OUP user menu

Different signalling pathways mediate glucose induction of SUC2, HXT1 and pyruvate decarboxylase in yeast

Mónica M. Belinchón, Juana M. Gancedo
DOI: http://dx.doi.org/10.1111/j.1567-1364.2006.00136.x 40-47 First published online: 1 January 2007


The glucose sensors Gpr1, Snf3 and Rgt2 generate the earliest signals produced by glucose in yeast. We showed that a lack of Gpr1 or Snf3/Rgt2 decreased by twofold the glucose induction of SUC2, but had no effect on the induction of pyruvate decarboxylase (Pdc). The induction of HXT1 was not affected by the absence of Gpr1. In an hxk1 hxk2 glk1 strain, high glucose fully induced SUC2, caused partial induction of HXT1 and had no effect on Pdc. In this strain, SUC2 induction was dependent on Gpr1, but HXT1 induction was not. Hxk2, required for the high expression of HXT1, was dispensable for the full induction of SUC2 or Pdc. These results indicate that glucose does not induce transcription through a single signalling pathway, but that several pathways may, in different combinations, regulate the transcription of different genes.

  • cAMP
  • glucose
  • Gpr1
  • hexokinase
  • transcriptional regulation
  • yeast


The yeast Saccharomyces cerevisiae may grow on diverse carbon sources, but is best adapted to use glucose and shows the highest rate of growth on this substrate. In the presence of glucose, the transcription of a number of genes involved in glucose utilization is induced (Müller, 1995; Boles, 1996; Johnston, 1999) and, amongst these, the best studied are the HXT genes encoding glucose transporters. It is well established that induction of the HXT genes requires the glucose transporter homologues Snf3 or Rgt2 (Özcan, 1996, 1998), and that the binding of glucose to Snf3 or Rgt2 initiates a series of reactions that result in the removal of the repressor Rgt1 from the corresponding promoters (Flick, 2003; Lakshmanan, 2003; Moriya & Johnston, 2004). This signalling pathway, however, appears to be mainly related to the control of the HXT genes (Kaniak, 2004).

Another important element in the initiation of glucose signalling is the G-protein-coupled receptor Gpr1, involved in glucose-induced cAMP synthesis (Yun, 1998; Kraakman, 1999). There is strong evidence that Gpr1 interacts directly with glucose (Lemaire, 2004), and it has also been established that full activation of cAMP synthesis by glucose requires its phosphorylation by any of the three identified glucose kinases, hexokinase 1 (Hxk1), hexokinase 2 (Hxk2) or glucokinase (Glk1) (Rolland, 2000). Hxk2, the main enzyme present during growth on glucose, also plays a specific role in glucose signalling, as it is required for the repression by glucose of some genes, although others are still controlled by glucose in an hxk2 mutant (Gancedo, 1998).

In this work, we examine how the lack of the glucose sensors Gpr1, Snf3 and Rgt2, or of glucose-phosphorylating enzymes, affects the induction by glucose of different systems: the SUC2 gene encoding invertase, the HXT1 gene encoding a low-affinity glucose transporter and the glycolytic enzyme pyruvate decarboxylase (Pdc). We conclude that different systems subject to induction by glucose respond to different combinations of signals.

Materials and methods

Plasmids and yeast strains

Plasmid pMMB1, containing a disrupted GPR1 gene, was constructed by replacing the 2-kb HpaI–BglII fragment from plasmid pGPR1-RS416 (Yun, 1998) with a 1.4-kb EcoRV–BglII fragment from plasmid pFA6a-KanMX4, that contains the KanMX gene (Wach, 1994). pCK169 is a multicopy plasmid containing an HXT1–lacZ fusion gene (Ko, 1993). The yeast strains used in this study are listed in Table 1. Strains with a disrupted GPR1 gene were constructed by transformation of strains W303-1A, MJL30 and THG1 with a SacI–ApaI fragment from pMMB1 and selection for resistance to geneticine. Correctness of the deletions was confirmed by PCR and Southern analysis of the transformants. Strains W303-1A, MMB3, MJL30, MMB4, THG1, MMB2 and THG1-HXK2 were transformed with plasmid pCK169 to yield strains MMB5–MMB11.

View this table:
Table 1

Yeast strains used in this work

StrainRelevant genotypeSource or reference
W303-1AMATαade2-1 his3-11,15 leu2-3,112 trp1-1 ura3-1Thomas & Rothstein (1989)
MMB3MATαade2-1 his3-11,15 leu2-3112 trp1-1 ura3-1 gpr1::KanMX4This study
MJL30MATαade2-1 his3-11,15 leu2-3112 trp1-1 ura3-1 snf3::HIS3 rgt2::LEU2 mth1::TRP1This laboratory
MMB4MATαade2-1 his3-11,15 leu2-3,112 trp1-1 ura3-1snf3::HIS3 rgt2::LEU2 mth1::TRP1 gpr1::KanMX4This study
THG1MATαleu2-1 ura3-1 lys1-1 hxk1::LEU2 hxk2::LEU2glk1::LEU2This laboratory
MMB2MATαleu2-1 ura3-1 lys1-1 hxk1::LEU2 hxk2::LEU2 glk1::LEU2 gpr1::KanMXThis study
THG1-HXK2MATαleu2-1 ura3-1 lys1-1 hxk1::LEU2 glk1::LEU2This laboratory
DFY1MATαlys1-1Walsh (1991)
DFY568MATαhxk1::LEU2 hxk2::LEU2 GLK1 leu2-1 lys1-1Walsh (1991)
DFY582MATαhxk1::LEU2 HXK2 glk1::LEU2 leu2-1 lys1-1Walsh (1991)
DFY583MATαHXK1 hxk2::LEU2 glk1::LEU2 leu2-1 lys1-1Walsh (1991)
MMB5MATαade2-1 his3-11,15 leu2-3,112 trp1-1 ura3-1 [HXT1-lacZ]This study
MMB6MATαade2-1 his3-11,15 leu2-3,112 trp1-1 ura3-1 gpr1::KanMX [HXT1-lacZ]This study
MMB7MATαleu2-3,112 his3-11,15 ade2-1 ura3-1 snf3::HIS3 rgt2::LEU2 mth1::TRP1 [HXT1-lacZ]This study
MMB8MATαleu2-3,112 his3-11,15 ade2-1 ura3-1 snf3::HIS3 rgt2::LEU2 mth1::TRP1 gpr1::KanMX [HXT1-lacZ]This study
MMB9MATαleu2-1 ura3-1 lys1-1 hxk1::LEU2 hxk2::LEU2glk1::LEU2 [HXT1-lacZ]This study
MMB10MATαleu2-1 ura3-1 lys1-1 hxk1::LEU2 hxk2::LEU2glk1::LEU2 gpr1::KanMX [HXT1-lacZ]This study
MMB11MATαleu2-1 lys1-1 hxk1::LEU2 glk1::LEU2 [HXT1-lacZ]This study

Growth conditions and induction of enzymes

The yeasts were grown at 30°C in a rich medium, YP (1% yeast extract and 2% bacteriological peptone), or in a synthetic medium, YNB [0.17% yeast nitrogen base (Difco) plus 0.5% ammonium sulphate, and the necessary supplements, at a final concentration of 20 mg L−1], with the carbon sources indicated in each case. The yeast cells were collected at the exponential phase of growth and kept frozen at −20°C until use. To induce invertase, cells grown in YP 2% glucose were washed with distilled water, resuspended at 1.5 mg wet weight mL−1 in YP 0.05% glucose, and incubated for the times indicated; incubation in YP 2% glucose was carried out as a control for repressed levels. To induce Pdc, ethanol-grown cells were washed with distilled water, resuspended in YP with 0.05% or 2% glucose at 4 mg wet weight mL−1, and incubated for the times indicated.

Extracts and enzyme assays

Extracts were prepared by shaking yeast cells with glass beads as described previously (Blázquez, 1993). To assay Pdc, extracts were performed with 0.1 M phosphate buffer, pH 6, containing 30 μM thiamine pyrophosphate and 5 mM cysteine. Invertase was assayed in whole cells as described previously (Goldstein & Lampen, 1975), with some modifications (Rodríguez & Gancedo, 1999), β-galactosidase was tested as described in Miller (1972) using noncentrifuged yeast extracts and centrifuging the samples before reading, and Pdc was measured spectrophotometrically (Singer & Pensky, 1952).

Northern blot analysis

Yeast cells were sampled by rapid filtration and RNA was extracted as described previously (Belinchón, 2004). The procedure for RNA separation by electrophoresis and the probe used for 25S rRNA are described in Belinchón (2004). As probe for SUC2, a BamH1–HindIII fragment (nucleotides 12–788 of the ORF) from plasmid pRB58 was used (Carlson & Botstein, 1982).


To examine the role of glucose sensors in the yeast plasma membrane on the process of induction of gene expression by glucose, a series of isogenic strains lacking either Gpr1, both Snf3 and Rgt2, or the three proteins, was constructed. As the absence of both Snf3 and Rgt2 severely impairs glucose metabolism, we deleted the gene MTH1 in the snf3 rgt2 strains, thus allowing the constitutive expression of some glucose transporter genes (Schmidt, 1999; Lafuente, 2000). The strains obtained (gpr1, snf3 rgt2 mth1 and snf3 rgt2 mth1 gpr1) grew in glucose medium at the same rate as the corresponding wild-type yeast.

Induction by glucose of SUC2 transcription requires either glucose metabolism or the Gpr1 sensor

To test whether expression of SUC2 in the strains snf3 rgt2 mth1 and snf3 rgt2 mth1 gpr1 took place in the absence of glucose, as reported for HXT2, HXT3 and HXT4 in an snf3 rft2 mth1 strain (Schmidt, 1999), we measured invertase in yeast cells grown in YNB ethanol. In both strains, the activity was less than 10 mU invertase mg−1 of yeast. Incubation of isogenic strains with different sets of glucose sensors in a medium with low glucose caused a strong induction that was found to be partially dependent on both Snf3/Rgt2 and Gpr1 (Fig. 1a). The effect on invertase activity of the lack of the glucose sensors was correlated with lower levels of the corresponding SUC2 mRNA (Fig. 1b). It should be indicated that, although invertase activity remained stable between 3 and 6 h after suspension of the yeast cells in a medium with low glucose, the glucose of the medium was exhausted after 90 min and there was a subsequent decrease in the amount of SUC2 mRNA (results not shown).

Figure 1

Effect of the lack of glucose sensors on the induction of the SUC2 gene, encoding invertase, by low glucose. Isogenic strains with the indicated genotype were grown in YP (1% yeast extract and 2% bacteriological peptone) 2% glucose (YPD) and resuspended in YP 0.05% glucose (Glu) or in YPD. (a) After 3 h of incubation, the yeasts were harvested and invertase was measured as described in Materials and methods. Data are averages (±SE) of at least three experiments. (b) Samples were taken from the YPD cultures and from the yeast suspensions in YP 0.05% glucose, incubated for different times, and SUC2 mRNA was probed by Northern blot (see Materials and methods for details). 25S rRNA was used as a loading control.

A possible role for Gpr1 in the expression of SUC2 was also investigated in the yeast strain THG1, which is unable to phosphorylate glucose. We observed that, in this strain (hxk1 hxk2 glk1), there was no induction of SUC2 by 0.05% glucose, but there was a strong induction by 2% glucose, detected as both an increase in mRNA and in invertase activity (Fig. 2). In the absence of both glucose-phosphorylating enzymes and Gpr1, no induction of SUC2 was observed, at either 0.05% or 2% glucose (Fig. 2). After reintroduction of HXK2 in the THG1 strain, catabolite repression of invertase was recovered (about 1 mU invertase mg−1 of yeast in the presence of 2% glucose) and invertase reached an activity of 380±20 mU mg−1 of yeast after derepression at low glucose. As the induction by glucose of genes encoding glucose transporters has been reported to be strongly dependent on Hxk2 (Özcan & Johnston, 1995), we tested whether Hxk2 played a similar role in the induction of invertase by low glucose. This was not the case, as invertase levels were slightly higher in strains expressing only Hxk1 or Glk1 than in the wild-type strain (Fig. 3).

Figure 2

Effect of the lack of Gpr1 on the induction of the SUC2 gene by glucose in a yeast strain unable to phosphorylate this sugar. (a) Strains THG1 (hxk1 hxk2 glk1) and its derivative MMB2 (hxk1 hxk2 glk1 gpr1) were grown in synthetic medium YNB [0.17% yeast nitrogen base (Difco) plus 0.5% ammonium sulphate, and the necessary supplements, at a final concentration of 20 mg L−1] with the carbon sources indicated, and collected during the exponential phase of growth. Data for invertase activity are averages (±SE) of at least three experiments. (b) SUC2 mRNA from strains THG1 and MMB2 grown in different media was analysed by Northern blot using methylene blue-stained 18S rRNA as a loading control.

Figure 3

Effect of different glucose-phosphorylating enzymes on the induced levels of invertase. Isogenic yeast strains with the indicated relevant genotype were grown in a synthetic medium with 2% ethanol and induced for 6 h in YP (1% yeast extract and 2% bacteriological peptone) 0.05% glucose. Invertase was assayed as described in Materials and methods. Data are the averages (±SE) of at least three experiments.

Induction by glucose of HXT1 and Pdc is independent of the Gpr1 sensor

Using strains lacking Gpr1 and/or Snf3/Rgt2, we observed that the reporter gene HXT1–lacZ was still fully induced by 4% glucose in the absence of Gpr1, whereas, in accordance with previous reports (Schmidt, 1999; Lafuente, 2000), the lack of Snf3 and Rgt2 (even in an mth1 background) blocked induction completely (Fig. 4a). In the absence of glucose phosphorylation, 2% glucose induced up to 10-fold the expression of a reporter HXT1–lacZ gene, but this induction was only 20% of that reached in an isogenic HXK2 strain (Fig. 4b). In the hxk1 hxk2 glk1 background, the lack of Gpr1 did not decrease the level of induction and may even increase it (Fig. 4b).

Figure 4

Effect of the lack of Gpr1 on the induction of an HXT1–lacZ fusion gene by high glucose. Two sets of isogenic Saccharomyces cerevisiae strains derived from W303-1A (a) or THG1 (b), carrying a plasmid with the HXT1–lacZ gene, were grown in synthetic medium YNB [0.17% yeast nitrogen base (Difco) plus 0.5% ammonium sulphate, and the necessary supplements, at a final concentration of 20 mg L−1] with the carbon sources indicated. Cells were harvested at the exponential phase of growth and β-galactosidase was tested as described in ‘Materials and methods’. Data are the averages (±SE) of at least three experiments.

The role of Gpr1 on the induction by glucose of Pdc was also tested. We found that a lack of Gpr1 and of the glucose sensors Snf3 and Rgt2 had no marked effect on the kinetics of induction by glucose of Pdc (Fig. 5a), and that Pdc activity after growth on glucose was similar in a wild-type strain and in strains lacking different sensors (Fig. 5b). It should be noted that, in glucose-grown cells, PDC1 is the gene responsible for most of the Pdc activity (Hohmann, 1993).

Figure 5

Effect of the lack of glucose sensors on the induction of pyruvate decarboxylase (Pdc) by high glucose. (a) Strains W303-1A (Wt) and its derivative MMB4 (snf3 rgt2 mth1 gpr1) were grown in YP (1% yeast extract and 2% bacteriological peptone) 2% ethanol and induced in YP 2% glucose as described in Materials and methods. (b) Isogenic strains with the genotype indicated were grown in YP 2% glucose and collected at the exponential phase of growth. Data are the averages (±SE) of at least three experiments.

We checked that the induction of Pdc by glucose was completely blocked in a strain that did not phosphorylate the sugar, as already reported by others (Gonçalves, 1997). Hxk2 was not specifically required for glucose induction, as a strain with Hxk1 only showed the same Pdc activity during growth on glucose as the wild-type yeast; however, in the absence of both Hxk1 and Hxk2, the induced levels of Pdc were decreased by 30–50%. In the presence of low glucose (0.05%), there was a twofold induction of Pdc that was only slightly affected by the absence of Hxk1 and Hxk2 (results not shown).


We have shown in this study that the induction by glucose of different systems has different requirements, thus indicating that the corresponding gene promoters are regulated by different signals.

In the case of SUC2, glucose modifies the capacity of several regulatory proteins to bind to its promoter, as shown schematically in Fig. 6. We have established that, in an snf3 rgt2 mth1 strain, SUC2 induction still requires glucose and that, at 0.05% glucose, expression is only 50% of that measured in a wild-type strain. These results indicate that the release of repression by Rgt1, a protein able to bind the SUC2 promoter (Hazbun & Fields, 2002), is required for SUC2 expression, but that the triggering of this release is not the only role of glucose. This agrees with an earlier observation showing that the deletion of RGT1 causes an increased expression of SUC2 in the absence of glucose, which does not reach, however, the full glucose-induced levels (Özcan, 1997). The fact that, in a genome-wide screening for targets of Snf3/Rgt2 (Kaniak, 2004), the SUC2 gene was not identified is consistent with a modest role for Rgt1 in the control of SUC2 transcription. Our results contrast with those of Schmidt (1999), who found that, in an snf3 rgt2 mth1 mutant, invertase was fully derepressed by low glucose and only partially repressed by high glucose. This discrepancy is probably due to differences in the genetic background of the strains used, as it has been observed that invertase expression varies widely in different snf3 mutants (Neigeborn, 1986; Özcan, 1997; Özcan, 2002).

Figure 6

Schematic representation of the promoters of SUC2 and HXT1 in the presence of different concentrations of glucose. Sko1, Rgt1, Mig1 and Sfl1 need to bind Cyc8 to act as repressors; however, for clarity, the Cyc8–Tup1 complex is not included in the diagram. On phosphorylation, the repressors dissociate from the promoters. Phosphorylation of the repressors is mediated by different systems: the Hog pathway for Sko1, binding of glucose to Snf3/Rgt2 for Rgt1, activation of Snf1 for Mig1, and activation of Tpk2 for Sfl1. ‘A’ represents a putative activatory protein, which may depend on cAMP. See Discussion for details. Distances between sites in the promoter of SUC2 are drawn to scale.

We have also shown that induction of SUC2 by low glucose is not dependent on a specific hexose kinase, and that induction can even take place in a strain unable to phosphorylate the sugar. In this strain, however, both a high concentration of glucose and the glucose sensor Gpr1 are required for SUC2 expression. This result parallels the observation that the increase in intracellular cAMP concentration triggered by glucose depends on signalling through both Gpr1 and the hexose kinases (Rolland, 2000). This points to an important role for cAMP in the control of SUC2 and shows that, in certain conditions, SUC2 may be induced in the absence of glucose metabolism. At low glucose, metabolism of the sugar is needed to achieve the required concentration of cAMP, and Gpr1 plays only a minor role, because, even in its absence, glucose triggers a moderate increase in the intracellular concentration of cAMP (Kraakman, 1999). In the absence of glucose phosphorylation, an increase in cAMP becomes dependent on Gpr1 and on a high concentration of glucose (Kraakman, 1999). One of the functions of cAMP may be to counteract repression by Sfl1 (Song & Carlson, 1998; see Fig. 6), as it has been shown that phosphorylation of Sfl1 by the protein kinase Tpk2 releases Sfl1 from the SUC2 promoter (Conlan & Tzamarias, 2001). Another function of cAMP could be the stimulation of a putative activation factor (indicated as A in Fig. 6) binding to two activating regions, SUC2A and SUC2B, in the promoter of SUC2 (Bu & Schmidt, 1998), as it has been reported that the expression of invertase at low glucose is much reduced in a mutant with a defective adenylate cyclase (Matsumoto, 1982). In the presence of high glucose, the repressor Mig1 binds to the SUC2 promoter and blocks its expression (Nehlin & Ronne, 1990). Repression of SUC2 by high glucose, however, does not take place in the absence of Hxk2 (Entian & Zimmermann, 1980; Ma & Botstein, 1986; Walsh, 1991). This is because Hxk2 is required for the inactivation of Snf1 triggered by high glucose, and an active Snf1 is able to phosphorylate Mig1, allowing it to dissociate from the SUC2 promoter (Treitel, 1998). This, in turn, explains our finding that high glucose induces SUC2 in an hxk1 hxk2 glk1 strain.

As shown in Fig. 6, there is a further regulatory element in the SUC2 promoter: a weak CRE site able to bind the repressor Sko1; however, it plays only a marginal role in SUC2 repression (Nehlin, 1992; Bu & Schmidt, 1998).

Transcription of HXT1 is regulated in a different way, although it shares with SUC2 the repressors Rgt1 and Sko1 shown in Fig. 6. Expression of HXT1 is completely dependent on the removal of the repressor Rgt1 from its promoter, mediated by the interaction of glucose with Snf3/Rgt2. In addition, as long as Sko1 is bound to the HXT1 promoter, the transcription rate of the gene is very low. Activation of the Hog pathway by the mild osmotic stress caused by 2% glucose produces the phosphorylation of Sko1 and its dissociation from the HXT1 promoter (Tomas-Cobos, 2004). The release of repression by Rgt1 and Sko1, however, is not sufficient for full expression of HXT1, which requires the Hxk2 isoenzyme (Özcan & Johnston, 1995). This may be related to a requirement for Hxk2 to inactivate the protein kinase Snf1 (Treitel, 1998), as it has been reported that activation of Snf1 leads to an inhibition of HXT1 expression (Tomas-Cobos & Sanz, 2002). However, as we measured in this work a degree of induction of HXT1 in the absence of glucose phosphorylation similar to that reported for an hxk2 mutant (Özcan & Johnston, 1995), it can be concluded that the release of repression by Rgt1 and Sko1 does not require glucose metabolism. We have also shown that the expression of HXT1 does not depend on the glucose sensor Gpr1, even in the absence of glucose phosphorylation, ruling out an activating role for cAMP.

With regard to induction by glucose of Pdc, the situation is again different. We have established that none of the glucose sensors Snf3, Rgt2 or Gpr1 plays a role in the process, but glucose metabolism is absolutely required. Our observation that induction is only partial at low glucose and depends on Hxk2 or Hxk1 when glucose is high is consistent with the proposal that the level of induction of Pdc is related to the glycolytic flux (Elbing, (2004), which, in turn, determines the intracellular concentrations of putative activating metabolites. Although a correlation has been found between Pdc activity and the concentration of triosephosphates (Boles & Zimmermann, 1993), and glucose 6-phosphate has been suggested to play a role in glucose repression (Gancedo, 1998; Vincent, 2001), the nature and mode of action of potential regulatory metabolites in mediating the different effects of glucose are still unknown. However, changes in the concentration of these metabolites are not uniquely related to glucose metabolism, as galactose or xylose may mimic to some degree diverse effects of glucose (Rodríguez & Gancedo, 1999; Belinchón & Gancedo, 2003).

A previous suggestion that glucose may induce transcription in the absence of glucose metabolism was based on the use of an hxt strain (Özcan, 2002). As this strain could grow on glucose, unless antimycin was added, the conclusion does not appear to be fully warranted. Our work with mutants lacking all the isoenzymes able to phosphorylate glucose indicates that, in the absence of glucose metabolism, a gene such as SUC2 may be fully induced by high glucose, but that there is only a partial induction of HXT1. We have also confirmed that the induction of Pdc is completely blocked in these conditions. There is a further difference between the different systems induced by glucose: although we have shown that either Hxk1 or Hxk2 may sustain full induction of Pdc by high glucose and that any of the glucose-phosphorylating enzymes allows induction of SUC2 or Pdc by low glucose, Hxk2 is specifically required for a strong induction of HXT1 by high glucose or of HXT4 by low glucose (Özcan & Johnston, 1995).


We thank H. Tamaki (Kyoto University) for plasmid pGPR1-RS416, D.G. Fraenkel (Harvard University) for a set of yeast strains and F. Portillo for plasmid pCK169. Critical reading of the manuscript by C. Gancedo and C. L. Flores is gratefully acknowledged. This work was supported by grants BMC2001-1690-C02-1 and BFU2004-02855-C02-1 from the Dirección General de Investigación Científica y Técnica (DGICYT) and by the European Union (EU) project BIO-HUG QLK3-CT-1999-00080. M.M.B. was supported by fellowships from the EU and the DGICYT.


  • Editor: Monique Bolotin-Fukuhara


View Abstract