The JI PBL Intereron Source
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     
 

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bolliger, L.
Right arrow Articles by Johansson, B.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bolliger, L.
Right arrow Articles by Johansson, B.
The Journal of Immunology, 1999, 163: 3867-3876.
Copyright © 1999 by The American Association of Immunologists

Identification and Functional Characterization of the {zeta}-Chain Dimerization Motif for TCR Surface Expression1

Luca Bolliger2 and Britt Johansson

Basel Institute for Immunology, Basel, Switzerland


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
We recognized a common dimerization motif between the transmembrane (TM) domain of {zeta}-chain family members and glycophorin A. We have shown that a glycine within the {zeta}-dimerization motif is critical for {zeta}-homodimerization and also for its association with the TCR/CD3 complex. Similarly, two residues within the CD3{delta}{gamma} TM domains have proven to be critical for their interaction with the {zeta}-homodimer. A three-dimensional homology model of the {zeta}-chain TM domain highlights potential residues preferentially involved either in the {zeta}2-CD3 or {zeta}2-TCR{alpha}ß association, confirming our experimental findings. These results indicate that, for symmetrical reasons, the {zeta}-homodimer participates in the TCR/CD3 complex assembly by interacting with CD3{gamma}{delta} TM domains, thereby masking their degradation signals located in the cytoplasmic tails.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
The oligomeric TCR complex (TCR/CD3) consists of two clonotypic TCR chains responsible for Ag recognition ({alpha}ß or {gamma}{delta}) and four noncovalently associated invariant subunits (CD3{delta}, {gamma}, {epsilon}, and {zeta}). The minimal stoichiometry for the TCR complex has been proposed as TCR{alpha}ßCD3{gamma}{delta}{epsilon}2{zeta}2 (1). The invariant subunits of the complex are not only responsible for efficient TCR{alpha}ß surface expression but are also essential for the signal transduction cascade and couple the TCR complex to the cytoplasmic signal transduction machinery via immunoreceptor tyrosine-based activation motifs (2, 3, 4).

The biogenesis of the TCR complex is a sequential and well-ordered process, submitted to the peer control of the cellular architectural editing that has been extensively studied (5, 6). In summary, these studies demonstrate that the domains involved in the biogenesis of the TCR complex are localized to the extracellular portion as well as to the transmembrane domain of TCR {alpha}ß-chains, the CD3 complex, and the {zeta}-homodimer (7, 8, 9, 10, 11, 12, 13). Cotransfection experiments further defined heterotypic interactions between the CD3 complex and TCR{alpha}ß, but left the {zeta}-homodimer as a biochemical "orphan" because no specific interaction with other components of the TCR complex could be detected (6, 14). The incorporation of the {zeta}-chain into TCR/CD3 partial complexes is the last step during the biogenesis and is also the rate-limiting step for the surface expression of the TCR complex (15). All in all, these observations suggest that the {zeta}-chain is likely to interact with at least two different TCR/CD3 components. We have addressed this question by studying the extracellular (EC)3 (16) and transmembrane (TM) domains of the {zeta}-homodimer by site-directed mutagenesis. Several reports studying the TM domain of the {zeta}-chain demonstrate that this domain mediates homodimerization (11, 17, 18) and that its association to CD16 and TCR is influenced by single point mutations within the TM domain (19).

We identified a glycophorin A (gp A) dimerization-like motif in the TM domain of the {zeta}-chain. Site-directed mutagenesis of critical residues revealed that the {zeta}-homodimer interacts via different sides of its TM domain with CD3 and TCR{alpha}ß. A three-dimensional homology model from the solved structure of gp A provided important information about the potential spatial localization and the nature of these interactions. In addition, we identified and characterized a complementary conserved 6-aa motif in the TM domain of CD3{delta} and CD3{gamma} that is likely to interact with the {zeta}-homodimer. These results suggest that the TM domain of the {zeta}-chain acts as a "structural glue" and keeps TCR{alpha}ß, and CD3{delta} and CD3{gamma} associated within the TCR complex, thereby masking their degradation signals (20).


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Cell lines

The 2B4 derivative MA5.8, which lacks endogenous {zeta} expression, was reconstituted with the {zeta}-mutants described in this study. The BW5147 cell line lacking CD3{delta} and {zeta} expression was reconstituted with {zeta} wild-type (WT) and CD3{delta} WT or mutant (21). The ecotropic packaging cell line Bosc23 was purchased from American Type Culture Collection (Manassas, VA). All cells were grown in IMDM supplemented with 5% heat-inactivated FCS, 2 mM L-glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin, and 50 µM 2-ME.

Mutations, transfections, infections, and FACS analysis

The strategy for producing the {zeta}-chain mutations was outlined in Bolliger et al. (22). The CD3{delta} cDNA was kindly donated by Dr. F. Letoruneur. Oligonucleotides were designed to introduce a Kozak element and an EcoRI restriction site at the 5' end (5'-TTGAATTCCACCATGGAACACAGCGGGATTCTG) and a FLAG epitope, a stop codon, and a BamHI site at the 3' end. The modified DNA was then cloned into pGEM3Z (Promega, Madison, WI). On oligonucleotide bearing the AG->LL point mutations was designed (5'-AGAAGGCCTTCCGGTCTCATGTCCTGCAAAGCAGTAGACCAACAAGAGCAGGAGCAG), and the PCR product with the 5' oligo described above was cloned in pGEMCD3{delta}cDNA as an EcoRI-StuI fragment. All mutations were introduced with PCR and were sequenced on an automated sequencing machine (Applied Biosystems, Foster City, CA). The correct constructs were cloned into a retroviral vector carrying the puromycin or neomycin resistance genes (LXSP or LXSN) (23, 24, 25). Large DNA preparations were performed according to standard procedures. Bosc23 cells were transfected using calcium-phosphate or fugen (Boehringer Mannheim, Mannheim, Germany). MA5.8 cell were infected as described previously (26), with the exception that DEAE-dextran was used (40 µg/ml).

MA5.8 or transfectants were stained with FITC-labeled anti-CD3{epsilon} (145-2C11) (27) or anti-TCRß mAbs (H57-597) (PharMingen, San Diego, CA). Dead cells were excluded by staining with 0.5 µg/ml propidium iodide (Molecular Probes, Eugene, OR). Acquisition was performed using a FACScan flow cytometer, and analysis was performed with CellQuest software (Becton Dickinson, San Diego, CA).

Cell surface biotinylation, immunoprecipitation, SDS-PAGE analysis, and Western blotting

All methods employed are as described in Bolliger and coworkers (16, 22). Briefly, cells were biotinylated in bicarbonate buffer (20 mM NaHCO3, 150 mM NaCl) with sulfo-NHS-biotin (100 µg/ml; Pierce, Rockford, IL) at 107 cells/ml (28) and lysed in lysis buffer (1% Triton X-100 (Sigma, Buchs, Switzerland), 50 mM Tris-HCl, pH 7.6, 150 mM NaCl, 1 mM PMSF, 10 µg/ml aprotinin, and 10 µg/ml leupeptin) for 20 min at 4°C. Immunoprecipitations were performed on the resulting cleared lysates using 2 µg mAb (anti-CD3{epsilon} or anti-{zeta}) and resolved by nonreducing SDS-PAGE. Proteins were transferred onto a nitrocellulose membrane (Bio-Rad, Mountain View, CA).

For Western blotting, the nitrocellulose membrane was blocked with 5% low-fat milk in 0.2% Tween-PBS. Blots were probed with 200 ng/ml H146-968 (mAb anti-{zeta}-chain (29)) for 1 h at room temperature in PBS-Tween. H146-968 mAbs were visualized with a polyclonal peroxidase-conjugated goat anti-hamster Ig conjugate (Southern Biotechnology Associates, Birmingham, AL). Biotinylated proteins or secondary biotinylated mAb were detected using streptavidin-peroxidase (Southern Biotechnology Associates). Complexes were visualized with an enhanced chemiluminescence system (Pierce). The mouse anti-FLAG mAb (Eastman Kodak, Rochester, NY) was visualized with a goat anti-mouse secondary HRP-labeled mAb (Upstate Biotechnology, Lake Placid, NY). The rabbit polyclonal antisera used for detection of CD3{delta} and CD3{gamma} in Western blots were kindly provided by Dr. T. B. Bäckström (Malhagan Institute, Wellington, New Zealand). A secondary goat polyclonal peroxidase-conjugated goat anti-rabbit Ig conjugate was used for detection (Southern Biotechnology Associates).

Modeling

Various structural models were produced. Swiss-PDBviewer (30), RASMOL (31), and Ribbons (32) graphic and modeling software were used to produce three-dimensional models and pictures. A hand alignment of the primary sequences of gp A TM and {zeta}-TM domains was produced. The published coordinates of gp A TM were used to model the structure of the {zeta}-TM domain (PDB access ID, AFO1 at the Protein Data Bank of the Brookhaven National Laboratory) (33). Modeling and energy minimization were performed with the Moloc software package (Ref. 34 , kindly provided by the Modeling Unit of Hoffmann-La Roche, Basel, Switzerland).


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
The TM of {zeta} and Fc{epsilon}R{gamma} contains a gp A-like dimerization motif

The {zeta} family of TCR- and FcR-associated proteins is a paradigm for proteins that are structurally essential for the TCR complex surface expression and signal transduction. The size of the EC and cytoplasmic domains of the {zeta}-chain with respect to the plasma membrane is in contrast to the other components of the TCR complex, which have very large EC domains and relatively short cytoplasmic tails. We assumed that the EC and TM domains of {zeta}-chain were responsible for setting the structural framework for the surface expression of the assembled receptors (11, 17, 18, 22). A recent and extensive study concluded that the TM domain of the {zeta}-chain mediated its disulfide homodimerization. However, homodimerization per se could not be associated with any consensus motif or specific amino acids within the TM domain (11).

Sequence analysis of the TM sequences of the {zeta} family members ({zeta}- and Fc{epsilon}R{gamma}-chain) indicated the presence of a well-characterized dimerization motif identified in the TM domain of gp A (Fig. 1Go, gp A) (35, 36, 37, 38). While the complete dimerization motif consists of the amino acid sequence LIxxGVxxGVxxT (conserved amino acids are represented with single letter code, while x indicates nonconserved positions), the {zeta} family members contain a minimized motif (LxxxxxGVxxT). The glycine residue (G) present in the second part of the motif (GVxxT) has been shown to be critical for driving the dimerization of gp A (38). In fact, the presence of the glycine at position 13 (according to the numbering used in Fig. 1GoA) is critical for the production of SDS-stable gp A homodimers (35, 36, 38, 39). By analogy, the high degree of conservation of the critical part of the motif suggested to us that the homodimerization of {zeta}-chain could be mediated mainly via this motif.



View larger version (46K):
[in this window]
[in a new window]
  		FIGURE 1. A, Sequence alignment of transmembrane domains of {zeta}-chain, Fc{epsilon}R{gamma} subunit, and gp A are compared (m, mouse; h, human; c, chicken). Positions are indicated, the cysteine forming the disulfide bond in {zeta} and FcR is at position 2, and the amino acids of the motif are boxed. B, Quantification of surface TCR expression of the {zeta}-dimerization mutants expressed in MA5.8 hybridoma according mean channel intensity fluorescence shifts after anti-CD3{epsilon} staining. In the front row, {zeta}-mutants lacking the interchain bond are shown, while the back row shows the same mutants containing an interchain disulfide bond. Three FACS profiles of bulk transfectants stained with anti-CD3{epsilon} mAb are shown as examples to assess for the quality of the data compiled in the graph.

 
Mutation of the {zeta}-dimerization motif affects TCR surface expression

To test the possibility that the conserved amino acid (G13, see Fig. 1GoA) in the {zeta}-chain TM domain was indeed involved in homodimerization (cf gp A), we replaced G13 with either hydrophobic amino acids (A, V, L, and F) or with a polar amino acid (S). To directly test the contribution of the interchain disulfide bridge on dimerization, we generated the same mutants by replacing the cysteine responsible for the covalent {zeta}-{zeta} dimerization with a glycine (Fig. 1GoA). The different {zeta}-chain mutants were used to reconstitute the {zeta}-deficient cell line MA5.8 (15) by retroviral gene delivery. The TCR surface expression of bulk transfectants was analyzed by FACS, and mean channel fluorescence intensity was quantified. In the presence of the disulfide bridge the effect of WT(G) to A, V, L, and F mutations impaired the TCR surface expression in a manner that was "proportional" to the size of the side chain (Fig. 1GoB, graph, back).

The absence of the disulfide bridge dramatically influenced the behavior of the mutants (Fig. 1GoB, graph, front). While some mutants showed a minimal reduction of the surface expression compared with their respective disulfide-bonded counterparts (compare WT(G) and G->S in the back row to the front row in Fig. 1GoB) or no difference (G->A), others completely abolished TCR surface expression (G->V, G->L, and G->F mutants). Most likely, the bulkiness of the side chains of these amino acids affected the homodimerization of the {zeta}-chain. The data suggests that there is a size threshold above which homodimerization in the absence of a disulfide bridge is unfavorable. This is in agreement with the results of gp A (35).

In conclusion, our results demonstrate that the mutated glycine present in the dimerization motif is critical for {zeta}-chain dimerization and that a {zeta}-dimer is required for TCR surface expression. In addition, we observed that the disulfide bridge can physically enforce {zeta}-dimerization, thereby partially overcoming or even reversing some of the negative effects of the amino acid substitutions tested.

Mutations in the {zeta}-dimerization motif also affect {zeta}2-CD3 interactions

To define the effect of the single amino acid exchanges on the stability of the complex, we subjected the different mutants expressing TCR complex at the surface to surface labeling with biotin and to immunoprecipitations (IPs) with various Abs (Fig. 2Go). To assess variations in the complex recovery during experiments, three different Abs were used (anti-TCRß-chain, anti-CD3{epsilon}, and anti-{zeta} chain). Immunoprecipitates were resolved on nonreducing SDS-PAGE gels and Western blotted (Fig. 2Go, A and B, upper and lower panels, respectively). The presence of some surface-labeled TCR/CD3 on the {zeta}-deficient cell line MA5.8 is consistent with some TCR cell-surface expression (2–5% of WT levels by FACS analysis; Fig. 2GoA and data not shown).



View larger version (65K):
[in this window]
[in a new window]
  		FIGURE 2. Biochemical analysis of the MA5.8 expressing the {zeta}-dimerization mutants. Following cell-surface biotinylation, each mutant, including the negative control, was immunoprecipitated with three different mAb (anti-TCRß, anti-CD3{epsilon}, and anti-{zeta}) followed by resolution on SDS-PAGE under nonreducing conditions and Western blotting. Blots were probed for biotin-labeled proteins (upper part, full gel) and specifically for {zeta}-chain, CD3{delta}, and {gamma} (lower part, gel strips). {zeta}-mutants lacking a disulfide bond are shown in A, while the covalently linked {zeta}-mutants are shown in B.

 
The {zeta}-mutants lacking the interchain disulfide bond, but supporting TCR surface expression, were included in this analysis (Fig. 2GoA; WT(G), G->S, and G->A). Our results showed that the lack of a disulfide bond affected the stability of the entire TCR complex, because a complete TCR/CD3/{zeta}2 complex could not be recovered neither in anti-TCRß nor in anti-{zeta}-chain IPs (Fig. 2GoA; WT(G)). However, the immunoprecipitation with anti-CD3{epsilon} mAb was more efficient because CD3 and TCR{alpha}ß, but not {zeta}, could be recovered. Interestingly, the {zeta}-chain could be recovered with some TCR {alpha}ß-chains, without any CD3 components, implying a direct interaction between {zeta} and TCR{alpha}ß (compare WT(G) and WT(G) without disulfide bond in Fig. 2GoA, upper and lower panels). These results further confirmed a peculiar role of the {zeta}-disulfide bond in stabilizing the TCR complex, but not in the signaling activity (11, 16). The point mutations, G->S and G->V lacking the disulfide bond, improved the stability of the TCR complex. In fact, all components could be similarly recovered in all three IPs (compare upper and lower panel of G->S and G->A to WT(G) in Fig. 2GoA). The yield of {zeta}-chain in immunoprecipitation with anti-{zeta}-chain mAb was clearly increased in comparison to the IPs using the other Abs; however, the amount of CD3 and TCR {alpha}ß-chains did not change across all IPs, suggesting a loose interaction between {zeta} and the TCR/CD3 complex.

The IPs of mutants G->S and G->A containing the interchain disulfide bond were as efficient as WT in recovering all TCR components as shown in Fig. 2GoB (compare upper and lower panels of WT, G->S, and G->A). In the G->V {zeta}-mutation, the recovery of CD3 components was less efficient. In fact, the amounts of CD3{delta} and {gamma} were clearly reduced in the anti-TCRß and anti-{zeta} IPs in comparison to WT (compare upper and lower panels in Fig. 2GoB to WT). Furthermore, in contrast to the amount of TCR {alpha}ß-chains, the recovery of {zeta}-{zeta} homodimers was decreased in anti-TCRß and anti-CD3{epsilon} IPs. This suggests that the {zeta}-{zeta} homodimer is loosely associated with the CD3 complex. The effect of the point mutation was more accentuated in the G->L mutation. In this case, a complete TCR complex could only be recovered by immunoprecipitating with an anti-CD3{epsilon} mAb, although similar amounts of TCR{alpha}ß were precipitated with all three Abs (compare TCR{alpha}ß in Fig. 2GoB, upper panel). The IPs with anti-TCRß and anti-{zeta} mAb efficiently recovered the TCR {alpha}ß-chains associated with the {zeta}-homodimer but none of the CD3 components. This was demonstrated by the lack of surface-labeled CD3 components and confirmed in Western blots probed specifically for CD3{delta} and {gamma} antisera (Fig. 2GoB, G->L, compare upper and lower panel, and similarly to WT(G) in Fig. 2GoA). Taken together, these results clearly showed that {zeta}2 and TCR {alpha}ß-chains have a specific site of interaction that is CD3 independent and that can be localized to the EC or/and TM domain of the {zeta}-chain (40, 41).

Finally, the G->F point mutation destabilized all TCR/CD3/{zeta}2 complex. In fact, TCR{alpha}ß could be efficiently recovered only with anti-TCRß-specific Abs, while the CD3 complex was recovered only with anti-CD3 specific Abs and the {zeta}-homodimer was recovered only with anti-{zeta} Abs. In this case, the insertion of a bulky residue (G->F) almost completely destabilized the TCR/CD3 complex, resulting in a modular desegregation (TCR{alpha}ß, CD3, and {zeta}2 were recovered separately). The outcome of the G->F mutation further highlights the critical role of the {zeta}-homodimer as a possible intermolecular "glue," and may explain previous observations where a lack of association of the {zeta}-chain to any single subunit of the TCR complex was reported (9, 14). Throughout this study, we noted that the anti-CD3{epsilon} mAb was more efficient in recovering surface labeled-CD3 components with respect to TCR {alpha}ß-chains and the {zeta}-chain, suggesting an Ab-mediated general stabilizing effect on the TCR complex or a consistent amount of clonotype-independent CD3 complexes present at the cell surface (42).

Theoretical three-dimensional model of the {zeta}-{zeta} TM domain

The presence of a primary sequence homology between the {zeta}-TM domain and the TM domain of gp A enabled us to model the {zeta}-chain homodimer on the solved three-dimensional structure of the gp A homodimer. The TM domain of gp A represents a classical double {alpha}-helical domain with a common interface, defined by the dimerization motif (33). Homology modeling of the {zeta}-TM domain using the software package Swiss-PDBviewer (30) showed that the two cysteines of the {zeta}-chain forming the interchain disulfide bond were facing each other. This suggested a possible spatial consensus between the dimerization motif and the covalent bond of the {zeta}-homodimer, in support of our initial hypothesis of a {zeta}-dimerization motif (Fig. 3GoA).



View larger version (65K):
[in this window]
[in a new window]
  		FIGURE 3. Three-dimensional model of the TM domain of {zeta}-chain derived from the homology modeling performed on the solved structure of gp A homodimer. A, C-{alpha} backbone of gp A and {zeta}-chain were visualized with the software package Rasmol (31 ). The opposing cysteines in the {zeta}-homodimer are visualized as spheres. B, The backbone ribbon with amino acids that undergo a conformational change are highlighted (F10, F'10, and V14; WT are shown in green, mutant G->F in yellow). The disulfide bond linking the two backbones is shown in green (S-S). C, The cavity defined by amino acids F'10-I16-Y20-V'14 described in the Results can be clearly seen. The side chain of the G->F point mutant (yellow) clearly fills the empty space of the cavity. (B and C were produced with the molecular graphics software Ribbons (32 )). D, Top view of the two {alpha}-helices ({alpha} and {alpha}'). The location of the mutated glycine (G13 and G'13) is indicated with an arrow. The position of F10 and F'10 are also indicated.

 
Using the software package Moloc, we were able to gain information about possible conformations of the side chains. This software applies a force field calculation algorithm to minimize the energy of the overall structure (34). The three-dimensional model of the {zeta}-chain TM dimer, shown in Fig. 3Go, B and C, revealed interesting details with respect to the mutants described in this report. The various mutations described were modeled and the resulting putative conformational changes were scored after applying the force field calculations, which produced the most stable conformations.

Mutation of the WT glycine residue in the TM domain to amino acids with large side chains (V, L, and F) induced the displacement of both backbone chains (data not shown). Substitution of WT glycine (G) by V, L, and F resulted in major changes of the surrounding amino acid side chains. Figs. 3Go, B and C show the impact of the G->F mutation. The amino acids undergoing the most significant changes are presented here (F10, V14, and symmetrically F'10, V'14, compare the WT, green, to the G->F mutant, yellow). The most striking result of the model was that when G was substituted with F, the bulky side chain of F was housed in a pocket that usually is occupied by the two hydrogen atoms of the glycine residue (Fig. 3GoB, G->F is indicated with a yellow arrow). This is consistent when G is substituted with any large amino acid. The pocket in the TM of the {zeta}-homodimer could be clearly visualized (Fig. 3GoC), and its boundaries could be defined by amino acids F'10, V'14, on one monomer, and I16 and Y20 (WT side chains are colored green). Because of the symmetry of the homodimer, the same pocket is present on the opposite site of the homodimer. This modeling data suggests that the size of the side chains of the mutant can interfere with the surrounding residues.

In conclusion, these modeling results suggest that the TM interaction within the {zeta}-dimer may occur via the dimerization motif and that the interaction to the TCR/CD3 complex may occur mainly via the side of the {zeta}-homodimer. These results, although theoretical, are in full agreement with the biochemical data presented above and might explain the reported phenomenon.

Interestingly, the negative charges (aspartic acid, D6; Fig. 1GoA) in the homodimer were almost facing each other without distorting the structure. In fact, two protonated carboxylic groups can stabilize each other via H bonds. This interaction may be stabilized by the presence of the disulfide bond and may be responsible for the reduction of the expression of TCR complex in the absence of the interchain {zeta}-disulfide bond (Fig. 1GoB). Aspartic acid, D6, has been reported to be involved in the interaction of {zeta}-dimers with CD16. However, its direct role in the interaction to CD16 and its role in {zeta}-dimerization remains tentative due to contrasting reports with respect to the importance of D6 (11, 17).

Modulation of TCR surface expression by bulky residues located on different sides of the {zeta}-chain dimer

The results of the molecular modeling of the {zeta}-chain TM domain and its phylogenetic conservation highlighted two bulky residues: a phenylalanine (F10) and a tyrosine (Y12) that may point toward the lipid bilayer (Fig. 3GoB and Fig. 4GoA). F10, together with V14, I'16, and Y'20, define one of the cavities between the two {alpha}-helices comprising the {zeta}-TM domain (Fig. 3GoC).



View larger version (28K):
[in this window]
[in a new window]
  		FIGURE 4. Structural and biochemical analysis of F10 and Y12: two bulky phylogenetically conserved {zeta}-TM residues. A, Graphical representation of the two residues in the three-dimensional model proposed. B, Mean channel fluorescence intensity of transfected MA5.8 obtained by directly labeled anti-CD3 FACS staining. -, mAb absent. +, mAb added to determine the residual TCR/CD3 expression in the absence of the {zeta}-chain. C, Western blot analysis after surface labeling of the Y12 mutants (upper panel) or {zeta}-specific probing (lower panel).

 
We mutated F10 either to amino acids with smaller side chains (A, V, or L) or to an {alpha}-helix breaker (P). The mutants were processed as described earlier, and TCR surface expression was quantified by FACS analysis (Fig. 4GoB). While the presence of the {alpha}-helix breaker proline completely abolished TCR surface expression, amino acid residues with increasingly smaller side chains (F->A and F->V) clearly improved its surface expression (Fig. 4GoB, compare rows WT, F->A, F->V, F->L, and F->P). Biochemical analysis of F->A, F->V, and F->L did not reveal any defect in the composition of the TCR complex. There was a generalized increase in the recovery of the components of the complex due to the augmented surface expression with F->A and F->V (around 50%; Fig. 4GoB). These results suggest that an increase in the size of the cavity (i.e., reducing the size of the side chain of F10, as visualized in the three-dimensional model; Fig. 3GoC) positively affects the surface expression of the TCR. A possible explanation is that the interaction of the {zeta}-chain with the CD3 complex (most likely CD3{delta} and CD3{gamma}) is more stable if F10 is replaced with a smaller amino acid, thereby masking the degradation signals in cytoplasmic tails of CD3 (20) and ultimately leading to more efficient TCR surface expression.

The three-dimensional model presented here also depicts the side chain of Y12 pointing toward the lipid bilayer (details in Fig. 4GoA). Due to its polar nature, it is unusual to find a Y residue in the middle of TM domains of monomeric proteins (L.B., unpublished observation); therefore, we postulate that this side of the {zeta}-TM domain may interact with an alternative component, not CD3. To test this hypothesis, we mutated Y12 to alanine, valine, leucine, and proline (Y->A, Y->V, Y->L, and Y->P) and analyzed the mutants as described previously. Surface expression was quantified (Fig. 4GoB), and, in contrast to the F10 mutants, substitutions of Y12 with amino acids with small side chains was found to impair the TCR surface expression (compare WT to Y->A, Y->V, and Y->L in Fig. 4GoB). As with the F->P mutation (unable to sustain surface TCR expression), the mutant Y->P completely abolished TCR surface expression, supporting the notion for the requirement of an {alpha}-helix in the TM of {zeta}. The biochemical analysis of mutations at Y12 (Fig. 4GoC) showed that, in contrast with the G13 or F10 mutations, a more general defect on the stability of the TCR complex, even if the {zeta}2-TCR{alpha}ß interaction was more strongly impaired than the {zeta}2-CD3 interaction. In fact, the amount of TCR{alpha}ß recovered in WT or Y->L in the IPs with several Abs was equivalent, while it was absent in the IPs with anti-CD3 or anti-{zeta} of the Y->A or Y->V mutants (compare surface-labeled bands in upper panel of Fig. 4GoC). However, the CD3 complex components (Fig. 4GoC, upper panel) and {zeta}-chain (Fig. 4GoC, lower panel) could still be recovered in the single immunoprecipitation. In the Y->P mutant, the {zeta}-chain was completely absent, suggesting its rapid degradation.

Y12 mutants produced a surface-labeled monomeric form of the {zeta}-chain that was recovered only in IPs with anti-{zeta} mAb. This supported our previous conclusion that the monomeric form of {zeta}-chain cannot associate with the TCR complex ({zeta}-monomer is indicated in the top panel as surface labeled, and as a specific band, lower panel, in Fig. 4GoC).

Conserved TM motif in CD3{gamma} and CD3{delta} are involved in the association to the {zeta}-chain

The fact that the mutation of the {zeta}-dimerization motif impaired the {zeta}2-CD3 suggested a loose interaction of those mutants with the TM domains of CD3. Therefore, we analyzed the TM domains of all known CD3 components. Results for the mouse homologues (mCD3) are presented in Fig. 5GoA. The TM domain of mCD3{epsilon} shares little identity with mCD3{delta} or mCD3{gamma}. This is consistent with CD3 components from other species (data not shown). The comparison of the TM domains of CD3{gamma} and CD3{delta} clearly highlighted a phylogenetically conserved motif of 6 aa (Fig. 5GoB, motif is boxed). However, it was surprising to see that the highest identity was condensed over such a short stretch of amino acids, because the TM domains of CD3{delta} and CD3{gamma} are equivalent and interchangeable for TCR function (12, 43). Because the two small residues within the motif (A and G) locate on opposite sides of the transmembrane negatively charged residue, which is important for the interaction with TCR {alpha}-chain (8, 44), we hypothesized a site of interaction between the {zeta}-chain and both CD3{delta} and CD3{gamma}.



View larger version (26K):
[in this window]
[in a new window]
  		FIGURE 5. Sequence comparison of the TM of mouse CD3{epsilon} with CD3{delta} and {gamma} (A), mouse CD3{delta} and {gamma}, and Xenopus and chicken hybrid CD3{delta}{gamma} (B). Identical residues are visualized with lines, and the putative region of CD3{delta} and {gamma} interacting with the {zeta}-homodimer is boxed.

 
We focused on the two small phylogenetically conserved amino acids within the motif (A and G) because, thanks to their small size, they could be located at contact sites between {alpha}-helices, thereby promoting chain-to-chain contact (e.g., dimerization motif). Because these residues could be part of a pocket, the effect of a nonconservative point mutation should be significant. The mutation of two small residues of CD3{delta} (alanine and glycine) to leucine produced a TM domain motif similar to the one present in CD3{epsilon} (Fig. 6GoA). Reconstitution of the hybridoma BW5147 ({zeta}-/- and CD3{delta}-/-) (21) with the WT {zeta}-chain and the WT or mutated CD3 {delta}-chain tagged with a C-terminal FLAG epitope (WT or CD3{delta}TM AG->LL) clearly showed that the CD3{delta}TM AG->LL was not able to reconstitute TCR surface expression (Fig. 6GoB). The biochemical analysis of the transfectants revealed that the mutant CD3{delta}TM AG->LL could assemble to CD3{epsilon}, but was loosely associated to the TCR {alpha}ß-chains and the {zeta}-homodimer, and similarly to the single transformants (compare the three IPs of the four different cell lines analyzed in Fig. 6GoC). Therefore, in the BW5147 cell line expressing {zeta} and the CD3{delta}TM AG->LL mutant all components needed for expression were present, but the two point mutations impaired the formation of stable complexes and thus TCR surface expression. Because the interaction of CD3{delta} to TCR {alpha}-chain during the TCR complex biogenesis are determined primarily by their EC domains (12), the experiments presented in this section clearly suggest that the motif in the CD3{delta} TM domain is critical for the interaction with the {zeta}-homodimer.



View larger version (33K):
[in this window]
[in a new window]
  		FIGURE 6. The analysis of the CD3{delta} TMAG->LL mutant. A, Alignment of the mutant with the different CD3 TM domains. B, FACS profiles by staining of BW5147 reconstituted with the different constructs with directly labeled anti-CD3 mAb. C, Biochemical analysis of the BW5147 transfectants. The upper panel represents a Western blot with mAb directed against the FLAG epitope present at the very C-terminal part of CD3{delta} WT and mutant, and the lower panel is the same blot probed for the {zeta}-chain.

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
A heterotypic interaction between the {zeta}-chain and any of the single TCR/CD3 subunits could not be detected in transient transfection experiments, suggesting that the {zeta}-chain is involved in a nondefined multicomponent interaction (14).

Sequence comparison of the {zeta}-TM domain with different TM domains revealed a dimerization motif homologous to the TM domain of gp A (36). The section of the motif that is present in the TM of the {zeta}-chain contains a glycine residue that is critical for the gp A dimerization, resulting in SDS-stable homodimers (Fig. 1Go; G13) (35, 37). By site-directed mutagenesis of the corresponding glycine in the TM domain of the {zeta}-chain (G13), we have shown that this position is critical for the homodimerization of the {zeta}-chain (Fig. 1GoB). Mutation of this residue also directly influences the surface expression and stability of the TCR complex. These effects are amplified in the absence of the interchain disulfide bond. The results also show that the {zeta}-chain must be in a dimeric form to transport the TCR complex to the cell surface. This is an interesting observation because the interchain disulfide bridge has been conserved throughout evolution (from chicken to humans) (45), although it is dispensable for the biogenesis of the TCR complex and for its function (11, 17, 18). From the phylogenetical point of view, it may be an advantage to have an interchain disulfide bond to cope with naturally occurring mutations. Although speculative, our experiments may support this view because the presence of the disulfide bridge minimized the effects of TM point mutations on TCR surface expression and function.

Dimerization motif variants can be found in two other molecules of immunological relevance: the TCR interacting molecule (TRIM) and DAP12. TRIM is a disulfide-linked homodimer recruiting intracellular signaling proteins to the plasma membrane (46), while DAP12 represents a disulfide-linked homodimer that is associated to activating NK cell receptors (47). Both molecules contain possible variants of the dimerization motif in their TM domains (GLxxxxGLxxV in TRIM and GIxxGDxxL in DAP12). The 4-aa spacing between the two GL pairs could support the formation of tetrameric complexes as experimentally shown with gp A TM mutants (37).

Some of the mutants presented in this study induced the biochemical segregation of the TCR complex in its modular components: the TCR{alpha}ß-{zeta}2 module segregated from the CD3 complex module in the WT(G) without disulfide bond, G->V and G->L mutants. A complete segregation of the single modules was induced in the G->F mutant (Fig. 2GoB). These results clearly show that the region defined by residues around the dimerization motif are also important for stabilizing the TCR/CD3 complex. Furthermore, they seem to be preferentially involved in an interaction with the CD3 complex rather than with TCR{alpha}ß. The three-dimensional model of the TM domain of the {zeta}-homodimer obtained by homology modeling from the three-dimensional structure of gp A gave insight into the effect that different point mutations could have on structure. The WT {zeta}-TM structure contains an empty cavity defined by the dimerization of the two TM domains (F'10, V'14, I16, and Y20; Fig. 3GoC). This is occupied by the side chains of the mutants (G->V, G->L, and G->F; Fig. 3GoC), which induced conformational changes of phylogenetically conserved bystander residues in the three-dimensional model (F10, F'10, V14, and V'14). Closer analysis of F10 (F'10 as well due to the symmetry of the TM domain) revealed its importance in the TCR biology, because point mutations to smaller amino acids, i.e., increasing the size of the cavity, had a positive effect on TCR surface expression (Fig. 6GoA, and in agreement with Fig. 2Go) without affecting the biochemical stability of the complex (data not shown). These biochemical and structural data supported each other and confirmed the critical role of amino acids surrounding the dimerization motif in the biogenesis of the TCR complex. In contrast, the mutation of the conserved Y12 residue (Fig. 4GoA), which is likely to point toward the membrane, had a negative effect on the TCR complex surface expression, as well as the dimerization of the {zeta}-chain (monomeric form of {zeta} present) (Fig. 4GoC). The mutation of Y12 impaired the interaction with the TCR {alpha}ß-chains and, to a minor extent, with the CD3 complex, suggesting its role in the overall stability of the complex. These results also indicated that residue Y12 was critical for the covalent dimerization of the {zeta}-chain, most likely because the presence of a bulky polar amino acid defines an axial asymmetry of the {alpha}-helix, which, in this specific case, would promote dimerization as a result of loss in entropy in the membrane (e.g., amphiphatic mitochondrial targeting sequences (48)).

Published data focusing on the specificity of CD3 domains during the biogenesis of the TCR complex clearly emphasized the central role of the EC domains, in contrast to the well conserved and exchangeable TM or cytoplasmic domains (12). In this study, a closer analysis of the CD3 TM domains highlighted a 6-aa motif, present in the TM domains of CD3{delta} and CD3{gamma} but not CD3{epsilon}, which has been conserved from Xenopus to human. Because of the composition of the motif (LALGVY) and its location in the lipid bilayer comparable to the dimerization motif, we hypothesized a role in the association to the {zeta}-homodimer. The reconstitution of a CD3{delta}-deficient cell line with a mutated CD3{delta} TM motif (CD3{delta} TM AG->LL, homologous to CD3{epsilon}) was not able to rescue TCR surface expression, even although all complexes could be detected intracellularly (Fig. 6GoC). These results suggested that this section of the TM of CD3{delta}, and by homology, of CD3{gamma}, might interact with the cavity produced by the homodimerization of the {zeta}-chain. The cavity may represent a docking site for a large hydrophobic residue present in the motif within the TM domains of the CD3{delta} and CD3{gamma} complex. The target amino acid on the side of CD3 could be the leucine in the middle of the motif (LALGVY). If this is correct, the occupied pocket in the {zeta}-chain dimerization mutants would weaken the interaction to the CD3 complex but not TCR{alpha}ß, thereby explaining our results (Fig. 2Go).

It has been proposed that the function of the {zeta}-homodimer, besides being a signaling component, is to maintain a tight interaction to CD3, thereby masking their protein kinase C-dependent internalization signals (4, 20, 49, 50). Extrapolating from our three-dimensional model, we think that this type of interaction might reflect a smaller version of a helical interaction described for other systems: a leucine bristle that has been proposed to mediate the assembly of MHC class II molecules (51), the methionine bristle in signal recognition particle-54 interacting with a signal sequences (52), or acidic receptors of the mitochondrial outer membrane (53).

This model gains further support from some recent studies in chicken and Xenopus where CD3 components have been cloned (54, 55, 56). Although a chicken CD3{epsilon} homologue was characterized, only a single CD3{delta}/{gamma} hybrid could be isolated. This shared homology with both CD3{delta} and CD3{gamma} from other species, including the TM motif. In fact, in chicken and Xenopus this CD3-hybrid seems to be the only other CD3 present, apart form CD3{epsilon} (T. Göbel, unpublished observation). The chicken {zeta}-chain, which also contains a {zeta}-dimerization motif, can functionally reconstitute a murine TCR complex lacking endogenous {zeta}-chain (45). We think that this species cross-equivalency is likely to be the result of the phylogenetical conservation of both motifs in the TM domains of {zeta}-chain, CD3{delta}, and CD3{gamma}.

In summary, we would like to propose a model (Fig. 7Go) in which the homodimeric TM domain of the {zeta}-chain is intimately involved in the stabilization of the TCR complex via its dimerization-motif face to CD3{delta} and probably, for symmetrical reasons, CD3{gamma}. At the same time, the Y12 side of the {zeta}-TM domain, together with its EC domain, might interact with TCR {alpha}ß-chains (Fig. 7Go and Ref. 16). In conclusion, this model provides an explanation why the {zeta}-homodimer has never been reported to interact with any single TCR complex components and proposes the {zeta}-homodimer as a "TCR glue."



View larger version (16K):
[in this window]
[in a new window]
  		FIGURE 7. The {zeta}-chain as a TCR "glue." Model summarizing the data and conclusions presented in this manuscript regarding the {zeta}-chain homodimer.

 


   Acknowledgments
 
We thank Drs. T. Göbel, H. Jacobs, K. Karjalainen, E. Palmer, L. Ramage, and S. Stotz for critically reading the manuscript. We thank Dr. T. Göbel for providing data prior publication and discussions. We also thank H. P. Stahlberger, H. Spalinger, and B. Pfeiffer for art work and photography.


   Footnotes
 
1 Address correspondence and reprint requests to Dr. Luca Bolliger and his present address: Hoffmann-La Roche, PRS Unit, 4070 Basel, Switzerland. E-mail address: Back

2 This work was supported by the Basel Institute for Immunology. The Basel Institute for Immunology was founded and is supported by Hoffmann-La Roche, Basel, Switzerland. Back

3 Abbreviations used in this paper: EC, extracellular; TM, transmembrane; gp A, glycophorin A; WT, wild type; IP, immunoprecipitation; m, mouse; TRIM, TCR interacting molecule. Back

Received for publication March 16, 1999. Accepted for publication July 19, 1999.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Klausner, R. D., J. Lippincott-Schwarz, J. S. Bonifacino. 1990. The T cell antigen receptor: insights into organelle biology. Annu. Rev. Cell. Biol. 6:403.
  2. Wegener, A. M. K., F. Letourneur, A. Hoeveler, T. Brocker, F. Luton, B. Malissen. 1992. The T cell receptor/CD3 complex is composed of at least two autonomous transduction modules. Cell 68:83.[Medline]
  3. Reth, M.. 1989. Antigen receptor tail clue. Nature 338:383.[Medline]
  4. Letourneur, F., R. D. Klausner. 1992. Activation of T cells by a tyrosine kinase activation domain in the cytoplasmic tail of CD3{epsilon}. Science 255:79.[Abstract/Free Full Text]
  5. Kearse, K. P., J. L. Roberts, A. Singer. 1995. TCR{alpha}-CD3{delta}{epsilon} association is the initial step in {alpha}ß dimer formation in murine T cells and is limiting in immature CD4+ CD8+ thymocytes. Immunity 2:391.[Medline]
  6. Hall, C., B. Alarcon, J. Sancho, T. Wileman, C. Terhost. 1991. Requirements for cell surface expression of the human TCR/CD3 complex in non-T cells. Int. Immunol. 3:359.[Abstract/Free Full Text]
  7. Manolios, N., J. S. Bonifacino, R. D. Klausner. 1990. Transmembrane helical interactions and the assembly of the T cell receptor complex. Science 249:274.[Abstract/Free Full Text]
  8. Cosson, P., S. P. Lankford, J. S. Bonifacino, R. D. Klausner. 1991. Membrane protein association by potential intramembrane charge pairs. Nature 351:414.[Medline]
  9. Manolios, N., F. Letourneur, J. S. Bonifacino, R. D. Klausner. 1991. Pairwise, cooperative and inhibitory interactions describe the assembly and probable structure of the T-cell antigen receptor. EMBO J. 10:1643.[Medline]
  10. Bonifacino, J. S., P. Cosson, N. Shah, R. D. Klausner. 1991. Role of potentially charged transmembrane residues in targeting proteins for retention and degradation within the endoplasmic reticulum. EMBO J. 10:2783.[Medline]
  11. Rutledge, T., P. Cosson, N. Manolios, J. Bonifacino, R. D. Klausner. 1992. Transmembrane helical interactions: {zeta} chain dimerization and functional association with the T cell antigen receptor. EMBO J. 11:3245.[Medline]
  12. Wegener, A. M. K., X. Hou, J. Dietrich, C. Geisler. 1995. Distinct domains of the CD3{gamma} chain are involved in surface expression and function of the T cell antigen receptor. J. Biol. Chem. 270:4675.[Abstract/Free Full Text]
  13. Dietrich, J., A. Neisig, X. Hou, A. M. K. Wegener, M. Gajhede, C. Geisler. 1996. Role of CD3{gamma} in T cell receptor assembly. J. Cell Biol. 132:299.[Abstract/Free Full Text]
  14. Manolios, N., O. Kemp, Z. G. Li. 1994. The T cell antigen receptor {alpha} and ß chains interact via distinct regions with CD3 chains. Eur. J. Immunol. 24:84.[Medline]
  15. Sussman, J., J. S. Bonifacino, J. Lippincott-Schwartz, A. M. Weissman, T. Saito, R. D. Klausner, J. D. Ashwell. 1988. Failure to synthesize the T cell CD3-{zeta} chain: Structure and function of a partial T cell receptor complex. Cell 52:85.[Medline]
  16. Johansson, B., E. Palmer, L. Bolliger. 1999. The extracellular domain of the {zeta}-chain is essential for TCR function. J. Immunol. 162:878.[Abstract/Free Full Text]
  17. Romeo, C., B. Seed. 1991. Cellular immunity to HIV activated by CD4 fused to T cell or Fc receptor polypeptides. Cell 64:1037.[Medline]
  18. Lanier, L. L., G. Yu, J. H. Phillips. 1991. Analysis of Fc{gamma}RIII (CD16) membrane expression and association with CD3{zeta} and Fc{epsilon}RI-{gamma} by site-directed mutation. J. Immunol. 146:1571.[Abstract]
  19. Kurosaki, T., I. Gander, J. V. Ravetch. 1991. A subunit common to an IgG Fc receptor and the T cell receptor mediates assembly through different interactions. Proc. Natl. Acad. Sci. USA 88:3837.[Abstract/Free Full Text]
  20. Dietrich, J., C. Geisler. 1998. T cell receptor {zeta} allows stable expression of receptors containing the CD3{gamma} leucine-based receptor sorting motif. J. Biol. Chem. 273:26281.[Abstract/Free Full Text]
  21. Letourneur, F., B. Malissen. 1989. Derivation of a T cell hybridoma variant deprived of functional T cell receptor {alpha} and ß chain transcripts reveals a nonfunctional {alpha}-mRNA of BW5147 origin. Eur. J. Immunol. 19:2269.[Medline]
  22. Bolliger, L., B. Johansson, E. Palmer. 1997. The short extracellular domain of the T cell receptor {zeta} chain is involved in assembly and signal transduction. Mol. Immunol. 34:819.[Medline]
  23. Miller, A. D., G. J. Rosman. 1989. Improved retroviral vectors for gene transfer and expression. Biotechniques 7:980.[Medline]
  24. Morgenstern, J. P., H. Land. 1990. Advanced mammalian gene transfer: high titre retroviral vectors with multiple drug selection markers and a complementary helper-free packaging cell line. Nucleic Acids Res. 18:3587.[Abstract/Free Full Text]
  25. Mullen, C. A., M. Kilstrup, R. M. Blaese. 1992. Transfer of the bacterial gene for cytosine deaminase to mammalian cells confers lethal sensitivity to 5-fluorocytosine: a negative selection system. Proc. Natl. Acad. Sci. USA 89:33.[Abstract/Free Full Text]
  26. Pear, W. S., G. P. Nolan, M. L. Scott, D. Baltimore. 1993. Production of high-titer helper-free retroviruses by transient transfection. Proc. Natl. Acad. Sci. USA 90:8392.[Abstract/Free Full Text]
  27. Leo, O., M. Foo, D. H. Sachs, L. E. Samelson, J. A. Bluestone. 1987. Identification of a monoclonal antibody specific for a murine T3 polypeptide. Proc. Natl. Acad. Sci. USA 84:1374.[Abstract/Free Full Text]
  28. Meier, T., S. Arni, S. Malarkannan, M. Poincelet, D. Hoessli. 1992. Immunodetection of biotinylated lymphocyte-surface proteins by enhanced chemiluminescence: a nonradioactive method for cell-surface protein analysis. Anal. Biochem. 204:220.[Medline]
  29. Rozdzial, M. M., R. T. Kubo, S. L. Turner, T. H. Finkel. 1994. Developmental regulation of the TCR {zeta}-chain: differential expression and tyrosine phosphorylation of the TCR {zeta}-chain in resting immature and mature T lymphocytes. J. Immunol. 153:1563.[Abstract]
  30. Guex, N.. 1996. Swiss-PDBviewer: a new fast and easy to use PDB viewer for Macintosh. Experimentia 52:A26.
  31. Sayle, R. A., W. E. Milner. 1995. RASMOL: biomolecular graphics for all. Trends Biochem. Sci. 20:374.[Medline]
  32. Carson, M.. 1991. Ribbons 2.0. J. Appl. Cryst. 24:958.
  33. MacKenzie, K. R., J. H. Prestegrad, D. M. Engelman. 1997. A transmembrane helix dimer: structure and implications. Science 276:131.[Abstract/Free Full Text]
  34. Gerber, P. R., K. Muller. 1995. MAB, a generally applicable molecular force field for structure modelling in medicinal chemistry. J. Comput. Aided Mol. Des. 9:251.[Medline]
  35. Lemmon, M. A., J. M. Flanagan, H. R. Treutlein, J. Zhang, D. M. Engelmean. 1992. Sequence specificity in the dimerization of transmembrane {alpha}-helices. Biochemistry 31:12719.[Medline]
  36. Lemmon, M. A., H. R. Treutlein, P. D. Adams, A. T. Brünger, D. M. Engelman. 1994. A dimerization motif for transmembrane {alpha}-helices. Nat. Struct. Biol. 1:157.[Medline]
  37. Mingarro, I., P. Whitley, M. A. Lemmon, G. VonHeijne. 1996. Ala-scanning mutagenesis of the glycophorin A transmembrane helix: a rapid way to map helix-helix interactions in integral membrane proteins. Prot. Sci. 5:1339.[Abstract]
  38. Mingarro, I., A. Elofsson, G. von Heijne. 1997. Helix-helix packing in a membrane like environment. J. Mol. Biol. 272:633.[Medline]
  39. Langosch, D., B. Brosig, H. Kolmar, H. J. Fritz. 1996. Dimerisation of the glycophorin A transmembrane segment in membranes probed with the ToxR transcription activator. J. Mol. Biol. 263:525.[Medline]
  40. Caspar-Bauguil, S., J. Arnaud, A. Huchenq, W. R. Hein, C. Geisler, B. Rubin. 1994. A highly conserved phenylalanine in the {alpha}ß-T cell receptor (TCR) constant region determines the integrity of TCR/CD3 complexes. Scand. J. Immunol. 40:323.[Medline]
  41. Geisler, C., B. Rubin, B. S. Caspar, E. Champagne, A. Vangsted, X. Hou, M. Gajhede. 1992. Structural mutations of C-domains in members of the Ig superfamily: consequences for the interactions between the T cell antigen receptor and the {zeta}2 homodimer. J. Immunol. 148:3469.[Abstract]
  42. Wiest, D. L., W. H. Burgess, D. McKean, K. Kearse, A. Singer. 1995. The molecular chaperone calnexin is expressed on the surface of immature thymocytes in association with clonotype-independent CD3 compexes. EMBO J. 14:3425.[Medline]
  43. Luton, F., M. Buferne, V. Legendre, E. Chauvet, C. Boyer, A. Schmitt-Verhulst. 1997. Role of CD3{gamma} and CD3{delta} cytoplasmic domains in cytolytic T lymphocyte functions and TCR/CD3 down-modulation. J. Immunol. 158:4162.[Abstract]
  44. Jacobs, H.. 1997. Pre-TCR/CD3 and TCR/CD3 complexes: decamers with differential signaling properties?. Immunol. Today 18:565.[Medline]
  45. Göbel, T. W., L. Bolliger. 1998. The chicken TCR {zeta}-chain restores the function of a mouse T cell hybridoma. J. Immunol. 160:1552.[Abstract/Free Full Text]
  46. Bruyns, E., A. Marie-Cardine, H. Kirchgessner, K. Sagolla, A. Shevchenko, M. Mann, F. Autschbach, A. Bensussan, S. Meuer, B. Schraven. 1998. T cell receptor (TCR) interacting molecule (TRIM), a novel disulfide-linked dimer associated with the TCR-CD3-{zeta} complex, recruits intracellular signaling proteins to the plasma membrane. J. Exp. Med. 188:561.[Abstract/Free Full Text]
  47. Lanier, L. L., B. C. Corliss, J. Wu, C. Leong, J. H. Phillips. 1998. Immunoreceptor DAP-12 bearing a tyrosine-based activation motif is involved in activation of NK cells. Nature 391:703.[Medline]
  48. Von Heijne, G.. 1986. Mitochondrial targeting sequences may form amphiphilic helices. EMBO J. 5:1335.[Medline]
  49. Dietrich, J., X. Hou, A. M. K. Wegener, C. Geisler. 1994. CD3{gamma} contains a phosphoserine-dependent di-leucine motif involved in down-regulation of the TCR. EMBO J. 13:2156.[Medline]
  50. Dietrich, J., J. Kastrup, B. Nielson, N. Odum, C. Geisler. 1997. Regulation and function of the CD3 {gamma} DxxxLL motif: a binding site for adaptor protein-2 in vitro. J. Cell Biol. 138:271.[Abstract/Free Full Text]
  51. Cosson, P., J. Bonifacino. 1992. Role of transmembrane domain interactions in the assembly of class II MHC molecules. Science 258:659.[Abstract/Free Full Text]
  52. Bernstein, H. D.