Research ArticleTCR Signaling

Dynamics of Subsynaptic Vesicles and Surface Microclusters at the Immunological Synapse

See allHide authors and affiliations

Science Signaling  11 May 2010:
Vol. 3, Issue 121, pp. ra36
DOI: 10.1126/scisignal.2000645


Imaging studies have identified clusters of kinases and adaptor proteins that serve as centers of signaling at the contact points between T cells and antigen-presenting cells (APCs). Here, we report that the kinase ZAP-70 and the adaptor proteins LAT and SLP-76 accumulated in separate clusters at the interface between T cells and coverslips coated with a stimulatory antibody against CD3, a component of the T cell antigen receptor complex. A fraction of LAT was detected in motile vesicles that repeatedly moved to surface microclusters of SLP-76 and the adaptor protein GADS (growth factor receptor–bound protein–related adaptor downstream of Shc), where they exhibited decreased motility. LAT molecules in which the residues tyrosine 171 and tyrosine 191 (which are required for the binding of LAT to GADS) were mutated to phenylalanine did not dwell at clusters of SLP-76. At immunological synapses, LAT-containing vesicles also colocalized with microclusters of SLP-76, as detected in experiments in which laser tweezers were used to position T cell–APC conjugates vertically for high-resolution imaging. Phosphorylation of LAT was most prominent when vesicular LAT colocalized with SLP-76. Indeed, the abundance of phosphorylated LAT within a microcluster of SLP-76 was greatest in those clusters that had more recent interactions with LAT-containing vesicles. Finally, negative signals by the inhibitory receptor ILT2 disrupted the assembly of SLP-76–containing microclusters. Together, these data show that the movement of LAT-containing vesicles is linked to the organization of protein microclusters and suggest an important role for vesicular LAT in the SLP-76 signalosome.


T cells become activated through the interaction of T cell antigen receptors (TCRs) with antigen-derived peptides that are bound to integral membrane proteins encoded by major histocompatibility complex (MHC) class I or class II genes. Ligation of the TCR induces phosphorylation of the TCR complex through CD4- and CD8-p56 Lck complexes, which leads to the recruitment of ζ chain–associated protein kinase of 70 kD (ZAP-70) and the phosphorylation of multiple downstream adaptor proteins including linker of activated T cells (LAT) and Src homology 2 (SH2) domain–containing leukocyte phosphoprotein of 76 kD (SLP-76), the latter of which associates with growth factor receptor–bound protein (Grb2)–related adaptor downstream of Shc (GADS) (16). Early imaging studies demonstrated that the activation of T cells is often accompanied by micrometer-scale clustering of proteins at the interface between the T cell and the antigen-presenting cell (APC) (7, 8), as is also visualized at contact points between other immune cell types (9, 10). More recent, higher-resolution imaging has revealed smaller aggregates of proteins, termed microclusters, within the T cell–APC interface, which is known as the immunological synapse (IS) (1117). TCR signaling is initiated in such microclusters (1518), and the signals are terminated while the microclusters move from the periphery to the center of the IS (14, 15). Signaling by inhibitory receptors on natural killer (NK) cells is also largely confined to small domains within the IS (18), which can be important for the mechanisms by which activating and inhibitory signals are integrated (19). The full range of the heterogeneity, localization, and dynamics of microcluster formation is at an early stage of discovery.

Building on the approach previously established by Bunnell, Samelson, their collaborators, and others (1517), we set out to address the heterogeneity and dynamics of microclusters by using multicolor, live-cell microscopy to visualize the supramolecular organization of fluorescently tagged ZAP-70, SLP-76, GADS, and LAT proteins during the activation of T cells. We found that ZAP-70, LAT, and SLP-76 formed separate clusters in response to ligation of the TCR, with ZAP-70 and SLP-76 found primarily at the cell surface, whereas a substantial fraction of LAT was present in subsynaptic vesicles. Vesicles rich in LAT trafficked rapidly between surface clusters of SLP-76, where they exhibited decreased motility and where phosphorylated LAT (at Tyr191 and Tyr226) was enriched. In this system of coverslips coated with stimulatory antibody against CD3, ZAP-70 formed an array of immobile clusters at the activating surface within which moved vesicles that contained LAT. We also determined that negative signaling through immunoglobulin (Ig)–like transcript 2 (ILT2), an inhibitory receptor found on a subset of mature T cells, disrupted the formation of microclusters of ZAP-70 and SLP-76. Thus, these data suggest that signaling at the IS can involve interactions between subsynaptic vesicles and cell-surface microclusters.


Discrete clusters of adaptor proteins assemble at the T cell IS

Microclusters of proteins that include the central adaptor protein SLP-76 are continuously generated at the periphery of the IS after the initial phase of T cell spreading in response to antigen (1417). Here, we followed the organization of proteins associated with the SLP-76 “signalosome” by cotransfecting Jurkat cells (a human CD4+ T cell leukemia cell line) with plasmid encoding yellow fluorescent protein (YFP)–tagged SLP-76 and a plasmid encoding either LAT or GADS tagged with the red fluorescent protein, mCherry. A monoclonal antibody (mAb) against CD3 (OKT-3) immobilized on glass coverslips was used to stimulate the transfected Jurkat cells, as has been performed by others (11, 13). Images were acquired after the initial antigen-induced spreading response in the Jurkat cells rather than at the initial time of stimulation so that we could monitor specifically how T cell signaling was sustained. Activation-induced clusters of mCherry-tagged GADS colocalized and co-migrated with microclusters of SLP-76-YFP (Fig. 1A and movie S1). This is consistent with the constitutive Src homology 3 (SH3) domain–mediated association of GADS and SLP-76 (6). In contrast, LAT and its binding partner SLP-76 segregated into separate puncta with distinct dynamics (Fig. 1B and movie S2). Clusters of LAT moved between distinct clusters of SLP-76 and were colocalized or juxtapositioned for prolonged periods of time (Fig. 1C and movie S3).

Fig. 1

Heterogeneity in T cell protein clusters. (A to C) Images show the distribution of protein clusters at the interface between transfected Jurkat cells and OKT-3–coated coverslips. (A) Distribution of SLP-76-YFP (green) and GADS-mCherry (red) in Jurkat cells. The cell was imaged 2 min after contact with the coverslip. (B) Distribution of SLP-76-YFP (green) and LAT-mCherry (red) in Jurkat cells. The interference reflection microscopy (IRM) image indicates the extent of the interface between the T cell and the coverslip. The cell was imaged 2.5 min after contact with the coverslip. (C) Movement of a cluster of LAT (red) between four separate clusters of SLP-76 (green) over a 3-min period. Arrows indicate the direction of movement of LAT. The panel on the right traces the movement of the LAT and SLP-76 clusters over the entire time course. (D and E) Distribution of proteins at the immune synapse formed between a transfected Jurkat cell and a SEE-pulsed Raji cell. Fixed conjugates were oriented in an upright position with optical tweezers to enable high-resolution imaging of the intercellular synapse. (D) Distribution at the immune synapse of SLP-76-YFP (green) and LAT-mCherry (red) or (E) of SLP-76-YFP (green) and GADS-mCherry (red). Pearson’s correlation coefficients (Rr) are shown in merged images as the mean ± SEM. Scale bars represent 10 μm, except in (C), where the scale bar represents 2 μm. Panels are representative of at least 10 cells each.

We next examined the organization of these same proteins within intercellular synapses between Jurkat cells and superantigen-pulsed Raji cells, a human B cell line that act as APCs. We used our previously described approach of positioning T cell–APC conjugates vertically with optical tweezers to obtain high-resolution images of intercellular synapses (20). In agreement with the images obtained from the experiments with antibody-coated slides, images of T cell–APC interfaces revealed that GADS colocalized with SLP-76, whereas LAT was organized within domains that were largely segregated from those that contained SLP-76 (Fig. 1, D and E). To measure the degree of colocalization of tagged signaling proteins at the IS, we performed Pearson’s correlation analysis of the red and green fluorescent images of T cell–APC conjugates. The Pearson’s correlation coefficient, Rr, which would be 1 for perfectly colocalized images and −1 for mutually exclusive images, was 0.76 ± 0.03 for GADS and SLP-76 but was only 0.19 ± 0.07 for LAT and SLP-76 (n > 10 conjugates), demonstrating a high degree of colocalization between GADS and SLP-76, but only a low degree of colocalization between LAT and SLP-76.

LAT accumulates within subsynaptic vesicles

The TCR cycles between the plasma membrane and vesicular compartments in resting and activated T cells (21). Similarly, TCR-associated signaling molecules, including LAT and SLP-76, are present at the plasma membrane and within intracellular vesicles. A fraction of SLP-76 accumulates in vesicles upon activation of T cells, whereas vesicles that contain LAT exist in resting T cells (3, 12, 2123). To investigate whether some of the observed protein “clusters” represented membrane-proximal cytoplasmic vesicles rather than proteins directly bound to the cell surface, we stained transfected Jurkat cells that expressed SLP-76 with the fluorescent lipid DiI (1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate), which inserts into the plasma membrane, and allowed the cells to rest for 12 hours to enable the internalization and dispersion of DiI into cytoplasmic vesicles (fig. S1). In this way, we showed that SLP-76 localized predominantly in clusters that were distinct from DiI-stained vesicles in Jurkat cells stimulated through CD3 (Fig. 2A). Instead, plasma membrane–proximal vesicles that were stained with DiI were seen to move between the clusters of SLP-76 with which they transiently colocalized. Analogous behavior was seen for plasma membrane–proximal vesicles stained with the fluorescent lipid DiD (1,1′-dioctadecyl-3,3,3′,3′-tetramethylindodicarbocyanine perchlorate) at intercellular immune synapses between conjugates of live Jurkat and Raji cells (Fig. 2B). A substantial fraction of LAT, but not of SLP-76, was colocalized with DiI-stained vesicles within a confocal “optical slice,” that is, within ~500 nm of the interface between the Jurkat cells and the antibody-coated, activating surface (Fig. 2C and movie S4). We found that 30 ± 7% of LAT existed within intracellular compartments in resting Jurkat cells, consistent with previous reports (22). We observed that this percentage increased upon stimulation of the Jurkat cells by the antigenic surfaces (fig. S2), which was consistent with the increased internalization of activated LAT that is mediated by the E3 ubiquitin ligase c-Cbl (24). In contrast, a small amount of phosphorylated LAT (pLAT) was detected within intracellular compartments in activated cells, confirming that active LAT was primarily present at the plasma membrane of the T cell (fig. S2).

Fig. 2

LAT accumulates within membrane-proximal cytoplasmic vesicles. (A) Distribution of SLP-76-YFP (green) and cytoplasmic vesicles marked by the membrane dye DiI (red) at the interface between transfected Jurkat cells and OKT-3–coated coverslips. (B) Distribution of SLP-76-YFP (green) and cytoplasmic vesicles (red) at the intercellular synapse between a transfected Jurkat cell and an SEE-pulsed Raji cell. Jurkat cells were stained with the membrane dye DiD to visualize cytoplasmic vesicles. (C) Distribution of SLP-76-YFP (blue), LAT-mCherry (red), and cytoplasmic vesicles (green) at the interface between Jurkat cells and OKT-3–coated coverslips. Jurkat cells were transfected with plasmids encoding SLP-76-YFP and LAT-mCherry, and vesicles were visualized with DiI. (D and E) Comparison of fixed cells visualized by TIRF microscopy (green) and wide-field microscopy (red) of cells transfected with plasmids encoding (D) LAT-mCherry or (E) SLP-76-YFP. Yellow indicates where fluorescence from both imaging modalities was detected. (F and G) Distribution of LAT (red) and (F) Rab7 (green) or (G) Rab8a (green) at the interface between fixed Jurkat cells and OKT-3–coated coverslips. Cells were cotransfected with plasmids encoding LAT-mCherry and either of the GFP-tagged Rab proteins. (H) TIRF image of SLP-76-YFP clusters in live transfected cells. Insets show the translocation of selected clusters over a 6-min period. Separate clusters (red and green asterisks) were seen to fuse (red/green asterisk). The maximum projection panel (Max. Project.) on the right shows the tracks of all of the clusters over the time course of the experiment. Scale bars, 10 μm; inset in (B), 2 μm; insets in (H), 5 μm. Panels are representative of at least 10 cells each.

Studies of T cell immune synapses by total internal reflection fluorescence (TIRF) microscopy have demonstrated that LAT accumulates in microclusters at the plasma membrane (15, 16). These surface microclusters are not easily revealed by confocal microscopy, because of the worse signal-to-noise ratio encountered compared to that of the thinner optical section obtained by TIRF microscopy. We therefore used TIRF microscopy to clarify whether the LAT detected by confocal microscopy was distinct from that of cell-surface microclusters. Transfected Jurkat cells expressing fluorescently tagged LAT or SLP-76 were viewed initially by TIRF microscopy (which has a penetration depth of ~130 nm) to identify membrane-associated proteins, and then by wide-field microscopy (which has a depth of focus of ~500 nm), which would additionally reveal surface-proximal vesicles in the same cells (Fig. 2, D and E). TIRF microscopy identified surface microclusters of LAT, as previously described (15, 16). Wide-field microscopy could not detect many of the LAT microclusters, because of increased background fluorescence within the thicker optical section, but instead identified a substantial accumulation of LAT deeper within the cell, that is, in membrane-proximal cytoplasmic vesicles that were not detected by TIRF microscopy (Fig. 2D). In contrast, most clusters of SLP-76 detected by wide-field microscopy (~90%) colocalized with clusters that were also imaged by TIRF microscopy, indicating that most of the cellular SLP-76 was found within surface microclusters (Fig. 2E).

To further determine the nature of the LAT-containing vesicles, we compared the location of LAT with those of markers for different vesicular compartments. We found that 70% of vesicles that contained LAT-mCherry also contained green fluorescent protein (GFP)–tagged Rab7, suggesting that LAT was localized in late endosomes (Fig. 2F). Some 22% of vesicles that contained LAT-mCherry also contained GFP-Rab8a, which indicated that some of the vesicles that contained LAT trafficked directly from the Golgi to the plasma membrane (Fig. 2F). This suggests that most of the vesicular LAT was internalized from the cell surface into Rab7-containing vesicles, whereas a smaller fraction was due to nascent LAT in Rab8a-containing vesicles.

Imaging analyses by others have suggested that clusters of SLP-76 are first internalized into cytoplasmic vesicles before becoming mobile, such that they are not detectable by TIRF microscopy (12). Consistent with this, we also found that some (<10%) SLP-76–containing microclusters disappeared from the plasma membrane, as imaged by TIRF microscopy, while they migrated from the periphery to the center of the IS. However, most microclusters of SLP-76 remained associated with the plasma membrane (that is, they were detectable by TIRF microscopy) and were motile, generally moving centripetally (Fig. 2H and movie S5). Clusters of SLP-76 aligned with a radial arrangement of microtubules at the activating interface (fig. S3), consistent with the centripetal movement of SLP-76 being directed by the cytoskeleton, although this remains to be confirmed experimentally.

Vesicular LAT associates with surface microclusters of SLP-76

Next, we tracked the speeds and trajectories of individual protein clusters with a view to assessing their interactions. Microclusters of SLP-76 formed at the periphery of the interface between the T cell and the coverslip and migrated toward the center (Fig. 3A), as described previously (11, 13, 15, 17). The trajectories of microclusters of GADS fully coincided with those of SLP-76 (Fig. 3A). In contrast, vesicular LAT and vesicles marked by DiI moved rapidly between clusters of SLP-76 (Fig. 3A). Both vesicular LAT and those vesicles stained by DiI moved substantially faster than did clusters of SLP-76 (Fig. 3, B and C, and fig. S4).

Fig. 3

Dynamics of microclusters and vesicles. (A) The movements of microclusters and vesicles were tracked in cells cotransfected with plasmid encoding SLP-76-YFP and either GADS-mCherry (left) or LAT-mCherry (middle), or that were transfected with plasmid encoding SLP-76-YFP and stained for vesicles with DiI (right). Tracks of SLP-76 are shown in shades of green and overlaid with the tracks of the other exogenous proteins or DiI in shades of red. Measurements were taken every 5 s. The dotted lines indicate the extent of the interface between the T cell and the coverslip. Arrowheads indicate the final trajectories of the tracks. (B) Distances traveled by individual microclusters within cells that were cotransfected with plasmids encoding SLP-76-YFP (green tracks) and either GADS-mCherry or LAT-mCherry or that were transfected with plasmid encoding SLP-76-YFP and stained for vesicles with DiI. The trajectories of secondary transfected proteins or membrane dye are shown in red. (C) Summary of the speeds of individual microclusters and vesicles as measured from time-lapse images. (D) Tracks of vesicular LAT in activated Jurkat cells that were cotransfected with plasmids encoding SLP-76-YFP and LAT-mCherry were analyzed. Specifically, the speed of each LAT-containing vesicle was determined when it was moving and was unconnected to clusters of SLP-76 (Free LAT), and this was compared to the speed of the LAT-containing vesicle when it overlapped with or was juxtaposed with clusters of SLP-76 (SLP-76–associated). (E) Plot shows the speeds of individual DiI-stained vesicles before and during contact with clusters of SLP-76. (F) Plot shows the speeds of individual DiI-stained vesicles before and during contact with clusters of SLP-76 at the intercellular immune synapse formed between a Jurkat cell and a SEE-pulsed Raji B cell. These data were obtained after optical tweezers had been used to orient the cell-cell conjugates for high-resolution imaging. (G and H) Kymographs depicting the movement of (G) a LAT cluster and (H) a DiI-stained vesicle and clusters of SLP-76. The left panels in each pair show single clusters of LAT-mCherry or a DiI-stained vesicle (red) and clusters of SLP-76-mYFP (green). Movement of the LAT cluster or the vesicle between the separate SLP-76 clusters is shown as a kymograph in the right panels. The arrow in (H) indicates the rapid translocation of a DiI-stained vesicle from one cluster of SLP-76 to another. Scale bar for the left panels, 2 μm. Scale bars for the right panels: vertical bar, 2 μm; horizontal bar, 1 min. (I) The LAT-deficient Jurkat cell line JCam2 was cotransfected with plasmids encoding SLP-76-YFP and either mCherry-tagged wild-type (WT) LAT or an mCherry-tagged mutant LAT that contained the substitutions Y171F and Y191F (Mut LAT). The plot compares the association time of SLP-76 with vesicles that contained WT or Mut LAT. Numbers above the bars indicate the numbers of interactions between LAT and SLP-76 that were observed to have the indicated duration of contact. (B to F) Graphs represent data from at least six cells each.

The speed of movement of vesicular LAT or vesicles stained by DiI was reduced by more than 50% when they were colocalized with microclusters of surface SLP-76 (Fig. 3, D and E). At intercellular synapses, the speeds of vesicles and microclusters were generally faster than those measured at the interface between cells and antibody-coated coverslips; however, vesicles that contained LAT similarly slowed when they colocalized with SLP-76 microclusters (Fig. 3F). This implied that vesicles that contained LAT slowed down when they were colocalized with surface clusters of SLP-76, as also revealed in kymographs (Fig. 3, G and H). This decreased motility could be caused by direct binding between vesicular LAT and microclusters that contain SLP-76, interactions that involve other proteins within LAT-rich vesicles, or, alternatively, confinement of vesicular LAT to domains in which the SLP-76 microclusters are located.

Thus, we next investigated whether the binding of LAT to SLP-76 could directly influence the association of LAT-rich vesicles with surface clusters of SLP-76. We generated a construct that encoded a mutated LAT protein that was unable to interact with the complex containing SLP-76 and GADS. Specifically, Tyr171 and Tyr191 of LAT (25) were substituted with phenylalanine, which also disrupts the interaction between LAT and Grb2. The Jurkat cell line, JCam2, which is deficient in LAT, was transfected with plasmids encoding mCherry-tagged wild-type or mutant LAT in addition to plasmid encoding SLP-76. Cells that contained mutant LAT had fewer clusters of SLP-76 than did cells that contained wild-type LAT, and these clusters did not move centripetally. In contrast, transfection of cells with plasmid encoding wild-type LAT restored the normal phenotype and dynamics of SLP-76 function. For those clusters of SLP-76 that did form, the duration of their contact with vesicles that contained mutant LAT was significantly less (P = 0.002) than that with vesicles that contained wild-type LAT (Fig. 3I). This suggests that the reduction in the speed of LAT vesicles that was observed when they colocalized with SLP-76 was directly influenced by the interaction between LAT and SLP-76. Thus, either the vesicles of LAT directly bind to surface microclusters of SLP-76, or their movement must be confined by a mechanism that is dependent on the interactions between LAT, SLP-76, and GADS.

The movement of subsynaptic vesicles is defined by the distribution of ZAP-70

LAT is phosphorylated by the protein tyrosine kinase ZAP-70 (3) and, therefore, we next investigated how ZAP-70 was organized in relation to vesicular LAT. In T cells activated by incubation on OKT-3–coated coverslips, ZAP-70 formed an array of clusters that did not show any directed movement, consistent with a previous study (26). The movement of subsynaptic vesicles stained by DiI and of vesicular LAT appeared to be restricted by the presence of ZAP-70 clusters (Fig. 4, A and B, and movies S6 and S7); specifically, vesicles moved along paths between clusters of ZAP-70. The activation of T cells by immobilized mAb against CD3 leads to the formation of immobile surface clusters of the TCR (27). Here, clusters of ZAP-70 colocalized with such stationary clusters of TCR (Fig. 4C). This is consistent with immobile clusters of ZAP-70 representing sites of TCR signaling. The movement of LAT-containing vesicles between such clusters demonstrates that these vesicles are tightly coupled to the architecture of the immune synapse and, in particular, to the organization of TCR–ZAP-70 signaling complexes.

Fig. 4

The movement of subsynaptic vesicles is defined by regions in which ZAP-70 forms clusters. The figure shows the relationship between clusters of ZAP-70 and movement of (A) subsynaptic vesicles or (B) vesicular LAT in Jurkat cells stimulated on OKT-3–coated coverslips. (A) Jurkat cells transfected with plasmid encoding ZAP-70-mCFP (green) were stained with DiI (red). (B) Jurkat cells were transfected with plasmids encoding ZAP-70-mCFP (green) and LAT-mCherry (red). The panels on the right show magnified views of the boxed regions and highlight the track of an exemplar vesicle or cluster of LAT (white line) that was tracked for at least 2 min. (C) Fixed Jurkat cells expressing ZAP-70-mCFP (red) and CD3ζ-YFP (green). Transfectants were activated on OKT-3–coated coverslips for 10 min before fixation. (A and B) Images were taken every 2.5 s. Because the clusters of ZAP-70 were extremely faint, the ZAP-70 images shown are the sum of all ZAP-70 frames acquired in each experiment and were processed in ImageJ to remove background signal from cytoplasmic ZAP-70. Scale bars: left panels of (A) and (B), 10 μm; right panels, 5 μm (C) 10 μm. Panels are representative of at least 10 cells each.

LAT is phosphorylated at clusters of SLP-76

To assess the activation state of LAT, we incubated transfected cells with antibodies against phosphorylated forms of LAT, specifically pTyr191 or pTyr226 (22). In transfected cells that contained fluorescently tagged SLP-76 and LAT, pLAT phosphorylated at Tyr191 or Tyr226 was detected in clusters in which LAT and SLP-76 colocalized (Fig. 5, A to D). Clusters of SLP-76 that were separate from large puncta of LAT were associated with lower amounts of pLAT, whereas any cluster of LAT protein that was separated from SLP-76 rarely contained any detectable pLAT (Fig. 5, C and D). We obtained similar results when cells were additionally stained with DiI, which identified LAT-containing “puncta” as subsynaptic vesicles (fig. S5A).

Fig. 5

Phosphorylation of LAT at the site of contact with clusters of SLP-76. Data and images from the interface between transfected Jurkat cells and OKT-3–coated coverslips. (A and B) Jurkat cell transfectants expressing SLP-76-YFP (green) and LAT-mCherry (red) were activated by OKT-3–coated coverslips. Immunofluorescent staining (blue) with mAbs against LAT phosphorylated on Tyr191 (pY191) (A) or Tyr226 (pY226) (B). (C and D) Plots show the fluorescence intensity of phosphorylated LAT (pTyr191 or pTyr226) staining in Jurkat cells (pTyr191 stain, n = 11 cells; pTyr226 stain, n = 7 cells) transfected with plasmids encoding SLP-76-YFP and LAT-mCherry after activation on coverslips coated with OKT-3. The abundance of pLAT was compared within clusters that contained LAT, SLP-76, or both (LAT/SLP-76). A region that lacked clusters was analyzed to determine the background (backgr.) intensity for each cell. (E) Jurkat cells transfected with plasmids encoding SLP-76-mYFP and LAT-mCherry were activated on coverslips coated with OKT-3. A single cell was selected and imaged every 3 s for 240 s. Cells were fixed and stained for pLAT (pTyr191), and the same cell that was visualized live was re-imaged after fixation and staining. The plot shows the fluorescence intensity above background of staining for pLAT within clusters of SLP-76. The live-cell data were used to determine the time that had elapsed since each SLP-76 cluster had last made contact with a cluster of LAT (n = 10 cells). Bars indicate the mean fluorescence intensity (MFI) ± SEM. Scale bars, 10 μm. (A and B) To enhance the weak staining for pLAT, corresponding images were processed in ImageJ to remove the background signal. (C to E) Fluorescence intensity was determined from unprocessed images.

We next related the extent of phosphorylation of LAT in clusters of SLP-76 to the history of the interactions between the SLP-76 clusters and the vesicles that contained LAT (Fig. 5E). The dynamics of SLP-76 and LAT were visualized in individual live cells, after which the cells were fixed and stained for the presence of pLAT. The intensity of LAT phosphorylation within individual clusters of SLP-76 was then related to the time that had elapsed since the last interaction between SLP-76 and puncta of LAT had occurred. This analysis revealed that the phosphorylation of LAT occurred to a greater extent within those clusters of SLP-76 that had had more recent interactions with puncta of LAT than in those clusters that had less recent interactions (Fig. 5E). Indeed, the extent of pLAT staining inversely correlated with the time that had elapsed since the last interaction among puncta of SLP-76 and LAT. We obtained similar results when the cells were additionally stained with DiI, which identified the puncta of LAT as vesicles (fig. S5B). These data suggested that the phosphorylation of LAT within clusters of SLP-76 was intimately linked to the transient colocalization of surface SLP-76 with LAT-containing vesicles.

Negative signaling through ILT2 inhibits the formation of microclusters

We next set out to test how negative signals affected the formation of microclusters and the localization of the subsynaptic vesicular compartment of LAT. The inhibitory receptor ILT2 (also known as LIR-1, LIRB1, and CD85j) is found on subsets of NK cells and T cells and binds to a broad spectrum of class I MHC proteins (28). In T cells, ligation of ILT2 reduces antigen-induced TCR signal transduction and actin polymerization (29). Ligation of ILT2, expressed in Jurkat transfectants, inhibited the formation of microclusters of ZAP-70 upon stimulation of the cells by surfaces coated with OKT-3 (Fig. 6A). Ligation of ILT2 also led to a large reduction in the number of SLP-76 clusters that were formed (Fig. 6B), which was likely a direct consequence of a block in the formation of clusters of ZAP-70. In the few cells that did form clusters of SLP-76, the movements of these clusters were restricted to small distances back and forth that were not directed toward the center of the interface (Fig. 6C). Furthermore, engagement of ILT2 reduced the speed of vesicles that contained LAT (Fig. 6D). No significant difference was detected between the speed of free LAT and that of LAT associated with clusters of SLP-76 when ILT2 was ligated. In live conjugates of superantigen-pulsed APCs with transfected Jurkat cells that expressed ILT2, ILT2 clustered at the interface between the T cell and the APC and inhibited the formation of SLP-76 microclusters at the IS (Fig. 6E). Inhibition of the formation of SLP-76 clusters was overcome by blocking the ligation of ILT2 with a mAb against class I MHC protein (Fig. 6F). Therefore, as a prototype for the action of inhibitory receptors, ILT2 interrupted the assembly of microclusters at the IS. Similarly, negative signaling upon the ligation of CTLA-4 inhibits the clustering of ZAP-70 (30), which indicates that the actions of inhibitory receptors on T cells may generally act upstream of the formation of protein microclusters.

Fig. 6

Negative signaling by ILT2 inhibits the formation of microclusters and interactions with LAT-containing vesicles. (A) Formation of clusters of ZAP-70 in Jurkat cells transfected with plasmids encoding ZAP-70-mCFP and ILT2-citrine. Cells were stimulated on slides coated with either antibody against CD3 (α-CD3) or a mixture of antibodies against CD3 and ILT2 (α-CD3/α-ILT2). To observe the faint clusters of ZAP-70, 10 consecutive exposures were summed, as described for Fig. 4. (B) Formation of clusters of SLP-76-YFP in Jurkat cells transfected with plasmid encoding SLP-76-YFP alone or with plasmids encoding SLP-76-YFP with ILT2-mCFP (indicated at the top of the graph). The number of SLP-76 clusters was determined at the interface between the transfected cells and the slides that were coated with either antibody against CD3 or a mixture of antibodies against CD3 and ILT2 (as indicated on the x axis). (C) Jurkat cells transfected with plasmids encoding LAT-mCherry, SLP-76-YFP, and ILT2-mCFP were stimulated on slides coated with a mixture of antibodies against CD3 and ILT2. The trace shows the movement of individual SLP-76-YFP (green) and LAT-mCherry (red) clusters over a 3-min period. The trace of the inset panel shows the movement of clusters over a further 3 min. (D) Plots showing the speed of movement of clusters of LAT in Jurkat cells that were transfected with plasmids encoding LAT-mCherry, SLP-76-YFP, and ILT2-mCFP and then stimulated on antibody-coated coverslips. Speed was measured separately for those clusters of LAT that were free and those that were associated with clusters of SLP-76. Cells were stimulated with either mAb against CD3 (right graph) or a mixture of antibodies against CD3 and ILT2 (left graph). (E) Live cell conjugates between SEE-pulsed Raji cells and transfected Jurkat cells expressing either SLP-76-YFP (top row) or SLP-76-YFP and ILT2-mCFP (bottom row). Distribution of ILT2-mCFP (red) is shown overlaid on the transmission image (bottom left panel). The extent of clustering of SLP-76-YFP at the synapse is highlighted by its fluorescence intensity being shown on the indicated glow scale (right panels). (F) Quantification of the number of conjugates of SEE-pulsed Raji cells and Jurkat cells that developed SLP-76–containing clusters at the IS. Jurkat cells were transfected with plasmid encoding SLP-76-YFP alone or with plasmids encoding SLP-76-YFP and ILT2-mCFP. The interaction between ILT2 and MHC protein was blocked in the presence of antibody (10 μg/ml) against class I MHC protein (W6/32). Cells lacking any punctate accumulation of SLP-76-YFP at the IS were scored as negative. Experiments were performed in triplicate, assessing a total of 270 conjugates. Error bars on the graph indicate the SEM. Scale bars, 10 μm, inset, 2 μm.


Here, we described the dynamics of distinct assemblies of the functionally connected TCR-proximal signaling components SLP-76, LAT, and ZAP-70 at the IS. The adaptor protein LAT is an integral plasma membrane protein that links antigen recognition by the TCR with several key downstream signaling components (1, 31). High-resolution imaging studies have shown that antigen receptors and LAT are organized in distinct domains at the plasma membrane, which are juxtaposed, but remain separate, upon antigenic stimulation (32, 33). LAT functionality therefore involves the spatiotemporal regulation of protein assemblies. Moreover, a substantial fraction of LAT accumulates within intracellular vesicles (3, 21, 22). We found that the dynamics of LAT in subsynaptic vesicles was strikingly governed by the architecture of the IS. Specifically, the movement of LAT-containing vesicles was confined to tracks between ZAP-70–containing microclusters that represented TCR activation sites. Also, LAT found in subsynaptic vesicles frequently associated with clusters of plasma membrane–associated SLP-76, where the movement of the LAT-containing vesicles was temporarily confined. These interactions correlated with domains in which LAT was phosphorylated, as visualized by immunofluorescence staining with phosphospecific mAbs. Indeed, the abundance of pLAT within a microcluster of SLP-76 was inversely correlated with the time since the last interaction between that cluster of SLP-76 and LAT-containing vesicles. Together, these data imply an important role for vesicular LAT in the SLP-76 signalosome. Specifically, these data suggest that the phosphorylation of LAT is concentrated where vesicular pools of LAT meet surface microclusters of SLP-76 at the plasma membrane.

There are several possible mechanisms by which vesicular LAT could slow down when colocalized with SLP-76. First, the movement of subsynaptic vesicles is reminiscent of the tracks of individual molecules of CD45, Lck, and surface LAT within the plasma membrane, which revealed membrane microdomains that result from protein-protein networks that can exclude or trap signaling molecules (27). Thus, specific protein-protein binding can create the observed confinement of vesicular LAT and this could be due to direct binding between LAT and the complex that contains GADS and SLP-76. In support of this, vesicles that contained mutant LAT proteins that lacked Tyr171 or Tyr191, which are critical for the binding of GADS and Grb2 to LAT, demonstrated significantly reduced association times with surface clusters of SLP-76, potentially resulting from the inhibition of the direct binding of vesicular LAT to surface SLP-76 or GADS. Alternatively, because TCR activation sites are directly linked to the cytoskeleton (11), the activating T cell surface may be partitioned into distinct microdomains through the involvement of actin-based membrane-skeleton “fences” (34). It is possible that such cytoskeletal corals could confine the movement of vesicles, potentially promoting interactions with surface clusters of protein trapped within the same microdomain; however, this remains to be tested as part of the broader aim to establish how LAT-containing vesicles are trafficked.

Although the correlations among the translocation of LAT to surface SLP-76, the phosphorylation of LAT, its decrease in motility, and the stable association of pLAT with SLP-76 are all consistent with the biochemical cascade outlined in previous studies, our data demonstrate that an interaction between the IS and the subsynaptic vesicular compartment could facilitate a previously uncharacterized physical basis for the spatiotemporal control of the active SLP-76 signalosome. Although the functional implications of the interaction between LAT-containing vesicles and surface clusters of SLP-76 remain to be clarified, it is possible that surface clusters of SLP-76 can “perceive” vesicular LAT differentially from LAT in the plasma membrane, perhaps because of the inverted presentation of the cytoplasmic domain of LAT when it is presented in trans (or antiparallel) compared to the conventional cis (parallel) interaction that would occur for proteins tethered to the same membrane. Regulating the interaction between vesicles and clusters of surface protein could therefore play an important role in limiting or promoting interactions between signaling proteins. Indeed, this is emerging to be the case in numerous cell biological systems (35).

Some studies have indicated that T cell costimulatory signals mediated by the α4β1 integrin very late antigen–4 (VLA-4) alter the migratory dynamics of key signaling molecules at the IS such that interactions between macromolecular clusters of SLP-76 and ZAP-70 are dramatically prolonged (26). Here, we demonstrated that the negative signal generated by the inhibitory receptor ILT2 inhibited the macromolecular assembly of clusters of ZAP-70 and SLP-76 as well as the dynamics and interactions of remaining microclusters of SLP-76 with subsynaptic LAT. These observations are consistent with the known biochemical effects of ILT2 in reducing the extent of phosphorylation of the TCR complex and disrupting cytoskeletal rearrangements. Thus, regulating the assembly and dynamics of such supramolecular interactions may be a common mechanism that is used by accessory receptors to modify the primary immune signaling cascade. Overall, an emerging new theme is that interactions between kinases and adaptors in immune cells can be controlled by the dynamics of supramolecular assemblies rather than by protein-protein interactions alone.

Materials and Methods

Cell culture and constructs

Jurkat cells [American Type Culture Collection (ATCC)] were grown in R10 medium [RPMI 1640 with 10% fetal calf serum (FCS), 2 mM l-glutamine, penicillin, and streptomycin]. LAT-deficient JCam2 cells were a gift from A. Weiss. Plasmids encoding fluorescently tagged constructs of Rab5, Rab8a, and Rab7 were gifts from M. C. Seabra. The coding sequences of the genes encoding GADS, ILT2, LAT, SLP-76, and ZAP-70 were amplified from human complementary DNA (cDNA). SLP-76 cDNA was cloned into the vector pEYFP-N1 (Clontech). Substitution of LAT residues Tyr171 and Tyr191 with phenylalanine was achieved by sequential site-directed mutagenesis by polymerase chain reaction (PCR) with the primers 5′-CATTGATGATTTCGTGAACGTTCCGGAGA-3′ and 5′-CATGGAGTCCATTGATGATTTCGTGAACG-3′ used to mutate Tyr171 and the primers 5′-AGCCGGGAGTTTGTGAATGTGTCCCAGGA-3′ and 5′-CTGGATGGCAGCCGGGAGTTTGTGAAT-3′ used to mutate Tyr191. GADS, ILT2, LAT, and ZAP-70 were tagged at the C terminus by subcloning their cDNAs into pcDNA3.1 (Invitrogen) encoding mCherry, Citrine, or mCFP. Cells were transfected by microporation (Digital Bio Technology) with a single pulse of 30 ms, 1380 V. All experiments investigating SLP-76-YFP were performed in wild-type Jurkat cells and in the SLP-76–deficient mutant Jurkat cell line, J14. Unless indicated otherwise, only the data-sets for wild-type Jurkat cells are shown, with data for J14 cells being equivalent.


Chambered coverslides (Nunc) that had been treated with poly-l-lysine (Sigma) were coated with the CD3-specific mAb OKT-3 (10 μg/ml; ATCC) or with both OKT-3 (10 μg/ml) and mAb (10 μg/ml) against ILT2 (R&D Systems) for 1 hour at 37°C to generate T cell–activating surfaces, and then blocked with R10 medium for 10 min. Cells were imaged by resonance scanning confocal microscopy with laser lines of 405, 488, 514, 546, and 594 nm and a 63× 1.2 numerical aperture (NA) water immersion objective (TCS SP5 RS, Leica). Images were acquired with Leica Application Suite Advanced Fluorescence (Leica) software. Simultaneous imaging of different fluorophores was enabled by sequential line scanning. Immunological synapses within T cell–APC conjugates were visualized at high resolution by orienting the cells with optical tweezers, as previously described (20). Such optical tweezers were generated by coupling an infrared laser beam (980 nm) into the non-scanned beam path of a commercially available confocal microscope (Leica TCS SP5), which had minimal effect on the cells because of the low absorption of water at this wavelength. Tweezers could be moved in all three dimensions by simple optics such as activatable mirrors (xy) and translatable lenses (z). Moreover, confocal imaging could be performed with no restriction during the manipulation of cell conjugates, which therefore enabled the imaging of intercellular T cell–APC ISs at full speed [>1 frame per second (fps)] and resolution (250 nm). TIRF microscopy was performed with laser lines 473 or 532 nm directed through an epifluorescence microscope (IX71, Olympus) fitted with a 60× 1.45 NA oil immersion objective. The penetration depth of the evanescent wave at the normalized intensity 1/e was calculated to be ~130 nm. To mark vesicles, cells were stained with 1 μM DiI (Vybrant DiI; Invitrogen) or DiD (Vybrant DiD; Invitrogen) in R10 medium for 30 min at 37°C. Cells were then washed and incubated overnight at 37°C to facilitate the internalization of DiI. To form conjugates between Jurkat and Raji B cells (ATCC), Raji cells were pulsed with Staphylococcus enterotoxin E (SEE) superantigen (100 ng/ml; Toxin Technology) for 1 hour. Cells that were activated for 10 min (Jurkat-Raji conjugates or Jurkat cells activated on OKT-3–coated surfaces) were fixed by incubation in 2% paraformaldehyde–0.1% glutaraldehyde (Sigma) for 2 min at room temperature. For antibody staining, fixed cells were permeabilized with 0.5% Triton–phosphate-buffered saline (PBS, Sigma) for 1 min and blocked for 1 hour with 5% horse serum. Cells were incubated for 1 hour at 4°C with rabbit polyclonal primary antibody (1 μg/ml), including antibody against pLAT Tyr191 (Upstate), antibody against pLAT Tyr226 (Millipore), or rabbit IgG isotype control (Invitrogen), followed by a 1-hour incubation at 4°C with Alexa Fluor 405–labeled goat secondary antibody against rabbit IgG (Invitrogen). Images were analyzed with Leica confocal software (Leica), Volocity (Improvision), or ImageJ (NIH) software. Live-cell images were acquired at 37°C, 5% CO2, with R10 as the imaging medium.

Statistical analysis

Pearson’s correlation coefficients (Rr) were calculated by intensity correlation analysis with ImageJ. Column statistics were performed with GraphPad software (Prism). Paired t tests were performed to analyze the data in Fig. 3, D to F. Contingency table analysis with a χ2 test was used to compare the data sets in Fig. 3I. Unpaired t tests were performed to analyze the data in Fig. 6D and fig. S4B. Data in Figs. 5, C to E, and 6B and figs. S5A and S5B were analyzed by one-way analysis of variance (ANOVA) between groups followed by a series of t tests. Mean values are shown and error bars represent the SEM. In the statistical analysis, P values greater than 0.05 are indicated as nonsignificant (ns), P values between 0.01 and 0.001 are indicated by double asterisks (**), and P values smaller than 0.001 are indicated by triple asterisks (***).


Acknowledgments: We thank M. J. Hannah (National Institute for Medical Research, London) for pointing out the possibility of differences in antiparallel versus parallel recognition of LAT. We are grateful to M. Seabra for the gifts of plasmids, as indicated. We thank N. Burroughs and members of our laboratories for helpful discussions and M. Spitaler for assistance in our core microscope facility (Facility for Imaging by Light Microscopy; FILM). Funding: This work was supported by the Biotechnology and Biological Science Research Council, the Medical Research Council, the Chemical Biology Centre, a Lister Institute Research Prize (to D.M.D.), and Wolfson Royal Society Research Merit Awards (to P.M.W.F. and D.M.D.). C.E.R. is the recipient of a Wellcome Trust Principal Research Fellowship, and work is supported by a program grant from the Wellcome Trust. Author contributions: M.A.P. performed most of the experiments; H.L., S.O., D.M.O., and S.V.P. generated reagents and performed experiments; M.A.P., C.E.R., and D.M.D. designed experiments; M.A.A.N. and P.M.W.F. helped design experiments that involved optical tweezers; and M.A.P. and D.M.D. wrote the manuscript with help from all of the coauthors. Conflicts of interest: The authors declare no conflicts of interest.

Supplementary Materials

Fig. S1. DiI-stained intracellular vesicles are positioned proximal to the IS.

Fig. S2. The intracellular distribution of LAT.

Fig. S3. Clusters of SLP-76 align with the microtubule cytoskeleton.

Fig. S4. Analysis of the speed of protein cluster at the interface between Jurkat cells and OKT-3–coated coverslips.

Fig. S5. LAT is phosphorylated at the sites of contact between SLP-76 microclusters and LAT-containing vesicles.

Movie S1. Dynamics of SLP-76 and GADS clusters.

Movie S2. Dynamics of SLP-76 and LAT clusters.

Movie S3. Dynamics of a single LAT cluster.

Movie S4. Dynamics of SLP-76, LAT, and intracellular vesicles.

Movie S5. Dynamics of SLP-76 clusters imaged by TIRF.

Movie S6. Dynamics of subsynaptic vesicles compared to those of clusters of ZAP-70.

Movie S7. Dynamics of vesicular LAT in relation to clusters of ZAP-70.

References and Notes

View Abstract

Navigate This Article