Generation of a Spindle Checkpoint Arrest from Synthetic Signaling Assemblies

Abstract

The spindle checkpoint acts as a mitotic surveillance system, monitoring interactions between kinetochores and spindle microtubules and ensuring high-fidelity chromosome segregation [1-3]. The checkpoint is activated by unattached kinetochores, and Mps1 kinase phosphorylates KNL1 on conserved MELT motifs to generate a binding site for the Bub3-Bub1 complex [4-7]. This leads to dynamic kinetochore recruitment of Mad proteins [8, 9], a conformational change in Mad2 [10-12], and formation of the mitotic checkpoint complex (MCC: Cdc20-Mad3-Mad2 [13-15]). MCC formation inhibits the anaphase-promoting complex/cyclosome (Cdc20-APC/C), thereby preventing the proteolytic destruction of securin and cyclin and delaying anaphase onset. What happens at kinetochores after Mps1-dependent Bub3-Bub1 recruitment remains mechanistically unclear, and it is not known whether kinetochore proteins other than KNL1 have significant roles to play in checkpoint signaling and MCC generation. Here, we take a reductionist approach, avoiding the complexities of kinetochores, and demonstrate that co-recruitment of KNL1^Spc7 and Mps1^Mph1 is sufficient to generate a robust checkpoint signal and prolonged mitotic arrest. We demonstrate that a Mad1-Bub1 complex is formed during synthetic checkpoint signaling. Analysis of bub3Δ mutants demonstrates that Bub3 acts to suppress premature checkpoint signaling. This synthetic system will enable detailed, mechanistic dissection of MCC generation and checkpoint silencing. After analyzing several mutants that affect localization of checkpoint complexes, we conclude that spindle checkpoint arrest can be independent of their kinetochore, spindle pole, and nuclear envelope localization.

Keywords: Bub; Mad; Mps1; cell cycle; checkpoint; kinetochore; mitosis; spindle; synthetic.

Graphical abstract

Figure 1

Co-tethering of Spc7-9TE and TetR-Mph1ΔN Generates a Robust Checkpoint Arrest (A) Schematic model of kinetochore-based checkpoint signaling versus the synthetic tetO platform. (B) TetR-Spc7-9TE is sufficient to recruit Bub1-GFP, Bub3-GFP, and Mad3-GFP to an array of Tet operators on a chromosome arm. Scale bar, 10 μm. See Figure S1 for TetR-Spc7wt and TetR-Spc7-9TA images. (C) Schematic summary of Spc7-9TE tethering. (D) Co-expression of TetR-Spc7-9TE and TetR-Mph1ΔN produces a robust mitotic arrest with short metaphase spindles. Scale bar, 10 μm. (E) Quantitation of arrested cells after 12, 14, 16, and 18 hr of Mph1 induction (cells grown without thiamine). The plus thiamine control culture does not arrest, containing just a few mitotic cells. 36 experiments were performed and data points are plotted along with the mean and SD. See also Figure S1.

Figure 2

Dependencies for Synthetic Checkpoint Arrest (A) Co-expression of TetR-Spc7-9TE and TetR-Mph1ΔN leads to a metaphase arrest with Mad2-GFP accumulating at the spindle poles (analyzed in detail in Figure 3). Scale bar, 10 μm. (B) Expression of either TetR-Spc7-9TE or TetR-Mph1ΔN alone is not sufficient for robust arrest. This experiment was repeated three times and is plotted as the mean ± SD. (C) Comparison of TetR-Spc7-9TE, TetR-wild-type Spc7 (Spc7-wt), and TetR-Spc7-9TA. The latter is unable to arrest cells, whereas the wild-type protein arrests better than Spc7-9TE. This experiment was repeated three times and is plotted as the mean ± SD. (D) The tetO array is not necessary for Mad2-GFP accumulation at spindle poles or metaphase arrest. The mitotic arrest, for both TetR-Spc7-wt and TetR-Spc7-9TE, was compared in strains containing either 112xtetO or no tet operators. This experiment was repeated three times and is plotted as the mean ± SD. (E) No arrest was observed when TetR was removed from the Mph1 fusion protein (Mad2-GFP did not accumulate at spindle poles). Scale bar, 10 μm. Anti-Flag (Mph1) immunoblot of whole cell extracts demonstrates that similar levels of Mph1 were expressed with and without TetR. (F) The mitotic arrest is Mad1, Mad2, Mad3, and Bub1 dependent, but independent of Bub3, Bub1 kinase activity, and Sgo2. These strains were analyzed at least three times and data plotted as the mean ± SD. The control strain (TetR-Spc7-9TE) on the left has Atb2-GFP as reporter and on the right Mad2-RFP. All strains contained the tetO array, apart from sgo2Δ and its corresponding control strain. Representative images are presented in Figure S2J. See also Figure S2.

Figure 3

Mad2p Accumulates at Spindle Poles in the Synthetic Arrest, but This Is Not Necessary for the Arrest (A) Cells were arrested with co-expression of TetR-Mph1ΔN and TetR-Spc7-9TE. Co-localization of the spindle pole marker Pcp1-RFP and Mad2-GFP is observed. The Mad2-GFP does not co-localize well with the kinetochore marker Fta3-RFP in arrested cells, although in a few cases kinetochores are close to the poles. Scale bar, 5 μm. Figure S3A demonstrates co-immunoprecipitation of Mad2 with gamma tubulin complex proteins. (B) Strains co-expressing Spc7 and Mph1 do not accumulate Mad2-GFP at spindle poles in strains containing the mad1-KAKA mutation that disrupts the Mad1-Cut7 kinesin motor interaction. Other motor mutants were analyzed (dynein, klp2Δ, klp5/6Δ) but found to have no effect on the arrest or Mad2-GFP localization to spindle poles (see Figure S3G). Scale bar, 10 μm. The mad1-ΔCC allele still arrests even though localization of Mad1 and Mad2 to the nuclear periphery/envelope and spindle poles is lost. This N-terminal coiled-coil domain also includes the Cut7 interaction site. (C) Quantitation of the mad1-KAKA and mad1-ΔCC mutant arrests. This experiment was repeated five times and data plotted as the mean ± SD. (D) The levels of Mph1 expression and Mad1 protein stability are not affected in these mad1 mutants. Time of Mph1 induction (after thiamine wash-out) is indicated. (E) Model with the Cut7 kinesin moving the Mad-Bub complex to spindle poles. This predicts that the movement of Bub1 to spindle poles is Bub3 independent, which was found to be the case (see Figure S3D). See also Figure S3.

Figure 4

Synthetic Checkpoint Arrest Requires Mad1-Bub1 Complex Formation and Is Inhibited by Bub3 (A) Cells containing Bub1-GFP, TetR-Spc7-9TE, and TetR-Mph1ΔN were arrested as above for 16 hr then cross-linked with DSP before harvesting. Mad1 co-immunoprecipitates with Bub1-GFP in the arrested cells, both in the presence and absence of the tetO array. On the immunoblots, Bub1-GFP was detected with anti-Bub1 antibody and Mad1 with anti-Mad1 (see Supplemental Experimental Procedures). (B) There is no arrest in the bub1-CD1 mutant, which disrupts the Bub1-Mad1 interaction (see Figure S4A for images). This experiment was repeated three times, and data were plotted as the mean ± SD. The mutant Bub1-CD1 protein is stable (see Figure S4B). (C) bub3Δ mutants containing TetR-Spc7-9TE arrest significantly faster than bub3⁺, TetR-Spc7-9TE. This experiment was repeated three times, and data were plotted as the mean ± SD. (D) TetR-Spc7-T9A combined with TetR-Mph1ΔN is able to arrest cells in the absence of Bub3. This experiment was repeated three times, and data were plotted as the mean ± SD. (E) Higher levels of the Mad1-Bub1 complex are generated in bub3Δ cells. These cells contained TetR-Mph1ΔN and TetR-Spc7-9TE and were harvested after 12 hr of TetR-Mph1 induction. On the immunoblots, Bub1-GFP was detected with polyclonal anti-Bub1 antibodies, and Mad1 with anti-Mad1 antibodies (see Supplemental Experimental Procedures). (F) Working model: diffusible heterodimers of TetR-Mph1ΔN and TetR-Spc7(1-666) actively produce a phospho-dependent Bub1-Mad1 complex, that than acts as an assembly platform for MCC production. In the absence of Bub3, shown to the right, Bub1 becomes hyperphosphorylated, which can enhance Bub1-Mad1 complex assembly and MCC production. See also Figure S4.