Proteins involved in establishment and maintenance of imprinted methylation marks (original) (raw)

Journal Article

Ruslan Strogantsev is currently a postdoctoral research associate at the laboratory of Prof Anne Ferguson-Smith. His main research interests are aimed at investigating mechanisms of establishment and maintenance of genomic imprints during mouse development with a particular focus on the KRAB zinc finger protein ZFP57. Ruslan originally obtained a BSc degree in Biochemistry in 2005 at the University of Bristol and then went on to do a PhD with Dr Adam West at Glasgow University, where he investigated novel chromatin insulator elements in the human genome.

Search for other works by this author on:

Anne C. Ferguson-Smith

Anne Ferguson-Smith is Professor of Developmental Genetics and Wellcome Senior Investigator in the Department of Physiology Development and Neuroscience at the University of Cambridge. She obtained her PhD from Yale University in 1989. Her postdoctoral research in Cambridge explored the identification, regulation and function of imprinted genes. She obtained an academic faculty position in 1994 and her current research uses genomic imprinting as a model system for analysing the epigenetic control of gene function in normal and abnormal development.

Search for other works by this author on:

Cite

Ruslan Strogantsev, Anne C. Ferguson-Smith, Proteins involved in establishment and maintenance of imprinted methylation marks, Briefings in Functional Genomics, Volume 11, Issue 3, May 2012, Pages 227–239, https://doi.org/10.1093/bfgp/els018
Close

Navbar Search Filter Mobile Enter search term Search

Abstract

Epigenetic phenomena are being increasingly recognized to play key roles in normal mammalian development and disease. This is exemplified by the process of genomic imprinting whereby despite identical DNA sequence, the two parental chromosomes are not equivalent and show either maternal- or paternal-specific expression at a subset of genes in the genome. These patterns are set up by differential DNA methylation marking at the imprinting control regions in male and female germ line. In this review, we discuss the specific mechanisms by which these methyl marks are established and then selectively maintained throughout pre-implantation development. Specifically, we discuss the recent findings of a critical role played by a KRAB zinc-finger protein ZFP57 and its co-factor KAP1/TRIM28 in mediating both processes.

INTRODUCTION

Eutherian mammals and, to a lesser extent, marsupials, possess a remarkable mechanism of gene regulation, termed genomic imprinting [1–4]. While the vast majority of autosomal genes are expressed from both parental chromosomes, imprinted gene expression is generally restricted to either the maternal or a paternal allele. The first evidence of the functional non-equivalence of the two parental genomes came from nuclear transfer experiments, demonstrating that parthenogenic and androgenic embryos, comprised of two maternal and two paternal genomes, respectively, do not survive past mid-gestation [5, 6]. The lethality in parthenogenetic conceptuses; however, can be rescued by restoring the dosage of key imprinted genes resulting in animals surviving to adulthood [7]. Genetic experiments using balanced translocations identified 13 subchromosomal regions, which, when both copies were of the same parental origin, result in phenotypic abnormalities including lethality in some cases [8]. These were subsequently found to harbour distinct imprinted gene clusters.

To date, there are approximately 150 confirmed imprinted genes in the mouse, with many also imprinted in humans, catalogued in (http://www.mousebook.org/catalog.php?catalog=imprinting) and (http://igc.otago.ac.nz). Imprinted genes have been shown to play roles in prenatal growth, lineage specification, neurological development, brain function and postnatal adaptations and metabolic homoeostasis. Similarly, deletions and translocations of imprinted domains in humans results in conditions such as Angelman Syndrome (AS), Prader–Willi syndrome (PWS), Beckwith–Wiedemann syndrome (BWS), Silver–Russell syndrome (SRS), transient neonatal diabetes (TND) and others, reviewed elsewhere [2, 9–12].

Only a few imprinted genes identified to date are singletons, unlinked to other imprinted genes. The majority are organized in clusters ranging in size and gene number, up to around 4 Mb in size. Each cluster typically contains a large imprinted non-coding RNA, and expressed protein coding genes, as well as non-imprinted genes in some clusters. Importantly, imprinted clusters harbour regions that are differentially methylated between the two parental chromosomes (so-called DMRs) that can regulate imprinting across the whole cluster. DMRs can be established either during gametogenesis (gDMRs) or post-fertilization in the soma (sDMRs) with germ line DMRs being the ones critical for imprinting control and somatic DMRs being acquired as a consequence of the gDMR. The majority of known gDMRs (17/21) are maternally methylated after birth in the growing oocyte, and occur at promoters of protein coding genes or large ncRNAs, while four are methylated embryonically in the developing male germ line and are intergenic [11, 13]. Maternal gDMRs essentially fit the criteria for CpG islands (CGIs) with respect to CpG density and GC-richness, while paternal gDMRs fall just short of these criteria. Several studies have demonstrated that gDMRs act as imprinting control regions (ICRs), whereby upon deletion of a particular gDMR, secondary sDMRs fail to establish and imprinting is abolished in the entire cluster associated with it. The ICR function of the gDMR has been demonstrated for Igf2r [14], Igf2-H19 [15], Gnas [16], Snrpn [17], Lit1 (KvDMR1 ) [18] and Dlk1– Dio3 (IG-DMR ) [19] clusters and is probably the case for gDMRs at the other clusters as well. Differential methylation of gDMRs in oocytes and sperm thus provide a stable somatically heritable mark, distinguishing otherwise identical parental chromosomes and providing the basis for imprinted gene expression. In this review, we discuss mechanisms by which this differential methylation is established in the gametes, and is then selectively maintained during the pre-implantation stages of development.

IMPRINTING METHYLATION CYCLE

There are four main stages in the DNA methylation cycle at the imprinted gDMRs: (i) erasure (in advance of re-setting) in primordial germ cells (PGCs), (ii) parental-specific establishment in the male or female germ line, (iii) selective maintenance during pre-implantation re-programming and (iv) general though not exclusive lifelong maintenance post-implantation, except for within the PGCs. The imprinting cycle needs to be viewed in light of global epigenetic changes that occur at each step. Here, we describe and compare DNA methylation dynamics globally and at the imprinted DMRs.

In mice, PGCs arise at E7.25 as a group of approximately 40 cells in the extra-embryonic region, marked by expression of Blimp1 and PGC7/Stella [20–22]. Epigenetic reprogramming of PGCs is initiated at E8.5 during their migration towards genital ridges and is then followed by widespread changes to chromatin architecture, loss of histone marks and global DNA hypomethylation including that of imprinted gDMRs between E11.5 and E12.5 [20, 23]. Loss of DNA methylation at this stage is likely to occur via an active mechanism perhaps linked to base excision repair (BER) [24, 25]. The process has been suggested to initiate by conversion 5-methyl cytosine (5mC) to 5-hydroxy-methylcytosine (5hmC) by the Ten-eleven translocation (TET) family of enzymes, whose expression is upregulated concomitantly with loss of 5mC [24]. DNA-repair driven demethylation has been proposed to directly trigger the chromatin remodelling observed concurrently with loss of 5mC [24–26].

In male mouse embryos, global de novo DNA methylation including of the four known paternally methylated gDMRs begins at E14.5 after re-setting of histone marks [20, 27, 28]. DNA methylation patterns are fully established before birth and are then maintained throughout large and variable numbers of mitotic cell divisions prior entering meiosis. Prolonged DNA methylation maintenance increases the possibility of spontaneous 5mC deamination into thymidine (C>T mutations) [29], which is likely to explain the lower CpG content and lower numbers of paternally (4) versus maternally (17) methylated gDMRs.

In females, however, re-methylation is only initiated after birth in growing oocytes in prophase stage of meiosis I and is mostly complete by Day 20 post-partum [27, 28]. Interestingly, while the paternally methylated gDMRs are methylated over a similar time course, maternal ICRs are methylated asynchronously at different stages of oocyte growth [30]. In addition, as the growing oocyte does not undergo any further divisions prior to fertilization, the risk of C>T mutations is much lower than at the paternally methylated gDMRs.

Fertilization triggers a second wave of epigenetic reprogramming that is characterized by the rapid loss of DNA methylation in the paternal pronucleus prior to the first cell division, followed by passive DNA demethylation of both parental pronuclei up to the 16-cell morula stage [27, 31]. Importantly, the methylation of gDMRs is selectively maintained during this period, allowing correct expression of imprinted genes later on. Several new findings have shed light on the likely mechanisms of the selective active demethylation of the paternal pronucleus and apparent protection of the maternal genome from this process. Work by Gu et al.. [32] and Wossidlo et al.. [33] have shown that the paternal pronucleus is selectively marked by 5hmC modification while 5mC is enriched in the maternal pronucleus. TET3 is the main enzyme expressed in growing oocytes capable of converting 5mC to 5hmC [33]. Embryos generated from TET3 null or siRNA depleted oocytes revealed absence of 5hmC and appearance of 5mC on the paternal pronucleus, while germ cell factor PGC7/Stella was found to be necessary to protect the maternal genome from TET3-mediated hydroxymethylation [32, 33]. Recent evidence suggests that the 5hmC is then lost in a passive replication-dependent manner [34, 35]. However, an active mechanism is also likely to take place as inhibition of the base excision repair components was also shown to increase 5mC levels in the paternal pronucleus [24]. The subsequent passive demethylation of both parental genomes is thought to occur by exclusion of maintenance DNA methyltransferase (Dnmt1) from the nucleus [36]. Given the apparent differences in the demethylation of the two parental pronuclei, paternally and maternally gDMRs may potentially employ different mechanism in order to maintain their methylation.

Global re-methylation of the genome is initiated just prior to the blastocyst stage and occurs predominantly in the inner cell mass (ICM), while the trophoblast lineage remains relatively hypomethylated [27]. It is important at this stage for the unmethylated allele of the imprinted gDMR to be protected against this de novo DNA methylation wave. Finally, the gDMRs are generally faithfully maintained throughout the life span of the organism except for the gonads, where imprints are erased in the PGCs and the cycle re-initiates. Below, we describe major factors identified to be critical for the establishment and maintenance of gDMRs.

FACTORS INVOLVED IN GERM LINE ESTABLISHMENT OF IMPRINTS

De novo DNMTs

Mammals possess two catalytically active DNMTs: DNMT3A and DNMT3B, which can catalyse de novo DNA methylation at distinct and overlapping locations during development [37–39]. Additionally, a homologous, but catalytically inactive co-factor DNMT3L can complex with DNMT3A and DNMT3B modulating their activity and sequence preference [40–45]. Expression analysis of these proteins revealed Dnmt3A and Dnmt3L to be highly expressed in male and female germlines and during early pre-implantation development up until the blastocyst stage [46–48]. In contrast, Dnmt3B expression is mainly zygotic initiating at two cell stage of the embryo and persisting beyond implantation [47]. Reverse genetic approaches have established critical functions for Dnmt3A and Dnmt3L in the establishment of DNA methylation at all imprinted gDMRs as well as at endogenous retroviral insertions (ERVs) such as LINEs (long intersperse elements) and IAPs (intracisternal type A particle) [41–43, 46, 49, 50]. However, the extent and severity of hypomethylation in the mutant animals differed between the male and female germ line.

Oocytes isolated from either Dnmt3A or Dnmt3L null females showed a complete loss of methylation at all maternal gDMRs, but only a mild and variable hypomethylation of LINE-1 and IAP elements [43, 47, 49, 50]. Embryos derived from these females died at ∼E10.5 with loss of imprinting at maternally linked gDMRs; however, repeat methylation was restored to normal levels. In contrast, ablation of these proteins in males led to severe azoospermia and thus complete sterility. Analysis of mutant spermatogonia revealed loss of methylation and upregulated expression of LINE-1 and IAP elements, meiotic failure and subsequent apoptosis of spermatocytes [41, 42]. Interestingly, the effect on the paternally methylated ICRs was less straight forward with different studies revealing different outcomes (Table 1). This variable effect was described as stochastic in nature and independent of the age of analysed spermatocytes perhaps reflecting partial redundancy in methylation mechanisms at these DMRs [43].

Table 1:

Reported hypomethylation as a result of Dnmt3 mutation among the three paternal DMRs investigated (Zdbf2 DMR not examined due to being only recently discovered [13])

References	Hypomethylation
Dnmt3A−/−	Dnmt3L−/−	Dnmt3B−/−
Bourc’his et al. [42]	ND	None	ND
Bourc’his and Bestor [41]	ND	H19-DMR (partial)	ND
Kaneda et al. [49]	H19-DMR, Dlk1–Dio3(IG-DMR)	H19-DMR	None
Kato et al. [43]	H19-DMR, IG-DMR, Rasgrf1 DMR (partial)	H19-DMR, IG-DMR, Rasgrf1 DMR	Rasgrf1 DMR (partial)

References	Hypomethylation
Dnmt3A−/−	Dnmt3L−/−	Dnmt3B−/−
Bourc’his et al. [42]	ND	None	ND
Bourc’his and Bestor [41]	ND	H19-DMR (partial)	ND
Kaneda et al. [49]	H19-DMR, Dlk1–Dio3(IG-DMR)	H19-DMR	None
Kato et al. [43]	H19-DMR, IG-DMR, Rasgrf1 DMR (partial)	H19-DMR, IG-DMR, Rasgrf1 DMR	Rasgrf1 DMR (partial)

Table 1:

Reported hypomethylation as a result of Dnmt3 mutation among the three paternal DMRs investigated (Zdbf2 DMR not examined due to being only recently discovered [13])

References	Hypomethylation
Dnmt3A−/−	Dnmt3L−/−	Dnmt3B−/−
Bourc’his et al. [42]	ND	None	ND
Bourc’his and Bestor [41]	ND	H19-DMR (partial)	ND
Kaneda et al. [49]	H19-DMR, Dlk1–Dio3(IG-DMR)	H19-DMR	None
Kato et al. [43]	H19-DMR, IG-DMR, Rasgrf1 DMR (partial)	H19-DMR, IG-DMR, Rasgrf1 DMR	Rasgrf1 DMR (partial)

References	Hypomethylation
Dnmt3A−/−	Dnmt3L−/−	Dnmt3B−/−
Bourc’his et al. [42]	ND	None	ND
Bourc’his and Bestor [41]	ND	H19-DMR (partial)	ND
Kaneda et al. [49]	H19-DMR, Dlk1–Dio3(IG-DMR)	H19-DMR	None
Kato et al. [43]	H19-DMR, IG-DMR, Rasgrf1 DMR (partial)	H19-DMR, IG-DMR, Rasgrf1 DMR	Rasgrf1 DMR (partial)

In summary, all three analysed paternal DMRs (H19, Dlk1 – Dio3 and Rasgrf1 DMRs) require Dnmt3A and Dnmt3L. In contrast, deletion of Dnmt3B either in oocyte or sperm resulted in no gross hypomethylation of gDMRs or repeat elements. Consequently, the progeny of Dnmt3B−/− males and females developed normally upon zygotic rescue of the protein [43, 49, 50]. Partial hypomethylation was only observed at the Rasgrf1 DMR in the Dnmt3B null sperm, where it may be utilized in the piRNA pathway uniquely employed for DNA methylation at this element [51].

Mechanistic insight into the DNMT3A/DNMT3L critical role in gDMR methylation came from solving the crystal structure of the two complexed together [44, 45]. The study has found that the two exist as a 2:2 heterotetramer with two 3A–3L and one 3A–3A interfaces. The latter interface is required to bind DNA, while the former provides two catalytic sites, capable of methylating CpGs spaced 8–10 bp apart on the opposite DNA strands. This periodicity of CpG nucleotides was suggested to be common in all maternally methylated gDMRs [45]; however this is not imprinting specific [52, 53]. In addition, recent oocyte methylome analysis revealed no differences in CpG periodicity between methylated and unmethylated CGIs in fully grown oocytes [30]. Thus, CpG periodicity is not sufficient to confer specific imprinting methylation in the germline.

Histone modifying enzymes

Despite identification of critical roles played by the DNMT3A–DNMT3L complex in germ line DNA methylation, their targeting to specific DMRs remained unclear. In a search for potential recruiting factors for this complex, Ooi et al. [54] have performed mass spectrometry analysis of DNMT3L interacting partners identifying besides DNMT3A and DNMT3B, four core histone proteins. Further experiments revealed that DNMT3L specifically interacts with the seven N-terminal amino acids of histone H3. Crucially, this interaction is abolished upon histone H3 lysine 4 methylation (H3K4me1,2,3—permissive chromatin mark), but was insensitive to H3K27 or H3K9 methylation (repressive chromatin marks) [54]. Subsequent studies by Ciccone et al. [55] reported that oocytes deficient for H3K4-specific histone lysine demethylase KDM1B fail to methylate four of seven maternal gDMRs examined, reinforcing a notion that unmethylated H3K4 is required to recruit the DNMT3A–DNMT3L complex or H3K4 methylation shields the CpG from methylation. The authors noted that KDM1B expression in growing oocytes coincided with the timing of acquisition of methylation at the four affected DMRs, while others may rely on a different histone lysine demethylase. ChIP-seq analysis of H3K4me3 in growing oocytes has revealed depletion of this mark at all gDMRs as well as other oocyte-methylated CGIs suggesting this to be a general pre-requisite for DNMT3A–DNMT3L-mediated methylation [30].

Analysis of several histone modifications at different gDMR elements in early embryos and embryonic stem (ES) cells has revealed that the normally methylated allele is associated with repressive H3K9 and H4K20 tri-methylation and symmetrical arginine H2A/H4R3 di-methylation, while the unmethylated allele is enriched for permissive H3K4 tri-methylation and H3K9 acetylation marks [56, 57]. In fact, the specific combination of H3K4, H3K9 and H4K20 tri-methyl peaks were found to be highly specific to germline DMR elements in ES cells [58]. The relationship between DNA methylation and histone modification marks remains of interest. For instance embryos derived from Dnmt3L null females, lack both repressive histone marks and DNA methylation at maternal gDMRs [57]. It is, however, likely that the two form a self-reinforcing loop critical for maintenance of both histone and DNA methyl marks.

Transcription traversing DMR elements

Further insights into targeting of DNA methylation to genomic imprints came from work of Chotalia and colleagues [59], who uncovered that methylation of the gDMRs at the Gnas locus is dependent upon transcripts emanating from the upstream Nesp gene. The authors have inserted a β-globin transcription termination cassette upstream of the Gnas XL and Exon 1A promoter DMRs, which upon maternal transmission resulted in hypomethylation of these elements. _Nesp_-associated transcripts emanated from the oocyte-specific promoter and could be detected before or just prior to the onset of DNA methylation at the gDMRs. Finally, the authors point out that the majority of maternally methylated DMRs are located within transcription units and identify alternative upstream start sites in oocytes. In support of a general transcription-driven mechanism of DNA methylation, the above mentioned survey of CGI methylation in oocytes revealed that methylated sequences in general, not just gDMRs, are preferentially intragenic and associated with RNA sequence tags [30]. In addition, some human imprinting disorders are caused by deletion of a promoter encompassing sequences upstream of the gDMR which is concordant with their hypomethylation [60, 61].

How might transcription target DNA methylation to specific elements? One hypothesis is that it creates an open chromatin environment at the site of the H3K4me0 DMR that allows access of the DNMT3A–DNMT3L complex to the DNA [59]. Alternatively, it may play a more direct role either in recruitment of the 3A–3L complex or through setting up histone marks. Indeed it has recently been reported that the PWWP domain of DNMT3A can specifically recognize the H3K36me3 mark, associated with transcriptional elongation [62]. In addition, studies in fission yeast have shown H3K36me3 to recruit histone deacetylase complexes, some of which are conserved in mammals, in order to silence cryptic promoter sites within the gene body [63, 64].

ZFP57

Previous studies have identified the KRAB zinc-finger protein ZFP57 as a factor required for the establishment of germ line DNA methylation at the Snrpn locus, as well as for imprint memory maintenance during pre-implantation epigenetic programming at other gDMRs [65]. As the majority of defects are related to pre-implantation maintenance functions, we discuss ZFP57 functions in that section (see below).

Germ line imprinted methylation put in perspective

It has been thought for a long time that imprinted gDMRs represent a unique set of CGIs in the genome specifically targeted by the methylation machinery in germ cells. This has, however, been recently challenged by Smallwood et al. [30], who systematically surveyed DNA methylation patterns at CGIs during different stages of oocyte growth (5 dpc, 20 dpc GV and MII oocytes) as well as sperm and blastocyst stage embryos. The authors identified more than 1000 CGIs methylated in mature oocytes, while 185 are methylated in sperm (of which 58 are unmethylated in oocyte). Interestingly, methylation of non-imprinted CGIs in oocytes share many features with the maternal gDMRs, such as location within actively transcribed units, depletion of H3K4me3 mark and a requirement for the DNMT3A–DNMT3L complex. However, only a small proportion (<15%) of these CGIs were found to fully retain their methylation in blastocyst including all the maternal gDMRs. Thus imprinted gDMR specification occurs both during germ line establishment and is most likely associated with selective maintenance during pre-implantation reprogramming.

FACTORS INVOLVED IN SELECTIVE MAINTENANCE OF METHYLATION IMPRINTS

As described above, reprogramming in pre-implantation embryos involves rapid loss of 5mC through conversion to 5hmC on the paternal genome and subsequent passive demethylation of both genomes in a replication-dependent manner. Imprinted DMRs must therefore employ processes to resist both mechanisms if they are to act as the epigenetic memory of parental origin.

Maintenance Dnmt1

Maintenance of genomic DNA methylation through mitotic divisions is principally governed by the DNA methyltransferase 1 (DNMT1) enzyme [66]. During S-phase of replication DNMT1 is recruited by its co-factor NP95(UHRF1) to the site of hemi-methylated DNA where it methylates the newly synthesized daughter strand [67–69]. Genetic knockout of Dnmt1 has been found to be lethal with global reduction in DNA methylation levels including to imprinted DMRs leading to loss of mono-allelic expression of imprinted genes [66, 70]. Up until recently, however, it was unclear to what relative extent DNMT1 and the de novo Dnmt1s are responsible for the maintenance of genomic imprints during pre-implantation.

Using a maternal and zygotic double mutant of DNMT3A and DNMT3B, Hirasawa et al. [47] demonstrated that these enzymes are dispensable for maintenance at the paternally methylated _H19_-DMR and Dlk1 – Dio3 (IG-DMR), with partial loss of methylation at the Rasgrf1 DMR, which as discussed above (Table 1) is the only DMR to utilize both of the enzymes. Maintenance of maternal imprints was not assessed due to establishment failure in DNMT3A null eggs (see above). In contrast, just the maternal deletion of Dnmt1 resulted in partial hypomethylation of both maternal and paternal DMRs, while deletion of maternal and zygotic stores resulted in a complete lack of methylation. The authors concluded that DNMT1 alone is sufficient to maintain methylation imprints in pre-implantation at most DMRs.

This finding was inconsistent with the general exclusion of the most abundant oocyte isoform of DNMT1 (DNMT1o) from the nucleus [36, 71]. However, two further studies have demonstrated a low level presence of what was considered the somatic (DNMT1s) isoform of the protein in the nucleus of late oocytes and early embryo [72, 73]. Furthermore, a DNMT1s-specific region of the protein has recently been found to be required for imprint-specific DNA methylation maintenance in ES cells [74]. Questions still remain about the role played by DNMT1o in imprint maintenance. DNMT1o was originally proposed to translocate to the nucleus of the 8-cell embryo and protect imprinted DMRs for single cell division. Specific deletion of this isoform in oocytes resulted in ∼50% hypomethylation at several gDMRs supporting this notion [36, 71]. Several recent studies, did not observe 8-cell nuclear localization, questioning its role at this particular stage [47, 72, 73, 75].

PGC7/Stella

PGC7/Stella was originally identified as a factor expressed early on in PGC specification at E7.25 and then later in growing oocytes of new-born mice [22]. Genetic knockout of PGC7 demonstrated a maternal effect pattern of inheritance: while homozygous null animals could be generated from heterozygous inter-crosses, homozygous embryos derived from PGC7−/− females rarely developed past the 4-cell stage. Even upon the zygotic rescue of PGC7 expression initiated at the two cell stage, very few embryos survived to birth [76, 77]. As the oocytes of PGC7−/− females do not show any apparent developmental defects, the lethality is most likely due to a critical requirement of this protein at the very early stages after fertilization. Indeed analysis of the late pronuclei stage zygotes derived from PGC7 siRNA depleted eggs revealed loss of 5mC and acquisition of 5hmC marks in both parental pronuclei—a mark that is normally restricted to the paternal genome [33].

PGC7 was also shown to be required to maintain DNA methylation at several maternal and paternal gDMRs (Peg1, Peg3, Peg10, H19, Rasgrf1 but not Snrpn, Peg5 or Dlk 1– Dio3 DMRs) in the early zygote [78]. Critically, the loss of methylation at these DMRs occurred before the first cell division and could not be explained by hydroxymethylation of the DNA as the bisulphite sequencing method employed in this analysis detects both 5mC and 5hmC modifications [79]. Hence the 5mC is actively demethylated perhaps by the BER pathway or by another mechanism such as rapid conversion to a state that is read as ‘thymidine’ upon bisulphite sequencing such as 5-carboxylcytosine (5caC)—a downstream product of 5hmC catalysed by TET enzymes [80].

KRAB zinc-finger protein ZFP57

ZFP57 has been originally identified in a gene trap screen as an ES cell-specific factor, downregulated upon differentiation [81]. A subsequently generated knockout mouse has uncovered its critical role in the establishment of DNA methylation at the Snrpn locus in the germ line as well as post-fertilization maintenance of multiple paternal and maternal methylation imprints [65]. In particular, embryos deficient for both the maternally and zygotically derived ZFP57 showed hypomethylation of four out of five imprints; the maternal (Snrpn, Peg1, Peg3 and Nnat/Peg5) gDMRs and the paternally methylated Dlk1 – Dio3 IG-DMR. The Igf2r and H19 DMRs were unaffected in the maternal-zygotic mutant suggesting either a selective influence on DMRs or redundancy with another currently unidentified protein. Hypomethylation of imprinted gDMRs was accompanied by loss of imprinting and resulted in complete lethality at ∼E14.5–E16.5. Presence of maternal ZFP57 stores alone or its zygotic expression could rescue this phenotype indicating that early post-fertilization stages represent the critical window for its function. MacKay et al. [82] independently uncovered a role for human ZFP57 in methylation imprint maintenance. These authors reported that in cases of TND linked to hypomethylation of the ZAC1 (PLAGL1) gDMR as well as variable loss of methylation at other imprinted loci; this was associated with recessive mutations in ZFP57. Maternal-zygotic effect mutations in human ZFP57 have not been described.

Recently, Quenneville et al. [83] have found that exogenously expressed tagged ZFP57 binds all known gDMRs in mouse ES cells, including those unaffected in embryos deficient for zygotic and maternal stores of the protein. The binding occurred on the normally methylated allele consistent with its in vitro preference for the methylated binding motif (TGC5mCGC). Targeted deletion of Zfp57 in cultured ES cells resulted in hypomethylation of several DMRs, which could not be rescued upon expression of exogenous Zfp57 [84]. Thus while necessary to maintain methylation at its target sites, ZFP57 itself is dependent on DNA methylation to bind its targets in the first place. It remains to be determined how ZFP57 participates in establishment of DNA methylation at the Snrpn DMR in female germ line.

The ZFP57 imprint maintenance mechanism: The role of KAP1 repressor complex

ZFP57 belongs to a large family of KRAB zinc-finger proteins, comprising several hundred members [85]. The KRAB domain of ZFP57 is responsible for interaction with the KAP1 (also called TIF1β and TRIM28) co-repressor complex, that serves as a scaffold molecule for recruitment of NuRD histone deacetylase complex, SETDB1 (ESET/KMT1E) histone H3 lysine 9 tri-methylase and a heterochromatin protein HP1 [86–88]. A complex between ZFP57 and KAP1 has recently been shown to interact with all three catalytically active DNMTs providing a likely mechanism for its function in the maintenance of DNA methylation [83, 84].

KAP1 has been found to play critical roles early in development with complete re-absorption of knockout embryos by E8.5, as well as failure of maintenance and self-renewal of mouse ES cells [89–91]. A maternal effect mutation in which KAP1 is depleted from oocytes results in epigenetic variation and imprint instability leading to fetal lethality due to diverse developmental defects. Phenotypes are presumably contributed by KAP1 acting at non-imprinted regions of the genome in addition to some imprints [92]. In ES cells KAP1 is required for the silencing of ERVs, which akin to imprinted regions, are protected from loss of DNA methylation at pre-implantation stages [93]. Similarly, derepression of ERVs was observed upon disruption of SETDB1 gene function in ES cells, while another KRAB zinc-finger protein ZFP809 represses the MLV class of ERVs [94, 95].

There is also mounting evidence, albeit some circumstantial, to further support a role for KAP1 in the regulation of some imprinted genes. For instance, F9 embryonic carcinoma cells, engineered to express a mutant form of KAP1 incapable of the heterochromatin protein HP1 interaction, exhibits loss of H3K9me3 and DNA methylation at the Peg1 gDMR, correlating with derepression of the Peg1 (Mest) gene [96, 97]. Similarly, inducible deletion of Kap1 in ES cells resulted in loss of H3K9me3 at multiple gDMRs, but not DNA methylation changes [83]. Perhaps more convincingly, only wild-type and not KRAB domain deleted ZFP57 could maintain methylation of the genomic imprints when exogenously expressed concomitant with the deletion of the endogenous protein [84]. Thus KAP1 is likely to be the principle co-factor of ZFP57 necessary for its role in the maintenance of genomic imprints.

Other DNA methylation maintenance factors

In addition to the above factors affecting maintenance of multiple imprinted regions, additional factors have been described that appear more specific to individual gDMRs. For instance, RBBP1/ARID4A and RBBP1L1/ARID4B were found to be involved in maintenance of genomic imprints at the Snrpn locus [98]. In another study, RNAi depletion of the methyl-binding domain (MBD3) protein in oocytes or one-cell stage embryos resulted in hypomethylation at the H19 gDMR [99]. Interestingly, both KAP1 and MBD3 can recruit repressive NuRD complex components and may thus provide some functional redundancy at this DMR. It could also explain the fact that while ZFP57 is directly bound at the H19 DMR, it is apparently dispensable for its methylation maintenance in the maternal-zygotic mutant embryos. Rasgrf1 is a gDMR that employs a piRNA pathway, normally utilized to repress retroviral elements, to establish and maintain its methylation [51]. A direct repeat within the Rasgrf1 DMR was found to promote transcription of an adjacent ERV element, that in turn is targeted by piRNA pathway culminating in DNMT3A/DNMT3B recruitment and DNA methylation.

LIFELONG MAINTENANCE OF METHYL IMPRINTS

Following the germ line establishment and preferential maintenance during pre-implantation development of imprinted regions evidence suggests that most gDMRs can retain their DMR status throughout the life of the organism, with the exception of germ cells and some cases of tissue-specific regulation [100, 101]. The two imprinted alleles are also known to carry secondary repressive and active epigenetic marks in order to maintain their DMR status. The unmethylated allele needs to be protected from the de novo wave of DNA methylation occurring concomitant with implantation. By virtue of being CGIs, they are likely to employ the same mechanisms used at other non-imprinted CGIs. Briefly, these might include zinc-finger transcription factors commonly bound at CGIs and a recently reported generic CpG-binding factor CFP1 conferring H3K4me3 mark at active genes, a mark that is depleted in the methylated regions [102–108].

In contrast, the opposite allele is locked in the self reinforcing loop between DNA methylation and repressive histone marks such as H3K9me3, H4K20me3 and H4R3me2s (symmetric arginine di-methylation). There are numerous studies demonstrating biochemical interactions between the histone modifying machinery laying down repressive marks and DNMTs [109–111]. Of direct relevance are recent reports that DNMT3A can be recruited by H4R3me2s at the human β-globin locus and can directly interact with SETDB1, the H3K9-specific histone methyl transferase [112, 113].

SUMMARY AND CONCLUDING REMARKS

DNA methylation has long been recognized as the primary mark distinguishing the two parental alleles at imprinted loci. In doing so it satisfies the four main mechanistic criteria of genomic imprinting: (i) it directly influences transcription of associated genes, (ii) it is heritable through somatic lineages by virtue of Dnmt1 maintenance activity, (iii) it is initiated on the two parental chromosomes at the time when they are separated in different cells (i.e. in the gametes) and (iv) it is erased in PGCs so that appropriate sex-specific methylation is established ensuring imprinting in the next generation. Herein we described the recent progress made in helping to understand the main mechanism governing the processes of establishment and maintenance of genomic imprints.

The establishment process is initiated at different times in male and female germ lines, may have evolved in each for different reasons, but largely involves the same mechanism (Figure 1A). De novo DNMT3A and its related DNMT3L co-factor are universally involved at establishment of DNA methylation at all gDMRs. The targeting of these is likely to be mediated by unmethylated H3K4 and/or transcription traversing the gDMR in germ cells. Many additional DNA elements acquire differential DNA methylation in germ cells, but very few maintain DMR status during genome wide erasure of 5mC marks early after fertilization [30]. The selective maintenance of methylated imprints during this stage is mediated by several factors. PGC7/Stella protects against rapid DNA demethylation at some imprints during the one-cell stage of development and the nuclear retaining of DNMT1s counteracts passive DNA demethylation due to replicative dilution (Figure 1B). The KRAB zinc-finger protein ZFP57 is an essential factor required for maintenance of genomic imprints at multiple gDMRs. ZFP57 can specifically recognize methylated DNA motifs and recruit the KAP1 repressor complex and DNMTs. The precise mechanisms underlying ZFP57 functions however, remain to be determined.

Figure 1:

Diagram summarizing main factors involved in establishment (A) and maintenance (B) of genomic methylated imprints. Permissive (H3K4me3), repressive (H3K9me3) and no (H3K4me0) histone modifications are indicated for each nucleosome drawn. Open and filled lollipops indicate unmethylated and methylated DNA. (A) Establishment of maternal methylation may require transcription from an upstream oocyte promoter (left most H3K4me3 modified nucleosome). Unmethylated H3K4 (right most nucleosomes) at the gDMR is required for binding of Dnmt3a-Dnmt3L tetrameric complex, whilst ZFP57 is also needed at least at the Snrpn locus. (B) Maintenance is mediated by ZFP57 binding to its methylated consensus motif and recruitment of the KAP1 complex as well as de novo and maintenance DNA methyl transferases. SETDB1 tri-methylates histone H3 lysine 9. PGC7/Stella mediates additional protection immediately after fertilization.

Many questions still remain relating to the interplay between the individual factors described here. Of interest is the unexplained phenomenon frequently observed at the Snrpn gDMR. Upon failure to establish methylation in Dnmt3L or ZFP57 null eggs, the methylation is sometimes reacquired in the embryo specifically on the maternal allele [65, 114]. This questions the notion that DNA methylation is the sole primary mark distinguishing the two parental chromosomes and suggests the presence of additional ‘memory’ factors that can confer parental origin identity at least at that locus.

Genomic imprinting results in a subset of genes being expressed from the paternally or maternally inherited chromosome, a process which is necessary for normal mammalian development.
Imprinted gene expression is determined by regions differentially methylated in the male and female germ cells (DMRs).
Germ line acquisition of methyl marks requires transcription through the DMR, unmethylated histone H3 lysine 4, ZFP57–KAP1 and the DNMT3A–DNMT3L de novo methyl transferase complex.
Selective maintenance of DNA methylation at gDMRs requires a maintenance Dnmt1 DNMT1, ZFP57 and KAP1.

FUNDING

Work associated with this review is supported by a grant from the BBSRC and FP7 EpiHealth. UK Biotechnology and Biological Sciences Research Council and European Commission Framework 7 Collaborative Project 278418 EpiHealth (BB/GO20930/1) to R.S.

References

Genomic imprinting in mammals: its life cycle, molecular mechanisms and reprogramming

Cell Res

2011

, vol.

(pg.

466

)

Genomic imprinting: the emergence of an epigenetic paradigm

Nat Rev Genet

2011

, vol.

(pg.

565

)

Mammalian genomic imprinting

Cold Spring Harb Perspect Biol

2011

, vol.

(pg.

)

et al. ,

The evolution of imprinting: chromosomal mapping of orthologues of mammalian imprinted domains in monotreme and marsupial mammals

BMC Evol Biol

2007

, vol.

pg.

157

Completion of mouse embryogenesis requires both the maternal and paternal genomes

Cell

1984

, vol.

(pg.

179

)

Development of reconstituted mouse eggs suggests imprinting of the genome during gametogenesis

Nature

1984

, vol.

308

(pg.

548

)

et al. ,

Birth of parthenogenetic mice that can develop to adulthood

Nature

2004

, vol.

428

(pg.

860

)

Differential activity of maternally and paternally derived chromosome regions in mice

Nature

1985

, vol.

315

(pg.

496

)

Transient neonatal diabetes mellitus type 1

Am J Med Genet Part C Semin Med Genet

2010

, vol.

154C

(pg.

335

)

Mechanisms of imprinting of the Prader-Willi/Angelman region

Am J Med Genet A

2008

, vol.

146A

(pg.

2041

)

Mechanisms regulating imprinted genes in clusters

Curr Opin Cell Biol

2007

, vol.

(pg.

281

)

Imprinting in clusters: lessons from Beckwith-Wiedemann syndrome

Trends Genet (TIG)

1997

, vol.

(pg.

330

)

et al. ,

A tripartite paternally methylated region within the Gpr1-Zdbf2 imprinted domain on mouse chromosome 1 identified by meDIP-on-chip

Nucleic Acids Res

2010

, vol.

(pg.

4929

)

et al. ,

Imprinted expression of the Igf2r gene depends on an intronic CpG island

Nature

1997

, vol.

389

(pg.

745

)

Deletion of the H19 differentially methylated domain results in loss of imprinted expression of H19 and Igf2

Genes Dev

1998

, vol.

(pg.

3693

702

)

et al. ,

Identification of an imprinting control region affecting the expression of all transcripts in the Gnas cluster

Nat Genet

2006

, vol.

(pg.

350

)

et al. ,

A mouse model for Prader-Willi syndrome imprinting-centre mutations

Nat Genet

1998

, vol.

(pg.

)

Regional loss of imprinting and growth deficiency in mice with a targeted deletion of KvDMR1

Nat Genet

2002

, vol.

(pg.

426

)

et al. ,

Asymmetric regulation of imprinting on the maternal and paternal chromosomes at the Dlk1-Gtl2 imprinted cluster on mouse chromosome 12

Nat Genet

2003

, vol.

(pg.

102

)

et al. ,

Germ line, stem cells, and epigenetic reprogramming

Cold Spring Harb Symp Quant Biol

2008

, vol.

(pg.

)

et al. ,

Blimp1 is a critical determinant of the germ cell lineage in mice

Nature

2005

, vol.

436

(pg.

207

)

A molecular programme for the specification of germ cell fate in mice

Nature

2002

, vol.

418

(pg.

293

300

)

et al. ,

Epigenetic reprogramming in mouse primordial germ cells

Mech Develop

2002

, vol.

117

(pg.

)

et al. ,

Genome-wide reprogramming in the mouse germ line entails the base excision repair pathway

Science

2010

, vol.

329

(pg.

)

et al. ,

Chromatin dynamics during epigenetic reprogramming in the mouse germ line

Nature

2008

, vol.

452

(pg.

877

)

Epigenetic reprogramming of mouse germ cells toward totipotency

Cold Spring Harb Symp Quant Biol

2010

, vol.

(pg.

211

)

et al. ,

Dynamic reprogramming of DNA methylation in the early mouse embryo

Dev Biol

2002

, vol.

241

(pg.

172

)

Epigenetic transitions in germ cell development and meiosis

Dev Cell

2010

, vol.

(pg.

675

)

Mutagenic and epigenetic effects of DNA methylation

Mutat Res

1997

, vol.

386

(pg.

107

)

et al. ,

Dynamic CpG island methylation landscape in oocytes and preimplantation embryos

Nat Genet

2011

, vol.

(pg.

)

Methylation levels of maternal and paternal genomes during preimplantation development

Development

1991

, vol.

113

(pg.

119

)

et al. ,

The role of Tet3 DNA dioxygenase in epigenetic reprogramming by oocytes

Nature

2011

, vol.

477

(pg.

606

)

et al. ,

5-Hydroxymethylcytosine in the mammalian zygote is linked with epigenetic reprogramming

Nat Commun

2011

, vol.

pg.

241

Replication-dependent loss of 5-hydroxymethylcytosine in mouse preimplantation embryos

Science

2011

, vol.

334

pg.

194

et al. ,

Generation and replication-dependent dilution of 5fC and 5caC during mouse preimplantation development

Cell Res

2011

, vol.

(pg.

1670

)

et al. ,

Dynamics of Dnmt1 methyltransferase expression and intracellular localization during oogenesis and preimplantation development

Dev Biol

2002

, vol.

245

(pg.

304

)

et al. ,

DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development

Cell

1999

, vol.

(pg.

247

)

The DNMT3 family of mammalian de novo DNA methyltransferases

Prog Mol Biol Transl Sci

2011

, vol.

101

(pg.

255

)

Eukaryotic cytosine methyltransferases

Ann Rev Biochem

2005

, vol.

(pg.

481

514

)

et al. ,

DNMT3L modulates significant and distinct flanking sequence preference for DNA methylation by DNMT3A and DNMT3B in vivo

PLoS Genet

2010

, vol.

pg.

e1001106

Meiotic catastrophe and retrotransposon reactivation in male germ cells lacking Dnmt3L

Nature

2004

, vol.

431

(pg.

)

et al. ,

Dnmt3L and the establishment of maternal genomic imprints

Science

2001

, vol.

294

(pg.

2536

)

et al. ,

Role of the Dnmt3 family in de novo methylation of imprinted and repetitive sequences during male germ cell development in the mouse

Hum Mol Genet

2007

, vol.

(pg.

2272

)

et al. ,

Formation of nucleoprotein filaments by mammalian DNA methyltransferase Dnmt3a in complex with regulator Dnmt3L

Nucleic Acids Res

2008

, vol.

(pg.

6656

)

et al. ,

Structure of Dnmt3a bound to Dnmt3L suggests a model for de novo DNA methylation

Nature

2007

, vol.

449

(pg.

248

)

et al. ,

Dnmt3L cooperates with the Dnmt3 family of de novo DNA methyltransferases to establish maternal imprints in mice

Development

2002

, vol.

129

(pg.

1983

)

et al. ,

Maternal and zygotic Dnmt1 are necessary and sufficient for the maintenance of DNA methylation imprints during preimplantation development

Genes Dev

2008

, vol.

(pg.

1607

)

et al. ,

Coordinate regulation of DNA methyltransferase expression during oogenesis

BMC Dev Biol

2007

, vol.

pg.

et al. ,

Essential role for de novo DNA methyltransferase Dnmt3a in paternal and maternal imprinting

Nature

2004

, vol.

429

(pg.

900

)

et al. ,

Genetic evidence for Dnmt3a-dependent imprinting during oocyte growth obtained by conditional knockout with Zp3-Cre and complete exclusion of Dnmt3b by chimera formation

Genes Cells

2010

, vol.

(pg.

169

)

et al. ,

Role for piRNAs and noncoding RNA in de novo DNA methylation of the imprinted mouse Rasgrf1 locus

Science

2011

, vol.

332

(pg.

848

)

Epigenetics: perceptive enzymes

Nature

2007

, vol.

449

(pg.

148

)

et al. ,

CG dinucleotide periodicities recognized by the Dnmt3a-Dnmt3L complex are distinctive at retroelements and imprinted domains

Mamm Genome

2009

, vol.

(pg.

633

)

et al. ,

DNMT3L connects unmethylated lysine 4 of histone H3 to de novo methylation of DNA

Nature

2007

, vol.

448

(pg.

714

)

et al. ,

KDM1B is a histone H3K4 demethylase required to establish maternal genomic imprints

Nature

2009

, vol.

461

(pg.

415

)

et al. ,

Genome-wide maps of chromatin state in pluripotent and lineage-committed cells

Nature

2007

, vol.

448

(pg.

553

)

et al. ,

Histone methylation is mechanistically linked to DNA methylation at imprinting control regions in mammals

Hum Mol Genet

2009

, vol.

(pg.

3375

)

Distinguishing epigenetic marks of developmental and imprinting regulation

Epigenet Chromatin

2010

, vol.

pg.

et al. ,

Transcription is required for establishment of germline methylation marks at imprinted genes

Genes Dev

2009

, vol.

(pg.

105

)

et al. ,

Deletion of the NESP55 differentially methylated region causes loss of maternal GNAS imprints and pseudohypoparathyroidism type Ib

Nat Genet

2005

, vol.

(pg.

)

et al. ,

A 5-kb imprinting center deletion in a family with Angelman syndrome reduces the shortest region of deletion overlap to 880 bp

Hum Genet

1999

, vol.

105

(pg.

665

)

et al. ,

The Dnmt3a PWWP domain reads histone 3 lysine 36 trimethylation and guides DNA methylation

J Biol Chem

2010

, vol.

285

(pg.

26114

)

et al. ,

Cotranscriptional set2 methylation of histone H3 lysine 36 recruits a repressive Rpd3 complex

Cell

2005

, vol.

123

(pg.

593

605

)

et al. ,

Histone H3 methylation by Set2 directs deacetylation of coding regions by Rpd3S to suppress spurious intragenic transcription

Cell

2005

, vol.

123

(pg.

581

)

et al. ,

A maternal-zygotic effect gene, Zfp57, maintains both maternal and paternal imprints

Dev Cell

2008

, vol.

(pg.

547

)

Targeted mutation of the DNA methyltransferase gene results in embryonic lethality

Cell

1992

, vol.

(pg.

915

)

Dnmt1 structure and function

Prog Mol Biol Transl Sci

2011

, vol.

101

(pg.

221

)

Structure and function of mammalian DNA methyltransferases

Chem Bio Chem

2011

, vol.

(pg.

206

)

Recruitment of Dnmt1 roles of the SRA protein Np95 (Uhrf1) and other factors

Prog Mol Biol Transl Sci

2011

, vol.

101

(pg.

289

310

)

Role for DNA methylation in genomic imprinting

Nature

1993

, vol.

366

(pg.

362

)

et al. ,

Genomic imprinting disrupted by a maternal effect mutation in the Dnmt1 gene

Cell

2001

, vol.

104

(pg.

829

)

et al. ,

Maintenance of genomic methylation patterns during preimplantation development requires the somatic form of DNA methyltransferase 1

Dev Biol

2008

, vol.

313

(pg.

335

)

et al. ,

Preimplantation expression of the somatic form of Dnmt1 suggests a role in the inheritance of genomic imprints

BMC Dev Biol

2008

, vol.

pg.

et al. ,

Identification of a region of the DNMT1 methyltransferase that regulates the maintenance of genomic imprints

Proc Natl Acad Sci USA

2009

, vol.

106

(pg.

20806

)

Safeguarding parental identity: Dnmt1 maintains imprints during epigenetic reprogramming in early embryogenesis

Genes Dev

2008

, vol.

(pg.

1567

)

et al. ,

Dppa3/Pgc7/stella is a maternal factor and is not required for germ cell specification in mice

BMC Dev Biol

2004

, vol.

pg.

et al. ,

Stella is a maternal effect gene required for normal early development in mice

Curr Biol

2003

, vol.

(pg.

2110

)

et al. ,

PGC7/Stella protects against DNA demethylation in early embryogenesis

Nat Cell Biol

2007

, vol.

(pg.

)

et al. ,

The behaviour of 5-hydroxymethylcytosine in bisulfite sequencing

PloS one

2010

, vol.

pg.

e8888

et al. ,

Tet proteins can convert 5-methylcytosine to 5-formylcytosine and 5-carboxylcytosine

Science

2011

, vol.

333

(pg.

1300

)

Identifying genes preferentially expressed in undifferentiated embryonic stem cells

BMC Cell Biol

2007

, vol.

pg.

et al. ,

Hypomethylation of multiple imprinted loci in individuals with transient neonatal diabetes is associated with mutations in ZFP57

Nat Genet

2008

, vol.

(pg.

949

)

et al. ,

In embryonic stem cells, ZFP57/KAP1 recognize a methylated hexanucleotide to affect chromatin and DNA methylation of imprinting control regions

Mol Cell

2011

, vol.

(pg.

361

)

et al. ,

Zinc finger protein ZFP57 requires its co-factor to recruit DNA methyltransferases and maintains DNA methylation imprint in embryonic stem cells via its transcriptional repression domain

J Biol Chem

2012

, vol.

287

(pg.

2107

)

et al. ,

KRAB zinc finger proteins: an analysis of the molecular mechanisms governing their increase in numbers and complexity during evolution

Mol Biol Evol

2002

, vol.

(pg.

2118

)

et al. ,

SETDB1: a novel KAP-1-associated histone H3, lysine 9-specific methyltransferase that contributes to HP1-mediated silencing of euchromatic genes by KRAB zinc-finger proteins

Genes Dev

2002

, vol.

(pg.

919

)

et al. ,

A novel nuclear receptor corepressor complex, N-CoR, contains components of the mammalian SWI/SNF complex and the corepressor KAP-1

J Biol Chem

2000

, vol.

275

(pg.

40463

)

et al. ,

The mammalian heterochromatin protein 1 binds diverse nuclear proteins through a common motif that targets the chromoshadow domain

Biochem Biophys Res Commun

2005

, vol.

331

(pg.

929

)

An RNAi screen of chromatin proteins identifies Tip60-p400 as a regulator of embryonic stem cell identity

Cell

2008

, vol.

134

(pg.

162

)

et al. ,

Mice lacking the transcriptional corepressor TIF1beta are defective in early postimplantation development

Development

2000

, vol.

127

(pg.

2955

)

et al. ,

A genome-wide RNAi screen identifies a new transcriptional module required for self-renewal

Genes Dev

2009

, vol.

(pg.

837

)

et al. ,

Trim28 is required for epigenetic stability during oocyte to embryo transition

Science

2012

, vol.

335

(pg.

1499

502

)

et al. ,

KAP1 controls endogenous retroviruses in embryonic stem cells

Nature

2010

, vol.

463

(pg.

237

)

Embryonic stem cells use ZFP809 to silence retroviral DNAs

Nature

2009

, vol.

458

(pg.

1201

)

et al. ,

Proviral silencing in embryonic stem cells requires the histone methyltransferase ESET

Nature

2010

, vol.

464

(pg.

927

)

et al. ,

Association of the transcriptional corepressor TIF1beta with heterochromatin protein 1 (HP1): an essential role for progression through differentiation

Genes Dev

2004

, vol.

(pg.

2147

)

et al. ,

Disruption of the interaction between transcriptional intermediary factor 1{beta} and heterochromatin protein 1 leads to a switch from DNA hyper- to hypomethylation and H3K9 to H3K27 trimethylation on the MEST promoter correlating with gene reactivation

Mol Biol Cell

2009

, vol.

(pg.

296

305

)

Deficiency of Rbbp1/Arid4a and Rbbp1l1/Arid4b alters epigenetic modifications and suppresses an imprinting defect in the PWS/AS domain

Genes Dev

2006

, vol.

(pg.

2859

)

et al. ,

Maintenance of paternal methylation and repression of the imprinted H19 gene requires MBD3

PLoS Genet

2007

, vol.

pg.

e137

100

et al. ,

Postnatal loss of Dlk1 imprinting in stem cells and niche astrocytes regulates neurogenesis

Nature

2011

, vol.

475

(pg.

381

)

101

et al. ,

Imprinted DNA methylation reprogramming during early mouse embryogenesis at the Gpr1-Zdbf2 locus is linked to long cis-intergenic transcription

FEBS Lett

2012

, vol.

(pg.

)

102

et al. ,

VEZF1 elements mediate protection from DNA methylation

PLoS Genet

2010

, vol.

pg.

e1000804

103

et al. ,

Sp1 elements protect a CpG island from de novo methylation

Nature

1994

, vol.

371

(pg.

435

)

104

et al. ,

Putative zinc finger protein binding sites are over-represented in the boundaries of methylation-resistant CpG islands in the human genome

PloS One

2007

, vol.

pg.

e1184

105

et al. ,

Sp1 sites in the mouse aprt gene promoter are required to prevent methylation of the CpG island

Genes Dev

1994

, vol.

(pg.

2282

)

106

et al. ,

CpG islands influence chromatin structure via the CpG-binding protein Cfp1

Nature

2010

, vol.

464

(pg.

1082

)

107

CpG islands and the regulation of transcription

Genes Dev

2011

, vol.

(pg.

1010

)

108

CpG islands—'a rough guide'

FEBS Lett

2009

, vol.

583

(pg.

1713

)

109

Epigenetic interplay between histone modifications and DNA methylation in gene silencing

Mutat Res

2008

, vol.

659

(pg.

)

110

DNA methylation and histone modifications: teaming up to silence genes

Curr Opin Genet Dev

2005

, vol.

(pg.

490

)

111

Caught in conspiracy: cooperation between DNA methylation and histone H3K9 methylation in the establishment and maintenance of heterochromatin

Biochem Cell Biol

2005

, vol.

(pg.

385

)

112

et al. ,

PRMT5-mediated methylation of histone H4R3 recruits DNMT3A, coupling histone and DNA methylation in gene silencing

Nat Struct Mol Biol

2009

, vol.

(pg.

304

)

113

et al. ,

The histone methyltransferase SETDB1 and the DNA methyltransferase DNMT3A interact directly and localize to promoters silenced in cancer cells

J Biol Chem

2006

, vol.

281

(pg.

19489

500

)

114

et al. ,

Stochastic imprinting in the progeny of Dnmt3L-/- females

Hum Mol Genet

2006

, vol.

(pg.

589

)