IN the yeast Saccharomyces cerevisiae, subtelomeric DNA is distinctly different from the rest of the chromosomal DNA in that it is repetitive and contains relatively few genes (Louis 1995). While subtelomeric sequences comprise ∼7% (∼25 kb × 32 ends/12,000 kb) of the genome, their function is unknown. During meiosis, homologs pair and segregate to reduce the chromosome number by half. Reciprocal recombination (crossing over) between homologs is essential for segregation. However, crossovers near the ends of S. cerevisiae chromosomes fail to promote efficient segregation (Rosset al. 1996) and might prevent it if they induce crossover interference (Suet al. 2000), a process that could prevent the occurrence of functional crossovers elsewhere on that chromosome. It also has been suggested that recombination near Drosophila chromosome ends fails to promote efficient segregation requiring utilization of the distributive segregation system (Carpenter 1973; Rasoolyet al. 1991; Mooreet al. 1994). Subtelomeric sequences from S. cerevisiae appear to exhibit an absence of meiotic double-strand-break sites (Kleinet al. 1996; Gertonet al. 2000) and undergo little reciprocal recombination (Steensmaet al. 1989; Goldman and Lichten 1996; Suet al. 2000). In this report, the low rate of recombination near the ends of chromosome I from S. cerevisiae are shown to be an inherent property of the subtelomeric sequences and not a telomere position effect (TPE). The few crossovers that occur are shown to produce crossover interference. As high levels of meiotic crossing over near chromosome ends could impair chromosome segregation, we suggest that subtelomeric DNA functions to prevent these crossovers from occurring. Therefore, subtelomeric regions help to ensure that crossovers occur where they promote segregation. Subtelomeric regions exhibit low levels of transcription and late DNA replication (Ferguson and Fangman 1992; Pryde and Louis 1997). Low levels of transcription near telomeres may be due in some part to a well-characterized TPE that was first revealed by an epigenetic repression of expression of genes inserted within or near telomeres (Gottschlinget al. 1990; Pryde and Louis 1997; Wyricket al. 1999). Late replication is also due to a TPE, but one that appears to involve a mechanism independent of the genes required to repress transcription (Ferguson and Fangman 1992; Pryde and Louis 1997). Investigation of meiotic recombination on the left end of chromosome I revealed that crossovers were negligible in the endmost 4–9 kb and then increased with distance from the telomere (Figure 1, A and B; Suet al. 2000). Depending on the composition of the subtelomeric region, similar gradients were observed for a TPE on transcription (Renauldet al. 1993).

The mechanism that keeps rates of meiotic recombination low in subtelomeric regions is unknown but could be due to either telomere proximity or properties inherent in subtelomeric DNA sequences. If the low rate of recombination is caused by telomere proximity, it might be related to the TPE on transcription or replication or it could be the result of the sequestration of telomeres into the bouquet array at meiotic leptotene (Trelles-Strickenet al. 1999). Such sequestration could prevent these sequences from participating in meiotic pairing and recombination.

To test whether low levels of meiotic recombination at the ends of chromosomes were the result of telomere proximity, genetically marked subtelomeric sequences from chromosome I were moved away from their natural telomere either by inserting ∼15 kb of S. carlsbergensis DNA at the left end or by reciprocally translocating a large fragment from chromosome II to the right end (Figure 1B). Reciprocal recombination was analyzed by tetrad analysis. The results (Table 1) showed only a small increase with respect to the controls in the left subtelomeric region, iLEU2-4-iHIS3-iTRP1-16(23)-pURA3 (INSERTION 1, 1A, 2, and 3), and a small decrease with respect to the controls in the right subtelomeric region, pLYS2-PHO11::ADE2 (TRANSLOCATION). The average total amount of recombination in the combined gene-poor iLEU2-4-iHIS3-iTRP1-16(23)-pURA3 region was almost identical in the insertions and the controls (6.1 ± 1.6 cM vs. 5.2 ± 1.1 cM). The lower rate of recombination in the pLYS2-PHO11::ADE2 interval of chromosome I in the translocation is consistent with an effect due to chromosome-size-dependent control of reciprocal recombination (Kabacket al. 1992). In all strains examined, the internal control intervals either were the same (pURA3-PYK1, control for the insertions) or showed the expected decrease (ADE1-pLYS2, control for the translocation) due to the large increase in chromosome size (Kabacket al. 1992). Thus, any TPE on meiotic recombination must be minimal and certainly does not affect most of the subtelomeric DNA.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

—Low rates of meiotic reciprocal recombination in subtelomeric regions of chromosome I are not due to a TPE. Recombination rates (centimorgans per kilobase) are from Table 1 for the noted heterozygous marker pairs in strains containing normal copies of chromosome I and those containing homozygous constructs that either separated the telomere from the marked subtelomeric region or moved a telomere to a new location in the middle of the chromosome. (A) Physical map of the subtelomeric regions of chromosome I showing location of open reading frames (shaded regions). (B) The left telomere was separated by inserting 15.5 kb of S. carlsbergensis (S. carlberg. or S. carl.) DNA (INSERTIONS 1, 2, and 3). The S. carlsbergensis DNA was further subdivided by inserting kanmx on one homolog (INSERTION 1A). The right telomere was separated using chromosome I-chromosome II reciprocal translocation II (TRANSLOCATION) (Kabacket al. 1999), which contains a breakpoint ∼1 kb to the right of the PHO11 gene (total length 800 kb). (C) A telomere was placed adjacent to YAL049 using chromosome I, bisection III (Kabacket al. 1999), and adjacent to CDC15 using deletion Δ3 (Bartonet al. 1997). Dashed lines denote deletion borders. Arrows indicate telomeres and ovals indicate centromeres. Dashed regions on chromosome maps and hatched regions of the bar graphs denote large physical distances that are not near telomeres and are not drawn to scale.

Next, reciprocal recombination was examined within the translocated telomere proximal S. carlsbergensis DNA, which originally came from the middle of chromosome III (iLEU2-iHIS3). The results showed that the rate was relatively high and about equal to that found in the middle of chromosome I (Kabacket al. 1989b). The S. carlsbergensis DNA insert was divided with an additional marker (kanmx; INSERTION 1A) into telomere proximal and distal halves. The telomere proximal half (iLEU2-kanmx) underwent recombination at a slightly higher rate compared to the telomere distal half (kanmx-iHIS3), consistent with idea that telomere proximity has little if any appreciable effect on meiotic recombination rates.

In these experiments, the intervals adjacent to the S. carlsbergensis insert, iLEU2-4-iHIS3, iHIS3-iTRP1-23, and iHIS3-iTRP1-16, showed a very small increase in recombination rates compared to the controls. While these results might be due to telomere removal and a small TPE, it is more likely that they are due to recombination events initiated in the actively recombining S. carlsbergensis DNA and resolved in the adjacent S. cerevisiae DNA. Note the absence of an observable increase in the more distal iTRP1-pURA3 interval. Irrespective of any small increase, recombination rates in these intervals were still very low and substantially less than elsewhere on chromosome I (Kaback 1989).

To further rule out a TPE, telomeres were placed adjacent to two regions that normally recombine at high rates and reciprocal recombination was measured (Figure 1C). Chromosome I bisection III places a cloned telomere from an artificial chromosome adjacent to YAL049 (Kabacket al. 1999). Deletion Δ3 places the natural chromosome IR telomere adjacent to PAU7, ∼2 kb from CDC15 (Bartonet al. 1997). Genes adjacent to the chromosome IR telomere have been shown to be under the control of a SIR-gene-dependent TPE (Bartonet al. 1997; A. B. Barton and D. B. Kaback, unpublished results). Nevertheless, rates of reciprocal recombination in both new telomere proximal intervals, YAL049::TRP1-CDC24 in the bisection and ADE1-PAU7:: LEU2::PHO11 in the deletion, were relatively high and did not vary significantly from the controls, YAL049::LEU2-CDC24 and ADE1-PAU7::LEU2, respectively, which were not adjacent to telomeres. The control for the bisection used a different selectable marker (TRP1 vs. LEU2) in otherwise isogenic strains. However, previous studies showed that heterozygous marker identity had no effect on reciprocal recombination in the adjacent homologous regions (Suet al. 2000). The rate of recombination in the 6.8-kb ADE1-PAU7::LEU2::PHO11 interval in deletion Δ3 equaled that found in the slightly smaller and well-investigated ADE1-CDC15 interval as well as the entire 47.0-kb ADE1-pHIS3 interval on normal copies of chromosome I (Mortimer and Schild 1980; Mortimer and Schild 1985; Kabacket al. 1989a). Therefore, the translocated telomeres did not lower recombination rates.

These results are in contrast to an ∼2.5-fold decrease in recombination observed adjacent to a translocated telomere in a different chromosome I bisection (Kabacket al. 1999). The cause of this apparent paradox is likely to be that the bisection that affected recombination rates removed a prominent DSB site normally adjacent to the interval (iARG4-fun30::LEU2) that had been examined (Kleinet al. 1996).

In total, these results show that telomere insertion cannot be correlated with any predictable repression of meiotic reciprocal recombination. Therefore, we conclude that the low recombination rates observed near both ends of chromosome I must be due mostly to the composition of the subtelomeric DNA sequences. Any repression due to telomere proximity must be less than twofold and is therefore incapable on its own of producing the observed low recombination rates throughout these 20- to 30-kb subtelomeric regions. It is still possible that a localized TPE prevents recombination in the outermost 2–4 kb of the chromosome end because this region was not examined. A telomere-dependent suppression of mitotic recombination has been observed (Stavenhagen and Zakian 1998). However, this effect appears to be specific to C_1-3A telomeric repeats, not to subtelomeric DNA. Furthermore, there was no apparent effect on recombination within the telomere-proximal S. carlsbergensis sequences. The absence of any telomere effect on meiotic recombination is further supported by the high level of ectopic recombination and recombination-induced marker loss observed near telomeres of chromosome fragments that are missing subtelomeric sequences (Arbelet al. 1999) and by the observation that low levels of recombination at the end of chromosome I were independent of SIR2 and SIR3, both of which are required to produce transcriptional TPEs (Suet al. 2000). The results presented here also are inconsistent with the possibility that recombination is prevented by physical constraints due to the association of telomere DNA with the nuclear periphery during meiotic prophase unless subtelomeric DNA itself is involved in the perinuclear localization of chromosomes.

Reciprocal recombination induces crossover interference. To determine whether crossovers in the iLEU2-4-pURA3 subtelomeric region exhibited interference on the adjacent pURA3-PYK1 interval, previously described tetrad data (Suet al. 2000) was further analyzed. Reciprocal recombination between pURA3 and PYK1 was significantly (P < 0.001) reduced from 38.0 ± 1.6 cM [447 parental ditypes (PDs), 28 nonparental ditypes (NPDs), and 724 tetratypes (TTs)] in tetrads that did not have a crossover in the subtelomeric region to 27.2 ± 4.0 cM (74 PDs, 2 NPDs, and 62 TTs) in tetrads that contained a crossover in the subtelomeric region. The same data were analyzed by ranking tetrads within the multiply marked subtelomeric intervals from iLEU2-pURA3 and by calculating interference with the Tetrads program (Mortimeret al. 1989). Crossovers within this subtelomeric region again exhibited crossover interference (k = 0.2 ± 0.1; based on 1303 no-crossover tetrads, 161 single-crossover tetrads, 2 double-crossover tetrads, and 0 >2-crossover tetrads, where k is equivalent to the coefficient of coincidence). In sum, these data indicate that a crossover in subtelomeric DNA induces crossover interference elsewhere on chromosome I.

Subtelomeric regions in yeast have been compared to telomeric heterochromatin of other eukaryotes and some noncoding structural role has been suggested (Busseyet al. 1995). As chiasmata that occur near the ends of chromosomes appear to be much less efficient at promoting segregation (Rosset al. 1996) but cause crossover interference, their occurrence is likely to lower the probability that a functional crossover occurs elsewhere on the chromosome. Indeed, it has been shown that interference can affect >200 kb, the approximate size of the two smallest yeast chromosomes (Kabacket al. 1999). The absence of subtelomeres would presumably lead to a normal rate of recombination within the endmost 30 kb of a chromosome. While the effect of losing a single subtelomeric sequence is expected to be relatively modest, certainly requiring the analysis of thousands of asci to quantitate, the loss of both subtelomeric regions is predicted to produce >1% meiotic nondisjunction for each of the small chromosomes (Suet al. 2000). Since this effect is expected to be additive, it is likely that a successful organism would not tolerate these high levels of nondisjunction. Therefore, we propose that subtelomeric regions function to prevent crossovers from occurring where they cannot promote segregation. The absence of subtelomeres would lead to high levels of crossing over near the ends of chromosomes and meiotic chromosome nondisjunction, especially on the smaller chromosomes.

It should be emphasized that these results were obtained with the smallest yeast chromosome. Since this chromosome has the fewest crossovers, it is likely to be most affected by the possible deleterious effects of high rates of recombination near telomeres. Thus, it is possible that the subtelomeric sequences on this chromosome are unique. Nevertheless, the apparent low level of DSB sites near the ends of all yeast chromosomes suggests that all subtelomeric DNA sequences may indeed play an active role in preventing nonfunctional meiotic recombination on all chromosomes.

The mechanism for lowering rates of recombination in these regions is not known but is clearly dependent on the subtelomeric sequences. Why subtelomeric DNA sequences are inherently nonrecombinogenic during meiosis may reflect the paucity of functional transcription promoter elements or cis-acting inhibitors capable of preventing recombination over a distance. It appears to be unrelated to ARS proximity as the S. carlsbergensis DNA inserted near the left telomere has two ARS elements that appear to be active in S. cerevisiae (Yanget al. 1999). Finally, some cases of human trisomy 21 (Down syndrome) that arose due to improper meiosis I chromosome segregation have been attributed to crossovers that occurred near the ends of chromosome 21, rather than to an absence of crossing over (Lambet al. 1997). These cases suggest that humans might also have a mechanism for preventing crossovers from occurring near telomeres.

• Received March 21, 2003.
• Accepted May 24, 2003.

• Copyright © 2003 by the Genetics Society of America