Centromeric Chromatin and Epigenetic Effects in Kinetochore Assembly

Abstract

The formation of centromeric chromatin can be divided into distinct, but related, steps of nucleation and propagation. We can envision three obvious mechanisms to control the establishment of centromeric chromatin. First, nucleation may be controlled by DNA sequence. In S. cerevisiae, DNA sequence is clearly of paramount importance, but even in higher cells centromere-specific DNA must play a role because kinetochores on normal chromosomes are never found at neocentromere-competent locations. Second, nucleation may be controlled by strictly limiting the amounts or activity of key trans-acting factors. In budding yeast for example, it appears that a key component of CBF3 is controlled by a cycle of coupled activation and destruction so that the number of CBF3 complexes able to initiate kinetochore assembly barely exceeds the number of chromosomes (Kaplan et al. 1997xKaplan, K.B, Hyman, A.A, and Sorger, P.K. Cell. 1997; 91: 491–500Abstract | Full Text | Full Text PDF | PubMedSee all ReferencesKaplan et al. 1997). In HeLa cells, the restriction of CENP-A expression to late S/G2 is essential for its correct incorporation into centromeric chromatin, since ectopic expression throughout S phase causes CENP-A to accumulate nonspecifically in the nucleus, and this accumulation appears to be toxic (Shelby et al. 1997xShelby, R.D, Vafa, O, and Sullivan, K.F. J. Cell Biol. 1997; 136: 501–513Crossref | PubMed | Scopus (215)See all ReferencesShelby et al. 1997). Third, nucleation could be spatially restricted along an entire chromosome by signals that emanate from particular chromosome structures such as kinetochores or telomeres (Williams et al. 1998xWilliams, B.C, Murphy, T.D, Goldberg, M.L, and Karpen, G.H. Nat. Genet. 1998; 18: 30–37Crossref | PubMed | Scopus (141)See all ReferencesWilliams et al. 1998). The recombinational enhancer on the left arm of budding yeast chromosome III provides a precedence for a chromosome-wide effect that is controlled by a localized DNA-bound structure (Szeto et al. 1997xSzeto, L, Fafalios, M.K, Zhong, H, Vershon, A.K, and Broach, J.R. Genes Dev. 1997; 11: 1899–1911Crossref | PubMedSee all ReferencesSzeto et al. 1997). The binding of Mcm1p protein to two specific sites activates recombination along the entire chromosome arm, and the association of the mating-type-specific repressor α2p with Mcm1p blocks this activation. Thus, two Mcm1p/α2p-DNA complexes can regulate, via an as-yet-unknown mechanism, an activity that occurs along a chromosome. An analogous system emanating from a kinetochore could, in a Robertsonian translocation, prevent the potentially destructive assembly of two kinetochores. In a marker chromosome, the absence of a normal kinetochore would make chromatin assembly permissive and allow the selection of a neocentromeric site. Once a kinetochore had assembled on this neocentromere, further kinetochore assembly on the same chromosome would then be inhibited.To explain how neocentromeres can be transmitted accurately during cell division, it seems necessary to postulate that centromeric chromatin is templated from preexisting chromatin on a parental chromosomes. The templating of centromeric chromatin may involve mechanisms analogous to those that control the propagation of active and inactive states of silenced or telomeric chromatin (seeGrunstein 1998xGrunstein, M. Cell. 1998; 93: 325–328Abstract | Full Text | Full Text PDF | PubMed | Scopus (230)See all ReferencesGrunstein 1998 [this issue of Cell). However, other chromatin-dependent phenomena are known to be metastable and to switch states every 100 divisions or so. Clearly, the requirement for one and only one centromere per chromosome is inconsistent with this degree of variability. Thus, centromeric chromatin must differ from other chromatin in the very tight control that is exerted over its formation and propagation. It is unclear whether an extension of existing models for heritable states of chromatin is sufficient to explain the high fidelity of kinetochore assembly, or whether additional mechanisms must be postulated. In the future we will have to understand how, when it comes to kinetochores, there can be only one.