Cyclin B3 is required for metaphase to anaphase transition in oocyte meiosis I

The meiotic cell cycle, which comprises two consecutive M phases, is crucial for production of haploid germ cells. In both mitotic and meiotic cell cycles, M phases share cyclin B-CDK1 as the key controller to ensure the reliability of cell cycle progression. During prometaphase (pro-MetI), spindle assembly checkpoint (SAC) proteins sequester Cdc20, the anaphase-promoting complex/cyclosome (APC/C) activator, and prevent it from promoting securin and cyclin B ubiquitylation (Thornton and Toczyski, 2003). In metaphase, when all kinetochores are attached to microtubules, Cdc20 liberates from SAC and leads to complete APC/C activity with degradation of both securin and cyclin B. Securin is an inhibitory chaperone of separase, and its destruction promotes separase cleavage of cohesin complexes, which initiates sister–chromatid separation and anaphase onset (Uhlmann et al., 1999). Meanwhile, the degradation of cyclin B reduces maturation-promoting factor or mitosis-promoting factor (MPF) activity and further improves the activity of separase and Cdh1-induced APC/C activation, which guarantees anaphase progression (Vázquez-Novelle et al., 2014).

Cyclin synthesis and degradation cooperate with cyclin-dependent kinases (CDKs) to control the progression of meiosis and mitosis. Although most of the basic cyclins used in the meiosis metaphase are analogous to those used in mitosis, the lingering question is whether the proofreading function of cyclins during mitosis are equally significant during meiotic division. The primary cyclins in metaphase are B-type cyclins, which contain at least three types of cyclin B (cyclin B1, B2, and B3) in mammals, and it appears that cyclin B1 (Ccnb1) is primarily responsible for MPF activity (Jones, 2004). Mice lacking Ccnb1 were not viable, whereas cyclin B2-null mice had no apparent defects (Brandeis et al., 1998). However, recent reports showed cyclin B2 could compensate for Ccnb1 in oocyte meiosis I (Li et al., 2018), which implies that there are specific modulations in the meiotic cell cycle regulation.

Cyclin B3 (Ccnb3) shares homology with A- and B-type cyclins (Gallant and Nigg, 1994) and is conserved during higher eukaryote evolution (Sigrist et al., 1995; Jacobs et al., 1998; Parry and O’Farrell, 2001; Lozano et al., 2002; Nguyen et al., 2002; Refik-Rogers et al., 2006; Tarailo-Graovac and Chen, 2012; Zhang et al., 2015). Previous studies have shown that females lacking Ccnb3 are sterile, with oocytes unable to complete meiosis I in Drosophila (Jacobs et al., 1998), implying that Ccnb3 may have a special role in meiotic regulation.

To clarify the function of Ccnb3 in meiosis in mammalian species, we generated Ccnb3 mutant mice via CRISPR/Cas9 and found that Ccnb3 mutation caused female infertility due to the failure of metaphase–anaphase transition in meiosis I. Ccnb3 was found to be necessary for APC/C activation to initiate anaphase I (AnaI), but not required for oocyte maturation, meiosis II progression, or early embryonic development. Our findings may shed light on the differential cell cycle regulatory mechanisms between meiosis and mitosis, as well as between male and female meiosis.

We first detected the expression pattern of Ccnb3 by quantitative PCR (Q-PCR) and found that its mRNA had a similar expression pattern with Ccnb1 during oocyte in vitro maturation (IVM), which implied that Ccnb3 may play an important role in meiosis cell cycle regulation (Fig. 1 A). To study this role of Ccnb3, we generated Ccnb3 mutant mice (referred to as Ccnb3^△/Y and Ccnb3^△/△ for male and female mutants, respectively) via CRISPR-mediated deletion of 29 bp in exon 3 of the Ccnb3 gene located on the X chromosome (Fig. S1 A). The genotypes and protein expression of Ccnb3 mutant mice were verified by PCR (Fig. 1 B) and Western blot (Fig. 1 C). By natural mating, we found that the Ccnb3^△/△ mice were infertile, while the Ccnb3^△/Y mice showed normal fertility (Fig. 1 D). To find the detailed mechanism of female infertility, we examined ovary development and folliculogenesis in Ccnb3^△/△ female mice. The H&E staining results showed that the Ccnb3^△/△ ovary development was normal (Fig. 1 E), and the number of oocytes superovulated from Ccnb3^△/△ mice was similar to that from WT female mice (referred to as Ccnb3^WT/WT; Fig. S1 B). To investigate whether the infertility was caused by embryonic lethality, we collected embryos from Ccnb3^△/△ female mice with vaginal plugs, after mating with Ccnb3^WT/Y (male WT) mice. All the collected fetuses were degenerated before embryonic day 7.5 (E7.5; Fig. 1 F). These results showed that Ccnb3 mutation leads to female infertility, while the defects were caused by embryonic lethality rather than abnormal follicular development.

Although the number of superovulated oocytes from Ccnb3^△/△ mice was identical to those from Ccnb3^WT/WT mice, the first polar body (PB1) was not observed in the Ccnb3^△/△ oocytes (Fig. S1, C and D). We suspected that the defects in Ccnb3^△/△ mice may occur at the time of meiosis progression. To confirm this hypothesis, we analyzed the meiosis maturation process of the Ccnb3^△/△ oocytes using IVM (Fig. 2 A and Fig. S2, A and B) and living cell–tracking assays (Fig. 2 B). We found that the fully grown germinal vesicle (GV)–stage Ccnb3^△/△ oocytes resumed meiosis with successive occurrences of GV breakdown (GVBD) and MetI spindle formation (Fig. 2 C). The efficiency of GVBD of Ccnb3^△/△ oocytes was equivalent to that of the WT (81.1 vs. 88.5%, respectively; Fig. 2 A). However, a successive blockade in MetI was exclusively generated in Ccnb3^△/△ oocytes, failing to extrude PB1 and still maintaining bivalent homologous chromosomes after 10 h of GVBD, when the Ccnb3^WT/WT oocytes had entered the MetII, characterized by visible PB1 and univalent sister chromatids (Fig. 2, B and D). We further observed the REC8 protein, a specific meiotic cohesin and found that REC8 was persistently presented on the chromosome arms with a typical feature of MetI when AnaI was initiated, which indicated that the Ccnb3^△/△ oocytes were arrested at MetI with inseparable homologous chromosomes (Fig. 2 D).

To evaluate the effect of Ccnb3 mutation on the developmental capacity of oocytes, we injected WT sperms into the Ccnb3^△/△ oocytes through intracytoplasmic sperm injection (ICSI). Interestingly, fertilized Ccnb3^△/△ oocytes were able to extrude the PB and develop into blastocysts (Fig. 2 E) with similar efficiency as those embryos in Ccnb3^WT/WT group (75.6 vs. 80.6%; Table S1), indicating that Ccnb3 mutation did not affect the preimplantation developmental capacity of embryos. We further established the embryonic stem cell (ESC) lines from these ICSI-derived blastocysts (Fig. 2 E) and analyzed the DNA content and karyotype of the Ccnb3^△/△ oocyte–derived ESCs. We found that the Ccnb3^△/△ ESCs were triploid and maintained a stable “57 + XXY” karyotype (Fig. 2, F and G). This triploid karyotype may account for the E7.5 degeneration (Fig. 1 F), which was consistent with previous reports (Niemierko, 1981).

As Ccnb3^△/△ oocytes can extrude PB after ICSI, we were curious about what kind of PB extrusion occurred in MetI-arrested oocytes after parthenogenetic activation (PA) or ICSI. In other words, the MetI-arrested oocytes may either continue to complete meiosis I by releasing the PB1 or directly skip the meiosis I to complete meiosis II by separating the sister chromatids and releasing the PB2. The genome-wide single nucleotide polymorphism (SNP) analysis on parental chromosomes is feasible for validating the genetic provenance of ESCs, which distinguished the PA-derived ESCs from fertilized ESCs (Kim et al., 2007; Mai et al., 2007). Hence, we analyzed the whole genome SNPs in Ccnb3^△/△ PA-derived ESCs to distinguish the homologous chromosome separation from the sister–chromatid separation. There were two types of ESCs established from Ccnb3^△/△ PA-embryos, one of which contains tetraploid set of chromosomes due to no PB extrusion by cytochalasin B (CB) treatment (referred to as 4N ESCs) and represents the whole genome SNPs of the CD-1 mice. The other contains diploid set of chromosomes due to one PB extrusion after PA (without CB treatment; referred to as 2N ESCs), which mimic the PB extrusion of MetI-arrested oocytes after ICSI or PA.

Whole genome DNA sequencing analysis showed that all SNPs in 4N ESCs were heterozygous. While in the 2N ESCs, if the segregation happened between homozygous chromosomes, SNPs would be almost homozygous, and only a fraction of the crossover parts would be heterozygous; and if segregation happened between sister chromatids, SNPs would be almost heterozygous, and only the crossover parts would be homozygous (Fig. 2 H). Our SNP analysis showed that in the 2N ESCs, SNPs were almost heterozygous, and only a fraction of crossover parts was homozygous, which implied that the sister chromatids of Ccnb3^△/△ oocytes were separated after PA (Fig. 2 H and Fig. S2 C). In other words, fertilization or PA triggered separation of sister chromatids instead of homozygous chromosomes in Ccnb3^△/△ oocytes.

To test whether the MetI arrest was due to the insufficiency of APC/C activity in Ccnb3^△/△ mice, we traced dynamic changes of the APC/C substrates, Ccnb1 and securin, during IVM by injecting mRNA of Ccnb1-EGFP or securin-EGFP into GV-stage oocytes (Thornton and Toczyski, 2003). The green fluorescence intensity remained almost unchanged during meiosis I progression in the Ccnb3^△/△ oocytes, while the WT oocytes experienced a significant decline with PB1 extrusion (Fig. 3, A–D). Securin is known for its role in inactivating the cohesin-cleaving enzyme, separase, until the metaphase–anaphase transition (Marangos and Carroll, 2008). We further confirmed inefficient degradation of endogenous securin protein by immunofluorescence staining (Fig. 3 E). These results indicated that there was no effective APC/C activity for Ccnb1 and securin degradation in Ccnb3^△/△ oocytes, and Ccnb3 is necessary for fully activated APC/C to induce the MetI–AnaI transition.

Fully activated APC/C activation requires the satisfaction of SAC (Nilsson et al., 2008). To evaluate the SAC dynamic changes during meiosis I, we detected the SAC proteins, BUB3 and MAD2, by immunofluorescence (Homer et al., 2005a,b; Li et al., 2009). Both BUB3 and MAD2 were deprived from the kinetochores during the pro-MetI to MetI transition, both in the WT and Ccnb3^△/△ oocytes (Fig. 3 F and Fig. S3 A), which suggested that Ccnb3 mutation does not affect the normal function of SAC.

As Ccnb3 mutation leads to feeble degradation of both Ccnb1 and securin at the beginning of anaphase, we wanted to confirm whether direct interference with APC/C substrates in MetI could rescue the phenotype of Ccnb3-deficient oocytes. We knocked down Ccnb1 and securin by injecting corresponding siRNA into GV-stage oocytes, respectively (Fig. S3 B). Surprisingly, knocking down either of them did not affect the development rate of GVBD (Fig. S3 C), while only reducing securin could restore the PB1 extrusion of Ccnb3^△/△ oocytes (Fig. 4, A and B). As CDK1 activity in the Ccnb3 mutant oocytes is significantly higher at AnaI (10 h after GVBD; Fig. S3 D), we tried to interfere with pan-CDK1 activity at MetI (6 h after GVBD) using the small molecule inhibitor, RO-3306. As expected, acute pharmacological inhibition of CDK1 partially rescued MetI arrest caused by Ccnb3 mutation and led to the PB1 extrusion (Fig. S3 E).

Our results collectively show that Ccnb3 mutant oocytes could not fully activate the APC/C and decline the CDK1 activity, which led to the failure of AnaI onset. Knockdown of securin or inhibition of CDK1 activity could rescue these defects. Securin and CDK1 worked synergistically to regulate the activity of separase, which cleaves the cohesin complexes to dissolve homologous chromosomes and sister chromatid at anaphase onset in meiosis and mitosis (Gorr et al., 2005; Stemmann et al., 2006). These findings indicated that Ccnb3 functioned to fully activate the APC/C to activate separase during MetI–AnaI transition by declining securin and CDK1 activity, resulting in a triumphant AnaI onset.

To investigate whether Ccnb3 protein could recover the MetI arrest in Ccnb3^△/△ oocytes, we injected Ccnb3 mRNA into Ccnb3^△/△ oocytes at GV stage (Fig. S3 F). Our results showed that replenishment of Ccnb3 mRNA could reverse the Ccnb3 mutant phenotypes by recovering PB1 extrusion (Fig. 4 C). The recovered PB1 extrusion, characterized by homologous chromosome separation and disappearance of bivalent chromosomes, was confirmed by chromosome spread assay (Fig. 4 D). Moreover, the rescued efficiency of PB1 release is comparable to that of the WT (Fig. 4 E), which further indicates the irreplaceable role of Ccnb3 in meiosis I. The APC/C activity in Ccnb3-rescued oocytes was further detected by tracing changes in fluorescence with exogenous injection mRNA of securin-EGFP. Securin protein was quickly degraded in time at the MetI–AnaI transition in Ccnb3 rescued oocytes (Fig. 4 F). These results suggest that Ccnb3 helps to fully activate the APC/C activity and promote the MetI–AnaI transition and PB1 extrusion.

As cyclins usually collaborate with CDKs to ensure precise operation of the cell cycle (Kishimoto, 2003; Brunet and Maro, 2005; Satyanarayana and Kaldis, 2009), we speculated that CCNB3 may work as a ligand combined with CDKs, like other cyclins. To explore the functional partner of CCNB3, we detected the interaction between CCNB3 and CDK1 by immunoprecipitation (IP) in cells coexpressing the two proteins, confirming that CCNB3 did interact with CDK1 (Fig. 4 G). We further confirmed this result strictly by testing endogenous protein interactions in Ccnb3^WT/Y and Ccnb3^△/Y testicular tissue, which showed CDK1 had interacted with CCNB3 in Ccnb3^WT/Y, but had no interaction in Ccnb3^△/Y testicular tissue (Fig. 4 H).

Cyclins play different roles in regulation of male and female meiosis. The infertility of female mice and contrasting fertility of male mice, both with Ccnb3 mutations, suggested that Ccnb3 only functioned in females meiosis, which may be related to the long duration of female meiosis I (Gemzell, 1962) and reflected different meiotic regulation mechanisms between the two sexes. Another possibility is that other sexual tendentious cyclins may compensate for the function of Ccnb3 in males, for instance, cyclin A1. Cyclin A1 is a male-specific cyclin in mice, and its deletion leads to a blocked spermatogenesis before the first meiotic division, whereas female mice remain normal following its deletion (Liu et al., 1998). Nonetheless, the gender bias of cyclins in meiotic regulation is unique, which needs to be further explored.

Successfully execution of metaphase–anaphase transition in eukaryotic cell division requires metaphase cyclins scheduled degradation at different times, especially cyclin B. Our findings that Ccnb3 played an essential role for MetI–AnaI transition during female meiosis I are consistent with the results in Drosophila (Yuan and O’Farrell, 2015) and Caenorhabditis elegans (Tarailo-Graovac and Chen, 2012), which implies that the specific role of Ccnb3 in female meiosis is conserved. We also found that knockdown of Ccnb1 without blocking GVBD could not recover the defects in oocytes lacking Ccnb3. A previous report showed that oocyte-specific knockout of Ccnb1 may lead to compensatory increasing of Ccnb2 with normally GVBD occurrence (Li et al., 2018). We speculated that Ccnb2 or cyclin A2 (Murphy et al., 1997) may complement Ccnb1 for maintaining high CDK1 activity to initiate GVBD and block the AnaI onset in oocytes lacking Ccnb3.

Beyond expectation, fertilization or PA directly switched the MetI-arrested oocytes into a fertilized status with PB2 instead of PB1 extrusion. How the spindle accomplishes assembly in the MetI-arrested oocytes will be an interesting question to study further. We verified the normal separation of sister chromatids in Ccnb3^△/△ oocytes after activation and the preimplantation developmental capacity of fertilized Ccnb3^△/△ oocytes have no defects, which implied that Ccnb3 is a meiosis-specific cyclin required for meiosis I, rather than meiosis II. In addition, the genome-wide SNP analysis provides a reliable application for analyzing the pattern of chromosome segregation.

Previous reports showed that mitotic CDKs phosphorylated APC/C subunits, which is required to allow APC activation by Cdc20 (Rudner and Murray, 2000; Chang et al., 2015). Combined with the above results, we speculated that Ccnb3 binding to CDK1 may function as a kinase which is necessary for full activation of APC/C by directly phosphorylating securin or other APC/C subunits for their timely degradation at AnaI onset. Securin is an inhibitor of separase, and we found that knockdown of securin recovered the MetI arrest in Ccnb3 mutant oocytes. Down-regulation of securin liberated separase, which may directly cleavage of cohesin to separate homologous chromosome. It has been reported that separase acts as repressor of CDK1/cyclin B, which inactivates MPF to ensure anaphase, progress in both mitosis and meiosis I (Gorr et al., 2005, 2006; Shindo et al., 2012). Meanwhile, both separase liberation and CDK1 inactivation may account for the resumption of MetI arrest in Ccnb3 mutant oocytes after securin knockdown, for which a concrete mechanism needs to be further explored. Besides, the meiosis-specific REC8, which is phosphorylated by different kinases (Alexandru et al., 2001; Lee and Amon, 2003; Yu and Koshland, 2005; Katis et al., 2010) to facilitate cohesin to cleavage from chromosomes, persistently remained on the homologous chromosome arm in Ccnb3^△/△ oocytes, which implies CCNB3/CDK1 complex may serve as an undiscovered candidate kinase for phosphorylation of REC8.

In this study, we found that Ccnb3 mutation caused female mouse infertility with the failure of metaphase–anaphase transition in oocyte meiosis I, while the Ccnb3 mutant male mice had normal fertility. Oocytes lacking Ccnb3 are characteristic of insufficient APC/C activation and normal preimplantation capacity after fertilization. Meanwhile, we found that Ccnb3 directly interacts with CDK1, which may exert an essential kinase activity to separate homologous chromosomes. Although we have elucidated the unique role of Ccnb3 in female meiosis, the direct target of Ccnb3 have not been found yet. The core regulatory machinery of Ccnb3 in metaphase to anaphase transition seems to be various in female and male meiosis. Similar findings were obtained independently with a different targeted mutation in Ccnb3 (Karasu et al., 2019). Our study reveals that Ccnb3 serve as a female meiosis-specific cyclin for metaphase–anaphase transition in meiosis I, which opens an avenue to elucidate the unique cell cycle regulation mechanisms for meiosis in mammals. The specific target of Ccnb3 need to be further studied to unraveling the differences of cyclins in female and male meiotic regulation.

Specific pathogen free–grade mice were obtained from Beijing Vital River Laboratories and housed in the animal facilities of the Chinese Academy of Sciences. All animal studies were performed in accordance with the Guidelines for the Use of Animals in Research issued by the Institute of Zoology, Chinese Academy of Sciences. The mice used in this study were CD-1 strains, and genome editing was performed by CRISPR/Cas9. Six sgRNAs were designed for targeting different sites of the mouse Ccnb3 gene (sgRNA-Ccnb3-1∼6). The efficiency of sgRNAs was verified by mouse fibroblast cell transfection, cell sorting, subsequent PCR identification, T7 endonuclease I (T7EI) assay, and Sanger sequencing. SgRNA-Ccnb3-1 showed the highest mutation ratio (12:20; 60%) according to T7 EI and Sanger sequencing, which was selected for in vitro transcription and injection. The injection concentration by ICSI of Cas9 mRNA/sgRNA-Ccnb3-1 was 100:50 ng/µl. The biallelic mutations of Ccnb3 in mouse embryos were produced with relatively high efficiency via zygote injection. Finally, we screened out the Δ29 bp mice for the next experiment. All sgRNAs are listed in Table S2.

Total RNA was extracted with TRIzol reagent (Invitrogen, 15596-018) from 20 oocytes in each group. Q-PCR was performed with Thunderbird SYBR qPCR Mix (Toyobo), using total cDNA as the template, in a total volume of 20 µl, including the primer, and ROX (Toyobo). The thermal cycling conditions of conventional real-time PCR were 1 min at 95°C, followed by 45 cycles of 15 s at 95°C, 15 s at 60°C, and 45 s at 72°C, and the melting curve conditions were 1 min at 95°C, 30 s at 60°C, and 30 s at 95°C. All reactions were performed using an Agilent Technologies Stratagene Mx3005P Real-Time PCR System (Applied Biosystems). Relative gene expression was analyzed based on the 2-ΔΔCt method with Actin as internal control. At least three independent experiments were analyzed. All primers are listed in Table S2.

Ovaries from WT and Ccnb3 mutant female mice were fixed in 4% paraformaldehyde at room temperature, overnight. The ovaries were dehydrated stepwise through an ethanol series (70, 80, 90, and 100% ethanol) and processed for paraffin embedding. 5-µm sections were cut with a Leica slicing machine (Leica Biosystems, RM2235) and mounted on poly-d-lysine–coated glass slides (Zhong Shan Golding Bridge Biotechnology). After dewaxing and hydration, the sections were stained with H&E, using standard methods and imaged with a Leica Aperio VERSA 8 microscope (Leica Biosystems).

We isolated GV stage oocytes from minced ovaries of 8–10-wk-old CD-1 female mice. For the in vivo method, oocytes collected after injection of 7.5 international units (IU) of pregnant mare’s serum gonadotropin, followed by 7.5 IU human chorionic gonadotropin. For the in vitro method, GV oocytes were superovulated from adult mice by injection of 7.5 IU pregnant mare’s serum gonadotropin after 44 h and cultured in vitro. The number of oocytes with PB1 at 8 h after GVBD when the PB1 had just been released. Oocytes were cultured in M2 medium for at least 12 h. The culture was conducted in an incubator under environmental conditions of 5% CO₂, 37°C, and saturated humidity. To evaluate the developmental capacity of oocytes, oocytes were stained with Hoechst 33342 at different time during in vitro culture. For the rescue experiment, oocytes were cultured in media containing RO-3306 at different concentrations and for different lengths of time as indicated in the figure legends.

The fluorogenic, cell-permeable reagent SiR-Tubulin (Cytoskeleton, CY-SC002) was used to image tubulin in living cells. GV-stage oocytes were injected with 10 pl of 100 ng/µl mRNA encoding securin-EGFP in M2 medium containing 250 mM dbcAMP using methods described elsewhere (Jaffe and Terasaki, 2004). Following mRNA injection, oocytes were cultured for 2–3 h at 37°C to allow securin-EGFP expression. Oocytes were cultured in dbcAMP-free M2 medium placed in a European Molecular Biology Laboratory environmental microscope incubator (GP106), allowing cells to be maintained in a 5% CO₂ atmosphere at 37°C with humidity control during imaging. Time-lapse image acquisitions were performed using a customized Zeiss LSM510 META confocal microscope equipped with a 532-nm excitation laser, a long-pass 545-nm emission filter, a 403 C-Apochromat 1.2 NA water immersion objective lens, and an in-house developed 3D tracking macro (Rabut and Ellenberg, 2004).

Chromosome spreads of mouse oocytes during meiotic maturation were prepared using methods previously described (Hodges and Hunt, 2002). Oocytes were briefly exposed to acid Tyrode’s solution (Sigma) to remove the zona pellucida under the microscope to avoid overdigestion. After a brief recovery in M2 medium, the oocytes were transferred onto glass slides and fixed in a solution of 1% paraformaldehyde in distilled H₂O (pH 9.2) containing 0.15% Triton X-100 and 3 mM dithiothreitol. The dried chromosome spreads were stained with 10 µg/ml Hoechst 33342 for DNA counterstaining, following immunofluorescence staining when required. The slides were dried slowly in a humidified chamber for several hours and then blocked with 2% BSA in PBS for 1 h at room temperature or overnight at 4°C and incubated with primary antibodies overnight at 4°C. After brief washes with washing buffer, the slides were then incubated with corresponding secondary antibodies for 1 h at room temperature. Rabbit anti-REC8 antibody (Abcam, ab192241) for marking homologous chromosomes was used as the primary antibody, and appropriate secondary antibody conjugated with Alexa Fluor Cy3 (Invitrogen) was used. The samples were observed under a laser-scanning confocal microscope (Zeiss LSM 780).

CD-1 mouse blastocysts were used to derive ESCs. The culture medium composed of N2B27 medium (Gibco), 1 mM l-glutamine, 0.1 mM β-mercaptoethanol, 50 U/ml penicillin, 50 µg/ml streptomycin, 3 µM CHIR99021 (Stemgent), 1 µM PD0325901 (Stemgent), and mLIF (ESGRO). The embryos were incubated at 37°C in a 5% CO₂ incubator for 4–5 d, and then the formed outgrowths (named passage 0) were picked separately using a glass pipette and dissociated with 0.25% trypsin-EDTA solution (Gibco). The cells were replanted and survived from the first trypsinization of the outgrowth were counted as passage 1. For routine passage, the passage ratio was ∼1:8, and mouse ESCs were passaged around every 3 d.

Ccnb3^△/△ ESCs were purified and analyzed by FACS. Single-cell suspensions were obtained by trypsin-EDTA digestion and repetitive pipetting and sieved through a 40-mm cell strainer. Cells were incubated with 10 µg/ml Hoechst 33342 (Invitrogen, H3570) for 15–20 min at 37°C before analysis. Data were collected with MoFlo XDP cell sorter (Beckman-Coulter). Ccnb3^WT/WT ESCs (2N) were used as a control.

Ccnb3^△/△ ESCs were incubated with 0.2 mg/ml nocodazole (Sigma, M1404) for 3 h. After trypsinization, the ESCs were suspended in 0.075 M KCl at 37°C for 30 min. Then, the cells were fixed in solution consisting of methanol and acetic acid (3:1 in volume) for 30 min and then were dropped onto precleaned slides. The cells were stained with Giemsa stain (Sigma, GS500ML) for 15 min after being incubated in 5 M HCl. More than 30 metaphase spreads were analyzed.

Collected oocytes were parthenogenetic activated in Ca²⁺-free CZB (Chatot Ziomek Bavister medium) containing 10 mM Sr²⁺ with (5 nM) and without CB for 6 h. PA-embryos were used to establish ESCs. Two ESCs from Ccnb3^△/△ PA-embryos were established, one of which contained 4N chromosomes with no PB extrusion (+CB, 4N ESCs) and the other containing 2N chromosomes with one PB extrusion after PA (−CB, 2N ESCs). At least 10⁷ ESCs were collected for DNA sequencing. DNA was isolated and checked using the Nano Photometer Spectrophotometer and Qubit 2.0 Fluorometer and 1 µg DNA template according to TruSeq DNA Sample Preparation Guide (Illumina, 15026486 Rev.C) method and process for library preparation. The Illumina HiSeq X Ten, PE150 strategy was used for sequencing the identified library. The clean reads were mapped to the mouse reference genome (mm9) using a Burrows-Wheeler Aligner (version 0.7.15; Rabut and Ellenberg, 2004). After removing duplicated reads, the base distribution for each chromosomal location was calculated using the Pysamstats (version 1.0.0), developed by A. Miles (University of Oxford, Oxford, England, UK). The single nucleotide variations in 4N ESCs sample with >10× coverage were used for the SNP identification. Sites selected for analysis were covered by two types of bases, located within nonrepeat genome regions (RepeatMasker, Institute for Systems Biology). Then, the sites with minor base taking >30% reads were used for the analysis. Next, we calculated the corresponding base frequency distribution in the 2N ESCs sample. During the homologous chromosome separation, the 2N ESCs sample showed the homozygous state in the identified single nucleotide variations sites; whereas, the cell showed sister chromatid separation and the homologous chromosomes were retained.

Oocytes were first fixed in 4% paraformaldehyde at room temperature for 30 min and then permeabilized in PBS containing 0.5% Triton X-100 for 20 min. Next, oocytes were blocked in PBS with 2% BSA for 1 h and then incubated with FITC-α-tubulin (Sigma, F2618) primary antibody at room temperature for 1 h. After washing three times with PBS, DNA was stained with Hoechst 33342. The samples were then mounted on slides and were observed under a laser-scanning confocal microscope (Zeiss LSM 780). At least 40 oocytes were examined in each treatment, and each treatment was repeated three times. All antibodies used in this study are listed in the Table S3.

Magnetic Beads Protein G were coated with 5 µg of the primary antibody in IP wash buffer (50 mM Tris-HCl, pH 7.4, 150 mM sodium chloride, 1 mM MgCl₂, and 0.05% NP-40) containing 5% BSA overnight with rotation at 4°C. Then, we collected mouse testes or specific cells and added IP lysis buffer (150 mM NaCl, 50 mM Tris-HCl, pH 7.4, 1 mM EDTA, 0.1% SDS, 1% NP-40, 0.5% sodium deoxycholate, 0.5 mM DTT, and 1 mM PMSF/cocktail) followed by incubation on ice for 10 min. Next, we centrifuged the IP lysate at 14,000 rpm for 10 min at 4°C, removed 100 µl of the supernatant and added this to 900 µl of the beads–antibody complex in IP Buffer (860 µl IP wash buffer and 35 µl 0.5 M EDTA) and incubated this with rotation overnight at 4°C. After washing, 50 µl elution buffer was added to the immunoprecipitate and the supernatant was used for Western blot. Western blotting was performed as described previously (Wu et al., 2005). The commercial antibodies used were anti-CDK1 (Abcam, ab32094) and anti-cyclin B3 (C517).

MPF concentration was measured by using mouse MPF assay ELISA kit (JiangLai Biology) according to the manufacturer’s instruction. In brief, cell lysates were added into the purified HRP-labeled cyclin B and CDK1 antibodies to form the antibody–antigen–enzyme–antibody complexes, and the complexes were catalyzed by HRP enzyme. After adding the tetramethylbenzidine substrate solution, MPF concentration was determined by measuring the absorbance of converted dye at a wavelength of 450 nm.

Statistical parameters including statistical analysis, statistical significance, and n value are reported in the figure legends. Statistical analyses were performed using Prism Software (GraphPad). For statistical comparison, Student’s t test was used. A value of P < 0.05 was considered significant, and NA stands for significant difference is not available with P > 0.05.

Fig. S1 describes the characters of Ccnb3 mutant mice. Fig. S2 shows the development of Ccnb3 mutant oocytes in meiosis. Fig. S3 exhibits the rescue of MetI arrest caused by Ccnb3 mutation. Table S1 shows the developmental capacity of Ccnb3^△/△ oocytes after ICSI. Table S2 lists the sgRNAs and primers designed in the study. Table S3 lists the reagent and resource used in the study.

We thank Scott Keeney (Howard Hughes Medical Institute; Memorial Sloan-Kettering Cancer Center) and Katja Wassmann (Sorbonne Université, Institut de Biologie Paris Seine) for generously sharing unpublished data and antibody information. We thank Lijuan Wang for assistance with live cell imaging; Shiwen Li and Xili Zhu for assistance with immunofluorescent staining; and Ming Ge for the microinjection of GV oocytes.

This study was supported by the National Key Research and Development Program (2017YFA0103803 and 2018YFC1004500), the China National Postdoctoral Program for Innovative Talents (BX201700243 to L. Wang), the National Natural Science Foundation of China (31621004 and 81571356 to Q. Zhou), and Key Research Projects of the Frontier Science of the Chinese Academy of Sciences (QYZDY-SSW-SMC002 to Q. Zhou).

The authors declare no competing financial interests.

W. Li, Q. Zhou, and Y. Zhang conceived and designed the experiments; L. Wang performed the ICS with the help of C. Liu and H. Sun; L. Zhang performed the Q-PCR, genotyping, and Western blot, with the help of J. Han and P. Li; G. Feng and J. Mao analyzed the DNA sequencing data; other experiments and analysis were performed by Y. Li and Z. He, with the help of X. Yuan and L. Jiang; J. Wang and Z. Li provided insightful suggestions for the manuscript and preparation. Y. Li wrote the manuscript with the help of other authors.