Fig 2: KIF4 and PRC1 repress chromosome congression when Cdk1 is partially active.(A) Experimental setup of a nocodazole washout assay. Thymidine blocked cells are released, nocodazole-arrested, and washed free from nocodazole to allow spindle assembly in presence of the proteasome inhibitor MG132, together with RO-3306. (B) Representative immunofluorescence images of mitotic cells, before and after nocodazole washout. Examples of different alignment defects are shown. Scale bar, 10 μm. (C) PRC1 and KIF4 prevent chromosome alignment when Cdk1 activity is compromised. Bar graph shows a quantification of the experimental setup shown in (A). (Mean ± s.e.m, n = 3, at least 150 cells/experiment). (D) Western blot analysis of the experiment in (C) shows the knockdown efficiency of the shRNAs targeting KIF4 or PRC1. (E) PRC1 present at microtubule overlaps partly localizes to astral microtubules after suppression of Cdk1 activity. U2OS cells were treated with or without 3 μM RO-3306 for 16 hours, fixed, and processed for immunofluorescence. Insets demonstrate PRC1 localization to astral microtubules in the vicinity of unaligned chromosomes when Cdk1 is inhibited. Scale bar, 10 μm. (F) PRC1 localization to astral microtubules is dependent on the presence of KIF4. U2OS WT or KIF4 KO cells were treated and processed as in (E). Prometaphase cells were scored for the presence of PRC1 on astral microtubules (MTs). Scale bar, 10 μm. (Mean ± s.e.m, n = 3, at least 30 cells/experiment; ****p < 0.0001, Student’s t test). (G) Schematic overview of the progression from prometaphase to metaphase with normal or compromised cyclin B-Cdk1 activity. Maximal Cdk1 activation directs chromosome alignment and the establishment of stable k-MT attachments in metaphase. Partial reduction in Cdk1 activity by RO-3306 treatment results in a prometaphase arrest that is dependent on KIF4 and PRC1. Due to a weakened spindle checkpoint, cells will exit from mitosis with lagging chromosomes leading to the formation of micronuclei. Western blots in panel (D) have been cropped and full-length gels can be viewed in Supplementary Fig. 6.

Fig 3: A shRNA barcode screen identifies PRC1 as critical mediator of the anti-proliferative effects of RO-3306.(A) Schematic outline of the RO-3306 barcode screen in U2OS cells. (B) Reconstitution of the RNAi-resistant human PRC1-1 cDNA in PRC1 knockdown cells. The pRS vector (shCTRL) was used as a control. (C) Western blot analysis of protein lysates corresponding to the cell lines shown in (B). shPRC1 Lib#1 targets only endogenous PRC1, while shPRC1 Lib#2 is able to target both endogenous and Venus-NT-PRC1 #1. (D) Western blot analysis of U2OS cells stably expressing individual Venus-PRC1 isoforms after transfection of siRNA duplexes specifically targeting the indicated splice variants. CTRL refers to cells lacking ectopic expression of Venus-PRC1. (E) Depletion of PRC1-1 rescues cell proliferation when Cdk1 activity is compromised. The functional phenotypes of the individual shRNAs targeting the PRC1 variants are indicated by the colony formation assay. The pRS vector (shCTRL) was used as a control. The knockdown efficiency of each individual shRNA was measured by examining the mRNA level of the PRC1-1 and PRC1-2 target genes by qRT-PCR. (Mean ± s.d. n = 3). (F) KIF4 is a PRC1 binding partner. Immunoprecipitations (IPs) with GFP-Trap_A beads were performed on extracts of U2OS WT (CTRL) or U2OS cells stably expressing Venus-PRC1-1 (PRC1-1). Aliquots of the IPs were analysed by SDS-PAGE. (G) Compromised Cdk1 activity in mitosis results in premature binding of PRC1 to KIF4. U2OS cells were synchronised in mitosis using thymidine and nocodazole. Cells arrested in mitosis were obtained by mitotic shake-off, released from the nocodazole and harvested after 60 minutes (G1) or, alternatively, treated with 5 μM MG132 with or without 3 μM RO-3306 and harvested after 30 minutes (M). Cell extracts were subjected to immunoprecipitation using a rabbit anti-PRC1. Aliquots of the immunoprecipitates were analysed by Western blotting. The signal intensity of the KIF4 band in the PRC1 IP was quantified and normalized to 1 for the G1 sample. *p < 0.05, Student’s t test. Western blots in panels (C,D,F,G) have been cropped and full-length gels can be viewed in Supplementary Fig. 6.

Fig 4: Model for specialized, short-lived reinforcement near chromosomes in the mammalian spindle.(A) K-fiber reinforcement in space: Microneedle (black circle) manipulation of the mammalian spindle reveals that k-fibers (light grey) are weakly coupled to their neighbors (thin dashed vertical line), strongly coupled to their sisters (thick horizontal line), and pivot around the pole (black arrow) but not around chromosomes. K-fibers are locally reinforced (dashed box, 3 μm) near chromosomes (spring) through specialized, non-uniform mechanisms requiring the microtubule crosslinker PRC1 (light blue squares). Other crosslinkers in the spindle are shown in dark grey. (B) K-fiber reinforcement in time: Local reinforcement near chromosomes is preserved over seconds (with a lifetime of 20 s, black star) yet remodels over minutes as molecules turn over in the spindle center. This allows the local architecture in the spindle center to persist under transient force fluctuations, and yet respond to sustained forces. Such short-lived reinforcement could help protect chromosome-to-spindle connections while allowing them to remodel (green arrow) as mitosis progresses. For simplicity, we only depict PRC1 turning over as time evolves (from light blue to dark blue PRC1 molecules), though microtubules and other crosslinkers also turn over.

Fig 5: The microtubule crosslinker PRC1 mediates the specialized and short-lived kinetochore-fiber reinforcement near chromosomes.See also Figure 4—figure supplements 1–2 and Figure 4—video 1. (A) Immunofluorescence images of a representative PtK2 mock RNAi (control) metaphase spindle showing where PRC1 is localized in the spindle (tubulin, yellow; PRC1, magenta). White box (bottom panel) shows the region in which intensity (B) was quantified. Scale bar = 5 μm. (B) Fluorescence intensity ratio of PRC1 to tubulin along the pole-pole axis (n = 6 cells), showing PRC1 localization in the spindle center. Plot shows mean ± SEM. (C) Immunofluorescence images of a representative PtK2 PRC1 RNAi metaphase spindle (tubulin, yellow; PRC1, magenta), showing PRC1 depletion. Scale bar = 5 μm. (D) Top: Timelapse images of a representative PtK2 metaphase PRC1 RNAi spindle (GFP-tubulin, grey) during a 60 s manipulation, showing microneedle position (white circle) and traced manipulated k-fiber (white). Scale bar = 5 μm. Time in min:sec. Bottom: Curvature mapped along traced k-fiber for each point in the top panel (blue, negative curvature; red, positive curvature), showing the absence of negative curvature near chromosomes without PRC1. (E) Local curvature of deformed k-fibers for normalized positions along the k-fiber (n = 12 k-fibers in 12 cells). Most k-fibers exhibit no negative curvature (grey) and one shows negative curvature similar to WT k-fibers (orange). Scatter plot of microneedle positions shown above (inset). (F) Left: Distribution of microneedle positions along the k-fiber in WT (n = 17 cells) and PRC1 RNAi (n = 12 cells) spindles, after datasets are minimally down-sampled to maximize microneedle position overlap between them. There is no significant difference in microneedle position in the two conditions (p=0.38, Mann-Whitney U test). Plot shows mean ± SEM. Right: Percentage of deformed k-fiber profiles showing negative curvature near chromosomes in WT and PRC1 RNAi manipulated spindles, showing loss of negative curvature without PRC1. (G) Schematic of the three measurements made in (H,I,J): Inter-kinetochore distance between the manipulated k-fiber and its sister, angle between the sister k-fiber plus-end (opposite the manipulated k-fiber) and the pole-pole axis, and the angle between sister k-fiber plus-end regions. (H) Change in inter-kinetochore distance in WT unmanipulated (control, n = 13 kinetochore pairs from 6 cells), WT manipulated (n = 8 kinetochore pairs from 8 cells) and PRC1 RNAi manipulated (n = 8 kinetochore pairs from 8 cells) spindles, measured over 60 s. Inter-kinetochore distance after manipulation is significantly higher in spindles with PRC1 RNAi than WT (p=0.003, Mann-Whitney U test). Plot shows mean ± SEM. (I) Change in angle of sister k-fiber plus-end with respect to the pole-pole axis in WT unmanipulated (control, n = 12 k-fibers from 6 cells) and WT manipulated (n = 11 k-fibers from 11 cells) and PRC1 RNAi manipulated (n = 9 k-fibers from 9 cells) spindles, measured over 60 s. The sister k-fiber moves less (smaller angle) towards the pole-pole axis after manipulation in PRC1 RNAi spindles compared to WT (p=0.004, Mann-Whitney U test). Plot shows mean ± SEM. (J) Distribution of the angle between sister k-fiber plus-end regions at the end of manipulation in WT (n = 20 cells) and PRC1 RNAi (n = 10 cells) spindles, measured between the chromosome-proximal regions of the k-fibers.Figure 4—source data 1.This spreadsheet contains the fluorescence intensity ratio of PRC1 to tubulin along the pole-pole axis of spindles acquired by immunofluorescence (Figure 4B), the local curvature along k-fibers manipulated over 60 s in PRC1 RNAi spindles (Figure 4E), microneedle positions from 60 s manipulations in WT and PRC1 RNAi such that their positions along the k-fiber maximally overlap (Figure 4F), the change in inter-kinetochore distance (Figure 4H) and angle of sister k-fiber plus-end from the pole-pole axis (Figure 4I) in unmanipulated and manipulated spindles, and PRC1 RNAi manipulated spindles, and the angle between sister k-fiber plus-end regions in WT and PRC1 RNAi PtK2 spindles (Figure 4J).