Structural basis for the recognition of histone H4 by the histone-chaperone RbAp46

Abstract

RbAp46 and RbAp48 (pRB-associated proteins p46 and p48, also known as RBBP7 and RBBP4, respectively) are highly homologous histone chaperones that play key roles in establishing and maintaining chromatin structure. We report here the crystal structure of human RbAp46 bound to histone H4. RbAp46 folds into a seven-bladed beta propeller structure and binds histone H4 in a groove formed between an N-terminal alpha helix and an extended loop inserted into blade six. Surprisingly, histone H4 adopts a different conformation when interacting with RbAp46 than it does in either the nucleosome or in the complex with ASF1, another histone chaperone. Our structural and biochemical results suggest that when a histone H3/H4 dimer (or tetramer) binds to RbAp46 or RbAp48, helix 1 of histone H4 unfolds to interact with the histone chaperone. We discuss the implications of our findings for the assembly and function of RbAp46 and RbAp48 complexes.

Figure 1

Structure of the RbAp46/Histone H4 Complex (A and B) Two different views of the structure obtained with the histone H4 peptide are shown at ∼90° to each other, as viewed from the top and side of the β-propeller structure. Residues 25–41 in histone H4 are shown in blue, whereas residues 9–410 in RbAp46 are shown in yellow. The disordered region of RbAp46 for which the electron density is unclear is shown as a dotted loop. The “Native 2” structure (PDB code: 3CFS; Tables S1 and S2) and the program PyMol (DeLano, 2002) were used to make this and the following figures. (C) Structure-based sequence alignment of the RbAp46, RbAp48, and p55 histone chaperones. The alignment is numbered as for human RbAp46. Sequences are labeled by species name (Hs, Homo sapiens; Dm, Drosophila melanogaster). The secondary structure is indicated by arrows for β strands and coils for α helices, whereas the positions of the seven blades in the β-propeller structure of RbAp46 are indicated by the numbers and letters above the β strands. Yellow stars and blue circles under the alignment indicate key residues involved in hydrophobic or hydrophilic/charged interactions with histone H4, respectively. Black squares indicate residues that were mutated. (D) Histone H4 residues (16–41) labeled as in (C). Residues in italics are not observed in the structure.

Figure 2

Histone H4 Recognition by the RbAp46-Binding Pocket (A) Electrostatic surface potential of RbAp46 contoured and color coded at −91 kT (red) and +91 kT (blue). The potential was calculated and displayed with the program PyMol (DeLano, 2002). The histone H4 peptide is shown as a stick model. The histone H4-binding pocket in RbAp46 is mainly formed by the negatively charged PP loop (which terminates in Pro-362 and Pro-363) and a hydrophobic surface on the N-terminal α helix (helix 1). The interaction of histone H4 residues (Gln-27, Lys-31, Ile-34, Arg-35, Arg-36, Leu-37, Arg-39, and Arg-40) is shown. (B) Detailed view showing the interactions of the hydrophobic Ile-34 and Leu-37 histone H4 residues with Phe-29 and Leu-30 in helix 1 of RbAp46, as well as the positively charged Arg-36, Arg-39, and Arg-40 histone H4 residues with the backbone carbonyl groups in the PP loop and a cluster of acidic residues (Glu-356, Asp-357, and Asp-360) in RbAp46. (C) Site-directed mutagenesis of RbAp46 in either the charged PP loop (E356Q + D357N + E359Q + D360N), the hydrophobic surface of helix 1 (L30Y), or both simultaneously all disrupt the interaction with histone H4 in pull-down experiments with GST-histone H4 1–48. (D) In reciprocal experiments, mutation of histone H4 residues interacting with either the charged PP loop (R39V + R40N), or of residues interacting with both helix 1 and the charged PP loop (I34T + L37D + R35S) and (L37D + R39V + R40N) also disrupt the binding. In both (C) and (D), the top panel shows an autoradiogram illustrating the amount of ³⁵S-labeled RbAp46 pulled down in each experiment, whereas the lower panel shows a Coomassie blue-stained gel indicating the amount of either GST or GST-histone H4 (1–48) used. In each experiment, the input lane contains 30% of the ³⁵S-labeled RbAp46 protein used in each of the pull-down assays. The experiments were carried out in 300 mM NaCl, 20 mM Tris (pH 8.0), 5 mM DTT, and 0.1% (v/v) NP-40.

Figure 3

Interaction of Histones H3 and H4 with RbAp46 (A) Analytical size-exclusion chromatography of the recombinant histone H3/H4 complex used in the biochemical experiments described in this paper. The inset panel shows a Coomassie blue-stained 4%–12% NuPAGE gel used to analyze the fractions. In the conditions used here (2 M NaCl and 20 mM HEPES [pH 7.5], on a Superdex 75 PC3.2/30 column), histones H3 and H4 are present as tetramers, but at lower ionic strengths (as used in the binding experiments) these dissociate to form a mixture of dimers and tetramers (Banks and Gloss, 2003). (B) Pull-down of either wild-type or mutant RbAp46 by histones H3 and H4 crosslinked to DynaBeads, in the absence or presence of the N-terminal histone H4 peptide (residues 16–41). (The experiments were carried out as described in Figure 2. The positively charged lysozyme protein was also crosslinked to beads in separate experiments and was used as a negative control.) (C) Comparison of the interactions of Ile-34, Leu-37, and Ala-38 in helix 1 of histone H4 with (i) the N-terminal helix of RbAp46 in the RbAp46/histone H4 peptide structure, (ii) α helices 2 of histone H3 and H4 in one (of the two) H3/H4 dimer in the nucleosome core particle (Davey et al., 2002; PDB code: 1KX5), and (iii) α helices 2 of histone H3 and H4 in the ASF1-histone H3/H4 complex (English et al., 2006; Natsume et al., 2007; PDB code: 2HUE). In (i), (ii), and (iii), the view is down the axis of helix 1 of histone H4. Because similar contacts are made between histones H3 and H4 in the complex with ASF1 and in both copies of histones H3 and H4 in the nucleosome core particle, it is likely that isolated histones H3 and H4 also interact with each other in a similar manner. Histone H4 is colored blue in all three structures, whereas histone H3 is yellow in the nucleosome core particle and pink in the ASF1 complex.

Figure 4

Interaction of Histones H3 and H4 with RbAp46 (A) Comparison of chemical crosslinking of histones H3 and H4 in the absence and presence of RbAp46. Three different 4%–12% NuPAGE gels were either (i) stained with Coomasie blue or blotted with (ii) anti-histone H4 or (iii) anti-histone H3 antibodies. By themselves, histones H3 and H4 appear to form a H3/H4 heterodimer (∼25 kDa), an H3₂ homodimer (∼30 kDa), and an H3₂/H4₂ heterotetramer (∼50 kDa), indicated by arrowheads. In the presence of RbAp46, two bands, which both contain histones H3 and H4 (also indicated by arrowheads), can be seen in positions roughly corresponding to the interaction of RbAp46 with histone H3/H4 dimers and tetramers (∼65–75 kDa). Control crosslinking experiments of RbAp46 with positively charged lysozyme were also carried out in the same conditions, but yielded no crosslinked species (data not shown). (In the H4 Western blot, staining with the anti-histone H4 antibody appears to be inhibited by the presence of histone H3, leading to white bands. In addition, at the higher concentrations of crosslinker, extensive crosslinking of the proteins results in the formation of large complexes that do not enter the gel—e.g., the band corresponding to the H3₂ complex becomes less intense. The antibodies may also recognize their epitopes less efficiently—as has been noted previously [Banks and Gloss, 2003].) (B) Analytical size-exclusion chromatography, on an Ettan LC system (GE Healthcare), of either 25 μM RbAp46 (blue trace) or 25 μM RbAp46/H3/H4 (red trace), in 200 mM KCl and 20 mM HEPES (pH 7.4), at 4°C on a Superdex 200 PC3.2/30 column. The fractions were collected and analyzed on 4%–12% NuPAGE gels.