Article

Generation of Cd3e^PRS/PRS mice

To address the importance of the nonredundant CD3 PRS in vivo without changing the position of the CD3 ITAM sequence relative to the plane of the plasma membrane in which CD3 and TCR are expressed, we replaced the RPPPVPNP sequence with the ASREKADA sequence that occupies an analogous position in the cytoplasmic tail of FcRI, another ITAM-containing subunit used by immunoreceptors (Supplementary Fig. 1 online). We derived 'knock-in' mice with the intended Cd3e^PRS mutation from Bruce-4 C57BL/6 embryonic stem cells; homozygous Cd3e^PRS/PRS mice were born at the expected frequencies and were healthy and fertile (Supplementary Fig. 2 online). Although contiguous to the CD3 ITAM, the substituted sequence did not affect ITAM tyrosine phosphorylation after TCR ligation; we confirmed this by studying the signaling properties of CD25-based chimeric molecules (Supplementary Fig. 3 online).

Analysis of Cd3e^PRS/PRS thymi showed they had a normal representation of DN, DP and SP subsets and a 30% decrease in cellularity (Fig. 1a and Supplementary Table 1 online). The frequency and absolute numbers of DN4 (CD44^- CD25^- ) cells were diminished in Cd3e^PRS/PRS thymi (Fig. 1b), which suggested that the Cd3e^PRS mutation resulted in less TCR selection. The expression of CD3 and TCR on the surface of Cd3e^PRS/PRS thymocytes was upregulated in a stage-specific way restricted to DP cells (Fig. 1c), and normal amounts of TCR-CD3 complexes were present on the surface of SP thymocytes (Fig. 1c). Cd3e^PRS/PRS DP thymocytes expressed normal amounts of CD4, CD8, CD69 and CD25 and slightly more CD5 and CD44 (Fig. 1c and data not shown). Spleen and lymph nodes of Cd3e^PRS/PRS mice had normal numbers of T cells (Fig. 1d,e, Supplementary Table 1 and data not shown); the variable -region repertoire used by mature T cells present in Cd3e^PRS/PRS mice showed no detectable alteration; and normal numbers of T cells expressing normal surface amounts of TCR were present (Supplementary Figs. 4 and 5 online). Therefore, the Cd3e^PRS mutation had a limited effect on the development of T cells expressing polyclonal TCR repertoire.

Figure 1: Higher TCR expression on DP thymocytes from Cd3e^PRS/PRS mice.

Flow cytometry of wild-type (WT) and Cd3e^PRS/PRS (PRS/PRS) thymocytes and splenocytes. (a) Expression of CD4 and CD8 on thymocytes. Numbers in outlined areas indicate percent DN cells (bottom left), DP cells (top right), CD4⁺ SP cells (top left) and CD8⁺ SP cells (bottom right). (b) Expression of CD44 and CD25 expression on purified TCR^- DN thymocytes. Numbers in outlined areas indicate percent DN2 plus DN3 cells (right) or DN4 cells (left). (c) Expression of TCR, CD3, CD5 and CD44 on DP, CD4⁺ SP and CD8⁺ SP thymocytes from wild-type mice (gray filled histograms) and Cd3e^PRS/PRS mice (solid lines). Dotted lines, isotype control staining. Numbers at top indicate geometric mean fluorescence of cells from wild-type mice (left) and Cd3e^PRS/PRS mice (right). (d) Expression of CD4 and CD8 on splenocytes. Numbers in outlined areas indicate percent CD4⁺ cells (top left) and CD8⁺ cells (bottom right). (e) Expression of TCR, CD3, CD5 and CD44 on CD4⁺ and CD8⁺ splenocytes from wild-type mice (gray filled histograms) and Cd3e^PRS/PRS mice (solid lines). Data are representative of at least three independent experiments.

Full size image (111 KB)

The Cd3e^PRS mutation abrogates Nck docking

To demonstrate that the ASREKADA sequence expressed by CD3^PRS subunits prevented their interaction with the Nck Src homology 3.1 (Nck-SH3.1) domain, we incubated thymocytes from wild-type and Cd3e^PRS/PRS mice in the presence or absence of antibodies to CD3. We then prepared cell lysates, precipitated proteins with beads coupled to fusion proteins of glutathione S-transferase and Nck-SH3.1 (GST–Nck-SH3.1) and assessed the presence of bound, fully assembled TCR-CD3 complexes by immunoblot analysis with antibody to CD3 (anti-CD3; Fig. 2a). In contrast to wild-type TCR complexes, those containing CD3^PRS subunits did not bind to the Nck-SH3.1 domain, which confirmed that the interaction between Nck-SH3.1 and CD3 required an intact PPVPNPDY segment^{23, 25}. Unexpectedly, we found interaction between wild-type TCR complexes and the Nck-SH3.1 domain in freshly isolated thymocytes, and stimulation with anti-CD3 resulted in a 1.2- to 1.5-fold gain over the amount in the basal state (Fig. 2a). To demonstrate that TCR-CD3 complexes expressed on the cell surface were detected by GST–Nck-SH3.1 beads, we biotinylated freshly isolated thymocytes and then precipitated proteins in total cell lysates with GST–Nck-SH3.1 beads. All the CD3 subunits of the TCR could be detected before and after stimulation with anti-CD3 (Fig. 2b and data not shown), which indicated that TCR complexes present on the surface of wild-type thymocytes contained CD3 PRS that constitutively bound Nck.

Figure 2: The Cd3e^PRS mutation abrogates Nck docking.

Analysis of wild-type and Cd3e^PRS/PRS thymocytes left untreated (UT) or incubated for 5 min (a,b) or for 10 or 30 min (c,d) at 37 °C with anti-CD3 (-CD3). (a) Precipitation of lysates with GST and GST–Nck-SH3.1 (SH3) beads and immunoblot (IB) analysis with anti-CD3 (-CD3) or anti--tubulin. kDa, kilodaltons. (b) Precipitation of lysates of unbiotinylated or surface-biotinylated wild-type thymocytes with GST or GST–Nck-SH3.1 (SH3) beads, followed by immunoblot analysis. IP (middle lane), immunoprecipitation of a lysate aliquot with anti-CD3 (shorter exposure time). Below, membranes striped and reprobed with anti-CD3 or anti--tubulin. (c) Flow cytometry of the expression of surface CD3 and intracellular APA1/1 staining in DP thymocytes. Dotted lines, isotype control. (d) Quantification of the data in c, presented as the geometric mean of CD3 surface expression or APA1/1 intracytoplasmic staining. Below, duration and temperature of incubation with anti-CD3. Data are representative of at least two independent experiments.

Full size image (76 KB)

To confirm the results reported above, we used the APA1/1 antibody^{26, 27} reported before to detect the CD3 PRS in T cells after TCR ligation. Freshly isolated, wild-type DP thymocytes showed unimodal APA1/1 staining identical to the staining in wild-type DP cells stimulated with anti-CD3 (Fig. 2c,d), which suggested that the accessibility of the CD3 PRS is not limited to the DP cells that express TCRs in the process of recognizing self-pMHC complexes²⁸. In addition, wild-type thymocytes cultured in suspension for 8 h at 37 °C to prevent interaction of TCRs with MHC molecules⁵ showed APA1/1 staining similar to that of freshly isolated wild-type thymocytes, and MHC class II–restricted DP cells in MHC class II–sufficient or MHC class II–deficient thymic environments showed similar APA1/1 staining (Supplementary Fig. 6 online). Thus, the CD3 PRS component of TCR complexes expressed on the surface of DP thymocytes is constitutively accessible to the Nck-SH3.1 domain independently of TCR-pMHC engagement.

TCR internalization and recycling in Cd3e^PRS/PRS cells

To investigate the mechanism of CD3 PRS–mediated TCR downregulation, we measured the kinetics of TCR cycling on DP thymocytes. First we found that Cd3e^PRS/PRS DP thymocytes internalized surface-bound anti-CD3 and anti-TCR with kinetics similar to those of wild-type thymocytes (Fig. 3a,b and data not shown). Next we incubated thymocytes with phycoerythrin-labeled anti-CD3 either on ice, to measure surface TCR-CD3 complexes, or at 37 °C, to measure the 'cycling pool' that corresponds to surface, internalized and recycled TCR complexes. Although the cycling pool of Cd3e^PRS/PRS DP cells was approximately 2.5-fold larger than that in wild-type DP cells, the cycling pool/surface pool ratio was similar in Cd3e^PRS/PRS DP thymocytes (2.4 0.2) and wild-type DP thymocytes (2.3 0.2; data not shown and Fig. 3c,d). Therefore, the lack of CD3 PRS resulted in larger surface TCR pools and intracellular cycling TCR pools in DP cells. Notably, the cell surface expression and cycling pools of TCR in CD4⁺ SP cells from wild-type and Cd3e^PRS/PRS thymi were similar (Fig. 3c,d).

Figure 3: TCR internalization and recycling in Cd3e^PRS/PRS DP thymocytes.

(a,b) Antibody-induced internalization of CD3 in wild-type (filled circles) and Cd3e^PRS/PRS (open circles) DP thymocytes. (a) Cell surface CD3, presented as the geometric mean of the fluorescence intensity. (b) Normalization of the data in a relative to CD3 expressed at time 0 (initial). (c) Flow cytometry of wild-type (gray filled histograms) and Cd3e^PRS/PRS (solid lines) DP and CD4⁺ SP thymocytes incubated with anti-CD3 for 60 min at 0 °C or at 37 °C to allow labeling of surface pools and cycling plus surface pools of TCR, respectively. Dotted lines, isotype control staining. (d) Quantification of the data in c, presented as the geometric mean (and s.e.m.) of CD3 expression relative to that in wild-type thymocytes. (e,f) Recycling of TCR-CD3 in wild-type (filled circles) and Cd3e^PRS/PRS (open circles) DP thymocytes incubated for 0–6 h in normal medium (dotted lines) or in hypertonic medium in the presence (dashed lines) or absence (solid lines) of cycloheximide. (e) CD3 expressed at the cell surface. (f) Normalization of the data in e relative to CD3 expressed at time 0 (initial). Data are representative of at least two independent experiments.

Full size image (71 KB)

By blocking TCR-CD3 internalization and protein synthesis with hypertonic medium and cycloheximide, respectively, the number of CD3 molecules present in the intracellular cycling pool can be determined¹². With this assay we found that wild-type and Cd3e^PRS/PRS DP thymocytes had similar kinetics of CD3 expression (Fig. 3e,f) and that more CD3 molecules recycled back to the surface in the absence of CD3 PRS, consistent with a larger intracellular cycling pool of CD3 (Fig. 3c,d). Notably, the larger TCR-CD3 cycling pool in Cd3e^PRS/PRS DP thymocytes did not result from more synthesis of TCR components. Therefore neither more neosynthesis nor a substantial alteration in constitutive CD3 internalization and recycling seems to account for the higher TCR-CD3 expression on the surface of Cd3e^PRS/PRS DP thymocytes.

Convergent functions of the CD3 PRS and of SLAP

Analysis of mice deficient in SLAP (Sla^{- /-} mice) has shown that this adaptor downregulates TCR expression on DP thymocytes^{3, 12, 13, 14, 15}. To determine whether SLAP and the CD3 PRS act together to decrease TCRs on the surface of DP cells, we analyzed DP thymocytes in conditions in which CD4–MHC class II interactions that 'trigger' the SLAP-based degradation mechanism were ablated. As described before^{4, 5, 6}, wild-type DP cells cultured at 37 °C in the absence of MHC class II–positive thymic stroma had a considerable increase in the expression of TCR at the cell surface (Fig. 4a); in contrast, Cd3e^PRS/PRS DP thymocytes showed only a modest increase in their already high expression of TCR. When we used the MHC class II–negative Tst-4–DL1 stromal cell line²⁹ to support in vitro T cell differentiation, we found no additive effect on TCR-CD3 expression on Cd3e^PRS/PRS DP cells (Fig. 4b and Supplementary Fig. 7 online). Reconstitution of recipient mice sufficient or deficient in MHC class II molecules with bone marrow from wild-type or Cd3e^PRS/PRS mice showed that the simultaneous disruption of CD4–MHC class II interactions and of the CD3 PRS had no additive effect on the amount of TCR expressed on Cd3e^PRS/PRS DP cells (Fig. 4c).

Figure 4: The MHC class II–CD4–Lck–SLAP pathway and CD3 PRS act together to downregulate TCR expression on DP cells.

(a) Flow cytometry of the surface expression of TCR on wild-type (gray filled histograms) or Cd3e^PRS/PRS (solid lines) DP thymocytes incubated for 12 h at 4 °C (left) or 37 °C (right). (b) TCR expression on freshly isolated wild-type (gray filled histograms) or Cd3e^PRS/PRS (solid lines) DP thymocytes (left) and on in vitro–produced DP thymocytes generated from wild-type (gray filled histograms) or Cd3e^PRS/PRS (solid lines) DN thymocytes cultured for 4 d with the MHC class II–negative Tst-4–DL-1 stromal cells (right). (c) Flow cytometry of cell surface expression of HY TCR on DP thymocytes from female Cd3e^5/5 mice sufficient (left) or deficient (right) in MHC class II molecules 10 weeks after reconstitution with bone marrow cells from wild-type (gray filled histograms) or Cd3e^PRS/PRS (solid lines) HY TCR-transgenic mice. Below plots (a–c), geometric mean of TCR normalized to those noted for DP thymocytes expressing wild-type CD3 subunits and developing in a MHC class II–sufficient environment (to facilitate comparison of TCR expressed in the various conditions). Dotted lines (a–c), isotype control staining. (d) Expression of TCR and CD3 subunits on the surface of DP cells (mouse genotypes, keys). Right two panels, Cd3e^PRS/PRS and Sla^{- /-} histograms overlaid on those corresponding to Cd3e^PRS/PRSSla^{- /-} and Cd3e^PRS/+Sla^{- /+} DP cells. Data are representative of two independent experiments.

Full size image (56 KB)

Next we directly compared the amount of TCR-CD3 expressed by thymocytes deprived of either CD3 PRS or SLAP. As reported before^{3, 13}, TCR and CD3 expression was approximately three- to fourfold higher on Sla^{- /-} DP cells than on wild-type DP cells, a result also noted on Cd3e^PRS/PRS DP thymocytes (Fig. 4d). DP cells deprived of both the CD3 PRS and SLAP had slightly more surface TCR than did DP cells deprived of either the CD3 PRS or SLAP. These data indicate that both the CD3 PRS and SLAP are required for efficient control of the amount of TCR expressed on DP cells, a conclusion supported by additional analyses of single- or compound-heterozygous mice that showed a distinct 'gene-dosage' effect (Fig. 4d).

Because SLAP regulates TCR expression on DP cells by targeting CD3 chains for degradation, we analyzed whether the CD3 PRS also participates in CD3 degradation. By evaluating total CD3 by immunoblot analysis and intracellular flow cytometry (Fig. 5a,b), we found that Cd3e^PRS/PRS DP thymocytes had approximately twofold more CD3 than did wild-type DP cells. In contrast, total amounts of TCR and CD3 chains were, if anything, only slightly greater in the absence of CD3 PRS (Supplementary Fig. 8 online). To evaluate directly the efficiency of CD3 degradation in the absence of CD3 PRS, we incubated total thymocytes with cycloheximide and monitored CD3 expression over time. In contrast to wild-type thymocytes, Cd3e^PRS/PRS and Sla^{- /-} thymocytes failed to degrade CD3 during the 8-hour assay (Fig. 5c,d). When we analyzed degradation of TCR and CD3 in parallel, we found no impairment in the absence of CD3 PRS or of SLAP (Fig. 5e). Therefore, both SLAP and the CD3 PRS are required for TCR downregulation through degradation of CD3.

Figure 5: The CD3 PRS and SLAP regulate CD3 degradation in DP thymocytes.

(a) Immunoblot (left) of wild-type and Cd3e^PRS/PRS whole-cell lysates (corresponding to 5.0 10⁶, 2.5 10⁶ and 1.3 10⁶ thymocytes; above lanes) with anti-CD3; -tubulin, loading control. Right, quantification of the immunoblot at left. (b) Flow cytometry of intracellular CD3 in DP and CD4⁺ T cells. Dotted lines, isotype control staining. Numbers in left plot indicate the geometric mean fluorescence of cells from wild-type mice (left) and Cd3e^PRS/PRS mice (right). (c) Immunoblot analysis of CD3 in thymocytes (genotype, above blot) incubated in the presence of cycloheximide (time, above lanes). -tubulin, loading control; AU, arbitrary luminescence units. (d) Quantification of the immunoblots in c by luminescence, presented as CD3 expression relative to expression at time 0 (initial). (e) Flow cytometry of intracellular expression of TCR and CD3 in DP thymocytes in the presence of cycloheximide, presented relative to expression at time 0 (initial). Data are representative of at least three independent experiments.

Full size image (59 KB)

Effect on thymic selection

The absence of SLAP improves the positive selection of DO11.10 TCR-transgenic thymocytes, an effect assigned to the larger amount of TCRs present on the DP thymocytes³. To analyze whether the Cd3e^PRS mutation had a similar effect on intrathymic selection, we backcrossed HY and Marilyn TCR-transgenic mice, models that permit an analysis of positive and negative selection among the same littermates^{30, 31}, with Cd3e^PRS/PRS mice. In the presence of CD3^PRS, we noted higher expression of the HY and Marilyn TCRs on DP cells, as detected with monoclonal antibody T3.70 (specific for the HY TCR) and by expression of V6, respectively (Fig. 6a,b and data not shown), results that recapitulated our earlier results with mice expressing a polyclonal TCR repertoire (Fig. 1c) and reinforced the view that the CD3 PRS regulates TCR in a stage-specific way. Despite the twofold more DP cells in HY-Cd3e^PRS/PRS female mice, their thymi contained 35% fewer CD8⁺ SP cells (Fig. 6a and Supplementary Table 2 online), consistent with less positive selection. We also noted much higher CD44 expression on all HY-Cd3e^PRS/PRS thymocytes (Fig. 6b), a finding similar to that obtained with Cd3e^PRS/PRS mice (Fig. 1c) and whose importance remains to be determined. Consistent with the lower positive selection in HY-Cd3e^PRS/PRS female thymi, T3.70⁺CD8⁺ splenocyte numbers were four- to sevenfold greater in spleens from female wild-type HY mice (2.9 10⁶ 0.8 10⁶; n = 4 mice) than in spleens from female HY-Cd3e^PRS/PRS mice (0.6 10⁶ 0.2 10⁶; n = 11 mice; Fig. 6c). Because of premature expression of the HY TCR transgene, most of the developing thymocytes in male HY mice are deleted before or at the transition to the DP stage³², resulting in much lower thymic cellularity and depletion of most T3.70⁺ DP and SP cells; in contrast, thymi from male HY-Cd3e^PRS/PRS mice had 1.3-fold more cellularity that resulted from the reappearance of T3.70⁺ DP cells and CD4^loCD8^lo cells (Supplementary Fig. 9 and Supplementary Table 2 online). Comparison of the expression of TCR-CD3 on the few DP cells in thymi from male HY-Cd3e^PRS/PRS and HY mice confirmed that the Cd3e^PRS mutation resulted in more TCR-CD3 on DP cells. Because none of the few T3.70⁺ DP cells in thymi from male HY-Cd3e^PRS/PRS mice transitioned to the T3.70^hiCD8^hi SP stage, HY-Cd3e^PRS/PRS male mice lacked peripheral T3.70^hiCD8^hi T cells (Supplementary Fig. 9). Therefore, in thymi from male HY-Cd3e^PRS/PRS mice, although negative selection of conventional T cells occurred, it did so with protracted kinetics. Analysis of the progeny of Marilyn female and male recombination-activating gene–deficient mice crossed with Cd3e^PRS/PRS mice showed they had normal negative selection, whereas there was less positive selection, which resulted in 50–70% fewer clonotype-positive CD4⁺ splenocytes (Supplementary Fig. 10 online). These data collectively show that elimination of the CD3 PRS had no net effect on negative selection, whereas positive selection was lower in both Marilyn and HY-Cd3e^PRS/PRS DP cells.

Figure 6: Impeded positive selection in Cd3e^PRS/PRS mice expressing the HY transgenic TCR.

(a–c) Flow cytometry of cells from female mice expressing the HY transgenic TCR on the wild-type and Cd3e^PRS/PRS background. (a) Expression of CD4 and CD8 on total thymocytes. Numbers in plots indicate percent cells in outlined windows. (b) Expression of T3.70, CD3, CD5 and CD44 on DP, DN and CD8⁺ SP thymocytes from wild-type mice (gray filled histograms) and Cd3e^PRS/PRS mice (solid lines). Dotted lines, isotype control staining. Numbers in top row indicate the geometric mean fluorescence of cells from wild-type mice (left) and Cd3e^PRS/PRS mice (right). (c) Expression of CD8 and T3.70 on total splenocytes. Numbers above outlined areas indicate percent T3.70^hiCD8^hi cells. (d) Flow cytometry of the expression of CD8 and T3.70 on total splenocytes from Cd3e^5/5 female mice sufficient (+/+) or deficient (/) in MHC class II molecules 10 weeks after reconstitution with bone marrow cells isolated from wild-type and Cd3e^PRS/PRS mice expressing the HY TCR. Numbers above outlined areas indicate percent T3.70⁺CD8⁺ T cells. Data are representative of at least three independent experiments.

Full size image (76 KB)

Finally, the mature T cells in the periphery of HY and Marilyn female mice expressing wild-type or mutant CD3 subunits had similar amounts of TCR, CD3, CD5, CD69, CD25 and CD44 (data not shown). Wild-type or CD4⁺ (Marilyn) T cells and CD8⁺ (HY) T cells expressing wild-type CD3 or CD3^PRS showed similar dose-response curves and identical proliferative responses (Supplementary Fig. 11 online and data not shown). Therefore, the CD3 PRS has no detectable function in antigen-driven activation of CD4⁺ and CD8⁺ T cells.

Divergent functions of the CD3 PRS and SLAP

The divergent effects of deficiency in SLAP or the CD3 PRS on positive selection may have resulted from different properties of the TCR-transgenic models (DO11.10 versus Marilyn and HY models). For example, the affinities in the Marilyn and HY TCRs might be near the threshold that distinguishes positive from negative selection and any increase in their density might convert positive selection into negative selection. It was thus important to test whether the CD3 PRS contributed to TCR signaling in DP cells regardless of its effect on TCR density. Given that the SLAP degradation pathway can be deactivated through elimination of MHC class II and that MHC class I–restricted HY T cells develop in the absence of MHC class II, we created bone marrow chimeras in which DP cells expressed HY TCR and wild-type CD3 in amounts similar to those found on Cd3e^PRS/PRS DP cells. We sublethally irradiated MHC class II–sufficient mice and MHC class II–deficient recipient mice and reconstituted then with wild-type HY or HY-Cd3e^PRS/PRS bone marrow cells. Consistent with the results obtained with HY mice (Fig. 6a,b), spleens from control chimeras reconstituted with wild-type HY bone marrow cells contained 3.5-fold more CD8⁺ T3.70⁺ cells than did MHC class II–sufficient hosts reconstituted with HY-Cd3e^PRS/PRS bone marrow cells (Fig. 6d and Supplementary Fig. 12a online). As expected, wild-type HY DP cells from MHC class II–deficient thymi expressed surface amounts of TCR similar to those noted on Cd3e^PRS/PRS DP cells from either MHC class II–sufficient or MHC class II–deficient thymi (Supplementary Fig. 12b). Consistent with our results reported above (Fig. 5), the higher TCR surface expression on wild-type HY DP cells in MHC class II–deficient thymi correlated with more total CD3, whereas TCR and TCR amounts were not modified much (Supplementary Fig. 12c). Regardless of the presence of more TCR at their surface, wild-type HY DP cells from MHC class II–deficient thymi generated mature T3.70⁺CD8⁺ T cells in numbers similar to those of HY DP cells from MHC class II–sufficient thymi (Fig. 6d and Supplementary Fig. 12a). These data demonstrate that increasing the expression of wild-type HY TCR complexes to that on HY-Cd3e^PRS/PRS DP cells had no negative effect on positive selection, which therefore supported the conclusion that the affinity of the HY TCR is near the threshold that distinguishes lack of selection from positive selection³³. The data also suggest that the CD3 PRS regulates TCR signaling in DP cells regardless of its effects on TCR expression.

Signaling defect in Cd3e^PRS/PRS DP cells

To confirm the data reported above, we measured changes in intracellular calcium in response to anti-CD3 by DP cells from Cd3e^PRS/PRS and Sla^{- /-} mice. We reasoned that because these cells express identical amounts of cell surface TCR (Fig. 7a), they should allow unambiguous conclusions to be drawn about the unique signaling function of the CD3 PRS. CD4 SP cells from Cd3e^PRS/PRS and Sla^{- /-} mice showed identical calcium responses to anti-CD3 cross-linked with streptavidin (Fig. 7b). In contrast, DP cells from Cd3e^PRS/PRS and Sla^{- /-} mice showed very different responses: most DP cells from Sla^{- /-} mice had more intracellular calcium after treatment with anti-CD3 alone, whereas Cd3e^PRS/PRS DP cells required further cross-linking with streptavidin (Fig. 7b). Compared with wild-type DP cells, the high expression of signaling-proficient TCRs on DP cells from Sla^{- /-} mice rendered them hyper-responsive to weak TCR stimuli; in contrast, despite similar high expression of TCR, Cd3e^PRS/PRS DP thymocytes were less responsive than were wild-type DP cells to weak TCR stimuli (Fig. 7b). These results are thus consistent with our hypothesis that CD3^PRS lowered the signaling potential of TCR expressed on DP cells in addition to its effect on the regulation of TCR expression.

Figure 7: The CD3 PRS enhances the responsiveness of DP thymocytes to weak TCR stimuli.

(a,b) Flow cytometry of cell surface TCR expression (a) and changes in intracellular calcium (b) for DP and CD4⁺ SP thymocytes (genotype, key). (b) Calcium fluxes on the whole cell populations (top and middle) and percent cells responding over a relative fluorescence threshold of 150 (bottom). The addition of ionomycin at 9 min ensured the cells were properly loaded with Indo-1. (c) Expression of CD4, CD8 and CD5 by thymocytes and splenocytes (genotype, left margin). Numbers in and adjacent to outlined areas indicate percent cells in specified windows. (d) Expression of TCR, CD3, CD24 and CD5 on DP and CD4⁺ SP thymocytes from Zap70^{- /-} mice (gray filled histograms), Zap70^{- /-} Sla^{- /-} mice (green lines) and Zap70^{- /-} Cd3e^PRS/PRS mice (blue lines). Data are representative of at least two independent experiments.

Full size image (91 KB)

In further experiments we used mice deficient in the tyrosine kinase Zap70 (in which T cell development is arrested at the DP stage) crossed with Cd3e^PRS/PRS and Sla^{- /-} mice. Consistent with a published report³, ablation of SLAP enabled Zap70-deficient DP cells to progress to the SP stage and to give rise to small numbers of mature CD4⁺ T cells in the periphery (Fig. 7c,d); presumably these effects were due to greater overall signaling from more signaling-competent TCRs. In contrast, the similarly considerable amount of TCR present on Cd3e^PRS/PRSZap70^{- /-} DP cells failed to compensate for the weaker TCR signals resulting from the lack of Zap70, and we found neither SP cells nor peripheral CD5⁺CD4⁺ T cells in Cd3e^PRS/PRSZap70^{- /-} mice (Fig. 7c,d). Therefore, although SLAP and the CD3 PRS acted together in the regulation of CD3 degradation to maintain low TCR expression on DP cells, only the CD3 PRS contributed to enhanced DP cell responsiveness to weak TCR ligands.

The CD3 PRS is required for CD3 phosphorylation

To determine how the CD3 PRS influences TCR responsiveness and SLAP-dependent CD3 degradation in preselection DP cells, we left Cd3e^PRS/PRS and Sla^{- /-} thymocytes unstimulated or stimulated them with anti-CD3 and then analyzed whole-cell lysates by immunoblot with antibody to phosphorylated tyrosine (anti-phosphotyrosine). After TCR cross-linking, Cd3e^PRS/PRS and Sla^{- /-} thymocytes had similar inducible phosphorylated species (Fig. 8a). However, in independent experiments we noted that anti-CD3-induced bands migrating at the position expected for c-Cbl, the adaptor Vav, Zap70 and the adaptor Lat were always stronger in Sla^{- /-} thymocytes than in Cd3e^PRS/PRS thymocytes; Sla^{- /-} thymocytes also had greater and more sustained phosphorylation of Erk1 and Erk2 kinases (Fig. 8b). Relative to Cd3e^PRS/PRS thymocytes, unstimulated and stimulated Sla^{- /-} thymocyte had a protein that migrated as a 21-kilodalton species and probably corresponded to the p21 tyrosine-phosphorylated CD3 isoform (Fig. 8c). Therefore, despite the presence of identical amounts of TCR complexes on the surface of Cd3e^PRS/PRS and Sla^{- /-} thymocytes, the lack of CD3 PRS resulted in less constitutive and TCR-induced phosphorylation of CD3 p21.

Figure 8: Tyrosine phosphorylation in thymocytes from wild-type, Cd3e^PRS/PRS and Sla^{- /-} mice after TCR cross-linking.

(a–c) Analysis of thymocytes (genotype, above blots) left untreated or stimulated with anti-CD3 for 1–30 min (above lanes) and immediately lysed; whole-cell lysates were separated by 8% SDS-PAGE (a,b) or 12.5% SDS-PAGE (c). (a,b) Immunoblot analysis with anti-phosphotyrosine (-phosphotyrosine (P-Tyr-100); a) or antibody to Erk1/2 phosphorylated at Thr202 and Tyr204 (-phospho-p44/p42 Erk1/2 (Thr202,Tyr204); b, top) and anti-Erk1/2 (-p44/p42 Erk1/2; b, bottom). Arrowheads indicate presumed proteins based on their expected molecular size. (c) Immunoblot analysis of the p21 form of tyrosine-phosphorylated CD3 subunit with anti-phosphotyrosine. Below, membranes stripped and reprobed with anti-CD3 and antibody to phospholipase C-1 (-PLC1). (d) Immunoassay of thymocytes left untreated or stimulated for 10 min with anti-CD3; lysate aliquots corresponding to equivalent amounts of proteins for each sample were immunoprecipitated with anti-CD3, separated by 12.5% SDS-PAGE and analyzed by immunoblot with anti-phosphotyrosine (4G10). Below, membranes stripped and reprobed with anti-CD3 (M-20) and anti-CD3- (H46-968) to confirm the position of the CD3 subunit and evaluate the ratio of the 16-kilodalton and p21 isoforms of CD3. Data are representative of at least two independent experiments.

Full size image (55 KB)

To confirm the results reported above, we left wild-type, Cd3e^PRS/PRS and Sla^{- /-} thymocytes untreated or stimulated them with anti-CD3, then precipitated TCR complexes with anti-CD3 and analyzed them by immunoblot with anti-phosphotyrosine, anti-CD3 or anti-CD3. Comparison of the Cd3e^PRS/PRS and Sla^{- /-} samples showed that engagement of TCR expressed on Sla^{- /-} thymocytes led to the generation of stronger CD3 p21, CD3 p23 and CD3 phosphorylated isoforms (Fig. 8d), and the anti-CD3 immunoblot confirmed that p21 was far less abundant in Cd3e^PRS/PRS thymocytes than in Sla^{- /-} thymocytes. The very distinct amounts of TCR expressed on the surface of wild-type and of mutant thymocytes complicated direct comparison of wild-type and Cd3e^PRS/PRS thymocytes (Fig. 7a); however, analysis of the anti-CD3 immunoblots indicated that although Cd3e^PRS/PRS thymocytes contained more CD3 subunits than did wild-type thymocytes, they had less abundant expression (in unstimulated conditions) or at most equal expression (in TCR-stimulated conditions) of the p21-phosphorylated isoform of CD3. These results collectively suggest that in preselection DP cells, the CD3 PRS controls the amount of CD3 phosphorylation, which therefore explains how the CD3 PRS affects SLAP-dependent CD3 degradation. Moreover, on a 'per-TCR' basis, the TCR-CD3 complexes expressed on Cd3e^PRS/PRS DP cells showed impaired signaling (Fig. 7) that could explain the lower positive selection noted in TCR-transgenic mice expressing CD3^PRS subunits (Supplementary Fig. 13 online).

Mice.

Mice were housed in specific pathogen–free conditions and were handled in accordance with French and European directives. The generation of Cd3e^PRS/PRS mice (B6-Cd3e^tm2Mal) is described in the Supplementary Methods online. Zap70^{- /-} mice, Sla^{- /-} mice, mice lacking the CD3 subunit (Cd3e^5/5 mice) and Cd3e^5/5 mice deficient in MHC class II molecules have been described^{3, 44, 45, 46}. HY TCR-transgenic mice express a monoclonal population of CD8⁺ T cells specific for Smcy738-746, a male antigen presented by H-2D^b molecules³⁰; Marilyn TCR-transgenic mice express a monoclonal population of V