Hypomethylation of paternal DNA in the late mouse zygote is not essential for development

Global demethylation of DNA which marks the onset of development occurs asynchronously in the mouse; paternal DNA is demethylated at the the zygote stage, whereas maternal DNA is demethylated later in development. The biological function of such asymmetry and its underlying mechanisms are currently unknown. To test the hypothesis that the early demethylation of male DNA may be associated with protamine-histone exchange, we ,used round spermatids, whose DNA is still associated with histones, for artificial fertilization (round spermatid injection or ROSI), and compared the level of methylation of metaphase chromosomes in the resulting zygotes with the level of methylation in zygotes obtained after fertilization using mature sperm heads (intracytoplasmic sperm injection or ICSI). In contrast to ICSI-derived zygotes, ROSI-derived zygotes possessed only slightly demethylated paternal DNA. Both types of zygotes developed to term with similar rates which shows that hypomethylation of paternal DNA at the zygotic metaphase is not essential for full development in mice. Incorporation of exogenously expressed histone H2BYFP into paternal pronuclei was significantly higher in ICSI-derived zygotes than in ROSI-derived zygotes. Surprisingly, in the latter the incorporation of histone H2BYFP into the paternal pronucleus was still significantly higher than into the maternal pronucleus, suggesting that some exchange of chromatin-associated proteins occurs not only after ICSI but also after ROSI. This may explain why after ROSI, some transient demethylation of paternal DNA occurs early after fertilization, thus providing support for the hypothesis regarding the link between paternal DNA demethylation and protamine/histone exchange.     

Parental origin of chromatin in human monopronuclear zygotes revealed by asymmetric histone methylation patterns, differs between IVF and ICSI

In mouse zygotes, many post-translational histone modifications are asymmetrically present in male and female pronuclei. We investigated whether this principle could be used to determine the genetic composition of monopronuclear human zygotes in conventional IVF and ICSI. First we determined whether male female asymmetry is conserved from mouse to human by staining polypronuclear zygotes with antibodies against a subset of histone N-tail post-translational modifications. To analyze human monopronuclear zygotes, a modification, H3K9me3, was selected that is present in the maternal chromatin. After IVF a total of 45 monopronuclear zygotes were obtained. In 39 (87%) of zygotes a nonuniform staining pattern was observed, proof of a bi-parental origin and assumed to result into a diploid conception. Two zygotes showed no staining for the modification, indicating that the single pronucleus was of paternal origin. Four zygotes contained only maternally derived chromatin. ICSI-derived monopronuclear zygotes (n = 33) could also be divided into three groups based on the staining pattern of their chromatin: (1) of maternal origin (n = 15), (2) of paternal origin (n = 8) or (3) consisting of two chromatin domains as dominating in IVF (n = 10). Our data show that monopronuclear zygotes originating from IVF generally arise through fusion of parental chromatin after sperm penetration. Monopronuclear zygotes derived from ICSI in most cases contain uni-parental chromatin. The fact that chromatin was of paternal origin in 24% of ICSI and in 4% of the IVF zygotes confirms earlier results obtained by FISH on cleavage stages. Our findings are of clinical importance in IVF and ICSI practice.     

The absence of a DNA replication checkpoint in porcine zygotes

It has been demonstrated that in the zygotes of some mammals a unique checkpoint controls the onset of DNA replication. Thus, DNA replication begins in the maternal pronucleus only after the paternal pronucleus is fully formed. In our experiments we have investigated whether this checkpoint also operates in porcine zygotes produced either by in vitro fertilization (IVF) or by intracytoplasmic sperm injection (ICSI). Our results show that the onset of DNA replication occurs in the maternal pronucleus even in the presence of an intact sperm head in zygotes produced by ICSI, as well as in polyspermic eggs where some sperm heads are intact or male pronuclei are not yet fully developed. We conclude that in porcine zygotes there is an absence of the DNA replication checkpoint that is typical for some other mammals.     

Full term development of rabbit oocytes fertilized by intracytoplasmic sperm injection

Intracytoplasmic sperm injection (ICSI) has been applied successfully in the treatment of male infertility in humans and in fertilization research in mice. However, the technique has had limited success in producing offspring in other species including the rabbit. The aim of this research was to test the in vitro and in vivo developmental of rabbit oocytes after ICSI. Sperm used for ICSI were collected from mature Dutch Belted buck and washed 2-3 times with PBS +0.1% polyvinyl alcohol (PVA) and then mixed with 10% polyvinyl pyrrolidone (PVP) prior to microinjection. Oocytes were collected from superovulated does 14-15 hr after hCG injection and were fertilized by microinjection of a single sperm into the ooplasm of each oocyte without additional activation treatment. After ICSI, the presumed zygotes were either cultured in KSOM +0.3% BSA for 4 days or transferred into oviducts of recipient does at the pronuclear or 2-cell stage. A high percentage of fertilization (78%, n = 114) and blastocyst development (39%) was obtained after ICSI. Control oocytes, receiving a sham injection, exhibited a lower activation rate (31%, n = 51) and were unable to develop to the blastocyst stage, suggesting that the blastocysts developed following ICSI were derived from successful fertilization rather than parthenogenetic development. A total of 113 embryos were transferred to six recipient does. Two recipients became pregnant and delivered seven live young. Our results demonstrated that rabbit oocytes can be successfully fertilized and activated by ICSI and can result in the birth of live offspring.     

Timing of the first cleavage post‐insemination affects cryosurvival of in vitro–produced bovine blastocysts

The time of the first cleavage of bovine zygotes during in vitro culture can affect the rate of development and cell number of the blastocysts. The aim of this work was to study the effect of the timing of first cleavage on the cryosurvival of the resulting blastocysts. Following standard IVM and IVF, zygotes were cultured in modified synthetic oviduct fluid (SOF), with 10% fetal calf serum (FCS) added 48 hr post insemination, in a humidified atmosphere of 5% CO_2, 5% O₂ and 90% N₂. Embryos which cleaved by 24, 27 30, 33, or 36 hr after insemination (IVF) were harvested and further cultured to the blastocyst stage (day 7 or day 8 post IVF). All developing blastocysts on days 7 and 8 were classified into three groups and were cryopreserved by vitrification. Group A consisted of blastocysts (<150 μm, small blastocysts); group B consisted of expanded or hatching blastocysts (>150 μm, large blastocysts); and group C consisted of morphologically poor quality blastocysts. The vitrification solution consisted of 6.5 M glycerol and 6% bovine serum albumin in PBS (VS3a). Thawed embryos were cultured further and survival was defined as the re‐expansion and maintenance of the blastocoel over 24, 48, and 72 hr, respectively. Overall survival and hatching at 72 hr post‐thawing was higher in blastocysts formed by day 7 than those formed by day 8 (60% vs. 40% survival; 63% vs. 45% hatching). Large blastocysts from day‐7 and day‐8 groups survived significantly better than small or poor quality blastocysts (76% vs. 63% and 31%; 72% vs. 30% and 26%, respectively; P < 0.05). Day‐7 blastocysts from the 27‐ and 30‐hr cleavage groups survived significantly better than those from the 36‐hr group (63% and 66% vs. 25%, P < 0.05). Day‐8 blastocysts from later cleaved (30 hr) zygotes had a higher survival than the 27‐hr cleavage groups (52% vs. 26%, P < 0.05). These results indicate that the day of blastocyst appearance, developmental stage, and timing of the first cleavage post‐insemination can influence the cryosurvival of bovine blastocysts following vitrification. Mol. Reprod. Dev. 53:318–324, 1999. © 1999 Wiley‐Liss, Inc.     

Microtubules and parental genome organisation during abnormal fertilisation in humans

We analysed the distribution of beta-tubulins, acetylated alpha-tubulins and chromatin configuration in 113 human zygotes showing abnormal fertilisation, 16-18 h after conventional in vitro fertilisation (IVF) or intracytoplasmic sperm injection (ICSI). After a first characterisation using phase contrast microscopy, immunofluorescence staining was performed in 67 IVF and 46 ICSI zygotes that developed one, three or more pronuclei and/or subnuclei, with or without extrusion of the second polar body. Independently of the number of pronuclei found, beta-tubulins were uniformly distributed throughout the cytoplasm of the abnormal zygotes. We did not observe any kind of microtubule alteration with respect of the ploidy level and/or its origin. The most frequent abnormal fertilisation pattern found after IVF was the presence of three or four pronuclei (74.6%). On the other hand, the presence of one pronucleus (63.0%) was the main pattern found after ICSI. No differences between the two groups were seen in terms of development of subnuclei. Anamolies detected after IVF and ICSI showed different aetiologies such as parthenogenetic activation, gynogenetic or androgenetic development, as well as digynic or diandric fertilisation.     

Genomic DNA damage in mouse transgenesis

Creating transgenic mammals is currently a very inefficient process. In addition to problems with transgene integration and unpredictable expression patterns of the inserted gene, embryo loss occurs at various developmental stages. In the present study, we demonstrate that this loss is due to chromosomal damage. We examined the integrity of chromosomes in embryos produced by microinjection of pronuclei, intracytoplasmic sperm injection (ICSI), and in vitro fertilization (IVF)-mediated transgenesis, and correlated these findings with the abilities of embryos to develop in vitro and yield transgenic morulas/blastocysts. Chromosomal analysis was performed after microinjection of the pronuclei in zygotes, as well as in parthenogenetic and androgenetic embryos. In all the pronuclei injection groups, significant oocyte arrest and increased incidence of chromosome breaks were observed after both transgenic DNA injection and sham injection. This indicates that the DNA damage is a transgene-independent effect. In ICSI-mediated transgenesis, there was no significant oocyte arrest. The observed chromosomal damage was lower than that after pronuclei microinjection in zygotes and was dependent upon the presence of exogenous DNA. The occurrence of DNA breaks, as measured by comet assay performed on the sperm prior to ICSI, showed that DNA damage was present in the sperm before fertilization. Embryonic development in vitro and transgene expression at the morula/blastocyst stage were higher in ICSI-mediated transgenesis than after microinjection of pronuclei into zygotes. Sperm-mediated gene transfer via IVF did not affect chromosome integrity, allowed good embryo development, but did not yield any transgenic embryos. The present study demonstrates that DNA damage occurs after both the microinjection of pronuclei and ICSI-mediated transgenesis, albeit through different mechanisms.     

Dynamics of DNA-demethylation in early mouse and rat embryos developed in vivo and in vitro

Virtually all mammalian species including mouse, rat, pig, cow, and human, but not sheep and rabbit, undergo genome-wide epigenetic reprogramming by demethylation of the male pronucleus in early preimplantation development. In this study, we have investigated and compared the dynamics of DNA demethylation in preimplantation mouse and rat embryos by immunofluorescence staining with an antibody against 5-methylcytosine. We performed for the first time a detailed analysis of demethylation kinetics of early rat preimplantation embryos and have shown that active demethylation of the male pronucleus in rat zygotes proceeds with a slower kinetic than that in mouse embryos. Using dated mating we found that equally methylated male and female pronuclei were observed at 3 hr after copulation for mouse and 6 hr for rat embryos. However, a difference in methylation levels between male and female pronuclei could be observed already at 8 hr after copulation in mouse and 10 hr in rat. At 10 hr after copulation, mouse male pronuclei were completely demethylated, whereas rat zygotes at 16 hr after copulation still exhibited detectable methylation of the male pronucleus. In addition in both species, a higher DNA methylation level was found in embryos developed in vitro compared to in vivo, which may be one of the possible reasons for the described aberrations in embryonic gene expression after in vitro embryo manipulation and culture.     

Fate of the first polar bodies in mouse oocytes

Both nuclear transfer and intracytoplasmic sperm injection (ICSI) practice necessitates studies on the spatial relationship between the MII spindle and the first polar bodies (FPB). Although recent observations have shown that the FPB position does not predict accurately the location of the meiotic spindle in metaphase II oocytes of monkey, hamster, and human, detailed studies on FPB deviation and its affecting factors are lacking. Since polar bodies can be used for genetic testing and oocyte quality grading, their life span under different conditions should be studied. The timing of formation and degeneration and the position relative to the MII spindle of the FPB and the factors affecting FPB deviation and degeneration during in vivo and in vitro aging of both in vivo and in vitro matured mouse oocytes were investigated in this study. Mice of the Kun-ming breed were used, and the intact and degenerated FPB were identified through microscopic morphology in combination with propidium iodide (PI) exclusion test and the chromosomes visualized by Hoechst staining. Results are summarized as follows: (i) oocytes started FPB extrusion at 8 hr after the onset of in vivo or in vitro maturation, but the number of FPB reached maximum much later in vitro (14 hr of culture) than in vivo (10 hr post hCG). (ii) Some FPB began to degenerate before ovulation and around 70% became degenerated within 6 hr after maximal nuclear maturation both in vivo and in vitro; they disappeared faster during in vivo than in vitro aging but turned from intact to degenerated at a similar tempo. (iii) Some FPB began to deviate from the MII spindle 10 hr after hCG injection or in vitro culture and the distance between FPB and the spindle increased with time during both in vivo and in vitro aging. (iv) FPB deviated more slowly in the in vitro matured oocytes than in in vivo matured. (v) Denudation performed after FPB extrusion markedly enhanced its deviation. (vi) The perivitelline space (PVS) increased with time during maturation and aging in vivo and in vitro and the values of PVS and the percentages of FPB adjacent to the spindle were significantly negatively correlated. (vii) Cytochalasin B and colchicine had no effect on FPB deviation. (viii) None of the more than 3,500 FPBs observed was found to be dividing or have divided into two cells at any time points before or after ovulation or in vitro maturation. Our results were consistent with the possibility that the displacement of the FPB was a time- and PVS-dependent process, indicating that PVS would increase with time and its formation and enlargement would facilitate the lateral displacement of the degenerating FPB.     

In vivo survival of domestic cat oocytes after vitrification, intracytoplasmic sperm injection and embryo transfer

We evaluated: (1) cleavage rate after IVF or intracytoplasmic sperm injection (ICSI) of in vivo- and in vitro-matured oocytes after vitrification (experiment 1); and (2) fetal development after transfer of resultant ICSI-derived embryos into recipients (experiment 2). In vivo-matured cumulus-oocyte complexes (COCs) were recovered from gonadotropin-treated donors at 24 h after LH treatment. In vitro-matured oocytes were obtained by mincing ovaries (from local veterinary clinics) and placing COCs into maturation medium for 24 h. Mature oocytes were denuded and cryopreserved in a vitrification solution of 15% DMSO, 15% ethylene glycol, and 18% sucrose. In experiment 1, for both in vivo- and in vitro-matured oocytes, cleavage frequencies after IVF of control and vitrified oocytes and after ICSI of vitrified oocytes were not different (P > 0.05). After vitrification, blastocyst development occurred only in IVF-derived, in vitro-matured oocytes. In experiment 2, 18 presumptive zygotes and four two-cell embryos derived by ICSI of vitrified in vitro-matured oocytes and 19 presumptive zygotes produced from seven in vivo- and 12 in vitro-matured oocytes were transferred by laparoscopy into the oviducts of two recipients, respectively. On Day 21, there were three fetuses in one recipient and one fetus in the other. On Days 63 and 66 of gestation, four live kittens were born. In vivo viability of zygotes and/or embryos produced via ICSI of vitrified oocytes was established by birth of live kittens after transfer to recipients.     

Fertilization of mouse oocytes from in vitro-matured preantral follicles using classical in vitro fertilization or intracytoplasmic sperm injection

Early preantral mouse follicles with a diameter of 110-160 microm were cultured in vitro for 10 or 12 days. Mature oocytes were retrieved following hCG, and fertilization was attempted either by in vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI). Two-cell and blastocyst formation rates and blastocyst cell numbers were compared between 10-day and 12-day in vitro-matured oocytes versus in vivo-matured oocytes. Uncleaved IVF oocytes were subjected to chromosome analysis. The 2-cell formation rate was significantly improved by ICSI compared with IVF both in 10-day (72.1% versus 56.1%; P = 0.03) and 12-day cultures (74.1% versus 54.5%; P = 0.028). Cytogenetic analysis of uncleaved MII oocytes following IVF showed that about 30% of MII oocytes showed no sign of sperm penetration. The blastocyst formation rate was significantly lower in 12-day versus 10-day cultures, whether fertilization was by IVF (40.7% versus 62.4%, P = 0.016) or by ICSI (32.5% versus 57.1%, P = 0.035). Blastocyst cell numbers from IVF and ICSI 10-day groups were similar and both significantly higher (P < 0.001) than from IVF 12-day cultures. All above expressed values were significantly higher for in vivo-matured oocytes. In conclusion, fertilization of oocytes from in vitro-matured mouse preantral follicles can be optimized with ICSI, giving significantly higher 2-cell formation rates than IVF. Blastocyst formation rate was not influenced by the technique of fertilization but rather by the extent of the in vitro culture period. Best results on preimplantation development of oocytes for in vitro-matured preantral follicles were obtained with ICSI on oocytes from 10-day in vitro cultures.     

Epigenetic abnormalities of the mouse paternal zygotic genome associated with microinsemination of round spermatids

Although round spermatid injection can be used to create progeny for males who do not produce mature sperm, the rate of successful embryogenesis after such procedures is significantly lower than that for similar procedures using mature spermatozoa. The mechanisms underlying this difference are unknown. In this study, we demonstrate that, unlike the normal paternal genome, the paternal zygotic genome derived from a round spermatid is highly remethylated before first mitosis after demethylation. Genomes from elongated spermatids exhibited an intermediate level of DNA methylation, between those of round spermatids and mature spermatozoa, suggesting that the male germ cell acquires the ability to maintain its undermethylated state in the paternal zygotic genome during this phase of spermiogenesis. In addition, treatment of zygotes with trichostatin A led to a significant reduction in DNA methylation, specifically in the spermatid-derived paternal genome, except for the pericentromeric regions enriched by trimethylation of Lys9 of histone H3. These data provide insight into epigenetic errors that may be associated with the poor development of embryos generated from immature spermatozoa.     

Maturity and fertility of rhesus monkey oocytes collected at different intervals after an ovulatory stimulus (human chorionic gonadotropin) in in vitro fertilization cycles

In rhesus monkeys undergoing ovarian stimulation for in vitro fertilization (IVF), a midcycle injection of human chorionic gonadotropin (hCG) substitutes for the LH surge and induces preovulatory oocyte maturation. The time interval between injection and oocyte collection, ideally, allows for the completion of oocyte maturation without ovulation, which would reduce the number of oocytes available for harvest. To evaluate the influence of this time interval on oocyte parameters following hCG administration, we conducted a series of gonadotropin treatment protocols in 51 animals in which the interval from hCG administration to follicular aspiration was systematically varied from 27 to 36 hr. Follicle number and size, evaluated prior to hCG administration by sonography, did not vary significantly or consistently with preovulatory maturation time. Oocytes were harvested by laparotomy or laparoscopy, and scored for maturity before insemination. The percentage of mature, metaphase II (MII) oocytes at recovery increased significantly with increasing preovulatory time and was inversely proportional to that of metaphase I (MI) oocytes. However, oocyte yield tended toward a progressive decrease with increasing preovulatory maturation times from a high of 27 oocytes at 27 hr to a low of 17 oocytes/animal at the 36 hr time interval. Fertilization levels declined significantly from a high of 50% at 27 hr to a low of 30% at 36 hr. Thus, although higher percentages of mature oocytes were recovered at the longer time intervals, optimal oocyte/embryo harvests were realized after the shorter time intervals (27 and 32 hr) and are most compatible with the goal of achieving high yields of fertile oocytes and embryos following gonadotropin stimulation in rhesus monkeys. © 1996 Wiley-Liss, Inc.     

Cytoskeletal and nuclear organization in mouse embryos derived from nuclear transfer and ICSI: a comparison of agamogony and syngamy before and during the first cell cycle

In this study, somatic cell nuclear transfer (SCNT) and intracytoplasmic sperm injection (ICSI) are used as models of agamogony and syngamy, respectively. In order to elucidate the reasons of low efficiency of somatic cell cloning, cytoskeletal and nuclear organization in cloned mouse embryos was monitored before and during the first cell cycle, and compared with the pattern of ICSI zygote. A metaphase-like spindle with alignment of condensed donor chromosomes was assembled within 3 hr after NT, followed by formation of pronuclear-like structures at 3-6 hr after activation, indicating that somatic nuclear remodeling depends on microtubular network organization. The percentage of two (pseudo-) pronuclei in cloned embryos derived from delayed activation was greater than that in immediate activation group (68.5% vs. 30.8%, P<0.01), but similar to that of ICSI group (68.5% vs. 65.5%, P>0.05). The 2-cell rate in NT embryos was significantly lower than that in zygotes produced by ICSI (64.8% vs. 82.5%, P<0.01). Further studies testified that the cloned embryos reached the metaphase of the first mitosis 10 hr after activation, whereas this occurred at 18 hr in the ICSI zygotes. Comparision of the pattern of microfilament assembly in early NT embryos with that in syngamic zygotes suggested that abnormal microfilamental pattern in cloned embryos may threaten subsequent embryonic development. In conclusion, agamogony, in contrast to syngamy, displays some unique features in respect of cytoskeletal organization, the most remarkable of which is that the first cell cycle is initiated ahead distinctly, which probably leads to incomplete organization of the first mitotic spindle, and contributes to low efficiency of cloning.     

Methylation reprogramming and chromosomal aneuploidy in in vivo fertilized and cloned rabbit preimplantation embryos

Active demethylation of the paternal genome but not of the maternal genome occurs in fertilized mouse, rat, pig, and bovine zygotes. To study whether this early demethylation wave is important for embryonic development, we have analyzed the global methylation patterns of both in vivo-fertilized and cloned rabbit embryos. Anti-5-methylcytosine immunofluorescence of in vivo-fertilized rabbit embryos revealed that the equally high methylation levels of the paternal and maternal genomes are largely maintained from the zygote up to the 16-cell stage. The lack of detectable methylation changes in rabbit preimplantation embryos suggests that genome-wide demethylation is not an obligatory requirement for epigenetic reprogramming. The methylation patterns of embryos derived from fibroblast and cumulus cell nuclear transfer were similar to those of in vivo-fertilized rabbit embryos. Fluorescence in situ hybridization with chromosome-specific BACs demonstrated significantly increased chromosomal aneuploidy rates in cumulus cell nuclear transfer rabbit embryos and embryos derived from nuclear transfer of rabbit fibroblasts into bovine oocytes compared with in vivo-fertilized rabbit embryos. The incidence of chromosomal abnormalities was correlated with subsequent developmental failure. We propose that postzygotic mitotic errors are one important explanation of why mammalian cloning often fails.     

Development of porcine embryos and offspring after intracytoplasmic sperm injection with liposome transfected or non-transfected sperm into in vitro matured oocytes.

The objective of this study was to evaluate in vitro and in vivo development of porcine in vitro matured (IVM) porcine oocytes fertilised by intracytoplasmic sperm injection (ICSI) and the possibility of producing transgenic embryos and offspring with this procedure. Activated ICSI oocytes had a higher pronuclear formation than non-activated ICSI oocytes (mean 64.8+/-17.3% vs 28.5+/-3.4%, p<0.05). When the zygotes with two pronuclei were cultured to day 2, there was no difference (p<0.05) in the cleavage rate (mean 60.0+/-7.0% vs 63.3+/-12.7%) between the two groups. The blastocyst rate in the activation group was significantly higher than that in the non-activation group (mean 30.0+/-11.6% vs 4.6+/-4.2%, p<0.05). After injection of the sperm transfected with DNA/liposome complex, destabilised enhanced green fluorescent protein (d2EGFP) expression was not observed on day 2 in either cleaved or uncleaved embryos. But from day 3, some of the embryos at the 2-cell to 4-cell stage started to express d2EGFP. On day 7, about 30% of cleaved embryos, which were in the range of 2-cell to blastocyst stage, expressed d2EGFP. However, for the IVF oocytes inseminated with sperm transfected with DNA/liposome complex, and for oocytes injected with sperm transfected with DNA/liposome complex, and for oocytes injected with DNA/liposome complex following insemination with sperm not treated with DNA/liposome complex, none of the embryos expressed d2EGFP. Sixteen day 4 ICSI embryos derived from sperm not treated with DNA/liposome complex were transferred into a day 3 recipient. One recipient delivered a female piglet with normal birthweight. After transfer of the ICSI embryos derived from sperm transfected with DNA/liposome complex, none of the four recipients maintained pregnancy.     

Epigenetic modification of histone 3 at lysine 9 in sheep zygotes and its relationship with DNA methylation

Background  

Previous studies indicated that, unlike mouse zygotes, sheep zygotes lacked the paternal DNA demethylation event. Another epigenetic mark, histone modification, especially at lysine 9 of histone 3 (H3K9), has been suggested to be mechanically linked to DNA methylation. In mouse zygotes, the absence of methylated H3K9 from the paternal pronucleus has been thought to attribute to the paternal DNA demethylation.