SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The cellular response of Saccharomyces cerevisiae to a linear plasmid encoded killer toxin from Pichia acaciae was analysed. As for the Kluyveromyces lactis zymocin, such toxin was recently shown to bind to the target cell's chitin and probably acts by facilitating the import of a toxin subunit. However, as distinct from zymocin, which arrests cells in G1, it provokes S-phase arrest and concomitant DNA damage checkpoint activation. Here, we report that such novel toxin type causes cell death in a two-step process. Within 4 h in toxin, viability of cells is immediately reduced to approximately 30%. Elevated mutation rates at the CAN1 locus prove DNA damaging mediated by the toxin. Cells arrested artificially in G1 or G2/M are very rapidly affected, while cells arrested in S loose their viability at a slower rate. S-phase arrest is, thus, a response of target cells to cope with DNA damage induced by the toxin. A second decline in viability requiring metabolically active target cells emerges upon toxin exposure over 10 h. During this phase, toxin treated cells develop abnormal nuclear morphology and react positive to terminal deoxynucleotidyl transferase-mediated nick end-labelling (TUNEL), indicative of DNA fragmentation. Furthermore, as judged from staining with fluorescein conjugated annexinV, cells expose phosphatidylserine at the outer membrane face and the formation of reactive oxygen species (ROS) is increased. ROS formation and concomitant cell death was heavily suppressed in a rho- derivative of the tester strain, while immediate reduction of viability was indistinguishable from the wild type. As a strain lacking the cellular target because of defects in the major chitinsynthase (Chs3) did not display such characteristic changes, the chitin binding and DNA-damaging P. acaciae toxin constitutes an apoptosis inducing protein. Both, DNA-damaging and apoptosis induction are unique features of this novel toxin type.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Production and secretion of cytotoxic compounds by microorganisms is a frequently applied principle to combat microbial competitors. In yeast, numerous killer systems have evolved. Among the strategies to kill or at least prevent growth of other yeasts is the inhibition of the target cell's transcription or replication machinery, membrane function or cell wall synthesis (reviewed in Magliani et al., 1997; Schmitt and Breinig, 2002; Schaffrath and Meinhardt, 2004). Such diverse targets require proteins specifically adapted and able to hijack normal cellular processes eventually causing a lethal effect. The K28 toxin, encoded by a dsRNA virus of Saccharomyces cerevisiae enters via endocytosis, and using the cell's secretory pathway, travels in reverse to the ER, exits to the cytoplasm and finally interferes with DNA synthesis in the nucleus (Schmitt et al., 1996; Eisfeld et al., 2000). The K1 killer toxin affecting membrane function is also encoded by a dsRNA virus and acts by activating Tok1 potassium channels (Ahmed et al., 1999; Schmitt and Breinig, 2002).

The lethal interaction between Kluyveromyces lactis and S. cerevisiae relies on zymocin, a heterotrimeric protein toxin consisting of α, β, γ subunits encoded by the linear plasmid pGKL1 of the dairy yeast (Gunge et al., 1981; Stark and Boyd, 1986; Stark et al., 1990). The α subunit binds to the target cell's chitin and the remarkably hydrophobic β-subunit is probably involved in uptake of the γ subunit into the target cell (Jablonowski et al., 2001a); γ interferes with RNA-polymerase II dependent transcription and requires the functional RNA-polymerase II Elongator complex to finally arrest target cells in G1 (Butler et al., 1991; Frohloff et al., 2001; Jablonowski et al., 2001b).

Besides the zymocin, three other linear plasmid encoded killer systems are known to date (Worsham and Bolen, 1990; Hayman and Bolen, 1991; Klassen and Meinhardt, 2002). The toxin encoded by pPin1–3 of Pichia inositovora functions analogously to zymocin, as it binds to chitin and is also dependent on RNA-polymerase II Elongator (Klassen and Meinhardt, 2003). However, toxins encoded by linear plasmids pWR1A of Wingea robertsiae and pPac1–2 of Pichia acaciae resemble each other but clearly differ from zymocin and the P. inositovora toxin (Klassen et al., 2004). As for the latter they indeed bind to the target cell's chitin and probably act by importing an effective toxin subunit, which however, differs structurally from zymocin γ and, strikingly, toxicity is independent from RNA-polymerase II Elongator. Consistently, the outcome of toxin action also differs for P. acaciae and W. robertsiae toxins when compared to zymocin: both former toxins provoke S-phase arrest and concomitant activation of the intra-S-phase DNA damage checkpoint (Klassen et al., 2004), whereas zymocin causes G1 arrest (Butler et al., 1991). The intra-S-phase DNA damage checkpoint can experimentally be activated by applying DNA damaging agents such as MMS (methyl methane sulphonate) and by replication fork stalling brought about by hydroxyurea, a potent ribonucleotide reductase inhibitor (Allen et al., 1994; Weinert et al., 1994; Elledge, 1996; Longhese et al., 2003). Such checkpoint activation results in transcription of damage inducible genes, and those of the environmental stress response as well as inhibition of late origin firing (Santocanale and Diffley, 1998; Gasch et al., 2001). As the novel type of killer toxin encoded by P. acaciae and W. robertsiae linear plasmids activates the checkpoint but still acts efficiently on checkpoint deficient cells, it possibly elicits DNA damages (Klassen et al., 2004). Here, we provide a detailed study of the cellular response of S. cerevisiae upon exposure to the P. acaciae toxin, which causes via a two-step process a time-dependent loss of viability.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Toxin treated cells loose viability in a two-step process

As killer assays for the P. acaciae toxin (Worsham and Bolen, 1990; Bolen et al., 1994; McCracken et al., 1994; Klassen et al., 2004) were based solely on measuring growth inhibition of S. cerevisiae, it still remained obscure as to whether it acts by reversibly inhibiting cell cycle progression or by provoking cell death as known for other killer systems (Butler et al., 1991; Schmitt and Breinig, 2002).

To broaden our knowledge of the toxin, we therefore had to determine the survival capability of S. cerevisiae LS20 and a corresponding chs3 mutant (Jablonowski et al., 2001a) in media containing selective concentrations of P. acaciae toxin [approximately fivefold minimum inhibitory concentration (MIC)] over an extended period of time, i.e. for 22 h. It is already known, that the P. acaciae toxin binds to chitin in vitro and loss of the major chitin synthase (Chs3) protects cells from the growth inhibiting effects of the toxin (Klassen et al., 2004). Consistently, in cultivations, the number of viable cells of the chs3 mutant increased despite presence of toxin, whereas viability of cells with an intact CHS3 allele (wt) drops during the first 4 h in toxin from 100% to approximately 30%, relative to the number of viable cells inoculated (Fig. 1A). Starting from the first reading point (5 min in toxin), loss of viability proceeded at a constant rate during this phase. Lethal events soon occur, even short contacts followed by removal of the toxin significantly affected viability (75% and 58% viability at 30 min and 1 h toxin exposure, respectively). Surprisingly, however, an intermediate survival rate of around 30% is maintained for a rather long period of time, i.e. for approximately 10 h (Fig. 1A). Subsequently, there is a final decline in viability, close to 1% after 22 h of toxin exposure (Fig. 1A).

image

Figure 1. Time-dependent loss of viability in toxin. A. Relative survival of S. cerevisiae LS20 (CHS3) and its Δchs3 derivative in medium containing P. acaciae toxin (5× MIC). B. Relative survival of S. cerevisiae KY117 arrested in G1, S or G2/M in the presence of P. acaciae toxin. Values given are means of triplicates and each experiment was repeated at least twice.

Download figure to PowerPoint

The lethal effect of P. acaciae toxin is cell cycle independent

FACS analysis previously revealed that S. cerevisiae transits from G1 into the S-phase in the presence of P. acaciae toxin. Cells arrest with partially replicated DNA and an activated intra-S-phase DNA damage checkpoint, as judged from Rad53 hyperphosphorylation. However, activating the Rad53 damage checkpoint cannot constitute the functional principle, as cells carrying the checkpoint deficient rad53-11 allele are not resistant but hypersensitive (Klassen et al., 2004). Hence, the problem arises whether the toxin's lethal action is bound to a particular phase of the cell cycle. To answer this question, the tester strain S. cerevisiae KY117 was pretreated with α-pheromone, hydroxyurea or nocodazole to induce G1, S or G2/M arrests, respectively. Subsequently, identical numbers of arrested cells were released to media containing the respective cell cycle arresting agent as well as P. acaciae toxin. Media containing the arresting compound only served as a control and changes in cell viability were determined over time in both, the cultures with the arresting agent and those containing in addition the toxin (Fig. 1B). Normalized to the toxin-free control, it became evident that G1 and G2/M arrested cells are very quickly killed by the toxin, whereas S-phase arrested cells are less rapidly affected. Thus, the toxin's lethal principle is not restricted to a specific stage of the cell cycle, moreover, viability decline of S-phase cells is slow compared to G1 or G2/M arrested cells. Such findings add up to a plausible explanation for S-phase arrest of asynchronously growing cells exposed to P. acaciae and W. robertsiae toxin.

Mutations accumulate in toxin survivors

As P. acaciae toxin speedily activates the DNA damage checkpoint in S. cerevisiae, toxin action may involve DNA damaging. Such action was checked directly at a specific site, i.e. by the canavanine resistance forward mutation assay, that was employed for determining mutation rates of the CAN1 gene (Whelan et al., 1979). For this purpose, we compared the number of canavanine resistant cells (can1R) arising in untreated cells and those exposed to P. acaciae toxin for 2–8 h. Spontaneous CAN1S to can1R mutation frequency was 1.35 per 106 viable cells in toxin free media, while toxin treated cells which survived toxin stress displayed a noticeably increased mutation frequency (8- to 15-fold) upon toxin exposure for 2, 4, 6 and 8 h (Fig. 2). Thus, DNA damaging occurs early during action of P. acaciae toxin and results in increased mutation rates measured at the CAN1 locus.

image

Figure 2. Mutation rates at the CAN1 locus with different toxin exposure times. The numbers of canavanine resistant colonies per 106 surviving cells are given. Values given are means of at least 7 independent experiments.

Download figure to PowerPoint

Toxin-mediated final decline of cell viability is reminiscent of apoptosis

FACS analysis of propidium iodide stained S. cerevisiae cells arrested with P. acaciae and W. robertsiae toxin revealed a significant proportion with a DNA content below 1C (Klassen et al., 2004), suggesting DNA fragmentation. Because a similar FACS profile was observed in S. cerevisiae during α-factor-mediated apoptosis (Severin and Hyman, 2002) and DNA fragmentation constitutes a hallmark of apoptosis per se (Madeo et al., 2002), it was tempting to check whether P. acaciae toxin treatment results in such programmed cell death.

We analysed the appearance of typical apoptotic markers in S. cerevisiae cells treated with P. acaciae toxin for 20–22 h. Among the final apoptotic cellular changes observable in yeast apoptosis is the cleavage of the nuclear scaffold, chromatin condensation and DNA fragmentation (Madeo et al., 2002). Severe changes in nuclear morphology, including elongation and fragmentation, can be observed in wild type (CHS3) S. cerevisiae toxin treated cells, but neither of the above is observed in similarly treated chs3 mutant nor in untreated cells of the wild type (Fig. 3).

image

Figure 3. DAPI stained cells of S. cerevisiae LS20 (CHS3) and its Δchs3 derivative, treated with P. acaciae toxin for 22 h (toxin) and untreated S. cerevisiae LS20 (no toxin).

Download figure to PowerPoint

Although DNA laddering during yeast apoptosis is not detectable on agarose gels (Madeo et al., 1997; Fröhlich and Madeo, 2000; Burhans et al., 2003) fragmentation of the chromosomal DNA correlates with cumulative appearance of 3′OH groups which are accessible to terminal deoxynucleotidyl transferase reactions. The frequently employed TUNEL- (terminal deoxynucleotidyl transferase-mediated dUTP nick end-labelling) test is based on in situ labelling of such accessible 3′OH groups with fluorescent dUTP (Madeo et al., 1997). Using wild type (CHS3) as well as chs3 cells treated with toxin and untreated wild type control cells, we observed a significant labelling of fixed protoplasts in the TUNEL assay, which depend clearly on toxin action and an intact CHS3 gene (Fig. 4), thereby supporting FACS based evidence for DNA fragmentation (Klassen et al., 2004).

image

Figure 4. DNA fragmentation analysis by the TUNEL test. Cells of S. cerevisiae LS20 (CHS3) and its Δchs3 derivative are shown, either treated with P. acaciae toxin for 22 h (toxin) or untreated S. cerevisiae LS20 (no toxin). Left panel: phase contrast micrographs of fixed protoplasts. Right panel: fluorescence of incorporated fluorescein-dUTP.

Download figure to PowerPoint

As for mammalian cells, phosphatidylserine is asymmetrically distributed in yeast membranes, being not detectable at the outer face (Cerbon and Calderon, 1991). Early during apoptosis, such asymmetry is disturbed and phosphatidylserine becomes exposed to the outer surface of the membrane, leaving, however, the cytoplasmic membrane functions unaffected. Flipping of phosphatidylserine has repeatedly been shown to occur during yeast apoptosis (Madeo et al., 1997; 1999; Ludovico et al., 2001). Thus, exposure of phosphatidylserine on the cell surface upon toxin treatment was checked by applying fluorescein conjugated annexinV, which binds to phosphatidylserine with high affinity. To assess membrane integrity, protoplasts were simultaneously stained with propidiumiodide, a dye, unable to pass intact membranes (Madeo et al., 1997; 1999; Ludovico et al., 2001). As shown in Fig. 5, there is a clear binding of annexin (green fluorescence) in membrane regions of toxin treated wild type cells. Exclusion of propidium iodide from the majority of annexin-positive cells confirmed that the annexin signal was not resulting from loss of cell integrity but represented appearance of phosphatidylserine on the outer leaflet of the plasma membrane. AnnexinV staining was never observed in chs3 cells treated with toxin nor in untreated controls, providing evidence that the detected exposure of phosphatidylserine in the wild type is a consequence of killertoxin exposure.

image

Figure 5. Analysis of phosphatidylserine exposure by staining with fluorescein-conjugated annexinV and propidium iodide to assess membrane integrity. Cells of S. cerevisiae LS20 (CHS3) and its Δchs3 derivative are shown, either treated with P. acaciae toxin for 22 h (toxin) or untreated S. cerevisiae LS20 (no toxin). Left panels: fluorescence of fluorescein-annexin superimposed on the phase contrast micrographs. Right panels: fluorescence of propidium iodide superimposed on the phase contrast micrographs.

Download figure to PowerPoint

Protein biosynthesis and generation of reactive oxygen species (ROS) are required for the final decline of viability

The appearance of intracellular ROS is one of the earliest markers detectable and functionally involved in yeast apoptosis (Madeo et al., 2002). For verification of ROS in toxin exposed cells, dihydrorhodamine123 was used, which via oxidization in the presence of intracellular ROS gives rise to a fluorescent derivative (Madeo et al., 1999). Staining of toxin treated wild type and chs3 cells (for 22 h) as well as untreated controls clearly revealed accumulation of ROS, a process, again strictly dependent on toxin treatment of cells carrying a functional CHS3 gene (Fig. 6A). However, accumulation of ROS is not detectable after 6 or 8 h in toxin when viability is already reduced to approximately 30% (Fig. 6B). As ROS formation depends on intact mitochondria and non-functional mitochondria were shown to suppress accumulation of ROS and concomitantly apoptosis (Kroemer and Reed, 2000; Ludovico et al., 2002; Severin and Hyman, 2002; Qi et al., 2003), we analysed the impact of mitochondrial dysfunction on toxin caused lethality. A rho- derivative of the toxin tester strain S. cerevisiae LS20, displayed clearly reduced ROS formation (after 22 h in toxin, Fig. 6B). However, such a strain is not protected with respect to the growth inhibitory function of the killer toxin (not shown) or loss of viability during the first 10 h in toxin (Fig. 6C), whereas, later on, compared to wild type survival of toxin treated cells significantly increases (Fig. 6C).

image

Figure 6. Involvement of reactive oxygen species (ROS) in the second phase of toxin action. A. Analysis of ROS formation by oxidation of dihydrorhodamine123. Cells of S. cerevisiae LS20 (CHS3) and its Δchs3 derivative are shown, treated either with P. acaciae toxin for 22 h (toxin) or untreated S. cerevisiae LS20 (no toxin). Left panel: phase contrast micrographs. Right panel: fluorescence of dihydrorhodamine123. B. Frequency of ROS positive cells in dependence of toxin exposure time and in the absence of intact mitochondria (rho-) or CHS3 gene. C. Time-dependent loss of viability in toxin, in the rho- strain, and in the presence of cycloheximide (CHX) compared to the wild type (rho+).

Download figure to PowerPoint

Apoptosis as an active cellular process requires protein biosynthesis, thus, cycloheximide has proven effective in suppression of yeast apoptosis phenomena (Madeo et al., 1999; Ludovico et al., 2001). For analysing the influence of cycloheximide on the time course of toxin-mediated cell death, to cells pretreated with cycloheximide subsequently toxin was added. As for the rho- strain displaying ROS-suppression, inhibition of protein biosynthesis by cycloheximide has no influence on the immediate loss of viability, but totally prevents the final decline of viability occurring in wild type cells exposed to the toxin for more than 10 h (Fig. 6C).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The P. acaciae killer toxin provokes S-phase arrest in S. cerevisiae and concomitant activation of the intra-S-phase DNA damage checkpoint, suggesting, the toxic effects may involve DNA-damaging (Klassen et al., 2004).

Here we show, that this novel type of killer toxin mediates a time-dependent loss of viability in S. cerevisiae. Because the toxin causes cell death in cell cycle phases other than S, and moreover, hydroxyurea arrested S-phase cells are partly protected from toxin lethality, it became evident that the toxic impact is not functionally restricted to a certain cell cycle phase. The S-phase arrest is therefore likely to be a collateral cellular response rather than a primary function of the toxin. Interestingly, G1 and G2/M arrested cells are killed more efficiently than asynchronously growing cells, while S-phase arrested cells are equally sensitive or, although slightly, more efficiently killed. Because checkpoint deficiency was previously shown to cause toxin hypersensitivity (Klassen et al., 2004), a likely explanation for the protection of S-phase arrested cells from lethal toxin effects is that the intra-S-phase checkpoint activation as a cellular response copes with the DNA damaging toxin. Possibly, S-phase arrested cells can partly antagonize toxin-mediated DNA damage by activation of the intra-S-phase checkpoint, which is disabled in G1 and G2/M arrested cells. Consistent with such assumption is the finding, that S. cerevisiae carrying the checkpoint deficient rad53-11 allele (GA1230; Shimada et al., 2002) is not protected from toxin lethality by pretreatment with hydroxyurea (data not shown). Furthermore, the cellular response to P. acaciae toxin strongly resembles effects of the DNA alkylating compound MMS which provokes S-phase arrest in wild type cells with concomitant activation of the intra-S-phase DNA damage checkpoint, whereas checkpoint deficient cells fail to arrest in S and display articulately increased sensitivity to MMS (Gasch et al., 2001).

Despite such evidence for DNA damaging mediated by the killer toxin, direct evidence for toxin-mediated DNA damage was lacking and that is why we checked the toxin's putative ability to induce mutations at the CAN1 locus. Indeed, mutation frequencies increased 8- to 15-fold compared to the untreated control when cells were exposed to P. acaciae toxin for 2–8 h, thus, providing direct evidence for a DNA damaging function. However, at present, it can hardly be decided whether DNA damage is the lethal principle or is an after effect of a hitherto unknown scenario causing cell death. Because DNA damage impairs cellular viability and DNA damage checkpoint activation as well as mutation induction are early events, toxin-mediated DNA damage likely contributes to the toxin's lethality.

Cell death of S. cerevisiae following extended toxin exposure (20–22 h) is accompanied by the occurrence of markers involved in apoptosis (Madeo et al., 2002; Burhans et al., 2003), such as changes in nuclear morphology (including fragmentation and elongation), fragmentation of the cellular DNA, appearance of ROS, and translocation of phosphatidylserine to the outer face of the plasma membrane. Because all such changes were seen to be toxin dependent, and loss of the cell's major chitin synthase (Chs3) not only protects from lethality but also suppresses apoptotic morphological changes, it became clearly evident that the chitin binding and DNA damaging P. acaciae toxin is a protein commiting S. cerevisiae to apoptotic cell death.

Interestingly, the toxin dependent loss of viability is a two-step process, in which viability is reduced to approximately 30% during the first 3–4 h, with such survival rate being maintained until a further decline of viability to approximately 1% is seen after 10–22 h in toxin. During the first 8 h, in which viability is significantly reduced and mutation frequencies are clearly increased, there is no detectable ROS formation. However, after 14 h, and more apparent after 22 h in toxin, reactive oxygen is formed, suggesting apoptosis not to be involved in initial steps affecting cellular viability but presumably being instrumental in the second (final) loss of viability seen after 10–22 h. Because ROS appear rather late during toxin action, not being detectable up to 6–8 h in toxin, they are unlikely the causative agents for the ascertained toxin-mediated DNA damages, which occur rather early (2 h) following toxin exposure. Consistently, a rho- derivative of the toxin tester strain LS20, displayed specific protection from the final loss of viability, whereas initial loss of viability (during the first 4 h) is indistinguishable from the wild type. Because the rho- strain was shown to exhibit a considerable reduction of ROS formation after 22 h in toxin, ROS are evidently functionally involved in the final decline of viability. Further evidence for apoptotic cell death occurring in long-term toxin treated cells is provided by the ability of cycloheximide to efficiently suppress such late decline of viability, whereas it fails to protect cells from the early toxin dependent loss of viability. Thus, inhibition of ROS formation and inhibition of protein biosynthesis by cycloheximide cause comparable effects, i.e. suppression of the final apoptotic cell death which occurs normally in cells treated with toxin for more than 10 h.

The P. acaciae killer toxin causes DNA damage checkpoint activation, induces mutations, and arrests cells in S phase. Taken into consideration that DNA damage and impaired DNA replication are known to induce apoptosis in S. cerevisiae (Del Carratore et al., 2002; Burhans et al., 2003; Qi et al., 2003), it is likely that DNA damage occurring early during toxin treatment finally triggers apoptosis in initially surviving cells. As mentioned before, the DNA alkylating agent MMS causes checkpoint activation and S-phase arrest in S. cerevisiae (Gasch et al., 2001), however, apoptotic cell death occurs only when cells are additionally defective in one of the subunits of the origin recognition complex (ORC) (Burhans et al., 2003). The latter forms pre-replicative complexes on chromatin in G1 and is required for establishment of replication forks in S-phase (Diffley, 2001). Lethality of the P. acaciae toxin is, however, clearly independent of DNA replication, as it efficiently kills cells arrested in G1 and G2/M.

Because of similarities of the effects caused by the P. acaciae toxin and DNA alkylating agents, both may share a similar lethal principle. Action of the P. acaciae toxin possibly involves cell cycle independent DNA modification leading to damage checkpoint activation in S, accompanied by 70% loss of viability of treated cells, with the surviving cells later undergoing an active cell death program. To our knowledge, the P. acaciae toxin represents the first example for a yeast killer toxin causing DNA damage and apoptotic cell death.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Strains, growth conditions, and preparation of killer toxin

Strains used in this study are listed in Table 1. Yeasts were routinely grown in YPD (1% yeast extract, 2% glucose, 2% peptone) at 30°C. For toxin production and the toxin assay YPD was adjusted to pH 6.9 and supplemented with 50 mM phosphate buffer (YPDPP). Crude killer toxin from P. acaciae was obtained as previously described (Klassen et al., 2004). A rho- derivative (LS20R-) of the toxin tester strain was obtained by growing S. cerevisiae LS20 in YNB minimal media (Difco, USA) containing 25 µg ml−1 ethidium bromide and selected by its inability to grow on medium containing 1% yeast extract, 2% glycerol, 2% peptone (Fox et al., 1991).

Table 1.  List of strains used in this study.
StrainDescriptionReference
Pichia acaciae NRRL Y-18665Wild type [pPac1-1+ pPac1-2+]Worsham and Bolen (1990)
Pichia acaciae PARK0NRRL Y-18665 [pPac1-1 pPac1-2]Klassen et al. (2004)
Saccharomyces cerevisiae LS20MATαleu2–3, 112 his3–11, 15 ura3Jablonowski et al. (2001a)
Saccharomyces cerevisiae LS20R-LS20, but rho-This study
Saccharomyces cerevisiae DY3LS20, but chs3::KlLEU2Jablonowski et al. (2001a)
Saccharomyces cerevisiae GA-1230MATa rad53-11 ade2-1 trp1-1 his3–11, 15 ura3-1 leu2-3112 can1-100 bar1::hisG ssd1-d2Shimada et al. (2002)
Saccharomyces cerevisiae KY117MATa ura3-52 trp1-Δ1 lys2-801amade2-101 his3-Δ200Butler et al. (1991)

Killer assay

Growth inhibition of S. cerevisiae by the P. acaciae killer toxin was determined by applying the microtitre plate assay (Klassen and Meinhardt, 2002). Serial dilutions of toxin preparation in YPDPP were used to determine the MIC. Two hundred microlitres of samples were inoculated with approximately 5 × 105 cells ml−1 and incubated in a microtitre plate at 30°C for 24 h. Subsequently, after adding 900 µl of water, turbidity was measured at 600 nm. As a control, cells were inoculated in media containing a mock preparation from a plasmid free P. acaciae strain (PARK0). Toxin preparations never exceeded 10% of the assay volume.

Survival tests

Fifty millilitres of YPD were inoculated with a single colony of the tester strain and grown overnight to early exponential phase. Three to five microlitres of this preculture were added to 200 µl of YPDPP containing toxin preparation (approximately fivefold MIC) or mock preparation in a microtitre plate. For determining the number of viable cells inoculated, 5 µl of mock treated control samples was removed immediately after resuspension and YPD was added to 1 ml of end volume followed by plating on YPD. For monitoring the effect of the toxin on cell viability, at different times 5 µl of the toxin containing culture was diluted 1:200 and plated likewise. Dilution leads to subinhibitory toxin concentrations in the medium. Colonies were counted following incubation for 48 h at 30°C. The same procedure was followed to analyse cells pretreated with cycloheximide (Sigma, Germany, 20 µg ml−1) for 1 h, which were subsequently inoculated in media containing cycloheximide and toxin. Relative survival was calculated as the ratio of the number of viable cells per millilitre in the toxin treated cultures referring to the number of inoculated viable cells per millilitre.

For monitoring toxin dependent loss of viability in cultures arrested at specific stages of the cell cycle, the preculture was pretreated for 2 h with 5 µg ml−1α-pheromone, 200 mM hydroxyurea, or 15 µg ml−1 nocodazole (Sigma, Germany) to induce G1, S or G2/M arrest, respectively. Three to five microlitres of arrested cultures were inoculated in 200 µl of YPDPP containing toxin plus the respective arresting agent and as a control YPDPP containing cell cycle arresting agents without the toxin. Viability determination was performed as outlined above. At timely intervals, both the dilutions of the toxin culture and the controls were plated to allow for determination of viability loss resulting from prolonged cell cycle arrest. Relative survival was calculated as the ratio of the number of viable cells per millilitre in cultures containing toxin plus cell cycle arresting agents referring to the number of viable cells per millilitre in the control samples, which contain cell cycle arresting agents exclusively. Viability tests were performed at least in triplicate.

Determination of mutation rates

Cells were left untreated or treated with killer toxin at different time intervals and were subsequently washed twice in sterile water. Following dilution, aliquots were plated on YPD to determine the titre of viable cells. For checking forward mutation rates at the CAN1 locus, cells were plated on YNB media containing canavanine (Sigma, Germany, 60 µg ml−1) as described by Morey et al. (2003). For each reading point, colonies were counted after 2 days at 30°C. Using freshly prepared cultures, this procedure was repeated at least sevenfold.

Detection of DNA breaks applying the TUNEL-assay

DNA fragmentation was detected using the terminal deoxynucleotidyl transferase-mediated dUTP nick end-labelling method essentially as described previously (Madeo et al., 1997; Fahrenkrog et al., 2004) employing the in situ cell death detection kit (Roche, Germany).

Cells were treated with toxin for 20–22 h, fixed in 3.7% formaldehyde for 30 min, converted to sphaeroplasts with zymolyase and applied to poly lysine coated slides (Madeo et al., 1997; 1999). The slides were rinsed with PBS (10 mM Na2HPO4/NaH2PO4, 150 mM NaCl, pH 7.5), incubated in permeabilization solution (0.1% triton X-100, 0.1% sodium citrate) for 2 min on ice and rinsed again with PBS. The labelling reaction was performed using 20 µl of TUNEL mixture (18 µl of TUNEL label containing fluorescein-dUTP, 2 µl of terminal deoxynucleotidyl transferase) for 1 h in a humid chamber in the dark. Subsequently, slides were washed with PBS and cells viewed under a fluorescence microscope.

Chromatin staining

Cells washed with PBS were incubated for 1 h in 70% ethanol. Following washing and rehydratization in PBS, cells were resuspended in PBS containing 1 µg ml−1 4,6 diamidino-2-phenylindole (DAPI, Sigma, Germany) and visualized under a fluorescence microscope.

Detection of intracellular ROS

Following incubation of cells at different times in toxin containing or toxin free media dihydrorhodamine123 was added at a concentration of 5 µg ml−1 and incubated for 2 h at 30°C (Madeo et al., 1999). Cells were resuspended in toxin free media containing the same concentration of dihydrorhodamine123 and finally observed under a fluorescence microscope.

Annexin V staining

To assess cellular integrity and exposure of phosphatidylserine, cells were converted to protoplasts and stained with propidium iodide and fluorescein-conjugated annexinV simultaneously (Annexin-V-FLUOS kit, Roche, Germany). Cells washed in sorbitol buffer (1.2 M sorbiol, 0.5 mM MgCl2, 35 mM potassium phosphate, pH 6.8) were protoplasted using zymolyase 20T (Seikagaku Corporation, Japan) and washed in annexin binding buffer (10 mM HEPES, 40 mM NaCl, 50 mM CaCl2, 1.2 M sorbit, pH 7.5). To cells, resuspended in 48 µl of annexin binding buffer, 1 µl of propidium iodide (PI) solution (Annexin-V-FLUOS kit, Roche, Germany) and 1 µl of fluorescein-annexin solution was added. Following incubation for 20 min in the dark, the cells were observed under a fluorescence microscope.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We thank Dirk Kemming, Münster for his help in establishing the TUNEL assay and John Paluszynski, Münster for reading the manuscript.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References