Fig 2: HSP70 chaperone activity attenuates cell sensitivity to Eg5 inhibitors.a–b HSP70 inhibitors enhanced the cytotoxicity of SB743921. CGL2 (a) and HeLaS3 (b) cells were incubated in medium containing SB743921 alone or in combination with HSP70 inhibitors, PES, VER or YM1, at the indicated concentrations for 72 h, after which cell viability was measured. Cell viability under HSP70 inhibitor alone was normalized to untreated cells. The mean ± SD from at least two independent experiments is shown. *p < 0.05 comparing HSP70 inhibitor-treated cells to cells without HSP70 inhibitor treatment by Student’s t test. c CGL2 cells harboring pFB-Neo empty vector or stably expressing FLAG-tagged HSP70-WT (FLAG-HSPA1A-WT) were incubated with the indicated concentration of SB743921 for 72 h before measurement of viability. Mean ± SD from at least three independent experiments is shown. *p < 0.05 comparing FLAG-HSPA1A-WT to pFB-Neo by two-way ANOVA. d Representative images of mitotic cells with the indicated types of spindle stained with CEP152 (green), a-tubulin (a-tub, red), and DAPI (blue). Cells were treated with 2.5 µM TriC or 5 µM BRD for 1 h and then fixed and immunostained to reveal mitotic spindles. e–k Percentages of cells with the indicated types of spindle are shown for the knock-down control (e, pLKO.1), HSP70-depleted (f, shHSP70), pFB-Neo-harboring (g), and FLAG-HSP70-WT (h, FLAG-HSPA1A-WT) and mutant-expressing (i–k) mitotic cells treated as indicated. At least 600 mitotic cells were counted for each experiment. The mean ± SD from at least three independent experiments is shown. *p < 0.05 comparing the respective spindle types and treatments between pLKO.1 and shHSP70 or between pFB-Neo and FLAG-HSPA1As by Student’s t test; N.S. no significance.

Fig 3: HSP70 may modulate Eg5 oligomeric assembly to control its localization in the spindle.a (left) Representative images show p-T926-Eg5 immunostaining (green) in untreated and PES- or VER-treated metaphase cells counterstained with a-tubulin (red) and DAPI (blue). (Right) Interquartile distribution of relative intensity of p-T926-Eg5 in the chromosome region was measured. Numbers in parentheses indicate the number of cells measured. b Representative GSD images show Eg5 immunostaining with corresponding wide-field images showing a-tubulin (green) and DAPI (blue) in untreated and PES- or VER-treated prometaphase and metaphase cells. c (Upper) The schematic illustration depicts the creation of a spindle mask by thresholding as described. (Lower) The Eg5 clusters were color-coded according to their sizes to reveal their distributions within spindle regions. d The percentages indicate Eg5 clusters within each indicated size range localizing to pole or inter-polar regions. The bar chart shows median ± 25th percentiles from at least 30 cells in four independent experiments. *p < 0.05 comparing inter-polar region to pole region by Student’s t test. e The percentages indicate Eg5 clusters within each indicated size range in the whole spindle of the untreated and PES- or VER-treated cells. The bar chart shows median ± 25th percentiles from at least 60 cells in four independent experiments. *p < 0.05 compared to untreated by Mann–Whitney Rank Sum test. f The percentages indicate the Eg5 clusters in the inter-polar regions of untreated and PES- or VER-treated prometaphase or metaphase cells. The bar chart shows median ± 25th percentiles from at least 20 cells in three independent experiments. *p < 0.05 comparing metaphase to prometaphase by Mann–Whitney Rank Sum test; N.S. no significance.

Fig 4: HSP70 inactivation enhances pole accumulation of Eg5 and prevents its inward movement.Representative images show the Eg5 (green) distribution within the spindle of untreated (a), PES-treated (b) and VER-treated (c) prometaphase and metaphase cells counterstained with a-tubulin (red) and DAPI (blue). Line-profile analysis of Eg5 intensity was performed as described. Eg5 intensity distributions were measured from at least 50 MT fibers in either prometaphase or metaphase cells. Data from cells imaged in two independent experiments were averaged to obtain the mean distribution curve and the 3 µm distance starting from spindle poles was plotted in (d) for untreated cells and in (e) and (f) for PES- and VER-treated cells, respectively. g Mean Eg5 distribution curves in untreated, PES- and VER-treated metaphase cells were plotted. Pairwise comparisons of the distribution curves were performed as described. *p < 0.05 compared to untreated. h–k Representative images show the Eg5 distribution within the spindle of the control (h. pLKO.1) and HSP70-depleted (i, shHSP70) prometaphase and metaphase cells. Eg5 intensity distributions were measured from at least 50 MT fibers in either prometaphase or metaphase cells. Data from cells imaged in two independent experiments were averaged to obtain the mean distribution curve and the 3 µm distance starting from spindle poles was plotted in (j) for control cells and in (k) for HSP70-depleted cells. l Mean Eg5 distribution curves in pLKO.1 control and HSP70-depleted metaphase cells were plotted. *p < 0.05 compared to pLKO.1.

Fig 5: HSP70 interacts with Eg5 within the spindle and is required for proper Eg5 mitotic function and localization.a Eg5 was co-immunoprecipitated with HSP70 in the mitotic cells. The immunoprecipitation using HSP70 antibody were performed for extracts of CGL2 cells that were cycling, synchronized at mitosis or arrested at mitosis by 16 h PES treatment as described and Eg5 in the immunoprecipitates were detected by Western blotting. b Immunoprecipitation by Eg5 antibody was also performed using synchronized mitotic control or HSP70-depleted CGL2 cell extracts and HSP70 in the immunoprecipitates were detected by Western blotting. c A PLA was performed, and representative images are shown for HSP70-Eg5 PLA signal (red) with counterstaining of ß-tubulin (ß-tub, green) and DAPI (blue) in untreated and PES- or VER-treated CGL2 cells. d Relative intensity of HSP70-Eg5 PLA signal in the whole cell was measured as described. The scatter plot shows the interquartile distribution of HSP70-Eg5 PLA intensity, and the numbers in parentheses indicate the number of cells measured. *p < 0.05 compared to untreated by Mann–Whitney Rank Sum test. e Representative images of mitotic cells with the indicated types of spindle are shown after staining with CEP152 (green), a-tubulin (a-tub, red) and DAPI (blue). f–g Percentages of control (f, pLKO.1) and HSP70-depleted (g, shHSP70) mitotic cells with the indicated types of spindle. At least 600 mitotic cells were counted for each experiment. The mean ± SD from three independent experiments is shown. *p < 0.05 comparing shHSP70 to pLKO.1 by Student’s t test. h Representative images are shown for control and PES- or VER-treated mitotic cells fixed without cold treatment. i Representative images of the spindles in control and PES- or VER-treated cells that were fixed after 5 min cold treatment. Left, cells contained cold-resistant spindle MT with Eg5 colocalization (Spindle MT+; k-MT-). Right, cells with spindle MTs disassembled and k-MTs revealed (Spindle MT-; k-MT+). j The bar chart shows the percentages of mitotic cells with Eg5-decorated MTs nucleating from the spindle pole. At least 300 mitotic cells were counted for each experiment. The mean ± SD from three independent experiments is shown. *p < 0.05 compared to untreated by Student’s t test. k The scatter plot shows the interquartile distribution of relative Eg5 intensities on the k-MT, measured as described. The numbers in parentheses indicate the number of the cells measured. *p < 0.05 compared to untreated by Mann–Whitney Rank Sum test.