Yeast pseudohyphal growth is regulated by GPA2, a G protein α homolog

Michael C. Lorenz, Joseph Heitman

Author Affiliations

  1. Michael C. Lorenz1 and
  2. Joseph Heitman*,1
  1. 1 Departments of Genetics and Pharmacology and the Howard Hughes Medical Institute, Duke University Medical Center, Durham, NC, 27710, USA
  1. *Corresponding author. E-mail: heitm001{at}
View Full Text


Pseudohyphal differentiation, a filamentous growth form of the budding yeast Saccharomyces cerevisiae, is induced by nitrogen starvation. The mechanisms by which nitrogen limitation regulates this process are currently unknown. We have found that GPA2, one of the two heterotrimeric G protein α subunit homologs in yeast, regulates pseudohyphal differentiation. Δgpa2/Δgpa2 mutant strains have a defect in pseudohyphal growth. In contrast, a constitutively active allele of GPA2 stimulates filamentation, even on nitrogen‐rich media. Moreover, a dominant negative GPA2 allele inhibits filamentation of wild‐type strains. Several findings, including epistasis analysis and reporter gene studies, indicate that GPA2 does not regulate the MAP kinase cascade known to regulate filamentous growth. Previous studies have implicated GPA2 in the control of intracellular cAMP levels; we find that expression of the dominant RAS2Gly19Val mutant or exogenous cAMP suppresses the Δgpa2 pseudohyphal defect. cAMP also stimulates filamentation in strains lacking the cAMP phosphodiesterase PDE2, even in the absence of nitrogen starvation. Our findings suggest that GPA2 is an element of the nitrogen sensing machinery that regulates pseudohyphal differentiation by modulating cAMP levels.


In response to severe nitrogen starvation, diploid cells of the budding yeast Saccharomyces cerevisiae undergo a dimorphic transition known as pseudohyphal differentiation (Gimeno et al., 1992). During pseudohyphal growth, cells adopt a unipolar budding pattern, elongate and can invade the growth substrate. In addition, cells remain attached following cytokinesis, resulting in chains of cells reminiscent of fungal hyphae. It has been suggested that this mode of growth may allow this non‐motile species to forage for nutrients under adverse conditions (Gimeno et al., 1992; Kron et al., 1994).

Pseudohyphal differentiation is regulated, in part, by elements of the MAP kinase cascade which also controls mating, namely the protein kinases STE20, STE11 and STE7 and the transcription factor STE12 (Liu et al., 1993). A heterodimer composed of STE12 and TEC1 binds to a filamentation response element (FRE), which may regulate expression of genes necessary for pseudohyphal differentiation (Laloux et al., 1994; Mösch et al., 1996; Madhani and Fink, 1997). A reporter gene under the control of an FRE [FG(TyA)::lacZ] is induced under nitrogen starvation conditions in a STE12‐ and TEC1‐dependent manner (Gavrias et al., 1996; Mösch et al., 1996; Madhani and Fink, 1997). The proximal elements that regulate mating, the pheromone receptors (STE2 and STE3) and the heterotrimeric guanine nucleotide binding protein (GPA1, STE4 and STE18), are not expressed in diploids and play no role in pseudohyphal differentiation (Liu et al., 1993). The mechanisms by which nitrogen starvation activates the MAP kinase cascade or triggers filamentous growth are not known, but the parallels between the signalling pathways in haploid and diploid cells led us to consider whether pseudohyphal growth might also be G protein regulated.

Heterotrimeric G proteins mediate a vast array of signalling processes in all eukaryotes, including mating response and development in fungi, vision in mammals and metazoans, and neurotransmitter and hormone action in mammals (reviewed in Neer, 1995; Borkovich, 1996). These regulatory complexes typically serve as a bridge between transmembrane receptors and effectors, including adenylate cyclase, phospholipases or protein kinases. A ligand‐bound receptor can either activate or inhibit effectors via the G protein. The activity of the G protein is controlled by a guanine nucleotide cycle in which an activated receptor promotes GDP to GTP exchange on the α subunit of the G protein. GTP binding stimulates dissociation of the α subunit from the βγ complex, and either the free α or free βγ dimer, or in some cases both, regulate downstream effectors. The intrinsic GTPase activity of the α subunit hydrolyzes GTP to GDP, promoting reassociation of the heterotrimer and attenuation of signalling.

The best characterized Gα subunit in S.cerevisiae, GPA1 (also known as SCG1), regulates mating response by binding to the pheromone receptors STE2 (α‐factor receptor) and STE3 (a‐factor receptor; reviewed in Kurjan, 1993; Bardwell et al., 1994). In this pathway, the free βγ subunit transmits the signal to a conserved MAP kinase cascade through the interaction of a large multimeric complex (Leeuw et al., 1995; Akada et al., 1996). Haploid Δgpa1 mutants are inviable due to constitutive activation of the pheromone response pathway and subsequent cell cycle arrest (Miyajima et al., 1987). G proteins also mediate mating response in the fission yeast Schizosaccharomyces pombe. In this organism two distinct Gγ subunits regulate the two environmental signals required for mating: the presence of pheromone is signalled by Gpa1 (Obara et al., 1991) and nitrogen starvation is signalled by Gpa2 (Isshiki et al., 1992). Morphogenesis triggered by environmental stimuli is regulated by G protein action in the fungi Coprinus congregatus (Kozak et al., 1995) and Neurospora crassa (Ivey et al., 1996).

The yeast genome project has identified only one protein other than GPA1 likely to encode a Gα subunit, GPA2, which was first isolated on the basis of its homology to a rat brain Gαs isoform (Nakafuku et al., 1988). Addition of glucose to glucose‐starved yeast cells induces a rapid but transient increase in cAMP levels. Expression of GPA2 from a high copy plasmid enhances this rise in cAMP (Nakafuku et al., 1988; Papasavvas et al., 1992). The glucose‐stimulated cAMP pulse is partially inhibited in the presence of mating pheromone; this inhibition requires GPA2 and appears to involve interactions between GPA2 and RAS2 (Arkinstall et al., 1991; Papasavvas et al., 1992). Yeast Ras modulates adenylate cyclase activity and cAMP levels (Toda et al., 1985; Field et al., 1988), and overexpression of GPA2 suppresses the growth defect of a temperature‐sensitive ras2 mutant (Nakafuku et al., 1988). High levels of cAMP activate the cAMP‐dependent protein kinase (protein kinase A, PKA); among its many functions, PKA mediates cellular responses to nutrient stress, heat shock and oxidative damage (reviewed in Broach and Deschenes, 1990). GPA2 and S.pombe Gpa2 are highly homologous (42% identity). The fission yeast Gpa2 protein is known to regulate mating in response to nitrogen starvation; S.pombe gpa2 mutant cells mate under nutrient‐rich conditions unlike wild‐type gpa2+ cells (Isshiki et al., 1992). This phenotype is shared by S.pombe mutants lacking adenylate cyclase (Maeda et al., 1990), and Gpa2 regulates cAMP levels in fission yeast (Isshiki et al., 1992).

Ras also regulates filamentous growth in budding yeast, and activated mutants of yeast RAS2 stimulate pseudohyphal differentiation (Gimeno et al., 1992). It was initially proposed that RAS2 would regulate pseudohyphal differentiation by modulating cAMP levels (Gimeno et al., 1992), a suggestion supported by the finding that overproduction of the cAMP phosphodiesterase PDE2 inhibits pseudohyphal growth in wild‐type strains and blocks hyperfilamentation induced by the dominant active RAS2Gly19Val mutant (Ward et al., 1995). Other work, however, has suggested that Ras acts upstream of the MAP kinase cascade (Mösch et al., 1996; Madhani and Fink, 1997). The interactions observed between RAS2, GPA2 and cAMP led us to test whether GPA2 is involved in filamentous growth. We report here that Δgpa2/Δgpa2 mutant strains have a defect in filamentous growth. Moreover, a dominant active GPA2 allele stimulates pseudohyphal differentiation, even on nitrogen‐rich media. Epistasis analysis and reporter gene experiments, however, indicate that GPA2 does not act upstream of the MAP kinase cascade. We also find that cAMP stimulates filamentation, and both RAS2Gly19Val and cAMP suppress the pseudohyphal defect conferred by the Δgpa2 mutation. We propose that GPA2 is an element of the nitrogen sensing machinery which signals pseudohyphal differentiation under starvation conditions via a signalling pathway involving cAMP.


The GPA2 Gα protein regulates pseudohyphal growth

Elements upstream of the pheromone responsive MAP kinase cascade which sense nitrogen starvation during pseudohyphal growth have not yet been identified. In fission yeast, the Gα subunit Gpa2 regulates mating in response to nitrogen limitation; this G protein is distinct from the pheromone‐stimulated Gpa1 protein (Obara et al., 1991; Isshiki et al., 1992). These studies led us to test whether nitrogen sensing in budding yeast might also be G protein‐mediated. Previous work found no role for S.cerevisiae GPA1 in pseudohyphal differentiation (Liu et al., 1993), thus we focused on GPA2, which has been implicated in the regulation of cAMP levels and Ras functions (Nakafuku et al., 1988).

We replaced the GPA2 open reading frame with a gene encoding resistance to G418 in the Σ1278b background (Grenson et al., 1966) commonly used for analysis of filamentous growth (Gimeno et al., 1992). Cells lacking GPA2 (gpa2‐Δ1) were viable and had no obvious defects in growth, mating or sporulation (data not shown). A homozygous Δgpa2/Δgpa2 diploid strain, however, exhibited a defect in pseudohyphal differentiation when grown on low nitrogen (50 μM ammonium sulfate) SLAD medium (Figure 1). Δgpa2/Δgpa2 mutant strains filament weakly, as do strains with mutations of MAP kinase elements (with the exception of Δste20 mutant strains, which do not filament at all; Liu et al., 1993). Expression of wild‐type GPA2 complemented and restored pseudohyphal growth in Δgpa2/Δgpa2 mutant strains (data not shown). The pseudohyphal defect conferred by the Δgpa2 mutation was also observed when cells were grown in the presence of limiting concentrations of proline, glutamine or urea (data not shown), indicating that GPA2 is a general regulator of filamentation.

Figure 1.

GPA2 regulates pseudohyphal growth. Diploid strains of the Σ1278b background with the indicated genotypes (wild‐type, MLY61; Δgpa2/Δgpa2, MLY132a/α; Δste11/Δste11, HLY506) were incubated on nitrogen limiting SLAD medium for 4 days at 30°C.

A dominant active GPA2 allele enhances pseudohyphal growth

Several mutations have been described in other Gα subunits that perturb proper function of the G protein and have been invaluable in dissecting G protein function. One such mutation, in which valine replaces the second glycine in the highly conserved GXGXXG motif, has been shown to reduce the GTPase activity of Gαs and the small G protein Ras (Graziano and Gilman, 1989; Masters et al., 1989), thus promoting the GTP‐bound and active forms of these proteins. The human H‐rasGly12Val mutation confers a transforming phenotype (reviewed in Lowy and Willumsen, 1993). In yeast, RAS2Gly19Val constitutively activates adenylate cyclase, raises intracellular cAMP levels (Toda et al., 1985) and promotes pseudohyphal differentiation (Gimeno et al., 1992). The corresponding mutation in GPA1, Gly50Val, increases basal activation of the pheromone response pathway, leading to a growth defect and impaired adaptation to prolonged pheromone exposure (Kurjan et al., 1991).

To characterize further the function of GPA2, we constructed an allele with the analogous mutation, Gly132Val (GPA2‐2), under the control of a galactose‐inducible promoter in a low‐copy, centromeric plasmid. Expression of the GPA2Gly132Val allele on nitrogen limiting medium containing galactose greatly enhanced pseudohyphal differentiation (Figure 2A). This stimulation was not observed with expression of the wild‐type allele or when expression was repressed with glucose (Figure 2A). In addition, this activated allele did not promote filamentous growth when expressed in a haploid strain (data not shown). Strains expressing the GPA2Gly132Val allele grew more slowly than wild‐type strains, and formed smaller colonies. This is not unusual in hyperfilamentous strains, as overexpression of STE12 confers a similar phenotype (Liu et al., 1993, and data not shown).

Figure 2.

A dominant GPA2 allele stimulates pseudohyphal growth. (A) Wild‐type diploid strain MLY61 with a control plasmid (vector), or plasmids expressing GPA2 (pML180) or GPA2Gly132Val (pML160) under repressing (glucose, left) or inducing (galactose, right) conditions were incubated on nitrogen limiting medium (SLAD or SLARG) for 4 days at 30°C. The colonies were photographed at 25× magnification. (B) The strains and GPA2 plasmids indicated in (A) were incubated on nitrogen‐rich medium containing 5 mM ammonium sulfate for 2 days at 30°C under repressing (glucose, left) or inducing (galactose, right) conditions. The colonies were photographed at 50× magnification.

If GPA2 is indeed a component of the nitrogen sensor, the GPA2Gly132Val allele might induce filamentation under nitrogen‐rich conditions in which this developmental pathway is normally repressed. Consistent with this hypothesis, expression of the GPA2Gly132Val mutation stimulated filamentous growth on media containing 5 mM ammonium sulfate (Figure 2B), a 100‐fold excess over standard pseudohyphal media (SLAD; 50 μM ammonium sulfate). Filamentation also occured, albeit to a lesser extent, on the synthetic medium YNB, which contains an even higher nitrogen concentration (38 mM ammonium sulfate; data not shown). Thus, constitutive activation of GPA2 relieves the requirement for nitrogen starvation to induce pseudohyphal growth.

A dominant negative GPA2 allele inhibits pseudohyphal growth

We next introduced a mutation in the hinge region of GPA2 which would be predicted to result in a dominant negative allele. The corresponding mutation in other Gα subunits prevents conformational changes normally induced by GTP binding that are required for βγ release and signalling (Miller et al., 1988), resulting in a dominant negative phenotype. This mutation in yeast GPA1 complements the lethality of the Δgpa1 mutation, but confers a semidominant sterile phenotype (Kurjan et al., 1991). The corresponding mutant allele, GPA2Gly299Ala (GPA2‐3), again under the control of a galactose‐inducible promoter, did not complement the pseudohyphal defect of Δgpa2 mutant strains, as expected (data not shown). Expression of GPA2Gly299Ala inhibited pseudohyphal differentiation in wild‐type diploid strains (Figure 3A). Thus, two mutations predicted to alter GPA2 function both have significant effects on filamentous growth and demonstrate that pseudohyphal differentiation is G protein regulated.

Figure 3.

A dominant negative GPA2 allele inhibits pseudohyphal growth. Wild‐type diploid strain MLY61 expressing GPA2 (pML180) or GPA2Gly299Ala (pML179) under repressing (glucose, left) or inducing (galactose, right) conditions on nitrogen limiting media was incubated for 4 days at 30°C.

β and γ subunits remain to be identified

In several G protein mediated signalling pathways, both the α and the βγ subunits positively contribute to signalling. The residual filamentation observed in Δgpa2 mutants could be indicative of a role for a βγ complex in signalling, although it is clear that the pheromone‐responsive βγ subunits (STE4 and STE18) do not regulate pseudohyphal growth (Liu et al., 1993). Using the fungal β proteins STE4 (S.cerevisiae) and Gpb1 (S.pombe) in BLAST searches, we identified eight candidate β genes in the S.cerevisiae genome; gene disruption experiments did not reveal a role for any of these genes in pseudohyphal differentiation (see Materials and methods). STE18 is the only known fungal γ subunit (Whiteway et al., 1989); none of the three candidate γ subunits we identified affected pseudohyphal growth (see Materials and methods). Other methods will be necessary to identify these components.

GPA2 does not regulate the MAP kinase cascade

Previous work has identified several genes and mutant alleles that promote filamentous growth (Gimeno et al., 1992; Liu et al., 1993; Gimeno and Fink, 1994), and these allowed us to test the point at which GPA2 acts by epistasis analysis. Both the dominant STE11‐4 allele and overexpression of STE12 stimulate pseudohyphal differentiation when expressed in wild‐type strains and suppress mutations of upstream components (Liu et al., 1993). If GPA2 were to act solely upstream of the MAP kinase cascade, expression of these alleles should suppress the pseudohyphal defect conferred by the Δgpa2 mutation. To our surprise, neither STE11‐4 nor increased STE12 expression resulted in filamentation in a Δgpa2/Δgpa2 mutant strain (Figure 4A). These observations suggest that GPA2 does not act upstream of the MAP kinase elements that regulate filamentous growth. Likewise, overexpression of PHD1, which induces hyperfilamentation in other strains (Gimeno et al., 1992; Gimeno and Fink, 1994), had no effect in Δgpa2/Δgpa2 strains. In contrast, expression of the dominant RAS2Gly12Val mutant did suppress the pseudohyphal defects conferred by the Δgpa2 mutation. This finding is consistent with a role for GPA2 in regulating cAMP levels, as has been suggested (Nakafuku et al., 1988). Others have proposed that RAS2 lies solely upstream of the MAP kinase elements (Mösch et al., 1996); however, we also find that RAS2Gly12Val suppresses the pseudohyphal defects of the Δste7, Δste11 and Δste12 mutants, findings which do not support this model. RAS2 may instead have multiple roles in regulating filamentous growth, as has recently been suggested (Mösch and Fink, 1997).

Figure 4.

GPA2 functions independently of the MAP kinase pathway. (A) Diploid wild‐type (MLY61) or Δgpa2/Δgpa2 (MLY132a/α) strains expressing RAS2Gly19Val, STE11‐4 or PHD1 were incubated on SLAD medium for 4 days at 30°C. The colonies were photographed at 25× magnification. (B) Diploid wild‐type (L3566), Δste12/Δste12 (HLY506) or Δste20/Δste20 (HLY492) strains expressing GPA2 (pML180) or GPA2Gly132Val (pML160) were incubated on SLARG medium for 4 days at 30°C. The colonies were photographed at 50× magnification.

The dominant active GPA2Gly132Val allele allowed us to address the relationship between GPA2 and the MAP kinase elements by another means. The GPA2Gly132Val allele was expressed in diploid strains carrying homozygous deletions of STE20, STE11, STE7 and STE12. None of the Δste mutations blocked the ability of GPA2Gly132Val to stimulate pseudohyphal differentiation (Figure 4B). Moreover, as in wild‐type STE+ strains, GPA2Gly132Val also allowed pseudohyphal growth of Δste mutant strains on nitrogen‐rich medium (data not shown). Filamentation in the Δste20/Δste20 strain expressing GPA2Gly132Val was somewhat reduced compared with wild‐type or other Δste mutants (Figure 4B), which is consistent with previous observations that the Δste20 mutant phenotype is more severe than that of other Δste mutants (Liu et al., 1993). These observations again support the conclusion that GPA2 does not regulate pseudohyphal differentiation via the MAP kinase cascade.

GPA2 alleles do not affect expression of a MAP kinase‐regulated reporter gene

To address an alternate explanation for these epistasis results, namely that GPA2 might regulate both the MAP kinase cascade and another pathway, we employed a reporter gene previously found to respond to MAP kinase activation under nitrogen starvation conditions (Laloux et al., 1994; Mösch et al., 1996). This reporter, FG(TyA):: lacZ, was first identified as a control element for the transposable element Ty1, and includes binding sites for TEC1 and STE12. It is induced upon nitrogen starvation but not in response to mating pheromone (Gavrias et al., 1996; Mösch et al., 1996; Madhani and Fink, 1997). Expression of the dominant active GPA2 allele had no affect on the activity of this reporter gene (Table I), and pseudohyphal growth induced by GPA2Gly132Val on nitrogen‐rich media (see Figure 2C) was not accompanied by an increase in reporter activity (Table I). Conversely, the dominant‐negative GPA2Gly299Ala allele did not prevent induction of FG(TyA)::lacZ under low nitrogen conditions, despite inhibiting pseudohyphal growth (Table I). In addition, alleles which induce expression of this reporter, such as STE11‐4 (Mösch et al., 1996), were not affected by deletion of GPA2 (Table II), even though STE11‐4 does not suppress the pseudohyphal defects of Δgpa2 mutant strains. Furthermore, the GPA2Gly132Val allele, while it suppressed the filamentation defect of Δste mutants, did not rescue reporter activity in these strains (Table III). Thus, the phenotypes conferred by mutant GPA2 alleles occur in the absence of any detectable difference in activation of the MAP kinase cascade, indicating that GPA2 does not regulate this pathway. Moreover, the activity of the FG(TyA)::lacZ reporter gene and the morphological response of filamentation can be separated under some conditions.

View this table:
Table 1. Activation of a FG(TyA)::lacZ reporter by GPA2 alleles
View this table:
Table 2. FG(TyA)::lacZ induction in GPA2 versus Δgpa2 strains
View this table:
Table 3. Effects of GPA2 alleles on FG(TyA)::lacZ expression in Δste strains

Pseudohyphal growth is regulated by cAMP

Earlier reports have suggested a connection between GPA2, Ras and cAMP metabolism (Nakafuku et al., 1988; Papasavvas et al., 1992). Our findings that GPA2 regulates filamentation and that the Δgpa2 mutation is suppressed by RAS2Gly19Val led us to test whether cAMP might also regulate this dimorphic transition. The efficacy of exogenous cAMP in yeast is greatly enhanced by mutations in the high affinity cAMP phosphodiesterase PDE2. Addition of 1 mM cAMP to nitrogen‐limiting media stimulated filamentation of diploid Δpde2/Δpde2 mutant strains, and filamentation was even more dramatic at 10 mM cAMP (Figure 5A). At 10 mM cAMP, even wild‐type (PDE2+) strains exhibit enhanced pseudohyphal growth (Figure 5A), and pseudohyphal differentiation also occured on nitrogen‐rich media (5 mM NH4+; Figure 6). These effects of cAMP were not observed with either AMP or cGMP (data not shown), indicating a specific role for the second messenger cAMP in regulating pseudohyphal differentiation.

Figure 5.

cAMP stimulates pseudohyphal differentiation. (A) Homozygous diploid strains with genotypes: wild‐type (MLY61), Δpde2/Δpde2 (MLY162a/α) or Δgpa2/Δgpa2 Δpde2/Δpde2 (MLY171a/α) were incubated on SLAD medium containing the indicated concentrations of cAMP for 4 days at 30°C. (B) Strains of genotypes Δpde2 MATα (MLY173), Δste7 Δpde2 MATa (MLY174) and Δste12 Δpde2 MATa (MLY175) expressing a plasmid‐borne copy of the opposite MAT locus were incubated on SLAD medium with or without 10 mM cAMP for 4 days at 30°C.

Figure 6.

cAMP allows pseudohyphal growth on nitrogen‐rich media. Homozygous diploid wild‐type (MLY61) and Δpde2/Δpde2 strains (MLY162a/α) were incubated on media containing 5 mM ammonium sulfate without (top) or with (bottom) 10 mM cAMP for 8 days at 30°C.

If the role of GPA2 is to regulate intracellular cAMP levels, as has been previously suggested (Nakafuku et al., 1988; Papasavvas et al., 1992), the pseudohyphal defects of Δgpa2 mutant strains might be suppressed by exogenous cAMP. To test this hypothesis, we constructed a homozygous diploid strain lacking both GPA2 and PDE2. As shown in Figure 5A, 100 μM cAMP restored filamentation to approximately wild‐type levels in a Δgpa2/Δgpa2 Δpde2/Δpde2 mutant strain. The effects of cAMP are not mediated by the MAP kinase pathway, as cAMP also stimulated filamentation in Δste7 Δpde2 and Δste12 Δpde2 mutant strains (Figure 5B), as did expression of the dominant active GPA2Gly132Val allele (Figure 4B and data not shown). Surprisingly, cAMP actually repressed expression of the FG(TyA)::lacZ reporter under nitrogen limiting conditions (Table IV). This is another example in which increased filamentous growth, as assayed morphologically, was not accompanied by increased expression of this reporter. These findings provide support for a model in which GPA2 regulates pseudohyphal differentiation by stimulating cAMP production and regulating a signalling pathway independent of the MAP kinase cascade.

View this table:
Table 4. Effects of cAMP on FG(TyA)::lacZ expression


Pseudohyphal differentiation is regulated by the Gα subunit GPA2

We find that cells lacking the Gα homolog GPA2 have a defect in pseudohyphal development. A constitutively active mutant allele, GPA2Gly132Val, which is predicted to have decreased GTPase activity, stimulates pseudohyphal growth; remarkably, this occurs even under nitrogen‐rich conditions. These findings suggest that GPA2 is a component of the nitrogen sensing machinery. A second mutation, GPA2Gly299Ala, which is predicted to prevent βγ release and interaction with signalling effectors, results in a dominant‐negative GPA2 allele that inhibits filamentation in wild‐type strains. We propose that GPA2 detects nitrogen starvation conditions through its interaction with an as yet unknown receptor (see below) and stimulates filamentous growth. Our data also indicate that GPA2 plays a positive role in signalling, in contrast to the pheromone response pathway in which the βγ complex transduces the signal and the GPA1 α‐subunit serves an inhibitory role.

GPA2 does not regulate the MAP kinase cascade

Our initial model was that GPA2 might act analogously to the G protein‐containing GPA1 and regulate the MAP kinase cascade in response to the appropriate stimuli. Epistasis analysis between GPA2 mutations and alleles of the MAP kinase cascade indicates that this is not the case. The Δgpa2 pseudohyphal defect is not suppressed by constitutive alleles of MAP kinase components, whereas the dominant active GPA2Gly132Val allele suppresses the filamentation defect in mutants lacking STE20, STE11, STE7 or STE12. Moreover, despite the dramatic morphological phenotypes, mutant GPA2 alleles had no effect on the expression of a filamentation response element regulated by STE12 and TEC1. Thus, the mechanism of activation of the MAP kinase cascade remains unknown. Nitrogen starvation does induce this reporter gene, even in the absence of GPA2, so it is likely that nutrient limitation regulates this pathway as well (Mösch et al., 1996; Madhani and Fink, 1997). Bypass of the MAP kinase pathway has been demonstrated in Candida albicans, in which single mutant strains lacking homologs of STE20, STE7 and STE12 (CST20, HST7 and CPH1, respectively) do undergo filamentous growth in response to some, but not all, environmental signals (Liu et al., 1994; Kohler and Fink, 1996; Leberer et al., 1996). As outlined in Figure 7, we propose that at least two signalling pathways, either parallel or partially interconnected, regulate pseudohyphal growth; one comprised by the MAP kinase cascade and the other including GPA2. Our findings, and the studies of others (Mösch et al., 1996; Madhani and Fink, 1997), suggest that both of these pathways may be nitrogen‐responsive; whether this occurs by a common mechanism is yet to be determined.

Figure 7.

A model for regulation of pseudohyphal differentiation. We propose that at least two pathways, one GPA2‐dependent and one MAP kinase‐dependent, regulate pseudohyphal differentiation in response to nitrogen starvation. In this model, Ras may participate in both signalling pathways. See text for more details.

A role for cAMP in filamentous growth

Our studies have also demonstrated that exogenous cAMP dramatically stimulates pseudohyphal differentiation, particularly in mutants lacking the cAMP phosphodiesterase PDE2. cAMP promotes filamentous growth on nitrogen‐rich media and suppresses the pseudohyphal defects of strains lacking STE7 or STE12, phenotypes shared with the dominant active GPA2 allele. In addition, cAMP restores filamentation to Δgpa2 Δpde2 mutant strains. The activity of cAMP provides one possible explanation for the hyperfilamentous phenotype of RAS2Gly19Val mutant strains, which hyperactivate adenylate cyclase, thus raising intracellular cAMP (Toda et al., 1985; Field et al., 1988). In addition, we find that RAS2Gly19Val suppresses the pseudohyphal defect of Δgpa2 mutant strains, while expression of the dominant STE11‐4 does not. Thus, the well‐established role for Ras in regulation of cAMP levels appears to also be operative in filamentous growth. This suggestion, made previously (Gimeno et al., 1992), is supported by the finding that overexpression of PDE2 suppresses the enhanced filamentation phenotype conferred by activated RAS2Gly19Val (Ward et al., 1995). However, RAS2 has also been implicated as an upstream regulator of the MAP kinase cascade (Mösch et al., 1996), as it is in signalling pathways in some multicellular organisms. Because cAMP suppresses the pseudohyphal defect of Δste mutant strains, the effects of cAMP are not mediated by the MAP kinase cascade. Interestingly, cAMP actually represses the FG(TyA)::lacZ reporter, which responds to MAP kinase activation. Further study will be necessary to clarify the role of the target of cAMP, PKA, in dimorphism in yeast. It seems likely that Ras regulates filamentous growth through multiple pathways, as has been recently proposed (Mösch and Fink, 1997), possibly one branch via cAMP and another via the MAP kinase cascade (see Figure 7).

Conserved role of G proteins and cAMP in fungal development

Regulation of fungal morphogenesis and development by G proteins and cAMP is not limited to S.cerevisiae. The requirement for nitrogen starvation for mating in S.pombe is bypassed by null mutations in Gpa2 (Isshiki et al., 1992), the fission yeast homolog of GPA2. Mating in rich media is also a phenotype of adenylate cyclase (cyr1) mutants in fission yeast, and this phenotype is suppressed by exogenous cAMP (Maeda et al., 1990). Fission yeast strains carrying dominant active Gpa2 mutations have elevated cAMP levels (Isshiki et al., 1992). In the corn pathogen Ustilago maydis, filamentous growth and pathogenicity are regulated by the mating type loci and result from conjugation of compatible cell types (Hartmann et al., 1996). The U.maydis Gα subunit Gpa3 has recently been shown to regulate both virulence and mating response (Regenfelder et al., 1997). G proteins also regulate development in N.crassa as mutants lacking the Gα homolog Gna1 are female‐sterile and defective in macroconidia formation (Ivey et al., 1996).

Changes in cAMP levels have been implicated in dimorphic transitions, particularly between the yeast form and the filamentous or mycelial form, in a number of fungi. In U.maydis, adenylate cyclase mutants have a constitutively filamentous phenotype which can be suppressed by exogenous cAMP or by a mutation in the protein kinase A regulatory subunit Ubc1 (Gold et al., 1994). In N.crassa, a PKA homolog, cot‐1, is required for hyphal elongation (Yarden et al., 1992). Elevated cAMP levels have been associated with mycelial growth or germ‐tube formation in the dimorphic species Histoplasma capsulatum, Blastomyces dermatitidis and C.albicans (Niimi et al., 1980; Medoff et al., 1986; Sabie and Gadd, 1992). The stimulation of filamentous growth by elevated cAMP in these organisms is similar to our findings in S.cerevisiae, whereas in U.maydis and Mucor racemosus, lower cAMP levels are found in the hyphal state (Paznokas and Sypherd, 1975; Gold et al., 1994). Thus, regulation of cAMP metabolism is closely linked to dimorphism in a diverse group of fungi.

Nitrogen sensing: GPA2‐coupled receptors?

While this work implicates GPA2 in the recognition of nitrogen starvation, we do not yet know the mechanism by which this occurs. Heterotrimeric G proteins typically bind directly to a transmembrane receptor; no such receptor has been identified that may sense nitrogen limitation. As a wide variety of structurally dissimilar compounds can satisfy cellular nitrogen requirements (reviewed in Magasanik, 1992), this signalling is potentially complex. The cell may sense the concentration of a common intracellular metabolite, in which case there may be a cytoplasmic binding protein that serves this role. Alternatively, the cell may have receptors for each potential nitrogen source, as it has multiple transmembrane permeases. Activity or inactivity of these permeases could transduce a signal through GPA2 to indicate the relative nitrogen abundance. Our recent findings support the latter model, as we find that an ammonia permease is required to promote pseudohyphal differentiation in response to limiting concentrations of ammonia (M.C.Lorenz and J.Heitman, manuscript in preparation). In carbon source metabolism, recent work also supports the model of permeases as receptors. Two glucose permease homologs participate in signalling glucose repression. The varying affinity of these permeases allows the cell to regulate the transcription of metabolic enzymes and other permeases based on the extracellular glucose concentration (Liang and Gaber, 1996; Ozcan et al., 1996).

Origins of the dual function of MAP kinase cascades

The extensive connections between mating, morphogenesis and nitrogen deprivation may provide an explanation for the dual role of the MAP kinase pathway in regulating both mating and filamentous growth in yeast. As S.cerevisiae exists in the environment primarily as diploids, this species may have evolved a cell type specificity to mating and differentiation, where responses to nutrient stress, such as sporulation and pseudohyphal differentiation, are confined to the more common diploid cell type. In contrast, in other fungi, including the fission yeast S.pombe and the basidiomycetous pathogen Cryptococcus neoformans, which are haploid in the environment, both pheromone and nitrogen starvation are required for mating. In these organisms, the two signalling pathways function coordinately in a single cell type to integrate nutrient responses with mating and conjugation. We hypothesize that these two signalling processes evolved to function coordinately in the evolutionary precursors of these diverse fungi and that this role has been maintained in some descendents (e.g. S.pombe and C.neoformans), but developed cell type specificity in others (e.g. S.cerevisiae).

Materials and methods

Yeast strains and media

Standard yeast media and microbiological techniques were used (Sherman, 1991). SLAD media contains 50 μM ammonium sulfate, 2% glucose and 2% Bacto‐agar (Gimeno et al., 1992). Our formulation of SLAD differs slightly from previous reports (Gimeno et al., 1992), and contains 0.17% yeast nitrogen base minus amino acids minus ammonium sulfate as opposed to the previously described 0.67% YNB. We have found no significant differences in pseudohyphal growth with these two formulations. Media used to induce the pGal‐GPA2 constructs (SLARG) contained 2% raffinose and 0.5% galactose.

Yeast strains are listed in Table V and are all derived from the Σ1278b background (obtained from M.Grenson). The Δgpa2::G418 mutant was created by the PCR mediated disruption technique of Wach et al. (1994), using the G418 resistance cassette from plasmid pFA6‐KanMX2 and bifunctional disruption primers (homology to the G418‐resistance cassette underlined): 5′‐TGTTACAGCACAAATCACGCGTATTTTCAAGCAAATATCATGGGTCAGCTGAAGCTTCGTACGC‐3′ and 5′‐GCATGCAGTTTTGTCTCTGTTTTAGCTGTGCATTCATGGTAACACGCAT‐ AGGCCACTAG TGGATCTG‐3′. Independently derived haploid strains (created in strains MLY40 and MLY41; see Table V) were mated to produce the homozygous diploid MLY132a/α. The Δpde2::G418 mutant strain was derived in the same manner to produce strain MLY162a/α using primers: 5′‐TTTGAGATCACTACTTAATTGAAGAAAACATAACCTATT GATCAGCTGAAGCTTCGTACGC‐3′ and 5′‐CAATGAATGGTACAAGAAATTTTGATATTCTTGCTATTGTG GGCATAGGCCACTAGTGGATCTG‐3′.

View this table:
Table 5. Yeast strains

The Δgpa2::G418 Δpde2:G418 double mutant strain MLY171 was created through a cross between MLY132α and MLY162a. Δpde2::G418 Δste strains were constructed by disrupting PDE2 in strains L3566 (wt, MATa/α), HLY351 (Δste7/Δste7) and HLY352 (Δste12/Δste12). Heterozygous Δpde2/PDE2 strains were sporulated and dissected; pseudohyphal growth was assayed in strains MLY173 (STE+ Δpde2 MATα), MLY174 (Δste7 Δpde2 MATa) and MLY175 (Δste12 Δpde2 MATa) by expressing a plasmid containing either the MATa or MATα locus; such pseudodiploid strains are competent to undergo pseudohyphal differentiation (Gimeno et al., 1992).

Strains used for the FG(TyA)::lacZ reporter assays were constructed to allow selection for the LEU2‐marked reporter plasmid pIL30‐LEU2. Haploid Δgpa2 strains MLY132a and MLY132α were made leu2 using the Δleu2::hisG disruption cassette from plasmid pNKY85 (Alani et al., 1987), then crossed to create MLY212a/α (see Table V). Strains lacking Δste20 and Δste12 were constructed through G418‐disruption in haploid strains MLY42 and MLY43. Oligonucleotides used to disrupt STE20 were 5′‐CACCCCATCCTAAATATCCCACAAGATCCTCGACTAATAC AAGAACAGCTGAAGCTTCGTACGC‐3′ and 5′‐GTACCCTGCTTGCTACGTTTACTTTTGTTTATCATCTTCAGT GCATAGGCCACT‐ AGTGGATCTG‐3′ and for STE12 were 5′‐ATAGCGGAACCGCTTTCTTTATTTGAATTGTCTTGTTCACCAA GGCAGCTGAAGCTTCGTACGC‐3′ and 5′‐AAAAATTATATTATATCAGGTTGCATCTGGAAGG‐ TTTTTATCGCATAGGCCACTAGTGGATCTG‐3′ (homology to G418 cassette underlined). After transformation with complementing plasmids to allow mating, these strains were crossed, forming MLY216a/α (Δste12::G418/Δste12::G418) and MLY219a/α (Δste20::G418/Δste20::G418). Strains MLY173 (Δpde2), MLY174 (Δste7 Δpde2) and MLY175 (Δste12 Δpde2) were converted to leu2 using pNKY85, resulting in MLY213, MLY214 and MLY215, respectively.


Yeast plasmids are listed in Table VI. The pGal‐GPA2 construct was created by PCR amplification of GPA2 from genomic DNA using primers: 5′‐TCCTCTAGAGCAAATATCATGGGTCTCTGC‐3′ and 5′‐CCCAAGCTTCTCAGAATGGTGCAAGTC‐3′. The resulting PCR product was cloned under the control of the GAL1,10 promoter in the XbaI–HindIII sites of pSEYC68 (CEN URA3). The pGal‐GPA2Gly132Val and pGal‐GPA2Gly299Ala alleles were created through PCR‐mediated overlap site‐directed mutagenesis (Ho et al., 1989) using outer primers (above) and mutagenic primers (mutation in bold), for Gly132Val: 5′‐CTACTGCTGGGTGCCGTTGAAAGTGGTAAGTCC‐3′ and 5′‐GGACTTACCACTTTCAACGGCACCCAGCAGTAG‐3′, and for Gly299Ala: 5′‐ATATACGACGTGGGTGCACAGCGTTCCGAAAGA‐3′ and 5′‐TCTTTCGGAACGCTGTGCACCCACGTCGTATAT‐3′.

View this table:
Table 6. Plasmids

Assay for FG(TyA)::lacZ activity

Plasmid pIL30‐LEU2, encoding the lacZ gene under the transcriptional control of the filamentation response element from Ty1 (Laloux et al., 1994; Mösch et al., 1996), was used for the reporter assays. To examine the effects of various GPA2 alleles, pIL30‐LEU2 was coexpressed in wild‐type (MLY97), Δste20/Δste20 (MLY219a/α) or Δste12/Δste12 (MLY216a/α) strains with control vector (pSEYC68), GPA2wt (pML180), GPA2Gly132Val (pML160) or GPA2Gly299Ala (pML179). To examine the effects of deletion of GPA2, pIL30‐LEU2 was coexpressed in wild‐type (MLY97) or Δgpa2/Δgpa2 (MLY212a/α) strains with control vector (Yeplac195), RAS2wt (pMW1), RAS2Gly19Val (pMW2), STE11‐4 (pSL1509) or PHD1 (pCG38). For the cAMP experiment, pIL30‐LEU2 was coexpressed with plasmids bearing the appropriate MAT locus in haploid strains MLY213 (Δpde2), MLY214 (Δste12 Δpde2) and MLY215 (Δste7 Δpde2). The assays were performed essentially as described (Mösch et al., 1996). Strains were grown in liquid selective media to mid‐logarithmic phase, collected by centrifugation and washed with water. The cells were plated to media containing 50 μM or 5 mM ammonium sulfate and 2% glucose (0.5% galactose and 2% raffinose were used when necessary to induce expression of the GPA2 alleles). After 48 h at 30°C, the cells were pooled and split; standard β‐galactosidase assays were performed on half of the pool, using chlorophenolred‐β‐d‐galactopyranoside (CPRG; Calbiochem) as substrate. The other half of the pool was disrupted by agitation with glass beads in lysis buffer (100 mM Tris–Cl, pH 8.0, 20% glycerol, 1 mM DTT, 100 U/ml aprotinin and 0.5 μM PMSF). Protein concentrations from the clarified extracts were determined using a Bradford assay (Bio‐Rad), with BSA as a standard. β‐galactosidase activities were normalized to protein concentration.


Representative colonies were photographed directly on agar plates using a Nikon Axiophot‐2 microscope with a 10× objective and 2.5× trinocular camera adaptor, except in Figure 2B and Figure 4B where a 20× objective and 2.5× trinocular adaptor were used.

Identification of Gβ and Gγ homologues

To find potential β subunits, BLAST searches of the yeast database were performed using S.cerevisiae STE4 (accession number Z75120; Whiteway et al., 1989) and S.pombe Gpb1 (L28061; Kim et al., 1996) sequences. Candidate genes, listed below (Table VII), were disrupted through the G418/PCR disruption protocol described above in diploid strain MLY61. Gγ candidates were identified based on small size (less than 150 amino acids) and a C‐terminal lipid modification motif (CAAX box) through the Yeast Protein Database (http://www.ypdhome.html).

View this table:
Table 7.

Note added in Proof

We have recently found that the Δgpa2 and Δras2 mutations exhibit a near synthetic lethal growth defect that is suppressed by a Δpde2 mutation, which increases intracellular cAMP levels (data not shown). Similar observations have recently been reported by Y.Xue and J.Hirsch (personal communication, manuscript submitted) and by Kubler et al., who also reported that GPA2 is required for pseudohyphal differentiation [ ].


We would like to thank Gerry Fink, Steve Garrett, Mary Ward and Marcelle Grenson for generously providing strains and plasmids. We also thank Steve Garrett, Pat Casey and Kara Dolinski for comments on the manuscript, members of the Heitman lab for helpful discussions and advice, and Scott Muir for technical assistance. Joseph Heitman is an assistant investigator of the Howard Hughes Medical Institute.


View Abstract