In eukaryotes, accurate chromosome segregation during mitosis and meiosis is coordinated by kinetochores, which are unique chromosomal sites for microtubule attachment. Centromeres specify the kinetochore formation sites on individual chromosomes, and are epigenetically marked by the assembly of nucleosomes containing the centromere-specific histone H3 variant, CENP-A. Although the underlying mechanism is unclear, centromere inheritance is probably dictated by the architecture of the centromeric nucleosome. Here we report the crystal structure of the human centromeric nucleosome containing CENP-A and its cognate α-satellite DNA derivative (147 base pairs). In the human CENP-A nucleosome, the DNA is wrapped around the histone octamer, consisting of two each of histones H2A, H2B, H4 and CENP-A, in a left-handed orientation. However, unlike the canonical H3 nucleosome, only the central 121 base pairs of the DNA are visible. The thirteen base pairs from both ends of the DNA are invisible in the crystal structure, and the αN helix of CENP-A is shorter than that of H3, which is known to be important for the orientation of the DNA ends in the canonical H3 nucleosome. A structural comparison of the CENP-A and H3 nucleosomes revealed that CENP-A contains two extra amino acid residues (Arg 80 and Gly 81) in the loop 1 region, which is completely exposed to the solvent. Mutations of the CENP-A loop 1 residues reduced CENP-A retention at the centromeres in human cells. Therefore, the CENP-A loop 1 may function in stabilizing the centromeric chromatin containing CENP-A, possibly by providing a binding site for trans-acting factors. The structure provides the first atomic-resolution picture of the centromere-specific nucleosome.

The nucleosome is the fundamental repeating unit of chromatin, via which genomic DNA is packaged into the nucleus in eukaryotes. In the nucleosome, two copies of each core histone, H2A, H2B, H3 and H4, form a histone octamer which wraps 146 base pairs of DNA around itself. All of the core histones except for histone H4 have nonallelic isoforms called histone variants. In humans, eight histone H3 variants, H3.1, H3.2, H3.3, H3T, H3.5, H3.X, H3.Y and CENP-A, have been reported to date. Previous studies have suggested that histone H3 variants possess distinct functions in the formation of specific chromosome regions and/or in the regulation of transcription and replication. H3.1, H3.2 and H3.3 are the most abundant H3 variants. Here, crystal structures of human nucleosomes containing either H3.2 or H3.3 have been solved. The structures were essentially the same as that of the H3.1 nucleosome. Since the amino-acid residues specific for H3.2 and H3.3 are located on the accessible surface of the H3/H4 tetramer, they may be potential interaction sites for H3.2- and H3.3-specific chaperones.

Histones are the protein components of the nucleosome, which forms the basic architecture of eukaryotic chromatin. Histones H2A, H2B, H3, and H4 are composed of two common regions, the “histone fold” and the “histone tail”. Many efforts have been focused on the mechanisms by which the post-translational modifications of histone tails regulate the higher-order chromatin architecture. On the other hand, previous biochemical studies have suggested that histone tails also affect the structure and stability of the nucleosome core particle itself. However, the precise contributions of each histone tail are unclear. In the present study, we determined the crystal structures of four mutant nucleosomes, in which one of the four histones, H2A, H2B, H3, or H4, lacked the N-terminal tail. We found that the deletion of the H2B or H3 N-terminal tail affected histone–DNA interactions and substantially decreased nucleosome stability. These findings provide important information for understanding the complex roles of histone tails in regulating chromatin structure.

CENP-A and CENP-B are major components of centromeric chromatin. CENP-A is the histone H3 variant, which forms the centromere-specific nucleosome. CENP-B specifically binds to the CENP-B box DNA sequence on the centromere-specific repetitive DNA. In the present study, we found that the CENP-A nucleosome more stably retains human CENP-B than the H3.1 nucleosome in vitro. Specifically, CENP-B forms a stable complex with the CENP-A nucleosome, when the CENP-B box sequence is located at the proximal edge of the nucleosome. Surprisingly, the CENP-B binding was weaker when the CENP-B box sequence was located in the distal linker region of the nucleosome. This difference in CENP-B binding, depending on the CENP-B box location, was not observed with the H3.1 nucleosome. Consistently, we found that the DNA-binding domain of CENP-B specifically interacted with the CENP-A-H4 complex, but not with the H3.1-H4 complex, in vitro. These results suggested that CENP-B forms a more stable complex with the CENP-A nucleosome through specific interactions with CENP-A, if the CENP-B box is located proximal to the CENP-A nucleosome. Our in vivo assay also revealed that CENP-B binding in the vicinity of the CENP-A nucleosome substantially stabilizes the CENP-A nucleosome on alphoid DNA in human cells.

CENP-B is a widely conserved centromeric satellite DNA-binding protein, which specifically binds to a 17-bp DNA sequence known as the CENP-B box. CENP-B functions positively in the de novo assembly of centromeric nucleosomes, containing the centromere-specific histone H3 variant, CENP-A. At the same time, CENP-B also prevents undesired assembly of the CENP-A nucleosome through heterochromatin formation on satellite DNA integrated into ectopic sites. Therefore, improper CENP-B binding to chromosomes could be harmful. However, no CENP-B eviction mechanism has yet been reported. In the present study, we found that human Nap1, an acidic histone chaperone, inhibited the non-specific binding of CENP-B to nucleosomes and apparently stimulated CENP-B binding to its cognate CENP-B box DNA in nucleosomes. In human cells, the CENP-B eviction activity of Nap1 was confirmed in model experiments, in which the CENP-B binding to a human artificial chromosome or an ectopic chromosome locus bearing CENP-B boxes was significantly decreased when Nap1 was tethered near the CENP-B box sequence. In contrast, another acidic histone chaperone, sNASP, did not promote CENP-B eviction in vitro and in vivo and did not stimulate specific CENP-B binding to CENP-A nucleosomes in vitro. We therefore propose a novel mechanism of CENP-B regulation by Nap1.