Drosophila centromeres are embedded deep in heterochromatin, amid large blocks of satellite DNAs and islands of transposable elements (64,66). A Drosophila centromere "signature" sequence, like the short modular CEN DNAs of budding yeast, has eluded detection despite impressive efforts to find one (64,66). Recent thinking has shifted to the idea that Drosophila centromeres (as well as those of most other animals) are defined by a specific and heritable localized chromatin modification, rather than by a DNA sequence per se (e.g., ref. 126). Indeed, the primary constriction of a metaphase chromosome, which defines the centromere cytologically, is an obvious manifestation of altered chromatin structure. Furthermore, observations in humans, Drosophila, and other organisms of noncentromeric DNAs being able to acquire stable centromere function, so-called "neocentromeres," supports epigenetic models of centromere identity and function (e.g., refs. 127,128).
Centromere identifier (Cid) in Drosophila (129) and CENP-A in vertebrates are histone H3-like proteins that mark DNA as "centromeric" when they are incorporated into nucleosomes in place of histone H3, independent of the normal replication-coupled nucleosome assembly mechanism (see ref. 130). On stretched centromeric chromatin fibers, blocks of Cid-containing nucleosomes can be seen interspersed with blocks of H3-containing nucleosomes (131,132). Although the mechanism of Cid deposition is not yet known, the DNA associated with Cid nucleosomes appears to be replicated earlier than that associated with the interspersed H3 nucleosomes (132), which suggests that origins of replication are somehow involved.
Models for how the linear arrangement of alternating Cid and H3 subdomains might form a functional centromere in three dimensions are nicely illustrated in (131). One possibility is that Cid and H3 nucleosomes are assembled into a solenoid, with Cid nucleosomes grouped on one side of the structure and H3 nucleosomes on the opposite side, analogous to the way hydrophobic and hydrophilic amino acids are distributed to opposite sides of an amphipathic helix. On such a centromere at metaphase, kinetochore proteins could attach to the Cid side, where they would capture spindle microtubules, and on the H3 side, cohesins could bind and tie together sister chromatids (131). Possible explanations for why centromeres are embedded in large blocks of repetitive DNAs can be found in refs. 131,132.
Was this article helpful?