Experiments were primarily performed in the Saccharomyces cerevisiae W303 background (Ronne and Rothstein, 1988) and are listed in Supplementary Table S1. The endogenously tagged CDK8-myc and kinase dead CDK8-myc strains were previously described (Chi et al., 2001). Strains used to monitor meiosis were derived from crosses between rapidly sporulating SK1 and W303, which provided efficient sporulation without premature meiotic induction (Cooper et al., 1997). Other strains were constructed using standard replacement methodologies (Janke et al., 2004). The endogenous RIM15 mutations (rim15 ^S24E , rim15 ^S24A , rim15 ^S3E and rim15 ^S3A ) were introduced using pop in, pop out, technology (Langle-Rouault and Jacobs, 1995). The URA3 integrating plasmids (listed in Supplementary Table S2) containing the mutations were linearized with BglII and transformed into wild-type cells (RSY10). These transformants were then counter-selected on 5-FOA, and resistant mutants confirmed by PCR. The Y2H assays were performed in the Y2H Gold strain (TaKaRa 630,489, PT4084-1; Matchmaker Gold Yeast Two-Hybrid System). The Saccharomyces cerevisiae Genome database identified members of the CDK8 module as SSN8/CNC1/UME3/SRB11, SSN3/CDK8/UME5/SRB10, SRB8/MED12/SSN5 and SSN2/MED13/UME2/SRB9. In this report, we will use CNC1, CDK8, MED12, and MED13 gene designations in accordance with the Mediator complex universal naming agreement (Bourbon et al., 2004). Plasmids used in this study are listed in Supplementary Table S2. The RIM15 integrating plasmids were constructed by first PCR-cloning wild-type RIM15 from RSY10 into pRS306 (Sikorski and Hieter, 1989). The rim15 ^S3A and rim15 ^S3E mutations were then introduced using site directed mutagenesis (Quick Change Kit, New England Biolabs). The Hsp26-mCherry reporter was constructed by first amplifying 1,000 bp upstream of the start of HSP26 through the open reading frame (ORF) from RSY0 and then inserting mCherry in frame at the C-terminal end of the ORF (pSW270) into pRS314. The pACT-T7 plasmid is a modification of the pACT2 plasmid (Van Criekinge and Beyaert, 1999) in which SDM was used to replace the HA epitope with T7. Additional plasmid construction details are available on request. All constructs were verified by DNA sequencing.

Yeast cells were grown in either rich, non-selective medium (YPDA: 2% [w:v] glucose, 2% w:v Bacto peptone, 1% w:v yeast extract, 0.001% w:v adenine sulfate) or synthetic minimal dextrose medium (SD: 0.17% w:v yeast nitrogen base without amino acids and ammonium sulfate, 0.5% w:v ammonium sulfate, 1x supplement mixture of amino acids, 2% w:v glucose) allowing plasmid selection as previously described (Cooper et al., 1997). For the nitrogen-starvation experiments, cells were grown to mid-log in SD medium, harvested by centrifugation, washed in 2x volume of water, and resuspended in SD-N medium (SD: 0.17% w:v yeast nitrogen base without amino acids and 2% w:v glucose) for the indicated time points. For rapamycin treatment, cells were grown to mid-log and rapamycin (Biovision, dissolved in 90% ethanol, 10% Tween-20) was added at 200 ng/ml for indicated time points. To monitor meiosis, diploid cells were plated onto SPIII medium (2% potassium acetate, 0.1% dextrose, 0.25% bacto-yeast extract, supplemented with essential amino acids at 7.5 mg/L (Klapholz et al., 1985). After 3 days at 30°C the cells were fixed in 70% ethanol and stained with DAPI diamidino-2-phenylindole (DAPI). At least 200 cells from three independent cultures were assayed for the appearance of bi- and tetra-nucleated cells to indicate the execution of one or two meiotic divisions.

For yeast survival plating assays, cells were grown to mid-log in SD media before being subjected to various stresses. The cells were then spotted on YPD using 10-fold serial dilutions. For nitrogen starvation, cells were washed in water before being resuspended in SD-N and plated at the time indicated. For stationary phase, the 0 h time point was taken at mid-log, then cells were allowed to continue growing, and plated at time points indicated. For rapamycin, H₂O₂ (Sigma, HX0635-3), and tert-Butyl hydroperoxide (Alfa Aesar, A13926) treatments, the respective chemicals were added to mid-log cells at the concentrations and time indicated, and then plated. For MMS (Sigma, M4016), Sorbitol, NaCl, Congo Red and SDS stresses, the cells were grown to mid-log and then plated onto YPD plates containing the stress at the indicated concentration. For survival at 37°C, the mid-log cells were incubated at 37°C for 48 h before being scored. For all other stresses, the results were recorded after 2 days at 30°C. The Hsp26-mCherry plating assays were executed as described (Luu and Macreadie, 2018). Cells harboring pSW270 were grown overnight in SD medium selecting for the plasmid. 5 µl of cells were then spotted to a SD plate and grown for 5 days at 30°C. The plate was then imaged using an iBright FL1500 imaging system (Thermo) using the TexasRed filter set. Glycogen accumulation assays were done as previously described (Chester, 1968). Three individual colonies from each strain tested were grown overnight, and then 3 µl plated onto YPD plates. After 6 days at 30 °C the plate was inverted over a glass jar containing 250 mg of iodine crystals (BeanTown Chemical, 219,030) so that the plate was three inches from the crystals, and an airtight seal was formed. The jar was then heated at 50°C until the WT cells started to develop a brown color, indicating glycogen staining by the iodine vapor. Fluorescent dityrosine assays to monitor spore wall formation were executed exactly as described (Briza et al., 1990).

RT-qPCR analyses were executed as previously described (Cooper et al., 2012). Oligonucleotides used during these studies are available upon request. All RT-qPCR studies were conducted with three biological samples in technical duplicates. The nitrogen starvation viability assays were executed in biological triplicate exactly as described (Willis et al., 2020) with 30,000 cells counted per timepoint using fluorescence cell analysis. p values were determined using the unpaired Students t-test. Data are mean ± standard deviation.

Western blot analysis was performed with protein extracts prepared using NaOH lysis of 25 ml of mid-log cells per timepoint exactly as described (Willis et al., 2018). To analyze cyclin C protein, extracts were prepared using glass bead lysis as previously described (Stieg et al., 2018). In short, 50 ml of mid-log cells were lysed in Ripa V buffer (50 mM Tris pH 8, 150 mM NaCl, 0.158% Sodium deoxylcholate, 1% NP-40) supplemented with 1 mM PMSF, 14 mM ß-mercaptoethanol, 1 μg/ml pepstatin, 1 μg/ml leupeptin, and 1x protease inhibitor (GoldBio, GB-333). 50 µg of total protein was subject to western analysis. To detect proteins, 1:5,000 dilutions of anti-MYC (UpState/EMD Millipore Corp,05–724), anti-HA (Abcam, ab9110), anti-mCherry (Novus, NBP1-96752), anti-T7 (EMD Millipore, 69,522), anti-GFP (Wako Pure Chemical Corp, 012–20,461), or anti-Pgk1 (Invitrogen, 459,250) antibodies were used. Western blot signals were detected using 1:5,000 dilutions of either goat anti-mouse (Abcam, ab97027) or goat anti-rabbit (Abcam, ab97061) secondary antibodies conjugated to alkaline phosphatase and CDP-Star chemiluminescence kit (Invitrogen, T2307). Signals were quantified relative to Pgk1 controls using CCD camera imaging (iBright, Thermo). All quantified assays were performed at least in duplicate, as indicated in the figures. Standard deviation and significance were calculated from the mean ± standard deviation using GraphPad Prism 7.

For co-immunoprecipitation experiments, 1 L of cells were grown to mid-log and treated with 200 ng/ml Rapamycin. Protein extracts were prepared using a glass bead lysis method in holoenzyme lysis buffer (50 mM HEPES pH 7.5, 250 mM potassium acetate, 5 mM EDTA, 0.1% NP-40) supplemented with protease inhibitors as above. One mg of total soluble protein was immunoprecipitated per timepoint from the whole cell lysate. Anti-Myc antibodies or anti-HA antibodies were used for immunoprecipitations. The immunocomplexes were then collected using Protein G beads (GoldBio, P-430) washed in IP buffer (25 mM Tris pH 7.4, 150 mM NaCl) then subjected to western analysis. For the input controls, 100 μg of protein was immunoprecipitated from whole-cell lysates.

The Rim15^PAS domain E. coli expression construct (pSW222) contains the coding sequence for amino-acids 1–221 amplified from wild-type cells (RSY10) and cloned into pGEX-4T1 (Sigma). Expression was induced in BL21 DE3 cells by addition of 0.5 mM IPTG (GoldBio, I2481C) at 16°C for 16 h. Cells were then pelleted and resuspended in 75 ml of GST lysis buffer (50 mM HEPES pH 7.5, 150 mM KCl, 14 mM ß-ME, 1 mM PMSF). The cells were then lysed by passing them through a fluid homogenizer (Microfluidizer LM10) two times at 10,000 psi. Cell debris was then pelleted at 40,000 x g for 30 min and the resulting supernatant was incubated with 1 ml of GST beads (Invitrogen, G2879) for 16 h with gentle shaking at 4 °C. The slurry was then centrifuged at 700 x g for 5 min at 4 °C, the supernatant discarded, and the beads washed twice in GST lysis buffer. The protein was then eluted using GST lysis buffer supplemented with 10 mM glutathione (GoldBio, G-155). The elution was then concentrated and desalted in an Vivaspin 500 column (LifeTech, 28-9322-25) and the final concentrated aliquots were supplemented to 10% glycerol prior to freezing at −80°C.

The in-vitro kinase assays were performed with Cdk8-myc and Cdk8-myc kinase Dead expressing cells (YC17 and YC17) as follows. Mid-log cells grown in SD medium were harvested and resuspended in kinase lysis buffer (30 mM HEPES pH 7, 100 mM potassium acetate, 1 mM EDTA, 1 mM MgCl₂, 10% Glycerol supplemented with 2 mM DTT, 200 mM potassium acetate, 1 mM PMSF, 1 μg/ml Leupeptin, 1 μg/ml Pepstatin and 1x protease inhibitor cocktail). The cells were lysed by glass bead disruption (8 × 30 s with 1 min on ice) and the lysated cleared by centrifugation at 20,000 x g for 15 min. For immuno-precipitations, 1 mg of protein extract (quantified by Bradford assay, BioRad) were first pre-cleared with protein A bead (GoldBio, P-400) for 1 h at 4°C with gentle agitation. One µg of anti-Myc antibody (Upstate) was added and incubated for 2 h at 4°C. The immune complexes were then collected using protein A beads for 1 h at 4°C. The beads were washed 4 times in kinase lysis buffer and the beads resuspended in kinase assay buffer (30 mM HEPES pH 7, 10 mM MgCl₂, 0.1 mM Sodium Orthovanadate, 5 mM B- Glycerophosphate) supplemented with 15 mM ATP, and 10 µCi of γ-³²P-ATP (Perkin Elmer, NEG502A) and 50 µg of purified Rim15^PAS. The reaction was incubated for 30 min at 25°C with the reaction terminated by the addition of 10x loading dye. The reaction was then run out on an SDS-PAGE gel and stained with InstantBlue (Expedeon, ISB1L) to visualize input, vacuum dried, and visualized by autoradiography.

For all microscopy studies, cells were grown to mid-log phase, washed, and resuspended in SD-N with rapamycin for the time points indicated. Deconvolved images were obtained using a Nikon microscope (Model E800) with a ×100 objective with ×1.2 camera magnification (Plan Fluor Oil, NA 1.3) and a CCD camera (Hamamatsu Model C4742). Data were collected using NIS software and processed using Image Pro software. All images of individual cells were optically sectioned (0.3-μM spacing), deconvolved and slices were collapsed to visualize the entire fluorescent signal within the cell. The nuclei were visualized in live cells using either the nuclear marker Nab2-mCherry or the nuclear pore marker Nup1-mCherry. Linear quantification analysis was measured using the NIS software. Single plane images were obtained using a Keyence BZ-X710 fluorescence microscope with a ×100 objective with ×1.0 camera magnification (PlanApoλ Oil, NA 1.45) and a CCD camera. Data were collected using BZ-X Analyzer software.

All representative results included at least two independent biological experiments. The actual number is given in the figure legends. p values were generated from Prism-GraphPad using unpaired Student’s t-tests; NS p ≥ 0.05; *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.005; ****p ≤ 0.001. All error bars indicate mean ± SD. For quantification of Hsp26-mCherry degradation kinetics, band intensities of each time point were first normalized to the unstressed, T = 0 band intensity. These values were then divided by Pgk1 loading control values that were also normalized to their T = 0 intensities. p-values shown are relative to wild-type T = 4 timepoints.

Previous studies have shown that cyclin C-Cdk8 and Rim15 repress or induce the transcription of an overlapping subset of stress responsive genes (Bartholomew et al., 2012; Willis et al., 2020). To investigate whether cyclin C-Cdk8 and Rim15 act in the same or separate pathways, epistasis analyses were conducted. Rim15 is required for entrance into, and recovery from, quiescence (Boy-Marcotte et al., 1998; Klosinska et al., 2011; Pfanzagl et al., 2018; Shi et al., 2010). Initially, we monitored viability after maintaining rim15∆, cdk8∆ or double mutants in stationary phase or in nitrogen depleted medium (SD-N) for 6 or 9 days. As expected, rim15∆ cells exhibited a dramatic loss in viability in stationary phase compared to the wild-type control or cdk8∆ mutants (Figure 2A). However, the rim15∆ cdk8∆ double mutant displayed a phenotype similar to the rim15∆ strain but not as severe. These results indicate that rim15∆ is epistatic to cdk8∆. To determine if the double mutant indeed was defective for entering quiescence, we tested for the presence of the storage carbohydrate glycogen, a marker of stationary phase (Reinders et al., 1998; Cao et al., 2016). Similar to the return to growth studies, both the rim15∆ and rim15∆ cdk8∆ mutants failed to accumulate glycogen (Figure 2B). Similar results were also obtained with the more stringent test for survival following prolonged incubation in SD-N (Figure 2C). Again, rim15∆ was epistatic to the cdk8∆ mutation. Although informative, these plating assays cannot distinguish between cells dying during prolonged starvation and those unable to reenter mitotic cell division. Therefore, we used an inviability dye (Phloxine B) to determine the percentage of dead cells in the population. These studies revealed that prolonged incubation in SD-N resulted in extensive cell death in rim15∆ and rim15∆ cdk8∆ (Figure 2D). Taken together, these results suggest that the Rim15 functions downstream or independent of Cdk8 repressor activity. However, the double mutant always partially phenocopied rim15∆ alleles. One possible explanation is that Rim15 is required for cyclin C destruction to fully inactivate Cdk8. However, cyclin C proteolysis kinetics were not altered in a rim15∆ mutant (Supplementary Figure S1). Therefore, the partial suppression observed in the double mutant may be due to Rim15-independent control of SRG transcription.

To directly measure the relationship between Cdk8 and Rim15 activities, we used RT-qPCR to monitor steady state mRNA levels of four genes (CTT1, HSP12, DDR2 and HSP26) that are both repressed by Cdk8 and activated by Rim15 (Barbet et al., 1996; Holstege et al., 1998). As anticipated, all four genes were derepressed in unstressed cdk8∆ cells while expression was at background levels in the rim15∆ strain (Figure 3A). However, deleting RIM15 in the cdk8Δ mutant reduced the transcription of these genes but not to background levels. These results indicate that the derepression observed in cdk8∆ mutants was mostly dependent on Rim15. In addition, the partial suppression observed for HSP12 and DDR2 argues that additional positive factors are present in unstressed cells. To determine if these changes in transcription also resulted in differences in protein levels, Western blot analysis revealed that Hsp26-mCherry levels were 2-3 fold higher in cdk8∆ cells but not in cdk8∆ rim15∆ (Figure 3B, Supplementary Figure S2A). Similarly, visualizing the fluorescent Hsp26-mCherry signal in cells after stationary phase entry (Luu and Macreadie, 2018) revealed a strong signal in wild-type and cdk8∆ cells, weaker in the cdk8∆ rim15∆ double mutant and absent in rim15∆ (Figure 3C). These results are consistent with the partial suppression of the cdk8∆ phenotype by deleting RIM15 observed in viability assays (Figure 2).

Finally, Atg8 is required for autophagy and destroyed in the vacuole. ATG8 transcription is repressed and activated by Cdk8 and Rim15, respectively (Bartholomew et al., 2012; Willis et al., 2020). We used a cleavage assay to monitor GFP-Atg8 levels and the production of free GFP, a hallmark of vacuole proteolysis. Compared to wild type, GFP-Atg8 levels were elevated in cdk8∆ mutants in replete medium (compare 0 h signals, Figure 3D). Although GFP-Atg8 levels are elevated in cdk8∆ mutants, no free GFP is observed indicating that the autophagic pathway had not initiated. Timepoints taken following nitrogen deprivation revealed characteristic GFP-Atg8 degradation with release of intact GFP (24 h timepoint). As expected, GFP-Atg8 was not induced in the rim15∆ mutant with corresponding reduction in free GFP compared to wild type or the cdk8∆ mutant (compare 24 h timepoints). The double mutant again displayed an intermediate level indicating a partial suppression of the cdk8∆ phenotype (Figure 3D, Supplementary Figures S2B,C). These results indicate that Cdk8 and Rim15 genetically interact with Rim15 functioning either downstream or independent of Cdk8 to control the expression of genes controlling multiple responses to nitrogen deprivation.

Cyclin C-Cdk8 is a nuclear protein kinase and our previous studies have shown that cyclin C is destroyed by the UPS following unfavorable environmental cues in including nitrogen starvation and rapamycin stress (Willis et al., 2020). On the other hand, Rim15 cycles between the nucleus and the cytoplasm being retained in the nucleus to activate gene expression only following stress (Wanke et al., 2005). This suggests a possible model in which cyclin-Cdk8 phosphorylation of Rim15 represses its nuclear activity in replete media (Figure 4A). Consistent with this idea, Cdk8 and Rim15 co-immunoprecipitated in unstressed cells and this interaction decreased after the addition of rapamycin (Figure 4B). To investigate if cyclin C-Cdk8 kinase can directly interact with Rim15, we used a two-hybrid approach to isolate the interaction domain between Cdk8 and Rim15. Rim15 is a large protein, containing several known domains and phosphorylation sites (Figure 4C). The two-hybrid studies revealed that Cdk8 interacts with the N-terminal PAS-ZnF domain of Rim15 (Figure 4D, Supplementary Figure S3). An interaction was also observed between the C-Terminal Disordered-Regulatory domain (DDII REG). However, this domain self-activated the ADE2 and HIS3 reporters, even at high levels of the 3-AT competitor. Therefore, this region was eliminated from further investigation as a false-positive (Supplementary Figure S3). Further analysis revealed that the PAS domain (amino acids 1–121) is sufficient to bind to Cdk8 using Y2H assays (Figure 4E). In vitro kinase assays using GST-Rim15¹⁻¹²¹, confirmed that Cdk8 phosphorylates this region, whereas the well characterized kinase dead mutant exhibited a reduced signal (Figure 4F). Taken together, these data support a model that Cdk8 directly phosphorylates Rim15 in replete media to inactivate the kinase thereby suppressing SRG transcription.

Rim15 function is tightly regulated by multiple kinases that modulate its subcellular address (Deprez et al., 2018). In unstressed nutrient replete cells, Rim15 is predominantly held in the cytoplasm through its interactions with the 14-3-3 binding protein Bmh2 (Wang et al., 2009). In addition, a population of Rim15 cycles through the nucleus with its exit dependent on phosphorylation by the nuclear Pho80-Pho85 protein kinase and the Msn5 exportin (Kaffman et al., 1994; Wanke et al., 2005). In these studies, due to the transient nature of Rim15’s passage through the nucleus, only kinase dead GFP-Rim15 (C1178Y), but not wild type, was detected in the nucleus in replete media (Wanke et al., 2005). GFP-rim15 ^C1178Y was also captured in the nucleus of cdk8∆ cells in replete media suggesting that Cdk8 kinase activity is not required for Rim15 nuclear localization (Figure 5A). By tagging endogenous Rim15 with mNeongreen, which is significantly brighter than GFP in yeast (Botman et al., 2019), we were also able to visualize endogenous Rim15 both in the cytoplasm and nucleus, in wild type and cdk8∆ (Figure 5B). Linear profile analysis was used to confirm the presence of Rim15 in the nucleus, which was marked with an inner nucleoporin, Nup1-mCherry (Supplementary Figures S4A,B) (Meszaros et al., 2015). Taken together, these results support the model that Cdk8 kinase activity prevents nuclear Rim15 from becoming active when it cycles through the nucleus in replete conditions, but does not affect its localization.

Cdk8, is a serine/threonine protein kinase, and phosphorylates SP/TP sites. The PAS region of Rim15 contains three SP/TP sites (amino acids S24, S68 and S84). Previous large scale phosphoproteomic studies identified S24 (Albuquerque et al., 2008) as a potential functional phosphorylation site that is also predicted to regulate protein complex architecture (Lanz et al., 2021). To test this, S24 was mutated to the phosphomimetic glutamic acid in the Rim15 endogenous loci. This substitution mimics S-phosphorylation by matching the phosphate oxygens of phosphorylated serine. It also provides the negative charge of this phosphoresidue and is of a similar volume (Perez-Mejias et al., 2020). As a control, we created an endogenous rim15 ^S24A strain in which the potential Cdk8 regulated serine residue was mutated to alanine residues which cannot be phosphorylated. Functional analysis of these mutants was assayed by monitoring (Figure 6A), glycogen storage (Supplementary Figure S5A) and Hsp26-mCherry accumulation (Supplementary Figure S5B), revealed that they do not alter the function of Rim15. As Cdk consensus motifs are frequently clustered in CDK substrates and CDK targets can be phosphorylated at multiple residues (Moses et al., 2007) we created a strain in which all three serines were mutated to either to glutamic acid (referred to as rim15 ^S3E ) or alanine (referred to as rim15 ^S3A ). The mutations did not affect the steady state levels of Rim15 (Supplementary Figure S5C, quantified in Supplementary Figure S5D). Similar to rim15∆, we observed a loss of viability following recovery from stationary phase in rim15 ^S3E but not rim15 ^S3A (Figure 6B). Likewise, only the rim15 ^S3E mutant behaved like rim15∆ when scored for glycogen storage in stationary phase (Figure 6C) and cell death in nitrogen deficient media (Figure 6D) whereas rim15 ^S3A behaved like wild-type cells. Lastly, inhibiting TORC1 using rapamycin had the same effect as did nitrogen depletion (Supplementary Figure S5E). Interestingly, the rim15 ^S3E mutation had no effect on the localization of Rim15 (Figure 6E). This is important as Rim15 kinase dead alleles are constitutively nuclear (see Figure 5A and previous studies (Wanke et al., 2005)). This suggests that even though rim15 ^S3E is a loss of function allele, it can be exported from the nucleus and retains kinase activity. Taken together, these results support the model that Cdk8 phosphorylation significantly represses Rim15 activity, as it traverses the nucleus in replete media. Derepression is of physiological relevance, as continuous inactivation of the PAS region results in starvation induced cell death.

To further test this model, we asked if the expression of genes controlled by the Cdk8-Rim15 axis changes following stress when the Cdk8 phosphomimetic mutant is the only source of Rim15. RT-qPCR analysis showed that two of the four genes we tested (CTT1 and HSP12) were induced at the same levels in rim15 ^S3E and wild-type cells. However, DDR2 and HSP26 showed a ∼50% decreased expression following 1 h rapamycin stress (Figure 7A). Confirming this, Hsp26-mCherry was only visualized in rim15 ^S3A but not rim15 ^S3E after 5 days in stationary phase (Figure 7B). In addition, a decrease in Hsp26-mCherry induction was seen in the rim15 ^S3E following treatment with rapamycin (Figure 7C, quantified Figure 7D). Likewise, induction of GFP-Atg8 was decreased in rim15∆ and rim15 ^S3E compared to wild type after 4 and 24 h in nitrogen starvation (Figures 8A,B). Cleavage of GFP from GFP-Atg8, which scores autophagic flux (Torggler et al., 2017), was significantly reduced when examined by Western blot analysis (Figure 8A and quantified in S2C). Similarly, significantly less free GFP was captured in vacuoles in rim15 ^S3E compared to wild-type after 24 h in SD-N (Figure 8C). Together this suggests Cdk8 phosphorylation of Rim15 is required to regulate a specific subset of genes controlled by Rim15.

Our previous studies have shown that cyclin C is destroyed in response to a subset of stress responses whereas it is unaffected by others (Cooper et al., 1999; Willis et al., 2020). Consistent with this model, recovery from the stresses that induced cyclin C degradation (oxidative stress induced by either 2 mM H₂O₂ or 2 mM tert-butyl hydroperoxide, 37°C heat shock and 0.1% methymethanesulfonate (MMS) induced DNA damage was decreased in the rim15 ^S3E mutant compared to both the S3A mutant and wild type controls (Figures 9A, Supplementary Figure S6A). Likewise, stresses which don’t affect cyclin C’s degradation (1 M sorbitol, 400 mM NaCl, 1 μg/ml Congo red and 25 μg/ml sodium dodecyl sulphate (SDS) did not affect the viability of rim15 ^S3E . (Figure 9B, Supplementary Figure S6B). These results support the model that inhibiting Cdk8 activity by cyclin C degradation activates Rim15 and promotes cell survival in response to a subset of specific environmental onslaughts.

Destruction of cyclin C is required for the full induction of several early meiotic genes including SPO13 (Cooper et al., 1997). Also Rim15 activity is required for activation of IME1, the master regulator of meiosis that activates early meiotic genes, via the Igo1/2 and Ume6 axis (Sarkar et al., 2014). Taken together, we predicted that Rim15 ^S3E diploids would be defective in entering the meiotic program. To test this, the number of cells completing meiosis was scored in wild-type and Rim15 ^S3E homozygous diploids. The results that show 60% of the Rim15 ^S3E diploids remain mono-nucleated, compared to 20% observed in wild type cells, indicative of a failure to complete any meiotic divisions. The Rim15 ^S3E diploids that were able to enter meiosis also were able to complete gametogenesis and form spores (arrow, Figure 9C, quantified in Figure 9D). Confirming this, Rim15 ^S3E diploids were positive when screened for the presence of the spore wall component dityrosine (Britza et al., 1986) whereas both haploid cells and a mutant deficient in spore wall assembly (ama1∆ (Cooper et al., 2000)) were negative (Figure 9E). These results support the model that inhibiting Cdk8 activity promotes Rim15 activation and is required for entry into the meiotic program.