Translational control of CAK and Cdk T-loop phosphorylation in response to growth in yeast

Cak1, Cdc28, uORF, stationary, chemostat .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 3

Cyclin-dependent kinases (Cdks) require activating T-loop phosphorylation, a modification that is considered constitutive. Here, we examined the regulation of the Cdk-activating kinase, Cak1, in budding yeast. We measured Cak1 levels and the activating T169 phosphorylation of Cdc28 (the budding yeast Cdk) across various nutrient environments. We found that the abundance of Cak1 and the T169 phosphorylation is significantly reduced in cells that are proliferating very slowly or have entered quiescence. A small upstream open reading frame (uORF) in the CAK1 transcript represses Cak1 synthesis, especially in poor growth conditions. Eliminating the uORF increased Cak1 levels but did not alter proliferation kinetics under most laboratory contexts. Instead, it reduced the viability of quiescent cells and the fitness of slowly proliferating chemostat cultures. In cells lacking several type 2C protein phosphatases, which remove the T169 phosphorylation, there was a pronounced acceleration of initiation of cell division in the absence of the uORF in CAK1. Our results suggest an unexpected layer of control, impinging on the activating phosphorylation of the Cdk. The uORF-mediated repression of Cak1 synthesis directly couples protein synthesis to the activity of the core cell cycle machinery. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 4 GRAPHICAL ABSTRACT .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 5

From late in the G1 phase of the cell cycle until anaphase, a rising Cdk activity drives cell cycle progression (Polymenis, 2022; Morgan, 2007). Active Cdks are serine/threonine protein kinases. The Cdks are inactive as monomers and require binding to a regulatory cyclin subunit to form an active Cdk-cyclin heterodimer. When a cyclin binds to the Cdk, the conformation of the Cdk's T-loop domain and a small nearby helix change, allowing protein substrates and ATP to access the active site of the kinase (Pavletich, 1999). In addition to cyclin binding, a conserved threonine residue in the T-loop must also be phosphorylated for Cdk activation. (Russo et al, 1996). T-loop phosphorylation is essential for Cdk function. Surprisingly, however, a mutation mimicking constant T-loop phosphorylation at the same site is also lethal unless the Cdk has more mutations elsewhere in the protein (Cross & Levine, 1998). Cells regulate the activity of Cdk complexes in many ways, as one would expect for the 'master' regulator of cell division. For example, interactions with other proteins activate or inhibit the kinase, while periodic phosphorylations at its N-terminus inhibit it (Polymenis, 2022; Morgan, 2007). Yet, cells do not periodically phosphorylate and dephosphorylate the T-loop to adjust their Cdk activity in the cell cycle (Kaldis, 1999; Cross & Levine, 1998). The seemingly constitutive nature of the activating T-loop phosphorylation has raised long-standing questions about its physiological role in cell proliferation (Blank et al, 2025). Cdk-activating kinases (CAKs) are the enzymes that phosphorylate Cdks at the T-loop (Kaldis, 1999; Blank et al, 2025). CAK themselves are part of the large Cdk family. In metazoans, Cdks that do not have direct cell cycle roles but instead transcriptional ones serve as CAKs (Fisher & Morgan, 1994; Fisher, 2005; Fisher et al, 1995; Shiekhattar et al, 1995). In budding yeast, Cak1 is the only CAK (Kaldis et al, 1996). Although Cak1 targets other proteins in meiosis (Whinston et al, 2013; Wagner et al, 1997), its single essential function is .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 6 phosphorylating the T169 residue of Cdc28, the main Cdk required for mitotic cell divisions in this organism. There is no evidence that the abundance, activity, or localization of Cak1 is regulated in proliferating cells, nor do post-translational modifications alter it (Kaldis et al, 1998). Likewise, the levels and activity of metazoan CAKs do not change during somatic (Tassan et al, 1994; Bartkova et al, 1996) or early embryonic cell divisions (Poon et al, 1994). The above observations add to the puzzle of why an essential modification and the 'writer' enzymes of that modification on the Cdk appear unregulated. Contrary to expectations of constitutive control of Cak1 and T-169 phosphorylation, here we show that Cak1 abundance and the activating T-loop phosphorylation of Cdc28 are regulated as cells transition from nutrient-rich to nutrient-poor environments or enter quiescence. A small upstream open reading frame (uORF) represses CAK1 mRNA translation. Cells lacking the uORF lose viability when starved and entering quiescence and are less fit in slowly proliferating steady-state chemostat cultures. Furthermore, the effects of the uORF-mediated control of Cak1 synthesis become very prominent in the absence of the counteracting phosphatases that remove the Cdk T169 phosphorylation. Our results suggest a previously unappreciated layer of Cdk regulation in coupling cell growth with cell division in poor conditions. Lowering T-loop phosphorylation may safeguard against mitotic cycling in unfavorable contexts. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 7

Cak1 levels and the activating T169-Cdc28 phosphorylation vary in different conditions We decided to test if Cak1 abundance and the activating phosphorylation of Cdc28 at T169 vary in different media because past studies overwhelmingly focused on cells continuously dividing in standard rich laboratory media. In addition, the only instance in the literature for any control over Cak1 levels in S. cerevisiae vegetative cells was when they were in the stationary phase, at which point the protein's abundance dropped significantly (Kaldis et al, 1998). We isolated Cdc28-TAP by immunoprecipitation from cells cultured in media with glucose, a preferred carbon source, or glycerol (Figure 1A). The isolated protein was then examined by mass spectrometry, to see if there were any changes in the fraction of the Cdc28 protein phosphorylated at T169. Indeed, in the glycerol-grown cells the fraction of T169-phosphorylated Cdc28 was reduced by about half (Figure 1B). To detect T169 phosphorylation in follow-up experiments, we then used a custom anti-pT169 antibody (see Materials and Methods) to provide independent validation of the mass spectrometry data and because it allows for faster, simpler, cheaper, and more sensitive detection. The custom antibody detected a protein of the expected size for Cdc28 (Figure 1C; compare the middle and right blots). The signal was also sensitive to phosphatase treatment (Figure 1C; middle blot, compare the left and right lanes) to the same extent as the signal from a previously validated phospho-specific antibody against another protein, Rps6 (Figure 1C; left blot, compare the left and right lanes). To monitor Cak1 levels, we used strains carrying an allele encoding Cak1 with a V5 epitope tag at its N-terminus (see Materials and Methods). We could then detect from the same samples the abundance of Cak1 (with an anti-V5 antibody), Cdc28 and Pgk1 (with commercial antibodies), and phospho-T169-Cdc28 (with our custom antibody). For each experiment, the .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 8 signal for each quantified target was internally normalized by dividing its intensity by the average intensity for that target across all conditions on the same gel, and then Log2 transformed (Figure 1D legend). Our results suggest that the levels of Cak1 were the highest in exponentially proliferating cells in glucose-containing media, lower in glycerol-containing ones, and, as expected based on (Kaldis et al, 1998), even lower in stationary phase cells (Figure 1D, left panel). In agreement with our mass spectrometry data (Figure 1B), T169 phosphorylation was lower in cells proliferating with glycerol as the carbon source, compared to cells proliferating with glucose, but undetectable in stationary phase cells (Figure 1D, second panel from the left). In contrast, total Cdc28 (and Pgk1; a glycolytic enzyme used as reference) levels did not change in all these conditions. We conclude that the levels of the activating Cdk phosphorylation (but not the levels of the Cdk itself) and the abundance of Cak1, the writer activating kinase, respond to growth conditions. CAK1 expression is regulated translationally, in part through a uORF To understand the regulation of Cak1 levels, we first quantified CAK1 mRNA abundance in the same growth conditions described above (Figure 1D). We used digital droplet PCR (ddPCR) because it is a quantitative, highly sensitive method well-suited for detecting differences in mRNA abundance (Quan et al, 2018). We found that steady-state levels of CAK1 mRNA were not significantly different among exponentially proliferating cells in glucose or glycerol-containing media, or in stationary phase cells (Figure EV1A, left). Hence, transcriptional control mechanisms are unlikely to explain the observed differences in Cak1 protein abundance. Since many cell cycle proteins are subject to regulated degradation, we next measured the apparent half-life of Cak1 in glucose- and glycerol-containing media. To study Cak1 protein stability, we constructed a high-copy (2μ) plasmid enabling copper-inducible (CUP1 promoter) expression of V5-CAK1 (Figure EV1B, top). As expected, V5-Cak1 synthesis was rapidly .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 9 induced upon the addition of CuSO4 (0.5mM) to cells harboring this plasmid (Figure EV1B, bottom). In separate experiments, we examined otherwise wild-type cells (strain BY4742) carrying this plasmid, growing with either glucose or glycerol as the carbon source. After a 20m induction to drive V5-Cak1 synthesis, we added cycloheximide (CHX) to block new protein synthesis. After normalizing the Cak1 signal (Figure EV1C, top) against the total protein signal at each time point (from the Ponceau-stained blots; Figure EV1C, bottom), there was little degradation of Cak1 in either the glucose- or glycerol-containing media, with an estimated half-life of Cak1 >4h in both conditions. These results argue against Cak1 destabilization as the primary mechanism explaining the observed differences in Cak1 protein abundance in glucose versus the poorer glycerol media. We note that the general notion that Cak1 is a very stable protein agrees with prior studies (Christiano et al, 2014; Kaldis et al, 1998). We then focused on potential mechanisms of translational control to explain the differing abundance of Cak1 under various growth conditions. We noticed that CAK1 has a uORF, encoding a 16-residue peptide, which initiates 37 nucleotides upstream of the CAK1 start codon and terminates downstream (marked as uORF-16 in Figure 2A, top). The uORF could explain the differences in the levels of Cak1 we described (Figure 1D). Ribosomes initiating at the uORF would bypass the start codon of CAK1, repressing Cak1 protein synthesis. The uORF-mediated repression of Cak1 synthesis would be disproportionately greater in poor growth conditions, when the cell’s ribosome content is lower, according to the classic kinetic model of translation (Lodish, 1974). To test the uORF’s role, we generated strains carrying a mutation at its AUG start codon to AAG (marked as uORF-0 in Figure 2A, second from bottom). Introducing the V5 epitope to monitor Cak1 abundance also extended the peptide produced by the uORF to 35 residues (marked as uORF-35 in Figure 2A, second from top). Lastly, to test whether it is the presence of a uORF in general, or specifically the uORF-encoded peptide that impacts Cak1 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 10 synthesis, we introduced a stop codon that terminates the uORF synthesis after only three residues (marked as uORF-3 in Figure 2A, bottom). In the uORF mutants, we monitored the Cak1, T169-phosphorylation, Cdc28, and Pgk1 levels by immunoblotting (Figure 2B). When uORF-0 cells proliferated with glycerol as the carbon source or were in stationary phase, the levels of Cak1 were derepressed, albeit not completely, compared to uORF-35 cells (compare the middle left panel to the bottom left panel). In contrast, in uORF-3 cells, Cak1 levels remain repressed as in uORF-35 cells (compare the top left panel to the bottom left panel). We conclude that the presence of the uORF represses Cak1 synthesis in a growth-dependent manner. The uORF-encoded peptide is unlikely to be responsible for this regulation. Instead, the repression likely arises from ribosomal re-initiation failure, whereby ribosomes that successfully translate the uORF-3 are less likely to re-initiate translation at the downstream CAK1 start codon. We note that the uORF-mediated repression was not the result of changes in the steady-state levels of the CAK1 transcript (Figure EV1A, right). Surprisingly, although Cak1 levels were derepressed in stationary phase uORF-0 cells, the levels of T169 phosphorylation were not, and remained undetectable (Figure 2B, second from the left column of panels -compare the middle to the bottom panel). The levels of Cdc28 and Pgk1 remained largely unaffected in all conditions and mutants (Figure 2B, 3rd and 4th column of panels from the left). These results suggest that the uORF regulates Cak1 levels in different growth conditions, but the levels of the activating phosphorylation on the Cdk are robust to those changes. To provide independent evidence of uORF’s expression and its impact on Cak1 levels, we generated plasmids encoding fluorescent proteins in their place (Figure 3A). In cells carrying these plasmids, mTurquoise2-associated fluorescence reported on uORF expression, while mRuby2-associated fluorescence reported on Cak1 (Figure 3B). These plasmids were maintained at a high copy number in cells, but the sequences driving expression of the .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 11 fluorescent reporters were from the CAK1 locus (see Materials and Methods). Our results show that mTurquoise2-associated fluorescence was readily detectable in all conditions we tested (Figure 3B,C), arguing that the start codon of the uORF is recognized by scanning ribosomes. Furthermore, the significant increase (~10-fold) in mRuby2-associated fluorescence when the uORF start codon was mutated (Figure 3C, compare the last two panels to the right) strongly supports the notion that the uORF represses Cak1 synthesis. We note that the magnitude of the repression appears greater with these reporters compared to our results from immunoblotting (Figure 2B, 2 to 3-fold), probably because these were high-copy plasmids. The experiments we described thus far were in cells proliferating with glucose or glycerol as the carbon source in batch cultures, where the doubling time varies ~2-fold (90m vs. 160m), and one cannot distinguish nutrient-specific effects from growth rate-dependent ones. In contrast, in chemostats, the growth rate is externally set by the rate of medium addition (D, dilution rate). It can be adjusted over a much greater range, independently of the limiting nutrient in the medium. For these reasons, we examined prototrophic uORF-35 and uORF-0 cells in chemostats under carbon or nitrogen limitation (Figure 4). Over a broad range of dilution rates, corresponding to a doubling time of 3.5h to 13.9h, Cak1 levels decreased linearly as the growth rate was lowered in uORF-35 cells, in both carbon- and nitrogen-limited cultures (Figure 4A, bottom; and Figure 4B, bottom two panels in the left column). Hence, Cak1 levels respond to changes in growth rate, but not in a nutrient-specific manner. This regulation was not evident in uORF-0 cells, where Cak1 levels were not elevated in higher dilution rates (Figure 4A, top; and Figure 4B, top two panels in the left column). There were also no growth rate or nutrient-specific changes in the abundance of Cdc28 and Pgk1 in these strains and conditions (Figure 4B, second and third columns of panels). These results argue that the repressive function of the uORF in the CAK1 transcript mediates the growth-dependent changes in Cak1 abundance. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 12 In contrast to the steady decline of Cak1 levels as the dilution rate was lowered, the levels of T169 phosphorylation displayed a different pattern: relatively constant until the cultures had a doubling time of 6-7 hours (D = 0.10 h-1), but dropping precipitously in cells proliferating at slower rates (Figure 4C). The growth-dependent reduction in T169 phosphorylation was slightly more acute in nitrogen-limited cells, occurring at doubling times of >5h (e.g., compare the two bottom panels in Figure 4C). In uORF-0 cells, the drop in T169 phosphorylation was still evident in low dilution rates, albeit it appears to have been pushed to even lower ones (e.g., in Figure 4C, compare the second from the top to the bottom panels). Phenotypes of cells lacking the CAK1 uORF The data we described above suggest that a sizable change in T169 phosphorylation in uORF-0 cells manifests only in very slowly proliferating cells in chemostats (Figure 4C), despite an increase in Cak1 levels. Consequently, we wondered if the 2-3 fold derepression of Cak1 synthesis in uORF-0 cells (Figure 2B) would lead to any phenotypes associated with cell proliferation. We first examined cells proliferating in batch cultures with different carbon or nitrogen sources (Figure EV2), comparing wild-type cells expressing the 16 residue-long uORF (uORF-16) with those that did not (uORF-0). There was no significant change in the doubling time (Td) between the two strains, or in budding and cell size (Figure EV2A), parameters often associated with altered cell cycle kinetics (Polymenis, 2022). Next, we used centrifugal elutriation to obtain highly synchronous early G1 cells from the wild-type (uORF-16) and uORF-0 strains growing with glycerol as the carbon source, and monitored them as they progressed through the cell cycle (Figure EV2B). In late G1, if budding yeast cells have grown enough in size, they commit to a new round of cell division at a point called START. Passage through START is followed by initiation of DNA replication and appearance of a bud on the cell surface (Hartwell & Unger, 1977; Johnston et al, 1977), which will eventually become the new daughter cell after cytokinesis. A parameter used to gauge START completion is the size at .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 13 which half of the cells are budded, known as the ‘critical size’ (Polymenis, 2022). We found that uORF-16 and uORF-0 cells had the same critical size (Figure EV2B), arguing that the kinetics of the G1/S transition are unaffected in cells lacking the uORF in CAK1. Hence, although in media with glycerol as the carbon source the uORF-0 cells have higher levels of Cak1 and T169 phosphorylation (Figure 2B), cell cycle progression was not measurably affected (Figure EV2). To gain insight into phenotypes associated with Cak1 overexpression in general, we used the copper-inducible plasmid driving ectopic CAK1 expression that we described earlier (Figure EV1B). Again, in proliferating cultures with either glucose or glycerol as the carbon source, there was no difference in the doubling time or cell size when CAK1 was over-expressed (not shown). Taken together, our results suggest that under standard laboratory conditions, Cak1 is not limiting and only a fraction of Cdc28 needs to be active for the timely completion of cell cycle transitions. However, in stationary phase cells, where Cak1 levels are typically very low (Figure 1D), we noticed that cells induced to overexpress CAK1 were smaller (Figure EV3A), albeit properly arrested in G1, based on their DNA content (Figure EV3B). Moreover, CAK1 overexpression resulted in a 3- to 4-fold decrease in viability in stationary-phase cells (Figure EV3C). We note that while some commonly used inducible systems (e.g., GAL-based) are sensitive to nutrients and the growth status of the cell, induction of CUP1-based expression is possible even in cells in the stationary phase (Hottiger et al, 1994). Following the above results, arguing that CAK1 overexpression may decrease the viability of quiescent cells, we next examined the viability of quiescent cells lacking the CAK1 uORF. Based on their ability to divide and form colonies on solid media, we found that uORF-0 cells lose viability at an accelerated rate compared to their uORF-16 counterparts (Figure 5A). Next, we measured the competitive fitness of uORF-0 cells over time in mixed populations (uORF-0 and uORF-16 started at equal proportions) in chemostat cultures, as there was a .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 14 pronounced drop in T169 phosphorylation levels at low dilution rates (see Figure 4C), concomitant with lower Cak1 levels (Figure 4A,B). Since uORF-0 cells carry the KanMX resistance marker, their proportion was scored based on their ability to grow on G418-containing solid media. Note that the fitness of cells carrying this marker is unaffected in similar chemostat settings (Baganz et al, 1997). Our results suggest that while at a higher dilution rate (D=0.13 h-1; Figure 5B, middle panel) the fitness of uORF-0 cells was unaffected, it progressively declined at a lower dilution rate (D=0.05 h-1), in cultures limited for carbon (Figure 5B, left panel) or nitrogen (Figure 5B, right panel). We conclude that while cells lacking the CAK1 uORF are indistinguishable from their wild-type counterparts in most proliferative settings under standard laboratory conditions, there are penalties in viability and fitness during quiescence and in cultures proliferating very slowly, respectively. Protein phosphatases counteract changes in cell cycle kinetics in cells lacking the CAK1 uORF Our data thus far suggest that eliminating the CAK1 uORF has limited phenotypic consequences in standard laboratory settings, likely because T169 phosphorylation levels are robust against a 2 to 3-fold change in Cak1 abundance. Since T169 phosphorylation is regulated by both 'writer' (Cak1) and 'eraser' (phosphatase) enzymes, we next examined the role of protein phosphatases. Type 2C phosphatases dephosphorylate Cdc28 at T169 (Cheng et al, 1999). The seven type 2C phosphatases in budding yeast (Ptc1-7) have overlapping roles with each other and with other phosphatase classes. While cells lacking all seven PTC genes are viable in standard conditions, specific mutant combinations are sensitive to various stresses (Sharmin et al, 2014). To test for functional interactions, we first constructed uORF-0, ptc2,3Δ triple mutants, because Ptc2 and Ptc3 were reported to dephosphorylate Cdc28 (Cheng et al, 1999). The overall proliferation of triple uORF-0, ptc2,3Δ cells was indistinguishable from that of .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 15 the uORF-16, ptc2,3Δ control on solid media with glucose or glycerol as the carbon source (Figure EV4A) or in liquid glucose batch cultures (Figure EV4B). To generate a more sensitized

that enables the detection of functional interactions arising from the higher abundance of Cak1 in uORF-0 cells, we then constructed uORF-0, ptc1,2,3Δ quadruple mutants. We chose the additional Ptc1 deletion because Ptc1 is involved in more processes than any other type 2C phosphatase in yeast (Ariño et al, 2019). Indeed, ptc1,2,3Δ mutants (with or without the CAK1 uORF) proliferated more slowly than wild type or ptc2,3Δ cells in glucose or glycerol-containing media (Figure EV4). We measured the DNA content of these cells proliferating exponentially in liquid batch cultures (Figure 6A). We noticed that a noticeable fraction of triple ptc1-3Δ cells (with or without the uORF) had a higher than 2N DNA content in glucose and glycerol, consistent with mitotic defects (Figure 6A, compare the right two panels to the others). There was a significant decrease (from 39% to 23%, p<0.0001 based on robust bootstrap ANOVA) in the fraction of cells with G1 DNA content in uORF-0, ptc1,2,3Δ cells compared to uORF-16, ptc1,2,3Δ cells (Figure 6A, compare the two rightmost panels at the bottom). A similar effect was also evident in glycerol cultures (Figure 6A, compare the two rightmost panels at the top). These results suggest that loss of the CAK1 uORF when ‘eraser’ phosphatases were also absent altered cell cycle kinetics significantly. We then examined synchronous cultures obtained by elutriation (Figure 6B). We found that ptc1,2,3Δ cells are difficult to elutriate because they proliferate slowly and the cultures contain many dead cells. Even in glucose media, among the elutriated G1 fraction of cells, only about two-thirds eventually budded, compared to >90% for the wild type (BY4742 strain background) cells (Figure 6B). Although uORF-16, ptc1,2,3Δ cells had the same critical size (the size at which half the cells were budded) as wild type cells (~37fL), there were nonetheless more budded smaller cells (Figure 6B; compare the left and middle panels), arguing that loss of .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 16 type 2C PPases may promote initiation of cell division. The effect was greatly enhanced in uORF-0, ptc1,2,3Δ cells, which had a much lower critical size (~30fL), consistent with a pronounced acceleration of START (Figure 6B, compare the right panel to the other two). Given that their overall proliferation rate was similar to that of uORF-16, ptc1,2,3Δ cells (see Figure EV4), we conclude that uORF-0, ptc1,2,3Δ cells have accelerated entry into the S phase but also delayed exit from mitosis, as expected for cells with higher Cdk activity. These results support a major role for phosphatases in counteracting the effect of higher Cak1 levels in uORF-0 cells. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 17

The results presented are significant for several reasons. Demonstrating growth-dependent control of Cak1 and T169 phosphorylation provides an answer to a long-standing question about the regulation of CAK activity. We describe the molecular mechanism underpinning this control as a cis element (the uORF) in the CAK1 transcript, making Cak1 synthesis disproportionately sensitive to the most central aspect of cell growth: the availability of ribosomes. We had reported a similar mechanism in the control of the early G1 cyclin Cln3 (Polymenis & Schmidt, 1997; Blank et al, 2018). Hence, when yeast cells encounter different nutrient environments, they couple the levels of two direct activators of the Cdk, Cak1 and Cln3, to the capacity of their protein synthesis machinery (see Figure 7 for a schematic). Cells were impervious to changes in Cak1 levels in most laboratory contexts. This was the result of counteracting type 2C protein phosphatases, which are apparently quite efficient in buffering T169 phosphorylation against changes in Cak1 levels mediated by the uORF mechanism we described (Figure 6). Nonetheless, our data also suggest that translational control of CAK1 may promote survival in the wild (Figures 4,5). In natural environments, where limited and changing resources are the norm, the uORF in CAK1 would promote fitness by efficiently inhibiting Cak1 synthesis and cell division when nutrient availability drops significantly. Finally, our results contribute to a unifying view of CAK across all eukaryotes. Observations in animal cells, such as the higher threshold of T-loop Cdk phosphorylation required for the G1/S transition in response to extracellular signals and the heightened involvement of counteracting phosphatases at that proliferative transition (Schachter et al, 2013; Blank et al, 2025), point to a general and conserved role for CAK in safeguarding cells before they commit to a new round of division. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 18

Strains and media All S. cerevisiae strains used in this study are listed in the Key Resources Table. Culture Conditions Unless indicated otherwise, the cells were cultured on the standard, rich, YPD medium (1% w/v yeast extract, 2% w/v peptone, 2% w/v dextrose), at 30°C. Throughout the manuscript, glucose refers to dextrose (D-glucose). In some experiments, as indicated, glycerol (2% w/v) replaced dextrose as the carbon source. Synthetic minimal media (SMM) contained 0.17% w/v yeast nitrogen base without amino acids and ammonium sulfate, 0.5% w/v nitrogen source (ammonium sulfate or proline, as indicated), 2% w/v dextrose, and were supplemented as needed for auxotrophies (Kaiser et al, 1994). For experiments involving the inducible expression of CAK1, the media were standard synthetic complete, lacking uracil (SC-Ura) (Kaiser et al, 1994), with 2% w/v dextrose or glycerol, as indicated, and the cells cultured at 30°C. For the experiments using the reporter plasmids, the media were also SC, but lacking both uracil and leucine. For plasmid construction and propagation in E. coli bacteria, cells were cultured on the standard lysogeny broth media (0.5% w/v yeast extract, 1% w/v tryptone, 1% w/v sodium chloride), at 37°C. Chemostat experiments Chemostat experiments were conducted using a BioFlo 110 bioreactor system (New Brunswick Scientific, Edison, NJ) with a working volume of 880 mL. To monitor the abundance of various proteins and T169 phosphorylation we used prototrophic strains with agitation set at 800 rpm for these experiments, at room temperature (Blank et al, 2018; Henry et al, 2010; Guo et al, 2004). The carbon-limited media were 0.17% w/v yeast nitrogen base without amino acids and ammonium sulfate, 0.5% w/v ammonium sulfate, and 0.08% w/v dextrose. The nitrogen-limited .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 19 media were 0.17% w/v yeast nitrogen base without amino acids and ammonium sulfate, 0.002% w/v ammonium sulfate, and 2% w/v dextrose. The dilution rate was adjusted as indicated for each experiment. The cell density remained >1E+07 cells/mL and did not vary more than threefold between the lowest and highest dilution rates. For experiments with auxotrophic strains (Figure 5), the media were supplemented as needed. Strain construction To monitor Cak1 protein abundance, we initially constructed strains carrying an allele encoding a TAP-tag fused to CAK1 at the C-terminus. However, these strains lost viability after a few generations and the cells had elongated, sausage-like morphology, indicative of mitotic defects (not shown). Consequently, we then constructed strains carrying N-terminal V5-tagged CAK1 alleles, expressing CAK1 from the endogenous chromosomal locus. We first inserted a KanMX6 cassette upstream, at position ChrVI:79250, with PCR-mediated methodology (Longtine et al, 1998), using primers VI:79250-79325-F1 and VI:79159-79249-R1 (see Key Resources Table), and plasmid pFA6a-KanMX6 as a template. We introduced the PCR product into strain BY4742 to generate strain MSP256. The KanMX6 insertion was verified by PCR, using primers VI:79600-FWD and VI:78815-REV. Then, we used genomic DNA from strain MSP256 as a template in a PCR reaction with primers V5-CAK1-REV and VI:79600-FWD. The PCR product was used to transform the prototrophic strain X2180-5A. The resulting strain (MSP258; denoted as uORF-35, V5-CAK1 throughout the manuscript) expressed a protein of the expected size for V5-Cak1, based on immunoblots, and the genomic area was also sequenced to verify the absence of any introduced mutations. Note that the uORF normally encodes a 16-residue peptide (Figure 2, top; denoted as uORF-16). By inserting the V5 sequences at the CAK1 locus, the uORF was extended to 35 residues (Figure 2, second from top; denoted as uORF-35). The doubling time and the percentage of budded cells between .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 20 strains X2180-5A (uORF-16, CAK1) and MSP258 (ChrVI:79250::KanMX6, uORF-35, V5-CAK1, X2180-5A otherwise) were indistinguishable, in glucose and glycerol media. To mutate the start codon of the uORF (AUG → AAG substitution) upstream of the CAK1 ORF, we used primers VI:79250-79325-F1 and VI:79159-uORFm-R1 in a PCR reaction with plasmid pFA6a-kanMX6 as a template. The PCR product was then used to transform strain BY4742, generating strain MSP257 (ChrVI:79250::KanMX6, uORF-0, CAK1, BY4742 otherwise). The KanMX6 insertion was verified by PCR, and the introduction of the mutation by sequencing. As above, we then used genomic DNA from strain MSP257 as a template in a PCR reaction with primers V5-CAK1-REV and VI:79600-FWD. The PCR product was used to transform the prototrophic strain X2180-5A, generating strain MSP259 (Figure 2, second from bottom; denoted as uORF-0, V5-CAK1), which was verified as described above for strain MSP258 (uORF-35, V5-CAK1). We also generated a strain carrying a truncated uORF, encoding only its first 3 amino acids. We used primers VI:79600-FWD and uORFm1-3-REV in a PCR reaction with genomic DNA from strain MSP256 as a template. The PCR product was used to transform strain X2180-5A, to generate strain MSP267. Introduction of the desired mutation was verified by sequencing as described above. To epitope-tag the CAK1 allele in this strain background, we then used genomic DNA of strain MSP267 as a template in a PCR reaction with primers V5-CAK1-REV and VI:79600-FWD. The PCR product was used to transform the prototrophic strain X2180-5A, generating strain MSP268 (uORF-3, V5-CAK1), verified by sequencing and immunoblotting to ensure that it expresses V5-Cak1 as expected. To generate a strain lacking PTC2 (encoding a type 2C protein phosphatase), we amplified the URA3 ORF using primers PTC2-URA3-FWD and PTC2-URA3-REV and genomic DNA from strain X2180-5A. We then used the resulting PCR product to transform strain BY4742, yielding a ptc2Δ::URA3 derivative (strain MSP269). We verified the correct replacement of PTC2 with URA3 using primers V:337690-FWD and V:335594-REV and the .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 21 genomic DNA of each putative transformant as a template. A haploid strain (HB499) lacking PTC3 was obtained by sporulation of the homozygous diploid ptc3Δ::KanMX/ptc3Δ::KanMX strain 33082. The double ptc2Δ::URA3 ptc3Δ::KanMX haploid strain (HB509) was a segregant of the cross between HB499 and MSP256. The ChrVI:79250::KanMX6, uORF-0, ptc2,3Δ haploid strain (HB505) was a segregant of the cross between HB509 and MSP257. Lastly, we crossed a ptc1Δ::HIS3 (BY4741 otherwise) strain described elsewhere (González-Rubio et al, 2023) with strain HB509, or HB505. Segregants of these crosses were selected for being triple ptc1-3Δ mutants without (HB531) or with the uORF-0 mutation (HB530), respectively, by co-segregation of the KanMXr, URA+, and HIS+ markers. Using genomic DNA from these strains in PCR reactions with primers VI:78815-REV and VI:79398-FWD, we validated the presence of the uORF-0 mutation (strain HB531) or not (strain HB530) by sequencing the PCR product with the VI:78815-REV primer. Since the PTC3 deletion was also marked with the KanMX marker, we confirmed that these strains were indeed ptc3Δ::KanMX through PCR, with primers II:113444-FWD and KanMX-REV, and genomic DNA as template. We also confirmed that they had no wild type PTC3 allele, in a PCR reaction with primers II:113444-FWD and II:114791-REV, and genomic DNA as template. Plasmids All enzymes used in plasmid construction were from New England Biolabs. To generate the plasmid driving copper-dependent expression of V5-CAK1, we used the single-pot approach and the plasmids included in the MoClo kit (Lee et al, 2015), as detailed below. A synthetic DNA fragment was made (Integrated DNA Technologies, Coralville, IA) containing the V5-CAK1 allele flanked with adaptor sequences compatible with the MoClo multipart assembly. The V5-CAK1 DNA fragment was digested with BsaI, electrophoresed through an agarose gel (1% w/v), and purified. It was then combined with plasmids YTK002, YTK067, YTK031, .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 22 YTK063, YTK074, YTK082, and YTK083 in a ligase reaction containing T4 ligase together with the BsaI restriction endonuclease. The reaction was carried out in a thermocycler at 37°C for 20m; (37°C for 5m, 16°C for 5m) for 50 cycles; 60°C for 5m; 80°C for 10m. An aliquot was used to transform E. coli (high-efficiency DH5alpha cells; New England Biolabs). Transformants were selected and verified to contain the high-copy (2 micron; 2μ) YEp-pCUP1-V5-CAK1 plasmid, where the copper-inducible pCUP1 promoter drives V5-CAK1 expression, the terminator is tADH1, the yeast selection marker is URA3, and AmpR-ColE1 are used for bacterial selection and propagation. The plasmid was sequenced with primers pBR322ori-FWD and AmpR-FWD (see Key Resources Table). Expression of a protein of the expected size upon induction with CuSO4 (0.5mM) was confirmed by immunoblotting for V5-Cak1 (see Figure EV1B). We also followed the same single-pot approach with reagents from the MoClo kit to generate several reporter constructs. Synthetic DNA inserts were made (Integrated DNA Technologies, Coralville, IA) containing CAK1 upstream sequences (starting at position chrVI:79401) fused to sequences encoding fluorescent reporter proteins and flanked with adaptor sequences compatible with the MoClo multipart assembly. These inserts were: ● uORF-mTurquoise2: chrVI:79401-79196 fused to the mTurquoise2 ORF. ● CAK1-mRuby2: chrVI:79401-79159 fused to the mRuby2 ORF. ● uORFm-CAK1-mRuby2 (chrVI:79401-79159 fused to the mRuby2 ORF, but with the uORF start codon mutated to AAG). The plasmids containing these synthetic insert DNA fragments were digested with BsaI, electrophoresed through an agarose gel (1% w/v), and the inserts isolated. The inserts were then combined with plasmids YTK002, YTK067, YTK031, YTK063, YTK074 (for the mRuby2 reporters) or YTK075 (for the mTurquoise2 reporter), YTK082, and YTK083 in a ligase reaction containing T4 ligase together with the BsaI restriction endonuclease. The reactions were .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 23 performed as above and used to transform E. coli and isolate the relevant plasmids. The plasmids were verified by sequencing as above, and with additional oligonucleotides that correspond to their insert sequences (VI:79398-FWD, mTurquoise2-FWD, and mRuby2-FWD; see Key Resources Table). The plasmids contained the tADH1 terminator, the 2 micron (2μ) element for high-copy propagation in yeast cells, and the AmpR-ColE1 elements for bacterial selection and propagation. The yeast selection marker is URA3 for the mRuby2 reporters, and LEU2 for the mTurquoise2 one. Cdc28-TAP immunoprecipitation and mass spectrometry for T169-Cdc28 phosphorylation To detect the T169 phosphorylation by mass spectrometry (Figure 1), we used a strain (BY4741 background) carrying a CDC28-TAP allele expressed from its endogenous chromosomal location (Ghaemmaghami et al, 2003) (see Key Resources Table). Overnight cultures growing in rich undefined media with glucose (YPDextrose) or glycerol (YPGlycerol) as the carbon source were diluted 1:100 into fresh 1L cultures and incubated at 30°C for several hours until they reached a cell density of ~2-5E+07 cells/mL in YPD, or ~8E+06 to 1E+07 cells/mL in YPGlycerol. The cultures were quenched with sodium azide (at 0.1% final concentration) and cycloheximide (100μg/mL). The cells were collected by centrifugation, and the pellets were stored at -80°C until needed. For immunoprecipitation, the cell pellets were washed in ice-cold lysis/IP buffer (10mM Tris-HCl, pH=8.0, 150mM NaCl), resuspended at ~1E+09 cells per reaction in 0.6mL of lysis buffer, containing protease and phosphatase inhibitor cocktails, and kept on ice. Glass beads (~0.35mL) were then added. The cells were broken in 8 cycles of vortexing at the maximum speed for 30s, followed by placement on ice for 30s. The samples were centrifuged at 5,000 g .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 24 for 5m, and the supernatants were collected and placed in fresh tubes on ice. To each reaction were added 50μL of IgG-Sepharose-6 fast-flow beads (pre-washed with lysis buffer), and incubated for 2h on a rotisserie mixer at 4°C. The beads were collected by centrifugation at 500 g for 1m, washed twice with lysis buffer and twice with 10mM ammonium bicarbonate buffer, prepared fresh on the day of the experiment. The beads were resuspended in 75μL of ammonium bicarbonate buffer and shipped at 4°C to the UC Davis proteomics core facility. At the core facility, samples underwent tryptic digestion and LC-MS/MS, and were processed according to their established protocols for phosphopeptide quantification. The signal in each sample for Cdc28 or pT169-Cdc28 was first normalized against the total ‘input’ signal in each case. The ratio of the two normalized values (i.e., pT169-Cdc28 : Cdc28) was used to compare the different samples (Figure 1B). All the data from this mass spectrometry experiment are in Dataset EV1. Protein surveillance Immunoblotting Protein extracts were made as described (Wallace et al, 2022), and resolved on 12% Tris‐Glycine SDS–PAGE gels, unless indicated otherwise. V5-Cak1 was detected using a mouse anti-V5 tag monoclonal antibody, conjugated with horseradish peroxidase (HRP), used at a 1:5,000 dilution (ThermoFisher; Cat#: R961-25). Loading was measured with a mouse anti‐Pgk1 primary antibody, followed by a secondary antibody (at 1:5,000; ThermoFisher; Cat#: 31431). To detect total amounts of Cdc28, we used a commercially available mouse monoclonal anti-Cdc28 antibody (at 1:1,000; Santa Cruz Biotechnologies #sc-515762). For detecting the activating phosphorylation of Cdc28, we obtained a custom rabbit polyclonal antibody (Biomatic, Kitchener, Ontario, Canada) raised against a peptide (RAFGVPLRAY(pT)HEIVT) encompassing T169 in the yeast Cdc28 protein. The antibody .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 25 (α-pT169) recognized a protein of the same apparent mass as one recognized by the anti-Cdc28 antibody (Figure 1C, compare the middle and right panels on the top). The signal was diminished (Figure 1C, middle panel on the top) when the samples were first treated for 30m at 30°C with 400u of Lambda Protein Phosphatase (New England Biolabs #P0753). The reduction in signal intensity was comparable to the reduction observed for ribosomal protein S6 phosphorylation (Figure 1C, compare the left and middle panels on the top). The latter was detected by a specific rabbit monoclonal antibody against Ser235/236 of the human protein (at 1:5,000; Cell Signaling, Cat#: 4858), followed by a secondary antibody (at 1:5,000; ThermoFisher; Cat#: 31466). V5-Cak1 Stability Assay To monitor the stability of V5-Cak1 (Figure EV1C), we transformed BY4742 haploid cells with the pCUP1-V5-CAK1 plasmid described above. The transformants were cultured in synthetic complete drop-out medium (SC-Ura), with glucose or glycerol as the carbon source. Cultures proliferating exponentially were induced with CuSO4 for 20m, and then treated with cycloheximide (CHX; 100μg/mL) to block protein synthesis. The levels of V5-Cak1 were then monitored by immunoblotting as described above, from samples taken at various time points. Image analysis Images were processed with the “Subtract Background” tool, and band intensities quantified using the “Measure” tool to obtain a mean intensity for each band, using the ImageJ software package. The area measured was kept constant for a particular sample series for each blot analyzed. The source data for all the immunoblots are in the supporting files. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 26 Transcript abundance using digital droplet PCR (ddPCR) For RNA surveillance, RNA extracts were made as described previously (Blank et al, 2024, 2020). Briefly, the ddPCR reaction mixture was prepared by following the manufacturer’s protocol (One-Step RT-ddPCR Advanced Kit for Probes), using the TaqMan® hydrolysis probes labeled with FAM-MBP for CAK1 and VIC-MBP for UBC6 reporter fluorophores. The abundance of the transcripts was obtained using the QuantaSoft™ Software. Transcript levels of CAK1 were normalized against the corresponding transcript levels of UBC6 in each sample. Centrifugal elutriation and cell size measurements In each experiment, a 250mL cell culture was grown to a density typical of mid-log cultures (~1-5E+07 cells/mL in YPD), then loaded onto a 40mL elutriator chamber at a flow rate of 35mL/min and centrifuged at 3,200rpm. The cells were washed twice with 250mL of medium, first at 2,800rpm, then at 2,400rpm. Finally, cells were elutriated with 250mL of the same medium at a pump speed of 38mL/min and centrifuge speed of 2,400rpm. This early G1 daughter cell fraction was isolated and monitored regularly as it progressed through the cell cycle. From live, unfixed cells, the percentage of budded cells was determined via phase microscopy, and cell size was measured using a Beckman Z2 channelyzer (Hoose et al, 2012; Soma et al, 2014). The Beckman Accucomp software was used to generate size histograms, from which the geometric mean size was calculated. In addition, the ‘birth’ size of asynchronous cultures was determined from the size histograms using established methods (Maitra et al, 2019; Truong et al, 2013). Briefly, the population to the left of the peak of the cell size histogram was assumed to contain daughter cells, and the size corresponding to the 10% smallest of these cells was taken as ‘birth size’. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 27 Doubling time measurements in batch cultures For experiments shown in Figure EV2, cell numbers were determined by direct counting using a hemocytometer. Natural log-transformed cell numbers were plotted against time (h). The specific growth rate (k, in h-1) was derived from the slope of the linear regression during the exponential phase. Doubling time (Td) was calculated using the equation: Td=ln2/k For experiments in Figure EV4, optical density at 600 nm was measured using an Agilent Synergy H1 plate reader as a proxy for cell density. Cultures (0.2mL per well) were incubated at 30°C with orbital agitation for 10s prior to hourly readings. Td estimates were calculated as described above, using the linear portion of the growth curves between the 2h and 10h timepoints. Flow cytometry and fluorescence microscopy The percentage of cells in the G1 phase of the cell cycle (%G1) was quantified as detailed in previously published protocols (Hoose et al, 2012, 2013). For the experiments in Figure 6A, we used a Cytek® Muse® Cell Analyzer. Because of S. cerevisiae's small genome size and the close spacing of DNA peaks in the histograms, which often makes the application of software developed for the analysis of DNA content in mammalian cells problematic, we used the following approach: Briefly, the area under the left half of the G1 peak in the DNA content histogram was measured, doubled to account for the entire G1 peak, and then divided by the total area under the histogram. This fraction, representing the proportion of cells with G1 DNA content (1N for haploid cells), was multiplied by 100 to express it as a percentage (%G1). For the experiments in Figure EV3B, we used a BD Accuri C6 Flow Cytometer at the Flow Cytometry Facility of the Texas A&M University College of Veterinary Medicine & Biomedical .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 28 Sciences. The calculated %G1 fraction in this case was from FlowJoTM flow cytometry software (BD Biosciences, Franklin Lakes, NJ) For the experiments in Figure 3, BY4743 cells carrying the mTurquoise2 and mRuby2 reporter plasmids were grown to mid-log phase at 30°C in synthetic complete (SC) medium lacking leucine and uracil, supplemented with either glucose or glycerol as the carbon source. To prepare samples, exponentially proliferating cultures were briefly sonicated to separate cell clumps. Cells were then collected by centrifugation at 5,000g for 1m, and the pellets were resuspended in phosphate-buffered saline (PBS) containing 0.5% w/v bovine serum albumin (BSA). Samples were protected from light and processed immediately for analysis, using a BD Accuri C6 Flow Cytometer at the Flow Cytometry Facility of the Texas A&M University College of Veterinary Medicine & Biomedical Sciences. For fluorescence microscopy, the samples were prepared as above and promptly imaged using a Nikon TS100 inverted microscope equipped with a Plan Apo 100X oil DIC

and a CoolSNAP DYNO camera. The filters used were an AT-DAPI filter set (excitation at λex ≅ 360-390nm) for imaging the mTurquoise2 reporter, and a DM575-TRITC filter set (excitation at λex ≅ 540-580nm) for imaging the mRuby2 reporter. Sample-size and replicates The number of samples in each experiment was not determined based on a formal power analysis. All replicates represented distinct biological samples from separate cultures. In experiments with a minimum of three replicates, statistical comparisons between groups were performed using a robust bootstrap ANOVA (t1waybt function, with the number of bootstrap samples set to 599) followed by post hoc tests (mcppb20 function) from the WRS2 R package (Mair & Wilcox, 2020). For experiments with at least four replicates, nonparametric statistical .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 29

were also employed, as indicated in each case. All collected data points, including any potential outliers, were included in the analyses. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 30 DATA AVAILABILITY Strains and plasmids are available upon request. The authors affirm that all data necessary for confirming the conclusions of the article are present within the article, figures, and tables.

This work was supported by the National Institutes of Health (NIH, grants R01 GM123139 and R35 GM161174 to M.P.). We thank Staci E. Hammer and Britt Faulk for technical assistance in some experiments. We also thank Dr. Humberto Martin (Universidad Complutense de Madrid) for his generous gift of the ptc1Δ::HIS3 strain. DISCLOSURE AND COMPETING INTERESTS STATEMENT The authors have no conflicts of interest to declare that could be perceived to influence the presentation or interpretation of the data. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 31

Ariño J, Velázquez D & Casamayor A (2019) Ser/Thr protein phosphatases in fungi: structure, regulation and function. Microb Cell Graz Austria 6: 217–256 Baganz F, Hayes A, Marren D, Gardner DC & Oliver SG (1997) Suitability of replacement markers for functional analysis studies in Saccharomyces cerevisiae. Yeast Chichester Engl 13: 1563–1573 Bartkova J, Zemanova M & Bartek J (1996) Expression of CDK7/CAK in normal and tumor cells of diverse histogenesis, cell-cycle position and differentiation. Int J Cancer 66: 732–737 Blank HM, Alonso A, Fabritius AS, Valk E, Loog M, Winey M & Polymenis M (2024) Translational control of MPS1 links protein synthesis with the initiation of cell division and spindle pole body duplication in Saccharomyces cerevisiae. Genetics 227: iyae069 Blank HM, Callahan M, Pistikopoulos IPE, Polymenis AO & Polymenis M (2018) Scaling of G1 Duration with Population Doubling Time by a Cyclin in Saccharomyces cerevisiae. Genetics 210: 895–906 Blank HM, No E-G & Polymenis M (2025) Cdk activation by phosphorylation: linking growth signals to cell cycle control. Biochem Soc Trans 53: BST20253004 Blank HM, Papoulas O, Maitra N, Garge R, Kennedy BK, Schilling B, Marcotte EM & Polymenis M (2020) Abundances of transcripts, proteins, and metabolites in the cell cycle of budding yeast reveal coordinate control of lipid metabolism. Mol Biol Cell 31: 1069–1084 Cheng A, Ross KE, Kaldis P & Solomon MJ (1999) Dephosphorylation of cyclin-dependent kinases by type 2C protein phosphatases. Genes Dev 13: 2946–2957 Christiano R, Nagaraj N, Fröhlich F & Walther TC (2014) Global proteome turnover analyses of the Yeasts S. cerevisiae and S. pombe. Cell Rep 9: 1959–1965 Cross FR & Levine K (1998) Molecular evolution allows bypass of the requirement for activation loop phosphorylation of the Cdc28 cyclin-dependent kinase. Mol Cell Biol 18: 2923–2931 Fisher RP (2005) Secrets of a double agent: CDK7 in cell-cycle control and transcription. J Cell Sci 118: 5171–5180 Fisher RP, Jin P, Chamberlin HM & Morgan DO (1995) Alternative mechanisms of CAK assembly require an assembly factor or an activating kinase. Cell 83: 47–57 Fisher RP & Morgan DO (1994) A novel cyclin associates with MO15/CDK7 to form the CDK-activating kinase. Cell 78: 713–724 Ghaemmaghami S, Huh WK, Bower K, Howson RW, Belle A, Dephoure N, O’Shea EK & Weissman JS (2003) Global analysis of protein expression in yeast. Nature 425: 737–41 González-Rubio G, Martín H & Molina M (2023) The Mitogen-Activated Protein Kinase Slt2 Promotes Asymmetric Cell Cycle Arrest and Reduces TORC1-Sch9 Signaling in Yeast .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 32 Lacking the Protein Phosphatase Ptc1. Microbiol Spectr 11: e0524922 Guo J, Bryan BA & Polymenis M (2004) Nutrient-specific effects in the coordination of cell growth with cell division in continuous cultures of Saccharomyces cerevisiae. Arch Microbiol 182: 326–30 Hartwell LH & Unger MW (1977) Unequal division in Saccharomyces cerevisiae and its implications for the control of cell division. J Cell Biol 75: 422–35 Henry KA, Blank HM, Hoose SA & Polymenis M (2010) The unfolded protein response is not necessary for the G1/S transition, but it is required for chromosome maintenance in Saccharomyces cerevisiae. PLoS One 5: e12732 Hoose SA, Rawlings JA, Kelly MM, Leitch MC, Ababneh QO, Robles JP, Taylor D, Hoover EM, Hailu B, McEnery KA, et al (2012) A systematic analysis of cell cycle regulators in yeast reveals that most factors act independently of cell size to control initiation of division. PLoS Genet 8: e1002590 Hoose SA, Trinh JT, Leitch MC, Kelly MM, McCormick RF, Spyrou CL, Smith R 3rd & Polymenis M (2013) Saccharomyces cerevisiae deletion strains with complex DNA content profiles. FEMS Microbiol Lett 345: 72–6 Hottiger T, Fürst P, Pohlig G & Heim J (1994) Physiological characterization of the yeast metallothionein (CUP1) promoter, and consequences of overexpressing its transcriptional activator, ACE1. Yeast Chichester Engl 10: 283–296 Johnston GC, Pringle JR & Hartwell LH (1977) Coordination of growth with cell division in the yeast Saccharomyces cerevisiae. Exp Cell Res 105: 79–98 Kaiser C, Michaelis S, Mitchell A, & Cold Spring Harbor Laboratory. (1994) Methods in yeast genetics : a Cold Spring Harbor Laboratory course manual 1994th ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press Kaldis P (1999) The cdk-activating kinase (CAK): from yeast to mammals. Cell Mol Life Sci CMLS 55: 284–296 Kaldis P, Pitluk ZW, Bany IA, Enke DA, Wagner M, Winter E & Solomon MJ (1998) Localization and regulation of the cdk-activating kinase (Cak1p) from budding yeast. J Cell Sci 111 ( Pt 24): 3585–3596 Kaldis P, Sutton A & Solomon MJ (1996) The Cdk-activating kinase (CAK) from budding yeast. Cell 86: 553–564 Lee ME, DeLoache WC, Cervantes B & Dueber JE (2015) A Highly Characterized Yeast Toolkit for Modular, Multipart Assembly. ACS Synth Biol 4: 975–986 Lodish HF (1974) Model for the regulation of mRNA translation applied to haemoglobin synthesis. Nature 251: 385–8 Longtine MS, McKenzie A 3rd, Demarini DJ, Shah NG, Wach A, Brachat A, Philippsen P & Pringle JR (1998) Additional modules for versatile and economical PCR-based gene .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 33 deletion and modification in Saccharomyces cerevisiae. Yeast 14: 953–61 Mair P & Wilcox R (2020) Robust statistical methods in R using the WRS2 package. Behav Res

52: 464–488 Maitra N, Anandhakumar J, Blank HM, Kaplan CD & Polymenis M (2019) Perturbations of Transcription and Gene Expression-Associated Processes Alter Distribution of Cell Size Values in Saccharomyces cerevisiae. G3 Bethesda 9: 239–250 Morgan DO (2007) The cell cycle: principles of control London, U.K.: New Science Press, Ltd Pavletich NP (1999) Mechanisms of cyclin-dependent kinase regulation: structures of Cdks, their cyclin activators, and Cip and INK4 inhibitors. J Mol Biol 287: 821–8 Polymenis M (2022) Two from One: A Short Introduction to Cell Division Mechanisms John Wiley & Sons Polymenis M & Schmidt EV (1997) Coupling of cell division to cell growth by translational control of the G1 cyclin CLN3 in yeast. Genes Dev 11: 2522–31 Poon RY, Yamashita K, Howell M, Ershler MA, Belyavsky A & Hunt T (1994) Cell cycle regulation of the p34cdc2/p33cdk2-activating kinase p40MO15. J Cell Sci 107 ( Pt 10): 2789–2799 Quan P-L, Sauzade M & Brouzes E (2018) dPCR: A Technology Review. Sensors 18: 1271 Russo AA, Jeffrey PD & Pavletich NP (1996) Structural basis of cyclin-dependent kinase activation by phosphorylation. Nat Struct Biol 3: 696–700 Schachter MM, Merrick KA, Larochelle S, Hirschi A, Zhang C, Shokat KM, Rubin SM & Fisher RP (2013) A Cdk7-Cdk4 T-loop phosphorylation cascade promotes G1 progression. Mol Cell 50: 250–260 Sharmin D, Sasano Y, Sugiyama M & Harashima S (2014) Effects of deletion of different PP2C protein phosphatase genes on stress responses in Saccharomyces cerevisiae. Yeast Chichester Engl 31: 393–409 Shiekhattar R, Mermelstein F, Fisher RP, Drapkin R, Dynlacht B, Wessling HC, Morgan DO & Reinberg D (1995) Cdk-activating kinase complex is a component of human transcription factor TFIIH. Nature 374: 283–287 Soma S, Yang K, Morales MI & Polymenis M (2014) Multiple metabolic requirements for size homeostasis and initiation of division in Saccharomyces cerevisiae. Microb Cell 1: 256–266 Tassan JP, Schultz SJ, Bartek J & Nigg EA (1994) Cell cycle analysis of the activity, subcellular localization, and subunit composition of human CAK (CDK-activating kinase). J Cell Biol 127: 467–478 Truong SK, McCormick RF & Polymenis M (2013) Genetic Determinants of Cell Size at Birth and Their Impact on Cell Cycle Progression in Saccharomyces cerevisiae. G3 Bethesda .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 34 3: 1525–30 Wagner M, Pierce M & Winter E (1997) The CDK-activating kinase CAK1 can dosage suppress sporulation defects of smk1 MAP kinase mutants and is required for spore wall morphogenesis in Saccharomyces cerevisiae. EMBO J 16: 1305–1317 Wallace RL, Lu E, Luo X & Capaldi AP (2022) Ait1 regulates TORC1 signaling and localization in budding yeast. eLife 11: e68773 Whinston E, Omerza G, Singh A, Tio CW & Winter E (2013) Activation of the Smk1 mitogen-activated protein kinase by developmentally regulated autophosphorylation. Mol Cell Biol 33: 688–700 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 35 FIGURES FIGURE 1. Cdc28 T169 phosphorylation and Cak1 protein levels vary with different growth conditions in batch cultures. (A) Schematic of the isolation of Cdc28-TAP by immunoprecipitation and processing for mass spectrometry to compare the extent of T169 phosphorylation in cells growing in media with glucose or glycerol as the carbon source. Created with BioRender.com. (B) Plots showing the relative abundance of recovered Cdc28 peptides with the T169 phosphorylation over the abundance of all Cdc28 peptides (y-axis). The ratio for cells growing with glucose was normalized to one, and the ratio for cells growing with glycerol is shown relative to this, from the indicated independent experiments (x-axis). The boxplot graphs were generated with R language functions. The central line inside the box represents the median of the data, while the box itself spans the interquartile range (IQR), and the whiskers extend to a range of 1.5 times the IQR in either direction. The replicates were all biological ones. The raw mass spectrometry data are in Dataset EV1. (C) An antibody against T169-Cdc28 phosphorylation. Protein extracts from exponentially proliferating cells (X2180-5A .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 36 background) in rich undefined medium (YPD) were split in half and either treated (+) or not (-) with lambda phosphatase. The samples were then run on an SDS-PAGE gel and transferred to a nitrocellulose blot. The blot was stained with Ponceau to indicate loading (bottom) and then processed with the antibodies indicated on top. The α-pT169 antibody was raised in rabbits, using the peptide RAFGVPLRAY(pT)HEIVT as an antigen, corresponding to Cdc28 (see

and Methods). (D) Representative immunoblots of V5-Cak1, pT169-Cdc28, Cdc28, and Pgk1 are shown at the top. The boxplot graphs below were drawn as in (B) and display the quantification of each target from four independent experiments. Band intensities were quantified using ImageJ software. For each experiment, the signal for each quantified target was divided by the average intensity for that target in the series across all conditions on the same gel. These ratios were then Log2 transformed. The Log2 transformed values are plotted on the y-axis. The x-axis indicates the respective growth conditions. The extracts were prepared from cells growing exponentially in batch cultures with glucose or glycerol as the carbon source or until they reached the stationary phase after being cultured for 3 days in media with glucose as the carbon source. In all cases, the cells expressed CAK1 from its endogenous chromosomal location but from an N-terminal V5 epitope-tagged CAK1 allele. The indicated p-values were determined using robust bootstrap ANOVA (t1waybt function) followed by post hoc tests (mcppb20 function) from the WRS2 R package. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 37 FIGURE 2. A uORF represses Cak1 synthesis, yet T169 phosphorylation is robust to changes in Cak1 levels. (A) A schematic of the CAK1 upstream genomic region. The wild-type uORF at position -37 encodes a 16-amino acid peptide. The engineered uORF changes are shown below, each with an N-terminal V5 epitope tag fused in-frame with the CAK1 ORF. (B) Boxplots showing the quantification of V5-Cak1, pT169-Cdc28, Cdc28, and Pgk1 in the indicated strains, drawn as in Figure 1. Data for the uORF-35 background is reproduced from Figure 1D for comparison. All experiments and analyses were performed as described in Figure 1D. All the strains were in the prototrophic X2180-5A background. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 38 FIGURE 3. The CAK1 uORF is expressed and represses downstream translation. (A) Schematic of high-copy reporter plasmids. Reporter 1 (top) replaces the uORF coding sequence with mTurquoise2 (uORF expression). Reporter 2 (middle) replaces the CAK1 open reading frame with mRuby2 and contains the uORF. Reporter 3 (bottom) is the same as Reporter 2 but with a mutated uORF start codon (corresponding to the uORF-0 allele in Fig. 2A). (B) Representative images of cells carrying the indicated reporters. Top row: Cells with Reporter 1 only (uORF-mTurquoise2, blue) show blue fluorescence, but not red. Middle/Bottom rows: Cells co-expressing Reporter 1 with either Reporter 2 (containing the uORF) or Reporter 3 (uORF-0). Note the marked increase in mRuby2 (red, CAK1 expression) in the Reporter 3 (uORF-0) co-transformants compared to Reporter 2, demonstrating uORF-mediated repression. The cells were imaged live. The exposure was 500ms for the blue channel and 200ms for the red channel. (C) Boxplots of single-cell mTurquoise2 (uORF) and mRuby2 (CAK1) fluorescence intensity from strain BY4743 co-transformed with Reporter 1 and either Reporter 2 (with uORF) .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 39 or Reporter 3 (uORF mutated). Cultures were grown in glucose or glycerol, as indicated. Fluorescence per cell was quantified with flow cytometry. Boxplots were drawn as in Figure 1. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 40 FIGURE 4. Cak1 levels and T169 phosphorylation respond to growth rate changes with a different dynamic range. (A) Representative immunoblots of V5-Cak1, pT169-Cdc28, Cdc28, and Pgk1, with antibodies indicated, from chemostat cultures limited for nitrogen (N; ammonium sulfate) or carbon (C; glucose), for the strains shown. Samples were collected at the indicated dilution (D) rates (in h-1). (B,C) The graphs display the quantification of each indicated target from independent chemostat experiments under the limitations shown for each strain. Band intensities were quantified and their Log2-transformed expressed ratios were calculated as in Figure 1, for each series across the different dilution rates, and plotted on the y-axis; the x-axis indicates the respective dilution rates. Loess curves are shown in red. All the strains were in the prototrophic X2180-5A background. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 41 FIGURE 5. Cells lacking the CAK1 uORF exhibit reduced viability during quiescence and are outcompeted by wild-type cells in slow-proliferating chemostat cultures. (A) Barplot showing the viability (y-axis) of quiescent cells over several days (x-axis), scored as described in the Materials and Methods. Cells without the CAK1 uORF (uORF-0, strain MSP257) had lower viability compared to cells with the uORF (uORF-16, strain MSP256). (B) Plots displaying the percentage of uORF-0 cells (y-axis, strain MSP257) against wild-type uORF-16 cells (strain BY4742) in mixed chemostat cultures over several days (x-axis) across the indicated dilution rates (D) and nutrient limitations. Loess curves are shown in red. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 42 FIGURE 6. Type 2C protein phosphatases buffer changes in cell cycle kinetics in uORF-0 cells. (A) DNA content analysis of the indicated strains (all in the BY4742 background) in glucose (YPD) or glycerol (YPGlycerol) media growing exponentially in liquid cultures, measured by flow cytometry. Histograms show cell counts (y-axis) vs. fluorescence per cell (x-axis). The average percentage (± SD) of cells in the G1 phase from three independent cultures is indicated in select strains. (B) Cell cycle progression of elutriated synchronous early G1 cells in YP-glucose. The percentage of budded cells is plotted against cell size (fL) for the .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 43 indicated strains. The plots show Loess curves with 95% confidence intervals. Red lines indicate the critical size for budding (the size at which 50% of cells have budded) for the wild type cells (uORF-16, PTC+; strain BY4742), while green lines indicate the critical size of the uORF-0, ptc1-3Δ cells (strain HB530). .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 44 FIGURE 7. Schematic of our findings. The uORF-mediated translational control of the transcripts encoding two core Cdk regulators, Cak1 and Cln3, couples cell division to ribosome availability. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 45 EXPANDED VIEW FIGURES FIGURE EV1. CAK1 expression is not regulated by transcription or proteolysis as a function of growth. (A) Plots of the relative steady-state levels of CAK1 mRNA from multiple independent experiments (y-axis), in the respective growth conditions (x-axis). Cells and growth conditions were the same as in Figure 1C. The boxplots were drawn as in Figure 1. The mRNA levels were measured by ddPCR, as described in the Materials and Methods. All values were normalized to the mean for exponentially growing uORF-35, V5-CAK1 cells (first box). No significant differences were found between the strains and conditions (p > 0.05, Kruskal-Wallis test). (B) A plasmid used for copper-inducible, ectopic V5-Cak1 synthesis is schematically shown on top and described in Materials and Methods. The immunoblot below is from cells (BY4742 background) carrying that plasmid showing the protein levels detected by the α-V5 antibody at the times indicated after induction with CuSO4. The band corresponds to a mass .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 46 consistent with the V5-Cak1 protein. (C) The V5-Cak1 protein is highly stable in glucose and glycerol media. Exponentially proliferating cells (BY4742 background) carrying the pCUP1-V5-CAK1 plasmid (described in A) were induced for 20 min with CuSO4 (added at 0.5 mM) and then treated with cycloheximide (CHX; 100 μg/mL) to block protein synthesis. The cultures were kept at 30°C, and the V5-Cak1 levels were monitored by immunoblotting (top) of samples taken at the indicated time points. Loading is shown at the bottom from the same blots stained with Ponceau. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 47 FIGURE EV2. Eliminating the uORF of CAK1 does not affect cell cycle kinetics in batch cultures. (A) Cell cycle parameters for the uORF-16 CAK1 (MSP256) and uORF-0 CAK1 (MSP257) strains shown on the x-axis, grown in the indicated media. The boxplots, drawn as described in Figure 1, show (from top to bottom): birth size, mean cell size, %budding, and doubling time (Td), on the y-axis. Values were obtained from independent experiments. No significant differences were found between the strains (p > 0.05, Wilcoxon rank sum test). (B) Cell cycle progression of elutriated synchronous early G1 cells in YP-glycerol. The percentage of budded cells is plotted against cell size (fL) for the same strains as in (A). The plots show Loess curves with 95% confidence intervals. Red lines indicate the critical size for budding (the size at which 50% of cells have budded). .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 48 FIGURE EV3. Overexpression of CAK1 in the stationary phase reduces cell size and viability. (A) Cell size distribution of stationary phase cells (BY4743) with (right panel) or without (left panel) the pCUP1-V5-CAK1 plasmid. Histograms show cell counts (y-axis) by size (x-axis) from 6-day cultures. Density lines represent cultures induced with 0.5 mM CuSO4 to express V5-Cak1 (yellow) versus mock-treated controls (blue). (B) DNA content analysis of the same samples shown in (A), measured by flow cytometry. Histograms show cell counts (y-axis) vs. fluorescence per cell (x-axis). The average percentage (± SD) of cells in the G1 phase from three independent cultures is indicated. (C) Representative cell viability plates for cultures carrying the pCUP1-V5-CAK1 plasmid, as shown in (A, right panel). Cultures were either mock-treated (left plates) or induced with CuSO4 (right plates) before plating on media lacking CuSO4. Viability dropped from 67% in mock-treated cultures to 19% in CuSO₄-treated cultures (χ2= 8123.4, df = 1, p < 2.2e-16). .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint 49 FIGURE EV4. Proliferation of type 2C protein phosphatase mutants with or without the uORF in CAK1. (A) Growth spot assays to assess the relative growth fitness of the indicated strains (all in the BY4742 background) on glucose (YPD) or glycerol (YPGlycerol) media. Overnight cultures were normalized to an initial cell density and aliquots were spotted at 10-fold serial dilutions. The plates were imaged after 2 (glucose) or 3 (glycerol) days at 30°C. (B) Growth curves of the strains shown in (A), cultured in YPD glucose medium. Ln-transformed absorbance values (600nm) are plotted on the y-axis against time (x-axis). Measurements were taken using a microplate reader and corrected for background absorbance of the sterile medium. Data are presented as Loess regression curves; shaded regions represent 95% confidence intervals (n=8 independent replicates per strain). .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.28.702203doi: bioRxiv preprint