TY - JOUR AU1 - Mersaoui, Sofiane Y AU2 - Bonnell, Erin AU3 - Wellinger, Raymund J AB - Abstract Cdc13 is an essential protein involved in telomere maintenance and chromosome capping. Individual domain analyses on Cdc13 suggest the presence of four distinct OB-fold domains and one recruitment domain. However, it remained unclear how these sub-domains function in the context of the whole protein in vivo. Here, we use individual single domain deletions to address their roles in telomere capping. We find that the OB2 domain contains a nuclear localization signal that is essential for nuclear import of Cdc13 and therefore is required for chromosome capping. The karyopherin Msn5 is important for nuclear localization, and retention of Cdc13 in the nucleus also requires its binding to telomeres. Moreover, Cdc13 homodimerization occurs even if the protein is not bound to DNA and is in the cytoplasm. Hence, Cdc13 abundance in the nucleus and, in consequence, its capping function is strongly affected by nucleo-cytoplasmic transport as well as nuclear retention by DNA binding. INTRODUCTION Functional telomeres, the terminal parts of linear chromosomes, are essential for safeguarding the genome. Their structure and organization are highly conserved amongst eukaryotic species. For example, telomeric DNA consists of short, repeated G-rich sequences generally ending with an extended overhang of the 3′-end (1,2). A certain minimal number of the species specific repeats is essential for telomere function as it is required for the binding of specialized telomeric proteins (1–3). The assembled functional telomeric complex protects ends from inappropriate resection, thus preventing DNA damage checkpoint activation and/or chromosome end to end fusions (1,2). A second function of telomeres is that of counteracting the progressive losses of terminal DNA that occurs during each round of replication due to the end-replication problem (4). Indeed, short telomeres are elongated by a specialized and conserved reverse transcriptase called telomerase (5,6). In budding yeast, both telomere elongation and the telomere capping function are orchestrated by a heterotrimeric complex composed of the essential proteins Cdc13, Stn1 and Ten1, respectively (abbreviated as CST; (1,7–9). CST complexes with similar functions exist in plants, fission yeast and mammals where they are associated with limiting telomerase-mediated telomere lengthening, lagging strand synthesis on telomeres and rescuing a stressed replication fork during replication of difficult sequences (10–16). There may be an analogous complex in the fruit fly (17,18), pointing to deeply rooted conserved features of chromosome capping mechanisms. In humans, mutations in either hCTC1 or hSTN1 have been shown to cause Coats Plus syndrome, characterized by telomere dysfunction not necessarily associated with short telomeres (13,19). Molecular modeling and direct structure function experiments on the budding yeast CST proteins revealed similarities with the heterotrimeric replication protein A (RPA) (20–22). Within CST, Cdc13 is thought to be the major hub for the various CST functions and it in particular determines DNA binding activity in vitro and in vivo (23–26). Loss of Cdc13 elicits a Rad9-dependent cell cycle arrest (9) and particularly the temperature-sensitive cdc13-1ts allele has been used extensively to study general DNA damage checkpoint function and cell cycle arrest mechanisms. However, the molecular details of why telomeres appear completely uncapped in cdc13-1 cells incubated at restrictive temperatures remained unknown. Studies on biochemical properties, including crystallization, of individual fragments revealed that Cdc13 may be composed of four distinct Oligosaccharide/oligonucleotide Binding (OB)-fold domains named OB1, OB2, OB3 (the DNA Binding Domain) and OB4, in addition to a recruitment domain (RD) that is in between OB1 and OB2 (Figure 1A; (21–24,27–32)). The N-terminal OB1 is proposed to mediate Cdc13 homodimerization (21). However, preventing this dimerization has little effect on capping and only slightly affects telomere homeostasis (21,29). A portion of Cdc13-OB1 also serves to anchor DNA Polymerase α-primase to the telomere to achieve C-strand synthesis (21,31,33,34). The RD domain just downstream of OB1 interacts with the telomerase subunit Est1, thereby mediating the link between telomerase and telomeres (25,35). Mutations that prevent this Cdc13–Est1 interaction will cause a loss of telomeric repeats and eventually a prolonged cell cycle arrest and lethality, the same phenotypes observed in cells that lack telomerase altogether (25,36–38). Figure 1. View largeDownload slide Cdc13 domain deletions and their effects on viability and telomere length. (A) Proposed domain map of the Cdc13-Myc protein and deletions used in this study. OB1, OB2, OB3 (DBD) and OB4 are described as OB (Oligonucleotide-oligosaccharide Binding)-fold domains, RD is the telomerase Recruitment Domain and DBD is the DNA Binding Domain. (B) Western blot showing protein expression of the Cdc13-Myc protein variants as in (A and C). RNA PolII is the loading control. (C) Viability test of a cdc13-1 strain transformed with plasmids expressing the indicated cdc13 alleles. Cells were incubated at indicated temperatures; –: no CDC13 allele on plasmid. (D) Southern blot assessing telomere lengths in cdc13-1 cells expressing the indicated alleles. Below the gel, when applicable for the expressed allele, the number of generations grown at 30°C is indicated, –: cells grown at 23°C. Ob2Δ and ob3Δ alleles did not grow at 30°C (see C). Y’-TRF indicates the diagnostic smear of telomeric bands which for a wt is around 1.3 kb (see lanes on the left). The gel was hybridized to a telomeric repeat probe pCT300 (43). Figure 1. View largeDownload slide Cdc13 domain deletions and their effects on viability and telomere length. (A) Proposed domain map of the Cdc13-Myc protein and deletions used in this study. OB1, OB2, OB3 (DBD) and OB4 are described as OB (Oligonucleotide-oligosaccharide Binding)-fold domains, RD is the telomerase Recruitment Domain and DBD is the DNA Binding Domain. (B) Western blot showing protein expression of the Cdc13-Myc protein variants as in (A and C). RNA PolII is the loading control. (C) Viability test of a cdc13-1 strain transformed with plasmids expressing the indicated cdc13 alleles. Cells were incubated at indicated temperatures; –: no CDC13 allele on plasmid. (D) Southern blot assessing telomere lengths in cdc13-1 cells expressing the indicated alleles. Below the gel, when applicable for the expressed allele, the number of generations grown at 30°C is indicated, –: cells grown at 23°C. Ob2Δ and ob3Δ alleles did not grow at 30°C (see C). Y’-TRF indicates the diagnostic smear of telomeric bands which for a wt is around 1.3 kb (see lanes on the left). The gel was hybridized to a telomeric repeat probe pCT300 (43). Partially overlapping with this Est1 binding in the Cdc13-RD is an association site for Stn1, which will limit telomerase mediated G-strand extension (27,35). The OB2 domain, downstream of the RD has been suggested to be required for CST assembly and involved in productive telomere capping (29). Of note, the point mutation in the cdc13-1 allele leading to the P371S amino acid change is in OB2 and cells harboring this allele display a very strong temperature-sensitive (ts) growth arrest phenotype at temperatures > 26°C (9). Next, OB3 is the DNA binding domain of Cdc13 which specifically binds single-stranded G-rich telomeric repeats such as the G-rich 3′ overhang on telomeres (24). DNA binding requires about 10–12 nucleotides and it is thought that this DNA association tethers Cdc13, and by extension, all CST functions to telomeres (23,24,26,32,35). Recent results show that the DNA binding characteristics of Cdc13 are fully determined by the OB3(DBD) alone and dimerization does not impinge on that activity (28). Finally, the available evidence suggests that Stn1 also interacts with the C-terminal OB4 (36,39). However, a deletion of this domain is tolerated in vivo and hence appears not essential for recruitment of Stn1 to telomeres (35). Analyses of isolated domains cannot completely address the functions of them in the context of the full length protein. Therefore, in particular the roles of the OB4 domain and the essential OB2 remained elusive. In this study, we take advantage of cdc13-independent strains (40,41) to analyze and characterize, in vivo, deletions of specific domains of Cdc13. The results show that in the absence of OB2, the Cdc13 protein is almost completely lost from the nucleus and localizes to the cytoplasm. Surprisingly, the same phenotype is observed using cells harboring the cdc13-1 allele and incubated at restrictive temperatures. Introducing an exogenous NLS sequence into Cdc13 without OB2 restores the nuclear localization of the mutated protein, suppresses the growth arrest phenotype of cdc13-1 cells at restrictive temperature and ensures functional telomere capping and cell viability on its own. In addition, nuclear retention of the Cdc13 protein also requires its ability to interact with DNA. Our data further argue that inside cells, Cdc13 occurs as a constitutive dimer, even if not bound to DNA. Thus, neither the cdc13-1 mutation, OB2 deletion, nor the loss of telomeric DNA interaction affected homodimer formation in vivo. Our results therefore for the first time explain an extensive literature on the phenotypes of the cdc13-1 mutation. In addition, we note that for the human CTC1 protein, which is part of hCST, cytoplasmic mislocalization is a feature of a number of mutations that are associated with Coats Plus syndrome (42). MATERIALS AND METHODS Plasmid constructions All plasmids are listed in Supplementary Table S1 and are derived from the original pVL438 (CEN, URA3, CDC13 (35)). All described mutations were achieved by PCR mutagenesis (primer sequences are supplied upon request) and verified by sequencing. All strains used in this study are listed in Supplementary Table S2. Note that all strains are annotated for where they were used in the experiments/figures. Southern blot and in-gel hybridization analysis DNA was extracted from saturated liquid cultures using a glass bead procedure (43). For native in-gel analyses, DNA was extracted from mid-log phase cultures. The DNA was digested with XhoI, fragments separated on agarose gels and the gels hybridized without any denaturation using a CA oligonucleotide probe as described in (44). For loading controls, after imaging of the signals from the native DNA gels, the DNA in the gels was denatured and gels re-probed with the CA oligonucleotide. For quantification, whole-lane signals from native hybridization were normalized first to the signal obtained from denatured, total DNA, and then compared to the control strains, as in (44). Experiments were done at least three times. Chromatin Immuno-Precipitation (ChIP) analysis ChIP was performed as described in (45,46) and the methodology is extensively detailed in the Supplemental Methods section. Briefly, strains were grown to OD = 1 and fixed with 1% (v/v) formaldehyde for 5 min. Cells were suspended in ChIP-lysis buffer containing PMSF. Flash frozen ‘Popcorn’ was obtained by dropping cells into liquid nitrogen and cell lysis was achieved in liquid nitrogen using a freezer mill method. After melting, samples were supplemented with a protease inhibitor cocktail and subjected to sonication. Soluble fractions were subjected to immunoprecipitation with Pierce™ Protein A/G magnetic beads using 2 μg of mouse monoclonal anti-Myc antibody for 2 h at 4°C. ChIP-beads were washed and crosslinks reversed overnight at 65°C. Samples were subjected to RNase (1 h) and Proteinase K (6 h) incubation. Then, DNA was extracted by a phenol/chloroform and purified by a Qiagen PCR kit purification. Immunopricipitates and Input DNA were analyzed by qPCR (see Supplemental Methods for details). Co-immunoprecipitation (Co-IP) and Western blot (WB) Protein extracts for WB were prepared using a trichloroacetic acid method. Proteins were separated by 6% SDS–PAGE for Cdc13-Myc blots and 10% for Rad53 blots as described in (40). Blotting was performed using indicated antibodies. Signals were revealed using horseradish peroxidase-conjugated secondary antibody with the enhanced chemiluminescence detection kit. CoIP was prepared as previously described (47). Cells from 0.5 OD cultures were pelleted and resuspended within appropriate lysis buffer supplied with protease inhibitors and PMSF. Flash frozen ‘Popcorn’ were made as above and ground in a Freezer/Mill. The powder was thawed on ice and cleared by centrifugation. Equivalents of 5 mg protein in the soluble fraction were adjusted for Co-IP in 1 ml of total volume. 20 μg of anti-Myc antibody were added and extracts incubated for 2 h at 4°C. IP was performed with 30 μl of a slurry with magnetic beads (Pierce™ Protein A/G) for another 2 h. For the HA IP experiments, 1 ml extract was added to 30 μl of a suspension of magnetic beads coated with anti-HA antibody. Beads were then removed and washed with TMG Lysis buffer. If indicated, beads were incubated with DNase I for 1 h at 37°C. Immunofluorescence and microscopy Cell cultures at OD 0.5 were fixed with 5% formaldehyde and suspended in 1 ml of lysis buffer for cell wall digestion (see Supplemental Methods for details). Cells were then resuspended and dropped onto immunofluorescence glass with a polylysine treated surface. After methanol fixation and acetone treatment, cells were air dried and treated with a PBS solution containing BSA and supplemented with a mix of primary antibody (anti-Myc and anti-tubulin or anti-PGK) for 2 h. Secondary antibodies (Alexa) conjugated to specific fluorophores were added in PBS buffer with BSA (1 mg/ml) for 2 h. After several washes, DNA was stained with DAPI and mounted on cover-slips. Pictures were acquired with a Zeiss mRm Axiocam mounted on an axio observer Z1 inverse microscope using Zen 2012 software as in (47). Localization and co-localization was determined manually. Signal intensity for ratio determinations was analyzed by Columbus software. Statistical t-test analyses are provided from at least three independent experiments. RESULTS Cdc13 OB2 and OB3 are essential for viability, while the RD and OB4 are essential for telomere length control In order to better understand the role of each domain in Cdc13, we generated a Myc tagged version expressed from a low copy plasmid (CEN/ARS) and in which each domain was deleted individually (see Figure 1A for details). Deleting the N-terminal OB1 caused a dramatic destabilization of the protein and causes a null phenotype (Figure 1B and Supplementary Figure S1D). However, that characteristic is not informative and OB1 deletions will not be discussed further. The remaining four alleles were first tested for their ability to complement a cdc13Δ allele in haploid spores. In line with previous results, two of the four new CDC13 alleles, cdc13-rdΔ and cdc13-ob4Δ, are not immediately essential for Cdc13-function (Supplementary Figure S1A). After further outgrowth, cultures of cells containing the cdc13-rdΔ allele displayed classical signs of senescence, similar in phenotype to tlc1Δ cells lacking telomerase altogether (Supplementary Figure S1B). Moreover, both cdc13-rdΔ and cdc13-ob4Δ complemented the temperature defect of the cdc13-1 strain at all temperatures tested (Figure 1C and Supplementary Figure S1C). In contrast, the cdc13-ob2Δ and cdc13-ob3Δ alleles could not complement an absence of Cdc13 and therefore, the two OB2 and OB3 domains are essential for Cdc13 function and cell viability (Supplementary Figure S1A). Transforming Cdc13-ob2Δ or Cdc13-ob3Δ separately into a strain that contains the hypomorphic cdc13–1 allele also did not suppress the thermosensitivity of this strain (Figure 1C). This lack of complementation is not due to protein expression levels: western blot analysis on total protein extracts showed that the expression level of the mutant proteins is comparable to the wildtype (Figure 1B). Southern blot analyses of telomeric restriction fragments using DNA derived from cells that harbor the cdc13-1 allele and the indicated deletion allele were performed after 120 generations of growth (Figure 1D). Cells with the Cdc13-ob2Δ or the Cdc13-ob3Δ protein do not grow at 30°C while when grown at permissive temperature, telomeres were of normal length (Figure 1D). As expected, cells expressing the Cdc13-rdΔ protein and grown at 30°C contained extremely short telomeres (Figure 1D). Remarkably however, the senescence phenotype and survivor formation was somewhat delayed from normally 100 generations to ∼220 generations in this situation (Supplementary Figure S1F). These results suggest that the cdc13-1 allele can partially trans-complement an allele that lacks the recruitment domain (cdc13-rdΔ), even if cells are grown at restrictive temperature. In the complete absence of telomerase, for example if cells lack the RNA component, a small subset of them will switch to a recombination-based telomere maintenance and such cells are called survivors. They show at least two distinct phenotypes in terms telomeric restriction fragments: type I cells display an amplification of telomere associated sequences and very short terminal repeat tracts, whereas type II cells display many bands of highly variable distinct sizes (for this latter, see Supplementary Figure S1E, lane 1) (1,48). In Cdc13-independent cells, also called adapted cdc13Δ cells that were derived from type-II survivors (41), the signals for telomeric restriction fragments are extremely smeary and almost lacking distinct bands (see example in Supplementary Figure S1E, lane 2). However, this phenotype reverts to the survivor pattern after reintroduction of a wt Cdc13 protein (Supplementary Figure S1E, lane 3; (41)). The Cdc13-rdΔ and Cdc13-ob4Δ proteins were sufficient to also revert this phenotype back to distinct bands (Supplementary Figure S1E, lanes 4, 7). However, introduction of the Cdc13-ob2Δ or Cdc13-ob3Δ proteins did not yield such a phenotype reversion, suggesting that these alleles caused a total loss of all functions (see lanes 5 and 6, Supplementary Figure S1E). As previously observed with the cdc13-2 allele that is located in the RD domain, the results with the Cdc13-rdΔ protein confirm that this domain does not contribute to the capping function (25), and by extension this also applies to the OB4 domain. Given that DNA binding is thought to be required for all functions of Cdc13, the deletion of OB3 (DBD) was expected to yield the complete loss of function phenotypes observed here. However, the fact that OB2 was equally essential was unexpected. Both OB2 and OB3 are required for DNA binding in vivo To better understand the role of the OB2 domain, we analyzed the functionality of the Cdc13-ob2Δ protein in telomere capping in more detail. One of the hallmarks of loss of Cdc13-mediated capping is extensive 5′-strand resection from the telomeres that can be detected via a native DNA in-gel hybridization assay (44). For the analyses here, cdc13–1 cells also expressing the indicated domain-deleted alleles were incubated at permissive temperature (– in Figure 2A) or shifted to 30°C, the restrictive temperature for cdc13-1, for 2h (+ in Figure 2A). Without any plasmid encoded version of Cdc13, 5′ resection and the accumulating ssDNA is readily detectable (Figure 2A, compare lane 2 to lane 4). The Cdc13-rdΔ and Cdc13-ob4Δ can fully complement this capping defect, as no increase in ssDNA signal is detectable as if a wt Cdc13 was expressed (Figure 2A, lanes 4, 6 and 12). However, on DNA derived from cells expressing the Cdc13-ob2Δ and Cdc13-ob3Δ proteins, a strong capping defect is detected (Figure 2A, lanes 8 and 10). This ssDNA accumulation results in Rad53 hyperphosphorylation which is indicative of an activation of the DNA damage checkpoint (Figure 2B lanes 2, 8 and 10; Supplementary Figure S2A). In order to address the possibility of a partial trans-complementation that could interfere with interpretations, the above experiment was performed in two additional ways. In the first, the endogenous Cdc13 is induced to be degraded via the degron system and therefore removed (49). Consistent with the results with the situation with the cdc13-1 allele above, expressing the Wt protein or the Cdc13-rdΔ or Cdc13-ob4Δ proteins fully protected the telomeres from resection (Supplementary Figure S2B). In contrast, expression of Cdc13-ob2Δ or Cdc13-ob3Δ revealed a strong capping defect and checkpoint activation after 3 h of degron induction (Supplementary Figure S2B–D). In the second situation, we performed the same analyses in an adapted cdc13Δ strain. Again as above, the Cdc13-rdΔ and the Cdc13-ob4Δ conserved the capping capacity and virtually no ssDNA accumulation was detected (Figure 2C), while the Cdc13-ob2Δ and Cdc13-ob3Δ expressing cells show a strong capping defect (Figure 2C). Hence, in all situations tested, cells expressing the Cdc13-ob2Δ protein behave like those expressing the Cdc13-ob3Δ, which cannot bind DNA. Therefore we assessed whether the Cdc13-ob2Δ can be associated with DNA in vivo by ChIP assays using cells that expressed the generated alleles in either adapted cdc13Δ or in cdc13-1 strains. As expected from previous data, both the Cdc13-rdΔ and Cdc13-ob4Δ proteins still bind DNA in vivo (Figure 2D and Supplementary Figure S2E) (24,25). In contrast, the Cdc13-ob2Δ protein behaved indistinguishably from the Cdc13-ob3Δ protein, suggesting that it does not associate with ssDNA in vivo (Figure 2D and Supplementary Figure S2E). Western blots after IP immunoprecipitation confirmed that the mutated proteins could be enriched equally with the anti-myc antibody (Figure 2E). In conclusion, the Cdc13-ob2Δ protein appeared not to be associated with DNA in vivo despite the fact that it contained the complete OB3 domain that was thought to be sufficient for all DNA binding activity of Cdc13 (28) and that the OB2 alone did not associate with DNA in vitro (24,29). Figure 2. View largeDownload slide Cdc13-ob2Δ and Cdc13-ob3Δ proteins lack DNA interaction and cause a loss of capping. (A) Native in-gel hybridization of DNA derived from cdc13-1 strains expressing indicated alleles of CDC13 incubated at permissive and restrictive conditions for 2 h. (–) Growth at 23°C; (+) growth at 30°C. sstDNA is a single stranded control DNA. Bottom panel: same gel re-hybridized after denaturation to control for DNA loading. Panel to the right: relative ssDNA signal detection, wt signal is set as 1. (B) Western blot for hyper phosphorylation status of Rad53 in each strain as indicated in (A); (–) growth at 23°C; (+) incubated at 30°C for 5 h. RNA PolII is the loading control. (C) Upper panel: Native In-gel hybridization of an adapted cdc13Δ strain expressing the indicated mutated Cdc13 proteins. Middle panel: EtBr stain as loading control. Lower panel: relative quantification ssDNA as in (A) (*P < 0.05, and **P < 0.01 obtained by t-test for A and C). (D) ChIP of indicated Myc-tagged Cdc13 proteins expressed in the adapted cdc13Δ strain. Relative quantification as % of input for the indicated strains. Black boxes: values for telomeric sequence enrichment; Gray box: values for an unrelated locus (CLN2) as negative control. (–) Indicates untagged version of Cdc13. (E) Western blot of IP Cdc13-Myc variants in used in (D). * Non-specific signal. Figure 2. View largeDownload slide Cdc13-ob2Δ and Cdc13-ob3Δ proteins lack DNA interaction and cause a loss of capping. (A) Native in-gel hybridization of DNA derived from cdc13-1 strains expressing indicated alleles of CDC13 incubated at permissive and restrictive conditions for 2 h. (–) Growth at 23°C; (+) growth at 30°C. sstDNA is a single stranded control DNA. Bottom panel: same gel re-hybridized after denaturation to control for DNA loading. Panel to the right: relative ssDNA signal detection, wt signal is set as 1. (B) Western blot for hyper phosphorylation status of Rad53 in each strain as indicated in (A); (–) growth at 23°C; (+) incubated at 30°C for 5 h. RNA PolII is the loading control. (C) Upper panel: Native In-gel hybridization of an adapted cdc13Δ strain expressing the indicated mutated Cdc13 proteins. Middle panel: EtBr stain as loading control. Lower panel: relative quantification ssDNA as in (A) (*P < 0.05, and **P < 0.01 obtained by t-test for A and C). (D) ChIP of indicated Myc-tagged Cdc13 proteins expressed in the adapted cdc13Δ strain. Relative quantification as % of input for the indicated strains. Black boxes: values for telomeric sequence enrichment; Gray box: values for an unrelated locus (CLN2) as negative control. (–) Indicates untagged version of Cdc13. (E) Western blot of IP Cdc13-Myc variants in used in (D). * Non-specific signal. The OB2 domain is required for nuclear localization of Cdc13 The above results suggested either a general misfolding with complete loss of function for the Cdc13-ob2Δ protein or a mislocalisation. In order to assess the latter hypothesis, we used the adapted cdc13Δ strains that grow even in the complete absence of Cdc13 and in which we localized tagged mutated proteins by immuno-fluorescence. When these cells express the wild type Cdc13-Myc protein, telomeres are capped (Figure 2C) and anti-myc staining revealed discrete staining almost exclusively in the nucleus, as expected (Figure 3A, second row). Quantification of signal intensities for Cdc13-Myc (red) versus DNA (blue) showed that more than 75% of the signal co-localized (Figure 3A, right, Supplementary Figure S3). The Cdc13-rdΔ and Cdc13-ob4Δ proteins were also found to localize to the nucleus (Figure 3A, Supplementary Figure S3), in line with the functional capping data above. In contrast, the Cdc13-ob2Δ and Cdc13-ob3Δ proteins showed differing and distinct phenotypes. The Cdc13-ob3Δ protein displayed a non-specific localization pattern: it is found equally abundant in the cytoplasm and nucleus (Figure 3A). Remarkably, the Cdc13-ob2Δ protein appears absent from the nucleus and localized almost exclusively to the cytoplasm (Figure 3A, Supplementary Figure S3B). This cytoplasmic localization suggested that the OB2 domain determines nuclear localization via import and/or the introduced mutations induced an export via a normally hidden nuclear export signal. Computational analyses of the Cdc13 protein actually predict a bipartite nuclear localization signal (NLS)(Figure 3B, (50)). In fact, this predicted NLS localizes on the N-terminal side of OB2 (a.a. 325–350). Indeed, while deleting a.a. 289–328 is well tolerated and results in a functional protein, deleting a.a. 329–368 results in a complete loss of function (Supplementary Figure S4A and B). If OB2 determined nuclear import of Cdc13, introduction of a strong exogenous nuclear localization signal into the Cdc13-ob2Δ protein should restore at least some functionality. Indeed, when we introduced the SV40 NLS between the RD domain and the OB3 of the Cdc13-ob2Δ, a construct called Cdc13-ob2Δ-NLS, >70% of the nuclear signal was restored (Figure 4A). As a control, we also added an NLS sequence in place of the OB3 between the OB2 and the OB4 and obtained a similar result (Figure 4A; Supplementary Figure S5A). Western blots confirmed that these proteins were expressed normally, even if the Cdc13-ob2Δ-NLS protein appears slightly less abundant (Supplementary Figure S4C). Nevertheless, the rescued localization of the Cdc13-ob2Δ-NLS protein to the nucleus is sufficient to allow cdc13–1 cells to grow at 30°C (Figure 4B) and prevent the generation of ssDNA by 5′-end resection (Figure 4C). Expression of this protein also abrogates Rad53 phosphorylation after incubation at 30°C (Figure 4D), but Rad53 still can be induced by incubation of the same strain with the radiomimetic Bleomycin, which induces DNA damage in a locus non-specific fashion (Supplementary Figure S4D). Consistent with these results obtained with cdc13-1 cells, the telomere capping function was also restored in the adapted cdc13Δ cells (Supplementary Figure S4E). These rescue phenotypes can be ascribed specifically to the introduced NLS signal peptide, because if this peptide contains an abrogating mutation (nlsm), all the complementing phenotypes were lost (Figure 4B and Supplementary Figure S4E). When the Cdc13-ob2Δ-NLS protein is expressed from a multicopy plasmid as the sole allele expressed, it actually allows cell growth (Figure 4E). Finally, telomeres in these cells harboring the Cdc13-ob2Δ-NLS multicopy plasmid had lengthened significantly beyond wt (Figure 4F). These findings strongly suggest that the N-terminal part of the OB2 domain is involved in nuclear import of the Cdc13 protein. In contrast to the results with the Cdc13-ob2Δ-NLS protein, the Cdc13-ob3Δ-NLS protein could not rescue the capping defect in adapted cdc13Δ cells (Supplementary Figure S4E), even though it was expressed at a level similar to wt (Supplementary Figure S4C) and properly localized to the nucleus (Figure 4A, Supplementary Figure S5A). These data corroborate the fact that the OB3 is required for DNA binding and hence all functions of Cdc13. Figure 3. View largeDownload slide Cdc13-ob2Δ mislocalizes to the cytoplasm. (A) Immunofluorescence using adapted cdc13Δ cells expressing the indicated Myc-tagged Cdc13 proteins. Blue: DAPI coloration, nuclear DNA; Red: Myc-tagged protein. Stippled White square: zoom-in. Graphics on the right represents the distribution of Cdc13-Myc protein in cells (Nucleus, Cytoplasm, N+C = localization in both nucleus and cytoplasm; (n) total number of cells analyzed in at least three independent experiments. Error bars represent ± SEM. Yellow arrowheads indicate particularly telling cells. (B) Graphic outline of the predicted cNLS localization in the Cdc13 protein as analyzed by cNLS Mapper (50). P value indicates the confidence level as calculated by NLS Mapper. Figure 3. View largeDownload slide Cdc13-ob2Δ mislocalizes to the cytoplasm. (A) Immunofluorescence using adapted cdc13Δ cells expressing the indicated Myc-tagged Cdc13 proteins. Blue: DAPI coloration, nuclear DNA; Red: Myc-tagged protein. Stippled White square: zoom-in. Graphics on the right represents the distribution of Cdc13-Myc protein in cells (Nucleus, Cytoplasm, N+C = localization in both nucleus and cytoplasm; (n) total number of cells analyzed in at least three independent experiments. Error bars represent ± SEM. Yellow arrowheads indicate particularly telling cells. (B) Graphic outline of the predicted cNLS localization in the Cdc13 protein as analyzed by cNLS Mapper (50). P value indicates the confidence level as calculated by NLS Mapper. Figure 4. View largeDownload slide Phenotype reversions after adding an NLS to Cdc13-ob2Δ. (A) Immunofluorescence of Myc-tagged Cdc13-ob2Δ-NLS and Cdc13-ob3Δ-NLS proteins in the adapted cdc13Δ strain as in Figure 3A. (B) Complementation of growth in the cdc13-1 at 30°C. The Cdc13-ob2Δ-nlsm protein harbors a mutated non-functional NLS. (C) In-gel hybridization analysis of telomeric DNA from the indicated strains. (D) Western blot for Rad53 phosphorylation in proteins derived from strains as in (C) grown at 23°C or 30°C as indicated. (E) Expression of the Cdc13-ob2Δ-NLS protein from a 2μ plasmid yields viable cells. Cells before FOA selection do contain a wtCdc13 gene on a URA3-plasmid (top row Yc-U-T; + pCDC13 (URA3)). After selection on FOA plates, cells contained TRP1 plasmids with the indicated CDC13 allele (empty = pRS424, no Cdc13; ob2Δ = pRS424+Cdc13-ob2Δ; ob2Δ-NLS = pRS424+Cdc13-ob2Δ-NLS; as indicated on top). (F) Telomere southern blot analysis of DNA from cells derived from panel (E). Cells before FOA selection do contain a wtCdc13 gene on a URA3-plasmid (lanes indicated with a – below gel); after FOA selection, cells contained the indicated allele (wtCdc13 or Cdc13-ob2Δ-NLS; as indicated on top of gel) and were grown for 30 or 130 generations (indicated below gel). Figure 4. View largeDownload slide Phenotype reversions after adding an NLS to Cdc13-ob2Δ. (A) Immunofluorescence of Myc-tagged Cdc13-ob2Δ-NLS and Cdc13-ob3Δ-NLS proteins in the adapted cdc13Δ strain as in Figure 3A. (B) Complementation of growth in the cdc13-1 at 30°C. The Cdc13-ob2Δ-nlsm protein harbors a mutated non-functional NLS. (C) In-gel hybridization analysis of telomeric DNA from the indicated strains. (D) Western blot for Rad53 phosphorylation in proteins derived from strains as in (C) grown at 23°C or 30°C as indicated. (E) Expression of the Cdc13-ob2Δ-NLS protein from a 2μ plasmid yields viable cells. Cells before FOA selection do contain a wtCdc13 gene on a URA3-plasmid (top row Yc-U-T; + pCDC13 (URA3)). After selection on FOA plates, cells contained TRP1 plasmids with the indicated CDC13 allele (empty = pRS424, no Cdc13; ob2Δ = pRS424+Cdc13-ob2Δ; ob2Δ-NLS = pRS424+Cdc13-ob2Δ-NLS; as indicated on top). (F) Telomere southern blot analysis of DNA from cells derived from panel (E). Cells before FOA selection do contain a wtCdc13 gene on a URA3-plasmid (lanes indicated with a – below gel); after FOA selection, cells contained the indicated allele (wtCdc13 or Cdc13-ob2Δ-NLS; as indicated on top of gel) and were grown for 30 or 130 generations (indicated below gel). Given these dramatic mis-localization findings with the Cdc13-ob2Δ protein, we wondered whether a similar effect could explain the phenotypes of the Cdc13-1 protein as well. The a.a. change in this mutation is in OB2 and we thus performed immuno-fluorescence studies with a Myc-tagged Cdc13-1 at different incubation temperatures (Figure 5). While the nuclear localization of the wt Cdc13 is not affected by temperature, there is a stark difference in the localization of the Cdc13-1-Myc protein between cells incubated at 23°C versus 35°C (Figure 5A and B, Supplementary Figure S5B). The predominantly nuclear staining of the Cdc13-1-Myc protein at 23°C is by and large lost for a very diffuse cytoplasmic staining at 35°C. This final phenotype is comparable to that displayed by the Cdc13-ob2Δ protein (see Figure 3A). Cells expressing the Cdc13-1-Myc protein and which are incubated at the restrictive temperature do induce Rad53 phosphorylation (Figure 5C) and arrest with a G2/M morphology, as expected. Moreover, in adapted cdc13Δ cells expressing the Cdc13-ob2Δ protein, the protein appears always in the cytoplasm, irrespective of the specific cell cycle phase of the analyzed cell (Figure 3A). Figure 5. View large Download slide The cdc13-1 mutation causes mislocalization. (A) Immunofluorescence on strains expressing the Cdc13-Myc, Cdc13-1-Myc or the Cdc13-ob2Δ-c-Myc protein. The cdc13-ob2Δ-c lacks a.a. 369–408 (see Supplementary Figure S4A). Analyses are as in Figure 3. Temperatures of experiments are indicated on the right. (n) Total number of cells analyzed. Error bars represent ± SEM. (B) Quantification of the nuclear signal versus total signal obtained with the anti-Myc antibody in the indicated strains and conditions. Untagged proteins are set as 1. Calculated Mean value is indicated on the top of each bar and error bars represent ± SEM. Indicated P-values (p) are provided by t-test between the two defined conditions. (C) Western blot for Rad53 phosphorylation of indicated strains as in panel (A). Figure 5. View large Download slide The cdc13-1 mutation causes mislocalization. (A) Immunofluorescence on strains expressing the Cdc13-Myc, Cdc13-1-Myc or the Cdc13-ob2Δ-c-Myc protein. The cdc13-ob2Δ-c lacks a.a. 369–408 (see Supplementary Figure S4A). Analyses are as in Figure 3. Temperatures of experiments are indicated on the right. (n) Total number of cells analyzed. Error bars represent ± SEM. (B) Quantification of the nuclear signal versus total signal obtained with the anti-Myc antibody in the indicated strains and conditions. Untagged proteins are set as 1. Calculated Mean value is indicated on the top of each bar and error bars represent ± SEM. Indicated P-values (p) are provided by t-test between the two defined conditions. (C) Western blot for Rad53 phosphorylation of indicated strains as in panel (A). A previous investigation on the reasons for the temperature-sensitivity of the cdc13-1 allele showed that an absence of the karyopherin beta Msn5 aggravated it (51). Given our results, we examined the localization of Cdc13-1 in msn5Δ cells at 23°C. As mentioned above, the cytoplasmic localization of Myc-tagged Cdc13-1 at this temperature is subtle, but noticeable, in MSN5wt cells. This phenotype becomes very much aggravated in msn5Δ cells, even at 23°C (Figure 6A, Supplementary Figure S5C). Remarkably, in msn5Δ cells, a certain level of cytoplasmic localization can be detected even for the wt Cdc13 protein (Figure 6, second row), suggesting that Msn5 does indeed play a role in the proper localization of Cdc13. Figure 6. View largeDownload slide Loss of nuclear signal for Cdc13 in msn5Δ cells. (A) Subcellular localization of Myc-tagged wt Cdc13 or Cdc13-1 was assessed in MSN5 or msn5Δ cells as indicated. All experiments were performed at 23°C and analyzed as above. (B) Quantification of the nuclear vs. cytoplasmic signal obtained with the anti-Myc antibody in the indicated strains and conditions as in (A). Untagged proteins are set as 1. Calculated Mean value is indicated on the top of each bar and error bars represent ± SEM. Indicated P-values (p) are provided by t-test between the two defined conditions. Figure 6. View largeDownload slide Loss of nuclear signal for Cdc13 in msn5Δ cells. (A) Subcellular localization of Myc-tagged wt Cdc13 or Cdc13-1 was assessed in MSN5 or msn5Δ cells as indicated. All experiments were performed at 23°C and analyzed as above. (B) Quantification of the nuclear vs. cytoplasmic signal obtained with the anti-Myc antibody in the indicated strains and conditions as in (A). Untagged proteins are set as 1. Calculated Mean value is indicated on the top of each bar and error bars represent ± SEM. Indicated P-values (p) are provided by t-test between the two defined conditions. Constitutive Cdc13 dimerization is independent of OB2 Previous results obtained with an isolated OB2 domain protein suggested a role in Cdc13 dimerization which may inhibit Cdc13-Stn1 interactions (29). In order to analyze the effect of an OB2 deletion on potential dimerization in the context of the complete Cdc13 protein, we performed in vivo co-immunoprecipitation experiments in a haploid strain that expressed two tagged versions of Cdc13 (13xMyc or 3xHA). When both Cdc13 proteins were wild-type versions, they were expressed at about equal levels and they can be co-immunoprecipitated, as expected (Supplementary Figure S6A–C). This interaction was not dependent on telomeric DNA, as dimerization was detectable even after DNase I treatment of the extracts and the IP fractions did not contain detectable levels of telomeric DNA sequences (Figure 7A). As expected, in this situation no detectable Rap1 protein was in the precipitate (Supplementary Figure S6D). Previous experiments demonstrated that the constitutive G-strand overhang on telomeres in mre11Δ cells were extremely short and most likely only a single Cdc13 molecule would fit on it (52). Cells with the mre11Δ allele and arrested in G0 for 48 h thus reflect a situation with minimal opportunity for Cdc13 to dimerize on the overhang DNA, yet the two Cdc13 versions still co-precipitated in extracts from these cells (Figure 7B). A previously described CDC13 allele called cdc13–4R contains 4 a.a. changes in OB1 and this allele was reported to have lost the ability to form dimers (21). Indeed, in the co-immunoprecipitation assays used here, dimerization of Cdc13-4R proteins could not be detected (Figure 7C, left panel). These setup experiments thus provided a solid platform with which to verify the interaction potential of the specific domain-deleted alleles. In order to avoid problems related to inviability of cells, for these next experiments we used the adapted cdc13Δ strain co-expressing the two differently tagged copies of the same alleles and compared co-immunoprecipitation profiles. Deletion of either the RD, the OB2 or the OB4 domains did not impair dimer formation (Supplementary Figure S6E). Consistent with this result, the hypomorphic cdc13–1 allele also did not impair the ability of the Cdc13–1 protein to be co-immunoprecipitated, even after 2 h of incubation of the cells at the restrictive temperature (Figure 7D). In contrast, the Cdc13-ob3Δ protein completely lost the ability to form a dimer (Supplementary Figure S6E). Remarkably, the Cdc13-ob3Δ also could not be co-immunoprecipitated with a wt protein (Supplementary Figure S6E, middle right). Given that the tagged Cdc13-ob2Δ protein still can be co-immunoprecipitated, yet mislocalizes to the cytoplasm and is not associated with telomeric DNA in vivo (Figures 2 and 3), these dimers form independently of DNA binding, confirming the DNase experiments above. Finally, we generated a new CDC13 allele (cdc13-ob3-13m), in which we changed 13 a.a. that were previously identified as key residues for the proper DNA binding of the OB3 domain (Figure 8A; (23). The Cdc13-ob3-13m protein is expressed in cells (Figure 7C), does not complement the cdc13-1 mutation at 30°C (Figure 8B), and does not bind telomeric DNA as assessed by ChIP (Figure 8C). However, tagged versions of this protein still can be co-immunoprecipitated and it can also form a heterodimer when co-expressed with a wt version of Cdc13 (Figure 7C right panels and Supplementary Figure S6E far right). We therefore conclude that homodimer formation is independent of DNA binding. This result allowed us to test whether DNA binding per se was required for nuclear retention of the Cdc13 protein, given that the Cdc13-ob3Δ version lost both, dimerization and nuclear retention. Immunofluorescence of Myc-tagged Cdc13-ob3-13m on cycling adapted cdc13Δ cells showed a diffuse signal localized in both nucleus and cytoplasm, essentially an indistinguishable phenotype as obtained with the Cdc13-ob3Δ protein (Figure 8D; Supplementary Figure S5D). This result suggests that it is Cdc13 binding to telomeric G-strand DNA that retains it in the nucleus and prevents its export to the cytoplasm. Figure 7. View large Download slide Co-immunoprecipitation of differentially tagged Cdc13 proteins. (A) Upper panel: Western blot probed with anti-HA antibodies after immunoprecipitation of protein extracts with anti-Myc. WCE: whole cell extract (before IP); IN: Input after centrifugal clearing of the WCE; FT: Flow through. Temperatures of the experiments and 1h DNase I treatment at 37°C before IP are indicated on right. Percentage indicates volume loaded relative to the Input. Lower panel: dot blot analysis after DNA extraction obtained from the same samples shown in the upper panel, and hybridized with telomeric probe PCT300. An aliquot of unlabeled PCT300 is used as a hybridization control (Telo-Prob). Signals of dots were quantified relative to the Input (R.IN) and the SEM from three independent experiments are indicated. Note the loss of DNA signal in the IPs with DNase treatment. (B) Co-immunoprecipitation of the two differentially tagged Cdc13 proteins in an mre11Δ background after extended G0 cell arrest. IPs are treated for 1h with DNase I at 37°C as indicated. (C) Left: Loss of Cdc13 dimerization if one of the proteins contains the 4R mutations. Western blot of precipitates from IP experiments derived from strains expressing the indicated proteins. Right: Co-immunoprecipitation of the Cdc13-ob3–13m dimer (see also reverse experiment in Supplementary Figure S4E). (D) Co-immunoprecipitation experiment of two cdc13-1 proteins tagged respectively with 13-Myc and 3xHA, and co-expressed from two independent plasmids. Strains are grown as indicated. Figure 7. View large Download slide Co-immunoprecipitation of differentially tagged Cdc13 proteins. (A) Upper panel: Western blot probed with anti-HA antibodies after immunoprecipitation of protein extracts with anti-Myc. WCE: whole cell extract (before IP); IN: Input after centrifugal clearing of the WCE; FT: Flow through. Temperatures of the experiments and 1h DNase I treatment at 37°C before IP are indicated on right. Percentage indicates volume loaded relative to the Input. Lower panel: dot blot analysis after DNA extraction obtained from the same samples shown in the upper panel, and hybridized with telomeric probe PCT300. An aliquot of unlabeled PCT300 is used as a hybridization control (Telo-Prob). Signals of dots were quantified relative to the Input (R.IN) and the SEM from three independent experiments are indicated. Note the loss of DNA signal in the IPs with DNase treatment. (B) Co-immunoprecipitation of the two differentially tagged Cdc13 proteins in an mre11Δ background after extended G0 cell arrest. IPs are treated for 1h with DNase I at 37°C as indicated. (C) Left: Loss of Cdc13 dimerization if one of the proteins contains the 4R mutations. Western blot of precipitates from IP experiments derived from strains expressing the indicated proteins. Right: Co-immunoprecipitation of the Cdc13-ob3–13m dimer (see also reverse experiment in Supplementary Figure S4E). (D) Co-immunoprecipitation experiment of two cdc13-1 proteins tagged respectively with 13-Myc and 3xHA, and co-expressed from two independent plasmids. Strains are grown as indicated. Figure 8. View largeDownload slide DNA association is not required for dimerization of Cdc13. (A) Sequence and position of the 13 a.a. residues changed in the DNA-binding defective cdc13-ob3-13m allele as compared to the wild type sequence. The modified residues are highlighted and the amino acid position within the protein is indicated on top. (B) Complementation test of the DNA-binding-defective cdc13-ob3-13m allele in the cdc13-1 strain at restrictive temperature. (C) Chromatin-IP experiment on the Cdc13-ob3-13m versus the Cdc13-ob3Δ variant expressed in the adapted cdc13Δ strains. (D) Immunofluorescence on cell of an adapted cdc13Δ strain expressing the Cdc13-ob3Δ or the Cdc13-ob3-13m protein. Analyses of the images as in Figure 3. Figure 8. View largeDownload slide DNA association is not required for dimerization of Cdc13. (A) Sequence and position of the 13 a.a. residues changed in the DNA-binding defective cdc13-ob3-13m allele as compared to the wild type sequence. The modified residues are highlighted and the amino acid position within the protein is indicated on top. (B) Complementation test of the DNA-binding-defective cdc13-ob3-13m allele in the cdc13-1 strain at restrictive temperature. (C) Chromatin-IP experiment on the Cdc13-ob3-13m versus the Cdc13-ob3Δ variant expressed in the adapted cdc13Δ strains. (D) Immunofluorescence on cell of an adapted cdc13Δ strain expressing the Cdc13-ob3Δ or the Cdc13-ob3-13m protein. Analyses of the images as in Figure 3. DISCUSSION The linearity of eukaryotic chromosomes requires specialized mechanisms to keep their ends stable (1–3). In mammals, a highly ordered six-member complex called shelterin is at the center for this function. In addition, during replication, the heterotrimeric hCtc1/hStn1/hTen1 complex (hCST) participates in limiting telomerase mediated G-strand extension as well as initiating lagging C-strand synthesis (13,42). However, it is equally important that these normal telomeric capping mechanisms not be active on chromosomal ends occurring elsewhere in the genome after accidental DNA breaks. In budding yeast, the trimeric Cdc13/Stn1/Ten (CST) complex is the key coordinating unit for telomeric sequence addition by telomerase and capping (1). CST has noted similarities with the trimeric ssDNA binding complex RPA and some of its functions may include those of a telomere specific RPA (22). Of the three CST proteins, Cdc13 is the one with high-affinity G-rich ssDNA binding and all telomeric functions depend on this capacity. Previous analyses of the full length Cdc13 protein or isolated fragments thereof have proposed an overall modular structure with four OB-fold domains and one recruitment domain between OB1 and OB2. Here we analyzed individual deletions of each of those domains and assessed the functions of the remaining protein in cells that harbor a ts allele of the gene (CDC13–1) or in cells that have adapted to grow without Cdc13 (40,41). Our immunolocalization studies showed deletion of all of OB2 (Δ290–493) causes a striking exclusive cytoplasmic localization (Figure 3). Predictive algorithms suggested that Cdc13 has a classical nuclear localization signal (cNLS) between a.a. 325 and 350 (Figure 3), a region previously ascribed as the N-terminus of OB2 (Figures 1 and 3B). Moreover, the addition of a classical SV40-derived NLS in place of this deletion corrects this mislocalization (Figure 4A), suppresses the thermo-sensitivity of the cdc13-1 allele at 30°C (Figure 4B–D) and, if expressed from a high copy plasmid, is sufficient to confer cellular viability on its own (Figure 4E). Furthermore, a deletion of only amino acids 329–368 also resulted in a non-functional Cdc13 (Supplementary Figure S4) and cytoplasmic localization (Figure 5A, bottom), suggesting that indeed, the essential function of the OB2 domain lies in mediating efficient import of Cdc13 into the nucleus and that this function is associated with the region 325–350 of OB2. Deleting the karyopherin Msn5 (Kap142) causes a significant shift of Cdc13 localization to the cytoplasm and an enhancement of the mislocalization of the Cdc13–1 protein, a phenotype that was also reported for the Rfa1 and Rfa2 proteins of RPA in msn5Δ cells (53). In addition, Msn5 had already been described as a phenotypic enhancer for the cdc13-1 allele (51). It is therefore reasonable to propose that nuclear import of Cdc13 is mediated by its NLS at a.a. 325–349 and that Msn5 contributes to it. However, Msn5-mediated nuclear localization can only be part of the import mechanism, given that a deletion of MSN5 is viable, yet nuclear localization of the Rfa1, Rfa2 and Cdc13 proteins is essential for their function. Currently, we do not know what other and redundant pathways ensure import in the absence of Msn5. Specifically for the Cdc13–1 protein, we cannot completely exclude an additional contribution of the San1-mediated protein degradation pathway to the observed phenotypes (51,54). Our co-immunoprecipitation approach confirmed previous suggestions of an area in OB1 (a.a. 80–120) that is important for dimerization of the protein (Figure 7C). In addition, the detailed analysis of the newly constructed alleles (OB2Δ, OB3-13m) shows that Cdc13 dimerization occurs constitutively and that dimerization is not dependent on DNA binding (Figures 7, 8 and Supplementary Figure S6). These data therefore predict that the stoichiometry of the CST complex is not a simple 1:1:1 ratio, as would be expected from a strict RPA parallelism. Rather, the results strongly suggest that Cdc13 functions as a dimer for the telomerase recruitment function, perhaps independently from the Stn1-Ten1 proteins (35). There is evidence for such an altered stoichiometry with at least two Cdc13 molecules in the CST complex also for the C. glabrata CST (30). While the data suggest a contribution to dimerization by the OB3 domain (Supplementary Figure S6E), it is possible that this particular protein is misfolded and cannot assume any function anymore. Nevertheless, the Cdc13-OB2Δ and the Cdc13–1 proteins, which both have a wt OB1 and OB3 do dimerize (Figure 7 and Supplementary Figure S6E) and both mislocalize to the cytoplasm (Figures 3 and 5). Therefore, dimerization occurs in the cytoplasm, presumably before nuclear import for the wt protein and is independent from DNA binding (Figures 3, 7 and 8). We speculate that dimerization of Cdc13 may be required to render the protein stable. Consistent with this possibility, mutations that delete the dimerization area in OB1 (Cdc13-ob1Δ, Figure 1) cause a destabilization of Cdc13 (Figure 1C, Supplementary Figure S1D). Interestingly, an inspection of the crystallized OB2 domain shows that in fact there is no structural information on the newly discovered NLS region (a.a. 325–350) in OB2. The first α-helical segment that is recognizable in the structure for OB2 starts at a.a. 348 and the classical OB fold is contained between a.a. 355 and 475 (29). Therefore, we propose that there is a new separate functional element, comprising the cNLS (325–350) between the RD domain and the actual OB2. This NLS domain mediates nuclear import and is essential for the in vivo function of Cdc13. Further, our data are consistent with the idea that the cdc13-1 (P371S) mutation interferes with NLS function. In the absence of structural information of the whole protein, one can only speculate on the changes that occur in the Cdc13–1 protein. We suppose a temperature-aggravated structural change of this area of the protein may cause an occlusion of the nearby NLS, leading to an absence of the protein from the nucleus and hence the observed phenotypes. Our results also show that DNA binding is required for nuclear retention of Cdc13 (Figures 3 and 8). Hence in a DNA binding mutant (the cdc13-ob3-13m allele), the protein displays a non-discriminate nuclear and cytoplasmic distribution. We interpret these data to mean that any Cdc13 protein not bound to DNA may be released to the cytoplasm, followed by cycles of import and export. For the wt protein, this would help prevent association of free Cdc13 to non-telomeric sites where it could stimulate inappropriate new telomere synthesis. Such non-telomeric DNA association is possible and Cdc13 phosphorylation has been proposed as one mechanism to limit off-target telomere formation (55). Our data here suggest that Cdc13 release to the cytoplasm has a major contribution to that genome stabilizing mechanism. Our conclusions neatly rationalize and explain a plethora of data in the relatively extensive literature that uses cdc13-1 as a DNA damage and checkpoint inducing system. In addition, mutations with similar cytoplasmic mislocalization have recently been described for the hCTC1 protein that is part of the human CST (42). The alleles causing cytoplasmic localization phenotype in human cells are directly associated with the rare, but severe Coats Plus disease, a syndrome that now is part of telomere associated diseases (56–58). Strikingly, disease-causing CTC1 alleles as well as the cdc13–1 allele cause cellular pathologies that are related to telomere dysfunction, but not necessarily with short telomeres (42). In addition, in both systems accumulation of single-stranded DNA appears to be a key characteristic associated with these alleles. Finally, certain hCTC1 alleles cause a loss of telomerase-mediated lengthening control, a feature that is also associated with cdc13–1 at semi permissive temperature or after the loss of the OB4 domain ((35); Figures 1 and 4). Therefore, given the overall structural and functional similarities between the budding yeast Cdc13 and the human CTC1, we propose that the yeast cdc13–1 allele could be used as a model system for investigations on treating the human disease caused by hCTC1 mutations. The results presented here would suggest that approaches stimulating nuclear import could become promising avenues in this regard. SUPPLEMENTARY DATA Supplementary Data are available at NAR Online. ACKNOWLEDGEMENTS We thank V. Lundblad for sharing a number of constructs and plasmids. M. Catala and E. Bajon helped with the initial microscopy techniques and the members of the Wellinger lab for stimulating discussions. FUNDING Canadian Institutes of Health Research [MOP 110982]; Canadian Research Chair in Telomere Biology; Sherbrooke RNA Group. Funding for open access charge: Canadian Institutes for Health Research. Conflict of interest statement. None declared. REFERENCES 1. Wellinger R.J., Zakian V.A. Everything you ever wanted to know about Saccharomyces cerevisiae telomeres: beginning to end. Genetics . 2012; 191: 1073– 1105. Google Scholar CrossRef Search ADS PubMed  2. Palm W., de Lange T. How shelterin protects mammalian telomeres. Annu. Rev. Genet.  2008; 42: 301– 334. Google Scholar CrossRef Search ADS PubMed  3. Schmutz I., de Lange T. Shelterin. Curr. Biol.  2016; 26: R397– R399. Google Scholar CrossRef Search ADS PubMed  4. Wellinger R.J. In the end, what's the problem. Mol. Cell . 2014; 53: 855– 856. Google Scholar CrossRef Search ADS PubMed  5. Wu R.A., Upton H.E., Vogan J.M., Collins K. Telomerase mechanism of telomere synthesis. Annu. Rev. Biochem.  2017; 86: 439– 460. Google Scholar CrossRef Search ADS PubMed  6. Greider C.W., Blackburn E.H. The telomere terminal transferase of Tetrahymena is a ribonucleoprotein enzyme with two kinds of primer specificity. Cell . 1987; 51: 887– 898. Google Scholar CrossRef Search ADS PubMed  7. Grandin N., Damon C., Charbonneau M. Ten1 functions in telomere end protection and length regulation in association with Stn1 and Cdc13. EMBO J.  2001; 20: 1173– 1183. Google Scholar CrossRef Search ADS PubMed  8. Grandin N., Reed S.I., Charbonneau M. Stn1, a new Saccharomyces cerevisiae protein, is implicated in telomere size regulation in association with Cdc13. Genes Dev.  1997; 11: 512– 527. Google Scholar CrossRef Search ADS PubMed  9. Garvik B., Carson M., Hartwell L. Single-stranded DNA arising at telomeres in cdc13 mutants may constitute a specific signal for the RAD9 checkpoint. Mol. Cell. Biol.  1995; 15: 6128– 6138. Google Scholar CrossRef Search ADS PubMed  10. Takikawa M., Tarumoto Y., Ishikawa F. Fission yeast Stn1 is crucial for semi-conservative replication at telomeres and subtelomeres. Nucleic Acids Res.  2017; 45: 1255– 1269. Google Scholar CrossRef Search ADS PubMed  11. Lee J.R., Xie X., Yang K., Zhang J., Lee S.Y., Shippen D.E. Dynamic interactions of Arabidopsis TEN1: stabilizing telomeres in response to heat stress. Plant Cell . 2016; 28: 2212– 2224. Google Scholar CrossRef Search ADS   12. Chastain M., Zhou Q., Shiva O., Whitmore L., Jia P., Dai X., Huang C., Fadri-Moskwik M., Ye P., Chai W. Human CST facilitates genome-wide RAD51 recruitment to GC-rich repetitive sequences in response to replication stress. Cell Rep.  2016; 16: 1300– 1314. Google Scholar CrossRef Search ADS PubMed  13. Chen L.Y., Lingner J. CST for the grand finale of telomere replication. Nucleus . 2013; 4: 277– 282. Google Scholar CrossRef Search ADS PubMed  14. Wellinger R.J. The CST complex and telomere maintenance: the exception becomes the rule. Mol. Cell . 2009; 36: 168– 169. Google Scholar CrossRef Search ADS PubMed  15. Surovtseva Y.V., Churikov D., Boltz K.A., Song X., Lamb J.C., Warrington R., Leehy K., Heacock M., Price C.M., Shippen D.E. Conserved telomere maintenance component 1 interacts with STN1 and maintains chromosome ends in higher eukaryotes. Mol. Cell . 2009; 36: 207– 218. Google Scholar CrossRef Search ADS PubMed  16. Miyake Y., Nakamura M., Nabetani A., Shimamura S., Tamura M., Yonehara S., Saito M., Ishikawa F. RPA-like mammalian Ctc1-Stn1-Ten1 complex binds to single-stranded DNA and protects telomeres independently of the Pot1 pathway. Mol. Cell . 2009; 36: 193– 206. Google Scholar CrossRef Search ADS PubMed  17. Cicconi A., Micheli E., Verni F., Jackson A., Gradilla A.C., Cipressa F., Raimondo D., Bosso G., Wakefield J.G., Ciapponi L.et al.   The Drosophila telomere-capping protein Verrocchio binds single-stranded DNA and protects telomeres from DNA damage response. Nucleic Acids Res.  2017; 45: 3068– 3085. Google Scholar CrossRef Search ADS PubMed  18. Zhang Y., Zhang L., Tang X., Bhardwaj S.R., Ji J., Rong Y.S. MTV, an ssDNA protecting complex essential for transposon-based telomere maintenance in Drosophila. PLoS Genet.  2016; 12: e1006435. Google Scholar CrossRef Search ADS PubMed  19. Simon A.J., Lev A., Zhang Y., Weiss B., Rylova A., Eyal E., Kol N., Barel O., Cesarkas K., Soudack M.et al.   Mutations in STN1 cause Coats plus syndrome and are associated with genomic and telomere defects. J. Exp. Med.  2016; 213: 1429– 1440. Google Scholar CrossRef Search ADS PubMed  20. Rice C., Skordalakes E. Structure and function of the telomeric CST complex. Comput. Struct. Biotechnol. J.  2016; 14: 161– 167. Google Scholar CrossRef Search ADS PubMed  21. Sun J., Yang Y., Wan K., Mao N., Yu T.Y., Lin Y.C., DeZwaan D.C., Freeman B.C., Lin J.J., Lue N.F.et al.   Structural bases of dimerization of yeast telomere protein Cdc13 and its interaction with the catalytic subunit of DNA polymerase alpha. Cell Res.  2011; 21: 258– 274. Google Scholar CrossRef Search ADS PubMed  22. Gao H., Cervantes R.B., Mandell E.K., Otero J.H., Lundblad V. RPA-like proteins mediate yeast telomere function. Nat. Struct. Mol. Biol.  2007; 14: 208– 214. Google Scholar CrossRef Search ADS PubMed  23. Mitton-Fry R.M., Anderson E.M., Hughes T.R., Lundblad V., Wuttke D.S. Conserved structure for single-stranded telomeric DNA recognition. Science . 2002; 296: 145– 147. Google Scholar CrossRef Search ADS PubMed  24. Hughes T.R., Weilbaecher R.G., Walterscheid M., Lundblad V. Identification of the single-strand telomeric DNA binding domain of the Saccharomyces cerevisiae Cdc13 protein. Proc. Natl. Acad. Sci. U.S.A.  2000; 97: 6457– 6462. Google Scholar CrossRef Search ADS PubMed  25. Nugent C.I., Hughes T.R., Lue N.F., Lundblad V. Cdc13p: a single-strand telomeric DNA-binding protein with a dual role in yeast telomere maintenance. Science . 1996; 274: 249– 252. Google Scholar CrossRef Search ADS PubMed  26. Lin J.J., Zakian V.A. The Saccharomyces CDC13 protein is a single-strand TG1-3 telomeric DNA-binding protein in vitro that affects telomere behavior in vivo. Proc. Natl. Acad. Sci. U.S.A.  1996; 93: 13760– 13765. Google Scholar CrossRef Search ADS PubMed  27. Wang M.J., Lin Y.C., Pang T.L., Lee J.M., Chou C.C., Lin J.J. Telomere-binding and Stn1p-interacting activities are required for the essential function of Saccharomyces cerevisiae Cdc13p. Nucleic Acids Res.  2000; 28: 4733– 4741. Google Scholar CrossRef Search ADS PubMed  28. Lewis K.A., Pfaff D.A., Earley J.N., Altschuler S.E., Wuttke D.S. The tenacious recognition of yeast telomere sequence by Cdc13 is fully exerted by a single OB-fold domain. Nucleic Acids Res.  2014; 42: 475– 484. Google Scholar CrossRef Search ADS PubMed  29. Mason M., Wanat J.J., Harper S., Schultz D.C., Speicher D.W., Johnson F.B., Skordalakes E. Cdc13 OB2 dimerization required for productive Stn1 binding and efficient telomere maintenance. Structure . 2013; 21: 109– 120. Google Scholar CrossRef Search ADS PubMed  30. Lue N.F., Zhou R., Chico L., Mao N., Steinberg-Neifach O., Ha T. The telomere capping complex CST has an unusual stoichiometry, makes multipartite interaction with G-Tails, and unfolds higher-order G-tail structures. PLoS Genet.  2013; 9: e1003145. Google Scholar CrossRef Search ADS PubMed  31. Mitchell M.T., Smith J.S., Mason M., Harper S., Speicher D.W., Johnson F.B., Skordalakes E. Cdc13 N-terminal dimerization, DNA binding, and telomere length regulation. Mol. Cell. Biol.  2010; 30: 5325– 5334. Google Scholar CrossRef Search ADS PubMed  32. Mitton-Fry R.M., Anderson E.M., Theobald D.L., Glustrom L.W., Wuttke D.S. Structural basis for telomeric single-stranded DNA recognition by yeast Cdc13. J. Mol. Biol.  2004; 338: 241– 255. Google Scholar CrossRef Search ADS PubMed  33. Hsu C.L., Chen Y.S., Tsai S.Y., Tu P.J., Wang M.J., Lin J.J. Interaction of Saccharomyces Cdc13p with Pol1p, Imp4p, Sir4p and Zds2p is involved in telomere replication, telomere maintenance and cell growth control. Nucleic Acids Res.  2004; 32: 511– 521. Google Scholar CrossRef Search ADS PubMed  34. Qi H., Zakian V.A. The Saccharomyces telomere-binding protein Cdc13p interacts with both the catalytic subunit of DNA polymerase alpha and the telomerase-associated est1 protein. Genes Dev.  2000; 14: 1777– 1788. Google Scholar PubMed  35. Chandra A., Hughes T.R., Nugent C.I., Lundblad V. Cdc13 both positively and negatively regulates telomere replication. Genes Dev.  2001; 15: 404– 414. Google Scholar CrossRef Search ADS PubMed  36. Pennock E., Buckley K., Lundblad V. Cdc13 delivers separate complexes to the telomere for end protection and replication. Cell . 2001; 104: 387– 396. Google Scholar CrossRef Search ADS PubMed  37. Lendvay T.S., Morris D.K., Sah J., Balasubramanian B., Lundblad V. Senescence mutants of Saccharomyces cerevisiae with a defect in telomere replication identify three additional EST genes. Genetics . 1996; 144: 1399– 1412. Google Scholar PubMed  38. Singer M.S., Gottschling D.E. TLC1: template RNA component of Saccharomyces cerevisiae telomerase. Science . 1994; 266: 404– 409. Google Scholar CrossRef Search ADS PubMed  39. Hang L.E., Liu X., Cheung I., Yang Y., Zhao X. SUMOylation regulates telomere length homeostasis by targeting Cdc13. Nat. Struct. Mol. Biol.  2011; 18: 920– 926. Google Scholar CrossRef Search ADS PubMed  40. Mersaoui S.Y., Gravel S., Karpov V., Wellinger R.J. DNA damage checkpoint adaptation genes are required for division of cells harbouring eroded telomeres. Microb. Cell . 2015; 2: 394– 405. Google Scholar CrossRef Search ADS PubMed  41. Larrivee M., Wellinger R.J. Telomerase- and capping-independent yeast survivors with alternate telomere states. Nat. Cell Biol.  2006; 8: 741– 747. Google Scholar CrossRef Search ADS PubMed  42. Chen L.Y., Majerska J., Lingner J. Molecular basis of telomere syndrome caused by CTC1 mutations. Genes Dev.  2013; 27: 2099– 2108. Google Scholar CrossRef Search ADS PubMed  43. Wellinger R.J., Wolf A.J., Zakian V.A. Origin activation and formation of single-strand TG1-3 tails occur sequentially in late S phase on a yeast linear plasmid. Mol. Cell. Biol.  1993; 13: 4057– 4065. Google Scholar CrossRef Search ADS PubMed  44. Dionne I., Wellinger R.J. Cell cycle-regulated generation of single-stranded G-rich DNA in the absence of telomerase. Proc. Natl. Acad. Sci. U.S.A.  1996; 93: 13902– 13907. Google Scholar CrossRef Search ADS PubMed  45. Fisher T.S., Taggart A.K.P., Zakian V.A. Cell cycle-dependent regulation of yeast telomerase by Ku. Nat. Struct. Mol. Biol.  2004; 11: 1198– 1205. Google Scholar CrossRef Search ADS PubMed  46. Taggart A.K., Teng S.C., Zakian V.A. Est1p as a cell cycle-regulated activator of telomere-bound telomerase. Science . 2002; 297: 1023– 1026. Google Scholar CrossRef Search ADS PubMed  47. Bajon E., Laterreur N., Wellinger R.J. A single templating RNA in yeast telomerase. Cell Rep.  2015; 12: 441– 448. Google Scholar CrossRef Search ADS PubMed  48. Teng S.C., Zakian V.A. Telomere-telomere recombination is an efficient bypass pathway for telomere maintenance in Saccharomyces cerevisiae. Mol. Cell. Biol.  1999; 19: 8083– 8093. Google Scholar CrossRef Search ADS PubMed  49. Vodenicharov M.D., Laterreur N., Wellinger R.J. Telomere capping in non-dividing yeast cells requires Yku and Rap1. EMBO J.  2010; 29: 3007– 3019. Google Scholar CrossRef Search ADS PubMed  50. Kosugi S., Hasebe M., Tomita M., Yanagawa H. Systematic identification of cell cycle-dependent yeast nucleocytoplasmic shuttling proteins by prediction of composite motifs. Proc. Natl. Acad. Sci. U.S.A.  2009; 106: 10171– 10176. Google Scholar CrossRef Search ADS PubMed  51. Addinall S.G., Holstein E.M., Lawless C., Yu M., Chapman K., Banks A.P., Ngo H.P., Maringele L., Taschuk M., Young A.et al.   Quantitative fitness analysis shows that NMD proteins and many other protein complexes suppress or enhance distinct telomere cap defects. PLoS Genet.  2011; 7: e1001362. Google Scholar CrossRef Search ADS PubMed  52. Larrivee M., LeBel C., Wellinger R.J. The generation of proper constitutive G-tails on yeast telomeres is dependent on the MRX complex. Genes Dev.  2004; 18: 1391– 1396. Google Scholar CrossRef Search ADS PubMed  53. Yoshida K., Blobel G. The karyopherin Kap142p/Msn5p mediates nuclear import and nuclear export of different cargo proteins. J. Cell Biol.  2001; 152: 729– 740. Google Scholar CrossRef Search ADS PubMed  54. Gardner R.G., Nelson Z.W., Gottschling D.E. Degradation-mediated protein quality control in the nucleus. Cell . 2005; 120: 803– 815. Google Scholar CrossRef Search ADS PubMed  55. Zhang W., Durocher D. De novo telomere formation is suppressed by the Mec1-dependent inhibition of Cdc13 accumulation at DNA breaks. Genes Dev.  2010; 24: 502– 515. Google Scholar CrossRef Search ADS PubMed  56. Polvi A., Linnankivi T., Kivela T., Herva R., Keating J.P., Makitie O., Pareyson D., Vainionpaa L., Lahtinen J., Hovatta I.et al.   Mutations in CTC1, encoding the CTS telomere maintenance complex component 1, cause cerebroretinal microangiopathy with calcifications and cysts. Am. J. Hum. Genet.  2012; 90: 540– 549. Google Scholar CrossRef Search ADS PubMed  57. Keller R.B., Gagne K.E., Usmani G.N., Asdourian G.K., Williams D.A., Hofmann I., Agarwal S. CTC1 Mutations in a patient with dyskeratosis congenita. Pediatr. Blood Cancer . 2012; 59: 311– 314. Google Scholar CrossRef Search ADS PubMed  58. Anderson B.H., Kasher P.R., Mayer J., Szynkiewicz M., Jenkinson E.M., Bhaskar S.S., Urquhart J.E., Daly S.B., Dickerson J.E., O'Sullivan J.et al.   Mutations in CTC1, encoding conserved telomere maintenance component 1, cause Coats plus. Nat. Genet.  2012; 44: 338– 342. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of Nucleic Acids Research. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com TI - Nuclear import of Cdc13 limits chromosomal capping JF - Nucleic Acids Research DO - 10.1093/nar/gky085 DA - 2018-04-06 UR - https://www.deepdyve.com/lp/oxford-university-press/nuclear-import-of-cdc13-limits-chromosomal-capping-ZF0yL3LQLx SP - 2975 EP - 2989 VL - 46 IS - 6 DP - DeepDyve ER -