Yeast caspase 1 links messenger RNA stability to apoptosis in yeast

Cristina Mazzoni, Eva Herker, Vanessa Palermo, Helmut Jungwirth, Tobias Eisenberg, Frank Madeo, Claudio Falcone

Author Affiliations

  1. Cristina Mazzoni*,1,
  2. Eva Herker2,
  3. Vanessa Palermo1,
  4. Helmut Jungwirth2,
  5. Tobias Eisenberg3,
  6. Frank Madeo3 and
  7. Claudio Falcone1
  1. 1 Department of Cell and Developmental Biology, Pasteur Institute–Cenci Bolognetti Foundation, University of Rome ‘La Sapienza’, Piazzale Aldo Moro 5, 00185, Rome, Italy
  2. 2 Institute for Physiological Chemistry, University of Tübingen, Hoppe‐Seyler‐Strasse 4, 72076, Tübingen, Germany
  3. 3 IMB, Karl‐Franzens University, Universitaetsplatz 2, 8010, Graz, Austria
  1. *Corresponding author. Tel: +(39 6) 4991 2257; Fax: +(39 6) 4991 2256; E-mail: cristina.mazzoni{at}


During the past years, yeasts have been successfully established as models to study the mechanisms of apoptotic regulation. We recently showed that mutations in the LSM4 gene, which is involved in messenger RNA decapping, lead to increased mRNA stability and apoptosis in yeast. Here, we show that mitochondrial function and YCA1, which encodes a budding yeast metacaspase, are necessary for apoptosis triggered by stabilization of mRNAs. Deletion of YCA1 in yeast cells mutated in the LSM4 gene prevents mitochondrial fragmentation and rapid cell death during chronological ageing of the culture, diminishes reactive oxygen species accumulation and DNA breakage, and increases resistance to H2O2 and acetic acid. mRNA levels in lsm4 mutants deleted for YCA1 are still increased, positioning the Yca1 budding yeast caspase as a downstream executor of cell death induced by mRNA perturbations. In addition, we show that mitochondrial function is necessary for fast death during chronological ageing, as well as in LSM4 mutated and wild‐type cells.


Apoptosis is a form of cellular suicide, particularly important for medical research, because misregulation of apoptosis can result in many human diseases (Steller, 1995). Recent studies have established yeasts as models to study the mechanisms of apoptotic regulation. In the yeast Saccharomyces cerevisiae, we detected cell death, with typical markers of apoptosis, such as DNA fragmentation, phosphatidylserine externalization and chromatin condensation, in a strain carrying a mutation in the AAA‐ATPase gene CDC48 (Madeo et al, 1997). Its mammalian orthologue VCP was subsequently linked to the regulation of apoptosis (Shirogane et al, 1999). In addition, several other pathways crucial for mammalian apoptosis are conserved in yeast; for example, a newly discovered proteasomal apoptotic pathway leading to the destruction of Cdc6 (Blanchard et al, 2002). Exposure to low doses of H2O2 or acetic acid, which are known to increase reactive oxygen species (ROS) production, induces apoptosis in wild‐type yeast cells, indicating that, as in metazoans, ROS are key regulators of yeast apoptosis (Madeo et al, 1999; Ludovico et al, 2001).

RNA metabolism and turnover is an important regulator of mammalian apoptotic events (Bellacosa & Moss, 2003). In a previous study, we showed that a truncated form of the essential gene LSM4, a component of a complex promoting pre‐messenger RNA splicing and mRNA decapping (Cooper et al, 1995; He & Parker, 2000), still supports growth in an Lsm4 devoid strain (Mazzoni & Falcone, 2001). In the stationary phase, expression of Kllsm4Δ1, the truncated form of LSM4, leads to an accumulation of mRNAs and the phenotypic markers of apoptosis. We also showed that other mutations in the yeast mRNA‐decapping pathway, such as those in the LSM1, LSM6, DCP1 and DCP2 genes, show increased mRNA stability and undergo apoptosis (Mazzoni et al, 2003b). Control of mRNA stability has been described as an important checkpoint during apoptosis, and, in addition, it has been reported that Lsm proteins are involved in autoimmune diseases and cancer (Schweinfest et al, 1997; Stinton et al, 2004). However, a mechanistic link between mRNA stability and apoptosis has not yet been described either in yeasts or in mammals.

It was previously shown that chronological ageing leads to apoptosis in yeast (Herker et al, 2004) and that a caspase‐like protease (Yca1) mediates cell death triggered by oxygen stress, salt stress or chronological ageing (Madeo et al, 2002; Wadskog et al, 2004). Here, we show that the yeast caspase Yca1 is also required for apoptosis induced by increased mRNA stability.

The disruption of YCA1 does not affect mRNA stability but rescues cells from apoptotic death induced by mutation of LSM4, a master regulator of mRNA degradation.

Results And Discussion

Apoptosis in the Kllsm4Δ1 mutant requires Yca1

Wild‐type yeast cells lose viability during long‐term cultivation (Longo et al, 1997), a process that is accompanied by apoptotic markers (Herker et al, 2004). As shown in Fig 1A, a mutant in the mRNA‐decapping protein Lsm4 (Kllsm4Δ1) shows a much higher and more rapid loss of viability than the wild‐type strain.

Figure 1.

Cell death in lsm4 mutants during chronological ageing is attenuated by deletion of the yeast caspase YCA1. Cells of wild‐type (WT), yca1Δ, Kllsm4Δ1, Klsm4Δ1/yca1Δ (A) and their derivative rho° strains (C) were grown in SD medium. Viability is expressed as a percentage of micro‐colony‐forming units (CFU). (B) Cell viability of Klsm4Δ1/yca1Δ mutants after the reintroduction of the YCA1 and KlLSM4 genes. Average and standard deviation of three independent experiments are shown.

To investigate whether the loss of viability observed in the mutant was mediated by caspases, which are prominent cell executioners in mammalian cells, we assessed the cell viability of a Kllsm4Δ1/yca1Δ double‐mutant strain.

Deletion of the yeast caspase gene YCA1 reduced the marked loss of viability observed in the Kllsm4Δ1 mutant (Fig 1A). The deletion of YCA1 by itself also increased the fraction of viable cells compared with the wild type, similar to results of a previous study that used strains in a different genetic background (Madeo et al, 2002). The absence of the caspase only delays but does not completely prevent cell death in the Kllsm4Δ1 mutant, indicating that alternative pathways are also involved in this process. The introduction of the YCA1 and the KlLSM4 genes into the Kllsm4Δ1/yca1Δ double mutant led to cell viability levels similar to those observed for the Kllsm4Δ1 and yca1Δ mutants, respectively (Fig 1B).

Although mitochondria are required for induction of apoptosis in many death scenarios observed in mammalian cells (Danial & Korsmeyer, 2004), it is not yet known whether the apoptotic phenotypes induced by mRNA perturbations in either yeasts or mammals depend on mitochondrial function. To address this question, we created a series of isogenic rho° strains and measured cell viability during the growth and stationary phases (Fig 1C). We observed that, as in rho+ cells, the loss of viability was much more pronounced in the Kllsm4Δ1 mutant and that the deletion of YCA1 suppressed this phenotype. Moreover, rho° strains showed significantly slower kinetics of death, probably reflecting a reduced production of stress factors.

Phenotypic markers of apoptosis include chromatin condensation, fragmentation and DNA breakage. To visualize chromatin structure by 4,6‐diamidino‐2‐phenylindole (DAPI) staining, we used rho° derivative strains to eliminate the mitochondrial DNA background fluorescence. As shown in Fig 2A (second column from the left), wild‐type cells had a normal nuclear morphology, whereas, in the Kllsm4Δ1 mutant, nuclei appeared fragmented and DNA condensed, phenotypes that are much less pronounced in the double mutant. The absence of Yca1 in the Kllsm4Δ1 mutant also had an effect on cell morphology (Fig 2A, first column to the left). Cells of the double mutant appeared almost normal in shape and size compared with the abnormal morphology of the Kllsm4Δ1 mutant. A similar aberrant cell morphology has also been observed in apoptotic cdc48 mutant strains (Madeo et al, 1997).

Figure 2.

The absence of Yca1 attenuates the apoptotic phenotypes of lsm4 mutants. (A) Wild‐type (WT), yca1Δ, Kllsm4Δ1 and Klsm4Δ1/yca1Δ cells were analysed for nuclear fragmentation (4,6‐diamidino‐2‐phenylindole (DAPI) staining), DNA strand breaks (dark cells in TdT‐mediated dUTP nick end labelling test) and reactive oxygen species production (bright cells after dihydrorhodamine 123 (DHR) staining). Left column: phase‐contrast pictures corresponding to DAPI staining. (B) Quantification of cell population showing apoptotic phenotypes.

Another hallmark of apoptosis is DNA breakage, which can be detected by the TdT‐mediated dUTP nick end labelling (TUNEL) assay (Gavrieli et al, 1992; Gorczyca et al, 1993). We previously reported that Kllsm4Δ1 cells contain a high percentage of free 3′‐OH ends generated by fragmentation of chromosomes (Mazzoni et al, 2003a, 2003b). As shown in Fig 2A (third column from the left), Kllsm4Δ1 cells showed DNA fragmentation that was almost completely suppressed in cells with the YCA1 gene deleted.

Another key event in triggering apoptosis in yeast is the production of ROS, which accumulate in senescent cells (Laun et al, 2001) or after oxidizing treatment, such as exposure to H2O2 (Fahrenkrog et al, 2004). Therefore, we determined the amount of ROS by incubation with dihydrorhodamine 123 (DHR; Schulz et al, 1996). Wild‐type and yca1Δ strains showed about 1% of ROS‐positive cells (Fig 2A). In contrast, about 40% of Kllsm4Δ1 cells showed intense intracellular staining with DHR. However, the fraction of DHR‐positive cells was reduced to about 10% in the double mutant that had the YCA1 gene deleted. These results are summarized in Fig 2B.

The double mutant is more resistant to apoptotic stress

Low doses of H2O2 and acetic acid are established triggers of apoptosis in yeast (Madeo et al, 1999; Ludovico et al, 2001). Therefore, we measured the cell viability of strains after exposure to different concentrations of H2O2. As already reported (Madeo et al, 2002), the yca1Δ strain survived better than the wild‐type cells, and, in addition, the double mutant Kllsm4Δ1/yca1Δ showed a higher percentage of viable cells than the Kllsm4Δ1 mutant (Fig 3A). The increased resistance of the double mutant to oxygen peroxide could also be visualized through the halo test (Fig 3B), showing that the halo size of the Kllsm4Δ1 mutant was about one‐third that of the double mutant.

Figure 3.

Deletion of YCA1 in the Klsm4Δ1 mutant suppresses the sensitivity to H2O2, recovers the lack of growth on acetic acid, caffeine and glycerol and relieves mitochondrial fragmentation. (A) Cell viability was measured after exposure of cells to H2O2 at the indicated concentration for 4 h. CFU, colony‐forming unit; WT, wild type. (B) Halo test: about 108 cells of Kllsm4Δ1 and Klsm4Δ1/yca1Δ strains were spread onto YPD plates. Whatman 003 paper discs were soaked with 10 μl of 30% H2O2 and placed on the surface of the plates. Halo formation was recorded after 2 days of growth at 28°C. (C) Cells were streaked on YPD, YPD+0.25% caffeine and YPD+70 mM acetic acid, and plates were incubated at 28°C for 3 days. (D) Cells were streaked on YP+3% glycerol (YPY) plates and incubated at 28°C for 2 days. (E) Mitochondrial morphology in Kllsm4Δ1/yca1Δ and Kllsm4Δ1 cells shown by mitochondria‐targeted green fluorescent protein (mitoGFP).

Similarly, the Kllsm4Δ1 mutant showed a higher sensitivity to acetic acid, as it was not able to grow on YPD plates containing 70 mM of the compound. Again, the observed sensitivity was strongly reduced when the YCA1 gene was also inactivated (Fig 3C).

In a previous work, we observed a pleiotropic phenotype in Kllsm4Δ1 cells of the related yeast Kluyveromyces lactis as growth was impaired by several drugs (Mazzoni et al, 2003a). These included caffeine, a drug that affects the protein kinase C (PKC)–mitogen‐activated protein kinase (MAPK) pathway.

Caffeine is also known to induce apoptosis in mammalian cells, and this event is mediated by the p53, Bax and caspase 3 pathways (He et al, 2003).

As can be seen in Fig 3C, S. cerevisiae cells expressing Kllsm4Δ1 also showed sensitivity to caffeine, which could be fully suppressed by inactivation of YCA1. This result suggests that in yeast, in which the homologues of p53 and Bax have not yet been identified, caffeine could trigger cell death through a caspase‐dependent pathway (supplementary Fig 1 online). Interestingly, the overexpression of SLT2/MPK1, a member of the PKC–MAPK pathway, suppressed the caffeine sensitivity of the Kllsm4Δ1 cells (data not shown).

Kllsm4Δ1 cells show aberrant mitochondrial morphology

In wild‐type yeast strains, mitochondria are arranged as a tubular network that is the product of an equilibrium between fusion and fission events.

Excessive mitochondrial fission and/or lack of fusion result in the breakdown of the mitochondrial network, leading to the fragmentation of mitochondria, respiratory deficiency, ROS generation and apoptosis in mammalian cells (Yaffe, 1999; Frank et al, 2001; Karbowski & Youle, 2003).

Similarly, Kllsm4Δ1 cells showed an accumulation of ROS (Mazzoni et al, 2003b) and growth arrest on respiratory carbon sources (Fig 3D; supplementary Fig 1 online). Again, the inactivation of YCA1 in this strain could rescue respiratory deficiency (Fig 3D). We analysed the mitochondrial morphology in the Kllsm4Δ1 and Kllsm4Δ1/yca1Δ mutants using mitochondria‐targeted green fluorescent protein (mitoGFP), a protein that specifically targets the mitochondrial matrix compartment (Westermann & Neupert, 2000). A total of 87% of the Kllsm4Δ1 cells showed an aberrant mitochondrial morphology with a punctuate distribution instead of the wild‐type tubular shape (Fig 3E). This could be the result of increased fission activity as well as of fusion deficiency (Mozdy et al, 2000; Westermann, 2002). Intriguingly, inactivation of YCA1 restored the tubular structures in 53% of cells, which suggested that the signals for mitochondrial fission in the Kllsm4Δ1 mutant are at least partially caspase dependent.

Mitochondrial morphology also changed in the wild‐type cells after H2O2 treatment. In this case, 73% and 92% of cells showed the punctuate morphology after 20 and 60 min of incubation with 3 mM H2O2, respectively. Under the same conditions, 42% and 78% of yca1Δ cells showed the punctuate morphology, suggesting that an increase in mitochondrial fission is a general feature of yeast apoptosis. In accordance with this, it has been published recently that mitochondrial fission increases after cell treatment with acetic acid (Fannjiang et al, 2004). Moreover, our data suggest that mitochondrial fragmentation after apoptotic stimuli is partially dependent on the presence of Yca1.

Levels of YCA1 transcript are higher in Kllsm4Δ1

We previously suggested that, in the Kllsm4Δ1 mutant, the stabilization of mRNAs could lead to the abnormal production of some proteins that trigger apoptosis (Mazzoni et al, 2003b). In fact, the rate at which a specific mRNA is degraded can be a significant contributing factor to the overall expression levels of the encoded protein, and it has been recently reported that mRNA decay mutants show defects in translational control (Holmes et al, 2004).

As the absence of YCA1 could suppress most of the apoptotic phenotypes in Kllsm4Δ1 cells, we investigated whether YCA1 mRNA was also stabilized in this mutant. We prepared total RNA from exponentially growing and stationary cultures of wild‐type and Kllsm4Δ1 cells and analysed the amount of YCA1 transcript. As shown in Fig 4A, we observed that the amount of YCA1 mRNA was more abundant in the mutant than in the wild type, both in exponential and stationary cells, suggesting the possibility that the stabilization of the transcript in the mutant could result in the progressive accumulation of Yca1 and trigger apoptosis.

Figure 4.

The Kllsm4Δ1 mutant shows a higher YCA1 transcript level, and the deletion of YCA1 in Kllsm4Δ1 does not affect messenger RNA turnover rate. (A) Total RNAs were prepared from exponentially (Exp.) growing and stationary (Stat.) cells and analysed for YCA1 expression. Ribosomal 26S and 18S were used as internal standards. WT, wild type. (B) Total RNAs were prepared from WT (lanes 1–5), Kllsm4Δ1 (lanes 6–10) and Kllsm4Δ1/yca1Δ (lanes 11–15) cells grown overnight in YPD at 24°C. SSA4 expression was induced by incubating cells for 15 min at 45°C. Thereafter, cultures were back‐shifted at 24°C and RNA samples, prepared after 30, 60 and 90 min, were probed with the SSA4 gene (upper panels). Ribosomal RNA 5S internal was used as a control (lower panels).

Yca1 is not involved in mRNA degradation defects

As the Kllsm4Δ1 mutant shows a delayed mRNA degradation (Mazzoni et al, 2003b), we investigated whether the increased survival and suppression of apoptotic phenotypes observed in the double mutant Kllsm4Δ1/yca1Δ were due to a restoration of mRNA degradation rate to the wild‐type level.

We previously detected in Kllsm4Δ1 cells a stabilization of the transcript of the SSA4 gene (Mazzoni et al, 2003b), the transcription of which can be switched on and off at 45 and 24°C, respectively (Boorstein & Craig, 1990).

Therefore, we determined SSA4 mRNA levels in wild‐type and mutant strains.

As can be seen in Fig 4B, the exposure to 45°C activated gene transcription, and 90 min after shifting cells back to 24°C, the amount of SSA4 mRNA declined to the basal level in wild‐type cells (lanes 3–5). In single and double mutants, the level of SSA4 transcript remained high 90 min after shifting cells back to 24°C (lanes 8–10, 13–15). Therefore, the absence of Yca1 did not suppress the defects in transcription degradation observed in the Kllsm4Δ1 mutant, positioning the budding yeast caspase Yca1 as a downstream executor of cell death induced by mRNA stability.

In conclusion, we showed that the deletion of the yeast caspase gene YCA1 inhibits mitochondrial fragmentation and most of the apoptotic phenotypes observed in the apoptotic Kllsm4Δ1 mutant.

It is not yet known how mRNA perturbations lead to apoptosis in mammalian cells (Bellacosa & Moss, 2003). Future studies in yeasts may show more details of the connection between mRNA stability and apoptosis.


Strains and plasmids. The Euroscarf YCA1 (Yor197w) disruption strain (Y02453) was crossed four times with the strain MCY4/313Kllsm4Δ1 (Mat α, ade1‐101, his3‐Δ1, trp1‐289, ura3, LEU2‐GAL1‐SDB23, pRS313/Kllsm4Δ1) to obtain the isogenic strains CML39‐11A (Mat a, ade1‐101, his3‐Δ1, leu2, ura3, trp1‐289), CML39‐9A (Mat a, ade1‐101, his3‐Δ1, leu2, ura3, trp1‐289, yor197w∷kanMX4) and CML39‐8D (Mat a, ade1‐101, his3‐Δ1, leu2, ura3, trp1‐289, LEU2‐GAL1‐SDB23, yor197w∷kanMX4, pRS313/Kllsm4Δ1).

Plasmid pRS313/Kllsm4Δ1 was obtained by cloning a SmaI–EcoRV fragment from plasmid pKS/KlLSM4 (Mazzoni & Falcone, 2001), containing the promoter and gene portion encoding the first 72 amino acids of KlLSM4, into the SmaI site of pRS313 (Sikorski & Hieter, 1989). This plasmid was transformed into the MCY4 strain to give the strain MCY4/313Kllsm4Δ1.

Plasmid pUG36ng/YCA1 was obtained by cloning a HindIII fragment from plasmid pKF788 (Madeo et al, 2002) into the HindIII of the pUG36 vector without an XbaI fragment containing yeast enhanced green fluorescent protein (yEGFP; pUG36ng).

Cells were grown in YP (1% yeast extract, 2% peptone) supplemented with 2% glucose (YPD) at 28°C or in minimal medium (0.67% yeast nitrogen base) containing 2% glucose (SD) supplemented with 10 μg/ml of the appropriate nutritional requirements according to the genotype of the strains.

rho° strains were obtained as described by Mazzoni et al (2003b).

Fluorescence microscopy. For nuclear morphology, exponentially growing cells were fixed with 70% (v/v) ethanol and stained with DAPI (1 μg/ml). The presence of ROS was detected with DHR (Sigma Aldrich Co. D1054), as described previously (Madeo et al, 1999). Free 3′‐OH ends were detected by TUNEL on cells grown for 9 days, as described (Madeo et al, 1997). For analysis of mitochondria morphology, we used plasmids pVT100U‐mtGFP and pYX232‐mtGFP (Westermann & Neupert, 2000) that target GFP into mitochondrial matrix. For image acquisition, we used an Axioskop2 fluorescence microscope (Carl Zeiss, Jena, Germany) equipped with a digital camera (micro‐CCD).

RNA isolation and analysis. Total RNA (10 μg) was loaded onto 1.2% agarose–MOPS gels containing formaldehyde (Schmitt et al, 1990). The SSA4 probe was obtained by PCR amplification with primers designed in the coding region. The YCA1 probe was an EcoRI–NotI fragment derived from plasmid pFM21. Hybridization experiments followed standard procedures (Sambrook et al, 1989).

For RNA analysis in the exponential and stationary phases, yeast strains were grown overnight in YPD, cells were diluted to an optical density (OD)600 of 0.065–0.07, cultured in YPD medium to an OD600 of 0.4–0.5 and an aliquot was taken for isolation of RNA.

The rest of the culture was grown to an OD600 of 5–7 and used for isolation of RNA from stationary phase cells.

Supplementary information is available at EMBO reports online (‐s1.pdf).

Supplementary Information

Supplementary Figure 1 [embor7400514-sup-0001.pdf]


We thank Professor J.D. Beggs for kindly providing the S. cerevisiae strain MCY4 and Dr B. Westermann for pVT100U‐mtGFP and pYX232‐mtGFP plasmids. F.M. is grateful for grants from the Deutsche Forschungsgemeinschaft and Fonds zur Förderung der wissenschaftlichen Forschung (S9304‐B05). We are also grateful to B. Burhans for critical reading of the manuscript.