The C-terminal Domain of the Largest Subunit of RNA Polymerase II Is Required for Stationary Phase Entry and Functionally Interacts with the Ras/PKA Signaling Pathway*

The Saccharomyces cerevisiae Ras proteins control cell growth by regulating the activity of the cAMP-de-pendent protein kinase (PKA). In this study, a genetic approach was used to identify cellular processes that were regulated by Ras/PKA signaling activity. Interest-ingly, we found that mutations affecting the C-terminal domain (CTD), of Rpb1p, the largest subunit of RNA polymerase II, were very sensitive to changes in Ras signaling activity. The Rpb1p CTD is a highly conserved, repetitive structure that is a key site of control during the production of a mature mRNA molecule. We found that mutations compromising the CTD were synthetically lethal with alterations that led to elevated levels of Ras/PKA signaling. Altogether, the data suggested that Ras/PKA activity was negatively regulating a protein that functioned in concert with the CTD during RNA pol II transcription. Consistent with this prediction, we found that elevated levels of Ras signaling caused growth and transcription defects that were very similar to those observed in mutants encoding an Rpb1p with a truncated CTD. In all, these data suggested that S. cerevisiae growth control and RNA pol II transcription might be coupled by using the Ras pathway to regulate CTD function. The was assessed after a shift to the nonpermissive temperature of 37 °C. 30 min at 37 °C, the poly(A) (cid:1) RNA had decreased 10-fold in this mutant. To analyze gene expression at the diauxic shift, log phase cultures were diluted a of 0.1 600 /ml and were incubated for 8–12 at 30 °C. The diauxic shift to that period of growth where have depleted the glucose in the culture and are beginning to utilize the by-products of the fermentation process. For these experiments, total was from in late log phase undergoing diauxic

The RAS genes encode small GTP-binding proteins that play an important role in regulating cell proliferation in many eukaryotes (1,2). Ras proteins typically function as molecular switches by oscillating between an active GTP-bound state and an inactive GDP-bound state (3,4). The budding yeast, Saccharomyces cerevisiae, contains two Ras proteins, Ras1p and Ras2p, that together control the activity of the cAMP-dependent protein kinase, PKA 1 (5,6). Ras/PKA signaling is a key regulator of growth in this yeast as mutations that inactivate this pathway cause a premature entry into the G 0 -like resting state, known as stationary phase (4). Conversely, elevated levels of Ras/PKA signaling result in a failure to arrest in a normal stationary phase (4,6). However, despite the obvious importance of this signaling pathway, only a few PKA sub-strates important for the control of yeast cell growth have been identified (7)(8)(9).
The entry into stationary phase is accompanied by broad changes in the patterns of gene expression that are controlled, in part, by the Ras/PKA pathway (10 -12). However, it is not yet known precisely how Ras activity influences the transcriptional apparatus. In S. cerevisiae, RNA polymerase (pol) II exists as a large holoenzyme complex that contains the 12subunit polymerase, the Mediator co-activator complex, the Srb8-11 protein complex, and several general transcription factors (13). This holoenzyme is recruited to active promoters as a result of specific interactions between Mediator subunits and the DNA-bound transactivators present at these promoters (14,15). Therefore, there are at least two potential targets for the Ras effects on RNA pol II activity as follows: the various transcription factors bound at the individual promoters, and the regulatory proteins associated with the RNA pol II holoenzyme. A number of studies already indicate that the Ras pathway regulates the activity of specific promoter-bound transcriptional regulators, like Msn2p and Msn4p (9,16). In contrast, although a few studies (12,(17)(18)(19) have hinted at the possibility, there have been no reports of signaling pathways directly targeting components within the RNA pol II holoenzyme.
Although RNA pol II transcription is controlled at multiple levels, it is clear that the C-terminal domain (CTD) of Rpb1p, the largest subunit of RNA pol II, is a key site of regulation. The Rpb1p CTD is a highly conserved, repetitive structure that is required for many aspects of mRNA production including pre-initiation complex formation, transcript elongation, and processing of the nascent transcript (20,21). In S. cerevisiae, the CTD consists of 26 or 27 repeats of the consensus heptamer, Tyr 1 -Ser 2 -Pro 3 -Thr 4 -Ser 5 -Pro 6 -Ser 7 (22,23). The CTD is phosphorylated by multiple protein kinases including Kin28p/Cdk7 in the general transcription factor, TFIIH, Srb10p/Cdk8 in the Srb complex, and Ctk1p/Cdk9 (24). These kinases are generally specific for either serine 2 or serine 5 of the individual repeats. Although all of the details are not yet understood, this phosphorylation clearly plays an important role in the regulation of CTD function (25,26). For example, the Mediator complex specifically associates with RNA pol II enzymes that contain a hypophosphorylated CTD (27)(28)(29). The subsequent phosphorylation of the CTD by Kin28p results in the release of the Mediator and the association of a complex of proteins important for transcriptional elongation (30). Therefore, a complete understanding of the mechanisms regulating mRNA production will require a thorough working knowledge of the Rpb1p CTD.
We are interested in defining the role played by the S. cerevisiae Ras/PKA signaling pathway in the control of stationary phase entry and general cell growth. In this study, we identified specific genetic interactions between mutations af-fecting the Ras pathway and those affecting the Rpb1p CTD or proteins that function through this domain. These genetic data suggested that the Ras/PKA pathway was negatively regulating activities associated with the Rpb1p CTD. Consistent with this prediction, we found that increased Ras signaling resulted in phenotypes similar to those observed in mutants that encode a Rpb1p with a truncated CTD. These CTD truncation mutants are viable but exhibit specific defects in growth and RNA pol II transcription. Altogether, the data suggest that S. cerevisiae growth control and RNA pol II transcription might be coupled by using the Ras pathway to regulate CTD function. The relative merits of a transcriptional control mechanism that directly targets proteins within the RNA pol II holoenzyme is discussed.

EXPERIMENTAL PROCEDURES
Growth Media-Standard Escherichia coli growth conditions and media were used throughout this study (31). Yeast YPAD, 5-fluoroorotic acid, and SC growth media were as described (32,33). YM-glucose medium refers to a yeast minimal medium containing 0.67% yeast nitrogen base (Difco), 2% glucose, and those growth supplements required for cell proliferation.
Plasmid Constructions-To construct the MET3-RAS2 Val19 plasmids, the MET3 promoter region was first cloned as a 550-bp SalI-EcoRV fragment from the pHAM8 plasmid (kindly provided by Dr. Harry Mountain) into pRS403 to form pPHY440. RAS2 Val19 was then cloned as a 1.3-kb BamHI fragment from pJW83.1 (kindly provided by Dr. Jennifer Whistler) into pPHY440 to form pPHY446. This RAS2 Val19 fragment contained the coding sequences and transcriptional terminator but lacked the RAS2 promoter. A MET3-RAS2 plasmid, pPHY442, was constructed in a similar fashion. The MET3-RAS2 Val19 hybrid gene from pPHY446 was subcloned into pRS416 to form pPHY795. The RAS2 Val19 plasmids, PHY453 and pJR1040, contain the RAS2 Val19 allele with its own promoter cloned into pRS415 and pRS316, respectively. The pRS plasmids were described previously (34,35).
Yeast Strain Constructions and Genetic Methods-The strains used in this study are listed in Table I. Unless otherwise noted, strains were from our lab collection or were derived during the course of this work. Standard yeast genetic methods were used for the construction of all strains (32). A plasmid shuffle procedure was used to construct the rpb1 strains described in Fig. 2. The different rpb1 alleles were present on LEU2-marked plasmids that were transformed into PHY2003; these plasmids were kindly provided by Dr. Richard Young. PHY2003 contains the wild-type RPB1 locus on a URA3 plasmid, pRY2112. The resulting transformants were plated to 5-fluoroorotic acid media to select for those cells that had lost the URA3 plasmid.
For the MET3-RAS2 Val19 experiments, strains carrying this inducible construct were typically grown to mid-log in YM-glucose minimal medium containing 500 M methionine. These growth conditions effectively repressed expression from the MET3 promoter. The cells were then collected by centrifugation and resuspended in the same growth medium lacking methionine to induce expression from the MET3 promoter.
To test for genetic interactions with RAS2 Val19 , the relevant strains were transformed with plasmids containing either RAS2 Val19 or the inducible MET3-RAS2 Val19 construct. We tested more than 100 different mutants defective for a wide variety of processes important for cell growth. These strains were either in our lab collection previously or were provided by scientists in the general yeast community. For strains with temperature-sensitive (ts) growth defects, the growth rates of the transformed strains were assessed at temperatures that were permissive or semi-permissive for the growth of the original mutant. In each case, the effects of RAS2 Val19 on the mutant were compared with the effects on an isogenic wild-type control. Strain information for all of the mutants listed in Table II is available upon request.
Stationary Phase Assays-The stationary phase viability assays were performed as described (12). For each strain tested, the number of survivors after 10 -14 days of growth in a minimal medium was normalized to that found for an isogenic wild-type control. The heat shock sensitivity assays were performed on cultures grown for 3 days in a YM-glucose minimal medium at 30°C. Cells were collected by centrifugation and resuspended in distilled water at a concentration of 10 A 600 units/ml. The suspensions were subjected to a series of 5-fold dilutions, and 200 l of each suspension was placed into a well of a microtiter plate. These suspensions were plated to solid medium with a 48-prong replicating block, and the plates were subjected to a 55°C heat shock for 20 or 30 min. The plates were then incubated for 3 days at 30°C, and the relative number of survivors was determined for each strain analyzed.
RNA Analyses-Total RNA was prepared from yeast cells by a hot phenol extraction method described previously (36). For Northern analyses, 20 g of total RNA per lane was loaded onto a formaldehydeagarose gel and subjected to electrophoretic separation. The gels were blotted to nylon membranes (Schleicher & Schuell) that were then hybridized with the appropriate 32 P-labeled probes (36). Typically, these probes were 0.7-1.0-kb PCR fragments that were prepared with the Oligo Labeling Kit (Amersham Biosciences). To measure the amount of total poly(A) ϩ RNA, 10 -20 g of total RNA was spotted onto a nitrocellulose membrane with the assistance of a slot blot apparatus (Schleicher and Schuell). The membrane was hybridized with a 32 Plabeled oligo(dT) oligonucleotide probe, and the relative signal present was quantified by a PhosphorImager analysis. As a control for this latter experiment, the amount of poly(A) ϩ RNA in an rpb1-1 mutant was assessed after a shift to the nonpermissive temperature of 37°C. After 30 min at 37°C, the poly(A) ϩ RNA signal had decreased more than 10-fold in this mutant. To analyze gene expression at the diauxic shift, log phase cultures were diluted to a density of 0.1 A 600 /ml and were incubated for 8 -12 h at 30°C. The diauxic shift refers to that period of growth where cells have depleted the glucose in the culture and are beginning to utilize the by-products of the fermentation process. For these experiments, total RNA was prepared from cells in late log phase and those undergoing the diauxic shift.
CTD Phosphorylation Assays-Cells were grown to mid-log in minimal medium, collected by centrifugation, and washed with 1 ml of ice-cold 20 mM NaF. Protein extracts for the Rpb1p CTD phosphorylation assays were prepared as described (37). The proteins were separated on a 5% SDS-polyacrylamide gel and were transferred to a nitrocellulose membrane (Hybond ECL, Amersham Biosciences) (36). The membrane was hybridized with antisera specific for the nonphosphorylated CTD repeat, monoclonal 8WG16 (Covance Research Products); phosphoserine at position 2 of the CTD repeats, monoclonal H5 (kindly provided by David Bregman); and phosphoserine at position 5, monoclonal H14 (Covance Research Products). Immunoreactive proteins were detected with anti-mouse IgG (Amersham Biosciences) used at a dilution of 1:2000. The Supersignal chemiluminescent substrate (Pierce) was subsequently used to illuminate the reactive bands.

Elevated Levels of Ras/PKA Signaling Were Synthetically
Lethal with Mutations That Affect RNA pol II Transcription-We are interested in understanding the role played by the Ras/PKA signaling pathway in the control of growth in S. cerevisiae. To this end, we have been examining the effects of Ras signaling activity on mutants defective in a variety of cellular processes. In particular, these experiments have been directed at identifying mutant phenotypes that were either suppressed or enhanced by increased levels of Ras signaling. The underlying rationale was that processes controlled by the Ras/PKA pathway would be the most sensitive to changes in Ras signaling activity. Interestingly, we found that many mutants defective for RNA pol II transcription exhibited a severe growth defect in the presence of increased Ras/PKA activity ( Fig. 1A and Table II). These Ras effects were rather specific for RNA pol II transcription as mutations affecting other cellular processes were relatively insensitive to increased Ras activity (Table II) (38). These latter mutations included those affecting protein translation, cell cycle progression, protein secretion, and mRNA export and processing.
These genetic studies have converged upon an ongoing analysis in our laboratory of the yeast rye mutants. These mutants were originally identified on the basis of a defective transcriptional response to nutrient deprivation (12). A subsequent analysis has shown that the rye mutants are defective for the control of growth and are unable to enter into a normal stationary phase (12,19). During the course of these studies, we found that rye1 mutations were synthetically lethal with a hyperactive allele of RAS2, known as RAS2 Val19 (19). This allele encodes a Ras2p with diminished GTPase activity, and thus the altered Ras2p is found more often in the active GTPbound form (5). The RYE1 gene was cloned and found to be identical to SIN4, a gene that encodes a component of the yeast Mediator complex (19,39). Thus, two independent lines of investigation have identified a genetic interaction between the Ras/PKA signaling pathway and the RNA pol II transcription apparatus.
In all, more than 100 different mutations were examined in combination with RAS2 Val19 and other alterations that affect the Ras/PKA signaling pathway (Table II and data not shown). Several examples of the synthetic growth defects observed with RAS2 Val19 are shown in Fig. 1. The presence of increased Ras signaling greatly exacerbated the growth defects associated with sin4, kin28, srb4-138, and rpb1-104 mutants. SRB4 encodes a component of the Mediator, and the rpb1-104 allele encodes a truncated Rpb1p protein that contains only 11 of the 26 CTD repeat units (40,41). These experiments employed an inducible allele of RAS2 Val19 in which the RAS2 promoter was replaced with the promoter from the yeast MET3 gene (19). The MET3 promoter is repressed by the presence of methionine in the growth medium, and RAS2 Val19 is therefore expressed only when methionine is absent (42,43). The growth defects observed with the MET3-RAS2 Val19 allele in Fig. 1 were specific to media lacking methionine (data not shown). Finally, all pairwise double mutant combinations between the RAS2 Val19 , kin28, rpb1-104, and sin4 mutations resulted in a severe growth defect (Fig. 1B) (44). This result indicated that each of these mutations was very sensitive to a further decrease in transcriptional efficiency and suggested that Ras/PKA signaling might also be influencing RNA pol II activity.
In addition to the above growth defects, the presence of increased levels of Ras/PKA signaling resulted in a rapid cell death in mutants that were defective for RNA pol II transcription. For example, with sin4 mutants, less than 0.01% of the cells were viable 2 h after induction of RAS2 Val19 expression (Fig. 1C). A similar induction of RAS2 Val19 in wild-type cells did not affect cell growth or viability. Because Sin4p is a transcriptional regulator, this lethality could have been due to a global defect in RNA pol II transcription or to defects at a subset of genes essential for cell viability. Two experimental observa- Essential subunit of Mediator sin4⌬ Nonessential subunit of Mediator gal11⌬ Nonessential subunit of Mediator rgr1-⌬2 Essential subunit of Mediator hrs1⌬ Nonessential subunit of Mediator med2⌬ Nonessential subunit of Mediator

Mutations not exhibiting a synthetic interaction with RAS2 Val19
Mutation Gene product function rpb1-1 Largest subunit of RNA pol II rpb5-9 Subunit of RNA pol II srb10⌬ CTD kinase; Cdk8 fcp1 CTD phosphatase med1⌬ Nonessential subunit of Mediator med9⌬ Nonessential subunit of Mediator ppr2⌬ pol II transcript elongation factor TFIIS cdc73⌬ Subunit of alternative RNA pol II holoenzyme swi2⌬ ATPase subunit of SNF/SWI complex cac3⌬ Subunit of chromatin assembly factor I gcn5⌬ Histone acetylase, subunit of SAGA complex taf90 TBP-associated protein (TAF) hat1⌬ Histone acetylase, type B ceg1-237 Subunit of mRNA capping enzyme prp18 Splicing factor prp22 Required for splicing dbp5 Required for export of mRNA cdc33-1 Cap-binding protein; translation factor cdc28-4 Cdc2 cyclin-dependent kinase homolog cdc37-1 Required for progression through G 1 sec18-1 Required for membrane vesicle fusion sec1-1 Required for transport from the trans-Golgi network to the plasma membrane vps34⌬ Phosphatidylinositol 3-kinase required for vacuolar protein sorting zds2⌬ Suppressor of sin4 tions indicated that this latter possibility was a more likely explanation. First, the expression of RAS2 Val19 did not affect the level of total poly(A) ϩ RNA in either wild-type or sin4 cells (Fig. 1D). Second, the steady-state levels of multiple mRNAs were essentially identical in wild-type and sin4 RAS2 Val19 mutants (19). In both of these experiments, RNA levels were assessed at a time when there were no viable cells present in the sin4 RAS2 Val19 mutant cultures (see Fig. 1C). Therefore, we interpret the lethality of the double mutants to be due to the misregulation of a limited number of promoters that have a specific function in growth control. The RAS2 Val19 Growth Defects Were Specific to Mutations Affecting the Rpb1p CTD-It is important to note that not all mutations affecting RNA pol II transcription were influenced by the presence of RAS2 Val19 (Table II). Instead, most of the mutations that were synthetically lethal with RAS2 Val19 affected proteins that interacted with the Rpb1p CTD. For example, RAS2 Val19 was synthetically lethal with mutations that affected the Mediator (sin4, gal11, med6, and srb4), the TFIIH complex (kin28, ccl1, rig2, and tfb1) and the CTD itself (Table  II) (45,46). The Mediator associates with the RNA pol II holoenzyme by directly interacting with the CTD, whereas TFIIH contains a major CTD kinase, Kin28p (28,47,48). The importance of the CTD for these RAS2 Val19 effects was perhaps best illustrated by the responses of different rpb1 alleles to increased Ras signaling. The presence of RAS2 Val19 greatly exacerbated the growth defects associated with those rpb1 alleles that encoded a truncated CTD ( Fig. 2A). In contrast, rpb1 mutations that affected other regions of Rpb1p were relatively insensitive to increased Ras activity (Fig. 2A). These latter mutations typically cause a severe ts defect in RNA pol II FIG. 1. Increased Ras signaling resulted in a rapid cell death in mutants defective for RNA pol II transcription. A, mutations that compromise RNA pol II function exhibited a synthetic growth defect with RAS2 Val19 . Yeast strains with the indicated mutations were transformed with either a control vector, pRS416 (RAS2), or a plasmid with RAS2 Val19 under the control of the inducible promoter from the yeast MET3 gene, pPHY795 (RAS2 Val19 ). The strains were grown to mid-log in the presence of methionine (repressed), and equal numbers of cells were then spotted to a growth medium lacking methionine (induced). Growth was assessed after 2 days at 30°C for wild-type, sin4, kin28, and srb4-138 and at 33°C for rpb5-9 and rpb1-104. The strains shown are wild-type (PHY1220), sin4 (PHY1454), kin28 (JGV4), srb4-138 (Z811), rpb5-9 (WY-186), and rpb1-104 (Z551). For all strains, the observed growth rate was compared with that of an isogenic wild-type control. B, schematic showing the synthetic lethal interactions observed between the indicated mutations. The lines with barbed ends indicate the synthetic lethal interactions detected in this study. The straight line indicates an interaction described previously (44). C, increased Ras signaling caused a rapid cell death in sin4 mutants. Yeast strains were grown to mid-log in a minimal medium containing methionine and were transferred to a medium lacking methionine to induce expression from MET3-RAS2 Val19 . At the indicated times, dilutions were plated to a rich growth medium and incubated for 3 days at 30°C. The number of colonies formed was a measure of the number of viable cells present in the culture. In each case, the numbers were normalized to those obtained for the isogenic wild-type strain. The strains analyzed were as follows: wild-type (PHY1837, triangles), RAS2 Val19 (PHY1834, circles), sin4 (PHY1647, diamonds), and sin4 RAS2 Val19 (PHY1649, squares). The RAS2 Val19 strains contained the MET3-RAS2 Val19 allele. D, increased Ras/PKA signaling did not cause a global defect in mRNA accumulation in sin4 mutants. Total poly(A) ϩ RNA levels in wild-type and sin4 RAS2 Val19 cells were measured by a slot blot analysis. Wildtype (PHY1837) and sin4 MET3-RAS2 Val19 (PHY1649) strains were grown to mid-log at 30°C in a YM-glucose minimal medium containing 500 M methionine. The cells were then transferred to a minimal medium lacking methionine to induce expression from the MET3-RAS2 Val19 gene. Total RNA was prepared at 0 and 4 h after the transfer to the medium lacking methionine, and 20 g of each sample was spotted onto a nitrocellulose membrane. The level of poly(A) ϩ RNA was then assessed by hybridization with a 32 P-labeled oligo(dT) probe. The relative signal was quantified with a PhosphorImager and is shown in terms of arbitrary PhosphorImager units.

FIG. 2. Evidence for a functional interaction between the Ras/ PKA signaling pathway and the RNA pol II CTD.
A, rpb1 alleles that encode CTD truncations were sensitive to increased levels of Ras/ PKA signaling. The indicated rpb1 strains with either a control vector (RAS2) or a plasmid containing MET3-RAS2 Val19 (RAS2 Val19 ) were grown to mid-log in the presence of methionine. Equal numbers of cells were then spotted to a growth medium lacking methionine to induce expression from the MET3 promoter and grown for 3 days at 33°C. The schematic indicates the relative positions of the mutations in the rpb1 alleles. B, decreased signaling through the Ras/PKA pathway suppressed the growth defects associated with kin28 and tfa1 mutants. Wild-type (GF262-2), kin28 (JGV4), and tfa1 (YSB0331) strains containing either a control vector or a high copy PDE2 plasmid were grown for 3 days on a YM-glucose minimal medium at 32°C for kin28 and 35°C for tfa1. activity and have been described previously (49). These data therefore suggested the existence of a functional relationship between Ras signaling activity and the Rpb1p CTD.
Because the CTD is essential for multiple steps during the production of the mature mRNA, we examined mutations that affected each of these different steps. This analysis indicated that the RAS2 Val19 lethality was limited to those mutations affecting RNA pol II transcription. Mutations that affected later steps, such as mRNA splicing and export, did not exhibit a genetic interaction with RAS2 Val19 (Table II). In addition, the Ras effects were specific to particular components of the RNA pol II machinery. Mutants defective for other activities, including other subunits of RNA pol II (rpb5), a CTD-specific phosphatase (fcp1), and a second CTD kinase (srb10), were not affected by increased levels of Ras/PKA signaling (Table II). In all, the specificity of the genetic interactions suggested that the Ras/PKA pathway might be targeting the CTD itself or perhaps a protein that functionally interacts with the CTD during RNA pol II transcription. Moreover, the specificity of the RAS2 Val19 effects was consistent with the possibility that a single PKA target was responsible for the genetic interactions observed in this study.
Decreased Ras/PKA Signaling Suppresses Mutations That Affect RNA pol II Transcription-The presence of RAS2 Val19 results in relatively high levels of PKA activity, levels higher than those typically seen during normal yeast cell growth (4,6). Therefore, we tested whether variations within the normal physiological spectrum of Ras activity would also have an effect upon RNA pol II transcription. Indeed, the above data suggest that decreasing the levels of Ras signaling might suppress mutations affecting CTD-related functions. Consistent with this prediction, we found that the ts growth defects associated with kin28 and tfa1 mutants were suppressed by lowering the activity of the Ras/PKA signaling pathway (Fig. 2B). TFA1 encodes the large subunit of TFIIE, and as indicated above, KIN28 encodes a CTD kinase associated with TFIIH (48,50). The general transcription factor, TFIIE, is required for the recruitment of TFIIH to the Rpb1p CTD (51)(52)(53). In these experiments, Ras/PKA activity was lowered by the introduction of a high copy plasmid containing the PDE2 gene. PDE2 encodes a cAMP phosphodiesterase that, when overproduced, results in lowered levels of cAMP and in decreased PKA activity (54). This PDE2 plasmid had no significant effect on the growth of wild-type cells but allowed kin28 and tfa1 ts strains to grow at elevated temperatures (Fig. 2B). In addition, a recent study (19) found that cyr1 mutations that lower cAMP levels also suppress transcription defects associated with sin4 mutants. CYR1 encodes the adenylyl cyclase that is activated by the yeast Ras proteins (55). Therefore, both raising and lowering the mitotic levels of Ras signaling had profound effects upon mutations that affect Rpb1p CTD function.

RAS2 Val19 Mutants Exhibited Growth and Transcription Defects Similar to Those Associated with CTD Truncation Mutants-
The above data can be explained by a model proposing that Ras activity negatively regulates aspects of Rpb1p CTD function. Such a model would predict that elevated levels of Ras signaling would produce phenotypes similar to those associated with rpb1 alleles that compromise CTD function. For these experiments, we used the rpb1-104 allele that encodes an Rpb1p with only 11 CTD repeats; the presence of at least 8 -10 heptad repeats is required for cell viability (56,57). Previous work has shown that rpb1-104 mutants grow poorly on media lacking inositol (49), and we have found that these mutants exhibit growth defects on galactose media (Fig. 3A). Here we found that RAS2 Val19 mutants exhibited similar defects on these growth media (Fig. 3A). In addition, these growth defects were exacerbated in RAS2 Val19 rpb1-104 double mutants (Fig.  3A). Note that this double mutant was viable at 30°C on glucose-based media but exhibited a severe growth defect on such media at both higher and lower temperatures (data not shown).
The RAS2 Val19 and rpb1-104 mutants also exhibited similar defects in activated transcription. CTD truncation mutants had been shown previously to be defective for transcription from the INO1 and GAL1 promoters (58). For both genes, the expression under inducing conditions was down 5-8-fold in the rpb1-104 mutant (Fig. 3B). A similar transcription defect was observed with RAS2 Val19 mutants, and these defects were greatly exacerbated in the RAS2 Val19 rpb1-104 double mutant (Fig. 3B). Moreover, RAS2 Val19 had no effect on the induced levels of other promoters, like those for PHO5 and MET19, that were also insensitive to CTD truncations ( Fig. 3B and data not  shown). Therefore, increased levels of Ras/PKA signaling re- For the INO1 assays, strains were grown to mid-log in a YM-glucose minimal medium supplemented with 400 M inositol and were transferred to a medium with 10 M inositol for 6 h. For GAL1, strains were grown to mid-log in a minimal medium containing 2% raffinose and were transferred to a medium containing 5% galactose for 6 h. For PHO5, strains were grown in a minimal medium containing 7.5 mM phosphate to mid-log and were transferred to a medium lacking phosphate for 6.5 h. Total RNA was then prepared, and the levels of the indicated mRNAs were assessed by Northern RNA blot analysis. sulted in growth and transcription phenotypes similar to those previously associated with CTD truncations.
The Rpb1p CTD Was Required for Entry into a Normal Stationary Phase upon Nutrient Deprivation-We also tested whether rpb1 alleles that truncate the CTD caused growth defects that were normally associated with increased levels of Ras signaling. This would be expected if some of the growth consequences of elevated Ras activity were due to the downregulation of CTD function. The primary defect associated with yeast RAS2 Val19 mutants is a failure to arrest within stationary phase when an essential nutrient becomes limiting (4). This defect in stationary phase entry has a number of consequences including a rapid loss of viability during growth in a nutrient-limiting medium, a failure to accumulate storage carbohydrates, like glycogen, and a diminished resistance to environmental stresses, including heat shock (4, 6, 10, 59). We tested whether rpb1-104 mutants were defective for stationary phase entry by examining each of these characteristics.
Wild-type stationary phase cells can survive for extended periods in a nutrient-depleted growth medium. In contrast, RAS2 Val19 mutants exhibit a rapid loss of viability following nutrient deprivation (4,6). For example, after 10 days of growth in minimal medium, the number of survivors in a RAS2 Val19 culture was less than 1% that observed with an isogenic wild-type strain (Fig. 4A). We found that rpb1-104, sin4, and kin28 mutants all exhibited a similar stationary phase viability defect (Fig. 4A). In each case, the survival rate of the mutant after 14 days of growth was more than 15-fold lower than that of the wild-type control (Fig. 4A). Thus, the presence of a wild-type Rpb1p CTD was required for the normal acquisition of this stationary phase characteristic.
Stationary phase cells also accumulate elevated levels of the storage carbohydrate, glycogen, and exhibit an increased resistance to elevated temperatures (10). We found that rpb1-104 mutants were defective for both of these stationary phase traits. First, rpb1-104 mutants, like RAS2 Val19 , failed to accumulate normal levels of glycogen following nutrient deprivation (Fig. 4B). For these assays, glycogen levels were assessed in wild-type, RAS2 Val19 , and rpb1-104 cultures grown for 2, 4, or 6 days in minimal medium at 30°C. The rpb1-104 storage defect was already apparent after only 2 days of growth in minimal medium, a time that roughly corresponds to the beginning of the post-diauxic phase of growth (12). Finally, we found that rpb1-104 mutants grown for 3 days in minimal medium were 5-6 times more sensitive than wild-type cells to a brief heat shock (Fig. 4C). In all, these data indicated that the Rpb1p CTD was required for entry into a normal stationary phase in S. cerevisiae. Moreover, these results were consistent with the proposition that Ras/PKA signaling negatively regulates functions associated with the Rpb1p CTD.
It is important to point out that the stationary phase defects associated with rpb1-104 mutants were generally less severe than those observed with RAS2 Val19 . For example, although RAS2 Val19 mutants exhibited a dramatic loss of viability after 5 days of growth in minimal medium, similar defects in the rpb1-104 strains were not observed until 10 -14 days of growth. One possible explanation for this difference is that the Ras/ PKA pathway targets multiple proteins important for stationary phase entry. Components associated with the Rpb1p CTD would represent only one of these regulatory targets. Alternatively, the less severe phenotype of the rpb1-104 mutant could be due to the presence of the 11 CTD repeats in the encoded Rpb1p; an Rpb1p lacking all repeats might be severely defective in stationary phase entry. Unfortunately, because the Rpb1p CTD is essential for cell viability, this latter possibility would be difficult to test directly.

RAS2 Val19 and rpb1-104 Mutants Exhibited Similar Defects in RNA pol II Transcription upon Nutrient Deprivation-
The entry into stationary phase is accompanied by broad changes in the patterns of gene expression. Although the levels of most transcripts decrease upon nutrient deprivation, specific sets of genes are induced at this time (10,11,60). Many of these latter genes are negatively regulated by Ras/PKA signaling activity (60,61). In this section, we tested whether the Rpb1p CTD was required for the full induction of seven genes known to be FIG. 4. Truncation of the Rpb1p CTD resulted in a failure to enter into a normal stationary phase. A, stationary phase survival of strains defective for functions associated with the Rpb1p CTD. Strains with the indicated genotypes were grown for 10 -14 days at 30°C in a YM-glucose minimal medium. Serial dilutions of these cultures were then plated and incubated for an additional 3-4 days at 30°C except for the kin28 strain which was incubated at 25°C. The number of colonies formed after this incubation was a measure of the number of survivors in the stationary phase cultures. The percent viability represents the number of survivors present in the mutant culture relative to the number present in a culture of an isogenic wild-type strain. The data represent the average of at least three different experiments, and the average error in each case was less that 15%. The strains analyzed were sin4 (PHY1454), kin28 (JGV4), rpb1-104 (PHY2193), and RAS2 Val19 (PHY2194). The inset shows the relative survival rate in the terminal dilutions of the following strains: wild-type (PHY2195), rpb1-104 (PHY2193), and RAS2 Val19 (PHY2194). B, rpb1-104 mutants exhibited defects in glycogen accumulation. Strains with the indicated genotypes were spotted to a YM-glucose minimal medium plate and incubated for 2 days at 30°C. These plates were then exposed to iodine vapors and photographed with a Sony Mavica MVC-FD25 digital camera. The intensity of the observed staining is proportional to the amount of glycogen present. The strains analyzed were rpb1-104 (PHY2193), RAS2 Val19 (PHY2194), and wild-type (PHY2195). C, stationary phase cultures of rpb1-104 mutants were sensitive to a brief heat shock treatment. Wild-type and rpb1-104 cells were grown for 3 days at 30°C in a YM-glucose minimal medium. Serial dilutions of these cultures were plated to solid medium, exposed to 55°C for the indicated times, and then incubated at 30°C for 3-4 days. The number of colonies formed was a measure of the fraction of cells resistant to the heat shock treatment imposed. The strains analyzed were wild-type (PHY2195) and rpb1-104 (PHY2193). regulated by nutrient limitation and the Ras/PKA signaling pathway. The expression of four of these genes, CTT1, GUT2, PGM2, and YOR173w, is dependent upon the Msn2p and Msn4p transcriptional regulators (60,62,63). Msn2p and Msn4p are key mediators of the yeast cell response to environmental stress and are themselves regulated by Ras/PKA activity (9,64). The other three genes, APE2, UGA1, and UGA2, are regulated by Ras signaling activity in a Msn2p/Msn4p-independent manner (60). These genes are all expressed at a low level during mitotic growth and are induced during the diauxic shift (data not shown) (11,60,61).
For this study, we examined the expression levels of these Ras-regulated genes at times that correlated roughly with late log phase and the diauxic shift (Fig. 5A). The levels present in the RAS2 Val19 and rpb1-104 mutants were compared with those in the isogenic wild-type strain. As expected, we found that all seven genes were induced at the diauxic shift in wildtype cells and that this induction was impaired in the RAS2 Val19 mutant (Fig. 5B). In general, the expression levels were 2-10-fold lower in the RAS2 Val19 strain (Fig. 5C). Interestingly, the induced level of all of these genes, except GUT2, was found to be diminished to a similar extent in the rpb1-104 mutant (Fig. 5, B and C). These data indicated that rpb1-104 mutants were defective in the initial transcriptional response to nutrient deprivation. The subsequent stationary phase phenotypes observed with these mutants are likely a consequence of these transcription defects.
Elevated Levels of Ras/PKA Signaling Did Not Affect the Phosphorylation State of the Rpb1p CTD-Because many of the functions of the Rpb1p CTD are regulated by serine phosphorylation (21,24), we examined the phosphorylation state of this domain in RAS2 Val19 mutants. Previous work (65) has shown that Rpb1p is found in two forms in vivo: a highly phosphorylated form, known as II 0 , and a hypophosphorylated form, known as II a . The degree of CTD phosphorylation is indicated by the relative amount of Rpb1p found in the II 0 form. We found that RAS2 Val19 and wild-type cells contained a similar fraction of Rpb1p in the highly phosphorylated II 0 form (Fig. 6A).
The phosphorylation state of the Rpb1p CTD was also assessed by immunoblotting experiments with antibodies that specifically recognize the phosphorylated forms of the serines at positions 2 and 5 of the CTD repeats (37,66). The relative signal in the immunoblot serves as a measure of the extent of phosphorylation at that particular serine residue. This analysis also indicated that RAS2 Val19 had no significant effect on Mid-log cultures of the following strains were diluted to an A 600 of 0.2 per ml and were incubated at 30°C: wild-type (PHY2195, squares), rpb1-104 (PHY2193, circles), and RAS2 Val19 (PHY2194, triangles). The T1 time point corresponds to late log phase, and T2 corresponds roughly to the beginning of the diauxic shift. Note that the RAS2 Val19 and rpb1-104 mutants attained a lower culture density at stationary phase than the isogenic wild-type strain. B, RAS2 Val19 and rpb1-104 mutants exhibited similar defects in the transcriptional response to nutrient deprivation. Total RNA was prepared from cells harvested at the T1 and T2 time points shown in A. The levels of the indicated mRNAs were assessed by Northern RNA blot analysis as described under "Experimental Procedures." The ACT1 levels served as a loading control for these experiments. C, PhosphorImager quantification of the Northern blot data presented in B. The mRNA levels at time point T2 in rpb1-104 (black bars) or RAS2 Val19 (gray bars) cells is expressed as a fraction of the level detected in the isogenic wild-type control.
FIG. 6. The presence of RAS2 Val19 did not have a significant effect on the phosphorylation state of the Rpb1p CTD. A, analysis of CTD phosphorylation in RAS2 Val19 mutants. The relative level of Rpb1p phosphorylation was assessed by immunoblot analysis with antibodies that recognize unphosphorylated CTD repeats (␣-CTD; monoclonal 8WG16) and phosphorylated serine 2 (␣-Ser-P2; monoclonal H5) or serine 5 (␣-Ser-P5; monoclonal H14) residues in the CTD repeats. Protein extracts were prepared from mid-log cultures of wild-type or RAS2 Val19 cells as described under "Experimental Procedures." Relative protein concentrations were assessed by Bradford assays and immunoblotting with antibodies to phosphoglycerate kinase (PGK) and Vps34p, Dbp5p, and carboxypeptidase Y (data not shown). The strains analyzed were PHY1220 containing either a control vector, pRS416 (RAS2), or a RAS2 Val19 plasmid, pJR1040 (RAS2 Val19 ). B, CTD phosphorylation levels decreased in kin28 mutants. Top, CTD phosphorylation was assessed by Western immunoblotting with the 8WG16 monoclonal antibody as described in A. Protein extracts were prepared from wild-type (WT) (GF262-2) and kin28 (JGV4) cells after a 0-and 30-min shift to the nonpermissive temperature of 39°C. Bottom, the levels of serine 5 phosphorylation in the CTD repeats were assessed by Western immunoblotting with monoclonal antibody H-14. Following a 30-min incubation at 39°C, protein extracts were prepared from wild-type (GF262-2; RAS2), RAS2 Val19 (GF262-2 with pJR1040; RAS2 Val19 ), and kin28 (JGV4; kin28) cells. The levels of phosphoglycerate kinase served as a loading control for this experiment. the relative amount of phosphorylation at either serine 2 or serine 5 (Fig. 6A). As a control for antibody specificity, CTD phosphorylation levels were also assessed in a kin28 ts mutant. KIN28 encodes a CTD kinase, and the inactivation of Kin28p results in a significant decrease in the level of CTD phosphorylation (44,67). Following a 30-min incubation at the nonpermissive temperature, we found that the CTD phosphorylation signal detected with the 8WG16, H5, and H14 antibodies was reduced to background levels in the kin28 strain ( Fig. 6B and data not shown). Therefore, the effects of elevated Ras signaling on RNA pol II transcription did not involve a gross alteration in the pattern of CTD phosphorylation.
The sin4 RAS2 Val19 Synthetic Lethality Was Suppressed by Mutations Affecting the Mediator-Previous studies have identified mutations that suppress the growth defects associated with truncations of the Rpb1p CTD. These SRB mutations were subsequently found to reside in genes encoding components of the Mediator (SRB2 and SRB4-7) and Srb (SRB8-11) complexes found in the RNA pol II holoenzyme (40,41,68). Both of these complexes physically interact with the Rpb1p CTD and influence RNA pol II transcription (24,69). We therefore tested whether these SRB mutations would also suppress the RAS2 Val19 rpb1-104 synthetic lethality. Interestingly, only those SRB mutations that affected Mediator components were able to suppress the Ras-induced lethality (Fig. 7A). These results indicated that the two classes of SRB mutation were functionally distinct and that simply suppressing the growth defects associated with CTD truncations might not be sufficient to alleviate the lethal effects of RAS2 Val19 . In addition, these data ruled out models proposing that the Ras effects on RNA pol II transcription observed here were due to the Ras/PKA stimulation of the inhibitory activity associated with the Srb complex. The Srb complex has been shown to function as a negative regulator of transcription at a select set of promoters (24).
Finally, to gain some insight into the nature of the Ras effects on CTD function, we tested whether the dominant SRB mutations would also suppress the growth defects associated with RAS2 Val19 . This suppression might be expected if Ras activity was directly affecting the Rpb1p CTD. However, we found that neither SRB4-1, SRB5-1, nor SRB6-1 was able to suppress the Ino Ϫ or Gal Ϫ growth defects of RAS2 Val19 mutants (data not shown). In addition, neither of these SRB alleles suppressed the growth defects associated with kin28 and tfa1 mutants (data not shown). In contrast, these kin28 and tfa1 defects were suppressed by decreasing the level of Ras/PKA signaling in these mutants (see above). Interestingly, the Srb2p, Srb4p, Srb5p, and Srb6p proteins appear to form a core subcomplex within the Mediator that is in close contact with the RNA pol II enzyme (70,71). This core complex may be responsible for relaying signals received from other Mediator proteins to the polymerase; these other Mediator proteins would be the components that physically interact with DNAbound transactivators. With such an architecture, our genetic data would suggest that the Ras/PKA pathway is targeting a component that functions either downstream of the Mediator or in an independent pathway that works through the Rpb1p CTD to regulate RNA pol II transcription. DISCUSSION This report describes a functional interaction that exists between the S. cerevisiae Ras/PKA signaling pathway and the C-terminal domain, or CTD, of Rpb1p, the largest subunit of RNA pol II. This connection was initially suggested by the lethality observed when Ras/PKA activity was elevated in mutants defective for RNA pol II transcription. The ensuing analysis indicated that the Rpb1p CTD was critical for these genetic interactions as mutations that compromised CTD function were the most sensitive to changes in Ras/PKA signaling activity. Altogether, the data suggested that Ras/PKA activity was influencing RNA pol II transcription at select promoters by regulating the function of the Rpb1p CTD. Because Ras/PKA signaling did not significantly affect the phosphorylation state of the CTD, we suggest that the Ras/PKA pathway is instead targeting proteins that function in concert with the Rpb1p CTD.
In the simplest sense, two types of models could be invoked to explain the genetic data presented here. In the first, the Ras/PKA pathway would be targeting some process unrelated to RNA pol II transcription. The observed growth defects would then be due to the additive effects of compromising this second process in a cell that is already defective for RNA pol II transcription. In the other model, the Ras/PKA pathway would be directly targeting a component of the RNA pol II transcription apparatus. In this case, the growth defects would be due to the synergistic effects of inhibiting transcription at two possibly related points. We feel that the data presented here are better explained by the latter model for the following reasons. First, the genetic interactions observed with the Ras/PKA pathway FIG. 7. A novel mechanism for the control of gene expression by the Ras/PKA signal transduction pathway. A, the RAS2 Val19 rpb1-104 synthetic lethality was suppressed specifically by SRB mutations in genes encoding components of the Mediator. rpb1-104 strains with the indicated SRB alleles were transformed with either a control vector (RAS2) or a RAS2 Val19 plasmid, pJR1040 (RAS2 Val19 ). These strains were then plated to a YM-glucose minimal medium and incubated for 3 days at 34°C. The strains analyzed were Z551 (SRB), Z814 (SRB2-2), Z552 (SRB4-1), Z553 (SRB5-1), Z554 (SRB6-1), Z695 (srb8-1), Z696 (srb9-1), and Z711 (srb11⌬). B, a model suggesting that signaling pathways target components within the RNA pol II holoenzyme. Top, a traditional model for the control of gene expression where the Ras/PKA pathway regulates the activity of specific DNA-bound transactivators. In this model, PKA directly phosphorylates each of the individual transactivators. Bottom, a schematic depicting a new model for the control of gene expression where Ras/PKA signaling regulates the activity of proteins associated with the RNA pol II holoenzyme or the Rpb1p CTD. In this model, PKA directly phosphorylates a holoenzyme protein that is responsible for mediating transcription by a number of different transactivator proteins. See the text for more details.
were limited to mutations affecting particular aspects of RNA pol II transcription; not all mutations affecting RNA pol II transcription were influenced by Ras signaling activity. In fact, many of the mutations that did not exhibit a synthetic lethal interaction with RAS2 Val19 result in a much more severe defect in transcription than do the Rpb1p CTD truncations. Second, increasing the levels of Ras/PKA signaling resulted in phenotypes reminiscent of those associated with a loss of CTD function and vice versa. Third, decreased signaling through the Ras/PKA pathway suppressed the growth defects associated with kin28 and tfa1 mutants. Both Kin28p and Tfa1p affect RNA pol II transcription, in part, by regulating the phosphorylation of the Rpb1p CTD (46). Fourth, not all of the srb mutations suppressed the RAS2 Val19 rpb1-104 synthetic lethality. All of these srb mutations suppress the growth defects associated with CTD truncations, and the first model would predict that these srb mutations would suppress the double mutant lethality as well. Altogether, these data suggest that the RAS2 Val19 lethality observed here is not merely the simple sum of a defect in transcription coupled with a defect in some other unrelated process important for cell growth. Instead, the data appear to support a model where Ras/PKA activity targets proteins important for RNA pol II transcription.
Clearly, one of the main priorities for future work will be the identification of the PKA substrate responsible for the genetic effects characterized in this study. Previous work with the yeast and mammalian PKA enzymes has identified a consensus target site of RRX(S/T)B, where X indicates any amino acid and B refers to those amino acids with a hydrophobic side chain (72,73). Unfortunately, there are few candidate proteins within the RNA pol II transcription apparatus that contain a perfect match to this consensus. One of the best potential sites we identified was an RRRSS sequence present within the Mediator protein, Med1p (74). However, our analysis of Med1p has indicated that this protein is not the PKA target responsible for the genetic interactions observed in this study. 2 This inability to identify the PKA substrate by sequence comparison is not altogether surprising. One of the primary difficulties encountered during the study of protein kinases is the variation typically found in the target sites that are phosphorylated (75). Therefore, the relevant PKA target could possess a phosphorylation site that varies enough from the consensus that it would be rather difficult to recognize a priori. As a result, we are presently carrying out more classical genetic approaches that do not require any prior knowledge of the nature of the PKA target. One example of such a strategy is a genetic screen that aims to identify mutations that suppress the synthetic lethality of the rpb1-104 RAS2 Val19 double mutant. Such genetic approaches should allow us to identify the PKA substrate responsible for the Ras effects on RNA pol II transcription that are described in this study.
In all, the data presented here suggest that the Ras/PKA pathway is negatively regulating proteins that function through the Rpb1p CTD. At first glance, this might appear to be counterintuitive as both Ras/PKA activity and the CTD are essential for yeast cell growth. However, a possible explanation is suggested by studies of rpb1 alleles that encode proteins with a truncated CTD. These rpb1 mutants, like rpb1-104, exhibit defects in expression from a distinct set of genes; the transcription from many, and perhaps most, promoters is apparently unaffected by the shortening of the CTD (58,76). This result indicates that particular genes require more CTD activity for their normal levels of expression. Therefore, the Ras pathway could affect the expression of specific sets of genes by down-regulating the activities of proteins associated with the Rpb1p CTD. Because Ras signaling is generally thought to serve as an indicator of environmental growth conditions (4,9,38), the affected genes should include those that would be preferentially expressed under conditions that are less than ideal for yeast cell growth. Clearly, this correlation holds for most of the genes examined in this study. In wild-type cells, the effects of the Ras pathway would presumably not be able to significantly dampen the expression of those genes required for growth. However, compromising the RNA pol II machinery by a second mutation could cause RAS2 Val19 to limit expression from essential genes and thus result in the synthetic lethality observed in this study.
The possibility that the Ras/PKA pathway could be regulating gene expression by targeting components within the RNA pol II holoenzyme is especially intriguing. Although no definitive examples have yet been described, the possibility of this type of a control mechanism has been alluded to in a number of previous studies (12,17,18,(77)(78)(79). In essence, this model suggests that in addition to targeting DNA-bound transcriptional regulators, signaling pathways might also regulate the activities of components within the RNA pol II holoenzyme (Fig. 7B). In this manner, signal transduction pathways would be able to control transcription from a large number of promoters in a single regulatory step. This type of an approach would be an efficient and parsimonious way to bring about coordinated changes in the patterns of gene expression. Indeed, this type of a mechanism might be used by the Snf1p protein kinase. Snf1p has been shown to associate physically with the RNA pol II holoenzyme and thereby elicit a significant change in the yeast transcriptosome (17). In addition, the holoenzyme proteins, Srb9p, Srb10p, and Srb11p, are likely targets of a signal transduction pathway responsible for coordinating yeast cell growth with nutrient availability (12,77). Finally, the data presented in this report suggest that the Ras/PKA signaling pathway influences RNA pol II transcription by modulating activities associated with the Rpb1p CTD. Further studies will be necessary to ascertain whether this type of a regulatory mechanism is indeed used to control gene expression in eukaryotic cells.