OUP user menu

Diagnosis of cell death induced by methylglyoxal, a metabolite derived from glycolysis, in Saccharomyces cerevisiae

Kazuhiro Maeta, Kaoru Mori, Yoshifumi Takatsume, Shingo Izawa, Yoshiharu Inoue
DOI: http://dx.doi.org/10.1016/j.femsle.2004.11.046 87-92 First published online: 1 February 2005


Methylglyoxal (MG) is a ubiquitous metabolite derived from glycolysis; however, this aldehyde kills all types of cell. We analyzed the properties of MG-induced cell death of the budding yeast Saccharomyces cerevisiae. The MCA1 gene encodes a caspase homologue that is involved in H2O2-induced apoptosis in yeast, although the disruption of MCA1 did not repress sensitivity to MG. In addition, the intracellular oxidation level did not increase under conditions in which MG kills the cell. Furthermore, the disruption of genes encoding antioxidant enzymes did not affect the susceptibility to MG. Here, we demonstrate that yeast cells killed by MG do not exhibit the characteristics of apoptosis in a TUNEL assay or an annexin V staining, but show those of necrosis upon propidium iodide staining. We demonstrate that MG at high concentrations provokes necrotic cell death without the generation of reactive oxygen species in S. cerevisiae.

  • Methylglyoxal
  • Yeast
  • Cell death
  • Necrosis
  • Reactive oxygen species

1 Introduction

Methylglyoxal (MG, CH3COCHO), a typical 2-oxoaldehyde, is a ubiquitous metabolite derived from glycolysis. The major source of MG in eukaryotic cells is a β-elimination reaction carried out by triosephosphate isomerase, a glycolytic enzyme [1,2]. Although a natural metabolite, MG reacts with macromolecules such as DNA, RNA and protein and subsequently disrupts cellular functions [3,4]. In fact, the overloading of MG has been demonstrated to kill all types of cells (for a review, see [3,4]). For example, MG induced apoptosis in Jurkat cells [5], but MOLT-4, HeLa and COS-7 cells died following MG treatment without exhibiting the apoptotic phenotype [6].

We have studied the metabolic route of MG using the budding yeast Saccharomyces cerevisiae as a model to gain insights into the physiological roles of MG, and found that MG is mainly metabolized to lactic acid through a glyoxalase system consisting of glyoxalase I and glyoxalase II. The former enzyme catalyzes the conversion of MG to S-d-lactoylglutathione in the presence of glutathione, while the latter hydrolyzes the glutathione thiol ester to d-lactic acid and glutathione. MG is also metabolized to lactic acid by a reduction/oxidation pathway, i.e., MG is reduced to l-lactaldehyde by NADPH-dependent methylglyoxal reductase and then oxidized to l-lactic acid by NAD+-dependent lactaldehyde dehydrogenase [7]. Although S. cerevisiae possesses two different metabolic routes from MG to lactic acid, yeast cells are most likely to use the glyoxalase system as the major metabolic pathway for this aldehyde, because a glyoxalase I-deficient (glo1Δ) mutant was hypersensitive to MG [8,9], but a methylglyoxal reductase-deficient mutant (gre2Δ) [10] was not [Maeta, K., Izawa, S. and Inoue, Y., unpublished data]. Additionally, a glyoxalase II-deficient (glo2Δ) mutant also exhibited sensitivity to MG [11].

In the cell death of Jurkat cells induced by MG, the generation of reactive oxygen species (ROS) is preceded by apoptosis [5]. On the other hand, it has been reported that yeast cells killed by H2O2 exhibit the characteristics of apoptosis [12]. The MCA1 (also known as YCA1) gene encodes a metacaspase in yeast, which is involved in apoptosis induced by H2O2, and therefore, the disruption of MCA1 reduces susceptibility to H2O2 [13]. To obtain further insight into the physiological role of MG, we examined the MG-dependent cell death of S. cerevisiae.

2 Materials and methods

2.1 Strains and media

All yeast strains of S. cerevisiae used in this study have the YPH250 background (MATatrp1-Δ1 his3200 leu2-Δ1 lys2-801 ade2-101 ura3-52). The construction of glo1Δ::HIS3 [14], gpx1Δ::HIS3, gpx2Δ::URA3, gpx3Δ::LEU2, tsa1Δ::TRP1 [15] and gsh1Δ::LEU2 [14] mutants was described previously. To disrupt the MCA1, PRX1, SOD1 and SOD2 genes, a part of the open reading frame was deleted and replaced with URA3, HIS3, LEU2 and TRP1, respectively, according to the method by Rothstein [16].

Media used were SD minimal medium (2% glucose, 0.67% yeast nitrogen base w/o amino acids; pH 5.5) with appropriate amino acids and bases, and YPD medium (2% glucose, 1% yeast extracts, 2% peptone; pH 5.5).

2.2 Spot assay

Cells cultured in YPD medium until A610 reached 0.1 were diluted serially (1:10) with a sterilized 0.85% NaCl solution, and each cell suspension (5 μl) was spotted onto YPD agar plates containing various concentrations of MG. Cells were cultured at 28 °C for 3 days.

2.3 MG treatment

Cells were cultured in SD medium until A610 reached 0.5, and various concentrations of MG were added. To determine whether newly synthesized protein(s) is necessary for the MG-induced cell death of S. cerevisiae, 10 μg/ml cycloheximide was added together with MG. Cells were incubated at 28 °C and a small portion of the culture was withdrawn periodically, diluted appropriately with a sterilized 0.85% NaCl solution, and spread onto YPD agar plates. Cells were cultured at 28 °C for 3 days, and surviving cells were counted.

2.4 Northern blotting

Cells were cultured in SD minimal medium until A610 reached approximately 0.8–1, and 0.8 mM H2O2 or 20 mM MG was added. Total RNA was prepared at the prescribed time by the method of Schmitt et al. [17]. To prepare the probe, MCA1 was amplified with the following primers; 5′-GGCCAGGTGATTGGATCCACCATCGACTAA-3′ and 5′-ATGCTTTGGAGACCAAGCATGCCAGAAAAG-3′. The resultant PCR fragment was digested with StyI and PstI. The DNA fragment containing the open reading frame of MCA1 (amino acid residue, 43–386) was labeled with [α-32P]dCTP.

2.5 TUNEL assay

A terminal deoxynucleotidyl transferase (TdT)-mediated dUTP nick-end-labeling (TUNEL) assay was performed essentially as described by Madeo et al. [18]. Briefly, cells were cultured in SD medium until A610 reached 0.5, then various concentrations of MG or 0.8 mM H2O2 were added (final volume, 5 ml), and the cells were incubated at 28 °C for 4 h. Cells were collected by centrifugation, and fixed with 5 ml of fixation buffer at room temperature for 1 h. Apoptotic cells were detected using a kit (In Situ Cell Death Detection Kit, AP, Roche Diagnostics) according to the manufacturer's specifications. Cells were observed with a fluorescence microscope (Olympus BX51).

2.6 Annexin V and propidium iodide staining

Annexin V staining and propidium iodide (PI) staining were performed using a kit (Annexin V-FITC Apoptosis Detection Kit, Bio Vision Research Products). Cells were cultured in SD medium until A610 reached 0.5, and various concentrations of MG or 0.8 mM (for annexin V staining) or 180 mM (for PI staining) H2O2 were added (final volume, 5 ml). After 4 h incubation, annexin V and PI staining were performed according to the manufacturer's directions.

2.7 Measurement of cellular oxidation levels

Cellular oxidation was measured using 2′,7′-dichlorofluorescin diacetate (DCFH-DA) as described by Davidson et al. [19]. Cells were cultured in 50 ml of SD medium in 200-ml flasks until A610 reached 0.5, and then various concentrations of MG or H2O2 were added. After 20 min at 28 °C, cells were collected by centrifugation, washed with the 0.85% NaCl solution, and resuspended in 10 ml of fresh SD medium containing 0.1 mM DCFH-DA. After 10 min, cellular oxidation levels were measured as described previously [15,20].

3 Results

3.1 MG kills yeast cells irrespective of ROS production

We monitored the effect of MG on the viability of yeast cells. As shown in Fig. 1(a), 8 mM MG temporarily inhibited the growth of wild-type cells but did not reduce viability, whereas glo1Δ cells died under the same conditions. At higher concentrations, wild-type cells died following MG treatment: approximately 90% died at 15 mM (data not shown) and 99% at 20 mM after 6 h. At a much higher concentration (50 mM MG), the viability of wild-type and glo1Δ cells rapidly decreased upon the addition of MG.

Figure 1

Effect of MG on cell death in yeast. (a) Wild-type (WT) and glo1Δ cells were treated with various concentrations of MG in SD medium. Open symbols, WT; closed symbols, glo1Δ. Concentrations of MG were: square, 0 mM; triangle, 8 mM; circle, 20 mM; and reverse triangle, 50 mM. (b) Wild-type (WT), mca1Δ, glo1Δ, and mca1Δglo1Δ cells were spotted onto YPD agar plates containing various concentrations of MG as indicated in the figure, and then cultured at 28 °C for 3 days. (c) Cells were cultured in SD medium until the log phase, and 0.8 mM H2O2 or 20 mM MG was added. Cells were collected periodically, and total RNA was prepared as described in the text. Twenty micrograms of RNA were applied to each lane. (d) Cellular oxidation levels following the treatment with 20 mM MG, 50 mM MG, 0.8 mM H2O2 or 180 mM H2O2 were measured using DCFH-DA as described in the text. The intensity of the fluorescence of the control cell (Con, untreated) was taken as 1.0.

In mammalian cells, ROS is one of the major causes of apoptosis [21]. Similarly, in S. cerevisiae, H2O2 induces apoptosis [12]. In the apoptotic process in mammalian cells, the activation of caspase is one of the critical steps. In S. cerevisiae, MCA1 encodes a homologue of caspase that is involved in H2O2-dependent apoptosis [13]; therefore, the disruption of MCA1 alleviates susceptibility to H2O2. However, susceptibility to MG was not affected by the disruption of MCA1 (Fig. 1(b)). Additionally, the introduction of MCA1 with a multicopy plasmid did not affect the susceptibility to MG (data not shown).

To further explore the role of Mca1 in cell death induced by MG, the expression of MCA1 was monitored. As shown in Fig. 1(c), mRNA levels of MCA1 were unchanged following the treatment with 20 mM MG. These results indicate that MCA1 is not likely to be involved in the MG-induced cell death.

Next, to determine whether ROS is formed during the death of yeast cells induced by MG, the cellular oxidation level was measured using the oxidant-specific fluorescent probe DCFH-DA. As shown in Fig. 1(d), the cellular oxidation level following MG treatment was substantially unchanged under conditions in which yeast cells died (viability after 20 mM MG treatment, 1.57 ± 0.78%). On the other hand, it markedly increased following treatment with H2O2 that yields almost the same killing rate as 20 mM MG (viability after 0.8 mM H2O2 treatment, 1.83 ± 0.45%).

To clarify the correlation between MG-induced cell death and ROS, we determined the susceptibility to MG of several mutants deficient in antioxidant enzymes. As shown in Fig. 2, no distinct differences were observed in susceptibility to MG between the wild-type and these mutants, except for a gsh1Δ mutant. The GSH1 gene product (γ-glutamylcysteine synthetase) is a rate limiting enzyme for glutathione biosynthesis and glyoxalase I, the major enzyme detoxifying MG, uses glutathione; therefore, the gsh1Δ mutant showed increased susceptibility to MG.

Figure 2

Effects of deficiency of antioxidant enzymes on susceptibility to MG. Cells of each mutant (TSA1, cytosolic thioredoxin perxidase; PRX1, mitochondrial thioredoxin peroxidase; SOD1, Cu,Zn-superoxide dismutase; SOD2, mitochondrial Mn-superoxide dismutase; GSH1, γ-glutamylcysteine synthetase; GPX1, GPX2 and GPX3, glutathione peroxidase) were spotted on YPD agar plates containing various concentrations of MG as indicated in figure. Cells were cultured at 28 °C for 3 days.

3.2 Newly synthesized proteins are not necessary for the yeast cell death induced by MG

Protein synthesis is involved in the apoptosis of many mammalian cells. Indeed, the H2O2-induced apoptosis of S. cerevisiae was reported to be repressed by cycloheximide [12]. Likewise, Ludovico et al. [22] reported that acetic acid induced the apoptosis of S. cerevisiae, which was blocked by cycloheximide. As shown in Fig. 3, cycloheximide did not repress the loss of viability among MG-treated cells. This result together with the data for the mca1Δ mutant suggests that MG-induced cell death does not involve apoptosis.

Figure 3

Effect of cycloheximide on viability following MG treatment. Wild-type cells were cultured in the SD medium until A610= 0.5, and then treated with 20 mM MG with or without 10 μg/ml cycloheximide (CHX). Cells were diluted appropriately, and spread on YPD agar plates to count the number of viable cells.

3.3 Diagnosis of MG-induced cell death

To determine more directly whether MG induces apoptosis, we examined cells treated with MG by TUNEL assay and annexin V staining. Madeo et al. [12] reported that the treatment of S. cerevisiae cells with 0.8 mM H2O2 induced apoptotic cell death. As shown in Fig. 4, we confirmed by TUNEL assay that yeast cells underwent apoptosis following exposure to 0.8 mM H2O2, i.e., chromosomal DNA strand breaks occurred and thus strong fluorescence derived from FITC-labeled dUTP was observed in the nucleus. Approximately 52% of cells (223/430) were found to be TUNEL-positive. On the other hand, only 4–5% of cells treated with 15–20 mM MG for 4 h were TUNEL-positive, of which approximately 90–99% died, a rate of death similar to that caused by 0.8 mM H2O2 (98.2% in our conditions). Approximately 7% (32/467) of untreated cells were also TUNEL-positive, hence the proportion of TUNEL-positive MG-treated cells is considered the background level.

Figure 4

Diagnosis of apoptosis. Wild-type cells were treated with 20 mM MG or 0.8 mM H2O2 for 4 h, and then subjected to a TUNEL assay and annexin V staining. Cells were observed by fluorescence microscopy. BF, bright field. Bars represent 5 μm.

Phosphatidylserine, a component of the cytoplasmic membrane, is externalized in the early stage of apoptosis and can be detected by annexin V [23]. Hence, we carried out the annexin V staining of cells treated with 20 mM MG or 0.8 mM H2O2. As shown in Fig. 4, the H2O2-treated cells were stained by FITC-labeled annexin V (approximately 55% of cells were stained), whereas the MG-treated cells were not. These results were well consistent with the data from the TUNEL assay. Taken together with our unpublished observation that genomic DNA fragmentation (DNA ladder) did not occur in cells treated with 20 mM MG, we conclude that apoptosis did not occur following the treatment with 20 mM MG for 4 h, even though the killing ratio was comparable with that induced by 0.8 mM H2O2.

Madeo et al. [12] reported that low concentrations of H2O2 (0.8 mM) caused apoptosis, whereas high concentrations of H2O2 (180 mM) induced necrosis. To determine whether high concentrations of MG induce necrosis, cells treated with 50 mM MG or 180 mM H2O2 for 4 h were stained with PI. As shown in Fig. 5, approximately 50% of the MG-treated cells were stained by PI. As was reported previously, the H2O2-treated cells were stained by PI.

Figure 5

Propidium iodide staining. Wild-type cells were treated with 50 mM MG or 180 mM H2O2 for 4 h at 28 °C and then subjected to propidium iodide (PI) staining. Cells were observed by fluorescence microscopy. BF, bright field. Bar represents 5 μm.

We measured oxidation levels of cells treated with 50 mM MG. As shown in Fig. 1(d), the cellular oxidation level was substantially unchanged following the treatment with 50 mM MG. These results suggest that higher concentrations of MG provoke a necrotic phenotype without the generation of ROS in yeast.

4 Discussion

In this study, we have demonstrated that yeast cells die following MG treatment irrespective of ROS production. In Jurkat cells, the generation of ROS is likely to be critical to the progression of apoptosis induced by MG [5]. Nakagawa et al. [24] reported that Jurkat cells are sensitive to H2O2 compared with other cell lines such as HeLa and U937. They have also reported that HL60 cells are sensitive to H2O [24] and undergo apoptosis following treatment with MG [25]. Taken together, MG treatment triggers the generation of ROS in some human leukemia cells, which in turn may induce the apoptosis of the ROS-sensitive cells. By contrast, in yeast, cellular oxidation levels did not increase following treatment with 20–50 mM MG (Fig. 1(d)). Additionally, the steady-state level of cellular oxidation in the wild-type and glo1Δ mutant was substantially unchanged (data not shown). Furthermore, mutants defective in some antioxidant enzymes did not show increased susceptibility to MG (Fig. 2).

Herker et al. [26] have reported that chronologically aged yeast cells die with the apoptotic phenotype. In cell death during long-term aging (>12 days old), the accumulation of ROS in cells is closely correlated with apoptosis. They have also reported that overexpression of YAP1, encoding a critical transcription factor responsible for the oxidative stress response in yeast [27], improved survival during chronological aging. On the other hand, we found that MG activates Yap1; however, ROS is not involved in the MG-dependent Yap1-activation system [28]. In addition, the expression of GLO1 increased after a diauxic shift, and concomitantly, the intracellular MG level decreased thereafter [28]. Therefore, cells in a stationary phase of growth are more resistant to MG than those in a log phase. Collectively, intracellular MG level seems little related to the production of ROS in yeast, and MG-induced cell death of yeast occurs irrespective of ROS.

The growth of wild-type cells was temporarily offset by 8 mM MG but cell viability did not decrease, whereas the viability of glo1Δ mutant cells dropped following the treatment with 8 mM MG (Fig. 1(a)); however, such cells were negative for TUNEL, annexin V staining and PI staining (data not shown). MG reacts with the guanine residue of DNA and RNA, and with the lysine and arginine residues of proteins to yield irreversible adducts that abolish function. The ability of glo1Δ cells to scavenge intracellular MG is impaired; therefore, the loading of 8 mM MG into glo1Δ cells may increase the cellular MG level causing the functions of intracellular components to be affected, leading to cell death. On the other hand, at higher MG concentrations, cells were stained with PI (approximately 50% at 50 mM MG, and 83% at 100 mM MG). In necrosis, the cytoplasmic membrane is injured by extracellular stimuli, thereby rendering it permeable to PI [29]. Hence, high concentrations of MG most likely culminate in damage to the membrane and a necrotic phenotype.

Necrosis is usually evoked by extreme nonphysiological stimuli. Indeed, the concentrations of H2O2 (180 mM) and MG (50 mM) used here are much higher than those in the cell under physiological conditions; however, the TNF-induced necrosis of murine fibrosarcoma L929 cells closely correlates with the increase in the intracellular MG level, which is linked to the phosphorylation of glyoxalase I [30]. This suggests that MG at metabolically produced levels may be involved in necrosis. Moreover, in necrosis, ion homeostasis and consequently cellular osmolarity are disrupted [31]. We have previously reported that the metabolism of MG and response to osmotic stress are closely correlated [7,14]. MG is a natural metabolite derived from glycolysis, and is involved in the pathology of various diseases including diabetes mellitus and complications thereof [4]; therefore, MG-induced cell death is of physiological interest.


This study was partially supported by grants from BRAIN.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
View Abstract