were washed three times in 1 ml of lysis buffer, and in in vitro kinase assays, an additional two washes were performed with 1 ml of Brij 96 assay buffer (25 mM HEPES (pH 7.5) and 0.1% (v/v) Brij 96). All washes were performed at 4°C.
For reprecipitation experiments, the beads containing the immune complexes were boiled for 5 min in 3% SDS and diluted 8-fold with lysis buffer. The beads were spun, and the supernatants were recovered and precleared for 30 min with 100 μl of 10% protein A-Sepharose beads. Proteins were precipitated with antisera plus 100 μl of 10% protein A-Sepharose beads for 90 min. Immunoprecipitates were washed three times with 1 ml of lysis buffer. Samples were boiled for 5 min in SDS buffer and run on 11% SDS-PAGE.
Detection of biotinylated cell surface Ags in precipitated immune complexes
Samples containing immunoprecipitates from surface-biotinylated cells were run on 11% SDS-PAGE and transferred to Hybond-C-super membranes (Amersham). Membranes were blocked overnight in Tris-buffered saline and 0.1% (v/v) Tween 20 (TBS-T) containing 5% (w/v) nonfat dried milk, washed once for 15 min and twice for 5 min each time with TBS-T, and incubated for 1 h at room temperature with ExtrAvidin peroxidase-conjugated (Sigma, Madrid, Spain; dilution, 1/7500 in TBS-T). Membranes were washed again for 15 min and an additional four times for 5 min each time with TBS-T, and biotinylated proteins were visualized by enhanced chemiluminescence (Amersham) and exposure to Biomax MR-1 films (Eastman Kodak, Rochester, NY).
Immune complex kinase assays
Brij 96 assay buffer (30 μl) containing 10 mM MnCl2, 1 mM Na3V04, 1 mM NaF, and 50 μCi (185 KBq) of [γ-32P]ATP was added to the beads containing the immune complexes, and in vitro kinase reactions were allowed to occur for 15 min at 25°C. Reactions were stopped by the addition of 30 μl of 2× SDS buffer, after which the samples were boiled for 5 min. Products were separated on 11% SDS-PAGE gels, and autoradiography of the dried gels was performed with Kodak X-OMAT S films (Eastman Kodak).
Approximately 3 × 107 Jurkat cells were used per sample. mAbs at 5 μg/ml and rabbit anti-mouse Ig at 30 μg/ml, or PHA at 10 μg/ml, were added to cells maintained at 37°C in RPMI (no FCS) and mixed by vortexing. After the times indicated, cells were briefly pelleted and lysed in lysis buffer, and the nuclear pellet was removed by centrifugation at 12,000 × g for 10 min at 4°C.
Cell lysates were denatured in 2× SDS buffer and run on SDS-PAGE. Samples were transferred to Hybond-C-super membranes by electroblotting. Membranes were blocked overnight in TBS-T containing 5% (w/v) nonfat dried milk, washed once for 15 min and twice for 5 min each time with TBS-T, and incubated for 1 h at room temperature with the primary Ab (1/5000 dilution). Membranes were washed again for 15 min and twice for 5 min with TBS-T, and incubated with goat anti-mouse or goat anti-rabbit Ig conjugated with peroxidase (1/20,000 dilution) for 1 h at room temperature. Membranes were washed again for 15 min and an additional four times for 5 min each time with TBS-T, and detection was accomplished using enhanced chemiluminescence (Amersham) and exposure to Biomax MR-1 films.
Measurement of [Ca2+]i
[Ca2+]i was determined as previously described (34). Briefly, cells were washed twice in RPMI containing 0.25% BSA and were resuspended at 2 × 107 cells/ml. Cells were incubated at 37°C for 10 min in the dark with 2 μM fura-2/AM and washed twice at room temperature in HBSS containing 1 mM CaCl2, 0.25% BSA, and 10 mM HEPES (pH 7.4). Before fluorometry, cells were diluted to 2 × 106 cells/ml, allowed to equilibrate to 37°C, and stimulated with mAbs at 2 μg/ml and with RAM at 25 μg/ml or PHA at 10 μg/ml.
Phosphatase assays were performed as described previously (35). Immune complexes precipitated by biotinylated anti-SHP-1 Abs and ImmunoPure avidin beads (Pierce) were washed three times in Nonidet P-40 lysis buffer and incubated for 3–4 h at 37°C in 25 mM HEPES (pH 7.5), 100 mM KCl, 3 mg/ml DTT, and 1 mg/ml p-nitrophenylphosphate (Sigma). Absorbance was measured at 413 nm.
Independent associations of CD2 and CD3 with CD5 in human T cells
Human PBMC were separated from whole blood, and surface labeled with biotin. Following cell lysis using the nonionic detergents Brij 96 and Nonidet P-40, immunoprecipitations of CD2, CD3, CD5, and CD45 were conducted. Immune complexes were separated by SDS-PAGE, and biotinylated proteins were visualized by enhanced chemiluminescence. When cells were lysed with Brij 96, CD5 was coprecipitated with CD3, as previously reported (22), as well as with CD2, as shown in Fig. 1,A. A protein of 67 kDa is clearly visible in CD2 and CD3 immune complexes from primary precipitates, and this protein was confirmed to be CD5 by reprecipitation using an anti-CD5 polyclonal serum (Fig. 1 B). However, when cells were lysed in lysis buffer containing 1% Nonidet P-40, CD5 was detected in CD2, but not in CD3, immunoprecipitates, indicating that the interaction between CD2 and CD5 in normal human T lymphocytes, although of an apparently lower stoichiometry, is stronger than the interaction between CD3 and CD5. These results also indicate that the interaction between CD2 and CD5 in normal human T cells can be independent of any contribution from CD3.
CD5 associates independently with CD2 and CD3, and is efficiently phosphorylated by kinases present in immune complexes of CD3 and of CD2 in Jurkat cells
In the Jurkat cell line JRT3-T3.5, which is negative for the expression of CD3 (Fig. 2), immunoprecipitation of CD2 again coprecipitated the CD5 Ag (Fig. 3,A). Conversely, the association between CD5 and CD3 did not require the presence of CD2 at the cell surface. In Jurkat cells devoid of CD2 (selected clone from E6.1), CD3 could precipitate CD5 (Fig. 3 B).
Not only could CD5 be coprecipitated with CD2 and CD3 independently, but it also could be phosphorylated by tyrosine kinases within the different immune complexes. We performed kinase assays on CD2, CD3, CD4, and CD5 immunoprecipitates from Jurkat cells not expressing CD3 complexes at the surface (Fig. 4,A) and on immunoprecipitates from CD2− Jurkat cells (Fig. 4,B). CD5 was present on CD2 immune complexes from the CD3-negative cell line and coprecipitated with CD3 in CD2− Jurkat cells, as the results from primary immunoprecipitations suggested and reprecipitations confirmed (Fig. 4). Despite the fact that CD3 Abs were able to precipitate some kinase activity and phosphoproteins from CD3− cells, possibly from endogenously produced CD3, the amount of CD5 coprecipitated was negligible compared with that of CD5 in CD2 immunoprecipitates, suggesting that CD3 has no role in CD2-mediated CD5 phosphorylation. Therefore, it seemed that both CD2 and CD3 have the potential to specifically associate with both CD5 and kinases that phosphorylate it.
Divergent patterns of CD5 phosphorylation following CD2 and CD3 cross-linking
We next wanted to study the possible role of CD5 in signal transduction via the CD3 and the CD2 pathways. For that purpose, we used a Jurkat cell variant, JKHM, that expresses both CD2 and CD3 at high levels. Through surface biotinylation and immunoprecipitation we determined that in these cells equivalent amounts of CD5 are precipitated with CD2 and CD3 (Fig. 5, A and B), and by in vitro kinase assays that CD5 is effectively phosphorylated by CD2 and CD3-associated kinases. Moreover, CD2 and CD3 associations with CD5 are independent of each other, as depletion of CD3 from cell lysates does not abrogate association between CD2 and CD5; conversely, preclearing of CD2 from cell lysates does not influence the level of CD5 coprecipitated with CD3 (Fig. 6).
The cells were functional in signaling, as calcium fluxes were generated following different stimulations with PHA and anti-CD3 and anti-CD2 Abs. Cross-linking CD2 using the RFT11 Ab gave the best signal, comparable to that of CD3, and significantly higher than the mitogenic combination of anti-CD2 monoclonals OKT11 and GT2 (Fig. 7). Using OKT11 cross-linking alone resulted in low calcium fluxes. Also, stimulation of CD2 was comparable to that emerging from the TCR/CD3 complex in inducing the phosphorylation of tyrosine residues in a number of substrates (data not shown). However, when we investigated the phosphorylation status of CD5 following different stimulations, we were unable to detect phosphorylation following cross-linking of CD2 (Fig. 8). Time-course experiments showed a rapid increase in the phosphorylation of CD5 following CD3 cross-linking, followed by a decline in the signal. However, none of the combinations of anti-CD2 Abs could produce a similar pattern. On the contrary, there seemed to be a rapid dephosphorylation of CD5 following CD2 stimulation, which was faster when using the Ab combinations that induced the highest calcium signals.
Differential usage of CD5 phosphorylation cannot be explained by differences in stoichiometry of CD5-CD3 association vs CD5-CD2 association
A number of features that are different between the CD2 and CD3 signaling pathways have been reported (14, 36, 37), one of the most significant being the report that stimulation of T cells via CD2 fails to induce the phosphorylation of CD3 ζ-chains and consequent docking of ZAP-70 to the CD3 complex (14). This may simply be due to the low level of association between CD2 and ζ. We performed kinase assays on immunoprecipitates of CD2 and CD3, following which ζ, CD5, and Lck were reprecipitated from the primary complexes. As displayed in Fig. 9, the amount of ζ associating with CD2 was just a tiny fraction of that associating with CD3. Following CD2 stimulation, the phosphorylation of ζ was perceptible (Fig. 10), but was so low that it may not have a physiological meaning compared with TCR stimulation. As expected, Lck was increasingly phosphorylated. By contrast, the amount of CD5 coprecipitating with CD2 was comparable to that coprecipitated by CD3 (Fig. 9), so it is striking that following CD2 triggering, CD5, contrary to ζ, became dephosphorylated (Fig. 10). The changes observed in the phosphorylation pattern of CD2 complexes following CD2 cross-linking reflect mainly changes in the phosphorylation level and not in the stoichiometry of associations, as CD2, Lck, and CD5 were present at similar amounts in both activated and nonactivated states (Fig. 10, lower panel). However, it seems that more ζ is associated with CD2 in activated cells, so it is possible that ζ is recruited to the CD2 complex following activation. Nevertheless, the amount of protein associated with CD2 was so low and difficult to detect by immunoblotting that the result could not be considered conclusive.
Changes in the activity of SHP-1 following activation via CD2 and CD3
It has been argued that CD5 can mediate negative or modulatory signals, possibly through its association with the protein tyrosine phosphatase 1C/SHP-1 (38, 39). Therefore, we measured the activity of that phosphatase following activation of Jurkat cells via CD3 or CD2. Cells were stimulated with OKT3 or RFT11 Abs, SHP-1 was specifically precipitated with biotinylated Abs and streptavidin beads, and the phosphatase activity of the immunoprecipitates was determined using a synthetic substrate, p-nitrophenylphosphate. Time-course experiments showed that there were consistently increments in the activity of SHP-1 following CD2 stimulation and a decline in the activity of the enzyme following TCR/CD3 triggering (Fig. 11). We therefore investigated whether SHP-1 could be found in association with CD2, which could explain why cross-linking of CD2 resulted in the enhancement of SHP-1 activity. Through in vitro kinase assays we were able to detect SHP-1 in immunoprecipitates of CD5 and CD3. However, we failed to detect any SHP-1 in CD2 immunoprecipitations (Fig. 12).
Some biochemical events following CD2-mediated activation appear to be very similar to those initiated by CD3, including calcium mobilization, activation of the tyrosine kinase Lck, appearance of a similar pattern of phosphopeptides after cross-linking either TCR/CD3 or CD2, and activation of the kinase of the Tec family, Itk (13, 15, 16, 17). It was therefore commonly accepted that CD2 would transduce signals of the same nature as those initiated by stimulation of the TCR/CD3 complex. However, recently it has been shown that features as central to CD3-mediated signaling as utilization of CD3 ζ-chains and ZAP-70 do not seem to be effectively used in signal transduction via CD2 (14), thus suggesting that CD2 and CD3 pathways may diverge at that level.
Although we have detected some phosphorylation of ζ upon CD2 stimulation, the fact remains that the level of association between CD2 and CD3-ζ in unstimulated cells is minimal, and increases in the phosphorylation of ζ-chains after CD2 cross-linking are so minute that this phosphorylation may be physiologically irrelevant compared with ζ phosphorylation after TCR/CD3 triggering. Therefore, we initially considered the hypothesis that activation via CD2, alternatively to using the ITAMs on CD3 chains, could proceed through the phosphorylation of CD5. Supporting our initial assumption was the finding that CD2 associates with CD5 in human T cells and cell lines independently of the TCR/CD3 complex, which was previously reported to closely interact with CD5 (22). Moreover, the CD5 fraction associating with CD2 had the potential to be phosphorylated by kinases contained in the CD2 immune complexes, again independently of the contribution of any kinase associated with the TCR/CD3 complex.
Interestingly, however, upon stimulation of T cells via CD2 we observed not an increase but, rather, a decline in the phosphorylation status of CD5, which was faster as the signal emerging from CD2 became stronger. This result was striking, as in contrast with the difference in the level of association between CD2 and ζ vs CD3 and ζ, discussed above, CD5 seemed to associate with both CD2 and CD3 at comparable levels.
Previous studies have also reported the absence of phosphorylation of CD5 or any other protein of similar molecular mass following CD2 stimulation (14, 40, 41, 42). The lack of phosphorylation of CD5 following CD2 triggering has a parallel in the cross-linking of CD5 alone, which, in contrast to TCR stimulation, does not induce phosphorylation of CD5 on tyrosine residues, although it is functional in other signaling pathways (43, 44). Therefore, it seems that CD5 may have a different role in signal transduction when coupled to the CD3 pathway or in its absence.
We show that CD2 is constitutively associated with CD5, and this association can be detected even under strong lysis conditions. Therefore, the lack of CD5 phosphorylation after CD2 triggering may have an unforeseen functional significance, possibly not reflecting only the lack of involvement in the CD2 pathway, but, instead, a key regulatory event. In this context, it is important to note that CD2 may functionally associate with CD5 in restraining the physiological activation through the TCR. In a mouse model where T cells express specific MHC class I-restricted TCRs, the absence of CD2 results in enhanced positive selection (45). A similar phenotype is observed in CD5 null MHC I-restricted TCR transgenic mice, which suggests that both CD2 and CD5 contribute to the modulation of signals during thymic selection. Furthermore, in mice deficient in both CD2 and CD5, the effect seems to be synergistic (45).
The negative role of CD5 is possibly due to its functional association with the tyrosine phosphatase SHP-1, as the absence of CD5 as well as of SHP-1 results in hyper-responsiveness upon TCR stimulation, and also in increased positive selection of thymocytes (38, 39). We had previously detected tyrosine phosphatase activity associated with CD5 in rat T cells (35), and in the present report a correlation between the phosphatase activity of SHP-1 following CD2 and CD3 activation and the status of phosphorylation of CD5 after the different stimuli was found. Although it was not proven that SHP-1 was the sole phosphatase responsible for the dephosphorylation of CD5, these results strongly suggest a differential role for this phosphatase following CD2 and CD3 stimulation. Moreover, the activity of the phosphatase seems to be specific for phosphorylated CD5, as when cells were stimulated via CD2 we could detect massive phosphorylation of Lck and, despite the low level of association, some degree of phosphorylation of ζ-chains, contrary to previously reported. By contrast, CD5 became dephosphorylated in the same complex.
It has been recently reported that SHP-1 constitutively associates with CD5, and the level of association increases following TCR engagement. The CD5 cytoplasmic membrane-proximal tyrosine residue, when phosphorylated, is a docking site for the SH2 domain of SHP-1 (46). We have not investigated whether SHP-1 dissociates from CD5 following CD2 cross-linking, but the overall activity of the phosphatase seems to increase. As CD2 does not associate with SHP-1, it is possible that CD2-associated Lck or other enzymes may regulate the activity of SHP-1, either by activating the phosphatase directly upon CD2 activation, thus explaining the observed dephosphorylation of CD5 upon CD2 triggering, or by maintaining a residual, but sustained, level of phosphorylation of CD5 and thus contributing to the constitutive binding of the SHP-1 to CD5.
Two levels of modulation of the TCR/CD3 signal are therefore considered. Firstly, TCR engagement induces the phosphorylation of CD5 in the SHP-1 binding tyrosine residue, thus recruiting SHP-1 to the membrane, where it controls the level of phosphorylation of the complex. Secondly, coactivation of CD5-associated CD2 may enhance the activity of SHP-1, thus modulating the overall phosphorylation status of the activation complex. This model could explain why the lack of CD2 and/or CD5 results in the increased reactivity of TCRs in the animal models discussed above (45).
Alternatively, it is possible that cross-linking of CD2 results in only the partial phosphorylation of the ITAMs present on CD3 chains. It has been suggested that the full positive signal involving the coupling of ZAP-70 to ITAMs requires both tyrosine residues to be phosphorylated. If only one of the residues becomes phosphorylated, the resulting signal may be inhibitory (P. Allen, unpublished observation). As SHP-1 can be coprecipitated with CD3, it may be that incomplete ITAM phosphorylation induced by CD2 could result in the recruitment of SHP-1, and not ZAP-70, to the TCR/CD3 complex. Following TCR-positive stimulation with complete ITAM phosphorylation, full occupancy of CD3 ITAMs by ZAP-70 could displace SHP-1 from the activation motifs, thus explaining the decrease in the activity of the phosphatase.
The present results together with the recent findings of the regulatory role of CD2 and CD5 in signal transduction and the parallelism in CD2 and CD5 expression observed during T cell ontogeny and after polyclonal activation (45, 47) support a functional role for the CD2/CD5 association described here in the regulation of signal transduction in T lymphocytes.
We thank Dr. S. P. Watson, Oxford University, for help with the measurement of [Ca2+]i, and Dr. M. H. Brown, Dr. D. Mason (Medical Research Council Cellular Immunology Unit, Sir William Dunn School of Pathology, University of Oxford, Oxford, U.K.), and Prof. M. de Sousa (Instituto de Biologia Molecular e Celular, Universidade do Porto, Porto, Portugal) for useful discussions and continuous support.
This work was supported by fellowships from the Fundação para a Ciência e a Tecnologia (to A.M.C. and M.A.A.C.) and a fellowship from the American Portuguese Biomedical Research Fund (to F.A.A.).
Abbreviations used in this paper: ITAM, immune receptor tyrosine-based activation motif; RAM, rabbit anti-mouse; [Ca2+]i, intracellular calcium concentration; SHP-1, SH2 domain-containing protein phosphatase 1.