INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Guanine nucleotide-binding proteins (G proteins) are important regulators of a wide spectrum of signal-transducing systems. G proteins consist of  and  subunits and the GTP-binding  subunit. The activity of these regulatory complexes is controlled by GDP-GTP exchange, which is accomplished by a transmembrane receptor and followed by dissociation of the  subunit from the subcomplex. Then, either the free  subunit or the dimer, or occasionally both, regulates downstream effectors. The signaling system is shut off by hydrolysis of GTP, followed by reassociation of the inactive complex. In higher eucaryotes, trimeric GTP-binding proteins are involved in the regulation of a large number of effectors, including adenylyl cyclase, phospholipase C, phospholipase A2, phosphoinositide 3-kinase, and ion channels (for reviews, see references 8 and 25).

In the yeast Saccharomyces cerevisiae, two  subunits (Gpa1p and Gpa2p), one  subunit (Ste4p), and one  subunit (Ste18p) have been characterized (22, 24, 38). The heterotrimeric complex composed of GPA1, STE4, and STE18 gene products has been shown to regulate the mitogen-activated protein kinase pathway in haploid cells upon pheromone stimulation via the pheromone receptor Ste2p or Ste3p (reviewed in references 1 and 18). The  and  subunits, which associate with the second G-protein  subunit, Gpa2p, are still unknown. Gpa2p plays a role in the regulation of cyclic AMP (cAMP) levels in cooperation with Ras2p (6, 19, 24, 26). Addition of glucose to glucose-starved yeast cells induces a transient peak of the intracellular cAMP level which correlates with the activation of adenylate cyclase by Ras proteins (4, 35). Furthermore, overexpression of Gpa2p causes an additional rise of the cAMP concentration and partially suppresses the growth defect of a temperature-sensitive ras2 mutant (19, 24).

Gpa2p is also involved in the pathway which signals pseudohyphal development under conditions of nitrogen limitation (17, 19). This signaling pathway is, at least in part, mediated by an increase of the cAMP level, leading to activation of protein kinase A (PKA). Whether Gpa2p activates adenylate cyclase directly or indirectly remains unknown (6, 19). In contrast to Ras2p, the Gpa2p signal transfer inducing pseudohyphal differentiation does not involve the mitogen-activated protein kinase cascade (19, 23). Recently, a membrane-spanning receptor, Gpr1p (G-protein-coupled receptor), has been shown to interact with Gpa2p in a two-hybrid assay (39, 41). This Gpa2p-coupled receptor initiates a Ras-independent signaling pathway and may be involved in the response of the cells to nutrients such as nitrogen, glucose, and other fermentable sugars (39, 42). The Gpr1p/Gpa2p pathway is thought to activate Sch9 protein kinase and to act in parallel with the Ras pathway (39). Both Ras2p and Gpa2p signals are likely required for cell growth control and pseudohyphal development.

To better understand the function of Gpa2p and its interplay with effectors, we sought to isolate proteins capable of physically interacting with Gpa2p. To this end, Gpa2p was used as a bait in the yeast interaction trap method (9). Among the genes isolated in this screen, one coded for Ime2p (Sme1p), a meiosis-specific protein kinase essential for the initiation of meiosis and sporulation under conditions of nutrient shortage (21, 30, 40). Kinase activity of Ime2p is required for the regulation of meiotic genes, presumably by phosphorylation of still unidentified substrates (16). In diploid cells, Ime2p expression is strongly induced by the transcriptional activator Ime1p upon exhaustion of nutrients (2, 15, 21, 31). Overexpression of Ime2p from a multicopy plasmid allows diploid cells to sporulate even in the presence of glucose and nitrogen (40).

Here we report that Gpa2p physically interacts with the C-terminal regulatory domain of the protein kinase Ime2p (Ime2CT) and inhibits sporulation when nitrogen or glucose is present. Inhibition of sporulation results, in part, from the inhibition of Ime2p kinase activity by the GTP-bound form of Gpa2p, effecting a switch to Gpa2p in the meiosis-pseudohyphal growth decision.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Yeast strains and media. Strains used in this study are listed in Table 1. SD medium contains 0.67% yeast nitrogen base without amino acids (Difco, Detroit, Mich.) and 2% glucose. YPD and YP lactate media contain 1% yeast extract, 2% Bacto Peptone (Difco), and either 2% glucose or 2% lactic acid adjusted to pH 5.5 with KOH. Sporulation (SPO) medium contains 1% potassium acetate (KAc). SLAD medium contains 50 µM ammonium sulfate, 2% glucose, and 0.17% yeast nitrogen base; SLADA medium contains 1% KAc in addition to glucose (19). Media were solidified with 2% agar (Difco).

Plasmids. Plasmids used for protein expression and oligonucleotides used for subcloning are listed in Tables 2 and 3, respectively.

Interaction trap. For selection of interaction partners by the two-hybrid system, S. cerevisiae Y190 (Clontech, Heidelberg, Germany) containing pGAL4-GPA2 as a bait was transformed with a yeast genomic interaction library (Clontech) by the lithium acetate method (14) to obtain 10⁶ transformants. Transformants were selected on SD plates lacking Trp, His, and Leu and containing 25 mM 3-amino-1,2,4-triazole. His-prototrophic colonies were assayed for -galactosidase activity as described by Fields and Song (9). Library plasmids were isolated as described by Hoffman and Winston (12) and analyzed by DNA sequencing.

Gene disruption. Gene disruption was performed by the one-step method of Rothstein (28). The GPA2 disruption cassette was constructed by replacing the GPA2 EcoRV DNA fragment with a blunt-ended EcoRI DNA fragment containing the TRP1 gene from plasmid Yrp7. The cassette was linearized with PvuII and DraI and used to disrupt GPA2 in strains W303 (7) and CEN.PK2 (27). For disruption of IME2, a BamHI-EcoRI fragment was replaced, after fill-in reaction, by a 2.4-kb HpaI DNA fragment containing LEU2. A 4.5-kbp NcoI and XhoI linear fragment was used to transform strains W303 and W303 gpa2/gpa2. Correct integration of the disruption cassettes was verified by Southern blot analysis or by PCR.

Construction of a homozygous tpk1^w1/tpk1^w1 diploid strain. A bcy1-tpk1^w1 haploid strain was transformed with a plasmid expressing the HO gene under the control of the GAL1 promoter. Transformants were grown on galactose medium, and zygotes were isolated by micromanipulation.

Purification of fusion proteins from E. coli. Recombinant proteins were expressed in E. coli BL21(DE3). Gpa2p-His₆ and Ime2p-His₆ were purified according to the protocol of the manufacturer, using Ni nitrilotriacetic acid (NTA) columns (Qiagen). GST and GST-Ime2p were purified as previously described (10).

Copurification experiments. Diploid strain MD211 cotransformed with pCUP1-His6GPA2 and with either pCUP1-HAIMECT or pRS423 was precultured in SD selective medium, diluted in lactate medium containing 1 mM CuSO₄, and induced for 6 h. Cells were harvested, resuspended in lysis buffer I (50 mM Tris-HCl [pH 7.5], 1% [vol/vol] Triton X-100, 0.1% [vol/vol] sodium dodecyl sulfate [SDS], 1 mM phenylmethylsulfonyl fluoride, 10 mM imidazole, and 1 µg each of aprotinin, leupeptin, and pepstatin A per ml), and disrupted with glass beads.

In vitro binding assay. Purified Gpa2p-His₆ was incubated with either 1 mM guanosine-5'-O-thiotriphosphate (GTPS) or 1 mM GDP (Sigma, Deisenhofen, Germany) for 30 min at 30°C in STE buffer (10 mM Tris-HCl [pH 8], 150 mM NaCl, 1 mM EDTA). The reaction was stopped by the addition of MgCl₂ to a final concentration of 20 mM. Gpa2p (500 ng) bound to GDP or to GTPS was diluted in 500 µl and incubated with glutathione-bound GST or GST-Ime2p at 4°C for 1 h. Columns were washed three times with STE buffer. Bound proteins were released from the columns by boiling, separated by SDS-PAGE, and analyzed by Western blotting with anti-Gpa2p antibodies and by Coomassie staining.

Protein kinase assay. Ni-NTA columns loaded with Gpa2p-His₆ were incubated with either 1 mM GTPS or 1 mM GDP in 20 mM Tris-HCl (pH 7.5). Columns were washed with reaction buffer (20 mM Tris-HCl [pH 7.5], 10 mM MgCl₂) to eliminate the excess of guanine nucleotides. After elution with 100 mM imidazole, Gpa2p-His₆ was used for kinase assays.

Miscellaneous. Antibodies directed against E. coli-purified Gpa2p-His₆ were raised in rabbits (Eurogentech). These antibodies were purified by affinity column chromatography using the purified Gpa2-His₆ recombinant protein coupled to cyanogen bromide-activated Sepharose. -Galactosidase assays were performed as previously described (11), using total cell extracts or permeabilized cells. -Galactosidase activities are expressed in Miller units. Protein concentrations were estimated by the method of Bradford (3).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Gpa2p interacts with Ime2p in a two-hybrid system. To identify proteins which interact with Gpa2p, we used the yeast two-hybrid system (9). Out of 20 independent positive clones, 17 were found to encode various overlapping parts of the 3' end of Ime2p, a protein kinase essential for meiosis and sporulation (40). These clones had in common a stretch of DNA coding for 33 amino acid residues of Ime2p (amino acids 445 to 478) (data not shown). To verify the specificity of the interaction between Gpa2p and Ime2p, BD-Gpa2p or AD-Gpa2p was assayed with Ime2CT (amino acids 445 to 645) or with full-length Ime2p. Ime2CT coexpressed with Gpa2p activated lacZ expression irrespective of whether the Ime2p part was fused to the Gal4p AD and Gpa2p was fused to the BD or vice versa (Table 4). About 1/40 of this activation potential was observed when the whole IME2 gene was used instead of the C-terminal segment. None of the plasmids, either alone or in combination with pLAM5'-1 as a control, led to activation of lacZ expression.

Ime2p interacts specifically with the active form of Gpa2p. GDP-bound G subunits are associated with subunits and thereby inactive. Addition of GTP to the G subunits triggers dissociation of the trimeric complex, leading to activation. Being a G subunit, Gpa2p may bind either GTP or GDP in response to specific signals in vivo. To test whether the GTP- or the GDP-bound form of Gpa2p was capable of interacting with Ime2p, transcriptional activation of the lacZ reporter gene was assayed, using a constitutively active Gpa2p^G132V or an inactive Gpa2p^G299A mutant as the target protein (19). Ime2p coexpressed together with the active form of Gpa2p led to activation of the reporter gene (Table 5). In contrast, Ime2p did not lead to activation of the lacZ reporter when coexpressed with the inactive form of Gpa2p. These findings demonstrate that interactions between the two proteins are specific and regulated by the activation status of Gpa2p. Most interestingly, Ime2pCT expressed together with either Gpa2^G132V or Gpa2p^G299A led to the same level of -galactosidase activity, which reveals that both GPA2 alleles are stably expressed. These results indicate that conformational changes of Gpa2p triggered by GTP or GDP binding are critical for the interaction with Ime2p whereas Ime2pCT interacts constitutively and independently of the activation status of Gpa2p.

Interactions between Gpa2p and Ime2p depend on the GTP-bound form of Gpa2p. We further investigated whether protein-protein interactions also occurred in yeast cells by using affinity copurification experiments. Since the GTP/GDP status of Gpa2p is not critical for its interaction with Ime2pCT, His-tagged Gpa2p and Ime2pCT fused to the HA epitope were coexpressed in a diploid strain disrupted for both copies of GPA2 and IME2 genes (strain MD211). Cell extracts were loaded onto Ni-NTA columns, and bound proteins were analyzed by Western blotting using anti-Gpa2 and anti-HA antibodies. HA-Ime2p was retained on the His₆-Gpa2p columns but not on the Ni-NTA matrix alone, demonstrating complex formation between the two proteins in vitro (Fig. 1).

View larger version (19K): [in this window] [in a new window]   FIG. 1.   Ime2CT specifically binds to Gpa2p. Strain MD211 was transformed with pCUP1-HAIME2CT together with an empty plasmid, pRS426 (lane 1), or with pCUP1-HisGPA2 (lane 2). Extracts from cells grown on lactate medium containing CuSO4 were purified on Ni-NTA columns. Bound proteins were analyzed by Western blotting using anti-Gpa2p and anti-HA antibodies.

View larger version (34K): [in this window] [in a new window]   FIG. 2.   Gpa2p bound to GTP physically interacts with Ime2p in vitro. Purified E. coli Gpa2p-His6 was preincubated with either GDP (lanes 2 and 4) or GTPS (lanes 3 and 5) and loaded onto glutathione resin carrying GST (lanes 2 and 3) or GST-Ime2p (lanes 4 and 5). Resin-bound proteins were analyzed by Western blotting with anti-Gpa2p antibodies and by Coomassie blue staining for GST and GST-Ime2p. Lane 1, Gpa2p-His6 purified from E. coli; lane M, molecular weight standards.

Gpa2p is a negative regulator of sporulation. Ime2p is essential for sporulation (40). Its expression is regulated positively by Ime1p and negatively by nutrients (21, 31). Deletion of IME2 abolishes sporulation, whereas its overexpression enables diploid cells to sporulate in the presence of glucose or nitrogen, conditions which normally inhibit spore formation. To address the physiological relevance of the interaction between Gpa2p and Ime2p, the IME2 gene was disrupted in wild-type cells and in gpa2/gpa2 homozygous diploid cells. Diploid mutant strains bearing the single deletion of GPA2 and IME2 sporulated normally, and both spore viability and germination were comparable to those for wild-type cells. The resulting haploid strains carrying the double disruption of GPA2 and IME2 displayed no obvious growth defects and had normal mating ability. These observations were not surprising, as IME2 expression is detected only in diploid strains under conditions of nutrient starvation (15). Finally, the ime2/ime2 gpa2/gpa2 homozygous diploid mutant showed the same defect in sporulation as the isogenic ime2/ime2 homozygous diploid single mutant (data not shown).

gpa2/gpa2

View larger version (17K): [in this window] [in a new window]   FIG. 3.   Gpa2p inhibits sporulation in the presence of nitrogen or glucose. (A) Wild-type (wt) and gpa2/gpa2 mutant strains were grown on glucose-rich medium, shifted to SPO medium containing 4 mM NH4Cl, and incubated for 10 days. Sporulation was measured by counting the asci. (B) The strain wild-type (), a gpa2/gpa2 strain () transformed with empty vectors, and a ime2/ime2 gpa2/gpa2 strain cotransformed either with pP584-IME2 and pG0304 () or with pP584-IME2 and an empty vector () were cultured on SD medium and shifted to SPO medium containing increasing concentrations of glucose as indicated. (C) Experiment performed as for panel B except that cells were shifted to SPO medium containing increasing concentrations of NH4Cl as indicated. Values represent averages of at least three independent experiments.

gpa2/gpa2

The negative effect of Gpa2p on sporulation is independent of cAPK. Modulation of PKA activity by cAMP interferes with entry of yeast cells into meiosis. Since overexpression of GPA2 leads to increased levels of cAMP, the negative regulation of sporulation by Gpa2p could be explained through the cAMP pathway (17, 24). To address this question, the effect of Gpa2p overproduction on sporulation was measured in a bcy1- tpk1^w1 mutant strain which lacks a cAMP-responsive protein kinase (cAPK) but still responds to nutrient starvation (5).

View larger version (27K): [in this window] [in a new window]   FIG. 4.   The constitutively active Gpa2p acts in a cAMP-independent pathway to regulate meiosis. A bcy1-tpk1w1 diploid strain transformed with either pVT-GPA2G132V (G132V), pVT-GPA2/G299A (G299A), or an empty plasmid (vector) was incubated in SPO medium for 2 days. Sporulation was scored by counting the asci. Values represent averages of at least three independent experiments.

Overproduction of Ime2p as well as the presence of acetate inhibits pseudohyphal development. Earlier reports have demonstrated the positive role of GPA2 in pseudohyphal development (17, 19). Our finding that Gpa2p also inhibited sporulation through its interaction with Ime2p led us to investigate whether IME2 overexpression or the presence of acetate may have an inhibitory effect on pseudohyphal development. The pseudohypha-forming diploid strain CEN.PK2 was transformed with pP584-IME2 expressing Ime2p and tested for filamentous growth. Wild-type cells developed pseudohyphae when grown on SLAD medium (Fig. 5A). In contrast, overexpression of Ime2p drastically reduced filamentous growth similarly to the gpa2/gpa2 mutation (Fig. 5B and 4C). Furthermore, the presence of 1% KAc in SLAD medium totally abolished filament formation (Fig. 5D). Thus, formation of pseudohyphae is prevented by overexpression of IME2 or by the presence of acetate. We conclude that pseudohyphal and meiotic developments are two mutually exclusive pathways which, in addition to RAS and cAMP, are regulated by Gpa2p and Ime2p.

View larger version (110K): [in this window] [in a new window]   FIG. 5.   Ime2p and the presence of acetate inhibit pseudohyphal development. Filamentous growth of the diploid strain CEN.PK2 transformed with an empty plasmid (A) or with pP584-IME2 overexpressing Ime2p (B) and of the isogenic gpa2/gpa2 mutant diploid strain (C) was inspected after 4 days of growth on SLAD medium. Strain CEN.PK2 strain was also tested for filamentous growth on SLADA medium (D).

Gpa2p does not affect the stability of Ime2p. Gpa2p binds to the acidic C-terminal tail of Ime2p, a domain which may be important for the proteolytic stability of Ime2p (16). It is, therefore, conceivable that Gpa2p regulates Ime2p activity by promoting its proteolytic degradation in vivo. To test this hypothesis, Western blot analyses were carried out under conditions where sporulation is repressed by Gpa2p. In these experiments, strain MD211 was transformed with pP584-IME2His6 and pG0304, expressing Ime2p-His6 and Gpa2p, respectively. Cells were grown on glucose medium and then shifted to SPO medium with or without nitrogen. Ime2p-His₆ could complement the ime2/ime2 mutation and allowed sporulation with wild-type efficiency, indicating that addition of the His₆ tag at the C-terminal of Ime2p had no effect on its physiological function (data not shown). Ime2p was detectable by immunodecoration when cells were grown on SPO medium (Fig. 6A, lanes 4 and 7) but not in YPD (Fig. 6A, lanes 3 and 6). In the presence of nitrogen, the level of Ime2p was down-regulated about threefold (Fig. 6A, lanes 2 and 4 or 5 and 7). These results are consistent with previous observations that transcriptional expression of the IME2 gene is regulated negatively by nutrients and positively by acetate (15). Coexpression of Gpa2p and Ime2p did not affect the level of Ime2p (Fig. 6A, lanes 2, 4, 5, and 7), indicating that Gpa2p does not induce proteolytic degradation of Ime2p.

View larger version (26K): [in this window] [in a new window]   FIG. 6.   Gpa2p does not promote proteolytic degradation of Ime2p. (A) Strain MD211 was transformed with pP584-IME2-His together with pRS426 as a control (Gpa2p) or with pG0304 (+Gpa2p). As a negative control, MD211 was transformed with plasmids pRS423 and pRS426 (Control). Transformants were grown in SPO medium containing 10 mM NH4Cl (+N), in SPO medium (N), or in YPD medium (G). After purification on an Ni-NTA column using the same protein amounts, bound proteins were analyzed by Western blotting using anti-His6 antibodies. (B) MD211 was transformed with either pRS426 as a control (Gpa2p) or pG0304 (+Gpa2p). Cells were grown in YPD medium (G), in SPO medium (N), or in SPO medium containing 10 mM NH4Cl (+N). Total extracts were analyzed by Western blotting using anti-Gpa2p antibodies. Control, Gpa2-His6 recombinant protein from E. coli.

Nitrogen induces interaction between Gpa2p and Ime2p. We further investigated whether the interaction of Gpa2p and Ime2p was regulated by the growth conditions. Gpa2p and Ime2p-His₆ were coexpressed in the diploid strain MD211, and cells were grown either on SPO medium or on SPO medium containing nitrogen. Cell extracts were loaded onto Ni-NTA columns, and bound proteins were analyzed by Western blotting using anti-Gpa2p antibodies (Fig. 7). Gpa2p alone did not bind to the Ni-NTA column (Fig. 7a, lanes 2 and 3). Gpa2p was retained on the column only when Ime2p-His₆ was purified from cells grown on KAc in the presence of nitrogen (Fig. 7a, lane 5), not when Ime2p-His₆ was purified from cells grown on KAc in the absence of nitrogen (Fig. 7a, lane 4). As nitrogen causes an about twofold increase in Gpa2p abundance (Fig. 7b), Gpa2p bound to Ime2p-His₆ could be directly proportional to the amount of Gpa2p present in the cell extracts. In contrast, Ime2p was expressed at higher levels when cells were grown in the absence of nitrogen, and recovery of Ime2p from the Ni-NTA columns was two to three times higher in SPO medium than in SPO medium supplemented with nitrogen (Fig. 7a). Therefore, copurification of Ime2p and Gpa2p should have been detected in both growth conditions if the interaction between the proteins was independent of the nitrogen source. As this was not the case, it may be reasonable to conclude that the presence of nitrogen regulates the association of Gpa2p and Ime2p in cells, i.e., under conditions when spore formation is inhibited by Gpa2p.

View larger version (34K): [in this window] [in a new window]   FIG. 7.   Nitrogen influences interaction between Gpa2p and Ime2p in vivo. Copurification of Gpa2p and Ime2p-His6 was performed as follows. Strain MD211 was cotransformed with pG0304 and either pRS423 as a control (lanes 2 and 3) or pP584-IME2His6 (lanes 4 and 5). Cells were grown in SPO medium (N; lanes 2 and 4) or in SPO medium containing nitrogen (+N; lanes 3 and 5). Total cell extracts were loaded onto an Ni-NTA column, and bound proteins were analyzed by Western blotting with purified anti-Gpa2p and anti-His6 antibodies (a). Aliquots of total cell extracts were also analyzed by Western blotting using purified Gpa2 antibodies (b). Lane 1, Gpa2p-His6 recombinant protein, purified from E. coli; lane M, molecular weight standards.

Gpa2p inhibits Ime2p kinase activity. Ime2p contains two characteristic domains. The amino-terminal domain (approximately 400 amino acids) harbors the kinase catalytic center and is required for initiation of meiosis. The carboxy-terminal peptide of 260 amino acids, which harbors six highly acidic subdomains, is not essential for kinase activity or sporulation but has a negative effect on either kinase activity or protein stability under certain nutritional conditions: deletion of almost the entire acidic tail has been shown to enhance sporulation efficiency on SPO plates containing glucose and to increase kinase activity in vitro (16). These observations suggest that Ime2p performs its physiological function through the phosphorylation of some unknown factors and that the C-terminal tail receives a signal to inhibit the kinase activity in the presence of nutrients.

View larger version (44K): [in this window] [in a new window]   FIG. 8.   Autoradiograms showing that Gpa2p-His6 recombinant protein bound to GTPS inhibits Ime2p kinase activity. (A) Histone H1 phosphorylation was assayed with GST-Ime2p purified from E. coli and [-32P]ATP in the presence of BSA (lane 2), Gpa2p bound to GDP (lane 3), or Gpa2p bound to GTPS (lane 4), washed free of unbound guanine nucleotides. GST was incubated under the same conditions in the presence of BSA (lane 1). As a control, histone H1 phosphorylation by PKA was assayed in the presence of either BSA (lane 5), Gpa2-GDP (lane 6), or Gpa2-GTPS (lane 7). (B) Ime2p-His6 purified from MD211 cells transformed with pP583-IME2His6 and grown in SPO medium was assayed for histone H1 phosphorylation in the presence of either BSA (lane 2), Gpa2p bound to GDP (lane 3), or Gpa2p bound to GTPS (lane 4). As a control, extracts from yeast cells transformed with an empty plasmid and grown in SPO medium were used (lane 1).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The G-protein  subunit Gpa2p has previously been shown to play a positive role in signaling pseudohyphal development (17, 19). Here we report that in addition to signaling pseudohyphal development, Gpa2p represses sporulation. Furthermore, Gpa2p interacts with the C-terminal regulatory domain of Ime2p, a protein kinase essential for spore formation, thereby regulating its activity.

Ime2p has been identified as a specific interaction partner of Gpa2p in the yeast two-hybrid system. Ime2p is a serine/threonine protein kinase of the CLK (Cdk-like kinase) subfamily and consists of two domains, an N-terminal kinase domain and an acidic C-terminal regulatory domain of about 260 amino acids (13). Both the regulatory C-terminal domain of Ime2p and the complete protein are able to interact with Gpa2p in the screen. Furthermore, using a constitutively active or inactive allele of GPA2, we show that the active conformation of Gpa2p associates with Ime2p whereas its inactive conformation does not. Similar results are obtained with the proteins purified from E. coli or S. cerevisiae in copurification experiments. GST-Ime2p purified from E. coli interacts more efficiently with Gpa2p bound to GTP than with Gpa2p bound to GDP. In contrast, when expressed alone in yeast, the acidic tail of Ime2p binds to Gpa2p independently of its GTP/GDP status. This latter observation explains why only overlapping clones bearing the C-terminal domain of Ime2p were isolated in the two-hybrid screen. Taken together, our data demonstrate that Gpa2p and Ime2p physically interact in yeast cells and that the GTP charge on Gpa2p is essential for this interaction.

The acidic tail of Ime2p has been shown to negatively regulate the kinase domain of Ime2p: removal of the C-terminal regulatory domain increases kinase activity in vitro (16). Direct protein-protein interactions with regulatory domains of protein kinases have previously been reported to be required for proper signaling (6, 32, 32, 37). We show here that GTP-bound Gpa2p inhibits Ime2p kinase activity whereas GDP-bound Gpa2p does not. This inhibition is specific for Ime2p kinase activity, as Gpa2p had no effect on PKA activity. Furthermore, Gpa2p is specifically phosphorylated by Ime2p, whereas it is not a substrate of PKA. It is conceivable that Ime2p modulates the activity or the location of the Gpa2p. These observations indicate that the active form of Gpa2p inhibits Ime2p kinase activity by interacting with the regulatory domain of Ime2p and suggest a feedback regulation of Ime2p on Gpa2p.

At least two signals are necessary for yeast cells to enter meiosis and sporulation. One is derived from heterozygosity at the mating-type locus; the other is produced by starvation for both nitrogen and fermentable carbon sources. The cAMP/PKA pathway signals both glucose and nitrogen proficiency in yeast. The Ras2p/cAMP pathway controls growth in response to the presence of fermentable sugar (4, 6, 34), whereas Gpa2p and cAMP are thought to be a component of the nitrogen sensing pathway (19, 20). The finding that gpa2/gpa2 ras2/ras2 double mutants display an additional growth defect compared to the single mutants demonstrates that Gpa2p and Ras2p have partially redundant functions (6, 23, 39). It has been demonstrated that overactivation of the Ras2/cAMP pathway in diploid cells impedes sporulation even in the absence of nitrogen and a fermentable carbon source, and it was assumed that this control was exerted via PKA. Our work provides evidence that at least part of this control is exerted through inhibition of Ime2p, the key regulator of entrance to the meiotic pathway, by its interaction with Gpa2p. Indeed, the sporulation efficiency of diploid cells expressing Ime2p from a multicopy plasmid in SPO medium containing glucose or nitrogen is 10- to 25-fold greater than that of wild-type cells. Coexpression of Gpa2p together with Ime2p reverts this effect three- to fivefold. Overexpression of Gpa2p decreases sporulation efficiency in the presence but not in the absence of nitrogen or glucose. Moreover, homozygous deletion of GPA2 allows diploid cells to sporulate in the presence of nitrogen at higher frequency than in wild-type cells. Finally, the active form of Gpa2p inhibits sporulation in a bcy1- tpk1^w1 strain. This strain, which contains disruptions of the catalytic subunits TPK2, TPK3, and the regulatory subunit BCY1 and encodes a functionally attenuated allele of TPK1, lacks cAPK (5). We conclude from these observations that at least one regulatory function of Gpa2p on meiosis is independent of the cAMP fluctuation and exerted through its interaction with the regulatory domain of Ime2p as long as nitrogen or glucose is available. These results are consistent with the previous observation that the acidic tail of Ime2p inhibits sporulation under partial starvation conditions: enhancement of sporulation in SPO medium containing glucose is observed in diploid cells expressing a truncated version of Ime2p lacking the C-terminal 207 residues (16).

Previous studies have revealed that Gpa2p is involved in the transfer cascade for induction of pseudohyphal growth upon exhaustion of the nitrogen source (17, 19). We show here that overexpression of Ime2p partially inhibits formation of pseudohyphae and that the presence of acetate, a sporulation inducer, completely abolishes this developmental program. These results suggest that sporulation and pseudohyphal development are mutually exclusive pathways and in part regulated by Gpa2p and Ime2p.

Interactions between Gpa2p and Ime2p are also linked to the supply of nitrogen and correlate with inhibition of sporulation. Indeed, Gpa2p interacts with Ime2p only in yeast cells grown on SPO medium containing nitrogen, not in cells grown on SPO medium alone. Our results indicate that the presence of nitrogen in SPO medium may lead to the activation of Gpa2p by GDP-GTP exchange and triggers interaction with Ime2p to inhibit sporulation. These observation are consistent with the finding that permanently active Gpa2p^R273A inhibits spore formation in SPO medium (39). Accordingly, an active Gpa2p relieves the requirement for the nitrogen signal and thereby inhibits sporulation even in SPO medium. Hence, wild-type Gpa2p is incapable of interacting with Ime2p in SPO medium without activation by nitrogen.

Gpa2p expression parallels Ime2p expression, in accord with a function in sporulation through regulation of Ime2p. On rich glucose medium, down-regulation of Ime2p activity is not required because Ime2p expression is repressed (15). Consistently, Gpa2p is expressed at low levels. Under starvation conditions, the expression of both Ime2p and Gpa2p is derepressed. After exhaustion of glucose, Ime2p is expressed through induction by IME1 but unable to allow sporulation as long as nitrogen is available. Gpa2p expression is induced about twofold under the same conditions. These findings support the idea that complete starvation for nutrients is required for sporulation to occur despite expression of Ime2p and that nitrogen inhibits sporulation-specific events downstream of Ime2p. Our data suggest that in wild-type cells, inhibition of Ime2p by Gpa2p is part of this down-regulating mechanism. This conclusion, however, does not exclude the contribution of the cAMP/PKA pathway to the regulation of meiosis and sporulation. On the contrary, the antagonistic action of Ime2p and cAMP/TPK kinases in pseudohyphal development and sporulation allows the abrupt switch from one developmental program to the other and prevents the occurrence of transition states, even under very gentle changes of solute concentration. This mechanism attributes the role of a switch between the two programs to Gpa2p.

In summary, two signals impinge on the control of sporulation: one in the presence of nitrogen or glucose through the inhibition of Ime2p kinase activity by Gpa2p, and the other through the cAMP/PKA and/or Sch9 pathway. This model is supported by the fact that inhibition of sporulation by the active GPA2^R273A allele is higher than that conferred by the active RAS2 allele, which, in contrast, causes a 10-fold-higher increase in heat shock sensitivity (39). These observations demonstrate that each of the two regulators, Gpa2p and Ras2p, must have a specific target(s) in addition to adenylate cyclase (6, 19, 39). Indeed, RAS2 promotes pseudohyphal growth through the activation of the Cdc42/Ste20/mitogen-activated protein kinase cascade. Alternatively, Gpa2p does not activate this cascade but appears to modulate the cAMP pool through an as yet unknown mechanism (17, 19, 23). It is still unclear whether, similarly to the mammalian Gs subunit, Gpa2p activates adenylate cyclase directly or indirectly (6, 19). Thus, Gpa2p interacts with different effectors under partial nutritional stress conditions to induce pseudohyphal development by activation of the cAMP/PKA and Sch9 pathways in cooperation with Ras2p. Simultaneously, Gpa2p prevents sporulation in part via inhibition of Ime2p kinase activity. The decision between the two alternatives is made through the availability of glucose. Glucose activates Ras2p and cAMP-dependent pathways and represses expression of IME1 and IME2. Thus, activation of the filamentous growth pathway excludes entrance to meiosis through the cooperation of both regulators Gpa2p and Ras2p. Further studies should focus on the identification of the subunits of the presumptive Gpa2p-containing trimeric G protein and their targets as well as their possible roles in Gpa2p-dependent signaling.