An evolutionarily conserved glycine motif located between the Basic-Rich Sequence (BRS) and the Proline-Rich Sequence (PRS) in the cytoplasmic tail of CD3ε is essential for T cell activation induced by the T Cell Receptor (TCR).
A comparison of the amino acid sequences in the cytoplasmic tail of CD3ε reveals a remarkable pattern of conservation across ontogeny, spanning from cartilaginous fishes to mammals (Fig. 1a). The most conserved feature is the Proline-Rich Sequence (PRS, PxxPxPDY), followed by the tyrosine residues of the Immunoreceptor Tyrosine-based Activation Motif (ITAM). Other sequences previously reported to be crucial for outside-in signal transmission within the TCR are also preserved, including the Basic-Rich Sequence (BRS)13,14 and the RK motif involved in recruiting Lck15. Additionally, here are other conserved amino acid residues located between the BRS and the PRS. Notably, we noticed the presence of a conserved proline residue (Pro161 in human CD3ε) and one or more glycine residues, forming a di-glycine motif (Gly169-Gly170 in human CD3ε), present in all vertebrates and conserved in mammals (Fig. 1a). Characterization of the structure of the cytoplasmic tail of murine CD3ε by Nuclear Magnetic Resonance (NMR) (PDB = 2K4F) revealed a random coil throughout the 57 amino acids of the polypeptide, except for two turns of an α-helix around the C-terminal tyrosine of the ITAM, an α-helix turn at the end of the PRS and N-terminal tyrosine of the ITAM, and an α-helix turn immediately after Pro161 (Fig. 1b). The residue corresponding to Gly169 is located within a long unstructured sequence.
To assess whether mutation to alanine of different residues in human CD3ε affected the TCR's ability to activate T cells in response to ligation with an anti-CD3 antibody, we first generated a CRISPR/Cas9 full knockout of CD3ε in the Jurkat T cell line. Subsequently, the resulting CD3εKO cell line was transfected with either wild-type CD3ε or alanine mutants corresponding to residues Arg155 and Arg156 of the BRS, residue Pro161, residues Gly169 and Gly170 of the di-glycine motif, and residues Arg192 and Lys193 of the RK motif. Stable clones were selected, and those exhibiting TCR expression at the cell surface similar to that of wild-type Jurkat cells (Fig. S1 in the Supplementary Information) were chosen for functional analysis. To this end, all Jurkat clones were stimulated with various concentrations of anti-CD3 antibody, and IL-2 present in the supernatant was measured by ELISA. As expected, the CD3εKO line failed to produce IL-2 in response to anti-CD3 stimulation (Fig. 1c). All mutants significantly reduced IL-2 production compared to the wild-type control. Of note, cells bearing the G169A mutation exhibited the least responsiveness to anti-CD3 stimulation, releasing IL-2 at levels as low as the CD3εKO control (Fig. 1c). In a separate set of experiments, Jurkat CD3εKO cells were transduced with lentiviral constructs expressing either wild-type CD3ε or the G169A or G170A mutants, and the resulting polyclonal population was stimulated with different doses of anti-CD3. The response to anti-CD3 stimulation was evaluated based on the expression of CD69 and CD25 as activation markers. Cells bearing either glycine mutation showed deficient responses to TCR triggering (Fig. 1d and 1e). These results demonstrate that mutation of glycine residues 169 and 170 to alanine in CD3ε diminishes the human TCR's ability to respond to stimulation with an anti-CD3 antibody.
The G169A mutation in CD3ε has a differential impact on the phosphorylation of specific tyrosine residues in CD3ζ.
To begin dissecting the mechanisms by which mutation of the di-glycine motif impairs TCR-triggered T cell activation, we first investigated some of the earliest events after TCR triggering, namely the tyrosine phosphorylation of the CD3ε and CD3ζ ITAMs. We focused our analysis on the G169A mutation, given that this has a stronger impact on IL-2 production (Fig. 1c). In anti-CD3 immunoprecipitates, we observed that the phosphorylation of the N-terminal tyrosine (Tyr166) of CD3ε was significantly inhibited by the G169A mutation (Fig. 2a). While CD3ε has just two phosphorytable tyrosine residues in its single ITAM, CD3ζ contains three ITAMs in its cytoplasmic tail, accounting for a total of six phosphorylatable tyrosines16. A diagram illustrating the residue numbers for these six tyrosine residues within the three ITAMs of human CD3ζ is presented in Fig. 2b. A deficiency in CD3ζ phosphorylation in Jurkat cells expressing the G169A mutant was noticed in Western blots of postnuclear supernatants of whole-cell lysates with a pan-phosphotyrosine-specific antibody (Fig, 2c). In addition to a putative defect on the tyrosine phosphorylation of CD3ζ, a clear defect in the phosphorylation of other proteins was noticed. According to their mobility in SDS-PAGE, those proteins could correspond to phospho-ZAP70 and phospho-Lck (Fig. 2c). Interestingly, the putative phospho-CD3ζ bands showed distinct high- and low-molecular-weight forms, and the G169A mutant reduced the phosphorylation of the upper form of apparent high molecular weight (Fig. 2c). To determine if the G169A mutation in CD3ε primarily affected the phosphorylation of CD3ζ to its highest molecular weight form, we examined the effect of the mutations on the tyrosine phosphorylation of CD3ζ using a phospho-specific antibody for Tyr142 (ITAM ζc). Tyrosine-phosphorylated CD3ζ in Jurkat cells with a wild-type TCR appeared as two protein bands of different molecular weight, with the slowest mobility one presumably corresponding to the fully phosphorylated form. While the G169A mutation did not abolish tyrosine phosphorylation of CD3ζ, it did lead to a noticeable reduction in the ratio of highly phosphorylated to low-phosphorylated forms (Fig. 2d).
We further examined the effects of the G169A mutation on CD3ζ phosphorylation using two additional available antibodies specific for Tyr83 (ITAM ζa) and Tyr111 (ITAM ζb) in postnuclear total cell lysates. Western blot analysis with the Tyr111-specific antibody revealed a protein band corresponding to the low-molecular weight form and a faint band corresponding to the high-molecular weight form, although only in WT cells (Fig. 2e). Intriguingly, western blotting with the Tyr83-specific antibody showed only one protein band corresponding to the high-molecular-weight form in wild-type cells (Fig. 2f). This single protein band was reduced in Jurkat cells expressing the G169A mutation compared with WT. Altogether, these findings suggest that the low-molecular weight form of phospho-CD3ζ is phosphorylated at Tyr111 and Tyr142 but not at Tyr83, whereas the high-molecular-weight form is phosphorylated at Tyr83, Tyr142, and, to a lesser extent, at Tyr111. Mutation G169A inhibited the phosphorylation of CD3ζ at Tyr83 but not at Tyr111 or Tyr142 (Fig. 2d-f), indicating that the mutation in CD3ε specifically impairs the phosphorylation of Tyr83 in the first ITAM of CD3ζ. Western blotting with antibody specific for total CD3ζ, phosphorylated and unphosphorylated, served as a loading control and to indicate that the unphosphorylated CD3ζ has the mobility of low apparent molecular weight one (Fig. 2g). The effect on the phosphorylation of the other three tyrosine residues, for which suitable antibodies were not available, remains unknown. Nevertheless, these results suggest that the G169A mutation has a specific impact on a tyrosine residue located in the upstream, membrane-proximal ITAM of CD3ζ.
The G169A mutation impairs proximal TCR signaling and the phosphorylation of Lck at Ser59 by ERK.
To investigate the consequences of the G169A mutation on TCR signaling downstream of the TCR, we decided to measure the phosphorylation of the tyrosine kinases Lck and Zap70, which are tightly associated with early signal transmission. A general cartoon depicting the hierarchy of phosphorylation events following TCR activation is shown in Fig. 3a. The activated TCR becomes phosphorylated on the tyrosine residues of its ITAMs by Lck. Different tyrosine phosphorylation sites have been identified in Lck, with two of the most important being the autophosphorylation site Tyr394 in its catalytic domain, which activates Lck, and the inhibitory site Tyr505 located at the negative regulatory tail17. Double-site tyrosine phosphorylation of the ITAMs by Lck creates docking sites for the cytoplasmic tyrosine kinase Zap70, which becomes recruited to the TCR. Once bound to the TCR, Zap70 becomes phosphorylated, including at the activating site Tyr31918. Active Zap70 phosphorylates, among other targets, the membrane-bound adapter protein LAT, which serves as a docking site for PLCγ1. Activation of PLCγ1 leads to activation of, among others, the Raf-ERK MAP kinase pathway. Raf activates the dual kinase MEK, which phosphorylates ERK at the Thr202 and Tyr204 sites. Active ERK is part of a feedback loop since it phosphorylates Ser59 in the unique region of Lck17. The effect of such phosphorylation on Ser59 is debated, with some data supporting a reduction in Lck activity19 and others supporting enhanced activity by inhibiting the recruitment of the phosphatase SHP-120.
Upon stimulation with anti-CD3 antibody in Jurkat cells bearing the G169A mutation phosphorylation at Tyr319 of Zap70 was strongly inhibited compared to that found in wild-type CD3ε Jurkat cells (Fig. 3b). Similar results were obtained for phosphorylation of LAT at Tyr132 (Fig. 3c); phosphorylation of ERK by MEK was inhibited as well (Fig. 3b).
We studied the effect of the G169A mutation on the phosphorylation of Lck at both the activating site Tyr394 and the inhibitory site Tyr505 in total cell lysates. We found that stimulation of both WT and G169A cells with anti-CD3 did not increase the total content of phospho-Lck, regardless of the anti-phosphosite antibody used (Fig. 3d), consistent with previous findings21. However, there was a clear increase in the apparent molecular weight of Lck induced by anti-CD3 stimulation. Specifically, a higher molecular weight band appeared after 5 minutes of stimulation, rather than at the earlier time point of 1 minute, and this band increased in abundance after stimulation at the expense of a low-molecular weight band (Fig. 3d). The G169A mutation reduced the stimulation-induced increase in the upper high-molecular weight form at the expense of the low-molecular weight form, as visualized by a diminished high/low ratio increase compared to WT cells (Fig. 3d, below). Since both Lck bands were equally detected by the anti-phosphoTyr394 and the anti-phosphoTyr505 antibodies, we investigated if the differences in both Lck bands could be due to phosphorylation at the Ser59 site. By probing whole cell lysates of WT Jurkat cells with the anti-phospho-Lck(Ser59) antibody, we found that phosphorylation at that site peaked at 5 minutes after stimulation with anti-CD3 antibody (Fig. 3e). Such induced phosphorylation was blocked in the presence of the MEK inhibitor UO126, thus proving that Ser59 phosphorylation was promoted by ERK when activated by MEK. Interestingly, the mobility of the Lck band detected by the anti-phospho-Lck(Ser59) antibody corresponded to the high-molecular weight form. Consequently, we investigated if the G169A mutation reduced the phosphorylation of Lck at Ser59 compared to WT Jurkat cells and found that this was the case (Fig. 3f). The same blotting membrane was reprobed with the anti-phosphoTyr394 antibody, which confirmed that the high-molecular weight Lck band corresponded to the form phosphorylated at Ser59 and that such phosphorylation was inhibited in cells bearing the G169A mutation. Together, these results show that the G169A mutation in CD3ε inhibits the activation of early downstream effectors of TCR signaling, including Zap70, LAT, and ERK, and that it has a selective inhibitory effect on the feedback loop of ERK on Lck phosphorylation at Ser59.
The G169A mutation impairs the outside-in transmission of conformational changes in the TCR, preventing the exposure of the PRS in CD3ε.
To elucidate a possible mechanism for the deficient phosphorylation of specific tyrosine residues in CD3ε and CD3ζ when the TCR bears the G169A mutation, we investigated if this mutation impaired the transmission of the conformational change from the extracellular domains of the TCR to the PRS of CD3ε. To this end, we performed a pull-down assay on detergent cell lysates of Jurkat WT and G169A cells with beads coated with GST protein fused to the SH3.1 domain of Nck1, as schematized in Fig. 4a. In WT cells, stimulation with anti-CD3 for 1 minute was already sufficient to induce the binding of the triggered TCR to the SH3.1 domain (Fig. 4b, uncropped blots are in the Supplementary Information). This induced binding was observed by Western blotting with the three antibodies used, specific for CD3ζ, CD3γ, and TCRα, indicating that the entire TCR shifted from a non-competent Resting state to a competent Active state. Compared with a WT TCR, the G169A mutation impaired the anti-CD3-induced binding of the TCR to the SH3.1 domain (Fig. 4b and 4c), suggesting that the G169A mutation impaired the transmission of the conformational change along the cytoplasmic tail of CD3ε.
Despite the inducible binding of the entire TCR to GST-SH3.1, some binding of CD3ζ, but not of the other subunits, was detected at time 0 (Fig. 4b). In another GST-SH3.1 pull-down experiment, the blotting membrane was incubated with the anti-phospho-CD3ζ(Tyr142) antibody. This showed some binding to the beads of the low-molecular weight form of CD3ζ with low phosphorylation even in lysates from resting cells (Fig. 4d). However, the binding of the high-molecular weight form of CD3ζ was exclusively detected after stimulation for one minute or longer, suggesting that the active TCR binding to the SH3.1 domain is fully tyrosine phosphorylated. The G169A mutation impaired such induced binding of the high-molecular weight, fully phosphorylated, form of CD3ζ (Fig. 4d) and suggests that adoption of the Active conformation by the TCR, including the cytoplasmic tail of CD3ε, is required for full phosphorylation of CD3ζ.