|Home | About | Journals | Submit | Contact Us | Français|
Signals transmitted by common components often elicit distinct (yet appropriate) outcomes. In yeast, two developmental options—mating and invasive growth—are both regulated by the same MAP kinase cascade. Specificity has been thought to result from specialized roles for the two MAP kinases, Kss1 and Fus3, and because Fus3 prevents Kss1 from gaining access to the mating pathway. Kss1 has been thought to participate in mating only when Fus3 is absent. Instead, we show that Kss1 is rapidly phosphorylated and potently activated by mating pheromone in wild-type cells, and that this is required for normal pheromone-induced gene expression. Signal identity is apparently maintained because active Fus3 limits the extent of Kss1 activation, thereby preventing inappropriate signal crossover.
Mitogen-activated protein kinases (MAPKs) are found in all eukaryotes and are expressed in virtually all mammalian tissues. The wide variety of stimuli that elicit MAPK activation and the large number of distinct responses that MAPKs regulate raise the issue of how they achieve specific coupling of signal to cellular response (Schaeffer and Weber, 1999; Widmann et al., 1999).
A particularly perplexing aspect of this problem of signal identity is when the same proteins regulate different responses in the same cell. For instance, rat PC12 cells proliferate when treated with epidermal growth factor (EGF) and differentiate when treated with nerve growth factor (NGF). Both stimuli lead to phosphorylation of the ERK2 MAP kinase, and ERK2 activation is required for both responses. A popular model suggests that the magnitude and duration of MAPK activation may specify signal identity in this case; EGF causes a transient activation of ERK2, whereas NGF causes a sustained increase in ERK2 phosphorylation (Marshall, 1995). How this differential MAPK activation is generated, how it is decoded by downstream components, and precisely what role it plays in signal identity are active areas of investigation (recently reviewed by Schaeffer and Weber, 1999; Tan and Kim, 1999).
Haploid cells of the budding yeast Saccharomyces cerevisiae halt cell division and reversibly differentiate into gamete-like cells when mixed with cells of the opposite mating type, or when exposed to purified mating pheromone. In contrast, nutrient-limited cells continue to divide and differentiate into filamentous cells capable of adhering to and invading beneath their substratum (Lengeler et al., 2000). Both mating and filamentous invasive growth require elements of the same MAPK cascade, including the MAPK/ERK kinase (MEK) Ste7 and the MEK kinase (MEKK) Ste11; cells lacking either Ste11 or Ste7 cannot mate and exhibit substantially reduced invasive growth (Roberts and Fink, 1994). The MAPK targets of Ste7 are Kss1 and Fus3. These two MAPKs share ~55% sequence identity with each other and ~50% identity with human ERK1 and ERK2.
How are these contrasting differentiation programs specified by signaling through the same MAPK cascade? A different mechanism than that proposed for PC12 cells has been thought to underlie specificity in yeast mating versus invasive growth (Madhani and Fink, 1998). First, it is thought that the MAPKs have specialized functions, with Kss1 regulating invasive growth and Fus3 regulating mating. Second, it is postulated that each MAPK is activated only by the appropriate upstream signal because it is sequestered into a pathway-dedicated signaling complex by protein-protein interactions.
Several observations indicate that Kss1 and Fus3 do indeed have specialized functions in mating and invasive growth. Kss1 appears to be the principal MAPK that regulates invasive growth. Ste7-mediated phosphorylation of Kss1 is required for efficient invasive growth, and strains deleted of the KSS1 gene are hypoinvasive (Cook et al., 1997; Madhani et al., 1997; Bardwell et al., 1998a). In contrast, Fus3 appears to have a function in invasive growth that is antagonistic to Kss1; strains deleted of the FUS3 gene are hyperinvasive (Roberts and Fink, 1994). Conversely, in the mating pheromone response, Fus3 appears to play the predominant role (Elion et al., 1991).
How might the nature of the signal dictate which MAPK becomes activated, given that both Kss1 and Fus3 are phosphorylated, and thereby activated, exclusively by the MEK Ste7? One possibility is that protein-protein interactions with pathway-specific upstream and downstream components might sequester shared components (such as Ste7) into pathway-dedicated signaling complexes (Whitmarsh and Davis, 1998). For example, the Ste5 scaffold/adaptor protein binds to Ste4, the β-subunit of the pheromone-receptor-coupled G protein, and to Ste11 (MEKK), Ste7 (MEK), and Fus3 (MAPK). As such, Ste5 could channel the pheromone signal from Gβ through the shared MEKK and MEK and into the “mating MAPK” Fus3. At the same time, Ste5 could insulate Fus3 from invasive growth signals. Similarly, a different scaffold could direct invasive growth signals to Kss1 and insulate it from mating signals (Madhani and Fink, 1998; See Figure 1A).
The idea that MAPK signaling specificity is determined by scaffold-like interactions and pathway-specialized MAPKs has also been proposed for other situations where parallel MAPK cascades share components (Posas and Saito, 1997). It has also been postulated to contribute to specificity in mammalian MAPK signaling (Whitmarsh and Davis, 1998). It has had a widespread influence in the primary literature, in reviews (e.g., Schaeffer and Weber, 1999; Widmann et al., 1999), and even in the recent edition of a popular cell biology textbook (Lodish et al., 1999). However, as a hypothesis it is largely untested and is mostly based on indirect evidence.
Several shortcomings are apparent when this model is applied to the paradigm of specificity in yeast mating versus invasive growth. First, no scaffold protein for the filamentous invasive growth cascade has been identified, nor is there evidence of selective pairwise interactions that could, in principal, provide scaffold-like functionality (Xia et al., 1998; Bardwell et al., 2001). Furthermore, Ste5, the scaffold protein for the mating pathway, binds to either Fus3 or Kss1, and does not appear to qualitatively discriminate between them (Choi et al., 1994). Most important, it is not clear that Kss1 is devoted only to invasive growth. Indeed, substantial evidence suggests that Kss1 and Fus3 share a partially overlapping role in mating (Elion et al., 1991; Cherkasova et al., 1999; Farley et al., 1999). For example, cells lacking either Kss1 or Fus3 mate effectively, whereas cells lacking both MAPKs are sterile. However, although cells lacking only Fus3 display some mating defects, cells lacking only Kss1 do not have any obvious deficiencies in pheromone response. It has thus been proposed that Fus3 is the true mating MAPK, whereas Kss1 is an “imposter” that substitutes for Fus3 when Fus3 is absent but normally has no role in mating (Madhani and Fink, 1998; Madhani et al., 1997). This modified model (see Figure 1B) is attractive because it explains why pheromone stimulation does not hyperactivate Kss1-regulated filamentation genes in cells containing Fus3. However, an important supposition of this model (that Kss1 is not activated during mating in cells containing Fus3) has not heretofore been directly examined biochemically.
Here we show that, in fact, Kss1 is phosphorylated and activated during mating in wild-type cells. Further, we confirm and extend a previous observation (Madhani et al., 1997) that Fus3 has a role as a specificity factor that prevents a Kss1-dependent “leak” from the mating pathway into the filamentation pathway. However, we find that this function of Fus3 requires its catalytic activity; this is not predicted by the model that Fus3 prevents signal crossover by physically blocking access of Kss1 to a mating pathway-specific component such as Ste5. Instead, we show that active Fus3 limits the magnitude and duration of Kss1 phosphorylation. We propose that transient versus sustained MAPK activation may encode specificity in this system. As such, signal identity in yeast mating versus invasive growth may be determined by mechanisms that are related to those mechanisms underlying specificity in proliferation versus differentiation in certain mammalian cells.
To examine Kss1 phosphorylation directly, we used an antibody that specifically recognizes the dually phosphorylated isoforms of MAPKs (Khokhlatchev et al., 1997). MAPK activation results from phosphorylation (by a cognate MEK) on a threonine and a tyrosine in a TXY motif located in a surface loop near the active site (the “activation loop”). An antibody raised against the phospho-epitope found in mammalian ERK1 and ERK2 also reacted specifically with the phosphorylated isoforms of Kss1 and Fus3 (Figure 2A, data not shown; see also Bardwell et al., 1998a). In clear contrast to the prediction of the model shown in Figure 1, we found that both Fus3 and Kss1 were rapidly phosphorylated in response to a brief (15 min) treatment with physiological levels of α factor mating pheromone (Figure 2A). Importantly, Kss1 phosphorylation was stimulated by pheromone in cells containing wild-type Fus3. Kss1 and Fus3 displayed very similar dose-response profiles, with roughly half-maximal activation at a concentration of 150 nM added pheromone (Figure 2B).
To further examine the kinetics of MAPK activation, a series of time course experiments were performed (an example is shown in Figure 2C). Over six independent time course experiments, several consistent patterns emerged, although there was some experiment-to-experiment variability in the duration of MAPK activation, perhaps due to subtle differences in the growth or nutrient status of the cells (Cullen and Sprague, 2000). First, both Kss1 and Fus3 were rapidly phosphorylated (within 5 min) following the addition of pheromone, with maximal activation achieved by 15 min; the kinetics of Kss1 and Fus3 induction were indistinguishable. Second, MAPK phosphorylation was usually still high 1 hr after pheromone addition, but was declining by 2–3 hr. Finally, Kss1 phosphorylation often declined roughly 1 hr sooner than Fus3 phosphorylation.
To examine MAPK activation during physiological mating, MAPK phosphorylation was monitored following the mixing of cultures of the two opposite mating types—MATa cells and MATα cells. Following mixing, both MAPKs were rapidly phosphorylated (Figure 2D), just as when synthetic α factor pheromone was added to a cells (e.g., Figures 2A–2C).
As a final means to verify the pheromone-stimulated activation of Kss1, the catalytic activity of Kss1 was directly monitored in an immunoprecipitation-kinase assay. An epitope-tagged version of Kss1 was expressed at a near-endogenous level (from its own promoter on a centromeric plasmid) in strains deleted of the chromosomal KSS1 gene (kss1Δ strains). Cell extracts were prepared, epitope-tagged Kss1 was immunoprecipitated, and its ability to phosphorylate a recombinant substrate was determined (Figure 2E). In a strain containing Fus3, the protein kinase activity of Kss1 was stimulated roughly 5-fold by a brief exposure to physiological levels of mating pheromone. Interestingly, however, in a strain lacking Fus3, the basal activity of Kss1 was increased by 2-fold (compare lane 1 to lane 3). Even more remarkably, pheromone-stimulated Kss1 activity was increased by a further 3- to 4-fold in cells lacking Fus3 relative to stimulated wild-type cells (compare lane 2 to lane 4).
To determine if Kss1 phosphorylation was required for a normal pheromone response, an unphosphorylatable mutant of Kss1 (Kss1AEF; Ma et al., 1995) was expressed at a near-endogenous level in a strain lacking endogenous Kss1. Kss1AEF contains mutations (T183A, Y185F) in the two target residues that are normally phosphorylated by Ste7 (MEK). The resulting FUS3+ kss1AEF strain displayed a clear defect in the expression of a pheromone-inducible reporter gene (FUS1-lacZ). The defect was present in unstimulated cells, across a wide range of pheromone doses, and when a cells were mixed with an equal number of α cells (Figure 3). Phosphorylation of Kss1 may primarily serve to counter the repression of transcription by unphosphorylated Kss1, because a FUS3+ kss1Δ strain displayed increased expression of the FUS1-lacZ reporter relative to wild-type, both in unstimulated cells (Bardwell et al., 1998b) and across a range of pheromone doses (data not shown). A similar mechanism has been proposed for Kss1 regulation of filamentation genes (Cook et al., 1997; Madhani et al., 1997; Bardwell et al., 1998a). Regardless of the precise mechanism, the finding that the prevention of Kss1 phosphorylation leads to impaired expression of mating genes is yet further evidence that Kss1 is normally phosphorylated during the mating pheromone response.
Signaling through the Ste11-Ste7 MAPK cascade is required for the expression of mating genes, such as FUS1, which are driven by pheromone response elements (PREs), and for the expression of filamentation genes, which are driven by filamentation response elements (FREs) (Hagen et al., 1991; Madhani and Fink, 1997). In wild-type haploid cells, FRE-driven transcription is constitutive at a level that is permissive for filamentation. In contrast, most PRE-driven mating genes are weakly expressed in unstimulated cells. Basically, FRE-driven expression is normally “on,” presumably due to a constitutive signal, whereas PRE-driven expression is “off” in the absence of pheromone. Reciprocally, stimulation with mating pheromone potently activates PRE-driven mating genes but does not hyperactivate FRE-driven transcription.
Cells lacking Fus3 are hyperinvasive (Roberts and Fink, 1994). Furthermore, pheromone stimulation results in hyperactivation of FRE-driven transcription in cells lacking Fus3 but not in their wild-type counterparts (Madhani et al., 1997). This aberrant gene-expression pattern has been proposed to result from a “leak” from the mating pathway (which is active at a low level even in the absence of pheromone stimulation) into the filamentous invasive growth pathway (Madhani and Fink, 1998). Indeed, expression of a filamentation-specific, FRE-driven reporter gene was dramatically increased in cells lacking Fus3 (Figure 4A; see also Madhani et al., 1997) and could be further amplified by treating these cells with a physiological concentration of mating pheromone (150 nM; Figure 4B) or by mixing them with the opposite mating type (Figure 4C). In contrast, FRE-driven transcription was not significantly elevated by pheromone treatment of wild-type cells (Figure 4B), even when concentrations as high as 5 μM were used (data not shown), nor was it elevated when wild-type a and α cells were mixed (Figure 4C). As an alternative means to activate the mating pathway, a weakly hyperactive allele of the mating-pathway-specific scaffold/adaptor protein STE5 was overexpressed in both the wild-type and fus3Δ strains (Figure 4D). In both strains, overexpression of STE5T52M caused an increase in the expression of a mating pathway-specific reporter gene (FUS1-lacZ), as previously reported (Hasson et al., 1994). Informatively, in the wild-type strain and particularly in the fus3Δ strain, the expression of the filamentation pathway-specific reporter (FRE-lacZ) was stimulated by the hyperactive STE5 allele. Hence chronic, low-level activation of the mating pathway can cause and potentiate the leak.
To further explore the role of mating-pathway-specific components in the aberrant activation of filamentation in cells lacking Fus3, the STE5 gene was deleted from both wild-type and fus3Δ strains. Deletion of STE5 abolished mating in both strains, as expected (data not shown). Also as previously reported (Roberts and Fink, 1994), we found that cells lacking only Ste5 exhibited undiminished invasive growth relative to wild-type (not shown), although we observed a modest reduction in FRE-driven gene expression (Figure 4D). Informatively, the abnormal high levels of FRE-driven expression seen in cells lacking Fus3 returned to wild-type levels when Ste5 was also absent (Figure 4D). Further, pheromone stimulation of FRE-driven expression was abolished in these cells (not shown). These data are consistent with previous results in which the mating-specific component STE4 (Gβ) was deleted (Madhani et al., 1997), and they indicate that the leak requires mating-specific components and thus originates from the mating pathway.
In cells lacking the FUS3 gene, there is an increase in the level of basal and pheromone-stimulated protein kinase activity of Kss1 (Figure 2E). Further, the hyperinvasiveness and increased FRE-driven expression of fus3Δ cells returns to wild-type levels in cells that also lack Kss1 (Madhani et al., 1997; Roberts and Fink, 1994). Reintroduction of the wild-type KSS1 gene on a plasmid into a fus3Δ kss1Δ double mutant restored the hyperinvasive phenotype (Figure 5A and data not shown). Reintroduction of the unphosphorylatable mutant of Kss1 (Kss1AEF), however, did not restore the hyperinvasive phenotype, nor did reintroduction of a phosphoryla-table but catalytically inactive mutant of Kss1 (Kss1Y24F; Ma et al., 1995) (Figure 5A). Collectively, these findings indicate that the leak from the mating pathway goes through Kss1, and that phosphorylation of Kss1 and its consequent protein kinase activity are required to transmit the leak.
To determine if phosphorylation and activation of Fus3 was required for its ability to block the leak, plasmid-borne alleles of the FUS3 gene were reintroduced into a fus3Δ strain (JCY120). These alleles included the null allele (or empty vector), the wild-type allele, an allele encoding an unphosphorylatable mutant (Fus3AEF, equivalent to Kss1AEF), and an allele encoding a phosphorylatable but catalytically inactive mutant (Fus3K42R) (Gartner et al., 1992). When expressed at near-endogenous levels, only wild-type Fus3 was able to complement the phenotype of the fus3Δ strain and restore wild-type levels of invasive growth and FRE-driven expression (data not shown). As this result conflicted with a previous report (Madhani et al., 1997), several approaches were used to verify it. First, we confirmed by immunoblotting that the plasmid-expressed wild-type and mutant Fus3 proteins were produced at levels roughly equivalent to endogenous Fus3 (data not shown). Second, we sequenced the complete open reading frames of the wild-type allele as well as both mutant alleles and found them to be identical to the previously reported nucleotide sequence (Elion et al., 1990) except, of course, for the bases encoding the K42R and T180A/Y182F mutations (data not shown). Third, we overproduced the wild-type and mutant alleles of Fus3 from a multicopy plasmid. Overproduction of wild-type Fus3 in the fus3Δ strain resulted in a marked inhibition of invasive growth, comparable to the minimal invasiveness of a ste7Δ strain (Figure 5B). In stark contrast, overproduction of either mutant of Fus3 in the fus3Δ strain had little or no effect on invasiveness or FRE-driven expression (Figure 5B, C). Immunoblotting revealed that the level of Fus3 overproduction achieved (for both the wild-type and mutant alleles) was between 50- and 200-fold (data not shown). Thus, we conclude that the catalytic protein kinase activity of Fus3 is required for its ability to prevent a leak from the mating pathway from stimulating invasive growth.
In principal, active Fus3 could inhibit filamentation by limiting Kss1 phosphorylation levels, by acting downstream of Kss1, or by acting on a parallel pathway. To assess the first of these possibilities, Kss1 phosphorylation was monitored in the presence and absence of Fus3. First, the basal phosphorylation level of Kss1 (i.e., in the absence of pheromone stimulation) was determined under conditions optimal for invasive growth (following 2 days growth on a plate). As shown in Figure 6A, phosphorylation of Kss1 was detected in an unstimulated wild-type strain. As expected, this phosphorylation was abolished when Ste7, the MEK that activates Kss1 (and Fus3), was absent. Notably, Kss1 phosphorylation was considerably increased when Fus3 was absent.
Second, the effect of Fus3 overproduction on Kss1 phosphorylation levels was monitored. Overproduction of wild-type Fus3 inhibited both the basal and pheromone-stimulated phosphorylation of Kss1, compared to fus3Δ strain carrying an empty vector as a control (Figure 6B, compare lanes 2 and 3 to lanes 4 and 5). Overproduction of “kinase-dead” Fus3K42R, in contrast, had no effect on Kss1 phosphorylation (lanes 6 and 7), correlating with its inability to block the mating → filamentation leak (see Figures 5B and 5C). Overproduction of wild-type Fus3 resulted in an increase in the absolute amount of phosphorylated Fus3, consistent with mass action (Figure 6B, compare lane 1 to 5). Revealingly, phosphorylation of the mutant Fus3K42R protein was markedly elevated relative to wild-type Fus3. These data indicate first, that Fus3 kinase activity partially inhibits the phosphorylation of Kss1, and second, that Fus3 kinase activity also partially inhibits its own phosphorylation. A likely explanation is that Fus3 activates a negative feedback mechanism(s) (Gartner et al., 1992).
Finally, the time course of MAPK phosphorylation in response to pheromone was monitored (Figure 6C). In the wild-type strain, both Kss1 and Fus3 phosphorylation were elevated at 1 hr after pheromone treatment but had declined by 2 hr, in agreement with the pattern discussed above with reference to Figure 2C. Notably, however, Kss1 phosphorylation was considerably elevated in the fus3Δ strain after 1 hr exposure to pheromone. Even more striking, the duration of Kss1 phosphorylation was substantially lengthened; Kss1 phosphorylation at 4 hr was still well above basal, and roughly comparable to the peak level seen in wild-type cells.
In summary, the hyperphysiological levels of invasive growth and filamentation-specific gene expression seen in the absence of Fus3 are accompanied by parallel increases in the extent of Kss1 phosphorylation, strongly suggesting that Fus3 inhibits invasive growth and maintains signal identity, at least in part, by limiting Kss1 phosphorylation (Figure 7).
The MAPK cascade regulating mating and invasive growth in S. cerevisiae provides a paradigm for similar pathways involved in growth, development, and disease in mammals, as well as for those pathways controlling differentiation and virulence in fungal pathogens of animals and plants (Widmann et al., 1999; Lengeler et al., 2000; Dohlman and Thorner, 2001). It is one of several important model systems for investigating aspects of specificity in MAPK cascade signaling (Schaeffer and Weber, 1999; Tan and Kim, 1999).
Herein we examined the regulation of the Kss1 and Fus3 MAPKs by mating signals and the consequences for specificity, and we presented three main findings. First, contrary to expectations, endogenous Kss1 was phosphorylated and activated in a wild-type strain following treatment with physiological concentrations of mating pheromone and during physiological mating (Figure 2); indeed, the phosphorylation of Kss1 was required for normal pheromone-stimulated gene expression (Figure 3). We conclude that Kss1 and Fus3 normally work together to regulate gene induction during mating.
Second, confirming and extending an important insight first provided by Madhani et al. (1997), we found that Fus3 acts as a specificity factor that blocks the mating signal from leaking, via Kss1, into the filamentous invasive growth pathway (Figure 4; Figure 5A). Our investigation, however, did not support the notion that Fus3 has a significant kinase-independent specificity function. Instead, we found that the catalytic activity of Fus3 was required to block the Kss1-dependent leak (Figures 5B and 5C). In a perhaps related phenomenon, active Fus3 inhibits the transposition of an endogenous retrovirus by negatively regulating the invasive growth pathway (Conte and Curcio, 2000).
Our third main finding was a series of observations which indicated that active Fus3 limits the magnitude and duration of Kss1 activation (Figure 6). Thus, signal identity may be maintained not by preventing access of Kss1 to the mating pathway but rather by limiting the extent of Kss1 activation by this pathway. In unstimulated wild-type cells, a low level of Kss1 phosphorylation was observed. This weak, sustained activation of Kss1 was apparently able to support normal levels of FRE-driven expression and invasiveness; deletion of STE7 eliminated Kss1 phosphorylation and crippled invasiveness and FRE-driven expression. A strain lacking Fus3 displayed a modest (2- to 3-fold) increase in basal Kss1 phosphorylation and activity. This low but sustained augmentation of Kss1 activity was apparently sufficient to support the increased invasiveness and FRE-driven expression characteristic of this strain. Yet, Kss1 phosphorylation and activation was increased to a greater extent (~5-fold) following pheromone stimulation of wild-type cells, but there was no corresponding increase in FRE-driven expression. In this case, however, Kss1 activation was relatively transient, typically returning to levels at or below baseline by about 2 hr after the addition of pheromone or the start of mating. In contrast, when fus3Δ cells were treated with pheromone, both the magnitude and duration of Kss1 activation were markedly expanded relative to cells containing Fus3. In this circumstance, signal identity was lost as FRE-driven expression was amplified by mating pheromone.
It appears, then, that a dose-dependent activation of Kss1 by pheromone is required for normal induction of mating genes, but too much (or too long an extent of) activation leads to a spillover of the mating signal that causes the erroneous activation of filamentation genes. The mechanism that prevents this signal crossover is robust, as it is able to resist considerably hyperphysiological levels of mating pheromone. Fus3 appears to be well suited to act as such a rheostat, as it is activated with a dose dependency that closely parallels that of Kss1. Active Fus3 initiates a feedback signal that downregulates the phosphorylation of both MAPKs. As a consequence, the extent of Kss1 activation by pheromone is restricted, preventing the inappropriate stimulation of filamentation genes and invasive growth (Figure 7). Fus3 may positively regulate the expression or activity of a negative regulator, such as a MAPK phosphatase, and/or may feed back on the pathway upstream of MAPK phosphorylation. That Fus3, but not Kss1, can activate this feedback control highlights the fact that differences in the substrate preferences of the two MAPKs also contribute to specificity.
Several mechanistic frameworks have been applied to the enigma of signal identity (how signaling through identical components can lead to distinct responses). When the same components regulate unique fates in different cells, tissue-specific factors can participate (Simon, 2000). This solution does not pertain to the sharing of components between pathways within a cell, however. In this case, one possibility is that a given response might require parallel signals through several branches of a signaling network (Schaeffer and Weber, 1999; Tan and Kim, 1999). This situation is not relevant to the example studied herein, since activation of the yeast mating/filamentation MAPK cascade (by expression of constitutively active mutants of Ste11 or Ste7) is sufficient to stimulate both PRE- and FRE-driven expression (Stevenson et al., 1992; Madhani et al., 1997). Another way that different signals can flow through the same components is if signal identity is encoded by the magnitude, duration, or frequency of the signal (Marshall, 1995; Frey et al., 2000). Yet a fourth possibility—that shared components may be sequestered into pathway-specific complexes by scaffolding or anchoring proteins (Pawson and Scott, 1997; Whitmarsh and Davis, 1998)—had heretofore been thought to largely account for the fidelity of yeast mating versus invasive growth MAPK signaling.
The results presented herein instead suggest that specificity is encoded in the magnitude and duration of the activation of the Kss1 MAP kinase. Hence, the yeast model may bear a closer resemblance to the paradigm of proliferation versus differentiation in mammalian cells, as exemplified by the PC12 system, than was previously appreciated. Indeed, in PC12 cells, MAPK-mediated feedback control is one of several possibilities that have been proposed to explain transient versus sustained MAPK activation (Brightman and Fell, 2000). MAPK-mediated feedback regulation may also help insulate the mating MAPK cascade from the osmotic stress pathway (O'Rourke and Herskowitz, 1998), again providing an alternative to the scaffold hypothesis (Posas and Saito, 1997).
Neither in mammals nor model organisms is it understood how transient versus sustained MAP kinase activity is interpreted in the nucleus to generate different patterns of gene expression. Presumably, the transcription factors that regulate mating and filamentation genes are differentially sensitive to the magnitude and duration of MAPK phosphorylation. PREs are mostly off when the magnitude of MAPK activation is low, but respond dramatically to a transient increase in MAPK phosphorylation. Conversely, FREs are mostly on when MAPK activity is low, and only respond to sustained changes in Kss1 phosphorylation. This discrimination is even more remarkable in view of the fact that the transcription-factor complexes bound to PREs and FREs apparently share most of their components (Cook et al., 1996; Madhani and Fink, 1997; Tedford et al., 1997; Bardwell et al., 1998b). As such, it is perhaps not surprising that the global gene expression patterns underlying distinct responses can be subtle (Fambrough et al., 1999; Roberts et al., 2000).
Strains JCY100 (MATa, STE+ KSS1+ FUS3+; Σ1278b lineage), and the isogenic derivatives JCY110 (kss1Δ::hisG), JCY120 (fus3Δ::TRP1), JCY130 (kss1Δ::hisG fus3Δ::TRP1), and JCY107 (ste7Δ::ura3) have been described (Cook et al., 1997). LFY105 (ste5Δ::LEU2) and LFY125 (fus3Δ::TRP1 ste5Δ::LEU2) were derived from JCY100 and JCY120, respectively, using a ste5::LEU2 allele obtained by PCR using genomic DNA from strain SY1492 (Stevenson et al., 1992) as template and primers LF15 (GATAACACGTTGTTT GAACATCGACAAG) and LF16 (CAGAACTAGACCTTCTGCAGAGA GATC). Strain DC17 (MATα his1) was used as the α strain for mixing experiments.
The FUS3 upstream region was amplified by high-fidelity PCR by the use of genomic DNA from strain JCY100 as the template and primers LB139 (GCGGGATCCATTATTTTCCTTTCTTTTTCTCTG) and LB140 (GCGAATTCAGGGGCGGTAGTGCTTGTAGTC). The resulting fragment was cut with EcoRI and BamHI and inserted into the corresponding sites in YCpU (Cook et al., 1997), thereby yielding YCpU-FUS3P. The wild-type FUS3 ORF and a catalytically inactive allele thereof (K42R) were amplified by PCR from plasmids YEpT-FUS3 (Bardwell et al., 1996) and pJB304 (Brill et al., 1994), respectively, using primers LB136 (GCGGGATCCACCATGCCAAAGAGA ATTGTATACA) and LB137 (CCGGTCGACCATATGTACATATGTAT ATGTGTACGTA). The resulting fragments were cut with BamHI and SalI and inserted into the corresponding sites of YCpU-FUS3P, thereby yielding YCpU-FUS3 and YCpU-fus3(K42R). YCpU-fus3(AEF) was constructed by oligonucleotide-directed mutagenesis (Quick-change; Stratagene) using primers LF10 (CAGCAAAGCGGCATGG CCGAGTTTGTGGCCACACG) and LF11 (CGTGTGGCCACAAACTC GGCCATGCCGCTTTGCTG). The three FUS3 alleles were subcloned into the high-copy (2 μ) vector YEpU (Bardwell et al., 1998b) as EcoRI-SalI fragments. All PCR and mutagenesis products were completely sequenced.
YCpU-KSS1 and the Y24F and T183A/Y185F (a.k.a. AEF) alleles thereof have been described (Bardwell et al., 1998a). YCpU-KSS1H6myc is a low copy (CEN) URA3-containing vector in which the KSS1 open reading frame, modified by a C-terminal hexahistidine and myc-epitope tag, is expressed from the KSS1 promoter. It was constructed by replacing the BamHI-SphI fragment in YCpU-KSS1 with the corresponding fragment excised from pGEM4Z-KSS1H6myc (Cook et al., 1996). YCpU-kss1(Y24F)H6myc was constructed by replacing the EcoRI-XbaI fragment in YCpU-KSS1H6myc with the corresponding region of YCpU-kss1(Y24F). Endogenous-level protein production from these constructs was verified by immunoblotting.
The mating-specific, PRE-driven reporter plasmids YEpU-FUS1Z (Bardwell et al., 1998b) and YEpL-FUS1Z (Bardwell et al., 2001) have been described; the former was used for the experiments shown in Figures 4B and 4C, and the latter in all other cases. The filamentation-specific, FRE-driven reporter YEpU-FT1Z was constructed following a method described previously (Cook et al., 1997) by the use of a double-stranded oligonucleotide containing the FRE from the upstream region of TEC1 (Madhani and Fink, 1997), which was generated by annealing TEC1FREw (TCGAGTCTAGATGAAACACGCA CATTCCG) to TEC1FREc (CTAGCGGAATGTGCGTGTTTCATCTA GAC). To make YEpL-FT1Z, the URA3 gene in YEpU-FT1Z was replaced by the LEU2 gene using a method described previously (Bardwell et al., 1998a). YEpU-FT1Z was used for the experiments shown in Figures 4B and 4C and part of of4D;4D; YEpL-FT1Z was used in all other cases.
pDJ174 (YEpU-STE5Hyp), containing the STE5T52M allele, is described elsewhere (Hasson et al., 1994).
Yeast were grown in standard media at 30°C with rotary shaking to mid exponential phase (A595 of ~0.8). Pheromone was added directly to growing cultures with no additional manipulation. Harvesting of yeast cultures for biochemical analysis and extract preparation was as described (Bardwell et al., 1996), except that a mixture of commercially available cocktails of protease and phosphatase inhibtors (Sigma catalog numberss P-8215, P-2850, and P-5726) was used. Portions (250 μg) of clarified cell extracts were precipitated with trichloroacetic acid with standard methods and resuspended, and 100 μg was loaded per lane onto a 10% SDS-PAGE gel. Electorphoresis, immunoblotting, and immunostaining were essentially as described (Bardwell et al., 1996). The antiactive MAPK antibody (NEB, catalog number 9101) was used at a dilution of 1:500 with overnight incubation at 4°C.
Reporter gene assays for cells growing in liquid cultures (Figures 3, 4B, and 4C) were as described (Cook et al., 1997). For the experiments shown in Figures 4A, 4D, and and5,5, cells were harvested from plates selective for plasmid maintenance after growth at 30°C for 24 hr. Extracts were then prepared and analyzed as above.
The coimmunoprecipitation kinase assay was performed as described elsewhere (Bardwell et al., 1996; Cook et al., 1996), except that the pheromone concentration was lowered to 150 nM. Kinase reactions contained 20 mM MOPS (pH 7.2), 10 mM MgCl2, 1 mM EGTA, 25 mM β-glycerophosphate, 1 mM sodium orthovanadate, 1 mM dithiothreitol (DTT), 10 μM ATP, 10 μCi [γ-32P]ATP, and 0.5 μM substrate. A fusion of glutathione-S-tranferase to the N-terminal third of Ste7 (GST-Ste71–172; Bardwell et al., 1996) was used as a recombinant substrate. Both Kss1 and Fus3 efficiently phosphorylate the N terminus of Ste7 (Bardwell et al., 1996), although the significance of this feedback phosphorylation remains unclear. Reactions were for 20 min at 30°C; they were analyzed by SDS-PAGE and quantified on a phosphoimager.
For generous gifts of reagents, we thank Gerald Fink, Duane Jenness, Hiten Madhani, George Sprague, and Jeremy Thorner. This work was supported by a Burroughs Wellcome Foundation New Investigator Award, by a Beckman Foundation Young Investigator Award, and by National Institutes of Health Grant GM60366 (all to L.B.), and by National Institutes of Health Training Grant GM07311 (to L.J.F.).