In vertebrates, centromeres lack defined sequences and are thought to be propagated by epigenetic mechanisms involving the incorporation of specialized nucleosomes containing the histone H3 variant centromere protein (CENP)-A. However, the precise mechanisms that target CENP-A to centromeres remain poorly understood. Here, we isolated a multi-subunit complex, which includes the established inner kinetochore components CENP-H and CENP-I, and nine other proteins, from both human and chicken cells. Our analysis of these proteins demonstrates that the CENP-H-I complex can be divided into three functional sub-complexes, each of which is required for faithful chromosome segregation. Interestingly, newly expressed CENP-A is not efficiently incorporated into centromeres in knockout mutants of a subclass of CENP-H-I complex proteins, indicating that the CENP-H-I complex may function, in part, as a marker directing CENP-A deposition to centromeres.

We previously identified a multisubunit complex (CENP-H/I complex) in kinetochores from human and chicken cells. We showed that the CENP-H/I complex is divided into three functional classes. In the present study, we investigated CENP-O class proteins, which include CENP-O, -P, -Q, -R, and -50 (U). We created chicken DT40 cell knockouts of each of these proteins, and we found that all knockout lines were viable, but that they showed slow proliferation and mitotic defects. Kinetochore localization of CENP-O, -P, -Q, and -50 was interdependent, but kinetochore localization of these proteins was observed in CENP-R-deficient cells. A coexpression assay in bacteria showed that CENP-O, -P, -Q, and -50 proteins form a stable complex that can associate with CENP-R. Phenotype analysis of knockout cells showed that all proteins except for CENP-R were required for recovery from spindle damage, and phosphorylation of CENP-50 was essential for recovery from spindle damage. We also found that treatment with the proteasome inhibitor MG132 partially rescued the severe mitotic phenotype observed in response to release from nocodazole block in CENP-50 -deficient cells. This suggests that CENP-O class proteins are involved in the prevention of premature sister chromatid separation during recovery from spindle damage.

CENP-C is a conserved inner kinetochore component. To understand the precise roles of CENP-C in the kinetochore, we created a cell line with a conditional knockout of CENP-C with the tetracycline-inducible system in which the target protein is inactivated at the level of transcription. We found that CENP-C inactivation causes mitotic delay. However, observations of living cells showed that CENP-C-knockout cells progressed to the next cell cycle without normal cell division after mitotic delay. Interphase cells with two nuclei before subsequent cell death were sometimes observed. We also found that approximately 60% of CENP-C-deficient cells had no Mad2 signals even after treatment with nocodazole, suggesting that lack of CENP-C impairs the Mad2 spindle checkpoint pathway. We also observed significant reductions in the signal intensities of Mis12 complex proteins at centromeres in CENP-C-deficient cells. CENP-C signals were also weak in interphase nuclei but not in mitotic chromosomes of cells with a knockout of CENP-K, a member of CENP-H complex proteins. These results suggest that centromere localization of CENP-C in interphase nuclei occurs upstream of localization of the Mis12 complex and downstream of localization of the CENP-H complex.

Centromere identity is thought to be determined by epigenetic mechanisms. The centromere-specific histone H3 variant CENP-A plays a central role in specifying the locus where the centromere is constructed. However, the precise mechanisms that target CENP-A to centromeric chromatin are poorly understood. Here, we show that facilitates chromatin transcription (FACT) localizes to centromeres in a CENP-H-containing complex-dependent manner. In conditional mutant cell lines for SSRP1, a subunit of FACT, centromere targeting of newly synthesized CENP-A is severely inhibited. The chromatin remodeling factor CHD1 binds to SSRP1 both in vivo and in vitro and associates with centromeres. The centromeric localization of CHD1 is lost in SSRP1-depleted cells. RNA interference knockdown of CHD1 leads to a decrease in the amount of centromere localized CENP-A. These findings indicate that the CENP-H-containing complex facilitates deposition of newly synthesized CENP-A into centromeric chromatin in cooperation with FACT and CHD1. INTRODUCTIONThe centromere is a specialized chromatin region that incorporates the centromere-specific histone H3 variant CENP-A. Although the chromatin that contains CENP-A is normally constructed on repetitive ␣-satellite DNA in human cells, the DNA sequences themselves are thought to be neither necessary nor sufficient for determination of centromere position (Cheeseman and Desai, 2008). Studies of neocentromeres have shown that functional centromeres can be ectopically constructed and stably maintained at loci that do not possess ␣-satellite sequences (Marshall et al., 2008). This indicates that the centromeric region is specified by epigenetic mechanisms rather than by cis-acting elements. One candidate for an epigenetic marker that specifies centromeres is CENP-A (Black and Bassett, 2008). CENP-A is localized at functional centromeres, including neocentromeres and is required for the centromeric localization of other centromere proteins (Warburton et al., 1997;Howman et al., 2000;Oegema et al., 2001;Goshima et al., 2003;Regnier et al., 2005;Liu et al., 2006). Ectopic mislocalization of human CENP-A, or its Drosophila counterpart CID, induces the colocalization of other centromere proteins to the noncentromeric loci (Van Hooser et al., 2001;Heun et al., 2006). Overall, the available information indicates that CENP-A is the key player for targeting centromere proteins to the correct sites by acting as an epigenetic marker. Thus, to understand how the centromere is specified in the correct chromosomal region, it is important to elucidate the mechanisms that target CENP-A to centromeric chromatin.Recent studies identified various proteins involved in the centromeric localization of CENP-A. RbAp46/48, hMis18␣/␤, and M18BP1/KNL2 are required for centromeric localization of CENP-A in vertebrates and nematodes (Hayashi et al., 2004;Fujita et al., 2007;Maddox et al., 2007). HJURP binds to CENP-A and is required for centromeric localization of CENP-A in vertebrates (Foltz et al., 2009;Dunleavy et al., 2009...

We identified CENP-50 as a novel kinetochore component. We found that CENP-50 is a constitutive component of the centromere that colocalizes with CENP-A and CENP-H throughout the cell cycle in vertebrate cells. To determine the precise role of CENP-50, we examined its role in centromere function by generating a loss-of-function mutant in the chicken DT40 cell line. The CENP-50 knockout was not lethal; however, the growth rate of cells with this mutation was slower than that of wild-type cells. We observed that the time for CENP-50-deficient cells to complete mitosis was longer than that for wild-type cells. Centromeric localization of CENP-50 was abolished in both CENP-H-and CENP-I-deficient cells. Coimmunoprecipitation experiments revealed that CENP-50 interacted with the CENP-H/CENP-I complex in chicken DT40 cells. We also observed severe mitotic defects in CENP-50-deficient cells with apparent premature sister chromatid separation when the mitotic checkpoint was activated, indicating that CENP-50 is required for recovery from spindle damage.

Chromatin structure plays a critical role in the regulation of transcription. Drosophila GAGA factor directs chromatin remodeling to its binding sites. We show here that Drosophila FACT (facilitates chromatin transcription), a heterodimer of dSPT16 and dSSRP1, is associated with GAGA factor through its dSSRP1 subunit, binds to a nucleosome, and facilitates GAGA factor-directed chromatin remodeling. Moreover, genetic interactions between Trithorax-like encoding GAGA factor and spt16 implicate the GAGA factor-FACT complex in expression of Hox genes Ultrabithorax, Sex combs reduced, and Abdominal-B. Chromatin immunoprecipitation experiments indicated the presence of the GAGA factor-FACT complex in the regulatory regions of Ultrabithorax and Abdominal-B. These data illustrate a crucial role of FACT in the modulation of chromatin structure for the regulation of gene expression.

GAGA factor is known to remodel the chromatin structure in concert with nucleosome-remodeling factor NURF in a Drosophila embryonic S150 extract. The promoter region of the Drosophila fushi tarazu (ftz) gene carries several binding sites for GAGA factor. Both the GAGA factor-binding sites and GAGA factor per se are necessary for the proper expression of ftz in vivo. We observed transcriptional activation of the ftz gene when a preassembled chromatin template was incubated with GAGA factor and the S150 extract. The chromatin structure within the ftz promoter was specifically disrupted by incubation of the preassembled chromatin with GAGA factor and the S150 extract. Both transcriptional activation and chromatin disruption were blocked by an antiserum raised against ISWI or by base substitutions in the GAGA factor-binding sites in the ftz promoter region. These results demonstrate that GAGA factor-and ISWI-mediated disruption of the chromatin structure within the promoter region of ftz activates transcription on the chromatin template.

The E3 ubiquitin ligase RNF20 regulates chromatin structure through ubiquitylation of histone H2B, so that early homologous recombination repair (HRR) proteins can access the DNA in eukaryotes during repair. However, it remains unresolved how RNF20 itself approaches the DNA in the presence of chromatin structure. Here, we identified the histone chaperone FACT as a key protein in the early steps of HRR. Depletion of SUPT16H, a component of FACT, caused pronounced defects in accumulations of repair proteins and, consequently, decreased HRR activity. This led to enhanced sensitivity to ionizing radiation (IR) and mitomycin-C in a fashion similar to RNF20-deficient cells, indicating that SUPT16H is essential for RNF20-mediated pathway. Indeed, SUPT16H directly bound to RNF20 in vivo, and mutation at the RING-finger domain in RNF20 abolished its interaction and accumulation, as well as that of RAD51 and BRCA1, at sites of DNA double-strand breaks (DSBs), whereas the localization of SUPT16H remained intact. Interestingly, PAF1, which has been implicated in transcription as a mediator of FACT and RNF20 association, was dispensable for DNA-damage-induced interaction of RNF20 with SUPT16H. Furthermore, depletion of SUPT16H caused pronounced defects in RNF20-mediated H2B ubiquitylation and thereby, impaired accumulation of the chromatin remodeling factor SNF2h. Consistent with this observation, the defective phenotypes of SUPT16H were effectively counteracted by enforced nucleosome relaxation. Taken together, our results indicate a primary role of FACT in RNF20 recruitment and the resulting chromatin remodeling for initiation of HRR.