Regulation of Genome Stability by TEL1 and MEC1, Yeast Homologs of the Mammalian ATM and ATR Genes

Corresponding author: Thomas D. Petes, University of North Carolina, Chapel Hill, NC 27599-3280., tompetes{at}email.unc.edu (E-mail)

Communicating editor: L. PILLUS

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In eukaryotes, a family of related protein kinases (the ATM family) is involved in regulating cellular responses to DNA damage and telomere length. In the yeast Saccharomyces cerevisiae, two members of this family, TEL1 and MEC1, have functionally redundant roles in both DNA damage repair and telomere length regulation. Strains with mutations in both genes are very sensitive to DNA damaging agents, have very short telomeres, and undergo cellular senescence. We find that strains with the double mutant genotype also have 80-fold increased rates of mitotic recombination and chromosome loss. In addition, the tel1 mec1 strains have high rates of telomeric fusions, resulting in translocations, dicentrics, and circular chromosomes. Similar chromosome rearrangements have been detected in mammalian cells with mutations in ATM (related to TEL1) and ATR (related to MEC1) and in mammalian cells that approach cell crisis.

CANCER cells often have elevated rates of genome instability of a variety of types including: (1) mutations and microsatellite alterations (MIN^- tumors), (2) chromosome nondisjunction (CIN^- tumors), and (3) translocations and other chromosome rearrangements (LENGAUER et al. 1998 ). Mutations in human genes in a variety of DNA repair pathways, including mismatch repair, DNA repair checkpoints, recombinational repair, and nucleotide excision repair result in increased genome instability at the cellular level and increased predisposition to cancer (SCHAR 2001 ).

The telomeres of normal human cells shrink with each cell division (HARLEY 1995 ). When the telomeres reach a certain critical length, the cells arrest in a state called "senescence" (HAYFLICK and MOORHEAD 1961 ; HARLEY 1995 ). When treated with transforming agents, some types of mammalian cells escape senescence (resuming cell division), although most of these cells will subsequently die at a stage called "crisis." The transition to crisis is associated with elevated genome instability including the formation of dicentric and ring chromosomes (COUNTER et al. 1992 ). These aberrations are thought to be a consequence of the very short telomeres in the cells that have escaped the senescent state.

One important gene in the regulation of genome stability in mammalian cells is ATM, the gene mutated in patients with ataxia telangiectasia (SHILOH 1997 ). The cellular phenotypes in human cells lacking ATM include sensitivity to DNA damaging agents, elevated levels of mitotic recombination (MEYN 1993 ), and increased end-to-end chromosome fusions associated with shortened telomeres (PANDITA et al. 1995 ). Patients with ATM are cancer prone (SHILOH 1997 ). Mutations in the structurally related gene ATR are associated with embryonic lethality and chromosomal fragmentation (BROWN and BALTIMORE 2000 ). ATM and ATR are proteins required for checkpoint responses to DNA damaging agents, such as the phosphorylation of a number of proteins that lead to cell cycle arrest in response to DNA damage (SHILOH 2001 ). It is likely that the elevated levels of genome instability observed in strains with ATM or ATR mutations reflect both inefficient repair of spontaneous DNA damage (such as double-stranded DNA breaks occurring during DNA replication) and defects in telomere replication.

The yeast Saccharomyces cerevisiae has genes related to ATM and ATR (CARR 1997 ; KEITH and SCHREIBER 1999 ), TEL1 (GREENWELL et al. 1995 ; MORROW et al. 1995 ), and MEC1/ESR1 (KATO and OGAWA 1994 ; WEINERT et al. 1994 ), respectively. Strains with single mutations in MEC1, but not TEL1, are sensitive to DNA damaging agents and fail to arrest the cell cycle in response to DNA damage or inhibition of DNA synthesis (WEINERT et al. 1994 ). Tel1p appears to have a role in the repair of DNA damage that is functionally redundant with that of Mec1p because tel1 mec1 double mutant strains are more sensitive to DNA damaging agents than are mec1 strains (MORROW et al. 1995 ; SANCHEZ et al. 1996 ; USUI et al. 2001 ) and an extra copy of TEL1 partially suppresses the DNA damage sensitivity of mec1 strains (MORROW et al. 1995 ). Strains with the tel1 mutation have greatly shortened telomeres, but do not undergo senescence (LUSTIG and PETES 1986 ). Although the effect of the mec1 mutation on telomere length is subtle, strains with the tel1 mec1 double mutant genotype undergo cellular senescence with approximately the same kinetics as observed in strains that lack telomerase (RITCHIE et al. 1999 ).

Strains with mec1 mutations have multiple phenotypes in addition to their sensitivity to DNA damaging agents including: (1) defective regulation of nucleotide pools (ZHAO et al. 1998 ), (2) defective meiotic checkpoints and regulation of recombination (KATO and OGAWA 1994 ; LYDALL et al. 1996 ; GRUSHCOW et al. 1999 ), (3) loss of telomeric silencing (CRAVEN and PETES 2000 ), (4) inability to redistribute silencing proteins from the telomeres to the sites of double-strand DNA breaks (MCAINSH et al. 1999 ; MILLS et al. 1999 ), (5) deficiency in regulating the firing of DNA replication origins in response to reduced nucleotide pools (SANTOCANALE and DIFFLEY 1998 ), (6) reduced ability to prevent breakdown of DNA replication forks stalled as a consequence of DNA damage (TERCERO and DIFFLEY 2001 ), and (7) accumulated small single-stranded DNA synthesis intermediates (MERRILL and HOLM 1999 ). At least some of the phenotypes associated with the mec1 mutation are likely to reflect the protein kinase activity associated with Mec1p, since a number of proteins involved in DNA replication and DNA repair, such as Rad53p (SANCHEZ et al. 1996 ) and replication protein A (BRUSH et al. 1996 ), are phosphorylated in a Mec1p- and/or Tel1p-dependent manner (LOWNDES and MURGUIA 2000 ). The observation that some substrates undergo both Mec1p-dependent and Tel1p-dependent phosphorylation is consistent with the finding that these proteins appear to have similar in vitro kinase activities (MALLORY and PETES 2000 ; PACIOTTI et al. 2000 ).

The possible functional redundancy of Tel1p and Mec1p in the repair of certain types of DNA damage is also argued by an analysis of gross chromosomal rearrangements (GCR; MYUNG et al. 2001B ). GCR are detected as events that lead to the simultaneous loss of two closely linked genes (CHEN et al. 1998 ). Among the GCR events detected in tel1 mec1 strains were deletions associated with addition of telomeric repeats and translocations with nonhomologous chromosomes (MYUNG et al. 2001B ). The rate of GCR in wild-type strains is very low, about 4 x 10^-10/division. This rate is not elevated in tel1 strains, but is elevated 200-fold in mec1 sml1 strains; strains with the tel1 mec1 sml1 genotype have an extremely high rate of GCR, 12,000-fold higher than that of wild type (MYUNG et al. 2001B ).

Below, we examine the effects of the tel1 and mec1 using several different assays for genome stability. We show that tel1 mec1 double mutant diploid strains have greatly elevated levels of chromosome loss and mitotic recombination. In addition, we demonstrate that tel1 mec1 haploid strains have a high rate of novel chromosome aberrations. Many of these aberrations involve the fusion of telomeric repeats, resulting in dicentric and circular chromosomes. We suggest that these aberrant chromosomes may be causally related to the elevated rates of chromosome instability and the senescence phenotype associated with the tel1 mec1 mutant.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strain constructions:
All of the strains used in this study, except DFS188, were isogenic with W303a (a ade2-1 his3-11,15 ura3-1 leu2-3,112 trp1-1 rad5-535; THOMAS and ROTHSTEIN 1989 ; FAN et al. 1996 ) except for changes introduced by transformation or crosses with isogenic strains; all strains were RAD5. The genotypes of these haploid strains are in Table 1. The genotype of the unrelated haploid DFS188 is a ura3-52 leu2-3,112 lys2 his3 arg8::hisG (SIA et al. 2000 ). Diploids TPY100, TPY101, and MD150 were made by crossing DFS188 to RCY308-10d:CR19, RCY308-10d:CR2, and RCY308-10d:CR18, respectively.

Diploids to monitor mitotic recombination and chromosome loss were constructed by the following crosses: RCY317 (RCY308-3a x RCY308-3d), RCY318 (RCY308-2b x RCY308-7b), RCY319 (RCY308-1d x RCY308-4a), RCY320 (RCY308-11a x RCY308-10d), RCY328 (RCY324-21c x RCY324-4c), RCY332 (RCY329-9b x RCY329-5a), RCY333 (RCY327-3c x RCY327-9a), RCY339 (RCY337-6c x RCY337-12c), RCY340 (RCY337-3d x RCY337-4b), and RCY342 (RCY337-26a x RCY337-24b).

Mitotic recombination and chromosome loss assays:
Mitotic recombination and chromosome loss were measured in diploids heterozygous for the recessive can1 and hom3 markers (Fig 1). For each rate estimate, we measured the frequencies of Can^r Thr⁺ (mitotic recombination events) and Can^r Thr^- (chromosome loss events) cells in 15–20 independent cultures. From these frequency measurements, we calculated rates of mitotic recombination and chromosome loss using the method of the median (LEA and COULSON 1949 ). We calculated 95% confidence intervals on these rate estimates as described previously (WIERDL et al. 1996 ). Similar methods were used in haploid strains to obtain the forward mutation rates at the CAN1 locus (KOKOSKA et al. 2000 ).

View larger version (22K): In this window In a new window Download PPT slide   Figure 1. Assay for chromosome loss and mitotic recombination in diploid yeast strains. Diploids heterozygous for the recessive mutations can1 and hom3 are sensitive to the drug canavanine and do not require threonine. (a) Loss of the chromosome containing the wild-type alleles for these markers results in derivatives that are Canr and Thr-. (b) A mitotic crossover between the CAN1 locus and the centromere, coupled with segregation of the homologs with the can1 mutation, will generate derivatives that are Canr, but Thr+. Other types of mitotic recombination events (gene conversion or break-induced recombination) can yield the same product.

Sequence analysis of can1 mutations:
To determine the nature of the mutation in canavanine-resistant isolates, we examined 10–20 independent canavanine-resistant derivatives of each strain. We first performed PCR analysis using primers derived from the 5'- and 3'-ends of CAN1 to determine whether the gene was present. If a DNA fragment of the expected size (1.7 kb) was present (class 1 mutation), the gene was sequenced at the UNC Automated Sequencing Facility. If no PCR product was produced (class 2 mutants), we did additional PCR reactions using primers derived from the region centromere distal to CAN1 and from the 10-kb region between CAN1 and PCM1 (the essential gene closest to the telomere of V_L). None of the mutant derivatives retained any of the DNA sequences distal to CAN1; all had breakpoints between CAN1 and PCM1. The sequences of the primer pairs used for this analysis are available on request (rolf{at}email.unc.edu).

The fusion partners of these translocated chromosomes were identified using an "arbitrary" PCR strategy (CHEN et al. 1998 ; MYUNG et al. 2001A , MYUNG et al. 2001B ). Ten PCR reactions were performed with each derivative. Each reaction contained one primer derived from the V_L breakpoint and one of the arbitrary primers designed by CHEN et al. 1998 and MYUNG et al. 2001B . The specific primers that were used were: ARB1, ARB4, ARB5, ARB6, ARBG, ARBA, ARBT, ARBC, ARBT1, and ARBT2. Purified genomic DNA was amplified under low-stringency conditions (5 cycles of 95°, 30 sec; 30°, 30 sec; 68°, 2.5 min) followed by higher-stringency amplification (30 cycles of 95°, 30 sec; 55°, 30 sec; 68°, 2 min). After the first round of PCR, the products were then amplified again using a primer directed to the "tag" sequence (ARB2; CHEN et al. 1998 ) and a second V_L primer located 100 bp closer to the chromosomal breakpoint.

Two sets of control reactions were performed: amplification of genomic DNA from a strain harboring an intact V_L chromosome and amplification of DNA from an isolate lacking the entire V_L region telomere proximal to the PCM1 gene. These controls reduced the number of false positives, reflecting PCR artifacts. DNA fragments that were specific to the class 2 canavanine-resistant mutant strains were excised from the gel, purified, and sequenced. Sequences were aligned using the BLAST sequence analysis program in the Stanford University Saccharomyces Genome Database (SGD).

Gel electrophoresis and Southern analysis:
For standard Southern analysis, DNA was isolated (GUTHRIE and FINK 1991 ) from cells grown in liquid (5 ml) YPD cultures at 30°. DNA was treated with various restriction enzymes and the resulting fragments were separated by electrophoresis using 0.8% agarose gel. Standard conditions of hybridization were used. Hybridization probes were prepared by PCR amplification of yeast genomic DNA. For analysis of the dicentric chromosome in RCY308-10d:CR19, we used PCR fragments derived from V_L (P1, generated using primers 5'-GGATGATCTTGGAGATCGC and 5'-GAGTCCAATTAGCTTCATCG) and XV_L (P2, generated using primers 5'-GGAATTTCGTTCCAACATCAATACC and 5'-CTAGTTAAGCGAGCATGTC); P1 includes chromosome V sequences 3207–3500, and P2 has chromosome XV sequences 41416–42014 (numbering system of SGD). For analysis of the circular chromosome in RCY308-10d:CR18, we used a PCR fragment derived from V_R (primers 5'-GAAAGTATAATGGAGCAC and 5'-TACACAGACCATACATTAG); this fragment includes chromosome V sequences 569035–569500. A hybridization probe specific for the Y' subtelomeric repeats was prepared as described by CRAVEN and PETES 1999 .

For analysis of intact chromosomes, we employed the methods described by GUTHRIE and FINK 1991 . Both transverse alternating-field electrophoresis (TAFE; Beckman GeneLine) and contour-clamped homogeneous electric field (CHEF; BioRad CHEF Mapper) setups were used. Yeast chromosome size standards were purchased from New England Biolabs (Beverly, MA). The chromosome V-specific hybridization probe was P1 (described above). The chromosome VIII-specific probe containing ARG4 sequences was obtained by PCR amplification of genomic DNA with the primers ARG4-BclI-F (5'-TGATCAAGTTGTTACCGATTTGAGA) and ARG4-BglII-R (5'-TTGTCCGAATCTCGAATCGATCTTTTG).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The purpose of these experiments was to monitor the effect of tel1 and mec1 mutations on the rates of several classes of genomic alterations including mitotic recombination, chromosome loss, and mutation. Mitotic recombination and chromosome loss rates were analyzed in diploid strains and mutation rates were determined in haploid strains. The genetic backgrounds of all strains were isogenic with W303a except for the mutational alterations described (Table 1). Since tel1 mec1 strains have very short telomeres and a senescent phenotype, as a control, we also examined the same classes of genomic alterations in tlc1 strains that lack telomerase (SINGER and GOTTSCHLING 1994 ). Both tlc1 and tel1 mec1 strains undergo cellular senescence (SINGER and GOTTSCHLING 1994 ; RITCHIE et al. 1999 ), and produce postsenescent "survivors" by a recombination-dependent mechanism (LUNDBLAD and BLACKBURN 1993 ; RITCHIE et al. 1999 ). Type I survivors have amplified subtelomeric repeats, whereas type II survivors have amplified poly (G_1-3T) sequences (TENG and ZAKIAN 1999 ). In our experiments, both tlc1 and tel1 mec1 strains were type I postsenescent survivors.

Mitotic recombination and chromosome loss:
We used a standard assay for measuring mitotic recombination and chromosome loss (HARTWELL and SMITH 1985 ). All diploids used in these experiments were heterozygous for mutations in CAN1 and HOM3, two markers located on opposite arms of chromosome V. Haploid strains with the can1 mutation are sensitive to the drug canavanine and hom3 strains require threonine. Since can1 and hom3 are recessive mutations, the diploid strains are sensitive to Can^s and Thr⁺. Loss of the copy of chromosome V with the wild-type alleles for these markers (Fig 1A) results in a strain that is Can^r and Thr^-. In contrast, mitotic recombination (Fig 1B) results in strains that are Can^r, but Thr⁺.

Two further points concerning the mitotic recombination assay should be mentioned. First, the event diagrammed in Fig 1B is a reciprocal mitotic crossover. Other types of mitotic recombination (gene conversion unassociated with crossingover or break-induced recombination) could produce a strain with the same phenotype (PAQUES and HABER 1999 ). Second, a Can^r Thr⁺ strain could also arise as a consequence of a mutation or deletion of the wild-type CAN1 gene. Since mutation and deletion rates are several orders of magnitude less than the observed mitotic recombination rates, the contribution of these mechanisms is negligible.

The rates of chromosome loss and mitotic recombination (Table 2) were determined as described in MATERIALS AND METHODS. We examined the effects of two different mec1 mutations: mec1-21, a haploid-viable allele (SANCHEZ et al. 1996 ) with a point mutation located outside of the kinase domain (CRAVEN and PETES 2000 ; MALLORY and PETES 2000 ), and mec1-, a complete deletion of MEC1 that is inviable without an accompanying sml1 mutation (ZHAO et al. 1998 ). The null mutation had a stronger effect on both mitotic recombination and chromosome loss rates, elevating both rates 10-fold. Similar effects have been observed in independent experiments by KLEIN 2001 . Strains with either tel1 or sml1 single mutations had approximately wild-type rates of chromosome loss and mitotic exchange, although tel1 mutants have slightly (3-fold) elevated rates of mitotic recombination and chromosome loss in other genetic backgrounds (GREENWELL et al. 1995 ). Strains with the tel1 mec1-21 or tel1 mec1- sml1 genotypes had 80- to 90-fold elevations in the rates of both mitotic recombination and chromosome loss. These rates are greater than expected for additive or multiplicative effects of the single mutations.

As described above, tel1 mec1 strains continuously shorten their telomeres and undergo cellular senescence (RITCHIE et al. 1999 ). To determine whether this phenotype was relevant to the genomic instability observed in tel1 mec1 cells, we also examined the rates of mitotic recombination and chromosome loss in strains lacking telomerase (tlc1 strains). The tlc1 mutation had only modest (threefold) effects, indicating that the elevated rates of genome instability in the tel1 mec1 cells are not solely attributable to a deficiency in telomere length regulation.

Since the mec1 mutation results in a DNA damage checkpoint deficiency, we also determined whether a mutation in a different checkpoint gene, RAD9 (WEINERT 1998 ), would have similar effects. Mutant rad9 strains had slightly elevated mitotic recombination rates and substantially elevated chromosome loss rates (Table 2), as reported previously (WEINERT and HARTWELL 1990 ). In contrast to the tel1 mec1 strain, however, the tel1 rad9 strain had nearly wild-type rates of chromosome loss and mitotic exchange. In addition, the tel1 mutation suppresses the elevated rate of chromosome loss observed in the rad9 strain. Although the interpretation of this observation is not clear, it is possible that the rad9 mutation results in DNA lesions that are transduced by competing Mec1p- and Tel1p-dependent pathways. If Mec1p transduces the signal more efficiently than Tel1p, then loss of Tel1p might result in a more efficient checkpoint response and a decreased rate of chromosome loss.

Mutator phenotype:
We also examined mutation rates in haploid tel1 mec1 strains and several other relevant genotypes using the standard CAN1 forward mutation assay that we have employed in previous studies (KOKOSKA et al. 2000 ). Mutations that inactivate Can1p, an arginine permease, result in strains that are resistant to canavanine. The mutation rate data are shown in Table 3. All rates are averages of rates from two experiments, involving at least 15 independent cultures/experiment.

Single mec1, tel1, or sml1 mutations had no significant mutator phenotype. In contrast, strains of the tel1 mec1-21 or tel1 mec1- sml1- genotypes had strong (50- to 200-fold) mutator effects. Strains with the tlc1 mutation had a substantial (8-fold) mutator phenotype, but the effects of tlc1 were much more modest than those observed for the tel1 mec1 strains. The rad9 mutation had no mutator phenotype as a single mutation or in combination with tel1.

We also analyzed sequence alterations at the CAN1 locus in a number of independent canavanine-resistant derivatives in the wild-type and mutant strains (Table 4). By PCR analysis (described in MATERIALS AND METHODS), we first determined whether the CAN1 locus was deleted. Class 1 mutants retained the locus, and subsequent sequence analysis indicated that these mutants had a point mutation or frameshift within CAN1. Class 1 alterations were the primary type of change observed in wild-type, mec1, tel1, and tlc1 strains. Class 2 mutants were strains with deletions of all or part of the CAN1 gene and loss of all DNA sequences distal to CAN1 on chromosome V. As discussed below, many of the class 2 alterations were fusions of the deleted copy of chromosome V to DNA sequences derived from a different homolog. Class 2 mutations represented most of the canavanine-resistant derivatives in the tel1 mec1 strains (Table 4).

We analyzed the class 2 mutations using the methods developed by CHEN et al. 1998 . The CAN1 locus is located on the left arm of chromosome V 32 kb from the telomere. Since no essential genes are distal to CAN1, deletions or chromosomal rearrangments that delete all of the sequences from CAN1 to the end of the chromosome are haploid viable. The first essential gene centromere proximal to CAN1 is PCM1 (located 44 kb from the telomere). Consequently, to determine the region of chromosome V that was retained in the class 2 mutations, we performed PCR using pairs of primers located centromere distal to CAN1 and at 1-kb intervals between PCM1 and CAN1. All class 2 mutant strains lacked all of the DNA sequences centromere distal to CAN1 and had a breakpoint somewhere between CAN1 and PCM1.

To identify the non-chromosome V sequences at the fusion breakpoint, we then performed 10 additional PCR reactions (details in MATERIALS AND METHODS). All reactions contained one primer in common, a primer with the most centromere-distal chromosome V sequences. The second primer (a pool of degenerate primers; CHEN et al. 1998 ; MYUNG et al. 2001B ) was unique to each reaction. For example, one such primer (ARB1; CHEN et al. 1998 ) had the sequence: 5'-GGCCACGCGTCGACTAGTACNNNNNNNNNNGATAT, with N representing an equal mixture of the four bases; the 20 base pairs at the 5'-end of the primer act as a tag for a second round of PCR. Following the first round of PCR with the 10 reactions, a second round of nested PCR was performed. DNA fragments that were unique to the class 2 mutants were sequenced.

The DNA sequences for 13 class 2 mutants are shown in Fig 2. In all of these strains, we observed DNA sequences derived from chromosome V_L (breakpoints varying between 32444 and 42860) fused to DNA sequences derived from a different chromosome or the opposite end of the same chromosome. As observed in other studies (CHEN et al. 1998 ; MYUNG et al. 2001A , MYUNG et al. 2001B ), some of these fusions involved a small amount of DNA sequence homology at the fusion junction (indicated by underlining in Fig 2), whereas other fusions involved no detectable homology; the lack of substantial sequence homology at the breakpoints indicates that the fusions involve nonhomologous end-joining (NHEJ) pathways rather than homologous recombination (PAQUES and HABER 1999 ). In 12 of the 13 class 2 mutants, the fusions involved chromosomal sequences. In one mutant (RCY337-3d:CR7), DNA sequences derived from the 2µ plasmid were fused to chromosome V sequences; a similar fusion was also observed by CHEN and KOLODNER 1999 . In 5 of the 12 class 2 mutants involving chromosomal sequences, the breakpoints were in the telomeric repeats [poly(G_1-3T)/poly(C_1-3A)] or in the subtelomeric X and Y' repeats.

View larger version (58K): In this window In a new window Download PPT slide   Figure 2. DNA sequences at translocation/fusion breakpoints in canavanine-resistant mutants derived from tel1 mec1 strains. Using methods described in the text, we determined the DNA sequences at the breakpoints between chromosome VL sequences and DNA sequences derived from other chromosomes. Numbers in parentheses indicate the position of the breakpoint on each of the two homologs (based on DNA sequence information from the Stanford Genome Database). Arrows indicate the orientation of the chromosomal DNA sequences before the fusion event and point toward the telomeres. Capital letters indicate the sequences found in the fusion fragment. Underscored letters show sequence identities. (a) DNA sequences derived from fusion fragments in RCY337-3d (tel1 mec1- sml1-). (b) DNA sequences derived from fusion fragments in RCY308-10d (tel1 mec1-21).

Another variable was the relative orientation of the fusions. Every single-stranded yeast DNA sequence (written 5' to 3') can be oriented toward the centromere or toward the telomere by using information in the Stanford Genome Database. On the basis of the orientation of the DNA sequences at the breakpoints, six of the fusions had sequences that were oriented in the same direction on each side of the breakpoint and, thus, would be expected to reflect the formation of two monocentric chromosomes. Six fusions had sequences oriented in opposite directions on each side of the breakpoint and would be expected to reflect the formation of a dicentric chromosome plus an acentric fragment or a circular chromosome (as discussed below).

Types of chromosome rearrangments that could lead to class 2 mutations are shown in Fig 3, a–d. In some cases, the chromosomes expected in the canavanine-resistant derivative depend on the location of the breakpoint. For example, in Fig 3B, if the breakpoint results in loss of essential DNA sequences from the dicentric chromosome, the nontranslocated derivative must segregate with the dicentric to recover a viable canavanine-resistant derivative. Alternatively, if the recombination/fusion involves telomeric sequences (Fig 3C), strains that have only the dicentric chromosome would be viable. It should be pointed out that two mechanisms could generate canavanine-resistant strains with a fusion between telomeric sequences and chromosome V_L: fusion with the telomere of a nonhomologous chromosome (Fig 3C) or an intrachromosomal fusion resulting in formation of a circular molecule (Fig 3D). Although (as described below) we demonstrated one example of both types of telomeric fusions, we have not distinguished these two possibilities for most of the telomeric fusions. Circularization of chromosomes has been observed in Schizosaccharomyces pombe strains with mutations in homologs of TEL1 and MEC1 (NAITO et al. 1998 ).

View larger version (24K): In this window In a new window Download PPT slide   Figure 3. Chromosome alterations leading to loss of CAN1 and distal DNA sequences. In a–c, chromosome V sequences are shown by thin lines (centromeres indicated by circles) and sequences from a nonhomologous chromosome are shown by thick lines; T indicates the telomeres. In a–c, the double-strand break that initiates exchange is on the nonhomologous chromosome and the double-headed arrows show the position of the exchange. We assume that the event is nonreciprocal and that DNA sequences that are unattached to centromeres are not stably maintained. In d, the VL sequences are shown by thin lines, and the VR sequences are represented by thick lines. For each event, only the canavanine-resistant derivative is selectable. All recombination events are assumed to involve little or no DNA sequence homology (consistent with the sequence analysis of breakpoints in Fig 2). (a) Loss of CAN1 as a consequence of formation of two monocentric translocations. (b) Loss of CAN1 as a consequence of formation of a single dicentric translocation. (c) Loss of CAN1 as a consequence of a fusion between VL sequences and telomeric repeats derived from a nonhomologous chromosome. (d) Loss of CAN1 as a consequence of a fusion between VL sequences and telomeric repeats derived from VR.

Physical and genetic analysis of a class 2 strain with a fusion between chromosome V_L and XV_L sequences (RCY308-10d:CR19) and a class 1 control strain (RCY308- 10d:CR2):
Since PCR amplification with degenerate primers might produce DNA fragments artifactually, we did several experiments to confirm our conclusions with one class 1 (RCY308-10d:CR2) and one class 2 (RCY308-10d:CR19) mutant. The orientation of the fusion sequences in RCY308-10d:CR19 (Fig 2B) would be expected to produce a dicentric chromosome with all of the essential DNA sequences of both chromosome V and XV translocation. We confirmed the structure of the fusion junction by Southern analysis (Fig 4). The junction had the restriction fragments expected from our sequence of the fusion junction and the restriction maps derived from the untranslocated chromosomes. In contrast, RCY308-10d:CR2 had the DNA fragments predicted from unrearranged copies of chromosomes V and XV (data not shown).

View larger version (14K): In this window In a new window Download PPT slide   Figure 4. Southern analysis of DNA isolated from RCY308-10d:CR19, a strain with a fusion between VL and XVL sequences, and RCY308-10d (control strain from which RCY308-10d:CR19 was derived). The DNA sequences at the breakpoint of the rearrangement in RCY308-10d:CR19 indicated a fusion at position 40705 of VL and 537 of XVL (within the subtelomeric X repeat of XV). P1 and P2 represent the positions of VL- and XVL-specific probes, respectively (probes described in MATERIALS AND METHODS). (a, top) Restriction maps of regions of the unrearranged VL and XVL chromosomes derived from Stanford Genome Database. (a, bottom) Expected restriction map from the fusion chromosome. (b and c) Southern analysis of DNA from RCY308-10d:CR19 (odd-numbered lanes) and RCY308-10d (even-numbered lanes) using P1 (b) and P2 (c) probes. In both b and c, the restriction enzymes used were BclI (lanes 1 and 2), SphI (lanes 3 and 4), SpeI (lanes 5 and 6), and PvuII (lanes 7 and 8). Note that the hybridization bands in DNA of RCY308-10d:CR19 are at the same positions in gels probed with P1 and P2.

We also showed genetic linkage between chromosome V and XV markers in a diploid (TPY100) formed by mating RCY308-10d:CR19 (can1 ARG8) with DFS188 (CAN1 arg8). Since the ARG8 gene is located 60 kb from the left telomere of chromosome XV, both heterozygous markers are near the putative fusion breakpoint. For strains heterozygous for two unlinked markers, in which at least one marker is also unlinked to the centromere, the expected ratio of parental ditype (PD) to nonparental ditype (NPD) to tetratype (T) tetrads is 1:1:4; linkage is indicated by a significant excess of PD tetrads over NPD tetrads. Of 18 tetrads with four viable spores derived from TPY100, 17 were PD tetrads and 1 was tetratype, indicating tight genetic linkage of markers on nonhomologous chromosomes. As a control, we mated a class 1 mutant (RCY308-10d:CR2) with DFS188. In this diploid strain (TPY101), we observed 6 PD, 3 NPD, and 10 T tetrads. The segregation pattern in TPY101 indicates that the linkage between CAN1 and ARG8 in RCY308-10d:CR19 is not a property of the progenitor RCY308-10d strain. In summary, these results confirm the existence of a fusion between chromosome V and XV sequences in RCY308-10d:CR19.

Another observation suggestive of a chromosome aberration in RCY308-10d:CR19 was the pattern of spore viability in TPY100. Of the tetrads dissected, the percentages of tetrads with 4, 3, 2, 1, and 0 viable spores were 7, 39, 32, 13, and 8, respectively; the total spore viability was 56%. In contrast, in the diploid TPY101, the percentages of tetrads with 4, 3, 2, 1, and 0 viable spores were 59, 19, 11, 6, and 5 (total spore viability of 81%). Poor spore viability is expected in strains containing dicentric chromosomes because the independent attachment of microtubules to two centromeres can lead to loss or fracture of the dicentric chromosome (HABER et al. 1984 ). Since similar patterns of spore inviability are also observed in strains heterozygous for translocations (MIKUS and PETES 1982 ), however, these data are suggestive rather than conclusive.

We also examined the sizes of the chromosomes in strains RCY308-10d:CR2 and RCY308-10d:CR19 by gel electrophoresis. The RCY308-10d:CR2 strain, as well as its TEL1 MEC1 derivative, has a chromosome V that is slightly shorter than the "standard" chromosome V (Fig 5A). If the V_L-XV_L chromosome fusion in RCY308-10d:CR19 involved all of the DNA sequences of chromosome XV (with loss of only the terminal 500 bp) and most of the DNA sequences of chromosome V (40705–576869), the dicentric would be 1600 kb. Using a probe for V_L sequences near the junction (P1, Fig 4A), by OFAGE analysis, we found that the V_L-XV_L fusion chromosome in RCY308-10d:CR19 was 840 kb (lane 2, Fig 5B). On the basis of these data, we suggest that, following the initial formation of the dicentric chromosome, there was a secondary chromosome rearrangment, yielding a V-XV translocation that lost some of the DNA sequences derived from XV.

View larger version (31K): In this window In a new window Download PPT slide   Figure 5. Analysis of chromosome V in strains RCY308-10d:CR2, RCY308-10d:CR18, and RCY308-10d:CR19. Intact chromosomes were examined by gel electrophoresis, using either TAFE (a) or CHEF (b and c) gels. The size standards were commercially available marker yeast chromosomes from New England Biolabs. Following gel electrophoresis, DNA molecules were transferred to nylon membranes and hybridized to chromosome-specific probes [YIp5 for chromosome V and a PCR fragment containing ARG4 sequences for chromosome VIII (primer sequences given in MATERIALS AND METHODS)]. (a) Gel analysis of chromosome V from strain YPH80 (lane 1, marker chromosomes), W303a (lane 2, control wild-type strain), and RCY308-10d:CR2 (lane 3, canavanine-resistant derivative with point mutation in CAN1). (b) Analysis of chromosome V from RCY308-10d (lane 1, control tel1 mec1-21 strain), RCY308-10d:CR19 (lane 2, V/XV translocation), and RCY308-10d:CR2 (lane 3). (c) Analysis of chromosomes V and VIII from RCY308-10d:CR2 (lanes 1 and 3) and RCY308-10d:CR18 (lanes 2 and 4). The nylon filter was first hybridized with the chromosome V-specific probe (lanes 1 and 2) and then stripped and reprobed with the chromosome VIII-specific probe (lanes 3 and 4).

Several additional points should be made concerning the chromosome aberrations of RCY308-10d:CR19. First, we do not know the exact nature of the V-XV translocation. It is not clear whether the 840-kb V-XV chromosome in RCY308-10d:CR19 is dicentric or became monocentric as a consequence of the secondary rearrangement; secondary rearrangements resulting in loss of one centromere from a dicentric have been described previously in yeast (NEFF and BURKE 1992 ). It is also unclear whether the spore inviability patterns observed in TP100 reflect recombination and segregation of a dicentric chromosome or recombination and segregation of another type of chromosomal rearrangement (such as a heterozygous translocation). We also do not know whether there are chromosome aberrations in RCY308-10d:CR19, in addition to the one involving chromosomes XV and V.

Physical and genetic analysis of a class 2 strain with a circular chromosome (RCY308-10d:CR18):
In RCY308-10d:CR18, V_L sequences were fused to poly(G_1-3T) sequences derived from V_R (Fig 2B). Such a fusion should result in a circular chromosome V. This conclusion was confirmed by two types of analysis. First, we examined chromosome V using a CHEF gel (Fig 5C). Although chromosome V from RCY308-10d:CR2 had the expected mobility (lane 1), chromosome V from RCY308-10d:CR18 remained trapped in the loading well (lane 2). The same membrane was rehybridized to a chromosome VIII-specific probe (lanes 3 and 4) to demonstrate that other chromosomes in RCY308-10d:CR18 had the expected mobility. Since the trapping of circular yeast chromosomes in gel wells during OFAGE has been described previously by GAME et al. 1989 , this result is consistent with the presence of a circular chromosome in RCY308-10d:CR18.

We also performed a standard Southern analysis using a V_R-specific probe (described in MATERIALS AND METHODS) that was located in single-copy sequences immediately centromere proximal to the X and Y' telomeric repeats. The expected (Fig 6A) and observed (Fig 6B) sizes of the restriction fragments for a circular chromosome were in good agreement. The confirmation of the SfiI site near the V_L-V_R junction is a particularly convincing argument, since SfiI cuts very infrequently in the yeast genome (LINK and OLSON 1991 ). In addition, since RCY308-10d:CR18 lacks the DNA fragments characteristic of the wild-type V_R, the chromosome rearrangement does not represent a V_R-V_L fusion between two linear copies of chromosome V. It should be noted that, although the strongest band of hybridization in the control strain is at the position expected for the V_R telomere (which has a single Y' element), there is a weak band of hybridization at approximately the size expected for a V_R telomere with two Y' elements. It is likely that the population of cells used for Southern analysis of the control strain had a small number of cells with a double Y' element. Such Y' amplification events have been observed previously in tel1 mec1 strains (RITCHIE et al. 1999 ).

View larger version (39K): In this window In a new window Download PPT slide   Figure 6. Southern analysis of DNA isolated from RCY308-10d:CR18, a strain with a circular derivative of chromosome V. (a) The placement of the restriction sites is based on information in the Stanford Genome Database. In addition to the depicted restriction sites, ApaI and BssHII cut 19 and 57.3 kb, respectively, centromere proximal to J1; SfiI cuts 392 kb centromere proximal to J2. The coordinate for the J2 junction indicated in parentheses is an extrapolation of the information in SGD, since the full sequence of chromosome VR is not present. The subtelomeric repeats of X and Y' are shown as shaded and hatched rectangles, respectively. The terminal poly(G1-3T) tract is shown as an open rectangle and the hybridization probe is indicated by a solid rectangle. (b) Southern blot of DNA isolated from RCY308-10d:CR18 (odd-numbered lanes) and the control strain RCY308-10d (even-numbered lanes) hybridized to a VR-specific probe (indicated by P in a). Restriction enzymes used in the analysis were: BssWI + SfiI (lanes 1 and 2), SfiI (lanes 3 and 4), BsaWI (lanes 5 and 6), BspEI (lanes 7 and 8), BssHII (lanes 9 and 10), AseI (lanes 11 and 12), and ApaI (lanes 13 and 14). The expected sizes of DNA fragments if RCY308-10d:CR18 has a circular chromosome (observed sizes in parentheses) are: BsaWI + SfiI, 8.9 kb (9 kb); SfiI, 392 kb (>20 kb); BsaWI, 12.2 kb (11.5 kb); BspEI, 16.1 kb (15 kb); BssHII, 66.5 kb (>20 kb); AseI, 11.1 kb (12 kb); and ApaI, 27.7 kb (>20 kb). The expected sizes of the DNA fragments for RCY308-10d (observed sizes in parentheses) are: BsaWI + SfiI, 8.5 kb (8.2 kb); SfiI, 392 kb (>20 kb); BsaWI, 8.5 kb (8.2 kb); BspEI, 8.5 kb (8.2 kb); BssHII, 9.5 kb (9.3 kb); AseI, 9.7 kb (10 kb); and ApaI, 8.9 kb (8.8 kb). (c) Southern blot of DNA isolated from RCY308-10d:CR18 (odd-numbered lanes) and the control strain RCY308-10d (even-numbered lanes) hybridized to a Y'-specific probe. The same samples used in b were treated with the same restriction enzymes and loaded in the same order on the gel. The bands shown as black dots in lanes containing RCY308-10d:CR18 are the sizes expected for the circular chromosome V. Lanes marked with open circles are the sizes expected for the linear chromosome V. The bands marked with triangles are approximately the sizes expected for integration of a second Y' at VR in a subset of cells with a linear chromosome V; their sizes also correspond to those of some of the weakly hybridizing bands observed in b.

We also crossed RCY308-10d:CR18 to the wild-type haploid strain DFS188 and examined the viability of spores of the resulting diploid strain (MD150). The percentages of tetrads with 4, 3, 2, 1, and 0 viable spores were 7, 13, 31, 32, and 17, respectively; the total spore viability was 40%. Thus, compared to the control diploid strain (TPY101, described above), there was a significant loss of spore viability. This type of inviability is expected in strains heterozygous for a circular chromosome, since meiotic recombination events between a linear and circular chromosome would be expected to result in formation of dicentric chromosomes and acentric fragments (HABER et al. 1984 ).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains with the double tel1 mec1 genotype have a much stronger phenotype than that observed in strains with the tel1 or mec1 single mutations in a variety of assays of genome stability. The specific assays for genome stability that have been used include sensitivity to DNA damaging agents (MORROW et al. 1995 ; USUI et al. 2001 ); measurements of telomere length and the rates of cellular senescence (RITCHIE et al. 1999 ); rates of deletions and other gross chromosomal rearrangements (MYUNG et al. 2001B ); and rates of chromosome loss, mitotic recombination, and forward mutation rates (as reported above).

The elevated levels of mitotic recombination and chromosome loss observed in tel1 mec1 strains suggest high levels of DNA damage. There are two nonexclusive explanations for the elevated rates of DNA damage. First, wild-type yeast strains may have intrinsically high levels of spontaneous DNA damage, but this level of DNA damage is efficiently repaired by a Mec1p-/Tel1p-dependent mechanism. Second, the tel1 mec1 strains may have a type of DNA damage that is not observed in wild-type strains. Although both of these mechanisms may contribute to the genomic instability observed in tel1 mec1 strains, on the basis of the types of chromosome alterations observed in the canavanine-resistant mutants derived in the tel1 mec1 genetic background, we favor the second explanation.

What types of DNA damage could be responsible for the elevated level of genome instability in tel1 mec1 strains? We suggest that there are two types leading to two classes of genome instability. First, the loss of telomeric repeats in tel1 mec1 strains results in the recognition of the telomeres as double-stranded DNA breaks (DSBs) that require repair. Use of telomeres as recombination substrates could result in formation of chromosome aberrations such as those shown in Fig 3C and Fig D. A similar hypothesis was previously used by MCEACHERN and IYER 2001 to explain the elevated levels of recombination of subtelomeric sequences in Kluyveromyces lactis strains with short telomeres. Second, we postulate that tel1 mec1 strains have elevated levels of DSBs throughout the genome. These DSBs could generate the types of chromosome aberrations shown in Fig 3, a and b. Consistent with this possibility, strains with mec1 (TERCERO and DIFFLEY 2001 ) and rad53 (LOPES et al. 2001 ) mutations are defective in maintaining blocked DNA replication forks, and it has been suggested that aberrant processing of blocked replication forks may lead to DSBs (LOPES et al. 2001 ). In addition, MERRILL and HOLM 1999 showed accumulated single-stranded DNA synthesis intermediates in mec1 strains and suggested that these intermediates would be processed into double-stranded DNA breaks during DNA replication.

Our studies are in good agreement with the very high rate of GCRs observed in tel1 mec1 cells and the low rate of GCR seen in tlc1 cells (MYUNG et al. 2001A , MYUNG et al. 2001B ). HACKETT et al. 2001 recently showed that yeast strains with a mutation in est1 had elevated levels of mutations in the CAN1 assay and high levels of chromosome rearrangments. Although the tlc1 mutation used in our study and that of MYUNG et al. 2001A , MYUNG et al. 2001B would be expected to result in a similar phenotype as the est1 mutation used by Hackett et al., one difference is that our experiments were performed with survivors, whereas those of Hackett et al. were done with est1 cells at various stages of the senescence process. They found that the est1 mutator phenotype was strongest in the cells that were growing poorly, and much of this phenotype was lost after survivors were generated. Since the experiments of MYUNG et al. 2001A , MYUNG et al. 2001B were done with tlc1 strains prior to the onset of survivors, however, there appears to be a significant difference in genomic instability in tlc1 and est1 strains.

Although tel1 mec1 strains have elevated rates of several types of genomic instability, the absolute rates of instability are quite different for the different assays (Table 2 and Table 3): rates (events per cell division) of about 10^-3 for chromosome loss, 10^-4 for mitotic recombination, and 10^-5 for mutations. Although the tel1 mec1-21 and tel1 mec1- sml1 strains have similar rates of mitotic recombination and chromosome loss, mutation rates in the tel1 mec1- sml1 strain are about fourfold higher than those observed in the tel1 mec1-21 strain. One interpretation is that the tel1 mec1- sml1 strain may have a slightly reduced ability (relative to the tel1 mec1-21 strain) to repair DNA breaks by homologous recombination pathways and, consequently, exhibit an elevation in repair by the NHEJ pathway associated with production of the chromosomal rearrangements. Since the rate of homologous mitotic recombination is much higher than the mutation rate at CAN1, an undetectable change in the mitotic recombination rate could substantially elevate the rate of mutation. Alternatively, compared to the tel1 mec1-21 strain, the tel1 mec1- sml1 strain may have elevated levels of DNA damage that can be repaired only by the NHEJ pathway.

It should be emphasized that some of the chromosome aberrations that we have observed would be expected to generate repeated cycles of chromosome rearrangements. For example, dicentric chromosomes, as a consequence of segregation problems during mitosis, may generate double-strand breaks at random positions. Such breaks would be expected to stimulate secondary rounds of mitotic recombination and chromosome loss. This type of mechanism has also been invoked by HACKETT et al. 2001 to explain chromosome rearrangments observed in yeast strains with an est1 mutation. Circular chromosomes, such as that demonstrated in RCY308-10d:CR18, would also be expected to be a source of genome instability. Recombination between circular and linear chromosomes either in meiosis or in mitotic diploid cells would produce dicentric linear chromosomes, and sister-sister strand recombination events would produce dicentric circular chromosomes.

Several more points concerning the comparison of tel1 mec1 and tlc1/est1 phenotypes should be mentioned. First, we observed a class 2 mutant in the tlc1 strain (Table 4), suggesting that chromosome rearrangements may occur in this genetic background. Second, since most of the mutants observed in the tlc1 survivors were point mutations (Table 4), there is a mutator phenotype associated with the telomerase-negative strain that is independent of chromosomal alterations. It is possible that tlc1 strains accumulate DNA damage that is repaired by error-prone DNA polymerases, although other explanations for the mutator phenotype also exist.

Our results indicate that the cell death observed in tel1 mec1 strains may involve more than a single mechanism. One mechanism may be loss of essential genes located near the telomere as a consequence of end-directed DNA degradation. A second mechanism may reflect the constant generation of DNA damage as a consequence of repeated cycles of breakage of dicentric chromosomes. As mentioned in the Introduction, mutations in ATM and ATR also lead to increased levels of end-to-end chromosome fusions, mitotic recombination, and chromosome breakage (MEYN 1993 ; PANDITA et al. 1995 ; BROWN and BALTIMORE 2000 ). Thus, the ATM family of protein kinases safeguards multiple aspects of chromosome stability in very different organisms.

	ACKNOWLEDGMENTS

We thank R. Rothstein, H. Klein, J. Mallory, K. Ritchie, and L. M. Curtis for strains used in the study and H. Klein, D. Gottschling, K. Myung, and R. Kolodner for communication of unpublished results and comments on the manuscript. We thank L. M. Curtis and L. Stefanovich for help with strain constructions, R. Kokoska and J. Merker for advice about data analysis, and K. Lobachev for help with the OFAGE analysis. The research was supported by National Institutes of Health grants to T.D.P. (GM-24110 and GM-52319) and R.J.C. (Building Interdisciplinary Research Careers in Women's Health scholar).

Manuscript received November 13, 2001; Accepted for publication February 19, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BROWN, E. J. and D. BALTIMORE, 2000  ATR disruption leads to chromosomal fragmentation and early embryonic lethality. Genes Dev. 14:397-402[Abstract/Free Full Text].

BRUSH, G. S., D. M. MORROW, P. HIETER, and T. J. KELLY, 1996  The ATM homologue MEC1 is required for phosphorylation of replication protein A in yeast. Proc. Natl. Acad. Sci. USA 93:15075-15080[Abstract/Free Full Text].

CARR, A. M., 1997  Control of cell cycle arrest by the Mec1^sc/Rad3^sp DNA structure checkpoint pathway. Curr. Opin. Genet. Dev. 7:93-98[Medline].

CHEN, C. and R. D. KOLODNER, 1999  Gross chromosomal rearrangements in Saccharomyces cerevisiae replication and recombination defective mutants. Nat. Genet. 23:81-85[Medline].

CHEN, C., K. UMEZU, and R. D. KOLODNER, 1998  Chromosomal rearrangements occur in S. cerevisiae mutator mutants due to intergenic lesions processed by double-strand-break repair. Mol. Cell 2:9-22[Medline].

COUNTER, C. M., A. A. AVILON, C. E. LEFEUVRE, N. G. STEWART, and C. W. GREIDER et al., 1992  Telomere shortening associated with chromosome instability is arrested in immortal cells which express telomerase activity. EMBO J. 11:1921-1929[Medline].

CRAVEN, R. J. and T. D. PETES, 1999  Dependence of the regulation of telomere length on the type of subtelomeric repeat in the yeast Saccharomyces cerevisiae.. Genetics 152:1531-1541[Abstract/Free Full Text].

CRAVEN, R. J. and T. D. PETES, 2000  Involvement of the checkpoint protein Mec1p in silencing of gene expression at telomeres in Saccharomyces cerevisiae.. Mol. Cell. Biol. 20:2378-2384[Abstract/Free Full Text].

CRAVEN, R. J. and T. D. PETES, 2001  The Saccharomyces cerevisiae suppressor of choline sensitivity (SCS2) gene is a multicopy suppressor of mec1 telomeric silencing defects. Genetics 158:145-154[Abstract/Free Full Text].

DESANY, B. A., A. A. ALCASABAS, J. B. BACHANT, and S. J. ELLEDGE, 1998  Recovery from DNA replication stress is the essential function of the S-phase checkpoint pathway. Genes Dev. 12:2956-2970[Abstract/Free Full Text].

FAN, H. Y., K. K. CHENG, and H. L. KLEIN, 1996  Mutations in the RNA polymerase II transcription machinery suppress the hyperrecombination mutant hpr1 of Saccharomyces cerevisiae.. Genetics 142:749-759[Abstract].

GAME, J. C., K. C. SITNEY, V. E. COOK, and R. K. MORTIMER, 1989  Use of a ring chromosome and pulse-field gels to study interhomolog recombination, double-strand DNA breaks and sister-chromatid exchange in yeast. Genetics 123:695-713[Abstract/Free Full Text].

GREENWELL, P. W., S. L. KRONMAL, S. E. PORTER, J. GASSENHUBER, and B. OBERMAIER et al., 1995  TEL1, a gene involved in controlling telomere length in S. cerevisiae, is homologous to the human Ataxia telangiectasia gene. Cell 82:823-829[Medline].

GRUSHCOW, J. M., T. M. HOLZEN, K. J. PARK, T. WEINERT, and M. LICHTEN et al., 1999  Saccharomyces cerevisiae checkpoint genes MEC1, RAD17 and RAD24 are required for normal meiotic recombination partner choice. Genetics 153:607-620[Abstract/Free Full Text].

GUTHRIE, C., and G. R. FINK (Editors), 1991 Guide to Yeast Genetics and Molecular Biology. Academic Press, San Diego.

HABER, J. E., P. C. THORBURN, and D. ROGERS, 1984  Meiotic and mitotic behavior of dicentric chromosomes in Saccharomyces cerevisiae. Genetics 106:185-205[Abstract/Free Full Text].

HACKETT, J. A., D. M. FELDSER, and C. W. GREIDER, 2001  Telomere dysfunction increases mutation rate and genomic instability. Cell 106:275-286[Medline].

HARLEY, C. B., 1995 Telomeres and aging, pp. 247–263 in Telomeres, edited by E. H. BLACKBURN and C. W. GREIDER. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

HARTWELL, L. H. and D. SMITH, 1985  Altered fidelity of mitotic chromosome transmission in cell cycle mutants of S. cerevisiae.. Genetics 110:381-395[Abstract/Free Full Text].

HAYFLICK, L. and P. S. MOORHEAD, 1961  The serial cultivation of human diploid cell strains. Exp. Cell Res. 25:585-621.

KATO, R. and H. OGAWA, 1994  An essential gene, ESR1, is required for mitotic cell growth, DNA repair and meiotic recombination in Saccharomyces cerevisiae.. Nucleic Acids Res. 22:3104-3112[Abstract/Free Full Text].

KEITH, C. T. and S. L. SCHREIBER, 1999  PIK-related kinases: DNA repair, recombination, and cell cycle checkpoints. Science 270:50-51.

KLEIN, H. L., 2001  Spontaneous chromosome loss in Saccharomyces cerevisiae suppressed by DNA damage checkpoint functions. Genetics 159:1501-1509[Abstract/Free Full Text].

KOKOSKA, R., L. STEFANOVIC, J. DEMAI, and T. D. PETES, 2000  Increased rates of genomic deletions generated by mutations in the yeast gene encoding DNA polymerase or by decreases in cellular levels of DNA polymerase . Mol. Cell. Biol. 20:7490-7504[Abstract/Free Full Text].

LEA, D. E. and C. A. COULSON, 1949  The distribution of the number of mutants in bacterial populations. J. Genet. 49:264-285.

LENGAUER, C., K. W. KINZLER, and B. VOGELSTEIN, 1998  Genetic instabilities in cancer. Nature 386:623-627.

LINK, A. J. and M. V. OLSON, 1991  Physical map of the Saccharomyces cerevisiae genome at 110-kb resolution. Genetics 127:681-698[Abstract].

LOPES, M., C. COTTA-RAMUSINO, A. PELLICIOLI, G. LIBERI, and P. PLEVANI et al., 2001  The DNA replication checkpoint response stabilizes stalled replication forks. Nature 412:557-561[Medline].

LOWNDES, N. F. and J. R. MURGUIA, 2000  Sensing and responding to DNA damage. Curr. Opin. Genet. Dev. 10:17-25[Medline].

LUNDBLAD, V. and E. BLACKBURN, 1993  An alternative pathway for telomere maintenance rescues est1-senescence. Cell 73:347-360[Medline].

LUSTIG, A. J. and T. D. PETES, 1986  Identification of yeast mutants with altered telomere structure. Proc. Natl. Acad. Sci. USA 83:1389-1402.

LYDALL, D., Y. NIKOLSKY, D. K. BISHOP, and T. WEINERT, 1996  A meiotic recombination checkpoint controlled by mitotic checkpoint genes. Nature 383:840-843[Medline].

MALLORY, J. C. and T. D. PETES, 2000  Protein kinase activities of Tel1p and Mec1p, two S. cerevisiae proteins related to the human ATM kinase. Proc. Natl. Acad. Sci. USA 97:13749-13754[Abstract/Free Full Text].

MCAINSH, A. D., S. SCOTT-DREW, J. A. MURRAY, and S. P. JACKSON, 1999  DNA damage triggers disruption of telomeric silencing and Mec1p-dependent relocation of Sir3p. Curr. Biol. 9:963-966[Medline].

MCEACHERN, M. J. and S. IYER, 2001  Short telomeres in yeast are highly recombinogenic. Mol. Cell 7:695-704[Medline].

MERRILL, B. J. and C. HOLM, 1999  A requirement for recombinational repair in Saccharomyces cerevisiae is caused by DNA replication defects of mec1 mutants. Genetics 153:595-605[Abstract/Free Full Text].

MEYN, M. S., 1993  High spontaneous intrachromosomal recombination rates in ataxia-telangiectasia. Science 260:1327-1330[Abstract/Free Full Text].

MILLS, K. D., D. A. SINCLAIR, and L. GUARENTE, 1999  MEC1-dependent redistribution of the Sir3 silencing protein from telomeres to DNA double-strand breaks. Cell 97:609-620[Medline].

MIKUS, M. and T. D. PETES, 1982  Recombination between genes located on non-homologous chromosomes in the yeast Saccharomyces cerevisiae.. Genetics 101:369-404[Abstract/Free Full Text].

MORROW, D. M., D. A. TAGLE, Y. SHILOH, F. S. COLLINS, and P. HIETER, 1995  TEL1, an S. cerevisiae homologue of the human gene mutated in Ataxia Telangiectasia, is functionally related to the yeast checkpoint gene MEC1.. Cell 82:831-840[Medline].

MYUNG, K., C. CHEN, and R. D. KOLODNER, 2001a  Multiple pathways cooperate i