domain–containing protein family required for assembly of CENP-A chromatin

Centromeric chromatin, in which centromere protein A (CENP-A) replaces histone H3, is organized during mitosis to form a surface for kinetochore assembly (Palmer et al., 1991; Sullivan et al., 2001; Cleveland et al., 2003). Studies of neocentromeres, which spontaneously form at low frequency on acentric chromosome fragments, have suggested that CENP-A nucleosomes rather than specifi c DNA sequences defi ne centromere activity (Choo, 2000). The mechanisms that load CENP-A nucleosomes onto chromatin and maintain their restricted distribution are topics of active investigation (Henikoff and Dalal, 2005). Work in budding and fi ssion yeasts has identifi ed proteins that contribute to CENP-A localization (Takahashi et al., 2000; Mellone and Allshire, 2003; Hayashi et al., 2004), including the general chromatin-remodeling complex CAF-1. However, because CAF-1 inhibition destabilizes CENP-A protein levels, it is unclear whether CAF-1 has a direct role in CENP-A loading (Hayashi et al., 2004). To identify proteins that are required to localize CENP-A, we took an unbiased functional genomic approach in the nematode Caenorhabditis elegans. C. elegans has holocentric chromosomes in which kinetochores form along the entire length of each sister chromatid instead of being confi ned to a localized chromosomal region. Fundamental similarities in the mitotic kinetochores of holocentric and monocentric chromosomes are indicated by both high resolution ultrastructural studies and conservation of the constituent proteins, from CENP-A on the DNA to components of the spindle microtubule interface (for review see Maddox et al., 2004). In both types of chromosome architectures, CENP-A is restricted to a subset of chromatin organized on opposing faces of sister chromatids to geometrically constrain kinetochore assembly and ensure correct attachment to the spindle (Buchwitz et al., 1999; Oegema et al., 2001). These similarities suggest that the mechanism of CENP-A loading is likely to be conserved between holocentric chromosomes of C. elegans and monocentric chromosomes of vertebrates. In this study, we provide evidence for this assertion by identifying and characterizing a protein family with a Myb-like DNA-binding domain that is specifi cally required for CENP-A loading in both nematodes and mammalian cells. Identifi cation of this protein class provides a starting Functional genomics identifi es a Myb domain–containing protein family required for assembly of CENP-A chromatin