APC/CCdh1 Enables Removal of Shugoshin-2 from the Arms of Bivalent Chromosomes by Moderating Cyclin-Dependent Kinase Activity

Abstract

In mammalian females, germ cells remain arrested as primordial follicles. Resumption of meiosis is heralded by germinal vesicle breakdown, condensation of chromosomes, and their eventual alignment on metaphase plates. At the first meiotic division, anaphase-promoting complex/cyclosome associated with Cdc20 (APC/C^Cdc20) activates separase and thereby destroys cohesion along chromosome arms. Because cohesion around centromeres is protected by shugoshin-2, sister chromatids remain attached through centromeric/pericentromeric cohesin. We show here that, by promoting proteolysis of cyclins and Cdc25B at the germinal vesicle (GV) stage, APC/C associated with the Cdh1 protein (APC/C^Cdh1) delays the increase in Cdk1 activity, leading to germinal vesicle breakdown (GVBD). More surprisingly, by moderating the rate at which Cdk1 is activated following GVBD, APC/C^Cdh1 creates conditions necessary for the removal of shugoshin-2 from chromosome arms by the Aurora B/C kinase, an event crucial for the efficient resolution of chiasmata.

Keywords: APC/C(Cdh1); aneuploidy; aurora kinase; cohesin; meiosis; shugoshin-2.

Graphical abstract

Figure 1

By Maintaining Low Levels of Cdc25B and Cyclin B1, the APC/C^Cdh1 Maintains GV Arrest, Prevents Premature Entry into Meiosis, and Ensures a Gradual Increase in Cdk1 Activity after GVBD (A) Oocytes harvested from Cdh1^f/f and Cdh1^f/fGdf9-iCre ovaries in the presence of IBMX that had already resumed meiosis upon isolation were quantified. GVBD rates for each of the indicated genotypes are plotted as a percentage of the total oocytes observed. Mean and SDs are displayed, and the number of females used is indicated (n). (B) GV-stage oocytes harvested in the presence of IBMX were released into the M16 medium. The kinetics of GVBD was captured using a time-lapsed confocal microscope. Microinjections were performed at GV stage in the presence of IBMX. The number of oocytes analyzed is indicated (n). (C and D) GV-stage oocytes harvested from ovaries isolated from 4-week-old females were immunoblotted for cyclin B1 (C) and Cdc25B (D). Actin was used as loading control. One hundred twenty GV-stage oocytes were pooled for each lane. (E) Fully grown GV-arrested oocytes were immunoblotted for phospho-Cdk1 (Tyr15) and Cdk1. For each lane, 120 GV-stage oocytes were pooled. (F) GV-stage oocytes harvested from the indicated genotypes were imaged for 18 hr. The kinetics of GVBD is displayed. The number of oocytes imaged is indicated (n). (G and H) Cdk1 kinase activity was estimated using in vitro phosphorylation of histone H1. Oocytes from Cdh1^f/f and Cdh1^f/fGdf9-iCre (G) and Cdc20^f/f, Cdc20^f/fZp3Cre, and Cdc20^f/fZp3Cre microinjected with Δ90-cyclin B1 (H) ovaries at GV, GVBD, and 20 and 40 min post-GVBD were incubated with radiolabeled ATP and histone H1. Samples were resolved on SDS-PAGE gels, and incorporated radioactivity was imaged. (I) GV-stage oocytes from each of the indicated genotypes depicted were immunoblotted for Cdh1 or Apc2 to confirm depletion. Actin was used as a loading control. Extracts from Apc2^f/f and Apc2^f/fZp3Cre were also immunoblotted for cyclin B1 and Cdc2-Y15P. For each lane, cell lysates from 120 GV-stage oocytes were loaded. Except for the Cdh1 Gdf9-iCre knockout and control ovaries, which were isolated at 4 weeks post-birth due to rapid depletion of the primordial follicles, ovaries from all other crosses were isolated at 6 weeks post-birth. For Apc2 knockout results, see Figure S1.

Figure 2

Following GVBD, Sgol2 Localizes on the Chromosome Arms and Kinetochores, and It Gradually Concentrates on Kinetochores during Late Metaphase (A) GV-stage oocytes harvested from Sgol2-deleted females were microinjected in M2 medium supplemented with IBMX with mRNA encoding GFP-Sgol2 and H2B-mCherry. After 1 hr of incubation, oocytes were released, and time-lapse confocal microscope images were captured for 12–14 hr following GVBD. Representative Z-projected images are displayed. (B) GV-stage oocytes harvested from wild-type control females were cultured for 2, 4, and 6 hr following GVBD. Chromosome spreads prepared at the indicated times following GVBD were stained for DNA (blue), Sgol2 (red), and CREST (green). See also Figure S2.

Figure 3

Entry into Meiosis with High Cdk1 Activity Prevents Removal on Sgol2 from Chromosome Arms (A) GV-arrested oocytes harvested from Apc2^f/fZp3Cre and Cdc20^f/fZp3Cre mice (first and second rows) were microinjected with GFP-Sgol2 (green) and H2B-mCherry (red) or co-injected with GFP-Sgol2 (green), H2B-mCherry (red), and Δ90-cyclin B1 mRNA (third row). After 1 hr in IBMX-containing media, microinjected oocytes were released in IBMX-free M16 medium, and a time-lapse confocal microscopy movie was started. Representative Z-projected time-lapse confocal microscopy images are displayed. Chromosomes and Sgol2 were visualized using H2B-mCherry (red) and GFP-Sgol2 (green), respectively. Times displayed are relative to the time of GVBD. The scale bar represents 10 μm. (B) GFP-Sgol2 intensity signal on chromosome arms at 14 hr was normalized by the GFP-Sgol2 signal intensity on chromosome arms at the prometaphase stage. Normalized intensity values of GFP-Sgol2 signal from Cdc20^f/fZp3Cre was compared to normalized GFP-Sgol2 intensities from Apc2^f/fZp3Cre (p < 0.0001) and Cdc20^f/fZp3Cre microinjected with Δ90-cyclin B1 mRNA (p < 0.0001). Mean and SDs are displayed. The number of oocytes studied is indicated (n). See also Figures S2 and S3. (C) GV-arrested oocytes from Apc2^f/fZp3Cre and Cdc20^f/fZp3Cre ovaries were released into IBMX-free medium and cultured for 14 hr following GVBD. Chromosome spreads were stained with DAPI (blue), anti-Sgol2 (red), and CREST (green). (D) Fluorescence intensity ratios of Sgol2 on arms and CREST at kinetochores were compared. Compared to Cdc20^f/fZp3Cre oocytes, Apc2^f/fZp3Cre retained more Sgol2 on chromosome arms at 14 hr post-GVBD (p < 0.0001). Upper and lower bars indicate 95^th and 5^th percentiles, respectively. The number of oocytes examined is indicated (n).

Figure 4

Removal of Sgol2 from Chromosome Arms Requires Aurora-B/C-Mediated Phosphorylation of Sgol2 (A) GV-arrested oocytes harvested from sgol2^Δ/Δ were microinjected with wild-type Sgol2 or T521AT600A Sgol2 mRNA in IBMX-containing media. After 1 hr in IBMX-containing media, oocytes were cultured in M16 medium for 12 hr following GVBD. Chromosome spreads were performed on oocytes that had extruded a polar body and were stained for DAPI (blue), CREST (green), and anti-Sgol2 (red). See also Figure S4. (B) GV-arrested oocytes harvested from Cdc20^f/fZp3Cre mice were microinjected with GFP-Sgol2 (green) and H2B-mCherry (red) mRNA (first row) or with GFP-T521AT600A-Sgol2 (green) and H2B-mCherry (red) mRNA (second row). To test whether the Aurora B/C kinase is required for the removal of Sgol2 from chromosome arms, oocytes were cultured in M16 medium supplemented with AZD1152 (100 nM; third row). Representative Z-projected, time-lapse confocal microscope images are displayed. Times displayed are relative to the GVBD. The scale bar represents 10 μm. (C) GFP-Sgol2 intensity on chromosome arms at 14 hr was normalized by GFP-Sgol2 intensity on chromosome arms at the prometaphase stage. Cdc20 knockout oocytes retained higher levels of GFP-T521A T600A Sgol2 (p < 0.0001) and wild-type GFP-Sgol2 when cultured with Aurora inhibitor (AZD1152; p < 0.0001). Mean and SDs are displayed, and the number of oocytes evaluated is indicated (n).

Figure 5

Aurora-B/C-Dependent Phosphorylation of Sgol2 Is Reduced in Apc2 Knockout Oocytes; Consequently, MCAK Localization Is Reduced and Inter-kinetochore Distance Is Increased in Apc2 Knockout Oocytes (A) GV-stage oocytes harvested from Apc2^f/f and Apc2^f/fZp3Cre ovaries were cultured for 4 hr following GVBD. Chromosome spreads were stained with DAPI (blue), anti-T521p-Sgol2 (red), and CREST (green). Fluorescence intensity ratios of T521p-Sgol2 on chromosome arms and CREST at kinetochores were compared (p < 0.0001). Mean and SDs are displayed, and the number of oocytes examined is indicated (n). (B) Chromosome spreads were prepared at 4 hr post-GVBD. Slides were stained for DNA (blue), MCAK (red), and CREST (green). Fluorescence intensity ratios of MCAK and CREST at kinetochores were compared (p < 0.0001). Mean and SDs are displayed, and the number of oocytes examined is indicated (n). (C) Oocytes were cultured for 4 hr following GVBD in IBMX-free medium. Chromosome spreads were stained with DAPI (blue), CREST (green), and anti-Aurora C (red). Fluorescence intensity ratios of Aurora C on chromosome arms and CREST at kinetochores from Apc2^f/f and Apc2^f/fZp3Cre oocytes were compared (p = 0.548). Mean and SDs are displayed, and the number of oocytes evaluated is indicated (n). (D) Oocytes harvested at GV stage from Apc2^f/f, Cdc20^f/fZp3Cre, and Apc2^f/fZp3Cre females were microinjected with H2B-mCherry (red) to visualize chromosomes and EB3-GFP (green) to visualize microtubules. Z-projected (12 Z slices acquired 1.5 μm apart), time-lapsed live-cell confocal images are displayed. (E) Ratios of chromosome length normalized by spindle length were calculated for metaphase I stage oocytes. Normalized chromosome lengths from Apc2^f/f were compared to normalized chromosome lengths from Cdc20^f/fZp3Cre (p = 0.2214) and Apc2^f/fZp3Cre (p < 0.0001). Mean and SDs are displayed, and the number of oocytes observed is indicated (n). See also Figure S5.

Figure 6

Retention of Sgol2 on Chromosome Arms during Anaphase I Causes Abnormal Attachments and Increases Aneuploidy in Apc2 and Cdh1 Knockout Oocytes (A) Apc2^f/f and Apc2^f/fZp3Cre oocytes were harvested in IBMX-containing media. To prevent metaphase arrest, Apc2^f/fZp3Cre oocytes were microinjected with Apc2 mRNA either at GV stage or after 3 or 4 hr post-GVBD. Oocytes were cultured for 12–14 hr following GVBD, and chromosome spreads were performed on oocytes that had extruded a polar body. Slides were stained with DAPI (blue), CREST (green), and anti-Sgol2 (red). Stacked bar plot indicates the incidence of aneuploidy in each group. The number of oocytes examined is indicated (n). (B) Oocytes harvested from the Cdh1^f/f and Cdh1^f/fGdf9-iCre ovaries were cultured in medium supplemented with IBMX. A group of Cdh1-deleted oocytes were microinjected with Cdh1 mRNA at the GV stage. Oocytes were matured in the M16 medium for up to 12 hr. Chromosome spreads, prepared from oocytes that had extruded the first polar body, were stained with DAPI (blue), CREST (green), and anti-Sgol2 (red). The frequency of aneuploidy was quantified. The number of oocytes analyzed is indicated by n. (C) Oocytes harvested at the GV stage from Apc2^f/fZp3Cre sgol2^Δ/Δ females were microinjected at 3 or 4 hr post-GVBD with Apc2 mRNA. Chromosome spreads were prepared from metaphase II stage oocytes from sgol2^Δ/Δ, and Apc2^f/fZp3Cre sgol2^Δ/Δ microinjected with Apc2 mRNA were stained with DAPI (blue), CREST (green), and anti-Sgol2 (red). The incidence of aneuploidy was quantified. The number of oocytes studied is indicated (n).

Figure 7

A Model for Meiotic Activation in Mouse Oocytes (A) Model wiring diagram showing interactions between components. Cdk1:CycB activity in early meiosis is determined by APC/C^Cdh1-mediated CycB degradation and inhibitory phosphorylation of Cdk1 by Wee1, counteracted by Cdc25. Cdk1 inhibits Wee1 and Cdh1 and activates Cdc25, creating positive/double-negative feedback loops. Cdh1 also promotes Cdc25 degradation. (B and C) Phase-plane analysis of the meiotic control network. The steady-state activities (nullclines) of Cdh1 as a function of Cdk1 (red) and Cdk1 as a function of Cdh1 (blue) are plotted for wild-type oocytes at GV arrest (B) and after release from IBMX (C). By definition, intersections of these curves correspond to stable (black circle) or unstable (empty circle) steady states of the whole system. The stable steady state in the upper left corner of (B) corresponds to GV arrest, whereas the one in the bottom right corner corresponds to prometaphase I. At meiotic resumption (C), the upper steady state is lost, leaving only the prometaphase state. The resulting transition to the prometaphase state occurs along a trajectory indicated by the dashed arrow. (D) Time course simulation of the transition described in (C). Species names correspond to the active form of the specified component: i.e., Cdk1 is the number of active (unphosphorylated) CycB:Cdk1 complexes and Cdc25 is the level of active, phosphorylated Cdc25. CycBT is the total of both free and Cdk1-bound CycB pools. Detailed analysis of Cdh1, Cdc25B, and Cdh1 Cdc25B double knockout is presented in Figure S6.