INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The ends of linear eukaryotic chromosomes form a special structure, the telomere. The telomeric DNA-protein complexes are essential for chromosome stability (49). They protect chromosomes from degradation and end-to-end fusion (54) and ensure their complete replication (41). In most eukaryotes, telomeric DNA contains a simple, repetitive sequence with the strand running toward the end of the chromosome being rich in G residues. For some organisms the configuration of the chromosome ends has been defined exactly. In hypotrichous ciliates the double-stranded region is followed by a 12- to 16-nucleotide-long single-stranded (ss) 3' overhang (22, 24), whereas mouse and human chromosomes contain ss termini of 45 to 200 nucleotides (36, 39, 60). In the yeast Saccharomyces cerevisiae the telomere repeats consist of 300 ± 75 bp of C_1-3A/TG_1-3 DNA. Detectable ss extensions of the G-rich strand are generated at telomeres specifically during S phase in a telomerase-independent process (11, 57, 58). A specialized enzyme, telomerase, performs synthesis of telomeric DNA by extending the 3' end of the G-rich strand of the telomere. Telomerase activity in S. cerevisiae depends on at least four protein subunits (encoded by EST1, EST2, EST3, and CDC13/EST4) (28, 34) and the RNA component (encoded by TLC1) (52). All subunits are essential for telomerase function in vivo, although only the catalytic subunit EST2 and the RNA template TLC1 are necessary for in vitro activity (6, 8, 30). Deletion of most individual components of the telomerase complex leads to inactivation of telomerase and thereby to a decrease in telomere length and to replicative senescence (28, 34).

However, deletion of CDC13/EST4 leads to immediate cell cycle arrest and cell death (56). This phenotype is triggered by the accumulation of telomeric single-stranded DNA (ssDNA) that activates an RAD9-dependent G₂ arrest (16). Therefore Cdc13p was proposed to provide protection of the telomere from nucleolytic degradation by DNA end binding. This role is consistent with the finding that Cdc13p binds ss telomeric DNA in vitro (29, 40) and binds exclusively to telomeric, but not to internal, C_1-3A/TG_1-3 repeat sequences (5). Very recently the DNA binding domain of Cdc13p has been mapped to amino acids 557 to 694. Heterologous expression in Escherichia coli of a small, CDC13-derived polypeptide containing this region results in a protein that binds, like the full-length Cdc13p, with high affinity to ss telomeric DNA (23). A single amino acid missense mutation within this region of Cdc13p causes thermolabile DNA binding, and consistent with the presumption that Cdc13p DNA binding is essential to protect chromosome ends, this mutant is temperature sensitive for growth (23).

Besides its role in chromosome end protection, Cdc13p is involved in recruiting telomerase to telomeric DNA. cdc13-2^est mutant cells exhibit a senescence phenotype but can be rescued by expression of a Cdc13-2^est-Est1 fusion protein (12). These data suggest that Cdc13p is essential for loading telomerase to the telomere and that this process is mediated via interaction with Est1p. Interaction of Cdc13p and Est1p has been shown by two-hybrid criteria. Additionally, hemagglutinin (HA)-tagged Cdc13p can be copurified with a glutathione S-transferase (GST)-Est1 fusion protein from yeast extracts if both proteins were overexpressed (45). Furthermore, Cdc13p seems to be involved in the accurate regulation of telomerase recruitment, as several CDC13 mutations, not yet mapped at the genomic level, confer either elongated telomeres (41, 18) or shortened telomeres (18).

In S. cerevisiae the steady-state level of telomeric GT repeat tract length seems to result from a balance between telomere elongation and telomere shortening (37). Many proteins involved in telomere length maintenance have been identified already. A major factor involved in negative telomere length regulation is the Rap1 protein, which binds with high affinity to specific sequences within the telomeric GT repeat tracts (7). Unregulated telomere elongation is prevented mainly by Rap1p and its interacting partners Rif1p and Rif2p (21, 31, 59). It has been proposed that a negative feedback mechanism determines the exact number of Rap1p molecules bound to telomeric DNA and regulates telomerase activity (37, 38). Recently, a model has been suggested in which a special folded structure prevents telomere elongation (46). In this model, the formation of the folded structure of the chromosome end depends on the length of the GT repeat tract and on the number of bound Rap1p.

At least two pathways are involved in positive telomere length control in S. cerevisiae. One pathway involves Tel1p and the Mre11-Rad50-Xrs2 complex, and disruption of any of these genes results in stable shortened telomeres (47). A second pathway affecting positive telomere length regulation involves the Yku heterodimer, which is also an essential component of the nonhomologous end-joining pathway (2-4, 44). As shown by in vivo cross-linking experiments, Ykup binds directly to telomeric DNA (19). Yku mutant strains display short but stable telomeres, and the ss telomeric overhang of the G-rich strand, usually restricted to S phase in wild-type cells, is present in Yku cells throughout the entire cell cycle (19).

Using a genetic approach we identified a new mutation in CDC13, designated cdc13-4, that is lethal in combination with a deletion of either subunit of the Yku heterodimer. The telomeres of cdc13-4 mutants shorten continuously for about 150 generations before eventually reaching a stable level comparable to the telomere length seen in Yku mutants. cdc13-4 causes no senescence phenotype, and a cdc13-4/rad52 double mutant is viable for at least several hundred generations. A cdc13-4/tel1 double mutant displays enhanced telomere shortening compared to either single mutant and Y' element amplification after 225 generations of growth. Coimmunoprecipitations reveal that HA₃-Cdc13-4p still associates with GST-Est1p when both proteins are overexpressed. In addition, in a cdc13 strain a Cdc13-4-Est1 fusion protein does not induce telomere elongation to the same extent as a wild-type Cdc13-Est1 fusion. The terminal chromosome configuration of cdc13-4 mutants seems, besides the telomere shortening, unchanged, since no ss G-rich overhang can be detected by native in-gel hybridization. Our data indicate that Cdc13p functions in telomere length regulation independent of its roles in chromosome end protection and telomerase recruitment.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

S. cerevisiae strains, media, growth conditions, and transformation. The strains used in this study are listed in Table 1. Cells were grown at 30°C using yeast extract-peptone-dextrose (YPD), yeast extract-peptone-galactose, or selective media as described elsewhere (14). Screening for synthetic lethal mutations was performed on YPD plates (9). For counterselection plates, 5-fluoroorotic acid (5-FOA) (bts) was added to selective media at a concentration of 1 mg/ml as described previously (9). To examine telomere length and the senescence phenotypes of strains over many generations, colonies derived from freshly germinated spores were streaked on YPD plates. After 48 h of incubation at 30°C, single colonies were restreaked on fresh YPD plates. This procedure was repeated up to nine times. Single colonies from different generations were then used for overnight inoculation and treated for DNA preparation. Yeast transformation was performed by the lithium acetate method (50).

Plasmids. The plasmid pCH-YKU70 used for the synthetic lethality screen was constructed as follows: a XhoI/EcoRI fragment containing a functional YKU70 gene was isolated from the plasmid pRS316-YKU70 (15) and blunted with Klenow enzyme. This fragment was then cloned into pCH1122 (26) linearized with SmaI. Expression of Yku70p from pCH-YKU70 was verified by complementation of the temperature sensitive phenotype of a yku70-deficient strain and by reconstitution of Yku heterodimer DNA binding activity in a gel retardation assay (15). The CDC13 expression plasmid pRS314-CDC13 was generated as follows: a 4.7-kb ApaI fragment containing 712 bp 5' of the start codon, the entire open reading frame (ORF) of CDC13, and 1,200 bp 3' of the stop codon was isolated from the library plasmid GP2a. This fragment was ligated to pRS314 (51), linearized with KpnI/SacI, and blunted with Klenow enzyme. To generate the plasmid pRS314-cdc13-4 expressing the mutated CDC13 allele, a 900-bp DNA fragment was amplified by PCR from genomic DNA of mutant LDM29 by using the primers CDC13-ATG (5'-ACG TGT CGA CCC GGG ATG GAT ACC TAG AAG AGC CTG AG-3') and CDC13-900 (5'-GAA ATA TTT CCC GGT AGA GGA GG-3'). The PCR product was subcloned into pZErO-2 (Invitrogen) and sequenced. A XhoI/NsiI fragment carrying the cdc13-4 point mutation was then excised from pZ-cdc13-4 and ligated to the vector pRS314-CDC13 digested with XhoI/NsiI. To generate several CDC13 disruption constructs, pRS314-CDC13 was digested with XhoI/AatII, thereby deleting the entire ORF of CDC13 except 57 bp at the 5' end. This fragment was replaced by a marker cassette of either KanMX4 or URA3, resulting in plasmid p-cdc13::KanMX4 or p-cdc13::URA3, respectively. The plasmid pRS-cdc13-4-KanMX4 was generated for genomic integration of the cdc13-4 allele by linearizing pRS314-cdc13-4 with AatII, blunting it with Klenow enzyme, and inserting the KanMX4 marker cassette.

Gene disruption. The yku70-deficient strain KL7 was generated by disruption of the YKU70 gene in K2348 as described previously (15). Gene disruption was verified by Southern blot analysis. To disrupt the CDC13 gene, plasmids pRS314-cdc13::URA3 and pRS-cdc13::KanMX4 were digested with ApaLI and KpnI, and the resulting linear disruption construct was used to transform several diploid strains to Ura⁺ or G418 resistance (Table 1). Disruption of the CDC13 gene was verified by Southern blot analysis. The yeast strain BMY13 carrying a genomic integrated cdc13-4 allele was generated by transforming LDY50 using the ApaI/KpnI fragment excised from pRS-cdc13-4-KanMX4. The transformed cells were plated on synthetic-dextrose minimal plates lacking uracil and containing 200 mg of G418/liter. Colonies arising from these plates were screened by PCR for correct integration of the marker gene. To verify the integration of the cdc13-4 point mutation, a PCR fragment spanning the corresponding part of the CDC13 gene was amplified and sequenced. BMY14 (W303a cdc13::URA3/cdc13-4::kanMX4 rad52::His3MX6/RAD52) was generated by replacement of the RAD52 ORF in BMY13 by PCR-based gene disruption (55). Sporulation of BMY13 and BMY14 resulted in haploid spores carrying the cdc13-4 point mutation (BMY17) and double mutant cdc13-4/rad52 (BMY18), respectively. Strain BMY56 was generated by crossing BMY17 with W303a. This strain was propagated for several generations and then used to introduce either a tel1 or est2 deletion. TEL1 was deleted by PCR-based replacement of the entire ORF with a His3MX marker (BMY57), and the EST2 gene was replaced by the TRP1 selection marker (BMY58). Transformants arising after incubation on selective media were screened by PCR for integration of the disruption constructs. Both heterozygous strains were then sporulated, and tetratype tetrads BMY59 and BMY60 were used for growth studies and analysis of telomere length phenotypes. To analyze expression of Cdc13-Est1 fusion proteins in a cdc13 strain, BMY62 was transformed with pRS-CDC13-EST1 or pRS316-cdc13-4-EST1 and sporulated on plates lacking uracil. BMY64 and BMY65 were isolated after tetrad dissection of BMY62+pRS-CDC13-EST1 and BMY62+pRS-cdc13-4-EST1, respectively.

Induced expression of HA₃-CDC13, HA₃-cdc13-4, and GST-EST1. For induced overexpression of HA₃-tagged CDC13 and cdc13-4, the GAL1 promoter together with the HA₃ tag was introduced in front of the genomic copy of CDC13 in W303a or cdc13-4 in BMY56. Integration of GAL1-HA₃ was performed by PCR-based methods as described previously (32) using the HIS3MX6 marker for selection. Correct integration of the HIS3MX-GAL1-HA₃ construct in the resulting strains HFY80 (HA₃-CDC13) and HFY84 (HA₃-cdc13-4) was verified by analytic PCR and sequencing of the PCR product. Expression of HA₃-Cdc13p and HA₃-Cde13-4p was analyzed by Western blotting using monoclonal anti-HA antibody 9F10 (Roche). The same PCR-based strategy was used to generate strains expressing GST::Est1 fusion protein under control of the GAL1 promoter. The TRP1-GAL1-GST construct was introduced in W303a, HFY80, and HFY84 resulting in the strains HFY81 (GST::EST1/EST1), HFY82 (HA₃-CDC13/CDC13 GST::EST1/EST1) and HFY86 (HA₃-cdc13-4/CDC13 GST::EST1/EST1). Correct integration was verified by analytic PCR and sequencing of the PCR product. Expression of GST::Est1p was analyzed by Western blotting using monoclonal anti-GST antibody (Sigma). Strains were grown on yeast extract-peptone media containing 2% galactose for induced expression of HA₃-CDC13, HA₃-cdc13-4, and GST::EST1. Strains HFY81, HFY82, and HFY86 were sporulated to generate haploid strains expressing the tagged Cdc13 and/or Est1 proteins. Tetrad analysis was performed on yeast extract-peptone plates containing galactose to allow expression of HA₃-Cdc13p, HA₃-Cdc13-4, and GST::Est1p. Spores expressing the tagged proteins were identified by marker analysis, and the resulting strains HFY81-8A (GST::EST1), HFY82-6B (HA₃-CDC13), HFY82-4C (HA₃-CDC13 GST::EST1), HFY86-3A (HA₃-cdc13-4 GST::EST1), and HYF86-9D (HA₃-cdc13-4 GST::EST1) were verified by Western blotting.

Immunoprecipitation. Coimmunoprecipitation experiments to analyze the interaction of GST-Est1p-HA₃-Cdc13p and GST-Est1p-HA₃-Cdc13-4p were performed using strains HFY82-4C, HFY86-3A, HFY86-9D, and, as controls, HFY81-8A and HFY82-6B. Crude extracts were prepared as follows: yeast strains were grown overnight in YPGal, diluted to an optical density at 600 nm (OD₆₀₀) of 0.2, and grown to an OD₆₀₀ of 0.8 to 1.2 in yeast extract-peptone-galactose. Cells were lysed in 20 mM Tris (pH 8.0)-200 mM NaCl-1 mM EDTA-1 mM dithiothreitol-0.01% NP-40-10% glycerol with one protease inhibitor cocktail tablet per 5 ml (complete, Mini, EDTA-free; Roche) in a bead beater. After centrifugation the soluble protein fraction was diluted 1:1 with lysis buffer containing 1% NP-40 and 0.2% Triton X-100. Crude extract (1,000 µg) was incubated with monoclonal anti-GST antibody, clone GST-2 (Sigma), for 1 h at 4°C, and then G-Sepharose (Pharmacia) was added. After incubation for 1 h at 4°C G-Sepharose beads were collected by centrifugation and washed twice with lysis buffer containing 0.5% NP-40-0.1% Triton X-100, twice with lysis buffer containing 1% NP-40-0.1% Triton X-100, and twice with lysis buffer containing 450 mM NaCl. The beads were treated with 1,000 U of DNase I/ml in lysis buffer containing 1 mM MgCl₂ and then washed twice with lysis buffer containing 450 mM NaCl and 350 mM potassium acetate. After 15 µl of Laemmli buffer was added, beads were heated 3 min at 95°C and the supernatant was loaded onto an 8% sodium dodecyl sulfate gel. Proteins were visualized by enhanced chemiluminescence Western blotting using anti-HA antibody 9F10 (Roche) and anti-GST antibody clone GST-2 (Sigma).

Synthetic lethality screen. Stationary phase cells of KL7 carrying the plasmid pCH-YKU70 were mutagenized with 3% ethyl methanesulfonate (EMS) for 90 min resulting in 15.6% survival. After EMS treatment, cells were plated on YPD plates containing 4% glucose to facilitate development of the red pigment. Uniformly red colonies were colony purified three times. Those which remained stably red under nonselective conditions were tested for sensitivity to 5-FOA. To test whether 5-FOA-sensitive cells were dependent on YKU70 expression rather than other components of the plasmid pCH-YKU70, the mutants were transformed with a second plasmid, pRS314-YKU70, expressing Yku70p and containing TRP1 for selection. As a control, mutants were transformed with pRS314. Mutants carrying pRS314-YKU70 or pRS314 were retested for their ability to form red-white sectors and their growth on 5-FOA. Out of 20,520 mutagenized cells, five mutants were clearly dependent on YKU70 expression. These mutants were stably red on YPD and sensitive to 5-FOA if transformed with the pRS314 vector control, but they displayed red-white sectoring colonies and growth on 5-FOA after transformation with pRS314-YKU70.

Complementation of YKU70 dependence. The mutant LDM29 was transformed using a single-copy genomic yeast library (ATCC 77164) and plated on Trp media. Out of 12,500 primary transformants, 15 plasmids were isolated leading to red-white sectoring colonies even after retransformation. In addition, these 15 plasmids enabled LDM29 cells to grow on 5-FOA-containing media, indicating that those cells were independent of YKU70 expression. Restriction analysis revealed the isolation of three different genomic fragments capable of complementing the dependence on YKU70. To identify the isolated fragments, the 5' and 3' ends of the fragments were sequenced using vector-specific primers.

Identification of the cdc13-4 mutation. The genomic mutation in LDM29 was mapped by gap repair (42). Plasmid pRS314-CDC13 was digested using different combinations of restriction enzymes. The resulting linear plasmids were transformed into LDM29. Generation of a functional CDC13 gene by gap repair results in cells independent of YKU70 expression, therefore displaying a red-white sectoring phenotype. Only cells transformed with pRS314-CDC13 with a XhoI/NsiI fragment spanning bp +57 to +830 of the CDC13 coding sequence deleted did not display red-white sectoring colonies and were sensitive to 5-FOA, indicating that a plasmid carrying the mutated allele of CDC13 was generated. To identify the mutation, a fragment corresponding to the mutated region in CDC13 was amplified by PCR from genomic DNA of LDM29 and was sequenced.

Yeast DNA extraction and analysis of telomeric DNA. Genomic DNA was isolated from 5- to 7-ml overnight cultures using the nucleon MiY DNA extraction kit (Amersham Life Science). For analysis of telomere length, genomic DNA was digested overnight using XhoI and separated on a 1% agarose gel in 1× Tris-acetate-EDTA buffer. DNA was transferred to nylon membranes (HybondN⁺) by vacuum blotting using 0.4 N NaOH. Detection of telomeric DNA fragments was performed as described elsewhere (2). Nondenaturing in-gel hybridization was performed as described previously (11).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of the cdc13-4 mutant. Yku mutant cells are temperature sensitive for growth (14, 15). To investigate the essential role of the Yku heterodimer at 37°C, we performed a synthetic lethality screen to isolate mutants in which YKU70 would be essential for viability. Therefore we disrupted the YKU70 gene in K2348 and tested the resulting mutant KL7 for phenotypes specific for Yku mutants. KL7 is temperature sensitive for growth at 37°C, deficient in nonhomologous end joining, slightly sensitive to methyl methanesulfonate, and displays shortened telomeres (data not shown). The YKU70 gene cloned into plasmid pCH1122 (pCH-YKU70) complemented the phenotypes of KL7, indicating a functional expression of YKU70 from the plasmid. KL7-pCH-YKU70 colonies grown on YPD displayed a red-white sectoring phenotype, showing that pCH-YKU70 was not essential for growth at 30°C under nonselective conditions.

Synthetic lethality of cdc13-4 with Yku. We isolated the cdc13-4 mutation in a synthetic lethality screen with YKU70. To verify the synthetic lethal phenotype we reintroduced the cdc13-4 mutation in the homozygous yku70 strain WaLU. Therefore we disrupted the CDC13 gene in WaUL and transformed the resulting strain LDY46, heterozygous for CDC13, with pRS-cdc13-4. As expected, we obtained only two colony-forming spores after sporulation and tetrad dissection of LDY46-pRS-cdc13-4 (data not shown). None of the viable spores was resistant to G418 (the KanMX marker gene was used for CDC13 disruption), indicating that all viable spores contain the wild-type allele of CDC13. The nonviable spores were examined by microscopy. We found many of these spores germinated but arrested at a two-cell stage. In a very few cases we observed microcolonies containing up to 20 cells, which lysed after 2 to 3 days of incubation at 30°C. To show that these phenotypes were not due to synthetic effects caused by the RAD5 mutation in the W303 background (13), we repeated the experiment in a CEN.PK2 strain. In this case we examined the synthetic lethality of cdc13-4 in a yku70- and a yku80-deficient CEN.PK2 strain, LDY54 and LDY55, respectively.

View larger version (67K): [in this window] [in a new window]   FIG. 1.   Synthetic lethality of cdc13-4 with the Yku heterodimer. Diploid strains CEN.PK2a yku70/yku70 cdc13/CDC13 and CEN.PK2a yku80/yku80 cdc13/CDC13 were transformed with plasmid pRS314-cdc13-4. Transformants were sporulated and tetrads were dissected using a Singer SMS Micromanipulator. Individual spores of each tetrad were placed down the columns on the YPD plates and incubated at 30°C for 3 to 4 days.

Telomeres of cdc13-4 mutants shorten to a steady-state level. To investigate the phenotype of a cdc13-4 single mutant we generated a haploid strain expressing the mutated CDC13 gene (LDY53). One allele of CDC13 was deleted in W303a, and the resulting heterozygous strain LDY50 was transformed using pRS314-cdc13-4. After sporulation and tetrad dissection, some tetrads were able to form three viable colonies (data not shown). Since disruption of CDC13 is lethal, the tetrads resulting in three colony-forming spores should contain one spore carrying a disrupted cdc13 allele and the plasmid expressing cdc13-4. All three tetrads tested formed two colonies unable to grow on uracil- or trytophan-lacking media and exhibited wild-type fragment size in a Southern blot. One colony was prototrophic for uracil and tryptophan and displayed a disrupted genomic CDC13 allele and Southern blot signals corresponding to the plasmid pRS314-cdc13-4 (data not shown). This colony corresponds to the cdc13-4 mutant LDY53.

View larger version (75K): [in this window] [in a new window]   FIG. 2.   Long-term analysis of the telomere length of a cdc13-4 mutant. Southern blot of genomic yeast DNA, probed with a telomere-specific poly(GT)20 oligonucleotide, is shown. The bracket indicates the telomeric GT repeat band derived from Y' element-containing chromosomes. Asterisks indicate terminal fragments derived from non-Y' element-containing chromosomes. W303 wild-type (wt) and W303 cdc13::URA3 + pRS314-cdc13-4 strains from one tetrad were propagated on YPD for 250 generations. Therefore colonies derived from freshly germinated spores were streaked on YPD plates. After 48 h of incubation at 30°C, single colonies were restreaked on fresh YPD plates. Cells were estimated to have undergone 20 to 25 divisions per streakout. Numbering at the top of the lanes (1x, 3x, etc) indicates the number of times of restreaking. Single colonies from different generations were then used for overnight inoculation and treated for DNA preparation. Genomic DNA was prepared as described in Materials and Methods. Lane 1, cdc13-4, 50 generations; lane 2, cdc13-4, 100 generations; lane 3, cdc13-4, 150 generations; lane 4, cdc13-4, 200 generations; lane 5, cdc13-4, 250 generations; lane 6, W303 wt, 50 generations; lane 7, W303 wt, 150 generations; lane 8, W303 wt, 250 generations; and lane 9, W303a yku70.

cdc13-4 mutants display no senescence phenotype. Telomerase deficiency results in replicative senescence. Telomeres shorten gradually with increasing generations, eventually leading to cell death (28, 34). However, a few survivors can arise in a senescent yeast culture. These survivors stabilize their telomeres by homologous recombination, adding Y' elements or GT repeats to the shortened chromosome ends (33). This process is detectable by an increase in intensity of the Y' element signals in a Southern blot. Deletion of RAD52 completely abolishes homologous recombination, and therefore no survivors appear in an est2/rad52-negative strain (27). To verify the observation that cdc13-4 mutant cells display shortened telomeres but no senescence phenotype, we generated the diploid strain BMY58, heterozygous for est2 and cdc13-4 mutation. After sporulation we compared growth of an est2 spore and a cdc13-4 mutant spore (Fig. 3).

View larger version (75K): [in this window] [in a new window]   FIG. 3.   Viability of cdc13-4, est2, and cdc13-4 est2 strains. After sporulation, cdc13-4, est2, and cdc13-4/est2 mutant cells from a single tetrad were successively streaked on YPD plates to test senescence.

View larger version (70K): [in this window] [in a new window]   FIG. 4.   Survivor formation in cdc13-4 and est2 mutants. Southern blot of XhoI-digested genomic yeast DNA probed with a poly(GT)20 oligonucleotide specific for telomeric repeats is shown. The bracket indicates the telomeric GT repeat band derived from Y' element-containing chromosomes. Asterisks indicate terminal fragments derived from non-Y' element-containing chromosomes. The arrows indicate restriction fragments corresponding to the subtelomeric Y' elements. After tetrad dissection, spores W303 cdc13-4 and W303 est2 from one tetrad were grown for 150 generations as described in Materials and Methods. Lane 1, W303a wild type (wt); lane 2, est2, 25 generations; lane 3, est2, 50 generations; lane 4, est2, 75 generations; lane 5, est2, 100 generations; lane 6, est2, 125 generations; lane 7, cdc13-4, 25 generations; lane 8, cdc13-4, 50 generations; lane 9, cdc13-4, 75 generations; lane 10, cdc13-4, 100 generations; lane 11, cdc13-4, 125 generations.

Cdc13p and Tel1p function in different pathways of telomere length maintenance. The synthetic lethality of the cdc13-4 mutation with a yku70 or yku80 deletion indicates that Cdc13p and the Yku heterodimer have independent but in some way overlapping functions at the telomere. Along with the Yku heterodimer and Cdc13p, a pathway comprised of Tel1p and the Mre11p-Xrs2p-Rad50p complex is involved in telomere length maintenance (47). To investigate if CDC13 is epistatic to TEL1 we generated the diploid strain BMY57, heterozygous for tel1 and cdc13-4. Telomeres of BMY57 cells were shorter than those of the diploid wild type (Fig. 5, compare lanes 1 and 2), indicating that reduced protein levels in the heterozygous strain already influence telomere length maintenance. For further analysis we used all four spores derived from a tetratype tetrad.

View larger version (87K): [in this window] [in a new window]   FIG. 5.   Telomere length of cdc13-4, tel1, and cdc13-4/tel1 mutants. Southern blot of genomic yeast DNA, probed with a telomere-specific poly(GT)20 oligonucleotide, is shown. Spores from a tetratype tetrad of BMY57 were propagated for 225 generations as described in Materials and Methods. The bracket indicates the telomeric GT repeat band derived from Y' element-containing chromosomes. Asterisks indicate terminal fragments derived from non-Y' element-containing chromosomes. The arrows indicate restriction fragments corresponding to the subtelomeric Y' elements. Lane 1, W303a wild type; lane 2, W303a cdc13-4/CDC13 tel1/TEL1 (BMY57); lane 3, W303 tel1, 50 generations; lane 4, W303 tel1, 125 generations; lane 5, W303 tel1, 225 generations; lane 6, W303a cdc13-4, 50 generations; lane 7, W303a cdc13-4, 125 generations; lane 8, W303a cdc13-4, 225 generations; lane 9, W303 tel1/cdc13-4, 50 generations; lane 10, W303 tel1/cdc13-4, 125 generations; lane 11, W303 tel1/cdc13-4, 225 generations; lane 12, W303a wild type, 50 generations; lane 13, W303a wild type, 225 generations.

Cdc13-4p is not altered in its binding to Est1p. Expression of a Cdc13-Est1 fusion protein complements a cdc13 or est1 deletion and, moreover, results in a dramatic increase in telomere length (12). These data suggest that the telomere-bound Cdc13p recruits telomerase via interaction with Est1p to the ssDNA overhang at chromosome ends. To examine if a reduced association with Est1 causes the telomere shortening phenotype of a cdc13-4 mutant, we analyzed the effect of expressing a Cdc13-4-Est1 fusion on telomere length. Therefore, a Cdc13-Est1 or Cdc13-4-Est1 fusion protein was expressed under the control of the CDC13 promoter from a single-copy plasmid in wild-type, cdc13-4, and cdc13 cells.

View larger version (78K): [in this window] [in a new window]   FIG. 6.   Influence of Cdc13-Est1 and Cdc13-4-Est1 fusion proteins on telomere length. (A) Southern blot of W303a wild-type (wt) and W303a cdc13-4 strains transformed with plasmids pRS316, p-CDC13-EST1, and p-cdc13-4-EST1 probed with a poly(GT)20 oligonucleotide. Transformants were cultured for 100 generations on selective media prior to DNA preparation. Lane 1, W303a wt + pRS316 control; lane 2, W303a wt + p-CDC13-EST1; lane 3, W303a wt + p-cdc13-4-EST1; lane 4, W303a cdc13-4 + pRS316 control; lane 5, W303a cdc13-4 + p-CDC13-EST1; lane 6, W303a cdc13-4 + p-cdc13-4-EST1. (B) Southern blot of W303a cdc13 strains transformed with plasmids p-CDC13-EST1 and p-cdc13-4-EST1 probed with a poly(GT)20 oligonucleotide. Transformants were cultured for 25 or 100 generations prior to DNA preparation.

View larger version (29K): [in this window] [in a new window]   FIG. 7.   Coimmunoprecipitation of HA-cdc13-4 with GST-Est1. Ten-microgram samples of crude extracts were loaded to compare the protein amounts of different mutants used for coimmunoprecipitation experiments (lanes 1 to 5). Coimmunoprecipitation of Cdc13p and Est1p was performed as described in Materials and Methods (lanes 6 to 10). Crude extract (1,000 µg) was incubated with 5 µg of anti-GST antibody, and G-Sepharose beads were used to isolate antibody and bound proteins. After intensive washing, G-Sepharose beads were heated to 95°C in Laemmli buffer and the supernatant was loaded onto an 8% sodium dodecyl sulfate gel. (A) Lane 1, HFY86-3A (HA-cdc13-4p, GST-Est1p); lane 2, HFY81-8A (Cdc13p, GST-Est1p); lane 3, HFY82-6B (HA-Cdc13p, Est1p); lane 4, HFY82-4C (HA-Cdc13p, GST-Est1p); lane 5, HFY86-9D (HA-cdc13-4p, GST-Est1p); lane 6, HFY86-3A (HA-cdc13-4, GST-EST1); lane 7, HFY81-8A (Cdc13p, GST-Est1p); lane 8, HFY82-6B (HA-Cdc13p, Est1p); lane 9, HFY82-4C (HA-Cdc13p, GST-Est1p); lane 10, HFY86-9D (HA-cdc13-4p, GST-Est1p). HA-Cdc13p and HA-Cdc13-4p were detected by anti-HA antibody. (B) The same blots as in panel A probed with an anti-GST antibody. Note that lanes 6 to 10 were exposed a significantly shorter time to detect the GST-Est1p signals than lanes 1 to 5 and blots probed with anti-HA antibody in panel A. Ab, antibody.

cdc13-4 mutation seems not to affect DNA binding. Very recently, different mutant alleles of CDC13 that cause stably shortened telomeres comparable to the cdc13-4 mutation have been described. These mutant Cdc13 proteins seem to display significantly reduced binding activity to telomeric DNA (18). Although expression of Cdc13-Est1p in cdc13-4 cells indicates that Cdc13-4p and Cdc13p compete for telomere binding, we wanted to determine if the DNA binding activity of Cdc13-4p is reduced compared to that of wild-type Cdc13p. Assuming that overexpression of Cdc13-4p should complement a reduced DNA binding activity, we analyzed telomere length in the diploid strain HFY82 expressing one wild-type copy of CDC13 and one copy of HA₃-cdc13-4 under control of the inducible GAL1 promoter.

View larger version (37K): [in this window] [in a new window]   FIG. 8.   Overexpression of cdc13-4 in a heterozygous mutant strain. Diploid strains W303a, BMY56 (cdc13-4/CDC13), and BMY17 (cdc13-4) were grown on galactose-containing media, whereas HFY84 (GAL1-cdc13-4/CDC13) and HFY86 (GAL1-cdc13-4/CDC13 GAL1-EST1/EST1) were simultaneously grown under inducing and noninducing conditions. Telomere length was investigated after 50 generations. Lane 1, W303a; lane 2, BMY17 (cdc13-4); lane 3, BMY56 (cdc13-4/CDC13); lane 4, HFY84 (GAL1-cdc13-4/CDC13) grown on glucose; lane 5, HFY84 (GAL1-cdc13-4/CDC13) grown on galactose; lane 6, HFY86 (GAL1-cdc13-4/CDC13 GAL1-EST1/EST1) grown on glucose; lane 7, HFY86 (GAL1-cdc13-4/CDC13 GAL1-EST1/EST1) grown on galactose. Brackets indicate terminal GT repeats, asterisks indicate non-Y' elements, and the arrows represent Y' element bands.

View larger version (78K): [in this window] [in a new window]   FIG. 9.   Telomeric end structure in cdc13-4 mutant cells. W303 wild-type (wt) and W303 cdc13-4 spores derived from one tetrad were streaked on YPD for 250 generations and used for genomic DNA preparation as described in Materials and Methods. (Left panel) After XhoI digestion, genomic DNA was separated by gel electrophoresis and analyzed by nondenaturing in-gel hybridization using a 22-mer C1-3A oligonucleotide as a probe. The arrow indicates terminal restriction fragments derived from Y' element-containing chromosomes. The strong signal in lane 6 corresponds to the elongated ssDNA overhang in yku80-deficient cells. Lane 1, 1-kbp ladder DNA; lane 2, W303 wt, 40 generations; lane 3, W303 wt, 260 generations; lane 4, W303 cdc13-4, 40 generations; lane 5, W303 cdc13-4, 260 generations; lane 6, yku80 mutant; lane 7, control ssDNA; lane 8, control double-stranded DNA. (Right panel) The same gel as in the left panel, after denaturation of the DNA in the gel and rehybridization to the same probe. The bracket indicates the telomeric GT repeat band derived from Y' element-containing chromosomes. Asterisks indicate terminal fragments derived from non-Y' element-containing chromosomes.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

S. cerevisiae CDC13 is an essential gene involved in chromosome end replication and protection. The cdc13-4 allele, which we isolated in a synthetic lethality screen with YKU70, causes a dramatic shortening of GT repeats at the telomeres but the strain remains viable. Telomere shortening proceeds slowly over approximately 150 generations; however, telomere length is stabilized at a short level after 200 generations (Fig. 2). This telomere phenotype is distinct from the senescence phenotype of a cdc13-2^est allele, which leads to progressive telomere shortening and eventually cell death (40). In a senescent yeast culture a few cells occasionally escape from cell death. These survivors stabilize their telomeres by adding either tandem copies of the subtelomeric Y' elements or C_1-3A/TG_1-3 repeats in a RAD52-dependent recombination process (53). cdc13-4 mutants do not display Y' element amplification in a Southern blot as observed in survivors of telomerase-negative yeast strains (Fig. 2 and 4). In addition, a cdc13-4/rad52 double mutant strain is viable for more than 250 generations while maintaining short telomeres (data not shown). Therefore, cdc13-4 mutants do not show characteristics of a senescent mutant and telomeres do not reach the critical length level which triggers telomere stabilization by homologous recombination. Compared to that of an est2 mutant, the rate of telomere shortening in a cdc13-4 mutant is clearly reduced (Fig. 4), indicating that telomerase activity is altered but not abolished. The stabilization of telomere length at a shorter level shows that telomerase is fully active at the new equilibrium length.

Mutations in TEL1 and TEL2 have been reported to cause a progressive telomere shortening phenotype comparable to that of cdc13-4. Telomeres in tel1-1 and tel2-1 mutants shorten to a stable level within 150 generations (35, 48), and a tel1-1/tel2-1 double mutant has no telomeres shorter than those of tel1-1 cells. This suggests that Tel1p and Tel2p function in the same pathway of telomere maintenance (25, 35). In contrast, a cdc13-4/tel1 double mutant displays slightly shorter GT repeat tracts after 125 generations compared to those of a tel1 or a cdc13-4 single mutant. In addition, Y' elements are amplified in the double mutant after 225 generations (Fig. 5), indicating that telomeres have shortened to a critical level. These data point toward a function of CDC13 in telomere maintenance independent of the TEL1 pathway.

The cdc13-4 mutation is synthetically lethal with yku70 or yku80 (Fig. 1). This might be explained by a reduced telomere capping ability of the Cdc13-4 protein, which becomes essential at the elongated ssDNA overhang in Yku mutants (19). This would then lead to the degradation of chromosome ends and cell cycle arrest. However, our data do not support such a model. Formation of microcolonies from double mutant spores (Fig. 1) makes it more likely that accelerated senescence is the reason for the synthetic lethality. Telomeres in Yku mutants are shortened severely, and any further GT repeat tract shortening by the cdc13-4 mutation would result in reaching a lethal level within a few generations.

Cdc13p has been proposed to control the susceptibility of chromosome ends to the specific degradation of the telomeric C_1-3A strand at the end of S phase (57), and therefore a reduced DNA binding activity of Cdc13-4p could possibly cause a progressive telomere shortening as seen in cdc13-4 cells. The cdc13-4 mutation at position 235 is not located in the DNA binding domain of Cdc13p (23) (Fig. 10), although this does not exclude a conformational change in the Cdc13-4 mutant protein resulting in reduced DNA binding activity. Cdc13p protects chromosome ends from degradation and thereby prevents the generation of telomeric ssDNA (16, 40). Therefore we would expect at least a slight increase in ssDNA at the telomeres in cdc13-4 cells if Cdc13-4p is reduced in DNA binding. However, native in-gel hybridization experiments revealed no increase in ssDNA formation in cdc13-4 mutants (Fig. 9).

View larger version (38K): [in this window] [in a new window]   FIG. 10.   Functional domains and selected mutations of CDC13 mutants. AA, amino acid; n.d., not determined.

The expression of a Cdc13-Est1 or Cdc13-Est2 fusion protein in cdc13 strains has been shown to complement for telomerase deficiency and additionally results in strongly elongated telomeres. Therefore, the expression of a mutant Cdc13-4-Est1 fusion protein should exhibit telomere elongation comparable to that of a wild-type Cdc13-Est1 fusion protein if DNA binding of Cdc13-4p is not reduced. In fact, telomere elongation was detected in strains expressing the mutant Cdc13-4-Est1 or the wild-type Cdc13-Est1 fusion protein (Fig. 6). Furthermore, GT repeat tract length in cdc13-4 cells expressing the wild-type Cdc13-Est1 fusion protein, although significantly elongated, did not reach the wild-type level after 100 generations (Fig. 6B), suggesting that endogenous Cdc13-4p can compete with the Cdc13-Est1 fusion protein for telomere binding, thereby partially preventing telomere elongation.

Additional evidence that DNA binding is unchanged is provided by the finding that heterozygous CDC13/cdc13-4 diploid yeast strains show reduced telomere length (Fig. 8), indicating an at least partially dominant phenotype of the cdc13-4 mutation. Telomere shortening in the heterozygous CDC13/cdc13-4 diploid strains is not caused by a reduced amount of functional Cdc13p since a CDC13/GAL1-HA₃-cdc13-4 strain exhibits wild-type telomere length (Fig. 8) on glucose where expression of HA₃-cdc13-4 by the GAL1 promoter is repressed. The dominant phenotype of Cdc13-4p is even more pronounced if overexpression of HA₃-cdc13-4 is induced in the heterozygous diploid (Fig. 8). This again indicates that Cdc13-4p competes with wild-type Cdc13p for telomere binding. Therefore, we present evidence that the mutant Cdc13-4 protein is capable of chromosome end binding with an affinity comparable to that of the wild-type Cdc13p. An alternative explanation for the partially dominant phenotype of the cdc13-4 mutation would be a competition of wild-type Cdc13p and mutant Cdc13-4p for a protein important for telomere elongation. Further experiments have to be performed to address this question.

The recruitment of telomerase to chromosome ends seems to take place via the interaction of Cdc13p and Est1p. Therefore, an attenuated interaction of Cdc13-4p and Est1p could cause telomere shortening to a stable level. The telomerase recruitment site of Cdc13p was recently mapped to amino acids 211 to 331 (43). The cdc13-4 mutation (P235S) is located near the border of this domain; thus, the interaction of the mutant Cdc13-4 protein and Est1p might be reduced. Nevertheless, we found no reduced interaction of Cdc13-4p-Est1p in coimmunoprecipitation experiments (Fig. 7). In addition, overexpression of Cdc13-4p or cooverexpression of Cdc13-4p and Est1p induces telomere shortening in a heterozygous diploid strain (Fig. 8) and did not complement the cdc13-4 mutation as we would suggest for a weakened interaction.

Significantly elongated telomeres, most likely the result of unregulated access of the active telomerase complex to the telomere, are detectable in yeast strains expressing a Cdc13-Est1 fusion protein (12). Although the Cdc13-4-Est1 fusion protein causes a dramatic telomere elongation in a cdc13 strain, the effect is not as pronounced as that observed for a wild-type Cdc13-Est1 fusion. Therefore, establishing a permanent interaction of Cdc13-4p and Est1p is not sufficient to complement the cdc13-4 mutation to wild-type level, indicating that a function independent of interaction with Est1p is affected in Cdc13-4p. The DNA binding domain of Cdc13p has been mapped to an internal part of the protein (23); nevertheless, the N-terminal 251 amino acids of Cdc13p associate in vivo with the telomere (5), indicating tight interaction with telomere bound proteins. This N-terminal domain partially overlaps the telomerase recruiting domain of Cdc13p (43) (Fig. 10) but seems not to be sufficient for Cdc13p-Est1p interaction. Thus, the cdc13-4 mutation might influence interaction with other proteins at the telomere, thereby preventing appropriate activation of telomerase activity.

In S. cerevisiae, telomere length seems to be maintained by the balance of two antagonistic processestelomere elongation and telomere shortening. Many proteins are necessary to maintain normal telomere length. Deletion of one Yku subunit (3, 44) or inactivation of a member of the TEL1 pathway, comprised of Tel1p, Mre11p, Xrs2p, and Rad50p, leads to telomere shortening to a stable level (2, 20). The additional telomere shortening seen in yku70/tel1 or yku70/rad50 double mutants indicates that the Yku heterodimer has a TEL1-independent role in telomere maintenance (47). In cdc13-4 mutant cells, telomerase seems to be inactive at normal telomere length, indicating that Cdc13p is involved in positive telomere length regulation by activating telomerase at short GT repeat levels. The further telomere shortening seen in cdc13-4/tel1 double mutants and the synthetic lethality of cdc13-4 with a Yku subunit deletion lead to the conclusion that at least three independent pathways are involved in positive telomere length regulation and that Ccd13p is an essential part of one of these pathways.

However, the addition of telomeric GT repeats to telomeric ends depends not only on telomerase but also on DNA polymerases Pol , Pol , and DNA primase, most likely by a coordinated regulation of C- and G-strand synthesis (10). Recently it has been shown that Cdc13p interacts with Pol1p, the catalytic subunit of DNA polymerase . Single point mutations in either CDC13 or POL1 that weaken the interaction of Cdc13p with Pol1p result in telomerase-dependent telomere lengthening (45). Therefore Cdc13p also seems to play an important role in negative telomere length control, presumably by coordinating telomeric C- and G-strand synthesis.

Until now three different functions of Cdc13p in telomere maintenance have been defined by CDC13 mutations (Fig. 10): (i) protection of chromosome ends from nucleolytic degradation (abolished in a cdc13-1^ts mutant at the restrictive temperature), (ii) loading of telomerase onto the ssDNA overhang at the telomere (prevented in cdc13-2^est cells), and (iii) regulation of telomere length. The role of Cdc13p in telomere length control seems to be multifaceted, since mutating CDC13 can cause either telomere lengthening, seen in cdc13-50 mutants (45) and different mutant CDC13 alleles (18), or telomere shortening to a new steady-state level, seen in newly identified CDC13 mutants (18) and the cdc13-4 mutant reported here. Our data present evidence that Cdc13p plays a key role not only in recruiting telomerase but also in modulating its access to the telomere, which might be influenced by additional regulatory proteins.