Spindle defects in HeLa cells lacking Sgt1. (A) CREST (red), α-tubulin (green) and DNA (blue) were visualized to investigate spindle structure. Panels A1–A3: Normal control cells at different stages of mitosis; panels A4–A9: spindle defects in interfered cells. (B) Centrosome problems revealed with an anti-γ-tubulin antibody (red). Panel B1: A normal control cell; panels B2 and B3: two interfered cells with fragmented and faint γ-tubulin staining.

A kinetochore defect in Sgt1-depleted HeLa cells. (A) Localization of Hec1 to kinetochores was compared in control and Sgt1-depleted cells in the absence of spindle poisons. The CREST antiserum stains the inner kinetochore. In control cells, Hec1 is localized at the periphery of CREST staining. There was no kinetochore localization of Hec1 after Sgt1 depletion. (B) Localization of CENP-E to the outer kinetochore is affected by loss of Sgt1. (C) Localization of CENP-F is also affected, whereas there is substantial residual cytoplasmic staining. (D) CENP-I is absent from the inner kinetochore of cells lacking Sgt1. (E) Localization of CENP-C is unaffected.

Silencing of Sgt1 affects progression through mitosis. (A) Western blot of HeLa cell lysates harvested at the indicated time points after transfection with siRNA duplexes against Sgt1. Control injections of anti-luciferase (Luc) siRNAs are also shown. Both splicing isoforms of Sgt1 are silenced. (B) DNA content was analysed by FC on PIstained cells. Cells lacking Sgt1 accumulate with 4N DNA, with an increase in a sub-G1 population 96-h after siRNA transfection. (C) A cell (white arrowhead) lacking Sgt1 enters mitosis and remains arrested for several hours, finally exiting without dividing. This cell eventually died around 75 h. (D) Trivariate FC analysis of cells lacking Sgt1. White bars: Cells with 4N DNA content; grey bars: cells with 4N DNA content that were H3-P positive (mitotic cells, M); black bars: cells with 4N DNA content that were positive for Cyclin B1 (total of G2- and M-phase cells). Note the increase of H3-P-positive cells following Sgt1 depletion, with a concomitant modest increase in G2–M cells at 72 h. More than 20% of the cell population had 4N DNA content but stained negative for mitotic markers, indicative of a tetraploid G1 population.

Mitotic delay of cells lacking Sgt1 is caused by the spindle checkpoint. (A) Mad1 is absent from the kinetochores of Sgt1-depleted cells despite evident lack of microtubule attachment. (B) Addition of nocodazole did not cause a return of Mad1 to kinetochores. (C) Also BubR1 was missing from the kinetochores of Sgt1-depleted cells. (D) Parallel silencing of Sgt1 and Mad2 by RNAi as revealed by western blot on total cell lysates 72 and 96 h after siRNAs transfection. Mad1, CENP-I and vinculin are shown as controls. (E) Bivariate FC analysis on the indicated samples. White bars: 4N cells; grey bars: 4N cells that were H3-P positive. Dual ablation of Mad2 and Sgt1 relieves the mitotic block caused by Sgt1 ablation. (F) Cells lacking Sgt1 are unable to maintain a mitotic arrest beyond an average of 11 h, even if nocodazole is added. Black bars: Duration of mitosis in cells filmed by time-lapse microscopy from 48 to 96 h after siRNA transfection. (G) Sgt1-depleted cells cannot mount a full-fledge MSC response in nocodazole. White bars: 4N cells; grey bars: 4N cells that were positive for H3-P. The lower percentage of 4N cells in the total cell population after Sgt1 depletion is explained by a G1 arrest, as documented in SF 2. Note that the scale of this figure is different from that used in (E).