CKAP5 stabilizes CENP-E at kinetochores by regulating microtubule-chromosome attachments

We first sought to characterize the chromosomal defects in response to the loss of CKAP5. CKAP5 was depleted by using esiRNA in cultured human cell lines, HeLa Kyoto, and U2OS. CKAP5 level was reduced by ~85% by esiRNA in HeLa Kyoto cells (Fig. EV1C). Depletion of CKAP5 led to severe chromosome congression defects in the majority of the mitotic cells. Consistent with earlier studies (Gergely et al, 2003), ~60% of mitotic cells showed partially aligned chromosomes at the metaphase plate and a few uncongressed chromosomes dispersed near the poles (Fig. EV1A,B). We then looked into the localization of KT proteins specific to different layers of KT in the CKAP5-depleted cells. It revealed that the fibrous corona and outer KT-associated protein, CENP-E was specifically lost from the partially aligned metaphase plate KTs in HeLa Kyoto cells upon depletion of CKAP5 (Fig. 1A,B). KT intensity of CENP-E in HeLa Kyoto cells was reduced by ~85% at the aligned KTs (Fig. 1C). CKAP5 depletion-induced loss of CENP-E was also apparent at the KTs of prometaphase cells (Fig. 1D–F). Prometaphase arrest was induced by treating the cells with Eg5 kinesin inhibitor, DMA (Dimethylenastron) (Tamura et al, 2015). CKAP5 depletion-induced CENP-E mis-localization from KTs was not due to any change of CENP-E expression as the overall cellular level of CENP-E was not altered upon CKAP5 depletion (Fig. EV1D). Similar results were obtained in U2OS cells depleted of CKAP5 by esiRNA and also in inducible Cas-9 expressing CKAP5 knockout HeLa cells (Fig. EV1E–G).

CKAP5 can localize to KTs even in the absence of MTs (Herman et al, 2020). To determine if the abrogation of CENP-E from KTs upon CKAP5 depletion can occur in the absence of microtubules, we checked CENP-E localization in cells by treating with 3.3 µM nocodazole (Sacristan et al, 2018), which completely depolymerizes MTs. KT localization of CENP-E was unaffected in the CKAP5-depleted cells upon depolymerization of the MTs (Fig. 1G), suggesting that CKAP5 depletion-induced CENP-E abrogation from the KTs involves KT-attached microtubules. Dynein plays an important role in removing fibrous corona proteins including CENP-E from MT-attached KTs (Howell et al, 2001; Sacristan et al, 2018). We next checked if CENP-E removal from the partially aligned KTs under CKAP5-depleted condition depends on dynein-mediated stripping by co-depleting dynein adaptor, Spindly along with CKAP5 in HeLa Kyoto cells. CENP-E levels were rescued significantly at the partially aligned KTs in the co-depleted cells, suggesting that CENP-E destabilization in CKAP5-depleted condition is due to CENP-E stripping by dynein (Fig. EV1H–J).

We next sought to identify the region of CKAP5 that could be involved in CENP-E regulation by rescue experiments using various CKAP5 deletion constructs (Fig. 1H). While the full length and CKAP5 consisting of TOG4, TOG5 and the C-terminus (CKAP5^853-2032) could rescue CENP-E localization at KTs, when co-expressed with CKAP5-shRNA in HeLa Kyoto cells, the C-terminal only domain (CKAP5^1429-2032) failed to do so (Figs. 1I,J and EV1K). Also, CKAP5^853-2032 expressing cells showed a reduced number of unaligned polar chromosomes (Fig. EV1L,M). The N-terminal only region of CKAP5 consisting of all the TOG domains (CKAP5^1-1428) did not show any specific localization to KTs or spindles (Fig. EV1N). Exogenous expression of CKAP5-GFP full length in endogenous CKAP5-depleted prometaphase cells also showed significant rescue of CENP-E localization as compared to only CKAP5-depleted prometaphase cells (Fig. EV2A,B). Taken together, the results indicate that MT-associated CKAP5 plays a critical role in stabilizing CENP-E at the aligned KTs and further, the region containing TOG4 and TOG5 is involved in this function.

Since CENP-E localization at attached KTs is known to be facilitated by its direct interaction with the Bub complex protein BubR1 (Chan et al, 1998; Huang et al, 2019; Legal et al, 2020; Mao et al, 2003), we next sought to examine CENP-E-BubR1 interaction in the CKAP5-depleted cells. Pulldown of MycGFP-CENP-E (by GFP Trap) from mitotic synchronized HEK 293T cells depleted of endogenous CKAP5 and expressed with MycGFP-CENP-E showed substantially reduced interaction of endogenous BubR1 with MycGFP-CENP-E as compared to the same in control cells (Fig. 2A,B). Depletion of Bub complex proteins (Bub1 and BubR1) has been shown to delocalize CENP-E from the KTs (Johnson et al, 2004; Legal et al, 2020). Therefore, to rule out the possibility that CKAP5 depletion-induced loss of CENP-E from KTs is due to defects in KT localization of BubR1/Bub complex, we checked BubR1 localization in the CKAP5-depleted cells. BubR1 levels at the aligned KTs were either not significantly affected or in some cases slightly increased but not reduced in the CKAP5 knockdown cells (Fig. 2C,D). Similar results were obtained for Bub1 localization as well (Fig. EV2C,D). Furthermore, since CKAP5^853-2032-GFP could rescue CENP-E localization at the aligned KTs under endogenous CKAP5-depleted condition, we next checked the effect of CKAP5^853-2032-GFP expression on CENP-E-BubR1 interaction in HEK 293T cells depleted of endogenous CKAP5. CENP-E-BubR1 interaction was substantially rescued upon expression of CKAP5^853-2032-GFP in these cells, supporting a direct role of CKAP5 in facilitating CENP-E-BubR1 interaction and thereby stabilizing CENP-E at the KTs (Fig. 2E,F).

Since MTs are essential for CKAP5 depletion-induced CENP-E abrogation from KTs, we characterized specific MT-associated defects, if any, in the CKAP5-depleted cells. To visualize specifically the KT-attached MTs, the non-KT MTs were depolymerized under cold condition (Wu et al, 2019). Immunofluorescence imaging of CKAP5-depleted cells showed MTs attached to the sister KTs with unequal lengths and reduced inter-KT distance between sister KT pairs of the aligned chromosomes (Fig. 3A–C). It was also revealed that CKAP5 depletion resulted in an increased incidence of the number of cells having at least one merotelically attached KT (Fig. 3D,E), and syntelic and monotelic attachments of the pole-proximal chromosomes (Fig. EV2E,F). Merotelic attachments, if not corrected, generate lagging chromosomes during anaphase (Sacristan et al, 2018). Presence of merotelic attachments was therefore, further confirmed by checking for the incidence of lagging chromosomes during anaphase. For this, CKAP5-depleted cells and control cells were treated with 250 nM Reversine (Mps1 inhibitor) for inducing anaphase (Sacristan et al, 2018) and it was observed that ~73% cells showed lagging chromosomes in anaphase, when CKAP5 was depleted suggesting the prevalence of merotelic attachments (Fig. EV2G,H). Interestingly, a significant number of KT-attached MTs (kMT) of CKAP5-depleted cells showed increased MT density (thicker MTs), suggesting enhanced stabilization of the kMTs (Inset, Fig. 3A,F). However, the increased MT stabilization phenotype was rescued by the expression of CKAP5^853-2032-GFP which encompasses the basic linker region which is important for its MT lattice binding (Figs. 3G and EV2I). Next, we sought to understand how increased MT stability can induce CENP-E removal from the aligned KTs. For this, MT hyperstabilization and low inter-KT tension phenotypes observed in CKAP5-depleted cells were mimicked in control cells by treating them with 10 µM paclitaxel for 30 min. Interestingly, CENP-E level at aligned KTs under such condition was intact as compared to DMSO-treated cells (Fig. 3H,I), suggesting that MT hyperstabilization alone is not solely responsible for CENP-E removal. Previous reports suggested that paclitaxel interferes with MT tracking ability of dynein (Forer et al, 2018; Gurden et al, 2018), which could also result in CENP-E stabilization at KTs. The role of MT hyperstability in CENP-E removal was further tested by partially destabilizing MTs in CKAP5-depleted cells using 3.3 µM nocodazole for 10 min (Fig. EV3A). A significant rescue in CENP-E levels at the partially aligned KTs was observed under partial nocodazole condition (Fig. 3J,K). This suggests that MT hyperstabilization in the absence of CKAP5 has indeed a significant role in CENP-E destabilization. Recent evidence suggests that point mutations (K1142, K1143) at the basic linker region of CKAP5 induce increased stability of microtubules attached to kinetochores of polar chromosomes similar to that observed in CKAP5-depleted cells (Herman et al, 2020). In order to test the role of this mutation on CENP-E levels, HeLa Kyoto cells were expressed with the basic linker region mutant of CKAP5, CKAP5^853-2032-KK/AA-GFP under endogenous CKAP5-depleted condition and the KT localization of CENP-E was assessed. The levels of CENP-E at the aligned KTs were significantly reduced in the KK/AA mutant-expressed condition as compared to CKAP5^853-2032-WT-GFP expression (Fig. 3L,M), but the extend of reduction in CENP-E levels at the KTs of cells expressing CKAP5^853-2032-KK/AA mutant was significantly less as compared to the near complete removal of CENP-E in CKAP5-depleted condition (Fig. 1A–C). Taken together, the results indicate that CKAP5 mediates CENP-E stabilization at KTs by regulating the spindle microtubule stability.

Previous studies showed that correct KT–MT end-on attachment stimulates PP1 activity at KTs leading to attachment stabilization (Emanuele et al, 2008; Kim et al, 2010; Liu et al, 2010; Nijenhuis et al, 2014). Moreover, PP1 also dephosphorylates CENP-E and supposedly induces its removal from KTs (Kim et al, 2010). Therefore, we checked how PP1 activity at the KTs is regulated in CKAP5-depleted cells. Immunofluorescent staining of PP1 in CKAP5-depleted cells showed its increased localization at the partially aligned metaphase KTs and also at prometaphase KTs (Fig. 4A–D). Consistent with enhanced PP1 recruitment, the level of phosphorylation of the outer KT protein, such as Hec1 showed a concomitant decrease at the KTs in prometaphase cells (Fig. EV3B), suggesting that PP1 activity at the KTs is increased in these cells. Interestingly, expression of CKAP5^853-2032-GFP which could rescue the enhanced MT stabilization phenotype, could also rescue the levels of PP1 at prometaphase KTs in endogenous CKAP5-depleted cells (Fig. EV3D,E). We then checked whether CENP-E delocalization from KTs in CKAP5-depleted cells could be attributed to increased PP1 activity at the KTs. Treatment of CKAP5-depleted cells with okadaic acid (0.25 µM) which inhibits PP1 activity resulted in nearly complete rescue of CENP-E at the KTs (Fig. 4E,F). Under similar conditions, a lower concentration of okadaic acid (0.12 µM) couldn’t rescue CENP-E localization, suggesting that PP1 but not PP2A is responsible for the CENP-E delocalization phenotype in CKAP5-depleted cells (Fig. EV3F). Furthermore, CENP-E-BubR1 interaction was rescued significantly in the CKAP5-depleted condition, when treated with okadaic acid in HEK 293T cells (Fig. 4G,H) substantiating the role of increased PP1 activity for the KT removal of CENP-E under CKAP5 depletion. Initial Aurora A/B-mediated phosphorylation followed by PP1-mediated dephosphorylation at the Aurora B-targeting site, Thr 422 of CENP-E facilitates metaphase chromosome alignment (Kim et al, 2010). To test the possibility that KT removal of CENP-E in CKAP5-depleted cells is associated with its PP1-mediated dephosphorylation at Thr 422, we checked the KT localization of a phospho-mimetic mutant of CENP-E specific to Thr 422 site (MycGFP-CENP-E-T422E). The KT intensity of CENP-E T422E mutant was checked and compared with that of CENP-E wildtype and a phospho-deficient T422A mutant under endogenous CENP-E depletion by siRNA in doxycycline-induced CKAP5 knockout HeLa cells (Fig. EV3G). Unlike GFP-CENP-E-WT, which was nearly completely lost from the KTs (Fig. 4I), GFP-CENP-E-T422E mutant was significantly stabilized at the partially aligned KTs under doxycycline-induced CKAP5 depletion condition (Fig. 4J,K). As reported in a recent study, the phospho-deficient mutant, CENP-E-T422A, could not localize to the partially aligned KTs under endogenous CENP-E-depleted condition (Eibes et al, 2023) irrespective of the presence or absence of CKAP5 (EV3 H). Since PP1 targeting to CENP-E and presumably PP1-mediated CENP-E dephosphorylation are critical for the removal of CENP-E from the KTs (Eibes et al, 2023; Kim et al, 2010), our results support the model that CKAP5 depletion induced MT over-stabilization, and the associated increased KT recruitment of PP1 could control the level of phosphorylated-CENP-E and its stabilization at the KTs.

We developed a basic mechanistic model of kMT–KT (kinetochore-attached microtubule–kinetochore) attachment error correction mechanism to comprehend the role of CKAP5 in establishing the correct end-on attachments. We only considered primary components of the mitotic spindle, including the astral MTs, kMTs, sister KTs, and centrosomes (Fig. 5A). The effect of CENP-E and other fibrous corona proteins, appear in the model as spring-like attachments between kMT tip and the KT (Fig. 5B,C) analogous to earlier studies (Powers et al, 2009; Sutradhar et al, 2015; Thomas et al, 2016). Depletion of CKAP5 leads to the loss of CENP-E at KTs in the experiments, and this is replicated in the model by a reduced strength of the kMT–KT spring-like attachments (Thomas et al, 2016). We assume the rate of the detachment of kMT–KT bond is proportional to the force applied on the kMT–KT springs for syntelic and merotelic attachments (Fig. 5D). We consider the force-dependent turnover of kMT–KT attachments based on earlier models of cargo hauling according to which an increased load force on molecular motors mediated by the cargoes enhances their detachment rate with the underlying MT track (Klumpp and Lipowsky, 2005; Kunwar et al, 2011). In addition, the detachment rate of a merotelic attachment is assumed to be proportional to the angle formed by the kMT–KT attachment with the KT–KT axis, considering that larger angles destabilize the kMT–KT attachment (Fig. 5F) (Renda et al, 2022), and the ratio of the number of MTs attached from the same pole to the number of MTs from the opposite pole (Lampson and Grishchuk, 2017). Each of these factors is thought to promote chromosomal bi-orientation by stabilizing kMTs that are more closely aligned to the KT–KT axis (Fig. 5E,F). Because of inter-KT tension stabilizing a pair of amphitelic sister KTs, no kMT detachment occurs in this case and new microtubule attachment to the KT pair is ignored. More details on the computational models are included in “Methods”, and the model parameters are noted in Table EV1.

The simulation was carried out for ~ 20 min, which is sufficiently long for the KTs in a control cell to achieve bi-orientation and congress to the metaphase plate (Paul et al, 2009). Our results revealed that significant erroneous attachments persisted in the CKAP5-depleted condition, while mostly the amphitelic attachments were formed in the control cells (Fig. 5G). CKAP5 depletion leads to a significant percentage of chromosomes with syntelic attachments (~33%) remaining after 20 min of simulation time with respect to a small percentage of merotelic attachments (~0.56%). Merotelic attachments were not observed in the control cells; however, a very small percentage of syntelic (~0.07%) and monotelic or unattached (~0.4%) chromosomes remained (Fig. 5G). The end-on attached sister KTs in the control cells were found near the spindle mid-zone that oscillates around it (Fig. 5H,I; Movie EV1). In the CKAP5-depleted cells, while the amphitelic KTs congressed mostly near the spindle center, syntelic attachments remained close to the poles and often formed behind the centrosomes (Fig. 5J,K; Movie EV2). Contrary to the control scenario, we found sister KTs with amphitelic attachments had a larger spread of the distribution away from the spindle center in the CKAP5-depleted cells (Fig. 5H–K). Interestingly, our simulation revealed that a relatively larger number of kMTs are associated with KTs in the CKAP5-depleted condition compared to the control scenario (Fig. 5L). We further noticed that varying the severity of CKAP5 depletion by reducing the force strengths of kMT–KT attachments yielded reduced amphitelic attachments and increased syntelic attachments (Fig. 5M). Overall, our computational study qualitatively captures the defects in chromosome distribution patterns and the occurrence of the erroneous attachments under the CKAP5-depleted condition as displayed in the experimental data (Figs. 3 and EV2E,F).

Having established the importance of CKAP5 activity for recruiting CENP-E to KTs, and the functional consequences of its inhibition, we sought to explore whether CENP-E expression and essentiality were associated with chromosome instability and with those of CENP-E in human cancer. For this, we analyzed mRNA and protein expression levels of 1408 and 339 human cancer cell lines, respectively, as well as CRISPR screens for 1078 human cancer cell lines, and associated them with the cells’ ploidy status, using data obtained from the Dependency Map database (Cohen-Sharir et al, 2021; Dempster et al, 2021; Nusinow et al, 2020; Tsherniak et al, 2017). The mRNA and protein expression levels of CKAP5 were similar between chromosomally stable, near-diploid cell lines and chromosomally unstable, highly aneuploid cell lines (Fig. 6A,B). However, the highly aneuploid cells were much more sensitive to the CRISPR-mediated knockout of CKAP5 (Fig. 6C,D), confirming a functional link between CKAP5 and chromosomal instability. Similar to other kinetochore-related genes (Fig. EV4), both expression and dependency levels were significantly correlated between CKAP5 and CENP-E (Fig. 6E–G), providing further support for their functional interaction. Taken together, these results suggest that the functional link between CKAP5 and CENP-E is ubiquitous in cancer, and that highly aneuploid, chromosomally unstable cancers might be more sensitive to CKAP5 depletion.

Dulbecco’s modified Eagle’s medium (DMEM) and fetal bovine serum (FBS) were purchased from ThermoFisher Scientific, Massachusetts, USA. Penicillin-streptomycin solution was obtained from Himedia, Mumbai, India. Thymidine, Calbiochem Insolution Okadaic acid, Eg5 kinesin inhibitor Dimethylenastron (DMA), EGTA, PIPES, HEPES, MG132, and DAPI were purchased from Sigma-Aldrich, St Louis, MO. In total, 16% PFA was obtained from Electron Microscopy Sciences, PA, USA. Transfection reagents Lipofectamine 3000 and RNAiMax and ProLong Gold antifade mountant were obtained from ThermoFisher Scientific, MA, USA. Transfection reagent PEI (Polyethyleneimine) was obtained from Polysciences, PA, USA. MISSION® esiRNA targeting FLUC (Cat # EHUFLUC) against Firefly Luciferase and MISSION® esiRNA against CKAP5 (Cat # EHU078221) were obtained from Sigma-Aldrich. GFP trap agarose beads (Cat # gta-20) for GFP pulldown were purchased from Chromotek, Munich, Germany. 3’ UTR siRNA against CENP-E (5′-CCACUAGAGUUGAAAGAUA-3′), custom siRNA against Spindly (5′-GAAAGGGUCUCAAACUGAA-3′) and CKAP5 (5′-CAUGCUCCACAGCAAACUCUC-3′) were purchased from GE Health care Dharmacon Inc. Rabbit polyclonal antibodies against CKAP5 (Cat # ab-86073), CENP-T (Cat # ab-220280), Knl1 (Cat # ab222055) were obtained from Abcam, Cambridge, MA, USA. Mouse monoclonal antibody against CENP-E (Cat # ab5093) and Bub1 (Cat # ab54893) were purchased from Abcam (MA, USA). Mouse monoclonal antibody against CENP-A (Cat # GTX13939) was obtained from GeneTex, CA, USA. Mouse monoclonal antibody against BubR1 (Cat # 612503) was obtained from BD Biosciences (California, USA). Mouse monoclonal antibodies for CENP-E (Cat # sc-376685), PP1 (Cat # sc-7482), GFP (Cat # sc-9996), C-Myc (Cat # Sc-40) were obtained from Santa Cruz Biotechnologies, CA, USA. Mouse monoclonal antibody against GFP (Cat # 632381) was purchased from Takara Bio Clontech, USA. Rabbit polyclonal antibody against pHecS55 (Cat # GTX70017) was obtained from GeneTex, CA, USA. Anti-centromere protein antibody, ACA (Cat # 15-234) was obtained from Antibodies Incorporated, CA, USA. Mouse monoclonal antibody against α-Tubulin (Cat # T6199) was obtained from Sigma-Aldrich, St Louis, MO. All HRP-conjugated secondary antibodies, Alexa 488, Alexa 647, and cy5-conjugated secondary antibodies were purchased from Jackson ImmunoResearch, PA, USA. Alexa 568-conjugated anti-rabbit secondary antibody was obtained from ThermoFisher Scientific, MA, USA. The dilutions of the primary antibodies used were: CENP-E (IF-1:200; WB-1:200), CENP-A (1:400), CENP-T (1:200), Knl1 (1:200), PP1 (IF-1:100; WB-1:500), ACA (1:700), α-Tubulin (IF-1:700; WB-1:2000), CKAP5 (IF-1:200; WB-1:2000), BubR1 (IF-1:200; WB-1:500), Bub1 (1:300), pHecS55 (1:200), GFP (WB-1:2000), and GST (WB-1:2000). Secondary antibodies were used at dilutions 1:1000 (IF) and 1:2000 (WB).

All truncated and mutated CKAP5 (ch-TOG) constructs (CKAP5^1-1428, CKAP5^853-2032, CKAP5 ^853-2032-KK/AA, CKAP5^1429-2032) were amplified from pBrain-ch-TOG KDP-GFP-shch-TOG (a kind gift from Stephen Royle, Addgene plasmid # 69113) and cloned to the same vector. The plasmid possesses shRNA against endogenous CKAP5 to which the CKAP5 cDNA is resistant. pCDNA3 eGFP empty vector (a gift from Doug Golenbock, Addgene plasmid # 13031) was used as empty GFP control. pCDNA5-FRT-TO-Myc-Lap-hCENP-E-WT, T422E and T422A mutants were obtained from Don W. Cleveland Lab, Department of Cellular and Molecular Medicine, University of California San Diego.

HeLa Kyoto cells were obtained from Sachin Kotak, Indian Institute of Science (IISc), Bangalore (originally provided by Daniel Gerlich, IMBA, Vienna). HEK 293T and U2OS cells were originally obtained from ATCC. CKAP5 conditional knockout HeLa cell line (cKMKO C5.2) was a kind gift from Ian M Cheeseman (Whitehead Institute, MIT, USA). The cell lines were maintained in DMEM high glucose media supplemented with 10% FBS at 37 °C in a humidified atmosphere with 5% CO₂. For conditional knockout cell lines, DMEM high glucose media supplemented with 10% Tet-free FBS was used. Cells were incubated in 2 µM doxycycline for 72 h to get effective CKAP5 knockout. For transient transfection of all plasmids to HeLa Kyoto cells and 293T cells, Lipofectamine 3000 or PEI (Polyethyleneimine) reagents were used. For mock and CKAP5 depletion, Luciferase esiRNA and CKAP5 esiRNA were transfected using Lipofectamine RNAiMax reagent. To ensure that the cells are in their first mitotic division after gene knockdown, single thymidine-arrested cells were released and transfected with esiRNA after 6 h of release, followed by a second thymidine arrest for 24 h and fixed at 9th hour after release. For pull-down experiments, cells were synchronized at prometaphase by double thymidine block followed by thymidine release to DMEM containing DMA (Dimethylenastron, Eg5 kinesin inhibitor) for 12 h. For metaphase synchronization, double thymidine blocked cells were either released into DMEM containing DMA for 12 h followed by DMA washout with fresh DMEM containing MG132 for 2 h or simply thymidine release followed by MG132 treatment at 8th hour after release for 2 h (HEK 293T cells). Drugs were used at the following final concentrations: Thymidine, 2 mM; DMA, 5 µM; MG132, 25 µM; Okadaic acid, 0.25 µM/0.12 µM; Reversine, 250 nM; Nocodazole, 3.33 µM; Doxycycline, 2 µM, Paclitaxel, 10 µM.

Cells were pre-extracted for 5 min using 0.1% Triton X 100 in PHEM buffer (60 mM PIPES (pH 6.8), 25 mM HEPES (pH 6.9), 10 mM EGTA and 2 mM MgCl₂) before fixation in some cases. Cells were then fixed using 4% PFA in PHEM for 15 min at room temperature, followed by blocking with 1% PBSAT (1× PBS, 1% BSA, 0.5% Triton x 100). For the cold stability assay, cells were incubated in ice for 5 min and fixed in ice-cold methanol for 20 min. For all CENP-E staining, cells were fixed in ice-cold methanol. Primary antibody incubations were done for 3 h or overnight at room temperature in a dark humidified chamber. Secondary antibody incubation was done for 1 h at room temperature. DNA was stained using DAPI, and the cells were mounted with ProLong Gold (Invitrogen). Images were acquired in Zeiss LSM 880 confocal microscope, Leica SP5 upright confocal microscope, and Olympus FluoView3000 with a 63× oil-immersion objective. For better visibility, image contrast was adjusted manually by changing the LUT values. Raw confocal images (no contrast adjustments) were quantified for individual KT intensity across z or from maximum projection images using ZEN blue software or ImageJ Fiji. Background intensity was measured from multiple regions within the cell away from the KTs, and average background intensities were subtracted from the KT intensities of proteins of interest. All mean KT intensities of required proteins were normalized with that of a KT marker quantified similarly. For MT plus end intensity, a line with 5 µm width was drawn across kMTs at the plus end, and the maximum intensity value was obtained from plot profile in Fiji (Liu et al, 2010). All graphs were plotted and analyzed for statistical significance using GraphPad Prism 9.

Fluorescence microscopic images were captured by a structured illumination method using an inverted Zeiss ELYRA PS1 microscope. Four lasers have been used for excitation: 405 nm (50 mW), 488 nm (200 mW), 561 nm (200 mW), and 642 nm (150 mW). Imaging was performed using a Zeiss oil-immersion objective (alpha Plan-apochromat DIC 63×/1.40 Oil DIC M27; numerical aperture (NA) 1.40 oil (Fig. 1B) and alpha Plan-apochromat DIC 100×/1.46 Oil DIC M27; numerical aperture (NA) 1.46 oil (Figs. 1E and 3D). Fluorescence light was spectrally filtered with emission filters (MBS– 405 + EF BP 420–495/LP 750 for laser line 405, MBS– 488 + EF BP 495–570/LP 750 for laser line 488, MBS– 561 + EF BP 570–650/LP 750 for laser line 561 and MBS–642 + EF LP 655 for laser line 642) and imaged using a PCO edge sCMOS camera. The respective bright-field images were taken using LED light sources. Acquired images have been processed using ZEN 2.0 structured illumination software.

Prometaphase or metaphase-arrested cells (HeLa Kyoto and HEK 293T) were lysed in mammalian cell lysis buffer (20 mM Tris, 50 mM NaCl, 1 mM EGTA, 1% Triton X 100 with 1× protease inhibitor cocktail). For all pull-down experiments, agarose beads were equilibrated in lysis buffer and blocked using 2% BSA in lysis buffer. For pulling down MycGFP-CENP-E and CKAP5-GFP the cell lysates were incubated with GFP nanobody-coated beads (GFP TRAP agarose) for 4 h at 4 °C. For CENP-E-BubR1 interaction rescue experiment, MycGFP-CENP-E was pulldown by incubating cell lysates with C-Myc antibody for 24 h at 4 °C followed by incubation with Protein A/G agarose for 2 h at 4 °C. The beads were then washed and processed for western blotting. The immunoblots were visualized in the ChemiDoc XRS+ Imaging System (Bio-Rad, Hercules, CA, USA). All protein band intensities were quantified using volume analysis tool in Image Lab software.

Data are presented as mean +/− SEM. The normally distributed data were analyzed with modified Student’s (Welch) t test at the 99% confidence level. For skewed datasets, Mann–Whitney test was used for analysis. Wherever applicable, one-way ANOVA followed by Tukey’s multiple comparison tests were performed. The data were plotted and analyzed using GraphPad Prism 9 software. The figures were organized using Adobe Photoshop and Adobe Illustrator.

The effects of CKAP5 depletion in HeLa cells were investigated in situ by a three-dimensional mechanistic model. Our model consisted of a spherical cell of radius R_cell with two static centrosomes placed d_cent distance away from the cell center on opposite sides. Initially, N_CH chromosomes (CHs) are placed randomly throughout the cell. Dynamically unstable microtubules (MTs) grow and shrink throughout the cell with their minus end attached to the centrosomes and search for the sister kinetochores (KTs) by their dynamic plus ends (Kirschner and Mitchison, 1986; Paul et al, 2009). The chromosome arms experience a polar ejection force away from the poles due to the astral microtubule tips hitting on them (Antonio et al, 2000; Brouhard and Hunt, 2005; Ke et al, 2009; Levesque and Compton, 2001; Rieder and Salmon, 1994). The kinetochores are captured by the astral microtubules with a certain probability once the MT hits the KT. The depolymerization force of the MT on the KT generates poleward tension and is assumed to be a spring force (Powers et al, 2009; Wei et al, 2007). The sister KTs are assumed to be attached by cohesion, which is modeled as a spring force. The formation of a proper bipolar spindle and the congregation of the chromosomes at the metaphase plate requires all the KT–MT attachments to be amphitelic (sister KTs connected to opposite spindle poles; see Fig. 5A). Other types of wrong attachments also occur during the formation of the mitotic spindle, namely syntelic attachments where both sister KTs are attached to MTs from the same pole (Fig. 5A,B) and merotelic attachments where one KT is attached to MTs from opposite poles (Fig. 5A,C). Correction of the merotelic and syntelic attachments is hampered by CKAP5 depletion. CKAP5 depletion is known to prematurely stabilize MT-KT attachments (Herman et al, 2020). We assume that the detachment rate of a kinetochore-attached MT (kMT) is proportional to the load force on the MT and that depletion of CKAP5 leads to the reduction of kMT load forces. All these components are included in the model in the following manner:

Human cancer cell line aneuploidy scores (AS), mRNA and protein expression datasets, and CRISPR-Cas-9-based gene dependency scores (determined by the “Chronos” algorithm) were obtained from DepMap release 22Q4 (https://depmap.org/portal/) (Cohen-Sharir et al, 2021; Dempster et al, 2021; Nusinow et al, 2020; Data ref: Tsherniak et al, 2017). Cell lines were split into two groups of semi-diploid and highly aneuploid cell lines, correlating to the top and bottom quartiles of aneuploidy scores. A two-sided Student’s t test was used to compare gene expression and CRISPR dependency values between the groups. Statistical analyses were performed in GraphPad Prism 9.1. To remove the effects of proliferation, we fitted linear regression models to the correlations and used the R “partialize” function to compute the residuals without the effect of MKI67 mRNA expression, which were then plotted.