Abstract
Germ cell-derived genomic structure variants not only drive the evolution of species but also induce developmental defects in offspring. The genomic structure variants have different types, but most of them are originated from DNA double-strand breaks (DSBs). It is still not well known whether DNA DSBs exist in adult mammalian oocytes and how the growing and fully grown oocytes repair their DNA DSBs induced by endogenous or exogenous factors. In this study, we detected the endogenous DNA DSBs in the growing and fully grown mouse oocytes and found that the DNA DSBs mainly localized at the centromere-adjacent regions, which are also copy number variation hotspots. When the exogenous DNA DSBs were introduced by Etoposide, we found that Rad51-mediated homologous recombination (HR) was used to repair the broken DNA. However, the HR repair caused the chromatin intertwined and impaired the homologous chromosome segregation in oocytes. Although we had not detected the indication about HR repair of endogenous centromere-adjacent DNA DSBs, we found that Rad52 and RNA:DNA hybrids colocalized with these DNA DSBs, indicating that a Rad52-dependent DNA repair might exist in oocytes. In summary, our results not only demonstrated an association between endogenous DNA DSBs with genomic structure variants but also revealed one specific DNA DSB repair manner in oocytes.
Introduction
For mammals, the oocyte growing from the primordial follicle oocyte to fully grown oocyte may spend several weeks or even several months (1–3). These growing oocytes are arrested at the dictyate stage (like the G2 stage in somatic cells) and contain two set of homologous chromosomes, each of which contains two sister chromatids and at least one chromatid is recombined with the homologous chromatid (1). During the growing of oocytes, both endogenous metabolite-induced reactive oxygen species (ROS) or DNA torsion stress and exogenous environmental factors (4) can lead to the DNA double-strand breaks (DSBs). These DSBs in growing oocytes would reduce the female fertility if not repaired (5) or induce genomic variant if repaired mistakenly (6). As low copy repeats (LCRs) are extreme abundant in genomes of human and other mammals (7, 8), when the DNA DSBs occur at the LCR-rich regions, RAD51-mediated DNA broken end invasion may mistakenly invade the homologous fragment in non-allelic region. Moreover, subsequent DNA repair would induce the DNA fragment deletion or duplication if the flanking LCRs were syntropic or induce the DNA inversion if the LCRs were in the opposite direction (9). The DNA DSB-induced genomic rearrangements might occur in somatic cells, which is a main cause of cancers (10), and they might also occur in germ cells such as gametes, which would induce hereditable de novo copy number variants (CNVs) in offspring.
For somatic cells, the DNA DSBs induced by cellular intermediate metabolites or environmental irradiation (11) can be firstly sensed by double-strand DNA ends binding proteins such as Mre11-Rad50-Nbs1 (MRN) complex or by single-strand DNA-binding protein RPAs (12), and then the DNA damage signals are spread by ATM/ATR and CHK1/2 kinases, which not only recruit proteins to repair the DNA DSBs but also block the cell cycle (13). It is intriguing that the oocytes can bypass the DNA damage checkpoint and resume meiosis or even finish meiosis and be parthenogenetically activated with broken DNA (14–16). Up to now, it is still not well known whether the fully grown oocyte has the ability to repair DNA DSBs and why oocyte can be so tolerated with DSBs.
Generally, the DNA DSBs could be repaired by two main manners: the homologous recombination (HR) and the non-homologous end joining (NHEJ) (13). For the HR-mediated DNA DSB repair, the broken DNA ends (or termed overhangs) could be bound by Rad51, and then Rad51 filaments mediate the single-strand DNA end invading into the allele region of sister chromatids and repair the DSBs by using the sister chromatid DNA as template (17). Compared with HR, the NHEJ is an error-prone manner that repairs the DNA DSBs through direct linkage of DSB ends, which may induce DNA sequence loss or ligation error (18). In somatic cells, the selection of DSB repair manners is associated with the cell cycle stage (19). The HR is active after DNA replication (S/G2), and NHEJ can occur during the entire interphase but can be repressed by the HR pathway (20, 21). In addition, DNA DSBs could also be repaired by a Rad52-dependent RNA-templated repair, in which RNAs are used as direct template to repair the DNA DSBs (22). In oocytes, it has been reported that Rad51 expresses during oocyte maturation, indicating that HR may play essential roles in oocyte DNA DSBs repair (23). However, whether HR plays roles in fully grown oocytes is not well known.
In our previous study, we found that the centromere-adjacent DNA regions in human oocyte might have experienced high frequency of DNA DSB repair-induced gene conversion events (24). However, it is not well known whether DNA DSBs exist at the centromere-adjacent regions of mammalian oocytes and by which means they are repaired. In this study, we analyzed the repair of endogenous and exogenous DNA DSBs in oocytes and investigated the effects of DNA repair on meiotic homologous chromosome segregation. Our results not only revealed the specific DNA DSBs in oocytes but also associated the oocyte DNA DSBs with human de novo genomic rearrangements.
Results
Endogenous DNA DSBs and its association with CNV hotspots in oocytes
Our previous data had showed that mature human oocytes might have experienced high frequency of DNA DSB repair at the centromere-adjacent regions (24). As the centromere-adjacent regions of human genome contain extremely abundant LCRs (8), which are molecular prerequisite for genomic rearrangement, we plotted the SNP genotype exchange rates (SERs) (24) in human oocytes with the CNV frequency and LCR (which length > 1 kb) data (Fig. 1A and Supplementary Material, Fig. S1). The results showed that both LCRs and CNV enriched at the centromere-adjacent regions of specific chromosomes. The correlation coefficients of accumulated LCR length and CNV boundaries in 14 chromosomes were larger than 0.4 and all correlation coefficients were larger than 0.1 (Supplementary Material, Fig. S2), indicating the CNVs positive correlated LCRs in specific chromosome regions. To analyze whether meiotic crossovers are also associated with the CNV hotspots, we further assessed the correlation between CNV hotspots and oocyte crossover hotspots but found no obvious relevance between these two hotspots (Supplementary Material, Fig. S3).
Endogenous DNA DSBs in oocytes. (A) The visualization of LCRs, oocyte SERs, human CNV frequency and oocyte crossover hotspots of human chromosome 9 (chr9). (B) Endogenous DNA DSBs localized near the centromeres in mouse oocytes, DSBs (white arrows) were marked by γH2A.X (red) and centromeres were marked by Crest (Green) antibody. (C) The number of γH2A.X foci in the growing (NSN) and fully grown (SN) mouse oocytes (ns, no significant). (D) TUNEL assay result in mouse oocytes. The fluorescein-dUTPs (green) were transferred to the broken DNA ends by terminal deoxynucleotidyl transferase. (E) Endogenous DNA DSBs still exist at the peri-GVBD stage but disappear at the late GVBD stage in mouse oocytes. NSN/SN oocytes without/with positive Hoechst ring-like structure surrounded nucleolus (green squares). Bar = 10 μm.
Similar to human genome, centromere-adjacent regions in mouse genome also contain LCRs (25), so we used mouse oocyte as a model to analyze the centromere-adjacent DNA DSBs in this study. To analyze whether or not DNA DSBs really occur at the centromere-adjacent DNA regions, we labeled the DNA DSBs with phosphorylated H2AFX (γH2A.X) antibody and the terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) method. The results from both methods showed that the growing oocytes (oocyte without Hoechst positive ring-like structure surrounded nucleolus, NSN stage oocytes) and fully grown oocytes (oocyte with Hoechst positive ring-like structure surrounded nucleolus, SN stage oocytes (26, 27)) contain DNA DSBs (Fig. 1B and D). We also labeled the centromere with Crest antibody and found that DNA DSB signal locations coincide with the centromere signals, which indicates that the DNA DSBs exist around the centromere regions of oocyte (Fig. 1B). Evidence had shown that most centromeres in oocyte colocalized with the bright Hoechst positive areas and the number of the bright Hoechst positive areas is mostly less than 20 (28–30). We counted the γH2A.X foci number and results showed that the γH2A.X foci number in both NSN and SN oocytes are around 14 (14.0 ± 5.2 in NSN oocytes and 14.8 ± 4.0 in SN oocytes, Fig. 1C), and there is no obvious difference between them.
To analyze whether the DNA DSBs persist during oocyte maturation, we labeled the γH2A.X and Crest antibody in oocytes in vitro matured for 1.5–2.5 h. As a result we found that the γH2A.X foci still exist at the germinal vesicle breakdown (GVBD) stage but gradually disappeared in the late GVBD stage (Fig. 1E). From the immunofluorescence results, we could also confirm that the DNA DSBs located beside the centromeres but not overlap with centromeres in oocytes (Fig. 1E, squares).
Repair of exogenous DNA DSBs impairs the meiotic chromosome segregation
To analyze whether HR-mediated DNA DSB repair exists in oocytes, we labeled the Rad51 in the Etoposide-treated oocytes (treatment for 1 or 3 h) and the control germinal vesicle (GV) stage oocytes (treated by Milrinone for 3 h). Although the DNA DSBs exist, there is no Rad51 focus formed in the control oocytes (Fig. 2A, up row, SN oocytes with Milrinone treatment 3 h). For the Etoposide-treated oocytes, we detected the Rad51 foci in NSN oocytes 1 h after Etoposide treatment, and SN oocytes 3 h after Etoposide treatment (Fig. 2A). In most species of mammals, NSN stage would take weeks or even months for the growing oocytes (31). Our results showed that with Etoposide treatment large Rad51 foci formed in the NSN oocytes not only located near the bright Hoechst positive sites but also other genomic positions (Supplementary Material, Fig. S4).
Rad51 participates in the exogenous but not endogenous DNA DSBs repair in mouse oocytes. (A) No Rad51 focus was detected in the control oocytes (SN stage acquired by Milrinone 3 h treatment) but Rad51 foci (yellow arrows) could be detected at the NSN and late SN oocytes after Etoposide treating for 1 and 3 h, respectively. (B) When the control and Etoposide-treated oocytes were released for IVM, Rad51 still could not be detected in the control GVBD and pre-MI oocytes, but detected in the Etoposide-treated oocytes (yellow arrows). (C) Etoposide treating did not affect the GVBD rate. (D) The first PBE rates in the control and Etoposide-treated oocytes after IVM for 14 h. Bar = 10 μm. *0.01 < P value < 0.05, **P < 0.01.
As DNA DSBs still exist in normal oocyte at the post-GVBD stage (Fig. 1E), we matured the normal oocytes and Etoposide-treated oocytes in vitro for 1.5–2.5 h and then labeled the Rad51. Results showed that Rad51 focus was not detected in the post-GVBD normal oocytes. However, although we did not detect Rad51 focus in SN oocytes with 1 h Etoposide treatment, the Rad51 foci could be detected in Etoposide-treated oocytes after 1.5–2.5 h maturation, not only in the cell cycle arrested SN oocytes but also in the post-GVBD oocytes (Fig. 2B).
The Rad51-mediated DNA DSB repair would form D-loop intermediates, which link the broken DNA end and the template DNA together (32, 33). Rad51 foci formed in peri-GVBD oocytes might induce chromosome to gather around. The in vitro maturation (IVM) experiments showed that the GVBD rates (2 h 45 min after IVM) between the Etoposide-treated oocyte (93.0%) and control oocytes (93.2%) have no obvious difference (Fig. 2C). However, with live cell imaging and chromosome spreading methods, we indeed found that the chromosomes of Etoposide-induced DSB oocyte showed an intertwined state (Fig. S5).
To analyze whether the intertwined chromosomes affect the meiotic progression of oocytes, we observed the polar body extrusion (PBE) rates of the control and Etoposide-treated oocytes. The results showed that the PBE rates significantly reduced in the Etoposide-treated oocytes (Fig. 2D and Supplementary Material, Fig. S5). Then we used the Rad51 inhibitor RI-1 (34) to inhibit the HR in oocytes. We found that 60 μm RI-1 exposure has no obvious effects on the PBE rates of normal oocytes (Fig. 3A). However, if we treated the oocytes with 60 μM RI-1 for 2 h and then induced DSBs by Etoposide before releasing the oocytes for maturation, the decreased PBE rates in Etoposide-treated oocytes could be partially rescued (Fig. 3B). After maturation for 16 h, the mean PBE rate of Etoposide-treated oocytes was 17.2% whereas the mean PBE rate of Rad51-inhibited and Etoposide-treated oocytes was 46.8% (P < 0.01, Fig. 3B).
Rad51 inhibitor partially rescued the chromosome segregation defects induced by the homologous recombination repair in mouse oocytes. (A) The PBE rates in the control and RI-1-treated oocytes. ns, no significant. (B) The PBE rates in the control and RI-1-trated oocytes after IVM for 11 to 16 h. RI-1 partially rescued the Etoposide-induced PBE rate decreasing. (C) EdU click assay of the Etoposide-treated and control post GVBD oocytes. When treated with Etoposide, EdU would be incorporated into the nuclear DNA though homologous recombination-mediated DNA strand replication. Yellow arrow, weak EdU signals on chromosomes. White dots outside chromosomes represent replicated mitochondria DNA. (D) The rates of different levels (as shown in C) of EdU incorporating into the nuclear DNA in the control, Etoposide-treated, RI-1 plus Etoposide-treated oocytes. Bar = 10 μm. *0.01 < P value < 0.05, **P value < 0.01.
During the HR repair of DNA DSBs, the broken DNA ends invade into the template DNA to synthesize a new DNA strand (33). To analyze whether RI-1 could block the Rad51-mediated DNA DSB repair, we labeled the new synthesized DNA with 5-ethynyl-2-deoxyuridine (EdU) click assay (35). We cultured the control, Etoposide-treated and RI-1 plus Etoposide-treated oocytes in M2 medium with 10 μM EdU and labeled the EdU with fluorescent dyes by using click reaction. As a result we found that EdU signals could not be found in chromosomes of normal oocytes but found in chromosomes of Etoposide-treated oocytes (Fig. 3C). Compared with the Etoposide-treated oocytes, the EdU signals became weak in RI-1 plus Etoposide-treated oocytes (Fig. 3D).
Rad52 and RNA:DNA hybrids were involved in the repairing of endogenous DNA DSBs in mouse oocytes
As we had not detected Rad51 focus or EdU signal in normal oocytes, so we consider whether the endogenous DNA DSBs in oocytes were repaired by non-HR pathways, which could also induce the gene conversion events. Evidence had shown that RNA-templated DNA DSB repair was a microhomology-mediated and Rad52-dependent pathway to repair the DNA DSBs. To make clear the repair mechanism of endogenous DNA DSBs, we labeled the Rad52 by anti-Rad52 antibody and the RNA:DNA hybrid by S9.6 (36) in the control and Etoposide-treated oocytes. As a result we found that both Rad52 and RNA:DNA hybrids were colocalized with γH2A.X in oocytes (Fig. 4).
Rad52 and RNA:DNA hybrids are involved in the oocyte endogenous DNA DSBs repair. (A) Colocalization of Rad52 (red) and γH2A.X (green) in the NSN and SN oocytes with or without Etoposide treating. Yellow arrows indicate Rad52 in the control oocytes. (B) Colocalization of RNA:DNA hybrids (recognized by S9.6 antibody, green) and γH2A.X (red) in the NSN and SN oocytes with or without Etoposide treating. White arrows indicate RNA:DNA hybrids in the control oocytes. Bar = 10 μm.
When oocytes were fully grown, the transcription activity would be globally silenced as indicated by 5-bromouracil (BrU) or 5-ethynyl uridine (EU) labeling (37–39). However, in our results we detected the RNA:DNA hybrids in mouse NSN, SN and even GVBD oocytes (Supplementary Material, Figs S6 and S7), which indicates that the RNA:DNA hybrid formation might be transcription independent. To verify whether the RNA:DNA hybrid formation is transcription independent, we labeled the nascent RNAs with EU click methods. As a result we found EU signal globally formed only in the control and Etoposide-treated NSN oocyte nucleus but not in the SN oocytes. Our results showed that exogenous DNA DSBs did not resume the transcription activity and the RNA:DNA hybrid formation might be transcription independent (Fig. 5). We also found the nascent RNA signals in the NSN and SN oocytes. However, these nascent RNA foci did not colocalized with γH2A.X foci (Fig. 5, white arrows). In addition, we also found the centromere-adjacent localization of RNA:DNA hybrids in diplotene stage spermatocyte, similar to oocytes (Supplementary Material, Fig. S8).
The formation of RNA:DNA hybrids are transcription-independent in mouse oocytes. (A) Representative images of EU (green) and γH2A.X (red) in the NSN and SN oocytes treated with/without Etoposide. Etoposide-induced DNA DSBs had not activated the transcription in SN oocytes. Yellow arrows, DNA DSB sites in control oocytes. (B) Representative images of EU (green) and RNA:DNA hybrids (recognized by S9.6 antibody, red) in the NSN and SN oocytes treated with/without Etoposide. RNA:DNA hybrids were formed in the Etoposide-treated SN oocytes although the global transcription were inactivated. Yellow arrows, RNA:DNA hybrid foci formed in the control oocytes. White arrows indicated the bright EU signals in the NSN and SN oocytes, which indicating transcription is still activated in the specific genome regions of SN oocytes. Bar = 10 μm.
Discussion
As the development of the next-generation sequencing technologies, more and more genomic variants in human were found associated with diseases (40). The genomic rearrangements such as DNA deletion, duplication, inversion and translocation could be generated in not only somatic cells but also germ cells. Evidence showed that de novo CNVs found in families increased the risk of autism (41). Human with pericentric inversion in chromosome 9 (chr9) had been shown with reduced fertility (42, 43). In addition, the de novo complex chromosome rearrangement had also been reported in human germ cells (44, 45). All these genomic variants could be formed in early embryos, sperm and oocytes. As DNA replication exist during adult spermatogenesis and early embryo development, so both ROS and DNA crosslink reagents might induce replication stress and DNA DSBs in male germ cells and early embryos, which might be the main source of pathogenic CNVs. However, as very few studies about endogenous oocyte DNA DSBs had been reported, the molecular mechanisms of oocyte-derived genomic rearrangement have not been considered before. In this study we showed that the endogenous DNA DSBs exist at the centromere-adjacent regions in mammalian oocytes. These peri-centric DNA DSBs might provide the molecular basis for the production of CNVs, inversions and even complex genomic rearrangement.
In this study we found the endogenous DNA DSBs in growing and fully grown oocytes. However, these DNA DSBs could not be repaired immediately by the Rad51-mediated HR system, indicating that oocytes have unknown mechanisms that block the HR pathway. Our results showed that oocytes could enter the pre-metaphase with endogenous DNA DSBs and these DNA DSBs could be repaired at peri-GVBD stage. These might explain why oocytes are less sensitive to the exogenous DNA DSBs (16). Evidences had shown that normal oocytes could not activate the ATM kinases (16), indicating endogenous centromere adjacent DNA DSBs did not activate ATM and their repair might be ATM independent. The DNA damage checkpoint needs to be repressed in oocytes so that oocyte could resume meiosis with endogenous DNA DSBs. Also, the endogenous DNA DSBs might be used to relieve the torsional stress of homologous chromosomes produced during the oocyte growth and meiosis. However, it is still not well known that the biological functions of these endogenous DNA DSBs in oocytes. Our results showed that oocyte DNA DSBs mostly located at the centromere-adjacent regions where enriched with LCRs that might be the source of de novo CNV formation; however, there is less knowledge about the source of these DNA DSBs. Evidences showed that microsatellite (simple repeats) instability to could induce genome rearrangement and cancer (46, 47). Microsatellite instability is generally induced by the deficiency of mismatch repair system (48); however, whether DNA DSBs in oocytes are associated with microsatellite instability and whether mismatch repair system plays roles in fully grown oocytes are not well known.
At the SN stage, oocytes have been fully grown and chromatins become condensed with reduced transcription. The specific chromatin structure might induce the non-allele homologous invasion of broken DNA end during HR repair. If HR occurred between different chromosomes, these HR repairs would make chromosomes intertwined together, which might delay the meiotic progression in oocytes. On the other hand, if the non-allele homologous invasion could not be released by mismatch repair system, DNA break-induced replication (49) might be activated and complex genomic rearrangement would be formed. In our results we found EU signals in normal SN oocytes indicating there were transcription events at specific genomic regions; whether these RNA transcription were associated with oocyte endogenous DNA DSBs and RNA:DNA hybrid formation should be further investigated.
Our results showed that Rad52 and RNA:DNA hybrids are involved in the endogenous DNA DSB repair. As the human oocyte sequencing data showed that the centromere-adjacent regions had high frequency gene conversion events (24) and HR had not been used to repair these DNA DSBs, our results indicate that oocyte may use an RNA-templated DNA repair method to repair the endogenous DNA DSBs. Evidence showed that metabolic disease such as diabetes and obesity could increase the ROS in oocytes (4, 50, 51), which might introduce exogenous DNA DSBs in oocytes. How oocyte could recognize the centromere-adjacent DNA DSBs and repair these DSBs is important for oocyte to maintain its genome stability. More efforts should be paid in this field since genome stability is essential for embryo development.
Materials and Methods
Ethic statement
All manipulations in this study were permitted by the ethics committee of Guangdong Second Provincial General Hospital and performed according the suggestions of animal application and welfare guidelines. All animals used in this study were purchased from SPF-Tsinghua company.
Data visualization
The human genome structure variant data were downloaded from dbVar (52). LCR data and other chromosome information of human and mouse were downloaded from UCSC genome browser (53). The data of human oocyte SERs (24) and crossover hotspots were extracted from literatures (24, 54). The human genome version hg19 and mouse mm10 were used in this study.
Oocyte isolation and treatment
The denuded fully grown GV stage oocytes were isolated from 6- to 8-week-old ICR strain mice. All the manipulations of oocytes before IVM were handled in the M2 medium (Sigma, M7167) with 2.5 μM milrinone (MCE, HY-14252). Fresh M2 medium was used for IVM of oocytes. To introduce DNA DSBs in oocytes, we treated the oocytes with 50 μM Etoposide (topoisomerase 2 inhibitor) (Beyotime, SC0173) for 1 or 3 h. To inhibit the activity of Rad51, we treated the oocytes with or without 40 μM RI-1 (MCE, HY-15317) for 2–3 h and all further manipulations were under RI-1 treatment (34). Then we induced DSBs in oocytes with Etoposide for 1 h and released the oocytes for IVM.
Immunofluorescence and TUNEL labeling
The primary antibodies used in this study were anti-Rad51 (Abcam, ab133534), anti-tubulin (Sigma, F2168), anti-Crest (Antibodies Incorporated, 15-234-0001), anti-γH2A.X (Santa, sc-517 348; and Bioworld, BS4760) and anti-RNA:DNA hybrid (Millipore, MABE1095). To label the specific proteins, the oocytes were firstly fixed with 4% paraformaldehyde for 40 min at room temperature. Secondly, oocytes were permeated by phosphate-buffered saline (PBS) with 0.3% triton X-100 and blocked by PBS with 3% bovine serum albumin. Then the oocytes were incubated with primary antibody at 4°C overnight and washed 3–8 times by PBS with 0.1% tween. If needed, oocytes were incubated with secondary antibodies and washed 3–8 times. Then the oocyte DNA was stained using 500 nm Hoechst for 1 h and mounted in DABCO reagent (0.25 g DABCO, 9 ml glycerol, 0.5 ml 1 M Tris pH8.0 and 0.5 ml water). Finally, the oocytes were observed under the Dragonfly 200 system (Andor). For TUNEL labeling we used the One Step TUNEL Apoptosis Assay Kit (Beyotime, C1086) and performed according to the instructions issued by the manufacturer.
Live cell imaging
To observe the GVBD and the chromosome dynamics of oocyte during its homologous chromosome segregation, the in vitro matured oocytes were real-time observed by using Dragonfly 200 system (Andor). For the control oocytes, we firstly in vitro cultured the oocytes for 0.5 h (to observe GVBD) or 8 h (to observe PBE of oocytes). Then the oocytes were cultured in M2 with 100 nm Hoechst33342 and observed under live cell station every 10 min for 4 h. The Etoposide-induced DSB oocytes were observed by using the same method. As the PBE time is delayed in Etoposide-treated oocytes, we observed the chromosome dynamics from IVM 12 h to IVM 16 h.
Chromosome spreading
For the chromosome spreading, the control oocytes and Etoposide-treated oocytes were in vitro matured for 6–7 h and then treated with 0.5% citric acid/50 mm sucrose solution for 30 min. Then oocytes were placed on the glass slide that was coated with 0.15% triton/1% paraformaldehyde. Before the drop dried at room temperature, the oocyte chromosomes were mounted with Hoechst33342/DABCO and observed under the Dragonfly system.
For the chromosome spreading of spermatocyte, we collected the spermatocytes from adult male ICR strain mice and treated with hypotonic solution (0.3% trisodium citrate dihydrate) for 30 min at room temperature. Then cells were added into 100 mm sucrose and dropped to slides. Then spermatocytes were fixed, permeated, blocked and labeled with appropriate antibodies. After mounting, slides were observed under Dragonfly system.
EdU and EU click reaction
To label the nascent DNAs or RNAs, we first cultured the oocytes in 10 μm EdU (Beyotime, C0071S) or 0.5 mm EU (Click Chemistry Tools, 1261-10) for 0.5 h or longer time as needed, then the click reactions were performed as the manufacturer’s descriptions (Beyotime, C0071S).
Statistics
At least three replicates were conducted for each experiment. For the comparison of oocyte PBE rates between the DSB oocytes and control oocytes, Student’s t test was used for hypothesis test by R software. The P values less than 0.05 (*) or 0.01 (**) were considered as significant or extremely significant.
Acknowledgements
We would like to thank X.H. Sun, X.H. Wang, C.F. Zhang and K. Peng in the Fertility Preservation Lab for their helps in animal care and instrument operations. We thank Dr Q.Q. Sha for sharing us the Crest antibody.
Conflict of Interest statement. None declared.
Funding
National Natural Science Foundation of China (98167145, 31401276 and 31571550)
