Automated image analysis and systematic identification of spindle-defective mutants using morphological profiles. (A) Illustration of image processing and segmentation approach. A representative fluorescent micrograph and corresponding segmented images from a high-content screen of cells expressing GFP-Tub1p are shown. The diagram on the right illustrates the attachment of specific labels to a region of interest (cell) to extract information for specific objects (the spindle) within the region. The numbers are automatically generated by the image analysis software. Bar, 5 µm. (B) Heat maps showing distribution of wild-type, bim1Δ, and bni1Δ cells for spindle orientation. 2D histograms of the distribution of ∼35,000 wild-type (left), and similar numbers of bim1Δ (middle) and bni1Δ (right) cells assessed for spindle orientation and budding index. The x axis represents the ratio of daughter to mother area (small to large budded cells) as an indication of cell cycle stage. The y axis represents the orientation of the spindle with respect to the orientation of the mother–bud axis of the cell. Colors on the heat map represent cell number as illustrated by the key to the right of the histograms. The yellow line is a contour line to indicate a boundary that contains the majority of the cells; with heat map representations, it is difficult to gauge the quantity of cells in lower density areas and some of the more subtle trends. To help visualize the overall trends in these heat maps and to make it easier to visually compare between the wild-type, bim1Δ, and bni1Δ plots, illustrative contour lines are plotted as yellow curves to indicate boundaries that roughly contain the majority of the cells below the curves. Given the binning of a 2D histogram, the curve is plotted so that the total number of cells in the bins above the curve is no more than 1,000, which is used so as to ignore the outliers with extreme spindle orientations. The curve is then smoothed by fitting the points using a linear/quadratic rational function. (C) Profiles of spindle orientation for rcr2Δ and dyn2Δ mutant cultures. The distribution of ∼60–100 budded cells from an rcr2Δ (control, left) or dyn2Δ (right) culture for spindle orientation relative to budding index is shown. Budded cells from the mutant culture are represented by blue “X’s” overlaid on an image of the 2D histogram from the wild-type culture (see B). Images corresponding to selected cells in each culture are shown. The increased representation of cells with misoriented spindles in the dyn2Δ mutant relative to the rcr2Δ culture is emphasized by the yellow contour line (see B). Axes are as in B. Bar, 5 µm. (D) Venn diagram summarizing the number of mutants with aberrant spindle morphology identified by high-content screens of single and double deletion collections.

Categorization of top-ranked spindle mutants identified in high-content screens. The list of mutant strains identified computationally as spindle defective is shown. Mutants were manually assessed in a secondary assay and classified into five major categories: (1) delayed nuclear positioning; (2) delayed anaphase; (3) binucleated; (4) misoriented spindle; and (5) fish hook spindle. Blue, a phenotype seen only in the indicated single deletion mutant; green, a phenotype seen in the relevant bni1Δ double deletion strain; red, a phenotype for the corresponding bim1Δ double deletion mutant. Mutants are organized according to gene ontology “process” annotations.

hnt3Δ, mcm21Δ, emi1Δ, and emi2Δ mutants have spindle elongation phenotypes and components of the CPC become partially mislocalized in mcm21Δ mutant cells. (A) Quantitation of number of cells with fish hook spindles in single and double mutant cultures. A histogram illustrating the percentage of cells in the indicated mutant cultures with fish hook spindles is shown. Cells were synchronized with α factor for 90 min, released into fresh medium, and assessed for fish hook spindles after 60 min; images of at least 100 anaphase cells for each culture were captured and quantified. Error bars show the standard deviation from three independent experiments. (B, top) Ipl1p-GFP localization to the elongating spindle and spindle midzone is disrupted in mcm21Δ mutant cells. Mid-log phase cells were treated with nocodazole for 2.5 h at room temperature and released into fresh medium before imaging. DIC images, fluorescent micrographs, and merged images are shown of representative single cells of wild-type (left) or mcm21Δ (right) strains expressing Ipl1p-GFP and RFP-Tub1p. Yellow arrowheads indicate the diffused localization of Ipl1p-GFP (for the mcm21Δ cells) and the spindle midzone. Bar, 5 µm. (B, bottom) Sli15p-GFP is mislocalized, and Bir1p-GFP localization is unaffected in mcm21Δ mutant cells. Mid-log phase cells were treated with nocodazole for 2.5 h and released into fresh medium before imaging. DIC images, fluorescent micrographs, and merged images of a representative wild-type or mcm21Δ anaphase cell expressing Sli5p-GFP/Bir1p-GFP and RFP-Tub1p are shown. Yellow arrowheads indicate the localization of Sli15p-GFP. (C and D) Quantitation of the defect in Ipl1p-GFP localization at the kinetochore during metaphase and the spindle midzone during anaphase in wild-type and mcm21Δ mutant cells. (C) For kinetochore localization, cells with spindles between <2 and 2–4 µm were quantified, as there was an increased mislocalization of Ipl1p-GFP/Sli15p-GFP during early anaphase. (D) For midzone localization, cells with spindles 4 µm and longer were assessed. Cells with Ipl1p-GFP signal throughout the spindle were also considered as enriched for midzone in the quantitation. n > 60–100 cells in each spindle length category.

Genetic interactions involving the FEAR network, MEN, and kinetochore components. (A) Diagram of key components of the FEAR network and MEN. Cdc14p is released from the nucleolus during early anaphase by the FEAR network and maintained in a released state by the MEN proteins. (B) Exacerbation of the fish hook spindle phenotype of mcm21Δ mutant cells by deletion of a FEAR network activator. Representative fluorescence micrographs of GFP-Tub1p in mcm21Δ, slk19Δ, and slk19Δ mcm21Δ cells are shown. Bar, 5 µm. (C) Quantitation of the fish hook spindle phenotype for the mutants shown in B. The plot shows the percentage of anaphase cells for the indicated mutant cultures with a clear fish hook spindle phenotype. Quantitation of the defect in an mcm21Δ mutant overexpressing MOB1 is also shown. Error bars show the standard deviation from three independent experiments counting ∼100 cells for each strain. (D) Quantitation of the fish hook spindle phenotype for single and double mutants of EMI2 with FEAR and MEN components. The plot shows the percentage of anaphase cells for the indicated mutant cultures with a clear fish hook spindle phenotype. Error bars show the standard deviation from three independent experiments counting ∼100 cells for each strain. (E) Quantitation of the percentage of cells with intact spindles that show premature resequestration of Cdc14p-GFP into the nucleolus. Cells were synchronized using α factor, released into fresh medium, and monitored using live-cell imaging every 15 min for 180 min. RFP-Tub1p was used to monitor the spindle and identify the cell cycle stage. Error bars show the standard deviation from three independent experiments counting ∼80 cells for each strain in all the time points. (F) Release of Cdc14p-GFP in mcm21Δ, emi2Δ, and dbf2Δ mutant cells. Cells were synchronized using α factor, released into fresh medium, and monitored using live-cell imaging every 15 min for 180 min. RFP-Tub1p was used to monitor the spindle and identify the cell cycle stage. Representative wild-type, emi2Δ, mcm21Δ, and dbf2Δ mutant cells in late anaphase are shown. Blue arrows indicate the premature sequestration of Cdc14p to the nucleolus before spindle disassembly. Note that unlike emi2Δ and dbf2Δ cells, mcm21Δ cells maintained the released state of Cdc14p even when the spindle was hyperelongated. Bar, 5 µm.

Sumoylation of Mcm21p is not required for kinetochore localization of the CPC. (A) Sumoylation of CTF19 complex components. Immunoprecipitation/Western blotting was performed from protein extracts of wild-type or ubc9-2 strains expressing Mcm21p-TAP, Ame1p-TAP, Ctf19p-TAP, or Okp1p-TAP. A strain expressing a version of TAP-tagged Mcm21p lacking 32 lysine residues (mcm21-32R) was also analyzed for SUMO modification. The asterisk indicates sumoylated Mcm21p protein. The anti-SUMO antibody detects both sumoylated and unmodified proteins due to the associated TAP tag. (B) Subcellular localization of wild-type Mcm21p-GFP and Mcm21p-32R-GFP. DIC images, fluorescent micrographs, and merged images of representative wild-type or mcm21Δ anaphase cells expressing either wild-type Mcm21p-GFP or Mcm21p-32R-GFP and RFP-Tub1p are shown. Note the formation of a fish hook spindle in the mcm21-32R mutant background. Bar, 5 µm. (C and D) Ipl1p-GFP (C) and Sli15p-GFP (D) localization to the kinetochore in wild-type, mcm21-32R, and mcm21Δ mutant cells. Mid-log phase cells were treated with nocodazole for 2.5 h at room temperature and released into fresh medium before imaging. Fluorescent micrographs and merged images are shown of representative single cells of wild-type, mcm21Δ, or mcm21-32R strains expressing Ipl1p-GFP and RFP-Tub1p. Bar, 5 µm. (E and F) Quantitation of the defect in Ipl1p-GFP and Sli15p-GFP localization to the kinetochore during metaphase in wild-type, mcm21Δ, and mcm21-32R mutant cells. Cells with spindles between <2 and 2–4 µm were quantified for Ipl1p-GFP (E) or Sli15p-GFP localization to the kinetochore (F) as indicated. n > 200 cells in each spindle length category.

Sumoylation of Mcm21p regulates the localization of the CPC to the spindle midzone. (A and B) Ipl1p-GFP and Sli15p-GFP localization to the spindle midzone in wild-type, mcm21-32R, and mcm21Δ mutant cells. Mid-log phase cells were treated with nocodazole for 2.5 h at room temperature and released into fresh medium before imaging. Fluorescent micrographs of representative single cells of wild-type, mcm21Δ, or mcm21-32R strains expressing Ipl1p-GFP (A), Sli15p-GFP (B), and RFP-Tub1p are shown. Numbers at the top of each panel denote the spindle length (µm) range that each micrograph represents, and the numbers at the bottom represent the spindle length (µm) of the cell shown. (C and D) Quantitation of the defect in Ipl1p-GFP and Sli15p-GFP localization to the spindle midzone in wild-type, mcm21Δ mutant cells, and mcm21-32R mutant cells. Cells with spindles 4 µm and longer were used to quantify Ipl1p-GFP (C) or Sli15p-GFP (D) localization to the spindle midzone as indicated. n > 60–100 cells in each spindle length category. Cells with Ipl1p-GFP or Sli15p-GFP signal throughout the spindle were also counted as mid-zone enriched for the quantitation. (E) Localization of Ipl1p-GFP and Sli15p-GFP in ubc9-2 strains. Cells were grown at 34°C for 3 h and imaged immediately (34°C panels) or shifted to 26°C for 2 h before samples were imaged (26°C panels). DIC images, fluorescent micrographs, and merged images are shown of representative single cells. Bar, 5 µm. (F) Quantitation of the defect in spindle localization of Ipl1p-GFP and Sli15p-GFP for mutants represented in E. n > 100 cells.

Model of mitotic exit and spindle disassembly. Major regulators of the CPC during mitosis are illustrated. For simplicity, MEN and FEAR pathway components are not shown. FEAR activates the partial release of Cdc14p during early anaphase, and MEN activates the complete release of Cdc14p during late anaphase. Cdc14p release from the nucleolus and sumoylation of Mcm21p at the kinetochore seem to be two important signals that regulate the dynamic relocalization of the CPC components to the spindle midzone from the kinetochore. We suggest that Mcm21p along with the members of the CTF19 complex may act as a scaffold for the CPC. Specifically, in cooperation with Cdc14p, sumoylation of Mcm21p may increase the efficiency of timed localization of the CPC to the spindle midzone for proper disassembly of the spindle.