Suppression of cdc13-2-associated senescence by pif1-m2 requires Ku-mediated telomerase recruitment

Abstract In Saccharomyces cerevisiae, recruitment of telomerase to telomeres requires an interaction between Cdc13, which binds single-stranded telomeric DNA, and the Est1 subunit of telomerase. A second pathway involving an interaction between the yKu complex and telomerase RNA (TLC1) contributes to telomerase recruitment but cannot sufficiently recruit telomerase on its own to prevent replicative senescence when the primary Cdc13-Est1 pathway is abolished—for example, in the cdc13-2 mutant. In this study, we find that mutation of PIF1, which encodes a helicase that inhibits telomerase, suppresses the replicative senescence of cdc13-2 by increasing reliance on the yKu-TLC1 pathway for telomerase recruitment. Our findings reveal new insight into telomerase-mediated telomere maintenance.


Introduction
Telomeres are composed of G/C-rich repetitive sequences at the termini of eukaryotic chromosomes and play a pivotal role in genome maintenance by "capping" chromosome ends, preventing them from unwanted nucleolytic degradation, homologous recombination, and fusion with neighboring chromosomes (Jain and Cooper 2010). In addition, to overcome progressive telomere shortening due to the end replication problem, telomeres are elongated by a specialized reverse transcriptase called telomerase. In the budding yeast Saccharomyces cerevisiae, telomerase is minimally composed of the protein subunit Est2 and the RNA subunit TLC1 (Singer and Gottschling 1994;Lingner et al. 1997). However, additional accessory proteins, Est1 and Est3, are required for telomerase activity in vivo and are thought to be involved in the recruitment and/or activation of telomerase (Wellinger and Zakian 2012). Eliminating any of the Est proteins or TLC1 results in an "ever shorter telomeres" (est) phenotype characterized by progressive telomere shortening that ultimately leads to replicative senescence (Lundblad and Szostak 1989;Singer and Gottschling 1994;Lendvay et al. 1996).
Maintaining telomere length homeostasis through the regulation of telomerase is essential for genome stability. Several lines of evidence suggest that the recruitment of telomerase to telomeres involves a direct interaction between the Est1 subunit of telomerase and Cdc13, a protein that binds single-strand telomeric DNA with high affinity (Lin and Zakian 1996;Nugent et al. 1996). Expression of a Cdc13-Est2 fusion protein can support telomere maintenance in an est1D null mutant, suggesting that the main function of Est1 is to bring telomerase to telomeres (Evans and Lundblad 1999). Cdc13 is essential for telomere capping, so a null mutation is lethal; however, an extensively studied point mutant, cdc13-2, is not capping defective but displays an est phenotype (Nugent et al. 1996). The amino acid mutated in cdc13-2, E252, lies within the recruitment domain (RD), which is able to recruit telomerase to telomeres when fused to the DNA-binding domain of Cdc13 (Pennock et al. 2001). The mutation (E252K) results in a charge swap and can be suppressed by est1-60, which encodes a mutant Est1 with a reciprocal charge swap (K444E), suggesting a direct physical interaction between the two proteins (Pennock et al. 2001). Consistent with this idea, purified fulllength Cdc13 and Est1 interact in vitro (Wu and Zakian 2011), and structural analysis revealed two conserved motifs within the Cdc13 RD, called Cdc13 EBM-N and Cdc13 EBM-C (referring to N-and C-terminal Est1-binding motifs, respectively), responsible for this interaction (Chen et al. 2018). The Cdc13 E252K mutation resides within the latter motif. Surprisingly, mutations in the Cdc13 EBM-C motif, including E252K, do not abolish the interaction between Cdc13 and Est1 in vitro despite causing a dramatic reduction in Est1 telomere association in vivo (Chan et al. 2008;Wu and Zakian 2011;Chen et al. 2018). Thus, the mechanism by which the Cdc13 EBM-C motif promotes telomerase-mediated telomere extension is still unclear.
In contrast, mutations in Cdc13 EBM-N abolish the Cdc13-Est1 interaction in vitro, yet only result in a modest reduction in Est1 telomere association and short, but stable, telomere length in vivo (Chen et al. 2018). This telomerase recruitment pathway works in parallel with a second pathway involving Sir4, the yKu complex, and TLC1. Double-strand telomeric DNA is bound by Rap1 (Buchman et al. 1988;Conrad et al. 1990), which interacts with Sir4 (Moretti et al. 1994). Sir4, in turn, interacts with the Yku80 subunit of the yKu complex (Roy et al. 2004), which binds to the tip of a 48-nt hairpin in TLC1 (Peterson et al. 2001;Stellwagen et al. 2003;Chen et al. 2018). Mutations that abolish the yKu-TLC1 interaction (e.g., tlc1D48 and yku80-135i) result in slightly short but stable telomeres (Peterson et al. 2001;Stellwagen et al. 2003), much like Cdc13 EBM-N mutations. Disrupting both the yKu-TLC1 interaction and Cdc13 EBM-N -Est1 interaction results in an est phenotype (Chen et al. 2018).
We previously showed that a double-strand break adjacent to at least 34 bp of telomeric sequence is efficiently extended by telomerase, resulting in the addition of a de novo telomere, but this does not occur in Cdc13 EBM-C mutants, such as cdc13-2 (Strecker et al. 2017). Surprisingly, we found that the lack of telomere addition in cdc13-2 cells can be suppressed by the pif1-m2 mutation (Strecker et al. 2017). In this study, we find that pif1-m2 suppresses the replicative senescence caused by the cdc13-2 mutation in a manner dependent on the yKu-TLC1 telomerase recruitment pathway. In addition, pif1-m2 suppresses the replicative senescence caused by disrupting both the yKu-TLC1 and Cdc13 EBM-N -Est1 interactions. These observations provide new insight into the complexity of telomerase-mediated telomere maintenance.

Liquid culture senescence assay
Liquid culture senescence assays were performed essentially as previously described (van Mourik et al. 2016). Each senescence assay started with diploid strains. Freshly dissected haploid spores were allowed to form colonies on YPD agar plates after two days of growth at 30 C. Cells from these colonies were serially passaged in liquid culture medium at 24-h intervals. For each passage, the cell density of each culture was measured by optical density (calibrated by cell counting using a haemocytometer), and the cultures were diluted back into fresh medium at a cell density of 2 Â 10 5 cells/ml. Cell density was plotted as a function of population doublings.

Telomere Southern blot
Telomere length analysis by Southern blotting was performed essentially as previously described (van Mourik et al. 2018). A 1.8-kb DNA fragment containing telomeric sequences generated from the BsmAI-digestion of plasmid pYt103 (Shampay et al. 1984) was loaded together with each sample. Southern blots were probed with a telomere-specific probe (5 0 -TGTGGGTGTGGTGTGTGGGT GTGGTG-3 0 ).

Results and discussion
Mutation of PIF1 suppresses the replicative senescence caused by the cdc13-2 mutation To investigate how telomere addition is possible in a cdc13-2 pif1-m2 genetic background, we serially passaged cells to determine whether they would senesce. For these experiments, we used strains from our previous study (Strecker et al. 2017): cdc13D or cdc13D pif1-m2 cells, kept alive by the presence of a high-copy plasmid expressing wild-type CDC13 and the URA3 gene, transformed with an additional high-copy plasmid containing either CDC13 or cdc13-2. These cells also carried a deletion of RAD52 to prevent homologous recombination-mediated telomere maintenance (Claussin and Chang 2015). We then counterselected the first plasmid by growing cells on media containing 5-fluoroorotic acid (5-FOA), which is toxic to cells expressing URA3. 5-FOA-resistant colonies were subsequently serially passaged on agar plates ( Figure 1A). Senescence was apparent for cdc13-2 PIF1 cells already after the first passage, whereas CDC13 and cdc13-2 pif1-m2 strains did not show any sign of senescence even after the fourth passage. We analyzed the telomere length of these strains and found that, consistent with previous studies, pif1-m2 has increased telomere length compared with wild type (Schulz and Zakian 1994) while the telomeres are very short in the cdc13-2 mutant (Lendvay et al. 1996;Nugent et al. 1996). Interestingly, cdc13-2 pif1-m2 telomeres are approximately wild-type in length, albeit more heterogeneous, and stable throughout the course of the experiment ( Figure  1B). Our findings indicate that telomerase-mediated telomere extension can occur in cdc13-2 pif1-m2 cells, allowing cells to maintain telomere length homeostasis and avoid replicative senescence. Because the strains used for this experiment have an unusual genotype (Table 1), relevant for our previous study (Strecker et al. 2017) but not for this study, we performed all subsequent experiments in a different strain background (W303), where none of the genes were overexpressed.
The yKu-TLC1 telomerase recruitment pathway is necessary to maintain telomere length in cdc13-2 pif1-m2 cells We hypothesized that the yKu-TLC1 pathway may become essential for telomere length homeostasis in cdc13-2 pif1-m2 strains. To test this possibility, haploid meiotic progeny derived from the sporulation of CDC13/cdc13-2 PIF1/pif1-m2 YKU80/yku80-135i and CDC13/cdc13-2 PIF1/pif1-m2 TLC1/tlc1D48 heterozygous diploids were serially propagated in liquid culture for several days (Figure 2, A and B). The yku80-135i and tlc1D48 alleles disrupt the interaction between the yKu complex and TLC1 (Peterson et al. 2001;Stellwagen et al. 2003). As expected, cdc13-2 cultures grew slower as the experiment progressed and cells senesced, but growth was eventually restored upon the emergence of survivors that utilize recombination-mediated mechanisms to maintain telomeres (Lendvay et al. 1996). In contrast, the cdc13-2 pif1-m2 strains did not senesce, confirming our previous observations in a different strain background (S288C for strains used in Figure 1 as opposed to W303 for all other strains used in this study). The cdc13-2 pif1-m2 yku80-135i and cdc13-2 pif1-m2 tlc1D48 triple mutants showed a pattern of senescence and survivor formation, indicating that the yKu-TLC1 telomerase recruitment pathway is required for telomere length homeostasis in cdc13-2 pif1-m2 cells. The yku80-135i and tlc1D48D alleles caused cdc13-2 and cdc13-2 pif1-m2 strains to senesce faster, but the reason for this is currently unclear.
The abundance of TLC1 RNA is reduced to $30% and $48% in yku80-135i and tlc1D48 mutants, respectively, compared with wild-type cells (Zappulla et al. 2011). In addition, disrupting the yKu-TLC1 interaction causes mislocalization of TLC1 to the cytoplasm (Gallardo et al. 2008;Pfingsten et al. 2012). It is possible that reduced abundance and/or mislocalization of TLC1, rather than disruption of the yKu-TLC1 telomerase recruitment pathway, is responsible for the senescence of cdc13-2 pif1-m2 yku80-135i and cdc13-2 pif1-m2 tlc1D48 triple mutants. Sir4 is also required for the yKu-TLC1 recruitment pathway, but deletion of SIR4 does not affect TLC1 abundance (Hass and Zappulla 2015), and there is no evidence that sir4D affects TLC1 localization. We find that cdc13-2 pif1-m2 sir4D triple mutants also senesce ( Figure 2C), although the "dip" in the senescence curve was more shallow (note the difference in scale on the y-axis). The shallow dip is consistent across multiple isolates of cdc13-2 pif1-m2 sir4D (nine isolates) as well as cdc13-2 sir4D (six isolates). The presence of the shallow dip in cdc13-2 sir4D indicates that this effect is due to sir4D, and is unrelated to the pif1-m2 suppression of cdc13-2 senescence. This effect of sir4D has also been observed with respect to the senescence of mre11D yku80D double mutants, which was attributed to increased recombination and amplification of Y 0 subtelomeric elements (Liu et al. 2021). While these experiments leave open the possibility that reduced abundance and/or mislocalization of TLC1 plays a role in the senescence of cdc13-2 pif1-m2 cells with an additional yku80-135i, tlc1D48, or sir4D mutation, the simplest interpretation of our findings is that recruitment of telomerase via the yKu-TLC1 pathway is indeed required for telomere length homeostasis in cdc13-2 pif1-m2 cells.

Strain name
Genotype Source

This study
In summary, these findings indicate that mutation of PIF1 allows sufficient telomerase recruitment to avoid replicative senescence caused by disruption of the Cdc13 EBM-C -Est1 interaction alone, or double disruption of both the Cdc13 EBM-N -Est1 and yKu-TLC1 interactions. However, suppression is not possible when both the Cdc13 EBM-C -Est1 and yKu-TLC1 interactions are abolished. Disruption of both the Cdc13 EBM-N -Est1 and Cdc13 EBM-C -Est1 interactions can be suppressed by mutation of PIF1 ( Figure 2E), suggesting that the Cdc13 EBM-N -Est1 interaction plays a more minor role, likely in support of the Cdc13 EBM-C -Est1 interaction. Our findings suggest that Pif1 inhibits telomerase regardless of how telomerase is recruited: mutation of PIF1 in cdc13-2 cells allows increased telomerase recruitment via the yKu-TLC1 pathway, while mutation of PIF1 in cdc13-F237A tlc1D48 cells allows increased telomerase recruitment via the Cdc13 EBM-C -Est1 pathway.
Mutation of PIF1 cannot suppress the replicative senescence of est1D The cdc13-2 mutation greatly reduces the recruitment of Est1 to telomeres (Chan et al. 2008), and the expression of Cdc13-Est2 fusion protein allows cells to stably maintain their telomeres in the absence of Est1 (Evans and Lundblad 1999). Therefore, it was possible that pif1-m2 suppresses the replicative senescence caused by cdc13-2 by somehow bypassing the need for Est1 for telomerase-mediated telomere extension. To test this idea, we sporulated an EST1/est1D PIF1/pif1-m2 heterozygous diploid and monitored growth of the haploid meiotic progeny ( Figure 2F). We find that est1D pif1-m2 double mutants senesce like est1D single mutants, indicating that mutation of PIF1 cannot bypass the need for Est1.
Tel1 acts through the Cdc13 EBM-C motif to regulate telomere length Because the cdc13-2 mutation normally results in a complete defect in telomerase-mediated telomere extension, it has not been possible to perform classical genetic epistasis experiments to determine which telomere length regulators act through the Cdc13 EBM-C -Est1 pathway. The viability and non-senescence of cdc13-2 pif1-m2 strains give us the opportunity to do so. The Rap1interacting factors, Rif1 and Rif2, negatively regulate telomerase (Hardy et al. 1992;Wotton and Shore 1997) while the Tel1 kinase is a positive regulator (Greenwell et al. 1995). We measured the telomere length of haploid strains propagated for over 100 population doublings after being generated from the sporulation of heterozygous diploids ( Figure 3A). We find that cdc13-2 pif1-m2 cells have short telomeres, which is in contrast to the more wildtype, but heterogeneous, length telomeres shown in Figure 1. The difference is most likely due to different genetic backgrounds (strains in Figure 1 are of the S288C background, with an additional deletion of RAD52, while all other strains used in  this study are of the W303 background; Table 1), but not due to cdc13-2 being expressed from a high-copy plasmid in Figure 1, because overexpression of neither CDC13 nor cdc13-2 affects telomere length ( Figure 3B). While deletion of RIF1 elongates cdc13-2 pif1-m2 telomeres, both cdc13-2 pif1-m2 rif2D and cdc13-2 pif1-m2 tel1D triple mutants have very similar telomere lengths compared with cdc13-2 pif1-m2, indicating that Rif2 and Tel1 function upstream and in the same pathway as the Cdc13 EBM-C motif (Figure 3) G cdc13-2 cdc13-2 pif1-m2 cdc13-2 sir4Δ cdc13-2 pif1-m2 sir4Δ Figure 2 Telomeres are maintained by the yKu-TLC1 pathway in cdc13-2 pif1-m2 cells. Senescence was monitored in liquid culture by serial passaging of haploid meiotic progeny derived from the sporulation of VSY20 (A), VSY7 (B), EFSY142 (C), EFSY73 (D), FRY867 (E), CAY2 (F), and MCY815 (G). Average cell density 6SEM of 3-9 independent isolates per genotype (except n ¼ 2 for cdc13-F237A, E252K) is plotted. telomere length, while the relationship between Tel1 and Rif1 is more complex and telomere-specific (Craven and Petes 1999;Sholes et al. 2021).
Tel1 often functions in concert with a related kinase, Mec1. Mutation of both MEC1 and TEL1 results in an est phenotype (Ritchie et al. 1999). Because Tel1 promotes telomerase activity through the Cdc13 EBM-C -Est1 interaction, we examined whether the same is true for Mec1. If so, the replicative senescence of mec1 tel1 double mutants, like cdc13-2, should also be suppressed by pif1-m2. We sporulated a MEC1/mec1-21 TEL1/tel1D PIF1/pif1-m2 diploid strain and monitored the growth of the mec1-21 tel1D and mec1-21 tel1D pif1-m2 haploid meiotic progeny. Both strains exhibited a similar rate of senescence ( Figure 2G), indicating that pif1-m2 cannot suppress the est phenotype of a mec1 tel1 double mutant, and that Mec1 functions in a different pathway than Tel1 to promote telomerase activity, as previously proposed (Ritchie et al. 1999;Keener et al. 2019).
In summary, our findings provide new insight into how telomerase is recruited to telomeres in S. cerevisiae. Further work is needed to determine how the Cdc13 EBM-C motif functions, what its relationship is with the Cdc13 EBM-N motif, and the role of Tel1 in promoting telomerase recruitment.

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. Wild type cdc13-2 pif1-m2 tel1Δ cdc13-2 pif1-m2 tel1Δ Figure 3 The cdc13-2 allele is epistatic to rif2D and tel1D with respect to telomere length regulation in a pif1-m2 background. Telomere Southern blot analysis of strains of the indicated genotypes. All strains were propagated for at least 100 population doublings before Southern blot analysis. Parental diploids 1, 2, and 3 are EFSY8, EFSY9, and EFSY31, respectively. The black arrowhead indicates a 1.8-kb DNA fragment generated from the BsmAIdigestion of plasmid pYt103 loaded as a control.