Removal of remodeling/reprogramming factors from oocytes and the impact on the full-term development of cloned embryos

The first mammalian clone, Dolly the sheep, was born more than 20 years ago; however, the success rate of animal cloning remains very low (Matoba and Zhang, 2018; Wilmut et al., 1997). This may be because of incomplete reprogramming of the donor nucleus by the oocyte cytoplasm. For example, a number of epigenetic modifications remain after nuclear transfer, which are termed reprogramming resistant regions (Matoba et al., 2014; Matoba and Zhang, 2018; Terashita et al., 2013), and these lead to the low success rate of animal cloning.

During animal cloning, the metaphase II (MII) spindle of the oocyte is removed before somatic cell nuclear transfer (‘enucleation’), after which a donor nucleus is fused/injected into an enucleated oocyte (Wakayama et al., 2019, 1998; Wilmut et al., 1997). These fused/injected somatic cell nuclei are remodeled, and form a newly assembled spindle using oocyte-remodeling factors. At the same time, the somatic cell nucleus is converted from a differentiated state to a totipotent state using oocyte-reprogramming factors. During enucleation, several important factors are removed from oocytes. For example, following the enucleation of bovine or mouse oocytes, the amounts of α-tubulin (Saraiva et al., 2015) or γ-tubulin (Van Thuan et al., 2006b) in the oocyte cytoplasm are reduced by more than half. Some kinds of histone deacetylases (HDACs), essential factors for embryo development, are probably removed from oocytes because they are colocalized with the spindle (Li et al., 2017). After nuclear transfer, nuclear-mitotic apparatus (NuMA), a matrix protein responsible for spindle pole assembly, and HSET (also known as KIFC1), a mitotic kinesin motor (Blangy et al., 1995; Mountain et al., 1999), are not detected in somatic-derived spindles of primate oocytes (Simerly et al., 2003). These results suggest that remodeling/reprogramming factors (RRFs) are also lost from oocytes during enucleation (Simerly et al., 2003), which may lead to incomplete remodeling/reprogramming or epigenetic abnormalities in the donor nucleus.

Conversely, the cytoplasm of oocytes contains surplus factors. For example, oocytes can accept up to four spermatozoa and form four male pronuclei in each oocyte (Clarke and Masui, 1987), or a quarter-size oocyte can accept only one spermatozoon (Tarkowski and Balakier, 1980; Wakayama and Yanagimachi, 1998). Although quarter-sized oocytes fail to develop after fertilization, half-sized oocytes can develop to full term without a reduction in the birth rate (Wakayama and Yanagimachi, 1998). If RRFs are similar to these factors, then it can be said that surplus RRFs also exist in each oocyte. Furthermore, inhibition of an HDAC with an inhibitor such as trichostatin A (TSA) did not decrease, but rather affected positively, the success rate of animal cloning (Kishigami et al., 2006a). Although it is not clear whether HDACs are RRFs, this finding supports the notion that surplus RRFs exist within the oocyte cytoplasm.

Controlling the amounts of RRFs in oocytes may help to improve our understanding of reprogramming, and subsequently increase the success rate of animal cloning. To date, several types of inhibitors have been used for nuclear transfer experiments, including cytochalasin B or latrunculin A (cytokinesis inhibitors) (Terashita et al., 2012; Wakayama and Yanagimachi, 2001), cycloheximide (protein synthesis inhibitor) (Liu et al., 1998), 6-dimethyl-aminopurine (protein phosphorylation inhibitor) (Susko-Parrish et al., 1994), and TSA or scriptaid (HDAC inhibitors) (Kishigami et al., 2006a; Van Thuan et al., 2009), which affect a wide range of factors and improve the success rate of cloning. Conversely, specific epigenetic abnormalities could be targeted via the injection of RNA, such as Xist siRNA or Kdm4 mRNA (Matoba et al., 2018). Although such RNA treatments could enhance the success rate of animal cloning, they can only be used for known factors. In addition, these inhibitors and reagents can damage the embryos if used at high concentrations or over a long period (Kishigami et al., 2006b). Thus, although it is important to manipulate RRFs to improve the success rate of animal cloning, currently no method is able to control the volume of RRFs without damaging the oocytes.

In this study, we developed a novel method for removing RRFs from oocytes by injecting somatic cell nuclei, which were subsequently removed from oocytes following the consumption of RRFs. Using this method, we observed an immediate decrease in the size of the oocyte's MII-spindle, suggesting a decrease in the levels of RRFs. This allowed us to examine the relationship between somatic cell reprogramming and the amount of RRFs in oocytes.

In the first series of experiments, we injected one to three somatic (cumulus) cell nuclei into oocytes and examined the size of the MII-spindle 6 h after injection by observations and live-cell imaging (Fig. 1A-D). The injected somatic-derived nuclei formed an irregular-shaped spindle [somatic cell nucleus-derived spindle (S-spindle)] with misaligned chromosomes, irrespective of the number of somatic cells injected, which was clearly distinguishable from the MII-spindle (Fig. 1D). The scattered chromosomes in the S-spindle were inevitable, because the spindle was formed from the cumulus cell nucleus at the G0/G1 stage; therefore, those chromosomes were unable to align on the equatorial region (Wakayama et al., 1998). The size of the MII-spindle decreased as the number of injected somatic cell nuclei increased (intact MII: 275 µm; injection of one cell: 197 µm; injection of two cells: 119 µm; injection of three cells: 87 µm) (Fig. 1E; Fig. S1; Table S1). Observation of the MII-spindle by live-cell imaging revealed that it began to reduce in size immediately after somatic nuclear injection, with ∼50% reduction within 2 h, irrespective of the number of injected somatic cell nuclei (Fig. 1F; Table S2; Movie 1). The γ-tubulin volume of the MII-spindle was determined by immunostaining. As shown in Fig. 1G,H, the γ-tubulin volume in the MII-spindle was also reduced in somatic cell-injected oocytes compared with that in intact oocytes (Fig. S2; Table S3).

The MII-spindle may shrink owing to several factors, such as kinases, phosphatases or chromatin modifiers derived from introduced somatic cytoplasm at the time of somatic nuclear injection rather than a reduction in RRFs. To avoid this possibility, first, we collected somatic cytoplasm from three or four somatic cells (Fig. S3A), which was then injected into intact oocytes, and the size of the MII-spindle was observed 4 h after injection (Fig. 2A). There was no significant change in the size of the MII-spindle between cytoplasm-injected oocytes and mocked control (PVP-injected) oocytes (Fig. 2F; Fig. S3F,G; Table S4). Next, we isolated denuded nuclei from brain (Fig. 2A; Fig. S3B-E) (Wakayama et al., 2008). Although those nuclei carry chromatin-associated proteins, they do not contain any cytoplasmic factors (Fig. S3B,C), and one or two nuclei were injected into oocytes, and then the size of the MII-spindle was observed 4 h after injection. The size of the MII-spindle significantly decreased with the increasing number of injected somatic cell nuclei (mocked control MII: 272 µm; injection of one cell: 188 µm; injection of two cells: 151 µm) (Fig. 2G; Fig. S3H-J; Table S5). Thus, although we cannot avoid the possibility that the chromatin-associated proteins of injected somatic nuclei caused this shrinking MII-spindle, the cytoplasmic factors did not affect the shrinking MII-spindle.

Next, to determine the inhibitory effect of the somatic nuclei, the S-spindle was removed from somatic nuclei-injected oocytes 2 h after injection, and then the size of the MII-spindle was examined after culture for 4 h (Fig. 2B). This experiment was performed under the assumption that if the S-spindle inhibits the recovery of the shrinking MII-spindle, its size will recover following removal of the S-spindle from reconstructed oocytes. The S-spindle could be easily removed from reconstructed oocytes without any staining owing to its relatively large size (Fig. 2D,E; Fig. S4). Consequently, the size of the MII-spindle did not recover, even in the absence of the S-spindle in oocytes (Fig. 2H,I; Table S6); this was confirmed by γ-tubulin staining (Fig. S2; Table S3). The results of these two experiments suggest that neither the cytoplasm of somatic cells nor the S-spindle inhibited the recovery of the shrinking MII-spindle.

Next, to examine the reversibility of MII-spindle shrinkage, the MII-spindle was removed from reconstructed oocytes, injected into an intact oocyte (Fig. 2C), and observed using a live-cell imaging system (Movie 2). As shown in Fig. 2J-L and Table S7, the shrinking MII-spindle recovered within 2 h of transfer into a fresh oocyte (from 50% to 85% of original), which suggested that the shrinking MII-spindle had the ability to recover its size. Thus, these results suggested that the injected somatic cell nuclei consumed RRFs from oocytes during the remodeling and reprogramming of the nuclei, consequently leading to the shrinkage of the MII-spindle in these oocytes lacking RRFs. In addition, the size of the original MII-spindle of the fresh oocyte was slightly shrunk by transplantation of a shrinking MII-spindle (92% of original), which suggest that introduced shrinking MII-spindles also consumed RRFs from fresh oocytes. Importantly, these RRF-reduced oocytes did not reproduce RRFs during the 3 h of study following treatment.

If the technique described above removed essential factors for embryo development along with RRFs, the subsequent oocytes would not be suitable for use in further experiments. In addition, although the injection of somatic cell cytoplasm into embryos does not affect the birth rate of offspring, it leads to placental abnormalities in the offspring (Van Thuan et al., 2006a).

First, to examine the normality of the shrinking MII-spindle for chromosome segregation, two somatic cell nuclei were injected into the oocyte and the S-spindle was removed 2 h later. Then, the RRF-reduced oocytes were parthenogenetically activated, and chromosome segregation was observed using a live-cell imaging system (Fig. 3A). Consequently, no shrinking MII-spindles presented abnormal chromosome segregation, consistent with the control (Fig. S5; Table S8).

Next, oocytes with reduced RRFs, in which the S-spindle was removed before use, were fertilized with spermatozoa, and their developmental potential was examined in vitro and in vivo (Fig. 3B). As a control, to mimic the damage induced by the injection and enucleation treatment, the MII-spindle of oocytes was enucleated and then injected into the same oocytes. When these embryos were cultured in vitro, 80% developed to the morulae stage 3 days after fertilization, which was similar to the development of the control oocyte (87%) (Fig. 3D,E; Table S9). However, when embryos were cultured for up to 96 h, 36% of embryos could develop to the blastocyst stage, which was lower than the value for the control (74%) (Fig. 3F). When embryos were transferred to a recipient female at the two-cell stage, although the birth rate was slightly but significantly decreased (39%) compared with that of the control (52%) (Fig. 3G; Table S10), 14 healthy offspring were obtained from embryos fertilized with RRF-reduced oocytes. This showed that the RRF-reduced oocytes were damaged slightly but not critically, and could undergo normal embryonic development after fertilization. In addition, this suggested that a normal-sized MII-spindle is not essential for full-term development. The average body weight and placental weight of newborns were slightly increased compared with those of the control (Table S10; Fig. S6).

Finally, we examined whether cloned offspring could be generated using RRF-reduced oocytes. In this study, one to three somatic cells were injected into the MII-spindle of oocytes, and 2 h later, one mixed large spindle (Fig. S7) was formed in all oocytes, irrespective of the number of injected somatic cell nuclei, which could be removed relatively easily using a large pipette. Then, another somatic cell nucleus was injected into the enucleated RRF-reduced oocytes (Fig. 3C). As a control, usual nuclear transfer was performed at the same time.

When γ-tubulin distribution in nuclear-transferred oocytes derived from RRF-reduced oocytes was compared with that in the control, there was no clear localization of γ-tubulin in either of the reconstructed oocytes (Fig. S8), suggesting lower levels of γ-tubulin in RRF-reduced oocytes. Conversely, when the epigenetic status of H3K4me3 and H3K27me3 (associated with specific abnormalities; Hormanseder et al., 2017; Liu et al., 2016; Matoba et al., 2018; Xie et al., 2016) of cloned embryos derived from RRF-reduced oocytes was observed, it was found that their methylation levels were slightly increased when RRF-reduced oocytes were generated by the one- or two-cell injection/removal method, but decreased when RRF-reduced oocytes were generated by the three-cell injection/removal method (Fig. 3H,J). However, there was no significant difference when comparing with their methylation levels in the control clone (Fig. 3H-K, with the exception of the two cell-injection/removal method of Fig. 3J; Table S11; Fig. S9).

When cloned embryos were cultured, most were able to develop to the two-cell stage, irrespective of the number of injected somatic cell nuclei (Table 1). After embryo transfer into recipient females, healthy cloned offspring were obtained (Fig. 3L). Interestingly, when RRF-reduced oocytes were generated via the one-somatic-cell injection/removal method, and subsequently used for nuclear transfer, the success rate of cloned mice was higher (4.4% versus control 3.7%; Fig. 3M). Furthermore, the body (Fig. 3N) and placental (Fig. 3O) weights were reduced in the cloned mice compared with those in the control. In addition, when RRF-reduced oocytes were generated by the two- or three-cell injection/removal method and used for nuclear transfer, healthy cloned offspring were obtained; however, the success rates were reduced compared with those of the control, and body and placental weights were increased, similar to those of the control clone.

The abnormalities observed in cloned embryos and the low birth rate of cloned animals are believed to result from incomplete reprogramming of donor nuclei after nuclear transfer (Matoba and Zhang, 2018); this incomplete reprogramming is probably caused by RRFs in the oocyte. However, the role of RRFs in oocytes is to reprogram the nucleus of the spermatozoa and oocyte, and not that of the somatic cell. Therefore, unmatched RRFs may inaccurately reprogram somatic cell nuclei. In this case, increasing the amounts of RRFs in oocytes would not improve the reprogramming of somatic cell nuclei. In fact, when we attempted to increase the amount of RRFs by fusing two oocytes, the concentration in the cytoplasm of giant oocytes was not changed; however, the total amount of RRFs in each oocyte was increased twofold and none of the cloned embryos developed to full term (Sayaka et al., 2008). Conversely, when HDACs were inhibited in oocytes with a low concentration of an HDAC inhibitor (e.g. 5 nM of TSA), some epigenetic abnormalities of cloned embryos were corrected and the birth rate of cloned mice was increased (Kishigami et al., 2006a; Ono et al., 2010; Van Thuan et al., 2009). Although it is unclear whether HDACs act as RRFs, these results suggest that an excess amount of unmatched RRFs in oocytes may lead to some abnormalities or damage, rather than improving the reprogramming of the donor nucleus.

A recent quantitative proteomics analysis identified thousands of proteins in mouse oocytes (Ma et al., 2013; Pfeiffer et al., 2015; Wang et al., 2010). To identify RRFs from this high number of proteins, it is necessary to compare intact oocytes with RRF-reduced oocytes. However, the HDAC inhibitor does not reduce the level of HDACs in oocytes, and half- or double-sized oocytes cannot be used for this purpose because there is no difference in the concentration of RRFs. Therefore, we developed a method in which some RRFs were removed from oocytes by injecting somatic cell nuclei into oocytes, and then subsequently removing these nuclei following the consumption of RRFs. This resulted in the immediate shrinkage of the MII-spindle in oocytes, which is a novel finding of this study. Although the MII-spindle shrinkage was reversible, spindle size never recovered in the oocytes, suggesting that once RRFs were removed from oocytes, they could not be regenerated, at least for a few hours after treatment. It is worth mentioning that the RRFs in this study probably also include spindle regulating factors.

In this study, the shrinking MII-spindle of oocytes was used as a marker of the reduced amount of RRFs, even though the relationship between these two has not yet been elucidated. However, donor cumulus cell nuclei, which were at the G0/G1-phase of the cell cycle (Schuetz et al., 1996), were absolutely remodeled, condensed, and formed the M-phase spindle in oocytes after nuclear injection (Fig. 1). This clearly suggests that at least remodeling and spindle regulating factors were removed from oocytes by this treatment and that, currently, only this method allows us to remove these factors without reducing the cytoplasm of oocytes.

Thus, it is important to determine whether RRF-reduced oocytes retain normal potential for embryo development after fertilization, or whether they can be used as a recipient oocyte for nuclear transfer. Previously, Kyogoku and Kitajima reported that when cytoplasmic factors were reduced by removing half the volume of the cytoplasm from oocytes, the size of the MII-spindle decreased and the birth rate of offspring also decreased (Kyogoku and Kitajima, 2017). However, in the present study, we demonstrated that the shrinking MII-spindle had normal potential for chromosome segregation, and that use of RRF-reduced oocytes only had a slight negative effect on fertilization and embryo development to full term.

Interestingly, when RRFs were removed from oocytes using the one-somatic-cell injection/removal method, cloned offspring could be generated with a relatively higher success rate than the control. Placentomegaly is a specific abnormality that is always observed in cloned mice (Tanaka et al., 2001); however, in this method, the weight of cloned placentae was reduced compared with that of the control cloned placentae. In addition, epigenetic modification of H3K4me3 and H3K27me3, which is a specific abnormality of cloned embryos, was slightly improved and was closer to that observed in fertilized embryos (Hormanseder et al., 2017; Liu et al., 2016; Matoba et al., 2018; Xie et al., 2016). This suggests that the oocytes possessed an excessive amount of RRFs, which perturbed rather than promoted the correct reprogramming of somatic cell nuclei. However, excessively reducing the amount of RRFs in oocytes using the two- to three-somatic-cell injection/removal method will negatively affect embryo development. This may suggest that an appropriate amount of RRFs is important for embryo development, similar to HDAC.

Enucleation was previously thought to explain, at least in part, the low success rate of cloned animals (Simerly et al., 2003). However, even if enucleation reduced the amount of only a few RRFs from oocytes, the effect would not be significant, and the reprogramming potential of oocytes would not be affected because oocytes possess abundant RRFs. Conversely, the karyoplast of the S-spindle derived using the three-somatic-cell injection/removal method possessed highly condensed RRFs, because these somatic nuclei absorbed RRFs from oocytes. The use of these karyoplasts for proteomics analysis may help identify RRFs. Thus, although further study is required, this method enables a more mechanistic study of the factors, rather than a descriptive study of the treatment of oocytes with reagents or inhibitors, without harmful effect to the oocytes. It therefore provides a new tool or a more valuable resource for researchers to identify RRFs in oocytes.

ICR and B6D2F1 (C57BL/6×DBA/2) mice, aged 8-10 weeks, were used to produce oocytes, spermatozoa and cumulus cells. The surrogate pseudopregnant females used as embryo transfer recipients were ICR strain mice, and were mated with vasectomized males of the same strain. B6D2F1 and ICR mice were purchased from Shizuoka Laboratory Animal Center (Hamamatsu, Japan). On the day of the experiments, or upon completion of the experiments, mice were euthanized by CO₂ inhalation or by cervical dislocation and used for subsequent analysis. 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 University of Yamanashi.

Female mice were super-ovulated via injection of 5 IU of equine chorionic gonadotropin, followed by 5 IU of human chorionic gonadotropin (hCG) 48 h later. Cumulus-oocyte complexes (COCs) were collected from the oviducts of females 14-16 h later and moved to a Falcon dish containing HEPES-CZB medium (Kimura and Yanagimachi, 1995). To disperse the cumulus cells, COCs were transferred to a 50 µl droplet of HEPES-CZB medium containing 0.1% bovine testicular hyaluronidase for 3 min. Cumulus-free oocytes were washed twice and moved to 20 µl droplets of CZB medium (Chatot et al., 1989) for culture. At the same time, the remaining cumulus cells were introduced into PVP medium (Kimura and Yanagimachi, 1995) on the manipulation chamber.

For somatic cell nuclear injection, donor cumulus cells were gently aspirated in and out of the injection pipette until their nuclei were largely devoid of a visible cytoplasmic membrane. For the injection of two or three cumulus cell nuclei into one oocyte, two or three nuclei were placed together inside the pipette. Then, one to three nuclei were immediately injected into an intact oocyte. These reconstructed oocytes were kept in an incubator until use.

Reconstructed oocytes were stained with Hoechst 33342 (0.1 mg/ml; Invitrogen, H1399) for 5 min, then observed using fluorescence microcopy (Olympus IX71). For immunostaining, the zona pellucidae of injected oocytes were removed by acetic Tyrode solution (Sigma-Aldrich, T1788) and then naked oocytes were fixed for 30 min at 25°C in 4% (w/v) paraformaldehyde. The fixed oocytes were washed three times in PBS-polyvinyl alcohol (0.1 mg/ml PVA, Sigma-Aldrich) for 10 min and stored overnight at 4°C in PBS supplemented with 1% (w/v) bovine serum albumin (BSA/PBS; Sigma-Aldrich) and 0.1% (v/v) Triton X-100 (Nacalai Tesque). For β-tubulin staining, oocytes were incubated with the relevant primary antibody, mouse monoclonal anti-β-tubulin antibody labeled with FITC (1:1000; BD Pharmingen, 556321), and cultured in 0.1% Triton X-100-1% BSA/PBS for 1 h at room temperature. After the oocytes had been washed three times in PBS-PVA for 10 min each, DNA was stained with propidium iodide (Sigma-Aldrich).

For γ-tubulin staining, oocytes were incubated with the relevant primary antibody, rabbit polyclonal anti-γ-tubulin antibody (1:10,000; Sigma-Aldrich, C7604), and cultured in 0.1% Triton X-100-1% BSA/PBS for 1 h at room temperature. After the oocytes were washed three times in PBS-PVA for 10 min, they were incubated at 25°C for 45 min with dye-conjugated secondary antibodies: Alexa Fluor 488-labeled goat anti-rabbit IgG (1:500; Thermo Fisher Scientific, A11034). After the oocytes had been washed three times in PBS-PVA for 10 min each, DNA was stained with 4′,6-diamino-2-phenylindole (DAPI; 3 μg/ml; Molecular Probes). Next, the oocytes were washed thoroughly, mounted on slides using Vectashield Mounting Medium (Vector Laboratories, H-1200), and observed under a confocal scanning laser microscope (Olympus FV1200).

For MII-spindle measurements, only spindles that were relatively flat and could be clearly observed were used. The S-spindle formed had an irregular shape with misalignment of chromosomes; therefore, the MII-spindle could be clearly distinguished from the S-spindle. Fused MII- and S-spindles or difficult to distinguish spindles were not measured. ImageJ software was used for the measurement.

As previously reported (Yamagata et al., 2009a,b), α-tubulin-EGFP and H2B-mCherry mRNA were injected into oocytes before somatic cell injection. The mRNA-injected oocytes were incubated for at least 3 h to allow the mRNA to be sufficiently translated for imaging. Then, live-cell imaging was performed using a CV1000 microscope (Yokogawa Electric).

Large cumulus cells were gently aspirated in and out of the injection pipette, and then the cytoplasm was collected inside the pipette. This was repeated for three to four cells and the pooled cytoplasm was injected into an oocyte. The size of the MII-spindle of oocytes was observed 4 h later by immunostaining, as described above.

The denuded nuclei were collected using the technique described previously (Wakayama et al., 2008). Briefly, a piece of brain tissue (frontal lobe) was placed in a 1.5 ml tube and homogenized gently. The isolated nuclei were collected using a relatively large micropipette (15 µm) and washed once in a PVP drop, and then one or two nuclei were injected into an oocyte. The size of the MII-spindle of oocytes was observed 4 h later by immunostaining, as described above.

To remove the S-spindle, treated oocytes were transferred to a droplet of HEPES-CZB medium containing 5 µg/ml cytochalasin B under mineral oil on a micromanipulation microscope stage. About 10 min later, neither the S-spindle nor the MII-spindle bump could be seen; however, it was possible to distinguish between S-spindle and the MII-spindle due to the size of the translucent region (Fig. S4). This allowed us to remove the S-spindle without any nuclear staining. The method used to remove the S-spindle was the same as that used for enucleation, as described previously (Wakayama et al., 1998). The size of the spindle of oocytes was observed 4 h later by immunostaining, as described above.

To transplant the shrinking MII-spindle into fresh oocytes, somatic cell-injected oocytes and intact oocytes were transferred to the same droplet of HEPES-CZB medium containing cytochalasin B. Then, the shrinking MII-spindle was removed from the oocytes and immediately injected into intact oocytes. In this case, the whole membrane of the oocytes was aspirated until it disappeared. Then, the size of the MII-spindle was examined by live-cell imaging. Before transplantation, EGFP-α-tubulin and H2B-mCherry mRNA were injected into both oocytes.

First, EGFP-α-tubulin and H2B-mCherry mRNAs were injected into reconstructed oocytes, and then the S-spindle was removed from the oocytes as mentioned above. Subsequently, the oocytes were parthenogenetically activated as described previously (Wakayama et al., 1998). Briefly, oocytes were cultured with Ca²⁺-free CZB medium containing 5 mM Sr²⁺ for 1 h, then washed with CZB medium three times and used for live-cell imaging for up to 6 h, as described above.

Two cumulus cell nuclei were injected into oocytes, and then the S-spindle was removed as described above. Then, ICSI was performed using the technique described by Kimura and Yanagimachi (Ito et al., 2019; Kimura and Yanagimachi, 1995). Briefly, a droplet of sperm suspension was mixed thoroughly with 20 µl of HEPES-CZB medium containing 12% (w/v) PVP (Mr 360 kDa) in the micromanipulation chamber. The sperm head was separated from the midpiece and tail by the application of one or more piezo pulses, and then injected into a shrinking MII-spindle oocyte. In the control, to mimic the injection and enucleation treatment, the MII-spindles of oocytes were enucleated and then injected into the same oocytes. After ICSI, fertilized embryos were cultured for up to 96 h to observe their potential development in vitro, or were transferred into recipient females to examine the potential for full-term development. At 18.5 days post-coitus (dpc), the offspring were delivered by cesarean section, and the weights of pups and placentae were determined.

One to three cumulus cell nuclei were injected close to the MII-spindle of oocytes. Most oocytes showed one large spindle 2 h after injection, with mixed somatic cell- and oocyte-derived chromosomes. Therefore, all chromosomes could be removed in one enucleation process. Then, another donor nucleus was injected into the enucleated oocytes, as described above. These reconstructed oocytes were activated and cultured for 1 day. Then, two-cell cloned embryos were transferred into recipient females, as detailed above. At 18.5 dpc, the offspring were delivered by cesarean section, and the weights of pups and placentae were determined. Some live fetuses were raised by lactating foster mothers (ICR strain).

Ten hours after activation, oocytes were fixed and permeabilized in PBS containing 4% paraformaldehyde and 0.2% (v/v) Triton X-100 (Nacalai Tesque) for 20 min. The activated zygotes were washed twice in PBS containing 0.25% (w/v) BSA (Nacalai Tesque) (PBS-BSA) for 15 min each. Zygotes were then incubated with the relevant primary antibodies, rabbit polyclonal anti-H3K4 me3 (1:1000; Upstate, 07-473) or rabbit polyclonal anti-H3K27 me3 (1:1000; Millipore, 07-449) in PBS-BSA overnight at 4°C. After the embryos were washed twice in PBS-BSA for 15 min, they were incubated at 25°C for 1 h with dye-conjugated secondary antibodies, Alexa Fluor 568-labeled goat anti-rabbit IgG (1:500; Molecular Probes, A11011). After washing twice in PBS–BSA, the embryos were mounted on a glass slide in Vectashield antibleaching solution (Vector Laboratories) containing 3 μg/ml DAPI (Molecular Probes). Subsequently, serial images were obtained using fluorescence confocal microscopy (FV-1000; Olympus). Relative H3K4me3 or H3K27me3 levels in embryos were measured using ImageJ software.

All experiments were repeated more than three times. These studies were performed by two or three experimentalists independently, and similar results were obtained irrespective of experimentalist. The rate of embryo development, the birth of offspring, body and placental weights were evaluated using Chi-squared tests. The size of the MII-spindle after injection of cytoplasm or purified nuclei was evaluated using a paired two-tailed t-test. Fluorescence levels were evaluated using one-way analysis of variance (ANOVA) followed by the Tukey-Kramer test, and P<0.05 was considered to represent a statistically significant difference.

We thank Miss C. Yamaguchi and Mr M. Nakamura for assistance in preparing the manuscript.

The peer review history is available online at https://dev.biologists.org/lookup/doi/10.1242/dev.190777.reviewer-comments.pdf