HBO1 HAT complexes target chromatin throughout gene coding regions via multiple PHD finger interactions with histone H3 tail

Abstract

The HBO1 HAT protein is the major source of histone H4 acetylation in vivo and has been shown to play critical roles in gene regulation and DNA replication. A distinctive characteristic of HBO1 HAT complexes is the presence of three PHD finger domains in two different subunits: tumor suppressor proteins ING4/5 and JADE1/2/3. Biochemical and functional analyses indicate that these domains interact with histone H3 N-terminal tail region, but with a different specificity toward its methylation status. Their combinatorial action is essential in regulating chromatin binding and substrate specificity of HBO1 complexes, as well as cell growth. Importantly, localization analyses on the human genome indicate that HBO1 complexes are enriched throughout the coding regions of genes, supporting a role in transcription elongation. These results underline the importance and versatility of PHD finger domains in regulating chromatin association and histone modification crosstalk within a single protein complex.

Figure 1. The interplay between the three PHD fingers within the HBO1 HAT complex

(A) Schematic representation of functional domains found within subunits of the HBO1 complex. Domains: MYST, MYST-family HAT domain; I and II, regions conserved in all MYST-ING complexes; Ser, serine-rich; ZNF, zing-finger; LZ, leucine zipper;. (B) Isoforms of the JADE1 protein and domains contained therein. (C) Association of ING5 and hEaf6 with the HBO1 complex requires full-length JADE1 protein (lane 1), while the JADE1S isoform only associates with HBO1 (lane 3). The PHD finger domain of JADE1 does not affect the composition of the complex (lane 2). Immunopurified JADE1 complexes from the indicated co-transfected cells were analyzed for subunit associations by gel staining. Subunit identifications were confirmed by western blot (not shown). (D) ING4/5 PHD fingers preferentially bind H3 peptides trimethylated on K4. Peptide pull-down assays with the indicated biotinylated peptides and recombinant PHD fingers fused to GST were analyzed by western blot with anti-GST. (E–F) Native ING4 and ING5 complexes preferentially acetylate histone H3 tails that are methylated on K4. Purified native ING4/5 complexes were tested in HAT assays with the indicated histone H3 peptides with different levels of methylation on K4 or K9. Reactions were spotted on membranes and counted by liquid scintillation. Values are based on duplicate assays with standard deviation. (G) JADE PHD finger domains interact with histone H3 tail peptides in vitro, but with different effect of K4 methylation. H3 peptide pull-down assays as in (D). (H) Affinities of JADE and ING PHD finger domains for histone H3 peptides and the effect of K4 trimethylation (measured by tryptophan fluorescence). * Taken from (Hung et al., 2009); ** taken from (Champagne et al., 2008). (I) JADE1 PHD1 is dominant over PHD2 for blocking interaction with H3K4me3 in the absence of ING PHD. H3 peptide pull-down assays as in (D). (J) Histone H3 tail is essential for HBO1 complexes to associate with native chromatin. The indicated JADE1L/S complexes (immunopurified from co-transfected cells as in (C)) were tested for binding to native chromatin purified from wild type or set1 (no H3K4me) yeast cells. Bound histones were visualized by western blot with the indicated H3 antibodies. The presence of the N-terminal truncation of H3 is shown on the right (lines separate samples because some lanes were removed).

Figure 2. Histone H3 tail-dependent binding to chromatin and HBO1 nucleosomal HAT activity depend on JADE1 PHD finger domains, while ING5 is required to target H3K4 methylated regions and modulates HAT specificity

(A) Immunopurified complexes from cells co-transfected with the indicated constructs were analyzed on gel to normalize relative amounts. (B) Binding assays with purified native chromatin from yeast and human cells. Chromatin binding with wild type and mutant JADE1 complexes immunopurified in (A) was analyzed as in Fig. 1J. (C) Efficient acetylation of H3 peptides by HBO1 requires JADE1 PHD fingers. HAT assays on H3 peptides as in Fig. 1E were performed with wild type and mutant JADE1 complexes immunopurified in (A). (D) HBO1 requires JADE1 PHDs to acetylate chromatin, while ING5 modulates histone tail specificity in the presence of H3K4me. HAT assays with the same complexes on yeast native chromatin with (WT) or without (set1) H3K4me mark. The reactions were loaded and acetylation revealed by fluorography. Coomassie staining of the gel is shown to ensure equivalent loading.

Figure 3. JADE1 PHD2 is essential for HBO1 binding to chromatin in vivo and for tumor suppressor activity

(A–B) HBO1 complexes require JADE1 PHD2 to associate with chromatin in vivo, even in the presence of ING5. Additionally, the HBO1-JADE1S complex does not associate with chromatin containing H3K4me in vivo (B). Immunopurified JADE1 complexes from the indicated co-transfected cells were analyzed for subunit associations and endogenous histone binding by gel staining (A). The full set of stoichiometric core histones is labeled on the right. Samples were also analyzed by western blot (B). (C) JADE1 complexes associate with K36me3-containing chromatin regions in vivo. JADE1L and JADE1S immunopurified complexes from the indicated co-transfections were analyzed by western with the indicated antibodies. The amount of co-purified chromatin is similar between JADE1L and JADE1S but there is significantly more H3K36me3 signal in JADE1S. (D) JADE1 PHD1 shows some preferential binding to H3K36me3 peptides. Peptide pull-down assays as in Fig. 1D. (E) Expression of JADE1L strongly inhibits HeLa cell growth while a protein lacking PHD2 does not. Indicated transduced cell lines were seeded at the same density and cells were counted over a period of 7 days. Values are based on duplicate counts of two independent cultures with standard deviation. (F) JADE1L inhibits anchorage independent cell growth while deletion of PHD2 stimulates it. HeLa cell clones expressing the different constructs were incubated in soft agar for three weeks and colony formation was measured. Values are based on quadruplicate counts of two independent cultures with standard deviation. Examples of image fields are shown at the bottom.

Figure 4. JADE1L HAT complexes are associated with the coding regions of genes in human cells

(A) Graphical summary of the data obtained by ChIP-chip assay with JADE1L wild-type and ΔPHD2 proteins in the presence or absence of doxorubicin. Numbers of significant detected peaks of binding are presented in relation to relative position on the body of genes, or distance from transcription start/end sites (TSS/TES) when located outside coding regions. (B–C) HBO1 colocalizes with JADE1L and histone H3 methylation marks in vivo. ChIP analysis of genes found to be bound by JADE1L in (A) using WT and mutant cells and also cells expressing tagged HBO1. IP/input signals obtained on the indicated loci with empty vector-transduced cells were subtracted from the signal obtained with the tagged proteins (B). H3K4me3 and H3K36me3 signal corrected for nucleosome occupancy (total H3) were also measured on the same loci (C). (D) ChIP-chip data as in (A) obtained on the Menin1 tumor suppressor gene. H3K4me3 and H3K36me3 signals obtained by ChIP-chip on the same ENCODE sequences in HeLa cells are also depicted (Koch et al., 2007; Lian et al., 2008). Values are presented as log2 ratio of IP/input. The part of the graph below a log2 ratio of 1 is shaded to highlight regions of significant binding.

Figure 5. Differential regulation of histone acetylation by JADE1-HBO1 complexes from transcription initiation to elongation

(A–B) ChIP analysis of histone H3 and H4 acetylation in cells stably expressing JADE1L wild type and ΔPHD2 or ING5, in the presence or absence of doxorubicin. Acetylation levels corrected for nucleosome occupancy were measured near EXT transcription start site and further downstream in the coding region. JADE1L affects acetylation at both loci while ING5 only affects the TSS region and favors H3 acetylation. Values are based on two independent experiments with standard error. (C) Model for HBO1-dependent acetylation, depicting how different HBO1-JADE HAT complexes can target initiation versus elongation steps of transcription (see text).