Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2008 Feb;28(3):1124-35.
doi: 10.1128/MCB.01361-07. Epub 2007 Nov 26.

CTCF is the master organizer of domain-wide allele-specific chromatin at the H19/Igf2 imprinted region

Affiliations

CTCF is the master organizer of domain-wide allele-specific chromatin at the H19/Igf2 imprinted region

Li Han et al. Mol Cell Biol. 2008 Feb.

Abstract

A paternally methylated imprinting control region (ICR) directs allele-specific expression of the imprinted H19 and Igf2 genes. CTCF protein binding in the ICR is required in the maternal chromosome for insulating Igf2 from the shared enhancers, initiation of the H19 promoter transcription, maintaining DNA hypomethylation, and chromosome loop formation. Using novel quantitative allele-specific chromatin immunoprecipitation-single-nucleotide primer extension assays, we measured the chromatin composition along the H19/Igf2 imprinted domain in cells with engineered mutations at the four ICR-CTCF binding sites. Abolishing CTCF binding in the ICR reduced normally maternal allele-specific H3K9 acetylation and H3K4 methylation at the H19 ICR and promoter/gene body and maternal allele-specific H3K27 trimethylation at the Igf2 P2 promoter and Igf2 differentially methylated regions (DMRs). Paternal H3K27 trimethylation and macroH2A1 became biallelic in the mutant cells at the H19 promoter while paternal H3K9 acetylation and H3K4 methylation became biallelic at the Igf2 DMRs. We provide evidence that CTCF is the single major organizer of allele-specific chromatin composition in this domain. This finding has important implications: (i) for mechanisms of insulation since CTCF regulates chromatin at a distance, involving repression by H3K27 trimethylation at the Igf2 locus independently of repression by DNA hypermethylation; and (ii) for mechanisms of genomic imprinting since point mutations of CTCF binding sites cause domain-wide "paternalization" of the maternal allele's chromatin composition.

PubMed Disclaimer

Figures

FIG. 1.
FIG. 1.
Validation of MEFs for chromatin analysis. (A) Features of the H19/Igf2 imprinted domain in wild-type 129 × CS MEFs. H19 and Igf2 are expressed from the maternal (M) or paternal (P) allele, respectively (large arrows). In the opposite alleles both genes are silent (X). The paternal allele of the ICR is methylated (black lollipops). The maternal ICR allele is unmethylated (white lollipops) and binds CTCF insulator protein (vertical ovals) at four sites. CTCF blocks the interaction between the downstream enhancers (small horizontal ovals) and the Igf2 promoters (only P2 is indicated for simplicity). The Igf2 DMRs (gray rectangles) are paternally hypermethylated (lollipops in shades of gray). Certain regions are indicated according to their distance (−8, −4, and −3 kb) from the H19 transcription start site. (B) Summary of the changes in CTCFm × CS MEFs. (C) In vivo CTCF binding in the ICR is present in 129 × CS MEFs (white bars) but is absent in CTCFm × CS MEFs (black bars) due to the point mutations at four CTCF binding sites. Results of a ChIP real-time PCR experiment are shown. The left (−4 kb) and right (−3 kb) halves of the ICR were tested. (D) Igf2 levels are about double while H19 levels are greatly reduced in CTCFm × CS MEFs compared to normal cells. Results of a Northern hybridization are shown. Mutant versus wild-type relative values after normalization to Gapdh values are shown above the mutant bands. (E) Igf2 expression becomes biallelic in CTCFm × CS MEFs due to lack of ICR insulation. Allele-specific gene expression was assessed by RT-PCR SNuPE experiments (37). Numbers above images are relative expression levels of the presumed inactive alleles, the maternal allele of Igf2 (M) or the paternal allele (P) of H19, as a percentage of total (M + P) expression. Quantitation controls (0, 50 and 100% CS/total) are shown on the left. (F) ICR DNA methylation is biallelic in CTCFm × CS MEFs, because CTCF binding no longer protects the ICR from de novo methylation (39). Bisulfite methylation analysis results are shown. Individual, maternally (M) or paternally (P) inherited normal chromosomes or maternally inherited CTCFm chromosomes (M CTCFm) are shown. Methylated and unmethylated CpGs are indicated by closed and open squares, respectively. (G) Bisulfite methylation analysis of the H19 promoter sequences. DNA methylation is paternal allele-specific in normal and CTCFm × CS MEFs. The transcription start site (horizontal arrow) is indicated.
FIG. 2.
FIG. 2.
CTCF binding in the ICR is required for hypomethylation of the Igf2 DMR1 and DMR2. Bisulfite methylation analysis of the Igf2 DMR1 (A) and DMR2 (B) in 129 × CS and in CTCFm × CS MEFs is shown. DNA from reverse cross-linked input chromatin preparations (N- and X-ChIP) was analyzed to match the ChIP experiments. Other details are as described in the legend of Fig. 1F. (C) Summary of the methylation levels. Maternal (black bars) and paternal (gray bars) allele-specific CpG methylation at the Igf2 DMR1 (primary data from panel A) and DMR2 (primary data from panel B) in MEFs was compared to kidneys and livers of 17.5 dpc fetuses (primary data not shown). The maternal alleles of DMR1 and DMR2 were relatively hypomethylated compared to paternal alleles in all normal 129 × CS but not in mutant CTCFm × CS samples. (D) Summary of the methylation changes. Mutant maternal alleles in CTCFm × CS MEFs were hypermethylated compared to normal maternal alleles in response to distant ICR CTCF site mutations. The differences are calculated from data shown in panel C. pat, paternal; mat, maternal.
FIG. 3.
FIG. 3.
Novel ChIP-SNuPE assays for the analysis of allele-specific chromatin modifications. (A) Outline of the SNuPE assay. The SNuPE assay takes advantage of SNPs between parental alleles (37). We found these SNPs by DNA sequencing of inbred 129 and CS mouse strains at the specific regions of interest. In the ChIP-SNuPE assay, ChIP is performed first with a specific antibody. The maternal and paternal alleles are immunoprecipitated indiscriminately from the chromatin. DNA from the precipitated chromatin is PCR amplified with region-specific primers spanning the polymorphic sites. PCR fragments are isolated and probed with a primer abutting the SNP, and the ratio of incorporation of allele-specific radionucleotides is measured and expressed as the percent allele of the total precipitation. (B) Mixing experiments at the regions of interest. Aliquots of 129- and CS-type DNA were combined in the relative ratios of the percent CS to the total (CS + 129), indicated on the top, and subjected to SNuPE with the region-specific oligonucleotides. The measured values are shown above each of the images. The samples 0, 50, and 100% were included in the experiments for quantitating alleles in samples subjected to ChIP. The locations of the region-specific assays are indicated in the legend of Fig. 1A. (C) Plot of the measured values in the mixing experiment. The assays were rigorously quantitative, using a small amount (10 ng) of total DNA. (D) Example of ChIP-SNuPE quantitation of histone composition. The ICR (−4 kb) region was analyzed. Maternal (M) or paternal (P) allele specificity was almost identical between the reciprocal mouse crosses. The numbers above the gel images are the percentages of the paternal allele's contribution to the total immunoprecipitate. CS and 129 alleles are shown on top and bottom, respectively.
FIG. 4.
FIG. 4.
Allele-specific chromatin composition along the H19/Igf2 imprinted domain. Allele-specific chromatin composition was quantitated by ChIP-SNuPE assays (Fig. 3) at the specific regions indicated in Fig. 1A. ChIP was done using antibodies recognizing specific histone modifications (indicated above graphs) to precipitate chromatin from 129 × CS or reciprocal CS × 129 MEFs (indicated at the right of each row of charts). The ratio of the allele-specific histone modification at a specific region was expressed as a percentage of maternal (black bars) or paternal (gray bars) alleles in the total (maternal + paternal, or 100%) of the immunoprecipitation. Missing data points mean lack of enrichment for a specific chromatin modification at the specific region when marked as not detected (nd). (A) CTCF binding is maternal allele specific in the ICR. (B) Active chromatin modifications localize to the maternal allele at the H19 regions and to the paternal allele at the Igf2 regions. (C) Complex pattern of repressive chromatin marks along the imprinted domain (see text for details). Almost no allele-specific chromatin differences exist at a “neutral” intermediary region −8 kb upstream of the H19 promoter. Reciprocal mouse crosses exhibit nearly identical allele-specific chromatin composition. pat, paternal; mat, maternal.
FIG. 5.
FIG. 5.
CTCF is responsible for region-specific enrichment of chromatin components at the H19 and Igf2 loci. The overall enrichment for specific chromatin components was compared between normal 129 × CS MEFs (white bars) and CTCFm × CS MEFs (black bars) by real-time PCR. The location of the regions is shown in Fig. 1A. At the H19 locus, the activating chromatin marks showed reduced precipitation in CTCFm × CS MEFs compared to normal cells, but the precipitation levels of repressing marks, H3K27m3 and macroH2A1, increased. At the Igf2 regions, H3K27m3 levels were greatly decreased in the mutant cells, while activating chromatin marks were more abundant at the DMR1 and at the Igf2 P2 promoter. There was no change at the −8-kb region. Average values are shown with standard deviations.
FIG. 6.
FIG. 6.
CTCF is required for allele-specific chromatin composition locally and at a distance. ChIP-SNuPE analyses of allele-specific chromatin composition reveal the consequences of ICR CTCF site mutations. (A) Repressive chromatin marks shift toward the maternal allele at the H19 locus. (B) Activating chromatin marks shift toward the maternal allele at the distant Igf2 locus. Chromatin was precipitated in duplicates from 129 × CS and CTCFm × CS MEFs with the specific antibodies indicated. SNuPE assays were performed with region-specific oligonucleotides. Allele-specific histone modification at a specific region was expressed as a percentage of maternal or paternal (dark or light color, respectively) alleles in the total immunoprecipitate. The average values are shown with standard deviations. pat, paternal; mat, maternal.
FIG. 7.
FIG. 7.
Summary of the results. (A) Specific regions analyzed in this study. (B) Allele-specific chromatin composition at the H19 and Igf2 loci in normal MEFs. M, maternal chromosome; P, paternal chromosome. Allele-specific activating (green) and repressing (red) signals are shown in the respective chromosomes. Chromatin components with low relative abundance are in parentheses. Note that repressive chromatin at the Igf2 locus is present at the hypomethylated allele. (C) Paternalization of chromatin composition along the H19/Igf2 imprinted region. The chromatin composition of the maternally inherited CTCFm chromosome became similar to that of the normal paternal chromosome. Vertical arrows in panels B and C indicate changes (decreases and increases, respectively) in enrichment for specific histone modifications at specific regions in the CTCFm × CS compared to 129 × CS MEFs.
FIG. 8.
FIG. 8.
The H19 promoter is unmethylated in the early embryo. The results of the bisulfite sequencing experiment are shown. The embryo proper was analyzed at 6.5 dpc. Other details are as described in the legend of Fig. 1.

References

    1. Bartolomei, M. S., A. L. Webber, M. E. Brunkow, and S. M. Tilghman. 1993. Epigenetic mechanisms underlying the imprinting of the mouse H19 gene. Genes Dev. 71663-1673. - PubMed
    1. Bell, A. C., and G. Felsenfeld. 2000. Methylation of a CTCF-dependent boundary controls imprinted expression of the Igf2 gene. Nature 405482-485. - PubMed
    1. Bell, A. C., A. G. West, and G. Felsenfeld. 1999. The protein CTCF is required for the enhancer blocking activity of vertebrate insulators. Cell 98387-396. - PubMed
    1. Brannan, C. I., and M. S. Bartolomei. 1999. Mechanisms of genomic imprinting. Curr. Opin. Genet. Dev. 9164-170. - PubMed
    1. Choo, J. H., J. D. Kim, J. H. Chung, L. Stubbs, and J. Kim. 2006. Allele-specific deposition of macroH2A1 in imprinting control regions. Hum. Mol. Genet. 15717-724. - PubMed

Publication types