Apoptosis Protection by Mcl-1 and Bcl-2 Modulation of Inositol 1,4,5-Trisphosphate Receptor-dependent Ca2+ Signaling*

Members of the Bcl-2 protein family play a central role in the regulation of apoptosis. An interaction between anti-apoptotic Bcl-xL and the endoplasmic reticulum (ER)-localized inositol trisphosphate receptor Ca2+ release channel (InsP3R) enables Bcl-xL to be fully efficacious as an anti-apoptotic mediator (White, C., Li, C., Yang, J., Petrenko, N. B., Madesh, M., Thompson, C. B., and Foskett, J. K. (2005) Nat. Cell Biol. 7, 1021–1028). Physiologically, Bcl-xL binds to the InsP3R to enhance its gating and Ca2+ signaling. Here we have discovered that structurally related proteins Bcl-2 and Mcl-1 function similarly. Bcl-2, Mcl-1 and Bcl-xL bind with comparable affinity to the carboxyl termini of all three mammalian InsP3R isoforms with important functional consequences. Stable expression of Bcl-2 or Mcl-1 lowered ER Ca2+ content and enhanced the rate of InsP3-mediated Ca2+ release in response to submaximal InsP3 stimulation in permeabilized wild-type DT40 cells but not in cells lacking InsP3R. In addition, expression of either Bcl-2 or Mcl-1 enhanced spontaneous InsP3R-dependent Ca2+ oscillations and spiking in intact cells in the absence of agonist stimulation. Bcl-2- and Mcl-1-mediated protection from apoptosis induced by staurosporine or etoposide was enhanced in cells expressing InsP3R, demonstrating that their interactions with InsP3R enable Bcl-2 and Mcl-1 to be fully efficacious anti-apoptotic mediators. Our data suggest a molecular mechanism that is shared by several anti-apoptotic Bcl-2 proteins that provides apoptosis resistance by direct interactions at the ER with the InsP3R that impinges on cellular Ca2+ homeostasis.

The Bcl-2 family of proteins are important regulators of programmed cell death with both pro-and anti-apoptotic members. In response to apoptotic stimuli, activation of the principle pro-apoptotic Bax and Bak causes increased mitochondrial membrane permeability and release of cytochrome c, a critical mediator in the commitment to cell death (1,2). The ability of pro-apoptotic Bcl-2 proteins to disrupt mitochondrial membrane integrity is held in check through heterodimeric interactions with anti-apoptotic members, including Bcl-2, Bcl-x L , and Mcl-1 (3,4).
In addition to mitochondria, both pro-and anti-apoptotic Bcl-2 proteins localize to the endoplasmic reticulum (ER), 2 where they regulate Ca 2ϩ fluxes across the ER membrane (5,6). The inositol 1,4,5-trisphosphate (InsP 3 ) receptor (InsP 3 R), a family of Ca 2ϩ release channels localized predominately in the ER, has been identified as a target for anti-apoptotic Bcl-2 and Bcl-x L (reviewed in Ref. 6). A direct interaction of Bcl-x L with all three isoforms of the InsP 3 R sensitizes InsP 3 R channel gating to extremely low [InsP 3 ] that exists in cells under resting conditions, and can result in reduced ER [Ca 2ϩ ] (7,8). The interaction modifies Ca 2ϩ -and InsP 3 -dependent regulation of channel activity in vitro and enhances Ca 2ϩ signaling in vivo. Enhanced Ca 2ϩ signaling in the form of spontaneous cytoplasmic Ca 2ϩ concentration ([Ca 2ϩ ] i ) oscillations or spiking, is correlated with increased apoptosis resistance (7,8). Similar effects have been observed in cells lacking both Bak and Bax, which had Bcl-x L or other anti-apoptotic Bcl-2 proteins more available to interact with the InsP 3 R (9). These data provide a molecular basis for the involvement of the ER as a major effector organelle in apoptosis, and they support a paradigm in which Bcl-x L is a direct effector of the InsP 3 R, increasing its sensitivity to InsP 3 and enabling ER Ca 2ϩ release to be more sensitively coupled to extracellular signals that enhances resistance to apoptotic stimuli.
Different pro-survival Bcl-2 proteins are differentially expressed in various cell types and engaged in responses to specific cellular stimuli (10,11). However, it is unknown whether the interaction of InsP 3 R with Bcl-x L , and the functional consequences for Ca 2ϩ signaling, ER Ca 2ϩ homeostasis, and apoptosis protection are more generally applicable to other antiapoptotic proteins. Bcl-2 family members can have distinct binding specificities (12,13). Furthermore, Bcl-2 has been reported to interact with the InsP 3 R (14) at different regions of the channel and with opposite functional effects on channel gating (14,15) compared with Bcl-x L (7,8). Thus, there is no a priori reason to assume that different anti-apoptotic Bcl-2 family members will interact with and affect InsP 3 R Ca 2ϩ release activity similarly. To address this, in the present study we explored the role of Bcl-2 and Mcl-1, anti-apoptotic Bcl-2 family members with structural homology to Bcl-x L . We have assessed the binding of Bcl-2, Bcl-x L and Mcl-1 to the carboxyl * This work was supported, in whole or in part, by National Institutes of Health termini of all three mammalian InsP 3 R isoforms and examined the relationship between ER Ca 2ϩ store content, [Ca 2ϩ ] i oscillations and apoptosis resistance conferred by the InsP 3 R-Bcl-2 protein interaction.
Biochemistry-Human Bcl-x L , Bcl-2, and Mcl-1 cDNAs were cloned into pFlag-C1 (Clontech). Sequences containing transmembrane helix 6 to the carboxyl terminus of rat InsP 3 R-1, InsP 3 R-2, and InsP 3 R-3 were subcloned into pEGX-6P-1 to create InsP 3 R-TM6ϩC constructs. Recombinant protein expression in E. coli and purification using glutathione beads were performed as described (8). The amounts of GST-InsP 3 R-TM6ϩC on beads were normalized by Western blotting using anti-GST antibody (GE Healthcare). Lipofectamine 2000 (Invitrogen) was used for Bcl-2 family member plasmid transfection into COS-7 cells. Concentrations of expressed Flagtagged Bcl-2 family member fusion proteins in COS-7 cell lysates were normalized using anti-Flag antibody M2 (Sigma). Pull-down assays were performed as described (8). Transfected COS-7 cells were washed twice with phosphate-buffered saline and harvested into 1 ml of phosphate-buffered saline containing 0.5% glycerol, 1% Triton X-100, 1 mM dithiothreitol, and protease inhibitor mixture (Sigma). After brief sonication (10 s) and centrifugation, the total protein concentration in the lysate was adjusted to 5 mg/ml and incubated with GST fusion protein (1 h, 4°C). Beads were centrifuged, washed three times, and prepared for Western blot. Western blot analyses were performed according to standard protocols. Flag-tagged Bcl-x L , Bcl-2, and Mcl-1 bound to GST-TM6ϩC beads were detected by SDS-PAGE and Western blotting using anti-Flag M2 antibody. We have determined that these protocols, with appropriate numbers of replicates and controls for amounts of proteins, exposure time, and gel loading, can discriminate affinities that differ by greater than a factor of 5. 3 Cytoplasmic [Ca 2ϩ ] Measurements-DT40 cells were plated onto a glass-bottomed perfusion chamber mounted on the stage of an inverted microscope (Olympus IX71) and incubated with Fura-2 AM (2 M; Invitrogen) for 30 min at room temperature in normal culture media. Cells were then continuously perfused with Hanks Balanced Salt solution (Sigma) containing (mM) 1.8 CaCl 2 and 0.8 MgCl 2 , pH 7.4. Fura-2 was alternately excited at 340 and 380 nm and the emitted fluorescence filtered at 510 nm was collected and recorded using a CCD-based imag-ing system running SimplePCI software (Hamamatsu Corp.). Dye calibration was achieved by applying experimentally determined constants in Equation 1.

ER [Ca 2ϩ
] Measurement-Cells were loaded with mag-Fura-2AM (5 M) in normal culture medium for 45 min at room temperature and then perfused with Hanks Balanced Salt solution for 15 min prior to recording. Cells were washed briefly with intracellular-like medium (ICM) containing: 125 mM KCl, 19 mM NaCl, 10 mM Hepes (pH 7.3 with KOH), and 1 mM EGTA and permeabilized by 2-3 min exposure to ICM containing ␤-escin (Sigma; 25 M). Store loading was achieved by switching to ICM containing Ca 2ϩ (free concentration 100 nM) and MgATP (1.5 mM). To induce Ca 2ϩ release, various concentrations of InsP 3 were applied in ICM without MgATP to prevent reuptake and containing a free [Ca 2ϩ ] of 1 M to produce optimal InsP 3 R activation (16). Data were acquired and calibrated as described for Fura-2.
Apoptosis Assays-Cells were dual-labeled with AlexaFluor488conjugated Annexin V and TOTO-3 (Invitrogen) following manufacturer's instructions and analyzed using flow cytometry (FACs Calibur; Beckton Dickinson) with CELLquest software. All experiments were performed at the RFUMS Flow Cytometry/Cell Sorting Research Support Laboratory.
Analyses and Statistics-In all experiments data from two independent clones expressing vector, Bcl-2 or Mcl-1 in WT or InsP 3 R-KO background were pooled and summarized as mean Ϯ S.E. For multiple comparisons, one-way ANOVA with Fisher's LSD (Least Significant Difference) posthoc comparisons were used to assess statistical significance of differences between means. Unpaired comparisons for normal and nonnormally distributed data employed the Student's t test and Mann-Whitney test, respectively. Differences between independent proportions were determined using a z-test. For all tests the differences between means were accepted as statistically significant at the 95% level (p Ͻ 0.05).

Bcl-x L , Bcl-2, and Mcl-1 All Bind with Similar Affinities to the Carboxyl Terminus of All Three InsP 3 R Isoforms-It was previously demonstrated that
Bcl-x L binds to all three mammalian InsP 3 R isoforms in a region localized to the carboxyl terminus extending from the luminal loop between transmembrane helices 5 and 6 to the carboxyl terminus (TM6ϩC) (8). This region contains the pore, including the putative pore helix, selectivity filter and TM6, and a distal 160 amino acids located in the cytoplasm. Therefore, we assessed whether or not Bcl-2 and Mcl-1 could also bind to this same channel region. GST-TM6ϩC fusion proteins of the rat types 1, 2, and 3 were immobilized on glutathione beads. The quantities of beads were titrated, and the amounts of GST-InsP 3 R fragments used for pull-down experiments, as well as the concentrations of Flagtagged Bcl-2 family member proteins expressed in COS-7 cell lysates, were adjusted to equivalent levels. The carboxyl-terminal fragments of each of the three InsP 3 R isoforms effectively pulled-down Mcl-1, Bcl-x L , and Bcl-2 with quantitatively similar (within a factor of 5) apparent affinities (Fig. 1A).

Bcl-2 and Mcl-1 Expression Alter Steady-State ER [Ca 2ϩ ]-
The functional effects of Bcl-2 and Mcl-1 on steady-state ER luminal [Ca 2ϩ ] ([Ca 2ϩ ] ER ) and ER Ca 2ϩ permeability were examined in stably transfected wild-type DT40 cells (DT40-WT) and in DT40 cells with all three InsP 3 R isoforms genetically deleted (DT40-InsP 3 R-KO) (17). The cell lines were characterized by Western blot and all subsequent experiments were carried out on two independent clones (shown in Fig. 1B). To monitor the ER store capacity directly, cells were loaded with the low affinity Ca 2ϩ indicator mag-Fura-2, which compartmentalizes in the ER lumen and cytoplasm. The plasma membrane was then permeabilized to remove cytoplasmic dye and the cells were bathed in a Ca 2ϩ -free intracellular solution (18). After equilibration, cells were exposed to MgATP and 100 nM free Ca 2ϩ to stimulate ER Ca 2ϩ uptake until a new steady-state filling level was achieved (Fig. 2, A-D). Of note, steady-state [Ca 2ϩ ] ER was markedly lower in DT40-WT cells expressing Bcl-2 and Mcl-1 compared with control cells expressing the vector only ( Fig. 2A). In contrast, steady-state [Ca 2ϩ ] ER was normalized in Bcl-2-and Mcl-1-expressing cells when ER Ca 2ϩ uptake was activated in the presence of the InsP 3 R inhibitor heparin (Fig. 2B). Because steady-state [Ca 2ϩ ] ER is determined by the balance between active uptake and passive release, these data suggest that reduced [Ca 2ϩ ] ER in the presence of Bcl-2 or Mcl-1 is due to increased Ca 2ϩ release mediated by InsP 3 R. To test this, similar experiments were performed in Bcl-2-or Mcl-1-expressing DT40-InsP 3 R-KO cells. In contrast to the effects observed in WT cells, Bcl-2 or Mcl-1 expression had no effect on [Ca 2ϩ ] ER in the InsP 3 -KO cells (Fig. 2, C and D) (Fig. 2E, inset). In response to subsaturating [InsP 3 ] (100 nM), a more complete store depletion was observed in the Bcl-2-and Mcl-1-expressing cells, which released about the same total Ca 2ϩ as vector controls, despite the reduced pool size prior to stimulation (Fig. 2F). These data suggest that the sensitivity of InsP 3 R-dependent Ca 2ϩ release is enhanced by the presence of Bcl-2 or Mcl-1.
The absence of MgATP during InsP 3 -induced Ca 2ϩ release prevents Ca 2ϩ reuptake into the ER (18 (Fig. 3A). In contrast, when Bcl-2 or Mcl-1 were overexpressed, the majority of cells displayed sustained [Ca 2ϩ ] i oscillations/spiking (Fig. 3, A and D), with fewer cells responding with a single [Ca 2ϩ ] i transient (ϳ55% for both Bcl-2 and Mcl-1; Fig. 3D). The amplitude of the initial [Ca 2ϩ ] i peak following anti-IgM was reduced in cells expressing Bcl-2, although the total Ca 2ϩ released during the 500 s after the initial [Ca 2ϩ ] i peak was unchanged (Fig. 3, B and C). Interestingly, in Mcl-1-ex- pressing cells the initial peak and total released Ca 2ϩ was larger than vector control cells (Fig. 3, B and C).
It was previously shown that expression of Bcl-x L enhances the frequency of spontaneous, InsP 3 R-dependent [Ca 2ϩ ] i oscillations in resting DT40-WT cells (7,8). This effect was attributed to the observed Bcl-x L -mediated enhanced sensitivity of InsP 3 R channels to low levels of InsP 3 that exist in un-stimulated cells. In the absence of stimulation, ϳ25% of DT40-WT cells expressing empty vector displayed spontaneous [Ca 2ϩ ] i oscillations/spiking (Fig. 3E). Expression of Bcl-2 increased this to nearly 40% and enhanced the proportion of cells displaying higher frequency oscillations (Fig. 3F) that resulted in an average oscillation frequency increase by nearly 2-fold compared with vector controls (Fig. 3G). Similarly, Mcl-1 expression also increased the number of oscillating cells (to nearly 50%) and high frequency oscillators, resulting in the mean oscillation frequency enhanced by 2.4-fold compared with control (Fig. 3H). Expression of Mcl-1 also increased the proportion of higher amplitude [Ca 2ϩ ] i spikes (Fig. 3F), elevating the mean spike amplitude by nearly 2-fold (Fig. 3G).
Taken together, the effects of Bcl-2 and Mcl-1 expression on InsP 3 -mediated [Ca 2ϩ ] i signals are similar to those observed for Bcl-x L , including enhanced InsP 3 R-mediated Ca 2ϩ flux (Fig. 2), InsP 3 R-dependent reduced steady-state [Ca 2ϩ ] ER (Fig. 2), and enhanced [Ca 2ϩ ] i oscillations in resting cells (Fig. 3). However, there were also distinctive features associated with the expression of each anti-apoptotic Bcl-2 protein. Bcl-2 expression reduced and Mcl-1 enhanced the amplitude of the initial Ca 2ϩ response to anti-IgM, whereas Bcl-x L has no effect (8). Neither Bcl-2 nor Bcl-x L (7,8) affected the amplitude of individual spontaneous [Ca 2ϩ ] i spikes in resting cells, whereas Mcl-1 increased the spike amplitudes. These differences between the effects of Bcl-2, Mcl-1, and Bcl-x L may be due to their different expression levels. Despite these differences, the predominant [Ca 2ϩ ] i signaling phenotypes induced by the expression of the three anti-apoptotic Bcl-2 proteins are qualitatively similar and consistent with their having a primary effect to increase the excitability of the InsP 3 R.  (19,20), cells were assayed at various times to assess apoptotic cell death. Cells labeled positive for annexin V or positive for annexin V and the vital stain TOTO-3 were considered to represent the total apoptotic cell population (Fig. 4A). In vectoronly expressing control cells, apoptosis initially progressed faster in WT than in InsP 3 R-KO cells, as indicated by a higher percentage of apoptotic cells at the 8-h and 2-h time points in STS-and ETS-treated samples, respectively (Fig. 4, B and C). This is consistent with previous reports identifying the InsP 3 R as a proapoptotic mediator during early apoptosis (7,19). Bcl-2 and Mcl-1 provided apoptosis protection when expressed in either the WT or InsP 3 R-KO backgrounds (Fig. 4, B-D). Bcl-2 provided better protection against STS than ETS (Fig. 4, B and  C), whereas MCl-1 was more effective against ETS (Fig. 4, D and  E). In general, protection was more effective when the two proteins were expressed in WT compared with InsP 3 R-KO cells. These data demonstrate that expression of the InsP 3 R enables both Bcl-2 and Mcl-1 to provide greater apoptosis resistance. The lack of difference between WT and InsP 3 -KO in the ability of Mcl-1 to inhibit apoptosis in response to STS (Fig. 4D) may indicate that the efficacy of protection afforded by this mechanism is stimulus dependent, suggesting that InsP 3 R interactions with different Bcl-2 proteins may couple to different downstream targets.

DISCUSSION
We previously identified a biochemical and functional interaction of Bcl-x L with the InsP 3 R channel carboxyl terminus that impinged on InsP 3 R gating activity and apoptosis protection (8). In the current study, we have determined that Bcl-2 and Mcl-1 also interact with this same region in all three rat InsP 3 R isoforms. By semi-quantitative pull-down assays, Bcl-x L , Bcl-2, and Mcl-1 appear to bind to this region in each isoform with approximately equal affinities. The similar binding to the three isoforms may not have been unexpected, as there is considerable sequence homology in the InsP 3 R region studied. However, the similar binding affinities of the three Bcl-2 family proteins is unexpected, and suggests that shared structural features among them are important for their interaction with the InsP 3 R carboxyl terminus. Direct binding of Bcl-2 to the InsP 3 R has been previously reported (14), whereas this is the first report of an interaction of Mcl-1 with InsP 3 R. Previous studies of Bcl-2 binding suggested weak binding to the InsP 3 R carboxyl terminus (21). However, the weak binding can be accounted for by the use of a carboxyl-terminal fragment that lacked critical binding determinants further upstream. 4 Bcl-2 was also reported to bind to a region within the channel regulatory domain (21). Both Bcl-x L and Mcl-1 bind weakly to this region, 4 whereas we show here that they, as well as Bcl-2, bind with similar and much higher affinity to the carboxyl-terminal construct employed in our studies. It was reported that the BH4 domain was necessary and sufficient in mediating the interaction between Bcl-2 and the InsP 3 R (22), whereas we have observed robust binding between the InsP 3 R carboxyl terminus and Mcl-1 that lacks a BH4 domain (23). Thus, Bcl-2 protein interactions with the InsP 3 R are likely to be complex and involve multiple binding sites within the channel. The interactions of Bcl-2 and Bcl-x L with the InsP 3 R have been previously demonstrated to affect cellular Ca 2ϩ homeostasis, but the molecular mechanisms are controversial. Expression of Bcl-2 and Bcl-x L attenuate InsP 3 R-dependent Ca 2ϩ release in response to maximal InsP 3 stimulation, but this has been attributed either to inhibition (14,22,24) or activation (7)(8)(9) of InsP 3 R channel activity by these proteins. In support of the former, recombinant Bcl-2 inhibited reconstituted InsP 3 R gating in planar lipid bilayers (14). In the latter case, reduced InsP 3 -mediated Ca 2ϩ signals were proposed to be caused by Bcl-2 protein-mediated reduction of steady-state [Ca 2ϩ ] ER (reviewed in Ref. 6). We previously identified the mechanism of this reduction by demonstrating that full-length, soluble recombinant Bcl-x L activated InsP 3 R single channel gating in native ER membranes, by an allosteric mechanism that sensitized the channel to low [InsP 3 ] (7,8). This activation mechanism is robust, because it was observed for the three mammalian InsP 3 R isoforms as well as the Sf9 insect cell InsP 3 R. We have been unable to perform similar studies with Bcl-2 because of its well-recognized poor solubility. Sensitization by Bcl-x L of the InsP 3 sensitivity of the InsP 3 R can account for reduced steady-state [Ca 2ϩ ] ER observed in some studies (7,8) because increased InsP 3 R activity resets the ER membrane Ca 2ϩ pumpleak balance. Here, we demonstrate directly, using an ER-localized Ca 2ϩ indicator, that both Bcl-2 and Mcl-1, like Bcl-x L , enhance the InsP 3 sensitivity of InsP 3 R-mediated Ca 2ϩ release that results in decreased steady-state [Ca 2ϩ ] ER . These effects of Mcl-1 and Bcl-2 are InsP 3 R-dependent because store filling is normalized by InsP 3 R inhibition, and they are not observed when Bcl-2 or Mcl-1 is expressed in InsP 3 R-KO cells. These results are significant because they demonstrate that reduced [Ca 2ϩ ] ER , observed when Bcl-2 is overexpressed in many studies (25)(26)(27), is caused by Bcl-2 interaction with the InsP 3 R. Our results therefore conflict with the reported inhibition of InsP 3 R channel gating by Bcl-2 (14,15). Furthermore, they reveal for the first time that Mcl-1 interacts with the InsP 3 R to produce qualitatively similar effects to Bcl-2 and Bcl-x L on ER Ca 2ϩ release and steady state [Ca 2ϩ ] ER .
Whereas our results suggest a common mechanism by which anti-apoptotic Bcl-2 family proteins impinge on Ca 2ϩ signaling, the mechanisms whereby enhanced InsP 3 R Ca 2ϩ signaling impinges on apoptosis protection are themselves controversial. Evidence that ER Ca 2ϩ release and subsequent uptake into mitochondria promotes cytochrome c release, and that reduced [Ca 2ϩ ] ER can minimize this effect and provide apoptosis protection, have supported the hypothesis that Bcl-2/Bcl-x L impinges on apoptosis by modulating [Ca 2ϩ ] ER (28). However, such a the model is not supported by observations in many studies of unchanged [Ca 2ϩ ] ER during Bcl-2-mediated apoptosis protection (5). Furthermore, co-expression of Bcl-x L with each individual InsP 3 R isoform in an InsP 3 R null background reduced [Ca 2ϩ ] ER only in cells expressing type-3 InsP 3 R whereas it provided InsP 3 R-dependent apoptosis protection in all of them. Thus, while lowering [Ca 2ϩ ] ER can provide apoptosis protection, it may not be the only or even most important mechanism involved. In contrast, a consistent feature associated with Bcl-x L /InsP 3 R-dependent apoptosis protection was the presence of enhanced InsP 3 R-dependent spontaneous [Ca 2ϩ ] i spiking or oscillations (7)(8)(9). We now show that expression of either Bcl-2 or Mcl-1 provides InsP 3 R-dependent apoptosis resistance and enhances InsP 3 R activity that is similarly manifested as an increase in the frequency and number of cells displaying of spontaneous [Ca 2ϩ ] i spiking. Similar enhanced, low-level InsP 3 R-dependent [Ca 2ϩ ] i signaling has been observed previously in other Bcl-2 expression cell models (15,27).
Our data suggest that a primary mechanism by which the ER impinges on apoptosis involves enhanced low-level spontaneous [Ca 2ϩ ] i signaling mediated similarly by Bcl-x L , Bcl-2, and Mcl-1 interactions with the InsP 3 R. The mechanisms that transduce enhanced spontaneous InsP 3 R [Ca 2ϩ ] i spiking into apoptotic resistance are unknown but are of obvious importance. Frequency modulation of [Ca 2ϩ ] i signals regulates numerous processes, including gene transcription (29). Periodic transient InsP 3 R-mediated [Ca 2ϩ ] i oscillations activate NF-B to afford apoptosis resistance (30). Thus, enhanced lowlevel spontaneous [Ca 2ϩ ] i signaling mediated by Bcl-x L , Bcl-2, and Mcl-1 may impinge on the nucleus to alter gene expression that promotes apoptosis resistance. Periodic InsP 3 R-mediated Ca 2ϩ release might also be an efficient mechanism to facilitate Ca 2ϩ delivery to mitochondria at a rate and magnitude that is optimal for stimulating mitochondrial bioenergetics but insufficient to trigger irreversible mitochondrial permeability transition and membrane disruption. In this model, increased apoptosis protection is linked to enhanced cellular bioenergetics (8,31). Future studies are necessary to identify the molecular mechanisms whereby constitutive low-level Ca 2ϩ signaling affords apoptosis protection.
In summary, we have discovered that Bcl-2, Bcl-x L , and Mcl-1, three structurally related anti-apoptotic proteins with differential expression and responses to different stimuli, bind with similar affinity to the carboxyl terminus of all three isoforms of the InsP 3 R. We have determined that both Bcl-2 and Mcl-1 have similar functional effects to those reported for Bcl-x L (7,8), in that they provide enhanced InsP 3 R activity and InsP 3 R-dependent spontaneous [Ca 2ϩ ] i signaling and apoptosis resistance. We conclude that multiple anti-apoptotic Bcl-2 homologues function at the ER by interacting with the InsP 3 R to generate pro-survival Ca 2ϩ signals that adapt cells to withstand apoptotic stimuli.