Nature Structural & Molecular Biology
11, 1068 - 1075 (2004)
Published online: 3 October 2004; | doi:10.1038/nsmb840
Intragenic DNA methylation alters chromatin structure and elongation efficiency in mammalian cellsMatthew C Lorincz1, David R Dickerson1, Mike Schmitt2
& Mark Groudine1, 21 Division of Basic Sciences, Fred Hutchinson Cancer Research Center, Seattle, Washington 98109, USA. 2 Department of Radiation Oncology, University of Washington School of Medicine, Seattle, Washington 98195, USA.
Correspondence should be addressed to Matthew C Lorincz mlorincz@fhcrc.orgTranscriptional silencing in mammals is often associated with promoter methylation. However, a considerable number of genomic methylated CpGs exist in transposable elements, which are frequently found in intronic regions. To determine whether intragenic methylation influences transcription efficiency, we used the Cre/loxP-based system, RMCE, to introduce a transgene, methylated exclusively in a region downstream of the promoter, into a specific genomic site. This methylation pattern was maintained in vivo, and yielded a clear decrease in transgene expression relative to an unmethylated control. Notably, RNA polymerase II (Pol II) was depleted exclusively in the methylated region, as was histone H3 di- and trimethylated on Lys4 and acetylated on Lys9 and Lys14. As the methylated region adopts a closed chromatin structure in vivo, we propose that dense intragenic DNA methylation in mammalian cells initiates formation of a chromatin structure that reduces the efficiency of Pol II elongation.The CpG dinucleotide is generally found at a lower than expected frequency in the mammalian genome, with the exception of the (G+C)-rich regions known as CpG islands, which have the statistically expected frequency of CpGs1. Analysis of the distribution of DNA methylation reveals that, whereas the majority of cytosines in the context of the CpG dinucleotide are methylated in normal adult somatic tissues, CpG islands, which encompass the promoters of about half of mammalian genes, typically remain methylation-free2. In contrast, transposable elements such as SINEs (short interspersed nuclear elements), LINEs (long interspersed nuclear elements) and endogenous retroviruses, which range in size from  0.3 to 10 kilobases (kb) and comprise >45% of the human genome3, are densely methylated in such tissues. These elements harbor a substantial proportion, if not the majority, of methylated CpGs in the genome4 and are frequently found in intronic regions3,
5,
6,
7. Thus, intragenic CpG methylation is generally found not in the promoter, but in the downstream transcribed region, where it may serve to prevent the activation of transposable elements.
Nevertheless, the majority of studies aimed at determining the influence of DNA methylation on transcription have focused on the methylation state of promoter regions, primarily because the CpG island promoters of tumor suppressor genes are aberrantly methylated in cancer. Such studies show a strong correlation between promoter methylation and transcriptional silencing. The observation that members of the methyl-DNA-binding domain (MBD) family of proteins, such as MeCP2, interact with histone deacetylase (HDAC)-containing complexes8,
9 indicates that DNA methylation may suppress transcription via deacetylation of histones. Indeed, methylated DNA is associated with hypoacetylated chromatin in vivo10,
11.
Analyses of methylated and unmethylated regions in the genome using endonucleases such as DNaseI and MspI clearly demonstrate that the former are relatively refractory to digestion, indicating a closed chromatin structure12,
13. Furthermore, microinjection experiments reveal that methylation-dependent transcriptional repression occurs by an indirect mechanism involving the time-dependent assembly of a repressive chromatin structure in murine cells14. Previously, we used the Cre/loxP-based recombinase-mediated cassette exchange (RMCE) system15 to target a transgene, unmethylated or in vitro−methylated at all CpG sites, into a specific genomic site and observed a methylation-dependent loss of DNaseI hypersensitivity and Pol II loading at the methylated transgene promoter16. Taken together, these results are consistent with a model in which DNA methylation triggers formation of a hypoacetylated, closed chromatin structure that precludes binding of critical transcription factors, and in turn, recruitment of Pol II to the promoter, ultimately blocking transcriptional initiation.
Analyses of episomal and transiently transfected constructs reveal that methylation downstream of an active promoter can also negatively influence transcription17,
18. Although transcriptional elongation is completely blocked by intragenic methylation in the filamentous fungus Neurospora crassa19, intragenic methylation clearly does not block elongation of endogenous genes in mammalian cells. DMR2, the intronic imprinting center of the mouse Igf2r gene, for example, is methylated on the active maternal allele20. However, it has yet to be determined whether the efficiency of transcriptional elongation is decreased by such methylation. The Igf2r gene (and imprinted genes in general) is not informative in this regard, because the paternal allele, although unmethylated in the DMR2 region, is transcriptionally silenced. An alternative approach is to compare the expression state of the same gene in different cell lines with distinct intragenic methylation patterns21,
22 in the downstream transcribed region21,
22. Expression analysis of several cancer cell lines showing different patterns of intragenic methylation in the endogenous INK4a-ARF locus clearly reveal that both the p16INK4a and p14ARF genes are active, regardless of the presence of intragenic methylation21. However, it is difficult to establish the influence of intragenic methylation on elongation efficiency by comparing transcription levels between different cell lines, as a number of factors may influence transcription in different cellular backgrounds. Furthermore, as endogenous genes typically contain heterogeneous levels of methylation, it is difficult to draw conclusions regarding the influence of methylation on expression, particularly in the absence of single-cell expression assays. Finally, although cell lines deficient in each of the DNA methyltransferases are available, the pleitropic effects associated with global DNA methylation deficiency preclude the use of such lines in analyzing the influence of methylation at specific loci.
To circumvent the complications discussed above, we used the RMCE system to target a transgene encoding the green fluorescent protein (GFP), either unmethylated or methylated exclusively in the intragenic region, to a specific genomic site. We show that the methylation pattern introduced in vitro is stably maintained in vivo, and reduces expression from the transgene relative to the unmethylated control integrated at the same site. This decrease in expression is accompanied by a reduction in the density of Pol II and the abundance of histone H3 acetylated on Lys9 and Lys14 and methylated on Lys4, exclusively in the methylated region. Based on these results, and parallel chromatin accessibility studies, we propose that DNA methylation in the transcribed region induces formation of a closed chromatin structure that reduces the efficiency of Pol II elongation.
Results
Intragenic methylation decreases expression
To determine whether DNA methylation in the intragenic region of a transcriptionally active gene influences Pol II elongation efficiency, we introduced a transgene containing the human p16INK4A promoter driving expression of 'humanized' GFP, either unmethylated or 'patch'-methylated in vitro exclusively in the intragenic region (Fig. 1a), into the RL5 integration site in murine erythroleukemia (MEL) cells by RMCE15,
16 (Fig. 1b). We chose to use the human p16INK4A gene promoter as a model, because the Alu repeat just 3' of this CpG island promoter (1 kb downstream of the transcription start region) is methylated in primary fibroblasts and mammary epithelial cells (data not shown). However, we replaced the CpG-rich Alu repeat with the CpG-rich, 'humanized' GFP gene so that we could study expression of the transgene in single cells. Ganciclovir-resistant subclones harboring the methylated cassette uniformly expressed GFP, as measured by flow cytometry (Fig. 2a). However, the median GFP expression level was 60% of that detected in subclones harboring the unmethylated cassette at the same integration site (Fig. 2b). This difference in expression level was consistently observed over several months in culture (data not shown) and was confirmed by quantitative RT-PCR of representative methylated (3M) and control unmethylated (4U) clones (Fig. 2c,d), using transgene-specific primers flanking the reporter gene intron (Fig. 1b). As the 5' primer we used anneals to the region just downstream of the p16 promoter start sites, the difference in expression from the methylated and unmethylated clones must be attributed to a difference in transcription from the p16 promoter itself, rather than the methylation-dependent silencing of an alternative or cryptic downstream promoter in the methylated clone.
 | | |
| | |
To establish the methylation status in vivo of the initially methylated and unmethylated cassettes, Southern analysis of both classes of clones was conducted using a GFP probe and genomic DNA digested with BamHI alone or in combination with the methylation-sensitive restriction enzyme HpaII (Fig. 3a). All clones digested with BamHI alone yielded a single band of 8 kb, indicative of successful integration in the RL5 site. In contrast, digestion with BamHI and HpaII yielded bands of 1.6 kb and 0.5 kb for the methylated and unmethylated clones, respectively (Fig. 3b). The former is indicative of maintenance of the initial methylation pattern, whereas the later reveals the absence of de novo methylation of the initially unmethylated cassette.
| | |
To characterize the methylation status of the methylated and unmethylated cassettes in greater detail, clones 3M and 4U were analyzed by bisulfite sequencing, using primers specific for the promoter, the 'methylation junction' and the downstream in vitro−methylated regions (Fig. 3c). Sequencing of cloned amplification products revealed that 'alleles' of the unmethylated clone remained virtually methylation free, whereas 'alleles' of the methylated clone retained the initial methylation state at the majority of CpG sites and showed no upstream spreading of methylation. Similar results were found for another methylated clone (data not shown). These data indicate that intragenic methylation is faithfully maintained in the actively transcribing transgene and that the associated decrease in expression occurs despite the absence of methylation within 1 kb of the promoter, raising the question: does this decrease in expression reflect a decrease in the efficiency of Pol II initiation or elongation?
Pol II density is lower in the methylated region
To determine whether intragenic DNA methylation affects Pol II recruitment to the p16 promoter at a distance, or the density of elongating Pol II in the methylated region itself, nuclear run-on analyses were conducted using probes specific for the promoter-proximal, intronic, and GFP regions (Fig. 4a). Comparison of the hybridization signal generated in the parent cell line (RL5) with that for clones 3M and 4U reveals a relatively strong transgene specific signal in the transcription start region (Fig. 4b). Analysis of four independent experiments revealed no difference in the transgene-specific signal between clones 3M and 4U in this region (Fig. 4c), suggesting that the methylation-associated decrease in expression does not reflect a decrease in the efficiency of Pol II initiation or promoter clearance in the 5' region of the cassette. In contrast, analysis of the intronic and GFP regions reveals no significant signal above background in clones 3M or 4U. The relatively high signal in the intronic region in the RL5 cell line indicates that this probe anneals to endogenous run-on products, whereas the weak signal in the GFP region reveals that less polymerase is traversing the downstream region than the promoter-proximal region, as we have observed in several endogenous genes23. Thus, the run-on assay is informative only for the promoter-proximal region of the transgene.
| | |
Given the ambiguous results obtained for the intronic and GFP regions in the run-on analyses, we chose to determine Pol II density across the transgene using an alternative experimental approach, namely chromatin immunoprecipitation (ChIP). Chromatin was isolated from clones 3M and 4U and ChIP analyses were carried out using antiserum specific for the N terminus of Pol II (three independent experiments were conducted in this and all subsequent ChIP experiments described here). Primers were designed to amplify the p16 promoter, intronic and GFP regions, the latter of which is in the methylated domain (Fig. 5a). PCR analysis of genomic DNA isolated from the RL5 line gave no amplification product with these primers, indicating that they are transgene-specific (data not shown). Quantitative duplex PCR was carried out using each of these primer pairs in combination with primers specific for the endogenous pancreatic amylase 2.1y (amy2.1) gene, which is transcriptionally silent in MEL cells. Relative to the amy2.1 internal control, no consistent difference in Pol II enrichment was found in the promoter or the intronic regions of the methylated versus the unmethylated cassette (Fig. 5b). These results indicate that polymerase loading is not affected by DNA methylation in the downstream region, consistent with results obtained by run-on analysis. In contrast, enrichment of Pol II in the GFP region was approximately two-fold greater in the unmethylated cassette than in the methylated cassette, consistent with the decrease in steady-state RNA in the latter clone (Fig. 2).
| | |
To determine how methylation in the downstream region affects the elongating form of Pol II, ChIP analyses were also conducted using a monoclonal antibody specific for the elongation-competent form of this polymerase, namely that phosphorylated at Ser5 (pSer5) of the C-terminal domain (CTD). Consistent with the Pol II ChIP experiments, no substantial difference in pSer5 Pol II was detected in the promoter or intronic regions (Fig. 5c), whereas the unmethylated clone consistently showed a higher level of pSer5 Pol II enrichment in the downstream methylated region than the methylated clone. Taken together, these results indicate that downstream DNA methylation exerts a negative effect on transcription at the stage of elongation through the methylated region itself, rather than an earlier step in transcription, such as PIC formation or promoter clearance.
Intragenic methylation alters chromatin structure
The N-terminal tails of histones associated with transcriptionally active genes are generally marked by a distinct set of covalent modifications, indicating that such modifications play a fundamental role in the processes of transcriptional initiation and/or elongation. As histone H3 di- or trimethylated on Lys4 is preferentially localized to promoter and promoter-proximal regions of transcriptionally active genes in higher eukaryotes24, we next determined whether DNA methylation affects these modifications locally and/or at a distance, using antisera specific for the Lys4 di- or trimethylated forms of H3. Three independent ChIP experiments were conducted using the same patch-methylated and control unmethylated clones used for the Pol II ChIP analyses (Fig. 6a,b). No difference in enrichment of these modified forms of H3 was detected in the promoter or intronic regions of the methylated transgene relative to the unmethylated transgene. In contrast, a two- to three-fold higher level of enrichment of both di- and trimethyl H3-Lys4 was consistently observed in the GFP region of the unmethylated relative to the methylated transgene, consistent with the relative enrichment of Pol II in this region.
| | |
Using a 'ChIP-CHIP' approach, we previously showed that H3 methylated on Lys79 marks transcriptionally active genes in higher eukaryotes25. To determine whether DNA methylation also influences covalent modification of this core residue, ChIP analyses were conducted using antiserum specific for this modification. A substantial enrichment of H3 dimethyl Lys79 in the downstream region of the transgene was detected, relative to the endogenous amy2.1 gene (Fig. 6c). However, in contrast to the results for H3-Lys4 methylation, H3 dimethyl Lys79 enrichment was no higher in the unmethylated clone than in the methylated clone in this region, indicating that the enzymes responsible for methylating these H3 residues are not influenced by DNA methylation in the same way.
Acetylation of lysines in the N termini of core histones is closely linked with transcriptional activation. Conversely, HDAC complexes targeted to methylated promoters maintain these residues in a hypoacetylated state, inhibiting transcription in the process10,
11. Acetylation of histone tails is of particular interest, as this modification neutralizes the positive charge on lysines and, in so doing, directs an allosteric change in nucleosome conformation that renders nucleosomal DNA more accessible to the transcription machinery26,
27. ChIP analysis using antiserum raised against an H3 peptide acetylated on Lys9 and Lys14 revealed a difference in enrichment of this modification exclusively in the GFP region, with the unmethylated clone showing approximately four-fold greater enrichment than the methylated clone in three independent experiments (Fig. 6d). Consistent with the results for H3-Lys4 methylation, these results indicate that the apparent reduction of Pol II traversing the methylated region of the transgene is associated with a distinct pattern of histone modifications exclusively in this region.
To assess directly whether DNA methylation in the body of the transcriptionally active transgene generates an altered chromatin structure, we analyzed the sensitivity of the promoter and GFP regions of the transgene in clones 3M and 4U using the MspI 'nuclear accessibility' assay13. Although MspI is not sensitive to CCGG methylation (where the underlined C is methylated) in vitro, methylated CCGGs are refractory to digestion with this enzyme in intact nuclei, presumably as a consequence of the methylation-dependent formation of a closed chromatin structure13. Primer sets were designed to amplify the promoter and CpG methylated regions of the transgene (Fig. 7a), each encompassing three MspI sites and a region in the endogenous amy2.1 gene lacking MspI sites. Duplex PCR analysis of genomic DNA isolated from MspI-digested nuclei isolated from clones 3M and 4U revealed no difference in sensitivity in the p16 promoter region of the transgene, relative to amy2.1 (Fig. 7b,c). In contrast, analysis of the downstream region revealed that the methylated transgene was substantially less sensitive to MspI digestion than the unmethylated transgene integrated at the same site. These results directly demonstrate that a patch of dense DNA methylation in an actively transcribed gene is sufficient to induce formation of a region of relatively compact chromatin and are consistent with the model that intragenic methylation reduces the efficiency of Pol II processivity and/or promotes premature dissociation of the polymerase.
| | |
Discussion
Using an integrated p16 promoter-GFP reporter transgene, either unmethylated or methylated exclusively in the downstream transcribed region, we observed an intragenic methylation-dependent decrease in elongation efficiency. Although the GFP gene is of foreign origin, several members of the MBD family of proteins have been shown to bind to methylated CpGs irrespective of sequence context outside of the CpG dinucleotide28,
29. Moreover, the GC content of transposable elements varies widely. Thus, the results reported here may be generally applicable to intragenic regions of dense DNA methylation.
The decreased level of engaged Pol II in the CpG methylated region coincided with a decrease in di- and trimethylated H3-Lys4 exclusively in this region. Recent genetic and biochemical studies in Saccharomyces cerevisiae reveal that the H3-Lys4 histone methyltransferase (HMTase) Set1 interacts with pSer5 Pol II and the pattern of H3-Lys4 tri-methylation strongly correlates with Set1 occupancy30. Thus, a simple explanation for our observation of a DNA methylation− dependent reduction in the level of H3-Lys4 methylation is that this mark depends on the local concentration of the elongating pSer5 Pol II−H3-Lys4 MTase complex. Given the presumed stability of histone methylation marks31, it would seem unlikely that a difference in Pol II−associated H3-Lys4 HMTase density would yield a difference in H3-Lys4 methylation in the methylated region of the transgene. However, our recent genome-wide analysis of histone modifications revealed a substantial correlation between the level of transcription and H3-Lys4 dimethylation in Drosophila melanogaster25, suggesting that the degree of this histone mark correlates with Pol II density. In contrast, H3 dimethyl Lys79 enrichment in the GFP region was no higher in the unmethylated clone than in the methylated clone. In S. cerevisiae, the activities of the HMTases specific for Lys4 and Lys79 depend on conjugation of ubiquitin to histone H2B32. Assuming that H2B ubiquitination is also required for the activity of these HMTases in mammalian cells, our results suggest that the effect of DNA methylation on covalent modification of Lys4 occurs downstream of H2B ubiquitination.
Although H3-Lys4 di- and trimethylation marks are associated with transcription in higher eukaryotes24, the role that these modifications play in the process of elongation remains to be determined. Recently, a human H3-Lys4 HMTase, termed Set9, was isolated and shown to potentiate transcriptional activation33. Using an in vitro assay, the authors showed that methylation of H3-Lys4 precludes binding of the HDAC-containing NuRD complex to the H3 tail, indicating that methylation of H3-Lys4 may promote transcription by inhibiting deacetylation of histone tails associated with actively transcribing genes. Consistent with this model, our experiments reveal that the reduction in H3-Lys4 methylation in the methylated intragenic region is accompanied by a decrease in acetylation of the N-terminal tail histone H3. Alternatively, recruitment of MBDs and associated HDAC complexes may be responsible for the decrease in acetylation in this region8,
9,
10,
11.
Several groups have shown that in higher eukaryotes, H3-Lys9 is dimethylated in euchromatic regions showing hypoacetylation of this residue. However, using several antibodies specific for H3 dimethyl Lys9, we did not detect any enrichment of this histone modification in the CpG-methylated region of the transgene, indicating that the effect on transcription may not depend on methylation of this residue. In S. cerevisiae, which lacks H3-Lys9 dimethylation altogether, hypoacetylation of histone H3 in coding regions has a more marked effect on transcription than does hypoacetylation in promoter regions34. Based on this observation, and the fact that purified, hyperacetylated nucleosomes support a higher rate of transcription than do nonmodified nucleosomes35, we propose that histone hypoacetylation, as triggered by dense DNA methylation, is responsible for the decreased efficiency of elongation in the intragenic region of the transgene studied here.
How might hypoacetylated of chromatin influence elongation efficiency? In vitro, acetylated nucleosomes show a lower level of compaction than do their native nonacetylated counterparts36, and are required to maintain the unfolded nucleosome structure associated with transcribing DNA37. Furthermore, acetylation of histone tails renders nucleosomal DNA more accessible to the transcription machinery26,
27. The hypoacetylated intragenic region of the transgene studied here adopts a closed chromatin structure in vivo, as do hypoacetylated intragenic regions of the endogenous INK4a-ARF locus21. Given that elongating Pol II is associated with complexes that include histone acetyltransferases38,
39, and a distinct set of proteins that disrupt nucleosomal structure40,
41, it has been suggested that acetylation of histone tails influences elongation efficiency by facilitating disruption of the nucleosome in front of the elongating holoenzyme42. If acetylation of histone tails in intragenic regions serves a similar purpose in mammals, then it is not surprising that the hypoacetylated and compact chromatin structure induced by DNA methylation leads to a reduction in elongation efficiency. Such a model predicts that a 'pileup' of Pol II might be observed upstream of the methylated region. However, it is also possible that decreased elongation efficiency through this region results in premature dissociation of Pol II from the transgene template, as has been suggested for the inefficiently elongated (A+T)-rich open reading frame (ORF) region of LINE-1 elements43. Because, relative to the unmethylated control cassette, we do not observe an increased density of polymerase 5' of the methylated region, we favor the latter model.
In higher eukaryotes, Lys4 methylation as well as Lys9 and/or Lys14 acetylation of histone H3 are typically detected exclusively within 1 kb of the transcription start sites of active genes44. Presumably, such marks play an important role in the early stages of elongation but are dispensable for elongation in distal regions. Dense CpG methylation proximal to the transcription start site may perturb this pattern of histone modifications and thereby negatively affect elongation efficiency. Notably, promoter-associated CpG islands often extend several hundred bases beyond the recognized transcription start sites. As the entire length of CpG islands is typically methylation-free, this arrangement may serve to protect genes from the inhibitory effect on elongation imparted by promoter-proximal DNA methylation. Determining whether proximity of a methylated region to the promoter is inversely correlated with the degree to which elongation is adversely affected is clearly of interest.
It has been suggested that the majority of DNA methylation in the mammalian genome is found in transposable elements, and serves to maintain such 'parasitic' elements in a silent state45. Indeed, the promoter regions of SINEs, LINEs and endogenous retroviruses are frequently methylated in somatic cells. Moreover, gene expression in the human genome is inversely correlated with intronic LINE element density46. Our results provide a link between these observations: intragenic DNA methylation may decrease the efficiency of elongation by inducing formation of a compact chromatin structure. The ORF region of LINE-1 elements, which was recently shown to impede transcriptional elongation43 (presumably by a methylation-independent mechanism), also may do so by altering chromatin structure.
As transposable elements are found in the introns of many genes3,
5, mechanisms have probably evolved to suppress the repressive effects of transcribing through densely methylated chromatin. However, for some genes, such as those involved in pattern formation, even minor changes in expression level may not be tolerated: the four homeobox gene clusters, for example, are nearly devoid of repeats3. In fact, whereas Alu repeats occupy 12.5% of total intragenic sequence in the human genome, 12% of intron-containing human genes are entirely free of these elements47, perhaps because the presence of such elements in these genes is deleterious.
Over time, spontaneous deamination of 5-methyl cytosine to thymidine, a relatively frequent mutagenic event in mammalian cells, may reduce the burden of intragenic methylation on elongation efficiency by purging intragenic regions of the methylated cytosines themselves. For genes harboring young, CpG-rich transposons, an antagonistic effect of intragenic methylation on elongation efficiency is presumably preferable to the alternative, namely activation and retrotransposition of these parasitic elements.
Methods
Generation and in vitro methylation of the L1-p16prohGFP-1L plasmid.
To generate the construct L1-p16prohGFP-1L, 835 bp of the 5' end of the gene encoding human p16INK4A (including the CpG island promoter and exon 1) was amplified by PCR and cloned into the L1-hGFP-1L reporter construct. To generate the patch-methylated cassette, this vector was digested with BglII and BamHI, generating two fragments: one containing the promoter and 1 kb of transcribed sequence, and the other the downstream transcribed region of the transgene (including the gene encoding GFP) and the plasmid backbone. The latter fragment was methylated in vitro with SssI and ligated to the unmethylated promoter fragment, generating a plasmid in which the transcription start sites of the p16 promoter are 1.0 kb from the 5' end of the methylated region. To determine that the methylation reaction was carried to completion, after organic extraction and ethanol precipitation, methylated DNA was digested with the methylation-sensitive enzyme HpaII and visualized by electrophoresis on an agarose gel (data not shown).
Tissue culture and RMCE.
Unmethylated and patch-methylated plasmid was introduced into RL5 MEL cells15 by electroporation, and selected in the presence of ganciclovir as described16.
Flow cytometry and RT-PCR.
Cells cultured at low density were processed as described for FACS analyses11. GFP+, ganciclovir-resistant cells were isolated by FACS, cloned by limiting dilution and screened for successful Cre-mediated exchange by Southern blotting. FACS data were analyzed using FlowJo (Treestar). Total RNA was isolated using the RNeasy kit (Qiagen) following the manufacturer's protocol and RT-PCR was carried out as described11. Primers specific for the human p16 promoter and the gene encoding GFP, which span the transgene intron, were used in a duplex PCR reaction with primers specific for the endogenous  -actin gene as an internal control (see Supplementary Table 1 online for all primer sequences used in this manuscript).
Southern blotting and bisulfite analysis.
Preparation of genomic DNA, restriction digests, membrane transfers and preparation of the DNA probe were done as described11. Clones harboring a single-copy integrant at the RL5 integration site were identified by digestion of genomic DNA (isolated at day 21 post-RMCE) with BamHI, which cuts once in the MFGhGFP provirus, followed by Southern blot analysis using the indirect end-labeling technique with the GFP probe. Methylation status of the provirus was determined by digestion of genomic DNA with BamHI alone or in combination with the methylation-sensitive enzyme HpaII. For bisulfite analysis, genomic DNA from MEL cell clones was digested with BamHI, purified by phenol-chloroform extraction, denatured and treated with bisulfite, as described48.
Nuclear run-on analysis.
Preparation of nuclei and nuclear run-on assays were done in the presence of low salt and 0.6% (v/v) sarkosyl, as described23, using [ -32P]UTP (NEN) as the label. Plasmid probes were generated by cloning amplification products of the promoter-proximal (449 bp), intronic (273 bp), and GFP regions (377 bp), into the pGEM-T Easy vector (Promega). A -actin amplification product (285 bp) was used as a control for the efficiency of the run-on reaction in each sample. Hybridization signal was quantified with a phosphorimager and ImageQuant software (Molecular Dynamics). Relative run-on signal in the p16 promoter region was determined by (i) subtracting vector signal from each sample (ii) normalizing to the -actin signal from the RL5 run-on sample (iii) subtracting the RL5 run-on signal from the transgene run-on signal for each probe, and (iv) taking the ratio of the normalized values for 3M and 4U in each experiment.
ChIP analysis.
To generate crosslinked chromatin for ChIPs with antibodies recognizing acetyl histone H3 Lys9 and/or Lys14 (Upstate Biotechnology), dimethyl-Lys4 (Upstate Biotechnology), trimethyl-Lys4 (Abcam) or dimethyl-Lys79 (provided by D. Gottschling, Fred Hutchinson Cancer Research Center), 4  107 exponentially growing MEL cells were incubated in the presence of 1% (v/v) formaldehyde for 40 min at 4 °C, and ChIP was conducted, with minor modifications, as described49. To generate crosslinked chromatin for ChIPs with antibodies recognizing Pol II, 1 108 exponentially growing MEL cells were incubated in the presence of 1% (v/v) formaldehyde for 40 min at 4 °C, and ChIP was conducted, as described50. Sonicated chromatin fragments typically ranged in size from 0.5 to 1 kb. Antibodies used include a polyclonal raised against the N terminus of the large subunit of Pol II (SC-899; Santa Cruz Biotechnology) and a monoclonal raised against the form of Pol II phosphorylated on Ser5 of the CTD (Covance).
Quantitative duplex PCR was carried out using a Perkin Elmer 9700 thermocycler, as described11. Conditions of linear amplification were determined empirically for all primer combinations. Each 25- l reaction was supplemented with 1 Ci of [-32P]dCTP (NEN). The reaction product was subject to electrophoresis on a 5% (v/v) nondenaturing polyacrylamide gel. Amplification products were quantified as described for the run-on analyses. To determine 'fold-enrichment' values for a given region in the cassette, the ratio of the two PCR products (transgene/control) was calculated for the antibody-bound fraction and normalized to the ratio obtained from the input material. Relative enrichment values were subsequently calculated by taking the ratio of the fold-enrichment values of the unmethylated/methylated samples in each of the regions analyzed.
Nuclease accessibility assay.
MspI digestion of 5 107 nuclei was carried out as described21. Briefly, nuclei were resuspended in 600 l of REB (NEB buffer II, 146 mM sucrose) and 10-l aliquots were digested with 0, 5, 10, 20, 30 or 50 units of MspI (NEB) for 30 min at 37 °C. Reactions were terminated by addition of 2 stop buffer (20 mM Tris, pH 7.4, 1% (w/v) SDS, 500 g ml-1 proteinase K) and incubated overnight at 55 °C. DNA was purified by phenol-chloroform extraction and resuspended in TE (40 l). Duplex PCR was conducted using 2 l of DNA and primers specific for the endogenous amy2.1 gene, in combination with primers specific for the transgene. Amplification products were quantified and normalized to the amy2.1 control, as described for the ChIP analyses.
Note: Supplementary information is available on the Nature Structural & Molecular Biology website.
Received 4 June 2004; Accepted 30 August 2004; Published online: 3 October 2004.
REFERENCES
- Takai, D. & Jones, P.A. Comprehensive analysis of CpG islands in human chromosomes 21 and 22. Proc. Natl. Acad. Sci. USA 99, 3740–3745 (2002). | Article | PubMed | ChemPort |
- Bird, A., Taggart, M., Frommer, M., Miller, O.J. & Macleod, D. A fraction of the mouse genome that is derived from islands of nonmethylated, CpG-rich DNA. Cell 40, 91–99 (1985). | Article | PubMed | ISI | ChemPort |
- Lander, E.S. et al. Initial sequencing and analysis of the human genome. Nature 409, 860–921 (2001). | Article | PubMed | ISI | ChemPort |
- Yoder, J.A., Walsh, C.P. & Bestor, T.H. Cytosine methylation and the ecology of intragenomic parasites. Trends Genet. 13, 335–340 (1997). | Article | PubMed | ISI | ChemPort |
- Smit, A.F. Interspersed repeats and other mementos of transposable elements in mammalian genomes. Curr. Opin. Genet. Dev. 9, 657–663 (1999). | Article | PubMed | ISI | ChemPort |
- Woodcock, D.M., Lawler, C.B., Linsenmeyer, M.E., Doherty, J.P. & Warren, W.D. Asymmetric methylation in the hypermethylated CpG promoter region of the human L1 retrotransposon. J. Biol. Chem. 272, 7810–7816 (1997). | Article | PubMed | ISI | ChemPort |
- Szak, S.T. et al. Molecular archeology of L1 insertions in the human genome. Genome Biol. 3, 0052–0058 (2002). | Article |
- Nan, X. et al. Transcriptional repression by the methyl-CpG-binding protein MeCP2 involves a histone deacetylase complex. Nature 393, 386–389 (1998). | Article | PubMed | ISI | ChemPort |
- Ng, H.H. et al. MBD2 is a transcriptional repressor belonging to the MeCP1 histone deacetylase complex. Nat. Genet. 23, 58–61 (1999). | Article | PubMed | ISI | ChemPort |
- Eden, S., Hashimshony, T., Keshet, I., Cedar, H. & Thorne, A.W. DNA methylation models histone acetylation. Nature 394, 842 (1998). | Article | PubMed | ISI | ChemPort |
- Lorincz, M.C., Schübeler, D. & Groudine, M. Methylation-mediated proviral silencing is associated with MeCP2 recruitment and localized histone H3 deacetylation. Mol. Cell. Biol. 21, 7913–7922 (2001). | Article | PubMed | ISI | ChemPort |
- Groudine, M., Eisenman, R. & Weintraub, H. Chromatin structure of endogenous retroviral genes and activation by an inhibitor of DNA methylation. Nature 292, 311–317 (1981). | Article | PubMed | ISI | ChemPort |
- Antequera, F., Macleod, D. & Bird, A.P. Specific protection of methylated CpGs in mammalian nuclei. Cell 58, 509–517 (1989). | Article | PubMed | ISI | ChemPort |
- Buschhausen, G., Wittig, B., Graessmann, M. & Graessmann, A. Chromatin structure is required to block transcription of the methylated herpes simplex virus thymidine kinase gene. Proc. Natl. Acad. Sci. USA 84, 1177–1181 (1987). | PubMed | ChemPort |
- Feng, Y.Q. et al. Site-specific chromosomal integration in mammalian cells: highly efficient CRE recombinase-mediated cassette exchange. J. Mol. Biol. 292, 779–785 (1999). | Article | PubMed | ISI | ChemPort |
- Schübeler, D. et al. Genomic targeting of methylated DNA: influence of methylation on transcription, replication, chromatin structure, and histone acetylation. Mol. Cell. Biol. 20, 9103–9112 (2000). | Article | PubMed | ISI | ChemPort |
- Hsieh, C.L. Stability of patch methylation and its impact in regions of transcriptional initiation and elongation. Mol. Cell. Biol. 17, 5897–5904 (1997). | PubMed | ISI | ChemPort |
- Keshet, I., Yisraeli, J. & Cedar, H. Effect of regional DNA methylation on gene expression. Proc. Natl. Acad. Sci. USA 82, 2560–2564 (1985). | PubMed | ChemPort |
- Rountree, M.R. & Selker, E.U. DNA methylation inhibits elongation but not initiation of transcription in Neurospora crassa. Genes Dev. 11, 2383–2395 (1997). | PubMed | ISI | ChemPort |
- Stoger, R. et al. Maternal-specific methylation of the imprinted mouse Igf2r locus identifies the expressed locus as carrying the imprinting signal. Cell 73, 61–71 (1993). | Article | PubMed | ISI | ChemPort |
- Nguyen, C.T., Gonzales, F.A. & Jones, P.A. Altered chromatin structure associated with methylation-induced gene silencing in cancer cells: correlation of accessibility, methylation, MeCP2 binding and acetylation. Nucleic Acids Res. 29, 4598–4606 (2001). | Article | PubMed | ISI | ChemPort |
- Pao, M.M. et al. The endothelin receptor B (EDNRB) promoter displays heterogeneous, site specific methylation patterns in normal and tumor cells. Hum. Mol. Genet. 10, 903–910 (2001). | Article | PubMed | ISI | ChemPort |
- Krumm, A., Hickey, L.B. & Groudine, M. Promoter-proximal pausing of RNA polymerase II defines a general rate-limiting step after transcription initiation. Genes Dev. 9, 559–572 (1995). | PubMed | ISI | ChemPort |
- Schneider, R. et al. Histone H3 lysine 4 methylation patterns in higher eukaryotic genes. Nat. Cell Biol. 6, 73–77 (2003). | Article | PubMed | ISI | ChemPort |
- Schübeler, D. et al. The histone modification pattern of active genes revealed through genome-wide chromatin analysis of a higher eukaryote. Genes Dev. 18, 1263–1271 (2004). | Article | PubMed | ISI |
- Hong, L., Schroth, G.P., Matthews, H.R., Yau, P. & Bradbury, E.M. Studies of the DNA binding properties of histone H4 amino terminus. Thermal denaturation studies reveal that acetylation markedly reduces the binding constant of the H4 "tail" to DNA. J. Biol. Chem. 268, 305–314 (1993). | PubMed | ISI | ChemPort |
- Lee, D.Y., Hayes, J.J., Pruss, D. & Wolffe, A.P. A positive role for histone acetylation in transcription factor access to nucleosomal DNA. Cell 72, 73–84 (1993). | Article | PubMed | ISI | ChemPort |
- Meehan, R.R., Lewis, J.D., McKay, S., Kleiner, E.L. & Bird, A.P. Identification of a mammalian protein that binds specifically to DNA containing methylated CpGs. Cell 58, 499–507 (1989). | Article | PubMed | ISI | ChemPort |
- Lewis, J.D. et al. Purification, sequence, and cellular localization of a novel chromosomal protein that binds to methylated DNA.
|
s.d.) in the promoter-proximal region is shown, normalized as described in Methods.
