Ε Phosphorylation of a Common Site in CD3 Domain Binding Via

Reciprocal Regulation of SH3 and SH2 Domain Binding via Tyrosine Phosphorylation of a Common Site in CD3ε

This information is current as Tapio Kesti, Anja Ruppelt, Jing-Huan Wang, Michael Liss, of October 2, 2021. Ralf Wagner, Kjetil Taskén and Kalle Saksela J Immunol 2007; 179:878-885; ; doi: 10.4049/jimmunol.179.2.878 http://www.jimmunol.org/content/179/2/878 Downloaded from

References This article cites 34 articles, 12 of which you can access for free at: http://www.jimmunol.org/content/179/2/878.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on October 2, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2007 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Reciprocal Regulation of SH3 and SH2 Domain Binding via Tyrosine Phosphorylation of a Common Site in CD3␧1

Tapio Kesti,* Anja Ruppelt,† Jing-Huan Wang,‡ Michael Liss,§ Ralf Wagner,§¶ Kjetil Taske´n,† and Kalle Saksela2*‡

Recruitment of cellular signaling proteins by the CD3 polypeptides of the TCR complex mediates T cell activation. We have screened a human Src homology 3 (SH3) domain phage display library for proteins that can bind to the proline-rich region of CD3␧. This screening identified Eps8L1 (epidermal growth factor receptor pathway substrate 8-like 1) together with the N- terminal SH3 domain of Nck1 and Nck2 as its preferred SH3 partners. Studies with recombinant proteins confirmed strong binding of CD3␧ to Eps8L1 and Nck SH3 domains. CD3␧ bound well also to Eps8 and Eps8L3, and modestly to Eps8L2, but not detectably to other SH3 domains tested. Interestingly, binding of Nck and Eps8L1 SH3 domains was mapped to a PxxDY motif Downloaded from that shared its tyrosine residue (Y166) with the ITAM of CD3␧. Phosphorylation of this residue abolished binding of Eps/Nck SH3 domains in peptide spot filter assays, as well as in cells cotransfected with a dominantly active Lck kinase. TCR ligation-induced binding and phosphorylation-dependent loss of binding were also demonstrated between Eps8L1 and endogenous CD3␧ in Jurkat T cells. Thus, phosphorylation of Y166 serves as a molecular switch during T cell activation that determines the capacity of CD3␧ to interact with either SH3 or SH2 domain-containing proteins. The Journal of Immunology, 2007, 179: 878–885. http://www.jimmunol.org/

cells recognize foreign peptides presented by MHC mol- sponses result in the activation of multiple signaling pathways, ecules of APCs. The TCR consists of a ␣␤ TCR het- changes in actin cytoskeleton, and induction of numerous genes. T erodimer (or ␥␦ TCR in ␥␦ T cells) responsible for ligand Although phosphorylation of ITAMs has been regarded as the binding and is associated with the signal-transducing CD3 com- earliest signaling event following TCR triggering, it has been plex composed of heterodimeric CD3␥-CD3␧ and CD3␦-CD3␧ shown that the TCR-CD3 complex undergoes a conformational and homodimer CD3␨ (see Ref. 1). The cytoplasmic tails of CD3 change that occurs even earlier and independently of phosphory- chains include tyrosine- and leucine-containing motifs called lation, which exposes a proline-rich region in the cytoplasmic tail ␥ ␦ ␧ ␧

ITAM (YxxL/I(x6–8)YxxL/I) (2). CD3 , CD3 , and CD3 have of CD3 (5–7). This alteration enables binding of Nck via its first by guest on October 2, 2021 one ITAM each, whereas CD3␨ contains three ITAM. These are SH3 domain (in the following referred to as Nck(I/III)-SH3) (5). phosphorylated upon TCR ligation by the Src family protein ty- This conformational change has been proposed to be essential for rosine kinase Lck (3), generating docking sites for Src homology T cell activation (5), and shown to correlate spatially and tempo- (SH)3 2 domain-containing proteins. Zap70 is the principal CD3 rally with negative selection of T cells (8), but the functional im- ITAM-binding SH2 protein, but other interacting proteins have portance of the proline-rich residues in CD3␧ has been recently also been reported (2). Recruitment of Zap70 leads to its phos- challenged (9). phorylation and activation, further tyrosine phosphorylation, and The SH3 domain is the most common modular protein binding recruitment of other kinases and adapter proteins (4). These re- domain in nature. It binds proline-rich sequences, which often contain the consensus sequence proline-x-x-proline (for reviews, see Refs. 10, 11). SH3 domains are typically found in proteins in- *Department of Virology, Haartman Institute, University of Helsinki and Helsinki volved in signal transduction, membrane trafficking, and cytoskel- University Central Hospital, Helsinki, Finland; †Biotechnology Centre, University of Oslo, Oslo, Norway; ‡Institute of Medical Technology, University of Tampere and etal organization. We have recently generated an essentially com- Tampere University Hospital, Tampere, Finland; §Geneart, Regensburg, Germany; plete collection (n ϭ 296) of human SH3 domains in the form of ¶ and Institute of Medical Microbiology and Hygiene, University of Regensburg, Re- a phage display library, and used this system for identification of gensburg, Germany preferred SH3 partners for different ligand proteins (12). Because Received for publication April 17, 2007. Accepted for publication May 1, 2007. studies by Gil et al. (5) suggested an important role for SH3 bind- 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 ing by CD3 in TCR function, but did not exclude the possibility with 18 U.S.C. Section 1734 solely to indicate this fact. that SH3 proteins other than Nck(I/III) could be involved, we de- 1 This work was supported by grants from the Academy of Finland, Medical Research cided to carry out a comprehensive and unbiased characterization Council of Tampere University Hospital, Medical Research Council of Helsinki Uni- of SH3 binding preferences of CD3␧. versity Hospital, and Sigrid Juselius Foundation (to K.S.), and from the Functional Genomics Program (FUGE), The Research Council of Norway, Norwegian Cancer Society, and Novo Nordic Foundation Committee (to K.T.). 2 Address correspondence and reprint requests to Dr. Kalle Saksela, Department of Materials and Methods Virology, Haartman Institute, University of Helsinki, Haartmaninkatu 3, FIN-0014 Reagents and cell culture Helsinki, Finland. E-mail address: kalle.saksela@helsinki.fi ␧ 3 Abbreviations used in this paper: SH, Src homology domain; TBP, TATA-box The expression constructs for CD3 and Eps8L1 (epidermal growth binding protein; HA, hemagglutinin. factor receptor pathway substrate 8-like 1) were generated from cDNA clones purchased from Deutsches Resourcenzentrum fu¨r Genomfors- Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 chung (CD3␧ (IRATp970D0474D) and Eps8L1 (IMAGp998E1911431Q

www.jimmunol.org The Journal of Immunology 879

and IRALp962L1134Q); RZPD). These cDNA clones and the GST constructs were cloned into pEBB (13), a mammalian expression vector with a strong EF-1 promoter. Plasmid pEBB-Nck-Myc was obtained from B. J. Mayer (University of Connecticut, Farmington, CT) and pLckY505F sfrom T. Mustelin (Burnham Institute, La Jolla, CA). The CD8/CD3␧ chimera consists of the ectodomain and transmembrane domain of human CD8␣ with two tandem copies of the Myc epitope inserted after the signal peptide FIGURE 1. Amino acid sequence of the cytoplasmic tail of human cleavage site, fused to the cytoplasmic tail of human CD3␧. In the Myc- CD3␧. These residues were expressed as a GST-fusion protein as indicated, ␧ ␣ TM-CD3 chimera, the ectodomain of CD8 was deleted, that is, the Myc or similarly fused to maltose-binding protein or the transmembrane and epitopes are between the CD8 signal peptide and transmembrane domain. ectodomains of CD8. The PxxDY motif is underlined and the ITAM ty- Details regarding all plasmid constructs are available upon request. 293FT cells were used for transfections; they were maintained in DMEM high rosine residues are in boldface type. glucose supplemented with 10% FBS and 2 mM glutamine. A standard calcium phosphate precipitation method was used for all transient trans- Analysis of mRNA expression fections. Typically, 2 ␮g of pEBBGST plasmids, 3 ␮g of pLck-DA, and 6 ␮g of plasmid for CD3␧ were used per a 10-cm dish. Immunoprecipita- Total RNA was isolated from 107 cells using GenElute Mammalian Total tions, electrophoresis, and immunoblotting were done by standard methods RNA Miniprep kit (Sigma-Aldrich). A total of 2.5-␮g samples were first (14). Cells were lysed in 1% Nonidet P-40 (14) buffer containing 50 mM treated with RNase-free DNase I (Fermentas UAB) to remove any con- octylglucopyranoside (Sigma-Aldrich). The human SH3 phage library and taminating DNA, and then either reverse-transcribed using random hex- the semiquantitative protein interaction assay have been described (12). amer primers (RevertAid H Minus First Strand cDNA Synthesis kit; Fer- mentas UAB) or mock-treated without reverse transcriptase. PCR primers Abs used were designed with Primer3 program (18) and had a melting temperature of 60°C. The primer sequences were as follows: Eps8 AGGACCAG Abs (with clone) were from the following sources: mouse anti- GAGAGGGTGTTT, TGCCTGCACCACCATATTTA; Eps8L1 TCACT Downloaded from hemagglutinin (HA) (F-7), mouse anti-Myc (9E10), mouse anti- CCACGTGAAAACGAG, CTCCAAGTCTCGGTGACCAT; GAPDH ␧ phosphotyrosine (PY20), and rabbit anti-CD3 (FL-207) from Santa Cruz GAGTCAACGGATTTGGTCGT, TTGATTTTGGAGGGATCTCG; and Biotechnology; rabbit anti-HA (Sigma-Aldrich); mouse anti-GST (GE TATA-box binding protein (TBP) GAATATAATCCCAAGCGGTTTG, Healthcare Bio-Sciences); mouse anti-phosphotyrosine (4G10; Upstate ACTTCACATCACAGCTCCCC. PCR was done using a LightCycler In- ␧ Biotechnology) used in one experiment (see Fig. 8); and mouse anti-CD3 strument (Roche Molecular Biochemicals) with Qiagen QuantiTect SYBR (OKT3; eBioscience). Streptavidin IRDye 800CW, goat anti-mouse IgG Green PCR kit. After initial activation of DNA polymerase at 95°C for 15 IRDye 800CW, and goat anti-rabbit IgG IRDye 680 were from LI-COR min, 50 amplification cycles were conducted as follows: denaturation at http://www.jimmunol.org/ Biosciences. Secondary HRP-conjugated Abs were from DakoCytomation. 94°C for 13 s, annealing at 55°C for 20 s, and extension at 72°C for 12 s. The PCR products were verified by melting curve analysis as well as re- Autospot peptide array striction digestion and gel electrophoresis. Peptide arrays were synthesized on cellulose paper by using MultiPep au- tomated multiple peptide synthesizer (INTAVIS Bioanalytical Instru- Results ments) as described (15). To identify SH3 domain-containing proteins that preferentially bind to human CD3␧, we expressed its intracellular domain as a Overlays GST fusion protein in Escherichia coli (see Fig. 1), and used this recombinant protein for panning of our human SH3 proteome

Interaction of spotted peptides with GST-fusion proteins was tested by by guest on October 2, 2021 overlaying the membranes with 1 ␮g/ml recombinant protein in TBST. phage display library (12). The CD3␧ fusion protein was found to Bound recombinant proteins were detected with anti-GST using a protocol serve as excellent “bait” in this system. It bound 10- to 100-fold where the filters were blocked in 5% nonfat dry milk in TBST for 30 min at room temperature, incubated1hatroom temperature or overnight at 4°C more SH3 phages than did plain GST, as judged by the number of with primary Abs, washed four times 5 min in TBST with 0.1% Tween 20, ampicillin-resistant bacterial colonies obtained by infection of E. and incubated with a HRP-conjugated secondary Ab. Blots were developed coli with phages associated with these proteins after washing (data using Supersignal West Dura Extended Duration Substrate or Supersignal not shown). The identity of the SH3 domains bound to GST-CD3␧ West Pico Chemiluminescent Substrate (Pierce). was determined by sequencing of the SH3 inserts of the selected Lentivirus production and transduction phage clones. The 15 SH3 clones identified after a single round of phage selection all represented Eps8L1 (8/15), Nck1(I/III)-SH3 The full-length Eps8L1 cDNA was fused with a biotin acceptor domain (1/15), or Nck2(I/III)-SH3 (6/15). Thus, these phage selection re- fragment (16) and cloned into pWPIneo, a lentiviral expression vector in which the GFP fragment of pWPI from D. Trono (The Swiss Federal In- sults supported the idea of Nck as an SH3 domain-containing stitute of Technology, Lausanne, Switzerland) was replaced with the G418 CD3␧-interacting protein as proposed by Gil et al. (5), but also resistance gene, to generate pWPIneoEps8L1. This plasmid was cotrans- indicated Eps8L1 as a preferred SH3-containing partner of CD3␧. fected into 293FT cells with pDELTA-8.9 and pVSVg to generate infec- To verify the interactions of these SH3 domains with CD3␧ in tious virus particles (17). Cell culture supernatant was used to infect Jurkat another system, they were produced as GST fusion proteins, and E-6 cells. Cells were grown in RPMI 1640 medium supplemented with ␧ 10% FBS and 2 mM L-glutamine in the presence of G418 at 600 ␮g/ml for tested for binding to the CD3 intracellular tail fused to maltose- 10 days to select for stable integrants. binding protein using a previously described multiwell assay (12). In addition to Eps8L1, Nck1(I/III), and Nck2(I/III), we also in- Cell stimulation and lysis cluded in these studies the SH3 domains of the three other related Jurkat E-6 cells stably expressing Eps8L1 and control cells were grown in Eps8 family proteins (Eps8, Eps8L2, and Eps8L3) as well as sev- RPMI 1640 medium, harvested, incubated in serum-free medium buffered eral unrelated SH3 domains. with 10 mM HEPES (pH 7.4) for 60 min, centrifuged, and suspended in the Fig. 2 shows binding signals obtained with 2-fold dilutions rang- 8 ␧ same medium at 10 cells/ml. Cells were stimulated with anti-CD3 Ab ing from 600 to 75 nM Nck(I/III) and Eps8 family SH3 domains, OKT3 at a concentration of 10 ␮g/ml. Freshly prepared pervanadate was added to a final concentration of 0.1 mM. Then, cells were diluted with 10 Hck-SH3, or plain GST. Apart from Eps8L2-SH3, which showed ml of ice-cold PBS and centrifuged. The 1 ml of lysis buffer (0.3% Brij 97, only weak binding under these conditions, all Nck(I/III)-SH3 and 20 mM Tris (pH 7.8), 2 mM orthovanadate, 1 mM NaF, and protease Eps8-SH3 domains bound well to CD3␧, and gave a significant 8 inhibitors; all from Sigma-Aldrich) was added per 10 cells. After 20 min interaction signal even when tested at 75 nM. In accordance with on ice, samples were centrifuged and supernatants were subjected to avidin pulldown (TetraLink Tetrameric Avidin Resin; Promega) or anti-CD3␧ our phage library screening data, however, Eps8L1-SH3 and immunoprecipitation. Odyssey Infrared Imaging System (LI-COR Bio- Nck(I/III)-SH3 domains showed the highest apparent affinity for sciences) was used for detection and quantification in Western blotting. CD3␧ in this assay. 880 PHOSPHORYLATION REGULATES BINDING OF Eps8L1 AND Nck TO CD3␧

FIGURE 2. Relative binding of CD3␧ to selected SH3 domains. CD3␧ (fused to maltose-binding protein) was immobilized to the bottom of multiwell plates and probed with 2-fold dilutions of different GST-SH3 domain proteins or plain GST as indicated. Specific binding was measured based on absorbance at 405 nm (y-axis) of a chromogenic reaction generated by a glutathione-HRP conjugate. Nck1 and Nck2 refer to the first of the three FIGURE 3. Real-time RT-PCR analysis of expression of Eps8 family SH3 domains in these proteins. genes in Jurkat cells. Total RNA purified from Jurkat T cells and DU145 prostate carcinoma cells was reverse transcribed into cDNA. Then, mRNA expression of Eps8, Eps8L1, and housekeeping genes GAPDH and TBP

was measured using a LightCycler. J, Jurkat cDNA; DU, DU145 cDNA; Downloaded from In contrast to Nck(I/III) and the Eps8 family, none of the other L1, Eps8L1; ϪRT, no reverse transcriptase. SH3 domains that we tested, including Hck, Lck, Src, Lyn, Fyn, Yes, nephrocystin, sorting nexins 9 and 30, endophilin1, Pacsin3, Grb2(I/II), ␣-PIX, ␤-PIX, Nck1(II/III), ArgBP2(III/III), and vin- coprecipitation studies (5). The 293FT cells were transfected with exin(III/III), showed significant binding to CD3␧ even when used expression vectors for HA-tagged CD3␧ together with Myc-tagged at concentrations up to 40 ␮M (data not shown). Each of these SH3 versions of Eps8L1, wild-type Nck, or a mutant form of Nck in http://www.jimmunol.org/ proteins has previously been shown to be functional in the same which a functionally critical tryptophan residue in each SH3 do- assay format (12). In summary, these data show that the intracel- main had been changed to lysine (20). Lysates of the transfected lular tail of CD3␧ is an avid and highly specific ligand for Nck(I/ cells were subjected to anti-Myc immunoprecipitations and ana- III) and Eps8 family (in particular Eps8L1) SH3 domains. lyzed by SDS-PAGE and immunoblotting. As shown in Fig. 4, Because little was known about expression of the Eps8 family CD3␧ was detectable in the immunoprecipitates in which Eps8L1 genes in T cells, we wanted to address this issue. We used the or wild-type Nck were present, but in none of the control samples DU145 prostate carcinoma cell line as a source of control mRNA (lacking CD3␧ or a Myc-tagged protein, or containing SH3-defi- because Eps8 genes, including Eps8L1, have been shown to be cient Nck). Thus, the interaction required functional SH3 domains, expressed in various epithelial cells and in the prostate (19). Initial and under these conditions Eps8L1 bound equally well or better by guest on October 2, 2021 analysis indicated that specific PCR products for Eps8 and Eps8L1 than Nck, consistent with the in vitro binding assays. could be amplified from cDNA prepared from DU145 cells as well It is known that singly expressed CD3 chains are not efficiently as Jurkat and primary human CD3-positive T lymphocytes, but translocated to the plasma membrane due to exposed endoplasmic not from mock cDNA preparations prepared in parallel with these reticulum retention signals (21). However, using a total cell sur- mRNAs without reverse transcriptase (data not shown). To get a face protein biotinylation approach (22) we have confirmed that at better idea of the relative expression levels of these mRNAs we least part of the transfected CD3␧ protein was located at the used quantitative real-time RT-PCR in which expression of two housekeeping genes, highly expressed GAPDH and low-abun- dance TBP, were analyzed as controls. Results from Jurkat and DU145 cells are shown in Fig. 3. As can be seen, Eps8L1 was expressed in similar amounts in both cell lines, whereas Eps8 was expressed at a higher level in DU145 cells. As judged by compar- ison with the GAPDH and TBP RT-PCR signals, expression of Eps8 and Eps8L1 mRNAs in Jurkat cells was not high but still easily detectable and possibly higher than that of TBP mRNA. All PCR signals from the primary human T cell cDNA were lower, but based on dilution of the cell line cDNA, the relative expression levels of Eps8 and Eps8L1 in Jurkat and primary human T cells were found to be very similar (data not shown). Thus, these anal- yses confirmed that mRNA for two members of the Eps8 family that showed strongest SH3-mediated binding to CD3␧ was indeed expressed in human T cells. Extension of these results to the pro- FIGURE 4. Binding of full-length CD3␧ to full-length Eps8L1 and tein level will require development of high quality Abs for immu- Nck1 proteins in transfected cells. Expression vectors for Myc- and HA- nological detection of these proteins. tagged proteins were introduced into 293FT cells as indicated. Transfected cells were lysed and proteins associated with anti-Myc immunocomplexes Next we wanted to verify that interactions involving the corre- were analyzed by SDS-PAGE and immunoblotting as indicated (top and sponding full-length proteins could be observed in live cells. Be- middle panels). Uniform expression of CD3␧ in all cells transfected with ␧ cause Eps8L1 showed the highest affinity to CD3 of all Eps8 CD3␧-HA was confirmed by probing the total lysates with anti-HA Abs family proteins in the initial assays, we focused these studies on (bottom panel). Nckmut is a mutant of Nck1 in which all three SH3 domains Eps8L1. As a control we used an expression vector for full-length have been rendered nonfunctional by changing a critical tryptophan residue Nck1, which has been shown to associate with CD3␧ in previous to lysine (W38K/W143K/W229K, (20)). The Journal of Immunology 881

plasma membrane and exposed to the extracellular space (data not shown). Moreover, we found that both CD3␧ biotinylation and coprecipitation with Eps8L1 were increased when an R183T change (21) was introduced to destroy the endoplasmic reticulum retention signal of the transfected CD3␧. These data support the idea that the interaction of the transfected CD3␧ and Eps8L1 proteins indeed took place at least in part if not exclusively at the plasma membrane. The Eps8 family of SH3 domains has been shown to bind to an atypical motif PxxDY instead of the classical PxxP consensus (23). Indeed, we noticed a PxxDY motif also in the proline-rich intracellular tail (aa 162–166) of CD3␧ (Fig. 1). To map the critical determinants in CD3␧ required for binding of Eps8-SH3 and Nck(I/III)-SH3 domains, progressive N- and C-terminal deletions (from residue 131 to 152 and from 187 to 169, respectively) were initially constructed and expressed as GST fusions in bacteria. Binding of Eps8L1-SH3 and Nck1(I/III)-SH3 domains to all these CD3␧ proteins was similar (data not shown). Although subtle effects on binding by the deleted regions could not be ruled out by this analysis, it indicated the core region 152–169 as the principal Downloaded from binding target for both SH3 domains. Next we synthesized a peptide overlapping this region (residues 155–179) immobilized on cellulose paper together with derivatives substituted with an increasing number of alanine residues starting from the N terminus or the C terminus (Fig. 5A), or with overlapping dialanine substitutions (Fig. 5B). This deletion and substitu- http://www.jimmunol.org/ tion analysis revealed that both Nck and Eps8L1 relied on the sequence PNPDY for binding to CD3␧. Furthermore, binding of Nck appeared somewhat more sensitive to substitutions N- and C-terminal to the PNPDY sequence, indicating that the Nck-SH3 domain requires additional determinants for high affinity binding. A peptide matrix was also made of the region 159–170 in which all natural amino acids were tested at each position, as shown in Fig. 6A. Probing of these peptide arrays with Eps8L1-SH3 iden- by guest on October 2, 2021 tified the peptide VPNPDY as the core binding site, and confirmed the critical role of the residues that form the PxxDY motif. Strik- ingly, probing of the same filters with Nck1(I/III)-SH3 showed a similar absolute dependence in binding for the PxxDY residues. Although the “footprint” of the Nck1(I/III)-SH3 on CD3␧ was larger (PPVPNPDY) the same PxxDY motif residues were essential for Nck-SH3 binding as well. To confirm the apparent requirement of the PxxDY motif for Nck/Eps8L1 binding in vivo, the cytoplasmic tail of CD3␧ was expressed as a CD8/CD3␧ chimera (consisting of a Myc-tagged ectodomain and transmembrane region of CD8 fused to the intracellular domain CD3␧), together with similar constructs in which FIGURE 5. Minimal regions in CD3␧ required for binding of Nck1(I/ Y166 (FY) or Y177 (YF) was changed to phenylalanine. These III)-SH3 and Eps8L1-SH3 domains. A, Membrane-immobilized peptides ␧ vectors were transiently coexpressed with GST-tagged Nck1-SH3 corresponding to residues 155–179 of CD3 , or variants of it with progressive N- or C-terminal alanine substitutions, were probed with GST- or Eps8L1-SH3 domains in 293FT cells. Cells were lysed 44 h fusion proteins containing Nck1(I/III)-SH3 or Eps8L1-SH3 domains as after transfection, and pulldowns were performed with glutathione- indicated. B, Residues 155–174 of CD3␧ were subjected to overlapping Sepharose beads, followed by SDS-PAGE and immunoblotting. dialanine substitutions as indicated, and probed with Nck1(I/III)-SH3 or Samples of unprocessed lysates of the transfected cells were ana- Eps8L1-SH3 as in A. lyzed in parallel. As shown in Fig. 6B, both Nck1-SH3 and Eps8L1-SH3 bound to the wild-type CD3␧ and CD3␧-YF but not to the CD3␧-FY. These data established that Y166, which is an form of Lck (Lck-Y505F) to phosphorylate the tyrosine residues of essential component of a functional PxxDY motif, was indeed nec- the CD3␧ ITAM. In these experiments we used a smaller CD8/ essary for Nck/Eps8 binding also in living cells, whereas the fol- CD3␧ chimera in which almost all of the CD8 ectodomain was lowing tyrosine residue (Y177) was dispensable. substituted with a Myc tag (Myc-TM-CD3␧) because this con- The tyrosine residue of the PxxDY motif (Y166) is shared with struct was more efficiently phosphorylated by Lck than the larger the ITAM element of CD3␧. This compact arrangement of binding CD8/CD3 chimera used in the previous experiments. As shown in motifs points to possible cross-regulation of CD3␧ tyrosine phos- Fig. 7A, coexpression of Lck-Y505F resulted in strong tyrosine phorylation and SH3/SH2 domain binding during T cell activation. phosphorylation of Myc-TM-CD3␧ detected in the lysate (Fig. 7A, To address this issue experimentally, we coexpressed Nck1-SH3 bottom panel), and abolished coprecipitation of Myc-TM-CD3␧ and a CD8/CD3␧ chimera with or without a dominantly active with Nck1-SH3 (Fig. 7A, compare two leftmost lanes). Because 882 PHOSPHORYLATION REGULATES BINDING OF Eps8L1 AND Nck TO CD3␧ Downloaded from http://www.jimmunol.org/

FIGURE 6. Mutation analysis of CD3␧ residues critical for SH3 binding. A, Essential role of a PxxDY motif for binding in vitro. A filter containing a synthetic peptide corresponding to CD3␧ residues 159–170 (single-letter code on top of the filter) and its variants in which all natural amino acids (left of the filter) were systematically tested at each position FIGURE 7. Tyrosine phosphorylation blocks PxxDY-mediated SH3 ␧ ␧ was probed with Nck1(I/III)-SH3 or Eps8L1-SH3 as in Fig. 4. B, An intact binding to CD3 . A, Lck-dependent phosphorylation of CD3 ITAM pre- by guest on October 2, 2021 PxxDY motif is required for coprecipitation from cells. The 293FT cells vents Nck1(I/III)-SH3 binding while enabling Zap70 SH2 binding. CD3␧ were transfected with vectors for GST-tagged Nck1(I/III)-SH3 or Eps8L1- was expressed in 293FT cells as a Myc-TM-CD3␧ chimera together with SH3 domains together with a Myc-tagged CD8/CD3␧ fusion protein (WT), a GST-tagged protein containing Nck1(I/III)-SH3, Hck-SH3, or tandem or mutants of it in which the PxxDY-defining (Y166, FY) or the other Zap70 SH2 domains, with or without a cotransfected dominantly active tyrosine of the CD3␧ ITAM (Y177, YF) had been replaced by a phenyl- Lck kinase (Lck-DA). Following lysis of the cells proteins precipitated by alanine. Lysates of transfected cells were subjected to gluthathione-agarose gluthathione-agarose beads were analyzed by Western blotting with anti- precipitation, and the associated GST- and Myc-tagged proteins were an- GST, anti-Myc, or anti-phosphotyrosine Abs, as indicated. Uniform ex- alyzed by SDS-PAGE and immunoblotting as indicated (top and middle pression of Myc-TM-CD3␧ in all transfected cells, and its tyrosine phos- panels). Whole cell extracts (WCE) were analyzed in parallel for expres- phorylation in cells transfected with Lck-DA was confirmed by probing sion of Myc-tagged CD8/CD3␧ to confirm uniform expression in all trans- total lysates with the corresponding Abs. B, Phosphorylation of tyrosine fected cells (bottom panel). residue 166 accounts for the loss of SH3 binding to CD3␧. Peptides spanning the PxxDY/ITAM region of CD3␧ were synthesized either in unmodified form or with Y166, Y177, or both substituted with a phosphotyrosine similar amounts of total Myc-TM-CD3␧ protein were expressed, residue. Filters were probed with Nck1(I/III)-SH3 and Eps8L1-SH3 do- and similar amounts of GST-tagged Nck1-SH3 were precipitated mains as in Fig. 4. in both cases, we conclude that CD3␧ ITAM phosphorylation negatively regulates binding of Nck1-SH3. ITAM binding of Zap70 is known to be phosphotyrosine-de- lanes). The observation of Nck-SH3 association with CD3␧ only in pendent. Therefore, we also did precipitations with a GST-tagged the absence of Lck phosphorylation was in agreement with our fragment of Zap70 containing its two tandem SH2 domains. As earlier data showing that introduction of a phenylalanine residue in expected, and in a striking contrast to the behavior of Nck1-SH3, place of Y166 in the CD3␧ PxxDY motif abolished binding of the Zap70-SH2 protein associated strongly with Myc-TM-CD3␧ in Nck1-SH3 and Eps8L1-SH3 domain. Indeed, in other experiments the presence, but not in the absence of coexpressed Lck-Y505F we could also show that phosphorylation of the tyrosine residue in (Fig. 7A, compare two middle lanes). The coprecipitated Myc-TM- CD3␧ PxxDY similarly prevented coprecipitation of Eps8L1-SH3 CD3␧ gave a strong signal when probed with an anti- (data not shown). phosphotyrosine Ab (Fig. 7A, bottom). As a negative control we To formally establish that phosphorylation of Y166 (rather than used Hck-SH3 domain, which did not bind to the intracellular tail Y177) in the ITAM was critical, and to confirm that other possible of CD3␧ in our in vitro protein-protein binding assay (see Fig. 2). effects of the coexpressed Lck-Y505F did not contribute to the loss No Myc-TM-CD3␧ could be coprecipitated with Hck-SH3 in the of coprecipitation of Myc-TM-CD3␧ with Nck1-SH3 or Eps8L1- GST pulldown independent of whether or not the CD3␧ ITAM was SH3 domains, peptides spanning the PxxDY-ITAM region of phosphorylated by Lck-Y505F (Fig. 7A, compare two rightmost CD3␧ were synthesized either in unmodified form, or with Y166, The Journal of Immunology 883

down samples (Fig. 8). Typically, the increase was 4- to 8-fold, as determined using the Odyssey Infrared Imaging System. This observation was in good agreement with the earlier conclusion by Alarcoń and colleagues (5) that an activation-induced conformational change is required to fully expose the SH3 binding site in CD3␧. However, supporting our cotransfection experiments with a dominantly active Lck kinase (see Fig. 7), when pervanadate was included to enhance TCR-triggered protein tyrosine phosphorylation, OKT3 stimulation failed to increase the amount of CD3␧ that precipitated with Eps8L1. As expected, control precipitations with avidin-coated beads from lysates of Jurkat cells not expressing biotinylation domain-tagged Eps8L1 contained no detectable CD3␧ regardless of OKT3 or pervanadate treatment of these cells. These results confirmed that transduced Eps8L1 can associate with endogenous CD3␧ in Jurkat T cells. Moreover, these results show that SH3 binding by CD3␧ is under dual regulation by TCR activation. TCR ligation, presumably via an activation-induced FIGURE 8. TCR activation-regulated binding of Eps8L1 to endogenous conformational change, promotes association of Eps8L1 with ␧ CD3 . Lentivirally transduced Jurkat cells expressing Eps8L1 tagged at the CD3␧. However, activation-induced CD3␧ tyrosine phosphoryla- C terminus with a biotin acceptor domain (J/Eps8L1-biotin) and control tion inhibits CD3␧ binding to Eps8L1, thereby providing a possi- Downloaded from Jurkat cells were stimulated with anti-CD3␧ Ab OKT3 in the absence or ble negative feedback regulation mechanism. presence of the phosphatase inhibitor pervanadate. Lysates of these cells were precipitated with avidin-coated beads and the amounts of Eps8L1 and associated CD3␧ were determined by Western blotting (WB). Total lysates Discussion were blotted with Abs against CD3␧ and phosphotyrosine. Odyssey Infra- Our phage display screens and in vitro binding assays shows that red Imaging System was used for detection. Eps8L1 and Nck1(I/III) and Nck2(I/III) are the preferred interaction partners for CD3␧ among human SH3 domains. Except for http://www.jimmunol.org/ Eps8L2, other members of the Eps8 family (Eps8 and Eps8L3) Y177, or both substituted with a phosphotyrosine residue (Fig. also bound well to CD3␧. These data support the results of Gil et al. 7B). Probing of filters containing these peptides with recombinant (5) who identified Nck as an SH3 partner of CD3␧ in a yeast Nck1-SH3 and Eps8L1-SH3 proteins clearly showed that phos- two-hybrid screening of a spleen cDNA library, but indicate the phorylation of Y166 alone or in combination with Y177 abolished Eps8 family proteins as alternative or additional regulators of in- all SH3 binding, whereas phosphorylation of Y177 alone had no tracellular signaling by CD3␧. effect. Thus, although ITAM phosphorylation created binding sites The selectivity of CD3␧ for Nck(I/III) and Eps8 was found to for SH2 domains of Zap70, the same phosphorylation event involve recruitment of these SH3 domains via an atypical PxxDY by guest on October 2, 2021 blocked binding of SH3 domains of Nck1 and Eps8L1, demon- binding motif in CD3␧. This motif has been previously character- strating that the PxxDY-ITAM region of CD3␧ binds SH3 and ized as the minimal consensus sequence for binding of the Eps8 SH2 domains in a mutually exclusive and phosphotyrosine-regu- family (23), but not known to serve as a docking site for Nck or lated manner. other non-Eps8 SH3 domains. Previous studies on binding of The studies described above as well as those reported by another Nck(I/III)-SH3 to CD3␧ (5, 9, 25) did not map the critical CD3␧ group (5) have shown that CD3␧ expressed ectopically in non-T residues involved in this interaction, but implicated a PxxP se- cells is constitutively competent for SH3 binding. However, bind- quence within a cluster of prolines in CD3␧ (see Fig. 1). Al- ing of Nck to endogenous CD3␧ in T cells has been shown to though our results show that the additional proline residues ad- require an activation-induced conformational change in the TCR- jacent to the PxxDY motif do contribute to Nck(I/III)-SH3 CD3 complex (5, 6, 24). Because TCR activation is also known to binding, the absolute requirement of the residues D165 and lead to ITAM phosphorylation (2, 3) it was of interest to study the Y166 in addition to P162 clearly define the PxxDY sequence as overall effect of TCR ligation on SH3 binding to CD3␧ in the the primary docking motif. context of the native TCR-CD3 complex in T cells. Considering the unexpected PxxDY specificity of Nck(I/III), it Because no Abs for immunological detection of native Eps8L1 is of interest to note that Cesareni and colleagues (26) have pro- were available, we generated Jurkat cell lines stably expressing a posed that a positively charged amino acid at position Ϫ2 relative lentivirally transduced Eps8L1 tagged at its C terminus with a to the first highly conserved tryptophan of the WW motif present biotin acceptor domain (J/Eps8L1-biotin), which can be efficiently in most SH3 domains would be important for PxxDY recognition precipitated and detected with avidin-based reagents (16). These by the Eps8-SH3 domain. Indeed, we noted that in addition to the cells and control Jurkat cells were stimulated with anti-CD3␧ Ab four members of the Eps8 family, the N-terminal SH3 of Nck1 and (OKT3) in the absence or presence of the tyrosine phosphatase Nck2 also have a lysine or an arginine residue in this position. inhibitor pervanadate. Aliquots of lysates from these cells were Although this feature is shared by 17 additional SH3 domains examined for their total phosphotyrosine content, and the rest was among the 296 that we have identified from the human genome used for precipitation of Ep8L1 with avidin-coated beads. As ex- (see Ref. 12), all of these differ from Eps8/Nck(I/III) in that they pected, TCR ligation in pervanadate-treated cells resulted in robust contain a negatively charged amino acid neighboring this Lys/Arg increase in phosphorylation of several substrate proteins of TCR- residue or lack the conserved WW motif (data not shown). Thus, activated tyrosine kinases (Fig. 8, bottom). Anti-phosphotyrosine the SH3 sequence X(K,R)XWW (where X can be any amino acid immunoblotting of anti-CD3 immunoprecipitates confirmed that except for E or D) appears to be correlated with PxxDY recogni- CD3␧ was indeed one of these phosphorylated proteins (data not tion. Future structural studies will be required to test the rele- shown). In the absence of pervanadate treatment, OKT3 stimula- vance and possibly explain the molecular basis of this correla- tion increased the amount of CD3␧ in the Eps8L1 (avidin) pull- tion. However, because some proteins that were reported to 884 PHOSPHORYLATION REGULATES BINDING OF Eps8L1 AND Nck TO CD3␧

bind to Nck(I/III)-SH3, such as SAM68 (27), do not contain a point to nanomolar KD values for the strongest ones of these in- PxxDY motif, binding of Nck(I/III)-SH3 may not be exclu- teractions (see Fig. 2). This unusually tight SH3 binding together sively restricted to ligands with a PxxDY motif. with the evolutionary conservation of the Eps8-Nck(I/III) target An intriguing aspect of the PxxDY motif of CD3␧ is that it site in CD3␧ (see below) would seem to argue for an important overlaps with the ITAM of this TCR-CD3 complex subunit. Our role for this interaction. Yet, in an apparent conflict with this idea, data confirmed that phosphorylation of the tyrosine residue (Y166) Szymczak et al. (9) recently reported that the defects in T cell shared by the ITAM and PxxDY motif of CD3␧ is required for development and function of CD3␧-deficient ␧⌬P mice (36) could recruitment of the tandem SH2 cassette of Zap70, which is a cru- be rescued using a mutant CD3␧ allele lacking the highly con- cial early event in T cell activation (2, 3). Conversely, we found served proline residues, including the PxxDY-defining proline. that tyrosine phosphorylation of the PxxDY motif abolishes Nck- However, considering that the region involved in Nck and Eps8 SH3 or Eps8-SH3 binding to CD3␧. Thus, phosphorylation of binding shows the highest degree of conservation among verte- Y166 serves as molecular switch that determines whether CD3␧ is brate CD3␧ orthologs and is identical in amino acid sequence from competent for SH2 or SH3 binding. There are several examples in flounder to man (data not shown), we believe that the results by which serine or tyrosine phosphorylation of an SH3 domain or its Szymczak et al. (9) are better explained by functional redundancy ligand negatively regulates binding (28–30), but to our knowledge, within the mouse TCR-CD3 complex than by irrelevance of the this example is the first of reciprocal regulation of SH3 and SH2 Eps8 or Nck(I/III) binding site of CD3␧. Development of novel binding via phosphorylation of a common binding site. approaches, such as combined mutations in the intracellular sig- This mutually exclusive binding of SH2 and SH3 proteins to naling domains of different CD3 subunits to reduce this redun- CD3␧ has obvious implications for the currently unclear role of dancy, are warranted to establish the role of SH3 binding by CD3␧ CD3␧-SH3 interactions in T cell development and function. Gil in T cell biology, and to identify the relative importance of Nck Downloaded from et al. (5) suggested that Nck binding to CD3␧ immediately fol- and Eps8 family proteins as cellular partners of CD3␧. lowing TCR ligation would contribute to optimal signaling output by the TCR-CD3 complex. This conclusion has been criticized Acknowledgments (31) because it was mainly based on ectopic overexpression of We thank Marika Va¨ha¨-Jaakkola, Marita Hiipakka, Satu Ka¨rkkaïnen, and Nck(I/III)-SH3, which conceivably could interfere with many Herma Renkema for assistance and advice at different phases of this other later signaling events involved T cell activation. Also, it has project, Sampsa Matikainen for T cell cDNA samples, Aldo Borroto and http://www.jimmunol.org/ been shown that at least the bulk of Nck recruitment to the TCR- Balbino Alarcoń for technical advice and discussions, and Didier Trono for CD3 complex does not occur in the absence of protein tyrosine the pWPI vector. kinase activity, and is instead mediated by binding of the Nck-SH2 domain to phosphorylated SLP-76 (32). In any case, assuming that Disclosures early SH3 binding to CD3␧ is required for maximal TCR signal- The authors have no financial conflict of interest. ing, our results argue that the interaction with Nck and/or Eps8 proteins would have to be transient and lost before Zap70 (or Syk) References binding could take place. Such transient binding of Nck or Eps8 1. Call, M. E., and K. W. Wucherpfennig. 2005. The T cell receptor: critical role of by guest on October 2, 2021 could help to recruit Lck or Fyn or otherwise facilitate phosphor- the membrane environment in receptor assembly and function. Annu. Rev. Im- munol. 23: 101–125. ylation of Y166 to create a docking site for Zap70. However, our 2. Pitcher, L. A., and N. S. van Oers. 2003. T-cell receptor signal transmission: who results could also be explained by involvement of Nck or Eps8 gives an ITAM? Trends Immunol. 24: 554–560. ␧ 3. Mustelin, T., and K. Tasken. 2003. Positive and negative regulation of T-cell binding to CD3 in other aspects of TCR-CD3 function, which activation through kinases and phosphatases. Biochem J. 371: 15–27. might take place only later after ITAM dephosphorylation, or in an 4. Jordan, M. S., A. L. Singer, and G. A. Koretzky. 2003. Adaptors as central alternative signaling pathway distinct from the canonical Lck/ mediators of signal transduction in immune cells. Nat. Immunol. 4: 110–116. 5. Gil, D., W. W. Schamel, M. Montoya, F. Sańchez-Madrid, and B. Alarcoń. 2002. Zap70/LAT/SLP-76 cascade. In particular, if Eps8L1 or other Recruitment of Nck by CD3␧ reveals a ligand-induced conformational change members of this family turn out to be more relevant partners of essential for T cell receptor signaling and synapse formation. Cell 109: 901–912. CD3␧ than Nck, the lessons learned from studies on Eps8 in reg- 6. Gil, D., A. G. Schrum, B. Alarcoń, and E. Palmer. 2005. T cell receptor engage- ment by peptide-MHC ligands induces a conformational change in the CD3 com- ulation of epidermal growth factor receptor signaling, internaliza- plex of thymocytes. J. Exp. Med. 201: 517–522. tion, and trafficking (33) might provide a framework for future 7. Risuenõ, R. M., D. Gil, E. Fernańdez, F. Sańchez-Madrid, and B. Alarcoń. 2005. investigations into this question. Also, the recently revealed actin Ligand-induced conformational change in the T-cell receptor associated with productive immune synapses. Blood 106: 601–608. capping activity of Eps8 (34, 35) could hint to alternative roles for 8. Risuenõ, R. M., H. M. van Santen, and B. Alarcoń. 2006. A conformational the Eps8L1-CD3␧ interaction in regulation of T cell actin change senses the strength of T cell receptor-ligand interaction during thymic selection. Proc. Natl. Acad. Sci. USA 103: 9625–9630. remodeling. 9. Szymczak, A. L., C. J. Workman, D. Gil, S. Dilioglou, K. M. Vignali, E. Palmer, The transmembrane and ectodomains of CD3␧ are required for and D. A. Vignali. 2005. The CD3␧ proline-rich sequence, and its interaction proper assembly of the TCR-CD3 complex, and are therefore crit- with Nck, is not required for T cell development and function. J. Immunol. 175: 270–275. ical for normal T cell development and function (reviewed in Ref. 10. Mayer, B. J., and K. Saksela. 2004. SH3 domains. In Structure and Function of 1). By contrast, the function of the intracellular signal-transducing Modular Protein Domains. G. Cesareni, M. Gimona, M. Sudol, and M. Yaffe, domain of CD3␧ appears to be highly redundant with signaling eds. Wiley-VCH, Weinheim, pp. 37–58. ␦ ␥ ␨ 11. Li, S. S. 2005. Specificity and versatility of SH3 and other proline-recognition output of the three other CD3 subunits (CD3 , CD3 , and CD3 ) domains: structural basis and implications for cellular signal transduction. Bio- (reviewed in Ref. 2), which complicates studies of signaling in- chem. J. 390: 641–653. teractions by any individual CD3 subunit. All CD3 subunits con- 12. Ka¨rkkaïnen, S., M. Hiipakka, J. H. Wang, I. Kleino, M. Va¨ha¨-Jaakkola, G. H. Renkema, M. Liss, R. Wagner, and K. Saksela. 2006. Identification of tain one or more ITAMs, whereas only CD3␧ carries a hitherto preferred protein interactions by phage-display of the human Src homology-3 known SH3 binding motif. Nevertheless, it is possible that the proteome. EMBO Rep. 7: 186–191. 13. Mizushima, S., and S. Nagata. 1990. pEF-BOS, a powerful mammalian expres- function provided to the TCR-CD3 complex via SH3 binding by sion vector. Nucleic Acids Res. 18: 5322. CD3␧ may also be redundant with protein interactions mediated by 14. Sambrook, J., and D. W. Russell. 2001. Molecular Cloning: A Laboratory Man- other CD3 subunits. ual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. 15. Frank, R. 1992. Spot-synthesis: an easy technique for the positionally address- The exact dissociation constants of the Eps8-Nck(I/III)-SH3- able, parallel chemical synthesis on a membrane support. Tetrahedron 48: CD3␧ complexes remain to be determined, but our current data 9217–9232. The Journal of Immunology 885

16. Cronan, J. E., Jr. 1990. Biotination of proteins in vivo: a post-translational mod- 27. Lawe, D. C., C. Hahn, and A. J. Wong. 1997. The Nck SH2/SH3 adaptor protein ification to label, purify, and study proteins. J. Biol. Chem. 265: 10327–10333. is present in the nucleus and associates with the nuclear protein SAM68. Onco- 17. Zufferey, R., D. Nagy, R. J. Mandel, L. Naldini, and D. Trono. 1997. Multiply gene 14: 223–231. attenuated lentiviral vector achieves efficient gene delivery in vivo. Nat. Biotech- 28. Duke-Cohan, J. S., H. Kang, H. Liu, and C. E. Rudd. 2006. Regulation and nol. 15: 871–875. function of SKAP-55 non-canonical motif binding to the SH3c domain of adhe- 18. Rozen, S., and H. Skaletsky. 2000. Primer3 on the WWW for general users and sion and degranulation-promoting adaptor protein. J. Biol. Chem. 281: for biologist programmers. Methods Mol. Biol. 132: 365–386. 13743–13750. 29. Zhao, Z. S., E. Manser, and L. Lim. 2000. Interaction between PAK and Nck: a 19. Tocchetti, A., S. Confalonieri, G. Scita, P. P. Di Fiore, and C. Betsholtz. 2003. In template for Nck targets and role of PAK autophosphorylation. Mol. Cell Biol. silico analysis of the EPS8 gene family: genomic organization, expression profile, 20: 3906–3917. and protein structure. Genomics 81: 234–244. 30. Park, H., M. I. Wahl, D. E. Afar, C. W. Turck, D. J. Rawlings, C. Tam, 20. Tanaka, M., W. Lu, R. Gupta, and B. J. Mayer. 1997. Expression of mutated Nck A. M. Scharenberg, J. P. Kinet, and O. N. Witte. 1996. Regulation of Btk function SH2/SH3 adaptor respecifies mesodermal cell fate in Xenopus laevis develop- by a major autophosphorylation site within the SH3 domain. Immunity 4: ment. Proc. Natl. Acad. Sci. USA 94: 4493–4498. 515–525. 21. Mallabiabarrena, A., M. A. Jimeńez, M. Rico, and B. Alarcoń. 1995. A tyrosine- 31. Levin, S. E., and A. Weiss. 2005. Twisting tails exposed: the evidence for TCR containing motif mediates ER retention of CD3-␧ and adopts a helix-turn struc- conformational change. J. Exp. Med. 201: 489–492. ture. EMBO J. 14: 2257–2268. 32. Barda-Saad, M., A. Braiman, R. Titerence, S. C. Bunnell, V. A. Barr, and 22. La Gruta, N. L., H. Liu, S. Dilioglou, M. Rhodes, D. L. Wiest, and D. A. Vignali. L. E. Samelson. 2005. Dynamic molecular interactions linking the T cell antigen 2004. Architectural changes in the TCR:CD3 complex induced by MHC:peptide receptor to the actin cytoskeleton. Nat. Immunol. 6: 80–89. ligation. J. Immunol. 172: 3662–3669. 33. Lanzetti, L., V. Rybin, M. G. Malabarba, S. Christoforidis, G. Scita, M. Zerial, 23. Mongiovi, A. M., P. R. Romano, S. Panni, M. Mendoza, W. T. Wong, and P. P. Di Fiore. 2000. The Eps8 protein coordinates EGF receptor signalling A. Musacchio, G. Cesareni, and P. P. Di Fiore. 1999. A novel peptide-SH3 through Rac and trafficking through Rab5. Nature 408: 374–377. interaction. EMBO J. 18: 5300–5309. 34. Croce, A., G. Cassata, A. Disanza, M. C. Gagliani, C. Tacchetti, M. G. Malabarba, M. F. Carlier, G. Scita, R. Baumeister, and P. P. Di Fiore. 2004. A 24. Minguet, S., M. Swamy, B. Alarcoń, I. F. Luescher, and W. W. Schamel. 2007. novel actin barbed-end-capping activity in EPS-8 regulates apical morphogenesis Full activation of the T cell receptor requires both clustering and conformational in intestinal cells of Caenorhabditis elegans. Nat. Cell Biol. 6: 1173–1179. changes at CD3. Immunity 26: 43–54. 35. Disanza, A., M. F. Carlier, T. E. Stradal, D. Didry, E. Frittoli, S. Confalonieri, Downloaded from 25. Park, S., K. Takeuchi, and G. Wagner. 2006. Solution structure of the first Src A. Croce, J. Wehland, P. P. Di Fiore, and G. Scita. 2004. Eps8 controls actin- homology 3 domain of human Nck2. J. Biomol. NMR 34: 203–208. based motility by capping the barbed ends of actin filaments. Nat. Cell Biol. 6: 26. Castagnoli, L., A. Costantini, C. Dall’Armi, S. Gonfloni, L. Montecchi-Palazzi, 1180–1188. S. Panni, S. Paoluzi, E. Santonico, and G. Cesareni. 2004. Selectivity and pro- 36. Wang, B., N. Wang, C. E. Whitehurst, J. She, J. Chen, and C. Terhorst. 1999. T miscuity in the interaction network mediated by protein recognition modules. lymphocyte development in the absence of CD3␧ or CD3␥␦␧␨. J. Immunol. 162: FEBS Lett. 567: 74–79. 88–94. http://www.jimmunol.org/ by guest on October 2, 2021