180
Xenopus
At face value, this major increase in H3K4me3 during
blastula stages appears to be at odds with nuclear transfer
experiments that suggest that histone modif cations can
be transmitted through early development. Ever since the
Nobel prize-winning work of Sir John Gurdon, Xenopus has
been an exceedingly powerful system to study the mechanisms underlying the reprogramming of somatic nuclei by
injection into eggs or oocytes. The major implication of the
reprogramming achievements is that the cellular changes of
differentiation are not genetic and irreversible but epigenetic
and reversible in nature (Gurdon 2013). Nuclear reprogramming, however, is not very effcient due to epigenetic barriers
that stabilize cellular identity. These barriers involve epigenetic memory of the donor cell type that manifests itself in
donor cell type-specifc gene expression in nuclear transfer
embryos. This is referred to as ON-memory, whereas OFFmemory refects a failure to activate genes in nuclear transfer
embryos that were repressed in donor cells. Reducing H3K4
methylation (mono-, di-, and tri-methylation) by overexpression of the Kdm5b demethylase in donor cells (endoderm)
reduced ON-memory (endoderm-specifc gene expression in
ectoderm) in nuclear transfer embryos and improved their
development (Hörmanseder et al. 2017). This raises the question of how ON-memory by H3K4me3 is transmitted through
the stages of short cell cycles before the blastula stage. Sperm
nuclei have both H3K4me3 and H3K27me3 at loci where they
also accumulate in blastula embryos (Oikawa et al. 2020).
The presence of H3K4me3 at the same loci in sperm and
blastula chromatin does not necessarily mean that the histone
modifcations are transmitted through the cleavage stages,
because H3K4me3 preferentially accumulates in DNA methylation-free regions (Bogdanovic et al. 2011; van Heeringen
et al. 2014; Hontelez et al. 2015). Conversely, H3K4me3 can
infuence DNA methylation. For example, RNAi of H3K4
methyl-transferase in human embryo carcinoma cells caused
CpG island hyper-methylation on a subset of loci in addition
to a loss of H3K4me3 (Putiri et al. 2014). If such cross-talk is
operational in donor nuclei, it is possible that H3K4me3 indirectly mediates ON-memory in nuclear transfer embryos by
stabilizing DNA methylation-free CpG islands, which subsequently stage preferential H3K4 re-methylation at the blastula stage. Interestingly, in zebrafsh embryos, “placeholder”
nucleosomes with a histone H2A variant and H3K4me1 (but
not H3K4me3) have been found to occupy regions lacking
DNA methylation in cleavage-stage embryos (Murphy et al.
2018). Upon ZGA, many of the placeholder-occupied regions
become active and acquire H3K4me3 (Murphy et al. 2018).
Loss of these placeholder nucleosomes, however, results in
the accumulation of DNA methylation on CpG islands. The
pre-ZGA placeholder nucleosomes have not been studied in
Xenopus but are consistent with acquisition of H3K4me3
and H3K27me3 in pre-existing hypomethylated regions
under the infuence of maternal factors during blastula and
gastrula stages in Xenopus (Bogdanovic et al. 2011; van
Heeringen et al. 2014; Hontelez et al. 2015). This is an attractive model (Figure 11.2B) that can potentially explain H3K4
methylation-dependent ON-memory in nuclear transfer
embryos. Other possibilities include the selective retention
of H3K4me3 on a small subset of genes during the cleavage
stages (exceptions to the dynamics shown in Figure 11.2A)
and differences between the Xenopus species used in these
experiments.
11.4.3. HETEROCHROMATIN MARKS AT REPETITIVE
ELEMENTS AND NEAR GENES
H3K9 methylation (me2, me3) and H4K20me3 decorate the
epigenome in a punctate manner. They co-occupy many
of the transposon sequences that are interspersed in the
genome and are enriched at putative centromeric and telomeric regions (van Kruijsbergen et al. 2017). The modif cations mark a subset of all transposons. Some retrotransposon
families gain H3K9me3 and H4K20me3 between blastula
and gastrula stages. By contrast, specifc DNA transposon
families are enriched for H4K20me3 and the Polycomb mark
H3K27me3 early in development, only to lose these marks
later in development. Interestingly, these DNA transposons
acquire H3K27me3 earlier than gene promoters do. Taking
intra-family diversity as a proxy for age, young transposable elements (high similarity within the sub-family) were
preferentially associated with H4K20me3 and H3K9me3, in
line with a role in repressing these genomic parasites (van
Kruijsbergen et al. 2017). Although repetitive elements may
be the primary drivers of heterochromatinization, the interspersed localization of transposable elements may cause
them to infuence gene regulation. The two major methyl
transferases responsible for H4K20me3, Kmt5b/c (Suv4–
20h1/h2), repress one of the three Oct4 paralogs (pou5f3.2,
also referred to as oct25) in X. laevis (Nicetto et al. 2013).
Knockdown of these enzymes caused defects in neural differentiation. Simultaneous morpholino targeting of pou5f3.2
rescued these defects in neural ectoderm gene expression.
In mouse embryonic stem cells, a similar role for Kmt5b/c
was identifed in downregulation of Oct4 during the exit of
pluripotency (Nicetto et al. 2013). These data show that the
role of H4K20me3 is not restricted to repression in the context of constitutive heterochromatin and that it contributes to
dynamic gene regulation and lineage commitment.
11.5. DNA METHYLATION: REPRESSION
VERSUS MODULATION OF
HISTONE MODIFICATIONS
11.5.1. DNA METHYLATION DYNAMICS
DNA methylation predominantly occurs at the majority of
CG (CpG) dinucleotides in vertebrate genomes. Deamination
of 5-methyl cytosine produces thymine, which explains why
CpG dinucleotides are relatively rare in the genome except
for sequences where CpGs are kept in an unmethylated
state, so-called CpG islands. Unmethylated CpG islands are
found at many promoters, and they are associated with large
H3K27me3-enriched domains (Bogdanovic et al. 2011;
Long et al. 2013).
Xenopus
At face value, this major increase in H3K4me3 during
blastula stages appears to be at odds with nuclear transfer
experiments that suggest that histone modif cations can
be transmitted through early development. Ever since the
Nobel prize-winning work of Sir John Gurdon, Xenopus has
been an exceedingly powerful system to study the mechanisms underlying the reprogramming of somatic nuclei by
injection into eggs or oocytes. The major implication of the
reprogramming achievements is that the cellular changes of
differentiation are not genetic and irreversible but epigenetic
and reversible in nature (Gurdon 2013). Nuclear reprogramming, however, is not very effcient due to epigenetic barriers
that stabilize cellular identity. These barriers involve epigenetic memory of the donor cell type that manifests itself in
donor cell type-specifc gene expression in nuclear transfer
embryos. This is referred to as ON-memory, whereas OFFmemory refects a failure to activate genes in nuclear transfer
embryos that were repressed in donor cells. Reducing H3K4
methylation (mono-, di-, and tri-methylation) by overexpression of the Kdm5b demethylase in donor cells (endoderm)
reduced ON-memory (endoderm-specifc gene expression in
ectoderm) in nuclear transfer embryos and improved their
development (Hörmanseder et al. 2017). This raises the question of how ON-memory by H3K4me3 is transmitted through
the stages of short cell cycles before the blastula stage. Sperm
nuclei have both H3K4me3 and H3K27me3 at loci where they
also accumulate in blastula embryos (Oikawa et al. 2020).
The presence of H3K4me3 at the same loci in sperm and
blastula chromatin does not necessarily mean that the histone
modifcations are transmitted through the cleavage stages,
because H3K4me3 preferentially accumulates in DNA methylation-free regions (Bogdanovic et al. 2011; van Heeringen
et al. 2014; Hontelez et al. 2015). Conversely, H3K4me3 can
infuence DNA methylation. For example, RNAi of H3K4
methyl-transferase in human embryo carcinoma cells caused
CpG island hyper-methylation on a subset of loci in addition
to a loss of H3K4me3 (Putiri et al. 2014). If such cross-talk is
operational in donor nuclei, it is possible that H3K4me3 indirectly mediates ON-memory in nuclear transfer embryos by
stabilizing DNA methylation-free CpG islands, which subsequently stage preferential H3K4 re-methylation at the blastula stage. Interestingly, in zebrafsh embryos, “placeholder”
nucleosomes with a histone H2A variant and H3K4me1 (but
not H3K4me3) have been found to occupy regions lacking
DNA methylation in cleavage-stage embryos (Murphy et al.
2018). Upon ZGA, many of the placeholder-occupied regions
become active and acquire H3K4me3 (Murphy et al. 2018).
Loss of these placeholder nucleosomes, however, results in
the accumulation of DNA methylation on CpG islands. The
pre-ZGA placeholder nucleosomes have not been studied in
Xenopus but are consistent with acquisition of H3K4me3
and H3K27me3 in pre-existing hypomethylated regions
under the infuence of maternal factors during blastula and
gastrula stages in Xenopus (Bogdanovic et al. 2011; van
Heeringen et al. 2014; Hontelez et al. 2015). This is an attractive model (Figure 11.2B) that can potentially explain H3K4
methylation-dependent ON-memory in nuclear transfer
embryos. Other possibilities include the selective retention
of H3K4me3 on a small subset of genes during the cleavage
stages (exceptions to the dynamics shown in Figure 11.2A)
and differences between the Xenopus species used in these
experiments.
11.4.3. HETEROCHROMATIN MARKS AT REPETITIVE
ELEMENTS AND NEAR GENES
H3K9 methylation (me2, me3) and H4K20me3 decorate the
epigenome in a punctate manner. They co-occupy many
of the transposon sequences that are interspersed in the
genome and are enriched at putative centromeric and telomeric regions (van Kruijsbergen et al. 2017). The modif cations mark a subset of all transposons. Some retrotransposon
families gain H3K9me3 and H4K20me3 between blastula
and gastrula stages. By contrast, specifc DNA transposon
families are enriched for H4K20me3 and the Polycomb mark
H3K27me3 early in development, only to lose these marks
later in development. Interestingly, these DNA transposons
acquire H3K27me3 earlier than gene promoters do. Taking
intra-family diversity as a proxy for age, young transposable elements (high similarity within the sub-family) were
preferentially associated with H4K20me3 and H3K9me3, in
line with a role in repressing these genomic parasites (van
Kruijsbergen et al. 2017). Although repetitive elements may
be the primary drivers of heterochromatinization, the interspersed localization of transposable elements may cause
them to infuence gene regulation. The two major methyl
transferases responsible for H4K20me3, Kmt5b/c (Suv4–
20h1/h2), repress one of the three Oct4 paralogs (pou5f3.2,
also referred to as oct25) in X. laevis (Nicetto et al. 2013).
Knockdown of these enzymes caused defects in neural differentiation. Simultaneous morpholino targeting of pou5f3.2
rescued these defects in neural ectoderm gene expression.
In mouse embryonic stem cells, a similar role for Kmt5b/c
was identifed in downregulation of Oct4 during the exit of
pluripotency (Nicetto et al. 2013). These data show that the
role of H4K20me3 is not restricted to repression in the context of constitutive heterochromatin and that it contributes to
dynamic gene regulation and lineage commitment.
11.5. DNA METHYLATION: REPRESSION
VERSUS MODULATION OF
HISTONE MODIFICATIONS
11.5.1. DNA METHYLATION DYNAMICS
DNA methylation predominantly occurs at the majority of
CG (CpG) dinucleotides in vertebrate genomes. Deamination
of 5-methyl cytosine produces thymine, which explains why
CpG dinucleotides are relatively rare in the genome except
for sequences where CpGs are kept in an unmethylated
state, so-called CpG islands. Unmethylated CpG islands are
found at many promoters, and they are associated with large
H3K27me3-enriched domains (Bogdanovic et al. 2011;
Long et al. 2013).