A Saccharomyces cerevisiae G-protein coupled receptor, Gpr1, is specifically required for glucose activation of the cAMP pathway during the transition to growth on glucose

Authors

  • Leon Kraakman,

    1. Laboratorium voor Moleculaire Celbiologie, Katholieke Universiteit Leuven, Institute of Botany and Microbiology, and,
    Search for more papers by this author
    • †These authors contributed equally to this study

  • Katleen Lemaire,

    1. Laboratorium voor Moleculaire Celbiologie, Katholieke Universiteit Leuven, Institute of Botany and Microbiology, and,
    2. Flanders Interuniversity Institute for Biotechnology — VIB, Kardinaal Mercierlaan 92, B-3001 Leuven-Heverlee, Flanders, Belgium.
    Search for more papers by this author
    • †These authors contributed equally to this study

  • Pingsheng Ma,

    1. Laboratorium voor Moleculaire Celbiologie, Katholieke Universiteit Leuven, Institute of Botany and Microbiology, and,
    Search for more papers by this author
  • Aloys W.R.H. Teunissen,

    1. Laboratorium voor Moleculaire Celbiologie, Katholieke Universiteit Leuven, Institute of Botany and Microbiology, and,
    Search for more papers by this author
  • Monica C.V. Donaton,

    1. Laboratorium voor Moleculaire Celbiologie, Katholieke Universiteit Leuven, Institute of Botany and Microbiology, and,
    Search for more papers by this author
  • Patrick Van Dijck,

    1. Laboratorium voor Moleculaire Celbiologie, Katholieke Universiteit Leuven, Institute of Botany and Microbiology, and,
    2. Flanders Interuniversity Institute for Biotechnology — VIB, Kardinaal Mercierlaan 92, B-3001 Leuven-Heverlee, Flanders, Belgium.
    Search for more papers by this author
  • Joris Winderickx,

    1. Laboratorium voor Moleculaire Celbiologie, Katholieke Universiteit Leuven, Institute of Botany and Microbiology, and,
    Search for more papers by this author
  • Johannes H. De Winde,

    1. Laboratorium voor Moleculaire Celbiologie, Katholieke Universiteit Leuven, Institute of Botany and Microbiology, and,
    Search for more papers by this author
  • Johan M. Thevelein

    1. Laboratorium voor Moleculaire Celbiologie, Katholieke Universiteit Leuven, Institute of Botany and Microbiology, and,
    Search for more papers by this author

Johan M. Thevelein. E-mail Johan.Thevelein@bio.kuleuven.ac.be; Tel. (+32) 16 321507; Fax (+32) 16 321979.

Abstract

In the yeast Saccharomyces cerevisiae the accumulation of cAMP is controlled by an elaborate pathway. Only two triggers of the Ras adenylate cyclase pathway are known. Intracellular acidification induces a Ras-mediated long-lasting cAMP increase. Addition of glucose to cells grown on a non-fermentable carbon source or to stationary-phase cells triggers a transient burst in the intracellular cAMP level. This glucose-induced cAMP signal is dependent on the G alpha-protein Gpa2. We show that the G-protein coupled receptor (GPCR) Gpr1 interacts with Gpa2 and is required for stimulation of cAMP synthesis by glucose. Gpr1 displays sequence homology to GPCRs of higher organisms. The absence of Gpr1 is rescued by the constitutively activated Gpa2Val-132 allele. In addition, we isolated a mutant allele of GPR1, named fil2, in a screen for mutants deficient in glucose-induced loss of heat resistance, which is consistent with its lack of glucose-induced cAMP activation. Apparently, Gpr1 together with Gpa2 constitute a glucose-sensing system for activation of the cAMP pathway. Deletion of Gpr1 and/or Gpa2 affected cAPK-controlled features (levels of trehalose, glycogen, heat resistance, expression of STRE-controlled genes and ribosomal protein genes) specifically during the transition to growth on glucose. Hence, an alternative glucose-sensing system must signal glucose availability for the Sch9-dependent pathway during growth on glucose. This appears to be the first example of a GPCR system activated by a nutrient in eukaryotic cells. Hence, a subfamily of GPCRs might be involved in nutrient sensing.

Introduction

In the yeast Saccharomyces cerevisiae, the addition of glucose to cells previously grown in the absence of glucose triggers a rapid, transient increase in the cAMP level, which sets off a protein kinase A (PKA)-mediated protein phosphorylation cascade. This cascade triggers a variety of physiological events, such as mobilization of trehalose and glycogen and rapid loss of stress resistance. Maintenance of these PKA-controlled characteristics during growth on glucose requires an additional pathway, the fermentable growth medium (FGM)-induced pathway, which involves at least in part the Sch9 protein kinase (Thevelein, 1994).

Adenylate cyclase activity in yeast is highly dependent on the activity of the Ras1 and Ras2 proteins, the activity status of which is dependent on the guanine nucleotide exchange proteins Cdc25 and Sdc25. In the absence of Ras or Cdc25, adenylate cyclase activity is very low and the resulting cAMP depletion causes growth arrest and permanent entry into the stationary phase G0 (Wigler et al., 1988; Broach and Deschenes, 1990; Tatchell, 1993). Genetic evidence indicated that the Ras and Cdc25 proteins were also required for the generation of the glucose-induced rise in the cAMP level and for the potent stimulation of cAMP synthesis by intracellular acidification (Mbonyi et al., 1988; Munder and Küntzel, 1989; Van Aelst et al., 1990; Van Aelst et al., 1991; Bhattacharya et al., 1995). However, we have recently shown that only intracellular acidification and not glucose triggers an increase in the GTP content of the Ras proteins. Stimulation of cAMP synthesis by glucose depends on another G-protein, Gpa2 (Colombo et al., 1998). We have now isolated the C-terminus of a putative G-protein-coupled receptor, Gpr1, in a yeast two-hybrid system using Gpa2 as bait, as recently reported by others (Yun et al., 1997; Xue et al., 1998), and we show that Gpr1 is essential for glucose activation of cAMP synthesis.

The initiation of fermentation by the addition of glucose to yeast cells growing on non-fermentable carbon sources using gluconeogenesis and respiration or to stationary-phase cells, triggers a rapid decrease in stress resistance (Van Dijck et al., 1995; de Winde et al., 1997). We have screened for mutants that are deficient in this decrease in stress resistance using a heat shock applied shortly after the initiation of fermentation as the selection protocol. The mutants isolated in this way are called ‘fil ’ mutants for fermentation-induced loss of stress resistance. The fil1 mutant has been identified as a partially inactivating mutation in the CYR1/CDC35 gene encoding adenylate cyclase (P. Van Dijck, P. Ma, M. Versele, M. F. Gorwa et al., unpublished). We now show that the fil2 mutant contains a nonsense mutation in the GPR1 gene.

Activation of the cAMP pathway by glucose is a transient phenomenon. Maintenance of the ‘activated PKA phenotype’ in the presence of glucose is closely correlated with continued growth on glucose and therefore requires the continued presence of all other essential nutrients (Thevelein, 1994). It has been demonstrated that readdition of nitrogen, phosphate or sulphate to yeast cells starved of these nutrients on glucose medium also causes activation of PKA targets, resulting in rapid re-establishment of the ‘activated PKA phenotype’ during reinitiation of growth (Hirimburegama et al., 1992). In this case nitrogen activation of trehalase and rapid repression of STRE-controlled messengers requires the activity of the Sch9 protein kinase (Crauwels et al., 1997). A large number of yeast genes that are induced by a variety of stress conditions contain one or more CCCCT elements in their promoter, which is responsible for the majority of the stress effects and is referred to as a STRE element (‘stress response element’) (Ruis and Schuller, 1995; Moskvina et al., 1998). The SCH9 gene was isolated as a multicopy suppressor of the lethality caused by cAMP depletion and is homologous to the TPK genes encoding the catalytic subunits of PKA (Toda et al., 1988). In this paper we show that glucose sensing for activation of the Sch9-dependent FGM pathway does not require the Gpr1–Gpa2 system. Hence, an alternative glucose-sensing system for activation of this pathway has to be present.

Results

Gpr1 interacts with Gpa2 in a two-hybrid screen

Because of the requirement of Gpa2 for glucose-induced activation of cAMP synthesis (Colombo et al., 1998), we have performed a two-hybrid screen to identify proteins that interact with Gpa2. We have isolated clones from five different genes, including a clone encoding the last 121 amino acids of the C-terminus of the GPR1 (YDL035c)-encoded protein, as recently reported by others (Yun et al., 1997; Xue et al., 1998). This protein has the typical structural features of a G-protein coupled receptor (GPCR), including seven transmembrane domains (TMs), extended N- and C-termini and a relatively large third intracellular loop (Dohlman et al., 1991; Strader et al., 1994) (Fig. 1A). Because Gpr1 also bares considerable sequence similarity to several types of GPCRs of higher eukaryotes, in particular in and around TM regions, mechanisms of ligand interaction and receptor activation may be similar. As an example, sequence similarity with adenosine-A2B receptors of mammalian and avian origin is shown in 1Fig. 1B and C.

Figure 1.

. The yeast G-protein-coupled receptor Gpr1 displays sequence similarity with GPCRs of higher eukaryotes. A. Putative structure modelling of Gpr1. The positions and orientations of the seven transmembrane domains are predicted by the Tmpred (Hofmann and Stoffel, 1993) and TopPred 2 (von Heijne, 1992) algorithms. Membrane-proximal amino acids are numbered according to their position in the protein sequence. A long poly asparagine track within the large third intracellular loop is indicated in yellow. The C-terminal 99-amino-acid peptide interacting with Gpa2 and the extension to the original 122-amino-acid peptide are indicated in dark and light pink respectively. The position of the nonsense mutation in fil2 is depicted in green. Sequence identities and similarities with adenosine A2B receptors, as shown in B and C are also depicted here using the same colour coding. B. Protein sequence alignment of adenosine A2B receptors (AA2B) from mammalian and avian origin, and yeast Gpr1. Alignments were obtained with blast2 and Psi-blast algorithms and extended manually. The predicted positions of the transmembrane domains are indicated with green boxes. An uncertainty about the C-terminal end of TM2 of Gpr1 is depicted with a stippled box (B). Identical amino acids are shown in dark blue and conservative substitutions in light blue. A tryptophan in TM6 and a tyrosine in TM7, which are highly conserved among many GPCRs (Baldwin, 1993), are shown in red (C).

Gpr1 is required for glucose-induced cAMP signalling

The addition of glucose to yeast cells grown on a non-fermentable carbon source, such as glycerol, triggers a rapid rise in the intracellular cAMP level. This cAMP response is largely absent in the gpr1Δ mutant and was observed in at least 10 independent experiments. A typical example is shown in 2Fig. 2A. As previously shown for the gpa2Δ strain (Colombo et al., 1998), the residual activation can be eliminated by preaddition of a small amount of glucose (data not shown) and is therefore probably due to interference with the strong stimulating effect of low intracellular pH on cAMP synthesis (Argüelles et al., 1990). The gpr1Δgpa2Δ double mutant behaved as the single gpr1Δ and gpa2Δ strains (Fig. 2A), indicating that both proteins act in the same pathway. Activation of cAMP synthesis by intracellular acidification (addition of 2,4-dinitrophenol at pH 6.0) was much less affected by deletion of GPR1 and/or GPA2 than the effect on glucose activation in the corresponding strains. This effect is probably indirect because there was always a clear cAMP response present upon acidification (Fig. 2B). The lack of a cAMP signal in the gpr1Δ strain was rescued by expression of the constitutively activated GPA2Val-132 allele (Fig. 2A), which we constructed on the basis of homology with mammalian G-proteins (Sprang, 1997). The gpr1ΔGPA2val-132 strain displayed a cAMP signal very similar to that in the wild-type strain. Hence, as has been described for other G alpha-proteins, Gpa2Val-132 functions independently of its cognate receptor (Gpr1) in stimulating cAMP synthesis. However, the constitutively activated Gpa2Val-132 allele fulfils only one of the two requirements for activation of cAMP synthesis by glucose. A second requirement is glucose phosphorylation (Beullens et al., 1988). Therefore, the cAMP level starts at a low value and increases only after the second requirement, glucose phosphorylation, has been fulfilled (see Discussion ).

Figure 2.

. Stimulation of cAMP accumulation by glucose and by intracellular acidification triggered by 2,4-dinitrophenol. Glucose (100 mM) (A) or 2 mM 2,4-dinitrophenol (B) was added to glycerol-grown cells of GPR1 and GPA2 mutants. Strains: wild-type (W303–1A) (●), gpr1Δ (LK5) (○), fil2 (KL1) (▴), gpa2Δ (PM731) (Δ), gpr1Δgpa2Δ (LK6) (▪), GPA2 val-132 (PM735) (□), gpr1ΔGPA2 val-132 (LK7) (♦).

The fil2 mutation is allelic with GPR1

A mutant allele of GPR1 was independently isolated in a screen for mutants that are specifically deficient in the rapid loss of high stress resistance in stationary-phase cells upon initiation of fermentation. This fil2 mutant was deficient in glucose-induced stimulation of cAMP synthesis (Fig. 2A), but not in stimulation by intracellular acidification (Fig. 2B). fil2 was genetically mapped close to the centromere of chromosome IV, and complementation and tetrad analysis confirmed that fil2 is allelic to GPR1 (results not shown). The fil2 mutation is an ochre nonsense mutation of codon 316 (Tyr) in the third intracellular loop of Gpr1 (Fig. 1A). Both the fil2 mutant and the gpr1Δ mutant are deficient in glucose-induced loss of heat resistance (Fig. 3A) and freeze resistance (results not shown). Replacement of Gpa2 by constitutively active Gpa2val-132 in the wild type, the gpr1Δ strain and the fil2 strain resulted in a lower level of heat resistance in all strains before and after the addition of glucose to derepressed cells (results not shown).

Figure 3.

. Effect of mutations in the Gpr1–Gpa2 system on glucose-induced loss of heat resistance and mobilization of trehalose and glycogen. A. Heat resistance (log survival percentage after a heat shock of 15 min at 51°C ). B. Trehalose and glycogen level. Strains: wild type (W303-1A) (●) gpr1Δ (LK5) (○), fil2 (KL1) (▴), gpa2Δ (PM731) (▵) and gpr1Δ gpa2Δ (LK6) (▪).

An important characteristic of the transition from the derepressed to the fermentative state is mobilization of the reserve carbohydrates trehalose and glycogen (van der Plaat, 1974; François et al., 1988). This process is controlled by the cAMP–PKA pathway and is delayed by inactivation of the Gpr1–Gpa2 GPCR system. The glucose-induced loss of heat resistance and mobilization of trehalose and glycogen were clearly faster in the wild-type strain than in all mutant strains (Fig. 3B). However, from the results shown in 4Fig. 4A, it is clear that the overall fluctuation of cAPK-controlled features (levels of trehalose, glycogen and heat resistance) during growth on glucose is still present in mutants lacking a functional Gpr1–Gpa2 system. This indicates that the GPCR system is not essential for the sustained response to glucose in the growth medium, but only for stimulation during the transition to growth on glucose. Two independent experiments were carried out in which all parameters were measured simultaneously for all strains. The results of a typical experiment are shown in Figs 3 and 4.

Figure 4.

. Effect of mutations in the Gpr1–Gpa2 system on PKA-controlled properties during diauxic growth on glucose. A. Growth (OD600), glucose consumption (glucose level in the medium), trehalose content, glycogen content and heat resistance (log survival percentage after 15 min at 51°C) of the cells are shown as a function of time. Strains: wild type (W303–1A) (●), gpr1Δ (LK5) (○), fil2 (KL1) (▴), gpa2Δ (PM731) (▵) and gpr1Δgpa2Δ (LK6) (▪). B. Northern blot analysis of the expression of the STRE-controlled genes HSP12 and SSA3 as well as the ribosomal protein gene RPL25 as a function of time during diauxic growth on glucose. 18S ribosomal RNA was used as standard. (In stationary phase all standards decline to some extent.) The samples for RNA extraction were taken from the same culture of which the biochemical parameters are shown in A.

Glucose repression of STRE-controlled genes and induction of ribosomal protein genes is delayed in gpr1 mutants

Another typical feature associated with glucose-induced activation of the Ras–cAMP pathway is the rapid shutdown of expression of genes that are controlled by STRE elements in their promoters (Ruis and Schuller, 1995; Moskvina et al., 1998). The glucose-dependent disappearance of SSA3 and HSP12 mRNAs is delayed in the gpr1Δ and gpr1Δgpa2Δ strains (Fig. 4B), as previously shown for the gpa2Δ mutant (Colombo et al., 1998). The addition of glucose to cells grown in the absence of glucose is known to trigger a rapid increase in ribosomal protein gene expression, which to some extent is also dependent on glucose activation of the cAMP pathway (Kraakman et al., 1993; Klein and Struhl, 1994; Griffioen et al., 1996; Crauwels et al., 1997). This increase in ribosomal protein gene expression (RPL25 ) is delayed in the gpr1Δ and gpr1Δgpa2Δ strains (Fig. 4B). The results show that the overall fluctuation in the expression of the STRE-controlled genes and the ribosomal protein genes during growth on glucose is still present in mutants in the Gpr1–Gpa2 system.

Gpr1 acts independently of the Sch9 protein kinase and nitrogen activation of the FGM pathway

Activation of the cAMP pathway by glucose is a transient phenomenon. Maintenance of the ‘activated protein kinase A phenotype’ (Thevelein, 1994) during growth on glucose is controlled by another glucose-activated pathway, the so-called ‘fermentable growth medium-induced (FGM) pathway’, which is dependent on protein kinase Sch9 (Crauwels et al., 1997). It has been proposed that Gpr1 acts upstream of Sch9 rather than upstream of adenylate cyclase, because a constitutively active GPA2 Ala-273 allele was unable to reduce heat resistance in an sch9Δ strain, in contrast to a constitutively active RAS2 val-19 allele (Xue et al., 1998). We now show that this conclusion does not appear to be correct.

Deletion of GPR1, GPA2, or both does not affect the growth rate of yeast cells, as opposed to deletion of SCH9 (Figs 4A and 5A and B). However, additional deletion of GPR1 or GPA2 in an sch9Δ strain nearly eliminates growth on glucose (Fig. 5A and B). These results make it unlikely that Gpr1–Gpa2 and Sch9 act in the same signal transduction pathway. When yeast cells are starved of nitrogen on a glucose-containing medium, they arrest in G0 and accumulate high amounts of trehalose. Readdition of a nitrogen source, e.g. asparagine, triggers rapid activation of trehalase and mobilization of trehalose, which is largely dependent on the presence of glucose but is not mediated by cAMP (Hirimburegama et al., 1992; Durnez et al., 1994). This activation of trehalase was still present in all GPR1–GPA2 mutants tested (Fig. 5C), in contrast to the deletion of SCH9, which abolishes the initial activation (Fig. 5C) (Crauwels et al., 1997). These results indicate that the Gpr1–Gpa2 GPCR system is not required for nitrogen activation of the Sch9-dependent FGM signalling pathway and, in particular, it is not required for the glucose dependency of this nitrogen activation process.

Figure 5.

. Gpr1 and Gpa2 do not act in the same signalling pathway as Sch9. A. Tetrad analysis of segregants obtained after sporulation of a heterozygous GPA2/gpa2ΔSCH9/sch9Δ strain (PM-D2). The colonies obtained from the spores were allowed to grow for 3 days or 7 days, as indicated. The genotype of the segregants was determined on the basis of the auxotrophic markers used to delete the genes. The undetectable (3 days) or smallest (7 days) colonies all had the gpa2Δsch9Δ genotype. The smallest (3 days), or second smallest (7 days) colonies all had the sch9Δ genotype. The gpa2Δ colonies had the same size as those of the wild-type strain. An example of the assignment of the genotype is shown at the right. B. Tetrad analysis of segregants obtained after sporulation of a heterozygous GPR1/gpr1ΔSCH9/sch9Δ strain (PM-D1). The genotype of the segregants was assigned in the same way as in A and the colony sizes of the gpr1Δ and the gpr1Δsch9Δ strains corresponded, respectively, to those of the gpa2Δ and the gpa2Δsch9Δ strains as shown in A. C. Nitrogen-induced activation of trehalase in nitrogen-starved glucose-repressed cells. l-Asparagine (10 mM) and the essential amino acids were added at time zero in a glucose-containing nitrogen starvation medium. Strains: wild type (W303–1A) (●), gpr1Δ (LK5) (○), fil2 (KL1) (▴), gpa2Δ (PM731) (Δ), gpr1Δgpa2Δ (LK6) (▪) and sch9Δ (PM-S1) (□).

Discussion

The Gpr1–Gpa2 GPCR system for glucose activation of the cAMP pathway

In this study we have identified a putative yeast GPCR that is required for activation of cAMP synthesis by glucose. Together with our previous results on the requirement of Gpa2 for glucose activation of cAMP synthesis (Colombo et al., 1998) and the evidence for physical interaction between Gpr1 and Gpa2 (Yun et al., 1997; Xue et al., 1998) (this work), the available data suggest that glucose sensing for activation of the cAMP pathway in S. cerevisiae is carried out by a G-protein coupled receptor system, consisting of the receptor Gpr1 and its associated G-protein Gpa2. Based on known mechanisms of GPCR functioning in yeast and other eukaryotes (Dohlman et al., 1991; Drayer and Van Haastert, 1994; Bourne, 1997), the most straightforward model explaining our results is that Gpr1 is activated by glucose and transmits a signal via Gpa2 to adenylate cyclase. Recent work in our laboratory has confirmed that Gpr1 is required specifically for the extracellular sensing of glucose. It has been possible to trigger activation of cAMP accumulation by the addition of glucose in a multiple glucose carrier deletion strain devoid of significant glucose uptake provided that a low level (< 1 mM) of maltose was added to the cells before the glucose. In this way the requirement of glucose phosphorylation for stimulation of cAMP synthesis by glucose was fulfilled (Rolland et al., manuscript in preparation). Further work, however, is needed to unravel the detailed mechanism of glucose sensing by this GPCR system and in particular to provide definite proof that Gpr1 binds glucose.

It has been well established that glucose activation of cAMP synthesis requires active sugar phosphorylation but no further metabolism of the glucose (Beullens et al., 1988). Because there is no correlation between the glucose-induced increases in sugar phosphates and cAMP, the sugar phosphates do not seem to act as ‘metabolic messengers’ for cAMP signalling. Therefore, much attention has been focused on a possible role of the sugar kinases and possible interacting proteins in the glucose-sensing mechanism for activation of the cAMP pathway (Thevelein, 1992). The recent, unexpected finding that a GPCR system is involved in glucose activation of cAMP synthesis raises the question of why and how glucose phosphorylation is required for signal transmission to adenylate cyclase. At present, the precise connection between sugar phosphorylation and the Gpr1–Gpa2 adenylate cyclase system remains unclear.

The double requirement for an active Gpr1–Gpa2 system and glucose phosphorylation is important for the proper interpretation of certain experimental results. For instance, the restoration of the glucose-induced cAMP signal in the gpr1Δ strain by expression of the constitutively active Gpa2val-132 allele is difficult to understand when based only on the known properties of GPCR systems. Normally, one would expect the cAMP level to be constitutively high and all glucose-induced cAMP increases to be absent. However, when the known requirement for glucose phosphorylation is taken into account, it is understandable that the initial cAMP level is low, because at that time there is no glucose phosphorylation. Before glucose addition, only the requirement for G-protein stimulation of adenylate cyclase is fulfilled by the presence of the dominant Gpa2val-132 allele. When glucose is added, it is taken up and phosphorylated, which fulfils the second requirement. As a result, the cAMP level rapidly increases and a cAMP signal resembling the signal in a wild type cell is generated.

Nutrient sensing by GPCRs

GPCRs are known to mediate cellular responses to diverse extracellular stimuli. To date, GPCRs have been identified for a wide range of hormones, neurotransmitters, pheromones, light, odorants and chemoattractants (Dohlman et al., 1991; Strader et al., 1994). No GPCRs involved in nutrient sensing have been reported. Hence, the glucose-sensing Gpr1–Gpa2 system for activation of the cAMP pathway in S. cerevisiae appears to be the first example of a nutrient-sensing GPCR system. If nutrient-sensing GPCRs were common to eukaryotic cells, they would provide a means for regulation of major signal transduction pathways by the nutrient status in the cellular environment. Glucose-sensing GPCRs might be involved in specific glucose detection by other cell types such as pancreatic beta cells (Matschinsky et al., 1998) and photosynthetic plant cells (Jang and Sheen, 1997; Smeekens and Rook, 1997).

Two different pathways mediate the transient and maintained effects on the PKA targets

The glucose-induced cAMP increase is a very transient and short-lived phenomenon, and there is at present no conclusive evidence that after prolonged incubation in the presence of glucose and, in particular, that during growth on glucose the cAMP level remains significantly higher than in cells deprived of glucose (Ma et al., 1997). On the other hand, several results have been obtained that support the notion that the maintenance of the PKA-controlled physiological changes during growth on glucose is not due to glucose activation of cAMP synthesis (for review see Thevelein, 1994). The recent demonstration that deletion of Gpa2 abolishes glucose-induced cAMP signalling without even reducing the fluctuation in the PKA-controlled properties during diauxic growth on glucose has provided further strong evidence against involvement of glucose activation of cAMP synthesis (Colombo et al., 1998). Our present results show that the complete Gpr1–Gpa2 glucose-sensing system can be eliminated without eliminating the glucose-induced changes in the PKA-controlled properties during growth on glucose. This further indicates that another glucose-sensing pathway must be responsible for the maintenance of these changes.

Our new results are in good agreement with the concept that two pathways are responsible for the changes in the PKA targets. A rapid response pathway, consisting of glucose activation of cAMP synthesis by the Gpr1–Gpa2 system, and a maintenance pathway, previously called the fermentable growth medium-induced pathway, which requires the functioning of the Sch9 protein kinase (Thevelein, 1994). The growth data with the deletion mutants indicate that Gpr1–Gpa2 and Sch9 do not act in the same signal transduction pathway, at least not in a simple linear fashion. Xue et al. (1998) suggested that Gpr1 acts upstream of Sch9 rather than upstream of adenylate cyclase. Presumably, overactivation by the GPA2 Ala-273 allele, which they used in their experiments, is weaker than overactivation by RAS2 val-19, which would explain why only the latter was able to suppress the heat-resistant phenotype of sch9Δ. The data on the changes in the PKA-controlled properties after the addition of glucose to the Gpr1–Gpa2 deletion mutants indicate that the Gpr1–Gpa2 GPCR system is only required for stimulation of these changes and not for their maintenance. Moreover, our isolation of a gpr1 mutant in a screen for mutants with a slower loss of heat resistance during the initiation of fermentation supports the physiological relevance of Gpr1 for rapid adaptation to the presence of glucose. Hence, the Gpr1–Gpa2 system is only responsible for triggering transient glucose-induced cAMP signalling, which is only involved in rapid adaptation effects, whereas Sch9 controls, at least in part, the long-term response during fermentative growth. Only in the absence of Sch9 does the Gpr1 pathway becomes important for growth. It is then apparently able to sustain residual growth of the cells. A model outlining the two distinct pathways for transient and sustained activation of the PKA targets is depicted in Fig. 6.

Figure 6.

. Model of the two pathways responsible for glucose control of the PKA targets in S. cerevisiae. The Gpr1–Gpa2 GPCR system activates cAMP synthesis in response to glucose immediately after its addition, and this effect is transient. The resulting increase in cAMP stimulates the rapid adaptation to the presence of glucose in the growth medium. The second pathway does not involve Gpr1–Gpa2 but another unknown glucose-sensing system. This so-called ‘fermentable growth medium-induced (FGM) pathway’ requires the Sch9 protein kinase and is responsible for sustained activation of the PKA targets during growth on glucose. In addition to glucose, it requires nitrogen, phosphate and all other nutrients essential for growth. The sensing mechanisms for the other nutrients and the precise connection between the Sch9 protein kinase and PKA are unclear.

cAMP signalling and fermentation-induced loss of stress resistance

The fil2 mutant described in this work was isolated in a screen for mutants that maintain a high stress resistance during active fermentation. Industrial yeast strains with the fil phenotype would be highly useful for commercial applications of yeast where high stress resistance and vigorous fermentation capacity are desired simultaneously (e.g. frozen doughs), a combination of properties that until now appeared ‘against biological design’ (Attfield, 1997). Genetic modification of the Gpr1–Gpa2 glucose-sensing system is therefore promising for the construction of industrial yeast strains with improved stress resistance during active fermentation.

Experimental procedures

Strains and growth media

Saccharomyces cerevisiae strains used in this study were W303-1A (MATaleu2-3112 ura3-1 trp1-92 his3-11,15 ade2-1 can1-100 GAL SUC mal ), isogenic haploid derivatives PM731 (W303-1A gpa2 ::URA3 ) (Colombo et al., 1998), PM735 (W303-1A GPA2Val-132), KL1 (W303-1A fil2 ), LK5 (W303-1A gpr1::LEU2 ), LK6 (W303-1A gpr1::LEU2 gpa2 ::URA3 ), LK7 (W303-1A gpr1::URA3 GPA2Val-132), PM-S1 (W303-1A sch9 ::URA3 ), and isogenic diploid derivatives PM-D1 (W303 MATa/α GPR1/gpr1::LEU2 SCH9/sch9 ::URA3 ) and PM-D2 (W303 MATa/αGPA2/gpa2 ::LEU2 SCH9/sch9 ::URA3 ). Yeast strains for two-hybrid analysis were pJ69-4A (MATaleu2-3112 ura3-52 trp1-901 his3-200 gal4Δ gal80Δ GAL2-ADE2 lys2 ::GAL1-HIS3 met2 ::GAL7-LacZ ) (James et al., 1996) and SFY526 (MATaleu2-3112 ura3-52 trp1-901 his3-200 ade2-101 lys2-801 canR gal4-542 gal80-538 ura3 ::GAL1-LacZ ) (Bartel and Fields, 1995).

Yeast-rich growth medium contained 1% (w/v) yeast extract, 2% (w/v) Bacto-peptone and either 4% (w/v) d-glucose (YPD), or 2% (w/v) ethanol, 2% (w/v) glycerol and 0.1% (w/v) d-glucose (YPEG). Alternatively, in two-hybrid assays, cells were grown on selective minimal medium, containing 0.67% (w/v) Difco yeast nitrogen base w/o amino acids and 2% (w/v) d-glucose, supplemented with appropriate amino acids or uracil depending on plasmid selection.

Two-hybrid constructs and screening

Constructs for yeast two-hybrid analysis (Bartel and Fields, 1995) were made by PCR cloning of the corresponding genes under study; PCR amplification was performed using the Takara ExTaq DNA Polymerase Kit or the Expand Long Template PCR Kit (Roche/Boehringer), according to the manufacturer's instructions with modifications. The complete coding regions of GPA1, GPA2 and STE18 were cloned as BamHI/Pst I fragments, and of STE4 as an EcoRI/BamHI fragment, into yeast two-hybrid fusion vectors pGBT9, pGAD424 and pAS2-1 (Clontech). Adequate cloning and expression of fusion proteins were checked by sequencing and Western blot analysis, using anti-Gal4-DBD antibody (Santa Cruz, SC-577). Two-hybrid fusion constructs were transformed into strain pJ69-4A (James et al., 1996) and protein–protein interaction was monitored by dilution spot assays on selective media lacking histidine or adenine. Interactions between Gpa1 and Ste4, and between Ste4 and Ste18 (Clark et al., 1993; Hirschman et al., 1997) were used as positive controls in subsequent screening procedures. Screens for proteins interacting with Gpa2 were performed by transformation of pJ69-4A containing pGBT9-GPA2, with mixed genomic libraries Y2HL-C1, -C2 and -C3 (James et al., 1996). Interaction candidates were obtained by selecting for histidine and adenine prototrophy. False positives were removed after plasmid curing. Specific interaction with Gpa2 was verified after retransformation of isolated library fusion constructs into pJ69-4A and SFY526 containing pGBT9-GPA2, stringent monitoring of adenine and histidine prototrophy, and assaying X-gal-induced β-galactosidase expression. Positive candidates were characterized by direct PCR cycle sequencing of genomic fragments fused in the library vector. In this way, Gpa2-interacting fragments were identified from five different genes (L. Kraakman and Johanes H. de Winde, unpublished results). A peptide consisting of the last 121 amino acids of Ydl035c/Gpr1 was found to interact specifically with Gpa2. The fused coding fragment in pGAD-C1 was truncated using an internal SmaI restriction site. The resulting peptide consisting of the last 99 amino acids of Gpr1 also interacted with Gpa2.

Mutant selection

Mutants deficient in fermentation-induced loss of stress resistance were obtained by treatment of culture samples with a heat shock of 30 min at 56°C, 90 min after the initiation of fermentation by addition of glucose. Subsequently, the cells were grown to stationary phase. This treatment was repeated five times, after which several much better surviving mutants were obtained (P. Van Dijck et al. unpublished).

Assays of stress resistance

In samples taken from the cultures at the indicated time points, heat shock and freeze resistance of yeast cells were determined after the addition of glucose to stationary-phase cultures. For the determination of heat resistance, cells were incubated for 15 min at 51°C. After cooling, aliquots were spread on YPD plates and surviving cells were counted as colony-forming units after 2 days of growth at 30°C. For assay of freeze resistance, cells were quickly frozen in ethanol at −30°C and kept at −30°C for 24 h, thawed at 30°C and frozen again for 24 h. This treatment was repeated four times, after which aliquots were spread on YPD plates and surviving cells were counted as colony-forming units after 2 days of growth at 30°C.

Biochemical determinations

The activity of neutral trehalase was determined in crude cell extracts as previously described (Pernambuco et al., 1996). The specific activity of trehalase is expressed as nmol of glucose liberated per minute per mg of protein. Intracellular levels of trehalose, glycogen and cAMP were determined as previously described (Colombo et al., 1998).

RNA extraction and Northern blot analysis

Isolation of total RNA and Northern analysis were performed essentially as previously described (Crauwels et al., 1997). Probes were PCR-generated fragments carrying HSP12, SSA3 or RPL25. 18S rRNA was used as internal standard. Northern blots were analysed using phosphorimager technology (Fuji, BAS-1000; software, PCBAS 2.0).

Footnotes

  1. †These authors contributed equally to this study

Acknowledgements

We are most grateful to Philip James (Madison) for supplying strains and libraries for two-hybrid screening. We also thank Willy Verheyden and Renata Wicik for excellent technical assistance. This work was supported by a fellowship from the Fund for Scientific Research — Flanders (Senior research assistant) to J.W., from the CNPq (Brazil) to M.C.V.D., and by grants from the Fund for Scientific Research — Flanders to J.W and J.M.T, the Research Fund of the Katholieke Universiteit Leuven (Concerted Research Actions), Interuniversity Attraction Poles network P4/30, the Flemish Ministry of Economy through the Institute for Scientific and Technological Research (IWT/EUREKA) and the Flanders Interuniversity Institute for Biotechnology — VIB to J.M.T.

Ancillary