Volume 154, Issue 5, Pages 1127-1139 (August 2013) S. cerevisiae Chromosomes Biorient via Gradual Resolution of Syntely between S Phase and Anaphase  Eugenio Marco, Jonas F. Dorn, Pei-hsin Hsu, Khuloud Jaqaman, Peter K. Sorger, Gaudenz Danuser  Cell  Volume 154, Issue 5, Pages 1127-1139 (August 2013) DOI: 10.1016/j.cell.2013.08.008 Copyright © 2013 Elsevier Inc. Terms and Conditions

Cell 2013 154, 1127-1139DOI: (10.1016/j.cell.2013.08.008) Copyright © 2013 Elsevier Inc. Terms and Conditions

Figure 1 WT Cells Show Bilobed Distribution of Kinetochores in Metaphase but Establish Bipolarity Only Gradually (A) Representative intensity projections of 3D image stacks of fixed WT cells coexpressing Spc42-CFP (red) marking SPBs and Ndc80-GFP (green) marking kinetochores. Scale bar, 0.5 μm. (B and C) Symmetrized Ndc80p distributions of n = 59 cells. Cells were ranked according to spindle lengths, which varied between 1 μm (bottom) and 2 μm (top) in (B). V plots show combined Ndc80-GFP intensity profiles as a function of spindle length (see Experimental Procedures). The SPB is located at the boundary of the trapezoidal intensity map. Diagonal black lines indicate loci of constant distance to the SPBs in steps of 0.2 μm. (D) Distribution of the relative intensity difference between Ndc80p bilobes (see Experimental Procedures). (E) Representative maximum intensity projections of CEN IV and SPB tags in cells with separated (top) and unseparated (bottom) sisters. Scale bar, 0.5 μm. (F and G) Time courses of distances between SPB A and SPB B (blue) and between SPB A and proximal (CEN A, green) and distal (CEN B, red) CEN IV tag. The cell in (F) has a spindle length of ∼1.4 μm with separated CEN IV tags. The cell in (G) has a spindle length of ∼1.75 μm with still-unseparated CEN IV tags. (H) Histograms of SPB-CEN distances in cells with unseparated tags (blue) and separated tags (red). For cells with unseparated tags, only the overall shortest SPB-CEN distance is included. (I) Normalized histogram of SPB-CEN distances from cells with spindle lengths between 1.5 and 2 μm. Black line, Ndc80p intensity distribution in an average spindle with a length of 1.75 μm (derived from the V plot in C). Red line, Ndc80p intensity distribution shifted by 100 nm, indicating the offset between Ndc80p and CEN IV tags. (J) Distribution of spindle lengths in CEN IV-snapshot assay and classification into unseparated (blue), separated (light red), and reclassified as separated (dark red) CEN IV tags. White solid and dashed lines indicate mean ± bootstrapped SD of cells with syntely. See Experimental Procedures and Figure S2 for details. (K) Time courses of SPB A-SPB B distances for cells entering anaphase. The time courses are aligned relative to one another with T = 0 s representing the frame in which the two chromosome tags start recoiling. (L) Normalized distribution of unseparated (blue), separated (light red), and reclassified as separated (dark red) CEN IV tag as a function of spindle length (derived from J). White lines indicate mean ± bootstrapped SD of cells with syntely. See also Figures S1, S2, and S3, Tables S1 and S2, and Movies S1, S2, S3, and S4. Cell 2013 154, 1127-1139DOI: (10.1016/j.cell.2013.08.008) Copyright © 2013 Elsevier Inc. Terms and Conditions

Figure 2 ipl1-321 Cells Show Asymmetric Kinetochore Distributions but Have Tightly Regulated Kinetochore-Microtubule Lengths (A) Representative intensity images (analogous to Figure 1A) of Spc42p and Ndc80p in ipl1-321 cells. Scale bar, 0.5 μm. (B) Distribution of the relative intensity difference between Ndc80p bilobes (analogous to Figure 1D for WT cells). Data are from n = 58 cells. Red line, distribution in WT cells. (C) Histograms of SPB-CEN distances in cells with unseparated tags (blue) and separated tags (red) (analogous to Figure 1H for WT cells). (D and E) Raw (D) and normalized (E) distributions of spindle lengths and classification into unseparated (blue) and separated (light red) CEN IV tags; ipl1-321 cells do not contain spindles with reclassified CEN IV tags. White solid and dashed lines indicate mean ± bootstrapped SD of syntely (see Figures 1J and 1L). See also Tables S1 and S2 and Movie S5. Cell 2013 154, 1127-1139DOI: (10.1016/j.cell.2013.08.008) Copyright © 2013 Elsevier Inc. Terms and Conditions

Figure 3 stu2-277 Cells Show Symmetric Kinetochore Distributions and Have Regulated Kinetochore-Microtubule Length but Establish Few Bipolar Attachments (A) Representative intensity images (analogous to Figure 1A) of Spc42p and Ndc80p in stu2-277 cells. Scale bar, 0.5 μm. (B) Symmetrized Ndc80p distributions of n = 55 stu2-277 cells (analogous to Figure 1C for WT cells). (C) Distribution of the relative intensity difference between Ndc80p bilobes (analogous to Figure 1D for WT cells). Red line, distribution in WT cells. (D) Histograms of SPB-CEN distances in cells with unseparated tags (blue) and separated tags (red) (analogous to Figure 1H for WT cells). (E and F) Raw (E) and normalized (F) distributions of spindle lengths and classification into unseparated (blue) and separated (light red) CEN IV tags; cells do not contain spindles with reclassified CEN IV tags. White solid and dashed lines indicate mean ± bootstrapped SD of cells with syntely (see Figures 1J and 1L). See also Figure S4, Tables S1 and S2, and Movie S6. Cell 2013 154, 1127-1139DOI: (10.1016/j.cell.2013.08.008) Copyright © 2013 Elsevier Inc. Terms and Conditions

Figure 4 cin8Δ Cells Show Symmetric Kinetochore Distributions and Have Less-Regulated Kinetochore-Microtubule Lengths, Leading to Transient Fusion of Bioriented Sister Kinetochores (A) Representative intensity images (analogous to Figure 1A) of Spc42p and Ndc80p in cin8Δ cells. Scale bar, 0.5 μm. (B) Symmetrized Ndc80p distributions of n = 104 cin8Δ cells (analogous to Figure 1C for WT cells). (C) Comparison of the WT (blue) and cin8Δ (red) Ndc80p distributions at 2 μm spindle length. (D) Distribution of the relative intensity difference between Ndc80p bilobes (analogous to Figure 1D for WT cells). Red line, distribution in WT cells. (E and F) Time courses of distances from SPB A to other tags (analogous to Figure 1F) for two cells with spindle midzone crossings by CEN IV tags. The example (F) shows a cell with transient separation and fusion of sister CEN IV tags, suggesting that the sisters have established biorientation but that the kinetochore-microtubule length is deregulated. (G) Kymograph aligned with respect to the position of SPB A of a cin8Δ cell entering anaphase at time ∼80 s. At T = 60 s, the CEN IV tags separate, but both cross the spindle midzone and then fuse again and cross back on the other side (time = 76 s), before they separate permanently. (H) Histograms of SPB-CEN distances in cells with unseparated tags (blue) and separated tags (red) (analogous to Figure 1H for WT cells). (I and J) Raw (I) and normalized (J) distributions of spindle lengths and their classification into spindles with unseparated (blue), separated (light red), and reclassified as separated (dark red) CEN IV tags. White solid and dashed lines indicate mean ± bootstrapped SD of cells with syntely (see Figures 1J and 1L). See also Tables S1 and S2 and Movie S7. Cell 2013 154, 1127-1139DOI: (10.1016/j.cell.2013.08.008) Copyright © 2013 Elsevier Inc. Terms and Conditions

Figure 5 Model for Mitosis Progression and Bipolarity Establishment in Budding Yeast (A) Schematic mitotic progression (from top to bottom) and bipolarity establishment in WT budding yeast cells. (B) Model changes in ipl1-321 cells. Cells have a large percentage of syntelic attachments to one SPB and, as a result, have an asymmetric Ndc80p distribution. (C) Model changes in stu2-277 cells. Cells have equal numbers of attachments to both SPBs but have a substantial fraction of monotelic attachments. The overall Ndc80p distribution is bilobed. (D) Model changes in cin8Δ cells. Cells establish bipolar attachments but do not constrain kinetochore microtubule lengths. This results in wider Ndc80p bilobes and frequent crossings of the spindle midzone by centromere tags. See also Figure S5. Cell 2013 154, 1127-1139DOI: (10.1016/j.cell.2013.08.008) Copyright © 2013 Elsevier Inc. Terms and Conditions

Figure S1 Analysis of WT Cells, Related to Figure 1 (A) 3D distribution of Ndc80p (green) and Spc42p (red) for a cell in mitosis. Planes for intensity projections are indicated by gray particles. (B) Distribution of the relative Ndc80p intensity difference for cells with spindle lengths 1.0–1.3 μm (left), 1.3–1.7 μm (middle), and 1.7–2 μm (right). (C) Model of a bilobed distribution of kinetochores mediated by complete bipolarity (D) Model of a bilobed distribution of kinetochores mediated by tensionless bipolar attachments (unseparated sisters). (E) Model of a bilobed distribution of kinetochores mediated by a mixture of bipolar attachments under tension, tensionless bipolar attachments, and syntelic attachments. (F) Four representative time-points of a movie with one hyper-stretched CEN IV-tag. The arrow heads indicates where the centromere-proximal tag recoils after t = 111 s. Cell 2013 154, 1127-1139DOI: (10.1016/j.cell.2013.08.008) Copyright © 2013 Elsevier Inc. Terms and Conditions

Figure S2 Classification of Spindles in CEN IV Assays, Related to Figure 1 (A) Scatter plot of spindle length versus ratio of mean CEN IV-tag intensity to mean SPB-tag intensity for all time points in 47 movies of WT cells filmed in the CEN IV-tracking assay. Red dots, time points with separated CEN IV tags; blue dots, time points with unseparated tags. Separated tags display a lower intensity ratio. (B) Probability distributions of mean CEN IV-tag intensity to mean SPB-tag intensity. Blue and red colors as in (A). Lines represent P(i|s), the probability of a given intensity for a tag in a state s = separated or s = unseparated. Best fits are obtained using a gamma distribution (C and D) Raw (C) and normalized (D) distributions of time points with unseparated (blue), separated (red) tags obtained from 47 movies as a function of spindle length. Black line in (D) defines P(s|l), i.e., the probability for detecting state s at a given spindle length l. (E) Bright-field image of wild-type cells overlaid by XY-positions of fluorescent tags detected in 3D. Different colors indicate tag association with individual spindles (F) Tag association to spindles by hierarchical clustering (G) Maximum intensity projection of the fluorescence images and clustered tag positions for cells in G1 (turquoise) and in mitosis (yellow and green) (H) Zoom into the clustering tree with leaves corresponding to the tags illustrated in (G). Cell 2013 154, 1127-1139DOI: (10.1016/j.cell.2013.08.008) Copyright © 2013 Elsevier Inc. Terms and Conditions

Figure S3 Analysis of Cells with Symmetric CEN IV Tag Design, Related to Figure 1 (A) Schematic of original asymmetric (top) and new symmetric (bottom) tag design. (2) denotes the centroid of the Tet operators in the asymmetric tag design. Assuming a chromatin density of 125 bp/nm (Bystricky et al., 2004), centroid (2) is ∼50 nm away from the actual position of the centromere. In the new symmetric tag design, centroids of each 6kb Tet operators are located in (3), and the overall centroid is located on top of the centromere, denoted by (1). Note that all centroids are within the Airy disk (green disk) centered on the centromere. (B) Schematic of two scenarios in which parts of the symmetric TetO/TetR-GFP arrays unravel. Top, for a short spindle (∼1.2 μm) with bipolar attachment the full intensity of the sister tags may be located within the Airy disk centered on the centromere. This fusion spot can be erroneously classified as unseparated. However, the probability of these misclassification decreases rapidly with longer spindles, in which the inter-kinetochore distance grows proportionately with the increasing SPB–SPB distance. Bottom, for a long spindle, when both tags are stretched, only ∼40% of the intensity is located within the Airy disk centered on the centromere, i.e., the resulting spot image does no longer have the properties of a diffraction-limited signal. This situation is robustly detected either by statistically testing the residuals of PSF-mixture fits to the spot (“CEN IV-tag tracking with super-resolution”), or in case of hyper-stretching by the intensity-based reclassification (“Reclassification of unseparated tags in CEN IV-snapshot assay”), which both result in a classification of the deformed spot signal as originating from bipolar attachment. (C) Scatter plot of spindle length versus ratio of mean CEN IV-tag intensity to mean SPB-tag intensity for all time points in 94 movies of WT cells filmed in the CEN IV-tracking assay. (D) Probability distributions of mean CEN IV-tag intensity to mean SPB-tag intensity. Blue and red colors as in (B). See Figure S2B for definition of fit lines. (E and F) Raw (E) and normalized (F) distributions of time points with unseparated (blue), separated (red) tags obtained from 94 movies as a function of spindle length. See Figure S2D for definition of fit line in (F). (G and H) Plots as in Figures 1J and 1L analyzing n = 925 mitotic cells with the symmetric CEN IV tag in the CEN IV-snapshot assay. (I) Classification of WT spindles with unseparated (blue), separated (light red), and reclassified as separated (dark red) CEN IV-tags. Data from Figure 1L remapped as a function of time since spindle pole separation. Black dots and error bars (Mean ± SD of 1,000 bootstrap samples), experimental data points for percentage of chromosomes with syntelic attachments. (J) Red dots are the predicted FRAP recovery based on our observed percentage of syntelic attachments in (I). Bars indicate standard deviation predicted from 1,000 bootstrap samples as in (H). Solid line represents our estimated total recovery at anaphase. Dashed lines represent standard deviations. Cell 2013 154, 1127-1139DOI: (10.1016/j.cell.2013.08.008) Copyright © 2013 Elsevier Inc. Terms and Conditions

Figure S4 Tubulin Signal for Wild-Type and Mutant Cells, Related to Figure 3 (A–C) Symmetrized tubulin distribution for: (A) wild-type cells, (B) stu2-277 cells and (C) cin8Δ cells. For comparison of the distributions the Tub1-GFP signal is normalized to a range [0,1] for each condition. Cell 2013 154, 1127-1139DOI: (10.1016/j.cell.2013.08.008) Copyright © 2013 Elsevier Inc. Terms and Conditions

Figure S5 Characteristic Time for the Resolution of One Syntelic Attachment in WT and ipl1-321 Cells, Related to Figure 5 (A and B) Dots represent the percentage of syntelic attachments as a function of time (see Extended Experimental Procedures) for WT (A) and ipl1-321 (B) cells. Lines are best fits of an exponential decay function reflecting resolution of syntely as a purely random process with characteristic time as fit parameter. Cell 2013 154, 1127-1139DOI: (10.1016/j.cell.2013.08.008) Copyright © 2013 Elsevier Inc. Terms and Conditions