Structure, assembly and reading of centromeric chromatin

Abstract

Centromeres are epigenetically defined chromatin domains marked by the presence of the histone H3 variant CENP-A. Here we review recent structural and biochemical work on CENP-A, and advances in understanding the mechanisms that propagate and read centromeric chromatin domains.

Figure 1

Canonical and centromeric nucleosomes. (A) Canonical H3-containing nucleosome (PBD ID 1KX5) [74], with histone H3 light blue, H4 green, H2A yellow, H2B red, and DNA gray (13 bp on each end colored magenta). (B) CENP-A-containing nucleosome (PDB ID 3AN2) [20], with CENP-A colored dark blue. Bottom: view of one DNA end, with the 13 bp of disordered DNA indicated by dashed magenta lines (figure adapted from [20]). (C) Loop 1 differences in CENP-A and H3 (adapted from [20]). Top: Sequence alignment showing the two-residue insertion in CENP-A loop 1. Bottom: closeup view of the loop 1 region, with CENP-A from [20] in dark blue and H3 from [74] in light blue. Shown in orange are CENP-A residues 79–83, which are ordered in the CENP-A monomer shown (potentially due to crystal packing interactions) and disordered in the second CENP-A monomer. An earlier CENP-A₂:H4₂ tetramer structure also showed a distinct conformation of CENP-A loop 1 with high crystallographic B-factors indicating flexibility [21]. While mutation or truncation of loop 1 mildly affects centromere targeting of CENP-A in vivo [20], the mechanism of this defect remains unknown.

Figure 2

CENP-A₂/Cse4₂:H4₂ tetramer structures, in the context of the full nucleosome (left) (PDB ID 3AN2) [20], and in isolation (middle: human CENP-A₂:H4₂ (PDB ID 3NQJ) [21], right: budding yeast (K. lactis) Cse4₂:H4₂ (PDB ID 2YFW) [26]), colored as in Figure 1. Secondary structure elements of one CENP-A monomer are labeled (left panel). Both tetramer structures determined without bound H2A:H2B and DNA show compaction of the tetramer by 10–14°, relative to the conformation of both the H3₂:H4₂ and CENP-A₂:H4₂ tetramers in the context of the full nucleosome (illustration adapted from [21]).

Figure 3

Structural analysis of the CENP-A chaperone HJURP/Scm3. (A) Diagram of S. cerevisiae Scm3 and H. sapiens HJURP. Conserved domains identified in [34] are shown: CENP-A/Cse4 binding region (Scm3 residues 90–142, HJURP residues 16–68) magenta, HJURP conserved domain (228–304) cyan, HJURP repeated regions (411–462 and 556–608) yellow. The DNA-binding region of Scm3 (1–113) identified by Xiao et al [23] is shown in gray. Constructs used for x-ray crystallography are indicated by thick lines. (B) Side and top views of the CENP-A:H4:HJURP trimer structure (PDB ID 3R45) [36], with HJURP shown in magenta. (C) The CENP-A:H4:HJURP trimer aligned with a second CENP-A:H4 dimer (molecular surface displayed), showing how HJURP partially occludes the CENP-A:CENP-A interface in the tetramer (yellow arrows). (D) Two structures of the budding-yeast Cse4:H4:Scm3 trimer. Left: structure obtained by X-ray crystallography of the proteins from K. lactis (PDB ID 2YFV) [26]. Right: structure obtained by NMR of the S. cerevisiae proteins (PDB ID 2L5A) [37]. All structures are oriented the same relative to the left-hand copy of CENP-A/Cse4 in Figure 2.