First published online August 4, 2003
doi: 10.1242/10.1242/dev.00625
Consequences of the depletion of zygotic and embryonic enhancer of zeste 2 during preimplantation mouse development
Sylvia Erhardt1,2,
I-hsin Su3,
Robert Schneider1,
Sheila Barton1,
Andrew J. Bannister1,
Laura Perez-Burgos4,
Thomas Jenuwein4,
Tony Kouzarides1,
Alexander Tarakhovsky3,* and
M. Azim Surani1,
1 Wellcome Trust/Cancer Research UK Institute, University of Cambridge, Tennis
Court Road, Cambridge CB2 1QR, UK
2 LBNL, MS 84-171, 1 Cyclotron Road, Berkeley, CA 94720, USA
3 Laboratory of Lymphocyte Signaling, the Rockefeller University, 1230 York
Avenue, New York, NY 10021, USA
4 IMP, Dr. Bohrgasse 7, A-1030 Vienna, Austria
View larger version (55K):
[in a new window]
|
Fig. 1. Asymmetric localisation of Ezh2/EED and the effect of Ezh2 depletion. (A)
Schematic representation of oocyte maturation and preimplantation development
(see left panels). The second meiotic division commences at fertilisation and
the two parental genomes remain separate as pronuclei until the first cleavage
division. Development can also be initiated by activation of oocytes without
fertilisation, followed by the suppression of second polar body extrusion by
cytochalasin B to generate diploid parthenogenetic embryos. To deplete the
oocytes of maternal Ezh2, transgenic mice with the Ezh2 conditional
alleles, Ezh2F/F, were crossed with ZP3 Cre recombinase
transgenic animals to delete the Ezh2F/F alleles
specifically in the growing oocyte (see far right panel). Zp3 is
expressed prior to the completion of the first meiotic division. Embryos
depleted of maternal Ezh2 (right panels) were compared with those lacking both
the maternal and embryonic Ezh2 (shown in the panel adjacent to the far-right
panel). (B) Schematic depiction of development of zygotes at 0-3, 3-6 and 6-10
hours post fertilisation (hpf) (top line), with the corresponding
immunostaining shown immediately below them. The haploid pronuclei inherited
from the sperm and the oocyte can be distinguished morphologically
(Hogan et al., 1994 ). Male and
female pronuclei, and the second polar body (PB) are marked. All images in
green show antibody staining, red shows DNA staining and yellow shows merged
images. Ezh2 is first associated preferentially with the female pronucleus and
the PB at 0-3 hpf. At  3-6 hpf, Ezh2 can also be detected in the paternal
pronucleus, and by 6-10 hpf, both male and female pronuclei show Ezh2 (white
arrow heads). (C) Depicts a zygote depleted of maternally inherited Ezh2.
Oocytes depleted of maternally inherited Ezh2 and fertilised by wild-type
sperm show Ezh2 by immunostaining at the four-cell stage, indicating
initiation of embryonic transcription of Ezh2. Note that the
pronuclei in Ezh2 depleted zygotes appear to be slightly larger and less
compact than in controls shown in 1B. (D) Eed is also asymmetrically localised
to the female pronucleus (right panels). However, in Ezh2-depleted zygotes,
asymmetric Eed localisation to the female pronucleus is highly reduced to
virtually absent. Thus, asymmetrical localisation of Eed is apparently
dependent on the maternal inheritance of Ezh2.
|
|
View larger version (66K):
[in a new window]
|
Fig. 2. Ezh2 depletion disrupts histone H3 methylation in zygotes and early
embryos. (A) Wild-type zygotes at 0-3 hpf show histone H3-me3-K27
methylation of the maternal pronucleus and of the PB. In the Ezh2-depleted
zygote, H3-K27 methylation is markedly reduced (right panels). The H3-K9 shows
a similar staining to H3-K27 but H3-K4 methylation is unchanged in Ezh2
depleted embryos (not shown). (B) A four-cell embryo with equal staining of
all nuclei for H3-K27. In embryos depleted of maternally inherited Ezh2
(middle panel), the H3-27 is substantially less than in wild-type embryos
(left panel). The H3-K27 levels are restored in these embryos at 16- to
32-cell stage (far right panel). Bottom row, green shows antibody staining;
middle row, DNA staining in red (propidium iodide); top row, merged
images.
|
|
View larger version (42K):
[in a new window]
|
Fig. 3. (A) Cryosection of a wild-type ovary. The ovary is stained with anti-Ezh2
(red; white arrow head) but not the surrounding somatic cells and follicle
cells (orange arrowhead). (B) The effect of depletion of maternally inherited
Ezh2 results in growth retardation even though Ezh2 transcription is
restored at about the four-cell stage from the paternal allele
(Fig. 1C), followed by a
delayed restoration of normal H3-K27 by the 16-cell stage
(Fig. 2B). The graph shows the
average wet weight in grams of pups from Ezh2-depleted oocytes (blue bars)
compared with pups with normal Ezh2 levels (red bars). (C) A summary of
phenotypes with different mutations of Ezh2. Normal maternal
inheritance of Ezh2 and one normal Ezh2 allele is necessary for
normal development [based on O'Carroll et al.
(O'Carroll et al., 2001) and
this study]. Mice from a conventional knockout approach die early during
embryogenesis, despite maternal supply of Ezh2, indicating that the embryonic
Ezh2 transcript is essential for development. Mice without maternal Ezh2
supply but embryonic transcription are viable and fertile, but display a
severe growth retardation until about the weaning age of 4 weeks.
|
|
View larger version (89K):
[in a new window]
|
Fig. 4. Ezh2 and Eed co-localise at the Xi at the blastocyst stage. (A)
Ezh2 distribution at the blastocyst stage of controls (fertilised or
PG+/+). In about 50% of blastocysts from fertilised oocytes, Ezh2
protein (green) is detected as a spot within the nucleus of TE cells (white
arrowhead). The inset shows a higher magnification of a TE cell in a
PG+/+ blastocyst. The accumulation of Ezh2-Eed co-localises with a
DNA dense region, presumably the inactivated X chromosome. As expected,
PG-/- blastocysts do not show Ezh2 staining above background (not
shown). (B) Eed distribution in PG+/+ blastocysts, which is similar
to the staining for Ezh2 shown in A. The inset shows that Eed stays associated
with one chromosome (DNA in red) during M-phase at the blastocyst stage,
presumably the inactive X chromosome. (C) Eed staining in PG-/-
blastocysts shows none of the localisation of Eed seen in PG+/+
blastocysts shown in B. (D) Ezh2 (red) and Eed (green) co-localise in
trophectoderm cells at the blastocyst stage, presumably at the inactivated X
chromosome. The image shows immunostaining of two TE cells from
PG+/+ blastocysts. (E) Eed co-localises with macro-H2A in
trophectoderm cells at the blastocyst stage in PG+/+ blastocysts.
Macro-H2A is highly enriched on the Xi
(Costanzi and Pehrson, 1998),
indicating that Ezh2 and Eed are also associated with the Xi.
|
|
View larger version (85K):
[in a new window]
|
Fig. 5. Histone methylation patterns are disturbed in the absence of Ezh2. (A)
Histone H3-me3-K9 modification is present mainly in pericentric
heterochromatic regions and can be detected as several foci in all cells of
controls (left). The number and intensity of these foci is highly reduced in
most cells of PG-/- embryos (far-right). (B) Histone
H3-me2-K27 accumulates to one region in cells of the TE and has a
stronger overall staining in cells of the presumptive ICM (left). By contrast,
this accumulation is not detectable in PG-/- embryos (far right).
(C) Histone H3-me3-K27 also accumulates in TE cells of control
embryos, which is abolished in PG-/- blastocysts (not shown). Eed
accumulation (in green) co-localises with H3-me3-K27 (in red) in TE
cells of controls, presumably at the Xi, whereas cells of the ICM
show a bright staining throughout the nucleus. DNA is counterstained with
Toto3 (blue). The insets show a TE cell at a higher magnification with a clear
co-localisation of Eed and H3-me3-K27. (D) Fluorescence in-situ
hybridisation (FISH) using a mouse X chromosome specific paint (red) combined
with immunofluorescence shows that H3-me3-K27 (green) is associated
with one X chromosome in a TE cell at the blastocyst stage. DNA is
counterstained with Toto3 (blue). Metaphase chromosomes stained for
H3-me3-K9 and H3-me1-K27 stained very brightly in
control embryos and in PG-/-, owing to a higher accessibility of
antibodies to metaphase chromosomes in early embryos. (E) Eed (green)
accumulation at the Xi does not co-localise with
H3-me3-K9 (red) in TE cells of controls. (F) Metaphase chromosomes
in control embryos stain brightly for H3-me3-K9 in the pericentric
heterochromatin (left, yellow staining). At later stages of development
(9.5dpc), when Ezh2/Eed do not co-localise to the Xi any more (not
shown), H3-me3-K9 specific antibodies stain one entire chromosome
(white arrow head) in addition to pericentric heterochromatin in female
embryos (middle). By contrast, me2- (and me3-) K27 is
highly associated with one chromosome in control embryos at the blastocyst
stage (right, white arrowheads), as well as in 50% of wild-type embryos at the
blastocyst stage (not shown). D-F show TE cells in interphase (D,E) or
metaphase (F) of normal fertilised embryos.
|
|
View larger version (43K):
[in a new window]
|
Fig. 6. The ICM with pluripotent epiblast cells shows a specific and characteristic
histone methylation pattern that is distinct from the pattern seen in TE
cells. (A) The ICM with pluripotent epiblast cells has high me1-,
me2- and me3-K27 methylation levels (pink, middle).
These cells show expression of Oct4-GFP, which is a marker of pluripotent
epiblast cells (green, right). This pattern of histone methylation differs
from that in trophectoderm cells shown in
Fig. 4. Cells that show
clustering of di- and tri-K27 methylation on the Xi do not express
Oct4-GFP (green), which is most obvious in single optical sections (right).
The white arrowheads in the middle and right images indicate an
Oct4-GFP-expressing cell in the ICM with high me3-K27
levels. The orange arrowhead indicates an Oct4-GFP negative cell that is
undergoing X-inactivation. (B) The high levels of H3-K27 methylation of the
ICM is similar to the staining at earlier stages of development, when all
cells are pluripotent. The images show a late morula stage just prior to the
blastocyst stage. (C) Deletion of Ezh2 from primary embryonic
fibroblasts (PEFs) show no detectable effects on growth. Growth of PEFs was
monitored in uninfected, GFP-infected and Cre-infected cells after three (1),
five (2), six plus one passages (3) and eight days (4) in culture after
infection. (F) Immunostaining of control GFP (upper panel) and experimental
Cre (lower panel)-infected PEFS. The Ezh2 protein level was already highly
reduced three days after Cre infection (lower panel) compared with
GFP-infected cells (upper panel) as shown by immunofluorescence (DNA in blue,
Ezh2 in red).
|
|
View larger version (31K):
[in a new window]
|
Fig. 7. Immunoprecipitation (IP) with Ezh2 antibodies shows histone
methyltransferase (HMTase) activity for K9 and K27. (A) HMTase activity was
tested on two different types of peptides with various combinations of lysine
methylation. (B) HMTase assay of Ezh2 IP from ES cell extract on peptides as
substrate. The immunoprecipitate showed highest HMTase activity for
K27-unmethylated peptides and a lower activity for peptides, which were
unmethylated at K9 but fully methylated at K4. (C) Western blot analysis of
IP-bound beads and supernatant (SN) showed that the Ezh2 antibodies
precipitate Ezh2 and that Eed and HDAC1 are bound to this complex.
|
|
View larger version (30K):
[in a new window]
|
Fig. 8. Summary of histone H3 methylation events and the role of Ezh2 at the onset
of mouse development. Shortly after fertilisation, Ezh2/Eed and H3 methylation
are predominantly associated with the female pronucleus, and establish an
epigenetic asymmetry between the two parental genomes. When this asymmetry is
disturbed after depletion of the maternally inherited Ezh2, there is a
long-term effect resulting in severe growth retardation of neonates even when
the embryonic transcription occurs at the four-cell stage. The epigenetic
asymmetry therefore has a significant effect on development and fetal growth.
During cleavage stages, Ezh2/Eed and H3 methylation levels are high. At this
stage, all cells are pluripotent. During differentiation of TE cells, there
are significant changes in the subnuclear localisation and levels of Ezh2/Eed
and H3 histone methylation. The pluripotent epiblast cells that continue to
express Oct4 retain a characteristic and a distinct histone
methylation pattern consistent with the epigenetic plasticity of these
cells.
|
|
© The Company of Biologists Ltd 2003
