Newly Generated T Cell Receptor Microclusters Initiate and Sustain T Cell Activation by Recruitment of Zap70 and SLP-76

ARTICLES

Newly generated T cell receptor microclusters initiate and sustain T cell activation by recruitment of Zap70 and SLP-76

Tadashi Yokosuka1, Kumiko Sakata-Sogawa2, Wakana Kobayashi1, Michio Hiroshima2, Akiko Hashimoto-Tane1, Makio Tokunaga2–4, Michael L Dustin5 & Takashi Saito1

T cell receptor (TCR) activation and signaling precede immunological synapse formation and are sustained for hours after initiation. However, the precise physical sites of the initial and sustained TCR signaling are not deﬁnitively known. We report here that T cell activation was initiated and sustained in TCR-containing microclusters generated at the initial contact sites and http://www.nature.com/natureimmunology the periphery of the mature immunological synapse. Microclusters containing TCRs, the tyrosine kinase Zap70 and the adaptor molecule SLP-76 were continuously generated at the periphery. TCR microclusters migrated toward the central supramolecular cluster, whereas Zap70 and SLP-76 dissociated from these microclusters before the microclusters coalesced with the TCR-rich central supramolecular cluster. Tyrosine phosphorylation and calcium inﬂux were induced as microclusters formed at the initial contact sites. Inhibition of signaling prevented recruitment of Zap70 into the microclusters. These results indicated that TCR- rich microclusters initiate and sustain TCR signaling.

Immune responses are triggered by the activation of T cells after However, the initial activation of T cells, including tyrosine phos- recognition of a specific antigen, which is mediated by the interaction phorylation, calcium mobilization and phosphoinositide metabolism, between antigen-presenting cells (APCs), such as dendritic cells, and occurs much earlier than immunological synapse formation10–13. antigen-specific T cells. Such interactions takes place through the Indeed, phosphorylation of the Src tyrosine kinases Lck and Zap70 10 2005 Nature Publishing Group Group 2005 Nature Publishing formation of a specific structure at the interface: the ‘immunological occurs earlier than mature immunological synapse formation .

© synapse’1–3. Within 10–15 min of cell-cell contact, dynamic rearrange- Analysis of CD2AP-deﬁcient mice has similarly indicated that ment occurs, accompanied by protein organization of both cell surface T cell activation is induced in the absence of mature immuno- and membrane-associated intracellular molecules. Immunological logical synapses14. Furthermore, partial agonist peptides can trigger synapse formation is dependent on T cell receptor (TCR)–mediated T cells in the absence of mature immunological synapses11,15.Thus, signaling that induces the cellular organization of cytoskeleton and immunological synapse formation may not be required for initial receptors and has effector functions4,5. For example, TCR stimulation T cell activation. induces speciﬁc segregation of surface molecules: the TCR-CD3 com- In contrast, the signaling pathways induced after TCR stimulation plex accumulates at the center of the interface as the central supra- have been analyzed extensively by biochemical and genetic molecular activation cluster (c-SMAC) together with other molecules, approaches. Antigen recognition by TCRs induces the phosphoryla- such as CD4-CD8, CD2 and CD28, whereas adhesion molecules such tion of the immunoreceptor tyrosine-based activation motif of as LFA-1 accumulate in a surrounding area called the peripheral CD3 molecules, particularly the CD3z chains, by Lck, which leads supramolecular activation cluster (p-SMAC). In addition, large mole- to recruitment of Zap70 to the phosphorylated immunoreceptor cules such as CD43 and CD45 are excluded from the immunological tyrosine-based activation motifs. Zap70 is activated and then phos- synapse and instead are located in the distal supramolecular activation phorylates several adaptor molecules, including two critical adaptors, cluster (d-SMAC)6–9. From these structural features of the immuno- LAT and SLP-76. The membrane adaptor molecule LAT, localized in logical synapse, particularly the accumulation of TCRs in the c-SMAC, lipid rafts, recruits SLP-76 through association with the adaptor it was thought that the c-SMAC mediates the antigen recognition and protein Gads. Phosphorylated LAT and SLP-76 then recruit many activation of T cells, whereas the p-SMAC serves to support cell ‘downstream’ signaling molecules, leading to the activation of adhesion and maintain the immunological synapse. various effector functions.

1Laboratory for Cell Signaling and 2Single-Molecule Immunoimaging, RIKEN Center for Allergy and Immunology, Tsurumi-ku, Yokohama, Kanagawa 230-0045, Japan. 3Structural Biology Center, National Institute of Genetics, and 4Department of Genetics, The Graduate University for Advanced Studies, Mishima, Shizuoka 411-8540, Japan. 5Program in Molecular Pathogenesis, Skirball Institute of Biomolecular Medicine and Department of Pathology, New York University School of Medicine, New York, New York 10021, USA. Correspondence should be addressed to T.S. ([email protected]). Received 11 July; accepted 6 October; published online 6 November 2005; doi:10.1038/ni1272

NATURE IMMUNOLOGY VOLUME 6 NUMBER 12 DECEMBER 2005 1253 ARTICLES

a 10 s 20 s 30 s 40 s 50 s 60 s 150 s 300 s b CD3ζ Zap70 SLP-76 ζ CD3 0–75 s 0–75 s SLP-76 Zap70 CD3ζ Zap70 SLP-76 c 0 0 0 35–70 s 0–35 s 75–150 s 1 1 1 75–150 s

2 2 2

3 3 3 Time (min) 70–105 s

4 4 4 150–225 s 150–225 s After-contraction Contraction Expansion 5 5 5

Figure 1 Generation of microclusters of CD3z, Zap70 and SLP-76 throughout the cell-bilayer contact area in a spatio-temporal way. (a) AND-Tg T cells expressing EGFP–CD3z, EGFP–Zap70 or EGFP–SLP-76, plated on a planar bilayer containing unlabeled I-Ek and ICAM-1 (prepulsed with PCC(88–104)). Images were obtained at a video rate (30 frames/s) using TIRFM (times, above images). (b) Representative CD3z, Zap70 and SLP-76 microclusters tracked after cell-bilayer contact; paths of movement are presented separately in three phases: expansion (top), contraction (middle) and after contraction with http://www.nature.com/natureimmunology c-SMAC (bottom). Real-time images are available in Supplementary Video 1 (EGFP–CD3z), Supplementary Video 2 (EGFP–Zap70) and Supplementary Video 3 (EGFP–SLP-76) online. (c) Inward movement and accumulation or disappearance of CD3z, Zap70, or SLP-76 microclusters, presented as horizontal elements in kymographs. Scale bars, 5 mm.

The dynamic recruitment of signaling molecules during the initial microclusters was necessary for inducing activation. After immuno- activation of T cells has been analyzed using Jurkat cells expressing logical synapse formation, signal-inducing microclusters were gener- various enhanced green ﬂuorescent protein (EGFP) fusion proteins ated only at the periphery of the immunological synapse, which and stimulation with immobilized antibody to CD3 (anti-CD3)16. suggested that sustained activation signals were produced from the Those studies have demonstrated that small clusters containing TCR, microclusters at the periphery of the immunological synapse and not kinase and adaptor molecules are induced immediately after stimula- from the c-SMAC. tion of the TCR. However, stimulation with immobilized anti-CD3 limits analysis to the initial phase of activation, and thus the entire RESULTS

2005 Nature Publishing Group Group 2005 Nature Publishing process up to the formation of the immunological synapse in more Generation of TCR microclusters at the initial contact sites

© physiological conditions remains unknown. The immunological Because we intended to analyze the microstructural basis of immu- synapse structure is necessary for the maintenance of sustained nological synapse formation, we used a combined system of a planar signals required for full T cell activation17. It is well known that bilayer that incorporated glycophosphatidylinositol (GPI)–anchored sustained signal transduction, including mitogen-activated protein intercellular adhesion molecule 1 (ICAM-1) and major histocompat- kinase activation, calcium mobilization and activation of Akt (protein ibility complex (MHC) class II (I-Ek) loaded with a pigeon cyto- kinase B), is required for T cell activation and induction of cytokine chrome c (PCC) peptide of amino acids 88–104 (PCC(88–104))18 and secretion and effector function. Immunological synapse structure is objective-type TIRFM to detect the dynamic recruitment of clustering maintained for hours17–20, and disruption of the immunological molecules after antigen stimulation of T cells. We obtained PCC(88– synapse immediately stops sustained signaling, which suggests that 104)-specific T cells from AND TCR-transgenic (AND-Tg) mice that the immunological synapse is required for the maintenance of had been activated in vitro and had been previously transduced by sustained signaling13,21–27. However, the actual physical location for EGFP-tagged CD3z, Zap70 and SLP-76 retroviruses and placed the the induction of such sustained signaling remains unclear. cells on a planar bilayer. Immunological synapse formation, visualized To analyze precisely and in real time the dynamic process of as TCR-enriched c-SMACs and LFA-1-enriched p-SMACs, as immunological synapse formation, particularly its relationship with described before1,2, occurred 5–10 min after cell-bilayer contact. the formation of the TCR signaling complex to induce activation When we analyzed c-SMACs in T cells expressing a fusion protein signals, we used a combined system of a planar bilayer for normal of an EGFP tag and CD3z (EGFP–CD3z) in this system, we found antigen-specific T cells and total internal reflection fluorescence that the c-SMACs were composed not simply of accumulated CD3z microscopy (TIRFM). This approach allowed us to analyze the but also many CD3z-containing microclusters, as suggested by other ‘nanoscale’ movement of signal molecules during T cell activation. reports6,16,28. We then sought to determine the origin of these Using this system, we have identified microclusters containing TCRs, microclusters by tracing back to the initial cell-bilayer contact (time kinases and adaptor molecules and have monitored the entire process 0). The CD3z microclusters were generated as early as 5 s after cell from immediately after cell-bilayer contact to the completion of contact with the planar bilayer (Fig. 1a and Supplementary Video 1 immunological synapse formation. We found that the microclusters online). During the initial 1–2 min, T cells expanded the contact area were sites for generating the initial and sustained activation signals and many microclusters were generated all over (expansion phase). and, moreover, that formation of TCR signaling complexes in Microcluster formation was induced by the interaction with the

1254 VOLUME 6 NUMBER 12 DECEMBER 2005 NATURE IMMUNOLOGY ARTICLES

a 1 min 3 min 30 min b 1 min 3 min 30 minc 1 min 3 min 30 min Zap70 SLP-76 SLP-76 ε ε CD3 CD3 Zap70 DIC DIC DIC

d 1 min 3 min 30 min e 1 min 3 min 30 min f 1 min 3 min 30 min http://www.nature.com/natureimmunology Zap70 Merge SLP-76 Merge SLP-76 Merge ε ε CD3 CD3 Zap70

Figure 2 Recruitment of Zap70 and SLP-76 to CD3 microclusters immediately after cell-bilayer contact. (a–c) Localization of CD3e together with Zap70 (a), CD3e with SLP-76 (b) and SLP-76 with Zap70 (c). (a,b) AND-Tg T cells expressing EGFP–Zap70 (a) or EGFP–SLP-76 (b) were plated on a planar bilayer containing I-Ek and ICAM-1; cells were fixed 1, 3 or 30 min later and were stained for CD3e (Alexa Fluor 633). EGFP fusion proteins, green; CD3e, red. (c) AND-Tg T cells expressing both EGFP–SLP-76 and mRFP–Zap70 were plated on a planar bilayer containing I-Ek and ICAM-1; cells were fixed 1, 3 or 30 min later. EGFP, green; mRFP, red. DIC, differential interference contrast. Scale bars, 5 mm. (d–f) Relative fluorescence intensities of EGFP–Zap70 (d, top), mRFP–Zap70 (f, bottom), EGFP–SLP-76 (e,f, top) and CD3e (d,e, bottom) on the diagonal yellow lines in a–c (bottom row). 2005 Nature Publishing Group Group 2005 Nature Publishing © peptide-MHC (pMHC), as microclusters were neither generated on the average moving distance was 0.84 ± 0.45 mm for the central the planar bilayer in the absence of antigen nor present in the cells, microclusters and 3.1 ± 1.2 mm for peripheral microclusters. indicating that the microclusters were not pre-formed. During the The number of EGFP–CD3z proteins in a single microcluster was expansion phase, the CD3z microclusters did not move, remaining at approximately 30–100 in this system, as estimated from the fluores- the same location (Fig. 1b, top left). After reaching maximum cell cence intensity. The expression of EGFP–CD3z was two- to threefold spreading, the CD3z microclusters began to move toward the center higher than that of endogenous CD3z, as assessed by intracellular and created a c-SMAC at 5 min (contraction phase; Fig. 1b, left staining (data not shown). Accordingly, we estimated that the number column, middle and bottom images). We determined the time course of CD3z proteins in one microcluster was approximately 40–150. using a kymograph (Fig. 1c, left). From that image, it was clear that We also quantified the area and relative fluorescence intensity of in- the size and intensity of the CD3z microclusters increased as the dividual CD3z microcluster (Supplementary Fig. 2 online) to deter- microclusters moved to the center, suggesting that the microclusters mine how extensively EGFP–CD3z accumulated in microclusters. The were fused on their way to the center. Traces of the movement of average sum of the area and the relative intensity of microclusters was individual microclusters containing CD3z showed that most of the 9.0% ± 3.0% and 32.4% ± 7.5% of whole contact area, compared with CD3z proteins were translocated into the c-SMAC. background. The fluorescence intensity inside the microclusters was By means of such individual tracing, we quantified the average 10.4 ± 4.7 times higher than outside the microclusters. These data may velocity and the moving distance (Supplementary Fig. 1 online). underestimate the number and fluorescence intensity of microclusters During the expansion phase, CD3z microclusters appeared first at the because of the possible existence of undetectable small clusters. center and then at the periphery, remaining at the same location. The average velocity was 6.8 ± 1.9 nm/s at the center and 19.0 ± 7.9 nm/s Initial microclusters contain TCRs, kinases and adaptors at the periphery. During the contraction phase, peripheral microclus- To understand the spatio-temporal dynamism of the TCR signaling ters in more distal areas moved much faster toward the center (92.5 ± complex during the generation and translocation of microclusters, 31.0 nm/s) than did those at the central area (4.4 ± 1.9 nm/s). After we analyzed T cells expressing EGFP–Zap70 as a CD3-associating the contraction phase, the velocity was reduced to a constant value critical kinase and EGFP–SLP-76 as a critical adaptor molecule for (24.9 ± 12.6 nm/s). Most CD3z microclusters reached the center, and transducing activation signals downstream, respectively. EGFP–Zap70

NATURE IMMUNOLOGY VOLUME 6 NUMBER 12 DECEMBER 2005 1255 ARTICLES

a 1 min 3 min 30 minbcd1 min 3 min 30 min 3 min 3 min ζ ζ ζ ζ CD3 CD3 CD3 CD3 pY pY pZap70 pZap70 Merge Merge Merge Merge DIC DIC DIC DIC

ef1 min 3 min 30 min 1 min 3 min 30 min http://www.nature.com/natureimmunology ζ ζ CD3 CD3 pY pZap70

Figure 3 Phosphorylated Zap70 and tyrosine-phosphorylated proteins are localized only in CD3z microclusters at the periphery of immunological synapse. (a,b) Staining for pZap70 (a)orpY(b) with Alexa Fluor 633 in AND-Tg T cells expressing EGFP–CD3z fixed at times above images. (c,d)Highermagni- fication of areas outlined in white in a (c)andinb (d). Arrowheads indicate microclusters containing both CD3z and pZap70 (c)orpY(d). Scale bars, 5 mm. (e,f) Relative fluorescence intensities of EGFP–CD3z (top), pZap70 (e,bottom)andpY(f, bottom) on the yellow diagonal lines in a and b (bottom row). 2005 Nature Publishing Group Group 2005 Nature Publishing © and EGFP–SLP-76 restored the TCR signaling for interleukin 2 To determine whether a single microcluster contained all the production in Zap70-deficient and SLP-76-deficient Jurkat cells, molecules required for T cell activation, we analyzed the colocalization respectively (data not shown). Zap70 and SLP-76 emerged as micro- of TCR, Zap70 and SLP-76 (Fig. 2). We fixed AND-Tg T cells clusters immediately after cell-bilayer contact, and the number of transduced by EGFP–Zap70 (Fig. 2a,d), and EGFP–SLP-76 microclusters increased until maximum cell spread was reached, (Fig. 2b,e) at various times after cell-bilayer contact, followed by similar to EGFP–CD3z (Fig. 1a and Supplementary Videos 2 and 3 staining with anti-CD3e. At initiation of contact and during the online). These microclusters did not move during the expansion phase expansion phase (1 min), Zap70 and CD3e and SLP-76 and CD3e (Fig. 1b,c, top row, middle and right images). However, their activity were colocalized in the same microclusters (Fig. 2a,b, left column). differed from that of CD3z after the contraction phase. Most Zap70 Zap70 and SLP-76 were mainly in the periphery at 3 min, when CD3e microclusters did not reach the center, stayed around the periphery began translocate to the center (contraction phase, Fig. 2a,b, center and disappeared (Fig. 1a,c), whereas some Zap70 accumulated in the column). Next, to analyze the localization of Zap70 and SLP-76 c-SMAC. Consistent with that finding, the total moving distances of together, we transfected monomeric red fluorescent protein Zap70 microclusters were rather small (Fig. 1b, middle column, (mRFP)–Zap70 and EGFP–SLP-76 fusion proteins together into middle and bottom images). In contrast, whereas the initial micro- AND-Tg T cells and plated cells on a planar bilayer. Zap70 and clusters containing EGFP–SLP-76 were formed like those containing SLP-76 localized together at the initiation of cell-bilayer contact CD3z and Zap70 during the expansion phase, some of SLP-76 (Fig. 2c,f). We further analyzed the colocalization of these three microclusters moved to the center during the early contraction molecules by counting all the microclusters on the eight representative phase (Fig. 1b, right), and almost all disappeared after the contraction cells in the three different combinations of proteins (Fig. 2a–e,left phase (Fig. 1c, right). These results demonstrate that Zap70 and SLP- column) and examining the localization of the two molecules together. 76 may have nonredundant functions in immunological synapse The percentage localization of Zap70 together with CD3e, SLP-76 with formation. The disappearance of microclusters containing Zap70 CD3e, or Zap70 with SLP-76 was 90.2% ± 7.7%, 85.3% ± 6.5% or and SLP-76 suggests that most Zap70 and SLP-76 microclusters 88.5% ± 7.2%, respectively (Supplementary Table 1 online). These were generated only at the periphery after the contraction phase results indicated that CD3z, Zap70 and SLP-76 were localized together and did not translocate to the c-SMAC, unlike CD3z. in most of the initial microclusters.

1256 VOLUME 6 NUMBER 12 DECEMBER 2005 NATURE IMMUNOLOGY ARTICLES

ab0 s 120 s 240 s 360 s 480 s 600 s 0 s 120 s 240 s 360 s 480 s 600 s

8 8 6 6 4 4

F405/F495 2

F405/F495 2 0 0 0 120 240 360 480 600 0 120 240 360 480 600 Time (s) Time (s)

Figure 4 Induction of calcium influx accompanies microcluster formation at the initial contact sites. (a)CD3z microclusters (top) and calcium influx (bottom) in AND-Tg T cells expressing EGFP–CD3z. Cells were loaded with Indo-1 and were plated on a PCC(88–104)-pulsed planar bilayer containing I-Ek and ICAM-1, and EGFP and Indo-1 were monitored simultaneously. Five representative cells are presented (one color per cell). (b) CD3z microclusters (top) and calcium influx (bottom) in AND-Tg T cells expressing EGFP–CD3z plated on a planar bilayer without peptide pulse, analyzed as in a. F405/F495, ratio of fluorescence at 405 nm to fluorescence at 495 nm.

We then examined whether these microclusters were generated in (Fig. 3b, left column), whereas only the microclusters at the periphery more physiological conditions. We analyzed two aspects: stimulation were stained after contraction began (Fig. 3a,b,centercolumns,and with low concentrations of an agonistic peptide, and microclusters c,d). The colocalization was restricted to the periphery and was not during cell-cell contact. When the concentration of PCC(88–104) was inside the cells, suggesting that phosphorylation of Zap70 and other http://www.nature.com/natureimmunology ‘titrated down’ to suboptimal doses for in vitro stimulation, both the molecules was induced in most of the microclusters during the size and the density of c-SMACs were decreased gradually. However, expansion phase and then was induced mainly in the microclusters all CD3z, Zap70 and SLP-76 microclusters were clearly detected at the at the periphery after the contraction phase. In a report suggesting the suboptimal PCC(88–104) concentration of 0.8 mM(Supplementary ‘resetting’ of T cell activation, the central clustering of pZap70 was Fig. 3 online). Quantitative analysis of CD3z microclusters showed shown transiently at 7 min but not at 3 min after the T cell–APC that the decrease in ﬂuorescence intensity of each microclusters contact6. Therefore, we carefully analyzed the kinetics of pZap70- seemed to be greater than that of the number of microclusters at containing microclusters at T cell–bilayer contacts throughout the low concentrations of antigen peptide (Supplementary Fig. 3 online), formation of mature immunological synapse. However, the accumula- suggesting that the increase in antigen concentration may have tion of pZap70 at around 7 min was not obvious (Supplementary accumulated more microcluster components into microclusters Fig. 6 online). Within 1 min, pZap70 was localized together with TCR instead of increasing the number of microclusters. We noted similar microclusters. The initial pZap70 microclusters soon disappeared and characteristics for Zap70 and SLP-76 microclusters at all concentra- only the peripheral pZap70 microclusters remained, probably because

2005 Nature Publishing Group Group 2005 Nature Publishing tions of PCC(88–104). Next, to analyze microcluster formation during Zap70 microclusters were continuously generated at the periphery and

© T cell–APC interaction, we generated Jurkat T cell lines expressing pZap70 microclusters were renewed there. EGFP–CD3z, EGFP–Zap70 or EGFP–SLP-76 and conjugated these To further verify the finding that microclusters containing TCRs, cells with Raji cells loaded with staphylococcus enterotoxin E. There kinases and adaptor molecules were the sites for generating the initial was microcluster formation of each of these molecules at the cell-cell activation signals, we simultaneously monitored the kinetics of contact sites (Supplementary Fig. 4 online). After the mature immu- microcluster formation and intracellular calcium. We labeled AND- nological synapse was formed, CD3z and a part of Zap70 were Tg T cells expressing EGFP–CD3z with the calcium indicator Indo-1 accumulated at c-SMAC and localized together with the TCR. SLP- and plated them on a PCC(88–104)-pulsed planar bilayer. Intracellular 76 was present mainly in the cytosol and not at the contact sites in the calcium increased rapidly within 30 s of cell-bilayer contact and this absence of stimulation. These results coincide with our data on T cell– was sustained throughout the experimental period (at least 10 min; bilayer interactions. We also assessed microcluster formation at the Fig. 4a). These calcium mobilizations were triggered in parallel with initiation of cell-cell contact using AND-Tg T cells expressing EGFP– the formation of several microclusters in the expansion phase. SLP-76 and PCC(88–104)-pulsed activated B cells (Supplementary The sustained calcium concentrations were unchanged by c-SMAC Fig. 5 online). Indeed, there were microclusters on the interface at the formation in the contraction phase and there were no oscillations in very beginning of cell contact. After the mature immunological intracellular calcium. When we plated the Indo-1-loaded cells on a synapse was formed, SLP-76 did not seem to be localized in the planar bilayer without a peptide pulse, the cells induced neither microclusters. Notably, only bright clusters and not small microclus- microcluster formation nor an increase in intracellular calcium ters could be detected in this system. (Fig. 4b). These results suggested that the microclusters were sufficient for initiating phosphorylation and calcium influx and TCR microclusters as the sites for initiating T cell activation that signaling molecules were rapidly recruited into the microclusters To identify the site for generating activation signals, we assessed after stimulation. tyrosine phosphorylation in T cells expressing EGFP–CD3z using monoclonal antibody to phosphorylated Zap70 (pZap70) and mono- Tyrosine kinase–dependent formation of functional microclusters clonal antibody to phosphotyrosine (pY). During the expansion The Src family kinases Lck and Fyn are critical tyrosine kinases for phase, most of the CD3z-containing microclusters were stained proximal TCR signaling by the phosphorylation of CD3z and the together with anti-pZap70 (Fig. 3a, left column) and anti-pY recruitment of Zap70. Therefore, we next examined microcluster

NATURE IMMUNOLOGY VOLUME 6 NUMBER 12 DECEMBER 2005 1257 ARTICLES

a 3 min 30 min Figure 5 Src family kinases regulate Zap70 recruitment to microclusters EGFP DIC EGFP DIC but not the formation of CD3z microclusters. AND-Tg T cells expressing EGFP–CD3z or EGFP–Zap70 were left untreated (a) or were pretreated with 10 mM PP2 (b), were plated on PCC(88–104)-pulsed planar bilayers ζ containing I-Ek and ICAM-1 and were ﬁxed 3 or 30 min after cell-bilayer

CD3 contact. Scale bars, 5 mm.

TIRFM after the completion of immunological synapse formation.

PP2 (–) Unexpectedly, many CD3z microclusters were continuously generated at the periphery, which is a lamellipodia-like structure, and continuously translocated to the center (Fig. 6, left, and Supplementary Zap70 Video 4 online). The estimated number of CD3z proteins in a single microcluster, when cells had a c-SMAC, was approximately 40–110 at b the periphery and 110–290 at the positions closer to the c-SMAC

ζ (Supplementary Fig. 8 online). In contrast, whereas Zap70 micro-

CD3 clusters were also generated at the periphery, most Zap70 microclusters were ‘dislocated’ on the way to the center, as shown in video images and short traces (Fig. 6, center column, and Supplementary

PP2 (+) Video 5 online). Similarly, SLP-76 microclusters were extensively and continuously generated at the periphery but completely disappeared

Zap70 (Fig. 6, right column, and Supplementary Video 6 online). We conﬁrmed those results by tracing and quantifying individual microclusters (Supplementary Fig. 9 online). The moving distances of Zap70 and SLP-76 microclusters (0.64 ± 0.43 mm and 0.59 ± 0.38 mm, http://www.nature.com/natureimmunology respectively) were much smaller than that of CD3z microclusters formation and the recruitment of Zap70 in the presence of the (2.18 ± 0.81 mm). Those results indicated that whereas CD3z accu- Src kinase inhibitor PP2. We treated AND-Tg T cells expressing mulated at the center, the intracellular signaling molecules Zap70 and EGFP–CD3z or EGFP–Zap70 with PP2 and plated the cells on a SLP-76 were associated mainly with TCR-CD3 only in the micro- planar bilayer. As expected, microcluster formation of Zap70 was clusters at the periphery of the immunological synapse. To conﬁrm blocked by PP2 both early and late compared with Zap70 micro- that, we examined the localization of CD3z and Zap70 together clusters in the untreated T cells (Fig. 5a,b, bottom). Unexpectedly, in the newly generated microclusters. These molecules localized CD3z microclusters were generated even in the PP2-treated T cells together in the same microclusters at the periphery after mature and translocated to the central area (Fig. 5b, top), although in a way different from that of the untreated cells. In the PP2-treated T cells, CD3z clusters moved multidirectionally and fused with each a CD3ζ Zap70 SLP-76 other, resulting in c-SMAC-like accumulation, although the cells

2005 Nature Publishing Group Group 2005 Nature Publishing did not spread (Supplementary Fig. 7 online). These results suggest

© that the CD3z microcluster formation can be initially triggered in a Src family kinase–independent way, whereas the subsequent generation of functional microclusters as well as the unidirectional translocation of microclusters to create c-SMAC proceeds in a kinase-dependent way. b Generation of sustained signals in new peripheral microclusters In contrast to the rapid assembly of the signaling microclusters, TCR- mediated signaling continues for hours after T cell–APC conjugation, and immunological synapse formation is required for the continuous TCR signaling3. To identify the sites at which signal transductions continued even after mature immunological synapse formation, we analyzed microclusters containing CD3z, Zap70 and SLP-76 by CD3ζ Zap70 SLP-76 c 0 0 0

Figure 6 CD3z, Zap70 and SLP-76 microclusters are continuously generated at the periphery after mature immunological synapse formation. 1 1 1

(a) AND-Tg T cells expressing EGFP–CD3z (left), EGFP–Zap70 (center) or Time (min) EGFP–SLP-76 (right), plated on a planar bilayer containing unlabeled I-Ek and ICAM-1 and prepulsed with PCC(88–104). Images at 15 min after cell-bilayer contact were collected at a video rate (30 frames/s) by TIRFM. d Real-time images are available in Supplementary Video 4 (EGFP–CD3z), Supplementary Video 5 (EGFP–Zap70) and Supplementary Video 6 (EGFP– SLP-76) online. (b) Surface plotting of the TIRFM images from a.(c)The 90-second movements of CD3z, Zap70 or SLP-76 microclusters, presented as kymographs. (d) Representative paths of movement of CD3z, Zap70 and SLP-76 microclusters. Scale bars, 5 mm.

1258 VOLUME 6 NUMBER 12 DECEMBER 2005 NATURE IMMUNOLOGY ARTICLES

Zap70TCR Merge DIC for the initial activation signals. We further extended our analysis to the dynamic movement and redistribution of the signaling molecules during the entire process of immunological synapse formation. We found that microclusters were generated continuously at the periphery even after mature immunological synapse formation, the peripheral microclusters contained phosphorylated signaling molecules and, therefore, these microclusters, but not the c-SMAC, probably served as the sites for generating activation signals. These microclusters might resemble the multifocal structure without a c-SMAC that has been reported in the antigen-independent synapse between T cells and dendritic cells, in terms of TCR patch formation on the whole interface of the cell-cell contact31. However, microclusters were actively induced along with cell spreading, as Figure 7 Localization of TCR and Zap70 together in newly generated suggested before16. Inhibition of actin polymerization by cytochalasin microclusters at the periphery after mature immunological synapse B resulted in considerable blockade of cell spreading and microcluster formation. AND-Tg T cells expressing EGFP–Zap70 were stained for TCRb (red) and were plated on a planar bilayer; cells were fixed 20 min after cell- formation, whereas small clusters in each microvillus on the contact bilayer contact. Bottom row, magnified images of regions outlined in white in interface were generated (unpublished observations). Those observa- top row. Arrowheads indicate microclusters contianing both pZap70 and tions suggest that microclusters may not be pre-formed but may be TCRb.Scalebars,5mm. induced along with cell spreading through TCR-pMHC interaction. We precisely analyzed the differential dynamics of CD3z, Zap70 and SLP-76 and found that the activities of these signaling molecules were immunological synapse formation (Fig. 7). Furthermore, pZap70 and different from those of receptors. The receptors at the contact sites pY staining was localized mainly at the periphery of the immunolo- were translocated into the c-SMAC, whereas most Zap70 and SLP-76 http://www.nature.com/natureimmunology gical synapse rather than at the center of the mature immunological molecules did not move to the c-SMAC. Most Zap70 disappeared synapse (Fig. 3a,c, right), suggesting that phosphorylation took place from the initial CD3z microclusters, whereas some moved to the mainly in the continuously generating new peripheral microclusters. c-SMAC. In contrast, SLP-76 completely disappeared during trans- These results collectively suggest that signaling molecules transiently location to the c-SMAC. This contrasts with the fate of SLP-76 associate with TCR microclusters at the periphery where the sustained microclusters in the case of immobilized antibody stimulation, signals proceed. which are internalized and transported to peri-Golgi region of the cell16,28. We found no evidence of stable accumulation of SLP-76 near DISCUSSION the Golgi apparatus. Whether the disappearance of Zap70 and SLP-76 The discovery of the immunological synapse and the specific segrega- reflects degradation, association with other molecules or recycling for tion of surface molecules at the interface between T cells and APCs has newly generated microclusters remains to be clarified. Zap70 associ- stimulated studies aimed at elucidating function of the immunological ates with Wiskott-Aldrich syndrome protein, and the complex is synapse in the recognition and activation of T cells. As the TCR recruited to the immunological synapse after TCR engagement32,33.

2005 Nature Publishing Group Group 2005 Nature Publishing complex is accumulated in the c-SMAC together with various key Similarly, SLP-76 also associates with Wiskott-Aldrich syndrome

© signaling molecules, it is believed that the c-SMAC functions in the protein and vasodilator-stimulated phosphoprotein via the adaptor recognition and activation of T cells. However, the p-SMAC and proteins Nck and ADAP, respectively34–36. Therefore, one possibility d-SMAC are thought to function merely in cell adhesion and the for the disappearance of Zap70 and SLP-76 is that these proteins are storage of molecules excluded from the c-SMAC, such as CD43 and involved in actin remodeling. Our finding of a bundle of SLP-76 and CD45, respectively7,9. Here we have done precise structural analysis by actin at the lamellipodia-like structure might support that idea molecular imaging of T cells expressing EGFP fusion proteins on (unpublished observation). pMHC-containing planar bilayers and detection with TIRFM. Starting We defined the site for the induction of tyrosine phosphorylation of from the observation that the c-SMAC was composed of many CD3z various signal molecules after TCR ligation. Zap70 almost completely microclusters, we found that microclusters were generated at the localized together with CD3z during the expansion phase, followed by periphery of the immunological synapse and served as the sites for loss of Zap70 from CD3z microclusters, which translocated to the initiating and sustaining TCR signals. These functional microclusters c-SMAC after the contraction phase and partially localized together were induced in physiological conditions, with suboptimal concentra- with CD3z during the sustained phase. The residual Zap70 in c-SMAC tions of antigen peptide (o1 mM) and at the T cell–APC interface. was not phosphorylated, which is consistent with a report that the Those findings change the present view of the function of immuno- pZap70 accumulates at the periphery before the mature immunological synapse and SMAC. logical synapse formation10. Our results suggest that like Lck, Zap70 The existence of a few clusters of CD3z at the interface between may also function as an adaptor, particularly in the c-SMAC. T cells and APCs has been noted before29,30. The initial events of T cell The dynamic generation of peripheral microclusters and their activation have been extensively analyzed, indicating that TCR stimu- critical involvement in both initial and sustained signaling suggest lation induces the dynamic formation of small clusters that are that microclusters are the cellular sites for the generation of activation enriched in TCR, Zap70, LAT, Grb2, Gads, SLP-76 and phospho- signals and indicate the importance of immunological synapse for proteins. However, those analyses were confined to initial events sustained signaling. These sustained signals are required for cumula- because they used stimulation with immobilized antibody. We con- tive effect and functional effector T cell differentiation17,19,20,37,38. firmed those findings by using normal T cells in our studies; we Whereas microclusters contain both receptors and signaling molecules constructed the initial TCR microclusters with receptors, kinases and at the beginning of cell-bilayer contact, the signaling molecules adaptor molecules during the expansion phase, which were sufficient dissociate from the receptors during the contraction phase. Therefore,

NATURE IMMUNOLOGY VOLUME 6 NUMBER 12 DECEMBER 2005 1259 ARTICLES

functional ‘signaling microclusters’ are continuously generated only at cells and maintaining cell-cell contact through strong interaction with the periphery, which induces sustained TCR signals. As biochemical pMHC. The balance between c-SMAC-mediated stabilization and analysis has shown that tyrosine phosphorylation is induced within a chemokine-induced polarization may determine the cell fate of stop- few minutes of TCR stimulation, this may be consistent with the ping versus moving. kinetics of the localization of TCRs and signaling molecules together in the peripheral microclusters. It is likely that phosphorylation and METHODS activation signals are induced when TCRs and signaling molecules Reagents. Staining antibodies and reagents were purchased from the following exist together in the microclusters, although it is possible that much suppliers: biotinylated anti-CD3e (2C11), eBioscience; anti-pZap70, Cell smaller and undetectable TCR clusters that may transit to detectable Signaling; anti-pY, Upstate; Alexa Fluor 633–streptavidin and Alexa Fluor microclusters could transduce activation signals. The estimated num- 647–anti-rabbit IgG, Molecular Probes; and indocarbocyanine–anti-mouse ber of CD3z proteins in the peripheral microclusters in the sustained IgG, Amersham Biosciences. Indo-1 acetoxylmethylester (Indo-1 AM) was purchased from Molecular Probes. PP2 was purchased from Calbiochem. phase was approximately 40–110. CD3z proteins in microclusters close to the c-SMAC numbered 110–290, probably because of fusion Plasmid construction. Mouse EGFP–CD3z, EGFP–Zap70, EGFP–SLP-76 and while moving toward c-SMAC. Those numbers are similar to those mRFP–Zap70 fusion proteins were generated by recombinant PCR with mouse generated immediately after the initial cell-bilayer contact (Supple- cDNA and pEGFP-N1 (BD Clontech) or pRSETB–mRFP1 (ref. 43; provided by mentary Fig. 8 online). Given a report that the engagement of R.Y. Tsien, University of California, San Diego, California) and were subcloned 5–10 TCRs may be sufficient to trigger calcium signaling11, we assume into retroviral vector pMXs (provided by T. Kitamura, Tokyo University, Tokyo, that activation signals can be induced in even smaller microclusters, Japan). The fusion proteins contained no linker sequences. which remain to be analyzed at the single-molecule level. As TCR Primary cell cultures and transductions. Constructs were transiently downregulation can continue over a period of hours, it is possible that transduced into Phoenix packaging cells (provided by G. Norlan, Stanford the TCR microclusters may become progressively smaller with increas- University) using Lipofectamine (Invitrogen). Retroviral supernatants were ing time of sustained signaling. centrifuged at 8,000g for 12 h and were concentrated tenfold. CD4+ T cells Treatment with PP2 inhibited the recruitment of Zap70 but not were purified from AND-Tg mice on a recombination-activating gene http://www.nature.com/natureimmunology the formation of CD3z microclusters, indicating that the TCR signal- 2–deficient background and were stimulated with 3 mM PCC(88–104) ing complex has the ability to form microclusters in a Src family (KAERADLIAYLKQATAK) and irradiated whole spleen cells from C3H/HeN kinase–independent way, consistent with a report describing mice. At 1 and 2 d after stimulation, the cells were suspended in the retroviral Zap70-independent clustering of CD3z in the immunological supernatant and were centrifuged at 1,000g for90mininthepresenceof synapse8. However, the activity of the clusters in that condition was 8 mg/ml of polybrene (Sigma) and 200 U/ml of human recombinant interleukin 2 (Ajinomoto). At day 4, 60–80% of T cells were transduced with EGFP different from the activity of clusters without PP2. They fused and and were sorted with a FACSAria (BD) to obtain populations homogenous dissociated randomly, suggesting that assembly with other signaling in fluorescence intensity. Cells were maintained for 5–14 d in RPMI molecules and the actin cytoskeleton regulates the proper formation medium plus 10% FCS supplemented with recombinant interleukin 2. In and translocation of the microclusters. Those results suggest that the Src kinase–blocking assay, AND-Tg cells expressing EGFP–CD3z or sequential and sustained signals in the microclusters are necessary EGFP–Zap70 were preincubated for 20 min at 37 1Cwith10mMPP2in for maintaining the structure of the immunological synapse. More- the culture medium. This drug was present in the assay medium throughout over, the continuous generation of signals delivered from each the experiment.

2005 Nature Publishing Group Group 2005 Nature Publishing microcluster might explain the ‘counting phenomenon’ associated Planar bilayers. The mouse MHC class II molecule I-Ek with a GPI ‘anchor’ © with TCR downregulation, as outlined in the published ‘serial (Ek–GPI) and mouse ICAM-1 with a GPI ‘anchor’ (ICAM-1–GPI) were triggering’ model39. purified from transfected Chinese hamster ovary and baby hamster kidney As the signaling sites for T cell activation are peripheral micro- cells, respectively, and were incorporated into dioleoyl phosphatidylcholine clusters, the issue of the function of the immunological synapse liposomes (Avanti Polar Lipids). Planar bilayers containing Ek–GPI and ICAM- remains. Our observation that signaling molecules associated with 1–GPI were formed in a flow cell chamber system (Bioptechs). For quantifica- the TCR complex only in the peripheral microclusters indicates that tion of I-Ek, activated B cells from C3H mice and silica beads (Bangs c-SMAC is not involved in signal transduction, as suggested Laboratories) loaded with dioleoyl phosphatidylcholine liposomes containing before10,40. Whereas the c-SMAC may contribute in some conditions Ek–GPI were stained with fluorescein isothiocyanate–labeled monoclonal k to the stabilization of a small amount of specific TCR-pMHC antibody to I-E (14-4-4) and were analyzed by flow cytometry. The density k interaction41, microclusters could also contribute to this process at a of I-E on the bilayers was 2.3-fold higher than that on activated B cells. Planar 1 m very low agonist pMHC density through the involvement of CD4 and bilayers were loaded overnight at 37 Cwith100 M PCC(88–104) in citrate buffer, pH 4.5; were blocked for 1 h at 37 1C with 5% nonfat dried milk in PB; self peptide–MHC. The ‘pseudo-dimer’ model, in which low-affinity 11,42 and were left open in the assay medium. The mobility of each GPI ‘anchor self peptide–MHC contributes to T cell signaling , may incorporate protein’ was checked before the experiments by analysis using the same proteins self peptide–MHC into TCR triggering and microcluster formation. labeled with fluorescence. All experiments in the planar bilayer were done in In contrast, it has also been proposed that the immunological synapse HEPES-buffered saline containing 1% human serum albumin, 2 mM MgCl2 is involved in TCR downregulation and endocytosis, which may and 1 mM CaCl2. be involved in the downregulation of excess activation10.Ouranalyses 44,45 suggest that the immunological synapse may function in continuous TIRFM imaging. Cells were imaged by TIRFM as described . A beam from a T cell activation by creating new signaling microclusters at the solid-state laser (488 nm and 20 mW; Sapphire 488-20-OPS; Coherent) and an inverted microscope (IX-81, Olympus, Japan) were used. Images were captured periphery and the c-SMAC may support late activation signals. with an electron-bombarded charge-coupled device camera (C-7190-23; Furthermore, given that whereas c-SMAC is relatively stable, Hamamatsu Photonics). Images were recorded and analyzed with AquaCosmos p-SMACs and d-SMACs show dynamic flow of microclusters, we software (Hamamatsu Photonics). The high sensitivity and resolution of our propose a possible function for the c-SMAC as an ‘anchor’ for the TIRFM system allowed us to monitor the formation of fluorescent micro- dynamic movement of cells. The c-SMAC may function in mediating clusters and the signaling process with minimal fluorescence decrease due to the microcluster-induced ‘stop signal’ by stabilizing the mobility of photo bleaching.

1260 VOLUME 6 NUMBER 12 DECEMBER 2005 NATURE IMMUNOLOGY ARTICLES

Fluorescence imaging with lipid bilayers. Cells on the planar bilayers were 5. Vicente-Manzanares, M. & Sanchez-Madrid, F. Role of the cytoskeleton during leuko- fixed at various time points and were incubated overnight with 4% parafor- cyte responses. Nat. Rev. Immunol. 4, 110–122 (2004). maldehyde. Then, the cells were stained for 30 min at 25 1C with primary and 6. Freiberg, B.A. et al. Staging and resetting T cell activation in SMACs. Nat. Immunol. 3, 911–917 (2002). secondary antibodies. Before being stained for intracellular molecules, 7. Allenspach, E.J. et al. ERM-dependent movement of CD43 defines a novel cells were treated for 1 min at 25 1C with 0.05% Triton X-100 in PBS and protein complex distal to the immunological synapse. Immunity 15, 739–750 for 30 min at 25 1C with 3% BSA in PBS for permeabilization and blocking, (2001). respectively. Antibodies were used at the following concentrations: anti-CD3e, 8. Blanchard, N., Di Bartolo, V. & Hivroz, C. In the immune synapse, ZAP-70 controls T cell polarization and recruitment of signaling proteins but not formation of the synaptic 5 mg/ml; anti-pZap70, 1 mg/ml; anti-pY, 2 mg/ml; anti-mouse IgG, 10 mg/ml; pattern. Immunity 17, 389–399 (2002). and anti-rabbit IgG, 10 mg/ml. All images were collected on a Leica DMIRES2 9. Delon, J., Kaibuchi, K. & Germain, R.N. Exclusion of CD43 from the immunological system and data were analyzed and prepared for presentation with Leica synapse is mediated by phosphorylation-regulated relocation of the cytoskeletal adaptor confocal software. moesin. Immunity 15, 691–701 (2001). 10. Lee, K.H. et al. T cell receptor signaling precedes immunological synapse formation. + Science 295, 1539–1542 (2002). Calcium induction assay. AND-Tg CD4 T cells expressing EGFP–CD3z were 11. Irvine, D.J., Purbhoo, M.A., Krogsgaard, M. & Davis, M.M. Direct observation of ligand incubated for 20 min at 37 1C with 20 mM Indo-1 in Hank’s balanced salt recognition by T cells. Nature 419, 845–849 (2002). solution buffer containing 1% BSA. Cells were washed twice, were resuspended 12. Harriague, J. & Bismuth, G. Imaging antigen-induced PI3K activation in T cells. Nat. in the assay medium and were loaded on the planar bilayers. Ratiometric Immunol. 3, 1090–1096 (2002). 13. Costello, P.S., Gallagher, M. & Cantrell, D.A. Sustained and dynamic inositol calcium measurements correlated with EGFP–CD3z images were collected by lipid metabolism inside and outside the immunological synapse. Nat. Immunol. 3, alternate visualization of Indo-1 through 405-nm and 495-nm filters in the cell 1082–1089 (2002). body every 15 s. All images were collected with an Olympus IX81 fluorescence 14. Lee, K.H. et al. The immunological synapse balances T cell receptor signaling and microscope, and fluorescence intensity was measured and calculated with degradation. Science 302, 1218–1222 (2003). 15. Purtic, B., Pitcher, L.A., van Oers, N.S. & Wulfing, C. T cell receptor (TCR) clustering MetaMorph (Universal Imaging). in the immunological synapse integrates TCR and costimulatory signaling in selected Tcells.Proc. Natl. Acad. Sci. USA 102, 2904–2909 (2005). T cell–B cell interactions. B cells were purified from C3H/HeN spleen cell 16. Bunnell, S.C. et al. T cell receptor ligation induces the formation of dynamically samples and were incubated in culture medium containing 10 mg/ml of regulated signaling assemblies. J. Cell Biol. 158, 1263–1275 (2002). lipopolysaccharide and 3 mM PCC. After 24–36 h, dead cells were excluded 17. Huppa, J.B., Gleimer, M., Sumen, C. & Davis, M.M. Continuous T cell receptor from lipopolysaccharide-stimulated B cell samples with Lympholyte-M (Cedar- signaling required for synapse maintenance and full effector potential. Nat. Immunol.

http://www.nature.com/natureimmunology 4, 749–755 (2003). + lane) and were used as APCs for AND-Tg CD4 T cells. B cells were applied on 18. Dustin, M.L. & Shaw, A.S. Costimulation: building an immunological synapse. Science poly-L-lysine-coated coverslips and T cells were subsequently loaded onto the 283, 649–650 (1999). coverslips. After conjugation, cells were fixed with 4% paraformaldehyde 19. Stoll, S., Delon, J., Brotz, T.M. & Germain, R.N. Dynamic imaging of T cell-dendritic at various times and were stained with biotinylated anti-CD3e and Alexa cell interactions in lymph nodes. Science 296, 1873–1876 (2002). 20. Miller, M.J., Wei, S.H., Parker, I. & Cahalan, M.D. Two-photon imaging of lymphocyte Fluor–labeled streptavidin. All images were collected with a Leica DMIRE2 motility and antigen response in intact lymph node. Science 296, 1869–1873 microscope. EGFP–CD3z, EGFP–Zap70 and EGFP–SLP-76 were transduced (2002). into wild-type, Zap70-deficient (p116) and SLP-76-deficient (J14) Jurkat 21. Hailman, E., Burack, W.R., Shaw, A.S., Dustin, M.L. & Allen, P.M. Immature + + T cells, respectively. Jurkat T cells were conjugated with Raji cells prepulsed CD4 CD8 thymocytes form a multifocal immunological synapse with sustained tyrosine phosphorylation. Immunity 16, 839–848 (2002). with staphylococcus enterotoxin E as described above. 22. Heissmeyer, V. et al. Calcineurin imposes T cell unresponsiveness through targeted proteolysis of signaling proteins. Nat. Immunol. 5, 255–265 (2004). Image processing and data analysis. The number of molecules in an 23. Iezzi, G., Karjalainen, K. & Lanzavecchia, A. The duration of antigenic stimulation microcluster was calculated as the fluorescence intensity of a single microcluster determines the fate of naive and effector T cells. Immunity 8, 89–95 (1998). + divided by the average of the fluorescence intensity of a single EGFP molecule. 24. van Stipdonk, M.J. et al. Dynamic programming of CD8 T lymphocyte responses. Nat. Immunol. 4, 361–365 (2003). Kymographs and surface plots were prepared for presentation with the

2005 Nature Publishing Group Group 2005 Nature Publishing 25. Faroudi, M., Zaru, R., Paulet, P., Muller, S. & Valitutti, S. Cutting edge: T lymphocyte Montage and Surface Plot ‘plug-ins’, respectively. Individual microclusters activation by repeated immunological synapse formation and intermittent signaling. © were tracked by ‘Manual tracking’, and the area and intensity of each J. Immunol. 171, 1128–1132 (2003). microcluster were calculated by ‘Analyze particles’ in ImageJ (National 26. Depoil, D. et al. Immunological synapses are versatile structures enabling selective T cell polarization. Immunity 22, 185–194 (2005). Institutes of Health). 27. Bunnell, S.C., Kapoor, V., Trible, R.P., Zhang, W. & Samelson, L.E. Dynamic actin polymerization drives T cell receptor-induced spreading: a role for the signal transduc- Note: Supplementary information is available on the Nature Immunology website. tion adaptor LAT. Immunity 14, 315–329 (2001). 28. Barda-Saad, M. et al. Dynamic molecular interactions linking the T cell antigen ACKNOWLEDGMENTS receptor to the actin cytoskeleton. Nat. Immunol. 6, 80–89 (2004). We thank R. Varma, G. Campi and T. Starr for instructions on the preparation 29. Krummel, M.F. & Davis, M.M. Dynamics of the immunological synapse: finding, of planar bilayers and for sharing data before publication, and H. Yamaguchi establishing and solidifying a connection. Curr. Opin. Immunol. 14, 66–74 (2002). for secretarial assistance. Supported by the Ministry of Education, Culture, 30. Krummel, M.F., Sjaastad, M.D., Wulfing, C. & Davis, M.M. Differential clustering of Sports, Science and Technology of Japan (T.Y., M.T. and T.S.), New Energy CD4 and CD3z during T cell recognition. Science 289, 1349–1352 (2000). Development Organization (M.T.), and the National Institutes of Health 31. Brossard, C. et al. Multifocal structure of the T cell - dendritic cell synapse. Eur. J. (AI043542 and AI044931 to M.L.D.). Immunol. 35, 1741–1753 (2005). 32. Gelkop, S. & Isakov, N. T cell activation stimulates the association of enzymatically COMPETING INTERESTS STATEMENT active tyrosine-phosphorylated ZAP-70 with the Crk adapter proteins. J. Biol. Chem. The authors declare that they have no competing financial interests. 274, 21519–21527 (1999). 33. Sasahara, Y. et al. Mechanism of recruitment of WASP to the immunological synapse and of its activation following TCR ligation. Mol. Cell 10, 1269–1281 Published online at http://www.nature.com/natureimmunology/ (2002). Reprints and permissions information is available online at http://npg.nature.com/ 34. Zeng, R. et al. SLP-76 coordinates Nck-dependent Wiskott-Aldrich syndrome protein reprintsandpermissions/ recruitment with Vav-1/Cdc42-dependent Wiskott-Aldrich syndrome protein activation at the T cell-APC contact site. J. Immunol. 171, 1360–1368 (2003). 35. Griffiths, E.K. & Penninger, J.M. Communication between the TCR and integrins: 1. Monks, C.R., Freiberg, B.A., Kupfer, H., Sciaky, N. & Kupfer, A. Three-dimensional role of the molecular adapter ADAP/Fyb/Slap. Curr. Opin. Immunol. 14, 317–322 segregation of supramolecular activation clusters in T cells. Nature 395, 82–86 (1998). (2002). 2. Grakoui, A. et al. The immunological synapse: a molecular machine controlling T cell 36. Krause, M. et al. Fyn-binding protein (Fyb)/SLP-76-associated protein (SLAP), Ena/ activation. Science 285, 221–227 (1999). vasodilator-stimulated phosphoprotein (VASP) proteins and the Arp2/3 complex link 3. Huppa, J.B. & Davis, M.M. T-cell-antigen recognition and the immunological synapse. T cell receptor (TCR) signaling to the actin cytoskeleton. J. Cell Biol. 149, 181–194 Nat. Rev. Immunol. 3, 973–983 (2003). (2000). 4. Dustin, M.L. & Cooper, J.A. The immunological synapse and the actin cytoskeleton: 37. Dustin, M.L. Stop and go traffic to tune T cell responses. Immunity 21, 305–314 molecular hardware for T cell signaling. Nat. Immunol. 1, 23–29 (2000). (2004).

NATURE IMMUNOLOGY VOLUME 6 NUMBER 12 DECEMBER 2005 1261 ARTICLES

38. Maldonado, R.A., Irvine, D.J., Schreiber, R. & Glimcher, L.H. A role for the immuno- 42. Wulfing, C. et al. Costimulation and endogenous MHC ligands contribute to T cell logical synapse in lineage commitment of CD4 lymphocytes. Nature 431, 527–532 recognition. Nat. Immunol. 3, 42–47 (2002). (2004). 43. Campbell, R.E. et al. A monomeric red fluorescent protein. Proc. Natl. Acad. Sci. USA 39. Valitutti, S., Muller, S., Cella, M., Padovan, E. & Lanzavecchia, A. Serial triggering of 99, 7877–7882 (2002). many T-cell receptors by a few peptide-MHC complexes. Nature 375,148–151 44. Axelrod, D. Cell-substrate contacts illuminated by total internal reflection fluorescence. (1995). J. Cell Biol. 89, 141–145 (1981). 40. Davis, S.J. & van der Merwe, P.A. The immunological synapse: required for T cell recep- 45. Tokunaga, M., Kitamura, K., Saito, K., Iwane, A.H. & Yanagida, T. Single molecule tor signalling or directing T cell effector function? Curr. Biol. 11, R289–R291 (2001). imaging of fluorophores and enzymatic reactions achieved by objective-type total 41. Delon, J. & Germain, R.N. Information transfer at the immunological synapse. Curr. internal reflection fluorescence microscopy. Biochem. Biophys. Res. Commun. 235, Biol. 10, R923–R933 (2000). 47–53 (1997). http://www.nature.com/natureimmunology 2005 Nature Publishing Group Group 2005 Nature Publishing ©

1262 VOLUME 6 NUMBER 12 DECEMBER 2005 NATURE IMMUNOLOGY