Signaling Efficiency of the T Cell Receptor Controlled by a Single Amino Acid in the β Chain Constant Region

The Vα2.1 and Vβ8.1 TCR cDNAs were isolated from the T cell hybridoma, 3BBM74, and confer reactivity to the I-A^bm12 alloantigen and the staphylococcal enterotoxin B (SEB) superantigen (16). The wild-type (wt) and mutated TCR constructs were generated using overlapping oligo nucleotides and PCR as previously described (15). All constructs were verified by DNA sequencing using the SequiTherm™ cycle sequencing kit (Epicentre Technologies Corp., Madison, WI) and the deduced amino acid sequences of the mutant β chains are shown in Fig. 1.

Retroviruses encoding wt or chimeric TCR chains were generated and used to infect the α⁻/β⁻ T cell hybridoma, 58hCD4, as previously described (15).

The 58hCD4 cell line has been previously described (15, 17) and was provided by O. Acuto (Pasteur Institute, Paris, France). The murine fibroblast cell line DAP.3 transfected with human HLA class II DR1 molecules (18) or the human, DR1-expressing, B lymphoblastoid cell line .221 (19) was used to present SEB. The use of the cytokine indicator cell lines HT-2 and FDC-P1 have been previously described (15).

The use of anti-Vα2 (B20.1), anti-Vβ8.1 (F23.1), and anti–CD3-ε (145-2c11) mAbs, as well as anti–CD3-γ, anti– CD3-δ, and anti–CD3-ε rabbit antisera have been previously described (20). The anti–ζ chain mAb, H146-968 (21), was purified from culture supernatants using protein G–Sepharose beads (Pharmacia Biotech AB, Uppsala, Sweden). The rabbit anti– ZAP-70 antiserum was provided by Dr. L. Samelson (National Institutes of Health, Bethesda, MD). The antiphosphotyrosine mAb 4G10 was purchased from Upstate Biotechnology Inc. (Lake Placid, NY).

The levels of cell-surface antigens were detected by immunofluorescence and analyzed on a FACScan^® using the CELLQuest analysis software (Becton Dickinson, Rutherford, NJ). T cell stimulation and IL-2 and IL-3 assays were performed as previously described (15).

T hybridoma cells (2 × 10⁷/sample) were lysed in buffers containing 1% digitonin and the TCR complex immunoprecipitated with 2 μg of B20.1 (anti Vα2). Immunoprecipitations, Western blotting, and detection of CD3 chains were carried out as previously described (20).

For superantigen stimulation, 10⁶ .221 APCs were incubated with or without SEB (10 μg/ml) for 2 h in 1 well of a 24-well plate. 10⁷ T hybridoma cells were added to the APCs, centrifuged at 1,000 rpm for 30 s, and then incubated at 37°C for 20 min. Cells were washed twice with ice cold PBS containing 0.4 mM Na₃VO₄ and 0.4 mM EDTA, and were then lysed in 0.5 ml lysis buffer (10 mM Tris/HCl, pH 7.5, 150 mM NaCl, 1 mM PMSF, 10 μg/ml leupeptin, and 10 μg/ml aprotinin) containing 2 mM Na₃VO₄, 25 mM NaF, and 1% Triton X-100. Lysates were centrifuged at 12,000 g for 10 min, and the relevant proteins in the supernatant were immunoprecipitated with 1.5 μg of H146-968 (anti-ζ) mAb or 3 μl of a rabbit polyclonal anti–ZAP-70 antiserum. Immunoprecipitates were recovered using protein G–Sepharose beads. The tyrosine phosphorylated proteins present in these immunoprecipitates were analyzed by Western blotting and detected using the antiphosphotyrosine mAb 4G10.

To better understand the functional relationships between the α/β heterodimer and the CD3 and ζ components, we made a number of chimeric TCR-β chains by replacing domains from the β chain constant region with homologous domains from the TCR-γ chain. We first generated a chimeric β chain, βVII, which contained TCR-β–derived V, D, J, and C region sequences up to and including the interchain Cys, followed by Cγ sequences encoding the COOH-terminal part of the CP domain, the transmembrane domain, and the cytoplasmic tail (Fig. 1,A). As seen in Fig. 1,B, this chimeric βVII chain was not expressed at the cell surface when paired with a wt α chain. However, Cβ and Cγ differ in the length of their CP domains. There are 13 amino acids between the interchains Cys₁₂₇ and Ala₁₄₁ in all β chains and 16 amino acids between Cys₁₁₇ and Ala₁₃₄ in all mammalian γ chains (Fig. 1 A and reference 22). Thus, the fact that the βVII chimera could not be expressed at the cell surface might have been due to the possibility that the Cγ length in this region (16 amino acids) was incompatible with surface expression in the context of an α/β TCR.

To test this idea, we constructed two additional chimeras, βV and βVI, which respected the conserved β chain length (13 amino acids) in this region (αwt; Fig. 1,A). Both the βV and the βVI chimeras were expressed at the cell surface in conjunction with a wt α chain (αwt; Fig. 1,B). In fact, the αwt/βV and the αwt/βVI chimeric TCRs were expressed at two- to threefold higher levels on the cell surface than were the αwt/βwt receptor. Hybridomas expressing these chimeric TCRs were stimulated with the superantigen, SEB bound to APCs. As seen in Fig. 2,A, the αwt/βV TCR responded to SEB about as well as the wt TCR, while the αwt/βVI TCR was clearly less sensitive (∼100-fold) to this superantigen. This signaling deficit was seen even more clearly when the SEB response of a hybridoma expressing the αwt/βVI mutant receptor was compared to that of an αwt/βwt hybridoma sorted for an equivalently high level of TCR expression (Fig. 2,B). Furthermore, transgenic mice expressing the αwt/βVI TCR were defective in responding to SEB and the I-A^bm12 alloantigen (data not shown). Thus, the signaling defect of this mutant TCR was not limited to superantigens. On the other hand, these chimeric TCRs could be activated by plate-bound anti-TCR mAbs (Fig. 2 C and data not shown), indicating that the hybridomas expressing the αwt/βVI TCR were not intrinsically defective. Therefore, TCRs comprised of the chimeric βVI chain seemed to be specifically deficient in transducing signals from antigenic ligands.

TCRs from these hybridomas were immunoprecipitated, and the associated polypeptides were analyzed by Western blotting (data not shown). The CD3-γ, -δ, and -ε chains as well as the ζ polypeptides were coprecipitated from hybridomas expressing either the wt TCR or the signaling-defective, αwt/βVI TCR (data not shown). In spite of a pronounced signaling defect (Fig. 2), the interactions with the CD3 and ζ chains were preserved in complexes containing the functionally defective, αwt/βVI heterodimer.

The tyrosine phosphorylation of the ζ chain was examined in these hybridomas as well. From the data in Fig. 3,A, it was apparent that the ζ chains in the αwt/βwt and αwt/βVI TCRs could be tyrosine phosphorylated and that both forms (p21 and p23) of the phosphorylated ζ chain could be generated. Although the tyrosine phosphorylation of ZAP-70 could not be demonstrated even in hybridomas expressing the αwt/βwt receptor (data not shown), an anti–ZAP-70 antiserum was used to evaluate ZAP-70 recruitment (Fig. 3 B). In superantigen-stimulated cells expressing either wt or mutant TCRs, the p21 and p23 forms of the ζ chain as well as the p34 phosphoprotein were coprecipitated with ZAP-70. These experiments indicated that even in hybridomas expressing the signaling-defective αwt/βVI TCR, ZAP-70 was nevertheless recruited to the phosphorylated ζ chains. Thus, the signaling defect in hybridomas expressing this mutant TCR was likely downstream from ZAP-70 recruitment.

Binding of an appropriate ligand to the TCR eventually leads to the activation of the protein kinase C/ras and the calcium intracellular signaling pathways (1, 7). Deficient activation of the PKC/ras pathway can be compensated for by the addition of the phorbol ester, PMA, whereas deficient activation of the calcium pathway can be compensated for by the addition of a calcium ionophore, such as ionomycin. In Fig. 4, hybridomas expressing the wt or the defective αwt/βVI TCR were stimulated with SEB in the presence of PMA or ionomycin. The impaired response of the αwt/βVI TCR was only marginally affected by several different concentrations of PMA, but was rescued in the presence of ionomycin used at several different concentrations (Fig. 4 and data not shown). These findings suggested that the mutant TCR was defective or inefficient in activating the calcium pathway.

The functional chimera, βV, contains the β chain sequence ASYQQ, which has been replaced with the γ chain sequence, LQFQF, in the nonfunctional chimeric chain, βVI (Fig. 1,A). The most striking amino acids in this region are Tyr₁₃₄ and Gln₁₃₆, both of which are encoded by Phe in murine Cγ1. Point mutations were introduced into the wt β chain cDNA, and the resulting mutant β chains were paired with an αwt chain and were tested for their reactivity to SEB. As seen in Fig. 5, A and B, the Tyr₁₃₄ → Phe mutation had a negligible effect on the response to SEB, whereas TCRs carrying the Gln₁₃₆ → Phe mutation had a marked and reproducible effect on SEB responsiveness. Considering the pronounced effect of the Gln₁₃₆ → Phe exchange on superantigen responsiveness, we examined the amino acid residues present at this position in all Cβ and Cγ sequences, some of which are shown in Fig. 5,C. Strikingly, the amino acids at this position in the β chain are always polar or charged. In contrast, γ chains always contain hydrophobic amino acids at this position (Fig. 5 C and reference 23).

We took advantage of the differences between α/β and γ/δ TCRs to generate signaling-defective receptor mutants. Since the TCR-β and -γ constant regions are structurally homologous, we could replace domains of Cβ with similar domains of Cγ. Most of the resulting chimeric chains paired with a wt α chain and were expressed on the cell surface (Fig. 1,B). Hybridomas expressing αwt/βVI TCRs are 100-fold less sensitive to antigenic signals than hybridomas expressing wt receptors (Fig. 2, A and B). The signaling defect seems specific for antigenic ligands, since αwt/ βVI TCRs respond fully to mAbs capable of receptor cross-linking (Fig. 2,C and data not shown). What distinguishes the functional βV chimera from the defective βVI chimera are the amino acids present from position 132 to 136 that are derived from Cβ and Cγ, respectively (Fig. 1,A). Site-directed mutagenesis defined Gln₁₃₆ as a critical residue in this region. This position has been conserved in the β chains of all known species from the skate to humans (23) and is either polar or charged (Fig. 5,C). In contrast, the analogous residue in the γ chain has been conserved as a hydrophobic amino acid (Fig. 5,C). This conservation is functionally significant since changing this residue from a hydrophilic to a hydrophobic amino acid is accompanied by a significant loss in signaling efficiency (Figs. 2 and 5).

There is evidence that the α/β and γ/δ heterodimers do not interact with the CD3 complex in the same way. For example, γ/δ TCRs can be expressed in the absence of CD3-δ (11). Furthermore, a CD3-ε epitope, contained within human α/β receptors and recognized by the mAb WT31 is masked in γ/δ receptors (9, 10, 24). Many of these differences might be attributed to the CP (membrane proximal) domains of the different TCR chains. We have previously identified a motif within the α CP domain which is strikingly absent from TCR-δ chains. This motif controls the proper association of CD3-δ and -ζ to the TCR complex and is required for antigen-driven signal transduction (15).

Examining the CP domain of the β chain, we have found a region within Cβ which controls the efficacy of antigen-driven signal transduction. Although alteration of this region does not lead to changes in the subunit composition of the TCR complex (data not shown), a clear biological defect was observed (Fig. 2). Mutation of this region had no observable effect on the amount of ζ chain phosphorylation nor on the recruitment of ZAP-70 to the ζ homodimer (Fig. 3). Thus, the signaling defect in this mutant is likely downstream from ZAP-70 recruitment. The signaling defect can be corrected or at least compensated in the presence of a calcium ionophore (Fig. 4), suggesting that the mutation within the β chain may lead to the inefficient induction of Ca²⁺ mobilization.

There is little structural information available for the CP domains since this region has been truncated from the TCR heterodimers used in crystallographic analyses (25, 26). Nevertheless, the CP domains of the α and β chains may function as a unit since amino acids within these domains have been conserved in α/β TCRs during the last 500 million yr (15, 22, 23). The length of the CP domains has been conserved as well. Considering the region between the interchain Cys and the transmembrane domain (defined by the beginning of the CART motif [22]), all known β chains contain 13 amino acids (22, 23). On the other hand, α chains contain from 17 to 20 amino acids in this region. Thus, in α/β TCRs, a length disparity between the α and β chain CP domains has been conserved throughout vertebrate evolution. In contrast, the lengths of the TCR-γ and -δ chain CP domains (between the interchain Cys and the transmembrane domain) are comparable in each species examined thus far (22, 23). Therefore, a length similarity between the δ and γ chains has been conserved in this region in γ/δ receptors during a similar period of vertebrate evolution.

Functionally, the CP domains may mediate an interaction with the ζ homodimer. In this regard, α/β TCRs compete more avidly for the ζ chain compared to a γ/δ TCR (27). Furthermore, mutation of the α-CPM leads to a TCR complex that is unstably associated with the ζ chains (15). Nevertheless, why the CP domains have diverged between α/β and γ/δ TCRs is puzzling. It is conceivable that the CP domains in the α/β TCR have a specialized function such as interacting with the CD4 and CD8 coreceptors or that they have evolved to optimize the efficiency of signal transduction from MHC encoded ligands.

The authors thank Drs. Luca Bolliger, Thomas Göbel, Wayne Hein, Michel Record, and Sabine Stotz for careful reading of the manuscript, Beatrice Pfeiffer for photography, Dr. L. Samelson for the rabbit anti– ZAP-70 antiserum, and O. Acuto for the 58hCD4 cell line.