Ethics approval

The use of immature unfertilized human oocytes in this study has been approved by the UK’s National Research Ethics Service under the REC Reference 11/EE/0346 (IRAS Project ID 84952). The porcine oocytes used in this study were obtained from a local abattoir as a waste product of the slaughtering process, and as such do not require ethics approval for usage in Germany where the study was conducted.

Primary porcine oocytes

All porcine oocytes were collected from pre-pubertal animals of female sex (exact age unknown). Porcine ovaries were obtained from a local abattoir and transported to the laboratory in a thermo-flask. Cumulus-oocyte complexes (COCs) were aspirated from 2–6-mm-large antral follicles using a 17-gauge needle affixed to a 1 ml disposable syringe. The aspired follicular fluid was collected into a 50 ml Falcon tube containing 2 ml of M2 medium with 1 mM dbcAMP. COCs were allowed to sediment for 15 min, and then washed extensively through four 35 mm Petri dishes, each containing 2 ml of pre-warmed M2 medium with 1 mM dbcAMP. Only fully grown oocytes with a homogeneous cytoplasm and at least three to five complete layers of compact cumulus cells were selected for the experiments. COCs were denuded using a series of transfer pipettes with a tip of defined diameter (175 µm, 145 µm and 135 µm) attached to an EZ-Grip denudation pipettor. Oocytes were cultured in M16 (5% CO2) or M2 medium at 38.5 °C. In most experiments, prophase arrest was maintained by supplementing the culture medium with dbcAMP. For experiments involving timed fixations and a few live cell imaging experiments specified below, cells were cultured for approximately 12 h in medium supplemented with 50 µM Butyrolactone I (Abcam, ab141520; 50 mM stock in ethanol). To induce the resumption of meiosis, oocytes were washed into dbcAMP/Butyrolactone I-free M2 medium.

Primary human oocytes

All human oocytes obtained for this study were sourced from patients of female sex undergoing fertility treatments at Bourn Hall Clinic (Bourn, Cambridgeshire, UK) between 12 January 2017 and 24 July 2019 after having obtained fully informed consent. All donations were anonymous. Patients were not monetarily compensated for their donation. Ages of human oocyte donors from Fig. 1 and Extended Data Fig. 1a are listed in Supplementary Table 1. Ages of human oocyte donors for data shown in Figs. 4b and 8b and Extended Data Figs. 3b and 4b are listed in Supplementary Table 2. Fifty-six germinal vesicle (GV) stage oocytes from 39 donors were obtained for immunofluorescence analysis. All donors underwent ovarian stimulation for intracytoplasmic sperm injection. Only oocytes that were immature and hence unsuitable for intracytoplasmic sperm injection were donated to the study. None of the oocytes used in this study were freeze-thawed. Oocytes were cultured in G-MOPS medium (Vitrolife, 10129) supplemented with 10% FBS (GIBCO, 16000044) under mineral oil (Merck, 8012-95-1) at 37 °C. Only oocytes that appeared morphologically healthy and underwent NEBD within 24 h upon retrieval were included in the study. NEBD of the GV was observed using a Primo Vision Evo+ timelapse camera (Vitrolife) installed inside the incubator. Thirty-nine oocytes were fixed at defined timepoints around NEBD. NEBD was scored by the disappearance of the GV observed using the Primo Vision Evo+ and confirmed under the stereomicroscope. Thirteen cells were fixed in the GV stage, 25 cells were fixed within 30 min of NEBD and 1 cell was fixed 1 h after NEBD. The remaining 17 oocytes had NEBD shortly before they were obtained from the IVF clinic. All 56 human oocytes were co-stained with an anti-α-tubulin antibody (MCA78G, Bio-Rad; 1:3,000) to visualize microtubules, an anti-LaminA/C antibody (Sigma-Aldrich, SAB4200236; 1:50) to visualize the nuclear lamina, 100 µM Hoechst 33342 (Molecular Probes; 20 mM stock) to stain DNA, and phalloidin conjugated with Alexa Fluor 488 or Rhodamine (Invitrogen; 1:20) to label F-actin. Twenty-seven oocytes had visible actin cables encapsulating the chromosomes, 21 oocytes had a microtubule meshwork encapsulating the chromosomes, four had already started to assemble a spindle, two had a bipolar spindle and two cells were not well fixed as evident from a poor preservation of the actin and microtubule cytoskeleton following immunofluorescence analysis.

The images and graphs in Fig. 1 and Extended Data Fig. 1a were generated by analysing videos from 35 live human oocytes that were recorded as part of a previous study5. All cells included in the analysis underwent NEBD within 24 hours of retrieval, formed MI spindles and progressed through anaphase I.

Expression constructs and mRNA synthesis

To generate the constructs for in vitro messenger RNA synthesis, previously published coding sequences were inserted into pGEMHE49 together with the miRFP50, mScarlet51, mClover3 (ref. 52) or SNAPf from pSNAPf (New England Biolabs) to generate pGEMHE-SNAPf, pGEMHE-RanWT-miRFP (Addgene #59750) (ref. 53), pGEMHE-RanT24N-miRFP (Addgene 61. 53), pGEMHE-H2B-SNAPf54, pGEMHE-SNAPf-FH2 (ref. 36), pGEMHE-SNAPf-KIND36, pGEMHE-mScarlet-UtrCH28 and pGEMHE-mClover3-hCenpC55. The expression constructs pGEMHE-H2B-mCherry30, pGEMHE-H2B-mScarlet56, pGEMHE-mEGFP-MAP4 (ref. 30), pGEMHE-EGFP-UtrCH29, pGEMHE-mScarlet-hCenpC55, pGEMHE-MyrGFP29 and pGEMHE-EGFP-Lamin B1 (ref. 29) were generated in previous studies. All mRNAs were synthesized with T7 polymerase (HiScribe T7 ARCA mRNA Kit) following the manufacturer’s instructions.

Microinjection of porcine oocytes and transient protein expression

Porcine oocytes were microinjected with 2 pl of mRNAs as previously described9, with the exception that an ‘injection shelf’ was generated by assembling two coverslips around a spacer consisting of two layers of double-sided sticky tape (Scotch; 3 M, 136R2) to accommodate ~120–130-µm-large porcine oocytes. Oocytes were microinjected immediately after isolation and denudation. mRNAs were injected at the following concentrations: mEGFP-MAP4 at 26 ng µl−1, H2B-mCherry and H2B-mScarlet at 1–5 ng µl−1, H2B-SNAPf at 10–30 ng µl−1, mClover3-hCenpC and mScarlet-hCenpC at 20 ng µl−1, EGFP-UtrCH and mScarlet-UtrCH at 180–360 ng µl−1, RanWT-miRFP at 1.25 µg µl−1, RanT24N-miRFP at 2.5 µg µl−1, SNAPf-FH2 at 2 µg µl−1, SNAPf-KIND at 2.5 µg µl−1, SNAPf at 1–1.25 µg µl−1 and EGFP-Lamin B1 at 34.5 ng µl−1. In most experiments, oocytes were allowed to express mRNAs for 3–4 h before release into dbcAMP-free medium for imaging. For three-colour live imaging of F-actin, kinetochores and chromatin, high-temporal-resolution imaging of F-actin after NEBD and in experiments with zeocin; cells were incubated in 50 µM Butyrolactone I for 12–15 h before they were released into Butyrolactone I-free medium for imaging.

Exoenzyme C3 (Cytoskeleton, CT03) was reconstituted in embryo tested water to obtain a 10 µM stock. Oocytes were microinjected with 10 pl of 2.4 µM exoenzyme C3, or the equivalent amount of bovine serum albumin (BSA, Sigma, A3311). Clostridium difficile toxin B (Sigma-Aldrich, SML1153) was dissolved in RNase free water to obtain a 1.5 µM stock. Oocytes were microinjected with 10 pl of 100 nM toxin B, or an equivalent amount of BSA.

Dextran labelling

Following the manufacturer’s instructions (MAN0001774|MP00143), 40 kDa amino dextran (Molecular Probes, D1861) was conjugated with Alexa Fluor 647 NHS Ester (Molecular Probes, A20006) in an equimolar reaction.

Drug addition and washout

All drugs were prepared in hybridoma-grade DMSO (Sigma-Aldrich, D2650). To depolymerize microtubules, oocytes were treated with 10 µM nocodazole (Sigma-Aldrich, M1404). Nocodazole powder was freshly dissolved before each experiment to obtain a 15 mM stock. To depolymerize actin, oocytes were treated with 5 µg ml−1 cytochalasin D (Sigma-Aldrich, C8273; 25 mg ml−1 stock stored at −20 °C). To stabilize actin filaments, oocytes were treated with 1 µM jasplakinolide (Invitrogen, J7473; 1 mM stock stored at −20 °C). Nocodazole and cytochalasin D were added to dbcAMP-free M2 medium 30 min before imaging. Jasplakinolide was added to dbcAMP-free M2 medium 1 h 30 min before imaging. To wash out cytochalasin D and nocodazole, cells were washed through four 35 mm Petri dishes, each containing 2 ml of M2 medium without drugs. All drug washouts were done approximately 1–3 h after NEBD. The activity of Arp2/3 was inhibited using 200 µM CK-666 (Sigma-Aldrich, SML0006; 130 mM stock stored at −20 °C). Myosin light chain kinase was inhibited using 6 µM ML-7 (Sigma-Aldrich, 475880; 22 mM stock stored at −20 °C). CK-666 and ML-7 were added to dbcAMP-free M2 medium 1 h before imaging.

Zeocin preparation

Zeocin powder (InvivoGen, ant-zn-1p) was dissolved in HEPES buffer at a stock concentration of 100 mg ml−1 (stock stored at 4 °C). To induce chromosome fragmentation, porcine oocytes were treated for 15 h with 500 µg ml−1 of zeocin diluted in M16 medium. To stop chromosome fragmentation, cells were transferred into zeocin-free M2 medium.

Cold-mediated microtubule depolymerization assays

Porcine oocytes were incubated on ice for 7 min and immediately fixed and processed for immunofluorescence microscopy as described below.

Immunofluorescence

To obtain oocytes at defined timepoints after NEBD, cells were monitored on a confocal microscope with transmission, or with a Primo Vision system. NEBD was determined as the moment when the nucleoli began to shrink.

Porcine and human oocytes were fixed in a solution containing 100 mM HEPES (pH 7.0, titrated with KOH; 1 M stock), 50 mM EGTA (pH 7.0, titrated with KOH; 0.5 M stock), 10 mM MgSO4 (1 M stock), 2% methanol-free formaldehyde (10% stock) and 0.5% Triton X-100 (10% stock; Sigma Aldrich, 93443) at 37 °C for 30 (porcine oocytes) or 60 min (human oocytes). Fixed oocytes were subsequently extracted in phosphate-buffered saline (PBS) supplemented with 0.5% Triton X-100 (PBT) overnight at 4 °C and thereafter blocked in PBT with 5% BSA (Fisher BioReagents BP1605-100, 11483823) (PBT-BSA) overnight at 4 °C. Porcine oocytes were then optically cleared as previously described56. All antibody incubations, F-actin and chromosome staining were performed in PBT-BSA. Primary antibodies were incubated overnight at 4 °C. Secondary antibody, fluorescently tagged phalloidin and Hoechst staining were performed for 1 h at room temperature. Primary antibodies used were rat anti-α-tubulin (MCA78G, Bio-Rad; 1:3,000), rabbit anti-Fmn2 (Atlas antibodies, HPA050649; 1:50), mouse anti-Lamin A/C (Sigma-Aldrich, SAB4200236; 1:50) and human anti-centromere antibody (ACA) (15–234, Antibodies Incorporated; 1:250). Secondary antibodies used were Alexa Fluor 488-, 564-, 568- or 647-conjugated anti-rat IgG, anti-rabbit IgG, anti-mouse IgG or anti-human IgG, all raised in goat or donkey (Molecular Probes; 1:400). DNA was stained with 100 µM Hoechst 33342 (Molecular Probes; 20 mM stock). F-actin was stained with phalloidin-Alexa Fluor 488 (Invitrogen; 1:20).

Confocal microscopy

Confocal imaging was performed in ~20 µl of M2 medium (for live oocytes) or PBS with 1% polyvinylpyrrolidone (for fixed oocytes) under paraffin oil in a 35 mm dish with a #1.0 glass coverslip. Images were acquired with LSM800, LSM880 and LSM900 confocal laser scanning microscopes (Zeiss) equipped with an environmental incubator box and a 40× C-Apochromat 1.2 NA water-immersion objective. Live samples were imaged at 38.5 °C, and fixed at 25 °C. Automatic 3D tracking was implemented for time-lapse imaging with a temporal resolution between 1.5 min and 10 min using AutofocusScreen57 or MyPiC58 on the LSM880, or by using a customized sample tracking solution provided by ZEISS Microscopy on LSM800 and LSM900 microscopes. In all experiments, control and experimental groups were imaged together, under identical imaging conditions on the same microscope. For some images presented in the figures, noise was reduced with a Gaussian filter in ZEN (Zeiss). Airyscan images were acquired using the airyscan module on LSM800, LSM880 and LSM900 confocal laser scanning microscopes (Zeiss) and processed in ZEN (Zeiss) after acquisition. Care was taken that imaging conditions (laser power, pixel-dwell time and imaging frequency and detector gain) did not cause phototoxicity (for live imaging), photobleaching or saturation. The actin cortex was saturated in experiments where F-actin in the cytoplasm was imaged.

Automatic detection of NEBD in porcine oocytes on a confocal microscope

The resumption of meiosis in porcine oocytes is asynchronous. Thus, to perform high-resolution spatial and temporal imaging of F-actin after NEBD, we designed a macro that allowed us to automatically detect NEBD.

To determine the timing of NEBD, we took advantage of a fluorescently labelled (Alexa Fluor 647 NHS Ester (succinimidyl ester); Invitrogen, A20006) 40 kDa dextran (Molecular Probes, A20006), which localizes to the cytoplasm in GV oocytes but upon NEBD quickly distributes within the nucleus. We recorded coarse resolution images of the dextran in the early stages of meiosis. The acquired images were read and analysed using the Bioformats59 package within a MATLAB Compiler Runtime Environment. To evaluate if NEBD had initiated, a sphere of 10 µm in diameter was placed at the centre of mass of the chromosome signal, and the average intensity of the dextran within this sphere was computed at every timepoint. A switch to high-resolution image acquisition was triggered when the intensity within the sphere became higher than a variable threshold. The threshold was determined for each oocyte at every timepoint individually, and was based on the median and the standard deviation of previous measurements. The script keeps the time interval for each imaged oocyte as the image acquisition interval set in the Zen software (Zeiss). The script was designed to work with the Zen Blue (Zeiss) image acquisition software.

Light sheet microscopy

Light sheet imaging was performed on an LS1 Live light sheet microscope system by Viventis. The illumination optics consist of two opposing Nikon CFI Plan Fluor air illumination objectives. We used a Nikon CFI75 Apochromat 25xC W NA 1.1 water immersion objective with switchable magnification between 18.75× and 37.5× (field of view 700 µm and 340 µm, respectively) as detection objective. The microscope had a fast motorized six-position filter wheel with emission filters in front of the camera. The following laser lines and emission filters were used:

Laser 488 nm, 60 mW diode laser, Semrock FF03-525/50-25 emission filter

Laser 515 nm, 50 mW diode laser, Semrock FF01-539/30-25 emission filter

Laser 561 nm, 50 mW DPSS laser, Semrock FF01-523/610-25 emission filter

Laser 638 nm, 100 mW diode laser, Chroma ZET405/488/561/640 m emission filter.

The camera used was an Andor Zyla 4.2 PLUS USB 3 sCMOS.

Immunoblotting

Thirty porcine GV oocytes per lane were washed in PBS and resuspended in 5 µl of PBS. NuPAGE LDS sample buffer (4×) (Thermo Fisher Scientific, NP0008) with 100 mM dithiothreitol was then added to the sample to a final 1× concentration, and the mixture was immediately snap-frozen in liquid nitrogen. Samples were thawed and frozen thrice more, followed by a 5 min incubation at 95 °C. Samples were resolved on a 10-well NuPAGE Tris-acetate protein gel of 1.0 mm thickness (Thermo Fisher Scientific, EA0375BOX) with NuPAGE Tris-Acetate SDS Running Buffer (Thermo Fisher Scientific, LA0041). Proteins were transferred onto an 0.45 µm Invitrolon PVDF Filter Paper Sandwich (Invitrogen, LC2005) with a NuPAGE Transfer Buffer (Thermo Fisher Scientific, NP00061) at 100 V for 1.5 h on ice. Blocking and antibody incubations were performed in PBS with 5% skimmed milk (w/v) (Sigma-Aldrich, LC2005) and 0.1% Tween-20. Primary antibodies were incubated overnight at 4 °C. We used rabbit anti-Fmn2 (Atlas antibodies, HPA05064; 1:1,000) as primary antibody and goat anti-Rabbit IgG (H + L) Cross-Adsorbed, HRP (ThermoFisher, 31462; 1:10,000) as secondary. Blots were developed with Super Signal West Femto Maximum Sensitivity Substrate (Thermo Fisher Scientific, 34094) and imaged with an Amersham Imager 600 (GE Healthcare).

General quantifications

Time-lapse videos of live porcine oocytes were analysed using Imaris 8.4.1 or Imaris 9.3.1 (Bitplane). NEBD in human and porcine oocytes was defined as the last time point before the nucleolus begins to visibly shrink/disappear.

Automatic creation of chromosome isosurfaces in Imaris

To consistently reconstitute the chromosome isosurfaces in Imaris, we used a MATLAB script published elsewhere56.

Automatic quantification of chromosome clustering speed

To calculate the speed of chromosome clustering, we first calculated the largest pairwise distance measured between all chromosomes for every timepoint. The change of the largest pairwise distance between the timepoints was then used to derive the speed of chromosome clustering.

To consistently and reliably quantify the largest pairwise distance between all chromosomes, we developed an in-house MATLAB script.

The script requires that isosurfaces are defined as vertices connected into polygons, and hence works only with versions older than Imaris 9. The script computes distances between all pairs of vertices via the ‘pdist’ MATLAB function. The largest distance is then selected and depicted in Imaris as a pair of spots. The largest distance is also added as a custom statistic to the isosurface object. For usage within Imaris, the script was compiled for MATLAB Compiler Runtime v95.

Automatic creation of a convex hull from 3D reconstituted chromosomes

To create a 3D convex hull, we used an in-house-developed MATLAB script. The script requires that isosurfaces are defined as vertices connected into polygons, and hence works only with versions older than Imaris 9. The script uses the MATLAB function ‘convhulln’ to identify the subset of vertices of a given isosurface making up the facets of the corresponding convex hull of this isosurface. For usage within Imaris, the script was compiled for MATLAB Compiler Runtime v93.

PIV analysis from high-temporal-resolution F-actin imaging after NEBD

To measure velocity vectors within the nuclear area, we first determined the boundaries of the nucleus based on EGFP-Lamin B1 signal. For this, the signal of the nuclear lamina was segmented using edge-aware local contrast enhancement followed by edge detection. The resulting fragmented and noisy segmentation was translated to polar coordinates and the outliers were removed. The centre of the polar coordinate system was defined by the centre of mass of the segmented nuclear lamina. Closing of the membrane was achieved by fitting a smoothing spline to the coordinates theta and rho of the remaining pixels. The smoothing spline was expanded by 1.5 μm or 2.5 μm per direction to produce a segmented image of the nuclear lamina.

The analysis was performed in MATLAB using ‘localcontrast’, ‘edge’ (with Canny algorithm), ‘rmoutliers’ and ‘fit’ functions. The exact parameters for these functions had to be adjusted in a few cases of lower signal-to-noise ratio. To import the images into MATLAB R2018b, we used the toolbox Bioformats59.

To distinguish movement of the entire cell or nucleus from movement of actin, the images were stabilized with the option ‘translation’ using the Zeiss Zen function ‘time alignment’ on the signal of the nuclear lamina.

Actin velocities were detected using PIVlab60 in MATLAB R2018b on the stabilized images. The Wiener2 denoise and low pass filters with 5 pixel window size were used for pre-processing. The particle image velocimetry (PIV) algorithm used was ‘FFT window deformation’. The interrogation windows were reduced in three steps of 64, 32 and 24 pixels width. The correlation robustness was set to ‘extreme’. All options for post-processing were disabled, which resulted in weak filtering of the velocity vectors.

The direction of the velocity vectors was set relative to the centre of the nucleus. The centre of the nucleus was determined on the basis of the centre of mass of the widened smoothing spline. The segmentation of the nucleus was used to identify which vectors were located inside the nucleus, but not to mask the nuclear membrane for the PIV analysis.

Circular statistics were computed using the Circular Statistics Toolbox61. Mean directions and the 95% confidence intervals are based on the velocity vectors with speeds exceeding two pixels per frame or 1.8 µm min−1.

Semi-automatic analysis of chromosome clustering upon chromosome fragmentation

H2B-positive foci were detected, counted and analysed using the surface function in Imaris 9.6.0 (Bitplane). Foci were classified manually as kinetochore-containing or kinetochore-free on the basis of the presence or absence of the hCenpC signal, respectively. The total time required for complete chromosome clustering was calculated as the time between NEBD and the first timepoint when only one H2B-positive surface was detected. Cells were scored as having completed chromosome clustering only if all the H2B-foci formed a single surface during the duration of imaging. Individual H2B foci with and without a kinetochore were tracked manually to estimate the time required for their clustering. Only foci that could be reliably tracked were analysed. Chromatin surrounding the nucleolus was excluded from the analysis of individual H2B foci.

Co-localization analysis for Formin-2 and F-actin

The specificity of the anti Formin-2 antibody was tested using the statistic ‘Shortest distance to Surface’ computed using Imaris 9.7.2 and Imaris 9.3.0 by Bitplane.

For this purpose, spots were created once on the Formin-2 signal and once on the Formin-2 signal mirrored along the x axis. These spots were compared relative to a surface segmentation of the actin signal (Phalloidin). Before the analysis, a region of interest of 10 μm × 10 μm × 10 μm was manually cropped. The region of interest was placed so that it included neither the nucleus nor the very edge of the image. Around 100 spots were created per cell to improve comparability.

The surface on actin was created with a batch script previously published elsewhere56 using Otsu’s method for thresholding and ensuring the volume of the actin surface to be within 30 μm³ to 1,000 μm³.

Manual quantification of the mean fluorescence intensity of actin filaments after NEBD

Quantification of the mean fluorescent intensity of actin filaments was done in ImageJ on two-dimensional images. To quantify the mean fluorescence intensity of actin filaments after NEBD, a region of interest of consistent size was placed next to the chromosome signal. Background was subtracted on the basis of intensity measurements in a region outside of the cell.

Statistics and reproducibility

Average (mean), standard deviation and statistical significance were calculated in GraphPad Prism (8.4.3). Statistical tests selected for each dataset are indicated in the figure legends. All data were tested for normal distribution in GraphPad Prism (8.4.3) before selection of the statistical tests. All graphs were generated in GraphPad Prism (8.4.3), except for the graphs in Fig. 4d, which were generated in Origin 2021b, and in Fig. 4e, which were generated in MATLAB R2018b. All data from porcine oocytes are from at least two independent experiments/multiple biological replicates. The number of experimental replicates for each assay is specified in the figure legends. All replication attempts were successful. No statistical methods were used to pre-determine sample size. Due to the large size and opaque nature of porcine oocytes, cells that had nuclei positioned in the top third of the cell could not be imaged well and were hence excluded from the analysis. Cells that died during imaging were excluded from the analysis. For each independent experiment, porcine oocytes were collected from multiple ovaries/animals and subsequently pooled together. The oocytes were then randomly split into a control group and an experimental group. The investigators were not blinded to allocation during experiments and outcome assessment.

Resource availabilityContact for reagent and resource sharing

Further information and requests for resources and reagents should be directed to and will be fulfilled by the corresponding author Melina Schuh (melina.schuh@mpinat.mpg.de).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.