MBIO.5.21.19

More Information About This Text

21 PROMOTERS AND ENHANCERS (Full Edition)

19 CpG islands are regulatory targets

A CpG island is a stretch of 1-2 kb in a mammalian genome that is rich in unmethylated CpG doublets.

CpG islands surround the promoters of constitutively expressed genes where they are unmethylated.
They are also found at the promoters of some tissue-regulated genes.
There are ~29,000 CpG islands in the human genome.
Methylation of a CpG island prevents activation of a promoter within it.
Repression is caused by proteins that bind to methylated CpG doublets.

The presence of CpG islands in the 5′ regions of some genes is connected with the effect of methylation on gene expression. These islands are detected by the presence of an increased density of the dinucleotide sequence, CpG.

The CpG doublet occurs in vertebrate DNA at only ~20% of the frequency that would be expected from the proportion of G•C base pairs. (This may be because CpG doublets are methylated on C, and spontaneous deamination of methyl-C converts it to T, introducing a mutation that removes the doublet.) In certain regions, however, the density of CpG doublets reaches the predicted value; in fact, it is increased by 10× relative to the rest of the genome. The CpG doublets in these regions are unmethylated.

These CpG-rich islands have an average G•C content of ~60%, compared with the 40% average in bulk DNA. They take the form of stretches of DNA typically 1-2 kb long. There are ~45,000 such islands altogether in the human genome (1443). Some of the islands are present in repeated Alu elements, and may just be the consequence of their high G•C-content. The human genome sequence confirms that, excluding these, there are ~29,000 islands. There are fewer in the mouse genome, ~15,500. About 10,000 of the predicted islands in both species appear to reside in a context of sequences that are conserved between the species, suggesting that these may be the islands with regulatory significance. The structure of chromatin in these regions has changes associated with gene expression (see Promoter activation involves an ordered series of events); there is a reduced content of histone H1 (which probably means that the structure is less compact), the other histones are extensively acetylated (a feature that tends to be associated with gene expression), and there are hypersensitive sites (as would be expected of active promoters) (710).

Figure 21.28

The typical density of CpG doublets in mammalian DNA is ~1/100 bp, as seen for a γ-globin gene. In a CpG-rich island, the density is increased to >10 doublets/100 bp. The island in the APRT gene starts ~100 bp upstream of the promoter and extends ~400 bp into the gene. Each vertical line represents a CpG doublet.

In several cases, CpG-rich islands begin just upstream of a promoter and extend downstream into the transcribed region before petering out. Figure 21.28 compares the density of CpG doublets in a "general" region of the genome with a CpG island identified from the DNA sequence. The CpG island surrounds the 5′ region of the APRT gene, which is constitutively expressed.

All of the "housekeeping" genes that are constitutively expressed have CpG islands; this accounts for about half of the islands altogether. The other half of the islands occur at the promoters of tissue-regulated genes; only a minority (<40%) of these genes have islands. In these cases, the islands are unmethylated irrespective of the state of expression of the gene. The presence of unmethylated CpG-rich islands may be necessary, but therefore is not sufficient, for transcription. So the presence of unmethylated CpG islands may be taken as an indication that a gene is potentially active, rather than inevitably transcribed. Many islands that are nonmethylated in the animal become methylated in cell lines in tissue culture, and this could be connected with the inability of these lines to express all of the functions typical of the tissue from which they were derived.

Methylation of a CpG island can affect transcription. Two mechanisms can be involved:

Methylation of a binding site for some factor may prevent it from binding. This happens in a case of binding to a regulatory site other than the promoter (see Oppositely imprinted genes can be controlled by a single center).
Or methylation may cause specific repressors to bind to the DNA.

Repression is caused by either of two types of protein that bind to methylated CpG sequences (for review see 2424). The protein MeCP1 requires the presence of several methyl groups to bind to DNA, while MeCP2 and a family of related proteins can bind to a single methylated CpG base pair. This explains why a methylation-free zone is required for initiation of transcription. Binding of proteins of either type prevents transcription in vitro by a nuclear extract (711).

MeCP2, which directly represses transcription by interacting with complexes at the promoter, is bound also to the Sin3 repressor complex, which contains histone deacetylase activities (see Figure 23.15). This observation provides a direct connection between two types of repressive modifications: methylation of DNA and acetylation of histones.

The absence of methyl groups is associated with gene expression. However, there are some difficulties in supposing that the state of methylation provides a general means for controlling gene expression. In the case of D. melanogaster (and other Dipteran insects), there is very little methylation of DNA (although there is gene potentially coding a methyltransferase), and in the nematode C. elegans there is no methylation of DNA. The other differences between inactive and active chromatin appear to be the same as in species that display methylation. So in these organisms, any role that methylation has in vertebrates is replaced by some other mechanism.

We have described three changes that occur in active genes:

A hypersensitive site (s) is established near the promoter.
The nucleosomes of a domain including the transcribed region become more sensitive to DNAase I.
The DNA of the same region is undermethylated.

All of these changes are necessary for transcription.

Last Revised on 12-16-2002

2424 Bird, A. (2002). DNA methylation patterns and epigenetic memory. Genes Dev. 16, 6-21. PubMed Journal

710 Bird, A. et al. (1985). A fraction of the mouse genome that is derived from islands of nonmethylated, Cp-G-rich DNA. Cell 40, 91-99. PubMed Journal
711 Boyes, J. and Bird, A. (1991). DNA methylation inhibits transcription indirectly via a methyl-CpG binding protein. Cell 64, 1123-1134. PubMed Journal
1443 Antequera, F. and Bird, A. (1993). Number of CpG islands and genes in human and mouse. Proc. Natl. Acad. Sci. USA 90, 11995-11999. PubMed Journal

Instructors: More Information About This Text | Jones and Bartlett Biological Science Titles