Here’s a research highlight that will appear soon on Nature Reports Stem Cells. This version has the addition of outside comment. It came in too late to be incorporated into the highlight, but I'm putting it here because I think it's interesting.
Multitasking methyltransferase: G9a silences gene expression two ways
As embryonic stem cells differentiate, the pluripotency gene known as Oct4 goes on lockdown. In fact, the guards to gene expression are doublelocked:the gene-encoding DNA strands are wound up into a structure called heterochromatin, in which the DNA is complexed with histones and other proteins in such a way that it is inaccessible to the transcriptional machinery. Furthermore, gene-expression machinery is kept at bay by chemical modifications to the DNA that signals the start of a gene. New work published in Nature Structural and Molecular Biology1 shows not only that both of these modifications are regulated by a single master protein, the histone methyltransferase G9a, but that this enzyme apparently brings about the inactivation of many early embryonic genes.
Led by Howard Cedar and Yehudit Bergman of Hadassah Hebrew University in Jerusalem, Israel, a team of researchers examined mouse embryonic stem (ES) cells as they began to differentiate. They used microarrays to compare cells at different stages of differentiation as well as cells that could and could not express G9a. This identified a number of genes besides Oct4 that were newly methylated during differentiation; in other words, as cells lost pluripotency, a set of genes was silenced by way of chemical modification to certain regions of DNA, and this methylation also seems to be under the control of G9a.
G9a helps convert chromatin into heterochromatin, in which gene expression is blocked.\Interestingly, DNA methylation and heterochromatinization seem to be independent of each other. The researchers blocked the heterochromatinization activity of G9a by introducing a mutation into the protein that prevents it from methylating a lysine 9 residue on histone protein H3 . This mutation did not, however, change patterns of gene methylation.
G9a seems to regulate DNA methylation by recruiting two well known DNA methyltransferases (Dnmt3a and Dnmt3b) to relevant sites in the genome, sites where G9a is helping to create heterochromatin. The authors speculate that other histone methyltransferases may also promote DNA methylation as well as heterochromatinization.
G9a directs both processes as mouse embryos transition between pre- and post-implantation stages. For a cell to be reprogrammed back to a pluripotent state, these states must be reversed, and while heterochromatin can be remodelled during cell division, DNA methylation patterns are often faithfully reproduced, and this, the authors believe, constitutes the main barrier to reprogramming. "G9a seems to serve as a master regulator involved in turning off pluripotency," says Cedar. His future work will be aimed at understanding how these processes reverse themselves when somatic cell genomes are reprogrammed to pluripotency.
1. Epsztejn-Litman, S. et al. De novo DNA methylation promoted by G9a prevents reprogramming of embryonically silenced genes. Nature Struct. Mol. Biol., advance online publication, doi:10.1038/msmb.1476 (26 October 2008). | Article |
When I write research highlights, I try to find other scientists to put that work in perspective. For the highlight below, I contacted the corresponding author of an earlier paper on the role of microRNAs and methylation in embryonic stem cells, which seemed to suggest another mechanism for silencing Oct4 as ES cells differentiate. I received a reply too late to work it into the highlight, so I am putting the responses to my four questions below. This is a joint response from Lasse Sinkkonen in Friedrich Miescher Institute for Biomedical Research in Basel and Petr Svoboda from Institute of Molecular Genetics in Prague.
1) How does this finding from Cedar and Bergman fit in with your work, particularly in terms of the methylation of Oct4?
Our work was actually inspired by their work because we saw these data
at Bergman's talk while ago and we decided to test if Dicer-/-ES cells
behave like G9a and DNMT3a/b-/-. Their work is in no contradiction to
our work as we show that miRNAs regulate DNMT3a/b, which are required
for G9a-induced methylation. We only saw a small effect of miRNAs on G9a mRNA levels (~1.5-fold) and since H3K9methylation of Oct-3 promoter
occurs without problems, we believe that miRNAs mainly affect
methylation downstream of G9a and independently of its DNMT3 binding.
2) How surprising/important is this most recent finding?
It is very interesting because it was known for a long time that there is a connection between histone modifications and DNA methylation. G9a work is important because it provides a mechanistical link between histone methylation and DNA methylation by dual function histone
methyltransferase. And it is surprising to see that if the histone
methylation activity of G9a is abolished, G9a can still repress its
target gene via DNA methylation.
3) What additional studies should be done before the paper's conclusions can be accepted?
There are three studies from October (two in EMBO J), which corroborate
the model, so I don't have any major problem with accepting NCB
conclusions.
4) What new questions/ applications does this finding open up?
DNA methylation has been perceived by many as a security lock on
transcriptionally silenced genes, which appears after histone
modifications. The latest results raise a question whether DNA
methylation could have a more active role in gene silencing and could be more of a parallel silencing system than a follow up guardian.
