The MEP2 ammonium permease regulates pseudohyphal differentiation in Saccharomyces cerevisiae

Michael C. Lorenz, Joseph Heitman

Author Affiliations

  1. Michael C. Lorenz1 and
  2. Joseph Heitman*,2
  1. 1 Department of Genetics and the Howard Hughes Medical Institute, Duke University Medical Center, 322 CARL Building, Research Drive, Durham, NC, 27710, USA
  2. 2 Department of Pharmacology and Cancer Biology and the Howard Hughes Medical Institute, Duke University Medical Center, 322 CARL Building, Research Drive, Durham, NC, 27710, USA
  1. *Corresponding author. E-mail: heitm001{at}


In response to nitrogen starvation, diploid cells of the budding yeast Saccharomyces cerevisiae differentiate into a filamentous, pseudohyphal growth form. This dimorphic transition is regulated by the Gα protein GPA2, by RAS2, and by elements of the pheromone‐responsive MAP kinase cascade, yet the mechanisms by which nitrogen starvation is sensed remain unclear. We have found that MEP2, a high affinity ammonium permease, is required for pseudohyphal differentiation in response to ammonium limitation. In contrast, MEP1 and MEP3, which are lower affinity ammonium permeases, are not required for filamentous growth. Δmep2 mutant strains had no defects in growth rates or ammonium uptake, even at limiting ammonium concentrations. The pseudohyphal defect of Δmep2/Δmep2 strains was suppressed by dominant active GPA2 or RAS2 mutations and by addition of exogenous cAMP, but was not suppressed by activated alleles of the MAP kinase pathway. Analysis of MEP1/MEP2 hybrid proteins identified a small intracellular loop of MEP2 involved in the pseudohyphal regulatory function. In addition, mutations in GLN3, URE2 and NPR1, which abrogate MEP2 expression or stability, also conferred pseudohyphal growth defects. We propose that MEP2 is an ammonium sensor, generating a signal to regulate filamentous growth in response to ammonium starvation.


Many fungi interconvert between multiple growth patterns, including yeast and filamentous forms, depending on environmental conditions. Filamentous growth may enable these non‐motile organisms to seek preferable environments. For example, conjugation of compatible cell types in the maize pathogen Ustilago maydis results in the formation of a filamentous heterokaryon, the cell type responsible for host infection (Banuett, 1991; Hartmann et al., 1996). Polymorphism between yeast, hyphal and pseudohyphal forms in the opportunistic human pathogen Candida albicans has been proposed to be a key means of tissue invasion and dissemination during infection, and non‐filamentous C.albicans mutants are avirulent in a mouse model (Lo et al., 1997). Similarly, in the human pathogen Cryptococcus neoformans, a filamentous growth form accompanied by spore formation (haploid fruiting) has only been observed in cells of the α mating type, which are more virulent than the a mating type (Kwon‐Chung et al., 1992; Wickes et al., 1996).

In the budding yeast Saccharomyces cerevisiae, severe nitrogen starvation induces diploid cells to differentiate into a filamentous, pseudohyphal growth form (Gimeno et al., 1992). This developmental pathway has been proposed to be a scavenging mechanism under nutrient limiting conditions (Gimeno et al., 1992). Similar to the filamentous states of other fungi, pseudohyphal cells are elongated and have an altered budding pattern and cell cycle compared with yeast form cells (Gimeno et al., 1992; Kron et al., 1994). In addition, pseudohyphal cells invade the growth substrate.

Pseudohyphal differentiation is regulated by GPA2, the α subunit of a guanine nucleotide binding protein (Kübler et al., 1997; Lorenz and Heitman, 1997). Diploid Δgpa2/Δgpa2 strains have a severe defect in filamentation, whereas a dominant active GPA2 allele stimulates pseudohyphal differentiation, even under conditions of nitrogen excess. GPA2 has been proposed to be a component of the nitrogen sensor that regulates this dimorphic switch (Lorenz and Heitman, 1997). G proteins play similar roles in other fungi, including U.maydis in which the Gα protein Gpa3 is required for mating, a necessary precursor to filamentous growth (Regenfelder et al., 1996), and in C.neoformans, in which the Gα protein Gpa1 regulates both mating and virulence (Tolkacheva et al., 1994; Alspaugh et al., 1997).

Several observations suggest that GPA2 and the small G protein RAS2 coordinately regulate vegetative and pseudohyphal growth. Mutational activation of either GPA2 or RAS2 stimulates filamentous growth (Gimeno et al., 1992; Lorenz and Heitman, 1997), both RAS2 and GPA2 modulate cAMP levels (Toda et al., 1985; Field et al., 1988; Nakafuku et al., 1988), and Δras2 and Δgpa2 mutations exhibit a synthetic growth defect (Kübler et al., 1997; Lorenz and Heitman, 1997; Y.Xue and J.Hirsch, personal communication). cAMP promotes filamentous growth, even under nutrient rich conditions, and increased cAMP levels suppress both the Δgpa2 pseudohyphal defect (Lorenz and Heitman, 1997) and the Δgpa2 Δras2 vegetative growth defect (Kübler et al., 1997). cAMP also regulates mating and dimorphism in other fungi, including U.maydis, C.albicans, Neurospora crassa, and Schizosaccharomyces pombe (Niimi et al., 1980; Maeda et al., 1990; Sabie and Gadd, 1992; Yarden et al., 1992; Gold et al., 1994; Kronstad, 1997).

Activation of a signaling pathway that is independent of the G protein/cAMP pathway and includes elements of the haploid pheromone response MAP kinase cascade is also required for pseudohyphal differentiation. Mutations in the STE20, STE11 and STE7 protein kinases and the STE12 transcription factor block filamentation in diploid cells (Liu et al., 1993). In response to nitrogen starvation the MAP kinase cascade stimulates the transcription of a reporter gene under the control of regulatory sequences from the transposon Ty1 (Laloux et al., 1994; Mösch et al., 1996). These sequences include a binding site for a STE12/TEC1 heterodimer (Madhani and Fink, 1997). In addition to the role of RAS2 in regulating cAMP synthesis, it has been suggested that the dominant active RAS2Val19 allele may also activate the MAP kinase cascade (Mösch et al., 1996).

Although pseudohyphal differentiation is induced by nitrogen starvation, the mechanisms by which this environmental signal is sensed are not known. GPA2 is involved, but potential GPA2‐linked receptors have not yet been reported. Nutrient‐based regulation has been extensively studied with respect to transcriptional and cell cycle control (reviewed in Grenson, 1992; Magasanik, 1992), yet little is known about the direct sensing of nutrient availability. This signaling is likely to be complex as yeast, like most microorganisms, can utilize a wide variety of compounds to satisfy nutritional requirements. Several recent studies have reported that transmembrane permeases have a role as receptors in such signaling pathways. In yeast, the glucose transporter homologs RGT2 and SNF3 are required for transcriptional induction of other glucose transporters based on glucose availability (Liang and Gaber, 1996; Ozcan et al., 1996), and dominant mutations in both RGT2 and SNF3 have been identified which signal in the complete absence of glucose (Ozcan et al., 1996). Despite the homology of RGT2 and SNF3 to glucose permeases, neither transports glucose (Liang and Gaber, 1996). A related glucose transporter homolog, Rco‐3, regulates conidiation in N.crassa in response to changes in sugar availability (Madi et al., 1997).

Transport through plasma membrane permeases is an early step in the metabolism of any nutrient and thus nutrient permeases are in a unique position to both sense and import their substrates. Yeast has specific, high‐affinity plasma membrane permeases for numerous nutrients (reviewed in Andre, 1995). Standard media for induction of pseudohyphal differentiation contains low concentrations (50 μM) of ammonium as the sole nitrogen source (Gimeno et al., 1992). We have therefore examined the role of the ammonium permeases, MEP1, MEP2 and MEP3 in the regulation of dimorphism in S.cerevisiae. We find that the high affinity ammonium permease MEP2 is required for pseudohyphal differentiation under standard conditions, whereas the homologous permeases MEP1 and MEP3 are not. Strains lacking MEP2 have no apparent defects in ammonium uptake, metabolism, or growth, even under the low ammonium concentrations that promote filamentation. The pseudohyphal defect of Δmep2/Δmep2 mutant strains is not suppressed by activation of the MAP kinase cascade, but is suppressed by activated alleles of GPA2 or RAS2, or by exogenous cAMP. We propose that MEP2 serves as an ammonium sensor to regulate pseudohyphal growth and functions in a signaling pathway upstream of GPA2, RAS2 and cAMP.


The MEP2 ammonium transporter is required for pseudohyphal differentiation

Previous studies had identified two ammonium‐specific permeases, MEP1 and MEP2 (Dubois and Grenson, 1979; Marini et al., 1994). A third permease, MEP3, was identified through BLAST searches during the course of this work and independently by Marini et al. (1997). These proteins share significant sequence identity: MEP1 and MEP3 share 80% identity; MEP2 is less similar, 41% identical with MEP1 and 39% identical with MEP3. MEP2 is the highest affinity permease, with a Km for ammonium of 1–2 μM whereas MEP1 (Km ∼5–10 μM) and MEP3 (Km ∼1.4–2.1 mM) are lower affinity ammonium transporters (Marini et al., 1997). Deletion of all three MEP permeases renders a cell inviable on media containing <5 mM ammonium as the sole nitrogen source (Marini et al., 1997; see Figure 1A and B), indicating that these proteins are the only specific ammonium permeases in yeast.

Figure 1.

MEP2 is required for pseudohyphal differentiation. (A) Homozygous diploid strains of the indicated genotypes (see Materials and methods; Table II) were incubated on low‐ammonium SLAD medium for 4 days at 30°C. (B) The diploid Δmep1/Δmep1 Δmep2/Δmep2 Δmep3/Δmep3 strain (MLY131a/α) with plasmids expressing no MEP gene, MEP1 (pML100), MEP2 (pML151) or MEP3 (pML113) were incubated on SLAD medium for 4 days at 30°C. Isogenic wild‐type (MLY61) and Δmep2/Δmep2 strains (MLY108a/α) are included as controls.

To test whether any of these permeases affect pseudohyphal differentiation, we constructed a series of homozygous diploid strains that each lack a single permease. Our initial hypothesis was that mutations in these permeases might impair ammonium uptake, thus increasing nitrogen starvation and possibly enhancing pseudohyphal differentiation. In contrast, we observed that strains lacking MEP2 have a severe defect in filamentous growth under limiting ammonium conditions (Figure 1A). Deletions of MEP1 or MEP3, however, had no effect on pseudohyphal growth (Figure 1A). As previously reported (Marini et al., 1997), strains lacking MEP2 had no apparent growth defect, even on low ammonium media. Δmep2/Δmep2 mutant strains do form filaments on media containing limiting concentrations of glutamine, proline, asparagine or arginine as the sole nitrogen source (data not shown), indicating that the pseudohyphal defect conferred by the Δmep2 mutation is only observed in the presence of its substrate, ammonium.

Next we constructed strains lacking multiple MEP permeases. Both the Δmep1 Δmep2 and Δmep2 Δmep3 homozygous diploid strains exhibited the pseudohyphal defect conferred by the Δmep2 mutation (Figure 1A). The Δmep1 Δmep2 strain also had a growth defect when grown under limiting ammonium conditions, consistent with previous studies (Figure 1A; Dubois and Grenson, 1979; Marini et al., 1994). Δmep1/Δmep1 Δmep2/Δmep2 cells apparently experience nitrogen starvation severe enough to result in a growth defect, yet still do not undergo filamentous growth. These findings suggest that MEP2 might have a role in regulating pseudohyphal differentiation in addition to its role in ammonium uptake.

When incubated on synthetic media containing 5 μM to 5 mM ammonium as the sole nitrogen source, the triple Δmep1 Δmep2 Δmep3 mutant strain completes a few cell divisions before arresting (Figure 1A), possibly when vacuolar stores of amino acids are exhausted. On YNB medium (38 mM ammonium), the triple mutant does grow, albeit with a significant growth defect (data not shown), suggesting that an additional non‐specific ammonium transport system exists. As shown in Figure 1B, expression of either MEP1, MEP2 or MEP3 from a plasmid is sufficient to complement the growth defect of the homozygous diploid Δmep1 Δmep2 Δmep3 mutant strain, demonstrating that all three proteins are functional ammonium transporters. In contrast, only expression of MEP2 also restored pseudohyphal differentiation in the diploid Δmep1 Δmep2 Δmep3 strain, indicating that MEP2 has a specific role in the regulation of filamentous growth that is not shared with the homologous permeases MEP1 and MEP3.

The Δmep2 mutation does not affect ammonium utilization

Although the Δmep2 mutation did not appear to have phenotypes indicative of defects in nitrogen metabolism, we wished to exclude the possibility that alterations in ammonium metabolism in Δmep2 mutant strains might contribute to the defect in pseudohyphal growth. Isogenic MEP2/MEP2 and Δmep2/Δmep2 strains grew at similar rates, even under low ammonium conditions, as determined from growth curves in liquid media (data not shown). We next directly assayed the ability of diploid Δmep strains to extract ammonium from the extracellular medium, a measure of the ammonium uptake capacity of these strains (see Materials and methods). As shown in Figure 2, only the Δmep1 Δmep2 and Δmep1 Δmep2 Δmep3 homozygous diploid strains exhibited reduced ammonium uptake capacity. These findings are consistent with the reduced growth rates of these strains on ammonium‐limiting media (Figure 1A and data not shown) and with a previous report (Marini et al., 1997). We also assayed the activity of the nitrogen metabolic enzymes glutamine synthase (GLN1) and glutamate dehydrogenase (GDH1 and GDH2) after growth at a range of ammonium concentrations from 50 μM to 50 mM. Activities of these enzymes were similar in wild‐type and Δmep2/Δmep2 strains (data not shown), indicating that neither ammonium uptake nor metabolism are significantly altered in cells lacking the MEP2 ammonium permease.

Figure 2.

Ammonium import in ammonium permease mutants. Ammonium uptake capacity was assayed as described in Materials and methods in homozygous diploid strains: wild‐type (MLY61), Δmep1 (MLY104a/α), Δmep2 (MLY108a/α), Δmep3 (MLY128a/α), Δmep1 Δmep2 (MLY115a/α) and Δmep1 Δmep2 Δmep3 (MLY131a/α). The ammonium remaining in the media is expressed as a percentage of the starting concentration (∼500 μM).

MEP2 regulation of pseudohyphal growth is independent of the MAP kinase cascade

The findings that the Δmep2 mutation does not alter ammonium uptake or metabolism yet blocks pseudohyphal differentiation led to the hypothesis that MEP2 might have a signaling function in the regulation of pseudohyphal differentiation. One previously characterized element of the signaling machinery is the pheromone responsive MAP kinase cascade. Expression of the dominant active STE11‐4 allele suppresses mutations in upstream components (e.g. Δste20) to restore either mating response or pseudohyphal growth, while overexpression of the STE12 transcription factor greatly enhances filamentation (Liu et al., 1993). These alleles allowed us to test the point of action of MEP2 with respect to the MAP kinase cascade through genetic epistasis. The pseudohyphal defect of the Δmep2/Δmep2 strain was not suppressed by the dominant active STE11‐4 allele or by overproduction of the STE12 transcription factor (Figure 3), suggesting that MEP2 does not function upstream of these proteins. Likewise, the Δmep2 mutation was not suppressed by overexpression of PHD1, a DNA binding protein that enhances pseudohyphal differentiation in both wild‐type strains (see Figure 3; Gimeno and Fink, 1994) and Δste mutant strains. The pseudohyphal deficiency conferred by the Δmep2 mutation was modestly suppressed by high copy expression of TEC1 (Figure 3), which encodes a transcription factor that dimerizes with STE12 to modulate expression of a nitrogen‐regulated reporter gene [FG(TyA)::lacZ; Laloux et al., 1994; Madhani and Fink, 1997]. This disparity between the STE12 and TEC1 epistasis results may result from lower expression of STE12, as high level overproduction of STE12 is lethal (Liu et al., 1993). To moderate STE12 expression we used a low concentration of glucose to reduce the activity of the galactose‐inducible promoter (Liu et al., 1993; see Materials and methods). Alternatively, TEC1 may have a STE12‐independent function also important in regulation of filamentation, or TEC1 may be regulated by a signaling pathway other than the MAP kinase cascade that activates STE12.

Figure 3.

MEP2 does not interact with the MAP kinase cascade. Epistasis tests with elements of the MAP kinase cascade were performed in homozygous wild‐type (MLY61) and Δmep2/Δmep2 (MLY108a/α) diploid strains. Each strain expresses a control vector, the dominant STE11‐4 allele (pSL1509), or high copy plasmids with TEC1 (p2.5‐2) or PHD1 (pCG38) grown on SLAD medium for 4 days at 30°C. Expression of the pGal‐STE12 allele (pNC252) on inducing SLADG media was assayed after 6 days at 30°C.

To examine the relationship between MEP2 and the MAP kinase pathway further, we employed the FG(TyA):: lacZ reporter known to be induced 2‐ to 10‐fold by nitrogen starvation (Mösch et al., 1996; Lorenz and Heitman, 1997; Madhani and Fink, 1997). Nitrogen starvation induced reporter gene expression to a similar extent in wild‐type, Δmep1, Δmep2 or Δmep1 Δmep2 homozygous diploid strains (Table I), thus MEP1 and MEP2 are not required for induction of the FG(TyA)::lacZ reporter. While this finding is in contrast to mutations in the MAP kinase cascade, which reduce induction of this reporter (Mösch et al., 1996; Madhani and Fink, 1997), expression of the FG(TyA)::lacZ gene is also unaffected by either loss of function or activated alleles of GPA2 (Lorenz and Heitman, 1997).

View this table:
Table 1. Δmep mutations do not alter FG(TyA)::lacZ induction by nitrogen starvation

The Δmep2 pseudohyphal defect is suppressed by activation of GPA2 or RAS2, or by exogenous cAMP

The findings presented above indicate that MEP2 regulates filamentous growth independent of the MAP kinase cascade. We recently identified a signaling pathway including GPA2, RAS2 and cAMP that also regulates pseudohyphal differentiation independent of this MAP kinase pathway (Lorenz and Heitman, 1997). We tested the relationship between MEP2 and GPA2 using a dominant active allele, GPA2‐2 (Gly132Val), which stimulates pseudohyphal differentiation, even in nutrient rich conditions, and suppresses the filamentation defect of Δste mutant strains (Lorenz and Heitman, 1997). Expression of GPA2‐2 restored filamentation in Δmep2/Δmep2 mutant strains (Figure 4), suggesting that MEP2 functions upstream of GPA2 in signaling pseudohyphal growth.

Figure 4.

Activated GPA2 suppresses the Δmep2/Δmep2 pseudohyphal defect. Wild‐type (MLY61) and Δmep2/Δmep2 (MLY108a/α) strains expressing a control vector (pSEYC68), wild‐type GPA2 (pML180), or the dominant, constitutively active GPA2‐2 allele (pML160) under the control of a galactose‐inducible promoter were incubated on SLARG medium for 4 days at 30°C.

If MEP2 acts in the same pathway as GPA2, conditions that restore filamentation in a Δgpa2/Δgpa2 strain should also suppress the pseudohyphal defect of Δmep2/Δmep2 strains. Indeed, one suppressor of Δgpa2, the dominant RAS2Gly19Val allele, also suppressed the Δmep2 mutation (Figure 5A). In strains lacking the high‐affinity cAMP phosphodiesterase PDE2, cAMP enhances filamentation and suppresses the Δgpa2 mutation (Lorenz and Heitman, 1997). Similarly, cAMP restored pseudohyphal differentiation in a Δmep2/Δmep2 Δpde2/Δpde2 strain (Figure 5B). These findings support a model in which GPA2, RAS2 and cAMP function downstream of MEP2 in a signaling pathway regulating filamentous growth.

Figure 5.

The Δmep2 pseudohyphal defect is suppressed by RAS2Val19 and by cAMP. (A) Wild‐type (MLY61) and Δmep2/Δmep2 (MLY108a/α) strains with a control plasmid, or expressing RAS2wt (pMW1) or RAS2Val19 (pMW2) were grown on SLAD medium for 4 days at 30°C. (B) Δpde2/Δpde2 (MLY162a/α) and Δmep2/Δmep2 Δpde2/Δpde2 (MLY229a/α) strains were grown on low ammonium SLAD media in the absence of cAMP or in the presence of 1 mM or 10 mM cAMP for 4 days at 30°C.

An intracellular loop is important for MEP function in pseudohyphal growth

To address the unique structural features of MEP2 required for pseudohyphal differentiation, we took advantage of the significant similarities between MEP1 and MEP2 in sequence (66% similarity), topology (10 predicted transmembrane domains), and function (as ammonium permeases), though MEP2 regulates filamentous growth while MEP1 does not. Deletion of the C‐terminus of MEP2, 50 amino acids in length and predicted to be cytoplasmic, had no effect on either the pseudohyphal growth or ammonium transport functions of MEP2 (Figure 6; pML153). We constructed a series of MEP1–MEP2 hybrids, with junctions in the predicted extracellular loops and expressed from the MEP2 promoter (Figure 6). Each of these hybrid permeases complemented the growth defect of the Δmep1 Δmep2 Δmep3 homozygous diploid strain on low ammonium media (50 μM), consistent with restoration of ammonium transport. A hybrid in which only the sequence at the N‐terminus is derived from MEP2 (through the first transmembrane domain; pML155) did not complement the pseudohyphal defect of Δmep2/Δmep2 mutant strains (Figure 6). This hybrid, composed almost entirely of MEP1 coding sequences, uses the MEP2 promoter; thus differences in expression are unlikely to account for the functional differences between MEP1 and MEP2. Expression of a hybrid protein including the three N‐terminal transmembrane domains of MEP2 fused to C‐terminal sequences of MEP1 (pML156) restored pseudohyphal differentiation in Δmep2/Δmep2 strains (Figure 6). These observations suggest that the region of MEP2 between the first and third transmembrane domains (e.g. the first intracellular loop) participates in the pseudohyphal regulatory function.

Figure 6.

MEP1/MEP2 hybrid analysis. The hybrids diagrammed above (MEP1 in grey, MEP2 in black) expressed from the high copy YEplac195 plasmid, each under control of the MEP2 promoter, were assayed for growth and filamentation in the Δmep1/Δmep1 Δmep2/Δmep2 Δmep3/Δmep3 strain (MLY131a/α). ‘+’ indicates complementation of either the pseudohyphal growth or ammonium uptake defects. Hybrids were constructed as described in the Materials and methods.

Next we constructed a fusion protein in which only the first intracellular loop plus the adjacent membrane spanning segments (87 amino acids) was derived from MEP2 (pML162). This hybrid protein complemented to restore both the pseudohyphal and ammonium transport functions of MEP2. Surprisingly, the reciprocal swap (pML161), in which the majority of the protein is MEP2 and only the first intracellular loop is derived from MEP1, also complemented both MEP2 functions. Thus the first intracellular loop is sufficient, but not strictly necessary, for the signaling function of MEP2; hence other regions of MEP2 are likely to also participate in signaling.

Regulation of MEP2 protein expression: nitrogen regulatory genes are required for filamentous growth

Next we addressed whether regulation of MEP2 expression could underlie a signaling function of this ammonium permease. Northern analysis demonstrated that MEP2 is preferentially expressed under nitrogen limiting conditions (Marini et al., 1997). Because many nutrient permeases are regulated both post‐translationally and transcriptionally (reviewed in Grenson, 1992), we analyzed MEP2 protein levels. To detect the MEP2 protein directly we used an integrative technique to tag the genomic MEP2 locus such that three repeats of the hemagluttinin epitope (HA) are fused to the C‐terminus (Schneider et al., 1995). This approach utilizes the native MEP2 promoter and terminator sequences. The epitope‐tagged protein complemented both the pseudohyphal and ammonium transport functions of the native MEP2 (data not shown).

The MEP2::HA protein was detected in cell lysates through Western blot analysis as shown in Figure 7A. MEP2 is present at very low levels in cells grown in rich medium (SD‐Ura). In media in which ammonium is the sole nitrogen source, MEP2 was expressed at high levels, and expression did not increase at lower ammonium concentrations (Figure 7A), in contrast with MEP2 mRNA which does accumulate at lower ammonium concentrations (Marini et al., 1997). Since MEP2 expression is not significantly different between high and low ammonium conditions, regulation of MEP2 protein levels does not explain its unique role in pseudohyphal differentiation.

Figure 7.

MEP2 protein expression requires GLN3, URE2, and NPR1. (A) Western blot analysis of the epitope tagged MEP2‐HA protein. Otherwise wild‐type MEP2::HA/MEP2::HA strain (MLY220), or Δgln3/Δgln3 (MLY194a/α), Δure2/Δure2 (MLY198a/α), and Δnpr1/Δnpr1 (MLY221) mutant strains were grown in minimal media containing the indicated concentration of ammonium sulfate for 4 h at 30°C. Proteins were detected using an α‐HA monoclonal antibody or α‐cyclophilin A (α‐CPR1) as a control for protein loading. (B) Homozygous diploid strains wild‐type (MLY61), Δgln3/Δgln3 (MLY139a/α), Δure2/Δure2 (MLY140a/α), and Δnpr1/Δnpr1 (MLY54a/α) were incubated on low ammonium media (SLAD) for 4 days at 30°C.

Marini et al. (1997) found that the nitrogen regulatory genes GLN3 and NIL1 are required for the transcription of MEP2. We found that MEP2 protein was markedly reduced in strains that lack GLN3 (Figure 7A). GLN3, a zinc‐finger DNA binding protein, is required for the transcriptional induction of a number of nitrogen catabolic genes, probably including MEP1 and MEP2 (Minehart and Magasanik, 1991; Stanbrough et al., 1995). URE2, a prion analog with homology to glutathione S‐transferases and which antagonizes GLN3 function in some cases (Coschigano and Magasanik, 1991; Wickner, 1994; Xu et al., 1995), was also required for MEP2 accumulation (Figure 7A). Finally, the MEP2 protein was also undetectable in Δnpr1/Δnpr1 mutant strains (Figure 7A). NPR1 is a protein kinase that maintains the activity of a variety of plasma membrane permeases, including MEP1 and MEP2, under conditions of nitrogen starvation (Grenson and Acheroy, 1982; Grenson and Dubois, 1982; Grenson, 1983b; Vandenbol et al., 1990).

One prediction from these findings is that the GLN3, URE2 and NPR1 regulatory proteins might also regulate pseudohyphal differentiation. This is indeed the case, as Δgln3/Δgln3, Δure2/Δure2 and Δnpr1/Δnpr1 mutant strains were all defective in pseudohyphal differentiation (Figure 7B). The filamentation defect of Δgln3/Δgln3, Δure2/Δure2 and Δnpr1/Δnpr1 strains was also observed when glutamine or proline were present as the sole nitrogen source (data not shown), in contrast to Δmep2/Δmep2 strains in which the pseudohyphal deficiency is limited to ammonium‐grown cells. When under the control of a heterologous promoter (the inducible Gal1,10 promoter), MEP2 complements the low ammonium growth defect of the mep triple mutant, indicating that MEP2 is expressed, but did not restore filamentation in Δgln3/Δgln3 or Δure2/Δure2 mutant strains. Thus these regulatory proteins are likely to have targets in addition to MEP2 that are critical for the regulation of dimorphism.

Other permeases may also regulate pseudohyphal differentiation

Next we addressed whether our findings with MEP2 could be extended to other nitrogen permeases. Since MEP2 does not alter pseudohyphal differentiation when nitrogen sources other than ammonium are present, we hypothesized that other permeases may play a similar role to MEP2 when in the presence of their substrates. Unfortunately, most other nitrogen sources are imported by a single specific uptake system, in addition to non‐specific systems such as the general amino acid permease. Analysis of other permeases, then, is more difficult than for the MEPs as the presence of multiple ammonium permeases made this system ideal for this type of analysis. A genetic screen was designed to address this issue. The NPR1 protein kinase, required for pseudohyphal growth and stable MEP2 expression, post‐translationally regulates many permeases, including those for glutamine (GNP1), proline (PUT4), general amino acids (GAP1) and several others (Grenson and Dubois, 1982; Grenson, 1983b; Vandenbol et al., 1987). These permeases are significantly less active in Δnpr1 mutant strains than in wild‐type strains; GAP1 activity, for example, is undetectable in the absence of NPR1 (Grenson, 1983b). Genetic evidence suggests that NPR1 antagonizes the activity of the essential NPI1/RSP5 protein–ubiquitin ligase (Grenson, 1983b; Hein et al., 1995; Huibregste et al., 1995). We capitalized on this phenotype to examine the role of other permeases in the regulation of dimorphism.

We isolated suppressors of the Δnpr1 growth defect, reasoning that some mutations might also suppress the Δnpr1 filamentation defect. UV‐induced mutants with improved growth on either ammonium‐limiting SLAD or tryptophan–citrulline (TC) media were isolated. Of 36 mutants identified (in seven allele groups), one mutant exhibited constitutively derepressed GAP1 activity, as measured by sensitivity to toxic levels of d‐histidine and l‐lysine under repressing (high ammonium) growth conditions, the behavior expected of an npi1/rsp5 mutant (Grenson, 1983b). This phenotype was complemented by expression of the wild‐type NPI1 from a low copy plasmid; moreover, the mutation is allelic with NPI1 based on tight linkage of the mutant phenotype to an NPI1‐URA3 tagged allele integrated at the mutant npi1 locus (see Materials and methods). The Δnpr1 growth and pseudohyphal defects were strongly suppressed by the npi1 mutation on glutamine and weakly on proline, but were not suppressed on media containing ammonium as the sole nitrogen source (Figure 8). Genetic studies have indicated that NPI1 targets the glutamine (GNP1) and proline (PUT4) permeases for degradation, but does not affect the MEP ammonium permeases (Dubois and Grenson, 1979; Grenson and Dubois, 1982; Grenson, 1983a,b). The only known targets for both NPR1 and NPI1 action are plasma membrane permeases; thus, this finding suggests that other permeases may play a role analogous to MEP2.

Figure 8.

Other nitrogen permeases may regulate pseudohyphal differentiation. Homozygous diploid wild‐type (MLY61), Δnpr1/Δnpr1 (MLY54a/α), Δnpr1/Δnpr1 npi1/npi1 (MLY92a/α), and npi1/npi1 (MLY96) strains were incubated on media containing 100 μM ammonium, glutamine or proline as the sole nitrogen sources, for 4 days at 30°C.


The downstream events in nutrient based signaling have been well studied—regulatory events at the transcriptional, translational, and post‐translational levels ensure the presence of the appropriate enzymatic systems to utilize the available nutrients (reviewed in Grenson, 1992; Magasanik, 1992). In contrast, the initial events that sense the quantity and quality of these nutrients are poorly understood. We have studied these processes during pseudohyphal differentiation, a nitrogen‐regulated developmental transition in budding yeast.

We find that the ammonium permease MEP2 is required for filamentous growth in yeast. These observations suggest MEP2 may function both to transport and to sense ammonium levels. First, MEP2 is one of three related ammonium permeases, any one of which are sufficient for growth on ammonium‐limiting media, yet MEP2 is the only permease required for pseudohyphal differentiation. The role of MEP2 in pseudohyphal growth exhibits signal specificity; the filamentation defect is observed only when ammonium is present as the sole nitrogen source in limiting concentrations. Finally, ammonium import and nitrogen metabolic pathways (as measured by the activity of GLN1, GDH1 and GDH2 in ammonium limiting media) are normal in Δmep2/Δmep2 mutant cells.

Based on these observations, we propose that MEP2 serves as both an ammonium transporter and as a component of a nitrogen sensor that signals when conditions are appropriate for pseudohyphal growth.

Our findings are consistent with a model (Figure 9) in which MEP2 functions upstream of the Gα protein GPA2 in the regulation of pseudohyphal growth. Mutational activation of GPA2 (the dominant GPA2‐2 allele) suppresses the Δmep2/Δmep2 pseudohyphal defect. Moreover, both RAS2Val19 and exogenous cAMP suppress the filamentation defect conferred by both the Δgpa2 and Δmep2 mutations. Importantly, this signaling pathway is independent of the pheromone responsive MAP kinase cascade, which also regulates pseudohyphal differentiation (see Figure 9).

Figure 9.

A model for MEP2 regulation of pseudohyphal differentiation. Ammonium starvation is sensed via MEP2 to produce a signal that activates GPA2 and a signaling pathway that regulates filamentous growth independently of the MAP kinase cascade.

While the model presented in Figure 9 is consistent with our data, there are other possible explanations as well. We have no evidence directly linking MEP2 and GPA2; thus MEP2 may function in a signaling pathway separate from either the GPA2/cAMP or MAP kinase pathways. A critical reagent to test this hypothesis would be a dominant MEP2 allele; we have been unable to isolate such an allele to date. Alternative models are that MEP2 might be required to secrete ammonium or a related compound with an extracellular signaling function, or to sequester ammonium in an intracellular compartment. There is some very recent evidence that ammonia secretion may facilitate intercolony communication in yeast (Palkova et al., 1997).

The suggestion that a permease such as MEP2 may function as a receptor is not unprecedented. Recent findings indicate a similar role for the glucose transporter homologs RGT2 and SNF3 in yeast. These proteins activate a signaling cascade to regulate transcription of additional hexose permeases and a similar role has recently been ascribed to the Rco‐3 glucose transporter during conidiation in N.crassa (Liang and Gaber, 1996; Ozcan et al., 1996; Madi et al., 1997). Dominant mutations in both RGT2 and SNF3 have been described that constitutively activate this signaling cascade, even in the complete absence of glucose (Liang and Gaber, 1996; Ozcan et al., 1996).

How might MEP2 function as an ammonium sensor? One possibility is that MEP2 exists in different conformational states depending on its transport status (active, idling or off) and that associated effector proteins recognize these different conformations. Extracellular ammonium concentration, then, would be communicated to effectors via the structure of MEP2. MEP2 may directly interact with GPA2; such an interaction would be novel in G protein signaling systems, typically regulated by seven‐transmembrane domain proteins such as the pheromone receptors, STE2 and STE3. Efforts using the two hybrid system have not provided evidence for a MEP2–GPA2 complex. Instead, other protein intermediates may couple MEP2 conformational changes to the activation of GPA2. As an alternative, MEP2 could be required to regulate the production of a signaling molecule, possibly a nitrogen metabolite, under nitrogen starvation conditions that would activate GPA2 through a more conventional signaling mechanism. The phenotypic differences between MEP2 and its homologs MEP1 and MEP3 could result from a direct interaction of MEP2 with ammonium assimilating enzymes or other enzymatic machinery that generates such a signaling molecule, or with other downstream effectors. The first intracellular loop of MEP2, which our evidence suggests plays a role in MEP2 signaling, may mediate such an interaction with downstream effectors.

The pseudohyphal defects of the Δmep2/Δmep2 mutant strain are only observed when ammonium, the substrate of MEP2, is the sole nitrogen source. In contrast, yeast cells initiate this dimorphic switch in response to general nitrogen starvation, regardless of the nitrogenous compounds present. Thus permeases for other compounds may have a regulatory role similar to MEP2 in the presence of their substrates. In accord with this hypothesis, we find that the NPR1 protein kinase is required for pseudohyphal differentiation on media containing ammonium (Figure 7B), proline (Figure 8) or glutamine (Figure 8). A mutation in the NPI1/RSP5 gene suppressed this Δnpr1 phenotype, but only with proline or glutamine as the sole source of nitrogen. Genetic evidence indicates that NPI1, an essential protein‐ubiquitin ligase (Hein et al., 1995; Huibregste et al., 1995), destabilizes the proline (PUT4) and glutamine permeases (GNP1) but not the MEP ammonium permeases (Dubois and Grenson, 1979; Grenson and Dubois, 1982; Grenson, 1983a,b). Our finding that NPR1 and NPI1 regulate pseudohyphal growth suggests that permeases other than MEP2 may regulate this dimorphic transition in the presence of their substrates. Like NPR1, the GPA2 Gα protein is required for filamentous growth in response to starvation for any nitrogen source (Lorenz and Heitman, 1997), and hence signaling from each of these permeases could be mediated by GPA2. The multiplicity of ammonium permeases in yeast made this system ideal for dissecting the role of the individual permeases; in the case of many other nitrogen sources there is only a single high affinity transporter, making similar studies more difficult.

Finally, as nitrogen starvation plays a broad role in regulating growth and differentiation pathways in diverse fungi, related transport proteins may have a conserved function in regulating these developmental events. As one example, mating in S.pombe and C.neoformans requires nitrogen starvation, which is signaled in both organisms via a conserved Gα protein with marked identity to yeast GPA2: GPA2 in S.pombe and GPA1 in C.neoformans (Isshiki et al., 1992; Tolkacheva et al., 1994; Alspaugh et al., 1997). The mechanisms by which these G proteins are activated by nitrogen starvation are not yet known but, based on our findings, may involve a conserved role for related nitrogen permeases. Experiments to test this hypothesis in C.neoformans are currently in progress.

Materials and methods

Yeast strains, media and microbiological techniques

Yeast strains are listed in Table II. Standard yeast media and genetic manipulations were as described (Sherman, 1991). Limiting nitrogen media (SLAD; Gimeno et al., 1992) contains 0.17% Yeast Nitrogen Base without amino acids or ammonium sulfate, 50 μM ammonium sulfate, 2% dextrose and 2% Bacto agar. SLARG, to induce GPA2 alleles, contains 0.5% galactose and 2% raffinose (Lorenz and Heitman, 1997). SLADG, to induce the pGal‐STE12 construct, contains 2% galactose and 0.13% glucose (Liu et al., 1993). YNB media contains 0.67% yeast nitrogen base (minus amino acids plus ammonium sulfate) and 2% glucose. Media which use alternative nitrogen sources contained proline, glutamine or ammonium sulfate at 100 μM and 2% Noble agar (Difco). TC media (Grenson, 1983b) contains 250 μg/ml tryptophan, 250 μg/ml citrulline, 2% glucose and 2% Bacto agar.

View this table:
Table 2. Saccharomyces cerevisiae strains

Yeast transformations were performed as described (Schiestl et al., 1993). Disruption alleles for MEP1, MEP2 and NPR1 were constructed by replacement of coding sequences by the LEU2 selectable marker. The Δmep1::G418, Δmep2::G418, Δmep3::G418, Δgln3::G418 and Δure2::G418 mutations were created through PCR mediated disruption (Wach et al., 1994) in either Leu+ (MLY40, MLY41; see Table II) or Δleu2::hisG (MLY42, MLY43) host strains.


Plasmids are listed in Table III. The plasmid‐borne MEP1 (pML100), MEP2 (pML151), MEP3 (pML113) and NPI1 (pML95) genes were derived from PCR amplification from genomic DNA of strain MLY54a/α. Plasmids for the analysis of MAP kinase function have been described (Liu et al., 1993; Gimeno and Fink, 1994; Mösch et al., 1996), as have GPA2 plasmids (Lorenz and Heitman, 1997).

View this table:
Table 3. Plasmid list

Hybrid analysis

Hybrids between MEP1 and MEP2 were constructed by PCR overlap. The fusion proteins are diagrammed in Figure 6. Each hybrid uses the native MEP2 promoter sequences. Junction points are in the predicted extracellular loops. The hybrids were cloned into the multicopy URA3 plasmid YEplac195 (Gietz and Sugino, 1988). The amino acid residues that comprise the junction points are shown in Table IV, numbered as in Marini et al. (1997).

View this table:
Table 4. Amino acid residues comprising junction points


All single colony photographs were taken directly from Petri plates using a Nikon Labophot‐2 microscope with a 10× primary objective (Zeiss) and a 2.5× trinocular camera adaptor (Nikon). Unless otherwise stated, colonies were incubated at 30°C for 4 days.

Ammonium import assays

Ammonium uptake was assayed as in Marini et al. (1994). The Δmep mutant strains were grown to late log phase in minimal proline media (Mpro; 0.1% proline, 0.17% yeast nitrogen base, 2% glucose). Cultures were diluted to an OD600 of 1.0 in Mpro and 10 mM ammonium sulfate was added to a final concentration of 500 μM. At the indicated times, a portion of the culture was removed and the cells removed by centrifugation. The ammonium concentration in the culture supernatant was assayed using a glutamate dehydrogenase‐linked assay as described (Tabor, 1971).

Growth rates of Δmep mutant strains were determined by OD600 at various time points in either liquid SLAD or SD‐Ura media. Assays for glutamine synthetase (GLN1) or glutamate dehydrogenase (GDH1 and GDH2) activity were performed on extracts from wild‐type or Δmep2/Δmep2 strains inoculated into liquid minimal media (0.17% yeast nitrogen base, 2% glucose) with increasing concentrations of ammonium (from 50 μM to 50 mM) and SD‐Ura and grown for 4 h at 30°C. Assays were performed as described (Doherty, 1970; Mitchell and Magasanik, 1983).

FG(TyA)::lacZ reporter gene assays

These assays were performed in strains MLY97 (wild‐type), MLY225a/α (Δmep2/Δmep2), MLY230a/α (Δmep1/Δmep1) and MLY231a/α (Δmep1/Δmep1 Δmep2/Δmep2) expressing the FG(TyA)::lacZ reporter gene from plasmid pIL30‐LEU2. Assays were performed as described (Lorenz and Heitman, 1997).

Epitope tagging and Western blotting

MEP2 was epitope tagged with the hemagluttinin (HA) epitope using the integrative method of Schneider et al. (1995). Three direct repeats of the nine amino acid HA epitope were fused in‐frame to the MEP2 C‐terminus. This fusion gene (MEP2::HA) is under the control of the endogenous MEP2 promoter.

Cultures of homozygous diploid MEP2::HA strains MLY220 (wild‐type), MLY221 (Δnpr1/Δnpr1), MLY194a/α (Δgln3/Δgln3) and MLY198a/α (Δure2/Δure2) were grown overnight in YNB, washed once with water and diluted into minimal media containing ammonium sulfate (50 μM to 50 mM; Figure 7A) for 4 h at 30°C. Extracts were prepared by glass bead agitation in lysis buffer (50 mM Tris–Cl, pH 8, 1 mM DTT, 1 mM EDTA, 100 U/ml aprotinin, 0.5 mM PMSF, 1 μg/ml TPCK, 1 μg/ml pepstatin, 1 μg/ml benzamidine). Protein concentrations were determined through a modified Bradford assay (Bio‐Rad). Equal amounts of protein were boiled in sample buffer (250 mM Tris–Cl, pH 6.8, 25% glycerol, 10% β‐mercaptoethanol, 5% SDS, 0.01% Bromophenol Blue) and separated by SDS–PAGE.

The MEP2::HA protein was detected by Western blot using an α‐HA monoclonal antibody (Berkeley), a horseradish peroxidase conjugated α‐mouse secondary antibody and the ECL system (Amersham). Cyclophilin A (CPR1) was used as a loading control and was detected with a polyclonal α‐CPR1 antibody (Cardenas et al., 1995) and an α‐rabbit secondary antibody.

Δnpr1 suppressor analysis

Δnpr1 strains grow poorly on SLAD (likely due to inhibition of MEP activity) and on TC media (tryptophan–citrulline, due to inhibition of GAP1; Grenson and Acheroy, 1982). To identify suppressors of the growth defect, 105 cells of strains MLY54a or MLY54α were plated to either SLAD + uracil or TC + uracil media and UV irradiated at 7500 μJ/m2 to 60% survival. Thirty six mutants in six recessive and one dominant allele groups were identified. The only mutant in this collection which did not suppress the Δnpr1 mutant phenotype on all nitrogen sources tested also showed the behavior expected of an npi1/rsp5 mutant; that is, derepression of GAP1 activity as assayed by the GAP1‐mediated uptake of toxic levels of l‐lysine and d‐histidine on derepressing (ammonium) media (Grenson, 1983b). A low copy vector (pML95) containing the NPI1 gene generated by PCR complemented this phenotype. We also cloned NPI1 in the yeast integrating vector YIp211, and used a linearized form to transform the npi1 mutant strain MLY141a. Transformants displayed NPI1+ phenotype in 15/16 Ura+ strains. The npi1 mutant phenotype did not reappear in 29 complete tetrads of a cross between the tagged NPI1‐URA3 strain and a wild‐type strain.


We thank G.Fink, M.Ward and S.Garrett for generously providing strains and plasmids, J.Nevins, R.Wharton and M.Cardenas for comments on the manuscript, and S.Garrett, D.Lew and the members of the Heitman lab for helpful discussions. J.H. is an assistant investigator of the Howard Hughes Medical Institute.