Clks 1, 2 and 4 prevent chromatin breakage by regulating the Aurora B-dependent abscission checkpoint

Abstract

When chromatin is trapped at the intercellular bridge, cells delay completion of cytokinesis (abscission) to prevent chromosome breakage. Here we show that inhibition of Cdc-like kinases (Clks) 1, 2 or 4 accelerates midbody resolution in normally segregating cells and correlates with premature abscission, chromatin breakage and generation of DNA damage in cytokinesis with trapped chromatin. Clk1, Clk2 and Clk4 localize to the midbody in an interdependent manner, associate with Aurora B kinase and are required for Aurora B-serine 331 (S331) phosphorylation and complete Aurora B activation in late cytokinesis. Phosphorylated Aurora B-S331 localizes to the midbody centre and is required for phosphorylation and optimal localization of the abscission protein Chmp4c. In addition, expression of phosphomimetic mutants Aurora B-S331E or Chmp4c-S210D delays midbody disassembly and prevents chromatin breakage in Clk-deficient cells. We propose that Clks 1, 2 and 4 impose the abscission checkpoint by phosphorylating Aurora B-S331 at the midbody.

Figure 1. Clk inhibition accelerates midbody disassembly in normally segregating cells.

(a) Midbody disassembly in the absence (control) or in the presence of TG003. Cells expressing mCherry:tubulin were analysed by time-lapse microscopy. Phase-contrast (right panels) and fluorescence images (left panels) are shown. Time from midbody formation to midbody disassembly is indicated. Midbodies are shown by solid arrows and midbody disassembly is indicated by broken arrows. (b) Frequency of bi-/multinucleate or midbody stage cells. Cells were transfected with negative siRNA (control), Clk1 siRNA (siClk1), Clk2 siRNA (siClk2), Clk4 siRNA (siClk4) or treated with TG003 for 5 h. Error bars show the s.d. from the mean from three independent experiments. A minimum of 300 cells were analysed per experiment. ***P<0.001 compared with control. The Student's t-test was used. (c) Examples of midbody stage cells (shown in brackets). Green, α-tubulin; blue, DNA. (d) Example of a binucleate cell after TG003 treatment. Green, actin; blue, DNA. (e) Localization of Clk1, Clk2 or Clk4 in cytokinesis. Green, Clk1, Clk2 or Clk4; red, CENP-A; blue, DNA. Twenty cells from two independent experiments were examined. (f) Localization of Clk2 in cells transfected as in b. Green, Clk2; red, α-tubulin; blue, DNA. Tubulin values indicate midbody thickness. Insets show magnified midbodies. (g) Mean midbody intensity of Clk2 in cells treated as in f. Data are from n cells from three independent experiments. Error bars show s.d. Values in control were set to 1. ***P<0.001 compared with control. The Mann–Whitney U-test was used. (h) Localization of Clk1:GFP, Clk2 or Clk4:GFP in purified midbodies. Green, GFP or Clk2; red, Aurora B. Scale bars, 5 μm.

Figure 2. Aurora B associates with Clks in cell lysates.

(a,b) Aurora B–S331 phosphorylation (pS331) in the absence (control) or presence of TG003 for 5 h. Green, pS331; red, α-tubulin; blue, DNA. (c) Mean midbody intensity of pS331 in cells treated as in a,b. Data are from n cells from three independent experiments. Values in control were set to 1. Error bars show s.d. ***P<0.001 compared with control. The Mann–Whitney U-test was used. (d) Localization of total Aurora B in cells treated as in b. Green, Aurora B; red, α-tubulin; blue, DNA. Thirty cells from three independent experiments were examined per treatment. Tubulin values indicate midbody thickness. Insets show magnified midbodies. Scale bars, 5 μm. (e) Clk2 in vitro kinase assay. Autoradiography analysis of Aurora B substrates (top) and Ponceau staining (bottom). (f) GST pull-down assay from cytokinesis-enriched cells. Lysates expressing Clk1:GFP, Clk2:GFP or Clk4:GFP proteins were incubated with 10 μg glutathione–agarose-bound WT GST–tagged Aurora B (GST-Aur B^WT) or GST. Associated GFP (top) or total GFP and actin (bottom) were detected by western blotting and GST proteins (middle) by Ponceau staining. (g) Immunoprecipitation assay from cytokinesis-enriched cells. Immunoprecipitated (IP) Aurora B (top) and total Aurora B and actin (bottom) were detected by western blotting.

Figure 3. Expression of the phosphomimetic S331E Aurora B or S210D Chmp4c mutants rescues the frequency of cells at midbody stage after Clk inhibition.

(a) Localization of phosphorylated Chmp4c-S210, S214 and S215 (pChmp4c) in the absence (control) or presence of TG003 for 5 h, or after transfection of cells with Chmp4c siRNA (siChmp4c). Green, pChmp4c; red, α-tubulin; blue, DNA. (b) Mean midbody intensity of pChmp4c in cells treated as in a. Data are from n cells from three independent experiments. Error bars show s.d. Values in control were set to 1. ***P<0.001. The Mann–Whitney U-test was used. (c) Localization of total Chmp4c in cells treated as in a. Green, Chmp4c; red, α-tubulin; blue, DNA. Thirty cells from three independent experiments were examined per treatment. Tubulin values indicate midbody thickness. Insets show magnified midbodies. (d,e) Frequency of cells at midbody stage. (d) Cells expressing siRNA-resistant forms of WT or S331E Myc-tagged Aurora B (Myc-Aurora B) were depleted of endogenous Aurora B by siRNA and treated as in a. (e) Cells expressing WT, S210D or S210A Chmp4c:GFP were treated as in a. Error bars show the s.d. from the mean from three independent experiments. A minimum of 200 cells were analysed per experiment. ***P<0.001. Student's t-test was used. (f) Breakage of intercellular bridges after Clk inhibition. HeLa cells expressing LAP2b:RFP were analysed by time-lapse microscopy in the absence (control) or presence of TG003. Time from formation of the LAP2b:RFP bridge is indicated. Higher magnification insets of LAP2b:RFP bridges are shown. (g) Intercellular bridges in fixed BE cells treated as in a. Thirty-five cells from three independent experiments were examined per treatment. Green, α-tubulin; red, actin; blue, DNA. Higher magnification insets of tubulin staining are shown. Intact intercellular or chromatin bridges are indicated by solid arrowheads and broken bridges by open arrowheads. Scale bars, 5 μm.

Figure 4. Clk inhibition correlates with chromatin breakage and generation of DNA damage.

(a) Cells were transfected with negative siRNA (control), Clk1 siRNA (siClk1), Clk2 siRNA (siClk2), Clk4 siRNA (siClk4) or treated with TG003 for 5 h. Intact DNA bridges are indicated by solid and broken bridges by open arrowheads. (b) Broken bridges index analysis in cells treated as in a. Error bars show the s.d. from the mean from three independent experiments. A minimum of 50 cells with chromatin bridges were analysed per experiment. (c) Cells expressing GFP or GFP:Vps4-K173Q were treated as in a. Broken bridges index shows the percentage of green cells with broken chromatin bridges/total green cells with chromatin bridges. Error bars show the s.d. from the mean from three independent experiments. A minimum of 50 green cells with chromatin bridges were analysed per experiment. (d) Lamin B-negative micronuclei (marked by arrow) in cells treated as in a. Green, lamin B2; red, actin; blue, DNA. (e) γ-H2AX-positive micronuclei (indicated by arrow) in cells treated as in a. Green, γ-H2AX; red, actin; blue, DNA. Scale bars, 5 μm. (f) Frequency of cells exhibiting lamin B-negative or γ-H2AX-positive micronuclei after treatment as in a. Error bars show the s.d. from the mean from three independent experiments. A minimum of 150 cells were analysed per experiment. (g) Frequency of cells at midbody stage. Cells were transfected with negative siRNA (control), Nup153 siRNA (siNup153) or combinations of Nup153 siRNA and Clk1 siRNA (siNup153+siClk1), Nup153 siRNA and Clk2 siRNA (siNup153+siClk2) or Nup153 siRNA and Clk4 siRNA (siNup153+siClk4). Error bars show the s.d. from the mean from three independent experiments. A minimum of 300 cells were analysed per experiment. ***P<0.001 compared with control. Student's t-test was used. (h) Westen blot analysis of Nup153 and lamin B2 in cells treated as in g.

Figure 5. Clks localize to the midbody in a mutually dependent manner.

(a) Cells expressing Clk1:GFP, Clk2:GFP or Clk4:GFP proteins were analysed in cytokinesis with chromatin bridges. Green, GFP; red, CENP-A; blue, DNA. A minimum of 22 cells from three independent experiments were examined per treatment. (b,c) Localization of Clk2. Cells were transfected with negative siRNA (control), Clk1 siRNA (siClk1), Clk4 siRNA (siClk4) or treated with TG003 for 5 h. Green, Clk2; red, CENP-A; blue, DNA. (d) Mean midbody intensity of Clk2 in cells treated as in b,c. (e) Localization of phosphorylated Aurora B–S331 (pS331) in cells treated as in b. Green, pS331; red, CENP-A; blue, DNA. Insets show magnified midbodies. Scale bars, 5 μm. (f) Mean midbody intensity of pS331 in cells transfected with Clk2 siRNA (siClk2) or treated as in b,c. Data are from n cells from three independent experiments. Error bars show s.d. Values in control were set to 1. ***P<0.001 compared with control. The Mann–Whitney U-test was used.

Figure 6. Clk inhibition reduces Chmp4c phosphorylation at the midbody in cells with chromatin bridges.

(a) Localization of total Aurora B. Cells were transfected with negative siRNA (control) or treated with TG003 for 5 h. Green, Aurora B; red, CENP-A; blue, DNA. (b) Mean midbody intensity of Aurora B in cells treated as in a or transfected with Clk1 siRNA (siClk1), Clk2 siRNA (siClk2) or Clk4 siRNA (siClk4). (c) Localization of phosphorylated Chmp4c-S210, S214 and S215 (pChmp4c) in the absence (control) or presence of TG003 for 5 h. Green, pChmp4c; red, CENP-A; blue, DNA. (d) Mean midbody intensity of pChmp4c. Data are from n cells from three independent experiments. Error bars show s.d. Values in control were set to 1. ***P<0.001 compared with control. The Mann–Whitney U-test was used. (e) Localization of total Chmp4c in cells treated as in c. Green, Chmp4c; red, CENP-A; blue, DNA. Insets show magnified midbodies. Scale bars, 5 μm. (f) Broken bridges index analysis. Cells expressing siRNA-resistant forms of WT or S331E Myc-tagged Aurora B (Myc-Aurora B) were depleted of endogenous Aurora B by siRNA and treated as in c. Error bars show the s.d. from the mean from three independent experiments. A minimum of 50 cells in cytokinesis with chromatin bridges were analysed per experiment. ***P<0.001 compared with control. Student's t-test was used.

Figure 7. Expression of the phosphomimetic mutant S331E Aurora B rescues Chmp4c phosphorylation at the midbody in Clk-deficient cells.

(a) Localization of phosphorylated Chmp4c-S210, S214 and S215 (pChmp4c). Cells expressing siRNA-resistant forms of WT or S331E Myc-tagged Aurora B (Myc-Aurora B) were depleted of endogenous Aurora B by siRNA in the absence (control) or presence of TG003 for 5 h. Green, Myc; red, pChmp4c; blue, DNA. Insets show magnified midbodies. Scale bars, 5 μm. (b) Mean midbody intensity of pChmp4c from cells treated as in a. Data are from n cells from three independent experiments. Values in control were set to 1. Error bars show s.d. ***P<0.001 compared with control. The Mann–Whitney U-test was used. (c) Broken bridges index analysis in cells expressing WT, S210D or S210A Chmp4c:GFP in the absence (control) or presence of TG003 for 5 h. Error bars show s.d. from the mean from three independent experiments. A minimum of 50 cells in cytokinesis with chromatin bridges were analysed per experiment. ***P<0.001 compared with control. Student's t-test was used. (d) Model for the role of Clks in abscission. p, phosphorylation.