Association of modified cytosines and the methylated DNA-binding protein MeCP2 with distinctive structural domains of lampbrush chromatin
We have investigated the association of DNA methylation and proteins interpreting methylation state with the distinctive closed and open chromatin structural domains that are directly observable in the lampbrush chromosomes (LBCs) of amphibian oocytes. To establish the distribution in LBCs of MeCP2, one of the key proteins binding 5-methylcytosine-modified DNA (5mC), we expressed HA-tagged MeCP2 constructs in Xenopus laevis oocytes. Full-length MeCP2 was predominantly targeted to the closed, transcriptionally inactive chromomere domains in a pattern proportional to chromomeric DNA density and consistent with a global role in determining chromatin state. A minor fraction of HA-MeCP2 was also found to associate with a distinctive structural domain, namely a short region at the bases of some of the extended lateral loops. Expression in oocytes of deleted constructs and of point mutants derived from Rett syndrome patients demonstrated that the association of MeCP2 with LBCs was determined by its 5mC-binding domain. We also examined more directly the distribution of 5mC by immunostaining Xenopus and axolotl LBCs and confirmed the pattern suggested by MeCP2 targeting of intense staining of the chromomeres and of some loop bases. In addition, we found in the longer loops of axolotl LBCs that short interstitial regions could also be clearly stained for 5mC. These 5mC regions corresponded precisely to unusual segments of active transcription units from which RNA polymerase II (pol II) and nascent transcripts were simultaneously absent. We also examined by immunostaining the distribution in lampbrush chromatin of the oxidized 5mC derivative, 5-hydroxymethylcytosine (5hmC). Although in general, the pattern resembled that obtained for 5mC, one antibody against 5hmC produced intense staining of restricted chromosomal foci. These foci corresponded to a third type of lampbrush chromatin domain, the transcriptionally active but less extended structures formed by clusters of genes transcribed by pol III. This raises the possibility that 5hmC may play a role in establishing the distinctive patterns of gene repression and activation that characterize specific pol III-transcribed gene families in amphibian genomes.
KeywordsLampbrush chromosomes Methylation 5-hydroxymethylcytosine Oocyte Transcription unit
Human influenza hemagglutinin epitope tag
Histone locus body
- mAb H5
Monoclonal antibody H5
5-Methylcytosine binding domain
Methyl-CpG-binding protein 2
Nuclear factor 7
Nuclear localization signal
Open reading frame
Polymerase chain reaction
- pol II
RNA polymerase II
- pol III
RNA polymerase III
Lampbrush chromosomes (LBCs; see http://www.projects.exeter.ac.uk/lampbrush) essentially comprise a transcriptionally hyperactive form of chromatin in which visible DNA loops bearing arrays of nascent transcripts extend from a chromosome backbone composed of relatively inactive and compact domains, the chromomeres (reviewed by Callan 1986; Morgan 2002; see Macgregor ibid.). Amongst other attributes therefore, LBCs have the useful feature of enabling morphologically discrete chromatin structural domains to be directly observed by light microscopy in the context of individual, identifiable chromosomes. The distinctiveness of lampbrush chromatin is usually interpreted as arising from adaptations associated with the transcriptional hyperactivity required for the construction by amphibian oocytes of a large “pre-embryo” within a single cell (Davidson 1986). However, it also seems likely that lampbrush chromatin possesses characteristics defined by the unique functions of gametes in particular and of germ cells in general, such as meiosis and the acquisition of totipotency. In the latter context, it is significant that amphibian oocytes can profoundly modify the chromatin of transferred somatic nuclei and reprogramme the patterns of gene activity of differentiated cells to more embryonic or stem-cell like states (Jullien et al. 2011). It may be therefore, that lampbrush chromatin has distinctive properties that uniquely reflect the dominant reprogramming potential of its natural environment.
Previous studies have identified some key properties of lampbrush chromatin such as the presence of particular modified core histones (Sommerville et al. 1993; Gaginskaya et al. 2009; Krasikova et al. 2009) and cohesin subunits (Austin et al. 2009), as well as the highly dynamic nature of the association of these proteins with LBCs (Stewart et al. 2006; Austin et al. 2009). However, lampbrush chromatin also appears distinctive with respect to the absence of typical chromatin components such as topoisomerase II (Hock et al. 1996), HMGN proteins (Korner et al. 2003) and condensin subunits (Beenders et al. 2003). Particularly intriguing is the apparent absence from lampbrush chromatin of H1 linker histones (Hock et al. 1993), given the many and varied roles that linker histones have in the structure and function of chromatin (Bustin et al. 2005). Reduced total levels of linker histones are also found in certain somatic cell types, notably mammalian neuronal cells and pluripotent embryonic stem cells (Woodcock et al. 2006). In the former case, recent data suggest that H1 histone may be replaced globally as a chromatin structural protein and general transcriptional repressor by the nuclear protein MeCP2 (Skene et al. 2010).
MeCP2 was first identified (Lewis et al. 1992) by its affinity for DNA containing 5-methylcytosine (5mC), the predominant modified base of vertebrate genomes. Methylation of DNA has long been associated with transcriptional repression and inactive chromatin as well as with the provision of a key mammalian epigenetic system for gene silencing (Klose and Bird 2006). At a chromosomal level, for instance, MeCP2 has been shown to be associated with highly methylated DNA in the pericentromeric regions of mouse metaphase chromosomes (Lewis et al. 1992). Recent genome-scale mapping of methylation has suggested that 5mC may also be found in other contexts, including in gene bodies (Jones 2012). In order to assess whether MeCP2 could be playing a structural role in any of the various distinctive morphological domains of lampbrush chromatin, we have investigated the targeting to LBCs of human MeCP2 constructs expressed in Xenopus laevis oocytes. Since the pattern of MeCP2 localization would be expected also to match the distribution of 5mC in lampbrush chromatin, we have investigated the latter in parallel. We have used the well-characterized antibodies now available to examine by immunostaining the LBCs of Xenopus and/or Ambystoma mexicanum (axolotl) with regard to the distribution of both 5mC and the oxidized 5mC derivative, 5-hydroxymethylcytosine (5hmC), which has recently been shown to exist at high levels in certain cell types (Kriaucionis and Heintz 2009; Tahiliani et al. 2009). We present evidence that MeCP2, 5mC, and 5hmC can all be associated with transcriptionally active structural domains of LBCs as well as with compact, transcriptionally inactive ones.
Materials and methods
Expression of HA-MeCP2 and mutants in Xenopus oocytes
A short sequence coding for the HA tag (YPYDVPDYA) was added in frame at the 5′ end of the open reading frame (ORF) coding for the X. laevis MeCP2. A short 5′ untranslated region sequence, routinely used in our lab for strong expression in frog oocytes (TGAGCCAGAACAATG) was then added by PCR immediately upstream of the HA tag. The resulting MeCP2-HA ORF was then cloned into the pCR®-Blunt II-TOPO® vector (Invitrogen, Carlsbad CA). The three MeCP2 deletion mutants were obtained by polymerase chain reaction (PCR), using the high fidelity Deep VentR® DNA Polymerase (New England BioLabs, Ipswich, MA) and the MeCP2-HA cDNA as a template. In addition, ∆C203-486 and MBD received an SV40 NLS (CCA AAG AAG AAG CGA AAG CTG) in their 3′ end to ensure that the corresponding proteins would enter the nucleus. The amplified DNA fragments were cloned into the pCR®-Blunt II-TOPO® vector (Invitrogen, Carlsbad CA), which contains an SP6 and a T7 promoter localized upstream and downstream of the multiple cloning site, respectively. In vitro transcriptions were performed as described in Beenders et al. 2007. Template DNAs were obtained by linearizing the clones described above downstream of their respective ORF, and either SP6 or T7 polymerases, as required, were used to synthesize capped, sense-strand RNAs. Newly made RNAs were phenol/chloroform extracted, precipitated with ethanol, and resuspended in water. Their concentration was adjusted to ∼1 mg/mL. RNAs were microinjected into the cytoplasm of stage IV–V oocytes under a dissecting microscope (S6 Leica), using a nanoject II (Drummond, Broomal PA) and glass pipettes prepared using a horizontal pipette puller P-97 (Sutter Instrument, Novato CA). Injections had a constant volume of 30 nL (30 ng of RNA).
Preparation and immunostaining of oocyte nuclear spreads
Oocyte nuclei (germinal vesicles; GVs) were manually dissected in GV isolation medium (83 mM KCl, 17 mM NaCl, 6.5 mM Na2HPO4, 3.5 mM KH2PO4, 1 mM MgCl2, 1 mM DTT; pH 7.0–7.2). Spread preparations of X. laevis GV contents were made using the procedure developed by Gall (Gall 1998; Gall and Wu 2010) and a slightly modified procedure used for axolotl GVs (Morgan 2008; see http://www.projects.exeter.ac.uk/lampbrush). After centrifugation to attach the GV contents to a microscope slide, preparations were fixed for a minimum of 15 min and a maximum of 1 h in 2 % paraformaldehyde made up in phosphate-buffered saline (PBS, 137 mM NaCl; 2.7 mM KCl; 10.2 mM Na2HPO4; 1.8 mM KH2PO4; pH 7.4) containing 1 mM MgCl2.
Prior to staining with primary antibodies directed against proteins, fixed preparations were rinsed in PBS and blocked by incubation in 10 % foetal calf serum in PBS for 30 min. The spreads were then incubated for 1 h at room temperature with primary antibodies, rinsed briefly with 10 % foetal calf serum and then incubated for 1 h with secondary antibodies diluted in PBS. Preparations were stained with DAPI (0.5 μg/ml in PBS) and mounted in 50 % glycerol/PBS. When co-staining with antibodies against modified cytosine was required, staining with DAPI and anti-protein antibodies, and sometimes imaging, were carried out prior to the denaturation/depurination pre-treatment used for staining with 5mC and 5hmC antibodies. The latter involved incubating preparations in 2 M HCl at 37 °C for 1 h, followed by four washes in PBS. Preparations were then immunostained as above for modified cytosines.
Primary antibodies were diluted in 10 % foetal calf serum as follows: mAb H5 (Warren et al. 1992) culture supernatant 1:50 dilution; α-RPC15 (Murphy et al. 2002) 1:500; mAb No34 (Murphy et al. 2002) 1:500; α-5mC mAb 33D3 (Active Motif) 1:200; α-5hmC mAb 59.1 (Active Motif) 1:4000; α-5hmC pAb (Active Motif 39769) 1:4000. Secondary antibodies, used at dilutions of 1–5 μg/ml, were Alexa 488-conjugated goat anti-mouse IgG, Alexa 594-conjugated goat anti-mouse IgM or goat anti-rabbit IgG (all Molecular Probes) and Cy2-conjugated goat anti-mouse Fc (gamma) fragment (Jackson Immunoresearch Laboratories). Phase contrast, DIC and fluorescence observations were made with an Olympus BX-60 microscope as described previously (Smith et al. 2003). Images were captured with a Princeton Instruments digital CCD camera (Roper Scientific) and processed using IPLab imaging software and Adobe Photoshop and Illustrator (Adobe Systems Inc).
Cytoplasms and nuclei were hand-isolated and homogenized in GV isolation medium. Crude extracts were centrifuged at 22,000×g at 4 °C for 10 min. Clarified extracts were collected and fractionated on 5–15 % polyacrylamide/SDS gradient gels, using the Mini-PROTEAN 3 Electrophoresis System (Bio-Rad, Hercules, CA). Proteins were transferred to Immobilon membranes (Millipore, Bedford, MA) for 2 h at 40 V in Tris–glycine buffer containing 20 % methanol. Membranes were blocked in 5 % dry milk–PBS for 1 h at room temperature and then incubated for 1 h at room temperature in PBS–Tween (PBST) containing the anti-HA antibody mAb 3 F10 at the concentration of 20 ng/mL. Membranes were washed 3× with 1× PBS containing 0.05 % of Tween 20 (PBS2T) and incubated with the secondary antibody in PBST for 1 h at room temperature. After three washes with PBS2T, detection was performed using the ECF™ kit (GE Healthcare, Fairfield CT) and the FLA3000 fluoroimager (Fuji Medical Systems, Stamford CT).
Targeting of MeCP2 to Xenopus lampbrush chromosomes
Immunolocalization of 5-methylcytosine in Xenopus and axolotl LBCs
Immunolocalization of 5-hydroxymethylcytosine in axolotl LBCs
Overall, we have shown that MeCP2 and the 5mC-modified DNA to which it binds are detected primarily in the compact, transcriptionally inactive structural domains of lampbrush chromatin, the chromomeres. However, we have also found evidence that both may also be associated with the extended, active domains exemplified by the pol II-transcribed lateral loops. Moreover, we have also found that there is an apparent enrichment of the oxidized derivative, 5hmC in a different active domain, namely the less extended one occupied by pol III-transcribed genes. What might be the significance of these patterns in the context of this distinctive type of chromatin that typifies amphibian oocytes?
MeCP2 and 5mC in chromomeres
Although the distribution of the endogenous MeCP2 was not defined here for the lack of a specific anti-Xenopus MeCP2 targeting antibody, the association of newly expressed HA-MeCP2 with LBCs, primarily with the more condensed chromomeric regions, is unlikely to be a fortuitous distribution pattern or based on protein–protein interactions. Rather, the association of MeCP2 with LBCs is likely to reflect a direct interaction with methylated DNA. Several lines of evidence support this conclusion. First, we showed that MBD behaves as an autonomous protein domain that is necessary and sufficient for targeting LBCs. Second, two distinct mutations, R106W and R133C, which have their binding affinities for methylated DNA reduced more than 100-fold (Ballestar et al. 2000), fail to associate with LBCs. It is not clear at the moment why a similar mutation, F155S, has little or no effect on the chromosomal distribution of MeCP2, but this discrepancy may simply reflect the structural difference between LBC chromatin and the DNA used to characterize F155S (Ballestar et al. 2000). Finally, the chromosomal pattern of 5mC and HA-MeCP2 are very similar.
There is abundant experimental evidence in support of a model in which MeCP2 represses transcription through binding to methylated DNA and recruiting histone deacetylases to modify the structure of chromatin (Jones et al. 1998; Nan et al. 1998). However, this is not the only mechanism by which MeCP2 may act as a dynamic transcriptional repressor, as it does not always require deacetylase activity (Nan et al. 1998; Kaludov and Wolffe 2000; Yu et al. 2000). In addition, MeCP2 may also act directly by preventing the formation of the transcriptional pre-initiation complex, likely through direct interaction with the basal transcription factor TFIIB (Kaludov and Wolffe 2000). Predictably then, we find that a newly expressed HA-MeCP2 interacts primarily with the condensed and transcriptionally inactive chromomeric regions of LBCs. Furthermore, the fact that the MBD alone associates with the chromomeres further supports the paradigm of an initial recruitment of MeCP2 followed by repression of transcription. Interestingly, in addition to being a repressor MeCP2 may also play an active role in regulating the architecture of silenced chromatin (Buhrmester et al. 1995; Weitzel et al. 1997; Andrulis et al. 1998; Horike et al. 2005; Singh et al. 2008). In fact, MeCP2 is part of a wider network of competing and co-operating chromatin binding proteins that constantly remodel nucleosomes and modulate the accessibility of chromatin fibres. This network includes HMGs A,B and N, UBF, and histone H1 (Bustin et al. 2005). Recently, MeCP2 was shown to bind methylated linker DNA in nucleosome arrays both in vivo and in vitro (Ishibashi et al. 2008) in a similar way to linker histones.
Regardless of MeCP2 roles within chromomeres, its rapid and efficient association with LBCs (it can be detected on LBCs as early as 1 h after injection of its corresponding transcript) indicates a very dynamic nature of its interactions with chromatin. This is in good agreement with the recent demonstration by fluorescence recovery after photo-bleaching that the interaction of MeCP2 with somatic chromatin is very transient (Kumar et al. 2008). It is noteworthy to recall here that LBCs lack histone H1 and the H1-like B4 protein (Smith et al. 1988; Hock et al. 1993). Thus, another possibility for explaining the efficient chromomeric targeting of HA-MeCP2 is a lack of a binding competition from histone H1 and variants. Interestingly, a HA-tagged histone H1 readily targets the chromomeric regions of LBCs, but not the lateral loops (Joseph Gall, personal communication). Clearly, more work will have to be carried out on LBCs to test whether endogenous MeCP2 is present on LBCs, can be exchanged for newly expressed HA-MeCP2 and whether its binding properties are affected by the presence of histone H1.
The association of MeCP2 and 5mC with lampbrush loops
Perhaps, one of our most exciting findings is the fact that both MeCP2 and 5mC can also be associated with the bases of extended lateral loops. Indeed, this distribution pattern is in very good agreement with the recent finding that more than 60 % of MeCP2 bound promoters in mammalian neurons belong to actively expressed genes (Yasui et al. 2007). In addition, while the expression level of thousands of genes in the hypothalamus of mice is affected by a dysfunction of MeCP2, it was found that more than 80 % of these genes appear to be activated by MeCP2 (Chahrour et al. 2008). Thus, the association of MeCP2 with the base of the loops could represent a population of the protein implicated in transcriptional activation, which would directly support the emerging new model presenting MeCP2 as both a transcriptional repressor and activator.
In some cases in which the long loops of axolotl LBCs allow a detailed observation of their bases, it appears that they are also non-transcribed, so the only distinction with the neighbouring chromomeric chromatin to which the loop is attached seems to be that these methylated regions are extended rather than compacted. Indeed it is feasible that the MeCP2-containing loop chromatin actually represent regions that normally lie on the surface of chromomeric domains and are then subject to disruption during preparation. A well-known feature of LBC preparations is the occasional formation by mechanical disruption of structures known as double-loop bridges, in which a break in the chromomeric axis is spanned by a pair of lateral loops (Callan 1986; see Macgregor ibid.). Perhaps the fibrillar extensions of MeCP2 and 5mC-containing loop bases represent a less severe stage of chromomeric disruption? On the other hand, another distinctive property of chromatin at the points at which loops insert into chromomeres has been described that concerns the presence of acetylated histone H4 (Sommerville et al. 1993). Antibodies against particular acetylated isoforms of H4 were found to produce intensely immunostained foci within the masses of chromomeric chromatin and these foci were generally located at points where the lateral loops were anchored. This type of chromatin mark is thought typically to be associated with active domains and its presence in chromomeres was considered unexpected (Sommerville et al. 1993). It seems possible that the somewhat paradoxical associations of MeCP2 and 5mC with loop bases and of acetylated H4 with chromomeric insertion points indicate that the chromatin states of the border regions between these structural domains are in flux and reflect a particularly dynamic region of lampbrush chromatin.
Although at this stage it is difficult to rule out the presence of 5mC in loop bases as arising from mild disruption of chromomeres during preparation, this does not account for the presence in axolotl LBC loops of interstitial segments containing 5mC. At the least, these regions show unambiguously that methylated DNA can exist in the extended, transcriptionally hyperactive structural domains represented by lateral loops. There seem several possible ways to explain the existence of interstitial methylated regions in loops, which may also be applicable to the similar regions at loop bases. One obvious explanation is that these regions represent methylated non-transcribed spacers separating multiple transcription units within a single loop domain. However, as discussed above in relation to Fig. 7; panel b, this is unlikely to be the general case for interstitial loop methylation. Similarly, the explanation that these regions of localized methylation simply correspond to densely methylated intragenic CpG islands as described within transcribed mammalian genes (Jones 2012), including those of oocytes (Smallwood et al. 2011), also appears incomplete. This is because, surprisingly, the interstitial 5mC regions of loops actually seem to be simultaneously non-transcribed (i.e., they are coincident with pol II- and nascent transcript gaps within transcription units that are otherwise maximally packed with transcription complexes). However, the precise correspondence of 5mC-stained loop regions with these gaps suggests that selective detection could provide a simple technical explanation for localized immunostaining. If loop DNA were entirely methylated then 5mC might only be detectable by immunostaining in gap regions because of the increased antibody accessibility resulting from the absence of pol II transcription complexes that normally mask the underlying DNA template. Detection of DNA in transcript gap regions could also be enhanced because such regions would be expected to form more condensed, nucleosomal chromatin (Scheer 1978, 1987). Interestingly, in recent studies assessing the presence of histone H4 in loop chromatin by immunostaining, discrete regions of loop H4 immunostaining also coincided with gaps in the pol II array and the nascent RNP matrix (Fig. 1 in Austin et al. 2009; Fig. 2 in Gaginskaya et al. 2009). Again, it could be argued that the localized immunostaining arises from the enhanced detectability of H4 in the transcript gaps in which DNP targets are not occluded by densely packed pol II. However, it is difficult to distinguish this kind of technical explanation for interstitial loop methylation from an alternative in vivo one; namely that 5mC occurs only in those internal regions of active genes in which a temporary absence of pol II provides opportunistic access for de novo DNA methylation. To distinguish these explanations, it will be important to assess whether the majority of loop DNA is in fact completely modified such that unmethylated CpG dinucleotides are typically absent from regions downstream of active promoters.
Regardless of loop methylation, a second question relevant to loop function in general concerns the significance of the gaps in the pol II and transcript arrays themselves. Simple technical explanations, such as their representing occasional physical breakage or extreme stretching of loops brought about during preparation, are difficult to sustain because 5mC- and histone H4-immunostaining demonstrate that chromatin is clearly detectable in the gaps. It is arguable that the selective stripping of pol II and attached nascent RNP from highly localized, short regions of loops during preparation could be responsible for the gaps revealed by pol II immunostaining. However, this seems unlikely given the stability of pol II attachment observed repeatedly in many investigations using “Miller-spreading” (Miller and Hamkalo 1972), which subjects LBCs to more disruptive preparative conditions than does preparation for light microscopy. Moreover, the attachment of polymerases to loop DNA in Miller spreads, unlike that of nucleosomes, is resistant even to treatment with detergents such as Sarkosyl (Scheer 1978). It seems more likely that the gaps in pol II immunostaining and nascent transcripts observable within loop transcription units are caused by a stochastic feature(s) of steady-state oocyte transcription. Whatever this feature might be, the gaps in the polymerase stream that result are presumably transient; otherwise, production of full-length transcripts by the affected transcription units would rapidly be shut off. The existence of stochastic fluctuations in transcription rate should be particularly amenable to study using lampbrush loop chromatin because their effects on polymerase packing density can be frozen at an instant in time and analysed directly in single transcription units.
Detection of 5hmC in oocyte chromatin
Immunostaining with two different antibodies against 5hmC provided clear evidence for the presence of this novel modified base in oocyte chromosomal DNA, although the detailed immunostaining patterns were distinctive for each antibody. The 5hmC mAb produced a widely distributed and homogeneous chromosomal staining pattern that was indistinguishable in detail from that of the 5mC mAb. This similarity could indicate a close intermingling of the two types of modified residues within localized regions of the genome and, given the potential role for hmC as an intermediate in the demethylation of 5mC, could reflect the existence of ongoing dynamic processes of de novo methylation and demethylation in the oocyte nucleus. Alternatively, the relatively coarse level of resolution provided by LBC staining might simply mean that at the molecular level 5hmC and 5mC are stably targeted to different DNA sequences even within a given co-stained chromosomal domain such as an individual chromomere, which contains on average about 5 Mb DNA.
By contrast, immunostaining with the 5hmC polyclonal antibody revealed a relatively small number of intensely stained foci on the chromosome axes. We think this distinctive pattern is at least partly explained by the known preference of the polyclonal antibody for multiple clustered 5hmC sites rather than isolated hydroxymethylated sites (Stroud et al. 2011). This explanation is suggested by the other notable feature of the 5hmC foci, namely that many coincide with sites of pol III transcription. Genes transcribed by pol III produce short non-coding RNAs such as 5S rRNA and tRNA (reviewed by White 2011) and typically occur in families composed of tandemly repeated units that in urodeles can be huge in number: in the newt Pleurodeles there are about 50,000 5S genes and 150,000 tRNA genes per haploid genome (Van den Eynde et al. 1989). Even though highly repeated, the 5S and tRNA genes of urodele genomes tend not to be highly dispersed but clustered at small numbers of chromosomal loci—for instance, there are thought to be about 75,000 5S genes at each of four loci in another newt, Notophthalmus viridescens (Pukkila 1975). Since each pol III gene has its own internal promoter, genomic regions containing these gene clusters can exhibit high densities of the dinucleotide CpG, the predominant target site for 5mC, and consequently for production of 5hmC (e.g. there are six CpGs in each 120 bp axolotl 5S gene, Morgan unpublished). Moreover, despite their high degree of transcriptional activity in oocytes, pol III transcriptional domains appear rather compact compared to the extended pol II-transcribed lateral loops. The presence at lampbrush pol III loci of an extraordinarily high spatial density of the sites preferentially recognised by the hmC polyclonal antibody could therefore be responsible for the apparent enrichment for 5hmC at these loci that is suggested by intense immunostaining. On the other hand, we did not detect foci of intense immunostaining with the polyclonal 5hmC antibody near the expected locations of tandemly repeated genes transcribed by other polymerases, namely the chromosomal rRNA genes at the nucleolus organizer and histone genes at HLB loci (Fig. 8; panel b). Since these tandem arrays might also be expected to produce a high chromosomal density of potential 5hmC sites, there could be an added functional significance to the presence of 5hmC in oocyte chromatin that is specific to pol III-transcribed genes.
A distinctive feature of the developmental regulation of amphibian pol III-transcribed genes was initially discovered for the 5S rRNA genes of X. laevis (reviewed by Wolffe and Brown 1988). Xenopus produces two major types of 5S RNA, a somatic type found in most cells and an oocyte-type synthesized and accumulated almost exclusively in oocytes. Each type of 5S RNA is transcribed from distinct genes, one comprising a large multigene family of oocyte 5S genes, and the other a much smaller family of somatic 5S genes. The two gene families exhibit contrasting patterns of expression, with oocyte genes becoming activated in oocytes and repressed in the somatic cells of the developing embryo, while somatic 5S genes are active in all cell types. Similar dual gene systems and patterns of expression have since been found for tRNA genes in Xenopus (Stutz et al. 1989) and the 5S genes of Pleurodeles (Van den Eynde et al. 1989). The establishment and propagation of the repressed state in Xenopus oocyte 5S genes during embryonic development was one of the earliest examples of an epigenetic mechanism to be recognised at a molecular level and factors contributing to the differential regulation of the two 5S gene families have been intensively investigated (Wolffe and Brown 1988; Bouvet et al. 1994; Howe et al. 1998). Xenopus oocyte-type 5S genes are thought to exhibit high levels of 5mC in somatic cells (Miller et al. 1978). This raises the possibility that production of the high-density 5hmC sites we detected in the pol III structural domains of axolotl lampbrush chromatin is part of a mechanism that marks oocyte-type, pol III-transcribed genes for transcriptional activation in oocytes and subsequent developmental repression in somatic cells.
We are grateful to Joseph Gall, Marion Schmidt-Zachmann and Alexey Ruzov for gifts of antibodies and to Andrew Johnson and Ian Mellor for oocytes. We are also very grateful to Natalya Novikova for expert technical assistance. MB is supported by a grant from the National Science Foundation.
This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.
- Beenders B, Jones PL, Bellini M (2007) The tripartite motif of nuclear factor 7 is required for its association with transcriptional units. Mol Cell Biol 27:2615–2624Google Scholar
- Davidson EH (1986) Gene activity in early development, 3rd edn. Academic, OrlandoGoogle Scholar
- Gall J (1998) Spread preparation of Xenopus germinal vesicle contents. In: Spector D, Goldman R, Leinwand L (eds) Cells: a laboratory manual, vol 1. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, pp 52.51–52.54Google Scholar
- Kumar A, Kamboj S, Malone BM, Kudo S, Twiss JL, Czymmek KJ, LaSalle JM, Schanen NC (2008) Analysis of protein domains and Rett syndrome mutations indicate that multiple regions influence chromatin-binding dynamics of the chromatin-associated protein MECP2 in vivo. J Cell Sci 121:1128–1137PubMedCrossRefGoogle Scholar
- Nan X, Ng HH, Johnson CA, et al. (1998) Transcriptional repression by the methyl-CpG-binding protein MeCP2 involves a histone deacetylase complex. Nature 393:386-389Google Scholar
- Scheer U (1987) Structure of lampbrush chromosome loops during different states of transcriptional activity as visualised in the presence of physiological salt concentrations. Biol Cell 59:33–42Google Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.