Molecular Microbiology (2000) 38(2), 348±358

Glucose-induced cAMP signalling in yeast requiresboth a G-protein coupled receptor system forextracellular glucose detection and a separablehexose kinase-dependent sensing process

Filip Rolland,1 Johannes H. de Winde,1

Katleen Lemaire,1,2 Eckhard Boles,3

Johan M. Thevelein1 and Joris Winderickx1*1Laboratorium voor Moleculaire Celbiologie, Katholieke

Universiteit Leuven, Kardinaal Mercierlaan 92,

B-3001 Leuven-Heverlee, Flanders, Belgium.2Flanders Interuniversity Institute for Biotechnology, VIB,

Kardinaal Mercierlaan 93, B-3001 Leuven-Heverlee,

Flanders, Belgium.3Institut fuÈr Mikrobiologie, Heinrich-Heine-UniversitaÈ t,

UniversitaÈ tsstr. 1, Geb. 26.12.01, D-40225 DuÈsseldorf,

Germany.

Summary

In Saccharomyces cerevisiae, glucose activation of

cAMP synthesis requires both the presence of the G-

protein-coupled receptor (GPCR) system, Gpr1-Gpa2,

and uptake and phosphorylation of the sugar. In a

hxt-null strain that lacks all physiologically important

glucose carriers, glucose transport as well as glu-

cose-induced cAMP signalling can be restored by

constitutive expression of the galactose permease.

Hence, the glucose transporters do not seem to have

a regulatory function but are only required for

glucose uptake. We established a system in which

the GPCR-dependent glucose-sensing process is

separated from the glucose phosphorylation process.

It is based on the specific transport and hydrolysis

of maltose providing intracellular glucose in the

absence of glucose transport. Preaddition of a low

concentration (0.7 mM) of maltose to derepressed

hxt-null cells and subsequent addition of glucose

restored the glucose-induced cAMP signalling,

although there was no glucose uptake. Addition of a

low concentration of maltose itself does not increase

the cAMP level but enhances Glu6P and apparently

fulfils the intracellular glucose phosphorylation

requirement for activation of the cAMP pathway

by extracellular glucose. This system enabled us to

analyse the affinity and specificity of the GPCR

system for fermentable sugars. Gpr1 displayed a

very low affinity for glucose (apparent Ka � 75 mM)

and responded specifically to extracellular a and b

D-glucose and sucrose, but not to fructose, mannose

or any glucose analogues tested. The presence of the

constitutively active Gpa2val132 allele in a wild-type

strain bypassed the requirement for Gpr1 and

increased the low cAMP signal induced by fructose

and by low glucose up to the same intensity as the

high glucose signal. Therefore, the low cAMP

increases observed with fructose and low glucose

in wild-type cells result only from the low sensitivity

of the Gpr1-Gpa2 system and not from the intracel-

lular sugar kinase-dependent process. In conclusion,

we have shown that the two essential requirements

for glucose-induced activation of cAMP synthesis

can be fulfilled separately: an extracellular glucose

detection process dependent on Gpr1 and an intra-

cellular sugar-sensing process requiring the hexose

kinases.

Introduction

In the yeast Saccharomyces cerevisiae, the hexoses

glucose, fructose and mannose not only serve as a

carbon and energy source, they also have a strong

regulatory effect on a range of physiological properties.

Addition of these rapidly fermented sugars to yeast cells

grown on other carbon sources triggers a wide variety of

processes directed towards the sole and optimal utiliza-

tion of these sugars and affects other seemingly unrelated

properties such as stress resistance, the level of storage

carbohydrates, sensitivity of the cell wall to lyticase, etc.

The glucose-induced transition is a multistep process

involving different signalling pathways. Although these

pathways have been studied for many years, resulting in

the characterization of many intermediate components,

only recently significant progress has been made in

elucidating the upstream sugar sensing mechanisms that

control activation of these pathways (Thevelein, 1994;

Ronne, 1995; OÈ zcan et al., 1996; 1998; de Winde et al.,

1996; 1997; Colombo et al., 1998).

Q 2000 Blackwell Science Ltd

Accepted 26 July, 2000. *For correspondence. E-mail joris.winderickx@bio.kuleuven.ac.be; Tel. (132) 16 321516 or 321500; Fax (132) 16321979.

One important signalling route that has been studied

in great detail is the Ras±adenylate cyclase pathway

(Broach and Deschenes, 1990; Thevelein, 1991; 1992;

Tatchell, 1993; Thevelein and de Winde, 1999). This

pathway regulates the production of cAMP, which is

synthesized by adenylate cyclase (Matsumoto et al.,

1984; Kataoka et al., 1985). Its activity is dependent on

the Ras proteins, which have been thought for many

years to replace the heterotrimeric Gs proteins that

control adenylate cyclase activity in mammalian cells

(Toda et al., 1985). In turn, cAMP causes activation of

cAMP-dependent protein kinase (cAPK) through binding

to the regulatory subunits resulting in dissociation and

concomitant activation of the catalytic subunits (Toda

et al., 1987a,b). cAPK is involved in the post-translational

regulation of a variety of proteins, for instance key

enzymes of gluconeogenesis and glycolysis. The gluco-

neogenic enzyme, fructose-1,6-bisphosphatase, is rapidly

inactivated after glucose-induced activation of cAPK,

whereas phosphofructokinase 2, an important regulator

of the major glycolytic enzyme phosphofructokinase 1, is

rapidly activated. Also, neutral trehalase and glycogen

phosphorylase are activated upon glucose addition,

resulting in mobilization of trehalose and glycogen. In

addition, cAPK exerts control at the transcriptional level,

for instance by repressing STRE-controlled genes and

inducing ribosomal protein genes (Thevelein, 1992; 1994;

Neuman-Silberberg et al., 1995; Ruis and Schuller, 1995).

Hence, the transient cAMP signal has been proposed to

trigger a fast resetting from gluconeogenic to fermentative

metabolism (Thevelein, 1991; Boy-Marcotte et al., 1996;

Jiang et al., 1998).

Only two stimuli are known to activate the Ras-

adenylate cyclase pathway in vivo. Addition of a rapidly

fermented sugar to cells growing on a non-fermentable

carbon source or stationary-phase cells (glucose-

derepressed cells) triggers a rapid, transient increase

in the cAMP level (van der Plaat, 1974; Purwin et al.,

1982; Tortora et al., 1982; Thevelein, 1984; Thevelein

et al., 1987a). Intracellular acidification, triggered for

instance by addition of the protonophore 2,4-dinitrophenol

(DNP) at low extracellular pH, triggers a more lasting

and higher cAMP increase (Caspani et al., 1985; Purwin

et al., 1986; Thevelein et al., 1987b). We have provided

evidence that fermentable sugars and intracellular acid-

ification affect adenylate cyclase activity through different

G proteins (Colombo et al., 1998). Intracellular acidifi-

cation, but not glucose, increases the GTP content on

the Ras proteins possibly through inhibition of the Ras-

GTPase activating proteins Ira1 and Ira2. The hetero-

trimeric Ga protein, Gpa2, is required for glucose-induced,

but not for intracellular acidification-induced activation of

cAMP synthesis. Recently, it has been shown that a

membrane-bound protein, Gpr1, interacts with Gpa2

(Yun et al., 1997; Xue et al., 1998; Kraakman et al., 1999).

Gpr1 displays structural and functional homology to

mammalian G-protein-coupled receptors indicating that it

may function as a receptor. Initially, it was suggested that

the GPR1-GPA2 encoded GPCR system would be

involved in nitrogen sensing (Xue et al., 1998), but

recently it was shown that this system is required for

activation of cAMP synthesis by glucose (Kraakman et al.,

1999).

In addition to Gpr1 and Gpa2, glucose activation of

cAMP synthesis also requires uptake and phosphorylation

of the sugar. The affinity of the glucose-sensing system is

relatively low (apparent Ka � 20 mM), in spite of the low

Km of glucose transport in derepressed cells (1±2 mM)

(Beullens et al., 1988). This might suggest that certain

low-affinity transporters play a regulatory role. Glucose

uptake in the yeast S. cerevisiae is generally considered

to occur via facilitated diffusion (Bisson et al., 1993). It is

catalysed by a series of glucose carriers encoded by the

HXT genes. Although more than 20 HXT genes have

been identified (AndreÂ, 1995; Kruckeberg, 1996; Boles

and Hollenberg, 1997), a strain deleted only for HXT1

to HXT7 does not show detectable glucose transport.

Therefore, this so-called hxt-null strain is unable to grow

on glucose (Reifenberger and Ciriacy, 1994; Reifenberger

et al., 1995). Expression of only one of the seven

transporters deleted in the hxt-null strain or derepression

of the galactose transporter Gal2 restores growth on

glucose of the hxt-null strain (Reifenberger et al., 1997).

All carriers also transport fructose albeit with a lower

affinity compared with glucose. Up to now, no evidence

has been found that one of the active glucose carriers

would have a regulatory role in the activation of glucose-

induced pathways such as for instance the main glucose

repression pathway (Reifenberger et al., 1997). On the

other hand, two of the HXT homologues for which no

active glucose transport could be demonstrated, Snf3 and

Rgt2, have been proposed to play a role, respectively, as

a low and high glucose sensor for glucose-induced

expression of some of the HXT genes (OÈ zcan et al.,

1996; 1998). Glucose phosphorylation can be carried out

by any one of the three known glucose kinases,

hexokinase PI, PII or glucokinase, and further metabolism

of the sugar beyond glucose phosphorylation is not

required for glucose-induced activation of cAMP synthesis

(Beullens et al., 1988; Pernambuco et al., 1996).

The connection between the requirements for uptake

and phosphorylation and the functioning of the G-protein

dependent receptor system is unclear. In addition, it is

also not known whether glucose uptake is only required

to sustain glucose phosphorylation, whether specific

glucose carriers play a role in the activation of cAMP

synthesis process and whether the apparent Ka for

activation by glucose is related to the affinity of the

Dual glucose-sensing system for cAMP induction 349

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 348±358

carriers expressed in derepressed cells. Therefore, we

have studied in more detail the connection between the

requirement of sugar uptake and phosphorylation and the

requirement of Gpa2 and Gpr1 for glucose- and fructose-

induced activation of the Ras±cAMP pathway. We show

that constitutive expression of the galactose permease in

an hxt-null strain restores glucose- and fructose-induced

cAMP synthesis making it unlikely that additional regula-

tory functions besides delivering the sugar are associated

with the hexose transporters. In addition, we developed a

system in which intracellular glucose phosphorylation

could occur in the absence of extracellular glucose by

provision of low amounts of maltose to derepressed hxt-

null cells. This system allowed us to demonstrate that the

glucose phosphorylation requirement for cAMP signalling

can be separated from the requirement for activation of

the GPCR system by glucose. We also demonstrate that

the GPCR system specifically responds to D-glucose and

sucrose but not to fructose or any other monosaccharide

tested and that it is responsible for the apparent low

affinity of glucose activation of cAMP synthesis. The much

weaker activation of the cAMP pathway by other sugars

like fructose seems only to be triggered by the intracellular

sugar phosphorylation dependent process.

Results

cAMP signalling is dependent on hexose transport but not

on a specific hexose transporter

To investigate the involvement of glucose transport and

phosphorylation in sugar-induced cAMP signalling, we

have monitored changes in cAMP and sugar phosphates

after addition of either glucose or fructose to derepressed

cells of a wild-type strain and two hxt-null strains. As

shown in Fig. 1A, addition of 100 mM glucose to wild-type

cells resulted in a sharp but transient increase in cAMP.

Addition of 100 mM fructose also resulted in an increase

in cAMP but the amplitude of the signal was only half

of that observed with glucose. The difference between

the glucose- and fructose-induced cAMP signal was not

specific for this genetic background but was consistently

observed in all wild-type strains tested (results not

shown). In spite of this difference in the amplitude of the

cAMP signal, both glucose and fructose addition resulted

in a comparable increase in Glu6P (Fig. 1B), Fru6P and

Fru1,6bP (results not shown). In both hxt-null strains,

neither addition of 100 mM glucose nor addition of

100 mM fructose triggered significant changes in cAMP

as shown in Fig. 1A. With glucose addition, a very slow

accumulation of Glu6P (Fig. 1B), Fru6P and Fru1,6bP

(results not shown) was observed indicating that the hxt-

null strains are still capable of taking up glucose albeit at a

rate which is too low to support growth or to trigger

activation of the Ras±cAMP pathway. Upon addition of

100 mM fructose to the same strains, the sugar phos-

phate concentrations did not change significantly (Fig. 1B

and results not shown).

We then tested whether constitutive expression of the

galactose permease, encoded by GAL2, could restore

glucose- and fructose-induced cAMP synthesis in the hxt-

null strains. It has been reported previously that GaI2 is

able to transport glucose (Liang and Gaber, 1996) and

fructose with high-affinity kinetics (Reifenberger et al.,

1997). Because GaI2 requires galactose induction and is

subject to glucose repression (Johnston and Carlson,

1992; Johnston et al., 1994), the permease is not involved

in glucose uptake during exponential growth on glucose in

wild-type cells. Also under our conditions it was appar-

ently largely absent as shown by the very slow increase in

the Glu6P level upon glucose addition to derepressed

hxt1-7D cells (Fig. 1B). We transformed the hxt-null

strains with a plasmid containing an ADH-GAL2 construct.

As opposed to the hxt-null strains, the transformants were

able to grow on glucose- or fructose-containing medium

(results not shown) and a clear accumulation of Glu6P

was observed upon addition of 100 mM glucose or

fructose to derepressed cells of the transformants.

Glucose addition also induced a marked cAMP increase

while fructose addition resulted only in a moderate cAMP

increase (Fig. 1A). These results demonstrate that the

galactose transporter can sustain the glucose-induced

cAMP signal. Hence, besides the transport of the sugar

no additional, regulatory function for cAMP induction

seems to be associated with the canonical hexose

transporters. The role of the hexose transporters appears

to be confined to maintaining a critical level of intracellular

sugar for the phosphorylation requirement.

The fructose- and low glucose-induced cAMP synthesis

reflects the sugar phosphorylation requirement and does

not require Gpr1 or Gpa2

As described above, addition of 100 mM fructose to wild-

type cells triggers a cAMP signal the amplitude of which is

about half of the cAMP signal obtained with 100 mM

glucose. In this way, fructose addition results in a cAMP

signal comparable to that obtained with 5 mM glucose as

shown in Fig. 2. We reported previously (Colombo et al.,

1998; Kraakman et al., 1999) that the cAMP increase after

addition of 5 mM glucose is not affected by deletion of

GPR1 or GPA2. They encode, respectively, the receptor-

like protein and the G-protein of the GPCR system that is

required for full activation of cAMP synthesis with high

(100 mM) glucose concentrations. This has now been

confirmed (Fig. 2A and C). In addition, Fig. 2B shows that

the cAMP increases obtained by addition of 100 mM

fructose to a wild-type strain, a gpr1D deletion strain or a

350 F. Rolland et al.

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 348±358

gpa2D deletion strain are comparable indicating also that

fructose sensing is independent of the GPCR system.

Deletion of the genes encoding the hexose kinases, i.e.

HXK1, HXK2 and GLK1, however, entirely eliminated the

cAMP signal induced by high fructose or low glucose

concentrations. Moreover, preaddition of 5 mM glucose

followed by addition of either 100 mM fructose or 100 mM

glucose eliminated the fructose-induced but not the

high glucose-induced cAMP signal (Fig. 2C). These data

indicate that activation by 5 mM glucose and 100 mM

fructose is proceeded by the same sugar phosphorylation

dependent mechanism.

The requirement for glucose activation of the Gpr1-Gpa2

GPCR system can be separated from the glucose

phosphorylation requirement

The glucose phosphorylation requirement complicates

investigation of the affinity and specificity of the Gpr1-

Gpa2 GPCR system. Therefore, we established a method

where the glucose-induced activation of cAMP synthesis

via the GPCR system can be studied independently of the

requirement for glucose phosphorylation. We reasoned

that the intracellular glucose required for glucose phos-

phorylation could also be provided by addition of maltose.

Maltose is transported by a specific maltose-induced

transport system and intracellularly converted by maltase

into glucose (Vanoni et al., 1989; Needleman, 1991). We

tested the effect of addition of low concentrations of

maltose to the wild-type strain MC996A and the hxt-null

strain. As shown in Fig. 3A, addition of 2.5 mM maltose

did not increase the Glu6P concentration in the wild-type

strain whereas a pronounced accumulation was observed

in the transporter deletion strain. This difference can

be explained on the basis that glucose transport is a

facilitated diffusion process and faster than maltose

uptake. Therefore, the intracellular glucose derived from

maltose might be exported rapidly to the medium in the

wild-type strain but not in the hxt-null strain so that in the

former the concentration of intracellular glucose available

for glucose phosphorylation would remain low. In the wild

type, addition of 2.5 mM maltose did not induce an

increase in cAMP. In the hxt-null strain, we could observe

a small increase in the level of cAMP upon addition of

2.5 mM maltose to derepressed cells which is in agree-

ment with the above described results that sugar

phosphorylation upon addition of fructose or low glucose

concentrations can give a partial cAMP signal (Fig. 3B).

Subsequently, we added 100 mM glucose 2 min after

the provision of maltose. As expected, this resulted in an

Fig. 1. Glucose- and fructose-induced accumulation of Glu6P and cAMP in vivo in wild type, two hxt-null strains and the hxt1-7D gal2DpADHGAL2 strain. The cAMP signal (A) and the Glu6P level (B) were measured as a function of time after addition of 100 mM glucose (X) or100 mM fructose (W) to derepressed cells of the wild-type strain MC996A, the hxt1-7D and hxt1-7D gal2D strains and the hxt1-7D gal2Dstrain constitutively expressing the Gal2 permease as indicated.

Dual glucose-sensing system for cAMP induction 351

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 348±358

increase in the Glu6P concentration and a concomitant

spike in cAMP in the wild-type strain. In the hxt-null strain,

glucose addition did not give a supplementary increase

in Glu6P but most interestingly, however, a significant

increase in cAMP was observed. We repeated the

experiment with the hxt-null strain but with preaddition of

0.7 mM maltose and added 100 mM glucose either 2, 4 or

6 min after the addition of maltose. No significant change

in the cAMP level was obtained with this very low maltose

concentration but each time glucose was added a

pronounced cAMP increase was observed as shown in

Fig. 4A. The intensity of the cAMP signal decreased

somewhat when the time interval between maltose and

glucose addition was longer most probably because of

gradual depletion of the low maltose level in the medium.

The hxt-null strain does not transport glucose at a

significant rate and the addition of glucose also did not

enhance the intracellular glucose or Glu6P level above

that observed with maltose alone. Therefore, the immedi-

ate and large increase in the cAMP level observed upon

addition of 100 mM glucose under these conditions could

only be caused by extracellular glucose detection via

the GPCR system. Indeed, no cAMP increases were

observed upon addition of 100 mM glucose after preaddi-

tion of maltose when GPR1 or GPA2 was deleted in the

hxt1-7D gal2D strain as shown in Fig. 3B and Fig. 4B.

We have shown previously that the constitutively active

Gpa2val132 allele is able to bypass the absence of Gpr1 for

glucose-induced activation of cAMP synthesis. To confirm

that the cAMP increase upon addition of 100 mM glucose

after maltose pretreatment is not due to residual glucose

uptake in the hxt1-7D gal2D strain, we tested whether

addition of 100 mM glucose to a hxt1-7D gal2D strain that

lacks the Gpr1 receptor but expresses Gpa2val132 allele

could trigger a cAMP increase. However, also with this

strain, no glucose-induced increase in cAMP could be

observed (data not shown). This confirms that the

extracellular glucose is not able to enter the cell at least

not at a rate able to sustain induction of the cAMP

signal.

Affinity and specificity of the Gpr1-dependent extracellular

glucose-sensing process

Using the approach described above, we determined the

specificity of the GPCR system. Addition of 100 mM of the

a- or b-anomer of D-glucose to cells of the hxt1-7D gal2D

strain after preaddition of 0.7 mM maltose triggered in

both cases an increase in the cAMP level. The b-anomer

of D-glucose was somewhat more potent than the a-

anomer (Fig. 5A). Addition of 100 mM sucrose also

triggered a similar Gpr1-Gpa2 dependent cAMP increase.

This was not due to extracellular cleavage of sucrose to

glucose and fructose by periplasmic invertase as inver-

tase mutants displayed the same sucrose-induced cAMP

signal (data not shown). However, addition of 100 mM

fructose (Fig. 5B) or 100 mM each of mannose, galac-

tose, xylose, L-glucose or the glucose analogues, 3-O-

methyl glucose, 6-deoxyglucose or 2-deoxyglucose

(results not shown) did not cause any significant change

in the cAMP level. Even with 300 mM fructose no cAMP

Fig. 2. High-fructose and low-/high-glucose-induced cAMPsignalling in Gpr1, Gpa2 and glucose kinase mutants. cAMP signalwith 100 mM glucose (A) or 100 mM fructose (B) in the wild-typestrain (X), the gpr1D (O), the gpa2D (K), the hxk1D hxk2D glk1Dgpr1D (B) and the hxk1D hxk2D glk1D gpa2D (A) deletion strainsand (C) cAMP signal with 5 mM glucose followed by 100 mMglucose (X) or fructose (W) in the wild-type strain and with 5 mMglucose followed by 100 mM glucose in the gpr1D (O), and thegpa2D (K) deletion strains

352 F. Rolland et al.

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 348±358

response was observed (Fig. 5B). Similar results were

obtained with derepressed wild-type cells when these

sugars and sugar analogues were added 2 min after

addition of 5 mM glucose. Addition of 100 mM maltose

with or without preaddition of 5 mM glucose did not trigger

a cAMP signal indicating that maltose cannot activate

the Gpr1-Gpa2 dependent system (results not shown).

Hence, the Gpr1-dependent extracellular-sensing system

appears to be specific for extracellular D-glucose, at least

among the monosaccharides tested, and for sucrose.

With the same experimental set-up, we also determined

the affinity of the Gpr1-dependent glucose detection system

using different glucose concentrations. As shown in Fig. 6A,

the maximum cAMP increase was observed with 400 mM

glucose. When the amplitude of the cAMP signal was

plotted vs. the concentration of glucose used, an apparent

Ka of about 75 mM was found as illustrated in Fig. 6B.

Gpa2val132 substitutes for the requirement of Gpr1-

dependent extracellular glucose sensing

To gain further support for the dual sensing system,

we compared the glucose- and fructose-induced cAMP

response of a wild-type strain with that of an isogenic

strain lacking the Gpr1 protein and expressing the

Gpa2val132 allele (Kraakman et al., 1999). In the latter,

activation of cAMP synthesis should be independent of

extracellular Gpr1-dependent sugar detection and should

only require sugar phosphorylation. As shown in Fig. 7A

and B, addition of different concentrations of glucose or

100 mM fructose all triggered an equally high cAMP

response in the gpr1D GPA2val132 strain. These results

confirm that Gpa2val132 indeed bypasses the need for

the high extracellular glucose concentrations normally

required for activation of the Gpr1-Gpa2 system and that

only the phosphorylation requirement has to be fulfilled for

cAMP signalling in a strain with a constitutively active

GPCR system. The data further support the idea that

fructose is not able to activate the Gpr1-Gpa2 system and

that the fructose-induced cAMP increase is largely, if not

exclusively, due to stimulation of cAMP synthesis by

intracellular glucose phosphorylation.

Fig. 4. Maltose-mediated fulfilment of theglucose phosphorylation requirement foractivation of cAMP synthesis by extracellularglucose.A. cAMP level after addition of 0.7 mMmaltose (X) alone or after addition of 100 mMglucose, 2 min (W), 4 min (O) and 6 min (K)after addition of 0.7 mM maltose toderepressed hxt1-7D gal2D cells.B. cAMP level after addition of 100 mMglucose, 4 min after addition of 0.7 mMmaltose to derepressed hxt1-7D gal2D (X),hxt1-7D gal2D gpr1D (W), hxt1-7D gal2Dgpa2D (O).

Fig. 3. Glu6P and cAMP level in the wild type, the hxt1-7D gal2Dand the hxt1-7D gal2D gpa2D strain after addition of maltose.Glu6P (A) and cAMP (B) levels in the wild type (X), the hxt1-7Dgal2D (W), and the and hxt1-7D gal2D gpa2D (O) after addition of2.5 mM maltose followed 2 min later by addition of 100 mM glucose.

Dual glucose-sensing system for cAMP induction 353

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 348±358

Discussion

The hexose carriers do not seem to play a specific

regulatory role in glucose activation of the cAMP pathway

Glucose activation of cAMP synthesis is known to require

glucose phosphorylation (Beullens et al., 1988) and the G-

protein-coupled receptor system consisting of Gpr1 and

Gpa2 (Kraakman et al., 1999). Our present result that the

glucose-induced cAMP signal is entirely absent in a yeast

strain without functional glucose carriers confirms that the

presence of the GPCR system is not enough to trigger

activation of the cAMP pathway by glucose. Transport

and phosphorylation of glucose are apparently required

for transduction of the glucose signal to adenylate

cyclase. Up to now, it had not been investigated whether

the glucose phosphorylation requirement reflected only

the need for glucose phosphorylation or whether one or

more specific glucose carriers in some way also played a

role. We have now demonstrated that the presence of

glucose-induced cAMP signalling does not require a

specific glucose carrier because the cAMP increase can

also be triggered in a hxt-null strain when the glucose is

transported solely via the galactose transporter or when

extracellular glucose is provided together with low

amounts of maltose. These data seem to indicate that it

is not important by which means the intracellular glucose

required for glucose phosphorylation is provided and that

therefore the transporters themselves do not seem to play

a regulatory role in glucose activation of cAMP synthesis.

Experiments with yeast strains expressing a single

functional glucose carrier have confirmed this conclusion

(F. Rolland et al.. unpublished results). In addition, the

glucose carrier homologues, Snf3 and Rgt2, previously

suggested to function as glucose sensors, are also not

required for glucose activation of cAMP synthesis (F.

Rolland, J.M. Thevelein and J. Winderickx, unpublished

results). The conclusion that the role of the glucose carriers is

confined to glucose uptake was also obtained for activation of

the main glucose repression pathway, which also requires

glucose phosphorylation (Reifenberger et al., 1997).

The exact nature of the signal given by the hexose

kinase dependent glucose phosphorylation process

remains to be elucidated. Data obtained in wild-type

strains and glucose carrier deletion strains clearly

indicated that there is no correlation between the Glu6P

increase and the amplitude of the glucose-induced cAMP

signal (Fig. 3, Beullens et al., 1988; F. Rolland et al.,

unpublished results). This can be interpreted as indicating

that Glu6P is not involved in the activation process or

alternatively that another factor is limiting the activation of

adenylate cyclase. Furthermore, results obtained with

glycolysis mutants and from experiments in which

Fig. 5. Specificity of the Gpr1-Gpa2dependent glucose detection process. cAMPafter addition of 100 mM a-D-glucose (X) orb-D-glucose (W) to derepressed hxt1-7Dgal2D cells 4 min after the addition of 0.7 mMmaltose (A) cAMP after addition of 100 mMglucose (X),100 mM fructose (W) and 300 mMfructose (O) to derepressed hxt1-7D gal2Dcells, 4 min after the addition of 0.7 mMmaltose (B).

Fig. 6. Affinity of the Gpr1-Gpa2 dependentglucose detection process. cAMP levelsobtained after addition of different glucoseconcentrations 4 min after addition of 0.7 mMmaltose to derepressed hxt1-7D gal2D cells.Glucose concentrations: 2 mM (X), 10 mM(W), 20 mM (O), 50 mM (K), 100 mM (B),300 mM (A), 400 mM (V), 500 mM (S),750 mM (P), 1000 mM (L) (A). Half-maximalcAMP accumulation is obtained with aglucose concentration of about 75 mM asindicated (B).

354 F. Rolland et al.

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 348±358

iodoacetate was used to inhibit ATP generation in

glycolysis, showed that the glucose-induced increase in

cAMP is not dependent on an increase in the ATP level.

The cAMP signal remains independent of the ATP level

provided that enough ATP is available as substrate for

adenylate cyclase (Beullens et al., 1988; F. Rolland et al.,

unpublished results). Hence, neither Glu6P nor ATP seem

to act as simple metabolic messengers for glucose

activation of cAMP synthesis

Extracellular glucose detection by the Gpr1-Gpa2

dependent mechanism can be separated from glucose

phosphorylation in the hxt-null strain

Because glucose activation of cAMP synthesis requires

glucose phosphorylation in addition to the presence of

the Gpr1-Gpa2 G-protein coupled receptor system, the

properties of the latter system cannot be studied in a

straightforward manner. For instance, it is difficult to

determine whether the receptor responds to extracellular

or intracellular glucose. One can also not test activation of

the GPCR system by glucose analogues because they

have to be able to sustain also the glucose phosphoryla-

tion requirement. For instance, it has been unclear

whether L-glucose can activate the GPCR system

because L-glucose cannot be transported and phosphory-

lated. 2-Deoxyglucose presents another problem: it is

efficiently transported and phosphorylated but not further

metabolized and therefore causes a precipitous drop in

the ATP level. Because cAMP synthesis requires ATP,

absence of a cAMP signal with 2-deoxyglucose might be

due to an ATP depletion effect. We have now established

a system in which the requirement for activation of the

Gpr1-Gpa2 detection mechanism and the requirement

for glucose phosphorylation can be fulfilled separately

(Figs 3 and 4). It makes use of the hxt1-7D gal2D strain

which is unable to take up glucose from the medium. The

strain is given maltose in a concentration too low to trigger

by itself an increase in the cAMP level but apparently high

enough to fulfil the glucose phosphorylation requirement,

because subsequent addition of glucose caused an

immediate spike in the cAMP level. This novel system

has led to several new conclusions. First of all, this

system uncouples for the first time the requirement

for extracellular glucose from the requirement for

intracellular glucose phosphorylation. Apparently, the

glucose detected by the Gpr1-Gpa2 system does not

have to be phosphorylated itself.

Secondly, this system clearly shows that Gpr1 is

required for the detection of extracellular glucose.

Although our data do not provide evidence for a direct

interaction between Gpr1 and glucose, such an interac-

tion offers by far the simplest explanation for the results

described in this paper. In the control experiment with the

hxt-null strain expressing the Gpa2val132 allele in the

absence of Gpr1, glucose addition was unable to trigger

any increase in cAMP. This indicates that any residual

glucose transport or any change in intracellular meta-

bolites or composition that might be caused by the

extracellular glucose in the hxt-null strain is unable to

affect the cAMP level. Whatever such residual changes

are, they are clearly not able to sustain enough glucose

phosphorylation to trigger the cAMP signal.

Thirdly, this system allows for the first time to study the

specificity of the GPCR system for the extracellular sugar

independent of the metabolism of this sugar. In this way

we have shown that Gpr1-dependent extracellular sugar

sensing is specific for D-glucose with a preference for

the b-anomer. None of the other hexoses or glucose

analogues tested was able to activate cAMP synthesis

under the same conditions as glucose. This explains why

deletion of Gpr1 or Gpa2 has no effect on the cAMP

increase observed after addition of 100 mM fructose

(Fig. 3). This increase is apparently only triggered by the

intracellular phosphorylation dependent process. When

the increase in cAMP was measured with different

extracellular glucose concentrations, an apparent Ka

of about 75 mM for the extracellular glucose detection

system was deduced. This apparent Ka is somewhat

higher than that previously obtained for glucose-induced

activation of cAMP synthesis in wild-type cells, which was

Fig. 7. Activation of cAMP synthesis in a gpr1D strain expressingthe constitutively active Gpa2val132 allele. cAMP after addition of5 mM (X), 25 mM (W), 50 mM (O) or 100 mM (K) glucose toderepressed W303-1A wild type and gpr1D Gpa2val132 cells (A).cAMP after addition of 100 mM glucose (X) or 100 mM fructose (W)to derepressed W303-1A wild type and gpr1D Gpa2val132 cells (B).

Dual glucose-sensing system for cAMP induction 355

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 348±358

around 25 mM (Beullens et al., 1988). However, it can be

considered in the same range in view of the different

strain and experimental set-up. Previously, we have

suggested that the low-affinity glucose uptake system

might be involved in glucose-activation of cAMP synthesis

because its Km of 20±25 mM is similar to the apparent Ka

of glucose activation of cAMP synthesis in wild-type cells.

Our present results show that the low-affinity system is

the Gpr1-Gpa2 system (or at least the Gpr1-dependent

glucose-sensing system) and not the low-affinity glucose

uptake system. The natural environment of yeast cells

(e.g. grape juice) can contain very high glucose concen-

trations which are fermented rather than respired. The

very low affinity of the GPCR system for its ligand might

therefore be related to its function in stimulating through

the cAMP pathway the switch from respirative/gluconeo-

genic growth to fermentative growth which only fully

occurs at glucose concentrations of around 20 mM

(Thevelein, 1991; Johnston and Carlson, 1992).

Gpa2val132 can fully substitute for the requirement of high

extracellular glucose

We have previously reported that for glucose-induced

cAMP signalling Gpr1 deletion can be bypassed by the

constitutively active Gpa2val132 allele (Kraakman et al.,

1999). In this paper, we have confirmed these results and

in addition we have shown that Gpa2val132 can substitute

for the requirement of high extracellular glucose. Indeed,

we demonstrated that in a gpr1D GPA2val132 strain cAMP

responses of similar magnitude are obtained with glucose

and fructose. This confirms that the fructose-induced

cAMP increase is only triggered by the sugar phospho-

rylation dependent process while the high glucose

induced increase is supported in addition by the Gpr1-

Gpa2 dependent process. We have further demonstrated

that in the gpr1D GPA2val132 strain low glucose concen-

trations trigger similar cAMP spikes as the high glucose

concentration. This further supports that the low glucose

sensitivity of activation of cAMP synthesis in wild-type

strains is due to the low sensitivity for glucose of the Gpr1-

Gpa2-dependent process. The glucose kinases have Km

values below 1.5 mM for both glucose and fructose (Lobo

and Maitra, 1977; Maitra and Lobo, 1983), which is much

less than the apparent Ka of the Gpr1-Gpa2 system for

glucose. Hence, in a wild-type strain, the latter system acts

as limiting factor for glucose activation of cAMP synthesis.

Conclusions

We have demonstrated that glucose activation of cAMP

synthesis consists of two processes: an extracellular

glucose-sensing process that is dependent on the G-

protein coupled receptor system consisting of Gpr1 and

Gpa2, and an intracellular glucose sensing process that is

dependent on glucose phosphorylation by the glucose

kinases. These two processes can be separated using a

strain in which all physiologically relevant glucose carriers

are deleted and in which the glucose phosphorylation

requirement is fulfilled through maltose transport and

hydrolysis. This shows that extracellular glucose detection

and glucose phosphorylation are independent processes

and that the extracellularly detected glucose molecules do

not need to be phosphorylated. We have also shown that

the requirement for the Gpr1-Gpa2 dependent extracel-

lular glucose sensing process can be fully substituted by

the constitutively active Gpa2val132 allele. This allows

ligands that are phosphorylated but not detected by the

Gpr1-Gpa2 system, such as fructose and low glucose, to

cause full activation of cAMP accumulation.

Experimental procedures

Strains and plasmids

The wild-type strain (MC996A) and the hxt-null strain hxt1-7D(RE700A) have been described by Reifenberger et al. (1995).The hxt1-7D gal2D (RE800A) strain has been described byReifenberger et al. (1997). We transformed the wild-typestrain (MC996A), the hxt1-7D (RE700A) and the hxt1-7Dgal2D (RE800A) strains with an episomal plasmid (pHL125)expressing GAL2 under the control of the ADH2 promoter, aconstruct kindly provided by Liang and Gaber (1996). Thestrains hxt1-7D gal2D gpa2D (FR801) and hxt1-7D gal2Dgpr1D (FR802) were obtained by deletion of, respectively,GPA2 and GPR1 in strain RE800A using a constructdescribed previously (Colombo et al., 1998; Kraakman et al.,1999). The strain hxt1-7D gal2D gpr1D GPA2val132 wasobtained by transforming strain FR802 with a centromericplasmid bearing the constitutive GPA2val132 allele. The wild-type strain W303-1A, the congenic triple hexose kinasedeletion strain and the gpr1D GPA2val132 have beendescribed before (de Winde et al., 1996; Kraakman et al.,1999). GPR1 and GPA2 were deleted in the triple hexosekinase deletion strain by replacing the coding region with theURA3 gene

Growth media

Yeast cells were grown at 308C in 1% yeast extract, 2%Bacto-peptone (YP) supplemented with either 2% maltose(YPM), 2% glucose (YPD), 2% fructose (YPF), 2% galactose(YPGal) or 3% glycerol (YPG) as indicated. To avoid thepossibility of generating suppressors, we tested the hxt1-7D ,the hxt1-7D gal2D and their derived deletion strains forgrowth on YPD medium before and after each experiment.For the strains carrying the pADHGAL2 plasmid, cells weregrown on nutrient supplementation media lacking uracil (SC-ura) that contained 0.67% Difco yeast nitrogen basesupplemented with 10 mM (NH4)2SO4 the required aminoacids and either 2% of the appropriate sugar (SCM, SCD,SCF, SCGal) or 3% glycerol (SCG).

356 F. Rolland et al.

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 348±358

Biochemical analyses

For the determination of the glycolytic metabolites and cAMP,cells were grown on YPG or SCG media and harvested bycentrifugation. The pellet was washed once, resuspendedand then incubated in 25 mM MES buffer pH 6 at 308C.Before and after the addition of sugars to this cell suspensionat a concentration of 100 mM unless otherwise mentioned,samples were taken at the time intervals indicated andimmediately quenched in 60% methanol at 2408C. Forglycolytic metabolite determination, cells were extractedusing chloroform at 2408C at neutral pH, essentially asdescribed by de Koning and van Dam (1992). Using the totalamount of protein in the sample as determined according toLowry et al. (1951), and the assumption of a yeast cytosolicvolume of 12 ml mg21 protein, cytosolic concentrations in mMwere calculated. For determination of the cAMP levels, cellswere extracted using perchloric acid. After neutralizationcAMP levels were determined using the cyclic AMP [3H]assay system from Amersham as described previously(Thevelein et al., 1987b). All experiments were performedat least three times and representative results are shown.

Acknowledgements

We are grateful to J. Rosseels and W. Verheyden forexcellent technical assistance. This work was supported byfellowships from the Institute for Scientific and TechnologicalResearch (IWT) to F.R. and the Fund for Scientific Research± Flanders to J.W. and by grants from the Fund for ScientificResearch ± Flanders and the Research Fund of theKatholieke Universiteit Leuven (Concerted Research Actions)to J.W. and J.M.T., the European Commission (Biotechnol-ogy BIO4-CT98-0562 to E.B. and J.M.T.), and InteruniversityAttraction Poles Network P4/30 to J.M.T. and the DeutscheForschungsgemeinschaft (grant BO 1517/1-1) to E.B.

References

AndreÂ, B. (1995) An overview of membrane transportproteins in Saccharomyces cerevisiae. Yeast 11: 1575±1611.

Beullens, M., Mbonyi, K., Geerts, L., Gladines, D., Detrem-erie, K., Jans, A.W.H., and Thevelein, J.M. (1988) Studieson the mechanism of the glucose-induced cAMP signal inglycolysis and glucose repression mutants of the yeastSaccharomyces cerevisiae. Eur J Biochem 172: 227±231.

Bisson, L.F., Coons, D.M., Kruckeberg, A.L., and Lewis, D.A.(1993) Yeast sugar transporters. Crit Rev Biochem Mol Biol28: 259±308.

Boles, E., and Hollenberg, C.P. (1997) The moleculargenetics of hexose transport in yeasts. FEMS MicrobiolRev 21: 85±111.

Boy-Marcotte, E., Tadi, D., Perrot, M., Boucherie, H., andJacquet, M. (1996) High cAMP levels antagonize thereprogramming of gene expression that occurs at thediauxic shift in Saccharomyces cerevisiae. Microbiology142: 459±467.

Broach, J.R., and Deschenes, R.J. (1990) The function ofRAS genes in Saccharomyces cerevisiae. Adv Cancer Res54: 79±139.

Caspani, G., Tortora, P., Hanozet, G.M., and Guerritore, A.(1985) Glucose-stimulated cAMP increase may bemediated by intracellular acidification in Saccharomycescerevisiae. FEBS Mircobiol Lett 186: 75±79.

Colombo, S., Ma, P., Cauwenberg, L., Winderickx, J.,Crauwels, M., Teunissen, A., et al. (1998) Involvement ofdistinct G-proteins, Gpa2 and Ras, in glucose- andintracellular acidification-induced cAMP signalling in theyeast Saccharomyces cerevisiae. EMBO J 17: 3326±3341.

Jiang, Y., Davis, C., and Broach, J.R. (1998) Efficienttransition to growth on fermentable carbon sources inSaccharomyces cerevisiae requires signaling through theRas pathway. EMBO J 17: 6942±6951.

Johnston, M., and Carlson, M. (1992) Regulation of carbonand phosphate utilization. In The Molecular and CellularBiology of the Yeast Saccharomyces, Vol. 2. Jones, E.W.,Pringle, J.R., and Broach, J.R. (eds). Cold Spring Harbor,NY: Cold Spring Harbor Laboratory Press, pp. 193±281.

Johnston, M., Flick, J.S., and Pexton, T. (1994) Multiplemechanisms provide rapid and stringent glucose repres-sion of GAL gene expression in Saccharomyces cerevi-siae. Mol Cell Biol 14: 3834±3841.

Kataoka, T., Broek, D., and Wigler, M. (1985) DNA sequenceand characterization of the S. cerevisiae gene encodingadenylate cyclase. Cell 43: 493±505.

de Koning, W., and van Dam, K. (1992) A method for thedetermination of changes of glycolytic metabolites in yeaston a subsecond time scale using extraction at neutral pH.Anal Biochem 204: 118±123.

Kraakman, L., Lemaire, K., Ma, P., Teunissen, A.W.R.H.,Donaton, M.C.V., Van Dijck, P., et al. (1999) A Sacchar-omyces cerevisiae G-protein coupled receptor, Gpr1, isspecifically required for glucose activation of the cAMPpathway during the transition to growth on glucose. MolMicrobiol 32: 1002±1012.

Kruckeberg, A.L. (1996) The hexose transporter family ofSaccharomyces cerevisiae. Arch Microbiol 166: 283±292.

Liang, H., and Gaber, R.F. (1996) A novel signal transductionpathway in Saccharomyces cerevisiae defined by Snf3-regulated expression of HXT6. Mol Biol Cell 7: 1953±1966.

Lobo, Z., and Maitra, P.K. (1977) Genetics of yeasthexokinase. Genetics 86: 727±744.

Lowry, O.H., Rosebrough, N.J., Farr, A.L., and Randall, R.J.(1951) Protein measurement with the Folin phenol reagent.J Biol Chem 193: 265±275.

Maitra, P.K., and Lobo, Z. (1983) Genetics of yeastglucokinase. Genetics 105: 501±515.

Matsumoto, K., Uno, I., and Ishikawa, T. (1984) Identificationof the structural gene and nonsense alleles for adenylatecyclase in Saccharomyces cerevisiae. J Bacteriol 157:277±282.

Needleman, R. (1991) Control of maltase synthesis. MolMicrobiol 5: 2079±2084.

Neuman-Silberberg, F.S., Bhattacharya, S., and Broach, J.R.(1995) Nutrient availability and the RAS/cyclic AMP path-way both induce expression of ribosomal protein genes inSaccharomyces cerevisiae but by different mechanisms.Mol Cell Biol 15: 3187±3196.

OÈ zcan, S., Dover, J., Rosenwald, A.G., WoÈ lfl, S., andJohnston, M. (1996) Two glucose transporters in S.cerevisiae are glucose sensors that generate a signal for

Dual glucose-sensing system for cAMP induction 357

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 348±358

induction of gene expression. Proc Natl Acad Sci USA 93:12428±12432.

OÈ zcan, S., Dover, J., and Johnston, M. (1998) Glucosesensing and signaling by two glucose receptors in the yeastSaccharomyces cerevisiae. EMBO J 17: 2566±2573.

Pernambuco, M.B., Winderickx, J., Crauwels, M., Griffioen,G., Mager, W.H., and Thevelein, J.M. (1996) Glucose-triggered signalling in Saccharomyces cerevisiae: differentrequirements for sugar phosphorylation between cellsgrown on glucose and those grown an non-fermentablecarbon sources. Microbiology 142: 1775±1782.

van der Plaat, J.B. (1974) Cyclic 3 0,5 0-adenosine monopho-sphate stimulates trehalose degradation in baker's yeast.Biochem Biophys Res Commun 56: 580±587.

Purwin, C., Leidig, F., and Holzer, H. (1982) Cyclic AMP-dependent phosphorylation of fructose 1,6-bisphosphatasein yeast. Biochem Biophys Res Commun 107: 1482±1489.

Purwin, C., Nicolay, K., Scheffers, W.A., and Holzer, H.(1986) Mechanism of control of adenylate cyclase activityin yeast by fermentable sugars and carbonyl cyanide m-chlorophenylhydrazone. J Biol Chem 261: 8744±8749.

Reifenberger, E., and Ciriacy, M. (1994) The roles of multiplehexose transporters in S. cerevisiae. Folia MicrobiolPrague 39: 541±542.

Reifenberger, E., Freidel, K., and Ciriacy, M. (1995)Identification of novel HXT genes in Saccharomycescerevisiae reveals the impact of individual hexose trans-porters on glycolytic flux. Mol Microbiol 16: 157±167.

Reifenberger, E., Boles, E., and Ciriacy, M. (1997) Kineticcharacterization of individual hexose transporters of Sac-charomyces cerevisiae and their relation to the triggeringmechanisms of glucose repression. Eur J Biochem 245:324±333.

Ronne, H. (1995) Glucose repression in fungi. Trends Genet11: 12±17.

Ruis, H., and Schuller, C. (1995) Stress signaling in yeast.Bioessays 17: 959±965.

Tatchell, K. (1993) RAS genes in the budding yeast Saccharo-myces cerevisiae. In Signal Transduction Prokaryotic andSimple Eukaryotic Systems. Kurjan, J., and Taylor, B.J.(eds). San Diego: Academic Press, pp. 147±188.

Thevelein, J.M. (1984) Cyclic-AMP content and trehalaseactivation in vegetative cells and ascospores of yeast. ArchMicrobiol 138: 64±67.

Thevelein, J.M. (1991) Fermentable sugars and intracellularacidification as specific activators of the Ras-adenylatecyclase signalling pathway in yeast: the relationship tonutrient-induced cell cycle control. Mol Microbiol 5: 1301±1307.

Thevelein, J.M. (1992) The RAS-adenylate cyclase pathwayand cell cycle control in Saccharomyces cerevisiae.Antonie Leeuwenhoek J Microbiol 62: 109±130.

Thevelein, J.M. (1994) Signal transduction in yeast. Yeast 10:1753±1790.

Thevelein, J.M., and de Winde, J.H. (1999) Novel sensingmechanisms and targets for the cAMP±protein kinase Apathway in the yeast Saccharomyces cerevisiae. MolMicrobiol 32: 1002±1012.

Thevelein, J.M., Beullens, M., Honshoven, F., Hoebeeck, G.,Detremerie, K., Griewel, B., et al. (1987a) Regulation ofthe cAMP level in the yeast Saccharomyces cerevisiae:the glucose-induced cAMP signal is not mediated by atransient drop in the intracellular pH. J Gen Microbiol 133:2197±2205.

Thevelein, J.M., Beullens, M., Honshoven, F., Hoebeeck, G.,Detremerie, K., den Hollander, J.A., and Jans, A.W.H.(1987b) Regulation of the cAMP level in the yeastSaccharomyces cerevisiae: intracellular pH and the effectof membrane depolarizing compounds. J Gen Microbiol133: 2191±2196.

Toda, T., Uno, I., Ishikawa, T., Powers, S., Kataoka, T.,Broek, D., et al. (1985) In yeast, Ras proteins arecontrolling elements of adenylate cyclase. Cell 40: 27±36.

Toda, T., Cameron, S., Sass, P., Zoller, M., Scott, J.D.,McBullen, B., et al. (1987a) Cloning and characterizationof BCY1, a locus encoding a regulatory subunit of thecyclic AMP-dependent protein kinase in Saccharomycescerevisiae. Mol Cell Biol 7: 1371±1377.

Toda, T., Cameron, S., Sass, P., Zoller, M., and Wigler, M.(1987b) Three different genes in Saccharomyces cerevi-siae encode the catalytic subunits of the cAMP-dependentprotein kinase. Cell 50: 277±287.

Tortora, P., Burlini, N., Hanozet, G.M., and Guerritore, A.(1982) Effect of caffeine on glucose-induced inactivation ofgluconeogenetic enzymes in Saccharomyces cerevisiae. Apossible role of cyclic AMP. Eur J Biochem 126: 617±622.

Vanoni, M., Sollitti, P., Goldenthal, M., and Marmur, J. (1989)structure and regulation of the multigene family controllingmaltose fermentation in budding yeast. Prog Nucleic AcidRes Mol Biol 37: 281±322.

de Winde, J.H., Crauwels, M., Hohmann, S., Thevelein, J.M.,and Winderickx, J. (1996) Differential requirement of theyeast sugar kinases for sugar sensing in establishing thecatabolite-repressed state. Eur J Biochem 241: 633±643.

de Winde, J.H., Thevelein, J.M., and Winderickx, J. (1997)From feast to famine: adaptation to nutrient depletion inyeast. In Yeast Stress Responses. Hohmann, S., andMager, W.H. (eds). Austin: Landes, pp. 7±52.

Xue, Y., Batlle, M., and Hirsch, J.P. (1998) GPR1 encodes aputative G protein-coupled receptor that associates withthe Gpa2p G subunit and functions in a Ras-independentpathway. EMBO J 17: 1996±2007.

Yun, C.W., Tamaki, H., Nakayama, R., Yamamoto, K., andKumagai, H. (1997) G-protein coupled receptor fromyeast Saccharomyces cerevisiae. Biochem Biophys ResCommun 240: 287±292.

358 F. Rolland et al.

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 38, 348±358