Aging Negatively Impacts DNA Repair and Bivalent Formation in the C. elegans Germ Line

Defects in crossover (CO) formation during meiosis are a leading cause of birth defects, embryonic lethality, and infertility. In a wide range of species, maternal aging increases aneuploidy and decreases oocyte quality. In C. elegans which produce oocytes throughout the first half of adulthood, aging both decreases oocytes quality and increases meiotic errors. Phenotypes of mutations in genes encoding double-strand break (DSB)-associated proteins get more severe with maternal age suggesting that early meiosis reflects a particularly sensitive node during reproductive aging in the worm. We observed that aging has a direct effect on the integrity of C. elegans meiotic CO formation, as observed by an increase of univalent chromosomes and fusions at diakinesis, with a considerable increase starting at 4 days. We also characterize the possible causes for the age-related changes in CO formation by analyzing both steady-state levels and kinetics of the ssDNA binding proteins RPA-1 and RAD-51. Profound reductions in numbers of both RPA-1 and RAD-51 foci suggests that both DSB formation and early meiotic repair are compromised in aging worms. Using laser microirradiation and γ-irradiation to induce exogenous damage, we show specifically that recruitment of these homologous recombination proteins is altered. Repair defects can be seen in two-and-one-half day-old adults making the loss of germline repair capacity among the earliest aging phenotypes in the worm.


INTRODUCTION
It has been shown that maternal age negatively impacts fertility across species. Particularly in humans, female reproductive capacity declines dramatically after the mid-30 s making it the earliest age-related decline that humans experience (Hassold and Hunt, 2001). A major cause of this reproductive decline is the increasing incidence of oocytes with chromosomal abnormalities, which leads to miscarriage and congenital defects. As more women opt to have children later in life, addressing the issue of female reproductive aging has become increasingly important. Recently, the nematode Caenorhabditis elegans has been developed as a model to study reproductive capacity decline with age (Hughes et al., 2007;Andux and Ellis, 2008;Luo et al., 2010;Quesada-Candela et al., 2021).
During meiosis, crossovers (COs), the physical exchange of genetic material between homologs, together with the action of cohesin proteins, ensure proper alignment and segregation of the chromosomes at meiosis I. Across species, age-related loss of cohesin is associated with increased chromosomal nondisjunction (Hodges et al., 2005;Subramanian and Bickel, 2008). In addition, specific CO configurations, specifically those with too few, too many, or misplaced COs (near centromeres and telomeres), are acutely sensitive to maternal age, although they can be segregated appropriately in younger animals (Hodges et al., 2005;Lamb et al., 2005). Historically, numerous studies documented changes in recombination rates in response to external factors, including maternal age (for example, Stern, 1926;Rose and Baillie, 1979;Hussin et al., 2011). As more chromosome-wide techniques have evolved, changes in crossover positioning have also been reported (Lim et al., 2008;Saini et al., 2020). Nevertheless, our understanding of the factors that drive these age-related changes in CO distribution and frequency is still limited.
In species which produce oocytes throughout adult life, defects in DSB induction and CO repair may further impair oocyte quality in older animals. COs are induced by the formation of meiotic double-strand breaks (DSBs) which are repaired by homologous recombination (HR) with specializations for meiosis. DSBs are induced by the evolutionarily conserved topoisomerase VI-like Spo11 (Keeney et al., 1997) in concert with accessory factors that influence the timing, placement, and extent of break formation. Endonucleolytic cleavage by the MRX/N complex releases Spo11 bound to short oligonucleotides and generates 3 single-stranded DNA (ssDNA) tails (Neale et al., 2005) that are rapidly resected and bound by the ssDNA binding protein complex RPA, protecting them from degradation. RPA is replaced by RAD51 and/or DMC1 which form (Gasior et al., 1998) filaments then can invade the homologous double stranded-DNA of either the sister chromatid or homologous chromosome. This single end-invasion can go on to form double Holliday Junctions (dHJ) (Shinohara et al., 1992;Sung, 1994), a subset of which are resolved to form COs (Schwacha and Kleckner, 1995). COs, together with cohesion, connect the homologous chromosomes, enabling their biorientation toward opposite spindle poles in metaphase I (reviewed in Sansam and Pezza, 2015).
The events of meiosis are tightly regulated to ensure both that each chromosome attains a crossover and that excessive DSBs are repaired and not passed along to offspring. Accordingly, non-HR pathways can step to repair DSBs if HR is ineffective. These include the classical non-homologous end joining (cNHEJ) pathway, microhomology-mediated end jointing (MMEJ), and single-strand annealing (SSA) (Macaisne et al., 2018). In C. elegans, DSB-promoting factors HIM-5 and CEP-1/p53 actively repress NHEJ functions allowing preferential use of HR (Mateo et al., 2016), intimating that alternative pathways can be available for DSB repair in early meiosis and are not simply late-acting salvage pathways that eliminate residual breaks.
Aging has a direct effect on the integrity of C. elegans meiotic CO formation as an increase in achiasmate chromosomes at diakinesis was observed at day 5 of adulthood, half-way through the worm's reproductive span (Luo et al., 2010;Meneely et al., 2012;Stamper et al., 2013;Achache et al., 2020). Also, it has been reported that maternal age exacerbates CO defects associated with mutations in SPO-11 accessory factors including dsb-2 and him-5 (Meneely et al., 2012;Stamper et al., 2013). This suggests that either the DSB break formation or repair machineries (or both) might be inherently sensitive to maternal age. However, the underlying cause(s) of this aging phenotype is not understood. In this study we tested the hypotheses that DSB formation and repair are impaired in aging. We analyzed the effect of maternal age on the recruitment of different HR proteins. To address this, we utilized two irradiation techniques (laser microirradiation and γ-irradiation) to induce exogenous damage in meiotic nuclei of young and aged worms Koury et al., 2018;Mahadevan et al., 2019). We observed that the total numbers of RPA-1 and RAD-51 foci were reduced by maternal age which results from both a decrease in HR-proficient DSBs and attenuated capability to recruit RPA-1 and RAD-51.

Strains and Genetics
All strains were derived from the wild-type Bristol strain N2 and were cultivated at 20 • C under standard conditions. Abbreviated names and full genotypes of the strains used in this study are listed in Table 1.

Chromosome Morphology Analysis of Diakinesis Oocytes
The numbers of DNA bodies present in diakinesis oocytes were assessed in intact adult hermaphrodites at days 1 through 10 of adulthood. Animals were picked as L4 larval and transferred each day until the cessation of egg-laying. Adults were fixed in Carnoy's fixative solution (60% ethanol, 30% chloroform, and 10% glacial acetic acid) and stained with 4 ,6-diamidino-2-phenylindole (DAPI) (10 mg/ml stock diluted 1:50,000 in 1X Phosphate-buffered saline (PBS) for 15 min. in a humidity chamber. Stain was removed and worms were mounted in Prolong Gold with DAPI, cured overnight at room temperature, and stored at 4 • C prior to imaging on a Nikon A1r confocal microscope. Diakinesis images were procured as 0.2µm per plain Z-stacks and visualized using Volocity 3-D Imaging Software (PerkinElmer Corp, now owned by QuorumTechnologies).

Immunostaining and Analysis of RAD-51 Foci
Gonads were dissected from 1-1 and 4-days adults in 1X sperm salts (50 mM PIPES pH 7.0, 25 mM KCl, 1 mM MgSO 4 , 45 mM NaCl, and 2 mM CaCl 2 ) 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 methanol at −20 • C for 2 min followed by 5 s in acetone at 4 • C. Slides were then washed in PBSTB [1XPBS with 0.1% Tween and 0.1% bovine serum albumin (BSA)], and incubated *The unc-119 gene is used as a selection marker for the transgenes. Prior studies of the cosa-1 strain suggest it does not appear to interfere with crossover formation (Yokoo et al., 2012).
overnight at 4 • C with primary antibody: rabbit anti-RAD-51 generated against peptide ASRQKKSDQEQRAA by Genscript for the Smolikove lab (Alleva et al., 2019), 1:30,000 in PBSTB). The next day, slides were washed 3X in PBSTB for 10 min and incubated with secondary antibody (goat anti-rabbit Alexa 568, 1:2,000 in PBSTB) for 2 h at room temperature in the dark. Slides were then washed 2 X 10 min in PBSTB, and 1 X 10 min with DAPI (10 mg/ml stock diluted 1:50,000 in 1X PBS). Slides were mounted in Prolong Gold with DAPI and put in the dark to cure overnight before imaging. Three-dimensional (3D) images of the whole germ lines were taken using a Nikon A1r confocal microscope with 0.2 µm step size and analyzed using Volocity 3D software (PerkinElmer). For wild-type worms, we divided the leptotene, zygotene, and pachytene regions into six zones and counted RAD-51 foci in every nucleus for a minimum of three germ lines/age. For rad-54 mutants, we counted RAD-51 foci in late pachytene nuclei for at least three germ lines/age. For irradiated worms, 15 midpachytene nuclei were analyzed.

Analysis of RPA-1 Foci
Gonads of gfp::rpa-1; spo-11 worms were dissected from 1-to 4days adults in M9 with 1 mM levamisole. Slides were then freezecracked and immersed in cold 100% EtOH for 2 min followed by 20 min in 4% paraformaldehyde/M9 in a humidity chamber. Slides were then washed in PBSTB 1×10 min, incubated with DAPI 1×10 min and washed again. After a quick dry, slides were mounted in Prolong Gold with DAPI.
Analysis of GFP::COSA-1 Foci gfp::cosa-1 and irradiated gfp::cosa-1; spo-11 worms were dissected in 1X sperm salts as described above. Slides were immediately freeze-cracked and immersed in 100% cold ethanol for 1 min, and fixed in 4% paraformaldehyde/1X sperm salts again for 10 min. Slides were washed 2X10 min in PBSTB, stained with DAPI in 1XPBS for 10 min followed by one wash with PBSTB for 10 min. Slides were mounted in Prolong Gold with DAPI. Images were acquired and analyzed as described above. GFP::COSA-1 foci in late pachytene nuclei (4 rows of nuclei before diplotene) were counted in at least five germ lines/age.

Western Blotting
For each age, 300 worms were transferred into a glass conical tube and washed thrice with 1X M9 buffer (22 mM KH 2 PO 4 , 42 mM Na 2 HPO 4 , 85.5 mM NaCl, 1 mM MgSO 4 ). The remaining liquid was removed, and the worm pellet was transferred to a 1.5 ml tube and flash frozen in liquid nitrogen. Pellets were thawed on ice, mixed with an equal volume of Laemmli Sample Buffer (Bio-Rad #161-0737) with 5% b-mercaptoethanol (Amresco M131), sonicated in a water bath for 2 min, heated at 95 • C for 10 min, then spun in a tabletop centrifuge for 5 min at maximum speed. Samples were resolved by 12% PAGE (TGX FastCast, Bio-Rad) and transferred to nitrocellulose in 20% ethanol. The membrane was blocked in 5% non-fat milk/TBST (50 mM Tris-HCl pH 7.4, 150 mM NaCl, and 0.1%Tween-20) for 1 h at RT and incubated in OLLAS-tag antibody, pAb, rabbit (GenScript, A01658, 1µg/ml in 5% milk/TBST) or mouse α-E7 (tubulin, Developmental Studies Hybridoma Bank, 1:2,000 in 5% milk/TBST) ON at 4 • C. The next day the membranes were washed in TBST and incubated in α-rabbit-HRP or α-mouse-HRP (1:2,000 in 5% milk/TBST) for 2 h at room temperature. Products were visualized by enhanced chemiluminescence (ECL) according to the manufacturer's instructions.

γ-Irradiation
1-and 4-days adult worms were exposed to 10Gy of γ-irradiation using a 137 Cs source (Gammacell1000 Elite; Nordion International). RAD-51 immunostaining was performed as described above on gonads dissected and fixed at 25 min, 4 and 8 h post-IR. RPA-1 detection was assessed at 25 min and 4 h post-IR and GFP::COSA-1 analysis were performed at 8 h post-IR.

Laser Microirradiation, Live Imaging, and Fixation
We followed the protocol outlined in Harrell et al. (2018) with the following modifications: for the live imaging experiments, five worms were imaged at a time at 2 min intervals for 45 min. Z stacks of 10 images at 1µm stages were taken at each interval. For the fixed sample experiments five worms were placed on a live imaging slide, microirradiated, recovered to an NGM plate until each specified time point, then placed in M9 on frosted microscope slides. The M9 was removed and the worms were fixed in 100% EtOH and mounted with VectaShield.
Endogenous RPA-1 Intensity Measurements gfp::rpa-1 adult worms were grown to 1, 2, 3, and 4 days old. Their gonads were dissected on charged slides, frozen at −80 • C, and placed in −20 • C EtOH for 1 min then mounted with 10 µl M9-DAPI and VectaShield. Whole intact gonads were imaged on the DeltaVision wide-field fluorescence microscope (GE Lifesciences) in seven 512 × 512 pixel zones with 100x/1.4 NA oil Olympus objective from the proximal pre-meiotic tip to the distal late pachytene phase of prophase I. Each image was the center slice of the nuclei in the upper section of the syncytial tube in both the DAPI (blue) and FITC (GFP) fluorescent channels. Measurement of the intensity of each nucleus was taken in FIJI in each zone in the FITC channel and corrected to cytoplasmic background.

Counting Recruitment Regions
Recruitment regions were counted in FIJI by visual inspection of each Z-slice of each time interval. A recruitment region was determined to be in response to microirradiation if it appeared in a location that did not contain a localization of RPA-1 fluorescence in the initial control image. Clusters were differentiated from foci when more than one distinct point of fluorescence was present but not separate from each other.

Statistical Analyses
All statistics were performed using GraphPad Prism 9. For all pairwise comparisons presented, the Mann-Whitney U-test was performed. For all binary data comparisons, the Fisher's Exact test of independence was used.

Ethics Statement
No studies in this paper required approval by institutional review boards.

Aging Increases Meiotic Defects in C. elegans
Multiple lines of evidence indicate that fertility in C. elegans declines with maternal age (Hughes et al., 2007;Luo et al., 2010). In both unmated and mated hermaphrodites, days 4-5 of adulthood marks a decline in reproductive capacity that is accelerated in the unmated animals due to lack of sperm (Luo et al., 2010). This timepoint is marked by increases in both inviable eggs and male offspring (Hughes et al., 2007;Luo et al., 2010) which can both be explained by the concomitant increase in CO errors seen in diakinesis oocytes. In order to determine when CO defects are first observed, we analyzed the integrity of chromosomes in N2 diakinesis nuclei throughout adulthood. In young adult animals (day 1), six bivalents were detected at diakinesis by fluorescence microscopy, corresponding to the six pairs of homologous chromosomes each held together by a chiasma (Figures 1A,B). A small fraction of nuclei (4%) showed 4 bivalents and 1 larger mass of DNA (Figure 1D), a configuration that arises in repair mutants and corresponds to a chromosomal fusion (Martin et al., 2005) but may result from two closely apposed chromosomes (for simplicity, we refer to these as fusions). Since C. elegans chromosomes are holocentric, fusion chromosomes would be stably transmitted during meiosis and the subsequent mitotic embryonic divisions, explaining the near 100% hatching rates of eggs. The small fraction of fusions was maintained in 2and 3-days-old adults. However, after this timepoint, nuclei with five bivalents and two univalents were also apparent, indicating that one pair of chromosomes failed to achieve a CO. As the worms aged, meiotic defects worsened. In 4-day-old adults, ∼15% of the analyzed diakinesis showed nuclei with achiasmate chromosomes and ∼14% exhibited fusions. Since day 4 of adulthood corresponds to the time when meiotic errors increase substantially ( Figure 1A) and when egg production begins to decrease (Hughes et al., 2007;Luo et al., 2010), we chose this timepoint as the "aged worms" for the remainder of the experiments described herein.
After day 4, meiotic errors continue to increase until reproductive senescence sets in at day 10. Curiously, older animals (>6-days-old) showed some irregular bivalents where the homologous chromosomes appeared to a be separating from one another ( Figure 1C). Similar separating bivalents were recently reported for aged worms grown at 25 • C, and the distinct univalent masses each contained the crossover marker, COSA-1 (Achache et al., 2020). At timepoints after egg-laying has ceased, days 10 and day 13, there remain a surprisingly large fraction of diakinesis oocytes with 6 DAPI bodies meaning that chromosomal abnormalities are not driving reproductive senescence.
To corroborate the impact of age on CO formation, we also examined GFP::COSA-1 foci, a cytological marker of CO commitment (Yokoo et al., 2012). As shown in Figure 1D, in 1 day-old adults, 5-6 COSA-1 foci are seen in the last 4 rows of nuclei prior to diplotene, consistent with prior studies (Yokoo et al., 2012;Machovina et al., 2016;Woglar and Villeneuve, 2018). By contrast, in day 4 adults, nearly 20% of nuclei had only 3 or 4 COSA-1 foci indicative of an impairment in CO formation. This decrease in COSA-1 foci could result from the combined effects of fewer DSBs, altered repair kinetics, as well as impaired recruitment of CO designation and commitment factors. Overall, these data support the conclusion that fewer COs are observed in older animals.

The Levels of Homologous Recombination Competent Double-Strand Breaks Decrease With Maternal Age
The increase in univalent (non-CO chromosomes) with maternal age could be explained by the decrease of meiotic-specific DSB Frontiers in Cell and Developmental Biology | www.frontiersin.org formation. Therefore, to determine if DSB formation is itself age-sensitive, we took advantage of RAD-51 immunostaining which marks early intermediates of homologous recombination repair, forming foci. Prior studies had shown that greater than 95% of meiotic DSBs can be marked with RAD-51 (Mets and Meyer, 2009). In day 1 adults, we observed the RAD-51 localization pattern expected for wild type C. elegans meiosis: RAD-51 foci began to appear during leptotene/zygotene, became numerous by mid-pachytene, and disappeared by late pachytene (Figures 2A,B), indicative of efficient DSB repair (Alpi et al., 2003;Mateo et al., 2016). On day 4 of adulthood, the same relative pattern was observed, although the total number of foci was reduced in each region of the early prophase germ line ( Figure 2B). This was seen most strikingly by the increase in the number of nuclei in the mid-pachytene regions that have zero RAD-51 and the decrease in nuclei with > 6 RAD-51 foci. A decrease in steady-state RAD-51 levels was also reported (Achache et al., 2020). In spo-11 mutant animals where meiotic DSBs are not made, few to no breaks are seen in Day 1 animals whereas a single RAD-51 focus can be seen in up to half of pachytene nuclei (Supplementary Figure 1), suggesting that if anything, the RAD-51 observed in wild-type 4 day-old adults is an over-representation of meiotically-induced breaks. These results suggest either that DSB numbers decrease with age, or that DNA repair kinetics change with age, or both.
Because DSBs are not made simultaneously and because RAD-51 is removed from the break sites as DSBs mature into COs, a static snapshot of the germ line cannot report the total number of DSBs that are made and undergo CO recombination. To assess total DSB numbers, we took advantage of the rad-54 mutant in which HR is stalled at a step after the recruitment of RAD-51 to resected DSBs, effectively trapping RAD-51 at all of the sites undergoing HR repair (Meneely et al., 2012). Consistent with prior studies, we observed that nearly 50% of day 1 nuclei accumulated between 20 and 40 RAD-51 foci, more than enough DSBs to ensure a CO on each chromosome ( Figure 2C). In contrast, on day 4 of adulthood, > 95% of the nuclei analyzed showed fewer than 20 RAD-51 foci, indicating that the total number of HR-competent breaks decreased with maternal age (Figure 2C). We previously showed that the him-5(ok1896) null mutant which fails to achieve COs on the X chromosome contains an average of 9.9 RAD-51 foci/nucleus (Meneely et al., 2012). By inference, the nuclei in older animals with fewer than 10 foci are likely to experience deficits in CO formation.

Homologous Recombination Is Less Efficient in Older Adult Germ Cells
Since the studies above suggest that DSB levels are decreased in older animals and/or may be repaired by non-HR mechanisms, we wanted to directly analyze HR efficiency in meiotic germ cells. To do this, we took advantage of a previously reported assay to examine repair kinetics after γ-radiation (IR)-induced DSBs (Li and Yanowitz, 2019). Post-IR, we assayed the dynamics of RPA-1 and RAD-51. RPA-1 is the ssDNA-binding protein that is the first to coat the resected DNA (Caldwell and Spies, 2020). Importantly, RPA-1 levels do not decrease with maternal age (Supplementary Figure 2B). Like RAD-51, RPA-1 is observed as chromatin-associated foci in fixed samples. These studies were performed in the spo-11 mutant background to eliminate the signal from endogenous meiotic breaks. Previous work determined that 10Gy IR is able to produce ∼20 DSBs per meiotic nucleus, which is sufficient to ensure a CO on each chromosome and is near the endogenous number of meiotic DSBs per nucleus (Mets and Meyer, 2009).
At the earliest time point post-IR (25 min), 99% of the day 1adult, pachytene nuclei recruited RPA-1 to at least one site, with an average of ∼4.3 foci/nucleus. In contrast, 64% of day 4-adult nuclei failed to recruit RPA-1. Although at 4 h post-irradiation both groups of worms showed elevated numbers of RPA-1 foci, the recruitment of RPA-1 was significantly lower in older animals ( Figure 3A and Supplementary Figure 3). To verify that the expression of GFP::RPA-1 is not reduced over time, thus leading to the reduction on RPA-1 recruitment, we performed measurements of the intensity of the nuclear GFP signal in young adults and 2, 3, and 4 day adults. These experiments were performed on fixed samples to allow analysis at higher resolution. We observed no decrease in overall GFP signal throughout aging (Supplementary Figure 2). Together these results suggest that aged animals are defective in recruitment of RPA-1 to DSB sites.
Similar defects in initial recruitment to DNA lesions are observed for RAD-51. At 25 min post-IR, significantly fewer RAD-51 foci are seen in day 4 compared to day 1 adults. RAD-51 foci are diminished in both groups at 4 and 8 h post-IR ( Figure 3B). However, RAD-51 loss appears slower in 4-day-old adults, as more foci were observed at 8 h post-IR in the day 4 compared to the day 1 adults. While we do not understand the basis for the differences in the patterns of RPA-1 and RAD-51, both sets of data point to altered kinetics of HR repair in older animals.
We also quantified the ability of IR-induced breaks to be converted to CO-committed intermediates by quantifying the recruitment of the COSA-1 to chromosomal foci in late pachytene nuclei (Yokoo et al., 2012). As shown in Figure 3C, we observed a remarkable reduction in GFP::COSA-1 foci in the aged animals 8 h post-IR. At this timepoint, nearly half of day 1 nuclei achieved 6 CO-competent foci, whereas almost no day 4 nuclei experienced full COSA-1 levels ( Figure 3C). Together with our observations that fewer early HR-intermediates are observed in the older animals, these results suggest that the ability to undergo CO recombination is compromised in older animals.
In addition to HR-mediated repair pathways, DSBs can also be repaired by other mechanisms including classical nonhomologous end joining (cNHEJ), micro-homology mediated end joining (MMEJ), single-strand annealing (SSA). Breakinduced replication (BIR) has not been shown to work in meiosis. We previously showed that in the absence of RAD-51 proteins, SSA, NHEJ, and MMEJ all contribute to repair in the meiotic germ line and that HIM-5 biases use of SSA over NHEJ and MMEJ, suggesting that active mechanisms inhibit use of these alternative DSB repair pathways (Macaisne et al., 2018). We therefore hypothesized that some of the abnormal chromosomal patterns that we observe in diakinesis oocytes from older animals might be explained if these non-HR pathways were to be used more frequently. To test this, we examined the diakinesis figures in cku-70 and polq-1 mutant animals that are impaired in NHEJ and MMEJ, respectively. As shown in Supplementary Figure 4, impairing NHEJ or MMEJ functions has little to no effect in FIGURE 3 | Dynamics of RPA-1 and RAD-51 are altered in aging, meiotic germ cells. spo-11 mutant worms at day 1 and day 4 of adulthood were exposed to 10Gy of γ-irradiation. (A) Average number of GFP::RPA-1 or B) or RAD-51 in mid-pachytene nuclei at indicated timepoints post-exposure. A total of 10 worms per age/timepoint were analyzed in gfp::rpa-1; spo-11 (A) or spo-11 (B) worms. (C) Percent of late pachytene nuclei containing the indicated number of COSA-1 foci in gfp::cosa-1, spo-11 worms at 8 h post-exposure. Statistical significance was determined using Chi-square test. ns p > 0.1, * p < 0.05, * * * * p < 0.0001. young animals. In contrast, in older animals loss of cku-70 or polq-1 increases the number of nuclei with 6 DAPI bodies. Thus, at least a subset of the chromosome morphology defects observed in diakinesis of older animals is explained by the aberrant activation of NHEJ or MMEJ.

Repair of Complex DSBs Is Severely Impaired in the Aging C.elegans Germ Line
Next, we tested whether the age-dependent defects in recruitment of RPA-1 could be aggravated when the DNA damage formed is more complex. Clustered DSBs are forms of complex DNA damage that are more challenging to repair (Nickoloff et al., 2020). Laser microirradiation induces targeted and complex DNA damage that recruits HR proteins and forms characteristic, clustered DSBs . To examine RPA-1 localization to DSBs in live worms, we utilized the GFP11 tagged RPA-1 strain  described above. In this strain endogenous RPA-1 foci are dim and can be clearly separated during analysis from microirradiation induced foci (Hefel et al., 2021).
To test if RPA-1 recruitment to complex DSBs is impaired with aging we exposed worms to microirradiation at days 1, 2, 2.5, 3, and 4 of adulthood and imaged mid-pachytene (MP) nuclei for 45 min (at 2 min intervals) post-exposure. This time frame was determined in our previous studies where we showed that most RPA-1 foci are recruited within the first 30 min in this strain (Hefel et al., 2021). RPA-1 is recruited to microirradiation induced breaks in a pattern of clusters and foci that cannot be resolved by live imaging  and is collectively referred to here as recruitment regions. Although the number of recruitment regions was no different between 1-and 2-day-old adults, recruitment in the 45 min window was impaired in 2.5-day-old adults and completely abrogated in 3-and 4-day-old adults (Figure 4). To examine if the defects in RPA-1 recruitment were unique to the MP region of the germ line, we performed microirradiation experiments in the other germline regions. We observed similar reductions in RPA-1 recruitment throughout the germ line of aged animals ( Figure 4C). These data altogether indicated that despite accumulating in the nucleus, RPA-1 fails to assemble on clustered DNA damage in the same kinetics in aged and nonaged worms.
Defects in recruitment of proteins following microirradiation can reflect an abrogation of RPA-1 recruitment or a delay in RPA-1 recruitment to clustered DSBs. Since live imaging data can only be acquired in a limited time window (due to expiration of the worms), analysis in large time scales can only be performed on fixed samples. For this analysis, we microirradiated worms, and then performed gonad dissections at 12 and 25 min, and 4 h timepoints post-microirradiation. We first quantified the percent of nuclei with RPA-1 recruitment regions out of the total number of microirradiated nuclei. 1-and 2-day old adults exhibited maximal recruitment of RPA-1 at the 25 min time point which declined slightly at 4 h (Figures 5A-C). In agreement with our live-imaging data, RPA-1 recruitment to microirradiated nuclei was severally impaired in 3-day-old adults and completely abrogated in 4-day-old adults. These data indicate that aged worms fail to recruit RPA-1 to complex DSBs. Aged worms also exhibited an increase in the proportion of foci over clusters (Figures 5D-F) which is consistent with inhibition of RPA-1 accumulation on microirradiation-induced DNA damage. Aged worms showed a small increase in number of foci per nucleus in later timepoints (Figure 5I, compared to 5G,H). Since for this analysis only nuclei with RPA recruitment regions were scored, this indicates that in the few nuclei in which RPA-1 succeeded in loading to DSBs RPA-1 did not unload in a timely manner.
Altogether our data indicated that, as observed with γ-irradiated worms, microirradiated worms also exhibit defects in RPA-1 recruitment. These defects were more severe in the microirradiated worms, likely due to the complexity of microirradiation induced damage that is more challenging to repair.

DISCUSSION
While it has been appreciated for decades that oocyte quality diminishes with maternal age, our understanding of the underlying deficits that arise in aged germ cells is still incomplete. We show here that the meiotic processes that control DNA repair and crossover formation deteriorate progressively with maternal age in C. elegans. As early as two and one-half days into adult life, just less than one-third of the way through its maximum reproductive period, wild-type worms have profound changes in their ability to recruit DSB-repair proteins to DNA lesions. These defects are the among the earliest harbingers of adult aging in the worm, preceding morphological and physiological somatic changes by several days (Son et al., 2019). In this regard, the oogenic germ line of the adult hermaphrodite may be similar to a human ovary, which also undergoes aging before the rest of the tissues in the body (Hassold and Hunt, 2001).
Since oocytes are made continuously in the worm, we were able to examine how meiotic processes changes in response to age. Using RAD-51 as a surrogate for DSBs, we observed a substantial decrease in HR-competent DSBs in aged germ lines. Since RAD-51 is downstream of resection, a decrease in focus formation could be explained by a bona fide decrease in DSB numbers and/or may reflect that DSBs are steered into non-HR (non-CO) pathways in older animals. Our data support defects in both DSB formation and in repair pathway choice.
We reasoned that if DSB efficiency did not decrease with age and DSB levels remained high, then to explain the observed decrease in RAD-51 foci, increasing numbers of chromosomes would be expected undergo non-HR repair. In fact, since most nuclei in day 4 adults have fewer than 10 RAD-51 foci (Figure 2B), this would mean that 10-20 DSBs would need to be repaired by alternative pathways. Not only would fewer bivalents be observed, but many larger fusions (and possibly fragments as well) would be seen at diakinesis. This is not what we observed. It was not until the cessation of reproductive life, that more than one fusion chromosome was observed ( Figure 1A). Instead, > 90% of nuclei appear to repair DSBs as intact, bivalent-like or univalent-like chromosomes throughout the reproductive span. Thus, we infer that substantially fewer DSBs are made as the animal ages. In support of this conclusion, when we trapped RAD-51 protein on the ssDNA filaments with the rad-54 mutation, RAD-51 levels were significantly reduced. Additionally, the preponderance of bivalent-like chromosomes in the older animals suggests that most chromosomes received a crossover despite low DSB levels, consistent with our previous data on CO distribution in aging animals (Lim et al., 2008). This data suggests that crossover homeostasis mechanisms (Martini et al., 2006) remain proficient in older animals. The reduction in DSBs that we observe either could be a result of an abrogation of SPO-11 activity or could reflect impairments in crossover feedback control mechanisms. The ATM-1 and ATR/ATL-1 kinases are activated by DSBs and downstream ssDNA intermediates and function in a regulatory feedback loop to retain DSB competency until a CO competent intermediate is achieved on all chromosomes  (Stamper et al., 2013;Lukaszewicz et al., 2018;Li and Yanowitz, 2019). One possibility is that these signaling pathways could be impaired in aging animals to prevent the maintenance of DSB competency.
The increase in fusion chromosomes in aged diakinesis nuclei supports the notion that breaks are being diverted into non-HR repair pathways ( Figure 1A). The impairment in recruitment of both RPA-1 and RAD-51 after IR and microirradiation also supports the notion that very early events in DSB repair are affected by age. In particular, these studies point to resection as the likely step in repair that is agesensitive, as processing of the DSB ends regulates both the competition between repair pathways and the ability to recruit RPA-1 (Marini et al., 2019). Chromosome fusions result from repair through NHEJ, MMEJ, or SSA. On day 1 of adulthood, HIM-5 and CEP-1/p53 proteins redundantly inhibit NHEJ during meiosis (Meneely et al., 2012). Although the CO defects of him-5 worsen with maternal age (Meneely et al., 2012), chromosome fusions were not observed, suggesting that the CEP-1 dependent block to NHEJ is still intact in older animals.
The proteins that prevent repair via MMEJ and SSA are currently unknown, so we can only postulate that one or both of these pathways may be accessed to induce the fusions see in aged nuclei.
HR-mediated repair is also severely compromised in aged germ cell nuclei. Using both γ-irradiation and microirradiation, we see significant deficits in the recruitment, accumulation, and turnover of both RPA-1 and RAD-51 at DSB sites (Figures 3-5). One caveat to these studies is that we do not know whether old and young worms' chromosomes are equally responsive to irradiation. However, since the nuclear envelope integrity deteriorates (Haithcock et al., 2005;D'Angelo et al., 2009) and chromatin decompacts in older animals (reviewed in Purohit and Chaturvedi, 2017;Sławińska and Krupa, 2021), we would anticipate that DNA damage would increase rather than decrease in older nuclei (Hernández et al., 2015). Further studies on the accessibility of chromatin to DNA damage and developing assay to directly monitor breaks in this system are an avenue of future studies. This impairment is seen in nuclei from the mitotic tip through late pachytene suggesting it reflects an overall loss of germ line capabilities with age. While this study has been focused on meiotic DNA repair, we note that the effects observed in the mitotic tip of the worm germ line may elucidate how aging impacts repair in replicative cells, including but not limited to the mammalian male germ cells which are maintained by continuous mitotic divisions. Although the general trends between IR and microirradiation are remarkably similar, there are notable differences particularly in regard to RPA-1 accumulation and turnover (Figures 3, 5A). This suggests that the source of DNA damage can dramatically change the persistence of repair intermediates and possibly the outcomes of repair. Microirradiation results in more complex form of DNA damage (clustered DSBs) that are more challenging to repair than dispersed DSBs, like the ones formed by γ-irradiation (Nickoloff et al., 2020). If repair processes are more challenging in older worms, microirradiation may result in more severe effects on recruitment of repair proteins to DSBs, compared to γ-irradiation. Further studies will be required to understand the molecular differences between these repair pathways and how they deteriorate with maternal age.
In the oldest oocytes (>day 6), we observed a novel appearance of some bivalents where the distance between univalents was increased. This separation of the bivalent might occur if, like in mammals and flies, cohesion between exchange homologs is lost with maternal age (Hodges et al., 2005;Subramanian and Bickel, 2008). Additional studies will be required to determine whether cohesin proteins expression or turnover is indeed impacted by maternal age in C. elegans and whether there is a functional consequence for the proper segregation of these bivalents.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
MR performed the studies in Figures 1-3, Supplementary  Figures 1, 3,   Supplementary Figure 3 | Percentage of middle pachytene (MP) nuclei containing the indicated number (key at far right) of GFP::RPA-1 foci (left) or RAD-51 foci (right) in spo-11 at timepoints post-10Gy of γ-irradiation. Numbers on top of bars indicate the total number of nuclei analyzed per age. A total of 10 worms per age/timepoint were analyzed. Statistical significance was determined using Chi-square test. ns p > 0.1, * p < 0.05, * * * * p < 0.0001.
Supplementary Figure 4 | Quantification of DAPI-bodies in cku-70 and polq-1 diakinesis nuclei from the indicated ages. Data analyzed in Figure 3 for N2 worms of the same ages was added to allow the comparison with the mutants. Numbers on top of bars indicate the total number of nuclei analyzed per age.