DNA isolated from cohesin subunit ChIPs and control input DNA was labeled with aminoallyl dUTP, conjugated to Cy5 (red) or Cy3 (green) fluorescent tags, and then hybridized competitively to microarrays. Although the samples were labeled by different fluorescent tags depending on the experiment, for the purposes of analysis, the ratios are converted such that the ChIP signal is represented by red and control DNA by green. The red-to-green (R:G) ratio for each ORF and intergenic region was calculated for cells arrested in mitosis using the cdc16 mutation and assigned a color, and the median value obtained for each element was then plotted on a map of the sixteen chromosomes to determine the chromosomal distribution of Mcd1–6HAp. Regions with a R:G ratio less than 1.8 are shown in gray, and those with ratios of 1.8 or higher are shown in red. The red shading is an indicator of the intensity of cohesin subunit binding, such that regions with larger R:G ratios have lighter shades of red. Hybridization data are unavailable for regions shaded in blue (see Materials and Methods), and genomic regions not present on the arrays are indicated in white. Centromere position is indicated by an asterisk.
(A) Chromatin isolated from strains containing Mcd1-6HAp (1377A1-4B, 1829-15B, and PMY270) was immunoprecipitated using anti-HA antibodies.
(B) Chromatin isolated from strains containing Smc3-6Mycp (PMY270 and 1839-3D) was immunoprecipitated with anti-Myc antibodies.

Cells containing Mcd1-6HAp were first staged in G1 using αF, and then released from G1 into medium containing nocodazole to arrest the cells in mitosis. For the centromere excision experiments (B–D), the cultures were divided in half after G1 arrest, and one half of each culture was treated with galactose for 2 h to induce centromere excision (see Materials and Methods). Both the induced (acentric) and uninduced (centric) control cultures were then released from the G1 arrest into fresh medium and rearrested in mitosis. Once arrested in mitosis, cells were fixed in formaldehyde and then processed for ChIP using antiserum against epitope-tagged Mcd1p (Mcd1-6HAp) as an indicator of the cohesin complex. DNA isolated from the ChIPs and diluted input DNA not subject to immunoprecipitation were then subjected to PCR analysis using oligonucleotide primer pairs that amplify approximately 300-bp fragments within the indicated regions. Quantitation of DNA in the Mcd1p ChIPs, expressed as a percentage of the input DNA, is plotted as a function of the locations of the midpoints of those DNA fragments based on the SGD coordinates. Centromere position is indicated by an oval (not drawn to scale). (A) The Mcd1p association profile for the CHRI pericentric region in strain 1377A1-4B is shown. Mcd1p binding adjacent to CEN1 is difficult to assess fully because of the presence of a moderately repetitive Ty element in the region from approximately 160 to 166 kb, indicated with the dashed line. Similarly, the Mcd1p binding profiles in the pericentric regions of CHRIII (B) and CHRXIV (C) are shown in the presence (black squares) and absence (gray circles) of CEN3 and CEN14 using strains PMY185 and PMY206, respectively. (D) The Mcd1p binding profiles for a centromere-distal region of CHRIII are shown for comparison in the presence (black squares) and absence (gray circles) of CEN3.

(A) CEN1 on CHRI was replaced with a CEN3-URA3 cassette flanked by head-to-tail-oriented site-specific recombination target sites (red arrows) for the R recombinase from Zygosaccharomyces rouxii, as described in Materials and Methods. This strain (1824-23B) contained the R recombinase under the control of a galactose-inducible promoter. Genomic DNA samples, taken prior to the addition of galactose to the culture medium (0) and at 0.5-h intervals for 4.5 h after the galactose addition, were digested to completion with PvuII (black arrows) and analyzed by Southern blot analysis using a 1.25-kb probe corresponding to CHRI SGD coordinates 151823 to 153080.
(B) The percentage of centromere excision was determined for the timecourse shown in (A). Briefly, a phosphorimage of the Southern blot and ImageQuant software were used to determine the pixel intensities of the unexcised and excised bands (top and bottom bands, respectively). The percent excision was then calculated as the pixel intensity present in the excised band divided by the total pixel intensities of both bands at each timepoint.
(C) A Southern blot analysis of centromere excision from CHRIII. The endogenous CEN3 on CHRIII was replaced by R-recombinase target-site-flanked CEN3 in strain 1829-15B, as described in Materials and Methods. The efficiency of centromere excision from CHRIII was determined by Southern blot analysis in two independent experiments using genomic DNA samples digested with SnaBI and a probe corresponding to CHRIII SGD coordinates 113799-114336. Lanes 1 and 3 represent uninduced controls, and lanes 2 and 4 represent the extent of centromere excision after 2 h of recombinase induction. The percent excision was determined as in (B).

Cultures of isogenic wild-type (1846-15A) and ndc10-42 mutant (1846-15C) cells were arrested in αF at 23 °C and then released into fresh medium containing nocodazole at 37 °C. After the cells arrested in mitosis (approximately 3 h), the cultures were crosslinked with formaldehyde and processed for ChIP using a monoclonal antiserum against epitope-tagged Mcd1p (Mcd1-6HAp) as an indicator of the cohesin complex. The cohesin association profiles in the pericentric regions of CHRIII (A) and CHRI (B) are shown for NDC10 (black squares) and ndc10-42 (gray circles) cultures. The positions of the centromeres are indicated by ovals (not drawn to scale). The dashed line in (B) indicates a region containing a Ty element. (C) To identify chromosomal regions depleted for cohesin binding in the absence of a functional kinetochore, the Mcd1p-ChIP-to-input fluorescence ratio obtained for each ORF and intergenic region in genomic microarray analyses of CHRV, CHRVI, and CHRIX in ndc10-42 cells was divided by the ratio obtained for NDC10 cells and plotted on a map of the chromosomes. Regions that demonstrated 2.5-fold or greater reduction in Mcd1p binding in the ndc10-42 mutant are shaded dark green, while lighter green hues represent further fold reductions in Mcd1p binding. Regions where the magnitude of Mcd1p binding was similar in NDC10 and ndc10-42 cells are shown in gray. Gaps in the chromosomal maps are genomic regions not represented on the microarrays, while regions shaded blue were present on the arrays but gave no data during hybridizations for reasons described in Materials and Methods. The location of the centromere on each chromosome is indicated by an asterisk.

The endogenous centromere on CHRIII was removed, and CEN6 was inserted at an ectopic location (SGD coordinate approximately 260 kb), producing yeast strain PMY318, as described in Materials and Methods. Cells containing the ectopic centromere and isogenic wild-type cells (1829-15B) were staged in G1 using αF, and then released into fresh medium containing nocodazole to arrest cells in mitosis. Cells were then fixed in formaldehyde and processed for ChIP using epitope-tagged Mcd1-6HAp as a marker for the cohesin complex.
(A) The Mcd1p binding profiles at the ectopic location on endogenous CHRIII (black squares) and in the presence of the ectopic centromere (gray circles) are shown. The location of the ectopic centromere is indicated by the black oval.
(B) The levels of Mcd1p binding in the region flanking the ectopically placed centromere were divided by those observed in the isogenic wild-type control strain to determine the fold increases in Mcd1p binding in the presence of the centromere. Data are plotted as a function of the SGD coordinates for this region.

The log₂ of the median of the R:G ratios for Smc3-6Mycp (triangles) and Mcd1-6HAp (squares) binding within approximately 60-kb pericentric regions of CHRV, CHRVI, and CHRIX in cdc16-arrested cells is plotted as a function of the indicated SGD coordinates. The relative position of the centromere within each pericentric region is indicated by the oval.

As a control for the shearing of chromatin, a precipitation of chromatin was performed in each experiment using a polyclonal antiserum specific for the kinetochore protein Mif2p, which has been shown to interact with centromeric DNA (Meluh and Koshland 1997). DNA crosslinked to Mif2 was then subjected to PCR analysis throughout a 2-kb region spanning CEN3, using a series of primer pairs that amplify 240 ± 21–bp fragments. CEN3 DNA spans SGD coordinates 114382 to 114498, indicated with the ovals. Data were plotted on a scale similar to cohesin subunit ChIP data for comparison, and the inset shows in detail the magnitude of binding within a 2-kb centromere-flanking region.

The Mcd1p binding profiles in the endogenous CHRIII pericentric region are shown in cells in which the centromere is absent, either because of centromere excision (gray circles, PMY185) or because of the movement of the centromere to an ectopic location on the right arm of CHRIII (black triangles, PMY318). PMY185 and PMY318 are highly related strains; PMY185 was one of the parental strains used to generate PMY318. In the centromere excision strain, Mcd1p binding was examined in the same cell cycle in which the centromere was lost, whereas in the ectopic centromere strain, Mcd1p association was determined many generations after centromere relocation (see text for further discussion). Mcd1p binding data from the centromere excision experiment are the same as those shown in Figure 2B, but are replotted here for clarity. CEN3 normally occupies the interval between SGD coordinates 114382-114498.

Cells containing an endogenous CHRI (1377A1-4B) and those in which CEN1 was replaced with CEN3 marked with URA3 (1824-23B), as described in Materials and Methods, were staged in G1 using αF and then released into fresh medium containing nocodazole. In the case of strain 1824-23B, the G1-arrested culture was split in half, and one half was treated with galactose to induce excision of CEN3 from CHRI prior to release into medium containing nocodazole. After reaching a mitotic arrest, the cultures were crosslinked with formaldehyde and processed for ChIP using a monoclonal antiserum against epitope-tagged Mcd1p (Mcd1-6HAp). The cohesin association profiles for the modified CHRI with and without CEN3 are shown (squares and circles, respectively). The position of the centromere is indicated by the oval (not drawn to scale). The dashed line indicates the region containing a Ty element. See Figure 2A for the Mcd1p association profile of the endogenous CHRI.