Weak but uniform enrichment of the histone variant macroH2A1 along the inactive X chromosome

Abstract

We studied the enrichment and distribution of the histone variant mH2A1 in the condensed inactive X (Xi) chromosome. By using highly specific antibodies against mH2A1 and stable HEK 293 cell lines expressing either green fluorescent protein (GFP)-mH2A1 or GFP-H2A, we found that the Xi chromosome contains approximately 1.5-fold more mH2A1 than the autosomes. To determine the in vivo distribution of mH2A1 along the X chromosome, we used a native chromatin immunoprecipitation-on-chip technique. DNA isolated from mH2A1-immunoprecipitated nucleosomes from either male or female mouse liver were hybridized to tiling microarrays covering 5 kb around most promoters or the entire X chromosome. The data show that mH2A1 is uniformly distributed across the entire Xi chromosome. Interestingly, a stronger mH2A1 enrichment along the pseudoautosomal X chromosome region was observed in both sexes. Our results indicate a potential role for macroH2A in large-scale chromosome structure and genome stability.

FIG. 1.

Preferential staining of mH2A1 in the Xi chromosome of HEK 293 cells. (A) The anti-mH2A1 immunopurified antibody recognizes a single band with a molecular mass of ∼42 kDa within an HEK 293 cell extract. The same amount of HEK 293 whole-cell extract was loaded on several wells of a 12% polyacrylamide gel containing SDS. After the completion of the electrophoresis, the gel was cut and one part of the gel was stained with Coomassie blue (lane 2), while the other part was used for Western blotting with anti-mH2A antibodies (lane 3). Lane 1, protein molecular mass marker. The same high specificity of the anti-mH2A1 antibody was observed when extract from mouse cells was used (not shown). (B) The mH2A1 antibody detects the Xi chromosome in HEK 293 cells at interphase. The top panel shows the nucleus of an HEK 293 cell stained with Hoechst, and the lower panel presents the immunostaining of the same cell with the mH2A1 antibody.

FIG. 2.

Measurements of the amount of mH2A1 associated with the Xi chromosome in HEK 293 cells. HEK 293 cell lines stably expressing either GFP-H2A or GFP-mH2A1.2 and naïve HEK 293 cells were used to estimate the ratio of mH2A1 to H2A in the Xi chromosome and the autosomes. To visualize mH2A1 in the naïve cells, an immunopurified anti-mH2A1 antibody was used. (A) Preferential staining of the Xi chromosome with GFP-H2A, GFP-mH2A1, and anti-mH2A1 antibody (first row). The second row shows the DAPI staining of the nuclei. The arrow indicates the Xi chromosome. (B) The Xi chromosome is enriched in mH2A1. Images were deconvolved into two dimensions and then segmented with Metamorph software (Universal Imaging Corporation). The intensity of the fluorescence of the GFP and the anti-mH2A1 staining of the Xi chromosome in the three types of cells was measured. This also was done for a randomly chosen region of the nucleus with the same dimensions as those of the Xi chromosome, and the ratio between the fluorescence intensities of the two signals was calculated for the respective cells. The same ratio was calculated for the DAPI signals. Eighty GFP-mH2A1.2 or GFP-H2A cells and 20 naïve HEK 293 cells were used in the experiments. The bars indicate the standard deviations.

FIG. 3.

ChIP of mH2A1 nucleosomes. Nuclei prepared from the liver of female mice were digested with microccocal nuclease, and mononucleosomes were purified on a sucrose gradient containing 0.6 M NaCl to remove linker histones. Mononucleosomes were immunoprecipitated by using anti-mH2A antibodies attached to protein A agarose beads. Decreasing amounts of input nucleosomes (lanes 1 to 4; the nucleosome amount in each successive lane is decreased by twofold), the bound fractions (lane 5), and the unbound fractions (lane 6) were loaded on a 5 to 15% gradient polyacrylamide gel containing SDS. Lane 7 was loaded with control beads, which did not contain anti-mH2A antibodies but that were incubated with the nucleosomes. After the completion of the electrophoresis, the proteins were transferred to a nitrocellulose filter. The filter first was revealed with an mH2A antibody and then with an anti-H2A antibody. Note that mH2A gives a much stronger signal than H2A in the bound fraction (lane 5). In contrast, the mH2A signals of the input (lanes 1 to 4) and the unbound (lane 6) fractions are much less than the respective signals for H2A. IgG, immunoglobulin G.

FIG. 4.

mH2A1 enrichment on the promoters of the X chromosome of female and male mouse livers. Chromatin from both female and male mouse livers was immunoprecipitated with immunopurified anti-mH2A.1 antibody, and the input and bound DNA fractions were used without amplification for hybridization on tiling arrays containing probes covering 5 kb (about −4000 to +1000) of 24,939 mouse promoters. (A) Distribution of the log₂ immunoprecipitated/input signal ratio of the promoter regions within a 6-Mb region encompassing the tandemly repeated Ott gene cluster. The top row shows genomic coordinates in megabase pairs (UCSC mm5 in NCBI build 33). The distribution of the signal that is generally centered around 0 is homogeneously shifted to higher values within the repeated region specifically in the female liver. (B) Comparative box plot analysis of the distribution of the mH2A enrichment levels in females and males on the different chromosomes. The ratios for both female and male bound (Ip)/input fractions and the female/male normalized input ratio are represented (indicated on the left and right, respectively). Data from all pooled autosomes (Auto) are compared to those from the X chromosome alone. Very similar distributions are observed when any individual autosome is analyzed.

FIG. 5.

mH2A1 enrichment on the whole length of the X chromosome of female and male mouse livers. Chromatin from both female and male mouse livers was immunoprecipitated with immunopurified anti-mH2A1 antibody, and the input and bound DNA (without PCR amplification) was used for hybridization on tiling arrays containing probes from the entire X chromosome and part of chromosome 17. M values (log₂ immunoprecipitated signal/input signal) were normalized to chromosome 17 using a weighted average (Tukey's biweight function). Arrays were scaled using mean absolute deviations from chromosome 17 values. (A) Comparative box plot analysis of the distribution of the mH2A enrichment levels in the female and male on chromosomes 17 and X. (B to D) Wavelet analysis (33) was used to transform the combined ChIP-on-chip data sets to highlight broad patterns of mH2A1 association on the chromatin. Wavelet-smoothened data at the 64-kb scale was obtained using the maximal overlap discrete wavelet transform. The horizontal axes show genomic positions, while the vertical axes show the wavelet coefficient at that position at the 64-kb scale. The top row shows genomic coordinates (from the NCBI m36 genome assembly). (B) Chromosome 17 (4657287 to 19251247 in the NCBI m36 genome assembly). (C) X chromosome. mH2A1 enrichment is seen across the entire female X chromosome but not on the male X chromosome. The signal on the female X chromosome is the average of the contributions from both the active X and the Xi chromosomes. (D) A high degree of mH2A1 enrichment is seen in the PAR (28), starting from within the Fxy/Mid-1 gene (depicted with a box; arrows show initiation sites of three annotated transcripts) of both the male and female X chromosomes.