Maternal BCAS2 protects genomic integrity in mouse early embryonic development

Organisms and cells have evolved complex systems for responding to DNA damage to ensure the survival and faithful transmission of genetic materials to subsequent generations. Depending on the type of DNA lesion, cells generally arrest cell cycle progression by activating the DNA damage response (DDR), which includes sensing DNA lesions and triggering cascades mediated by the ATM (ataxia telangiectasia mutated) and ATR (ataxia telangiectasia and Rad3 related) kinases to repair damaged DNA (Ciccia and Elledge, 2010; Jazayeri et al., 2006). RPA (replication protein A), an eukaryotic single stranded DNA (ssDNA)-binding heterotrimeric protein composed of subunits RPA1, RPA2 and RPA3, plays crucial roles in DNA repair and replication stress to maintain genome integrity (Fanning et al., 2006). ssDNA can be recognised by RPA, which recruits the ATR/ATRIP complex and generates a platform to activate the ATR signalling cascade (Zou and Elledge, 2003). Defects in DDR affect genome instability and cell survival (Jazayeri et al., 2006).

Genome instability arising from DDR defects in early embryos or transmitted from germ cells might result in human diseases, including cancers and abnormal reproductive outcomes (O'Driscoll and Jeggo, 2006). Because of a lack of efficient DNA damage checkpoints, the mouse oocyte is comparatively tolerant of the accumulation of DNA lesions (Marangos and Carroll, 2012). Genetic lesions induced by DNA-damaging agents in mouse sperm are repaired in fertilised eggs (Generoso et al., 1979). DNA demethylation is accompanied by the appearance of the DNA damage signal in mouse zygotes (Hajkova et al., 2010; Wossidlo et al., 2010). Thus, DDR might contribute to mouse early embryonic development by cooperating with important events within the programmed implantation window (Kojima et al., 2014; Zeng and Schultz, 2005).

DDR depends on maternal mRNAs and proteins in newly fertilised zygotes, in which transcription activity is essentially lacking (Brandriff and Pedersen, 1981; Li et al., 2010). DDR can be detected almost immediately after fertilisation and as early as the G1 phase (Toyoshima et al., 2005). Preventing the RPA complex from binding to DNA by mutating RPA1 results in embryonic lethality at the blastocyst stage (Wang et al., 2005). Genetic mutations in Atr and other important DNA damage repair genes such as Chk1 (checkpoint kinase 1, also known as Chek1 – Mouse Genome Informatics) and Rad50 (RAD50 homologue) lead to early embryonic lethality during the perimplantation stage (Brown and Baltimore, 2000; Liu et al., 2000; Luo et al., 1999; Takai et al., 2000). The embryonic lethality of these mutants impedes investigation into the maternal function of DDR genes. Until now, only maternal PMS2 (postmeiotic segregation increased 2), a mammalian homologue of a bacterial DNA mismatch repair protein, is reported to play a role in DDR in mouse development (Larson et al., 2004). Embryos lacking maternal UBE2A (ubiquitin-conjugating enzyme E2A), a homologue of the ubiquitin-conjugating enzyme RAD6 involved in DNA repair in yeast, arrest development at the two-cell stage. However, whether UBE2A is involved in DNA repair in mouse early embryos has not been analysed (Roest et al., 2004). Currently, elucidation of how maternal DDR functions in early mouse development remains elusive.

Bcas2 (breast carcinoma amplified sequence 2) was originally identified by gene amplification in human breast cancer cells (Nagasaki et al., 1999). Subsequently, BCAS2 has been identified as a core component of PRP19, a pre-mRNA splicing complex (Ajuh et al., 2000). BCAS2 downregulates p53 (transformation related protein 53, also known as TRP53 – Mouse Genome Informatics or TP53 – HUGO Gene Nomenclature Committee) in the presence of DNA damage in MCF-7 cells (Kuo et al., 2009). BCAS2 interacts directly with RPA in DNA damage repair in HEK293T and HeLa cells (Maréchal et al., 2014; Wan and Huang, 2014). However, the physiological roles of BCAS2 in mammals are unknown. In the current study, we find that BCAS2 responds to endogenous and exogenous DNA damage in mouse zygotes and demonstrate its vital role in the maintenance of genomic integrity during mouse preimplantation development, potentially through interaction with RPA.

Bcas2 mRNA and protein profiles were examined by quantitative RT-PCR (qRT-PCR) and western blot in mouse oocytes and preimplantation embryos. Bcas2 transcript and protein levels were comparatively high in germinal vesicle (GV) oocytes, decreased in metaphase II (MII) oocytes, further reduced after fertilisation until the two-cell stage, elevated in morulas, and increased in blastocyst stage embryos (Fig. 1A,B). BCAS2 protein localisation in mouse oocytes and preimplantation embryos was further investigated by immunofluorescent staining. BCAS2 localised to nuclei in oocytes and at all stages of mouse preimplantation embryos except for MII oocytes (Fig. 1C).

BCAS2 is involved in pre-mRNA splicing and DDR as described above. However, splicing is not inefficient in the mouse zygote (Abe et al., 2015). To determine whether BCAS2 was involved in DNA repair in mouse zygotes, the correlation between BCAS2 and γH2AX, a marker of DNA damage, was assessed in zygotes of relevant pronuclear stages (PN) by staining. Mouse zygotes at various PN stages were obtained 23 h (PN2), 25 h (PN3), 28 h (PN4), and 30 h (PN5) post-hCG at embryonic day (E)0.5 and stained with traditional staining conditions (PT conditions, PFA–Triton). BCAS2 appeared in PN2-5 stage embryos, and BCAS2 staining intensity was stronger in paternal pronuclei than in maternal pronuclei in PN3 to PN5 zygotes, which correlated with γH2AX staining (Fig. 2A). Furthermore, under TP (Triton–PFA) conditions, BCAS2 foci colocalised with γH2AX in the pronuclei of PN3 to PN5 zygotes (Fig. 2B), suggesting its potential role in DDR in mouse zygotes.

Because γH2AX and BCAS2 foci colocalised more frequently in the paternal pronuclei of mouse zygotes, we induced DNA double strand breaks (DSBs) in MΙΙ oocytes to examine whether BCAS2 responds to exogenous DNA damage in zygotes (Fig. 2C). MΙΙ oocytes were treated with ultraviolet light (UV), a common treatment to induce DSBs, and fertilised with normal sperm (Bradley and Taylor, 1981; Masui and Pedersen, 1975). Few mouse zygotes developed into two-cell embryos when MΙΙ oocytes were irradiated with a high dosage of UV irradiation (0.42 J/cm²) (data not shown). After lowering the UV dosage (0.21 J/cm²), the majority of zygotes underwent the pronuclear stage and developed into two-cell embryos with a severe delay (data not shown). UV-induced DSBs were evidenced by increased staining of γH2AX, Ser15-phosphorylated p53 and CHK1 in maternal pronuclei (Fig. S1). Having determined the proper UV dosage (0.21 J/cm²), irradiated MΙΙ oocytes were fertilised with normal sperm. Subsequently, zygotes were cultured for 10 h and stained with BCAS2 and γH2AX antibodies. Compared with the controls, BCAS2 and γH2AX foci accumulated in the maternal pronuclei of zygotes (Fig. 2D). These data show that BCAS2 responds to UV-induced DSBs by localizing to DNA damage sites, supporting the potential role of BCAS2 in DNA repair in mouse zygotes.

To investigate whether BCAS2 functions in DNA repair in early embryos, a mouse line with a conditionally targeted Bcas2 allele was established by flanking its exons 3 and 4 with LoxP sites (Fig. S2). Bcas2^+/fl-Neo (fl, flox) mice were viable and fertile. However, no Bcas2^{fl-Neo/fl-Neo} mice resulted from mating Bcas2^+/fl-Neo males and females (data not shown), suggesting that the inserted Neo cassette might disrupt BCAS2 function and result in embryonic lethality. After deleting the Neo cassette from Bcas2^+/fl-Neo mice by mating with Rosa26-Flp mice (Rodriguez et al., 2000), Bcas2^fl/fl mice were obtained.

To investigate BCAS2 function in mouse oocytes, Bcas2^fl/fl females were mated with Zp3-Cre males (Lewandoski et al., 1997). Bcas2^fl/fl;Zp3-Cre mice were obtained by mating Bcas2^fl/+;Zp3-Cre mice and genotyped with PCR (Fig. S2, Fig. S3A). Bcas2^fl/fl;Zp3-Cre females and males grew to adults and appeared grossly normal. However, Bcas2^fl/fl;Zp3-Cre females did not deliver offspring after mating with normal males over three months (Fig. 3A). Unless otherwise stated, Bcas2^fl/+;Zp3-Cre and Bcas2^fl/fl;Zp3-Cre oocytes and embryos are hereafter designated control and Bcas2^mNull, respectively. Compared with controls, the Bcas2 mRNA level was reduced by 95% (Fig. 3B), and BCAS2 protein was almost absent in Bcas2^mNull oocytes (Fig. 3C,D). These data show that BCAS2 plays important roles in female mouse fertility.

To examine the potential role of BCAS2 in oogenesis, ovaries from control (Bcas2^fl/+;Zp3-Cre) and Bcas2^fl/fl;Zp3-Cre females were fixed and sectioned for histological analysis. Ovaries from Bcas2^fl/fl;Zp3-Cre females included all stage follicles and corpora lutea and were indistinguishable from controls (Fig. S3B). In addition, superovulation resulted in similar MΙΙ oocyte morphology and number between the control and Bcas2^mNull groups (Fig. S3C,D). These data suggest that BCAS2 might be not essential for oocyte maturation and ovulation in Bcas2^fl/fl;Zp3-Cre females.

Next, mouse embryos were recovered at E0.5, E1.5, E2.5 and E3.5 after mating Bcas2^fl/+;Zp3-Cre and Bcas2^fl/fl;Zp3-Cre females with wild-type males. Compared with controls, Bcas2^mNull zygotes and two-cell embryos were almost normal in morphology and number (E0.5 and E1.5, Fig. 3E,F). However, the majority of Bcas2^mNull embryos arrested at the two- to four-cell stages at E2.5 and E3.5, when most control embryos were at the morula and blastocyst stages, respectively (E2.5 and E3.5, Fig. 3E,F).

To further examine the defects, embryos were recovered at E0.5 and E1.5 and cultured in vitro to develop into the next cleavage stage. The rate of Bcas2^mNull zygotes progressing into the two-cell stage was similar to controls (Fig. 3G), but the rate of two-cell embryos progressing into the four-cell stage was much lower than in controls (Fig. 3H), suggesting a delayed transition from the two- to the four-cell stage in Bcas2^mNull embryos. These results indicate that maternal BCAS2 is required for mouse preimplantation development.

The observed accumulation of BCAS2 and γH2AX foci in the maternal pronuclei of zygotes exposed to exogenous DNA damage (Fig. 2D) prompted us to investigate the role of BCAS2 in DNA repair. To address this, two-cell embryos were recovered at 33 h (G0), 37 h (mid/late S-phase) and 48 h post-hCG (late G2 phase) at E1.5 (Artus and Cohen-Tannoudji, 2008) and stained with anti-γH2AX antibody. The γH2AX staining was mostly diffuse in the nuclei of control (n=10) and Bcas2^mNull two-cell (n=12) embryos recovered at 33 h post-hCG (Fig. 4A). At 37 h post-hCG, γH2AX foci increased in both control and Bcas2^mNull two-cell embryos (Fig. 4A,B). The number of γH2AX foci was similar in control (n=13) and Bcas2^mNull embryos (n=34). However, large γH2AX foci were frequently observed in Bcas2^mNull embryos (Fig. 4A,B). At 48 h post-hCG, γH2AX foci were dramatically reduced in control embryos (n=65) but remained at significantly higher levels in Bcas2^mNull embryos (n=46) (Fig. 4A,B).

To investigate whether increased γH2AX expression in Bcas2^mNull two-cell embryos was caused by developmental delay, we stained Bcas2^mNull embryos (48 post-hCG) with phosphorylated H3S10, a marker of late G2/M-phase (Goto et al., 1999). When normalised to DNA staining by Hoechst, both controls (n=30), and Bcas2^mNull embryos (n=30) shared a similar intensity of phosphorylated H3S10 signal (Fig. 4C,D), suggesting that the majority of Bcas2^mNull two-cell embryos entered G2-phase at 48 post-hCG with damaged DNA in the same way as control two-cell embryos. This would indicate that increased γH2AX is not a result of developmental delays in Bcas2^mNull two-cell embryos, and together, these data suggest the BCAS2 plays an important role in DNA repair in mouse early embryos.

The damaged DNA in Bcas2^mNull two-cell embryos might activate p53 to delay cell cycle progression (Shiloh, 2001). To test this, Bcas2^mNull embryos recovered at 48 h post-hCG were stained with anti-phosphorylated p53 antibody. Compared with control, phosphorylated p53 increased in Bcas2^mNull two-cell embryos (Fig. 5A). Western blotting further confirmed the increased level of γH2AX and phosphorylated p53 (Fig. 5B). In response to DNA damage, the activation of p53 usually triggers the expression of p21 (cyclin-dependent kinase inhibitor 1A, also known as Cdkn1a – Mouse Genome Informatics) and Gadd45a (growth arrest and DNA-damage-inducible 45 alpha) to arrest the cell cycle (Brugarolas et al., 1995; Siafakas and Richardson, 2009). Thus, these p53 targets in Bcas2^mNull embryos were examined by qRT-PCR. Compared with controls, the expression of Gadd45a was not significantly changed. However, p21 mRNA significantly increased in Bcas2^mNull two-cell embryos (Fig. 5C).

Subsequently, we tested whether p53 inhibition could alleviate the two-cell arrest of Bcas2^mNull embryos. To this end, in vitro fertilised Bcas2^mNull zygotes were treated with Pifithrin-α, a specific inhibitor of p53-dependent transcriptional activation, and cultured for 43 h. Compared with 0.1% DMSO treatment (control), Pifithrin-α significantly decreased the percentage of arrested Bcas2^mNull two-cell embryos (Fig. 5D,E). Furthermore, Pifithrin-α significantly decreased the expression of p21 and Gadd45a in Bcas2^mNull two-cell embryos (Fig. 5F). These results suggest that the delayed or arrested development of Bcas2^mNull two-cell embryos is partially as a result of damaged DNA and the subsequent activation of p53.

Considering the relatively limited ability of Pifithrin-α to alleviate the developmental arrest of Bcas2^mNull two-cell embryos, other pathways might also be involved in the arrest of Bcas2^mNull embryos. DNA damage activates the WEE1/CDC25 pathway that is regulated by CHK1, which arrests the cell cycle at interphase by stimulating the CDC2-cyclin B complex (O'Connell et al., 1997; Sanchez et al., 1997). In normal mouse two-cell embryos, CHK1 was phosphorylated at S-phase and then decreased at the late G2 stage during cell cycle progression (Fig. S4). To examine whether CHK1/CDC25 was activated in Bcas2^mNull two-cell embryos, control and Bcas2^mNull two-cell embryos recovered at 48 h post-hCG were immunostained for pCHK1. pCHK1 staining was almost absent in controls (n=30) but was observed in 48% of Bcas2^mNull two-cell embryos (n=25) (Fig. 5G).

In addition to DNA damage checkpoints, malfunction of zygotic genome activation (ZGA) also causes two-cell stage arrest (Bultman et al., 2006; Posfai et al., 2012; Wu et al., 2003). When the expression of ZGA genes (Zeng et al., 2004; Zeng and Schultz, 2005) was examined by quantitative RT-PCR, the mRNA levels of ZGA genes such as Tdpoz1, Tdpoz3, Tdpoz4 increased. However, the expression of other ZGA genes (Gm13043 and Ccnjl) decreased (Fig. S7) or remained unaffected (Zscan4a, H2-Q6 and Prss8, data not shown), suggesting that ZGA is partially affected in Bcas2^mNull two-cell embryos and might also contribute to delayed or arrested development.

Unrepaired DNA damage endangers cells with risks of genome instability, including abnormal micronuclei (Takai et al., 2000). To examine if micronuclei were present in Bcas2^mNull embryos, DNA was stained with Hoechst. Micronuclei were not observed in control embryos. However, 40% (n=40) of Bcas2^mNull embryos surviving beyond the two-cell stage had micronuclei (Fig. 5H). These data suggest that BCAS2 disruption leads to the accumulation of damaged DNA and activates checkpoints in Bcas2^mNull two-cell embryos, which might arrest mouse development at early embryonic stages.

To examine DNA repair defects in mouse zygotes, embryos were recovered at 23, 25, 28 and 30 h post-hCG (PN2, PN3, PN4, PN5, respectively) and stained with anti-γH2AX antibody. In control zygotes, γH2AX foci initially appeared at PN2, dramatically increased at PN3 and PN4, and almost disappeared at late PN5 (Fig. 6A, left panel). The number of γH2AX foci did not significantly increase during PN2 to PN4 (data not shown). However, enlarged γH2AX foci were observed in Bcas2^mNull PN3 zygotes (Fig. 6A, right panel). Strikingly, compared with the control at late PN5 in which γH2AX foci almost disappeared, γH2AX foci persisted in Bcas2^mNull zygotes (Fig. 6A,B). DNA fragmentation was further confirmed in Bcas2^mNull PN5 zygotes by TUNEL assay (Fig. S6).

Next, we tested whether re-expression of BCAS2 could repair the damaged DNA in Bcas2^mNull zygotes. Bcas2^mNull zygotes were recovered at 23 h post-hCG and microinjected with EGFP-tagged Bcas2 mRNA or control (EGFP). The injected zygotes were cultured for 7 h until the PN5 stage and stained with anti-γH2AX antibody. γH2AX foci were partially reduced in the Bcas2^mNull zygotes injected with EGFP-Bcas2 mRNA but not in the Bcas2^mNull zygotes injected with EGFP (Fig. 6C,D). These results further support a BCAS2 function in DNA repair in mouse zygotes.

DNA replication mainly initiates at PN3 in mouse zygotes (Ferreira and Carmo-Fonseca, 1997). The observation of enlarged γH2AX foci at PN3 in Bcas2^mNull zygotes (Fig. 6A, arrows) led us to suspect that the defects in DNA repair might disrupt DNA replication in Bcas2^mNull zygotes. To test this, control and Bcas2^mNull zygotes were recovered at 25, 28 and 30 h post-hCG, and pronuclear formation was examined. Most control and Bcas2^mNull zygotes developed into PN3 (25 h), PN4 (28 h) and PN5 (30 h). Then, these embryos were cultured with EdU for 30 min prior to fixation and stained with EdU. In controls, the EdU signal was almost equally diffused in both parental pronuclei at PN3 and in the peripheral regions of both pronuclei at PN4 (Fig. 7A, left panel), consistent with a previous report (Wossidlo et al., 2010). However, the EdU signal was either weak or absent in PN3 and PN4 Bcas2^mNull zygotes (Fig. 7A, right panel; Fig. 7B). PN3 and PN4 zygotes labelled with EdU were further quantified. In controls, 87% (n=32) of PN3 and 83% (n=45) of PN4 zygotes were labelled with EdU. However, only 51% (n=26) of PN3 and 38% (n=31) of PN4 Bcas2^mNull zygotes were labelled with EdU (Fig. 7B). At the late PN5 stage, control zygotes (n=15) were not labelled with EdU. However, 64% (9/14) of Bcas2^mNull zygotes were still labelled with EdU (Fig. 7B). These data demonstrate that Bcas2^mNull embryos have impaired DNA replication, suggesting that unrepaired DNA breaks cause replication stress during pronuclear stages in mouse zygotes.

To further investigate the role of BCAS2-mediated DNA repair in recovering from replication stress, zygotes were treated with hydroxyurea (HU), a general DNA damaging agent that stalls or breaks replication forks. Control and Bcas2^mNull zygotes were recovered at 25 h post-hCG and exposed to HU for 3 h. After HU release, these embryos were cultured for 23 h to develop into two-cell embryos. Approximately 55% of control zygotes (n=121) developed into two-cell embryos, whereas only 22% of Bcas2^mNull zygotes (n=132) progressed to the two-cell stage (Fig. 7C). These data suggest that the capability to recover from replication stress is impaired in Bcas2^mNull zygotes.

The RPA complex, a well-known sensor in response to replication stress, is composed of RPA1, RPA2 and RPA3 (Fanning et al., 2006; Zou and Elledge, 2003). BCAS2 lacking the 139 N-terminal bases fails to interact with RPA1 (Wan and Huang, 2014), prompting us to examine whether the interaction between BCAS2 and RPA1 was required for DNA repair in mouse zygotes. Normal zygotes at PN2 were injected with Bcas2 mRNA lacking the N-terminal 139 bases (referred as ΔN-EGFP), cultured until PN5 and stained with γH2AX. Compared with controls injected with EGFP mRNA or full-length Bcas2-EGFP, zygotes injected with ΔN-EGFP exhibited increased γH2AX foci, suggesting that BCAS2 lacking the interaction with RPA1 might function in DNA repair in mouse zygote as a dominant-negative mutant (Fig. 6C,E). Furthermore, γH2AX foci were not reduced by the injection of ΔN-EGFP in Bcas2^mNull zygotes compared with Bcas2^mNull zygotes injected with normal Bcas2-EGFP mRNA (Fig. 6C,D). These data suggest that the BCAS2 interaction with RPA1 is important for DNA repair in mouse zygotes.

In response to DNA damage, RPA2 is phosphorylated at Ser33, Ser4/Ser8, Ser12, and Thr21, promoting the binding of the RPA complex to DNA to activate DNA damage checkpoints (Anantha et al., 2007; Block et al., 2004; Oakley et al., 2003; Shi et al., 2010; Wang et al., 2001). Mutated Ser33A, Ser4/Ser8 and Thr21A phosphorylation sites in RPA2 lead to defective recovery from replication stress (Ashley et al., 2014; Olson et al., 2006). Decreased BCAS2 affects the phosphorylation of RPA2 Ser4/Ser8 (Wan and Huang, 2014), which is phosphorylated after Thr21 and Ser33 phosphorylation (Anantha et al., 2007; Liu et al., 2012). To further investigate whether BCAS2 is involved in DNA repair through the RPA complex, the activity of RPA2 in response to DNA damage was investigated in control and Bcas2^mNull zygotes treated with HU. The phosphorylated RPA2 formed numerous foci that colocalised with γH2AX in control zygotes (n=20) (Fig. 8A, upper panel), indicating that RPA2 is effectively activated in normal zygotes. However, few phosphorylated RPA2 foci were observed in Bcas2^mNull zygotes (n=19) (Fig. 8A, lower panel). Compared with control, phosphorylated RPA2 foci were significantly decreased in Bcas2^mNull zygotes (Fig. 8B). These data suggest a defect of RPA activity in responding to replication stress in the Bcas2^mNull zygotes.

To examine the involvement of the RPA complex in DNA repair at the first mouse cell cycle, normal zygotes were recovered at 23 h post-hCG (PN2) and either injected with Rpa2 mutant (S4A/S8A, Thr21A or S33A) mRNA or treated with TDRL-505, an inhibitor blocking the formation of RPA-ssDNA (Anciano Granadillo et al., 2010; Shuck and Turchi, 2010). After culturing for 7-8 h, the embryos were stained for γH2AX. Zygotes injected with normal Rpa2 mRNA (control) had few γH2AX foci (n=51). However, embryos injected with mutant Ser4A/Ser8A (n=22), Thr21A (n=28) or Ser33A (n=23) mRNA had increased γH2AX foci in both parental pronuclei (Fig. 8C,D). In addition, TDRL-505 treatment dramatically increased γH2AX foci in PN5 zygotes (Fig. 8E,F). TDRL-505 treatment also delayed zygote development into two-cell embryos, accompanied by accumulated γH2AX foci (Fig. S7A-C). These data suggest that the RPA complex plays important roles in DNA repair in mouse zygotes.

Together, our data suggest that BCAS2 protects mouse zygotes from replication stress and DNA damage by influencing RPA function during the first cell cycle in mouse development.

BCAS2 plays important roles in alternative splicing and DNA repair in cells. Knockdown of BCAS2 by siRNA in HEK293T and HeLa cells results in no obvious defects in DNA repair, but cells are defective in responses to campotothecin (CPT)-induced DNA lesions (Wan and Huang, 2014). In the present study, we demonstrated that BCAS2 responds to both exogenous and endogenous DNA lesions in mouse zygotes. Oocyte-specific ablation of BCAS2 results in the developmental arrest of cleavage-stage embryos accompanied by accumulated DNA damage and defects in DNA replication. Thus, our results suggest that BCAS2-mediated DNA repair is important to maintain the genome integrity of mouse early embryos. Persistent DNA damage in Bcas2^mNull zygotes and two-cell embryos but not in BCAS2 knockdown cells suggests that early mouse embryos might be more sensitive to damaged DNA than normal cells.

We previously showed that BCAS2 responds to CPT-induced DNA damage by interacting with RPA (Wan and Huang, 2014). Consistently, BCAS2 mutants unable to bind RPA1 failed to repair damaged DNA in mouse zygotes. Furthermore, phosphorylated RPA2 did not form foci in response to replication stress in Bcas2^mNull zygotes. In addition, disruption of the activity of the RPA complex resulted in accumulated DNA damage in mouse zygotes. Our data suggest that the maternal BCAS2 pathway potentially protects genomic integrity through the RPA complex in mouse early embryonic development.

In somatic cells, deficient DNA repair usually activates cell cycle checkpoints such as the CHK1 and p53-p21-cyclin-CDKs pathways (Arias-Lopez et al., 2006; El-Deiry et al., 1992; Espinosa et al., 2003; Jazayeri et al., 2006; Levine, 1997; Macleod et al., 1995; Miyashita and Reed, 1995). Although many mutant strains display embryonic lethality at the preimplantation stage, few mutant mouse lines exhibit embryonic lethality at early cleavage stages (Friedberg and Meira, 2006). Because of the maternal contribution of these genes, the embryonic lethal phenotypes observed in mutants deficient in DDR might underscore the importance of DDR during preimplantation development. Our finding that both the p53 and CHK1 pathways are activated in response to damaged DNA in Bcas2^mNull two-cell embryos suggests that mouse early embryos could arrest development in response to DNA damage via these pathways as early as the two-cell stage.

Bcas2^mNull zygotes cleaved into two-cell embryos without obvious developmental delays but arrested at the zygotic stage with increased DNA damage upon HU treatment, supporting the notion that the checkpoint is activated in a dosage dependent manner in mouse zygotes (Adiga et al., 2007; Grinfeld and Jacquet, 1987). Thus, it is possible that damaged DNA does not activate the checkpoint in Bcas2^mNull zygotes, which cleave into two-cell stage embryos. As DNA lesions accumulate and checkpoints gradually activate at the two-cell stage, Bcas2^mNull embryonic development eventually arrests at late cleavage stages. In addition, the malfunction of ZGA might also contribute to two-cell arrest in Bcas2^mNull two-cell embryos.

Because BCAS2 is involved in multiple functions in cells and as a result of the limitations of current models and analyses, we could not rule out the possibility that BCAS2 might also indirectly affect the function of RPA in DNA repair in mouse zygotes or the roles of BCAS2 during oogenesis and early development. However, established mouse embryos lacking maternal BCAS2 not only allow us to study the physiological function of BCAS2 in DNA repair during early embryonic development but might also provide useful models for elucidating the roles of BCAS2 in ZGA, DNA demethylation and other important events involved in mouse early embryogenesis.

In compliance with the guidelines of the Animal Care and Use Committee of the Institute of Zoology at the Chinese Academy of Sciences, the Bcas2^loxp/loxp;Zp3-Cre mouse line was maintained on a mixed background (129/C57BL/6). To construct a conditional mouse line for targeting the Bcas2 allele, exons 3 and 4 of Bcas2 were flanked by LoxP sites (Fig. S2, Table S1). The recombined embryonic stem cells (ESCs) were screened by PCR with two primers (Neo-F and screen-R, Table S1), further confirmed by sequencing (data not shown), and injected into C57BL/6 blastocysts. One chimera transplanted the targeted allele through the germline, and the Neo cassette was deleted by mating Bcas2^+/fl-Neo and Rosa26-Flp mice to produce Bcas^+/fl mice (Fig. S2). Mice were genotyped using tail DNA as PCR templates with three oligo-primers (F-WT, F-KO, R, Table S1; Fig. S3A). The conditional mouse line was established by crossing Bcas2^fl/fl mice with Zp3-Cre mice to specifically ablate BCAS2 in growing oocytes (Fig. S2). Bcas2^mNull embryos were obtained by mating Bcas2^fl/fl;Zp3-Cre females with normal males.

Fertility assessment was conducted by 2:1 caging of Bcas2^fl/+;Zp3-Cre and Bcas2^fl/fl;Zp3-Cre females (n>5) with wild-type males for over three months. Successful mating was verified by the presence of a post-coital vaginal plug, after which mice were caged separately. For histology analysis, ovaries from adult females were fixed in Bouin's solution, sectioned, rehydrated, and stained with hematoxylin and 1% eosin (ZSGB-BIO).

For in vitro fertilisation (IVF), MΙΙ oocytes were harvested with M2 medium from 7- to 8-week-old females after injections of PMSG and hCG (Ningbo Second Hormone Factory) and transferred into human tubular fluid (HTF) buffer. Next, 5×10⁶/ml capacitation sperm were added to HTF containing MΙΙ oocytes, which were then cultured in a humidified atmosphere of 5% CO₂ for the next 6 h and then transferred to KSOM at the designated stages (Fraser and Drury, 1975).

For harvesting oocytes or embryos, females were intraperitoneally injected with PMSG and hCG. Oocytes or embryos were manipulated in M2 at the indicated time points after hCG injection: MΙΙ oocytes, 13 h; zygotes, 23-30 h; two-cell embryos, 33-48 h; morula, 71 h; blastocyst, 98 h. Germinal vesicle stage (GV) oocytes were obtained 48 h after PMSG.

MΙΙ oocytes were manipulated with M2 and transferred to HTF media covered with mineral oil. UV irradiation was performed using UV-C lamps (254 nm). After UV irradiation, MΙΙ oocytes were transferred to another HTF drop and fertilised in vitro with normal sperm. For UV irradiation two-cell embryo experiments, embryos were obtained at 37 h post-hCG, irradiated for 1 min, and cultured 3 h prior to fixation and staining.

For Pifithrin-α (Sigma) treatment, Bcas2^mNull zygotes from IVF were cultured for 15 h, treated with either 0.1% DMSO or 100 µM Pifithrin-α for 43 h, and imaged. At 48 h post-hCG, 25 Bcas2^mNull two-cell stage embryos that were treated with either 0.1% DMSO or 100 µM Pifithrin-α were collected for qRT-PCR. For HU and TDRL-505 treatment, control and Bcas2^mNull zygotes were recovered at 25 h post-hCG and exposed to 4 mM HU (Sigma) for 3 h or 100 µM TDRL-505 (Calbiochem) for 7-8 h. Zygotes were either fixed for immunostaining with anti-phosphorylated-RPA2 or anti-γH2AX antibody or cultured for 48 h post-hCG for assessing the percentage of embryos progressing to the two-cell stage.

For PT (PFA–Triton) conditions, embryos were fixed in 4% PFA for 30 min at 37°C, washed three times with PBS, and permeabilised with 0.5% Triton X-100 in PBS (5 min, RT). Embryos were washed three times with PBST (PBS with 0.1% Triton X-100), blocked with 5% donkey serum (1 h, RT), incubated in primary antibodies (Table S2, overnight at 4°C), washed with PBST, and incubated with secondary antibody (Table S2, 1 h at RT). The images were obtained with the LSM780 (Zeiss).

For TP (Triton–PFA) conditions, the zona pellucida of embryos were removed with acid Tyrode's solution (Sigma). Embryos were permeabilised with 0.1% Triton X-100 in PBS (PBST) for 1 min on ice and then fixed with 4% PFA (37°C, 30 min). The embryos were stained as described above.

Immunofluorescence images were analysed with ZEN lite 2011 (Zeiss). A z-series was performed to cover the maximal radius of individual pronuclei. For data presentation, the separated images of both pronuclei were merged.

Mouse oocytes or early embryos were washed with PBS before freezing in 10 µl RIPA containing protease inhibitor cocktail (Roche) and stored at −80°C. The samples were separated on NuPage 4-12% Bis-Tris gels with MOPS SDS running buffer and transferred to PVDF membranes (Invitrogen). The blots were pre-treated with Superblock Blocking Buffer (Thermo Fisher Scientific), incubated with primary antibodies (Table S2) overnight at 4°C, washed six times with PBST (5 min), incubated with horseradish peroxidase-conjugated secondary antibodies (Table S2, 1 h at RT), detected with SuperSignal West Dura Extended Duration Substrate (Thermo Fisher Scientific) and imaged with a LAS-3000 (FujiFilm).

Embryos were recovered at E0.5 or E1.5 and cultured in KSOM containing 50 mM EdU at 37°C for 30 min prior to fixation. Detection of incorporated EdU was performed using the Cell-Light EdU Apollo 567 In Vitro Imaging Kit (RiboBio) according to the manufacturer's protocol. Immunofluorescent staining was performed as described above. Control and Bcas2^mNull zygotes were examined by TUNEL assay according to the manufacturer's protocol (Roche).

Bcas2 and Rpa2 were amplified from mouse ovary cDNA with specific primers (Table S3). Bcas2 and Rpa2 cDNA were inserted into modified pEGFP-N1 and Pcs2⁺-myc vectors, respectively. Rpa2 point mutations were generated by mutating Rpa2 (Table S3) in the Rpa2-pcs2⁺-myc vector (Toyobo). Linearised plasmid was used for in vitro transcription with the mMESSAGE mMACHINE Kit (Ambion) and purified with an RNeasy MinElute Cleanup Kit (Qiagen). mRNA was dissolved in nuclease-free water at 100 ng/µl. Mouse zygotes before the PN3 stage were microinjected with 5-10 pl of mRNA and cultured until the next processes.

The mRNA was isolated from oocytes or embryos using the Dynabeads mRNA Direct Micro Kit (Ambion). Reverse transcription of RNA was performed using PrimeScript RT Reagent Kit (Takara). Quantitative PCR was performed using EvaGreen qPCR MasterMix (Applied Biological Materials) with specific primers (Table S4). Measurements were performed with three independent biological replicates.

Mouse zygotes and two-cell embryos from control and Bcas2^mNull females were stained as described above. Z-series images were obtained to cover the maximal radius of individual pronuclei (nuclei) and were merged for each embryo by ZEN lite 2011 (Zeiss). Each picture was processed with Enhanced Contrast by saturated pixels 0.4% using ImageJ software. The foci in each picture were counted with Find Maxima by altering Noise Tolerance. The data were presented as the means±s.e.m.

Quantitative analyses were carried out with GraphPad Prism software with at least 3-5 independent biological samples and expressed as the means±s.e.m. P-values of comparisons between two groups were calculated using Student's t-test. Embryo developing beyond two-cell stage was the dichotomous dependent variable and Pifithrin-α treatment was covariates in the model using binary logistic repression analysis of SPSS software. P <0.05 was considered as significant.

We thank Shiwen Li (Institute of Zoology, Chinese Academy of Sciences) for technical assistance and members of Li's laboratory for discussions and critical comments on the manuscript.