T cell activation requires mitochondrial translocation to the immunological synapse

Abstract

T helper (T_h) cell activation is required for the adaptive immune response. Formation of the immunological synapse (IS) between T_h cells and antigen-presenting cells is essential for T_h cell activation. IS formation induces the polarization and redistribution of many signaling molecules; however, very little is known about organelle redistribution during IS formation in T_h cells. We show that formation of the IS induced cytoskeleton-dependent mitochondrial redistribution to the immediate vicinity of the IS. Using total internal reflection microscopy, we found that upon stimulation, the distance between the IS and mitochondria was decreased to values <200 nm. Consequently, mitochondria close to the IS took up more Ca²⁺ than the ones farther away from the IS. The redistribution of mitochondria to the IS was necessary to maintain Ca²⁺ influx across the plasma membrane and Ca²⁺-dependent T_h cell activation. Our results suggest that mitochondria are part of the signaling complex at the IS and that their localization close to the IS is required for T_h cell activation.

T helper (T_h) cell activation is central to the adaptive immune response. The initiation of a productive T_h cell-based immune response can be divided into several steps. Foreign antigens are transported to T_h cell zones of draining lymph nodes by antigen-presenting cells. Antigens are then presented by antigen-presenting cells through class II MHC complexes until they encounter the corresponding naïve antigen-specific T_h cells (1, 2). Molecular rearrangements after antigen recognition lead to the formation of an organized immunological synapse (IS), which consists of a central cluster of T_h cell receptors (TCR) [central supramolecular activation clusters (c-SMAC)] that is surrounded by a ring of adhesion molecules (peripheral supramolecular activation clusters) (3–5). Whereas spatial molecular reorganization after IS formation has been analyzed in detail, almost nothing is known about mitochondrial redistribution and its function during and after formation of the IS.

After formation of the IS, several signaling cascades are initiated in T_h cells that finally lead to the secretion of IL-2 and other cytokines and an extensive clonal expansion and differentiation (6–8). A necessary step for the activation of T_h cells after TCR engagement is the stimulation of Ca²⁺ entry across the plasma membrane through the opening of calcium release-activated calcium (CRAC)/Orai1 channels (9–12). CRAC/Orai1 channels are gated by Stim1 after depletion of intracellular Ca²⁺ stores (13, 14). It is widely agreed that maximal store depletion will cause maximal CRAC/Orai1 activation, maximal Ca²⁺ signals, maximal IL-2 production, and maximal T_h cell activation (15, 16).

Results

Efficient Ca²⁺-Dependent T_h Cell Activation After Formation of the IS Requires Mitochondrial Ca²⁺ Uptake.

Analyzing Ca²⁺ signals in the T_h cell line Jurkat after formation of an IS with anti-CD3 beads [supporting information (SI) Fig. 6 A and B], we found that increases in intracellular Ca²⁺ concentration ([Ca²⁺]_i) were more sustained after IS formation than after stimulation with the same anti-CD3 antibody or the commonly used anti-CD3 antibody OKT-3 in solution (Fig. 1A), both of which do not induce IS formation. This was unexpected because anti-CD3 antibodies in solution and anti-CD3 beads were similarly efficient to induce the TCR-dependent signaling cascade that leads to Ca²⁺ store depletion (Fig. 1B), which should therefore activate CRAC/Orai1 channels to the same extent. We found similar differences of sustained [Ca²⁺]_i as shown in Fig. 1A with anti-CD3/anti-CD28 beads compared with the same antibodies in solution in primary human T_h cells isolated from peripheral blood (SI Fig. 6C). Accordingly, formation of the IS was also more efficient to induce proliferation of primary T_h cells than the same antibodies in solution (Fig. 1C). The sarcoendoplasmic Ca²⁺ ATPase inhibitor thapsigargin (TG) has been shown to slowly but maximally and irreversibly deplete Ca²⁺ stores (17), which leads to maximal activation of CRAC/Orai1 channels and maximal [Ca²⁺]_i increases (18, 19). Although CRAC channels were maximally activated under these conditions, TG-induced increases of [Ca²⁺]_i were also lower than increases of [Ca²⁺]_i after formation of the IS (Fig. 1A).

Fig. 1.

Efficient Ca²⁺-dependent T_h cell activation after formation of the IS requires mitochondrial Ca²⁺ uptake. (A) Average [Ca²⁺]_i of Jurkat T cells. The [Ca²⁺]_i plateau (sustained [Ca²⁺]_i) at 800–900 s was analyzed for anti-CD3 beads (1,361 ± 7 nM, 183 cells), 1 μM TG (735 ± 39 nM, 1,525 cells), 5 μg/ml OKT-3 (470 ± 15 nM, 321 cells), or 5 μg/ml anti-CD3 mAbs (464 ± 8 nM, 170 cells). Bead-induced [Ca²⁺]_i was significantly higher than TG-, OKT-3-, or anti-CD3-induced [Ca²⁺]_i (unpaired Student's t test, P < 2.5 × 10⁻¹⁶). (B) Average [Ca²⁺]_i of Jurkat T cells in 0 mM Ca²⁺ Ringer's solution to assess the efficiency of Ca²⁺ store depletion. (C) Average of cell proliferation of T_h cells from four blood donors after a 5-day incubation with 2 μg/ml phytohemagglutinin plus 10 units/ml IL-2 (positive control), anti-intercellular adhesion molecule beads (negative control), 5 μg/ml anti-CD3 and anti-CD28 mAbs (costimulation) in solution, or anti-CD3/CD28 beads. Data are shown as percentage of T_h cell proliferation regarding the cell proliferation measured at day 0. Bead stimulation was significantly better than antibody stimulation (Mann–Whitney test, P = 0.000032). (D) Average [Ca²⁺]_i of Jurkat T cells in the presence of 1 μM carbonyl cyanide m-chlorophenylhydrazone (CCCP). No significant difference was found between the sustained Ca²⁺ signals (P > 0.1). (E) Infrared images and rhod-2 fluorescence pictures of cells stimulated by either TG or anti-CD3 beads. Warmer colors indicate higher rhod-2 fluorescence. (F) Statistical analysis of the normalized rhod-2 fluorescence from T cells stimulated by either TG (37 mitochondrial spots) or anti-CD3 beads (90 mitochondrial spots).

The sustainability of the [Ca²⁺]_i rise induced by the IS was completely abolished by inhibition of mitochondrial Ca²⁺ uptake with carbonyl cyanide m-chlorophenylhydrazone (Fig. 1D), a combination of antimycin/oligomycin (SI Fig. 6D), ruthenium red (SI Fig. 6E), or ruthenium 360 (SI Fig. 6F). Under these conditions, formation of the IS reduced steady state [Ca²⁺]_i to the same degree as did TG or anti-CD3 antibodies in solution. This indicates that mitochondrial Ca²⁺ uptake plays an essential role to maintain an efficient Ca²⁺-dependent T_h cell stimulation after formation of the IS. Mitochondrial Ca²⁺ uptake has previously been shown to sustain CRAC/Orai1 channel activity by reducing its Ca²⁺-dependent Ca²⁺ inactivation after nonphysiological activation of channels (20, 21). Mitochondria likely dissipate the microdomain of high [Ca²⁺]_i beneath the plasma membrane and thereby prevent slow CRAC/Orai1 channel inactivation (20, 22). For this reason, mitochondria should take up a larger amount of Ca²⁺ after IS formation than after TG stimulation, which in turn would result in a more efficient reduction of Ca²⁺-dependent CRAC/Orai1 channel inactivation and thereby sustain Ca²⁺ influx through channels for an extended period. Using the mitochondrial Ca²⁺ indicator rhod-2/AM, we tested this hypothesis. To facilitate rhod-2 measurements, cells were electroporated after rhod-2/AM loading, which removes most of the remaining cytosolic rhod-2, allowing a better resolution of mitochondrial Ca²⁺ measurements. As expected, mitochondria took up significantly more Ca²⁺ after IS formation than after TG stimulation (Fig. 1 E and F). Mitochondrial modulation of CRAC/Orai1 channel activity has already been reported after a directed translocation of mitochondria toward the plasma membrane, which allows mitochondria to take up a larger amount of incoming Ca²⁺ directly beneath the mouth of channels and thereby significantly reduces the Ca²⁺-dependent CRAC/Orai1 channel inactivation (23). Therefore, we analyzed mitochondrial localization within T_h cells after IS formation.

The IS and Mitochondria Form a Signaling Complex.

Usually, mitochondria appear to be evenly distributed throughout the cytosol in most cell types. This is, for instance, the case in HeLa cells, in which many thin tubular mitochondrial structures have been observed by using confocal microscopy (24–26) (Fig. 2A). Mitochondrial localization in T cells is different, because they are often localized preferentially in one area of the cell as seen in the confocal and EM pictures (Fig. 2 B and C). This polarized mitochondrial localization may be caused by the rather large nucleus in T cells. It is difficult to resolve the dense interconnected mitochondrial network in T cells even with confocal microscopy. To get an estimate of the localization of all mitochondrial structures within T cells over longer times without significant bleaching, we used epifluorescence measurements such as the one depicted in Fig. 2D.

Fig. 2.

Different intracellular localization profile of mitochondria in HeLa and T cells. (A) Confocal fluorescence image from a single MitoTracker Green/AM-loaded HeLa cell. (B) Confocal fluorescence image of MitoTracker/AM/di-8-ANEPPS colabeled Jurkat T cells. (C) Two examples of EM images obtained from Jurkat T cells. N, nucleus; M, mitochondria; PM, plasma membrane. (D) Infrared and MitoTracker fluorescence images from MitoTracker Green/AM-loaded Jurkat T cells obtained by epifluorescence microscopy. Scale bars indicate magnifications.

Fig. 3A shows that mitochondrial localization was changed after formation of the IS. In almost all Jurkat or primary human T_h cells we observed a directed movement of mitochondria to the plasma membrane as exemplified in Fig. 3A. Quantitative analysis revealed that many mitochondrial structures moved into an area not farther than 1 μm away from the plasma membrane as indicated by the yellow lines in Fig. 3A and the bar graphs in Fig. 3B. After IS formation, mitochondria were, however, not only translocated to the plasma membrane but preferentially to the IS itself, as is evident from the epifluorescence and two-photon pictures in Fig. 3C. Approximately 50% (in case of Jurkat T_h cells) or even 70% (in case of primary human T_h cells) of the mitochondria were found in the small area around the IS marked in red (Fig. 3D). To analyze how intimate this potential interaction between the IS and mitochondria is, we used total internal reflection microscopy (TIRF), which limits detection of fluorescence to the immediate vicinity (<200 nm) of the plasma membrane. T_h cells were seeded onto coverslips coated with either anti-CD3 (to induce IS formation on the coverslip) or IgG (as control) antibodies. Mitochondrial translocation to the IS was observed with anti-CD3-but not with anti-IgG-coated coverslips (Fig. 3E). Quantification of data from all cells revealed a clear difference between anti-CD3 antibody-coated coverslips compared with anti-IgG antibody-coated coverslips (Fig. 3F). To analyze whether mitochondrial translocation depends on Ca²⁺ influx through CRAC/Orai1 channels, we separated Ca²⁺ release and influx during TIRF experiments. Ca²⁺ stores were depleted for 15 min by allowing contact with the anti-CD3-coated coverslip in the absence of extracellular Ca²⁺. At time 0, Ca²⁺ influx was initiated by the readdition of 20 mM Ca²⁺. We observed a large translocation of mitochondria to the IS, which was significantly reduced in the presence of the CRAC/Orai1 channel blockers 2-APB or BTP2 (Fig. 3G). Therefore, we conclude that mitochondrial translocation into the immediate vicinity (distance <200 nm) of the IS depends on Ca²⁺ influx.

Fig. 3.

Focal stimulation of TCR activates mitochondrial translocation not only toward the plasma membrane but also to the IS. (A) Infrared and fluorescence images from MitoTracker Green/AM-loaded Jurkat T cells before and 15 min after stimulation with anti-CD3 beads. (B) Statistical analysis of the subplasma membrane localization of mitochondria (<0.99 μm beneath the plasma membrane) in Jurkat (31 cells) and CD4⁺ (18 cells) T cells before (resting) and 15 min after stimulation with anti-CD3 beads. Bead stimulation was significantly different from resting conditions (paired Student's t test; Jurkat, P = 7 × 10⁻⁵; CD4⁺, P = 3 × 10⁻⁶). (C) Epifluorescence microscopy and infrared pictures of MitoTracker Green during anti-CD3 bead stimulation. The yellow ring labels the plasma membrane. Shown are two-photon microscopy pictures of a Jurkat T cell stained with MitoTracker/AM and anti-CD45-Alexa Fluor 488 mAb (for the plasma membrane). (D) Percentage of mitochondria at the IS (see Inset, area limited by red line) compared with the net subplasma membrane MitoTracker fluorescence (<0.99 μm beneath the plasma membrane) in Jurkat and CD4⁺ T cells after anti-CD3 bead stimulation. (E) TIRF microscopy pictures of single MitoTracker Green/AM-loaded Jurkat T cells that were settled on anti-CD3 or anti-IgG mAb-coated coverslips in the absence of extracellular Ca²⁺ solution for 4 min and then (at 0 min) exposed to 20 mM Ca²⁺. (F) Statistical analysis of the normalized MitoTracker fluorescence from 21 and nine cells analyzed as the one in E. (G) Statistical analysis of the normalized MitoTracker fluorescence from TIRF experiments. Jurkat T cells were settled on anti-CD3 mAb-coated coverslips in the absence of extracellular Ca²⁺ solution for 15 min and then (at 0 min) exposed to 20 mM Ca²⁺ solution. Control (47 cells), 2-APB-treated (50 μM, 10 cells) or BTP2-treated (100 nM overnight incubation, 20 cells) cells are shown. (H) Infrared images and rhod-2 pictures of T cells stimulated by anti-CD3 beads. Red arrow indicates the position of a mitochondrial cluster beneath the IS, and green arrows indicate the position of mitochondrial populations farther away from the IS. Also shown is kinetic analysis of the normalized rhod-2 fluorescence from mitochondria localized either close to or farther away from the IS (seven cells).

The intimate contact between mitochondria and IS should lead to an enhanced uptake of Ca²⁺ by mitochondria in the vicinity of the IS. We tested this in rhod-2-loaded cells. Whereas in most T cells mitochondria were only found close to the IS, we found cells in which some mitochondrial structures were close to the IS and others farther away. Fig. 3H depicts such a cell, and it is obvious that mitochondria close to the IS had higher rhod-2 fluorescence than the mitochondria farther away from the IS. The analysis of rhod-2 fluorescence over time of all cells reveals a clear difference between mitochondria close to and farther away from the IS. The higher intramitochondrial Ca²⁺ is probably caused by the intimate contact between mitochondria and the IS, which exposes mitochondria to higher microdomains of [Ca²⁺]_i.

Coupling Between the IS and Mitochondria Is Required for Ca²⁺ Signaling.

We have previously shown that microtubules are involved in the translocation of mitochondria toward the plasma membrane after artificial stimulation of Ca²⁺ influx through CRAC/Orai1 channels with TG. After nocodazole treatment to disrupt microtubule-based transport, mitochondrial translocation toward the plasma membrane after TG stimulation was completely abolished (23). In contrast, after bead stimulation, intact microtubule-based transport was not necessary for the translocation of mitochondria toward the IS (Fig. 4A Left and B) and also not required for the Ca²⁺-dependent T_h cell activation (Fig. 4C). This shows that, after formation of the IS, other transport mechanisms for mitochondria must exist. Microfilaments (also called actin filaments or actin cytoskeleton) are also used for mitochondrial transport within the cell (27, 28). Latrunculin B is known to inhibit actin polymerization, thereby disrupting actin-dependent cellular functions (29). The treatment of T_h cells with latrunculin B significantly reduced mitochondrial translocation toward the plasma membrane (Fig. 4 A Right and B) and sustained Ca²⁺ signals after bead stimulation (Fig. 4C), demonstrating the importance of mitochondrial localization relative to the plasma membrane for Ca²⁺-dependent T_h cell activation.

Fig. 4.

Sustained [Ca²⁺]_i after IS formation requires actin cytoskeleton rearrangement-dependent mitochondrial translocation to the IS. (A) Infrared and MitoTracker fluorescence images from single MitoTracker Green/AM-loaded Jurkat T cells before and after stimulation with anti-CD3 beads in the presence of 2 μM nocodazole after a preincubation period for 35 min or in the presence of 10 μg/ml latrunculin B after a preincubation period for 15 min. (B) Statistical analysis of the subplasma membrane localization of mitochondria in Jurkat T cells (<0.99 μm beneath the plasma membrane): resting (86 cells), beads (31 cells), beads plus nocodazole (36 cells), and beads plus latrunculin B (19 cells). Bead stimulation was significantly different from either resting conditions (unpaired Student's t test, P = 2 × 10⁻⁶) or cells stimulated in the presence of latrunculin B (P = 0.009). (C) Average [Ca²⁺]_i of Jurkat T cells stimulated with anti-CD3 beads (control, 1,361 cells) or with anti-CD3 beads in the presence of nocodazole (178 cells) or latrunculin B (106 cells). (D) Average [Ca²⁺]_i of Jurkat T cells stimulated with anti-CD3 beads (control, 104 cells) or with anti-CD3 beads in the presence of nocodazole (52 cells) or latrunculin B (38 cells). (E) Infrared, fluorescence, and merged images from single Jurkat T cells in which the actin cytoskeleton was stained by Texas red phalloidin after the stimulation with anti-CD3 beads (control) or with anti-CD3 beads in the presence of nocodazole or latrunculin B. (F) Average [Ca²⁺]_i of Jurkat T cells stimulated with TG or TG plus anti-CD3 beads in the presence (290 or 270 cells, respectively) or absence (260 or 300 cells, respectively) of latrunculin B. (Inset) The complete experiments. The influx and sustained phase of the Ca²⁺ signals are magnified.

Although we did not find any side effect of latrunculin B on TCR-dependent store-operated Ca²⁺ signals using anti-CD3 antibodies in solution, CRAC channel activity, or mitochondrial Ca²⁺ uptake (SI Fig. 7 A–D), latrunculin B clearly reduced IS-induced Ca²⁺ release (Fig. 4D), which could in part explain the low Ca²⁺ signals in Fig. 4C, because reduced Ca²⁺ release from stores may not activate CRAC channels efficiently enough to allow sustained [Ca²⁺]_i elevations. The reduced Ca²⁺ release was likely due to a lower number of TCR that could be focally stimulated by the anti-CD3 bead at the IS, because latrunculin B disrupted the actin rearrangement and subsequent TCR translocation toward the cell–bead contact point (Fig. 4E) as reported previously (29). A reduced TCR engagement probably decreased the InsP₃ production and thereby the subsequent InsP₃ receptor-dependent Ca²⁺ release, which in turn lead to a less efficient CRAC channel activation.

To analyze the importance of the actin cytoskeleton-dependent translocation of mitochondria to the IS independent of the amount of Ca²⁺ store release, we repeated the Ca²⁺ imaging experiments after maximal store depletion with TG (Fig. 4F Inset). Stores were depleted in the presence of TG in 0-Ca²⁺ solution, and Ca²⁺ influx was assessed by the change to 1 mM Ca²⁺ solution. Under these conditions, peak [Ca²⁺]_i signals were generally higher (23) because Ca²⁺ ATPases in the plasma membrane initially cannot counterbalance the massive simultaneous Ca²⁺ influx through fully activated and opened CRAC channels, because the activity of Ca²⁺ ATPases is only up-regulated slowly (30). The important result, however, is that the sustained Ca²⁺ signals (magnified in Fig. 4F) have the highest amplitude if cells were stimulated by TG and beads in the absence of latrunculin B. In this case, in addition to maximal store depletion, a functional IS is generated and mitochondria are translocated to the IS. With latrunculin B present, the IS cannot form (Fig. 4E), and sustained Ca²⁺ signals are lower despite maximal store depletion. TG, in the presence or absence of latrunculin B, did of course also not induce IS formation, and, consequently, sustained Ca²⁺ signals were also lower. We therefore conclude that an actin cytoskeleton-dependent translocation of mitochondria to the IS is required for sustaining Ca²⁺ signals that are necessary for efficient Ca²⁺-dependent T_h cell activation and subsequent cell proliferation.

Discussion

The main result of this study is that efficient Ca²⁺-dependent activation and clonal expansion of T_h cells require an intimate coupling between mitochondria and the IS. Formation of a matured IS does not only induce sustained TCR stimulation by confining the TCR and agonist peptide–MHC complex to a small region of the plasma membrane (4, 31), but it also activates and sustains mitochondrial relocalization to the IS. Both process are required to sustain Ca²⁺ signals through CRAC/ORAI1 channels, which in turn allows the efficient production of cytokines that occurs over a period of hours (3, 32).

Although a quick, directed, and active mitochondrial translocation along microtubules toward the plasma membrane was already observed after Ca²⁺ influx through CRAC channels in T cells (23), we never detected significant increases of mitochondria into an area within 200 nm from plasma membrane after TG or OKT3 stimulation (data not shown) as described here after formation of IS. Such an intimate contact between mitochondria and plasma membrane allows mitochondria to reduce the local accumulation of Ca²⁺ close to the sites that govern CRAC channel inactivation more efficiently and thereby prolong Ca²⁺ entry for an extended period. This intimate contact was mediated by the actin cytoskeleton. In agreement with these results, disruption of actin cytoskeleton reduced Ca²⁺ signals in T cells only after formation of the IS but not after TG or OKT3 stimulation despite maximal store depletion (Fig. 4D and SI Fig. 7). We propose that mitochondria move actively along microtubules toward the membrane until they reach the cortical actin cytoskeleton, where mitochondria are transferred from microtubules to actin cytoskeleton and then translocated to the IS in an actin cytoskeleton-dependent manner.

The efficiency of an antigen-coated surface to activate T cells compared with antigens in solution has been mostly explained by a prolonged TCR stimulation. A faster down-regulation of TCR induced by antibodies in solution has been observed, leading to a less sustained TCR signaling (33, 34). However, it has also been reported that the down-regulation of TCR in the plasma membrane is not significantly altered by antibodies in solution compared with antibodies coated on a surface (32). Liu et al. (35) demonstrated that reduction of TCR on the cell surface is caused by the disruption of TCR-containing vesicle recycling to the membrane rather than enhancement of TCR internalization and subsequent degradation. The down-regulation of ≈20% of TCR at the plasma membrane took ≈30 min (35). In agreement with this result, we did not find any differences in the efficiency of Ca²⁺ store depletion by antibodies in solution compared with bead stimulation. Thus, the decreased Ca²⁺ signals and the inefficient T_h cell activation by antibodies in solution are not explained by a faster down-regulation of TCR or by disruption of the signalosome.

Recent evidence suggests that TCR microclusters rather than the c-SMAC initiate and sustain TCR signaling (5, 29, 36, 37). TCR microclusters are dynamically generated in the peripheral supramolecular activation clusters and move to the c-SMAC where their signaling is switched off after a while. These results raise the question of why T_h cells allow the formation of c-SMAC in the first place, if it is needed neither to initiate nor to sustain TCR signaling. Our data suggest a role for the c-SMAC during the activation of T_h cells. We postulate that one important function of the c-SMAC is to maintain a high polarization of T_h cells to initiate and sustain the actin cytoskeleton-dependent localization of mitochondria in the immediate vicinity of the IS (Fig. 5). The IS–mitochondria signaling complex sustains CRAC/Orai1 activity, which is required for the Ca²⁺-dependent activation of transcription factors and proper T_h cell function. Hence, CRAC/Orai1 channels localized at the IS would be longer active than those far away from the IS. In agreement with this prediction we found that mitochondria close to the IS take up significantly more Ca²⁺ than mitochondria farther away from the IS. In addition, it is likely that CRAC/Orai1 channels are enriched at the IS as is the case for other membrane proteins after cell polarization. Such an enrichment of CRAC/Orai1 channels would make sense because their long-lasting activity requires an efficient reduction of their Ca²⁺-dependent inactivation, which is achieved only by the mitochondrial Ca²⁺ uptake in the immediate vicinity of the channel.

Fig. 5.

Model for Ca²⁺-dependent T cell activation. After formation of the IS, many mitochondria are translocated into the vicinity of the IS. This close coupling facilitates a larger and more sustained Ca²⁺ influx and the concomitant activation of transcription factors like NFAT, AP1, and NF-κB by reducing the Ca²⁺-dependent CRAC/Orai1 channel inactivation of channels localized close to or within the IS. CRAC/Orai1 channels localized far away from the IS (and most likely not close to mitochondria) inactivate much more rapidly and do not contribute to sustained Ca²⁺ signals.

Our results stress the importance of the actin cytoskeleton to coordinate the relative distance between IS and mitochondria. Rather than being part of the signal transduction machinery itself, as has been proposed in T cells and other cells, the actin cytoskeleton may exert much of its “signaling” power through changing the relative distances between signalosomes and/or organelles. Mitochondrial relocalization during IS formation can thus be used to control the Ca²⁺-dependent activation of transcription factors (38).

Experimental Procedures

Research carried out for this study with human material was approved by the local ethics committee.

Cells.

Human Jurkat T cell lines were grown as described previously (39). Human peripheral blood lymphocytes and CD4⁺ T cells were isolated and cultured as previously described (40). To avoid prestimulation, T_h cells (CD4⁺) were negatively purified with the T Cell Negative Isolation Kit from Invitrogen (Karlsruhe, Germany) according to the manufacturer's instructions.

Reagents.

All chemicals not specifically mentioned were of the highest grade from Sigma–Aldrich (Deisenhofen, Germany). Other reagents used in our experiments include fura-2/AM, rhod-2/AM, MitoTracker Green FM, latrunculin B, TG (Invitrogen), OKT3 (ATCC CRL-8001), mouse anti-human CD43-FITC mAb (Dako Cytomation, Hamburg, Germany), ruthenium 360 (Calbiochem, Darmstadt, Germany), CGP37157 (Tocris, Ellisville, MO), and BTP2 (Nycomed, Konstanz, Germany).

Fluorescence Microscopy and Ca²⁺ Imaging.

Ca²⁺ and mitochondrial imaging were carried out as described (23) with a ×40 (Uplan/Apo, N.A. 1.0 oil) or ×100 (Uplan/Apo, N.A. 1.35 oil) objective. TG (1 μM), OKT3 (5 μg/ml), anti-human CD3 mAbs (5 μg/ml), or anti-CD3 beads were used to stimulate Jurkat T cells. An anti-human CD3 mAb/anti-human CD28 mAb (5 μg/ml) mixture or anti-CD3/anti-CD28 beads were used to stimulate peripheral blood lymphocytes. Carbonyl cyanide m-chlorophenylhydrazone (1 μM), an antimycin (2 μM)/oligomycin (1 μM) mixture, ruthenium red (100 μM), or ruthenium 360 (50 μM, 45-min preincubation) were used for disrupting mitochondrial Ca²⁺ uptake. Because of the reported incapability of ruthenium red to cross the plasma membrane (41), cells were electroporated as previously reported (42) to introduce it into cells.

For mitochondrial Ca²⁺ measurements, cells were loaded at 22–23°C for 45 min with 10 μM rhod-2/AM in culture medium with 10 mM Hepes added, washed with fresh medium, electroporated as reported (42) to remove the cytosolic rhod-2, washed with fresh medium twice, stored at room temperature for 10 min, and used immediately. Cells were illuminated at 550 nm with the Polychrome IV Monochromator (TILL Photonics, Gräfelfing, Germany) using SP 546/10 as excitation filter and DCLP 555 as dichroic mirror. The fluorescence emissions at ≈580 nm (HQ 575/30) were captured with a CCD camera (TILL Photonics), digitized, and analyzed using TILL Vision software. Cells were illuminated every 5 s. To facilitate Ca²⁺ measurements in mitochondria with rhod-2, mitochondrial Ca²⁺ export was inhibited by 10 μM CGP 3715. Rhod-2 fluorescence was background-subtracted and normalized to the initial fluorescence values.

Bead Stimulation.

We followed the standard procedure for absorbing proteins on polystyrene microparticles (size >0.5 μm) established by Polysciences (Eppelheim, Germany). Two alterations in the procedure were required to optimize our results. In step 1 we used 100 μl of a 2.5% suspension of beads (≈2.1 × 10⁷ beads), and in step 7 we added 50 μg of the protein to be absorbed. Chemical composition and pH of buffers were the same as recommended in the protocol.

Azid free anti-human CD3, CD28, or intercellular adhesion molecule mAbs (Euroclone, Lugano, Switzerland) were passively coupled to microparticles (diameter = 5.83 μm) to produce stimulatory (anti-CD3 beads, 50 μg), costimulatory (anti-CD3/CD28 beads, either 25 μg/25 μg or 12.5 μg/37.5 μg), and nonstimulatory (anti-intercellular adhesion molecule beads, 50 μg) beads, respectively. They were stored at 4°C in the specified storage buffer until use. Beads were washed twice with PBS before resuspending them in the Ringer's solution used for the experiments. The ratio of beads to cells was between 2:1 and 1:1.

Cell Proliferation.

Proliferation experiments were carried out in 96-well cell culture plates (black/transparent bottom, catalog no. 353948; Becton Dickinson, Heidelberg, Germany). Data points were measured in triplicates. A total of 50,000 cells per well were used in a total volume of 200 μl in each well. Plates were incubated for 4 h (day 0) or 120 h (day 5) at 37°C, 5% CO₂, and 95% humidity. After incubation time, the number of living cells was determined by the CellTiter-Blue assay (catalog no. G8081; Promega, Madison, WI) with a GeniosPro universal microplate reader (Tecan, Crailsheim, Germany) following the manufacturer's instructions, and results are expressed as percentage of relative fluorescence units. Cells were stimulated by phytohemagglutinin (2 μg/ml, PHA-P, catalog no. L9132; Sigma–Aldrich) and human IL-2 (10 units/ml, catalog no. 1204700; Roche, Mannhein, Germany) as a positive control, anti-CD3/CD-28 beads, 5 μg/ml anti-CD3, and 5 μg/ml anti-CD28 antibodies in solution (not coated on the plate). As negative control anti-intercellular adhesion molecule-coated beads or unstimulated cells were analyzed.

Actin Staining.

For labeling F-actin, stimulated T_h cells were precipitated and washed two times with Ringer's solution before allowing them to adhere to polyl-ornithine-coated (0.1 mg/ml in distilled water; Sigma–Aldrich) glass coverslips for 10–15 min at room temperature. Cells were washed twice with PBS (pH 7.4), fixed in 3.7% formaldehyde solution for 10 min at room temperature, washed two or more times with PBS, treated with 0.1% Triton X-100 for 3–5 min, washed two or more times with PBS, and stained with Texas red-X phalloidin (catalog no. T7471; Invitrogen) for 20 min at room temperature. After the staining period, cells were washed at least twice before taking pictures.

Two-Photon Imaging.

Two-photon imaging was carried out exactly as described previously (23).

Evanescent-Wave Imaging.

The TIRF setup was based on an IX70 microscope (Olympus, Hamburg, Germany) equipped with a ×100/1.45 N.A. Plan Apochromat Olympus objective, a TILL-TIRF condenser (TILL Photonics), and an argon 180 laser (Spectra Physics, Darmstadt, Germany) emitting at 488 nm. Images were acquired with a Micromax 512BFT camera (Princeton Instruments, Ottobrunn, Germany) controlled by MetaMorph (Visitron, Göttingen, Germany). The acquisition rate was 0.5 Hz, and the exposure time was 200 ms. Pixel size was 130 nm. The penetration depth was measured as follows. A low density of 100-μm Ø TetraSpec Beads (Invitrogen) was suspended in water containing 0.05% agar to immobilize them. Other beads were attached to a glass coverslip, and a thin layer (≈0.5 mm) of the bead suspension was poured on the glass. This preparation was then viewed in epifluorescence, and the position of the beads was measured from a deconvoluted Z stack that was acquired starting from the glass coverslip up to 1 μm in the bead suspension with an interval of 0.1 μm. Then one image of the beads was made by using TIRF illumination, and the brightness of the beads was plotted against their Z position. The curve was fitted with the exponential equation I_Z = I₀ exp^−Z/d_p, where I is the intensity, Z is the distance, and d_p is the penetration depth. The measured penetration depth was 230 nm with an illumination at 488 nm. Note that this method can be used because the depth of field of the objective (348 nm at λ = 488 nm) is deeper then the laser penetration depth.

Data Analysis and Statistics.

Data were analyzed by using TILL Vision (TILL Photonics), Igor Pro (Wavemetrics, Lake Oswego, OR), NIH ImageJ, and Excel (Microsoft). All values are given as mean ± SEM (number of cells). More than five independent experiments were performed for each experimental condition. In case data points were normally distributed, an unpaired two-sided Student's t test was used. If normal distribution could not be confirmed, a nonparameterized test (Mann–Whitney) was carried out. P values are stated in the figure legends.

Abbreviations

IS

immunological synapse

T_h

T helper

CRAC

calcium release-activated calcium

TG

thapsigargin

TCR

T_h cell receptor

c-SMAC

central supramolecular activation cluster

TIRF

total internal reflection microscopy.