Expression of the glyoxalase I gene of Saccharomyces cerevisiae is regulated by high osmolarity glycerol mitogen-activated protein kinase pathway in osmotic stress response.

Methylglyoxal is a cytotoxic metabolite derived from dihydroxyacetone phosphate, an intermediate of glycolysis. Detoxification of methylglyoxal is performed by glyoxalase I. Expression of the structural gene of glyoxalase I (GLO1) of Saccharomyces cerevisiae under several stress conditions was investigated using the GLO1-lacZ fusion gene, and expression of the GLO1 gene was found to be specifically induced by osmotic stress. The Hog1p is one of the mitogen-activated protein kinases (MAPKs) in S. cerevisiae, and both Msn2p and Msn4p are the transcriptional regulators that are thought to be under the control of Hog1p-MAPK. Expression of the GLO1 gene under osmotic stress was completely repressed in hog1Delta disruptant and was repressed approximately 80 and 50% in msn2Delta and msn4Delta disruptants, respectively. A double mutant of the MSN2 and MSN4 gene was unable to induce expression of the GLO1 gene under highly osmotic conditions. Glucose consumption increased approximately 30% during the adaptive period in osmotic stress in the wild type strain. On the contrary, it was reduced by 15% in the hog1Delta mutant. When the yeast cell is exposed to highly osmotic conditions, glycerol is synthesized as a compatible solute. Glycerol is synthesized from glucose, and a rate-limiting enzyme in glycerol biosynthesis is glycerol-3-phosphate dehydrogenase (GPD1 gene product), which catalyzes reduction of dihydroxyacetone phosphate to glycerol 3-phosphate. Expression of the GPD1 gene is also under the control of Hog1p-MAPK. Methylglyoxal is also synthesized from dihydroxyacetone phosphate; therefore, induction of the GLO1 gene expression by osmotic stress was thought to scavenge methylglyoxal, which increased during glycerol production for adaptation to osmotic stress.

Several environmental stresses are known to trigger intracellular alterations in organisms; e.g. synthesis of some stressinducible proteins or the cellular responses against extracellular signals. Organisms of all types show the synthesis of stressinducible proteins, and the most advanced understandings of the stress-inducible proteins have been obtained from the study of heat shock protein (HSP). 1 A sudden increase in the temperature of the environments in which cells are growing induces increased synthesis of a set of heat shock mRNAs and proteins in the cells. Heat shock response in eukaryotes is different from that in bacterial cells, although the basic mechanisms are similar among eukaryotes. A heat shock transcription factor, which is synthesized constitutively, is trimerized, modified, and translocated by heat shock and binds to a ciselement (the heat shock element) in the promoter region of HSP genes to activate transcription (for review, see Ref. 1). Some HSP genes in Saccharomyces cerevisiae, such as HSP104, HSP26, and HSP12, have another cis-element, termed stress response element (STRE) (5Ј-AGGGG-3Ј), in addition to the heat shock element (2,3). In addition to these HSP genes in S. cerevisiae, the CTT1 gene encoding cytosolic catalase has also been known to have the STRE in its 5Ј-flanking region (4). Expression of the CTT1 gene is induced by a wide variety of stresses, such as osmotic stress and oxidative stress, as well as heat shock, and these stress signals are thought to be focusing on the STRE (2). Stress response mechanisms through the STRE are fairy broad; therefore, elucidation of molecular mechanisms for stress response, including screening an appropriate specific marker gene, is of considerable interest.
In addition to heat shock response, one of the well-known stress response systems in S. cerevisiae is an osmotic stress response (5)(6)(7). Increased osmolarity of environment surrounding the cells induces rapid increase in expression of various kinds of genes. The osmosensing system has been extensively studied in a bacterial system, and existence of a two-component regulatory system was proved (8,9). S. cerevisiae also has a bacterial-like two-component osmosensing system consisted of Sln1p and Ssk1p (10 -13). When the yeast cell is exposed to a highly osmotic environment, rapid efflux of water from the cell shrinks cell and decreases its turgor pressure. As one of adaptive responses to the osmotic stress, the yeast cell produces glycerol as a compatible osmolyte. A key enzyme in biosynthesis of glycerol is glycerol-3-phosphate dehydrogenase, which is encoded by the GPD1 gene. The GPD1 gene is essential for survival under highly osmotic conditions (14). Expression of the GPD1 gene under highly osmotic conditions is under the control of Hog1p (14). Hog1p is one of the mitogen-activated protein kinases (MAPKs) in S. cerevisiae. Both residues of Thr174 and Tyr176 of the Hog1p are phosphorylated by a MAPK kinase (MAPKK), Pbs2p. Pbs2p is phosphorylated by redundant MAPKK kinases, Ssk2p and Ssk22p. The hog1⌬ knockout mutant showed lethality under highly osmotic conditions (2,15).
The substrate for glycerol-3-phosphate dehydrogenase (Gpd1p) is dihydroxyacetone phosphate, an intermediate of glycolytic pathway, and the enzyme catalyzes reduction of dihydroxyacetone phosphate to glycerol 3-phosphate in the presence of NADH. Some phosphatases, such as Gpp2p, hydrolyze glycerol 3-phosphate to glycerol (16). Dihydroxyacetone phosphate is also a substrate for methylglyoxal syn-thase that converts dihydroxyacetone phosphate to methylglyoxal (17)(18)(19)(20). Methylglyoxal is also synthesized during the triosephosphate isomerase reaction (␤-elimination) (21,22). Methylglyoxal (CH 3 COCHO) is a typical 2-oxoaldehyde in organisms, and it has two carbonyl carbons, i.e. ketone and aldehyde. Methylglyoxal can react with various biological compounds in cells, such as protein, DNA, and RNA, to inactivate them (23,24), and thus it shows cytotoxicity. To avoid overaccumulation of methylglyoxal, the cells from prokaryotes to higher eukaryotes have several defense systems. The glyoxalase system is a ubiquitous scavenging system for methylglyoxal. The glyoxalase system consisted of two enzymes, i.e. glyoxalase I (EC 4.4.1.5), which converts methylglyoxal to S-D-lactoylglutathione in the presence of glutathione, and glyoxalase II (EC 3.1.2.6), which hydrolyzes S-D-lactoylglutathione to D-lactic acid and glutathione. We have been systematically studying the glyoxalase system in various microorganisms, and we have proved that glyoxalase I is required for detoxification of methylglyoxal (25,26).
We found that the GLO1 gene had two STREs in its 5Јflanking region, and expression of the GLO1 gene was specifically induced by osmotic stress. In this paper, we also describe expression of the GLO1 gene in several gene disruptants, the gene products of which are involved in the HOG (high osmolarity glycerol) MAP kinase pathway, and discuss why the GLO1 gene is expressed under highly osmotic conditions.
Construction of GLO1-lacZ-To construct the GLO1-lacZ fusion gene, PCR primers were designed to amplify the 5Ј-flanking region of the GLO1 gene and 15 amino acid residues from N-terminal methionine. The PCR primers used were 5Ј-CTGAAGGAGGTGCCCGGGG-GATAAACCTAC-3Ј and 5Ј-ATCATTCCCGGGTTTCTCAATCTGAAT-TGG-3Ј. Both primers were designed to contain the SmaI site (shown by italic letters). A PCR fragment (992 bp) amplified using these primers was digested with SmaI and then introduced to SmaI site of pMC1871, which contained the lacZ gene of Escherichia coli without its original promoter region and N-terminal eight amino acids (31). The resultant plasmid (pMCGlac18) was digested with SalI, and the GLO1-lacZ cassette was cloned to the SalI site of the centromere plasmid pRS415 (32) to yield pRSGlac415.
Construction of Disruptants-The HOG1 gene with its 5Ј and 3Ј regions was cloned by PCR using primers HOG1S (5Ј-GTTGTTAG-GAAAGCATGCTTTATCTCCAAG-3Ј) and HOG1R (5Ј-CCTTTTATGG-GATCCTAATTTCTTAAGGAG-3Ј). Both primers were designed to contain recognition site for SphI for HOG1S and BamHI site for HOG1R, respectively; these sites are indicated by italic letters. The PCR fragment (2340 bp) was cloned between the SphI and BamHI sites of pUC19 to form pUCHOG1. To construct the hog1⌬::URA3 mutant, pUCHOG1 was digested with BalI and HincII, and a 400-bp fragment in the open reading frame of the HOG1 gene was replaced with the URA3 gene to yield pUHOG⌬Ura3. The resultant plasmid was then digested with SphI and BamHI, and then the hog1⌬::URA3 fragment was introduced to S. cerevisiae YIT2 (cta1⌬) (33) to construct the strain YYH1 (cta1⌬ hog1⌬::URA3). The MSN2 gene of YIT2 was disrupted using a plasmid pt32-⌬XB::HIS as described by Estruch and Carlson (27), and the disruptant was named YYM2 (cta1⌬ msn2⌬). The MSN4 gene was cloned by PCR using primers MSN4S (5Ј-CGCCACACCAACATG-CAACTTCTCCCAAGA-3Ј) and MSN4R (5Ј-GCTCTTCCAACCAAGC-CTCATTGCTCCTTG-3Ј). Primer MSN4S corresponded to the region between 533 and 562 bp downstream from the ATG codon, and the MSN4R corresponded to the region between 2653 and 2682 bp from the ATG codon of the MSN4 gene. The PCR fragment (2150 bp) was digested with SphI and EcoRI and then cloned between the SphI and EcoRI sites of pUC19. The resultant plasmid (pUCmsn4) was digested with EcoRV and AflII to delete the 593-bp fragment, which contained the zinc-finger motif of Msn4p (27), and then replaced with the URA3 gene to construct pUmsn4⌬Ura3. The plasmid was digested with SphI and EcoRI, and the DNA fragment containing the msn4⌬::URA3 cassette was introduced to strains YIT2 and YYM2 to construct YYM4 (cta1⌬ msn4⌬) and YYM24 (cta1⌬, msn2⌬, and msn4⌬), respectively. The GLO1 gene disruption plasmid (pUG⌬His3) was constructed as follows. A plasmid containing the GLO1 gene, pE24GLO1 (26), was digested with SphI, and the GLO1 gene was recloned to the SphI site of pUC19 to form pUGLO1. The plasmid was digested with EcoRV and HapI to delete the 850-bp fragment containing the open reading frame of the GLO1 gene, and it was replaced with the HIS3 gene to form pUG⌬His3. The pUG⌬His3 was digested with SphI, and the glo1⌬::HIS3 fragment was introduced to strain YGS1 (gsh1⌬::LEU2). The strain YGS1 was constructed by using the GSH1-disruption plasmid pYOG1211 (29) and strain YPH250 as a host. The double mutant of the GSH1 and GLO1 genes was named YGSL1 (gsh1⌬ glo1⌬). The yap1⌬ disruptant was constructed by disrupting the YAP1 gene of YPH250 using pSM27 (yap1⌬::HIS3) (28). Disruption of each gene was verified by PCR or Southern analysis, and transformation of yeast was carried out as described in our previous paper (26).
Construction of the YAP1-overexpressing Strain-A plasmid, pSc-Cer1, carrying the YAP1 gene (30), was digested with EcoRV and ClaI, and the fragment with the YAP1 gene was cloned to the BamHI site of YEp13 after blunting each cohesive end with a Klenow fragment. The resultant plasmid, YEp-YAP1, was introduced into S. cerevisiae YPH250. Overexpression of the YAP1 gene was confirmed by monitoring the growth on SD (2% glucose, 0.67% yeast nitrogen base) minimal agar plate containing 0.2 g/ml cycloheximide as reported by Hertle et al. (34).
Stress Experiments-Each strain of S. cerevisiae carrying pRS-Glac415 was cultured in a test tube containing 5 ml of SD medium (pH 5.5) with appropriate amino acids and bases at 28°C with shaking for 30 h. A small portion of the culture was transferred to a 200-ml flask containing 50 ml of YPD medium (2% glucose, 2% peptone, 1% yeast extract; pH 5.5) and cultured at 28°C with shaking for approximately 18 h. When the optical density of the culture at 610 nm (A 610 ) reached approximately 1.0, several chemicals, such as H 2 O 2 , NaCl, and ethanol were added, and then culture was continued for another 1 h. For heat shock experiment, the flask containing the culture was transferred to an incubator preheated to 37°C, and cultivation was continued for another 1 h. After stress treatment, cells were collected, and cell extracts were prepared as described below.
Preparation of Cell Extracts-Cells were collected by centrifugation (500 ϫ g at 4°C for 10 min), washed twice with 0.85% NaCl solution, and suspended in 300 l of 10 mM potassium phosphate buffer (pH 7.0). Cells were transferred to an Eppendorf tube containing an approximately equal volume of glass beads and then agitated with a vortex mixer at maximum speed for 3 min. Cell homogenates were centrifuged at 14,000 rpm for 15 min at 4°C, and resultant supernatants were used as cell extracts.
Enzyme Assay-Glyoxalase I activity was assayed as described pre-  (35). One unit of the activity was defined as the amount of enzyme forming 1 mol of S-D-lactoylglutathione per min at 25°C. Catalase activity was measured according to the method of Roggenkamp et al. (36). One unit of the activity was defined as the amount of enzyme decomposing 1 mol of H 2 O 2 per min at 25°C. ␤-Galactosidase activity was measured as described by Miller (37). One unit of the activity was defined as the amount of enzyme increasing A 420 by 1000 per min at 30°C. Protein was determined by the method of Bradford (38). Northern Blotting-Cells of S. cerevisiae YPH250 were cultured in YPD medium until A 610 reached approximately 1.0, and then 0.5 M NaCl was added. After 30 min incubation, total RNA was prepared according to the method of Schmitt et al. (39). To analyze the effect of the newly synthesized protein for expression of the GLO1 gene in osmotic stress response, cells were pretreated with 50 g/ml cycloheximide for 15 min before the addition of NaCl. RNA was separated by an agarose gel containing formaldehyde as described by Sambrook et al. (40). The GLO1 probe was prepared by digesting pUGLO1 with BamHI, and the resultant fragment (837 bp) containing the open reading frame of the GLO1 gene was purified by the low melting point agarose gel electrophoresis, and labeled by [␣-32 P]dCTP (Amersham Corp.) using a kit (Takara).
Measurement of Glucose-Cells of YIT2 and YYH1 were cultured in a 200-ml flask containing 50 ml of YPD medium until A 610 reached approximately 1.0, and NaCl was added to bring the final concentration of the culture to 0.5 M. 1.5 ml of culture was withdrawn periodically, and the A 610 and glucose concentrations of the culture were measured using a kit (Glucose B-test Wako).
Measurement of Methylglyoxal-Preparation of cell extracts was essentially same as described above, except that cells were suspended in distilled water instead of 10 mM potassium phosphate buffer (pH 7.0). Cell extracts were used as a source of methylglyoxal in glyoxalase I reaction. The reaction mixture (1 ml) contained 100 M potassium phosphate buffer (pH 7.0), 2 mM glutathione, 56 units/ml glyoxalase I (Sigma), and various concentrations of methylglyoxal, or cell extracts.
Chemicals-Methylglyoxal and glyoxalase I was purchased from Sigma. Glutathione was obtained from Kohjin (Tokyo, Japan). Restriction enzymes, DNA modification enzymes, and the DNA labeling kit were purchased from Takara Shuzo (Kyoto, Japan). [␣-32 P]dCTP (6000 Ci/mmol) and nylon membrane (Hybond-N) were obtained from Amersham-Japan (Tokyo, Japan). The glucose assay kit was purchased from Wako Chemicals (Kyoto, Japan). Plasmid pRS415 was purchased from Stratagene (La Jolla, CA).

RESULTS
Effect of Yap1p on Glyoxalase I Activity-Yap1p is one of the transcriptional regulators in S. cerevisiae, and it closely correlates with glutathione metabolism as well as drug resistance (41)(42)(43)(44)(45). Because detoxification of methylglyoxal is performed by glyoxalase I in the presence of glutathione, we supposed that expression of the GLO1 gene might be regulated by the Yap1p. We then constructed both of the YAP1 overexpressing strain and yap1⌬ disruptant and measured the glyoxalase I activity. However, glyoxalase I activity was not affected by the copy number of the YAP1 gene product (wild type, 0.128 Ϯ 0.015 units/mg protein; YAP1 overexpression, 0.145 Ϯ 0.018 units/mg protein; yap1⌬, 0.142 Ϯ 0.022 units/mg protein).
Expression of the GLO1 Gene Is Induced by Osmotic Stress-By an analysis of the 5Ј-flanking region of the GLO1 gene, we found two STREs 432 and 229 nucleotides upstream of the initiation (ATG) codon (Fig. 1). Several genes have been reported to have the STRE, such as CTT1, HSP104, HSP26, and DDR2, and expression of such genes is induced by several stresses, such as heat shock, oxidative stress, osmotic stress, ethanol stress, and so on (2). The CTT1 gene encodes cytosolic catalase, and it was widely used as a marker gene for experiments for stress response analysis through STRE; therefore, we also used CTT1 catalase as a marker. Because S. cerevisiae has another catalase in peroxisome that is encoded by the CTA1 gene (46), we disrupted the CTA1 gene to specifically monitor the CTT1 catalase activity. The GLO1-lacZ fusion gene was also constructed to quantify the expression of GLO1 gene under several environmental stresses. As shown in Fig. 2A, expression of GLO1-lacZ was specifically induced by NaCl, and other stresses that could induce expression of the CTT1 gene were not effective for induction of GLO1-lacZ expression. Glyoxalase I activity was also increased when the cells were .5% ethanol (EtOH) were added to the culture. For the heat shock experiment, a flask containing the culture was transferred to an incubator preheated to 37°C. Cells were then cultured for another 1 h, and cell extracts were prepared as described in the text. Black bars indicate ␤-galactosidase (GLO1-lacZ) activity, and white bars indicate CTT1 catalase activity. B and C, S. cerevisiae YIT2 carrying pRSGlac415 was cultured at 28°C in YPD medium until A 610 reached approximately 1.0, and 0.5 M NaCl, 0.5 M KCl, and 1.0 M sorbitol were added to the culture. Cells were then cultured for another 1 h, and cell extracts were prepared as described in the text. B, ␤-galactosidase (GLO1-lacZ) activity; C, CTT1 catalase activity. treated with 0.5 M NaCl, although other stresses did not affect the enzyme activity (data not shown). As shown in Fig. 2B, expression of GLO1-lacZ was induced by not only NaCl but also by other osmotic stresses, including sorbitol. Therefore, induction of GLO1-lacZ expression under high NaCl concentrations was thought to be caused by osmotic stress rather than salt stress. Fig. 3 showed the comparison of optimal concentration of NaCl for induction of GLO1-lacZ and CTT1. Induction of both gene expression was observed between the range of 0.3 M and 0.7 M NaCl. Maximum induction of GLO1-lacZ was observed at 0.5 M NaCl, whereas that of CTT1 was observed at 0.3 M. At 1.4 M NaCl, no induction was observed in both genes.
Northern Blotting-To confirm that an increase of ␤-galactosidase activity from the GLO1-lacZ fusion gene was due to the increased expression of the GLO1 gene, Northern blot analysis was done. Because the optimal concentration of NaCl for induction of GLO1-lacZ was 0.5 M (Fig. 3), S. cerevisiae YPH250 was treated with 0.5 M NaCl, and the mRNA level of the GLO1 gene was measured. As shown in Fig. 4, the GLO1 mRNA level increased by osmotic stress. Therefore, increased activity of ␤-galactosidase from GLO1-lacZ by osmotic stress was confirmed to reflect the expression of GLO1 gene. To investigate whether de novo synthesis of some factors is required for induction of the GLO1 gene transcription, the cells were pretreated with cycloheximide to block protein synthesis and then exposed to osmotic stress. As shown in Fig. 4, mRNA level of the GLO1 gene was also increased, even though protein synthesis was blocked. Therefore, induction of GLO1 gene expression was not dependent upon the newly synthesized protein(s) during osmotic stress treatment.
Involvement of GLO1 Gene Expression in HOG-MAPK Pathway-S. cerevisiae has been reported to have a bacterium-like two-component system for sensing high osmolarity of environments. Sln1p and Ssk1p are, respectively, a sensor kinase for high osmolarity and a regulator to mediate the signal to the redundant MAPKK kinases Ssk2p and Ssk22p. Activated MAPKK kinases phosphorylate the Pbs2p MAPKK. The Pbs2p can also receive a high osmolarity signal from the second osmosensor, Sho1p. The activated Pbs2p then phosphorylates Hog1p-MAPK. Therefore, in both pathways, signals focus on Hog1p. The Mig1p-like zinc finger proteins Msn2p and Msn4p are thought to be under the control of Hog1p in osmotic stress response. We then disrupted HOG1, MSN2, MSN4, and both MSN2 and MSN4 genes by gene replacement in the cta1⌬ background and measured expression of the GLO1 and CTT1 genes under highly osmotic conditions. As shown in Fig. 5A, induction of the GLO1 gene expression under highly osmotic conditions was almost completely repressed in the hog1⌬ mutant (YYH1). In the msn2⌬ mutant (YYM2), induction was approximately 20% compared with strain carrying the wild type MSN2 allele (YIT2). On the other hand, the msn4⌬ mutant (YYM4) could induce approximately 50% more expression of the GLO1 gene than YIT2. In the case of double mutant of the MSN2 and MSN4 genes (YMM24), induction of GLO1 expression was not observed. These results strongly suggest that expression of the GLO1 gene under highly osmotic conditions is controlled by the HOG-MAPK pathway. On the other hand, expression pattern of the CTT1 gene was different from that of the GLO1 gene (Fig. 5B). In the strain YYH1 (hog1⌬), induction of the CTT1 gene expression under highly osmotic conditions was extremely reduced compared with that of YIT2, although the CTT1 was still induced by osmotic stress. In the strains YYM2 (msn2⌬) and YYM4 (msn4⌬), the CTT1 gene expression was also induced as observed in the case of GLO1-lacZ; however, the most conspicuous difference was seen in the behavior of the msn2⌬ msn4⌬ double mutant (YMM24). The CTT1 gene was still induced in the strain YMM24 when the cells were exposed to high osmotic stress.
Effect of Osmotic Stress on Consumption of Glucose-S. cerevisiae synthesizes glycerol as a compatible solute when the To block the protein synthesis, cycloheximide (CHX) was added prior to the addition of NaCl. cells are exposed to highly osmotic environments. Glycerol is synthesized from glucose through dihydroxyacetone phosphate and glycerol 3-phosphate. On the other hand, methylglyoxal is also synthesized from dihydroxyacetone phosphate by methylglyoxal synthase (17)(18)(19)(20) or from triosephosphate isomerase reaction (␤-elimination) (21,22) during catabolism of glucose. Therefore, glucose consumption is expected to be increased to synthesize glycerol during the adaptive response to highly osmotic environments. We measured the concentration of glucose in culture medium to evaluate the consumption speed of glucose by yeast cells. When the cells were exposed to highly osmotic conditions, intracellular water efflux was increased, and it resulted in causing cell shrinkage. Therefore, the A 610 of the culture suddenly increased (Figs. 6, A and B); however, it recovered during the adaptation period (5-30 min), and growth was restarted (60 -120 min). The generation time of the strain YIT2 (HOG1) without osmotic stress was 295.6 Ϯ 5.22 min (0 -120 min), whereas it increased to 375.6 Ϯ 13.7 min after recovery from cell shrinkage (60 -120 min). In the case of strain YYH1 (hog1⌬), generation time (276.3 Ϯ 7.74 min) without osmotic stress during the 0 -120-min period was similar to that of YIT2. On the other hand, once the cells were exposed to high osmolarity, generation time during the 60 -120 min period increased to 589.6 Ϯ 18.7 min. During the 5-30-min period, yeast cells seemed to recover the cell size to adapt osmotic pressure. Glucose consumption speed of YIT2 under the osmotic conditions during the 5-30-min period was approximately 30% higher than that during the same period without osmotic stress (Fig. 6C). After recovery of the growth (60 -120 min), glucose consumption rate of YIT2 under highly osmotic conditions was 10% higher than that of the same strain during the 60 -120-min period. In the case of hog1⌬ mutant (YYH1), the glucose consumption rate during the 5-30-min period was decreased approximately 15% under highly osmotic conditions compared with that in the same strain in normal osmolarity. The glucose consumption rate of the hog1⌬ cells treated or not treated by NaCl during the 60 -120-min period was similar (Fig. 6D).
We then investigated whether intracellular concentration of methylglyoxal increased when the cells were exposed to highly osmotic environments. Methylglyoxal rapidly reacts with glutathione nonenzymatically to form hemimercaptal (47), and the compound is the intrinsic substrate for glyoxalase I; i.e. glyoxalase I catalyzes conversion of hemimercaptal to S-D-lactoylglutathione. Therefore, it has been thought that methylglyoxal may not exist free in the cell (25). Furthermore, because methylglyoxal is toxic, it is rapidly detoxified by glyoxalase I (k cat ϭ 4.53 ϫ 10 4 min Ϫ1 ) (26). We thus disrupted both GSH1 and GLO1 genes to block degradation of methylglyoxal. The gsh1⌬ mutant cannot produce glutathione at all (29), and no glyoxalase I activity was detected from the glo1⌬ cell (26). The steady state level of methylglyoxal in the YGSL1 (gsh1⌬ glo1⌬) cell without treatment by NaCl was 2.24 mol/g-wet cell, whereas that in the cells treated with 0.5 M NaCl for 1 h was 2.75 mol/g-wet cell. Fold increase of intracellular methylglyoxal content was similar to that of glucose consumption rate after osmotic stress. However, we could not detect methylglyoxal content in the wild type strain (YPH250) using the enzymatic method that we adopted in this study. This was presumably due to rapid degradation of methylglyoxal by the actions of Glo1p and glutathione.

Regulation of GLO1 Gene Expression in Osmotic Response by the HOG-MAPK Pathway-The YAP1/PAR1/PDR4/SNQ3
gene product is a transcriptional activator. A key enzyme for glutathione biosynthesis is the GSH1 gene product, ␥-glutamylcysteine synthetase, and expression of the GSH1 gene is positively regulated by the Yap1p. Glutathione recycling is also important for cells to maintain the intracellular redox state, and the GLR1 gene encoding glutathione reductase is also under the control of Yap1p. Glutathione is involved in the glyoxalase I reaction. Furthermore, overproduction of Yap1p renders yeast cells to resist against many structurally unrelated drugs, such as 1,10-phenanthroline (PAR1) (48), cycloheximide (PDR4) (49), nitrosoguanidine (SNQ3) (34), cadmium (YAP1) (28), and H 2 O 2 (YAP1) (30). Because methylglyoxal is a metabolic cytotoxic compound in the cell, we thought that expression of the GLO1 gene may also be regulated by Yap1p. However, no alterations of glyoxalase I activity were observed in the yeast cells overexpressing the YAP1 gene or in the yap1⌬ mutant. On the other hand, we found two STREs in the 5Јflanking region of the GLO1 gene, and expression of the GLO1 gene was specifically induced when the cells were exposed to highly osmotic environments. Gounalaki and Thireos (50) reported that the TPS2 gene encoding trehalose phosphate phosphatase had the STREs in its promoter region, and expression of the TPS2 gene was induced by several stresses, such as heat shock, osmotic stress, and metabolic inhibitors. They also reported that Yap1p was required for transcriptional regulation of the TPS2 gene through the STRE. Yap1p is a functional homolog of mammalian AP-1, and the consensus sequence for recognition site of the Yap1p (Yap1p recognition element) was reported to be 5Ј-TTAGT(C/A)A-3Ј (51). The Yap1p recognition element sequence was not found in the promoter region of the GLO1 gene, and the increased copy number of the Yap1p in the cells did not affect the glyoxalase I activity. Furthermore, glyoxalase I activity was increased in the yap1⌬ mutant by osmotic stress (data not shown). Therefore, we concluded that expression of the GLO1 gene was not dependent upon the Yap1p.
Expression of the CTT1 gene encoding cytosolic catalase has been known to be induced by many stressful conditions, such as heat shock, osmotic stress, oxidative stress, ethanol, sorbate, low pH, and so on (2,3). The CTT1 gene has STREs in its promoter region, and such environmental signals have been thought to focus on the STRE (2). Instead, we confirmed that CTT1 catalase activity was increased by heat shock, osmotic stress, oxidative stress, and ethanol (Fig. 2). On the other hand, expression of the GLO1 gene, which had two sets of STREs, was specifically induced by osmotic stress. However, at 1.4 M NaCl, no induction was observed in either the CTT1 or GLO1 gene. Recently, Norbeck and Blomberg (52) identified several genes, the expression of which was enhanced by 1.4 M NaCl in S. cerevisiae. They presented a putative cis-element that responded to 1.4 M NaCl stress, i.e. they proposed 5Ј-TATGC-CTCT-3Ј as the consensus sequence. Neither the CTT1 nor the GLO1 gene has this consensus sequence in the promoter region. Expression of the CTT1 gene under highly osmotic conditions was reported to be regulated by HOG-MAPK pathway via STRE, and zinc finger proteins Msn2p and Msn4p proved to be required for transcriptional activation (53). As shown in Fig.  5, catalase activity was not enhanced in the hog1⌬ mutant as much as in the cells carrying the wild type HOG1 allele. In the strain YMM24 (msn2⌬ msn4⌬), catalase was still induced by osmotic stress; however, the GLO1 gene expression was almost completely repressed in the YMM24 mutant, as well as in the hog1⌬ mutant. These results suggested that some factors other than Msn2p and Msn4p may be concerned with transcriptional regulation of the CTT1 gene in osmotic stress response in S. cerevisiae; however, such factors might not be involved in osmotic response of the GLO1 gene. Because expression of the GLO1 gene specifically corresponds to osmotic stress and is involved in the HOG-MAPK pathway, the GLO1 gene could be a good model for analysis of MAPK cascade in S. cerevisiae. Physiological Significance for Induction of the GLO1 Gene in Osmotic Stress Response-As shown in Fig. 6, the glucose consumption rate transiently increased by osmotic stress in the wild type strain. Norbeck and Blomberg (52) reported that expression of both of the HXT1 gene, encoding hexose transporter, and the GLK1 gene, encoding glucokinase, was enhanced under highly osmotic conditions. S. cerevisiae synthe-FIG. 7. Speculative model for induction of the GLO1 gene under highly osmotic conditions. Expression of the HXT1 (hexose transporter), GLK1 (glucokinase), and GPP2 (glycerol-3-phosphate phosphatase) genes has been reported to be enhanced by osmotic stress (16,52). The GPD1 gene, which encodes glycerol-3-phosphate dehydrogenase, is under the control of Hog1p. Influx of glucose is increased if S. cerevisiae is exposed to highly osmotic environments, and the glucose is metabolized to dihydroxyacetone phosphate. The Gpd1p catalyzes reduction of dihydroxyacetone phosphate to glycerol 3-phosphate in the presence of NADH, and the Gpp2p catalyzes dephosphorylation of glycerol 3-phosphate to glycerol. On the other hand, methylglyoxal is synthesized from dihydroxyacetone phosphate by methylglyoxal synthase or by ␤-elimination in triosephosphate isomerase reaction. The GLO1 gene is expressed by the HOG-MAPK pathway to scavenge methylglyoxal that may be increased by osmotic stress. FBP, fructose-1,6-bisphosphate; DHAP, dihydroxyacetone phosphate; G3P, glyceraldehyde 3-phosphate; MG, methylglyoxal; S-LG, S-D-lactoylglutathione; and GSH, glutathione (reduced form). sizes glycerol as a compatible solute when the cells are exposed to highly osmotic environments. Glycerol is synthesized via glycolysis, and the rate-limiting step for glycerol production is a glycerol-3-phosphate dehydrogenase reaction. Glycerol-3phosphate dehydrogenase is encoded by the GPD1 gene, and the gene expression under highly osmotic conditions is regulated by Hog1p (14). The gpd1⌬ mutant cannot grow in a medium containing a high concentration of NaCl (14). Expression of the GPP2 gene, encoding glycerol-3-phosphate phosphatase, was also reported to be enhanced by osmotic stress (16). To adapt to high osmolarity, S. cerevisiae cells produce glycerol from glucose, and dihydroxyacetone phosphate is a precursor. On the other hand, methylglyoxal is also synthesized from glycolysis, and dihydroxyacetone phosphate is a substrate for methylglyoxal synthase (17)(18)(19)(20). Methylglyoxal is also inevitably produced from the triosephosphate isomerase reaction (21,22). Therefore, an increased flux of glucose to glycolysis may cause the enhancement of intracellular methylglyoxal content. Actually, we found that the steady state level of methylglyoxal in the gsh1⌬ glo1⌬ mutant cells treated by 0.5 M NaCl increased approximately 23% compared with that in the untreated cells. Glucose consumption was also increased approximately 30% by osmotic stress (Fig. 6C). Therefore, the physiological purpose for increasing GLO1 gene expression under highly osmotic conditions may be to scavenge methylglyoxal that is increased in the adaptive response to high osmolarity (Fig. 7).
We also speculated about the physiological reason that the CTT1 gene is also expressed in osmotic stress response. Under aerobic conditions, NADH in cytosol cannot transfer electrons to acetaldehyde in the alcohol dehydrogenase reaction; however, dihydroxyacetone phosphate can accept electrons from NADH by the action of Gpd1p, and dihydroxyacetone phosphate is reduced to glycerol 3-phosphate. Glycerol 3-phosphate can pass through the mitochondrial membrane and transfer electrons to FAD to form FADH 2 , and glycerol 3-phosphate itself is oxidized to dihydroxyacetone phosphate. Dihydroxyacetone phosphate in mitochondria is then transformed back to cytosol (via the glycerol phosphate shuttle). An electron is transferred to ubiquinone (CoQ) from FADH 2 in the electron transport chain; therefore, increased influx of glycerol 3-phosphate from cytosol to the mitochondria may cause enhancement of flux of the electron transfer in respiratory chain, and it may increase occurrence of the reactive oxygen species. The SOD2 gene encodes Mn-superoxide dismutase (Mn-SOD), and Mn-SOD (Sod2p) is localized to mitochondria. Mn-SOD is one of the antioxidant enzymes in this organelle. We found that the sod2⌬ mutant could not grow on the YPD agar plate containing 0.7 M NaCl. 2 H 2 O 2 is also reactive oxygen species, and it can freely pass the biological membrane; therefore, expression of the CTT1 gene, which encodes cytosolic catalase, is induced by osmotic stress to scavenge H 2 O 2 that may leak out from mitochondria to cytosol.