Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2016 Aug 19;44(14):6693-706.
doi: 10.1093/nar/gkw258. Epub 2016 Apr 15.

Protection of CpG Islands From DNA Methylation Is DNA-encoded and Evolutionarily Conserved

Affiliations
Free PMC article

Protection of CpG Islands From DNA Methylation Is DNA-encoded and Evolutionarily Conserved

Hannah K Long et al. Nucleic Acids Res. .
Free PMC article

Abstract

DNA methylation is a repressive epigenetic modification that covers vertebrate genomes. Regions known as CpG islands (CGIs), which are refractory to DNA methylation, are often associated with gene promoters and play central roles in gene regulation. Yet how CGIs in their normal genomic context evade the DNA methylation machinery and whether these mechanisms are evolutionarily conserved remains enigmatic. To address these fundamental questions we exploited a transchromosomic animal model and genomic approaches to understand how the hypomethylated state is formed in vivo and to discover whether mechanisms governing CGI formation are evolutionarily conserved. Strikingly, insertion of a human chromosome into mouse revealed that promoter-associated CGIs are refractory to DNA methylation regardless of host species, demonstrating that DNA sequence plays a central role in specifying the hypomethylated state through evolutionarily conserved mechanisms. In contrast, elements distal to gene promoters exhibited more variable methylation between host species, uncovering a widespread dependence on nucleotide frequency and occupancy of DNA-binding transcription factors in shaping the DNA methylation landscape away from gene promoters. This was exemplified by young CpG rich lineage-restricted repeat sequences that evaded DNA methylation in the absence of co-evolved mechanisms targeting methylation to these sequences, and species specific DNA binding events that protected against DNA methylation in CpG poor regions. Finally, transplantation of mouse chromosomal fragments into the evolutionarily distant zebrafish uncovered the existence of a mechanistically conserved and DNA-encoded logic which shapes CGI formation across vertebrate species.

Figures

Figure 1.
Figure 1.
Hypomethylated regions (HMRs) on human chromosome 21 are largely recapitulated in the transchromosomic mouse model. (A and B) Profiles of non-methylated DNA (BioCAP-seq) at three regions on human chromosome 21 in human (upper) and in Tc1 mouse (lower, inverted) liver (A) and testis (B) tissues. Genes are depicted above the BioCAP traces. (C and D) Heatmaps depicting BioCAP (left) and H3K4me3 ChIP-seq (right) signal across human chromosome 21 HMRs in the human and Tc1 mouse liver (C) and testis (D) tissues. Signal is ranked according to HMR length and aligned to the centre of the HMR. Scalebar in kb. (E) Scatterplots of BioCAP-seq reads for human and Tc1 mouse at all human HMRs from liver (left) and testis tissue (right).
Figure 2.
Figure 2.
Identification of species-specific HMRs. (A and B) BioCAP-seq traces across two human-specific (A) and two Tc1-specific (B) HMRs on human chromosome 21. Species-specific HMRs (ssHMRs) are indicated by a horizontal bar below the BioCAP-seq traces (upper). Bisulfite sequencing at species-specific HMRs confirms alterations in methylation at these sites (lower). Bisulfite amplicons (BA) are depicted by a horizontal black bar, CpG dinucleotides by a vertical line and the methylation status of each CpG in human or Tc1 liver is depicted as a vertical line between 0 and 100%. (C) Scatter plot of BioCAP-seq reads for human and Tc1 mouse at all HMRs to illustrate human and Tc1-specific HMRs in liver (upper) and testis (lower). (D and E) Heatmaps of BioCAP-seq signal in human and Tc1 mouse liver (D) and testis (E) tissues illustrate that a subset of HMRs are differentially methylated. Heatmaps are ranked according to HMR length and aligned to the centre of the HMR with shared (upper), human-specific (middle), and Tc1-specific (lower) sites clustered together. Scalebar in kb. (F) Venn diagrams depicting the overlap between human-specific HMRs (upper) or Tc1-specific HMRs (lower) from different tissues.
Figure 3.
Figure 3.
The majority of species-specific HMRs are distal to gene promoter regions. (A) Venn diagrams depicting the overlap between HMRs in the human and Tc1 mouse tissue. HMRs are segregated into those located at transcription start site (TSS ± 500bp) (upper) and those found away from TSSs (lower). Human-specific HMRs are coloured in dark grey, shared HMRs in grey and Tc1-specific HMRs in light grey. (B) Boxplots depicting fold-change in BioCAP signal between Tc1 mouse and human at Tc1-specific HMRs located at a gene TSS or elsewhere in the chromosome. Tc1-specific HMRs away from gene TSSs exhibit a greater fold-change in BioCAP signal compared to those at gene TSSs. Significance values calculated using Welch's T-test (*P < 0.05).
Figure 4.
Figure 4.
Tc1-specific HMRs are CpG-rich and often associated with young repetitive DNA elements. (A and B) CpG density (A) and GC content (B) plots of liver (upper panel, red) and testis (lower panel, blue) HMRs that are shared between human and Tc1 mouse (dark line) or are Tc1-specific (pale line). A background control is indicated (grey). (C) Boxplots depicting age of repeats (in million years) associated with HMRs in liver (left) and testis tissue (right). HMRs are segregated into those that are shared in human and Tc1 mouse (purple) and are Tc1-specific (pink). The difference in repeat age between shared and Tc1-specific HMRs is significantly different as calculated by a Mann–Whitney U test (*P < 5 × 10−4). (D and E) Snapshots of two Tc1-specific HMRs associated with a repetitive element. (F) A snapshot of a gene promoter on human chromosome 21 with a shared HMR, and which is associated with RNA PolII in both the Tc1 mouse and endogenous human nuclear environment. (G) RNA PolII enrichment at human-Tc1 shared TSS-associated HMRs (upper) and Tc1-specific non-TSS-associated HMRs (lower) in human and Tc1 mouse liver.
Figure 5.
Figure 5.
Species-specific transcription factor binding is widely associated with species-specific HMR formation. (A and B) CpG density (A) and GC content (B) plots of liver (upper panel, red) and testis (lower panel, blue) HMRs that are shared between human and Tc1 mouse (dark line) or are human-specific (dashed line). A background control is indicated (grey). (C) A snapshot illustrating a human-specific HMR corresponding to a human-specific transcription factor binding event as indicated by CEBPA and HNF4A ChIP-seq signal. (D) A 4-way Venn diagram comparing human-specific HMRs to human-specific transcription factor binding sites (TFBSs) for CEBPA, HNF4A and CTCF on human chromosome 21. 31% of human-specific HMRs overlapped a human-specific TFBS. (E) A snapshot illustrating a Tc1 mouse-specific HMR corresponding to a Tc1 mouse-specific transcription factor binding event as indicated by HNF4A and CTCF ChIP-seq signal. (F) A 4-way Venn diagram comparing Tc1-specific HMRs to Tc1-specific transcription factor binding sites for CEBPA, HNF4A and CTCF on human chromosome 21. 28% of Tc1-specific HMRs overlapped a Tc1-specific TFBS.
Figure 6.
Figure 6.
The DNA encoded principles that underpin HMR formation are conserved across vast expanses of divergent evolution. (A) BioCAP-seq profile of a mouse chromosomal DNA fragment introduced into the zebrafish genome and analysed at 28–30 h post-fertilisation (hpf). The BioCAP signal from three representative mouse cell-types ES cells, liver and testis (green, red and blue traces) and the BioCAP trace observed for this locus in the developing zebrafish (grey) is indicated. CpG density and GC content are depicted in black. All four mouse HMRs form on the mouse BAC DNA in the zebrafish embryo and a cluster of repetitive LTR elements in the centre of the mouse BAC form zebrafish-specific HMRs. (B) A snapshot of a promoter-associated mouse HMR. (C) A snapshot of a broadly hypomethylated region. (D) A snapshot of a zebrafish-specific HMR region that forms at a CpG and GC-rich mouse exonic region that is normally methylated in mouse tissues. (E) A snapshot illustrating a cluster of mouse LTR elements which are CpG dense and become hypomethylated in zebrafish. (FH) Scatterplots comparing BioCAP-seq read counts at mouse HMRs in zebrafish with mouse liver (F), testis (G) and embryonic stem cells (H).

Similar articles

See all similar articles

Cited by 17 articles

See all "Cited by" articles

References

    1. Jones P.A. Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat. Rev. Genet. 2012;13:484–492. - PubMed
    1. Klose R.J., Bird A.P. Genomic DNA methylation: the mark and its mediators. Trends Biochem. Sci. 2006;31:89–97. - PubMed
    1. Sharif J., Muto M., Takebayashi S., Suetake I., Iwamatsu A., Endo T.A., Shinga J., Mizutani-Koseki Y., Toyoda T., Okamura K., et al. The SRA protein Np95 mediates epigenetic inheritance by recruiting Dnmt1 to methylated DNA. Nature. 2007;450:908–912. - PubMed
    1. Bird A. DNA methylation patterns and epigenetic memory. Genes Dev. 2002;16:6–21. - PubMed
    1. Jones P.A., Liang G. Rethinking how DNA methylation patterns are maintained. Nat. Rev. Genet. 2009;10:805–811. - PMC - PubMed

Publication types

MeSH terms

Feedback