Fig 2: Nek5 loss leads to retention of centrosome linker proteins and reduced MT nucleation in mitosis. (a) Staining of DNA, a-tubulin, and pericentrin in GL2- and Nek5-depleted cells at the mitotic stages indicated. (b) Histogram of cells in different stages of mitosis, as indicated, after 48 h of mock, GL2, or Nek5 depletion in U2OS cells. (c) Metaphase GL2- and Nek5-depleted cells stained for DNA, a-tubulin, and pericentrin. The yellow line indicates the plane of the metaphase plate; the white line indicates the pole-to-pole spindle axis. (d) Dot plot of the angle between the spindle axis and metaphase plate for mock-, GL2-, and Nek5-depleted cells. Each dot represents a single cell with data collected over two independent experiments, bars represents mean ± SD. (e and f) Staining of C-Nap1 or rootletin and ?-tubulin in prophase U2OS cells after mock, GL2, or Nek5 depletion. (g) Histogram of relative intensities of ?-tubulin, C-Nap1, rootletin, Nek2, and phospho-C-Nap1 (pC-Nap1) in mock-, GL2-, Nek5-, and Nek2-depleted prophase U2OS cells. Each protein was normalized to its intensity in mock-treated cells. (h) Staining of pC-Nap1 and Nek2 in prophase U2OS cells that were mock, GL2, or Nek5 depleted. The ratios of pC-Nap1/Nek2 intensities are indicated. (i) Staining of centrin, Cdk5Rap2, Cep192, and pericentrin in GL2- and Nek5-depleted prophase cells. (j) Histogram of relative intensities of centrin, Cdk5Rap2, Cep192, and pericentrin in prophase U2OS cells that were mock-, GL2-, or Nek5-depleted for 48 h. Each protein was normalized to its intensity in mock-treated cells. (k) Staining of pericentrin, EB1, and a-tubulin in GL2- and Nek5-depleted prophase cells treated with nocodazole for 5 h and 60 s washout to allow MT regrowth. (l) Histogram of the diameter of MT asters formed after 60 s of regrowth in mock-, GL2-, Nek5-, and Cep192-depleted cells. Bars: (a, c, e, f, h, and i, main images) 10 µm; (c) 5 µm; (e, f, h, i, and k, insets) 1 µm. Histograms show mean + SD (error bars); n = 3, 40 centrosomes (l) or 100 cells (b, g, and j) per experiment. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Fig 3: HSP70 accumulates at the spindle pole in a PLK1-dependent manner and colocalizes with PCNT, CEP215 and ?-tubulin. A Representative images of immunostained HSP70 (green) at the spindle pole in the CGL2 cells. Cells were treated with 20 µM PLK inhibitor III or BI2536 30 min before fixation. Spindle poles were identified by either a-tubulin (a-tub, red) or CEP152 (red) staining. B Relative HSP70 intensity at the spindle pole (SP) was measured. Box plots show the 10th, 25th, 50th, 75th, 90th percentiles, and numbers in parenthesis (n) indicate the number of spindle poles examined. *p < 0.05, significant decrease compared to untreated by Mann–Whitney rank sum test. C–F PLK1 (C green) and HSP70 (E green) accumulation at the spindle pole (specified by a-tubulin, red) in cells transduced with control vector (pFB-Neo), FLAG-PLK1-WT or D176N. D, F Relative intensity of PLK1 (D) and HSP70 (F) at the spindle pole (SP). Numbers of examined spindle poles (n) are indicated. Mann–Whitney rank sum test was performed. *p < 0.05 comparing WT to pFB-Neo; #p < 0.05 comparing D176N to WT. G Colocalization of HSP70 (green) with CEP215 (a), PCNT (b) and ?-tubulin (c) (red) was revealed by 3D-SIM imaging. The mitotic centrosomes in the red-boxed regions in the upper panels were magnified and are presented as wide-field (middle) and SIM (bottom) images. H Percentages of each indicated PCM component that were colocalized with HSP70 with or without HSP70 depletion are shown in the box plot with the number of spindle poles (n) indicated. *p < 0.05 by Mann–Whitney rank sum test comparing shHSP70 to CGL2

Fig 4: HSP70 is required for the interaction between PCNT and CEP215. A GSD images of PCNT (green) and CEP215 (red) in control (a) and HSP70-depleted (d) mitotic cells. Lines of different directions across the center are shown in merged images. (b, c, e and f) The intensities along each line within a central 1.8-µm region were plotted. At least 20 spindle poles were analyzed and five lines were drawn in each pole. Mean distribution curves are shown in green for PCNT and red for CEP215. The distance between two peak regions (mean ± SD) and the peak-center intensity ratio (median), which were determined as described, are presented and were tested by Student’s t test and Mann–Whitney rank sum test, respectively. *p < 0.05 comparing shHSP70 to pLKO.1. B PLA was performed as described in “Methods” to measure the interaction between PCNT and CEP215. Left, representative images of PLA signal generated using heat-inactivated (-) or normal activity (+) PLA reagents. Right, representative confocal images of cells depleted of HSP70 or treated with 40 µM PES 30 min before fixation are shown. C Relative PCNT and CEP215 PLA intensity at the spindle pole with the number of examined spindle poles (n) indicated. Mann–Whitney rank sum test was performed. *p < 0.05 comparing shHSP70 to pLKO.1 or PES to untreated. D Representative GSD images of colocalization between PCNT (green) and CEP215 (red) at the spindle pole of mitotic cells from control or HSP70-depleted cultures. E The level of PCNT and CEP215 colocalization was measured by Pearson’s correlation coefficient, and the coefficient ranking is shown with the number of examined spindle poles (n) indicated. Mann–Whitney rank sum test was performed. *p < 0.05 comparing shHSP70 to pLKO.1

Fig 5: HSP70 is required for the 3D assembly of CEP215, PCNT and ?-tubulin at the mitotic centrosome. A 2D projections of the averaged 3D-SIM images of each indicated centrosome marker in control interphase (a) and mitotic cells (c) and HSP70-depleted interphase (b) and mitotic (d) cells. Merged images show the relative distribution of each centrosome marker with respect to ?-tubulin. B, C Box plots show the relative diameter and volume of each marker in control and HSP70-depleted mitotic cells with the number of samples (n) indicated. Mann–Whitney rank sum test was performed. *p < 0.05 comparing shHSP70 to pLKO.1. D Surface-rendered images of the centrosome with or without HSP70. E (a) Representative wide-field image of a HSP70-depleted mitotic cell with multipolar spindle. Two poles with centriole marker (CEP164) are indicated with red boxes, and the other extra spindle poles without CEP164 are indicated by white boxes. (b) Averaged SIM images of the extra spindle poles