Department of Biotechnology, Instituto de Agroquímica y Tecnología de los Alimentos, Consejo Superior de Investigaciones Científicas, Poligono de la Coma, s/n, P.O. Box 73, Burjassot Valencia 46100, Spain
Jose Antonio Prieto
Department of Biotechnology, Instituto de Agroquímica y Tecnología de los Alimentos, Consejo Superior de Investigaciones Científicas, Poligono de la Coma, s/n, P.O. Box 73, Burjassot Valencia 46100, Spain
Methylglyoxal (MG), a glycolytic by-product, is an extremely toxic compound. This fact suggests that its synthesis and degradation should be tightly controlled. However, little is known about the mechanisms that protect yeast cells against MG toxicity. Here, we show that in Saccharomyces cerevisiae, MG exposure increased the internal MG content and activated the expression of GLO1 and GRE3, two genes involved in MG detoxification; GPD1, the gene for glycerol synthesis; and TPS1 and TPS2, the trehalose pathway genes. This response was specific as demonstrated by the analysis of marker genes and effectors of the general stress response. Physiological experiments with MG-treated cells showed that this compound triggers the overproduction of glycerol. Furthermore, a gpd1 gpd2 double mutant showed enhanced MG contents compared with the wild-type. Overall, these results appeared to indicate that up-variations in the intracellular content of the toxic compound are perceived by the cell as a primary signal to trigger the transcriptional response. In agreement with this, MG-instigated GPD1 activation was enhanced in strains lacking GLO1, and this effect correlated with the internal MG content. Finally, induction of GPD1, TPS1 and GRE3, and enhanced MG contents were also observed in low-glucose-growing cells subjected to a sudden increase in glucose availability. The implications of this regulatory mechanism on protection against MG are discussed.
Methylglyoxal (MG), an intrinsic intermediate of glycolysis , is an extremely toxic compound, able to react with, and modify, different molecular targets . MG is formed by fragmentation of free solution forms of both dihydroxyacetone phosphate (DHAP) and glyceraldehyde-3-phosphate (GA3P) . In bacteria, MG can also be formed biosynthetically by the action of MG synthase . The homologous enzyme from Saccharomyces cerevisiae has been isolated and characterised , but its sequence has not been reported, nor a gene encoding the enzyme has been localised.
Cells of different origin accumulate MG under physiological conditions resulting in a loss of control over carbohydrate metabolism. Thus, Escherichia coli cells overproduce MG by deregulation of the transport and metabolism of glucose . In bovine endothelial cells, incubation in 30 mM glucose increases MG . Enhanced MG levels upon osmotic and oxidative stress in mammalian [8,9] and yeast  cells have also been reported. Consistent with this, increased consumption of glucose during osmotic adaptation in yeast has been demonstrated .
In yeast, there are three established pathways for MG detoxification. First, MG conversion into l-lactaldehyde by the action of MG reductase . Second, the glyoxalase system, consisting in two enzymes, glyoxalase I and glyoxalase II (encoded by the GLO1 and GLO2 genes, respectively), whose activity converts MG into d-lactic acid in the presence of glutathione . Finally, the yeast aldose reductase, encoded by the GRE3 gene , which transforms MG into 1,2-propanediol in a two-step reaction, dependent on NADPH .
Expression of GLO1 is induced by osmotic stress and regulated by the high-osmolarity glycerol (HOG) response pathway , the main signalling cascade that contributes to the up-regulation of osmotically induced genes [14,15]. Transcription of GRE3 is up-regulated under a variety of stress conditions, and several factors have been found to control its expression, between them Msn2p and Msn4p , the two factors mediating the general stress response pathway in Saccharomyces. Stress-triggered induction of GRE3 increases aldose reductase activity, leading to a drop in the intracellular level of MG, a circumstance that is not observed in cells of gre3 and glo1 gre3 mutants .
The stress-instigated transcriptional activation of MG degradative pathways suggests that this regulatory mechanism could function in other physiological situations that result in MG overproduction. Thus, there is evidence that yeast cells activate MG-degradative pathways in response to exogenous MG, since a glo1 mutant shows a clear phenotype of MG sensitivity . The strong toxicity of MG also suggests that its synthesis should be tightly controlled. In bacteria, this control is established by allosteric regulation of the MG synthase activity . However, this control level remains unclear in eukaryotic cells because, in them, non-enzymatic formation appears to be the primary source of MG . Neither the overall protective mechanisms involved in the yeast response to increased MG levels, nor the metabolic signals mediating such response have still been elucidated.
In this paper, we show that expression of GRE3 and GLO1 genes from Saccharomyces is specifically up-regulated in response to increased intracellular MG levels. We also show that the same stimulus provokes the induction of GPD1, TPS1 and TPS2, three genes involved, directly or indirectly, in control of the level of glycolytic intermediates. Evidence suggesting that yeast cells sense intracellular MG levels is presented, and the implication of the nuclear factors Msn2p/Msn4p is discussed.
2 Materials and methods
2.1 Strains and culture conditions
The S. cerevisiae strains used in this work are listed in Table 1. Yeast cells were cultured in YNB minimal media [0.67% yeast nitrogen base without amino acids (DIFCO) plus 2% glucose] supplemented with the appropriate concentrations of essential nutrients . Cells were grown routinely in 250- or 1000-ml Erlenmeyer flasks at 30 °C on an orbital shaker (200 rpm).
Saccharomyces cerevisiae strains used in this study
Reference or source
MATa URA3 LEU2 trp1-289 his3-Δ1 MAL2-8cSUC2
MATa ade2-1 his3-11,15 leu-2-3,112 trp1-1 ura3-1 can1-100 GAL mal SUC2
W303 msn2 msn4
MATa ade2-1 his3-11,15 leu-2-3,112 trp1-1 ura3-1 can1-100 GAL mal SUC2 msn2-Δ3::HIS3 msn4 1::TRP1
W303 gpd1 gpd2
MATa ade2-1 his3-11, 15 leu-2-3, 112 trp1-1 ura3-1 can1-100 GAL mal SUC2 gpd1 ::TRP1 gpd2 ::URA3
MATa ura 3-52 his3-Δ200 leu2-Δ1 trp1-Δ1 lys2-801 ade2-101
Yeast Genetic Stock Center
MATa ura 3-52 his3-Δ200 leu2-Δ1 trp1-Δ1 lys2-801 ade2-101 glo1Δ::HIS3
For MG experiments, cells were grown to a mid-exponential phase (OD600=0.3–0.5) and MG (Sigma) was added at various final concentrations. For osmotic-shock treatment, cells were collected and shifted to 0.3 M NaCl-containing medium. Samples were taken at the indicated times.
2.2 RNA purification and Northern blot analysis
Total RNA from yeast cells was prepared as described . Equal amounts of RNA (30 μg) were separated in 1% (w/v) agarose gels, containing formaldehyde (2.5% v/v), transferred to a Nylon membrane and hybridised with 32P-labelled probes. Probes were obtained as follows: PCR-amplified DNA fragments containing sequences of GRE3 (whole ORF), GPD1 (+23 to +848), GLO1 (+124 to +506) and TPS1 (whole ORF) genes; 2.75 kb NotI fragment of plasmid pTPS2  containing TPS2 sequence; 1.1 and 1.3 kb fragments of the CTT1 gene sequence obtained by EcoRI restriction of plasmid pRB322-5109 ; 1.4 kb fragment of the HSP26 sequence obtained from a SphI/BglII restriction of the plasmid pVZ26 ; and HSP104 1.2 kb EcoRI fragment of plasmid pUZ1 . Probes were radiolabelled with the Random-primer Kit Ready-to-Go (Pharmacia Biotech) and [α32P]dCTP (Amersham). For control, filters were hybridised with a PCR-amplified probe of the IPP1 (whole ORF) gene . We used this gene, instead of ACT1 or rDNA, because preliminary experiments had indicated that its expression is not altered by addition of MG. The filters were analysed by autoradiography. For quantification, films were scanned and the images quantified with the QuantityOne software (BioRad) with no modification. Spot intensities were evaluated with respect to the mRNA level of IPP1 and presented as percentages of the maximal amount of induction.
2.3 Cell sampling and determination of metabolites
For MG assays, 40 ml of yeast culture (OD600=0.4–0.5) were filtered and the cell cake was resuspended immediately in 1.0 ml of acid extraction buffer (2.3 M HClO4, 90 mM imidazole). Then, the mixture was poured into a screw-cap Eppendorf tube containing 0.3 g of glass beads (acid-washed, 0.4-mm diameter) and frozen in liquid nitrogen. This protocol was found to be faster in obtaining the cell extracts than the traditional methanol-quenching procedure , especially for the large amount of sample volume (40 ml). Moreover, no significant differences were found in the MG measurements of control cells sampled with both methods (data not shown). Consequently, the latter was only chosen when removal of medium from the cells was required (MG-shocked cells). In this case, the yeast culture was centrifuged at 3000g for 2 min (−10 °C) and sprayed into 35 ml of buffered 60% methanol solution  kept at −40 °C. The mixture was centrifuged and the pellet was washed again under the same conditions. Finally, the cell cake was resuspended as above and frozen in liquid nitrogen. The suspension was allowed to thaw on ice and extraction was carried out by three freeze-thaw cycles with shaking in a Mini-bead beater (30 s at maximum speed) between each cycle. After centrifugation (5 min at 8000g, 0 °C), the supernatant was analysed immediately for MG content. For glycerol measurements, cells were collected by filtration. The filter was transferred quickly to a cold (4 °C) tube containing 1 ml of distilled water and the yeast suspension was boiled for 10 min, cooled on ice and centrifuged at 15,300g for 10 min (4 °C). Finally, the supernatant was collected and used for the assays.
Glycerol was measured using a standard commercial kit (Roche), following manufacturer's instructions. Methylglyoxal was determined by HPLC as 2-methylquinoxaline, according to the conditions described by Chaplen et al. . Samples (0.5 ml) were derivatised using 1,2-diaminobenzene, as described by Cordeiro and Ponces Freire , except that the reaction mixture was incubated for 3 h at room temperature. 5-Methylquinoxaline was used as internal standard. For quantification, a linear methylglyoxal calibration curve was obtained by plotting known methylglyoxal concentrations against ratios of analyte peak height to internal-standard peak height. An intracellular water value of 1.4 μl ml−1 per unit of OD600 was used to calculate the molar concentration of internal MG. For the trehalose assay, samples were taken and treated with trehalase as described in . The glucose liberated was measured colorimetrically, by the glucose oxidase and peroxidase method, in the presence of 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS, Sigma), following manufacturer's instructions.
3.1 MG triggers transcriptional activation of GRE3 and GLO1
We analysed by Northern blot the expression of GRE3 in cells of the wild-type yeast strain CEN.PK113-11A exposed to MG. As shown in Fig. 1(a), expression of GRE3 was induced by sub-lethal concentrations of MG. Like for many other genes activated by stress or chemical stimuli , the GRE3 response to MG was time- and dose-dependent, reaching a maximum level of mRNA after 60 min of exposure at 1 mM MG under the assay conditions (Fig. 1(a)). Furthermore, the MG-imposed transcriptional response was also observed for GLO1 (Fig. 1(b)), the gene for glyoxalase I.
Expression of GRE3 and GLO1 is up-regulated by MG. YNB-grown cells of the CEN.PK113-11A strain were exposed to a final concentration of 0.5 (○), 1 (△) or 2 (◻) mM MG (a), or 1 mM MG (b,c). Samples were taken at the indicated times and analysed for GRE3 (a) and GLO1 (b) transcript levels, or MG content (c). Northern blots and MG determinations were performed as described in Section 2. Values in (c) are given as means ± SD of three independent experiments. The graph in panel (a) represents quantification of the mRNA levels of GRE3, relative to those of IPP1.
We further investigated whether externally added MG was internalised. For this purpose, we analysed the kinetics of intracellular MG variations after addition of 1 mM MG (final concentration). As shown in Fig. 1(c), the internal concentration of MG showed an immediate increase when cells were exposed to the toxic compound, reaching a maximum level, around 60 μM, after 10–20 min of exposure. Later, the amount of intracellular MG decreased, with values at 40 min only threefold higher than the basal value (6.9 ± 1.4 μM). In contrast, external MG levels did not vary significantly during the time course of the experiment: 0.98 ± 0.3 mM just after the addition to 0.95 ± 0.2 mM at 40 min. Therefore, a shift to MG-containing medium causes a transient increase in the internal content of the toxic agent and the activation of genes implied in its degradation.
3.2 MG activates a specific transcriptional response
MG is a toxic compound able to cause cell death at high doses and to delay growth at sublethal concentrations. Therefore, the up-regulation of GRE3 and GLO1 by MG could be explained as a result of the activation by this metabolite of the general stress response. In order to investigate this possibility, RNA samples from cells exposed to 1 mM MG were hybridised with probes of HSP26, HSP104 and CTT1, three marker genes widely used in studies on the general stress response . As expected from previous reports, transcription of these genes was found to be fully derepressed upon osmotic- or heat-shock (Fig. 2(a)). However, their expression was completely indifferent to a 30-min treatment with 1 mM MG. Neither was induction found when longer times or higher MG concentrations were tested (data not shown).
MG exposure does not trigger the general stress response. (a) Cell cultures of the CEN.PK113-11A strain were heat-shocked (39 °C) or shifted to 0.7 M NaCl- or 1 mM MG-containing medium. (b) YNB-grown cells of the W303-1A wild-type (♢) and msn2 msn4 mutant (△) strain, were transferred to 4 mM MG-containing medium. Samples were taken after 30 min or at the indicated times and analysed by Northern blot as described in Section 2. Filters were probed for HSP26, HSP104 and CTT1 (a) or GRE3 (b) mRNA. The graph in panel (b) represents quantification of the mRNA levels of GRE3, relative to those of IPP1. (c) Cells from mid-exponential phase (OD600=0.3–0.4) were diluted from 1.0 to 10−3, spotted (5 μl) onto YNB plates containing MG at the indicated concentrations and incubated at 30 °C for 2–4 days.
This result suggested that MG does not activate the S. cerevisiae general stress response. In order to confirm this idea, we analysed the induction of GRE3 in a double-disruption mutant msn2 msn4 and the parental W303-1A strain. The transcriptional factors MSN2/MSN4 have been reported to mediate the general stress response in S. cerevisiae and in particular the activation by different stress conditions of GRE3 and GLO1[10,29,30]. As for the reference strain CEN.PK113-11A, MG-triggered activation of GRE3 was observed for cells of the W303-1A background, although maximal mRNA levels in this strain were found when cells were challenged to 4 mM MG (Fig. 2(b)). Under these conditions, induction of GRE3 was delayed, but not abolished, in the msn2 msn4 double mutant. Thus, the nuclear factors, Msn2p and Msn4p, appear to play an indirect role in the response to MG.
To further analyse this possibility, we tested the effect of the MSN2/MSN4 disruption on MG resistance. As shown in Fig. 2(c), no obvious phenotype was found when wild-type and msn2 msn4 mutant cells were spotted on plates containing 4 mM MG, a concentration that impairs growth in S. cerevisiae. Similar results were observed when a range of MG concentrations at 4–8 mM were tested (data not shown). Hence, the MSN2/MSN4 gene function is not essential to trigger the protective response to MG in S. cerevisiae.
3.3 Genes involved in glycerol and trehalose pathways respond to MG
The results described above indicated that yeast cells respond to MG by activating the expression of genes which provide high MG-degradative capacity. We were also interested to know if this protective response includes the control of MG formation. In S. cerevisiae, the synthesis of this metabolite appears to depend on the internal content of triose phosphates. Therefore, we analysed the MG-induced response of GPD1, the gene encoding glycerol-3-phosphate dehydrogenase [31,32]. This activity transforms DHAP in glycerol and, thus, redirects a part of the carbon flux away from glycolysis. As depicted in Fig. 3(a), addition of 4 mM MG to a W303-1A yeast culture induced the expression of GPD1 to a level similar to that observed for GRE3 (Fig. 2(b)). We also tested if the addition of MG could affect the transcript level of TPS1 and TPS2, the genes coding for the enzymes of the trehalose biosynthetic pathway [33,34]. Indeed, the expression of TPS1 and TPS2 was up-regulated by transferring cells to a MG-containing medium (Fig. 3(b)). Finally, we studied if the effects observed in the GRE3 expression by deletion of MSN2 and MSN4 could be extended to other MG-responsive genes. As shown in Fig. 3(a), the MG-induced up-regulation of GPD1 was again impaired, but not abolished, in the msn2 msn4 mutant strain. Similar results were observed for TPS1 (Fig. 3(b)) and TPS2 (Fig. 3(c)). In the latter, a full induction in the mutant strain was even observed after 60 min of MG exposure. Thus, Msn2p/Msn4p appears to affect the timing of the induction but not the magnitude of the response.
The glycerol and trehalose pathway genes are induced by MG. Cells of the W303-1A wild-type (♢) and msn2 msn4 mutant (△) strain, were transferred to 4 mM MG-containing medium. Samples were taken at the indicated times and analysed by Northern blot for mRNA levels of GPD1 (a), TPS1 (b) and TPS2 (c) as described in Section 2. Graphs represent mRNA levels relative to those of IPP1.
3.4 The level of MG depends on the activity of the glycerol pathway
The transcriptional activation of GPD1, TPS1 and TPS2 in response to MG suggested that the activity of glycolytic sinks at the upper part of glycolysis could regulate the formation of this compound. To test this hypothesis, we first examined the pattern of glycerol production following a shift to 4 mM MG-containing medium (Fig. 4(a)). As can be seen, the transfer triggered the overproduction of glycerol after 3 h of MG exposure. This observation is consistent with a role of the glycerol pathway in the control of MG synthesis. To further reinforce this idea, we analysed the levels of intracellular MG in a yeast strain lacking GPD1 and GPD2. GPD2, the gene homologous to GPD1, is responsible for the synthesis of glycerol under anaerobic conditions . Thus, no glycerol production takes place in a gpd1 gpd2 double mutant. As shown in Fig. 4(b), internal MG levels were significantly higher, 24.2 ± 1.8 versus 9.6 ± 1.3 μM, in the gpd1 gpd2 mutant than in the reference strain grown in YNB medium. These differences were much more pronounced after an osmotic-shock, a condition that stimulates MG formation . Indeed, osmotic stress caused a moderate and transient increase of MG in cells of the wild-type (Fig. 4(b)). This variation was also observed in NaCl-stressed cells of the gpd1 gpd2 mutant but, in this case, the overproduction of MG was maintained longer (Fig. 4(b)).
The levels of MG depend on the activity of the glycerol pathway. (a) Cells of the wild-type (W303-1A) strain were grown in YNB and exposed to 4 mM MG, final concentration (black bars). Untreated samples were used as control (grey bars). At the indicated times, samples were taken and analysed for intracellular glycerol. (b) YNB-grown cells of the W303-1A (grey bars) and gpd1 gpd2 mutant (black bars) strains were subjected to osmotic stress (0.3 M NaCl) and analysed for intracellular MG content. MG and glycerol assays were performed as described in Section 2. Values in panels correspond to a representative experiment. Three independent repetitions displayed the same tendencies.
We also tested if MG exposure could affect the level of trehalose. This disaccharide is produced from glucose-6-phosphate, so its synthesis could function, like that of glycerol, as a metabolic sink for intermediates at the upper part of glycolysis. However, overproduction of trehalose was not detected in MG-treated cells, even after 3 h of MG exposure (data not shown). Attempts to determine MG levels in a tps1 strain were also discontinued due to the inability of mutant cells to grow on glucose.
3.5 The MG response correlates with the intracellular content of the toxic agent
The results shown above indicated that MG addition enhances the intracellular levels of this compound, suggesting that this variation could be the signal to activate the transcriptional response. Further evidence about this was obtained by analysing GPD1 mRNA levels and MG content in mutant cells of MG pathways (YPH250 background), exposed to the same concentration of external MG. As expected, addition of MG to YNB-grown cells of the YPH250 wild-type strain caused a rapid increase in the intracellular MG content (Fig. 5(a)), although the maximum value reached, around 20 μM, was clearly much lower than that found for other reference strains (Fig. 1(c)). In consonance with this, the induction level of GPD1 was also weaker (Fig. 5(b)) than that depicted above (Fig. 2(a)). Interestingly, lack of GLO1 resulted in a more pronounced transcriptional activation of GPD1 (Fig. 5(b)). This effect correlated again with the intracellular level of MG, which in these cells reached values eightfold higher than those observed for the wild-type strain after 40 min of MG stimulation (Fig. 5(a)). Similar results were also found for the glo1 gre3 double disruption mutant (data not shown). Therefore, the increase of the intracellular MG level and not the mere presence of exogenous MG appeared to trigger the MG-induced transcriptional response. These results also suggested that the metabolism of MG was not required to activate this regulatory mechanism. To confirm this point, we tested if acetol or 1,2-propanediol, the main products of the aldose reductase pathway, were able to induce the expression of MG-responsive genes. None of these compounds caused any transcriptional activation, even when concentrations of 4–16 mM were analysed (data not shown).
Internal MG levels modulate the genetic response to the toxic agent. YNB-grown cells of the wild-type (YPH250, grey bars, ♢) and glo1 (YGL1, black bars, ◻) mutant strains were exposed to 4 mM MG (final concentration) and analysed for intracellular MG (a) and GPD1 mRNA levels (b). Samples were taken at the indicated times. MG determinations, total RNA preparation and mRNA quantification were carried out as described in Section 2. Values in (a) are given as means ± SD of three independent experiments. The graphs in (b) represent quantification of the mRNA levels of GPD1, relative to those of IPP1.
3.6 Enhanced glucose uptake triggers the MG response
Overproduction of MG has been reported to occur under conditions resulting in a loss of control over carbohydrate metabolism. According to this, it would be expected that the MG response could also take place in such situations, and not only after the addition of exogenous MG. To obtain further evidence about this idea, we measured the levels of MG and mRNA of MG-responsive genes in a yeast culture shifted from 0.2% to 2% glucose. It is well known from previous studies that glucose concentrations below 20 mM (i.e., 0.36%), activate the expression of the high-affinity glucose transport system  and that a sudden increase in glucose availability induces an accelerated glucose uptake . As shown in Fig. 6(a), the level of intracellular MG of low-glucose-grown cells displayed a sharp increase after the transfer, achieving its highest value, 17.3 ± 1.9 μM, at 10 min. Later, the amount of MG started to decrease reaching values around 10–11 μM after 20 min. This amount is similar to that observed for 2%-glucose-grown cells of the CEN.PK113-11A strain (Fig. 1(c), control). Correlated with this variation, we observed an induction in the expression of GPD1, TPS1 and GRE3 (Fig. 6(b)), suggesting again that the elevation in the MG levels triggers the transcriptional response. However, the three genes followed different induction kinetics. GPD1 up-regulation was early, around fivefold at 5 min, and remained high, threefold, during the time-course of the experiment. TPS1 showed a gradual induction, with a maximum, fourfold, at 40 min. GRE3 was scarcely induced, and no variation was found for the GLO1 mRNA.
Enhanced glucose uptake triggers the MG-dependent response. A YNB-(0.2% glucose) growing W303-1A yeast culture was pulsed with glucose (2%, final concentration). At the indicated times, cell samples were taken and analysed for MG content (a) or GPD1 (♢), TPS1 (◻), GRE3 (^) and GLO (△) mRNA levels (b). MG determinations and Northern-blot analysis were performed as described in Section 2. Values in (a) are means ± SD of three independent experiments. The graph in panel (b) represents quantification of the mRNA levels of each gene, relative to those of IPP1.
In this work, we show that MG exposure triggers a specific cell response, focused on a particular set of genes with functions in MG protection. GRE3 and GLO1 are two key elements of the MG detoxification mechanism in yeast, and lack of them produces MG sensitivity [4,10].
MG exposure activated also the expression of GPD1, TPS1 and TPS2. The up-regulation of GPD1 is logical if the main purpose of this activation is to control the level of triose-phosphates, from which MG is formed. As we demonstrated, lack of glycerol production in a gpd1 gpd2 double mutant resulted in enhanced MG contents in cells growing under non-stress conditions. Furthermore, this effect was more pronounced in cells subjected to osmotic shock, a stress situation that activates glucose metabolism  and glycerol formation . Consistent with this, overproduction of glycerol was stimulated in MG-exposed cells. Thus, our results suggest that up-regulation of glycerol synthesis provides a basic mechanism to control the intracellular content of MG.
We noted, however, that MG exposure activated the change in the glycerol pools only 3–4 h after MG addition, whereas GPD1 induction peaks at 30 min. Analysis of internal MG levels in MG-exposed cultures (Fig. 1) also demonstrated that yeast cells control the MG content within 30 min of MG exposure. Thus, these results cast doubts on the functional role of the GPD1 activation. However, we rationalise that this response could obey to the need to maintain a long-term potential of defence against MG. Indeed, similar features can be found in the well-known osmotic response of yeast. First, the maximum level of glycerol synthesis is observed around 2 h after the shift to high-osmolarity environments, whereas the highest induction of GPD1 is found much earlier, 15–30 min [23,31]. Second, although the osmotic equilibrium is reached in a relatively short time, the overproduction of the osmolyte is kept high till the osmo-stress condition is suppressed . As we showed, extracellular MG levels remained almost invariable after the addition of MG to a yeast culture. Thus, the induction of the glycerol pathway presumably enhances the cell's ability to manage a hostile high-MG environment after adaptation to growth under these conditions.
Nevertheless, the induction by MG of GPD1 could also provide additional mechanisms to control the level of this compound. Indeed, the onset of GPD1 expression, and in particular of TPS1 and TPS2, could suggest the existence of glycerol and trehalose cycles, since MG exposure did not trigger the biosynthesis of the disaccharide. Although the existence of these futile cycles is unclear, activation of glycerol and trehalose cycles has been claimed to play a protective role under different stress conditions by down-regulating carbon flux through the upper part of glycolysis [38–41]. Like in these situations, the activation of futile cycles in MG-exposed cells could provide a mechanism to control carbon flux by reducing the rate of MG formation. More work is, however, required to demonstrate the function of futile cycles in MG-shocked cells.
The existence in Saccharomyces of a protective response against MG suggested that dangerous internal MG levels could be perceived as a signal to trigger the transcriptional regulation. As we show, the exposition of yeast cells to exogenous MG transiently increased the internal content of the toxic agent. However, none of the main degradation products of MG, acetol or 1,2-propanediol, was able to induce the expression of MG-responsive genes. This suggested that no further metabolism of MG was required to activate the transcriptional response. In agreement with this, glo1 mutant cells showed enhanced expression of GPD1 when compared with the wild-type under the same stimulus. Moreover, this response correlated with higher levels of MG in the strain lacking GLO1 than in the reference strain. Thus, our results suggest that intracellular rather than extracellular MG is the signal to trigger the transcriptional response. In this respect, it is interesting to note that MG is not a substance that cells can easily encounter in the environment, at least at the concentrations assayed in this study. It seems unlikely, therefore, that a specific protective mechanism against external MG has been evolutionary developed.
Additional data evidencing that yeast cells could monitor alterations in the internal MG level were obtained by analysis of MG-responsive genes in low-glucose-growing cells pulsed with 2% glucose, a situation that sharply enhances glucose consumption , leading to a glycolytic overflux . As we demonstrated, a sudden increase in glucose availability caused MG over-accumulation. This is consistent with previous reports showing enhanced MG levels by overexpression of glycolytic genes, like PGI1 and PFK1. Again, this variation activated the expression of TPS1, TPS2 and GRE3. Moreover, this result indicates that the MG response is triggered not only by the external addition of MG, but also by small internal MG changes that do not detectably impair viability and growth. Indeed, the maximal value of internal MG reached after addition of 1 mM MG was around 65 μM for the CEN.PK113-11A strain, whereas a concentration not higher than 12–13 μM was detected after pulsing with glucose. Thus, these differences could account for the lack of induction of GLO1 under such conditions. Like for other stimuli and environmental conditions , the MG-response shows a dose- and gene-dependency.
Overall, the results reported here led us to propose a model for the MG-triggered defensive response (Fig. 7). High internal MG levels are interpreted by the cell as a signal that MG synthesis is too high and should be decreased. This triggers a dual response: first, enhancing the cell machinery to eliminate the excess of the toxic agent, and second, preventing new formation of MG by activating the glycerol and trehalose pathways. How the MG level is perceived as a signal and how it is transduced to induce the expression of genes are open questions. Our results indicate that MG exposure does not activate the general stress mechanism and that MSN2/MSN4, the two nuclear factors mediating this response in Saccharomyces, have only an indirect effect on the induction of MG-responsive genes. Similar results have been found for subsets of genes induced by different stress conditions, like heat shock or H2O2. In addition, no effects on MG resistance were found by MSN2/MSN4 deletion. Thus, our results implicate additional signal transduction pathways and regulators of the MG response.
The yeast's protective response against MG. An increase in the intracellular concentration of MG is perceived by the cell as a primary signal to trigger a specific response, focused on a subset of genes, GLO1 and GRE3, two key elements of the MG-detoxifying cellular machinery, and GPD1, TPS1 and TPS2, the genes for the glycerol and trehalose pathways. These could indirectly control de novo synthesis of MG by eliminating glycolytic intermediates or by down-regulating carbon flux (see text for details.)
Whether MG is able to interact directly with elements of a signal transduction pathway, or whether the mechanism is indirect is also unknown. Nevertheless, the ability of MG to modify proteins at the molecular level  make this metabolite an obvious candidate to modulate putative targets of signal transduction pathways.
We thank F. Randez-Gil and F. Estruch for the critical reading of this manuscript. We are also very grateful to Y. Inoue, K.-D. Entian and F. Estruch for providing the plasmids and yeast strains. This work was supported by the Comisión Interministerial de Ciencia y Tecnología (Project AGL2001-1203).
(1998) Expression of the glyoxalase I gene of Saccharomyces cerevisiae is regulated by high osmolarity glycerol mitogen-acivated protein kinase pathway in osmotic stress response. J. Biol. Chem. 273, 2977–2983.
(2000) The transcriptional response of Saccharomyces cerevisiae to osmotic shock. Hot1p and Msn2p/Msn4p are required for the induction of subsets of high osmolarity glycerol pathway-dependent genes. J. Biol. Chem. 275, 8290–8300.
(1994) GPD1, which encodes glycerol-3-phosphate dehydrogenase, is essential for growth under osmotic stress in Saccharomyces cerevisiae, and its expression is regulated by the high-osmolarity glycerol response pathway. Mol. Cell. Biol. 14, 4135–4144.
(1993) Disruption of TPS2, the gene encoding the 100-kDa subunit of the trehalose-6-phosphate synthase/phosphatase complex in Saccharomyces cerevisiae, causes accumulation of trehalose-6-phosphate and loss of trehalose-6-phosphate phosphatase activity. Eur. J. Biochem. 212, 315–323.
(1997) Effects of various types of stress on the metabolism of reserve carbohydrates in Saccharomyces cerevisiae: genetic evidence for a stress-induced recycling of glycogen and trehalose. Microbiology 143, 1891–1900.