Involvement of CTCF in transcription regulation of EGR1 at early G1 phase as an architecture factor

Early growth response 1 (EGR1) is a transcription factor and regulates cellular processes such as proliferation, differentiation, and apoptosis. The expression of EGR1 is rapidly induced in response to several stimuli, and it activates the expression of downstream target genes involved in signaling cascades. EGR1 gene is also known to be transcribed in early G1 phase. However, the regulation of EGR1 transcription in early G1 phase is not clarified well. Here we found that CCCTC-binding factor (CTCF), a chromatin binding protein, is required to transcribe EGR1 gene at the onset of early G1 phase. We found that CTCF mediated the formation of higher-order chromatin structures among CTCF binding sites located in the EGR1 locus. Disruption of the CTCF-dependent higher-order chromatin structure using nuclease-dead Cas9 (dCas9)-mediated interference reduced the EGR1 transcription in early G1 phase. Collectively, we propose that CTCF has functional roles for the temporal expression of EGR1 in early G1 phase through regulation of higher-order chromatin structure organization.

Immediate early genes (IEGs) are a set of cellular genes, and transcription of mRNAs from which is rapidly induced by both extracellular and intracellular signals through several factors which does not requires de novo protein synthesis 1 . Most IEGs encode transcription factors involved in initiation of signaling cascades by modulating transcription of the target genes. EGR1 belongs to IEGs and is transcribed rapidly and transiently in response to various kinds of stimuli 2 . EGR1 functions as both an activator and a repressor for transcriptional regulation of numerous genes, including EGF, TGF-β, and Bax, in concert with several cofactors 3,4 to control cell growth, tumorigenesis 5 , and apoptosis 6 . Analyses of the EGR1 promoter revealed that a number of cis-elements are located within 720 bp upstream of the EGR1 transcription start site (TSS) 7 . It has been proposed that several serum response elements (SREs) located at approximately 300 bp upstream of the EGR1 TSS have a crucial role for the expression of EGR1 7 . Serum response factor (SRF) binds to SREs by interacting with TCF co-regulator proteins in a phosphorylation-dependent manner 8 . It has been reported that the expression of EGR1 was induced in early G1 phase 9 . Because SRF-TCF complex is activated in early G1 phase by growth factors to induce genes involved in G1 progression 10 , EGR1 is thought to be regulated by SRF-TCF complex in early G1 phase. From functional analyses of CTCF in the EGR1 expression, CTCF was thought to function as a negative regulator during mouse myeloid cell differentiation or in LPS-stimulated macrophages 11 . CTCF binding motif is located at approximately 1.2 kb upstream of the EGR1 TSS 11 . CTCF is a DNA binding protein possessing C2H2 zinc finger motifs and was originally found as a repressor of c-myc 12 and lysozyme 13 . CTCF is a highly conserved multifunctional protein in vertebrates. CTCF functions not only as a transcription factor with a variety of binding partners 14 , but also a regulator of higher-order chromatin structure through the formation of chromatin loops that define the boundary between transcriptionally active and inactive chromatin 15 . Recently we reported that CTCF is highly phospholylated in mitosis and the phosphorylation of CTCF impairs its DNA binding activity 16 . In parallel with the reformation of nucleus at telophase, dephosphorylation of CTCF and its recruitment on chromatin might occur, although a phosphatase(s) involved in this process is Unknown. Higher order chromatin structures, including structures possibly regulated by CTCF, are disrupted in mitosis and reformed in G1 phase 17 . Regulation of gene expressions through higher-order chromatin structures are observed through whole Figure 1. CTCF was associated with EGR1 promoter and stimulated its transcription in early G1 phase. (A) Expression level of CTCF in CTCF KD cells. HeLa S3 cells were transfected with shEGFP expression plasmid as a control (shEGFP, lanes 1-3) or shCTCF#1 expression plasmid (shCTCF#1, lanes 4-6), or shCTCF#2 expression plasmid (shCTCF#2, lanes 7-9). Lysates from desired cell numbers (0.25 × 10 5 , 0.5 × 10 5 , and 1 × 10 5 ) were subjected to western blot analyses using anti-CTCF and anti-α-tublin antibodies. The original blots are presented in Supplementary Fig. S9. (B) Effect of CTCF KD on transcription in early G1 phase. HeLa S3 cells transfected with shEGFP (blue line and diamonds) or shCTCF#1 (red line and squares), or shCTCF#2 (green line and triangles) expression plasmid were synchronozied and arrested with nocodazole. Total RNAs were extracted from cells collected at indicated time points after nocodazole release and subjected to qRT-PCR using specific primers for EGR1 pre-mRNA, CCND1 pre-mRNA, and 28 S rRNA. The amounts of pre-mRNA levels were normalized by the amount of 28 S rRNA, and then relative expression levels were shown. (C) Expression level of EGR1 in CTCF KD cells in early G1 phase. HeLa S3 cells were transfected with shEGFP expression plasmid (shEGFP, lanes 1-4) or shCTCF#1 expression plasmid (shCTCF#1, lanes 5-8). Lysates from cells collected at indicated time points were subjected to western blot analyses using anti-EGR1 and anti-α-tublin antibodies. The original blots are presented in Supplementary Fig. S9 cell cycles. In particular, early G1 phase may be a time in which profound reformation of higher-order chromatin structure and transcription restart simultaneously proceed. Thus, CTCF might be involved in the expression of early G1 genes, such as EGR1. However, the precise function of CTCF in the EGR1 expression in early G1 phase is not well known.
Here, we have shown that CTCF is required for the transcription of EGR1 in early G1 phase. Chromatin Immunoprecipitation (ChIP) and Chromosome Conformation Capture (3 C) analyses indicated that CTCF-mediated higher-order chromatin structure is formed among the promoter and the upstream and the downstream CTCF-binding sites of the EGR1 gene locus after mitotic exit. dCas9-mediated interference of the formation of higher-order chromatin structure in early G1 phase also reduced EGR1 transcription. Collectively, these results suggest that CTCF is important for the temporal transcription regulation of EGR1 through its function in the organization of higher-order chromatin structure.

Results
CTCF is required for the expression of the EGR1 gene in early G1 phase. To know whether CTCF is involved in the expression of EGR1 in early G1 phase, we examined the effect of CTCF knockdown (KD) on the EGR1 transcription level in early G1 phase. In CTCF KD cells, using plasmids expressing shRNA against CTCF (shCTCF#1 and #2), the expression level of the CTCF proteins was less than 25% of that in the control cells ( Fig. 1A). At 63 h post transfection of the shRNA expression plasmid, HeLa S3 cells were treated with 165 nM of nocodazole for 6 h, as described in the Experimental procedures. The expression levels of CTCF were not affected by cell cycle synchronization ( Supplementary Fig. S1). After removal of the drug, the cells were incubated at 37 °C to synchronize the cell population in early G1 phase. Total RNAs were isolated from the cells and subjected to qRT-PCR using the primers that span the exon-intron junctions. Along with the progression of G1 phase, the expression level of EGR1 pre-mRNA was peaked at 2 h post release and quickly reduced at 3 h post release in the control cells (Fig. 1B). In contrast, the EGR1 transcription level in CTCF KD cells had decreased to less than 30% of that in the control cells at 2 h post release (Fig. 1B). These results indicate that CTCF is a positive regulator of EGR1 transcription in early G1 phase. We also examined the pre-mRNA level of CCND1 gene which is also expressed in G1 phase and has putative CTCF binding sites 18 . The amount of CCND1 pre-mRNA was reduced in CTCF KD cells compared with that of control cells, suggesting that CTCF also regulates CCND1 transcription in G1 phase. Similar results were obtained from shCTCF#1 and shCTCF#2. The cell cycle progression profiles of the control and CTCF KD cells were not significantly changed ( Supplementary Fig. S2). Notably, the expression of EGR1 protein also reduced in CTCF KD cells in early G1 phase (Fig. 1C). To clarify the role(s) of CTCF in the transcriptional regulation of EGR1, we examined the binding of CTCF to a CTCF binding site located 1.2 kb upstream 11 of the EGR1 TSS during early G1 phase. ChIP assays were performed using lysates prepared from HeLa S3 cells at 0, 1, 2 and 3 h post release from nocodazole treatment. As expected, CTCF interacted with the CTCF binding site in the EGR1 promoter after nocodazole release and its binding was observed during cell cycle progression (Fig. 1D).
CTCF does not affect early G1 transcription from the EGR1 promoter in a reporter plasmid system. It has been reported that CTCF plays a role in transcriptional regulation through interactions with transcription-related factors independently of higher-order chromatin structure formation 19,20 . To reveal the function(s) of CTCF in the EGR1 transcription in early G1 phase, we used a luciferase reporter system to analyze the effect of CTCF bound to the promoter on EGR1 transcription in early G1 phase. A DNA fragment corresponding to the promoter region between −720 bp and + 120 bp of the EGR1 TSS was amplified by PCR and cloned into a luciferase reporter plasmid (pEGR1 pro-720/+120 Luc). We also constructed a luciferase reporter plasmid containing a DNA fragment between −1358 bp and + 120 bp of the EGR1 TSS (pEGR1 pro-1358/+120 Luc) containing a promoter-proximal CTCF binding site located at the nucleotide position between −1227 bp and −1208 bp of the EGR1 TSS. pEGR1 pro-1358/+120ΔCTCF BS Luc has mutated nucleotide sequences of the promoter-proximal CTCF binding site (Figs 2A and S3A). The amounts of CTCF on pEGR1 pro-1358/+120ΔCTCF BS Luc decreased to 25% of that on pEGR1 pro-1358/+120 Luc ( Supplementary Fig. S3B). However, no significant differences in the luciferase activity were found among cells transfected with plasmids pEGR1 pro-720/+120 Luc, pEGR1 pro-1358/+120 Luc, or pEGR1 pro-1358/+120ΔCTCF BS Luc (Fig. 2B).
These results indicate that, in the reporter plasmid-based assay system, the promoter-proximal CTCF binding site does not have any functions to the EGR1 expression. This indicates that CTCF-mediated EGR1 activation requires other CTCF binding sites in the EGR1 locus in addition to the promoter-proximal binding site. Based on ChIP-sequence data from ENCODE project 18 , it is assumed that CTCF may bind to upstream and downstream of the EGR1 gene locus ( Supplementary Fig. S4). To confirm that CTCF binds to these loci, ChIP assays were performed using lysates prepared as described in Fig. 1. Figure 3B shows that CTCF binds to upstream (15 kb from the EGR1 TSS) and downstream (38 kb from the TSS) sites in the EGR1 gene locus in early G1 phase. Both the upstream and downstream CTCF binding sites showed similar increment patterns of CTCF binding as seen for quantification of the level of CTCF associated with the promoter-proximal CTCF binding site in early G1 pahse. Eary G1 cells were collected at indicated time points after release, and ChIP assays were performed using with anti-CTCF antibody and rabbit normal IgG. qPCR was performed with primers for amplification of DNA including CTCF binding sites located the 1.2 kb upstream of the EGR1 gene transcription start site (TSS). The relative amount of DNA co-immunoprecipitated with each antibody was shown as % of input. Data information: Each error bar shows standard deviation of three independent experiments. Student's two tailed t-test, *P < 0.05, ns, P > 0.05.
in the promoter-proximal CTCF binding site in early G1 phase (Fig. 1D). These results led us to examine the long-range chromatin interaction among the upstream, promoter, and downstream CTCF binding sites in the EGR1 gene locus (Fig. 4).

CTCF is involved in the formation of higher-order chromatin organization in the EGR1 locus.
To address whether long-range chromatin interactions exist in the EGR1 locus, we performed 3 C analyses using primer sets indicated in Table 1 to identify a physical interaction between two separate genomic regions 21 . After crosslinking, the genomic DNA was digested with a restriction enzyme and subjected to ligation followed by PCR using primers corresponding to regions between 15 kb upstream and the promoter, 38 kb downstream and the promoter, and 15 kb upstream and 38 kb downstream (Fig. 4A). In control cells, amplified PCR products derived from any combinations among the EGR1 upstream, promoter, and downstream regions were observed in a time-dependent manner after nocodazole release, indicating the close proximity of each these two separate regions forms in early G1 phase (Fig. 4B, lanes 1-4, 9-12, and 17-20). We did not observe any amplified PCR products using primers corresponding to the Igf2 downstream and EGR1 promoter regions ( Supplementary  Fig. S5), suggesting that the observed PCR signals in the EGR1 locus reflect specific interactions between the two separate genomic regions. Igf2 is a gene known to be regulated by CTCF and located on chromosome 11 which is different that of EGR1. Thus, Igf2 and EGR1 might be not close each other in the nucleus. The PCR products were reduced or not observed in CTCF KD cells (Fig. 4B, lanes 5-8, [13][14][15][16][21][22][23][24], indicating that the long-range chromatin interactions observed in the EGR1 locus were dependent on CTCF. From these results, it is possible that CTCF regulates EGR1 transcription in early G1 phase as a chromatin organizer. However, the results observed in the CTCF KD condition were potentially caused by an indirect effect due to the abundant existence of CTCF binding sites in whole genomic DNA. To investigate whether CTCF-mediated long-range chromatin interactions in the EGR1 locus are required for transcription in early G1 phase, we developed a dCas9-mediated interference method to inhibit CTCF function in the EGR1 locus. CRISPR-associated endonuclease Cas9 is recruited through RNA-DNA base pairing between Cas9-guide RNA (gRNA) and a specific target DNA sequence, resulting in the induction of a double-strand DNA break 22 . Because dCas9 has two alanine substitutions at amino acid positions 10 and 840 (aspartic acid and histidine, respectively), dCas9 has gRNA-dependent DNA binding activity but not nuclease activity. Thus, the dCas9-gRNA complex is recruited to a specific chromatin site without DNA damages 23 . To bring about interference of CTCF at each CTCF binding site, we designed gRNA sequences covering each of the CTCF binding sites. Cells were transfected with a plasmid construct expressing both 3× FLAG tagged dCas9 and gRNA. The gRNA targets were either a control nonspecific sequence (non-homologous with any CTCF binding sites in the EGR1 locus), the upstream, downstream, or promoter proximal CTCF binding site in the EGR1 locus ( Supplementary Fig. S6). The cells were synchronized in prometaphase with nocodazole. At 0, 1, 2 and 3 h post release from nocodazole arrest, total RNAs were isolated and subjected to qRT-PCR for quantification of EGR1 pre-mRNA. Compared with the cells expressing dCas9 with nonspecific gRNA, the cells expressing dCas9 with specific gRNA against the upstream, downstream, or promoter CTCF binding site showed reduced EGR1 transcription in early G1 phase at 2 h post release (Fig. 5A). It is noteworthy that the effect of gRNA expression was specifically observed in transcription of EGR1 but not CCND1, suggesting that the effect of gRNAs is highly specific for EGR1 gene. The ChIP assays of dCas9 demonstrated that the gRNAs specifically interacted with each target site in the EGR1 locus (Fig. 5B). To analyze whether dCas9 competes with CTCF for these binding sites, sequential ChIP analyses (termed re-ChIP) were performed. dCas9-bound chromatin was concentrated by immunoprecipitation with anti-FLAG antibody and then were subjected to further immunoprecipitation with control IgG, ani-CTCF, and anti-FLAG antibodies, respectively ( Supplementary Fig. S7). We found that CTCF was not observed on the upstream ( Supplementary  Fig. S7, upper graph, lanes 5), the downstream ( Supplementary Fig. S7, middle graph, lanes 8), and the promoter proximal CTCF binding sites ( Supplementary Fig. S7, lower graph, lanes 11) by expressing gRNAs targeting to these sites, respectively. We also examined the EGR1 transcription in cells transfected with two or three gRNAs targeting to the CTCF binding sites (Fig. 5C). The level of EGR1 pre-mRNA but not CCND1 pre-mRNA was reduced more than that with single gRNA transfected cells (Figs 5C and S8). To examine whether dCas9 expression also disrupts higher-order chromatin interactions at the EGR1 locus, 3 C analyses were performed with cells expressing dCas9 and gRNAs against the CTCF binding sites in the EGR1 locus (Fig. 6). We observed that the interactions of the upstream-downstream regions (Fig. 6C,D) and the promoter-downstream regions (Fig. 6E,F), but not the upstream-promoter regions (Fig. 6A,B), decreased by expressing dCas9-gRNA against CTCF binding sites in EGR1 locus. Further, we also found that all three CTCF binding sites are required for the long-range interactions of the upstream-downstream and promoter-downstream regions. The CTCF binding sites located on upstream and promoter in the EGR1 locus are on the forward strand (the coding strand of EGR1), while the downstream CTCF binding site of the EGR1 locus is on the reverse strand (the template strand of EGR1). Because, it is reported that two CTCF binding sites located in convergent orientation interacts more often than that of located in the same orientation 24 . The 3 C interaction of the upstream-promoter region seems weaker compared with others (Fig. 6B). This might be due to the orientation of CTCF motif on each CTCF binding site in the EGR1 locus. Notably, the interaction between the upstream and downstream regions of Igf2 was not affected by dCas9-gRNAs for the EGR1 locus (Fig. 6B, bottom graph). These results suggest that the dCas9-gRNA complexes interfere with the binding of CTCF to the EGR1 upstream, downstream, or promoter CTCF binding site, and consequently the formation of 3 C interaction was disturbed. Therefore, it is concluded that CTCF mediates the long-range chromatin interactions among the CTCF binding sites in the EGR1 locus. The interactions between upstream-downstream and promoter-downstream might be important for the expression of EGR1 in early G1 phase.

Discussion
EGR1 regulates several cellular processes such as cell proliferation and apoptosis as a transcription factor 2 . It has been reported that EGR1 is transcribed in early G1 phase 9 . In general, transcription factors are the primary determinants of the transcription regulation. However, the molecular mechanism of the cell cycle-specific transcription of EGR1 is not well understood. It has been reported that CTCF binds to the promoter-proximal CTCF binding site of EGR1 and negatively regulates its transcription during myeloid cell differentiation or LPS-stimulated macrophages 11 . Here, we examined whether CTCF is involved in cell cycle-dependent transcription of the EGR1. We found that the expression of EGR1 pre-mRNA in early G1 phase is reduced by knockdown of CTCF, suggesting an involvement of CTCF in the cell-cycle specific transcription of EGR1 as a positive regulator (Fig. 1B). CTCF is known as a multi-functional protein 15 and activates mRNA synthesis through its interaction with transcription-related factors such as the largest subunit of RNA polymerase II, RBP1/POLR2A 19 , or the general transcription factor TFII-I 20 . However, the promoter-proximal CTCF binding site did not activate the transcription from transiently transfected reporter plasmid in early G1 phase (Fig. 2). The inhibition of CTCF by shRNA or dCas9-gRNA resulted in reduced expression of EGR1 pre-mRNA in early G1 phase (Figs 1 and 5) and impairment of chromatin-loop structures (Figs 4 and 6). The expression level of EGR1 pre-mRNA showed sharp increment and decrement during early G1 phase (Figs 1B and 5A,C). During the period from 0 to 2 h post release from nocodazole block, CTCF-mediated higher-order chromatin structures within the EGR1 gene locus might be required for its maximum transcription in early G1 phase. After that, the expression levels of EGR1 pre-mRNA were decreased at 3 h post nocodazole release. However, binding levels of CTCF on each CTCF binding site (Figs 1C and 3B) and the formation of higher order chromatin structures in the EGR1 locus (Fig. 4B) did not show the similar fluctuating patterns to that of the EGR1 pre-mRNA levels. It might suggest that there is the inhibitory mechanism(s) of the EGR1 expression which is controlled independently of CTCF-mediated higher-order chromatin structures. The gRNA against the promoter proximal CTCF binding site reduced the expression of EGR1 in early G1 phase, suggesting that there is a functional significance of the promoter proximal CTCF binding site on the EGR1 expression, even though it can not be reconstituted in transiently transfected plasmid lacking additional CTCF binding sites. These results suggest that higher order interactions between the promoter proximal CTCF binding site and other CTCF binding sites are important for its proper function. It has been reported that nucleosomes surrounding a CTCF binding site are well positioned 25 . Hence, we can not exclude the possibility that the local chromatin structure around the promoter proximal CTCF binding site might be important for its function in the EGR1 locus because nucleosome positioning on transiently transfected DNA is usually different from that of native chromatin 26 . It has been reported that in mouse erythroblast cells, chromatin loop structure is generally disrupted in mitosis and increased sharply until early G1 phase (60-90 min post nocodazole release) 27 . Reformation of chromatin loop structures occur independently of its gene expression patterns during cell cycle, although the functional relationship between chromatin loop structures and its direct contribution to a gene expression is not clear 27 . Our findings suggest that the formation of higher order chromatin structures in the EGR1 locus in early G1 phase may contribute to the expression of the EGR1 pre-mRNA in early G1 phase. Enhancers regulate transcription of their target genes through the binding of transcription factors in a distance-independent manner 28 . It has been suggested that distal enhancers and their target promoters are closely positioned in three dimensional space through higher order chromatin structures to establish a microenvironment to regulate efficient gene expressions 29 . The enhancer database predicts a possibility of the presence of cell type specific putative enhancer elements in the human EGR1 gene locus 30 . These putative enhancer elements may contribute to an optimal induction of EGR1 in early G1 phase mediated by CTCF.
Not only EGR1 but also other early G1 genes 9 have CTCF binding sites in their gene loci (Supplementary  Table S1), which suggests possible contribution of CTCF to expression of those genes via higher-order chromatin structure formation. However, it remains unclear whether our findings of CTCF involvement in the expression of the EGR1 gene in the early G1 phase through higher-order chromatin structure formation could be expanded to other early G1 expression genes. Because of the technological limitation of PCR-based one-by-one analysis of higher-order chromatin interaction in being unable to handle multiple loci, several modified 3C-based methods have been developed to overcome this problem. Among these methods, the Hi-C technique 31 is a practical methods to analyze genomewide chromatin interactions, especially in megabase-scale chromatin domain structures. Recently, it was reported that a modified Hi-C technique, called the in situ Hi-C protocol 24 , can be used to make a kilobase-scale high resolution interaction map. Analyses using such techniques may allow us to obtain more insights into the functional contribution of CTCF to the early G1 phase gene expression through higher-order chromatin structure formation.

Methods
Cell culture, transfection and synchronization. HeLa S3 cells were maintained at 37 °C in DMEM supplemented with 10% FBS. Cells were transfected with pU6-puro-shCTCF or pU6-puro-shEGFP 32 plasmid using NEON transfection system (Thermo Fisher Scientific). After 20 h post transfection, puromycin was added in the medium at the concentration of 2 μg/ml, and cells were incubated further for 20 h. After selection in the presence of puromycin, mitotic cells synchronized with nocodazole were collected. Briefly, cells were treated with 2.5 mM thymidine (Sigma) in the culture medium for 16 h. Thymidine blocked cells were released into a fresh culture medium and incubated for 7 h, so that cells stayed in late S and G2 phases. Then, cells were treated with nocodazole (Sigma) at a final concentration of 165 nM and further incubated for 6 h. Mitotic cells were collected by gentle shaking of cell culture dishes. For release of cells from mitosis, nocodazole-arrested cells were reattached to poly-L-lysine coated culture dishes, then washed with PBS, and cultured for indicated periods.

Preparation of plasmids.
To generate pU6-puro-shCTCF#1 plasmid, which expresses 21 nucleotides-long hairpin-type shRNA against human CTCF, corresponding to the sequence between nucleotide positions 868 and 888 (NCBI Reference Sequence: NM_006565.3), 5′-CACCGTGTCTAAAGAGGGTCTTGTGGTGTGCTGTCCCG CAAGGCCCTCTTTAGACACTTTTTT-3′ and 5′-GCATAAAAAAGTGTCTAAAGAGGGCCTTGCGGGACA GCACACCACAAGACCCTCTTTAGACAC-3′ oligonucleotides were phosphorylated, annealed each other, and cloned into pU6-puro 32 that had been digested with BspM I. Oligonucleotides corresponding to the sequence between nucleotide positions 2,819 and 2,839 (NM_006565.3), 5′-CACCGTAAATTATGGAGTGTTCTGATTG ATATCCGTCAGAACATTCCATGATTTGCTTTTTT-3′ and 5′-GCATAAAAAAGCAAATCATGGAATGTTCT GACGGATATCAATCAGAACACTCCATAATTTAC-3′ were used to generate pU6-puro-shCTCF#2 plasmid. The human EGR1 promoter was amplified by PCR using the genomic DNA extracted from HeLa S3 cells as template    of uncut genome, Dpn II digested genome without ligation, and ligated genome DNA are shown in Fig. S5A. A control template for 3 C assay was prepared from the bacterial artificial chromosome (BAC) clone RP11-113M21 containing EGR1 locus on human chromosome 5, pBS-Igf2-upstream, and pBS-Igf2-downstream. BAC DNA and plasmid DNA are mixed in equimolar amounts and digested with Dpn II. Digested DNA was randomly ligated by T4 DNA ligase. The random ligated DNA was purified and used as a standard template for semi-quantitative PCR. PCR was performed with primers shown in Table 1. Primer pairs for each PCR are shown in Table 2. The position of the restriction enzyme recognition site, CTCF motif, and position of primers in each of the nucleotide sequence is shown in Fig. S6. PCR products were separated in a 6% polyacrylamide gel and visualized by staining with ethidium bromide. Gel images were taken by Gel Documentation System (ATTO). Band intensities were measured by ImageJ 35 .