Activation of T Cell Calcium Influx by the Second Messenger ADP-ribose*

Stimulation of Jurkat T cells by high concentrations of concanavalin A (ConA) induced an elevation of the endogenous adenosine diphosphoribose (ADPR) concentration and an inward current significantly different from the Ca2+ release-activated Ca2+ current (ICRAC). Electrophysiological characterization and activation of a similar current by infusion of ADPR indicated that the ConA-induced current is carried by TRPM2. Expression of TRPM2 in the plasma membrane of Jurkat T cells was demonstrated by reverse transcription-PCR, Western blot, and immunofluorescence. Inhibition of ADPR formation reduced ConA-mediated, but not store-operated, Ca2+ entry and prevented ConA-induced cell death of Jurkat cells. Moreover, gene silencing of TRPM2 abolished the ADPR- and ConA-mediated inward current. Thus, ADPR is a novel second messenger significantly involved in ConA-mediated cell death in T cells.


Stimulation of Jurkat T cells by high concentrations of concanavalin A (ConA) induced an elevation of the endogenous adenosine diphosphoribose (ADPR) concentration and an inward current significantly different from the Ca 2؉ release-activated Ca 2؉ current (I CRAC ). Electrophysiological characterization and activation of a similar current by infusion of ADPR indicated that the ConA-induced current is carried by TRPM2. Expression of TRPM2 in the plasma membrane of Jurkat T cells was demonstrated by reverse transcription-PCR, Western blot, and immunofluorescence. Inhibition of ADPR formation reduced ConA-mediated, but not store-operated, Ca 2؉ entry and prevented ConA-induced cell death of Jurkat cells. Moreover, gene silencing of TRPM2 abolished the ADPR-and ConA-mediated inward current. Thus, ADPR is a novel second messenger significantly involved in ConA-mediated cell death in T cells.
Ca 2ϩ is a universal messenger from bacterial to mammalian cells since its concentration gradients across both organelle and plasma membranes can be efficiently used to communicate biological signals. Therefore, the control of the intracellular free Ca 2ϩ concentration [Ca 2ϩ ] i 3 is of crucial importance for the regulation of many cellular functions, including proliferation, contraction, fertilization, motility, apoptosis, and cell death (1). Receptor-mediated Ca 2ϩ influx from the extracellular space is one important mechanism to control [Ca 2ϩ ] i in non-excitable cells, e.g. leukocytes (2). Although the molecular machinery underlying Ca 2ϩ entry is still poorly defined, cation channels of the transient receptor potential (TRP) family that includes several subfamilies (3)(4)(5) are likely candidates for Ca 2ϩ entry pathways under the control of membrane receptors. TRPM2 (formerly LTRPC2 and TRPC7) is a member of the TRPM subfamily. TRPM2 forms non-selective Ca 2ϩ -permeable cation channels and is mainly expressed in brain and in cells of the immune system (6 -8). Opening of the channel is induced by intracellular ADP-ribose (ADPR; Refs. 6 and 7) and enhanced by increased cytosolic Ca 2ϩ (9). Whether NAD also activates TRPM2 currents is still controversial (6,7,10,11). The nudix box in the cytosolic C-terminal region of TRPM2, a conserved motif of enzymes with nucleotide pyrophosphatase activity, seems to be responsible for gating of TRPM2 by ADPR and possibly by NAD (6,7,12). An involvement of TRPM2 in cellular signaling processes has been proposed since the expression of TRPM2 confers susceptibility to oxidant-induced cell death (11).
A key question in the field relates to the potential role of the NAD metabolite ADPR as a second messenger. A function of NAD or ADPR as the missing link between specific extracellular signals and Ca 2ϩ influx mediated by TRPM2 has been hypothesized (8,13,14), but experimental proofs are missing so far. To test this hypothesis directly, we developed a method to measure intracellular levels of ADPR in Jurkat T cells (15). We report that cytosolic ADPR concentrations are raised in response to concanavalin A (ConA) and induce Ca 2ϩ entry through TRPM2, thereby significantly increasing [Ca 2ϩ ] i . Inhibition of intracellular ADPR formation or gene silencing of TRPM2 efficiently diminished receptor-mediated Ca 2ϩ influx carried by TRPM2. Moreover, blockade of ADPR formation also efficiently blocked ConA-induced cell death.

EXPERIMENTAL PROCEDURES
Electrophysiology-Membrane currents were recorded in the wholecell configuration of the patch clamp technique (16) or the perforatedpatch configuration with nystatin (17). An EPC9 patch clamp amplifier was used in conjunction with the PULSE stimulation and data acquisition software (HEKA Elektronik, Lamprecht, Germany). The patch electrodes were made from 1.5-mm diameter borosilicate glass capillaries and filled with intracellular solution. Data were low pass-filtered at 1 kHz and compensated for both fast and slow capacity transients. Series resistance was compensated by 50 -90%. All experiments were performed at room temperature with T cells slightly attached for the perforated patch mode, as described for the microinjection experiments, and firmly attached to high molecular weight poly-L-lysin for the wholecell mode, respectively. For the whole-cell configuration, the pipette solution contained 145 mM potassium glutamate, 8 mM NaCl, 1 mM MgCl 2 , and 10 mM EGTA, adjusted to pH 7.2 with KOH and to a free Ca 2ϩ of 100 nM with CaCl 2 . The external solution contained 145 mM NaCl, 2 mM MgCl 2 , 1 mM CaCl 2 , 2.8 mM KCl, 10 mM HEPES, 10 mM glucose, adjusted to pH 7.2 with NaOH. The cells were held at Ϫ60 mV, and current-voltage (I-V) relations were obtained using 250-ms voltage * This study was supported by grants from the Deutsche Forschungsgemeinschaft (to A. H. G. and J. R. S) and from the Werner-Otto-Foundation, the Wellcome Trust, and the Gemeinnü tzige Hertie-Stiftung (all to A. H. G.). This article is based in part on doctoral studies by A. G. and R. F. in the faculties of Chemistry and Biology, University of Hamburg. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1   For Western blot analysis of TRPM2 expression, membrane proteins from Jurkat cells (100,000 ϫ g centrifugation pellet; Ref. 20) were separated by a 7% SDS-PAGE (90 g/lane) and transferred by tank blotting onto a nitrocellulose membrane (Hybond ECL; Amersham Biosciences). The membrane was blocked with 5% dry milk protein in Trisbuffered saline. The primary antibody against TRPM2 (Abcam, Cambridge, UK, catalog number AB11167) was used at a concentration of 2 g/ml for 2 h, and the secondary antibody was a goat anti-rabbit horseradish peroxidase conjugate (Dianova, Hamburg, Germany) at 0.12 g/ml for 1 h. The blocking peptide was incubated with the anti-TRPM2 monoclonal antibody at 5-fold excess at 4°C overnight prior to addition to the blot membrane. Detection was done with Hyperfilm ECL and the ECL kit (Amersham Biosciences), according to the instructions of the manufacturer.
For immunostaining of TRPM2, Jurkat cells were fixed and permeabilized with p-formaldehyde and methanol as described (21). The primary antibody against TRPM2 was used at 20 g/ml for 1 h, and a fluorescein isothiocyanate-conjugated goat anti-rabbit antibody (Abcam, catalog number AB7086) at 10 g/ml (1 h of incubation) was used for detection. An Improvision imaging system (Improvision, Tübingen, Germany) was used at 100-fold magnification (Leica objective type HCX APO ϫ100/1.3 OIL U-V-I; numerical aperture 1.3), built around a fluorescence microscope (Leica DM IRE2). Images were taken at room temperature with a grayscale CCD camera (type C4742-95-12ER; Hamamatsu). Using a single stained cell, 125 images were recorded from bottom to top with a distance of 0.2 m between each layer.
Determination of Endogenous ADPR-Jurkat T cells were cultured as described (20). 5 ϫ 10 7 cells were incubated in extracellular buffer (22) for 30 min at 25°C and stimulated for different periods of time with 100 g/ml ConA Type IV (Sigma). In some experiments, the cells were incubated in the presence of 100 M Cibacron blue 3GA (Sigma) in extracellular buffer for 20 min before ConA stimulation. Cellular amounts of ADPR were determined as described (15). Briefly, cells were lysed in trichloroacetic acid, and cell debris was removed. The samples were divided into identical halves, to one of which genuine ADPR (2.5 nmol) was added for determination of the recovery. Protein precipitates were removed, and the supernatants were neutralized by extraction with diethyl ether. The samples were then purified by solid phase extraction and analyzed by reversed-phase HPLC using a Multohyp BDS C18 column (250 ϫ 4.6 mm, particle size 5 m; Chromatographie Service, Langerwehe, Germany) as described (15). The recoveries were calculated for each twin sample, and the amounts of ADPR were corrected for the respective recovery of each sample.
Microinjection and Single Cell Ca 2ϩ Imaging-Ratiometric Ca 2ϩ imaging was performed as described (21). Briefly, glass coverslips coated with bovine serum albumin and poly-L-lysine (0.1 mg/ml) were mounted on the stage of the Leica microscope (Leica DM IRE2). An Improvision imaging system as described above for immunocytochemistry was used. Illumination at 340 and 380 nm was performed using a monochromator system (polychromator IV; TILL Photonics). Images were acquired using Openlab software (v3.09; Improvision). Raw data images were stored on hard disk, and ratio images (340/380 nm) were constructed pixel-by-pixel.
Gene Silencing of TRPM2-Inserts coding for short hairpin RNAs (shRNAs) under control of the human U6 snRNA promoter were cloned into the vector pDsRed2-C1 (Clontech). One shRNA was targeted against the sequence 5Ј-CAG GCG CAT CCC ACT CTA T-3Ј (positions 4450 -4468 in TRPM2 cDNA, GenBank TM accession number NM_003307), the other was targeted against the sequence 5Ј-CAT CTG CAG GAA GAG CAT AAA-3Ј (positions 4099 -4119), and both contained the loop sequence 5Ј-TTC AAG AGA-3Ј. The vector pDsRed2-C1 without insert was used as control. Transfection was carried out using a Bio-Rad Genepulser (4-mm cuvettes, 2 ϫ 10 7 cells in 400 l, 200 V, 960 microfarads, 40 g of plasmid DNA). Intact cells were purified from debris using a Ficoll gradient. Cells were analyzed for fluorescence 24, 48, and 72 h after transfection. Only cells showing a distinct DsRed2 fluorescence were selected for current recordings.
Cell Death Assay-Jurkat cells were washed with RMPI 1640 and incubated at 10 6 cells/ml, either without or with different concentrations of 3GA (30 -300 M) for 30 min at 37°C. Then, ConA (final concentration, 100 g/ml) or a vehicle control was added. After a further 15-min incubation, 7.5% newborn calf serum was added. In some experiments, 1 mM EGTA was present in all solutions. The cells were incubated for 165 min at 37°C and thoroughly resuspended, and cell numbers were determined using an automated cell counter (Casy 1 TT, Schärfe System, Reutlingen, Germany).

RESULTS
Stimulation of intact Jurkat T cells by a high ConA concentration (100 g/ml) resulted in a slowly developing, large inward current (Fig.  1A) characterized by an almost linear I-V relationship (Fig. 1B), as determined in the perforated-patch mode in the presence of charybdotoxin to block Ca 2ϩ -activated K ϩ channels (26). The classical inward current triggered by T cell activation via low concentrations of mitogenic lectins (Ͻ10 g/ml) is I CRAC , which is characterized by high selectivity for FIGURE 1. Stimulation of Jurkat T cells with a high concentration of ConA induced both a membrane current and a sustained Ca 2؉ influx pathway distinct from the classical activation of I CRAC . A, membrane current amplitudes in intact Jurkat T cells were measured in the perforated-patch configuration in the presence of charybdotoxin (100 nM) at a membrane potential of Ϫ80 mV, either with (F) or without (E) stimulation by ConA (100 g/ml) at the time point indicated. Currents were analyzed in two batches of cells; ConAinduced currents were observed in 3 out of 10 and 5 out of 5 cells, respectively. B, the I-V relation of the ConA-induced current in A was obtained by subtraction of the current trace before stimulation from the current trace at the maximal amplitude. C, fura-2-loaded Jurkat cells were stimulated in the absence of extracellular Ca 2ϩ with either ConA (100 g/ml, upper panel) or thapsigargin (100 nM, lower panel) to activate I CRAC . Ca 2ϩ influx was then induced by the readdition of 1.35 mM extracellular Ca 2ϩ either in the presence or in the absence of the I CRAC inhibitor Gd 3ϩ (1 M). D, the sensitivity of the sustained Ca 2ϩ entry toward Gd 3ϩ was derived from the tracings in C. The differences in [Ca 2ϩ ] i at the plateau phase and before the readdition of Ca 2ϩ were calculated, corrected for the leak influx at the same concentration of Gd 3ϩ , and normalized to controls. E, Western blot analysis of TRPM2 expression in membrane preparations from Jurkat cells. The detection was done as indicated with both primary and secondary antibody, with an additionally blocking peptide against the primary antibody, or with the secondary antibody alone. The faint upper band represents either fulllength TRPM2 or its ⌬C form (161 kDa), whereas the lower band is indicative of SSF-TRPM2 (138 kDa). Unspecific bands and degradation products are marked by large and small arrowheads, respectively. F, localization of TRPM2 in Jurkat cells was analyzed by confocal immunofluorescence using the same primary antibody as in E and a fluorescein isothiocyanate-conjugated secondary antibody. The left panel shows TRPM2 localization at the largest diameter of the cell, whereas the right panel shows TRPM2 localization 2.6 m above the position of the left panel. Control images without the addition of the primary antibody were taken under essentially the same conditions. G, ADPRinduced membrane current in Jurkat cells were measured in the whole-cell configuration using repetitive voltage ramp commands starting from a holding potential of Ϫ60 mV. Depicted are representative current amplitudes at Ϫ80 mV either in the absence (E, n ϭ 21) or in the presence (F, n ϭ 12) of ADPR (300 M). H, the I-V relation of the ADPR-induced current was obtained by subtraction of the current trace measured directly after establishment of the whole-cell configuration from the fully activated current. In panels A, B, G, and H, data from current measurements in a single cell are displayed. FEBRUARY 3, 2006 • VOLUME 281 • NUMBER 5

JOURNAL OF BIOLOGICAL CHEMISTRY 2491
Ca 2ϩ , very low unitary conductance, inward rectification, remarkable sensitivity toward Gd 3ϩ , and activation by store depletion, e.g. by thapsigargin (27)(28)(29). Thus, the current observed here upon the addition of a high ConA concentration (100 g/ml) is different from I CRAC in terms of conductance and I-V relationship. To further distinguish Ca 2ϩ signals based on the novel current or I CRAC , the sensitivity of the sustained Ca 2ϩ entry phase toward Gd 3ϩ was analyzed upon activation by either thapsigargin or ConA. Thapsigargin-mediated Ca 2ϩ entry was effectively inhibited by Gd 3ϩ (Fig. 1C) with an IC 50 of ϳ0.1 M and full inhibition at 1 M (Fig. 1D) as described (30). In contrast, ConA-mediated sustained Ca 2ϩ entry was unaffected at 0.1 M Gd 3ϩ (Fig. 1D), and the IC 50 was about 1 order of magnitude higher as compared with activation by thapsigargin (Fig. 1D).
The characteristics of the novel current, especially the linear I-V relationship, the high current amplitudes, and the reduced sensitivity to Gd 3ϩ , indicated that it may be carried by the cation channel TRPM2. Indeed, expression of TRPM2 in cells of the immune system has been published (7). To confirm these published data, the expression and functional properties of TRPM2 were investigated in Jurkat T cells. Reverse transcriptase-PCR analysis confirmed the expression of TRPM2 in Jurkat T cells (data not shown). The use of different primer pairs revealed the expression of differentially spliced mRNAs: (i) an amplicon indicative of a TRPM2 isoform (termed SSF, see Ref. 19) significantly truncated in the N terminus; (ii) two amplicons indicative of the full-length mRNA; and (iii) an amplicon indicative of a TRPM2 isoform slightly truncated in the C terminus by deletion of exon 27 (see Ref. 10). Western blot analysis resulted in a major band at ϳ138 kDa and a faint band at 161 kDa (Fig. 1E). Comparison of the theoretically expected sizes deduced from the above mentioned mRNAs indicates that the 138-kDa protein represents the N-terminally truncated version (SSF), whereas the 161-kDa protein is full-length TRPM2. The specificity of the commercially available monoclonal antibody used was confirmed by blockade of specific staining of the 161-and 138-kDa bands by excess of the antigenic peptide used for monoclonal antibody preparation (Fig. 1E). Confocal immunostaining revealed localization of TRPM2 mainly on the cell surface; in the absence of primary antibody, only modest background staining was observed (Fig. 1F). The functional properties of TRPM2 in Jurkat T cells were analyzed by whole-cell mode patch clamp experiments. Upon intracellular application of ADPR, a slowly developing inward cation current was recorded at a membrane potential of Ϫ80 mV (Fig. 1G), whereas infusion of intracellular buffer did not induce any currents (Fig. 1G). The I-V relation of the current activated by ADPR was linear (Fig. 1H) and comparable with the current seen upon ConA stimulation in the perforated patch mode (Fig. 1,  A and B).
These data indicate that the large inward current observed upon ConA stimulation may be carried by the non-selective, non-rectifying cation channel TRPM2. Since TRPM2 is activated by ADPR (Fig. 1, G  and H), an HPLC method for the quantification of cellular amounts of ADPR was developed (15). Cellular ADPR levels were analyzed in Jurkat T cells either not stimulated or activated by ConA ( Fig. 2A, upper versus  lower panel). Under stimulation conditions, an increase in cellular ADPR concentrations was found consistently (Fig. 2A, lower panel). Time course analysis upon stimulation of Jurkat T cells with ConA revealed a statistically significant ϳ1.5-fold increase (Fig. 2B). A plateau phase was reached after 1 min of stimulation and remained elevated for at least 15 min (Fig. 2B). The increase of [Ca 2ϩ ] i upon ConA stimulation of these cells (Fig. 2B, solid line) tightly followed the rise of cellular ADPR levels (Fig. 2B, dashed line), indicating a causal relation between these two signaling events.
Although ConA induced a significant Ca 2ϩ entry (Fig. 1C), it was not clear whether TRPM2, as an unspecific cation channel supposed to mainly carry Na ϩ ions under physiological conditions, may contribute sufficiently to the ConA-induced Ca 2ϩ signal. The currents shown in Fig. 1, A and G, likely mainly reflect the entry of Na ϩ through TRPM2 since upon replacement of extracellular Na ϩ by the non-permeant cation N-methyl-D-glucamine, hardly any currents were resolved (10). To investigate directly whether ADPR-induced openings of TRPM2 are sufficient to elevate [Ca 2ϩ ] i by the entry of extracellular Ca 2ϩ , single cell Ca 2ϩ imaging experiments were carried out using intact Jurkat T cells (20,21). Upon microinjection of ADPR, a concentration-dependent increase of [Ca 2ϩ ] i was observed, whereas injection of intracellular buffer did not induce Ca 2ϩ signals (Fig. 3). The average tracings (Fig. 3, right panels) demonstrate that the velocity of the increase in [Ca 2ϩ ] i , rather than the amplitude of the [Ca 2ϩ ] i plateau, correlated with the ADPR concentration present in the pipette. It is important to note here that the actual intracellular concentrations of injected compounds are roughly 100-fold less as compared with the pipette concentration (which is given in Fig. 3); this is due to the fact that the injection volume is ϳ1% of the cell volume (23). Co-injection of ryanodine at a high concentration, known to block Ca 2ϩ release from ryanodine receptor (RyR), did not significantly reduce the effect of ADPR (Fig. 3). Microinjection of ADPR in the absence of extracellular Ca 2ϩ was almost without effect, confirming that Ca 2ϩ entry is the major, if not the only, mechanism operated by intracellular ADPR (Fig. 3, lower panels).
To further investigate the causal relationship between stimulation of cells, ADPR formation, and TRPM2 activation, two independent approaches were performed. First, expression of TRPM2 was suppressed by gene silencing using RNA interference. Jurkat T cells were transiently transfected with plasmids expressing two different shRNAs specifically targeted to TRPM2 mRNA. The usefulness of the selected shRNAs was initially analyzed by patch clamp recordings in the wholecell mode. Fig. 4B demonstrates that 24 and 72 h after transfection, current densities observed upon intracellular infusion of ADPR were largely suppressed (Fig. 4, B and C), whereas in control cells, the ADPRinduced current was fully developed (Fig. 4, A and C). Similar results were obtained with a second shRNA construct (data not shown). Then, recordings in the perforated patch mode were conducted upon cell stimulation by ConA. Although ConA induced the typical slowly developing, large inward current in control cells, largely reduced or no currents were observed in cells expressing the shRNA directed against TRPM2 (Fig. 4, D and E). To analyze the effect of TRPM2 gene silencing on Ca 2ϩ signaling, Ca 2ϩ imaging experiments were performed. Upon ConA addition, control cells responded with kinetics similar to those found in cell suspensions (compare Fig. 4F, with Fig. 2B; increase to a plateau level within ϳ150 s), albeit with a slightly decreased amplitude (compare Fig. 4F with Fig. 2B). Importantly, the cells expressing the shRNA directed against TRPM2 responded much more weakly (Fig. 4F).
In a completely independent approach toward the causal relation between cell stimulation and TRPM2 activation, the competitive NAD glycohydrolase (NADase) inhibitor Cibacron blue 3GA (31-33) was used. Preincubation of Jurkat T cells with 3GA reduced the ConAinduced increase of ADPR down to control levels (Fig. 5A). Ca 2ϩ influx into Jurkat cells was assayed using the same Ca 2ϩ -free/Ca 2ϩ -readdition protocol as introduced in Fig. 1C. Inhibition of ADPR formation by pretreatment of the cells with 3GA significantly reduced the peak and almost completely abolished the plateau phase of the ConA-induced Ca 2ϩ influx (Fig. 5, B and C). In contrast, the activation of store-operated Ca 2ϩ influx with the sarco/endoplasmic reticulum calcium ATPase (SERCA) inhibitor thapsigargin (34) was not affected by 3GA (Fig. 5C), ruling out an unspecific, generalized inhibition of Ca 2ϩ entry by 3GA.
Enzymatic conversion of NAD to ADPR may be catalyzed by the type II transmembrane protein CD38 (35). CD38 displays mainly NADase and, as a side activity, also ADP-ribosyl cyclase activity. However, since the active site of CD38 is located extracellularly (reviewed in Ref. 36), the involvement of this ecto-enzyme in the formation of intracellular Ca 2ϩmobilizing messengers is still controversial. The possible function of ecto-CD38 in the formation of intracellular ADPR in Jurkat cells was investigated using the non-membrane-permeant CD38 inhibitor nicotinamide 2Ј-deoxy-2Ј-fluoroarabinoside adenine dinucleotide (Ref. 18). Inhibition of extracellular CD38 activity had no significant effect on the ConA-induced [Ca 2ϩ ] i plateau (Fig. 5C). In contrast, blockade of the src family tyrosine kinases p59 fyn and p56 lck by PP2 (37) significantly, but not completely, suppressed the ConA-induced Ca 2ϩ plateau (Fig. 5C). Combined inhibition of ADPR formation by 3GA and tyrosine kinase activity by PP2 completely abrogated the ConA-induced [Ca 2ϩ ] i plateau ( Fig. 5C; compare with the leak influx induced by buffer addition).

Ca 2؉ Entry Mediated by the TRPM2 Agonist ADP-ribose
To further exclude a direct effect of 3GA on TRPM2 channel opening, ADPR-induced membrane currents in Jurkat cells were quantified in the presence or absence of 3GA in the patch pipette. TRPM2 currents were not significantly altered by the presence of 3GA (currents at 180 s: 300 M ADPR without 3GA, Ϫ129 Ϯ 25 pA (n ϭ 21, mean Ϯ S.E.); 300 M ADPR in the presence of 100 M 3GA, Ϫ143 Ϯ 53 pA (n ϭ 9, mean Ϯ S.E.)), confirming that 3GA inhibited ADPR formation but not the effects of ADPR on TRPM2.
To test whether ADPR-mediated Ca 2ϩ influx plays a functional role, we studied cell death in Jurkat cells. Treatment of Jurkat cells with high concentrations of ConA, a condition known to induce apoptosis in various cell types (38,39), resulted in a rapid decrease of the cell number (Fig. 6). This effect of ConA was completely reversed by the NADase inhibitor 3GA in a concentration-dependent manner as well as by chelation of extracellular Ca 2ϩ by EGTA (Fig. 6). In controls, 3GA on its own had no significant effect on the cell number (Fig. 6). In the presence of EGTA alone, a slight reduction of the cell number was observed, indicating a small cytotoxic effect of complete chelation of extracellular Ca 2ϩ by EGTA; however, the protective effect of EGTA in the presence of ConA was clearly evident (Fig. 6).

DISCUSSION
This study was designed to examine a potential role of ADPR as a second messenger mediating ConA-induced Ca 2ϩ entry in T-lymphocytes. Evidence for such a role is provided by the following findings. (i) ConA induced formation of ADPR and activated an inward current characteristic for TRPM2; (ii) TRPM2 is expressed in Jurkat T cells; (iii) physiological concentrations of intracellular ADPR are sufficient to elevate [Ca 2ϩ ] i significantly; (iv) suppression of TRPM2 expression by gene silencing almost completely abolished the ConA-induced inward current and Ca 2ϩ signaling; and (v) inhibition of ADPR formation not only diminishes Ca 2ϩ entry but also blocks ConA-mediated cell death in T cells.
One hallmark of our study is the receptor-mediated increase of cellular ADPR as analyzed by HPLC. The method applied here (15) has  Fig. 1. C, maximal current densities of the ADPR-induced currents in cells transfected with a vector control or a vector expressing an shRNA against TRPM2 24 or 72 h after transfection were calculated (n ϭ 5-8, mean Ϯ S.E., * , p Ͻ 0.05). D, Jurkat cells transiently transfected with a vector control or a vector expressing the shRNA4450 against TRPM2 were stimulated 72 h after transfection with ConA (100 g/ml) as indicated by the arrow, and membrane currents were measured in the presence of charybdotoxin (100 nM) in the perforated patch configuration as detailed in the legend for Fig. 1. E, maximal current densities of the ConA-induced currents 72 h after transfection with either a vector control or a vector expressing an shRNA against TRPM2 were calculated (n ϭ 4 -5, mean Ϯ S.E., #, p ϭ 0.068). F, Jurkat cells were loaded with fura-2, and single cell Ca 2ϩ imaging was performed at ϫ40 magnification as described under "Experimental Procedures." Cells transiently transfected with a vector control (control) or a vector expressing the shRNA4450 against TRPM2 (shRNA 4450) were analyzed 72 h after transfection. Images were acquired at ϳ1 image/1.5 s. ConA (100 g/ml) was added as indicated. Data are combined from three independent experiments and are displayed as mean Ϯ S.E. (n ϭ 25 (control), n ϭ 23 (shRNA4450)).
been modified from a two-step HPLC method for the quantification of cADPR (22). Similar methods have been used to quantify cellular nucleotides in the same concentration range (40,41). Using this technique, we found that upon stimulation of Jurkat cells with ConA, intracellular ADPR levels increased ϳ50 -60% (Fig. 2B). Based on cell volume determinations, and assuming an even distribution inside the cell, intracellular concentrations of ADPR in response to ConA were estimated to be in the range of 60 -90 M, depending on the time point after stimulation. These concentrations may be locally higher, due to biosynthesis at cer-tain sites of the cell, e.g. close to signaling protein complexes in the vicinity of the plasma membrane. Since such localized high concentrations of ADPR cannot be determined directly, their consideration may appear speculative at present. Nevertheless, there is convincing evidence that even the lower concentrations determined here are sufficient to gate TRPM2. The finding that in addition, Ca 2ϩ and cADPR act as co-modulators of TRPM2 (9, 42) fits into our model since it has long been known that stimulation of T cells by ConA activates rapid formation of InsP 3 and likely also of cADPR (20), which in turn leads to rapid Ca 2ϩ release from intracellular stores (43,44). This elevation of [Ca 2ϩ ] i is likely to represent the co-stimulus, allowing efficient TRPM2 channel opening even at suboptimal ADPR concentrations. Once TRPM2 channel activity is induced by ADPR, the influx of Ca 2ϩ ions through TRPM2 itself provides further supply with the co-stimuli in a feed-forward mechanism, even if the InsP 3 -and cADPR-driven Ca 2ϩ release terminates. The duration of TRPM2-mediated Ca 2ϩ entry, therefore, appears to be controlled mostly by the intracellular ADPR concentration. The kinetic data obtained for T cells showed a rather constant elevation of ADPR, at least for 15 min after stimulation. Obviously, in the context of induction of cell death by ConA, a sustained ADPR signal is required. Importantly, stimulation by ConA not only increased the cellular ADPR concentration but also induced a slowly developing inward current with the characteristic linear I-V relationship. Moreover, both inhibition of ADPR formation by 3GA and suppression of TRPM2 expression by gene silencing inhibited the ConA-induced formation of ADPR, ConA-FIGURE 5. The ConA-induced, but not the storeoperated, Ca 2؉ influx was reduced upon inhibition of ADPR formation by 3GA. A, Jurkat T cells were preincubated with or without the NAD glycohydrolase inhibitor 3GA (100 M) for 20 min and then stimulated for 5 min with either ConA (100 g/ml) or a vehicle control as indicated. ADPR levels were quantified by reversed phase HPLC (n ϭ 7-12, mean Ϯ S.E., * , p Ͻ 0.05) as detailed in the legend for Fig. 2. B, fura-2-loaded Jurkat cells were incubated with or without 3GA as in A, and Ca 2ϩ influx was quantified by Ca 2ϩ readdition as detailed in the legend for Fig. 1. C, fura-2-loaded Jurkat cells were preincubated with 3GA (100 M), the CD38 inhibitor nicotinamide 2Ј-deoxy-2Ј-fluoroarabinoside adenine dinucleotide (200 nM), and/or the Tyr-kinase inhibitor PP2 (1 M) as indicated. Ca 2ϩ influx upon stimulation with either ConA (100 g/ml, left panel) or thapsigargin (Tg) (100 nM, right panel) was quantified as in B and normalized to controls without inhibition (n ϭ 4 -20, mean Ϯ S.E., * , p Ͻ 0.05 to control). Note the leak influx in the absence of any stimulus (left bar). ara-F-NAD, nicotinamide 2Ј-deoxy-2Ј-fluoroarabinoside adenine dinucleotide. D, Mn 2ϩ quenching of fura-2 fluorescence was used to quantify Ca 2ϩ influx without secondary effects by elevation of [Ca 2ϩ ] i . Fura-2-loaded Jurkat cells were preincubated with or without 3GA (100 M) as indicated and then stimulated with ConA (100 g/ml) in the presence of extracellular Mn 2ϩ (25 M). Fluorescence was detected at the isosbestic wavelengths of fura-2, smoothed using a seven-point Savitsky-Golay algorithm, and normalized to the time point of stimulation. E, the initial quenching of fura-2 fluorescence after stimulation by ConA (100 g/ml) or thapsigargin (100 nM) either with or without 3GA preincubation was quantified by exponential decay curve fitting of the tracings displayed in D (n ϭ 5, mean Ϯ S.E., * , p Ͻ 0.05). The tracings shown in panels B and D are characteristic examples for buffer controls, stimulation by ConA, or preincubation with 3GA followed by stimulation with ConA. FIGURE 6. Inhibition of ADPR formation by 3GA reversed ConA-induced cell death in Jurkat T cells. Jurkat cells were cultured for 3 h in the presence or absence of ConA (100 g/ml), different concentrations of 3GA, or EGTA (1 mM) as indicated. After the incubation, the cell numbers were determined by an automated CASY cell counter and then normalized to controls (n ϭ 3-5, mean Ϯ S.E., * , p Ͻ 0.05 to control). mediated Ca 2ϩ entry, and ConA-induced inward current. In primary rat striatal cells, a comparable approach with similar results was conducted by Fonfria et al. (45). Instead of the endogenous TRPM2 agonist ADPR, the striatal cells were stimulated by extracellular H 2 O 2 , and instead of using an NADase inhibitor, Fonfria et al. (45) used the PARP inhibitor SB-750139. These results indicate that the origin of ADPR may be different in different cell types, e.g. either stemming from NAD in T cells or stemming from poly(ADPR)residues on proteins in striatal cells.
Another hallmark of our study is the first direct evidence that activation of TRPM2 by ADPR allows sufficient Ca 2ϩ ions to enter the cytosol to markedly increase [Ca 2ϩ ] i , as shown by simultaneous Ca 2ϩ imaging and microinjection of ADPR. For the interpretation of these effects, it should be considered that intracellular RyRs possess adenine nucleotide binding sites and thus are potential additional targets for ADPR (46). To exclude such an effect, we employed ryanodine at an inhibitory concentration. Under such conditions, the Ca 2ϩ signals evoked by ADPR were slightly reduced. Likely, this slight reduction was due to a block of Ca 2ϩinduced Ca 2ϩ release from RyR, but we cannot completely exclude that ADPR may exert minor effects on RyR in addition to its main effect on TRPM2. However, the fact that almost all Ca 2ϩ signaling was abolished in the absence of extracellular Ca 2ϩ suggests that the main, if not the only, target of ADPR is TRPM2.
Taken together, these data suggest a signaling pathway starting with cross-linked cell surface receptors, formation of ADPR, channel opening of TRPM2, and elevation of [Ca 2ϩ ] i followed by cell death. To further analyze the causal relationship of these building blocks, several inhibitors were employed. Firstly, the NADase inhibitor 3GA almost completely blocked ADPR formation. Importantly, 3GA specifically blocked all downstream events in the signaling chain, e.g. sustained Ca 2ϩ signaling and cell death. This strongly suggests that a causal relation between these individual building blocks exists. Furthermore, unrelated signaling modules, such as the capacitative Ca 2ϩ entry machinery, were not affected by 3GA. Finally, even TRPM2 channel gating by ADPR was almost unchanged in the presence of 3GA. A major role of capacitative Ca 2ϩ entry was further excluded by an ϳ10-fold higher sensitivity of this mechanism to Gd 3ϩ (30) as compared with the major Ca 2ϩ entry pathway operated by ConA. A major role for the ecto-NADase CD38 was ruled out by showing that the non-permeant inhibitor nicotinamide 2Ј-deoxy-2Ј-fluoroarabinoside adenine dinucleotide (18) did not markedly reduce ConA-mediated Ca 2ϩ signaling.
In conclusion, we propose that the second messenger ADPR and the Ca 2ϩ -permeable cation channel TRPM2 are key players in the regulation of cellular functions in T-lymphocytes. ADPR is formed in response to cross-linking of membrane receptors by ConA, induces Ca 2ϩ and Na ϩ entry through TRPM2, and thereby mediates Ca 2ϩ -dependent cell death.