Phospholipase C Binds to the Receptor-like GPR1Protein and Controls Pseudohyphal Differentiation inSaccharomyces cerevisiae *

The hormone receptor-like protein Gpr1p physically interacts with phosphatidylinositol-specific phospholipase C (Plc1p) and with the Gα protein Gpa2p, as shown by two-hybrid assays and co-immune precipitation of epitope-tagged proteins. Plc1p binds to Gpr1p in either the presence or absence of Gpa2, whereas the Gpr1p/Gpa2p association depends on the presence of Plc1p. Genetic interactions between the null mutations plc1Δ,gpr1Δ, gpa2Δ, and ras2Δsuggest that Plc1p acts together with Gpr1p and Gpa2p in a growth control pathway operating in parallel to the Ras2p function. Diploid cells lacking Gpr1p, Plc1p, or Gpa2p fail to form pseudohyphae upon nitrogen depletion, and the filamentation defect of gpr1Δ and plc1Δ strains is rescued by activating a mitogen-activated protein kinase pathway via STE11-4 or by activating a cAMP pathwayvia overexpressed Tpk2p. Plc1p is also required for efficient expression of theFG(TyA)::lacZ reporter gene under nitrogen depletion. In conclusion, we have identified two physically interacting proteins, Gpr1p and Plc1p, as novel components of a nitrogen signaling pathway controlling the developmental switch from yeast-like to pseudohyphal growth. Our data suggest that phospholipase C modulates the interaction of the putative nutrient sensor Gpr1p with the Gα protein Gpa2p as a downstream effector of filamentation control.

Phosphatidylinositol-specific phospholipase C (PI-PLC) 1 hydrolyzes the membrane phospholipid phosphatidylinositol 4,5bisphosphate (PIP 2 ) to produce inositol 1,4,5-trisphosphate and diacylglycerol. In animal cells, these cleavage products serve as important second messengers; 1,4,5-trisphosphate triggers an intracellular Ca 2ϩ release, whereas diacylglycerol activates protein kinase C (1). Furthermore, the PI-PLC substrate PIP 2 itself is an important signal modulating the activity of membrane-bound proteins (2). The 10 known mammalian PI-PLC isozymes can be divided into three subtypes ␤, ␥, and ␦, differing in their structural organization and in their mode of activation by heterotrimeric G proteins and G protein-coupled hormone receptors (3).
The budding yeast Saccharomyces cerevisiae contains a sin-gle phospholipase C gene (PLC1) encoding a ␦ type PIP 2 -specific enzyme (4 -6). In most yeast strains the PLC1 gene is not essential for viability at 25°C, but Plc1p-deficient mutants arrest at temperatures above 35°C as multibudded enlarged cells unable to complete cytokinesis, they are sensitive against osmotic stress and nitrogen starvation, they do not sporulate as homozygous diploids, and they are defective in the utilization of nonfermentable carbon sources, suggesting that the hydrolysis of PIP 2 is required for a number of nutritional and stressrelated responses (4,7). The Plc1p-catalyzed formation of 1,4,5trisphosphate is stimulated by nitrogen feeding of starved cells; this response depends on a functional Ras GDP-GTP exchange factor, Cdc25p (8).
To learn more about specific PI-PLC functions in yeast, we have performed a two-hybrid screen with Plc1p as bait and identified several different prey peptides physically interacting with Plc1p (9). One of the prey peptides was detected in three independent clones and identified as a carboxyl-terminal region of the GPR1 gene product, a hormone receptor-like plasma membrane protein that was previously found in a similar screen using the G␣ protein Gpa2p as bait (10). A mutational analysis has demonstrated that Gpr1p acts upstream of Gpa2p, an activator of adenylate cyclase, in a Ras-independent growth control chain (10). The deletion of GPR1 abolishes the hyperactivation of adenylate cyclase upon glucose feeding, suggesting a glucose-sensing function of Gpr1p (11). The activation of adenylate cyclase by Gpa2p is required for another nutrient signaling pathway, the induction of pseudohyphal growth in diploid cells upon nitrogen depletion (12)(13)(14)(15). Here we report the physical interactions between Gpr1p, Plc1p, and Gpa2p, and we demonstrate important functions of Gpr1p and Plc1p in nitrogen-controlled signaling pathways leading to pseudohyphal growth.
Yeast strains were usually grown in rich medium (YPD, yeast extract, peptone, and dextrose) or in selective medium (SD, synthetic dextrose) complemented by auxotrophic requirements (20). YPGal contained 2% galactose instead of glucose. Synthetic low ammonia dextrose medium (SLAD) was prepared as described (12). Standard methods for yeast transformations and genetic analyses were used (20).
Plasmid Constructions- Table II summarizes the codes and features of plasmids. The standard vectors pUC19 and YEp24 were obtained from New England BioLabs, the two-hybrid vectors pEG202 and pJG4 -5 from R. Brent, and pRS416 from EUROFAN. A 3.3-kb BamHI fragment of pTY4 (5) containing the PLC1 coding and flanking regions was inserted into the sites of pRS416 (21) and YEp24 to obtain pKA42 and pKA43, respectively. The same BamHI PLC1 fragment was subcloned in a pUC19 derivative lacking a NdeI site to obtain pMS56. The NdeI site within the PLC1 coding region of pMS56 was converted to an EcoRI site by introducing the adaptor TAGGGAATTCCC to obtain pMS57. A 2.4-kb EcoRI/SalI fragment containing the PLC1 codons 79 -896 was inserted between the EcoRI and XhoI sites of the twohybrid vectors pJG4 -5 and pEG202 (22) to obtain pMS58 and pMS61, respectively. The GPR1 flanking and coding region was isolated as a 5.3-kb ClaI fragment from a cosmid (ATCC, clone 70974) and inserted into the single ClaI site of pRS416 to obtain pKA47. The same region was recovered from pKA47 as a 5.3-kb BamHI/XhoI fragment and inserted between the respective sites of YEp24 to obtain pKA49. pMS72 was isolated as prey plasmid from a genomic library in pJG4 -5, using pMS61 as bait. A 1-kb SalI/HindIII fragment of pKA47 was subcloned into pUC19, and the HindIII site of the resulting plasmid pEH13 was converted to a XhoI site by the adaptor AGCTGGCTCGAGCCA to obtain pEH14. A 1-kb EcoRI/XhoI fragment of pEH14 was inserted between the respective sites of pJG4 -5 to obtain pEH21. A GPR1 fragment (residues 273-622) was isolated by PCR, using the primers CGGAATTCGCCACCAGTGAAAGTAAAAGAATTAAAGCGCAAAT-TGG and CGCTCGAGTGCCCTTAGATTCTTTTGAATTTGTGCCCT. The 1-kb PCR product was trimmed by EcoRI and XhoI digestion and inserted between the repective sites of pJG4 -5 (pKA50). A 2.6-kb SalI/NheI fragment containing the GPA2 region was obtained by PCR and inserted between the respective sites of YEp24 to obtain pEH4. A MluI site was introduced between the EcoRI and XhoI sites of pJG4 -5 by inserting the linker sequence GAATTCCACGCGTCCCGGGAC TAGTCTCGAG, to obtain pEH2. A 1.8-kb MluI/XhoI pEH2 fragment was inserted between the MluI and NheI sites of pEH4 (pMS81) or between the EcoRI and SalI sites of pEG202 (pMS82) after repairing the non-cognate ends. The loxP-kanMX4-loxP plasmid pUG6 (19) was used to insert the GAL1 promoter and three copies of the HA epitope (pFM 224) or Myc epitope (pFM225). 2 A 2-kb BglII fragment containing the TPK2 gene was isolated by PCR and inserted into the BamHI site of YEp24 to obtain pEH31. Plasmids B2616 (containing the STE11-4 allele in the centromeric vector YCp50; Ref. 30), B2255 (YCp50-RAS2  ) and pFG(TyA)::lacZ-HIS3 (15,28) were obtained from H.-U.

Mösch.
Yeast Two-hybrid Methods-A two-hybrid screen with Plc1p as bait was performed as described (22,23) by using a yeast genomic (S288c) library in the vector pJG4 -5 (kind gift of R. Brent). Plasmids were rescued from positive clones expressing the reporter genes LEU2 and lacZ upon galactose induction, and inserts were identified by sequence determination. Two-hybrid interactions between fusion proteins were quantified by measuring ␤-galactosidase activity (24).
Immunological Methods-Yeast strains producing epitope-tagged proteins were grown in YPD to early log phase. Cells were collected by 2 M. Farkasovsky, unpublished observation.  centrifugation, washed twice with YP buffer, resuspended in YPGal and grown to A 600 ϭ 1.5. All subsequent steps were performed at 4°C. Cells were collected, washed twice with IP buffer (50 mM HEPES, pH 8, 50 mM NaCl, 2 mM EDTA, 1 mM NaN 3 , 5% glycerol, 1 mM phenylmethylsulfonyl fluoride, 1 tablet of Roche Molecular Biochemicals protease inhibitor mixture Complete/25 ml), and resuspended in an equal volume of IP buffer. The cell suspension was homogenized with 1 volume of acid-washed, sterilized, and chilled glass beads (0.45 mm) by vortexing at maximum speed for 30 s, followed by chilling on ice for 60 s. This step was repeated 10 times. Supernatants were then removed by centrifugation and pooled together with supernatants obtained after washing glass beads twice with 2 to 4 volumes of IP buffer containing 0.1% Triton X-100. The lysate was then centrifuged for 10 min at 10,000 ϫ g to remove cell debris. Protein concentration in the supernatant was determined by a Bio-Rad (Bradford) protein assay. Immunoprecipitation was performed in 450 l of IP buffer containing 0.1% Triton X-100 and 2 mg of total protein by adding 3 g of purified monoclonal anti-HA antibodies (BABCO, catalog no. 16B12) or anti-Myc antibodies (Santa Cruz Biotechnology, catalog no. 9E10). After incubation for 1 h at 4°C, a suspension of protein A-conjugated Sepharose beads (Amersham Pharmacia Biotech) in equal volume of IP buffer containing 2% bovine serum albumin (40 l final volume) was added, and the mixture was incubated for 2 h at 4°C with constant mixing. The beads were pelleted at 1000 ϫ g and washed four times with 700 l of IP buffer for 5 min at 4°C under constant mixing. The washed beads were resuspended in 20 l of 2ϫ Laemmli sample buffer and incubated for 5 min at 100°C. The supernatant (15 l) was analyzed by SDS-PAGE and blotted onto Hybond ECL membranes (Amersham Pharmacia Biotech) for 24 h at 4°C in transfer buffer (25 mM Tris, 195 mM glycine, 0.1% SDS, 10% methanol). Nonspecific binding sites were blocked by incubating the membrane for 1 h at room temperature in PBST (2.7 mM KCl, 1.5 mM KH 2 PO 4 , 137 mM NaCl, Na 2 HPO 4 , 0.1% Tween 20, pH 7.5) containing 5% nonfat dry milk. Immunodetection was accomplished by using biotinylated anti-mouse antibodies (1:2000 dilution) and streptavidin-biotinylated horseradish peroxidase complex (1:3000 dilution) with the enhanced chemiluminescence Western blotting detection system (ECL, Amersham Pharmacia Biotech).

Two-hybrid Interactions between Gpr1p, Plc1p, and
Gpa2p-We have used a two-hybrid screen (22,23) to identify proteins physically interacting with phospholipase C. A functional Plc1 protein lacking only the dispensable first 78 residues (5) was fused to the DNA-binding LexA protein, and the resulting plasmid pMS61 (see Table I) was co-transformed as bait together with a genomic yeast library in the activating domain fusion vector pJG4 -5. Three independent prey clones giving strong signals (expression of LEU2 and lacZ reporter genes) were found to contain the carboxyl-terminal residues 821-961 of the G protein-coupled receptor (GPR1) gene product (9). Almost identical Gpr1p regions (residues 863-961 and 839 -961) were previously identified as preys by using the G␣ protein Gpa2p as bait (10). As shown in Fig. 1A, the GPR1 gene product exhibits seven membrane-spanning domains, a feature characteristic for G protein-coupled hormone receptors including the two yeast pheromone receptors Ste2p and Ste3p (10). The two largest Gpr1p domains predicted to face the cytosolic side of the plasma membrane are the third intracellular loop (residues 273-621) and the carboxyl-terminal domain (residues 671-961).
The two-hybrid interactions of three Gpr1p peptides ( Fig.  1A) with almost intact Plc1p (residues 79 -869) and intact Gpa2p (Fig. 1, B and C) are summarized in Table III. Both Plc1p and Gpa2p interact strongly with the two carboxyl-terminal Gpr1p regions and, to a much lesser extent, with the third intracellular loop. We note that the longer carboxyl-terminal domain of Gpr1p is more efficient than the short one, if combined with Plc1p, whereas Gpa2p interacts more efficiently with the short C terminus of Gpr1p. There is also a weak but significant interaction between Plc1p and Gpa2p.
Co-immune Precipitation of Gpr1p, Plc1p, and Gpa2p-To confirm a physical association of Gpr1p with Plc1p and Gpa2p by co-immune precipitation, we have inserted a PCR-based cassette containing the kanamycin resistance gene kanMX4, the GAL1 promoter, and the epitope tags 3xMyc or 3xHA at the translational starts of the genes GPR1, PLC1, and GPA2. Heterozygous diploid strains co-expressing HA-Gpr1p and Myc-Plc1p (KAY320), HA-Gpr1p, and Myc-Gpa2p (SMY294) or HA-Plc1p and Myc-Gpa2p (SMY302) were grown in galactose media and analyzed by immunoblotting of crude extracts, using monoclonal antibodies against HA and c-Myc. Fig. 2 demonstrates that galactose-induced cells produce immunoreactive proteins corresponding in size to the calculated molecular mass of the tagged proteins. Myc-Gpa2p (55 kDa, lanes B and C) and HA-Gpr1p (110 kDa, lanes D and F) migrate as a single bands, whereas Myc-Plc1p or HA-Plc1p form a minor 92-kDa band in addition to the major 105-kDa species (lanes A and E), probably reflecting some proteolysis. Fig. 3 shows the results of co-immunoprecipitation experiments using heterozygous diploid strains expressing either HA-Gpr1p (SMY246), Myc-Plc1p (FMY525), or both tagged proteins (KAY320). Crude lysates of cells grown in galactose medium were treated with purified monoclonal anti-Myc antibodies, and the immuno-complex was absorbed to protein A-conjugated Sepharose beads. The immunoprecipitate was then probed by Western blotting using the two specific antibodies. Fig. 3, lane A, demonstrates that HA-Gpr1p co-precipitates with Myc-Plc1p from lysates of strain KAY320, whereas HA-Gpr1p is absent from anti-Myc precipitates of cells expressing FIG. 1. Protein maps of Gpr1p (A), Plc1p (B), and Gpa2p (C). A, the seven potential membrane-spanning domains of Gpr1p (10) are shown as black boxes. The three indicated fragments were used for two-hybrid assays. aa, amino acids. B, the two catalytic domains (X and Y) and a carboxyl-terminal C2 domain of Plc1p (4-6) are shaded and a potential EF-hand Ca 2ϩ binding site is indicated as a black circle. A functionally dispensable amino-terminal domain (residues 1-78) is indicated by broken lines. C, two conserved Gpa2p regions involved in GTP binding and hydrolysis (25) are indicated as open circles 1 and 2, respectively.  Finally, the Plc1p/Gpa2p association was tested in the presence or absence of Gpr1p. According to Fig. 5 Myc-Gpa2p copurifies with HA-Plc1p both in the presence (lane A) and absence of (lane D) of Gpr1p.
Growth of haploid strains containing either single or double null mutations was tested on solid SC medium at 25 and 37°C, as shown in Fig. 6. The deletion of PLC1 in the YEK107 background only slightly impairs growth at 25°C, but it leads to an arrest at 37°C, as observed in several other backgrounds (4 -6), whereas strains lacking GPR1, GPA2, or RAS2 remain viable at both temperatures. The reduced growth rate of these mutants at 37°C appears to be a property of the ⌺1278b background (HMC372 and YEK107). The double mutant gpr1⌬ gpa2⌬ has previously been shown to grow as efficiently as the single mutants, whereas combinations with the ras2⌬ mutation (gpr1⌬ ras2⌬ and gpa2⌬ ras2⌬) lead to a synthetic slow growth phenotype (10,14). Fig. 6 demonstrates that combinations with the plc1⌬ null mutation lead to similar results; the double mutants plc1⌬ gpr1⌬ and plc1⌬ gpa2⌬ have the growth phenotype of plc1⌬ (slightly impaired growth at 25°C, temperature sensitivity), whereas the plc1⌬ ras2⌬ double mutant exhibits extremely slow growth at 25°C. The growth defects of all PLC1-deficient strains could be rescued by adding the PLC1-containing plasmids pKA42 and pKA43 (data not shown). We have also confirmed the synthetic slow growth phenotype of gpr1⌬ ras2⌬ and gpa2⌬ ras2⌬ double mutants in the ⌺1278b background (data not shown). Our findings suggest that Plc1p acts in the same growth control pathway as Gpr1p and Gpa2p (no synthetic growth defects of double mutants) and in parallel with a Ras2p-controlled pathway (synthetic growth defect of plc1⌬ ras2⌬). Gpr1p and Plc1p Are Required for Pseudohyphal Growth under Nitrogen Depletion-Gpa2p has previously been shown to control the diploid-specific dimorphic transition from yeastlike to pseudohyphal growth in response to nitrogen starvation (13,14). This observation has prompted us to test a possible role of the Gpa2p-interacting proteins Gpr1p and Plc1p in the same nutrient-dependent signaling process.  A and D), SMY302  (lanes B and E), and SMY294 (lanes C and F) was obtained by lysis in 2 N NaOH containing 5% mercaptoethanol and trichloroacetic acid precipitation. Protein was analyzed by SDS-PAGE and immunoblotting using anti-Myc (lanes A-C) or anti-HA (lanes D-F). Protein size standard (␤-galactosidase, bovine serum albumin, and carbonic anhydrase) was from Bio-Rad.  SMY294 (lanes A and E), SMY246 (lanes B  and F), SMY193 (lanes C and G), and SMY296 (lanes D and H) were analyzed by SDS-PAGE and immunoblotting using anti-HA (lanes A-D) or anti-Myc (lanes E-H) antibodies. A and E), FMY559 (lanes B  and F), SMY193 (lanes C and G), and SMY306 (lanes D and H) were analyzed by SDS-PAGE and immunoblotting using anti-HA (lanes A-D) or anti-Myc (lanes E-H) antibodies. after 5 days incubation at 30°C on plates containing nitrogendepleted synthetic low ammonia dextrose medium (12). Fig. 7A demonstrates the formation of agar-invading pseudohyphal filaments at the periphery of colonies from strain HMC372 (containing the "empty" plasmid YEp24). This filamentation effect is completely abolished by the deletion of the GPR1 gene, as seen in Fig. 7B (strain KAY232 carrying YEp24). The filamentation defect of the gpr1⌬ strain is rescued if the "empty" plasmid YEp24 is replaced by pKA49, a YEp24 derivative containing the wt GPR1 gene (Fig. 7C). Furthermore, filamentation is also restored by the dominant-active RAS2 Val-19 allele (Fig. 7D) by increasing the dosage of the TPK2 gene, which encodes one of the three catalytic subunits of cAMP-dependent protein kinase (26,27), or by introducing the dominant-active STE11-4 allele (29), a component of the MAPK cascade involved in filamentation control (13,14,16), as seen in Fig. 7, panels E and F, respectively.

FIG. 5. Co-immune precipitation of Myc-Gpa2p and HA-Plc1p in the presence and absence of Gpr1p. Anti-Myc immunoprecipitates obtained from strains SMY302 (lanes
Similar results were obtained with homozygous diploid strains lacking the PLC1 gene, as shown in Fig. 8. The plc1⌬ strain KAY322 exhibits a somewhat ragged colony morphology but does not form filaments (Fig. 8B). The filamentation defect of plc1⌬ cells is rescued by a multicopy TPK2 plasmid (Fig. 8E) or a centromeric STE11-4 plasmid (Fig. 8F) but not by the RAS2 Val-19 allele (Fig. 8D), suggesting a role of Plc1p downstream of Ras2p (see "Discussion"). We do not have an explanation why the filamentation defect of plc1⌬ cells is only partially restored by the multicopy PLC1 plasmid pKA43 (Fig. 8C) or by the centromeric PLC1 plasmid pKA42 (data not shown). Differential Effects of GPR1 and PLC1 Deletions on the Transcription of the Reporter Gene FG(TyA)::lacZ-The expression of the reporter FG(TyA)::lacZ, a fusion of the Ty1 transposon to the ␤-galactosidase gene, depends on the Ste12p transcription factor (31) and correlates well with pseudohyphal growth induced by nitrogen starvation, mainly reflecting the activity of the Ste20p-MAPK cascade involved in the filamentation process (14 -16, 32). We have therefore used this reporter to test the role of Gpr1p and Plc1p in filamentation control by an alternative approach.
The data of Table IV indicate a strong influence of Plc1p on the reporter gene expression; a 5-fold reduction of lacZ activity is observed upon the plc1⌬ deletion, and the activity is almost completely recovered by adding back a plasmid-borne PLC1 gene. The 2-fold reduction of lacZ activity in the ras2⌬ strain YEK44 corresponds to the effects of deleting components of the MAPK pathway (e. g. ste20⌬, ste7⌬, ste12⌬), as observed by others (15). In contrast, the lacZ activity remains relatively high upon the GPR1 deletion (1.2-fold reduction), suggesting that Plc1p is more important than Gpr1p for controlling the MAPK pathway (see "Discussion").

DISCUSSION
We have shown that the hormone receptor-like plasma membrane protein Gpr1p (10) interacts physically with both Plc1p (4 -6) and with the G␣ protein Gpa2p (25, 10), using two independent methods: two-hybrid interactions of fusion proteins and co-immune precipitation of epitope-tagged proteins. Furthermore, we demonstrate that all three proteins play important roles in controlling the switch between yeast-like and pseudohyphal growth of diploid cells upon nitrogen depletion (12)(13)(14)(15)(16).
The simplified model shown in Fig. 9 may help to discuss some details of these findings. The GPR1 gene sequence predicts the intracellular orientation of two large protein regions (see Fig. 1A): the loop between the membrane-spanning domains 5 and 6 (349 residues) and the carboxyl-terminal 291 residues. A shorter carboxyl-terminal region (141 residues), which was identified as prey during two-hybrid screens with intact Gpa2p (10) or near-intact Plc1p (9) as bait, interacts more efficiently with Gpa2p (relative activity 1334 units, see Table III) than the long carboxyl terminus (919 units), whereas the long carboxyl terminus is more efficient in binding Plc1p (340 units) than the short one (Ref. 28.7 units). Gpa2p and Plc1p also interact with each other (12 units) and with the Gpr1p loop region (7 and 9 units, respectively) but with lower efficiency.
Although two-hybrid data do not necessarily reflect in vivo interactions, we tentatively conclude that Plc1p and Gpa2p associate preferentially with the carboxyl-terminal Gpr1p domain, whereby the upstream region of the long carboxyl terminus is more important for Plc1p binding than for Gpa2p binding. We have confirmed and extended the physical interaction studies by co-immune precipitation of Myc-or HA-tagged proteins, carefully ruling out nonspecific interactions. It turned out that the Gpr1p/Plc1p complex is formed either in the presence or absence of Gpa2p, and the Plc1p/Gpa2p complex is formed either in the presence or absence of Gpr1p, whereas Gpa2p associates with Gpr1p only in the presence of Plc1p but not in its absence.
These findings may suggest that PIP 2 -specific phospholipase C is required to expose a Gpa2p binding site at the carboxylterminal Gpr1p domain, perhaps by modulating the interaction between basic residues of the Gpr1p polypeptide and the acidic PIP 2 head groups at the inner side of the plasma membrane. Such interactions have been shown to be critically important in exposing an ATP binding site at the carboxyl-terminal domain of a mammalian ATP-sensitive potassium channel (34). Gpa2p is known to be required for the induction of pseudohyphal growth in diploid cells upon nitrogen starvation (13,14), by activating a cAMP-dependent filamentation pathway (13,14,25). Here we show that Gpr1p and Plc1p act upstream of Gpa2p in the same process; filamentation is prevented by deleting either GPR1 or PLC1 (in homozygous diploids), and is restored by the corresponding wt genes.
Furthermore, the filamentation defect of the null mutations gpr1⌬ and plc1⌬ is suppressed by activating the cAMP pathway via overexpression of the cAMP-dependent protein kinase Tpk2p or by activating the MAPK pathway via the dominant active STE11-4 allele. Both signaling pathways have previously been shown to control pseudohyphal growth upon nitrogen starvation (13-16, 27, 35, 36), converging in the transcriptional control of filamentation genes such as FLO11 (35). The unusually large FLO11 promoter (spanning at least 2.8 kb) is regulated by the cAMP-dependent pathway via the transcription factor Flo8p and by the MAPK pathway via the transcription factors Ste12p and Tec1p. Both filamentation pathways can replace each other by overexpression or constitutive activation of single components (e.g. overexpressed Flo8p suppresses the loss of Ste12p and vice versa) (35).
The MAPK filamentation pathway can be monitored in a more specific way by the transcriptional reporter FG(TyA)::lacZ, which responds much more to dominant-activated STE11-4 (8-fold stimulation of lacZ activity) than to the activation of the cAMP pathway (less than 2-fold stimulation) (15). Using this reporter we find that the lacZ activity strongly depends on the Plc1p function (5-fold reduction in plc1⌬ extracts), but it is less dependent on the Gpr1p function (1.2-fold reduction in gpr1⌬ extracts). In comparison, the lacZ activity is reduced 2-fold in extracts of ras2⌬ (see Table IV) and ste20⌬ mutants (15).
This somewhat unexpected result appears to indicate that the MAPK filamentation pathway requires the Plc1p function, whereas Gpr1p may operate mainly through the Gpa2p/cAMP pathway. In addition, we find that the dominant-active RAS2 Val-19 mutation suppresses the filamentation defects of gpr1⌬ (Fig. 7D) and gpa2⌬ strains (data not shown) but does not rescue the plc1⌬ defect (Fig. 8D). This observation suggests that Plc1p has a second function downstream of Ras2p, perhaps that of controlling the activation of the MAPK pathway via Cdc42p (15) in addition to activating the cAMP pathway by modulating the Gpr1p/Gpa2p interaction. According to the model of Fig. 9, the G protein-coupled receptor-like Gpr1 pro-  9. A model for filamentation control in diploid S. cerevisiae cells. The receptor-like plasma membrane protein Gpr1p is suggested to act as a nitrogen sensor regulating the switch between yeastlike and pseudohyphal growth by binding phospholipase C (Plc1p) and the G␣ protein Gpa2p at its intracellularly oriented carboxyl-terminal domain. Gpa2p has previously been shown to activate a cAMP-dependent filamentation pathway, which operates in parallel with a Cdc42p/ Ste20p/MAPK pathway (12)(13)(14)(15). Ras2p activates both signaling pathways via Cyr1p and Cdc42p, respectively, during diploid filamentation (12)(13)(14)(15), but also during invasive growth, a related differentiation process (33). tein functions as a nitrogen sensor, which activates a cAMPdependent filamentation signaling pathway by subsequently binding Plc1p and Gpa2p. A similar nutrient-sensing function has been suggested for Mep2p, a high affinity ammonium permease required for filamentation control (36). Although the filamentation defect of mep2⌬/mep2⌬ diploids is suppressed by dominant active GPA2 or RAS2 mutations, there is no evidence directly linking Mep2p and Gpa2p, and it is possible that Mep2p functions in a signaling pathway separate from either the Gpa2p/cAMP or MAPK pathway (36). The relationship between the Grp1p-and Mep2p-controlled filamentation routes thus remains to be established.
Previous studies have implied a general role for yeast Plc1p in regulatory pathways necessary for adaptation to changing nutrient and temperature conditions (4, 6, 7), and some phenotypic properties of plc1⌬ mutants (loss of viability upon nitrogen starvation, sporulation defect of homozygous diploids) point to the role of Plc1p in nitrogen-controlled signaling pathways (4). Our data are in agreement with this view by placing the Plc1p function within a well defined nitrogen signaling pathway involved in filamentation control. The interaction of Plc1p with a receptor-like protein (Gpr1p) and a G␣ protein (Gpa2p) is reminiscent of the association of mammalian PLC-␦ with agonist-bound ␣ 1 -adrenergic receptors and with receptorcoupled G h ␣, a multifunctional GTP-binding protein having transglutaminase activity (3,37).
Other phenotypic properties of yeast plc1⌬ strains such as cytokinesis defects (4) or aberrant chromosome segregation (6) point to multiple functions of Plc1p. Indeed, we observe a physical interaction of Plc1p with Num1p, a cortical protein controlling nutrient-dependent nuclear migration (38,39), with Bni4p, a bud neck protein controlling septum formation and cytokinesis (40), as well as with a few functionally unknown transmembrane proteins (9). Our data support the view that yeast Plc1p has multiple roles in modulating membrane/protein interactions by cleavage of the lipid PIP 2 .