Synergy of Eed and Tsix in the repression of Xist gene and X-chromosome inactivation
2008; Springer Nature; Volume: 27; Issue: 13 Linguagem: Inglês
10.1038/emboj.2008.110
ISSN1460-2075
AutoresShinwa Shibata, Takashi Yokota, Anton Wutz,
Tópico(s)CRISPR and Genetic Engineering
ResumoArticle29 May 2008free access Synergy of Eed and Tsix in the repression of Xist gene and X-chromosome inactivation Shinwa Shibata Corresponding Author Shinwa Shibata Department of Stem Cell Biology, Graduate School of Medicine, Kanazawa University, Takara-machi, Kanazawa, Ishikawa, Japan Search for more papers by this author Takashi Yokota Takashi Yokota Department of Stem Cell Biology, Graduate School of Medicine, Kanazawa University, Takara-machi, Kanazawa, Ishikawa, Japan Search for more papers by this author Anton Wutz Anton Wutz Research Institute of Molecular Pathology, Vienna, Austria Search for more papers by this author Shinwa Shibata Corresponding Author Shinwa Shibata Department of Stem Cell Biology, Graduate School of Medicine, Kanazawa University, Takara-machi, Kanazawa, Ishikawa, Japan Search for more papers by this author Takashi Yokota Takashi Yokota Department of Stem Cell Biology, Graduate School of Medicine, Kanazawa University, Takara-machi, Kanazawa, Ishikawa, Japan Search for more papers by this author Anton Wutz Anton Wutz Research Institute of Molecular Pathology, Vienna, Austria Search for more papers by this author Author Information Shinwa Shibata 1, Takashi Yokota1 and Anton Wutz2 1Department of Stem Cell Biology, Graduate School of Medicine, Kanazawa University, Takara-machi, Kanazawa, Ishikawa, Japan 2Research Institute of Molecular Pathology, Vienna, Austria *Corresponding author. Department of Stem Cell Biology, Graduate School of Medicine, Kanazawa University, 13-1 Takara-machi, Kanazawa, Ishikawa 920-8640, Japan. Tel.: +81 76 265 2208; Fax +81 76 234 4238; E-mail: [email protected] The EMBO Journal (2008)27:1816-1826https://doi.org/10.1038/emboj.2008.110 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info X-chromosome inactivation (XCI) depends on the noncoding Xist gene. Xist transcription is negatively regulated by its antisense partner Tsix, whose disruption results in nonrandom XCI in females. However, males can maintain Xist in a repressed state without Tsix, indicating participation of additional factor(s) in the protection of the single male X from inactivation. Here, we provide evidence that the histone methyltransferase Eed is also involved in the process. Male embryonic stem cells with Eed-null and Tsix mutations (XΔY Eed−/−) showed Xist hyperactivation upon differentiation, whereas cells with either mutation alone did not. Impaired X-linked gene expression was observed in the XΔY Eed−/− ES cells at the onset of differentiation. The Xist promoter in the XΔY Eed−/− cells showed elevated histone H3-dimethyl lysine 4 modifications and lowered CpG methylation, which are characteristics of open chromatin. Hence, we identified Eed as an additional major player in the regulation of Xist expression. The synergy of Polycomb group proteins and antisense Tsix transcription in Xist gene regulation explains why males can repress Xist without Tsix. Introduction X-chromosome inactivation (XCI) is a sex chromosome dosage compensation mechanism employed by female mammals. During the process, one of two active X-chromosomes (Xa) in female embryonic cells is randomly chosen and inactivated during development (Lyon, 1961; reviewed by Heard and Disteche, 2006). The noncoding gene Xist (Brockdorff et al, 1992; Brown et al, 1992) has been shown to be critical for XCI (Penny et al, 1996). It is encoded on the X-chromosome and is transcribed at a very low level in the undifferentiated condition in both females and males (Panning and Jaenisch, 1996; Lee et al, 1999). Upon differentiation, it is exclusively expressed from the inactive X-chromosome (Xi) and coats Xi in females (Clemson et al, 1996), whereas Xist transcription is soon terminated on the future Xa and in males. The choice of Xi is achieved by Xist upregulation in cis (Wutz and Jaenisch, 2000). Xist is believed to function as an RNA entity, because of its characteristic repeat sequence (Wutz et al, 2002) and distribution pattern in the nucleus. Xist is negatively regulated by its antisense partner Tsix, which overlaps the Xist gene (Lee et al, 1999; Sado et al, 2001; Shibata and Lee, 2003). Xist is always upregulated at the mutant Tsix allele in heterozygous female embryonic stem (ES) cells, resulting in nonrandom inactivation of the Tsix-mutated X-chromosome (Lee and Lu, 1999; Luikenhuis et al, 2001; Sado et al, 2001; Shibata and Lee, 2004). In contrast, Tsix mutation does not lead to Xist expression in male ES cells upon differentiation. Previous reports described ectopic Xist accumulation in a minor portion of Tsix-mutant male ES cells (0–13%) (Lee and Lu, 1999; Luikenhuis et al, 2001; Sado et al, 2002), whereas another study observed ectopic Xist accumulation more frequently (39%) (Vigneau et al, 2006). Importantly, male embryos carrying a Tsix mutation on the single X-chromosome develop to term when the extraembryonic tissues are complemented by wild-type tetraploid cells (Ohhata et al, 2006), indicating that most embryonic cells in males can maintain Xist gene repression without Tsix. These observations suggest the presence of additional or alternative factor(s) that inhibit the activation of Xist gene in male embryos. Recent studies have shed light on the role of Tsix in regulating chromatin structure in the Xist locus. Sado et al (2005) indicated that disruption of Tsix caused impaired establishment of repressive chromatin structure at the Xist promoter and exon 1 in developing embryos. Navarro et al showed that the Xist promoter region, flanked by CTCF-binding sites, was maintained in a heterochromatic state by Tsix. Tsix truncation resulted in altered modification at lysine 4 of histone H3 (H3K4) and lysine 9 to resemble a pseudoeuchromatic state (Navarro et al, 2006). Sun et al (2006) reported that Tsix downregulation induced a transient heterochromatic state, characterized by histone H3 trimethyl-lysine 27 (H3K27m3) modification in undifferentiated female ES cells. These reports suggest that Tsix transcription influences the chromatin structure at the Xist promoter in different ways depending on the differentiation stage and position within the locus. We focused on H3K27m3, because this modification is clearly elevated when Tsix transcription is absent in both female and male undifferentiated ES cells (Navarro et al, 2006; Sun et al, 2006; Shibata and Yokota, 2008). In addition, the biological significance of the regulation by Tsix of the H3K27m3 modification is still unclear. The H3K27m3 modification is generally considered to be a repressive chromatin mark; however, the loss of Tsix transcription paradoxically results in Xist gene activation in females. Methylation of the histone H3 lysine 27 (H3K27) is conferred by the Polycomb repressive complex 2 (PRC2), which is composed of the Eed, Ezh2 and Suz12 proteins (Cao and Zhang, 2004). Eed is essential for the histone methyltransferase (HMTase) activity, because Eed−/− ES cells lack the H3K27m3 modification (Montgomery et al, 2005). Eed is necessary for development (Faust et al, 1995) and regulates developmental control genes as well as a subset of imprinted genes (Mager et al, 2003). In ES cells, PRC2 occupies genes encoding transcription factors crucial for development, and Eed mutations result in their premature expression (Azuara et al, 2006; Boyer et al, 2006). These loci were termed bivalent domains due to the special modification pattern consisting of trimethylated H3K27 and H3K4, a repressive and active chromatin mark, respectively (Bernstein et al, 2006). Hence, Eed has an important function in gene regulation in undifferentiated and differentiating cells in conjunction with other chromatin factors. Eed and H3K27 methylation are also involved in the establishment and maintenance of XCI. Eed localizes on the Xi in female trophoblast stem cells, and reactivation of Xi is found in Eed−/− trophoblast stem cells when they are differentiated (Kalantry et al, 2006). Recruitment of Eed and H3K27 methylation are also observed on Xi in female embryos and ES cells at early stages of XCI (Plath et al, 2003; Silva et al, 2003). However, recent findings indicate that Eed is dispensable for the initiation of random XCI (Kalantry and Magnuson, 2006; Schoeftner et al, 2006). XCI without Eed is explained by a contribution of Polycomb repressive complex 1 (PRC1) that ubiquitinates histone H2A (Schoeftner et al, 2006). These reports focused mainly on the role of Eed in inducing global heterochromatin formation on Xi, but as was shown by Sun et al (2006), Eed is likely to have additional roles in the regulation of local Xist chromatin structure in concert with Tsix. Therefore, we disrupted Tsix in an Eed−/− male ES cell line to investigate the role of Eed in regulating Xist chromatin structure and to examine the biological significance of the H3K27m3 modification that is observed when Tsix transcription is absent. The role and relationship of Eed and Tsix in the regulation of Xist are discussed. Results Generation of male Tsix mutant ES cells with Eed−/− background Tsix mutant ES cell lines are summarized in Figure 1A. Firstly, we targeted the clone36 Eed−/− male ES cell line (XY Eed−/−) (Schoeftner et al, 2006) and truncated Tsix transcription to generate male Tsix mutant ES cells with an Eed−/− background (XΔY Eed−/−) (Figure 1B–D). This type of Tsix mutation has been shown to eradicate its function in repressing Xist in female ES cells (XΔX) (Shibata and Lee, 2004). We then rescued Eed in the XY Eed−/− and XΔY Eed−/− cells by transgenic expression of an enhanced green fluorescent protein (EGFP)–Eed fusion protein (XY Eed-TG and XΔY Eed-TG, respectively) (Figure 1E and F). In addition to the western blot for H3K27m3, a quantitative chromatin immunoprecipitation (ChIP) assay was used to examine known H3K27m3-labeled sites in undifferentiated ES cells, the Sox9 and Gata6 promoters (Boyer et al, 2006), and confirmed that the Eed activity was sufficiently rescued in the XΔY Eed-TG cells for these promoters (Supplementary Figure S1). Although the clone36 Eed−/− ES cells have an additional Xist cDNA transgene (Tg) under control of tetracycline-inducible promoter on chromosome 11, the Xist Tg has been shown to be inactive without induction (Wutz and Jaenisch, 2000). Figure 1.Generation of Tsix-trap male ES cells in the Eed−/− background (XΔY Eed−/−). (A) Relationship of ES cell lines generated in this study. tetOP-Xist, tetracycline-inducible promoter and Xist cDNA (Tg) with Mus spretus repeat polymorphism; 11, chromosome 11; X, chromosome X. (B) Targeting construct for Tsix. Large open and small gray rectangles show Xist and Tsix exons, respectively. Numbered arrows represent primers for genomic PCR. S, SpeI; E, EcoRI; B, BamHI restriction enzyme sites. (C) SpeI-digested Southern blot showing correct recombination of 5′-homology arm in two independent XΔY Eed−/− clones. All lanes were derived from the same gel. (D) Genomic PCR confirming proper recombination of 3′-homology arm. Primer pair 1–2 was used for mutant (mt) and 1–3 for wild-type (wt) amplification. All lanes were derived from the same gel. (E) Rescuing Eed by transfecting the pEGFP-Eed plasmid (XY Eed-TG and XΔY Eed-TG). Expression of the fusion protein was confirmed by flow cytometry. Results were from the cells at the passage of less than 4 (XΔY Eed-TG) or 8 (XY Eed-TG) after their derivation. (F) Western blot demonstrating reversion of the H3K27m3 modification in the transgenic cell lines. Western blotting was done using cells at the passage of less than 4 (XΔY Eed-TG) or 8 (XY Eed-TG). Four lanes in the right and left panels were derived from the same gel, respectively. Download figure Download PowerPoint XDY Eed−/− cells display Xist hyperactivation upon differentiation We examined Xist RNA expression in the XΔY Eed−/− cells by fluorescent in situ hybridization (FISH) using a strand-specific riboprobe. We found strong Xist expression in the XΔY Eed−/− cells, but not in the XY Eed−/− embryoid bodies (EB) differentiated for 3 days (Figure 2A and B). The number of Xist-positive nuclei was found significantly elevated in two independent XΔY Eed−/− lines and was similar to that of differentiating XΔX ES cells (Figure 2C and Supplementary Table I). Polymorphism of an Xist RT–PCR product confirmed that the ectopic Xist expression in the XΔY Eed−/− EB was from the endogenous Xist allele, not from the Xist cDNA Tg, which is also present in all clone36-derived ES cells (Figure 2D). The amount of Xist RNA expressed during the course of XCI was further quantified by real-time PCR. The results were normalized to glyceraldehyde-3-phosphate dehydrogenase (Gapdh) expression, and the amount of Xist RNA expression, relative to undifferentiated wild-type female (XX) ES cells, is shown (Figure 2E and Supplementary Table II). Interestingly, the XΔY Eed−/− ES cells showed elevated Xist RNA level even in the undifferentiated condition, and upon differentiation, they expressed five to ten times more Xist RNA than XX cells. The XY Eed−/− EB also displayed elevated Xist level when compared with undifferentiated XY Eed-TG cells, but far less than XX and XΔY Eed−/− EB. We then investigated whether the loss of Eed was the cause of this result, because Xist activation does not generally occur in male Tsix mutant ES cells. We looked for suppression of ectopic Xist hyperactivation in the XΔY Eed-TG cells and confirmed that the rescued Eed successfully inhibited Xist hyperactivation (Figure 2F). Therefore, both Tsix and Eed contribute to the repression of Xist gene, but either of the two is sufficient for preventing ectopic Xist activation in males. Figure 2.XΔY Eed−/− cells display Xist hyperactivation upon differentiation. (A) Xist RNA-FISH using strand-specific riboprobe (red) in the XY Eed−/− and (B) XΔY Eed−/− ES cells differentiated for 3 days. (C) The count of Xist-positive nuclei in FISH. More than 80 nuclei for the XY Eed−/− cell line and more than 180 nuclei for other lines were counted. The XΔY Eed−/−1 and XΔY Eed−/−2 are independent clones. (D) Xist cDNA Tg is inactive in the XΔY Eed−/− ES cells, shown by the polymorphism of Xist RT–PCR product digested with PstI restriction enzyme. tetOP-Xist, Xist from the Tg with Mus spretus (M. sp) sequence. RT–PCR was performed using RNA obtained from the XY Eed-TG cells cultured in the presence of doxycyclin for Tg induction. Endo. Xist, endogenous Xist. RT–PCR in the XΔX EB cells in which Xist is predominantly expressed from the Mus musculus (M. mus) allele. XΔY Eed−/−, RT–PCR in the XΔY Eed−/− EB cells differentiated for 12 days. Shown below is a schematic representation of the PCR products with M. mus or M. sp repeat polymorphism (open boxes). P, PstI sites. (E) Quantitative RT–PCR for Xist. Relative amount (mean) of Xist RNA to undifferentiated (undif.) wild-type female (XX) ES cells normalized to Gapdh is shown. Error bars represent s.d. (F) Rescuing Eed inhibited ectopic Xist expression. Relative amount of Xist RNA in the XΔY Eed−/− day 4 EB to that in the XΔY Eed-TG is shown. Download figure Download PowerPoint Xist hyperactivation in the XDY Eed−/− cells leads to partial XCI upon differentiation We next investigated the consequence of Xist hyperactivation in the XΔY Eed−/− cells by differentiating the mutant ES cells in vitro. The XΔY Eed−/− ES cells in the undifferentiated condition displayed round, well-packed colony morphology typical of mouse ES cells (Figure 3A). The XΔY Eed−/− EB cells, after long adherent culture, contained flattened cells suggesting differentiation (Figure 3B), but it was not clear whether their X-chromosomes were inactivated or not. The growth of XY Eed−/− EB cells were retarded as compared with the EB cells with intact Eed (Figure 3C). Expression of transgenic Eed rescued their poor growth (Supplementary Figure S2). The growth of XΔY Eed−/− EB cells was further retarded compared with the XY Eed−/− EB cells: they spread less and their EB size was smaller than that of the XY Eed−/− EB cells (Figure 3C and Supplementary Figure S2). This observation suggested that a substantial amount of differentiating XΔY Eed−/− cells were lost from culture due to the inactivation of their single X-chromosome. To examine if XCI occurs in the XΔY Eed−/− cells during differentiation, we studied the expression of X-linked Mecp2, Pgk1 and Chic1 genes by quantitative RT–PCR (qRT–PCR) (Figure 3D). The expression of these genes decreased immediately upon differentiation, and interestingly, the reduction became less obvious at the late stage of EB day 12. We also examined colocalization of the Xi chromatin marker, histone H4 monomethyl-lysine 20 (H4K20m1) (Kohlmaier et al, 2004), with Xist RNA in immuno-FISH. Although Xist RNA deposition was frequently found in the XΔY Eed−/− EB nuclei, co-localization of Xist and condensed H4K20m1 was never detected in the late stages of differentiation in contrast to the XΔX nuclei (Figure 3E and Table I). Careful examination of EB cells at day 2 or 4 revealed weak H4K20m1 staining with Xist paint in a maximum of 10% of the XΔY Eed−/− nuclei (Supplementary Figure S3 and Supplementary Table III). Taken together, these findings indicate that the Xist hyperactivation in the XΔY Eed−/− cells induced partial XCI at the onset of differentiation. However, it was incomplete, presumably due to the absence of Eed, and a substantial number of cells survived and restored their X-linked gene expression after the critical time window for silencing by Xist RNA (Wutz and Jaenisch, 2000) (Supplementary Figure S4). Figure 3.Differentiation and XCI of the XΔY Eed−/− ES cells. (A) Compact colony morphology of the XΔY Eed−/− ES cells in undifferentiated condition. (B) Morphology of the XΔY Eed−/− EB differentiated for 12 days. (C) Gross appearance of day 12 EB in the XΔY, XY Eed−/− and XΔY Eed−/− background. (D) Relative amount of X-linked Mecp2, Pgk1 and Chic1 mRNA in the XΔY Eed−/− cells (colored columns) to those in the XY Eed−/− cells (gray columns) in undifferentiated or differentiating conditions. Error bars show s.d. Asterisks demonstrate statistically significant reduction of the gene expression in the XΔY Eed−/− cells (*P<0.05; **P<0.0005). (E) Immuno-FISH for H4K20m1 (green) and Xist RNA (red) in the XΔX and XΔY Eed−/− EB (days10–12). Download figure Download PowerPoint Table 1. The number of nuclei showing colocalization of condensed H4K20m1 and Xist RNA in Xist-positive nuclei of the mutant EB Genotype EB differentiation H4K20m1 colocalization No. of Xist-positive nuclei counted XΔX 7 days 44 (40.4%) 109 XΔY Eed−/− 1a 7 days 0 (0%) 142 XΔY Eed−/− 1a 18 days 0 (0%) 143 XΔY Eed−/− 2b 18 days 0 (0%) 114 a XΔY Eed−/− cell clone 1. b XΔY Eed−/− cell clone 2. Deregulated antisense transcription in the Xist gene body of the XDY Eed−/− ES cells We confirmed that Tsix transcription was successfully truncated in the XΔY Eed−/− ES cells by a northern blot of poly-A purified RNA, using a probe residing in the Xist promoter (Figure 4A and B). However, antisense RNA was detected in the double mutant cells by strand-specific RT–PCR, and it disappeared when Eed was rescued (Figure 4C). To eliminate the possibility that the transcript originated from the Xist cDNA Tg, we performed qRT–PCR for Tsix in amplicons that do not amplify the Tg (Figure 4D). The amplicon at the 3′-end of the Tsix (no. 4) antisense transcript was not detected in the XΔY Eed−/− cells, whereas in the amplicon spanning the Xist introns (no. 5) could be detected. We suggest that the loss of Eed in the Tsix-deficient background resulted in an open chromatin structure that led to deregulated antisense transcription from cryptic promoters to various degrees in the Xist gene body. The absence of an antisense transcript at the 3′-end of Tsix suggested that the transcript was terminated by multiple poly-A signals in the antisense orientation residing near the Xist transcription start site (Shibata and Lee, 2003). Average Tsix expression levels observed by qRT–PCR were lower in the XY Eed−/− and XY Eed-TG lines than the wild-type male ES cells, but the difference was not statistically significant. Figure 4.Northern blot and strand-specific or quantitative RT–PCR for Tsix. (A) Positions of northern blot probe (filled rectangle 1), strand-specific RT–PCR amplicons (2 and 3) and qRT–PCR amplicons (4 and 5). (B) Northern blot for Tsix. The XΔY Eed−/−1 and XΔY Eed−/−2, and the XΔY Eed-TG1 and XΔY Eed-TG2 are independent clones. After the initial northern blot for Tsix, the same membrane was stripped and reprobed for Gapdh. (C) Strand-specific RT–PCR for Tsix at the amplicons 2 and 3. (D) qRT–PCR for Tsix at amplicons 4 and 5. Relative Tsix expression levels to the wild-type male (XY) ES cells are shown. Results were from more than three independent samples and error bars indicate s.d. Filled circles in the Tsix RNA (5) graph represent Tsix levels in individual samples of the XΔY Eed−/−1 and XΔY Eed−/−2 lines. Download figure Download PowerPoint The XDY Eed−/− ES cells display loss of CpG methylation at the Xist promoter DNA in the Xist locus has been shown to be methylated in undifferentiated male ES cells (Norris et al, 1994). We examined the methylation level of the Xist locus in undifferentiated XΔY Eed−/− ES cells by Southern blot using methylation-sensitive restriction enzymes. The SacII site at the Xist promoter displayed lowered CpG methylation in the XΔY Eed−/− cells (Figure 5A). This was also the case in the HpaII site within Xist exon 1, which was revealed by comparing the intensity of methylated bands (Figure 5B). Note that the unmethylated band in Figure 5B represents both endogenous Xist and Xist cDNA Tg. Extra bands observed in Figure 5B originated from the Xist cDNA Tg, which was obvious to identify due to the absence of an EcoRI site and the presence of multiple HpaII sites in the tetracycline inducible promoter and its flanking sequence. Rescuing Eed in the XΔY Eed−/− cells (XΔY Eed-TG) resulted in partial reversion of CpG methylation, which is in contrast to a previous report showing that a Tsix mutation did not affect the methylation status of Xist locus in males (Sun et al, 2006). Given that the Xist locus in the XΔY Eed−/− ES cells takes an open chromatin configuration, as was shown by reduced CpG methylation and Xist hyperactivation upon differentiation, we suggest that, once opened, the chromatin cannot easily reset to a repressed condition by rescuing Eed activity. Figure 5.Methyl-CpG-sensitive Southern blot at the Xist promoter and exon 1. (A) SacII-digested Southern blot at the Xist promoter. Me, methylated; UnMe, unmethylated; E, EcoRI site. Position of the probe is shown in the map. Arrow 1, nonspecific band found in the XY Eed−/− and its derivative lines. (B) HpaII-digested Southern blot in the Xist exon 1. Arrows 2, 3 and 4 indicate bands originated from Xist cDNA Tg. All lanes were derived from the same gel (B) or from twin gels run in parallel (A). Download figure Download PowerPoint The XDY Eed−/− ES cells display elevated H3K4 methylation at the Xist promoter To gain further insight into the role of Eed and Tsix in Xist chromatin structure regulation, we examined the methylation of H3K4 and H3K27 and the recruitment of transcription factor IIB (TFIIB) by the ChIP assay. Here, it must be considered again that the XY Eed−/− cells and their derivatives contain an Xist cDNA Tg that includes Xist-GB1 and Xist-GB2 amplicons, but not others (Figure 6A). The H3K27m3 modification was no longer found in the XΔY Eed−/− cells, confirming that Eed is responsible for the modification that appears when Tsix is absent (Figure 6B). Rescuing Eed resulted in a clearly elevated H3K27m3 level at the Xist promoter (Xist-P) in the XΔY Eed-TG cells (P<0.0005). PRC1 and its product monoubiquitinated histone H2A (UbH2A) are linked to Xi (de Napoles et al, 2004; Schoeftner et al, 2006) and they contribute to the control of developmental regulator genes (Stock et al, 2007). We examined if UbH2A modification at the Xist promoter is affected by Tsix mutations in both Eed+/+ and −/− cell lines (Supplementary Figure S5). In all cases, the modification was nearly to the background level, and we did not detect a statistically significant difference between the wild-type and Tsix-deficient lines. The loss of Tsix in the Eed−/− background resulted in a significantly increased dimethyl-H3K4 (H3K4m2) level at the Xist promoter and gene body (Xist-GB1). Both of these amplicons are within the previously reported CTCF-flanked region, and the level of H3K4m2 was comparable with that in the report (Navarro et al, 2006) (Figure 6C). Such augmented H3K4m2 level in the XΔY Eed−/− cells was not clear outside of the CTCF-flanked region (Xist-GB2) or at 5′-portion of Tsix (5′-Tsix). We also found significantly elevated TFIIB recruitment at the Xist promoter in the XΔY Eed−/− cells (Figure 6D), although it might be just reflecting the large transcription difference of Xist gene. Upon differentiation, the H3K4m2 level at the Xist promoter persisted to be higher in the XΔY Eed−/− EB than in XY Eed−/− EB. Enhanced H3K4m2 levels are consistent with the observed Xist hyperactivation (Figure 6E). Because the H3K4m2 modification in the Xist locus disappears upon differentiation in both wild-type and Tsix mutant male ES cells (Shibata and Yokota, 2008), we conclude that the Tsix mutation in the absence of Eed resulted in persistent high H3K4m2 level around the Xist promoter. Figure 6.ChIP in male ES cells with mutations in Eed and/or Tsix. (A) Positions of PCR amplicons for ChIP. Those amplicons overlapping with Xist cDNA Tg are underlined. ChIP results (mean %IP to input) for (B) H3K27m3, (C) H3K4m2 and (D) TFIIB in undifferentiated condition. (E) ChIP for H3K4m2 in day 12 EB. Error bars represent s.d. Asterisks in the graphs indicate statistically significant difference between the XY Eed−/− and XΔY Eed−/− lines (***P<0.001; **P<0.005; *P<0.05). Differences between other cell lines are not shown for simplicity. Download figure Download PowerPoint Tsix and Eed have a synergistic role in repressing Xist Taken together with data from ChIP, RT–PCR and CpG methylation analyses, we present a summary illustrating the roles of Tsix and Eed in the regulation of Xist chromatin structure (Figure 7). Xist chromatin is most condensed in the XΔY (or XΔY Eed-TG) ES cells, followed by XY (or XY Eed-TG) and XY Eed−/− cells, and becomes highly opened in the XΔY Eed−/− ES cells. The opened Xist chromatin configuration in the XΔY Eed−/− ES cells allows Xist hyperactivation upon differentiation. These results suggest a model that Tsix transcription negatively regulates both PRC2 and H3K4 HMTase at the Xist promoter and exon 1. It has been shown that Tsix transcription prevents Eed/PRC2 recruitment to the Xist promoter in cis (Sun et al, 2006). Tsix transcription has also been reported to inhibit H3K4 methylation (Navarro et al, 2006), whereas the difference in H3K4m2 levels between the XY Eed-TG and XΔY Eed-TG cells was not prominent (Figure 6C). The increased PRC2 recruitment or elevated H3K27m3 modification may inhibit H3K4 HMTase localization or the activity at the Tsix-deficient allele. The loss of Eed alone does not result in highly opened chromatin because Tsix still inhibits H3K4 HMTase. When both Tsix and Eed are absent, augmented H3K4 HMTase activity confers highly elevated H3K4 methylation that induces ectopic Xist activation upon differentiation. The mechanism of Xist hyperactivation in the XΔY Eed−/− cells is in clear contrast to that of the physiological Xist activation in female future Xi. In the latter, Xist RNA yield is limited, and in females, Xist transcription is activated at Tsix-deficient alleles with elevated H3K27m3 modification (Sun et al, 2006). Hence, we suggest that Eed contributes in male cells to inhibit ectopic Xist activation during differentiation when Tsix transcription goes down. Figure 7.Summary of the results and a suggested model. Schematic representation of Xist chromatin structure in (A) XY (or XY Eed-TG), (B) XΔY (or XΔY Eed-TG), (C) XY Eed−/−, (D) XΔY Eed−/− ES cells and (E) female future Xi at the onset of XCI. Thick column, Xist exon 1; thin column, Xist promoter; open lollipops, H3K4m2; filled hexagons, H3K27m3; filled rectangles, methylated CpG. The darkness of the columns represents closed chromatin structure. Download figure Download PowerPoint Discussion Inability of the XDY Eed−/− ES cells to repress Xist despite intact counting We demonstrated that the male ES cells with both Eed-null and Tsix mutations underwent ectopic Xist hyperactivation upon differentiation. This result can be attributed to either defective X-chromosome counting or dysfunction in Xist gene regulation. The Tsix mutant allele generated in this study does not lose any DNA elements necessary for X-chromosome counting, because the female ES cells heterozygous for the mutation did not show aberrant counting such as two Xi or no Xi (Shibata and Lee, 2004). The counting function has been ascribed to Xite (Ogawa and Lee, 2003) and an additional region at the 5′-portion of Tsix (Morey et al, 2004; Lee, 2005). Both DNA elements are completely conserved in the XΔY Eed−/− cells, and they are not included in the Xist cDNA Tg. Similarly, the DNA elements required for homologous chromosome p
Referência(s)