OUP user menu

Glucose controls multiple processes in Saccharomyces cerevisiae through diverse combinations of signaling pathways

Mónica M. Belinchón, Juana M. Gancedo
DOI: http://dx.doi.org/10.1111/j.1567-1364.2007.00236.x 808-818 First published online: 1 September 2007

Abstract

We have studied how the lack of glucose sensors in the plasma membrane, or of the enzymes Hxk1, Hxk2, Glk1, which catalyze the first intracellular step in glucose metabolism, affect the different responses of Saccharomyces cerevisiae to glucose. Lack of the G-protein-coupled receptor Gpr1 or of Snf3/Rgt2 did not affect glucose repression of different genes or activation by glucose of plasma membrane ATPase, whereas lack of Gpr1 decreased, in an additive manner with lack of Mth1, the degradation of fructose 1,6-bisphosphatase that takes place in the presence of glucose. In an hxk1 hxk2 glk1 strain, unable to phosphorylate glucose, all of these responses to the sugar were suppressed or strongly reduced. In the absence of Hxk2 (or Hxk1 and Hxk2), glucose repression of SUC2, GAL1 and GDH2 was relieved, but that of FBP1 and ICL1 was maintained. Hxk1 or Hxk2 were needed for activation of plasma membrane ATPase but not for degradation of FbPase.

Keywords
  • cAMP
  • glucose
  • Gpr1
  • hexokinase
  • Snf3
  • yeast

Introduction

Glucose is not only the best carbon source for the yeast Saccharomyces cerevisiae, but is also a signaling molecule that triggers many different processes. When there is a high concentration of glucose in the medium, many genes involved in the metabolism of alternative carbon sources are repressed (Gancedo, 1998; Carlson, 1999), the turnover of a number of mRNAs is increased (Lombardo et al., 1992; Mercado et al., 1994; Cereghino & Scheffler, 1996), and different enzymes and sugar transporters are inactivated (Gancedo & Gancedo, 1997). On the other hand, glucose induces the transcription of a variety of genes (Müller, 1995; Boles, 1996; Johnston, 1999), the stabilization of some mRNAs (Yin, 2003), and the activation of several enzymes (Portillo, 2000).

Theoretically, the diversity of processes triggered by glucose might depend on a single early signal targeting different regulatory pathways, or on different signals, each controlling a distinct process. However, even the induction of different genes by glucose, which could be considered as a single process, is not regulated by the same early signals. The glucose sensors Snf3/Rgt2 play an important role in the control of the genes encoding glucose transporters (Özcan, 1998), but are dispensable for the induction of pyruvate decarboxylase and have only a small effect on the induction of invertase (Belinchón & Gancedo, 2007). Also, the lack of the G-protein-coupled glucose receptor, Gpr1 (Yun, 1998; Rolland, 2000), or of the glucose-phosphorylating enzymes Hxk1, Hxk2 and Glk1, which catalyze the first step in glucose intracellular metabolism, have different effects on the induction of the glucose transporter Hxt1, pyruvate decarboxylase or invertase (Belinchón & Gancedo, 2007).

It therefore appeared interesting to investigate how the lack of the glucose sensors Snf3, Rgt2 and Gpr1, or of Hxk1, Hxk2 and Glk1, affected other types of process triggered by glucose, such as glucose repression and activation or inactivation of enzymes. For this, we have examined a set of genes subject to glucose repression that belong to different regulatory circuits (Gancedo, 1998): SUC2 and GAL1, which encode invertase and galactokinase, required for the metabolism of other sugars; FBP1 and ICL1, which encode the gluconeogenic enzymes fructose-1,6-bisphosphatase (FbPase) and isocitrate lyase (Icl); and GDH2, which encodes the NAD-dependent glutamate dehydrogenase (GlutDH). We have taken the plasma membrane H+-ATPase as the model for an enzyme activated by glucose, and FbPase as an example of an enzyme inactivated in the presence of glucose.

We found that most of the glucose effects studied are independent of the glucose sensors Snf3/Rgt2 and Gpr1, and that all require glucose metabolism, some of the effects being specifically dependent on Hxk2 (or Hxk1/Hxk2).

Materials and methods

Yeast strains and growth conditions

The yeast strains used in this work are listed in Table 1. CLF62 and MJL30 are derived from strain W303-1B (Thomas & Rothstein, 1989); they were constructed as described for CLF61 and CJM285, respectively, which are similar strains derived from W303-1A (Lafuente, 2000). To generate strain CJM479, strains CLF62 and MMB3 were crossed, the diploid was sporulated, and a recombinant was selected as resistant to geneticin and prototrophic for tryptophan. THG1 was constructed by crossing strains W303-1A and DFY568, sporulating the diploid, selecting an hxk1 hxk2 recombinant unable to grow on fructose, and crossing it with DFY582. From the diploid obtained, a recombinant unable to grow on glucose was selected, and it was determined that extracts from this strain, THG1, had no glucose-phosphorylating activity. The strains that we have used belong to three different backgrounds, W303, DFY1, and THG1. The reference strain, THG1-HXK2, isogenic with THG1, was constructed by replacing the interrupted hxk2::LEU2 gene by the intact HXK2 gene. The yeasts were grown at 30°C in a rich medium, YP (1% yeast extract and 2% bacteriologic peptone), with the carbon sources indicated in each case at 2% unless indicated otherwise. The yeast cells were collected at the exponential phase of growth and kept frozen at −20°C until use. To derepress invertase, cells grown in YPglucose (YPD) were washed with distilled water, resuspended at 1.5 mg wet weight mL−1 in YP 0.05% glucose, and incubated for 3 h.

View this table:
1

Yeast strains used in this work

StrainRelevant genotypeSource or reference
W303-1AMAT a ade2-1 his3-11,15 leu2-3,112 trp1-1 ura3-1Thomas & Rothstein (1989)
W303-1BMATα ade2-1 his3-11,15 leu2-3,112 trp1-1 ura3-1Thomas & Rothstein (1989)
MMB3MAT a ade2-1 his3-11,15 leu2-3,112 trp1-1 ura3-1 gpr1::KanMX4Belinchón & Gancedo (2007)
CLF62MATα ade2-1 his3-11,15 leu2-3,112 trp1-1 ura3-1 mth1::TRP1This laboratory
CJM479MAT a ade2-1 his3-11,15 leu2-3,112 trp1-1 ura3-1 gpr1::KanMX4 mth1::TRP1This work
MJL30MATα ade2-1 his3-11,15 leu2-3,112 trp1-1 ura3-1 snf3::HIS3 rgt2::LEU2 mth1::TRP1This laboratory
MMB4MATα ade2-1 his3-11,15 leu2-3,112 trp1-1 ura3-1 snf3::HIS3 rgt2::LEU2 gpr1::KanMX4 mth1::TRP1Belinchón & Gancedo (2007)
DLY1901MAT a ade2-1 his3-11,15 leu2-3,112 trp1-1 ura3-1 rgt1::LEU2Palomino (2006)
THG1MATα leu2-1 ura3-1 lys1-1 hxk1::LEU2 hxk2::LEU2 glk1::LEU2This laboratory
MMB2MATα leu2-1 ura3-1 lys1-1 hxk1::LEU2 hxk2::LEU2 glk1::LEU2 gpr1::KanMX4Belinchón & Gancedo (2007)
THG1-HXK2MATα leu2-1 ura3-1 lys1-1 hxk1::LEU2 glk1::LEU2This laboratory
DFY1MAT a lys1-1Walsh (1991)
DFY568MATα hxk1::LEU2 hxk2::LEU2 GLK1 leu2-1 lys1-1Walsh (1991)
DFY582MATα hxk1::LEU2 HXK2 glk1::LEU2 leu2-1 lys1-1Walsh (1991)
DFY583MATα HXK1 hxk2::LEU2 glk1::LEU2 leu2-1 lys1-1Walsh (1991)

Derepression of FbPase, inactivation of FbPase, and activation of ATPase

To follow the kinetics of derepression of FbPase, cells grown in YPgalactose were washed with distilled water, suspended at 20 mg wet weight mL−1 in YPethanol or YPethanol+2% glucose, and incubated at 30°C; samples were taken at the times indicated.

To measure inactivation of FbPase by glucose, the yeasts were grown in YPethanol until exponential phase, and glucose was then added to the culture medium at a final concentration of 2%. Samples were taken at the times indicated.

To measure activation of the plasma membrane H+-ATPase, yeast cells grown in YPD were collected, washed twice with distilled water, and resuspended in water at 130 mg wet weight mL−1. The suspension was incubated for 5 min at 30°C, and samples were collected immediately and 10 min after adding glucose (2% final concentration). In the case of strain hxk1 hxk2 glk1, the yeast was grown in YPgalactose.

Extracts and enzyme assays

Extracts were prepared by shaking yeast cells with glass beads as previously described (Blázquez, 1993), and the activities of FbPase, Icl, GlutDH and galactokinase were measured spectrophotometrically (Gancedo & Gancedo, 1971; Dixon & Kornberg, 1959; Doherty, 1970; Maitra & Lobo, 1971). Invertase was assayed in intact cells as described previously (Goldstein & Lampen, 1975), with some modifications (Rodríguez & Gancedo, 1999). H+-ATPase activity was measured at pH 6.5, with 2 mM ATP (Serrano, 1988).

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 probes used for the 25S rRNA gene, FBP1, ICL1 and GDH2 are described in Belinchón (2004).

Western blot analysis

This was performed as previously described (Zaragoza, 2001), except that we used as secondary antibody goat anti-rabbit peroxidase conjugate (Santa Cruz Biotechnology, CA, USA) diluted 1 : 5000, and that the antibodies were visualized with a solution containing 1 M Tris-HCl (pH 8.8), 50 mM p-coumaric acid, 50 mM luminol, and 3.6% H2O2.

Results

Glucose repression is independent of the glucose sensors Snf3, Rgt2 and Gpr1

Glucose repression is not relieved by the absence of Snf3 and Rgt2, as long as glucose can be metabolized efficiently (Lafuente, 2000). This result does not eliminate these glucose sensors as participants in glucose repression, as there is a further glucose sensor, Gpr1, that might play a redundant role. We therefore measured the levels of different enzymes subject to catabolite repression and belonging to different regulatory circuits (Gancedo, 1998) in cells grown in 2% glucose or in derepressed conditions, using a series of isogenic strains lacking either Gpr1, both Snf3 and Rgt2, or the three proteins. As shown in Table 2, in all strains the enzymes tested were strongly repressed in glucose cultures, indicating that the glucose sensors are completely dispensable for long-term glucose repression. It should be noted that the snf3 rgt2 strains used also had a deletion in MTH1, which encodes a repressor of the HXT genes, to avoid impairment of glucose metabolism due to reduced expression of the glucose transporters in those strains (Özcan, 1996, 1998).

View this table:
2

Effect of the lack of glucose sensors on the expression of different enzymes

Relevant genotypeInvertase [mU (mg yeastww)−1]Glutamate dehydrogenase [mU (mg protein)−1]Fructose-1,6-bisphosphatase [mU (mg protein)−1]
2% Glu0.05% Glu2% Glu2% Et+3% Gly2% Glu2% Et+3% Gly
Wild-type (W303-1A)<5230 ± 1710 ± 3124 ± 1<129 ± 1
gpr11120 ± 1612 ± 2108 ± 12<126 ± 5
snf3 rgt2 mth17112 ± 612 ± 1101 ± 6<128 ± 3
snf3 rgt2 gpr1 mth1170 ± 1414 ± 3107 ± 7<125 ± 4
  • Yeasts were grown in repressing or derepressing conditions, with glucose (Glu), ethanol (Et) or glycerol (Gly) as carbon sources, and enzymatic activities were assayed as described in ‘Materials and methods’. Data are averages (± SE) of a least three independent experiments. ww, wet weight.

Absence of the sensors did not affect the kinetics of repression by glucose, as there was a similar decline in the mRNA levels of different genes after exposure of the various strains to glucose (Fig. 1). We conclude, therefore, that the mechanisms through which glucose triggers catabolite repression are independent of the plasma membrane glucose sensors. It may be noted that, in this particular experiment, the levels of GDH2, FBP1 and ICL1 mRNAs in derepressing conditions are lower in the mutants than in the wild-type strain. However, in other experiments where derepressed levels of mRNAs were measured, there were no consistent differences between strains, as found also for the enzymatic activities of GlutDH and FbPase (Table 2).

1

Repression of different genes by glucose is not affected by the lack of glucose sensors. Isogenic strains with the genotype indicated were grown in derepressing carbon sources (Gly, glycerol; Et, ethanol), and samples were taken before (0) and after the addition of glucose (Glu) to the medium (7 and 14 min). The corresponding mRNAs were probed by Northern blot (see ‘Materials and methods’ for details). Samples from yeasts grown in YPD were run as controls. Numbers below the gel lanes indicate relative amount of mRNA, normalized to the ACT1 gene, given as the percentage of the values at time 0.

Glucose repression of different enzymes does not respond in the same way to the absence of diverse kinases able to phosphorylate glucose

The isoenzyme Hxk2 is required for repression of the SUC2 gene by high glucose levels, but other genes are still fully glucose-repressed in hxk2 strains (Gancedo, 1998). We have studied the effects of individual glucose-phosphorylating enzymes on the repression of different enzymes. As some of the enzymes studied also required an inducing carbon source to be expressed (Fernández, 1993), the activity of those enzymes was also tested in yeasts grown in the presence of the inducer. The results (Table 3) show that there is not a single pattern relating the presence of a certain kinase with the extent of glucose repression. Any of the kinases allowed full repression of FbPase and Icl, whereas GlutDH was weakly repressed and galactokinase not repressed at all when Glk1 was the only phosphorylating enzyme. Hxk1 produced a marked repression of GlutDH, although not as strong as that caused by Hxk2, and only residual repression of galactokinase. It should be noted that the low activities of FbPase and Icl in extracts from the hxk1 hxk2 strain grown in glucose cannot be ascribed to the lack of an inducing carbon source, as the same low activities are observed when the yeast is grown in a mixture of glucose and ethanol. In the case of GlutDH, it is surprising to find that, in the hxk1 hxk2 strain grown in the presence of glucose together with ethanol, the enzymatic activity was always lower than in the yeast grown in either carbon source alone. This result is consistent with a previous observation showing that GlutDH activity was lower after growth on glycerol+ethanol than after growth on glycerol alone (Belinchón & Gancedo, 2003), but no straightforward explanation can be offered.

View this table:
3

Effect of glucose-phosphorylating enzymes on catabolite repression

EnzymesCarbon sourcesStrains
Wild-typehxk1 HXK2 glk1HXK1 hxk2 glk1hxk1 hxk2 GLK1
Galactokinase [mU (mg protein)−1]Galactose560 ± 80 (4)785 ± 58 (3)535 ± 9 (3)
Galactose+glucose≤20 (3)384 ± 34 (3)540 ± 100 (5)
Glut DH [mU (mg protein)−1]Glucose8 ± 2 (3)10 ± 2 (3)44 ± 9 (3)120 ± 15 (3)
Glucose+ethanol11 ± 1 (2)12 ± 1 (3)43 ± 11 (3)71 ± 20 (3)
Ethanol157 ± 10 (3)173 ± 8 (3)169 ± 21 (3)188 ± 29 (4)
FbPase [mU (mg protein)−1]Glucose≤1 (2)≤1 (3)≤1 (3)3 ± 1 (3)
Glucose+ethanol≤1 (3)≤1 (3)≤1 (3)≤1 (4)
Ethanol70 ± 11 (5)59 ± 3 (3)65 ± 3 (3)69 ± 14 (6)
Icl [mU (mg protein)−1]Glucose≤1 (2)≤1 (3)≤1 (2)2 ± 1 (2)
Glucose+ethanol≤1 (2)≤1 (3)≤1 (2)2 ± 1 (3)
Ethanol148 ± 17 (5)154 ± 29 (3)133 ± 23 (3)138 ± 15 (4)
  • Strains expressing different glucose-phosphorylating enzymes in the DFY1 background were grown in YP with the carbon sources indicated (2%), and enzymatic activities were assayed as described in ‘Materials and methods’. Data are averages (± SE) of independent experiments (number of cultures in parentheses).

In the absence of the three glucose-phosphorylating enzymes, the activities of the enzymes tested were similar when the yeast was grown on a gluconeogenic carbon source, such as ethanol, in the absence or presence of glucose (Table 4). The partial repression by glucose, however, is significant, as glucose caused also a delay in the kinetics of derepression of FbPase (supplementary Fig. S1). To test whether this effect of glucose was mediated through Gpr1, we also measured the effect of glucose in an hxk1 hxk2 glk1 gpr1 strain. As shown in Table 4 and supplementary Fig. S1, glucose acted similarly in the two strains, indicating that its small repressive effect was independent of Gpr1.

View this table:
4

Effect of glucose on the activities [mU (mg protein)−1] of different enzymes in yeast strains unable to phosphorylate glucose

hxk1 hxk2 glk1hxk1 hxk2 glk1 gpr1
EthanolEthanol+glucoseEthanolEthanol+glucose
GlutDH152 ± 6 (3)135 ± 16 (3)215 ± 38 (6)125 ± 16 (6)
FbPase68 ± 4 (3)60 ± 1 (3)68 ± 5 (6)54 ± 6 (5)
Icl170 ± 2 (3)145 ± 6 (3)122 ± 6 (6)86 ± 12 (7)
  • Strains THG1 (hxk1 hxk2 glk1) and MMB2 (hxk1 hxk2 glk1 gpr1) were grown in YP with the carbon sources indicated (2%), and enzymatic activities were assayed as described in ‘Materials and methods’. Data are averages (± SE) of independent experiments (number of cultures in parentheses).

Activation of plasma membrane H+-ATPase by glucose is not affected by lack of glucose sensors but is strongly impaired when both Hxk1 and Hxk2 are absent

Activation of plasma membrane H+-ATPase by glucose is a well-known phenomenon, but is still poorly understood (Portillo, 2000). This activation occurred in the absence of glucose sensors (Table 5), but full activation showed a requirement for either Hxk1 or Hxk2. As shown in Table 6, there was only a 2.5-fold activation in a strain expressing only Glk1 vs. a c. 10-fold activation in the control strains. A yeast strain lacking all glucose-phosphorylating enzymes exhibited only a small activation of the ATPase by glucose (Table 7). This was not due to an intrinsic defect in the capacity of ATPase to be activated in this strain, as galactose activated the enzyme up to sixfold (results not shown). We also checked that when an isogenic strain able to phosphorylate glucose was grown in galactose, activation by glucose was strong, although not as strong as that observed for the glucose-grown yeast (Table 7).

View this table:
5

Effect of the lack of glucose sensors on the activation of the H+-ATPase by glucose

Relevant genotypeATPase [mU (mg yeastww)−1]
Starved yeast+Glucose
Wild-type (W303-1A)0.12/0.312.4/2.5
s nf3 rgt2 gpr1 mth10.16/0.212.3/2.1
  • Yeasts were grown in YP 2% Glu, and plasma membrane H+-ATPase was activated and tested as described in ‘Materials and methods’. Results of two independent experiments are shown.

View this table:
6

Effects of glucose-phosphorylating enzymes on the activation of the H+-ATPase by glucose

Relevant genotypeATPase [mU (mg yeastww)−1]
Starved yeast+Glucose
Wild-type (DFY1)0.23/0.263.2/2.7
HXK1 hxk2 glk10.27/0.332.8/2.7
hxk1 HXK2 glk10.19/0.212.5/1.9
hxk1 hxk2 GLK10.32 ± 0.080.86 ± 0.22
  • Yeasts were grown in YP 2% Glu, and plasma membrane H+-ATPase was activated and tested as described in ‘Materials and methods’. Experiments were performed in duplicate or in triplicate; in the latter case, data are given as averages (± SE).

View this table:
7

Effects of the glucose-phosphorylating enzymes and of the carbon source used for growth on the activation of the H+-ATPase by glucose

Relevant genotypeCarbon sourceATPase [mU (mg yeastww)−1]
Starved yeast+Glucose
hxk1 HXK2 glk1Glucose0.40/0.677.5/7.7
hxk1 HXK2 glk1Galactose0.44/0.513.1/3.8
hxk1 hxk2 glk1Galactose0.10 ± 0.030.16 ± 0.04
  • Yeast strains THG1-HXK2 and THG1 were grown in YP with the carbon source indicated (2%), and plasma membrane H+-ATPase was activated and tested as described in ‘Materials and methods’. Experiments were performed in duplicate or in triplicate; in the latter case, data are given as averages (± SE).

Inactivation of FbPase by glucose is partially dependent on the glucose sensor Gpr1 and requires glucose phosphorylation but not specifically the Hxk1/Hxk2 isoenzymes

Inactivation of FbPase by glucose proceeds in two phases: a very rapid reversible inactivation due to phosphorylation, followed by a slower proteolytic degradation (Gancedo & Gancedo, 1997). Lack of glucose sensors did not markedly affect the first phase of FbPase inactivation, but long-term inactivation was slightly slowed down in snf3 rgt2 and gpr1 mutants, and strongly delayed in an snf3 rgt2 gpr1 mutant (Fig. 2a). As the snf3 rgt2 mutants had an additional MTH1 deletion, to allow the expression of glucose transporters, it was necessary to examine the effect of the lack of Mth1 itself on the kinetics of FbPase inactivation. We found that in the mth1 mutant, there was a decrease in the inactivation rate similar to that observed in the snf3 rgt2 mth1 mutant (Fig. 2a and b), thus suggesting that Snf3 and Rgt2 do not play a role in FbPase inactivation but that Mth1 is involved in the process. As the only known role of Mth1, alone or together with Std1, is to allow Rgt1 to exert its repressive effect on the expression of glucose transporters (Moriya & Johnston, 2004), we also measured FbPase inactivation in an rgt1 mutant. As shown in Fig. 2b, the lack of Rgt1 delayed the first phase of the inactivation and blocked the second phase. This strong effect of the rgt1 mutation supports the conclusion that the role of Mth1 in the inactivation of FbPase is related to its interaction with Rgt1.

2

Effect of the lack of glucose sensors on the inactivation of FbPase triggered by glucose. Isogenic strains, W303-1A (Wt), MMB3 (gpr1), MJL30 (snf3 rgt2 mth1), MMB4 (snf3 rgt2 gpr1 mth1), CLF62 (mth1), CJM 479 (mth1gpr1) and DLY1901 (rgt1), were grown in YPethanol, and samples were taken at time 0 and at 1, 30, 60 and 90 min after the addition of glucose to the cultures (final concentration 2%). See ‘Materials and methods’ for details. Data are given as percentage of the activity at time 0. (a) Representative experiment. (b) Mean values and SD of at least three experiments.

We also measured the rate of inactivation of FbPase in a double mutant mth1 gpr1, and found that the two mutations had an additive effect (Fig. 2a and b). The second phase of FbPase inactivation, between 30 and 90 min, proceeded at a similar rate in strains mth1 gpr1 and mth1 gpr1 snf3 rgt2 (Fig. 2b). Although a small difference between the strains was observed in the extent of inactivation at 1 min (Fig. 2b), its significance is not clear.

Inactivation of FbPase was followed in a yeast strain lacking all the glucose-phosphorylating enzymes. As shown in Fig. 3, the rapid inactivation step, resulting from phosphorylation by a cAMP-dependent kinase (Gancedo, 1983), is only partially impaired, but the subsequent, slower inactivation is blocked. To test whether Gpr1 was required for the inactivation observed, the effect of glucose was checked in an hxk1 hxk2 glk1 gpr1 mutant. The degree of inactivation was indeed decreased, reflecting the inability of glucose to increase cAMP levels in this type of mutant (Rolland, 2000), but residual inactivation still took place (Fig. 3), suggesting the existence of an additional regulatory mechanism.

3

Effect of the lack of glucose-phosphorylating enzymes on the inactivation of FbPase by glucose. Strains THG1 (hxk1 hxk2 glk1), MMB2 (hxk1 hxk2 glk1 gpr1) and THG1-HXK2 (hxk1 HXK2 glk1), as control, were grown in YPethanol, and samples were taken at time 0 and at 1, 60 and 90 min after the addition of glucose to the cultures (final concentration 2%). See ‘Materials and methods’ for details. Data are given as percentage of the activity at time 0.

We observed that any of the glucose-phosphorylating enzymes allowed a normal course of inactivation (Fig. 4a). As it had been reported that, in an hxk2 strain, glucose did not trigger proteolysis of FbPase (Horak, 2002), we also measured the amount of the corresponding protein, and found that both in a wild-type strain and in a strain expressing only Glk1, the second phase of inactivation of the enzyme and the degradation of the protein proceeded in parallel (Fig. 4b).

4

Inactivation of FbPase by glucose in strains with different phosphorylating enzymes. Strains DFY1(Wt), DFY583 (HXK1 hxk2 glk1), DFY582 (hxk1 HXK2 glk1) and DFY568 (hxk1 hxk2 GLK1) were grown in YPethanol, and samples were taken at time 0 and at the times indicated after the addition of glucose to the cultures (final concentration 2%). See ‘Materials and methods’ for details. (a) Data for activity are given as percentage of the activity at time 0. (b) Extracts from strains DFY1 and DFY568 were subjected to Western analysis as described in ‘Materials and methods’.

Discussion

We have shown that the glucose sensors Snf3, Rgt2 and Gpr1 are not required for long-term glucose repression of the enzymes that we examined (Table 2). This conclusion holds true despite the MTH1 deletion present in the snf3 rgt2 strains, as snf3 rgt2 mth1 mutants show, like the wild-type strain, a large difference in enzyme levels depending on the presence or absence of a high concentration of glucose in the medium. For the different genes studied, both short-term and long-term glucose repression were unaffected by the lack of any of the glucose sensors (Fig. 1). This is in contrast with the observation that glucose repression of the stress-controlled genes SSA3 and HSP12 is delayed in a gpr1 strain (Kraakman, 1999a). This difference may be explained if the short-term repression of these genes is strongly dependent on a marked increase in intracellular cAMP. Although cAMP has been shown to inhibit the transcription of some glucose-repressed genes (Zaragoza, 1999), the control of transcription mediated by cAMP is often redundant with other regulatory mechanisms (Zaragoza, 2001).

Our results also show that in strains where glucose cannot be phosphorylated, it does not trigger substantial repression of transcription (Table 4), thus supporting the importance of glucose metabolism for catabolite repression, already suggested by the fact that reduced glycolytic flux decreases the repression caused by glucose (Gamo, 1994; Ye, 1999; Lafuente, 2000; Otterstedt, 2004). Although it has been stated that catabolite repression can take place in the absence of glucose uptake, the results of the study were not really conclusive, as the hxt strain used showed residual growth on glucose (Liang & Gaber, 1996).

Repression of some genes by glucose requires not only metabolism of the sugar, but also the presence of specific kinases. Our results show that lack of Hxk2 strongly decreases repression of GAL1 and has only a moderate effect on glucose repression of GDH2, confirming earlier work (Zimmermann & Scheel, 1977; Coshigano et al., 1991), and they indicate that simultaneous lack of Hxk1 and Hxk2 has different effects on different genes: glucose repression of GAL1 is completely relieved, repression of GDH2 is very weak, and repression of FBP1 or ICL1 is normal, as also reported for PCK1 (Yin, 1996). It appears, therefore, that catabolite repression of the gluconeogenic enzymes has peculiar characteristics.

We have shown that activation of plasma membrane H+-ATPase by glucose was independent of the presence or absence of Snf3, Rgt2 and Gpr1 (Table 5). Although it has been reported that ATPase activation is decreased in some snf3 mutants (Souza, 2001; Kotyk, 2003; Trópia, 2006), in other genetic backgrounds Snf3 is dispensable (Kotyk, 2003; this study). The irrelevance of Gpr1 for the glucose-induced activation of plasma membrane H+-ATPase supports previous results showing that this activation does not depend on cAMP changes (Mazón, 1989). Interestingly, whereas deletion of Gpr1 did not affect H+-ATPase activation by glucose (Souza, 2001; this study), deletion of Gpa2 decreased activation two-fold (Souza, 2001; Trópia, 2006), thus suggesting that, although Gpr1 interacts with Gpa2 (Yun, 1997), there is some signaling through Gpa2 that takes place independently of Gpr1. Taken together, the available results indicate that none of the plasma membrane glucose sensors plays a major role in triggering glucose activation of plasma membrane H+-ATPase. On the other hand H+-ATPase activation is very much impaired in the absence of glucose phosphorylation (Souza, 2001; this study), and is also compromised in a yeast strain lacking both Hxk1 and Hxk2 (Table 6) (F. Portillo, personal communication). This suggests that the activation depends on the glycolytic flux and possibly on changes in the concentration of intermediary metabolites. Such an interpretation is consistent with the observation that galactose is also able to activate the plasma membrane H+-ATPase (Rodríguez & Gancedo, 1999; this study).

Although the first stage of the inactivation of FbPase by glucose depends on phosphorylation of the enzyme by a cAMP-dependent protein kinase (Rittenhouse, 1987), it was not much affected by the lack of Gpr1. This indicates that although the increase in the intracellular concentration of cAMP, triggered by glucose, in a gpr1 mutant is smaller than that taking place in a wild-type strain (Kraakman, 1999a), it is sufficient to induce rapid phosphorylation of FbPase. In mutants unable to phosphorylate glucose, the cAMP increase is substantially reduced (Rolland, 2000), and, accordingly, FbPase is only partially inactivated. The small amount of inactivation observed in the hxk1 hxk2 glk1 gpr1 mutant, in conditions where cAMP levels have been reported not to change (Rolland, 2000), would depend on a different regulatory mechanism. This mechanism may involve the Rgt1 and Mth1 proteins, as the process of short-term inactivation is slowed down in a double mutant mth1gpr1 and in an rgt1 strain.

The finding of a role for Gpr1, complementary to that of Mth1, in the second phase of FbPase inactivation is consistent with a report of the involvement of a cAMP signaling pathway in the vacuolar-dependent degradation of FbPase (Hung, 2004). On the other hand, the role of Rgt1 in FbPase inactivation cannot be related to the known regulatory functions of Rgt1. Formally, Rgt1 could be required for the expression of elements of the pathways involved in degradation of FbPase, or it could repress elements inhibiting proteolysis. What is clear is that for the function of Rgt1, Mth1 is partially redundant with Std1, as occurs for the transcription of most HXT genes (Moriya & Johnston, 2004), and this would explain why the lack of Mth1 has a weaker effect on FbPase inactivation than the lack of Rgt1. It should be noted that, in the course of a systematic two-hybrid analysis (Ito, 2001), Mth1 was found to interact with Vid22, a protein involved in the uptake of FbPase from the cytoplasm to vesicles that will later fuse with the vacuole (Brown, 2002). It is not yet clear, however, whether the lack of Mth1 interferes with this function of Vid22.

Degradation of different proteins, triggered by glucose, has been shown to be dependent on glucose phosphorylation (Jiang, 2000; Hung, 2004), and our findings on FbPase inactivation are consistent with this. There have, however, been contradictory reports on the specific role of Hxk2 in the process; in one of them, Hxk2 was deemed necessary for proteolysis of FbPase to occur (Horak, 2002), whereas in the other, Hxk2 was not found to be necessary (Hung, 2004). In our conditions, we found that in the absence of both Hxk1 and Hxk2, the two phases of FbPase inactivation proceeded normally and that FbPase was proteolytically degraded. The different results obtained may be related to the use of different experimental conditions, as it has been demonstrated that, depending on the metabolic state of the yeast cells, FbPase is degraded through alternative pathways, via either the proteasome or the vacuole (Hung, 2004). The differences observed would be explained if the two pathways had different requirements for Hxk2.

Although most glucose effects require glucose metabolism, in an hxk1 hxk2 glk1 triple mutant some genes are still induced by glucose (Belinchón & Gancedo, 2007) and others are slightly repressed by glucose (Table 4). This may suggest that intracellular glucose itself can play a signaling role, but there is still no conclusive evidence for such a role, in the absence of hexokinases. In a wild-type strain the situation is different; glucose would modify the conformation of Hxk2, and could thereby affect its regulatory function (Kraakman, 1999b). An alternative explanation for the behavior of the hxk1 hxk2 glk1 triple mutant could be that the EMI2 (YDR516C) gene, a strong homolog of GLK1 (Lutfiyya, 1998), encodes a protein with some capacity to phosphorylate glucose. This seems unlikely, however, as the triple mutant does not grow on glucose, no glucose-phosphorylating activity can be detected in the corresponding extracts, and several amino acids that are strongly conserved in a variety of kinases able to phosphorylate glucose are different in the sequence reported for Emi2. Another possibility could be that the small amount of repression exerted by nonmetabolized glucose takes place through a cAMP signal, as appears to be the case for induction of SUC2 by glucose (Belinchón & Gancedo, 2007). However, whereas the SUC2 induction that took place in an hxk1 hxk2 glk1 mutant was blocked in an hxk1 hxk2 glk1 gpr1 mutant (Belinchón & Gancedo, 2007), in such a mutant FbPase, GlutDH and Icl were still partially repressed by glucose, indicating the existence of an as yet unidentified glucose signal independent of metabolism and able to cause a low level of repression.

The requirement for glucose metabolism observed in most processes triggered by glucose supports the idea that changes in the concentration of some intracellular metabolite(s) constitute an important factor in glucose signaling, although such metabolites have not been yet characterized. Although it has been proposed that glucose 6-phosphate could mediate glucose repression (Gancedo, 1998; Vincent, 2001), and it has been found that the intracellular concentration of fructose 1,6-bisphosphate is highest in the presence of glucose (Rodríguez & Gancedo, 1999; Belinchón & Gancedo, 2003), there is no correlation between the intracellular concentrations of these hexose phosphates and the degree of repression of FbPase by different carbon sources (Rodríguez & Gancedo, 1999; Belinchón & Gancedo, 2003).

We summarize in Fig. 5 how different signaling pathways mediate the effects of glucose presented here and in an earlier paper (Belinchón & Gancedo, 2007). It can be seen that the different changes triggered by glucose in yeast cannot be explained as the response to single signals, such as an induction signal, generated in a receptor-mediated process, or a repression signal, requiring the uptake and metabolism of glucose (Özcan, 2002). Instead, the picture that emerges is that different processes are dependent on the combination of a variety of signals.

5

The effects of glucose on different processes in yeast depend on different combinations of early signals. Induction of SUC2 by low glucose depends partially on the glucose sensors, and requires glucose metabolism; in the absence of metabolism, induction requires high glucose and is fully dependent on Gpr1. Induction of HXT1 requires Snf3 or Rgt2, and is partially dependent on Hxk2. Induction of PDC1 is independent of the glucose sensors, has some dependence on Hxk2 or Hxk1, and requires glucose metabolism. Repression of different genes is independent of the glucose sensors; for SUC2, GAL1 and GDH2, Hxk2 or Hxk1 are required, and presumably glucose metabolism; for FBP1 and ICL1, glucose metabolism is necessary, but any glucose-phosphorylating enzyme sustains repression. In the absence of glucose metabolism, phosphorylation of FbPase is impaired, and the effect is reinforced by lack of Gpr1; degradation of FbPase depends on glucose metabolism, and is decreased when Gpr1 is lacking. Activation of H+-ATPase is independent of the glucose sensors, and has a strong dependence on Hxk2 or Hxk1; it is nearly blocked in the absence of glucose metabolism. The interactions between the initial elements of signaling and their targets are indicated by arrows; the width of the arrows is related to the strength of the signal, and putative interactions are indicated by dashed lines. Embedded Image, induction; Embedded Imagerepression; Embedded Imageactivation; Embedded Imageinactivation (by phosphorylation or degradation).

Supplementary material

Fig. S1. Derepression kinetics of FbPase in yeast strains unable to phosphorylate glucose. Strains THG1 (hxk1 hxk2 glk1) and MMB2 (hxk1 hxk2 glk1 gpr1) were grown in YPgalactose and derepressed in YPethanol or YPethanol + 2% glucose as described in ‘Materials and methods’.

Acknowledgements

We thank H. Tamaki (Kyoto University) for plasmid pGPR1-RS416, D.G. Fraenkel (Harvard University) for a set of yeast strains, and F. Moreno (Universidad de Oviedo) for strain DLY1901. We are indebted to F. Portillo for advice in measuring activation of the H+-ATPase and for communication of unpublished results. Critical reading of the manuscript by C. Gancedo and C.L. Flores is gratefully acknowledged. This work was supported by grants BMC2001-1690-CO2-1 and BFU2004-02855-C02-01 from the Dirección General de Investigación Científica y Técnica, and by the EU project BIO-HUG QLK3-CT-1999-00080. M.M. Belinchón was supported by fellowships from the EU and the DGICYT.

Footnotes

  • Editor: Monique Bolotin-Fukuhara

References

View Abstract