Ca2+ entry induced by cyclic ADP-ribose in intact T-lymphocytes.

Cyclic ADP-ribose (cADPr) is a potent Ca2+-mobilizing natural compound (Lee, H. C., Walseth, T. F., Bratt, G. T., Hayes, R. N., and Clapper, D. L. (1989) J. Biol. Chem. 264, 1608-1615) which has been shown to release Ca2+ from an intracellular store of permeabilized T-lymphocytes (Guse, A. H., Silva, C. P., Emmrich, F., Ashamu, G., Potter, B. V. L., and Mayr, G. W. (1995) J. Immunol. 155, 3353-3359). Microinjection of cADPr into intact single T lymphocytes dose dependently induced repetitive but irregular Ca2+ spikes which were almost completely dependent on the presence of extracellular Ca2+. The Ca2+ spikes induced by cADPr could be blocked either by co-injection of cADPr with the specific antagonist 8-NH2-cADPr, by omission of Ca2+ from the medium, or by superfusion of the cells with Zn2+ or SK-F 96365. Ratiometric digital Ca2+ imaging revealed that single Ca2+ spikes were initiated at several sites (“hot spots”) close to the plasma membrane. These hot spots then rapidly formed a circular zone of high Ca2+ concentration below the plasma membrane which subsequently propagated like a closing optical diaphragm into the center of the cell. Taken together these data indicate a role for cADPr in Ca2+ entry in T-lymphocytes.

Intracellular Ca 2ϩ signaling is one of the major events transducing extracellular signals into many different types of living cells. In Jurkat T-lymphocytes it is well accepted that D-myoinositol 1,4,5-trisphosphate (IP 3 ) 1 releases Ca 2ϩ from an intracellular store located in the endoplasmic reticulum via its specific receptor (1)(2)(3). In addition, in T-lymphocytes a sustained long-lasting Ca 2ϩ entry can be observed in response to stimulation of the T cell receptor-CD3 complex (4 -6). This Ca 2ϩ entry is necessary for clonal expansion of T cells and therefore essential for a functional immune response. One of the basic mechanisms underlying Ca 2ϩ entry in electrically non-excitable cells appears to be the "capacitative" mechanism (reviewed in Refs. 7 and 8). The central idea of the capacitative mecha-nism is that a decrease in intraluminal Ca 2ϩ , e.g. as a result of IP 3 -induced Ca 2ϩ release, in turn leads to Ca 2ϩ entry. It is less clear, however, how this information from the intracellular stores is then transduced to the plasma membrane to activate opening of Ca 2ϩ channels (reviewed in Refs. 7 and 8).
In addition to IP 3 , cyclic ADP-ribose (cADPr) has been shown to be a potent natural ligand for mobilization of Ca 2ϩ from intracellular stores in sea urchin eggs (9). Now, there is increasing evidence that in addition to some other higher eukaryotic cell systems (10 -15), cADPr is also involved in intracellular Ca 2ϩ signaling in T-lymphocytes. We and others have recently demonstrated cADPr-induced Ca 2ϩ release from a non-thapsigargin sensitive intracellular store of T-lymphocyte cell lines (16,17). These stores were isolated from mouse Tlymphoma cells as light-density membrane vesicles carrying ryanodine receptors being different from the endoplasmic reticulum and localized close to the plasma membrane (17). In addition, the expression of brain-type ryanodine receptors also in human Jurkat T cells was observed (18), although in certain Jurkat subclones the expression of ryanodine receptors was not detectable (19).
To determine the function of cADPr in Ca 2ϩ signaling in intact T cells, especially a potential link between a decrease in the intraluminal Ca 2ϩ concentration in the cADPr-sensitive Ca 2ϩ stores and Ca 2ϩ entry, we microinjected cADPr into intact cells while recording digital Ca 2ϩ images of Fura2-loaded cells.
Digital Ca 2ϩ Imaging and Analysis of Data-Jurkat T-lymphocytes were cultured in RPMI 1640 medium supplemented with fetal bovine calf serum (10%, v/v), penicillin (100 units/ml), and streptomycin (50 g/ml). The cells were loaded with Fura2/AM as described (3). For parallel microinjection and Ca 2ϩ imaging experiments, the T cells had to be firmly attached to thin (0.2 mm) glass coverslips (to keep them in a defined position while forcing them with the microinjection pipette), but to stay inactivated in terms of Ca 2ϩ signaling. To achieve this, several different coating methods for the glass surface were tested, e.g. bovine serum albumin at 0.1, 0.5, or 1.0 mg/ml, agarose at 0.01, 0.05, 0.1, or 0.25 mg/ml, GelRite (Roth, Karlsruhe, Germany) at 0.05 or 0.1 mg/ml and poly-L-lysine at 0.01, 0.05, or 0.1 mg/ml. Many of these conditions resulted in activation of Ca 2ϩ signals, except low concentration of bovine serum albumin and poly-L-lysine. Of these different approaches, only the following combination of coatings proved to be the most useful: the glass coverslips were coated first with bovine serum albumin (5 mg/ml), and then with poly-L-lysine (0.1 mg/ml). Then, a small chamber (80 l volume) was fixed onto the coverslip and the Fura2-loaded cells were added and superfused with a buffer containing 140 mM NaCl, 5  T-lymphocytes was analyzed with a digital ratiometric imaging station (PhotoMed GmbH/Photon Technology, Wedel, Germany). The coated coverslips with cells were mounted on the stage of an inverted Axiovert 100 fluorescence microscope (Zeiss, Oberkochen, Germany). The excitation light source beam was split using an optical chopper, then passed through either a 340-or 380-nm optical filter, and guided into the microscope via fiber optics. The fluorescence intensity was filtered at 510 nm and then monitored using a CCD camera at a resolution of 525 ϫ 487 pixels (type C2400 -77, Hamamatsu, Garching, Germany). The data sampling rate usually was 1 ratio/5 s, in some experiments 1 ratio/s. Re-analysis of Ca 2ϩ image data was carried out using the so-called region-of-interest function of the ImageMaster software (Pho-toMed GmbH/Photon Technology, Wedel, Germany). Region of images were set either to cover the whole cell or subregions of the cell. The numerical median ratios and the corresponding free Ca 2ϩ concentrations were calculated by the software using external calibration.
Microinjection-Microinjections were done with an Eppendorf system (transjector type 5246, micromanipulator type 5171; Eppendorf-Netheler-Hinz, Hamburg, Germany) using Femtotips II as pipettes. The compounds to be microinjected were diluted to their final concentrations in intracellular buffer (20 mM Hepes, 110 mM KCl, 2 mM MgCl 2 , 5 mM KH 2 PO 4 , 10 mM NaCl, pH 7.2) and filtered (0.2 m) directly before filling into the Femtotips. Injections were made using the semiautomatic mode of the Eppendorf system at a pipette angle of 45°and the following instrumental settings: injection pressure 80 hPa, compensatory pressure 60 hPa, injection time 0.5 to 1 s, and velocity of the pipette 700 m/s. Under such conditions the injection volume was 1-1.5% of the cell volume as measured by microinjection of a fluorescent compound (Fura2-free acid) and subsequent determination of its quantity in a spectrofluorimeter at 360 nm (excitation) and 500 nm (emission).

Ca 2ϩ
Signaling Induced by Microinjection of cADPr-In response to microinjection of cADPr, the free intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ) increased in an oscillatory manner; either two to several irregular Ca 2ϩ spikes or a combination of spikes and sustained elevated Ca 2ϩ levels were observed ( Fig.  1, B, C, and D). The effect of cADPr was dose-dependent showing no increases in [Ca 2ϩ ] i when intracellular buffer was injected, and increasing responses when cADPr in the pipette was elevated stepwise to 100 M (Fig. 1). During every injection about 1-1.5% of the cell volume was injected. We assumed a fast dilution of cADPr in the cell being comparable to the diffusion reported for IP 3 in cytosolic extracts of Xenopus laevis (23). Considering the delay between injection and onset of the signal (Fig. 1, B, C, and D), a dilution factor of at least 50 must be assumed when estimating the effective intracellular concentration. Thus, at 1 M cADPr in the pipette, which was the threshold concentration (Fig. 1, B and F), the intracellular concentration of cADPr should be about 20 nM. Ca 2ϩ spikes with amplitudes up to 1.5 M [Ca 2ϩ ] i were observed at a pipette concentration of 10 M cADPr, amounting to an effective concentration of about 200 nM cADPr in the cell (Fig. 1, C and G). At 100 M cADPr in the pipette (about 2 M cADPr in the cell) usually 1 or 2 fast and large spikes were observed followed by a decline to an elevated level which was maintained for a relatively long period of time (Fig. 1, D and H). A pipette concentration of 100 M cADPr as compared with 10 M resulted in a considerably faster onset of the Ca 2ϩ signal (compare Fig. 1, G and H); however, later after microinjection, e.g. after 600 s, high Ca 2ϩ spikes were still observed at 10 M, but not at 100 M cADPr in the pipette (Fig. 1, C and D). At a pipette concentration of 1 mM cADPr (data not shown), the Ca 2ϩ signals were very similar to the ones obtained at 100 M indicating saturation of the dose-response relationship.
Ca 2ϩ Signaling Induced by cADPr and IP 3 Can Be Blocked Specifically-The effect of cADPr was specific, because it could be blocked by co-injection of a 10-fold excess of the specific antagonist 8-NH 2 -cADPr (Refs. 24 and 25; Fig. 2, A and B). For comparison, IP 3 was also microinjected. The resulting Ca 2ϩ spike pattern was somewhat more regular and usually not as long-lasting as compared with cADPr (Fig. 2C). Specificity was demonstrated by inhibition of the effect by co-injection with a 10-fold excess of the partial antagonist D-myo-inositol 1,4,6phosphorothioate (Fig. 2D).
Ca 2ϩ Entry Induced by cADPr-Since in most of the experiments, microinjection of cADPr did not result in a single Ca 2ϩ spike, but in longer lasting trains of spikes (Figs. 1, 2A, and  3A), the involvement of Ca 2ϩ entry in these sustained signals was investigated. Such cADPr-induced Ca 2ϩ signals were nearly completely abolished when the cells were superfused with extracellular buffer containing no Ca 2ϩ (Fig. 3B). Moreover, when cells superfused with 1 mM Ca 2ϩ and microinjected with 10 M cADPr were challenged with Zn 2ϩ (1 mM) after the first Ca 2ϩ spike, further spikes were completely inhibited (Fig.  3C). The effect of Zn 2ϩ could be washed out (data not shown), indicating that Zn 2ϩ acted by blocking Ca 2ϩ entry at the plasma membrane level. Further evidence for Ca 2ϩ entry in response to microinjected cADPr was obtained by blocking the Ca 2ϩ signals with the drug SK-F 96365 (Ref. 26; Fig. 3D). The IP 3 -induced Ca 2ϩ spikes also were largely dependent on the presence of external Ca 2ϩ (data not shown).
Spatial Development of Ca 2ϩ Spikes Induced by cADPr-The spatial development of a single Ca 2ϩ spike in response to microinjection of cADPr is characterized by (i) a rapid increase of [Ca 2ϩ ] i to a slightly elevated level within 5 to 10 s throughout the cell (Fig. 4, upper panel, image 1-3), (ii) a slowly generated wave of high [Ca 2ϩ ] i initiating from distinct sites ("hot spots") which then formed a circular zone close to the plasma membrane (Fig. 4, upper panel, images 3 and 4) and propagating like a closing optical diaphragm into the center of the cell (Fig.   FIG. 2 4 -6). At the top of the spike very high [Ca 2ϩ ] i were observed in the central part of the cell where the nucleus is located (Fig. 4, upper panel, image 6). At a slightly slower velocity, Ca 2ϩ was removed from the central part of the cell reaching a level that was slightly elevated (Fig. 4, lower panel, images [7][8][9]. This level in the central part of the cell was then maintained for some 100 s, while Ca 2ϩ oscillations were observed during this period of time in the subplasmalemmal cytoplasm (Fig. 4, lower panel, images 10 and 11). Finally, the cytoplasm also nearly reached basal Ca 2ϩ levels (Fig. 4, lower  panel, image 12), before the next Ca 2ϩ spike appeared (data not shown). Such cADPr-induced Ca 2ϩ waves propagating at high [Ca 2ϩ ] throughout the whole cell including the large nucleus of T-lymphocytes were not seen in all cases; in a number of cells Ca 2ϩ signals of high amplitude developed mainly in the cytoplasm reaching at the top of the spike a status comparable to image 4 in Fig. 4. Then [Ca 2ϩ ] i in this circular zone returned to basal concentrations. However, irrespective of whether cADPr might have been injected into the cytoplasm or in the nucleus, Ca 2ϩ waves were always initiated in the subplasmalemmal space pointing toward a crucial role of Ca 2ϩ influx in both cases.

DISCUSSION
The mechanism of Ca 2ϩ entry in T-lymphocytes as well as in other electrically non-excitable cells is not well understood. The trigger to switch on the capacitative Ca 2ϩ entry has been reported to be likely the depletion of the IP 3 -sensitive intracel-lular Ca 2ϩ pool in T cells (4 -6). The nature of the subsequent signal to stimulate Ca 2ϩ entry itself is discussed to be either the soluble calcium-influx factor CIF, protein-protein interaction between the IP 3 -receptor and the Ca 2ϩ entry channel(s), or a G-protein-mediated process (reviewed in Refs. 7 and 8). We confirmed the potential role of IP 3 in this process by showing that microinjection of this second messenger induced a train of Ca 2ϩ signals which depended on extracellular Ca 2ϩ (Fig. 2C).
As a novel finding we now add cADPr as a compound activating Ca 2ϩ entry in response to microinjection. During our experiments in nominally Ca 2ϩ -free medium only very small Ca 2ϩ signals were observed regardless whether cADPr or IP 3 was microinjected. However, these Ca 2ϩ signals were more pronounced at the site of microinjection, as compared with the signals averaged from the whole cell. Our data indicate that both ligands induced Ca 2ϩ release from their distinct target Ca 2ϩ stores, as has been clearly demonstrated for IP 3 and cADPr in permeabilized cells (16) and in vesicular membrane subfractions separated by density gradient centrifugation (17). The relatively small and local Ca 2ϩ release induced by both cADPr and IP 3 is also explained by the dilution of cADPr and IP 3 within the cell after microinjection (pipette concentrations 10 and 4 M; assumed intracellular effective concentrations about 200 and 80 nM). In the presence of extracellular Ca 2ϩ , the small Ca 2ϩ release then was followed for both ligands, IP 3 and cADPr, by a secondary Ca 2ϩ entry of much higher magnitude. This interpretation is in agreement with recent reports show- ing fast abrogation of Ca 2ϩ spikes induced by extracellular stimuli by omission of extracellular Ca 2ϩ in individual T cells (27)(28)(29). Also, the type of Ca 2ϩ wave (Fig. 4) argues for Ca 2ϩ influx since the wave started in a circular manner close to the plasma membrane all around the cell and propagated like a closing optical diaphragm into the center of the cell.
The mechanism by which cADPr stimulated Ca 2ϩ entry is not yet clear. However, at least three models are possible: first, microinjected cADPr released Ca 2ϩ from its target Ca 2ϩ store, namely membrane vesicles which are located close to surface receptor-capped structures (17). This primary event then led to activation of a Ca 2ϩ entry mechanism similar to the capacitative mechanism. The fact that SK-F 96365 inhibited both cADPr-and IP 3 -mediated Ca 2ϩ entry 2 may argue for this possibility. Second, cADPr may have opened directly Ca 2ϩ channels in the plasma membrane. This possibility appears to be less likely, since no such action of cADPr has ever been described. Furthermore, this hypothetical cADPr-responsive Ca 2ϩ channel then must have the same pharmacological properties, e.g. inhibition by 8-NH 2 -cADPr, which also is not very likely. As a third possibility, cADPr may have activated IP 3 receptors. However, all experiments carried out in permeabilized T cells suggest that cADPr acts completely independent of IP 3 receptors (16).
In conclusion, we add as a novel observation that Ca 2ϩ release by cADPr can induce Ca 2ϩ entry by a mechanism obviously being analogous to the capacitative mechanism induced by microinjection of IP 3 . Thus, cADPr, an intracellular ligand which is modulated in its activity indirectly by the T cell receptor-CD3 complex (30), may represent a new intracellular tool to control Ca 2ϩ entry in T cells. The recent discovery and molecular cloning of a whole family of human homologues of the Drosophila photoreceptor trp-channel points toward diversity in the structural basis for Ca 2ϩ entry (31) and opens up the possibility of identifying the Ca 2ϩ entry channel involved in cADPr-mediated Ca 2ϩ signaling in the future.