Three of the DSB proteins appear to have a biased effect on X chromosome CO formation: null alleles of xnd-1 and him-5 predominantly show a loss of CO intermediates on the X, increases in univalent X chromosomes at diakinesis in half or almost all nuclei, respectively, and an increase in XO male offspring resulting from nondisjunction of the X chromosome during the meiotic division; the hypomorphic him-17(e2806) mutation shows increased incidence of one unattached homolog pair in diakinesis and an increase in male production consistent with an X-specific defect (Reddy and Villeneuve, 2004). To determine if these genes function through a common underlying mechanism, we first performed Y2H assessing interactions with both high and low stringency. We observed a strong Y2H interaction between XND-1 and HIM-17, and a weak one between HIM-17 and HIM-5 (Figure 1A, Figure 1—figure supplement 1). To validate these interactions, we also performed immunoprecipitations (IPs) of HIM-17::GFP followed by mass spectrometry (MS, Figure 1—figure supplement 1B). This identified XND-1, as well as 16 additional strong interactors of HIM-17::GFP (Figure 1—figure supplement 2). Moreover, endogenous XND-1 was identified in western blots of IPs performed independently using 3xHA-tagged HIM-17 epitopes, indicating a robust interaction between these proteins (Figure 1B, Figure 1—figure supplement 1C). In contrast, HIM-5 was not detected in either the HIM-17::GFP co-IPs or IP/MS, suggesting that the interaction observed by Y2H either exists very transiently or was bridged by a protein expressed in yeast.

Prior studies from our group indicate that xnd-1 regulates HIM-5 transcriptionally since ectopic expression of him-5::GFP using the pie-1 germline promoter, but not expression using the endogenous him-5 promoter, rescues the X chromosome nondisjunction defects conferred by the xnd-1 mutation (Wagner et al., 2010; Meneely et al., 2012; McClendon et al., 2016). Since RNA-Seq studies showed that him-17 regulates many germline genes including him-5 (Carelli et al., 2022), we set out to determine if ectopic expression of him-5 could also rescue him-17 null mutants. Similar to what we observed with xnd-1 mutants, Ppie-1::him-5—but not Phim-5::him-5—rescued the X chromosome defects seen by the increase of XO male offspring from XX hermaphrodites (known as the high incidence of males or ‘Him’ phenotype) (35% of males in him-17(ok424) vs. 7.02% in Ppie1::him-5;him-17(ok424)) (Figure 1C). Consistent with the lack of rescue, there was no statistically significant change in univalent formation in Phim-5::him-5;him-17(ok424) (Figure 1—figure supplement 3). These results cannot be explained by HIM-5 protein levels which are lower in Ppie-1::him-5 than in Phim-5::him-5 (McClendon et al., 2016). The lack of rescue from Phim-5::him-5 supports the conclusion that HIM-17 directly regulates him-5 expression. Since the pie-1 promoter appears to be independent of HIM-17, the Ppie-1::him-5 construct can—at least partially—bypass this regulation and suppress the him-17 mutant phenotype. These findings suggest that both XND-1 and HIM-17 promote X chromosome DSB and CO formation by regulating him-5 expression. Moreover, xnd-1 and him-17 have distinct roles in DSB/CO formation since only null alleles of him-17, but not xnd-1 or him-5, confer a severe loss of DSBs (Meneely et al., 2012; Reddy and Villeneuve, 2004; Wagner et al., 2010), while only xnd-1 mutations influence the timing of DSB induction/repair (Wagner et al., 2010; McClendon et al., 2016).

We note that ectopic HIM-5 expression strongly rescued the embryonic lethality conferred by the him-17 null allele (Figure 1C). While a small fraction of the lethality is attributed to X chromosome nondisjunction, most of the lethality is a consequence of autosomal missegregation. Thus, even though him-5 null mutant animals exhibit a profound loss of DSBs/COs on the X chromosomes with rare CO defects on autosomes (Meneely et al., 2012), him-5 expression can restore CO formation to both autosomes and X chromosomes. This strongly supports a general role for HIM-5 in CO formation and further indicates that HIM-5 is sufficient to restore COs on the X chromosome.

To further characterize the interaction between XND-1 and HIM-17, we also sought to determine whether they colocalize on meiotic chromosomes using confocal microscopy. Prior studies have shown that HIM-17 associates with meiotic chromatin (Reddy and Villeneuve, 2004). Co-staining the HIM-17::HA transgenic line with anti-HA and anti-XND-1 antibodies showed that HIM-17, like XND-1, is enriched on autosomes (Figure 1D; Wagner et al., 2010). Both proteins showed no overlap with the chromosome axes (marked with HTP-3; Figure 1D) or 4′,6-diamidino-2-phenylindole (DAPI; Figure 1—figure supplement 4B), although we cannot rule out that a small fraction of the population, either below the level of detection or transiently, localizes to the axes. These immunolocalization studies also show a strong juxtaposition of HIM-17 and XND-1, with no detectable overlap between their staining patterns in germ lines from the mitotic region through late pachytene, when XND-1 disappears from the nucleus (Figure 1E, Figure 1—figure supplement 4B). Structural modeling of the XND-1 and HIM-17 interaction (Abramson et al., 2024; Comeau et al., 2004a) provides a potential explanation for this observation: HIM-17 is a very large protein, and the HA-tag positioned at its amino terminus is predicted to be located at least 180 Å away from the protein:protein interaction interface (Figure 1F).

Since HIM-5 appears to be central to DSB/CO formation on the X chromosome and the presence of DSB-1 and DSB-2 is required for all meiotic breaks, we next sought to determine whether HIM-5 interacts with DSB-1, its paralog DSB-2, or DSB-3. We found that DSB-1 and DSB-2 show robust Y2H interactions and that DSB-1 and DSB-3 exhibit weak Y2H interactions (Figure 2A, Figure 2—figure supplement 1A), confirming results from a prior study (Hinman et al., 2021). We also found that HIM-5 shows weak Y2H interactions with DSB-1 (Figure 2A). We confirmed the interaction between DSB-1 and HIM-5 with co-IP, pulling down GFP::DSB-1 and probing for HIM-5::3xHA by western blotting (Figure 2B).

We next wanted to understand the functional relevance of the interaction between DSB-1 and HIM-5. To this end, we examined the localization of HIM-5 proteins in the dsb-1 mutant background. Co-staining of the chromosome axis (anti-HTP-3) and either HIM-5::GFP transgene (anti-GFP) or endogenously tagged HIM-5::3xHA (anti-HA) revealed that HIM-5 nuclear localization begins just prior to the transition zone in both wild-type and dsb-1 mutants (Figure 2C, Figure 2—figure supplement 1A, arrows). However, whereas in wild-type, HIM-5 protein remains nuclear throughout early-mid pachytene and becomes further enriched on meiotic chromatin; in dsb-1 mutants, HIM-5 nuclear localization is lost (Figure 2C, Figure 2—figure supplement 1B). From this change in HIM-5 protein distribution, we infer that DSB-1 (and/or its DSB-promoting function) is required to retain HIM-5 in meiotic nuclei.

Since HIM-5 levels could also be affected by dsb-1 and/or dsb-2, we next asked whether ectopic expression of HIM-5 could suppress defects in bivalent formation due to the loss of dsb-1 or dsb-2. As shown in Figure 2D, ectopic HIM-5 reduced the number of univalents in both mutant backgrounds. Since a small fraction of dsb-2; Ppie-1::him-5 animals exhibited six or seven DAPI bodies, we assayed whether the transgene could improve hatching rates and reduce male production in dsb-2 (Figure 2E). Whereas the Ppie-1::him-5 transgene has no impact on hatching and male frequencies, the more highly expressed Phim-5::him-5 transgene (McClendon et al., 2016) conferred a mild suppression of the embryonic lethality. This ability to suppress hatching—but not male production—further supports a role for HIM-5 in regulating total DSB numbers.

Having established interactions between subsets of the SPO-11 accessory factors in C. elegans, we next wanted to perform more extensive genetic analyses between the known DSB factors. To this end, we created double mutant combinations of alleles with partial DSB defects. These included either weak loss-of-function alleles of essential DSB genes (him-17 and mre-11) or null alleles of genes that partially reduce DSBs (cep-1, dsb-2, rec-1, him-5, and parg-1). Diakinesis oocytes were analyzed in the single and double mutants by whole-mount fixation and DAPI staining followed by confocal microscopy and 3D visualization. The chromosomes are highly condensed at diakinesis, and six bivalents can be detected in wild-type worms. Univalents, fusions, and DNA fragments can be seen in mutants with defects in DSB formation or their repair into a CO, respectively (Chin and Villeneuve, 2001; Guo et al., 2022; Rosu et al., 2013). We hypothesized that double mutant strains carrying mutations in genes that function in different aspects of DSB induction would show more severe defects and have an increase in univalent formation compared to single mutants. On the other hand, genes that function together or at the same step of break induction should exhibit no or only mild enhancement of the DSB defects.

Double mutant analysis was precluded with dsb-1 and dsb-3 since the single mutants already show a complete absence of breaks (Hinman et al., 2021; Stamper et al., 2013). By contrast, dsb-2 mutant nuclei manifest an age-dependent loss in DSB capacity (Rosu et al., 2013; Stamper et al., 2013) allowing us to analyze interactions between dsb-2 and cep-1, parg-1, him-17, and rec-1 (Figure 3A, Figure 3—figure supplement 1A–D). As 1-day-old adults, dsb-2 single mutants had ~50% of diakinesis nuclei that contained 6 bivalents, ~30% with 7 DAPI-stained bodies (5 bivalents and 2 univalents), and ~20% with 8 or more DAPI-stained bodies. Single mutants of cep-1, parg-1, and rec-1 each had >90% of diakinesis nuclei with 6 bivalents, consistent with prior studies (Chung et al., 2015; Janisiw et al., 2020; Mateo et al., 2016). In animals carrying him-17(e2806) mutation, a weak allele of the otherwise essential locus for DSB formation, only ~50% of mutant nuclei had six bivalents; the other 50% had seven DAPI-stained bodies, reflecting the requirement for this gene in CO formation on the X chromosome (Reddy and Villeneuve, 2004). When dsb-2 was combined with each of these mutations, an increased number of DAPI-stained bodies and a concomitant decrease in the number of chiasmata were observed (Figure 3A, Figure 3—figure supplement 1). These results strongly suggest that dsb-2 functions cooperatively with all four loci for the regulation of meiotic CO formation.

Given that rec-1, parg-1, and cep-1 fall in a separate functional group from dsb-2, we next addressed whether these three genes function together or independently. All double mutant permutations of rec-1, parg-1, and cep-1 were analyzed, and in each case, the single and double mutants were nearly indistinguishable from the single mutants, that is, six bivalents (Figure 3—figure supplement 1L, M). Therefore, we infer that these genes likely function to control the same step in DSB/CO induction (Figure 3B). Consistent with this interpretation, when rec-1, parg-1, or cep-1 were combined with other DSB gene mutations, including him-17, they had similar effects, and the numbers of univalents were increased (Figure 3—figure supplement 1A, C, D for Group 1; Figure 3—figure supplement 1E–G for Group 3; and Figure 3—figure supplement 1J, K for Group 4). Similar results have previously been reported from our laboratories for double mutants with him-5 (Chung et al., 2015; Janisiw et al., 2020; Mateo et al., 2016).

mre-11 functions in both DSB formation and subsequent DNA end resection (Chin and Villeneuve, 2001), the latter as part of the MRE11–RAD50–NBS1/Xrs2 (MRN/X) resection complex, the other proteins of which are encoded by the worm rad-50 and nbs-1 genes (Girard et al., 2018; Hayashi et al., 2007). Null mutations in mre-11 and rad-50 completely abrogate DSB induction. However, we were able to take advantage of a separation-of-function allele, mre-11(iow1), that cannot perform its end-resection functions but can induce DSBs (Yin and Smolikove, 2013). The iow1 mutation causes chromosome fusions to appear in diakinesis oocytes resulting from the aggregation of unprocessed DNA ends. While this allele makes enough DSBs to cause aggregation of all chromosomes, we hypothesized that it might have a minor impairment in DSB induction that would be revealed when combined with other DSB mutations. If this were the case, the double mutants would make fewer breaks, leading to fewer chromosome fusions and an increase in the number of DAPI-stained bodies/univalents. Consistent with this hypothesis, mre-11(iow1) double mutants with each of the other DSB mutations led to fewer fusions and more DAPI-stained bodies (Figure 3A, Figure 3—figure supplement 1H–K). Thus, we propose that mre-11, and by extension rad-50, function in their own branch of the DSB/CO pathway.

Surprisingly, a unique phenotype was observed in a subset of the cep-1; mre-11 double mutants. Our prior studies showed that cep-1 functions redundantly with him-5 to both ensure DSB formation and promote downstream homologous recombination (HR) repair by preventing NHEJ (Mateo et al., 2016). In that study, we reported that cep-1(lg12501) is a separation-of-function allele that is defective in DSB formation. In the cep-1(lg12501); mre-11(iow-1) double mutants, we observed mixed phenotypes: some nuclei had 12 DAPI-stained bodies, while others exhibited an exacerbation of the fusion defect (Figure 3—figure supplement 2). We infer that the univalents indicate a synergistic effect between mre-11 and cep-1 for DSB formation; the increase in fusions, on the other hand, may suggest that cep-1(lg12501) is not fully wild-type for the DNA repair functions of CEP-1.

While each of the genes that we analyzed has reported roles in DSB formation, we wanted to confirm that the observed increases in univalents in the double mutants were in fact due to impairment in DSB induction and not to deficits in later steps of CO formation. Therefore, we relied on an established assay to distinguish between these possibilities: we asked whether the addition of exogenous breaks from gamma irradiation (IR) could suppress the univalent phenotype by serving as a surrogate for SPO-11-induced DSBs (Dernburg et al., 1998). Recent work has shown that 10 Gy IR is able to produce up to ~20 DSBs per meiotic nucleus (Lascarez-Lagunas et al., 2023), which is sufficient to ensure a CO on each chromosome (Mets and Meyer, 2009). We analyzed the number of DAPI bodies with and without IR in a subset of the double mutants where we observed synergistic effects (dsb-2; parg-1, dsb-2; rec-1, and rec-1; him-17). When each of these double mutant strains was exposed to IR, the wild-type phenotype of 6 bivalents was largely restored (Figure 3—figure supplement 3), supporting the conclusion that the CO deficiency observed is triggered by defects in DSB formation.

The genetic analyses described above defined four functional groups that contribute to break formation (Figure 3B). Using Y2H, we next set out to determine whether interactions between and within groups could explain how the different complexes come together to regulate DSB induction (Figure 3C, D, Figure 1—figure supplement 1A). CEP-1 showed auto-activation in both vectors and therefore was removed from further analysis. DSB-1, HIM-17, and REC-1 also showed self-activation in the binding domain vector; therefore, only the activation domain fusions were analyzed. Pairwise interactions between all remaining combinations were then analyzed. HIM-5 displayed the most promiscuous Y2H interactions, with strong interactions seen with MRE-11 and weak interactions as already discussed with HIM-17 and DSB-1. XND-1, apart from interacting with HIM-17, also showed strong Y2H interactions with REC-1. A recent study provided Y2H evidence supporting the idea that SPO-11 interacts with DSB-1 (Hinman et al., 2021). Critically, our work supports and extends this observation to show that DSB-1 is the only known accessory factor that interacts with SPO-11 in Y2H assays, cementing its role as a key determinant for DSB induction and explaining its role in feedback controls (Stamper et al., 2013; Hinman et al., 2021; Guo et al., 2022).

To analyze more in-depth the Y2H interactions between DSB-1 and DSB-2, DSB-3, and SPO-11, we looked for protein domains in DSB-1 that could be involved in these interactions. Interestingly, DSB-1 has five ‘sticky’ helices (helices 2–6 in Figure 3—figure supplement 4) that have the potential to be implicated in protein–protein interactions. We first deleted all of these helices together by introducing a deletion after L114 (DSB-1 (1–114) in Figure 3E, F). We found that this truncated version of DSB-1 loses the ability to interact with DSB-2 and DSB-3 but is still able to associate with SPO-11. Moreover, when we only deleted the helices 3–6 but kept the helix 2 (DSB-1 (1–145)), we obtained the same results, suggesting that the Y2H interactions between DSB-1 and DSB-2/DSB-3 are mediated by the C-terminal helices 3–6 and not helix 2. To prove this, we created a truncated DSB-1 protein that only contains amino acids from P145 to the stop codon (DSB-1(145-Ct)). Consistent with the deletion analysis, we observed that the C-terminal domain is sufficient to promote associations with DSB-2 and DSB-3. This domain was also not sufficient to promote the DSB-1/SPO-11 interaction (Figure 3E). These experiments complement the results published by Villeneuve lab where they observed that the DSB-1 N-terminus is the domain necessary for the DSB-1/SPO-11 (but not for the DSB-1/2/3) interaction (Hinman et al., 2021).

We next used AlphaFold and structural modeling to investigate whether the C. elegans REC114/MEI4 homolog could assemble in the same fashion as their distant orthologues (Claeys Bouuaert et al., 2021; Daccache et al., 2023) in which two helices of REC114 dimer surround a MEI4 helix. DSB-1 is predicted to form a stable homodimer that can assemble into a highly stable trimeric complex with DSB-3 threaded through the middle (Figure 3G–I) analogous to what is predicted for Rec114-Mei4 (Daccache et al., 2023). By contrast, DSB-2 cannot form homodimers, although it can replace one of the DSB-1 monomers to form the DSB-1:DSB-2:DSB-3 ternary complex. Free energy calculations (Champ and Camacho, 2007) show that the DSB-1:DSB-2:DSB-3 complex is less stable (–3.8 kcal/mol) than a DSB-1:DSB-1:DSB-3 complex consistent with the more limited role for DSB-2 in break formation, at least in young worms (Rosu et al., 2013; Stamper et al., 2013).

Studies from multiple systems have elucidated mechanisms to coordinate the timing of DSBs with the completion of meiosis, for example by CDK and Ddf4 phosphorylation in yeast (Murakami and Keeney, 2008; Carelli et al., 2022). In worms, REC-1 was previously shown to be a target of CDK, and therefore, it was hypothesized that such a phosphorylation event would be critical to ensure that break induction occurs after the completion of S phase (Chung et al., 2015). REC-1 and HIM-5 evolved from a gene duplication event in the C. elegans lineage, whereas the ancestral gene in related nematodes has features of both proteins, although it shares greater homology with HIM-5 (Chung et al., 2015). Thus, we hypothesize that the single ancestral REC-1/HIM-5 protein would be subject to CDK phosphorylation, which would then promote the assembly of the complexes at the chromosome axis to promote break formation.

Timing also appears to be critical for CO outcomes. Studies in yeast show different outcomes for early and late breaks (Joshi et al., 2015), and recent studies in worms indicate that later breaks are needed for CO formation (Hicks et al., 2022). We previously showed that xnd-1 mutants have altered DSB kinetics, showing elevated levels of DSB formation in leptotene/zygotene (McClendon et al., 2016; Wagner et al., 2010). This function of XND-1 is mediated, at least in part, by its effect on chromatin architecture. One possibility is that the physical interaction that we show here between XND-1 and REC-1 couples the completion of S phase to the activation of breaks within the correct chromatin environment, thus ensuring their timely formation. In the absence of XND-1, increases in global histone acetylation (Gao et al., 2015; Wagner et al., 2010) might allow SPO-11 easier access to the DNA. Another possibility is that upon phosphorylation of REC-1 by CDK-1, REC-1 through PARG-1 transiently brings XND-1 (the timer) to the axis, an event that restrains DSB formation.

We provide strong evidence for direct interactions between XND-1 and HIM-17, both by Y2H and co-IP. HIM-17 appears to have a role in transcriptional regulation of DSB factors. Here, we show that ectopic expression of HIM-5 from the pie-1 promoter, but not overexpression from the endogenous him-5 promoter, suppressed him-17 phenotypes. This is similar to the rescue we saw of xnd-1 (McClendon et al., 2016) and suggests that both proteins contribute to the expression of him-5. This supports findings that him-17 regulates the expression of ~300 germline-enriched genes, including him-5, rec-1, and dsb-2 (Carelli et al., 2022), whose concomitant loss would be expected to abrogate break induction (Chung et al., 2015 and Figure 3A, Figure 3—figure supplement 1). While its only role may be in transcriptional regulation, a growing literature and data presented herein suggest a more direct role for HIM-17 in meiosis. We presented Y2H data for a weak interaction between HIM-17 and HIM-5. Although these interactions were not confirmed by IP, it remains possible that these complexes are transient associations, are unstable during IP, and/or are not abundant. Together with observed changes in chromatin structure and both DSB and CO distribution in him-17 mutants (Nadarajan et al., 2021; Reddy and Villeneuve, 2004), it is tempting to speculate that together with XND-1, HIM-17 helps to create an environment permissive for SPO-11 binding and/or break formation. XND-1 and HIM-17 may function together in this role or redundantly, perhaps through XND-1 interacting with REC-1 and HIM-17 interacting with HIM-5, or in related species through interactions with the single REC-1/HIM-5 paralog.

One of the conundrums arising from our data is in the function of the HIM-5 protein. him-5 was identified based on its preferential impact on CO formation on the X chromosome (Broverman and Meneely, 1994; Hodgkin et al., 1979). The X-specific phenotypes conferred by the him-5 mutation could suggest that the X chromosome is simply more sensitive to a reduction in the number of active SPO-11 complexes; the fact that mutations in dsb-2 also reduce DSBs but without an X chromosome bias argues against this interpretation. We show herein that HIM-5 is sufficient to ensure that COs occur on the X chromosome. Ectopic expression of him-5 can substantially rescue both the autosomal and X chromosomal defects of him-17 null mutations—despite expression levels lower than endogenous him-5 (McClendon et al., 2016). The X chromosome-specific defects of him-5 mutations might arise due to defects in recruitment of DSB-1-2-3/SPO-11 to specific sites on the X or to an impaired ability to coordinate the assembly of the DSB machinery on the—late-replicating (Jaramillo-Lambert et al., 2007)—axis of the X chromosome, although other models can be envisioned.

Our data and a growing literature insinuate a more general and central role for HIM-5 in DSB formation and downstream CO repair: (1) rec-1;him-5 double mutants give an age-dependent severe loss of DSBs (like dsb-2 mutants) suggesting that the ancestral function of the protein may have a more profound effect on break formation (Chung et al., 2015); (2) total DSBs are reduced to nearly 50% of wild-type and are severely delayed in formation in him-5 mutants (Meneely et al., 2012); (3) a recent study reports a role for him-5 in CO formation in males where the X chromosome does not engage in CO exchange (Engebrecht et al., 2025); (4) HIM-5 is required for the normal distribution of COs on the autosomes (Meneely et al., 2012); (5) redundantly with CEP-1, HIM-5 promotes DSB formation and HR-mediated repair and inhibits NHEJ (Mateo et al., 2016), and (6) our data herein show that HIM-5 interacts not only with DSB-1, but also components of the other functional groups that regulate DSB induction. Together, we believe this data supports a role for HIM-5 in bridging the accessory factors with SPO-11 to ensure the proper timing, placement, and number of breaks.

Our Y2H and IP results show very strong protein–protein interactions between the different DSB factors; however, their localization along the gonads does not frequently = coincide suggesting temporal and spatial regulation throughout meiosis. Further studies are necessary to elucidate how these interactions arise and change in vivo and how these are integrated with both cell cycle and regulatory feedback controls.

Worms were cultured on MyoB plates (Burns et al., 2006) seeded with OP50 at 20°C, unless otherwise noted (Brenner, 1974). Mutant strains used in this study are listed in Supplemental Material, Supplementary file 1. All strains were derived from the wild-type Bristol strain N2.

The numbers of DAPI-stained bodies present in diakinesis oocytes were assessed in intact adult hermaphrodites at day 1 of adulthood (24 hr post-L4, unless otherwise specified). Adults were fixed in Carnoy’s fixative solution (60% ethanol, 30% chloroform, and 10% glacial acetic acid) and stained with DAPI (10 mg/ml stock diluted 1:50,000 in 1X phosphate-buffered saline; PBS) for 15 min in a humid chamber. Worms were mounted in Prolong Gold with DAPI, cured overnight at room temperature, and stored at 4°C prior to imaging either on a Nikon A1r confocal microscope or Leica Stellaris 5 confocal with an integrated White Light Laser (WLL). Diakinesis images were procured as 0.2 µm per plain Z-stacks and visualized using Volocity 3-D Imaging Software for Nikon images and LAS X for Leica images. Fisher tests were performed to analyze the DAPI-stained chromosome morphology data in diakinesis oocytes. This approach was used to compare the distribution of DAPI-stained bodies, focusing on differences in the frequency of univalents and bivalents. Each experiment was repeated three times and statistical analyses were conducted using R software, with a significance threshold set at p < 0.05.

Worms of specified genotypes were exposed to 10 Gy of γ-irradiation using a ¹³⁷Cs source (Gammacell1000 Elite; Nordion International). Analysis of diakinesis oocytes by DAPI staining was performed as described above at 28 hr post-irradiation.

Gonads from day 1 adults of the appropriate genotypes were dissected in 1X sperm salts (50 mM PIPES pH 7.0, 25 mM KCl, 1 mM MgSO₄, 45 mM NaCl, and 2 mM CaCl₂) with 1 mM levamisole and fixed in 1% paraformaldehyde diluted in 1X sperm salts for 5 min in a humid chamber. Slides were then frozen on a metal block on dry ice for at least 10 min prior to flicking off the cover slip and immersing in 100% ethanol at –20°C for 2 min. Slides were then washed 3X for 10 min each in PBSTB [1XPBS with 0.1% Tween and 0.1% bovine serum albumin (BSA)], and prior to overnight incubation with primary antibody also diluted in PBSTB. Primary antibodies were as follows: mouse anti-GFP 1:500 (Invitrogen Cat. #33-25—Lot 683990A), rabbit anti-phosphoHTP-3^S285 1:2000 (Das et al., 2022), mouse anti-HA 1:1000 (mouse anti-HA-tag (6E2) Cell Signaling Cat. #23675, Lot 5); guinea pig anti-XND-1, 1:10,000 (Wagner et al., 2010). Overnight incubations were performed at 4°C except for anti-HA, which was performed at room temperature. The next day, slides were washed 3X in PBSTB for 10 min each and incubated with Alexa-conjugated secondary antibodies from Molecular Probes: anti-mouse Alexa 488, anti-rabbit Alexa 633, anti-guinea pig Alexa 568 (all diluted 1:2000 in PBST/BSA 0.1%). Incubation was for 2 hr at room temperature in the dark. Slides were then washed 2 × 10 min in PBSTB, and 1 × 10 min with DAPI (10 mg/ml stock diluted 1:50,000 in 1X PBS). Slides were mounted in Prolong Diamond with DAPI and put in the dark to cure overnight before imaging. Images were acquired on a Leica Stellaris 5 confocal and Lightning image processing during acquisition. Z-stacks were acquired with 0.2 micron or smaller step size and visualized in 3D with Leica LAS X software. Each set of experiments was repeated three times.

All confocal microscopy images generated in this study have been deposited in the BioImage Archive under accession number S-BIAD2881 and S-BIAD2907.

cDNAs were amplified from germline-specific cDNA libraries and were cloned by Gibson Assembly into the GAL4 activating domain (pGAD-C1) and binding domain (pGBD-C1)-expressing vectors. Plasmids with the desired cDNA were verified by Sanger sequencing prior to use. Plasmids were co-transformed into the PJ69-4a yeast strain and selected by growth on SC-LEU-TRP medium. Three to five transformants were grown to early log phase (OD₆₀₀ = 0.2), and 5 µl of culture was spotted onto selection plates:

Plates were incubated at 30°C for 72 hr and then imaged. The experiment was performed in triplicate, and empty vectors in the activating and binding domains were included as negative controls. Data are presented as spots on a plate and are from three to four individual transformants per interaction tested and are not individual colonies. The experiment was repeated in triplicate from different transformations.

In order to perform co-immunoprecipitation experiments, synchronized GFP::dsb-1; him-5::3xHA and him-5::3xHA or him-17::3xHA;Phim-5::him-5::GFP::3xFLAG and Phim-1::him-5::GFP::3xFLAG strains were grown until young adult stage (24 hr post-L4) and nuclear protein fractionation was executed as previously shown (Silva et al., 2014). The chromatin-bound fraction was generated upon Benzonase digestion (25 U/100 μl of extract for 1 hr at 4°C). 2 mg of nuclear extract (nuclear-soluble and chromatin-bound fractions were pooled together) was incubated with Agarose GFP traps (Chromotek) or anti-HA Affinity Matrix (Roche) in Buffer D (20 mM HEPES pH 7.9, 150 mM KCl, 20% glycerol, 0.2 mM EDTA, 0.2% Triton X-100 and 1x complete Roche inhibitor) overnight at 4°C. The following day, the agarose beads were recovered by centrifugation at 7500 rpm for 2′ at 4°C and extensively washed with Buffer D at room temperature. After the final wash, the beads were resuspended in 40 μl of 2x Laemmli Buffer and boiled for 10′, after which they were spun at maximum speed for 1′ and the whole eluate was loaded on a precast 4–20% gradient acrylamide gel. Proteins were transferred on a nitrocellulose membrane for 90 min at 4°C at 100 V and blocked for 1 hr at room temperature in 1x TBST containing 5% BSA. Mouse monoclonal anti-HA (Cell Signaling), chicken polyclonal anti-GFP (AbCam), mouse monoclonal anti-FLAG HRP-conjugated (Sigma), and polyclonal anti-XND-1 (Wagner et al., 2010) antibodies were diluted in blocking solution at 1:1000, 1:5000, 1:2000, and 1:2500, respectively, and left to incubate overnight at 4°C. Washes were performed in 1x TBST and anti-mouse, anti-chicken, and anti-guinea pig HRP-conjugated secondary antibodies (Thermo Fisher) were diluted 1:10,000 in 1x TBST containing 5% milk for 1 hr at room temperature. After several washes in 1xTBST, the membrane was incubated with Clarity Max ECL (Bio-Rad) and imaged with a G-Box (Syngene).

IPs were performed on stock AV280: unc-119(e2498) III; him-17(ok424) V; meIs5[him-17::GFP +unc-119(+)]. Preparation and freezing of synchronized worms were carried out as in Cheeseman et al., 2004 except that worms were grown on 8P plates seeded with the E. coli strain NA22 at room temperature. Worm lysis was carried out by grinding worm pellets in liquid nitrogen, and subsequent sonication (Diagenode) was performed with the following settings: 3x 5 cycles on high with 30 s on and 30 s off. Lysate was centrifuged for 45 min at 15,000 rpm at 4°C. To avoid unspecific protein binding, the worm lysate was pre-cleared by incubation to uncoupled agarose beads (Chromotek, bab-20) for 1 hr at 4°C. Pre-cleared worm lysate was incubated with GFP-trap (Chromotek, gta-20) for 3 hr at 4°C. Beads were washed three times with lysis buffer containing proteinase inhibitors (Roche) and seven times with lysis buffer without proteinase inhibitors. Proteins were eluted by adding 40 μl 2x LSB and boiling for 5 min at 95°C. Protein precipitation was carried out using ProteoExtract Protein Precipitation Kit (EMD Millipore, 539180-1KIT) according to the manufacturer’s protocol and submitted for MS analysis to the Taplin Biological Mass Spectrometry Facility. The MS proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository (Perez-Riverol et al., 2022) with the dataset identifier PXD057629 and 10.6019/PXD057629.

GFP pulldowns of HIM-17::GFP for IP/MS were performed in duplicate and corrected against unspecific binding of proteins bound to GFP trap beads. Specifically, potential false positive hits, defined as such because they were also detected in wild-type (N2), AIR-2::GFP and AIR-2KD::GFP pulldowns, were removed from this list. Then, only the remaining hits detected on both independent biological repeats, and for which 2 or more peptides were detected in each experiment, made it to this final table. The only exception is KU-80, which made it to the list after subtraction of the unspecific/false positive hits, but for which only 1 peptide was detected in the December pulldown (14 peptides were detected in the Feb 2014 pulldown). Given that we found KU-70 both times and the coverage/depth was better for the February 2014 pulldown, we decided to keep KU-80 on the final table. Figure 1—figure supplement 2 shows all the hits that remained after subtraction of the unspecific binding proteins.