Proper chromosome segregation and formation of viable gametes depend on synapsis and recombination between homologous chromosomes during meiosis. Previous reports have shown that the synaptic structures, the synaptonemal complexes (SCs), do not occur in yeast cells with the SPO11 gene removed. The Spo11 enzyme makes double-strand breaks (DSBs) in the DNA and thereby initiates recombination. The view has thus developed that synapsis in yeast strictly depends on the initiation of recombination. Synapsis in some other species (Drosophila melanogaster and Caenorhabditis elegans) is independent of recombination events, and SCs are found in spo11 mutants. This difference between species led us to reexamine spo11 deletion mutants of yeast. Using antibodies against Zip1, a SC component, we found that a small fraction (1%) of the spo11 null mutant cells can indeed form wild-type-like SCs. We further looked for synapsis in a spo11 mutant strain that accumulates pachytene cells (spo11Δ ndt80Δ), and found that the frequency of cells with apparently complete SC formation was 10%. Other phenotypic criteria, such as spore viability and homologous chromosome juxtaposition measured by FISH labeling of chromosomal markers, agree with several previous reports of the spo11 mutant. Our results demonstrate that although the Spo11-induced DSBs obviously promote synapsis in yeast, the presence of Spo11 is not an absolute requirement for synapsis.
MEIOTIC chromosome pairing and recombination are two processes that are generally necessary for proper segregation of homologous chromosomes and mixing of parental genomes in the first meiotic division. Both processes are closely linked to the appearance of a proteinaceous structure, the synaptonemal complex (SC), which forms between the homologs along their entire length (synapsis; reviewed in Roeder 1997). Although valid for a large majority of eukaryotes, the simple and straightforward view that all the meiotic events (pairing, recombination, and synapsis) are needed to assure proper chromosome segregation is not true for all organisms. There are examples of accurate chromosome division without either synapsis or recombination. In fission yeast, Schizosaccaromyces pombe, both crossing over and correct chromosome segregation can be achieved without SCs. In Bombyx mori females, a modified SC can act as glue between the homologs and thus assure proper chromosome segregation in the absence of crossing over (Rasmussen 1976). It is even possible to accomplish proper segregation without either synapsis or recombination: in Drosophila, homologous chromosomes that have not recombined can be paired and segregated by a mechanism called distributive segregation, in which heterochromatic pairing plays an important role (reviewed by Walker and Hawley 2000).
In organisms that require pairing, recombination, and synapsis for normal chromosome segregation there are examples of mutants with obstructed, but not blocked, meiotic progress. In a mutant strain of budding yeast that lacks a major component of the SC, Zip1, some recombination still occurs, showing that the SC is not absolutely required for recombination in yeast (Sym et al. 1993; Storlazzi et al. 1996). The spore viability in the zip1 null mutants is 60%, showing that the chromosome segregation is reasonably good (Tung and Roeder 1998). While it seems as if the SC is not required for recombination, it has been generally believed that the initial step of recombination, the formation of double-strand breaks (DSBs) in the DNA, is an absolute prerequisite for successful synapsis in yeast (reviewed in Keeney 2001). DSBs are generated by a topoisomerase-related meiosis-specific enzyme called Spo11 (Sun et al. 1989; Cao et al. 1990; Bergerat et al. 1997; Keeney et al. 1997). Together with Spo11, at least 10 other genes are needed for initiation of recombination by DSB formation in yeast (reviewed in Keeney 2001). Studies show that the sites on the chromosomes where recombination is initiated by DSBs are also the sites where synapsis starts. Several recombination enzymes colocalize with the first synaptic protein, Zip3, which recruits the proteins Zip2 and Zip1 that complete synapsis (Agarwal and Roeder 2000). The amount of Spo11-induced DSBs in different spo11 mutants has been correlated with the level of SC formation, showing that initiation of synapsis is indeed induced by DSBs (Henderson and Keeney 2004). In this study it was also shown that the number of Zip3 sites decreases if the frequency of DSBs decrease.
Both yeast and mouse spo11 null mutants are defective in DSB formation as well as synapsis (Giroux et al. 1989; Weiner and Kleckner 1994; Mahadevaiah et al. 2001). There are, however, observations of SC-like structures in both mouse and yeast spo11 mutants. In yeast, traces of SCs have been reported in spo11 null mutants, and there is an observation of complete SC formation in a yeast strain with a spo11 point mutation (spo11-1; Klapholz et al. 1985; Loidl et al. 1994; Malkova et al. 2000). In contrast to the yeast and mouse phenotype, neither the Spo11 protein nor DSBs are required for synapsis in Drosophila melanogaster and Caenorhabditis elegans (Dernburg et al. 1998; McKim et al. 1998). Thus the formation of DSBs by Spo11 promotes synapsis in, e.g., yeast and mouse, while this system does not seem to be necessary in some other organisms.
The differences in the synapsis and initiation of recombination processes between species, and the uncertainty about the yeast phenotype, led us to reinvestigate the yeast spo11 mutant. We have examined spo11 mutants of Saccharomyces cerevisiae using immunofluorescence with the aim of gaining a better understanding of the relationship between the initiation of meiotic recombination and synapsis in yeast. We have found complete and homologous synapsis (SC formation) in a small fraction of the cells (1%). In a spo11 mutant that arrests in pachytene (spo11Δndt80Δ), we found a 10-fold increase of cells with full SC complement compared to the single mutant. Other aspects of the spo11 mutant phenotype fit with data previously published. Spore viability is close to zero. A moderate but notable level of colocalization of both subtelomeric and interstitial parts of homologous chromosomes occurs in the spo11 mutant cells, which is in agreement with several previously published reports (Loidl et al. 1994; Weiner and Kleckner 1994; Peoples et al. 2002). Although initiation of recombination via Spo11 normally precedes synapsis and clearly strongly catalyzes SC formation, the present result demonstrates that chromosome synapsis in yeast can occur in the absence of Spo11.
MATERIALS AND METHODS
The yeast (S. cerevisiae) strains used in this study are as follows: NKY2967 (a/α ho::LYS2/ho::LYS2 lys2/lys2 ura3/ura3 leu2::hisG/leu2::hisG his4X::LEU2/his4B::LEU2 spo11Δ::hisG-URA3-hisG/spo11Δ::hisG-URA3-hisG), NKY1551 (a/α ho::LYS2/ho::LYS2 lys2/lys2 ura3/ura3 leu2::hisG/leu2::hisG arg4-Nsp/arg4-Bgl his4X::LEU2(Bam)-URA3/his4B::LEU2), NKY2310 (a/α ho::LYS2/ho::LYS2 lys2/lys2 ura3/ura3 leu2::hisG/leu2::hisG ndt80Δ::LEU2/ndt80Δ::LEU2 his4X::LEU2-MluI-URA3/his4B::LEU2-MluI), NKY 2535 (a/α ho::LYS2/ho::LYS2 lys2/lys2 leu2::hisG/leu2::hisG ura3/ura3 spo11Δ::hisG-URA3-hisG/spo11Δ::hisG-URA3-hisG ndt80Δ::LEU2/ndt80Δ::LEU2), and S2888 (a/α his4-280/ his4-260 leu2-27/leu2-3,112 arg4-8/+ thr1-1/thr1-4 cyh10/+ ade2-1/ade2-1 ura3-1/ura3-1 trp1-1/trp1-289 spo11-ADE2/spo11-ADE2).
The cells were cultured as described in Schmekel (2000). Briefly, yeast strains were grown to saturation in YPAD medium and then diluted 1:200 in presporulation medium (2% Bacto-peptone, 1% Bacto-yeast extract, 1% potassium acetate, and 0.003% adenine sulfate) and allowed to grow to a density of 1.5–2.0 × 107 cells/ml. The cells were then transferred to sporulation medium with a density of 3.0–4.0 × 107 cells/ml. Unless otherwise mentioned, the cells were grown at 30° and harvested after 5 or 8 hr.
Immunofluorescence and fluorescence in situ hybridization:
Immunofluorescence experiments were carried out with spread preparations on glass slides (Loidl et al. 1998) as described in Bhuiyan et al. (2003). The Red1 antibody, which was a kind gift from Shirleen Roeder's laboratory, was used in dilution 1:1000 (Smith and Roeder 1997). For antibody signal strength comparison, the Zip1 antiserum published in Sym and Roeder (1995) was used. Fluorescence in situ hybridization (FISH) was carried out according to Weiner and Kleckner (1994) with minor modifications (described in Bhuiyan et al. 2003), using one probe that binds to an ∼35-kb-long subtelomeric region of chromosome III (c9171; American Tissue Culture Collection (ATCC) no. 70884) and another probe that binds to an ∼26-kb-long interstitial part of chromosome VIII (c9315; ATCC no. 71216). Images from fluorescence and confocal microscopy were processed using Adobe Photoshop.
Both wild-type (NKY1551) and spo11 mutant (NKY2967) strains were sporulated for 3 days. Tetrads were dissected using a Singer MSM System Series 200 dissecting apparatus and spores were allowed to grow on YPAD plates.
A fraction of spo11 mutant cells exhibit wild-type-like SCs:
To evaluate the level of SC formation, we performed immunofluorescence experiments on spread nuclear preparations of a spo11 deletion mutant (NKY2967) and on the wild-type strain (NKY1551). SC structures were identified using a polyclonal antibody that recognizes the Zip1 protein (the antiserum is described in Bhuiyan et al. 2003). We classified the different appearances of SC staining into four classes, A–D. The SC staining pattern that fully resembled the wild-type SC pattern (Figure 1a) is class A. We believe that the SCs in these cells correspond to homologously synapsed chromosomes. Each cell contained ∼16 SCs, we could usually discern most of the individual bivalents, and we rarely saw structures that could correspond to forked SCs. The class A Zip1-staining pattern was found in 1% of the spo11 mutants harvested after 5 hr in sporulation medium, which is the time when the highest portion of wild-type cells are in pachytene (see Table 1 and Figure 2, a–c). Many of the mutant nuclei appeared with what we called an “SC-like” staining pattern (class B; Figure 1b). The Zip1-staining structures in these nuclei were short, often faint, and discontinuous. In addition, a large fraction of the spo11 mutant nuclei showed a more diffuse staining (class C, Figure 1c). Nuclei that lacked anti-Zip1 staining were placed into class D. We saw the same patterns of Zip1 staining in the wild-type nuclei, although in this case the frequencies were different (Table 1).
Yeast cells that lack the SPO11 gene are not affected by meiotic checkpoint arrest or delay, and they sporulate after incomplete meiosis. Because the cell populations are not in perfect synchrony, only a fraction of the cells are in fact in pachytene at the “pachyetene time point.” We realized that the number of SC-forming cells in the spo11 mutant for this reason may have been underestimated and wanted to investigate how many cells were able to create full synapsis totally. We therefore examined the level of SC formation in the double mutant spo11Δndt80Δ, in which pachytene cells accumulate. Ndt80 is a transcription factor needed for meiotic cell-cycle progress, and ndt80 mutants arrest in pachytene (Xu et al. 1995). Cells were collected after 5 and 8 hr in sporulation medium. We found that 10% of the cells in the double-mutant strain formed wild-type SCs (class A) after 5 hr in sporulation medium and 11% after 8 hr (Figure 2, d–f). This is a 10-fold increase compared to the single spo11 mutant and is one-sixth of the SC content in ndt80Δ cells in which ∼66% of the cells in these experiments form class A SCs. With time, class B cells increased more than class A cells in frequency, while the class C cells decreased in relative numbers. A summary of the SC frequency data in spo11Δ, spo11Δndt80Δ, ndt80Δ, and wild type is given in Table 2.
The occurrence of SCs in the spo11 mutant cells was also investigated using an antibody against another SC protein, Red1, which is located in the lateral elements of the SC. The immunolabeling of the SCs in wild-type cells with the Red1 antibody appeared similar to the labeling with the Zip1 antibody, like a group of ribbons. However, since the Red1 antibody stains both axial elements and synapsed chromosomes, and the Red1 staining often is more discontinuous than the Zip1 staining, we could not perform the classification analysis that was done using the Zip1 antibody. A fraction of the Red1-stained nuclei showed more continuous SCs. Complete synapsis is more reliably confirmed if the SC labeling is continuous, and this class was used for investigation of the presence of Red1 staining SCs in the spo11 mutant cells. The Red1 antibody stained a fraction of the spo11ndt80 double-mutant cells in the characteristic wild-type-like SC pattern (Figure 3, a vs. b), a result that supports the finding of SCs in spo11 mutant cells.
The literature concerning the phenotype of spo11 mutants in yeast is not consistent as to whether any SCs are formed. We were concerned that our result would not be due to SC formation that had been artificially introduced into the cells. Exogenously induced DSBs may give rise to such artifactual SCs. We have taken extensive measures to detect any such effect. For example, we have changed all components in the growth media and cell preparation solutions, including the water source, to other brands; we have cultivated the cells on premises other than our own laboratory; we have grown cells in the dark; we have received new cells and restarted the cell strains; and we have measured the radiation in the laboratory. The SCs were present under all of these conditions.
SCs have also been observed in another spo11Δ strain (S2888 of BR background; data not shown) using the Zip1 antibody. The yeast strains used in this study are all widely used, well defined, and described in materials and methods (for source see Acknowledgment).
Several other aspects of the spo11 mutant phenotype are in agreement with those previously described:
The phenotype of spo11Δ includes aneuploidy and inviable spores, due to missegregation of homologs. The number of viable spo11Δ spores would probably increase if DSBs were introduced. We therefore performed spore viability tests to determine the phenotype of the spo11 mutant cells under the conditions that we use. Tetrads of wild type and spo11Δ were dissected and allowed to grow on YPAD. The spore viability of the spo11Δ cells was 0.2% (n = 440) and that of the wild-type cells was 96.0% (n = 360).
We performed FISH experiments using chromosome-specific DNA probes on cells that were surface spread to determine the extent to which homologous chromosomes had been brought into juxtaposition (pairing). One probe, derived from a cosmid clone (ATCC no. 70884), recognizes one end of chromosome III and another probe, ATCC no. 71216, recognizes an interstitial part of chromosome VIII. Wild-type cells were used as positive controls. Unpaired homologs appear as two separate dots, while two homologs that are paired appear very close together and often fuse into one dot. Homologs were classified as paired if the two FISH signals were fused (or separated from each other by <0.7 μm; Weiner and Kleckner 1994). Table 3 shows that the subtelomeric probe was highly paired in both spo11 mutant strains independent of time of harvesting. The pairing was almost as high in the mutants as in the wild-type cells. The number of fused signals of the interstitially situated probe was lower than that of the subtelomeric probe (26% in spo11Δ and 34% in spo11Δndt80Δ), but considerably higher than what can be expected if no homologs are paired. A summary of the chromosome pairing data is given in Table 3.
SC formation depends on physical growth conditions:
The difference between our results showing that the SCs form in spo11 mutant cells and those in previous reports may arise from technical differences in the preparation of the cells. It is possible that slight differences in growth conditions give rise to differences in SC occurrence. We have therefore altered the growth conditions to determine the effect of these conditions on SC occurrence. We cultivated spo11Δndt80Δ and wild-type cells under different conditions (26°, 30°, and 34°, in five times higher and five times lower cell concentration than normal, and in two different incubators), harvested the cells at the time for normal pachytene, and recorded differences in SC frequency and sporulation frequency (Table 4). Almost all variations from the standard conditions gave considerably lower SC frequency (in the order of 10–50% of normal SC frequency). Low cell concentration had the greatest effect on the SC count. The low number of cells with SCs in some cases was due to an obvious inability to develop meiotic cells (which resulted in low sporulation frequency) while in other cases it was due to delayed or accelerated meiotic development (resulting in normal sporulation frequency). Surprisingly, two different shaking incubators gave a significantly different SC count. Growth conditions thus clearly influence the occurrence of pachytene cells at a certain time point and sporulation. Notably, these experiments were performed using the double mutant that under optimized conditions has SCs in 10% of the cells. Although the SCs were never totally abolished in these preparations, the same degree of reduction in SC count could be expected in the single spo11 mutant, which would result in ∼0.1–0.5% cells with SCs. This level could possibly be considered “nondetectable.”
SCs can form in spo11 null mutants:
The Spo11 protein forms DSBs that act as initiation sites for recombination. It seems as if formation of DSBs by Spo11 promotes pairing as well as synapsis (e.g., yeast, Keeney et al. 1997; and mouse, Mahadevaiah et al. 2001) if homologous chromosomes are not brought together by other mechanisms, like prealignment of chromosomes as in Drosophila (e.g., Fung et al. 1998) or meiotic pairing centers as in C. elegans (Villeneuve 1994). The fact that homologous chromosomes may be synapsed in the absence of Spo11 in some organisms suggests that DSBs do not take active part in the mechanism of synapsis. It is not clear whether any isolated recombination step is directly involved in initiation of synapsis in yeast. Nevertheless, in yeast the chromosomes are synapsed temporally just after the DSB formation (Padmore et al. 1991). Early as well as late recombination enzymes in yeast, like Rad51, Dmc1, and Msh4, may be removed with the chromosomes still capable of synapsis, although the process may be delayed (Rockmill et al. 1995; Novak et al. 2001). The removal of yeast Spo11 or any of at least four other proteins that are necessary for DSB formation, on the other hand, has been reported to completely block SC formation (reviewed by Burgess 2002). Although observations of yeast clearly show that SC formation is strongly promoted by Spo11-induced DSBs, spo11 null mutants occasionally form SC segments (Loidl et al. 1994) or SC-like structures (0.2–1% of recorded cells; Malkova et al. 2000). Furthermore, Klapholz et al. (1985) found wild-type-like SC structures in the nuclei of a spo11 mutant (spo11-1) by electron microscopy. However, spo11-1 is a point mutation and it is unclear how strongly Spo11 function is affected.
Contrary to previously published results, we have found that a small fraction of pachytene spo11 null mutant cells have fully synapsed chromosomes. Thus, even if the Spo11 DSBs are strong promoters of synapsis, and synapsis does not occur efficiently in the absence of DSBs, our result demonstrates that synapsis in yeast may occur without Spo11 and probably without DSBs. We were concerned with the fact that we reproducibly found small fractions of wild-type-like SCs while other investigators have not detected SCs in spo11 mutant cells. One possible artifact could be exogenously introduced DSBs. These DSBs would induce recombination and possibly SC formation and thus lead to a higher frequency of cells that undergo proper segregation, resulting in an elevated frequency of spore viability. In a study in which Δspo11 cells were moderately X-ray irradiated (15 krad), spore viability increased 40 times, which was 3% of the nonirradiated wild-type level (Thorne and Byers 1993). We expected the spore viability level to be at least somewhat elevated compared to previously reported levels if we had induced DSBs in our samples. Our spore viability tests, however, showed lower viability than previously reported (0.2% vs. 1%; Klapholz and Esposito 1982; Weiner and Kleckner 1994). We thus believe that the SC formation that we have found is not due to exogenously introduced DSBs.
spo11 yeast mutants go through meiosis but produce aneuploid spores. Because of poor cell synchrony and transient synapsis, only a fraction of the SCs that are totally produced during meiosis are present in a cell preparation at a given time point. To get an estimate of the total number of spo11 mutant cells that are able to synapse their homologous chromosomes, we used a spo11 mutant that arrests in pachytene (spo11Δndt80Δ). We found a 10-fold increase of SC-forming cells in this strain compared to the single spo11 mutant strain. Surprisingly, we found that the double mutant showed almost the same number of SC-forming cells after 5 hr as after 8 hr in sporulation medium (10% vs. 11%). Probably this is due to the “pachytene peak” occurring somewhat earlier than at 5 hr in these experiments and to the fact that the double mutant already has accumulated most pachytene cells before 5 hr. Since the spo11 single mutants only transiently synapse, and thus have only a small number of SC-forming cells at a certain time (1%), it seems reasonable that the total number of cells that have synapsed during meiosis is the same in the single spo11 mutant and in the double-mutant populations (10%).
spo11 mutants show some homologous pairing, but results vary:
We, as well as several other investigators, have measured the degree of juxtaposition of homologous chromosome loci in yeast spo11 mutants by FISH (here called pairing; Loidl et al. 1994; Weiner and Kleckner 1994; Cha et al. 2000; Neale et al. 2002; Peoples et al. 2002). The absolute level of pairing varies between the investigations. Cha et al. (2000) and Neale et al. (2002) did not find any meiotic pairing in spo11 deletion mutants, whereas the other studies show moderate capacity of spo11 mutant cells to bring homologs together. Subtelomeric markers seem to show significantly higher pairing level [using the same probe, Weiner and Kleckner (1994) find 42% pairing while the present study shows 47% pairing in spo11 mutants]. Using interstitial probes, different reports show different levels of pairing in spo11 yeast mutants, probably depending on the location of the probe: 4, 12, and 14% (Weiner and Kleckner 1994); 9% (Loidl et al. 1994); 26% (this report); and 41–46% in a double-mutant spo11Δndt80Δ (Peoples et al. 2002). Interestingly, Loidl and coworkers found that the level of pairing rose from 9% in all spo11 mutant cells to 26% if only pachytene cells were counted. Together these reports indicate that meiotic spo11 mutant cells to some extent are able to bring homologs into juxtaposition. We do not believe that the pairing that we see is remnant Spo11 independent premeiotic pairing because in the strain background that we use (SK1) disruption of premeiotic pairing occurs after 2.5 hr in sporulation medium (Cha et al. 2000), and we harvest our cells well after this time, after 5 hr in sporulation medium. The premeiotic pairing explanation seems especially unlikely since we do not see any difference in pairing level in the double-mutant cells harvested after 5 hr and those harvested after 8 hr in sporulation medium. The bottom line is that the pairing phenotype of the spo11 mutant shown in this report does not significantly differ from previously published reports.
It is also interesting that Peoples et al. (2002) find pairing levels that are 4 to 10 times higher than what has been reported for the single-mutant spo11Δ (Loidl et al. 1994; Weiner and Kleckner 1994; this report) when they perform FISH pairing measurements on the double-mutant spo11Δndt80Δ. Thus, the pairing pattern of the spo11Δndt80Δ vs. the spo11Δ cells are in accordance with the SC-forming cell pattern that we see in our experiments; prolonged pachytene, which is achieved by cell-cycle arrest in the ndt80 mutant cells, leads to an elevated incidence of both pairing and synapsis in spo11 mutants in the absence of Spo11. It should, however, be noted that in the study of Neale et al. (2002), where a spo11ndt80 double-mutant strain was used, no SCs were found.
Variations in preparation may lead to different results:
The difference between our results that show SC formation in spo11 mutant cells and other reports where SCs were not detected may arise from technical differences in the preparation of the cells. Although the different meiosis laboratories use similar methods, it is possible that slight differences in growth conditions could give rise to differences in SC occurrence or detection level. Börner et al. (2004) show that the ability of several recombination mutants to carry through both synapsis and recombination steps is temperature dependent. In our experiments, altered temperatures, cell concentrations, and change of incubator resulted in lower SC counts. In these experiments the double-mutant spo11Δ ndt80Δ, which under normal conditions contains a relatively large amount of SC-containing cells (10%), was used. So, although SCs were present to some extent in preparations from all conditions, the interesting observation is that almost all alterations resulted in only 10–50% of normal SC count. The same level of reduction can be expected in spo11 single-mutant cells. In these cells we saw 1% SC-containing cells under normal growth conditions, and reduction in the same order may be considered undetected. Another reason for the contradictory result may be the quality of the antibody. Different Zip1 antisera give signals of different strength. In a comparison with another Zip1 serum, we found our antibody to give a particularly strong signal (data not shown). The differences in SC result in the spo11 mutant may also be due to differences in nuclear spreading and detection techniques. We always screened at least five slides, all with cells spread using different quantities of the paraformaldehyde and the Lipsol solutions (see materials and methods), and all samples were screened using the Zip1 antibody, which we have found more reliable than silver staining.
DSB-independent pairing may aid synapsis:
Several studies show involvement of telomeres in the pairing of homologs during meiosis. Prior to pairing of the entire chromosomes, telomeres often cluster in one area on the nuclear envelope (bouquet), a process that in yeast is mediated by the telomere-binding protein Ndj1 (Trelles-Sticken et al. 2000). Rockmill and Roeder (1998) show in a genetic study that telomeres as well as ends of chromosomes interact to promote homolog pairing. Further, it has recently been shown that telomerase-deficient mice fail in synapsis and recombination (Liu et al. 2004). It is interesting to note that spo11 mutants show a considerably higher level of homologous pairing in telomeres than in interstitial sites of chromosomes monitored by FISH (Weiner and Kleckner 1994; this study). It thus seems as if homologous pairing of telomeres is much less dependent on DSB formation than are the interstitial parts of the chromosomes. It is possible that the telomere pairing facilitates the SC formation that we detect in our spo11 preparations.
Weiner and Kleckner (1994) show that homologous chromosomes pair before meiotic S-phase, but disassociate during replication and then reassociate as a first step in the meiotic pairing process. Peoples et al. (2002) could distinguish stable meiotic from unstable premeiotic pairing by the detection of “close, stable homolog juxtaposition” in a Cre/loxP-based recombination assay, which requires close chromosome contact. Several recombination mutant strains that were investigated for the Cre/loxP interaction showed much lower than wild-type levels of contact between the homologs during meiosis. Not surprisingly, the spo11 null mutant strain showed the lowest recombination frequency in the assay. Interestingly, the Cre/loxP interaction, however, was not abolished, but was found to be at a significantly higher level than the ectopic recombination. Peoples et al. suggest that the early, DSB-independent homolog interactions are not sufficient to establish connections that are stable enough to perform recombination, while meiotic pairing that involves DSBs bring the homologs close together. They further suggest that both DSB-dependent and DSB-independent pairing mechanisms may be involved in meiotic pairing. If DSB-independent pairing also takes part in linking the homologous chromosomes together in normal pachytene, these forces should also be present if the DSB-dependent pairing is absent, as in spo11 mutants. The low but significant FISH pairing level seen in spo11 mutants, and the notable level of close contact shown by Peoples et al. (2002), may be remnants of the second round of DSB-independent pairing that has not been disrupted by chromatin rearrangement or spreading force. The remnant DSB-independent pairing could contribute to the formation of the SCs that we found in this study.
Similar to the distributive segregation described in Drosophila, Guacci and Kaback (1991) found evidence for an alternative segregation pathway in yeast. In their experiments nonhomologous chromosomes or chromosome fragments segregate from each other in 90% of the cells. It would be tempting to regard the incidents described above as parts in a salvage pathway to rescue a cell that has been unsuccessful in recombination. Telomeric pairing that normally occurs, DSB-independent pairing that obviously allows some close connection also in spo11 mutants, and DSB-independent synapsis that we show in this study, as well as distributive segregation, could all possibly contribute to an alternative way of proper chromosome segregation. However, although we find fully formed SCs in a fraction of the spo11 mutant cells, we also find that the meiotic products are inviable. Thus the hypothetical salvage pathway is unable to rescue all 16 nonrecombinant chromosome pairs from nondisjunction.
Our results show that even if the majority of meiotic yeast cells fail in chromosome synapsis in the absence of Spo11, a fraction of the cells manage to synapse in a wild-type-like manner. So, even if the Spo11-induced DSBs are strong promoters of synapsis in yeast and the absence of Spo11 strongly reduces synapsis, it seems as if the presence of Spo11 is not an absolute requirement for synapsis.
The authors thank the laboratories of Nancy Kleckner for yeast strains and Shirleen Roeder for antibodies and yeast strains. We are grateful to Gunilla Dahlfors and George Farrants for critical reading of the manuscript. This study was performed with financial support from the Swedish Cancer Society, the Carl Trygger Foundation, and the Magn. Bergvall Foundation.
Communicating editor: R. S. Hawley
- Received April 5, 2004.
- Accepted June 25, 2004.
- Genetics Society of America