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.

A histone H3 variant, H3T, is highly expressed in the testis, suggesting that it may play an important role in the chromatin reorganization required for meiosis and/or spermatogenesis. In the present study, we found that the nucleosome containing human H3T is significantly unstable both in vitro and in vivo, as compared to the conventional nucleosome containing H3.1. The crystal structure of the H3T nucleosome revealed structural differences in the H3T regions on both ends of the central α2 helix, as compared to those of H3.1. The H3T-specific residues (Met71 and Val111) are the source of the structural differences observed between H3T and H3.1. A mutational analysis revealed that these residues are responsible for the reduced stability of the H3T-containing nucleosome. These physical and structural properties of the H3T-containing nucleosome may provide the basis of chromatin reorganization during spermatogenesis.

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.

NASP (nuclear autoantigenic sperm protein) is a member of the N1/N2 family, which is widely conserved among eukaryotes. Human NASP reportedly prefers to bind to histones H3⅐H4 and the linker histone H1, as compared with H2A⅐H2B, and is anticipated to function as an H3⅐H4 chaperone for nucleosome assembly. However, the direct nucleosome assembly activity of human NASP has not been reported so far. In humans, two spliced isoforms, somatic and testicular NASPs (sNASP and tNASP, respectively) were identified. In the present study we purified human sNASP and found that sNASP efficiently promoted the assembly of nucleosomes containing the conventional H3.1, H3.2, H3.3, or centromere-specific CENP-A. On the other hand, sNASP inefficiently promoted nucleosome assembly with H3T, a testis-specific H3 variant. Mutational analyses revealed that the Met-71 residue of H3T is responsible for this inefficient nucleosome formation by sNASP. Tetrasomes, composed of the H3⅐H4 tetramer and DNA without H2A⅐H2B, were efficiently formed by the sNASP-mediated nucleosome-assembly reaction. A deletion analysis of sNASP revealed that the central region, amino acid residues 26 -325, of sNASP is responsible for nucleosome assembly in vitro. These experiments are the first demonstration that human NASP directly promotes nucleosome assembly and provide compelling evidence that sNASP is a bona fide histone chaperone for H3⅐H4.

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.