Report

Histone H4-K16 Acetylation Controls Chromatin Structure and Protein Interactions

See allHide authors and affiliations

Science  10 Feb 2006:
Vol. 311, Issue 5762, pp. 844-847
DOI: 10.1126/science.1124000

Abstract

Acetylation of histone H4 on lysine 16 (H4-K16Ac) is a prevalent and reversible posttranslational chromatin modification in eukaryotes. To characterize the structural and functional role of this mark, we used a native chemical ligation strategy to generate histone H4 that was homogeneously acetylated at K16. The incorporation of this modified histone into nucleosomal arrays inhibits the formation of compact 30-nanometer–like fibers and impedes the ability of chromatin to form cross-fiber interactions. H4-K16Ac also inhibits the ability of the adenosine triphosphate–utilizing chromatin assembly and remodeling enzyme ACF to mobilize a mononucleosome, indicating that this single histone modification modulates both higher order chromatin structure and functional interactions between a nonhistone protein and the chromatin fiber.

DNA in eukaryotes is present as chromatin, which is an assembly of histones, DNA, and chromatin-associated proteins. The basic building block of chromatin is the nucleosome, which contains two copies of histones H2A, H2B, H3, and H4 (1). Fifteen to 38 amino acids from each histone N terminus form the histone “tails,” providing a platform for posttranslational modifications that modulate the biological role played by the underlying DNA (2). One prevalent modification is H4-K16Ac (3), which has roles in transcriptional activation and the maintenance of euchromatin (4, 5).

Recent work has focused on the ability of histone marks to modulate the binding of nonhistone proteins to the chromatin fiber, such as the yeast silencing factor Sir3 and the Drosophila chromatin-remodeling enzyme ISWI (6, 7). We were interested in testing whether histone modifications might control higher order chromatin structures. Indeed, random hyperacetylation of histone tails (>6 acetates per octamer) disrupts intramolecular folding of nucleosomal arrays into compact, 30-nm-thick fibers (8). Additionally, the H4 tail, and particularly residues 14 to 23, are uniquely important for the formation of these fibers (1, 9). The acetylation of H4-K16 occurs within this region, providing a potential mechanism to regulate chromatin folding.

We used a native chemical ligation strategy to generate recombinant histone H4 homogeneously acetylated at K16 (10, 11). In this strategy, an H4 N-terminal peptide (amino acids 1 to 22), with a C-terminal thioester and an acetylated lysine 16, was synthesized. A recombinant C-terminal fragment of histone H4 (amino acids 23 to 102), in which H4 arginine 23 (R23) had been changed to a cysteine (R23C), was expressed and purified (fig. S1). Chemical ligation of these components yielded full-length H4-K16Ac (Fig. 1A). This product is the same length as unacetylated histone H4 (Fig. 1B) but demonstrates an expected reduction in charge (fig. S2). Three additional H4 polypeptides were expressed and purified (fig. S1): (i) wild-type (WT) H4; (ii) H4-R23C, which harbors the same cysteine substitution present in ligated H4; and (iii) H4-ΔN, in which the N-terminal tail (residues 1 to 19) has been deleted. All H4 polypeptides were incorporated into histone octamers containing recombinant H2A, H2B, and H3 (1), and neither H4-K16Ac nor ligation interfered with octamer assembly (Fig. 1B).

Fig. 1.

Histone H4-K16Ac is incorporated into nucleosomal arrays. (A) Ligation of K16Ac-containing peptide and H4 C-terminal fragment. A ligation reaction is separated on 18% SDS-PAGE gel and stained with Coomassie (27). (B) Shown are peak fractions of H4-WT–, H4-R23C–H4-K16Ac–, or H4-ΔN–containing histone octamers isolated by gel filtration, separated on 18% SDS-PAGE gel, and stained with Coomassie. (C) Nucleosomal arrays containing histone H4-K16Ac prepared in the presence of 174–base pair (bp) 5S DNA. Nucleosomal arrays before (pre) and after (post) precipitation with 4.0 mM MgCl2 analyzed on native 4% PAGE and stained with ethidium bromide are shown. Shown is a typical gel obtained with saturated arrays containing histone H4-R23C or histone H4-K16Ac. Mono and Naked indicate 5S mononucleosomes and free DNA, respectively.

Using step-wise salt dialysis, the four distinct histone octamers were assembled into nucleosomal arrays with a DNA template that harbors 12 copies of the 177–base pair “601” nucleosome positioning sequence (601-177-12) (9, 12). In order to ensure that each DNA template was saturated with 12 nucleosomes, octamers were added in slight excess to the number of 601 repeats, and mononucleosomal-length DNA was included in the reconstitution reactions to act as an “octamer sink” (Fig. 1C) (9). After assembly, 601-177-12 nucleosomal arrays were purified from mononucleosomes by selective MgCl2 precipitation (Fig. 1C), and array saturation was confirmed (fig. S2).

We used sedimentation velocity in conjunction with van Holde–Weischet analysis (13) to ascertain the distribution of sedimentation coefficients for the population of nucleosomal arrays within a sample. Saturated arrays containing H4-WT (Fig. 2A, diamonds), H4-R23C (circles), H4-K16Ac (squares), and H4-ΔN octamers (triangles) were sedimented in a buffer lacking divalent cations, conditions in which nucleosomal arrays adopt an extended “beads-on-a-string” conformation. Each of the four arrays showed nearly identical distributions of sedimentation coefficients (S), with expected midpoints between 34 and 36 S (Fig. 2A, open symbols) (9, 14).

Fig. 2.

H4-K16Ac abolishes higher order chromatin structure. (A) H4-K16Ac abolishes folding of nucleosomal arrays. Integrated sedimentation coefficient distributions of nucleosomal arrays in the presence and absence of Mg2+ were determined by using sedimentation velocity and van Holde-Weischet analysis. Arrays containing WT, H4-R23C, H4-K16Ac, and H4-ΔN histones are depicted by diamonds, circles, squares, and triangles, respectively. Arrays analyzed in the absence (0.1 mM ethylenediaminetetraacetic acid) or presence of Mg2+ (1.0 mM MgCl2) are shown as open or solid symbols, respectively. S20°C,W is the sedimentation coefficient corrected to water at 20°C and adjusted for the difference in mass of the H4-ΔN histone. The data shown are representative of 3 to 5 array reconstitutions. (B) Subsaturated arrays containing H4-K16Ac. Analysis was performed as in (A). Data shown are representative of at least 3 to 5 array reconstitutions. (C) H4-K16Ac disrupts array oligomerization. Nucleosomal arrays were incubated with varying concentrations of MgCl2 at room temperature for 15 min, followed by centrifugation in a microfuge. The fraction of array remaining in the supernatant is plotted as a function of MgCl2 concentration. Arrays containing WT, H4-R23C, H4-K16Ac, and H4-ΔN histones are depicted by solid circles, open circles, open squares, and solid squares, respectively.

When wild-type (Fig. 2E, solid diamonds) and H4-R23C (solid circles) arrays were incubated in a buffer containing 1.0 mM MgCl2, they formed more compact fibers that shifted the sedimentation coefficient distributions to midpoints between 53 and 54 S. These results are consistent with the formation of compact, 30-nm–like fibers, and the H4-R23C substitution does not disrupt this condensation reaction. In contrast, and consistent with a previous report (9), the array reconstituted with H4 that lacks an N-terminal tail (H4-ΔN) (Fig. 2C, solid squares) was unable to condense fully, reaching a midpoint of only 44S. The array reconstituted with H4-K16Ac displayed an identical defect in MgCl2-dependent compaction (Fig. 2A, solid triangles), suggesting that the acetylation of a single lysine leads to a chromatin folding defect equivalent to deletion of the entire H4 tail.

We also analyzed a large set (>10) of different subsaturated arrays (<12 nucleosomes per template) reconstituted with the four different octamers (Fig. 2B) (15). In every case, the addition of MgCl2 to the wild-type and H4-R23C arrays led to large increases in the sedimentation coefficient distributions, consistent with salt-dependent compaction. However, arrays reconstituted with the H4-ΔN and H4-K16Ac octamers were again equally defective for MgCl2-dependent compaction at every level of saturation (Fig. 2B) (15).

As the concentration of MgCl2 was increased beyond 1.5 mM, arrays underwent reversible self-association that is believed to mimic fiber-fiber interactions that stabilize higher order chromosomal domains (16). For the 601-177-12 arrays, the deletion of the H4 tail disrupted self-association (9). To test whether H4-K16Ac affects this intermolecular interaction, arrays reconstituted with H4-WT, H4-R23C, H4-K16Ac, and H4-ΔN octamers were assayed. Self-association of the H4-WT and H4-R23C arrays (Fig. 2C, solid and open circles, respectively) occurred at magnesium concentrations of 1.5 to 2.0 mM, comparable to previously reported values (9). In contrast, self-association of the H4-ΔN and H4-K16Ac arrays (Fig. 2C, solid and open squares, respectively) both required higher concentrations of MgCl2 (2.5 to 3.0 mM). Thus, as in the case for intramolecular folding, the single acetylation of H4-K16 cripples the self-association of arrays and shows defects equivalent to the loss of the H4 tail.

Next we investigated whether H4-K16Ac is associated with decondensed chromatin structures in vivo. HeLa nuclei were digested with micrococcal nuclease (Mnase), and the released chromatin fractionated into MgCl2-soluble and MgCl2-insoluble components. Samples were separated on an SDS–polyacrylamide gel and on an acid-urea-triton (AUT) gel, and the abundance of H4-K16Ac was analyzed by Western blotting (Fig. 3). Consistent with the biochemical studies, H4-K16Ac was enriched in the MgCl2-soluble chromatin fractions (Fig. 3, A and B) that are also enriched in transcriptionally active gene sequences (17). Furthermore, the MgCl2-soluble fractions contained an H4 tail that is exclusively monoacetylated at H4-K16 (Fig. 3B).

Fig. 3.

H4-K16Ac is enriched in MgCl2-soluble chromatin. HeLa nuclei were digested with micrococcal nuclease, and the solubilized chromatin was incubated with MgCl2 at different final concentrations of 0.5, 1, or 2 mM for 10 min on ice. Histones from the magnesium soluble (S) and insoluble (P) chromatin fractions were electrophoresed on (A) 15% SDS-PAGE gels or (B) 15% AUT-PAGE gels. The AUT gel separates the acetylated histone isoforms. The top panels display Coomassie blue–stained gels and the bottom panels show Western analyses with antibodies to H4-K16Ac (Upstate, Charlottesville, VA).

We also tested whether nucleosomal H4-K16Ac affects interactions with chromatin-associated proteins, specifically the Drosophila ISWI-containing adenosine triphosphate (ATP)–utilizing chromatin assembly and remodeling enzyme (ACF) complex. This complex hydrolyzes ATP to mediate nucleosome sliding in cis along DNA (18). This activity requires residues 16 to 19 of the H4 tail (19, 20), and in vitro peptide competition assays have suggested that H4-K16Ac may reduce the interaction of ISWI with nucleosomes (7). To test directly the effect of H4-K16Ac on ISWI activity, we reconstituted end-positioned mononucleosomes with wild-type, H4-R23C, and H4-K16Ac octamers. Each of these mononucleosomes was incubated with the ACF complex, and ATP-dependent mobilization of nucleosomes was analyzed by native polyacrylamide gel electrophoresis (PAGE). When wild-type nucleosomes were incubated with ACF and ATP, a slower-migrating species accumulated with time (half-time t1/2 = ∼7.5 min) (Fig. 4.). This shift in electrophoretic mobility depended on the presence of both ACF and ATP, and it likely represents sliding of the nucleosome to a more central position. This ACF activity was also detected on mononucleosomes reconstituted with H4-K16Ac, but in this case, sliding was slower as compared to the WT and H4-R23C substrates (t1/2 = ∼20 min). These results are consistent with previous peptide studies, demonstrating that H4-K16Ac regulates the functioning of a chromatin-remodeling enzyme independent of its effects on chromatin higher order structure.

Fig. 4.

H4-K16Ac inhibits ACF-mediated nucleosome sliding. (A) End-positioned nucleosomes containing indicated histone H4 were incubated with ACF and ATP at 30°C. Samples were taken from reactions at indicated times and quenched by the addition of EDTA and by placing them on ice. Samples were separated by electrophoresis on 4% native polyacrylamide gels. Lanes 1, 6, and 11 show mononucleosomes without incubation with ACF. (B) Profiles of gel lanes shown in (A). Peaks corresponding to the predicted end-positioned and centrally positioned nucleosomes are indicated by arrows.

The structural effect of H4-K16Ac may directly contribute to regions of decondensed chromatin in eukaryotic organisms. In budding yeast, over 80% of H4 is acetylated at lysine 16, and most of the genome exists in a decondensed state (3, 21). Likewise, evidence suggests that the transcriptionally enhanced X chromosome of male flies, a site of ubiquitous H4-K16Ac, is decondensed (22). Such decondensation of chromatin may contribute to the establishment of transcriptionally active euchromatic regions. In vitro transcription studies suggest that the adoption of higher order chromatin structure reduces gene transcription (23). In contrast, acetylation of H4-K16 increases gene transcription both in vitro and in vivo (24), and the decompaction resulting from such modification may increase the accessibility of factors that promote transcription.

Do other histone marks regulate chromatin folding? The phosphorylation of H3-S10 (S, serine) does not disrupt chromatin folding (25), and triacetylation of the H3 tail by Gcn5p does not disrupt chromatin compaction (26). Similarly, residues 1 to 13 of histone H4 that include three sites of acetylation are dispensable for folding of 601-177-12 arrays (9). Thus, H4-K16 is likely to be a unique acetylation site of histone tails, which function as a dual switch for higher order chromatin structure and protein-histone interactions, promoting chromatin function in a mutually reinforcing manner.

Supporting Online Material

www.sciencemag.org/cgi/content/full/311/5762/844/DC1

Materials and Methods

Figs. S1 to S3

References

References and Notes

View Abstract

Navigate This Article