Biology Reference
In-Depth Information
A second mechanism relies on the presence of GC-rich sequences,
which are often bound by Polycomb components (
Ku et al., 2008;
Mendenhall et al., 2010
). Such sequences do not necessarily need to be
CpG islands, whose methylation is also associated with repressed gene activ-
ity. The majority of CpG islands are not bound by PRC2 (
Ku et al., 2008
)
and the binding of PRC2 can precede their methylation (
Mohn et al., 2008
).
Furthermore, CpG islands in the
HOXC
cluster are mostly unmethylated in
human brain cells (
Illingworth et al., 2008
) similar to CpG islands around
HOX
gene promoters in human lung fibroblast cells (
Weber et al., 2007
).
The dynamic recruitment of PRC2 to GC rich sequences may thus com-
plement the repressive effect of DNA methylation.
A third mechanism of targeted Polycomb recruitment involves the func-
tion of large noncoding RNAs, as exemplified by HOTAIR (
Gupta et al.,
2010; Rinn et al., 2007
). HOTAIR is transcribed from a region between the
HOXC12
and
HOXC11
genes. It was shown to recruit PRC2 in
trans
,to
many target sites in the genome. HOTAIR knockdown in cellular systems
mildly upregulated several genes, including the 5
0
-located
HOXD
genes
(
Gupta et al., 2010; Rinn et al., 2007
). However, the direct impact of
HOTAIR over the repression of posterior
HOXD
genes
in vivo
, via the
recruitment of PRC2, could not be observed in mice carrying a deletion
including this LncRNA (
Schorderet & Duboule, 2011
). Furthermore,
HOTAIR binds the same unique site in the
HOXD
cluster in two different
cell types, despite the fact that different
HOXD
genes are repressed in these
cell types (
Chu, Qu, Zhong, Artandi, & Chang, 2011
). Also, this RNA is
poorly conserved between human and mouse (
Schorderet & Duboule,
2011
) and, hence, its function may be in part specific to humans. Altogether,
while this RNA may be of importance for recruiting PRC2 at many geno-
mic loci, it is likely of little relevance in the dynamic recruitment of PRC2 at
the
HOXD
locus during collinearity.
To our knowledge, a systematic mapping of Trithorax components at
mammalian
Hox
clusters during development has not yet been reported.
At these genomic loci, the H3K4me3 mark is deposited by the
COMPASS-likeMLL1 andMLL2 complexes that contain the uniqueMenin
subunit, which distinguishes them from the MLL3 and MLL4 complexes
(
Wang et al., 2009
). Menin has been proposed to target Trithorax complexes
to
Hox
clusters via its association either with LEDGF, a factor required for full
activity of
HOXA9
in cancer cells (
Yokoyama & Cleary, 2008
), or with
ASH2L, a COMPASS-like subunit that binds the
HOXA
cluster in human
foreskin fibroblasts (
Chen et al., 2011; Hughes et al., 2004
). In addition,
