DNA methylation and functional characterization of the XIST gene during in vitro early embryo development in cattle

ABSTRACT

XIST, in association with the shorter ncRNA RepA, are essential for the initiation of X chromosome inactivation (XCI) in mice. The molecular mechanisms controlling XIST and RepA expression are well characterized in that specie. However, little is known in livestock. We aimed to characterize the DNA methylation status along the 5’ portion of XIST and to characterize its transcriptional profile during early development in cattle. Three genomic regions of XIST named here as promoter, RepA and DMR1 had their DNA methylation status characterized in gametes and embryos. Expression profile of XIST was evaluated, including sense and antisense transcription. Oocytes showed higher levels of methylation than spermatozoa that was demethylated. DMR1 was hypermethylated throughout oogenesis. At the 8–16-cell embryo stage DMR1 was completed demethylated. Interestingly, RepA gain methylation during oocyte maturation and was demethylated at the blastocyst stage, later than DMR1. These results suggest that DMR1 and RepA are transient differentially methylated regions in cattle. XIST RNA was detected in matured oocytes and in single cells from the 2-cell to the morula stage, confirming the presence of maternal and embryonic transcripts. Sense and antisense transcripts were detected along the XIST in blastocyst. In silico analysis identified 63 novel transcript candidates at bovine XIST locus from both the plus and minus strands. Taking together these results improve our understanding of the molecular mechanisms involved in XCI initiation in cattle. This information may be useful for the improvement of assisted reproductive technologies in livestock considering that in vitro conditions may impair epigenetic reprogramming.

Introduction

In mammals X chromosome inactivation (XCI) occurs during early embryo development and is an evolutionary strategy chosen to compensate the imbalance in gene production between females and males [1,2]. In this process, one of the X chromosomes in the female cells is randomly inactivated [1] and the established pattern of inactivation is maintained in the daughter cells after mitosis, generating a mosaicism in female adult tissues [2]. In mice, XCI is imprinted in extraembryonic annexes, in which the paternal X chromosome is preferentially inactivated [3]. However, in the inner cell mass, the process of inactivation is random [4,5]. Nevertheless, in other species, such humans and rabbits, XCI is exclusively random [5].

Many factors are involved in XCI, including ncRNA, protein complexes and pluripotency genes [6,7]. In mice, the X-inactive specific transcript (XIST) is one of the most important factors for the initiation of XCI [6,8]. The XIST gene is located in the X inactivation centre (XIC) locus and acts in cis along the chromosome to be silenced, recruiting enzymatic complexes, such as PRC2, which are responsible for the establishment of the H3K27me3 histone repressive mark [6,9,10]. The DNA methylation pattern in the XIST promoter has been considered the trigger to XIST expression in mouse embryos [11]. In cattle, however, the differentially methylated region presented in exon 1 (DMR1) has been associated with the regulation of XIST expression [12,13].

XIST has been subjected to alternative splicing, producing some different isoforms in a tissue- or sex-specific manner [14,15]. Its genomic structure is complex, possessing six groups (A to F) of tandem repeats regions along the gene, which are conserved among different species of mammals [16]. Repeat A (RepA) seems to be the most important repeat for the mechanism of initiation and maintenance of X inactivation [17]. RepA is located in XIST exon 1 and produces a short RNA transcript that is responsible for distributing the XIST RNA in cis along the X chromosome to be inactivated through the association of its transcript with the PRC2 complex [17,18].

In addition to XIST, XIC also produces an XIST antisense ncRNA named TSIX [19]. Its transcription is carried out exclusively by the active X chromosome [19] and has a XIST antagonist transcription pattern [19]. In mice, TSIX expression is responsible for enriching the XIST promoter with heterochromatin factors on the future active X chromosome [20,21]. Despite its importance in mice, TSIX is considered a pseudogene in many species, such as the cow, which means that its structure is not conserved among mammals [12,22].

The process of XCI is well understood in mice, but very little is known about this process in other species, especially livestock. Due to the relevance of assisted reproductive technologies (ART) for livestock production and considering that XCI is a mechanism essentially controlled by epigenetic factors occurring during the initial embryogenesis, improving our understanding of this process in domestic animals is key in order to improve the efficiency of ARTs. Cycles of epigenetic reprogramming and imprinting establishment occur during gametogenesis [23,24] and initial embryogenesis [23,25]. XCI is established in this window of development. Thus, accurate epigenetic reprogramming is directly correlated to oocyte [26] and embryo quality [27]. In this sense, we aimed to characterize the molecular profile of the bovine XIST locus during early development, evaluating DNA methylation, gene expression, and antisense transcription.

Results

DNA methylation profile of XIST DMR1 during oogenesis

The methylation profile of the XIST DMR1 in oocytes during bovine oogenesis is shown in Figure 1. Immature oocytes of all categories of follicles showed a hypermethylated pattern and no significant differences in the DNA methylation percentage were found among the groups.

Figure 1.

Dynamics of DNA methylation of XIST DMR1 during bovine oogenesis. DNA methylation pattern of immature oocytes from primordial, final secondary, small antral, and large antral follicles. Each line represents one individual clone and each circle represents one CpG dinucleotide (17 CpGs). White circles represent unmethylated CpGs, filled black circles represent methylated CpGs and grey circles represent a CpG that could not be analysed. Methylation percentage is shown in bars graphs (mean ± SEM). Letter ‘a’ for all analysed groups indicates no different means (p ≥ 0.05).

DNA methylation profile of XIST DMR1 in matured gametes and initial embryogenesis

The XIST DMR1 methylation pattern was determined in mature gametes as well as in embryos and the placenta (Figure 2). Different methylation patterns were found for gametes, with sperm showing a hypomethylated pattern (3.33 ± 1.05%) compared to immature and MII oocytes (79.54 ± 6.56 and 89.59 ± 2.31, respectively), which showed a hypermethylated pattern (p < 0.001). Moreover, the DMR1 was found to be already completely demethylated in initial embryogenesis, as observed for 8–16-cell embryos. At least up until the blastocyst stage, this DMR did not start reprogramming, showing hypomethylation in morula (0.84 ± 0.57%), and in ICM (1.07 ± 0.72%) and TE (3.93 ± 1.24%) cells. The hypomethylated pattern found in trophectoderm cells, however, was not maintained in allantochorion, in which a significant increase of DNA methylation was observed throughout gestation (Figure 2).

Figure 2.

Dynamics of DNA methylation of XIST DMR1 in gametes and embryos of cattle. DNA methylation pattern of sperm, immature and matured oocytes, embryos in different stages (8–16-cell, morula and inner cell mass and trophectoderm cells of blastocyst) and allantochorion of foetal placenta. Each line represents one individual clone and each circle represents one CpG dinucleotide (17 CpGs). White circles represent unmethylated CpGs, filled black circles represent methylated CpGs and grey circles represent a CpG that could not be analysed. Methylation percentage is shown in bars graphs (mean ± SEM). Different letters indicate different means (p < 0.05).

DNA methylation profile of RepA in matured gametes and initial embryogenesis

RepA methylation was also analysed in the gametes, embryos, and placenta (Figure 3). As observed in DMR1, RepA also presented different methylation patterns between the sperm and MII oocytes (10.26 ± 3.73% and 74.27 ± 8.77%, respectively; p < 0.001), but not for immature oocytes (12.29 ± 4.21%) as found in DMR1. Moreover, DNA methylation was found to increase in oocytes during maturation (12.29 ± 4.21% and 74.27 ± 8.77% for immature and MII oocytes, respectively; p < 0.001). In contrast to DMR1, RepA lost methylation only at the blastocyst stage, with 8–16-cell embryos (92.18 ± 2.22%) and with the morula cells (95.33 ± 0.51%) hypermethylated and the ICM (1.97 ± 1.41%) and TE cells (0.00 ± 0.00%) becoming completely demethylated. However, a hypermethylated pattern was re-established in allantochorion, except for in one female placenta sample (26.92 ± 11.27 and 89.24 ± 2.50).

Figure 3.

Dynamics of DNA methylation of RepA in gametes and embryos of cattle. DNA methylation pattern of spermatozoa, immature and matured oocytes, embryos in different stages (8–16-cell, morula and inner cell mass and trophectoderm of blastocysts) and allantochorion of foetal placenta. Each line represents one individual clone and each circle represents one CpG dinucleotide (18 CpGs). White circles represent unmethylated CpGs, filled black circles represent methylated CpGs and grey circles represent a CpG that could not be analysed. Methylation percentage is shown in bars graphs (mean ± SEM). Different letters indicate different means (p < 0.05).

DNA methylation profile of the XIST promoter in matured gametes

DNA methylation of the XIST promoter was analysed in the sperm and MII oocytes (Figure 4). As shown for DMR1 and RepA, the methylation pattern for this region is also different between sperm (3.12 ± 1.68%) and MII oocytes (46.67 ± 10.58%) (p = 0.008).

Figure 4.

Dynamics of DNA methylation of XIST promoter in bovine gametes. DNA methylation pattern of sperm and matured oocytes. Each line represents one individual clone and each circle represents one CpG dinucleotide (six CpGs). White circles represent unmethylated CpGs, and filled black circles represent methylated CpGs. Methylation percentage is shown in bars graphs (mean ± SEM). Different letters indicate different means (p < 0.05).

Single-cell expression of the XIST locus during early development

Single MII oocytes and single blastomeres of female embryos were used to evaluate gene transcription in the XIST locus. For this, 5 single oocytes, 10 single blastomeres from 2-cell embryos, 15 single blastomeres from 4-cell embryos, 24 single blastomeres from 8–16-cell embryos, and 17 single blastomeres from 2 morula were analysed. The results are shown in Figure 5. All oocytes and blastomeres from 2-cell embryos presented both XIST and GAPDH expression. In the 4-cell embryos, three of five embryos presented XIST and GAPDH expression in all cells, while two embryos showed some GAPDH⁻ blastomeres. Two 8–16-cell embryos also showed some GAPDH⁻ blastomeres. Seventeen individual cells from morula were analysed and 82.3% of them were XIST⁺ (Figure 5).

Figure 5.

XIST expression in single oocytes and blastomeres from female bovine embryos of different stages of development. a – Individual oocytes and blastomeres of embryos in different stages (2-cell, 4-cell, 8–16-cell and morula) were analysed. XIST and/or GAPDH positive cells were coloured represented according to the figure legend. Circles representing individual cells of 8–16-cell embryos and morula are not representing the actual number of blastomeres analysed, but the colour pattern is representing the actual proportion of cells XIST⁺/GAPDH⁺. b-GAPDH e XIST expression in MII oocytes and individual blastomeres of different bovine embryo stages (2-cell, 4-cell, 8–16-cell, and morula). GAPDH⁺ – cells expressing GAPDH; XIST⁺ – cells expressing XIST.

Strand-specific expression throughout the XIST locus

Expanded blastocysts and testicular tissue were used to search for sense and antisense transcription throughout the bovine XIST locus. Both sense and antisense transcription were found at the beginning and at the end of the XIST locus in the embryos and testicle (Figure 6). Amplicon identities were confirmed by their size in agarose gel (175 and 159 bp for the first and the last exons of XIST, respectively), by their melting temperature in qPCR (~79°C and ~73°C for the amplicons from the first and the last exons, respectively) and by their sequence homology compared with reference sequences from GenBank (Figure 7). Although an amplicon of 1,318 bp was expected in the 5’ portion of XIST (exons 1 and 2), corresponding to the XIST X2 variant, it was not detected in qPCR. Oligo(dT) primers, as a positive control, confirmed the presence of the specific amplicons and no RT control attested for the absence of genomic DNA contamination (Figure 6).

Figure 6.

Strand-specific real time PCR for sense and antisense transcription analysis throughout the bovine XIST locus (TSIXP primers – exons 1 and 2; TSIXU primers – last exon). a – Amplification curves and threshold line (above) and Ct (cycle threshold) for each class of cDNA (below) for TSIXP and TSIXU in bovine embryos. b – 2% agarose gel using 1 kb Plus DNA Ladder (Invitrogen) showing amplicons for the forward, reverse, oligo(dT), RT- and negative PCR for the first and second exons of XIST (175 bp; left side in gel) and for the last exon (159 bp; right side in gel) in bovine embryos. c – Amplification curves and threshold line (above) and Ct (cycle threshold) for each class of cDNA (below) for TSIXP in bovine testicle. d – 2% agarose gel using 1 kb Plus DNA Ladder (Invitrogen) showing amplicons for the forward, reverse, oligo(dT), RT- and negative PCR for the first and second exons of XIST (TSIXP; 175 bp) in bovine testicle. e – Amplification curves and threshold line (above) and Ct (cycle threshold) for each class of cDNA (below) for TSIXU in bovine testicle. f – 2% agarose gel using 1 kb Plus DNA Ladder (Invitrogen) showing amplicons for the forward, reverse, oligo(dT), RT- and negative PCR for the last exon of XIST (TSIXU; 159 bp) in bovine testicle.

Figure 7.

Alignment of amplicons generated from exons 1 and 2 (TSIXP) and the last exon of XIST (TSIXU) of bovine embryos with sequences from GenBank. a – Alignment of an amplicon generated from the first and second exons of XIST with X2 and X3 XIST variants from GenBank. b – Alignment of an amplicon generated from the last exon of XIST with the X1, X2 and X3 XIST variants from GenBank.

Identification of novel XIST locus transcript candidates

We identified 63 novel transcript candidates at bovine XIST locus from both the plus and minus strands (16 foetal plus strand, 37 foetal minus strand, 4 embryonic minus strand, and 4 embryonic plus strand) and two transcripts with no strand information at the embryonic stages analysed (Figure 8). After the alignment of the primer sequences used for amplification of the exon 1–2 regions, we noted that the Ex1-F primer overlapped with exon 1 of the predicted Bos taurus XIST transcript variant 3 (XR_001495595.2) and that the Ex2-R aligns to exon 2 from the predicted transcripts 1–3 (XR_001495594.2; XR_001495596.2, XR_001495595.2 – respectively). All deposited sequences were recovered from GenBank. In terms of the length of the amplified RT-qPCR fragment for this primer pair (Figure 6), the transcript variant was identified as the XIST transcript variant 3 (X3). Therefore, we decided to use this sequence, denoted only as XIST, as the reference sequence in the analysis of our in silico results. Both primers pairs for exon 1 and exon 6 overlap with at least one transcript from the sense strand and with the exon of adopted XIST. In addition, we compared the genomic position corresponding to amplicon fragments amplified by RT-qPCR primers for bovine XIST exons 1–2 and exon 6. Consequently, we found that both the primer sets overlap with sense and antisense transcripts candidates (Figure 8). Overlapping the exon 1 region, we found a set of six transcripts (MSTRG.26262.1; MSTRG.26262.2; MSTRG.26262.3; MSTRG.26262.4; MSTRG.26262.5; MSTRG.26262.6) predicted to be antisense to XIST from the StringTie results. Moreover, interspecies comparative analysis of this region indicated that it corresponds to the mouse homolog region where Xist Activating RNA (XistAR) (GenBank: KJ440524.1) and also presents an overlap with bovine XIST RepA (GenBank: AF104906.5). The fragments identified had a similar length to mouse XistAR, approximately 2.7 kb. According to the multiple alignment results (Supplementary Material), these six transcripts had a similarity ranging from 54.50% to 55.75% with mouse XistAR (Supplementary Material 1). As XistAR and RepA are overlapping in mice, we investigated if the repeats at the bovine RepA region are present in these antisense transcripts. We found these repeats in all six transcript sequences with similarities up to 90% with mice homologs (Supplementary Material). However, Ex2-F but not Ex1-R sequence aligned with the six antisense transcripts with homology to murine XistAR. An antisense transcript with sequence complementarity for both XIST exon1–2 primers was not found by our in silico analysis, suggesting that other antisense not detected by StringieTie analysis may exist.

Figure 8.

UCSC graphical results, showing all predicted transcript candidates found in the bovine XIST locus overlapping the bovine XIST gene (Bta_XIST – transcript reference sequence: XR_001495595.2). All custom tracks were uploaded to UCSC genome Browser for the Bos taurus genome assembly bosTau7. The custom track 1mmf_embryonic_merged_strand_unknown represents the transcript candidates derived from embryonic developmental stages analysed where the strands were not identified. The custom tracks 1mmf_embryonic_merged_plus and 1mmf_embryonic_merged_minus represent the sense antisense transcript candidates respectively, derived from embryonic developmental stages analysed. The custom tracks 1mmf_foetal_day_105_merged_plus and 1mmf_foetal_day_105__minus represent the positions of sense and antisense transcript candidates respectively, derived from foetal (day 105) tissues analysed. Blue vertical bands represent the genomic positions of the primers used for strand specific RT-PCR (SS-RT-PCR), Ex1-F, and Ex2-R. Red vertical bands represent the Ex6-F/R primers.

For both primers used to detect XIST transcription by primer pairs located in the exon 6 sequence, alignment was found for all three aforementioned predicted XIST isoforms. The primers Ex6-F and Ex6-R overlapped with the antisense transcript composed of two exons, with the predicted start site mapping a greater than 10-kb downstream region from the last XIST exon. Another two antisense transcript candidates (MSTRG.271419.4 – embryonic and MSTRG.26263.16 – foetal) have their predicted start sites mapped over a greater than 19-kb downstream region of the bovine XIST 3’ end. The second exon of these transcripts was mapped to intron 1 of XIST. The majority of the predicted antisense transcripts were mapped at the first intron of Bta-XIST. Unfortunately, we were not able to evaluate if the transcripts found at intron 1 were spliced isoforms from the transcripts with start sites mapped downstream to the XIST 3’ end. The only exon-intron structure similarity with human and mice among the antisense transcript candidates for bovine TSIX were an overlap among the 3’-end of transcript candidates MSTRG26264.1 and MSTRG26264.2 and the last exon of bovine XIST. Using the tool ‘view in other genomes’ from UCSC Genome Browser [28], we found that the human homologs of this region were also overlapping between TSIX and XIST. The annotation to the human homolog region was chrX: 73,041,511–73,047,965 (GRCh37/hg19) and corresponds to an overlap region between TSIX and the last exon of human XIST. The alignment of the bovine MSTRG26264.1 and MSTRG26264.2 sequences with the human homolog region presented a high conservation with similarity of 70.23% and 68.32%, respectively (Supplementary Figure 2).

We did not identify an antisense transcript candidate with more than two predicted exons, encompassing also the sequences of intron 1, exon 2 and the last exon of bovine XIST. We did identify the predicted sense transcripts overlapping with at least five XIST exons. Four of these transcripts (MSTRG.26261.1, MSTRG.26261.2, MSTRG.26261.3, and MSTRG.26261.4) presented putative start sites 55 kb upstream of the stat site of our bovine XIST. These four predicted sense transcripts overlapped the other transcripts identified by transcriptome assembly protocol (Figure 8).

Discussion

The long non-coding RNA XIST is essential for the initiation of XCI and its maintenance in female mammalian cells [29,30]. As XCI progresses until the blastocyst stage in mice [5,31], it is important to characterize the XIST expression profile throughout this window of development in order to gain a better understanding of the process of X inactivation. Our laboratory was interested in evaluating the influence of assisted reproductive technologies (ARTs) on DNA methylation patterns of imprinted genes involved in bovine embryo development [32–35]. As XCI is imprinted in mice [5,31,36] and is not as well understood in cattle, we searched for specific epigenetic marks related to XCI in cattle to subsidize future studies about the influence of ARTs on XCI and consequently on embryo quality.

Although XCI may seem to be conserved in some placental mammals, such as mice, humans and rabbits [37], little is known about this mechanism in cattle [12,22,33]. Therefore, the characterization of methylation and the expression profiles of XIST is important considering its pivotal role in XCI [12,29,30,38]. In this study, we first found that XIST DMR1 is hypermethylated and is not reprogrammed during oogenesis, from the primordial follicles until MII oocytes (Figures 1 and 2). In mice, a DMR, located both in the XIST promoter region and in the 5’ portion of the first exon is responsible for XIST imprinted expression [12], which suggests that the DNA methylation pattern of bovine XIST DMR1 may also control XIST expression in cattle. During female PGCs migration to the genital ridge and the start of oogenesis, a cycle of epigenetic reprogramming begins and the inactive X chromosome is reactivated [30,39,40]. Then, both XIST alleles are found in an active state at the time of fertilization. The hypermethylated state that we found in oocytes since primordial follicles may indicate a precocious reprogramming or a protection against reprogramming that may be indicative of a species-specific pattern, as suggested by Colosimo et al [41]. This pattern of hypermethylation during oogenesis may indicate that DNA methylation in this region of XIST is one of the events that enables X chromosome reactivation during bovine oogenesis, inhibiting XIST expression and resulting in an active state for the maternal X at the time of fertilization [42,43]. Since DMR1 is located on XIST gene body, the hypermethylation at oocytes may indicate the transcriptional active state as indicated by the presence of XIST transcripts in female gametes demonstrate in our PCR results (Figure 5). Gene body methylation at X-linked genes has been associate with the active X (Xa) chromosome [44].

Here, we also found that the three regions that we evaluated in XIST, DMR1, RepA, and the promoter were hypermethylated in MII oocytes differently than in sperm, which was demethylated (Figures 2, 3, and 4). Other studies also characterized XIST DMR in cattle [43,45], mice [46] and sheep [47]. These studies showed that this genomic region of XIST is a germline differentially methylated region and suggested that this pattern is conserved among eutherian mammals. This suggests that this region is involved in differential chromatin organization between the maternal and paternal alleles and in XIST expression [5,31]. The other two genomic regions, RepA and the promoter, had not yet been epigenetically characterized in cattle. The mouse Repeat A region of XIST transcribes a sense shorter ncRNA known as RepA, which is essential for XCI initiation [17]. As in mice, XIST and RepA are two sense-transcribed ncRNA showing different mechanisms of transcription control [17]. Our results showed that RepA and promoter are also gDMRs, considering the hypomethylated and hypermethylated patterns found for sperm and MII oocytes, respectively (Figures 3 and 4). Regarding the methylation profile for the promoter (Figure 4), our results are in accordance with the pattern of mouse gametes [48]. In mice, a DMR present in the promoter region and in the 5’ portion of XIST gene controls XIST imprinted expression [12], which suggests that the region that we evaluated here may be involved in the control of XIST expression in early embryonic development in cattle. The hypomethylated pattern found in sperm for the three analysed genomic regions of XIST suggested a poised pattern that is permissive to transcription, but it does not means that the sperm cell is transcribing the gene. Other epigenetic factors than DNA methylation, such as the presence of protamines, may be enough to silence the gene even with the DNA demethylated.

Soon after fertilization, a second cycle of epigenetic reprogramming initiate in both male and female pronuclei [23]. However, the exact moment it occurs depends on the different genomic regions or genes [25]. Our results showed that at the 8–16-cell embryos, DMR1 was already demethylated and remained hypomethylated at least until the blastocyst stage in both TE and ICM cells (Figure 2). This result is in accordance with a global DNA demethylation that occurs in both pronucleus after fertilization in mammals [23,49,50]. Although the association between XIST methylation and gene expression was not tested in this study, it is reasonable to speculate that, at some level, at least some XCI-related events are already occurring at the 8–16-cell embryo stage in cattle, which is most likely completed around the elongated blastocyst stage (days 14–15 of development) [51]. However, RepA showed hypermethylation until the morula stage. This suggests that the paternal allele gained methylation in early embryogenesis since the sperm was hypomethylated (Figure 3). This result is different from what is consolidated in the literature, in which the male pronucleus is actively demethylated soon after fertilization [52]. We suggest that this result could be due to species-specific patterns. The difference in the time of reprogramming between DMR1 and RepA that we showed in this study (Figures 2 and 3) suggests that these two regions of bovine XIST may be involved in the control of expression of different transcripts.

The DNA methylation patterns that control mouse XIST expression have already been identified in gametes [48], preimplantation embryos [5,31], and trophectoderm cells [36,53]. On the other hand, in humans and rabbits, a non-imprinted pattern of XIST expression was observed [54]. In cattle, the literature is controversial. Although an imprinted pattern for XCI has been suggested in bovine placenta [12], other studies show a non-imprinted pattern of XIST expression in placenta [45]. The pattern of hypomethylation of DMR1 and RepA in the trophectoderm cells of blastocysts and the hypermethylated pattern of these regions in female placenta (allantochorion) (Figures 3 and 4) suggest that these regions are not responsible for an imprinted expression pattern of XIST in cattle, because, if so, ~50% of methylation would be found in the placenta [55].

In addition to characterizing the DNA methylation patterns of XIST, we also evaluated the expression profiles of the XIST locus during early development. The literature is controversial about XIST expression in mammalian oocytes and embryos. Murine oocytes do not show XIST expression [56]. In human, the results are more divergent. While a study described XIST expression in embryos starting at the zygote stage [57], a more recent study detected XIST expression from the 4-cell stage [58]. In cattle, while a study did not detect XIST expression in in vitro matured oocytes [59], other studies detected XIST transcription from the one-cell embryo stage [51,60,61]. These are very conflicting results considering that transcripts that are detected in one-cell embryos are from maternal heritance [62]. In this study, we detected XIST transcripts from the oocyte to blastocyst stage (Figure 5 and Figure 6). Other studies also detected XIST expression in bovine embryos [59,63], corroborating our results. In all the aforementioned studies in which XIST expression was analysed in oocytes or embryos, oligo(dT) and/or random primers were used to synthetize cDNA. These strategies were not able to distinguish between sense or antisense transcription. Even our methodology of single-cell gene expression analysis did not discriminate transcripts originating from the sense or antisense DNA strands, being used here just as an initial screening tool. SS-RT-PCR is a reliable methodology to distinguish sense and antisense transcripts, and here we showed both sense and antisense transcription in female embryos and testes (Figs. 6 and 7). Based on what is known regarding mice and humans, the sense transcription detected here could be that of XIST and/or other shorter transcripts, such as RepA [17]. Based on the sequences deposited in GenBank, in this study we only detected the XIST X3 isoform using TSIXP primers (amplicon of 175 bp; Figure 6), considering that the amplicon generated by this pair of primers on the X2 isoform would be of 1318 bp. It is possible that we were not able to detect this isoform due to the difficultly in detecting a large amplicon using real-time PCR.

The methylation and transcription patterns of XIST in embryos found in this study suggest some inferences of XIST expression regulation by DNA methylation. Until 8–16-cell embryo stage, XIST RNA is exclusively from the oocyte, independently of DNA methylation patterns. From morula, XIST RNA that was detected is from embryo cells. DMR1 region is already demethylated in morula stage (Figure 2) and RepA (Figure 3) region demethylated in blastocyst stage (ICM and TE). Then, it is possible that XIST transcriptional activation occurs in the transition from morula to blastocyst, considering that some cells from morula are expressing XIST and others did not (Figure 5). However, although XIST is active from morula stage, XCI only initiate in later stages in cattle [51], which may suggest that either other sense or antisense RNA were detected in SS-RT-PCR assay, or the XIST RNA do not act in embryos in the earlier stages.

Although a recent study showed that XIST is exclusively expressed in female cells and is repressed in male cells [64], another study detected XIST expression in male cells until day 3 of differentiation in mice [65], which may corroborate the XIST expression which we detected in the testes. One hypothesis is that XIST transcription is non-effective in male cells, being immediately degraded after transcription [66].

The most well-known antisense transcript in the mouse XIST locus is TSIX, a negative modulator to XIST [19]. In cattle, however, antisense transcription is not well understood. Until this study, TSIX was believed to be a pseudogene in cattle [12,22]. XIST antisense transcription was then detected in the gonads and somatic tissue of cattle [67], but only evaluating exon 1 of the XIST gene. Here, we showed the presence of antisense transcripts from the XIST locus, first at the beginning of the gene, in exons 1 and 2, and then in the last exon, both in female embryos and the testes.

By analysing our in silico results, we identified a set of sense and antisense transcript candidates which overlapped the homolog regions of XistAR and Tsix/TSIX in mice and humans. The evolutionary conservation of six transcripts overlapping the bovine XIST exon 1 constitutes strong evidence that these transcript candidates are the bovine homolog of mouse XistAR ncRNA [68]. For these transcripts, an identity percent greater than 50% was found after comparing with the mouse XistAR sequence. This similarity was higher in the repeat-A motifs (from bovine RepA sequence) overlapped by these transcripts (Supplementary Figure 1). The A-repeats from RepA have been proposed to recruit PRC2 to Xi, and to participate in XIST processing by binding to alternative splicing factors ASF/SF2 [17,18,69]. NcRNAs show low sequence conservation among the different species but may present patches of higher conservation along their sequences [70,71]. Although, we were unable to detect the expression of these transcripts by PCR using the exons 1–2 primer pair, the high similarity, especially with regards to functional motifs, corroborates the existence of a bovine XISTAR homolog transcript. Mice and human Tsix/TSIX have a distinct exon-intron structure, with human TSIX predicted to be unspliced [72]. The only exon-intron structure similarity we found among the anti-sense transcript candidates to TSIX were an overlap at the 3’ end of the transcript candidates MSTRG26264.1 and MSTRG26264.2 and in the last exon of bovine XIST. This overlap also occurs in humans [72]. The exon-intron structure of four antisense transcripts with putative start sites 10–19 kb downstream of the 3’ end of bovine XIST corroborated with our hypothesis that bovine TSIX is transcriptionally functional. The exact size and gene structure of a TSIX gene in cattle could not be determined by the detection of higher transcripts using our SS-RT-PCR primers. Comparing the in silico and in vitro analysis results, we suggest the existence of a functional TSIX and XistAR homologs in cattle. In addition to the antisense transcription results, our in silico analysis of the sense transcript candidates suggests that these overlap with the last four exons of bovine XIST. We believe that these transcripts may represent alternative splicing isoforms of this ncRNA, as described in other species [15,73,74]. Our results suggest that bovine XIST may be comprised of more alternative splicing isoforms than mice or humans [73] and of that reported in other species, and that these isoforms may be expressed during the embryonic and foetal periods of development. Based on the same comparison results, we also propose that bovine XIST may present more transcript variants and a still unexplored regulatory promoter structure (Figure 8).

To our knowledge, this is the first study to characterize the DNA methylation profile of RepA and the promoter of the XIST locus in cattle during this window of development. Moreover, our results showed, for the first time, the complete characterization of the methylation pattern of DMR1, describing its pattern throughout bovine oogenesis and initial embryogenesis. Considering the differences in methylation found between RepA and DMR1, it is reasonable to suggest that these regions may control the expression of different transcripts. Moreover, it is the first time that sense and antisense transcription is described both at the beginning and at the end of the XIST locus in cattle. Combining the PCR and in silico analysis results, we present the first consistent evidence of the occurrence of bovine XISTAR and TSIX homologs. We also detected possible new variants of the bovine XIST gene. Taken together, the information obtained in this study could contribute to a better understanding of the epigenetic mechanisms involved in XCI during initial embryogenesis in cattle.

Materials and methods

Experimental design

The XIST gene was epigenetically and functionally characterized. Epigenetic characterization was performed by analysing the 5’ portion of the gene. Sodium bisulphite-treated DNA samples were subjected to PCR for three different genomic regions of the XIST gene: a CpG island classically found in exon 1 of XIST (named here as DMR1) [43]; Repeat A region (named here as RepA), also in exon 1; and the XIST promoter region (named here as promoter). The GenBank accession number for all regions is AJ421481.1 and the molecular structure of the 5’ portion of bovine XIST gene that was analysed in this study is illustrated in Figure 9. The DNA methylation pattern was determined in the sperm and oocytes (from preantral and antral follicles), as well as in female embryos, from the 8–16-cell to the blastocyst stage (inner cell mass (ICM) and trophectoderm (TE)). In addition, female and male placenta samples were used as the controls as TE cells were also analysed in this study and an imprinting pattern has already been suggested for the bovine placenta [12]. It was collected placenta biopsies (allantochorion) from four Nellore (Bos taurus indicus) parturitions (2 females and 2 males) produced by artificial insemination using non-sexed semen, immediately after delivery.

Figure 9.

Experimental design: DNA methylation of XIST was characterized in sperm, immature, and in vitro matured oocytes, 8–16-cell and morula stage embryos, inner cell mass and trophectoderm cells and allantochorion. An initial 5’ portion of the XIST locus of cattle was analysed. In the top (A) of the figure there is a schematic representation of this region, in which the wide horizontal black line indicates the genomic sequence, while the larger grey rectangles represent exons 1 and 2. The two inclined lines between exons 1 and 2 indicate that the first intron is not totally represented in the figure. Green and yellow rectangles represent the possible promoter and TATA box, respectively. The narrow vertical red rectangles in exon 1 indicate the 8.5 repeat sequences that consist of Repeat A region, while the purple rectangle in exon 1 indicates DMR1. The large green arrow indicates the XIST transcription start site, while the narrow black arrows indicate the localization and position of the different regions analysed in this study. The large black arrows indicate the three regions chosen for DNA methylation analysis, in which white circles represents the amount of CpG dinucleotides analysed. The DMR1 region is located in exon 1, according to GenBank XIST X2 RNA variant (XR_001495596.2) [12,43,79] (a). Gene expression analysis was conducted in two experiments. Single-cell analysis was performed in mature oocytes and single embryo blastomeres (2-cell, 4-cell, 8–16-cell, and morula embryos). SS-RT-PCR was analysed in female embryos (expanded blastocysts) and testicular parenchyma. All bovine XIST transcripts were analysed, and the three RNA isoforms (X1, X2, and X3) are represented at the bottom of the figure, in which the wide horizontal black line indicates the genomic sequence, while the larger grey rectangles represent the exons. The green rectangles indicate the portion of the last exon analysed in single-cell experiment, while the yellow and light red rectangles indicate the initial portion and last exon, respectively, analysed in SS-RT-PCR (b).

Gene expression experiments were also performed. Single-cell analysis was conducted for a first screening of the XIST expression profile in oocytes and in single blastomeres of female embryos from the 2-cell to morula stage (produced using sexed semen). A region in the last exon was chosen because this exon is common to all three bovine XIST RNA isoforms (X1, X2, X3; GenBank accession numbers XR_001495594.2, XR_001495596.2 and XR_001495595.2, respectively). Female embryos (expanded blastocysts) were also used to run a strand-specific RT-PCR (SS-RT-PCR) experiment. Using this strategy, it was possible to search for sense and antisense transcription in the XIST locus, both at the beginning (present in the X2 and X3 isoforms) and at the end (in the last exon, present in all three XIST isoforms). In both the X2 and X3 variants, reverse primer annealing was carried out in first exon, whereas forward primer annealing was carried out in exon 2, whose transcripts contained 1,318 base pairs and 175 base pairs, respectively. In the SS-RT-PCR analysis, testicular tissue was used as a male control. The experimental design is shown in detail in Figure 9.

Oocyte recovery and in vitro maturation

The ovaries of crossbred cows (Bos taurus indicus x Bos taurus taurus) were collected immediately after their slaughter at a local abattoir (Qualimaxima, Luziânia, Goiás, Brazil). They were immediately transported to the laboratory in saline solution (0.9% NaCl) supplemented with penicillin G (100 IU/mL) and streptomycin (100 µg/mL; Sigma, St. Louis MO, USA) at 34–36°C. Cumulus oocyte complexes (COCs) were recovered and classified according to procedures previously established in our laboratory [35,75,76]. Briefly, the ovarian cortex was separated with a scalpel blade and cut longitudinally, transversally and obliquely with a Tissue Chopper (The Mickle Laboratory Engineering Co. Ltd., Gomshall, Surrey, England). Cuts of 150, 200, 250, 300, and 350 μm thicknesses were performed. The entire process was performed using phosphate-buffered saline (PBS) containing 10% foetal calf serum (FCS; Gibco BRL, Burlington, ON, Canada). Ovarian fragments were placed in 50 mL conical tubes along with approximately 5 mL of PBS supplemented with 10% of FCS. A 3 mL Pasteur pipette was used to mechanically dissociate the oocytes with successive suspension (10–40 times). The resulting material was filtered using a 500 and 245 µm nylon mesh to collect the large and small oocytes, respectively. After decantation, 1 mL of the pellet was analysed using an inverted microscope (Axiovert 135M, Zeiss, Germany). Oocytes with homogeneous cytoplasm and free of granulosa cells (denuded by pipetting) were transferred to a 10 mL drop of tissue culture medium-199 (TCM-199) supplemented with Hank’s Balanced Salt (Gibco BRL, Burlington, ON, Canada). After many washes to remove any granulosa cells and impurities, the isolated oocytes were photographed and measured using the Motic Images Plus 2.0 programme (Motic China Group Co. Ltd., Xiamen, China). The diameter of the oocytes was measured by excluding the zona pellucida. Oocytes <20 μm, 65–85 μm, 100–120 μm (from 1–3 mm follicles), and >128 μm (from >6 mm follicles) in diameter were classified as oocytes from primordial follicles, final secondary follicles, small antral follicles, and large antral follicles, respectively [75]. Oocytes from primordial and final secondary follicles represented the preantral phase of bovine folliculogenesis [75]. Antral groups were selected according to a study that was performed in our laboratory, which showed that oocytes from 1–3 mm follicles are less competent for embryo production than oocytes from follicles >6 mm in size [76]. Oocytes were washed four times with PBS without calcium or magnesium. One hundred and forty oocytes for each group (primordial, final secondary, small antral and large antral) were used for DNA methylation analysis. These were divided in two pools of 70 oocytes. Then, the oocytes were stored at −80°C until DNA isolation.

For in vitro maturation (IVM), COCs from follicles of 3–8 mm in diameter, which are routinely used for IVP, were aspirated. Only COCs with homogeneous granulated cytoplasm and at least three layers of compact cumulus cells were used. After selection, COCs were washed and transferred to drops of 150 µL (25–30 oocytes) of maturation medium, covered with silicone oil and incubated for 24 h at 38.5°C with 5% CO₂. The maturation medium consisted of TCM-199 (Invitrogen, CA, USA) supplemented with 10% FCS (Gibco BRL, Burlington, ON, Canada), 10 µg/mL FSH (Sigma, St. Louis MO, USA), 1.0 µg/ml L-glutamine (Sigma, St. Louis MO, USA), 0.1 mM cysteamine, 0.2 mM sodium pyruvate and 250 mg/mL amikacin sulphate. Following IVM, COCs were incubated with 0.2% hyaluronidase for 10 min. and then denuded by repeated pipetting. Only oocytes that had extruded their first polar body were considered matured (MII oocytes) and used for DNA isolation. Five individual MII oocytes were used for single cell expression analysis. Both immature and matured oocytes were denuded with vigorous pipetting and the absence of cumulus cells was observed in stereomicroscope. Oocytes for both groups were washed four times in PBS without calcium or magnesium before storage at −80°C. Two pools of 60 oocytes per group (immature or MII oocytes) were used for DNA methylation analysis.

In vitro embryo production

COCs submitted to IVM were transferred to 50 µL of fertilization medium Tyrode’s albumin lactate pyruvate [TALP], supplemented with penicillamine 2.0 mM, hypotaurine 1.0 mM, 250 mM epinephrine, and heparin 10 µg/mL, covered with silicone oil and incubated for 24 h at 38.5°C with 5% CO₂. For in vitro fertilization (IVF), female sex-sorted semen from a Gyr bull was used (ABS Pecplan) with proved in vitro fertility. Sperm cells were selected with a 45:90% Percoll (GE® Healthcare, Piscataway, NJ, USA) gradient method. COCs and sperm were co-incubated for 18 h at 38.5°C with 5% CO₂. Zygotes were washed and transferred to drops of 150 µL of SOFaaci medium supplemented with essential and non-essential amino acids, trisodium citrate 0.34 mM, myo-inositol 2.77 mM and 5% of FCS. Embryos were cultured at 38.5°C with 5% CO₂ and were then collected at different development stages, according to previous study [33]: 2-cell [32 h post insemination (p.i.)], 4-cell (48 h p.i.), 8–16-cell (72 h p.i), morula (144 h p.i.) and expanded blastocysts (168 h p.i.). Embryos were stored in PBS at −80°C for genomic DNA isolation or in RNAlater (Ambion, Austin, TX, USA) at −80°C for RNA isolation. For DNA methylation analysis, a pool of 17 embryos of 8–16-cell, two pools of 5 embryos in morula stage and 10 biopsies of expanded blastocysts were used, isolating the inner cell mass and trophectoderm. For single-cell analysis, five embryos of 2-cell, five embryos of 4-cell, three embryos of 8–16 cell and two embryos in morula stage were used. For Strand-Specific RT-PCR analysis, one pool of 40 embryos in expanded blastocyst stage was used.

Micromanipulation for trophectoderm isolation

Expanded blastocysts (n = 10) were micromanipulated to separate the trophectoderm (TE) from the inner cell mass (ICM) [77]. Biopsies were performed manually using a micromanipulator M&M (M&M – The Micromanipulator Microscope Company, Escondido, CA, USA) and stainless-steel blades at a 15 degree angle (Bio-Cut®-Blades Feather, Feather Safety Razor Co, Chome Kita-Ku, Osaka, Japan). Embryos were micromanipulated on a Petri dish containing 150 µL PBS supplemented with 2% FCS, under a stereomicroscope (Stemi SV6, Zeiss®, Göttingen, Germany). The biopsies of the different embryos were pooled according to their origins (ICM or TE) and stored at −80°C. For DNA methylation analysis, one pool of 10 biopsies of ICM and one pool of 10 biopsies of TE were used.

Sperm processing and genomic DNA isolation

Sperm DNA from a sexually mature Nellore (Bos taurus indicus) bull of proven fertility and routinely used for IVP in our laboratory was isolated and used as a control for the DNA methylation analysis. Three straws from three different semen collections were used. Sperm processing and DNA isolation was conducted as described by Carvalho et al [78].

DNA isolation from oocytes and embryos

Pronase E (10 mg/ml; Sigma®) was added to tubes containing pooled oocytes or embryos in order to digest the zona pellucida. The tubes containing Pronase E were incubated at 37°C for 45 min. Following incubation, the enzyme was inactivated at 85°C for 15 min. After digestion with Pronase E, genomic DNA was isolated by cellular lysis using heat shock, in which samples were frozen in liquid nitrogen and immediately heated at 95°C in a thermocycler for 1 min. This procedure was repeated five times. For TE and ICM samples, the same procedure was used, except that Pronase E digestion was omitted.

DNA isolation from placenta

Placenta biopsies (allantochorion) of 1 cm² were used to isolate genomic DNA using the QIAmp DNA Mini and Blood Mini kit (Qiagen), according to manufacturer’s instructions.

Sodium bisulphite treatment, PCR amplification, and bisulphite sequencing for DNA methylation analysis

Genomic DNA from sperm, oocytes, embryos, and placenta were treated with sodium bisulphite using the EZ DNA Methylation-Lightning kit (Zymo Research, Irvine, CA, USA), according to manufacturer’s procedure. Sodium bisulphite-treated DNA were stored at −80°C until PCR amplification.

Nested PCR for DMR1 or hemi-nested PCR for RepA and the promoter were run using a Mastercycler Gradient thermocycler (Eppendorf, Hamburg, Germany). The primers used for DMR1 amplification were obtained from a previous study [43]. Primers for RepA and promoter were designed for this study, based on Repeat A XIST description [79] and XIST promoter prediction for cattle [12], respectively. Primers for RepA and promoter were designed using the Bisulfide Primer Seeker tool (http://www.zymoresearch.com/tools/bisulfite-primer-seeker). Primer sequences, their position and amplicon size are listed in Table 1.

Table 1.

Gene identification, primer sequences, primer positions, amplicon size, and number of CpG sites.

Genomic region	Primers sequences (5’-3’)	Primers position	Amplicon size	Number of CpG sites
DMR1 out	F: GGGTGTTTTTGTTTTAGTGTGTAGTA	+1127 to +1252	482 pb	17
	R: CTTTAATACCACCCACTAAAATTAATAC	+1581 to +1608
DMR1 inner	F: TTGTTATATAGTAAAAGATGGT	+1169 to +1190	405 pb
	R: ACCAATCCTAACTAACTAAATA	+1552 to +1573
RepA out	F: TTTGGTTGTTTTTTTTGGGTTTTTTGTGG	+84 to +112	634 pb	18
	R: ACCTAACTACAAAATCATCTCCCAA	+693 to +717
RepA inner	R: TAAATCCACTCACACAACACAAAC	+660 to +683	600 pb
Promoter out	F: TTT GAA GTT ATG GTT TTT GGA TTA GAA ATG	−411 to −382	320 pb	6
	R: TAAAATCTAAAAAATATTCCAAAAAAAACCACAC	−125 to −92
Promoter inner	F: TTGTTATTTTTTTGAATTTTTTTTTTTTTGTTATTGGG	−346 to −309	255 pb

For DMR1, PCR was performed in a final volume of 20 µL using 1X Taq buffer, 1.5 mM MgCl₂, 0.4 mM dNTPs, 1 U Platinum Taq polymerase (Invitrogen, CA, USA), and 1 µM of each primer (forward and reverse). The amplification conditions for all samples (sperm, oocytes, embryos, and placenta) are specified in Table 2. Final extension for both rounds of amplification was conducted at 72°C for 15 min.

Table 2.

Nested PCR conditions for the DMR1.

Reaction	DNA template (µL)	Initial denaturing temperature	40 cycles
			Denaturing	Annealing	Extension
First	3.0	94°C 7 min.	94°C 45 s.	47°C 1 min. and 30 s.	72°C 1 min.
Second	0.5*	94°C 4 min.	94°C 40 s.	42°C 45 s.	72°C 45 s.

* Volume of amplicon from the first PCR

For RepA, PCR was performed in a final volume of 20 µL using 1X Taq buffer, 0.4 mM dNTPs, 1 U Platinum Taq polymerase (Invitrogen, CA, USA), and 1 µM of each primer (forward and reverse). The amplification conditions for all samples (sperm, oocytes, embryos and placenta) were performed with an initial denaturing step at 94°C for 4 min and final extension at 72°C for 15 min. MgCl₂ concentration and amplification conditions for each sample for the first and second rounds of amplification are listed in Table 3.

Table 3.

Nested PCR conditions for RepA.

Sample	Round of amplification	MgCl2 (mM)	DNA template (µL)	45 cycles
				Denaturing	Annealing	Extension
Sperm	First	1.0	3.0	94°C 40 s.	55°C 1 min.	72°C 1 min.
	Second	1.0	0.25*	94°C 40 s.	55°C 1 min.	72°C 1 min.
Placenta	First	1.0	3.0	94°C 40 s.	55°C 1 min.	72°C 1 min.
	Second	1.0	0.25*	94°C 40 s.	64.5°C 1 min.	72°C 1 min.
Oocytes and embryos	First	1.0	3.0	94°C 40 s.	54°C 1 min.	72°C 1 min.
	Second	1.5	0.5*	94°C 40 s.	65°C 1 min.	72°C 1 min.

* Volume of amplicon from the first PCR

For XIST, promoter PCR was performed in a final volume of 20 µL using 1X Taq buffer, 0.4 mM dNTPs, 1 U Platinum Taq polymerase (Invitrogen, CA, USA) and 1 µM of each primer (forward and reverse). All the PCR amplifications were performed with an initial denaturing step at 94°C for 4 min and a final extension at 72°C for 15 min. The MgCl₂ concentration and amplification conditions for all samples (sperm and oocytes) are listed in Table 4.

Table 4.

Nested PCR conditions for the XIST promoter.

Reaction	MgCl2 (mM)	DNA template (µL)	40 cycles for the first round and 45 for the second round
			Denaturing	Annealing	Extension
First	2.0	3.0	94°C 40 s.	53.5°C 1 min.	72°C 1 min.
Second	1.5	0.5*	94°C 40 s.	59°C 1 min.	72°C 1 min.

* Volume of amplicon from the first PCR

After PCR, the amplicons were purified from agarose gel using the Wizard® SV Gel and PCR Clean-Up System (Promega, Madison, WI, USA), according to the manufacturer’s procedure. Thus, the purified amplicons were cloned into the TOPO TA Cloning® vector (Invitrogen, CA, USA) and transferred into DH5α cells using a heat shock procedure. Plasmidial DNA was isolated using the QIAprep Spin Miniprep Kit (Qiagen, CA, USA) and individual clones were sequenced using BigDye® cycle sequencing chemistry and an ABI3100 automated sequencer. The sequencing quality was analysed using Chromas®, and the methylation pattern was analysed using the BiQ Analyser programme [80]. DNA sequences were compared with GenBank AJ421481 for the Bos taurus XIST DNA sequence. Only sequences originating from clones with ≥95% of homology and cytosine conversion were used in the analysis [32,34,35,78].

Single-cell gene expression

Matured oocytes (MII) (n = 5) and female embryos (2-cell, n = 5; 4-cell, n = 5; 8–16-cell, n = 3; morula, n = 2) were produced as previously described. Individual oocytes or embryos were sequentially washed in PBS with 20% and 0% of FCS. Then, samples were incubated at 37°C for 5 min to digest zona pellucida using Pronase E (Sigma, St. Louis MO, USA) at a final concentration of 10 mg/mL. Zona pellucida degradation was monitored using a stereomicroscope and thereafter samples were washed twice in PBS with 2% of FCS. Individual oocytes were placed into 0.2-mL microtubes. Single blastomeres were disaggregated by a mechanical method using a glass pipette. Individual blastomeres were also placed in the 0.2-mL microtubes. Cell lysis, reverse transcription, cDNA pre-amplification and real-time PCR amplification were conducted according to the Single Cell-to-CT™ kit (Ambion, Austin, TX, USA), using the manufacturer’s procedure based on a TaqMan assay and oligo(dT) primers (Applied Biosystems). A pair of primers was used to amplify the region of the last exon of XIST, which is common to the three known isoforms of bovine XIST RNA. GAPDH was used as endogenous control. qPCR was run in an ABI 7500 Fast Real-Time PCR system (Applied Biosystems, Foster City, CA, USA) thermocycler in the following conditions: 50°C for 2 min.; 95°C for 10 min. and 45 cycles of 95°C for 5 sec. and 60°C for 1 min.

Strand-specific RT-PCR

One pool of 40 expanded blastocysts (Bx; 168 h p.i.) were used in this assay. Total RNA was isolated using the RNeasy Plus Micro kit (Qiagen, Hilden, Germany), according to manufacturer’s procedure. Total RNA was treated with 5 U of RQ1 RNase-Free DNase (Promega, Madison, WI, USA) and cDNA synthesis were conducted using the SuperScript™ III First-Strand Synthesis SuperMix kit (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s procedure. cDNA was performed using forward and reverse primers for the exon 1-2 and the exon 6 of XIST . cDNA was also performed using Oligo(dT) primers as control. The primers used in this experiment are listed in Table 5.

Table 5.

Gene identification, primers sequences and concentration, and amplicon size for SS-RT-PCR analysis.

Gene	Primers sequence (5’-3’)	Primers concentration	Amplicon size
XIST exon 1–2	Ex2-F: CAGGCTTCACTCCACCTAAA	100 nM	175 pb
	Ex1-R: GTTAGGCTAGAGGGTTGGTTAG
XIST exon 6	Ex6-F: GGACCAGACTTCACCAAGAAA	200 nM	159 pb
	Ex6-R: GAAATGGGCCTAGTCTAAAGGG

A cDNA synthetized with Oligo(dT) primers was used as a positive control and a cDNA synthesis using Oligo(dT) primers in the absence of reverse transcriptase was used as a negative control. Therefore, six different cDNA samples were synthetized. Each real-time PCR was run using cDNA relative to two embryos and was performed using Fast Sybr Green Master Mix (Applied Biosystems, Foster City, CA) in an ABI 7500 Fast Real-Time PCR system thermocycler (Applied Biosystems, Foster City, CA). The qPCR conditions were as follows: 95°C for 10 min, 50 cycles of 95°C for 15 s and 60°C for 1 min. The size of the amplicons was confirmed using a 2% agarose gel. They were purified from the gel using a Wizard® SV Gel and PCR Clean-Up System kit (Promega, Madison, WI, USA), according to manufacturer’s procedure. Thereafter, the amplicons were sequenced using BigDye® cycle sequencing chemistry and an ABI3100 automated sequencer and the resulting sequences were compared to sequences deposited in GenBank using the BLASTN tool (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=blastn&PAGE_TYPE=Blast Search&LINK_LOC = blasthome).

Testicular tissue was also used to isolate RNA to be used as the male control. Total RNA was isolated using the TRIzol™ Plus RNA Purification kit (Ambion, Austin, TX, USA). Total RNA was treated with 5U of RQ1 RNase-Free DNase (Promega, Madison, WI, USA). RNA from testicular tissue was used to synthetize the same cDNA samples as embryos and an amount relative to 50 ηg of RNA was used in a qPCR reaction.

Identification of novel transcripts candidates at the bovine XIST locus

To verify the presence of possible sense and antisense transcription by RT-qPCR at the bovine XIST locus, we performed a transcriptome assembly analysis using the RNA-seq data obtained from the Short Read Archive (SRA) of the NCBI (www.ncbi.nlm.nih.gov/sra/). The SRA files used corresponded to RNA-seq data from different bovine developmental stages such as blastocyst (SRR2927479, SRR2927480, SRR2927481, SRR2927482, SRR2927483, SRR2927484, SRR2927485, SRR2927486, SRR1217126, SRR1217128, SRR1217129, SRR1217130), hatched embryos (SRR1217104, SRR1217105, SRR1217106, SRR1217107, SRR1217108, SRR1217109), and elongated embryos (SRR1217110, SRR1217111, SRR1217112, SRR1217114) [81,82].

Since antisense transcription has been previously reported at the XIST locus for foetal tissue samples in cattle [67], we also performed transcriptome assembly using the RNA-seq data corresponding to tissues derived from foetal day 105 from the work of [83] (SRR1658361, SRR1658365, SRR1658373 – brain; SRR1658363, SRR1658367, SRR1658375 – liver; SRR1658364, SRR1658368, SRR1658379 – skeletal muscle) . Therefore, all SRA files were downloaded from the SRA database and converted into FASTQ format using SAMtools [84]. The FASTQ files were trimmed using Trim Galore – Version 0.4.4 (https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/). Subsequently, trimmed files were analysed using FastQC software and only files with very good quality calls (following the FatQC parameters) were used for further analysis (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/).

After processing the reads, we performed alignments against the BosTau7 reference genome using the HISAT2 aligner [85] followed by the conversion of the SAM files into BAM format and subsequently into the bam.sorted format files. These final files were used for transcriptome assembly using the StringTie assembler [86]. The GTF (Gene transfer format) files generated from each developmental stages or foetal tissues were merged to create a transcriptome representation of each sample.

To make facilitate the analysis of the in silico results, we divided the annotation results for the transcript candidates of the bovine XIST loci files into embryonic and foetal tissues groups and into the plus and minus derived transcripts (Figure 8). Moreover, these specific GTF files were uploaded onto UCSC custom track and the XIST locus transcripts were downloaded in separated fasta files containing the plus and minus strand derived transcripts using UCSC Table Browser [28]. Finally, the similarities of the transcript sequences were compared with homolog ncRNAs from the XIST loci from human and mice using the online version of Clustal Omega software [87].

Statistical analysis

Methylation data were compared using ANOVA and the Tukey test or the Student’s t-test and the Kruskal-Wallis and Mann-Whitney U-tests for normal and not normal data, respectively. Gene expression and strand-specific expression data were presented as a descriptive analysis. p < 0.05 was considered to be statistically significant difference. All analyses were performed using Systat 10.2 (2002, Inc., Richmond, CA, USA).

Funding Statement

This study was financed in part by the Coordination for the Improvement of Higher Education Personnel - Brazil (CAPES) - Finance Code 001 and Embrapa Genetic Resources and Biotechnology, Brazil.

Acknowledgments

We would like to thank Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brazil (CAPES) and Embrapa Genetic Resources and Biotechnology, Brazil for the support they provided to this study. Alexandre Rodrigues Caetano, Margot Alves Nunes Dode, and Maurício Machaim Franco were CNPq research fellows.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed here.