Signposting genes: the pull of an epigenetic signal
Intriguing new work from researchers in the Biochemistry department along with collaborators at Harvard University has finally shed light on the function of a feature of vertebrate genomes that has eluded researchers for over 20 years.
Dr Rob Klose and colleagues have shown how CpG islands - enigmatic regions of the genome associated with genes - may play a proactive role in regulating genes. They have published their findings in Molecular Cell (1).
Chemistry on the CpG island - how the methyl group is removed from lysine by KDM enzymes (Click to enlarge)
CpG islands are genomic regions that have higher than expected levels of the dinucleotide sequence CG, or 'CpG' as it is called. CpG sequences at islands are resistant to a chemical modification known as methylation which is seen at CpG sequences elsewhere in the genome. CpG islands are spread throughout the genome, largely in the regulatory or 'promoter' sequences near genes. Despite this intriguing location, researchers have not been able to pin down any role for the islands in regulating genes.
Dr Klose's interest in these elusive genomic regions was prompted by an unexpected finding. He was studying a chemical modification of one of the constituent proteins of chromatin, the complex combination of DNA and proteins that makes up chromosomes. The protein is Histone H3 which can acquire methyl groups on its lysine amino acid residues.
Dr Klose explains: 'We know that there is histone lysine methylation at promoters of every gene and also in the body of the genes. When we looked at the genome-wide profiles of this, what we noticed is that they quite often correlated with boundaries of CpG islands. We thought there must be some relationship - the profiles matched each other almost perfectly.'
A clue to what that relationship might be came from some studies carried out 10 years ago. Researchers had identified a short protein sequence that specifically recognises nonmethylated CpG dinucleotides in vitro. A family of proteins recently found to carry this sequence are enzymes called histone demethyases that remove methyl groups from lysine residues in histone H3. But, says Dr Klose, nobody asked whether these or other proteins carrying the sequence are binding to CpG islands in the genome.
'It's been so long with so little headway made on CpG islands. It's a really tough question. But our finding has opened up a whole new way of thinking about what a CpG island is'
The researchers confirmed that KDM2A, a histone demethyase enzyme, specifically binds to nonmethylated CpG DNA both in the test-tube and in cells. The enzyme is unable to bind when the CpG DNA is methylated. In a key series of experiments which tapped into the bioinformatics expertise of collaborators Peter Park and Michael Tolstorukov at Harvard, the group went on to show that the enzyme associates with CpG islands in the genome.
'We used a technique called ChIP sequencing, on a genomic scale, to see how widespread the phenomenon is', explains Dr Klose. 'We've really demonstrated that this is something that's occurring genome-wide.'
In further experiments, they showed that KDM2A, once bound to nonmethylated CpG islands, appears to remove the methyl groups from histone H3 lysine residues in these regions. This suggests that KDM2A is marking out the islands from neighbouring DNA.
'It's a relatively straightforward mechanism of using the epigenetic signal in DNA to force these enzymes to this region, and then translating that DNA-encoded information to create a surrounding chromatin environment', says Dr Klose. 'This isn't directly impinging on gene activation - it's really creating a chromatin environment which is different from the rest of the genome.'
He explains why the cell would need to signpost these regions of the genome: 'You've got a really big mammalian genome that's full of junk DNA and other structural regions, and these are regions where you don't want transcription factors to be working. So we think that this chromatin environment is a way of highlighting regulatory elements.'
Diagrammatic representation of a CpG island showing DNA wrapped around histone H3 protein, a CpG island and transcription start site (TSS), and the enzyme KDM2A removing methyl groups (K36) on lysine residues of H3 at the island.
The researchers would like to understand more about how this important chromatin architecture is built up. They hope to achieve this eventually by perturbing the enzymes that create this unique chromatin signature, but a more fundamental understanding of the properties of the system is required before the group can embark on these sorts of studies.
'There are many intersecting pathways leading to this architecture,' comments Dr Klose. 'We're starting with so little information about CpG islands or the system itself. There's a lot of groundwork we need to do to build on understanding how the system works before we can really go after perturbing it in the right way.'
One thing that they do know about the chromatin architecture is that it changes as cells move through the cell division cycle. CpG chromatin is set up following cell division, preparing the genome for the transcriptional tasks ahead of it. Removal of histone lysine methyl groups may play a part in creating a chromatin environment that allows transcription to take place.
'We're trying to examine the cell cycle to see if there's an opportunity to understand how the system is built. But we're also doing a lot of in vitro work trying to understand at the reductionist biochemical level exactly how the proteins interface with the chromatin structure.'
In the meantime, the findings from Dr Klose and his colleagues are likely to stimulate interest across a wide spread of researchers.
'It's been so long with so little headway made on CpG islands. It's a really tough question. But our finding has opened up a whole new way of thinking about what a CpG island is. The paper hits to many people - the CpG island crowd, the transcription field, and also the people who are interested in chromatin structure.'