Cytoplasmic competition between separate parental pronuclei in zygotes

Abstract

Embryogenesis begins with a zygote—a single cell with two pronuclei that separately enclose maternal and paternal chromosomes. The functional significance of the separation of parental chromosomes into distinct pronuclei remains unexplored, despite the fact that one-pronuclear biparental zygotes are used clinically^1–3. Here, using a combination of mouse zygote manipulation, quantitative imaging and theoretical approaches, we show a cytoplasm-mediated competition mechanism between separate parental pronuclei that ensures developmental potential. This mechanism limits pronuclear volume and prevents epigenetic mark dysregulation, including loss of trimethylated histones. One-pronuclear biparental zygotes lack this mechanism, resulting in a reduced rate of development to term. This low developmental potential can be partially rescued by competition-based or drug-based restoration of epigenetic marks. This study provides a spatial mechanism linking fertilization to the establishment of the full developmental potential for the next generation, highlighting caveats in clinical use of one-pronuclear biparental zygotes.

Drug-based and manipulation-based strategies provide evidence of pronuclear competition for finite cytoplasmic resources between separate parental pronuclei in zygotes, suggesting a cell biological mechanism for the cytoplasm-mediated remote parental interaction that regulates epigenetic marks.

Main

Canonical cells have a single nucleus containing both maternal and paternal genomes. In contrast, in mammalian zygotes (or fertilized eggs), which are formed by the fusion of an egg and a sperm at the beginning of a new life, the maternal and paternal genomes are contained separately in two compartments called pronuclei. This two-pronuclear (2PN) state is maintained until the pronuclear envelopes break down at the entry of the first mitosis^4–7. The parental chromosomes are thus spatially separated at mitotic entry and are therefore at risk of chromosome segregation errors^4–6. Although the risks that arise from the 2PN state have been described, the benefits of the 2PN state remain unclear. Previous studies using haploid parthenotes and androgenotes show that both parental genomes are required for mammalian zygotes to establish chromatin regulation and developmental potential^8–10. However, no studies have examined one-pronuclear (1PN) biparental diploid zygotes. It is therefore unclear whether the separation of the parental genomes into two pronuclei is functionally important.

In clinical assisted reproduction, 1PN zygotes are commonly observed, accounting for approximately 2–8% of in vitro fertilization (IVF) and intracytoplasmic sperm injection (ICSI) cases^11,12. A considerable proportion of these 1PN zygotes are biparental diploids, a fraction of which can develop to term with no apparent abnormalities^1,2. Based on these results, 1PN biparental zygotes have been considered recently as a source of clinically useful embryos. However, clinical studies suggest that 1PN biparental zygotes are less likely to develop to term^2,3. Whether and how the 1PN biparental state affects the developmental potential of zygotes has not been studied experimentally.

Generation of 1PN biparental zygotes

To investigate the functional significance of the separation of the parental genomes into two pronuclei, we established an ICSI-based method to generate 1PN biparental zygotes using mice as a model (Fig. 1a). Live imaging showed that the cell-cycle length from anaphase II to the first M-phase entry in 1PN biparental zygotes was not significantly different from that in control 2PN zygotes (Extended Data Fig. 1a,b). Three-dimensional analysis of 1PN biparental zygotes at the late G2 stage (10 h after ICSI) showed that the volume of the biparental pronucleus was almost equal to the total volume of the maternal and paternal pronuclei of control 2PN zygotes (Fig. 1b). The biparental pronucleus consisted of two distinct hemispheres, each showing enrichment of 5-methylcytosine (5mC) or 5-hydroxymethylcytosine (5hmC) on DNA (Fig. 1c and Extended Data Fig. 1c), markers for maternal and paternal chromosomes, respectively¹³. Imaging of the first mitosis of 1PN zygotes revealed that 5mC-enriched chromosomes were separated spatially from 5hmC-enriched chromosomes, as observed in the mitosis of 2PN zygotes (Extended Data Fig. 1d)^4–6. Thus, 1PN biparental zygotes form an enlarged pronucleus with little intermixing of parental chromosomes.

Fig. 1. One-pronuclear biparental zygotes show reduced trimethylated histones in an enlarged pronucleus.

a, Schematic diagram of the production of 1PN biparental zygotes. Zygotes were fixed 10 h after ICSI. b, 1PN biparental zygotes show an enlarged pronucleus. Pronuclear volume was determined from three-dimensional reconstructed images of histone H3. c, Spatial separation of parental chromosomes in 1PN zygotes. Immunostaining of 5mC (magenta; maternal genome) and 5hmC (green; paternal genome). d, Representative images of trimethylated histones (green) H3K4me3 (left), H3K27me3 (middle) and H3K9me3 (right). Histone H3 was stained for reference. Colour-coded images show the signal intensity ratio. e, Quantification of total fluorescence intensities of H3K4me3 (left), H3K27me3 (middle) and H3K9me3 (right), normalized to control 2PN zygotes. f, Schematic of regions of interest (ROIs) used for intensity measurements (‘Interior’ and ‘Peri-NPB’). g,h, Maternal pronuclear intensities of H3K4me3 (left panels) and H3K27me3 (right panels) relative to histone H3 intensity within interior (g) and peri-NPB (h) ROIs, normalized to control 2PN zygotes. Pronuclear volume is shown in Extended Data Fig. 1g. Number of oocytes from three independent experiments. Statistical test, Mann–Whitney U test (b,g,h). Error bars, s.e.m. 3D, three-dimensional; biPN, biparental pronucleus; H3, histone H3; mPN, maternal pronucleus; pPN, paternal pronucleus. Scale bars, 5 μm.

Source data

Extended Data Fig. 1. 1PN zygote.

(a) Live imaging of pronuclear formation in 1PN biparental zygotes. Maximum z-projection images of maternally-encoded H2B-GFP (green) and microinjected mCherry-MBD-NLS (magenta). Note that H2B-GFP and mCherry-MBD-NLS temporarily serve as markers for maternal and paternal chromosomes, respectively, after fertilization, but they gradually mix between parental chromosomes. Time after the anaphase onset of meiosis II (Ana II, hr:mm). Magenta, cyan and white arrowheads show maternal chromosomes, paternal chromosomes and the polar body, respectively. (b) Cell-cycle length. Duration from anaphase onset of meiosis II (Ana II) to the entry to the 1st mitosis (NEBD, nuclear envelope breakdown) was determined. Number of oocytes from four independent experiments. Mann-Whitney U test was performed. Error bars, SEM. (c) 5mC and 5hmC signal intensities normalized to that of control 2PN zygotes. Number of oocytes from three independent experiments. Mann-Whitney U test was performed. Error bars, SEM. (d) Spatial separation of parental chromosomes in the first mitosis of 1PN biparental zygotes. Immunostaining of 5mC (magenta; maternal genome marker) and 5hmC (green; paternal genome marker). (e,f,j) Immunostaining of trimethylated histones and methylated DNA in 1PN biparental zygotes. Representative images of H3K4me3 (green) and 5mC (magenta) (e), of H3K27me3 (magenta) and 5hmC (green) (f), and H3K27ac (green) and 5hmC (magenta) (j). The line plots show the signal intensities of H3K4me3 (green) and 5mC (magenta), of H3K27me3 (magenta) and 5hmC (green), and H3K27ac (green) and 5hmC (magenta) along the white line, respectively. Representative images are shown. Similar results were obtained from two independent experiments. (g) Pronuclear volume of zygotes using immunostaining analysis in Fig. 1d,e. Number of oocytes from three independent experiments. Mann-Whitney U test was performed. Error bars, SEM. (h,i) Immunostaining of H3K4me3 in 1PN biparental zygotes after pre-extraction treatment. Representative images of H3K4me3 (green) and of their signal intensity ratio (H3K4me3/H3, fire colour code) are shown in (h). H3K4me3/H3 ratio values within interior and peri-NPB ROIs, were normalized to that of control 2PN zygotes (i). Number of oocytes from four independent experiments. Mann-Whitney U test was performed. Error bars, SEM. (k,l) Immunostaining of H3K27ac in 1PN biparental zygotes. Representative images of H3K27ac (green) and of their signal intensity ratio (H3K27ac/H3, fire colour code) are shown in (k). H3K27ac/H3 ratio values within interior and peri-NPB ROIs, were normalized to that of control 2PN zygotes (l). Number of oocytes from two independent experiments. Mann-Whitney U test was performed. Error bars, SEM.

Source data

We then asked whether the 1PN biparental state affects intranuclear regulation. One regulation characteristic of zygotes is the asymmetric regulation of epigenetic marks between maternal and paternal chromosomes. Specifically, trimethylated histones such as H3K4me3, H3K9me3 and H3K27me3 are inherited from the egg and their global levels are relatively maintained on maternal chromosomes throughout the zygote (1-cell) stage, whereas they are largely absent in sperm and gradually acquired on paternal chromosomes after fertilization^14–22 while undergoing substantial reprogramming of their genome-wide patterns on both parental genomes^23–26. These epigenetic marks play critical roles in embryonic development^14–16,21. To obtain an estimate of whether the global levels of these modifications are altered in 1PN biparental zygotes, we measured the total immunofluorescence intensity of modification-specific antibodies throughout the entire volume of the biparental pronucleus in 1PN zygotes, and compared it with the combined total intensity of both pronuclei in control 2PN zygotes at the late G2 stage. This analysis detected significantly reduced levels of H3K4me3 and H3K27me3 but not H3K9me3 in 1PN biparental zygotes (Fig. 1d,e), raising the possibility that the global levels of H3K4me3 and H3K27me3 are dysregulated.

To achieve robust quantification of H3K4me3 and H3K27me3, we calculated the signal ratio between modification-specific antibodies and pan-H3 antibodies, focusing on the maternal regions, which consistently exhibited sufficient signal intensities for reliable measurement, in contrast to the paternal regions in our experimental conditions (Fig. 1d). To ensure consistency and to avoid bias from structurally distinct regions, we measured signal intensities in two defined categories within the maternal hemisphere: (1) the nuclear interior away from the nucleolus precursor body (NPB) and (2) the peri-NPB region, while avoiding the estimated maternal–paternal boundary in 1PN biparental zygotes (Fig. 1f). This quantification showed significantly lower levels of H3K4me3 and H3K27me3 in the maternal hemisphere interior of 1PN zygotes compared with that of the maternal pronucleus of 2PN zygotes (Fig. 1d,g and Extended Data Fig. 1e–g). Peri-NPB regions of 1PN zygotes also showed significantly lower maternal levels of these marks (Fig. 1d,h). We confirmed the decrease in H3K4me3 levels in the maternal hemisphere using 1PN biparental zygotes fixed after pre-extraction (Extended Data Fig. 1h,i). Furthermore, H3K27ac levels were also decreased significantly in the maternal hemisphere of 1PN biparental zygotes (Extended Data Fig. 1j–l), indicating that the 1PN state affects not only H3K4me3 and H3K27me3 but also other modified histones.

Together, these results suggest that 1PN biparental zygotes are defective in regulating global levels of several epigenetic marks, including maternal H3K4me3, H3K27me3 and H3K27ac.

Abnormalities in epigenomic profiles

To investigate genome-wide epigenetic profiles, we mapped H3K4me3 and H3K27me3 in 2PN and 1PN zygotes by using low-input carrier DNA-assisted chromatin immunoprecipitation followed by sequencing (CATCH-seq)^27,28. Our CATCH-seq data were highly reproducible between replicates and showed high correlations with previously reported datasets (Extended Data Fig. 2a,b). For this analysis, we used hybrid zygotes (BDF1 × JF1), which allowed us to determine the parental origin of modified-histone-bound DNA based on single-nucleotide polymorphisms (SNPs). H3K4me3 and H3K27me3 were more enriched on the maternal genome than on the paternal genome in control 2PN zygotes, consistent with previous reports^23,29 (Extended Data Fig. 2c) and our immunofluorescence (Fig. 1d). In 1PN zygotes, the overall maternal–paternal ratio of H3K4me3 was not significantly different from that in control 2PN zygotes (Extended Data Fig. 2c), although immunofluorescence quantification indicated H3K4me3 reductions in the maternal regions (Fig. 1d–h). This observation implied that its paternal levels may also be reduced in 1PN zygotes. Indeed, 1PN zygotes exhibited lower enrichment of H3K4me3 at TSS regions on the paternal allele when compared with 2PN zygotes (Extended Data Fig. 2d,e,h), suggesting inefficient establishment of H3K4me3 on the paternal allele after fertilization. By contrast, noncanonical broad peaks of maternal H3K4me3 (refs. ^23–25) retained in 1PN zygotes, with a few peak regions being increased or decreased (Extended Data Fig. 2i,j). The distribution of H3K27me3 was also changed in 1PN zygotes: it was lowered significantly at the paternal allele of 1PN zygotes when compared with 2PN zygotes (Extended Data Fig. 2k–m). Consistently, the overall maternal bias of H3K27me3 was pronounced in 1PN zygotes (Extended Data Fig. 2c). Given that H3K27me3 is newly deposited on the paternal chromatin after fertilization³⁰, this result suggests that de novo H3K27me3 establishment is impaired in 1PN zygotes. Unlike the paternal genome, reduction of histone methylation at the maternal genome was not clearly detected, possibly due to random loss occurring across the maternal genome or below a detection limit of our epigenomic profiling technology. Nevertheless, these results suggest that de novo establishment of paternal H3K4me3 and H3K27me3 marks is inefficient in 1PN biparental zygotes.

Extended Data Fig. 2. Validation of genome-wide profiling of epigenetic marks with CATCH-seq.

(a) Correlation between CATCH-seq and public ChIP-seq^23,29 data of 2PN at the PN5 stage. (b) CATCH-seq. Scatter plot showing the correlation between biological replicates. Spearman correlation is also shown. (c) H3K4me3 and H3K27me3-bound sequences detected by CATCH-seq were subjected to parental allele classification. Differences in read counts between two alleles (paternal – maternal) normalized by the total number of allelic reads (paternal + maternal) for H3K4me3 and H3K27me3 between 1PN and 2PN zygotes. Two and three biologically independent experiments were performed for H3K4me3 and H3K27me3, respectively. Each dot, one experiment. Lines connect data from the same experiments. (d,f) Heatmap showing the enrichment levels of H3K4me3 around TSSs of all genes (left) and the averaged plot of each cluster (right). All genes were categorized into three clusters according to the signal levels. Welch t-test was performed. (e,g) Genome browser snapshot of H3K4me3 profiles. (h) GO-term analysis of Cluster1 genes was performed using DAVID (one-sided Fisher’s exact test, EASE). Bars show −log10(unadjusted P value); no multiple-comparison adjustment. (i) Heatmap showing the enrichment levels of maternal H3K4me3 around all broad peaks (>5 kb). The rows were ordered by the 1PN/2PN ratio of their peak regions. (j) Genome browser snapshot of broad H3K4me3 profiles at the maternal allele. (k) Genome browser snapshot of H3K27me3 profiles. (l) Heatmap showing H3K27me3 signal enrichment at the maternal and paternal allele-biased regions. (m,n) Box plot showing log2 fold change [1PN/2PN (m) and P1PN/2PN (n)] of H3K27me3 levels at the maternal and paternal allele-biased regions (maternal: n = 13,750 regions; paternal: n = 12,789 regions). Values were calculated from pooled CATCH-seq data of 2 biologically independent replicates. The centre line indicates the median. The box edges, upper, and lower whiskers indicate the interquartile range (IQR; from the 25th to 75th percentile), the largest value smaller than 1.5× the IQR above the 75th percentile, and the smallest value larger than 1.5× the IQR below the 25th percentile, respectively.

Epigenetic defects without intermixing

One possible mechanism for the dysregulation of epigenetic marks in 1PN biparental zygotes is the intermixing of nucleoplasmic factors derived from maternal and paternal chromosomes. To test this, we first used haploid parthenotes, which carry one enlarged pronucleus containing only maternal chromosomes (Fig. 2a,b). Quantification of immunofluorescence signals revealed that, even without paternal chromosomes, the interior and peri-NPB regions of the pronucleus of haploid parthenotes showed significantly decreased levels of maternal H3K4me3 (Fig. 2c,d) and H3K27me3 (Fig. 2e,f), as observed in 1PN biparental zygotes. Second, we used haploid androgenotes, which carry only one paternal pronucleus (Fig. 2a). CATCH-seq analysis detected reductions in paternal H3K4me3 levels in androgenotes (Extended Data Fig. 2f,g; ‘AG’), as observed in 1PN biparental zygotes. In contrast, reductions in paternal H3K27me3 levels were not detected (Extended Data Fig. 2k,l; ‘AG’), unlike 1PN biparental zygotes. These results suggest that the reduction in paternal H3K27me3 observed in 1PN biparental zygotes is due to biparental nucleoplasmic intermixing, whereas those of maternal H3K4me3 and H3K27me3 and paternal H3K4me3 are probably due to other mechanisms linked to the 1PN state.

Fig. 2. Dysregulation of the levels of maternal epigenetic marks is probably not due to nucleoplasmic intermixing.

a, Schematic diagram of the production of parthenotes and androgenotes. b, Pronuclear volume of parthenotes and androgenotes. Pronuclear volume was determined from three-dimensional reconstructed images of histone H3. c,e, Immunostaining of parthenotes and androgenotes. Representative images of trimethylated histones (green) H3K4me3 (c), H3K27me3 (e). Histone H3 was stained for reference. Colour-coded images show the signal intensity ratio. d,f, Ratio values within interior and peri-NPB ROIs normalized to that of the maternal chromosomes of control 2PN zygotes (H3K4me3 (d), H3K27me3 (f)). Number of oocytes from three independent experiments. Statistical test, Mann–Whitney U test (b–f). Error bars, s.e.m. Scale bars, 5 μm.

Source data

Cytoplasm limits pronuclear volume

A clue to the mechanism underlying the loss of H3K4me3 and H3K27me3 in 1PN biparental zygotes came from the observation in the above experiments: haploid parthenotes were similar to 1PN biparental zygotes not only in decreased trimethylated histones, but also in increased pronuclear volume (Figs. 1b and 2b). This suggests a link between epigenetic loss and pronuclear enlargement.

This finding led us to search for strategies to control pronuclear volume in zygotes. In many experimental models of mononuclear cells, the size of the cytoplasm scales with the size of the nucleus³¹. To investigate whether this general rule applies to 2PN zygotes, we took advantage of our established techniques for manipulating the cytoplasm of mouse oocytes³². We generated metaphase II (MII) oocytes carrying half the normal cytoplasmic volume and oocytes carrying twice the normal cytoplasmic volume (Fig. 3a and Extended Data Fig. 3a–c). We then fertilized these oocytes by ICSI or IVF to obtain ‘halved’ and ‘doubled’ zygotes (Fig. 3a and Extended Data Fig. 3a). Three-dimensional measurement of pronuclei at the late G2 stage showed that both maternal and paternal pronuclear volumes increased significantly in doubled zygotes, whereas both pronuclear volumes decreased in halved zygotes (Fig. 3b). These results suggest that both maternal and paternal pronuclear volumes are limited by a finite cytoplasmic pool of components.

Fig. 3. Cytoplasmic competition mechanism regulates pronuclear volume in zygotes.

a, Experimental scheme of the production of halved and doubled zygotes. The image shows MII oocytes used for ICSI or IVF. b, Pronuclear volume scales with cytoplasmic volume. Pronuclear volume was determined from three-dimensional reconstructed images of histone H3. c, Paternal-to-maternal ratio of pronuclear volume in halved and doubled zygotes. d, Faster nuclear import in paternal pronuclei. EGFP-NUP153 (nuclear pore marker, inverted grey scale) and IBB-DiHcRed (nuclear import marker, fire colour code) were imaged. Pronuclei of the same zygote are shown. Time after the onset of pronuclear growth is shown (h:min). e,f, Quantification of EGFP-NUP153 (e) and IBB-DiHcRed (f) signals. g, Model predicting the kinetics of pronuclear growth. h, Live imaging of pronuclear growth. Pronuclear volume was determined from three-dimensional reconstructed images. Number of oocytes given are from four and five independent experiments using doubled and halved zygotes, respectively (b,c) or from one of three independent experiments (e,f). Statistical test, Mann–Whitney U test (b,c). Error bars in b and c, s.e.m. Data in h are presented as mean values ± s.d. Scale bars, 100 μm (a); 5 μm (b,d).

Source data

Extended Data Fig. 3. Cytoplasm-mediated pronuclear-to-pronuclear competition mechanism.

(a,b) Experimental scheme of the production of halved and doubled zygotes. (c) Measurement of the volumes of halved and doubled meiosis II (MII) oocytes. Number of oocytes from three independent experiments. Data are presented as mean values ± SD. (d) Immunostaining of nuclear pores in 2PN zygotes. Signal intensity of mab414 (nuclear pore complex proteins, green and colour code) at the nuclear rim was measured. Normalized values are shown. Number of oocytes from three independent experiments. Mann-Whitney U test was performed. Error bars, SEM. (e) Immunostaining of NUP98 in 2PN zygotes. Signal intensity of NUP98 (green and colour code) at the nuclear rim was measured. Normalized values are shown. Number of oocytes from two independent experiments. Mann-Whitney U test was performed. Error bars, SEM. (f) Experimental scheme of pronuclear isolation for LC-MS/MS analysis. The image shows isolated pronuclei. (g) Paternal-to-maternal ratios of pronuclear volume and surface. Pronuclear volume and surface were measured after 3D reconstruction images of zygotes stained with Histone H3. Number of oocytes from three independent experiments. Error bars, SEM. (h) Quantitative measurement of the amount of nuclear pore proteins with MS-QBiC. Relative amounts of the peptides of the indicated proteins were calculated from individual peak intensities shown in (i–l). Three biologically independent experiments were performed. Each dot, one experiment. Welch’s t test was performed. (i–l) MS-QBiC data details. Extracted ion chromatograms (XIC) of the main molecular ions (red), the M + 1 ions (blue), and the M + 2 ions (green) are shown. The analysis was performed three times for each pronuclei (pPN-1, 2, 3 and mPN-1, 2, 3). The main molecular ions correspond to the light peptides originated from the samples (upper panels) and the heavy peptides originated from the spiked internal standards (lower panels), respectively. The XICs of the peptide derived from NUP35 (i), NUP93 (j), and NUP98 (k, l) are shown.

Source data

Competition between parental pronuclei

Maternal and paternal pronuclei probably share volume-limiting components because (1) the volume ratio between maternal and paternal pronuclei was not affected significantly in halved and doubled zygotes (Fig. 3c) and (2) the pronuclear volume of haploid parthenotes and 1PN biparental zygotes was comparable with the total volume of maternal and paternal pronuclei of 2PN zygotes (Figs. 1b and 2b).

Based on these observations, we hypothesised that pronuclear volume is regulated by cytoplasm-mediated pronuclear-to-pronuclear competition for limiting components. The limiting component mechanism in frog egg extracts and embryos is dependent on nuclear import activity^33,34. Live imaging and immunostaining analyses of 2PN zygotes showed that, during pronuclear growth, the paternal pronucleus exhibited a faster enrichment of nuclear pore components and faster nuclear import than the maternal pronucleus (Fig. 3d–f, Extended Data Fig. 3d,e and Supplementary Video 1). Consistently, mass spectrometry-based quantification by isotope-labelled cell-free product (MS-QBiC)³⁵ showed that the paternal-to-maternal ratios of nuclear pore components NUP35, NUP93 and NUP98 abundance were 1.98, 1.89 and 1.53, respectively, greater than the paternal-to-maternal ratio of pronuclear surface area, 1.34 (Extended Data Fig. 3f–l).

Based on these results, we developed a theoretical model assuming that (1) pronuclei increase their volume by incorporating components from the limited cytoplasmic pool through their nuclear pores and (2) the rate of nuclear pore production at the surface of the paternal pronucleus is higher than that of the maternal pronucleus. This model predicted a kinetic pattern of pronuclear growth (Fig. 3g), which was consistent with the live observation of zygotes (Fig. 3h). A key prediction of this model is that the pronuclear growth of haploid parthenotes should (1) be slower in the early phase, (2) be faster in the late phase and (3) result in a larger volume compared with that of the paternal pronucleus in 2PN zygotes (Fig. 3g). The results of pronuclear volume measurements in live parthenotes and zygotes were consistent with the prediction (Fig. 3h, Extended Data Fig. 4a and Supplementary Video 2). Furthermore, the model predicted the kinetics of pronuclear growth in doubled and halved zygotes, which were consistent with live observation (Extended Data Fig. 4b,c). Thus, our theoretical model and quantitative measurements suggest that cytoplasm-mediated pronuclear-to-pronuclear competition regulates pronuclear volume in zygotes.

Extended Data Fig. 4. A competition-based theoretical model predicts the kinetics of pronuclear growth.

(a) Live imaging of pronuclear growth. Maximum z-projection images of maternally-encoded H2B-GFP (green) and microinjected mCherry-MBD-NLS (magenta). Note that H2B-GFP and mCherry-MBD-NLS temporarily serve as markers for maternal and paternal chromosomes, respectively, after fertilization, but they gradually mix between parental chromosomes. Parthenotes were imaged using mCherry-MBD-NLS without maternally-encoded H2B-GFP. Time after the onset of pronuclear growth (hr:mm). Magenta, cyan and white arrowheads showed maternal chromosomes, paternal chromosomes and polar body, respectively. (b) Theoretical model predicts the kinetics of pronuclear growth in halved and doubled zygotes. (c) Live imaging of pronuclear growth in halved and doubled zygotes. Pronuclear volume was determined from 3D-reconstructed images. Number of oocytes analysed. Halved, Control, and Doubled samples were obtained from 3 of 4, 5 of 7, and 1 of 3 independent experiments, respectively. Data are presented as mean values ± SD.

Source data

Larger pronuclei, lower epigenetic marks

Using doubled and halved zygotes carrying pronuclei of different volumes, we investigated whether altered pronuclear volume is associated with loss of epigenetic marks. Notably, quantification of immunofluorescence signals showed that the interior and peri-NPB regions of the enlarged maternal pronucleus of doubled zygotes exhibited decreased global levels of H3K4me3 and H3K27me3 (Extended Data Fig. 5a,b), recapitulating the observations in the maternal hemisphere of the enlarged pronucleus of 1PN biparental zygotes (Fig. 1d–h). Conversely, the interior and peri-NPB regions of the small maternal pronucleus of halved zygotes showed increased global levels of H3K4me3 and H3K27me3 (Extended Data Fig. 5a,b). Thus, maternal pronuclear volume was associated inversely with the levels of maternal H3K4me3 and H3K27me3. Such an inverse association was also observed for maternal H3K27ac (Extended Data Fig. 5c). Thus, pronuclear volume is associated with loss of the global levels of several epigenetic marks including but not limited to maternal H3K4me3 and H3K27me3.

Extended Data Fig. 5. Pronuclear volume is associated with decrease in maternal trimethylated histones.

(a–c) Immunostaining of H3K4me3, H3K27me3 and H3K27ac in halved and doubled zygotes. Halved and doubled zygotes were immunostained for H3K4me3, H3K27me3 or H3K27ac (green). Histone H3 was stained for reference. Representative images of H3K4me3, H3K27me3, or H3K27ac (green) in the maternal pronucleus of the same zygotes are shown. Color-coded images show the signal intensity ratio (H3K4me3/H3, H3K27me3/H3 or H3K27ac/H3). Ratio values within interior and peri-NPB ROIs, are normalized to that of control 2PN zygotes and shown in the plots. Number of oocytes from four independent experiments for the H3K4me3 group, and from three independent experiments for all other groups. Mann-Whitney U test was performed. Error bars, SEM.

Source data

Competition-based epigenetic rescue

Based on these findings, we hypothesised that suppressing pronuclear enlargement could rescue the loss of epigenetic marks in 1PN biparental zygotes. To test this hypothesis, we introduced an extra pronucleus by inhibiting the extrusion of the second polar body using Cytochalasin B during the procedure to create a biparental pronucleus. We reasoned that the extra pronucleus would act as a competitor in the cytoplasm and suppress the growth of the biparental pronucleus. After completion of pronuclear growth at late G2, we removed the extra pronucleus from the cytoplasm (Fig. 4a; ‘post-competition 1PN zygote’, P1PN). Consistent with our model, post-competition 1PN zygotes exhibited a significantly reduced volume of the biparental pronucleus (Fig. 4b,c). Notably, quantification of immunofluorescence signals showed that the interior and peri-NPB regions of the biparental pronucleus of post-competition 1PN zygotes had strongly rescued levels of H3K4me3 and H3K27me3 in the maternal hemisphere (Fig. 4c–f). Furthermore, CATCH-seq analysis showed that post-competition 1PN zygotes had recovered the promoter H3K4me3 enrichment on the paternal genome (Extended Data Fig. 2f,g). In contrast, paternal H3K27me3 levels remained largely unrecovered (Extended Data Fig. 2k,l,n), consistent with the idea that this mark is reduced by biparental nucleoplasmic intermixing (Extended Data Fig. 2k,n). Overall, these results suggest that transient introduction of an extra pronucleus during pronuclear growth can partially rescue the loss of epigenetic marks in 1PN biparental zygotes.

Fig. 4. Manipulations that suppress pronuclear enlargement rescue epigenetic dysregulation in 1PN biparental zygotes.

a, Schematic diagram of the production of post-competition P1PN biparental zygotes. b, P1PN biparental zygotes show suppressed pronuclear enlargement. Pronuclear volume was determined from three-dimensional reconstructed images of histone H3. c,e, Immunostaining of P1PN biparental zygotes. Representative images of trimethylated histones (green) H3K4me3 (c) and H3K27me3 (e). Histone H3 was stained for reference. Colour-coded images show the signal intensity ratio. d,f, P1PN biparental zygotes show rescued maternal H3K4me3 and H3K27me3. Ratio values within interior and peri-NPB ROIs normalized to that of control 2PN zygotes (H3K4me3 (d), H3K27me3 (f)). g–j, Cytoplasmic halving rescues H3K4me3 and H3K27me3 in 1PN biparental zygotes. Representative images of trimethylated histones (green) H3K4me3 (g) and H3K27me3 (i). Histone H3 was stained for reference. Colour-coded images show the signal intensity ratio. Ratio values within interior and peri-NPB ROIs normalized to that of control 2PN zygotes (H3K4me3 (h), H3K27me3 (j)). k, Cytoplasmic halving suppresses pronuclear enlargement of 1PN zygotes. Pronuclear volume was determined from three-dimensional reconstructed images of histone H3. Number of oocytes from five (b,d) or three (f–k) independent experiments. Statistical tests, one-way analysis of variance with Dunn’s multiple-comparison test (b,d,f); Mann–Whitney U test (k). Error bars, s.e.m. Scale bars, 5 μm (c,e); 10 μm (g,i).

Source data

The rescue effect of the extra pronucleus on the epigenetic mark levels of the biparental pronucleus could be mediated by (1) induced pronuclear-to-pronuclear competition for a limited cytoplasmic resource or (2) transport of factors from the extra pronucleus to the biparental pronucleus. To distinguish between these possibilities, we mimicked the cytoplasmic effect of pronuclear-to-pronuclear competition by reducing the cytoplasm by micromanipulation. In support of the hypothesis that competition mediates epigenetic rescue, cytoplasmic reduction significantly rescued the immunofluorescence levels of H3K4me3 and H3K27me3 in the interior and peri-NPB regions of the maternal hemisphere of 1PN biparental zygotes (Fig. 4g–j). Consistent with our model, this effect was associated with a significant suppression of the enlargement of the biparental pronucleus (Fig. 4g,k). We conclude that induction or mimicry of pronuclear-to-pronuclear cytoplasmic competition can partially rescue the levels of epigenetic marks in 1PN biparental zygotes.

Low developmental ability of 1PN zygotes

We then investigated the developmental consequence of 1PN biparental zygotes. When cultured in vitro, 1PN biparental zygotes developed to the morula/blastocyst stage at a rate comparable with that of control 2PN zygotes (Fig. 5a and Extended Data Fig. 6a). Notably, however, when two-cell-stage embryos derived from 1PN biparental zygotes were transplanted, they resulted in a significantly reduced rate of full-term development compared with control 2PN zygotes (Fig. 5b,c). Offspring derived from 1PN zygotes showed no detectable sex ratio bias or body weight abnormalities, although there was a small increase in placental weight (Extended Data Fig. 6b,c). These results provide experimental evidence that 1PN biparental zygotes are not fully competent for embryonic development and are prone to loss after implantation.

Fig. 5. Rescue of the developmental potential of 1PN biparental zygotes.

a, Pre-implantation development. Proportions of embryos reaching the two-cell and morula/blastocyst stages at the days indicated. b,c, Manipulation-based rescue of full-term developmental potential. Proportions of embryos reaching the two-cell stage and developing to full term after transfer (b). Derived pups (c). d,e, TSA rescues H3K4me3. Representative images of H3K4me3 (green) and their signal intensity ratio (H3K4me3/H3, fire) in 1PN zygotes (d). H3K4me3/H3 ratios within interior and peri-NPB ROIs, normalized to control 2PN zygotes (e). f, Drug-based rescue of developmental potential. Proportions of embryos reaching the two-cell stage and developing to term after transfer. g,h, Kdm5b knockdown. Representative images of KDM5b signal (g). KDM5b signal intensities normalized to control zygotes (h). i,j, Kdm5b knockdown partially rescues H3K4me3. Representative images of H3K4me3 (green) and their signal intensity ratio (H3K4me3/H3, fire) in 1PN zygotes (i). H3K4me3/H3 ratios within interior and peri-NPB ROIs normalized to control 2PN zygotes (j). k, Kdm5b knockdown partially rescues developmental potential. Proportions of embryos reaching the two-cell stage and full-term development after transfer are shown. l, Summary, in normal 2PN zygotes, separate parental pronuclei compete for limited cytoplasmic factors, limiting the volume of each pronucleus and allowing proper regulation of epigenetic marks including trimethylated histones, to ensure full developmental potential. In 1PN zygotes, competition is absent, forming an enlarged pronucleus with dysregulated epigenetic marks and reduced developmental potential. Number of oocytes from three (a,i,j), seven (f), two (g,h) or at least four (k) independent experiments. Each dot, one experiment (a,b,f,k). Statistical tests, Fisher’s exact test (a,b,f,k); Mann–Whitney U test (e,h,j). Error bars, s.e.m. KD, knockdown. Scale bars, 100 μm (a); 5 μm (d,g,i).

Source data

Extended Data Fig. 6. Development of embryos derived from 1PN and P1PN zygotes.

(a) Pre-implantation development. In vitro development of embryos derived from 2PN (control), 1PN, and post-competition 1PN (P1PN) zygotes were monitored. Developmental stages categorized at the indicated days are shown. Number of oocytes from three independent experiments. (b) Offspring sex ratio derived from 1PN biparental zygotes. Number of pups from three independent experiments. (c) Body and placenta weights of offspring derived from 1PN biparental zygotes. Number of pups and placentas from fourteen independent experiments. Mann-Whitney U test was performed. Error bars, SEM. (d) Schematic diagram of RNA-seq. B6 oocytes and JF1 sperm were used for the production of zygotes. (e) Volcano plot showing 36 differentially expressed genes. Significantly differentially expressed genes (DEGs) in 1PN (FDR < 0.05) are indicated in blue. Differential expression analysis was performed using a two-sided exact test based on the negative binomial distribution implemented in edgeR. P values were adjusted for multiple comparisons using the false discovery rate (FDR) method (Storey’s q-value approach). (f) Validation of parental allele classification. 2-cell embryos derived from [B6 oocyte × JF1 sperm] and [JF1 oocyte × B6 sperm] zygotes were subjected to single-embryo RNA-seq. Embryos derived from parthenotes were used as a reference. The assigned read ratio (B6/JF1) is shown. Number of oocytes from three experiments. Error bars indicate SEM. Sequences detected in B6 parthenotes were predominantly assigned to B6, as were those in JF1 parthenotes to JF1, whereas those in zygotes were more evenly assigned. (g) Maternal-paternal ratios of each gene. Average CPM value ratio of each gene was plotted to compare 1PN and 2PN. The diagonal red line indicates equal amounts. Red dots indicate ZGA genes. Significantly differentially expressed genes (DEGs) in (e) are indicated in blue dots.

Source data

Comparison of late-two-cell-stage embryos derived from 1PN biparental zygotes with those from control 2PN zygotes using single-embryo RNA sequencing (RNA-seq) analysis (Extended Data Fig. 6d) revealed 36 differentially expressed genes (DEGs) (Extended Data Fig. 6e and Supplementary Table 1), including genes involved in zygotic gene activation (ZGA), such as Zscan4 and Obox4^36–38. Parental allele classification (Extended Data Fig. 6d,f) showed no systematic shift in the maternal-to-paternal ratio of RNAs including DEGs and ZGA genes (Extended Data Fig. 6g). These results suggest that 1PN biparental zygotes are partially defective in ZGA.

Rescue strategies of developmental potential

We then investigated whether our strategy to rescue the epigenetic defects of the biparental pronucleus improves the developmental potential of 1PN zygotes. When cultured in vitro, post-competition 1PN biparental zygotes developed to the blastocyst stage at a rate comparable with that of control 2PN zygotes (Fig. 5a). When two-cell embryos derived from post-competition 1PN biparental zygotes were transplanted, they developed to term at a significantly increased rate compared with competition-free 1PN biparental zygotes (Fig. 5b). Thus, a manipulation-based strategy that induces pronuclear-to-pronuclear competition improves the developmental potential of the biparental pronucleus.

This improvement was probably mediated by the rescue of epigenetic dysregulation. To investigate a causal link between the epigenetic and developmental defects of 1PN biparental zygotes, we used the histone deacetylase inhibitor Trichostatin A (TSA)³⁹, which has several epigenetic effects, including increased H3K4me3 levels in zygotes⁴⁰. As expected, treatment with TSA significantly increased H3K4me3 levels in the interior and peri-NPB regions of the maternal hemisphere of 1PN biparental zygotes (Fig. 5d,e). Notably, this treatment significantly rescued the ability of 1PN biparental zygotes to develop to term, whereas it did not significantly improve the development of 2PN zygotes (Fig. 5f). Furthermore, RNAi-mediated knockdown of the H3K4me3 demethylase KDM5b, which significantly increased the global level of maternal H3K4me3 in the interior and peri-NPB regions of the pronucleus (Extended Data Fig. 5g–j), reduced the ability of 2PN zygotes to develop to term^24,25, whereas it partially but significantly rescued the developmental ability of 1PN biparental zygotes (Fig. 5k). These results indicate that the developmental defects of 1PN biparental zygotes are due to several epigenetic defects, and demonstrate a drug-based strategy to improve the developmental potential of 1PN biparental zygotes.

Discussion

Taken together, these results demonstrate the physiological significance of the separation of the parental genomes into two distinct pronuclei in zygotes (Fig. 5l). The two-pronuclear state establishes competition between the parental pronuclei for a finite resource of the shared cytoplasm. This competition limits pronuclear volume, which is linked to the ability of zygotes to properly regulate several epigenetic marks, including H3K4me3 and H3K27me3. Competition-mediated regulation is required for zygotes to ensure developmental potential. Thus, the spatial separation of the parental genomes into distinct pronuclei is not simply a consequence of fertilization but plays an active role in embryonic development. However, a drawback of parental spatial separation is that it facilitates chromosome segregation errors during the first embryonic division^4–6.

The parental asymmetry in pronuclear volume (the paternal pronucleus is larger than the maternal one in mouse and human zygotes) is unlikely to be explained by parental competition per se, but may be attributable to faster recruitment of nuclear pore proteins to the paternal pronucleus (Fig. 3d–f). Molecular mechanisms linking limited pronuclear volume to epigenetic regulation should be investigated in future studies. One possibility is that competition-mediated restriction of pronuclear volume limits the intranuclear dilution of factors that regulate the global levels of epigenetic marks, such as those that protect against histone demethylation and those that promote methylation. Alternatively, parental pronuclei may compete for cytoplasmic factors that act to reduce the global levels of epigenetic marks, as they do for factors that lead to pronuclear enlargement. Candidate factors subject to competition may include proteins located at cytoplasmic lattices, which contain numerous chromatin modifiers and their regulators^41,42. In either scenario, this study demonstrates that pronuclear-to-pronuclear competition for the finite cytoplasmic resource is a cell biological mechanism for cytoplasm-mediated remote parental interaction that regulates epigenetic marks.

This healthy parental competition is lacking in 1PN biparental zygotes, resulting in dysregulation of epigenetic marks and thus reduced developmental potential (Fig. 5l). The partial rescue by KDM5b knockdown (Fig. 5k) supports the idea that the low developmental potential of 1PN zygotes is probably attributed to defects in regulation of several epigenetic marks, including those for trimethylated histones and others. Reductions in the levels of H3K4me3 and H3K27me3 in 1PN biparental zygotes were detected on maternal chromosomes by immunofluorescence quantification. Furthermore, CATCH-seq analysis detected clear defects in the establishment of H3K4me3 and H3K27me3 marks on the paternal genome in 1PN biparental zygotes. Treatment with TSA and transient induction of pronuclear-to-pronuclear competition probably had a broad effect on several epigenetic marks of both maternal and paternal chromosomes, resulting in the partial rescue of the developmental potential of 1PN biparental zygotes. Although these drug-based and manipulation-based strategies present approaches for rescuing 1PN biparental zygotes of mice, this study highlights the need for in-depth studies on human zygotes and important caveats in the use of 1PN zygotes for assisted reproductive technology.

Methods

Animals

All animal experiments conformed to the Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Committee of Laboratory Animal Experimentation of the RIKEN Centre for Biosystems Dynamics Research. C57BL/6 (B6), JF1/Ms (JF1), BDF1 (C57BL/6 × DBA/2) and R26R-H2B-EGFP^+/− (C57BL/6 × DBA/2 background)⁴³ mice, aged 8–10 weeks, were used to produce oocytes and sperm. Surrogate pseudopregnant females used as embryo transfer recipients (see ‘Embryo transfer’) were ICR strain mice mated with vasectomized males of the same strain. BDF1 mice were purchased from Japan SLC Inc.

Oocyte collection

Mature oocytes were collected from the oviducts of 8- to 10-week-old female mice that had been induced to superovulate with 5 IU of equine chorionic gonadotropin (ASKA Pharmaceutical) followed by 5 IU of human chorionic gonadotropin (ASKA Pharmaceutical) 48 h later. Cumulus-oocyte complexes were collected from the oviducts approximately 16 h after human chorionic gonadotropin injection. Cumulus-oocyte complexes were placed in M2 medium and treated with 0.1% (w/v) bovine testicular hyaluronidase. After several minutes, the cumulus-free oocytes were washed twice and then transferred to Chatot, Ziomek and Bavister medium (CZB). Mature MII oocytes were subjected to ICSI.

Micromanipulation

For cytoplasmic removal (Extended Data Fig. 3a,b), micromanipulation was performed as described previously⁴⁴ with some modifications. Briefly, oocytes at the MII stage were transferred to M2 medium supplemented with 5 μg ml⁻¹ cytochalasin B (Sigma-Aldrich) for 10 min. Then, the zona was cut using the LYKOS laser system (Hamilton Thorne) in a micromanipulation chamber placed on a warmed stage (37 °C) in an inverted microscope (Olympus). After cutting the zona, we held each oocyte with a holding pipette at the 9-o’clock position, and then rotated the oocyte until the hole of the zona was at the 3-o’clock position. After inserting a fire-polished injection pipette (inner diameter 30 µm) through the hole of the zona, we aspirated half the cytoplasmic volume and pinched off (Supplementary Video 3). The volume of the aspirated cytoplasm was controlled by an ocular micrometer. To obtain control oocytes, we aspirated half the total cytoplasmic volume and returned it to the oocyte.

To generate doubled oocytes (Extended Data Fig. 3a,b), we performed micromanipulation based on previously described methods⁴⁵ with some modifications. Briefly, oocytes at the MII stage were transferred to M2 medium supplemented with 5 μg ml⁻¹ cytochalasin B. A hole was made in the zona pellucida by applying several piezo pulses using an enucleation pipette. The MII spindle was aspirated into the pipette with a minimal volume of ooplasm. After enucleation, oocytes were cultured in CZB medium for at least 15 min for recovery. The enucleated oocytes and intact oocytes were treated with acid Tyrode’s solution (pH 2.1–2.5, Irvine Scientific) for 30 s to dissolve the zona. After several washes, the zona-free oocytes were cultured in CZB medium for at least 30 min, and then transferred to M2 medium containing 0.05 mg ml⁻¹ phytohaemagglutinin (Wako Pure Chemical Industries) for 1 min. The oocytes were transferred to M2 medium on a micromanipulation chamber, in which an intact oocyte and an enucleated oocyte were attached to each other by micropipettes. The attached oocytes were cultured in CZB medium for at least 15 min, washed three times with fusion medium (0.3 M mannitol, 0.1 mM MgSO₄ and 0.1% polyvinyl alcohol) then transferred to the same solution placed between parallel electrodes and separated by 0.5 mm in a chamber. Current was applied from an electric cell fuser (LF101; BEX Co.) (15 V alternate current at 1 MHz for 1 s to line up the attached oocytes, 30 V direct current pulse for 20 µsec to induce cell fusion). The oocytes were washed and then cultured in CZB medium for 30 min for recovery and cell rounding (Supplementary Video 4).

Fertilization

IVF was performed according to the manufacturer’s instructions using CARD MEDIUM (Cosmo Bio). To limit the time of fertilization, we set the insemination time to 1 h.

ICSI with sperm heads was carried out as described previously⁴⁶. Briefly, the sperm head was separated from the tail by applying several piezo pulses to the neck region, and then injected into an oocyte. After 20 min of recovery at room temperature, injected oocytes were cultured in CZB.

To generate 1PN biparental zygotes, we injected the sperm head around the midpoint between the maternal spindle and the oocyte centre. Placing the sperm head away from the plasma membrane helped to prevent its extrusion into the polar body. The sperm head was still extruded occasionally into the polar body, resulting in the production of haploid parthenotes, which were distinguished by the immunostaining patterns of the polar body and pronucleus and excluded from further analysis. In rescue experiments, zygotes were treated with 50 nM TSA for 10 h after ICSI.

Parthenogenetic and androgenetic zygotes

To generate haploid parthenogenetic embryos, we activated oocytes using CZB supplemented with 10 mM SrCl₂, 2 mM EGTA for 10 h. To generate androgenetic embryos, a sperm head was injected into the enucleated MII oocytes. Injected oocytes were cultured in CZB medium for 30 min for recovery, and then cultured in activation medium.

Embryo transfer

Embryo transfer was carried out as described previously⁴⁷. Embryos at the two-cell stage were transferred at 0.5 days post-coitum into the oviducts of pseudopregnant female mice of the ICR strain that had been mated with a vasectomized male the night before transfer. At 18.5 days post-coitum, the offspring were delivered by Caesarean section because the mother sometimes eats the pups. Mice without implantation marks were excluded from the data. Surviving pups were fostered by an ICR foster mother that had given birth on the same day.

Live cell imaging

After linearization of the template plasmid, mRNA was synthesized using the mMESSAGE mMACHINE KIT (Ambion). Synthesized RNAs were stored at −80 °C until use. In-vitro-transcribed mRNAs (2.0 pl of 1,000 ng µl⁻¹ EGFP-Nup153, 2.0 pl of 850 ng µl⁻¹ importin β-binding domain of importin α (IBB)-DiHcRed and 1.2 pl of 1,000 ng µl⁻¹ mCherry-methyl CpG binding domain (MBD)-nuclear localization signal (NLS)) were microinjected into MII oocytes. IBB-DiHcRed and mCherry-MBD-NLS serve as markers for nuclear import⁴⁸ and post-fertilization paternal chromosomes⁴⁹, respectively. The microinjected oocytes were cultured for 1 h, and then ICSI or IVF was performed.

Live cell imaging was performed as described previously³² with some modifications. Briefly, a Zeiss LSM710 or LSM880 confocal microscope, operated using ZEN software (ZEN 2011 SP3 (black) v.8.1.11.484 for LSM710 and ZEN 2.3 SP1 FP1 (black) v.14.0.13.201/v.14.0.9.201 for LSM880) and equipped with a ×40 C-Apochromat 1.2 numerical aperture water immersion objective lens (Carl Zeiss) was controlled by a multi-position autofocus macro⁵⁰. For imaging, 29 confocal z-sections (every 3.0 µm) of 512 × 512 pixel xy images covering a total volume of 84.85 × 84.85 × 84.00 µm were acquired at 5-min intervals for at least 14 h after fertilization.

Kdm5b knockdown

To knock down Kdm5b, 2.5 pl of 5 nM Kdm5b (catalogue no. 4390771, product ID s93702) and negative control (catalogue no. 4390843) Silencer Select siRNAs (Thermo Fisher Scientific) were microinjected into MII oocytes. The microinjected oocytes were cultured for 1 h then used for ICSI.

Three-dimensional analysis

To measure pronuclear volume (Fig. 3 and Extended Data Fig. 4), we performed three-dimensional surface rendering with the signals of mCherry-MBD-NLS or H2B-GFP with Imaris software (Bitplane, v.7.4.2). The volume was obtained after ellipsoid fitting.

Immunostaining

Zygotes were fixed 10 h after fertilization (PN5 stage) in 2% paraformaldehyde in PBS-polyvinyl alcohol (PVA) (pH 7.4) for 30 min. For the 5hmC and 5mC staining, fixed zygotes were incubated in 4 N HCl solution at room temperature for 10 min, followed by neutralization (1 M Tris-HCl, pH 8.0) for 20 min. After blocking and permeabilization in PBS-PVA containing 1 mg ml⁻¹ bovine serum albumin (BSA) (PBS-PVA-BSA) and 0.1% Triton X-100, oocytes were incubated with appropriate primary antibodies overnight at 4 °C, washed several times in PBS-PVA-BSA and incubated with secondary antibodies for 120 min at room temperature. DNA was counterstained with 40 µg ml⁻¹ of Hoechst 33342. Finally, the oocytes were washed and transferred to BSA-PVA for imaging with a Zeiss LSM780 confocal microscope, operated using ZEN software (ZEN 2011 SP3 (black) v8.1.11.484). Images were reconstructed in three dimensions with Imaris software.

The following primary antibodies were used: rabbit anti-histone H3K4me3 (1:200, Upstate Biotechnology, catalogue no. 07-473), rabbit anti-histone H3K9me3 (1:200, Upstate Biotechnology, catalogue no. 07-442), rabbit anti-histone H3K27me3 (1:200, Upstate Biotechnology, catalogue no. 07-449), rabbit anti-histone H3K27ac (1:200, Abcam, catalogue no. ab177178), mouse anti-histone H3 (1:200, Abcam, catalogue no. ab195277), mouse anti-nuclear pore complex (mab414) (1:200, Abcam, catalogue no. ab24609), rat anti-NUP98 (1:800, Abcam, catalogue no. ab50610), rabbit anti-5hmC (1:500, Active Motif, catalogue no. 39769), mouse anti-5mC (1:500, Eurogentec, catalogue no. BI-MECY-0100) and rabbit anti-KDM5b (1:200, Abcam, catalogue no. ab181089) antibodies. The secondary antibodies were Alexa Fluor 488 Goat Anti-Mouse IgG (H+L) (catalogue no. A11029), Alexa Fluor 488 Goat Anti-Rabbit IgG (H+L) (catalogue no. A11034), Alexa Fluor 488 Goat Anti-Rat IgG (H+L) (catalogue no. A11006) and Alexa Fluor 555 Goat Anti-Mouse IgG (H+L) (catalogue no. A21424) (1:400, Molecular Probes).

Quantification of fluorescence signals

Total fluorescence intensity throughout the entire pronucleus was measured using Imaris. The pronuclear volume was defined based on H3 signals in three dimensions, and the total signal intensity of the modified histone within the defined volume was obtained. Fluorescence intensity of a cytoplasmic region of equal volume was measured and subtracted to yield corrected total pronuclear intensity values.

To quantify the relative levels of H3K4me3, H3K27me3 or H3K27ac to histone H3, the mean signal intensity for each modified histone (I_{me_pro}) within ROIs in two defined categories: (1) the pronuclear interior away from the NPB and (2) the peri-NPB region. NPBs were defined as spherical areas of depleted H3 signal within the pronucleus. ROIs for the interior regions were placed more than 1 µm away from both the pronuclear and NPB rims and from the estimated maternal–paternal boundary within the maternal pronucleus or hemisphere. ROIs for the peri-NPB regions were generated using the ‘Make band’ (band = 1) function in Fiji (v.1.54p) after delineating the periphery of the NPB manually within the maternal pronucleus or hemisphere. For comparisons of fluorescence levels across samples, ROIs of identical size were applied consistently: fixed at 3.82 µm × 3.82 µm under normal conditions and adjusted to 3.32 µm × 3.32 µm for smaller pronuclei (for example, halved zygotes). The mean cytoplasmic signal intensity (I_{me_cyto}) from a region near the pronuclei was subtracted (I_{me_pro}− I_{me_cyto}). Similarly, the histone H3 intensity was determined as (I_{H3_pro} − I_{H3_cyto}). The ratio between the two values, (I_{me_pro} − I_{me_cyto})/(I_{H3_pro} − I_{H3_cyto}) was then calculated.

Single-embryo RNA-seq

C57BL/6 × JF1/Ms zygotes were selected 10 h after fertilization (late G2 or PN5 stage) and the second polar body was removed using a micromanipulator. Two-cell embryos were collected 34 h after fertilization. Single-embryo RNA-seq libraries were prepared with SMART-Seq v.4 Ultra Low Input RNA Kit (Takara Bio) and Nextera XT DNA Library Preparation Kit (Illumina).

The RNA-seq reads were filtered using fastp (v.0.20.0)⁵¹ to remove Illumina adaptor sequences, and to eliminate low-quality bases with the ‘-3 -q 15 -l 15’ options. The filtered reads were mapped to the mouse genome (mm10) using Hisat2 with default parameters except for the ‘--sp 1000,1000’ option. The featureCounts v.2.0.0⁵² was used to generate counts of reads aligned to annotated genes in the Ensembl GRCm38.99 with options ‘-p -O–fraction’.

Before conducting subsequent downstream analyses, samples were subjected to stringent quality filtering using the following threshold: embryos with a number of detected genes greater than 10,000, and a ratio of mitochondrial genes less than 0.2 were retained. This filtering excluded three samples. Samples were also subjected to parental allele classification (‘Parental allele classification in bioinformatics analysis’) to exclude parthenotes (paternal allele ratio of less than 0.1) that could not be distinguished from 1PN biparental zygotes by microscopic observation.

The amount of poly(A) RNA for each embryo was estimated from the inferred linear regression parameters based on the ratio of spike-in derived reads using the ‘relative2abs’ function implemented in the monocle package⁵³. In each experiment, embryos derived from 2PN zygotes were used as a reference for data normalization. Differentially expressed genes (false discovery rate (FDR) < 0.05) between the PN types were explored using the edgeR (v.3.40.2)⁵⁴ with the trimmed mean of M values normalization, using the dataset of genes expressed in at least ten samples.

Carrier-assisted chromatin immunoprecipitation followed by sequencing (CATCH-seq)

CATCH-seq is an improved method of the ultra-low-input native chromatin immunoprecipitation followed by sequencing (ChIP-seq) (ultra-low-input micrococcal nuclease-based native ChIP (ULI-NChIP))⁵⁵; its original protocol was described in ref. ²⁸ and further optimized in ref. ²⁷. BDF1(C57BL/6×DBA2) × JF1/Ms zygotes were selected 10 h after fertilization (late G2 or PN5 stage) and the second polar body was removed using a micromanipulator. Zygotes were collected in 9 µl of Nuclei EZ lysis buffer (Sigma, NUC-101) supplemented with a complete EDTA-free protease inhibitor cocktail and 1 mM phenylmethanesulfonyl fluoride. The samples were snap-frozen in liquid nitrogen and stored at −80 °C until use. On the day of library construction, the samples were thawed on ice and added with 1 µl of the lysis buffer containing a constant number of Drosophila melanogaster S2 cells (Thermo Fisher Scientific, catalogue no. R69007) for spike-in normalization purposes. Then, 1 µl of a 1% Triton X-100 (Merck, catalogue no. 93443) and 1% deoxycholate (Nacalai, catalogue no. 10712-54) mixture solution was added to the samples, which then sat on ice for 5 min. The chromatin was fragmented by 2 U µl⁻¹ micrococcal nuclease (MNase) (catalogue no. M0247S, NEB) in 1× MNase buffer supplemented with 1% PEG6000 (Hampton Research, catalogue no. HR2-533) and 2 mM dithiothreitol (Nacalai) at 21 °C for 7.5 min. The MNase reaction was stopped by adding 1/10 volume of 100 mM EDTA and 1/12 volume of the 1% Triton X-100 and 1% deoxycholate mixture, then the samples were rested on ice for 15 min. The chromatin lysates were then added to freshly prepared immunoprecipitation buffer⁵⁵, and 5% volume was kept for input library construction. To the rest of the lysate, 30 ng of annealed I-SceI carrier DNA was added²⁷. The forward and reverse strands of the carrier DNA were as follows: /5AmMC6/Gtagggataacagggtaattagggataacagggtaattagggataacagggtaattagggataacagggtaattagggataacagggtaattagggat aacagggtaat*c/3AmMO/ and /5AmMC6/Gattaccctgttatccctaattaccctgttatccctaattaccctgttatccctaattaccctgttatccctaattaccctgttatccctaattaccctgttatcc cta*c/3AmMO/, respectively, where asterisks represent phosphorothioate bonds. The oligonucleotides were synthesized by Integrated DNA Technologies. For each immunoprecipitation reaction, 0.25 µl of rabbit anti-H3K4me3 (Active Motif, catalogue no. 39159), 500 ng of rabbit anti-H3K27me3 (Diagenode, catalogue no. C15410069) conjugated to precleared Dynabeads Protein A (Thermo Fisher Scientific, catalogue no. 10006D) and G (Thermo Fisher Scientific, catalogue no. 10007D) mixture was used. The specificity of these antibodies was validated in previous studies (H3K4me3 (ref. ⁵⁶), H3K27me3 (ref. ⁵⁷)). After immunoprecipitation at 4 °C overnight, the chromatin-Dynabeads were washed twice each with low and high salt wash buffers, and the chromatin was eluted in the freshly prepared ChIP elution buffer at 65 °C for 1 h⁵⁵. DNA was recovered by ChIP DNA Clean and Concentrator Kit (Zymo Research). Adaptor ligation was performed by NEBNext Ultra II DNA Library Prep Kit for Illumina (catalogue no. E7645, NEB) in a half scale of the manufacturer’s instruction, and the libraries were purified by 1.8× SPRIselect beads (catalogue no. B23318, Beckman Coulter). The DNA was amplified by KAPA Hifi 2× mater PCR mix (catalogue no. KK2605) for 13–15 PCR cycles with dual indexing primers (NEBNext Multiplex Oligos for Illumina, catalogue no. E6440). After purification with 0.9× SPRIselect beads, the samples were digested by I-SceI (5 U µl⁻¹, NEB, catalogue no. R0694) at 37 °C for 2 h followed by heat inactivation at 65 °C for 20 min and purified by 0.9× SPRIselect beads. The second amplification was not performed. The libraries were sequenced on Nextseq500 (single-end reads) or Nextseq2000 (paired-end reads) (Illumina).

CATCH-seq data analysis

For CATCH-seq, all sequencing reads after trimming by fastp were aligned to the mouse (mm10) using Bowtie 2 v.2.3 (ref. ⁵⁸) with ‘-N 1 -L 25’ options. For allele-specific analysis of CATCH-seq in the BDF1 × JF1 F1 hybrid embryos, parental allele classification was described below (‘Parental allele classification in bioinformatics analysis)’. For visualization of CATCH-seq using the Integrative Genomics Viewer (v.2.16.1)⁵⁹, genome coverage tracks were generated using bamCoverage from deepTools (v.3.5.1)⁶⁰ with the parameters ‘–binSize 50–scaleFactor 1–normalizeUsing RPKM–numberOfProcessors 28–extendReads 200–ignoreDuplicates–smoothLength 100’. SNPsplit (v.0.3.4)⁶¹ was applied to assign uniquely aligned reads to their parental origins. To compare replicates, read counts over each bin (10 kb) across the whole genome were calculated and the reads per million values for each bin were used to calculate the Spearman correlation coefficient. Averaged reads per kilobase million (RPKM) between replicates was calculated using bigwigAverage from the deepTools suite (v.3.5.3). To compare correlation between CATCH-seq and STAR ChIP-seq data²³, read counts over each bin (10 kb) across the whole genome were calculated and the reads per million values for each bin were used to calculate the Spearman correlation coefficient using deepTools suite and visualized using the R function pheatmap.

For analyse the H3K27me3 allelic-biased region analysis, using the H3K27me3 ChIP-seq datasets of mourla stage embryos⁵⁷, the 10-kb bins (no slide) were divided into ‘K27me3 Mat-biased’ and ‘K27me3 Pat-biased’ groups. Each 10-kb bin contained more than 50 SNPs. The average RPKM of H3K27me3 in each bin was calculated using ComputeMatrix in deeptools (v.3.5.1). Bins with RPKM > 2 were used as positive regions. Then, the allelic ratio of average RPKM in each group’s bin was calculated and fold change > 2 defined as allelic-biased regions.

Definition of H3K4me3 broad peaks

Previous H3K4me3 ChIP-seq data in fully grown oocytes⁶² were aligned to mm10 and the data were processed as described above.

H3K4me3 peaks were called MACS2 (v.2.2.9.1)⁶³ with the parameters ‘-g mm --nomodel --nolambda --broad’. To identify broad H3K4me3 domains, those with a distance shorter than 5 kb were merged using the ‘merge’ function from BEDTools (v.2.26.0), as described previously²⁴.

Parental allele classification in bioinformatics analysis

To distinguish the parental alleles of the read sequences of BDF1 × JF1/Ms and C57BL/6 × JF1/Ms embryos, diagnostic SNPs (dSNPs) were developed using the deposited deep sequence data obtained from JF1/Ms⁶⁴. After mapping the JF1/Ms reads to the mouse genome (mm10) using BWA-MEM v.0.7 with the default parameters, SNPs were called using HaplotypeCaller implemented in GATK v.4.1.4.1 (ref. ⁶⁵). After filtering out low-confidence SNPs with ‘QD < 2.0, QUAL < 30.0, SOR > 3.0, FS > 60.0, MQ < 40.0, MQRankSum < −12.5, and ReadPosRankSum < −8.0’, differently fixed 20,885,964 sites were used as dSNPs. For the reference of allele-specific analysis, dSNP sites on the mouse genome (mm10) were masked as N. Each aligned read (see ‘Single-embryo RNA-seq’ and ‘CATCH-seq data analysis’ above) was subjected to assign its parental origin using the dSNPs with SNPsplit v.0.3.4 (ref. ⁶¹).

Preparation of heavy-labelled peptides by cell-free synthesis for MS-QBiC

Stable isotope-labelled peptides were synthesized using an Escherichia coli reconstituted cell-free protein synthesis system, PURE system, containing ¹³C₆ ¹⁵N₄ l-arginine and ¹³C₆ l-lysine (Thermo Scientific). Details of the synthesis procedure are described previously⁶⁶ with slight modifications. In brief, DNA templates for the PURE system were prepared by a two-step PCR using Taq DNA polymerase (PCR Master Mix, Promega) with forward primers containing the T7 promoter and the appropriate reverse primers listed in Supplementary Table 2. The amino acid sequences of the synthesized peptides are shown in Supplementary Table 2. Subsequently, synthetic peptides underwent reductive alkylation, trypsin digestion, and desalting using an in-house-made C18 stage-tip. The peptides were then dried with a Speedvac, dissolved in 0.1% trifluoroacetic acid/2% acetonitrile, and used for mass spectrometry.

Liquid chromatography–tandem mass spectrometry experiments

To perform liquid chromatography–tandem mass spectrometry (LC-MS/MS) analysis, pronuclei were isolated by micromanipulator (Supplementary Video 5). Approximately 170 maternal and paternal pronuclei were pooled for the analysis. Biological replicates were performed in triplicate. Each pronucleus was suspended in 50 mM ammonium bicarbonate and reduced and alkylated with tris(2-carboxyethyl)phosphine and iodoacetamide. Trypsin digestion was performed using the SP3 method⁶⁷. For mass spectrometry, the tryptic digests of pronuclei were mixed with the heavy-labelled peptides and iRT-kit 11 peptides (Biognosys AG) as a retention time marker.

Mass spectra were obtained with an Orbitrap Eclipse (Thermo Fisher Scientific) coupled to a nanoflow UHPLC system (Vanquish; Thermo Fisher Scientific). The peptide mixtures were loaded onto a C18 trap column (PepMap Neo Trap Cartridge, ID 0.3 mm × 5 mm, particle size 5 μm, Thermo Fisher Scientific) and then fractionated by the C18 analytical column (Aurora, ID 0.075 × 250 mm, particle size 1.7 μm, IonOpticks). The peptides were eluted at a flow rate of 300 nl min⁻¹ using the following gradient, with the percentage of solvent B indicated: a linear gradient from 0% to 2% over 1 min, followed by a linear increase from 2% to 5% over 2 min, a linear gradient from 5% to 16% over 58.5 min, another linear gradient from 16% to 25% over 22.5 min, a linear gradient from 25% to 35% over 13 min, a sharp increase to 95% over 1 min, holding at 95% for 5 min and finally re-equilibration at 5%. The composition of solvent A and solvent B is 100% H₂O, 0.1% formic acid and 100% acetonitrile, 0.1% formic acid, respectively. The Orbitrap was operated in a data-dependent mode with a cycle time of 3 s. The MS1 scan was collected at 120,000 resolution with the mass range for 335–1,500 m/z using a standard automated gain control. Intensity threshold of 20,000 and charge states 2–7 were taken for MS/MS starting from high intensity precursors. Quadrupole was used for precursor isolation with 1.6 m/z isolation width and fragmented with normalized higher energy dissociation energy of 30% and resulting fragment ions were recorded in Orbitrap analyser. The MS2 scan was collected at 30,000 resolution using a standard automated gain control target and maximum injection time of 54 ms. Dynamic exclusion was set to 20 s.

The raw data files were searched against the Mus musculus dataset (Uniprot Proteome UP000005640, 20,230,207 downloaded) with the common Repository of Adventitious Proteins (cRAP, ftp://ftp.thegpm.org/fasta/cRAP) to recognize the contaminant proteins. The search was conducted through MASCOT v.2.8 (Matrix Science) using Proteome discoverer 2.5 (Thermo Fisher Scientific) with the FDR for peptide and protein identification set at 1%. The following settings were used for search parameters: enzyme, trypsin; FDR, 0.01; precursor mass tolerance, 10 ppm; product tolerance, 0.02 Da; fixed modification, carbamidomethylation of cysteine; variable modification, oxidation of methionine and acetylation of protein N-termini; maximum missed cleavages, 2. The peak area calculation was performed using Skyline (v.23.1.0.380) (MacCoss Lab Software)⁶⁸ with setting MS1 filter to a count of three (M, M + 1 and M + 2). Extracted ion chromatograms processed using Savitzky-Golay filters provided by Skyline software are presented.

Theoretical model

We introduced our theoretical model for pronucleus growth controlled by a limiting factor, assuming pronucleus pores are active channels that selectively transport the limiting factor from the cytoplasm into pronucleus. The limiting factor transportation is associated with the volume flux, the rate of which is assumed to be dependent linearly on the number of the pronucleus pores. We assume that the number of pores $A_i$ of pronucleus $i$ increases with the size of pronucleus as $A_i=a_i{V}_i^{2/3}$, where $a_i$ is the production rate of the pronucleus pores and $V_i$ is the volume of the pronucleus and the exponent $2/3$ converts the volume to the membrane area. Then, the time-evolution equation of the number of the limiting factor is given by

$$\frac{\mathrm{d}\left(V_i{n}_i\right)}{\mathrm{d}t}=k{A}_i{n}_{\mathrm{cyto}}$$ 1

where $k$ is the transportation rate of the limiting factor through a pronucleus pore and $n_{\mathrm{cyto}}$ is the density of the limiting factor in cytoplasm. As the limiting factors are transported to pronuclei, that in the cytoplasm decreases as

$$\frac{\mathrm{d}\left(V_{\mathrm{cyto}}{n}_{\mathrm{cyto}}\right)}{\mathrm{d}t}=-\sum _{i}k{A}_i{n}_{\mathrm{cyto}}$$ 2

where $V_{\mathrm{cyto}}$ is the volume of the cytoplasm and the summation is calculated over the pronuclei that share the cytoplasmic limiting factor. By writing the volume flux rate associated with the transportation of a limiting factor as $v$, the volume of the pronucleus $i$ and the cytoplasm changes as

$$\frac{\mathrm{d}{V}_i}{\mathrm{d}t}=v{A}_i{n}_{\mathrm{cyto}},$$ 3

$$\frac{\mathrm{d}{V}_{\mathrm{cyto}}}{\mathrm{d}t}=-\sum _{i}v{A}_i{n}_{\mathrm{cyto}}$$ 4

where the summation is calculated over the pronuclei. Here we assume that the total number of the limiting factor and the total volume are conserved in zygote; ${\sum }_i{n}_i{V}_i+n_{\mathrm{cyto}}{V}_{\mathrm{cyto}}=N_{\mathrm{total}}$ and ${\sum }_i{V}_i+V_{\mathrm{cyto}}=V_{\mathrm{total}}$. By applying these conservation rules to equations (1) and (2), the time-evolution equations of the concentration of the limiting factor become

$$\frac{\mathrm{d}{n}_i}{\mathrm{d}t}=\left(k-v{n}_i\right)a_i{V}_i^{-\frac{1}{3}}{n}_{\mathrm{cyto}}$$ 5

$$\frac{\mathrm{d}{n}_{\mathrm{cyto}}}{\mathrm{d}t}=-\sum _i\left(k-v{n}_{\mathrm{cyto}}\right)a_i{V}_i^{\frac{2}{3}}\frac{1}{V_{\mathrm{cyto}}}{n}_{\mathrm{cyto}}$$ 6

To solve equations (3)–(6) numerically, we fixed the parameters as $k=5$ and $v=0.1$, and the time increment is set to $\mathrm{d}t=0.01$. The number of pronuclei and the pore production rate $a_i$ are varied to reproduce the experimental conditions. The maternal and paternal pronuclei are distinguished by the pore production rate: $a_{\mathrm{mat}}=100$ for the maternal pronucleus and $a_{\mathit{pat}}=150$ for the paternal pronucleus. For the control case, the total volume and limiting factor are set as $V_{\mathrm{total}}=100$ and $N_{\mathrm{total}}=1$, and the initial condition is set as $V_{\mathrm{mat}}\left(t=0\right)=V_{\mathrm{pat}}\left(t=0\right)=$$0.1,n_{\mathrm{mat}}\left(t=0\right)=n_{\mathrm{pat}}\left(t=0\right)=0$, $V_{\mathrm{cyto}}\left(t=0\right)=V_{\mathrm{total}}-V_{\mathrm{mat}}\left(t=0\right)-V_{\mathrm{pat}}\left(t=0\right)$ and $n_{\mathrm{cyto}}\left(t=0\right)=N_{\mathrm{total}}/V_{\mathrm{cyto}}\left(t=0\right)$. For the parthenote, the paternal pronucleus is omitted, and as a consequence the initial condition of the cytoplasmic volume is modified to $V_{\mathrm{cyto}}\left(t=0\right)=V_{\mathrm{total}}-V_{\mathrm{mat}}\left(t=0\right)$. For the doubled and halved zygotes, the total volume and limiting factor are replaced by $V_{\mathrm{total}}=200$ and $N_{\mathrm{total}}=2$ and by $V_{\mathrm{total}}=50$ and $N_{\mathrm{total}}=1/2$, respectively.

Statistical analysis

Statistical analyses were performed using GraphPad Prism (v.7.02). Statistical methods used are described in the figure legends. Two-sided tests were performed for all analyses, except for Extended Data Fig. 2h where one-sided Fisher’s exact test was used. No statistical methods were used to predetermine sample size. Oocytes collected from several donor mice were pooled and allocated randomly to experimental groups. Investigators were not blinded to group allocation during experiments or analysis.

Reporting summary

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

Online content

Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at 10.1038/s41586-026-10417-7.

Author contributions

H.K and T.S.K. conceived the study. H.K. performed most experiments. M.T. and T.S. performed theoretical modelling. M.M., M.K. and A.I. performed CATCH-seq and analysed the data. T.M., A.H. and Y.O. helped to analyse sequence data. R.N. and Y.S. performed MS-QBiC analysis. T.S.K. supervised the project. T.S.K., A.I. and H.K. wrote the manuscript with input from all authors.

Peer review

Peer review information

Nature thanks Gavin Kelsey and the other, anonymous, reviewers for their contribution to the peer review of this work. Peer reviewer reports are available.

Data availability

All RNA-seq and CATCH-seq datasets have been deposited at GEO: GSE279786 is publicly available. Previously published data were downloaded from GEO: H3K4me3 ChIP-seq data (GSE71434)²³; H3K27me3 ChIP-seq data (GSE76687)²⁹. All image datasets are available in SSBD with the identifier 10.24631/ssbd.repos.2026.02.492 (ref. ⁶⁹). Source data for all main and extended data figures are provided as separate Excel files with this study. All mass spectrometry datasets: The accession numbers are PXD049709 for ProteomeXchange and JPST002943 for jPOST. All data necessary to evaluate the conclusions of this study are included in the paper or deposited in public repositories. Additional materials are available from the corresponding author upon reasonable request. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Extended data

is available for this paper at 10.1038/s41586-026-10417-7.

Supplementary information

The online version contains supplementary material available at 10.1038/s41586-026-10417-7.