CDK-1 Inhibition in G2 Stabilizes Kinetochore-Microtubules in the following Mitosis

Abstract

Cell proliferation is driven by cyclical activation of cyclin-dependent kinases (CDKs), which produce distinct biochemical cell cycle phases. Mitosis (M phase) is orchestrated by CDK-1, complexed with mitotic cyclins. During M phase, chromosomes are segregated by a bipolar array of microtubules called the mitotic spindle. The essential bipolarity of the mitotic spindle is established by the kinesin-5 Eg5, but factors influencing the maintenance of spindle bipolarity are not fully understood. Here, we describe an unexpected link between inhibiting CDK-1 before mitosis and bipolar spindle maintenance. Spindles in human RPE-1 cells normally collapse to monopolar structures when Eg5 is inhibited at metaphase. However, we found that inhibition of CDK-1 in the G2 phase of the cell cycle improved the ability of RPE-1 cells to maintain spindle bipolarity without Eg5 activity in the mitosis immediately after release from CDK-1 inhibition. This improved bipolarity maintenance correlated with an increase in the stability of kinetochore-microtubules, the subset of microtubules that link chromosomes to the spindle. The improvement in bipolarity maintenance after CDK-1 inhibition in G2 required both the kinesin-12 Kif15 and increased stability of kinetochore-microtubules. Consistent with increased kinetochore-microtubule stability, we find that inhibition of CDK-1 in G2 impairs mitotic fidelity by increasing the incidence of lagging chromosomes in anaphase. These results suggest that inhibition of CDK-1 in G2 causes unpredicted effects in mitosis, even after CDK-1 inhibition is relieved.

Fig 1. CDK-1 inhibition in G2 promotes bipolar spindle maintenance without Eg5 in RPE-1 cells.

A) Schematic of drug treatments used in B-E. B) Representative images of spindles following drug treatments in (A). Tubulin is shown in green, and DNA is shown in blue. Scale bar, 5 μm. C-E) Quantification of the percentage of mitotic cells with monopolar spindles following treatment with MG-132 and STLC (MG-STLC; C), MG-132 and DMSO (MG-DMSO; D), or STLC only (E) as diagrammed in (A). A statistically significant difference exists between DMSO- and RO-3306-treated cells only when cells were treated with MG-STLC (panel C; p < 0.001), but not when cells were treated with MG-DMSO or STLC only (panels D and E; p > 0.05). F) Quantification of the time dependence of the effect of RO-3306. Cells were treated as in (A), except that the length of time in media without drug and the length of time in RO-3306 were varied after thymidine washout. A statistically significant difference exists between DMSO- and RO-3306-treated cells at 3 h (p < 0.05), but a statistically significant difference was not found between DMSO- and RO-3306-treated cells at 1 h or 2 h (p > 0.05). G) Schematic of drug treatments used in H. H) Quantification of the percentage of mitotic cells with monopolar spindles following MG-132 and STLC as diagrammed in (G). A statistically significant difference exists between 0 h vs 3, 6, 9, or 12 h (p ≤ 0.001). I) Schematic of drug treatments used in J. J) Quantification of the percentage of mitotic cells with monopolar spindles following the drug treatments diagrammed in I. EtOH, ethanol; Purv, purvalanol A. A statistically significant difference exists between ethanol- and purvalanol A-treated cells (p < 0.05). For C-F, H, and J, bars represent the mean and error bars represent the standard error of the mean (s.e.m.) of at least 180 cells from at least 3 experiments.

Fig 2. The bipolarity-protective effect of CDK-1 inhibition in G2 depends on Kif15.

A) Representative images of spindles following transfection with control or Kif15 siRNA, followed by the drug treatments described in Fig 1A. In overlay images, Kif15 is shown in magenta, tubulin is shown in green, and DNA is shown in blue. Scale bar, 5 μm. B) Quantification of the percentage of mitotic cells with monopolar spindles following MG-STLC treatment as in (A). Bars represent the mean and error bars represent the standard error of the mean (s.e.m.) of at least 227 cells from 3 experiments. Among cells treated with DMSO, a statistically significant difference exists between cells subjected to control and Kif15 RNAi (p < 0.001); among cells treated with RO-3306, a statistically significant difference exists between cells subjected to control and Kif15 RNAi (p < 0.001); among cells subjected to control RNAi, a statistically significant difference exists between DMSO- and RO-3306-treated cells (p < 0.001); but among cells subjected to Kif15 RNAi, a statistically significant difference was not found between DMSO- and RO-3306-treated cells (p > 0.05). C and D) Lengths of spindles in cells transfected with control (C) or Kif15-specific siRNA (D), followed by the drug regimen described in 1A, and fixed at indicated time points after STLC application. Each bar represents at least 40 cells from at least 3 experiments. Among cells subjected to control RNAi, a statistically significant difference exists between DMSO- and RO-3306-treated cells (p < 0.001), and statistically significant differences exist among time points (p < 0.05) except between 10 and 20 min (p > 0.05). Among cells subjected to Kif15 RNAi (panel D), a statistically significant difference was not found between DMSO- and RO-3306-treated cells (p > 0.05), but statistically significant differences exist across time points (p < 0.001). E) Normalized immunofluorescence intensity of Kif15 relative to MTs on the spindle for untransfected cells treated with a double thymidine block, washout, 5 h recovery, 3 h RO-3306 or DMSO, washout, and 90 min MG-132 before fixation. Each bar represents 29 cells from 3 experiments. A statistically significant difference was not found between DMSO- and RO-3306-treated cells (p > 0.05). For C-E, box-and-whisker plots indicate the 10^th, 25^th, 50^th, 75^th, and 90^th percentile, as well as outliers.

Fig 3. CDK-1 inhibition in G2 increases K-MT stability.

A) Quantification of the percentage of mitotic cells with monopolar spindles following drug treatment as in 1A with or without the addition of 0.5 nM Taxol during MG-132 and MG-132 + STLC. Bars represent the mean and error bars represent the s.e.m. of at least 827 cells from 5 experiments. A statistically significant difference exists between DMSO- and RO-3306-treated cells (p < 0.001), and a statistically significant difference exists between cells treated or not treated with Taxol (p < 0.001). B) Representative images of RPE-1 cells treated with a double thymidine block, washout, 5 h recovery, 3 h RO-3306 or DMSO, washout, 90 min MG-132 +/- taxol, and 6 min nocodazole +/- taxol. In images presented as a heat map, black indicates low intensity and yellow-white indicates high intensity. In overlay, tubulin is shown in green, kinetochores (CREST staining) in magenta, centrin in yellow, and DNA in blue; scale bar, 5 μm. C) Quantification of the sum fluorescence intensity of MT polymer remaining in cells after the treatments described in B, normalized to DMSO without taxol. Box-and-whisker plots indicate the 10^th, 25^th, 50^th, 75^th, and 90^th percentile as well as outliers from at least 60 cells from at least 4 experiments. A statistically significant difference exists between DMSO- and RO-3306-treated cells (p = 0.003), and a statistically significant difference exists between cells treated or not treated with Taxol (p < 0.001).

Fig 4. High K-MT stability is required for the bipolarity-protective effect of RO-3306.

A) Representative images of spindles following transfection with control or HURP-targeting siRNA, followed by the drug treatments described in Fig 1A. In overlay images, HURP is shown in magenta, tubulin is shown in green, and DNA is shown in blue; scale bar, 5 μm. B) Quantification of the percentage of mitotic cells with monopolar spindles following MG-STLC treatment as in (A). Bars represent the mean and error bars represent the s.e.m. at least 61 cells from at least 3 experiments. Among cells treated with DMSO, a statistically significant difference exists between cells subjected to control and HURP RNAi (p < 0.001); among cells treated with RO-3306, a statistically significant difference exists between cells subjected to control and HURP RNAi (p < 0.001); among cells subjected to control RNAi, a statistically significant difference exists between DMSO- and RO-3306-treated cells (p < 0.001); but among cells subjected to HURP RNAi, a statistically significant difference was not found between DMSO- and RO-3306-treated cells (p > 0.05). C) Quantification of the percentage of mitotic cells with monopolar spindles following MG-STLC treatment as in (A), but with the inclusion of 0.5 nM Taxol during MG-STLC treatment. A statistically significant difference exists between cells subjected to control RNAi and HURP RNAi (p < 0.001), and a statistically significant difference exists between DMSO- and RO-3306-treated cells (p = 0.005). D) Normalized immunofluorescence intensity of HURP relative to MTs on the spindle for untransfected cells treated with a double thymidine block, washout, 5 h recovery, 3 h RO-3306 or DMSO, washout, and 90 min MG-132 before fixation. Bars represent at least 39 cells from 4 experiments. Box-and-whisker plots indicate the 10^th, 25^th, 50^th, 75^th, and 90^th percentile as well as outliers. A statistically significant difference exists between DMSO- and RO-3306-treated cells (p < 0.05).

Fig 5. CDK-1 inhibition in G2 increases the duration of mitosis and the frequency of mitotic errors.

A) Frames from DIC movies of representative cells undergoing mitosis following a double thymidine block, washout, 5 h recovery, 3 h RO-3306 or DMSO, and washout. The frames of nuclear envelope break down (NEBD) and anaphase onset (AO) are indicated; numbers indicate minutes from NEBD; scale bars, 5 μm. B) Histogram of the time from NEBD to AO in cells treated as in A. Each population includes at least 116 cells from 3 experiments. A statistically significant difference exists between DMSO- and RO-3306-treated cells (p < 0.001). C) Histogram of the time from NEBD to AO in cells treated as in A, but with the addition of 1 μM Reversine after washout of DMSO or RO-3306. Each population includes at least 80 cells from at least 3 experiments. A statistically significant difference was not found between DMSO- and RO-3306-treated cells (p > 0.05). D and E) Illustrative images of cells treated as in A with DMSO (D) or with RO-3306 (E) and fixed during early anaphase (left) or late anaphase/telophase (right). In overlays, kinetochores (CREST) are shown in magenta, and DNA is shown in blue. Although cells with lagging chromosomes were a minority population, they are shown to illustrate their morphology. Scale bars, 5 μm. F) Quantification of the percentage of anaphase/telophase cells with lagging chromosomes following the treatments described in C. Box plots indicate the 25^th, 50^th, and 75^th percentile from at least 945 cells from 4 experiments. A statistically significant difference exists between DMSO- and RO-3306-treated cells (p < 0.01).