Phosphorylation Regulates the p31^Comet-Mitotic Arrest-deficient 2 (Mad2) Interaction to Promote Spindle Assembly Checkpoint (SAC) Activity

Background: p31^Comet antagonizes the SAC effector Mad2 to promote mitotic progression.

Results: p31^Comet phosphorylation modulates binding to Mad2.

Conclusion: Phosphorylation attenuates binding of Mad2 by p31^Comet to promote SAC activity.

Significance: Our study provides the first description of p31^Comet regulation and new insight into the control of SAC activity.

Abstract

The spindle assembly checkpoint (SAC) ensures the faithful segregation of the genome during mitosis by ensuring that sister chromosomes form bipolar attachments with microtubules of the mitotic spindle. p31^Comet is an antagonist of the SAC effector Mad2 and promotes silencing of the SAC and mitotic progression. However, p31^Comet interacts with Mad2 throughout the cell cycle. We show that p31^Comet binds Mad2 solely in an inhibitory manner. We demonstrate that attenuating the affinity of p31^Comet for Mad2 by phosphorylation promotes SAC activity in mitosis. Specifically, phosphorylation of Ser-102 weakens p31^Comet-Mad2 binding and enhances p31^Comet-mediated bypass of the SAC. Our results provide the first evidence for regulation of p31^Comet and demonstrate a previously unknown event controlling SAC activity.

Introduction

The spindle assembly checkpoint (SAC)² is an evolutionarily conserved and essential surveillance mechanism that ensures that chromosome segregation during mitosis proceeds with high fidelity. The SAC monitors the attachment of chromosomes to microtubules of the mitotic spindle and mediates inhibition of the anaphase-promoting complex (APC) ubiquitin ligase. SAC activity persists until each kinetochore of the sister chromosome pairs have achieved attachment to microtubules emanating form opposite poles of the spindle and are under tension. Kinetochores that are unattached or that are not under tension recruit SAC components that ultimately lead to the activation of the effector molecules Mad2 and BubR1. Upon activation, Mad2 and BubR1, along with Bub3, bind the APC activator/adapter protein Cdc20 to form the mitotic checkpoint complex (MCC), resulting in inhibition of APC^Cdc20 activity (1).

The Mad2 protein exists in two conformations: closed C-Mad2 (active) and open O-Mad2 (inactive) (2–10). Activation of the checkpoint recruits the C-Mad2/Mad1 heterodimer to the kinetochore. C-Mad2, in this complex, then dimerizes with cytoplasmic Mad2 to catalyze the formation of additional C-Mad2 molecules that directly bind Cdc20. Binding of Cdc20 by Mad2 primes it for binding by BubR1/Bub3 to form the MCC and prevent mitotic progression (11).

Once all kinetochores have achieved proper attachment, the checkpoint is satisfied and mitotic progression resumes. Recovery from SAC activity occurs in two steps: termination of MCC formation and disassembly of preexisting MCCs. p31^Comet is a key factor in the recovery of cells from SAC activity (12–16). p31^Comet specifically binds C-Mad2 via the same binding interface as O-Mad2 (9, 12, 17, 18). p31^Comet silences SAC activity in vitro and in vivo in a manner that requires this interaction with Mad2 (18). p31^Comet interacts with C-Mad2 in both the MCC and the C-Mad2·Mad1 complex and has been demonstrated to antagonize both (9, 12, 17–19).

We and others (20–23) have shown that APC^Cdc20 remains inhibited in extracts generated from SAC-active cells, despite the presence of p31^Comet in these extracts. These data suggest that p31^Comet is activated upon satisfaction of the SAC to facilitate recovery from the checkpoint. Consistent with this idea, the presence of p31^Comet in Cdc20 immunocomplexes has been shown to increase upon release from nocodazole (i.e. during SAC recovery) (12). However, p31^Comet and Mad2 interact throughout the cell cycle, including early mitosis when the SAC is active (15, 16, 19).

We have examined the mechanism that permits SAC activity despite the interaction of a key effector molecule with its antagonist. We show that p31^Comet interacts with Mad2 solely as an inhibitor. Moreover, we demonstrate that although the two proteins can be found in complex through out the cell cycle, phosphorylation of p31^Comet during mitosis weakens the interaction with Mad2, as judged by competition assays, to allow SAC activity.

EXPERIMENTAL PROCEDURES

Cell Culture

HeLa and HCT116 cells were obtained from ATCC and maintained in DMEM supplemented with 10% FBS. Cells were synchronized as described (20). Nocodazole was added 5 h after release from thymidine. Cells were transfected with Transit-LT1 (Mirus) or RNAiMax (Invitrogen) per manufacturer's instructions. Where indicated, cells were treated with 100 ng/ml nocodazole. HeLa Flp-In T-Rex cells were a gift of Dr. Stephen Taylor (15). Isogenic p31^Comet WT and S102A cell lines were generated with the use of pGLAP2 as described (24).

Antibodies

Anti-Myc (9E10) was produced at the Lerner Research Insitute. Commercial antibodies were as follows: cyclin B (GNS1), p31^Comet (4RE23), and Mad2 (107–276-3) from Santa Cruz Biotechnology; actin (AC-15), FLAG (M2), and p31^Comet (4G11) from Sigma; p31^Comet (EPR9584) from Abcam.

Plasmids and Recombinant Proteins

A gateway-compatible open reading frame for p31^Comet was obtained from Origene. Subcloning was performed using LR clonase to transfer the ORF into pCS2, pGLAP2, and pGex6P1-derived destination vectors. Mutagenesis was performed using the QuikChange mutagenesis strategy. His₆-p31^Comet Δ35 produced as described (20).

Western Blotting, Immunoprecipitation, and Phos-tag Analysis

Cell extracts were generated in EBC buffer (50 mm Tris (pH 8.0), 120 mm NaCl, 0.5% Nonidet P-40, 1 mm DTT, 25 mm β-glycerophosphate, 5 mm NaF, 1 mm NaVO₄, and leupeptin, pepstatin, and chymotrypsin, each at 10 μg/ml. For immunoprecipitation, equal amounts of cell lysates were incubated with the indicated antibodies for 2–12 h and washed in EBC buffer, including inhibitors. Immunoprecipitation samples or equal amounts of whole cell lysates were resolved by SDS-PAGE. For analysis of phosphorylation, proteins were resolved on 8% SDS-PAGE gels incorporating 40 μm Phos-tag reagent (Wako) and 80 μm MnCl₂. Proteins were transferred to PVDF membranes (Millipore) probed with the indicated antibodies and visualized with the LI-COR Odyssey near infrared imaging system.

In Vitro Phosphorylation and Mass Spectrometry

HCT116 cells were synchronized and harvested in mitosis after a thymidine-nocodazole block or in late M/early G₁, 2 h after release from a thymidine-nocodazole block. Extracts were then prepared by resuspension in extract buffer (20 mm Tris-HCl, pH 7.2, 2 mm DTT, 0.25 mm EDTA, 5 mm KCl, 5 mm MgCl₂) followed by two rounds of freeze-thaw and passage through a needle. Extracts were supplemented with ATP and an energy-regenerating system. Mitotic extracts were supplemented with non-destructible cyclin B1 and MG132 to maintain the mitotic state. GST-p31^Comet was then incubated in extract for 1 h at 30 °C and then captured on glutathione-agarose. After washing, the proteins were resolved on Phos-tag gels as above and visualized with Gelcode Blue (Pierce). Protein bands were digested with trypsin, extracted in 50% acetonitrile and 5% formic acid. After evaporation, peptides were resuspended in 1% acetic acid and analyzed on a Finnigan LTQ-Obitrap Elite hybrid mass spectrometer system. The HPLC column was a Dionex (15 cm, × 75 μm inner diameter) Acclaim Pepmap C18 (2 μm) 100 Å reversed phase capillary chromatography column.

Live Cell Microscopy

Isogenic HeLa cell lines were plated in 24-well dishes and synchronized as above. Expression of p31^Comet proteins was induced after release from thymidine. Eight hours post-release, nocodazole was added, and cells were transferred to Leica DMIRB inverted microscope equipped with a Roper Scientific CoolSNAP HQ Cooled CCD camera (Roper Scientific, Tucson, AZ), temperature controller (37 °C) and CO₂ (5%) incubation chamber (Leica Microsystems GmbH), and a PeCon incubator (PeCon GmbH, Erbach, Germany) controlled by MetaMorph Software (Molecular Devices, Downingtown, PA). Images of multiple fields per well were collected at 10-min intervals and analyzed using ImageJ and Photoshop.

Statistical Analysis

One-way analysis of variance with Bonferroni post tests were performed using GraphPad Prism.

RESULTS

p31^Comet and Mad2 Interact throughout the Cell Cycle

APC^Cdc20 remains inhibited by the MCC in extracts derived from mitotic, SAC-active cells, despite the presence of p31^Comet in these extracts (20–22). In addition, association of p31^Comet with Cdc20 increases during release from a nocodazole block (12). Together, these data suggest that an activating event promotes the interaction of p31^Comet with Mad2 followed by subsequent silencing of SAC activity. To identify this event, we sought to confirm that the binding of p31^Comet to Mad2 follows this model, that is, that p31^Comet binds and antagonizes Mad2 during mitotic exit. To analyze the interaction of p31^Comet and Mad2, we examined both p31^Comet and Mad2 immunocomplexes for the presence of the interacting partner as well as the Mad2 interactor Mad1. Unexpectedly, we found that p31^Comet interacted with Mad2 as well as the Mad2 heterodimeric partner Mad1 at all stages of the cell cycle (Fig. 1A). Notably, p31^Comet associates with Mad2 and Mad1 in cells arrested in mitosis by treatment with nocodazole, in which the SAC is maximally active and the interaction does not increase upon mitotic exit (Fig. 1, A and B). Several recent reports have confirmed our findings (15, 16, 19). These data pose a critical question: how does a cell mount a SAC-mediated mitotic arrest, whereas the critical SAC effector Mad2 is readily found in complex with its inhibitor p31^Comet?

FIGURE 1.

p31^Comet and Mad2 interact throughout the cell cycle. A, HCT116 cells were synchronized by a double thymidine block and released into nocodazole (Noc)-containing medium. Cells in S, G₂, and M phases were harvested at 2, 8, and 14 h, respectively. Lysates and p31^Comet or Mad2 immunoprecipitates (IP) were resolved by SDS-PAGE and blotted with the indicated antibodies. Cyclins A and B were probed as markers for the indicated cell cycle phases. B, HCT116 cells were synchronized in mitosis by thymidine-nocodazole block and released (Rel.) into fresh medium. Cells harvested at the indicated time points were resolved as in A. Cyclin B levels were probed to monitor the efficiency of mitotic exit. C, asynchronous lysates were immunoprecipitated with control IgGs or the indicated antibodies and analyzed as described in A.

p31^Comet Associates with Mad2 via a Single Interface

Analysis of the p31^Comet ΔN35-C-Mad2 crystal structure demonstrated that p31^Comet and O-Mad2 bind C-Mad2 via the same interface (18). Given this knowledge, the function of p31^Comet, and the interaction of p31^Comet with Mad2 throughout the cell cycle, we propose a model for the regulated inhibition of Mad2 by p31^Comet (Fig. 2). This model assumes that p31^Comet interacts with the C-Mad2:Mad1 heterodimer solely via the O-Mad2 binding interface of C-Mad2. To confirm the validity of this assumption, we asked whether the p31^Comet Q83A/F191A mutant was capable of binding to the C-Mad2:Mad1 heterodimer. Consistent with its inability to interact with C-Mad2, p31^Comet Q83A/F191A did not co-precipitate Mad1 (Fig. 3A). However, as no other direct p31^Comet-interacting proteins are known, it remained possible that the Q83A/F191A mutant is deficient in interactions that are independent of the O-Mad2 interface. Given the crystal structure data, we reasoned that such additional p31^Comet-Mad2 interactions would be mediated by the N terminus of p31^Comet, which is not present in the structure. p31^Comet deleted for the N-terminal 50 amino acids (Δ50) showed no defect in interacting with Mad2 or Mad1 (Fig. 3B). To exclude the possibility that the interaction with the C-Mad2:Mad1 heterodimer is mediated by a p31^Comet-Mad1 interaction not predicted by structural modeling, we determined that p31^Comet is not able to interact with Mad1 in the absence of Mad2 (Fig. 3C). Finally, we confirmed that immunoprecipitation of Mad2 using an antibody that specifically binds C-Mad2 via the O-Mad2/p31^Comet binding surface (see Ref. 16) does not co-precipitate p31^Comet as would be expected if an additional binding interface exists (Fig. 3D). Taken together, these data indicate that p31^Comet interacts with C-Mad2 and the C-Mad2:Mad1 heterodimer solely via the O-Mad2 binding interface. We conclude that p31^Comet associates with Mad2 in an inhibitory manner throughout the cell cycle.

FIGURE 2.

Models of p31^Comet-Mad2 interaction. Apparent p31^Comet Activity, during G₂/interphase, p31^Comet may actively prevent the generation of low levels of C-Mad2 (15). p31^Comet activity is low during prometaphase to allow SAC activity. Upon attachment of all chromosomes at metaphase, p31^Comet binds active Mad2, preventing further Mad2 activation and silences the SAC. Regulated Affinity Model, our proposed model. The affinity of p31^Comet for Mad2 is diminished in early mitosis, due to post-translational modification(s), to promote SAC activity. Loss of early mitotic modifications and/or late mitosis-specific modifications enhance p31^Comet affinity for Mad2 to promote SAC silencing.

FIGURE 3.

p31^Comet only binds Mad2 as an inhibitor. A, HeLa cells were transfected with FLAG-tagged p31^Comet or the Mad2-binding mutant Q83A/F191A. Lysates and FLAG-immunoprecipitates (IP) were probed for the presence of Mad2 and Mad1. B, cells were transfected with FLAG-p31^Comet or a mutant lacking the N-terminal 50 residues (Δ50) and analyzed as described in A. C, cells were transfected with siRNAs targeting Mad2. Lysates and p31^Comet-IPs were analyzed as described in A. D, cell lysates were immunoprecipitated with antibodies recognizing total Mad2 or C-Mad2 and probed for the presence of Mad1 and p31^Comet.

The p31^Comet-Mad2 Interaction Is Weakened in Early Mitosis

We next tested our model, which predicts that the interaction of p31^Comet with Mad2 is weakened in mitosis (Fig. 2). We utilized an in vitro capture-chase assay. Recombinant GST-Mad2 bound to glutathione-agarose beads was incubated with extracts from cells that were SAC-active (nocodazole-arrested) or SAC-inactive (late mitotic/early G₁, nocodazole-released). We then tested the relative strength of p31^Comet-Mad2 binding by incubating the bead-bound complexes with increasing amounts of a recombinant p31^Comet with the N-terminal 35 amino acids deleted, p31^CometΔ35. After washing, we probed the bound complexes for the presence of the full-length, captured p31^Comet with an N-terminal specific-antibody. As predicted by model 3, p31^Comet captured from mitotic extracts was more readily displaced from Mad2 than the p31^Comet from SAC-inactive extracts (Fig. 4A). The altered p31^Comet affinity depicted in Fig. 2 is the likely result of a post-translational modification. We therefore tested whether phosphorylation affected the affinity of p31^Comet for Mad2. We determined that p31^Comet is phosphorylated in mitosis by comparing the migration of p31^Comet in extracts of S, G₂, or mitosis on Phos-tag gels (25). These analyses indicated that a substantial portion of p31^Comet is phosphorylated during mitosis as evidenced by diminished electrophoretic migration (Fig. 4B). We confirmed that this altered mobility was indeed due to phosphorylation by treating mitotic lysates with λ-phosphatase prior to Phos-tag electrophoresis. Phophatase treatment resulted in a loss of the slower migrating forms of p31^Comet, confirming that p31^Comet is phosphorylated in mitosis (Fig. 4C).

FIGURE 4.

Phosphorylation of p31^Comet weakens the interaction with Mad2 during mitosis. A, left panel, p31^Comet was captured by recombinant Mad2 from extracts of nocodazole-arrested (M) or nocodazole-released (M/G₁) cells. Mad2-p31^Comet complexes were incubated with increasing amounts of recombinant p31^Comet Δ35 (N-terminal deletion) and retention of cellular p31^Comet was determined. Middle panel, quantification of the percent of captured p31^Comet that was displaced by Δ35. Right panel, cyclin B was probed in the same extracts to confirm mitotic status. B, HCT116 cells were synchronized as in 1A and lysates were resolved by Phos-tag SDS-PAGE to examine the phosphorylation status of p31^Comet. C, mitotic lysates were treated with or without λ-phosphatase and resolved by Phos-tag SDS-PAGE. D, p31^Comet was captured from asynchronous lysates by C-Mad2 in complex with GST-Cdc20^111–138. The relative strength of the interaction was tested by challenge with purified Mad2 (predominantly O-Mad2). Bound proteins were resolved by Phos-tag SDS-PAGE. E, FLAG-p31^Comet was phosphorylated in mitotic extract and then captured, challenged, and analyzed as described in D. Right panel, the mean and S.E. for the percent of p31^Comet displaced by increasing amounts of Mad2 was quantified (n = 3). The effect of phosphorylation on displacement of p31^Comet by Mad2 was analyzed by linear regression.

To further determine that phosphorylation weakened the interaction between p31^Comet and Mad2, we utilized Phos-tag to analyze capture-chase assays similar to those above. To more closely recapitulate the physiological competition between p31^Comet and O-Mad2 (or other C-Mad2 binding partners, e.g. BubR1) for C-Mad2, we captured p31^Comet from extract with C-Mad2 bound to the Mad2-binding region of Cdc20 (GST-Cdc20^111–138) and then challenged binding with the monomeric Mad2 (O-Mad2) (5, 17). We first captured endogenous p31^Comet from asynchronous extracts, which contained a minor but readily detectable mitotic phosphorylated form. When challenged with increasing molar ratio of O-Mad/C-Mad2, the phosphorylated form was lost more readily (Fig. 4D). To more closely examine the situation in mitosis, we phosphorylated a FLAG-p31^Comet in mitotic extracts and allowed a molar excess of this protein to bind the GST-Cdc20^111–138 and then performed the chase assays as above. The phosphorylated form was more readily displaced once the O-Mad2/C-Mad2 increased over 1:1 (Fig. 4E). These data indicate that phosphorylation lowers the affinity of p31^Comet for Mad2 during mitosis.

Phosphorylation of Ser-102 Regulates the Mad2 Interaction

We next sought to identify the mitosis-specific phosphorylation event(s) that mediate the weakened binding to Mad2. To facilitate the purification of p31^Comet in quantities sufficient for these studies and circumvent issues with obtaining mitotic, SAC-active cells upon ectopic expression of p31^Comet, we phosphorylated recombinant p31^Comet in mitotic extracts. Recombinant GST-p31^Comet was incubated in SAC-active or SAC-inactive extracts, purified, and resolved on Phos-tag gels to facilitate analysis of the phosphorylated protein (Fig. 5A). Mass spectrometry analysis of phosphorylated p31^Comet band identified phosphorylation of Ser-102 in the slower migrating band, whereas the unmodified peptide was identified in the lower, faster migrating band (Fig. 5, B and C).

FIGURE 5.

p31^Comet Ser-102 is phosphorylated in mitosis. A, recombinant p31^Comet was incubated in mitotic (M) or G₁ extracts and resolved by Phos-tag SDS-PAGE. Cyclin B levels were monitored to confirm the mitotic status of the extract. B and C, LC-MS/MS analysis of the p31^Comet tryptic digests from A identified the ¹⁰⁰KPSPQAEEMoLK¹¹⁰ peptide in both the unmodified and Ser-102 phosphorylated forms. The CID spectra for the unmodified peptide (B) contains several sequence-specific ions, including the y9 and y8 ions whose mass difference is 87 Da, consistent with an unmodified Ser residue. Although the CID spectra of the modified peptide (C) looks similar to the unmodified form, there are several differences including the presence of several H₃PO₄ neutral loss peaks, and the mass difference between y9 and y8 ions is 167 Da, consistent with Ser-102 as the site of phosphorylation.

p31^Comet Ser-102 Phosphorylation Status Modulates SAC Activity in Vivo

Phosphorylation of p31^Comet Ser-102 is a strong candidate for modulating p31^Comet-Mad2 binding. Ser-102 lies in a flexible loop adjacent to the Mad2-interacting residue Gln-83 and is thus physically poised to regulate this interaction (18). To examine the importance of Ser-102 phosphorylation in vivo, we constructed isogenic HeLa cells harboring tetracycline-inducible wild type p31^Comet or the S102A mutant (Fig. 6A). We first tested whether phosphorylation of Ser-102 is required for the decreased electrophoretic mobility we observed for endogenous p31^Comet in mitosis (Fig. 4C). Overexpression of p31^Comet abrogates Mad2-dependent mitotic arrest. To facilitate the accumulation of exogenous p31^Comet-expressing cells in mitosis that is required for analyzing phosphorylation in intact cells, we induced p31^Comet proteins in cells overexpressing Mad2 to buffer the effects of increased p31^Comet levels and then treated the cells with nocodazole. Mutation of Ser-102 prevented the mitotic phosphorylation shift of p31^Comet (Fig. 6B).

FIGURE 6.

Phosphorylation of Ser-102 attenuates the antagonism of Mad2 by p31^Cometin vivo. A, expression of p31^Comet wild type (WT) or S102A mutant proteins was induced with doxycycline (Dox) in isogenic HeLa cell populations. B, cells from A were transfected with Mad2-encoding constructs and arrested in mitosis by a thymidine-nocodazole protocol. Mitotic cells were collected by shake-off and exogenous p31^Comet proteins were analyzed by Phos-tag SDS-PAGE and probed for the FLAG epitope. C, cells were treated as described in B, and mitotic index was determined. The mitotic index of uninduced cells was set at 100%. Error bars represent S.D. (n = 3). **, p < 0.01; ***, p < 0.001. D and E, cells were synchronized by thymidine-nocodazole block. Mitotic duration was analyzed by time-lapse microscopy. D, representative images of mitotic progression in the indicated cell populations. The time (in min) relative to mitotic entry is indicated. E, plot showing the cumulative number of cells in mitosis at the indicated times for each cell population. CTRL, control.

In conjunction with our data above, these results suggest that phosphorylation of Ser-102 alters the affinity of p31^Comet for Mad2 in mitosis. Thus, we postulated that the increased affinity of p31^Comet S102A for Mad2 would enhance its ability to silence the SAC. We tested the ability of exogenous p31^Comet in these cells to silence nocodazole-induced activation of the SAC in asynchronous cells. These analyses revealed a modest increase in SAC silencing by the p31^Comet S102A relative to wild type-expressing cells (data not shown). We repeated these analyses in synchronized cells. p31^Comet expression was induced (or not) upon release from a thymidine block. Cells were then challenged with nocodazole and the mitotic index was determined. Whereas control cells efficiently accumulated in mitosis, expression of exogenous p31^Comet proteins resulted in a dramatic decrease in the mitotic index. Importantly, expression of S102A produced a significantly stronger bypass of SAC activity in comparison with wild type (Fig. 6C).

Our experiments thus far indicate that phosphorylation of Ser-102 does play a role in diminishing the antagonism of Mad2 by p31^Comet during mitosis. Because phosphorylated p31^Comet still binds Mad2, overexpression of these proteins should bypass Mad2-mediated arrests, as has been shown for overexpression of wild type p31^Comet. Under these conditions, a single time point may not be sufficient to fully reveal the importance of phosphorylating Ser-102. We thus repeated the experiments above and monitored the duration of mitotic arrest in induced or uninduced cells in real time. Consistent with our above results, these analyses confirmed that exogenous p31^Comet proteins are capable of bypassing Mad2-mediated arrest. However, analysis of mitotic duration revealed that failure to phosphorylate Ser-102 enhances the ability of p31^Comet to silence SAC activity (Fig. 6, D and E). Together, our results indicate that phosphorylation of p31^Comet Ser-102 is an important event in the regulation of Mad2 activity in vivo.

DISCUSSION

Recent studies from a number of laboratories have highlighted the importance of mechanisms that promote recovery from the activation of the SAC, which occurs in every cell cycle. Although several proteins and events have been shown to promote recovery from SAC-mediated mitotic arrest, the regulation of these processes is unknown. p31^Comet is the best characterized component of the SAC recovery machinery (12–16). p31^Comet is required for efficient progression beyond metaphase and timely resumption of mitotic progression upon release from spindle poisons such as nocodazole.

Recovery from SAC activity requires two steps: 1) termination of SAC signaling and MCC formation and 2) disassembly of preexisting MCCs. p31^Comet specifically binds C-Mad2 via the same binding interface as O-Mad2 (9, 12, 17, 18). p31^Comet silences SAC activity in vitro and in vivo in a manner that requires this interaction with Mad2 (18). p31^Comet interacts with C-Mad2 in both the MCC and the C-Mad2·Mad1 complex, and it has been suggested that it may participate in both steps of SAC inactivation (9, 12, 17–19). However, it is currently unclear which step p31^Comet participates in in vivo as evidence for and against both models has been presented.

p31^Comet is coordinately expressed with Mad2 during the cell cycle.³ Whereas Mad2 is required in early mitosis, p31^Comet is required for efficient progression beyond metaphase, suggesting that it becomes active to inhibit Mad2 at this point. Consistent with this idea, APC^Cdc20 remains inhibited in extracts derived from cells arrested in prometaphase by nocodazole treatment, despite the presence of p31^Comet. We have shown that this inhibition is mediated by the MCC, suggesting that p31^Comet is not active in prometaphase (20). However, p31^Comet interacts with Mad2 throughout the cell cycle. We show that this binding reflects inhibition of Mad2. Thus, regardless of which step(s) of SAC silencing p31^Comet promotes, how the SAC is activated despite inhibition of Mad2, is a critical question.

We have determined that although the interaction of p31^Comet with Mad2 is readily detectable when the SAC is active, the binding between these two proteins is weakened in these cells. Phosphorylation of p31^Comet mediates this weakened interaction. Specifically, we identified phosphorylation of Ser-102 in the mitosis-specific phosphorylated form of p31^Comet. Ser-102 is positioned near a key residue, Gln-83, in the p31^Comet-Mad2 interaction. To address the impact of phosphorylating p31^Comet Ser-102 on its interaction with Mad2, we utilized competitive binding assays in vitro. These analyses revealed that phosphorylation of Ser-102 is sufficient for inducing a subtle but significant weakening of the binding of p31^Comet and Mad2 in vitro. Importantly, preventing phosphorylation of Ser-102 enhances the ability of exogenous p31^Comet to antagonize Mad2 in vivo. Given the dynamic nature of SAC activation in vivo (13, 19, 21) and the potential involvement of other factors and modifications (e.g. Mad2 phosphorylation) in attenuating the p31^Comet-Mad2 interaction, it is likely that our in vitro assay may not fully recapitulate the impact of phosphorylating p31^Comet on Mad2 activation in an intact mitotic cell. Together, our data suggest a model of SAC regulation based on modulating the strength of the interaction of p31^Comet with Mad2. Although our data strongly support Ser-102 as a key regulatory site, we note that this residue, as with many reported phosphosites in human p31^Comet, is not highly conserved in p31^Comet from other species. However, we note that these loop regions generally contain potential phosphosites, suggesting that the mechanism is conserved.

Attenuating the interaction of p31^Comet with C-Mad2 would enhance the ability of O-Mad2 and/or BubR1 to compete with p31^Comet for binding to C-Mad2 at the kinetochore or in the MCC, respectively, promoting SAC activity (15, 16, 18, 26). Additionally, attenuating constant inhibition of C-Mad2, rather than activating inhibition of Mad2 at metaphase, would ensure rapid APC^Cdc20 activation and progression from metaphase to anaphase upon satisfaction of the SAC. However, our analyses indicate that the entire pool of p31^Comet is not phosphorylated during mitosis, even in nocodazole-treated cells in which the SAC is maximally active. This finding is in line with data showing that p31^Comet functions during prometaphase to prevent maximal SAC activation (19). In addition, these findings suggest that antagonistic kinase and phosphatase activities are required for regulation of p31^Comet in mitosis. It will be important to identify these enzymes to fully understand the importance of this p31^Comet phosphoregulation. Future studies will pursue these enzymes and compare the interplay of p31^Comet phosphorylation with additional events, including phosphorylation of critical interacting proteins such as Mad2 or Cdc20, which are known to contribute to regulation of the SAC (27, 28).