ABSTRACT
In response to perturbed DNA replication, ATR (ataxia telangiectasia and Rad3-related) kinase is activated to initiate the checkpoint signaling necessary for maintaining genome integrity and cell survival. To better understand the signaling mechanism, we carried out a large-scale genetic screen in fission yeast looking for mutants with enhanced sensitivity to hydroxyurea. From a collection of ∼370 primary mutants, we found a few mutants in which Rad3 (ATR ortholog)-mediated phospho-signaling was significantly compromised. One such mutant carried an uncharacterized mutation in tel2, a gene encoding an essential and highly conserved eukaryotic protein. Previous studies in various biological models have shown that Tel2 mainly functions in Tel2-Tti1-Tti2 (TTT) complex that regulates the steady-state levels of all phosphatidylinositol 3-kinase-like protein kinases, including ATR. We show here that although the levels of Rad3 and Rad3-mediated phospho-signaling in DNA damage checkpoint were moderately reduced in the tel2 mutant, the phospho-signaling in the DNA replication checkpoint was almost completely eliminated. In addition, the tel2 mutation caused telomere shortening. Since the interactions of Tel2 with Tti1 and Tti2 were significantly weakened by the mutation, destabilization of the TTT complex likely contributes to the observed checkpoint and telomere defects.
INTRODUCTION
DNA replication can be perturbed by various endogenous and exogenous factors. If undetected, perturbed replication forks collapse, causing chromosomal DNA damage or even cell death. To maintain the genome integrity, eukaryotes have evolved a surveillance mechanism called the DNA replication checkpoint (DRC) to monitor fork progression during normal S phase or under stress (see reviews in references 1 and 2). The DRC senses the problems and activates cellular responses such as increased production of deoxynucleoside triphosphates (dNTPs), cell cycle delay, fork stabilization, and suppression of late firing origins, which all work in concert to minimize the mutation rate and to ensure accurate duplication of the genome. Consistent with its importance in genome integrity, the DRC is highly conserved from yeasts to humans, and defects in the pathway cause a wide range of developmental and cancer predisposition syndromes. Although debatable, mutations generated by errors in DNA replication followed by mistakes in repair likely contribute significantly to sporadic cancers (3).
Studies in the past decades have identified a related set of sensor proteins in all eukaryotes that assemble at perturbed forks for the DRC signaling. Among the sensors, ATR (ataxia telangiectasia and Rad3-related) kinase works with its cofactor ATRIP (ATR-interacting protein) and the 9-1-1 (Rad9-Rad1-Hus1) complex to initiate the signaling by phosphorylating various substrates, including the checkpoint mediators and effector kinase (4–6). The activated mediators channel the signal to the effector kinase (7, 8). Once activated, the effector kinase diffuses away from the fork and relays the signal to various cellular structures to stimulate the responses mentioned above. Despite its importance in genome integrity and extensive studies in the past, the DRC signaling mechanism, however, remains incompletely understood (9–11).
To better understand the mechanism by which the checkpoint signaling is initiated at perturbed forks (9, 10), we have searched for new DRC mutants in the fission yeast Schizosaccharomyces pombe, an established model for studying the cellular mechanisms that are conserved in humans. By random mutation of the genome, a large-scale genetic screen has been carried out to look for mutants with enhanced sensitivity to hydroxyurea (HU). HU has been of clinical and scientific interest for ≥100 years (12, 13). It perturbs DNA replication by inhibiting ribonucleotide reductase (RNR), a highly conserved enzyme required to provide dNTPs for DNA replication and repair (14). RNR contains a catalytically essential diferric tyrosyl radical inside its smaller subunit. HU quenches the tyrosyl radical and thus suppresses RNR, which slows down polymerase movement at the forks (15, 16). Consistent with this mechanism, HU resistance has been observed in cells overexpressing RNR or expressing a mutant RNR (17–19). DNA replication can also be perturbed by DNA damage such as those caused by methyl methanesulfonate (MMS) or UV light. Unlike HU, which slows forks globally, DNA damage pauses a subset of on-going forks at the damage sites on the leading strand template (20). In addition, DNA damage, if occurs outside S phase, provokes the DNA damage checkpoint (DDC) responses (Fig. 1A). Therefore, HU, if properly used, specifically induces replication stress, and studies in various eukaryotic organisms have shown that the primary response to HU treatment is the activation of DRC (2, 21, 22).
Sensitivity of the newly identified hus227 mutant to HU and DNA damage. (A) Schematics of Rad3 kinase signaling in the DRC (left) and the DDC (right) pathways of S. pombe. Rad3-specific phosphorylation sites are indicated. When DNA replication is perturbed, Rad3 phosphorylates Mrc1 and Cds1 to activate the DRC (8, 11, 44). When DNA damage occurs outside the S phase or the forks collapse, such as that in HU-treated mrc1 or cds1 cells, Rad3 phosphorylates Crb253BP1/Rad9 and Chk1 to stimulate the DDC responses (46, 48, 49). Phosphorylation of Rad9 in the 9-1-1 complex is required for activation of both the DRC and the DDC (11, 45). While phosphorylated Rad9 recruits Rad4TopBP1/Dpb11 to promote Crb2 and Chk1 activation, it remains unclear how phosphorylated Rad9 promotes Cds1 activation (dash line). (B) Sensitivities of hus227 mutant (YJ1496) to HU and MMS and UV and bleomycin (BLM) were examined by using a standard spot assay. Logarithmically growing cells were diluted in 5-fold steps and spotted onto YE6S plates as the control and for UV treatment or YE6S plates containing the indicated drugs. The plates were incubated at 30˚C for 3 days and then photographed. Wild type (TK48) and rad3 (NR1826), mrc1 (YJ15), cds1 (GBY191), and chk1 (TK197) checkpoint mutants were used as controls. The experiment was repeated three times, and a representative result is shown. (C) The ts phenotype of hus227 was examined by using a spot assay. (D) Sensitivity of hus227 to acute HU treatment. The strains used in panel B were incubated in YE6S medium containing 15 mM HU. At the indicated time points, the cells were spread onto YE6S plates to recover for 3 days. Colonies were counted, and the results are presented as percentages. Error bars represent means and standard errors of the mean (SEM) of triplicates (see Table S4, sheet 1, in the supplemental material). (E) Overexpression of the RNR small subunit Suc22 rescued hus227. Suc22 was expressed in S. pombe on a vector under the control of its own promoter. The metabolic erg11-1 mutant was used as a negative control (38). V, vector alone.
We report here our identification of a previously uncharacterized mutation in tel2 by the hus (HU sensitive) screen that significantly sensitizes S. pombe to HU and the DNA damage agents MMS, UV, and bleomycin. Tel2 is an essential and a highly conserved protein among eukaryotes (23, 24). It was originally identified in Caenorhabditis elegans and Saccharomyces cerevisiae ≥30 years ago (25–29). The current model suggests that Tel2 functions in the TTT (Tel2-Tti1-Tti2) complex as a cochaperone that regulates the protein levels of all phosphatidylinositol 3-kinase-like protein kinases (PIKKs), including ATR kinase, and hence multiple cellular processes (30–33). We found that the level of Rad3ATR was moderately reduced in this mutant similar to that in S. cerevisiae tel2-1 mutant (34, 35). Consistent with the reduced Rad3 level, the signaling in the DDC pathway was moderately compromised. Surprisingly, the mutation almost completely eliminated the signaling in the DRC pathway. Interestingly, unlike this S. pombe tel2 mutant and the two C. elegans mutants that are sensitive to replication stress (29), the S. cerevisiae tel2-1 mutant is insensitive to HU and DNA damage (34). Similar to the S. cerevisiae tel2-1 mutant (25), the mutation also caused telomere shortening in S. pombe. Because the mutation significantly weakened the interactions of Tel2 with Tti1 and Tti2, it is likely that the destabilized TTT complex causes the defects in checkpoint signaling and telomere maintenance.
RESULTS
Screening of a fission yeast mutant hus227 with enhanced sensitivities to HU and DNA damage.We have carried out a large-scale hus screen in S. pombe looking for new DRC mutants and accumulated ∼370 primary mutants. The mutants were backcrossed three times to remove bystander mutations. After removing known mutations by crossing with DRC mutants and other hus mutants, a small set of new hus mutants was screened that likely causes DRC defects. One such mutant is hus227. Preliminary results suggested that phosphorylation of the DRC mediator protein Mrc1Claspin by Rad3 (8, 11) was eliminated. We therefore decided to investigate further on hus227.
We first examined the sensitivities of hus227 to HU and DNA damage by using standard spot assay. Rad3 is the master checkpoint kinase responsible for activation of both DRC and DDC in S. pombe (36) (Fig. 1A). The ATM (ataxia telangiectasia mutated) ortholog Tel1 contributes minimally to checkpoint in fission yeast. Deletion of rad3 sensitizes S. pombe to both HU and DNA damage due to a lack of checkpoint functions. As shown in the top panels of Fig. 1B, while S. pombe cells lacking Rad3, Mrc1, or Cds1 were highly sensitive to HU, the cells lacking Chk1, the effector kinase of the DDC, were less sensitive, suggesting that the replication stress induced by HU is mainly dealt by the DRC. Under similar conditions, the hus227 mutant was found to be sensitive to HU, and the sensitivity was higher than those of mrc1 and cds1 mutants but less than that of the rad3 mutant. We then examined the sensitivity of the hus227 mutant to DNA damage caused by MMS and UV (Fig. 1B, middle panels). Unlike the HU treatment, the chk1 mutant was more sensitive to MMS and UV than the cds1 mutant, indicating that the DNA damage that occurs at G2, the major cell cycle stage in S. pombe, is mainly dealt by the DDC. The hus227 mutant was also sensitive to MMS and UV, and the sensitivity was higher than that of the chk1 mutant but less than that of the rad3 mutant. We then examined the sensitivity of the hus227 mutant to bleomycin (Fig. 1B, bottom panels), an antibiotic that generates single-strand and double-strand breaks in chromosomal DNA (37). Interestingly, the hus227 mutant was highly sensitive to bleomycin and the sensitivity was even higher than that of the rad3 mutant that lacks both DDC and DRC. This suggests that the mutation may also affect DNA repair (see below). We also found that the hus227 mutant was unable to grow at 37°C (Fig. 1C), showing that it is a ts mutant and that the mutated gene is likely essential or required for cell growth at 37°C.
We have recently screened a new set of hus mutants with remarkable HU sensitivities that are caused by mutations in various metabolic pathways and not the DRC (38, 39). These mutants are much more sensitive to chronic exposure than to acute HU treatment, and they cannot be rescued by upregulation of Suc22, the small subunit of RNR and the major regulatory target of the DRC in fission yeast. We then examined the sensitivity of the hus227 mutant to acute HU treatment and found that the hus227 mutant was sensitive and that the sensitivity was comparable to that of the mrc1 mutant but less than that of the rad3 mutant (Fig. 1D). We next examined the rescuing effect of Suc22 on the hus227 mutant. Unlike the metabolic erg11-1 mutant in ergosterol biosynthesis that could not be rescued (38), upregulation of Suc22 rescued the hus227 mutant similar to the rad3 mutant (Fig. 1E). We then performed tetrad dissection on crosses of the hus227 mutant with all known DRC mutants, as well as S. pombe lacking the nuclear membrane protein Lem2 (40). All crosses generated spores with wild-type HU resistance (see Fig. S1 in the supplemental material). This suggests that the hus227 mutant is not allelic to any of the tested mutants and is likely a new DRC mutant in S. pombe.
hus227 carries a single missense mutation in tel2.To identify the mutated gene, we transformed hus227 with S. pombe genomic DNA expression libraries carrying an ura4 marker. Colonies grown on plates lacking uracil were replicated onto plates containing HU to screen those with conferred HU resistance. Plasmids recovered from the yeast colonies were subjected to restriction enzyme digestions and subsequent DNA sequencing, which identified tel2. Expression of wild-type tel2 on a vector fully rescued hus227 (Fig. 2A). Under the same conditions, mutant tel2 rescued to a much lesser degree. Sequencing of the PCR product generated by using a high-fidelity polymerase and genomic DNA purified from hus227 identified a single G-to-A mutation, causing a Cys307-to-Tyr amino acid change in Tel2 (Fig. 2B).
The hus and ts phenotypes of hus227 are caused by a single mutation in tel2. (A) Wild-type and mutant tel2 was expressed in the indicated strains on a vector under the control of its own promoter. The HU sensitivity was examined by using a standard spot assay. (B) DNA sequencing of tel2 in hus227 identified a G-to-A mutation that converts Cys307 to Tyr in Tel2. (C) Strategy for integrating the tel2 mutation at its genomic locus (upper panel). HA and nmtT represent the HA epitope and nmt1 terminator, respectively. Integrants were screened by colony PCRs (red lines) and backcrossed to ensure single copy integration in the genome. Genomic DNAs were purified from the integrants (YJ1513 and YJ1515) and wild-type cells (TK48) for PCRs to confirm the integration at tel2 locus (green line). The PCR products were analyzed by agarose gel electrophoresis (lower panel). (D) HU sensitivities of wild type (YJ1513) and mutant tel2 integrants (YJ1515) were determined by using a standard spot assay at 30°C (upper panels) or 25°C (lower panels). The same strains were also spotted on YE6S plates and incubated at 37°C for 2 days to examine the ts phenotype (upper right panel). (E) Expression on a vector increased the protein levels of Tel2. Wild-type (YJ1513) and mutant (YJ1515) tel2-HA integrants carrying a vector or the vector expressing wild-type or mutant tel2-HA were lysed by using the TCA method and analyzed by Western blotting. A section of Ponceau S-stained membrane was used as the loading control. The intensities of the Tel2 bands were quantified, and the results are shown as percentages at the bottom of the figure.
To see whether the mutation is linked to the hus and ts phenotypes of hus227, we integrated wild type and the mutant tel2 tagged with a hemagglutinin (HA) epitope at the genomic locus in wild-type S. pombe by using the method diagrammed at the top of Fig. 2C. After screening by colony PCRs and backcrosses to ensure single-copy integration in the genome, the integrants were confirmed by PCRs using Phusion polymerase (Fig. 2C, lower panel) and Western blotting with anti-HA antibody (see Fig. S2 in the supplemental material). We then examined the HU sensitivities of the integrants and found that while the integrant of wild-type tel2 was resistant to HU, the integrant of mutant tel2 was sensitive (Fig. 2D). The mutant integrant not only showed that the HU sensitivity similar to hus227, but the integration also caused a similar ts phenotype (Fig. 2D, top right panel). We also assessed the HU sensitivity at 25°C and found that both hus227 and the mutant integrant were sensitive, although to a lesser degree (Fig. 2D, lower panels). This suggests that the mutation uncouples the essential role of Tel2 in cell growth and the function in HU resistance. We then quantified the protein levels of Tel2 by Western analysis using Ponceau S staining as the loading control because it is linear with protein concentration (R2 = 0.99) (41) (Fig. 2E). The results showed that when cultured at 30°C, the level of endogenous mutant Tel2 was reduced to 80.2% ± 4.1% (n = 4) of the wild-type level. At 25°C, the level of mutant Tel2 was similar to that in wild-type cells (see Fig. S2 and S7 in the supplemental material). When expressed on a vector, both wild-type and mutant Tel2 were increased by >2-fold, which may explain the partial rescuing effect of mutant Tel2 shown in Fig. 2A. Because both hus and ts phenotypes of hus227 were caused by the single mutation in tel2, we here renamed hus227 tel2-C307Y. Throughout the rest of this study, all experiments were carried out in YE6S medium at 30°C except when the use of different temperatures is specified.
Tel2 was first identified in S. cerevisiae (24, 25) and C. elegans (26–28, 42) with a wide range of defects in genome integrity, maintenance of telomeres, life span, and development. Previous studies in various models have shown that Tel2 mainly functions, together with Hsp90 and several cofactors, in regulating protein stability of all PIKKs and hence the multiple biological processes (30–33). Although Tel2 is conserved in eukaryotes, the similarity in overall amino acid sequences is limited (Fig. S3). Similar to the mutated residues Ser129 in S. cerevisiae tel2-1 and Cys135 and Cys772 in C. elegans rad-5 and clk-2, respectively, the mutated Cys307 residue in tel2-C307Y is not highly conserved. It was therefore not a surprise that the expression of human Tel2 failed to rescue tel2-C307Y (Fig. S4). Considering that the mutation may perturb metabolism via other PIKKs and thus sensitizes S. pombe to HU, we examined the HU sensitivity of tel2-C307Y cells containing various auxotrophic markers. The result showed that auxotrophies contributed minimally to HU sensitivity (Fig. S5). By measuring the bulk phosphorylation of Mrc1 in a tel2 shutoff strain, an earlier study suggests that Tel2 may function in DRC in S. pombe (43). However, depletion of Tel2 minimally sensitized S. pombe to HU and UV and, more importantly, the cut or “cell untimely torn” cells, which occur in all HU-treated DRC mutants, were rarely observed, which raises a concern regarding the observed DRC defect (43). In addition, the DRC defect was not examined in details due to a lack of phospho-specific antibodies. Since our newly screened tel2-C307Y mutant is highly sensitive to HU and DNA damage, we decided to further investigate this mutant and its potential defect in DRC by using the set of phospho-specific antibodies that we have made available in the previous studies (11, 44).
HU arrests tel2-C307Y cells in S phase.As mentioned above, HU arrests some of the metabolic hus mutants such as erg11-1 and hem13-1 at the G2/M phase, not the S phase, since they are killed by HU via a mechanism unrelated to replication stress (38, 39). We first examined the cell cycle progression of tel2-C307Y by flow cytometry (Fig. 3A). When incubated with 15 mM HU for 1 to 3 h, wild-type cells were increasingly arrested at S phase. Further incubation did not completely stop DNA synthesis since the cells managed to continue the S-phase progression and eventually finished the bulk of DNA synthesis in about 7 or 8 h. rad3 and mrc1 cells were also arrested in S phase after 3 h of HU treatment. However, these checkpoint mutants, and the rad3 mutant in particular, could not properly synthesize DNA in HU. Under similar conditions, HU arrested the majority of tel2-C307Y at the S phase, although a small number of cells remained at G2/M after 3 to 4 h of incubation. Unlike wild-type cells, however, tel2-C307Y failed to continue DNA synthesis in HU similar to rad3 and mrc1. This result strongly suggests a DRC defect in tel2-C307Y.
DRC defect in tel2-C307Y. (A) Cell cycle analysis of tel2-C307Y. The wild-type, rad3, mrc1, and tel2-C307Y cells used in Fig. 1B were incubated with 15 mM HU and analyzed by flow cytometry every hour during the incubation. Red lines indicate 1C and 2C DNA contents. (B) tel2-C307Y cells undergo cell division in the presence of HU. The cells were incubated with HU as in panel A and fixed in 2.5% glutaraldehyde at each time point. The fixed cells were stained with Hoechst and Blankophor for microscopic examination. Next, ≥150 cells were counted for each sample, repeated three times (Table S4, sheet 2). The cells with a septum are reported as percentages. (C) tel2-C307Y cells showed the cut phenotype in HU. Cells were treated with 15 mM HU for 6 h as in panel A, fixed onto glass slides by brief heating, and stained as in panel B for microscopic examination. Arrows indicate the cut cells. (D) tel2-C307Y cells failed to recover from HU arrest. The strains used in panel A were treated with 15 mM HU for 4 h and then released in fresh medium. Cell cycle progression was monitored every hour during the release.
DRC defect in tel2-C307Y.In the presence of HU, the DRC is activated to delay mitosis so that the cells have adequate time to complete DNA synthesis before they divide. The DRC mutants, however, proceed into mitosis, generating the cut phenotype in which two daughters do not have equal amounts or lack detectable genomic DNAs. We stained the cells with Hoechst for genomic DNAs and Blankophor for the septum and examined cell septation during the course of HU treatment (Fig. 3B). After wild-type cells were treated with 15 mM HU for 3 h, cell division was almost completely suppressed and remained suppressed during the rest of HU treatment, suggesting an activated DRC. In contrast, rad3 mutant showed a robust cell division activity in HU. More than 50% of cells underwent cell division after 3 h of HU treatment and continued to divide during the rest of the treatment. As a result, most of the cells were short and showed the cut phenotype (Fig. 3C, arrows). The mrc1 mutant initially slowed down cell division during the first 1 to 3 h. After 3 to 4 h in HU, the cells began to divide, generating the cut cells (Fig. 3C). However, unlike the HU-treated rad3 cells that were short due to the lack of both DRC and DDC, the mrc1 cells elongated because the collapsed fork activates DDC (8, 40). Similar to rad3 and mrc1 mutants and consistent with the cell-killing effect shown in Fig. 1D, tel2-C307Y cells continued to divide in the presence of HU, and the septation index was higher than for mrc1 cells but lower than for rad3 cells (Fig. 3B). Furthermore, since the HU-treated tel2-C307Y cells were shorter than mrc1 cells but longer than rad3 cells, the mutation may also affect the DDC (see below).
Next, we examined cell recovery from HU (Fig. 3D). After 4 h of treatment with 15 mM HU, almost all wild-type, rad3, and mrc1 cells and the majority of the tel2-C307Y cells were arrested at S phase. When released in fresh medium, wild-type cells could fully recover and returned to normal cell cycle in ≤4 h. In contrast, rad3 and mrc1 cells could not finish DNA synthesis and failed to recover. Similarly, tel2-C307Y cells also failed to recover from the HU arrest. Together, we show that the tel2-C307Y mutation causes a profound DRC defect, which sensitizes S. pombe to the replication stress induced by HU or DNA damage.
Elimination of Rad3 phosphosignaling in the DRC.Under replication stress, Rad3 phosphorylates two functionally redundant residues Thr645 and Thr653 in Mrc1 and Thr412 in Rad9 of the 9-1-1 complex (Fig. 1A, left). Phosphorylation of Mrc1 and Rad9 facilitates the phosphorylation of Thr11 in Cds1CHK2/Rad53 by Rad3, which promotes the autophosphorylation of Cds1-Thr328 in the activation loop. Phosphorylation of Thr328 directly activates Cds1, leading to full activation of the DRC in S. pombe (8, 11, 44). When DNA damage occurs outside S phase or forks collapse causing strand breaks, such as the situations in HU-treated mrc1 or cds1 cells, Rad3 phosphorylates Rad9 and Crb253BP1/Rad9 (45–47), which in turn facilitate the phosphorylation of Chk1-Ser345 by Rad3 (48, 49), leading to the activation of DDC (Fig. 1A, right).
To investigate the DRC defect in tel2-C307Y, we examined Rad3-mediated phosphorylation of Rad9, Mrc1, and Cds1 by using phospho-specific antibodies (8, 11, 44). Under physiological conditions, Rad9 is phosphorylated at a basal level mainly by Rad3 in wild-type cells (compare wild-type cells with rad3, tel1, and rad3 tel1 mutants in Fig. 4A). After the cells were treated with 15 mM HU for 3 h, phosphorylation of Rad9 was significantly increased. In the tel2-C307Y mutant, the basal Rad9 phosphorylation remained detectable under normal conditions. After HU treatment, however, the phosphorylation was not increased similar to that in HU-treated rad3 cells. Next, we examined the Rad3-specific phosphorylation of Mrc1-Thr645, a representative of the two redundant sites in Mrc1 (Fig. 4B) (8). After HU treatment, Mrc1 phosphorylation was significantly increased in wild-type cells. Under similar conditions, Mrc1 phosphorylation was reduced to an undetectable level in tel2-C307Y cells. Similar to rad3 cells, the protein level of Mrc1 in HU-treated tel2-C307Y cells was lower than in wild-type cells, which is consistent with the lack of DRC because DRC upregulates Mrc1 (50). We also examined Mrc1 phosphorylation after the cells were treated with 0.01% MMS for 90 min. Similar results were observed (Fig. 4C). Finally, we examined Rad3-dependent phosphorylation of Cds1 and found that the phosphorylation was almost completely eliminated in HU (Fig. 4D) or MMS (Fig. 4E). Together, these results show clearly that the tel2-C307Y mutation almost completely abolished Rad3 kinase signaling in DRC.
Elimination of Rad3 kinase signaling in the DRC pathway. (A) HU-induced phosphorylation of Rad9 by Rad3 was eliminated in tel2-C307Y. Wild-type and mutant S. pombe cells were treated (+) or not treated (–) with 15 mM HU for 3 h. Rad9-HA was immunoprecipitated and separated by SDS-PAGE for Western blotting with anti-HA antibody (lower panel). The same blot was stripped and reprobed with the phospho-specific antibody (upper panel). The Rad9 phosphorylation bands were quantified, and the band intensities in comparison to HU-treated wild-type cells are shown at the bottom of the panel. (B) HU-induced phosphorylation of Mrc1 by Rad3 was reduced to an undetectable level in tel2-C307Y cells. Wild-type and mutant cells used in Fig. 1B were treated with 15 mM HU for 3 h. Phosphorylation of Mrc1 (top panel) was detected with phospho-specific antibodies in whole-cell lysates prepared by the TCA method. The same blot was stripped and reprobed with anti-Mrc1 antibodies (middle panel). A section of Ponceau S-stained membrane is shown (bottom panel). (C) MMS-induced phosphorylation of Mrc1 was also reduced to an undetectable level in tel2-C307Y. The cells were treated with 0.01% MMS for 90 min and then lysed for Western blotting as in panel B. (D) HU-induced phosphorylation of Cds1 by Rad3 was eliminated in tel2-C307Y. Wild-type, rad3, and tel2-C307Y cells were treated (+) or not treated (–) with HU. Cds1-HA was immunoprecipitated and then analyzed by Western blotting with anti-HA antibody (bottom panel). The same membrane was stripped and reprobed with the phospho-specific antibody (top two panels). The Cds1 phosphorylation bands were quantified, and the results are shown at the bottom. (E) MMS-induced phosphorylation of Cds1 was also eliminated in tel2-C307Y cells. The cells were treated with 0.01% MMS for 90 min and analyzed as in panel D.
Reduced Rad3 phosphosignaling in DDC.Next, we examined the Rad3-dependent phosphorylation of Rad9 and Chk1 in the DDC pathway (Fig. 1A, right). Similar to the HU treatment (Fig. 4A), treatment with 0.01% MMS for 90 min significantly increased Rad9 phosphorylation by Rad3 in wild-type cells (Fig. 5A). However, Rad9 phosphorylation was not readily increased in MMS-treated tel2-C307Y cells. We then examined the phosphorylation of Chk1-Ser345 using the phospho-specific antibody developed by the Walworth lab (48) (Fig. 5B). To our surprise, unlike the Cds1 phosphorylation shown in Fig. 4D and E, ≥50% of Chk1 phosphorylation remained detectable in MMS-treated tel2-C307Y cells. To confirm this result, we examined Chk1 phosphorylation in the presence of increasing concentrations of MMS (Fig. 5C). Quantitation results clearly showed that although reduced, the majority of the Chk1 phosphorylation remained in tel2-C307Y cells (Fig. 5D). Since Tel2 functions in multiple cellular processes, it is possible that other kinases participate in Chk1 phosphorylation in tel2-C307Y cells. We then examined the phosphorylation in tel2-C307Y cells lacking either Rad3 or Tel1 (Fig. 5E). The result showed clearly that the major kinase responsible for Chk1 phosphorylation in tel2-C307Y cells was Rad3, which excluded this possibility. Together, these results show that although Rad3 signaling was almost completely eliminated in the DRC, it remained partially functional in the DDC. This result is consistent with the observation in Fig. 3C in which the HU-treated tel2-C307Y cells were slightly longer than rad3.
Moderate reduction of Rad3 kinase signaling in DDC. (A) MMS-induced phosphorylation of Rad9 by Rad3 was significantly reduced in tel2-C307Y cells. The cells were treated with 0.01% MMS for 90 min, analyzed, and quantified as in Fig. 4A. (B) MMS-induced phosphorylation of Chk1 by Rad3 was only moderately reduced in tel2-C307Y cells. Wild type, rad3, and tel2-C307Y cells were treated with MMS as in panel A. Chk1-HA was immunoprecipitated from an equal number of cells for Western blotting with anti-HA antibody (lower panel). The same blot was stripped and reprobed with the phospho-specific antibody (48) for phosphorylated Chk1 (upper panel). The Chk1 phosphorylation bands were quantified, and the results are shown at the bottom. (C) Chk1 phosphorylation was examined and quantified as in panel B after the cells were treated with increasing concentrations of MMS for 90 min. (D) Intensities of the Chk1 phosphorylation bands in panel C were quantified, and the results are shown as percentages. (E) MMS-induced phosphorylation of Chk1 in tel2-C307Y was dependent on Rad3. Chk1 phosphorylation was examined and quantified as in panel B in wild-type, tel2-C307Y, or tel2-C307Y cells lacking rad3 or tel1.
Moderate reduction of Rad3 and Tel1 in tel2-C307Y cells.Studies in multiple organisms have shown that Tel2 regulates the stability of all PIKKs, including ATR and ATM (30, 33–35, 51). We next measured the levels of Rad3 and Tel1 in S. pombe. Although not vigorously examined, the numbers of Rad3 and Tel1 molecules in an S. pombe cell are much lower than in S. cerevisiae and mammalian cells (52). We tagged the Rad3 N terminus with the myc epitope at its genomic locus. The tagging does not affect the function of Rad3 since the tagged strain showed a wild-type HU resistance (Fig. 6A). To ensure accurate quantification, three separate samples were analyzed by using Ponceau S staining as the loading controls (41). We found that Rad3 in tel2-C307Y cells was reduced to 60.1% ± 1.3% (n = 3) of the wild-type level (Fig. S6). To see whether the kinase activity of Rad3 contributes to its instability in tel2-C307Y, we examined the kinase-inactive rad3-D2249E cells (36) and found that the kinase activity was unrelated to the instability (Fig. 6C). Since Rad3 complexes with Rad26ATRIP/Ddc2, we also examined Rad26 and found that in contrast to Rad3, the level of Rad26 was significantly increased by >2-fold in tel2-C307Y cells (Fig. 6D; see also Fig. S6 in the supplemental material). This shows that Tel2 specifically regulates the stability of Rad3, not its cofactor Rad26, which is different from human cells in which depletion of Tel2 decreases the levels of both ATR and ATRIP (53).
Despite showing only moderate reduction in Rad3 and Tel1 protein levels, tel2-C307Y cells show a telomere length maintenance defect. (A) Inactivation of Rad3 caused synergistic growth defect in tel2-C307Y. Wild type (YJ1138) and kinase-inactive (kd) rad3-D2249E (YJ1139) were tagged with myc epitope at the genomic locus. After crossing with wild-type and tel2-C307Y strains, colonies were selected for examining cell growth under normal conditions or in the presence of HU by using a standard spot assay. (B) Western analysis of Rad3 levels in colonies containing wild-type or mutant tel2. Cell lysates were prepared by using the TCA method for Western blotting. Rad3 was detected by anti-myc antibody and quantified, and the results are shown as percentages at the bottom of this panel and in Fig. S6 in the supplemental material. The asterisk indicates a cross-reactive material. Ponceau S staining was used as the loading control. (C) Reduction of Rad3 level in tel2-C307Y was unrelated to the kinase activity of Rad3. rad3-D2249E was tagged, crossed into tel2-C307Y cells, and analyzed as in panel B. (D) The Rad26 level increased in tel2-C307Y. Rad26 was similarly tagged and analyzed as Rad3 in panel B. (E) Tel1 was tagged with myc epitope at the genomic locus in the wild type (LS8284) or tel2-C307Y mutant (YJ1541). Equal numbers of the cells were collected in triplicates, lysed, and immunoprecipitated. An untagged strain was used as the control. The relative intensities of the Tel1 bands are shown on the bottom and in Fig. S6 in the supplemental material. (F) Shortened telomeres in tel2-C307Y mutant. Genomic DNAs were purified from wild-type (TN3783 and TN3784) and tel2-C307Y (YJ1495 and YJ1496) cells, digested with EcoRI, and processed for Southern blotting with a telomere probe (54). A representative gel from three independent experiments is shown. (G) Southern blots were quantified with ImageQuant (Fig. S8). The telomere lengths in the indicated strains were calculated by comparing with the standard markers (Table S4, sheet 4). The means and SEM in each column are derived from five or six samples. ***, P < 0.001. (H) ChIP analysis to monitor telomere association of Rad26 and Tel2. Real-time PCR was used to quantify immunoprecipitated telomeric DNA relative to input DNA samples. The means and SEM are calculated from seven and five independent experiments for Rad26 and Tel2, respectively (Table S4, sheets 5 and 6). **, P < 0.01; ***, P < 0.001.
We then tagged and examined Tel1, the ATM ortholog in S. pombe. Because our preliminary data suggested that the number of Tel1 per cell was even lower than Rad3 and could not be reliably detected by Western blotting in whole-cell extracts, we examined Tel1 after enrichment by immunoprecipitation (IP). Multiple samples of equal number of cells were collected, lysed, and immunoprecipitated for Western blotting (Fig. 6E). The results showed that similar to Rad3, Tel1 was reduced to 38% ± 33.7% (n = 5) of its wild-type level in tel2-C307Y (see Fig. S6 in the supplemental material). Together, these data showed that both Rad3 and Tel1 were moderately reduced in tel2-C307Y cells. Whether the mutation affects the protein stability of other PIKKs, although likely, remains to be investigated. Interestingly, the double mutant containing tel2-C307Y and kinase-inactive rad3 grew significantly more slowly than the single mutants under normal conditions (Fig. 6A). This indicates that the remaining Rad3 in tel2-C307Y cells is active that plays an important role in promoting the cell growth (see below).
Shortened telomeres in tel2-C307Y cells.Since Rad3 and Tel1 play redundant but essential roles in recruitment of telomerase to telomeres in S. pombe (54, 55), we expected that tel2-C307Y cells might carry shorter telomeres. In S. cerevisiae, Tel2 directly binds to telomeric DNA and regulates telomere length (25, 56). In the two C. elegans tel2 mutants, reports showed clear changes or no change in telomere length (29, 42, 57). Overexpression of Tel2 in human cells gradually lengthens telomeres (58). Although acute depletion of Tel2 in S. pombe does not alter telomere length (43), the tel2-C307Y mutation may make an accumulative effect on telomere through generations. To investigate, we examined telomere length in tel2-C307Y cells after restreaking on plates five times at either 25 or 32°C. Genomic DNAs were prepared, digested with EcoRI, and processed for Southern blotting (Fig. 6F). We found that tel2-C307Y cells showed significant telomere shortening for both temperatures. Quantification results from three independent experiments showed that at 25°C, the average telomere length in tel2-C307Y was 290 bp (n = 5), which was significantly shorter than that in wild-type cells (385 bp, n = 6) (Fig. 6G). At 32°C, the length further shortened to 156 bp (n = 6) in tel2-C307Y compared to the wild-type length of 317 bp (n = 6).
Consistent with the previous in vitro data of S. cerevisiae Tel2 (56) that found direct binding of Tel2 to telomeric DNA, chromatin immunoprecipitation (ChIP) assays detected weak telomere association of Tel2 that is eliminated in tel2-C307Y cells (Fig. 6H). In addition, we found that Rad26 associated with telomeres and the association was also significantly reduced to almost the background level in tel2-C307Y cells, suggesting that association of Rad3-Rad26 with telomeres is dependent on Tel2. Strong reduction in Rad26 binding in tel2-C307Y cells is especially remarkable as Rad26 protein level is actually increased in this mutant (Fig. 6D and S6). Taken together, these results clearly demonstrate that Tel2 is required to maintain wild-type telomere length in S. pombe by modulating Rad3-Rad26 action at telomeres.
Destabilization of the TTT complex.Tel2 binds to Tti1 and Tti2 to form the TTT complex, which interacts directly with newly synthesized PIKKs and functions as a cochaperone for proper folding of these large protein kinases (30, 31). Since the levels of Rad3 and Tel1 were reduced in tel2-C307Y cells, we then investigated whether the mutation affects the interactions of Tel2 with Tti1 and Tti2 by coimmunoprecipitation (co-IP). Tti1 and Tti2 were tagged with myc epitope at their genomic loci following the similar strategy used for tel2 integration (Fig. 2C). The tagged strains were crossed into S. pombe expressing endogenously HA tagged wild-type or mutant Tel2. The co-IP experiments were carried out multiple times, and representative results are shown in Fig. 7. We found that all three proteins could be successfully immunoprecipitated. When Tti1 and Tti2 were immunoprecipitated (Fig. 7A and C), Tel2 was coimmunoprecipitated. When Tti1 and Tti2 were immunoprecipitated from tel2-C307Y cells, much less Tel2 was coimmunoprecipitated, suggesting weaker interactions. The intensities of the coimmunoprecipitated Tel2 bands were quantified and normalized with the inputs (Table S4, sheet 7). After removal of nonspecific bindings, we found that the average levels of coimmunoprecipitated Tel2 were reduced to 14.2% (n = 3) and 26.5% (n = 4) of wild-type levels for Tti1 and Tti2, respectively (Fig. 7E). In the reciprocal coimmunoprecipitations, the mutation reduced the levels of coimmunoprecipitated Tti1 (Fig. 7B) and Tti2 (Fig. 7D) to 26.6% (n = 5) and 15.1% (n = 4) of the wild-type levels, respectively (Fig. 7E and sheet 7 in Table S4 in the supplemental material). Clearly, the TTT complex was significantly destabilized by the tel2-C307Y mutation.
Weakened interactions of Tel2-C307Y with Tti1 and Tti2. (A) Co-IP of Tel2 with Tti1 was significantly compromised by tel2-C307Y mutation. Tti1 was myc tagged at the genomic locus (SK6), crossed into S. pombe containing HA-tagged wild-type (SK7) or mutant (SK9) tel2. Tti1 was immunoprecipitated (upper panel) using anti-myc antibody to detect coimmunoprecipitated Tel2 (lower panel) as described in Materials and Methods. Untagged Tti1 strain (YJ1513) was used as the negative control. Totals of 2.4% of the cell extracts were loaded as the inputs (left three lanes). A section of the Ponceau S-stained inputs is shown. Tel2 specifically coimmunoprecipitated with Tti1 was quantified and normalized with the inputs, and the results are shown underneath the Tel2 blot. A representative result from five independent co-IP analyses is shown. (B) Reduced amount of Tti1 coimmunoprecipitated with Tel2 in tel2-C307Y. Tel2 was immunoprecipitated in the same strains as in panel A with anti-HA antibody to detect the coimmunoprecipitated Tti1. (C) Similar to the experiment in panel A, Tti2 was myc tagged in wild-type (YJ1549) and tel2-C307Y (YJ1550) cells. Tel2 coimmunoprecipitated with Tti2 in wild-type and tel2-C307Y cells was quantified and normalized, and the results are shown under the Tel2 blot. A representative result from three independent co-IP analyses is shown. (D) A reciprocal co-IP of Tti2 with Tel2 was carried out in wild-type and tel2-C307Y cells similar to the analysis described in panel B. (E) The coimmunoprecipitated bands in columns A, B, C, and D were quantified and normalized with the inputs. After the nonspecific bindings were removed, the intensities of the co-IP bands from tel2-C307Y cells are shown as percentages (brown) relative to those from wild-type cells (blue). The means and SEM are derived from five and three independent experiments for Tti1-Tel2 and Tti2-Tel2 co-IP analyses, respectively (Table S4, sheet 7).
Other potential functions of Tel2 in S. pombe.Since Rad3 kinase activity promotes normal growth of tel2-C307Y cells (Fig. 6A), we then crossed tel2-C307Y cells with other checkpoint mutants and examined the cell growth of the double mutants under normal conditions or in the presence of MMS, UV, or HU (Fig. 8A). The results showed that among all double mutants tested, the rad3Δ tel2-C307Y cells had the most severe growth defect, suggesting that Rad3 might play additional role other than checkpoint activation that is especially important in promoting the cell growth of tel2-C307Y. Conversely, since the rad3Δ tel2-C307Y double mutant and all other tested double mutants were more sensitive to MMS, UV, and HU, Tel2 may have other functions that are genetically separable from the checkpoint function. Interestingly, the double mutants containing crb2 or chk1 in the DDC were significantly more sensitive to MMS and UV than were the double mutants containing mrc1 and cds1 in the DRC pathway. Because HU specifically induces S-phase stress and MMS and UV mainly cause DNA damage at G2, this result is consistent with the observed major defect of tel2-C307Y in the DRC and not in the DDC (Fig. 4 and 5). Furthermore, with a compromised DDC, DNA damage caused by MMS and UV may be carried over to the next cell cycle (59), exacerbating the replication stress and thus enhancing the lethality. Since DNA strand breaks that occur at G2 are less likely to be carried over to next cell cycle (60, 61), we then examined the sensitivity of chk1Δ tel2-C307Y double mutant to bleomycin (Fig. 8B) with the expectation that this mutant might show less synthetic lethality compared to that in MMS or UV treatment. Surprisingly, similar to the MMS and UV treatments, the chk1Δ tel2-C307Y mutant, as well as the double mutants containing rad3 or cds1, was significantly more sensitive to bleomycin than the single mutants. It is possible that with a compromised DDC, DNA strand breaks, particularly single-strand breaks, can also be carried over to next cell cycle. The reduced Tel1 level in the double mutants may also explain the higher sensitivities because Tel1 can activate Chk1 under certain specific conditions (62, 63). As mentioned above, because tel2-C307Y cells were more sensitive to bleomycin than was the rad3 mutant, Tel2 may also regulate DNA strand break repair or other undefined cellular processes.
Increased drug sensitivities of the double mutants containing tel2-C307Y and checkpoint mutations. (A) Sensitivities of the single and double mutants containing tel2-C307Y and the indicated checkpoint mutations to MMS, UV, and HU were examined by using a standard spot assay. Dashed lines indicate discontinuity. (B) Sensitivities of the indicated single and double mutants containing tel2-C307Y, rad3, and cds1 in DRC and chk1 in DDC to bleomycin.
Because the tel2-C307Y mutant is a ts mutant, we examined the cell cycle progression after the cultures were shifted to 37°C. At 25°C, the mutant behaved almost like wild-type cells (see Fig. S7A in the supplemental material). After culture at 37°C, Tel2 gradually decreased to a minimally detectable level within ∼12 h (Fig. S7B). Meanwhile, the DNA contents began to increase, and a fraction of the mutant cells showed 8C to 16C DNA contents after 4 to 6 h, which was minimally detected in wild-type cells (Fig. S7A). This suggests DNA rereplication or misregulation of cell division. We then examined the cells under microscope after staining with Hoechst and Blankophor (Fig. S7C). Most wild-type cells contained a condensed nucleus before or after incubation at 37°C for 12 h. In contrast, the nuclei in some mutant cells were less condensed or relaxed after culturing at 37°C. Since cells with multiple septa were never observed, acute depletion of Tel2 may cause DNA rereplication, although other possibilities remain. Consistent with this result, a previous study showed that S. pombe lacking tel2 had a defect in S-phase entry (43). Together, these results suggest a potential role of Tel2 in DNA replication. Further studies are needed to investigate this possibility.
DISCUSSION
Our genetic screen has identified a previously uncharacterized tel2-C307Y mutation that significantly sensitizes S. pombe to HU and DNA damage. At the cellular level, the mutation causes cut cells in HU, and the number of cut cells corroborates the cell killing effect of HU in this mutant. At the molecular level, the mutation almost completely eliminates the Rad3 kinase signaling in the DRC, which explains the cut phenotype, as well as the hypersensitivities of the mutant to all tested replication stresses. We also measured the protein level of Rad3 and found it was moderately reduced. Although this moderate reduction of Rad3 may explain the partial signaling defect in the DDC, it remains unclear why the DRC is preferentially affected. Nonetheless, although the budding yeast tel2-1 mutant shows an almost wild-type resistance to replication stress (34), the DRC defect in S. pombe tel2-C307Y mutant is likely conserved in higher eukaryotes since enhanced sensitivity to replication stress has also been described in the two C. elegans tel2 mutants (29) and in mammalian cells (53, 64) in which Tel2 is depleted.
Tel2 interacts with Tti1 and Tti2 to form the TTT complex, which interacts with R2TP (Rvb1-Rvb2-Tah1-Pih1) complex (65, 66) to bring Hsp90 to PIKKs for their maturation. Interruption of the interactions between Tel2 and its client kinases reduces their steady-state levels, leading to a wide range of phenotypes, including the checkpoint defect. Consistent with this model, we found that although the levels of Rad3 and Tel1 were moderately reduced in tel2-C307Y, the mutation significantly weakened the interactions of Tel2 with Tti1 and Tti2. Thus, the reduced Rad3 level is likely caused by destabilization of the TTT complex. The remaining Rad3 in tel2-C307Y is clearly functional because Rad3-dependent Chk1 phosphorylation was still detectable. Furthermore, the kinase activity of Rad3 promotes the growth of tel2-C307Y cells, and its cofactor Rad26 is significantly increased in the mutant, which likely compensates for the partial loss of Rad3. The reduced DDC signaling may leave some DNA lesions “unnoticed” at G2 that are likely carried over to the next cell cycle and thus aggravate the replication stress. Therefore, the hypersensitivity of tel2-C307Y cells to DNA damage is likely caused by the lack of DRC exacerbated by a compromised DDC. It is also possible that the mutation affects other functions of Tel2, such as DNA repair, leading to the hypersensitivity to DNA damage, particularly those caused by bleomycin.
The most striking observation in this study is that although a reasonable amount of functional Rad3 remains in tel2-C307Y, its signaling in the DRC is almost completely abolished. The hypersensitivity of the mutant to HU and DNA damage suggests that Tel2 may function at or close to the checkpoint sensors. The decreased Rad3 level is clearly related to the checkpoint defect. However, why a moderate reduction of Rad3 has such a great impact on the DRC and not the DDC remains unclear. Here, we propose several hypotheses that need further investigations: first, in S. cerevisiae, the checkpoint is more tolerant to DNA damage or damage-like structures during the S phase (67). A similar threshold mechanism may also exist in S. pombe. When the Rad3 level is reduced, the threshold is raised to a level where the DRC completely “ignores” the replication stress. Second, as a cochaperone, Tel2 may help to assemble DRC proteins to the forks. For example, the highly charged and disordered Mrc1 is specifically expressed during G1/S transition (8) and a nonessential component of replisome (68–70). The tel2-C307Y mutation may affect the assembly of Mrc1 and/or other checkpoint proteins and hence the DRC defect. Consistent with this possibility, overexpression or acute depletion of Tel2 in mammalian cells causes checkpoint defects and increased sensitivity to replication stress even before ATR level is reduced (53, 58, 64). Third, the data shown in Fig. S7 and a previous study in S. pombe lacking tel2 (43) suggest that Tel2 may regulate DNA replication. Defects in this process may increase the checkpoint threshold in the S phase (67) mentioned above and explain the slower S-phase arrest of tel2-C307Y cells in HU (Fig. 3A and D). Finally, as described above, Tel2 may function in DNA repair, which is known to influence checkpoint signaling.
Tel2 physically interacts with telomeric DNA and regulates telomere length in S. cerevisiae (25, 56). Although it remains controversial whether Tel2 functions in the maintenance of telomere length in C. elegans, overexpression of Tel2 in human cells gradually lengthens telomeres (58). Our results show that similar to what occurs in S. cerevisiae, Tel2 associates with telomeres and functions in the homeostasis of telomere length in S. pombe. Furthermore, Tel2 was found to be important for recruitment of Rad26 to telomeres. Although Rad3 and Tel1 play redundant roles in phosphorylating telomere protein Ccq1 (54, 55), Rad3-Rad26 complex is the primary kinase that phosphorylates Ccq1 to maintain normal telomere length. Thus, rad3Δ and rad26Δ cells show significantly shorter but stable telomeres due to the residual action of Tel1. On the other hand, tel1Δ cells do not show any telomere shortening, and association of Tel1 at telomere cannot be detected in S. pombe cells (71). Thus, short but stable telomeres observed in the tel2-C307Y mutant, combined with the significant loss of Rad26 association at the telomeres, suggest that Tel2’s role in telomere length regulation is primarily through regulation of the Rad3-Rad26 kinase complex, although we cannot entirely rule out the possibility that Tel2 also affects Tel1 function at telomeres.
MATERIALS AND METHODS
Yeast strains and plasmids.The S. pombe strains were usually cultured at 30°C in YE6S (0.5% yeast extract, 3% dextrose, and six supplements) or in EMM medium lacking the appropriate supplements following standard methods (72). Yeast strains, plasmids, and PCR primers used in this study are listed in Tables S1, S2, and S3, respectively, in the supplemental material. All cloned genes and mutations were confirmed by DNA sequencing (Retrogen).
hus screen.Screening for new hus mutants was carried out as previously described (38, 39). Briefly, wild-type S. pombe was mutagenized with methylnitronitrosoguanidine and saved at 4°C (73). Approximately 3,000 to 5,000 mutagenized cells were spread on each YE6S plate. The colonies were replicated onto YE6S plates containing 5 mM HU to screen those that were sensitive to HU (Sigma). The primary hus mutants were crossed with known hus mutants to identify novel mutants and then backcrossed three times to remove bystander mutations.
Integration of tel2-C307Y mutation.The tel2-C307Y expression cassette was cloned into the pBS cloning vector between BamHI and MluI sites. HA tags linked with an ura4 marker were inserted in frame at the 3′ end of tel2 (see diagram in Fig. 2C). After digestion with PstI and XhoI, the 6,470-bp integration fragment was gel purified and then transformed into the wild-type TK7 strain. Colonies formed on plates lacking uracil were replicated onto HU plates. Integrants that were sensitive to HU were screened by colony PCRs for both 5′ and 3′ ends. The screened integrants were also backcrossed once to ensure single-copy integration in the genome. Genomic DNAs were then purified from the integrants for PCRs using the primers SpTel2(P)f and SpTel2(T)MluI-b and Phusion polymerase (NEB) to confirm the integration at the tel2 locus. Western blotting with anti-HA antibody confirmed the proper tagging and integration.
Drug sensitivity.Sensitivities to HU and various DNA-damaging agents were determined by standard spot assay or in liquid medium as described in our previous studies (8, 11, 38, 39). Briefly, for the spot assay, 2 × 107 cells/ml of logarithmically growing S. pombe were diluted in 5-fold steps and spotted in 3-μl portions onto YE6S plates or YE6S plates containing the drugs at indicated concentrations. The YE6S plates spotted with the cells were dried before the treatment with UV (Stratalinker 2400). The plates were incubated at 30°C for 3 days or 25°C for 4 days and then photographed. All spot assay experiments were repeated at least once. To examine the sensitivity to acute HU treatment (73), HU was added to liquid YE6S medium at 15 mM. At each time point, an equal number of cells were removed, diluted 1,000-fold, spread onto three YE6S plates, and incubated at 30°C for 3 days for cell recovery. Colonies were counted and are presented as percentages of the untreated cells.
IP and co-IP.A total of 1 × 108 logarithmically growing cells were harvested and saved at –20°C in a 1.5-ml screw-cap tube. The frozen cell pellets were lysed by using a mini-bead beater in buffer containing 25 mM HEPES-NaOH (pH 7.5), 50 mM NaF, 1 mM NaVO4, 10 mM NaP2O7, 40 mM β-glycerophosphate, 0.1% Tween 20, 0.5% NP-40, and protease inhibitors. The lysates were centrifuged at 16,000 × g at 4°C for 5 min to make the cell extract. Anti-HA or anti-myc antibody–agarose resins (Santa Cruz) were washed with Tris-buffered saline containing 0.05% Tween 20 (TBS-T) three times and incubated with 5% BSA in TBS-T for ≥30 min at 4°C. The cell extract was incubated with the prewashed antibody resins by rotating in 2-ml tubes at 4°C for 2 h. The resins were washed with TBS-T at 4°C for 20 min, repeated three times. The immunoprecipitated samples were separated by SDS-PAGE, followed by Western blot analyses with anti-HA or anti-myc antibodies.
Western blotting.The phospho-specific antibodies against phosphorylated Mrc1-Thr645, Rad9-Thr412, and Cds1-Thr11, and their specificities were described in our previous studies (8, 11, 44). The phospho-specific antibody against phosphorylated Chk1-Ser345 was kindly provided by N. C. Walworth (48). Western analyses of phosphorylated Rad9-Thr412, Mrc1-Thr645, and Cds1-Thr11 have been described in our previous study (11). Rad3, Tel1, Rad26, Tel2, Tti1, and Tti2 were tagged with myc or HA epitope at their genomic loci and examined by Western blotting by using mouse monoclonal antibodies against the myc (9E10; Thermo Scientific) or HA (12CA5; Sigma). For the Western analyses, 1 × 108 logarithmically growing cells were fixed in 15% trichloroacetic acid (TCA) on ice for ≥3 h and then lysed by using a mini-bead beater. The lysates from 2 × 106 to 4 × 106 cells were separated by SDS-PAGE. Rad3, Rad26, and Tel2 were directly detected in whole-cell lysates by Western blotting with Ponceau S staining as the loading control. Tel1 was immunoprecipitated before the Western analysis. The blotting signal was detected by electrochemiluminescence by using a ChemiDoc XRS imaging system (Bio-Rad). The intensities of the specific bands were quantified and analyzed by ImageLab (Bio-Rad).
Flow cytometry.A total of 1 × 107 logarithmically growing cells were collected, fixed in ice-cold 70% ethanol, and then analyzed by using an Accuri C6 flow cytometer, as described in our previous studies (38, 39).
Microscopy.The cells were fixed directly onto uncoated glass slides by brief heating at 75°C for 30 s or in medium containing 2.5% glutaraldehyde at 4°C for ≥3 h. The glutaraldehyde-fixed cells were washed with phosphate-buffered saline by centrifugation at 2,300 × g for 30 s, stained in the same buffer with 5 μg/ml Hoechst 33258 (Sigma-Aldrich), and diluted 1:100 in Blankophor working solution (MP Biochemicals). The stained cells were examined using an Olympus EX41 fluorescence microscope. Images were captured with an IQCAM camera (Fast1394) using Qcapture Pro 6.0 software. Approximately 150 cells were counted for each sample, and counts were repeated three times. Images were also extracted into Photoshop (Adobe) to generate Fig. 3C, as well as Fig. S7C in the supplemental material.
ChIP assay.Logarithmically growing cells were grown at 30°C, processed for ChIP assay using monoclonal anti-myc or anti-HA antibodies, and analyzed by quantitative PCR as previously described (74). The primers used for ChIP assays (75) are listed in Table S3 in the supplemental material.
Telomere length analysis by Southern blotting.Wild-type and tel2-C307Y cells were restreaked ≥5 times on YES plates (>120 generations) at either 25 or 32°C. Logarithmically growing cells at either temperature were collected to prepare genomic DNAs, which were digested with EcoRI and then processed for Southern blot analyses with telomere probe as previously described (54). The gels of three independent experiments were quantified with ImageQuant software (see Fig. S8 in the supplemental material) to calculate the telomere length (Table S4, sheet 4). Since the EcoRI site is ∼750 bp away from the telomere repeat track in fission yeast, 750 bp was subtracted to estimate the length of telomere repeats.
ACKNOWLEDGMENTS
We thank Tom Kelly and Paul Russell for sharing the yeast strains, Nancy Walworth for Chk1 phospho-specific antibody, and Michael Kemp and three anonymous reviewers for critical reading of the manuscript. Other members of the Xu lab are acknowledged for their help and support.
This study was supported by NIH RO1 grants GM110132 to Y.J.X. and GM078253 to T.M.N.
FOOTNOTES
- Received 17 April 2019.
- Returned for modification 9 May 2019.
- Accepted 19 July 2019.
- Accepted manuscript posted online 22 July 2019.
Supplemental material for this article may be found at https://doi.org/10.1128/MCB.00175-19.
- Copyright © 2019 American Society for Microbiology.