(Received for publication, November 15, 1996, and in revised form, March 7, 1997)
From the Institute for Biochemistry and Molecular Cell Biology, Ludwig Boltzmann Forschungsstelle, University of Vienna, Dr. Bohr, Gasse 9, 1030 Vienna, Austria
Basal and induced transcription of pheromone-dependent genes is regulated in a cell cycle-dependent way. FUS1, a gene strongly induced after pheromone treatment, shows high mRNA levels in mitosis and early G1 phase of the cell cycle, a decrease in G1 after START and again an increase in S phase. Overexpression of CLN2 was shown to repress the transcript number of pheromone-dependent genes (1). We asked whether the activities of components of the mating pathway fluctuate during the cell cycle. We were also interested in determining at what level Cln2 represses the signal transduction machinery. Here we show that the activity of the mitogen-activated protein kinase Fus3 indeed fluctuates during the cell cycle, reflecting the oscillations of the gene transcripts. CLN2 overexpression represses Fus3 kinase activity, independently of the phosphatase Msg5. Additionally, we show that the activity of the MEK Ste7 also fluctuates during the cell cycle. Increased Cln2 levels repress the ability of hyperactive STE11 alleles to induce the pathway. G protein-independent activation of Ste11 caused by an rga1 pbs2 mutation is resistant to high levels of Cln2 kinase. Therefore our results suggest that Cln2-dependent repression of the mating pathway occurs at the level of Ste11.
Mating between cells of the two opposite haploid cell types of Saccharomyces cerevisiae requires recognition of the mating partner through secreted pheromones. In response to pheromone both mating partners arrest their cell cycle in G1 (2, 3) and induce a set of pheromone-dependent genes (4). They undergo morphological changes (shmoo formation), which enables two cells of opposite mating type to fuse and form a diploid zygote. Mating is restricted to a short period in the G1 phase of the cell cycle (5-8). Cells which have passed a point called START in G1 are committed to undergo a new cell cycle and cannot respond to mating pheromone until they reach G1 again (9, 10). Transition through START is mediated by the function of the G1 cyclins Cln1, Cln2, and Cln3, which are the regulatory subunits of the Cdc28 kinase (11-15). In response to mating pheromone the activity of the Cdc28-Cln kinase gets inhibited, so that the cell cannot pass START (16). Mating pheromone induces a MAPK1 pathway which ultimately results in the phosphorylation of the putative Cdk inhibitor (cyclin-dependent kinase inhibitor) Far1 (17, 18). Phosphorylation of Far1 causes its association with the Cdc28-Cln1 and Cdc28-Cln2 kinases and thereby inhibits their ability to drive the cell through START (17-19). Both mating partners are arrested in the same stage of the cell cycle when they fuse. This mechanism ensures the correct ploidy of the zygote.
Transcription of pheromone-inducible genes such as FUS1, STE2 (1, 19), SST2, STE12, and MFA2 (20) has been shown to fluctuate during the cell cycle. It reaches its maximum during late mitosis and early G1, decreases drastically in late G1 around START, and increases again in the G2 phase of the cell cycle (1, 20). The transcript of FAR1, which encodes a putative cyclin-dependent kinase inhibitor, also fluctuates during the cell cycle, but in contrast to other pheromone-inducible genes, its transcription in G2/M is Mcm1-dependent (20). The stability of the Far1 protein is also regulated during the cell cycle. Far1 protein can be detected only in early G1 cells and is rapidly degraded in other stages of the cell cycle (10). This ensures that the cell cycle arrest in response to pheromone occurs in early G1 only.
The cell cycle regulated restriction of pheromone-dependent gene transcription may reflect an important mechanism ensuring that the responsiveness of the cell to mating pheromone is maximal in early G1. The decrease of pheromone-dependent gene transcription in late G1 correlates with an increase in the appearance of G1 cyclin. Overexpression of CLN2 represses FUS1 transcription (1). Preliminary epistasis experiments demonstrated that the repression occurs downstream of the receptor (1) and that it involves neither Sst2 (implicated in recovery from pheromone-induced G1 arrest) nor the carboxyl-terminal part of the pheromone receptor (implicated in desensitization).
Potential targets of Cln2-mediated repression are the components of the
pheromone-induced MAPK pathway. This signal transduction pathway has
been the focus of several reviews (4, 21, 22). In short, activation is
mediated by the binding of pheromone to a seven-transmembrane receptor
coupled to a heterotrimeric G protein. Upon pheromone induction G
dissociates and releases G
(23-27). G
transmits the signal
to the Ste20 kinase (28, 29) by an as yet unknown mechanism involving
the GTPase Cdc42, the guanine nucleotide exchange factor Cdc24 (30, 31)
and the GTPase-activating protein Rga1 (32). Downstream of Ste20 the
sequential activation of several protein kinases, tethered together by
the scaffold protein Ste5, propagates the signal (28). These kinases
belong to the family of highly conserved MAPK pathways (21) and
function in the linear order Ste11, Ste7, and Fus3/Kss1. The MAPKs Fus3 and Kss1 are thought to activate the transcription factor Ste12, which
mediates the transcriptional induction of
pheromone-dependent genes (33-35).
In this work we investigate which components of the pheromone-induced signal transduction pathway are regulated in a cell cycle-dependent way. We show that the activity of the MAPK Fus3 fluctuates during the cell cycle. Furthermore, Fus3 kinase activity is repressed by overexpression of CLN2. The repression occurs independent of the function of the phosphatase MSG5, which dephosphorylates Fus3 (36). The activity of the MEK Ste7 was shown to respond to the cell cycle as well. By the use of hyperactive alleles of STE11 and a deletion of RGA1, we conclude that CLN2 represses the mating pathway at the level of Ste11. The studies in this work represent an important contribution to the understanding of the regulation of a signaling cascade by the cell cycle and the phenomenon described here may be relevant in the regulation of MAPK pathways in higher eucaryotic cells.
The S. cerevisiae strains utilized
in this study are listed in Table I. Standard yeast
techniques were used as described in Sherman et al. (37).
YEP (complete) medium, minimal synthetic medium, and supplements are
described by Hicks and Herskowitz (38). Yeast transformations were
performed either as described by Ito et al. (39) and Beggs
(40), or by one-step transformation (41). Induction of S. cerevisiae cells with pheromone required 0.3 µg/ml -factor
for bar1 strains and 1 µg/ml for BAR1 strains. Unless otherwise indicated cells were induced for 10 min with pheromone. For the expression of the galactose inducible GAL1-10 CLN2 construct cells were pregrown on YEP medium plus 2%
raffinose. The GAL1-10 promoter was induced for 2.5 h
with galactose, the control was either grown in YEP plus 2% raffinose,
or, where indicated, glucose was added to a final concentration of
2%.
|
Standard DNA manipulations were performed according to Sambrook et al. (42) or Ausubel et al. (43). In Table II the plasmids used in this study are listed. The GAL1-10 CLN2 plasmid K2573 was cut with EcoRV and integrated into the genome. The integration was checked by Southern blotting.
|
Far1 was expressed in E. coli from
the plasmid GA1896, which contains the NH2-terminal
fragment of Far1 (amino acids 50-301) under the control of the
isopropyl -D-thiogalactopyranoside-inducible T7
promoter. Expression and purification of the protein are described in
Peter et al. (17). Fus3-R42 was expressed as a
glutathione S-transferase fusion protein from the plasmid
GA1944 in E. coli and purified as described in Errede et al. (44).
Fus3 kinase assays were done as
described in Peter et al. (17), except that the reactions
were done in 6 µl of HBII buffer, where 3 µCi of
[-32P]ATP (6000Ci/mmol), 0.5-2 µl (0.1 µg/µl)
of Far1 substrate, and 25 mM MOPS, pH 7.2, up to a final
volume of 16 µl were added. The reactions were carried out at
30 °C for 30 min.
The strains used contain the Myc
epitope-tagged Ste7 wild-type protein (pNC318) or an inactivated
version of Ste7 (pNC318-R220) on a centromeric plasmid under the
control of the CYC1 promoter (45). 70-ml cultures were grown
to an OD600 around 0.8 in YEP plus 2% sucrose.
cdc15-2 and cdc4-1 cells were arrested as
described. Unless otherwise indicated everything was done at 4 °C.
Cells were harvested by centrifugation, washed with kinase extract
buffer (KEB) (50 mM Tris, pH 7.5, 150 mM NaCl,
5 mM EDTA, O.1% Nonidet P-40, 50 mM sodium
fluoride, 30 mM
Na2H2P2O7, 15 mM 4-nitrophenyl phosphate, O.1 mM
orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 40 µg/ml aprotinin, and 20 µg/ml leupeptin), and resuspended in 200 µl of breakage buffer (KEB without sodium fluoride and
Na2H2P2O7). Cells were
lysed by vortexing them with glass beads 3 times for 4 min. The
extracts were cleared by centrifugation once for 5 min and twice for 15 min at 14,000 rpm. The protein concentration was determined using the
Bio-Rad protein assay as recommended by the manufacturer. 300 µg of
protein extract were used for the immunoprecipitation. Ste7 was
precipitated with approximately 10 µg of anti-Myc 9E10 antibody for
1.5 h. 20 µl of protein A-Sepharose beads in a 1:1 suspension
(preincubated in breakage buffer) were added. After 1 h incubation
the beads were washed 5 times in breakage buffer, and 3 times in HEPES,
25 mM, pH 7.5. For the kinase assay, 6 µl of kinase assay
buffer (25 mM HEPES, pH 7.5, 15 mM
MgCl2, 5 mM EGTA, 1 mM
dithiothreitol, 0.1 mM orthovanadate, 15 mM
phenylmethylsulfonyl fluoride, 15 mM 4-nitrophenyl
phosphate, 40 µg/ml aprotinin, and 20 µg/ml leupeptin) were added
to the beads. 0.5 µg of E. coli purified
Fus3-R42 as a substrate, 0.2 µl of
[-32P]ATP (10 µCi/µl), 1 µl of HEPES, 250 mM, 2 µl of 1 mM ATP, and H2O up
to a final volume of 20 µl were added. The kinase assay was incubated
at 30 °C for 30 min. The reaction was stopped by adding SDS sample
buffer (46) and boiling the extracts for 4 min. Phosphorylation of the
substrate was detected by SDS-polyacrylamide gel electrophoresis
analysis on a 9.5% polyacrylamide (28:2, acrylamide-bisacrylamide) gel
(46).
Yeast total RNA
preparation was performed as described by Cross and Tinkelenberg (47).
RNA was separated on formaldehyde-agarose gels and transferred to
Boehringer Mannheim nylon membranes (as recommended by the
manufacturer). Northern blots were processed as described by Price
et al. (48). The DNA fragments used as probes were labeled
by random priming and isolated as follows: SST2, a
ClaI/HapI fragment was cut out from the
SST2 coding region; CMD1, with the oligos
5-GTCCTCCAATCTTACCGAAG-3
and
5
-TTACAAGTAGAATCCATTTAGATAACAAAGCAGCG-3
the CMD1 coding
region was isolated by PCR from genomic DNA; CLN2, a
SalI/ClaI fragment was cut out from the
CLN2 coding region; CDC28, an
EcoRI/PstI fragment from the plasmid K2354,
containing a GAL1-10 CDC28 construct (by A. Amon) was
isolated; FUS1, the HindIII/SalI
fragment of the FUS1 coding region was isolated; STE3, a HindIII fragment from the STE3
coding region was isolated; ACT1 was isolated as a
XhoI/KpnI fragment from the ACT1
coding sequence.
DNA content of propridium iodide-stained cells was measured on a Becton-Dickinson FACScan as described by Lew et al. (49) and Epstein and Cross (50).
cdc15-2 and cdc4-1 Cell Cycle ArrestCells were grown at permissive temperature at 24 °C in the appropriate medium from OD600 0.2 to 0.8. The cultures were diluted to OD600 0.4 and arrested for 3-4 h at 37 °C. The arrest was checked by microscopic examination and determination of the budding index.
cdc15-2 Release Experimentscdc15-2 cells were arrested as described above. The cultures were released by diluting them with ice-cold YEPD to OD600 O.5 and further cooling them down on ice till the temperature was 24 °C. They were released at 24 °C. Aliquots were taken at the indicated time points for kinase assays, FACS analysis, RNA preparation, protein extracts for Western blots, and for the determination of the budding index.
The
transcripts of pheromone-dependent genes fluctuate during
the cell cycle (1, 19, 20). To determine whether the transcriptional
fluctuations are the consequence of differences in the activity of the
MAPK Fus3, we expressed FUS3 constitutively from the
TPI1 promoter to examine kinase activity independently of
its transcriptional regulation during the cell cycle. Kinase assays
from distinct cell cycle arrested cells were performed with a Myc
epitope-tagged Fus3. Fus3 was immunoprecipitated with 9E10 anti-Myc
monoclonal antibody. Cells harboring a temperature-sensitive mutation
in the CDC15 or the CDC4 gene were arrested at
the restrictive temperature in late mitosis or late G1,
respectively. In vitro Fus3 kinase assays using bacterial
recombinant Far1 as a substrate were performed under pheromone-induced
conditions (Fig. 1A) and uninduced conditions
(data not shown). According to the fluctuations observed for
pheromone-dependent genes, Fus3 MAPK activity was expected
to be high in cells arrested in late mitosis, and low in cells arrested
in late G1. As can be seen in Fig. 1A, Fus3 kinase activity shows the same cell cycle-dependent
pattern as described for the FUS1 transcript (1). Fus3
activity is high in cdc15-2 arrested cells (lane
3), and low in cdc4-1 arrested cells (lane
5).
To further analyze the cell cycle-dependent regulation of basal Fus3 kinase activity we performed cdc15-2 release experiments in the absence of pheromone. The cultures were arrested at restrictive temperature in late mitosis and released from the block at permissive temperature. The arrest was checked by microscopic examination and determination of the budding index (data not shown). At the indicated time points, samples were taken for kinase assays (Fig. 1B) and Northern analysis (Fig. 1C). The activity of Fus3 was examined by in vitro kinase assays. As seen in Fig. 1B the activity of Fus3 shows the same cell cycle-dependent regulation which had been observed for the transcripts of mating dependent genes.
Differences in the activity of Fus3 could have been due to differences in Fus3 protein levels. To rule out this possibility, Western blot analysis was performed. Aliquots from the cdc15-2 release experiment in Fig. 1, B and C, were used for the preparation of cell extracts. The Western blot was probed with anti-Fus3 and anti-Cdc28 polyclonal antibodies. The Fus3 protein concentration stays constant throughout the cell cycle and is not responsible for the fluctuations in kinase activity or in expression of pheromone-dependent genes (data not shown).
A second release experiment to study the timing of Fus3 activity during the cell cycle was performed. The release was checked by FACScan analysis (Fig. 1E). Fus3 activity and CLN2 mRNA fluctuate antagonistically (Fig. 1D). These data are in accordance with previous studies showing that the FUS1 and CLN2 transcripts fluctuate antagonistically (1). Basal and pheromone-induced activity of the MAPK Fus3 are cell cycle- regulated, but the protein levels of Fus3 stay constant. The activity of the MAPK Fus3 reaches its maximum in late mitosis and early G1, drops after START, is low in late G1, and increases again in S phase. The pattern of its cell cycle regulation is the same as observed for the FUS1 transcript and reaches its maximum when CLN2 levels are low.
Overexpression of CLN2 Represses Fus3 Kinase ActivityOverexpression of the G1 cyclin genes
CLN1 and CLN2 represses transcription of
FUS1 (1). To see whether CLN2 overexpression reduces Fus3 kinase activity, CLN2 was overexpressed from
the GAL1-10 promoter in cycling cells and Fus3 activity was
examined by in vitro kinase assays. Myc epitope-tagged Fus3,
expressed from the TPI1 promoter, was immunoprecipitated
with anti-Myc 9E10 antibodies. We confirmed that the steady state
protein levels of Fus3M in these strains are not affected by
overexpression of CLN2 using Western blot analysis with Fus3
specific polyclonal antibodies (data not shown). A bacterial
recombinant NH2-terminal fragment of Far1 served as a
substrate for Fus3 in immunokinase assays. Fig.
2A shows that Fus3 activity is strongly
repressed in cells overexpressing CLN2 from the
GAL1-10 promoter (lanes 2 and 4)
compared with cells where the promoter is not induced (lanes
1 and 3). This holds true not only for basal levels of Fus3 activity but also for pheromone-induced levels (lanes 3 and 4).
To confirm that the down-regulation of Fus3 activity was due to Cln2 expression and not to distinct growth conditions as a result of different carbohydrate sources we compared strains harboring the GAL1-10 CLN2 construct and the empty vector alone, respectively, in a Northern blot experiment. The same levels of the pheromone-inducible genes SST2 and STE3, irrespectively of the source of carbohydrates used in the media, were observed (Fig. 2C, lanes 1 and 2 for STE3, lanes 7 and 8 for SST2). At the same time, cells harboring the GAL1-10 CLN2 construct repress SST2 and STE3 transcription on galactose containing medium (Fig. 2C, compare lanes 3 and 4 for STE3, compare lanes 5 and 6 for SST2), as it was already shown for FUS1 transcription (1). Assuming that the transcription of SST2 and STE3 reflects the kinase activity of Fus3, the experiment demonstrates that repression of Fus3 activity occurs because of elevated Cln2 levels and not because of different carbohydrate sources.
CLN2-dependent Repression of Fus3 Activity Does Not Require Msg5 FunctionEither positive or negative factors could
mediate a cell cycle-dependent fluctuation in Fus3
activity. It was demonstrated by Doi et al. (36) that the
protein phosphatase Msg5 dephosphorylates Fus3 and thereby inactivates
it. Loss of Msg5 activity causes an increase in Fus3 activity.
Repression of Fus3 activity through overexpression of CLN2
could be regulated by increased Msg5 activity. We wanted to elucidate
the potential role of Msg5 in mediating the down-regulation of Fus3
activity in cells with elevated Cln2 levels. For this purpose, the
effect of CLN2 overexpression on Fus3 activity was examined
in a MSG5 deletion strain. Fus3 kinase assay were performed
as described above. Fig. 3 shows that repression of Fus3
by Cln2 does not require the Msg5 phosphatase (compare also with Fig.
2) and therefore must work through a different mechanism (Fig. 3,
lanes 1 and 2 (no -factor treatment) and
lanes 3 and 4 (
-factor treatment for 10 min)).
The studies described above indicate that the cell cycle machinery
represses the mating pathway upstream of the MAPK Fus3.
Ste7 MEK Activity Fluctuates during the Cell Cycle
Experiments were performed to see whether the activity of
the MEK Ste7 is also cell cycle-regulated. Cells were arrested with either cdc15-2 or cdc4-1, and Ste7 kinase
assays were performed under pheromone-uninduced conditions. Myc
epitope-tagged Ste7 was expressed constitutively from the
CYC1 promoter to avoid a potential cell cycle regulation of
the STE7 transcript. As a control wild-type cells harboring
the same STE7 construct and cells with a construct
containing a mutated version of STE7 were used. These cells
were treated with pheromone. Fig. 4 shows that the
activity of Ste7 is high in late mitosis (cdc15-2 arrest,
lane 4) and low in late G1 (cdc4-1
arrest, lane 3), as is the case for Fus3.
Epistasis Experiments with Hyperactive Alleles of STE11
To further see which components in the pathway are regulated by the cell cycle, studies with hyperactive alleles of the MEKK Ste11 were done. STE11-1 and STE11-4 hyperactive alleles induce the pheromone pathway independent from pheromone treatment and independent from G protein function (51). In this case, downstream targets of Ste11, like the MEK Ste7 and the MAPK Fus3, are induced and activate pheromone-dependent genes. Down-regulation of Fus3 activity or pheromone-dependent transcripts by overexpression of CLN2 in cells induced by an hyperactive STE11 allele would indicate that Cln2 regulates the pathway through Ste7, Ste11, or an event that controls the activity of the hyperactive Ste11 protein.
Fig. 5A shows that overexpression of
CLN2 represses the increased transcription of
SST2 in strains harboring the hyperactive STE11
alleles (compare lanes 3 and 4 (STE11-4 allele) and lanes 5 and 6 (STE11-1 allele) with lanes 1 and 2 (wild-type allele)). The transcription of another pheromone-inducible
gene, STE3, was also repressed (data not shown). Fus3 kinase
assays were performed as described above and demonstrate that the
activity of Fus3 in cells harboring STE11-1 is still
down-regulated by overexpression of CLN2 (Fig.
5B, compare lanes 3 and 4, and the
wild-type GA145, lanes 1 and 2). In conclusion,
the experiments described above indicate that the repression of the
pheromone-dependent signal transduction pathway occurs at a
step at the same level as Ste11 or downstream of Ste11.
Repression of the Hyperactive STE11-1 Allele Does Not Require G Protein Function
To test whether components at the level of the G
protein are involved in Cln2-dependent down-regulation of
the pathway the following experiments were performed: deletion of
GPA1, the subunit of the G protein, constitutively
activates the mating pathway due to a release of the
subunit
(23, 24, 27, 52). Overexpression of GAL1-10 CLN2 suppresses
the lethal phenotype of a gpa1 strain (1), suggesting that
Cln2 represses the mating pathway downstream of the
subunit. To
rule out that components upstream of STE11, which are not
necessary for Ste11-1 activity, are involved in
CLN2-dependent repression of the mating pathway, we examined the effect of GAL1-10 CLN2 induction in
ste4 STE11-1 cells. STE4 encodes the
subunit
of the G protein. It was shown that activation of the pathway by
STE11-1 does not require STE4 function. The
effect of CLN2 overexpression on the transcription of
SST2 (Fig. 5A, lanes 7 and
8) and STE3 (data not shown) was examined. Cln2
represses the mRNA levels of SST2 and STE3.
The Cln2-dependent down-regulation of the mating pathway
works independently of the G protein. It can be concluded that Cln2
represses the mating pathway downstream of the G protein.
To further see at what point of the signaling cascade Cln2 represses the mating pathway a strain carrying a deletion in the RGA1 gene was used. RGA1 was first characterized by Stevenson et al. (32) as a putative GTPase-activating protein for Cdc42 (32). Deletion of rga1 increases signaling of the pheromone pathway, an effect which is enhanced by a PBS2 deletion. A deletion in RGA1 and PBS2 activates the pheromone-dependent signal transduction pathway independently of the G protein, but requires function of Ste20 and Ste11. It functions upstream, or at the level of Ste11 (32). Fig. 5C shows the transcriptional level of SST2 and STE3 in strains deleted for STE4 which normally has very low basal activity of the pheromone-dependent pathway. Indeed, in the single mutant ste4 strain, hardly any SST2 transcription can be detected (Fig. 5C, lanes 1 and 2). In a ste4 rga1 pbs2 strain the pathway is more active (Fig. 5C, lanes 3, 5, 7, and 10). The activation is comparable to basal levels in a wild-type strain (32).
We wanted to know whether CLN2 overexpression represses transcription of SST2 and STE3 in a ste4 rga1 pbs2 mutant: CLN2 overexpressed from the GAL1-10 promoter is not able to repress the activation of the mating pathway due to an rga1 deletion (Fig. 5C, compare lanes 5 and 6, lanes 7 and 8, and lanes 9 and 10).
The following conclusions can be drawn from the experiments described above: Cln2-dependent repression does not seem to occur downstream of Ste11, as this would enable Cln2 to repress the signal obtained in a ste4 rga1 pbs2 strain. Cln2-dependent repression does not require function of the G protein, as the hyperactive STE11-1 allele is down-regulated in a ste4 strain. The repression seems to occur at the same level as Ste11. Possible targets left are, for example, the Cdc42 complex, Ste5 and Ste20. Other potential targets are Bem1 (54) and Cla2 (55, 56), proteins involved in polarized growth, an event which is induced at the level of Cdc42.
Cln2-dependent Repression Does Not Require Bem1 and Cla2 FunctionBem1 was shown to be important for cell polarity
during budding. This protein associates with Cdc24 (57). Bem1 also
forms a complex with Ste20, Ste5, and actin and thus seems to be
necessary for polarized rearrangement of the actin cytoskeleton prior
to mating (54). We thought of Bem1 as a candidate for
Cln2-dependent repression of the mating pathway as it may
act at the same level as STE11. Bem1 might have been either
a direct target of Cln2 or mediator of Ste5 or Ste20 repression. To
answer the question whether Bem1 was involved in the repression of the
pathway by Cln2, the following experiment was performed: in a
bem1 strain CLN2 was overexpressed from the
constitutive TPI1 promoter and the effect on FUS1
transcription was investigated. The bem1 strain is
temperature-sensitive. CLN2 repressed FUS1
mRNA accumulation on permissive (Fig. 6, lanes
5-8) and restrictive temperature (data not shown). We conclude
that Bem1 is not necessary for Cln2-dependent repression of
the mating pathway.
CLA2 codes for a GTPase-activating protein and is necessary for budding in cln1 cln2 cells (55). It was also isolated as BUD2 (56). It plays a role in localization and formation of buds and seems to be the GTPase for Bud1/Rsr1 (55, 56). GTP-bound Rsr1 is thought to interact with Cdc24 (58) and thus ensures polarity establishment at appropriate locations (59). Cdc24 may form a complex with Rsr1-GTP at the presumptive bud site and may thereby get activated in a Cdc28 kinase-dependent manner (53). The formation of this complex might involve Bem1 (60). Cla2 could therefore be a target for Cln2, whose activation would prevent activation of Rsr1. Loss of CLA2 function was found to be synthetic lethal with cln1 cln2 (55). Repression of the mating pathway by Cln2 could involve the activation of bud emergence. For this reason a cla2 strain was compared with the corresponding wild-type strain, both overexpressing CLN2 from the TPI1 promoter, to see whether Cla2 is necessary for the pathway repression. As it is shown in Fig. 6 (lanes 1-4), a CLA2 deletion has no influence on the down-regulation of the pheromone-dependent gene FUS1 by CLN2.
In summary it can be stated that the putative targets left for
Cln2-dependent repression of the mating pathway are Ste11, Ste5, Ste20, and the Cdc42 complex (Fig. 7). Further
studies will concentrate on the questions of which of these proteins
interact with Cln2, how this interaction occurs, and if Cln2 represses one of the components of the mating pathway directly or indirectly.
Messenger RNA levels of pheromone-inducible genes fluctuate during the cell cycle (1, 19). They reach their maximum in late mitosis and early G1, drop drastically at START in late G1, and increase again in S phase. In response, pheromone cells have to arrest their cell cycle in G1 before START (2, 3). Cells which have passed START are resistant to mating pheromone (9). Up-regulation of the transcription of pheromone-dependent genes in early G1 may reflect a mechanism that ensures the proper timing of the cell cycle arrest. Expression of the Far1 protein, which is thought to specifically inhibit the G1 cyclins Cln1 and Cln2, is restricted to early G1 (17, 18, 61). After START, Far1 is degraded and the cell is committed to undergo a mitotic cell division cycle (62). Improper expression of a truncated, nondegradable version of Far1 by the GAL1-10 promoter causes cell cycle arrest also in other stages of the cell cycle (budded-cell-arrest phenotype). The cell cycle regulation of Far1 transcription and proteolysis in G1 is one of the mechanisms ensuring the specificity of pheromone arrest in early G1 (62).
The fluctuations observed for transcript numbers of pheromone-dependent genes are oscillating in such a way that they reach their maximum when CLN1 and CLN2 transcripts are down-regulated and drop at the time when CLN1 and CLN2 reach their maximum. Oscillations of the G1 cyclins Cln1 and Cln2 and the accumulation of mRNA of pheromone-dependent genes are counteracting. Oehlen and Cross (1) showed that in fact increased Cln2 levels repress the transcript numbers of pheromone-dependent genes. Cells expressing CLN2 constitutively from the strong GAL1-10 promoter are resistant to pheromone. Cln2-dependent repression of pheromone-induced genes was shown to occur downstream or at the level of the G protein, and does not depend on either SST2 (part of the recovery response, 63) or the NH2 terminus of the pheromone receptor (important for desensitization) (64, 65). We were interested in assessing which components of the mating pathway are affected by a cell cycle-dependent down-regulation and at what level Cln2 represses the pathway.
To address the first question, studies were undertaken to examine the activity of the MAPK Fus3 during the cell cycle. They demonstrate that Fus3 activity is cell cycle-regulated, reflecting the fluctuations in transcript numbers of pheromone-dependent genes. Both basal and pheromone-induced Fus3 activity reach their maximal levels in mitosis and early G1. After START, Fus3 activity is almost absent and increases again during S phase. The cell cycle-dependent regulation of the transcripts of pheromone-induced genes reflects fluctuations in the activity of the MAPK Fus3.
The activity of the MAPK Fus3 is down-regulated in cells overexpressing CLN2 from the induced GAL1-10 promoter, indicating that the pathway and the cell cycle interact at a point upstream of the transcriptional regulation of pheromone-dependent genes. A potential regulator of the observed fluctuations in Fus3 activity seemed to be the Msg5 protein phosphatase. From our experiments the possibility that Msg5 is the mediator of Cln2-dependent repression of Fus3 activity can be excluded.
The whole kinase cascade is cell cycle-regulated in its activity. The activity of Ste7, which lies upstream of Fus3 in the kinase cascade, is subjected to the same cell cycle-dependent regulation as Fus3. Experiments with hyperactive STE11 alleles (51) indicate that Cln2-dependent repression of the mating pathway occurs through Ste11. Induced activities caused by STE11 hyperactive alleles are repressed by increased Cln2 levels. To further analyze at what point Cln2 represses the signal transduction pathway a strain with a deletion in the RGA1 gene was used (32). RGA1 is thought to act at or before Ste11 in the signaling cascade. Deletion of RGA1 increases the activity of Cdc42. Cdc42 was shown not only to have a role in establishing cell polarity (66) but also in pheromone-dependent induction of the mating pathway (30, 31). A deletion in RGA1 induces the pathway independently of the G protein, but requires the activity of the MAPK cascade. Transcription of pheromone-dependent genes can be observed in ste4 rga1 pbs2 cells, whereby pbs2 enhances the effect of an RGA1 disruption. Overexpression of CLN2 in ste4 rga1 pbs2 cells had no effect on the transcription of pheromone-dependent genes.
There are several conclusions from the experiments with strains harboring the hyperactive STE11 alleles and strains with a deletion in the RGA1 gene: repression of the hyperactive STE11 alleles demonstrates that the site of Cln2 dependent down-regulation has to be downstream or at the level of Ste11. Down-regulation of a STE11-1 allele in a ste4 background indicates that G protein function is not necessary for Cln2-dependent repression of the mating pathway. Experiments with rga1 strains suggest that the repression does not occur downstream of Ste11, because the transcription observed in these strains is not down-regulated by high levels of Cln2. This would be the case if either Fus3 or Ste7 were targets of Cln2. Thus the target of Cln2 seems to be a component involved in Ste11 activity.
Epistasis experiments elucidated the relative positions of components
of the mating pathway. It was shown that Ste11 acts upstream of Ste7
and Fus3, as a deletion of STE7, FUS3, or
STE12 suppresses the activation of the mating pathway by an
hyperactive STE11 allele (51). On the other hand, it was
shown that Ste11 acts downstream of Ste3 and Ste4, as a deletion of
STE3 (encodes the -factor receptor) and STE4
(encodes the
subunit of the G protein) had no effect on the
capacity of an hyperactive STE11 allele to induce the
pathway (51). Ste5 is necessary for the optimal activity of the
hyperactive Ste11 kinase and thereby acts at the level or downstream of
STE11 (51). Overexpression of STE5 suppresses
mutant alleles of pathway components downstream of Ste11. Consistent
with these genetic data it was shown that Ste5 interacts with Ste11,
Ste7, and Fus3/Kss1 in two hybrid assays. This indicates that it has a
role as a scaffold protein, holding together the components of the
kinase cascade (21). Rga1 works upstream or at the level of Ste11, as
it was shown by the following epistasis experiments: a deletion in
STE20, but not in STE5, reduces the signal
obtained by an RGA1 deletion (32). A deletion in STE11 abolishes the increased transcription in cells deleted
for RGA1. From these data we can think of a linear kinase
cascade downstream of Ste11, but at the level of Ste11 different
components in parallel may be involved in transmitting the signal from
the G protein to the kinase cascade. The hyperactive alleles were not
tested in a ste20 background because at this time the Ste20 kinase was not identified yet. The pathway seems to be much more complex at the level of Ste11 than it was initially thought. For this
reason it is not possible to exactly define the component of the mating
pathway interacting with Cln2.
The different possible interactions between Cln2 and the mating pathway are illustrated in Fig. 7. Cln2 induces budding without pheromone treatment, and morphological changes resulting in shmoo formation are inhibited (18). For this reason experiments determining the role of Bem1 and Cla2 in Cln2-dependent repression of the pathway were done. They were both shown not to be involved in Cln2-dependent repression of the mating pathway.
Further experiments will try to exactly define the point where Cln2 interacts with the mating pathway and how this interaction occurs. We exclude the possibility that the observed repression of the mating pathway is due to a shortened G1 phase and budding at a smaller cell size in cells overexpressing CLN2 since Cross (67) demonstrated that DAF1-1, a mutant allele of the G1 cyclin CLN3, which causes a shorter G1 phase, accumulates FUS1 mRNA to wild-type levels. Cells bud at a smaller cell size, but the intensity of the FUS1 transcript remains the same as in wild-type cells. We expect that overexpression of CLN2 shortens G1, but that this effect is not causing the down-regulation of the gene transcript numbers.
From Fig. 7 it can be seen that the possible targets left are Ste5, Ste20, Ste11, and the Cdc42 complex. Cln2 seems to down-regulate one of these components of the mating pathway either directly or indirectly. On the other hand, it is still possible that Cln2 is repressing several components of the mating pathway concomitantly. It could down-regulate the kinase cascade by repressing Ste11 activity, and promote budding by repressing polarized growth for shmoo formation. The combination of different hyperactive mutants of components involved in the pheromone response will address this issue.
The biological relevance of the studies presented here lies in the demonstration that the cell cycle-dependent regulation of the mating pathway determines its responsiveness and thereby the correct timing of the cell cycle arrest. Also in higher eucaryotic cells the correct timing of responses to external signals has to be in a cell cycle context. Edelmann et al. (68) demonstrated that the activities of the p42mapk/p44mapk also fluctuate during the cell cycle in Swiss mouse 3T3 fibroblasts. It is not known whether upstream components of this pathway underlie the same cell cycle regulation. The studies presented here may be relevant for the proper timing of cell growth and proliferation and may help to elucidate the general mechanism ensuring the correct timing of different responses to external stimuli.
We thank members of G. Ammerer's and K. Nasmyth's lab for providing yeast strains, probes, antibodies, technical advice, and helpful discussions. We thank B. Errede, G. Sprague, B. Stevenson, and J. Pringle for yeast strains and plasmids. K. W. is especially grateful to M. Baccarini and A. Gartner for critical reading of the manuscript and to B. Errede for help in Chapel Hill.