Transient cyclical methylation of promoter DNA

Vol 452 | 6 March 2008 | doi:10.1038/nature06640

LETTERS Transient cyclical methylation of promoter DNA Sara Kangaspeska1*, Brenda Stride1*{, Raphae¨l Me´tivier2, Maria Polycarpou-Schwarz1, David Ibberson1, Richard Paul Carmouche1, Vladimir Benes1, Frank Gannon1{ & George Reid1

Methylation of CpG dinucleotides is generally associated with epigenetic silencing of transcription and is maintained through cellular division1–3. Multiple CpG sequences are rare in mammalian genomes, but frequently occur at the transcriptional start site of active genes, with most clusters of CpGs being hypomethylated4. We reported previously that the proximal region of the trefoil factor 1 (TFF1, also known as pS2) and oestrogen receptor a (ERa) promoters could be partially methylated by treatment with deacetylase inhibitors5, suggesting the possibility of dynamic changes in DNA methylation. Here we show that cyclical methylation and demethylation of CpG dinucleotides, with a periodicity of around 100 min, is characteristic for five selected promoters, including the oestrogen (E2)-responsive pS2 gene, in human cells. When the pS2 gene is actively transcribed, DNA methylation occurs after the cyclical occupancy of ERa and RNA polymerase a –259 –201

NucE

–354 –301 HpaII

RE 2E 1/ RE E 2 1/

–388–370 –400

–167

NucT

–24 –23

–84HpaII qPCR

–74 –78 –43

Addition of E2 (min)

c

+1

e 100 80

HpaII –9

–181 –84 –56 HpaII

–439 –354HpaII qPCR –285 –236

–30

–20 –13 HpaII

+17

CpG –20 Undig.

+10

–20 HpaII qPCR

+60

GST–MBD qPCR

Input (%)

–266

0 10 20 40 60 80 90 10 0 11 0 12 0 13 0 14 0

–405 –393 –409

II (polII). Moreover, we report conditions that provoke methylation cycling of the pS2 promoter in cell lines in which pS2 expression is quiescent and the proximal promoter is methylated. This coincides with a low-level re-expression of ERa and of pS2 transcripts. Cell-specific patterns of cytosine methylation occur on approximately 60% to 90% of CpG dinucleotides within the somatic genomes of vertebrates6. The pattern of methylation on DNA is maintained through cellular replication, predominantly by the action of the methyltransferase DNMT1, which preferentially acts on hemimethylated CpG substrates. DNMT3a and DNMT3b have been assigned as the cellular de novo methyltransferases because they target unmethylated and hemimethylated substrates with comparable efficiencies7. We and others have shown that ERa cyclically engages with target gene promoters to initiate the sequential recruitment of

MspI CpG –20

0

PPIA pS2

40 20 0 20 40 60 80 100120140160180 Addition of E2 (min) Pull-down

pS2

Input

PPIA

Pull-down Input

80

Control CpG –354 CpG –84 CpG –20

60 40

Input (%)

100

100

60

0 20 40 60 80 100 120140160 180 Doxorubicin release (min)

f

d

80

0

40

+40

Undigested (%)

Input (%)

100

60

20

Undig.

Control

b

PPIA pS2

80 60

pS2

poIII H1 AcH4 ERα

40 20

20

0

0 0 20 40 60 80 100120140160180 Addition of E2 (min)

Figure 1 | The methylation status of the pS2 promoter cycles on release from synchronization. a, A schematic overview of relevant features of the pS2 promoter is shown. NucE and NucT are phased nucleosomes associated with the oestrogen response element (ERE) and TATA box, respectively. The positions of CpG dinucleotides are marked by vertical lines, and those within the sequence CCGG are marked as HpaII sites (filled circles) with the TSS denoted as 11. b, MCF-7 cells were cultured for 72 h in oestrogen-free media. E2 (1 3 1028 M) was then added and genomic DNA subsequently prepared at ten-min intervals. After fragmentation by sonication, DNA fragments containing methylated CpG dinucleotides were selectively isolated using the MBD of human MECP2 fused to GST. The proportion of pS2 and PPIA promoter DNA selectively captured, compared with total input DNA, was monitored by qPCR (b, top panel) and by endpoint PCR (b, bottom panel). c, The methylation status of CCGG sites close to the TSS of the pS2 promoter were then kinetically evaluated using the

0 20 40 60 80 100120140160180 Doxorubicin release (min)

methylation-sensitive restriction enzyme HpaII in conjunction with PCR. After synchronization by release from E2 deprivation, genomic DNA was digested with HpaII or with the methylation-insensitive isoschizomere MspI, and subjected to PCR with primers flanking the enzyme-recognition site encompassing the CpG at position 220 or a control region in the 59 region of the pS2 promoter that contains no HpaII sites. Undig., undigested. d, This procedure was also characterized by qPCR; however, in this case, primers were used that allow the individual HpaII sites at 220, 284 and 2354 bp to be probed. e, f, MCF-7 cells were treated for 1 h with 300 nM doxorubicin, which was then removed by washing and media replacement. Serial samples were processed by the GST–MBD pull-down procedure (e) or by ChIP (f) with the antibodies as indicated. Primers specific for the pS2 and PPIA promoters were used to determine the proportion of promoter DNA captured by each affinity isolation; mean values 6 s.e.m. are shown, n 5 2.

1 European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany. 2SPARTE, UMR CNRS, 6026, Universite´ de Rennes I, Baˆtiment 13, 35042 Cedex, Rennes, France. {Present addresses: Phenex Pharmaceuticals AG, J542N Werksgela¨nde BASF, 67056 Ludwigshafen, Germany (B.S.); Science Foundation Ireland, Wilton Park House, Wilton Place, Dublin 2, Ireland (F.G.). *These authors equally contributed to this work.

112 ©2008 Nature Publishing Group

LETTERS

NATURE | Vol 452 | 6 March 2008

(GRM4, Fig. 2c) and the inwardly rectifying potassium channel J8 (KCNJ8, Fig. 2d)—demonstrate cyclical changes in promoter methylation after promoter synchronization with doxorubicin treatment, as analysed either by qPCR (middle panels) or by endpoint PCR (lower panels). The discovery that periodic changes in the methylation status occur on the pS2 promoter in oestrogen-dependent cells prompted an evaluation of the plasticity of pS2 promoter methylation in cells in which high levels of methylation may contribute to a restriction of gene expression. High levels of pS2 mRNA are present in the ERapositive cell line MCF-7, but not in the ERa-negative cell line MDAMB-231 (Fig. 4a). This correlates with an increase in the proportion of methylated CpG dinucleotides close to the TSS, as determined by bisulphite sequencing (Fig. 3a and Supplementary Fig. 2). Transient exposure to doxorubicin resulted in a reduction of methylation of the proximal pS2 promoter in MDA-MB-231 cells, with hypomethylation occurring approximately 45 min after the addition of doxorubicin (Fig. 3b, left panel). ChIP analysis indicated that demethylation occurs after an exclusion of DNMT1, recruitment of DNMT3a and DNMT3b and the transient recruitment of thymine-DNA glycosylase (TDG; Fig. 3b, right panel). These findings are consistent with events proposed to induce cytosine demethylation13. Promoter release from doxorubicin blockade resulted in cyclical changes in the methylation status of the proximal promoter of the pS2 gene in MDA-MB-231 cells (Fig. 3c, top left panel). Similar demethylation in response to doxorubicin treatment as well as cyclical methylation/demethylation in response to release from blockade were also observed on the pS2 promoter in HeLa cells (Supplementary Fig. 5). Kinetic ChIP analysis indicates that histone H1 remained associated with the pS2 promoter after release from doxorubicin synchronization, although cyclical changes in the acetylation status of histone H4 and low-level recruitment of polII did occur (Fig. 3c, top right panel). Synchronization by a-amanitin14 induces cyclical methylation of the 220 CpG dinucleotide in MCF-7 cells (Fig. 3c, bottom left panel). However, unlike doxorubicin, a-amanitin did not induce demethylation of pS2 a

+1

+95

b

+163 +219 Exon 1

100

Input (%)

Input (%)

+1

+72 +177 +303 Exon 1

100

80 60 40 20

80 60 40 20

0

0

0 20 40 60 80 100 120 140160 180

0 20 40 60 80 100 120 140 160180

Doxorubicin release (min)

Doxorubicin release (min) Pull-down Input

c

–24

d

+1 +161 +170 Exon 1

–166

–42 +1 Exon 1

Input (%)

80 60 40

80 60 40 20

20 0

Pull-down Input

100

100 Input (%)

transcription factor complexes that promote and then limit transcription of responsive genes8–10. Here we show that, similarly, cyclical methylation of promoter DNA occurs on a timescale of tens of minutes. Our initial observations were made on the E2-responsive pS2 promoter, a schematic representation of which is shown in Fig. 1a. A rapid, quantitative, affinity isolation procedure using the methylbinding domain (MBD) of methyl CpG-binding protein 2 (MECP2) was developed to characterize the methylation status of DNA (Supplementary Fig. 1). In brief, genomic DNA was isolated and fragmented by sonication to an average size of approximately 300 bp, and DNA containing symmetrically methylated 5-MeCpG was selectively isolated using the glutathione S-transferase (GST)–MBD fusion protein. The ERa-positive, E2-dependent cell line MCF-7 was maintained in oestrogen-free medium for three days to minimise activity of the pS2 promoter. E2-responsive promoters were then synchronously activated by the addition of E2. Cells were collected at ten-minute intervals for three hours and subjected to the GST– MBD pull-down assay, which was evaluated using primers encompassing the pS2 promoter close to the transcriptional start site (TSS, Fig. 1a). The pS2 promoter is initially unmethylated (Fig. 1b and Supplementary Fig. 2), with two cycles of DNA methylation observed in the three hours after addition of E2, and broad peaks occurring between 20 and 50 min and between 120 and 150 min (Fig. 1b). By contrast, methylation of the peptidylprolyl isomerase A (PPIA) promoter, constitutively expressed in MCF-7 cells, was not observed (Fig. 1b). The methylation-sensitive restriction endonuclease HpaII, which does not cleave hemimethylated or fully methylated C5-MeCGG sequences, was used to evaluate cycling of the CCGG site at 220 relative to the TSS of the pS2 promoter on release from E2 deprivation. Periodic methylation, similar to that observed using the GST– MBD fusion procedure, was detected by endpoint PCR evaluating the methylation status of the CpG sequence at position 220 (Fig. 1c) and by quantitative PCR (qPCR) using primer sets that individually evaluate the methylation status of CpGs at positions 2354, 284 and 220 relative to the TSS (Fig. 1d). We next developed a promoter-synchronization methodology designed to facilitate the evaluation of E2-independent genes. Cells were treated for one hour with 300 nM doxorubicin, which prevents transcription by intercalating into GC-rich regions of DNA and by compromising topological changes required for the function of RNA polymerase II and topoisomerases I and II (ref. 11). Additionally, by intercalating into DNA, doxorubicin perturbs the action of DNMT1 (ref. 12) and potentially may affect the action of other components of the transcriptional machinery. Synchronous release from doxorubicin blockade was then achieved by washout. The methylation status of the 59 proximal pS2 promoter in doxorubicin-treated MCF-7 cells showed cyclic changes with kinetics similar to that seen after release from E2 deprivation (Fig. 1e and Supplementary Fig. 3). Additionally, doxorubicin synchronization induced cyclical changes in the occupancy of the pS2 promoter by ERa, polII and acetylated H4 (AcH4, a marker of transcriptional permissiveness on histones), as determined by chromatin immunoprecipitation (ChIP), albeit with different kinetics to those induced when promoters are synchronized by a-amanitin, an inhibitor of polII (refs 9 and 10). Histone H1, associated with condensed chromatin, was never present on the pS2 promoter in MCF-7 cells. Promoter methylation is coincident with, but persists longer than, the initial productive cycle. A subsequent round of methylation then occurs after the productive wave of the second cycle (Fig. 1f). Doxorubicin promoter synchronization was then used to evaluate proximal promoter methylation of an additional four genes that are highly expressed in MCF-7 cells and that are subject to rapid turnover at the messenger RNA level (Supplementary Fig. 4). Kinetic analysis using the GST–MBD pull-down procedure was performed after doxorubicin blockade and release. All promoters analysed—ERa (Fig. 2a), TFF3 (Fig. 2b), glutamate receptor, metabotropic 4

0 20 40 60 80 100120140160 180

Doxorubicin release (min)

0

0 20 40 60 80 100120140 160180

Doxorubicin release (min) Pull-down Input

Pull-down Input

Figure 2 | Cyclical changes in methylation status occur on promoters of several genes. Serial samples of genomic DNA after synchronization with doxorubicin were subjected to GST–MBD affinity isolation, and the proportion of methylated DNA determined relative to input levels by qPCR (middle panels) and by end-point PCR (lower panels) were measured for: a, ERa 59 transcribed region; b, TFF3 59 transcribed region; c, GRM4 proximal promoter; and d, KCNJ8 proximal promoter. The schemes at the top of each panel indicate the TSS, the occurrence of CpGs, the position of exon 1 and the position of the PCR primers used. 113

©2008 Nature Publishing Group

LETTERS


20

60

15

40

10

20

5

0 0 20 40 60 80 100 120 140 160 180


20 0

80 60 40

b

1.6 1.2


20 0

a

0.8 0.4

ERα

MCF-7

Control ERα

MDA-MB-231

Control

0 pS2

ERa

U

Undigested (%)

MCF-7

40

0 20 40 60 80 100120140160180

Doxorubicin release (min) MDA-MB-231 100

100 60

0

ed

80

80

poIII H1 AcH4

25

0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140 α-amanitin release (min)

Figure 3 | Doxorubicin induces cyclical changes in the methylation status of transcriptionally inactive promoters. a, Bisulphite sequencing of the upper and lower strands of the pS2 proximal promoter in MCF-7 (left) and in MDA-MB-231 (right) cells was performed to determine the proportion of CpG sites methylated at each occurrence of CpG. b, After addition of 300 nM doxorubicin to MDA-MB-231 cells, the methylation status of the pS2 promoter was characterized by the GST–MBD isolation procedure (b, left panel). Serial samples of MDA-MB-231 cells treated with 300 nM doxorubicin were subject to ChIP analysis with the indicated antibodies; mean values 6 s.e.m., n 5 2, are shown (b, right panel). c, The methylation status of the pS2 promoter in MDA-MB-231 cells cycles after release from doxorubicin blockade, characterized by selective purification using the GST–MBD fusion protein (c, top left). AcH4 and, to a lesser extent, polII, cyclically associate with the pS2 promoter on release from doxorubicin synchronization, as determined by ChIP with the indicated antibodies (c, top right; mean values 6 s.e.m., n 5 2 are shown). Whereas synchronous promoter release from a-amanitin treatment induces cycling in MCF-7 cells (c, bottom left), it does not induce either demethylation or cyclical methylation of the pS2 promoter in MDA-MB-231 cells (c, bottom right).

c

SO

30 MDA-MB-231

60

re at

0 0 10 20 30 40 50 30 40 50 60 Time after addition of doxorubicin (min)

MDA-MB-231 pS2

100 Input (%)

20

0

24

48

72

Time after release from doxorubicin (h)

120 100 80 0.2

0 M nt CFre 7 at e DM d SO Do x

c

10

U

0

MDA-MB-231

0

DM

10

20

Per cent expression relative to MCF-7; normalized to PPIA

40

nt

20

60

MDA-MB-231

80

MCF-7

30

100 Input (%)

AcH3 DNMT3a DNMT3b DNMT1 TDG

Lower strand

MCF-7

b

MDA-MB-231

-301 -181 -84 -56 -20 -13 -9 +10 +17

0 20 40 60 80 100

-301 -181 -84 -56 -20 -13 -9 +10 +17

Methylated (%)

MCF-7 100 80 pS2 Upper strand 60 40 20 0

Methylated (%)

a

doxorubicin-induced methylation cycling in MDA-MB-231 cells has on gene expression. There is no detectable expression of ERa or of pS2 in MDA-MB-231 cells (Fig. 4a), and the promoter DNA is extensively methylated (Supplementary Figs 2 and 5); however, lowlevel re-expression of ERa protein (Fig. 4b) and pS2 mRNA (Fig. 4c) occur after exposure to doxorubicin for one hour. Active cytosine demethylation in response to glucocorticoid-induced development that is dependent on topoisomerase activity has been described for the tyrosine aminotransferase gene20. By analogy, transient inhibition of topoisomerase II activity by doxorubicin may induce similar effects on a subset of promoters, indicating that methylation cycling alone is insufficient to provoke robust mRNA expression. This is reflected in the limited re-expression of ERa protein and pS2 mRNA, despite most of promoters in the cell population exhibiting cyclical changes in their methylation status. Kinetic ChIP analysis of the pS2 promoter in MDA-MB-231 cells after doxorubicin synchronization suggests that local chromatin remains predominately associated with histone H1 and is likely to be condensed. By contrast, cyclical changes in the acetylation status of histones H3 (not shown) and H4 occur. The observation that broadly similar kinetics of transitory methylation were obtained using different methodologies to synchronize the pS2 promoter (release from E2 deprivation and from a-amanitin and doxorubicin blockade) and using five different promoters synchronized by doxorubicin treatment (pS2, ERa, TFF3, GRM4 and KCNJ8) may indicate that common sets of events mediate promoter cycling. Further insights into the mechanism and timing of dynamic methylation events on the pS2 promoter in response to E2 and a-amanitin treatment are provided in a related paper13. Transitory modifications in the methylation status of promoters may provide a mechanistic link between transcription and longer term changes in the epigenetic status of promoters. In particular, the finding that doxorubicin can be used to modify the expression of transcribed and quiescent promoters provides a general methodology to analyse the role of cyclical methylation in gene expression.

Expression relative to PPIA

promoters in MCF-7 cells before washout and, in agreement with its inhibitory effect on polII, did not induce cyclical methylation in MDA-MB-231 cells (Fig. 3c, bottom right panel). Transcription generally opposes a repressive environment in which multiple regulatory restrictions have to be overcome before a gene is expressed15–17. Functionally, cyclical changes in promoter CpG methylation may parallel changes that occur to the pattern of covalent modifications to histone tails, known as the histone code, which is a well characterized example of covalent marks that dynamically influence transcription18,19. We evaluated the effect that

0

24 48 72

Time after release from doxorubicin (h)

Figure 4 | ERa and pS2 are re-expressed at low levels after transitory treatment with doxorubicin. a, The relative expression of ERa and pS2 mRNA was monitored by reverse transcription (RT)–qPCR on RNA isolated from MCF-7 and MDA-MB-231 cells, shown as mean values 6 s.d., n 5 3. Levels were normalized to PPIA mRNA, which shows no changes in methylation status or expression under the conditions used. b, ERa, detected by HC-20 (Santa-Cruz) in western blot analysis, shows a small, transient reexpression after release from 1 h treatment with doxorubicin. c, Similarly, pS2 mRNA shows transient, low-level re-expression on release from 1 h doxorubicin blockade; mean values 6 s.d., n 5 3, are shown. For c, MDAMB-231 cells were used.


LETTERS


METHODS SUMMARY Genomic DNA isolation, GST–MBD pull-down assays and ChIP. Genomic DNA was isolated as described21. GST–MBD fusion protein preparation and the GST–MBD pull-down were performed as detailed in the Methods. Quantitative PCR (ABI Prism 7500 System, Applied Biosciences) was used to characterize regions of interest, with primers used reported in Supplementary Table 1 (see Methods). ChIP was performed as described previously10. Methylation-sensitive restriction enzyme analysis and bisulphite sequencing. Restriction endonuclease analysis and bisulphite sequencing were performed as described22. Primer sets used for PCR of digested samples and bisulphite sequencing are reported in Supplementary Table 1. Three biological replicates were used for each bisulphite conversion reaction, with a minimum of 12 clones per cell line, with both DNA strands subjected to sequencing. RT–PCR and qPCR. For reverse transcription PCR, ,1 3 103 cells were lysed in TRIzol reagent (Invitrogen), and RNA isolated according to the manufacturers’ instructions. RNA was reverse transcribed using the Superscript III First-Strand Synthesis System (Invitrogen) and subjected to qPCR (ABI Prism7500 System, Applied Biosciences), with primers shown in Table 1 (see Methods). Full Methods and any associated references are available in the online version of the paper at www.nature.com/nature. Received 19 September 2007; accepted 10 January 2008. 1.

Jones, P. A. & Takai, D. The role of DNA methylation in mammalian epigenetics. Science 293, 1068–1070 (2001). 2. Li, E. Chromatin modification and epigenetic reprogramming in mammalian development. Nature Rev. Genet. 3, 662–673 (2002). 3. Siegfried, Z. et al. DNA methylation represses transcription in vivo. Nature Genet. 22, 203–206 (1999). 4. Costello, J. F. & Plass, C. Methylation matters. J. Med. Genet. 38, 285–303 (2001). 5. Reid, G. et al. Multiple mechanisms induce transcriptional silencing of a subset of genes, including oestrogen receptor a, in response to deacetylase inhibition by valproic acid and trichostatin A. Oncogene 24, 4894–4907 (2005). 6. Ng, H. H. & Bird, A. DNA methylation and chromatin modification. Curr. Opin. Genet. Dev. 9, 158–163 (1999). 7. Duncan, B. K. & Miller, J. H. Mutagenic deamination of cytosine residues in DNA. Nature 287, 560–561 (1980). 8. Shang, Y. et al. Cofactor dynamics and sufficiency in estrogen receptor-regulated transcription. Cell 103, 843–852 (2000). 9. Reid, G. et al. Cyclic, proteasome-mediated turnover of unliganded and liganded ERa on responsive promoters is an integral feature of estrogen signaling. Mol. Cell 11, 695–707 (2003). 10. Metivier, R. et al. Estrogen receptor-a directs ordered, cyclical, and combinatorial recruitment of cofactors on a natural target promoter. Cell 115, 751–763 (2003). 11. Gewirtz, D. A. A critical evaluation of the mechanisms of action proposed for the antitumor effects of the anthracycline antibiotics adriamycin and daunorubicin. Biochem. Pharmacol. 57, 727–741 (1999).

12. Yokochi, T. & Roberson, K. D. Doxorubicin inhibits DNMT1, resulting in conditional apoptosis. Mol. Pharmacol. 66, 1415–1420 (2004). 13. Me´tivier, R. et al. Cyclical DNA methylation of a transcriptionally active promoter. Nature doi:10.1038/nature06544 (this issue). 14. Metivier, R., Reid, G. & Gannon, F. Transcription in four dimensions: nuclear receptor-directed initiation of gene expression. EMBO Rep. 7, 161–167 (2006). 15. Spector, D. L. The dynamics of chromosome organization and gene regulation. Annu. Rev. Biochem. 72, 573–608 (2003). 16. Khorasanizadeh, S. The nucleosome: from genomic organization to genomic regulation. Cell 116, 259–272 (2004). 17. Dillon, N. Gene regulation and large-scale chromatin organization in the nucleus. Chromosome Res. 14, 117–126 (2006). 18. Mellor, J. Dynamic nucleosomes and gene transcription. Trends Genet. 22, 320–329 (2006). 19. Jenuwein, T. & Allis, C. D. Translating the histone code. Science 293, 1074–1080 (2001). 20. Kress, C., Thomassin, H. & Grange, T. Active cytosine demethylation triggered by a nuclear receptor involves DNA strand breaks. Proc. Natl Acad. Sci. USA 103, 11112–11117 (2006). 21. Nelson, J. E. & Krawetz, S. A. Purification of cloned and genomic DNA by guanidine thiocyanate/isobutyl alcohol fractionation. Anal. Biochem. 207, 197–201 (1992). 22. Metivier, R. et al. Transcriptional complexes engaged by apo-estrogen receptor-a isoforms have divergent outcomes. EMBO J. 23, 3653–3666 (2004).

Supplementary Information is linked to the online version of the paper at www.nature.com/nature. Acknowledgements The authors thank L. Cortesi, S. Denger and J. Lewis for their critical reading of this manuscript, and J. Blake for his assistance in expression array analysis. This work was supported by the EC 6th framework programme grant CRESCENDO and by the European Molecular Biology Organisation (EMBO). Author Contributions G.R. and F.G. conceived the study, designed the experimental strategy and analysed data; S.K., B.S. and R.M. contributed to the experimental design. S.K. and B.S. prepared the GST–MBD fusion protein, validated the pull-down method, performed the MBD pull-downs and RT–PCR, and prepared DNA and chromatin for methylation analysis and for ChIP. R.M. performed restriction-sensitive methylation analysis and ChIP. M.P.-S. generated RNA for expression analysis and conducted the western blot studies. D.I., R.P.C. and V.B. produced the bisulphite sequencing results and expression array data from material provided by S.K., B.S. and M.P.-S. G.R., R.M. and F.G. wrote the paper, S.K. prepared the Supplementary Information, and all authors discussed the results and commented on the manuscript. F.G. and G.R. are joint senior authors. Author Information The gene expression data set reported has been deposited in the NCBI Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE10145. Reprints and permissions information is available at www.nature.com/reprints. Correspondence and requests for materials should be addressed to G.R. ([email protected]).


doi:10.1038/nature06640

METHODS Cell culture. MCF-7, MDA-MB-321 and HeLa cells were maintained in DMEM supplemented with 10% fetal calf serum (Invitrogen), 2 mM glutamine (Invitrogen) and penicillin/streptomycin (Invitrogen) at 37 uC and 5% CO2. Cells were free of mycoplasma. Generation of the GST–MBD fusion construct. The MBD of MECP2 (ref. 23) was amplified by PCR from cDNA prepared from MCF-7 cells with the following primers: forward primer 59-GGATCCTCTGCCTCCCCCAAACAGCGGCGCT; reverse primer 59-GAATTCTCATCTCCCAGTTACCGTGAAGTCA. The amplified product was cloned into the PCR2.1 vector (Invitrogen), sequence validated and then subcloned as a BamHI/EcoR1 fragment into the pGEX-6p1 expression vector (Invitrogen). GST and GST–MBD fusion protein preparation. Bacterial cultures were incubated at 20 uC until a D600 nm of between 0.6 and 0.8 was achieved. Protein expression was induced by the addition of 0.4 mM isopropyl b-D-1-thiogalactopyranoside. Bacteria were collected 16 h later by centrifugation and were resuspended in extraction buffer (1 mM DTT (dithiothreitol), 1 mM EDTA, 20 mM Tris (pH 8), 100 mM NaCl and 0.5% Triton-X 100) containing a cocktail of protease inhibitors (Roche), and were lysed by sonication (3 3 1 min on maximum power; Sonifier Cell Disruptor) followed by three passes through a French press (SLM Aminco, SLM instruments) at a pressure of between 500 and 1,000 pounds per square inch. The lysate was clarified by centrifugation at 20,000g for 30 min. Five millilitres of a 50% slurry of glutathione agarose beads (Sigma Aldrich) equilibrated in pulldown buffer (20 mM Tris (pH 8.0), 100 mM NaCl, 1 mM EDTA) were added to the cleared supernatant from 3 litre of culture, with rotation for 1 h at 4 uC. The beads were washed until the eluate had a D280 nm , 0.04. Protein was eluted from the beads by incubation for 30 min at 4 uC in pull-down buffer containing 10 mM glutathione. Glutathione was removed from the protein by extensive dialysis against buffer containing 10 mM Tris (pH 8) and 100 mM NaCl. Purified protein was aliquoted, snap-frozen in liquid nitrogen and stored at 280 uC until use. Preparation of genomic DNA and GST–MBD pull-down experiments. Genomic DNA from ,2 3 106 cells was prepared according to the method described previously21 with small modifications. In brief, cells were lysed in guanidine thiocyanate buffer (3.6 M guanidine thiocyanate, 0.05% SDS, 50 mM Tris, pH 8), the lysate was extracted with phenol/chloroform and then precipitated twice with ethanol. Genomic DNA (20 mg in 200 ml) was then sonicated to an average size of around 300 bp (1 min at maximum power, interval 0.2 using a CosmoBio Bioruptor). Sonicated DNA (5 mg) was then incubated with 5 mg of GST–MBD fusion protein or GST alone in a final volume of 200 ml pull-down buffer. Reactions were incubated for 2 h at 4 uC with constant

rotation. DNA–protein complexes were then captured by the addition of 100 ml of a 50% slurry of glutathione agarose beads, with a further incubation at 4 uC for 1 h. Beads were washed four times with 500 ml of pull-down buffer. Bound GST or GST–MBD–DNA complexes were eluted with 10 mM glutathione in pull-down buffer. This eluate was purified using a Qiaquick PCR Purification Kit (Qiagen). Purified DNA was then used as template in qPCR (ABI Prism 7000 System, Applied Biosciences) to amplify the promoter regions of genes of interest. Primers are shown in Supplementary Table 1. Untreated sonicated DNA was used define the input amount of DNA. PCR with reverse transcription. For PCR with reverse transcription, ,1 3 106 cells were lysed in TRIzol reagent (Invitrogen) and RNA isolated according to the manufacturer’s instructions. RNA was reverse transcribed using Superscript III First-Strand Synthesis System for RT–PCR (Invitrogen) and subjected to PCR with the primers shown in Table 1. PCR products were resolved on a 1% agarose gel. Bisulphite modification and sequencing. Genomic DNA (10 mg) from MCF-7, MDA-MB-231 or HeLa cells was diluted in 50 ml H2O and denatured with 3 M NaOH at 37 uC for 15 min. Hydroquinone (33 ml of 20 mM, Sigma Aldrich) and 530 ml of 4.3 M sodium bisulphite (Sigma Aldrich) were added and samples were incubated at 50 uC for 16 h. Converted DNA was purified using a PCR Purification Kit (Qiagen) and the conversion was completed by adding 5.7 ml of 3 M NaOH followed by incubation at 37 uC for 20 min. DNA was then precipitated and subjected to strand-specific PCR using primers specific for bisulphite-converted material, as shown in Supplementary Table 1. All PCR products were isolated using Montage Gel Extraction Kit (Millipore), and were then precipitated, cloned and sequenced. Three biological replicates were used for each bisulphiteconversion reaction, and 12 clones per cell line were sequenced. Methylation-sensitive restriction enzyme analysis. To analyse the methylation status of individual CpG dinucleotides contained within the sequence CCGG on various promoters, 5 mg of genomic DNA was digested for 2 h at 37 uC with 10 units of the methylation-sensitive restriction enzyme HpaII or its methylation-insensitive isoschitzomere MspI (Roche) (where 1 unit digests 1 mg of plasmid DNA in 1 h). Control samples were undigested. All samples were then precipitated and subjected to qPCR with primer sets that flank one or more individual 59-CCGG-39 sites and/or flank a control site without a HpaII/MspI recognition sequence. Primers are shown in Supplementary Table 1. The digested sample was then compared with the equivalent undigested sample and normalized to the control site. 23. Nan, X. & Meehan, R. R. Dissection of the methyl-CpG binding domain from the chromosomal protein MecP2. Nucleic Acids Res. 21, 4886–4892 (1993).

©2008 Nature Publishing Group