The Coxsackievirus 2B Protein Suppresses Apoptotic Host Cell Responses by Manipulating Intracellular Ca 2 (cid:1) Homeostasis*

Enteroviruses, small cytolytic RNA viruses, confer an antiapoptotic state to infected cells in order to suppress infection-limiting apoptotic host cell responses. This antiapoptotic state also lends protection against cell death induced by metabolic inhibitors like actinomycin D and cycloheximide. The identity of the viral antiapoptotic protein and the underlying mechanism are unknown. Here, we provide evidence that the coxsackievirus 2B protein modulates apoptosis by manipulating intracellular Ca 2 (cid:1) homeostasis. Using fluorescent Ca 2 (cid:1) indicators and organelle-targeted aequorins, we demonstrate that ectopic expression of 2B in HeLa cells decreases the Ca 2 (cid:1) content of both the endoplasmic reticulum and the Golgi, resulting in down-regulation of Ca 2 (cid:1) signaling between these stores and the mitochondria, and increases the influx of extracellular Ca 2 (cid:1) . In our studies of the physiological importance of the 2B-induced alterations in Ca 2 (cid:1) signaling, we found that the expression of 2B suppressed caspase activation and apoptotic cell death induced by various stimuli, including actinomycin D and cycloheximide. Mutants of 2B that were defective in reducing the Ca 2 (cid:1) content of the stores failed to suppress apoptosis. These data implicate a functional role of the perturbation of intracellular Ca 2 (cid:1) compartmentalization in the enteroviral strategy to suppress intrinsic apoptotic host cell responses. The putative down-regulation of an endoplasmic reticulum-depend-ent apoptotic pathway is discussed. activation, cells were fixed and stained as described (17) with an anti-cleaved caspase-3 (Asp-175) antibody (Cell Signaling Technology), which specifically recognizes active caspase-3 but not its inactive zymogen, and a Texas Red-conjugated goat anti-rabbit antibody. The percentage of caspase-3-positive GFP-expressing cells was determined de visu by counting at the microscope. To assay cell survival, cells were washed thoroughly to eliminate apoptotic cells after incubation with the apoptotic drugs. The percentage of fluorescent cells before and after a challenge with the apoptotic drugs was deter- mined de visu by counting at the microscope.

Many viruses are endowed with the potential to manipulate cell death pathways in order to prevent premature abortion of the infectious cycle (1,2). The molecular mechanism by which enteroviruses manipulate the life span of their host cell is, as yet, poorly understood. Enteroviruses (coxsackievirus, poliovirus, and ECHO virus) are nonenveloped, cytolytic RNA viruses that replicate their genome at the secretory pathway-derived membrane vesicles that accumulate in the cytoplasm of the infected cell (3). Classically, lytic viral replication is believed to induce canonical cellular necrosis by destruction of the plasma membrane, causing the collapse of ionic gradients. However, evidence is accumulating that the issue of how enteroviruses induce cell death is much more complex. Enterovirus infection leads to the development of the so-called cytopathic effect (CPE), 1 a necrosis-like type of cell death that is characterized by rounding up of the infected cells, distortion and displacement of the nuclei, condensation of chromatin, and increased plasma membrane permeability. This type of cell death is the result of a complex interplay between apoptosisinducing and apoptosis-suppressing functions encoded by the enterovirus genome (4 -6). Early in infection, i.e. upon translation of the enterovirus RNA genome, sufficient quantities of a putative pro-apoptotic function are produced to trigger an apoptotic response. Concomitantly with the onset of viral replication, however, the implementation of the virus-induced apoptotic program is abruptly interrupted, suggesting that enteroviruses also encode an antiapoptotic function (6). This antiapoptotic function also renders infected cells resistant against non-viral apoptotic stimuli like cycloheximide and actinomycin D (4). The apoptosis-suppressing function dominates upon productive infection (i.e. conditions that allow efficient virus replication). Under these conditions, only at late stages (i.e. after the development of CPE) can some signs of apoptosis be detected (7). Virus replication and CPE, however, are not sensitive to caspase inhibitors or the overexpression of Bcl-2 (5). The apoptosis-inducing function dominates upon non-permissive infection (i.e. conditions that restrict virus growth). The full-blown apoptosis that is induced under these latter conditions is efficiently suppressed by caspase inhibitors as well as by Bcl-2 overexpression (5).
Recent studies have shed some light on the identity of the putative apoptosis-inducing enterovirus proteins. Individual expression of the viral proteinase 2A pro , which inhibits cap-dependent translation of cellular mRNAs, or the expression of 3C pro , which shuts off host cell RNA transcription, results in apoptotic cell death (8,9). Little is known about the identity of the apoptosis-suppressing factors. The 3A protein suppresses the extrinsic apoptotic pathway by eliminating labile receptors from the cell surface (10) through its ability to inhibit protein secretion (11). The enteroviral protein that is responsible for the suppression of the intrinsic apoptotic host cell responses has not yet been identified.
Ca 2ϩ is one of the most versatile and universal signaling mediators in cells and is required for the activation of many cellular processes. Increasing evidence indicates that alterations in the finely tuned intracellular Ca 2ϩ homeostasis and compartmentalization can lead to cell death, either through apoptosis or necrosis (12). The switch from the control of physiological functions to the death program most likely involves alterations in the tightly regulated spatio-temporal Ca 2ϩ pattern or alterations at the level of organelles (e.g. mitochondria or ER/Golgi) or effector proteins (e.g. calpain or calcineurin) that are activated by Ca 2ϩ (13). Enteroviruses have a profound effect on intracellular Ca 2ϩ homeostasis (14,15). We have previously shown that infection of HeLa cells with the coxsackievirus results in a reduction of the amount of Ca 2ϩ that can be released from the intracellular stores. In parallel, a gradual increase in the cytosolic Ca 2ϩ concentration ([Ca 2ϩ ] cyt ) is observed due to the influx of extracellular Ca 2ϩ (15).
The enterovirus 2B protein has been implicated in the virusinduced alterations in intracellular Ca 2ϩ homeostasis (15,16). The 2B protein is a small (97-99 aa) membrane-integral replication protein (17) that, in infected cells, is localized at the surface of the ER-and Golgi-derived membrane vesicles at which viral replication takes place (3,18,19). All enterovirus 2B proteins contain two hydrophobic regions, of which one is predicted to form a cationic amphipathic ␣-helix (20,21). This amphipathic ␣-helix displays characteristics typical for the group of membrane-lytic ␣-helical peptides that can build membrane-integral pores by forming multimeric transmembrane bundles (22,23). Homomultimerization reactions of 2B have been demonstrated by yeast and mammalian two-hybrid (24,25) and biochemical approaches (26) and in living cells by using fluorescence resonance energy transfer (FRET) microscopy (27). These data strongly suggest that 2B is responsible for the reduction of the [Ca 2ϩ ] in ER and Golgi ([Ca 2ϩ ] ER and [Ca 2ϩ ] Golgi ) by building pores in the membranes of these organelles. Direct evidence that 2B indeed causes a reduction in the [Ca 2ϩ ] of the stores is, however, still lacking.
In this study, we investigated Ca 2ϩ homeostasis in 2B-expressing cells and evaluated the physiological importance of the alterations in Ca 2ϩ signaling for the implementation or suppression of the different cell death programs. Using fluorescent Ca 2ϩ indicators and organelle-targeted aequorins (AEQs; genetically encoded Ca 2ϩ sensors), we demonstrate that 2B indeed reduces [Ca 2ϩ ] ER and [Ca 2ϩ ] Golgi in HeLa cells. We show that this leads to a decrease in the amount of Ca 2ϩ that can be released from these stores and, as a consequence, results in the stimulus-induced rise of [Ca 2ϩ ] in the mitochondria ([Ca 2ϩ ] mt ). Moreover, the influx of Ca 2ϩ from the extracellular medium is increased, and, thus, the [Ca 2ϩ ] cyt responses are larger. In our studies of the functional importance of the 2B-induced manipulation of intracellular Ca 2ϩ signaling, we found that the expression of 2B suppressed apoptosis induced by certain stimuli, including actinomycin D and cycloheximide. 2B mutants that were unable to reduce the Ca 2ϩ content of the stores failed to protect against apoptosis. These data implicate a functional role of the 2B-induced perturbation of intracellular Ca 2ϩ compartmentalization in the enteroviral strategy to suppress premature abortion of the viral life cycle and provide a physiological example of the regulatory role of Ca 2ϩ signaling in the modulation of apoptotic cell death.
The coxsackievirus B3 used in this study was derived from plasmid pCB3/T7, which contains a cDNA of coxsackievirus B3 (strain Nancy) behind a T7 RNA polymerase promoter. Virus titers were determined by end point titration as described (20).
For the Fura-2/AM measurements, 3 ϫ 10 5 CHO cells were transfected with 2 g of either GFP (S65T) or 2B-GFP (S65T) using Lipofectin (Invitrogen). After 24 h, 5 ϫ 10 4 cells were seeded onto 24-mm coverslips and grown for another 16 h before [Ca 2ϩ] cyt measurements. For the aequorin measurements, HeLa cells were seeded onto 13-mm coverslips and cotransfected using the Ca 2ϩ phosphate technique with 1 g of erAEQ, GoAEQ, cytAEQ, or mtAEQ, together with 3 g of 2B-EGFP or in combination with the same amount (3 g) of mt-GFP (which served as negative control). For the cell death experiments, HeLa cells were seeded onto 24-mm coverslips and transfected with 4 g of 2B-EGFP or mt-GFP.
Fura-2/AM Measurements-The fluorescent Ca 2ϩ indicator Fura-2 was used to measure [Ca 2ϩ] cyt at the single cell level essentially as described (15). Briefly, cells were incubated in medium supplemented with 2.5 M Fura-2/AM for 30 min, washed with HT buffer to remove the extracellular probe, supplied with preheated HT buffer (supplemented with 1 mM CaCl 2 ), and placed in a thermostated (37°C) incubation chamber on the stage of an inverted fluorescence microscope (Nikon). Dynamic video imaging was performed using the MagiCal hardware and TARDIS software (Joyce Loebl, Tyne and Wear, UK). Fluorescence was measured every 2 s with the excitation wavelength alternating between 340 and 380 nm and the emission fluorescence being recorded at 492 nm. At the end of the experiment, a region free of cells was selected, and one averaged background frame was collected at each excitation wavelength for background correction. To measure the [Ca 2ϩ ] of the stores, the amount of thapsigargin-releasable Ca 2ϩ was determined. After recording the cells in Ca 2ϩ -containing HT buffer, the medium was replaced with Ca 2ϩ -free HT buffer supplemented with 0.5 mM EGTA. After 10 min, the cells were challenged with 1 M thapsigargin, and [Ca 2ϩ ] cyt was measured as described above.
Aequorin Measurements-Organelle-targeted aequorin constructs were used to measure the [Ca 2ϩ ] in the cytosol and different organelles at the cell population level (at 36 h post-transfection). For the mtAEQ and cytAEQ measurements, the cells were incubated with 5 M coelenterazine for 1-2 h in DMEM supplemented with 1% FBS and then transferred to the perfusion chamber. To reconstitute the erAEQ and GoAEQ with high efficiency, the luminal [Ca 2ϩ ] of the ER and Golgi first must be reduced. This was achieved by incubating the cells for 1 h at 4°C in KRB supplemented with 5 M coelenterazine, the Ca 2ϩ ionophore ionomycin, and 600 M EGTA. After this incubation, cells were extensively washed with KRB supplemented with 2% bovine serum albumin. The coverslip with transfected cells was placed in a perfused, thermostated chamber located in the close proximity of a low noise photomultiplier, with built-in amplifier-discriminator. All aequorin measurements were carried out in KRB supplemented with Cells were loaded with Fura-2/AM, GFP-positive cells were identified, the fluorescence at 340 and 380 nm was analyzed, and the ratio 340/380 nm was calculated. B, the amount of thapsigargin-releasable Ca 2ϩ . The average peak increase in the thapsigargin-induced 340/380 nm ratio is shown (i.e. the increase in the 340/380 nm ratio relative to the basal 340/380 nm ratio in the Ca 2ϩ -free medium that was recorded just before the addition of thapsigargin; see panel C). C, representative traces showing the peak increases in thapsigargin response of a control cell, a GFP expressing cell, and three 2B-GFP expressing cells. D, average peak increase in the thapsigargin-induced 340/380-nm ratio plotted against the initial 340/380-nm ratio of cells in the presence of extracellular Ca 2ϩ . Note that the average amount of thapsigargin-releasable Ca 2ϩ of 2B-GFP-expressing cells that exhibited an initial 340/380-nm ratio of 0.6 -0.9 (which is similar to that of the control cells) was significantly lower than that of control cells. In each experiment, the 340/380-nm ratio of a large number of cells from different coverslips was analyzed. The average Ϯ S.D. of five independent experiments is shown. *, p Ͻ 0.05; **, p Ͻ 0.01. type and mutant aequorins (30,33).
Cell Death Analysis-To assay the antiapoptotic effects in coxsackievirus-infected cells, cells grown on coverslips were either mock-infected or infected with coxsackievirus (30 min at room temperature) at a multiplicity of infection of 50. At the indicated times (specified in the relevant figure legends), cells were challenged with guanidine hydrochloride (GuHCl) (2 mM), cycloheximide (100 g/ml), or actinomycin D (0.5 g/ml) for the indicated times. For the analysis of the nuclear morphology, cells were fixed with 4% paraformaldehyde, stained with DAPI (10 g/ml), viewed under an Axiovert epifluorescence-inverted microscope (Carl-Zeiss GmbH), and imaged with a Nikon Coolpix 995 digital camera equipped with an MDC lens. Caspase activation was analyzed using the Apo-ONE TM homogeneous caspase-3/7 assay according to the instructions of the manufacturer (Promega). Analysis of DNA fragmentation was performed as described by Tolskaya et al. (4). Incubations were performed in serum-free DMEM except for the experiments in which caspase-3 activation was assayed (because of the high caspase-3 background in serum-free DMEM in control cells and infected cells).
To investigate the antiapoptotic effects in 2B-expressing cells, cells grown on coverslips were transfected either with 2B-GFP or mt-GFP. At 36 h post-transfection, cells were treated with the apoptosis-inducing drugs for 4 h (for analysis of caspase-3 activation) or 16 h (for analysis of cell survival). Incubation with cycloheximide, actinomycin D, and etoposide was performed in DMEM containing 10% FBS. Incubation with ceramide was performed in KRB buffer supplemented with 1 mM CaCl 2 . To assay caspase-3 activation, cells were fixed and stained as described (17) with an anti-cleaved caspase-3 (Asp-175) antibody (Cell Signaling Technology), which specifically recognizes active caspase-3 but not its inactive zymogen, and a Texas Red-conjugated goat antirabbit antibody. The percentage of caspase-3-positive GFP-expressing cells was determined de visu by counting at the microscope. To assay cell survival, cells were washed thoroughly to eliminate apoptotic cells after incubation with the apoptotic drugs. The percentage of fluorescent cells before and after a challenge with the apoptotic drugs was determined de visu by counting at the microscope.

RESULTS
Ca 2ϩ Homeostasis in 2B-expressing Cells-To investigate the effects of 2B on subcellular Ca 2ϩ homeostasis, we made use of a fusion protein of 2B and GFP. We have shown previously that the fusion of EGFP at the C terminus of the coxsackievirus 2B protein does not interfere with its membrane-active function and localization (17). Because the EGFP fluorescence heavily contaminates that of the fluorescent Ca 2ϩ indicator Fura-2 (data not shown) (34), we made use of GFP (S65T), a nonenhanced GFP protein that exhibits 30-fold reduced fluorescence intensity relative to EGFP. Fig. 1A shows the [Ca 2ϩ ] cyt of nontransfected control cells and cells expressing either GFP (S65T) or 2B-GFP (S65T). The [Ca 2ϩ ] cyt in GFP-expressing cells was slightly lower than that of control cells (suggesting that there is some minor contamination of the Fura-2 fluorescence by GFP). Expression of 2B-GFP, however, resulted in a significant increase in [Ca 2ϩ ] cyt (p Ͻ 0.01), shown previously to depend on extracellular Ca 2ϩ (15).
Next, we addressed the relationship between the [Ca 2ϩ ] cyt and the [Ca 2ϩ ] of the intracellular stores (i.e. ER and Golgi) using thapsigargin, a specific inhibitor of the sarco/endoplasmic reticulum Ca 2ϩ ATPase (SERCA) pump. The activity of the SERCA is required to compensate for the continuous leakage of Ca 2ϩ that takes place from the stores under normal conditions. Upon inhibition of the SERCA, the Ca 2ϩ that leaks from the stores is not resequestered and accumulates in the cytosol. The size of the thapsigargin-induced transient increase in [Ca 2ϩ ] cyt reflects the [Ca 2ϩ ] of the stores. Fig. 1B shows that the amount of thapsigargin-releasable Ca 2ϩ in cells expressing GFP was similar to that observed in nontransfected control cells. Cells expressing 2B-GFP, however, exhibited a significant decrease (p Ͻ 0.01) in the amount of thapsigargin-releasable Ca 2ϩ . i.e. the higher the initial [Ca 2ϩ ] cyt , the lower the amount of Ca 2ϩ that can be released from the stores (Fig. 1, C and D). Fig.  1D also demonstrates that the filling state of the thapsigarginsensitive stores of 2B-GFP-expressing cells was always lower than that of control cells, even when the initial [Ca 2ϩ ] cyt was similar. Taken together, these results show that the 2B-induced reduction in the [Ca 2ϩ ] of the stores precedes the increased influx of extracellular Ca 2ϩ , providing evidence that the stores are the primary target of 2B.
Reduced  (Fig. 2). In 2B-GFP-expressing cells, lower steady state levels (ϳ30 -40% reduction) were observed in both compartments (ϳ255 and ϳ215 M in the ER and Golgi, respectively). The addition of histamine to the cells resulted in a rapid decrease in [Ca 2ϩ ] in both organelles, confirming that the sensitivity to agonists of the two Ca 2ϩ stores was retained, although a smaller amount of Ca 2ϩ could be released in 2Bexpressing cells. Taken together, these data indicate that 2B is responsible for a reduction of the luminal [Ca 2ϩ ] in both ER and Golgi.
Mitochondrial and Cytosolic Ca 2ϩ Handling in 2B-expressing Cells-Mitochondria play an important role in intracellular Ca 2ϩ homeostasis. The mitochondria are localized in close proximity of inositol 1,4,5-trisphosphate (IP3)-gated channels and are capable of taking up the Ca 2ϩ that is released by IP3-generating agonists, thereby buffering the [Ca 2ϩ ] cyt (35). We hypothesized that the 2B-induced reduction in the steady state [Ca 2ϩ ] ER and [Ca 2ϩ ] Golgi levels and the ensuing reduction of IP3-induced Ca 2ϩ release should decrease the uptake of Ca 2ϩ by the mitochondria. To test this hypothesis, HeLa cells were transfected with mtAEQ and either 2B-GFP or mt-GFP and then challenged with histamine (in the presence of extracellular Ca 2ϩ ). Fig. 3A shows that the peak mitochondrial response is markedly reduced (almost ϳ35%) in 2B-GFP-expressing cells compared with the control cells. Similar results were obtained when cells were challenged with ATP, an agonist that drives production of IP3 through its action on a Gq-coupled P2Y receptor, or when CHO cells were used as model system (data not shown). Taken together, these data indicate that the 2B-induced reduction of [Ca 2ϩ ] in the stores leads to a reduction in the stimulus-induced mitochondrial Ca 2ϩ uptake.
The [Ca 2ϩ ] Golgi levels should also be reflected in a decrease in the rise of the [Ca 2ϩ ] cyt upon stimulation with IP3-generating agonists. To investigate this supposition, HeLa cells were cotransfected with cytAEQ and 2B-GFP or mt-GFP and then challenged with histamine (in the presence of extracellular Ca 2ϩ ). Surprisingly, we found no difference in the amplitude of [Ca 2ϩ ] cyt between 2B-GFP-expressing cells and control cells (Fig. 3B). On the contrary, upon releasing Ca 2ϩ from the stores by a passive process (i.e. ionomycin treatment in Ca 2ϩ -free conditions), a significant reduction in 2B-GFP-expressing cells was observed (data not shown). A likely explanation for the results shown in Fig. 3B is that a reduced amount of Ca 2ϩ is released by histamine in 2B-GFP-expressing cells but that this effect is masked by an increased influx of Ca 2ϩ from the extracellular medium due to an additional effect of 2B. To find support for this hypothesis, we evaluated the relative efficiency of the two pathways in HeLa and CHO cells for Ca 2ϩ increase in the cytosol; i.e. the Ca 2ϩ released from stores and Ca 2ϩ influx from the extracellular medium. In these experiments, the cells were first challenged with histamine in HeLa (or ATP in CHO) in Ca 2ϩ -free medium; under those conditions the increase of the [Ca 2ϩ ] cyt will be due only to the release of Ca 2ϩ from the stores. The following re-addition of Ca 2ϩ to the extracellular medium causes a second [Ca 2ϩ ] cyt rise due to the influx through the plasma membrane channels. Fig. 3C demonstrates that an increased influx of Ca 2ϩ in 2B-expressing cells indeed equilibrates the reduced amount of Ca 2ϩ that is released from the stores in these cells as compared with control cells.

Coxsackievirus Infection Protects against Cycloheximide and Actinomycin D-induced Apoptosis-Poliovirus infection of HeLa cells under permissive conditions results in CPE,
whereas infection under nonpermissive conditions (e.g. in the presence of GuHCl, an inhibitor of viral replication) results in apoptotic cell death (4). Moreover, productive poliovirus infec-tion has been shown to confer an antiapoptotic state to HeLa cells that protects them against apoptosis induced by actinomycin D and cycloheximide (at concentrations of 0.5 g/ml and 100 g/ml, respectively) (4). We tested whether these features are conserved in coxsackievirus. HeLa cells were infected with coxsackievirus in the absence or presence of GuHCl, and the type of cell death that was induced was assayed by examining the nuclear morphology. Fig. 4A shows that the productive coxsackievirus infection of HeLa cells (i.e. in the absence of GuHCl) resulted in the canonical CPE (crescent-shaped nuclei with condensed chromatin), whereas under nonpermissive conditions (i.e. in the presence of absence of GuHCl) a typical nuclear apoptotic response, i.e. condensation of chromatin and fragmentation into apoptotic bodies, was observed in a large portion of the cells, similarly as described for poliovirus (4). To test whether coxsackievirus also initiates an antiapoptotic program, mock-infected and coxsackievirus-infected cells were challenged with actinomycin D and cycloheximide in the concentrations described above. Fig. 4B shows that treatment of mock-infected cells with these drugs resulted in a typical nuclear apoptotic response. Upon the addition of these drugs to coxsackievirus-infected cells, however, no signs of apoptosis were observed, but signs typical of CPE were manifested instead. Together, these findings suggest that the antiapoptotic function is conserved in coxsackievirus.
To further demonstrate the existence of a viral antiapoptotic function, we monitored the ability of coxsackievirus to suppress caspase activation. To this end, HeLa cells that were either mock-infected or coxsackievirus-infected were treated with actinomycin D. Caspase activation was monitored both by incubating HeLa cell lysates with a caspase-3 fluorescent substrate and by analyzing internucleosomal DNA degradation ("laddering"). Fig. 5 shows that both caspase activation and DNA laddering are strongly reduced in coxsackievirus-infected cells. These findings provide further evidence for the existence of an antiapoptotic function in coxsackievirus.
Expression of 2B Protects against Cycloheximide and Actinomycin D-induced Apoptosis-Next, the possible role of the coxsackievirus 2B-induced alterations in Ca 2ϩ homeostasis in the suppression of apoptosis was investigated. For this investigation the following protocol was used. Cells were transfected with 2B-GFP, and the number of fluorescent cells was counted at the microscope before and after a challenge with an apoptotic drug. The rationale, as reported previously for other proteins such as Bcl-2 (36) and VDAC (37), is that if the transfected protein increases the sensitivity to apoptotic agents, the number of fluorescent cells will be lower after the apoptotic challenge (because more transfected cells die), whereas the opposite will happen if the protein has antiapoptotic activity.
To investigate whether 2B is involved in conferring an antiapoptotic state to infected cells, 2B-GFP-expressing cells and mt-GFP-expressing cells were challenged with actinomycin D and cycloheximide at 36 h posttransfection. The results obtained with actinomycin D are shown in Fig. 6A. When mt-GFP was transfected, ϳ40 -45% of the cells showed GFP fluorescence. 16 h after the addition of actinomycin D, the number of viable cells was drastically reduced, but the fraction of fluorescent cells remained the same. Conversely, when 2B-GFP was transfected, the same fraction of fluorescent cells was identified before actinomycin D treatment, but it markedly increased 16 h after the addition of the apoptotic agent (ϳ70%). These data indicate that the 2B-induced alterations in Ca 2ϩ homeostasis confer protection against actinomycin D-induced apoptotic cell death (p Ͻ 0.005). Enhanced survival was also observed upon treatment with cycloheximide (p Ͻ 0.01) (Fig. 6B).
To test whether the expression of 2B also protects against other apoptotic stimuli, we tested the possible suppressive ef-fect against apoptosis induced by 10 M ceramide, a lipid signaling mediator that releases Ca 2ϩ from intracellular stores and induces apoptosis via a Bcl-2 sensitive pathway (38), and by 20 M etoposide, a drug that causes DNA damage by inhibiting topoisomerase II, leading to apoptosis through p53 via a Bcl-2-sensitive, Bax-dependent pathway. Again, the number of fluorescent cells before and after a treatment with the apoptotic drugs was determined. No change in the percentage of living fluorescent cells was observed in HeLa cells transfected with mt-GFP. Cells expressing 2B-GFP were efficiently protected against apoptosis induced by ceramide (p Ͻ 0.005) (Fig.  6C) but not by etoposide (Fig. 6D). The failure of 2B to suppress etoposide-induced apoptosis was not due to a lower apoptotic efficacy of this drug (etoposide induced apoptosis in ϳ60% of the cells, similar to what was observed for actinomycin D and cycloheximide; data not shown). Taken together, these data indicate that 2B protects against some, but not all, Bcl-2sensitive apoptotic pathways.
To demonstrate that the cytoprotective effect of 2B is indeed due to a reduced activation of caspases, cells expressing either 2B-GFP or mt-GFP were treated either with actinomycin D, cycloheximide, or etoposide and then stained with an antibody that specifically recognizes the active form of caspase-3 but not its inactive zymogen (Fig. 7). In control cells, the vast majority of cells showed extensive caspase-3 activation upon the addition of each of these apoptotic drugs (similarly as in nontransfected cells; data not shown). In 2B-expressing cells, however, caspase-3 activation induced by actinomycin D and cycloheximide, but not by etoposide, was potently suppressed. These findings are in agreement with the results described in Fig. 6 and provide evidence that the enhanced survival observed in 2B-expressing cells is due to the suppression of caspase-3 activation.
Effects on Intracellular Ca 2ϩ Homeostasis and Apoptosis of 2B Mutants-Our data strongly suggest that the 2B protein lends protection against apoptosis through its ability to manipulate intracellular Ca 2ϩ homeostasis. However, it cannot be excluded that the effects of 2B on Ca 2ϩ homeostasis and its anti-apoptotic ability represent two distinct, unrelated functions. To investigate whether the antiapoptotic activity of 2B is functionally related to its ability to manipulate intracellular Ca 2ϩ fluxes, we characterized two 2B-GFP mutants with mutations in the hydrophobic regions that are implicated in pore formation (17). In the first mutant, the amphipathic character of the ␣-helix formed by the first hydrophobic region (aa 37-54) is disturbed by the substitution of lysine residues 41, 44, and 48 with hydrophobic leucine residues (mutation K41L/K44L/ K48L) (25). In the second mutant, the hydrophobic nature of the second hydrophobic region (aa 63-80) is disturbed by substitution of two hydrophobic residues (isoleucine 64 and valine 66) by polar serine residues (mutation I64S/V66S) (20).
HeLa cells were cotransfected with erAEQ and the indicated 2B-GFP mutants or mt-GFP, and the [Ca 2ϩ ] ER was compared. Fig. 8A shows that both mutants 2B-K41L/K44L/K48L and 2B-I64S/V66S failed to reduce the luminal [Ca 2ϩ ] in the ER. Consistent with this, measurement of the [Ca 2ϩ ] mt upon challenge with histamine in cells transfected with mtAEQ showed that the peak mitochondrial Ca 2ϩ response in cells expressing the mutants was similar to that in control cells (Fig. 8B). These data provide evidence that both mutants are defective in manipulating intracellular Ca 2ϩ fluxes. To investigate the antiapoptotic activity of these mutants, HeLa cells expressing the 2B-GFP mutants were challenged with either actinomycin D or cycloheximide, and caspase-3 activation was analyzed as described above. Fig. 9 shows that the mutants 2B-K41L/K44L/ K48L and 2B-I64S/V66S were unable to suppress either acti- nomycin D-induced or cycloheximide-induced caspase-3 activation. Consistent with this, both mutants also failed to increase cell survival after treatment with these apoptotic drugs (data not shown). Together, these findings demonstrate that the antiapoptotic activity of 2B is functionally linked to its ability to manipulate intracellular Ca 2ϩ homeostasis.

DISCUSSION
Viruses induce a number of alterations in the metabolism and structure of their host cell to ensure efficient reproduction. Some of these alterations can be sensed by the host cell and turn on a defensive apoptotic reaction that is aimed at curtailing virus replication. Many viruses have developed countermeasures to prevent premature abortion of the viral life cycle. Enteroviruses are small cytolytic RNA viruses that cause a necrosis-like type of cell death, called CPE, which is the ultimate result of a complex interplay between apoptosis-inducing and apoptosis-suppressing functions encoded by the enterovirus genome (4). In this report, we provide evidence that the enterovirus 2B protein plays a major role in suppressing apoptotic host cell responses by manipulating intracellular Ca 2ϩ homeostasis.
Using both chemical and genetically encoded Ca 2ϩ indicators, we demonstrated that the expression of 2B results in the following effects: (i) a reduction in the luminal [Ca 2ϩ ] in both ER and Golgi; (ii) a decrease in the amount of Ca 2ϩ that can be released from these organelles using either thapsigargin or physiological, IP3-generating stimuli like histamine and ATP; (iii) a reduction in the stimulus-induced amount of Ca 2ϩ that is taken up by mitochondria; and (iv) an increase in the influx of extracellular Ca 2ϩ , leading to a rise in [Ca 2ϩ ] cyt . The reduction in the Ca 2ϩ -filling state of the stores preceded the increase in the influx of extracellular Ca 2ϩ , indicating that the stores are the primary target of 2B. Similar results were obtained in HeLa cells and CHO cells, indicating that it is unlikely that the effects of 2B are cell type-specific. These data are consistent with the idea that 2B decreases [Ca 2ϩ ] ER and [Ca 2ϩ ] Golgi by increasing the passive leakage of Ca 2ϩ ions from these stores, most likely through the formation of membrane-integral pores, and thereby accounts for a reduction in the amount of releasable Ca 2ϩ and the down-regulation of Ca 2ϩ fluxes between stores and the mitochondria. The increased influx of Ca 2ϩ is most likely due to the increased plasma membrane permeability that is observed in 2B-expressing cells, which allows the passage of ions and normally nonpermeant low molecular weight compounds (11,39,40).
The 2B-induced perturbation of intracellular Ca 2ϩ distribution and signaling was identified as an important component of the enteroviral strategy to suppress infection-limiting apoptotic host cell responses. These apoptotic responses are most likely triggered by the action of the viral proteinases 3C pro and 2A pro , which inhibit cellular transcription and cap-dependent translation, respectively (8,9). Enteroviruses interrupt this apoptotic program by encoding functions that implement an antiapoptotic program (4). The antiapoptotic state that is conferred to infected cells can even suppress nonviral apoptotic stimuli such as cycloheximide and actinomycin D (whose effects resemble those of 3C pro and 2A pro , respectively). In our studies of the physiological relevance of the 2B-induced alterations in intracellular Ca 2ϩ homeostasis, we observed a role for 2B in suppressing apoptosis. The expression of 2B suppressed caspase-3 activation and apoptotic cell death induced by actinomycin D and cycloheximide. Mutants of 2B that were defective in manipulating intracellular Ca 2ϩ fluxes failed to protect against apoptosis. Together, these findings strongly suggest that 2B, by modulating intracellular Ca 2ϩ homeostasis, plays a major role in conferring the antiapoptotic state to infected cells and, thereby, in extending the life span of the host cell.
The enterovirus 2B protein represents one of the first antiapoptotic proteins of small genome RNA viruses. Most of our knowledge about viral apoptosis-suppressing functions comes from the group of large genome DNA viruses (herpesviruses, adenoviruses, and poxviruses), which often encode multiple antiapoptotic proteins, including Bcl-2 homologs, caspase suppressors, and cell cycle and transcription mediators (1, 2). Relatively little is known about apoptosis-suppressing functions in small genome RNA viruses that exhibit a relatively fast replication cycle and have little, if any, genetic capacity to develop individual antiapoptotic functions. It has been suggested that the antiapoptotic activities of RNA viruses may be byproducts of the cellular alterations that are induced by the viral replication proteins to efficiently replicate their RNA genome. Obviously, the ability of 2B to suppress apoptotic host cell responses is not its sole or primary function. Mutations in the hydrophobic domains of 2B, which most likely interfere with pore formation (as shown in this study for the mutants 2B-K41L/K44L/K48L and 2B-I64S/V66S), cause primary defects in viral RNA replication (20,21). The exact function of 2B is as yet unknown. The ability of 2B to form pore-like structures in secretory pathway membranes may be required for the ability of the 2BC precursor to cause the accumulation of se-cretory pathway-derived membrane vesicles at which viral RNA replication takes place (3,41). Thus, the ability of 2B to reduce the Ca 2ϩ filling state of the stores most likely serves two different functions, namely formation of the viral replication complex and suppression of apoptotic host cell responses.
How can the 2B-induced alterations in intracellular Ca 2ϩ signaling lend protection against certain apoptotic stimuli? The finding that 2B failed to suppress etoposide-induced apoptosis indicates that the alterations in Ca 2ϩ homeostasis suppress a specific apoptotic pathway rather than confer a general blockage of apoptosis. Recent studies on the Bcl-2 oncogene have provided evidence for the coexistence of spatially different apoptotic pathways in the same cell, one ER-dependent and one mitochondrion-dependent, which eventually converge at the mitochondria (42,43). Bcl-2 is localized at the outer mitochondrial membrane, the nuclear envelope, and the ER membrane. Most of the results published so far have emphasized the importance of Bcl-2 at the mitochondria, where it antagonizes the ability of pro-apoptotic Bcl-2 family members to induce cytochrome C release (44). The function of Bcl-2 at the ER membrane is less clear. An ER-restricted Bcl-2 mutant suppressed apoptosis induced by various stimuli (including ceramide) but failed to protect against etoposide, a drug that acts directly at the mitochondria by causing membrane translocation of Bax (42,43). The antiapoptotic activity of ER-restricted Bcl-2 argues for the existence of a signaling mechanism between the ER and the mitochondria. Growing evidence indicates that alterations in Ca 2ϩ fluxes between these organelles have a modulatory effect on apoptosis (13). Key events occurring in the mitochondrial matrix such as ATP production, an important source of reactive oxygen species production, and possibly also the opening of the permeability transition pore leading to swelling of the outer mitochondrial membrane and the release of proapoptotic proteins like cytochrome C, are sensitive to increases in [Ca 2ϩ ] mt (13,45). Accordingly, down-regulation of Ca 2ϩ fluxes between the ER and the mitochondria can protect the mitochondria from cytotoxic rises in [Ca 2ϩ ] mt . Increasing evidence indicates that Bcl-2 exerts some of its antiapoptotic effects from the ER by reducing [Ca 2ϩ ] ER and down-regulating Ca 2ϩ fluxes between the ER and mitochondria (36,46,47), possibly through the ability of Bcl-2 to form ion channels in the ER membrane (48). Moreover, conditions that lowered [Ca 2ϩ ] ER protected HeLa cells from ceramide-induced apoptosis, whereas conditions that increased [Ca 2ϩ ] ER had the opposite effect (38). The findings that 2B down-regulated Ca 2ϩ fluxes between the stores and the mitochondria and suppressed apoptosis induced by ceramide, but not etoposide, strongly suggest that 2B specifically targets a Ca 2ϩ -sensitive, ER-dependent apoptotic pathway.
Is the ability of 2B to decrease Ca 2ϩ signaling between the ER and mitochondria sufficient for conferring an antiapoptotic state to enterovirus-infected cells? In both poliovirus-infected cells (4, 6) and coxsackievirus-infected cells (this study), the antiapoptotic function is expressed early in the infection (i.e. at ϳ2 h postinfection). Previously, we demonstrated that coxsackievirus infection results in a ϳ50% reduction of the Ca 2ϩ content of the thapsigargin-sensitive stores within the first 2 h of infection (15). Thus, the virus-induced alterations in intracellular Ca 2ϩ signaling coincide with the implementation of the anti-apoptotic state, lending support to the idea that these events are causally linked. It cannot be excluded that the rise in [Ca 2ϩ ] cyt also contributes somehow to the suppression of apoptosis in infected cells. A rise in [Ca 2ϩ ] cyt may lead to the activation of calpains, Ca 2ϩ -sensitive proteases that can exert antiapoptotic effects by cleaving caspases, including key caspases 8 and 9 (49,50). Interestingly, aberrant processing of caspase-9 (albeit at a low level) was observed in HeLa cells upon productive poliovirus infection but not upon abortive infection (51). However, aberrant processing of caspase-9 was not reported upon productive coxsackievirus infection (52). Our observation that the 2B-induced alterations in Ca 2ϩ homeostasis failed to protect against etoposide-induced caspase-9-de-pendent apoptosis suggests that it is unlikely that aberrant processing of caspase-9 can completely account for the antiapoptotic activity of 2B. Moreover, significant rises in [Ca 2ϩ ] cyt of coxsackievirus-infected cells take place only after 5 h. Therefore, we propose that the 2B-induced reduction of the Ca 2ϩ content of the ER and Golgi and the resulting down-regulation of Ca 2ϩ fluxes between these stores and the mitochondria is the major component of the viral antiapoptotic program. It should be kept in mind, however, that signs of apoptosis (cytochrome C release and caspase activation) become apparent in enterovirus-infected cells (7,51,52) later in the infection, and it has been suggested that this may contribute to the killing of the host cell and the release of virus progeny (52). Thus, the antiapoptotic activity of the 2B protein most likely serves to delay apoptotic responses, thereby providing the virus the time required for genome replication, rather than to completely prevent all signs of apoptosis.
Further research is required to define the antiapoptotic mechanism of 2B and the role of Ca 2ϩ therein. A better understanding of the molecular mechanism used by enteroviruses to manipulate cell death may lead to a better insight into the pathogenesis of viral disease and may contribute to our understanding of the critical role of Ca 2ϩ signaling in the modulation of apoptosis.