Research ArticleCalcium signaling

Frontrunners of T cell activation: Initial, localized Ca2+ signals mediated by NAADP and the type 1 ryanodine receptor

See allHide authors and affiliations

Science Signaling  13 Oct 2015:
Vol. 8, Issue 398, pp. ra102
DOI: 10.1126/scisignal.aab0863

Calcium signals down to the millisecond

Engagement of the T cell receptor (TCR) stimulates Ca2+ signaling, which is required for T cell activation. The earliest Ca2+ signals are short-lived and localized near the sites of TCR stimulation; later events are longer-lasting and more widespread. Wolf et al. used a combination of fluorescent indicator dyes and microscopy to perform high-resolution imaging of Ca2+ signals that occurred within milliseconds of the TCR stimulation of live mouse and human T cells. Microinjection of cells with the second messenger NAADP, which is generated upon T cell activation, produced a similar spatiotemporal pattern of Ca2+ signals in the absence of TCR activation. Both TCR- and NAADP-dependent signals were markedly reduced by depletion of ryanodine receptors, which are localized in the endoplasmic reticulum, implicating this internal calcium store as a source for the early Ca2+ signals required for T cell activation.

Abstract

The activation of T cells is the fundamental on switch for the adaptive immune system. Ca2+ signaling is essential for T cell activation and starts as initial, short-lived, localized Ca2+ signals. The second messenger nicotinic acid adenine dinucleotide phosphate (NAADP) forms rapidly upon T cell activation and stimulates early Ca2+ signaling. We developed a high-resolution imaging technique using multiple fluorescent Ca2+ indicator dyes to characterize these early signaling events and investigate the channels involved in NAADP-dependent Ca2+ signals. In the first seconds of activation of either primary murine T cells or human Jurkat cells with beads coated with an antibody against CD3, we detected Ca2+ signals with diameters close to the limit of detection and that were close to the activation site at the plasma membrane. In Jurkat cells in which the ryanodine receptor (RyR) was knocked down or in primary T cells from RyR1−/− mice, either these early Ca2+ signals were not detected or the number of signals was markedly reduced. Local Ca2+ signals observed within 20 ms upon microinjection of Jurkat cells with NAADP were also sensitive to RyR knockdown. In contrast, TRPM2 (transient receptor potential channel, subtype melastatin 2), a potential NAADP target channel, was not required for the formation of initial Ca2+ signals in primary T cells. Thus, through our high-resolution imaging method, we characterized early Ca2+ release events in T cells and obtained evidence for the involvement of RyR and NAADP in such signals.

INTRODUCTION

During intracellular Ca2+ signaling, early localized Ca2+ signals precede an increase in global free cytosolic Ca2+ concentration ([Ca2+]i). These initial, short-lived Ca2+ signals are associated with the opening of either single intracellular Ca2+ release channels, termed fundamental Ca2+ release events, or clusters of these channels, termed elementary Ca2+ release events. The principal Ca2+ release channels mediating such events are the d-myo-inositol 1,4,5-trisphosphate receptor (IP3R) and the ryanodine receptor (RyR) (1, 2). The latter are a family of three channels—RyR1, RyR2, and RyR3—which have homology to the IP3R. The plant alkaloid ryanodine shows high affinity for the RyR, and it exerts an inhibitory effect at micromolar concentrations. Well-studied among initial, short-lived Ca2+ signals are Ca2+ sparks, elementary Ca2+ release events that are most frequently associated with the opening of clusters of RyRs in cardiac myocytes. Ca2+ sparks form by the opening of small clusters of RyRs, and they are characterized by an amplitude of ~170 nM, a duration of signal of ~10 to 100 ms, and a spatial spread of ~1 to 3 μm (36). The fundamental Ca2+ release events mediated by RyRs in cardiac myocytes, termed Ca2+ quarks, are believed to originate from the opening of single RyRs (6), and they have been previously characterized as having an amplitude of 37 ± 6 nM and having diameters at half-maximal amplitude of 0.85 ± 0.2 μm or less (7). Elementary Ca2+ release events that are mediated by IP3Rs, which are termed “Ca2+ puffs,” typically have amplitudes of 50 to 600 nM, a diameter of ~6 μm, and a total duration of about 1 s (5).

During the activation of T cells, Ca2+ signaling is initiated by the engagement of a T cell receptor (TCR) on the surface of a T cell with a peptide-bound major histocompatibility complex on the surface of an antigen-presenting cell (APC) at the contact point between the two cells, which is known as the immunological synapse. We previously demonstrated an important role for the second messenger nicotinic acid adenine dinucleotide phosphate (NAADP) in T cell activation. In Jurkat cells, NAADP is formed within ~10 s of their activation (8), and it evokes Ca2+ release at low nanomolar concentrations (9, 10). Furthermore, NAADP signaling is essential for the reactivation of effector T cells (11, 12). However, since the discovery of NAADP in 1987 (13), different candidate target channels, including RyRs, two-pore channels, transient receptor potential channel, subtype mucolipin 1 (TRPML1) or subtype melastatin 2 (TRPM2), have been proposed to lead to an NAADP-dependent increase in [Ca2+]i (14). Nevertheless, the rapid formation of NAADP in different cell types (8, 15, 16) suggests that it plays a major role in early, local Ca2+ signals (17). Indeed, early local Ca2+ signals have been observed upon microinjection of NAADP into T cells (10, 18) and pulmonary arterial smooth muscle cells (19); however, in general, data showing elementary Ca2+ signaling events evoked by NAADP are scarce (18).

In contrast to Ca2+ imaging in large, adherent cells, such as cardiac myocytes, Ca2+ imaging in small cells, such as T cells, is challenging in different aspects. During immunological synapse formation, T cell morphology changes or T cells move. Furthermore, the spherical shape of T cells results in differential fluorescence intensities at central and peripheral parts of the cell. To obtain reliable images of local [Ca2+]i, so-called ratiometric analysis must be performed. For this, two images representing different properties of a Ca2+ dye, or a combination of two Ca2+ dyes, are acquired. To detect local Ca2+ signals, ratio calculation (ratioing) is performed with these two images, taken at almost the same time or, as the best way, at the identical time point. Because we expected local Ca2+ signals to appear within tens or hundreds of milliseconds of cell activation, high temporal resolution was required. Similarly, the Ca2+ imaging system needed to provide spatial resolution that was high enough to detect elementary Ca2+ signals, which originate from small clusters of channels, or even fundamental Ca2+ signals, which originate from the opening of single channels.

Here, we (i) established and optimized a high-resolution Ca2+-imaging technique suited to monitor rapid and local changes in [Ca2+]i during T cell activation; (ii) characterized these short-lived, initial Ca2+ signals; and (iii) obtained evidence of critical roles for RyRs and NAADP in these local signals. The method was developed in experiments performed with human Jurkat cells stimulated by beads coated with anti-CD3 antibody, which were used to mimic T cell activation by an APC during immunological synapse formation. The method was then applied to primary T cells from wild-type, RyR1−/−, and TRPM2−/− mice and to Jurkat cells and was combined with the intracellular application of NAADP. In summary, we developed a high-resolution Ca2+-imaging technique that was used to characterize the nature of the initial Ca2+ signals in activated T cells.

RESULTS

Limited applicability of single-wavelength dyes

To establish and optimize high-resolution Ca2+ imaging, we tested Ca2+ indicators for their photostability and fluorescence dynamics in T cells upon activation of a strong cytosolic Ca2+ signal. Comparison of Fluo-3, Fluo-4, and Fluo-8 indicated that Fluo-4 was the most stable dye during fast image acquisition [~38 frames/s (fps)] and that it was the dye with the highest fluorescence dynamics upon cell activation with the anti-CD3 monoclonal antibody (mAb) OKT3 (fig. S1). When using single-wavelength dyes to determine [Ca2+]i, ratioing of subsequent frames to the first frame (F/F0, where F0 is the first frame) is a typical procedure (20); however, during rapid cellular movement or morphological changes, calculation of F/F0 leads to systematic errors (fig. S2A). The errors occur because of the ratioing of frames with altered cell position to the initial cell position (F0). Thus, some regions of interest (ROIs) close to the cell edge could not be quantified correctly (fig. S2B). In contrast, shape changes and movements were fully visible with cells coloaded with Fluo-4 and Fura Red and subsequent ratiometric analysis (fig. S2A, lower image panel), which enabled accurate analysis of ROIs (fig. S2, C and D). This finding shows that F/F0 ratioing is not applicable to determine local Ca2+ signals in T cells during activation.

Dual-wavelength dyes

As alternatives for single-wavelength dyes, dual-excitation or dual-emission wavelength dyes were considered. The time delay for the excitation of Fura-2 at 340 and 380 nm makes this dye unsuitable for fast imaging. In contrast, application of dual-emission wavelength dyes enables simultaneous acquisition of two (emission) wavelengths if an emission beam splitter is used. However, the emission shift indicators Indo-1 and ACaR (Asante Calcium Red) showed either strong photobleaching during fast acquisition, low intrinsic fluorescence intensity, or weak fluorescence dynamics (fig. S1) and therefore were not suitable for detecting local T cell Ca2+ signals at the highest temporal and spatial precision.

To maintain the principal advantage of emission shift ratioing and the photostability and brightness of Fluo dyes, we coloaded the cells with the two Ca2+ indicators Fluo-4 and Fura Red. To circumvent the different intrinsic fluorescence intensities of the two dyes because of differences in quantum efficiency, Fura Red and Fluo-4 were applied in a molar 2:1 ratio (21). In a cell-free, buffered solution, both dyes showed a very similar distribution pattern to that in Jurkat or primary T cells (fig. S3), suggesting that noise in the ratio values to a very large extent was systemically caused and was not likely a result of unequal distribution of the two dyes throughout the cytosol and nucleus. Thus, pixel-by-pixel ratioing was possible when suitable thresholds were applied. Furthermore, Fura-2 images showed very similar intensity distributions compared to those of Fluo-4 and Fura Red, and ratioing resulted in very similar data sets (fig. S4, A and B). Thus, we suggest that the imaging approach described here is suitable to detect local Ca2+ signals at the highest temporal and spatial precision.

Identification and analysis of local Ca2+ signals

To identify short-lived, initial Ca2+ signals, we used 100 frames from the live-cell imaging analysis, starting from the activating event (contact between the cell and the anti-CD3–coated bead) and stopping before initiation of the global Ca2+ signal, using the maximum intensity tool of Fiji software [Fiji1.47v (22)]. The maximum intensity tool generates one output image with the maximum pixel value over all images in a stack at the particular pixel location (http://imagejdocu.tudor.lu/doku.php?id=gui:image:stacks). Subsequently, pixels with the highest signal intensity values (≥90% intensity range of individual cells) were selected and used to identify regions with the highest local Ca2+ concentrations within each cell. To construct Ca2+ tracings, ROIs with a minimum diameter of 0.65 μm were defined.

Initial, short-lived Ca2+ signals in Jurkat cells

Activation of Jurkat cells (control clone E2) (23) by anti-CD3–coated beads resulted in very rapid local Ca2+ events (Fig. 1A). Sites of local Ca2+ signaling were selected by the maximum intensity projection tool, and ROIs were defined; near and distant ROIs not selected by the maximum intensity projection tool were used as controls (Fig. 1B). Overall, two groups of signals within the cell were observed within the first 15 s after contact between the cell and the bead was made. Within 1 s of contact, local Ca2+ signals were detected near the bead contact zone (Fig. 1A). Within the first 130 ms, local Ca2+ signals merged together and formed connected, but still localized, areas of Ca2+ signaling with peak amplitudes >115 nM Ca2+ (a Fluo-4/Fura Red ratio of >1), especially very close to the plasma membrane. Between 400 and 780 ms after bead contact, these Ca2+ signals declined in intensity and became disconnected (Fig. 1A, middle). Afterwards, locally separated Ca2+ signals more distant from the plasma membrane occurred as separate events (Fig. 1, A to C; ROIs: 1, 2, and 4). In the time period between 8.06 and 8.84 s after bead contact, these signals increased further (Fig. 1A, bottom) before they expanded into global Ca2+ signaling throughout the whole cell at ≥9 s (Fig. 1A, top panel). Ca2+ signals were characterized by amplitudes of 79 ± 3 nM Ca2+ (a Fluo-4/Fura Red ratio of 0.79; n = 93 signals from 11 cells). Because of the 2-Hz filter applied for the analysis of the corresponding Ca2+ tracings, the duration or time taken to reach the peaks of these signals could not be determined; we estimated this time to be ≤500 ms.

Fig. 1 Analysis of short-lived, initial Ca2+ signals in Jurkat cells.

(A to C) Jurkat cells (clone E2) were loaded with Fluo-4 AM (acetoxymethyl ester) and Fura Red AM, and Ca2+ imaging was performed as described in Materials and Methods. Stimulation was performed by the addition of a bead coated with the anti-CD3 mAb OKT3. The bead contact sites are indicated schematically. Images are from a single cell and are representative of 11 cells. (A) Top: Frame scan of initial localized Ca2+ release events in a Jurkat cell upon bead contact. Middle: Frame scan of a magnified region (8 × 8 μm) of the same cell near the bead is presented from the time of bead contact until ~800 ms after contact. Bottom: Frame scan of images from the same cell from 8.06 until 8.84 s after bead contact. In this presentation, five sequential images were merged to generate an average image to decrease noise (compare to fig. S2D). (B) ROIs were selected by maximum intensity projection (Fiji version 1.47) of the first 100 frames. Relevant signals (shown in red) were identified with a threshold of >90% of [Ca2+]i max, and relevant ROIs, as well as control ROIs nearby and distant from the relevant Ca2+ release sites, were constructed. (C) Data from the selected ROIs were subjected to 2-Hz low-pass filtering and are plotted as Ca2+ tracings.

To gain insights into the mechanism(s) underlying the short-lived and initial local Ca2+ signals, we used Jurkat cell clone #10, which had stable expression of an antisense construct silencing the expression of all three RyR subtypes, hereafter termed RyR-deficient Jurkat cells (23). Previously, we reported that RyR-deficient Jurkat cells exhibit decreased Ca2+ signaling upon T cell activation by soluble anti-CD3 as compared to control cells (24). Here, we confirmed this finding, because upon localized activation with an anti-CD3–coated bead, the global Ca2+ response during extended recording periods of ~80 s was much lower in RyR-deficient Jurkat cells than in control cells (Fig. 2 and fig. S5). When imaged at high spatiotemporal resolution, local Ca2+ signals in RyR-deficient Jurkat cells were largely decreased (Fig. 2A); ROIs selected by maximum intensity projection (Fig. 2B) showed similar kinetics as those of control ROIs (Fig. 2C). In summary, short-lived, initial Ca2+ signals in RyR-deficient Jurkat cells had an amplitude of 61 ± 4 nM (Fluo-4/Fura Red ratio, 0.68; n = 46 signals from six cells; Fig. 3A). Furthermore, the numbers of cells that showed short-lived, initial Ca2+ signals were statistically significantly lower for RyR-deficient Jurkat cells (6 of 11 cells, Fig. 3A) than for control cells (local signal in 11 of 11 cells). Indeed, the number of short-lived, initial Ca2+ signals per cell within 15 s of cell stimulation was decreased in RyR-deficient Jurkat cells (4.2 ± 2.2) compared to that in control cells (8.6 ± 2.1; Fig. 3A). This suggests that the loss of RyRs prevented the efficient initiation of short-lived, initial Ca2+ signals, which resulted in largely delayed and diminished global Ca2+ signals upon activation. Similarly, in control Jurkat cells (clone E2), 50 μM ryanodine diminished initial Ca2+ signals in comparison to the vehicle control (n = 10 cells per group; Table 1). The percentage of responsive cells was decreased to 64%, and the amplitude of the initial Ca2+ signals was statistically significantly reduced from 121 ± 5 nM to 92 ± 4 nM by ryanodine (the Fluo-4/Fura Red ratio was reduced from 1 to 0.87; Table 1). However, the number of signals per cell was not affected (6.5 ± 2.0 signals per cell for the control versus 5.9 ± 2.0 signals per cell for ryanodine-treated cells). In contrast to ryanodine, 10 μM xestospongin C, an inhibitor of the IP3R, did not decrease number of responding cells or number of initial Ca2+ signals (Table 1).

Fig. 2 Effect of knockdown of RyRs on short-lived, initial Ca2+ signals.

(A to C) RyR-deficient Jurkat cells (clone #10) were loaded with Fluo-4 AM and Fura Red AM, and Ca2+ imaging was performed as described in Materials and Methods. Stimulation was performed by the addition of a bead coated with the anti-CD3 mAb OKT3. The bead contact site is indicated schematically. Images are from a single cell and are representative of five cells. (A) Top: Frame scan of images over time. Bottom: Magnification of the bead contact zone (6 × 6 μm). In comparison to Jurkat clone E2 (Fig. 1), no short-lived, initial Ca2+ release events were detectable in RyR-deficient Jurkat cells. Similar results were obtained in 5 of 11 cells. In this presentation, five subsequent images were merged to decrease noise (compare to fig. S2D). Note that in 6 of 11 cells, short-lived, initial Ca2+ release events were detected, as detailed in Results. (B) ROIs were selected by maximum intensity projection (Fiji version 1.47) of the first 100 frames. Relevant signals (shown in red) were identified with a threshold of >90% of [Ca2+]i max, and relevant ROIs, as well as control ROIs nearby and distant from the relevant Ca2+ release sites, were constructed. (C) Data from the selected ROIs were subjected to 2-Hz low-pass filtering and are plotted as Ca2+ tracings.

Fig. 3 Short-lived, initial Ca2+ signals in RyR-deficient Jurkat cells, RyR1−/− primary T cells, and wild-type T cells.

(A to C) Primary CD3+ T cells were freshly isolated from the spleens of wild-type (WT) or RyR1−/− mice by negative selection. The cells were loaded with Fluo-4 AM and Fura Red AM, and Ca2+ imaging was performed as described in Materials and Methods. Stimulation was performed by the addition of beads coated with anti-CD3 and anti-CD28 mAbs. The bead contact site is indicated schematically. Data are from single cells and are representative of 14 WT T cells and 8 nonresponding RyR1−/− T cells (of a total of 16 cells analyzed). (A) Characteristics of short-lived, initial Ca2+ signals after the localized activation of Jurkat cells (clones E2 and #10) and primary T cells from WT and RyR1−/− mice by anti-CD3– and anti-CD3/anti-CD28–coated beads. As a negative control, anti-IgG–coated beads were incubated with WT cells (WT neg.) Data are means ± SEM of 9 to 16 cells analyzed. *P < 0.05, **P < 0.01, ***P < 0.001 by the Mann-Whitney U test (for Jurkat cells) or the Kruskal-Wallis test (for mouse primary T cells). (B and C) Top: Frame scans of initial localized Ca2+ release events in (B) a WT T cell and (C) a RyR1−/− T cell upon contact with a bead. Bottom: Magnified regions (4 × 4 μm) of the cell near the bead contact zone were imaged from the time of bead contact until ~800 ms after contact was initiated. In this presentation, five subsequent images were merged to decrease noise (compare to fig. S2D).

Table 1 Summary of short-lived, initial Ca2+ signals in Jurkat cells and primary mouse T cells upon activation by beads coated with anti-CD3 or anti-CD3/anti-CD28 antibodies.

DMSO, dimethyl sulfoxide; XeC, xestospongin C.

View this table:

Role of RyR1 in short-lived, initial Ca2+ signals in primary T cells

Next, we analyzed local Ca2+ signals in primary CD3+ T cells from wild-type and RyR1−/− mice after activation with beads coated with anti-CD3 and anti-CD28 antibodies. Similar to our findings from experiments with Jurkat cells, short-lived, initial Ca2+ signals occurred in 12 of 14 primary wild-type mouse T cells within 1 s after stimulation (Fig. 3, A and B). A second time period of local Ca2+ signaling started at ~8 s after stimulation (Fig. 3A). The number of local Ca2+ signals per cell within 15 s of cell stimulation was 11.4 ± 2.7 (Fig. 3, A and B). These Ca2+ signals were localized very close to the plasma membrane (Fig. 3B) and had an amplitude of 180 ± 5 nM (Fluo-4/Fura Red ratio, 0.9; n = 160 signals from 14 cells) and a diameter very close to the calculated spatial resolution of 368 nm. The kinetics and amplitude of localized Ca2+ signals showed some variations between individual cells (fig. S6). Similarly to RyR-deficient Jurkat cells, initial localized Ca2+ signals of RyR1−/− T cells were largely reduced as compared to those in wild-type T cells (Fig. 3C). The number of responding RyR1−/− T cells was 50% (n = 8 of 16 cells; Fig. 3A). The amplitude of the Ca2+ signals in these cells was reduced to 147 ± 6 nM (Fluo-4/Fura Red ratio, 0.84; n = 66 signals from eight cells), and there were 4.1 ± 1.5 local Ca2+ signals per cell (Fig. 3A). Furthermore, in those RyR1−/− T cells that responded to stimulation, local Ca2+ signals were mainly observed within 1 s of stimulation, but there was no second time period of signaling after ~8 s after stimulation (fig. S7). To rule out the possibility that the T cells were nonspecifically activated by contact with beads, anti–immunoglobulin G (IgG)–coated beads were used as a negative control in experiments with primary T cells from wild-type mice (Fig. 3A). The addition of anti-IgG–coated beads resulted in 0.3 ± 0.2 signals per cell with an amplitude of 57 ± 7 nM (Fluo-4/Fura Red ratio, 0.52; n = 3 signals from 10 cells; Fig. 3A).

Mean responsiveness

To facilitate the comparison of different cell types or cell treatments with respect to short-lived, initial Ca2+ signals, we multiplied the number of signals per cell by the amplitude of the signals to calculate a combined parameter, termed “mean responsiveness.” Knockdown of RyRs or knockout of RyR1 in Jurkat cells or primary mouse T cells, respectively, resulted in a statistically significant decrease in mean responsiveness compared to that of control cells (Fig. 4). Furthermore, measurements in nominally Ca2+-free medium (−[Ca2+]ex) resulted in a substantial decrease in the mean responsiveness of control Jurkat cells (E2 cells); signals per cell were reduced to 2.0 ± 0.7, and the amplitude decreased to 55 ± 5 nM (Fluo-4/Fura Red ratio, 0.7 ± 0.05; n = 39 signals from 20 cells) (Fig. 4A). In the absence of any stimulus, the mean responsiveness of Jurkat cells was very low. Similarly, mechanical stimulation of T cells from wild-type mice by beads coated with an irrelevant IgG exerted only a very low mean responsiveness (Fig. 4B and Table 1).

Fig. 4 Comparison of short-lived, initial Ca2+ signals in RyR-deficient Jurkat cells, RyR1−/− primary T cells, and WT cells.

(A and B) Initial Ca2+ signals in control (clone E2) and RyR-deficient (clone #10) Jurkat cells, primary T cells from RyR1−/− cells, and primary T cells from WT mice that were treated with IgG-coated beads (WT neg.) were characterized by the number of signals and signal amplitude as shown in Fig. 3A. Mean responsiveness was calculated as the product of the number of signals and the amplitude for the indicated (A) Jurkat cells and (B) primary T cells. Data are means ± SEM. *P < 0.05, ***P < 0.001 by the Kruskal-Wallis test.

NAADP and initial, short-lived Ca2+ signals

In previous work, we demonstrated the rapid formation of NAADP within the first 10 s of Jurkat cell activation (8). Furthermore, data from our studies and others provide evidence of a role for type 1 RyR in Ca2+ signaling evoked by NAADP (11, 12, 25, 26). Thus, we next used control and RyR-deficient Jurkat cells in experiments that combined high-resolution Ca2+ imaging with microinjection of NAADP. In cells microinjected with NAADP, no low-pass filtering or image averaging was performed because of the fast propagation of Ca2+ signals. Initial signals were determined by identification of the highest [Ca2+]i signals (highest 20% ratio values in each frame) with a diameter of ≥0.65 μm.

Microinjection of NAADP (60 nM) into control Jurkat cells caused localized Ca2+ signals within 20 to 25 ms close to the site of injection in 73% of cells (Fig. 5A). Most responding cells achieved global Ca2+ signaling after more than 200 ms after injection, whereas in very fast responding cells, global Ca2+ signals had already appeared within 200 ms. In contrast, in most RyR-deficient Jurkat cells, microinjection of NAADP evoked neither local Ca2+ signals (Fig. 5B) nor global Ca2+ waves (fig. S8). Furthermore, the percentage of RyR-deficient Jurkat cells that responded to injection with NAADP was similar to the small percentage of cells (control or RyR-deficient) that responded to injection with buffer (Fig. 5C). In contrast, local and global Ca2+ signals observed upon microinjection of IP3 (4 μM) were unaffected in RyR-deficient Jurkat cells (Fig. 5C and fig. S8). Typical local Ca2+ signals in control Jurkat cells evoked by NAADP within 20 to 25 ms of injection had a diameter of ~1.06 μm with a maximum amplitude of [Ca2+]i of ~372 nM (Fluo-4/Fura Red ratio, 1.56). The number of local Ca2+ signals per cell was between ~1.5 and 1.7. Together, these data suggest that localized Ca2+ signals evoked by NAADP within 20 to 25 ms of injection depend on RyRs.

Fig. 5 Subcellular Ca2+ signals evoked by microinjection of NAADP.

(A to C) Jurkat cells (control clone E2 and RyR-deficient clone #10) were loaded with Fluo-4 AM and Fura Red AM, and Ca2+ imaging was performed as described in Materials and Methods. During Ca2+ imaging, NAADP (pipette concentration, 60 nM) or IP3 (pipette concentration, 4 μM) was microinjected into the cells with the semiautomatic mode of the microinjection apparatus as detailed in Materials and Methods. For NAADP, data from two representative cells of 80 (44 clone E2 cells and 36 clone #10 cells) are shown. Arrows indicate the time of injection. (A) Top: Frame scan of a control Jurkat cell (clone E2). Bottom: Magnification of the area of microinjection. Black lines in frames at 20 and 100 ms represent local Ca2+ signals. (B) Frame scan of a RyR-deficient Jurkat cell (clone #10). (C) Percentages of the indicated cells that responded upon microinjection of intracellular buffer, NAADP or IP3. Data from 36 to 44 cells (for NAADP) and 15 to 18 cells (for IP3) were analyzed.

Lack of involvement of TRPM2 in short-lived, initial Ca2+ signals in primary mouse T cells

Because TRPM2 has been discussed as a putative target of NAADP, we analyzed local Ca2+ signals in primary CD4+ T cells from TRPM2−/− mice after activation with anti-CD3/CD28–coated beads and compared them to those in activated cells from wild-type mice. Within the first few seconds after contact with the beads, initial Ca2+ signals occurred in wild-type and TRPM2−/− T cells (Fig. 6, A and B). These signals were localized very close to the plasma membrane (Fig. 6, A and B) with amplitudes of 124 ± 5 nM (n = 111 signals from 20 cells) and 134 ± 4 nM (n = 142 signals from 22 cells) for wild-type and TRPM2−/− cells, respectively (Fig. 6C). The number of signals, the amplitudes, and the mean responsiveness values were not statistically significantly different between wild-type and TRPM2−/− T cells (Fig. 6C and Table 1), which suggests that TRPM2 was not involved in short-lived, initial Ca2+ signals in primary mouse T cells.

Fig. 6 Short-lived, initial Ca2+ signals in primary T cells from WT and TRPM2−/− mice.

(A to C) Primary CD4+ T cells were freshly isolated from the spleens of WT and TRPM2−/− mice by negative selection. The cells were loaded with Fluo-4 AM and Fura Red AM, and Ca2+ imaging was performed with beads coated with anti-CD3 and anti-CD28 mAbs. The bead contact site is indicated schematically. Data are from single cells and are representative of 20 WT and 22 TRPM2−/− T cells. (A and B) Top: Frame scans of initial localized Ca2+ release events in (A) a WT T cell and (B) a TRPM2−/− T cell upon contact with a bead. Bottom: Magnified regions (4 × 4 μm) near the bead contact zone were imaged from the time of bead contact until ~800 ms after contact was initiated. In this presentation, five subsequent images were merged to decrease noise (compare to fig. S2D). (C) Characteristics of short-lived, initial Ca2+ signals in primary mouse WT and TRPM2−/− T cells after localized activation by anti-CD3– and anti-CD28–coated beads. Data are means ± SEM of 20 to 22 cells. *P < 0.05, **P < 0.01, ***P < 0.001 by the Mann-Whitney U test.

DISCUSSION

Here, we described a systematic investigation by high-resolution ratiometric Ca2+ imaging that enabled the subcellular analysis of Ca2+ signals in T cells upon activation. We selected suitable Ca2+ indicators, established a system for rapid data acquisition, proposed procedures for image analysis, and used T cells from relevant knockout mice to identify and characterize short-lived, initial Ca2+ signals.

Advantages of the application of Fluo-4/Fura Red in subcellular Ca2+ imaging

Although Fluo dyes hold many advantages, such as broad dynamic range, applicability to local and global Ca2+ signals (27), low photobleaching (20), and high fluorescence intensity (fig. S1), cellular movements and shape changes substantially limit their applicability of single-wavelength dyes and F/F0 ratioing (fig. S2). However, the combination of Fluo dyes with Fura Red is suitable for the detection of subcellular signals in cardiac myocytes (28) and in T cells (29). When using a pair of dyes, such as Fluo-4 and Fura Red, they ideally should have identical or very similar Ca2+-binding kinetics and subcellular distributions (30). Furthermore, ideally, their rates of photobleaching should be low and comparable. Indeed, very similar Ca2+-binding kinetics were previously shown for Fluo-3 and Fura Red (21). Floto et al. identified problems with determining subcellular Ca2+ signals because of strong intercellular loading variations of Fluo-3 and Fura Red (30). In our cells, we did not observe any marked intercellular differences in loading; in addition, intracellular variations of dye distribution were considerably low (fig. S3). We determined the dissociation constant (Kd) of the indicators in situ with a 10-point calibration curve in our cell system (fig. S9). We describe here the Kd for the dye combination of Fluo-4 and Fura Red in situ (Kd = 408 nM), which is comparable to published values for the combination of Fluo-3 and Fura Red (Kd = 381 nM) (21).

Intracellular concentrations of Fluo-4 and Fura Red were kept as low as possible to circumvent interference with Ca2+ signaling, for example, because of buffering of short-lived or low-amplitude Ca2+ signals (31). The Ca2+-imaging approach described here is technically limited by the detection limits, for example, amplitude differences as low as 18 nM (Fluo-4/Fura Red ratio, ~0.1) at a temporal resolution of ~40 fps and a calculated spatial resolution of ~0.368 μm. A study of endothelial smooth muscle cells used frame scanning of Fluo-4–loaded cells at 30 to 60 fps to detect the sparklet activity of TRPV4 (transient receptor potential cation channel subfamily V member 4); under these conditions, it was even possible to optically record single-channel openings (32). The Ca2+-imaging system established in our study is similar to that described previously (32) but is especially suited to cope with artifacts that occur because of cellular shape changes, for example, during activation at an immunological synapse. Up to now, short-lived, low-amplitude Ca2+ signals in T cells in response to directed, localized activation have not been investigated in detail; indeed, previous studies focused on T cell Ca2+ signals in the range of seconds or tens of seconds after activation (3337). Thus, we established a method that enables the detailed analysis of initial short-lived, localized Ca2+ signals in T cells.

Elementary Ca2+ signaling events in T cells and their underlying mechanism(s)

How do the Ca2+ signal data determined here compare to previously described elementary Ca2+ release events? Local Ca2+ signals obtained by the microinjection of NAADP or IP3 into Jurkat cells showed diameters of 1.07 to 1.15 μm, which are more consistent with sparks (6) in cardiomyocytes than with puffs (5). Furthermore, the amplitudes of Ca2+ signals upon microinjection of cells with NAADP or IP3 were ~374 and 151 nM, respectively (Fluo-4/Fura Red ratios of 1.6 and 1.1, respectively). For IP3, this value falls in the range of 50 to 600 nM that was described for puffs (5), but it seems to be too high for sparks. Nevertheless, knockdown of RyRs abrogated rapid local Ca2+ signals 25 ms after injection with NAADP, which suggests that even these very fast signals depend on functional RyRs. Thus, it seems to be difficult to compare values obtained for elementary Ca2+ signaling events in certain cell types to those in other cell types. Cell type–specific architectures of Ca2+ release units composed of Ca2+ channels and perhaps neighboring Ca2+ pumps and Ca2+-binding proteins may largely influence parameters such as signal diameter and amplitude.

As mentioned earlier, initial, short-lived, localized Ca2+ release events observed upon microinjection of cells with NAADP depended on RyRs. This is noteworthy because the “two-pool model” of NAADP function suggests that NAADP initially acts on two-pore channels on acidic stores to stimulate the release of Ca2+ that, in turn, activates RyRs (or IP3Rs), which results in amplification of the primary (local) Ca2+ signal (38). In contrast, our kinetic data indicate that RyRs are the initially targeted channels involved in very fast NAADP-evoked Ca2+ release events. During the first 15 s of activation of Jurkat cells by anti-CD3–coated beads, we detected Ca2+ signals with amplitudes of ~79 nM (Fluo-4/Fura Red ratio, 0.79) and diameters of ~0.43 μm. Thus, these signals are considerably smaller in amplitude and diameter than are sparks, which suggests that these short-lived, initial Ca2+ signals more closely resemble quarks (7). Quarks typically have an amplitude of 37 ± 6 nM Ca2+ and are 0.85 ± 0.2 μm in diameter, although it was speculated that the diameter of these signals might more likely be ~0.4 μm (7). In primary T cells, the diameters of Ca2+ signals were even smaller than those in Jurkat cells. However, the signal amplitude in the primary T cells (193 ± 8 nM) was about twofold greater than that in Jurkat cells, which is suggestive of either activation of larger clusters of RyRs or additional amplification mechanisms.

Knockout of RyR1 in primary mouse T cells or knockdown of RyR in Jurkat cells had very comparable effects. In ~45% of such cells, short-lived, initial Ca2+ signals were not observed. Furthermore, in responding cells, the number of these signals was reduced about twofold compared to that in wild-type or control cells. However, about half of the cells in which RyRs were knocked down or in which RyR1 was knocked out responded with short-lived, initial Ca2+ signals that were similar in spatial spread to those in control cells but somewhat diminished in amplitude. Regarding the combined parameter termed mean responsiveness, a marked and statistically significant decrease was observed between control cells and RyR knockdown and RyR1−/− cells. These results suggest that short-lived, initial Ca2+ signals initiated by beads coated with anti-CD3 (and anti-CD28) antibodies do not strictly depend on RyRs but that RyRs (RyR1 in primary T cells) substantially contributed to the initiation of such signals. Knockout of another target channel candidate for NAADP, TRPM2, did not affect short-lived, initial Ca2+ signals. Together, the rapid formation of NAADP within a few seconds upon anti-CD3 stimulation (8) and the strong dependence of NAADP-evoked Ca2+ signals in T cells on RyRs, but not TRPM2, suggest that short-lived, initial Ca2+ signals are (co)activated by NAADP acting on RyRs.

What other systems might be involved in the generation of short-lived, initial Ca2+ signals? A role for IP3 is not very likely, because its formation proceeds within minutes rather than seconds or milliseconds (39). Neither the number of Ca2+ signals nor the number of responding cells upon stimulation by anti-CD3–coated beads was reduced by the inhibition of IP3R with 10 μM xestospongin C (Table 1); there were slight increases compared to those in control cells. These data suggest that the opening of IP3Rs is not a major event in this early phase of Ca2+ signaling. However, inhibition of IP3Rs resulted in a decrease in the amplitude of Ca2+ signals, which may be explained by IP3Rs acting as enhancing elements. Finally, evidence for a component of Ca2+ entry was obtained, because all the parameters of short-lived, initial Ca2+ signals were decreased in control Jurkat cells that were activated in the absence of extracellular Ca2+. Thus, overall, additional Ca2+ channels appear to play pivotal roles in short-lived, initial Ca2+ signals during T cell activation too. In summary, short-lived, initial Ca2+ signals were characterized in a T cell line and in primary T cells. By knocking out RyR1 and knocking down RyRs, we demonstrated a role for RyRs in the generation of such Ca2+ signals; however, additional mechanisms, such as Ca2+ entry, appear to be involved and need to be investigated in the future.

MATERIALS AND METHODS

Materials

AMs of Indo-1, Fluo-3, Fluo-4, and Fura Red were obtained from Life Technologies. ACaR and Fluo-8 AM were obtained from TEFLabs. All dyes were dissolved in DMSO, divided into aliquots, and stored at −20°C until required for use. Anti-human CD3 mAb (clone OKT3) was generated as described previously (40). NAADP was obtained from Biolog GmbH.

Cell culture and isolation of primary T cells

Jurkat cells (subclone JMP) were cultured in RPMI 1640 containing 25 mM Hepes and GlutaMAX-I (Gibco, Life Technologies), supplemented with 7.5% newborn calf serum (Biochrom, Merck Millipore), penicillin (100 U/ml), and streptomycin (100 μg/ml). The Tet-On Jurkat cell clones pTRE2-EGFP/E2 (abbreviated as clone E2) and pTRE2-240/10 (RyR knockdown cells, abbreviated as clone #10) were generated and cultured as described previously (24). Primary T cells were obtained from the lymph nodes and spleens of naïve C57Bl/6 mice. CD3+ T cells were isolated by negative selection with the EasySep Mouse T Cell Enrichment Kit (STEMCELL Technologies Inc.) according to the manufacturer’s instruction. CD4+ T cells were isolated by negative selection with biotinylated anti-CD44 antibody (BioLegend) and streptavidin microbeads (Miltenyi Biotec GmbH) as well as an EasySep Mouse T Cell Isolation Kit (STEMCELL Technologies Inc.) according to the manufacturer’s instructions. Cell purity (which was typically >95% T cells) was assessed by immunostaining with fluorescein isothiocyanate–conjugated anti-mouse TCRβ antibody (clone H57-597, BioLegend) and measured with a FACSCalibur flow cytometer (BD Biosciences). Cells were kept in Dulbecco’s modified Eagle’s medium containing 25 mM Hepes for up to 6 hours until they were required for use. Mice with inactivation of one copy of the Ryr1 gene were generated on a mixed 129Sv × C57Bl/6 genetic background by Takeshima et al. (41). This mouse strain was backcrossed to the C57Bl/6N background for 10 generations and further crossed to Rosa26-tdRFP fluorescent reporter strain (42). To overcome the postnatal lethality of RyR1−/− animals, fetal liver chimeras were produced. RyR1+/− animals were mated, and the embryos at embryonic days 13 to 14.5 were used to isolate fetal livers containing hematopoietic progenitor cells. Homogenized livers were cultured in StemSpan serum-free medium for the expansion and culture of hematopoietic cells (STEMCELL Technologies) supplemented with mouse stem cell factor (50 ng/ml), mouse interleukin-3 (IL-3; 20 ng/ml), and human IL-6 (50 ng/ml) for several hours to enable polymerase chain reaction–based genotyping to be performed before cell transfer. All cytokines were obtained from PeproTech. Rag1-deficient recipient mice (43) were irradiated with a 2.5-Gy dose and received 5 × 106 fetal liver cells by injection into the tail vein. Three to 4 weeks later, the peripheral blood of the fetal liver chimeric mice was assessed by immunostaining with antibodies specific for TCRβ and B220, followed by flow cytometric analysis to test for reconstitution efficiency. Chimeric mice with sufficient numbers of RFP+ T cells in their blood were sacrificed 6 to 8 weeks after transfer, and their T cells were isolated from spleens and lymph nodes and maintained on ice until dye loading was performed. TRPM2−/− mice were provided by Y. Mori (44). Animal experiments were performed in accordance with local regulations for animal welfare in the German federal state of Lower Saxony and the City of Hamburg.

Single-cell Ca2+ imaging

Primary T cells or Jurkat cells were incubated with the membrane-permeable AM esters of the Ca2+ dyes Fluo-3, Fluo-4, Fluo-8, Indo-1, ACaR (in 0.04% Pluronic F-127, Sigma-Aldrich) (all at 10 μM), or 20 μM Fura Red. Therefore, about 2.5 × 106 cells were centrifuged at 500g for 5 min and resuspended in 500 μl of RPMI medium supplemented with fresh serum and containing the appropriate dye. The cells were incubated for 50 to 55 min, during which 2 ml of fresh medium was added after 20 min of incubation. After centrifugation, the cells were washed and resuspended in Ca2+ buffer [140 mM NaCl, 5 mM KCl, 1 mM MgSO4, 1 mM CaCl2, 20 mM Hepes (pH 7.4), 1 mM NaH2PO4, 5 mM glucose] and then kept for 20 min for de-esterification. All steps were performed at room temperature, and the cells were kept in the dark after the dyes were added. For chemical inhibition of RyRs or IP3Rs, Jurkat cells were preincubated with 50 μM ryanodine (Sigma-Aldrich), 10 μM xestospongin C (Merck Millipore), or 0.5% DMSO as a vehicle control for 20 min before being analyzed. In experiments performed in the absence of extracellular Ca2+, cells were centrifuged and resuspended in Ca2+-free buffer [140 mM NaCl, 5 mM KCl, 1 mM MgSO4, 20 mM Hepes (pH 7.4), 1 mM NaH2PO4, 5 mM glucose] immediately before measurements were performed. Coverslips were coated with bovine serum albumin (5 mg/ml, Sigma-Aldrich) and poly-l-lysine (0.1 μg/ml, Sigma-Aldrich). Chamber slides were prepared by mounting rubber O-rings onto the slides with silicon grease (23). T cells were added and allowed to adhere before measurement. Slides were mounted onto a Leica IRBE2 microscope (100-fold magnification). A Sutter DG-4 was used as a light source, and frames were acquired with an electron-multiplying charge-coupled device camera (C9100, Hamamatsu). Images (512 × 512 pixels) were acquired in 14-bit mode with twofold binning. Measurements of emission-shift wavelength dyes such as Indo-1 were made with a Dual-View module (Optical Insights, PerkinElmer Inc.). Exposure time was 25 ms (40 fps); however, in microinjection experiments, exposure time was reduced to 20 ms (50 fps). The following filter sets (AHF analysentechnik) were used for the measurement of the respective indicators [excitation (ex), beam splitter (bs), and emission (em), all in nanometers]: Fluo-3, Fluo-4, Fluo-8 (ex, bp480/40; bs, Q505LP; em, bp535/50); Indo-1 (ex, 360/12; bs, 380 DCLP; em1, 405/30; bs, 440; em2, 485/25); ACaR (ex, 494/20; bs, 495; em1, 542/50; em2, 650/57); and Fluo-4/Fura Red (ex, 480/40; bs, 495; em1, 542/50; em2, 650/57).

Determination of the Kd of Fluo-4 and Fura Red and calculation of [Ca2+]i

The Kd of Fluo-4 and Fura Red was determined with a Ca2+ Calibration Buffer kit (Life Technologies) according to the manufacturer’s instructions. In brief, Jurkat cells were loaded with the two Ca2+ indicators as described earlier. Cells (1 × 106 cells/ml) were lysed in different Ca2+-EGTA buffers with 10 different Ca2+ concentrations (ranging from 0.017 to 1.35 μM) and Triton X-100 (0.1%). Fluorescence was detected with a Tecan microplate reader. A mean Kd of 408 ± 12 nM was determined in 15 experiments (fig. S9). Calculation of [Ca2+]i was performed as described by Grynkiewicz et al. (45) using in situ–determined Rmin [using the lowest ratio (R) and fluorescence (F) after EGTA chelation] and Rmax.(maximum R and F) in single-cell measurements. For calibration, data were subjected to image post-processing. Because bleaching correction of Fura Red was not feasible in single-frame acquisitions, it was assumed that the initial FFura Red was about as high as the FFura Red after correction for bleaching (see also fig. S1D).

Measurement of localized Ca2+ signals: Anti-CD3–coated beads and microinjections

To initiate subcellular Ca2+ signals, either incubation with beads coated with anti-CD3 or anti-CD3/anti-CD28 antibodies (for primary cells; Becton, Dickinson and Company) or microinjection with NAADP was performed. Protein G Beads (Merck Millipore) were coated with antibodies according to the manufacturer’s instructions. To stimulate cells, soluble antibody or beads were added after 1 min of slow acquisition (1 frame/5 s). Microinjection was performed as described previously (10) with an Eppendorf FemtoJet and a Micromanipulator type 5171 with Femtotips II as microinjection pipettes (Eppendorf GmbH). NAADP and IP3 were dissolved in water and diluted to their final injection concentrations of 60 nM and 4 μM, respectively, in intracellular buffer [20 mM Hepes (pH 7.2), 110 mM KCl, 2 mM MgCl2, 5 mM KH2PO4, 10 mM NaCl]. This solution was subjected to Chelex resin to remove Ca2+ ions directly before the microinjection pipette was filled. Microinjection was performed using the semiautomatic mode with the following settings: compensatory pressure, 25 hPa; injection pressure, 40 hPa; pipette speed, 700 μm/s; injection time, 0.3 s.

Image processing

Volocity software (version 6.6.2; PerkinElmer Inc.) was used for the acquisition and analysis of mean global Ca2+ signals. Frames taken with a Dual-View module were split into the two wavelength emission channels with Fiji 1.47v (22). Furthermore, Fiji was used for background correction of Fura Red and Fluo-4, as well as for the bleaching correction of Fura Red. For this purpose, we used an ImageJ plug-in for bleaching correction assuming that the decay in fluorescence intensity followed a biexponential function (46). The plug-in is a modification of the original Bleach Corrector plug-in in Fiji by K. Miura and J. Rietdorf (http://cmci.embl.de/downloads/bleach_corrector) written in Jython (www.jython.org/). It requires a series of frames with a constant time interval. To correct for background fluorescence for each frame, the mean intensity from an area devoid of fluorescent cells was subtracted from the intensity value of each pixel of that frame. To determine the parameters of the biexponential decay function, the mean fluorescence intensity for an ROI in a fluorescent cell was normalized to the mean intensity of the same ROI of the first frame, and the CurveFitter function of Fiji was then used to fit a biexponential function to the normalized fluorescence intensities. Data from four cells were used to obtain parameters for the biexponential function. With these parameters, correction factors were calculated for each frame. Afterward, each frame was corrected by dividing each pixel by the correction factor (fig. S1D). Openlab software (version 5.5.2, PerkinElmer Inc.) was used for post-measurement registration, image deconvolution, and ratioing of the two emission wavelengths. Image deconvolution was performed as described previously (4) to obtain digital confocal images. Applying the no-neighbor deconvolution, stray light was reduced by 45%, and the gain was set to 30. The ratios were median-filtered (3 × 3 pixels), and a variable threshold against background noise was applied. Analysis of short-lived Ca2+ signals was performed with the maximum intensity tool of Fiji (as described in Results). Therefore, the time point of bead contact was estimated as follows: magnetic beads are characterized by a weak autofluorescence and could therefore be identified during analysis. Bead autofluorescence had no effect on the analysis of Ca2+ signals.

Noise reduction

Given the large amount of data acquired during high-resolution Ca2+ imaging, two strategies were pursued for noise reduction: For rapid noise-reduced visualization of the image data and initial detection of single ROIs exhibiting large Ca2+ signals, a variant of the computationally cheap moving average filter was implemented and applied to the temporal image series (averaging intensity values for each pixel over five frames; implemented in MATLAB version 7.14, www.mathworks.de). For detailed analysis, the time-discrete data from tracings of single ROIs were subjected to Fourier transformation with the fast Fourier transformation function of the NumPy/SciPy stack (version 0.12.0, www.scipy.org) for Python (version 2.7.5, www.python.org). The spectra indicated that only the frequency range below 2 Hz deviated from white noise with similar amplitude. Consequently, the data from the tracings were low-pass–filtered (fifth-order Butterworth filter) with a cutoff frequency of 2 Hz with a Python script importing and exporting data from and to Microsoft Excel files. The filter was applied once forward once backward (using the filtfilt function of SciPy) to avoid phase shifting.

Noise and detection limits in Fluo-4/Fura Red ratioing experiments

With Fluo-4 and Fura Red in T cells, an acquisition velocity of up to 48 fps was feasible. Spatial resolution was both diffraction- and sampling bandwidth–limited and was calculated to be 368 nm at 100-fold magnification. To analyze the background Ca2+ signal noise under such conditions, homogeneous buffers containing Fluo-4 and Fura Red were measured. Therefore, Fluo-4 AM (10 μM) and Fura Red AM (20 μM) were de-esterified by esterase (1200 U/ml treated with Chelex 100 to remove possible Ca2+ contaminations; both from Sigma-Aldrich). Tracings in fig. S2D represent Ca2+ signals in ROIs of similar size to those in living cells. Obtained from a homogeneous cell-free solution at 100 nM [Ca2+]free, the tracings resulted in a noise amplitude of 21 nM Ca2+. Two strategies were pursued for noise reduction (see Noise reduction for details): For rapid noise-reduced visualization of the image data and areas exhibiting Ca2+ signals, a variant of a moving average filter was applied to the temporal image series. This reduced noise in ROIs to ~9 nM Ca2+ (fig. S2D). For detailed analysis, tracings of single ROIs were low-pass–filtered with a cutoff frequency of 2 Hz. Noise was reduced to ~6 nM Ca2+ (fig. S2D). Subsequent identification of subcellular Ca2+ signals was performed by applying a detection limit of threefold noise amplitude (a Fluo-4/Fura Red ratio of ~0.1).

Mean responsiveness

A combined parameter to describe the responsiveness of individual cells was defined: responsiveness = number of signals × amplitude. This combined parameter was used for the calculation of the mean responsiveness (±SEM) over all cells from a specific group.

Statistical analysis

Data are single-cell measurements from three separate sets of experiments and are shown as means ± SEM. Statistical analysis was performed with GraphPad Prism 6 (GraphPad Software Inc.) by applying Mann-Whitney U tests and Kruskal-Wallis tests with Dunn post hoc tests, respectively, to the nonnormally distributed data. The normality of distribution was tested by Shapiro-Wilk.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/8/398/ra102/DC1

Fig. S1. Properties of Ca2+ indicators.

Fig. S2. Comparison of F/F0 Fluo-4 ratioing to Fluo-4/Fura Red ratioing during localized T cell activation.

Fig. S3. Comparison of the Fluo-4/Fura Red ratio in Jurkat cells and primary mouse T cells as compared to that in a cell-free solution.

Fig. S4. Comparison of the intracellular distributions of the Ca2+ indicators Fluo-4 and Fura Red with that of Fura-2 in Jurkat cells.

Fig. S5. Effect of knockdown of RyRs on global Ca2+ signaling in Jurkat cells.

Fig. S6. Kinetics and amplitude of localized Ca2+ signals in individual primary mouse T cells.

Fig. S7. Short-lived, initial Ca2+ signals in a responding primary T cell from RyR1−/− mice.

Fig. S8. Effect of knockdown of RyRs on Ca2+ signaling evoked by microinjection of NAADP into Jurkat cells.

Fig. S9. Calculation of the Kd for Fluo-4/Fura Red in situ.

REFERENCES AND NOTES

Acknowledgments: We are grateful to H. Takeshima (Kyoto University, Japan) for the RyR1+/− mice and to Y. Mori (Showa University School of Pharmacy, Tokyo, Japan) for the TRPM2−/− mice. Funding: This study was supported by the Deutsche Forschungsgemeinschaft (grants GU 360/15-1 and 360/16-1 to A.H.G., and grants FL 377/2-1 and TRR-SFB43-TP B11 to A.F.), the Forschungszentrum Medizintechnik Hamburg (to I.M.A.W. and R.W.), Förderfonds Medizin of the University Medical Center Hamburg-Eppendorf (grant NWF 15/13 to I.M.A.W.), the Hertie Foundation (grant P1130072 to A.F. and D.L.), and the Landesforschungsförderung of the City of Hamburg (Research Group ReAd Me to A.H.G., I.M.A.W., and H.-W.M.). Author contributions: I.M.A.W., B.-P.D., E.G., and F.C. designed and performed the measurements and subsequent analyses of calcium signals; D.S. and R.W. designed and implemented the image noise-reduction and stacking tool; J.K. isolated the TRPM2−/− cells; H.-W.M. and V.S. provided mice and isolated TRPM2−/− T cells; M.v.O., D.L., and A.F. generated and isolated RyR1−/− murine T cells; R.F. wrote the bleaching correction tool and the noise filter; A.H.G. designed the study and the experiments and evaluated measurements; and all authors wrote the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: Use of the TRPM2−/− mice requires a material transfer agreement from Showa University School of Pharmacy, Tokyo, Japan.
View Abstract

Navigate This Article