|
Resolution: standard / high Figure 1.
A model for the maintenance of chromatin silencing. (a) Replication-coupled maintenance of a silent region of chromatin. Solid lines indicate
DNA; cylinders represent nucleosomes (light, old; dark, newly added); circles represent
other proteins or protein complexes; flags indicate histone methylation; M indicates
DNA methylation (with new methylation in bold); the large oval represents DNA polymerase.
Before replication (1), a silenced region is marked by histone methylation and DNA
methylation on CG and CNG motifs. As the polymerase moves along the leading strand
from left to right (2), methylation on CG dinucleotides is passively maintained behind
the replication fork by MET1 (3). Old nucleosomes are randomly distributed between
the two chromatids and new nucleosomes are added by the CAF1 chaperone complex (4).
In the top chromatid in the diagram (A), there are two adjacent nucleosomes that are
methylated on H3 K9, thus providing cues for CAF1 to deposit a new nucleosome that
is methylated on H3 K9 by KYP (5). On the bottom chromatid (B), however, the nucleosome
distribution leads to loss of epigenetic information at the edge of the silent domain,
so new nucleosomes are deposited by CAF1 without H3 K9 methylation (6). CMT3 is therefore
able to use the cues provided by H3 K9 methylation to properly maintain CNG methylation
on the top chromatid (7), but not the bottom (8). Chromatin remodeling by DDM1 enables
both DNA and histone methylation, perhaps by allowing access of other proteins to
the DNA. (b) RNA-based reinforcement of silencing. The bottom (B) chromatid from (a) is shown after
the replication fork has passed completely. Now, siRNAs homologous to the silent region
guide H3 K9 methylation by KYP and DNA methylation by DRM. H3 K9 methylation also
allows the maintenance of CNG methylation by CMT3. The problem shown in (a) is thus
solved: the silent domain is fully maintained, despite random nucleosome distribution
during replication.
Zilberman and Henikoff Genome Biology 2004 5:249 doi:10.1186/gb-2004-5-12-249 |