Roles of Heterochromatin and Telomere Proteins in Regulation of Fission Yeast Telomere Recombination and Telomerase Recruitment*

When the telomerase catalytic subunit (Trt1/TERT) is deleted, a majority of fission yeast cells survives by circularizing chromosomes. Alternatively, a small minority survives by maintaining telomeric repeats through recombination among telomeres. The recombination-based telomere maintenance in trt1Δ cells is inhibited by the telomere protein Taz1. In addition, catalytically inactive full-length Trt1 (Trt1-CI) and truncated Trt1 lacking the T-motif and reverse transcriptase (RT) domain (Trt1-ΔT/RT) can strongly inhibit recombination-based survival. Here, we investigated the effects of deleting the heterochromatin proteins Swi6 (HP1 ortholog) and Clr4 (Suv39 family of histone methyltransferases) and the telomere capping complex subunits Poz1 and Ccq1 on Taz1- and Trt1-dependent telomere recombination inhibition. The ability of Taz1 to inhibit telomere recombination did not require Swi6, Clr4, Poz1, or Ccq1. Although Swi6, Clr4, and Poz1 were dispensable for the inhibition of telomere recombination by Trt1-CI, Ccq1 was required for efficient telomere recruitment of Trt1 and Trt1-CI-dependent inhibition of telomere recombination. We also found that Swi6, Clr4, Ccq1, the checkpoint kinase Rad3 (ATR ortholog), and the telomerase regulatory subunit Est1 are all required for Trt1-ΔT/RT to inhibit telomere recombination. However, because loss of Swi6, Clr4, Rad3, Ccq1, or Est1 did not significantly alter the recruitment efficiency of Trt1-ΔT/RT to telomeres, these factors are likely to enhance the ability of Trt1-ΔT/RT to inhibit recombination-based survival by contributing to the negative regulation of telomere recombination.

Telomeres, the ends of eukaryotic chromosomes, must fulfill two essential functions to achieve stable inheritance of intact chromosomes. First, telomeres must protect chromosome ends from uncontrolled degradation, end-to-end fusion, and recombination. Second, telomeres must allow complete replication of linear chromosome ends, which cannot be fully replicated by replicative DNA polymerases (1).
In most eukaryotic organisms, telomeric DNA is composed of short GT-rich repeat sequences and extended by telomerase, which utilizes its tightly bound RNA subunit as a template for de novo telomeric repeat DNA addition (2). On the other hand, recombination-based telomerase-independent mechanisms can also extend telomeric GT-rich repeats in various model organisms when telomerase is inactivated and in ϳ10% of human tumors (3).
In multicellular organisms including humans, expression levels of telomerase subunits and overall telomere repeat length as well as composition and modification status of various telomere bound proteins are carefully regulated based on tissue types and developmental stages (4). In fact, studies have uncovered connections between dysfunctional telomeres and various human age-related diseases and cancer in recent years (4,5).
Although most of the telomeric GT-rich repeats are composed of double-stranded DNA, telomeric DNA terminates with a 3Ј GT-rich single-stranded DNA, commonly referred to as G-tail. Because telomerase cannot act on blunt ends (6), the G-tail is essential for telomere extension by telomerase. Both the G-tail and the double-stranded DNA portion of the GTrich telomere repeats are coated by various sequence specific telomere-binding proteins (7), which are critical to prevent telomere-bound DNA repair and DNA damage checkpoint proteins from causing telomere fusions and permanent cell cycle arrest (8,9).
Interestingly, various DNA repair and checkpoint factors, such as Ku70⅐Ku80, Mre11⅐Rad50⅐Nbs1, ATM, and ATR⅐ ATRIP, play critical roles in telomere maintenance (10 -12). In addition, the formation of heterochromatin structure at telomeres has been observed in many organisms, and the regulation of heterochromatin formation has been suggested to contribute to the proper protection of telomeres (13). However, it is still not fully understood how heterochromatin structures might affect the ability of telomere-specific factors and DNA damage response proteins to regulate telomere functions. Therefore, we decided to investigate how recombinationbased telomere maintenance and recruitment of telomerase are affected by loss of various telomere-associated proteins or proteins involved in the formation of telomere heterochromatin in fission yeast Schizosaccharomyces pombe. Fission yeast cells utilize telomere proteins that are highly conserved with mammalian telomere proteins (7). Moreover, the mechanism of heterochromatin formation is very well conserved between fission yeast and mammalian cells (14).
The catalytic subunit of telomerase, known as TERT (telomerase reverse transcriptase) (2), is encoded by the trt1 ϩ gene in fission yeast (15). The Trt1 subunit forms a stable complex with its regulatory subunit Est1 and a telomerase RNA TER1 (16 -18). All three subunits are essential for telomere extension by telomerase in fission yeast. In addition, Ccq1, a subunit of the Pot1 telomere-capping complex (composed of Pot1⅐Tpz1⅐Poz1⅐ Ccq1), is critical for the recruitment of telomerase to telomeres and the inhibition of recombination at telomeres (19,20) (see Fig. 1 and supplemental Table S1).
When Trt1 is deleted, fission yeast cells progressively lose their telomeric DNA and viability. However, trt1⌬ cells can eventually generate survivors either by circularizing chromosomes or by maintaining telomeric repeats through recombination among telomeres (21). Chromosome circularization is a much more frequently used mode of survival in fission yeast (21); however, rare survivors, which utilize a homologous recombination (HR) 2 -based mechanism to maintain linear chromosomes, can be selected in serially diluted trt1⌬ liquid cultures, as "linear" trt1⌬ survivors have a selective advantage in competitive growth conditions given that they grow faster than circular survivors (21).
The linear mode of telomerase-independent survival in fission yeast requires the HR protein Rad22 Rad52 , Tel1 ATM , the Mre11⅐Rad50⅐Nbs1 complex, Rqh1 RecQ-like DNA helicase, and the telomere protein Rap1 (22,23). Thus, linear survivors appear to resemble budding yeast Type II recombination survivors or the mammalian alternative lengthening of telomeres (ALT) mode of telomere maintenance (23,24). Moreover, we have previously shown that the telomeric GT-rich repeat-specific double-stranded DNA-binding protein Taz1 plays a critical role in inhibiting recombination-based telomere maintenance in trt1⌬ cells (23) and that taz1⌬ trt1⌬ cells sustain robust growth with stable linear chromosomes (21,23).
Reintroduction of Taz1 into taz1⌬ trt1⌬ linear survivor cells strongly induces chromosome circularization due to inhibition of recombination-based telomere maintenance (23). Similarly, reintroduction of catalytically inactive Trt1 (Trt1-CI) or a C-terminal-truncated Trt1, which lacks the telomerase-specific T-motif and the reverse transcriptase domain (Trt1-⌬T/ RT), also causes circularization of chromosomes in taz1⌬ trt1⌬ survivor cells, uncovering a RT-independent role for Trt1 in the inhibition of telomere recombination (23). Moreover, we have established that the non-homologous end-joining (NHEJ) DNA repair protein complex Ku70⅐Ku80 is essential for inhibition of telomere recombination by Trt1-CI, whereas Ku70⅐Ku80 is dispensable for Taz1-dependent inhibition of telomere recombination (23).
Fission yeast Taz1 protein is thought to represent the counterpart of the mammalian telomere proteins TRF1 and TRF2 and binds specifically to the double-stranded DNA portion of telomeric repeats (7,25). In addition to inhibiting telomere recombination in trt1⌬ cells, Taz1 is essential for protection of telomeres against NHEJ-dependent telomere fusion in G 1 phase (26) and efficient replication of telomeric GT-rich repeats by replicative DNA polymerases (27). Moreover, Taz1 is important for the recruitment of the telomeric proteins Rap1 and Rif1 to telomeres, and deletion of taz1, rap1, or rif1 leads to telomerase-dependent expansion of the GT-rich repeat-tract in fission yeast, indicating that they are involved in the negative regulation of telomerase (21,25,28).
Studies have further shown that Taz1 contributes to the formation of heterochromatin structures at telomeres (25,29). Taz1, Ccq1, and the RNAi machinery redundantly contribute to the formation of telomeric heterochromatin by promoting the recruitment of the Snf2/Hdac-containing Repressor Complex (SHREC) (30). The assembly of heterochromatin in fission yeast involves the methylation of histone H3 on lysine 9 (H3 K9me) by Clr4, an ortholog of the mammalian Suv39 family of histone methyltransferases (31). Moreover, fission yeast heterochromatin is enriched for Swi6, a HP1 ortholog that specifically recognizes and binds H3 K9me (32) (see Fig. 1). Deletion of clr4 or swi6 has been suggested to elevate recombination among sub-telomeric regions (33). However, the contribution of heterochromatin in the regulation of recombination-based telomere maintenance has not been investigated in fission yeast.
Here, we tested if Taz1-or Trt1-dependent inhibition of telomere recombination requires the presence of heterochromatin proteins (Swi6 and Clr4) or Pot1 complex components (Ccq1 and Poz1). Our results establish that Taz1 and Trt1-CI can efficiently inhibit telomere recombination in the absence of telomeric heterochromatin or the intact telomere-capping complex. On the other hand, our investigations utilizing Trt1-⌬T/RT implicate a subtle contribution of heterochromatin and the checkpoint kinase Rad3 ATR in repression of telomere recombination in fission yeast.
Pulsed-field Gel Electrophoresis (PFGE)-Chromosomal DNA samples were prepared in agarose plugs from extensively restreaked strains as previously described (10). NotI-digested DNA samples were fractionated in a 1% agarose gel with 0.5ϫ TAE buffer (20 mM Tris acetate, 0.5 mM EDTA) at 14°C using the CHEF-DR III system (Bio-Rad) at 6 V/cm (200 V) and a pulse time of 60 -120 s for 24 h. The probes specific for telomeric C, I, L, and M NotI fragments (see Fig. 2A) were prepared as previously described (21).
Southern Blot Analysis-EcoRI-digested DNA was prepared from fission yeast cells and separated in a 1% agarose gel at 100 V for 4 h. DNA was then transferred to a Hybond XL membrane (Amersham Biosciences) for 2 h in transfer buffer (1.5 M NaCl, 0.02 M NaOH). The membrane was then hybridized with a telomeric repeat DNA probe (10).

Chromatin Immunoprecipitation (ChIP) Analysis and Dot Blot-Ex-
ponentially growing cells were processed for ChIP as previously described (39). For Trt1-myc ChIP, monoclonal anti-Myc antibody (9B11, Cell Signaling) was used, and ChIP data were quantified using dot blot hybridization. First, ChIP and input DNA samples were denatured by boiling at 100°C for 10 min in 0.4 M NaOH and 10 mM EDTA, snapchilled on ice, and blotted onto a Hybond XL membrane (GE Healthcare). Dot blots were then hybridized with a telomeric repeat DNA probe (10) and exposed to a phosphorimaging cassette (GE Healthcare), and hybridization signals were quantified by using Image-Quant software. For Rhp51 ChIP, polyclonal anti-Rad51 antibody (A-92, Santa Cruz) was used, and quantitative real-time PCR was used to analyze ChIP samples. Percent precipitated DNA values were calculated based on ⌬Ct between input and immunoprecipitation samples after performing several independent triplicate SYBR Green-based real-time PCR (Bio-Rad) using TAS1 primers jk380 and jk381 (39). Two-tailed Student t tests were performed, and p values Յ0.05 were considered as statistically significant differences.

Clr4, Swi6, Poz1, and Ccq1 Are Dispensable for Recombination-based Maintenance of Telomeres in taz1⌬ trt1⌬ Cells-
The telomeric repeat binding protein Taz1 is essential for proper maintenance of telomeric heterochromatin in fission yeast (25,29). Taz1 can promote the formation of heterochromatin near telomeric GT-rich repeat sequences by promoting the accumulation of the mammalian HP1 ortholog Swi6 to telomeric repeats even when telomeric repeats are inserted in the middle of chromosomes (29). Accumulation of Swi6 at telomeric and sub-telomeric regions is dependent on the Suv39 family histone H3 lysine 9 methyltransferase Clr4 (29).
Thus, it appears that the ability of Taz1 to inhibit telomerase is not dependent on heterochromatin formation at telomeres and sub-telomeres. However, because formation of heterochromatin has been reported to be important for the inhibition of recombination at centromeres and mating type loci as well as sub-telomeres in fission yeast (33,41), we decided to test if the ability of Taz1 to inhibit recombination-based telomere main-FIGURE 1. A model of fission yeast telomere proteins discussed in this study. The complex formed by Taz1, Rap1, Poz1, Tpz1, Pot1, and Ccq1 is thought to resemble the mammalian telomere complex shelterin. Taz1, Rap1, and Poz1 are important for the negative regulation of telomerase, whereas Ccq1, Tpz1, and Pot1 are implicated in the recruitment of telomerase to telomeres. Ccq1 is also associated with the SHREC complex, and it promotes the Clr4-dependent methylation of histone H3 lysine 9 (H3 K9me) and the accumulation of Swi6 at telomeres. tenance in trt1⌬ cells might require Taz1-dependent heterochromatin formation at telomeres. Accordingly, we decided to generate taz1⌬ trt1⌬ swi6⌬ and taz1⌬ trt1⌬ clr4⌬ triple mutant strains to examine whether the reintroduction of Taz1 would lead to chromosome circularization. Additionally, we tested whether Clr4 or Swi6 is required for the inhibition of recombination-based telomere maintenance by the catalytically inactive Trt1 (Trt1-CI) or the N-terminal Trt1 fragment lacking the T-motif and RT domain (Trt1-⌬T/RT) (see Fig. 3A).
It was possible that Swi6-and Clr4-dependent formation and spreading of heterochromatin at telomeres/sub-telomeres might be essential for telomere maintenance in taz1⌬ trt1⌬ cells. To test this, we first generated multiple independent triple mutant strains, restreaked them extensively on YES plates, and tested their telomere status by PFGE. Both taz1⌬ trt1⌬ swi6⌬ and taz1⌬ trt1⌬ clr4⌬ cells were able to stably maintain telomeres (see Fig. 2, B and C). Thus, we concluded that neither Swi6 nor Clr4 is required for recombination-based telomere maintenance observed in taz1⌬ trt1⌬ cells.
Previously, it has been proposed that loss of proper telomere capping might allow mammalian tumor cells to survive more efficiently through the recombination-based ALT telomere maintenance mechanism (3). Therefore, in addition to Swi6 and Clr4, we decided to test if loss of the Pot1 telomere-capping complex subunits Poz1 and Ccq1 might affect the ability of Taz1 or Trt1 to inhibit recombination-based telomere maintenance (Fig.  1). We did not test the roles of Pot1 or Tpz1 because these proteins are essential for telomere capping, and deletion of these genes leads to immediate chromosome circularization (19).
Ccq1 associates with the SHREC heterochromatin effector complex, and it has been proposed that Ccq1 collaborates with Taz1 in promoting sub-telomeric recruitment of the SHREC complex (30). Moreover, Ccq1 has been implicated in the inhibition of recombination at telomeres (19,20). Similar to taz1⌬ and ccq1⌬ cells, poz1⌬ cells are also defective in transcriptional silencing of a marker gene inserted near telomeres, indicative of a failure in proper heterochromatin formation (35). In addition, poz1⌬ cells carry massively elongated telomeres, suggesting that Poz1 is required for the negative regulation of telomerase (19). We next tested the possibility that the presence of Poz1 or Ccq1 is required for telomere maintenance in taz1⌬ trt1⌬ cells. However, because multiple independently derived taz1⌬ trt1⌬ poz1⌬ and taz1⌬ trt1⌬ ccq1⌬ strains all stably maintained telomeres after extensive restreaking on YES plates (Fig. 2, D and E, and data not shown), we concluded that the presence of the intact Pot1 telomere capping complex is not essential for recombinationbased telomere maintenance in taz1⌬ trt1⌬ cells.
Taz1-dependent Inhibition of Telomere Recombination Does Not Require Swi6, Clr4, Poz1, or Ccq1-Having established that Swi6, Clr4, Poz1, and Ccq1 are not essential for telomere maintenance in taz1⌬ trt1⌬ cells, we next tested whether Taz1-dependent inhibition of telomere recombination might require intact telomeric heterochromatin or the telomere capping complex. We found that reintroduction of Taz1 induced chromosome circularization in taz1⌬ trt1⌬ swi6⌬, taz1⌬ trt1⌬ clr4⌬, taz1⌬ trt1⌬ poz1⌬, and taz1⌬ trt1⌬ ccq1⌬ cells (Fig. 3,  B-E). Thus, we concluded that heterochromatin formation at telomeres or the presence of an intact Pot1 complex is dispensable for Taz1-dependent inhibition of telomere recombination. In addition, because reintroduction of Taz1 was able to cause telomere fusions in all mutant backgrounds, we also concluded that Swi6, Clr4, Poz1, and Ccq1 are not required for fusion of telomeres.
In contrast, a previous study reported rapid telomere loss for taz1⌬ ccq1⌬ cells when Taz1 was eliminated from ccq1⌬ cells to generate taz1⌬ ccq1⌬ cells (20). In our hands we observed that taz1⌬ ccq1⌬ cells generated by genetic cross of single mutant strains initially undergo a very low viability phase but quickly generate a mixture of survivors that carry either circular chromosomes or stably linear chromosomes. 3 The taz1⌬ ccq1⌬ linear chromosome survivors behaved similarly to taz1⌬ trt1⌬ ccq1⌬ cells carrying either wild-type Trt1 or Trt1-CI plasmids.
In any case, because we found that recruitment of wild-type Trt1 was greatly reduced (but not completely abolished) in taz1⌬ ccq1⌬ cells compared with taz1⌬ cells (Fig.  5C), we concluded that efficient recruitment of telomerase to telomeres, promoted by Ccq1, is important for inhibition of telomere recombination by Trt1-CI. On the other hand, because Ccq1 was previously found to be involved in the inhibition of telomere recombination (19) and taz1⌬ trt1⌬ ccq1⌬ cells showed slightly longer telomeres than taz1⌬ trt1⌬ cells (Fig. 4), the ability of Ccq1 to repress telomere recombination might also contribute to Trt1-CI-induced chromosome circularization in taz1⌬ trt1⌬ cells.

Trt1-⌬T/RT Is Unable to Inhibit Recombination-based
Telomere Maintenance in the Absence of Swi6, Clr4, and Ccq1-The C-terminal-truncated Trt1 (Trt1-⌬T/RT), which lacks both the T-motif and RT domain (Fig. 3A), can also efficiently inhibit recombination-based telomere maintenance (23). Thus, we next examined if Trt1-⌬T/RT could still induce chromosome circularization in the absence of Swi6, Clr4, Poz1, or Ccq1.  We found that loss of Swi6 completely abolished Trt1-⌬T/ RT-induced chromosome circularization (Fig. 3B). On the other hand, we found that chromosomes in ϳ30% of taz1⌬ trt1⌬ clr4⌬ cells (4 of 13 examined) became circular after extensive restreaking on YES plates ( Fig. 3C and data not shown). We are unsure why the telomere status among independent taz1⌬ trt1⌬ clr4⌬ cells is mixed after Trt1-⌬T/RT reintroduction. However, we can conclude that Trt1-⌬T/RT requires the presence of both Swi6 and Clr4 to efficiently induce chromosome circularization (Fig. 3, B-C).
Rad3 ATR Is Required for Trt1-⌬T/RT-dependent Inhibition of Telomere Recombination-We had previously assumed that Trt1-CI and Trt1-⌬T/RT would show identical genetic requirements for the efficient inhibition of telomere recombination. Thus, we had not tested how reintroduction of Trt1-⌬T/RT would affect recombination-based telomere maintenance in taz1⌬ trt1⌬ est1⌬, taz1⌬ trt1⌬ pku70⌬, or taz1⌬ trt1⌬ rad3⌬ (23). However, because we now observed that Swi6 and Clr4 are uniquely required for Trt1-⌬T/RT to efficiently inhibit telomere recombination, we next tested how loss of the telomerase regulatory subunit Est1, the NHEJ DNA repair protein Ku70, and the checkpoint kinase Rad3 ATR affect the ability of Trt1-⌬T/RT to induce chromosome circularization in taz1⌬ trt1⌬ survivor cells.
Our previous analyses indicated that Est1 is required for efficient telomere recruitment of Trt1 and Trt1-dependent inhibition of telomere recombination (23). On the other hand, loss of Ku70 abolished Trt1-dependent inhibition of telomere recombination without affecting recruitment of Trt1 to telomeres. Rad3 ATR appeared not to contribute to the negative regulation of telomere recombination as Trt1-CI was still able to efficiently induce chromosome circularization in taz1⌬ trt1⌬ rad3⌬ cells (23).
Trt1-⌬T/RT Shows Reduced but Significant Ccq1-and Est1independent Telomere Association Compared with Wild-type Trt1-Quantitative ChIP analysis revealed that Trt1-⌬T/RT was less efficient in precipitating telomeric repeat DNA than wild-type Trt1, indicating that Trt1-⌬T/RT is bound less efficiently to telomeres (Fig. 5, C and D). Moreover, loss of Swi6, Clr4, or Rad3 ATR did not significantly reduce telomere association of either wild-type Trt1 or Trt1-⌬T/RT (Fig. 5, C and D). Thus, we concluded that Swi6, Clr4, and Rad3 ATR are likely to contribute to the inhibition of telomere recombination, and they are, thus, needed for the less efficient Trt1-⌬T/RT-induced chromosome circularization. However, they are dispensable for the strong inhibition of telomere recombination imposed by Taz1 or Trt1-CI.
For wild-type Trt1, we observed a significant reduction in precipitation efficiency of telomeric repeat DNA in taz1⌬ trt1⌬ ccq1⌬ and taz1⌬ trt1⌬ est1⌬ cells compared with taz1⌬ trt1⌬ cells by ChIP assays, consistent with previous studies demonstrating that Ccq1 and Est1 promote the recruitment of telomerase to telomeres (20,23,35). However, we observed a residual Trt1 association significantly above the untagged Trt1 control strain even in the absence of Ccq1 or Est1, suggesting that telomerase can be recruited to telomeres independently of Ccq1 and Est1, at least in a taz1⌬ background (Fig. 5C). The effect of expressing wild-type Trt1 on telomere length was also examined. Genomic DNA was digested with EcoRI, fractionated in a 1% agarose gel, and processed for Southern blot analysis. A probe specific to telomeric repeat DNA was used in hybridization. kb, kilobase(s); pld, plasmid.
We have concluded in our previous paper that the loss of Est1 abolishes the recruitment of Trt1 to telomeres in taz1⌬ est1⌬ cells (23). However, in our earlier experiments we used quantitative real-time PCR to detect telomerase recruitment to telomeres. The primer pairs used in those PCR are located in the sub-telomeric region immediately adjacent to the telomeric repeat sequences. Thus, although PCR primers are only 250 -300 base pairs away from the 3Ј ends of telomeres in wild-type cells, they are several kilobases away from the 3Ј ends in taz1⌬ cells due to Trt1-dependent telomere elongation. We had partially corrected for the fact that PCR primers are farther away from telomere ends by reducing the number of sonication cycles, but this probably led to an underestimation of telomere-bound Trt1. In the current study, we hybridized a telomeric repeat DNA probe to ChIP samples spotted on a nylon membrane to measure Trt1 recruitment to telomeres and, thus, improved the sensitivity of Trt1 detection in taz1⌬ cells. Therefore, although both real-time PCR and hybridization methods clearly indicate that Est1 is crucial for the efficient recruitment of Trt1 even in the absence of Taz1, we can now detect Est1independent recruitment of Trt1 significantly above the untagged background control.
Because the efficiency for telomeric repeat DNA precipitation by wild-type Trt1 was comparable among taz1⌬ trt1⌬ ccq1⌬, taz1⌬ trt1⌬ est1⌬, and taz1⌬ trt1⌬ ccq1⌬ est1⌬ cells (Fig. 6C), we could exclude the possibility that Ccq1 and Est1 were examined by anti-Myc Western blots and found to be comparable among the various mutant backgrounds. Western blots with anti-Cdc2 antibody were used as loading controls. C, recruitment of wild-type Trt1 to telomeres was monitored by quantitative ChIP assays. Percent precipitation of input DNA was determined for each ChIP sample based on quantification by dot blot hybridization with a telomeric repeat DNA probe (representative dot blots are shown below). Average % precipitation values from at least three independent experiments are plotted, and error bars represent S.D. For all strains tested, Trt1-myc showed statistically significant enrichment of telomeric DNA over no tag control (p ϭ 0.003 for taz1⌬ trt1⌬ ccq1⌬, p ϭ 0.006 for taz1⌬ trt1⌬ est1⌬, and p Ͻ 0.0002 for other triple mutant strains). Compared with taz1⌬ trt1⌬, only taz1⌬ trt1⌬ ccq1⌬ (p ϭ 0.0001) and taz1⌬ trt1⌬ est1⌬ (p ϭ 0.0007) showed statistically significant reductions in Trt1 recruitment. D, recruitment of Trt1-⌬T/RT to telomeres was monitored by quantitative ChIP assays with dot blot hybridization using a telomeric repeat DNA probe (representative dot blots are shown below). Average % precipitation values from at least three independent experiments are plotted, and error bars represent S.D. For all strains tested, Trt1-⌬T/RT-myc showed statistically significant enrichment of telomeric DNA over no tag control (p Ͻ 0.016). Compared with taz1⌬ trt1⌬, none of the triple mutant strains showed statistically significant changes in % precipitation values (p Ͼ 0.14).
are redundantly required for recruitment of telomerase to telomeres. In addition, our results implicate the existence of a telomerase recruitment mechanism that is independent of Ccq1 and Est1 in fission yeast.
We were surprised to find that Trt1-⌬T/RT association with telomeres was not significantly reduced by loss of Ccq1 and/or Est1 (Figs. 5D and 6D). Because we could not detect significant changes in the recruitment efficiency of Trt1-⌬T/RT to telomeres by deleting Ccq1 or Est1, loss of Trt1-⌬T/RT-induced chromosome circularization could not be solely explained by roles of Ccq1 and Est1 in promoting the efficient recruitment of Trt1 to telomeres. We have observed that taz1⌬ trt1⌬ est1⌬ (23) and taz1⌬ trt1⌬ ccq1⌬ (Fig. 4) cells maintain a slightly longer average telomere length than taz1⌬ trt1⌬ cells. Moreover, Ccq1 has been shown to be important for preventing telomere recombination (19,20). Thus, our results are consistent with the notion that Est1 and Ccq1 also contribute to inhibi-tion of recombination-based survival in fission yeast, independently of Taz1 and Trt1. If this hypothesis is indeed true, one might be able to obtain separation of function mutants for Est1 and Ccq1, which fail to support telomerase recruitment but can still contribute to inhibition of telomere recombination. Such results would then provide independent experimental support for the existence of Trt1-independent roles for Est1 and Ccq1 in inhibition of telomere recombination.
Rad3 ATR Is Involved in Inhibition of Rhp51 Rad51 Accumulation at Telomeres in taz1⌬ trt1⌬ Cells-Because our results suggest that Swi6, Clr4, and Rad3 ATR are uniquely required for Trt1-⌬T/RT-induced chromosome circularization by contributing to the inhibition of telomere recombination, we next decided to test if loss of Swi6, Clr4, and Rad3 ATR would cause an increase in the recruitment of DNA repair factors involved in telomere recombination. Because we have previously established that the HR protein Rad22 (Rad52 ortholog) is essential for the maintenance of linear chromosomes in taz1⌬ trt1⌬ cells (23), we initially wanted to test if the recruitment of Rad22 Rad52 to telomeres is increased when Swi6, Clr4, or Rad3 ATR are eliminated in taz1⌬ trt1⌬ cells by ChIP assays. Unfortunately, when we introduced the Myc-tagged Rad22 Rad52 , which we had previously deemed largely functional based on DNA damage sensitivity (39), into taz1⌬ trt1⌬ cells, we discovered that the resulting cells were unable to maintain stable linear chromosomes (data not shown). Therefore, we turned to another HR repair protein Rhp51 (Rad51 ortholog), because we can utilize an antibody raised against mammalian Rad51 to monitor the recruitment of Rhp51 to DNA by ChIP (39,42).
We first established that Rhp51 Rad51 is essential for linear chromosome maintenance in taz1⌬ trt1⌬ cells by creating several independently derived taz1⌬ trt1⌬ rhp51⌬ strains, extensively restreaking them on YES plates, and then examining their telomere structure by PFGE (Fig. 7A). We also established that the recruitment of Rhp51 Rad51 to telomeres is significantly enhanced in taz1⌬ trt1⌬ cells compared with wild-type cells (Fig. 7B).
We then asked if the elimination of Swi6, Clr4, or Rad3 ATR would increase Rhp51 Rad51 recruitment to telomeres in taz1⌬ trt1⌬ cells. Additionally, we examined if the elimination of FIGURE 6. Wild-type Trt1 and Trt1-⌬T/RT can be recruited to telomeres in the absence of Taz1, Est1, and Ccq1. A and B, expression levels of wild-type Trt1 (A) or Trt1-⌬T/RT (B) were examined by anti-Myc Western blots and found to be comparable among the various mutant backgrounds. Western blots with anti-Cdc2 antibody were used as loading controls. C and D, recruitment of wild-type Trt1 (C) or Trt1-⌬T/RT (D) to telomeres was monitored by quantitative ChIP assays with dot blot hybridization using a telomeric repeat DNA probe (representative dot blots are shown below). Average % precipitation values from at least four independent experiments are plotted, and error bars represent S.D. For all strains tested, Trt1-myc and Trt1⌬T/RT-myc showed statistically significant enrichment of telomeric DNA over no tag control (p Ͻ 0.006). E, telomere status analysis by PFGE indicates that taz1⌬ trt1⌬ est1⌬ ccq1⌬ cells stably maintain telomeres.

DISCUSSION
In this study we took advantage of the unique ability of fission yeast to survive telomere dysfunction by circularizing their chromosomes to understand how telomere recombination and telomerase recruitment is regulated. Because taz1⌬ trt1⌬ cells are healthy and maintain stable telomeres by utilizing a recombination-based mechanism, mutations that lead to chromosome circularization in taz1⌬ trt1⌬ cells might identify positive regulators of telomere recombination. Conversely, because we had previously established that Taz1, Trt1-CI, and Trt1-⌬T/RT strongly inhibit telomere recombination, mutations that suppress chromosome circularization upon reintroduction of Taz1, Trt1-CI, or Trt1-⌬T/RT into taz1⌬ trt1⌬ cells may identify factors that contribute to the inhibition of telomere recombination. In addition, a subset of mutations that can suppress chromosome circularization induced by Trt1-CI or Trt1-⌬T/RT may identify factors involved in the recruitment of telomerase to telomeres.
We examined how loss of proteins involved in the formation of sub-telomeric heterochromatin and loss of two Pot1 telomere-capping complex subunits affect telomere recombination and/or telomerase recruitment to telomeres. We also reexamined the roles of the checkpoint kinase Rad3 ATR and the telomerase regulatory subunit Est1 in inhibition of telomere recombination. Our results imply roles for Swi6, Clr4, Rad3 ATR , Ccq1, and Est1 in the protection against telomere recombination, which we propose is crucial for the induction of chromosome circularization in taz1⌬ trt1⌬ cells upon reintroduction of Trt1-⌬T/RT. However, it should be noted that contributions made by Swi6, Clr4, and Rad3 ATR are minor compared with major inhibitors of telomere recombination, such as Taz1, Trt1, and Ku70.
We also established that Ccq1 and Est1 are both important for the efficient recruitment of Trt1 to telomeres even in the absence of Taz1, the negative regulator of telomerase. However, because we could still detect a reduced but significant recruitment of Trt1 even in the absence of both Ccq1 and Est1, we propose the existence of a telomerase recruitment mechanism that could function in the absence of Taz1, Ccq1, and Est1.
We have reported previously that taz1⌬ est1⌬ cells cannot maintain stable linear telomeres and circularize chromosomes when they were generated by deleting est1 ϩ from taz1⌬ cells (23). In contrast, when est1 ϩ was deleted from taz1⌬ trt1⌬ cells, the resulting triple mutant cells were able to stably maintain telomeres (23). Based on these observations, we speculated that Trt1 might be able to contribute to the inhibition of recombination-based telomere maintenance in the absence of Est1. That we can now detect residual Trt1 recruitment to telomeres even in taz1⌬ est1⌬ background further supports the notion that Trt1 indeed could have Est1-independent roles at telomeres.
Based on the results presented in the current and previous papers, we can begin to establish a hierarchical order of the numerous proteins involved in telomere maintenance both for the inhibition of telomere recombination (Fig. 8A) and for the telomerase-dependent telomere extension (Fig. 8B). We do not imply the existence of linear pathways for the various factors indicated in Fig. 8, but their placement is meant to summarize the genetic requirements for either the inhibition or the promotion of telomere recombination and telomerase-dependent telomere elongation.
For inhibition of telomere recombination (Fig. 8A), the inhibitory sign directly points from Taz1 to a "telomere recombination" box, as we have yet to find the mutation(s) that can abolish the Taz1-induced chromosome circularization. On the other hand, for telomerase-dependent telomere extension (Fig.  8B), Rap1 and Poz1 are placed below Taz1, as Rap1 and Poz1 are required for the Taz1-dependent inhibition of telomere extension by telomerase (19,28).
Similarly, factors placed below Trt1 or Trt1-⌬T/RT in Fig.  8A are those we have shown to be required for chromosome circularization induced by Trt1-CI or Trt1-⌬T/RT, respectively. For Fig. 8B, Est1 and Ccq1 are placed below Trt1 as they are both essential for telomerase-dependent telomere maintenance (16,19,20). Rad3 ATR and Mre11⅐Rad50⅐Nbs1-Tel1 are placed in branched arrows, as they are redundantly required for FIGURE 7. Recruitment of Rad51 to telomeres monitored by quantitative ChIP assays. A, telomere status analysis by PFGE indicates that Rhp51 Rad51 is essential for telomere maintenance in taz1⌬ trt1⌬ cells. B, recruitment of Rhp51 Rad51 to telomeres was monitored by quantitative ChIP assays using real-time PCR. Average % precipitation values from at least four independent experiments are plotted, and error bars represent S.D. Comparable levels of Rhp51 were expressed among the different mutant strains based on anti-Rad51 Western blots. Western blots with anti-Cdc2 antibody were used as loading controls. wt, wild type.
It is particularly intriguing that Taz1, Rap1, and Poz1 show quite different phenotypes with regard to the regulation of recombination at telomeres despite the fact that deletions of these three factors result in massive telomerase-dependent telomere elongation and loss of transcriptional repression for markers inserted adjacent to the telomeric repeats (28,35,44). In the case of Poz1, we did not find any evidence that this protein plays a role in the repression of telomere recombination. In addition, the fact that taz1⌬ trt1⌬ poz1⌬ cells can stably maintain telomeres indicates that Poz1 is not involved in the promotion of recombination-based telomere maintenance. In contrast, we have previously demonstrated that Rap1 is essential for recombination-based telomere maintenance in taz1⌬ trt1⌬ cells (23). We were surprised to find that Rap1 can contribute to recombination-based survival in the absence of Taz1, as efficient recruitment of Rap1 to telomeres has been shown to depend on Taz1 in fission yeast (23,28,45) (see also Fig. 1). However, the recent finding that Rap1 directly interacts with Poz1 (19) raised the possibility that a Poz1-dependent recruitment of Rap1 to telomeres might be critical for the Rap1 role in the promotion of recombination-based telomere maintenance. However, because taz1⌬ trt1⌬ poz1⌬ cells can stably maintain telomeres, our results indicate that the positive contribution of Rap1 in recombination-based telomere maintenance is independent of both Taz1 and Poz1.
By analyzing the spontaneous rearrangement of sub-telomeric regions, the Ku70⅐Ku80 complex and the formation of sub-telomeric heterochromatin have previously been implicated in the repression of recombination at telomeres (33,42,46). Our chromosome circularization assays using Taz1, Trt1-CI, and Trt1-⌬T/RT have enabled us to detect the contribution of these factors in the regulation of telomere recombination. In addition, the use of Trt1-⌬T/RT in the chromosome circularization assay has allowed us to identify a previously unknown contribution of Rad3 ATR in the repression of telomere recombination. Factors that affect the recruitment efficiency of telomerase to telomeres, such as Est1 and Ccq1, have also been identified as necessary for the suppression of chromosome circularization by Trt1-CI.
Therefore, the chromosome circularization assay we have developed should be useful in identifying new factors that contribute to the regulation of telomere recombination and telomerase-dependent telomere extension. Because proteins involved in telomere regulation are well conserved between fission yeast and mammalian cells, careful evaluation of factors that affect telomere recombination and telomerase recruitment in fission yeast may provide new insight into the regulation of mammalian telomere recombination and telomerase recruitment.