CDC25B is required for the metaphase I-metaphase II transition in mouse oocytes

Mammalian oocytes enclosed in follicles are arrested in prophase of meiosis I (prophase I) until ovulation. Ovulation is triggered by an increase in luteinizing hormone, which triggers a cascade of signaling events that lead to resumption of meiosis (Solc et al., 2010). Cyclin-dependent kinase 1 (CDK1) is essential for resumption of meiosis (Adhikari et al., 2012) because conditionally deleting Cdk1 in mouse oocytes results in infertility due to the inability of Cdk1-deficient oocytes to resume meiosis. CDK1 activity is inhibited during prophase arrest as a consequence of inhibitory phosphorylation of T14 and Y15 (Adhikari et al., 2016), which are mediated by WEE1B (also known as WEE2) and MYT1 (Han et al., 2005; Oh et al., 2010); inhibiting WEE1B and MYT1 in prophase I-arrested oocytes induces resumption of meiosis (Oh et al., 2010). Activation of CDK1 is triggered by the activation of CDC25, a dual specificity protein phosphatase (Sebastian et al., 1993). Three Cdc25 homologs (a, b and c) are present in the mammalian genome. CDC25B (Lincoln et al., 2002) but not CDC25C (Chen et al., 2001) is essential for CDK1 activation and resumption of meiosis in mouse oocytes. The third homolog, CDC25A, cooperates with CDC25B during resumption of meiosis, and its degradation following resumption of meiosis is required for the meiosis I (MI)-meiosis II (MII) transition (Solc et al., 2008).

A high intracellular concentration of cyclic AMP (cAMP) is required to maintain prophase I arrest in mouse oocytes. cAMP activates protein kinase A (PKA), which catalyzes activating phosphorylation of WEE1B and an inhibiting phosphorylation of CDC25B. Luteinizing hormone, which acts on the surrounding follicle cells, induces resumption of meiosis by decreasing intracellular oocyte cAMP concentration by a mechanism that ultimately results in an increase in phosphodiesterase 3A (PDE3A) activity in the oocyte (Jaffe and Egbert, 2017). The increase in PDE3A activity decreases oocyte cAMP concentration and in turn PKA activity, which results in a decrease in WEE1B activity and an increase in CDC25B activity, the next result being CDK1 activation (Jaffe and Egbert, 2017; Solc et al., 2010).

In human somatic cells, CDC25B cooperates with CDC25A to induce mitotic entry (Lindqvist et al., 2005). CDC25B is required for centrosomal microtubule nucleation in prophase by activating the cytoplasmic pool of CDK1 (Gabrielli et al., 1996). siRNA-mediated Cdc25b knockdown in somatic cells delays centrosome separation (Lindqvist et al., 2005), and acute inhibition of CDC25s with BN82685, a specific small-molecule inhibitor in prometaphase cells, delays spindle assembly (Cazales et al., 2007). In addition, Skp1-cullin-1-F-box protein (SCF)-dependent CDC25B degradation during the metaphase-anaphase transition is required for correct chromosome segregation and mitotic exit (Thomas et al., 2010). Although these studies indicate that CDC25B is important not only for mitotic entry but also for mitotic progression in somatic cells, a role, if any, for CDC25B following resumption of meiosis is not known.

Here, we use Cdc25b^−/− mouse oocytes that cannot resume meiosis, thereby leading to infertility (Lincoln et al., 2002), to ascertain whether CDC25B plays a role following resumption of meiosis. We report that experimentally activating CDK1 in Cdc25b^−/− oocytes results in resumption of meiosis but the oocytes tend to arrest in MI. In addition, although inactivation of CDK1 during the presumed time of anaphase I (Ana I) occurs in these oocytes, the anaphase-promoting complex/cyclosome (APC/C) is only partially activated; APC/C activation is required for a successful MI-MII transition. Thus, CDC25B is essential not only for resumption of meiosis but also for the MI-MII transition.

CDC25B is essential for resumption of meiosis in females (Lincoln et al., 2002). To ascertain whether CDC25B plays a role following resumption of meiosis, we assessed the ability of Cdc25b^+/− oocytes to mature in vitro. The rationale was that a reduced amount of CDC25B in these oocytes would support resumption of meiosis but may be insufficient for subsequent events. We first established by immunoblotting that the amount of CDC25B was reduced in Cdc25b^+/− oocytes (HET) compared to wild-type (WT) or Cdc25b^−/− (KO) oocytes. As anticipated, the amount of CDC25B was reduced by ∼40% in HET oocytes relative to WT (Fig. 1A,B); two bands corresponding to CDC25B were absent in KO oocytes. The two CDC25B bands probably correspond to two Cdc25b mRNA transcript variants (accession numbers NM_001111075.4 and NM_023117.4). Analysis of previously published RNA-seq data of oocytes from C57BL/6 mice (Abe et al., 2015) confirmed the presence of these two transcript variants that differ by the presence or the absence of exon 7 (data not shown). We then assessed the ability of HET oocytes to mature relative to WT and KO oocytes. In contrast to WT oocytes that successfully matured to MII, as evidenced by emission of a polar body, and as expected KO oocytes that did not, approximately a third of the HET oocytes did not emit a polar body and were presumably arrested at MI (Fig. 1C,D). These results suggest a role for CDC25B in the MI-MII transition, in addition to its documented role in initiation of maturation (Lincoln et al., 2002).

To further explore a potential function of CDC25B after resumption of meiosis, we activated CDK1 to initiate resumption of meiosis in KO oocytes by three different experimental approaches. We first pharmacologically inhibited the WEE1 and MYT1 kinases with PD0166285 (PD) (Potapova et al., 2009, 2011; Wang et al., 2001). Because WEE1 and MYT1 are responsible for the inhibitory phosphorylations of CDK1 during prophase I arrest (Han et al., 2005; Oh et al., 2010) – CDC25B presumably removes these phosphorylations during resumption of meiosis (Lincoln et al., 2002) – PD treatment of KO oocytes resulted in resumption of meiosis as expected but 61% of these treated oocytes arrested in MI (Fig. 2A). Both WT oocytes and PD-treated WT oocytes normally matured to MII (Fig. 2A). To confirm that MI arrest in KO oocytes was due to the absence of CDC25B and not a non-specific effect of PD treatment, we monitored meiotic maturation of WT oocytes simultaneously treated with 2.5 µM milrinone (an inhibitor of meiotic maturation) and 0.5 μM PD. The results indicated that PD treatment could rescue milrinone-induced prophase I arrest of all WT oocytes and that ∼90% of these oocytes progressed to MII, suggesting a specific role for CDC25B in the MI-MII transition (Fig. S1).

In the second approach, we expressed CDC25A by microinjecting an Egfp-Cdc25a complementary (c)RNA. CDC25A cooperates with CDC25B during resumption of meiosis to activate CDK1, but after resumption of meiosis, CDC25A degradation is required for the MI-MII transition (Solc et al., 2008). Injecting ∼10 pl of 5 ng/µl of Egfp-Cdc25a cRNA did not compromise maturation to metaphase II in WT oocytes, but injecting ∼10 pl of 10 ng/µl induced MI arrest (Fig. S2A), a finding consistent with the requirement of CDC25A destruction for the MI-MII transition (Solc et al., 2008). Accordingly, we used the lower concentration of Egfp-Cdc25a cRNA for rescue experiments with KO oocytes. Approximately 75% of KO oocytes microinjected with Egfp-Cdc25a cRNA resumed meiosis but 53% arrested in MI, and only 22% reached MII (Fig. 2A). These results support a role for CDC25B in the MI-MII transition. Note that, as expected, microinjection of Egfp-Cdc25b cRNA completely rescued meiotic maturation in KO oocytes (Fig. S2B-D).

In a third approach, we used Trim-Away to promote EGFP-CDC25B degradation after germinal vesicle breakdown (GVBD) in KO oocytes (details in the Materials and Methods section). Light-sheet microscopy confirmed expression of mCherry-TRIM21 and efficient depletion of EGFP-CDC25B, as fluorescence of EGFP was reduced to the background level (Fig. S2E, Movie 1). In contrast to the PD-treated KO oocytes (Fig. 2A), we found that although EGFP-CDC25B was degraded, only ∼25% of these KO oocytes were arrested in MI (Fig. S2F). We suspect that the narrow time window required to deplete CDC25B after GVBD to prevent its function subsequent to GVBD is not sufficient, i.e. that although CDC25B is degraded post-GVBD, there nevertheless remains sufficient amounts of CDC25B to execute its post-GVBD function prior to its complete degradation.

We next analyzed the kinetics of meiotic maturation and spindle formation in KO oocytes after PD treatment or Egfp-Cdc25a cRNA microinjection by confocal time-lapse imaging of oocytes (Fig. S3A). The oocytes were microinjected with an H2b-mCherry cRNA to visualize chromosomes and stained with a fluorogenic probe SiR-tubulin to visualize microtubules. Whereas inhibition of WEE1 and MYT1 did not have a significant effect on the timing of GVBD, the process of GVBD initiated by Egfp-Cdc25a cRNA microinjection was significantly slower in comparison to WT oocytes (Fig. 2B). KO oocytes treated with PD or microinjected with Egfp-Cdc25a cRNA formed a normal bipolar spindle (Fig. S3A), suggesting that MI arrest was not due to a defective spindle. Dispensability of CDC25B in oocytes for spindle formation is in contrast to human somatic cells in which centrosome-localized CDC25B regulates spindle formation (Gabrielli et al., 1996; Löffler et al., 2006; Mirey et al., 2005).

Although the majority of KO oocytes treated with PD or microinjected with Egfp-Cdc25a cRNA arrested in MI, ∼39% and ∼22%, respectively, still entered Ana I and extruded the first polar body (Fig. 2A). Interestingly, although KO oocytes microinjected with Egfp-Cdc25a cRNA that extruded the first polar body entered Ana I with similar kinetics as WT oocytes, KO oocytes treated with PD were significantly delayed (Fig. 2C). These results suggest that coordination between WEE1, MYT and CDC25B is important for the normal timing of Ana I onset, and when both players are missing, Ana I is significantly delayed. Note that WT oocytes and PD-treated WT oocytes had a similar timing of Ana I entry (Fig. 2C), which contrasts to budding yeast in which SWE1 (budding yeast WEE1 homolog) depletion accelerates anaphase entry because SWE1 is a component of the morphogenesis checkpoint (Lianga et al., 2013).

To assess chromosome behavior and microtubule organizing centers (MTOCs) clustering in KO oocytes treated with PD at high resolution, we microinjected oocytes stained with SiR-tubulin with cRNAs encoding an MTOC marker, mEGFP-mCDK5RAP2, and H2B-mCherry, to visualize chromosomes, as described previously (Blengini et al., 2021), and followed meiotic maturation and spindle formation by time-lapse light-sheet microscopy (Fig. 2D; Movie 2). These experiments indicated no significant difference in the percentage of oocytes with segregation errors between WT and KO oocytes (Fig. S3B). At MI, WT oocytes contained a bipolar spindle with fully clustered MTOCs and centrally aligned chromosomes on the metaphase plate (Fig. 2D). At the same time, in the majority of KO oocytes, chromosomes were also aligned at the spindle equator (Fig. 2D,E). In KO oocytes, fragmented MTOCs formed two dominant MTOC clusters similar to WT oocytes (Fig. 2D,F). We also observed that the onset of spindle formation and spindle elongation (bipolarization) was not affected in KO oocytes (Fig. S3C,D). These results provide further evidence that KO oocytes arrest in MI with normal bipolar spindles.

To further explore the role of CDC25B in the MI-MII transition, we microinjected KO oocytes with a cRNA encoding constitutively active CDK1 (Tischer and Schuh, 2016) in which T14 and Y15 residues are replaced by nonphosphorylatable A and F, respectively (CDK1^af) (Adhikari et al., 2016; Pomerening et al., 2008). KO oocytes microinjected with Cdk1^af cRNA (70 ng/µl) resumed meiosis (76%) but 45% remained arrested in MI (Fig. S4A). In contrast, ∼90% of WT oocytes microinjected with the same amount of Cdk1^af cRNA matured to MII. Microinjection of half the concentration of Cdk1^af cRNA into KO oocytes was less effective at inducing maturation and led to significant MI arrest (Fig. S4B); microinjected WT oocytes were unaffected.

Time-lapse imaging of oocytes expressing H2B-mCherry demonstrated that although KO oocytes co-microinjected with Cdk1^af cRNA (70 ng/µl) were significantly delayed in resumption of meiosis (Fig. S4C), those that extruded the first polar body entered Ana I earlier than WT oocytes microinjected with the same amount of Cdk1^af cRNA (Fig. S4D). Because there was a significant delay in resumption of meiosis in KO oocytes microinjected with Cdk1^af cRNA, we determined whether oocytes that resumed meiosis arrested in MI. Results of these experiments did not show any significant difference in the timing of resumption of meiosis in KO oocytes at the MI versus MII stage. We conclude that acceleration of meiotic resumption does not affect MII progression in KO oocytes (Fig. S4E).

In summary, forced activation of CDK1 in KO oocytes by Egfp-Cdc25a cRNA microinjection, WEE1 and MYT1 pharmacological inhibition or Cdk1^af cRNA microinjection bypassed the requirement of CDC25B for resumption of meiosis and thereby unmasked a requirement of CDC25B for the MI-MII transition.

A requirement of CDC25B for the MI-MII transition led us to investigate whether MI arrest in KO oocytes is due to the inability to activate APC/C or inactivate CDK1. Accordingly, we used KO oocytes treated with PD because this approach was the most effective at inducing resumption of meiosis in the absence of CDC25B. To monitor APC/C activation, we microinjected oocytes with a cRNA encoding securin-EGFP, which is degraded by APC/C shortly before Ana I entry (McGuinness et al., 2009). Chromosomes were visualized by H2b-mCherry cRNA microinjection and microtubules by SiR-tubulin staining. In WT oocytes and PD-treated WT oocytes, securin-EGFP was normally degraded before Ana I onset (Fig. 3A,B; Movie 3). In KO oocytes treated with PD, securin-EGFP was also degraded but its rate of degradation was reduced in oocytes failing to extrude the first polar body (Fig. 3C). To confirm reduced APC/C activity in MI-arrested KO oocytes, we monitored the destruction of a non-CDK1-binding cyclin B1, another APC/C substrate, which is already partially destroyed in prometaphase I (Levasseur et al., 2019). Oocytes stained with SiR-tubulin were microinjected with H2B-mCherry and a non-CDK1-binding cyclin B1 reporter cyclin B-Y170-Vfp cRNA, and degradation of the cyclin B1 reporter was monitored over time by live imaging (Fig. S5A-C, Movie 4). These experiments demonstrated that the rate of cyclin B degradation exhibits similar kinetics of destruction as securin in MI-arrested KO oocytes, thereby providing further evidence of reduced APC/C activity in KO oocytes.

APC/C activation (Fig. 3A,B), normal bipolar spindle formation and chromosome alignment (Fig. 2D) in PD-treated KO oocytes suggest that MI arrest is not due to persistent spindle assembly checkpoint (SAC) activity, although MI-arrested oocytes did not reach maximal APC/C activity (Fig. 3C). SAC monitors the attachment of microtubules to kinetochores, and its activation caused by loss of kinetochore-microtubule (K-MT) attachments prevents APC/C activation and delays anaphase entry. In addition, inappropriate K-MT attachment could result in chromosome misalignment (Musacchio and Salmon, 2007). PLK1 is essential for full APC/C activation, which is independent of satisfying the SAC in mouse oocytes (Solc et al., 2015). When PLK1 is acutely inhibited at the end of MI (Solc et al., 2015), oocytes remain arrested in MI, and degradation of securin exhibits a similar pattern as CDC25B KO oocytes treated with PD. Although CDC25B is widely phosphorylated by PLK1 (Lobjois et al., 2011), the role of PLK1-mediated CDC25B phosphorylation in Ana I onset remains unknown and requires further investigation.

Ana I entry requires not only APC/C-mediated proteolysis of cyclin B1 and securin (Herbert et al., 2003) but also a WEE1- and MYT1-dependent decrease in CDK1 activity (Zhou et al., 2019). To measure CDK1 activity, we used an in vitro biochemical assay employing the fragment of lamin A containing a CDK1 phosphorylation site (Fig. 3D). As expected, in WT oocytes, CDK1 activity transiently decreased during the MI-MII transition. We have seen a similar transient decrease in PD-treated WT oocytes, although CDK1 activity was lower in metaphase I (8 h) compared to WT oocytes. Importantly, KO oocytes treated with PD had a similar transient decrease in CDK1 activity during the MI-MII transition, which is consistent with the activation of APC/C.

Taken together, these data show that MI arrest in KO oocytes treated with PD is not caused by an inability to activate APC/C or an inability to transiently decrease CDK1 activity. In addition, because CDC25B is a CDK1 activator, a decrease in CDK1 activity would be expected when CDC25B is absent or reduced, but this is not the case. Because CDK1 activity in MI oocytes is similar between PD-treated WT oocytes and PD-treated KO oocytes, oocytes may compensate for a decrease in CDC25B function by increasing the activity of CDC25A/C (Kanatsu-Shinohara et al., 2000; Solc et al., 2008).

To further investigate a relationship between partial APC/C activation and MI arrest in KO oocytes treated with PD, and to avoid the situation when resumption of meiosis must be experimentally induced in KO oocytes by PD treatment or Egfp-Cdc25a cRNA microinjection, we analyzed APC/C activation in HET oocytes that have a reduced amount of CDC25B protein (Fig. 1B). Although HET oocytes normally resume meiosis and do not require forced activation of CDK1, approximately a third remained arrested in MI (Fig. 1A), despite having a similar CDK1 activity in MI as control WT oocytes (Fig. 4A). Live imaging of oocytes microinjected with Securin-Egfp and H2b-mCherry cRNAs, and stained with SiR-tubulin (Fig. 4B; Movie 5) showed that, as expected, HET oocytes had a significant delay in resumption of meiosis compared to WT oocytes (Fig. 4C). The Ana I entry was not significantly delayed (Fig 4D). WT and HET oocytes that matured to MII had similar kinetics of destruction of securin-EGFP, demonstrating normal APC/C activity (Fig. 4B,E,F). HET oocytes that arrested in MI initiated destruction of securin-EGFP but the degradation rate was significantly reduced, suggesting partial APC/C activation. These data confirm that the normal expression of CDC25B is important for full activation of APC/C and Ana I entry.

To further confirm the role of SAC in MI arrest, we examined the localization of MAD2 (also known as MAD2L1), an essential component of SAC signaling in mouse oocytes, by immunofluorescent staining (Fig. 5A). First, we detected the MAD2 level on kinetochores at prometaphase I (4 h after meiotic resumption) in WT oocytes. We showed that MAD2 was enriched at the kinetochores in WT oocytes in this time interval. The relative protein expression levels (fluorescence intensity) of MAD2 were markedly reduced at kinetochores in WT oocytes treated with PD at MI (8 h after meiotic resumption). Although there was oocyte-to-oocyte variability, no significant difference in MAD2 intensity was measured at MI between PD-treated WT and PD-treated KO oocytes, suggesting that SAC activity in KO oocytes is not elevated (Fig. 5B).

We then ascertained whether KO oocytes could progress to MII when SAC signaling is removed by treating oocytes simultaneously with PD and reversine to inhibit the SAC kinase MPS1 (also known as TTK). Meiotic progression and securin degradation were monitored by light-sheet microscopy (Fig. 5C; Movie 6). As expected, reversine treatment accelerated progression to Ana I, with securin degradation beginning at ∼4 h after meiotic resumption in WT oocytes (Fig. 5D). In the presence of reversine, KO oocytes treated with PD exhibited rates of securin degradation similar to WT oocytes (Fig. 5E). All KO oocytes entered Ana I in the presence of reversine. Some of these oocytes either failed to extrude a polar body and complete MII, or had a cytokinesis failure and retracted the polar body into the cytoplasm (Fig. 5C,F). Taken together, the comparable reduction of MAD2 at kinetochores in KO oocytes and failure to complete MII when the SAC is inhibited suggests that CDC25B has an additional role in promoting MII progression beyond SAC satisfaction.

The results of experiments reported here demonstrate that CDC25B regulates not only resumption of meiosis but also the MI-MII transition in mouse. This requirement is similar to that observed in somatic cells in which CDC25B is involved not only in mitotic entry but also mitotic progression (Gabrielli et al., 1996; Lindqvist et al., 2005), although its requirement is not absolute because KO mice are viable (Lincoln et al., 2002). In intestinal epithelial stem cells and progenitor cells, the absence of CDC25B can be compensated by CDC25A but not CDC25C (Lee et al., 2011). In oocytes, the CDC25B requirement is associated with the initial CDK1 activation and resumption of meiosis (Lincoln et al., 2002), and the results described here strongly implicate a function subsequent to resumption of meiosis, namely the MI-MII transition.

CEP55, a centrosome- and midbody-associated protein involved in cytokinesis (Fabbro et al., 2005), is required for reduced expression of CDC25B after the meiotic resumption in MI (Zhou et al., 2019). siRNA-mediated CEP55 knockdown induces MI arrest characterized by an increased amount of CDC25B and increased CDK1 activity (due to decreased CDK1 phosphorylation) but normal APC/C activity. Partial pharmacological inhibition of CDC25s or inhibition of CDK1 in MI in CEP55-depleted oocytes rescues Ana I entry. Thus, Ana I entry requires not only APC/C-mediated proteolysis but also phosphorylation-dependent CDK1 inhibition that is linked to regulating the amount of CDC25B (Zhou et al., 2019). Here, we show an opposite situation, namely, that when CDC25B activity is too low or even absent in MI oocytes, this decreased activity is also detrimental for Ana I onset. In conclusion, our work, together with another recent study (Zhou et al., 2019), shows that CDC25B activity must be precisely regulated after meiotic resumption to allow MI-MII transition.

Animals were treated and used according to the guidelines of the Expert Committee for the Approval of Projects of Experiments on Animals of the Academy of Sciences of the Czech Republic. To obtain Cdc25b ^−/− mice on the inbred C57BL/6 genetic background, we used the Cdc25b^−/− mice described previously (Lincoln et al., 2002) and bred heterozygous males with WT C57BL/6 females (Charles River) for six consecutive generations. Heterozygous offspring were then bred among themselves to produce the first generation of Cdc25b^−/− mice on the inbred C57BL/6 genetic background. Further experimental Cdc25b^−/− (KO), Cdc25b^+/− (HET) and Cdc25b^+/+ (WT) females were generated from breeding between HET parents from the previous breeding.

In all experiments, germinal vesicle (GV)-stage oocytes were collected at room temperature from 12-20-week-old females 44 h after injection with PMSG/eCG (5IU, ProSpec-Tany TechnoGene). Oocytes were collected and microinjected in M2 medium (Sigma-Aldrich) containing 2.5 µM milrinone (Sigma-Aldrich) to prevent meiotic maturation. Oocytes were cultured in minimum essential medium (MEM; Sigma-Aldrich) supplemented with 1.14 mM sodium pyruvate (Sigma-Aldrich), 4 mg/ml bovine serum albumin (BSA; Sigma-Aldrich) and penicillin-streptomycin (75 U/ml; 60 µg/ml, Sigma-Aldrich). Culture and live imaging were performed at 37°C in a 5% CO₂ atmosphere. A final concentration of 1 μM reversine (Sigma-Aldrich) was added to the oocytes for securin degradation analysis.

We used previously described plasmids for in vitro transcription of histone H2b-mCherry (Kitajima et al., 2011), Egfp-Cdc25a (Solc et al., 2008), Egfp-Cdc25b (Solc et al., 2008), Egfp-Cdk1^af (Tischer and Schuh, 2016), securin-Egfp (Kudo et al., 2006), mEgfp-mCdk5rap2 (Balboula et al., 2016), Y170A cyclin B1-Vfp (Levasseur et al., 2019) and mCherry-mTrim21 (Clift et al., 2017). pGEMHE-mCherry-mTrim21 was a gift from Melina Schuh, Max Planck Institute for Multidisciplinary Sciences, Gottingen, Germany (Addgene plasmid 105522). Plasmids were linearized and in vitro transcribed using mMESSAGE mMACHINE T3 and T7 kits (Ambion) to prepare cRNAs. cRNAs were not polyadenylated. Before storage at −80°C, all cRNAs were purified using an RNAeasy kit (Qiagen). GV oocytes were microinjected with ∼10 pl of a cRNA solution (50 ng/µl H2b-mCherry, 50 ng/µl securin-Egfp, 5 or 10 ng/µl Egfp-Cdc25a, 75 ng/µl Egfp-Cdc25b, 35 or 70 ng/µl Egfp-Cdk1^af, 125 ng/µl mEgfp-mCdk5rap2, 75 ng/µl Y170A cyclin B1-Vfp and 200 ng/µl mCherry-mTrim21 cRNAs). Oocytes were cultured for 2 h in MEM medium supplemented with milrinone to allow protein expression after cRNA microinjection.

For CDC25B immunoblotting, 80 GV oocytes were used. Immunoblotting was performed as described previously (Mayer et al., 2016). CDC25B was detected with the CDC25B C20 antibody (Santa Cruz Biotechnology, sc-326, 1:500). A similar amount of protein loading was verified by total protein staining (Pierce Reversible Protein Stain Kit for Nitrocellulose Membranes, Thermo Fisher Scientific, 24580) of the membrane after immunoblotting.

Following meiotic maturation, oocytes were fixed in PHEM (60 mM PIPES, 25 mM HEPES, 10 mM EGTA and 2 mM MgCl₂) containing 2% paraformaldehyde for 20 min followed by three consecutive washes with blocking buffer [PBS plus 0.3% (w/v), BSA plus 0.1% (v/v) and Tween 20]. Prior to immunostaining, oocytes were permeabilized for 20 min in PBS containing 0.1% (v/v) Triton X-100 and 0.3% (w/v) BSA, followed by 10 min in blocking buffer. Immunostaining was performed by incubating cells in primary antibody for 2 h in a dark humidified chamber at room temperature, followed by three consecutive 10-min incubations in blocking buffer. After washing, secondary antibodies were diluted 1:100 in blocking solution, and the sample was incubated for 1 h at room temperature. After washing, the cells were mounted in Mowiol with 4′,6-diamidino-2-phenylindole, dihydrochloride (Invitrogen). The primary antibodies ACA (1:30; Antibodies Incorporated, 15-234) and MAD2 (1:100; BioLegend, PRB452C) were used for immunofluorescence experiments. The secondary antibodies donkey anti-rabbit IgG Alexa Fluor 488 (Thermo Fisher, A-21206) and Alexa Fluor 647 AffiniPure donkey anti-human IgG (Jackson ImmunoResearch, 709-605-149) were used at a dilution of 1:100 for immunofluorescence experiments.

To measure CDK1 kinase activity in mouse oocytes, we adapted a previously described assay used in human somatic cells (Müllers et al., 2017). Oocytes (5 per sample) were collected at the indicated time points after milrinone wash out. Before collection, oocytes were stained with 50 ng/ml Hoechst 34580 (Sigma-Aldrich) for 1 h, and chromosome segregation was evaluated by fluorescence microscopy to distinguish MI and Ana I oocytes. Samples were washed in PBS and stored at −80°C in a minimum volume of PBS. The samples were incubated in kinase buffer [10 mM HEPES (7.4 pH), 2.5 mM β-glycerol phosphate, 1 mM EGTA, 1 mM EDTA and 4 mM MgCl₂] supplemented with 0.2% Triton X-100, 100 µM ATP and purified GST-LAMS22 substrate (2 µg) for 30 min at 30°C. The reaction was stopped by adding 1× Reducing SDS Loading Buffer (Cell Signaling Technology) and boiling for 4 min. Proteins were separated by SDS-PAGE (gradient 4-12%), and phosphorylation of GST-LAMS22 was detected by immunoblotting using phospho Ser22 lamin A/C Ser22 antibody (Cell Signaling Technology, D2B2E) in a 1:1000 dilution. Band intensity was quantified using ImageLab Software.

EGFP-CDC25B depletion by Trim-Away was performed as described previously (Clift et al., 2017). The anti-GFP antibody used was a rabbit polyclonal anti-GFP (Abcam, ab6556), which was concentrated using Amicon Ultra-0.5 100-kDa centrifugal filter devices (Millipore, UFC510024). Prior to microinjection, the antibody was diluted in PBS containing 0.05% NP40 to 0.29 mg/ml. GV oocytes stained with 75 nM SiR-DNA (Spirochrome, SC007) were co-injected with cRNA encoding for Egfp-Cdc25b and mCherry-Trim21. Once the expressed EGFP-CDC25B induced GVBD in KO oocytes, half of the oocytes were microinjected with ∼10 pl of the anti-GFP antibody to induce EGFP-CDC25B depletion. Immediately after microinjection, oocytes were scanned using light-sheet microscopy.

Time-lapse acquisitions in Fig. 3A, Fig. 4B and Fig. S3A were performed using a Leica TCS SP5 microscope with an HCX PL Apo Lambda Blue 40×1.25 oil objective. In Fig. S3A, 1× zoom was used, and 14 confocal 5-µm sections were taken with a 1024×1024 pixel image resolution using 10 min time intervals. In Fig. 3A and Fig. 4B, tracking function of the matrix screen module in LAS AF software (Leica Microsystems) was used, and 16 confocal 5-µm optical sections with a 4.8× zoom on the area of individual oocytes in 512×512 pixel image resolution using a 1000-Hz speed bi-directional scan were acquired. For spindle visualization, oocytes were stained with 100 nM SiR-tubulin as described previously (Balboula et al., 2016). Time-lapse image acquisitions in Fig. 2D, Fig. 5C, Fig. S2E and Fig. S5A were performed using a Viventis LS1 Live light-sheet microscope system (Viventis Miscoscopy Sarl, Switzerland) with a Nikon 25X NA 1.1 detection objective with 1.5× zoom. Thirty-one 2-μm optical sections were taken with a 750×750-pixel image resolution using 10-min time intervals. EGFP, VFP, mCHERRY and SiR fluorescences were excited by 488, 515, 561 and 638 nm laser lines, respectively. EGFP and VFP emissions were detected using 525/50 (bandpass, BP) and 539/30 (BP) filters, respectively. For the detection of mCHERRY and SiR fluorescences, a 488/561/640 (triple bandpass, TBP) filter was used.

Fiji software (Schindelin et al., 2012) was used for image analysis. Securin-EGFP and cyclin B1-VFP reporter mean intensity was measured on non-adjusted middle optical stack in every time frame, and the measured values were normalized to the time frame with a maximum mean intensity value in every oocyte. The normalized intensities were smoothed using the loess smoother function with a span parameter of 0.25. The securin-EGFP degradation rate was determined as described previously (Solc et al., 2015).

The data for most of the results were obtained from at least three independent experiments and the data pooled. NCSS 11 software (NCSS, Kaysville UT, USA) was used for statistical analysis. Comparisons of multiple groups were calculated using Kruskal–Wallis one-way ANOVA. P-values for multiple comparisons with control (WT oocytes) were calculated using Tukey–Kramer's test. When only two groups were analyzed, the Mann–Whitney test was used. Samples with P<0.05 were considered as statistically significant. In figures, *P<0.05, **P<0.01 and ***P<0.001.

We dedicate this article to the memory of Assoc. Prof. Dr Petr Solc, who passed away during the revision. We thank Melina Schuh, Max Planck Institute for Multidisciplinary Sciences, Gottingen, Germany, for providing the Egfp-Cdk1^af plasmid and Suzanne Madgwick, Newcastle University, UK, for providing the Y170A cyclin B1-Vfp plasmid.

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.252924.