A Novel cis Regulatory Element Regulates Human XIST in a CTCF-Dependent Manner

ABSTRACT

The long noncoding RNA XIST is the master regulator for the process of X chromosome inactivation (XCI) in mammalian females. Here, we report the existence of a hitherto-uncharacterized cis regulatory element (cRE) within the first exon of human XIST, which determines the transcriptional status of XIST during the initiation and maintenance phases of XCI. In the initiation phase, pluripotency factors bind to this cRE and keep XIST repressed. In the maintenance phase of XCI, the cRE is enriched for CTCF, which activates XIST transcription. By employing a CRISPR-dCas9-KRAB-based interference strategy, we demonstrate that binding of CTCF to the newly identified cRE is critical for regulating XIST in a YY1-dependent manner. Collectively, our study uncovers the combinatorial effect of multiple transcriptional regulators influencing XIST expression during the initiation and maintenance phases of XCI.

INTRODUCTION

X chromosome inactivation is a gene dosage compensatory phenomenon in mammals. XIST, the inactive X (Xi)-specific transcript, is the long noncoding RNA (lncRNA) that coordinates the process of X chromosome inactivation (XCI) in eutherian mammalian females (1–3). The phenomenon of XCI corrects for the X-linked gene dosage disparity between the males (XY) and females (XX) of mammalian species (4). The first and the foremost event for the initiation of XCI is the monoallelic and sustained upregulation of Xist which occurs in the epiblast cells during the implantation of mouse embryos. At this stage, one of the two X chromosomes is randomly chosen for silencing (5, 6). Xist/XIST lncRNA physically coats the chosen Xi in cis, initiating epigenetic reprogramming characterized by the sequential removal of active chromatin marks, RNA polymerase II (PolII) exclusion and the establishment of repressive modifications (7–10), macroH2A recruitment (11), and CpG methylation of gene promoters (12, 13). Consequently, the entire X chromosome undergoes heterochromatinization, rendering it stably inactive in a mitotically heritable manner (14–22). Since Xist/XIST lncRNA is the central player for establishing the process of chromosome-wide transcriptional silencing, it is essential to regulate its levels and function to ensure proper initiation and maintenance of XCI.

Several studies have demonstrated that the temporal activation of Xist during a specific window of development is brought about by the concerted action of a network of activators and repressors either encoded from the X-inactivation center (Xic) locus or regulating Xic (23). In addition to Xist, the Xic locus codes for a number of lncRNAs that participate in regulating Xist. One of the most critical cis regulators of Xist is its antisense lncRNA Tsix, which, unlike Xist, is exclusively produced from the future active X (Xa) and negatively regulates Xist by modulating the chromatin architecture of its promoter (24–28). Xic is partitioned into two distinct topologically associated domains (TADs)—(i) Xist TAD (∼550 kb) containing Xist and its positive regulators such as Ftx and Jpx and (ii) Tsix TAD (∼300 kb) harboring negative regulators of Xist such as Tsix and Xite. Both these TADs show opposite transcriptional behavior on the chosen Xi, with the expression of genes on Xist TAD increasing and that on Tsix TAD decreasing during differentiation of mouse embryonic stem cells (mESCs) (29, 30). A recent report highlighted the role of promoter of another long noncoding RNA, Linx, as a cis regulatory silencer of Xist. Linx is located on the Tsix TAD and serves as a silencer of Xist independent of its transcript or transcription or Tsix. However, when it is placed in Xist TAD, it serves as an enhancer of Xist (31). Besides the X-chromosomal cis-acting modulators, the autosomally encoded trans-acting factors such as the core pluripotency factors—Oct4, Sox2, Nanog, Klf4, and c-Myc—link the two developmental events of differentiation and XCI (32), by regulating Xist (33, 34), Tsix (35, 36), Rnf12 (37), and Linx (31). Thus, it is well established that Xist expression is robustly regulated by a multitude of cis and trans factors acting either synergistically or independently to ensure accurate execution of the developmentally important process of XCI.

For the past 2 decades, mouse has been the preferred model system to study the molecular pathways leading to the initiation and establishment of XCI. Our understanding of the mechanism(s) regulating XIST and the molecular dynamics of XCI in other eutherian mammals is rather limited. Deciphering the XCI pathways in multiple systems is important to address the question of conservation and evolution of the process of XCI. Although mouse and human Xist/XIST were discovered almost simultaneously (1–3), our understanding of human XIST regulation has remained poor compared to its mouse counterpart. Human and mouse XIST/Xist are functionally conserved since ectopic insertion of human XIST in murine and human cells induces XCI (17, 18, 38, 39). However, there is only 49% conservation at the sequence level, with the maximum homology observed in the first exon which harbors the repeat elements (A to F) (40–42). Most notably, Tsix, the key negative regulator of Xist in mouse, is a pseudogene in humans, and the potential regulatory sequences of XIST also do not show any conservation between humans and mouse (43–45). Hence, apart from a plethora of other factors regulating Xist, the critical mechanism governed by Tsix does not seem to be conserved. It is known that not only is Xist/XIST induced in both male and female blastocysts from the maternal as well as paternal X chromosomes but it also coats the X chromosomes, leading to partial silencing of X-linked genes in mouse but not in humans (46–48). This discrepancy can be attributed to the human-specific lncRNA, XACT, which specifically coats Xa and coaccumulates with XIST in human preimplantation as well as human embryonic stem cells (hESCs), possibly tampering with XIST silencing function (49, 50). This suggests that although the process of XCI is conserved across eutherians and is dependent on XIST RNA, it could be manifested in diverse ways in different species.

A few studies have attempted to address the regulation of human XIST. Hendrich et al. compared the putative promoter sequences from human, horse, rabbit, and mouse and discovered that the first 100 bp upstream of the transcription start site (TSS) exhibits maximum conservation and hence assigned it as the promoter of human XIST (51). Through a series of in vitro biochemical assays, the authors identified three transcription factors—SP1, YY1, and TBP—as potential regulators of XIST. The question whether any of these factors can bind XIST promoter in cells and regulate its expression was addressed by two independent studies wherein YY1 was uncovered as the key transcription factor activating XIST transcription (52, 53). Here, we demonstrate for the first time that besides YY1, XIST transcription is also under the control of pluripotency factors—OCT4, SOX2, and NANOG as well as CTCF. More specifically, XIST is repressed by the pluripotency factors and activated by YY1 and CTCF. We report the existence of a novel cis regulatory element (cRE) at the XIST locus located in the first exon that seems to be bound by the aforementioned factors in a sex-specific and X inactivation-dependent manner. Further, this element could presumably act as a crucial determinant of transcriptional outcome from the XIST promoter during the initiation as well as maintenance phases of XCI.

RESULTS

Cell line models provide the context for initiation and maintenance phases of XCI.

XCI can be broadly categorized into two stages: (i) initiation, when differentiating embryonic stem cells undergo XCI for the first time, and (ii) maintenance, where any cell carrying an inactive X ensures continued expression of XIST RNA from the inactive X upon subsequent cell divisions. In the case of human XCI, YY1 is the only known regulatory factor that influences XIST expression from the inactive X. However, the question of alternate regulatory molecules affecting XIST expression is still unanswered. This is especially of importance in light of the fact that the antisense RNA TSIX is truncated/nonfunctional in humans (43, 44).

To study the regulation of XIST by factors other than YY1 in the contexts of initiation and maintenance phases of XCI, we employed a human male embryonic carcinoma cell line, NTERA-2 clone D1 (NT2/D1), and differentiated cells such as HEK293T. NT2/D1 cells express pluripotency factors and small amounts of XIST as assessed by RNA fluorescent in situ hybridization (FISH) (54) and respond to retinoic acid (RA)-mediated differentiation cues to give rise to neuronal progenitors (55). Therefore, this cell line provides a good system to probe for the dynamic pattern of XIST promoter activity during the initial expression of XIST RNA. HEK293T is an epithelial cell line of female origin and exhibits consistent expression of XIST denoting the maintenance phase of XCI (56).

XIST expression is detected only in the female cell line HEK293T and not in the cell lines of male origin, NT2/D1 cells and DLD1 (epithelial cell line) (Fig. 1A). To understand the regulation of XIST during the initiation phase, we performed RA-mediated differentiation of NT2/D1 cells and show that the levels of pluripotency factors, OCT4 (POU5F1), SOX2, and NANOG, decline and that the expression of neuronal progenitor-specific marker, PAX6, is upregulated during differentiation, thus corroborating the previous report (55) (Fig. 1B and C). In this differentiation paradigm of NT2/D1 cells, XIST expression is upregulated progressively (Fig. 1D to F). RNA FISH analysis indicated that over 20% of NT2/D1 cells exhibit robust XIST RNA FISH signal upon differentiation of NT2/D1 for 5 days (Fig. 1E and F). We verified that NT2/D1 cells harbor two intact and one broken X chromosome copies by performing X chromosome paint for metaphase spreads as well as interphase nuclei (Fig. 1E). It is important to note that the pattern remains the same in undifferentiated and 5-day RA-treated cells. Despite harboring two intact X chromosomes, only about 20% of NT2/D1 cells differentiated for 5 days show XIST RNA FISH signal. Whether this is due to differential kinetics of XIST RNA expression in NT2/D1 or nonpersistent expression and/or maintenance of XIST RNA as seen in various hESC lines (57–60) awaits further investigation. Nevertheless, based on the results presented in Fig. 1A to F, we believe that NT2/D1 cells can serve as a good model to address this question, especially since hESCs have not served as an ideal model for the purpose so far.

FIG 1.

NT2/D1 and HEK293T cells provide the contexts for initiation and maintenance phases of XCI. (A) Semiquantitative RT-PCR for mature and premature XIST using cDNA prepared from HEK293T (female), NT2/D1 (male), and DLD1 (male) cells. 18S rRNA and β-actin serve as controls. (B) qRT-PCR depicting a decrease in the levels of OCT4 and NANOG and increase in PAX6 levels upon RA-mediated differentiation of NT2/D1 cells for 6 days. The x axis represents the differentiation time point, and the y axis represents the fold change normalized to 18S rRNA. Each point on the graph represents values from 5 independent experiments, and error bars represent ±standard errors of the mean (SEM). (C) Immunoblotting showing a decrease in OCT4, SOX2, and NANOG levels upon RA-mediated differentiation of NT2/D1 cells for 6 days. γ-Tubulin serves as an equal loading control. (D) qRT-PCR depicting an increase in the levels of XIST upon RA-mediated differentiation of NT2/D1 cells for 6 days. The x axis represents the differentiation time point, and the y axis represents the fold change normalized to 18S rRNA. Each point on the graph represents values from 3 independent experiments, and the error bars represent ±SEM. (E) X chromosome paint for metaphase spread and interphase nuclei from undifferentiated (0-day) and 5-day RA-treated NT2/D1 cells. RNA FISH for mature XIST for undifferentiated (0-day) and differentiating (5-day, RA) NT2/D1 cells. Arrowheads indicate the FISH signal. (F) Quantification for the RNA FISH signals in 0-day and 5-day RA-treated NT2/D1 cells. n = 3; 200 nuclei were counted for each replicate, and statistical significance was ascertained by Student’s t test.

Transcription from XIST promoter is governed by the promoter as well as exon 1 of XIST.

The pioneering study attempting to characterize the promoter of human XIST restricted its analysis to +50 to −50 bp from the TSS, since it was found to be conserved across four mammalian species (51). We chose to test the larger promoter region to uncover the unique potential regulatory elements for human XIST by cloning up to 4,408 bp upstream of XIST TSS into a promoterless luciferase reporter vector (Fig. 2A). These DNA constructs were then transfected into NT2/D1 and HEK293T cells and assayed for the presence of promoter by measuring the induced firefly luciferase activity. Similar to the first study on characterizing XIST promoter elements (51), we observed that the region of +50 bp to −51 bp was sufficient to drive the transcription of the luciferase gene (Fig. 2B and C). Also, the promoter activities for all other fragments with increasing distance from the TSS (except for the fragment +50 bp to −260 bp, bar 3 in Fig. 2B and C) remained constant compared to the vector control. As a control, +50 bp to −1,050 bp was cloned in the antisense (AS) orientation. Since XIST promoter is unidirectional, the fragment cloned in the AS direction failed to transcribe the reporter gene and hence did not exhibit any measurable reporter activity (AS, Fig. 2B and C). We also measured the reporter activities in differentiating NT2/D1 cells (0 to 2 days) and observed a significant decrease in luciferase levels (Fig. 2D). Expression of previously identified transcriptional activators of XIST, SP1 and YY1 (51), also decreased upon differentiation of NT2/D1 cells, reaffirming their roles in regulating XIST (Fig. 2E).

FIG 2.

Induction of XIST during differentiation of NT2/D1 cells is governed by the promoter as well as exon 1 of XIST. (A) The red boxes denote the fragments of the region upstream to XIST clone into pGL3 Basic luciferase reporter for testing potential promoter activity. The blue arrow indicates the first exon of XIST with the arrowhead denoting the direction of transcription and the tail indicating the TSS. The numbers flanking each red box denote the relative location of the fragment with respect to the transcription start site (TSS) of XIST. (B and C) Luciferase reporter activities for XIST promoter constructs transfected into NT2/D1 cells (B) or HEK293T (C) compared to pGL3 Basic. Firefly luciferase activities represented here are normalized to Renilla luciferase activity, which serves as an internal control. The x axis indicates DNA constructs transfected, and the y axis represents the normalized fold change in firefly luciferase activity. Each bar represents values from three independent experiments. Error bars represent ±SEM. Asterisks represent the significance over vector control per Student’s t test (****, P value < 0.0001; ***, P value < 0.001; **, P value < 0.01; ns, nonsignificant). (D) Luciferase reporter activities of +50 bp to −51 bp, +50 bp to −1,050 bp, and +50 bp to −4,408 bp promoter constructs decrease upon RA-mediated differentiation of NT2/D1 for 3 days. Firefly luciferase activities represented here are normalized to Renilla luciferase activity, which serves as an internal control. The x axis indicates differentiation time points, and the y axis represents the normalized fold change in the firefly luciferase activity. Each bar represents values from three independent experiments. Error bars represent ±SEM. Asterisks represent the significance over vector control per Student’s t test (**, P value < 0.01; *, P value < 0.05; ns, nonsignificant). (E) SP1 and YY1 proteins decrease upon RA-mediated differentiation of NT2/D1 for 6 days. γ-Tubulin serves as a loading control. (F) ChIP-seq peaks for OCT4 (SRR5642847) and NANOG (SRR5642845) for the XIST locus in human embryonic stem cell line H9 and OCT4 ChIP-seq peak for NT2/D1 cells (SRR1640282). The arrow indicates direction of XIST transcription. All peaks were confirmed to have a P value of <0.05 as reported by MACS2 callpeak function. (G) ChIP-qPCR analysis showing a decrease in OCT4, SOX2, and NANOG enrichment on the XIST at +4.5 kb (as shown in the schematic above) in undifferentiated (0-day) versus 5-day differentiated NT2/D1 cells. (H) Positive control for the ChIP of OCT4/POU5F1 in NT2D1. The positive control is selected based on the location of two existing peaks in NT2D1 cells. (I) ChIP-qPCR analysis for the control region showing enrichment of pluripotency factors on NANOG promoter. Each bar represents values from 2 independent experiments. Error bars represent the ±SEM. Asterisks represent the significance over DLD1 per Student’s t test (*, P value < 0.05; ns, nonsignificant). (J) ChIP-qPCR analysis demonstrating a change in the occupancies of YY1 on the XIST promoter-proximal region (+1.5 kb) (as shown in the schematic above) in undifferentiated (0-day) versus 5-day differentiated NT2/D1 cells. The x axis represents the immunoprecipitated factor, and the y axis represents the enrichment calculated as percent input. Each bar represents values from 2 independent experiments. Error bars represent ±SEM. Asterisks represent the significance over undifferentiated cells (0 day) per Student’s t test (*, P value < 0.05; ns, nonsignificant). (K) ChIP-qPCR analysis for YY1 on XIST promoter-proximal region (∼+1.5 kb) (as shown in the schematic above) in HEK293T (female) and DLD1 (male) cells. Each bar represents values from 3 independent experiments. Error bar represents ±SEM. Asterisks represent the significance per Student’s t test (*, P value < 0.05).

The tested promoter fragments exhibited a similar trend in both the cell lines tested. This implies that the assessed transcription activity is independent of the cellular context wherein all the necessary transcription factors are present in both cell types. This raises the question as to why XIST fails to express in the undifferentiated NT2/D1 cells. In the mouse system, the pluripotency factors negatively regulate Xist, and therefore, NT2/D1 cells, which mimic the undifferentiated embryonic stem cells and exhibit a negative correlation between XIST and pluripotency factor expression, were the logical candidates for assessing their role. Toward this, we scanned the XIST locus for binding sites of pluripotency factors by reanalyzing the available chromatin immunoprecipitation sequencing (ChIP-seq) data sets for OCT4 and NANOG in H9 hESCs and OCT4 in NT2/D1 cells and observed their enrichment on the first exon of XIST, ∼+4.5 kb from the TSS in both the cell types independent of sex (Fig. 2F). We validated their binding at this specific location by ChIP-quantitative PCR (qPCR) and observed a higher enrichment of OCT4 and SOX2 in undifferentiated (0-day) compared to 5-day-differentiated NT2/D1 cells (Fig. 2G). This indicates that binding of the pluripotency factors might serve to repress XIST in an undifferentiated state. The NANOG promoter region comprising ChIP-seq peaks for OCT4 and SOX2 served as a positive control for pluripotency factor ChIP-qPCR assay (Fig. 2H and I).

The next prominent question was to identify the factors positively regulating XIST in this scenario. YY1 has been shown to bind at the site +1.5 kb from the TSS and serve as a transcriptional activator of XIST (52, 53). Therefore, we determined YY1 binding at this site by ChIP-qPCR and observed a significant enrichment for day 5 differentiated NT2/D1 cells (Fig. 2J). We also confirmed YY1 enrichment at this site specifically in the female cells by comparing ChIP-qPCR results between female cells (HEK293T) and male cells (DLD1) (Fig. 2K). These results suggest that the enrichment of the pluripotency factors 4.5 kb downstream of the TSS negatively correlates with XIST expression and that occupancy of YY1, the known transcriptional activator of XIST, at +1.5 kb correlates with the induction of XIST in the NT2/D1 differentiation model. Therefore, it seems likely that similarly to the mouse system, the pluripotency factors act as the potential repressors of human XIST.

Pluripotency factors negatively regulate XIST.

To validate if pluripotency factors indeed repress XIST, two approaches were employed: (i) depleting their levels in NT2/D1 cells, which express high levels of OCT4, SOX2, and NANOG, do not express XIST, and hence provide the initiation phase context, and (ii) forced overexpression of pluripotency factors in HEK293T cells that do not endogenously express these factors and exhibit high expression of XIST (maintenance phase). Knockdown of OCT4, SOX2, or NANOG in undifferentiated NT2/D1 cells (Fig. 3A) led to a significant increase in XIST levels (Fig. 3C) when either OCT4 alone was depleted or both SOX2 and NANOG were depleted (Fig. 3C). Increase in the expression of the neuronal marker PAX6 is expected upon knockdown of pluripotency factors and hence serves as a control (Fig. 3B). We would like to point out that a significant upregulation in XIST expression was observed only when the levels of all the three factors were highly reduced, which is under the condition where OCT4 is depleted.

FIG 3.

Pluripotency factors repress XIST by binding to exon 1 (+4.5 kb) site. (A) Immunoblotting to determine the knockdown efficiencies of OCT4, SOX2, and NANOG in NT2/D1 cells. β-Actin serves as an equal loading control. (B) qRT-PCR for PAX6 upon siRNA-mediated knockdown of OCT4, SOX2, and NANOG in NT2/D1 cells. The x axis represents siRNA transfected, and the y axis represents the fold change normalized to 18S rRNA. Each bar represents values from 3 independent experiments. Error bars represent the ±SEM. Asterisks represent the significance over vector control per Student’s t test (**, P value < 0.01; *, P value < 0.05; ns, nonsignificant). (C) qRT-PCR for mature XIST upon siRNA-mediated knockdown of OCT4, SOX2, and NANOG in NT2/D1 cells. The x axis represents siRNA transfected, and the y axis represents the fold change normalized to 18S rRNA. Each bar represents values from 3 independent experiments. Error bars represent the ±SEM. Asterisks represent the significance over vector control per Student’s t test (*, P value < 0.05; ns, nonsignificant). (D) Experimental scheme to overexpress OCT4, SOX2, and NANOG in NT2/D1 cells differentiated for 4 days. (E) Immunoblotting for FLAG to confirm the overexpression of OCT4, SOX2, and NANOG in NT2/D1 cells differentiated for 4 days. γ-Tubulin serves as an equal loading control. (F) qRT-PCR showing a significant reduction in mature XIST upon overexpression of pluripotency factors in NT2/D1 cells differentiated for 4 days. The x axis represents transfected DNA, and the y axis represents the fold change normalized to 18S rRNA. Each point on the graph represents values from 3 independent experiments, and error bars represent ±SEM. Asterisks represent the significance over vector control per Student’s t test (****, P value < 0.0001; ***, P value < 0.001; *, P value < 0.05). (G) Immunoblotting for FLAG to confirm the overexpression of OCT4, SOX2, and NANOG in HEK293T cells. γ-Tubulin serves as an equal loading control. (H) qRT-PCR showing a significant reduction in mature and premature XIST upon overexpression of pluripotency factors in HEK293T cells. The x axis represents the mature or premature XIST, and the y axis represents the fold change normalized to 18S rRNA. Each point on the graph represents values from 5 independent experiments, and error bars represent ±SEM. Asterisks represent the significance over vector control per Student’s t test (****, P value < 0.0001; ***, P value < 0.001; *, P value < 0.05). (I) ChIP-qPCR showing occupancies of OCT4, SOX2, and NANOG on the exon 1 (+4.5 kb) site upon their overexpression in HEK293T cells. The x axis represents the transfected DNA, and the y axis represents the enrichment calculated as percent input. Each point on the graph represents values from 2 independent experiments, and error bars represent ±SEM. Asterisks represent the significance over vector control per Student’s t test (*, P value < 0.05).

To further test the repressive role of pluripotency factors, we overexpressed OCT4, SOX2, or NANOG in 4-day-differentiated NT2/D1 cells per the scheme depicted in Fig. 3D. The 4-day differentiation time point was chosen since pluripotency factors begin to decline (Fig. 1B and C) and XIST expression is induced (Fig. 1D) at this time point. A significant decline in XIST expression was observed under these experimental conditions, thereby strengthening the proposed repressive role for OCT4, SOX2, and NANOG (Fig. 3E and F). These results led us to conclude that OCT4, SOX2, and NANOG act as the repressors of XIST in embryonic stem cells, wherein it is necessary to prevent premature induction of XIST.

The counterexperiment of ectopically overexpressing OCT4, SOX2, or NANOG in HEK293T cells (Fig. 3G) led to a significant decrease in mature XIST expression as assessed by quantitative real-time PCR (qRT-PCR) primers designed at the exon-exon junction (Fig. 3H). A similar trend of decrease was obtained for premature XIST, assessed by primers targeting the intronic region, which suggests that the observed effect is indeed an outcome of transcription and not a posttranscriptional mechanism (Fig. 3H). The overexpressed OCT4 and SOX2 proteins are enriched at the +4.5-kb region on exon 1 of XIST, providing clear evidence that the pluripotency factors negatively regulate transcription of XIST by directly binding to this site (Fig. 3I). Considering the effect of OCT4, SOX2, and NANOG on XIST expression in HEK293T cells, we postulate that occupancy of the +4.5-kb site by the pluripotency factors could exert a more potent effect on the expression of XIST in conjunction with the binding of the known activator YY1 on the promoter-proximal region.

The pluripotency factor-bound regulatory site on XIST exon 1 is a potential cis regulatory element (cRE).

The results obtained thus far provided compelling evidence to further characterize the regulatory region of the human XIST exon 1 (+4.5 kb from the TSS). It seemed likely that this element harbors regulatory potential whose function is modulated by the binding of pluripotency factors. To assess the functional significance of this site, we performed ChIP-qPCR for active (H3K27ac) and inactive (H3K27me3) histone modifications in female (HEK293T) as well as male (DLD1) cells and observed a significant enrichment of active chromatin marks only in the female cells, indicating that we are indeed scoring for the XIST locus from the Xi (Fig. 4A). Additionally, there was no significant difference between H3K27me3 enrichment on this site in the male and female cells suggesting it to be enriched on the XIST locus of Xa (Fig. 4A). Another feature of the newly identified cRE is the significant enrichment of CTCF specifically in female cells (MCF7, HeLa, HEK293, and female skin epithelium) versus male cells (DLD1 and male skin epithelium) as evident from the analysis of available ChIP-sequencing data sets (61, 62) (Fig. 4B). The inclusion of primary cell data lends confidence in the findings and excludes the possibility of observing CTCF occupancy merely as an artifact of cultured cell lines. A previous report ruled out the role of CTCF in regulating XIST since the authors observed no appreciable difference in its enrichment on the part of XIST locus tested between male and female cells (52). Nonetheless, this report has not examined the site identified in our study, and hence, we investigated its role in governing XIST transcription.

FIG 4.

Pluripotency factor binding element is the potential cRE. (A) ChIP-qPCR analysis showing enrichment of active histone mark H3K27ac and repressive histone mark H3K27me3 for XIST cRE (+4.5 kb, as shown in the schematic above) on exon 1 in HEK293T (female, blue bar) and DLD1 (male, brown bar) cells. The x axis represents the antibodies used for ChIP, and the y axis represents the enrichment calculated as percent input. Each bar represents values from 3 independent experiments. Error bars represent ±SEM. Asterisks represent the significance over DLD1 ChIP per Student’s t test (**, P value < 0.01; ns, nonsignificant). (B) ChIP-seq peaks for OCT4 (SRR5642847) and NANOG (SRR5642845) for the XIST locus in human embryonic stem cell line H9 and OCT4 ChIP-seq peak for NT2/D1 cells (SRR1640282). CTCF ChIP-seq for male cell lines, DLD1 (DRR014660) and skin epithelium (SRR6213724), and female cell lines, MCF7 (SRR577680, SRR577679), HeLa (SRR227659, SRR227660), HEK293 (DRR014670), and skin epithelium (SRR6213076) lines. The arrowhead indicates the direction of XIST transcription, and the tail denotes the TSS. All peaks were confirmed to have a P value of <0.05 as reported by MACS2 callpeak function. (C) Immunoblotting to confirm siRNA-mediated knockdown of SP1, YY1, and CTCF in HEK293T cells. γ-Tubulin serves as an equal loading control. (D) qRT-PCR demonstrating reduction in XIST (both mature and premature) levels upon knockdown of YY1 or CTCF in HEK293T cells. The x axis represents the mature or premature XIST, and the y axis represents the fold change normalized to 18S rRNA. Each point on the graph represents values from 4 independent experiments, and error bars represent ±SEM. Asterisks represent the significance over vector control per Student’s t test (****, P value < 0.0001; **, P value < 0.01; *, P value < 0.05; ns, nonsignificant). (E) Immunoblotting to confirm overexpression of SP1, YY1, and CTCF in NT2/D1 cells. γ-Tubulin serves as an equal loading control. (F) qRT-PCR demonstrating a significant increase in XIST expression upon overexpression of YY1 or CTCF in NT2/D1 cells. The x axis represents the transfected DNA, and the y axis represents the fold change normalized to 18S rRNA. Each point on the graph represents values from 3 independent experiments, and error bars represent ±SEM. Asterisks represent the significance over vector control per Student’s t test (****, P value < 0.0001; *, P value < 0.05; ns, nonsignificant). (G) (i and ii) ChIP-qPCR analysis showing enrichment of CTCF on XIST cRE (+4.5 kb) on exon 1 in HEK293T (female (i) and DLD1 (male (ii) cells. (iii) Sequential ChIP in HEK293T cells. (iv and v) ChIP-qPCR analysis for CTCF (iv) and H3K27ac and H3K27me3 (v) upon siRNA-mediated knockdown of CTCF in HEK293T cells (siCTCF). Each bar represents values from 3 (for panel i) or 2 (for panels ii to v) independent experiments. Error bars represent ±SEM. Asterisks represent the significance per Student’s t test (*, P value < 0.05). (H) Immunoblotting for OCT4 and SOX2 upon the knockdown of OCT4 and for FLAG to confirm the overexpression of SP1, YY1, and CTCF in NT2D1 cells. γ-Tubulin serves as a loading control. (I) qRT-PCR for mature XIST upon knockdown of OCT4 and overexpression of SP1, YY1, and CTCF in NT2D1 cells. The x axis represents DNA and siRNA transfected, and the y axis represents the fold change normalized to 18S rRNA. Each bar represents values from 2 independent experiments. Error bars represent ±SEM. Asterisks represent the significance over vector control per Student’s t test (**, P value < 0.01; *, P value < 0.05; ns, nonsignificant).

Perturbing the levels of SP1, YY1, and CTCF in HEK293T cells caused a significant decrease in both mature and premature XIST levels upon knockdown of YY1 or CTCF (Fig. 4C and D). Overexpression of SP1, YY1, or CTCF in undifferentiated NT2/D1 cells which express XIST at very low levels led to a significant increase in XIST levels upon overexpression of YY1 or CTCF (Fig. 4E and F). These results strongly suggest that in addition to YY1, CTCF is an essential regulator of XIST. Although SP1 was identified to bind near the XIST TSS (51), it does not play any role in transcriptional regulation of XIST. Furthermore, we validated CTCF ChIP-sequencing analysis findings by performing ChIP-qPCR for CTCF in HEK293T and DLD1 cells (Fig. 4G, i and ii) and observed enrichment on the cRE in the female cells by an order of magnitude, which positively correlates with XIST expression (Fig. 4D). To confirm whether CTCF is indeed bound to cRE only on the Xi, we performed sequential ChIP-qPCR in HEK293T cells using anti-H3K27ac antibody (mark enriched at +4.5-kb site only in female cells as per Fig. 4A) or anti-H3K27me3 antibody (mark enriched at +4.5-kb site in both female and male cells as per Fig. 4A) for the first ChIP followed by a second ChIP using anti-CTCF antibody. We observed a higher enrichment of CTCF when the first ChIP was performed using H3K27ac antibody (Fig. 4G, iii, blue bar) compared to H3K27me3 antibody (Fig. 4G, iii, brown bar). Further confirmation for the direct role of CTCF in regulating XIST transcription is evidenced by the decreased enrichment of CTCF (Fig. 4G, iv) and increased H3K27me3 levels at the cRE observed upon knockdown of CTCF in HEK293T cells (Fig. 4G, v).

We investigated the interplay between transcriptional activators and repressors of XIST by overexpressing potential transcriptional activators, SP1 or YY1 or CTCF coupled with the knockdown of OCT4 in NT2/D1 cells. Interestingly, we observed a significantly higher upregulation of XIST when the overexpression of YY1 or CTCF was coupled with the knockdown of OCT4 (Fig. 4H and I), compared with only the overexpression of YY1 or CTCF (Fig. 4E and F) or the knockdown of pluripotency factors alone (Fig. 3A and B). This suggests that the balance of activators (YY1 and CTCF) and repressors (OCT4, SOX2, and NANOG) governs the transcriptional activation of XIST. It is noteworthy that the small interfering RNA (siRNA)-mediated knockdown of OCT4 depletes not only OCT4 but also SOX2 (Fig. 4H). Altogether, these results indicate the existence of a cRE on exon 1 of XIST whose function is modulated by CTCF or pluripotency factors, which in turn determines transcriptional status of XIST.

CTCF binding to cRE assists in stabilizing YY1 binding at the XIST promoter-proximal site.

YY1 has been demonstrated to be a conserved transcriptional activator of XIST (52, 53), which is also recapitulated in our study. In order to test the functional significance of CTCF in regulating XIST, we assessed occupancy of YY1 at a previously identified promoter-proximal region (+1.5 kb) upon knockdown of CTCF. Interestingly, while the levels of YY1 do not change upon depleting CTCF in HEK293T cells (Fig. 4C), a significant decrease in its enrichment can be observed at the promoter-proximal region (Fig. 5A). Furthermore, we show that the downregulation in XIST observed upon depleting CTCF can be rescued by overexpressing YY1 in HEK293T cells (Fig. 5B and C). This suggests that CTCF assists in either the recruitment or maintenance of YY1 binding at the promoter-proximal region. Based on these results, we envisaged the significance of CTCF binding to cRE as an important determinant of XIST transcription. To test this possibility, we employed the CRISPR-dCas9-KRAB-directed approach to interfere with CTCF binding and the active chromatin mark at the cRE (Fig. 5D) and observed a significant downregulation in mature as well as premature XIST levels (Fig. 5E). ChIP-qPCR analysis indicated enrichment of dCas9 (monitored by FLAG ChIP) and H3K9me3 mark at cRE, suggestive of efficient targeting (Fig. 5F). This led to a concomitant decrease in CTCF occupancy at cRE (Fig. 5F). In accordance with our CTCF knockdown results (Fig. 5A), YY1 occupancy at the promoter-proximal region declined under these conditions as well (Fig. 5G). We also scored for H3K9me3 at the promoter-proximal region to determine if the observed XIST downregulation upon targeting CRISPR-dCas9-KRAB to cRE is an outcome of spreading of the repressive mark to the promoter. Surprisingly, we observed a decrease in the H3K9me3 mark at the promoter-proximal region. The reason for this is not fully understood and needs further investigation. Based on these results, we infer that binding of CTCF to the newly identified cRE regulates XIST in a YY1-dependent manner. It is imperative to point out that the primers used to score for mature XIST and premature XIST by qRT-PCR are downstream of the dCas9-KRAB target site (cRE). Therefore, the possibility of dCas9-KRAB acting as a roadblock to the elongating RNA PolII as reported previously (63, 64) cannot be ruled out. However, we also observed a significant decrease in YY1 binding at the promoter-proximal region (Fig. 5G), which strongly supports the role of CTCF-bound cRE in assisting in either recruitment or maintenance of YY1 binding. Most notably, the magnitude of decrease in the levels of mature and premature XIST strongly correlates with YY1 enrichment at the promoter-proximal region (compare Fig. 5A and C with Fig. 5E and G). Hence, the transcriptional downregulation of XIST upon dCas9-KRAB targeting to cRE is most likely the direct consequence of perturbing CTCF binding to cRE and not an outcome of blocking elongating RNA PolII.

FIG 5.

CTCF assists in the recruitment or maintenance of YY1 binding to the promoter-proximal region of XIST. (A) ChIP-qPCR analysis showing a reduction of YY1 occupancy on the promoter-proximal site (+1.5 kb) upon siRNA-mediated knockdown of CTCF in HEK293T cells (siCTCF). The x axis represents the antibody used for ChIP, and the y axis represents the normalized fold enrichment over control siRNA. Each bar represents values from 3 independent experiments. Error bars represent ±SEM. Asterisks represent the significance over YY1 ChIP for vector control per Student’s t test (*, P value < 0.05). The x axis represents the transfected siRNA, and the y axis represents the normalized fold enrichment over control siRNA. (B) Immunoblotting to confirm knockdown of CTCF and overexpression of FLAG-tagged YY1 in HEK293T cells. γ-Tubulin serves as an equal loading control. (C) qRT-PCR for mature XIST upon knockdown of CTCF and/or overexpression of FLAG-YY1 in HEK293T cells. The x axis represents the DNA and siRNA transfected, and the y axis represents the normalized fold enrichment over 18S rRNA. Each bar represents values from 2 independent experiments. Error bars represent ±SEM. Asterisks represent the significance over control (first bar) per Student’s t test (***, P value < 0.001; ns, nonsignificant). (D) Schematic depicting dCas9-KRAB-based repression strategy. (E) qRT-PCR demonstrating reduction in XIST (both mature and premature) levels upon transfecting dCas9 + sgRNA in HEK293T cells. The x axis represents the mature or premature XIST, and the y axis represents the fold change normalized to dCas9-only control. Each point on the graph represents values from 6 independent experiments, and error bars represent ±SEM. Asterisks represent the significance over vector control per Student’s t test (****, P value < 0.0001). (F) ChIP-qPCR analysis showing enrichment of dCas9, CTCF, and H3K9me3 at cRE (+4.5 kb, as shown in the schematic above) on exon 1 in HEK293T cells transfected with just dCas9 or dCas9 + sgRNA. The x axis represents the antibodies used for ChIP, and the y axis represents the enrichment calculated as percent input. Each bar represents values from 3 independent experiments. Error bars represent ±SEM. Asterisks represent the significance over dCas9 control per Student’s t test (*, P value < 0.05). (G) ChIP-qPCR analysis showing enrichment of YY1 and H3K9me3 at promoter-proximal region (+1.5 kb, as shown in the schematic above) on exon 1 in HEK293T cells transfected with just dCas9 or dCas9 + sgRNA. The x axis represents the antibodies used for ChIP, and the y axis represents the enrichment calculated as percent input. Each bar represents values from 3 independent experiments. Error bars represent ±SEM. Asterisks represent the significance over dCas9 control per Student’s t test (****, P value < 0.0001; *, P value < 0.05).

DISCUSSION

XIST is the critical regulator of XCI, a mechanism to correct for the gene dosage imbalance between males and females of the placental mammals (1–4). Being central to the process of XCI, it is essential to modulate XIST levels temporally to ensure faithful execution of the process and hence the developmental programs. Despite significant conservation of X-linked genes, chromosomal synteny as well as the process of XCI guided by XIST lncRNA between several eutherian mammals, XCI is initiated in diverse ways in different species. One of the possible explanations could be evolutionary alterations in the developmental programs across species necessitating robustness in the regulation of XCI to accommodate these changes (46). Extensive studies using the mouse system have highlighted the multiple ways in which the transcription of Xist is regulated during the initiation and the maintenance phases of XCI. Thus far, only YY1 has been discovered to be the common factor regulating XIST in mouse and human (52, 53). Here, we have uncovered a complex regulatory network involving pluripotency factors, YY1, and CTCF that shapes the transcriptional outcome from the XIST promoter (Fig. 6).

FIG 6.

A model illustrating the role of CTCF-bound cRE in dictating the transcription from XIST promoter. (A) In undifferentiated ES cells, cRE (+4.5 kb) of XIST is bound by the pluripotency factors (OCT4, SOX2, and NANOG) keeping XIST repressed. Upon differentiation, levels as well as enrichment of pluripotency factors on the cRE decrease. Subsequently, YY1 now occupies the promoter-proximal site, leading to induction of XIST. (B) In differentiated cells, CTCF binding to cRE (+4.5 kb) enables either the recruitment or maintenance of YY1 at the promoter-proximal region, maintaining persistent expression of XIST. Abrogation of CTCF binding to cRE by CRISPR-dCas9-KRAB interference disrupts YY1 binding to the promoter-proximal region, causing downregulation of XIST.

Unlike mouse ES cells, the human counterpart has not served as a tractable model to address the aspects of XCI and XIST regulation. Multiple groups have reported considerable variation in XIST expression as well as XCI status between human ES lines over passages (58, 65). However, more recent studies describe a method of culturing human ES cell lines, exhibiting consistent XIST expression as well as XCI upon differentiation (66, 67). The difficulty of human ES cells as a model notwithstanding, we have worked with NT2/D1 and HEK293T as models of initiation and maintenance phases of XCI, respectively. Moreover, we have reanalyzed the data relevant to our study from multiple cell types including a human ES line and primary tissues of female and male origin to discover physiologically relevant Xi-specific factors contributing to XIST regulation.

To address XIST regulation in the context of the initiation phase of XCI, we have extensively worked with NT2/D1 cells, a human male embryonic carcinoma cell line. The NT2/D1 cell line provides the context for the initiation phase of XCI since it harbors properties similar to the embryonic stem cells and has been extensively used to understand the human stem cell biology (68–70). More importantly, it also expresses low levels of XIST (54), analogous to those seen for human and mouse ES cells as well as blastocysts (5, 46–48, 71, 72). We show for the first time that XIST expression can be induced in these cells upon RA-mediated differentiation, and hence, NT2/D1 can serve as a model to determine the induction of XIST. The fact that there is a negative correlation between XIST induction and loss of OCT4, SOX2, and NANOG during differentiation prompted us to speculate their role in governing XIST transcription. Moreover, there have been studies reporting pluripotency factors to be important positive and negative regulators of mouse Xist and Tsix, respectively (33, 35, 36). Upon reanalyzing published ChIP-sequencing data sets and the experimental validation, we provide evidence that pluripotency factors repress XIST. Interestingly, although levels of YY1, a known transcriptional activator of XIST, decline during differentiation, its occupancy on the previously reported promoter-proximal site (+1.5 kb) increases, causing a significant increment in XIST expression by day 5 of differentiation (52, 53). We believe that regulation mediated by the pluripotency network could be an essential way to control upregulation of XIST temporally. These findings provide compelling evidence suggestive of this gene body regulatory region being a potential Xi-specific cis regulatory element. Indeed, a higher enrichment of activation-associated histone marks—H3K27ac on the region of interest specifically in the female—strongly suggests it to be the case.

We also report that cRE is specifically enriched for CTCF specifically in female cells. A role of CTCF in XIST regulation was ruled out based on the observation that it does not exhibit a differential enrichment between male and female fibroblasts or XIST⁺ and XIST⁻ human ES cells (52). However, the results obtained in our study unequivocally demonstrate it to be specifically enriched on the cRE only on Xi. By performing knockdown and overexpression experiments, we show that XIST is positively regulated by CTCF in addition to YY1. Interestingly, XIST levels increase dramatically when overexpression of YY1 or CTCF is coupled with the knockdown of OCT4 and SOX2 in NT2/D1 cells, suggesting that the balance of activators (YY1 and CTCF) and repressors (pluripotency factors) determines XIST transcriptional status.

Since overexpression- and knockdown-based results can be confounded by indirect effects, we made use of a CRISPR-dCas9-KRAB-based interference approach to specifically target cRE and address its significance more directly. Findings obtained with this strategy corroborated the significance of the newly identified cRE in manifesting transcriptional regulation of XIST in a CTCF- and YY1-dependent manner. Therefore, we propose that at the onset of XCI, the cRE is kept repressed by pluripotency factors. However, in the differentiated cells which do not express pluripotency factors, cRE on the Xi is now bound by CTCF (Fig. 6 legend). As a result, XIST continues to be synthesized from Xi and is involved in maintaining silencing. Whether the inconsistency in XIST expression observed in various female human ES lines could plausibly be a consequence of the variability in the cRE function would be a promising line of future investigation.

In conclusion, we have identified a novel cRE for XIST/Xist that is controlled by developmentally regulated factors, OCT4, SOX2, and NANOG as the repressors and the chromatin organizer CTCF/YY1 as activators. Since all the identified regulators are expressed to similar extents in male and female cells, it is intriguing that they are able to regulate the XIST expression in a Xi-specific manner. This is suggestive of additional factors/mechanisms contributing toward such specificity, which warrants further investigation. Determining if this feature is conserved in other eutherians as well will provide useful clues regarding the conserved mechanisms governing XIST transcription and provide key insights into evolutionary conservation of the phenomenon of X inactivation.

MATERIALS AND METHODS

Cell culture.

Human embryonic carcinoma cell line NT2/D1 (NTERA2-clone D1) was obtained as a kind gift from Peter Andrews, University of Sheffield, United Kingdom. The cells were grown in Dulbecco’s modified Eagle’s medium with sodium pyruvate, high glucose (DMEM; Sigma-Aldrich); supplemented with 10% fetal bovine serum, 2 mM l-glutamine, and penicillin-streptomycin (Invitrogen); and maintained at 37°C under a 5% CO₂ atmosphere. NT2/D1 cells were passaged upon reaching 70% confluence by gentle scraping, and trypsin was not used for passaging. Human embryonic kidney cells (HEK293T) were grown in DMEM without sodium pyruvate, high glucose; supplemented with 10% fetal bovine serum and penicillin-streptomycin (Invitrogen); and maintained at 37°C under a 5% CO₂ atmosphere. HEK293T cells were maintained passaged upon reaching 70% confluence using 0.05% trypsin. DLD1 cells were grown in RPMI supplemented with 10% fetal bovine serum and penicillin-streptomycin and maintained at 37°C under a 5% CO₂ atmosphere. DLD1 cells were passaged upon reaching 70% confluence using 0.05% trypsin.

RA-mediated differentiation of NT2/D1 cells.

All-trans-retinoic acid (RA) (procured from Sigma-Aldrich) used for inducing differentiation of NT2/D1 cells was reconstituted at a concentration of 5 mg/ml in dimethyl sulfoxide (DMSO) and stored at −80°C. For differentiation experiments, NT2/D1 cells were harvested using 0.05% trypsin, resuspended in fresh medium, and seeded at a density of 0.15 × 10⁶ cells in a 6-well plate or 1 × 10⁶ cells in a 100-mm tissue culture dish (Corning). Cells were allowed to grow for 24 h, following which a few cells were harvested for RNA and protein extractions or cross-linked for ChIP as 0-day control. RA was added to the remaining wells/plates at a concentration of 13.7 μM for the remaining 6 days. Each day, cells were either replenished with fresh medium and RA or harvested as day 1, 2, 3, 4, 5, or 6 samples.

Molecular cloning.

To characterize the XIST promoter, genomic regions ranging from +50 bp to −4,408 bp upstream of the XIST TSS were cloned into a promoter-less reporter vector, pGL3 Basic, procured from Promega. XIST genomic regions were PCR amplified using the genomic DNA extracted from NT2/D1cells using specific primer pairs (Table 1) followed by restriction enzyme digestion of the vector and insert and ligation using T4 DNA ligase (New England Biolabs). The positive clones were confirmed by automated Sanger sequencing.

TABLE 1.

List of oligonucleotide primers used^a

Description	Sequence (5′ to 3′)
Primers used for molecular cloning
+50 to −51 bp, forward	CTGCAGCAGCGAATTGCAG
+50 to −51 bp, reverse	CAAAGATGTCCGGCTTTC
+50 to −260 bp, forward	CTGCAGCAGCGAATTGCAG
+50 to −260 bp, reverse	GCAGTTTATGGAGGATTTTAGC
+50 to −560 bp, forward	CTGCAGCAGCGAATTGCAG
+50 to −560 bp, reverse	GGAATGGGAAGTCCCTTGAAG
+50 to −717 bp, forward	CTGCAGCAGCGAATTGCAG
+50 to −717 bp, reverse	GCCATTCTATGAAATGTCTTTC
+50 to −1,050 bp, forward	CTGCAGCAGCGAATTGCAG
+50 to −1,050 bp, reverse	GCCAGTGGGAGGGTAATGTA
+50 to −1,050 bp, antisense, forward	CTGCAGCAGCGAATTGCAG
+50 to −1,050 bp, antisense, reverse	GCCAGTGGGAGGGTAATGTA
+50 to −2,624 bp, forward	CTGCAGCAGCGAATTGCAG
+50 to −2,624 bp, reverse	TGGAGCCAAGCAGTAGTGAA
+50 to −4,408 bp, forward	CTGCAGCAGCGAATTGCAG
+50 to −4,408 bp, reverse	TTCCTCCCTCTCCCTAGTGTTT
Human SP1, forward	ATGAGCGACCAAGATCACTCC
Human SP1, reverse	TCAGAAGCCATTGCCACTG
Human YY1, forward	ATGGCCTCGGGCGACACCCTC
Human YY1, reverse	TCACTGGTTGTTTTTGGCC
Human CTCF, forward	ATGGAAGGTGATGCAGTCG
Human CTCF, reverse	TCACCGGTCCATCATGCTGAG
Human OCT4, forward	ATGGCGGGACACCTGGCTTC
Human OCT4, reverse	TCAGTTTGAATGCATGGGAG
Human SOX2, forward	ATGTACAACATGATGGAGAC
Human SOX2, reverse	TCACATGTGTGAGAGGGGCAG
Human NANOG, forward	ATGAGTGTGGATCCAGCTTG
Human NANOG, reverse	TCACACGTCTTCAGGTTGC
Primers used for qRT-PCR
Human GAPDH, forward	CTGCACCACCAACTGCTTAG
Human GAPDH, reverse	GTCTTCTGGGTGGCAGTGAT
Human 18S rRNA, forward	CGCCGCTAGAGGTGAAATTCT
Human 18S rRNA, reverse	CGAACCTCCGACTTTCGTTCT
Human mature XIST, forward	ACATGCCTGGCACTCTAGCA
Human mature XIST, reverse	AAACATGGAAATGGGTAAGACACA
Human premature XIST, forward	TGCTTTAGCATCAAAGCCCT
Human premature XIST, reverse	GCCTTAGATTCCCAGTTCCA
Human OCT4, forward	AGCAAAACCCGGAGGAGT
Human OCT4, reverse	CCACATCGGCCTGTGTATATC
Human SOX2, forward	TGCTGCCTCTTTAAGACTAGGAC
Human SOX2, reverse	CCTGGGGCTCAAACTTCTCT
Human NANOG, forward	CCTGAACCTCAGCTACAAACAG
Human NANOG, reverse	GCTATTCTTCGGCCAGTTGT
Human PAX6, forward	GGCACACACACATTAACACACTT
Human PAX6, reverse	GGTGTGTGAGAGCAATTCTCAG
Primers used for ChIP-PCR
Human XIST locus promoter proximal (+1.5 kb) (hg19, chrX 73070859 + 73071010), forward	TGCTAATTCACCCAGGTCTTC
Human XIST locus promoter proximal (+1.5 kb) (hg19, chrX 73070859 + 73071010), reverse	GAGAAAAGGTGGGATGGACA
Human XIST locus cRE (+4.5 kb) (hg19, chrX 73,067,862–73,068,061), forward	AGGGAAGTGAGTGGGGTCTT
Human XIST locus cRE (+4.5 kb) (hg19, chrX 73,067,862–73,068,061), reverse	TTACAGCAGGGGGTACTTGG
Human Nanog promoter (hg19, chr12 8044771–8044882), forward	GCTGGGTTTGTCTTCAGGTT
Human Nanog promoter (hg19, chr12 8044771–8044882), reverse	TCCCGTCTACCAGTCTCACC
sgRNA sequences targeting cRE
Constant oligonucleotide	AAAAGCACCGACTCGGTGCCACTTTTTCAAGTTGATAACGGACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAAC
sgRNA1	TAATACGACTCACTATActttagtgcacaggaagaggGTTTTAGAGCTAGAAATAGCAAGb
sgRNA2	TAATACGACTCACTATAgttagtgcacaagatcgtggGTTTTAGAGCTAGAAATAGCAAGb

^aGAPDH, glyceraldehyde-3-phosphate dehydrogenase; chr, chromosome.

^bThe CrisprRNA sequence targeting the XIST locus is in lowercase.

Transfection of DNA, siRNA, and gRNA.

NT2/D1 or HEK293T cells were transfected with the equimolar amounts of DNA constructs using Lipofectamine 2000 and siRNA using Lipofectamine RNAiMax (Invitrogen) per manufacturer’s guidelines. HEK293T cells were transfected with plasmid for dCas9-KRAB and guide RNA (gRNA Table 1) mix using Lipofectamine 3000 per manufacturer’s guidelines.

Luciferase reporter assay.

Luciferase reporter assays were performed using the dual luciferase assay kit from Promega. Cells were transfected with firefly luciferase and Renilla luciferase DNA constructs. After harvesting, the cells were lysed using 1× passive lysis buffer per manufacturer’s instructions. The lysates and substrates were mixed in the optical-bottom 96-well plate (ThermoFisher Scientific) according to the guidelines provided by Promega. The reporter activities were measured using luminometry mode on the Varioskan machine (ThermoFisher Scientific). In all the assays, Renilla luciferase activity measurement served as an internal control. The fold change was calculated with respect to either the vector control or 0-day control as and when mentioned.

RNA extraction and cDNA synthesis.

Total RNA was extracted using the TRIzol reagent (Invitrogen). One to 2 μg of RNA was used for DNase treatment per the protocol provided by the manufacturer (Promega). This was followed by cDNA synthesis using reverse transcriptase kits from Applied Biosystems (high-capacity cDNA synthesis kit). Additionally, minus-reverse transcriptase (RT) control was set up to verify the efficiency of DNase treatment. The synthesized cDNA was used to set up quantitative real-time PCR (qRT-PCR).

Quantitative real-time PCR.

Quantitative real-time PCR was performed using SYBR green chemistry (Roche) with a specific set of primer pairs (Table 1) using the ViiA7 thermal cycler (Applied Biosystems). Changes in threshold cycles (C_T) were calculated by subtracting the C_T values of the gene of interest from that of housekeeping control (for qRT-PCR) [C_T(target genes) − C_T(18S rRNA)]. ΔC_T values of specific target genes from the experimental samples were then subtracted from their respective control samples to generate ΔΔC_T values. The fold changes were calculated using the following formula: 2^[−(ΔΔC_T value)]. For quantification after ChIP, DNA recovered post-ChIP and percent input are used as the templates for the PCR with a specific set of primers. For quantification of enrichment, the efficiency of chromatin immunoprecipitation of particular genomic locus was calculated as follows:

$$\%\mathtt{\left(}\mathrm{ChIP/total\hspace{0.17em}input}\mathtt{\right)}=2^\mathtt{\left\{}\mathtt{\left[}{C}_T\mathtt{\left(}X\%\hspace{0.17em}\mathrm{input}\mathtt{\right)}\mathrm{\hspace{0.17em}-\hspace{0.17em}}\log \mathtt{\left(}x\%/{\log }_2\mathtt{\right)}\mathrm{\hspace{0.17em}-\hspace{0.17em}}{C}_T\mathtt{\left(}\mathrm{ChIP}\mathtt{\right)}\mathtt{\right]}\mathtt{\right\}}\times 100\%$$

Relative occupancy was calculated as a ratio of specific signal over background: occupancy = % input (specific loci)/% input (background loci).

Protein extraction and immunoblotting.

Cell pellets were resuspended in RIPA buffer (10 mM Tris [pH 8.0], 1 mM EDTA [pH 8.0], 0.5 mM EGTA, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, 140 mM NaCl) containing 1× protease inhibitors (procured from Roche) and lysed by repeated freeze-thaw cycles. The lysates were centrifuged at 14,000 rpm, 4°C, 30 min, to eliminate the cellular debris. The supernatant was collected in the fresh microcentrifuge tube. The concentrations of protein were determined by performing bicinchoninic acid (BCA) assay (ThermoFisher Scientific). Equal amounts of protein lysates were boiled in 1× Laemmli buffer (0.5 M Tris-HCl [pH 6.8], 28% glycerol, 9% SDS, 5% 2-mercaptoethanol, 0.01% bromophenol blue) for 10 to 15 min and subjected to electrophoresis on a polyacrylamide gel. The separated proteins were transferred onto a polyvinylidene difluoride (PVDF) membrane (Millipore) using phosphate-based transfer buffer (10 mM sodium phosphate monobasic, 10 mM sodium phosphate dibasic) at 4°C, 600 mA, 2 h. After the completion of transfer, membranes were blocked in 5% skim milk and incubated overnight at 4°C with the primary antibodies prepared in 5% BSA. The membranes were washed three times with the buffer containing 20 mM Tris buffer (pH 7.4), 500 mM NaCl, and 0.1% Tween 20 (TST) the next day and incubated with the appropriate secondary antibodies conjugated with horseradish peroxidase (HRP) for an hour at room temperature. Following this, the membranes were again washed three times with TST buffer. The blots were developed using Immobilon Western chemiluminescent HRP substrate (Millipore) and detected using ImageQuant LAS 4000 (GE Healthcare) according to the manufacturer’s instructions.

Chromatin immunoprecipitation.

ChIP was performed as per X-ChIP protocol. Briefly, 10 μg of sonicated chromatin (average length 150 to 400 bp) was incubated with 1 μg of specific antibody and incubated on an end-to-end rocker at 4°C overnight. Each ChIP was performed at least 3 times as indicated in the figure legends, and quantitative real-time PCR was set up. Primers used for the ChIP-qPCR analysis are listed in Table 1.

Sequential ChIP assay.

The method followed for the sequential ChIP assay was same as that described above. ChIP was started with 200 μg of chromatin, and the first elution was done using the buffer containing 0.1 M NaHCO₃, 1% SDS, and 10 mM dithiothreitol (DTT). The eluate was subjected to a second ChIP with either rabbit IgG or CTCF antibodies.

RNA FISH.

The probe for RNA FISH was prepared with 2 μg of bacterial artificial chromosome (BAC) RP183A17 using the nick translation kit purchased from Roche. Salmon sperm DNA and cot1 DNA were added to the probe to mask the repetitive sequence. Prior to hybridizations, the cells on the coverslips were washed with RNase-free phosphate-buffered saline (PBS) followed by permeabilization using freshly made cystoskeletal buffer {10 mM PIPES [piperazine-N,N′-bis(2-ethanesulfonic acid)] (pH 7.0), 100 mM NaCl, 300 mM sucrose, 3 mM MgCl₂} containing 0.5% Triton X-100 for 4 to 5 min on ice. Nuclei were fixed using 4% paraformaldehyde for 10 min at room temperature. The coverslips were washed twice with 70% ethanol and then subjected to dehydration by sequentially incubating the coverslips in 80%, 95%, and 100% ethanol for 3 min each, followed by air drying. The nuclei were then subjected to the hybridization step in the dark and humidified chamber at 37°C for 24 h. After incubation, the coverslips were washed three times with 50% formamide + 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) buffer (pH 7.2) for 5 min each at 42°C, followed by washing three times with 2× SSC buffer for 5 min each at 42°C, and counterstained with 4′,6-diamidino-2-phenylindole (DAPI) and washed twice with 2× SSC buffer. The coverslips were then mounted on a clean slide for microscopic observation.

Metaphase spread preparation.

A semiconfluent culture of NT2D1 cells (∼60% confluence) was subjected to metaphase arrest by treating these cells with 0.1 μg/ml colcemid for 90 min at 37°C. Cells were harvested by trypsinization, followed by hypotonic treatment (0.075 M KCl) for 30 min at 37°C, which was terminated by fixing cells with 4 to 5 drops of chilled fixative solution (methanol-acetic acid, 3:1), followed by centrifugation at 1,000 rpm for 10 min at 4°C. The cell pellet was washed three times with fixative solution to remove cell debris and resuspended in fresh fixative solution, based on the amount of cell pellet recovered. This suspension was dropped from a height onto a clean glass slide and air dried.

Two-dimensional in situ hybridization.

Whole-chromosome paint for human X chromosome (WCP-X) was obtained from Applied Spectral Imaging (ASI), Israel. WCP-X (5 μl) was equilibrated at 37°C for 5 min, denatured at 80°C for 5 min, and quick-chilled on ice for 2 min followed by a preannealing at 37°C for 30 min. The denatured probe (3 to 4 μl) was spotted onto a premarked area of a glass slide containing fixed interphase nuclei and metaphases (as prepared above), covered with an 18-mm by 18-mm coverslip, sealed with nail polish, codenatured at 72 to 74°C for 1.5 min, and hybridized for 36 to 48 h in a humidified box at 37°C. Posthybridization, the coverslip was removed carefully and the glass slides were washed in 50% formamide–2× SSC (pH 7.4), three times for 5 min each at 45°C, followed by three washes for 5 min each in 1× SSC at 45°C with gentle agitation. Slides were briefly dipped in 0.1% Tween 20–4× SSC, counterstained with DAPI for 2 min, washed in 2× SSC, mounted in antifade solution, and stored at 4°C until they were imaged. Metaphases and interphase nuclei were imaged under an Axioimager Z2 immunofluorescence microscope (Zeiss) (63× objective, numerical aperture [NA] 1.4).

gRNA generation for dCas9-KRAB-based repression of cRE.

Two single guide RNAs (sgRNA) for cRE were designed using the CHOPCHOP webtool (73). For each sgRNA, a 60-base oligonucleotide (“gene-specific oligo”) containing (i) a promoter for in vitro transcription, (ii) the 20-base spacer region specific to the target site, and (iii) an overlap region that anneals to the constant oligonucleotide was synthesized. sgRNA were generated by in vitro transcription (IVT). For IVT, a primer carrying the T7 promoter site and sgRNA sequence was annealed with the constant oligonucleotide and used for IVT. IVT was performed with the MAXIscript T7 transcription kit (ThermoFisher) using the protocol provided with the kit.

Antibodies, siRNAs, and other reagents.

SP1 (5931S) antibody for immunoblotting was procured from Cell Signaling Technology (Danvers, MA, USA), YY1 (ab12132) (sc-7341) antibodies for Western blotting were purchased from Abcam, SOX2 (AF2018) and NANOG (AF1997) antibodies for immunoblotting and ChIP were purchased from R&D (Menomonie, WI, USA), OCT4 antibody (sc-9081) for Western blotting and ChIP and goat-HRP secondary antibody were purchased from Santa Cruz Biotechnology, and CTCF antibody (07-729) procured from Millipore/Upstate was used for all of the ChIP experiments. CTCF antibody (sc-21298) purchased from Santa Cruz Biotechnology was used for immunoblotting, Normal rabbit IgG (12-370) and normal mouse IgG (12-371) for ChIP were purchased from Millipore, and β-actin (VMA00048) and γ-tubulin primary antibodies and mouse-HRP and rabbit-HRP secondary antibodies were purchased from Bio-Rad Laboratories. Histone modification antibodies used for the ChIP were as follows: H3K27ac (Abcam, ab4729), H3K27me3 (Abcam, ab6002), H3K4me3 (Millipore, 17-614), and H3K9me3 (Millipore, 07-523) The following antibodies were used for ChIP-qPCRs following dCas9-KRAB targeting: anti-FLAG (Sigma; F-3165), anti-CTCF (Abcam; 612148), anti-H3K9me3 (Millipore; 07-442), and anti-YY1 (Santa Cruz; sc-7341).

Sequencing analysis for ChIP-seq.

SRR/DRR/ERR IDs of the data sets used are provided below. Data sets used are tabulated by cell line, and the files can be downloaded from the link https://www.ebi.ac.uk/ena.

All Illumina data sets were aligned with Bowtie 2 (73) to the hg19 reference assembly, following which MACS2 (74) was used to identify significant peaks relative to control where available. The P value cutoff was set at 0.05, and MACS2 callpeak uses the t test to determine the significance of the enrichment at a certain location over control. The resulting narrowPeak files were used to plot the figures regarding ChIP-seq peaks using the GViz package in Rstudio. The SOLiD 3.0 data set for the NT2/D1 OCT4 ChIP-seq was aligned to the hg19 reference genome using Bowtie 1, following which the downstream processing was performed as indicated above.

Data availability.

Files for the data sets used for analysis can be downloaded from https://www.ebi.ac.uk/ena by providing the following IDs as search terms: H9 cells, OCT4 (SRR5642847), Nanog (SRR5642845); HEK293 cells, CTCF (DRR014670); DLD1 cells, CTCF (DRR014660); female breast epithelium cells (ENCODE), CTCF (SRR6213076, input SRR6213541); male breast epithelium cells (ENCODE), CTCF (SRR6213724, input SRR6214303); MCF7, CTCF (SRR577680, SRR577679) (ENCODE), HeLa, CTCF (SRR227659, SRR227660); NT2/D1, OCT4 (SRR1640282). (This data set is generated on the platform SOLiD3.0.)