Fig 2: Genome-wide CRISPR screening reveals synthetic growth defects between the catalytic subunit of PP6 and the kinetochore protein NDC80. (A) A schematic depicting the role of TPX2 in stabilization of the active pool of Aurora A at the mitotic spindle. Aurora A switches between inactive unphosphorylated (T-form) and active phosphorylated (P-form) conformations. PP6 dephosphorylates the Aurora A–TPX2 complex and promotes Aurora A inactivation. Removal of PP6 thus results in amplified Aurora A activity. (B) A cartoon outlining how the enlarged spindles in PPP6C KO cells fail to maintain the compact array of chromosomes seen in parental cells during metaphase and anaphase. Escaped chromosomes in PPP6C KO cells go on to form micronuclei and cause other nuclear shape defects. (C) Pooled genome-wide CRISPR screens were performed in parental and PPP6C KO eHAP cells. Data from two independent screens were combined and analyzed using Fisher’s method to calculate Fisher’s combined P value (FCP). Significance (-log10FCP) is plotted against the LFC in PPP6C KO compared with the parental cells. Significantly positively (orange) and negatively (blue) selected genes, P < 0.01 in both screens with mean LFC < 0.25 or >0.25 are highlighted on the plot.

Fig 3: Comparative functional genomics screening for synthetic growth defects in parental and PPP6C KO eHAP cells. (A) Workflow for genome-wide CRISPR KO screens using the GeCKO V2 libraries, with data analysis in MaGeCK comparing gene selection in PPP6C KO to parental haploid eHAP cells. (B) DNA sequence of the PPP6C genomic locus showing the sequence of candidate PPP6C KO alleles in three candidate haploid eHAP cell clones. (C) Western blot of parental eHAP and candidate PPP6C KO alleles showing loss of PPP6C protein and elevation of active Aurora A pT288. (D) Parental eHAP and PPP6C KO clone #1 stained for Aurora A pT288, tubulin, and DNA. Representative cells in metaphase are shown, with arrowheads to mark the spread of active Aurora A on the mitotic spindle. Circled areas and numbers indicate the spindle diameter in µm (ø). (E) Parental eHAP and PPP6C KO clones cells stained for NUPs and DNA. Groups of interphase cells are shown, with arrowheads to indicate micronuclei. (F) Frequency plot of median enrichment of all genes and a selected set of core essential genes (CEG) in screens 1 and 2. (G) Clonogenic survival assays for three gRNA sequences targeting NDC80 in parental and PPP6C KO eHAP cell lines. Example images of survival assays are shown. (H) Clonogenic survival assays for NDC80 gRNAs g1–g3 relative to the control gRNA in eHAP cells (mean ± SD; n = 3–5). Statistical significance was analyzed using an unpaired two-tailed t test with Welch’s correction (*, P < 0.05; **, P < 0.01). (I) Western blot validation of NDC80 depletion by the NDC80 g1–g3 gRNAs in eHAP cells. Note g2 results in reduced expression and a ladder of truncated NDC80 protein species. (J) Parental and PPP6C KO HeLa cells were treated for 48 or 72 h with siRNA for the indicated negatively selected genes identified by genome-wide screening as candidates for synthetic lethality with PPP6C KO. The proportion of morphologically abnormal nuclei is plotted in the graph. Source data are available for this figure: SourceData FS2.

Fig 4: PP6 and Aurora A regulate the size of the mitotic spindle. (A) Metaphase spindle size (mean ± SD; n = 12–13) in parental and PPP6C KO HeLa cell lines stained for active Aurora A pT288, tubulin, and DNA. Statistical significance was analyzed using an unpaired two-tailed t test with Welch’s correction (***, P < 0.001). (B) Metaphase spindle size (mean ± SD; n = 15–29) in parental and PPP6C KO HeLa cell lines after 30 min treatment in the presence (+) or absence (-) of Aurora A inhibitor (AurA-i). Statistical significance was analyzed using a Brown-Forsythe ANOVA (**, P < 0.01; ***, P < 0.001; ****, P < 0.0001). (C) Parental and PPP6C KO cell lines were treated with STLC for 3 h to arrest cells in mitosis with monopolar spindles and then treated for 30 min in the absence (Control) and presence of Aurora A (AurA-i) or Aurora B (AurB-i) inhibitors. The cells were then stained for DNA and CENP-C. Monoastral spindle diameter (mean ± SD; n = 9–21) is shown for the different conditions. Statistical significance was analyzed using a Brown-Forsythe ANOVA (****, P < 0.0001). (D) Time-lapse imaging of DNA segregation in parental and PPP6C KO cells. NEBD was taken as the start of mitosis. Anaphase is shown with higher time resolution with arrows to mark anaphase spindle defects in PPP6C KO cells. Arrows indicate chromosomes escaping the anaphase spindle. (E and F) Mitotic progression from NEBD to anaphase onset (E; the line marks the median value) and cumulative mitotic index in parental and PPP6C KO cells (F; n = 26–28). PPP6C KO cells show extended mitosis and delayed mitotic exit. (G) Mitotic chromosome spreads from parental and PPP6C KO HeLa cell lines. Arrows indicate broken or unpaired chromosomes. (H) Flow cytometry was used to measure cell cycle distribution (mean ± SD; n = 3) and ploidy of parental and PPP6C KO HeLa cells. Plots show counts of DNA content with dotted lines to mark 2c and 4c in the parental control cell line.

Fig 5: Characterization of PPP6C KO HeLa cell lines. (A) Parental and candidate PPP6C KO clones were stained for DNA and NUP153. Enlarged insets show examples of nuclear morphology in the parental cells and defects in the different KO clones. Arrowheads indicate micronuclei or nuclear morphology defects. (B) To confirm PPP6C was deleted, parental HeLa and the candidate PPP6C KO clones were Western blotted for PPP6C and tubulin as a loading control. (C) Parental and PPP6C KO clone#34 in asynchronous culture or arrested in mitosis for 15 or 18 h with nocodazole were Western blotted for PPP6C, cyclin B, Aurora A, and the activating pT288 phosphorylation on Aurora A. Tubulin was used as a loading control. Source data are available for this figure: SourceData FS1.