Apoptosis but not other activation events is inhibited by a mutation in the transmembrane domain of T cell receptor beta that impairs CD3zeta association.

The transmembrane domain of T cell receptor (TCR) β contains a conserved immunoreceptor tyrosine-based activation-like motif consisting of a duplicated YXXL sequence. The motif is also present in TCRγ, the equivalent chain to TCRβ in γδ T lymphocytes but is absent in TCRα and TCRδ. To determine the putative role of this sequence in TCR·CD3 complex assembly and signal transduction, a TCRβ chain cDNA was mutated in the C-terminal tyrosine of the motif, cloned in an expression vector, and transfected into TCRβ-negative Jurkat cells. Transfectants of the mutated chain (MUT) expressed, on average, much less TCR·CD3 complex on the membrane than wild type TCRβ transfectants. Radiolabeling experiments suggested that the mutation caused a loose association of the CD3ζ chain resulting in a defective assembly. However, stimulation of high TCR·CD3 expressing wild type and MUT clones with monoclonal antibodies and Staphylococcus aureus enterotoxin B resulted in similar levels of CD25 and CD69 expression, interleukin-2 secretion, and TCR·CD3 complex down-regulation. By contrast, MUT cells were clearly resistant to activation-induced cell death, and they did not express CD95-ligand upon activation. These results suggest a differentiated intracellular signaling pathway leading to apoptosis in which Tyr-TM11 of the immunoreceptor tyrosine-based activation motif-like motif and CD3ζ appear to be involved.

The T cell antigen receptor⅐CD3 complex (TCR⅐CD3) 1 is formed by a clonotypic heterodimer (␣␤ or ␥␦), which provides ligand specificity, non-covalently linked to at least four invariant chains (CD3⑀, -␥, -␦, and -) (for review see Refs. [1][2][3]. Assembly occurs by pairwise interactions (4), and as a result, the TCR⅐CD3 complex is formed by four dimeric components as follows: (a) clonotypic TCR␣ and TCR␤ chains that are covalently linked via a single extracellular disulfide bond; (b) the non-covalent CD3␥⑀ dimer; (c) the non-covalent CD3␦⑀ dimer; and (d) a disulfide-linked family dimer consisting of any of the five defined members of this family, , , , , and the ␥ chain of the high affinity Fc⑀ receptor (5). In addition to these interactions, stable pairwise associations can also be observed between single clonotypic and CD3 chains (4,6,7). Nevertheless, the CD3 dimer will only assemble in the complex if all the other subunits are present, thus explaining why CD3 is the last component to be integrated into the TCR⅐CD3 complex during assembly (8 -10). The stoichiometry, as well as the possible formation of alternative TCR⅐CD3 complexes, is still a question of debate.
The ability of antigen receptors to transduce signals to multiple biochemical cascades is the central event of immune cell activation (11). Engagement of the multicomponent TCR⅐CD3 complex with its antigen/MHC ligand, agonist mAbs, or superantigens results in several biochemical processes critical for the functional activation of T lymphocytes, including cellular proliferation, cell differentiation, and programmed cell death. Using chimeric molecules and reconstituted receptors, several reports have provided evidence for a redundant signal transduction domain present in the CD3 chains (12). This domain, which contains a pair of YXXL/I sequences (where X corresponds to a variable residue), is known as antigen recognition activation motif, activation receptor homology-1, tyrosinebased activation motif, or immunoreceptor tyrosine-based activation motif (ITAM) (13). This sequence, which is triplicated in and present as a single copy in each of the other CD3 chains, has an important role in coupling the TCR⅐CD3-mediated signaling to intracellular signal transduction molecules (14). However, how the TCR␣␤ clonotypic receptor is able to deliver signaling events to the other chains of the complex after its stimulation by an antigen remains to be solved. Due to the short length of the cytoplasmic tails and to the apparent lack of inherent signaling activity of both clonotypic chains, it is assumed that the TCR heterodimer transmits the signals produced upon antigen binding through the CD3 chains. TCR and CD3 domains involved in TCR⅐CD3 complex assembly could also be good candidates to mediate interactions required for signal transduction. In this regard, transmembrane domains of the TCR and CD3 chains have been shown to play a crucial role in assembly, partly due to charge neutralization between the basic residues of TCR␤, TCR␣ (or TCR␥ and TCR␦), and the acidic residues of each one of the CD3 chains (1)(2)(3). However, additional residues must participate to add specificity to the interaction.
The location in the transmembrane region of TCR␤ of a conserved ITAM-like motif, which is also present in TCR␥ but not in TCR␣ or TCR␦, prompted us to search for the role this sequence could play in assembly, surface expression, and function of the TCR⅐CD3 complex, as a possible first signal transducer after TCR stimulation. Although clones that expressed high levels of the TCR⅐CD3 complex on the cell surface were obtained, the mutation of the C-terminal tyrosine of the ITAMlike motif in TCR␤ resulted in an impaired CD3 assembly. Surprisingly, stimulation of MUT clones through the TCR⅐CD3 complex resulted specifically in defective induction of programmed cell death, whereas other activation read outs were normal, including IL-2 secretion, IL-2R␣ (CD25), and CD69 expression and down-regulation of the TCR⅐CD3 complex. These results suggest a specific role of CD3 in TCR⅐CD3induced apoptosis.

PCR Procedures and DNA Cloning
The HA1.7 ␤ chain cDNA (V␤3) was a gift from Dr. M. Owen (EMBL/ GenBank accession number X63456). It was derived from the HA1.7 human CD4 ϩ T cell clone specific for the influenza hemagglutinin (HA) peptide 307-319 in the context of HLA-DR1. For cloning into the pSR␣ expression vector, XhoI and BamHI restriction sites were introduced at the 5Ј-and 3Ј-ends, respectively, of the HA1.7 ␤ chain cDNA. Full WT cDNA was PCR-amplified by using, as amplification primers, the 5Јsense oligonucleotide CCCCTCGAGCCATGGGAATCAGGCTC (oligo 1), which included the XhoI site, and the 3Ј-antisense oligonucleotide GGGGGATCCAGGGCTGCCTTCAG (oligo 4), which included the BamHI site.
The construction of the full MUT cDNA involved the following steps. 1) Generation of a PCR-amplified product from the 5Ј-end of the molecule to the nucleotide 892. Mutations at positions 881 and 882 were introduced by using the 3Ј-antisense oligonucleotide CAGCACAG-CAAGCAGGGTGGC (oligo 2) which included the changes (underlined and boldface) A/T and T/C to produce the desired Tyr/Leu change. Oligo 1 was used as sense primer for the amplification. 2) Generation of a PCR-amplified product from position 872 to the 3Ј-end of the cDNA. It was carried out by using the oligonucleotide GCCACCCTGCTTGCT-GTGCTG (oligo 3) as 5Ј-sense primer which is complementary to oligo 2, and the oligo 4 previously described as 3Ј-antisense primer. 3) Generation of the full cDNA MUT. It was achieved by mixing purified PCR products derived from steps 1 and 2, taking into account that they have complementary tails, and performing a PCR with oligo 1 and oligo 4 as amplification primers.
In all cases, 25 amplification cycles were performed, each consisting of 1 min at 94°C, 2 min at 55°C, 2 min at 72°C, and a single final extension of 5 min at 72°C. Full PCR-derived XhoI/BamHI-tailored cDNAs, WT and MUT, were isolated, cloned into pSR␣ expression vector previously digested with XhoI and BamHI, and fully sequenced in Sequagel gels (National Diagnostics, Hessle, UK) by the Sequenase (U.S. Biochemical Corp.) method.

Transfection, Selection, and Surface Expression Analysis
DNA-mediated gene transfer into 31.13 cells was accomplished by electroporation. 5 ϫ 10 6 cells were mixed with 40 g of DNA in a sterile disposable cuvette (Bio-Rad) in a volume of 0.8 ml of RPMI 1640 and electroporated in a Bio-Rad gene pulser unit at 250 V and 960 microfarads. After electroporation, the cells were immediately resuspended in 10 ml of RPMI 1640 (Bio-Whitakker) containing 10% fetal calf serum. Forty-eight hours later, the cells were plated in 96-well flat bottom plates at 2 ϫ 10 4 cells/well in selective medium containing 1 mg/ml G418 (Sigma). After 3-4 weeks, the growing cells were screened by flow cytometry for surface expression of CD3 (UCHT1), V␤3 (JOVI-3), and V␤8 (JR2).

Flow Cytometry
10 5 -10 6 cells were incubated on ice for 30 min with a specific mAb (2-4 g/ml) in PBS, washed, and incubated for additional 30 min with FITC-conjugated anti-mouse IgG antibody. After washing, the cells were analyzed in a flow cytometer (EPICS-XL MCL, Coulter).

Functional Assays
IL-2 Production Assay-10 5 cells/well were stimulated with soluble UCHT1 or JOVI-3 at concentrations ranging from 0.5 to 5 g/ml in the presence of 10 ng/ml PMA. After 24 h, 100 l of culture supernatant were removed from each well and frozen to ensure that no viable cells remained. The IL-2 content of these supernatants was determined in a CTLL2 proliferation assay. CTLL2 were seeded at 10 4 cells/well in a 1:2 dilution of the culture supernatants, and 24 h later the cells were pulsed for 6 h with 1 Ci of [ 3 H]thymidine (Amersham Corp., Buckinghamshire, UK) per well. Cells were harvested, and [ 3 H]thymidine incorporation was determined by liquid scintillation counting.
Cell Viability Assays-2 ϫ 10 4 cells/well were stimulated with plastic-bound mAbs UCHT1, JOVI.3, or soluble SEB at different concentrations. After 48 h, viable and dead cells were counted in a hemocytometer in the presence of 0.25% trypan blue. Antibody-coated plastic wells were prepared by overnight incubation of 96-well plates (Costar, Cambridge, MA) with various concentrations of purified mAb in PBS at 4°C.
Down-regulation of CD3-2.5 ϫ 10 5 cells/well were incubated at 37°C for 20 h with different concentrations of soluble SEB or for 4 h with 10 g/ml JOVI-3. Cells were washed, stained with a FITC-conjugated CD3 mAb, and analyzed by flow cytometry.
Flow Cytometry Analysis of DNA Degradation-2 ϫ 10 4 cells/well were stimulated for 48 h with soluble SEB, ranging from 1 to 100 g/ml, or with anti-CD95 mAb at 500 ng/ml. They were then harvested from culture and permeabilized in 500 l of 100 g/ml propidium iodide (Sigma), 0.05% Nonidet P-40, 10 g/ml RNase, in PBS. Alternatively, cells were stimulated with 1 M A23187 plus 15 ng/ml PMA for 24 h. After vortexing, samples were allowed to equilibrate at 4°C in the dark for at least 1 h before being analyzed. Propidium iodide fluorescence analysis was performed by flow cytometry (EPICS-XL MCL, Coulter).

Metabolic Labeling and Immunoprecipitation
20 ϫ 10 6 cells/ml were washed twice with Dulbecco's modified Eagle's medium and incubated for 4 h in Dulbecco's modified Eagle's medium without cysteine and methionine in the presence of [ 35 S]methionine (0.5 mCi/ml). After labeling, the cells were lysed in 1% Brij 96 (Sigma) lysis buffer containing protease inhibitors. The lysates were pre-cleared twice with preimmune serum and protein A-Sepharose beads. Immunoprecipitation was carried out using specific antibodies pre-bound to protein A-Sepharose beads as described (16).

Cell Surface Radioiodination
To enhance the detection of , the cells were pretreated with the water-soluble Bolton-Hunter reagent before radioiodination. Briefly, 20 ϫ 10 6 cells were washed with PBS, resuspended in 1 ml of PBS plus 200 ng/ml of sulfosuccinimidyl-3-(4-hydroxyphenyl)-propionate (Pierce), and incubated at room temperature for 30 min. The reaction was stopped by diluting the cells with 10 ml of 10 mM L-lysine in PBS. The cells were centrifuged and 125 I-labeled by the lactoperoxidase method and subsequently lysed and immunoprecipitated as above.

Expression of a TCR-␤ Chain Mutated in a Transmembrane
Tyrosine Residue Results in Low TCR/CD3 Expression-The comparison between the amino acid sequences of the transmembrane domains of TCR␤ chains from different species (Fig.   1) shows a high degree of homology, therefore suggesting that this domain plays a role other than anchoring the chain in the cell membrane. The extent of residue conservation in this domain is much lower when TCR␤ is compared with TCR␥, its analogue chain in ␥␦ T cells. Interestingly, an ITAM-like motif, as defined by the sequence (YXXLXXXXXXYXXL/I), is present only in the transmembrane domains of TCR␤ and TCR␥, whereas the tyrosine residues of the motif are absent in TCR␣ and TCR␦. It is significant, however, that in TCR␣ both tyrosine residues are conservatively substituted by phenylalanine residues.
To investigate the putative role of the TCR␤ ITAM-like motif, its C-terminal tyrosine residue (named Tyr-TM11) was mutated to leucine using a PCR-based mutagenesis protocol. WT and MUT versions of TCR␤ HA1.7 cDNA were transfected into the 31-13 TCR␤ minus Jurkat mutant cells previously described (15). After selection in the presence of G-418, WT and MUT TCR␤ transfectants were analyzed by flow cytofluorim- ]Methionine-labeled WTB7 and MUTC2 clones were immunoprecipitated with CD3⑀ mAb SP34 or TCR␤ mAb JOVI-3. The immunoprecipitates were subjected to two-dimensional electrophoresis, first under nonreducing conditions, followed by reducing conditions. The positions of the most relevant proteins are indicated. ␣ m and ␤ m , mature forms; ␣ i and ␤ i , immature forms. B, mutated TCR␤ chain mRNA is overexpressed in MUTC2 cells. Total RNA from WTG7 or MUTC2 cells was transferred to nylon membranes after denaturing electrophoresis and probed with V␤3, TCR␣, and CD3-chains, and ribosomal 28 S specific probes. V␤3 messenger RNA shows five times more expression in MUTC2 than in WTG7 cells. etry using UCHT1 mAb to determine whether transfection of TCR␤ chain cDNA had reconstituted the expression of the TCR⅐CD3 complex. As shown in Fig. 2, WT clones expressed surface TCR/CD3, with a mean percentage of positive cells of 58%. On the other hand, MUT clones expressed, on average, only 25% positive cells. In this regard, Fig. 2 shows a clustering of the WT clones in the 60 -80% positive region while MUT clones remain in the 10 -30% region. The reduced surface expression of the TCR⅐CD3 complex in clones containing the mutated TCR␤ suggests that this protein does not efficiently reconstitute the TCR⅐CD3 complex. To further analyze the possible functional implication of Tyr-TM11, three MUT clones expressing high levels of TCR⅐CD3 were used in comparative experiments with three WT clones expressing similar levels of the complex. All six clones were positive not only with UCHT1 but also with the anti-V␤3 mAb JOVI-3, which recognizes the variable region of the transfected TCR␤ HA1.7. On the contrary, none of the clones were stained with a specific V␤8 antibody, showing that TCR⅐CD3 expression in these clones had not resulted from a reversion and reexpression of endogenous Jurkat's TCR␤. Other T cell markers were equally expressed in all clones, and, as expected, HLA class II molecules were not detected on the surface of Jurkat, 31-13, and transfected cells (data not shown).
Mutation of Tyr-TM11 from TCR␤ Results in an Impaired Association of CD3-To investigate the causes of the possible assembly defect, WTB7 and MUTC2 cells were metabolically labeled with [ 35 S]methionine for 4 h, and their lysates were immunoprecipitated with CD3⑀ and TCR␤-specific mAbs. As shown in Fig. 3A, JOVI-3 immunoprecipitated a TCR␣/␤ het-erodimer from WTB7 cells that had the mobility of the mature Golgi processed form. However, in MUTC2 cells, JOVI-3 mainly immunoprecipitated a TCR␣/␤ heterodimer that migrated with the characteristic mobility of the immature heterodimer located in the endoplasmic reticulum (ER) (19). Nevertheless, some mature TCR␣/␤ heterodimer was also detected in MUTC2 cells, accounting for the surface staining of these cells. In addition to the mature and immature heterodimer, large amounts of unassociated TCR␤ chain were detected in the MUTC2 clone (spot in the diagonal, Fig. 3A) but not in WTB7 cells. The immunoprecipitation with the CD3⑀-specific mAb SP34 showed that in MUTC2 cells the immature form of the TCR␣/␤ heterodimer was CD3-associated. This result suggests that mutation of Tyr-TM11 does not affect the ability of TCR␤ to associate with TCR␣ or CD3␥, -␦, and -⑀ but results in an impaired exit of the TCR⅐CD3 complex from the ER. The ER location of most of the TCR⅐CD3 complex in MUTC2 cells is suggested by the dominant presence in the immunoprecipitate of immature TCR␣/␤ heterodimer, which has been previously shown to be endo-H-sensitive (19). In addition, confocal microscopy of JOVI-3-stained WTB7 and MUTC2 cells confirmed that in MUTC2 the TCR⅐CD3 complex was mostly located in the ER (data not shown). Nevertheless, enough complex should leak out of the ER to allow high expression on the cell surface. The export deficit produced by the mutation in TCR␤ could be compensated in MUTC2 cells by the overwhelming expression of the TCR␤ chain. In this regard, an RNA blot analysis with a V␤3-specific probe showed that the transfected mRNA is expressed at 5-fold higher levels in MUTC2 than in WTB7 cells (Fig. 3B).
The hampered transport of the TCR⅐CD3 complex from the ER in MUTC2 is compatible with a deficit in the association of CD3 chain which has been suggested to be the last step in the assembly of the TCR⅐CD3 complex (9,20). Indeed, although some CD3 homodimer was detected in metabolic labelings of WTB7 cells (Fig. 3A), none was detectable in MUTC2. However, due to the poor labeling of CD3, we could not rule out the existence of this subunit. To further explore this possibility, WTB7 and MUTC2 cells were surface iodinated by the lactoperoxidase method prior to exposing the cells to sulfosuccinimidyl-3-(4-hydroxyphenyl)-propionate, previously shown to enhance the detection of CD3 (16). After iodination, the cell lysates were immunoprecipitated with CD3⑀and CD3-specific mAbs. As shown in Fig. 4, immunoprecipitation with CD3 antibody N39 from the WTB7 lysate resulted in the coprecipitation of CD3␥, -␦, and -⑀ as well as the TCR␣/␤ heterodimer. On the other hand, although immunoprecipitation with anti-CD3 from MUTC2 revealed similar, if not higher, amounts of CD3 to those found in WTB7, the amounts of coprecipitated CD3␥, -␦, and -⑀ chains and TCR␣/␤ were highly diminished. Conversely, the immunoprecipitation with a CD3⑀ mAb showed that the levels of CD3␥, -␦, -⑀, and TCR␣/␤ were similar in WTB7 and MUTC2 cells, while the levels of complex-associated CD3 were much lower in MUTC2 than in the WTB7 clone (Fig.  4). These experiments suggest that, on the MUTC2 cell surface, CD3 is loosely associated in the TCR⅐CD3 complex and mostly present as an independent molecule.
Mutation of Tyr-TM11 Abrogates TCR⅐CD3-induced Programmed Cell Death but Not Other Activation Events-To test the effect of Tyr-TM11 mutation on TCR/CD3-mediated signaling, MUT and WT clones were stimulated with PMA and different concentrations of soluble UCHT1 mAb; after 24 h, IL-2 production was measured in a CTLL2 proliferation assay. Although some intrinsic variability was found among the different clones, there was not a clear effect of the mutation in the ability of the TCR⅐CD3 complex to promote IL-2 secretion (Fig.  5A). Similar results were obtained when soluble JOVI-3 mAb; or plastic-bound UCTH-1 and JOVI-3 mAbs were used (data not shown). As expected, the parental TCR⅐CD3-negative Jurkat mutant 31-13 was unresponsive. However, a very clear effect of Tyr-TM11 mutation on cell viability was observed. Consistently, stimulation with plastic-bound UCHT1 or JOVI-3 mAbs resulted in significant losses of cell viability, in WT clones but not in MUT ones, as measured by trypan blue dye exclusion (Fig. 5B).
Upon stimulation with different ligands of the TCR⅐CD3 complex, a clear dose-response effect on cell viability was observed. That this phenomena is a dose-dependent effect is shown in Fig. 6. Clearly, clone MUTC2 was much less sensitive to stimuli-induced cell death than clone WTB7. Interestingly, soluble Staphylococcus aureus enterotoxin B (SEB), which has been shown to be a ligand for human TCR V␤3 (21) even in the absence of class II molecules (17,22), was able to induce a similar effect (Fig. 6C). The mutation of Tyr-TM11 in TCR␤, therefore, abrogated TCR⅐CD3-induced cell death but did not affect IL-2 secretion. T cells have been described to undergo apoptosis when stimulated with TCR/CD3 ligands (23)(24)(25). To determine whether apoptosis is the mechanism underlying cell death in our cell system, WTG7 and MUTC2 cells were stimulated with various concentrations of soluble SEB, and the level of apoptosis was measured by propidium iodide incorporation in the DNA of permeabilized cells. As shown in Fig. 7, SEB stimulation leads to an increase in the number of cells with sub-G 1 amounts of DNA in the WTG7 clone, whereas no effect was observed in MUTC2 cells. These results suggest that, upon stimulation, genomic DNA was only degraded in WT cells. Similar results were observed in WT and MUT clones obtained from independent transfections (data not shown).
To ascertain whether other activation events were affected by the mutation in TCR␤, clones were stimulated with anti-TCR␤ mAb JOVI-3. The expression of activation markers IL-2R␣ chain (CD25) and CD69 was then analyzed. As shown in Fig. 8A, JOVI-3 stimulated CD25 and CD69 expression in both MUTC2 and WTB7 cells. Similar results were obtained when these clones were stimulated with UCHT1 or SEB (data not shown). In addition, ligand-induced receptor internalization experiments showed that both JOVI-3 mAb (Fig. 8B) and SEB (not shown) induced down-regulation of the TCR⅐CD3 complex in WT and MUT cells with similar efficiencies. Taken together, all previous data indicate that mutation of Tyr-TM11 in TCR␤ results in a specific inhibition of TCR⅐CD3-induced apoptosis but has no effect in IL-2 production, CD25 and CD69 expression, and TCR⅐CD3 complex down-regulation.

TCR⅐CD3 Stimulation of MUT Cells Results in Reduced Induction of CD95 Ligand Expression Although the Cells Remain
Fully Sensitive to CD95-induced Apoptosis-The pathways for activation of apoptosis are multiple and distinct in different cell types. Recently, it has been described that in T cells, TCR⅐CD3 cross-linking induces both CD95 (also known as APO-1 and Fas) and CD95 ligand (CD95-L) expression and that TCR⅐CD3-induced programmed cell death is mediated by engagement of CD95 by its ligand (23)(24)(25). In Jurkat cells, CD95 expression is constitutive, and TCR/CD3-induced cell death is exclusively mediated by the induction of CD95-L expression (23). We decided to test whether the different sensitivity to TCR ⅐ CD3induced apoptosis in WT and MUT cells was dependent on differences in the CD95 pathway. As shown in Fig. 9A, WTB7 and MUTC2 clones expressed constitutively comparable levels of CD95 on the cell surface. In addition, the levels of expression of CD95 remained stable in both cell types upon TCR/CD3 stimulation with specific mAbs (data not shown). The signaling pathway through CD95 was functional because similar levels of DNA degradation could be induced when either WT or MUT cells were activated via CD95 with a specific anti-CD95 mAb, CH-11, which is able to induce apoptosis (Fig. 9B). Furthermore, stimulation by a TCR-independent mechanism, phorbol ester plus calcium ionophore (Fig. 9C), confirmed that MUT cells are not intrinsically defective in the apoptotic machinery.
On the other hand, soluble DX2, an anti-CD95 mAb previously shown to block CD95/CD95-L interaction, was able to prevent SEB (Fig. 9D) and UCHT1-or JOVI-3 (not shown) induced cell death in WTG7 cells. These results suggested that, as previously shown in Jurkat (23), apoptosis in WT cells was mediated via CD95/CD95-L interaction. Therefore, it was compelling to test whether induction of CD95-L expression was produced upon TCR⅐CD3 triggering. As shown in Fig. 9E, stimulation with SEB resulted in expression of CD95-L mRNA in WTG7, whereas in MUTC2 cells such expression was highly reduced. The same results were obtained when other WT and MUT clones were tested (data not shown). However, stimulation with phorbol ester plus ionophore resulted in similar induction of CD95-L in WT and MUT clones (data not shown) suggesting, again, that signaling to apoptosis through the CD95/CD95-L pathway takes place in MUT cells when a stimulus that bypasses TCR⅐CD3 is used.
These results suggest that mutation of Tyr-TM11 in TCR␤ results in defective CD95-L induction in Jurkat transfectants and that this is the basic mechanism underlying the resistance to TCR⅐CD3-induced apoptosis of these cells. DISCUSSION We have investigated the effects of replacing a highly conserved tyrosine residue in the transmembrane region of TCR␤ chains, on assembly and signal transduction. Interestingly, this residue lies in a sequence that resembles consensus ITAM motifs previously described in the cytoplasmic tails of multisubunit receptors including these of the CD3␥, -␦, -⑀, andchains of the TCR⅐CD3 complex (13). Two major effects were observed upon transfection of the MUT TCR␤ chain cDNA in Jurkat TCR␤ negative cells. First, an impaired intracellular association of CD3 was observed that apparently resulted in the surface expression of a TCR⅐CD3 complex loosely associated with the CD3 chain. Second, and more interesting, a specific inhibition of apoptosis was induced upon TCR⅐CD3 stimulation. As all other measured T cell activation events were not affected by the mutation, our results suggest the existence of an independent intracellular signaling pathway for apoptosis in which CD3seems to be involved.
The conspicuous location of the ITAM-like motif makes it difficult to explain how Tyr-TM11 could play a role in transmembrane signaling. Recently, however, a conserved antigen receptor transmembrane (CART) motif has been described (26), which shares homology with the ITAM-like motif we describe here. Based on theoretical grounds, the authors suggested that several amino acids, including Tyr-TM11, could be important in antigen receptor function. The mutational studies performed on the transmembrane domain of surface IgM implicating polar FIG. 8. Stimulation of the TCR⅐CD3 complex in WT and MUT clones causes its down-regulation as well as the induction of CD25 and CD69 expression. A, expression of CD25 and CD69 antigens. WTB7 and MUTC2 clones were stimulated for 24 h with 10 g/ml plastic-bound V␤3 mAb JOVI-3. The expression of CD25 and CD69 was analyzed by flow cytofluorimetry (thick line). Dashed lines represent CD25 and CD69 expression before TCR/CD3 stimulation. B, WTG7 and MUTC2 clones were stimulated for 4 h at 37°C with JOVI-3 mAb (10 g/ml). TCR/CD3 expression was analyzed by flow cytometry after staining the cells with a FITC-conjugated anti-CD3 mAb. Broken lines represent negative controls of staining as determined by using irrelevant FITC-conjugated antibodies. sequences in intersubunit interactions and cell activation support this idea (27)(28)(29). Curiously, it has been recently reported that the selective point mutation of a similar tyrosine residue included in an intracytoplasmic ITAM motif completely abrogated the ability of this motif to mediate cell death signals (30). However, the expected behavior of the ␤ chain transmembrane domain as a target for a protein tyrosine kinase has not been reported, and our attempts to demonstrate such potential susceptibility have been unsuccessful. If this failure reflects either technical problems or natural incapacity of the TCR␤ chain to be phosphorylated will require additional and more specific experiments.
Although the mutation of Tyr-TM11 could directly affect the association with a not yet defined protein that would mediate triggering of apoptosis, a more plausible explanation for the presented data is that the impaired association of CD3 in the complex is the cause of defective apoptosis. Hence, this subunit would be directly involved in the activation of the cell death program. This idea is reinforced by the data of Vignaux et al. (31) showing that cross-linking of a Tac/ chimera, containing the cytoplasmic tail of CD3, induces apoptosis in a murine T cell hybridoma. Furthermore, it has recently been shown that in vitro cross-linking of Tac/ induces double-positive thymocyte cell death from Tac/ transgenic RAG2Ϫ/Ϫ mice (32). The same effect was observed by cross-linking of Tac/⑀ chimera in Tac/⑀ transgenic RAG2Ϫ/Ϫ mice. If these data reflect either a redundant role for the ⑀ and chains in delivering qualitatively similar signals for inducing apoptosis or the result of nonphysiological stimuli remains controversial. On this line of evidence it has been reported that the role of individual ITAM-containing chains may have qualitatively different functions that can control the TCR⅐CD3 triggering under different physiological situations depending upon the activation status of the cell (33).
It has been recently described that the route leading to apoptosis triggering by the TCR⅐CD3 complex in murine hybridomas and in Jurkat cells is mediated by the CD95 pathway. While in Jurkat cells CD95 is constitutively expressed, CD95-L is rapidly up-regulated, after SEB or anti-CD3 stimulation, leading to induction of cell death (23)(24)(25)34). However, we have shown that while apoptosis triggering by direct stimulation of CD95 is not affected by the mutation, the induction of CD95-L upon TCR/CD3 triggering is inhibited. Thus, the CD95-L induction seems to be the basic phenomenon affected by Tyr-TM11 mutation and probably by the impaired association of CD3. This idea is supported by the fact that cross-linking of a Tac/ chimera was capable of inducing CD95-L expression (31).
On the other hand, the mutation of Tyr-TM11 of TCR␤ did not affect other activation events, implying that a CD3 minus TCR⅐CD3 complex may be sufficient to trigger other activation programs. In this regard, a CD3 minus module has previously mAb DX2 and analyzed by flow cytometry. Dashed lines represent the fluorescence intensity for 97% of mock stained cells. B, DNA degradation. WTG7 and MUTC2 cells were treated with 500 ng/ml anti-CD95 mAb, CH11. After 48 h, DNA degradation was measured by propidium iodide staining. The percentage of cells with sub-G 1 amounts of DNA is indicated in each panel. C, DNA degradation induced upon addition of phorbol ester plus calcium ionophore. WTB7 and MUTC2 cells were incubated with 1 M calcium ionophore A23187 plus 15 ng/ml PMA. After 24 h, DNA degradation was measured as in B. D, inhibition of SEB-induced cell death with anti-CD95 mAb DX2. DNA degradation was measured by propidium iodide staining in WTG7 and MUTC2 cells treated for 48 h with 10 g/ml SEB in the presence of 500 ng/ml DX2. The percentage of cells with sub-G 1 amounts of DNA is indicated in each panel. E, differential expression of Fas-L (CD95-L) in WT and MUT clones. Total RNAs were extracted from WTB7 and MUTC2 cells stimulated with 10 g/ml soluble SEB for the indicated times. Blots were hybridized with a CD95-L cDNA fragment and rehybridized with a 28 S ribosomal probe as loading control.
FIG. 9. Failure in CD95 ligand induction is responsible for the resistance to SEB or anti-TCR/CD3 mAb-induced programmed cell death in MUT clones. A, expression of CD95 in nonstimulated WTG7 and MUTC2 clones. Cells were stained with 4 g/ml anti-CD95 been shown to be sufficient to induce IL-2 secretion in a murine hybridoma (35). Also consistent with these data, TCR complexes containing a CD3 chain deprived of all ITAM motifs have been shown to promote normal T cell maturation in transgenic mice (36). Moreover, double-positive thymocytes from these mice are able to transduce signals resulting in up-regulation of CD69 surface expression, up-regulation of CD5 mRNA, and down-regulation of RAG mRNAs.
The three ITAM motifs of CD3 and the one of CD3⑀ have been shown to bind with different affinities to the SH2 domains of proteins involved in the activation pathway, such as ZAP-70, Shc, and phosphatidylinositol 3-kinase (12,37,38). It is, therefore, reasonable to assume that the role of CD3 in apoptosis could take place through the specific recruitment, by any of its ITAM motifs, of signal transducing proteins needed for apoptosis triggering but not, for instance, for IL-2 secretion. Alternatively, apoptosis triggering may require a signal threshold that would not be reached in the absence of the six ITAM motifs included in the CD3 homodimer. Therefore, the lack of associated CD3 may have a quantitative rather than a qualitative effect. Further experiments are required to distinguish between both possibilities.
The data presented in this study are consistent with the idea that the association of CD3 is the last step in the assembly of the TCR⅐CD3 complex because it requires the previous association of CD3␥, -␦, and -⑀ with TCR␣ and TCR␤. Thus, transient transfection studies performed in COS cells indicated that CD3 does not interact directly with TCR or CD3 chains alone, but it requires the previous association of both TCR␣ and -␤ to the other CD3 chains in order to provide adequate conformation for binding of CD3 (8,20).
It has been shown that the electrostatic interactions between basic residues in the transmembrane region of TCR␣ (lysine and arginine), TCR␤ (lysine), and single acidic residues in the transmembrane domains of each of the different CD3 chains play an important role in assembly (7). The tyrosine residue mutated in TCR␤ is in position ϩ4 of the lysine residue which, assuming an ␣-helix conformation for the transmembrane domain, would be placed on the same face of the helix. Thus, lysine and tyrosine could form a hydrophilic surface in the transmembrane domain of TCR␤ that mediates assembly with the CD3 subunits. Nevertheless, while the basic residues of TCR␤ and -␣ chains appear to be involved in the association with CD3␥, -␦, -⑀, and perhaps -, Tyr-TM11 seems to be specifically involved in assembly of the CD3 chain. Interestingly, mutation of two contiguous hydrophilic amino acids, including a tyrosine, located in the transmembrane domain of the chain of the B cell receptor resulted in loss of association with Ig␣ and Ig␤ chains (29), pointing to a general role of transmembrane tyrosine residues in multisubunit receptor assembly and expression. The effect of Tyr-TM11 mutation on CD3 association could be due to a direct interaction between this residue and specific amino acids of CD3, or alternatively, to steric or allosteric effects on other CD3 chains.
Most TCR␤ MUT clones expressed lower levels of TCR⅐CD3 complexes on the cell surface than WT ones. This effect was likely due to a hampered exit of TCR⅐CD3 complexes from the ER. The inefficient association of CD3 chain to the TCR␣␤/ CD3␥␦⑀ core in the ER probably avoids the transport of complete complexes to the cell surface (9,10), and only in those clones where mutated TCR␤ is overexpressed would the defective association of CD3 be compensated. Interestingly, CD3 was only partially associated to the remaining TCR⅐CD3 complex on the cell surface of the MUTC2 clone expressing high levels of the complex, suggesting the existence of an association/dissociation equilibrium both inside the cell and on the plasma membrane. The association of CD3 would be necessary for the TCR⅐CD3 complex to leave the ER but once by-passed the ER retention checkpoint could CD3 again dissociate. Our results suggest, therefore, that the composition of the TCR⅐CD3 complex could fluctuate among several forms at equilibrium. Non-TCR-associated CD3 could either remain as a free chain on the cell surface or become associated to other proteins such as the transferrin receptor (16). Interestingly, it has been described that tumor infiltrating T lymphocytes show a marked alteration in the structure of their TCR⅐CD3 complexes, namely the absence of chain (39 -41).
The structural and/or functional significance of the ITAMlike motif located in the transmembrane domains of TCR␤ and TCR␥ chains is unknown. Our data point out that the Cterminal tyrosine residue of such a motif is implicated in the association with CD3 and in the TCR⅐CD3-mediated signaling, probably via CD3, that leads to apoptosis. The putative role of other amino acids of the motif in assembly and signal transmission through the TCR⅐CD3 complex will require further investigation.