CD6 is a type I membrane glycoprotein expressed on thymocytes, mature T and B1a lymphocytes, and CNS cells. CD6 binds to activated leukocyte cell adhesion molecule (CD166), and is considered as a costimulatory molecule involved in lymphocyte activation and thymocyte development. Accordingly, CD6 partially associates with the TCR/CD3 complex and colocalizes with it at the center of the mature immunological synapse (IS) on T lymphocytes. However, the signaling pathway used by CD6 is still mostly unknown. The yeast two-hybrid system has allowed us the identification of syntenin-1 as an interacting protein with the cytoplasmic tail of CD6. Syntenin-1 is a PDZ (postsynaptic density protein-95, postsynaptic discs large, and zona occludens-1) domain-containing protein, which functions as an adaptor protein able to bind cytoskeletal proteins and signal transduction effectors. Mutational analyses showed that certain amino acids of the most C-terminal sequence of CD6 (-YDDISAA) and the two postsynaptic density protein-95, postsynaptic discs large, and zona occludens-1 domains of syntenin-1 are relevant to the interaction. Further confirmation of the CD6-syntenin-1 interaction was obtained from pull-down and coimmunoprecipitation assays in mammalian cells. Image analyses also showed that syntenin-1 accumulates at CD6 caps and at the IS. Therefore, we propose that syntenin-1 may function as a scaffolding protein coupling CD6 and most likely other lymphocyte receptors to cytoskeleton and/or signaling effectors during IS maturation.
Human CD6 is a type I glycoprotein of 105–130 kDa expressed on thymocytes, mature T and B1a cells (1), and various brain regions, especially basal glia and cortex cerebellum (2). CD6 belongs to the scavenger receptor cysteine-rich (SRCR) 3 superfamily of protein receptors based on the presence in its extracellular region of three 100- to 110-aa-long cysteine-rich domains characteristic of that family (3, 4). The most membrane-proximal SRCR domain of CD6 is directly involved in the interaction with the CD6 ligand, CD166/activated leukocyte cell adhesion molecule (ALCAM) (5). CD166/ALCAM is found expressed on a great variety of cell types such as activated T and B lymphocytes, lymphomas, activated monocytes, keratinocytes, thymic epitelium, neurons, and brain cortex (6). Because both CD6 and CD166/ALCAM are expressed on thymocytes, the CD6-ALCAM interaction has been implicated in thymocyte adhesion (6, 7, 8) and thymocyte development (9).
Available evidence indicates that CD6 is an accessory molecule capable of delivering costimulatory signals to T lymphocytes (10, 11, 12, 13, 14, 15). However, the precise function of CD6 in lymphocyte activation and differentiation still remains elusive. Recently, it has been demonstrated that on resting T cells CD6 partially associates with CD5 (16), a close member of the SRCR superfamily, and with the TCR/CD3 complex (17). Moreover, CD6 accumulates at the central part of the mature immunological synapse (IS), where it colocalizes with TCR/CD3 and CD5 (17). More importantly, CD6 has been implicated in early T cell-APC contacts influencing IS maturation (17) and further T cell proliferative responses (17, 18). In B cells, it has been reported that CD6 ligation protects from anti-IgM-mediated apoptosis through bcl-2 induction (19).
The signaling pathway used by CD6 to influence T and B cell activation and maturation is mostly unknown. CD6 has a long intracytoplasmic region devoid of intrinsic activity, but containing several consensus sequences related to signal transduction (20). CD6 is constitutively phosphorylated, and it is hyperphosphorylated by serum and protein kinase C (PKC) activators, which causes a weight shift from 105 to 130 kDa (12, 21). CD6 also becomes transiently tyrosine phosphorylated upon CD3 stimulation either alone or by cocross-linking with CD2 or CD4 (22, 23). Only the two most C-terminal tyrosine residues (Tyr629 and Tyr662) on the cytoplasmic tail of CD6 seem to be critical for its TCR-induced tyrosine phosphorylation (23). It has been described that the CD6-mediated effects on T cell proliferation involve tyrosine kinase activity, which is dependent on PKC activation (24).
To identify molecular partners of CD6 involved in its signaling pathway, we performed yeast two-hybrid screening using the cytoplasmic region of CD6 as a bait. We found that syntenin-1 is an intracellular binding partner for the C-terminal region of CD6, and provide evidence for its colocalization not only at CD6 caps, but also at the IS. Syntenin-1 is a 32-kDa scaffolding protein containing two postsynaptic density protein-95, discs large, and zona occludens-1 domains (PDZ), which are multifunctional protein-binding modules first identified in PDZ proteins (25). PDZ family proteins are involved in the localization of receptors and cytosolic effectors to specific membrane sites and in linking extracellular signals to the cytoskeleton and intracellular signaling pathways. Syntenin-1 was originally identified as a syndecan-binding PDZ protein (26), and later on has been shown to interact with about one dozen proteins, most of which are membrane receptors (27). Interestingly, syntenin-1 has been found to interact with both pre- and postsynaptic neuronal receptors (28, 29, 30, 31). Therefore, our data open the possibility that syntenin-1 may be relevant not only to neuronal, but also IS physiology.
Materials and Methods
Chemicals and Abs
PBL were obtained from buffy coats by centrifugation on a standard Ficoll gradient (d = 1077). COS-7 cells were obtained from the American Tissue Culture Collection. The mouse anti-CD6 mAb 161.8 (IgG1) and anti-HLA-DR Edu-1 (IgG2a) were produced in our laboratory by R. Vilella (Hospital Clinic, Barcelona, Spain), and were assigned during the Sixth International Workshop on Leukocyte Antigens (32). The rabbit polyclonal antiserum against the extracellular region of human CD6 region was produced in our laboratory by immunization with four fortnightly i.m. injections (50 μg each) of a recombinant soluble form of CD6 (17) in CFA (first) and IFA (next) (Invitrogen Life Technologies). The mouse anti-syntenin mAbs 4D12 and 4F6 and the rabbit polyclonal anti-syntenin antiserum have been reported elsewhere (33). The anti-hemagglutinin (HA) mAb HA.11 was from Babco, and the biotin-labeled anti-HA mAb was from Santa Cruz Biotechnology. The FITC-labeled anti-CD6 (M-T605, IgG1) and anti-CD3 (UCHT1, IgG1) mAbs were from BD Pharmingen. The HRP-labeled streptavidin and the HRP-labeled goat anti-rabbit (GARIg) and anti-mouse Ig (GAMIg) antisera were from Amersham Biosciences. The FITC-labeled rabbit anti-mouse Ig was from Sigma-Aldrich, and the cyanine 3 (Cy3)-labeled GARIg was from Jackson ImmunoResearch Laboratories. The Alexa 488-labeled GARIg was from Molecular Probes. Staphylococcal enterotoxin E (SEE) was from Toxin Technology.
The CD6 construct E1213 (coding from A613 to the natural stop codon) was obtained by PCR amplification with the sense 5′-TAACGTCGACTGCAGGGCCCCCGGCTGATGAC-3′ and antisense 5′-AAGAATGCGGCCGCTAGGCTGCGCTGATGTCATCGT-3′ oligonucleotides, using the pHβ-CD6.wt expression construct (16) as a template. The PCR product was cloned into SalI/NotI-restricted pPC62 vector (BD Clontech) (34) as in-frame fusion with the Gal4 binding domain (BD). The CD6 construct E8910 (from the A431 to E527) was obtained in a similar way using the sense 5′-AGCAGTCGACAGCCCTCCCCGTAATGGTG-3′ and antisense 5′-AAGAATGCGGCCGCTATTCCTCCAAGGGTGGCAT-3′ oligonucleotides. The deletion and point-mutation derivatives of the E1213 construct were generated by oligo-directed mutagenesis using the sense 5′-TAACGTCGACTGCAGGGCCCCCGGCTGATGAC-3′ oligonucleotide in combination with one of the following antisense oligonucleotides: 5′-TAATGCGGCCGCTAGGATGCGCTGATGTCATCGT-3′ (A668S), 5′-TAATGCGGCCGCTAGGCTGAGCTGATGTCATCGT-3′ (A667S), 5′-AAGAATGCGGCCGCTAGGCTGCTGCGATGTCATCGT-3′ (S666A), 5′-AAGAATGCGGCCGCTAGGCTGCGCTGGCGTCATCGT-3′ (I665A), 5′-AAGAATGCGGCCGCTAGGCTGCGCTGATGGCATCGT-3′ (D664A), 5′-AAGAATGCGGCCGCTAGGCTGCGCTGATGTCAGCGTAG-3′ (D663A), 5′-AAGAATGCGGCCGCTAGGCTGCGCTGATGTCATCGGCGTC-3′ (Y662A), and 5′-AAGAATGCGGCCGCTAGTCATCGTAGTCATCGTTGTC-3′ (Δ4). The pPC62-CD5cy construct (from Y378 to natural stop codon) was obtained, as previously reported (35).
The full-length mouse syntenin-1 cDNA was isolated from the two-hybrid assay screen as in-frame fusion with the Gal4 activation domain (AD) cloned into pPC86 (syntenin.wt). The syntenin.PDZ1 construct (from M1 to R193) was obtained by PCR amplification using the sense 5′-ATTAGTCGACATGTCTCTTTATCCATCTCTTGA-3′ and antisense 5′-AGAATGCGGCCGCTTACCTGTCACGGATCGTCATG-3′ oligonucleotides, followed by further cloning into SalI/NotI-restricted pPC86 vector. The syntenin.PDZ2 construct was obtained by deleting the PDZ1 domain by PCR overlap extension of two cDNA fragments, one coding from M1 to K109 and amplified with the 5′-ATTAGTCGACATGTCTCTTTATCCATCTCTTGA-3′ and 5′-ACTGTCCGTTGAAAGGGCTTAATCTTCGTTCTCCG-3′ oligonucleotides, and another coding from P194 to the natural stop codon and amplified with the 5′-CGGAGAACGAAGATTAAGCCCTTTGAACGGACAGT-3′ and 5′-GAATGCGGCCGCTTAAACTTCAGGAATGGTGTGAT-3′ oligonucleotides. The two cDNA fragments were mixed and subjected to PCR with the above underlined primer pair. The resulting syntenin.PDZ2 construct was further cloned into the SalI/NotI-restricted pPC86 vector.
The constructs coding for HA-tagged syntenin (HA-syntenin.wt) and PDZ1 (HA-syntenin.PDZ1) were obtained by PCR amplification with the sense 5′-ATGAATTCATGTCTCTTTATCCATCTCTT-3′ oligonucleotide and the antisense 5′-ATACTCGAGTTAAACTTCAGGAATGGTGTG-3′ and 5′-ATACTCGAGTTACCTGTCACGGATCGTCAT-3′ oligonucleotides, respectively, using pPC86-syntenin.wt as a template. The two products were then cloned into EcoRI/XhoI-restricted pMT2-HA vector. The HA-syntenin.PDZ2 construct lacking PDZ1 was obtained by PCR with the 5′-GGGATCCGTGAAGTTATTATGCATAAGGACAGC-3′ and 5′-GCTGTCCTTATGCATAATAACAATAACTTCACGGATCCC-3′ oligonucleotide pair.
The pHβ-CD6.Δ4, pHβ-CD6.P527stop, and pHβ-CD6.V463stop constructs were generated by cloning into EcoRI/BamHI-restricted pHβ-CD6.wt construct (16) the PCR products obtained with the sense 5′-GTCACTATAGAATCTTCTGTG-3′ oligonucleotide and the antisense 5′-AAAGGATCCCTAGTCATCGTAGTCATCGTTGTC-3′, 5′-TTTGGATCCCTAGGGTGGCATCTGGAACCTG-3′, and 5′-GTTGGATCCTTAAACTTCTTTGGGGATGGTGAT-3′ oligonucleotides, respectively.
Yeast two-hybrid library screening and direct interaction tests
The cDNA library obtained from the IgG2a-expressing K46 mouse B cell lymphoma (36) was kindly provided by K. Campbell (Fox Chase Cancer Center, Philadelphia, PA). This cDNA library was directionally cloned into pPC86 (34) as in-frame fusion with the Gal4 AD. The pPC86-cDNA library was cotransformed with the pPC62-E1213 construct as a bait into the yeast strain HF7c (37), with integrated growth selection reporter gene HIS3, using the lithium acetate procedure (38). Positive clones were isolated by growth on medium lacking histidine (-H), and the cDNA plasmids were released, sequenced, and retrotransformed with pPC62 constructs for the E8910, E1213, or CD5cy inserts. For direct interaction tests, the HF7c yeasts were cotransformed with the indicated constructions and analyzed by growth in -H selective medium.
GST protein production
The GST-syntenin.ΔNt fusion protein was obtained by cloning the EcoRI-NotI fragment from the pPC86-Syntenin.wt construct into appropriately restricted pGEX-5X3 vector. Expression of GST fusion proteins and their immobilization to glutathione-Sepharose 4B beads were performed following manufacturer’s instructions (Amersham Biosciences).
COS-7 cells were transiently transfected with the indicated constructs using Lipofectamine 2000 (Invitrogen Life Technologies) following manufacturer’s instructions. After 24 h, cells were solubilized for 30 min on ice in a lysis buffer containing 10 mM Tris-HCl, pH 7.6, 140 mM NaCl, 5 mM EDTA, 140 mM NaF, 0.4 mM orthovanadate, 5 mM pyrophosphate, 1 mM PMSF, Complete protease inhibitor mixture tablets (Roche Diagnostics), and 1% Nonidet P-40 (Roche Diagnostics).
The J77 and CD6-deficient 2G5 Jurkat cells were stably transfected with the pHβ-CD6.wt, pHβ-CD6.P527stop, and pHβ-CD6.V463stop constructs, and further selected with Geneticin (G418), as described (39).
Samples of Nonidet P-40 detergent solubilizates from COS-7 cells transfected with pHβ-CD6.wt were incubated for 2 h at 4°C with equal amounts of GST alone or GST-Syntenin.ΔNt immobilized to glutathione-Sepharose 4B beads (Amersham Biosciences). Precipitates were washed three times with 1 ml of 1% Nonidet P-40 lysis buffer and run into 10% SDS-PAGE. Samples (30 μl) of the lysates were run as transfection control. Proteins were transferred (at 0.4 A, 100 V for 1 h) to nitrocellulose membranes (Bio-Rad). Filters were blocked for 30 min at 37°C with blocking solution (5% nonfat milk powder in PBS), and then incubated for 30 min at room temperature with a 1/1000 dilution of rabbit polyclonal antiserum raised to the intracytoplasmic region of human CD6 (16) in blocking solution. After three washes with PBS plus 0.1% Tween 20, the membranes were incubated with HRP-labeled GARIg and washed again three times. Membranes were developed by chemiluminescence with SuperSignal West Dura Extended Duration Substrate (Pierce) and exposure to X-OMAT films (Kodak) or LAS3000 luminiscent imager (Fuji Photo Film).
Brij 58 and Nonidet P-40 detergent solubilizates from either PBL or COS-7 cells cotransfected with the indicated constructs were immunoprecipitated for 2 h at 4°C with specific mAb plus 20 μl of 50% protein A-Sepharose CL-4B beads (Amersham Biosciences). Precipitates and 30-μl samples from each cell lysate were run in 10% SDS-PAGE and transferred to nitrocellulose membranes, as indicated above. Membranes from COS-7 cells were blotted with either anti-HA mAb HA.11 plus HRP-labeled GAMIg or biotin-labeled anti-HA mAb plus HRP-labeled streptavidin. Once developed, the membranes were subjected to stripping for 5 min at room temperature with 200 mM NaOH and blotted again with rabbit polyclonal antiserum to the intracytoplasmic region of CD6 plus HRP-labeled GARIg. Membranes from PBL lysates were split in two sections: the upper section was blotted with rabbit polyclonal antiserum to the cytoplasmic region of CD6 plus HRP-labeled GARIg, and the bottom section with rabbit polyclonal anti-syntenin antiserum plus HRP-labeled GARIg. The membranes were developed by chemiluminescence, as indicated above.
Double immunofluorescence studies
PBL and 2G5 Jurkat transfectants stably expressing the pHβ-CD6.P527stop and pHβ-CD6.V463stop constructs were incubated for 10 min at 4°C with saturating amounts of anti-CD6 mAb (161.8) plus FITC- or Alexa 488-labeled GAMIg. The cells were washed once with PBS and then either kept on ice or incubated at 37°C for 30 min. CD6 capping induction was stopped with ice-cool PBS. Cells were plated onto coverslips pretreated with poly-l-lysine (PLL) (Sigma-Aldrich), fixed for 10 min with 2% paraformaldehyde, and stopped with TBS. Later, coverslips were permeabilized for 1 min with 1% Triton X-100 (Merck). Syntenin-1 was detected with a rabbit polyclonal anti-syntenin antiserum plus Cy3-labeled GAMIg. The images were obtained with a Leica TCS-SP confocal scanning laser microscope (Leica Microsystems) and analyzed with the Image Processing Leica Confocal Software and Photoshop 4.0 (Adobe Systems).
T-B cell conjugate formation
The use of the human Vβ8 TCR-expressing J77c120 Jurkat cell line and the human B cell line Raji for T-B cell conjugate formation has been reported elsewhere (40). Raji cells were incubated with the fluorescent cell tracker chloromethyl derivative of aminocoumarin (Molecular Probes), washed twice with RPMI 1640 medium, and incubated for 20 min at 37°C with 2 μg/ml SEE. Raji cells were mixed with J77c120 cells (5 × 104) stably expressing CD6.wt and then incubated for 15 min at 37°C. The cell mixture was plated onto PLL-coated slides, and incubated for 10 min at 37°C. Next, cells were fixed, blocked with saturating amounts of human polyclonal IgG1 (Sigma-Aldrich), and stained with the following Ab combinations: FITC-labeled anti-CD3 mAb and rabbit polyclonal antiserum to the extracellular region of CD6 plus Cy3-GARIg; FITC-labeled anti-CD3 mAb and rabbit polyclonal anti-syntenin antiserum plus Cy3-GARIg; and FITC-labeled anti-CD6 mAb with rabbit polyclonal anti-syntenin antiserum plus Cy3-labeled GARIg. Before syntenin staining, cell conjugates were permeabilized as above. Series of optical sections of J77-Raji conjugates were obtained with a Leica TCS-SP confocal scanning laser microscope analyzed with the Image Processing Leica Confocal Software and Photoshop 4.0 (Adobe Systems). Differential interference contrast images were also obtained.
Yeast two-hybrid screening for CD6-interacting proteins
To identify conserved CD6-interacting proteins involved in signal transduction, we performed a two-hybrid screening using a mouse B cell cDNA library fused to the AD of the Gal4 transcription factor as a prey (36), and the cytoplasmic region of human CD6 fused to the BD of Gal4 as a bait. The whole intracellular region of CD6 (from aa A431 to the natural stop codon) could not be used as a bait because of spontaneously induced trans activation of the His reporter gene (data not shown). The cDNA coding for C-terminal region of CD6, from the residue A613 to the natural stop codon, did not show such an activity and was finally used as a bait (pPC62-E12E13). This C-terminal region is highly conserved among humans and mice (Fig. 1) and has been reported to contain the two only CD6 tyrosine residues phosphorylatable upon CD3-cross-linking (23), as well as several consensus motifs for serine/threonine phosphorylation and for binding of Src homology 3- and Src homology 2-containing proteins. A positive clone containing an insert corresponding to the full-length syntenin-1 cDNA was identified. Syntenin-1 is a scaffolding intracellular protein characterized by the presence of a tandem of PDZ domains (PDZ1 and PDZ2) flanked by N- and C-terminal fragments. Syntenin-1 is a highly conserved protein (95% of amino acid similarity between mice and humans), being the PDZ domains, the region involved in most of the reported interactions for syntenin-1, where this conservation reaches its highest value (99% of amino acid similarity) (Fig. 1).
The specificity of the CD6-syntenin interaction was probed by direct two-hybrid assays, in which the full-length cDNA of mouse syntenin-1 cloned into pPC86 (pPC86-syntenin.wt) was cotransformed with different CD6 and CD5 cDNA constructs cloned into pPC62 (pPC62-E1213, pPC62-E8910, pPC62-CD5cy). As shown in Fig. 2, yeast growth in the -H was detected when using the C-terminal region of human CD6 (E1213), but not a more membrane-proximal region of CD6 encompassing from 431 to 527 aa (E8910). Similar negative growth results were obtained for the whole cytoplasmic region of human CD5 (CD5cy), encompassing from amino acid Y378 to the natural stop codon.
Mapping of the CD6-syntenin-1 interaction
Most of the protein interactions reported for syntenin-1 are mediated through its PDZ domains, which bind to short hydrophobic C-terminal amino acid motifs of the target proteins (26). The PDZ domains of syntenin-1 may function in a cooperative fashion and exhibit degenerate specificity because they may bind to several C-terminal peptide sequences (27, 41). Therefore, we investigated whether the hydrophobic C-terminal amino acid sequence present in human CD6 (-ISAA) is responsible for the interaction with syntenin-1, as well as which PDZ domain of syntenin-1 is involved on it. Direct two-hybrid yeast assays were performed in which several mutant variants of the pPC62-E1213 and pPC86-syntenin.wt constructs were used (Fig. 3, A and B, respectively). As illustrated by Fig. 3,C, the ability to growth in -H was prevented when yeasts were cotransformed with wild-type syntenin-1 (pPC86-syntenin.wt) and a deletion-mutant CD6 construct (pPC62-E1213.Δ4) lacking the last 4 C-terminal aa (-ISAA). To further analyze which amino acids were relevant to the interaction, we similarly tested different point-mutation variants of the pPC62-E1213 construct (A668S, A667S, S666A, I665A, D664A, D663A, Y662A) (Fig. 3,A). The results presented in Fig. 3,C show that mutation of any of the last four C-terminal residues of CD6 (A668S, A667S, S666A, and I665A), as well as of Y662 impaired the interaction with syntenin-1. On the contrary, the mutation of amino acids D664 and D663 neutrally affected the interaction (Fig. 3 C). These results indicate that the -Y-X-X-I-S-A-A sequence is involved in the binding of CD6 to syntenin-1.
To investigate which PDZ domain(s) of syntenin-1 is involved in the interaction with CD6, yeast were cotransformed with pPC62-E1213 and deletion-mutant variants of syntenin-1 devoid of either PDZ1 or PDZ2 domain (pPC86-syntenin.PDZ2 and pPC86-syntenin.PDZ1, respectively) (Fig. 3,B). From the growth analysis in -H shown in Fig. 3 D, it can be concluded that the simultaneous presence of the two PDZ domains of syntenin-1 is necessary for the interaction to occur. This requirement for the complete tandem is reminiscent of that reported for the interaction of syntenin-1 with other transmembrane proteins (26, 30, 42, 43, 44).
Pull-down and coimmunoprecipitation analyses of the CD6-syntenin-1 interaction in mammalian cells
To further confirm the CD6-syntenin-1 interaction in a mammalian cell system, we first performed pull-down experiments in COS-7 cells. In these experiments, COS-7 cells were transiently transfected with an expression construct coding for the full-length cDNA of human CD6 (pHβ-CD6.wt). Samples of detergent solubilizates from these cells were then incubated with glutathione-Sepharose beads coupled to GST either alone or fused in-frame to mouse syntenin-1 devoid of its N-terminal region (GST-syntenin.ΔNt), and the resulting precipitates were blotted with a polyclonal antiserum raised to the intracytoplasmic region of CD6. As shown in Fig. 4 A, the presence of CD6 was detected in GST-syntenin.ΔNt, but not in GST precipitates, thus indicating both the specificity of the CD6-syntenin-1 interaction and the dispensability of the N terminus of syntenin-1 for it.
Next, we wanted to confirm whether the -ISAA sequence of CD6 cytoplasmic region functions as a syntenin-1 BD in mammalian cells. With this aim, COS-7 cells were cotransfected with an expression construct for HA-tagged full-length mouse syntenin-1 (HA-syntenin) in conjunction with expression constructs for either full-length human CD6 (CD6.wt) or two cytoplasmic tail-truncated CD6 forms, one lacking the -ISAA sequence (CD6.Δ4) and another lacking the 141 most C-terminal residues of CD6 (CD6.P527stop). As shown in Fig. 4 B, the presence of HA-syntenin-1 was detected in CD6 immunoprecipitates from COS-7 cells cotransfected with CD6.wt, but not CD6.Δ4 or CD6-P527stop. These data confirm that the last 4 aa (-ISAA) of CD6 are needed for its interaction with syntenin-1.
The requirement for integrity of the tandem of PDZ was also investigated in COS-7 cells. These cells were cotransfected with expression constructs for full-length human CD6 (CD6.wt) and HA-tagged full-length mouse syntenin-1 or its individual PDZ domains (HA-syntenin, HA-PDZ1, or HA-PDZ2, respectively). As illustrated in Fig. 4 C, the presence of HA-tagged proteins in CD6 immunoprecipitates was only detected when COS-7 cells were cotransfected with full-length CD6 and syntenin-1 cDNAs. Neither HA-tagged PDZ1 nor PDZ2 could be detected in CD6 immunoprecipitates, thus indicating that both PDZ domains in tandem are needed for the interaction of syntenin-1 with the intracytoplasmic region of CD6.
The interaction between CD6 and syntenin-1 was further investigated by coimmunoprecipitation studies of normal untransfected human lymphocytes. To this purpose, Brij 58 detergent solubilizates from human PBL were subjected to immunoprecipitation with mouse mAbs to HLA-DR, CD6, or syntenin-1. The resulting immunoprecipitates were analyzed for the presence of human syntenin-1 and CD6 by Western blot using specific rabbit polyclonal antisera. As shown in Fig. 4 D, a band of 32 kDa corresponding to syntenin-1 was detected in CD6 immunoprecipitates. A similar band was detected in syntenin-1, but not HLA-DR immunoprecipitates used as positive and negative control, respectively. These results indicate that the interaction between CD6 and syntenin-1 detected in yeast and COS-7 cells may be of physiological relevance for normal human lymphocytes. The lack of detection of CD6 in syntenin-1 immunoprecipitates does not detract from this conclusion and may indicate that multiple protein interactions are displayed by syntenin-1 in human lymphocytes.
Colocalization studies of CD6 and syntenin-1 in human lymphocytes
Functional protein interactions require physical colocalization between the interacting proteins. Thus, we studied the normal distribution of CD6 and syntenin-1 and their possible colocalization in normal T lymphocytes by using double immunofluorescence assays. To this purpose, PBL were doubly stained with mouse anti-CD6 mAb plus FITC-GAMIg and rabbit polyclonal anti-syntenin-1 serum plus Cy3-GARIg, and then analyzed by confocal microscopy. As shown in Fig. 5,A, CD6 almost exclusively distributes along the plasma membrane, while syntenin-1 shows both membrane and cytosolic distribution. This cell distribution pattern is compatible with a partial colocalization of the two proteins at the plasma membrane. To further evidence a putative physical association between CD6 and syntenin-1 on the surface of T lymphocytes, the ability of the two molecules to cocap was analyzed. Cocapping was explored by incubating PBL with an anti-CD6 mAb plus FITC-GAMIg for 30 min at 37°C, followed by fixation and staining for syntenin-1, as detailed above. As illustrated in Fig. 5 A, syntenin-1 systematically accumulated at CD6 caps, thus supporting the partial cocapping of the two proteins. Taken together, the coimmunoprecipitation and cocapping results on intact lymphocytes are compatible with a physical relationship between human syntenin-1 and CD6 under basal and capping conditions.
To analyze whether the presence of syntenin in CD6 caps was exclusively due to its interaction with the C-terminal region of CD6, we performed similar cocapping experiments with 2G5 Jurkat transfectants stably expressing cytoplasmic tail-truncated CD6 forms (CD6.P527stop and CD6.V463stop). As shown in Fig. 5 B, syntenin-1 accumulated at caps resulting from cross-linking of CD6.P527stop and CD6.V463stop molecules, which are devoid of the C-terminal CD6 PDZ-binding motif. These data suggest that syntenin-1 may be a normal constituent of cap structures independently of its association with CD6. In fact, we also detected the accumulation of syntenin at CD43 caps (data not shown). This is not surprising because it is known that syntenin-1 interacts with proteins binding to the actin cytoskeleton (27, 28).
CD6 is recruited to the center of Ag-induced mature IS, where it physically associates and colocalizes with the TCR/CD3 complex and CD5 (16, 17). To further validate the in vivo relevance of the CD6/syntenin-1 interaction, we investigated a possible colocalization of syntenin-1 at the mature IS. To this purpose, J77 Jurkat T cells were stably transfected with a construct coding for wild-type CD6 (pHβ-CD6.wt) to increase its constitutive low-CD6 surface expression. Transfectant J77 cells were incubated for 30 min at 37°C with SEE-loaded Raji B cells, and then cell conjugates were stained for syntenin-1 and analyzed by confocal microscopy. As illustrated by Fig. 6, syntenin-1 systematically accumulates at the contact site between T-B cells, where it partially colocalizes with CD6 and CD3, the latter used as a marker for mature IS. This accumulation is compatible with a putative role for syntenin-1 in the recruitment of CD6 (and most likely other lymphocyte surface receptors) to the IS. This fact, however, does not exclude the possibility that syntenin-1 is also a normal constituent of the IS, playing a general role in the cytoskeleton reorganization that takes place at this relevant cellular structure.
In this study, we provide evidence on the association of CD6 with the scaffolding PDZ-containing protein syntenin-1, as well as on their colocalization at the IS. This represents not only the first reported protein-protein interaction involving the intracytoplasmic region of CD6, but also the first description of the involvement of syntenin-1 in the protein-protein interactions taking place during IS maturation. CD6 is a lymphocyte surface receptor, which mediates cell-to-cell interactions relevant to IS maturation (17) and T cell proliferative responses (17, 18). However, the intracellular interactions responsible for these CD6-mediated effects are mostly unknown to date. By using yeast two-hybrid screens, syntenin-1 has been identified in this study as a CD6-binding partner, and this interaction has been further confirmed by biochemical studies on normal human lymphocytes. These results indicate that syntenin-1 may function as a scaffolding protein coupling CD6 (and very likely other lymphocyte receptors) to cytoskeleton and/or to downstream signaling intermediaries during IS formation and maturation.
Syntenin-1 is a cytosolic 298-residue-long protein, originally identified as a molecule linking syndecan to the cytoskeleton (26). Subsequently, syntenin-1 also has been found to play a role in protein trafficking (44, 45), cell adhesion (33), and transcription factor activation (46, 47). These diverse biological functions reported for syntenin-1 are a result of its interactions with numerous targets (27). Most of the binding partners of syntenin-1 reported to date are membrane proteins, except for the transcription factors Sox4 and eIF5A (46, 47), and the cytosolic actin-binding protein merlin, a member of the protein 4.1 superfamily, which also includes ezrin, moesin, and radixin (27). These observations suggest that syntenin-1 may function as an adaptor-like protein coupling membrane receptors to the cytoskeleton or to cytosolic downstream signal effectors, and this also could be the case for the CD6 lymphocyte surface receptor.
The amino acid sequence of syntenin-1 reveals the presence of a tandem of PDZ domains (PDZ1 and PDZ2) preceded by a N-terminal fragment of 112 aa and followed by 24 C-terminal residues. PDZ domains are ubiquitous signaling domains of 80–90 aa in length, with over 400 distinct copies in the human genome (48, 49). PDZ domains are widespread among signaling and cytoskeletal proteins and are thought to function as molecular scaffolds for the assembly of multiprotein complexes and to localize them to specialized membrane and subcellular compartments (50). Thus, PDZ proteins are pivotal for the correct organization of protein complexes in cell-cell junctions (51), in phototransduction cascades (52), in neuronal synapses (53), and for sorting/targeting proteins to specific subcellular sites (54, 55). PDZ domains generally recognize short sequence motifs present at the last four to five C-terminal residues on their target proteins and are typically grouped into three different classes: class I PDZ domains recognize the motif -X-(S/T)-X-φ, class II PDZ domains the motif -X-φ-X-φ, and class III PDZ domains the motif D/E-X-φ (φ represents a bulky hydrophobic amino acid, and X any amino acid) (49). An additional class of PDZ domains would also recognize the motif -X-X-C (56). Examples outside this paradigm are well documented, and some PDZ domains show degenerate specificity (57). In the case of syntenin-1, there are putative binding partners belonging to all classes of target proteins, suggesting that its PDZ domains may also exhibit degenerate specificities. The C-terminal sequence present on human CD6, -ISAA, fits well with the consensus class I PDZ-binding sequence (-X-(S/T)-X-φ). The mutational analysis performed in this study confirmed this assumption, but also revealed the contribution of other nearby residues (-YXXISAA). The analysis showed that the four last C-terminal residues of CD6 together with the tyrosine residue at −7 position (Y662) are relevant to the interaction with syntenin-1, while two aspartic acid residues at position −5 and −6 are irrelevant (-YDDISAA). The contribution of neighboring residues to the binding is not surprising and has been demonstrated for the residues −6 and −7 of merlin (27). Interestingly, Y662 is one of the few tyrosine residues of CD6 that are phosphorylated upon T cell activation (22). This opens the possibility that binding of CD6 to syntenin-1 could be regulated through tyrosine phosphorylation.
It has been reported that the two PDZ domains of syntenin-1 function in a cooperative fashion, this meaning that the integrity of the complete tandem is required for targeting interacting partners (26, 30, 42, 43, 44). Consistent with this, direct two-hybrid assays in yeast and coimmunoprecipitation assays in mammalian transfectants show that the PDZ domains of syntenin-1 bind to CD6 when in tandem, but not when isolated. Nevertheless, recent biophysical binding experiments demonstrate that each PDZ domain of syntenin-1 is able to target different proteins, and that the properties of the tandem are the sum of the binding properties of the individual PDZ domains (27). This would imply that syntenin-1 could bind to CD6 simultaneously to other binding partners, as it has been proposed for the coupling of syndecan to merlin through syntenin-1 (27). In this case, syntenin-1 would mediate the colocalization of syndecan-1 through PDZ2 and merlin through PDZ1. Alternatively, it has been reported that the N-terminal region of syntenin-1 is responsible for its interaction with Sox-4 and eIF5A (46, 47). Thus, syntenin-1 could bind to CD6 through the tandem of PDZ domains and to other signaling proteins through the N-terminal region.
The localization and clustering of cell surface receptors to specific subcellular positions can be critical for their proper functioning in the sending and receiving of signals. It has been described that syntenin-1 accumulates at cell-cell contacts between epithelial cells, colocalizing with F-actin and β-catenin (33). This suggests a role for syntenin-1 in cell adhesion, cytoskeletal dynamics, and cell polarity, and pointing syntenin-1 as an adaptor protein binding transmembrane proteins to the cell actin cytoskeleton (26, 27). Accordingly, the results presented in this work demonstrate the accumulation of syntenin-1 at two highly ordered structures, the caps and the IS, in which settlement and arrangement of the submembranous actin cytoskeleton play an important role (58, 59). The cap structures consist of the assembly of receptors and signaling molecules involved in lymphocyte activation, which resembles that of supramolecular activation clusters at the interfaces of physical contact between T cells and APCs, also known as the IS (58). Although the biological significance of caps is questionable, that of the supramolecular activation clusters is indubitable.
The presence and accumulation of syntenin-1 at the IS are new, but not surprising, given the close relationship existing between the neural and the IS (60). During the last years, syntenin-1 has emerged as a protein involved in neural synaptic regulation through its binding to a number of pre- and postsynaptic membrane proteins such as merlin/neurofibromatosis type 2 protein/schwannomin (28), B class ephrins (29), neurofascin (31), glycine transporter 2, and syntaxin1A (30), as well as 2-amino-5-hydroxy-5-methyl-4-isoxazolepropion acid, kainite, and metabotropic glutamate receptors (61, 62, 63). Apart from syntenin-1, other PDZ-containing proteins such as protein interacting with PKC (63), the membrane-associated guanylate kinases calcium, calmodulin-associated serine/threonine kinase and discs large (64), the serine/threonine kinase Unc5.1 (65), and the adaptor proteins glutamate receptor-interacting protein and ephrin-interacting protein (29) are important for the synaptic trafficking and clustering of receptors. Regarding the IS, the presence of syntenin-1 has not been reported previously and will deserve further studies on its functional meaning. The only indirect evidence for its possible relevance at the IS comes from the finding that CD148/protein tyrosine phosphatase-η, an integral-membrane protein tyrosine phosphatase that is excluded from the IS and down-regulates prolonged T cell signaling (66), has been found to associate with syntenin-1 (43). As reported for the neural synapse, other PDZ-containing proteins distinct from syntenin-1 may play a role at the IS. Indeed, it has been reported recently that discs large is recruited to cortical actin in T cell, forms complexes with early participants in the signaling process, and functions as a negative regulator of T cell activation (67).
In conclusion, we have identified syntenin-1 as the first intracellular partner reported for human CD6, a lymphocyte surface receptor involved in proper IS maturation (17) and T cell proliferation (17, 18). The accumulation of syntenin-1 at the IS recalls that previously reported at the neural synapse and suggests that syntenin-1 (alone or in conjunction with other PDZ-containing proteins) may be relevant for processes that occur not only at the neural, but also the IS.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by grants from Ministerio de Ciencia y Tecnología (SAF 2001-1832 and SAF2004-03251). I.G., M.F., and R.F. are recipients of fellowships from Ministerio de Sanidad y Consumo-Institut D’Investigacions Biomèdiques August Pi i Sunyer (Contrato post-formación sanitaria especializada), Institut D’Investigacions Biomèdiques August Pi i Sunyer, and Ministerio de Ciencia y Tecnologia (FP-2001-0621), respectively.
Abbreviations used in this paper: SRCR, scavenger receptor cysteine rich; AD, activation domain; ALCAM, activated leukocyte cell adhesion molecule; BD, binding domain; Cy3, cyanine 3; GAMIg, goat anti-mouse Ig; GARIg, goat anti-rabbit Ig; -H, histidine-lacking medium; HA, hemagglutinin; IS, immunological synapse; PDZ, postsynaptic density protein-95, discs large, and zona occludens-1 domain; PKC, protein kinase C; PLL, poly-l-lysine; SEE, staphylococcal enterotoxin E.