Maximal Stimulation of Meiotic Recombination by a Yeast Transcription Factor Requires the Transcription Activation Domain and a DNA-Binding Domain

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

Communicating editor: R. S. HAWLEY

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The DNA sequences located upstream of the yeast HIS4 represent a very strong meiotic recombination hotspot. Although the activity of this hotspot requires the transcription activator Rap1p, the level of HIS4 transcription is not directly related to the level of recombination. We find that the recombination-stimulating activity of Rap1p requires the transcription activation domain of the protein. We show that a hybrid protein with the Gal4p DNA-binding domain and the Rap1p activation domain can stimulate recombination in a strain in which Gal4p-binding sites are inserted upstream of HIS4. In addition, we find recombination hotspot activity associated with the Gal4p DNA-binding sites that is independent of known transcription factors. We suggest that yeast cells have two types of recombination hotspots, (transcription factor dependent) and ß (transcription factor independent).

RECOMBINATION events are not distributed evenly along eukaryotic chromosomes. Regions with high (hotspots) and low (coldspots) levels of recombination exist (reviewed by LICHTEN and GOLDMAN 1995 ). In the yeast Saccharomyces cerevisiae, meiotic recombination hotspots are usually located between, rather than within, genes and are associated with high levels of local meiosis-specific double-strand DNA breaks (DSBs). These breaks occur within regions of "open" chromatin (WU and LICHTEN 1994 ), although the patterns of DSBs do not reflect in detail the patterns of nuclease sensitivity in the chromatin (FAN et al. 1995 ; WU and LICHTEN 1995 ).

The HIS4 hotspot region contains binding sites for the transcription activators Rap1p, Gcn4p, Bas1p, and Bas2p (ARNDT et al. 1987 ; TICE-BALDWIN et al. 1989 ; DEVLIN et al. 1991 ). Mutations that eliminate the binding sites for Rap1p or Bas2p cause loss of hotspot activity (WHITE et al. 1991 , WHITE et al. 1993 ). In addition, loss of Bas1p or Bas2p causes loss of hotspot activity (WHITE et al. 1993 ), whereas loss of Gcn4p has little effect (WHITE et al. 1992 ). Although these results suggest a possible linkage between transcription and recombination, deletion of the TATAA sequence, required for efficient expression of HIS4, has no effect on hotspot activity (WHITE et al. 1992 ). Thus, transcription factors are required for hotspot activity in a role that is not directly linked to transcription (WHITE et al. 1993 ). Similar conclusions have been reached in recent studies of the M26 hotspot of S. pombe (KON et al. 1997 ).

A number of observations indicate that the transcriptional activator Rap1p is important for hotspot activity at the HIS4 locus. First, as described above, alteration of the Rap1p-binding site upstream of HIS4 eliminates hotspot activity (WHITE et al. 1991 ). Second, overproduction of Rap1p stimulates hotspot activity and this effect is eliminated in strains with a mutation of the Rap1p-binding site (WHITE et al. 1991 ). Third, when the normal transcription-factor-binding sites upstream of HIS4 are replaced with a 51-bp insertion of telomeric DNA, strong hotspot activity is detected (WHITE et al. 1993 ; FAN et al. 1995 ); telomeric sequences bind Rap1p (LONGTINE et al. 1989 ; LIU and TYE 1991 ) as well as a number of other proteins. The DSBs associated with the telomeric insertion upstream of HIS4 occur immediately adjacent to, but not within, the telomeric repeats (XU and PETES 1996 ). In addition, duplication of the Rap1p site that is present in a single copy in the sequences upstream of HIS4 in wild-type strains strongly stimulates recombination, and mutation of one of these sites results in loss of this stimulation (WHITE et al. 1993 ; FAN et al. 1995 ). Thus, Rap1p, in the absence of the other transcriptional activators of HIS4 expression, can stimulate hotspot activity.

Rap1p has diverse cellular roles (reviewed by SHORE 1994 ). Rap1p is an essential DNA-binding protein, and this binding induces a bend in DNA in vitro (GILSON et al. 1993 ). Rap1p is required for the activation of transcription at a number of loci (SHORE 1994 ) including HIS4 (DEVLIN et al. 1991 ). At other loci (the silent MAT loci and the telomeres), Rap1p represses transcription (KURTZ and SHORE 1991 ; HARDY et al. 1992 ; MORETTI et al. 1994 ). As described above, Rap1p binds to telomeres and different mutant Rap1 proteins result in either telomere lengthening or shortening (reviewed by SUSSEL and SHORE 1991 ; KYRION et al. 1992 ); on the basis of these and other results, it has been suggested that telomere length is regulated in yeast by a mechanism that "counts" the number of Rap1p molecules bound to the telomere (MARCAND et al. 1997 ). Rap1p localizes in clusters at the nuclear periphery, presumably reflecting these telomeric associations (KLEIN et al. 1992 ). Finally, Rap1p can stimulate the formation of quadruplexes of telomeric DNA (GILSON et al. 1994 ; GIRALDO and RHODES 1994 ).

Some of the functions described above require Rap1p to interact with other proteins, including Sir3p, Sir4p, Rif1p, and Rif2p (HARDY et al. 1992 ; KLEIN et al. 1992 ; MORETTI et al. 1994 ; WOTTON and SHORE 1997 ). The interactions with Sir3p and Sir4p are required for silencing of gene expression at the telomeres (MORETTI et al. 1994 ) and localization of Rap1p clusters to the nuclear periphery (PALLADINO et al. 1993 ), whereas the interactions with Rif1p and Rif2p are required for normal telomere length regulation as well as telomere silencing (HARDY et al. 1992 ; WOTTON and SHORE 1997 ).

The domains of the 827-amino-acid Rap1p required for the activities described above are diagrammed in Figure 1. The region of the protein between amino acids 43 and 279 is required for the bend induced in DNA by Rap1p binding (GILSON et al. 1993 ). The DNA-binding region extends from amino acids 361 to 596 (HENRY et al. 1990 ). The acidic transcriptional activation domain (amino acids 630 to 695) is adjacent to the DNA-binding domain (HARDY et al. 1992 ). The carboxy terminal quarter of the protein contains the domains required for telomeric silencing (amino acids 667–827; BUCK and SHORE 1995 ) and the regulation of telomere length (amino acids 726–827; SUSSEL and SHORE 1991 ; LUSTIG and PETES 1993 ). This region also has binding sites for Sir3p, Sir4p, Rif1p, and Rif2p (HARDY et al. 1992 ; MORETTI et al. 1994 ; WOTTON and SHORE 1997 ).

View larger version (36K): [in this window] [in a new window] [Download PPT slide]   Figure 1. Rap1p functional domains in Rap1p deletion and Gal4p/Rap1p fusion proteins. The full-length Rap-1p is 827 amino acids. The various functional domains of the proteins [summarized in SHORE 1994 and references therein] are shown at the top of the figure. The extent of the deletions encoded by the rap1-17 and rap1-N mutant alleles and the portion of Rap1p in various Gal4p/Rap1 fusion proteins (plasmids encoding these fusions indicated on the right) are shown at the bottom of the figure. All fusion proteins contain the DNA-binding domain of Gal4p (amino acids 1–147 of Gal4p, indicated by zigzag lines).

Given the diverse cellular functions of Rap1p, a number of explanations of the recombination-stimulating effect of the protein are possible. First, the binding of Rap1p may open the chromatin, allowing access of the enzyme complex that forms the initiating double-strand DNA breaks; this complex presumably contains Spo11p (KEENEY et al. 1997 ), which has been shown to possess DNA double-strand endonucleolytic activity during meiosis. Second, Rap1p may interact directly with proteins of the recombination-initiating complex, tethering the complex to the hotspot region. Third, the DNA bend induced by Rap1p binding may act as a recombinogenic signal. Fourth, binding of Rap1p to the hotspot may result in localization of this region to a cellular compartment that has high levels of recombination. As described below, by analyzing strains with various deletions of Rap1p and by examining Gal4p/Rap1p fusions, we show that the maximal recombination-stimulating activity of Rap1p requires both DNA binding and activation domains. The domains of the protein required to bend DNA, to control telomere silencing and telomere length, and to localize Rap1p to the nuclear periphery are not required to stimulate recombination, eliminating explanations 3 and 4 above. Further, as these dispensable domains contain the regions of the protein required for interactions with Sir3p, Sir4p, Rif1p, and Rif2p, none of the known Rap1p-interacting proteins are involved in meiotic recombination initiation at HIS4.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Media:
Standard media were utilized (GUTHRIE and FINK 1991 ). Sporulation plates contained 1% potassium acetate, 0.1% yeast extract, 0.05% glucose or galactose, 6 µg/ml adenine, and 2% agar.

Plasmids:
The plasmids pAJL1000 and pDTK109 contain altered forms of RAP1 inserted into the CEN-containing vector pRS316 (SIKORSKI and HIETER 1989 ). The plasmid pAJL100 (provided by A. Lustig) has the rap1-17 allele, encoding Rap-1p with a C-terminal deletion, whereas the plasmid pDTK109 contains the rap1-N allele, encoding a Rap1p lacking amino acids near the N terminus (Figure 1). The pDTK109 plasmid was constructed by joining an EcoRI-XbaI fragment containing the rap1-N allele [derived from pM672 (provided by D. Shore)] with EcoRI-XbaI-treated pRS316.

The plasmids pDTK116-BC, pDTK118-BAC, pDTK121-BA contain fusion genes encoding proteins with the DNA-binding domain from Gal4p (amino acids 1–147 of Gal4p) attached to various Rap1p-derived domains inserted in the vector pRS316; the synthesis of the fusion gene in these plasmids is controlled by the RAP1 promoter. The portion of the Rap1p represented in each plasmid is shown in Figure 1. The plasmid pDTK116-BC was constructed by transferring the EcoRI-XbaI fragment from p136 (BUCK and SHORE 1995 ) into EcoRI/XbaI-treated pRS316. The plasmid pDTK118-BAC was constructed by transferring a KpnI-NotI fragment encoding the fusion protein (derived from p292; BUCK and SHORE 1995 ) into KpnI/NotI-treated pRS316.

The plasmid pDTK121-BA was derived from pDTK118-BAC. pDTK118-BAC was treated with BamHI and NotI, deleting the DNA sequences encoding the Rap1p portion of the fusion protein, and a BamHI-NotI fragment encoding amino acids 630–695 of Rap1p was inserted. This DNA fragment was produced by using the oligonucleotides 739 (5' CAGTAGGATCCTATCTTACGCTATACCAGAAAAC 3') and 706 (5' GTATTGCGGCCGCTTATGGTGGAAAGCTTATGGTATC 3') in a PCR reaction with undeleted pDTK118-BAC DNA; this fragment was then treated with BamHI and NotI. The plasmid pDTK122-B (encoding a protein with the Gal4p DNA-binding domain but no Rap1p protein sequences) was constructed by related procedures. The double-stranded oligonucleotide formed by annealing 13030 (5' GATCTAATGATAA 3') and 13031 (5' GGCCTTATCATTA 3') was ligated to BamHI/NotI-digested pDTK118-BAC. The altered regions of pDTK121-BA and pDTK122-B were confirmed by sequencing.

The plasmid pQF1 was used to replace the transcription-factor-binding sites located upstream of the wild-type HIS4 gene with duplicated Gal4p-binding sites. Oligonucleotides with the sequences 5' TCGACGGAAGACTCTCCTCCGCACTGTATTACCATAGTACCGCGGAAGACTCTCCTCCGC 3' and 5' TCGAGCGGAGGAGAGTCTTCCGCGGTACTATGGTAATACAGTGCGGAGGAGAGTCTTCCG 3' were annealed together and inserted into the XhoI site of pPD5 (a YIp5 derivative; WHITE et al. 1993 ). The DNA sequences shown in boldface represent the Gal4p-binding sites (FELDMANN et al. 1994 ) and the DNA sequences located between the two binding sites are the same as those used previously to separate two Rap1p-binding sites (WHITE et al. 1993 ); these intervening sequences have been previously demonstrated to lack hotspot activity in the absence of Rap1p-binding sites.

The plasmids used to generate mutations of various yeast genes were the following: (1) pLSD64 (provided by D. Shore; rap1::LEU2 allele), (2) pKD18 (provided by E. T. Young; adr1-2::URA3 allele), (3) pBM2387 (provided by M. Johnston; gal4::hisG-URA3-hisG allele), (4) pNKY85 (provided by S. Keeney; leu2::hisG-URA3-hisG allele), and (5) pNKY349 (provided by N. Kleckner; rad50S).

Strains:
All strains used in this study were derived from the haploid strains AS13 (MATa leu2 ura3 ade6 rme1) and AS4 (MAT trp1 arg4 tyr7 ade6 ura3 spt22; STAPLETON and PETES 1991 ). The relevant genotypes of the haploid and diploid strains used in our study are summarized in Table 1 and Table 2.

The diploid strains DTK438 and DTK453 were used to examine the effects of rap1-N and rap1-17 on HIS4 hotspot activity. To construct an AS13-derived haploid strain with the rap1-N mutation, we transformed the haploid strain HF4U (FAN et al. 1995 ) with pDTK109 (encoding rap1-N) to yield DTK342. HF4U is isogenic with AS13 except for the following alterations: (1) replacement of HIS4 with his4-lopc, (2) replacement for wild-type RAD50 gene with the rad50S allele (ALANI et al. 1990 ), and (3) a spontaneous mutation of the URA3 gene used to insert rad50S sequences (FAN et al. 1995 ). The wild-type RAP1 gene in DTK342 was then inactivated by transformation with BglII/XbaI-treated pLSD64, a plasmid containing a rap1::LEU2 allele. The net result of these manipulations was DTK354, in which RAP1 was replaced with a plasmid-borne rap1-N allele.

To create an AS13 derivative bearing rap1-17, we crossed DTK354 (described above) to DTK355. DTK355 was derived from DNY25 (which contains a his4-lopc mutation; NAG et al. 1989 ) by using pGAL-HO to perform a mating-type switch (HERSKOWITZ and JENSEN 1991 ). The diploid DTK356 is a derivative of the DTK354 x DKT355 diploid that had lost the pDTK109 plasmid. The plasmid pAJL1000 (encoding rap1-17) was introduced into this strain, creating DTK359. This diploid was sporulated, and a plasmid-bearing spore with the genotype a rap1::LEU2 rad50S was identified (DTK366).

A similar process was used to create AS4-derived rap1-17 and rap1-N strains. A rad50S derivative of AS4 (DNY107; NAG and PETES 1993 ) was transformed with a BglII fragment derived from pNKY85 (strain DTK387) and a Ura^- derivative of the resulting transformant was selected on medium contain 5-fluoroorotate. This strain (DTK388) had the leu2::hisG allele. DTK388 was mated to LS1-leu2. LS1-leu2 was derived from LS1 (SYMINGTON and PETES 1988 ) by a two-step transplacement with the KpnI-treated plasmid pleu2-RI (provided by G. S. Roeder); this plasmid contains a mutant leu2 gene caused by "fill-in" of an EcoRI site in the coding sequence cloned in YIp5. LS1 is isogenic with AS4 except LS1 has the opposite mating type and lacks a number of restriction site alterations in the region between LEU2 and CEN3 (SYMINGTON and PETES 1988 ; STAPLETON and PETES 1991 ). The diploid formed by crossing DTK388 and LS1-leu2 (DTK389) was transformed with a BglII-XbaI fragment derived from pLSD64 to generate DTK400 (heterozygous for rap1::LEU2). DTK410 and DTK411 were created by transformation of DKT400 with pDTK109 (encoding rap1-N) and pAJL1000 (encoding rap1-17), respectively. DTK410 was sporulated and dissected, and a spore containing pDTK109 with the genotype rap1::LEU2 was isolated (DTK431). Similarly, a spore derived from DTK411 containing pAJL1000 with the genotype rap1::LEU2 was isolated (DTK451). The diploid DTK438 was constructed by mating DTK354 and DTK431, and the diploid DTK453 was generated by mating DTK366 and DTK451.

In the diploid strain QFY120, the transcription-factor-binding sites located upstream of HIS4 were replaced by duplicated Gal4p-binding sites. In addition, the strain was heterozygous for the his4-lopc allele. This strain was constructed by mating QFY32 and QFY33. The haploids QFY32 and QFY33 were derived from PD80 and PD63 (DETLOFF et al. 1992 ), respectively, by two-step transplacement with PvuII-treated pQF1d; the resulting substitution is his4-301. The genotypes and progenitor strains for PD80 and PD63 are shown in Table 1. Derivatives of QFY32 (QFY38) and QFY33 (QFY39) with mutant gal4 genes were constructed by transformation of the parental strains with SacI/KpnI-treated pBNM2387 DNA, followed by selection of Ura^- derivatives of the resulting Ura⁺ transformants. The diploid QFY122 was obtained by mating QFY38 and QFY39. Derivatives of QFY38 (QFY51) and QFY39 (QFY52) with a deletion of ADR1 were generated by transformation of the parental strains with SacI/SphI-treated pKD18. The diploid QFY132 was obtained by crossing QFY51 and QFY52.

To examine the recombination-stimulating effects of various Gal4p/Rap1p fusion proteins, we transformed QFY122 with the following plasmids (name of transformed strain in parentheses): pRS316 (DTK191), pDTK116-BC (DTK192), pDTK118-BAC (DTK193), pDTK121-BA (DTK197), and pDTK122-B (DTK206). A related set of strains was constructed by transforming the rad50S diploid QFY135 with the same plasmids: PRS316 (DTK216), pDTK116-BC (DTK217), pDTK118-BAC (DTK218), pDTK121-BA (DTK219), and pDTK-122-B (DTK220). QFY135 was constructed by mating the haploids QFY44 and PG31, gal4::hisG derivatives of QFY36 and QFY37, respectively. QFY36 and QFY37 were Ura⁺ rad50S transformants obtained by transformation of HF5U (QFY36) and HF6U (QFY37) with EcoRI/BamHI-treated pNKY349.

Meiotic analysis:
For genetic analysis, diploids were sporulated on plates at 18° for 4–6 days and dissected by standard methods onto plates containing rich growth medium; the colonies were then replica plated to various omission media. Spores derived from strains heterozygous for the his4-lop and his4-lopc alleles have high levels of postmeiotic segregation, which are detectable as sectored His⁺/His^- colonies on medium lacking histidine (NAG et al. 1989 ). To ensure that all sectored colonies were detected, we used a microscope to examine all colonies on histidine omission medium. For those strains containing an autonomously replicating plasmid, all data were derived from tetrads in which at least one spore colony retained the plasmid. For physical studies of meiotic cells (determination of the level of meiosis-specific DSBs), cells were sporulated in liquid medium at 24°, as described previously (NAG and PETES 1993 ).

Isolation and analysis of DNA and RNA:
DNA was isolated from meiotic or premeiotic cells by methods described previously (NAG and PETES 1993 ). To monitor DSBs at the HIS4 hotspot, we treated DNA samples with BglII and examined the resulting fragments by Southern analysis. The hybridization probe was an XhoI-XbaI fragment of plasmid pDN42 containing sequences derived from the 5' end of HIS4 (NAG and PETES 1993 ). Previous studies (FAN et al. 1995 ) show that the DNA fragment resulting from the DSB at the hotspot is ~2 kb and the unbroken BglII fragment is ~3 kb. The amount of hybridization to these two fragments was determined using the PhosphorImager (Molecular Dynamics, Sunnyvale, CA). The DSB percentage was calculated by dividing the amount of hybridization in the 2-kb fragment by the sum of the amounts of hybridization to the 2- and 3-kb fragments. Because the percentage of DSBs is affected by the efficiency of sporulation, we normalized the percentage of DSBs at the HIS4 hotspot to the percentage of DSBs at the ARG4 hotspot; DSBs at the ARG4 locus were detected using an EcoRV-BglII fragment derived from plasmid pAK1 as a hybridization probe for BglII-treated yeast DNA samples (FAN et al. 1995 ). The normalization was done by multiplying the percentage of DSBs at the HIS4 locus in each individual experiment by the correction ratio (maximum percentage of DSBs at the ARG4 locus divided by the percentage of DSBs at the ARG4 locus in each individual experiment).

Total RNA was prepared (SOLOMON et al. 1992 ) from cells grown in liquid cultures. Standard Northern analysis was done (SAMBROOK et al. 1989 ). The nylon filter with the transferred RNA was hybridized to a mixture of two probes, the same HIS4 probe used in the Southern analysis (described above) and an actin gene probe; the ACT1 probe was a 1-kb gel-purified HindIII-XhoI fragment of plasmid pGAL1-ACT1 (provided by R. Sia, Duke University). Levels of hybridization were quantified using a PhosphorImager (Molecular Dynamics).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

As described in the Introduction, Rap1p is required for the meiotic recombination hotspot observed upstream of HIS4. Rap1p has diverse cellular roles and the protein domains required for these roles have been characterized (Figure 1). To determine which protein domains of Rap1p were required for hotspot activity, we examined the recombination-stimulating properties of deletion derivatives of Rap1p. In addition, we studied the effects of fusion proteins with the Gal4p-binding domain and various segments of Rap1p on recombination in strains in which the normal HIS4 upstream sequences were replaced with Gal4p-binding sites. These two studies show that maximal hotspot activity requires proteins that have both a DNA-binding domain and a transcription activation domain.

Analysis of the effects of aminoterminal and carboxy-terminal deletions of Rap1p on HIS4 hotspot activity:
One convenient method of monitoring HIS4 hotspot activity is by measuring the frequency of aberrant segregation (non-2:2 segregation) of a heterozygous mutation within the HIS4 coding sequence by tetrad analysis. In many of our previous studies, we have used mutant alleles resulting from 26-bp palindromic insertions (his4-lopc or his4-lop) near the beginning of the HIS4 gene. In one such strain, DNY26 (NAG et al. 1989 ; FAN et al. 1995 ), the frequency of aberrant segregation of his4-lopc was 51% (Table 3). As described in the Introduction, this very high level of meiotic recombination is dependent on the binding of a number of transcription factors in the region upstream of HIS4, including Rap1p, Bas1p, and Bas2p. As shown in Table 3, a deletion that removes all of these transcription-factor-binding sites (his4-52, strain PD81) or the Rap1p-binding site (his4-51, strain MW118) substantially reduces the frequency of aberrant segregation (WHITE et al. 1991 ; DETLOFF et al. 1992 ); these alterations also eliminate meiosis-specific DSBs associated with hotspot activity (FAN et al. 1995 ). In addition, overproduction of Rap1p further stimulates HIS4 recombination (WHITE et al. 1991 ). Replacement of the wild-type HIS4 promoter with two copies of the Rap1p-binding site is sufficient to initiate recombination at wild-type levels, while three copies of the binding site create the strongest meiotic recombination initiation site yet observed (his4-202, MW154; Table 3). In general, the frequency of aberrant segregation of markers within the HIS4 coding sequence is correlated with the frequency of crossovers between HIS4 and the linked LEU2 gene (Table 3).

To determine the role of the nonessential functional domains of Rap1p in stimulating HIS4 hotspot activity, two diploid strains were constructed, each with a different mutant allele of RAP1. The strain DTK453 contained the rap1-17 mutant allele (KYRION et al. 1992 ), encoding a Rap1p with a C-terminal deletion (Figure 1). The diploid DTK438 contained rap1-N (GILSON et al. 1993 ), encoding a Rap1p with a deletion near the N terminus. Both strains were heterozygous for the his4-lopc allele. These strains were sporulated, tetrads were dissected, and we monitored the frequency of aberrant segregation of his4-lopc (Table 3).

The aberrant segregation of his4-lopc in DTK453 was 38%, significantly (P < 0.001 by Fisher exact test for both comparisons) above the level of aberrant segregation seen in a strain lacking the HIS4 upstream activation region (PD81) or in a strain that lacks only the Rap1p DNA-binding site (MW118). The aberrant segregation of his4-lopc in DTK438 (37%) was nearly identical to DTK453 and significantly greater than observed for PD81 or MW118. For both DTK453 and DTK438, the frequencies of aberrant segregation for his4-lopc were significantly reduced compared to the frequency observed in the wild-type strain DNY26.

These results suggest that the portions of Rap1p that are deleted in Rap1-17p and Rap1p-Np are not essential, but may contribute, to hotspot activity. Alternatively, because the mutated forms of Rap1p are expressed from a plasmid in DTK453 and DTK438, it is possible that the level of Rap1p is quantitatively different from the wild-type level of Rap1p. The strains with the mutated forms of Rap1p do not have a global reduction in meiotic recombination relative to the wild-type DNY26 strain. The gene conversion frequencies at the ARG4 locus were 18% (DTK438), 12% (DTK453), and 9% (DNY26). The recombination distances between ARG4 and the centromere in the three strains were 12 cM (DTK438), 16 cM (DTK453), and 12 cM (DNY26).

Strains with a null mutation of RAP1 are not viable. If this lack of viability is a consequence of the role of Rap1p as a transcription factor, Rap1-17p and Rap1-Np should be capable of stimulating HIS4 transcription. We confirmed this prediction by measuring the level of HIS4 mRNA (relative to an actin control mRNA) by Northern analysis (Figure 2). Control strains AS13 and HF4U (wild-type promoter sequences and wild-type Rap1p) had similar levels of HIS4 mRNA. A strain with a deletion of the Rap1p-binding site (HF3) had a level of HIS4 transcript at least 10-fold reduced from the control strains, whereas strains with the rap1-17 or rap1-N mutations had HIS4 mRNA levels that were reduced about 4-fold. Thus, the proteins encoded by these mutant genes retain some, but not all, of their transcription-stimulating activity.

View larger version (78K): [in this window] [in a new window] [Download PPT slide]   Figure 2. Rap1-Np and Rap1-17p stimulate HIS4 transcription. Haploid strains were grown in rich medium and total RNA was isolated from each strain. RNA was examined by Northern analysis using probes to HIS4 and ACT1. AS13 and HF4U have wild-type promoter regions at HIS4, whereas HF3 bears the his4-51 mutation (a disruption of the Rap1p-binding site in the HIS4 upstream activation region). DTK354 and DTK366 have wild-type HIS4 promoter sequences but contain the rap1-N and rap1-17 mutations, respectively.

As shown in Figure 1, Rap1-17p lacks the silencing domain, the domain required for telomere length maintenance, and the binding sites for Sir3p, Sir4p, and Rif1p. The results shown in Table 3 and Figure 2 indicate that none of these functions are essential for the hotspot-activation or transcription-activation functions of Rap1p. As expected from this conclusion, a deletion of SIR3 had no effect on DSB formation at the HIS4 hotspot (data not shown). Similarly, the region of Rap1p that generates a DNA bend upon binding (missing in Rap1-Np) is not required for hotspot activity or transcription activation. These results suggest that the essential recombination-stimulating activity of Rap1p might be the DNA-binding and transcription activation domains. Because Rap1p is essential, we could not examine the effects of deletions removing these domains by the same approach. Consequently, we replaced the Rap1p-binding sites upstream of HIS4 with Gal4p-binding sites and examined the transcription-stimulating effects of various activators that had the Gal4p DNA-binding domain and portions of Rap1p.

Stimulation of HIS4 gene expression and meiotic recombination activity by insertion of Gal4-binding sites:
As in the studies described above, we used diploid strains that were heterozygous for the allele his4-lopc and homozygous for various sequences upstream of HIS4. We first examined the effects of these alterations on HIS4 gene expression. As shown in Figure 3, strains with the wild-type upstream sequences (DNY26) or with a telomeric insertion (MW154, containing three Rap1p-binding sites replacing the wild-type upstream sequences) have strong His⁺ phenotypes (WHITE et al. 1993 ). A strain (PD81) that lacks transcription-factor-binding sites is His^- (NAGAWA and FINK 1985 ). We constructed the strain QFY120 in which the wild-type transcription-factor-binding sites were replaced by two Gal4p-binding sites separated by a short spacer (see MATERIALS AND METHODS). This substitution (his4-301) activated HIS4 gene expression weakly (Figure 3).

View larger version (28K): [in this window] [in a new window] [Download PPT slide]   Figure 3. HIS4 gene expression in strains with various transcription-factor-binding sites upstream of HIS4. Diploid strains were grown in rich growth medium to stationary phase. Undiluted samples were placed in one column of wells in a microtiter dish. Each sample was serially diluted (1:10) in water in adjacent wells within each row. A multipronged device was then used to transfer a small amount of the samples from the microtiter dish to a plate containing solid medium with or without histidine; the strains were then grown at 30° for (a) 3 and (b) 5 days. The relevant features of the strains are the following: DNY26 (wild-type HIS4 upstream sequences), MW154 (telomeric insertion with three Rap1p-binding sites replacing wild-type upstream sequences), PD81 (deletion of wild-type upstream sequences), QFY120 (replacement of wild-type upstream sequences with duplicated Gal4p-binding sites; GAL3 ADR1), QFY122 (replacement of wild-type upstream sequences with duplicated Gal4p-binding sites; gal4 ADR1), and QFY132 (replacement of wild-type upstream sequences with duplicated Gal4p-binding sites; gal4 adr1).

The activation of HIS4 gene expression in QFY120 in glucose-containing medium is somewhat surprising for two reasons. First, expression of Gal4p is repressed four- to seven-fold in cells grown on glucose-containing medium (GRIGGS and JOHNSTON 1991 ). Second, although Gal4p binds to DNA when cells are grown in the absence of galactose, the activation domain is blocked from functioning because of its interactions with Gal80p (JOHNSTON et al. 1987 ; MA and PTASHNE 1987 ), an interaction that is lost when the cells are grown in galactose (BHAT and HOPPER 1992 ).

To investigate in more detail the possibility of Gal4p-independent activation of HIS4 expression by the duplicated Gal4p-binding sites, we constructed a derivative of QFY120 (QFY122) that was homozygous for a deletion of GAL4. In addition, because the Gal4p-binding site shares sequence homology with the binding site recognized by the transcription factor Adr1p (CHENG et al. 1994 ), we constructed a derivative of QFY120 (QFY132) that lacked both GAL4 and ADR1 genes. Both QFY122 and QFY132 were weakly His⁺ (Figure 3), confirming that the Gal4p-binding sites confer some activation of HIS4 transcription that is independent of the binding of either Gal4p or Adr1p.

In these strains, we monitored the level of aberrant segregation for the his4-lopc allele (as a genetic measure of hotspot activity), as well as the frequency of crossing-over between HIS4 and the linked LEU2 gene (Table 3). QFY120, homozygous for the insertion of duplicated Gal4p-binding sites (his4-301), had a level of aberrant segregation (53%) for the his4-lopc allele that was close to that observed in the wild-type strain DNY26. Sporulation of QFY120 in medium containing galactose, instead of glucose, did not significantly elevate the frequency of aberrant segregation or crossovers above the frequency observed in cells sporulated in glucose-containing medium. In addition, derivatives of QFY120 in which GAL4 (QFY122) or GAL4 and ADR1 (QFY132) were mutated had about the same level of aberrant segregation and crossovers as QFY120 (Table 3). In summary, when duplicated Gal4p-binding sites were inserted upstream of HIS4, we observed recombination hotspot activity. Both the frequency of aberrant segregation of his4-lopc and the level of HIS4-LEU2 crossovers were elevated relative to the values observed in PD81, a strain that lacks hotspot activity. The hotspot activity of the Gal4p-binding sites, however, did not require the binding of Gal4p or Adr1p.

At least two possibilities exist for the hotspot activity observed in QFY120 and its derivatives. First, the duplicated Gal4p-binding sites might bind an unknown transcription factor that stimulates meiotic recombination at HIS4. Alternatively, some property of the DNA sequence of the insertion may be responsible for hotspot activity. These possibilities will be addressed in more detail in the DISCUSSION.

Analysis of the recombination-stimulating properties of fusion proteins containing the DNA-binding domain of Gal4p and activation domains of Rap1p in strains with Gal4p-binding sites upstream of HIS4:
Although Gal4p does not appear to contribute to hotspot activity (as described above), we previously concluded that Rap1p was capable of contributing to HIS4 hotspot activity (WHITE et al. 1991 , WHITE et al. 1993 ). To determine whether the differences in the behavior of these two transcription factors resided in the activation domain of the proteins, we examined the recombination-stimulating properties of proteins with the Gal4p DNA-binding domain and the Rap1p activation domain in strains with two Gal4p-binding sites upstream of HIS4. We transformed QFY122 (homozygous for duplicated Gal4p-binding sites and the gal4 mutation) with five different plasmids: (1) pRS316 (vector alone), (2) pDTK116-BC (binding domain of Gal4p and C terminus of Rap1p), (3) pDTK118-BAC (binding domain of Gal4p, complete activation domain of Rap1p and C terminus of Rap1p), (4) pDTK121-BA (binding domain of Gal4p and complete activation domain of Rap1p), and (5) pDTK122-B (binding domain of Gal4p without any Rap1p). The regions of Rap1p in each plasmid are shown in Figure 1. Although the fusion protein encoded by pDTK116-BC has part of the activation domain of Rap1p, this protein is incapable of activating transcription (HARDY et al. 1992 ).

These strains were sporulated and tetrads were dissected. As above, we monitored the aberrant segregation frequency of his4-lopc and crossovers between HIS4 and LEU2; only tetrads in which at least one spore retained the plasmid (signalling retention of the plasmid in the diploid cell giving rise to the spores) were included in our analysis (Table 4). As expected, QFY122 with the vector alone had the same frequency of recombination as QFY122 without any plasmid (compare Table 3 and Table 4). Three types of effects were seen in the derivatives of QFY122 with the fusion plasmids. In the strain with pDTK116-BC, both the frequency of aberrant segregation and the level of HIS4-LEU2 crossovers were significantly (P < 0.01) reduced compared to QFY122 with the vector alone. In the strains with pDTK118-BAC and pDTK121-BA, both the frequency of aberrant segregation and the level of crossovers between HIS4 and LEU2 were significantly (P < 0.01) elevated compared to QFY122 with only the vector. In the strain with the plasmid pDTK122-B, the frequency of aberrant segregation was significantly (P < 0.02) elevated, but the frequency of crossovers was not (P = 0.8), compared with QFY122 with the vector alone.

These results indicate that fusion proteins with the Gal4p-binding domain and a complete Rap1p activation domain (encoded by plasmids pDTK118-BAC and pDTK-121-BA) stimulate meiotic recombination more effectively than other fusion proteins. A protein with the Gal4p-binding domain with no Rap1p (encoded by pDTK122-B) had a smaller stimulatory effect on recombination, whereas a fusion protein with the Gal4p-binding domain, a truncated activation domain, and the C-terminal sequences derived from Rap1p (encoded by pDTK116-BC) suppressed recombination. These effects are unlikely to reflect differing amounts of different Gal4p/Rap1p proteins, because Western analysis indicates similar levels of expression for the different fusions (HARDY et al. 1992 ).

As first shown at the ARG4 locus (SUN et al. 1989 ), meiotic recombination hotspots in S. cerevisiae are sites for DSB formation. At HIS4, the level of DSB formation is roughly proportional to the level of hotspot activity, although a basal level of recombination is observed in the absence of a detectable DSB (FAN et al. 1995 ). To examine the level of DSBs at HIS4 associated with various fusion proteins, we introduced the same five plasmids described above into a derivative of QFY122 that was homozygous for rad50S (QFY135), a mutation that prevents the processing of DNA ends resulting from DSB formation (ALANI et al. 1990 ). We monitored DSBs at the HIS4 locus in each strain as described in MATERIALS AND METHODS. The frequencies of HIS4 DSBs (as a percentage of the total HIS4 DNA) are shown in Figure 4.

View larger version (43K): [in this window] [in a new window] [Download PPT slide]   Figure 4. Meiosis-specific DSBs at the HIS4 locus in strains in which the wild-type HIS4 upstream region was replaced with duplicated Gal4p-binding sites. All diploid strains depicted in this gure were homozygous for insertions of duplicated Gal4p-binding sites upstream of HIS4, gal4, and rad50S. In addition, each strain contained a plasmid with vector sequences alone (pRS316) or insertions of genes encoding various Gal4p/Rap1 fusion proteins (Figure 1). All plasmids of the latter class have the Gal4p DNA-binding domain. pDTK122-B encodes a protein with only the Gal4p DNA-binding domain. Other plasmids contain the following domains of Rap1p: pDTK116-BC (silencing and telomere length regulation domains; aa 653–827), pDTK118-BAC (activation, silencing, and telomere length regulation domains; aa 563–827); pDTK121-BA (activation domain; aa 630–695). Maximal levels of DSB formation require the Rap1p activation domain.

As expected, the highest levels of DSB formation (~25%) were observed in strains with the plasmids pDTK118-BAC and pDTK121-BA; these same plasmids were associated with the highest levels of recombination. The levels of DSB in the other three strains were considerably lower, between 3 and 8%. Although this result is consistent with the lower level of recombination observed in these strains, the level of DSBs does not appear to be directly proportional to the level of recombination. For example, although the frequency of aberrant segregation differs by less than a factor of two between the strain with the plasmid pDKT118-BA and the strain with the vector alone, the level of DSBs differs by about a factor of three. In summary, the analysis of DSBs confirms that the highest level of recombination in the QFY122 background requires both the Gal4p-binding domain and a complete Rap1p activation domain. Some hotspot activity (measured by genetic and physical techniques), however, results from the duplicated Gal4p-binding sites in the absence of the binding of known transcription factors.

Analysis of the effects of fusion proteins containing the DNA-binding domain of Gal4p and activation domains of Rap1p on gene expression in strains with Gal4p-binding sites upstream of HIS4:
We examined the QFY122 derivatives with the same plasmids used to study recombination for growth on medium with and without histidine. Although all strains grew at approximately the same rate on medium with histidine, growth on medium without histidine varied in the different strains (Figure 5). Strains with the plasmids pDTK118-BAC or pDTK121-BA formed colonies much more readily on medium lacking histidine than the strains with pRS316, pDTK116-BC, or pDTK122-B. All of these strains, unlike PD81, however, eventually formed colonies on the histidine-deficient medium. Thus, the same fusion proteins that maximally stimulate HIS4 meiotic recombination also maximally stimulate HIS4 gene expression.

View larger version (39K): [in this window] [in a new window] [Download PPT slide]   Figure 5. Analysis of the effects of various Gal4p/Rap1p fusion proteins on HIS4 gene expression in strains with duplicated Gal4p-binding sites upstream of HIS4. Strains were grown to stationary phases in medium lacking uracil (to force retention of plasmids) or rich medium (other strains). The resulting cultures were diluted and plated (as described in Figure 3 legend) on solid medium (a) containing or (b) lacking histidine. The relevant features of each strain are the following: DNY26 (wild-type HIS4 upstream sequences), PD81 (deletion of wild-type upstream sequences), QFY122 (replacement of wild-type upstream sequences with duplicated Gal4p-binding sites; gal4 ADR1), DTK191 (QFY122 + pRS-316 vector), DTK192 (QFY122 + pDTK116-BC), DTK193 (QFY122 + pDTK118-BAC), DTK197 (QFY122 + pDTK121-BA), and DTK206 (QFY122 + pDTK122-B).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The main conclusions of this study are the following: (1) maximal rates of meiotic recombination hotspot activity stimulated by the binding of transcription factors require both the DNA binding and activation domains of these proteins, and (2) certain DNA sequences appear to stimulate recombination and gene expression without requiring any known bound transcription factors. We suggest that yeast cells have two types of recombination hotspots: hotspots, in which hotspot activity involves an interaction of the recombination machinery with the activation domain of bound transcription factors, and ß hotspots, in which the recombination machinery (without contacting transcription factors) acts on "naked" DNA. Each of these conclusion will be discussed further below.

Transcription-factor-dependent stimulation of recombination activity:
We previously concluded that the activity of the wild-type HIS4 recombination hotspot required the binding of three transcription factors, Rap1p, Bas1p, and Bas2p (WHITE et al. 1991 , WHITE et al. 1993 ). Replacement of the wild-type upstream binding sites with duplicated or triplicated Rap1p-binding sites also resulted in hotspot activity (WHITE et al. 1993 ), suggesting that Rap1p, in the absence of the other transcription factors, could stimulate meiotic recombination. Because (as described in the Introduction) Rap1p has a wide variety of cellular roles unrelated to its function as a transcription factor, we performed deletion analysis and examined fusions of Gal4p and Rap1p to clarify which portions of the protein were required for hotspot activity. This analysis indicates that maximal recombination activity requires both the DNA binding and the transcription activation domains.

Previous studies demonstrate that meiotic hotspot activity at HIS4 is not directly related to the rate of transcription (WHITE et al. 1992 ). Two alternative models to explain the recombination-stimulating effect of transcription factors are the following: (1) chromatin is remodeled by the transcription factor to allow accessibility of DNA sequences to the recombination machinery [which presumably includes Spo11p (KEENEY et al. 1997 ) as well as other proteins], and (2) the activation domain of the bound transcription factor directly contacts the recombination machinery to recruit it to the DNA (as shown in Figure 6A). The second model is similar to the mechanism by which transcription factors are thought to stimulate gene expression (PTASHNE and GANN 1997 ). Although neither of these models can be ruled out by existing data, we favor the second, because the degree of nuclease sensitivity of chromatin (presumably reflecting DNA accessibility) at hotspots does not correlate well with the degree of hotspot activity (WU and LICHTEN 1995 ; FAN and PETES 1996 ). If this model is correct, one interesting issue is whether the recombination and transcription machinery compete for interactions with the transcription factors.

View larger version (18K): [in this window] [in a new window] [Download PPT slide]   Figure 6. - and ß-meiotic-recombination hotspots. (a) -hotspot. A transcription factor (indicated by cross-hatching) with an activation domain (AD) and DNA-binding domain (BD) binds to DNA and interacts with the recombination complex. This interaction targets (and/or tethers) the complex to the DNA, allowing DSB formation. In this diagram, the large white circles indicate nucleosomes bound to coding sequences, and the white rectangles indicate binding sites for the transcription factor. (b) ß-hotspot. A region that is devoid of nucleosomes and is not bound by transcription factors is directly accessible to the recombination complex, allowing local DSB formation.

A number of variants of the second model are possible. The activation domain could contact the recombination machinery directly or indirectly through connector proteins. Once the connection is made, the transcription factor could tether the recombination complex to the DNA until the initiating DSB is made or the connection could simply facilitate the entry of the recombination machinery onto the chromosome by a transient association. Because DSBs occur in a local region of ~50–100 bp near the hotspot, rather than at a single position (DE MASSY et al. 1995 ; LIU et al. 1995 ; XU and KLECKNER 1995 ; XU and PETES 1996 ), a transient association of the transcription factor and the recombination machinery may be more likely. The DSB sites associated with insertion of the triple Rap1p-binding sites upstream of HIS4 occur in a 50-bp region immediately adjacent to, but not within, the Rap1p-binding sites (XU and PETES 1996 ).

Different transcription factors appear to have different abilities to stimulate recombination. HIS4 hotspot activity in strains with duplicated Gal4p-binding sites upstream of HIS4 was approximately the same with or without Gal4p, even when cells were sporulated in galactose-containing medium (Table 3). Although we did not monitor synthesis of Gal4p in meiosis, KOLODKIN et al. 1986 showed meiotic expression of Gal4p-dependent genes when strains were sporulated in galactose-containing medium, suggesting that Gal4p is present under these conditions. The results shown in Table 3, therefore, indicate that Gal4p, in contrast to Rap1p, does not effectively stimulate hotspot activity. We previously found that deletion of Gcn4p, which binds upstream of HIS4, has no effect on HIS4 hotspot activity (WHITE et al. 1992 ). In addition, many intergenic regions in yeast do not represent strong meiotic recombination hotspots (BAUDAT and NICOLAS 1997 ), although most of these regions contain transcription-factor-binding sites.

The results obtained with the Gal4p/Rap1p fusions (Table 4) add several complicating factors. First, in the strain with the plasmid pDTK116-BC (fusion protein with the Gal4p DNA-binding domain and the C terminus of Rap1p lacking the activation domain), we found a significant reduction of recombination compared to the level of exchange found in the absence of Gal4p. Second, the strain with the plasmid pDTK122-B (Gal4p-binding domain alone) had slightly elevated levels of aberrant segregation for his4-lopc without an elevated level of crossovers or DSBs. Although we have no simple explanation for these results, some of these complications may represent interactions between transcription-factor-dependent and transcription-factor-independent hotspot activity (as described below).

The finding that transcription factors stimulate meiotic recombination at the HIS4 locus (as described above and in WHITE et al. 1991 , WHITE et al. 1993 ) is consistent with the finding that most meiotic recombination hotspots in S. cerevisae map between, rather than within, genes (reviewed by LICHTEN and GOLDMAN 1995 ). In S. pombe, activity of the artificial ade6-M26 hotspot also requires binding of the transcription factor Aft1/Pcr1 (KON et al. 1997 ). Although detailed information about meiotic recombination hotspots is limited for higher organisms, a hotspot located within the major histocompatibility complex (MHC) of the mouse contains transcription-factor-binding motifs and has DNase I-hypersensitive sites (SHENKAR et al. 1991 ).

In summary, meiotic recombination hotspots induced by the action of transcription factors may be common in eukaryotes. We term these -recombination hotspots.

Possible transcription-factor-independent stimulation of recombination activity and gene expression at HIS4:
In addition to naturally occurring meiotic recombination hotspots, foreign DNA sequences introduced into yeast cells will occasionally stimulate meiotic recombination. For example, high levels of meiotic recombination are associated with integrated pBR322 sequences (STAPLETON and PETES 1991 ; WU and LICHTEN 1995 ) and with an insertion of tandem BamHI sites (XU and KLECKNER 1995 ). We describe above evidence that duplicated Gal4p-binding sites also have hotspot activity even in the absence of Gal4p. One explanation of this result is that the duplicated Gal4p-binding sites represent an -recombination hotspot for which the transcription factor has not yet been identified. An alternative possibility is that the duplicated binding sites represent a different type of hotspot (ß-hotspot) in which hotspot activity reflects DNA that is accessible to the recombination machinery independent of the binding of a transcription factor (Figure 6B); this DNA may be "naked," that is, unbound by any cellular protein. We favor the second explanation, because it seems more consistent with the observation that insertions of foreign DNA of diverse sequence often have hotspot activity.

Two other issues related to the hypothesis of two types of hotspots should be discussed. First, most recombination hotspots in yeast are located in regions of open chromatin (WU and LICHTEN 1994 ). This result is consistent with the expected location of both - and ß-hotspots. Second, we suggest that naturally occurring hotspots in yeast are likely to be rather than ß. Because the chromosome that initiates recombination is the recipient of information during meiotic exchange (reviewed by PETES et al. 1991 ), if a diploid strain is heterozygous for a hotspot, meiotic gene conversion events will preferentially result in loss of the hotspot. Because -hotspots are transcription-factor-binding sites, it is likely that other selective forces will balance the loss caused by meiotic recombination. Any ß-hotspot that is not associated with a cellular function, however, would be expected to be eventually lost as a consequence of recombination.

There are two modes of stimulating transcription that may be related to our proposal of two types of recombination hotspots. High levels of gene expression in yeast, as in other eukaryotes, usually require the binding of transcription factors to the upstream region (reviewed by STRUHL 1995 ; PTASHNE and GANN 1997 ). IYER and STRUHL 1995 , however, showed that poly(A) and poly(G) sequences inserted upstream of the coding sequences stimulate transcription in yeast in the absence of the binding of known transcription factors. They suggested that an intrinsic structure of these simple DNA sequences increased the accessibility of transcription factors to nearby chromosomal regions, stimulating transcription. This explanation is very similar to our explanation of ß-hotspots. Our finding that the duplicated Gal4p-binding sites stimulate gene expression of HIS4 in strains lacking Gal4p (Figure 3) suggests that DNA sequences allowing transcription-factor-independent gene expression may also function as ß-hotspots. In addition, we have recently found that tandem arrays of the repeat CCGNN, a nucleosome-excluding sequence, stimulate both meiotic recombination and transcription in yeast (D. T. KIRKPATRICK, Y.-H. WANG, M. DOMINSKA, J. D. GRIFFITH, and T. D. PETES, unpublished data).

Minisatellite instability:
Mammalian genomes contain regions of DNA (minisatellites) in which short (20 to several hundred bp) repeats are tandemly arranged (ARMOUR and JEFFREYS 1992 ). These regions are often meiotically unstable and the types of alterations that are observed indicate that the instability reflects meiotic recombination (JEFFREYS et al. 1994 ). One explanation for the high levels of meiotic instability observed for some minisatellites is that they represent either - or ß-hotspots. For example, the minisatellite located near Ha-Ras that binds the NF-B transcription factor (TREPICCHIO and KRONTIRIS 1993 ) could be an hotspot. Previously, WAHLS 1998 suggested that hyper-variable minisatellites may represent preferred sites for the resolution of Holliday junctions, a late step in recombination. We prefer the hypothesis that the minisatellites are hotspots for the initiation of recombination.

	ACKNOWLEDGMENTS

We thank D. Shore (and members of his lab), A. Lustig, S. G. Gasser, M. Johnston, and E. T. Young for yeast strains, plasmids, and/or useful discussions. We thank M. Dominska and D. Terrill for help with tetrad dissections and E. Sia and R. Kokoska for comments on the manuscript. This work was supported by National Institutes of Health grant GM-24110. D.T.K. is a Special Fellow of the Leukemia Society of America.

Manuscript received April 1, 1998; Accepted for publication January 28, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALANI, E., R. PADMORE, and N. KLECKNER, 1990  Anaylsis of wild-type and rad50 mutants of yeast suggests an intimate relationship between meiotic chromosome synapsis and recombination. Cell 61:419-436[Medline].

ARMOUR, J. A. L. and A. J. JEFFREYS, 1992  Biology and applications of human minisatellite loci. Curr. Opin. Genet. Dev. 2:850-856[Medline].

ARNDT, K. T., C. A. STYLES, and G. R. FINK, 1987  Multiple global regulators control HIS4 transcription in yeast. Science 237:874-880[Abstract/Free Full Text].

BAUDAT, F. and A. NICOLAS, 1997  Clustering of meiotic double-strand breaks on yeast chromosome III. Proc. Natl. Acad. Sci. USA 94:5213-5218[Abstract/Free Full Text].

BHAT, P. J. and J. E. HOPPER, 1992  Overproduction of the GAL1 or GAL3 protein causes galactose-independent activation of the GAL4 protein: evidence for a new model of induction for the yeast GAL/MEL regulon. Mol. Cell. Biol. 12:2701-2707[Abstract/Free Full Text].

BUCK, S. W. and D. SHORE, 1995  Action of a RAP1 carboxy-terminal silencing domain reveals an underlying competition between HMR and telomeres in yeast. Genes Dev. 9:370-384[Abstract/Free Full Text].

CHENG, C., H. KACHEROVSKY, K. M. DOMBEK, S. CAMIER, and S. K. THUKRAL et al., 1994  Identification of potential target genes for Adr1p through characterization of essential nucleotides in UAS1. Mol. Cell. Biol. 14:3842-3852[Abstract/Free Full Text].

DE MASSY, B., V. ROCCO, and A. NICOLAS, 1995  The nucleotide mapping of DNA double-strand breaks at the CYS3 initiation site of meiotic recombination in Saccharomyces cerevisae.. EMBO J. 14:4589-4598[Medline].

DETLOFF, P., M. A. WHITE, and T. D. PETES, 1992  Analysis of a gene conversion gradient at the HIS4 locus in Saccharomyces cerevisae.. Genetics 132:113-123[Abstract].

DEVLIN, C., K. TICE-BALDWIN, D. SHORE, and K. T. ARNDT, 1991  RAP1 is required for BAS1/BAS2- and GCN4-dependent transcription of the yeast HIS4 gene. Mol. Cell. Biol. 11:3642-3651[Abstract/Free Full Text].

FAN, Q. and T. D. PETES, 1996  Relationship between nuclease-hypersensitive sites and meiotic recombination hot spot activity at the HIS4 locus of Saccharomyces cerevisiae.. Mol. Cell. Biol. 16:2037-2043[Abstract].

FAN, Q., F. XU, and T. D. PETES, 1995  Meiosis-specific double-strand DNA breaks at the HIS4 recombination hot spot in the yeast Saccharomyces cerevisae: control in cis and trans.. Mol. Cell. Biol. 15:1679-1688[Abstract].

FELDMANN, H., M. AIGLE, G. ALJINOVIC, B. ANDRE, and M. C. BACLET et al., 1994  Complete DNA sequence of yeast chromosome II. EMBO J. 13:5795-5809[Medline].

GILSON, E., M. ROBERGE, R. GIRALDO, D. RHODES, and S. M. GASSER, 1993  Distortion of the DNA double Helix by RAP1 at silencers and multiple telomeric binding sites. J. Mol. Biol. 231:293-310[Medline].

GILSON, E., T. MULLER, J. SOGO, T. LAROCHE, and S. M. GASSER, 1994  RAP1 stimulates single- to double-strand association of yeast telomeric DNA: implications for telomere-telomere interactions. Nucleic Acids Res. 22:5310-5320[Abstract/Free Full Text].

GIRALDO, R. and D. RHODES, 1994  The yeast telomere-binding protein RAP1 binds to and promotes the formation of DNA quadruplexes in telomeric DNA. EMBO J. 13:2411-2420[Medline].

GRIGGS, D. W. and M. JOHNSTON, 1991  Regulated expression of the GAL4 activator gene in yeast provides a sensitive genetic switch for glucose repression. Proc. Natl. Acad. Sci. USA 88:8597-8601[Abstract/Free Full Text].

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

HARDY, C. F. J., L. SUSSEL, and D. SHORE, 1992  A RAP1-interacting protein involved in transcriptional silencing and telomere length regulation. Genes Dev. 6:801-814[Abstract/Free Full Text].

HENRY, Y. A. L., A. CHAMBERS, J. S. H. TSANG, A. J. KINGSMAN, and S. M. KINGSMAN, 1990  Characterization of the DNA binding domain of the yeast RAP1 protein. Nucleic Acids Res. 18:2617-2623[Abstract/Free Full Text].

HERSKOWITZ, I. and R. E. JENSEN, 1991  Putting the HO gene to work: practical uses for mating-type switching. Methods Enzymol. 194:132-146[Medline].

IYER, V. and K. STRUHL, 1995  Poly(dA:dT), a ubiquitous promoter element that stimulates transcription via its intrinsic DNA structure. EMBO J. 14:2570-2579[Medline].

JEFFREYS, A. J., K. TAMAKI, A. MACLEOD, D. G. MONCKTON, and D. L. NEIL et al., 1994  Complex gene conversion events in germline mutation at human minisatellites. Nature Genetics 6:136-145[Medline].

JOHNSTON, S. A., J. M. SALMERON, and S. S. DINCHER, 1987  Interaction of positive and negative regulatory proteins in the galactose regulon of yeast. Cell 50:143-146[Medline].

KEENEY, S., C. N. GIROUX, and N. KLECKNER, 1997  Meiosis-specific DNA double-strand breaks are catalyzed by Spo11, a member of a widely conserved protein family. Cell 88:375-384[Medline].

KLEIN, F., T. LAROCHE, M. E. CARDENAS, J. F.-X. HOFMANN, and D. SCHWEIZER et al., 1992  Localization of RAP1 and topoisomerase II in nuclei and meiotic chromosomes of yeast. J. Cell Biol. 117:935-948[Abstract/Free Full Text].

KOLODKIN, A. L., A. J. S. KLAR, and F. W. STAHL, 1986  Double-strand breaks can initiate meiotic recombination in S. cerevisiae.. Cell 46:733-740[Medline].

KON, N., M. D. KRAWCHUK, B. G. WARREN, G. R. SMITH, and W. P. WAHLS, 1997  Transcription factor Mts1/Mts2 (Aft1/Pcr1, Gad7/Pcr1) activates the M26 meiotic recombination ho