Initiation of recombination in Saccharomyces cerevisiae

Proc. Natl. Acad. Sci. USA Vol. 91, pp. 11929-11933, December 1994 Genetics

Initiation of recombination in Saccharomyces cerevisiae haploid meiosis (double-strand break/hotspot/paring/sister chromatid/yeast)

BERNARD DE MASSY*t, FREDERIC BAUDATt, AND ALAIN NICOLASt Institut de G6ndtique et Microbiologie, Batiment 400, Universit6 Paris-Sud, 91405 Orsay Cedex, France

Communicated by Nancy Kleckner, August 23, 1994 (received for review July 10, 1994)

ABSTRACT In most eukaryotes during prophase I of meiosis, homologous chromosomes pair and recombine by coordinated molecular and cellular processes. To directly test whether or not the early steps of the initiation of recombination depend on the presence of a homologous chromosome, we have examined the formation and processing of DNA double-strand breaks (DSBs, the eariest physical lamark of recombinatlon initiation) in various haplold Saccharomyces cerevisiae strains capable of entering meloss. We find that DSBs occur in haploid meiosis, showing that the presence of a homolog Is not required for DSB formation. DSBs occur at the same positions in haploid and diploid meloses. However, these two types of melosis exhibit subtle differences with respect to the timing of formation and levels of DSBs. In haploid melosis, a dower rate ofDSB formation and a reduction in the frequency of DSB (at one of the three sites analyzed) were observed. These results might indcate that interactions between homologs play a role in the formation of meiotic DSBs. Furthermore, hpid strains exhibit a pronounced delay in the disappearance of meiotic DSBs compared to diploid strains, which suggests that sister chromatid interactions for DSB repair are inhibited in hapioid melosis.

During meiosis, a complex series of events is required to ensure the proper segregation of the chromosomes from a diploid cell to its four haploid progeny (1). This is accomplished by a single round of DNA replication followed by two successive divisions. The first division is reductional and ensures the separation of homologous chromosomes (homologs) to opposite poles, whereas the second division is similar to an equational mitotic division in that sister chromatids segregate. In most organisms, the proper reductional segregation of chromosomes at division I requires the recognition of homologs, synapsis by the formation of the synaptonemal complex (SC), and at least one reciprocal exchange per pair of homologs to provide a physical connection between them (2). How these events are regulated with respect to each other is not yet known. A major question is the mechanism by which homologs find each other and how this is related to other events such as recombination and synapsis (2-5). Recent data in the yeast Saccharomyces cerevisiae suggest that many meiotic recombination events are initiated by localized DNA double-strand breaks (DSBs), which are detected at many locations in the genome (6-10). Detection of these intermediates has permitted the ordering of events at the DNA level temporally with respect to other processes: in particular, DSBs can occur before or during initiation of tripartite SC formation (11) and pairing between homologs takes place before SC formation (12). A model presented by Kleckner et al. (4) proposes that interactions between homologs precede and are required for

the initiation of recombination during meiosis, suggesting a temporal and functional order between the successive steps of homology search, pairing between homologs, DSBs, and synapsis (for review, see ref. 13). To directly test whether or not the initiation of meiotic recombination depends on the presence of a homologous chromosome, we have examined the formation and processing of meiotic DSBs in haploid strains of S. cerevisiae able to enter meiosis. In wild-type diploid meiosis, the DNA fragments diagnostic of DSBs are heterogenous in size, due to variable resection of strands that have their 5' ends at the DSB site (14, 15), and are transient, since they are ultimately converted to mature recombinant molecules (16). In diploid cells homozygous for the radSOS mutation, the DSB ends are neither processed nor repaired (7, 17) and DSB fragments accumulate as discrete species. This facilitates the mapping of break sites and the measurement of the total amount of breakage. The present study, performed in radSOS or RADSO strains at three initiation sites of recombination, demonstrates the occurrence of DSBs in haploid meiosis at the same positions that it occurs in diploid meiosis. Some differences between haploid and diploid meioses in the frequency and timing of DSB formation and in the repair of DSBs were observed. Implications of these results with respect to the interactions between homologous chromosomes and between sister chromatids during meiosis are discussed.

MATERIALS AND METHODS Plasmids. D79 was a gift from D. Shore (Columbia University) and contains the SIR4 gene disrupted at its BamHI site by a Bgl II fragment carrying the LEU2 gene. pMY239 contains an EcoRI-Sal I SIR4 fragment from LSD321 (gift from D. Shore) cloned into the EcoRI and Sal I sites of pRS316 (18). pMY242 was obtained by cloning a 4.1-kbp EcoRI-BamHI Mata fragment from YCpS-Mata (19) into pRS303 (18). Yeast Strains and Sporulation. Strains used in this study are listed in Table 1. ORT463, which contains a disruption of the SIR4 gene (sir4::LEU2), was obtained by transformation of the strain MGD131-2C with a purified 5-kbp Sac I-HindIH fragment from D79 plasmid DNA. The plasmid pMY239 (S1R4, ARSH4, CEN6, URA3) was transformed into ORT463, which could then be mated to introduce the sir4-::LEU2 disruption into other strains by genetic crosses. YSC83AHS225 is a 225-bp Hpa I-SnaBI deletion at the 3' end of YSC83 gene (20, 21). ura3::DupARG4 is an integration of the 12-kbp Apa I-SnaBI fragment containing ARG4 into the Apa I and Stu I sites of the URA3 gene (22). Integration of Mata at the HIS3 locus was achieved by transformation of Abbreviations: DSB, double-strand break; SC, synaptonemal complex. *To whom reprint requests should be addressed. tPresent address: Institut Curie, Section de Biologie, 26 rue d'Ulm, 75231 Paris Cedex 05, France.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 11929

11930

Genetics: de Massy et al. Table 1. Yeast strains Strain Ploidy n MGD131-2C n ORT463 n ORD1743-14D n ORD2113-6D n 0RD1173-13D n 0RD2124-5B n ORT502 n ORT499 ORD307

2n

ORD893

2n

0RD2106

2n

0RD1763

2n

Proc. Natl. Acad. Sci. USA 91 (1994)

Genotype Mata arg4A2060 leu2-3,112 trpl-289 ura3-52 cyhr Mata sir4::LEU2 arg4A2060 leu2-3,112 trpl-289 ura3-52 cyhr Mata sir4::LEU2 polylARG4 YSC83AHS225 radSOSKI8J trpl-289 ura3-52 cyhr ade2 Mata sir4::LEU2 polylarg4RV leu2-3,112 trpl -289 ura3::Duppolylarg4Bg cyhr Mata polylARG4 YSC83AHS225 radSOSKI81 leu2-3,112 his3Al trpl-289 ura3-52 Mata polylarg4RV leu2-3,112 his3Al ura3::Duppolylarg4Bg cyhr Mata his3AJ-Mata-HIS3 polylARG4 YSC83AHS225 radSOSKI81 leu2-3,112 trpl-289 ura3-52 Mata his3l-Mata-HIS3 polylarg4RV leu2-3,112 trpl-289 ura3::Duppolylarg4Bg cyhr Mata arg4RV leu2-3,112 his3Al trpl-289 ura3-52 CYHS ADE2 Mata arg4Bg LEU2 HIS3 trpl-289 ura3-S2 cyhr ade2 Mata polylarg4RV leu2-3,112 his3Al trpl-289 ura3-52 cyhr ADE2 Mata polylarg4Bg radSOSKI81 LEU2 HIS3 trpl-289 ura3-52 cyhr ade2 Mata sir4::LEU2 polylARG4 YSC83AHS225 trpl-289 ura3-52 CYHS ADE2 Mata sir4::LEU2 polylARG4 YSC83AHS225 radS5OSKI81 trpl-289 ura3-52 cyhr ade2 Mata sir4::LEU2 arg4A&2060 leu2-3,112 his3Al trpl-289 ura3::Duparg4Bg cyhr ADE2 ura3-52 cyhr ade2 Mata sir4::LEU2 arg4RV leu2-3,112 HIS3 trpl-289

radSOSKI81

ORD2124-5B or ORD1173-13D strains with pMY242 DNA linearized at the Pst I restriction site. All constructions were verified on Southern blots. For sporulation, cells were grown at 30°C in presporulation medium (23) to 4-5 x 107 cells per ml for diploids and 5-7 x 107 cells per ml for haploids. At these concentrations, the proportion of cells with small buds was 15 ± 4%. The cells were then washed in water and incubated at the same density in sporulation medium (1% potassium acetate supplemented with required bases and amino acids) at 30°C with vigorous shaking. The efficiencies of entry into meiosis were 60 ± 10% in haploid and diploid strains, as determined by 4',6-diamidino-2-phenylindole staining. Detection and Quantification of DSBs. DNA extraction, Southern blot hybridization, and quantification of bands with the SOFT radioimager (Quartz et Silice, Nemours, France) were as described (20). The frequency of DSBs corresponds to the amount of the DSB fragment relative to the total amount of parental plus DSB fragments. From measurements of the DSB frequency in isogenic radSOS and RAD50 strains, the life span (d) of DSBs in RAD50 strains was calculated by the formula: d = [S - (tmax - to)yo1/ymax, where S is the integral of the measurements of the DSB frequencies in RAD50 strains calculated from the first time point before the appearance of meiotic DSB to the time point of disappearance of meiotic DSB for diploids (i.e., from 4 to 14 h) or to the last time point in the experiment for haploids (i.e., from 4 to 24 h). tma,, is the time when the maximal DSB frequency (yn.) is reached in radSOS strains. to, where the DSB frequency is yo, is the first time point before the increase in the DSB frequency. In haploid strains, since the absolute values of the slope of appearance and disappearance of DSBs differ, we conclude that the DSB life span is not uniform within the population. The formula described above gives, therefore, only an estimate of the median value for d in the population of haploid cells. In addition, in haploids, the value of S and, therefore, of d is underestimated since at the last time point in the experiment (i.e., 24 h), the DSB frequency has not decreased to yo.

RESULTS Experimental Systems. In S. cerevisiae, entry into meiosis requires the coexpression of mating types a and a (24). A haploid strain normally expresses only one mating type. Three types of genetic alterations permit haploid cells to enter meiosis: (i) mutations in the sir genes (SIR-SIR4) that result in the derepression of normally silent mating type loci HMR and HML and lead to coexpression of a and a mating types (25); (ii) introduction of the Mata gene in a Mata

radSOSKI81

haploid strain (or vice versa); or (iii) mutation in the RME1 gene, which encodes for a repressor of meiosis (26, 27). We describe the analysis of meiotic recombination in sir2, sir4, rmel haploids and Mata haploid strains containing an ectopic integration of the Mata gene at the HIS3 locus. DSBs were analyzed at three hot spots of meiotic recombination (ARG4 and DED81-DED82 on chromosome VIII and CYS3 on chromosome I) in RAD50 and radSOS strain backgrounds. The features of the ARG4 initiation site and its associated DSB have been studied in detail in diploid cells (6, 14, 20-22, 28-30). The DED81-DED82 site of DSB displays properties similar to the ARG4 site (29). The CYS3 locus is a hot spot of meiotic recombination (31). Meiotic DSBs were detected at two sites in the promoter region of the CYS3 gene (M. Vedel and A.N., unpublished data). Meiotic DSBs at CYS3 in sir4 radSOS Strains. The CYS3 meiotic DSBs formed in a SlR4 raJSOS homozygous diploid are shown in Fig. 1A, lane 10, as control. In this experiment, cells were taken 12 h after transfer in the sporulation medium, and chromosomal DNA was extracted, digested with HindIII, and probed with an EcoRI fragment located at the 3' end of the CYS3 gene (Fig. 1D). The two adjacent CYS3 meiotic DSBs are =150 bp apart. The frequency of the two CYS3 DSBs, which reached its maximal level by 12 h, was 7 ± 2% of total DNA (average value from several experiments). The frequencies of the CYS3-distal DSB (slower migrating species) and the CYS3-proximal DSB were 4.7 ± 1.3% and 2.3 ± 0.7%, respectively, as inferred from the ratio (two) of the peak values of each signal. Time course analyses of the CYS3 DSBs in sir4 rad5OS haploid and sir4 radSOS homozygous diploid are presented in Fig. 1 A and B, respectively. 'In both cases, the two CYS3 meiotic DSBs were detected at positions (±50 bp) similar to those in the SIR4 radSOS diploid, demonstrating that the formation of these meiotic DSBs can occur in the absence of a homolog. Quantification of the DSB frequencies in the sir4 rad5OS haploid and diploid strains is shown in Fig. 1C. In the sir4 rad5OS diploid, DSBs accumulated from 4 to 10 h to reach a maximum frequency of 7%, with the CYS3-distal/CYS3proximal DSB ratio equal to 1.4 to 2.0. The kinetics and the DSB levels were identical to those obtained in an isogenic 51R4 radS5OS diploid (see above and M. Vedel and A.N., unpublished data). Thus, the sir4 mutation has no effect on the formation of the CYS3 DSBs in diploid meiosis. In the sir4 radSOS haploid, the kinetics and the efficiency of DSB formation differed from those observed in the sir4 radSOS diploid: the rate of formation of DSBs was on average 1.5 times slower in the sir4 radS5OS haploid than in the sir4 radSOS diploid, as determined by the slopes of the curves


Genetics: de Massy et al.

8

11931

'C U

-u

m

U)

4

0

A

B

2n rad5OS 12 h

n sir4 rad5OS

2n sir4 rad5OS

0 2 4 6 8 1012 24

aU

h a

2.. I 0* *N-

0M

. . ;;

-

0

5

10

15

25

20

Hours in sporulation medium

D

DSBs *3

a*

DDSBs

DSBs

H

RI

*-

RI H

l_

CYS3 3

_5

6

8

9

probe

1 Kbp

10

FIG. 1. Formation of meiotic DSB at the CYS3 site in haploid and diploid sir4 rad5OS strains. (A) Time course of the CYS3 DSBs in a sir4 rad5OS haploid (ORD1743-14D) (n sir4 radSOS). At various times after transfer to sporulation medium, chromosomal DNA was extracted, digested with HindIII, and analyzed by Southern blot hybridization with the EcoRI-EcoRI probe. Analysis of chromosomal DNA from a rad5OS homozygous diploid (ORD307) 12 h after transfer to sporulation medium is shown as a control (2n rad5OS). (B) Time course of the CYS3 DSBs in a sir4 radSOS homozygous diploid (ORD2106). Digestion and probing of chromosomal DNA were as in A. (C) DSB frequency in 0RD1743-14D (A) (open squares), 0RD2106 (B) (solid squares), and ORT502 (Fig. 3A) (solid diamonds). The two CYS3 DSBs were quantified together. (D) Physical map of the CYS3 region, indicating positions of the HindJII sites (H), the EcoRI-EcoRI (RI) probe, and the CYS3 DSBs.

after normalization for the frequency of DSB at its maximal level; the first increase in the frequency of DSBs was observed at 8 h (compared to 6 h in the sir4 radSOS diploid), therefore, showing a delay of -2 h in the time of appearance of meiotic DSBs in this haploid meiosis; the frequency of the sum of the two CYS3 DSBs reached 3% in the sir4 radSOS haploid compared to 7% in the sir4 radSOS diploid. The ratio of the peak intensities of the two CYS3 DSB signals was 1 0.1 in the sir4 radSOS haploid, with each DSB frequency -1.5%, indicating a significant reduction of the frequency of the CYS3-distal DSB in the sir4 rad5OS haploid (1.5 vs. 4.7 1.3% in the diploid). Meiotic DSBs at CYS3 in sir4 RADSO Strains. To determine whether resection of DSB ends and subsequent repair occur in the absence of homologs, CYS3 DSB fragments were examined in a sir4 RAD50 haploid (Fig. 2A) and compared to a sir4 RAD50 homozygous diploid (Fig. 2B). We found that the pattern of processing and repair of DSBs was identical in the sir4 RAD50 diploid and the SIR4 RAD50 diploid (M. Vedel and A.N., unpublished data), demonstrating that the sir4 mutation has no effect on these events. In both cases, the diagnostic fragments were transient (for the quantitative ±

A

n

0

2

4

sir4 RAD50 2n rad5OS 6 8 10 12 14 24 24 h

analysis, see Fig. 2C), heterogenous in size, and smaller than the unprocessed fragments detected in rad5OS strains. In the sir4 RAD50 haploid, meiosis-specific heterogenous fragments were also detected with a distribution of sizes similar to the intermediates detected in the sir4 RAD50 diploid. We conclude that the resection of CYS3 DSBs occurs similarly in the presence or absence of a homolog. However, the curves of the appearance and disappearance of the CYS3 DSBs differed in the RADSO haploid and diploid strains (Fig. 2C). The slope of appearance of DSBs depends both on the level and rate of DSB formation. Indeed, the difference in these slopes between the sir4 RAD50 diploid and haploid can be explained by the different DSB levels measured in rad5OS derivatives ofthese two strains. More precise measurements in RAD50 strains would be, however, required to detect small variations in the rates of DSB formation. The other parameters of these curves (time of maximum frequency and maximum frequency) are dependent on the kinetics, the levels and the life span of DSBs. DSB life span can be calculated using the integral of the curves and the maximum frequency of DSBs reached in the isogenic rad5OS strains. In the sir4 RAD50 diploid, the life span-of DSBs was 4

0

C

.

2n sir4 RAD50

B 2

4

6

8 10 12 14 24

h

3

*

U) co 0

:

2

06 a

a)

L0 LL

DSBs

-

DSBs

0 0

5

10

1

'0

25

Hours in sporulation rnedium

FIG. 2. Formation of meiotic DSBs at the CYS3 site in haploid and diploid sir4 RAD50 strains. Digestion and probing of chromosomal DNA in Fig. 1A. (A) Time course of CYS3 DSBs in a sir4 RAD50 haploid (ORD2113-6D) (n sir4 RAD50). Chromosomal DNA from a radSOS homozygous diploid (ORD893) 24 h after transfer to sporulation medium is the control (2n rad5OS). (B) Time course of CYS3 DSBs in a sir4 RADSO diploid (ORD1763). Four times more DNA was loaded at the 6-h point compared to the 8-h time point. (C) DSB frequency in 0RD2113-6D (A) (open squares), ORD1763 (B) (solid squares), and ORT499 (Fig. 3B) (solid diamonds).

were as

11932

Genetics: de Massy et A


2.7 h (another experiment with the same strain gave a life span of 1.7 h). In the sir4 RADSO haploid, the median value of the life span was =11 h. This increased life span of DSBs accounts for the delayed time at which the maximum frequency of DSBs was detected in the haploid (12 h) compared to the diploid (8 h). The maximum frequencies of DSBs measured in the sir4 RAD50 haploid (4%) and the sir4 radSOS haploid (3%, see above) were similar. This is precisely what is expected when DSB life span is equal to or greater than the time interval during which DSBs are generated. The increased life span we observed in the sir4 RAD50 haploid could result from a slower repair of DSBs or from the appearance of DSBs at late time points (between 14 and 24 h). We consider this last possibility unlikely, since measurements of DSB frequencies in the isogenic radSOS strains showed no increase after 14 h (see Fig. 1C). This high DSB life span of the sir4 RAD50 haploid showed, therefore, that the repair (if any) of DSBs in haploid meiosis was much slower than in diploid meiosis (either sir4 or SIR4). The decrease in the amount of DSBs at 24 h in the haploid strain could be due to homologous repair, DNA degradation, or other processing events. We conclude that recombination between sister chromatids, the only homologous partners in the haploid cells, is inefficient in sir4 haploid meiosis. Meiotic DSBs at CYS3 in MataHIS3::Mata Haploid Strains. To confirm that the DSB features in haploid meiosis described above were not a specific effect of the sir4 mutation, we studied Mata haploid strains (radSOS or RADSO) carrying an ectopic Mata integration. The formation of the CYS3 DSBs at its normal position in the Mata HIS3::Mata radSOS haploid (Fig. 3A) confirmed the result obtained with the sir4 radSOS haploid and the conclusion that meiotic DSBs at the CYS3 sites could occur in the absence of homologs. The quantification reported in Fig. 1C revealed a 1.5 times slower rate of DSB formation compared to diploid meiosis, in agreement with the results obtained in sir4 radSOS haploid meiosis. No significant delay, relative to diploid meiosis in the time at which the meiotic DSBs appear was observed. The maximal amount of the two adjacent DSBs was 5% (at 12 h) with a ratio of 1 ± 0.1. The amount of the CYS3-distal DSB was, therefore, reduced in this haploid meiosis (2.5% compared to 4.7 ± 1.3% in the diploid), as seen in the sir4 radSOS haploid. The processing and repair of the CYS3 DSBs was analyzed in a Mata HIS3: :Mata RAD50 haploid (Fig. 3B) and quantified (Fig. 2C). As in sir4 haploid meiosis, CYS3 DSB fragments were heterogenous in size and inefficiently repaired, with a median life span of =9 h. Study of Other DSB Sites in Haploid Melosis. To generalize our observations of DSB formation at the CYS3 locus, we then examined meiotic DSBs at the ARG4 (containing the

DISCUSSION The present data demonstrate the formation of meiotic DSBs at three initiation sites of meiotic recombination (CYS3, ARG4, and DED81-DED82) in sir4 and Mata HIS3::Mata haploid meiosis. In parallel experiments done in sir2 and rmel haploid strains, meiotic DSBs at CYS3, ARG4, or DED81-DED82 sites were also detected (data not shown). Meiotic DSBs at the ARG4 site have also been observed in sir3 haploid meiosis (40). We conclude that the presence of a homolog is not required for the occurrence of DSBs, the earliest physical landmark of recombination initiation. The detailed studies of sir4 and Mata His3::Mata haploid strains revealed subtle differences between haploid and diploid meioses in the formation ofDSBs: in haploid meiosis, the rate of DSB formation was slower and the frequency of one of the A

4i

U

2

(t 'r 7-v

Io7.

4

24

ii.....:.: _4 m4

B

m""i;

:.-

._

;

!.

_w

AR Cia

1.

B

A i Mati.

poly] construct; ref. 20) and DED81-DED82 sites in sir4 radSOS (Fig. 4A) and Mata HIS3::Mata rad5OS (Fig. 4B) haploid strains. At both sites, we observed that meiotic DSBs occurred at the same positions as in diploid strains. Quantification was done on the ARG4 DSB, where the maximum frequency was 7 ± 2%, similar to diploid strains (20). The rates of DSB formation were 1.3 (in Mata HIS3: :Mata radSOS haploid) to 1.8 (in sir4 rad5OS haploid) times slower relative to radSOS diploid meiosis. We also examined the features of the ARG4 DSBs in a sir4 RAD50 haploid strain. We observed that as for the CYS3 sites, the DSB fragments were heterogenous in size and inefficiently repaired (data not shown).

rn Mati. HIS3::Mata HIS3..: Mata rad5OS

2

4

6

RA 05C0

;m.

.ll!''.:

4:XOF

01* ~ $ A

"T

Ai g! A I

1;x DSBs

SBs ..

"

8 10 12 14 24 h

it

FIG. 3. Formation of meiotic DSB at the CYS3 site in Mata HJS3::Mata radSOS orRAD50 haploid strains. Digestion and probing of chromosomal DNA were as described in Fig. 1A. (A) Time course of the CYS3 DSBs in a Mata HIS3::Mata radSOS haploid (ORT502). Quantification of DSB frequency is in Fig. 1C. (B) Time course of the CYS3 DSBs in a Mata HJS3::Mata RAD50 haploid (ORT499). Quantification of the DSB frequency is in Fig. 2C.

FIG. 4. Formation of meiotic DSB at the DED82-DED81 and ARG4 sites in sir4 radSOS and Mata HIS3::Mata radSOS haploid strains. DSBs at these sites were detected from the same DNA samples used for the analysis of CYS3 DSBs described in Figs. 1A and 3A. The strains used contain the polylARG4 construct, a more efficient initiation site than the wild-type ARG4 gene (20). (A) Time course of the DED82-DED81 and ARG4 DSBs in a sir4 radSOS haploid (ORD1743-14D). Chromosomal DNA was digested with SnaBI and probed with the ARG4 EcoRV-Bgl II internal fragment. (B) Time course of the DED82-DED81 and ARG4 DSBs in a Mata HIS3::Mata radSOS haploid (ORT502). Digestion and probing of chromosomal DNA were as in A.

Genetics: de Massy

et

al.

two CYS3 DSBs was reduced relative to diploid meiosis. The interpretations of these similarities and differences are presented below. Control of DSB Formation During Meiosis. The features of position, amount, and timing of formation of meiotic DSBs were defined by the analysis of rad5OS haploid or diploid strains. The positions of DSBs at the CYS3, ARG4, and DED81-DED82 sites were found to be identical in haploid and diploid meioses at our resolution level (±50 bp). The frequency of DSBs was quantified at the two most efficient DSB sites (CYS3 and ARG4). At the ARG4 site, these frequencies were similar in haploid and diploid meioses, whereas at the CYS3 site (which includes two adjacent DSB sites), a small reduction was detected specifically at the CYS3-distal DSB (frequencies of 1.5 and 2.5% in haploid meioses vs. 4.7 ± 1.3% in diploid meiosis). The reduction of the frequency of the CYS3-distal DSB in haploid meiosis indicates that all sites are not functionally identical. The rates of DSB formation (both at the CYS3 and ARG4 sites) were 1.3- to 1.8-fold slower in both sir4 and Mata HIS3::Mata haploid meioses relative to diploid meiosis. From the similarities of DSB formation in haploid and diploid meioses, we conclude that the presence of a homolog is not required for the occurrence of meiotic DSBs. The differences between haploid and diploid meioses on DSB formation can be interpreted in several ways. They could reflect a difference at the level of the substrate or of factors acting upon the substrate. For example, a slower and/or less synchronous progression of haploid cells through meiosis and in particular through the stage(s) playing a role in DSB formation could explain the slower rate of DSB formation in haploids. The reduced frequency of DSB at the CYS3-distal site would, however, not be explained by such a mechanism. We therefore favor the view that meiotic DSBs are influenced by the presence or absence of homologs. This conclusion would then support the model in which interactions and pairing between homologs occur before DSBs and play an important role in the control of DSB formation (4, 13). The formation of DSBs in haploid meiosis could reflect the ability of the enzymatic machinery involved in this process to act upon unpaired chromosomes and possibly interacting sister chromatids at this stage, although with a lower efficiency (at the CYS3-distal site). The Behavior of Sister Chromatids: The Repair of DSBs in Haploid Meiosis. An important issue, related to the understanding of the chromosome behavior during meiosis, is the relationship between sister and homologous nonsister chromatids. We found that the repair of DSBs at the CYS3 and ARG4 sites is inefficient in haploid meiosis compared to diploid meiosis. This could be explained if some components of the recombination machinery are defective in meiotic haploid cells. It seems more likely, however, that DSBs are stable in haploid cells because interactions between sister chromatids are constrained. In these cells, the sister chromatid is the only homologous partner available to repair a DSB. Thus, the failure of DSBs to proceed along the recombination pathway could be explained if the interaction with the sister chromatid was blocked. It is known that in diploids, sister chromatids are tightly associated early in meiosis (32) and that crossing-over between sister chromatids occurs at a lower frequency than between homologs (33). Similarly, the recent physical characterization of recombination intermediates corresponding to interhomolog and intersister joint molecule has established that joint molecules form between sister chromatids at =10% of the interhomolog level (34, 35), providing direct evidence for a specific inhibition of sister chromatid interactions in diploid meiosis. An implication of our studies of haploid meiosis is that the specific inhibition of recombination between sister chromatids is independent of


11933

the presence of homologs and is, therefore, an intrinsic property of meiotic sister chromatids. The genetic observation that sister chromatid exchanges are 10 times higher in haploid meiosis than in diploid meiosis has been interpreted to indicate that the inhibition of sister chromatid exchange observed in diploids was due to competition with the homolog (36). In its absence, sister chromatids would then recombine efficiently. However, as discussed by Schwacha and Kleckner (35), it is also possible that the high level of sister chromatid exchanges is a passive consequence of the absence of homologs and that the timing of these exchanges might be delayed compared to those occurring between homologs. In addition, cytological studies have reported SC formation in yeast sir4 haploid meiosis (37) and in plant haploid meiosis (38). Further studies should determine the temporal and functional relationship between SC and the processing of DSBs in yeast haploid meiosis. We thank D. Shore for providing sir4 and sir2 plasmids; A. Mitchell for providing rmel plasmids; M. Costa, M. Poncelet, and G. Stahl for performing some experiments; V. Rocco and D. Zickler for stimulating discussions and comments on the manuscript. We thank S. Le Bilcot and D. Borowski for excellent technical assistance. This work was supported by grants from the Centre National de la Recherche Scientifique (Unite de Recherche Associde 1354), the Universitd Paris-Sud, the Association pour la Recherche sur le Cancer, the Ligue Nationale Contre le Cancer, and the Institut National de la Santd et de la Recherche Mddicale. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39.

John, B. (1990) Meiosis (Cambridge Univ. Press, Cambridge, U.K.). Roeder, G. S. (1990) Trends Genet. 6, 385-389. Loidl, J. (1990) Genome 33, 759-778. Kleckner, N., Padmore, R. & Bishop, D. K. (1991) Cold Spring Harbor Symp. Quant. Biol. 56, 729-743. Atcheson, C. L. & Esposito, R. E. (1993) Curr. Opin. Genet. Dev. 3, 736-744. Sun, H., Treco, D., Schultes, N. P. & Szostak, J. W. (1989) Nature (London) 338, 87-90. Cao, L., Alani, A. & Kleckner, N. (1990) Cell 61, 1089-1101. Nag, D. K. & Petes, T. D. (1993) Mol. Cell. Biol. 13, 2324-2331. Zenvirth, D., Arbel, T., Sherman, A., Goldway, M., Klein, S. & Simchen, G. (1992) EMBO J. 11, 3441-3447. Wu, T.-C. & Lichten, M. (1994) Science 263, 515-518. Padmore, R., Cao, L. & Kleckner, N. (1991) Cell 66, 1239-1256. Schertan, H., Loidl, J., Schuster, T. & Schweizer, D. (1992) Chromosoma 101, 590-595. Hawley, R. S. & Arbel, T. (1993) Cell 72, 301-303. Sun, H., Treco, D. & Szostak, J. W. (1991) Cell 64, 1155-1161. Bishop, D. K., Park, D., Xu, L. & Kleckner, N. (1992) Cell 69, 439-456. Goyon, C. & Lichten, M. (1993) Mol. Cell. Biol. 13, 373-382. Alani, E., Padmore, R. & Kleckner, N. (1990) Cell 61, 419-436. Sikorski, R. S. & Hieter, P. (1989) Genetics 122, 19-27. Heude, M. & Fabre, F. (1993) Genetics 133, 489-498. de Massy, B. & Nicolas, A. (1993) EMBO J. 12, 1459-1466. Rocco, V., Daly, M. J., Matre, V., Lichten, M. & Nicolas, A. (1993) Yeast 9, 1111-1120. Nicolas, A., Treco, D., Schultes, N. P. & Szostak, J. W. (1989) Nature (London) 338, 35-39. Resnick, M. A., Stasiewick, S. & Game, J. C. (1983) Genetics 104, 583-601. Herskowitz, I. (1988) Microbiol. Rev. 52, 536-553. Rine, J. & Herskowitz, I. (1987) Genetics 116, 9-22. Mitchell, A. P. & Herskowitz, I. (1986) Nature (London) 319, 738-742. Covitz, P. A., Herskowitz, I. & Mitchell, A. P. (1991) Genes Dev. 5, 1982-1989. Schultes, N. P. & Szostak, J. W. (1991) Mol. Cell. Biol. 11, 322-328. Schultes, N. P. & Szostak, J. W. (1990) Genetics 126, 813-822. Rocco, V., de Massy, B. & Nicolas, A. (1992) Proc. Nat. Acad. Sci. USA 89, 12068-12072. Cherest, H. & Surdin-Kerjan, Y. (1992) Genetics 130, 51-58. Moens, P. (1987) Meiosis (Academic, Orlando, FL), pp. 1-17. Jackson, J. A. & Fink, G. R. (1985) Genetics 189, 303-332. Collins, I. & Newlon, C. S. (1994) Cell 76, 65-75. Schwacha, A. & Kleckner, N. (1994) Cell 76, 51-63. Wagstaff, J. E., Klapholz, S., Waddell, C. S., Jensen, L. & Esposito, R. E. (1985) Mol. Cell. Biol. 5, 3532-3544. Loidl, J., Nairz, K. & Klein, F. (1991) Chromosoma 100, 221-228. Gillies, C. B. (1984) Crit. Rev. Plant Sci. 2, 81-116. Gilbertson, L. A. & Stahl, F. W. (1994) Proc. Natl. Acad. Sci. USA 91, 11934-11937.