Replication Protein A1 is essential for DNA damage repair during mammalian oogenesis

Persistence of unrepaired DNA damage in oocytes is detrimental and may cause genetic aberrations, miscarriage, and infertility. RPA, an ssDNA-binding complex, is essential for various DNA-related processes. Here we report that RPA plays a novel role in DNA damage repair during postnatal oocyte development after meiotic recombination. To investigate the role of RPA during oogenesis, we inactivated RPA1 (replication protein A1), the largest subunit of the heterotrimeric RPA complex, specifically in oocytes using two germline-specific Cre drivers (Ddx4-Cre and Zp3-Cre). We find that depletion of RPA1 leads to the disassembly of the RPA complex, as evidenced by the absence of RPA2 and RPA3 in RPA1-deficient oocytes. Strikingly, severe DNA damage occurs in RPA1-deficient GV-stage oocytes. Loss of RPA in oocytes triggered the canonical DNA damage response mechanisms and pathways, such as activation of ATM, ATR, DNA-PK, and p53. In addition, the RPA deficiency causes chromosome misalignment at metaphase I and metaphase II stages of oocytes, which is consistent with altered transcript levels of genes involved in cytoskeleton organization in RPA1-deficient oocytes. Absence of the RPA complex in oocytes severely impairs folliculogenesis and leads to a significant reduction in oocyte number and female infertility. Our results demonstrate that RPA plays an unexpected role in DNA damage repair during mammalian folliculogenesis.


41
As the largest cell in the body of mammals, the oocyte is crucial for female fertility and early 42 embryogenesis. To acquire maturity and competency for fertilization and subsequent embryonic 43 development, female germ cells undergo a series of developmental processes. In mammals, following 44 premeiotic DNA replication, oocytes enter an extended prophase of meiosis I during fetal development.

45
In mice, oocytes progress through leptotene and zygotene stages, and reach the pachytene stage by 46 embryonic day 18.5 (E18.5) (Borum, 1961;Pepling, 2006). At postnatal day 5 (P5), the majority of oocytes 47 reach and arrest at the late diplotene (or dictyate) stage, with individual oocytes contained in the 48 primordial follicles where they remain arrested until ovulation (Cohen et al., 2006;Pepling and Spradling, 49 2001). In mice, cohorts of primordial follicles are periodically recruited to enter a 3-week growth phase, 50 progressing through primary and secondary follicles (Amleh and Dean, 2002). After puberty, these pre-51 antral follicles are further stimulated by follicle-stimulating hormone (FSH), leading to the formation of 52 antral follicles (Barnett et al., 2006;Kumar et al., 1997). Throughout these stages, oocytes remain 53 arrested at the diplotene or germinal vesicle (GV) stage due to their interactions with the surrounding 54 granulosa cells (Coticchio et al., 2015). On the other hand, oocytes also control ovarian follicular 55 development and the proliferation of granulosa cells (Eppig, 2001). Finally, triggered by the luteinizing 56 hormone (LH) surge, fully grown immature oocytes are released from the GV stage arrest, resume 57 meiosis, and progress through metaphase I (MI) and metaphase II (MII), leading to ovulation (Richards

61
Because the prolonged arrest of mammalian oocytes at the diplotene or GV stage can last for months in 62 mice and even decades in humans, there is a potential for the accumulation of DNA damage during this 63 extended period (Carroll and Marangos, 2013). Alongside single-strand breaks (SSBs) and double-strand 64 breaks (DSBs), other types of DNA lesions, such as base mismatches and alterations in DNA structure, 3 2008; Suh et al., 2006); 3) Fully grown GV oocytes in antral follicles exhibit a weakened G2/M checkpoint, 74 allowing oocytes with DNA damage to still progress into the M phase due to a limited ability to activate 75 the ataxia telangiectasia mutated (ATM) kinase (Marangos and Carroll, 2012) and/or a delayed response 76 to the DNA damage (Subramanian et al., 2020); 4) Oocytes with DNA damage that enter the M phase 77 normally arrest at MI due to the activation of the spindle assembly checkpoint (SAC) (Collins et al., 2015).

79
During DNA replication and transcription, the double-stranded DNA (dsDNA) must be unwound and 80 separated to act as a template for DNA or RNA synthesis. This unwinding exposes single-stranded DNA 81 (ssDNA) regions, which are more susceptible to DNA damage factors and can accumulate DNA lesions 82 (Saini and Gordenin, 2020). To safeguard and stabilize the exposed ssDNA, cells employ various ssDNA 83 protecting complexes. Among these, the replication protein A (RPA) complex, composed of RPA1, RPA2, 84 and RPA3, is the primary ssDNA-binding heterotrimeric complex involved in numerous DNA metabolic 85 pathways, including DNA replication, recombination, repair, and DNA damage checkpoints (Wold, 1997; 86 Zou et al., 2006). RPA binds ssDNA with high affinity, thereby preventing the formation of secondary 87 structures and protecting ssDNA from the action of nucleases. During these processes, RPA also directly 88 interacts with other DNA processing proteins (Deng et al., 2015). For instance, during homology-directed 89 repair of DSBs, ssDNA tails are initially bound by RPA, which is then replaced by RAD51 recombinase 90 (Stauffer and Chazin, 2004). During DNA damage checkpoints, RPA is required for localization of ATR 91 (ataxia telangiectasia mutated and Rad3-related) kinase to the DNA damage sites and for activation of 92 ATR-mediated phosphorylation of the downstream targets (Zou and Elledge, 2003). In addition to its role

99
While RPA has been investigated in yeast and cell lines, its role in vivo during mammalian development 100 remains largely unexplored due to the early embryonic lethality resulting from Rpa1 mutation in mice 101 (Wang et al., 2005). Although RPA has been demonstrated to be crucial for meiotic recombination in 102 mouse spermatogenesis (Shi et al., 2019), its role in female germ cell development has yet to be 103 elucidated. In this study, we generated conditional knockouts of Rpa1 in oocytes using two distinct Cre 104 drivers (Zp3-Cre and Ddx4-Cre). Surprisingly, we discovered that RPA plays an essential role in DNA 105 damage repair after meiotic recombination and is required for follicular development during oogenesis.

106
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made   was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made     166 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

176
Follicle quantification was carried out following the previously described protocol (Myers et al., 2004).

177
Ovaries were sectioned at 8-µm intervals and subjected to hematoxylin and eosin staining. In each mouse,

222
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023. ; https://doi.org/10.1101/2023.07.04.547725 doi: bioRxiv preprint

225
Rpa1 oocyte-specific conditional knockout (OcKO) female mice were generated as described in the 226 Methods section (Fig. 1A). The deletion of exon 8 causes a frame shift in the resulting mutant transcript.

227
To validate this, we performed immunofluorescence (IF) analysis using an RPA1-specific antibody that 228 has been previously validated in both mouse cell lines (Flach et al., 2014) and mouse oocytes (Yueh et  was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

262
We proceeded to investigate the underlying cause of sterility in Rpa1 OcKO females. At approximately 263 3.5 weeks of age, the size of ovaries from Rpa1-ZcKO females was comparable to that of the control 264 group. However, by 5 weeks of age, the OcKO females exhibited smaller ovarian size, and no further 265 growth was observed beyond 10 weeks of age ( Fig. 2A). Furthermore, even after hormone PMSG 266 treatment, very few large follicles could be observed in the OcKO ovaries at 10 and 16 weeks of age ( Fig.   267 2A). To further confirm these observations, histological ovarian sections were analyzed, and follicle 268 quantification was performed across various age groups, ranging from 2 months to 8 months (Fig. 2B).

269
At 2 months of age, the number of antral follicles was significantly decreased in the OcKO females (Fig.   270   2C). Starting at 4 months of age, both secondary and antral follicles exhibited a significant decrease in 271 numbers (Fig. 2D). These findings indicate that the severity of oocyte depletion increases with age in  was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023. ; https://doi.org/10.1101/2023.07.04.547725 doi: bioRxiv preprint To investigate the underlying cause of the substantial loss of secondary and antral follicles in Rpa1-ZcKO 294 ovaries, immunofluorescence analysis was performed on ovarian sections. Interestingly, although minor 295 DNA strand breaks detected by TUNEL were shown in granulosa cells of the control ovaries, massive 296 TUNEL signals were observed specifically in the granulosa cells of the antral follicles, but not other 297 follicles, in the Rpa1-ZcKO ovaries (Fig. 3A, B). Given that the Cre-recombinase activity under Zp3-Cre    in Rpa1-OcKO oocytes (Fig. 4D, E). Additionally, the TUNEL assay, capable of detecting both SSBs and 331 DSBs, did not detect apparent DNA strand breaks in Rpa1-OcKO oocytes (Fig. S1). In conclusion, Rpa1-

332
ZcKO oocytes displayed severe DNA damage, which is not attributed to DNA breaks.  In order to assess the impact of RPA loss in fully grown GV oocytes, we focused on p53 (known as 353 TRP53 in mice), since p63 expression is absent in antral follicles (Livera et al., 2008;Suh et al., 2006).

354
Notably, Rpa1-OcKO oocytes exhibited a significant increase in phosphorylation of p53 at Ser15, 355 indicating activation of the p53 pathway in the absence of the RPA complex (Fig. 5A, B). However, there 356 was no noticeable change in the acetylation status of p53 at Lys379 (Fig. 5C, D), suggesting that the loss 357 of RPA complex may not directly influence p53 acetylation or that alternative post-translational 358 modifications of p53 are involved. We also examined the retinoblastoma tumor suppressor protein Rb,

359
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023. ; https://doi.org/10.1101/2023.07.04.547725 doi: bioRxiv preprint 15 which plays a key role in DNA damage response (Wang et al., 2001). We observed a significant increase 360 in phospho-Rb (Ser807/811) levels in Rpa1-OcKO oocytes (Fig. 5E, F)

394
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023.

406
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023.

411
Immunofluorescence analysis of nuclear spreads revealed the presence of RPA1 foci on the 412 chromosomes of Rpa1-Ddx4-Cre OcKO oocytes at the pachytene stage from E18.5 embryos, indicating 413 that RPA1 was not depleted in the oocytes (Fig. 7A). Furthermore, immunofluorescence of synaptonemal 414 complex proteins showed normal chromosomal synapsis in control and Rpa1-Ddx4 OcKO oocytes (Fig.   415   7A, B). The persistence of RPA1 in E18.5 oocytes following Ddx4-Cre-mediated deletion is expected.  (Fig. 7C, D) and the 422 metaphase II stage (MII) (Fig. 7E, F), indicating that the DNA damage resulting from RPA loss in GV 423 oocytes persisted and led to chromosome misalignment during meiotic divisions.

424
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023. Although RPA has been studied in yeast and cell lines, its role in mammalian development remains 438 largely unexplored due to the early embryonic lethality resulting from Rpa1 mutation (Wang et al., 2005) 439 and null allele  in mice. In this study, we utilized two Cre drivers and conditional KO 440 strategy specifically in oocytes to investigate the consequences of RPA1 loss. We find that depletion of  (Table S1) show no changes in their transcript levels.  was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023. ; https://doi.org/10.1101/2023.07.04.547725 doi: bioRxiv preprint G2/prophase oocytes. It is generally accepted that fully grown GV oocytes have a weakened G2/M 470 checkpoint due to limited activation of the ATM kinase (Marangos and Carroll, 2012). However, in our 471 study, we observed robust activation of ATM in Rpa1-OcKO oocytes (confirmed by two different 472 commercial antibodies), suggesting that non-DSB-mediated DNA damage might be more efficient than 473 DSBs in activating ATM in G2/prophase oocytes. Nevertheless, despite the activation of ATM, these 474 Rpa1-OcKO oocytes still progressed to the MI stage (Fig. 7)

483
Although the precise mechanisms underlying this phenomenon require further investigation, our RNA-484 seq Gene Ontology (GO) analysis revealed that actin cytoskeleton organization was most significantly 485 altered in response to the loss of RPA in oocytes (Table S3)

491
The age-dependent progressive oocyte depletion phenotype in Rpa1-OcKO females is intriguing. As 492 these females age, the phenotypes worsen, including a decline in oocyte numbers and the loss of 493 secondary and antral follicles under the Zp3-Cre driver ( Fig. 1 and Fig. 2). Similarly, when the earlier 494 driver Ddx4-Cre was used, females exhibited an even earlier and more severe phenotype, with the loss 495 of almost all follicles, including primordial and primary follicles, as early as 4-6 weeks of age (Fig. S2).

496
These findings suggest that it takes some time for the pre-existing Rpa1 mRNA and protein to degrade 497 fully and for the phenotypes to manifest after Cre-mediated deletion. This explains the detection of RPA1 was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023. ; https://doi.org/10.1101/2023.07.04.547725 doi: bioRxiv preprint 22 and the loss of follicles following depletion of RPA in oocytes. We hypothesize that GV oocytes carrying 504 severe DNA damage due to RPA loss fail to drive the growth and/or differentiation of certain granulosa 505 cells through paracrine factors, but the specific mechanisms require further investigation.

507
Our study provides valuable insights into the role of the RPA complex during mammalian oogenesis. We 508 have demonstrated that RPA is crucial for DNA damage repair, and its loss leads to extensive and 509 persistent DNA damage responses, as evidenced by the activation of key signaling molecules including     . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023. . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023. ; https://doi.org/10.1101/2023.07.04.547725 doi: bioRxiv preprint