Endoplasmic reticulum stress-induced programmed cell death in soybean cells

Programmed cell death (PCD) plays an essential role in many aspects of animal and plant development and in response to stresses (Beers and McDowell, 2001; Katoch et al., 2002). The main signaling pathways of PCD have been described mainly in animal cells. In plants, PCD has been observed during tracheary elements differentiation (Kuriyama and Fukuda, 2002), aerenchyma formation (Gunawardena et al., 2001), hypersensitive response (HR) (Heath, 2000) and as a consequence of several biological and chemical stresses (Beers and McDowell, 2001). Although there are few reports concerning the mechanisms of PCD in plant cells, recent studies have shown several morphological and biochemical similarities between apoptosis and plant PCD (Danon et al., 2000). In particular, similar to animal apoptosis, the involvement of calcium ions, production of reactive oxygen species, activation of caspase-like proteases, cytochrome c release from mitochondria, cytoplasmic shrinkage, chromatin condensation and DNA fragmentation have been observed to occur during plant PCD (Levine et al., 1996; Danon et al., 2000; McCabe and Leaver, 2000; Neill et al., 2002). On the whole, these data, even though not demonstrated simultaneously in the same plant system, suggest that PCD in plants and animals may be based on a common cell death process (Hoeberichts and Woltering, 2003).

The main PCD signaling pathways illustrated in animal cells involve plasma membrane receptors and mitochondria. Only recently has the importance of the endoplasmic reticulum (ER) in triggering a specific program of cell death been recognized (Nakagawa et al., 2000; Nakagawa and Yuan, 2000). The ER is the major compartment for Ca²⁺ sequestration and signaling (Meldolesi and Pozzan, 1998). The high lumenal Ca²⁺ concentration acts as a pool for releasable Ca²⁺ during cell signaling (Berridge et al., 2000) and also regulates the functioning of numerous ER lumenal proteins (Brostrom and Brostrom, 1998). The main role of the ER, regulated by a number of Ca²⁺-dependent proteins, is the control of synthesis, folding and export of proteins (Meldolesi and Pozzan, 1998). Prolonged ER Ca²⁺ depletion impaired ER functions causing the so called ER stress (Pahl, 1999). ER stress in animal cells has received increased attention for its involvement in pathologically relevant apoptosis (Aridor and Balch, 1999; Paschen and Doutheil, 1999). Stress in the ER has been shown to induce different apoptotic pathways involving a cross-talk between ER and mitochondria, the activation of caspase 12 and/or cytochrome c release by mitochondria, causing caspase 9 activation (Hacki et al., 2000; Zuppini et al., 2002; Jimbo et al., 2003). Procaspase 12 resides in the cytosolic side of the ER membrane and, upon its activation by ER stress, is released into the cytosol in an active form (caspase 12) (Nakagawa et al., 2000; Nakagawa and Yuan, 2000). The ER apoptotic pathway seems to be triggered by a specific ER stress: caspase 12 knockout mice undergo apoptosis induced by stimuli that do not involve ER stress, whereas they are resistant to ER stress-induced apoptosis (Nakagawa and Yuan, 2000).

Several studies have suggested that many stimuli that alter the cytosolic Ca²⁺ concentration ([Ca²⁺]_cyt) and the storage of Ca²⁺ in the intracellular organelles induce apoptosis in a variety of cells (Arnaudeau et al., 2002; Grebenová et al., 2003). Molecules that inhibit the sarcoplasmic/ER Ca²⁺-ATPase (SERCA) such as thapsigargin (TG) and cyclopiazonic acid (CPA) cause depletion of intracellular Ca²⁺ stores with consequent increase in [Ca²⁺]_cyt and induce apoptosis (Thastrup et al., 1990; Mason et al., 1991; Nguyen et al., 2002). The data so far available on ER stress-induced apoptosis by TG or CPA are from animal cells. There is no evidence for a similar mechanism in plant cells.

We have used a soybean cell line stably expressing the Ca²⁺-sensitive photoprotein aequorin, to analyze the pathway of cell death induced by a definite ER stress. Although the ER has a smaller Ca²⁺ storage capacity than the vacuole, its importance in Ca²⁺ release during signaling has recently been suggested (Navazio et al., 2000; Navazio et al., 2001). In plant cells two distinct subtypes of Ca²⁺-ATPases have been found in the ER: ECA1, a type IIA ER-type Ca²⁺/Mn²⁺ pump, which is sensitive to CPA (Sze et al., 2000; Pittman and Hirschi, 2003); and ACA2, a CaM-regulated type IIB Ca²⁺ pump, which is insensitive to CPA (Harper et al., 1998; Pittman and Hirschi, 2003). We investigated the role of CPA, a specific blocker of plant ER-type IIA Ca²⁺ pumps (Sze et al., 2000), in triggering cell death. We found that inhibition of ER Ca²⁺-ATPase by CPA causes PCD involving a [Ca²⁺]_cyt increase. Our findings indicate that in soybean cells CPA induces an ER stress, which triggers a PCD program involving cytochrome c release from mitochondria and caspase-like activities. The participation of both ER and mitochondria suggests the existence of a cross-talk between these two compartments to achieve the program of cell death.

CPA, 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (acetoxymethil ester) (BAPTA-AM) and bafilomycin A₁ were obtained from Sigma-Aldrich (St Louis, USA). Coelenterazine was purchased from Molecular Probes Inc. (Eugene, OR).

Soybean (Glicine max L.) cell-suspension culture, transformed with the Ca²⁺-sensitive photoprotein aequorin (Mithöfer et al., 1999; kindly provided by G. Neuhaus, Freiburg, Germany), were grown in a Murashige-Skoog liquid medium supplemented with 5% sucrose, 10 μg/ml naphthlaleneacetic acid, 2 μg/ml kinetin, 1 μg/ml thiamine-HCl and 10 μg/ml kanamycin. Cells were subcultured every 3 weeks by making a 1:10 dilution in 40 ml of fresh medium and maintained at 24°C with a 16-hour photoperiod with constant shaking. Treatments with CPA were performed 3 weeks after reinoculation, during the exponential growth phase of the cells.

Cells were harvested by centrifugation (5 minutes, 1600 rpm) and solubilized with a solution containing 0.1 M Tris, 0.2 M NaCl, 0.2% Triton X-100, 1 mM EDTA, 1 μM leupeptin and 0.5 mM phenylmethylsulphonylfluoride (PMFS). Protein extracts were separated by SDS-PAGE (10% polyacrylamide) (Laemmli, 1970) and transferred to Immobilon-P transfer membrane (Millipore, Bedford, MA) for chemiluminescence detection, according to the method of Towbin et al. (Towbin et al., 1979). Filters were incubated with 5% skimmed milk in TBS for 1 hour and immunoblotted with the specific antibodies for 2 hours. Membranes were probed with polyclonal antisera raised against the tobacco binding protein (BiP) (1:5000, from A. Vitale, Milano, Italy) and spinach calreticulin (CRT) (1:2000, Navazio et al., 1995). Moreover, a polyclonal antibody against murine caspase 12 was used (1:5000, Oncogene, Boston, MA). After incubation, filters were washed twice with 0.05% Tween/TBS followed by one wash in TBS before incubation with the secondary antibody. Labeled proteins were detected using CDP-star™ (Biolabs, Hitchin, UK) and autoradiography films (Sigma-Aldrich, St Louis, USA). The exposed films were quantified by densitometric analysis using Quantity One and Molecular Analyst software (BioRad, Oakland, CA).

Protein content in extracts was determined by the method of Bradford (Bradford, 1976) using bovine serum albumin as standard.

Reconstitution of the Ca²⁺-sensitive photoprotein aequorin was performed in vivo by incubating soybean cells with 5 μM coelenterazine overnight in darkness. Treatments with CPA were carried out by injecting 50 μl of twofold concentrated CPA solution (dissolved in the basal cell culture medium) into the same volume of reconstituted cell-suspension culture (about 3 mg fresh weight) using a light-tight syringe. For experiments carried out in the absence of extracellular Ca²⁺, cells were extensively (10 vol) washed three times with Ca²⁺-free culture medium and then resuspended in the same medium containing 100 μM EGTA. The residual aequorin was completely discharged by adding 0.33 M CaCl₂ in 10% ethanol. Luminescence data were collected and converted into [Ca²⁺]_cyt by a computer algorithm based on the Ca²⁺ response curve of aequorin (Brini et al., 1995).

The viability of cells was determined by incubation with 0.05% Evans Blue (Levine et al., 1996). Cells were then extensively washed with deionized water to remove excess and unbound dye. Dye bound to dead cells was solubilized in 50% methanol/1% SDS for 30 minutes at 50°C and quantified by absorbance at 600 nm. CPA was applied dissolved in dimethyl sulfoxide (DMSO, final solvent concentration, 0.1% v/v). DMSO had no effect on the viability of cells. The caspase inhibitors Ac-DEVD-CHO and Ac-LEHD-CHO (BioSource International, Inc., Camarillo, CA, USA) were dissolved in DMSO and applied either to cell suspension together with CPA or to cytosolic protein extracts.

Hydrogen peroxide (H₂O₂) release by soybean cells was measured as described by Wolff (Wolff, 1994). Cells were treated and harvested at different time points to measure H₂O₂ release. The assay is based on a colorimetric reaction caused by the peroxide-mediated oxidation of Fe²⁺ followed by the reaction of Fe³⁺ with xylenol orange. 0.5 ml assay solution (0.5 mM ammonium ferrous sulfate, 50 mM H₂SO₄, 0.2 mM xylenol orange, 200 mM sorbitol) was added to 0.5 ml of control and treated soybean cells and the absorbance at 560 nm was detected after 45 minutes incubation.

Cytosolic protein extracts were prepared by incubating cells (10 minutes on ice) with 75 mM NaCl, 1 mM NaH₂PO₄, 8 mM Na₂HPO₄, 250 mM sucrose and 190 μg/ml digitonin. Cytosolic fractions were then separated by centrifugation (5 minutes, 13,000 g, 4°C), loaded (10 μg) onto a 12% acrylamide gel, transferred to Immobilon-P transfer membrane (Millipore, Bedford, MA) and probed with a rabbit polyclonal cytochrome c antibody (1 μg/ml dilution, Oncogene, Boston, MA) as described previously.

Caspase 3 and caspase 9 activities were measured using the `caspase 3 colorimetric activity assay kit' and the `caspase 9 colorimentric activity assay kit' (Chemicon International, Inc., Temecula, CA), respectively, as described by the manufacturer. The assays are based on spectrophotometric detection of the chromophore p-nitroaniline (pNA) after the caspase-dependent cleavage from the labeled substrates DEVD-pNA or LEHD-pNA by active caspase 3 or 9, respectively. 15 μg of cytosolic protein extracts from soybean cells were incubated for 2 hours at 37°C with DEVD-pNA or LEHD-pNA, in the presence or absence of 20 μM caspase inhibitors Ac-DEVD-CHO (for caspase 3) or Ac-LEHD-CHO (for caspase 9). Caspase-like activity was determined by spectrophotometric quantification of the free pNA (λ_em=405 nm).

To analyze nuclear morphological changes induced by CPA, cells were stained with the fluorescent dyes Hoechst 33342 (HO) and propidium iodide (PI). Cells were incubated in darkness with 8 μg/ml HO (Sigma-Aldrich, St Louis, USA) and 5 μg/ml PI (Sigma-Aldrich, St. Louis, USA) at room temperature for 10 minutes and observed by fluorescence microscopy using an excitation light of 350 nm and 570 nm, respectively for the two dyes.

Data are presented as mean±s.d. of three independent experiments and the differences between groups were assessed with Student's t-test; statistically significant at P<0.05 (one asterisk) and P<0.005 (two asterisks).

Previous reports indicate that the ER stress response produces an increase in the expression of some ER chaperones, such as BiP (Rao et al., 2001) and CRT (Llewellyn et al., 1996). Moreover caspase 12, the most recently characterized member of the caspase family, is activated only by ER stress inducers (Nakagawa et al., 2000). We investigated whether CPA treatment causes an ER stress by analyzing both the rise in protein level of the ER chaperones BiP and CRT and the activation of ER-resident caspase 12.

As shown in Fig. 1, incubation of cells (24 hours) with CPA (50 μM) increased both BiP [(1.5±0.1)-fold] (Fig. 1A) and CRT [(2.0±0.03)-fold] (Fig. 1B) expression levels, compared with untreated cells. Moreover, although up to now no functional homologs of animal caspases have been identified in plants, by using an antibody against the murine caspase 12 we found a very specific cross-reacting protein band in protein extracts from soybean cells. This protein has an expected molecular mass of about 53 kDa and its expression is augmented by (1.7±0.2)-fold after CPA treatment (Fig. 1C). In addition, the procaspase 12-like seems to be processed after the CPA treatment as indicated by the presence of proteolytic fragments (Fig. 1C) of comparable molecular mass to active animal caspase 12 (Nakagawa et al., 2000). Taken together the data highlight the occurrence of an ER stress in soybean cells caused by CPA treatment.

We investigated whether CPA treatment affects [Ca²⁺]_cyt. Addition of CPA (50 μM) to soybean suspension cells was found to induce, in temporal terms, a biphasic Ca²⁺ signal: an immediate sharp elevation in [Ca²⁺]_cyt, which dissipated within 5 minutes, was followed by a second flattened Ca²⁺ increase (Fig. 2A). The first Ca²⁺ rise peaked at 0.44±0.05 μM 100 seconds after CPA administration, whereas the second Ca²⁺ elevation reached 0.30±0.03 μM after 10 minutes and did not return to basal level in the time evaluated (15 minutes).

To analyze whether extracellular Ca²⁺ influx contributed to CPA-induced Ca²⁺ changes, cells were extensively washed and resuspended in Ca²⁺-free medium supplemented with 100 μM EGTA to chelate any residual Ca²⁺ ions. Removal of the external Ca²⁺ pools did not markedly affect the biphasic response (Fig. 2A). The observed limited reduction of the first Ca²⁺ peak (from 0.44 to 0.33 μM) is probably due to a partial depletion of intracellular Ca²⁺ stores following the EGTA treatment (Cessna and Low, 2001), rather than a participation of extracellular Ca²⁺ in the Ca²⁺ rise induced by CPA. Thus, the increase in [Ca²⁺]_cyt seems to be mainly due to Ca²⁺ release from intracellular stores.

Addition of 200 nM bafilomycin A₁, a highly specific inhibitor of vacuolar-type H⁺-ATPase inducing a reduction of the storage capacity of the vacuole (Takahashi et al., 1997), resulted in the complete abolition of the second Ca²⁺ peak (Fig. 2B). This suggests the possibility that Ca²⁺ release from the vacuole may account for the second Ca²⁺ transient.

Soybean suspension cells were treated with 50 μM CPA and cell viability was assessed with Evans Blue staining (Levine et al., 1996), as lethally damaged cells are unable to exclude the dye. Exposure of cells for 24 hours to CPA significantly decreased cell viability: CPA induces a rise in cell death of about (1.7±0.08)-fold compared to untreated cells (Fig. 3). Thus a continuous exposure of cells to CPA triggers cell death within 24 hours.

Treatment of cells with the intracellular Ca²⁺ chelator BAPTA-AM (10 μM for 1 hour) prior to CPA treatment (24 hours), although causing an effective chelation of [Ca²⁺]_cyt (data not shown), did not produce any significant change in the percentage of cell death compared with CPA-treated cells (Fig. 3).

To examine the events that link ER stress to cell death in soybean cells challenged with CPA we monitored some of the recognized hallmarks of PCD such as the accumulation of H₂O₂, cytochrome c release, caspase-like activity and chromatin condensation.

A tightly controlled generation of reactive oxygen species is crucial in the apoptotic pathway (Jabs, 1999). In particular, H₂O₂ functions as a signaling molecule, which could drive PCD induced by various stimuli in plant cells (Beers and McDowell, 2001). Treatment of soybean cells with CPA resulted in a transient accumulation of H₂O₂ in the culture medium. The H₂O₂ concentration started to increase rapidly within 2 minutes after the treatment, reaching a maximum (169 μM ±1.66 μM) at about 25 minutes, and declined more slowly without getting back to the baseline level within the considered time lapse (Fig. 4). Untreated cells show a baseline production of H₂O₂ of less than 16 μM in the medium (Fig. 4).

Mitochondria play a central role in the initiation of apoptosis in animal cells by releasing cytochrome c into the cytosol with the resulting activation of caspase 9 (Desagher and Martinou, 2000). We evaluated the release of cytochrome c on cytosolic protein extracts by western blot analysis. Fig. 5 shows CPA-induced release of cytochrome c in soybean cells. Cytosolic extracts from untreated cells do not contain any detectable cytochrome c (Fig. 5). The presence of caspase 9-like activity was monitored using the synthetic substrate LEHD conjugate with the chromophore pNA. We found that cytosolic extracts from CPA-treated cells show about twofold increase in caspase 9-like activity compared with the control (Fig. 6). In addition, application of the caspase 9 inhibitor Ac-LEHD-CHO (20 μM) together with CPA, blocked the protease activation (Fig. 6).

The release of cytochrome c from mitochondria leads to the formation of a cytoplasmic complex with Apaf-1 and the caspase 9 protease, which in turn promotes a cascade of caspase activation, including caspase 3 (Cain et al., 2002). In our study a synthetic tetrapeptide sequence (DEVD), based on the caspase 3 cleavage site and conjugated to pNA, was used for the assay of caspase 3-like protease activity in soybean cells. As shown in Fig. 7, caspase 3-like activity was significantly activated after 24 hours of incubation with CPA. Cytosolic fractions of CPA-treated soybean cells show about twofold increase in caspase 3-like activity compared to untreated cells (Fig. 7); moreover, this activity was blocked in the presence of 20 μM of the caspase 3 inhibitor Ac-DEVD-CHO (Fig. 7).

In an attempt to further define the involvement of caspase-like activities in CPA-induced cell death, the same tetrapeptide inhibitors of caspase 9 and caspase 3 (20 μM) were administered to cell cultures 30 minutes prior to CPA treatment. Inhibition of caspases failed to block CPA-induced cell death, as tested with the Evans Blue cell viability assay (data not shown). Similar results have been previously obtained in different animal experimental systems, and interpreted as a redirection of cells towards necrosis when the apoptotic pathway is blocked by caspase inhibitors (Vercammen et al., 1998a; Vercammen et al., 1998b; Hetz et al., 2002).

To further explore the nature of CPA-induced cell death, we evaluated the cell morphological changes by light microscopy and double staining of the nuclei with the fluorophores HO and PI. Untreated cells show a HO fluorescence homogenously distributed in the cell nuclei without chromatin condensation (Fig. 8B) and a negative PI staining (Fig. 8C). This staining pattern occurs in about 95% of the control cells analyzed. The nuclei of CPA-treated (24 hours) cells exhibit condensed chromatin with regions of nonhomogeneous HO fluorescence (Fig. 8E,H), a staining pattern consistent with apoptosis in animal cells (Maruyama et al., 2000). The presence of both early (HO⁺ and PI^– nuclei) and late (HO⁺ and PI⁺ nuclei) PCD stages is observable in the same treated cell population (Fig. 8E,F,H,I) probably because of its asinchronicity. Fig. 8G shows the detachment of the plasma membrane from the cell wall, which is considered as a hallmark of PCD in plant cells (Beers and McDowell, 2001). In CPA-treated cells this alteration occurs progressively as the PCD program advances (Fig. 8D,G). This morphology is not present in untreated cells (Fig. 8A). The analysis of genomic DNA extracted from CPA-treated soybean cells by agarose gel did not show the presence of DNA fragmentation within the 24 hours of treatment (data not shown).

In this paper, we demonstrate that in soybean cells CPA determines an ER stress, which in turn induces a PCD program involving a biphasic increase in [Ca²⁺]_cyt, generation of H₂O₂, release of mitochondrial cytochrome c, activation of caspase-like proteases and chromatin condensation.

Accumulating evidence shows that besides its crucial role as an important Ca²⁺ store and in protein maturation, folding and transport, the ER is emerging as a central player in apoptosis (Rizzuto et al., 1998; Berridge, 2002). The data so far available comes mainly from mammalian cells, although the precise involvement of the ER in the apoptotic response has not yet been completely unraveled. In animal cells disturbance of ER functions have been shown to trigger apoptotic pathways involving a cross-talk between ER and mitochondria, which is regulated by Bcl-2 protein, and the activation of ER-resident caspase 12 as mediator of death signaling (Hacki et al., 2000; Nakagawa and Yuan, 2000; Nakagawa et al., 2000). In particular, caspase 12 knockout mice are resistant to ER stress-induced apoptosis while their cells undergo apoptosis in response to stimuli (i.e. staurosporine) that do not involve ER stress (Nakagawa and Yuan, 2000). Compared with animal systems, relatively scarce knowledge is available on the detailed mechanism of PCD in plants. However, some aspects of the molecular machinery of PCD seem to be conserved between plants and animals. So far, the role of the ER in plant PCD has been suggested in sycamore cell cultures treated with tunicamycin and brefeldin A (Crosti et al., 2001). The authors argue that chemicals that interfere with ER functions induce PCD with apoptotic features.

In a wide variety of animal cell types, inhibiting ER Ca²⁺-ATPase activity with TG or structurally unrelated Ca²⁺-ATPase antagonists, such as CPA, induces apoptosis, which is triggered by ER Ca²⁺ depletion (Zhou et al., 1998; Nguyen et al., 2002; Zuppini et al., 2002). We found that in soybean cell suspension cultures CPA induces cell death when applied at 50 μM for 24 hours. Our data suggest that lowering ER Ca²⁺ concentration by CPA treatment is a sufficient signal to initiate a specific PCD program. Incubation of cells with CPA causes an ER stress response, highlighted by an enhanced transcription of CRT and BiP. Both proteins have been shown to increase their expression in response to stimuli that disrupt ER homeostasis.

In animal systems apoptosis is mediated by the key enzymatic activity of a class of specific proteases: the caspases (Nicholson, 1999). Although no direct homologues of caspase genes have been identified in the plant genome so far, caspase-like activities have been reported in PCD induced by different treatments and during HR cell death (del Pozo and Lam, 1998; Sun et al., 1999). Recently, based on homology searches, a family of caspase-related proteases (the metacaspases) has been identified in Arabidopsis (Uren et al., 2000). In the investigations reported here, western blot analysis on protein extracts from soybean cells with an antibody raised against murine caspase 12 showed the presence of a cross-reacting protein band, which seems to increase in its expression after the CPA treatment and to be processed into proteolytic fragments of a molecular weight comparable to mammalian active caspase 12. Although a molecular approach is needed to firmly establish the nature of this protein, our data suggest the involvement of a protein structurally related to animal caspase 12 (one of the most important mediator of ER stress-induced apoptosis) in plant ER-mediated PCD.

The role of cytosolic Ca²⁺ as pro-apoptotic messenger involved in triggering apoptosis and in regulating death-specific enzymes has been ascertained in animal and plant systems (Levine et al., 1996; Nicotera and Orrenius, 1998). In this study we show that inhibition of ER Ca²⁺-ATPase activity after the application of CPA causes, in soybean cells, a biphasic increase in [Ca²⁺]_cyt as a result of the ion release mainly from intracellular stores (ER and vacuole). The participation of the vacuolar compartment in the CPA-induced Ca²⁺ signaling is suggested by the annulment of the second rise in [Ca²⁺]_cyt by bafilomycin A₁. PCD induction seems to be signaled in response to ER Ca²⁺ lowering rather than [Ca²⁺]_cyt elevation. Indeed, inhibiting the increase of cytosolic Ca²⁺ with BAPTA-AM did not prevent CPA-induced cell death.

Mitochondria function as major checkpoint in dictating cell survival or cell death by integrating Ca²⁺ and apoptotic stimuli in animal cells. Mitochondrial Ca²⁺ accumulation seems to trigger the opening of permeability transition pore with the consequent release of apoptogenic proteins (Bernardi et al., 1999). Previous data on endothelial cells have shown an accumulation of Ca²⁺ within mitochondria as a result of CPA treatment (Jornot et al., 1999). Several reports also suggest the key role of mitochondria in plant PCD induced by various stimuli (Lam et al., 1999; Jones, 2000). Particularly, the requirement of Ca²⁺ localization in the mitochondrial matrix seems to be important for the permeability transition pore opening and cytochrome c release (Virolainen et al., 2002). Cytochrome c leakage and activation of caspase-like proteases were detected in our study during CPA-induced PCD, suggesting the involvement of mitochondria in the death pathway. Our results indicate an increase in caspase 9-like activity in cytosolic extracts of soybean cells treated with CPA, which is blocked by the specific inhibitor Ac-LEHD-CHO of animal caspase 9. Furthermore, the presence of caspase 3-like protease activity was detected from the specific cleavage of the caspase 3 substrate in cells treated with CPA. Also in this case, the elevation in the protease activity was successfully prevented by the use of the caspase 3 inhibitor Ac-DEVD-CHO. The existence of a caspase-like proteolytic activity and its essential role in plant PCD have been strongly suggested in several plant systems (Sun et al., 1999; Mlejnek and Prochàzka, 2002). Caspase-like activity has been reported in tobacco tissue undergoing HR (del Pozo and Lam, 1998) and specific inhibitors of animal caspase 1 and 3 prevent menadione-induced PCD in tobacco protoplasts (Sun et al., 1999).

A number of recent studies have tried to prove whether plant PCD induced by various elicitors occurs with morphological and physiological features similar to those observed during apoptotic cell death in animal cells (Mlejnek and Prochàzka, 2002; Hoeberichts and Woltering, 2003). One of such hallmarks is the condensation of chromatin followed by DNA fragmentation into 180 bp fragments or multiples of these (ladders). Damage of DNA with ladder formation has been observed in CPA-induced cell death in animal cells (Zhou et al., 1998). In plants, DNA fragmentation has been reported in several PCD processes (Sun et al., 1999; Mlejnek and Prochàzka, 2002), although plant PCD occurring without laddering has also been observed (Dangl et al., 1996; Fukuda, 2000). We provide evidence that CPA induces chromatin condensation in soybean cells, but no DNA fragmentation into ladders was found within 24 hours of treatment (data not shown). Similar results have been obtained for PCD induced by other stimuli. Nitric oxide for example induces apoptosis in mammalian cells occurring with DNA laddering (Pinsky et al., 1999), whereas in Arabidopsis suspension cultures a condensation of nuclear chromatin without DNA laddering has been observed during nitric oxide-induced PCD (Clarke et al., 2000). Thus, despite the recognized evolutionary conservation of some mechanisms underlying PCD in plants and animals, some features seem to rely on the considered cell type and PCD-inducing stimulus.

In summary, our data show that inhibition of the ER Ca²⁺-ATPase by CPA causes stress in the ER of soybean cells. In response to the induced stress, cells transiently increase [Ca²⁺]_cyt and undergo a PCD program involving mitochondria and caspase-like activities. We propose that the PCD pathway activated by CPA might involve a cross-talk between ER and mitochondria to fulfil the cell death program.

We thank G. Neuhaus (Freiburg, Germany) for kindly providing aequorin-expressing soybean cell cultures and A. Vitale (Milano, Italy) for the kind gift of the tobacco anti-BiP antibody. This work was supported by Progetto Giovani Ricercatori of the University of Padova (to A.Z.) and by FIRB 2002 prot. RBNE01KZE7.