Functional Ryanodine Receptor Expression Is Required for NAADP-mediated Local Ca2+ Signaling in T-lymphocytes*

Nicotinic acid adenine dinucleotide phosphate (NAADP) is a potent Ca2+-mobilizing nucleotide involved in T cell Ca2+ signaling (Berg, I., Potter, B. V. L., Mayr, G. W., and Guse, A. H. (2000) J. Cell Biol. 150, 581–588). The objective of this study was to analyze whether the first subcellular Ca2+ signals obtained upon NAADP stimulation of T-lymphocytes depend on the functional expression of ryanodine receptors. Using combined microinjection and high resolution confocal calcium imaging, we demonstrate here that subcellular Ca2+ signals, characterized by amplitudes between ∼30 and 100 nm and diameters of ∼0.5 μm, preceded global Ca2+ signals. Co-injection of the ryanodine receptor antagonists ruthenium red and ryanodine together with NAADP abolished the effects of NAADP, whereas the d-myo-inositol 1,4,5-trisphosphate antagonist heparin and the Ca2+ entry blocker SKF&96365 were without effect. This pharmacological approach was confirmed by a molecular knock-down approach. Jurkat T cell clones with largely reduced expression of ryanodine receptors did not respond to microinjections of NAADP. Taken together, our data suggest that the Ca2+ release channel sensitive to NAADP in T-lymphocytes is the ryanodine receptor.

Nicotinic acid adenine dinucleotide phosphate (NAADP) 1 is an endogenous nucleotide in eukaryotic cells and to date represents the most powerful Ca 2ϩ -releasing compound. This compound was discovered by Lee and co-workers as an impurity of NADP in 1987 (1), and its structure was obtained in 1995 (2).
In contrast to the other known Ca 2ϩ -releasing compounds, D-myo-inositol 1,4,5-trisphosphate (InsP 3 ) (for review see Ref. 3) and cyclic ADP-ribose (cADPR) (for review see Ref. 4), the Ca 2ϩ channel sensitive to NAADP, is still a matter of debate. Pharmacological Ca 2ϩ release data obtained in sea urchin egg homogenates suggest that neither the InsP 3 receptor nor the ryanodine receptor (RyR) represents the molecular target for NAADP (2). In addition, NAADP appears not to act on the classical rapidly exchanging Ca 2ϩ store, the endoplasmic reticulum. Indeed, a NAADP-sensitive Ca 2ϩ pool was separated from the cADPR-and InsP 3 -sensitive one by stratification of sea urchin eggs (5). Subsequently, this store was identified as the reserve granule of eggs, a lysosome-related organelle (6). Recently, a lysosome-related acidic Ca 2ϩ store sensitive to NAADP was also detected in higher eukaryotic cells, e.g. pancreatic acinar cells and the MIN6 pancreatic ␤-cell line (7), and in rat cortical neurons (8).
Although the pharmacological characterization of NAADPinduced Ca 2ϩ signaling in sea urchin eggs resulted in the conclusion that a novel Ca 2ϩ channel unrelated to the known intracellular Ca 2ϩ release channels is involved, a number of reports from heart and skeletal muscle and pancreatic acinar cells suggest that RyR are the Ca 2ϩ channels mediating the effect of NAADP (9 -11).
Accordingly, different models resulting from these conflicting data have been postulated. The two-pool model (12) consists of two separate Ca 2ϩ pools: a lysosome-related Ca 2ϩ pool with the novel NAADP receptor giving rise to spatiotemporally restricted trigger Ca 2ϩ and an endoplasmic reticulum-related Ca 2ϩ pool with InsP 3 receptors and RyR, which then respond to the trigger Ca 2ϩ by Ca 2ϩ -induced Ca 2ϩ release (CICR). The other model reduced the number of Ca 2ϩ channels sensitive to NAADP to the RyR but postulated different binding proteins for NAADP and cADPR, mediating their effects at the RyR (11). We have recently shown that global Ca 2ϩ signaling induced by NAADP in human Jurkat T-lymphocytes depends on functional expression of RyR and also on Ca 2ϩ entry (13).
Because that report was compatible with both models, the present study was conducted to analyze the very initial subcellular Ca 2ϩ release events observed upon NAADP stimulation and, in particular, to understand whether the RyR is necessary also for these spatiotemporally restricted Ca 2ϩ signals.
Ratiometric Ca 2ϩ Imaging-The cells were loaded with Fura-2/AM as described by Kunerth et al. (16) and kept in the dark at room temperature until use. Thin glass coverslips (0.1 mm) were coated with bovine serum albumin (5 mg/ml) and poly-L-lysine (0.1 mg/ml). Silicon grease was used to seal small chambers consisting of a rubber O-ring on the glass coverslips. 60 l of buffer A containing 140 mM NaCl, 5 mM KCl, 1 mM MgSO 4 , 1 mM CaCl 2 , 1 mM NaH 2 PO 4 , 5.5 mM glucose, and 20 mM HEPES, pH 7.4, and a 40-l cell suspension (2 ϫ 10 6 cells/ml) suspended in the same buffer were added into the small chamber (16).
The coverslip with cells slightly attached to the bovine serum albumin/ poly-L-lysine coating was mounted on the stage of a fluorescence microscope (Leica DM IRE2).
Ratiometric Ca 2ϩ imaging was performed as described recently (16). We used an Improvision imaging system (Tü bingen, Germany) built around the Leica microscope at 100-fold magnification. Illumination at 340 and 380 nm was carried out using a monochromator system (Polychromator IV, TILL Photonics, Grä felfing, Germany). Images were taken with a gray-scale CCD camera (type C4742-95-12ER; Hamamatsu, Enfield, United Kingdom) operated in 8-bit mode. The spatial resolution was 512 ϫ 640 pixel at 100-fold magnification. Camera exposure times were 12 (at 340 nm) and 4 ms (at 380 nm). The acquisition rate was ϳ1 ratio m 160 ms. Raw data images were stored on a hard disk. Confocal Ca 2ϩ images were obtained by off-line noneighbor deconvolution using the volume deconvolution module of the Openlab software as described recently for 3T3 fibroblasts (17). The deconvolved images were used to construct ratio images (340/380). Finally, ratio values were converted to Ca 2ϩ concentrations by external calibration. To reduce noise, ratio images were subjected to median filter (3 ϫ 3) as described previously (16). Data processing was performed using Openlab software, version 1.7.8, 3.0.9, or 3.5.2 (Improvision, Tü bingen, Germany).

RESULTS
Subcellular Ca 2ϩ signaling upon microinjection of 100 nM NAADP into Jurkat T-lymphocytes (clone JMP) remained almost unchanged for a very short, initial period of a few hundred milliseconds (  1C). These control data, together with the fact that in the very initial period after microinjection almost no changes in Ca 2ϩ signaling were observed, suggest that (i) the subcellular signals were caused by NAADP and that (ii) the temporal resolution of image acquisition was sufficient to analyze the very initial NAADP-induced Ca 2ϩ signals. A magnified view of the images showing NAADP-induced subcellular Ca 2ϩ signals revealed that, in case of the 100 nM injection, the signals were more concentrated in one part of the cell, almost merging into one much bigger Ca 2ϩ signal ( Fig. 2A,  arrow; diameter of the merged signal ϳ2 m). However, individual small Ca 2ϩ signals with diameters characteristic for Ca 2ϩ quarks (diameter ϳ0.5 m) were still observed in close vicinity to the bigger signal ( Fig. 2A, arrowheads). Upon 30 nM NAADP microinjections, mainly the smaller Ca 2ϩ signals were observed (Fig. 2B, arrowheads).
Very recently, we have shown that global Ca 2ϩ signals evoked by microinjection of NAADP in T cells were completely dependent on functional expression of RyR (13). However, these data left open the possibility that NAADP may initially activate a Ca 2ϩ channel different from the RyR, giving rise to local Ca 2ϩ signals, which then by the process of CICR would trigger RyR, according to the 2-pool model proposed originally by Cancela et al. (12). To distinguish between the 2-pool model and a model where RyR directly (or via an additional binding protein) responds to NAADP (11), subcellular NAADP-induced Ca 2ϩ signals as described in Figs. 1 and 2 were analyzed under further experimental conditions. Ca 2ϩ signaling was analyzed in T cells co-injected with the following: (i) NAADP and the RyR antagonists ruthenium red and ryanodine; (ii) NAADP and the InsP 3 antagonist heparin; or (iii) NAADP and the Ca 2ϩ entry blocker SKF 96365.
As reported previously (13), the global Ca 2ϩ signal observed upon NAADP injection was completely abolished when ruthenium red was co-injected (Fig. 3A). An analysis of small subcellular regions of interest (ROI) set either close to the plasma membrane (pm), the cytosol (cyt), or the nucleus (nuc) revealed an oscillatory behavior of the local [Ca 2ϩ ] (Fig. 3B). Importantly, the amplitude of local Ca 2ϩ signals increased with time in the presence of NAADP (Fig. 3B, black tracings). In contrast, when NAADP was co-injected with ruthenium red, the subcellular signals remained at similar amplitudes compared with buffer injections (Fig. 3B, red and blue tracings). Quantitative evaluation of an array of microinjection and preincubation data revealed the following. Co-injection of the antagonists of RyR, ruthenium red and ryanodine (the latter at a high inhibitory concentration), completely blocked subcellular Ca 2ϩ signals evoked by NAADP, regardless of the subcellular localization of the signals (Fig. 4). In contrast, neither co-injection of the InsP 3 antagonist heparin nor preincubation with the Ca 2ϩ entry blocker SKF 96365 antagonized subcellular signals evoked by NAADP (Fig. 4). Rather, a weak stimulatory effect of both heparin microinjection and preincubation with SKF 96365 in the absence and presence of NAADP (co-)injection was ob- served (Fig. 4). The weak agonistic effect of heparin at the RyR and the weak Ca 2ϩ -mobilizing properties of SKF 96365 are well documented (21,22) and thus explain the observed effects. Autoinactivation of the NAADP signaling system by injection of a high concentration of NAADP resulted in signal amplitudes comparable to buffer injections. In addition, subcellular Ca 2ϩ signals evoked by the established RyR agonist cADPR were fully blocked by this procedure (Fig. 4).
Analysis of the spatiotemporal pattern of Ca 2ϩ signaling evoked by NAADP compared with the pattern induced by the established RyR agonist cADPR did not result in significant differences (Fig. 5) as can be seen from the images (Fig. 5A) and the ROI tracings close to the pm, in the cyt, or in the nuc (Fig. 5B).
Although the pharmacological data and the spatiotemporal characteristics of NAADP-mediated subcellular Ca 2ϩ signals suggest involvement of RyR, the use of inhibitors may be misleading due to insufficient specificity. Therefore, in a second independent approach, microinjection and imaging experiments were conducted in RyR knock-down T cell clones (13,15,18). These RyR knock-down Jurkat T cell clones were obtained by stable transfection with plasmids expressing antisense RNA against RyR (15). Clone 25 expressed antisense RNA specific for the type 3 RyR, whereas clone 10 expressed an antisense RNA fragment targeting all three types of RyR. As control, a Jurkat T cell clone with stable expression of an antisense fragment directed toward a protein not expressed in T cells was chosen (clone E2; fragment directed against enhanced green fluorescent protein). Largely reduced expression of RyR on the protein level was analyzed by Western blot experiments and is documented by Langhorst et al. (13) and Schwarzmann et al. (15).
Microinjection of NAADP into T cells of the control clone E2 resulted in comparable subcellular signals as described above for the Jurkat T cell clone JMP (Fig. 6A). In stark contrast, no changes in subcellular Ca 2ϩ signaling were observed in the RyR knock-down T cells, neither in the pan knock-down clone 10 (Fig. 6B) nor in the type 3-specific knock-down clone 25 (Fig.  6C). Quantitative evaluation of individual ROI set close beneath the pm, in the cyt, or in the nuc resulted in an even clearer picture (Fig. 7) compared with the experiments described in Figs. 3 and 4. Rapidly after microinjection of NAADP, the amplitude of subcellular Ca 2ϩ signals in all three ROIs increased in control clone E2 but remained almost unchanged in knock-down clones 10 and 25 (Fig. 7B). Quantitative evaluation of such experiments showed that NAADP induced a statistically significant increase of the signal amplitude in all three ROI in control clone E2, whereas in RyR knock-down clones 10 and 25, no significant differences to buffer injections were observed. DISCUSSION Here we report that the initial subcellular Ca 2ϩ signals induced by microinjection of NAADP depend on the functional expression of RyR in human T cells. Cancela et al. (23) were the first to show that Ca 2ϩ spiking evoked by NAADP could be efficiently blocked by an inhibitory concentration of ryanodine (23). Obviously influenced by the fact that inhibition of RyR did not affect NAADP-induced Ca 2ϩ release in the sea urchin egg system (2), Cancela et al. (23) interpreted their data not as a direct effect but rather as inhibition of CICR via RyR secondary to the "trigger Ca 2ϩ " provided by a separate and novel receptor/ Ca 2ϩ release channel sensitive to NAADP.
In a recent report (13), we demonstrated that both RyR inhibition and the down-regulation of its expression almost completely abolished NAADP-mediated global Ca 2ϩ signaling in T cells. These data, although pronouncing the role of RyR in NAADP-induced Ca 2ϩ signaling in T cells, were still compatible with a separate and novel receptor/Ca 2ϩ release channel sensitive to NAADP. Accordingly, the task of this putative channel would then be to initiate CICR via RyR by evoking local pacemaker Ca 2ϩ signals. The alternative model is simpler and has been proposed for heart and skeletal muscle and for pancreatic acinar cells (9 -11). NAADP either directly or via a specific binding protein binds to RyR to induce opening of its Ca 2ϩ channel. In the experiments described in the current report, inhibition by the antagonists of RyR, ryanodine (at high concentration) or ruthenium red, abolished any NAADP-mediated subcellular Ca 2ϩ signals above intrinsic background. In contrast, inhibition of InsP 3 receptors by heparin or inhibition of Ca 2ϩ entry by SKF 96365 did not block but rather enhanced the subcellular signals. Such agonistic properties of heparin on RyR are well documented (21). Likewise, already in the original publication the inventors of SKF 96365, Merrit et al. (22) stated that "in some conditions in either intact or permeabilized cells, SKF 96365 appeared to cause some discharge of intracellular Ca 2ϩ stores" (22). Thus, the slight increase in the amplitudes of subcellular Ca 2ϩ signals in heparin-injected or SKF 96365preincubated cells is not unexpected. In the context of the experiments described here, these positive controls confirm the specificity of the antagonistic effects seen with high ryanodine concentrations and ruthenium red.
Most importantly, comparable inhibitory effects on NAADP-mediated subcellular Ca 2ϩ signals were also obtained in RyR knock-down T cell clones. Thus, the relevance of RyR for these subcellular Ca 2ϩ signals evoked by NAADP was confirmed by a completely independent molecular approach. The NAADP-mediated subcellular Ca 2ϩ signals analyzed were very small in diameter (starting from ϳ0.5 m) and also small in amplitude (average between 50 and 100 nM), indicating that they may represent an opening of small clusters of RyR, as proposed recently (24). Microinjection of the established RyR agonist cADPR produced subcellular Ca 2ϩ signals with similar spatiotemporal characteristics compared with NAADP, further indicating that both adenine nucleotides act on the same target Ca 2ϩ channel.
Taken together, our data suggest that RyR are, in addition to their central role in the global Ca 2ϩ signals, also responsible for the very initial subcellular Ca 2ϩ signals evoked by NAADP in human T-lymphocytes. Our data do not rule out that NAADPsensitive RyR may be expressed in acidic lysosomal Ca 2ϩ stores in T-lymphocytes or that a specific binding protein for NAADP mediating its effect at the RyR exists. However, evidence for a novel NAADP-sensitive Ca 2ϩ channel in T cells was not obtained using state-of-the-art Ca 2ϩ -imaging technology. ] in response to injection of NAADP (30 nM) into control clone E2 or RyR knock-down clone 10 or 25 is shown. In B, local responses were recorded in the cytosol close to the pm, in the cyt, or in the nuc and expressed as the difference between the maximal peak amplitude and the amplitude of the surrounding area. Arrows indicate time point of injection. C, peak amplitudes as in B expressed as the mean Ϯ S.E. (n ϭ 8 -11 cells for each condition); p values obtained from the Student's t test as indicated. In the condition "NAADPϪ Ϫ," cells were microinjected with intracellular buffer.