Accumulation of Mad2-Cdc20 complex during spindle checkpoint activation requires binding of open and closed conformers of Mad2 in Saccharomyces cerevisiae

Abstract

The spindle assembly checkpoint (SAC) coordinates mitotic progression with sister chromatid alignment. In mitosis, the checkpoint machinery accumulates at kinetochores, which are scaffolds devoted to microtubule capture. The checkpoint protein Mad2 (mitotic arrest deficient 2) adopts two conformations: open (O-Mad2) and closed (C-Mad2). C-Mad2 forms when Mad2 binds its checkpoint target Cdc20 or its kinetochore receptor Mad1. When unbound to these ligands, Mad2 folds as O-Mad2. In HeLa cells, an essential interaction between C- and O-Mad2 conformers allows Mad1-bound C-Mad2 to recruit cytosolic O-Mad2 to kinetochores. In this study, we show that the interaction of the O and C conformers of Mad2 is conserved in Saccharomyces cerevisiae. MAD2 mutant alleles impaired in this interaction fail to restore the SAC in a mad2 deletion strain. The corresponding mutant proteins bind Mad1 normally, but their ability to bind Cdc20 is dramatically impaired in vivo. Our biochemical and genetic evidence shows that the interaction of O- and C-Mad2 is essential for the SAC and is conserved in evolution.

Figure 1.

The mad2^RA and mad2^QA point mutant alleles do not complement the deletion of the MAD2 gene in S. cerevisiae. Strains with the indicated genotypes were grown to log phase, arrested in G1 by α factor, and released in fresh medium containing nocodazole. At the indicated times, cell samples were withdrawn for FACS analysis of DNA contents (A) and to score the percentage of budded and rebudded cells as well as the percentage of sister chromatid separation (B). wt, wild type.

Figure 2.

The O and C conformers of ScMad2^wt form a complex that requires Arg126 and Gln127. (A) GST (lanes 1–4), GST-Mad1^563–590 (lanes 5–8), and GST-Cdc20^184–210 (lanes 9–12) were immobilized on GSH beads at ∼1.0 μM and incubated with ∼5 μM Mad2^wt or Mad2^ΔC (lanes 13 and 14) for 1 h. GSH beads were collected by centrifugation, washed, and bound proteins were analyzed by SDS-PAGE. For samples in lanes 8 and 12, beads were incubated with ∼5 μM Mad2^wt, washed, incubated with the same concentration of Mad2^ΔC for an additional hour, and analyzed by SDS-PAGE. (B) The experiment was performed as in A but with Mad2^RQEA and Mad2^ΔC. Although Mad2^RQEA binds GST-Mad1 and Cdc20 as well as Mad2^wt, it is unable to bind Mad2^ΔC (lanes 8 and 12). (C) 50 μl of a 20-μM solution of Mad2^ΔC was analyzed by SEC on a Superdex-200 PC 3.2/30 column and found to elute as a monomer. The content of 14 30-μl consecutive fractions eluting between 1.4 and 1.82 ml was analyzed by SDS-PAGE and was Coomassie stained. (D) To generate C-Mad2^wt, 200 μM Cdc20^195–211 synthetic peptide was incubated with 20 μM Mad2^wt for 1 h. The sample was analyzed by SEC as in C. (E) As in D but with Mad2^RQEA and the Cdc20^195–211 peptide. (F) C-Mad2^wt–Cdc20^195–211 was mixed with Mad2^ΔC for 1 h before separation by SEC. Dimerization of O and C conformers was revealed by a shift in elution volume relative to O- and C-Mad2. (G) The same experiment with C-Mad2^RQEA–Cdc20^195–211 rather than with Mad2^wt shows that this double point mutant protein is unable to bind O-Mad2. AU, arbitrary unit.

Figure 3.

Monomers and dimers of ScMad2^wt. Different Mad2 species (at a concentration of 20 μM) were analyzed using a Superdex-75 PC 3.2/30 SEC column. For each panel, 14 30-μl fractions between 0.94 and 1.36 ml were analyzed by SDS-PAGE. For each panel, the elution volumes of C-Mad2–O-Mad2 dimers (D) and the O-Mad2 monomer (M′) or C-Mad2 monomer (M″) are marked. In D and E, the black dotted lines mark the elution volume of M′ (O-Mad2), which is shifted relative to M″. (A) Pure ScMad2 eluted as a monomer. (B) Upon the addition of Cdc20^195–211 at 1/4 of the Mad2 concentration, ∼50% of Mad2 is shifted into a dimer peak. (C) Cdc20^195–211 at 1/2 of the Mad2 concentration causes most Mad2 to shift into a dimer peak. (D) Upon the addition of superstoichiometric Cdc20, a C-Mad2 monomer accumulates. (E) The process of the creation of C-Mad2–Cdc20 is complete. AU, arbitrary unit.

Figure 4.

O-Mad2 binds the Mad1–Mad2 core complex. (A) Mad1–Mad2 forms a stable tetrameric assembly, the Mad1–Mad2 core complex (Sironi et al., 2002). A recombinant yeast complex containing the C-terminal region of Mad1 (residues 529–750 lacking the N-terminal kinetochore-binding domain of Mad1) was coexpressed in bacteria with Mad2^wt, and the resulting complex was purified to homogeneity (see Materials and methods). 50 μl Mad1–Mad2 complex (10 μM) was analyzed by SEC on a Superdex-200 PC 3.2/30 column. Fractions between 1.15 and 1.85 ml were analyzed by SDS-PAGE. (B) SEC profile of Mad2^wt covalently labeled with AlexaFluor488. The content of the elution fractions was analyzed after SDS-PAGE on a UV trans-illuminator (top) and by Coomassie staining (bottom). (C and D) AlexaFluor488-Mad2^ΔC (C) and AlexaFluor488-Mad2^RQEA (D) was analyzed as in B. (E) Mad2^wt (O-Mad2) was incubated stoichiometrically with Mad1^529–750–Mad2^wt, and the resulting sample was analyzed by SEC. Most of the AlexaFluor488 signal associated with Mad2^wt was incorporated in a high molecular weight complex, indicating binding to Mad1–Mad2. (F) The same experiment was repeated using AlexaFluor488-Mad2^ΔC. Also in this case, the AlexaFluor signal was shifted to a high molecular weight complex with Mad1–Mad2. (G) AlexaFluor488-Mad2^RQEA fails to bind Mad1–Mad2, indicating that Arg126 and Gln127 are part of the binding interface. (H) Mad2^wt was incubated stoichiometrically with Mad1^529–750–Mad2^wt in the presence of Cdc20^195–211. The AlexaFluor488 signal associated with Mad2^wt is released from the Mad1–Mad2 complex. (I) As in H but with AlexaFluor488-Mad2^ΔC, which does not bind Cdc20 and is not released from Mad1–Mad2. AU, arbitrary unit.

Figure 5.

The O–C interaction of Mad2 is conserved in evolution. (A) Human O-Mad2 binds the Mad1–C-Mad2 complex. 50 μl of human Mad1–Mad2 complex (10 μM of divalent complex) was combined stoichiometrically with 20 μM of human AlexaFluor488-Mad2^ΔC and analyzed by SEC on a Superdex-200 PC 3.2/30 column. Fractions between 1.15 and 1.85 ml were analyzed by SDS-PAGE. Human Mad1–Mad2 and Mad2 were expressed and purified as described previously (De Antoni et al., 2005a). (B) AlexaFluor488-Mad2^ΔC from S. cerevisiae was incubated with human Mad1–Mad2 and analyzed as in A. (C) AlexaFluor488-ScMad2^wt was incubated with ScCdc20^195–212. This yeast C-Mad2–Cdc20 complex did not bind human Mad1–Mad2. (D) AlexaFluor488-HsMad2^ΔC was incubated stoichiometrically with yeast Mad1–Mad2 and analyzed as in A. (E) AlexaFluor488-HsMad2^wt was incubated with a synthetic peptide encompassing the Mad2-binding segment of HsCdc20 (Cdc20^111–138). This human C-Mad2–Cdc20 complex did not bind yeast Mad1–Mad2. AU, arbitrary unit.

Figure 6.

Mad2^ΔC has a dominant-negative effect on the checkpoint. Strains with the indicated genotypes were grown to log phase in YEPR medium, arrested in G1 by α factor, and released into YEPRG medium containing nocodazole. 1% galactose was added to the cultures half an hour before the release to induce the GAL1 promoter. At the indicated times, cell samples were withdrawn for FACS analysis of DNA contents (A) and to score the percentage of budded and rebudded cells as well as the percentage of sister chromatid separation (B).

Figure 7.

Mutations in the binding interface between O- and C-Mad2 impair Cdc20 binding. (A) Protein extracts were prepared from cycling cells of untagged wild-type (W303; lanes 1 and 6), MAD1-myc18 MAD2 (ySP2218; lanes 2 and 7), and MAD1-myc18 mad2Δ strains carrying the MAD2^wt (ySP5314; lanes 3 and 8), mad2^RA (ySP5316; lanes 4 and 9), and mad2^QA (ySP5318; lanes 5 and 10) alleles integrated at the LEU2 locus. Total extracts and anti-myc IPs were analyzed by Western blotting (WB) to detect Mad1-myc18 and Mad2. (B) Cycling cultures of untagged wild-type (W303; lanes 1 and 6), myc18-CDC20 MAD2 (ySP1413; lanes 2 and 7), and myc18-CDC20 mad2Δ strains carrying, respectively, the MAD2^wt (ySP5311; lanes 3 and 8), mad2^RA (ySP5355; lanes 4 and 9), and mad2^QA (ySP5356; lanes 5 and 10) alleles integrated at the LEU2 locus were arrested in G1 by α factor and released in the presence of nocodazole. After 80 min, cells were collected, and protein extracts were used for IPs with anti-myc antibodies. FACS analysis (not depicted) confirmed that cells were in G2/M. Total extracts and immunoprecipitates were analyzed by Western blotting to visualize myc18-Cdc20 and Mad2. (C) The same extracts as in B were used for IPs with anti-Mad2 polyclonal antibodies and analyzed by Western blotting as in B.

Figure 8.

Implications of the Mad2 template model. (A) There are two pools of Mad2: a cytosolic O-Mad2 pool and a C-Mad2 pool bound to Mad1. The latter is a template required to create C-Mad2 bound to Cdc20, the “copy.” In the absence of Mad1–Mad2, the Mad2–Cdc20 complex does not form efficiently. We speculate that the binding reaction, which implies a big conformational change for Mad2, is slow. (B) O-Mad2 is recruited from the cytosolic to Mad1–Mad2 at the kinetochore. Within the core complex, C-Mad2 is responsible for the interaction with O-Mad2. (C) Recruitment to the Mad1–Mad2 complex is impaired if Mad2 contains mutations such as RA and QA that affect its ability to bind O-Mad2. Under these conditions, the checkpoint cannot be activated. (D) The O-Mad2 molecule bound to C-Mad2 binds Cdc20 to create a new C-Mad2 conformer. The green circle enclosing O-Mad2 signifies that this monomer, not the C-Mad2 monomer bound to Mad1, is transferred to Cdc20. The representation of this monomer as O-Mad2 is possibly a simplification. Because we presume that prior binding of O-Mad2 to C-Mad2 accelerates binding to Cdc20 relative to cytosolic O-Mad2, this monomer might be characterized by a partially unfolded conformation of the C-terminal tail of Mad2, representing a transition state from the open to the closed conformation. (E) The C-Mad2–Cdc20 complex is a copy of the C-Mad2–Mad1 complex (the template). The decisive difference between these complexes is likely that Mad1–Mad2 is very stable, whereas Cdc20–Mad2 exists transiently and its concentrations can be reversed. (F) The Mad2 template hypothesis postulates that C-Mad2–Cdc20 acts in the cytosol like C-Mad2–Mad1 at the kinetochore. The reaction is similar to that shown in B. In comparison with C, it is easy to see that the RA and QA Mad2 mutants will also be unable to promote this step. The hypothetical reaction shown in this panel has the character of a positive feedback loop.