Characterization of Dnmt1 Binding and DNA Methylation on Nucleosomes and Nucleosomal Arrays

Abstract

The packaging of DNA into nucleosomes and the organisation into higher order structures of chromatin limits the access of sequence specific DNA binding factors to DNA. In cells, DNA methylation is preferentially occuring in the linker region of nucleosomes, suggesting a structural impact of chromatin on DNA methylation. These observations raise the question whether DNA methyltransferases are capable to recognize the nucleosomal substrates and to modify the packaged DNA. Here, we performed a detailed analysis of nucleosome binding and nucleosomal DNA methylation by the maintenance DNA methyltransferase Dnmt1. Our binding studies show that Dnmt1 has a DNA length sensing activity, binding cooperatively to DNA, and requiring a minimal DNA length of 20 bp. Dnmt1 needs linker DNA to bind to nucleosomes and most efficiently recognizes nucleosomes with symmetric DNA linkers. Footprinting experiments reveal that Dnmt1 binds to both DNA linkers exiting the nucleosome core. The binding pattern correlates with the efficient methylation of DNA linkers. However, the enzyme lacks the ability to methylate nucleosomal CpG sites on mononucleosomes and nucleosomal arrays, unless chromatin remodeling enzymes create a dynamic chromatin state. In addition, our results show that Dnmt1 functionally interacts with specific chromatin remodeling enzymes to enable complete methylation of hemi-methylated DNA in chromatin.

Fig 1. Dnmt1 requires symmetric linker DNA in order to bind to nucleosomal DNA.

(A) DNA fragments containing the nucleosome positioning sequence (NPS) located either in the center or at the border of the DNA molecules were assembled into mono-nucleosomes using the salt dialysis method. The DNA templates are named according to the size and location of the DNA linkers with respect to the NPS. (B, C) Different combinations of nucleosomal substrates (50 nM each) were mixed in a 1:1 ratio (lanes 2 and 6) and incubated with increasing concentrations of Dnmt1 (100 nM– 500 nM; lanes 3–5 and 7–9). Reactions were analyzed by native polyacrylamide gel electrophoresis next to a molecular weight marker (M). (D) Monitoring the Dnmt1 binding affinity to mono-nucleosomes containing asymmetrical and symmetrical DNA linkers (77-NPS-0 and 77-NPS-77). Reactions were analyzed on 6% (left panel) and on 4.5% (right panel) native polyacrylamide gels. The position of the Dnmt1-nucleosome complex is indicated (lanes 7–10). (E) Binding of Dnmt1 does not disrupt the nucleosome. The 77-NPS-77 nucleosome (lane 3) was incubated with Dnmt1 to form the stable nucleosome–Dnmt1 complex (lane 4). This complex was incubated with increasing concentrations of competitor DNA to compete Dnmt1 off the nucleosome (lane 5–7). Reaction products were analyzed on a native polyacrylamide gel. The positions of nucleosomes and Dnmt1-nucleosome complexes are indicated. Lane 2 shows the 6 kb competitor plasmid DNA. (F) (upper part) Summary of the results of the Dnmt1 band shift assays and (lower part) a cartoon showing how Dnmt1 (grey box) could bind on the entry/exit sites of the nucleosome (red: DNA; bluish: histones).

Fig 2. DNA binding properties of Dnmt1.

(A) DNA fragments of different length (size range from 15 to 250 bp) were incubated with increasing concentrations of Dnmt1 (lanes 2–5). DNA and nucleoprotein complexes were separated on native polyacrylamide gels and stained with ethidium bromide. The asterisk indicates the positions of the Dnmt1-DNA complexes. (B) Quantification of the Dnmt1 binding assay shown in (C), using an equimolar mixture of fluorescently labeled DNA fragments from 15 to 60 bp in length. The remaining free DNA was quantified and plotted. (C) Competitive electromobility shift assay with a mixture of differently sized fluorescently labeled double stranded (ds) oligonucleotides (4 pmol each). The DNA fragments (15–60 bp; 30 bp: lane 2–7, 45 bp: lane 8–14; 15 bp and 60 bp: lane 15–21) were incubated with increasing concentrations of Dnmt1 (0.1 μM- 0.5 μM, and complex formation was analyzed on a 15% native polyacrylamide gel.

Fig 3. Dnmt1 exhibits similar binding affinities towards hemi- and non-methylated DNA.

(A) Scheme showing the setup of a microscale thermophoresis (MST) assay. The aqueous solution inside the capillary is locally heated with a focused IR-laser and coupled with an epifluorescence microscope by using an IR mirror. (B) MST is based on the directed movement of molecules along temperature gradients, an effect termed thermophoresis. The fluorescence inside the capillary is measured for each different concentration of the unlabeled molecule, and the normalized fluorescence in the heated spot is plotted against time. The IR laser is switched on at t = 5 s, the fluorescence decreases as the temperature increases, and the labeled molecules move away from the heated spot because of thermophoresis. When the IR laser is switched off, the molecules diffuse back. The thermophoretic movement for a single ligand concentration is shown in (B). (C, D) Quantitative analysis of Dnmt1 binding to hemi- and non-methylated DNA by MST. The indicated fluorescently labeled DNA substrate of 60 bp in length was incubated with increasing concentrations of Dnmt1 (300 nM—9.4 μM), in the presence or absence of SAM. Binding curves were normalized to the fraction of bound molecules. All MST curves were fitted using the Hill-equation, and EC50 values as well as Hill coefficients (n) are given. (C) shows the analyses of the measurements for the Cy5 labeled DNA substrates, (D) for the Cy3 labeled DNA substrates.

Fig 4. Dnmt1 binds to the DNA linkers at the entry/exit sites of the nucleosome.

(A) Experimental setup (left) and quality analysis (right) of the DNaseI protection assay. The 77-NPS-77 template (301 bp long) was fluorescently labeled by PCR using 5’ labeled oligonucleotides (5'FAM and 5'HEX) and was assembled into nucleosomes by salt gradient dialysis. Free DNA (lanes 2, 3 and 5) and nucleosomes (Nuc; lanes 6 and 8) and nucleosome-Dnmt1 complexes (Nuc-D1; lane 9) were partially digested with 0.1U DNaseI (DN; lanes 3, 6 and 9). The reaction was stopped by the addition of 5 mM EDTA to inactivate the DNaseI and subsequently separated on a native polyacrylamide gel. DNaseI treated complexes were extracted from the gel and analyzed by capillary electrophoresis. (B) Comparison of the electropherograms of DNaseI treated free DNA (red line) with nucleosomal DNA (blue line). The position of the nucleosome core is indicated (grey bar). The electropherograms for both directions (5’HEX: on top and 5’FAM: on bottom) are shown. (C) As in (B), the electropherograms of the DNase I treated nucleosome (blue line) are compared with the Dnmt1-nucleosome complex (red line). The electropherograms for both directions (5’HEX: on top and 5’FAM: on bottom) are shown. The protected regions are highlighted in boxes. RFU: relative fluorescence units.

Fig 5. Nucleosomes inhibit the Dnmt1 dependent DNA methylation in vitro.

(A) Experimental setup of the bisulfite sequencing and the radioactive methyltransferase assay. Black triangles indicate CpG sites and the oval indicates the position of the nucleosome. (B) Analysis of the C91-NPS2-C104 template (342 bp fragment harbouring 27 CpG sites) as free (lane 2) and nucleosomal DNA (lane 3) on a native polyacrylamide gel. (C) Enzymatic activity of Dnmt1 on the NPS2 (no linker DNA in nucleosomal form) and the C91-NPS2-C104 template both as free and nucleosomal DNA analyzed in the radioactive methyltransferase assay using [³H]-SAM (360 nM) as substrate. The incorporation of the [³H]-modified CH₃ group was quantified, indicated as counts per minutes (cpm). (D) Bisulfite analysis of the DNA methylation reaction of Dnmt1, using the C91-NPS2-C104 DNA as free DNA and nucleosomal substrate. The DNA methylation efficiency of Dnmt1 at individual CpG sites of the free (white bars) and mononucleosomal DNA (black bars) is given. The (+) and (-) strands are shown and CpG sites are marked as arrows. The ellipse illustrates the position of the nucleosome. In both assays, the bisulfite sequencing and the radioactive methyltransferase assay, the C91-NPS2-C104 and NPS2 templates were use as as non-methylated substrates.

Fig 6. DNA methylation in the context of nucleosomal arrays and chromatin remodeling enzymes.

(A) Schematic illustration of the method used to generate hemi-methylated DNA. (B) Non-methylated and hemi-methylated nucleosomal arrays generated by salt gradient dialysis were analyzed by partial MNase digestion for 10 to 90 sec (lanes 2–4 and 5–7). The purified DNA was analyzed by agarose gel electrophoresis and ethidium bromide staining. The DNA marker is loaded in lane 1 (M). (C) SDS-PAGE showing the purified, recombinant chromatin remodeling enzymes ACF and Brg1. (D) Testing the activity of the purified remodeling enzymes in nucleosome remodeling assays. Brg1 was incubated with a nucleosomal substrate positioned at the center of the DNA fragment (lanes 1 to 3). ACF was tested with a nucleosome positioned at the border of a DNA fragment of 208 bp in size (lanes 4–6). Reactions contained 1mM ATP and were stopped after 60 min by the addition of competitor DNA. Nucleosome positions were analyzed by native PAGE. (E) Analysis of nucleosome remodeling and Dnmt1 dependent DNA methylation on nucleosomal arrays. Non-methylated (black bars) or hemi-methylated (grey bars) nucleosomal arrays were incubated with Dnmt1, ATP, Brg1 and ACF as indicated. The incorporation of [³H]-labeled CH₃ was determined by scintillation counting, shown as counts per minutes (cpm).