Ash1l methylates Lys36 of histone H3 independently of transcriptional elongation to counteract polycomb silencing

Abstract

Molecular mechanisms for the establishment of transcriptional memory are poorly understood. 5,6-dichloro-1-D-ribofuranosyl-benzimidazole (DRB) is a P-TEFb kinase inhibitor that artificially induces the poised RNA polymerase II (RNAPII), thereby manifesting intermediate steps for the establishment of transcriptional activation. Here, using genetics and DRB, we show that mammalian Absent, small, or homeotic discs 1-like (Ash1l), a member of the trithorax group proteins, methylates Lys36 of histone H3 to promote the establishment of Hox gene expression by counteracting Polycomb silencing. Importantly, we found that Ash1l-dependent Lys36 di-, tri-methylation of histone H3 in a coding region and exclusion of Polycomb group proteins occur independently of transcriptional elongation in embryonic stem (ES) cells, although both were previously thought to be consequences of transcription. Genome-wide analyses of histone H3 Lys36 methylation under DRB treatment have suggested that binding of the retinoic acid receptor (RAR) to a certain genomic region promotes trimethylation in the RAR-associated gene independent of its ongoing transcription. Moreover, DRB treatment unveils a parallel response between Lys36 methylation of histone H3 and occupancy of either Tip60 or Mof in a region-dependent manner. We also found that Brg1 is another key player involved in the response. Our results uncover a novel regulatory cascade orchestrated by Ash1l with RAR and provide insights into mechanisms underlying the establishment of the transcriptional activation that counteracts Polycomb silencing.

Figure 1. Basic characterization of Ash1l mutant ES cells and Hox gene expression in response to RA.

(A) Schematic representation of the strategy used for targeted disruption of the Ash1l gene. Exons 11–12 encoding the AWS-SET domain with their flanking introns were floxed by loxP sequences. Cre-mediated germ-line recombination resulted in the generation of the ΔSET allele. (B) A PCR primer-pair of SET_flank_F/R was used to distinguish between expression from the wild-type allele and that from the ΔSET allele as shown in (A). A PCR primer-pair of Ab3_F/R was used to determine total expression from both alleles. The PCR primer-pairs are listed with their sequences in Table S4. (C) Comparison of Ash1l mRNA expression among E10.5-litter-mate embryos by conventional RT-PCR. Expression levels of Gapdh mRNA are shown as controls. (D) Protocol for RA-induced differentiation of ES cells. (E) Conventional RT-PCR analyses of Hoxb4 and Hoxd4 mRNA expression levels in response to various concentrations of RA. (F and G) Quantitative RT-PCR analysis of Hoxd4 mRNA expression levels. RA-titration analysis after 48 hours of induction (F), and a time-course analysis using 1 nM RA (G). The results are represented as relative expression levels between wild-type and ΔSET ES cells. Wild-type cells (blue bars or line) and homozygous ΔSET ES cells (orange bars or line) are shown. These results represent means and standard deviations (s.d.) of three independent PCR reactions (Student's t-test, *P<0.05).

Figure 2. Comprehensive gene expression analyses of differentiating ES cells.

ES cells were treated with 10-Seq analyses. (A) RA-responsive genes (those whose expression levels were increased more than 5-fold compared with undifferentiated ES cells: 543 genes out of 14,255 eligible annotated genes) were plotted on the graph using modified fragments/kb of transcript/million fragments mapped (FPKM) values. The x-axis corresponds to expression levels of each gene (shown as log2 transformation of each FPKM value plus 0.1), and the y-axis corresponds to fold change in gene expression levels between ΔSET ES cells and wild-type (shown as Δlog2 transformation). The encircled area was enriched for some Hox and Wnt family genes (arrows, see also Figure S2). (B) Gene ontology analysis of ΔSET-impaired 152 genes. A subset of RA-responsive genes demonstrating a greater-than-2-fold decrease in the modified FPKM values in differentiating ΔSET ES cells was analyzed. Gene enrichment P-values were calculated by Chi-square test. The numbers of genes in each group are shown in parentheses. (C) Bar chart showing relative ratios of the numbers of genes carrying trimethylation of Lys27 , a hallmark of the Polycomb-regulated genes. Genes showing a difference greater than 2-fold were analyzed (up- or down-regulated). In addition, RA-responsive genes in the down-regulated genes were further extracted and analyzed [the last group, same as in (B)]. Chi-square testing was used for calculation of P-values against the total gene set, *P<0.001. The numbers of genes in each group are shown in the table below. (D) Radar chart showing relative ratio of status of chromatin signatures for indicated gene group as in (C). (E) Venn diagram showing the relationship between Ash1l-target genes and ΔSET-affected genes. The numbers of genes in each compartment are shown. The total number of annotated genes analyzed was 18,724.

Figure 3. Ash1l specifically methylates Lys36 of histone H3 both in vitro and in vivo.

(A) Structure of full-length Ash1l (left) and GST-fusion Ash1l constructs (right). NF: the N-terminal flanking region of the AWS domain, H2213K: amino acid substitution of His2213 to Lys. (B and C) Histone methyltransferase assays with GST-fusion Ash1l (rEnzyme). rTet on DNA: recombinant (H3–H4)₂ tetramers bound to plasmid DNA. rNucleosomes: recombinant nucleosomes containing wild-type or mutant histone H3 as indicated. CBB: Coomassie brilliant blue staining for histone polypeptides separated on a SDS-PAGE gel. Position of histone H3 in the autoradiogram is indicated in each assay (arrowheads). (D) Diagram of the Hoxd4 gene. Black and grey boxes represent exons and a 3′ retinoic acid responsive element (RARE), respectively. Black bars under the diagram indicate the regions analyzed by ChIP assays. TSS: transcription start site. (E and F) ChIP assays of various regions in differentiating ES cells before (E) or after (F) RA treatment (see the culture protocol shown in Figure 1D). The results are represented as mean values relative to input (% input). Error bars represent s.d. (Student's t-test, *P<0.05). The antibodies used are indicated above each graph. Broken lines indicate approximate levels of ChIP signals in the Il2ra promoter as a control.

Figure 4. Lys36me2 occurs independently of the Ser2-phosphorylation of RNAPII, while Lys36me3 occurs in a context-dependent manner.

ChIP assays of differentiating ES cells either with (+) or without (−) DRB treatment. Promoter-proximal coding regions of either Hoxd4 (A) or Gapdh (B) were analyzed. DRB was added during RA treatment, and then ES cells were cultured for another 16 hours. The antibodies used are indicated above each graph. The results are represented as means and s.d. (Student's t-test, *P<0.05). Broken lines indicate approximate levels of ChIP signals in either Il2ra promoter as controls. (C) In the presence or absence of DRB, bulk histones in differentiating ES cells were analyzed by immunoblot using the indicated antibodies. (D) Distributions of Lys36me2/3 ChIP-Seq read counts relative to a metagene in the presence or absence of DRB. The y-axis corresponds to relative read counts to base-read counts (shown as log2 transformation). TSS: transcription start site, TES: transcription end site. (E and F) Genomic profiles for Lys36me2/3, RAR, and Ash1l ChIP-Seq signals of numbers of genomic loci in differentiating ES cells (10 nM RA treatment for 2 days). DRB was used during culturing of cells where indicated. The x-axis corresponds to genomic locations; the y-axis corresponds to the normalized ChIP-Seq signal density. Representations of UCSC genes are shown on the bottom, in which Hox loci (E) and the other representative loci (F) are shown. RAR ChIP-Seq datasets (GSE19409) were downloaded from the NCBI Short Read Archive database. (G) Bar charts showing relative ratios of the numbers of RAR-associated genes in indicated gene groups in differentiating ES cells. The gene groups are classified according to fold change in Lys36me2/3 levels in response to DRB: Δlog2 transformation of normalized reads/kb/million mapped (RPKM) values of each gene in DRB (+) over those in (−). Total numbers of genes in each gene group are indicated on the right side of each graph. Chi-square testing was used for calculation of P-values where indicated, *P<0.001. (H) Distributions of Lys36me2/3 ChIP-Seq read counts relative to RAR binding sites in ES cells in the presence or absence of DRB.

Figure 5. Lys36me2/3 and exclusion of the PRCs occur independently of productive transcriptional elongation.

(A) Protocol for RA-induced differentiation of ES cells under DRB pretreatment. DRB was added to the culture medium 1 hour before the addition of RA, then the ES cells were cultured for another 16 hours in the presence of 100 nM RA. (B) Nuclear run-on assay in combination with RT-qPCR analyses of Hoxd4 mRNA expression either with or without DRB pretreatment. The results are represented as values relative to Gapdh mRNA in each cell type. Error bars represent s.d. (Student's t-test, *P<0.05). ND: not detected. (C and D) ChIP assays of differentiating ES cells either with (+) or without (−) DRB pretreatment. A promoter-proximal coding region of Hoxd4 was analyzed. The antibodies used are indicated above each graph. The results are represented as means and s.d. (Student's t-test, *P<0.05). Broken lines indicate approximate levels of ChIP signals in either Il2ra promoter (C) or Gapdh coding region (D) as controls.

Figure 6. Functional links of Ash1l to the Tip60, Mof, and Brg1 complexes.

(A–C and F–I) ChIP assays of Hoxd4 and Gapdh in differentiating ES cells either with (+) or without (−) DRB treatment. Regions that were analyzed were divided into two parts as indicated in each panel: promoter-proximal (proximal) and distal (distal) coding regions. The antibodies used are indicated above each graph or in panels. The results are represented as means and s.d. (Student's t-test, *P<0.05). Broken lines indicate approximate levels of ChIP signals in Il2ra promoter as a control. (D) A diagram of the Gapdh gene is shown. Black bars under the diagram indicate the regions analyzed by ChIP assays. (E) Whole-cell extracts were analyzed by immunoblot using the antibodies against the indicated histone modifications.

Figure 7. Posterior shifts of the expression boundaries of Hoxb4 and Hoxd4 mRNAs.

Whole-mount in situ hybridization analyses of Hoxb4 and Hoxd4 mRNA expression in E10.5 embryos. Shown are normal (white arrows) and affected (orange arrow-heads) anterior expression boundaries at the paraxial mesoderm in wild-type and ΔSET mutant embryos, respectively.

Figure 8. Typical skeletal phenotypes of Ash1l ΔSET mice, and a genetic interaction between Ash1l and Mel18.

Lateral views of the cervico-thoracic region of the axial skeleton are shown. (A) The C2-to-C1 transformation in a ΔSET mouse (C2*), deformities of the anterior arch of the atlas (aaa* at C1*) and an ectopic rib (arrow-head) on the C7 vertebra (C7*). (B) A genetic interaction between Mel18 mutant allele and Ash1l ΔSET allele. The C2-to-C3 transformation (C2**) in a Mel18 mutant mouse is partially suppressed by an additional Ash1l ΔSET allele (C2#).

Figure 9. A proposed role of Ash1l with RAR in the establishment of transcriptional activation.

In the upper panel, Ash1l is preloaded in the promoter-proximal coding region with the condensed bivalent chromatin (thick and short gene body) that mainly generates immature short transcripts (represented in orange on the gene body). Enzymatic activity of Ash1l is inactivated under the condition (light-red). During the establishment of transcriptional activation, retinoic acid and its receptor (RA&RAR) promote activation of Ash1l (dark-red), as well as association of the other Lys36-methylases with the target chromatin. These Lys36-methylases, including Ash1l, orchestrate the downstream mechanisms directly or indirectly, thereby further promoting RA response through alleviating the repressive effect of the PRCs and opening the condensed chromatin (represented by the extended shape of the gene body in the bottom panel) independently of transcriptional elongation. The Brg1 complex may indirectly target Lys36me2/3 through Lys16ac.