Frequent Meiotic Recombination Between the Ends of Truncated Chromosome Fragments of Saccharomyces cerevisiae

Corresponding author: Giora Simchen, Department of Genetics, The Hebrew University, Jerusalem 91904, Israel., simchen{at}vms.huji.ac.il (E-mail)

Communicating editor: M. LICHTEN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

A single truncated chromosome fragment (TCF) in diploid cells undergoes frequent ectopic recombination during meiosis between markers located near the ends of the fragment. Tetrads produced by diploids with a single TCF show frequent loss of one of the two markers. This marker loss could result either from recombination of the TCF with one of the two copies of the chromosome from which it was derived or from ectopic recombination between the ends of the TCF. The former would result in shortening of a normal chromosome and lethality in one of the four spores. The high frequency of marker loss in tetrads with four viable spores supports recombination between the TCF ends as the main source of marker loss. Most of the spore colonies that display TCF marker loss contained a TCF with the same marker on both ends. Deletion of most of the pBR322 sequences distal to the marker at one of the subtelomeric regions of the TCF did not reduce the overall frequency of recombination between the ends, but affected the loss of one marker significantly more than the other. We suggest that the mechanism by which the duplication of one end marker and loss of the other occurs is based on association and recombination between the ends of the TCF.

THE telomeric regions of eukaryotic chromosomes are specialized structures. They have a special mode of replication, ensuring the complete replication of the ends, and they possess special protein assemblies, possibly required for the protection of the ends. Broken chromosomes lacking telomeric structures at their ends are very unstable, due not only to incomplete replication but also to degradation and end-to-end fusion (DUNN et al. 1984 ; PLUTA and ZAKIAN 1989 ; LOUIS and HABER 1990A , LOUIS and HABER 1990B ; LOUIS 1995 ).

Most organisms also possess telomere-associated middle repetitive sequences like the He-T element of Drosophila melanogaster (BIESSMANN et al. 1990 ) and the Y' and X elements of the Saccharomyces cerevisiae (ZAKIAN and BLANTON 1988 ). Several lines of evidence suggest that telomere-associated sequences on different (nonhomologous) chromosomes are involved in recombination processes (CHARRON et al. 1989 ). The Y' elements on different chromosomes are much more homologous within strains than between strains. The number of Y' elements at specific chromosomes was shown to change during mitotic growth via unequal sister chromatid exchange and ectopic recombination, namely, between homologous sequences at different genomic locations (LOUIS and HABER 1990A ). Acquisition of telomeres by linear plasmids involves recombination of both the telomeric (C–A_[1–3])_n sequence, and the telomere-associated Y' element (SZOSTAK and BLACKBURN 1982 ; DUNN et al. 1984 ).

Chromosome fragments (CFs) were found to associate and recombine during meiosis with the chromosomes from which they were derived, thus interfering with their meiotic segregation and resulting in increased meiotic nondisjunction (NDJ; ARBEL 1992 ; GOLDWAY et al. 1993 ). This increase was correlated with the length of homology shared by the chromosome and the CF and was not affected by the position of the centromere on the CF (ARBEL 1992 ). However, although NDJ frequency was increased up to 100-fold, it never exceeded 1% for even the longest CF.

We constructed a new type of CFs derived from chromosome VII of S. cerevisiae and named them truncated chromosomal fragments (TCFs). These fragments were created from "regular" CFs, that run from a certain site on chromosome VII to the telomere of the same arm, by shortening that arm. Both the generation of the original CFs and their shortening were based on the chromosome fragmentation method of VOLLRATH et al. 1988 . Similar to CFs, the TCFs affect the segregation of the chromosomes to which they are homologous, and the length of the TCF determines the magnitude of this effect. The two ends of the TCFs carry sequences from the fragmentation vectors and two different genetic markers, which enable us to follow the presence of the TCF and its involvement in recombination events. Tetrad analysis of diploid yeast strains that contain TCFs revealed tetrads with unexpected segregation of the two genetic markers at the two ends of these fragments. In a detailed physical analysis of these tetrads we show that this unexpected segregation results from the replacement of the genetic marker at one end by sequences of the genetic marker from the other end of the same TCF (R. SHEMESH and G. SIMCHEN, unpublished results). Our results demonstrate intensive meiotic recombination between the two termini of the TCF. We suggest that these processes are initiated by the telomeric and/or the telomere-associated sequences.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Plasmids:
Fragmentation plasmids for generating the CFs were constructed by cloning restriction fragments into the fragmentation vector YCFT4 [YCF4 of VOLLRATH et al. 1988 , modified by removing the sup11 gene]; YCFT4 carries the bla gene and the origin of replication of pBR322, the gene URA3, 1.3 kb from the Y' (the yeast telomere-associated sequence), the centromere of chromosome IV, and a polylinker. To create pTA695, a 2.2-kb BamHI-BglII fragment containing part of the gene CYH2 (from YEpCYH1, a gift from J. Warner) was cloned into BamHI-digested YCFT4. To create pTA773, a 1.8-kb EcoRI-BglII fragment containing the 3' end and the region distal to the gene LEU1 was cloned into EcoRI-BglII-digested YCFT4. The fragmentation plasmids for shortening the CFs and generating the TCFs were based on the vector YCFT3, which is identical to YCFT4 except that it lacks the centromere (ARBEL 1992 ). To create pTA697, the same BamHI-BglII fragment used to create pTA695 was cloned into BamHI-BglII-digested YCFT3. To create pTA699, a 1.8-kb BglII fragment containing the gene CUP2 was cloned into BamHI-BglII-digested YCFT3. To create pTA744, a 1.7-kb BamHI-BglII fragment containing part of the gene TRP5 (M. GOLDWAY, personal communication) was cloned into BamHI-digested YCFT3.

To replace the genetic markers on the CFs and the TCFs, pTA738, pTA776, and pTA777 were constructed in the following way. The overhangs of the 3.2-kb BamHI fragment containing the gene ADE2 (from R854, a gift from G. S. Roeder) were filled in using the Klennow component of DNA polymerase I, and the fragment was then inserted between the sites PvuII and EcoRV of YCFT3 to create pTA738. Similarly, a 3.2-kb BamHI fragment containing the gene TRP5 (V. ZAKIAN, personal communication) was cloned between the PvuII and EcoRV sites of YCFT3 to create pTA776. Plasmid pTA777 was constructed by cloning the 1.5-kb PstI-BamHI fragment from YCFT3, which contains the Y', between the PvuII and NsiI sites of YIp5. YCFT3 was first digested with BamHI, the overhangs filled in by the Klennow fragment of DNA polymerase I, and then was digested with PstI.

Yeast strains:
DA10-F is a diploid strain whose parental haploids are HA19-F (MATa his4-t2 can1 ade2-d1 ura3-52 cyh2 leu1::HIS4) and 91H-3 (MAT lys2-201 his4-t2 ade2 ura3-52 trp5d); all the diploid yeast strains are isogenic to DA10-F. The different diploid strains (names starting with D) were constructed by transforming the haploid strain 91H-3 or its derivatives and then crossing them to HA19-F.

Strains carrying different TCFs were constructed in three steps. Diploid strain DC699: 91H-3 was transformed with pTA695 that had been digested with BamHI and BglII to generate the CF C-CYH2. Next, the URA3 marker on the CF was replaced by the ADE2 gene by transformation with pTA738 digested with PvuI and SalI. Finally, the CF was shortened by transformation with pTA699 digested with BglII to generate the TCF T-CUP2 (Figure 1). Strains DL697 and DL744: 91H-3 were transformed with pTA773 digested with BglII to generate the CF C-LEU1, and the marker URA3 was replaced by the gene ADE2 as described for C-CYH2. For DL697, the CF was shortened by transforming with pTA697 digested with BglII to generate the TCF L-CYH2 (Figure 1), and for DL744 the CF was shortened by transforming with pTA744 digested with BamHI and BglII to generate the TCF L-TRP5 (Figure 1). For strain DC699-R, the marker URA3 on the TCF T-CUP2 was first replaced by the gene TRP5 by transformation with pTA776 digested with PstI, followed by transformation with pTA777 digested with BglII and SalI. The final TCF carries the gene URA3 at the end of its long arm but lacks all the pBR322 sequences between URA3 and the Y', including the gene bla, coding for ß-lactamase (Figure 1).

View larger version (27K): In this window In a new window Download PPT slide   Figure 1. TCFs used in this work. All TCFs were created by shortening the left arm of chromosome VII. CFs were created by shortening the chromosome from proximal markers, using sequences from chromosome VII, cloned in YCFT4. TCFs were created by truncating the CFs from distal markers using sequences from chromosome VII, cloned in YCTF3. Strain DC699 contains the TCF T-CUP2 (165 kb), which carries chromosome VII sequences between the genes CYH2 and CUP2. Strain DL697 contains the TCF L-CYH2 (180 kb), which carries chromosome VII sequences between the genes LEU1 and CYH2. Strain DL744 contains the TCF L-TRP5 (45 kb), which carries chromosome VII sequences between the genes LEU1 and TRP5. Strain DC699-R contains the same TCF as DC699, with the difference that the bla containing sequences of pBR322 between URA3 and Y' (3.2 kb) were deleted. Restriction sites are indicated for the TCF ends: P, PstI; X, XbaI.

The generation of the CFs and the TCFs was confirmed by separating chromosome-size DNA on the CHEF-II apparatus (Bio-Rad, Richmond, CA). Replacements of genetic markers on CFs and TCFs were confirmed by Southern blots and hybridization of both chromosome-size DNA separated by CHEF gel electrophoresis and digested DNA separated on standard agarose gels.

Media and general procedures:
Yeast cells were grown vegetatively in the nonselective YEPD medium (1% yeast extract, 2% Bacto-pepton, and 2% dextrose); for selective growth, cells were grown on SC medium (synthetic complete, 0.17% yeast nitrogen base, 0.5% ammonium sulfate, and 2% dextrose) that contained all nutrients prescribed by the auxotrophic mutations, except one (i.e., SC-Ura, SC-Trp etc.). For plates, 1.5% Bacto-agar was added. Cells were sporulated by replica plating to sporulation (SPO) plates (0.25% yeast extract, 1.5% potassium acetate, 0.05% dextrose, 1.5% Bacto-agar, and supplemented with the required nutrients) and incubated at 30° for 4–5 days.

All yeast transformations were performed by the lithium acetate procedure of ITO et al. 1983 . Chromosome-size DNA was prepared according to SCHWARTZ and CANTOR 1984 and was separated in 1% SeaKem (FMC, Rockland, ME) agarose gels on the CHEF-DR II apparatus (Bio-Rad). Gels were blotted to Hybond-N (Amersham, Piscataway, NJ) membranes according to the manufacturer's recommendations. Radioactive probes were prepared with a random-primed DNA labeling kit (Boehringer Mannheim, Mannheim, Germany) according to the manufacturer's recommendations. For rehybridization, probes were removed from blots by boiling in a 0.1% SDS solution.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Construction of TCFs:
Using a modified version of the yeast fragmentation vectors developed by VOLLRATH et al. 1988 (and see MATERIALS AND METHODS), we constructed fragments that were derived from chromosome VII; we named these fragments TCFs because they were shortened from the telomeric side. The truncated fragments were generated in three steps. In the first step, "regular" CFs are generated, running from a certain point on the chromosome through the telomere of the same arm. In the second step, the URA3 marker on the fragment is replaced by ADE2. Finally, the fragments are shortened from the telomeric end to generate the truncated versions. The TCFs carry the sequences derived from the fragmentation vectors on both termini and are marked with two different genetic markers: ADE2 on the short arm and URA3 at the end of the long arm (Figure 1).

Expectations of recombination events involving TCFs:
During meiosis of a diploid containing a single TCF, the fragment can either recombine with chromosome VII or remain nonrecombinant. In the latter case the tetrad is expected to have two Ura⁺ Ade⁺ spores and two Ura^- Ade^- spores. The outcome of a reciprocal recombination event between the TCF and chromosome VII is expected to be a fragment that has regained the full-length arm, all the way to the chromosomal telomere, and a shortened chromosome; haploid progeny that carry only this shortened chromosome are expected to be nonviable (Figure 2). It means that when the TCF recombines with chromosome VII, four-viable-spore tetrads will be recovered only if the recombinant fragment and the shortened chromosome migrate to the same pole in meiosis I (Figure 2). The outcome of such an event is a tetrad with two Ura⁺ Ade⁺ spores and two Ura^- Ade^- spores, which is phenotipically indistinguishable from tetrads where the TCF did not recombine with chromosome VII. Among the tetrads with four viable spores of diploid strains containing a TCF, we do not expect any Ura^- Ade⁺ and/or Ura⁺ Ade^- spores. Such spores, however, could be expected in tetrads with only three viable spores (Figure 2B and Figure C). Such three-viable-spore tetrads were examined for strains with a TCF and the frequency of recombination between the TCF and chromosome VII was found to be frequent, although variable between strains (Table 1).

View larger version (24K): In this window In a new window Download PPT slide   Figure 2. Recombination between the TCF and chromosome VII. Three types of tetrads that can rise from recombination of the TCF with one of the two homologs. The occurrence of each type depends on the segregation of the homologs and the TCF. Markers are represented by A for ADE2 and U for URA3.

Analysis of tetrads with four viable spores derived from strains with TCFs:
In a tetrad analysis of strains with a TCF that was marked with URA3 and ADE2 at the two ends, up to 10% of the tetrads with four viable spores contained Ura^- Ade⁺ and/or Ura⁺ Ade^- spores (Table 2). The unusual tetrads were of three different types, as presented in Table 2; they all have two Ura^- Ade^- spores and at least one spore that is either Ura^- Ade⁺ or Ura⁺ Ade^-. In addition to the markers URA3 and ADE2 on the TCF, the strains have also six heterozygous markers on four chromosomes. These markers segregated 2:2 in almost all the regular or unusual tetrads; as expected from known frequencies of gene conversion, some 3:1 and 1:3 segregations of various markers were recovered at low frequency (data not shown). We found three types of unusual tetrads (Table 2). Type I tetrad contains one spore of Ura⁺ Ade^- phenotype and another of Ura^- Ade⁺. Type II tetrad contains one spore that is Ura⁺ Ade^- and one Ura⁺ Ade⁺. Type III tetrad has one spore Ura^- Ade⁺ and one Ura⁺ Ade⁺.

How did the unusual tetrads arise? As noted above, they could not be the outcome of reciprocal recombination between the TCF and chromosome VII. The following five possible explanations for the formation of the unusual tetrads were tested.

Alternative possible explanations for the formation of unusual tetrads:
(1) A nonreciprocal recombination process has occurred between the TCF and the homologous region on one of the chromosomes VII. The fragment has been lengthened to the chromosomal telomere by a long-range gene conversion event, with no reciprocal shortening of the chromosome. This may explain the Ura^- Ade⁺ spores, but not the Ura⁺ Ade^- ones, as there is no homology between the short arm of the TCF, where the ADE2 gene is located, and chromosome VII (HARRIS et al. 1993 ). (2) The genes at the termini of the TCFs in the unusual spore colonies were turned off, probably by a mechanism similar to the one suggested for the silencing of genetic markers at sites next to the telomeric sequences in chromosomes VII and V (GOTTSCHLING et al. 1990 ; PALLADINO and GASSER 1994 ). (3) The wild-type genes in the TCF were converted from the mutant alleles at their "native" locations. (4) One of the mutant genes at the native locations was converted from the wild-type gene on the TCF and the fragment itself was subsequently lost. (5) The TCF's termini interacted with other telomeres and lost genetic markers through gene conversion or reciprocal exchange. These interactions could occur between the two ends of the TCF, between an end of the fragment and a chromosomal telomere, or both. Such a gene conversion event could also transfer the genetic marker from the TCF to a chromosomal telomere and then could be followed by loss of the TCF (R. SHEMESH and G. SIMCHEN, unpublished results).

Haploid progeny that express one of the genetic markers of the TCF contain a TCF:
To distinguish among the above explanations, we first asked whether the haploid progeny that express only one of the genetic markers of the TCF actually contained the fragment. Of the 15 progeny checked by pulsed-field gel electrophoresis (CHEF) of chromosome-size DNA, all carried fragments. In one of the progeny the fragment reverted to the original length of the CF (before shortening), while the remaining 14 retained the size of the TCF (Figure 3). In one case, the fragment with the length of the original CF came from a Ura^- Ade⁺ spore colony, from a type III tetrad (Table 2). Having a fragment in all the unusual progeny that we examined excluded the possibility that one of the mutant genes at the "native" locations was converted from a wild-type gene on the TCF and that the TCF was subsequently lost (possibility 4).

View larger version (52K): In this window In a new window Download PPT slide   Figure 3. TCFs in spore colonies that express only one of the genetic markers. DNA was prepared from spore colonies derived from strain DC699 that express only one of two TCF markers (Table 2). The chromosomal DNA was separated using CHEF gel electrophoresis (see MATERIALS AND METHODS) and was transferred to a nylon membrane. Membranes were hybridized to either URA3 (a) or ADE2 (b). Only one of the two probes hybridized to the TCF, except in lane 1, where no TCF is present, and lane 2, where the original TCF is detected by both probes. The native chromosome, which carries the original, mutated copy of each marker, also hybridized to the probe (chromosome V for URA3 and chromosome XV for ADE2). (Lane 1) A haploid carrying no TCF. (Lane 2) A haploid carrying the TCF T-CUP2 (160 kb). (Lanes 3 and 4) Haploids from a type III tetrad (in lane 3 the TCF has regained the original length of the CF from which it was derived). (Lanes 5 and 6) Haploids from a type II tetrad. (Lanes 7–10) Haploids from type I tetrad (see Table 2). Variation in fragment length is probably due to different recombination sites, variation in telomere length due to recombination, or technical factors in the running of the gel, such as slight variations in salt concentrations of the plugs.

The unexpressed genes are not present on the TCFs:
To see whether the TCFs carried sequences of the unexpressed genes, the pulsed-field gels were blotted and the membranes were hybridized with the URA3 and ADE2 genes as probes. As can be seen in Figure 3, a and b, the TCFs from the Ura⁺ Ade^- colonies hybridize with URA3 but not with ADE2. The converse situation is found for the fragments from Ura^- Ade⁺ colonies: they hybridize with the ADE2 probe and not with URA3. As no sequences of the unexpressed genes could be detected on the TCFs, we excluded the possibilities that the unexpressed genes were either silenced by some epigenetic mechanism (2) or converted from the mutant alleles at the native sites (3). In all of the colonies examined, both URA3 and ADE2 hybridized only to the TCF and to their native chromosomes (namely, URA3 hybridized to the band representing chromosome V and ADE2 to the chromosome XV band; Figure 3, a and b).

The singly expressed marker is present at both ends of the TCF:
How are the genes URA3 and ADE2 eliminated from the TCFs? If the elimination occurred by recombinational interactions between the two ends of the same fragment, it was expected to involve transfer of genetic information from one terminus to the other. In this case, we should be able to detect sequences of the expressed genetic marker at both ends of the TCF. This would not be the case if the elimination occurred through interaction between the end of the TCF and a chromosomal telomere. Such an interaction would lead to loss of one marker at the end of the TCF, without the gain of the other marker at this particular end.

The 160-kb TCF T-CUP2, which spans the CYH2-CUP2 interval, contains a unique site for the restriction enzyme NotI (ARBEL 1992 ). Digestion of this TCF with NotI was expected to generate two short fragments, one that would hybridize with URA3 and a second that would hybridize with ADE2. Comparing NotI-digested DNA from haploid strain HC699 (carrying the T-CUP2 TCF) to its parental strain 91H-3 (without a fragment) reveals a new 115-kb fragment that hybridizes to the gene URA3 and a new 45-kb fragment that hybridizes to the gene ADE2 (Figure 4, a and b, lanes 1 and 2).

View larger version (39K): In this window In a new window Download PPT slide   Figure 4. TCFs that express only one marker carry the same marker on both ends. The chromosomal size genomic DNA presented in Figure 3 was fully digested with the restriction enzyme NotI, separated on 1% agarose gel using CHEF gel electrophoresis (see MATERIALS AND METHODS), and transferred to a membrane. Lanes 1–9 correspond to the lanes in Figure 3, and DNA was prepared from the same spore colonies. Membranes were hybridized to either URA3 (a) or ADE2 (b). Only one of the two probes hybridized to the TCF, except in lane 1, where no TCF is present, and lane 2, where the original TCF is detected by both probes. Except for lane 3, all other digested TCFs show the same marker on both fragments (45 and 115 kb). Note that the NotI fragment from the native chromosome, which carries the original, mutated copy of each marker, also hybridized to the probe [i.e., a chromosome V fragment (230 kb) for URA3 (a) and a chromosome XV fragment (220 kb) for ADE2 (b)]. Variation in fragment length is probably due to different recombination sites, variation in telomere length due to recombination, or technical factors in the running of the gel, such as slight variations in salt concentrations of the plugs.

NotI digestion of DNA from the unusual haploid progeny, and hybridization with either URA3 or ADE2, revealed that in all cases the Ura⁺ Ade^- haploids gave 115- and 45-kb bands, both of which hybridized with the URA3 gene, and the Ura^- Ade⁺ haploids gave two bands of these sizes that hybridized with the ADE2 gene (Figure 4, a and b). We conclude that, in almost all of these cases, there has been recombination between the two ends of the TCF without involvement of the chromosomal telomeres. The one exception is the Ura^- Ade⁺ spore that contains a fragment that was of the original CF size (Figure 3, lane 3). In this case, ADE2 hybridized only to a 45-kb band (Figure 4, lane 3).

Deletion of pBR322 sequences at the end of the long arm of the TCF:
The telomeres of the different chromosomes consist of homologous telomeric and subtelomeric sequences. The two ends of the TCF have a different structure and share with each other pBR322 sequences that were derived from the fragmentation vectors. We assumed that the homology between those pBR322 sequences could be the primary source for the very high rate of recombination events between the ends. It was shown previously that the bla gene of pBR322 (coding for ß-lactamase) functions as a hotspot for meiotic recombination when inserted in homologous positions in chromosome III (STAPLETON and PETES 1991 ). To examine this possibility, the whole pBR322 region between the Y' sequence and the gene URA3 (including bla) was deleted from the end of the long arm of the TCF in strain DC699 (see MATERIALS AND METHODS) to create CD699-R. The deleted region comprises ~60% of the nontelomeric homology between the two ends of the TCF. As seen in Table 2, this reduction in homology between the ends did not reduce the frequency of unusual tetrads.

The unusual haploid progeny of strain DC699-R were examined in the same way as those of strain DC699 and revealed exactly the same results. All 8 haploids examined (out of 18 four-viable-spore tetrads that showed TCF marker loss; see Table 2) contained a fragment the size of the TCF, and the fragments hybridized either to URA3 (DNA from Ura⁺ Ade^- spores) or to ADE2 (DNA from Ura^- Ade⁺ spores), but not to both probes (data not shown). Digestion of DNA with NotI and hybridization with URA3 and ADE2 revealed that in this case, too, all the fragments were carrying sequences of the expressed marker on both ends (Figure 5). These results indicate that homology for the pBR322 sequences distal to the genetic marker (bla) is not required for the induction of the recombination processes between the two ends of the TCFs. The proximal pBR322 sequence (1.2 kb including tet, Figure 1) could play a role in recombination between the ends of the TCF. Disruption of the homology between the remaining pBR322 sequence near the URA3 marker (tet) and the sequence near the ADE2 marker, which is disrupted by CEN4, seems to affect the direction of end-to-end recombination of the TCF (Table 2). Fewer type III than type II tetrads were found for DC699-R compared with DC699 (a contingency ²_[1] is 4.79 without Yates correction, 0.05 > P > 0.025, or 3.06 with the correction, and 0.10 > P > 0.05). The results, thus, do not exclude the involvement of these sequences in the recombination events (see DISCUSSION).

View larger version (36K): In this window In a new window Download PPT slide   Figure 5. A TCF from which most pBR322 sequences were deleted behaves like the original TCF. Chromosomal DNA was produced from spore colonies of DC699-R, a strain derived from DC699, in which the TCF was modified. The bla gene and all the pBR322 sequences between the URA3 marker and Y' were deleted. The spore colonies expressed only one of the two TCF markers. DNA was partially digested by NotI for 5 hr and separated using CHEF gel electrophoresis (see MATERIALS AND METHODS). Gel was blotted to a membrane for hybridization. Membranes were hybridized to either URA3 (a) or ADE2 (b). The TCFs hybridized to one of the two probes except in lane 1, where a normal TCF is detected by both probes, and each TCF fragment hybridized to a different probe. In all other lanes (2–9) the full-length TCF (160 kb) carries only one of the two markers, and the digested TCFs carry the same marker, either URA3 or ADE2 on both fragments (45 and 115 kb).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

A novel type of CF was constructed by shortening the long arm of "regular" CFs, using the fragmentation method of VOLLRATH et al. 1988 . These TCFs carry at both ends sequences that originated from the fragmentation vector, and the ends are marked with two different genetic markers, the gene ADE2 on the short arm and URA3 at the end of the long arm.

During meiosis of diploid cells carrying a TCF, the two termini of these fragments interact to give a high level of meiotic recombination. These events were observed in up to 10% of the four-viable-spores tetrads. In these cases one or two of the spores contained a TCF that had lost one of its terminal genetic markers. None of the tetrads had two spores with TCFs that lost the same marker (i.e., two Ura⁺ Ade^- or two Ura^- Ade⁺ spores). Such tetrads are expected from recombination between the two ends of a TCF during the mitotic growth prior to meiosis. Therefore, we could conclude that all the events occurred during meiosis.

In only one case did an unusual haploid progeny (from type II tetrad, Table 2) not result from recombination between the two termini of the TCF. In this case, a gene conversion-like event between the TCF and chromosome VII has probably occurred, reverting the fragment to the original size of the CF before its shortening. As this tetrad had four viable spores, the full-length CF could not result from reciprocal recombination with chromosome VII. Separating chromosome-size DNA from the four haploid progeny of this tetrad confirmed that each carried a full-length chromosome VII (data not shown). Except for this case, all the unusual tetrads are the result of recombination between the two termini of a TCF.

There are three types of unusual tetrads (Table 2), in all of which (as in the "regular" tetrads) two of the spores do not contain a fragment (and are Ura^- Ade^-). The two other spores in each of the tetrads contain a TCF. In type I tetrads, there is one Ura⁺ Ade^- spore and one Ura^- Ade⁺ (Table 2); the TCF in one spore carries the gene UAR3 at both termini while the TCF in the other spore carries ADE2 at both termini. In types II or III tetrads (Table 2), one of the spores carries the two genetic markers while the second carries the same marker at both termini. We suggest that during meiosis the single TCF folds on itself, enabling the homologous sequences at the two ends to interact. Type III tetrads result from reciprocal recombination between the ends of different sister chromatids, while types II and III are the outcome of nonreciprocal transfer (gene conversion) of sequences from one end of the TCF to the other. Reciprocal recombination between the ends of the same chromatid would give rise to TCF with the gene ADE2 at the end of the long arm and the gene URA3 at the short arm, but such an event would be phenotypically indistinguishable from the regular tetrads. This means that the actual number of reciprocal recombination events is probably twice the number recovered and that recombination between the two termini of a TCF actually occurs in more meioses than identified in our experiments.

Three different TCFs were examined (Table 2), and all gave high proportions of the unusual tetrads. These three TCFs are homologous to different regions of the left arm of chromosome VII, so each has different unique sequences adjacent to the sequences of the fragmentation vector. Apparently the high-level recombination is not greatly influenced by the size of the TCF, but it might differ due to the presence of chromosome VII unique sequences. Higher differences between different TCF-carrying strains regarding the frequency of recombination between the TCF and chromosome VII were observed (Table 1).

There are three types of homologies between the two ends of the TCFs. These are sequences of the bacterial plasmid pBR322 and part of the subtelomeric element Y', both originating from the fragmentation vectors (Figure 1), and the telomeric (C–A_[1–3])_n elements that were acquired by the ends after the fragmentation (VOLLRATH et al. 1988 ). The largest homology between the ends consists of the pBR322 sequences. By deleting the pBR322 sequences between the Y' and the gene URA3 from the end of the long arm of the TCF, the pBR322 homology between the ends was reduced by ~60%. The remaining homology is effectively even smaller as the proximal pBR322 segment is interrupted by the 1.4-kb CEN4 fragment, leaving ~1.2 kb on the short arm, and the remaining 0.3 kb on the other side of the centromere. Deleting the region between Y' and the gene URA3 at the end of the long arm of the TCF deletes the gene bla. This bacterial sequence was shown to function as a hotspot for meiotic recombination when inserted at homologous positions into chromosome III, but not when present at one homolog only (STAPLETON and PETES 1991 ). We expected that if the pBR322 homology was the main source for recombination between the ends of the TCF, this deletion would reduce the recombination level markedly. As shown in Table 2, there is no reduction in the overall end-to-end recombination level in the TCF that was deleted for this region of the long arm. This means that the deleted pBR322 sequences do not induce the high level of recombination between the ends of the TCF. However, there is a significant reduction in the frequency of type III tetrads in the deleted strain. This could imply an involvement of the deleted pBR322 sequences in gene conversion, in that their absence restricts the donation of ADE2 from the short arm of the TCF to the end of the long arm (but not the reciprocal donation of URA3). A "hot" region for recombination was found at the tet gene of pBR322 (LICHTEN and GOLDMAN 1995 ), which might influence both the high frequency of end-to-end recombination of the TCF and its directional character after deletion of the distal, pBR322 bla sequence. The pBR322 sequences might also play a role in determining the recipient during gene conversion. If the chromatid that initiates recombination is the recipient of genetic information in gene conversion events (HABER et al. 1991 ), the deleted end is expected to be unable to serve as a recipient. In addition to the pBR322 sequences, two other regions may serve as initiators of recombination between the ends of the TCF, namely, the subtelomeric Y' element and/or the telomeric sequences.

An alternative explanation to the effect of the deletion might be that it somehow stabilizes the interaction and/or pairing between the ends, thus leading to an increase in reciprocal exchanges instead of the two gene conversion types. Type I of unusual tetrads (Table 2) is most likely a result of reciprocal recombination that resolves in the region of pBR322 homology, proximal to the genetic markers (Figure 1).

The occurrence of double strand breaks (DSBs) during meiosis in S. cerevisiae has been documented at several distinct sites that are associated with elevated meiotic recombination (recombination hotspots; see ZENVIRTH et al. 1992 , ZENVIRTH et al. 1997 ; KLEIN et al. 1996A , KLEIN et al. 1996B ). Lately it was shown that meiosis-induced DSB sites are highly correlated with regions that are hypersensitive to DNAse I in chromatin obtained from vegetatively growing cells (LICHTEN and GOLDMAN 1995 ). A strong hypersensitive site for DNAse I was shown to occur in the junction between the telomeric (C–A_[1–3])_n sequence and the adjacent subtelomeric region, suggesting a hotspot for meiotic recombination in this region. So the junction between the telomeric and subtelomeric regions might initiate high-level recombination between the ends of both normal chromosomes and TCFs. In this case it might have been expected that the ends of the TCF would interact with each other as well as with the chromosomal telomeres. We find, however, that all the recombination events but one appear to have involved the two termini of the TCF, without involvement of the chromosomal telomeres.

There are two main families of Y' subtelomeric elements, differing primarily in their sizes (LOUIS and HABER 1990B ), A comparison of Y' sequences from different chromosomes and different yeast strains revealed high structural similarity between similar size Y's within strains. On the other hand, Y's from different size classes were as different within strains as between strains (LOUIS and HABER 1990B ). The absence of recombination between the TCF and other chromosomes suggests that Y's belonging to different classes on different chromosomes are less likely to interact. The termini of the TCFs carry a relatively small portion of the same Y' element. The general structure of the two ends of a TCF is therefore similar, but very different from the structure of the chromosomal ends. Thus the observed interactions between the ends of the TCF are in agreement with the similarity between the Y' elements they contain and the pattern of interactions observed between Y' elements at the chromosomal ends. It is possible that the enhanced recombination observed between the termini of the TCFs is also influenced by the small size of the TCF (T-CUP2 is 160 kb long, while the shortest yeast chromosome is 240 kb long). This might suggest that the ends of the TCF may easily find each other by the TCF folding on itself. However, similar recombination events were also observed with larger TCFs (485 kb long, R. SHEMESH and G. SIMCHEN, unpublished results).

We believe that even though the telomeres have an effect on recombination, for the most part recombination does not occur between telomeres or between subtelomeric sequences of nonhomologous chromosomes. When a single TCF is present, the subtelomeric sequences and telomeres might mediate interaction between its ends.

	FOOTNOTES

¹ These authors contributed equally to this work.

	ACKNOWLEDGMENTS

This work was supported by a grant from the U.S.-Israel Binational Science Foundation. We thank Drs. Michael Lichten, Shoshana Klein, and two anonymous reviewers for suggestions and comments on the manuscript.

Manuscript received July 13, 1998; Accepted for publication August 12, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ARBEL, T., 1992 Meiotic nondisjunction of homologous chromosomes in the yeast Saccharomyces cerevisiae. Ph.D. Thesis. The Hebrew University of Jerusalem, Jerusalem, Israel.

BIESSMANN, H., J. M. MASON, K. FERRY, M. D'HULST, and K. VALGEIRSDOTTIR et al., 1990  Addition of telomere-associated HeT DNA sequences "heals" broken chromosome ends in Drosophila.. Cell 61:663-673[Medline].

CHARRON, M. J., E. READ, S. R. HAUT, and C. A. MICHELS, 1989  Molecular evolution of the telomere-associated MAL loci of Saccharomyces cerevisiae.. Genetics 122:307-316[Abstract/Free Full Text].

DUNN, B., P. SZAUTER, M. L. PARDUE, and J. W. SZOSTAK, 1984  Transfer of yeast telomeres to linear plasmids by recombination. Cell 39:191-201[Medline].

GOLDWAY, M., T. ARBEL, and G. SIMCHEN, 1993  Meiotic nondisjunction and recombination of chromosome III and homologous fragments in Saccharomyces cerevisiae.. Genetics 133:159-169[Abstract].

GOTTSCHLING, E. D., O. M. APARICIO, B. L. BILLINGTON, and V. A. ZAKIAN, 1990  Position effect at S. cerevisiae telomeres: reversible repression of POL II transcription. Cell 63:751-762[Medline].

HABER, J. E., W. Y. LEUNG, R. H. BORTS, and M. LICHTEN, 1991  The frequency of meiotic recombination in yeast is independent of the number and position of homologous donor sequences: implication for chromosome pairing. Proc. Natl. Acad. Sci. USA 88:1120-1124[Abstract/Free Full Text].

HARRIS, S., K. S. RUDNICKI, and J. E. HABER, 1993  Gene conversions and crossing over during homologous and homologous ectopic recombination in Saccharomyces cerevisiae.. Genetics 135:5-16[Abstract].

ITO, H., Y. FUKUDA, K. MURATA, and A. KIMURA, 1983  Transformation of intact yeast cells treated with alkali cations. J. Bacteriol. 153:163-168[Abstract/Free Full Text].

KLEIN, S., D. ZENVIRTH, V. DROR, A. B. BARTON, and D. B. KABACK et al., 1996a  Patterns of meiotic double-strand breakage on native and artificial yeast chromosomes. Chromosoma 105:276-284[Medline].

KLEIN, S., D. ZENVIRTH, A. SHERMAN, K. RIED, and G. RAPPOLD et al., 1996b  Double-strand breaks on YACs during yeast meiosis may reflect meiotic recombination in the human genome. Nat. Genet. 13:481-484[Medline].

LICHTEN, M. and A. S. H. GOLDMAN, 1995  Meiotic recombination hotspots. Annu. Rev. Genetics 29:423-444[Medline].

LOUIS, E. J., 1995  The chromosome ends of Saccharomyces cerevisiae.. Yeast 11:1553-1573[Medline].

LOUIS, E. J. and J. E. HABER, 1990a  Mitotic recombination among subtelomeric Y' repeats in Saccharomyces cerevisiae.. Genetics 124:547-559[Abstract].

LOUIS, E. J. and J. E. HABER, 1990b  The subtelomeric Y' repeat family in Saccharomyces cerevisiae: an experimental system for repeated sequence evolution. Genetics 124:533-545[Abstract].

PALLADINO, F. and S. M. GASSER, 1994  Telomere maintenance and gene repression: a common end? Curr. Opin. Cell Biol. 6:373-379[Medline].

PLUTA, A. F. and V. A. ZAKIAN, 1989  Recombination occurs during telomere formation in yeast. Nature 337:429-433[Medline].

SCHWARTZ, D. C. and C. R. CANTOR, 1984  Separation of yeast chromosome-sized DNA molecules by pulsed field gradient gel electrophoresis. Cell 37:67-75[Medline].

STAPLETON, A. and T. D. PETES, 1991  The Tn3 ß-lactamase gene acts as a hotspot for meiotic recombination in yeast. Genetics 127:39-51[Abstract].

SZOSTAK, J. W. and E. H. BLACKBURN, 1982  Cloning yeast telomeres on linear plasmid vectors. Cell 29:245-255[Medline].

VOLLRATH, D., R. W. DAVIS, C. CONNELLY, and P. HIETER, 1988  Physical mapping of large DNA by chromosome fragmentation. Proc. Natl. Acad. Sci. USA 85:6027-6031[Abstract/Free Full Text].

ZAKIAN, V. A. and H. M. BLANTON, 1988  Distribution of telomere-associated sequences on natural chromosomes in Saccharomyces cerevisiae.. Mol. Cell. Biol. 8:2257-2260[Abstract/Free Full Text].

ZENVIRTH, D., T. ARBEL, A. SHERMAN, M. GOLDWAY, and S. KLEIN et al., 1992  Multiple sites for double-strand breaks in whole meiotic chromosomes of Saccharomyces cerevisiae.. EMBO J. 11:3441-3447[Medline].

ZENVIRTH, D., J. LOIDL, S. KLEIN, A. ARBEL, and R. SHEMESH et al., 1997  Switching yeast from meiosis to mitosis: double-strand break repair, recombination and synaptonemal complex. Genes Cells 2:487-498[Abstract].

This article has been cited by other articles:

				A. B. Barton, Y. Su, J. Lamb, D. Barber, and D. B. Kaback A Function for Subtelomeric DNA in Saccharomyces cerevisiae Genetics, October 1, 2003; 165(2): 929 - 934. [Abstract] [Full Text] [PDF]