Calcineurin Is Downstream of the Inositol 1,4,5-Trisphosphate Receptor in the Apoptotic and Cell Growth Pathways*

The inositol 1,4,5-trisphosphate receptor (IP 3 R) is a calcium (Ca 2 1 ) release channel found on the endoplas-mic reticulum of virtually all types of cells. Human T lymphocytes (Jurkat) that are made deficient in IP 3 R do not generate Ca 2 1 signals in response to T cell receptor stimulation, fail to translocate the nuclear factor for activated T cells to the nucleus, and are remarkably resistant to induction of apoptosis with CD95 (Fas), dex-amethasone, g irradiation, and T cell receptor stimulation using anti-CD3 antibody. Expression of constitutively active calcineurin A in IP 3 R-deficient T cells restored nuclear factor for activated T cells translocation to the nucleus and dephosphorylation of Bad and rendered the cells sensitive to apoptotic inducers. Induction of apoptosis required both active calcineurin A ( D CnA) and activation-dependent colocalization of CnA with its substrate. Thus, the Ca 2 1 -dependent phosphatase calcineurin (CnA) is downstream of the IP 3 R in both the cell growth and apoptotic signaling pathways. 2 1 i work the IP 3 R 1 is necessary for both cell growth and that one of the key downstream targets of IP 3 Ca 2 1 in the pathway 3 inverted fluorescence accomplished

The inositol 1,4,5-trisphosphate receptor (IP 3 R) is a calcium (Ca 2؉ ) release channel found on the endoplasmic reticulum of virtually all types of cells. Human T lymphocytes (Jurkat) that are made deficient in IP 3 R do not generate Ca 2؉ signals in response to T cell receptor stimulation, fail to translocate the nuclear factor for activated T cells to the nucleus, and are remarkably resistant to induction of apoptosis with CD95 (Fas), dexamethasone, ␥ irradiation, and T cell receptor stimulation using anti-CD3 antibody. Expression of constitutively active calcineurin A in IP 3

R-deficient T cells restored nuclear factor for activated T cells translocation to the nucleus and dephosphorylation of Bad and rendered the cells sensitive to apoptotic inducers. Induction of apoptosis required both active calcineurin A (⌬CnA) and activation-dependent colocalization of CnA with its substrate. Thus, the Ca 2؉ -dependent phosphatase calcineurin (CnA) is downstream of the IP 3 R in both the cell growth and apoptotic signaling pathways.
Elevation of [Ca 2ϩ ] i signals diverse functions in T cells including cell growth and apoptosis (1)(2)(3)(4). Previous work has shown that the IP 3 R 1 is necessary for both cell growth (5) and cell death (3). It has been proposed that one of the key downstream targets of IP 3 -mediated Ca 2ϩ flux in the growth pathway is calcineurin A (CnA) (6). Moreover, sustained Ca 2ϩ elevation is required for CnA-dependent dephosphorylation of NF-AT and the subsequent translocation of this important transcription factor into the nucleus of T cells (7). Because IP 3 R-deficient cells are resistant to apoptosis (3) and FK506, a CnA inhibitor, prevents apoptosis (8 -10) we reasoned that CnA could be a Ca 2ϩ -dependent enzyme downstream of IP 3 R in the apoptotic and cell growth signaling pathways.
CnA is a serine-threonine phosphatase regulated by Ca 2ϩ and the Ca 2ϩ calmodulin complex (11). It is sensitive to immunosuppressive drugs such as FK506 and cyclosporin A. CnA inhibition results in the failure to translocate NF-AT to the nucleus, preventing activation of several cytokine genes, in-cluding interleukin-2, that are required for T cell growth. It has been shown that T cell activation requires a sustained increase in cytosolic [Ca 2ϩ ] to activate interleukin-2 gene transcription. IP 3 receptor-deficient Jurkat T cells are defective in intracellular Ca 2ϩ release and are resistant to apoptosis (3).
We now show that ⌬CnA restores NF-AT translocation and the apoptotic phenotype in IP 3 R-deficient T cells. The identification of CnA as a target for intracellular Ca 2ϩ release suggests that Ca 2ϩ -dependent dephosphorylation of downstream targets plays a role in the biochemical events triggering apoptosis and cell growth pathways.

EXPERIMENTAL PROCEDURES
Cell Culture and Apoptosis Induction-IP 3 R-deficient T cells (3) were cultured in RPMI 1640 medium containing 10% fetal bovine serum and 0.6 mg/ml hygromycin, whereas vector-and ⌬CnA-transfected cells were cultured with 0.6 mg/ml hygromycin and 0.4 mg/ml G418. To induce apoptosis, both vector-transfected and ⌬CnA-transfected cells were washed three times with serum-free RPMI 1640 medium and cultured in 10% RPMI 1640 medium in the presence and absence of plate-bound ␣CD3 (10 g/ml) as described (3). After the indicated time periods, the cells were washed and analyzed as described previously (3). In addition, the apoptotic cells were identified by the annexin V-fluorescein staining as per the manufacturer's instructions (CLONTECH Laboratories, Palo Alto, CA). The annexin assay and the cell death assays were used as complementary techniques to determine apoptotic cell death; both assays were performed, and representative results are shown.
Transfections-A 1.2-kilobase Sma-SalI fragment containing the coding region of the constitutively active CnA was cloned into pEGFP-C3 (CLONTECH Laboratories). 3.2 g of either vector or vector containing CnA DNA was used to transfect IP 3 R-deficient cells with the Lipofectin reagent (Life Technologies, Inc.). 36 h after transfection with either GFP vector alone or GFP vector containing ⌬CnA, transfected cells were pelleted by centrifugation at 1500 rpm for 5 min, resuspended in 1.5 ml of RPMI 1640 medium, layered over an equal volume of a Ficoll-Paque gradient (Amersham Pharmacia Biotech), and centrifuged for 20 min at 1500 rpm. The viable cells at the interphase were collected, washed in RPMI 1640 medium, and used for the experiments. The viability of these cells was determined using the trypan blue exclusion test. To induce apoptosis viable cells were cultured on ␣CD3coated plates as described (3). In all experiments only viable cells expressing levels of ⌬CnA comparable with the levels of CnA in the parental cells (as determined by immunoblotting; e.g. see Fig. 1D) were used. The relatively low expression of ⌬CnA was similar in all cells used for all experiments.
Immunofluorescence-Cells were activated for 15 min in the presence of the indicated stimuli, fixed with 3.7% paraformaldehyde, permeabilized with 0.2% Triton X-100 in buffer with divalent (BWD) buffer (125 mM NaCl, 5 mM KCl, 1 mM KH 2 PO 4 , 5 mM glucose, 10 mM NaHCO 3 , 1 mM Mgcl 2 , 1 mM CaCl 2 , and 20 mM HEPES, pH 7.4), and immunostained as described (12). Anti-NF-ATp antibody (Upstate Biotechnology Inc., Lake Placid, NY) and rhodamine conjugated to rabbit IgG were used for NF-AT detection. After washing, the cells were mounted and examined at ϫ 100 magnification using a scanning laser confocal attachment (Zeiss LSM 410) mounted on a Zeiss Axiovert 100 TV inverted fluorescence microscope equipped for fluorescence and transmittedlight imaging. Sample excitation and confocal image collection are accomplished using an argon-krypton laser and photomultiplier detectors. For green fluorescent protein, the excitation and barrier filters * This work was supported by National Institutes of Health (NIH) Grants AI39794, HL56180, and HL61503 (to A. R. M.), the American Heart Association (to A. R. M. and T. J.), the Richard and Lynne Kaiser Family Foundation (to A. R. M.), and by NIH Shared Instrumentation Grants 1S10 RR10506 and 5P30 CA13696. 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.
were set at 488 and 510 nm, respectively. Red fluorescence of NF-AT was measured at an excitation wavelength of 568 and emission was measured at 590 nM; the optical section thickness was 1 m. Specificity of immune staining was confirmed by the absence of fluorescence in cultures incubated with secondary antibodies alone or with pre-immune sera.
Western Blotting-Cells were pelleted by centrifugation and washed with phosphate buffered saline and then lysed with radioimmune precipitation assay buffer (1% Nonidet P-40, 1% deoxycholate, 0.1% SDS supplemented with 10 g/ml aprotinin, 2 g/ml leupeptin). After clearing cellular debris using centrifugation at 14,000 ϫ g, protein concentration in the supernatant was determined by Bradford assay (Bio-Rad). 100 g of protein from each lysate was denatured with SDS sample buffer and separated by 12% SDS-polyacrylamide gel electrophoresis. Following size fractionation using SDS-polyacrylamide gel electrophoresis, proteins were transferred to nitrocellulose membranes using a wet tank apparatus (Bio-Rad mini wet transfer unit) overnight at 4°C. Transfer membranes were then incubated with blocking solution (5% dry milk containing 0.5% Tween 20) for 1 h at room temperature followed by incubation with 1 g/ml primary antibody for 3 h. Membranes were washed four times with Tris-buffered saline with Tween 20 and incubated with the appropriate secondary antibody (1/ 1000 dilution) for 1 h followed by washing four times. Signal detection was performed with a chemiluminescence kit (Amersham Pharmacia Biotech). The primary antibodies were anti-CnA antibody (PharMingen; antibody recognizes both CnA and ⌬CnA) and anti-Bad antibody (Transduction Laboratories, Lexington, KY).

RESULTS
CnA is a heterodimer comprised of a catalytic A subunit and a Ca 2ϩ -binding regulatory B subunit (11). Constitutively active (Ca 2ϩ -independent) CnA (13) was stably expressed in IP 3 Rdeficient T cells (Jurkat) (Fig. 1A). Jurkat transfected with GFP vector alone exhibited a diffuse fluorescent pattern (Fig.  1B), whereas cells transfected with GFP/⌬CnA exhibited a cytoplasmic localization of the fluorescent signal (Fig. 1C) consistent with the cytoplasmic distribution of ⌬CnA. Typically both parental and IP 3 R-deficient T cells expressing high levels of ⌬CnA underwent spontaneous cell death. This finding was consistent with previous reports showing that expression of ⌬CnA induces cell death (14). However, it was possible to select cells that expressed low enough levels of ⌬CnA to remain viable (Fig. 1D). Immunoblot analysis of endogenous CnA and transfected ⌬CnA expression in IP 3 R-deficient T cells transfected with GFP vector or ⌬CnA showed that in the cell lines selected for the present study the level of ⌬CnA expression was comparable with that of the endogenous CnA (Fig. 1D). The growth rate of cells stably expressing low levels of CnA was the same as that of vector-transfected controls (Fig. 2).
To determine whether activation of CnA is a downstream signal required for apoptosis in Jurkat cells, sensitivity to inducers of apoptosis was examined using IP 3 R-deficient cells that are resistant to apoptosis and that express low levels of ⌬CnA (Fig. 3A). Stimulation with ␣CD3 induced apoptosis in a concentration-dependent manner in IP 3 R-deficient Jurkat cells expressing ⌬CnA but not in control cells (Fig. 3A). Annexin V assays, which are specific for apoptosis, showed that there was no increase in apoptotic cell death after stimulation with ␣CD3 in IP 3 R-deficient cells not expressing ⌬CnA (Fig. 3B). Thus, Fig. 3, A and B shows that expression of ⌬CnA, but not vector alone, rendered the previously resistant cells sensitive to induction of apoptosis by ␣CD3. This finding suggests that the resistance to apoptosis exhibited by IP 3 R-deficient cells (3) is causally linked to their inability to release intracellular Ca 2ϩ (5), which is required to activate CnA (7).
It has been shown that the activation state of NF-AT is critically dependent on CnA activity (15). NF-AT activation is required to coordinate the immune response (16). The immunosuppressive drugs cyclosporin A and FK506, which are potent inhibitors of CnA, inhibit NF-AT translocation (15). Because FK506 is also known to inhibit activation-induced apoptosis in T cells, the requirement of CnA in both cellular activation and apoptosis raises the immediate question of whether alternate pathways for cellular activation or cell death share common Ca 2ϩ -dependent signaling steps. IP 3 R-deficient T cells expressing low levels of ⌬CnA, but not those expressing vector alone, exhibited translocation of NF-AT in response to stimulation with ␣CD3 (Fig. 4A). In the cell expressing GFP vector alone NF-AT remains localized to the cytoplasm after stimulation with ␣CD3, as in the unstimulated cells, whereas in the cells expressing ⌬CnA the NF-AT signal is detected in the nucleus following stimulation with ␣CD3 (Fig. 4A). The Ca 2ϩ ionophore, ionomycin, which increases [Ca 2ϩ ] i , induced translocation of NF-AT in both vector-and ⌬CnA-transfected IP 3 R-deficient T cells (Fig. 4B). FK506 (1 M) blocked NF-AT translocation in both constitutively active ⌬CnA-expressing cells and in ionomycin-treated cells (Fig. 4B). These findings, coupled with our earlier reports showing that IP 3 R-deficient cells are resistant to apoptosis (3) and cannot undergo TCRmediated activation (5), suggest that IP 3 -induced intracellular Ca 2ϩ release is required for activation of apoptosis and for NF-AT translocation. The Bcl-2 family of proteins are important regulators of programmed cell death (17)(18)(19) and are comprised of both antiand pro-apoptotic members. It is generally believed that the ratio of pro-apoptotic and anti-apoptotic members of the Bcl-2 family members regulates cell fate by the ability of the mem-bers to form a complex with Bcl-2 or Bcl-X L (17,20). Bad is a member of the Bcl-2 family that is pro-apoptotic when it is associated with Bcl-2 or Bcl-X L . Recent studies have suggested that phosphorylation of Bad by the serine-threonine kinase, Akt (21), causes its release from Bcl-2 or Bcl-X L and eventual degradation after binding to 14-3-3 proteins (22). Thus, phosphorylation of Bad by Akt is anti-apoptotic, and dephosphorylation of Bad is pro-apoptotic. Akt is an important mediator of the growth-promoting pathways that are linked to growth factor receptor pathways such as those involved in insulin signal-

FIG. 4. Expression of ⌬CnA restores NF-AT translocation in IP 3 R-deficient T cells. A, IP 3 R-deficient cells transfected with GFP
vector or GFP plus constitutively active CnA (GFP-⌬CnA) either unactivated (Unstimulated; panels on left) or activated (Stimulated; panels on right) with ␣CD3 and stained with ␣NF-AT antibody. B, NF-AT translocation was determined by gating nuclear and cytoplasmic fluorescence signals before and after activation with ␣CD3 using IP 3 Rdeficient cells transfected with GFP vector (open bars) or GFP plus constitutively active CnA (GFP-⌬CnA; filled bars). The nuclear to cytoplasmic signal ratio from unactivated cells was subtracted from activated cells and was used as an indicator for translocation. Data represent three or more experiments. ing. Thus, signals such as growth factors and cytokines promote cell survival in part by activating pathways that result in Bad phosphorylation by Akt (23).
We hypothesized that Bad might be a downstream target of CnA. TCR-mediated activation of cells expressing low levels of ⌬CnA resulted in dephosphorylation of Bad protein (Fig.  5A). In contrast, dephosphorylated Bad was not observed in cells expressing ⌬CnA in the absence of ␣CD3 stimulation or in cells treated with ionomycin (3 M) alone (without ␣CD3 stimulation), suggesting that active CnA alone in the absence of an apoptotic inducer is insufficient to dephosphorylate Bad. This was in contrast to the translocation of NF-AT, which is also dependent on activation of CnA (15). We showed that ionomycin (3 M) alone (without ␣CD3 stimulation) caused NF-AT translocation. TCR-mediated activation of IP 3 R-deficient T cells that were not expressing ⌬CnA failed to dephosphorylate Bad (Fig. 5A). These findings suggested to us that when ⌬CnA was present at physiological levels in cells (as opposed to overexpressed) an additional signal was required to achieve colocalization of ⌬CnA with Bad. Indeed, ␣CD3-mediated TCR stimulation did induce the colocalization of ⌬CnA and Bad (Fig. 5B). These data suggest that the resistance to apoptosis in IP 3 R-deficient T cells is linked to the inability to dephosphorylate Bad because of insufficient CnA activation. DISCUSSION Elevation of intracellular Ca 2ϩ levels can influence a wide variety of biochemical processes including gene transcription. Activation of CnA and the resulting dephosphorylation of NF-AT are downstream events triggered by a rise in [Ca 2ϩ ] i . Modulation of CnA activation can alter the immune response. Indeed, viruses have evolved the ability to evade host defense systems by inhibiting CnA activity (24). NF-AT4-deficient thymocytes display heightened sensitivity to apoptosis mediated through the TCR, indicating that NF-AT4 might control the up-regulation of survival genes (25). It should be noted that NF-AT4 first associates with CnA in the cytoplasm and then both are translocated into the nucleus, where CnA may continue to act to counter the effects of nuclear NF-AT kinases (26). Our data are in agreement with the earlier findings that NF-AT translocation is defective in triple negative (type 1,2,3 IP 3 R-deficient) cells in response to B cell antigen receptor cross-linking (27). Taken together our data and those of others indicate that Ca 2ϩ -dependent signaling via CnA can activate cell growth pathways, e.g. through NF-AT, and cell death pathways, e.g. through Bad.
Cell death signals may be regulated by both transcriptional (e.g. dexamethasone-induced apoptosis of lymphocytes) and post-translational modification. It is not clear whether CnA regulates apoptosis solely via post-translational modifications or via transcriptional effects as well. Our data are consistent with the findings that dephosphorylation of Bad by CnA neutralizes the anti-apoptotic function of Bcl-2. The other possibilities could be that by dephosphorylating the transcription factors present in cytosol, CnA acts to translocate them to the nucleus to turn on the genes necessary for apoptosis. The corollary of this possibility is that survival factor-induced phosphorylation (such as by Akt) would retain the transcription factors in the cytosol and thus would prevent the activation of apoptotic genes. In this regard, it has been shown recently that Akt promotes cell survival by phosphorylating a Forkhead transcription factor, FKHRL1. Interestingly, dephosphoryl-ation of this transcription factor results in translocation to the nucleus and activates Fas ligand (28).
In summary, our study establishes that IP 3 R-mediated Ca 2ϩ release is required for CnA activation, NF-AT translocation, and Bad dephosphorylation. Phosphorylation of Bad promotes cell survival by protecting the anti-apoptotic function of Bcl-2, and CnA acts as a negative regulator of Bad via dephosphorylation and thus neutralizes Bcl-2 activity. Given this model, the magnitude and spatio-temporal activation of CnA can determine the specificity of a given response (e.g. either cell growth or cell death). The activation of specific scaffold, anchoring, and adapter proteins that recruit and place active CnA close to its substrate (29) could provide specificity. A mammalian scaffold complex has recently been identified that accounts for specificity in the mitogen-activated protein kinase signaling pathway (30). However, active CnA in the absence of ␣CD3 activation failed to induce apoptosis in IP 3 R-deficient T cells unless the levels of active CnA were higher than those observed in normal cells. Interestingly, raising [Ca 2ϩ ] i by itself (without ionomycin) was sufficient to induce translocation of NF-AT but not dephosphorylation of Bad. Thus, in addition to the elevation of [Ca 2ϩ ] i another signal is required for apoptotic but not necessarily for cell growth pathways. These data suggest that in addition to Ca 2ϩ -dependent activation of CnA, apoptotic inducers must also trigger localization of CnA to substrates to initiate apoptosis.