Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Developmental biology

Two paths to silence merge

To maintain their identity across generations, specialized cells must heritably repress swathes of genes — keeping active only genes necessary for the cell's purpose. Now it seems two repressive pathways join forces.

Even though cells of all developmental stages carry the same DNA, they have their own identity, defined by the combination of proteins expressed in each cell. These expression patterns, although set early during development, are reproduced in each mature, specialized cell later in life, over many cell divisions. This phenomenon is referred to as ‘cellular memory’. Any perturbation of this fine-tuned system is a hazard to proper development and health.

Cellular memory is thought to be regulated by two ‘epigenetic’ mechanisms, which heritably change the characteristics of the cell without altering its DNA sequence. These mechanisms do this by altering the chromatin (the DNA and its associated proteins), which changes the availability of the genes to be expressed. One mechanism involves a group of enzymes that attach methyl groups to the DNA, the DNA methyltransferases (DNMTs); the other involves Polycomb group (PcG) proteins, which modify the histone proteins around which the DNA is wrapped. On page 871 of this issue, Viré et al.1 show a direct interaction between DNMTs and the Polycomb protein EZH2 that hints at how these complex systems might collaborate to set up cellular memory.

DNA methylation helps to reorganize chromatin into a ‘silent’ state in which genes are not expressed. In mammals, DNMT3A and -3B are mainly responsible for establishing methylation at previously unmethylated sites, whereas DNMT1 is the major maintenance methyltransferase, reproducing existing methylation patterns during cell division2.

PcG proteins were originally identified for their role in maintaining the repressed state of certain developmental genes in the fruitfly Drosophila. EZH2 is a histone methyltransferase that can methylate lysine 27 of the tail of histone H3 (H3-K27) or lysine 26 of histone H1 (H1-K26)3. This enzyme functions as part of the Polycomb repressive complexes 2 and 3 (PRC2/3), and whether it methylates H1 or H3 depends on which other proteins are in the complexes3. These histone modifications can attract other repressive complexes, such as Polycomb repressive complex 1 (PRC1), that then propagate the silenced state4,5.

In their study, Viré et al. show convincingly that EZH2 interacts with all three DNMTs in vitro and in vivo (Fig. 1). DNMTs have not been found previously in purified PcG complexes from Drosophila or mammals, perhaps indicating that the interactions observed by Viré et al. are transient. How could these interactions affect the gene-expression programme of the cell? The authors find that proper repression of a few genes targeted by EZH2 requires both EZH2 and DNMTs. Moreover, EZH2 is needed to bring DNMTs to the regulatory regions (‘promoters’) of the EZH2 target-genes. Depletion of EZH2 disturbs this recruitment, de-repressing the genes and allowing expression. At the same time, there is a decrease in the repressive histone H3-K27 methylation marks.

Figure 1: Making epigenetic methylation marks.

The Polycomb group protein EZH2 acts as part of the Polycomb repressive complexes 2 and 3 (PRC2/3), which add methyl groups to histone proteins, primarily to lysine 27 of histone H3 (small blue circles). Viré et al.1 show that EZH2 interacts with DNA methyltransferases (DNMTs) to recruit them to the regulatory region of certain target genes. The DNMTs methylate DNA (small green circles), either setting up new methyl marks during development, or maintaining the marks after each round of cell division. These marks are correlated with gene repression, and they could have a role in recruiting further repressive complexes, such as PRC1 or histone deacetylase co-repressors, through methyl-binding domain proteins.

These results reveal that EZH2, as part of the PRC2/3 complexes, can physically recruit DNMTs to certain target-genes and that this process is essential for silencing the genes. Surprisingly, reducing the levels of any of the DNMTs separately results in similar de-repression of the EZH2 target-genes. This is unexpected because the different DNMTs by and large seem to fulfil distinct roles in development and cellular viability2. It is possible that various EZH2-containing complexes might exist that can interact with the de novo DNMTs or the maintenance DNMT under different circumstances. This would allow PcG complexes to take part in initiating and maintaining transcriptional programmes depending on the chromatin environment and stimuli.

Viré et al. show that overexpression of EZH2 increases the DNA methylation of the promoters of its target genes, and that, remarkably, when EZH2 is downregulated the methylation decreases rapidly. This highlights the dynamic nature of DNA methylation and suggests that, just as histones are associated with separate enzymes for methylation and demethylation of their tails6, DNA demethylases may exist to actively remove the methyl marks, although such enzymes remain elusive.

How is DNA methyltransferase activity of DNMTs regulated by EZH2 histone methyltransferase activity? The authors propose a model whereby DNMTs require the EZH2-mediated histone H3-K27 methylation to methylate the DNA at target-gene promoters. This is analogous to the link proposed between the H3-K9 histone methylation system and DNA methylation in certain silent chromatin regions7. One possibility is that DNMTs recognize the H3-K27 methylation marks directly or indirectly through the interaction with another yet unknown partner that can regulate DNMT accordingly. Another possibility is that the H3-K27 methylation mark is not the engaging signal, but that DNMTs could be activated by EZH2 directly, perhaps by methylation.

DNMT recruitment to EZH2 target-genes not only adds another layer of repression to ‘lock in’ the silent state by methylating the surrounding DNA, but it might also help to recruit PRC1 proteins8. It is also possible that EZH2 acts independently of PRC1 by recruiting co-repressors that associate with DNMTs. In both scenarios, DNA methylation of EZH2 target-genes can lead to the recruitment of proteins that bind to the methyl-DNA marks, and the formation of more repressive complexes (Fig. 1). So are all EZH2 target-genes also silenced by DNA methylation, and are they overlapping with or separate from genes repressed by the EZH2-PRC1 interaction? Notably, the mechanism by which PcG-induced silencing is inherited is still not known. Perhaps, for a select group of target genes, EZH2-coupled DNA methylation can participate in preserving PcG-related cellular memory during DNA replication, through the maintenance DNMT (Fig. 2).

Figure 2: Inheritance of epigenetic memory.

A possible mechanism for inheritance of Polycomb-related epigenetic memory through DNA methylation during DNA replication. As the DNA is replicated, the maintenance DNA methyltransferase, DNMT1, adds a methyl mark (open green circles) to the newly replicated strand (purple), thus passing the DNA methylation mark to the progeny of the dividing cell through DNA replication. Viré et al.1 report that DNMT1 interacts with the histone methyltransferase EZH2, contained in the PRC2/3 complexes, possibly recruiting the complexes to sites of newly methylated DNA, to ensure that the histones too are in a repressive state.

Although Viré et al. provide a simple and attractive model, their findings cannot easily be generalized to other PcG target-genes. For example, during X-chromosome inactivation, when one entire X chromosome is silenced in female mammals, EZH2 and another PcG protein known as EED accumulate on the inactive X chromosome at the onset of the process — well before there is any DNA methylation9. Moreover, paternal-specific repression of the Kcnq1 gene domain in mouse placenta depends mainly on recruitment of EZH2 and EED with no apparent promoter DNA methylation10,11. These and other examples12,13 imply that although PcG-related silencing and DNA methylation cooperate in many cases, they are not always coupled. One intriguing possibility is that DNMTs could be inactive when they are recruited by EZH2, only being activated later by some as yet unknown regulatory signal.

Finally, the core members of the PRC2 complex are evolutionarily conserved from plants to mammals5. Yet, the nematode Caenorhabditis elegans seems to lack any detectable DNA methylation or genomic sequences resembling genes encoding DNMTs14, and fly DNA has very little methylation5,15. So, the connection between PcG proteins and DNA methylation must have evolved relatively recently.

Notwithstanding the questions that remain, the demonstration that the PcG protein EZH2 and DNA methylation can be directly coupled will provoke new hypotheses regarding the dynamics and heritability of epigenetic marks that can now be explored.


  1. 1

    Viré, E. et al. Nature 439, 871–874 (2006).

    ADS  Article  Google Scholar 

  2. 2

    Li, E. Nature Rev. Genet. 3, 662–673 (2002).

    CAS  Article  Google Scholar 

  3. 3

    Kuzmichev, A., Jenuwein, T., Tempst, P. & Reinberg, D. Mol. Cell 14, 183–193 (2004).

    CAS  Article  Google Scholar 

  4. 4

    Lund, A. H. & van Lohuizen, M. Curr. Opin. Cell Biol. 16, 239–246 (2004).

    CAS  Article  Google Scholar 

  5. 5

    Ringrose, L. & Paro, R. Annu. Rev. Genet. 25, 413–434 (2004).

    Article  Google Scholar 

  6. 6

    Yu-ichi, T. et al. Nature advance online publication doi:10.1038/nature04433 (2005).

  7. 7

    Lehnertz, B. et al. Curr. Biol. 13, 1192–1200 (2003).

    CAS  Article  Google Scholar 

  8. 8

    Hernández-Muñoz, I., Taghavi, P., Kuijl, C., Neefjes, J. & van Lohuizen, M. Mol. Cell. Biol. 25, 11047–11058 (2005).

    Article  Google Scholar 

  9. 9

    Heard, E. Curr. Opin. Cell Biol. 16, 1–9 (2004).

    Article  Google Scholar 

  10. 10

    Umlauf, D. et al. Nature Genet. 12, 1296–1300 (2004).

    Article  Google Scholar 

  11. 11

    Lewis, A. et al. Nature Genet. 12, 1291–1295 (2004).

    Article  Google Scholar 

  12. 12

    Yu-Wei, L. et al. Cancer Res. 64, 8184–8192 (2004).

    Article  Google Scholar 

  13. 13

    Feldman, N. et al. Nature Cell Biol. advance online publication doi:10.1038/ncb1353 (2006).

  14. 14

    Grace Goll, M. & Bestor, T. H. Annu. Rev. Biochem. 74, 481–514 (2005).

    Article  Google Scholar 

  15. 15

    Marhold, J., Kramer, K., Kremmer, E. & Lyko, F. Development 131, 6033–6039 (2004).

    CAS  Article  Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Taghavi, P., van Lohuizen, M. Two paths to silence merge. Nature 439, 794–795 (2006).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing