| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
McGill University and Montreal Children's Hospital Research Institute5
Departments of Pediatrics,6
Human Genetics,7 and
Pharmacology & Therapeutics,8McGill University, Montreal, Quebec, Canada H3H 1A1
Centre de Recherche en Reproduction Animale,9 Faculté de Médecine Vétérinaire, Université de Montréal, St-Hyacinthe, Quebec, Canada J2S 7C6
ABSTRACT
Findings from recent studies have suggested that the low survival rate of animals derived via somatic cell nuclear transfer (SCNT) may be in part due to epigenetic abnormalities brought about by this procedure. DNA methylation is an epigenetic modification of DNA that is implicated in the regulation of imprinted genes. Genes subject to genomic imprinting are expressed monoallelically in a parent of origin-dependent manner and are important for embryo growth, placental function, and neurobehavioral processes. The vast majority of imprinted genes have been studied in mice and humans. Herein, our objectives were to characterize the bovine SNRPN gene in gametes and to compare its methylation profile in in vivo-produced, in vitro-produced, and SCNT-derived Day 17 elongating embryos. A CpG island within the 5' region of SNRPN was identified and examined using bisulfite sequencing. SNRPN alleles were unmethylated in sperm, methylated in oocytes, and approximately 50% methylated in somatic samples. The examined SNRPN region appeared for the most part to be normally methylated in three in vivo-produced Day 17 embryos and in eight in vitro-produced Day 17 embryos examined, while alleles from Day 17 SCNT embryos were severely hypomethylated in seven of eight embryos. In this study, we showed that the SNRPN methylation profiles previously observed in mouse and human studies are also conserved in cattle. Moreover, SCNT-derived Day 17 elongating embryos were abnormally hypomethylated compared with in vivo-produced and in vitro-produced embryos, which in turn suggests that SCNT may lead to faulty reprogramming or maintenance of methylation imprints at this locus.
assisted reproductive technology, bovine, DNA methylation, early development, embryo, epigenetic, gamete biology, genomic imprinting, in vitro culture, in vitro fertilization, SCNT, somatic cell nuclear transfer, SNRPN
INTRODUCTION
Since Dolly, somatic cell nuclear transfer (SCNT) has been used to successfully "clone" several other species, including cows, pigs, goats, cats, mice, rats, rabbits, and horses (reviewed by Rhind et al. [1]). SCNT involves the transfer of a nucleus from a differentiated somatic cell into an enucleated oocyte, followed by oocyte activation, an in vitro culture period, and transfer of a preimplantation embryo to a recipient surrogate mother. In bypassing the normal course of fertilization, successful embryonic development of an SCNT embryo depends on the ability of the oocyte cytoplasm to reprogram the somatic cell nucleus back to an embryonic state. The low success rate of SCNT procedures across species and regardless of donor somatic cell nucleus reflects the inefficiency of this reprogramming process.
Most SCNT embryos die before birth, and of those that are born, many die during the perinatal period. In addition to the high incidence of prenatal and neonatal lethality, surviving SCNT-derived offspring have physiological problems as adults (reviewed by Wilmut et al. [2]). For example, mice produced using SCNT have been shown to be significantly larger than their normal littermates [3]. Placental abnormalities also appear to be common in SCNT-derived animals, especially in sheep, mice, and cattle [4–6].
In addition to pathological phenotypes, studies have examined the nature of the molecular reprogramming defects underlying the low viability of SCNT embryos (reviewed by Rhind et al. [1]). Abnormalities in epigenetic mechanisms are of particular interest because of the dynamic nature of DNA methylation and chromatin modification patterns during early embryogenesis and the requirement of their proper inheritance for normal development.
To date, the most widely investigated epigenetic modification is DNA methylation. The methylation of DNA at cytosine residues within CpG dinucleotides is associated with transcriptional repression and is implicated in maintaining genomic stability and in silencing repetitive elements; it is the best understood epigenetic mark regulating genomic imprinting. Imprinted genes are unique in that they are expressed exclusively from only one parental allele. DNA methylation is essential in facilitating this monoallelic expression by allowing one allele to be distinguished from the other. Approximately 75 imprinted genes have been identified to date in humans and mice, and many play essential roles in regulating fetal growth, placental function, and postnatal behavior, and several are linked to human disease (reviewed by Reik and Walter [7]).
The paternally expressed bicistronic gene SNURF-SNRPN, hereafter referred to as SNRPN, encodes a splicing factor and is among the best studied imprinted genes to date. In mice and humans, its imprinting has been shown to be controlled by a well-characterized differentially methylated region (DMR), with SNRPN acquiring a maternally derived methylation imprint during oogenesis [8, 9]. Methylation abnormalities within the SNRPN DMR have been linked to the pathogenesis of Prader-Willi and Angelman syndromes, two neurodevelopmental disorders that affect children. Recently, findings from studies have suggested an increased incidence of Angelman syndrome and Beckwith Wiedemann syndrome (another imprinting disorder) in children conceived using assisted reproductive technologies; in each case in which it was tested, the two diseases were associated with decreased levels of maternal allele methylation (reviewed by Gosden et al. [10] and by Lucifero et al. [11]).
Abnormalities in global and repetitive element DNA methylation patterns have been well described in SCNT-derived embryos [12–16]. In mice, abnormalities in imprinted gene expression [17], DMR methylation [17], and DNA methyltransferase expression and localization [18] have been observed after SCNT. However, in cattle an understanding of the consequence of SCNT on imprinted gene methylation and regulation is lacking, perhaps because of the challenges involved in characterizing appropriate sequences. In this study, our objectives were twofold. First, we were interested in identifying the bovine SNRPN region corresponding to the mouse and human DMR and in determining whether its methylation profile was conserved across additional species. Second, we wished to investigate the effects of in vitro production and SCNT on the DNA methylation status of SNRPN in individual bovine Day 17 elongating embryos.
MATERIALS AND METHODS
Establishment of Nuclear Donor Cell Cultures
Fetal fibroblast cell cultures were established from a 60-day-old crossbred fetus produced by artificial insemination (AI) of a Holstein (Bos taurus) heifer with semen from a Nelore (Bos indicus) bull. Fetal tissues were minced and digested with 0.25% trypsin and 0.02% EDTA (Gibco BRL, Burlington, ON, Canada) at 37°C for 10 min. Isolated cells were washed and cultured in Dulbecco modified Eagle medium (DMEM; Gibco BRL) supplemented with 10% fetal bovine serum (FBS; Gibco BRL) and 0.5% antibiotics (penicillin 10 000 U ml–1–streptomycin 10 000 µg ml–1; Gibco BRL) at 37°C in 5% CO2. When the cultures were confluent, primary passage cells were frozen in culture media supplemented with 10% dimethyl sulfoxide and stored in liquid nitrogen.
Production of Host Oocytes
Ovaries were collected from a local abattoir and transported to the laboratory in saline at 30–35°C within 2 h after the slaughter. Follicles with diameters between 2 and 8 mm were punctured with a 19-gauge needle, and cumulus-oocyte-complexes (COCs) with several layers of cumulus cells and homogeneous oocyte cytoplasm were washed in Hepes-buffered tissue culture medium (TCM-199; Gibco BRL) supplemented with 10% (vol/vol) FBS. Groups of 20 COCs were placed in 100 µl of bicarbonate-buffered TCM-199 supplemented with 10% FBS, 50 µg ml–1 LH (Ayerst, London, ON, Canada), 0.5 µg ml–1 FSH (Folltropin-V; Vetrepharm, St-Laurent, PQ, Canada), 1 µg ml–1 estradiol 17-ß (Sigma-Aldrich, St. Louis, MO), 22 µg ml–1 pyruvate (Sigma-Aldrich), and 50 µg ml–1 gentamicin (Sigma-Aldrich). After 20–24 h of in vitro maturation, cumulus cells were removed by vortexing the COCs in PBS and 0.2% hyaluronidase (Sigma-Aldrich) and were selected for the presence of a polar body.
Production of In Vivo Embryos by AI
Nonlactating Holstein cows were superovulated by intramuscular injection of Folltropin-V given every 12 h in decreasing doses starting at Day 9–10 of the estrous cycle (Day 0 = estrous). Cows received an injection of 500 µg of cloprostenol (Estrumate; Schering-Plough Animal Health, Pointe-Claire, QC, Canada) and were artificially inseminated at 52 h and 86 h after the initiation of superovulation [19]. Semen used for the production of the Day 60 fetus for the donor cell line and for AI, as well as the in vitro-produced embryos, was from the same sire. Experiments were performed in compliance with the guidelines set by the Canadian Council for Animal Care.
In Vitro-Produced Embryos
In vitro-matured oocytes were fertilized in vitro using standard protocols [20]. Briefly, 20–25 COCs were placed in 100-µl drops of Tyrode medium supplemented with 0.6% BSA (fraction V; Sigma-Aldrich), lactate, pyruvate, gentamicin, and heparin (10 µg/ml). Frozen-thawed spermatozoa were washed and centrifuged through a Percoll gradient and diluted to 106 live spermatozoa/ml. At 20 h following the start of incubation with spermatozoa, COCs were denuded of cumulus cells by brief shaking, and the presumptive zygotes were transferred to 50-µl drops of synthetic oviduct fluid containing amino acids at oviductal fluid concentrations (SOF modified medium [21]) and cultured for 8 days under the same conditions used for the SCNT embryos.
Embryo Reconstruction by Nuclear Transfer
Selected oocytes were placed in PBS containing 7.5 µg ml–1 cytochalasin B (Sigma-Aldrich) and enucleated by removing a small fraction of the cytoplasm surrounding the first polar body. After microsurgery, oocytes were placed in medium containing 5 µg ml–1 Hoechst 33342 for 15 min and then briefly exposed to ultraviolet light to confirm the complete absence of chromatin. Nuclear donor cells were thawed at 37°C, immediately washed in 10 ml of culture media (DMEM, supplemented with 10% FBS and 0.5% antibiotics), and then placed in culture using the same media. Nuclear transfer was performed using confluent cells that were maintained in culture for 2–5 passages. A single nuclear donor cell was introduced into the perivitelline space of the enucleated oocyte, and the resulting couplet was placed in a 0.3 M mannitol solution containing 0.1 mM MgSO4 and 0.05 mM CaCl2 and exposed to a 1.5-kV electric pulse lasting 70 µsec. After electrical stimulation, couplets were washed and cultured in 50-µl drops of SOF modified medium [21] supplemented with 0.8% BSA-V fatty acid free (Sigma-Aldrich) under equilibrated mineral oil at 39°C in a humidified atmosphere of 5% CO2 and 5% O2. After 1–2 h in culture, couplets were examined to determine fusion and were exposed to 5 µM ionomycin (Sigma-Aldrich) for 4 min to induce oocyte activation and were then replaced in culture for an additional 8 days.
Production of Day 17 Elongating Embryos
The estrous cycle of Holstein heifers was synchronized by the injection of 500 µg of the prostaglandin F2
analogue cloprostenol (Estrumate). Six to eight days after the standing heat, Day 8 in vitro-produced or reconstructed blastocysts were transferred to the uterine horn ipsilaterally to the presence of the ovary corpus luteum. Embryos were washed with TCM-199 Hepes-buffered medium supplemented with 20% of FBS, loaded into a 250-µl straw, and transferred. Five to fifteen embryos were transferred to each recipient female. At Day 17 after embryo reconstruction, the transferred embryos were nonsurgically recovered by flushing the uterus of the recipient females with PBS using a Foley catheter. Embryos were separated from the flushing media and microscopically inspected to separate those that were recovered intact. After selection, the embryos were washed three times in PBS and individually frozen to –70°C in 0.2 ml of distilled water. Development to the blastocyst stage was 32% for SCNT embryos (based on the total number of fused or reconstructed embryos) and 35% for in vitro-produced embryos (based on the total number of oocytes placed in in vitro fertilization). The survival rate of the embryos produced by SCNT with the fetal fibroblast cells was 11% (3 calves from 28 transferred blastocysts) (Bordignon and Smith, unpublished results). Only those embryos that were recovered intact were used for the experiments. The mean length of intact Day 17 embryos was 5.36 ± 1.03 cm for SCNT and 3.67 ± 0.22 cm for in vitro-produced embryos. Day 17 elongating bovine embryos are composed mostly of trophectoderm-derived tissue. The embryonic disc could not be definitively identified in the intact Day 17 embryos used for the DNA methylation studies.
Search for the Bovine SNRPN DMR
Genome walking was used to identify the position of the DMR of the bovine SNRPN gene. Total genomic DNA was extracted from fibroblasts of a Day 60 B. taurus fetus using a DNeasy Tissue Kit (Qiagen). Four DNA libraries were obtained using the Universal Genome Walker Kit (Clontech), which requires two nested PCR reactions per library. Briefly, the primers for the PCR reaction consisted of adaptor primers provided with the kit and gene-specific primers. The Snrpn DMR is localized around exon 1 in most species analyzed to date. Therefore, we designed our gene-specific primers in exon 2 using sequences obtained from GenBank (AF101040) and moved in the 5' direction. The protocol produced a series of four fragments (4.5, 0.4, 1.5, and 0.6 kb) that covered the entire intron 1, exon 1, and part of the 5'-promoter region. Each PCR fragment was cloned (pGEM-T easy) and sequenced for analysis. The cloned nucleotide sequence matches sequence data now available in the B. taurus 6x Ensembl assembly (gene ID ENSBTAG00000002692).
DNA Isolation and Bisulfite Sequencing
DNA was isolated using Trizol reagent (Invitrogen) according to the manufacturer's instructions. Approximately 100 ng of embryonic, sperm, or somatic cell DNA was digested using HindIII (Gibco BRL), and bisulfite treatment was carried out as previously described [22, 23]. In the case of oocytes, approximately 300 germinal vesicle (GV) stage oocytes were collected and pronase treated to ensure complete removal of any somatic cells. Oocyte DNA was isolated and bisulfite treated as previously described [22]. Primers specific for bisulfite-converted DNA for SNRPN were designed. Because the bisulfite protocol used was optimized for small amounts of DNA (
100 ng) and bisulfite treatment at pH 5.0 leads to high levels of DNA degradation, nested PCR amplification was necessary. Primers were designed according to bisulfite standards (no CpG sites within primers and 
2 cytosines within primer sequence to select for converted sequences) without the help of a computer program. For the outside nested PCR, the primer sequences were as follows: Forward 5'-GGAAAGTTTGAGGAAATTTGAATAAGG-3'; Reverse 5'-CAAATACCCCCAAAACCTAACAAAAC-3'. The primers used for the inside nested reaction were as follows: Forward 5'-TTGGGAGGTATTATTTTGGGTTGAAG-3'; Reverse 5'-AAAAAATCAATCCAACCCCAAACCTC-3'. Each 25-µl PCR reaction contained 4 µl of bisulfite-treated DNA, 25 ng of each primer, 2.5 µl (100 µM) deoxynucleotide triphosphates (Invitrogen), 5 µl 5X PCR buffer (300 mM Tris-HCl, 7.5 mM ammonium sulfate, and 12.5 mM MgCl2 [pH 8.5]) (Invitrogen), and 1.25 U of DNA Taq polymerase (Invitrogen). First-round PCR was performed under the following conditions: 4 min at 94°C, 2 min at 55°C, and 2 min at 72°C for two cycles, followed by 35 cycles of PCR consisting of 1 min at 94°C, 2 min at 55°C, and 2 min at 72°C. For the second round of PCR, 4 µl of the first-round sample were used, and the conditions for the PCR were the same as the first-round conditions, except that the first two cycles were omitted. Clones containing the appropriate inserts were sequenced using an ABI 310 sequencer. Sodium bisulfite converts all unmethylated cytosines, whether or not they are in CpG dinucleotides to guanines. Only sequences with greater than 95% bisulfite conversion efficiency were used for analysis (i.e., to avoid false overestimation of methylated CpGs). Sequence differences (polymorphisms) between clones with similar CpG methylation profiles were verified to ensure that unique clones were represented. We examined 39 CpG sites in a 548-bp fragment of SNRPN. Absence of strain-specific single nucleotide polymorphisms prevented the parental origin of the sequenced strands from being determined.
RESULTS
Characterization of the Bovine Putative SNRPN DMR
A bovine genomic library was created and used to amplify and obtain the 5' SNRPN sequence. Similar to the syntenic mouse and human SNRPN promoter sequences, the bovine upstream sequence obtained by genome walking was CpG rich and was found to be homologous to the sequence upstream of SNRPN available in the B. taurus 6x Ensembl assembly. The bovine SNRPN gene has 10 exons and spans approximately 23 kb (Fig. 1a). CpGPlot analysis (http://www.ebi.ac.uk/emboss/cpgplot/) of the full Ensembl genomic sequence showed that the region surrounding exon 1 is marked by three CpG islands, the first of which is contained within the region we analyzed using bisulfite sequencing (Fig. 1a).
|
To determine whether the bovine methylation profile was conserved with that seen in mouse and human, bisulfite sequencing was used to investigate the methylation status of a number of CpG sites within the upstream region we initially obtained. Our primer design allowed us to determine the methylation status of 39 CpG sites spanning 548 bp (Fig. 1a). This approach also permitted us to simultaneously investigate methylation patterns within promoter, exon 1, and intron 1 sequences.
Alignment of these 548 bp with the homologous mouse and human sequences showed several areas of exact identity, with the most obvious spanning the sequence of exon 1, as would be expected (Fig. 1b). Several CpGs (13, 14, 26, 28, 30, 33, and 35) appear to be conserved across all three species, which suggests that these sites may have evolutionarily conserved functional significance. With this reasoning, the bovine sequence was analyzed for the presence of transcription factor binding sites using Transcription Element Search System, or TESS (http://www.cbil.upenn.edu/tess/). Possible binding sites for transcription factors with previously identified sensitivity to CpG methylation such as YY1 [24] and SP1 [25] are shown in Figure 1b. However, neither of these factors had binding sites that localized to CpG sites conserved across the three species.
SNRPN DNA Methylation Profiles in Gametes, Somatic Cells, and AI Embryos
We first investigated the methylation patterns of CpG island 1 in GV oocytes and sperm (Fig. 2, a and b). In keeping with mouse and human profiles, alleles amplified from bovine GV oocytes were highly methylated (88%, defined as percentage strands with >50% CpGs methylated) (Table 1). In contrast, sperm alleles were completely unmethylated (0%), suggesting that the region analyzed may be within the bovine SNRPN DMR. Because we could not definitively identify the parental alleles in embryos, we will refer to the bovine SNRPN region we examined as the "putative" SNRPN DMR.
|
|
Before our analysis of the putative SNRPN DMR methylation in in vitro-produced and SCNT Day 17 embryos, we examined the methylation profile of three Day 17 in vivo-produced embryos conceived via AI. In keeping with the presence of maternal and paternal alleles, approximately half of the alleles were hypermethylated (Fig. 3a [i–iii] and Table 1). Further supporting its imprinting status in the bovine model, single-stranded conformational polymorphism analysis was used to show SNRPN monoallelic expression from the paternal allele in two in vivo-produced Day 17 embryos (Supplementary Figure 1, available online at http://www.biolreprod.org). We also determined the methylation profile of a fetal liver sample that was taken from an in vivo-produced Day 60 fetus and again found approximately half of the strands to be methylated (Fig. 3b and Table 1).
|
The methylation profile of SNRPN alleles in additional somatic cell samples was investigated. Bisulfite sequencing results for adult liver (data not shown) and fetal fibroblast samples (Fig. 3c and Table 1) showed approximately half of the SNRPN alleles to be highly methylated. This further suggested the CpG-rich SNRPN sequence amplified to be a putative DMR, as it appeared to retain the germline-derived differential methylation profile in somatic cells. The fetal fibroblast sample analyzed was the source of the nuclear donor cells used to create the SCNT embryos in this study.
Because parental-specific single nucleotide polymorphisms (which would have allowed distinction of the parental alleles and confirmation of the sequence as a DMR) were not found, Fisher exact text was used to indicate that all somatic cell results shown were not significantly different than 50%, the expected percentage of alleles methylated (Table 1).
SNRPN Methylation Profile in In Vitro-Produced Embryos
Having identified the putative SNRPN bovine DMR, we were interested in determining its methylation profile in Day 17 in vitro-produced embryos. DNA from eight individual embryos was isolated and subjected to bisulfite mutagenesis. Results for four embryos were similar to the patterns seen for the somatic and in vivo embryo samples analyzed in that approximately half of the alleles sequenced were methylated (Fig. 4, a–d). We saw somewhat lower than expected frequencies of methylated alleles in the remaining four of these eight embryos (Fig. 4, e–h). Overall, the percentage of strands with greater than 50% of CpGs methylated ranged from 21% to 46% (Table 1).
|
SNRPN Methylation Profile in SCNT Embryos
While our analysis of in vitro-produced bovine preimplantation embryos showed methylation imprint patterns inherited from the gametes to be maintained for the most part in the embryos examined, we were also interested in determining whether SCNT affected the putative methylation imprint on SNRPN. Again, embryos were analyzed at Day 17, and DNA methylation analysis was carried out on individual embryos. Of eight SCNT embryos that were analyzed, seven were dramatically hypomethylated (Fig. 5, j–p and Table 1). Only one SCNT embryo appeared to have a "normal" methylation profile (Fig. 5i), suggesting that SCNT can affect SNRPN methylation in SCNT-derived bovine Day 17 elongating embryos.
|
DISCUSSION
In this study, we identified a putative DMR within a 5' CpG island of the bovine SNRPN gene and assessed its methylation status in gametes, in somatic cells, and in single Day 17 elongating embryos using bisulfite sequencing. In addition to looking at the effect of in vitro production on SNRPN methylation, we addressed whether reprogramming abnormalities in SCNT-derived embryos include DNA methylation imprint defects at this locus.
Characterization of a Putative Bovine SNRPN DMR
With the exception of a small number of genes, our knowledge of imprinted gene regulation is restricted to two species, mouse and human. Although interest is growing as a result of concerns with regard to the effects of in vitro culture and SCNT on epigenetic reprogramming, in ruminants little is known about imprinting mechanisms. IGF2R, a maternally expressed gene that inherits a methylation imprint from the oocyte [26], has been characterized in sheep with large offspring syndrome; the syndrome is associated with methylation abnormalities within the IGF2R DMR [27]. The imprinting status of H19, a maternally expressed and paternally methylated gene, has also been identified in sheep and cattle [28, 29].
Our interest in investigating SNRPN stems from its having a well-defined DMR that not only controls its imprinting but also regulates a number of other imprinted genes clustered on mouse central chromosome 7 and on human chromosome 15q11–13 [30]. Methylation abnormalities within the SNRPN DMR are associated with Angelman and Prader-Willi syndromes, the former syndrome putatively linked to assisted reproductive technologies or infertility [10]. Mouse investigations of the effects of in vitro culture on imprinted gene expression revealed SNRPN expression to be unperturbed compared with other imprinted genes [31]. However, studies in mice looking at the effects of in vitro culture on other imprinted genes such as H19 found abnormalities in imprinted expression and in DMR methylation [31, 32]. These studies highlight the need for additional animal models to examine the potential of in vitro production of embryos to perturb imprinting.
As in the case of H19 and IGF2R in ruminants, we found the SNRPN methylation imprint in cattle to be well conserved with DNA methylation profiles previously observed in mouse. In the mouse, the methylation of Snrpn required for functional imprinting (i.e., monoallelic expression) in the offspring is acquired during oocyte growth [9, 33]. Two studies on the timing of DNA methylation of SNRPN in human oocytes using bisulfite sequencing reached different conclusions. In the earlier study using unfertilized oocytes from a fertility center, the SNRPN DMR was found to be largely unmethylated in oocytes, leading the authors to suggest that the SNRPN methylation imprint is established after fertilization [34]. A later study was conducted using GV stage, metaphase I, and metaphase II oocytes that were unsuitable for transfer and were donated by patients for research purposes; the SNRPN DMR was found to be highly methylated in all oocytes examined [8]. The results of the latter study fit well with the earlier data from the mouse; however, because of limitations in collecting sufficient numbers of healthy human oocytes, more studies are needed to clarify the timing of methylation imprint establishment in human oocytes. Therefore, our finding that the bovine putative SNRPN DMR is methylated in GV oocytes agrees with mouse and some human findings and provides additional evidence for the importance of the postnatal oocyte growth phase in methylation imprint establishment. The bovine results will need to be extended to additional imprinted genes and times during oocyte development, as has been done in the mouse. Because the SNRPN methylation profile in cattle appears to be conserved compared with that seen in other species, it provides an additional tool for studies geared toward understanding the mechanisms of imprinted gene regulation and will allow for a more thorough evaluation of vulnerability to in vitro technologies and SCNT.
SNRPN DNA Methylation Profile in In Vitro-Produced and SCNT Embryos
While global methylation profiles in bovine in vitro-cultured and SCNT-derived preimplantation embryos have been investigated, detailed methylation profiling at the single gene level has been impeded by the lack of bovine genomic sequence. Having obtained and characterized a putative SNRPN DMR and using an approach that enabled us to look at 39 CpG sites within a region spanning the promoter, first exon, and first intron of the gene, we were interested in determining whether in vitro production or SCNT affected the maintenance of the SNRPN imprint.
Investigations have indicated that the methylation of imprinted genes is maintained during the zygote to blastocyst window at a time when much of the rest of the genome becomes demethylated [7]. To determine if methylation was maintained or recapitulated in in vitro-produced and SCNT embryos, we were interested in examining embryos at an early stage of development. We chose Day 17 elongating embryos as a time when there were sufficient numbers of cells for accurate analysis of DNA methylation by bisulfite sequencing in individual embryos. For accurate assessment of methylated cytosines in a given sample, bisulfite treatment must be titrated to convert a high percentage (>95%) of the unmethylated cytosines in the genome to uracil; during the process, it has been estimated that 84%–96% of the starting DNA is degraded following bisulfite treatment [35]. Intact Day 17 bovine embryos could be recuperated before implantation by uterine flushing. An important caveat for the interpretation of our results is the fact that the Day 17 bovine embryo is mostly composed of extraembryonic tissue. However, previous studies have indicated that the methylation of at least some imprinted genes, including Snrpn in mouse [36] and IGF2R in sheep [27], is maintained in the extraembryonic tissues, at least at an early stage of development.
Half of the Day 17 in vitro-produced elongating embryos showed approximately 50% methylation of SNRPN. However, in the additional four Day 17 in vitro-produced bovine embryos examined, SNRPN was somewhat hypomethylated. This finding could reflect the possibility that methylation imprints do not need to be maintained as stringently on imprinted genes in the extraembryonic tissues compared with the embryo. This suggestion is supported by data from midgestation mouse embryos in which the Snrpn and H19 methylation was found to be lower (but not dramatically so) in some placentas compared with the corresponding embryos [36]. Alternatively, it is possible that techniques associated with in vitro production may adversely affect bovine embryos. Examination of embryos and placentas at later stages of gestation would be useful in helping to distinguish between these possibilities, keeping in mind the fact that those embryos with abnormal imprints may die during gestation.
Multiple studies have indicated that one of the underlying problems with SCNT is the inefficiency of the somatic nucleus to be correctly epigenetically reprogrammed. Genomewide abnormalities in DNA methylation patterns or cytosine methylation levels after SCNT have been observed in mice, cattle, and sheep [12, 13, 15, 16, 37]. For the most part, analysis of imprinted gene expression and DNA methylation profiles in SCNT-derived embryos has been limited to studies in mice [15, 17, 31, 38, 39]. Mann et al. [17] found multiple imprinted genes to be aberrantly expressed in SCNT-derived mouse blastocysts. They also showed parental allele-specific loss of DNA methylation at the Snrpn and H19 DMRs in samples of pooled SCNT embryos [17].
In sheep produced by SCNT, analysis of IGF2, IGF2R, and H19 expression and DNA methylation status using a restriction enzyme-based approach found IGF2R DMR2 methylation to be most severely affected by SCNT, with only one of 13 SCNT-derived lambs exhibiting loss of methylation at H19 [28]. Investigations into imprinting status after SCNT in cattle have focused mainly on expression analysis. One such study found H19 to be biallelically expressed in three of four SCNT-derived calves that died during the perinatal period because of physiological abnormalities [29]. Dindot et al. [40] examined the expression profiles of XIST, IGF2, and GTL2 in Day 40 SCNT fetuses and placental tissue and showed that only XIST expression in the chorion was dysregulated.
Herein, we determined the DNA methylation profile of the putative SNRPN DMR in individual Day 17 SCNT-derived elongating embryos. This approach allowed us to assess and compare DNA methylation profiles at the single conceptus level. Of the eight SCNT embryos examined, seven were strikingly hypomethylated, suggesting that SCNT may adversely affect DNA methylation imprints. This result is in agreement with the mouse SCNT study in which hypomethylation at the Snrpn DMR was also observed [17]. As already suggested for the in vitro-produced embryos, future studies on SCNT embryos at a later stage of gestation, in which DNA methylation analyses of the embryo proper and extraembryonic tissue are carried out separately, should allow for the examination of lineage-specific vulnerabilities in methylation imprint maintenance after SCNT.
Taken together, findings from DNA methylation studies on bovine embryos suggest regional hypermethylation after SCNT [13, 14], while sequence-specific investigations generally show hypomethylation of various genes [41 and the present study]. Paradoxically, this finding of global hypermethylation and single-copy gene hypomethylation contrasts with the changes seen in cancer cells in which global hypomethylation and tumor suppressor gene hypermethylation are predominantly observed [42]. This suggests that the deregulation of the enzymes responsible for these contrasting observations occurs in opposite directions in SCNT embryos and in cancer cells.
In addition to correlating the DNA methylation abnormalities that we observed with imprinted expression profiles, future studies should be directed at trying to understand the mechanisms underlying the loss of DNA methylation seen in SCNT-derived bovine embryos. After mouse SCNT, abnormalities have been described in the expression and localization of the major maintenance DNA methyltransferase DNMT1o in preimplantation embryos [18]. Maintenance and de novo DNA methyltransferases have been identified and found to be expressed in bovine preimplantation embryos [43], and their dysregulation after in vitro production or SCNT may explain the loss of methylation observed [44]. A further important goal will be to characterize other imprinted genes in cattle and to identify their DMRs. This will provide additional tools to help us understand the mechanisms underlying imprinted gene regulation and determine whether in vitro culture and SCNT result in specieswide imprinting errors.
FOOTNOTES
1 Supported by the Fonds de la Recherche en Santé du Québec (FRSQ) and by the Natural Sciences and Engineering Research Council of Canada. This work was part of the Program on Oocyte Health (http://www.ohri.ca/oocyte) funded by grant HGG62293 under the Healthy Gametes and Great Embryos Strategic Initiative of the Canadian Institutes of Health Research (CIHR) Institute of Human Development, Child and Youth Health. D.L. is the recipient of FRSQ and CIHR studentships and a European Molecular Biology Organization long-term fellowship. J.M.T. is a William Dawson Scholar of McGill University and a Scholar of the FRSQ. 
2 Correspondence: Jacquetta M. Trasler, McGill University and Montreal Children's Hospital Research Institute, 2300 Tupper Street, Montreal, Quebec, Canada H3H 1P3. FAX: 514 412 4331; jacquetta.trasler{at}mcgill.ca
3 Current address: Laboratory of Developmental Genetics and Imprinting, The Babraham Institute, Cambridge CB2 4AT, United Kingdom.
4 Current address: Department of Animal Science, McGill University, Macdonald Campus, Ste-Anne-de-Bellevue, Quebec, Canada H9X 3V9.
Received: 13 February 2006.
First decision: 3 March 2006.
Accepted: 13 June 2006.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
