INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Phosphorylation of the G protein-coupled receptors (GPCRs) family generally serves to negatively regulate receptor activity, i.e., functioning in desensitization. For the ₂-adrenergic receptor (₂AR), a well-studied member of the GPCR family (30, 36), desensitization is initiated through Ser-Thr phosphorylation of the cytoplasmic, C-terminal regulatory tail domain (CTD) of the receptor by the dedicated kinase, the ₂AR kinase (ARK). The binding of -arrestin protein to the phosphorylated CTD then serves to attenuate receptor responsiveness both by uncoupling the liganded receptor from the G protein and by directing the binding of the liganded receptor into clathrin-coated pits, allowing it to be sequestered by endocytosis to the inside of the cell. This feedback control is regulated through the regulation of ARK. ARK is activated through interactions with the liganded receptor, through interactions with plasma membrane lipids, and through a binding interaction with the subunit of the heterotrimeric G protein. Free subunit, released from the heterotrimer through the actions of the liganded receptor, remains plasma membrane associated due to C-terminal prenylation of the  subunit and thus serves as a plasma membrane binding site for cytosolic ARK. In addition to this ARK-mediated, so-called homologous desensitization, ₂AR is also subject to heterologous desensitization, i.e., phosphorylation by protein kinase C and protein kinase A (PKA), which may be activated through the actions of a variety of different signaling pathways. PKA also mediates direct feedback control of ₂AR; as adenylate cyclase is the primary downstream ₂AR effector, agonistic stimulation of ₂AR elevates cytoplasmic cyclic AMP levels and serves to activate PKA. ₂AR provides a well-characterized paradigm for GPCR desensitization. The GPCR family, however, is huge and evolutionarily diverse, encompassing hundreds of different receptors in organisms throughout the eukaryotic kingdom. The few cases where the desensitization mechanism has been analyzed have revealed both a diversity of mechanisms and a reiteration of common themes (36). While ligand-stimulated CTD phosphorylation generally plays a central role, the molecular mechanisms that regulate this phosphorylation often stray from the ₂AR paradigm. The work presented herein focuses on the mechanism underlying the ligand-induced phosphorylation of a yeast GPCR.

In the yeast Saccharomyces cerevisiae, two GPCRs function in the sexual conjugation of the two haploid mating types, a and . Identifying and locating a potential mate involves tracking the signature pheromone of the mating partner to its source.  cells serve as the source for the secreted peptide pheromone -factor, while a cells secrete the farnesylated peptide a-factor. The two GPCRs serve to detect these pheromones. The -factor receptor (Ste2p), localized to the surface of a cells, detects -factor and thereby its  cell mate, while the a-factor receptor (Ste3p), located on  cells, detects a-factor from nearby a cells. The ensuing cell-cell contact and cell fusion depend on a chemotropic response wherein new cell growth is polarized into a tapered mating projection directed toward the pheromone source. The pheromone receptors, through their detection of the external pheromone gradient, play a central role in this chemotropic response. Both receptors have relatively large CTDs which, although dispensable for the gross functioning of the receptor (ligand binding and G protein coupling), are required for receptor regulation and are required for the production of a normal mating projection (25) (A. F. Roth and N. G. Davis, unpublished results). The regulatory functions controlled by the pheromone receptor CTDs, namely, desensitization and endocytosis (2, 8, 25, 40, 42), likely help to shape this chemotropic response.

The elements which comprise the pheromone signaling pathway are now quite well understood (29, 47). Liganded receptors couple to a typical heterotrimeric G protein, which in turn activates a downstream mitogen-activated protein kinase (MAPK) cascade. Elements of the MAPK cascade, the MAPK kinase kinase Ste11p, the MAPK kinase Ste7p, and the MAPK Fus3p, are coordinated upon the scaffolding protein Ste5p. Propagation of the signal along the pathway involves a scheme of sequential activation wherein each kinase is activated through phosphorylation by the kinase immediately upstream. Two of the main outcomes of signaling, G₁ cell cycle arrest and the transcriptional induction of a set of pheromone-responsive genes, are triggered by Fus3p-mediated activation of Far1p, the effector of cell cycle arrest, and Ste12p, the transcriptional activator for pheromone-responsive genes.

For the yeast G protein-coupled signaling pathway, the free subunit of the G protein, not the G subunit, initiates downstream signaling. Reminiscent of ARK activation (described above), recent work with yeast indicates that a key signaling step may be the G-directed translocation of the Ste5p-MAPK complex to the plasma membrane (37). Once localized to the plasma membrane, the Ste5p-MAPK complex may be activated through an interaction with other plasma membrane-localized signaling components. Relevant in this regard is the PAK kinase homologue Ste20p, as well as both the Rho-like GTPase Cdc42p and its guanine nucleotide exchange factor Cdc24p.

Although the two pheromone receptors, Ste2p and Ste3p, have no obvious sequence homology, both activate the same downstream signal transduction pathway (1). Furthermore, regulation of the two receptors appears, superficially at least, to be similar. Both are subject to CTD phosphorylation: both show a constitutive level of CTD phosphorylation which increases upon exposure of the cells to the appropriate pheromone ligand (8, 40). Similarly, both receptors are also subject to two modes of ubiquitin-directed endocytosis (16, 43)a constitutive, ligand-independent mode as well as a ligand-dependent uptake mode (8, 23, 45).

As with ₂AR, multiple Ser-Thr residues within the pheromone receptor CTDs apparently serve as phosphoryl acceptors. For Ste2p, phosphorylation occurs at at least 10 distinct CTD sites (40, 56). As with ₂AR, pheromone receptor CTD phosphorylation likely functions in desensitization: receptors with deletions of portions of the CTD or having mutated CTD serines manifest a hypersensitive response to pheromone challenge (2, 3, 25, 40). Details of this process, however, may stray from the ₂AR paradigm, as clear homologues either for arrestin or for a ARK-like kinase are notably absent from the completed yeast genomic database.

For Ste2p, -factor-induced phosphorylation does not require the instigation of G protein-mediated signal transduction from the liganded receptor: ligand-induced phosphorylation of Ste2p proceeds normally in cells lacking the heterotrimeric G protein (56). We find in the present work that quite a different story applies to the ligand-induced phosphorylation of the a-factor receptor. This phosphorylation does require the G protein heterotrimer. Indeed, not only is the G protein required but also required are many of the downstream signal transduction components. Based upon these results, we have proposed a model for the regulation of Ste3p phosphorylation wherein the key step is the localization of activated MAPK to the plasma membrane.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Plasmids. pND733 is a pRS316-based URA3/CEN/ARS plasmid (46) carrying a GAL1p-STE2-HA construct. Ste2p expressed from this construct has a single hemagglutinin (HA) epitope tag fused to its C terminus following Thr₄₂₅ and replacing the Ste2p C-terminal six amino acid residues. Three Ycp50-based URA3/CEN/ARS plasmids carrying either wild-type STE12 (p650) (13) or the equivalent plasmid with one of two in-frame deletions within the STE12 coding sequence were used. The ste12253-335 allele removes STE12 residues 253 through 335, and the ste12255-354 allele removes residues 255 through 354 (24). pND882 carries the GAL1p-STE5-CTM construct from pGS5-CTM (37) on pRS316. The Ste5-CTM fusion has the single C-terminal transmembrane domain (CTM) of Snc2p fused to the Ste5p C terminus. pC1-H6 is a URA3/CEN/ARS GAL1p-STE12 isolate from a library of GAL1-driven yeast genomic fragments (P. Pryciak, personal communication). pRD-STE11-H3 is a URA3/CEN/ARS GAL1p-GST-STE11N plasmid (32).

Strains. Table 1 shows the strains used in this study. The GAL1p-STE3(M271I, M304I) allele was constructed by two sequential oligonucleotide site-directed mutageneses (26) of the GAL1p-STE3 URA3-integrating plasmid pSL1904. pSL1904 is identical to pSL1839 (44) except that the 760-bp GAL1,10 promoter fragment of pSL552 (1) substitutes for sequences 417 to 110 bp upstream of the STE3 open reading frame. To construct NDY867, the GAL1p-STE3(M271I, M304I) allele was integrated at the STE3 locus of NDY344 (44), displacing ste3::LEU2 via the two-step gene replacement method (43).

ste2::TRP1

ste3::LEU2

MAT HIS3_P::STE3

ste20::URA3

ste11::URA3

ste12::LEU2

fus3::LEU2

fus3::LEU2

kss1::URA3

Cell labeling, pheromone treatment, extract preparation, and immunoprecipitation. Cells were pulse-labeled for 10 min with a mixture of [³⁵S]methionine and [³⁵S]cysteine and then chased for various times with excess cold amino acids as described previously (43). For a-factor treatment, a concentrated culture supernatant from yeast cells which overproduce a-factor was prepared (43). The a-factor activity of this preparation was estimated to be 400-fold more concentrated than that of an unconcentrated supernatant from a saturated culture of wild-type MATa cells. Pulse-labeled cells were treated for 15 min with 20 µl of one of three different a-factor dilutions per labeling reaction: concentrated a-factor (1×), a 1:10 dilution (0.1×), or a 1:100 dilution (0.01×) (YPD medium was used for dilutions). Mock-treated controls were treated with a concentrated supernatant prepared in parallel from isogenic mfa1 mfa2 cells (43). Following the a-factor treatment of the labeled cultures, cells were collected by centrifugation, extracts were prepared via glass bead disruption, and Ste3p was immunoprecipitated, all as described previously (43).

Phosphatase treatment of labeled protein extracts. In some cases, ³⁵S-labeled protein extracts were treated with phosphatase prior to immunoprecipitation. For this, the glass bead protein extraction method (43) was altered to reduce the sodium dodecyl sulfate (SDS) concentration within the final labeled protein extract. Culture aliquots corresponding to approximately 2 × 10⁷ cells, pulse-labeled and chased as described above, were collected by centrifugation, resuspended in 15 µl of extraction buffer (40 mM Tris-Cl [pH 8.0], 8 M urea, 0.1 mM EDTA, 1% -mercaptoethanol), transferred to a tube containing a 10-µl volume of acid-washed glass beads (212 to 300 µm in diameter; Sigma Chemical Co., St. Louis, Mo.), and vortexed for 15 s. An additional 15 µl of extraction buffer supplemented with 1.5% SDS was added, and cell lysis was completed with an additional 1 min of vortexing. Samples were immediately heated at 120°C for 5 min and then stored for several hours at 0°C prior to initiation of the phosphatase treatment. Just prior to the phosphatase treatment, samples were reheated at 100°C for 5 min, following which a 9-µl portion of the extract was diluted into 0.8 ml of phosphatase digestion buffer (20 mM sodium citrate [pH 6.0], 50 mM NaCl, 0.5 mM phenylmethylsulfonyl fluoride, 1 µg of leupeptin per ml, 1 µg of pepstatin per ml). Digestion was done with 0.5 U of potato acid phosphatase (Boehringer Mannheim Biochemicals, Indianapolis, Ind.) for 40 min at 25°C. Mock-digested control samples were diluted and incubated in parallel. Reactions were terminated, and samples were prepared for immunoprecipitation with the addition of 0.2 ml of concentrated immunoprecipitation buffer (0.5 M Tris-Cl [pH 8.0], 0.36% SDS, 10 mM EDTA, 0.5% Triton X-100). Immunoprecipitation was done as described previously (43).

Immunoblots. Log-phase cultures of cells carrying the GAL1-regulated receptor genes, either plasmid STE2-HA (pND733) or chromosomally integrated GAL1p-STE3, were grown overnight in 2% raffinose medium as described previously (44). Receptor expression was induced with the addition of galactose to 2% for various times and subsequently repressed with the addition of glucose to 3%. At various times after glucose addition, cells were treated with pheromone. a-Factor treatment and mock pheromone treatment were as described previously (8), except that 0.5 volume of pheromone supernatant was added. For -factor treatment, synthetic -factor peptide (Sigma) was added to either 10⁵ M or 10⁷ M. Preparation of protein extracts for Western blotting, SDS-polyacrylamide gel electrophoresis (PAGE), and transfer to nitrocellulose were done as described previously (8). Ste2p was detected with monoclonal antibody (MAb) HA.11 (BAbCo, Berkeley, Calif.) specific for an HA epitope tag C-terminally fused to Ste2p (pND733). Ste3p was detected with affinity-purified rabbit polyclonal antibodies directed against the Ste3p CTD (43). Blots were developed with the ECL chemiluminescence system (Amersham Corp., Arlington Heights, Ill.). In some cases, prior to Western blotting, protein extracts were treated with potato acid phosphatase at a final concentration of 0.06 U/ml as described previously (43).

CNBr treatment. Protein extracts were prepared for immunoblotting from MAT GALp-STE3(M271I, M304I) NDY867 cells essentially as described above except that 5 × 10⁸ cells were used to make the extracts more concentrated. Extracts (20 µl) were digested in a 200-µl reaction mixture containing 70% formic acid and 2.5 M CNBr. Digestion was done at 25°C for 16 h in the dark. Protein was precipitated with trichloroacetic acid and prepared for SDS-PAGE and Western blot analysis as described previously (43). Some samples were treated with phosphatase as described above prior to SDS-PAGE.

-Galactosidase assays. MAT FUS1-LacZ cells harboring the GAL1-driven constitutively active STE gene constructs either as CEN plasmids or as a chromosomally integrated construct (rad16::GAL1p-STE4) were grown as overnight log-phase cultures in minimal synthetic medium supplemented with 0.2% yeast extract, as for pulse-labeling (43), with the exception that 2% raffinose was substituted for glucose. To induce the expression of the signaling component, 2% galactose was added for 2 h. Culture aliquots were then collected by centrifugation, cells were permeabilized, and -galactosidase activity was determined as described previously (22).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Ligand-induced phosphorylation of the Ste3p CTD. Ste3p is a short-lived protein: in cells growing at 30°C, Ste3p is degraded with a half-life of about 15 min (8). Rapid turnover occurs via rapid constitutive endocytosis, which removes the receptor from the cell surface and delivers it to the vacuole for degradation. This process is ligand independent, occurring whether or not the a-factor ligand is present. A consequence of such dynamic membrane trafficking is that, at any particular moment, much of the Ste3p in the cell is in transittraveling through either the secretory pathway on its way to the cell surface or the endocytic pathway on its way to the vacuole. For studying ligand-induced phosphorylation, we routinely use a pulse-chase protocol to focus analysis on the subset of newly synthesized receptor proteins that is surface resident and therefore available for interaction with pheromone. In resting, unstimulated cells, Ste3p shows heterogeneous gel mobility (Fig. 1A). Treatment of these cells with the a-factor ligand results in a reduction in Ste3p mobility, with much of the receptor now running coherently as a single species (Fig. 1A). When protein extracts are subjected to phosphatase treatment prior to immunoprecipitation, the receptor from both a-factor-treated and untreated cells collapses to a single, faster-migrating species (Fig. 1A). Our conclusion is that in resting cells, Ste3p is subject to a constitutive level of phosphorylation that increases when cells are treated with a-factor.

View larger version (44K): [in this window] [in a new window]   FIG. 1.   The Ste3p CTD is subject to both constitutive and ligand-dependent phosphorylation. (A) a-factor treatment leads to increased Ste3p phosphorylation. MAT HIS3p-STE3 yeast cells (NDY414), which have the constitutive HIS3 promoter replacing the pheromone-regulated STE3 promoter, were pulse-labeled for 10 min with [35S]methionine, chased with excess cold methionine and cysteine for 1 min, and either mock treated () or treated with a 1× concentration (see Materials and Methods) of a-factor (a-F) (+) for an additional 15 min. Protein extracts prepared from the labeled cells were treated with potato acid phosphatase (P'ase) (+) or were mock treated () in parallel. Finally, Ste3p was immunoprecipitated with Ste3p-specific antibodies and subjected to SDS-PAGE and autoradiography. Quantitative phosphorimager analysis of this and parallel experiments indicates that the a-factor treatment results in no significant increase or decrease in the amount of radioactive Ste3p recovered, implying no major effect of the pheromone treatment on the rate of Ste3p turnover. (B) Ste3p retained at the cell surface in end4 mutant cells is subject to ligand-induced phosphorylation. Receptor synthesis was induced from a log-phase culture of MAT GAL1p-STE3(M271I, M304I) end4-1 yeast cells (NDY867) with a 45-min period of galactose addition. Glucose was then added to repress further synthesis. After 10 min, cultures were treated for 15 min with a-factor (+) or were mock treated () as described in Materials and Methods. Cells were grown at 30°C; while this temperature is permissive for end4-1 cell growth, endocytosis of both pheromone receptors remains strongly impaired (39, 43). Protein extracts prepared from these cells were subjected to SDS-PAGE and Western analysis with Ste3p-specific antibodies. (C) The Ste3p CTD provides a major locus for both constitutive and ligand-induced phosphorylation. The protein extracts prepared for panel B were treated with CNBr (see Materials and Methods) prior to SDS-PAGE and Western analysis. The M271I and M304I mutations remove methionyl sites of CNBr cleavage and augment the visualization of the CTD fragment: M304I removes a cleavage site internal to the CTD, and M271I eliminates a partial fragment from position 271 to the C terminus. Neither mutation perturbs Ste3p functioning in terms of mating, endocytosis, or phosphorylation (data not shown). To identify the electrophoretic position of a CTD that lacks phosphoryl modification, an additional protein extract was prepared in parallel to the two prepared for panel B. In this case, the requirement for increased concentrations of Ste3 antigen within the protein extract to be subjected to phosphatase digestion was accommodated with a 2-h galactose induction period prior to the 15 min a-factor treatment. (The intact Ste3p derived from this extract showed electrophoretic mobility identical to that of the Ste3p derived from the a-factor-treated culture shown in Fig. 1B) (data not shown). Following CNBr treatment and prior to SDS-PAGE and Western analysis, this sample was treated with phosphatase. For the SDS-PAGE analysis of the CNBr-treated protein extracts, 12% acrylamide gels were used (as opposed to the 8% gels used for the visualization of full-length Ste3p).

Ligand-induced phosphorylation of Ste3p requires the G subunit of heterotrimeric G protein. For the GPCRs, the first cytoplasmic signaling step initiated by the liganded receptor is the dissociation of the heterotrimeric G protein into the  and subunits. The ligand-dependent phosphorylation of the -factor receptor (Ste2p) does not require this step and proceeds unimpaired in cells deficient for the G protein components (56). To test if this is also true for Ste3p, we have used pulse-chase analysis to examine both constitutive and ligand-dependent Ste3p phosphorylation in ste4 cells. STE4 encodes the G component of the G protein. As the subunit is the primary transducer of the ligand signal in yeast, ste4 cells fail to mount a pheromone response. For this analysis, we have used strains that have the endogenous, pheromone-responsive STE3 promoter replaced by the constitutive HIS3 promoter. This strategy not only blocks the induced increase in transcription with a-factor treatment (15) but also, more importantly, eliminates the fivefold reduction in basal STE3 expression seen in resting cells with disruptions of elements of the pheromone signal transduction pathway; the resting activity of this pathway is required for the basal transcription of many pheromone-inducible genes (12). The HIS3 promoter is a relatively weak promoter (52), and HIS3-driven Ste3p expression is found to be two- to threefold higher than expression from the unstimulated STE3 promoter (A. Roth and N. Davis, unpublished results). STE4⁺ and ste4 HIS3_P-STE3 MAT cells were mock treated or treated with one of three different a-factor concentrations for 10 min (Fig. 2A). In the STE4⁺ background, ligand-induced mobility shifts consistent with induced phosphorylation were equivalently seen over the entire 100-fold a-factor concentration range (these concentrations of the pheromone give a graded response when tested for induction of the pheromone-induced transcriptional response; see Fig. 8A) (data not shown). In ste4 cells, the a-factor-induced mobility shift was completely blocked at all concentrations (Fig. 2A). Thus, the ligand-induced phosphorylation of Ste3p quite clearly requires functional G.

View larger version (50K): [in this window] [in a new window]   FIG. 2.   The ligand-dependent phosphorylation of the two pheromone receptors exhibits different requirements for the G subunit. (A) The G subunit is required for the a-factor-induced phosphorylation of Ste3p. MAT HIS3P-STE3 cells (wild type; NDY414) and isogenic ste4 cells (NDY675) were subjected to a pulse-chase protocol similar to that described in the legend to Fig. 1. Following a 10-min pulse with [35S]methionine and a 10-min chase with excess cold amino acids, cultures were treated with four different concentrations of a-factor (a-F) for 15 min: concentrated a-factor (1), a 10-fold dilution of concentrated a-factor (.1), a 100-fold dilution of concentrated a-factor (.01), and no a-factor (0). Ste3p was immunoprecipitated from the labeled protein extracts and then subjected to SDS-PAGE and autoradiography. As a control for the specificity of the anti-Ste3p antibodies used for immunoprecipitation, extracts from the isogenic ste3 strain SY2638 (3) were processed in parallel. For this experiment, electrophoresis was done on a 20-cm gel format, as opposed to the 7.5-cm minigel format used in the other experiments. On this extended gel format, the presentation of Ste3p from pheromone-treated and untreated cells is much different from that on the minigel format (compare wild-type samples with those of Fig. 1A); with the extended format, heterogeneously modified receptor species are more widely spread. (B) The G subunit is not required for -factor-induced modifications of the -factor receptor. The GAL1p-STE2-HA centromeric plasmid pND733 was introduced into the MATa end4-1 strain RH268-1C (wt) or the isogenic ste4 strain NDY795. Following 1 h of galactose-induced receptor expression, glucose was added, and cultures were further treated for 10 min with -factor (-F) added to 106 M (+) or were mock treated () in parallel. Cultures were maintained at 30°C throughout (see the legend to Fig. 1B). Protein extracts prepared from these cells were subjected to SDS-PAGE and then to Western analysis with an anti-HA.11 monoclonal antibody. As a control for cross-reaction of the antibody, extracts from RH268-1C cells transformed by the empty centromeric plasmid vector (no HA) were processed in parallel. The brackets at right indicate the positions of pheromone-dependent Ste2p modifications likely corresponding to ubiquitin-modified Ste2p (16). (C) The pheromone-induced modifications of Ste2p include phosphorylation. Protein extracts from the ste4 cells of panel B were treated with phosphatase (P'ase) (+) or mock treated in parallel with no added phosphatase () as described in Materials and Methods. Extracts were then subjected to SDS-PAGE and Western analysis as described for panel B.

Ste3p phosphorylation is induced when the signaling pathway is artificially activated. To assess if the STE4 dependence reflected a need for induction of the downstream signaling pathway, we have tested if Ste3p phosphorylation could be induced in trans via -factor induction of the Ste2p signaling. While the two receptors normally function in different cell types, both couple to the same downstream signaling components (1). We reasoned that if the activation of the downstream signaling cascade was all that was required for the induced phosphorylation of Ste3p, then its activation through -factor stimulation of Ste2p should also lead to Ste3p phosphorylation. For this experiment, we have constructed the MAT strain NDY598, which artificially expresses both pheromone receptors: Ste3p expressed from its natural, -specific promoter and Ste2p artificially expressed from the HIS3 promoter. In addition, NDY598 also has disruptions of the two -factor-encoding loci, MF1 and MF2, obviating any autocrine Ste2p stimulation (Fig. 3A). Again, we have monitored Ste3p phosphorylation via immunoprecipitation from in vivo pulse-labeled cells. Added -factor is without effect on Ste3p phosphorylation in control MAT strains lacking the -factor receptor (Fig. 3B). In cells that express Ste2p, however, a clear induction of Ste3 phosphorylation is seen (Fig. 3B). Indeed, the induced changes in Ste3p gel mobility induced at two concentrations of the -factor ligand are identical to those seen when the a-factor ligand is added (Fig. 3B). In addition to indicating a role for the activated signaling pathway in the induced phosphorylation of Ste3p, this experiment also demonstrates that liganded a-factor receptor is not the obligate substrate for phosphorylation: phosphorylation proceeds normally even when the normal receptor ligand is not present.

View larger version (21K): [in this window] [in a new window]   FIG. 3.   Ste3p phosphorylation may be induced in trans via -factor stimulation of an ectopically expressed -factor receptor. (A) Experimental design. At left is shown an approximate mechanism for the induced phosphorylation of Ste3p that occurs in wild-type MAT cells treated with a-factor. The binding of a-factor pheromone (gray squares) to the a-factor receptor elicits a G-dependent mechanism which leads to receptor phosphorylation. At right is outlined the logic of an experiment that tests if G protein-dependent signaling activated by the binding of -factor (gray triangles) to Ste2p may substitute for the a-factor requirement in inducing Ste3p phosphorylation. This experiment was done with the MAT HIS3P-STE2 mf1 mf2 strain NDY598. Thus, these cells constitutively express both yeast pheromone receptors but neither of the two pheromones. (B) trans-Activation of G protein signaling bypasses the a-factor requirement for induced phosphorylation of Ste3p. NDY598 cells were labeled within a pulse-chase regimen identical to that used in the experiment shown in Fig. 2A, i.e., 10 min of pulse-labeling followed by 10 min of chase, followed by 15 min of pheromone treatment. Pheromone concentrations used were as follows: concentrated a-factor (aF), 0.1 × 106 M -factor (.1), 10 × 106 M -factor (10), or no pheromone (0). As a control, NDY544 (wt), a strain isogenic to NDY598 but lacking the HIS3P-STE2 construct, was pulse-labeled and treated with 10 × 106 M -factor (10) or mock treated (0). Ste3p was immunoprecipitated from the labeled protein extracts and then subjected to SDS-PAGE and autoradiography. As a control for antibody specificity, an extract from the isogenic ste3 strain NDY746 (3) was processed in parallel.

View larger version (35K): [in this window] [in a new window]   FIG. 4.   Ligand-independent activation of the signaling pathway via G overproduction leads to induced phosphorylation of Ste3p. Two isogenic MAT HIS3P-STE3 strains were used: NDY711 (wt) and NDY692, a version having an integrated GAL1p-STE4 construct. Galactose (2%) was added to cells growing in raffinose medium to induce Ste4p overproduction. At 20 min prior to the indicated times, culture aliquots were removed and subjected to 10 min of pulse-labeling with [35S]methionine, followed by 10 min of chase with excess cold methionine and cysteine. Cultures labeled for the 0-h time point were never exposed to galactose. Radiolabeled Ste3p was immunoprecipitated from protein extracts derived from these cells and then subjected to SDS-PAGE and autoradiography. As a control for antibody specificity, an extract from the isogenic ste3 strain NDY691 (3) was processed in parallel. For this experiment, SDS-PAGE was done with the extended gel format (see the legend to Fig. 2A).

Requirement for the distal components of the signaling pathway. A key element within the pheromone signaling pathway is the MAPK Fus3p. Through phosphorylation of the transcriptional activator Ste12p and the Dig1p-Dig2p corepressor system, Fus3p activates pheromone-dependent transcription (7, 20, 49). Through phosphorylation of the Cdk inhibitor Far1p, Fus3p serves to institute pheromone-dependent cell cycle arrest in G₁ (34). Fus3p is activated as an end point of a cascade of sequential kinase activation steps. Elements within this cascade include the PAK kinase Ste20p, the MAPK kinase kinase Ste11p, and the MAPK kinase Ste7p. Fus3p, together with two kinases immediately upstream, Ste11p and Ste7p, is coordinately bound to the kinase scaffolding protein Ste5p. Fus3p is considered to be the dedicated MAPK for the pheromone signaling pathway (31). However, in fus3 cells, Fus3p function can be partly replaced by Kss1p, a homologous MAPK apparently dedicated to the filamentation response. Due to the redundant actions of Kss1p, testing of the MAPK requirement for various pheromone responses generally is done with strains having both MAPK genes deleted, i.e., fus3 kss1 cells.

fus3 kss1

fus3 kss1

fus3 kss1

View larger version (64K): [in this window] [in a new window]   FIG. 5.   Participation of the pheromone signal transduction chain components in the ligand-induced phosphorylation of Ste3p. The signaling-competent wild-type MAT HIS3P-STE3 strain NDY414 and isogenic strains with disruptions of various elements of the pheromone signaling pathway were pulse-labeled and treated with the concentrations of a-factor described in the legend to Fig. 2A. Following pheromone treatment, labeled Ste3p was immunoprecipitated from protein extracts and subjected to SDS-PAGE and autoradiography. As in previous experiments, the ste3 strain SY2638 (3) was processed in parallel as a control for the specificity of the Ste3p antibodies. (A) The redundant pair of MAPKs, Fus3p and Kss1p, is required for the ligand-induced phosphorylation of Ste3p. The gel mobilities of Ste3p isolated from a-factor-treated wild-type cells, ste4 cells (NDY675), or fus3 kss1 cells (NDY1022) were compared. Four different a-factor (a-F) concentrations were used: concentrated a-factor (1), 10-fold-diluted a-factor (.1), 100-fold-diluted a-factor (.01), or no a-factor (0). (B) Involvement of the PAK kinase homologue Ste20p, the MAPK kinase kinase Ste11p, and the pheromone-dependent transcriptional activator Ste12p in Ste3p ligand-induced phosphorylation. The gel mobilities of Ste3p isolated from a-factor-treated or mock-treated wild-type cells, ste4 cells (NDY675), ste20 cells (NDY647), ste11 cells (NDY693), or ste12 cells (NDY708) were compared. Pulse-labeled cells for each were treated with concentrated a-factor (1) or were mock treated (0). Wild-type cells were additionally treated with a 10-fold dilution of concentrated a-factor (.1). (C) Far1p, a protein required both for pheromone-induced G1 arrest and for morphogenesis, is not required for the ligand-induced phosphorylation of Ste3p. The gel mobilities of Ste3p isolated from a-factor-treated or mock-treated wild-type and far1 cells (NDY953) were compared as described for panel A.

fus3 kss1

Lack of a requirement for pheromone-induced gene expression. A major outcome of pheromone signaling is the induction of new gene transcription mediated by the transcriptional activator Ste12p. Most of the known pheromone responses, e.g., G₁ arrest, the development of a mating projection, and cell-cell agglutination, require this induced transcription. For these responses, ste12 null mutations show the same absolute block as that seen with null mutations in the other pathway genes (47). In Fig. 5B, however, the effect of the ste12 mutation appeared to differ from that of ste4, ste20, or ste11 in terms of the impact upon the induced phosphorylation of Ste3p: ste12 cells showed only a partial block. In the experiment shown in Fig. 6A, we have examined the ste12 defect more extensively, comparing Ste3p phosphorylation in ste12 mutant cells to that in both isogenic wild-type cells and ste11 mutant cells. For ste11 cells, as with the other signaling mutants tested in Fig. 5, induced Ste3p phosphorylation was completely blocked at all pheromone concentrations (Fig. 6A). ste12 cells clearly differed: ligand-induced changes in Ste3p phosphorylation were apparent even at the lowest concentration of a-factor used (Fig. 6A). Thus, in terms of Ste3p phosphorylation, ste12 cells are capable of mounting a significant response to the pheromone. Nonetheless, this response is somewhat impaired relative to that of wild-type cells: in ste12 cells, a portion of the newly synthesized receptor resisted shifting to the hyperphosphorylated form (slow electrophoretic species) over the entire 100-fold a-factor concentration range (Fig. 6A).

View larger version (42K): [in this window] [in a new window]   FIG. 6.   Pheromone-induced gene expression is not required for induced phosphorylation of Ste3p. (A) Ste12p is only partially required. Cells from the MAT HIS3P-STE3 strain NDY414 as well as from the isogenic ste12 (NDY708) and ste11 (NDY693) strains were pulse-labeled, treated with a-factor (a-F), and processed for immunoprecipitation as described in the legend to Fig. 2A. As a control for antibody specificity, a labeled extract from the ste3 strain SY2638 (3) was processed in parallel. (B) Ligand-induced Ste3p phosphorylation does not require new protein synthesis. Following 10 min of pulse-labeling of MAT HIS3P-STE3 cells (NDY753) with [35S]methionine, the culture was split: half was treated with 50 µg of cycloheximide per ml (+ chx) added together with cold chase amino acids for 10 min prior to the 15-min a-factor treatment, and half received no cycloheximide (no chx) but was otherwise treated identically. A labeled extract from the ste3 strain NDY746 was processed in parallel (3). As a test of the efficacy of the cycloheximide treatment, an aliquot of the NDY753 culture was pretreated with 50 µg of cycloheximide per ml 10 min prior to [35S]methionine pulse-labeling (chx35S). Strain NDY753 (isogenic to W303-1B) was used for this experiment instead of strain NDY414, due to the relative insensitivity of NDY414 cells to cycloheximide treatment. NDY753 cells differ somewhat from NDY414 cells in terms of the dose response for a-factor-induced Ste3p phosphorylation: NDY753 cells require at least 10-fold-higher a-factor levels to achieve Ste3p mobility shifts equivalent to those seen in the NDY414 background (compare to Fig. 2A or 5). Panels labeled "no chx" and "+ chx" were derived from the same gel exposed for different periods of time. For the former panel, a threefold-longer autoradiographic exposure was required to attain the level of band intensity equivalent to that in the latter panel. This result likely reflects an effect of cycloheximide treatment on the normally rapid Ste3p turnover rate. Previous work with the -factor receptor (Ste2p) indicated that -factor-induced vacuolar turnover was dramatically slowed in cycloheximide-treated cells, with the receptor accumulating within a prevacuolar endosomal compartment (17). (C) ste12 mutants, which maintain the basal rate of pheromone transcription, restore wild-type Ste3p phosphorylation to ste12 cells. MAT HIS3P-STE3 cells that were either ste4 (NDY675) or ste12 (NDY708) were subjected to the a-factor treatment and pulse-chase described in the legend to Fig. 2A. In addition, some of the ste12 cells were transformed with one of three plasmids: the centromeric plasmid p650, which carries wild-type STE12, or the equivalent plasmid carrying one of two in-frame ste12 deletion mutants (253-335 or 255-354).

Effects of different dominantly activated STE gene constructs. We have shown (Fig. 4) that activation of the signaling pathway by Ste4p overproduction induces Ste3p phosphorylation. Next, we tested the effects on Ste3p phosphorylation of pathway activation at other steps. Three activating constructs, in addition to GAL1p-STE4, were obtained: a GAL1p-STE5-CTM construct, a GAL1p-STE11N construct, and a GAL1p-STE12 construct. Like that of Ste4p, Ste12p overproduction also leads to pheromone-independent activation of the pheromone response (9). The GAL1p-STE11N and GAL1p-STE5-CTM constructs overproduce dominantly activated mutant alleles of STE11 and STE5. STE11N lacks the N-terminal Ste11p regulatory domain (32, 38), and STE5-CTM adds a plasma membrane-targeting transmembrane domain to the Ste5p C terminus (37). All four activating constructs are grossly similar in their effects: all lead to pheromone-independent G₁ arrest, transcriptional activation of FUS1, and partial suppression of the sterility associated with deletion of the pheromone receptor gene.

GAL1p-STE11N

View larger version (36K): [in this window] [in a new window]   FIG. 7.   Effect of dominantly activated STE gene constructs on both pheromone-dependent transcription and induced Ste3p phosphorylation. Four different GAL1-driven constructs were used: the GAL1p-STE4 construct was chromosomally integrated, while the other three were introduced into an isogenic strain as CEN/ARS plasmids: GAL1p-STE5-CTM (pND882), GAL1p-STE11N (pRD-STE11-H3), and GAL1p-STE12(pC1-H6). The wild-type control strain, which lacked an activated signaling allele, carried the CEN/ARS vector pRS316 (46). (A) Each of the activating constructs strongly induces the expression of the pheromone-responsive gene FUS1. The MAT FUS1-LacZ strain SY1937 transformed by the four plasmids mentioned above or the isogenic rad16::GAL1p-STE4 strain NDY787 was treated with 2% galactose for 2 h, following which the amount of -galactosidase expressed was measured (see Materials and Methods). -Galactosidase activity measured from cells expressing the four different dominantly activated STE gene constructs was compared to the baseline level derived from SY1937 cells transformed with pRS316 (con). Reported values represent the average of two separate determinations (for each data point, the two activity measurements differed from one another by less than 20%). A 1-h treatment of SY1937 cells with a 0.1× concentration of a-factor (see Fig. 2A) resulted in the expression of 35 U of -galactosidase activity (data not shown). (B) Differential effect of the activated alleles on the induced phosphorylation of Ste3p. Cells of the MAT HIS3P-STE3 strain NDY711 transformed by the four plasmids mentioned above or of the isogenic HIS3P-STE3 rad16::GAL1p-STE4 strain NDY692 growing in raffinose medium were induced by the addition of 2% galactose for 2 h. The final 20 min of the galactose induction period consisted of 10 min of pulse-labeling and then 10 min of chase with excess cold amino acids. Ste3p was immunoprecipitated from protein extracts prepared from these cells and then subjected to SDS-PAGE and autoradiography. Cells expressing the four different dominantly activated STE gene constructs were compared to NDY711 cells transformed by the empty vector pRS316 (wt). As a control, isogenic ste3 cells (SY1817) were labeled and processed in parallel. The panel on the right is a longer autoradiographic exposure of the two lanes expressing the activated alleles of STE4 and STE11 (poor Ste3p labeling is routinely seen in cells expressing the STE11N construct).

GAL1p-STE11N

GAL1p-STE11N

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Different kinase systems operate on the two pheromone receptors. While the two yeast pheromone receptors share many functional similarities, the present work indicates a clear difference between the two in terms of the mechanisms used for inducing ligand-dependent phosphorylation. For Ste2p, the -factor-induced increase in receptor phosphorylation does not depend upon the activation of the downstream signaling pathway; pheromone-induced phosphorylation is unimpeded in cells with disruptions of the individual G protein subunits or when receptor expression is artificially forced in a/ diploid cells which lack the expression of key signaling components: Gpa1p (G), Ste4p (G), Ste18p (G), and Ste5p (56). For Ste3p, a-factor-induced phosphorylation depends not only on the G protein  subunit (Ste4p) (Fig. 2A) but also on most of the downstream signaling components, including Ste20p (the PAK kinase homologue), Ste5p (the kinase cascade scaffolding protein), and the individual components of the MAPK cascade (i.e., Ste11p, Ste7p, and Fus3p/Kss1p). Gene disruptions of each of these pathway elements share the same phenotype in terms of Ste3p phosphorylationinduced phosphorylation is wholly blocked. For the induced phosphorylation of Ste3p, we conclude that the pheromone signal must be transduced along the signaling pathway from the G protein through the MAPK.

yck1 yck2-2^ts

yck1 yck2

yck1 yck2-2^ts

Regulation of ligand-induced phosphorylation. While Ste3p phosphorylation clearly responds to the action of the pheromone signaling pathway, this response stands out in terms of its lack of a requirement for pheromone-induced gene expression. The transcriptional response is a major outcome of pheromone signaling, and most of the other pheromone responses (e.g., G₁ arrest, polarized morphogenesis, and cell-cell agglutination and fusion) depend on it. Three results presented here indicate that Ste3p phosphorylation differs. First, while gene disruption of most of the components of the pheromone signaling pathway abolishes induced phosphorylation, disruption of STE12 does not: ste12 cells clearly respond to a-factor treatment with increased Ste3p phosphorylation (Fig. 6A). Indeed, the partial impairment seen in ste12 cells apparently does not result from failed transcriptional induction (Fig. 6C). Second, induced Ste3p phosphorylation is not blocked by cycloheximide, indicating that this response does not depend on pheromone-induced new protein synthesis (Fig. 6B). Third, while GAL1p-STE11-N and GAL1p-STE12 constructs strongly and rapidly induce pheromone-responsive transcription (Fig. 7A), neither functions to induce Ste3p phosphorylation (Fig.