ABSTRACT
Two roles for the Saccharomyces cerevisiae Cdc13 protein at the telomere have previously been characterized: it recruits telomerase to the telomere and protects chromosome ends from degradation. In a synthetic lethality screen with YKU70, the 70-kDa subunit of the telomere-associated Yku heterodimer, we identified a new mutation in CDC13, cdc13-4, that points toward an additional regulatory function of CDC13. AlthoughCDC13 is an essential telomerase component in vivo, no replicative senescence can be observed in cdc13-4 cells. Telomeres of cdc13-4 mutants shorten for about 150 generations until they reach a stable level. Thus, incdc13-4 mutants, telomerase seems to be inhibited at normal telomere length but fully active at short telomeres. Furthermore, chromosome end structure remains protected in cdc13-4mutants. Progressive telomere shortening to a steady-state level has also been described for mutants of the positive telomere length regulator TEL1. Strikingly, cdc13-4/tel1Δ double mutants display shorter telomeres than either single mutant after 125 generations and a significant amplification of Y′ elements after 225 generations. Therefore CDC13, TEL1, and the Yku heterodimer seem to represent distinct pathways in telomere length maintenance. Whereas several CDC13 mutants have been reported to display elongated telomeres indicating that Cdc13p functions in negative telomere length control, we report a new mutation leading to shortened and eventually stable telomeres. Therefore we discuss a key role of CDC13 not only in telomerase recruitment but also in regulating telomerase access, which might be modulated by protein-protein interactions acting as inhibitors or activators of telomerase activity.
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 yeastSaccharomyces cerevisiae the telomere repeats consist of 300 ± 75 bp of C1–3A/TG1–3 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 byEST1, 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, C1–3A/TG1–3 repeat sequences (5). Very recently the DNA binding domain of Cdc13p has been mapped to amino acids 557 to 694. Heterologous expression inEscherichia 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.cdc13-2est mutant cells exhibit a senescence phenotype but can be rescued by expression of a Cdc13-2est–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 severalCDC13 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 inCDC13, 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 acdc13-4/rad52 double mutant is viable for at least several hundred generations. A cdc13-4/tel1Δ 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 cdc13Δ 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
S. cerevisiae strains, media, growth conditions, and transformation.The strains used in this study are listed in Table1. 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).
Yeast strains used in this studya
Plasmids.The plasmid pCH-YKU70 used for the synthetic lethality screen was constructed as follows: aXhoI/EcoRI fragment containing a functionalYKU70 gene was isolated from the plasmid pRS316-YKU70 (15) and blunted with Klenow enzyme. This fragment was then cloned into pCH1122 (26) linearized withSmaI. Expression of Yku70p from pCH-YKU70 was verified by complementation of the temperature sensitive phenotype of ayku70-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) ofCDC13, 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 thecdc13-4 point mutation was then excised from pZ-cdc13-4 and ligated to the vector pRS314-CDC13 digested withXhoI/NsiI. To generate several CDC13disruption constructs, pRS314-CDC13 was digested withXhoI/AatII, thereby deleting the entire ORF ofCDC13 except 57 bp at the 5′ end. This fragment was replaced by a marker cassette of either KanMX4 or URA3, resulting in plasmid p-cdc13Δ::KanMX4 or p-cdc13Δ::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.
To generate CDC13-EST1 and cdc13-4–EST1 fusion constructs, the EST1 gene was amplified from genomic DNA from strain W303a using primers Est1SacI (forward: 5′-GAG CTC ATG GAT AAT GAA GAA GTT AAC G-3′) and Est1SalISmaI (reverse: 5′-GTC GAC CCC GGG TCA AGT AGG AGT ATC TGG CAC-3′). A C-terminal fragment ofCDC13 was amplified using primers Cdc13-P3 (5′-CTG GTG CCA GGC GTC AAT TGC-3′) and Cdc13-P4Sma (5′-ATC CCG GGC GAG GTG GGA ACG GCT CCG-3′) and cloned into plasmid pZErO-2. TheEST1 fragment was digested usingSmaI/HpaI and ligated into pZ-CDC13-P3P4Sma linearized with SmaI. The correct orientation of the construct was verified by restriction analysis. The plasmid was then cut with SacII/PstI, and the DNA fragment containing C-terminal CDC13-EST1 was isolated. pRS314-CDC13 was digested with SacII/NotI, and a 3.1-kb fragment containing the N-terminal part of CDC13 and theCDC13 promoter was isolated. Both fragments were then ligated to pRS314 NotI/PstI, resulting in pRS314-CDC13-EST1. To delete the HindIII vector site, pRS314-CDC13-EST1 was cut with PstI/ApaI, treated with T4 polymerase, and religated. The religated vector was cut withNotI/KpnI, and the CDC13-EST1 fragment was isolated and ligated to pRS316 NotI/KpnI to obtain pRS316-CDC13-EST1. The plasmid pRS316-cdc13-4-EST1 was generated by restriction of pRS316-CDC13-EST1 withHindIII and replacing the resulting internalCDC13 HindIII fragment by the corresponding cdc13-4 HindIII fragment. The correct orientation of the cdc13-4 HindIII fragment was checked by restriction analysis, and sequencing confirmed the single base pair exchange in pRS316-cdc13-4-EST1.
Gene disruption.The yku70-deficient strain KαL7 was generated by disruption of the YKU70 gene in K2348α as described previously (15). Gene disruption was verified by Southern blot analysis. To disrupt the CDC13gene, plasmids pRS314-cdc13Δ::URA3 and pRS-cdc13Δ::KanMX4 were digested with ApaLI andKpnI, 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 (W303aαcdc13::URA3/cdc13-4::kanMX4 rad52Δ::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/rad52Δ (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 acdc13Δ 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 andcdc13-4, the GAL1 promoter together with the HA3 tag was introduced in front of the genomic copy ofCDC13 in W303aα 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. TheTRP1-GAL1-GST construct was introduced in W303aα, HFY80, and HFY84 resulting in the strains HFY81 (GST::EST1/EST1), HFY82 (HA3-CDC13/CDC13GST::EST1/EST1) and HFY86 (HA3-cdc13-4/CDC13GST::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-cdc13-4GST::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 KαL7 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 YKU70expression 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 YKU70expression. 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 usingXhoI 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
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 K2348α and tested the resulting mutant KαL7 for phenotypes specific for Yku−mutants. KαL7 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 KαL7, indicating a functional expression of YKU70 from the plasmid. KαL7-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.
After EMS mutagenesis of KαL7-pCH-YKU70 we isolated five stably red mutants, which clearly required YKU70 expression for viability (see Materials and Methods). To identify the mutated gene causing the requirement for YKU70 expression we transformed one mutant, LDM29, with a single-copy yeast library and screened for sectoring colonies indicating that pCH-YKU70 was no longer essential for viability. Plasmids isolated from 15 sectoring colonies revealed three independent clones, two of them carrying a DNA fragment containing the full-length YKU70 gene. The third plasmid, GP2a, contained a fragment of chromosome IV from YDL57269 to YDL68607. This fragment encoded five ORFs, among them YDL220c coding forCDC13/EST4.
Cdc13p, like the Yku heterodimer, has been shown to be an important factor for telomere maintenance. Therefore we subcloned theCDC13 gene from plasmid GP2a into the single-copy vector pRS314 (51). After transformation with the resulting plasmid pRS314-CDC13, LDM29 displayed a clear sectoring phenotype indicating that a mutation in CDC13 caused dependence on Yku70p expression (data not shown). Using the gap repair method (42) we identified a 773-bp fragment near the 5′ end of the CDC13 gene carrying the mutation. Sequencing of this fragment revealed the presence of a single point mutation (at position 703, changing a cytosine to a thymine), thereby leading to the amino acid exchange proline 235 to serine (P235S). Since this mutation differs from the CDC13 mutants already described in the literature, we designated it cdc13-4.
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 homozygousyku70 strain WaLαU. Therefore we disrupted theCDC13 gene in WaUαL 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 ayku80-deficient CEN.PK2 strain, LDY54 and LDY55, respectively.
Again we found only two colony-forming spores for most of the dissected tetrads. In some cases one or two microcolonies arose (Fig.1). Cells from these microcolonies were not viable after restreaking on YPD plates (data not shown). Our data show that cdc13-4 is synthetic lethal with eitheryku70 or yku80 deletion. Therefore we suggest that a cdc13-4 mutant is dependent on a functional Yku heterodimer.
Synthetic lethality of cdc13-4 with the Yku heterodimer. Diploid strains CEN.PK2aα yku70/yku70 cdc13/CDC13 and CEN.PK2aα yku80/yku80 cdc13/CDC13were 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 mutatedCDC13 gene (LDY53). One allele of CDC13 was deleted in W303aα, 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 expressingcdc13-4. 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 genomicCDC13 allele and Southern blot signals corresponding to the plasmid pRS314-cdc13-4 (data not shown). This colony corresponds to thecdc13-4 mutant LDY53.
One important role of Cdc13p in telomere maintenance is loading telomerase to its ss template at chromosome ends. Incdc13-2est mutants the loading function is abolished, presumably by inhibition of the Cdc13p-Est1p interaction, thereby resulting in progressive telomere shortening and senescence (40). To investigate the effect of the cdc13-4mutation on telomere stability we performed long-term growth experiments using strain LDY53. No growth reduction was observed for the cdc13-4 mutant for more than 250 generations (data not shown), suggesting that this mutation causes no senescence phenotype. To better understand the effect of the cdc13-4 mutation, we examined telomere length in this mutant after various generations (Fig.2).
Long-term analysis of the telomere length of acdc13-4 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 W303cdc13::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 cells displayed a significant shortening of the telomeric GT repeat tracts after approximately 50 generations (Fig. 2, lane 1). After 150 generations (Fig. 2, lane 3), telomeric GT repeat tracts were almost as short as those observed foryku70-deficient strains (Fig. 2, lane 9). However, no further telomere shortening was observed after an additional 100 generations (Fig. 2, lane 5) and after several hundred generations (data not shown). Introduction of the cdc13-4 mutation in a CEN.PK2 genetic background resulted in a comparable telomere length phenotype (data not shown).
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 anest2/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 est2Δ and cdc13-4mutation. After sporulation we compared growth of an est2Δ spore and a cdc13-4 mutant spore (Fig.3).
Viability of cdc13-4, est2Δ, andcdc13-4 est2Δ strains. After sporulation, cdc13-4, est2Δ, and cdc13-4/est2Δ mutant cells from a single tetrad were successively streaked on YPD plates to test senescence.
Whereas est2-negative cells displayed a significant growth reduction after 50 generations and survivor formation occurred after 75 generations, cdc13-4 mutant cells grew normally over the entire time frame tested (Fig. 3). We analyzed telomere repeat sequences in BMY59-6C (est2Δ) and BMY59-6B (cdc13-4) cells by Southern blotting after growth for 25, 50, 75, 100, and 125 generations (Fig.4).
Survivor formation in cdc13-4 andest2Δ 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 est2Δ from one tetrad were grown for 150 generations as described in Materials and Methods. Lane 1, W303a wild type (wt); lane 2,est2Δ, 25 generations; lane 3, est2Δ, 50 generations; lane 4, est2Δ, 75 generations; lane 5,est2Δ, 100 generations; lane 6, est2Δ, 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.
As expected, telomeric GT repeat tracts shortened dramatically in anest2Δ mutant within 50 generations (Fig. 4, lanes 2 and 3). Survivor formation became obvious by the appearance of randomly sized telomeric fragments after 75 generations and the significant amplification of Y′ elements after 100 and 125 generations (Fig. 4, lanes 4 to 6). In contrast, the rate of GT repeat shortening was clearly reduced in a cdc13-4 mutant (Fig. 4, lanes 7 to 11) compared to est2Δ cells and telomeres did not reach the critical length where Y′ element amplification starts inest2Δ strains. We observed no increase in Y′ element signals in BMY59-6B cells after growth for 125 generations (Fig. 4, lane 11) or 250 generations (Fig. 2, lane 5). Furthermore, acdc13-4/rad52 double mutant, BMY18, displayed no growth reduction after several hundred generations (data not shown). Telomeres were as short as those observed for the single cdc13-4mutant and stayed stable at this short level (data not shown).
To investigate whether the rate of telomere shortening is increased in a cdc13-4/est2 double mutant, we compared the growth behavior of an est2Δ spore and acdc13-4/est2Δ spore from a tetrad of strain BMY58. The double mutant displayed significant growth reduction after 50 generations and survivor formation after 75 generations comparable to that of the est2Δ single mutant (Fig. 3). In addition, telomere shortening was not accelerated and Y′ element amplification occurred in both strains after 75 generations (data not shown).
Cdc13p and Tel1p function in different pathways of telomere length maintenance.The synthetic lethality of the cdc13-4mutation 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 tel1Δ 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.
Telomere length of cdc13-4, tel1Δ, andcdc13-4/tel1Δ 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, W303aα wild type; lane 2, W303aα cdc13-4/CDC13tel1Δ/TEL1 (BMY57); lane 3, W303α tel1Δ, 50 generations; lane 4, W303α tel1Δ, 125 generations; lane 5, W303α tel1Δ, 225 generations; lane 6, W303acdc13-4, 50 generations; lane 7, W303a cdc13-4, 125 generations; lane 8, W303a cdc13-4, 225 generations; lane 9, W303α tel1Δ/cdc13-4, 50 generations; lane 10, W303α tel1Δ/cdc13-4, 125 generations; lane 11, W303α tel1Δ/cdc13-4, 225 generations; lane 12, W303a wild type, 50 generations; lane 13, W303a wild type, 225 generations.
As shown in Fig. 5, the rate of GT repeat loss was accelerated intel1Δ cells (Fig. 5, lane 3) compared to that incdc13-4 mutant cells (Fig. 5, lane 6). At the steady-state level, telomeres of tel1Δ cells were significantly shorter than those of cdc13-4 mutant cells (Fig. 5, lanes 4, 5, 7, and 8). The rate of telomere shortening in thecdc13-4/tel1Δ double mutant strain (Fig. 5, lane 9) seemed not to be accelerated compared to that in tel1Δ (Fig. 5, lane 3), but the telomeres of the double mutant were shorter than the telomeres of either single mutant after 125 generations (Fig. 5, lanes 4, 7, and 10). After 225 generations we observed a dramatic increase in Y′ element signals in the cdc13-4/tel1Δ double mutant (Fig. 5, lane 11), indicating that telomeres were stabilized by Y′ element amplification. Although the growth ofcdc13-4/tel1Δ mutants seemed to be reduced after 100 generations, cells did not cease growth completely and no fast-growing survivors occurred. Instead, colonies of the double mutant formed during a further 100 generations of growth were significantly smaller than those of either single mutant or wild type (data not shown).
Cdc13-4p is not altered in its binding to Est1p.Expression of a Cdc13-Est1 fusion protein complements a cdc13 orest1 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 acdc13-4 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 cdc13Δ cells.
Expression of either fusion protein resulted in significant telomere elongation in wild-type and cdc13-4 strains (Fig.6A). We observed no differences in telomere elongation between mutant Cdc13-4–Est1p- and wild-type Cdc13-Est1p-expressing cells (Fig. 6A, lanes 2, 3, 5, and 6), indicating that both proteins bind with comparable affinities to chromosome ends. The effects of the fusion proteins on telomere length were not as pronounced as expected, and although the expression of the Cdc13-Est1 fusion protein in cdc13-4 cells leads to GT tract elongation, the telomeres of these cells did not reach wild-type level. These results point toward a competition between the fusion protein and cellularly expressed Cdc13p alleles. When the influence of the fusion proteins on telomere maintenance was examined in a cdc13Δ strain, dramatically elongated telomeric GT repeat tracts were observed after 100 generations. However, the Cdc13-4–Est1 fusion protein (Fig.6B, lane 5) did not induce telomere elongation to the same extent as a Cdc13-Est1 fusion (Fig. 6B, lane 4). These data suggest that the Cdc13-4–Est1 fusion is capable of binding the telomeric ends and provokes deregulated telomere elongation. Hence, since the Cdc13-4–Est1 fusion does not lead to telomere elongation as observed for Cdc13-Est1p, the establishment of a permanent interaction between Cdc13-4p and Est1p, thereby tethering telomerase to the telomere, seems not sufficient to complement the cdc13-4mutation.
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)20oligonucleotide. 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 cdc13Δ strains transformed with plasmids p-CDC13-EST1 and p-cdc13-4-EST1 probed with a poly(GT)20oligonucleotide. Transformants were cultured for 25 or 100 generations prior to DNA preparation.
To verify that the Cdc13p-Est1p interaction is not altered in acdc13-4 mutant, we performed coimmunoprecipitation experiments. Recently it was reported (45) that a Cdc13-Est1 interaction can be detected biochemically if both proteins are overexpressed. Therefore, we generated strains expressing the chromosomal copy of CDC13 or cdc13-4 as an N-terminal HA3-tagged protein and Est1p as a GST fusion protein under control of the inducible, strong GAL1promoter. Anti-GST monoclonal antibodies were used to precipitate GST-Est1 fusion proteins, and precipitates were analyzed by Western blotting for HA3-Cdc13p or HA3-Cdc13-4p.
Although only a small portion of the GST-Est1 fusion protein interacts with Cdc13p, no differences in the amount of coimmunoprecipitated HA3-Cdc13p (Fig. 7, lane 9) or HA3-Cdc13-4p (Fig. 7, lanes 6 and 10) protein were detectable. We did not observe cross-reaction of HA3-Cdc13p with the anti-GST antibody (Fig. 7, lane 8), and no signal was detectable when GST-Est1p was immunoprecipitated from extracts containing wild-type Cdc13 without the HA tag (Fig. 7, lane 7). From these data we conclude that the Cdc13-4 mutant protein is not altered in its ability to interact with Est1p.
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 cdc13-4mutation have been described. These mutant Cdc13 proteins seem to display significantly reduced binding activity to telomeric DNA (18). Although expression of Cdc13-Est1p incdc13-4 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 inducibleGAL1 promoter.
After growth on glucose-containing media, the telomere length of HFY82 cells was comparable to that of wild type (Fig.8, lanes 1 and 4), indicating that one wild-type copy of CDC13 was sufficient for telomere stability. Strikingly, after growth under inducing conditions on galactose for approximately 50 generations, telomeres were significantly shorter than those of the wild type (Fig. 8, lane 5). Telomere shortening was already obvious in the heterozygous strain BMY56, where Cdc13p and Cdc13-4p were expressed from the nativeCDC13 promoter (Fig. 8, lane 3), even though GT repeat tract loss was not as pronounced as seen in a haploid cdc13-4mutant (Fig. 8, lane 2). In addition, cooverexpression of HA3-Cdc13-4p and GST-Est1p could not restore wild-type telomere length, but it did induce telomere shortening (Fig. 8, lane 7). Therefore, Cdc13-4p is at least in part dominant on Cdc13p and might compete with Cdc13p for telomere binding. These data indicate that neither DNA binding activity nor interaction with Est1 is reduced in a Cdc13-4 mutant protein.
Overexpression of cdc13-4 in a heterozygous mutant strain. Diploid strains W303aα, 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, W303aα; 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/CDC13GAL1-EST1/EST1) grown on galactose. Brackets indicate terminal GT repeats, asterisks indicate non-Y′ elements, and the arrows represent Y′ element bands.
It has been proposed that Ccd13p protects chromosome ends from degradation by binding to the ss 3′ GT overhang. At the restrictive temperature, cdc13-1ts cells exhibit an increased amount of ssDNA in telomeric and subtelomeric regions (16). To investigate whether a cdc13-4 mutant displays an accumulation of ssDNA at the telomeres, we performed nondenaturing in-gel hybridization using a synthetic oligonucleotide specific for telomeric GT repeats. As a control we usedyku80 mutant cells that have been shown to contain a long ss overhang of the G-rich strand throughout the cell cycle (19).
In contrast to yku80 mutants, the ssDNA signal of thecdc13-4 mutant remained at a wild-type level after growth for 40 (Fig. 9, lanes 2 and 4) and 260 (Fig. 9, lanes 3 and 5), generations. Therefore, chromosome ends still seem to be protected from nucleolytic degradation by the Cdc13-4 mutant protein.
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
S. cerevisiae CDC13 is an essential gene involved in chromosome end replication and protection. The cdc13-4allele, which we isolated in a synthetic lethality screen withYKU70, 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 acdc13-2est 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 C1–3A/TG1–3 repeats in aRAD52-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 est2Δ 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 ofcdc13-4. Telomeres in tel1-1 andtel2-1 mutants shorten to a stable level within 150 generations (35, 48), and a tel1-1/tel2-1double mutant has no telomeres shorter than those of tel1-1cells. This suggests that Tel1p and Tel2p function in the same pathway of telomere maintenance (25, 35). In contrast, acdc13-4/tel1Δ double mutant displays slightly shorter GT repeat tracts after 125 generations compared to those of atel1Δ 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 withyku70 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 thecdc13-4 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 C1–3A 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. Thecdc13-4 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 incdc13-4 mutants (Fig. 9).
Functional domains and selected mutations ofCDC13 mutants. AA, amino acid; n.d., not determined.
The expression of a Cdc13-Est1 or Cdc13-Est2 fusion protein incdc13Δ 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/cdc13-4diploid 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). Thecdc13-4 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 cdc13Δ 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 processes—telomere 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 cdc13-4mutant 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/tel1Δ 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 α, Pol δ, 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 α. Single point mutations in either CDC13 orPOL1 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 cdc13-1ts mutant at the restrictive temperature), (ii) loading of telomerase onto the ssDNA overhang at the telomere (prevented in cdc13-2est cells), and (iii) regulation of telomere length. The role of Cdc13p in telomere length control seems to be multifaceted, since mutatingCDC13 can cause either telomere lengthening, seen incdc13-50 mutants (45) and different mutantCDC13 alleles (18), or telomere shortening to a new steady-state level, seen in newly identified CDC13mutants (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.
ACKNOWLEDGMENTS
We thank E. Schiebel for yeast strains and plasmids. We are grateful to R. J. Wellinger for ssDNA analysis. We also thank S. Eimer for helpful discussions during this project and S. Eimer and B. Lakowski for valuable comments on the manuscript.
B. Meier and L. Driller contributed equally to the work.
This work was supported by Deutsche Forschungsgemeinschaft grant Wi 319/11-3, Project 7.
FOOTNOTES
- Received 26 February 2001.
- Returned for modification 28 March 2001.
- Accepted 13 April 2001.
- Copyright © 2001 American Society for Microbiology
