The yeast RAS1 and RAS2 genes appear to be involved in control of cell growth in response to nutrients. Here we show that this growth control also involves a signal mediated by the heterotrimeric G protein α subunit homolog encoded by GPA2. A GPA2 null allele conferred a severe growth defect on cells containing a null allele of RAS2, although either mutation alone had little effect on growth rate. A constitutive allele of GPA2 could stimulate growth of a strain lacking both RAS genes. Constitutive GPA2 conferred heat shock sensitivity on both wild‐type cells and cells lacking RAS function, but had no effect in a strain containing a null allele of SCH9, which encodes a kinase related to protein kinase A. The GPR1 gene was isolated and was found to encode a protein with the characteristics of a G protein‐coupled receptor. Double Δgpr1 Δras2 mutants displayed a severe growth defect that was suppressed by expression of the constitutive allele of GPA2, confirming that GPR1 acts upstream of GPA2. Gpr1p is expressed on the cell surface and requires sequences in the membrane‐proximal region of its third cytoplasmic loop for function, as expected for a G protein‐coupled receptor. GPR1 RNA was induced when cells were starved for nitrogen and amino acids. These results are consistent with a model in which the GPR1/GPA2 pathway activates the Sch9p kinase to generate a response that acts in parallel with that generated by the Ras/cAMP pathway, resulting in the integration of nutrient signals.
Signal transduction pathways that regulate growth in response to nutrients are essential for cell viability; these pathways, however, are not well understood at the molecular level. In yeast, one aspect of nutrient‐induced growth control is thought to be mediated by the RAS/cAMP pathway, which responds to the addition of a fermentable sugar by a transient increase in cAMP concentration (Broach, 1991; Thevelein, 1994). The presence of the sugar generates a signal that has not been defined but appears to impinge on the Ras proteins. Once activated, the Ras proteins stimulate adenylyl cyclase to produce cAMP (Kataoka et al., 1985; Toda et al., 1985). cAMP binds the regulatory subunit of protein kinase A (PKA), which results in the release of active catalytic kinase subunits that then phosphorylate targets involved in energy metabolism and cell growth (Matsumoto et al., 1982; Toda et al., 1987a, b). This pathway is essential, because deletion of both of the genes that encode the Ras proteins, RAS1 and RAS2, or of the single gene that encodes adenylyl cyclase, CYR1, results in inviable cells. The inviability of cells containing deletions of the RAS genes or of CYR1 is suppressed by overexpression of the SCH9 gene, which encodes a protein kinase related to PKA (Toda et al., 1988). However, it is not clear whether the SCH9 gene product functions directly in the RAS/cAMP pathway or whether it functions in a parallel pathway.
Yeast cells express many different classes of signaling molecules in addition to guanine nucleotide‐binding proteins of the Ras family. One such class is composed of heterotrimeric G proteins made up of α, β and γ subunits. The identification of G protein α subunits in yeast was accomplished originally by screening a yeast genomic library with a probe made from cDNA clones encoding mammalian Gαi2 and Gαo (Nakafuku et al., 1987, 1988). Genes encoding two Gα subunits, GPA1 and GPA2, were isolated by this procedure. Sequencing of the entire Saccharomyces cerevisiae genome has now established that GPA1 and GPA2 are the only two Gα subunit genes encoded in this genome, indicating that there are two pathways in yeast that signal through Gα subunits. GPA1 subsequently was shown to be involved in mediating the pheromone response signal transduction pathway (Dietzel and Kurjan, 1987; Miyajima et al., 1987). Whereas Gpa1p is capable of productively coupling to the α‐factor receptor, Gpa2p is not (Blumer and Thorner, 1990).
GPA2 was proposed to function in the regulation of cAMP levels, based on the finding that its overexpression causes a 2‐fold increase in the level of cAMP induced by glucose (Nakafuku et al., 1988). Overexpression of GPA2 also restores the cAMP response to a strain containing a temperature‐sensitive mutation in the RAS2 gene and suppresses the growth defect of this strain. These results support the idea that the Gpa2p protein is involved in regulating the production of cAMP by adenylyl cyclase, in the same way that mammalian Gαs stimulates adenylyl cyclase activity (Simon et al., 1991). However, it was shown that deletion of the GPA2 gene had no effect on cAMP levels (Nakafuku et al., 1988), which made it difficult to establish whether GPA2 is involved in this process. The phenotype of the GPA2 deletion is inconclusive because deletion of genes that are known to be involved in cAMP regulation, such as the phosphodiesterase genes, has little effect on cAMP levels. This phenomenon is due to feedback control on cAMP levels by the Ras proteins over many orders of magnitude (Nikawa et al., 1987). Therefore, changes in cAMP levels are difficult to document in cells with intact RAS genes.
In the study described here, we show that the GPA2 and RAS pathways have partially redundant functions that act in parallel, and that cells with defects in both pathways are impaired for growth. We have also isolated a gene encoding a putative G protein‐coupled receptor that appears to provide the upstream signal that activates Gpa2p, which will ultimately allow one of the ligands that initiates the nutrient signaling pathway to be identified.
Activated Gpa2p confers phenotypes similar to those produced by high cAMP levels
Overexpression of the G protein α subunit encoded by GPA2 increases the level of cAMP in cells and suppresses the phenotype of a ras2ts mutant (Nakafuku et al., 1988). One interpretation of these observations is that the normal function of Gpa2p is to regulate the concentration of cAMP. Activated alleles of RAS2, which are known to increase cAMP levels, confer characteristic phenotypes such as decreases in sporulation efficiency and accumulation of storage carbohydrates (Toda et al., 1985). To test whether activation of Gpa2p has similar effects on cell physiology, a constitutive allele of GPA2 was constructed by replacing the arginine at position 273 with an alanine. The same change in the corresponding residue of mammalian Gαs, which is also the site modified by cholera toxin, results in a protein that constitutively activates adenylyl cyclase and displays a 100‐fold decreased rate of GTP hydrolysis (Freissmuth and Gilman, 1989). Diploid cells transformed with the constitutive GPA2R273A allele or with an activated allele of RAS2 were incubated in liquid sporulation medium for 3 days, and the degree of sporulation was determined. Whereas the percentage sporulation of control cells was 42.1 ± 4.0, that of cells containing the constitutive GPA2R273A allele was 11.1 ± 0.9 (Figure 1A). The percentage sporulation of cells containing the activated RAS2 allele was 18.8 ± 0.4, slightly higher than the value obtained for cells containing constitutive GPA2. These results indicate that activation of Gpa2p causes a significant sporulation defect that is comparable with the defect conferred by activated Ras2p.
Another aspect of cell physiology that is affected by elevated cAMP levels is the ability to survive heat shock (Toda et al., 1987a). To test the effect of constitutive GPA2 on thermotolerance, stationary phase cultures of haploid cells transformed with the constitutive GPA2R273A allele or with the activated allele of RAS2 were exposed to a heat shock, and the percentage of surviving cells was determined. Cells containing constitutive GPA2 were ∼60‐fold more sensitive to heat shock than wild‐type cells (Figure 1B). An even larger effect was seen with activated RAS2, which caused a 10‐fold increase in heat shock sensitivity compared with constitutive GPA2. Thus, in contrast to the sporulation results, the degree of heat shock sensitivity is substantially greater in cells containing activated Ras2p than in cells containing activated Gpa2p.
In summary, the constitutive GPA2 allele caused changes in cell physiology that are consistent with a role for Gpa2p either in cAMP regulation or in a pathway that is redundant with the cAMP/PKA pathway. These experiments do not distinguish between cAMP‐dependent and cAMP‐independent mechanisms because the cellular functions that were measured can be regulated by both mechanisms (Cameron et al., 1988).
GPA2 and RAS2 are functionally related
Previous results have shown that overexpression of GPA2 suppresses the growth phenotype of a ras2ts mutant (Nakafuku et al., 1988), suggesting that there is a functional relationship between GPA2 and the RAS genes. The potential relationship between these genes was explored further by investigating whether null alleles of GPA2 and RAS genes display genetic interactions. To determine the phenotype of Δgpa2 Δras2 double mutants, a diploid strain heterozygous for GPA2 and RAS2 deletion alleles was sporulated and tetrads were dissected. Although strains containing single Δgpa2 or Δras2 mutations grew normally, strains containing both Δgpa2 and Δras2 mutations displayed a severe growth defect (Figure 2A). Colonies of cells containing double Δgpa2 Δras2 mutations were barely visible after 2 days of growth, and were still quite small after 3 days of growth. Therefore, the phenotype of the double mutant uncovers a requirement for a growth function that can be supplied by either GPA2 or RAS2. This function is specific to RAS2, because strains containing double Δgpa2 Δras1 mutations displayed little or no growth defect (data not shown). The synthetic growth defect of Δgpa2 Δras2 strains has also been seen by investigators studying the involvement of GPA2 in pseudohyphal development (Kübler et al., 1997; Lorenz and Heitman, 1997).
In S.cerevisiae, RAS2 plays a role in at least two different processes, nutrient signaling and completion of mitosis. Both of these processes are involved in regulating cell growth, but the mitotic function of RAS does not act through cAMP generation (Morishita et al., 1995). To determine whether the redundant function of GPA2 and RAS2 involves cAMP regulation, a triple mutant was constructed that contained deletion alleles of GPA2, RAS2 and PDE2, which encodes a high affinity phosphodiesterase (Wilson and Tatchell, 1988). Deletion of PDE2 restored normal growth to a Δgpa2 Δras2 strain (Figure 2B), indicating that elevation of the in vivo cAMP concentration compensated for the lack of GPA2 and RAS2. It is therefore possible that the redundant function of these genes involves positive control of cAMP levels or that the function of GPA2 controls a pathway that is redundant with the Ras/cAMP pathway.
GPA2 acts in parallel to the RAS function
The synthetic slow growth phenotype observed in a Δgpa2 Δras2 strain is consistent with two models of signaling by Ras proteins and Gpa2p. One possible model is that Gpa2p acts upstream of Ras1p and Ras2p in the same signaling pathway. Upstream activation of this pathway would stimulate Gpa2p to transmit the signal to the Ras proteins by increasing their activity. If this were the case, the Δgpa2 Δras2 phenotype would be due to the low basal activity of Ras1p, which would be insufficient for full stimulation of adenylyl cyclase. An alternative model is that Gpa2p acts through a pathway that is independent of the Ras proteins. In this case, the Δgpa2 Δras2 phenotype would be due to the lack of stimulatory inputs from both of the parallel pathways. These two models can be distinguished by testing whether the effects of constitutive GPA2 occur in a strain that lacks all Ras proteins.
The constitutive GPA2 allele did not suppress the inviability of a Δras1 Δras2 strain (data not shown). The finding that activated Gpa2p cannot compensate for the lack of Ras function can be interpreted in a manner consistent with either of the models described above. If Gpa2p acts upstream of the Ras proteins, then activated Gpa2p would have no effect in the absence of Ras. Alternatively, if the Gpa2p and Ras pathways act independently, activation of Gpa2p alone could fail to maintain viability.
To determine the effect of constitutive Gpa2p in a RAS null strain, GPA2R273A was overexpressed in a strain containing deletions of the RAS1, RAS2 and PDE2 genes. In this strain, the lethal phenotype associated with RAS null alleles is suppressed by a deletion of the phosphodiesterase gene (Wilson and Tatchell, 1988). Overexpression of GPA2R273A conferred a growth advantage in a Δras1 Δras2 Δpde2 background when compared with vector alone (Figure 3A). Measurement of growth rates indicated that the strain containing the vector had a doubling time of ∼225 min, whereas the strain containing GPA2R273A had a doubling time of ∼102 min, a difference of ∼2‐fold. This result is consistent with the idea that GPA2 and RAS act in independent pathways. However, because the GPA2R273A allele did not have a very dramatic effect on the growth rate of a RAS null strain, another experiment was performed to confirm this result.
In wild‐type cells, the GPA2R273A allele conferred a significant decrease in heat shock resistance (Figure 1B), demonstrating that this parameter is a sensitive measure of GPA2 function. The effect of GPA2R273A on thermotolerance was therefore tested in a RAS null strain. Whereas Δras1 Δras2 Δpde2 cells containing vector displayed 85% survival after heat shock, the same cells overexpressing GPA2R273A displayed only 3% survival (Figure 3B). These results indicate that RAS function is not required for the physiological effects of constitutive GPA2, and strongly suggest that the GPA2 and RAS genes act in independent pathways.
GPA2 function requires SCH9
The results presented above demonstrate that, although constitutive GPA2 confers phenotypes similar to those produced by high cAMP levels, it does not act through the RAS genes. These characteristics are similar to those of the SCH9 gene, which also acts in parallel to RAS (Toda et al., 1988). SCH9 encodes a kinase related to cAMP‐dependent kinase, and its overexpression compensates for the lack of RAS function. SCH9 has been shown recently to be required for the response of nitrogen‐starved cells to the re‐addition of nitrogen (Crauwels et al., 1997). To investigate whether the GPA2 function requires SCH9, the effect of GPA2R273A on heat shock resistance was determined in a Δsch9 strain. In contrast to the result obtained with a strain lacking RAS function, expression of GPA2R273A in a Δsch9 strain had no effect on heat shock resistance (Figure 3B). This finding suggests that SCH9 is required for the function of GPA2, and is consistent with a model in which the Sch9p kinase acts downstream of Gpa2p in a signaling pathway that does not include Ras. To test whether RAS2 function also requires SCH9, the effect of an activated allele of RAS2 on heat shock resistance was determined in a Δsch9 strain. Expression of activated RAS2 in either a wild‐type or Δsch9 strain conferred the same degree of heat shock sensitivity (Figure 3C), suggesting that RAS2 function is independent of SCH9.
Isolation of GPR1, a putative receptor gene
To isolate other components of the GPA2 signaling pathway, a two‐hybrid protein interaction screen (Fields and Song, 1989) was performed using GPA2 as the bait. Screening of a yeast genomic library with a GPA2 fusion construct resulted in the isolation of plasmids containing short segments of an uncharacterized gene that was given the name GPR1. The full‐length GPR1 gene encodes a protein of 961 amino acids (DDBJ/EMBL/GenBank accession No. Z74083) that is predicted to contain seven membrane‐spanning domains, a feature characteristic of G protein‐coupled receptors (Figure 4A). The putative structure of this protein indicates that it would contain a very large third cytoplasmic loop of ∼346 amino acids, and a large cytoplasmic tail of ∼281 amino acids. The third cytoplasmic loop contains two copies of a short, basic sequence; one copy is present at the N‐terminal end of the loop and the other copy is present at the C‐terminal end (Figure 4A and B, boxed). The third cytoplasmic loop also contains a polyasparagine stretch of unknown function.
In contrast to the pheromone receptors, which have no homology to other receptors of this class, Gpr1p can be aligned with the G protein‐coupled receptor superfamily (Baldwin, 1993). In particular, Gpr1p contains several amino acids in its transmembrane domains that are conserved within this superfamily. Of these, the most highly conserved residues are the alanine at position 193 in transmembrane domain 4, the phenylalanine at position 262 in transmembrane domain 5, the tryptophan at position 634 in transmembrane domain 6, and the tyrosine at position 676 in transmembrane domain 7 (Figure 4A). When these residues are positioned with respect to the predicted arrangement of the receptor transmembrane α‐helices (Baldwin, 1993), they all face away from the surrounding membrane lipid and toward the center of the molecule or the other helices. Intramolecular interactions between these transmembrane α‐helices are thought to maintain the structure of the receptor in the membrane and allow it to bind the G protein.
Two GPR1‐containing plasmids were isolated in the two‐hybrid screen; one contained the coding region for the C‐terminal 122 amino acids and the other contained the coding region for the C‐terminal 99 amino acids (Figure 4B). The cytoplasmic tail regions of several mammalian G protein‐coupled receptors have also been shown to interact with Gα subunits, although in these cases the membrane‐proximal region of the cytoplasmic tail contains the Gα‐binding activity (O'Dowd et al., 1988; König et al., 1989; Münch et al., 1991; Ohyama et al., 1992; Hawes et al., 1994). The finding that the C‐terminal end of the Gpr1p cytoplasmic tail interacts with Gpa2p suggests that other Gα subunits may also interact with this region of their associated receptors. If the assays used previously to measure α‐subunit–receptor binding are less sensitive than the two‐hybrid assay, this area of contact could have been overlooked.
GPR1 acts upstream of GPA2
To determine if GPR1 acts in the same signaling pathway as GPA2, a diploid strain heterozygous for GPR1 and RAS2 deletion alleles was sporulated and tetrads were dissected. Strains containing single Δgpr1 or Δras2 mutations grew normally, but strains containing both Δgpr1 and Δras2 mutations displayed a severe growth defect (Figure 5A). The slow growth rate of Δgpr1 Δras2 strains was essentially identical to that of Δgpa2 Δras2 strains (Figure 2A), suggesting that GPR1 and GPA2 function in the same process. An experiment was therefore performed to determine the effect of different GPA2 alleles on the growth rate of a Δgpr1 Δras2 strain. A single copy plasmid containing GPA2 had no effect on the growth rate of the Δgpr1 Δras2 strain; however, multicopy GPA2 partially suppressed the growth defect of this strain (Figure 5B). Moreover, the constitutive GPA2R273A allele in single copy completely suppressed the growth phenotype of the Δgpr1 Δras2 strain. The most straightforward interpretation of these results is that Gpa2p acts downstream of Gpr1p in the same signaling pathway, as would be expected for a Gα subunit and its associated receptor. In addition, the finding that Δgpa2 and Δgpr1 mutations produce the same degree of growth inhibition in a Δras2 strain suggests that Gpr1p is the only receptor that is coupled to Gpa2p. The idea that Gpr1p and Gpa2p act in the same signaling pathway is also supported by the finding that the growth defect of a Δgpa2 Δgpr1 Δras2 strain is no more severe than that of a Δgpa2 Δras2 strain (data not shown).
GPR1 is not required for GPA2 expression or basal activity
The genetic experiment that places the function of GPR1 upstream of GPA2 is consistent with more than one possible relationship of their gene products. As mentioned above, a likely possibility is that GPR1 encodes the receptor that couples to Gpa2p. However, alternative possibilities are that the GPR1 gene product is required for the expression of the GPA2 gene or that it is required to maintain the stability or activity of the Gpa2p protein. To test these possibilities, the effect of GPA2 on heat shock sensitivity was determined in cells lacking GPR1 function.
Wild‐type cells carrying a single copy plasmid with the constitutive GPA2R273A allele under its own promoter (scGPA2*, Figure 5C) were 13‐fold more sensitive to heat shock than cells carrying vector alone. Expression of GPA2R273A conferred a similar increase in heat shock sensitivity on Δgpr1 cells. If the function of the GPR1 gene product was to promote efficient expression of the GPA2 gene, then a null allele of GPR1 would be expected to decrease the expression of GPA2R273A and thus decrease its ability to confer heat shock sensitivity. These results therefore demonstrate that GPR1 is not required for efficient expression of GPA2.
Overexpression of the wild‐type GPA2 gene by expressing it from the GAPDH promoter on a multicopy plasmid conferred a modest 2‐fold increase in heat shock sensitivity on wild‐type cells (mcGPA2, Figure 5C). Overexpression of GPA2 also conferred an ∼2‐fold increase in heat shock sensitivity on Δgpr1 cells when compared with vector alone in the same cells. Therefore, the basal activity of Gpa2p is maintained in the absence of a functional GPR1 gene, suggesting that GPR1 is not required for the stability or activity of the Gpa2p protein.
Strains containing the Δgpr1 mutation and wild‐type GPA2 were slightly more sensitive to heat shock than the corresponding GPR1 strains, suggesting that the GPA2 pathway is activated to a low level in Δgpr1 strains. This phenotype can be compared with deletion of the pheromone receptor gene STE3, which confers an ∼2‐fold increase in the basal activity of the pheromone response pathway (Boone et al., 1993). Therefore, these results are entirely consistent with the assignment of Gpr1p as the receptor that couples to Gpa2p.
Gpr1p is localized to the cell surface
If Gpr1p is a member of the G protein‐coupled receptor family, then it should be located at the cell surface. To determine the subcellular location of Gpr1p, the GPR1 gene was fused with the coding sequence of green fluorescent protein (GFP; Chalfie et al., 1994) and transformed into wild‐type cells. The GPR1–GFP construct complemented the growth defect of a Δgpr1 Δras2 strain, demonstrating that the fusion gene is fully active (data not shown). Cells expressing GPR1–GFP showed a cell surface staining pattern, demonstrating that Gpr1p is localized at the plasma membrane (Figure 6). In addition to cell surface staining, a portion of the signal appeared in discrete foci within cells, suggesting that Gpr1p may also be located on intracellular vesicles.
Membrane‐proximal regions of the Gpr1p third cytoplasmic loop are required for function
A number of studies have demonstrated that G protein‐coupled receptors contain sequences in the membrane‐proximal regions of their third cytoplasmic loops that are required for coupling to the G protein (Baldwin, 1994). The Gpr1p third cytoplasmic loop contains the sequences KRIKAQIG near its N‐terminal end and KKRRAQIQ near its C‐terminal end (Figure 4A and B, boxed). A related sequence is present in the third cytoplasmic loop of the S.cerevisiae pheromone receptors, which have very short third cytoplasmic loops. Mutation of some these residues in the α‐factor receptor affects its ability to couple to the Gpa1p Gα subunit without affecting its ability to bind ligand (Clark et al., 1994). Likewise, the pheromone receptors from Schizosaccharomyces pombe have a related sequence in their third cytoplasmic loops (Kitamura and Shimoda, 1991; Tanaka et al., 1993). An alignment of these sequences is shown in Figure 7A.
To test whether the membrane‐proximal regions of the third cytoplasmic loop of Gpr1p are required for its function, each of these regions was deleted individually from the GPR1 coding sequence. GPR1 mutations in which the coding region contained a deletion of eight amino acids at the N‐ (residues 277–284) or C‐terminal region (residues 610–617) of the third cytoplasmic loop were unable to complement the growth defect of a Δgpr1 Δras2 strain (Figure 7B). The third cytoplasmic loop of Gpr1p also contains a long stretch of polyasparagine residues (Figure 4B, boxed). To determine whether this asparagine‐rich sequence is required for Gpr1p function, a GPR1 mutation containing a deletion of this region (residues 490–586) was also constructed. The GPR1d490–596 gene was able to complement the growth defect of a Δgpr1 Δras2 strain (Figure 7B).
The abundance and localization of the mutant Gpr1p proteins was investigated by tagging each construct with GFP. To determine the relative abundance of the mutated versions of Gpr1p, an immunoblot containing cell extracts from strains expressing each of the GPR1 deletions was probed with anti‐GFP antiserum. The GPR1d277–284, GPR1d490–;596 and GPR1d610–617 constructs all expressed proteins at a level equal to or higher than the wild‐type expression level (Figure 7C, lanes 2–5). Localization of the mutated versions of Gpr1p was determined by observing cells expressing the GFP‐tagged versions of the proteins by fluorescence microscopy. The GPR1d277–284, GPR1d490–596 and GPR1d610–617 constructs all expressed proteins that were localized to the cell surface (Figure 7D). These results demonstrate that deletion of an internal region of the Gpr1p third cytoplasmic loop that encompasses 97 amino acids has no effect on the function of Gpr1p, but that deletion of the membrane‐proximal regions of this loop abolishes the function of Gpr1p. This evidence supports the idea that Gpr1p is a member of the G protein‐coupled receptor family.
It was also of interest to determine whether the region of GPR1 that encodes the cytoplasmic tail of the protein is required for its function because this portion of GPR1 was isolated in the two‐hybrid screen based on its interaction with Gpa2p. However, constructs that deleted most of the Gpr1p cytoplasmic tail (residues 694–954) or the smallest region that was isolated in the two‐hybrid screen (residues 841–954) did not produce a protein product that was detectable by immunoblot (Figure 7C, lanes 6 and 7), so this region appears to be important for some step in the production or stabilization of Gpr1p.
GPR1 RNA is induced in response to starvation for nitrogen and amino acids
To test whether the GPR1 gene is regulated by the availability of nutrients, the effect of nitrogen starvation on the abundance of GPR1 RNA was determined. RNA samples were isolated from cells in log phase, from cells that had been starved for nitrogen and essential amino acids for 24 h, and from starved cells to which asparagine and essential amino acids had been added back for 2 h. The abundance of GPR1 RNA increased to a very high level in cells starved for nitrogen and amino acids compared with its abundance in cells growing in log phase (Figure 8, lanes 1 and 2). Addition of essential amino acids and asparagine, an efficient nitrogen source, to starved cells caused a decrease in the abundance of GPR1 RNA (Figure 8, lane 3). Cells starved for a carbon source did not display induction of GPR1 RNA (data not shown), suggesting that this induction is not a general response to growth arrest. In addition, there was no difference in the abundance of GPR1 RNA in cells growing on a fermentable carbon source compared with cells growing on a non‐fermentable carbon source (data not shown). To determine whether induction of GPR1 RNA requires amino acid starvation, a strain that is prototrophic for all amino acids was starved solely for nitrogen. Likewise, an auxotrophic strain was starved for nitrogen in the presence of essential amino acids. In both of these cases, GPR1 RNA was not induced (Figure 8, lanes 4–9), indicating that amino acid starvation is necessary for this response. The induction of GPR1 RNA therefore appears to be a specific response to nitrogen and amino acid deprivation.
The Ras/cAMP pathway has long been thought to play a role in detecting and responding to nutrients, although the connection between pathway activation and nutrient availability has remained obscure. This work shows that the G protein α subunit Gpa2p functions in a signaling pathway that acts parallel to Ras and upstream of the Sch9p kinase. It also describes the isolation of the GPR1 gene, which encodes a putative G protein‐coupled receptor that is proposed to initiate the Gpa2p signaling pathway. The following observations support the idea that Gpr1p is a G protein‐coupled receptor and that it couples to Gpa2p: (i) Gpr1p shares sequence homology with members of the G protein‐coupled receptor family and is predicted to contain seven transmembrane domains; (ii) Gpr1p was shown to be located on the cell surface; (iii) Gpr1p physically associated with Gpa2p in the yeast two‐hybrid assay; (iv) the Δgpa2 and Δgpr1 mutations caused essentially identical severe growth defects in a Δras2 strain; (v) GPA2 acts downstream of GPR1 by genetic criteria because an activated allele of GPA2 compensated for the loss of GPR1; (vi) GPR1 was not required for the efficient expression of the GPA2 gene or for the basal activity of the Gpa2p protein; and (vii) Gpr1p contains short stretches of amino acids in the membrane‐proximal region of its third cytoplasmic loop that are homologous to sequences in other yeast G protein‐coupled receptors and that are required for its function. These results make a strong case for the assignment of Gpr1p as the receptor that couples to Gpa2p, and suggest that an extracellular ligand binds to Gpr1p and causes the activation of Gpa2p.
The results described here are consistent with two alternative models for the role of Gpa2p in growth control (Figure 9). In both cases, binding of ligand to the Gpr1p receptor is expected to stimulate guanine nucleotide exchange on Gpa2p, resulting in its activation. Both models require that Gpa2p does not act through Ras because the effects of a constitutive allele of GPA2 are not affected by deletion of both RAS genes. Therefore, it is likely that Gpa2p and the Ras proteins act independently and in parallel to elicit common responses. The two models propose different pathways for transmission of the signal from Gpa2p to downstream components. One possibility is that Gpa2p acts in a cAMP‐independent pathway that leads to activation of the Sch9p kinase (Figure 9, labeled A). The other possibility is that Gpa2p directly stimulates adenylyl cyclase, as is seen for Gαs activation of mammalian adenylyl cyclase (Figure 9, labeled B). In the latter case, it is necessary to propose that the function of Sch9p is at least partially required downstream of adenylyl cyclase activation (Figure 9B, dotted arrow) to explain the observation that deletion of SCH9 blocks the effect of activated Gpa2p on heat shock sensitivity. A regulatory link between SCH9 and the PKA pathway has been demonstrated by the finding that deletion of SCH9 causes an increase in PKA activity (Crauwels et al., 1997). Direct activation of adenylyl cyclase by Gpa2p is also consistent with the previous finding that overexpression of GPA2 causes a small increase in the level of cAMP induced by glucose (Nakafuku et al., 1988). In addition, activated Gpa2p causes changes in cell physiology that are similar to the changes caused by elevated PKA activity, such as increased heat shock sensitivity and reduced sporulation efficiency. However, we strongly favor the possibility that Gpa2p acts in a cAMP‐independent pathway that activates Sch9p (Figure 9). This possibility is supported by the following observations. First, deletion of GPA2 has no effect on the level of cAMP (Nakafuku et al., 1988). Second, the physiological parameters affected by constitutive Gpa2p can be regulated in a cAMP‐independent manner (Cameron et al., 1988), in addition to their known regulation through the cAMP/PKA pathway. Finally, all the genetic evidence indicates that GPA2 and RAS2 function in different pathways. For example, deletion of SCH9 completely blocks the heat shock sensitivity of strains expressing a constitutive allele of GPA2, but has no effect on the heat shock sensitivity of strains expressing a constitutive allele of RAS2. This result suggests that SCH9 functions downstream of GPA2 but does not function downstream of RAS2. Furthermore, constitutive GPA2 does not affect the growth rate of a Δsch9 strain (data not shown), although it does increase the growth rate of a Δras1 Δras2 Δpde2 strain. In contrast, a constitutive allele of RAS2 suppresses the growth defect of a Δsch9 strain (Toda et al., 1988), suggesting that SCH9 is not downstream of RAS2. Sch9p is most closely related to yeast and mammalian PKA, so it would not be surprising if the substrate specificity of these two kinases overlaps. Common substrates of these downstream kinases could account for the finding that mutational activation of GPA2 and RAS2 has similar physiological effects. The severe growth defect of strains that contain a Δras2 mutation in combination with a Δgpa2 or Δgpr1 mutation could be due to the fact that both the Ras and Gpa2p pathways are compromised in their ability to activate downstream kinases in the double mutant strains.
The model for the RAS and GPA2 pathways proposes that they are partially redundant. One question raised by this model is why RAS1 and RAS2 can compensate completely for the lack of GPA2 function, but GPA2 cannot compensate for the lack of all RAS function. A rationale for this observation is that the RAS pathway detects one type of nutrient, such as carbon, and the GPA2 pathway detects another type of nutrient, such as nitrogen. If cells can maintain a slow growth rate in the presence of a carbon source by using internal stores of nitrogen, it might be advantageous to have both an essential carbon detection pathway and a non‐essential nitrogen detection pathway. The nitrogen pathway would then contribute to growth control when the essential requirement for carbon is met. This mechanism is consistent with the finding that the presence of glucose is required for the response of nitrogen‐starved cells to the addition of nitrogen, as measured by trehalase activation (Thevelein, 1994). Nutrient‐mediated growth control in yeast has been proposed to involve the integration of signals from several different sensing pathways (Broach, 1991; Thevelein, 1994). The uncovering of a signaling pathway that acts parallel to the Ras pathway and is partially redundant with it suggests a way in which the cell could sum up the input from different sensors. If each signal activates kinases that have some common substrates, cell cycle progression could occur when a critical level of substrate phosphorylation is reached.
Several observations suggest that the GPR1/GPA2 pathway has a potential role in nutrient sensing, either in nitrogen detection or in a broader function of detecting nutrients other than carbon (Thevelein, 1994). First, the abundance of GPR1 RNA was found to increase under conditions of nitrogen and amino acid starvation, suggesting that the GPR1/GPA2 pathway is involved in regulating growth in response to the presence of these nutrients. The response of nitrogen‐starved cells to the addition of nitrogen and amino acids is independent of functional Ras proteins (Durnez et al., 1994), consistent with the involvement of the GPR1/GPA2 pathway. Secondly, the nitrogen response is defective in strains containing a null allele of SCH9 (Crauwels et al., 1997), which acts downstream of GPA2. Finally, it has been shown recently that GPA2 is required for pseudohyphal growth, which occurs in response to nitrogen starvation (Kübler et al., 1997; Lorenz and Heitman, 1997). However, it should be noted that the response to nitrogen also requires PKA (Durnez et al., 1994), although it is not associated with an increase in cAMP levels (Hirimburegama et al., 1992). Therefore, this response appears to involve both the PKA and the Sch9p kinases, and is probably not the result of a simple, linear signaling pathway, but rather involves cross‐pathway interactions. Interactions between the two pathways is one explanation for the observation that overexpression of GPA2 causes a small increase in the level of cAMP induced by glucose (Nakafuku et al., 1988). Although the PKA and Sch9p kinases are involved in the response to nitrogen, strains containing null alleles of these genes are defective for growth in the presence of all nutrients. Therefore, these kinases must also play a role during normal growth in complete medium. Similarly, Δras2 strains containing null alleles of GPR1 or GPA2 are defective for growth in the presence of all nutrients, suggesting that Gpr1p and Gpa2p also play a role in normal growth.
Previous work aimed at identifying specific molecules that act as growth control signals has been complicated by the metabolic requirement for nutrients. Thus the signal that ultimately causes Ras to be activated could be initiated by a molecule that functions both outside the cell as a ligand and inside the cell as a substrate for metabolic processes. Similarly, the ligand that binds the Gpr1p receptor could function both extracellularly and intracellularly. The isolation of a cell surface receptor that is likely to be involved in growth control by nutrients will allow a definitive approach for identifying an extracellular signal of this class.
Materials and methods
The GPA2 gene was cloned by amplifying a 1.6 kb fragment from yeast genomic DNA by polymerase chain reaction (PCR) using primers oGPA2‐1, 5′‐CCGGATCCCAGCTGCGCCCAAATGATTC‐3′ and oGPA2‐4, 5′‐CCGGATCCGCTGTGCATTCATTGTAACAC‐3′ (genomic sequences are underlined in all primers), each of which contains a flanking BamHI site. This fragment was cloned into the BamHI site of YCplac33 (Gietz and Sugino, 1988) to create pGPA2‐33.1. To construct a TRP1 disruption of GPA2, a 0.9 kb NruI–SmaI fragment from D759 was cloned into the MluI–BssHII sites of pGPA2‐33.1, which had been blunt‐ended using Klenow fragment, to produce pgpa2‐1::TRP1. To construct a multicopy plasmid with GPA2 under the control of the GAPDH promoter, a 1.4 kb fragment was amplified from yeast genomic DNA by PCR using primers oGPA2‐3, 5′‐CCGGATCCGCGAGCCTTATTGTTACAGC‐3′ and oGPA2‐4, each of which contains a flanking BamHI site. This fragment was cloned into the BamHI site of YEplac112 (Gietz and Sugino, 1988) under the control of the GAPDH promoter to produce pGPA2‐112.1. The GAPDH promoter was subcloned into YEplac112 from vector pAB23BXN as a 0.4 kb BamHI–BglII fragment. Site‐directed mutagenesis with primer oCTGPA‐2, 5′‐CTGACGTCATCTGTGCCGATCT‐3′ (changed nucleotides are in bold) was used to change the arginine at position 273 in GPA2 to an alanine (Transformer kit, Clontech). The altered gene was subcloned as a 1.4 kb BamHI fragment into vector YEplac112 under the control of the GAPDH promoter to produce pG2CT‐112.2. A single copy GPA2R273A plasmid was constructed by replacing the 1 kb MluI–BssHII fragment in pGPA2‐33.1 with the corresponding fragment containing the GPA2R273A allele to produce pG2CT‐33.2. The GPA2 construct used in the two‐hybrid screen was made by amplifying a 1.4 kb fragment from yeast genomic DNA by PCR using primers oGPHYB, 5′‐CCGGATCCTGGGTCTCTGCGCATCTTCA‐3′ and oGPA2‐4, each of which contains a flanking BamHI site. This fragment was cloned into the BamHI site of vector pGBT9 (Clontech) to produce pGBT9‐GPA2.
To construct a pde2::HIS3 allele, a 1.6 kb fragment containing the PDE2 gene was amplified from yeast genomic DNA by PCR using primers 5‐PDE2, 5′‐CGTCTAGAGATCACTACTACTTAATTG‐3′ and 3‐PDE2, 5′‐CGGTCGACACAATGAATGGTACAAGA‐3′, that contain an XbaI site or a SalI site. This fragment was cloned into XbaI–SalI‐digested pUC19 to create pUC19‐PDE2. The disruption construct was made by cloning a 1.8 kb HincII–SmaI fragment from pUC18‐HIS3 into the HpaI–EcoRV sites of pUC19‐PDE2, to produce ppde2‐1::HIS3. To construct a sch9::URA3 allele, a 1.4 kb fragment containing the SCH9 gene was amplified from yeast genomic DNA by PCR using primers o5SCH9, 5′‐CCGGATCCGAATAACATCAGAAAATGCC‐3′ and o3SCH9, 5′‐GGGGATCCAATCGCAAAGAGCGATGTTA‐3′, that contain BamHI sites. This fragment was cloned into BamHI‐digested pUC19 to create psch9.19. The disruption construct was made by cloning a 1.2 kb XbaI–BamHI fragment from pura3.BS into the NheI–BglII sites of psch9.19, to produce psch9.19::URA3.
To construct a gpr1::HIS3 allele, a 1.4 kb fragment containing the 5′ end of the GPR1 gene was amplified from yeast genomic DNA by PCR using primers 5TH110, 5′‐CGCTGCAGATGATAACTGAGGGATTT‐3′ and 3TH110, 5′‐GTCGCTGTTATCGTTCTT‐3′. This fragment was digested with PstI, which cuts within the primer, and with XbaI, which cuts within the GPR1 insert. The PstI–XbaI fragment was cloned into PstI–XbaI‐digested pUC19 to create pUC19‐GPR1N. The pgpr1‐1::HIS3 disruption construct was made by cloning a 1.8 kb HincII–SmaI fragment from pUC18‐HIS3 into the HpaI–BstBI sites of pUC19‐GPR1N, which had been blunt‐ended using Klenow fragment. The GPR1–GFP fusion gene was constructed using a GPR1 genomic clone (pGPR1‐50.1) that was obtained by screening bacterial colonies containing DNA from the YCp50 yeast library 3JDAF2 (Hirsch and Cross, 1993) with the 32P‐labeled 1.4 kb XbaI–PstI fragment from pUC19‐GPR1N. A 2.5 kb SacI–SalI fragment containing the GPR1 promoter and an N‐terminal portion of the coding sequence was subcloned from pGPR1‐50.1 into SacI–SalI‐digested YEplac112 to produce pGPR1N‐112.1. The 3′ end of the GPR1 gene from the unique SalI site to the end of the coding region was amplified by PCR using primer 5‐GPR1C, 5′‐AGTTGTCTCGTCGACGTCATT‐3′, which includes this SalI site, and primer 3‐GPR1C, 5′‐CGCTGCAGGCGGCCGCATAATGGTCCATTTCTTAAGAAG‐3′, which contains PstI and NotI sites. The product was digested with SalI and PstI and subcloned into the SalI–PstI sites of pGPR1N‐112.1 to produce pGPR1‐112.2, which reconstructs the entire GPR1 gene. The same fragments were used to construct a CEN plasmid containing the 3.4 kb SacI–PstI fragment that includes GPR1, which was called pGPR1‐22.2. To construct an in‐frame fusion between the GPR1 and GFP coding regions, a 0.7 kb NotI fragment containing the GFP gene was cloned into the NotI site at the end of the GPR1 coding region in pGPR1‐112.2 to create pGPR1‐GFP.1. The 3.4 kb SacI–PstI fragment from pGPR1‐112.2 was cloned into the SacI–PstI sites of pUC19 to create pGPR1‐19.2.
GPR1 deletions were made with the QuikChange kit (Stratagene) using either pGPR1‐GFP.1 or pGPR1‐22.2 as the template with the following primers: for GPR1d277–284, primer oGPR1DEL1, 5′‐TTCATTACCAGTGAAAGTGACTTTAACCATAACGTA‐3′, and its reverse complement; for GPR1d490–586, primer oGPR1DEL2, 5′‐AAGGAAAAAGGAGGCATCGACAGATGCGAAAATTCA‐3′, and its reverse complement; and for GPR1d610–617, primer oGPR1DEL3, 5′‐CAAACCTACAAACAAATGAAGAATCTAAGGGCAATA‐3′, and its reverse complement. GPR1d694–954 was made using the Transformer kit (Clontech) with primer oGPR1DEL4, 5′‐TGGGCAAAAACAGAATCAAAATTCTTAAGAAATGGACCA‐3′ using pGPR1‐19.2 as the template. The resulting deletion construct was cloned into the SacI–PstI sites of YCplac22 and YEplac112 to create pGPR1d694–954‐22.2 and pGPR1d694–954‐112.2, respectively. A 0.7 kb NotI fragment containing the GFP gene was cloned into the NotI site of pGPR1d694–954‐GFP.1 112.2 to create pGPR1‐GFP.1. GPR1d841–954 was made using the Transformer kit (Clontech) with primer oGPR1DEL5, 5′‐ATTCCAATGCTTGGCGGATTCTTAAGAAATGGACCATTA‐3′ using pGPR1‐19.2 as the template. pGPR1d841–954‐22.2, pGPR1d841–954‐112.2 and pGPR1d841–954‐GFP.1 were made as described for the GPR1d694–954 constructs.
Strain construction and media
Strains used in this study are listed in Table I. The GPA2 null allele was made by transformation of cells with the 1.4 kb BamHI fragment from pgpa2‐1::TRP1. The RAS2 null allele was made using the ras2::LEU2 construct from p530 as described (Tatchell et al., 1984). RAS1 null alleles were made either using the ras1::URA3 construct from p545 as described (Tatchell et al., 1984) or by transformation of cells with the 4 kb HindIII fragment from pras1‐1::LEU2. The PDE2 null allele was made by transformation of cells with the 3 kb SphI–SacI fragment from ppde2‐1::HIS3. The SCH9 null allele was made by transformation of cells with the 1.9 kb BamHI fragment from psch9.19::URA3. The sch9::TRP1null allele was made by transformation of an sch9::URA3 strain with a 3.8 kb SmaI fragment from marker swap plasmid pUT11 (Cross, 1997). The GPR1 null allele was made by transformation of cells with the 2.2 kb SphI–SacI fragment from pgpr1‐1::HIS3. Diploid H91 was made by crossing strain YX1B to strain YX2. Diploid H96 was made by crossing strain YX6B to strain YX2. Prototrophic strain W3031B.TLH was made by transforming strain W3031B with plasmids containing the TRP1, LEU2 and HIS3 genes.
Strains were grown on YEPD (2% glucose) or YEP‐Gal (3% galactose), and strains under selection were grown on synthetic dropout media, as described (Guthrie and Fink, 1991).
Two‐hybrid screen and yeast methods
pGBT9‐GPA2 was transformed into reporter strain HF7c (Clontech) and the resulting strain was transformed individually with each of three yeast genomic DNA fusion libraries, Y2HL‐C1, Y2HL‐C2 and Y2HL‐C3 (James et al., 1996). Transformation mixtures were plated on medium lacking histidine, and positive transformants were retested for β‐galactosidase expression by incubation in the presence of 0.3 mg/ml X‐gal. GPR1‐containing plasmids TH110 and TH112 were both isolated from library Y2HL‐C1. Controls for non‐specific protein interactions included co‐expression of pGBT‐GPA2 with a plasmid expressing a GAL4 activation domain fusion with SV40 large T‐antigen and co‐expression of TH1‐10 with a plasmid expressing a GAL4‐binding domain fusion with p53, both of which gave background levels of β‐galactosidase activity.
Yeast cells were sporulated by resuspending 0.1 ml of a saturated culture into 2.5 ml of sporulation medium (1% potassium acetate, 0.1% yeast extract, 0.05% glucose, 0.1 mM tryptophan, 0.2 mM leucine, 0.03 mM histidine, 0.05 mM uracil and 0.07 mM adenine) and incubating them at 30°C with shaking for 3 days.
Heat shock assays were performed by diluting an overnight saturated culture 1:20 into fresh medium and incubating it at 30°C with shaking for 2 days. Then 1 ml of this culture was removed into a glass tube which was placed in a 50°C water bath for 20 min. Heat‐shocked and non‐heat‐shocked cultures were then diluted and plated for counting.
Yeast cells were starved for nitrogen by growing them to log phase in YEPD and transferring them to medium containing 4% glucose, 0.26 mM adenine and 1.7% Difco yeast nitrogen base without amino acids and ammonium sulfate for 24 h, as described (Hirimburegama et al., 1992). Addition of nitrogen to starved cells was performed by adding asparagine and essential amino acids to the following final concentrations: 10 mM asparagine, 0.4 mM tryptophan, 0.9 mM leucine and 0.13 mM histidine.
Yeast transformations were performed by the lithium acetate method (Ito et al., 1983) modified as described previously (Hirsch and Cross, 1993). Yeast RNA was extracted from cells as described previously (Cross and Tinkelenberg, 1991).
Cell lysates were prepared by harvesting 12 ml of log phase cells, washing once with cold TE and resuspending in 150 μl of lysis buffer [50 mM Tris–HCl (pH 8.0), 1% SDS, 1 mM phenylmethylsulfonyl fluoride (PMSF), 1 μg of (apoprotin, leupeptin, chymostatin and pepstatin) per ml]. The mixture was added to acid‐washed glass beads (0.5 mm) and shaken at high speed for 10 min. Glass beads and cell debris were separated from the lysate by centrifugation in a microfuge for 2 min. The protein concentration of the samples was determined using a bicinchoninic acid protein assay kit (Pierce) and equal amounts were loaded onto SDS–polyacrylamide gels (10% polyacrylamide). Separated proteins were transferred to nitrocellulose and the blot was probed with anti‐GFP rabbit polyclonal antiserum at a dilution of 1:1000 or with anti‐phosphoglycerate kinase (PGK) rabbit polyclonal antiserum at a dilution of 1:300 000. Donkey anti‐rabbit immunoglobulin conjugated to horseradish peroxidase (Amersham) was used at a dilution of 1:10 000, and immune complexes were detected with an enhanced chemiluminescence kit (Amersham).
RNA was transferred to a nitrocellulose membrane after formaldehyde–agarose gel electrophoresis as described (Lehrach et al., 1977). The membranes were UV cross‐linked using a Stratalinker UV box. Pre‐hybridization and hybridization were done at 65°C in a buffer containing 0.9 M NaCl, 0.09 M sodium citrate, 0.1% Ficoll, 0.1% polyvinylpyrrolidone, 0.1% bovine serum albumin, 33 mM sodium pyrophosphate and 50 mM sodium phosphate monobasic. The probes used were gel‐purified DNA restriction fragments 32P‐labeled by random primer labeling using a Prime‐It kit (Stratagene). The fragments used were a 1.4 kb XbaI–MluI fragment from plasmid TH1‐10 and a 0.5 kb BamHI–XbaI fragment from pPGK1, which encodes phosphoglycerate kinase.
Cells containing the Gpr1p–GFP fusion protein were grown at room temperature and viewed using either the fluorescein isothiocyanate (FITC) filter for fluorescence microscopy or Nomarski optics for differential interference contrast microscopy on a Zeiss Axiophot microscope. They were photographed with a 100× objective.
We thank P.James and E.Craig for the yeast two‐hybrid libraries, K.Tatchell for plasmids used in this work, J.Kahana and P.Silver for anti‐GFP antiserum, and M.Lorenz and J.Heitman for communicating results prior to publication. We also thank an anonymous reviewer for helpful suggestions for experiments. This project was supported by a Grant‐In‐Aid from the American Heart Association, New York City Affiliate.
- Copyright © 1998 European Molecular Biology Organization