![]() |
|
| |||||||||||||||||||
Epigenomics: The first genome-wide methylome
Almost 19% of the A. thaliana genome is methylated, and besides the expected degree of methylation in heterochromatin, a considerable amount is also found in euchromatin. The first genome-wide DNA methylation map has been published. Analysis of the methylation landscape in Arabidopsis thaliana and how it changes, and with what consequences, in methylation-deficient mutants has provided important insights into how this DNA modification regulates gene expression.
The groups of Steve Jacobsen and Joseph Ecker used a combination of biochemical approaches and whole-genome tiling microarrays to map the methylated components of the A. thaliana genome. The resulting DNA methylation map has a resolution of 35 bp and reveals that almost 19% of the A. thaliana genome is methylated. As expected, much of this methylation occurs in heterochromatin, including centromeres, which harbours transposons and repetitive elements. But a considerable amount of methylation is also found in euchromatin. Unsurprisingly, the highest levels are seen in pseudogenes and unexpressed genes, but around 5% of expressed genes have methylated promoters and 33% are methylated within a transcribed region (the body-methylated genes). A bias against methylation in the 5' and 3' ends of expressed genes indicates that methylation might interfere with transcription initiation and termination. A comparison of DNA methylation sites with microarray expression data from a number of tissues and conditions revealed that body-methylated genes are more highly expressed than unmethylated genes, with promoter-methylated genes showing the lowest levels of expression. Promoter-methylated genes tend to be expressed in a tissue-specific manner, whereas body-methylated genes tend to be constitutively expressed. To gain insights into the processes that are regulated by DNA methylation, the authors analysed changes in methylation pattern and gene expression in the methyltransferase mutants: met1 (which lacks almost all CG and much of the non-CG methylation that is observed in the wild type) and the triple mutant drm1 drm2 cmt3 (which specifically lacks most non-CG methylation). Although methylation patterns in the triple mutant and the wild type are very similar (only 7% of the mutant genome loses methylation), 64% of the genome loses methylation in met1 mutants. The data are consistent with genic methylation occurring mostly at CG sites and most likely being associated with small interfering RNAs. Moreover, because previously silenced transposons and pseudogenes become transcribed in met1 mutants, CG methylation seems to be chiefly responsible for this silencing. Interestingly, non-CG methylation might be more important for the regulation of functional genes, as indicated by expression analysis in the triple mutants. The results also reveal that DNA methylation (mainly at CG sites) has an important role in regulating the expression of non-coding RNAs. All of the data from this impressive study are available online for further scrutiny. The wealth of information that this tour de force has yielded issues a challenge for those who work in other systems. As the authors point out, whole-genome tiling arrays are now available for many organisms, so the stage is set for other methylome analyses. Magdalena Skipper References
| |||||||||||||||||||
![]() | |||||||||||||||||||
HOME | SIGNALING UPDATE | MOLECULE PAGES | DATA CENTER | ABOUT US | |||||||||||||||||||