In budding yeast, at least 10 proteins are required for formation of the double-strand breaks (DSBs) that initiate meiotic recombination. Spo11 is the enzyme responsible for cleaving DNA and is found in a complex that also contains Ski8, Rec102, and Rec104. The Mre11/Rad50/Xrs2 complex is required for both DSB formation and DSB processing. In this article we investigate the functions of the remaining three proteins—Mer2, Mei4, and Rec114—with particular emphasis on Mer2. The Mer2 protein is present in vegetative cells, but it increases in abundance and becomes phosphorylated specifically during meiotic prophase. Mer2 localizes to distinct foci on meiotic chromosomes, with foci maximally abundant prior to the formation of synaptonemal complex. If DSB formation is blocked (e.g., by a spo11 mutation), dephosphorylation of Mer2 and its dissociation from chromosomes are delayed. We have also found that the Mei4 and Rec114 proteins localize to foci on chromosomes and these foci partially colocalize with each other and with Mer2. Furthermore, the three proteins co-immunoprecipitate. Mer2 does not show significant colocalization with Mre11 or Rec102 and Mer2 does not co-immunoprecipitate with Rec102. We propose that Mer2, Mei4, and Rec114 form a distinct complex required for DSB formation.
IN most sexually reproducing organisms, meiotic recombination plays a critical role in promoting the accurate segregation of chromosomes at the first meiotic division. In the absence of crossing over, homologous chromosomes move independently at meiosis I, resulting in gamete inviability due to aneuploidy.
Meiotic recombination events initiate with double-strand breaks (DSBs) catalyzed by the Spo11 protein, a homolog of archaeal topoisomerase VI (Keeney 2001). However, Spo11 does not act alone. In budding yeast, at least nine other gene products are absolutely required for meiotic DSB formation: Rec102, Rec104, Rec114, Mei4, Mer2, Rad50, Mre11, Xrs2, and Ski8 (Keeney 2001). Hereafter, these gene products will be referred to as DSB proteins.
Mutations eliminating any of the DSB proteins confer the same phenotype in meiotic cells (Keeney 2001). Due to the defect in DSB formation, there is no meiotic recombination; chromosomes therefore undergo wholesale nondisjunction resulting in the production of inviable spores. In addition, stable homolog pairing is significantly reduced, and there is a complete failure of synaptonemal complex formation.
Of the 10 DSB proteins, only Spo11 has a defined biochemical function in DSB formation (Keeney 2001). Several observations suggest that Spo11 acts together with Ski8, Rec102, and Rec104, with all four proteins participating in the assembly of a higher-order complex. Synthetic interactions between non-null alleles and multicopy suppression of conditional mutants connect Spo11 with Rec102 and Rec104 (Rieger 1999; Salem et al. 1999; Kee and Keeney 2002). In addition, the three proteins co-immunoprecipitate (Kee and Keeney 2002; Jiao et al. 2003; Kee et al. 2004). Spo11 and Ski8 are mutually dependent for their association with chromatin and both are required for chromosomal localization of Rec102 and Rec104 (Arora et al. 2004). Arora et al. (2004) have proposed that Ski8 promotes an interaction between Spo11 and a Rec102/Rec104 complex.
The budding yeast Rad50, Mre11, and Xrs2 proteins form a well-characterized complex that is required both for the formation of meiotic DSBs and for the subsequent processing of these breaks to expose single-stranded tails (Haber 1998). Although the complex has double-strand exonuclease and single-strand endonuclease activities that may be involved in DSB processing, the role of the complex in DSB formation remains elusive. The Rad50/Mre11/Xrs2 complex also plays important roles in vegetative cells, including telomere maintenance, the cell cycle response to DNA damage, and DSB repair by both homologous recombination and nonhomologous end joining (Keeney 2001).
Even less is known about the functions of the Mer2, Mei4, and Rec114 proteins. MER2 gene expression is regulated in an unusual manner. The gene is transcribed in both vegetative and meiotic cells, but the transcript is spliced efficiently only in meiosis because splicing is largely dependent on the meiosis-specific Mer1 protein (Engebrecht et al. 1991). Here, we have examined the accumulation, phosphorylation, and chromosomal localization of the Mer2 protein throughout meiotic prophase. We have also investigated the interaction of Mer2 with Mei4 and Rec114 and with components of the Spo11/Ski8/Rec102/Rec104 and Mre11/Rad50/Xrs2 complexes. Our results suggest that Mer2 forms a complex with Mei4 and Rec114, and this complex is distinct from the complexes characterized previously.
MATERIALS AND METHODS
pJ88 encoding MBP-Mer2 was constructed by M. Odell and B. Rockmill (Yale University, New Haven, CT) as follows. The full-length MER2 coding region was amplified such that an EcoRI site was introduced just before the start codon and an SalI site was introduced directly after the stop codon. The resulting EcoRI–SalI fragment was inserted at the EcoRI–SalI sites of the expression vector pMAL-2 (New England Biolabs) such that MER2 is downstream of the sequences encoding the maltose-binding protein (MBP) and fused in frame.
Yeast manipulations were performed and media were prepared as described by Sherman et al. (1986). Gene disruptions were confirmed by PCR analysis. All strains used are homozygous diploids unless otherwise indicated. Both haploid parents of all diploids are isogenic with BR1919-8B (Rockmill and Roeder 1990). Diploids are heterozygous at MAT (MATa/MATα) and homozygous for his4-260, leu2-3,112, thr1-4, ade2-1, ura3-1, and trp1-289. Diploid strains carrying mutations and/or producing tagged proteins were generated by mating appropriate haploids, which were constructed by transformation and/or genetic crosses.
Using a PCR strategy (Schneider et al. 1995), Mer2 was tagged after the 248th codon and Rec114 was tagged after the start codon with three copies of HA or Myc. Using other PCR templates (Longtine et al. 1998), Mei4 was tagged before the stop codon with 13 copies of Myc or with GFP. A strain producing Rec102-9Myc was constructed using pKK2 as described by Kee and Keeney (2002). In every case, the gene encoding a tagged protein was used to replace the corresponding gene on the chromosome. All genes encoding tagged proteins are fully functional, on the basis of assays of sporulation efficiency and spore viability. Spore viability was 100% for MER2-HA, 91% for MER2-MYC, 96% for MEI4-13MYC, 100% for MEI4-GFP, 96% for REC114-HA, 96% for REC114-MYC, 96% for MER2-HA MEI4-13MYC, 98% for MEI4-GFP REC114-MYC, 95% for MEI4-13MYC REC114-HA, 98% for MER2-MYC REC114-HA, and 98% for REC102-MYC MER2-HA 98% (22 tetrads dissected for each strain).
rec114∷kanMX4, ski8∷kanMX4, mre11∷kanMX4, xrs2∷kanMX4, mec1∷TRP1, sml1∷kanMX4, mek1∷kanMX4 disruptions were made by the PCR strategy of Longtine et al. (1998). Plasmids for introducing the following mutations have been described: mer2∷ADE2 (Engebrecht et al. 1991), rec102∷LEU2 (Bhargava et al. 1991), rec104∷LEU2 (Hollingsworth and Johnson 1993), mei4∷ADE2 (Menees and Roeder 1989), tel1∷URA3 (Usui et al. 2001), spo11-Y135F (Cha et al. 2000), and rad50-K181∷URA3 (Alani et al. 1990). rad50-K181∷URA3 is referred to as rad50S elsewhere in the text.
Anti-Mer2 antibody production and purification:
An MBP-Mer2 fusion protein was purified from Escherichia coli containing pJ88 by the methods of Guan and Dixon (1991). Purified MBP-Mer2 was used as antigen to raise antibodies in rabbits at the Pocono Rabbit Farm and Laboratory (Canadensis, PA). Serum was affinity purified according to Snyder (1989). A 1:100 dilution of the purified antibody was used in immunofluorescence experiments.
Analysis of chromosome spreads:
Meiotic chromosomes were spread and stained with antibodies as described (Bailis and Roeder 1998). Primary antibodies, rabbit anti-Zip1 (Sym et al. 1993), rabbit anti-Red1 (Smith and Roeder 1997), mouse anti-Myc 9E10 (Covance), mouse anti-HA 11 (Covance), guinea pig anti-Mre11 (Usui et al. 1998), guinea pig anti-GFP, and rabbit anti-GFP antibodies (AbCam) were used at 1:100 dilution. For secondary antibodies, anti-rabbit antibodies conjugated with FITC or Texas Red, anti-mouse antibodies conjugated with FITC or Texas Red, or anti-guinea pig antibodies conjugated with FITC (Jackson Immuno Research Labs) were used at 1:200 dilution. Chromosomal DNA was stained with 1 μg/ml DAPI. A fluorescence microscope (Nikon Eclipse E800, Plan Apo100X/1.4 oil objective) equipped with fluorescence optics was used to observe antibody-stained preparations. Images were captured with a Sensys CCD camera and a Coolsnap HQ camera (Photometrics, Tucson, AZ).
To assess overlap between different pairs of Mer2, Mei4, and Rec114 proteins, nuclei from cells harvested after 14 hr in sporulation medium were used, and at least 15 nuclei were examined for each strain and antibody combination tested. Student t-tests were performed to compare the extent of colocalization in wild type with that observed in mutants (Zar 1999). Fortuitous colocalization was determined by rotating an overlay corresponding to one of the two signals by 90° or 180° and determining the amount of overlay between the misoriented images as described by Gasior et al. (1998). In all cases, the frequency of fortuitous overlap was ∼10%.
Denatured yeast cell extracts were prepared as described by Kee and Keeney (2002). Immunoprecipitation, Western blotting, phosphatase treatment, and preparation of nondenatured cell extracts were carried out as described by Hong and Roeder (2002). Proteins were fractionated by electrophoresis on a 10% SDS–polyacrylamide gel at 100 V for 2 hr. The blot was then incubated with enhanced chemifluorescence substrate and exposed to the STORM PhosphorImager (Molecular Dynamics, Sunnyvale, CA). Mouse anti-HA 11 (Covance), rat anti-HA 3F10 (Roche), mouse anti-Myc 9E10 (Covance), rabbit anti-Myc (Abcam), and rat anti-tubulin antibodies (Sera-Lab) were used at 1:1000 dilution.
ImageQuant software (Molecular Dynamics) was used to view and quantify Western blots. Quantification of gel band intensities was carried out in ImageQuant following the manufacturer's instructions. Signal intensities of tubulin and Mer2 bands were determined using volume analysis with local-average background correction applied. The intensity of the Mer2 band was then normalized relative to tubulin.
Mer2 associates with chromosomes early in meiotic prophase:
To gain insight into the function of Mer2, anti-Mer2 antibodies were raised and used to localize the protein within meiotic cells. Meiotic nuclei that are surface spread and stained with anti-Mer2 antibodies display numerous foci, indicating that Mer2 associates with chromosomes (Figure 1A). This staining pattern is specific for Mer2, as no antibody staining is observed in spread nuclei from a mer2 null mutant (data not shown).
The timing of Mer2 localization was assessed by staining meiotic chromosomes from wild-type cells with antibodies to both Mer2 and Zip1 (Figure 1). Zip1 is a component of the synaptonemal complex and thus can be used as a marker for progression through meiotic prophase (Sym et al. 1993). Mer2 chromosomal foci are maximally abundant during the leptotene stage of meiotic prophase, prior to Zip1 localization to chromosomes. As chromosomes begin to synapse during zygotene, the number of Mer2 foci declines. By pachytene, when chromosomes are fully synapsed, the number of Mer2 foci on chromosomes is ∼15% of the number observed at leptotene. DSBs occur at late leptotene or very early zygotene (Padmore et al. 1991).
Mer2 localization is independent of other DSB proteins:
To determine whether any of the other DSB proteins are required for the association of Mer2 with chromosomes, Mer2 localization was examined in spread nuclei prepared from various null mutants. For this purpose, a gene encoding Mer2 tagged with three copies of the HA epitope was used to replace the wild-type MER2 gene on the chromosome and Mer2 was detected with anti-HA antibodies. Since chromosomes do not synapse in spo11 strains, Red1 (instead of Zip1) staining was used for staging. Red1 associates with the cores of meiotic chromosomes and displays punctate staining during early prophase (leptotene/zygotene) but fairly continuous staining at midprophase (pachytene) (Smith and Roeder 1997).
Mer2 localizes to chromosomes in the ski8, spo11, and mre11 null mutants, as shown in Figure 2, A–C. Mer2 also localizes to chromosomes in the absence of Rec102, Rec104, Rec114, Mei4, Xrs2, or Rad50 (data not shown). These results indicate that the chromosomal association of Mer2 is independent of all other DSB proteins.
Whereas the number of Mer2 foci in wild type declines significantly from leptotene to pachytene, this does not appear to be the case in mutants that fail to make DSBs. Mer2 foci were quantified in wild-type and spo11 strains using Red1 staining to stage nuclei. In contrast to the situation in wild type, Mer2 foci in spo11 are equally abundant at early and midprophase (Figure 2, B and D). Moreover, the number of Mer2 foci in spo11 is increased in early prophase and even more at midprophase (P < 0.05) compared to wild type at the corresponding stages. These results suggest that Mer2 foci do not dissociate from chromosomes with normal timing when DSB formation is impaired.
Mer2 increases in abundance and is phosphorylated during meiosis:
To compare Mer2 protein levels during meiosis and mitosis, extracts were prepared from both vegetative and meiotic cells producing Mer2-HA and protein levels were analyzed on Western blots. Under vegetative conditions, the anti-HA antibody specifically recognizes one band with an apparent molecular weight of 47 kDa (Figure 3A), which is somewhat higher than the predicted molecular weight of Mer2 (37 kDa) plus three copies of HA (3 kDa). Under meiotic conditions, an additional band of a slower mobility is apparent (Figure 3A), indicating that Mer2 is modified in a meiosis-specific manner.
To investigate the possibility that the modification is due to phosphorylation, Mer2-HA was immunoprecipitated from meiotic cells and treated with calf intestinal phosphatase (Figure 3B). Phosphatase treatment converted most of the slower migrating form of Mer2 to the faster migrating form. When phosphatase inhibitor was added to the phosphatase reaction, the slower migrating form persisted. These data indicate that the modified form of Mer2 present in meiotic cells is the result of phosphorylation.
Phosphorylation of Mer2 is independent of Mek1, Mec1, and Tel1:
Which kinase phosphorylates Mer2? Mek1 is a meiosis-specific kinase that is a component of chromosome cores and plays a role in the pachytene checkpoint (Bailis and Roeder 2000). The Mec1 and Tel1 kinases act upstream of Mek1 in the checkpoint pathway (Bailis and Roeder 2000). To determine whether phosphorylation of Mer2 depends on any of these kinases, extracts from wild-type, mek1, mec1, and tel1 strains producing Mer2-HA were analyzed by Western blotting. In all strains, a significant fraction of Mer2 is phosphorylated (Figure 3C), suggesting that Mer2 is not modified by Mek1, Mec1, or Tel1.
Normal degradation and dephosphorylation of Mer2 depends on DSBs:
To examine the abundance and phosphorylation of Mer2 throughout meiosis, Western blot analysis was carried out at different time points after the introduction into sporulation medium. Both wild-type and spo11 strains were examined to determine the effect of DSBs on Mer2 accumulation and phosphorylation.
In wild type, Mer2 protein reaches maximum abundance at 12 hr (Figure 4, A and C), corresponding to the leptotene stage, when the number of Mer2 foci on chromosomes is maximal. At this time point, the amount of Mer2 protein is approximately sevenfold higher than in vegetative cells. In spo11 cells (Figure 4, B and C), the kinetics of Mer2 accumulation are similar to wild type up to 12 hr, but Mer2 continues to accumulate until 16 hr (when most wild-type cells are in pachytene). The final amount of Mer2 is two- to threefold higher than the maximum level observed in wild type. Thus, degradation of Mer2 is substantially delayed in the absence of Spo11.
In wild type, the fraction of Mer2 protein that is phosphorylated reaches a maximum at 12 hr and then declines gradually (Figure 4, A and D). In contrast, in spo11 cells the fraction of Mer2 that is phosphorylated continues to increase until 16 hr and remains fairly constant thereafter (Figure 4, B and D). Although the fraction of Mer2 that is phosphorylated remains constant between 16 and 24 hr in spo11, the actual amount of phosphorylated protein decreases three- to fourfold (by 24 hr, both meiotic divisions have been completed).
To determine whether dephosphorylation of Mer2 depends on DSBs or on the Spo11 protein per se, a time-course analysis of Mer2 phosphorylation was carried out in a strain carrying a non-null allele of SPO11, known as spo11-Y135F. This point mutation in the Spo11 catalytic site prevents DSB formation, but allows the production of a full-length protein that can perform the pre-DSB functions of Spo11 (Cha et al. 2000). The pattern of Mer2 phosphorylation in spo11-Y135F is indistinguishable from that of the spo11 null mutant (data not shown), arguing that dephosphorylation of Mer2 depends on DSBs.
Mer2, Mei4, and Rec114 partially colocalize:
Two DSB proteins about which little is known are Mei4 and Rec114. We observed a strong interaction between these two proteins in the yeast two-hybrid protein system (data not shown). To investigate further the functions of these proteins, we determined whether they localize to chromosomes.
An MEI4 gene encoding Mei4 tagged with GFP was used to replace the wild-type MEI4 gene. When meiotic nuclei from MEI4-GFP cells were surface spread and labeled with anti-GFP antibodies, Mei4 was found to localize to foci on early prophase chromosomes (Figure 5, A and B). The number of Mei4 foci decreases from early to midprophase (Figure 6E), although this decline is not as pronounced as that observed for Mer2.
A REC114 gene encoding Rec114 tagged with three copies of the Myc epitope was used to replace the wild-type REC114 gene. Rec114-Myc localizes to chromosomes as foci (Figure 5, A and C), with the number of foci decreasing from early to midprophase (Figure 6G).
To test for colocalization of Mei4 and Rec114, spread nuclei from a REC114-MYC MEI4-GFP strain were stained with antibodies to both GFP and Myc. Colocalization was assessed at 14 hr, when most cells are in zygotene. This analysis revealed that Rec114 partially colocalizes with Mei4 (Figure 5A). About 70% of Rec114 foci contain Mei4 and ∼55% of Mei4 foci contain Rec114 (Figure 5D); these frequencies are significantly greater than the frequency of fortuitous overlap (∼10%) (see materials and methods). Thus, Rec114 and Mei4 interact with each other, as evidenced both by colocalization and by two-hybrid protein analysis.
Is there any overlap between Rec114/Mei4 foci and Mer2 foci? Staining spread nuclei from a MEI4-GFP strain with antibodies to Mer2 and GFP revealed overlap between Mei4 and Mer2 (Figure 5, B and D). Similarly, staining nuclei from a REC114-MYC strain with anti-Myc and anti-Mer2 antibodies demonstrated partial colocalization of Rec114 and Mer2 (Figure 5, C and D). In both cases, the extent of overlap is similar to that observed for Mei4 and Rec114 (Figure 5D) and significantly greater than the frequency of fortuitous overlap. These data suggest that Mer2, Mei4, and Rec114 are part of the same complex.
Independent localization of Mer2, Mei4, and Rec114:
To investigate the requirements for chromosomal localization of Mer2, Mei4, and Rec114, the proteins were localized in meiotic mutants lacking one or both of the other proteins. Both Mer2 and Mei4 localize in the rec114 mutant (Figure 6A). Mer2 and Rec114 both localize in mei4 strains (Figure 6B) and Mei4 and Rec114 localize in the absence of Mer2 (Figure 6C). With the exception of Mei4 in mer2 (described below), the number of foci on chromosomes in the mutants is similar to, or greater than, the number observed in wild type. In several cases, the number of Mer2/Mei4/Rec114 foci does not decline significantly as cells progress from early to midprophase (Figure 6, E–G), consistent with the results obtained for Mer2 in the spo11 mutant.
In an effort to determine whether proteins are properly localized in the mutants, colocalization was assessed. The extent of overlap is similar, or somewhat increased, compared to wild type, suggesting that the remaining proteins still associate with each other in the absence of the third protein (Figure 6D).
Localization of the Mer2/Mei4/Rec114 proteins was also tested in double mutants. The number of Mer2 foci in the mei4 rec114 double mutant is similar to that observed in the mei4 and rec114 single mutants (Figure 6F). Similarly, Rec114 localizes to the same extent in the mer2 mei4 double as it does in the mer2 and mei4 single mutants (Figure 6G). In the mer2 rec114 double mutant, Mei4 localization is similar to that observed in the mer2 single mutant (Figure 6E).
Taken together, these results demonstrate that the Mer2, Mei4, and Rec114 proteins localize to chromosomes largely independently of each other.
The mer2 mutation decreases chromosomal association of Mei4:
Although Mei4 localizes normally in the absence of Rec114, it does not localize properly in the absence of Mer2. When a mer2 strain carrying GFP-tagged Mei4 was stained with anti-GFP antibodies, the number of foci observed was significantly less than in wild type at both early and midprophase (Figure 6, C and E). Furthermore, a Mei4 protein tagged with Myc was not detectable on spread chromosomes from a mer2 strain, although it localized normally in wild type (data not shown). These data suggest that Mer2 facilitates loading of Mei4 onto chromosomes and/or stabilizes its association with chromosomes.
Mer2, Mei4, and Rec114 co-immunoprecipitate:
Co-immunoprecipitation experiments were carried out to investigate further the possibility that Mer2, Mei4, and Rec114 form a complex. Nondenatured meiotic cell extracts prepared from cells producing Mer2-HA and Mei4-Myc were immunoprecipitated with anti-HA antibodies and then probed for Mei4 using anti-Myc antibodies. As shown in Figure 7A, Mei4-Myc and Mer2-HA co-immunoprecipitate. Similarly, Mei4-Myc and Rec114-HA co-immunoprecipitate (Figure 7B) and Mer2-Myc and Rec114-HA co-immunoprecipitate (Figure 7C).
Extracts from the mer2, mei4, and rec114 mutants were tested for co-immunoprecipitation of the remaining two components of the putative Mer2/Mei4/Rec114 complex. Mer2-HA and Mei4-Myc co-immunoprecipitate in rec114 (Figure 7A). Similarly, Mei4-Myc and Rec114-HA co-immunoprecipitate in mer2 (Figure 7B) and Rec114-HA and Mer2-Myc co-immunoprecipitate in mei4 (Figure 7C). Note that a larger portion of Mer2 is phosphorylated in the rec114 mutant (Figure 7A), consistent with the finding that dephosphorylation of Mer2 depends on DSB formation. These data demonstrate that Rec114, Mei4, and Mer2 co-immunoprecipitate in all pairwise combinations and each interaction is independent of the third protein.
The Mer2/Mei4/Rec114 complex is distinct from previously characterized complexes:
To examine the relationship of the Mer2/Mei4/Rec114 complex to the Mre11/Rad50/Xrs2 and Spo11/Ski8/Rec102/Rec104 complexes, spread nuclei were double stained for components of two different complexes. Mre11 was examined as a representative of the Mre11/Rad50/Xrs2 complex, Rec102 as a representative of the Spo11/Ski8/Rec102/Rec104 complex, and Mer2 as a representative of the Mer2/Mei4/Rec114 complex.
When an mre11 strain carrying a plasmid producing Mre11 tagged with 13 copies of the Myc epitope (D'Amours and Jackson 2001) was surface spread and stained with anti-Myc antibodies, Mre11-Myc showed a dotty staining pattern on early prophase chromosomes (Figure 8A). This is the first time that Mre11 has been shown to localize to chromosomes in wild-type cells of budding yeast. The extent of overlap between Mre11-Myc and Mer2 (Figure 8A) is not significantly greater than the level of fortuitous overlap.
These data suggest that the Mer2-containing complex localizes to different chromosomal locations than the Mre11- and Spo11-containing complexes. However, an alternative interpretation is that the complexes localize to the same locations, but at different points in time. In this case, it might be possible to demonstrate overlap by blocking DSB repair and thereby perhaps blocking the dissociation of complexes from chromosomes. Colocalization of Mer2 with Mre11 or Rec102 was tested in a rad50S mutant, in which DSBs are made but their processing to expose single-stranded tails is blocked (Alani et al. 1990). The rad50S mutation did not increase the degree of overlap between Mer2 and Mre11 (Figure 8C) or between Mer2 and Rec102 (Figure 8D). Colocalization between Mer2 and Rec102 was also tested in the spo11-Y-135F point mutant. No significant colocalization of Mer2 and Rec102-Myc was observed (Figure 8E).
Co-immunoprecipitation between Mer2 and Rec102 was also tested. Nondenatured meiotic cell extracts from cells producing Mer2-HA and Rec102-Myc were immunoprecipitated with anti-HA antibody and probed for Rec102-Myc with anti-Myc antibody. Mer2-HA does not co-immunoprecipitate with Rec102 (Figure 7D).
Taken together, the cytological and immunoprecipitation data indicate that the Mer2-containing complex is distinct from the Mre11- and Rec102-containing complexes.
Mer2, Rec114, and Mei4—a novel complex:
Although Mer2, Mei4, and Rec114 were identified some time ago, it has remained unclear whether these proteins are directly involved in the formation of DSBs. They might influence DSB formation indirectly, perhaps by regulating the production or activity of other proteins more directly involved in the initiation of recombination. Our demonstration that Mer2, Mei4, and Rec114 localize to meiotic chromosomes suggests that these proteins either are directly involved in DSB formation or promote a chromosome structure that is conducive to breakage. Previous studies have shown that DSBs are formed just prior to or simultaneously with the initiation of SC formation (Padmore et al. 1991). Thus, the timing of Mer2/Mei4/Rec114 localization to chromosomes (in leptotene/zygotene) is consistent with a role in the initiation of recombination.
Several observations indicate that Mer2, Mei4, and Rec114 form a complex. First, these proteins partially colocalize with each other on meiotic chromosomes. Second, these proteins co-immunoprecipitate from meiotic cell extracts. Third, we observed a strong interaction between Mei4 and Rec114 in the two-hybrid protein system. Arora et al. (2004) also observed this interaction, as well as interactions between Mer2 and Rec114 and between Mer2 and Mei4.
Additional evidence for an interaction between Mer2 and Mei4 comes from examining the genetic requirements for protein localization. Chromosomal localization of tagged versions of Mei4 is impaired in the absence of Mer2, even though the tagged proteins localize normally in wild type. Two different tagged versions of Mei4 differ in terms of their dependence on Mer2, despite the fact that both proteins are fully functional in wild type (on the basis of spore viability). Although both tags are located at the carboxyl terminus of the protein, they may have differential effects on the ability of Mei4 to interact stably with its partners. The genetic interaction between mer2 and the MEI4-MYC and MEI4-GFP mutations is consistent with the notion that Mer2 and Mei4 interact physically. Although Mei4-Myc was not detected on chromosomes in the mer2 mutant, Mei4-Myc did precipitate with Rec114-HA from a mer2 extract. This result suggests either that the interaction between Mei4 and Rec114 does not depend on their chromosomal association or that Mei4-Myc is present on chromosomes in the mer2 mutant, but this association is not sufficiently stable to survive the spreading procedure.
Requirements for localization of Mer2, Mei4, and Rec114 to chromosomes:
Mer2 localizes to chromosomes normally in the absence of Mei4 and Rec114. Similarly, localization of Rec114 is independent of both Mei4 and Mer2 and Mei4 localizes properly in the absence of Rec114. Although localization of tagged Mei4 is compromised in the absence of Mer2, it is possible that localization of wild-type (i.e., untagged) Mei4 does not depend on Mer2. Thus, the three proteins in the complex appear to localize to chromosomes largely independently of each other. In single mutants, the remaining two proteins still colocalize with each other, implying that they are assembling at the proper chromosomal locations.
How are Mer2, Mei4, and Rec114 recruited to chromosomes? It is possible that one component (or more) of the Mer2/Mei4/Rec114 complex remains to be identified. Perhaps a fourth component (protein X) interacts with all three proteins, thus accounting for their colocalization and coprecipitation. In the case of the Mre11/Rad50/Xrs2 complex, Mre11 is the bridge that holds Rad50 and Xrs2 together (Usui et al. 1998). Perhaps the Mer2, Mei4, and Rec114 proteins localize to chromosomes by virtue of their interaction with protein X, and this association is stabilized by interactions with each other. In this case, the tagged Mei4 proteins might have weakened interactions with protein X and thus depend on Mer2 for its stable association with chromosomes.
Protein X need not be a meiosis-specific protein. Two-hybrid interactions between Mer2 and Mei4 and between Mei4 and Rec114 are observed in both vegetative and meiotic cells (Arora et al. 2004) (data not shown), implying that the postulated bridging protein is present in vegetative cells. It is also possible that Mer2, Mei4, and Rec114 localize to chromosomes, not by virtue of interaction with a particular protein, but due to interaction with a specific chromatin or DNA structure.
Mer2, Mei4, and Rec114 form a distinct complex:
A couple of observations suggest that the Mer2/Mei4/Rec114 complex is distinct from the Spo11/Ski8/Rec102/Rec104 and Mre11/Rad50/Xrs2 complexes. First, Mer2 does not colocalize with Mre11 or Rec102 on chromosomes. Second, Mer2 does not co-immunoprecipitate with Rec102.
These observations do not rule out the possibility that the Mer2/Mei4/Rec114 complex localizes to the same chromosomal locations as one or both of the other complexes. The complexes might act at the same sites, but in a specific temporal sequence such that one complex is released before the next complex is loaded onto chromosomes. We looked into this possibility by testing for colocalization in a rad50S mutant, in which DSBs are made but not processed to expose single-stranded tails (Alani et al. 1990). We observed no colocalization of Mer2 with either Mre11 or Rec102 in the rad50S background. We also tested for colocalization in the spo11-Y135F mutant, in which a full-length Spo11 protein is present but enzymatically inactive (Cha et al. 2000). We saw no evidence of colocalization of Mer2 with Mre11 or Rec102 in spo11-Y135F. These observations suggest that the Mer2/Mei4/Rec114 complex goes to different sites on chromosomes than the other two complexes, although they do not rule out the possibility of localization to the same sites at different times. The presence of the different complexes might be mutually exclusive, such that the Mer2/Mei4/Rec114 complex cannot occupy the same site at the same time as a Spo11/Ski8/Rec102/Rec104 or Mre11/Rad50/Xrs2 complex.
A number of observations indicate that the Spo11- and Mre11-containing complexes localize to the sites of DSBs. In a rad50S mutant, Spo11 remains covalently bound to DSB ends (Keeney et al. 1997) and Spo11 and Mre11 show extensive colocalization on chromosomes under these conditions (Prieler et al. 2005). Furthermore, both Spo11 and Mre11 have been shown to localize to DSB hotspots by chromatin immunoprecipitation (Borde et al. 2004; Prieler et al. 2005). Taken together, these observations suggest that the Mer2/Mei4/Rec114 complex does not localize to the sites of meiotic DSBs.
Dephosphorylation and chromosomal dissociation of Mer2 depends on DSBs:
The Mer2 protein is phosphorylated specifically in meiotic cells. This phosphorylation does not depend on DSB formation or on the presence of Mei4 or Rec114. Phosphorylated Mer2 accumulates in the spo11 mutant, suggesting that this form of Mer2 is active in DSB formation.
Accumulation, phosphorylation, and chromosomal association of Mer2 are tightly linked to each other. When DSB formation is blocked by the spo11 mutation, the Mer2 protein accumulates to a higher level and reaches maximum abundance at a later time than in wild type. In addition, a greater fraction of the protein is phosphorylated and the protein remains phosphorylated even after nuclear division is complete. The same is true in the spo11-Y135F point mutant, arguing that the effect results from a failure to make DSBs rather than from the absence of the Spo11 protein. Cytological analysis demonstrates that Mer2 remains on chromosomes much later in spo11 cells compared to wild type; Mer2 is still on chromosomes at the time of the meiosis I division (data not shown). Formation of DSBs may trigger dephosphorylation of Mer2, which may in turn lead to its dissociation from chromosomes and its subsequent degradation.
Possible functions for the Mer2/Mei4/Rec114 complex:
Despite many years of study in numerous laboratories, still very little is understood about the functions of the DSB proteins. Only Spo11, the endonuclease responsible for cleavage, has an assigned function in DSB formation (Keeney et al. 1997). The function of the remaining proteins is not simply to localize Spo11 to chromosomes. If Spo11 is targeted to induce cuts at the GAL promoter by fusion to the Gal4 DNA-binding domain, cleavage still depends on the other nine DSB proteins (Pecina et al. 2002). Thus, these proteins are required to activate Spo11 and/or to prepare an appropriate chromosomal substrate for cleavage by Spo11.
Several observations suggest that special features of chromatin structure or chromosome organization are important determinants of meiotic DSB formation. Sequences that serve as hotspots for DSB formation become more accessible (as evidenced by an increase in nuclease sensitivity) just prior to recombination initiation (Ohta et al. 1994). Spo11 cannot cleave DNA that has not undergone premeiotic DNA replication and the associated transition in chromatin structure (Borde et al. 2000; Cha et al. 2000; Smith et al. 2001; Murakami et al. 2003). Sister-chromatid cohesion is regulated in a meiosis-specific manner and cohesion depends on proteins produced specifically in meiotic cells (Petronczki et al. 2003). Mutations affecting the chromosomal core proteins, Red1 and Mek1, lead to precocious separation of sister chromatids and also cause a substantial decrease in DSB formation (Leem and Ogawa 1992; Schwacha and Kleckner 1997; Bailis and Roeder 1998). The importance of histone modifications for efficient DSB formation provides additional evidence that chromatin structure plays an important role in recombination initiation (Reddy and Villeneuve 2004; Sollier et al. 2004; Yamada et al. 2004; Yamashita et al. 2004).
Our data suggest that the Mer2/Mei4/Rec114 complex is not located at DSB sites, leading us to propose that this complex can act at a distance to promote DNA cleavage by Spo11. Perhaps the complex triggers a change in chromatin structure that is transmitted along the chromosome to the sites of DSB hotspots. Ohta et al. (1994) have shown that the meiosis-specific increase in nuclease sensitivity observed at DSB sites is diminished in the mre2 mutant. Since the Mre2 protein is involved in splicing MER2 pre-mRNA (Nakagawa and Ogawa 1997), the phenotype of mre2 presumably reflects a defect in Mer2 function. Another possibility is that the Mer2/Mei4/Rec114 complex serves to restrict Spo11 activity to hotspots by binding to the intervening cold regions. A third, not mutually exclusive, possibility is that the Mer2/Mei4/Rec114 complex is involved in sister-chromatid cohesion, either promoting cohesion or recognizing the appropriate chromosome structure and transmitting this information to the DSB machinery. Communication with the Spo11-containing complex might involve a direct, but transient, interaction. Alternatively, Mer2 may be phosphorylated when the proper chromatin structure is recognized and this phosphorylation may serve as a signal to trigger DSB formation.
We thank S. Keeney and S. Jackson for providing plasmids. Oligonucleotides were synthesized by the W. M. Keck Foundation Biotechnology Resource Laboratory. We thank members of the Roeder lab for helpful discussion and comments on the manuscript. This work was supported by grant GM28904 from the United States Public Health Service to G.S.R. and by the Howard Hughes Medical Institute.
Communicating editor: P. J. Pukkila
- Received March 30, 2006.
- Accepted June 6, 2006.
- Copyright © 2006 by the Genetics Society of America