Ybp2 Associates with the Central Kinetochore of Saccharomyces cerevisiae and Mediates Proper Mitotic Progression

The spindle checkpoint ensures the accurate segregation of chromosomes by monitoring the status of kinetochore attachment to microtubules. Simultaneous mutations in one of several kinetochore and cohesion genes and a spindle checkpoint gene cause a synthetic-lethal or synthetic-sick phenotype. A synthetic genetic array (SGA) analysis using a mad2Δ query mutant strain of yeast identified YBP2, a gene whose product shares sequence similarity with the product of YBP1, which is required for H2O2-induced oxidation of the transcription factor Yap1. ybp2Δ was sensitive to benomyl and accumulated at the mitotic stage of the cell cycle. Ybp2 physically associates with proteins of the COMA complex (Ctf19, Okp1, Mcm21, and Ame1) and 3 components of the Ndc80 complex (Ndc80, Nuf2, and Spc25 but not Spc24) in the central kinetochore and with Cse4 (the centromeric histone and CENP-A homolog). Chromatin-immunoprecipitation analyses revealed that Ybp2 associates specifically with CEN DNA. Furthermore, ybp2Δ showed synthetic-sick interactions with mutants of the genes that encode the COMA complex components. Ybp2 seems to be part of a macromolecular kinetochore complex and appears to contribute to the proper associations among the central kinetochore subcomplexes and the kinetochore-specific nucleosome.


Introduction
The centromere interacts with the spindle fibers to ensure the correct segregation of chromosomes during mitotic and meiotic cell divisions. It requires DNA sequence elements and structural and regulatory proteins for its activity and coordination within the cell cycle. Kinetochores are specialized protein complexes that assemble on the centromeric (CEN) DNA and mediate the attachment of duplicated chromosomes to the mitotic spindle; they also generate signals to arrest cell cycle progression if metaphase is not achieved properly (reviewed in [1,2]). After spindle pole body duplication, sister centromeres separate transiently and oscillate along the spindle axis until anaphase, when they separate permanently [3][4][5][6].
The budding yeast Saccharomyces cerevisiae has short 125-bp point centromeres and its kinetochores bind to single microtubules. The kinetochore consists of 3 protein layers (inner, central, and outer) that assemble hierarchically onto the CEN DNA (reviewed in [7][8][9]). The inner kinetochore proteins are in direct contact with the CEN DNA. In S. cerevisiae, the inner kinetochore consists of the CBF3 complex and a specialized nucleosome containing the conserved histone H3-like protein Cse4 (Cnp1 in fission yeast and CENP-A in higher eukaryotes). The CBF3 complex is composed of 4 essential proteins: Ndc10/Ctf14/Cbf2, Cep3/Cbf3b, Ctf13, and Skp1. The inner kinetochore complex binds to CEN DNA and is required for the centromeric association of all other kinetochore proteins. The central kinetochore consists of at least 3 major subcomplexes: the COMA complex (Ctf19, Okp1, Mcm21, and Ame1); the MIND complex (Mtw1, Nnf1, Nsl1, and Dsn1); and the Ndc80 complex (Ndc80, Nuf2, Spc24, and Spc25). The central kinetochore complex connects the inner kinetochore to various microtubule-binding proteins. The outer kinetochore consists of the Dam1 complex, which contains 9 or more subunits. The outer kinetochore proteins associate with microtubules for their kinetochore association. Recent studies have implicated the conserved Ndc80 complex in several essential outer kinetochore functions, including microtubule binding and control of a safety device known as the spindle checkpoint [10][11][12][13].
In addition to kinetochore proteins, numerous other proteins are integral to chromosome stability, including spindle checkpoint proteins, motor proteins, microtubule-associated proteins, regulatory proteins, and proteins implicated in CEN DNA chromatin dynamics, structure, and sister chromatid cohesion [1,7,14,15]. Proteomic approaches have identified structural components of the kinetochore, and genetic approaches have identified various proteins important for chromosome segregation in yeast. For instance, a chromosome transmission fidelity (CTF) screen has identified mutations in genes encoding DNA replication, cohesion, and kinetochore proteins [16]. Synthetic dosage lethal (SDL) screens in which mutants that cannot tolerate overexpression of kinetochore proteins have identified chromosome stability genes, many of which are not components of the kinetochore (e.g., chromatin-modifying or tubulin-folding proteins) [17][18][19][20][21][22][23]. Synthetic genetic array (SGA) analysis has enabled investigators to perform systematic genome-wide genetic screens in yeast [24].
The spindle checkpoint ensures accurate chromosomal segregation by monitoring that anaphase is initiated only after proper kinetochore-microtubule associations of all sister chromatids occur. It thereby safeguards against missegragation events during mitosis and meiosis; recent studies have shown that spindle checkpoint dysfunction might facilitate tumor progression (reviewed in [25,26]). Chief components of the spindle checkpoint complex in S. cerevisiae include the mitotic arrest-defective (MAD) genes MAD1-3 [27] and the budding-uninhibited-by-benzimidazole (BUB) genes BUB1 and BUB3 [28]. Simultaneous mutations in one of several kinetochore and cohesion genes, such as CTF19 and CTF8, and a spindle checkpoint gene cause a synthetic-lethal or synthetic-sick phenotype [29][30][31]. Thus, a synthetic lethality screen that uses a spindle checkpoint mutant should identify genes whose functions are monitored by the spindle checkpoint. We therefore characterized the mitotic function of Ybp2 that was identified by an SGA screen by using the mad2-deletion (mad2D) strain.

Results
Synthetic Genetic Array (SGA) Screen of mad2D Mutants Identified YBP2 as a Mitotic Factor To identify genes that regulate mitosis in yeast, an SGA screen of the mad2D strain as a query mutation was performed against a complete set of deletion-mutant yeast strains [24]. SGA allows crossing of the query mutation into the set of viable deletion mutants, thereby allowing the resulting double mutants to be screened for synthetic-lethal or synthetic-sick interactions. The mad2D SGA screen detected previously described kinetochore proteins, microtubule-binding proteins, chromatin proteins, and cohesion proteins [24]. Mutants of most known nonessential outer and central kinetochore genes were synthetic-lethal or synthetic-sick when combined with the mad2D mutation, confirming that the screen worked as expected. Of the 32 genes detected, 4 have not been previously identified or characterized as mitotic factors. One of them has recently been named Ybp2 because of its homology with Ybp1 (Yap1-binding protein 1), a protein required for oxidation of the transcription factor Yap1 in response to H 2 O 2 [32].
We found that the ybp2D strain was sensitive to benomyl ( Figure 1A) and accumulated at the G2/M stage of the cell cycle ( Figure 1B), traits common to most kinetochore mutants. These results suggest that YBP2 is required for mitotic function.

Phenotypic Analysis of the ybp2D Mutant
To further examine the potential synthetic lethality of ybp2D with mad2D, we constructed a ybp2Dmad2D double mutant carrying the MAD2 CEN-URA3 plasmid. We used the 5fluoroorotic acid (FOA) that selects against strains that contain the URA3 gene. The double mutant grew on a 5-FOA plate (unpublished data), indicating that the interaction between YBP2 and MAD2 is not a straightforward synthetic-lethal interaction. Next, we constructed a ybp2Dmad2D double mutant in our background (YPH499) to assess the double mutant for a more subtle synthetic fitness defect. The ybp2Dmad2D double-mutant cells grew relatively normally at 30uC but failed to grow at 37uC ( Figure 1C). The double mutants of ybp2D and other spindle checkpoint mutants (ybp2Dbub1D, ybp2Dmad1D, ybp2Dmad3D, and ybp2Dbub3D) showed the same temperature sensitivity ( Figure 1C). Thus, YBP2 has a temperature-sensitive conditional syntheticlethal interaction with spindle checkpoint genes.
The ybp2D mutant arrested with largely G2/M DNA content in haploid cells containing a nonessential marked chromosome fragment ( Figure 1B) and in diploid cells (Supplemental Figure  S1), but, surprisingly, not in normal haploid cells that did not contain an extra chromosome fragment (unpublished data). Quantification of cell and nuclear morphology in haploid ybp2D cells containing the chromosome fragment revealed that ,45% of the cells were large budded, with an undivided nucleus positioned at or near the neck between the mother and daughter cells ( Figure 1D). The G2/M delay caused by ybp2D was abolished in the absence of Mad2 (Figure 2A), which indicates that the G2/M delay is dependent on the spindle checkpoint.
We examined chromosome stability in ybp2D mutant cells by a colony color assay, which measures the stability of a marker chromosome fragment [33]. The ybp2D mutant cells showed only a moderate chromosome-missegregation phenotype (chromosome fragment loss 1.0%) (Figures 2B and D and Supplemental Figure  S2). However, ybp2Dmad2D exhibited an enhanced chromosomemissegregation phenotype (chromosome fragment loss: 42%) as compared with either single mutant ( Figure 2B). These results suggest that the mitotic defect caused by deletion of YBP2 is monitored by the spindle checkpoint, and thus the ybp2D cells are protected from chromosome missegregation.
The synthetic lethal screen is generally employed to identify genes encoding proteins that function in a parallel pathway or with overlapping functions within an essential pathway. Because we used the spindle checkpoint mutant mad2D as a query, we examined whether Ybp2 functions in the spindle checkpoint. A major phenotype of spindle checkpoint mutants is repeated budding without mitotic arrest in the presence of the microtubule-depolymerizing drug nocodazole. However, ybp2D mutant cells arrested in the G2/M phase of the cell cycle (Supplemental Figure S3A), indicating that the spindle checkpoint is activated in response to unattached kinetochores.
Ctf8 is required for proper sister chromatid cohesion [30]. In ctf8 mutants, microtubules can attach to kinetochores but the tension at kinetochores is reduced because the linkage between sister chromatids is compromised. We arrested ctf8D, ctf8Dybp2D, and wild-type cells in G1 with a-factor at 30uC and then released them into a YPD medium without a-factor. There was a delay in Pds1 degradation in ctf8Dybp2D and ctf8D cells, indicating spindle checkpoint activation (Supplemental Figure S3B). These results indicate that Ybp2 is not required for spindle checkpoint function.
In addition, we found that ybp2D cells did not exhibit sister chromatid cohesion defects by the cohesion assay, using a strain an array of lactose operators integrated at the TRP1 locus, 12 kb from the CEN of chromosome IV and expressing a GFP-lactose repressor (GFP-lacI) [34,35] (Supplemental Figure S4).

Ybp1 does not Have a Mitotic Function
Recent studies suggest that Ybp2, which is homologous (35% identity) to Ybp1, might be required for oxidative stress tolerance because Ybp1 is a positive regulator of oxidative stress tolerance [32,36]. Therefore, we examined whether Ybp1 is also required for a mitosis function. The ybp1D mutant cells showed no benomyl sensitivity, and there was no enhanced benomyl sensitivity in the ybp2Dybp1D double mutant ( Figure 2C). Also, ybp1D cells did not show a chromosome-missegregation phenotype, and deletion of YBP1 did not enhance the missegregation phenotype of ybp2Dybp1D cells ( Figure 2D). These results suggest that Ybp1 does not play a significant role in the mitotic function of Ybp2.  Ybp2 Interacts with Several Central Kinetochore Proteins and the Centromeric Histone Cse4 To examine the interaction between Ybp2 and known kinetochore proteins, we generated antibodies against Ybp2 and strains that express several myc-tagged kinetochore proteins under their own promoters. We also confirmed that all tagged proteins were functional (unpublished data). Ybp2 or Ybp2-myc coimmunoprecipitated with all components of the COMA complex (Ctf19, Okp1, Mcm21, and Ame1; Figure 3A); 3 of 4 components of the Ndc80 complex (Ndc80, Nuf2, and Spc25 but not Spc24; Figure 3B); and 1 of 4 components of the MIND complex (Nsl1 but not Mtw1, Nnf1, or Dsn1; Figure 3C). Ybp2 coimmunoprecipitated with Ctf3, Mcm16, Mcm22, Chl4, and Spc105 ( Figure 3D). Interestingly, Ybp2-myc interacted with Cse4, the centromeric histone H3 (yeast CENP-A homolog) ( Figure 3E). Ybp2 did not coimmunoprecipitate with Ipl1, Slk19, Mif2, Ndc10 (Supplemental Figure S5), Duo1, Nkp1, or Skp1 (unpublished data). These results indicate that Ybp2 associates with the central kinetochore complexes, especially the COMA and Ndc80 complexes, and the centromeric nucleosomes ( Figure 4G).
Because previous chromatin-immunoprecipitation (ChIP) assays have shown that all the central kinetochore proteins characterized associate with CEN DNA [1], we performed ChIP assays by using Ybp2-myc strains. Myc-tagged Ybp2 was immunoprecipitated from cross-linked extracts with anti-myc-conjugated beads. The coimmunoprecipitated DNA was analyzed by PCR with primers specific to centromeric regions of chromosomes I and XVI and to the noncentromeric region PGK1 as a control for binding specificity. The Okp1 myc-tagged strain was used as positive control. Ybp2 coimmunoprecipitated specifically with CEN DNA ( Figure 4A and Supplemental Figure S6A), indicating that it associates with CEN DNA in vivo. This association was abolished in ndc10-1 mutant cells ( Figure 4B and Supplemental Figure S6B), indicating that the association of Ybp2 with CEN DNA is dependent on Ndc10 and thereby the CBF3 complex. These results reveal that Ybp2 is part of the macromolecular kinetochore complex.
We proposed a model that summarizes the immunoprecipitation and ChIP results ( Figure 4C). Ybp2 interacts with proteins in the COMA and Ndc80 complexes, but does not substantially interact with those in the MIND complex. Therefore, these 3 complexes appear to have a 3-dimensional surface. We hypothesized that Ybp2 might be localized among the 3 complexes ( Figure 4C) to bridge COMA and Ndc80 onto the centromeric nucleosome.

Interactions among the 3 Central Kinetochore Subcomplexes Are Increased in ybp2D Mutant Cells
Given the interaction of Ybp2 with the central kinetochore subcomplexes (COMA, MIND, and Ndc80), we expected that Ybp2 might be important for the interaction between the Ndc80 and the COMA complexes ( Figure 4C). As per this hypothesis, the Ndc80-COMA association would be disrupted in the ybp2D background but the Ndc80-MIND and the COMA-MIND associations would not ( Figure 5A). To test this hypothesis, we performed coimmunoprecipitation assays using ybp2D mutant cells. The Spc25-Mtw1 interaction for the Ndc80-MIND association, the Mtw1-Ctf19 interaction for the MIND-COMA association, and the Ndc80-Ctf19 interaction for the Ndc80-COMA association ( Figure 5A) have been previously reported [37,38]. First, we tested the interaction between Spc25 in the Ndc80 complex and Mtw1 in the MIND complex. We constructed an Spc25-myc Mtw1-HA double-tagged strain in the wild-type or ybp2D strain, performed an anti-myc immunoprecipitation assay, and identified the Mtw1-HA tagged protein by Western blot. The interaction of Spc25 with Mtw1 was slightly increased or unaffected in the absence of Ybp2 ( Figure 5B). Next, we checked whether the lack of Ybp2 disrupted the interaction between Ctf19 in the COMA complex and Mtw1 in the MIND complex. The Ctf19-Mtw1 interaction was slightly increased in the absence of Ybp2 ( Figure 5C). Finally, we checked whether the lack of Ybp2 disrupted the interaction between Ctf19 in the COMA complex and Ndc80 in the Ndc80 complex. Unexpectedly, the Ctf19-Ndc80 interaction was increased in the absence of Ybp2 ( Figure 5D). Together, these results indicate that the lack of Ybp2 enhances the interactions among the central kinetochore subcomplexes (COMA, MIND, and Ndc80).
Next, we examined the association of the central kinetochore proteins with CEN DNA in the absence of Ybp2. The association of Ndc80, Ctf19 and Mcm21 but not Mtw1, Nuf2 or Dsn1 with CEN DNA increased in ybp2D cells ( Figures 5E and F), which is consistent with the protein-protein interaction results. Taken together, it appears that Ybp2 may push the central kinetochore subcomplexes away rather than hold them together.

YBP2 Genetically Interacts with the Genes that Encode the COMA Components
Because Ybp2 appears to coprecipitate more efficiently with Ctf19 and Mcm21 ( Figure 3A) than with other kinetochore proteins and because the CEN binding of Ctf19 and Mcm21 was dramatically affected in the absence of Ybp2 ( Figures 5E and F), we tested the genetic interaction between YBP2 and CTF19 or MCM21. All deletion mutants (ybp2D, ctf19D, and mcm21D) grew normally on YPD at 25uC or 37uC, but the ybp2Dctf19D and ybp2Dmcm21D double mutants and the ybp2Dctf19Dmcm21D triple mutant failed to grow at 37uC ( Figure 5G, top and middle). On the other hand, the ctf19Dmcm21D double mutant grew normally on YPD at 37uC ( Figure 5G, top and middle). Thus, a ybp2D mutation genetically interacts with mutations in CTF19 or MCM21, which suggests that Ybp2 supports the essential function of the COMA complex. Consistent with these observations, ybp2D exhibited temperature-sensitive conditional synthetic-lethal interactions with ame1-4 or okp1-5, respectively ( Figure 5G, bottom).

Discussion
In this study, we characterized the mitotic function of Ybp2, which was identified by an SGA screen by using the mad2-deletion (mad2D) strain. ybp2D mutants displayed phenotypes commonly observed in kinetochore mutants, such as sensitivity to the benomyl, G2/M accumulation when cycling, and chromosome missegregation. Ybp2 associates with several central kinetochore proteins and the centromeric histone Cse4. It also associates with CEN DNA in an Ndc10-dependent manner. YBP2 shows genetic interactions with the kinetochore genes CTF19 and MCM21. These results strongly support the role of Ybp2 in kinetochore function.

SGA Screens That Use a Spindle Checkpoint Mutant Identified Mitotic Factors
An SGA screen that uses a mad2D mutant strain identified the 4 novel mitotic genes (YBR194W, YDR359C, YGL060W, YJL064W) [24]. When the screen identified the genes, no gene had been previously characterized. However, while we were characterizing these genes, studies on YDR359C and YGL060W were published by other investigators, who designated these genes as VID21 [39] and YBP2/YBH1 [36], respectively. Vid21 is required for bulk histone H4 acetylation and is functionally linked to the histone H2A variant Htz1 and the Swr1 ATPase complex (SWR-C) that assembles Htz1 into chromatin [39]. Ybp2 The indicated strains were grown to log phase at 30uC, lysed, and anti-myc immunoprecipitations were performed. Total lysate (T), supernatant (S), and the immunoprecipitated fraction (IP) were subjected to SDS-PAGE, and Western blots were used to detect Ybp2, Ctf19, Cse4, or myc-tagged proteins with the respective antibodies. For the quantification of Ybp2, aliquots of the total lysate (1/500 of the total), supernatant (1/500 of the total), and IP fraction (2/5 of the total lysate) were loaded. Arbitrary number is defined as the ratio of the amount of coprecipitated protein to the amount of input protein. Intensity of the total lysate is influences H 2 O 2 tolerance [36], though its role in this function appears marginal. The functions of the other genes are yet unknown. However, YJL064W is dubious and overlaps with DLS1; therefore, DLS1 may be the real target of Mad2. Genetic interactions between YBP2 and MAD2 have also been identified by Daniel et al. [40], thus providing an independent confirmation of our findings. Importantly, none of these genes, including YBP2, has been identified as a mitotic factor by previous proteomic methods. Thus, genetic screens complement proteomic approaches and hold promise for identifying novel mitotic factors that are functionally related to known mitotic factors.

Kinetochore Function of Ybp2
Based on protein-protein and protein-DNA coimmunoprecipitation results, we describe a Ybp2 network of physical interactions in the kinetochore. Ybp2 interacts with proteins in the COMA and Ndc80 complexes, but not substantially with those in the MIND complex. We hypothesized that the 3 complexes have a 3dimensional surface by assuming that Ybp2 is a structural component of the central kinetochore. However, coimmunoprecipitation data generated by using the ybp2D mutant suggest that lack of Ybp2 enhances the interactions among the central kinetochore subcomplexes. ChIP assays showed that CEN binding of Ctf19, Mcm21, and Ndc80 also increased in ybp2D cells. Interestingly, Ybp2 localizes throughout the cell (Supplemental Figure S7, [36]). One possibility is that Ybp2 may not be merely a structural component of the kinetochore complex, but have a chaperone function by helping arrange the central kinetochore subcomplexes.
When the spindle checkpoint is active, loss of Ybp2 does not cause a substantial mitotic defect but does cause a mitotic delay. When the spindle checkpoint is inactive, the loss of Ybp2 results in substantial chromosome loss, suggesting that the spindle checkpoint detects the mitotic defect and suppresses it by delaying mitosis.
We also observed a synthetic sickness phenotype for ybp2Dctf19D and ybp2Dmcm21D double mutants. Apart from Okp1 and Ame1, Ctf19 and Mcm21 are nonessential in the COMA complex. Because we did not see any growth defect in the ctf19Dmcm21D double mutant, Ybp2 may support the essential function of the COMA complex more specifically than it supports that of the other central kinetochore complexes.
The ybp2D mutant arrested with G2/M DNA content in haploid cells containing a nonessential marked chromosome fragment and in normal diploid cells, but not in normal haploid cells. This interesting finding suggests that there is some specific mechanistic difference between the mitotic chromosome segregation of aneuploid haploids and normal diploid cells, but further investigation is required to address this issue.
Ybp2 has 3 sequence blocks that are homologous to the S. cerevisiae kinetochore protein Slk19 (Supplemental Figure S8A). These blocks are also conserved in the human homologue of Slk19, CCDC73 (unpublished data). Slk19 is important for proper chromosome segregation and is found at kinetochores in metaphase and at the spindle midzone in anaphase [41]. Slk19 was identified as a component of the FEAR (Cdc Fourteen Early Anaphase Release) network, which promotes Cdc14 release from the nucleolus during early anaphase [42]. We found that the ybp2D mutation genetically interacts with the slk19D mutation (Supplemental Figure S8B). Therefore, Ybp2 may be functionally related to Slk19.

YBP2 is a Polymorphic Gene
We have found polymorphisms in YBP2. A sequence comparison of YBP2 of BY4741 (SGA background) and YPH499 (our background) has revealed 7 different bases in the open reading frame (1926 bp), of which 3 affect the amino acid sequences (Supplemental Figure S9). We have not found any functional relevance of these polymorphisms yet.

Yeast strains and medium
Supplemental Table S1 presents the genotypes of yeast strains used for this study. The medium for yeast growth and sporulation has been previously described [43]. Yeast strains were transformed by the method of Ito and coworkers [44]. Strains that expressed myc-tagged proteins were generated by the procedure of Longtine and coworkers [45]. For the microtubule-depolymerizing drug sensitivity assay, benomyl (DuPont, Wilmington, DE) was added at the indicated concentration to the YPD medium; only dimethyl sulfoxide was added to the YPD medium as a control.

Colony color assay
The colony color assay was performed as previously described [33,46]. In brief, each strain containing a single SUP11-marked chromosome fragment was plated at a density of ,200 colonies per plate on minimal (SD) medium containing a limiting amount of adenine (6 mg/mL) and grown at 30uC. A colony consists of cells containing the chromosome fragment (white) and cells that have lost it (red), resulting in a white-and-red sectored phenotype.

ChIP assay and coimmunoprecipitation analyses
Immunoprecipitation using yeast lysates was performed as described previously [47]. ChIP assays of yeast lysates were performed by a previously described method [20], with the  Anti-myc chromatin immunoprecipitation (ChIP) assays were performed from log-phase cells at 30uC (A) or 25uC (B). Total lysate (T, 1/800 of the total) and coimmunoprecipitated DNA (IP, 1/10 of the total) were analyzed by PCR with primers specific to centromeric regions of chromosome I and XVI and to a noncentromeric region (PGK1) as a control for binding specificity. The myc-tagged Okp1 strain was used as positive control. For the quantification of CEN1 and CEN16, aliquots of the total lysate (2/25 of the total) and IP fraction (2/5 of the total) were loaded on an 8% acrylamide gel. Arbitrary number is defined as the ratio of the amount of coprecipitated DNA to the amount of input DNA. Intensity of the total lysate is defined as 1 in each strain. The yeast strains used were (A) untagged (YPH499), Ybp2-myc (Y1689), and Okp1-myc (Y1707); and (B) untagged (YPH499), Ybp2-myc (Y1689), Ybp2-myc ndc10-1 (Y1838), and Okp1-myc (Y1707). (C) A model of the role of Ybp2 in the central kinetochore. Ybp2 interacts strongly with proteins in the COMA and Ndc80 complexes (shown in red) but not as well with those in the MIND complex and other kinetochore proteins (shown in brown). We propose that Ybp2 is localized to bridge the COMA and Ndc80 complexes onto the centromeric nucleosome. doi:10.1371/journal.pone.0001617.g004  [50]. The expected sizes of PCR products were 302 bp (CEN1), 344 bp (CEN16), and 288 bp (PGK1). The total chromatin added varied from 1/1500 to 1/600 and the immunoprecipitated chromatin template varied from 1/10 to 1/30 of the available template, depending on the linear range for PCR. Band intensities of X-ray films and photographs were quantified with the Quantity One (Version 4.6.2) software. Binding activity was defined as the ratio of the amount of coprecipitated protein or DNA to the amount of input protein or DNA.