Impaired Resection of Meiotic Double-Strand Breaks Channels Repair to Nonhomologous End Joining in Caenorhabditis elegans

Abstract

Repair of double-strand DNA breaks (DSBs) by the homologous recombination (HR) pathway results in crossovers (COs) required for a successful first meiotic division. Mre11 is one member of the MRX/N (Mre11, Rad50, and Xrs2/Nbs1) complex required for meiotic DSB formation and for resection in Saccharomyces cerevisiae. In Caenorhabditis elegans, evidence for the MRX/N role in DSB resection is limited. We report the first separation-of-function allele, mre-11(iow1) in C. elegans, which is specifically defective in meiotic DSB resection but not in formation. The mre-11(iow1) mutants displayed chromosomal fragmentation and aggregation in late prophase I. Recombination intermediates and crossover formation was greatly reduced in mre-11(iow1) mutants. Irradiation-induced DSBs during meiosis failed to be repaired from early to middle prophase I in mre-11(iow1) mutants. In the absence of a functional HR, our data suggest that some DSBs in mre-11(iow1) mutants are repaired by the nonhomologous end joining (NHEJ) pathway, as removing NHEJ partially suppressed the meiotic defects shown by mre-11(iow1). In the absence of NHEJ and a functional MRX/N, meiotic DSBs are channeled to EXO-1-dependent HR repair. Overall, our analysis supports a role for MRE-11 in the resection of DSBs in middle meiotic prophase I and in blocking NHEJ.

INTRODUCTION

Meiotic recombination is initiated by the formation of meiotic double-strand breaks (DSBs), which are then repaired by the meiotic homologous recombination (HR) pathway. A meiosis protein, SPO-11, along with other auxiliary proteins catalyzes the formation of meiotic DSBs and remains covalently bound to the 5′ ends of the DSBs (1). These DSBs are converted to long 3′ single-stranded DNA (ssDNA) by a process called resection. Meiotic DSB resection is initiated by endonuclease activity to release oligonucleotide-bound SPO-11 and generate a short 3′ overhang (2). The short 3′ overhang is further processed by a 5′-3′ exonuclease activity to generate long 3′ ssDNA (3, 4). The ssDNA is rapidly bound by the ssDNA binding protein RPA, which is then displaced by RAD-51, forming a filament that coats the ssDNA (5). RAD-51 then mediates the invasion of the ssDNA into the homologous double-strand DNA (dsDNA) (6, 7). Using the homologous sequence as the template, new DNA is synthesized from the invading 3′ ends to fill in the region lost by end resection. Ligation of the invading strands forms a double Holliday junction (dHJ), which can be resolved to form crossovers (COs) (8). COs connect the homologous pairs of chromosomes, enabling their biorientation toward opposite spindle poles in metaphase I.

Mre11 acts in the context of the MRX/N complex, which is composed of Mre11, Rad50, and Nbs1/Xrs2. The MRX/N complex has dual roles in the early steps of the HR pathway in some organisms. First, MRX/N acts as a cofactor for Spo11 and is required for meiotic DSB formation in Saccharomyces cerevisiae (9). Second, MRX/N is required for the resection of meiotic DSBs in S. cerevisiae, Schizosaccharomyces pombe, and Arabidopsis sp. (10–13). The N terminus of the Mre11 protein contains five evolutionarily conserved phosphoesterase motifs responsible for the endonuclease and 3′-5′ exonuclease activity of the Mre11 protein (10, 14–16). Point mutations in those motifs disrupt the MRX/N complex function in DSB resection but not in DSB formation (10, 14–16). The endonuclease activity of Mre11 is critical for the initiation of DSB resection and the removal of SPO-11 in meiosis (2, 10, 15–18). The 3′-5′ exonuclease activity of Mre11 is implicated in further resection of DSB ends in addition to other 5′-3′ exonucleases, such as Exo1 and Dna2 (3, 4, 19). In contrast to evidence from yeast, the meiotic function of MRX/N in metazoan systems is poorly understood, due to the lethality caused by the disruption of these genes (20–23). The study of a nuclease-deficient allele of Mre11 in mice suggests a role for MRX/N in DSB resection in mitosis (24). However, little is known about MRX/N complex function in meiosis in these organisms.

An alternative mechanism for repair of DSBs is the nonhomologous end joining (NHEJ) pathway, which results in random joining of two dsDNA ends. This pathway is initiated by the binding of the Ku70/Ku80 heterodimer to dsDNA ends. Ku70/80 binding serves to protect the dsDNA ends from degradation as well as recruit DNA ligase IV and its accessory proteins. Ligase IV can then join DNA ends regardless of their homology (25). NHEJ is an error-prone pathway that frequently results in addition or removal of base pairs and/or chromosomal rearrangement (26–28). The NHEJ pathway is used primarily in the G₁ phase of the cell cycle, when a homolog or sister chromatid is not nearby (29). The role of the MRX/N complex in NHEJ is controversial. The MRX/N complex, but not its nuclease activity, has been implicated in the NHEJ pathway for repair of mitotic DSBs in S. cerevisiae (16, 28, 30). However, evidence suggests that the MRX/N complex is not required for NHEJ in S. pombe and vertebrates (31, 32). In Caenorhabditis elegans, NHEJ is exclusively used in nondividing somatic cells for DNA repair, but normally it is not active in germ line cells (33). Nevertheless, NHEJ has been shown in C. elegans to repair meiotic DSBs in the germ line when HR is not available (34, 35).

In C. elegans, MRX/N is required to form meiotic DSBs, since mutants defective for Mre11 or Rad50 show no recombination intermediates as defined by the presence of RAD-51 or crossover formation and the finding that chromosomes remain intact (36, 37). These phenotypes are indistinguishable from those of spo-11 null mutants, indicating a great reduction in meiotic DSB formation (37, 38). However, unlike spo-11 mutants, the introduction of exogenous DSBs into mre-11 or rad-50 null mutants fails to restore crossovers and results in chromosomal fragmentation and aggregation at diakinesis, similar to mutants that are perturbed in later stages of meiotic DSB repair (e.g., rad-51, com-1, or Cebrc-2 [34, 39, 40]). These studies suggest a role for MRX/N downstream of DSB formation in the repair of exogenous DSBs (36, 37). This downstream role may be specific to irradiation-induced DSBs, since meiotic DSBs have a distinct structure (i.e., SPO-11 is covalently bound to the 5′ end of the DSBs).

Here we report the analysis of meiotic recombination in mre-11(iow1), the first separation-of-function allele of mre-11 in C. elegans, which is competent to form meiotic DSBs but not process them. Our analysis indicated that in the presence of a defective MRX/N, some meiotic DSBs are repaired via NHEJ. In the absence of a functional MRX/N and NHEJ, an EXO-1-dependent resection becomes active in middle to late prophase I. This EXO-1-mediated resection can partially bypass the requirement of MRX/N.

MATERIALS AND METHODS

Strains.

All C. elegans strains were cultured under standard conditions at 20°C (60). Strain N2 worms were utilized as the wild-type background, while Hawaiian CB4856 wild-type worms were used for single-nucleotide polymorphism (SNP) mapping in the process of mapping mre-11(iow1). The following mutations and chromosome rearrangements were used (33, 38, 46, 51): LGIII, cKu-80(ok861), exo-1(tm1842); LGIV, spo-11(ok79), msh-5(me23), him-6(ok4128), rad-51(lg8701), nT1[qIs50](IV, V); and LGV, mre-11(ok179), mre-11(iow1). The following transgenic line was used (49, 61): opIs263(RPA-1::YFP), meIs8(GFP::COSA-1).

Isolation of mre-11(iow1).

We isolated the iow1 allele in a forward genetic screen by using ethyl methanesulfonate as the mutagen. We screened for worms that showed a maternally rescued embryonic lethality, since worms with meiotic defects can survive to adulthood but fail to produce viable progeny. Then we performed a cytology-based secondary screen to identify worms with aberrant chromosome morphology at diakinesis. Allele iow1 was linked to SNPs located in a region of chromosome V. In addition, iow1 failed to complement mre-11(ok179). The point mutation was identified by Sanger sequencing of the mre-11 gene (including 5′ and 3′ untranslated regions [UTRs]), by using the following primers: forward primers GGATTCCAACATGCGAGATTGTAG, CAGTCTCGACGAGGAGACGAG, GCTACTTCGCTCACTCCGGAAG, and GATAGGAATTGGAAGAGCTTTAGAGC and reverse primers CGCGATTGGATTCCCATTCAAGC, CATCTATGTATCGGCCATCCGTGTG, GATCGTTAATCGGAGGCAGAGGC, and CACTCCCTAATAGTCTTGCACTCCC.

Western blotting.

Rabbit anti-MRE-11 antibody was a generous gift from Simon Boulton. Adult homozygote worms were picked or selected via fluorescence-activated cell sorting (62). The rabbit anti-MRE-11 antibody was used as primary antibody (1:1,000). Mouse antitubulin (1:1,000; DSHB) was used as the loading control. Secondary antibodies used were anti-rabbit antibody conjugated to horseradish peroxidase (HRP; 1:10,000) and anti-mouse antibody–HRP. Tween–Tris-buffered saline (1×)–5%milk was used for incubation and blocking.

Apoptosis analysis.

Germ cell corpses were scored in adult hermaphrodites 20 h post-L4 stage, as described in reference 46. Statistical comparisons between genotypes were performed using the two-tailed Mann-Whitney test, with 95%confidence intervals (CI).

Immunostaining and microscopy.

Whole-worm preparation was performed as described in reference 42. Oocytes in whole-worm preparations were scored in adult hermaphrodites 20 h post-L4 stage. Immunostaining for RAD-51 and MRE-11 was performed as described in reference 42. Primary antibodies were rabbit anti-RAD-51 (1:10,000; ModEncod) and rabbit anti-MRE-11 (1:100; Simon Boulton). The secondary antibodies used were Alexa Fluor 488–anti-rabbit antibody (1:500) and Alexa Fluor 555–anti-rabbit antibody. YFP::RPA-1 and GFP::COSA-1 focus microscopy was performed as follows. Adult hermaphrodites 20 h post-L4 were dissected to release gonads on slides. Slides were frozen on dry ice and dipped into methanol for 1 min, followed by 15 min of fixation in 4%formaldehyde in the dark. Then, slides were washed in phospate-buffered saline–Tween (PBST) for 5 min, stained with 4′,6-diamidino-2-phenylindole (DAPI; 1:2,000 dilution of a 5-mg/ml DAPI stock in PBST) for 5 min, and washed in PBST for 5 min. Vectashield was added to slides.

The images were acquired using the DeltaVision wide-field fluorescence microscope system (Applied Precision) with Olympus 100×/1.40-numerical aperture lenses. Optical sections were collected at 0.20-μm increments with a coolSNAP_HQ camera (Photometrics) and softWoRx software (Applied Precision) and deconvolved using softWoRx 5.0.0 software, except for Fig. S1 in the supplemental material, which is not deconvolved and is a single optical section. Images are projections through three-dimensional data stacks of whole nuclei (15 to 30 0.2-mm slices/stack).

Focus quantification.

Quantification of RAD-51 foci was performed for all seven zones composing the premeiotic tip to late pachytene regions of the germ line, as described in reference 42. The total numbers of nuclei scored per zone (arranged sequentially from zone 1 to zone 7) from three gonads from each genotype without irradiation was as follows: wild-type, 96, 120, 103, 84, 102, 60, and 56; spo-11, 127, 135, 117, 113, 114, 91, and 74; mre-11(iow1), 159, 149, 115, 94, 84, 64, and 68; cKu-80, 81, 101, 88, 90, 87, 75, and 75; mre-11(iow1); cKu-80, 133, 127, 104, 99, 51, 67, and 72; exo-1, 81, 117, 108, 88, 80, 64, and 60; exo-1; cKu-80 [mre-11(iow1)/nT1; exo-1; cKu80 were used as exo-1; cKu-80 double mutants], 80, 84, 88, 62, 54, 52, and 46; mre-11(iow1); exo-1, 58, 84, 96, 93, 75, 68, and 53; mre-11(iow1);exo-1; cKu-80, 52, 60, 68, 68, 67, 61, and 50; mre-11(ok179), 79, 110, 94, 88, 92, 89, and 71; him-6, 92, 110, 94, 74, 102, 71, and 60; mre-11(iow1); him-6, 75, 84, 82, 77, 70, 71, and 56; and mre-11(iow1); him-6; cKu-80, 75, 92, 93, 84, 83, 67, and 51. Due to the high number of RAD-51 foci per nucleus, only clearly nonoverlapped nuclei were counted from three gonads, except for mre-11(iow1); exo-1. The total numbers of nuclei scored per zone (arranged sequentially from zone 1 to zone 7) from three gonads from each genotype with irradiation were as follows: wild-type, 58, 55, 40, 51, 47, 58, and 32; spo-11, 67, 50, 58, 55, 55, 38, and 38; mre-11(iow1), 57, 90, 108, 108, 92, 73, and 66; exo-1, 69, 87, 69, 62, 55, 36, and 41; and mre-11(iow1); exo-1 (data from 6 gonads), 120, 175, 179, 154, 139, 124, and 109. Statistical comparisons between genotypes were performed using the two-tailed Mann-Whitney test, with 95%CI.

Quantification of RPA-1::YFP foci was performed in the same manner as for RAD-51 foci, with the following total numbers of nuclei scored per zone (arranged sequentially from zone 1 to zone 7) from the three gonads: rpa-1::yfp, 63, 73, 52, 55, 45, 38, and 39, and mre-11(iow1); rpa-1::yfp, 51, 81, 76, 65, 61, 56, and 44. Statistical comparisons between genotypes were performed using the two-tailed Mann-Whitney test, with 95%CI.

Quantification of GFP::COSA-1 foci was performed for zone 8 and zone 9, composed of nuclei from late pachytene through diplotene, with the zone sizes as described in reference 42. The total numbers of nuclei scored for each type were as follows: gfp::cosa-1, 87 (3 gonads); mre-11(ok179); gfp::cosa-1, 89 (4 gonads); mre-11(iow1); gfp::cosa-1, 77 (3 gonads); mre-11(iow1)/nT1; cKu-80; gfp::cosa-1, 91 (5 gonads); mre-11(iow1); cKu-80; gfp::cosa-1, 133 (5 gonads); mre-11(iow1)/nT1; msh-5; gfp::cosa-1, 78 (4 gonads); and mre-11(iow1); msh-5; gfp::cosa-1, 97 (3 gonads). Statistical comparisons between genotypes were performed using the two-tailed Mann-Whitney test, with 95%CI.

MRE-11 fluorescence intensity quantification.

The fluorescence intensity of MRE-11 was quantified by using the ImageJ program. The fluorescence intensities (arbitrary signal intensity per area unit) of 12 nuclei and 12 cytoplasmic areas in the transition zone were measured per gonad and were normalized to the background intensity (arbitrary signal intensity per area unit). The quantification was done with the middle panel of the three-dimensional stacks, with a fluorescence threshold of 256, using ImageJ. The total number of nuclei and the cytoplasmic area scored for each type were as follows: wild type, 180 (15 gonads); mre-11(ok179), 120 (10 gonads); and mre-11(iow1), 60 (5 gonads). Statistical comparison between genotypes were performed using the two-tailed Mann-Whitney test, with 95%CI.

RESULTS

mre-11(iow1) is a hypomorph of mre-11 with defects downstream of meiotic DSB formation.

mre-11(iow1) was isolated from a forward genetic screen (see Materials and Methods) designed to identify meiotic genes involved in chiasmata formation. mre-11(iow) failed to complement the null deletion mutation of mre-11, mre-11(ok179), indicating iow1 is allelic to mre-11 (Table 1) (36). Sequencing of mre-11(iow1) revealed a C-T transition, expected to result in substitution of threonine (T) with isoleucine (I) at amino acid 72 (Fig. 1A and B). This threonine residue is conserved in all Mre-11 orthologs excluding that in S. pombe (Fig. 1B) and is located in the first phosphoesterase motif immediately upstream of the residue corresponding to D16 in S. cerevisiae (Fig. 1A and B). The D16 mutation causes a separation-of-function phenotype in S. cerevisiae; studies of the D16 mutation in S. cerevisiae have demonstrated that the first phosphoesterase motif is essential for the nuclease activity of Mre-11 (among other phenotypes [14, 41]).

Table 1.

Embryonic lethality and HIM rate analysis

Genotype	Avg ± SE no. of eggs laid (no. of broods)	%inviable eggs (total no. of eggs)	%malesa (total no. of adults)
Wild type	217 ± 43 (6)	0 (1,303)	0 (715)
mre-11(ok179)	156 ± 23(8)	97.4 (1,247)	25 (12)
mre-11(iow1)	201 ± 22 (3)	99.8 (605)	NA
cKu-80	178 ± 43 (7)	0 (1,245)	0.4 (810)
mre-11(ok179)/mre-11(iow1)	213 ± 37 (5)	99.9 (908)	NA
mre-11(iow1); cKu-80	163 ± 77 (6)	87 (975)	16.7 (48)

^aThe percentage of males was determined separately from embryonic lethality. NA, not applicable.

Fig 1.

Defects in DSB repair in the mre-11(iow1) mutant. (A) Protein architecture of the predicted C. elegans MRE-11 protein. Black segments indicate the five conserved phosphoesterase motifs (14). The asterisk marks the point mutation in mre-11(iow1). (B) C. elegans MRE-11 predicted protein sequence of the first phosphoesterase motif, aligned with orthologs from human, mouse, Drosophila melanogaster, S. pombe, and S. cerevisiae. The asterisk marks the point mutation of mre-11(iow1). (C to J) Single DAPI-stained oocyte nuclei of the indicated genotypes at diakinesis 1. Each panel shows a projection of a three-dimensional stack of the entire nucleus. The arrowhead indicates a chromosomal aggregate, and the arrow indicates chromosomal fragments. Bars, 2 μm. (K and L) Frequency distribution of DAPI bodies at diakinesis 1.

We analyzed the morphology of meiotic chromosomes by staining DNA with DAPI (DAPI bodies). DSBs are generated when nuclei enter meiosis (transition zone) and are processed to form COs by late pachytene (42). However, these COs (chiasmata) cannot be observed cytologically until the diakinesis stage, when chromosomes are highly condensed. In C. elegans, the six pairs of homologous chromosomes (bivalents) are visualized as six DAPI bodies most clearly in diakinesis 1 (at the last oocyte stage before fertilization) (Fig. 1C). A complete loss of crossovers results in 12 DAPI bodies, each representing one univalent (one chromosome, e.g., spo-11 [38]). Defects in the repair of DSBs lead to a different outcome; DAPI bodies of varied sizes and numbers between 1 and >12 are observed (e.g., rad-51, com-1, and Cebrc-2 [34, 39, 40]). We considered DAPI bodies as those that by visual inspection appeared smaller than the size of a univalent as fragments, while ones that appeared bigger than a bivalent were considered aggregates. We reasoned that oocytes with fewer than six DAPI bodies must have at least one aggregate. The mre-11(iow1) mutants showed a combination of chromosomal aggregates and fragments (which ranged from 1 to 14), sharply contrasting with the 6 bivalents or 12 univalents observed in the wild type and the null mre-11(ok179) mutant, respectively (Fig. 1C, D, and E). The 12 univalents observed in mre-11(ok179) mutants indicated a great reduction in meiotic DSB formation, while the aggregates and fragments of mre-11(iow1) mutants also have been found in mutants that fail to repair meiotic DSBs via HR (e.g., rad-51, com-1, and Cebrc-2 [34, 39, 40]). The mre-11(iow1) allele is recessive to the wild-type allele of mre-11, because the diakinesis oocytes of mre-11(iow1)/+ worms contained six DAPI bodies (data not shown). The mre-11(iow1) allele failed to complement the null mre-11(ok179) allele, because oocytes of mre-11(iow1)/mre-11(ok179) mutants showed chromosomal aggregates and fragments, a phenotype similar to the mre-11(iow1) homozygous mutant (Fig. 1F). The observation that oocytes of mre-11(iow1)/mre-11(ok179) mutants were similar to mre-11(iow1) mutants strongly suggests that a residual function of MRE-11 is retained by the protein encoded by the iow1 allele.

An inability to form bivalents leads to chromosome missegration in meiosis I. Nondisjunction of the X chromosome in meiosis I results in an increased percentage of males (XO) with the HIM (high incidence of males) phenotype. Nondisjunction of autosomes leads to inviable aneuploid embryos (36, 43). Compared to the wild type, mre-11(iow1) and mre-11(ok179) mutants both showed elevated levels of embryonic lethality (Table 1) (Fisher's exact test, P < 0.0001). The embryonic lethality of mre-11(iow1) progeny was significantly higher than that of mre-11(ok179) progeny (Table 1) (Fisher's exact test, P < 0.0001), consistent with the severe defects of chromosomal morphology observed in oocytes of mre-11(iow1). Most lethality of mre-11(ok179) progeny can be attributed to aneuploidy due to chromosome missegregation in meiosis (36). The even greater lethality of mre-11(iow1) progeny presumably reflects the combined consequences of missegregation and defective repair of meiotic DSBs. The almost-complete embryonic lethality has also been observed in meiotic DSB repair mutants (e.g., rad-51, com-1, and Cebrc-2 [34, 39, 40]). The embryonic lethality of mre-11(iow1)/mre-11(ok179) progeny was indistinguishable from that of mre-11(iow1) progeny, in agreement with the dominance of the mre-11(iow1) allele over that of the null mre-11(ok179) allele (Table 1) (Fisher's exact test, P = 1). The embryonic lethality analysis and cytological examination both suggested that mre-11(iow1) is a hypomorph of mre-11 with defects downstream of meiotic DSB formation.

The defects observed in mre-11(iow1) mutants could be attributed to a reduction in the level of the MRE-11 protein. Alternatively, they could be due to specific perturbation of one of the activities of MRE-11, while retaining others (separation of function). To test whether protein levels of MRE-11 were affected by the mre-11(iow1) missense mutation, we performed Western blotting and immunohistochemical analysis using a C. elegans MRE-11-specific antibody. Western blot analysis of MRE-11 revealed an ∼90-kDa band in protein extracts from wild-type worms, as previously found (see Fig. S1 in the supplemental material) (44). This band was absent from the null mre-11(ok179) mutants but still present in similar levels in mre-11(iow1) mutants, indicating that the iow1 mutation does not affect the levels of the MRE-11 protein. The same antibody localized to germ line nuclei as multiple foci in wild-type worms (see Fig. S1D to F and J) (44). The nuclear localization, but not the cytoplasmic staining, was mostly absent in the mre-11(ok179) null mutant (see Fig. S1A to C and J), indicating that the nuclear localization is specific to MRE-11. MRE-11 nuclear foci were abundant in the germ line of mre-11(iow1) mutants (see Fig. S1G to J), consistent with the protein level of MRE-11 not being affected in mre-11(iow1) mutants. These data support the hypothesis that the meiotic defects of mre-11(iow1) mutants are due to perturbation of the resection activity of the MRE-11 protein and not due to the transcription, translation, and/or protein stability of MRE-11.

Recombination intermediates fail to form in mre-11(iow1) mutants.

The resection activity of Mre11 is required to resect meiotic DSBs, which generate ssDNA, promoting subsequent repair by the HR pathway to yield crossovers in yeast (10–13). If the C. elegans mre-11(iow1) mutants are proficient in DSB formation but not in repair due to the perturbation of resection activity, we would expect that spo-11 (required for DSB formation) would be epistatic to mre-11(iow1) (34, 40). In spo-11 null mutants, DSBs were not formed and no crossovers were present, resulting in 12 univalents (Fig. 1G) (38). A spo-11 mutation suppressed the aggregation and fragmentation phenotypes observed in mre-11(iow1) mutants; oocytes from mre-11(iow1); spo-11 double mutants displayed a majority of nuclei with 11 or 12 DAPI bodies (Fig. 1H and K). This result was not significantly different from that for spo-11 single mutants (two-tailed Mann-Whitney test, P = 0.48) but was significantly different from that for mre-11(iow1) mutants (two-tailed Mann-Whitney test, P < 0.0001). MSH-5 is a member of the ZMM proteins, which are needed to promote CO formation (45), and C. elegans MSH-5 functions after the formation of the RAD-51 nucleofilament but upstream of dHJ resolution (46). Diakinesis 1 oocytes of msh-5 mutants showed 12 univalents due to defects in CO formation, while they retained the ability to repair DSBs through other homology-dependent pathways (Fig. 1I and L) (46). As expected, the msh-5 mutation did not suppress the phenotype observed in mre-11(iow1) mutants. The number of DAPI bodies in oocytes of mre-11(iow1); msh-5 double mutants showed a wide range of distribution, from 3 to 11 (Fig. 1J and L), sharply contrasting with the 12 univalents in the msh-5 single mutants (Fig. 1I and L) (two-tailed Mann-Whitney test, P < 0.0001). Statistical analysis indicated a subtle but significant difference of DAPI body counts between the mre-11(iow1); msh-5 double mutants and the mre-11(iow1) mutants (two-tailed Mann-Whitney test, P = 0.037). This could be due to some residual repair of DSBs via the HR pathway of mre-11(iow1), which was eliminated by the msh-5 mutation. All results taken together indicate that mre-11(iow1) acts downstream of DSB formation and upstream of CO formation, as expected for a mutant defective in DSB resection.

C. elegans meiotic chromosomes lack strong DSB hot spots (47, 48), and therefore, no molecular assays are available to determine the levels of DSB formation and resection. However, cytological markers for proteins associated with DSBs are extensively used for this purpose (34, 40). RPA-1 loads onto the newly generated ssDNA concurrently with DSB resection. RPA-1 loading then facilitates the association of RAD-51 with the ssDNA (34), which forms a stable filament. To determine if ssDNA was generated in the germ line of mre-11(iow1) mutants, we measured the numbers of RPA-1 and RAD-51 foci in wild type, mre-11(iow1), and spo-11 mutants. Worm germ lines were divided into 7 zones (as described in reference 42) (Fig. 2I), and numbers of RPA-1 and RAD-51 foci were assessed in mitosis (zones 1 and 2), transition zone (zone 3), and pachytene (zone 4 to zone 7).

Fig 2.

Lack of RPA-1 foci in mre-11(iow1) mutants. (A to H) Micrographs of YFP-tagged RPA-1 foci in the nuclei of indicated stages and indicated genotypes. Images are projections through three-dimensional data stacks. RPA-1::YFP foci are shown in green, and DAPI-stained chromosomes are shown in blue. Bar, 2 μm. (I) Graph of a gonad, representing division of 7 zones from the premeotic tip to late pachytene. PM, premeiotic tip; TZ, transition zone; EP, early pachytene; MP, middle pachytene; LP, late pachytene. Quantitative time course analysis of RPA-1::YFP foci during meiotic prophase I. The position along the x axis corresponds to zones in the germ line.

To test the hypothesis that resection is greatly reduced in mre-11(iow1) mutants, we examined the loading of yellow fluorescent protein (YFP)-tagged RPA-1 onto meiotic chromosomes of the wild type and mre-11(iow1) mutants. In the wild-type germ line, we observed an increase in the number of RPA-1 foci starting in early pachytene, eventually peaking at late pachytene (Fig. 2C, D, and I) (34). However, we did not observe an increased number of RPA-1 foci in pachytene in the mre-11(iow1) mutant germ line (Fig. 2G, H, and I). This is consistent with the hypothesis that resection is greatly reduced in the mre-11(iow1) mutant germ line.

We observed that the number of RAD-51 foci in the wild-type germ line rose upon entrance into the transition zone (Fig. 3A and 4B; see also Table S1 in the supplemental material), peaked at early-middle pachytene (Fig. 3A′ and 4B; see also Table S1), and were mostly absent by late pachytene (Fig. 3A″ and 4B; see also Table S1), as previously demonstrated (42). The number of RAD-51 foci was low through the germ line of spo-11 mutants (Fig. 3B and 4C; see also Table S1) due to absence of meiotic DSBs (38). The number of RAD-51 foci in the mre-11(iow1) mutant germ line remained low from the transition zone to late pachytene, with most nuclei exhibiting no RAD-51 foci (Fig. 3D and 4E; see also Table S1), similar to what was observed in spo-11 controls.

Fig 3.

RAD-51 foci at meiotic prophase I in germ lines of different mutants. Immunofluoresence micrographs show RAD-51 foci of germ line nuclei of the indicated genotypes. PM, premeiotic tip; E/M P, early-middle pachytene; LP, late pachytene. Images are projections through three-dimensional data stacks. RAD-51 foci are shown in green, and DAPI-stained chromosomes are shown in blue. Bar, 2 μm.

Fig 4.

Quantitative time course analysis of RAD-51 foci at meiotic prophase I. (A) Graph of a gonad, representing division of the 7 zones from the premeotic tip to late pachytene. PM, premeiotic tip; TZ, transition zone; EP, early pachytene; MP, middle pachytene; LP, late pachytene. (B to J) Levels of RAD-51 foci are indicated by color coding, as indicated to the left of panel B, showing the quantification of RAD-51 foci in the germ line of the indicated genotypes. The positions along the x axes correspond to zones in the germ line. The percentage of nuclei falling into each color-coded category is indicated on the y axes.

Crossover formation is greatly reduced in the mre-11(iow1) mutant germ line.

COSA-1, a cyclin-related protein conserved in metazoans, is required for meiotic COs and localizes to the single CO site in each homolog in late prophase I (49). We tested whether CO formation was defective in the germ line of mre-11(iow1) mutants by examining the loading of green fluorescent protein (GFP)-tagged COSA-1 onto meiotic chromosomes in wild-type, mre-11(ok179), and mre-11(iow1) worms. The average number of COSA-1 foci for wild-type worm nuclei (from pachytene through diplotene) was 5.95 ± 0.04 (average ± standard error [SE]), with most nuclei showing 6 bright COSA-1 foci corresponding to the 6 CO sites per nucleus (Fig. 5A, A′, and F) (49). In the mre-11(ok179) null mutant germ line at the same stages, the average number of foci per nucleus was 0.56 ± 0.07 (average ± SE) (Fig. 5B, B′, and F). In mre-11(iow1) mutants, the number of average foci per nucleus was 0.58 ± 0.08 (Fig. 5C, C′, and F). These numbers were significantly lower than for the wild type (two-tailed Mann-Whitney test, P < 0.0001), but the results were indistinguishable from mre-11(ok179) mutants (two-tailed Mann-Whitney test, P = 1). This indicates that CO formation is severely reduced in mre-11(iow1) mutants.

Fig 5.

COSA-1 loading is reduced in the mre-11(iow1) mutant germ line but partially restored in mre-11(iow1); cKu-80 mutants. (A to E) Immunofluoresence micrographs of GFP::COSA-1 foci of germ line nuclei of the indicated genotypes. Images are projections through three-dimensional data stacks. GFP::COSA-1 foci are shown in green, and DAPI-stained chromosomes are shown in blue. Bar, 2 μm. (F) Quantitative analysis of GFP::COSA-1 foci observed in the different worms.

To examine if the residual COSA-1 foci in mre-11(iow1) mutants resulted from repair of a few DSBs via HR, we examined if the absence of MSH-5 could eliminate those COSA-1 foci in mre-11(iow1) mutants. The average number of COSA-1 foci per nucleus for mre-11(iow1); msh-5 mutants was 0.06 ± 0.02 (see Fig. S3C and D in the supplemental material), significantly lower than in the mre-11(iow1) mutants (0.56 ± 0.07) (Fig. 5C and F) (two-tailed Mann-Whitney test, P < 0.0001), but indistinguishable from msh-5 mutants (see Fig. S3A and D) (two-tailed Mann-Whitney test, P = 0.85). This elimination of the residual COSA-1 foci by the removal of MSH-5 in mre-11(iow1) mutants indicates that there is a small amount of HR-mediated repair in mre-11(iow1) mutants.

The absence of NHEJ partially suppresses the meiotic defects of mre-11(iow1) mutants.

The chromosome aggregation observed in the oocytes of mre-11(iow1) mutants could be explained by repair of unresected meiotic DSBs via NHEJ. In the C. elegans germ line, it has been shown that some meiotic DSBs are repaired via NHEJ when HR is defective (e.g., Cebrc-2 [34]). Therefore, we tested the hypothesis that DSB repair via NHEJ in the mre-11(iow1) mutant accounts for the chromosomal aggregates in this mutant by creating double mutants of mre-11; cKu-80. cKU-80 is the worm ortholog of Ku80 and is essential for NHEJ-mediated DSB repair (33). If some meiotic DSBs were acted on by NHEJ in mre-11(iow1) mutants, we would expect that removal of NHEJ by a cKu-80 mutation would suppress the chromosome aggregation in mre-11(iow1) mutants. The diakinesis 1 oocytes of cKu-80 mutants showed 6 DAPI bodies, indistinguishable from the wild type (Fig. 6A, C, and K). This argues that the C. elegans germ line with functional HR does not use NHEJ (33). The frequency of oocytes with aggregates in mre-11(iow1); cKu-80 double mutants was significantly lower than oocytes of mre-11(iow1) single mutants, and some DAPI bodies in the oocytes of the double mutants appeared to be bivalent (Fig. 6E and K) (Fisher's exact test, P < 0.01). These data suggest that removal of cKu-80 partially suppresses the chromosome aggregation contained by mre-11(iow1) mutants and supports our hypothesis that some chromosome aggregation in mre-11(iow1) mutants is due to the repair of meiotic DSBs via NHEJ.

Fig 6.

Chaismata are partially restored in mre-11(iow1); cKu-80 mutants but eliminated by depletion of RAD-51 and EXO-1. (A to E, M to O, and Q to S) DAPI-stained nuclei oocytes at diakinesis 1 of the indicated genotype. In the mre-11(iow1); cKu-80 double mutant (E), bivalent-like DAPI bodies were observed (arrows). (F to J) DAPI-stained nuclei oocytes at diakinesis 1 post-gamma irradiation of the indicated genotypes. Bars, 2 μm. (K, L, P, and T) Frequency distributions of DAPI-stained oocyte nuclei at diakinesis 1 under the indicated conditions.

The high level of embryonic lethality in mre-11(iow1) mutants is partially suppressed by the removal of NHEJ.

To examine in more detail the effect of NHEJ removal on mre-11(iow1) mutants, we compared embryonic lethality and the HIM rate in the double mutants, single mutants, and wild-type worms. Compared to the single mre-11(iow1) mutants, the double mutant mre-11(iow1); cKu-80 showed a significant decrease in embryonic lethality; 13.0%of the double mutant embryos laid hatched (Table 1) (Fisher's exact test, P < 0.0001), consistent with partial rescue in chromosomal aggregation in late prophase I. Among those hatched F1 progeny of mre-11(iow1); cKu-80 mutants, the percentage of males was significantly higher than that of wild-type and cKu-80 worms (Table 1) (Fisher's exact test, P < 0.0001), indicating defects in chromosome segregation.

The cKu-80 mutation reduces the high level of germ line apoptosis in mre-11(iow1) mutants.

In C. elegans, most recombination-defective mutants that exhibit chromosome aggregates and fragments in late prophase I also show a high level of apoptosis at late pachytene that is triggered by the DNA damage checkpoint (e.g., rad-51 and Cebrc-2 [34, 50]). Since mre-11(iow1) mutants showed chromosome aggregates and fragments at late prophase I, we investigated whether germ line apoptosis was elevated in mre-11(iow1) mutants by staining germ line nuclei undergoing apoptosis with acridine orange (46). We found a 2-fold increase in the level of apoptotic nuclei in late pachytene of mre-11(iow1) mutants compared to the wild-type worms (Fig. 7) (two-tailed Mann-Whitney test, P < 0.001). Since removal of NHEJ partially suppressed the chromosome aggregates in mre-11(iow1) mutants, we asked if removal of NHEJ would also suppress the high level of germ line apoptosis in mre-11(iow1) mutants. The levels of apoptosis in mre-11(iow1); cKu-80 double mutants (Fig. 7) were significantly lower than those of mre-11(iow1) single mutants (Fig. 7) (two-tailed Mann-Whitney test, P < 0.0001) but significantly higher than in wild-type and cKu-80 worms (Fig. 7) (two-tailed Mann-Whitney test, P < 0.001). The intermediate level of apoptosis in mre-11(iow1); cKu-80 double mutants was consistent with the partial rescue of chromosome morphology of oocytes at diakinesis in the absence of NHEJ.

Fig 7.

Germ line apoptosis is elevated in mre-11(iow1) mutants and is partially suppressed by blocking NHEJ. Acridine orange-stained apoptotic germ line nuclei were scored.

DSBs are repaired via the HR-mediated pathway in the mre-11(iow1); cKu-80 mutant germ line.

A possible explanation for the partial suppression of the mre-11(iow1) mutant phenotypes by the cKu-80 mutation is that, when NHEJ is not available, DSBs are repaired via the HR pathway. The HR pathway is completely dependent on RAD-51 in C. elegans (no DMC1 is present). We therefore examined if a rad-51 null mutation would eliminate bivalent formation in the mre-11; cKu-80 mutants. The mre-11(iow1); rad-51; cKu-80 triple mutants showed a high level of chromosomal fragmentation by DAPI bodies count (Fig. 6O and P), significantly different from mre-11(iow1); cKu-80 (Fig. 6E and K) (two-tailed Mann-Whitney test, P < 0.0001) and from rad-51;cKu-80 (Fig. 6N and P) (two-tailed Mann-Whitney test, P < 0.0001). This supports our hypothesis that repair of DSBs in mre-11; cKu-80 depends on HR.

To further analyze the repair of DSBs in mre-11(iow1); cKu-80, we analyzed RAD-51 localization in mre-11(iow1); cKu-80 double mutants. In agreement with our hypothesis, we observed RAD-51 loading onto chromosomes in the double mutants (Fig. 3E and 4F; see also Table S1 in the supplemental material). Moreover, RAD-51 loading was delayed in mre-11(iow1); cKu-80 worms compared to the wild-type and cKu-80 worms, in agreement with late activation of the HR pathway. RAD-51 foci did not load onto chromosomes in the germ line of mre-11(iow1); cKu-80 worms through early pachytene (Fig. 3E and E′ and 4F; see also Table S1). RAD-51 foci started to increase around middle pachytene and peaked at late pachytene (Fig. 3E″ and 4F; see also Fig. S4B and Table S1 in the supplemental material). The number of RAD-51 foci on germ line chromosomes in late pachytene of mre-11(iow1);cKu-80 worms was significantly higher than the foci at the corresponding stage of wild-type, mre-11(iow1), or cKu-80 worms (Fig. 3 and 4; see also Fig. S4 and Table S1) (pairwise comparison of foci in zone 7, two-tailed Mann-Whitney test, P < 0.0001). The loading of RAD-51 in late meiotic prophase I in mre-11(iow1); cKu-80 worms suggests DSBs are resected in mre-11(iow1) mutants when NHEJ is not available for their repair, albeit in later stages than observed in the wild-type worms.

To further demonstrate that DSBs are repaired by the HR pathway to give rise to COs in mre-11(iow1); cKu-80 double mutants, we examined the loading of GFP-tagged COSA-1 onto the chromosomes in the double mutants. The number of COSA-1 foci per nucleus from late pachytene through diplotene in the germ line of mre-11; cKu-80 double mutants was 4.58 ± 0.16 (average ± SE) (Fig. 5E, E′, and F). This COSA-1 foci level in mre-11; cKu-80 double mutants was significantly higher than that of mre-11(iow1) single mutants (Fig. 5C and F) (two-tailed Mann-Whitney test, P < 0.0001) but significantly lower than cKu-80 and wild-type worms (Fig. 5A, D, and F) (two-tailed Mann-Whitney test, P < 0.0001). The partial restoration in the COSA-1 loading in mre-11(iow1); cKu-80 double mutants suggests formation of COs in most nuclei and is consistent with the partial suppression of mre-11(iow1) by cKu-80 mutation for other phenotypes.

The repair of DSBs in mre-11(iow1); cKu-80 mutant germ line requires EXO-1.

The delayed loading of RAD-51 in the mre-11(iow1); cKu-80 mutant germ line can be interpreted as resection of meiotic DSBs by an alternative nuclease, which is active from middle to late prophase I. Exo1 and BLM (the orthologs in worms are EXO-1 and HIM-6) have been shown to be involved in resection activity in DSB ends in yeast (3, 4). To examine if Exo1 and BLM contribute to the repair of meiotic DSBs in mre-11(iow1);cKu-80 mutants, we constructed mre-11(iow1); exo-1; cKu-80 and mre-11(iow1); him-6; cKu-80 triple mutants.

Most diakinesis 1 oocytes of exo-1; cKu-80 double mutants showed 6 DAPI bodies (Fig. 6R and T), a result indistinguishable from the wild type (Fig. 6A and K) (two-tailed Mann-Whitney test, P = 0.87) and exo-1 worms (data not shown; two-tailed Mann-Whitney test, P = 0.64). The diakinesis 1 oocytes of mre-11(iow1); exo-1 mutants showed chromosomal aggregation and fragmentation (Fig. 6Q and T), similar to mre-11(iow1) mutants (Fig. 6D and K) (two-tailed Mann-Whitney test, P = 1). The majority of mre-11(iow1); exo-1; cKu-80 triple mutant oocytes showed 11 or 12 DAPI bodies (Fig. 6S and T), sharply contrasting with the mre-11; cKu-80 mutants (Fig. 6E and K) (two-tailed Mann-Whitney test, P < 0.0001). This indicates that the repair of DSBs in mre-11(iow1); cKu-80 via HR is greatly reduced in the absence of exo-1.

We also examined if the absence of EXO-1 could eliminate or reduce the RAD-51 foci observed from middle to late prophase I in mre-11(iow1); cKu-80 mutants. The germ line of exo-1 mutants exhibited wild-type loading patterns of RAD-51 (transition zone to late pachytene) (Fig. 3F and 4G; see also Table S1 in the supplemental material); however, it did not disappear in late pachytene (Fig. 3F″ and 4G; see also Table S1). This indicated defects in dissolution of RAD-51 in the exo-1 mutant germ line. The germ line of exo-1; cKu-80 mutants showed a similar RAD-51 loading pattern as exo-1 mutants (Fig. 3G and 4H; see also Table S1). The RAD-51 foci remained low throughout the germ line of mre-11(iow1); exo-1 mutants (Fig. 3H and 4I; see also Table S1), similar to mre-11(iow1) mutants (Fig. 3D and 4E; see Table S1). Removal of exo-1 in the mre-11(iow1); cKu-80 background eliminated RAD-51 loading through pachytene (Fig. 3E and I and 4F and J; see also Table S1) (pairwise comparison of foci in zone 6 and 7, two-tailed Mann-Whitney test, P < 0.0001). Overall, our data support a role for EXO-1 in the resection of DSBs from middle to late pachytene in mre-11; cKu-80 mutants. Interestingly, we observed RAD-51 foci in nuclei in diplotene to diakinesis in mre-11; exo-1; cKu-80 triple mutants (see Fig. S4C in the supplemental material), which indicates another resection activity independent of EXO-1 and MRE-11 in these very late stages of meiotic prophase I.

To examine the role of HIM-6 in resection, we quantitated the DAPI bodies and RAD-51 loading in mre-11(iow1); him-6; cKu-80 triple mutants. The majority of oocytes of mre-11(iow1); him-6;cKu-80 triple mutants showed 11 or 12 DAPI bodies (data not shown). However, the RAD-51 focus loading pattern of mre-11(iow1); him-6; cKu-80 was not different from that of mre-11; cKu-80 mutants (Fig. 3E and 4F; see also Fig. S2D and H and Table S1 in the supplemental material). This indicates him-6 does not contribute to resection in mre-11(iow1); cKu-80 mutants but has a role downstream of resection, as previously reported (51).

The repair of gamma irradiation-induced DSBs in mre-11(iow1) mutants depends on NHEJ and EXO-1.

Unlike programmed meiotic DSBs, DSBs introduced by ionizing irradiation (IR) are not bound by Spo11. IR-induced DSBs can be repaired via the HR pathway to yield COs in the C. elegans germ line (37). Previous studies showed that null mutations of the MRX/N complex genes confer a profound defect in the response of germ cells to IR-induced DNA damage (36, 37). To test whether it is the resection activity of mre-11 that is required in the IR-induced DNA damage response, we used gamma irradiation to examine the mre-11(iow1) mutants.

We examined chromosome morphology in diakinesis 1 oocytes 24 h post-gamma irradiation. Under these conditions, wild-type and cKu-80 worms showed 6 DAPI bodies, comparable to their nonirradiated counterparts (Fig. 6A, F, C, H, K, and L). Irradiated spo-11 mutants had 84%oocytes with 6 DAPI bodies (Fig. 6B, G, K, and L). This indicated that DSBs introduced under these experimental conditions were sufficient to restore chiasmata in spo-11 mutants. The occurrence of oocytes with aggregates in irradiated mre-11(iow1) mutants was significantly higher than in the nonirradiated counterparts (Fig. 6D, I, K, and L) (Fisher's exact test, P < 0.0001). The occurrence of oocytes with aggregates in irradiated mre-11(iow1) mutants was significantly reduced by the cKu-80 mutation (Fig. 6I, J, K, and L) (Fisher's exact test, P < 0.0001). Our data suggest that exogenous DSBs are repaired via NHEJ, resulting in severe chromosome aggregation in mre-11(iow1) mutants. This supports a requirement for functional MRE-11 to repair exogenous DSBs in the germ line. Interestingly, diakinesis oocytes of irradiated mre-11(iow1); cKu-80 double mutants did not show a significant difference in chromosomal morphology as measured by DAPI body counts compared to nonirradiated counterparts (Fig. 6E, J, K, and L) (two-tailed Mann-Whitney test, P = 0.97). This suggests efficient repair of exogenous DSBs in mre-11(iow1); cKu-80 double mutants by the HR pathway or an alternative pathway.

RAD-51 loading onto gamma-irradiated DSBs requires RAD-50 (and presumably MRE-11) in early to mid-meiotic prophase (transition zone to late pachytene), but it is dispensable for late prophase nuclei and mitotically dividing nuclei (37). Thus, the C. elegans germ line switches between RAD-50/MRE-11-independent and -dependent loading of RAD-51 as cells enter meiosis. This is speculated to be due to a requirement for the resection of DSBs by the MRX/N complex only in early to midprophase. To investigate if the resection activity of MRE-11 was required for RAD-51 loading at DSBs sites generated by IR, we analyzed RAD-51 localization in gamma-irradiated mre-11(iow1) mutants. Germ line nuclei from mitosis to meiotic prophase I in irradiated wild-type and spo-11 worms are capable of loading RAD-51 (Fig. 8A, B, F, and G; see also Table S1 in the supplemental material). In contrast with the wild-type and spo-11 worms, RAD-51 foci were absent from the transition zone to mid- to late pachytene in irradiated mre-11(iow1) worms; their loading was reduced in the premeiotic zone and in late pachytene (Fig. 8A, B, D, F, G, and H; see also Table S1). Similar numbers of RAD-51 foci were also observed in mre-11(ok179) null mutants (see Fig. S5A in the supplemental material). The failure to accumulate RAD-51 from the onset of meiosis to midpachytene in irradiated mre-11(iow1) worms indicates MRE-11 is required to resect DSBs at these stages, supporting previous findings (37). As expected, cKu-80 mutants were indistinguishable from the wild type in their response to IR, consistent with NHEJ being dispensable for germ line DSB repair in the presence of functional HR (see Fig. S5D). In agreement with our observation of the role of NHEJ in repair of meiotic DSBs under HR-defective conditions, irradiated mre-11(iow1); cKu-80 double mutants showed an increase in RAD-51 foci on exogenous DSBs in late prophase I compared to the irradiated mre-11(iow1) single mutants, but RAD-51 loading at earlier stages of meiosis was unaffected (Fig. 8D and H; see also Fig. S5E).

Fig 8.

RAD-51 foci at meiotic prophase I in germ lines following gamma irradiation. (A to E) Immunofluoresence micrographs of RAD-51 foci of germ line nuclei following irradiation. PM, premeiotic tip; E/M P, early-middle pachytene; LP, late pachytene. Images are projections through three-dimensional stacks. RAD-51 foci are shown in green, and DAPI-stained chromosomes are shown in blue. Bar, 2 μm. (F to J) Histograms showing quantification of RAD-51 foci in the germ line of the indicated genotypes. Positions along the x axes correspond to zones in the germ line. The percentage of nuclei falling into each color-coded category is indicated on the y axes.

As RAD-51 focus levels were reduced but not abrogated in the premeiotic zone and in late pachytene in mre-11(iow1) mutants, we hypothesized that the resection activity of MRE-11 is dispensable in mitosis and in late prophase I due to the redundant resection activities of EXO-1 or HIM-6. To examine whether EXO-1 or HIM-6 plays a role in resection of irradiation-induced DSBs in mre-11(iow1) mutants, we analyzed RAD-51 loading in irradiated mre-11(iow1); exo-1 and mre-11; him-6 double mutants. Compared to irradiated mre-11(iow1) mutants, irradiated mre-11(iow1); exo-1 worms showed greatly reduced RAD-51 foci in the premeiotic tip (Fig. 8D, E, H, and J; see also Table S1 in the supplemental material) (pairwise comparison of foci in zones 1 and 2, two-tailed Mann-Whitney test, P < 0.0001) and showed elimination of RAD-51 foci in late pachytene (Fig. 8D″, E″, H, and J; see also Table S1) (pairwise comparison of foci in zone 7, two-tailed Mann-Whitney test, P < 0.0001). Nevertheless, removal of HIM-6 in mre-11(iow1) mutants did not result in a reduction in the premeiotic tip and late pachytene (see Fig. S5B and C in the supplemental material). These results indicated that EXO-1 is responsible for the resection of irradiation-induced DSBs in late pachytene when MRE-11's resection activity is impaired.

DISCUSSION

A conserved function for the nuclease domain of MRE-11 in DSB resection.

Here we identified the first separation-of-function allele of the MRX/N complex in C. elegans and provided clear evidence of a role for MRE-11 in the resection of meiotic, SPO-11-induced DSBs in C. elegans. Several lines of evidence indicated that iow1 mutation affects resection rather than DSB formation: (i) mre-11(iow1) meiotic defects appeared downstream of DSB formation; (ii) spo-11 deletion was able to suppress the chromosomal abnormalities of mre-11(iow1) mutants; (iii) elimination of NHEJ in mre-11(iow1) mutants largely restored RAD-51 loading and CO formation. Meiotic DSBs cannot be properly repaired in the mre-11(iow1) mutant, and DSBs are likely not efficiently resected, as evidenced by defects in the loading of RPA-1 and RAD-51. The mre-11(iow1) mutant harbors a point mutation in the N-terminal phosphodiesterase domain, a domain found to be required for the nuclease activity of Mre11 in other species (10, 14–16). Specifically, the yeast “nuclease dead” separation-of-function allele Mre11-D16A/N mutation is in the same phosphodiesterase motif as the iow1 mutation (14), suggesting that the nuclease activity of MRE-11 is affected in C. elegans mre-11(iow1) mutants. While we were unable to directly test whether mre-11(iow1) nuclease activity was completely abrogated when we used biochemical approaches, our data clearly argue for a conserved function for MRE-11 in meiotic DSB resection in metazoan meiosis.

The mre-11(iow1) mutation is most similar to the S. cerevisiae Mre11-D16A and Mre11-D16N mutations, which exhibit compromised MRX/N complex formation (as well as DNA damage sensitivity and telomere length defects [41]). Therefore, it is possible that MRX/N activity in the mre-11(iow1) mutant is affected due to compromised complex stability. The observation that MRE-11 foci are bigger and preferentially enriched in the chromosome of transition zone nuclei in wild-type worms but smaller and evenly distributed between chromosome and cytosol in mre-11(iow1) mutants could be a manifestation of comprised MRX/N stability. Alternatively, this may be explained by disruption of DNA binding activity of the complex. Although we were not able to test complex stability directly in the mre-11(iow1) mutant in the present study, the evidence indicates that the MRX/N complex assembles in mre-11(iow1) mutants: meiotic DSBs (which in C. elegans are dependent on the MRX/N complex [36, 37]) are formed at normal levels in mre-11(iow1) mutants. Additionally, the level of the MRE-11 protein in mre-11(iow) mutants is similar to that found in the wild type, suggesting that there is no aberrant turnover of the mutant protein. Whether the mre-11(iow1) mutation affects complex stability or not, the biological outcome is a separation of function of the two key activities, DSB formation and resection, of the MRX/N complex.

DSB repair pathway choice in the C. elegans germ line.

NHEJ and HR are the two major pathways used to repair DSBs in most organisms (29). Results in S. cerevisiae and S. pombe have indicated that initiation of resection by MRX/N and Com1/Sae2 at mitotic DSB ends is critical to promote HR repair by dissociating KU from those ends (52–54). A recent report implied a role for COM-1 in antagonizing KU, preventing meiotic DSB repair via NHEJ in the C. elegans germ line (55). Prior to our work, it was unknown whether MRE-11 is required for antagonizing NHEJ in C. elegans. Our data suggest that MRE-11 is a critical component in committing repair of germ line DSBs to the HR pathway, as opposed to NHEJ, because the latter is used to repair both meiotic and irradiation-induced DSBs in the absence of a functional MRX/N. Repair of meiotic DSBs in mre-11(iow1) mutants via NHEJ indicated that SPO-11 is removed from DSB ends, because KU cannot act on DSB ends bound by SPO-11. The removal of SPO-11 from meiotic DSB ends in mre-11(iow1) mutants can be explained by a redundant resection activity of COM-1 (see Fig. S6 in the supplemental material). An alternative explanation would be that mre-11(iow1) has some residual nuclease activity that is sufficient for SPO-11 removal but not for generation of a sufficient amount of ssDNA that can be cytologically observed (RAD-51 and RPA-1 binding). Even though we do not yet possess the tools to test these two possibilities in C. elegans, our data still suggest a role for MRX/N in channeling repair of germ line DSBs to HR, which is critical for the maintenance of genomic integrity.

Alternative pathways for DSB resection in C. elegans.

Work in yeast has indicated that meiotic DSB resection is initiated by Mre11 and Com1/Sae2 to generate short ssDNA that can be extensively resected by two resection activities: those of EXO-1 and Blm1/Sgs1-Dna2 (3, 4, 10, 14–16, 19, 56, 57). In a wild-type background, EXO-1 is not required for CO formation (this study and reference 55). In the MRE-11/RAD-50 null mutants, meiotic DSBs and therefore COs don't form. Thus, the position of EXO-1 in the pathway cannot be assessed. Our separation-of-function allele provides such an opportunity and uncovers a role for EXO-1 in resecting meiotic DSBs when MRX/N is not functional and NHEJ is blocked. Similarly, in the absence of COM-1 and NHEJ, EXO-1-dependent HR repair can restore CO formation, indicating a role for EXO-1 in the resection of meiotic DSBs when COM-1 is absent (55). However, the effects of EXO-1 depletion are different in worms deficient in a functional MRX/N and of COM-1; the deletion of EXO-1 can abrogate a larger fraction of COs in the com-1; cKu-80 mutants compared to the mre-11(iow1); cKu-80 mutants. This implies that the resection activity of EXO-1 is independent of COM-1 but partially dependent on a functional MRX/N complex. Another interesting feature of EXO-1-dependent DSB resection is its different temporal activation (based on the RAD-51 focus kinetics) in backgrounds deficient in COM-1 (and NHEJ) and MRX/N (and NHEJ). In com-1; cKu-80 worms, RAD-51 foci start to appear during early pachytene, similar to the pattern that is observed in the wild type (55). Depletion of EXO-1 in com-1; cKu-80 mutants eliminates RAD-51 foci beginning in early pachytene. However, in mre-11(iow1); cKu-80 worms, RAD-51 foci do not start to appear until midpachytene. These results indicate that the resection activity of EXO-1 is dependent on a functional MRX/N only in early prophase I but not in later stages. We propose two explanations for this temporal dependence on MRX/N: (i) the impaired MRX/N resection activity reduces the amount of DNA substrates that EXO-1 can act on, and therefore EXO-1-dependent resection occurs slowly and RAD-51 does not load until midpachytene; or (ii) MRX/N can stimulate EXO-1's activity directly in early prophase I, independently from its effect on ssDNA formation. We prefer the second argument, because this is consistent with the physical interaction between Mre11 and Sgs1, and with a role for MRX/N in stimulating Sgs-Exo-1-mediated DSB end resection in vitro (56, 58, 59). Furthermore, when MRX/N is absent, EXO-1 still cannot act on free DSB ends produced by ionizing irradiation in early prophase I (zones 3 to 6 [this study and reference 37]). This supports a role for MRX/N in stimulating EXO-1 independently of its resection activity. Here, we propose that in C. elegans, MRX/N is required for meiotic DSB resection, for antagonizing KU proteins, and for efficient recruitment of EXO-1 in early pachytene (see Fig. S6 in the supplemental material).

We found no evidence for a role for HIM-6 in resection, downstream of MRX/N, in C. elegans meiosis. This contrasts with the observation that human MRN enhances BLM-DNA2 resection activity in vitro (56). These findings may indicate a different role for HIM-6 in C. elegans meiosis (51) compared to that in other organisms, or it may be that HIM-6's action is masked by the redundant resection activity of EXO-1.

Overall, the mre-11(iow1) mutant provides unique insight into how defects in DSB resection affect events in meiosis and created an opportunity to disentangle various redundant pathways regulating the repair of meiotic DSB in C. elegans.