Damage-induced phosphorylation of BRC-1/BRD-1 in meiosis preserves germline integrity

Abstract

Multiple DNA repair pathways have evolved to safeguard genome integrity and ensure organismal viability in the face of DNA damage. Errors in DNA repair processes in meiosis can lead to aneuploidy and developmental defects, but the processes that protect the germline from DNA damage remain poorly understood. Here we report a DNA damage-induced phosphorylation of the BRC-1/BRD-1 heterodimer that is essential for germline integrity in Caenorhabditis elegans. Failure to phosphorylate BRC-1/BRD-1 in response to DNA damage results in meiotic double-strand breaks (DSBs) accumulation, chromosome breakage, catastrophic diakinesis, and loss of fecundity. We further show that these defects are driven by the activity of C. elegans Bloom and Mus81, which catalyze Holliday junction dissolution and resolution, respectively. Hence, we propose that phosphorylation of BRC-1/BRD-1 in response to ionizing radiation-induced DSBs constitutes a key regulatory step that ensures the proper resolution of recombination intermediates required to preserve germline integrity.

Graphical Abstract

Graphical Abstract.

Introduction

The tumour suppressor proteins BRCA1 (Breast Cancer 1) and BARD1 (BRCA1-associated RING domain protein 1) perform roles in replication fork protection, checkpoint signalling, and DNA repair by homologous recombination (HR) [1]. Mutations in BRCA1 and BRCA2 have been linked to an increased lifetime risk of developing certain types of cancer, including breast, ovarian, and prostate cancers in humans [2, 3]. A body of evidence has implicated BRCA1 in regulating the resection of DNA double-strand breaks (DSBs), whereas BRCA2 facilitates the recruitment and initial nucleation of the Rad51 recombinase onto processed DSBs. During meiosis, programmed DSBs are generated to initiate HR, which is required to promote the formation of inter-homolog crossovers (COs) needed for faithful chromosome segregation. These joint DNA molecules must disengage in order to segregate, and this is achieved by redundant endonucleases (called resolvases) and the BTR complex [4]. In Caenorhabditis elegans, MUS-81 and SLX-1 have overlapping roles with the Bloom syndrome helicase ortholog, HIM-6, in processing recombination intermediates [5–7]. The C. elegans Bloom ortholog HIM-6 also suppresses heterologous recombination in the germline, which could lead to translocations and genome rearrangements [8].

Studies in mice have shown that hypomorphic mutations or deficiencies in BRCA1 lead to defects in meiotic recombination, resulting in chromosome abnormalities and impaired fertility [9]. In C. elegans, BRC-1 and BRD-1 have also been implicated in meiosis and DNA repair during meiotic recombination. The BRC-1/BRD-1 complex localizes to the synaptonemal complex (SC) in an interdependent manner [10, 11]. During meiosis, the SC holds homologous chromosomes together, serving as a scaffold for HR and ensuring accurate chromosome segregation [12]. BRC-1 and BRD-1 are dispensable for meiotic CO formation but are required for meiotic homolog-independent DSB repair [13, 14]. Under defective meiosis, BRC-1 and BRD-1 mediate RAD-51 filament stability and CO formation [10, 11, 13]. Moreover, it has been shown that BRC-1 and BRD-1 prevent recombination between heterologous templates and repress error-prone DSB repair through non-homologous end joining (NHEJ) and polymerase theta-mediated end joining (TMEJ) [15, 16]. The mechanisms that regulate BRC-1 and BRD-1 activation after exogenous DNA damage in meiosis remain unclear.

The DNA-damage-activated kinases ATM (ataxia-telangiectasia-mutated) and ATR (ataxia-telangiectasia-related) play central roles in DSB sensing and repair in mitotic cells [17–19]. ATM and ATR kinases also localize to meiotic chromosomes and have been implicated in promoting HR, repair-template choice, and CO control [20]. In C. elegans the ATR kinase, ATL-1, is essential for mitotic cell cycle arrest and induction of apoptosis in response to DNA damage but shows no obvious meiotic defects in SC assembly [21]. Conversely, C. elegans ATM, ATM-1, is necessary for the restoration of chromatin and re-synapsis of the chromosome axes after irradiation [22]. ATM-1 seems to bias the choice of repair template to the homologous chromosome, with both checkpoint kinases acting redundantly to promote RAD-51 accumulation at DSBs [23]. Recently, both kinases have been shown to function at meiotic entry to reshape the genome through cohesins to promote interhomolog interactions and meiotic recombination [24]. Specifically, ATM and ATR phosphorylate components of the SC to maintain architectural integrity, promoting DSB repair pathway choice to avoid NHEJ [25, 26].

Here we investigated how exogenous DNA damage is repaired in the meiotic germline of C. elegans. We present evidence that ionizing radiation (IR)-induced DSBs result in IR-dependent phosphorylation of the BRC-1 and BRD-1 heterodimer, which regulates the correct resolution of recombination intermediates. We identify a cluster of S/T-Q motifs that form DNA damage-induced phosphorylation sites in the BRC-1 and BRD-1 proteins. The corresponding non-phosphorylatable (BRC-1^4A and BRD-1^3A) mutants exhibit IR sensitivity and defects in DSB repair, which are exacerbated by exogenous DNA damage, demonstrating that failure to phosphorylate the heterodimer BRC-1/BRD-1 impairs proper DSB repair. Surprisingly, we show that phosphorylation is dispensable for inter-homolog DSB repair but is required to avoid inappropriate processing of recombination intermediates by Bloom and Mus81. Hence, our data reveal a mechanism by which the meiotic checkpoint acts to protect the germline from exogenous DNA damage and genetic instability.

Materials and methods

Experimental model maintenance and in vivo assays

Strains and maintenance

Standard methods were used for the maintenance and manipulation of C. elegans strains [27, 28]. Nematode strains were provided by the Caenorhabditis Genetics Center, which is funded by the NIH National Center for Research Resources, except for 1960 ollas strain, polq-1,mus-81, and him-6 knockout strains, generated and/or kindly provided by Verena Jantsch group [11]. The strains with transgenic brc-1^(4A), brd-1^(3A), and polq-1;brc-1^(4A) alleles were generated using CRISPR/Cas9 [29, 30] by microinjection into N2 (WT). The double mutants brc-1^(4A);GFP::msh-5,brd-1^(3A);GFP::msh-5,mus-81;brc-1^(4A),him-6;brc-1^(4A),msh-5;brc-1^(4A), and syp-1;brc-1^(4A) were generated by crossing the corresponding strains. All strains are listed in Supplementary Table S1.

Embryonic lethality

Embryonic lethality was scored by comparing the number of eggs that hatch to produce viable progeny versus the total number of eggs laid. Briefly, L4 hermaphrodites grown at 20°C were individually plated. The animals were transferred to new plates once every 24 h until the egg laying stopped. Eggs laid were immediately counted. When each brood reached adulthood, the total number of live animals per brood was counted and checked against the egg count to give the total brood size and an estimate of the embryonic lethality frequency. The number of arrested larval and male progeny animals was also noted. A minimum of three experiments were performed for each strain. The total number of single hermaphrodites analysed is indicated in Table 1.

Table 1.

Viability analysis of brc-1^(4A)and brd-1^(3A)mutant alleles

Genotype	Average brood + SD (n)a	Percentage viable embryos (n)b	Percentage male (n)c
N2 (WT)	292.73 + 16.05 (19)	98.69 (5649)	0.04 (5598)
brc-1 (4A)	297.33 + 12.39 (11)	100 (3262)	0.03 (3262)
brd-1 (3A)	290.5 + 25.27 (8)	99.37 (2353)	0 (2338)

^aParentheses indicate the total number of singled hermaphrodites for wich entire brood size were scored.

^bParentheses indicate the total number of fertilized eggs scored.

^cParentheses indicate the total number of adults scored.

For brood analysis after irradiation, L4 animals were exposed to the indicated Gy doses of γ-ray from BioBeam8000. After 24 or 48 h, five post-irradiation P0 worms were plated to lay eggs for 10 h. Thirty-six hours later the number of hatched F1 larvae, dead embryos, and males were counted [31]. Three plates were counted for each strain and condition, and the experiment was repeated four times.

Generation of nematode strains

Generation of brd-1 non-phosphorylable mutant and polq-1;brc-1^(4A) double mutant by CRISPR/Cas9 genome editing

The generation of brc-1^(4A) mutant was ordered to SunyBiotech. brd-1^(3A) mutant worms were generated in a two-step genome edition by CRISPR/Cas9 as in [30]. N2 was used for injection and dpy-10 co-edition was used as a positive control of Cas9 activity. Injection mixes contained Cas9 (250 ng/ml), ALT-R tracrRNA (141.97 ng/ml), ALT-R dpy-10 crRNA (14.39 ng/ml), ALT-R target gene crRNA (59 ng/ml), ssODN dpy-10 (cn64) repair template (28 ng/ml), ssODN target gene repair template (116 ng/ml), and nuclease-free H₂O. polq-1;brc-1^(4A) mutant worms were also generated by CRISPR/Cas9. In this case, brc-1^(4A) was used for injection, and two different ALT-R target gene crRNAs were used in the injection mix. General reagents were acquired from IDT and are listed in Supplementary Table S2. The resulting transformants (roller or dumpy phenotype worms) were transferred to new plates and genotyped. All primers and RT are listed in Supplementary Table S3.

Worm genotyping

The resulting transformants were checked by single-worm PCR using MyTaq DNA polymerase and restriction enzyme digestion. Genomic DNA was obtained by single worm lysis and used in PCRs. Restriction enzyme digestion was used to check for the integration of the brd-1^(3A) repair template (BioLabs restriction enzymes PstI and PvuII are listed in Supplementary Table S2). Homozygote phospho-allele candidates were sequenced to check the integration of the expected mutations and resultant strains were backcrossed twice.

The double mutant strains were also checked by single worm PCR (primers are listed in Supplementary Table S3). To check for brc-1^(4A) and brd-1^(3A) alleles in gfp::msh-5 homozygote candidates, a fragment of the respective gene was amplified by PCR and sequenced. The presence of gfp::msh-5 allele was determined by taking advantage of the fluorescence the mutation generates. To check for brc-1^(4A) allele in polq-1;brc-1^(4A),mus-81;brc-1^(4A),him-6;brc-1^(4A),msh-5;brc-1^(4A), and syp-1;brc-1^(4A), specific primers for brc-1^(4A) and brc-1 wild-type allele amplification were designed. Restriction enzyme digestion was used to check for the presence of msh-5 and syp-1 knockout alleles (BioLabs restriction enzymes Hyp188I and BstAPI are listed in Supplementary Table S2). The primers used for all strains of genotyping are listed in Supplementary Table S3.

Molecular biology

sgRNA guides design

To design the sgRNA recognition sites and the repair templates for brd-1^(3A) mutant, we followed [29] protocol. To select PAM sites, we used CRISPOR website (http://crispor.tefor.net/). The designed sRNA guides are indicated in Supplementary Table S3.

Peptide arrays and kinase assays

For the peptide array studies, 18-mer peptides were made by solid‐phase synthesis and purified by high‐performance liquid chromatography, and their sequences were verified by mass spectroscopy. The 18-mer peptides were juxtaposed by three amino acids until scanning the complete BRC-1 and BRD-1 proteins. All peptides contained an N‐terminal biotin group with an aminohexanoic spacer to be spotted onto cellulose membrane. The membrane was activated by soaking in methanol for 2 min and washed twice with kinase buffer supplemented with 3% bovine serum albumin (BSA). In vitro phosphorylation was performed by incubating the membrane in 5 ml of kinase buffer supplemented with N2 worm extracts (protein concentration of 10 mg/ml) and 100 μCi of [32P] γ-ATP. After adding stop buffer, the membrane was washed sequentially in 1 M NaCl, then 1% sodium dodecyl sulphate, and finally 0.5% phosphoric acid solution. After washing in 96% ethanol, the membrane was dried and exposed to autoradiography film.

Gene knockdown

For gene knockdown assays, RNAi depletion by the feeding method was performed as described in [32], with modifications. HT115 bacteria containing the pL4400 vector or the corresponding RNAi feeding construct were seeded on LB plates containing ampicillin and tetracycline and incubated overnight at 37°C. A bacteria colony was suspended in 2 ml of LB, containing ampicillin, and incubated overnight at 37°C. The bacteria inoculum was seeded in NGM six-well plates with ampicillin, tetracycline, and 6 mM isopropylthio-β-d-galactosidase and incubated at RT for 36 h. Ten late L4 worms for each strain and condition were transferred to wells and irradiated at 75 Gy. Worms were incubated for 48 h at 20°C before DAPI staining of the germlines.

RNA extraction and RT-qPCR

N2 worms were treated as described earlier, recollected with H₂O-DEPC, and freeze-thawed three times in dry ice before being broken with a tissue grinder. RNeasy mini kit (QIAGEN) was used for RNA extraction. The expression of the target genes was determined by reverse transcription quantitative polymerase chain reaction (RT-qPCR), and the 2-ΔΔCt method was employed to calculate the relative mRNA expression levels of the genes. QuantiTect Reverse Transcription Kit (QIAGEN) and SYBR Green Master Mix reagent were used to perform the RT-qPCR. The primers used for the target genes are presented in Supplementary Table S3.

Cellular biology

Immunostaining

For RAD-51 and SYP-1 immunostaining, worms were treated as described in [33], with modifications. One day post-L4 adult gonads were dissected in 0.1% TBSTw (1 TBS, 0.1% Tween 20) on a Superfrost Plus slide (VWR) and were fixed for 5 min in 1% paraformaldehyde. Gonads were then flash frozen in liquid N₂ and the coverslip was removed. The slide was placed in −20°C MeOH for 10 min and washed three times in 0.1% TBSTw for 10 min before being placed in block (1 TBS, 0.1% Tween 20, 0.7% BSA) for 1 h. Slides were incubated overnight at 4°C in a dark humidifying chamber with diluted primary antibody to stain. Dilutions used were rabbit α-RAD-51 (1:5000) and guinea pig α-SYP-1 (1:500) in 0.1% TBSTw. Next day gonads were rinsed and then washed three times in 0.1% TBSTw, each for 10 min at RT, and incubated for 2 h with the secondary antibody in 0.1% TBSTw (α-RABBIT 1:200, α-GUINEA PIG 1:500) in a dark humidifying chamber. Gonads were rinsed and 1 μg/ml DAPI in 0.1% TBSTw was added to each slide and incubated for 10 min. Slides were washed four times in 0.1% TBSTw for 10 min and mounted with Vectashield. They were maintained at 4°C for 2–3 days prior to imaging. For irradiation experiments, one-day post-L4 adults were irradiated with 75 Gy dose of γ-rays from BioBeam8000, and 48 h post-irradiation gonads were dissected for immunostaining as described.

For BRC-1 and BRD-1 antibody worms were treated as described in [25], with modifications. Four hours after irradiation, worms were dissected in PBS on poly lysine slides, fixed for 15 min in 4% paraformaldehyde, and replaced for 5 min in TBSBTx (TBSB + 0.1% TX100). The slides were washed twice for 10 min with 0.1%TBSTw and one more time for 30 min with TBSB (TBS + 0.5% BSA). They were incubated overnight at 4°C with rabbit α-BRC-1 or rabbit α-BRD-1 (1:100) and chicken α-SYP-1 (1:300) antibody dilution. Next day gonads were rinsed and then washed three times in 0.1% TBSTw, each for 10 min at RT, and incubated for 2 h with the secondary antibodies (α-RABBIT 1:500 and α-CHICKEN 1:500). Gonads were rinsed and then washed three times for 10 min with 2 mg/ml DAPI in 0.1% TBSTw and mounted with 10 ml Vectashield (with 1 mg/ml DAPI) per sample for further analysis.

SYTO12 for apoptotic corpses quantification

For apoptotic corpses (AP) analysis, 24 h post-L4 animals were incubated in the dark with a 40 mM aqueous solution of the genotoxic agent SYTO-12. After 4 h, to facilitate the visualization of apoptotic corpses, worms were transferred to new dishes with OP50 bacteria and were incubated for 1 h at 25°C to metabolize the excess of SYTO-12 present in the intestine. After the indicated time, worms were transferred to slides with agarose pads to score; under the microscope, the presence of fluorescent bodies indicated cells undergoing apoptosis [31]. The experiment was repeated a minimum of three times and a minimum of 30 total worms for each strain were scored. For irradiation experiments, L4 worms were irradiated with 75 Gy dose of γ-ray from BioBeam8000, and 12-, 24-, and 36-h post-irradiation gonads were treated as described earlier.

GFP::MSH-5 foci quantification

One day post-L4 adults were irradiated with 75 Gy dose of γ-ray from BioBeam8000. Forty-eight hours post-irradiation, GFP::MSH-5 foci were detected as described in [34], with modifications. For GFP::MSH-5 detection, worms were dissected in 0.1% TBSTw and directly frozen in liquid nitrogen. After freeze-cracking, slides were incubated in methanol at −20°C for 5 min. Gonads were immediately washed with 0.1% TBSTw for 5 min and fixed with 4% PFA in 100 mM K₂HPO₄. Slides were washed three times in 0.1% TBSTw for 10 min and incubated with g/ml DAPI in 0.1% TBSTw for 10 min. Slides were mounted with Vectashield and germlines were examined with fluorescence microscopy. Ten to twelve nuclei were counted for late-pachytene region for a minimum of four germlines per genotype and condition.

Diakinesis nuclei quantification

For visualization and quantification of diakinesis nuclei, germlines were dissected and stained with DAPI as described for GFP::MSH-5 detection, except that paraformaldehyde was used at 1% and fixation was performed before liquid nitrogen freeze-cracking. DAPI bodies from a minimum of 2–4 diakinesis nuclei were counted or morphologically checked for 10–15 germlines per genotype and condition.

Fluorescence microscopy

Leica DM6000B was used to examine the germlines with 100X HCXPL-APO/1.40 OIL lens, and images were captured using Leica LAS-AF computer software for RAD-51 and SYP-1 immunostaining and GFP::MSH-5 foci quantification. Between 150 and 220 confocal planes of 0.2 μm distance were taken for each gonad (depending on the thickness of the gonad), and non-saturated laser conditions were adjusted for each experiment. Analysis was performed for half of the germline along the dorsal-ventral axis, using the maximum projection for RAD-51 quantification.

Nikon SMZ-645 was used to examine the germlines with 60X PL-APO/1.45 OIL lens for apoptotic corpse analysis.

For BRC-1 and BRD-1 immunostaining and diakinesis nuclei quantification, Zeiss AxioImager.2 was used to examine the germlines with Plan-Apochromat 63× and 100×/1.40 oil lenses, and images were captured using Zeiss Zen2 computer software. Between 70 and 130 confocal planes of 0.2 μm distance were taken for each gonad (depending on the thickness of half the gonad or diakinesis stage), and non-saturated laser conditions were adjusted for each experiment.

RAD-51 foci quantification

Analysis of RAD-51 foci was performed as described in [33]. Each germline was divided into six regions, corresponding to mitosis division zone, transition zone, early pachytene, medium pachytene, and late pachytene zones, and diplotene zone. The number of foci per nucleus in each region of the germline was quantified. A minimum of four germlines per genotype and condition were analysed. Data show the % of nuclei of the different categories based on the number of foci/nuclei. A minimum of four germlines per genotype were analysed.

Quantification and statistical analysis

Quantification was always performed in raw images. After quantification, for beauty purposes in the images shown a background subtraction plugging was applied using Fiji software. Statistical significance was determined with unpaired t-test or 2-way ANOVA using PRISM software (Graphpad Software Inc.). Specific replicate numbers (n) for each experiment can be found in the corresponding figure legends. In all figures, means are plotted, and standard deviation (SD) or standard error of the mean (SEM) is represented as error bars.

Results

DNA damage phosphorylation sites in BRC-1 and BRD-1

We have previously described a DNA damage-induced phosphorylation of the SC component SYP-1 that channels the repair of excessive DSBs through the activity of BRC-1 [25]. Given the importance of BRCA1 in regulating pathway choice in mitotic cells, we considered the possibility that phosphorylation of BRC-1 and/or BRD-1 may play an analogous regulatory role during meiotic HR repair. To explore this possibility, we examined damage-induced phosphorylation events on BRC-1 and BRD-1 using peptide arrays against both proteins. Protein extracts from untreated worms or worms treated with 75 Gy of IR were used for in vitro kinase assays, which revealed DNA damage-induced phosphorylation events on both BRC-1 and BRD-1. Notably, even in the untreated controls, several peptides displayed detectable signals, most containing consensus motifs for kinases. Clearer in the BRD-1 peptide array. After irradiation, some of these pre-existing sites showed heightened signal intensity, and, importantly, additional peptides (absent in the non-irradiated samples) became phosphorylated, indicating irradiation-specific kinase activity.

Specifically, we searched for peptides with irradiation-specific signals that harbour the Ser/Thr-Gln (S/T-Q) motif, the canonical phosphorylation site recognized by both ATR and ATM kinases. We identified an SQ motif surrounded by several serine and threonine residues between amino acids 436 and 443 within the BRCT domain of BRC-1 (Fig. 1A). In the case of BRD-1, we identified a cluster of three TQ motifs within 104–132 amino acids adjacent to the RING domain (Fig. 1B). Both BRCT and RING domains of the worm proteins are highly conserved with human and Xenopus orthologs BRCA1 and BARD1 [35]. These results demonstrate that these residues are modified under in vitro conditions.

Figure 1.

In vitro DNA damage phosphorylation of BRC-1 and BRD-1 proteins. (A) Left, the BRC-1 peptide array with N2 (WT) protein extracts without DNA damage (top) and with N2 (WT) extracts after 75 Gy (bottom). Each of the spots represents an 18-mer peptide fragment juxtaposed by three amino acids (aa) scanning the complete BRC-1 protein. Positive serial spots (detected by autoradiography) corresponding to the specific DNA damage-phosphorylated region are boxed in red. Scheme depicts the phosphorylation site established by the peptide array data, with the possible phosphorylation residues highlighted in red. Right, the predicted structure of BRC-1 BRCT domain was determined by AlphaFold. In red, the putative phosphorylated S/T-Q motif for BRC-1 (S438). (B) Left, the BRD-1 peptide array with N2 (WT) protein extracts without DNA damage (top) and with N2 (WT) extracts after 75 Gy (bottom). Each of the spots represents an 18-mer peptide fragment juxtaposed by three amino acids (aa) scanning the complete BRD-1 protein. Positive serial spots (detected by autoradiography) corresponding to the specific DNA damage-phosphorylated region are boxed in red. Scheme depicts the phosphorylation site established by the peptide array data, with the possible phosphorylation residues highlighted in red. Right, the AlphaFold predicted structure of BRD-1 showing the RING domain (yellow). In red, the putative phosphorylated residues for BRD-1 (T104, T111, and T131).

To examine the structural context of the phospho-motifs, we used ChimeraX to predict the structure of C. elegans BRC-1 and BRD-1 BRCT and RING domains, respectively [36]. We used the predicted structures determined by AlphaFold of BRC-1 BRCT domain (Fig. 1A) and BRD-1, focusing on the RING domain (yellow) and its continuous region (grey), where the motifs of interest are located (Fig. 1B). The predicted conformations show that all phospho-motifs are solvent exposed on the surface of the predicted protein structure.

To investigate the biological relevance of the putative damage-induced phosphorylation sites in BRC-1 and BRD-1 proteins, we used CRISPR–Cas9 to generate non-phosphorylatable alleles in which the putative phosphorylated residues were changed to alanine. For BRC-1, we generated the S438A mutation in addition to S436A, S441A, and T442A mutations to ensure there were no possibilities for compensatory phosphorylation. This resulted in the brc-1^4A allele. For BRD-1 we generated the brd-1^3A in which we changed each of the threonines (T104, T111, and T131) to alanine. To eliminate possible off-targets of Cas9, the resulting transgenic lines were then back-crossed with the wild-type [N2 (WT)] and brd-1 knockout, respectively. We first analysed the effect of these mutations on worm development by performing viability assays. Similar to what has been described for the loss of brc-1 or brd-1 [13, 35], both mutated alleles are viable and do not show overt differences in brood size with respect to the wild-type (Table 1). These data indicate that disruption of BRC-1 and BRD-1 putative phosphorylation does not negatively impact development.

The lack of BRC-1 and BRD-1 phosphorylation results in mild meiotic DSB repair defects

In previous studies, it was shown that the loss of brc-1 or brd-1 leads to a mild meiotic defect, supported, amongst other observations, by the appearance of males, which results from loss of an X-chromosome [13, 35]. During the viability assays we did not observe an increase of males in the progeny of the brc-1^4A and brd-1^3A mutants (Table 1). Nevertheless, we also performed a cellular analysis of the germline, which allows for temporal and spatial analyses of meiotic progression through prophase I [37], with markers of the key steps in meiosis. Homologous chromosome synapsis can be studied by immunostaining of the SC central region protein SYP-1 [38]; using an antibody against SYP-1, we observed that its localization between paired homologous chromosomes in the brc-1^4A and brd-1^3A mutant strains is indistinguishable from the wild-type (Supplementary Fig. S1A), indicating that synapsis is unaffected in these mutants.

Next, we studied programmed meiotic DSB repair using an antibody against RAD-51, which catalyzes the strand invasion and exchange steps during HR [39]. During normal meiosis in wild-type worms, RAD-51 foci are observed at sites of SPO-11-induced meiotic DSBs, with RAD-51 foci first appearing in the transition zone and progressively increasing to a maximum number in mid-pachytene and finally disappearing in late pachytene (Fig. 2A and B) [38]. In the case of brc-1^4A and brd-1^3A mutant germlines, we observed that RAD-51 foci start to appear in the transition zone but are slightly increased when compared to wild-type germlines (Fig. 2A and B and Supplementary Fig. S1B). This was more prominent in brc-1^4A, where at later regions we observed 27% of nuclei with persistent RAD-51 foci at unrepaired DSBs. Since the accumulation of unrepaired DNA damage leads to apoptosis, we also scored germ cell apoptosis and found that apoptotic corpses were significantly increased in brc-1^4A and brd-1^3A mutant strains when compared to wild-type (Fig. 2C). These data suggest that the inability to phosphorylate BRC-1 or BRD-1 leads to mild defects in the repair of programmed meiotic DSBs.

Figure 2.

brc-1 and brd-1 phospho-mutants show mild defects in meiosis DSB repair. (A) Representative images of mid-pachytene nuclei (region 4) stained with anti-RAD-51 (red) and DAPI (blue) for the indicated genotypes. Scale bar, 5 μm. (B) Quantification of recombination marker RAD-51 foci in the indicated strains in normal conditions. A minimum of four gonads from four independent experiments were analysed and 70–80 nuclei were scored for each region per gonad. The top diagram of a hermaphrodite gonad indicates the regions in which the number of RAD-51 foci was scored: (i) mitotic division zone; (ii) transition zone; (iii) early pachytene; (iv) mid-pachytene; (v) late pachytene; and (vi) diplotene. Data are represented as mean ± SD. (C) Quantification of apoptotic corpses stained with SYTO-12 at late pachytene stage in the indicated strains in normal conditions. A minimum of 30 gonads from 4 independent experiments were analysed. Data are represented as mean ± SD and P-values for unpaired t-tests are indicated. (D) Representative images of mid-pachytene stage nuclei stained with the antibody against SC protein SYP-1 (red), BRC-1 (green on top), or BRD-1 (green on bottom) and counterstained with DAPI for the indicated genotypes. Scale bar, 3 μm. (E) Quantification of the number of DAPI-stained bodies in the diakinetic oocytes of the indicated strains in normal conditions. Light grey represents 6 DAPI bodies, dark grey represents 7–11 DAPI bodies, and orange represents 12 DAPI bodies. −4 to −1 oocytes from a minimum of 10 gonads from 3 independent experiments were analysed for each strain and condition. Data are represented as mean of the different experiments ± SEM. A 2-way ANOVA with Tukey’s multiple comparisons test was performed. No significant P-values were obtained.

Finally, it has been shown that both proteins BRC-1 and BRD-1 are dependent on each other for mutual stability and co-localize on the SC during prophase I in C. elegans germline [11]. Since BRC-1/BRD-1 localization is essential for DSB repair upon induction of exogenous damage during gametogenesis, we investigated the localization of both proteins in our mutant alleles. Detection was performed by employing previously characterized antibodies against BRC-1 and BRD-1 [40]. Immunofluorescences with αBRC-1 in brc-1^4A germlines or αBRD-1 in brd-1^3A germlines showed that both proteins localize at meiotic chromosomes with a pattern similar to that of the SC (Fig. 2D), as previously described [11]. To further prove that our alanine mutants are not behaving as a null mutation, we generated double mutants of brc-1^4A with syp-1 and msh-5. Loss of SYP-1 generates univalents because homologous chromosomes fail to remain synapsed, thereby disrupting HR and eliminating CO formation [41]. Likewise, msh-5 mutants are CO-deficient, as MSH-5 is an essential component of the pro-CO machinery [42]. It has been shown that loss of brc-1 in backgrounds where CO formation is prevented leads to chromosome fragmentation at diakinesis [13]. In contrast, our double mutants show the expected 12 univalents at diakinesis (Fig. 2E). Since the non-phosphorylatable alleles do not mimic these null mutants’ phenotypes, the observed defects likely result from loss of BRC-1 phosphorylation rather than complete loss of function.

Loss of BRC-1 and BRD-1 phosphorylation leads to IR sensitivity

Since we observed in vitro phosphorylation of BRC-1 and BRD-1 under DNA damage conditions, we next analysed whether the phosphorylation status of BRC-1 or BRD-1 impacts the ability of nematodes to respond to exogenous DNA damage induced by IR. First, we determined IR sensitivity by scoring survival of the resulting F1 progeny 24–36 h after irradiation of L4 stage hermaphrodites with different doses [31]. Irradiation with low doses resulted in a slight decrease in survival of 22% in wild-type, 31% in brc-1^4A mutant allele, and 40% in brd-1^3A allele (Fig. 3A). However, at higher doses (75 Gy and 100 Gy), both mutant alleles exhibited reduced survival relative to the wild-type strain (Fig. 3A). Loss of brc-1 or brd-1 has been reported to confer marked sensitivity to DNA damage [35]. To determine whether our point mutants mimic this null phenotype in the DNA-damage response, we incorporated brc-1- and brd-1-null strains into the survival assay. As previously published, both null mutants displayed pronounced hypersensitivity to IR (Fig. 3A). These findings show that strains bearing the non-phosphorylatable alleles are more sensitive to IR than the wild-type strain, yet their sensitivity remains intermediate between wild-type and the corresponding brc-1 or brd-1 null mutants.

Figure 3.

Loss of BRC-1 and BRD-1 phosphorylation leads to IR sensitivity. (A) Sensitivity of L4-stage worms from the indicated strains to different doses of irradiation (IR). Survival percentage of offspring is shown. Data are represented as average percentage ± SD from three independent experiments with 15 worms each. A 2-way ANOVA with Dunnett’s multiple comparisons test was performed. Significant P-values are indicated. (B) Representative images of mid-pachytene nuclei stained with anti-RAD-51 (red) and DAPI (blue) for the indicated genotypes, 48 h after irradiation at doses indicated. Scale bar, 5 μm. (C) Quantification of recombination marker RAD-51 foci in the indicated strains 48 h after irradiation at doses indicated on top of each graph. A minimum of three gonads from three independent experiments were analysed and 70–80 nuclei were scored for each region per gonad. On the right, a diagram of a hermaphrodite gonad, indicating the regions in which the number of RAD-51 foci was scored: (i) mitotic division zone; (ii) transition zone; (iii) early pachytene; (iv) mid-pachytene; (v) late pachytene; and (vi) diplotene. Data are represented as average ± SD. (D) Quantification of apoptotic corpses stained with SYTO-12 in late pachytene stage in the indicated strains at different times after 75 Gy irradiation. A minimum of 20 gonads from 4 independent experiments were analysed. Data are represented as average ± SD, and P-values for significant unpaired t-tests are indicated.

We showed earlier that brc-1^4A and brd-1^3A show mild defects in repairing programmed meiotic DSBs. To determine how brc-1^4A and brd-1^3A alleles respond to an excess of DSBs, we analysed the intensity and distribution of RAD-51 foci 48 h post-treatment with 75 Gy IR since at 24 h we still saw mitotic cell cycle arrest. In the C. elegans germ line, a checkpoint network senses the damage and enforces cell-cycle arrest, safeguarding the diploid precursor cells in the distal mitotic region and therefore the future developing gametes in meiotic prophase. Under non-treated conditions we could observe similar results as before, in which both mutants show modestly increased RAD-51 foci until late pachytene when compared with the wild-type, suggesting a defect in DSB repair (Fig. 3B and C and Supplementary Figs S2 and S3). Consistent with an intermediate phenotype, the non-phosphorylatable brc-1 and brd-1 strains accumulate RAD-51 foci from pachytene through diplotene under untreated conditions to a lesser extent than the corresponding null mutants (Fig. 3B and C and Supplementary Figs  S2 and  S3). In germlines from wild-type treated worms, we observed a mild increase in the levels of RAD-51 at the transition zone and an increase in the percentage of nuclei with RAD-51 foci compared to germlines from non-irradiated wild-type worms (Fig. 3B and C and Supplementary Figs S2 and S4), implying that at this time point the wild-type has less DSB repair efficiency. Both null mutants and phosphomutants displayed similar increased RAD-51 foci tendency, although not statistically significant (Fig. 3B and C and Supplementary Figs S2 and S4). In contrast with the wild-type, we observed more nuclei in the category of >12 foci or RAD-51 stretches (at regions 4 and 5) in the brc-1^4A and brd-1^3A mutant strains, only shared with the brd-1 null mutant. We also quantified apoptotic corpses at different time points after L4-stage hermaphrodites IR treatment. As can be observed in the graph, brc-1^4A and brd-1^3A mutant strains showed a significant increase of apoptosis throughout the time course compared to the control wild-type (Fig. 3D). These data indicate that the repair of IR-induced DSBs is initiated in the presence of the unphosphorylated versions of BRC-1 and BRD-1 proteins.

IR-dependent meiosis catastrophe in non-phosphorylatable brc-1 and brd-1 alleles

While performing the RAD-51 analysis of L4-stage hermaphrodites at 48 h after irradiation, we noticed an increase in diakinesis nuclei with abnormalities. In C. elegans successful meiotic recombination and CO formation results in six bivalents at diakinesis; however, mutants defective in DSB repair usually show altered DAPI bodies at diakinesis, particularly after exogenous DNA damage. Therefore, we analysed the DNA structures of diakinesis nuclei along the z-stacks of the acquired images (videos of several examples can be found as supplementary material). In unchallenged conditions, DAPI staining of diakinesis cells showed no significant differences between the wild-type and brc-1^4A and brd-1^3A mutant strains: the majority of diakinesis presented six DAPI-stained bodies (Fig. 4A and B and Supplementary Fig. S5). Intriguingly, occasional chromosome fusion events were detected in nuclei of the phospho-mutant strains, whereas such aberrations were never observed in the corresponding null backgrounds (Fig. 4B). In all the strains, including the wild-type, we detected an increase in aberrant diakineses following IR treatment (75 Gy) (Fig. 4A and B and Supplementary Fig. S5). Strikingly, however, brc-1^4A and brd-1^3A strains showed a dramatic increase in chromosome aberrations, including altered numbers of DAPI bodies (Fig. 4B), chromosome fusions, fragmentation, or both [Fig. 4A panels (a, b, d) and 4B]. In addition, 5.8% of brc-1^4A and 2.9% of brd-1^3A diakinesis nuclei exhibit diffuse chromosome masses [Fig. 4A panel (c) and 4B]. This observation indicates that failure to repair IR-induced DSBs results in aneuploidy when BRC-1 and BRD-1 cannot be phosphorylated.

Figure 4.

IR-dependent meiosis catastrophe in non-phosphorylatable brc-1 and brd-1 alleles. (A) Representative diakinesis-stage nuclei of the indicated strains and conditions stained with DAPI showing normal karyotype (six bivalents) and defective diakinesis products 48 h after 75 Gy irradiation. Arrowheads mark the described phenotypes: (a) chromosome fragmentation, (b) chromosome fusions, (c) masses, and (d) chromosome fusions and fragmentation. Scale bar, 3 μm. (B) Quantification of the different diakinesis nuclei phenotypes in the diakinetic oocytes of the indicated strains in normal conditions and 48 h after 75 Gy irradiation. −4 to −1 oocytes from a minimum of 10 gonads from 3 independent experiments were analysed for each strain and condition. Data are represented as the mean of the different experiments ± SEM. (C) Quantification of pro-CO marker GFP::MSH-5 foci in the indicated strains 48 h after 75 Gy irradiation. A minimum of four gonads from four independent experiments were analysed and 12 nuclei from late pachytene stage were scored per gonad. Individual values are represented. P-values for unpaired t-test weren’t significant. (D) Representative images of late-pachytene stage nuclei stained with pro-CO marker MSH-5 fused to GFP (green) and DAPI (blue) for the indicated strains 48 h after 75 Gy irradiation. Scale bar, 5 μm.

The high percentage of diakinesis nuclei with abnormal chromosomes did not initially correlate with the survival curves observed in our IR sensitivity assay. It is important to note that these experiments were conducted at different developmental stages and timepoints following irradiation. Specifically, the original survival assay analysed the progeny produced within 24 h post-irradiation of L4-stage hermaphrodites, whereas immunofluorescence was performed 48 h after irradiation of 24-h post-L4 adults. To address this discrepancy, we performed an additional IR sensitivity assay, this time analysing the progeny produced 48 h post-irradiation of L4-stage hermaphrodites. Under these conditions, we observed a marked decrease in survival in the wild-type strain as well as in both phospho-mutant strains (Supplementary Fig. S10). Interestingly, we also found that in the 24-h assay, brood size was reduced by approximately half following IR treatment, whereas this reduction was not seen in the 48-h assay (Supplementary Table S1). These differences suggest that worm age at the time of IR exposure may influence both germline sensitivity and reproductive output, potentially confounding the interpretation of survival data. Thus, this could be in part related to the age of the nematodes and could be masking the data.

Phosphorylation of BRC-1 regulates HR intermediate processing

Repair of meiotic DSBs is crucial for proper formation of CO; C. elegans exhibit extreme interference with only one CO formed per homolog pair. It has been shown that loss of brc-1 and brd-1 does not affect the number of CO but leads to an altered CO landscape [10]. Nevertheless, we checked the number of CO in the phospho-mutant alleles; to this end, we combined the alleles with the CO marker gfp::msh-5 [11] and analysed MSH-5 foci, a marker of COs. In normal conditions, we observed six CO in wild-type, brc-1^4A, and brd-1^3A strains at late pachytene. In germlines challenged with IR we observed no significant increase in the number of MSH-5 foci in wild-type worms and the same was observed in the strains carrying the non-phosphorylatable alleles (Fig. 4C and D). These data indicate that the number of COs is not overtly affected in brc-1^4A and brd-1^3A even if excess DSBs are introduced by IR and inter-homolog recombination is preserved in our mutants’ alleles.

In C. elegans the NHEJ repair pathway requires the activities of the CeKU70-CeKU80 heterodimer and lig-4 (Ligase IV). Previous studies have shown that illegitimate activation of NHEJ is in part responsible for the meiotic abnormalities observed in HR mutants [43]. Having observed that IR-induced DSBs lead to defective diakinesis in the phospho-mutant strains, we wanted to address whether NHEJ contributes to this phenotype. To this end, we analysed diakinesis nuclei with and without IR in wild-type, brc-1^4A, and brd-1^3A strains subject to lig-4 RNAi. Under normal conditions, wild-type, brc-1^4A, and brd-1^3A strain worms presented with six bivalents at diakinesis, and this was modestly altered by lig-4RNAi. After IR treatment, the presence of aberrant chromosomal fusion and fragmentations in the brc-1^4A and brd-1^3A germ lines was exacerbated irrespective of the status of LIG-4 (Fig. 5A and Supplementary Fig. S6). To confirm the effectiveness of the RNAi treatment, we assessed lig-4 transcript levels by RT-qPCR. The results demonstrate successful depletion of lig-4 following RNAi (Supplementary Fig. S6B). These data indicate that NHEJ does not make a significant contribution to the improper repair of IR-induced DSBs in these strains, suggesting that alternative meiotic DSB repair processes are driving misrepair in the absence of phosphorylation of BRC-1 and BRD-1.

Figure 5.

Meiosis catastrophe of the phospho-mutants is dependent on HIM6/MUS81. (A) Quantification of the different diakinesis nuclei phenotypes in the diakinetic oocytes of the indicated strains with/without LIG-4 RNAi depletion in normal conditions and 48 h after 75 Gy irradiation. −4 to −1 oocytes from a minimum of 10 gonads from 3 independent experiments were analysed for each strain and condition. Data are represented as the mean of the different experiments ± SEM. (B) Quantification of the different diakinesis nuclei phenotypes in the diakinetic oocytes of the indicated strains 48 h after the irradiation dose indicated on top of each graph. −4 to −1 oocytes from a minimum of 10 gonads from 3 independent experiments were analysed for each strain and condition. Data are represented as the mean of the different experiments ± SEM. (C) Quantification of the different diakinesis nuclei phenotypes in the diakinetic oocytes of the indicated strains 48 h after the irradiation dose indicated on top of each graph. −4 to −1 oocytes from a minimum of 10 gonads from 3 independent experiments were analysed for each strain and condition. Data are represented as the mean of the different experiments ± SEM.

It has been proposed that TMEJ acts as an alternative repair mechanism to HR in the brc-1 mutant [15, 16]. To test the involvement of TMEJ, we generated double mutants of brc-1^4A and polq-1. While we observed no overt differences in the polq-1;brc-1^4A strain between the single mutants in the absence of IR (Fig. 5B and Supplementary Fig. S7), IR treatment led to a synergetic effect in the polq-1;brc-1^4A strain. Analysis of the double mutant polq-1;brc-1^4A strain showed that IR treatment led to a general increase of the aberrant diakinesis nuclei, with almost no presence of nuclei showing six bivalents and an increase in chromosome fusions and chromosome masses (an extreme degree of fusion plus de-condensation) compared with the single mutant (Fig. 5B and Supplementary Figs S7 and S8). These results indicate that TMEJ is not responsible for the abnormal repair of IR-induced DSBs that lead to fusion events in the brc-1^4A strain.

Although COs are not affected in our phospho-alleles, it has been shown that BRC-1 is required for efficient processing of recombination intermediates. We therefore examined diakinesis in the absence of later HR intermediate processing factors, including the resolvase MUS-81 and Holliday junction resolution factor HIM-6 (Bloom). Again, no changes were observed in the double mutants in the absence of DNA damage treatment. In contrast, in gonads of mus-81;brc-1^4A and him-6;brc-1^4A exposed to IR, we observed a general reduction in aberrant diakinesis nuclei, with a suppression of chromatin masses and a significant decrease in the number of nuclei that showed chromosome fusions and fragmentation (Fig. 5B and Supplementary Figs S7 and S8). Moreover, in the double mutant with him-6 the reduction in aberrations results in an increase in the univalents after IR (Fig. 5B and Supplementary Fig. S7). These data reveal that the phosphorylation of BRC-1/BRD-1 after IR is important for the correct regulation of HR intermediate resolution/dissolution.

To further investigate whether the defects observed in our phospho-mutant alleles are specifically related to the processing of HR intermediates, we performed diakinesis analysis following IR in genetic backgrounds deficient in CO formation, such as syp-1 and msh-5, as described earlier. These mutants are unable to generate COs and therefore do not present HR intermediates. Visibly, in these CO-deficient backgrounds, the presence of masses and the frequency of nuclei showing chromosome fragmentation and fusions were reduced in the double mutants carrying the brc-1 phospho-alleles compared to single brc-1^4A mutant (Fig. 5C and Supplementary Fig. S9). This finding supports the notion that BRC-1 phosphorylation plays a key regulatory role in the proper resolution of HR intermediates.

The reduction in diakinesis chromosome masses when combining brc-1^4A allele with the resolvases prompts us to check whether this could alleviate the IR sensitivity observed in the non-phosphorylatable brc-1 allele. For that we included the mus-81;brc-1^4A strain in the sensitivity assay performed at 48 h post irradiation on progeny of L4-stage hermaphrodites. First we found that mus-81 is extremely sensitive to IR (Supplementary Fig. S10B). Additionally, the brood size reduction was exacerbated, particularly in the mus-81;brc-1^4A mutant (Supplementary Fig. S10A and C). However, in the mus-81;brc-1^4A strain we detected an increase of survival to the single mutant brc-1^4A levels. These data suggest that both act in the same pathway to deal with excessive DNA damage.

Discussion

In response to exogenous DSBs, the cellular response to DNA damage is initiated by the activation of ATM/Tel1 and ATR/Mec1 kinases, which are conserved throughout eukaryotes [44, 45]. In C. elegans it has been shown that both kinases also regulate key steps of meiosis, phosphorylating several target proteins involved in DSB formation and repair, cohesion, and synapsis between chromosomes [23–26, 46]. In previous work from our group, we described that IR-induced phosphorylation of SYP-1, a central SC component, is required for the repair of excessive DSBs relying on brc-1 activity [25], and it has been recently shown that BRC-1 regulates homolog-independent processing [14]. In this work we describe damage-induced phosphorylation of BRC-1/BRD-1, which we propose impacts template bias by controlling HR intermediates resolution. In contrast to brc-1/brd-1 null mutants, the non-phosphorylatable BRC-1 and BRD-1 alleles do not result in an increase of males in the progeny nor overt defects in meiotic DSB repair in the scenario where CO is impeded, indicating that DSB repair through the homolog is not compromised in the phospho-alleles [11, 13, 35]. However, brc-1^4A and brd-1^3A strains do exhibit a robust DNA damage sensitivity phenotype, which we attribute to DSB repair defects that lead to meiotic catastrophe and aberrant diakinesis.

Similar to their human counterparts, C. elegans BRC-1 and BRD-1 proteins interact through the N-terminal RING finger domains [40, 47]. Notably, the phosphorylation site identified in the BRD-1 protein, a cluster of three TQ, is located adjacent to the RING domain, which constitutes the core structural and functional element that enables BRD-1 to act as partner of BRC-1. Despite the proximity to the RING domain, phosphorylation of the TQ cluster is unlikely to affect heterodimerization, as the mutant strains exhibited normal BRC-1/BRD-1 localization on the SC. Indeed, it has been described that the BRC-1/BRD-1 heterodimer is interdependent in its localization along the SC and that the loss of either protein in synapsis-deficient backgrounds leads to embryonic lethality [10]. A recent study reported that mutations within residues 124–270 of human BARD1 cause a pronounced resection defect, reflected by a sharp decrease in RPA1- and RAD51-positive foci [48]. By contrast, our BRD-1 point mutant displays the opposite phenotype: RAD-51 foci accumulate more than in the wild-type, indicating that DNA-end resection remains largely intact. Our findings suggest that the BRD-1 variant may perturb a later step in double-strand-break repair, likely the recruitment or stabilization of downstream repair-pathway factors, rather than the initial resection stage and RAD-51 loading.

The phospho-site of BRC-1 is in one of the two BRCT repeat domains. BRCT domains are found in many DNA replication and repair proteins, are required to bind phosphorylated proteins, and are also involved in DNA and PAR binding [49]. In the case of human BRCA1, the BRCT domain is essential for tumour suppressor activity [50]. Recent studies have shown that the E3 ubiquitin ligase activity of the human BRCA1–BARD1 complex is essential for efficient HR, particularly during the early steps of DNA end resection. In C. elegans, this function appears to be more prominently associated with BRC-1 during meiosis. Moreover, E3 ligase activity has been implicated in the proper nuclear accumulation of the BRCA1–BARD1 complex [36]. Based on these observations, we considered whether the phenotype observed in our phospho-mutant alleles might not reflect impaired ubiquitin ligase function. We find that germline chromatin localization of the BRC-1–BRD-1 complex is not affected in the phospho-mutants, suggesting that E3 ligase activity is not impaired in our mutants. Therefore, we have ruled out a ubiquitin ligase-defective phenotype in these alleles. While the precise function of phosphorylation of BRC-1 within the BRCT domain is unknown, it is tempting to speculate that this post-translational modification may regulate the protein’s ability to interact with critical repair substrates or cofactors. Such modulation could fine-tune BRC-1 activity during key steps of HR, possibly influencing the recruitment or turnover of downstream repair factors. Further biochemical and structural studies will be required to elucidate the precise mechanistic impact of this modification.

Previous research has suggested that BRC-1 activity can stimulate non-CO-mediated repair and that in its absence, alternative repair mechanisms may be activated to repair intermediates [13, 51]. BRCA1 has been shown to suppress NHEJ in order to promote the repair of DSBs through HR. Since NHEJ has been linked to meiotic aberrations observed in HR mutants, we hypothesized that NHEJ repair may be used as a last resource to deal with the additional damage caused by IR in our phospho-mutants. However, contrary to our expectations, aberrant meiotic diakinesis was exacerbated after IR in the phospho-mutants lacking lig-4. We believe that while the inactivation of NHEJ is dependent on BRCA1, this role does not rely on these BRC-1 and BRD-1 specific phosphorylation sites. We also examined alternative NHEJ, also known as TMEJ, which functions to repair resected DSBs in the absence of HR [52], and it has been shown to be the source of genome structural variations in a BRCA1- or BARD1-deficient scenario [15]. Similarly to what we observed with NHEJ, TMEJ loss in the phospho-mutants conferred an increase in chromosomal abnormalities in IR-treated animals. These data indicate that neither NHEJ nor TMEJ is responsible for the abnormal diakinesis we observe.

Meiotic DSB repair via HR is believed to result in the formation of double Holliday junction (dHJ) intermediates, which can be resolved through cleavage by HJ resolvases, which results in either a CO or non-CO (NCO) product or dissolution by the BTR complex, consisting of the HIM-6/BLM helicase, topoisomerase IIIα, RMI1, and RMI2 [53, 54]. Importantly, CO designation is unaltered in the phosphomutants when there is excess DNA damage. In C. elegans it has been proposed that MUS-81 and HIM-6 act redundantly to resolve joint molecules early in meiosis, presumably to form NCO [5, 6, 55]. Strikingly, we show that the aberrant meiotic chromosome masses observed in the BRC-1 phospho-mutants after IR treatment are largely suppressed by removing HIM-6 and MUS-81. Moreover, in CO-deficient backgrounds, the frequency of nuclei showing chromosome fusions and fragmentation and chromosome masses was reduced, further supporting a role for BRC-1 phosphorylation in the proper resolution of HR intermediates.

In summary, we have identified in vitro DNA damage-dependent phosphorylation events in BRC-1 and BRD-1; the in vivo analysis of unphosphorylatable mutants has shown that these putative modifications protect germline cells from meiotic catastrophe upon IR treatment. We have excluded that this putative phosphorylation is to control alternative repair pathways as NHEJ or TMEJ, since the double mutants do not suppress the abnormal chromosomal rearrangements that we observe in the non-phosphorylatable mutants. Unexpectedly we have uncovered a role of BRC-1 in balancing the resolution/dissolution of recombination intermediates between homologous chromosomes, since removal of HIM-6/BLM or MUS-81 suppresses fusion that leads to chromosome masses of the brc-1^4A mutant. Hence, we propose that phosphorylation of BRC-1/BRD-1 in response to excess DNA damage regulates HR intermediate processing by MUS-81 and HIM-6/BLM. Our data support a model (Fig. 6) in which damage-induced BRC-1/BRD-1 phosphorylation is involved in the processing of NCOs. In this way, the cell regulates the excess inter-homolog COs, which could hinder correct chromosome segregation at the first meiotic division. Considering the conserved nature of most repair processes across species, it is possible that BRCA1 and BARD1 play similar roles in meiosis and DNA repair in higher eukaryotes. Indeed, the protein sequence of the human orthologs contains similar S/T-Q motifs. It will also be interesting to consider whether this regulatory mechanism is sex-specific since it has been shown in the nematode that BRC-1/BRD-1 differently regulates meiotic DSB repair and CO determination in male and female meiosis [16].

Figure 6.

Proposed model. BRC-1/BRD-1 phosphorylation dependent on DDR activation by ATM-1/ATL-1 regulates HIM-6/MUS-81 function, correctly balancing the resolution/dissolution of recombination intermediates arising between homolog chromosomes. Through this mechanism, the cell regulates the processing and repair of excess inter-homolog COs, which could hinder correct chromosome segregation at the first meiotic division.

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

None declared.

Funding

The publication of this article is funded by Spanish Ministry of Science and Innovation, the Spanish Research Agency, and the European Regional Development Fund—Proyecto PID2021-123850N-I00 financiado por MCIN/AEI/10.13039/501100011033/y por FEDER Una manera de hacer Europa. García-Muse’s lab work is supported by grants from Spanish Ministry of Science, Innovation and Universities (PGC2018-101099) and Spanish Ministry of Science and Innovation (PID2021-123850N). M.C. was holder of postdoctoral fellowship from the Universidad de Sevilla. N.F-F. was holder of formation grant PJUS5 from the Universidad de Sevilla (dentro del Marco del Sistema Nacional de Garantía Juvenil y del Programa Operativo de Empleo Juvenil 2014–2020). Funding to pay the Open Access publication charges for this article was provided by Universidad de Sevilla.

Data availability

Further information and requests for resources and reagents should be directed to and will be fulfilled upon reasonable request by the Lead Contact, Tatiana Garcia-Muse (tatiana.muse@cabimer.es).