Vigilin interacts with CCCTC-binding factor (CTCF) and is involved in CTCF-dependent regulation of the imprinted genes Igf2 and H19

Authors

  • Qiuying Liu,

    1. Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China
    Search for more papers by this author
    • These authors contributed equally to this work
  • Bo Yang,

    1. Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China
    Current affiliation:
    1. The Third Affiliated Hospital, Zhengzhou University, Zhengzhou, China
    Search for more papers by this author
    • These authors contributed equally to this work
  • Xiaoyan Xie,

    1. Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China
    Search for more papers by this author
  • Ling Wei,

    1. Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China
    Search for more papers by this author
  • Wenquan Liu,

    1. Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China
    Search for more papers by this author
  • Wenli Yang,

    1. Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China
    Search for more papers by this author
  • Yajun Ge,

    1. Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China
    Search for more papers by this author
  • Qiang Zhu,

    1. Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China
    Search for more papers by this author
  • Junzhe Zhang,

    1. Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China
    Search for more papers by this author
  • Lei Jiang,

    1. Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China
    Search for more papers by this author
  • Xiaoqin Yu,

    1. Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China
    Search for more papers by this author
  • Wenyan Shen,

    1. Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China
    Search for more papers by this author
  • Ran Li,

    1. Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China
    Search for more papers by this author
  • Xuejiao Shi,

    1. Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China
    Search for more papers by this author
  • Bo Li,

    1. Department of General Surgery, West China Hospital, Sichuan University, Chengdu, China
    Search for more papers by this author
  • Yang Qin

    Corresponding author
    1. Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China
    2. Sichuan University ‘985 Project – Science and Technology Innovation Platform for Novel Drug Development’, Chengdu, China
    • Correspondence

      Y. Qin, Department of Biochemistry and Molecular Biology, West China School of Preclinical and Forensic Medicine, Sichuan University ‘985 Project – Science and Technology Innovation Platform for Novel Drug Development’, Sichuan University, Chengdu 610041, China

      Fax: +86 028 85503204

      Tel: +86 028 85501283

      E-mail: qin_1@sina.com

    Search for more papers by this author

Abstract

CCCTC-binding factor (CTCF), a highly conserved zinc finger protein, is a master organizer of genome spatial organization and has multiple functions in gene regulation. Mounting evidence indicates that CTCF regulates the imprinted genes Igf2 and H19 by organizing chromatin at the Igf2/H19 locus, although the mechanism by which CTCF carries out this function is not fully understood. By yeast two-hybrid screening, we identified vigilin, a multi-KH-domain protein, as a new partner of CTCF. Subsequent coimmunoprecipitation and glutathione S-transferase pulldown experiments confirmed that vigilin interacts with CTCF. Moreover, vigilin is present at several known CTCF target sites, such as the promoter regions of c-myc and BRCA1, the locus control region of β-globin, and several regions within the Igf2/H19 locus. In vivo depletion of vigilin did not affect CTCF binding; however, knockdown of CTCF reduced vigilin binding to the H19 imprinting control region. Furthermore, ectopic expression of vigilin significantly downregulated Igf2 and upregulated H19, whereas depletion of vigilin upregulated Igf2 and downregulated H19, in HepG2, CNE1 and HeLa cells. These results reveal the functional relevance of vigilin and CTCF, and show that the CTCF–vigilin complex contributes to regulation of Igf2/H19.

Structured digital abstract

Abbreviations
CCD

centrally conserved DNase I-hypersensitive domain

ChIP

chromatin immunoprecipitation

CTCF

CCCTC-binding factor

DMR

differentially methylated region

GFP

green fluorescent protein

GST

glutathione S-transferase

HP1

heterochromatin protein 1

ICR

imprinting control region

IGF2

insulin-like growth factor 2

IP

immunoprecipitation

qPCR

quantitative PCR

RFLP

restriction fragment length polymorphism

shCTCF

CCCTC-binding factor small hairpin RNA

shGFP

green fluorescent protein small hairpin RNA

shRNA

small hairpin RNA

shvigilin

vigilin small hairpin RNA

SRA

steroid receptor RNA activator

Introduction

CCCTC-binding factor (CTCF) is a highly conserved 11-zinc-finger protein that is ubiquitously expressed in higher eukaryotes. CTCF is implicated in diverse roles in gene regulation via binding to DNA, and mediates protein–protein interactions that are also linked to human development and diseases, including cancer [1, 2]. Several recent genome-wide analyses have identified substantial numbers of CTCF target sites in the human and mouse genomes [3, 4]. CTCF was reported to bind to the β-globin locus [5], the Igf2/H19 imprinting locus [6] and the DM locus [7] for insulator function, and to promoter regions of c-myc [8] and BRCA1 [9] for transcriptional repression. Moreover, numerous potential CTCF-binding partners have been reported [10-18]. Although the association of CTCF with YB1 [10], Kaiso [11] and Sin3A [12] may confer transcription regulatory activity, interaction of CTCF with the SNF2-like chromodomain helicase protein CHD8 [13] and/or the DEAD-box RNA helicase p68 (DDX5) with its associated noncoding RNA, steroid receptor RNA activator (SRA) [14], is required for insulator function. A multiprotein complex, cohesin, which is well known for its ability to hold sister chromatids together during mitosis and DNA repair, has been shown to interact with CTCF at many genomic locations to mediate extensive long-range intrachromatin or interchromatin interactions [15-18].

The insulin-like growth factor (IGF2) gene and H19 are reciprocally imprinted genes in allele-specific expression, leading to paternal expression of Igf2 and maternal expression of H19 in most tissues [19]. Igf2 encodes a fetal growth factor, whereas H19 encodes a long noncoding RNA that is a precursor of miR-675. Parental allelic expression is controlled by an imprinting control region (ICR) [or differentially methylated region (DMR)] that lies upstream of the H19 promoter. CTCF plays a vital role in the current model of the mechanism underlying this reciprocal imprinting. It has been demonstrated that CTCF binding to the unmethylated maternal ICR prevents the distal enhancers from activating the Igf2 promoter, resulting in silencing of the maternal Igf2 allele. In contrast, CTCF is prevented from binding to the methylated paternal ICR, so the ICR loses insulator activity, leading to activation of Igf2 from the paternal allele by interaction of Igf2 promoters with enhancers. Both p68/SRA and cohesin were detected at the H19 ICR as well as at other genomic CTCF-binding sites [14, 16, 18]. In vivo depletion of CTCF, p68/SRA or cohesin reduced insulator activity at the H19 ICR, and increased levels of Igf2 expression while decreasing levels of H19 transcription [14, 16]. Recent studies have shown that, cohesin, together with CTCF, binds at several CTCF-bnding sites, such as the ICR, CTCF-binding sites immediately adjacent to the Igf2 DMR0 (AD1 and AD2), the CTCF-binding site in the centrally conserved DNase I-hypersensitive domain (CCD) region and a CTCF-binding site downstream of the H19 enhancers (DS) to organize the allele-specific higher-order chromatin conformation at the Igf2/H19 locus [15, 20]. However, it remains largely unknown what other factors may be involved in regulating gene expression together with CTCF.

We previously showed that CTCF played an important role in imprinting regulation of Igf2 [21], and that there were three methylation profiles for the H19 ICR in human hepatocellular carcinoma. Both the hypomethylated and hypermethylated profiles were associated with aberrant imprinting of Igf2 and H19. We speculated that CTCF has an equally complex set of parameters in regulating genomic imprinting in human cancer. In this study, we have identified and validated a novel CTCF partner, vigilin, that could bind to several known CTCF-binding sites and cooperate with CTCF in regulating expression of Igf2 and H19.

Results

Vigilin interacts with CTCF

To identify CTCF-associated proteins, a yeast two-hybrid screen was performed. The ORF of CTCF was constructed as a translational fusion of a DNA-binding domain, and used as the bait for screening of a human liver yeast two-hybrid cDNA library. Six genes were identified as candidate CTCF interaction partners in this screen. They were the genes encoding vigilin (high-density lipoprotein-binding protein), chromodomain helicase DNA-binding protein 1-like, DEAD-box RNA helicase p72, also known as DDX17, Ras-related protein A, NADH dehydrogenase subunit 4, and ribosomal protein L31. To confirm the protein interaction, we repeated the yeast two-hybrid assay by exchanging the candidate gene fragment inserts from pGADT7 to pGBKT7 vector, likewise, CTCF from pGBKT7 to pGADT7 vector. CTCF showed strong positive hybridization to vigilin, p72 and NADH dehydrogenase subunit 4, but weak hybridization to chromodomain helicase DNA-binding protein 1-like, Ras-related protein A and ribosomal protein L31 (data not shown). Interestingly, the DEAD-box RNA helicase p72 and the partially homologous protein p68, together with the noncoding RNA, SRA, were found in the CTCF–cohesin complex that was required for the regulation of Igf2/H19 [14].

To verify interaction between CTCF and vigilin in vivo, we performed immunoprecipitation (IP) experiments with HepG2 and CNE1 cells. Western blot analysis with antibodies against CTCF and vigilin identified endogenous CTCF at ~ 130 kDa (Fig. 1A) and vigilin at ~ 150 kDa (Fig. 1B) in all cultured mammalian cells. Immunoprecipitates obtained with antibody against CTCF were probed by western blotting with antibodies against CTCF and vigilin. As shown in Fig. 1C, both CTCF and vigilin were observed in the immunoprecipitates of CTCF, but not in those of control IgG. Similarly, both vigilin and CTCF were present in the immunoprecipitates of antibody against vigilin but not in those of control IgG (Fig. 1D). To confirm a direct interaction between vigilin and CTCF, we performed glutathione S-transferase (GST) pulldown assays with GST–CTCF fusion proteins expressed and purified from Escherichia coli BL21 cells. GST and GST–CTCF proteins were immobilized on glutathione–agarose beads and incubated with HeLa cell nuclear extracts. As shown in Fig. 1E, full-length GST–CTCF but not GST could pull down endogenous vigilin in HeLa nuclear extracts. These data suggest that vigilin interacts with CTCF. Thus, vigilin is a new protein partner of CTCF, a finding that has not been reported before.

Figure 1.

Vigilin interacts with CTCF. (A, B) Protein expression of CTCF and vigilin, respectively, in cultured cells. Nuclear lysates from cultured cells were analyzed by SDS/PAGE and western blotting with antibodies against CTCF (A) or vigilin (B). β-Actin was used as an endogenous control. (C) Lysates prepared from HepG2 (left) and CNE1 (right) cells were immunoprecipitated with antibody against CTCF or mouse IgG, and the precipitates were analyzed by western blotting with antibodies against CTCF and vigilin. (D) Cell lysates from HepG2 (left) and CNE1 (right) cells were subjected to IP with antibody against vigilin or rabbit IgG, and this was followed by western blot analysis with antibodies against vigilin or CTCF. (E) Western blot of input, GST-associated and GST–CTCF-associated proteins isolated from HeLa nuclear extracts probed with antibodies against vigilin.

Vigilin colocalizes with CTCF at known CTCF-binding sites in vivo

CTCF is able to bind to different targets in the regulatory regions of many genes, and tens of thousands of potential CTCF target sites have been identified in the human and mouse genomes by the use of high-throughput chromatin immunoprecipitation (ChIP)–chip analyses [3, 4]. The H19 ICR the promoter regions of c-myc and BRCA1 and the 5′-DNase-hypersensitive site 5 of the β-globin locus control region have been previously shown to contain CTCF-binding sites [13]. To investigate whether vigilin coexists with CTCF at these known CTCF-binding sites, HepG2 cells were harvested for ChIP assays with antibodies against either CTCF or vigilin. Purified DNA samples were amplified by PCR with specific primers based on several known CTCF-binding sites: the H19 ICR the c-myc and BRCA1 promoter regions, and the 5′-DNase-hypersensitive site 5 of the β-globin locus control region [13]. Figure 2A clearly shows that CTCF did bind at the H19 ICR the promoter regions of c-myc and BRCA1, and the β-globin locus control region, but not at the negative control site, the exon 4/5 region of the H19 gene. Unexpectedly, vigilin bound to these CTCF target sites in HepG2 cells. However, there was still a small amount of residual binding of vigilin to the exon 4/5 region of H19. These results indicated that vigilin and CTCF can bind to the same CTCF-binding sites.

Figure 2.

Vigilin exists at known CTCF-binding sites in chromatin. (A) ChIP was performed with antibodies against CTCF, vigilin, mouse IgG, rabbit IgG, or just beads, shown at the top of the panel, and PCR amplification was performed with primers for the H19 ICR BRCA1, the β-globin locus, c-myc, and exon 4/5 of H19, shown on the left of the panel. Genomic DNA in the input HepG2 cell lysates was used as a positive control. (B, C) CTCF-ChIP DNA samples (B) and vigilin-ChIP DNA samples (C) obtained from HepG2 and CNE1 cells were analyzed by qPCR with primers specific for different regions of the Igf2/H19 locus. β-Actin was used as the negative control site. Results are presented as means ± standard errors of the mean; n = 3.

The imprinted Igf2/H19 locus plays a causative role in several embryonic growth disorders and various cancers, and the transcriptional insulator protein CTCF has been shown to regulate Igf2 and H19 by binding to several CTCF-binding sites within the Igf2/H19 locus [20]. To investigate whether vigilin coexists with CTCF at the CTCF-binding sites within this locus in HepG2 and CNE1 cells, we performed ChIP assays with antibodies against CTCF or vigilin, and followed these with quantitative PCR (qPCR) with primers within the two CTCF-binding sites adjacent to DMR0 (CTCF AD1 and CTCF AD2), the CCD, and the ICR. As shown in Fig. 2B, ChIP with CTCF revealed several CTCF-binding sites within the Igf2/H19 locus that were significantly enriched relative to IgG controls; these results were consistent with those of recent studies [20]. Remarkably, ChIP with vigilin was also highly enriched at AD1, AD2, the CCD, and the ICR (Fig. 2C). Neither CTCF nor vigilin were bound to the negative control site β-actin. The species-matched IgG was used for IP as a negative control.

Vigilin binding at the H19 ICR is CTCF-dependent

To confirm that vigilin recruitment to the H19 ICR is dependent on CTCF, ChIP analysis of the H19 ICR was performed with antibodies against CTCF and vigilin following knockdown of CTCF or vigilin with small hairpin RNAs (shRNAs) [vigilin shRNA (shvigilin); CTCF shRNA (shCTCF)]. We found that CTCF knockdown in HepG2 and CNE1 cells had no effect on the protein expression levels of vigilin (Fig. 3A). As expected, the recruitment of CTCF to the H19 ICR was reduced by CTCF knockdown in HepG2 and CNE1 cells as compared with that in control cells with green fluorescent protein (GFP) shRNA (shGFP) (P < 0.01) (Fig. 3B). Interestingly, the recruitment of vigilin to the H19 ICR was also significantly reduced by CTCF knockdown in HepG2 cells (P < 0.05) and CNE1 cells (P < 0.05) (Fig. 3C). Knockdown of vigilin did not affect CTCF protein expression (Fig. 3D). Vigilin knockdown decreased the binding of vigilin to the H19 ICR (P < 0.01) (Fig. 3E), but did not significantly affect the binding of CTCF to the H19 ICR (P > 0.05) (Fig. 3F). These results show that vigilin recruitment to the H19 ICR may require CTCF, but that CTCF binding to the H19 ICR does not depend on vigilin.

Figure 3.

Binding of vigilin to the H19 ICR is CTCF-dependent. (A) Western blot analysis of CTCF and vigilin levels in nuclear extracts from shCTCF-transfected or shGFP-transfected HepG2 and CNE1 cells. β-Actin is shown as a control. (B, C) ChIP qPCR analysis of crosslinked DNA samples immunoprecipitated with antibody against CTCF (B) or antibody against vigilin (C) at the H19 ICR in CTCF-depleted HepG2 and CNE1 cells. (D) Western blot analysis of CTCF and vigilin in shvigilin-transfected HepG2 and CNE1 cells. (E, F) ChIP qPCR analysis of isolated DNA associated with vigilin (E) and CTCF (F) at the H19 ICR in vigilin-depleted HepG2 and CNE1 cells. Data are presented as means ± standard errors of the mean; n = 3. *P < 0.05 and **P < 0.01 as compared with control, Student's t-test.

Vigilin regulates expression of the imprinted genes Igf2 and H19

Previous studies from our laboratory indicated that CTCF plays an important role in regulation of the imprinted genes Igf2 and H19 [21]. Together with the finding that vigilin and CTCF bound to the CTCF-binding sites (AD1, AD2, CCD, and ICR) within the Igf2/H19 locus in vivo, this prompted us to examine the effect of vigilin ectopic expression on Igf2 and H19 transcript levels. HeLa cells have been used to examine imprinting of the Igf2/H19 locus following knockdown of CTCF, cohesin, p68, or SRA [14, 16]. HepG2 cells were reported to show monoallelic Igf2 expression with P3 transcription, whereas the P3 promoter is highly active in fetal tissues but is downregulated shortly after birth in normal tissues [22]. Also, CNE1 cells were identified as being heterozygous for Igf2 by screening of several cell lines (Fig. 4). To this end, HepG2, CNE1 and HeLa cells were transiently transfected with vigilin or CTCF construct, and RT-qPCR was performed to detect the transcript levels of Igf2 and H19. We determined the transfection efficiencies at the protein level of both CTCF and vigilin (Fig. 5A,C,E). The results of both western blotting and RT-qPCR analysis (Fig. 5B,D,F) showed that CTCF ectopic expression had no obvious effect on the expression levels of vigilin, and vice versa. Igf2 transcript levels were decreased by approximately 0.75-fold to 0.8-fold and 0.7-fold to 0.75-fold in CTCF-overexpressing and vigilin-overexpressing cells, respectively, whereas H19 transcript levels were increased by approximately 2.5-fold to three-fold in CTCF-overexpressing cells and by approximately 2.5-fold to 2.8-fold in vigilin-overexpressing cells (Fig. 5B,D,F).

Figure 4.

Genotypes of the Igf2 polymorphism and H19 polymorphism in cultured cells. PCR of Igf2 digested with ApaI yielded both a 236-bp fragment and a 173-bp fragment in the heterozygous cases. PCR of H19 digested with AluI yielded a 247-bp fragment or a 142-bp and a 105-bp fragment in the homozygous cases. (A) gDNA samples from HepG2, SMMC7721, CNE1, HeLa, MCF-7, ZR-75-30 and MDA-MB-231 cells were amplified for Igf2, and products were digested with ApaI. Heterozygosity for Igf2/ApaI RFLP was observed in CNE1, MCF-7, ZR-75-30 and MDA-MB-231 cells, and homozygoity for Igf2/ApaI RFLP was obserbed in the other three cultured cell lines. (B) gDNA samples were amplified for H19, and products were digested with AluI. All cells showed monoallelic expression of H19.

Figure 5.

Expression of Igf2 and H19 transcripts in CTCF-overexpressing or vigilin-overexpressing cells. (A, C, E) Western blot analysis of CTCF and vigilin levels in nuclear extracts from HepG2 (A), CNE1 (C) or HeLa (E) cells transfected with the CTCF overexpression construct, the vigilin overexpression construct, or control empty vectors. β-Actin is shown as a control. (B, D, F) HepG2 (B), CNE1 (D) and HeLa (F) cells were transfected with the CTCF overexpression construct, the vigilin overexpression construct, or control empty vectors. After transfection for 48 h, cells were harvested for RNA collection and subjected to RT-qPCR with primers based on CTCF, vigilin, Igf2, and H19. Transcript levels were normalized to β-actin and then the control sample. Data are presented as means ± standard errors of the mean; n = 3. **P < 0.01 as compared with control, Student's t-test.

We also knocked down the endogenous expression of both vigilin and CTCF (as a positive control) by using either shvigilin or shCTCF, and examined the expression of Igf2 and H19 in these three cell lines. We confirmed that both protein and mRNA levels of CTCF and vigilin were decreased in cells transfected with shCTCF and shvigilin (Fig. 6). As shown in Fig. 6, vigilin depletion in these three cell lines had no obvious effect on the expression levels of CTCF, and vice versa. As expected, depletion of CTCF resulted in an increase in Igf2 transcript levels of approximately 1.5-fold to two-fold, and significantly decreased H19 expression by approximately 0.35-fold to 0.5-fold as compared with that in control cells (Fig. 6B,D,F), in agreement with previous findings [14, 16]. Remarkably, shRNA-mediated knockdown of vigilin increased Igf2 mRNA levels by approximately 1.5-fold to 1.7-fold and decreased H19 mRNA levels by approximately 0.4-fold to 0.5-fold as compared with that in control cells (shGFP) (Fig. 6B,D,F).

Figure 6.

Expression of Igf2 and H19 transcripts in CTCF-depleted or vigilin-depleted cells. (A, C, E) Western blot analysis of CTCF and vigilin levels in nuclear extracts from HepG2 (A), CNE1 (C) or HeLa (E) cells transfected with shCTCF, shvigilin or shGFP. β-Actin is shown as a control. (B, D, F) RT-qPCR analysis of CTCF, vigilin, Igf2 and H19 expression in shCTCF-transfected and shvigilin-transfected HepG2 (B), CNE1 (D) and HeLa (F) cells. Transcript levels were normalized to β-actin and then the control shGFP sample. Data are presented as means ± standard errors of the mean; n = 3. *P < 0.05 and **P < 0.01 as compared with control, Student's t-test.

Discussion

CTCF is involved in diverse roles in gene regulation, including enhancer blocking, X chromosome inactivation, and genomic imprinting through sharing chromatin-binding sites with various partners in different contexts [1, 2]. CHD8 has been shown to be present at several known CTCF-binding sites, such as the H19 ICR, the promoter regions of BRCA1 and c-myc, and the β-globin locus control region [13]. RNA interference-mediated knockdown of CHD8 abolishes the insulator activity of the H19 ICR leading to reactivation of Igf2 from chromosomes of maternal origin [13]. Studies have recently shown that cohesin colocalizes at CTCF-binding sites and is required for higher-order chromatin at the imprinted Igf2/H19 locus [15, 20]. In this study, we utilized a yeast two-hybrid screening method with full-length CTCF cDNA as bait to screen the human liver cDNA library; this method resulted in the identification of several interacting factors, including vigilin and other proteins. The coimmunoprecipitation and GST pulldown experiments confirmed that vigilin did indeed interact with CTCF. Vigilin bound to several known CTCF-binding sites, and binding of vigilin to the H19 ICR was CTCF-dependent. Our results indicated that vigilin, a multi-KH-domain protein, interacts with CTCF at several CTCF-binding sites, and participates in CTCF-dependent regulation of the imprinted genes Igf2 and H19 in vivo.

Vigilin, also known as DDP1 in Drosophila and Scp160p in Saccharomyces cerevisiae, is a ubiquitous and highly conserved protein in all eukaryotic organisms. A most striking feature of vigilin is the presence of 15 tandem hnRNP KH domains, which are involved in nucleic acid binding and protein–protein interactions. Vigilin was shown to act in the cytoplasm in regulating RNA metabolism, and to act in the nucleus to maintain heterochromatin structure and chromosome segregation [23-30]. Wang et al. [30] found that, in cell extracts, vigilin bound to promiscuously adenosine-to-inosine (A-to-I) edited RNAs. Furthermore, nuclear vigilin was found in complexes containing not only the editing enzyme ADAR1 but also Ku70/86 autoantigen, DNA-dependent protein kinase catalytic subunits, and RNA helicase A. The vigilin nuclear complex was probably involved in heterochromatin formation through an RNA-mediated pathway.

DDP1 had previously been shown to colocalize with heterochromatin protein 1 (HP1) at pericentric heterochromatin and euchromatic sites [27]. Depletion of DDP1 in Drosophila leads to apparent mislocalization of HP1 and defects in chromosome segregation. It has been reported that CTCF colocalizes with HP1 in the nuclear compartment [3]. Another vigilin partner, Ku70, was shown to coprecipitate with CTCF and bind to two sites within the H19 ICR, although these sites did not appear to play an important role in the establishment or maintenance of imprinting at the Igf2/H19 locus [31, 32]. Taken together with our data showing interaction between vigilin and CTCF, all of these findings support the idea that there is a possible link between CTCF and heterochromatin [3, 4].

It has been reported that CTCF-binding sites are abundant in both human and mouse genomes, spreading over euchromatic and heterochromatic domains [3, 4]. These CTCF-binding sites have been found in various types of repeat sequences, such as long terminal repeats, LINEs, SINEs, microsatellite repeats, and CpG islands [3, 4]. In higher eukaryotes, repetitive DNA elements play an important role in heterochromatin formation. As an important factor in the formation and maintenance of heterochromtin, DDP1 was originally isolated as an ssDNA-binding protein that interacts specifically with the unstructured dodeca-satellite C-strand of Drosophila pericentric heterochromatin [27]. In addition, in human HEK293T cells, vigilin was reported to interact with the highly repeated DNA sequence comprising heterochromatic α and β satellites [30]. Thus, vigilin binding to these CTCF-binding sites may occur for two reasons: one is that these sites are enriched in repetitive DNAs; and the other is that vigilin recognizes different sequences via different binding modes, probably because of interactions of different sets of protein factors, as we also found that vigilin binding to the H19 ICR was CTCF-dependent, as discussed below. For the phenomenon whereby vigilin shows residual binding to the exon 4/5 region of the H19 gene to which CTCF does not bind, the possible explanation is that different functions of vigilin and CTCF enable vigilin to bind to the DNA sequences outside of the CTCF-binding sites. Our finding here that vigilin colocalized with CTCF at these CTCF-binding sites also supports the notion of a link between CTCF and heterochromatin [3, 4].

Previous studies have shown that cohesin and p68 become delocalized in cells depleted of CTCF, whereas depletion of cohesin or p68 does not affect CTCF binding at most sites [14, 18], similarly to our finding that loss of CTCF significantly decreases vigilin binding to the H19 ICR; however, depletion of vigilin leads to loss of CTCF binding but not significantly. A possible explanation for this phenomenon is that alteration of vigilin has a marginal effect on CTCF expression.

CTCF regulates many genes, and the results reported here show that vigilin colocalizes with CTCF at the H19 ICR the promoter regions of c-myc and BRCA1, and the 5′-DNase-hypersensitive site 5 of the β-globin locus control region; it will be of great interest to explore whether CTCF-mediated recruitment of vigilin is necessary mechanism in regulating these targeted genes. CTCF is critical for regulating expression of the imprinted genes Igf2 and H19 through binding at the H19 ICR [1, 2]. In our preliminary experiment, we found that mRNA expression levels of both vigilin and H19 are cell cycle-dependent, and knockdown of vigilin resulted in an increase in Igf2 expression and a decrease in H19 expression in HepG2 cells [33]. We report here that the vigilin–CTCF complex is indeed involved in the regulation of Igf2 and H19, as demonstrated by transient expression experiments with their overexpression constructs or specific shRNAs in three different cell lines. Overexpression of vigilin results in an increase in H19 expression and a decrease in Igf2 expression, and depletion of vigilin increases Igf2 expression and decreases H19 expression. Similar effects are observed in cells upon overexpression or depletion of CTCF, and these results are consistent with other reports [14, 16, 20]. Thus, we suggest that vigilin, synergistically with CTCF, regulates Igf2 and H19 expression.

Collectively, our results demonstrate that a new partner of CTCF, vigilin, coexists with CTCF at several CTCF-binding sites, and participates in CTCF-dependent regulation of the imprinted genes Igf2 and H19. More work will be needed, however, to determine the molecular details of how vigilin works with CTCF. In future work, we will seek to identify the functional regions of both vigilin and CTCF that are responsible for their interaction. Given the known properties of vigilin in heterochromatin formation and RNA processing, and of CTCF in the organization of higher-order chromatin architecture and gene regulation, how vigilin interacts with CTCF, how vigilin and CTCF organize chromatin structure to further facilitate the dynamic balance between heterochromatin and euchromatin and how they regulate gene expression will also be of considerable interest.

Experimental procedures

Cell lines and antibodies

HepG2, SMMC-7721, MCF-7, ZR-75-30, MDA-MB-231 and HeLa cells were cultured in DMEM, and CNE1 cells were grown in RPMI-1640 medium (Gibco). All cell lines were routinely cultured in medium supplemented with 10% newborn calf serum at 37 °C in a humidified incubator with 5% CO2. The primary antibodies used were as follows: mouse mAb against CTCF (ab37477; Abcam, Cambridge, MA, USA), mouse (G3A1) mAb IgG1 isotype control (Cell Signaling Technology, Beverly, MA, USA), antibody against high-density lipoprotein-binding protein/vigilin (PN051PW; MBL, Nagoya, Japan), mouse mAb against β-actin (TA-09; ZSGB-BIO, Beijing, China), and normal rabbit IgG (prepared and purified in our laboratory).

Yeast two-hybrid screening

The two-hybrid bait plasmid pGBKT7–CTCF was constructed by cloning full-length human CTCF cDNA (provided by W.W. Quitschke, State University of New York) between the SalI and BamHI sites of vector pGBKT7. The yeast strain AH109, containing four reporter genes (GAL1UAS-GAL1TATA-HIS3, GAL2UAS-GAL2TATA-ADE2, MEL1UAS-MEL1TATA-lacZ, and MEL1UAS-MEL1TATA-MEL1), was transformed with the recombinant bait plasmid pGBKT7–CTCF mated to BD MatchmakerhLiver in yeast (The Pretransformed Library) according to the Matchmaker Two-Hybrid System 3 protocol (Clontech). Transformants were selected on medium lacking tryptophan, leucine, histidine, and adenine. Subsequently, β-galactosidase activity was tested with the filter lift assay to identify the positive clones. Positive clones were chosen, PCR-amplified with primers complementary to the sequence of the pGADT vector, digested with HaeIII, and subjected to sequence analysis. Sequences were compared with the latest release of the GenBank database by the use of blast online (http://www.ncbi.nlm.nih.gov/blast).

DNA extraction and PCR–restriction fragment length polymorphism (RFLP) analysis

The ApaI polymorphic site in exon 9 of Igf2 and the AluI polymorphic site in exon 5 of H19 were used to evaluate the genotypes for Igf2 and H19, separately. Genomic DNA of HepG2, SMMC7721, CNE1, HeLa, MCF-7, ZR-75-30 and MDA-MB-231 cells was analyzed for the above single-nucleotide polymorphisms by PCR and RFLP to identify heterozygous samples of Igf2 and H19. PCR amplification was performed with 30 cycles under the following conditions: 30 s at 94 °C, 30 s at the optimal annealing temperature (55 °C for Igf2, and 58 °C for H19), and 30 s at 72 °C, with specific primers as follows: for Igf2, 5′-CTTGGACTTTGAGTCAAATTGG-3′ and 5′-CCTCCTTTGGTCTTACTGGG-3′; and for H19, 5′-TCAGGAATCGGCTCTGGAGG-3′ and 5′-ATGATGTGGTGGCTGGTGGT-3′. The PCR products of Igf2 were digested with ApaI and those of H19 were digested with AluI, and they were then electrophoresed for analysis of genotypes of Igf2 and H19 separately.

Plasmid construction

The full-length CTCF insert was moved from the pGEM-7Zf(–)/CTCF to the pCS-CG vector (Addgene, Cambridge, MA, USA) to construct pCS-CG/CTCF (CTCF). The full-length vigilin cDNA in the pCS-CG/vigilin vector was constructed in our laboratory and confirmed by sequencing. The empty vector was used as a control.

For RNA interference experiments, two pairs of shRNA sequences were designed and blast searched against the human genome to ensure that they were specific for CTCF or vigilin, respectively: CTCF1 forward oligonucleotide, 5′-GATCCGTCACCCTCCTGAGGAATCATTCAAGAGTGATTCCTCAGGAGGGTGACTTTTTG-3′; CTCF1 reverse oligonucleotide, 5′-AATTCAAAAAGTCACCCTCCTGAGGAATCACTCTTGAATGATTCCTCAGGAGGGTGACG-3′; CTCF2 forward oligonucleotide, 5′-GATCCGGTGTCTAAAGAGGGCCTTGTTCAAGAGCAAGGCCCTCTTTAGACACTTTTTG-3′; CTCF2 reverse oligonucleotide, 5′-AATTCAAAAAGGTGTCTAAAGAGGGCCTTGCTCTTGAACAAGGCCCTCTTTAGACACCG-3′; vigilin1 forward oligonucleotide, 5′-GATCCGTCCCAACACAAGTATGTCATTCAAGAGTGACATACTTGTGTTGGGACTTTTTG-3′; vigilin1 reverse oligonucleotide, 5′-AATTCAAAAAGTCCCAACACAAGTATGTCACTCTT GAATGACATACTTGTGTTGGGACG-3′; vigilin2 forward oligonucleotide, 5′-GATCCGGCTCGGAAGGACATTGTTGTTCAAGAGCAACAATGTCCTTCCGAGCCTTTTTG-3′; and vigilin2 reverse oligonucleotide, 5′-AATT CAAAAAGGCTCGGAAGGACATTGTTGCTCTTGAA CAACAATGTCCTTCCGAGCCG-3′. As a control for nonspecific effects, GFP sequences were as follows: GFP forward oligonucleotide, 5′-GATCCGCGCAAGCTGACCCTGAAGTTTCAAGAGACTTCAGGGTCAGCTTGCGTTTTTG-3′; and GFP reverse oligonucleotide, 5′-AATT CAAAAAGCGCAAGCTGACCCTGAAGTCTCTTGAAACTTCAGGGTCAGCTTGCGCG-3′. These oligonucleotides were synthesized by Invitrogen. The specific shRNA-coding DNA duplexes were inserted into the BamHI and EcoRI double-digested pSIREN vector (BD Bioscience Clontech, USA) to generate shCTCF, shvigilin, and shGFP.

Transient transfections

Cells were seeded in six-well dishes for 24 h, and transfected with construct plasmids for 48 h by the use of Lipofectamine 2000 (Invitrogen), according to the manufacturer's protocol. Transfection tests were performed in triplicate.

RNA extraction and RT-qPCR

Total RNA was extracted from cells by the use of Trizol (Invitrogen), according to the manufacturer's protocol. RNA preparations were treated with RNase-free DNase I (Fermentas) for 30 min at 37 °C to remove any contaminating DNA. Reverse transcription was performed with Moloney murine leukemia virus reverse transcriptase (TaKaRa), according to the manufacturer's instructions. Amplification of cDNA was monitored with SYBR Premix Ex Taq (TaKaRa) on the My iQ Real-Time PCR Detection System (Bio-Rad). RT-qPCRs were performed in triplicate for each sample. Transcript levels were normalized to β-actin levels. The specific primers used in this experiment were as follows: CTCF forward primer, 5′-TTACACGTGTCCACGGCGTTC-3′; CTCF reverse primer, 5′-GCTTGTATGTGTCCCTGCTGGCA-3′; vigilin forward primer, 5′-CCCCAAATACCATCCCAAGA-3′; vigilin reverse primer, 5′-TGCTCAAGTTCACCCACAAT-3′; Igf2 forward primer, 5′-CTTGGACTTTGAGTCAAATTGG-3′; Igf2 reverse primer, 5′-CCTCCTTTGGTCTTACTGGG-3′; H19 forward primer, 5′-TCAGGAATCGGCTCTGGAGG-3′; H19 reverse primer, 5′-ATGATGTGGTGGCTGGTGGT-3′; β-actin forward primer, 5′-TCATCACCATTGGCAATGAG-3′; and β-actin reverse primer, 5′-CACTGTGTTGGCGTACAGGT-3′.

Western blot analyses

Total protein from cells was extracted with RIPA buffer (1 mm MgCl2, 10 mm Tris/HCl, pH 7.4, 1% Triton X-100, 0.1% SDS, 1% NP-40) (Beyotime, Jiangsu, China) with a complete protease inhibitor cocktail tablet (Roche, Mannheim, Germany) for 5 min on ice. Proteins were separated by SDS/PAGE, and western blotted with antibodies against CTCF, vigilin, and β-actin. Protein bands were detected with an enhanced chemiluminescence western blotting detection system (Beyotime) and imaging system (ChemiDoc XRS; Bio-Rad Laboratories).

Immunoprecipitation

Cells (1 × 107) were washed with ice-cold NaCl/Pi, and lysed with RIPA buffer containing protease inhibitors (Roche) on ice for 10 min. The lysates were collected and centrifuged for 5 min at 26 g. The supernatant was incubated with 3 μg of the indicated antibodies plus Protein A+G agarose beads (Beyotime, Jiangsu, China) with rotation overnight at 4 °C. Beads were washed three times with cell lysis buffer. The bound protein was then eluted in 20 μL of loading buffer and subjected to western blot analysis.

GST pulldown assay

GST and full-length GST–CTCF were expressed in Ecoli strain BL21 (Invitrogen). GST fusion proteins were purified with glutathione–Sepharose 4B (GE Healthcare, Piscataway, USA), and then incubated with HeLa cell nuclear extracts in RIPA buffer supplemented with protease inhibitors (Roche) for 4 h at 4 °C. After washing, proteins on beads were eluted by boiling in SDS loading buffer and examined by western blotting.

ChIP

ChIP was performed as described previously [14]. ChIPs were performed with the indicated antibodies, and the purified DNA samples were used for PCR or qPCR. The normal PCR products were amplified with the following specific primers [13]: H19 ICR forward primer, 5′-GAACAATGAGGTGTCCCAGTTCCA-3′; H19 ICR reverse primer, 5′-GGATAATGCCCGACCTGAAGATCT-3′; c-myc forward primer, 5′-GTTTTAAGGAACCGCCTGTCCTTC-3′; c-myc reverse primer, 5′-GGATTGCAAATTACTCCTGCCTCC-3′; BRCA1 forward primer, 5′-ACGATTAGCTGTCCGGAGACACGG-3′; BRCA1 reverse primer, 5′-GACTAGTTACTGTCTTTGTCCGCC-3′; β-globin 5′HS5 forward primer, 5′-TTCCCACAGTCTGTTGGTCACAG-3′; β-globin 5′HS5 reverse primer, 5′-AGATGTCCTGTCCCTGTAAGGTG-3′; exon 4/5 of H19 forward primer, 5′-TGCTGCACTTTACAACCACTGCAC-3′; and exon 4/5 of H19 reverse primer, 5′-TCATCCCGCTGGAGGAGCTCAGCT-3′. The primers based on the H19 ICR or H19 exon 4/5 were used as positive or negative controls, respectively. qPCR analysis was carried out with the My iQ Real-Time PCR Detection System (Bio-Rad) with SYBR Premix Ex TaqT (TaKaRa) and primers based on different portions of the Igf2/H19 locus [20]: CTCF AD1 forward primer, 5′-CTCTTCTCTCAATTCCCAAGGTTT-3′; CTCF AD1 reverse primer, 5′-CACCTTTCTGAAGCATCCGTTT-3′; CTCF AD2 forward primer, 5′-ACCCATGGCACCCTG ACA-3′; CTCF AD2 reverse primer, 5′-ACCAAGCTTTCGCCATTTAGC-3′; CCD forward primer, 5′-GGAGGAGGACAGAGGCAAGAG-3′; CCD reverse primer, 5′-AACAAAATTTCAGCCGGTTCA-3′; ICR forward primer, 5′-TGAATTTGCCCACAGGTGTTC-3′; and ICR reverse primer, 5′-GCCTTGGGTCACCTTCAGACT-3′. The results are presented as the percentage of input chromatin that was precipitated.

Acknowledgements

We thank W. W. Quitschke (State University of New York) for the pGEM-7Zf(–)/CTCF vector, M. Li (Sichuan University) for the RNAi-ready-pSIREN vector, and J. Huang (Sichuan Cancer Hospital) for the CNE1 cell line. This work was supported by the Doctoral Education Fund of Education Ministry of China (grant no. 20100181110051), National Natural Science Foundation of China (grant no. 30870757), National Natural Science Foundation of China (grant no. 81172372), and the China Medical Board of New York.

Author contributions

YQ conceived the study. QL performed most experiments. BY, XX, LW, WL, WY, YG, QZ, JZ, LJ conducted Yeast two-hybrid screening experiments. XY, WS, RL, XS performed Plasmid construction and pull down experiments. BL and YQ analysed the data, and QL wrote the paper. All authors discussed the results and commented on the manuscript.

Ancillary