Early and Dynamic Polarization of T Cell Membrane Rafts and Constituents Prior to TCR Stop Signals

Cell culture and preparation

D10 T cells and CH27 B cells were maintained as described previously (27). All mice were purchased from The Jackson Laboratory. Primary T cells collected from spleens of B10.BR mice were enriched using nylon wool. Briefly, harvested cells were first passed through a 70-μm nylon cell strainer, followed by one wash with 0.14 M NH₄Cl in 17 mM Tris (pH 7.2) and two washes with complete medium containing DMEM supplemented with 10% FCS and 0.01% 2-ME. The cells were then applied to a prepared nylon wool column, incubated for 1 h at 37°C, and then eluted using complete medium. MHC II^−/− B cells were isolated by negative selection (StemCell Technologies) from the spleen of MHC-deficient mice (strain B6.129-H2^d1Ab1-Ea/J), which lack the entire MHC class II (MHC II) locus (28).

Gene construction and expression

L₁₀-GFP and S₁₅-GFP in pWay20 have been described elsewhere (29). Cherry-labeled actin was provided by Dr. D. Knecht (University of Connecticut, Storrs, CT). A fusion protein containing the pleckstrin homology (PH) domain of Akt fused to the C terminus of enhanced GFP (GFP-PH_Akt) was cloned by amplifying the sequence encoding the PH domain of Akt using RT-PCR (iScript cDNA Synthesis Kit; Bio-Rad). Mouse total RNA was the template and the primers were: AACAACAACAACAACCCGATCACCATCAACAACTTCCGT (coding), and CTAATTCATGGTCACGCGGTGCTTGG (noncoding). The PH_Akt PCR product was subcloned into the SmaI site of the GFP expression vector pWay20 (30). CD4 fusion genes containing CFP were generated by subcloning PCR amplification products into the SmaI site of pWay20_cyan fluorescent protein (CFP), which contains the enhanced GFP coding sequence of pWay20 replaced with that of CFP. Wild-type CD4 and the C422,424S mutant were amplified using cDNA provided by Dr. J. K. Rose (Yale University School of Medicine, New Haven, CT) as the template and the following primers: AGAATGAACCGGGGAGTCCCTTT (coding) and AATGGGGCTACATGTCTTCT (noncoding). A tailless CD4 encoding residues 1 through 418 and lacking the membrane proximal palmitoylation sites was generated using the above coding primer and the following primer for the noncoding strand: GAAGAAGATGCCTAGCCCAA. The TCR ζ-chain-YFP fusion gene was constructed by first amplifying DNA encoding ζ-chain by RT-PCR using mouse total RNA as the template. The primers used for the amplification were: ACATCATCTAGAATGAAGTGGAAAGTGTCT (coding) and TCTAGTGAATTCGTTGTTGTTGTTGTTGAGCTGGGAGCTGCTACT (noncoding). The underlined residues represent an XbaI and EcoRI site that was introduced into the coding and noncoding primers, respectively. Following amplification, the product was subcloned into pcDNA3.1 (Invitrogen Life Technologies) using the XbaI and EcoRI sites of the vector. DNA encoding YFP was placed downstream and in-frame to ζ-chain, flanked by the XbaI and KpnI sites of the vector. Primers used for amplifying YFP have been described previously (11). All clones were verified by DNA sequencing.

Cell conjugation

T cell-B cell conjugates were prepared as described elsewhere (27). For stimulation of D10 T cells with conalbumin-pulsed B cells, CH27 B cells were pulsed with Ag and incubated with D10 T cells as previously described (27). Pretreatments of D10 cells included the following: 10 μM 4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine (PP2; Calbiochem) for 15 min, 10 μM wortmannin (Calbiochem), 5 μM latrunculin B (Lat B; Calbiochem) for 1 h, 10 mM methyl-β-cyclodextrin (MβCD; Sigma-Aldrich) for 30 min, and 5 μg/ml filipin (Sigma-Aldrich) for 30 min. All pretreatments were performed at 37°C at a cell density of 10⁷/ml in RPMI 1640 containing 50 mM HEPES (pH 7.4). Additions were made from 100× stock solutions in DMSO. Blocking of I-A^k in CH27 cells was performed by incubating 10⁶ cells with 1.5 μg of Ab (11-5.2; BD Biosciences) in 100 μl of DMEM containing 50 mM HEPES (pH 7.4) at 37°C for 10 min, followed by washing and mixing with D10 cells.

Fluorescence labeling of cells

T cells were labeled with Indo-1 by incubating cells at a density of 5 × 10⁶/ml for 20 min at 37°C in HBSS (pH 7.4) that contained 1.0 g/L glucose plus 1.0 mM CaCl₂, and Indo-1/AM (Invitrogen Life Technologies) at a final concentration of 6 μM, before washing and resuspending cells in RPMI 1640 containing 10% FCS and 50 mM HEPES (pH 7.4). D10 T cells were labeled with DiIC₁₂ or DiIC₁₆ (Invitrogen Life Technologies) by incubating 10⁶ cells for 5 min on ice with 4 μg/ml of either dye in 100 μl of RPMI 1640 containing 50 mM HEPES (pH 7.4).

For immunofluorescence staining, the samples were first fixed using 2% paraformaldehyde. Staining was performed using 0.5–1.0 μg of Ab in 1.0 ml of PBS containing 10 mM glycine for 45 min. Phosphotyrosine residues were labeled using biotinylated 4G10 (Upstate Cell Signaling Solutions), followed by Texas Red-conjugated streptavidin (Jackson ImmunoResearch Laboratories). GM1 and CD4 were stained using biotinylated cholera toxin B subunit (Ctx; Sigma-Aldrich) and mAb (clone RM4-5; BD Biosciences), respectively, and Texas Red-conjugated streptavidin (Jackson ImmunoResearch Laboratories). Vβ8 was detected by staining with a FITC-conjugated mAb (F23.1; BD Biosciences). TCRζ was detected by staining with a FITC-conjugated mAb (6B10.2; Santa Cruz Biotechnology).

Fluorescence microscopy and image analysis

Confocal microscopy was performed using a Zeiss LSM 510 META microscope. Image processing and measurement were performed using IP Labs software (BD Biosceinces). For detection of GFP and Alexa 488, the samples were excited at 488 nm and wavelengths 530–560 nm were collected in the emission. Texas Red and Cherry were imaged by 543-nm excitation and collection of wavelengths 600–660 nm. Detection of the Ca²⁺ changes by Indo-1 fluorescence was performed by sample excitation at 364 nm and collecting the Indo-1 fluorescence in two channels: channel A (382–478 nm) and channel B (489–542 nm). Increases in the intracellular Ca²⁺ were quantitated by dividing channel A by channel B following subtraction of the respective background signals. In some instances, images were collected by wide-field fluorescence microscopy (see Fig. 9, D–F) using a Nikon Eclipse E400 microscope equipped with a 100× objective (aperture 1.3), and an Olympus DP70 camera for detection. GFP fluorescence was excited between 465 and 495 nm, CFP fluorescence at 425–455 nm, and Texas Red between 540 and 580 nm. Emission wavelengths collected for detection were 505–555, 470–500, and 600–660 nm emission for the GFP, CFP, and Texas Red, respectively. The images were deconvoluted using the IP Labs software.

Capping of the fluorescent probes in the fixed conjugates was quantitated by measuring the average fluorescence intensity of the outer membrane at the cell interface versus that of the remaining plasma membrane. For scoring of capping in the conjugates, a cap was defined as when the average fluorescence intensity at the cell interface was at least 1 SD greater than the average intensity of the remaining membrane.

Induction and preparation of anergic and nonanergic CD4⁺ T cells

Seven week-old sex-matched C57BL/6 mice were injected i.p. with 50 μg of staphylococcal enterotoxin B (SEB; Sigma-Aldrich) or PBS alone as described (31, 32). CD4⁺ T cells were isolated by negative selection on an AutoMACS (Miltenyi Biotec) magnetic cell sorter using a CD4 T cell isolation kit (Miltenyi Biotec). The untouched CD4⁺ T cells were resuspended in DMEM containing 50 mM HEPES. T cell proliferation was tested as described using either plate-bound anti-Vβ8-specific mAb or anti-CD3 Abs (125-C11; BD Biosciences) (33).

Specific capping of T cell membrane rafts bound to unpulsed B cells

Association of T cells with APCs include receptor-ligand interactions other than the TCR and pMHC, and these may provide sufficient signals to polarize T cell membrane rafts in the absence of cognate Ag. We tested this hypothesis by measuring for an Ag-independent capping of T cell membrane rafts by B cell APCs using D10 cells bound to unpulsed CH27 cells. D10 T cells recognize residues 134–146 of hen egg conalbumin in the context of I-A^k (34), which is expressed on the CH27 B cells. We prepared the samples by first cosedimenting the cells, followed by a brief 2-min incubation at 37°C, and then fixation. Flow cytometry showed that ∼10% of the T cells conjugated to the B cells in these conditions, and this was approximately half the conjugation frequency measured for samples containing conalbumin-pulsed B cells and incubated for 20 min before measurement (data not shown). Similar conjugation frequencies for T cells bound to B cell APCs have been noted by others (35).

For detection by confocal microscopy, the T cells were labeled by expressing fusion genes encoding GFP and a minimal membrane-anchoring signal, represented by either the first 10 aa of Lck (L₁₀) or the first 15 aa of Src (S₁₅). The L₁₀ sequence targets GFP to the detergent-resistant membrane raft fraction and the S₁₅ signal limits GFP to the Triton X-100-soluble (TX-100), nonraft membrane pools (29). Approximately 40% of the L₁₀-GFP and <1% of the S₁₅-GFP associated with the detergent-resistant raft fraction in D10 T cells (data available at http://www.omrf.ouhsc.edu/ OMRF/Research/15/LMSF/homepage.html).

Data presented in Fig. 1⇓ show a specific polarization or capping of L₁₀-GFP in D10 cells conjugated to the unpulsed B cells. For example, the confocal images in Fig. 1⇓A (left and middle) show enrichment of L₁₀-GFP, but not S₁₅-GFP, at the contact site between separate D10 and CH27 cells. Quantitating the capping in populations of cell conjugates (Materials and Methods) showed the capping frequency of the T cells labeled with L₁₀-GFP was approximately three times that of the T cells labeled with S₁₅-GFP (Fig. 1⇓B). Furthermore, in Fig. 1⇓B are data showing pretreating the D10 cells with MβCD inhibited the capping, thus demonstrating it was cholesterol dependent. However, returning treated T cells to medium containing serum resulted in a capping frequency for the L₁₀-GFP similar to untreated samples, indicating that the T cells were viable following the initial cholesterol extraction by MβCD. Note the small but significant increase in the frequency of capping for the samples in the add-back experiment (Fig. 1⇓B), which may arise from activation of the T cells following serum starvation during incubation with the MβCD. Similar to the D10 cells, we observed a cholesterol-dependent capping of Ctx in labeled mouse primary T cells conjugated to unpulsed CH27 cells (Fig. 1⇓, A and B). Thus, the results with the D10 cells were not unique to this T cell line.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Visualization of T cell raft capping in fixed conjugates with unpulsed B cells. A, Confocal and differential interference contrast (DIC) images of fixed conjugates containing D10 cells (left and middle) or primary T cells (right) bound to unpulsed CH27 cells. The D10 cells expressed either L₁₀-GFP or S₁₅-GFP. The primary T cell was stained with biotinylated Ctx before conjugation and then Texas Red-labeled streptavidin following fixation. The samples were incubated for 2 min following conjugation and before fixation. B, Capping frequencies of L₁₀-GFP and S₁₅-GFP in D10 cells and Ctx in primary cells conjugated to unpulsed CH27 cells. Capping was scored based on the relative enrichment of fluorophore at the cell interface (Materials and Methods), measured for 60–70 cell conjugates, and averaged from three or more separate trials. Each conjugate was scored as exhibiting capping only when the average fluorescence intensity of the T cell plasma membrane was 1 SD or greater than the average intensity of the remaining outer membrane. Pretreatments of the D10 cells were either 10 mM MβCD for 30 min or MβCD followed by serum add-back (MβCD + serum). The untreated cells contained vehicle alone. The bars represent the average capping frequency for each sample ± SEM. **, p < 0.01; ***, p < 0.001, determined by Student’s t test using the capping frequency of the untreated L₁₀-GFP and untreated Ctx samples as the reference for the D10 and primary T cells, respectively. C, D10 cells double-labeled with L₁₀-GFP and either DiIC₁₆ (top) or DiIC₁₂ (bottom) before conjugation. The graph on the right of each image set in is a plot of the fluorescence intensity of the plasma membrane for L₁₀-GFP (green) and DiIC₁₆ or DiIC₁₂ (red) that was measured around the perimeter of the corresponding T cell. The y-axis represents relative fluorescence intensity of the pixels and has arbitrary units that range in value between 0 and 255. The x-axis represents the distance around the periphery of the cell. The vertical lines indicate the region of the T cell that contacted the B cell. D (left), Confocal and DIC images of detergent-resistant membranes in a T cell conjugated to an unpulsed B cell. Jurkat T cells expressing L₁₀-GFP were conjugated to unpulsed Raji B cells for 2 min and then extracted with 1% TX-100 for 10 min at 4°C. To measure for extraction of the TX-100-soluble fraction, the sample was fixed and immunostained with Ab specific to CD45 and Texas Red-conjugated secondary Ab. (right) Detection of CD45 in untreated Jurkat cells. The arrowheads in A–D indicate where the T cells contacted the B cells. The scale bars represent 5 μm.

We performed additional controls to determine whether the T cell capping by unpulsed B cells is specific to membrane rafts and not an artifact due to membrane ruffling or changes in cell morphology. First, D10 cells expressing L₁₀-GFP were double-labeled with either DiIC₁₆ or DiIC₁₂. These dyes preferentially label the raft and nonraft fractions of cell membranes, respectively (36), and we observed cocapping of only the DiIC₁₆ with L₁₀-GFP (Fig. 1⇑C). In a separate experiment, we treated conjugates containing unpulsed B cells with a chilled 1% TX-100 solution to visualize the detergent-resistant membrane fraction (9). Fig. 1⇑D shows the L₁₀-GFP was retained at the cell interface following extraction. Note in Fig. 1⇑D the staining with Ab to CD45 shows the treatment with TX-100 was effective at removing proteins in the TX-100-soluble fraction.

Capping in T cells occurs independently of MHC

APCs can stimulate T cells with noncognate Ag or self-ligands (37), which could account for the capping in Fig. 1⇑ by unpulsed B cells. We therefore tested this possibility by determining whether the capping that we observed in conjugates with conalbumin-free B cells was MHC II-dependent. First, D10 T cells were conjugated to primary B cells isolated from MHC II-deficient mice, in which case capping of L₁₀-GFP in the T cells was still evidenced (Fig. 2⇓A), and the frequency of capping was unchanged relative to conjugates containing B cells expressing MHC II (Fig. 2⇓C). Moreover, cocapping of I-A^k in the CH27 cells with T cell L₁₀-GFP, which is a signature of Ag presentation (Fig. 2⇓B, top) (27), was absent in conjugates containing unpulsed B cells (Fig. 2⇓B, bottom). In another experiment, CH27 cells were preincubated with Ab to I-A^k to block interactions with the TCR, and this treatment also had no effect on the frequency of capping in T cell-B cell conjugates (Fig. 2⇓C). Importantly, the conditions for Ab blocking of I-A^k were sufficient as evidenced by effective inhibition of a conalbumin-dependent increase in capping frequency (20-min conjugations with Ag; Fig. 2⇓C) and L₁₀-GFP enrichment in the caps (Fig. 2⇓D). These data therefore establish that early T cell capping in the absence of added Ag occurs independently of MHC II.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

MHC II-independent capping of rafts in T cells. A and B, Confocal images of D10 cells expressing L₁₀-GFP and bound to either a (A) MHC II^−/− primary B cell or (B) conalbumin-pulsed (top) or unpulsed (bottom) CH27 cells. The samples in B were stained with Ab to I-A^k and a Texas Red-labeled secondary Ab following conjugation and fixation. The unpulsed and pulsed conjugates were incubated for 2 and 20 min, respectively, before fixing. C, Measured capping frequencies of D10 cells expressing L₁₀-GFP and conjugated to either Ag-pulsed or unpulsed CH27 cells. In one sample, the D10 cells were conjugated to MHC II-deficient primary B cells (MHC II^−/−). αI-A^k represents samples where MHC II was blocked with Ab before conjugation to the T cells. Each sample represents the measurement of ∼70 conjugates in three separate trials. The 2-min conjugates containing unpulsed CH27 cells and control samples without pretreatments are presented in Fig. 1⇑. The error bars represent ± SEM. **, p < 0.01, determined by Student’s t test using the capping frequency of the untreated L₁₀-GFP sample as reference. D, Histograms of the relative L₁₀-GFP enrichment at the cell interface measured in D10 cells conjugated to either unpulsed (−Ag) or Ag-pulsed (+Ag) CH27 cells and incubated for either 2 or 20 min following conjugation and before fixation. The relative enrichment of L₁₀-GFP at the cell interface was calculated for each conjugate by dividing the average fluorescence intensity of the T cell outer membrane where it contacted the B cell by the average intensity of the remaining plasma membrane.

MHC-independent capping of T cell membrane rafts requires actin polymerization and signals from Src kinases and PI3K

Because capping of T cell rafts is actin dependent (11), we next measured for an Ag-independent capping of T cell actin by unpulsed B cells. Using D10 cells coexpressing Cherry-labeled actin (Cherry-actin) and L₁₀-GFP, we observed a cholesterol-dependent cocapping of these molecules by unpulsed CH27 cells (Fig. 3⇓A). Furthermore, pretreating the T cells, unpulsed B cells, or both with latrunculin B (Lat B) before conjugation effectively inhibited the Ag-independent capping of L₁₀-GFP in the D10 cells (Fig. 3⇓B). The inhibition mediated by pretreating the B cells with Lat B correlates with earlier studies showing that T cell activation requires an intact APC cytoskeleton (27, 38). Altogether, these data show that the Ag-independent capping in T cells requires an intact actin cytoskeleton in both cells of the conjugates.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Ag-independent capping of L₁₀-GFP is actin dependent and requires signals from Src family kinases and PI3K. A, Confocal and DIC images of D10 cells that coexpressed L₁₀-GFP and Cherry-actin. The T cells were conjugated to an unpulsed (−Ag) CH27 cell for 2 min or an Ag-pulsed (+Ag) CH27 cell for 20 min. In the bottom row, the T cell was pretreated with MβCD before conjugating to an unpulsed CH27 cell for 2 min. The arrowheads indicate where the T cells contacted the B cells, and the scale bar represents 5 μm. The accompanying graphs (right) are a plot of the GFP (green) and Cherry (red) fluorescence intensity values measured around the circumference of the T cells. The vertical bars indicate the boundaries of where the T cells contacted the B cells. Note in the MβCD-treated sample an increase in the cytoplasmic signal of the Cherry-actin. This is representative of other cells in the sample because the average cytoplasmic signal of the Cherry-actin increased from 40 to 80% following pretreatment with MβCD and suggests that removal of cholesterol disrupts signals that are necessary for maintaining the actin cytoskeleton. B, Capping frequencies of L₁₀-GFP in D10 cells bound to unpulsed CH27 cells. The T cells, B cells, or both were treated with Lat B prior to conjugation. The values represent the average of three separate trials totaling >70 conjugates for each sample. The error bars represent ± SEM. C, D10 T cells expressing L₁₀-GFP and bound to unpulsed CH27 cells. Following conjugation and fixation, the samples were immunostained with biotinylated 4G10 and Texas Red-labeled streptavidin. The samples were either untreated (left) or pretreated with PP2 (right). These data are representative of ∼30 separate T cell-B cell conjugates, where the PP2 caused a 4-fold decrease in the frequency of conjugates containing phosphotyrosine-enriched caps from 63% in untreated samples to 13% in those treated with the PP2. D, Capping frequencies of L₁₀-GFP in D10 cells pretreated with either PP2 or wortmannin (WM) and bound to unpulsed CH27 cells. The values represent the average of three separate trials totaling >70 conjugates for each sample. For B and D, the untreated samples are also represented in Fig. 1⇑. ***, p < 0.001 and **, p < 0.01 determined by Student’s t test, using L₁₀-GFP in untreated samples as the reference. E, Confocal and DIC images of D10 cells expressing GFP-PH_Akt. The T cells were conjugated to either unpulsed (middle) or Ag-pulsed (bottom) or CH27 cells. The top panel contains a control showing a D10 cell alone. The arrowheads indicate where the T cell contacted the B cell. Approximately 25% of the total fluorescence signal occurred in the outer membrane in T cells conjugated to either pulsed or unpulsed B cells, but only 10% of the signal was membrane associated in the unconjugated T cells.

One interpretation of the data in Fig. 3⇑, A and B, is that the unpulsed B cells promote T cell signaling events that drive capping of the T cell actin cytoskeleton and cocapping of membrane rafts (4, 9). Consistent with this hypothesis, a PP2-sensitive enrichment of phosphotyrosine signals cocapped with L₁₀-GFP where T cells contacted unpulsed CH27 cells (white arrowheads, Fig. 3⇑C). We also observed that capping of the phosphotyrosine signals was cholesterol dependent, because pretreating T cells with filipin reduced the frequency of conjugates with phosphotyrosine-enriched caps from 63% in untreated cells (n = 25) to only 16% (n = 25) in the treated samples. Finally, the PP2 also inhibited the capping of L₁₀-GFP in T cells bound to unpulsed B cells (Fig. 3⇑D), thus indicating the Ag-independent capping requires the phosphotyrosine signals detected in Fig. 3⇑C.

We also observed an inhibition of the Ag-independent L₁₀-GFP capping by pretreating the T cells with wortmannin (Fig. 3⇑D), thus suggesting it is PI3K dependent. We therefore further tested for PI3K activation during Ag-independent T cell-B cell interactions by measuring the relative phosphatidylinositol 3,4,5-triphosphate (PIP₃) levels in the T cell plasma membrane. Particularly, a GFP fusion protein that contained the pleckstrin homology (PH) domain of Akt (GFP-PH_Akt), which specifically binds to PIP₃ (39) was expressed in T cells. In Fig. 3⇑E are representative data showing recruitment of GFP-PH_Akt to the outer membrane of only the D10 cells bound to either Ag-pulsed or unpulsed B cells, thus indicating an increase in PIP₃ in these conditions. In some cases, we observed enrichment of the probe at the cell interface in the conjugates (Fig. 3⇑E, middle), but this was not a consistent finding. Altogether, Fig. 3⇑, C–E shows T cells undergo activation signals when bound to unpulsed B cells and this results in an actin-dependent capping of the T cell rafts.

MHC-independent capping of T cell membrane rafts is transient and precedes TCR stop signals

To better resolve the kinetics of the Ag-independent capping in relation to Ag-dependent events, we performed live cell imaging experiments using L₁₀-GFP- and S₁₅-GFP-labeled D10 cells added to either unpulsed or conalbumin-pulsed CH27 cells. In experiments with unpulsed B cells, the T cells adhered briefly to the B cells for 60–80 s, after which they disengaged and migrated to another cell (supplemental videos 1–3⁴). In contrast, in experiments with Ag-pulsed B cells, the T cells remained adhered to the B cells for many minutes (supplemental video 4), thus indicating that the brief T cell adherence to unpulsed B cells is due to a lack of pMHC-dependent TCR stop signals. Nonetheless, the T cell interactions with unpulsed B cells were sufficient to cause a transient capping of the L₁₀-GFP, but not S₁₅-GFP (Fig. 4⇓, A and B, Fig. 5⇓, and supplemental videos 1 and 2). Pretreating the T cells with filipin showed that the transient capping was cholesterol dependent (Fig. 4⇓C, Fig. 5⇓, and supplemental video 3). Statistically, we observed that ∼65% of the T cells expressing L₁₀-GFP and that bound to unpulsed B cells longer than 30 s exhibited the transient capping, and Student t tests showed that the enrichment of L₁₀-GFP was significant compared with that of the zero time point (p < 0.001; Fig. 5⇓).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

L₁₀-GFP undergoes a specific and cholesterol-dependent transient capping during initial T cell interactions with B cells. A–D, DIC (top) and confocal (bottom) images of labeled D10 cells during binding to unpulsed (A–C) or conalbumin-pulsed CH27 cells (D). In the confocal images, the fluorescence intensities are represented by pseudocolor, and the color assignment for intensity values is shown in the bottom left panel. The D10 cells were labeled with either L₁₀-GFP (A, C, and D) or S₁₅-GFP (B). The sample in C was pretreated with 5 μg/ml filipin for 30 min before adding to the B cells. The arrowheads indicate the region of the T cell outer membrane contacting the B cell. The indicated times are from when the cells first contacted. The scale bar represents 5 μm.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Kinetics of L₁₀-GFP clustering during early T cell-B cell interactions. Averaged results of live cell imaging experiments measuring time-dependent changes in the fluorescence enrichment of labeled D10 T cells by CH27 B cells. The D10 cells were labeled with either L₁₀-GFP (grey, black, dashed grey) or S₁₅-GFP (dashed black). In one sample, the T cells were pretreated with filipin (blue). In another sample, the B cells were pulsed with conalbumin (black). The relative enrichment of L₁₀-GFP at the cell interface was calculated for each frame in the experiment as described in Materials and Methods. The plots represent the averaged values of ∼15 separate trials for each experiment. Time 0 represents when the cells made initial contact. The error bars represent ±SEM. ***, p < 0.001; **, p < 0.01; and *, p < 0.05 determined by Student’s t test. The reference for the samples with cells expressing L₁₀-GFP (grey, black, and dashed grey) was the corresponding time point of the S₁₅-GFP sample (dashed black). The reference for the remaining experiments was the corresponding time in the untreated sample expressing L₁₀-GFP and without Ag. The inset is the sample with Ag extended to >200 s and showing substantially greater enhancement of L₁₀-GFP occurred with Ag stimulation of the T cells.

Interestingly, we observed that Ag had no effect on either the kinetics or the amount of the transient capping relative to that in experiments with unpulsed samples (Fig. 4⇑D, Fig. 5⇑, and supplemental video 4). Nevertheless, samples with Ag exhibited a recapping of the L₁₀-GFP at later time points, which evolved into a morphologically distinct IS. These data are therefore evidence that the transient capping occurs independent of pMHC-dependent signals in the T cell. Similarly, a statistically significant increase in the intracellular Ca²⁺ (p < 0.05), indicative of TCR-dependent signaling, did not occur until immediately before the recapping and after the transient capping was complete (Fig. 6⇓, A and B, and supplemental video 5). This finding is consistent with the Ca²⁺ providing signals for the robust actin rearrangements and membrane capping that form the IS (23), but shows that it is not necessary for the initial transient capping. Finally, pretreating the T cells with filipin blocked the sustained adhesion and recapping (Fig. 7⇓, A and B, supplemental video 6), when Ag was present suggesting that the Ag-independent capping is necessary for generation of TCR-dependent stop signals.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Transient capping of T cell rafts does not require the pMHC-dependent Ca²⁺ flux. A, DIC (top), confocal (middle), and Indo-1 ratio images (bottom) of an L₁₀-GFP-labeled D10 cell during stimulation by an Ag-pulsed CH27 cell. The times represent after the cells first contacted. B, Averaged results from 11 separate experiments measuring time-dependent changes in L₁₀-GFP capping and intracellular Ca²⁺ during Ag presentation. The Indo-1 ratio was calculated as described in Materials and Methods and the values were normalized by dividing by the maximum value measured in each trial. The x-axis represents time after the cells first contacted. The error bars represent ±SEM. Student t tests were performed using time 0 for each set as the reference and ***, p < 0.001; **, p < 0.01; and *, p < 0.05.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Inhibition of early T cell raft capping and TCR stop signals by filipin. A, DIC (top) and confocal (bottom) images of a D10 T cell expressing L₁₀-GFP and added to an Ag-pulsed CH 27 cell. The T cell was pretreated with filipin. The averaged results from 15 separate trials are plotted in B (red) with the data from Fig. 6⇑ using untreated T cells (black). The arrowheads indicate the interface between the T cell and B cell. The error bars represent ±SEM. ***, p < 0.001; **, p < 0.01; and *, p < 0.05, determined by Student’s t test using the corresponding time in the untreated sample as the reference.

Ag-independent capping of CD4 and TCRζ

Because the TCR coreceptor CD4 associates with membrane rafts (18), we tested whether the Ag-independent capping functioned in concentrating CD4 at the APCs in cell conjugates. First, immunostaining of T cells conjugated to unpulsed B cells showed a cholesterol-dependent capping of CD4 (Fig. 8⇓, A and B). Next, we measured for Ag-independent capping of a CD4 tailless which lacks its cytoplasmic domain and does not associate with membrane rafts (18). In Fig. 8C⇓ are data showing the tailless mutant exhibited significantly less Ag-independent capping than wild-type protein. Next, to determine whether the CD4 capping in Fig. 8⇓, A and B, was Lck dependent, we measured the C422,424S mutant of CD4, which does not associate with Lck but maintains raft association (18). Fig. 8⇓C shows that efficient capping of the C422,424S occurred in T cell conjugates with unpulsed B cells. Thus, CD4 association with membrane rafts, but not its association with Lck, is necessary for efficient capping during Ag-independent interactions with B cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

Ag-independent capping of CD4 and TCR ζ-chain. A, Confocal and DIC images of primary T cells conjugated to unpulsed CH27 cells. Following fixation, the T cells were stained with Ab specific to CD4 and Alexa 488-conjugated secondary Ab. B, Capping frequencies of CD4 in untreated or MβCD-treated (+MβCD) primary T cells bound to unpulsed CH27 cells. C, Capping frequencies of wild-type CD4, CD4 tailless, and C422,424S mutants. Each CD4 contained CFP fused at its C terminus for detection. The CD4 genes were expressed in Jurkat T cells and conjugated to unpulsed Raji B cells. B and C, the error bars represent ±SEM and ***, p < 0.001 and *, p < 0.05. D, Fluorescent and DIC images of D10 cells conjugated to either CH27 cells (top, third, and bottom rows) or a MHC II-deficient primary B cell (second row). Only the sample in the third row contained an Ag-pulsed B cell. The T cell in the bottom row was pretreated with 10 mM MβCD. ζ-Chain was detected either by immunostaining with FITC-labeled Ab to ζ-chain (top, third, and bottom rows) or by expressing a YFP-labeled ζ-chain (second row). The filled arrowheads indicate where the T cells contacted the B cells and the open arrowheads indicate microclusters of ζ-chain at the cell interface. The accompanying graphs (right) are plots of the fluorescence intensity values measured around the circumference of the T cells. The white bars in A and D represent 5 μm.

We also observed an Ag-independent membrane enrichment of TCR ζ-chain in conjugates with unpulsed B cells and B cells deficient in MHC II, but which occurred as microclusters rather than the broad caps often observed with CD4 and other membrane raft markers (Fig. 8⇑D). Fig. 8⇑D (second row) also shows Ag-independent microclusters of a YFP-labeled ζ-chain, demonstrating that the microclustering of ζ-chain was not an artifact from immunostaining. T cells conjugated to either MHC II-deficient B cells (Fig. 8⇑D, second row) or MHC II-blocked CH27 cells (data not shown) also exhibited ζ-chain microclusters. Thus, the microclustering occurred independent of pMHC II complexes. Similar microclusters of the TCR ζ-chain have been reported in T cells stimulated by binding to planar membranes containing pMHC (40, 41) and the microclusters were also evident in Ag-stimulated D10 cells (Fig. 8⇑D, third row). The microclusters were the same size in the stimulated T cells and those bound to unpulsed B cells, but they were more abundant in the stimulated cells (average of 10 for pulsed samples and 3 for unpulsed). Similarly, an average of three microclusters were present in T cells bound to either MHC II-blocked CH27 cells or MHC II-deficient primary B cells. Interestingly, the ζ-chain microclusters were not sensitive to MβCD (Fig. 8⇑D, bottom) and the clustering occurred as frequently in the MβCD-treated T cells (61%, n = 25) as that in untreated samples (60%, n = 28).

Ag-independent capping of T cell membrane rafts is inhibited in anergic T cells

Anergic T cells unresponsive to Ag exhibit a poor recruitment of protein kinase θ and LAT to the IS (22). To determine whether T cell anergy also affects the Ag-independent capping, Ag-free conjugates were prepared using anergic primary T cells. T cell anergy was generated by exposing B6 mice to SEB superantigen by i.p. injection, which results in the nondeleted CD4⁺Vβ8⁺ cells becoming anergic as defined by a failure to respond to cross-linking of Vβ8 (Fig. 9⇓A) (33). Mice injected with PBS alone served as the control. We isolated CD4⁺ T cells from control and SEB-injected mice by depletion 1 wk postinjection and detected Vβ8⁺ cells by immunostaining. By staining the rafts with Ctx, we observed that the Ag-independent capping of rafts was inhibited in anergic Vβ8⁺ T cells compared with control groups represented by Vβ8⁺ T cells from PBS-injected mice (Fig. 9⇓B) and Vβ8⁻ T cells from SEB-injected animals (Fig. 9⇓C). Conjugates with anergic T cells also exhibited an inhibition in the Ag-independent clustering of phosphotyrosine signals (Fig. 9⇓D). Quantitation showed 59% (n = 30) of the conjugates with control Vβ8⁺ T cells exhibited capping of phosphotyrosine signals, but only 22% of the Vβ8⁺ conjugates from anergic T cells (n = 32). From these experiments, we conclude that the distinct properties of anergic T cells include an inhibition of Ag-independent capping and associated phosphotyrosine signals.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 9.

Ag-independent capping of L₁₀-GFP was inhibited in anergic T cells. Measurement of total CD4⁺ splenic T cells isolated by negative selection from B6 mice. The mice were injected with either PBS or SEB 7 days before each experiment. A, [³H]thymidine proliferation assays of purified CD4⁺ T cells from PBS (▪)- or SEB (▦)- injected mice. Data are reported as average cpm ± SD and the measurements were performed in triplicate. Cells were bound to plates containing either no Ab (untreated) or Ab to either CD3 or Vβ8. B, Fluorescence and DIC images of purified CD4⁺ T cells bound to unpulsed CH27 cells. The T cells were doubled-labeled with Texas Red-conjugated Ctx and FITC-conjugated Ab specific to Vβ8. The T cells were from either PBS- or SEB (+SEB)-injected mice. The T cell in the lower panel was Vβ8⁻, thus indicating that it was not part of the anergic population of T cells in the sample. The arrowheads indicate the T cell outer membrane where it contacted the B cell. C, Capping frequencies of GM1 in either Vβ8⁺ or Vβ8⁻ primary T cells from SEB-injected mice or Vβ8⁺ T cells from mice injected with PBS. The error bars represent ±SEM and **, p < 0.01, determined by Student’s t test with the PBS sample as the reference. D, Primary T cells from either PBS- or SEB-injected mice conjugated to unpulsed CH27 cells and immunostained with biotinylated 4G10 and Texas Red-labeled streptavidin. Vβ8 was detected by costaining as in B. The scale bars represent 5 μm.