Defects in N-glycosylation induce apoptosis in yeast


*E-mail; Tel. (+49) 941 943 3043; Fax (+49) 941 943 3352.


N-glycosylation in the endoplasmic reticulum is an essential protein modification and highly conserved in evolution from yeast to man. Defects of N-glycosylation in humans lead to congenital disorders. The pivotal step of this pathway is the transfer of the evolutionarily conserved lipid-linked core-oligosaccharide to the nascent polypeptide chain, catalysed by the oligosaccharyltransferase. One of its nine subunits, Ost2, has homology to DAD1, originally characterized in hamster cells as a defender against apoptotic death. Here we show that ost mutants, such as ost2 and wbp1-1, display morphological and biochemical features of apoptosis upon induction of the glycosylation defect. We observe nuclear condensation, DNA fragmentation as well as externalization of phosphatidylserine. We also demonstrate induction of caspase-like activity, both determined by flow cytometric analysis and in cell-free extracts. Similarly, the N-glycosylation inhibitor tunicamycin in combination with elevated temperature is able to challenge the apoptotic cascade. Heterologous expression of anti-apoptotic human Bcl-2 diminishes caspase activation, improves survival of cells and suppresses the temperature-sensitive growth defect of wbp1-1. Furthermore, accumulation of reactive oxygen species occurs in response to defective glycosylation. As deletion of the metacaspase YCA1 does not seem to abrogate glycosylation-induced apoptosis, we postulate a different proteolytic process to be involved in this death pathway.


Apoptosis is a highly co-ordinated form of programmed cell death characterized by distinct morphological and biochemical changes and is crucial for the development, maintenance and disease control of multicellular organisms (Kerr et al., 1972; Kerr, 2002). There is increasing evidence that apoptotic-like cell death pathways also exist in unicellular organisms such as yeast (Madeo et al., 1997; Skulachev, 2002; Burhans et al., 2003; Hardwick and Cheng, 2004; Madeo et al., 2004; Ivanovska and Hardwick, 2005) conferring selective advantage in adapting to adverse environmental conditions and thus ensuring survival of the clone (Herker et al., 2004).

First indication for apoptosis in Saccharomyces cerevisiae was the observation that a mutation in the AAA-ATPase gene CDC48, involved in vesicle fusion associated with secretion and cell division (Latterich et al., 1995), causes cell death exhibiting the typical morphological hallmarks of apoptosis: chromatin condensation, DNA fragmentation and exposure of phosphatidylserine at the outer leaflet of the plasma membrane (Madeo et al., 1997). Since then homologues of CDC48 have been found to regulate also metazoan apoptosis (Shirogane et al., 1999; Wu et al., 1999). Meanwhile various different stress stimuli like H2O2 (Madeo et al., 1999), acetic acid (Ludovico et al., 2001; Giannattasio et al., 2005), exposure to high salinity (Huh et al., 2002; Wadskog et al., 2004) or infection with killer viruses (Ivanovska and Hardwick, 2005; Reiter et al., 2005) were shown to induce apoptosis-like cell death in yeast. In addition, ageing represents a physiological scenario for yeast apoptosis, as altruistic death of aged cells might be beneficial for the cell community (Laun et al., 2001; Fabrizio et al., 2004; Herker et al., 2004).

Recently the caspase-like protein Yca1/Mca1, belonging to a new family of caspases, the so-called metacaspases (Uren et al., 2000), was found to be required for H2O2- and ageing-induced (Madeo et al., 2002; Herker et al., 2004) apoptotic pathways in Saccharomyces cerevisiae. Since then, yeast apoptosis induced by salt stress, killer toxins, misregulation of silencing or mitochondrial fission has been shown to be Yca1-dependent (Bettiga et al., 2004; Fannjiang et al., 2004; Wadskog et al., 2004; Ivanovska and Hardwick, 2005; Reiter et al., 2005). Nevertheless, also metacaspase-independent programmed death seems to occur (Ivanovska and Hardwick, 2005). In addition, ectopic expression of anti-apoptotic and pro-apoptotic proteins of the Bcl-2/Bax-family in yeast prevents or causes cell death, suggesting that certain ancient reaction steps of the apoptotic machinery of higher eukaryotes may be conserved (Greenhalf et al., 1996; Longo et al., 1997; Manon et al., 1997; Tao et al., 1997; Ligr et al., 1998; Matsuyama et al., 1998; Priault et al., 2002).

N-glycosylation is one of the most common types of eukaryotic protein modifications. The pathway is highly conserved from yeast to man and essential for viability and development (Kornfeld and Kornfeld, 1985; Tanner and Lehle, 1987; Knauer and Lehle, 1999; Helenius and Aebi, 2001). The key step of this pathway is the transfer of the evolutionarily conserved core-oligosaccharide GlcNAc2Man9Glc3 linked to dolichyl pyrophosphate to selected asparagine residues of the nascent polypeptide chain, catalysed by the hetero-oligomeric membrane complex oligosaccharyltransferase (OST). Biochemical and molecular genetic studies in yeast have identified nine proteins, which are assembled into a complex, consisting of Ost1p, Stt3p, Wbp1p, Ost3p, Ost6p, Swp1p, Ost2p, Ost5p and Ost4p (Silberstein and Gilmore, 1996; Knauer and Lehle, 1999). One of the subunits, Ost2, is 40% identical to Dad1, a highly conserved protein in higher eukaryotes including plants. DAD1 (defender against apoptotic death) was originally isolated as a cDNA complementing a ts baby hamster kidney (BHK) cell line that dies by apoptosis (Nakashima et al., 1993; Niederer et al., 2005) at its non-permissive temperature, probably by disturbing cellular functions as a result of impairment of the essential process of N-linked glycosylation. Moreover, Dad1 was shown to interact with Mcl1, a member of the Bcl2 family (Makishima et al., 2000) and mice lacking Dad1 undergo apoptosis-associated embryonic death (Brewster et al., 2000; Hong et al., 2000). Inhibition of N-glycosylation by tunicamycin was also reported to induce apoptosis in mammalian cells, but with conflicting results (Perez-Sala and Mollinedo, 1995; Makishima et al., 1997; Hacki et al., 2000). In human, defects in N-glycosylation are the cause for congenital disorders of glycosylation (CDG), a new family of genetic diseases with a severe, multisystemic clinical picture, including central and peripheral nervous involvement as well as psychomotor retardation (Jaeken and Carchon, 2004).

Here, we demonstrate that yeast N-glycosylation mutants with defects in the oligosaccharyltransferase undergo apoptosis-like cell death accompanied by the occurrence of typical morphological apoptotic hallmarks, by the induction of a caspase-like activity as well as the production of reactive oxygen species (ROS). As deletion of the metacaspase YCA1/MCA1 does not seem to affect this activity, we postulate a different proteolytic process to be involved in the N-glycosylation-induced cell death. On the other hand expression of human Bcl-2 diminishes caspase-like activity, improves cell survival and restores the growth defect observed in connection with the N-glycosylation defect, suggesting involvement of Bcl-2-related steps in the cascade.


A defect in N-glycosylation exhibits typical apoptotic cellular phenotypes

When the N-glycosylation-defective, temperature-sensitive mutants ost2-3 or wbp1-1 were shifted from 25°C to the non-permissive temperature of 37°C an underglycosylation of glycoproteins occurs (te Heesen et al., 1992; Silberstein et al., 1995) and cells start dying. Under these conditions typical morphological markers of apoptosis could be observed. By 4,6-diamidino-2-phenylindole (DAPI) staining, chromatin fragmentation or distributed nuclear fragments were detectable (Fig. 1A–C). Likewise DNA cleavage occurred, as proven by a strong TdT-mediated dUTP-X nick end labelling (TUNEL)-positive phenotype (Fig. 1D–F). Moreover, exposure of phosphatidylserine at the outer leaflet of the plasma membrane could be demonstrated by Annexin V-FITC staining (Fig. 1G–I), under conditions where membrane integrity was still retained, as indicated by exclusion of propidium iodide co-staining.

Figure 1.

Apoptotic phenotypes induced by defective N-glycosylation. DAPI-stained cells reveal nuclear breakage (A–C); DNA fragmentation was monitored by TUNEL staining (D–F) and Annexin V – propidium iodide staining demonstrates exposition of phosphatidylserine (G–I). The arrow in I indicates a double-stained cell, indicative of advanced apoptosis or necrosis that could be observed during later stages of apoptosis induction. ost2 (B, E and H) and wbp1-1 (C, F and I) mutants were grown at 25°C, shifted to 37°C and processed as described in Experimental procedures. In A, D and G, corresponding images of isogenic wild-type (SS330) cells are shown. Bars correspond to 5 µm (A–C; G–I) and 10 µm (D–F).

Detection of caspase-like activity by cell flow cytometry and in cell-free extracts of N-glycosylation-defective cells

Core effectors of apoptosis encompass proteolytic enzymes of the caspase family, although there is recent increasing evidence of other, caspase-independent routes of apoptotic pathways (Susin et al., 1999; Cande et al., 2002). We asked, whether a caspase-like activity can be measured in vivo as well as in vitro in response to an N-glycosylation defect. To this purpose wbp1-1 cells were incubated with cell permeable FITC-labelled valyl-alanyl-aspartyl-[O-methyl]-fluoromethylketone (FITC-VAD-FMK) that binds to the active centre of metazoan caspases allowing the determination of cells with active caspases by flow cytometry. As shown in Fig. 2A, a shift of the wbp1-1 culture from 25°C to the restrictive temperature of 37°C caused induction of caspase activity (36% and 66%, respectively, of the cells after 4 and 8 h), whereas in cells at 25°C only a slight activity could be measured. Also isogenic wild-type cells, when shifted to 37°C, did not provoke caspase activity (Fig. 2B). The enzyme activity was reduced in the presence of the broad spectrum caspase-inhibitor carbobenzoxy-valyl-alanyl-aspartyl-[O-methyl]-fluoromethylketone (zVAD-FMK) or by addition of the anti-oxidant N-acetyl-l-cysteine (NAC). NAC was reported to block or delay animal cell death (Mayer and Noble, 1994). Treatment of wbp1-1 cells with zVAD-FMK or NAC during exposure at 37°C resulted also in a partial, but significant, better survival compared with cells without these compounds (Fig. 2C). In Fig. 3 it is shown that caspase activity was dependent on protein synthesis, as in the presence of cycloheximide the amount of apoptotic cells was reduced. This indicates that cell death process caused by the N-glycosylation defect requires cellular metabolism, as reported for H2O2- (Madeo et al., 1999) or acetic acid-mediated apoptosis (Ludovico et al., 2001) of yeast cells. As it was reported that yeast cells may bind under certain conditions non-specifically FITC-labelled VAD-FMK (Wysocki and Kron, 2004) we also determined caspase activity in cell-free extracts in a fluorimetric caspase assay in the presence of different caspase substrates (Fig. 4). Whereas N-acetyl-Asp-Glu-Val-Asp-7-amino-4-trifluoro-methyl-coumarin (Ac-DEVD-AFC) and N-acetyl-Tyr-Val-Ala-Asp-7-amino-4-trifluoro-methyl-coumarin (Ac-YVAD-AFC) were poor substrates, N-acetyl-Val-Glu-Ile-Asp-7-amino-4-methyl-coumarin (Ac-VEID-AMC) and to some extent also N-acetyl-Ile-Glu-Thr-Asp-7-amino-4-trifluoro-methyl-coumarin (Ac-IETD-AFC) were efficiently cleaved. The activities could be inhibited in part by the respective specific aldehyde-inhibitors (Ac-VEID-CHO, etc.) or by the pancaspase inhibitor zVAD-FMK, when added to the reaction mixture. In contrast to the wbp1-1 mutant, in isogenic wild-type cells, when shifted to 37°C, the activity was not increased (data not shown). The assay was linearly dependent on time up to 24 h and on the amount of protein added (data not shown). The activity of extracts isolated from cells grown at 25°C may reflect some upregulation of the apoptotic cascade in ost mutants already at the permissive temperature. This finding is in agreement with the observation that, depending on the type of ost mutant and growth conditions, some underglycosylation of proteins, indicative for endoplasmic reticulum (ER) stress, can be observed already at the permissive temperature (te Heesen et al., 1992). We also noticed that the relative activity of caspase activity of cells at 25°C compared with 37°C was higher in the in vitro assay, as when measured in vivo by flow cytometry (Fig. 2). The reason is not clear. One possibility could be that in cell-free extracts from yeast also some unspecific cleavage of the common vertebrate caspase substrates occurs by proteases other than caspases (see also Discussion).

Figure 2.

Flow cytometry analysis of caspase-like activity.
A. wbp1-1 cells were grown overnight at 25°C in YEP-glucose media to an OD600 of 5 × 10−6. The culture was split and then incubated further at 25°C and 37°C, respectively, for 4 h or 8 h. Pancaspase inhibitor zVAD-FMK (20 µM) or NAC (10 mM) were added at the time point of temperature shift and a second time 0.5 h prior to measurement as indicated.
B. Isogenic wild-type cells (SS330) were shifted to 37°C for 8 h and analysed by flow cytometry.
C. Survival of wbp1-1 cells. Cultures were shifted for 4, 6, or 8 h to the restrictive temperature of 37°C in the absence or presence of zVAD-FMK and NAC respectively. Single cells were sorted with the MoFlo cell sorter, spotted onto YEPD plates and incubated at the permissive temperature of 25°C. The experiments have been performed three or more times with similar values.

Figure 3.

Caspase-like activity is dependent on protein synthesis. Cell cultivation and measurement of the activity was as in Fig. 2. Cycloheximide (50 µg ml−1) was added at the time point of temperature shift. Data were recorded after 4 h. The values shown are representative for three independent experiments.

Figure 4.

Caspase-like activity in vitro. Cell-free extracts were prepared from wbp1-1 cultures grown at 25°C or shifted to 37°C for 6 h as described in Experimental procedures. Extracts were tested for their ability to cleave the fluorogenic caspase substrates Ac-DEVD-AFC, Ac-YVAD-AFC, Ac-VEID-AMC or Ac-IETD-AFC in the absence or presence of the respective aldehyde-inhibitors. The data presented in the right panel was obtained in the presence of the pancaspase inhibitor zVAD-FMK. The values are the average of three independent experiments.

Defects in N-glycosylation lead to production of ROS

One of the key mechanisms by which cells trigger programmed cell death is the production of reactive oxygen species (ROS). Dihydrorhodamine 123 (DHR123) was used as a probe for the detection of this apoptotic marker, which is oxidized to the fluorescent chromophore rhodamine by ROS. As shown by confocal microscopy, 31% and 87%, respectively, of wbp1-1 cells appeared fluorescent after a 4 h and 16 h shift to the restrictive temperature (Fig. 5A, B, I and J). A similar number of cells was also positive, when caspase-activity was measured using FITC-VAD-FMK for staining, which amounted to 40% and 86% respectively (Fig. 5E, F, M and N). In contrast, wild-type cells under these conditions showed hardly neither ROS (Fig. 5C, D, K and L) nor caspase staining (Fig. 5G, H, O and P). The results suggest the involvement of ROS in the apoptotic cascade induced by defective N-glycosylation.

Figure 5.

Cells displaying caspase activity produce also reactive oxygen species. wbp1-1 or isogenic wild-type cells (SS330) were shifted to 37°C for 4 h (A–H) and 16 h (I–P), respectively, and stained with FITC-VAD-FMK for caspase activity (E–H, M–P) or with dihydrorhodamin 123 for ROS (A–D, I–L). Both confocal images (panels 1 and 3) and overlays with DIC images (panels 2 and 4) are shown. Experiment was performed at least three times independently.

Disruption of YCA1 or osmotic stabilization does not attenuate apoptosis-induced N-glycosylation defects

It has been shown in the case of H2O2-, acetic acid- or age-induced apoptosis (Ludovico et al., 2001; Madeo et al., 2002; Herker et al., 2004) that disruption of the metacaspase YCA1 abrogates and retards, respectively, apoptosis pointing to a functional participation of YCA1 in this process. As depicted in Fig. 6A, deletion of YCA1 in wbp1-1 did not affect caspase activity, when measured by flow cytometry. In fact the amount of apoptotic cells in the YCA1-deleted strain was slightly but consistently higher, for reasons not known. Furthermore, there was also no delay in the survival rate of wbp1-1yca1Δ cells compared with the wbp1-1 single mutant (Fig. 6B). These results suggest that YCA1 does not seem to be involved in this type of programmed cell death and another caspase-like activity must be responsible for the proteolytic activity. As defective N-glycosylation may affect cell wall integrity, which can be suppressed by osmotic support (Heinisch et al., 1999), we tested whether supplementation of the medium with 1 M sorbitol as stabilizer can prevent or retard induction of apoptosis and viability. As shown in Fig. 7, no protection of these processes occurred in the presence of sorbitol.

Figure 6.

Disruption of YCA1 does not abrogate apoptosis induced by defective N-glycosy-lation.
A. Flow cytometric measurement of caspase activity of wbp1-1 and wbp1-1yca1Δ cells shifted to 37°C for 4 and 8 h respectively.
B. Survival of wbp1-1 and wbp1-1yca1Δ strains. Cells were shifted to the restrictive temperature of 37°C for 4 or 8 h, sorted for single cells with the MoFlo cell sorter, spotted onto YEPD plates and incubated at 25°C. The data shown are representative for at least three independent experiments.

Figure 7.

Osmotic stabilization by sorbitol does not abrogate apoptosis. wbp1-1 cells were grown at 25°C in YEPD medium without (A) and with 1 M sorbitol (B), shifted to 37°C for 4 and 8 h, respectively, and analysed by flow cytometric analysis for caspase activity.
C. Survival of wbp1-1. Cultures from A and B were sorted for single cells with the MoFlo cell sorter, spotted onto YEPD and YEPD-sorbitol plates, respectively, and incubated at the permissive temperature of 25°C. The data shown are representative for at least three independent experiments.

Expression of human Bcl-2 in wbp1-1 abrogates apoptosis induced by N-glycosylation

Ectopic expression in yeast of a variety of apoptosis-regulatory proteins derived from animal or human sources has been used for different purposes in apoptosis research. We asked, whether the known apoptosis blocker Bcl-2 is able to prevent apoptotic phenotypes caused by defective N-glycosylation. Human Bcl-2 was expressed under the control of the galactose inducible GAL1 promoter in the wbp1-1 mutant. As shown in Fig. 8A, in wbp1-1 cells harbouring Bcl-2 the amount of apoptotic cells was reduced from 56% to 19%, compared with the wbp1-1 cells without Bcl-2. It should be noted that under the growth conditions used, i.e. minimal medium and galactose as carbon source, induction of apoptosis is somewhat slower (compare with Fig. 2). In contrast to wild-type Bcl-2, expression of non-functional Bcl-2ΔCp with a C-terminal deletion did not restore enzyme activity. We also found that Bcl-2 improved survival of wbp1-1 cells upon induction of apoptosis (Fig. 8B), as well as was able to suppress the temperature growth defect of the wbp1-1 mutant (Fig. 8C).

Figure 8.

Heterologous expression of human BCL2 abrogates caspase activation (A), improves survival (B) and suppresses the temperature sensitivity (C) of the wbp1-1 mutant.
A. Flow cytometry analysis of caspase activity in wbp1-1 and wbp1-1 expressing BCL2 and BCL2ΔC, respectively, under the control of the Gal-inducible promoter. Cells were cultivated in YNB minimal medium containing galactose and shifted for 12 h to 37°C.
B. Survival of cells upon 12 h exposure to the restrictive temperature of 37°C.
C. In total, 3 µl of a serial 1:10 dilution of 0.1 OD600 cells were plated on YEP-galactose and incubated at 25°C and 31°C respectively. The data shown are representative for at least three independent experiments.

Tunicamycin induces apoptosis in wild-type yeast

N-glycosylation can be inhibited by tunicamycin by interfering with the formation of dolichol-pyrophosphate-N-acetylglucosamine, the first step in the biosynthesis of the lipid-linked oligosaccharide precursor (Lehle and Tanner, 1976). We asked, whether apoptosis can be induced also in wild-type cells, when N-glycosylation is prevented by tunicamycin. As shown in Fig. 9, when wild-type cells were grown for 8 h at 37°C in the presence of tunicamycin 54% of the cells were apoptotic. Similarly as in the case of the wbp1-1 mutant (Fig. 2), NAC or zVAD-FMK partially protected the tunicamycin effect (Fig. 9). Surprisingly the defect caused by tunicamycin was dependent on the growth temperature, as at 30°C no increase of the caspase activity occurred (see right panel). A temperature shift to 32°C and 35°C in the presence of tunicamycin for 8 h caused 13% and 50%, respectively, of the cells to go into apoptosis (data not shown). These results indicate that the N-glycosylation defect needs in addition a temperature stimulus to exert the cell death programme.

Figure 9.

Inhibition of N-glycosylation by tunicamycin in wild-type cells. Cells (strain BY4741) were cultivated at 25°C and shifted to 37°C for 8 h in the presence of 10 µg ml−1 tunicamycin (TM). Where indicated in addition the pancaspase inhibitor zVAD-FMK (20 µM) or NAC (10 mM) were added. In the right panel, TM was added, but cells were kept at 30°C. The data shown are representative for at least three independent experiments.


The ER is an organelle in which secretory as well as resident proteins of the secretory pathway are synthesized, folded and modified. The most common type of covalent modification in the ER is N-linked glycosylation that has been shown to be essential for viability of cells and development of multicellular organisms (Huffaker and Robbins, 1983; Ioffe and Stanley, 1994).

The ER is very sensitive to perturbations of its environment. To survive ER stress situations, this organelle responds by triggering specific signalling pathways including the adaptive unfolded protein response (UPR), first discovered in yeast (Patil and Walter, 2001). Recent experimental evidence in mammalian and yeast cells indicates that prolonged ER stress may ultimately lead to apoptotic cell death (Breckenridge et al., 2003; Haynes et al., 2004; Rao et al., 2004; Rutkowski and Kaufman, 2004). Besides alterations in Ca2+ homeostasis or oxidative stress, also misfolded proteins cause disruption of ER function. One of the highly diverse roles attributed to N-glycan chains of proteins in the ER is their contribution to proper folding and degradation of misfolded glycoproteins, respectively, via ER-associated degradation (ERAD) (Helenius and Aebi, 2001). Therefore, also defects in N-glycosylation can be expected to contribute to ER stress. We have shown here that yeast mutants with a defect in the key enzyme of protein N-glycosylation, the oligosaccharyltransferase, catalysing the transfer of the conserved core-oligosaccharide to the nascent polypeptide chain, leads to apoptotic-like cell death. We were able to demonstrate, as a consequence of defective N-glycosylation, the appearance of typical morphological features of apoptotic cell death, such as chromatin condensation, DNA fragmentation and translocation of phosphatidylserine, as well as the induction of caspase-like activity and the formation of reactive oxygen species.

With respect to the established caspase-like activity in response to defective N-glycosylation, one should mention that in yeast to date only one metacaspase, Yca1/Mca1p, has been identified and shown to be implicated in H2O2-, ageing- (Madeo et al., 1999), in salt stress- (Wadskog et al., 2004), or killer virus-induced (Ivanovska and Hardwick, 2005) apoptosis events. On the other hand, there is recent evidence for Yca1-independent apoptotic pathways identifying the HtrA-like protein Nma111 as a nuclear serine protease and the apoptosis-inducing factor AIF as mediators of yeast apoptosis (Fahrenkrog et al., 2004; Wissing et al., 2004). The mammalian homologue of HtrA2/Omi was recently reported to mediate apoptosis in a serine protease-dependent manner owing to its ability to antagonize the inhibitor of apoptosis protein XIAP (Suzuki et al., 2001; Hedge and Williams, 2002; Verhagen et al., 2002). Furthermore, serine proteases, different from HtrA2/Omi, were found to act during ER stress signalling both in caspase-dependent and caspase-independent pathways (Egger et al., 2003). This study provides evidence that the caspase-like activity seems to be independent of Yca1p, as deletion of YCA1 did not influence the amount of apoptotic cells in vivo when measured with FITC-VAD-FMK, nor did it affect the survival of cells. The caspase-like activity determined in cell-free extracts revealed a substrate preference for the VEID- and diminished also for the IETD-peptide, and was inhibited significantly, albeit not complete, by the corresponding specific inhibitors as well as by the non-specific pan-caspase inhibitor zVAD-FMK. In mammalian cells these two compounds are used as substrates for the effector caspase-3 and caspase-6 respectively. As an overlapping activity between these compounds has been observed, an accurate classifying of the protease activity is not possible. Thus the yeast caspase-like activity responsible for the cleavage of these vertebrate-type caspase substrates as well as its own target substrate(s) remain to be identified. As it has been shown that some proteases different from the caspase family are able to cleave caspase substrates and are prone to interact with small peptide caspase inhibitors also, such a possibility cannot be excluded (Turk et al., 2002; Rozman-Pungercar et al., 2003). In this context it is also worth mentioning that the recently discovered plant vacuolar protease, VPE, mediating virus-induced hypersensitive programmed cell death is structurally unrelated to caspases, although it has caspase-1 activity and an ability to bind caspase-1 inhibitors (Hatsugai et al., 2004). In addition to the above-mentioned serine protease-mediated death signalling in the ER, caspase-3 and caspase-9 (Egger et al., 2003) as well as caspase-12 (Nakagawa et al., 2000; Morishima et al., 2002; Rao et al., 2002) were shown to be activated in mammalian cells during ER stress, pointing to a rather complex, not yet understood network, to sense and to cope with ER abnormalities.

Another finding of this study was the generation of ROS and that addition of anti-oxidants or heterologous expression of human Bcl-2 reduced caspase activity and increased viability of cells. Human Bcl-2 was shown before to protect yeast also against superoxide-induced death and to delay death when entering stationary phase (Longo et al., 1997). We have not shown yet, where ROS originates from and how it integrates into the cell death process. As sources for ROS are conceivable both mitochondria and the UPR-regulated oxidative protein folding machinery in the ER (Tu and Weissman, 2002; Harding et al., 2003). Regarding Bcl-2, most of the results published to date emphasize its importance at the mitochondria. However, several new observations suggest also a function for Bcl-2 at the ER (Rudner et al., 2002; Schinzel et al., 2004). Thus, Bcl-2 was shown to associate with the ER and nuclear envelope and to interfere with apoptosis induction, e.g. by Bax, ceramides or serum withdrawal. Moreover, recent studies uncovered an apoptotic cross-talk between the ER and mitochondria. When ER function was disturbed by brefeldin, a drug-causing ER dilation due to inhibition of secretion, release of cytochrome c from mitochondria occurs that is controlled by a Bcl-2 variant exclusively targeted to the ER (Hacki et al., 2000; Annis et al., 2004). Furthermore, cytochrome c release seems to be caused by a release of Ca2+ from the ER (Boya et al., 2002) controlled by BH3-only class proteins (Zong et al., 2003), followed by rapid uptake in mitochondria through regions that are associated with the ER membrane (Filippin et al., 2003). Other proteins recently implicated in mammalian ER stress-induced apoptosis include the tyrosine kinase c-Abl (Ito et al., 2001) and the integral ER protein BAP31 interacting with Bcl-2 and procaspase-8L (Breckenridge et al., 2003; Reimertz et al., 2003). But altogether the mechanistic link between ER stress and apoptosis remains to be elucidated. Whether similar scenarios hold true also in yeast cannot be answered, until corresponding functional homologues will be identified.

Finally we showed that also in wild-type yeast, cell death could be induced, when N-glycosylation was prevented by tunicamycin, as it was observed, but in part with conflicting results, in animal cells (Perez-Sala and Mollinedo, 1995; Makishima et al., 1997; Hacki et al., 2000; Niederer et al., 2005). Likewise, as already shown for wbp1-1, both NAC and zVAD-FMK were able to retard tunicamycin-mediated apoptosis. But we found that tunicamycin alone was not sufficient, at least under the conditions and time frame used here, to provoke the death cascade. A temperature stimulus was needed in addition. Under the assumption that the glycosylation defect contributes to protein folding, additional temperature stress may prevent the cell to cope with this already deteriorated situation.

Compared with our knowledge, how mitochondria are involved in intrinsically induced programmed cell death, the signalling pathways by which ER stress contributes to apoptosis is still in its infancy. As ER stress in human is considered a cause of pathologically relevant apoptosis, particularly of neurodegenerative disorders (Aridor and Balch, 1999; Paschen and Frandsen, 2001; Kouroku et al., 2002; Hetz et al., 2003), ER-mediated cell death programmes has received growing attention. Even though realizing that apoptosis in yeast differs from that of multicellular organisms, yeast with its both biochemical and genetic power may well help to isolate and identify ancient components and targets to delineate ER-mediated apoptotic cascades.

Experimental procedures

Yeast strains, media and genetics methods

The following strains were used: SS330 (MATa ade2-101 ura3-52 his3Δ200 tyr1), MA7-B (MATa ade2-101 ura3-52 his3Δ200 lys2-801 wbp1-1), BY4741 (MATa his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0), ost2-3 (MATα ade2-101 ura3-52 his3Δ200 lys2-801 leu2-Δ1 trp1-Δ1 Δost2::LEU2 [pRS314-ost2-3]). Strains were grown in YEP (1% yeast extract, 2% bacto-peptone) in the presence of 2% glucose or galactose, or in selective YNB (yeast nitrogen base) dropout medium supplemented with amino acids and nucleotide bases and 2% glucose and galactose, respectively, at the temperatures indicated. To induce apoptosis, fresh cultures were grown to a cell density of 5 × 106 cells ml−1 at 25°C and divided into two flasks and then further incubated for the indicated time at 25°C or 37°C.

Disruption of YCA1 was carried out with a kanMX-cassette amplified by polymerase chain reaction (PCR) from the plasmid pUG6 (Guldener et al., 1996), using 5′fw primer 5′-GCGTCCGGGTAATAACAACTATTGAAAAAGCATGGCTTCG CATTAATAGGAGCCcagctgaagcttcgtacgc-3′ and 3′rev primer 5′-TACATAATAAATTGCAGATTTACGTCAATAGGGTGTGAC GATGATAATTGTGGgcatagcccactagtggatct-3′. The cassette was transformed into yeast strain MA7-B. Recombinants were selected for resistance to G418 sulphate and positive clones were verified by PCR. Plasmids encoding human BCL2 or BCL2ΔC under the control of the inducible GAL1 promoter were kindly provided by S. Manon (Bordeaux) and transformed into yeast using standard techniques (Guthrie and Fink, 1991).

Detection of apoptotic markers

Chromatin staining.  For nuclear staining with diaminophenylindole (DAPI, Sigma), cells were washed with PBS, fixed in 70% ethanol for 30 min and washed again. Samples were incubated 20 min with 0.5 µg ml−1 DAPI in PBS and analysed after washing by fluorescence microscopy (Zeiss Axioskop with Axiovision software).

Annexin V staining.  Phosphatidylserine was labelled with annexin V-FITC (ApoAlert Annexin V-FITC Apoptosis Kit, Clontech) after cell wall digestion. Cells were washed in sorbitol buffer containing 1.2 M sorbitol, 35 mM potassium phosphate (pH 6.8), 0.5 mM MgCl2 and digested with 4.2% glusulase (NEN Life Science Products) and 0.15 mg ml−1 zymolyase 100T (Seigagaku Kogyo) in the same buffer for 1 h at 30°C. Spheroblasts were carefully washed with 10 mM HEPES-NaOH (pH 7.4), containing 1.2 M sorbitol, 35 mM potassium phosphate, 0.5 mM MgCl2, 140 mM NaCl, 2.5 mM CaCl2, and incubated in the same buffer with 2 µg ml−1 Annexin V-FITC and 50 ng ml−1 propidium iodide for 15 min [final volume 50 µl for 1 OD (optical density)]. For microscopy, cells were harvested and resuspended in washing buffer.

TUNEL assay.  DNA fragmentation was demonstrated by labelling free 3′-OH termini with FITC-coupled deoxyuridine, an anti-fluorescein antibody conjugated to peroxidase (In Situ Cell Death Detection Kit, POD; Roche) and DAB as a peroxidase substrate (Roche). Yeast cells were harvested, fixed in 4% paraformaldehyde for 1 h and digested as described above for annexin V-labelling. Spheroblasts were applied to a silane-coated slide and processed according to the manufacturer's instructions. After rinsing with PBS, endogenous peroxidases were blocked by incubation with 0.3% H2O2 in methanol for 30 min. Slides were then washed with PBS three times and covered with 0.1% Triton X-100, 0.1% sodium citrate for 2 min at 4°C for permeabilization of cells. Cells were washed again three times and incubated with 10 µl of TUNEL reaction mixture [terminal deoxynucleotidyl transferase 200 U ml−1, 10 mM FITC-dUTP, 25 mM Tris-HCl (pH 7.2), 200 mM sodium cacodylate, 5 mM CoCl2] at 37°C for 1 h. After washing, 10 µl of Converter-POD (anti-fluorescein antibody conjugated to horseradish peroxidase) was added for 30 min at 37°C. Slides were stained with DAB-substrate solution for 10 min at room temperature. Coverslips were mounted with Kaiser's glycerol gelatin (Merck).

Caspase and ROS staining of cells

In total, 5 × 106 cells were harvested, washed in PBS and resuspended in 200 µl of staining solution containing 10 µM FITC-VAD-FMK (CaspACE FITC-VAD-FMK in situ marker, Promega). After incubation for 20 min at room temperature with low agitation in darkness, cells were centrifuged, washed twice in 1 ml of PBS and resuspended in 200 µl of PBS. Free intracellular radicals (ROS) were detected by DHR123, which was added 1.5–2 h before harvesting the cells from a 2.5 mg ml−1 stock solution in ethanol to a final concentration of 5 µg ml−1. For confocal laser scanning microscopy (LSM 510-META, Zeiss) FITC-VAD-FMK stained cells were detected with an argon laser (excitation 488 nm).

Assay of caspase activity by flow cytometry and in cell-free extracts

For flow cytometric analysis cells were stained as described above with FITC-VAD-FMK and analysed using a FACS Calibur equipment (Becton Dickinson) and CellQuest data software or with the MoFlo (Cytomation) high-speed sorter and Summit v3.1 software with excitation and emission settings of 488 nm and 525–550 nm respectively (filter FL-1).

A fluorimetric assay was established to measure caspase activity in vitro. An exponential cell culture of an OD600 of 0.5 grown at 25°C was split and incubated further for 6 h at 25°C and 37°C respectively. Cells of 15 OD were harvested, washed with 5 ml of Tris-HCl (pH 7.5) and resuspended in 200 µl lysis buffer, containing 20 mM HEPES (pH 7.3), 0.5% Nonidet P40, 84 mM KCl, 10 mM MgCl2, 0.2 mM EDTA, 0.2 mM EGTA, 1 mM DTT, 5 µg ml−1 aprotinin, 1 µg ml−1 leupeptin, 1 µg ml−1 pepstatin and 1 mM PMSF. Cells were lysed with glass beads (0.45–0.55 mm) at 4°C. The homogenate was centrifuged for 30 min at 20 000 g and the supernatant was used as enzyme source. Protein extract (50 µg protein) was incubated with 40 µM Ac-DEVD-AFC, Ac-YVAD-AFC, Ac-IETD-AFC (Biomol, Germany) or Ac-VEID-AMC (Calbiochem) in 50 mM HEPES-NaOH (pH 7.3) buffer containing 100 mM NaCl, 10% saccharose, 0.1% CHAPS and 10 mM DTT in a final volume of 1 ml. Assay was linear with time for more than 24 h and dependent on the amount of protein added. Caspase inhibitors were added at a concentration of 20 µM, when indicated. Measurements were accomplished in a Jobin Yvon-Spex Fluoromax-2 spectrofluorimeter with DataMax software. Release of 7-amino-4-trifluoromethylcoumarin (AFC) was detected using an excitation wavelength of 400 nm and an emission wavelength of 489 nm; for release of 7-amino-4-methylcoumarin (AMC) an excitation wavelength of 370 nm and an emission wavelength of 440 nm were applied.

Test for determination of survival

To determine survival, cells were separated by the MoFlo cell sorter and spotted onto YEPD plates. For each time point 250 single cells were plated and colonies were determined after 4–6 days’ incubation at the indicated temperature.


We are grateful to Stephen Manon for generously providing Bcl-2 plasmids, Reid Gilmore for the ost2 strain and Michael Thomm for access to the MoFlo instrument. We also thank Wolfgang Forster and Marit Hoffmann for their help in flow cytometry measurements. This work was supported by grants from the Deutsche Forschungsgemeinschaft and the Körber-Stiftung.