Abstract

Maintenance of X-inactivation is achieved through a combination of different repressive mechanisms, thus perpetuating the silencing message through many cell generations. The second human X–Y pseudoautosomal region 2 (PAR2) is a useful model to explore the features and internal relationships of the epigenetic circuits involved in this phenomenon. Recently, we demonstrated that DNA methylation plays an essential role for the maintenance of X- and Y-inactivation of the PAR2 gene SYBL1; here we report that the silencing of the second repressed PAR2 gene, SPRY3, appears to be independent of DNA methylation. In contrast to SYBL1, the inactive X and Y alleles of SPRY3 are not reactivated in cells treated with a DNA methylation inhibitor and in cells from ICF (immunodeficiency, centromeric instability, facial anomalies) syndrome patients, which have mutations in the DNA methyltransferase gene DNMT3B. SPRY3 X- and Y-inactivation is associated with a differential enrichment of repressive histone modifications and the recruitment of Polycomb 2 group proteins compared to the active X allele. Another major factor in SPRY3 repression is late replication; the inactive X and Y alleles of SPRY3 have delayed replication relative to the active X allele, even in ICF syndrome cells where the closely linked SYBL1 gene is reactivated and advanced in replication. The relatively stable maintenance of SPRY3 silencing compared with SYBL1 suggests that genes without CpG islands may be less prone to reactivation than previously thought and that genes with CpG islands require promoter methylation as an additional layer of repression.

INTRODUCTION

X-chromosome inactivation (XCI) is the most studied long-range mechanism of epigenetic silencing, affecting most of the genes of a randomly selected X-chromosome in mammalian females (1). The establishment and maintenance of XCI depends on a number of molecular mechanisms, including XIST RNA coating, Polycomb 2 group protein recruitment, repressive histone modifications and DNA methylation (2).

Several genes escape XCI (3,4), many of which are located on two regions shared by X and Y chromosomes, the pseudoautosomal regions (PARs) (5,6). The PAR regions play a role in driving sex assortment in human male meiosis. Given the recombination between the X and Y in these regions, it was expected that their genes would escape XCI, being expressed also from the Y chromosome. Although true for the short-arm PAR (PAR1) genes, only the most telomeric of the long-arm PAR (PAR2) genes escape inactivation (IL9R and cXYorf1), whereas the centromeric ones (SYBL1 and SPRY3) are X and Y inactivated (7). Such different epigenetic behaviour correlates with the evolutionary history of the region: the inactivated genes arose from a more ancient X-chromosome acquisition than the escaping genes (8).

The unusual finding of Y-inactivation, however, has not been fully explored for mechanistic clues. We previously described a key role for DNA methylation in the allele-specific silencing of SYBL1 in XCI and Y-inactivation (9). We also found that repressive histone modifications and late replication play key roles in this repression (10,11). Although such epigenetic modifications are often inter-related, several studies suggest some independence between methylation, late replication and histone modification as repressive mechanisms (10,12,13).

Here we describe the novel finding of X-inactivation maintenance that is independent of DNA methylation. SPRY3, the proximal PAR2-inactivated gene that lacks a canonical CpG island, is not reactivated at both Xi and Y alleles, following treatment with DNA methylation inhibitors, nor in cells from human syndromes affecting DNA methylation. As in the case of SYBL1, a clear difference in histone methylation is evident between the promoters of active and repressed alleles; in addition, H3K27 histone methyltransferase E(Z)H2 together with the PRC2 Polycomb complex binds the inactive alleles. In contrast to SYBL1, however, late replication and repression of the Y allele was maintained in ICF (immunodeficiency, centromeric instability, facial anomalies) cells where a relatively sharp early–late replication border was seen in the 99 kb region between the escaping and repressed genes. We suggest that genes containing CpG island promoters may be more prone to reactivation than those without, explaining the need for XCI and other silencing systems, such as genomic imprinting, to adopt promoter methylation as an additional layer of silencing.

RESULTS

DNA methylation does not affect the allele-specific expression of the human PAR2 gene SPRY3

SPRY3 is localized at the centromeric end of PAR2, about 99 kb upstream from the gene encoding the synaptobrevin like protein 1, SYBL1 (14). Similar to SYBL1, inactive X (Xi) and Y chromosome alleles of SPRY3 are inactivated, in spite of their identical sequences owing to the relative high frequency of recombination. We began a systematic analysis of the elements controlling SPRY3 transcription in order to gain insights into its allele-specific silencing.

The gene is composed of two exons, with the initial codon residing in exon 2 (Fig. 1). Genomic sequencing and expressed sequence tag analysis identified a long 3′-untranslated region, and 5′-RACE experiments positioned the start site 434 bases upstream of translation initiation. Using a promoter scanning strategy, the 1 kb region (Fig. 1) spanning the transcription start site has been found by in vitro assay using a luciferase reporter to contain the strongest promoter activity. An internal, minimal promoter region from nucleotides −180 to +21, however, is capable of sustaining transcription (De Bonis, unpublished data).

Given the sequence identity between active and inactive alleles, we sought to determine the role of epigenetic mechanisms, in primis DNA methylation, in the maintenance of allele-specific silencing at SPRY3 locus. To determine whether there is a CpG island associated with the gene, we scanned a 20 kb region centered on the promoter region for the presence of CpG clusters (using the EMBOSS CpG software at: http://www.ebi.ac.uk/emboss/cpgplot/). No canonical CpG islands were found, even using low stringency parameters. To determine the role of DNA methylation in SPRY3 regulation, if any, we examined the methylation status of all the clusters surrounding the promoter and first exon by digestion of genomic DNA from somatic cell hybrids with two methylation-sensitive enzymes (HpaII and BssHII), followed by PCR analysis: no signs of differential methylation were seen between expressed and silent alleles (data not shown). In contrast, the adjacent gene, SYBL1, which is subject to allele-specific DNA methylation (11), did show allele-specific amplification of repressed alleles following methylation sensitive genomic restriction (data not shown).

A more thorough analysis was performed by examining the effect on SPRY3 allelic expression of genomic DNA hypomethylation induced either by 5aCdr treatment or by using the genetic model of cell lines from patients affected by ICF syndrome, which are deficient in the DNMT3B cytosine methyltransferase (15).

Somatic cell hybrids retaining the human Xi or the Y chromosomes as a sole human counterpart were treated with 5aCdr and examined at various times after treatment for reactivation of SPRY3 and SYBL1 by RT–PCR, using the escaping gene MIC2 as an internal control (Fig. 2). As previously shown (9), SYBL1 reactivation on both the Xi and Y chromosomes is readily detected under these conditions; however, SPRY3 reactivation could not be detected at any time point. In addition, treatment of a male lymphoblastoid cell line with 5aCdr resulted in a partial reactivation of SYBL1, but no sign of SPRY3 reactivation was detected (data not shown).

ICF syndrome is an autosomal recessive human disease caused by germline mutations affecting DNMT3B gene (1618). Patients show hypomethylation at heterochromatic loci, including the inactive X-chromosome (19). Among many genes on the X-chromosome showing hypomethylation of their associated CpG island, but retaining their monoallelic status of expression, SYBL1 shows hypomethylation of its CpG island and advanced replication time, resulting in biallelic expression (10). Therefore, we checked the allelic expression of SPRY3 gene using two cell lines from ICF male patients, which are polymorphic for a TaqI restriction fragment length polymorphism (RFLP), located at its 3′-UTR. The product, 1374 bp in length, was digested with TaqI endonuclease, which cuts to yield a 919 bp fragment in all samples and a 455 bp fragment that will be cleaved into 337 and 118 bp fragments in the allele with the polymorphic site (Fig. 2B). SPRY3 expression behaves quite differently from SYBL1; in fact, the monoallelic expression of SPRY3 remains unaltered in both cell lines (with the uncut allele being active in PT5 and the cut allele active in PT1).

SPRY3 monoallelic expression is associated with specific histone changes

We next explored the role of histone modification in the allelic regulation of SPRY3 expression. Histone H3 methylation was of particular interest given its important DNA methylation-independent role in X-inactivation (2) and genomic imprinting in the placenta (20,21).

Chromatin immunoprecipitation experiments were performed using antibodies against methylated lysines at positions 4, 9 and 27 of histone H3. An antibody against the acetylated H3 lysine 9 was also used. Experiments on somatic cell hybrids showed specific association of dimethylated H3K9 and di-trimethylated H3K27 with the repressed inactive X and Y alleles, whereas dimethylation at H3K4 and acetylation at H3K9/14 specifically marked the expressed active X allele (Supplementary Material, Fig. S1).

To make the scoring of the results easier when two human sex chromosomes were present, we adopted a strategy for allele-specific PCR, called MAMA–PCR (22). Taking advantage of allele-specific single nucleotide polymorphisms (SNPs) in this way, we examined histone modifications at the SPRY3 promoter (Fig. 1). We developed specific primers to amplify selectively the active X allele (GA primer), or the Xi or Y alleles (GC primer; see Material and Methods). Chromatin immunoprecipitations followed by allele-specific PCR were carried out on male and female cell lines that are polymorphic for the A/C nucleotide at position −585 from +1 (NM_005840) (Fig.1).

We have quantified relative levels of immunoprecipitated and input fractions and have presented all results as histograms corresponding to percentage of input (Fig. 3). Allele-specific ratios are shown for all the antibodies tested. It is worth noting the complementary behaviour of H3K4 and H3K27, which mark the expressed and repressed alleles, respectively.

We also analyzed the enrichment of allele-specific histone modifications in the upstream region of the gene, as well as in the coding region and in the long 3′-UTR using polymorphic RFLPs (Fig. 1). The differential modifications were also found in the body of the SPRY3 gene (Supplementary Material, Fig. S3), but were lost outside the gene (data not shown).

To fully understand the role of histone acetylation in maintaining SPRY3 repression, we treated cells with histone deacetylase (HDACs) inhibitors, such as trichostatin A (TSA) and butyric acid (BA). Given that X-linked genes are not often responding to HDACs inhibitors (23), we have chosen as model system cell from fragile X syndrome male patients. In fact, reactivation of FMR1 gene has been linked to both 5aCdr and HDACs inhibitors treatment, with a synergistic effects between these two drugs (24).

S1 cell line, from a fragile X male patient containing roughly 400–500 fully methylated CGG repeats has been used. This cell line is heterozygous for SPRY3 TaqI RFLP at 3′-UTR (discussed earlier), thus permitting again to check allele-specific expression before and after treatment. In Figure 4, the effect of treatment on SPRY3 and FMR1 genes with various drugs is shown. Both genes escape inactivation upon treatment; but, reactivation of SPRY3 is visible only after TSA or BA treatment (see Materials and Methods), whereas FMR1 is reactivated upon TSA, BA or 5aCdr treatment, as previously reported (24).

Recruitment of PRC2 proteins in the maintenance of SPRY3 X- and Y-inactivation

Recent results highlight the involvement of Polycomb group proteins in the early maintenance of both imprinting and random X-inactivation (2). The binding of PRC2 complex components, such as EZH2 (enhancer of zeste homolog-2, which is the histone methyltransferase specific for the H3K27) and EED1 (embryonic ectoderm development) and SUZ12 (suppressor of zeste-12), have also been demonstrated to take part in the repression of placenta-specific imprinted genes by a mechanism possibly involving histone deacetylases (25).

Interestingly, an association of methylated H3K27 to silenced alleles has been revealed for SYBL1 gene (Supplementary Material, Fig. S2). Preliminary results identify a specific binding of PRC2 complex components to SYBL1 promoter (data not shown; Matarazzo et al., manuscript in preparation).

We investigated the association of the PRC2 complex and HDAC1 with the SPRY3 gene regulatory region by chromatin immunoprecipitation followed by allele-specific PCR in the promoter region (MAMA GA/GC), both in male and female cell lines (Fig. 5). Normalization and allelic ratios are as shown in Figure 3. A specific binding on Xi and Y repressed alleles of male and female lymphoblasts is clearly discernible; the allelic ratios consistently indicate the specific enrichment of PRC2 proteins on repressed alleles.

Analysis of SPRY3 replication time in normal and ICF cell lines

Advanced replication time is a key feature of reactivated alleles of SYBL1 in hypomethylated ICF cells (10). To examine the involvement of replication time in regulation of SPRY3, we analyzed replication time of the locus and its genomic context in lymphoblasts from normal male and ICF patients (Fig. 6). In normal male lymphoblasts, SPRY3 and SYBL1 on the active X replicate in mid-S (peaks at S2 and S3), whereas on the Y, they replicate primarily in late S (S4 and G2/M). This pattern can also be seen at the PAR2 DXS1107 locus (Fig. 1), where there is an informative CA repeat polymorphism in this normal male, with the larger band representing the late replicating Y allele (Fig. 6A).

PT5 ICF male lymphoblasts show the normal expression pattern of only the active X allele of SPRY3, but abnormally express both X and Y alleles of SYBL1 (Fig. 6B) (10). Replication of the Y allele of SPRY3 is normal, but that of SYBL1 is advanced to an active X-like pattern. DXS1107 marker, although not informative, is only 33 kb centromeric of the advanced SYBL1 allele and appears to have a somewhat intermediate pattern of advanced replication for the Y allele (decreased G2/M peak, but strong S4 peak). A second ICF cell line from patient PT1, although not informative for SYBL1 XhoI polymorphism, shows the same replication time advance, as in PT5 (data not shown).

We also analyzed replication features for SPRY3 and other PAR2 markers in fibroblasts from normal male and female individuals and from ICF female fibroblasts (Fig. 6C and D). Replication of all the markers analyzed is generally later on the active X in fibroblasts (S3 and S4 peaks) than in lymphoblasts (S2 and S3 peaks), as seen in normal male fibroblast at SPRY3 locus. The Y allele replicates primarily in S4 with some G2/M, similar to the lymphoblast pattern, but resulting in much less resolution between the replication of the active X and Y in fibroblasts. As in lymphoid cells, the normal fibroblast inactive X and Y alleles replicate very late. This pattern remains essentially unaltered in the female PT4 ICF FB clone, with both alleles showing late replication (S4 peak). In PT4 ICF female fibroblasts, the inactive X allele of DXS1107 appears to be somewhat advanced in replication, as seen by the lower G2/M signal. The strong S4 peak, however, indicates that replication remains in the second half of S-phase in PT4 lymphoblasts, similar to the replication of G6PD and several other inactive X loci that remain silent in this cell line (10). The analysis of a second ICF female fibroblast culture (PT3) also suggests late replication for SPRY3 on the inactive X (data not shown).

DISCUSSION

Soon after XIST-mediated initiation and propagation of silencing, additional epigenetic mechanisms act to maintain XCI, such as Polycomb 2 group protein association, histone modification, late replication time, macro H2A association and DNA methylation (2). As result, a stable maintenance of the repressive signal is ensured. It is thought that all the elements of this complex array of signals are required to maintain the silencing of genes subject to X-inactivation. Failure in the Dnmt1 maintenance methylation system destabilizes embryonic XCI (26). In addition, de novo DNA methylation has been shown to be required for the stable maintenance of XCI, but not for initiation or propagation status (27).

Our studies of histone modification, DNA methylation and replication timing in the PAR2 region under conditions of DNA hypomethylation suggest that DNA methylation is not important for the maintenance of XCI for the non-CpG island gene, SPRY3. Given that it is also inactivated at the identical Y allele (14), these experiments further support the equivalence of the X- and Y-inactivation maintenance mechanisms, as already demonstrated for SYBL1 gene (11). The role of DNA methylation has been recently discussed (28) in the case of the partially escaping X-linked TIMP1 gene. Interestingly, these authors proposed that histone modifications rather than DNA methylation act to maintain X-inactivation.

Histone modifications associated with SPRY3 repression seem to play a major role for the maintenance of repression. This issue has been further confirmed by the reactivation of SPRY3 expression upon HDAC inhibitors treatment (Fig. 4). It is interesting to note that X-linked genes are not prone to reactivation upon HDACs inhibition alone (23). Thus, FMR1 and SPRY3 are examples of this class of reactivated X-linked genes, with or without promoter methylation (see also Supplementary Material, Table SI for further information).

The mechanistic link between DNA and histone methylation reported by us for SYBL1 (11) and by others (29) is absent in SPRY3 regulation, supporting the notion that such modifications may also occur in a DNA methylation-independent way.

Histone methylation in K9 and K27 is common to promoters of X-linked genes, as elsewhere suggested (30). Such modifications could be secondary targets of complexes containing distinct methyltransferases and chromodomains that are able to distinguish between these methyl marks or they could be recognized in combination. The recruitment of the methyltransferase EZH2 and the other members of Polycomb group to SPRY3 seems to support the first hypothesis (Fig. 5). These complexes may act independently to silence transcription either in a redundant manner or synergistically such that neither H3K9 nor H3K27 alone can efficiently repress expression. In our experiments, however, we could not distinguish the specific contribution of each of two methyl marks to the silencing.

Because H3K27 methylation is achieved through the function of the EZH2 histone methyltransferase, this implies the involvement of Polycomb group protein complex PRC2, the mammalian ortholog of Drosophila Esc/E(z) complex, composed of EED1, EZH2 and SUZ12 proteins. This complex has previously been implicated in the maintenance of extra-embryonic X-inactivation (31); further studies associated the same complex in the early maintenance random X-inactivation in the embryo proper (32,33). Transcriptional repression mediated by PRC2 involves histone deacetylase activity, interacting with the complex (25). Here we demonstrated the association of PRC2 and HDAC1 to stably silenced alleles of SPRY3 (Fig. 5). Notably, a similar association has been established also for the other PAR2 inactivated gene, SYBL1 (Matarazzo, unpublished data).

To study the stability of repression for inactive alleles of SPRY3, we treated somatic cell hybrids and human lymphoblastoid cells with the DNA methylation inhibitor 5aCdr and found no reactivation for SPRY3 (Fig. 2) (data not shown), but the nearby SYBL1 gene was readily reactivated, as previously described (9). In addition, inactive alleles of SPRY3 do not escape in ICF syndrome cells that are deficient in X- and Y-linked CpG island methylation (Fig. 2B). These data also match well with those obtained by analyzing SPRY3 expression in uremic/hyperhomocysteinemic (Hcy) patients (13). This metabolic defect causes a plasma increase in the methionine intermediate homocysteine, which leads to a competitive inhibition of cellular methyltransferases (34), resulting in genomic DNA hypomethylation (13,35). SYBL1, as well as H19 and IGF2 allelic expression were profoundly altered in Hcy patients, whereas SPRY3 expression was not (13).

It appears that the DNMT3B deficiency in ICF prevents CpG island promoter methylation associated with XCI (10). Although gene silencing is less stable under these conditions, most genes do not escape XCI, XIST RNA association and repressive histone modifications are normal in ICF for most of the inactive X (36). Spontaneous reactivation only occurred when replication timing was markedly advanced, as seen for SYBL1 and G6PD in specific cell lines (10). Late replication time appears to be an epigenetic silencing mechanism that can be independent of promoter methylation. Conversely, promoter methylation can silence transcription in the absence of late replication in several systems (12,3739).

It is interesting that the abnormal late–early replication transitions found in ICF cells between the Y-linked reactivated SYBL1 and inactivated SPRY3 genes of the PAR2 region can occur over relatively short distances (30–90 kb; Fig. 6) (data not shown). The rather sharp transition suggests that there may be an insulator element between these loci. The CTCF protein appears to function as a methylation-dependent insulator in imprinted domains (40), as well as in X-linked domains that escape XCI (41). Therefore, CTCF is a good candidate factor for separation of the SYBL1SPRY3 domains upon demethylation of recognition elements in the intergenic region.

SPRY3 regulation likely shares silencing features with a large number of genes on the Xi that are known to lack CpG islands (4,42). Although we did not observe SPRY3 reactivation, other non-CpG island genes could be envisioned to reactivate under demethylation conditions as a result of domain-wide effects such as advanced replication time induced by demethylation of extragenic sequences in the domain. Advanced replication timing alone, however, is not likely to be sufficient for reactivation of non-CpG island genes, as repressive histone modifications and other silencing mechanisms must also be removed, just as in the case of CpG-island genes. In fact, we propose that non-CpG island genes are less prone to reactivation than CpG island genes because the basal transcription of such genes is generally lower.

Genes with CpG island promoters are much more likely to be widely expressed than those without (43). In addition, the maintenance of repression for several tissue-specific genes with CpG islands appears to involve island methylation (44). For many tissue-specific genes with CpG islands that remain unmethylated, the absence of tissue-specific transcription factors prevents high levels of transcription; even in these cases, however, there is evidence that the islands may provide low-level transcription in a variety of cell types (45). Under conditions of promoter hypomethylation on the Y and Xi, such basal transcription is likely to promote full reactivation for CpG island-containing genes.

Genomic imprinting and XCI have been linked in several ways (46,47). Both systems evolved from mammalian-specific pathways, both are controlled by epigenetic mechanisms such as CpG island methylation and histone modification and they both have untranslated RNAs that play key roles in silencing. Stable maintenance of imprinting in placenta has recently been shown to be independent of DNA methylation, relying on specific histone modifications and Polycomb complex PRC2 association (20). Considering that XCI and imprinting in marsupials occur via mechanisms independent of DNA methylation, the authors postulated that the ancestral mechanism of allelic inactivation was dependent only on specific histone modifications. Perhaps, this mechanism was retained in the placenta of eutherian species because the placenta is a transient organ, which may tolerate a less stable inactivation. Indeed, XCI in marsupials apparently depends only on histone modifications and shows a greater rate of inactive X-reactivation than in eutherian mammals (48). In contrast, SPRY3 X-inactivation is quite stable under the destabilizing conditions of somatic cell hybridization (Fig. 2) (4), where XIST RNA association with Xi is disrupted (49,50), and spontaneous reactivation of hypomethylated CpG island genes on the ICF Xi occurs (10). In addition, DNA hypomethylation in ICF or uremia patients or by 5aCdr treatment of normal cells or hybrids also failed to reactivate SPRY3. Further studies will be required to understand the evolution of SPRY3-type of XCI maintenance and whether it may be linked to genomic imprinting.

Our PAR2 region data provide an interesting look into the multiple layers of silencing in XCI and suggest that ‘Y-inactivation’ shares many features of XCI. This system will be useful in further elucidating whether or not differences in silencing stability between the CpG island and non-CpG island gene classes reflect differences in intrinsic promoter characteristics and/or developmental establishment or maintenance of repressive mechanisms. Other layers of repression, which need to be studied include such possibilities as differential XIST RNA association in the case of the inactive X and differential subnuclear localization for both Xi- and Y-linked repression (51,52).

MATERIALS AND METHODS

Sequence analysis and SNPs searching

We obtained the human sequence for SPRY3 mRNA (NM_005840) and 5′ and 3′ sequences from the UCSC Genome Bioinformatics Server (http://www.genome.ucsc.edu/) using the May 2005 freeze assembly. For the SNPs analysis, we used the SNP track of the UCSC genome browser and sequences were obtained from dbSNP at NCBI (http://www.ncbi.nlm.nih.gov/SNP).

Cells culture conditions

Epstein–Barr virus-transformed lymphoblast cell line AB was derived from a normal male individual. GM07693 (NIGMS) is a fibroblast cell line from a female with highly skewed X-inactivation (>95:5). The ICF cell lines analyzed in this study, all derived from cases with classical ICF features, include lymphoblast cells from PT1 and PT5 males and PT4 female, untransformed PT3 fibroblasts, PT4 fibroblast and h-TERT-transformed PT3 (10,36). Lymphoblasts and fibroblasts were cultured as previously described (10,11).

Treatments with 5aCdr, TSA and BA

BA (Sigma, St Louis, MO, USA) was resuspended in sterile water to a concentration of 500 mm and stored at −80°C in aliquots. TSA (Sigma) was resuspended in ethanol to a concentration of 1 mg/ml stock and stored at −20°C in small aliquots. Finally, a 1.75 mm stock solution of 5aCdr (Sigma) was prepared in sterile water and stored at −80°C in aliquots.

For 5aCdr treatment, cells were counted, split and seeded to an initial concentration of 2.5–3×105 cells/ml in a total volume of 10 ml per flask. A daily dose of 1 µm of 5aCdr was added to the flasks and was thoroughly resuspended, whereas a control flask was left untreated. Cells were harvested after 120 h treatment. The medium was changed every 48 h. RNA was extracted with RNAqueous kit (Ambion), according to manufacturer's protocol. Single-strand cDNA were synthesized with SuperScriptII (Invitrogen) according to manufacturer's protocol.

For TSA and BA treatments, cells were counted, split and seeded to the initial concentration of 2.5–3×105 cells/ml in a total volume of 10 ml per flask. A single dose of either BA or TSA was added to the flasks and was thoroughly resuspended. A control flask for each patient cell line was left untreated or received a mock treatment with a comparable volume of ethanol in the case of the TSA treatment. Cells were harvested after 48 h and the RNA extracted as above.

To assess the expression status of SPRY3 gene after 5aCdr treatment in hybrid cells, we performed RT–PCR using the following primers: F7 5′-ATGCCACTACAGGTTTGACTCG-3′ and RT2 5′-GACACTGCAGGAACAAAAGAGA-3′. PCR amplification for the control genes SYBL1 and MIC2 was performed using primers described previously (9).

PCR amplification of SPRY3 has been carried out after TSA, BA and 5aCdr treatment of cells derived from fragile-X S1 patient with primers F5 5′-GTGGGCCTAGTTCTTGCTTT-3′ and RT2 (described above). Here, PCR product, 362 bp in length, was digested with TaqI restriction enzyme, giving the following fragments, 264 and 98 bp: primers described previously have been used to amplify the control gene FMR1 (9). GAPDH has been amplified to normalize cDNA templates with the following primers: GAPDH-F 5′-ACATGTTCCAATATGATT-3′ and GAPDH-R 5′-TGGACTCCACGACGTACTCAG-3′.

To assess the allelic SPRY3 expression status in ICF cell lines, cDNAs derived from males polymorphic for a TaqI RFLP in the 3′-UTR of the gene (G/T; dbSNP rs306889) were amplified with primers F7 and RT2 (described earlier). PCR amplification was done at 64°C and the product, 1374 bp in length, was digested with TaqI endonuclease, which cuts in two different positions, nt 919 and nt 1256, being polymorphic only in the latter. As a result, the digestion of every allele will produce a 919 bp fragment. The remaining 455 bp band will be cleaved into 337 and 118 bp fragments in the allele with the polymorphic site.

ChIP assay, allele-specific PCR (MAMA) and 5′–3′ allele-specific modifications

Chromatin immunoprecipitation has been performed as previously described (11). Polyclonal antisera have been used, according to the manufacturer's instructions (Upstate), to detect the following modifications of histone H3 tails at PAR2 loci: acetylation of lysine 9 and 14, dimethylation of lysine 4 and 9, di-trimethylation of lysine 27. Moreover, commercial antisera against Eed, Ezh2 and Suz12 were used to analyze PAR2 associations (Abcam anti-Ezh2, Upstate anti-Eed, Abcam anti-Suz12). A rabbit IgG antiserum (Upstate, Inc.) was used as a negative control.

After immunoprecipitation, allele-specific semi-quantitative PCR was performed as described previously (22,53). Briefly, using MAMA technology, we have designed two primer pairs spanning the known rs306882 SNP (C/A), the first one annealing specifically to the ‘A’ allele and the other one to the ‘C’ allele. The primers and annealing temperatures were MAMA-GA: 5′-AGCTACAGCTCCGGGAAAGga-3′/MAMA-rev: 5′-GTTGGCAGTGTGTTCAGCTTAG-3′ at 62°C and MAMA-GC: 5′-AGCTACAGCTCCGGGAAAGgc-3′/MAMA-rev at 67°C. For semi-quantitative PCRs we quantified, after electrophoretic separation, the intensity of each immunoprecipitated and input band using QuantityOne (BioRad) software. We performed each immunoprecipitation in triplicate. All results are presented as percent of precipitation, obtained by dividing averaged values of the precipitated sample by the averaged values of the corresponding input chromatin. Histone modifications in the coding region were examined with the F5 and RT3 primers (RT3 5′-CTGAAGAGGAGTGGGATAGATA-3′), annealing at 60°C.

For chromatin immunoprecipitation on SYBL1 promoter, we used di-trimethylated H3K27 (D. Reinberg) and acetylated H3K9 (Upstate, Inc.). Amplification has been performed as in Matarazzo et al. (11).

Replication timing analysis

Replication time was determined by analysis of newly replicated DNA labeled with bromodeoxyuridine (BrdU), which was isolated from cells sorted into specific cell cycle stage by flow cytometry as previously described (54,55). The BrDU-labeled DNA was isolated with an anti-BrdU antibody, and the newly replicated DNA fractions were then analyzed for the presence of specific loci by semi-quantitative PCR (54,55).

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG Online.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge M. Cuccurese and G. Chirico for help in the initial phases of the work, R. Feil and S. Gartler for comments, D. Reinberg for the gift of H3K27 di-trimethylated antibody. Also, we acknowledge the help of Genome Browser available at UCSC (http://genome.ucsc.edu). Research in MD'E laboratory is sponsored by Telethon grant GGP02308 and EEC contract no. LSHB-CT-2004-503243, (SAFE NoE). R.S.H. was supported by grant HD16659 from the NIH. G.N. and P.C. gratefully acknowledge Telethon for funding.

Conflict of Interest statement. None declared.

The authors wish it to be known that, in their opinion, the first three authors should be regarded as joint First Authors.

Figure 1. Map of PAR2 region. Position of the genes identified, represented as black bold arrows, is reported in a total length of 330 kb. SYBL1 and SPRY3 genes are separated by ~99 kb. A magnification of SPRY3 gene is shown. Patterned and large black bars indicate the promoter and the coding region, respectively, and horizontal and vertical lines represent primer pairs used and SNPs, respectively. Shown also is the polymorphic site, where specific MAMA primers are located in the promoter region, and RFLPs used for ChIP analysis.

Figure 1. Map of PAR2 region. Position of the genes identified, represented as black bold arrows, is reported in a total length of 330 kb. SYBL1 and SPRY3 genes are separated by ~99 kb. A magnification of SPRY3 gene is shown. Patterned and large black bars indicate the promoter and the coding region, respectively, and horizontal and vertical lines represent primer pairs used and SNPs, respectively. Shown also is the polymorphic site, where specific MAMA primers are located in the promoter region, and RFLPs used for ChIP analysis.

Figure 2. Reactivation of SPRY3. (A) Effect of 5aCdr treatment, which reactivates SYBL1, does not reactivate SPRY3 expression in inactive X (Xi) and Y hybrids. MIC2 expression has been used as control. Hours of treatment are indicated. (B) SPRY3 does not escape from silencing in ICF cells. Allele-specific expression analysis by RT–PCR in lymphoblastoid cells from two ICF patients (PT5 and PT1) heterozygous for TaqI RFLP. The product of the primers F7/RT2, 1374 bp in length, was digested with TaqI endonuclease, which cuts in two different positions, giving a 919 bp fragment in all samples and a 455 bp fragment that will be cleaved into 337 and 118 bp fragments in the allele with the polymorphic restriction site (see the map for details).

Figure 2. Reactivation of SPRY3. (A) Effect of 5aCdr treatment, which reactivates SYBL1, does not reactivate SPRY3 expression in inactive X (Xi) and Y hybrids. MIC2 expression has been used as control. Hours of treatment are indicated. (B) SPRY3 does not escape from silencing in ICF cells. Allele-specific expression analysis by RT–PCR in lymphoblastoid cells from two ICF patients (PT5 and PT1) heterozygous for TaqI RFLP. The product of the primers F7/RT2, 1374 bp in length, was digested with TaqI endonuclease, which cuts in two different positions, giving a 919 bp fragment in all samples and a 455 bp fragment that will be cleaved into 337 and 118 bp fragments in the allele with the polymorphic restriction site (see the map for details).

Figure 3. Different histone modifications mark Xi/Y alleles repression and Xa allele expression. Allele-specific semi-quantitative PCR of the promoter region (MAMA GA/GC) was performed after chromatin immunoprecipitation on lymphoblastoid cell line AB (male) and GM07693 (skewed female fibroblast). Following electrophoretic separation of bands, the relative IP and input intensities of allele-specific bands were quantified. The relative level of immunoprecipitation is presented in the bar graph as a fraction of the input, obtained by dividing averaged values of the precipitated material by the averaged values of the corresponding input chromatin. Dimethyl-K4 and acetyl-K9/14 clearly mark the expressed allele; di/trimethyl-K27 and dimethyl-K9 mark repressed ones. For each antibody immunoprecipitation, we used an IgG rabbit antiserum as a control for specificity. Numbers reported in the table under the graphics correspond to the allelic ratio (Xa/Xi and X/Y).

Figure 3. Different histone modifications mark Xi/Y alleles repression and Xa allele expression. Allele-specific semi-quantitative PCR of the promoter region (MAMA GA/GC) was performed after chromatin immunoprecipitation on lymphoblastoid cell line AB (male) and GM07693 (skewed female fibroblast). Following electrophoretic separation of bands, the relative IP and input intensities of allele-specific bands were quantified. The relative level of immunoprecipitation is presented in the bar graph as a fraction of the input, obtained by dividing averaged values of the precipitated material by the averaged values of the corresponding input chromatin. Dimethyl-K4 and acetyl-K9/14 clearly mark the expressed allele; di/trimethyl-K27 and dimethyl-K9 mark repressed ones. For each antibody immunoprecipitation, we used an IgG rabbit antiserum as a control for specificity. Numbers reported in the table under the graphics correspond to the allelic ratio (Xa/Xi and X/Y).

Figure 4. Treatment with HDACs inhibitors (TSA, BA) reactivates SPRY3 Y repressed allele. S1 fragile-X cell line, heterozygous for SPRY3 TaqI RFLP, were treated with different drugs altering histone acetylation (TSA, BA) or DNA methylation (5aCdr). Top: RNA from treated cells has been extracted then reverse transcribed and digested with TaqI (see Materials and Methods). Lanes 1 and 2, untreated (U) sample; lanes 3 and 4, S1 TSA treated; lanes 5 and 6, BA treated; lanes 7 and 8, 5aCdr treated. RT+ and RT− samples are reported. Note the HDACs treatments bring to SPRY3 biallelic expression; conversely, after 5aCdr treatment, no reactivation and monoallelic expression are visible. Below: same cDNAs from treated and untreated cells amplified by using FMR1 primers (24). Note a slight reactivation upon treatment of BA and TSA, robust after 5aCdr treatment. Normalization with GAPDH primers has been carried out; WT, normal control sample.

Figure 4. Treatment with HDACs inhibitors (TSA, BA) reactivates SPRY3 Y repressed allele. S1 fragile-X cell line, heterozygous for SPRY3 TaqI RFLP, were treated with different drugs altering histone acetylation (TSA, BA) or DNA methylation (5aCdr). Top: RNA from treated cells has been extracted then reverse transcribed and digested with TaqI (see Materials and Methods). Lanes 1 and 2, untreated (U) sample; lanes 3 and 4, S1 TSA treated; lanes 5 and 6, BA treated; lanes 7 and 8, 5aCdr treated. RT+ and RT− samples are reported. Note the HDACs treatments bring to SPRY3 biallelic expression; conversely, after 5aCdr treatment, no reactivation and monoallelic expression are visible. Below: same cDNAs from treated and untreated cells amplified by using FMR1 primers (24). Note a slight reactivation upon treatment of BA and TSA, robust after 5aCdr treatment. Normalization with GAPDH primers has been carried out; WT, normal control sample.

Figure 5.In vivo binding of EZH2, SUZ12, EED1 and HDACI on the silenced Xi/Y SPRY3 promoters. Chromatin immunoprecipitation on lymphoblastoid cell line AB (male) and GM07693 (skewed female fibroblast) was followed by allele-specific PCR (MAMA) as described in Figure 3. Histograms show specific enrichment of these proteins on repressed alleles.

Figure 5.In vivo binding of EZH2, SUZ12, EED1 and HDACI on the silenced Xi/Y SPRY3 promoters. Chromatin immunoprecipitation on lymphoblastoid cell line AB (male) and GM07693 (skewed female fibroblast) was followed by allele-specific PCR (MAMA) as described in Figure 3. Histograms show specific enrichment of these proteins on repressed alleles.

Figure 6. Replication timing in the PAR2 region, present on both the X and Y chromosomes, was examined using a BrdU incorporation method. The X and Y alleles of SPRY3 and SYBL1 can be distinguished in informative males using polymorphic restriction sites; the chromosomal assignment of alleles can be done in males using RT–PCR analysis because the Y alleles of these genes are usually silent. In females, we cannot distinguish the active from the inactive allele in uncloned populations. LB, lymphoblasts.

Figure 6. Replication timing in the PAR2 region, present on both the X and Y chromosomes, was examined using a BrdU incorporation method. The X and Y alleles of SPRY3 and SYBL1 can be distinguished in informative males using polymorphic restriction sites; the chromosomal assignment of alleles can be done in males using RT–PCR analysis because the Y alleles of these genes are usually silent. In females, we cannot distinguish the active from the inactive allele in uncloned populations. LB, lymphoblasts.

Table 1.
Mechanism of reactivation Genes affected References 
X-inactivated genes reactivated after azacytidine treatment (extensive genomic demethylation) Many (such as: SYBL1, PGK1, HPRT, FMR1)  
X-inactivated genes with variable inactivation after azacytidine treatment TIMP 1 Anderson and Brown (2005) 
X-inactivated genes reactivated in ICF syndrome (scarce hypomethylation) SYBL1, G6PD Hansen et al. (2000) 
X-linked genes reactivated upon HDACS treatment SPRY3 This article 
Mechanism of reactivation Genes affected References 
X-inactivated genes reactivated after azacytidine treatment (extensive genomic demethylation) Many (such as: SYBL1, PGK1, HPRT, FMR1)  
X-inactivated genes with variable inactivation after azacytidine treatment TIMP 1 Anderson and Brown (2005) 
X-inactivated genes reactivated in ICF syndrome (scarce hypomethylation) SYBL1, G6PD Hansen et al. (2000) 
X-linked genes reactivated upon HDACS treatment SPRY3 This article 

References

1
Lyon, M.F. (
1961
) Gene action in the X-chromosome of the mouse (Mus musculus L.).
Nature
 ,
190
,
372
–373.
2
Heard, E. (
2005
) Delving into the diversity of facultative heterochromatin: the epigenetics of the inactive X chromosome.
Curr. Opin. Genet. Dev.
 ,
15
,
482
–489.
3
Carrel, L., Cottle, A.A., Goglin, K.C. and Willard, H.F. (
1999
) A first-generation X-inactivation profile of the human X chromosome.
Proc. Natl Acad. Sci. USA
 ,
96
,
14440
–14444.
4
Carrel, L. and Willard, H.F. (
2005
) X-inactivation profile reveals extensive variability in X-linked gene expression in females.
Nature
 ,
434
,
400
–404.
5
Graves, J.A., Wakefield, M.J. and Toder, R. (
1998
) The origin and evolution of the pseudoautosomal regions of human sex chromosomes.
Hum. Mol. Genet.
 ,
7
,
1991
–1996.
6
D'Esposito, M., Matarazzo, M.R., Ciccodicola, A., Strazzullo, M., Mazzarella, R., Quaderi, N.A., Fujiwara, H., Ko, M.S., Rowe, L.B., Ricco, A. et al. (
1997
) Differential expression pattern of XqPAR-linked genes SYBL1 and IL9R correlates with the structure and evolution of the region.
Hum. Mol. Genet.
 ,
6
,
1917
–1923.
7
D'Esposito, M., Ciccodicola, A., Gianfrancesco, F., Esposito, T., Flagiello, L., Mazzarella, R., Schlessinger, D. and D'Urso, M. (
1996
) A synaptobrevin-like gene in the Xq28 pseudoautosomal region undergoes X inactivation.
Nat. Genet.
 ,
13
,
227
–229.
8
Charchar, F.J., Svartman, M., El-Mogharbel, N., Ventura, M., Kirby, P., Matarazzo, M.R., Ciccodicola, A., Rocchi, M., D'Esposito, M. and Graves, J.A. (
2003
) Complex events in the evolution of the human pseudoautosomal region 2 (PAR2).
Genome Res.
 ,
13
,
281
–286.
9
Huber, R., Hansen, R.S., Strazzullo, M., Pengue, G., Mazzarella, R., D'Urso, M., Schlessinger, D., Pilia, G., Gartler, S.M. and D'Esposito, M. (
1999
) DNA methylation in transcriptional repression of two differentially expressed X-linked genes, GPC3 and SYBL1.
Proc. Natl Acad. Sci. USA
 ,
96
,
616
–621.
10
Hansen, R.S., Stoger, R., Wijmenga, C., Stanek, A.M., Canfield, T.K., Luo, P., Matarazzo, M.R., D'Esposito, M., Feil, R., Gimelli, G. et al. (
2000
) Escape from gene silencing in ICF syndrome: evidence for advanced replication time as a major determinant.
Hum. Mol. Genet.
 ,
9
,
2575
–2587.
11
Matarazzo, M.R., De Bonis, M.L., Gregory, R.I., Vacca, M., Hansen, R.S., Mercadante, G., D'Urso, M., Feil, R. and D'Esposito, M. (
2002
) Allelic inactivation of the pseudoautosomal gene SYBL1 is controlled by epigenetic mechanisms common to the X and Y chromosomes.
Hum. Mol. Genet.
 ,
11
,
3191
–3198.
12
Hansen, R.S., Canfield, T.K., Fjeld, A.D. and Gartler, S.M. (
1996
) Role of late replication timing in the silencing of X-linked genes.
Hum. Mol. Genet.
 ,
5
,
1345
–1353.
13
Ingrosso, D., Cimmino, A., Perna, A.F., Masella, L., De Santo, N.G., De Bonis, M.L., Vacca, M., D'Esposito, M., D'Urso, M., Galletti, P. et al. (
2003
) Folate treatment and unbalanced methylation and changes of allelic expression induced by hyperhomocysteinaemia in patients with uraemia.
Lancet
 ,
361
,
1693
–1699.
14
Ciccodicola, A., D'Esposito, M., Esposito, T., Gianfrancesco, F., Migliaccio, C., Miano, M.G., Matarazzo, M.R., Vacca, M., Franze, A., Cuccurese, M. et al. (
2000
) Differentially regulated and evolved genes in the fully sequenced Xq/Yq pseudoautosomal region.
Hum. Mol. Genet.
 ,
9
,
395
–401.
15
Wijmenga, C., van den Heuvel, L.P., Strengman, E., Luyten, J.A., van der Burgt, I.J., de Groot, R., Smeets, D.F., Draaisma, J.M., van Dongen, J.J., De Abreu, R.A. et al. (
1998
) Localization of the ICF syndrome to chromosome 20 by homozygosity mapping.
Am. J. Hum. Genet.
 ,
63
,
803
–809.
16
Hansen, R.S., Wijmenga, C., Luo, P., Stanek, A.M., Canfield, T.K., Weemaes, C.M. and Gartler, S.M. (
1999
) The DNMT3B DNA methyltransferase gene is mutated in the ICF immunodeficiency syndrome.
Proc. Natl Acad. Sci. USA
 ,
96
,
14412
–14417.
17
Okano, M., Bell, D.W., Haber, D.A. and Li, E. (
1999
) DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development.
Cell
 ,
99
,
247
–257.
18
Xu, G.L., Bestor, T.H., Bourc'his, D., Hsieh, C.L., Tommerup, N., Bugge, M., Hulten, M., Qu, X., Russo, J.J. and Viegas-Pequignot, E. (
1999
) Chromosome instability and immunodeficiency syndrome caused by mutations in a DNA methyltransferase gene.
Nature
 ,
402
,
187
–191.
19
Bourc'his, D., Miniou, P., Jeanpierre, M., Molina Gomes, D., Dupont, J., De Saint-Basile, G., Maraschio, P., Tiepolo, L. and Viegas-Pequignot, E. (
1999
) Abnormal methylation does not prevent X inactivation in ICF patients.
Cytogenet. Cell Genet.
 ,
84
,
245
–252.
20
Lewis, A., Mitsuya, K., Umlauf, D., Smith, P., Dean, W., Walter, J., Higgins, M., Feil, R. and Reik, W. (
2004
) Imprinting on distal chromosome 7 in the placenta involves repressive histone methylation independent of DNA methylation.
Nat. Genet.
 ,
36
,
1291
–1295.
21
Umlauf, D., Goto, Y., Cao, R., Cerqueira, F., Wagschal, A., Zhang, Y. and Feil, R. (
2004
) Imprinting along the Kcnq1 domain on mouse chromosome 7 involves repressive histone methylation and recruitment of Polycomb group complexes.
Nat. Genet.
 ,
36
,
1296
–1300.
22
Li, B., Kadura, I., Fu, D.J. and Watson, D.E. (
2004
) Genotyping with TaqMAMA.
Genomics
 ,
83
,
311
–320.
23
Riggs, A.D., Xiong, Z., Wang, L. and LeBon, J.M. (
1998
) Methylation dynamics, epigenetic fidelity and X chromosome structure.
Novartis Found. Symp.
 ,
214
, 214–225; discussion
225
–232.
24
Chiurazzi, P., Pomponi, M.G., Pietrobono, R., Bakker, C.E., Neri, G. and Oostra, B.A. (
1999
) Synergistic effect of histone hyperacetylation and DNA demethylation in the reactivation of the FMR1 gene.
Hum. Mol. Genet.
 ,
8
,
2317
–2323.
25
van der Vlag, J. and Otte, A.P. (
1999
) Transcriptional repression mediated by the human polycomb-group protein EED involves histone deacetylation.
Nat. Genet.
 ,
23
,
474
–478.
26
Sado, T., Fenner, M.H., Tan, S.S., Tam, P., Shioda, T. and Li, E. (
2000
) X inactivation in the mouse embryo deficient for Dnmt1: distinct effect of hypomethylation on imprinted and random X inactivation.
Dev. Biol.
 ,
225
,
294
–303.
27
Sado, T., Okano, M., Li, E. and Sasaki, H. (
2004
) De novo DNA methylation is dispensable for the initiation and propagation of X chromosome inactivation.
Development
 ,
131
,
975
–982.
28
Anderson CL and Brown, C. (
2005
) Epigenetic predisposition to expression of TIMP1 from the human inactive X chromosome.
BMG Genet.
 ,
6
,
48
–62.
29
Tamaru, H. and Selker, E.U. (
2001
) A histone H3 methyltransferase controls DNA methylation in Neurospora crassa.
Nature
 ,
414
,
277
–283.
30
Rougeulle, C., Chaumeil, J., Sarma, K., Allis, C.D., Reinberg, D., Avner, P. and Heard, E. (
2004
) Differential histone H3 Lys-9 and Lys-27 methylation profiles on the X chromosome.
Mol. Cell. Biol.
 ,
24
,
5475
–5484.
31
Wang, J., Mager, J., Chen, Y., Schneider, E., Cross, J.C., Nagy, A. and Magnuson, T. (
2001
) Imprinted X inactivation maintained by a mouse Polycomb group gene.
Nat. Genet.
 ,
28
,
371
–375.
32
Plath, K., Fang, J., Mlynarczyk-Evans, S.K., Cao, R., Worringer, K.A., Wang, H., de la Cruz, C.C., Otte, A.P., Panning, B. and Zhang, Y. (
2003
) Role of histone H3 lysine 27 methylation in X inactivation.
Science
 ,
300
,
131
–135.
33
Silva, J., Mak, W., Zvetkova, I., Appanah, R., Nesterova, T.B., Webster, Z., Peters, A.H., Jenuwein, T., Otte, A.P. and Brockdorff, N. (
2003
) Establishment of histone h3 methylation on the inactive X chromosome requires transient recruitment of Eed–Enx1 polycomb group complexes.
Dev. Cell
 ,
4
,
481
–495.
34
Clarke S, B.K. (
2001
)
S-Adenosylmethionine-Dependent Methyltransferases.
  Cambridge University Press, Cambridge, UK.
35
Yi, P., Melnyk, S., Pogribna, M., Pogribny, I.P., Hine, R.J. and James, S.J. (
2000
) Increase in plasma homocysteine associated with parallel increases in plasma S-adenosylhomocysteine and lymphocyte DNA hypomethylation.
J. Biol. Chem.
 ,
275
,
29318
–29323.
36
Gartler, S.M., Varadarajan, K.R., Luo, P., Canfield, T.K., Traynor, J., Francke, U. and Hansen, R.S. (
2004
) Normal histone modifications on the inactive X chromosome in ICF and Rett syndrome cells: implications for methyl-CpG binding proteins.
BMC Biol.
 ,
2
,
21
.
37
Kawame, H., Gartler, S.M. and Hansen, R.S. (
1995
) Allele-specific replication timing in imprinted domains: absence of asynchrony at several loci.
Hum. Mol. Genet.
 ,
4
,
2287
–2293.
38
Cerrato, F., Dean, W., Davies, K., Kagotani, K., Mitsuya, K., Okumura, K., Riccio, A. and Reik, W. (
2003
) Paternal imprints can be established on the maternal Igf2-H19 locus without altering replication timing of DNA.
Hum. Mol. Genet.
 ,
12
,
3123
–3132.
39
Gribnau, J., Hochedlinger, K., Hata, K., Li, E. and Jaenisch, R. (
2003
) Asynchronous replication timing of imprinted loci is independent of DNA methylation, but consistent with differential subnuclear localization.
Genes Dev.
 ,
17
,
759
–773.
40
Lewis, A. and Murrell, A. (
2004
) Genomic imprinting: CTCF protects the boundaries.
Curr. Biol.
 ,
14
,
R284
–R286.
41
Filippova, G.N., Cheng, M.K., Moore, J.M., Truong, J.P., Hu, Y.J., Nguyen, D.K., Tsuchiya, K.D. and Disteche, C.M. (
2005
) Boundaries between chromosomal domains of X inactivation and escape bind CTCF and lack CpG methylation during early development.
Dev. Cell
 ,
8
,
31
–42.
42
Ke, X. and Collins, A. (
2003
) CpG islands in human X-inactivation.
Ann. Hum. Genet.
 ,
67
,
242
–249.
43
Schug, J., Schuller, W.P., Kappen, C., Salbaum, J.M., Bucan, M. and Stoeckert, C.J., Jr. (
2005
) Promoter features related to tissue specificity as measured by Shannon entropy.
Genome Biol.
 ,
6
,
R33
.
44
Song, F., Smith, J.F., Kimura, M.T., Morrow, A.D., Matsuyama, T., Nagase, H. and Held, W.A. (
2005
) Association of tissue-specific differentially methylated regions (TDMs) with differential gene expression.
Proc. Natl Acad. Sci. USA
 ,
102
,
3336
–3341.
45
Cuadrado, M., Sacristan, M. and Antequera, F. (
2001
) Species-specific organization of CpG island promoters at mammalian homologous genes.
EMBO Rep.
 ,
2
,
586
–592.
46
Reik, W. and Lewis, A. (
2005
) Co-evolution of X-chromosome inactivation and imprinting in mammals.
Nat. Rev. Genet.
 ,
6
,
403
–410.
47
Delaval, K. and Feil, R. (
2004
) Epigenetic regulation of mammalian genomic imprinting.
Curr. Opin. Genet. Dev.
 ,
14
,
188
–195.
48
Wakefield, M.J., Keohane, A.M., Turner, B.M. and Graves, J.A. (
1997
) Histone underacetylation is an ancient component of mammalian X chromosome inactivation.
Proc. Natl Acad. Sci. USA
 ,
94
,
9665
–9668.
49
Hansen, R.S., Canfield, T.K., Stanek, A.M., Keitges, E.A. and Gartler, S.M. (
1998
) Reactivation of XIST in normal fibroblasts and a somatic cell hybrid: abnormal localization of XIST RNA in hybrid cells.
Proc. Natl Acad. Sci. USA
 ,
95
,
5133
–5138.
50
Clemson, C.M., Chow, J.C., Brown, C.J. and Lawrence, J.B. (
1998
) Stabilization and localization of Xist RNA are controlled by separate mechanisms and are not sufficient for X inactivation.
J. Cell Biol.
 ,
142
,
13
–23.
51
Cremer, T. and Cremer, C. (
2001
) Chromosome territories, nuclear architecture and gene regulation in mammalian cells.
Nat. Rev. Genet.
 ,
2
,
292
–301.
52
Mahy, N.L., Perry, P.E., Gilchrist, S., Baldock, R.A. and Bickmore, W.A. (
2002
) Spatial organization of active and inactive genes and noncoding DNA within chromosome territories.
J. Cell Biol.
 ,
157
,
579
–589.
53
Cha, R.S., Zarbl, H., Keohavong, P. and Thilly, W.G. (
1992
) Mismatch amplification mutation assay (MAMA): application to the c-H-ras gene.
PCR Methods Appl.
 ,
2
,
14
–20.
54
Hansen, R.S., Canfield, T.K., Lamb, M.M., Gartler, S.M. and Laird, C.D. (
1993
) Association of fragile X syndrome with delayed replication of the FMR1 gene.
Cell
 ,
73
,
1403
–1409.
55
Hansen, R.S., Canfield, T.K., Fjeld, A.D., Mumm, S., Laird, C.D. and Gartler, S.M. (
1997
) A variable domain of delayed replication in FRAXA fragile X chromosomes: X inactivation-like spread of late replication.
Proc. Natl Acad. Sci. USA
 ,
94
,
4587
–4592.