Logo of ajhgLink to Publisher's site
Am J Hum Genet. Oct 2002; 71(4): 923–932.
Published online Sep 13, 2002. doi:  10.1086/342931
PMCID: PMC378545

Histone Modifications Depict an Aberrantly Heterochromatinized FMR1 Gene in Fragile X Syndrome

Abstract

Fragile X syndrome is caused by an expansion of a polymorphic CGG triplet repeat that results in silencing of FMR1 expression. This expansion triggers methylation of FMR1's CpG island, hypoacetylation of associated histones, and chromatin condensation, all characteristics of a transcriptionally inactive gene. Here, we show that there is a graded spectrum of histone H4 acetylation that is proportional to CGG repeat length and that correlates with responsiveness of the gene to DNA demethylation but not with chromatin condensation. We also identify alterations in patient cells of two recently identified histone H3 modifications: methylation of histone H3 at lysine 4 and methylation of histone H3 at lysine 9, which are marks for euchromatin and heterochromatin, respectively. In fragile X cells, there is a decrease in methylation of histone H3 at lysine 4 with a large increase in methylation at lysine 9, a change that is consistent with the model of FMR1's switch from euchromatin to heterochromatin in the disease state. The high level of histone H3 methylation at lysine 9 may account for the failure of H3 to be acetylated after treatment of fragile X cells with inhibitors of histone deacetylases, a treatment that fully restores acetylation to histone H4. Using 5-aza-2′-deoxycytidine, we show that DNA methylation is tightly coupled to the histone modifications associated with euchromatin but not to the heterochromatic mark of methylation of histone H3 at lysine 9, consistent with recent findings that this histone modification may direct DNA methylation. Despite the drug-induced accumulation of mRNA in patient cells to 35% of the wild-type level, FMR1 protein remained undetectable. The identification of intermediates in the heterochromatinization of FMR1 has enabled us to begin to dissect the epigenetics of silencing of a disease-related gene in its natural chromosomal context.

Introduction

Fragile X mental retardation is caused by mutation of the FMR1 gene (MIM 309550) located at Xq27.3. Of the cases of fragile X, >95% are due to the expansion of a CGG triplet repeat tract located at the 5′ end of FMR1 (Warren and Sherman 2001). In the general population, there is a range of 6–54 repeats in this tract, with a mode of 30 (Fu et al. 1991; Snow et al. 1993). In patients with fragile X, the CGG repeat tract expands to >200 CGG repeats and can be as large as 1,000 repeats (Kremer et al. 1991; Verkerk et al. 1991). This expansion triggers methylation of cytosines in the CpGs in the repeat tract and in the flanking sequence, including the FMR1 gene promoter, resulting in the silencing of FMR1 transcription (Pieretti et al. 1991; Sutcliffe et al. 1992; Hornstra et al. 1993). Treatment of fragile X cells with the DNA methylation inhibitor 5-aza-2′-deoxycytidine (azadC) results in the loss of methylation at expanded FMR1 DNA and reactivates transcription of the gene (Chiurazzi et al. 1998, 1999; Coffee et al. 1999). This result suggests that DNA methylation, not the expansion itself, is the major factor in the loss of transcriptional activity of FMR1 in patients with fragile X.

DNA methylation has long been associated with the transcriptional repression of many genes (Eden and Cedar 1994; Li 1999). The effect of DNA methylation on gene silencing can be direct, by inhibition of transcription factor binding, or indirect, through the induction of changes in local chromatin structure at the site of methylation. One model suggests that DNA methylation induces changes in chromatin architecture through the recruitment of histone deacetylases (HDACs) by a methyl CpG binding (MBD) protein such as MeCP2 (Jones et al. 1998; Nan et al. 1998). Histone deacetylation is thought to result in a tighter association between the histone amino termini and DNA, resulting in a condensed chromatin structure that excludes critical transcription factors. In addition, it has been suggested that removal of the acetyl groups from the histones facilitates nucleosome-nucleosome contacts, thus allowing stacking of the nucleosomes and formation of higher-order structures (Bestor 1998).

In addition to lysine acetylation, other modifications of histones, such as methylation of lysine and arginine, ubiquitination, and phosphorylation, have been identified (Strahl and Allis 2000; Jenuwin and Allis 2001). In addition to a direct influence on the structure of chromatin, these modifications may also serve as “marks” for interactions with other chromatin proteins (Strahl and Allis 2000; Jenuwin and Allis 2001). For example, the transcription factors PCAF and TAFII250 selectively interact with acetylated histone tails through their bromodomains (Dhalliun et al. 1999; Jacobson et al. 2000). Likewise, proteins containing chromodomains, such as HP1, interact with histone H3 tails methylated at lysine 9, a mark associated with heterochromatin (Bannister et al. 2001; Lachner et al. 2001). HP1 has been shown to dimerize in solution, leaving the chromodomains at the ends of flexible linkers, free for association with histone H3 methylated at lysine 9, providing an attractive model for how higher-order chromatin structure can be organized in regions marked by histone H3 methylation at lysine 9 (Brasher et al. 2000; Jenuwin and Allis 2001). Interestingly, a recent report shows that disruption of methylation of histone H3 at lysine 9 perturbs DNA methylation in Neurospora crassa, indicating that histone methylation can direct DNA methylation (Tamaru and Selker 2001). Methylation of histone H3 at lysine 4, on the other hand, is a conserved euchromatic mark that is associated with transcriptionally active genes but not with inactive facultative heterochromatin (Strahl et al. 1999; Litt et al. 2001; Boggs et al. 2002; Peters et al. 2002). Observations such as these have led to the “histone code” hypothesis, according to which a specific constellation of modified histone residues are thought to regulate unique biological outcomes through specific interactions with other components of chromatin (Strahl and Allis 2000; Jenuwin and Allis 2001).

Elsewhere, we demonstrated that there is a loss of acetylation of histone H3 and H4 at FMR1 in fragile X cells compared with normal cells (Coffee et al. 1999). Here, we demonstrate that naturally occurring fragile X alleles, which carry various numbers of CGG repeats, possess a level of H4 acetylation that is inversely proportional to their CGG repeat tract length and to the cells' delayed responsiveness to azadC treatment. However, increased acetylation of histone H4 associated with these FMR1 alleles, either naturally occurring or induced by the HDAC inhibitor trichostatin A (TSA), was not sufficient to relax the condensed chromatin structure at the locus. We also identify fragile X–specific alterations in two additional histone modifications at FMR1: methylation of histone H3 at lysine 4 (a mark for euchromatin) and methylation of lysine 9 (a mark for heterochromatin). Histone H3 at FMR1 in fragile X cells is methylated at lysine 9 and undermethylated at lysine 4. The modification is similar to the code that is found in heterochromatic regions of the genomes of model organisms and to the code that is imprinted both on human genes and on the inactive X chromosome in all mammals (Strahl et al. 1999; Litt et al. 2001; Xin et al. 2001; Boggs et al. 2002; Peters et al. 2002). Treatment of fragile X cells with HDAC inhibitors has little, if any, effect on these modifications, suggesting that histone H3 methylation at lysine 9 blocks the acetylation of this residue. Finally, we show that FMR1 DNA methylation, altered by treatment with azadC, is tightly coupled to changes in histone H3 and H4 acetylation, and to methylation of histone H3 at lysine 4, all marks for euchromatin. However, methylation of histone H3 at lysine 9, a mark for heterochromatin, did not follow drug-induced changes in DNA methylation, indicating that it may operate independently of DNA methylation.

Material and Methods

Cell Lines, Cell Culture, and Drug Treatment

Epstein-Barr virus (EBV)–transformed lymphoblastoid cell lines E3 and E4 derived from male patients with fragile X and carrying FMR1 alleles of 230 CGG repeats or 410 CGG repeats (data not shown), respectively, were a generous gift from G. Neri (Universita Cattolica, Rome) (Chiurazzi et al. 1999). The EBV-transformed lymphoblastoid cell line J-1, carrying an unmethylated 30–CGG repeat allele, was derived from an unaffected male and GM3200A (available from Coriell Cell Repositories), carrying a methylated 530–CGG repeat allele, was derived from a male patient with fragile X. The cells were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium supplemented with 10% fetal bovine serum and 100 U/ml penicillin and 100μg/ml streptomycin.

TSA (330 nM; Sigma) or sodium butyrate (10 mM; Sigma) was added to the media for 24 h; azadC (1 μM; Sigma) was added to the media and was replenished every 48 h by replacement of half the conditioned media with fresh media and drug. In the azadC withdrawal experiment, cells were synchronized with two 8-h blocks of thymidine prior to addition of azadC (Chiurazzi et al. 1998).

Chromatin Immunoprecipitation

We performed chromatin immunoprecipitation as described elsewhere (Coffee et al. 1999). The chromatin solution was used for immunoprecipitation with antibodies directed against histone H4 acetylated at lysines 5, 8, 12, and 16 (Upstate Biotech), histone H3 acetylated at lysines 9 and 14 (Upstate Biotech), histone H3 dimethylated at lysine 9 (Upstate Biotech), and histone H3 dimethylated at lysine 4 (Upstate Biotech). Prior to PCR analysis, the IP-DNAs were digested with XhoI to separate the CGG repeat tract, which interferes with amplification. The PCR conditions and primers used were as described elsewhere (Coffee et al. 1999). The PCR products were electrophoresed, and the image was captured using the Eagle Eye imaging system and was analyzed with Molecular Dynamics ImageQuant software.

RT-PCR and LightCycler RT-PCR

Total RNA was isolated using Trizol (Gibco BRL) or RNeasy (Qiagen), according to the manufacturers' instructions. Conventional RT-PCR was performed using the GeneAmp RNA PCR kit (Perkin Elmer), as described elsewhere (Coffee et al. 1999). LightCycler RT-PCR was performed using the LightCycler RNA amplification kit (Roche), according to the manufacturers' suggested conditions for one-step RT-PCR. Total RNA (100 ng) was used for each sample. The PCR conditions were as follows: RT at 55°C for 10 min, followed by an initial denaturation at 95°C for 30 s, then 45 cycles of 95°C for 1 s, 60°C for 10 s, and 72°C for 30 s. SYBR Green I (Roche) was used as the fluorescent marker to monitor DNA accumulation. The primers used were 5′-GATGAAGATACCTGCACATTC-3′ and 5′-TAGCTCCAATCTGTCGCAACTGC-3′, yielding a 593-bp product. Melting-curve analysis of the reactions was done using the LightCycler software (Roche).

Western Blot Analysis

Protein was isolated and quantified using the Bradford reagent (BioRad) and amounts specified in figure 1 were loaded onto a 7.5% SDS polyacrylamide gel. The resolved protein was transferred to Hybond-P membrane (Amersham), was stained with PonceauS (Sigma) to confirm transfer, and was then blocked in 10% milk (Carnation) in PBS/0.2% Tween-20 for 1 h. The membrane was probed with the anti-FMR1 protein (FMRP) monoclonal antibody 1FM 1AC hybridoma supernatant at 1:10 (Devys et al. 1993) for 3 h, washed, and then probed with an anti-mouse Ig, Fc-specific-antibody coupled to HRP (Sigma) for 1 h. After four 15-min washes, the blot was developed using a chemiluminescent substrate (Amersham) and was then exposed to film.

Figure  1
Fragile X cell lines treated with azadC. A, Light cycler RT-PCR quantitation of FMR1 transcript following azadC treatment of fragile X cell lines harboring 230 CGGs, 410 CGGs, or 530 CGGs. The amount of FMR1 transcript in 100 ng of total RNA (expressed ...

Nuclease Accessibility

Approximately 107 EBV-transformed lymphoblastoid cells were washed with PBS and permeabilized with L-α-lysophosphatidylcholine, as described elsewhere (Eberhart et at. 1996). The cells were collected and suspended in NEB2 restriction enzyme buffer (New England Biolabs), and the indicated amounts of MspI were added. The samples were incubated for 60 min at 37°C and treated with proteinase K, and the DNA was isolated, digested with HindIII, and analyzed by Southern blot hybridization as described elsewhere (Pieretti et al. 1991).

Southern Blot Analysis

DNA for Southern blot analysis was isolated from the cells by use of a salting-out procedure (Miller et al. 1988). Genomic DNA (10 μg) was digested with the restriction enzyme pair of HindIII and BssHII. The products of the digestion were separated by electrophoresis on an 0.8% agarose gel and transferred to Hybond N+ membrane (Amersham). The membrane was hybridized with a 32P-labeled 5.2-kb fragment of FMR1 DNA isolated from plasmid pE5.1 (Pieretti et al. 1991).

Results

Histone H4 Acetylation at FMR1 Is Proportional to CGG Repeat Tract Length and to Responsiveness to azadC Treatment

Patients with fragile X harbor CGG repeat expansions that vary from ~200 repeats to >1,000 repeats. Previous work suggested that fragile X cell lines carrying shorter CGG repeat tracts were more responsive to azadC (Chiurazzi et al. 1999). Using light-cycler RT-PCR, we carefully measured the amount of FMR1 transcript induced by azadC over time in three different male fragile X cell lines harboring alleles with ~230, ~410, or ~530 CGG repeats. These alleles are fully methylated at the diagnostic EagI and BssHII restriction sites located in the FMR1 promoter (data not shown). FMR1 transcripts became readily detectable in the 230-CGG repeat cell line at day 4, with the longer-repeat cell lines taking until day 6 to produce an equivalent amount of FMR1 transcript (fig. 1). FMR1 mRNA continued to accumulate in all the cell lines throughout this course of treatment, with the amount of FMR1 transcript present at day 8 being inversely proportional to CGG repeat tract length (fig. 1).

It has been reported that CGG repeat expansions, of the size usually found in patients, in the FMR1 transcript can impair translation by impeding the migration of the 40S subunit along mRNA (Feng et al. 1995). However, FMRP has been detected after azadC treatment of fragile X syndrome cells (Chiurazzi et al. 1998). Since we achieved such high levels of FMR1 transcript during reactivation of the short allele–containing cells with azadC, we used western blot analysis to determine whether FMRP also accumulated. In normal cells, FMRP could be detected with as little as 3 μg of total cell protein lysate loaded on the gel (fig. 1B). However, in fragile X cells, FMRP could not be detected, even when 20-fold more protein was analyzed. Therefore, in the 230-CGG repeat cell line in which transcription was reactivated to a level >35% of normal after 8 d of azadC treatment, FMRP remained undetectable (<5% of normal), indicating that FMRP translation is inhibited by CGG expansion.

The finding of differential reactivation led us to test the idea that histone acetylation at FMR1 differed as a function of repeat size. Using chromatin immunoprecipitation (ChIP), we found that histone H3 was significantly hypoacetylated and was barely detectable in the three fragile X cell lines (fig. 2). The average ratios of histone H3 acetylation, relative to the constitutively expressed, X-linked G6PD gene, for the fragile X cell lines carrying repeats of 230 CGGs, 410 CGGs, or 530 CGGs were 0.07, 0.03, and <0.01, respectively, versus 1.14 for normal cells (fig. 2). The average ratios of histone H4 acetylation at FMR1 were 0.63, 0.23, and 0.13, respectively, versus 1.11 for normal cells. Therefore, there was a graded loss of histone acetylation, especially histone H4, that was proportional to CGG repeat length and to the cell line's responsiveness to azadC, suggesting that alleles with longer CGG repeat tracts are further down a pathway of transcriptional repression.

Figure  2
Histone H3 and histone H4 acetylation at FMR1 in fragile X cell lines harboring 230 CGGs, 410 CGGs, or 530 CGGs. A, A representative multiplex PCR analysis of DNAs immunoprecipitated with either anti-acetyl histone H3 or anti-acetyl histone H4 antibodies ...

Acetylation of Histone H4 Alone Does Not Relax the Condensed Chromatin Structure at Fragile X FMR1 Alleles

Relative to normal cells, chromatin at FMR1 in permeabilized fragile X cells is resistant to MspI digestion, indicating that it is organized into a condensed state (Eberhart et al. 1996). The availability of cell lines harboring FMR1 alleles with a spectrum of histone H4 acetylation led us to test whether the higher levels of histone H4 acetylation are a reflection of a more open chromatin structure. Using the MspI nuclease accessibility assay (Eberhart et al. 1996), we gauged the compaction of chromatin at FMR1 associated with various levels of histone H4 acetylation. In permeabilized normal cells, cleavage at the 16 MspI sites at the FMR1 promoter (fig. 3A) was readily observed at all MspI concentrations (fig. 3B, lanes 1–4). The same sites were much less accessible to MspI in permeabilized fragile X cells with 530 CGG repeats, where histone H4 was almost completely deacetylated (fig. 3B, lanes 9–12). FMR1 DNA associated with an increased level of histone H4 acetylation at the naturally occurring 230-CGG repeat allele, where H4 acetylation was ~50% of normal, was also highly resistant to MspI digestion (fig. 3B, lanes 5–8). Finally, TSA treatment of the 530-CGG repeat cell line, which restores H4 acetylation to normal levels but has little or no effect on H3 acetylation (Coffee et al. 1999; also see fig. 4), did not observably increase accessibility of FMR1 DNA to MspI (fig. 3B, lanes 13–16). This indicates that even a wild-type level of histone H4 acetylation is not sufficient to open the condensed heterochromatin-like structure at FMR1 in patients with fragile X, which may explain, in part, why TSA does not efficiently reactivate transcription of FMR1 alleles in fragile X cells (Chiurazzi et al. 1999; Coffee et al. 1999).

Figure  3
Nuclease accessibility analysis of FMR1 DNA associated with various levels of acetylated histone H4. A, Map of the 5′ end of FMR1, showing the location of the relevant restriction sites, the transcriptional start site, and the CGG repeat tract. ...
Figure  4
ChIP analysis of histone H4 and H3 acetylation, H3 methylation at lysine 9 and histone H3 methylation at lysine 4 in a normal and a fragile X cell line carrying a 530-CGG repeat allele without HDAC inhibitor treatment (lanes 1 and 2) and the fragile X ...

Histone H3 Methylation Is Increased at Lysine 9 and Is Decreased at Lysine 4 at FMR1 in Fragile X Cells

Recently, it has been shown that methylation of histone H3 at lysine 4 or lysine 9 is selectively enriched in euchromatic and heterochromatic regions of chromosomes, respectively (Strahl et al. 1999; Rea et al. 2000; Boggs et al. 2002; Peters et al. 2002). We used ChIP analysis to determine whether these chromatin marks were altered at FMR1 in fragile X cells. Histone H3 methylation at lysine 4 was significantly reduced at FMR1 in fragile X cells (fig. 4, “MeH3-K4,” lanes 1 and 2). Conversely, the amount of methylation of histone H3 at lysine 9 at FMR1 in fragile X cells was significantly greater than the amount observed at FMR1 in normal cells (fig. 4, “MeH3-K9,” lanes 1 and 2). As expected, methylation of histone H3 at lysine 9 was not observed at the constitutively active G6PD gene. These results demonstrate a complementary pattern of histone H3 methylation at FMR1 in normal and fragile X cells, in which the euchromatic marks in normal cells are replaced with heterochromatic marks in fragile X cells.

An individual lysine 9 of histone H3 can be acetylated or methylated but cannot possess both groups simultaneously. To determine whether methylation of histone H3 at FMR1 changed in a coordinated fashion with acetylation, we used ChIP to assess alterations in these modifications at FMR1 in fragile X cell lines after treatment with the HDAC inhibitors sodium butyrate or TSA. After drug treatment, histone H4 acetylation at FMR1 increased to wild-type levels, as shown previously (fig. 4, “AcH4”) (Coffee et al. 1999). There was a small increase in methylation of histone H3 at lysine 4 with sodium butyrate (fig. 4, “MeH3-K4”). Methylation of H3 lysine 9 was not reduced, and may slightly increase after the drug treatment (fig. 4, “AcH3” and “MeH3-K9”). These data suggest that methylation of histone H3 at lysine 9 blocks its acetylation, preventing the transition of FMR1 from a heterochromatic-like state to a euchromatic-like state.

Acetylation of Histone H3 and H4 Is Tightly Coupled to DNA Methylation

Elsewhere, we and others have demonstrated that treatment of fragile X cells with azadC resulted in transcriptional reactivation of FMR1 (Chiurazzi et al. 1998, 1999; Coffee et al. 1999). This reactivation is accompanied by an increase of both histone H3 and H4 acetylation at the 5′ end of the gene (Coffee et al. 1999; and shaded portion of fig. 5). To determine whether the changes in DNA methylation and the concomitant changes in histone acetylation and transcription are stable without the continual presence of azadC, the drug was removed from the media, and the cells were passaged for an additional 30 d (unshaded portion of fig. 5). Samples were collected at various intervals and were assayed for DNA methylation, by digestion with methylation-sensitive restriction enzymes, histone acetylation, and FMR1 mRNA levels (fig. 5). Immediately following withdrawal of azadC, there was an increase in DNA methylation with a parallel decrease in histone acetylation. Steady-state FMR1 mRNA levels began to decrease several days later. Twenty days after the removal of azadC, FMR1 DNA became completely remethylated at the diagnostic BssHII site in the FMR1 promoter, histone acetylation returned to pretreatment levels, and FMR1 transcript became undetectable (fig. 5). Thus, the changes in DNA methylation and histone acetylation induced by azadC at FMR1 are not maintained without the continuous presence of the drug, and the heterochromatic nature of FMR1 in fragile X cells is tightly coupled to DNA methylation.

Figure  5
DNA methylation, ChIP, and RT-PCR analysis of FMR1 after addition and withdrawal of azadC. The fragile X cell line harboring 530 CGG repeats was treated with 1 μM azadC for 5 d (shaded area), followed by removal of the drug from the media. The ...

Methylation of Histone H3 at Lysine 9 Does Not Follow Drug-Induced Loss of DNA Methylation and Transcriptional Reactivation of FMR1

Methylation of histone H3 at lysine 9 and methylation of cytosines in CpG dinucleotides are features of heterochromatin (Jenuwin and Allis 2001). The molecular link between these two epigenetic changes, however, remains to be defined. To determine whether histone H3 methylation is coupled to DNA methylation at FMR1, we used the azadC-mediated demethylation of FMR1 DNA as a model to assess the changes in histone H3 methylation that occurs during the switch from a transcriptionally inactive heterochromatin-like state to a transcriptionally active euchromatin-like state. DNA demethylation became detectable (as assessed by cleavage at the methyl-sensitive BssHII restriction sites located in the FMR1 promoter) as early as 2 d after addition of azadC, and it continued to increase until approximately day 8, when it plateaued at ~13%–14% (fig. 6). FMR1 transcript became detectable by RT-PCR at approximately day 6 and increased steadily throughout the remainder of the treatment (fig. 6). Like the euchromatic marks of histone H3 and H4 acetylation, histone H3 methylation at lysine 4 increases at FMR1 steadily throughout the treatment, doubling from 45% of the normal level in the untreated fragile X cell to 90% of normal after 16 d of azadC treatment (data not shown). Methylation of histone H3 at lysine 9 decreased initially at FMR1, dropping to about half the amount seen in untreated cells after 2 d of treatment, but it reproducibly recovered and by day 16 was at or above the level seen in untreated cells (fig. 6). These results demonstrate that the heterochromatic mark of methylation of histone H3 at lysine 9 at FMR1 did not track with DNA demethylation, as did the euchromatic marks of H3 methylation at lysine 4 and histone acetylation, suggesting that this histone modification is related in a more complex way to DNA methylation.

Figure  6
DNA methylation, RT-PCR, and ChIP analysis of FMR1 during the course of azadC treatment. The fragile X cell line carrying 410 CGG repeats was treated with 1 μM azadC for 16 d. Equal numbers of cells were harvested at the indicated times for DNA, ...

Discussion

Elsewhere, we showed that treatment of fragile X cells with TSA resulted in acetylation of histone H4 at FMR1 but had little, if any, effect on histone H3 acetylation and failed to reactivate transcription (Coffee et al. 1999). Loss of DNA methylation after azadC treatment resulted in an increase in both histone H3 and H4 acetylation and transcriptional reactivation. This suggested that histone H3 deacetylation, but not histone H4 deacetylation, is linked to DNA methylation and transcriptional silencing (Coffee et al. 1999). Here, we extend this initial observation by showing that acetylation of histone H4 at naturally occurring fragile X FMR1 alleles can vary between 10% and 50% of normal levels. The extent of histone H4 deacetylation correlated with increasing CGG repeat length. One possible basis for this observation could be an increasing abundance of HDACs at the locus, resulting from a larger number of binding sites for methylcytosine-binding proteins in the longer CGG repeat alleles. Histone H3 was more profoundly and uniformly deacetylated across the spectrum of CGG repeat sizes, suggesting that loss of this modification is more important in transcriptional silencing. A similar linkage of DNA methylation to histone H3 deacetylation, but not histone H4 deacetylation, was found at the imprinted Snrpn and U2af1-rs1 genes in mice (Gregory et al. 2001).

The spectrum of histone H4 acetylation found at FMR1 in fragile X cells suggested that there is a gradation in the possible heterochromatin-like states that FMR1 can populate. Supporting this notion is the observation shown here, and also by others (Chiurazzi et al. 1999), that FMR1 transcription is reactivated earlier by azadC treatment in fragile X cells with shorter CGG repeat expansions. The earlier azadC reactivation of FMR1 that possesses partially acetylated histones is reminiscent of the synergistic reactivation achieved by dual treatment with HDAC inhibitors and a subthreshold dose of azadC (Chiurazzi et al. 1999). This relatively large restoration of FMR1 mRNA is not accompanied by a proportional increase in FMRP accumulation. This is consistent with prior work indicating that expansion-containing mRNA is refractory to efficient translation (Feng et al. 1995).

Partial or even full acetylation of histone H4, however, did not increase the accessibility of FMR1 chromatin to nucleases, indicating that H4 acetylation is not sufficient to relax chromatin condensation at FMR1. However, the observation that histone H4 acetylation—either at the naturally occurring 230–CGG repeat allele or induced by HDAC inhibitor treatment of cells with large repeat tracts—augments azadC-mediated transcriptional reactivation suggests that the extent of histone H4 acetylation correlates with the preparation of this locus to become transcriptionally active. Whether histone H4 acetylation is necessary for transcription of FMR1 remains to be determined.

Complementary patterns of histone H3 methylation at lysines 4 and 9 are emerging as a common theme in the difference between transcriptionally active and inactive loci. For example, at the chicken β-globin locus, transcriptionally active genes are associated with histone H3 methylated at lysine 4, and transcriptionally repressed genes are associated with methylation of histone H3 at lysine 9 (Litt et al. 2001). Similarly, there is a complementary pattern of histone H3 methylation at lysines 4 and 9 at the Prader-Willi imprinting centers in humans (Xin et al. 2001). Here we show that there is also a complementary pattern of histone H3 methylation at lysine 4 and 9, associated with FMR1 in normal and fragile X cells, that correlates with gene expression. In normal cells, histone H3 associated with FMR1 is methylated at lysine 4 but not at lysine 9. In fragile X cells, histone H3 associated with FMR1 is undermethylated at lysine 4 and hypermethylated at lysine 9.

Methylation of histone H3 at lysine 9 at expanded FMR1 alleles suggests that a histone methyltransferase, such as SUV39H1 (Rea et al. 2000), is being targeted to the locus. The chromodomain of the heterochromatin protein HP1 has been shown to have an affinity for histone H3 tails methylated at lysine 9, suggesting a model in which methylation of histone H3 at lysine 9 targets HP1 to a locus resulting in heterochromatin formation and transcriptional repression (Bannister et al. 2001; Lachner et al. 2001). Indeed, methylation of histone H3 at lysine 9, directed by Rb, targets HP1 to the cyclin E promoter, correlating with the repression of its transcription (Nielsen et al. 2001). This model may also apply to FMR1 in fragile X syndrome. Maintenance of methylation of histone H3 at lysine 9 after TSA treatment would explain why FMR1 remains in a condensed state, despite an increase of histone H4 acetylation to normal levels, since a chromodomain-containing protein, such as HP1, could remain targeted to the gene.

We also show that, after removal of azadC, there is a coordinated remethylation of FMR1 DNA, renewed deacetylation of histones, and transcriptional silencing. This result demonstrates that these changes are not stable without the continuous presence of azadC. It also reinforces the observation that these changes are tightly coupled to the methylation state of the DNA. We emphasize that, although suggestive, the data do not prove that these chromatin changes are causally involved in transcription.

Like the marks of histone H3 and H4 acetylation, methylation of histone H3 at lysine 4 increased after azadC treatment, indicating that this modification is also coupled to DNA methylation of FMR1. In contrast to these changes, methylation of H3 at lysine 9 was transient, decreasing to ~50% of the starting level but recovering to the original level by day 16. FMR1 transcripts continued to accumulate throughout the course of azadC treatment, indicating that the chromatin changes due to drug treatment may override the repressive effects of methylation of histone H3 at lysine 9, possibly by inhibiting binding of proteins, such as HP1. Since there is an increase in acetylation but no net change in methylation of H3 lysine 9 after 16 d, the increase in acetylation must be due to acetylation of other accessible lysine 9s of different H3 tails that reside in the same nucleosome, on neighboring nucleosomes, or even on different chromosomes in the population of cells.

The continued association of partially demethylated FMR1 DNA with histone H3 methylated at lysine 9, coupled with the observation that FMR1 DNA becomes remethylated after withdrawal of azadC, suggests that undermethylated FMR1 DNA associated with histone H3 methylated at lysine 9 could be a target for DNA methyltransferase. This is consistent with a recent model in which methylation of histone H3 at lysine 9 is a determinant of DNA methylation (Tamaru and Selker 2001). In our working model, histone H3 associated with expanded, unmethylated CGG repeat tracts would become a target for a histone methyltransferase such as SUV39H1. The association of FMR1 DNA containing expanded CGG tracts with histone H3 methylated at lysine 9 would then result in the targeting of a DNA methyltransferase to FMR1 by association with a chromodomain-containing protein. There is precedent in model organisms for the association of a DNA methyltransferase activity with a chromodomain. For example, in Arabidopsis thaliana, DNA methylation and histone methylation are linked by a set of chromomethylase proteins, which contain both a chromodomain and the catalytic domain of a DNA methyltransferase (Lindroth et al. 2001). It also has been shown recently that, in N. crassa, DNA methylation is dependent upon histone H3 methylation at lysine 9 and that disruption of histone methylation leads to DNA demethylation (Tamaru and Selker 2001). In our working model, after establishment of DNA methylation, HDACs and other components of the transcriptional repression machinery would be targeted to FMR1 by association with methyl DNA-binding proteins, such as MeCP2, which condenses the chromatin at the locus, repressing transcription of the gene. HDAC inhibitors are able to partially reverse this process by inhibiting H4 deacetylation at the locus, but, because histone H3 is methylated at lysine 9, acetylation of this key amino acid would be blocked, preventing the transition from a transcriptionally silent condensed chromatin state back to a transcriptionally active open chromatin state. This model suggests the possibility that disruption of histone H3 methylation at lysine 9 at FMR1 in fragile X cells could also lead to DNA demethylation, resulting in chromatin opening and transcriptional reactivation of the gene.

Acknowledgments

We thank J. Boss, P. Vertino, and P. Wade for helpful discussions. This work is supported by a postdoctoral research grant from the FRAXA Research Foundation and by National Institutes of Health grant HD35576.

Electronic-Database Information

The accession number and URL for data presented herein are as follows:

Online Mendelian Inheritance in Man (OMIM), http://www.ncbi.nlm.nih.gov/Omim/ (for FMR1 [MIM 309550])

References

Bannister AJ, Zegerman P, Partridge JF, Miska EA, Thomas JO, Allshire RC, Kouzarides T (2001) Selective recognition of methylated lysine 9 on histone H3 by the HP1 chromodomain. Nature 410:120–124 [PubMed]
Bestor TH (1998) Methylation meets acetylation. Nature 393:311–312 [PubMed]
Boggs BA, Cheung P, Heard E, Spector DL, Chinault AC, Allis CD (2002) Differentially methylated forms of histone H3 show unique association patterns with inactive human X chromosomes. Nat Genet 30:73–76 [PubMed]
Brasher SV, Smith BO, Fogh RH, Nietlispach D, Thiru A, Nielsen PR, Broadhurst RW, Ball LJ, Murzina NV, Laue ED (2000) The structure of mouse HP1 suggests a unique mode of single peptide recognition by the shadow chromodomain dimer. EMBO J 19:1587–1597 [PMC free article] [PubMed]
Chiurazzi P, Pomponi MG, Pietrobono R, Bakker CE, Neri G, Oostra BA (1999) Synergistic effect of histone hyperacetylation and DNA demethylation in the reactivation of the FMR1 gene. Hum Mol Genet 8:2317–2323 [PubMed]
Chiurazzi P, Pomponi MG, Willemsen R, Oostra BA, Neri G (1998) In vitro reactivation of the FMR1 gene involved in fragile X syndrome. Hum Mol Genet 7:109–113 [PubMed]
Coffee B, Zhang F, Warren ST, Reines D (1999) Acetylated histones are associated with FMR1 in normal but not fragile X syndrome cells. Nat Genet 22:98–101 [PubMed]
Devys D, Lutz Y, Rouyer N, Bellocq JP, Mandel JL (1993) The FMR-1 protein is cytoplasmic, most abundant in neurons, and appears normal in carriers of the fragile X premutation. Nat Genet 4:335–340 [PubMed]
Dhalliun C, Carlson JE, Zeng L, He C, Aggarwal AE, Zhou M-M (1999) Structure and ligand of a histone acetyltransferase bromodomain. Nature 399:491–496 [PubMed]
Eberhart DE, Warren ST (1996) Nuclease sensitivity of permeabilized cells confirms altered chromatin formation at the fragile X locus. Somat Cell Mol Genet 22:435–441 [PubMed]
Eden S, Cedar H (1994) Role of DNA methylation in the regulation of transcription. Curr Opin Genet Dev 4:255–259 [PubMed]
Feng Y, Zhang F, Lokey L, Chastain JL, Lakkis L, Eberhart D, Warren ST (1995) Translation suppression by trinucleotide repeat expansion at FMR1. Science 268:731–734 [PubMed]
Fu Y-H, Kuhl DPA, Pizzuti A, Pieretti M, Sutcliffe JS, Richards S, Verkerk AJMH, Holden JJA, Fenwick RG, Warren ST, Oostra BA, Nelson DL, Caskey CT (1991) Variation of the CGG repeat at the fragile X site in genetic instability: resolution of the Sherman paradox. Cell 67:1047–1058 [PubMed]
Gregory RI, Randall TE, Johnson CA, Khosla S, Hatada I, O'Neill LP, Turner BM, Feil R (2001) DNA methylation is linked to deacetylation of histone H3, but not histone H4, on the imprinted genes Snrpn and U2af1-rs1. Mol Cell Biol 21:5426–5436 [PMC free article] [PubMed]
Hornstra IK, Nelson DL, Warren ST, Yang TP (1993) High resolution methylation analysis of the FMR1 gene trinucleotide repeat region in fragile X syndrome. Hum Mol Genet 2:1659–1665 [PubMed]
Jacobson RH, Ladurne AGR, King DS, Tjian R (2000) Structure and function of a human TAFII250 double bromodomain module. Science 288:1422–1425 [PubMed]
Jenuwin T, Allis CD (2001) Translating the histone code. Science 293:1074–1080 [PubMed]
Jones PL, Veenstra GJC, Wade PA, Vermaak D, Kass SU, Landsberger N, Strouboulis J, Wolffe AP (1998) Methylated DNA and MeCP2 recruit histone deacetylase to repress transcription. Nat Genet 19:187–191 [PubMed]
Kremer EJ, Pritchard M, Lynch M, Yu S, Holman K, Bakker E, Warren ST, Schlessinger D, Sutherland GR, Richards RI (1991) Mapping of DNA instability at the fragile X to a trinucleotide repeat sequence p(CGG)n. Science 252:1711–1714 [PubMed]
Lachner M, O'Carroll D, Rea S, Mechtler K, Jenuwein T (2001) Methylation of histone H3 lysine 9 creates a binding site for HP1 proteins. Nature 410:116–120 [PubMed]
Li E (1999) The mojo of methylation. Nat Genet 23:5–6 [PubMed]
Lindroth AM, Cao X, Jackson JP, Zilberman D, McCallum CM, Henikoff S, Jacobsen SE (2001) Requirement of CHROMOMETHYLASE3 for maintenance of CpXpG methylation. Science 292:2077–2080 [PubMed]
Litt MD, Simpson M, Gaszner M, Allis CD, Felsenfeld G (2001) Correlation between histone lysine methylation and developmental changes at the chicken β-globin locus. Science 293:2453– 2455 [PubMed]
Miller SA, Dykes DD, Polesky HF (1988) A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res 16:1215 [PMC free article] [PubMed]
Nan X, Ng H-H, Johnson CA, Laherty CD, Turner BM, Eisenman RN, Bird A (1998) Transcriptional repression by the methyl-CpG-binding protein MeCP2 involves a histone deacetylase complex. Nature 393:386–389 [PubMed]
Nielsen SJ, Schneider R, Bauer U-M, Bannister AJ, Morrison A, O'Carroll D, Firestein R, Cleary M, Jenuwein T, Herrera RE, Kouzarides T (2001) Rb targets histone H3 methylation and HP1 to promoters. Nature 412:561–565 [PubMed]
Peters AHFM, Mermoud JE, O'Carroll D, Pagani M, Schhweizer D, Brockdorff N, Jenuwin T (2002) Histone H3 lysine methylation is an epigenetic imprint of facultative heterochromatin. Nat Genet 30:77–80 [PubMed]
Pieretti M. Zhang F, Fu Y-H, Warren ST, Oostra BA, Caskey CT, Nelson DL (1991) Absence of expression of the FMR1 gene in fragile X syndrome. Cell 66:817–822 [PubMed]
Rea S, Eisenhaber F, O'Carroll D, Strahl BD, Sun Z-W, Schmid M, Opravil S, Mechtler K, Ponting CP, Allis CD, Jenuwein T (2000) Regulation of chromatin structure by site-specific histone H3 methyltransferases. Nature 406:593–599 [PubMed]
Snow K, Doud LK, Hagerman R, Pergolizzi RG, Erster SH, Thibodeau SN (1993) Analysis of a CGG sequence at the FMR-I locus in fragile X families and in the general population. Am J Hum Genet 53:1217–1228 [PMC free article] [PubMed]
Strahl BD, Allis CD (2000) The language of covalent histone modifications. Nature 403:41–45 [PubMed]
Strahl BD, Ohba R, Cook RG, Allis CD (1999) Methylation of histone H3 at lysine 4 is highly conserved and correlates with transcriptionally active nuclei in Tetrahymena. Proc Natl Acad Sci USA 96:14967–14972 [PMC free article] [PubMed]
Sutcliffe JS, Nelson DL, Zhang F, Pieretti M, Caskey CT, Saxe D, Warren ST (1992) DNA methylation represses FMR-1 transcription in fragile X syndrome. Hum Mol Genet 1:397–400 [PubMed]
Tamaru H, Selker EU (2001) A histone H3 methyltransferase controls DNA methylation in Neurospora crassa. Nature 414:277–283 [PubMed]
Verkerk AJ, Pieretti M, Sutcliffe JS, Fu Y, Kuhl DP, Pizzuti A, Reiner O, Richards S, Victoria MF, Zhang F (1991) Identification of a gene (FMR-1) containing a CGG repeat coincident with a breakpoint cluster region exhibiting length variation in fragile X syndrome. Cell 65:905–914 [PubMed]
Warren ST, Sherman SL (2001) The fragile X syndrome. In: Scriver CR (ed). The metabolic and molecular basis of inherited disease, 8th ed. McGraw-Hill, New York, pp 1257–1289
Xin Z, Allis CD, Wagstaff J (2001) Parent-specific complementary patterns of histone H3 lysine 9 and H3 lysine 4 methylation at the Prader-Willi syndrome imprinting center. Am J Hum Genet 69:1389–1394 [PMC free article] [PubMed]

Articles from American Journal of Human Genetics are provided here courtesy of American Society of Human Genetics
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...