- Email: [email protected]
Mps1 Phosphorylation by MAP Kinase Is Required for Kinetochore Localization of Spindle-Checkpoint Proteins
Current Biology 16, 1764–1769, September 5, 2006 ª2006 Elsevier Ltd All rights reserved
DOI 10.1016/j.cub.2006.07.058
Report Mps1 Phosphorylation by MAP Kinase Is Required for Kinetochore Localization of Spindle-Checkpoint Proteins Yong Zhao1 and Rey-Huei Chen2,* 1 Graduate Field of Biochemistry Molecular and Cell Biology Cornell University Ithaca, New York 14853 2 Institute of Molecular Biology Academia Sinica Taipei 11529 Taiwan
Summary The spindle checkpoint delays anaphase onset until all chromosomes have achieved bipolar attachment to the spindle microtubules. Unattached kinetochores activate the spindle checkpoint by recruiting several spindle-checkpoint proteins, including Mps1, Mad1, Mad2, Bub1, Bub3, and BubR1 (Mad3 in yeast) [1]. In vertebrate cells, active MAP kinase (MAPK) is also enriched at unattached kinetochores and is required for the spindle checkpoint [2–5]. It has been shown that the kinase activity of Mps1 is required for the spindle checkpoint [6] and for kinetochore localization of Bub1, Bub3, Mad1, and Mad2 [6, 7]. We herein demonstrate that MAPK phosphorylates Mps1 at S844 in Xenopus egg extracts. Interestingly, changing S844 to unphosphorylatable alanine (S844A) has no effect on the kinase activity of Mps1, although it abolishes the checkpoint function of Mps1. Biochemical and immunofluorescence studies show that S844A mutation perturbs kinetochore localization of Mps1 and other spindle-checkpoint proteins, whereas the phosphorylation-mimicking S844D mutant restores their functions. Our studies suggest that Mps1 phosphorylation by MAPK at S844 might create a phosphoepitope that allows Mps1 to interact with kinetochores. In addition, our results indicate that active Mps1 must localize to kinetochores in order to execute its checkpoint function. Results and Discussion MAPK Contributes to Mps1 Phosphorylation Mps1 is a dual-specificity protein kinase that functions in spindle-pole-body duplication and the spindle checkpoint [8–11]. The kinase activity of Mps1 is required for the spindle checkpoint [6] and targets Mad1 in budding yeast [12]. Mps1 is a phosphoprotein [10, 13], and phosphorylation of human Mps1 is regulated by the cell cycle [10, 13]. However, it is not clear how phosphorylation might regulate the kinase activity and the checkpoint function of Mps1. We first examined Mps1 in the egg extracts from the frog Xenopus laevis. Western blotting of Mps1 showed a slower gel mobility of the protein in
*Correspondence: [email protected]
metaphase and spindle-checkpoint-active extracts than in interphase extracts (Figure 1A, lanes 1–3), indicating that Mps1 was differentially modified at interphase and M phase. These proteins were indeed Mps1 because immunodepletion with Mps1 antibodies abolished the signal (Figure 1B, lane 2). The proteins were restored after the addition of in-vitro-synthesized RNA corresponding to the wild-type or kinase-dead Mps1 (Figure 1B, lanes 3 and 4). Both wild-type and kinase-dead Mps1 showed similar gel mobility (Figure 1B, lanes 3 and 4) that increased upon phosphatase treatment (Figure 1B, lanes 5 and 6). The result suggests that Mps1 in egg extracts was phosphorylated mostly by some other kinase(s), rather than through autophosphorylation. Interestingly, adding the phosphatase inhibitor microcystin to the samples as the control experiment resulted in a further mobility retardation of wild-type Mps1 but not the kinase-dead protein (Figure 1B, lanes 7 and 8), suggesting that autophosphorylation of Mps1 in egg extracts was balanced by a phosphatase activity. The in vitro immunocomplex kinase assay of Mps1 from various egg extracts showed similar levels of autophosphorylation activities (Figure 1A, lanes 4–6). The autophosphorylated Mps1 proteins migrated approximately as 160 kDa proteins that were indeed Mps1 because we were able to reimmunoprecipitate the denatured proteins with Mps1 antibodies (data not shown). The result suggests that the increased phosphorylation of Mps1 at M phase had no significant effect on the kinase activity of Mps1 in the egg extract. In addition, Western blotting of the in vitro autophosphorylation reactions showed an extensive mobility shift of the wild-type, but not the kinase-dead, Mps1 (Figure 1B, lanes 9 and 10), indicating that purified Mps1 was readily autophosphorylated in vitro. Xenopus Mps1 protein contains a number of serine/ threonine-proline sequences that are potential targets for Cdc2 and MAPK. Active MAPK is enriched at kinetochores of misaligned chromosomes and is dissociated from kinetochores after mid-anaphase [4, 5]. The spindle checkpoint requires active MAPK in Xenopus egg extracts and tissue-cultured cells [2, 3]. To determine whether Mps1 might be a target for MAPK, we examined the effect of U0126, an inhibitor for MAPK kinase (MEK). In U0126-treated metaphase egg extracts, Mps1 appeared in more-rapidly migrating species in comparison with that in control extracts, indicating reduced phosphorylation (Figure 1C) and suggesting that MAPK contributes to Mps1 phosphorylation. We performed two-dimensional tryptic-phosphopeptide mapping to determine the phosphorylation sites. We used kinase-dead Mps1 for the analysis in order to avoid any background phosphorylation generated through autophosphorylation. We first compared the maps of the proteins labeled with 32P in interphase and metaphase extracts. The map from the metaphase sample revealed a major phosphopeptide that was absent in the interphase sample, whereas most of the other phosphopeptides
MAPK-Dependent Kinetochore Targeting of Mps1 1765
Figure 1. MAPK Contributes to Mps1 Phosphorylation (A) Mps1 is differentially phosphorylated during the cell cycle. Mps1 immunoblotting (lanes 1–3) and in vitro Mps1-autophosphorylation reactions (lanes 4–6) were performed with interphase (lanes 1 and 4), metaphase (lanes 2 and 5), and spindle-checkpointactive (lanes 3 and 6) Xenopus egg extracts. An autoradiography of in-vitro-phosphorylated Mps1 is shown (lanes 3–6). The migration of molecular-size standards is indicated on the right. (B) Both wild-type and kinase-dead Mps1 are phosphorylated. Western blotting of Mps1 was performed with CSF-arrested extract (lane 1), Mps1-depleted extract (lane 2), and depleted extract supplemented with wildtype (lanes 3) or kinase-dead (lane 4) Mps1. Extracts containing wild-type or kinasedead Mps1 were treated with l protein phosphatase (lanes 5 and 6), treated with microcystin (lanes 7 and 8), or subjected to in vitro autophosphorylation reactions (lanes 9 and 10) prior to the Western blot analysis. (C) Inhibition of MAPK activity reduces Mps1 phosphorylation at M phase. Metaphase extracts were treated with DMSO (lane 1) or U0126 (lane 2). The samples were immunoblotted with anti-Mps1 (upper panel), antiMAPK (middle panel), and anti-phosphorylated MAPK (lower panel) antibodies. (D) Two-dimensional tryptic-phosphopeptide mapping of Mps1. Kinase-dead Mps1 was labeled with 32P in interphase or metaphase extracts in the absence or presence of U0126, as indicated. The map of kinasedead Mps1 that also contains the S844A mutation is shown in the lower right panel. (E) In vitro phosphorylation of Mps1 by MAPK. GST fusion of Mps1 (amino acids 756–882) (lanes 1 and 3) or the same region containing the S844 mutation (lanes 2 and 4) were subjected to in vitro phosphorylation with control (lanes 1 and 2) or MAPK (lanes 3 and 4) immunoprecipitations. An autoradiography of 32P incorporated into GST-Mps1 (upper panel) and the Coomassie Blue staining of the gel (lower panel) are shown.
were present in both interphase and metaphase extracts (Figure 1D). Interestingly, this mitotic phosphopeptide was greatly reduced when the extract was treated with U0126 (Figure 1D). The weak signal of the mitotic phosphopeptide in U0126-treated extracts might be due to residual MAPK activity in the extracts. These results indicate that the major mitosis-specific phosphopeptide was likely generated by MAPK. Alignment of Xenopus, mouse, and human Mps1 protein sequences revealed eight evolutionarily conserved MAPK recognition sites in Xenopus Mps1: S283, S459, T476, T491, T506, S733, T753, and S844. In order to identify the phosphorylation site, we constructed eight mutants that each contained a single mutation of alanine or valine in place of the conserved serine or threonine, respectively. Phosphopeptide mapping of these eight mutants showed that S844A abolished the major mitosis-specific phosphopeptide (Figure 1D, lower right panel), suggesting that S844 was likely the major target for MAPK. To test whether Mps1 is a substrate for MAPK in vitro, we generated GST-fusion proteins with truncated Mps1 that contained amino acid 756 to the C terminus. In vitro phosphorylation reactions showed that MAPK readily phosphorylated the wild-type protein but not the S844A mutant (Figure 1E), suggesting that Mps1 indeed was a substrate for MAPK.
The Spindle Checkpoint Requires Phosphorylation at S844 We next determined whether phosphorylation at S844 of Mps1 was involved in the spindle-checkpoint function of Mps1. The egg extracts are normally arrested at metaphase by the cytostatic factor activity (CSF), and the arrest is released upon calcium addition. The spindle checkpoint is typically activated in the extract by incubation with a sufficient amount (9,000–15,000/ml extracts) of sperm nuclei and the microtubule inhibitor nocodazole. After the checkpoint is activated, the extract remains at M phase even after calcium addition, as indicated by sustained Cdc2 activities (Figure 2A, mock depletion). In Mps1-depleted extracts incubated with sperm nuclei and nocodazole, Cdc2 activities declined rapidly upon calcium addition, indicating that the spindle checkpoint was lost in the absence of Mps1 (Figure 2A). When wild-type Mps1 was added back to extracts depleted for the endogenous protein, Cdc2 activities were maintained (Figure 2A), indicating that the exogenously added Mps1 was functional. However, addition of either kinase-dead or S844A mutants failed to sustain Cdc2 activities (Figure 2A), indicative of a defective spindle checkpoint. On the other hand, the S844D mutant, with aspartic acid in place of serine to mimic phosphorylation, was able to support the spindle
Current Biology 1766
Figure 3. S844A Affects Binding of the Checkpoint Proteins to Unattached Chromosomes, but Not the Kinase Activity of Mps1 Figure 2. S844A Mutation Abolishes the Spindle-Checkpoint Function of Mps1 (A) Sperm nuclei were incubated with mock-depleted extracts or Mps1-depleted extracts supplemented with wild-type (WT), kinasedead (KD), S844A, or S844D Mps1 proteins as indicated on the left. Nocodazole was added to samples as indicated on the right. Calcium chloride was then added to override the CSF activity. Cdc2 activities in the samples, as determined by in vitro histone H1 phosphorylation, were measured at the times indicated after calcium chloride addition. Autoradiographies of histone H1 phosphorylation are shown. An immunoblot of Mps1 in the samples at time 0 is shown in (B).
checkpoint (Figure 2A). These results suggest that both the kinase activity and phosphorylation at S844 are essential for the spindle-checkpoint function of Mps1. In addition, extracts containing the S844D mutant in the absence of nocodazole did not maintain Cdc2 activities after calcium addition (Figure 2A), indicating that S844D protein was unable to constitutively activate the spindle checkpoint without spindle disruption. Because the kinase activity of Mps1 is critical for the spindle checkpoint [6], we tested whether S844A mutation might affect the kinase activity of Mps1. Metaphase extracts depleted of endogenous Mps1 were supplemented with wild-type, kinase-dead, or S844A-mutant Mps1 to the endogenous levels. Mps1 proteins were immunoprecipitated for in vitro autophosphorylation
(A) Wild-type Mps1 (lane 1), kinase-dead (lane 2), or S844A (lane 3)mutant proteins were immunoprecipitated from metaphase egg extracts, and in vitro autophosphorylation reactions followed. An autoradiography of 32P incorporated into Mps1 is shown (upper panel). An immunoblot of Mps1 proteins in the extracts is shown in the lower panel. (B) Sperm nuclei and nocodazole were incubated with egg extracts that were mock depleted (lanes 1 and 6) or both depleted for the endogenous Mps1 and supplemented with mock (lanes 2 and 7), wildtype (lanes 3 and 8), kinase-dead (lanes 4 and 9), or S844A (lanes 5 and 10) Mps1 proteins as indicated. Chromosomes in the samples were isolated by centrifugation through a sucrose cushion. The extracts (lanes 1–5) and chromosomal fractions (lanes 6–10) were immunoblotted for proteins indicated on the left.
reactions. As expected, kinase-dead Mps1 had no autophosphorylation activity (Figure 3A, lane 2). Interestingly, the activity of the S844A mutant was comparable to that of the wild-type protein (Figure 3A, compare lanes 1 and 3), indicating that phosphorylation at S844 is not required for the kinase activity of Mps1. This result suggests that the checkpoint defect of the S844A mutant was not due to a lack of the kinase activity. Phosphorylation at S844 Is Required for Kinetochore Localization of the Spindle-Checkpoint Proteins Because Mps1 is required for kinetochore localization of other spindle-checkpoint proteins [6, 7], we tested the
MAPK-Dependent Kinetochore Targeting of Mps1 1767
Figure 4. S844 Phosphorylation Is Required for Kinetochore Localization of Mps1 (A) Sperm nuclei and nocodazole were incubated with mock-depleted extracts or Mps1-depleted extracts supplemented with mock or various Mps1 proteins as indicated on the left. The chromosomes were fixed and isolated for immunofluorescence staining with Mps1 and Mad1 antibodies. (B) Extracts were treated with DMSO or U0126 as indicated on the left before incubation with sperm nuclei and nocodazole for immunofluorescence staining of Mps1 and Mad1. Scale bars represent 10 mm.
possibility that S844A mutation might affect this function. We purified the chromosomes from nocodazoletreated egg extracts containing various forms of Mps1 and then analyzed the spindle-checkpoint proteins by Western blotting. The result showed that Mps1 depletion caused a great reduction in the levels of BubR1, Bub1, Bub3, Mad1, and Mad2 in the chromosomal fractions (Figure 3B). Addition of wild-type Mps1, but not kinase-dead or S844A mutants, to the depleted extracts restored the levels of the checkpoint proteins as well as hyperphosphorylation of Bub1 and BubR1 in the chromosomal fractions (Figure 3B). This result indicates that both the kinase activity of Mps1 and phosphorylation at S844 are required for efficient
recruitment of the spindle-checkpoint proteins to unattached chromosomes. We were unable to detect Mps1 in the chromosomal fractions by Western blotting, probably because of a low level of Mps1 at kinetochores or a very dynamic interaction of the protein with kinetochores. In fact, fluorescence recovery after photobleaching (FRAP) analysis of GFP-Mps1 reveals a fast turnover rate at unattached kinetochores with a half-life of approximately 10 s [14]. We thus examined Mps1 localization by immunofluorescence staining of isolated chromosomes. We found that Mps1 and Mad1 antibodies costained kinetochores of chromosomes prepared from mock-depleted extracts treated with nocodazole (Figure 4A, top panels). Both
Current Biology 1768
Mps1 and Mad1 disappeared from kinetochores in Mps1-depleted extracts (Figure 4A). Adding back wildtype Mps1 or the S844D mutant, but not the S844A mutant, restored both anti-Mps1 and anti-Mad1 staining at kinetochores (Figure 4A). Consistent with the notion that MAPK phosphorylates S844 in Mps1, U0126 treatment diminished both Mps1 and Mad1 at kinetochores (Figure 4B). These results suggest that kinetochore targeting of Mps1 requires phosphorylation at S844 and that the spindle-checkpoint defect of the S844A mutant is due to its inability to associate with kinetochores. In addition, kinase-dead Mps1 was able to localize to kinetochores, although the staining level was lower than that of wild-type and S844D proteins (Figure 4A). The result indicates that the kinase activity of Mps1 is not absolutely required for its kinetochore targeting but that phosphorylation of some kinetochore or spindle-checkpoint proteins, or both, by Mps1 might in turn enhance kinetochore association of Mps1. In this study, we demonstrate that Mps1 in Xenopus egg extracts is phosphorylated differentially during the cell cycle, and these results are similar to previous findings in human cells [10, 13]. We show that the increased phosphorylation at M phase is dependent on MAPK. MAPK is required for the spindle checkpoint, and phosphorylation of Cdc20 by MAPK is required for Cdc20 to associate with spindle-checkpoint proteins [15]. In addition, MAPK phosphorylates and activates Bub1 at kinetochores, and this activity facilitates the spindle checkpoint in egg extracts [16]. We herein identify Mps1 as another target for MAPK in the spindle checkpoint. Together, these findings suggest that MAPK controls several major steps in the spindle checkpoint. We map the major MAPK-dependent phosphorylation site to S844 and demonstrate that phosphorylation at this site is essential for the spindle checkpoint. The unphosphorylatable S844A mutant is defective in the spindle checkpoint, whereas the phosphorylation-mimicking S844D mutant is functional. The S844A mutant retains the normal kinase activity but is unable to localize to the kinetochore. These results suggest that active Mps1 must associate with kinetochores in order to exert its function in the spindle checkpoint. It is possible that Mps1 might phosphorylate some kinetochore components in order for spindle-checkpoint proteins to bind. Alternatively, Mps1 might efficiently phosphorylate its substrate when both Mps1 and the substrate are brought to close proximity by the unattached kinetochore. In budding yeast, it has been shown that overexpression of Mps1 constitutively activates the spindle checkpoint in the absence of any kinetochore or spindle defect [12, 17, 18]. Overexpression of Mps1 might allow efficient phosphorylation of its substrates and thus bypass kinetochore targeting. Furthermore, the S844A mutant cannot localize to the kinetochore, suggesting that phosphorylation of S844 by MAPK might create some phosphoepitope that allows Mps1 to interact with kinetochore components. It has also been shown that Aurora-B/INCENP complex is required for kinetochore localization of Mps1 [7]. It remains to be determined if there is any interplay between Aurora-B/INCENP and MAPK. Upon docking at the kinetochore, Mps1 then phosphorylates other kinetochore or spindle-checkpoint proteins, or both, to recruit other spindle-checkpoint
proteins to kinetochores. The nature of the kinetochore component that associates with phosphorylated Mps1 and the substrates for Mps1 remains to be determined. Supplemental Data Supplemental Data include Experimental Procedures and can be found with this article online at http://www.current-biology.com/ cgi/content/full/16/17/1764/DC1/. Acknowledgments This work was supported by grants from the National Institutes of Health (USA), the David and Lucile Packard Foundation (USA), the National Science Council (Taiwan), and Academia Sinica (Taiwan). The authors declare that there is no financial conflict of interest related to this work. Received: May 16, 2006 Revised: July 1, 2006 Accepted: July 13, 2006 Published: September 5, 2006
References 1. Taylor, S.S., Scott, M.I., and Holland, A.J. (2004). The spindle checkpoint: A quality control mechanism which ensures accurate chromosome segregation. Chromosome Res. 12, 599–616. 2. Minshull, J., Sun, H., Tonks, N.K., and Murray, A.W. (1994). MAPkinase dependent mitotic feedback arrest in Xenopus egg extracts. Cell 79, 475–486. 3. Wang, X.M., Zhai, Y., and Ferrell, J.E., Jr. (1997). A role for mitogen-activated protein kinase in the spindle assembly checkpoint in XTC cells. J. Cell Biol. 137, 433–443. 4. Shapiro, P.S., Vaisberg, E., Hunt, A.J., Tolwinski, N.S., Whalen, A.M., McIntosh, J.R., and Ahn, N.G. (1998). Activation of the MKK/ERK pathway during somatic cell mitosis: Direct interactions of active ERK with kinetochores and regulation of the mitotic 3F3/2 phosphoantigen. J. Cell Biol. 142, 1533–1545. 5. Zecevic, M., Catling, A.D., Eblen, S.T., Renzi, L., Hittle, J.C., Yen, T.J., Gorbsky, G.J., and Weber, M.J. (1998). Active MAP kinase in mitosis: Localization at kinetochores and association with the motor protein CENP-E. J. Cell Biol. 142, 1547–1558. 6. Abrieu, A., Magnaghi-Jaulin, L., Kahana, J.A., Peter, M., Castro, A., Vigneron, S., Lorca, T., Cleveland, D.W., and Labbe, J. (2001). Mps1 is a kinetochore-associated kinase essential for the vertebrate mitotic checkpoint. Cell 106, 83–93. 7. Vigneron, S., Prieto, S., Bernis, C., Labbe, J.C., Castro, A., and Lorca, T. (2004). Kinetochore localization of spindle checkpoint proteins: Who controls whom? Mol. Biol. Cell 15, 4584–4596. 8. Lauze, E., Stoelcker, B., Luca, F.C., Weiss, E., Schutz, A.R., and Winey, M. (1995). Yeast spindle pole body duplication gene MPS1 encodes an essential dual specificity protein kinase. EMBO J. 14, 1655–1663. 9. Weiss, E., and Winey, M. (1996). The S. cerevisiae SPB duplication gene MPS1 is part of a mitotic checkpoint. J. Cell Biol. 132, 111–123. 10. Stucke, V.M., Sillje, H.H., Arnaud, L., and Nigg, E.A. (2002). Human Mps1 kinase is required for the spindle assembly checkpoint but not for centrosome duplication. EMBO J. 21, 1723– 1732. 11. Fisk, H.A., Mattison, C.P., and Winey, M. (2003). Human Mps1 protein kinase is required for centrosome duplication and normal mitotic progression. Proc. Natl. Acad. Sci. USA 100, 14875–14880. 12. Hardwick, K.G., Weiss, E., Luca, F.C., Winey, M., and Murray, A.W. (1996). Activation of the budding yeast spindle assembly checkpoint without mitotic spindle disruption. Science 273, 953–956. 13. Liu, S.T., Chan, G.K., Hittle, J.C., Fujii, G., Lees, E., and Yen, T.J. (2003). Human MPS1 kinase is required for mitotic arrest induced by the loss of CENP-E from kinetochores. Mol. Biol. Cell 14, 1638–1651.
MAPK-Dependent Kinetochore Targeting of Mps1 1769
14. Howell, B.J., Moree, B., Farrar, E.M., Stewart, S., Fang, G., and Salmon, E.D. (2004). Spindle checkpoint protein dynamics at kinetochores in living cells. Curr. Biol. 14, 953–964. 15. Chung, E., and Chen, R.H. (2003). Phosphorylation of Cdc20 is required for its inhibition by the spindle checkpoint. Nat. Cell Biol. 5, 748–753. 16. Chen, R.H. (2004). Phosphorylation and activation of Bub1 on unattached chromosomes facilitate the spindle checkpoint. EMBO J. 23, 3113–3121. 17. Fraschini, R., Beretta, A., Lucchini, G., and Piatti, S. (2001). Role of the kinetochore protein Ndc10 in mitotic checkpoint activation in Saccharomyces cerevisiae. Mol. Genet. Genomics 266, 115–125. 18. Poddar, A., Daniel, J.A., Daum, J.R., and Burke, D.J. (2004). Differential kinetochore requirements for establishment and maintenance of the spindle checkpoint are dependent on the mechanism of checkpoint activation in Saccharomyces cerevisiae. Cell Cycle 3, 197–204.
DOI 10.1016/j.cub.2006.07.058
Report Mps1 Phosphorylation by MAP Kinase Is Required for Kinetochore Localization of Spindle-Checkpoint Proteins Yong Zhao1 and Rey-Huei Chen2,* 1 Graduate Field of Biochemistry Molecular and Cell Biology Cornell University Ithaca, New York 14853 2 Institute of Molecular Biology Academia Sinica Taipei 11529 Taiwan
Summary The spindle checkpoint delays anaphase onset until all chromosomes have achieved bipolar attachment to the spindle microtubules. Unattached kinetochores activate the spindle checkpoint by recruiting several spindle-checkpoint proteins, including Mps1, Mad1, Mad2, Bub1, Bub3, and BubR1 (Mad3 in yeast) [1]. In vertebrate cells, active MAP kinase (MAPK) is also enriched at unattached kinetochores and is required for the spindle checkpoint [2–5]. It has been shown that the kinase activity of Mps1 is required for the spindle checkpoint [6] and for kinetochore localization of Bub1, Bub3, Mad1, and Mad2 [6, 7]. We herein demonstrate that MAPK phosphorylates Mps1 at S844 in Xenopus egg extracts. Interestingly, changing S844 to unphosphorylatable alanine (S844A) has no effect on the kinase activity of Mps1, although it abolishes the checkpoint function of Mps1. Biochemical and immunofluorescence studies show that S844A mutation perturbs kinetochore localization of Mps1 and other spindle-checkpoint proteins, whereas the phosphorylation-mimicking S844D mutant restores their functions. Our studies suggest that Mps1 phosphorylation by MAPK at S844 might create a phosphoepitope that allows Mps1 to interact with kinetochores. In addition, our results indicate that active Mps1 must localize to kinetochores in order to execute its checkpoint function. Results and Discussion MAPK Contributes to Mps1 Phosphorylation Mps1 is a dual-specificity protein kinase that functions in spindle-pole-body duplication and the spindle checkpoint [8–11]. The kinase activity of Mps1 is required for the spindle checkpoint [6] and targets Mad1 in budding yeast [12]. Mps1 is a phosphoprotein [10, 13], and phosphorylation of human Mps1 is regulated by the cell cycle [10, 13]. However, it is not clear how phosphorylation might regulate the kinase activity and the checkpoint function of Mps1. We first examined Mps1 in the egg extracts from the frog Xenopus laevis. Western blotting of Mps1 showed a slower gel mobility of the protein in
*Correspondence: [email protected]
metaphase and spindle-checkpoint-active extracts than in interphase extracts (Figure 1A, lanes 1–3), indicating that Mps1 was differentially modified at interphase and M phase. These proteins were indeed Mps1 because immunodepletion with Mps1 antibodies abolished the signal (Figure 1B, lane 2). The proteins were restored after the addition of in-vitro-synthesized RNA corresponding to the wild-type or kinase-dead Mps1 (Figure 1B, lanes 3 and 4). Both wild-type and kinase-dead Mps1 showed similar gel mobility (Figure 1B, lanes 3 and 4) that increased upon phosphatase treatment (Figure 1B, lanes 5 and 6). The result suggests that Mps1 in egg extracts was phosphorylated mostly by some other kinase(s), rather than through autophosphorylation. Interestingly, adding the phosphatase inhibitor microcystin to the samples as the control experiment resulted in a further mobility retardation of wild-type Mps1 but not the kinase-dead protein (Figure 1B, lanes 7 and 8), suggesting that autophosphorylation of Mps1 in egg extracts was balanced by a phosphatase activity. The in vitro immunocomplex kinase assay of Mps1 from various egg extracts showed similar levels of autophosphorylation activities (Figure 1A, lanes 4–6). The autophosphorylated Mps1 proteins migrated approximately as 160 kDa proteins that were indeed Mps1 because we were able to reimmunoprecipitate the denatured proteins with Mps1 antibodies (data not shown). The result suggests that the increased phosphorylation of Mps1 at M phase had no significant effect on the kinase activity of Mps1 in the egg extract. In addition, Western blotting of the in vitro autophosphorylation reactions showed an extensive mobility shift of the wild-type, but not the kinase-dead, Mps1 (Figure 1B, lanes 9 and 10), indicating that purified Mps1 was readily autophosphorylated in vitro. Xenopus Mps1 protein contains a number of serine/ threonine-proline sequences that are potential targets for Cdc2 and MAPK. Active MAPK is enriched at kinetochores of misaligned chromosomes and is dissociated from kinetochores after mid-anaphase [4, 5]. The spindle checkpoint requires active MAPK in Xenopus egg extracts and tissue-cultured cells [2, 3]. To determine whether Mps1 might be a target for MAPK, we examined the effect of U0126, an inhibitor for MAPK kinase (MEK). In U0126-treated metaphase egg extracts, Mps1 appeared in more-rapidly migrating species in comparison with that in control extracts, indicating reduced phosphorylation (Figure 1C) and suggesting that MAPK contributes to Mps1 phosphorylation. We performed two-dimensional tryptic-phosphopeptide mapping to determine the phosphorylation sites. We used kinase-dead Mps1 for the analysis in order to avoid any background phosphorylation generated through autophosphorylation. We first compared the maps of the proteins labeled with 32P in interphase and metaphase extracts. The map from the metaphase sample revealed a major phosphopeptide that was absent in the interphase sample, whereas most of the other phosphopeptides
MAPK-Dependent Kinetochore Targeting of Mps1 1765
Figure 1. MAPK Contributes to Mps1 Phosphorylation (A) Mps1 is differentially phosphorylated during the cell cycle. Mps1 immunoblotting (lanes 1–3) and in vitro Mps1-autophosphorylation reactions (lanes 4–6) were performed with interphase (lanes 1 and 4), metaphase (lanes 2 and 5), and spindle-checkpointactive (lanes 3 and 6) Xenopus egg extracts. An autoradiography of in-vitro-phosphorylated Mps1 is shown (lanes 3–6). The migration of molecular-size standards is indicated on the right. (B) Both wild-type and kinase-dead Mps1 are phosphorylated. Western blotting of Mps1 was performed with CSF-arrested extract (lane 1), Mps1-depleted extract (lane 2), and depleted extract supplemented with wildtype (lanes 3) or kinase-dead (lane 4) Mps1. Extracts containing wild-type or kinasedead Mps1 were treated with l protein phosphatase (lanes 5 and 6), treated with microcystin (lanes 7 and 8), or subjected to in vitro autophosphorylation reactions (lanes 9 and 10) prior to the Western blot analysis. (C) Inhibition of MAPK activity reduces Mps1 phosphorylation at M phase. Metaphase extracts were treated with DMSO (lane 1) or U0126 (lane 2). The samples were immunoblotted with anti-Mps1 (upper panel), antiMAPK (middle panel), and anti-phosphorylated MAPK (lower panel) antibodies. (D) Two-dimensional tryptic-phosphopeptide mapping of Mps1. Kinase-dead Mps1 was labeled with 32P in interphase or metaphase extracts in the absence or presence of U0126, as indicated. The map of kinasedead Mps1 that also contains the S844A mutation is shown in the lower right panel. (E) In vitro phosphorylation of Mps1 by MAPK. GST fusion of Mps1 (amino acids 756–882) (lanes 1 and 3) or the same region containing the S844 mutation (lanes 2 and 4) were subjected to in vitro phosphorylation with control (lanes 1 and 2) or MAPK (lanes 3 and 4) immunoprecipitations. An autoradiography of 32P incorporated into GST-Mps1 (upper panel) and the Coomassie Blue staining of the gel (lower panel) are shown.
were present in both interphase and metaphase extracts (Figure 1D). Interestingly, this mitotic phosphopeptide was greatly reduced when the extract was treated with U0126 (Figure 1D). The weak signal of the mitotic phosphopeptide in U0126-treated extracts might be due to residual MAPK activity in the extracts. These results indicate that the major mitosis-specific phosphopeptide was likely generated by MAPK. Alignment of Xenopus, mouse, and human Mps1 protein sequences revealed eight evolutionarily conserved MAPK recognition sites in Xenopus Mps1: S283, S459, T476, T491, T506, S733, T753, and S844. In order to identify the phosphorylation site, we constructed eight mutants that each contained a single mutation of alanine or valine in place of the conserved serine or threonine, respectively. Phosphopeptide mapping of these eight mutants showed that S844A abolished the major mitosis-specific phosphopeptide (Figure 1D, lower right panel), suggesting that S844 was likely the major target for MAPK. To test whether Mps1 is a substrate for MAPK in vitro, we generated GST-fusion proteins with truncated Mps1 that contained amino acid 756 to the C terminus. In vitro phosphorylation reactions showed that MAPK readily phosphorylated the wild-type protein but not the S844A mutant (Figure 1E), suggesting that Mps1 indeed was a substrate for MAPK.
The Spindle Checkpoint Requires Phosphorylation at S844 We next determined whether phosphorylation at S844 of Mps1 was involved in the spindle-checkpoint function of Mps1. The egg extracts are normally arrested at metaphase by the cytostatic factor activity (CSF), and the arrest is released upon calcium addition. The spindle checkpoint is typically activated in the extract by incubation with a sufficient amount (9,000–15,000/ml extracts) of sperm nuclei and the microtubule inhibitor nocodazole. After the checkpoint is activated, the extract remains at M phase even after calcium addition, as indicated by sustained Cdc2 activities (Figure 2A, mock depletion). In Mps1-depleted extracts incubated with sperm nuclei and nocodazole, Cdc2 activities declined rapidly upon calcium addition, indicating that the spindle checkpoint was lost in the absence of Mps1 (Figure 2A). When wild-type Mps1 was added back to extracts depleted for the endogenous protein, Cdc2 activities were maintained (Figure 2A), indicating that the exogenously added Mps1 was functional. However, addition of either kinase-dead or S844A mutants failed to sustain Cdc2 activities (Figure 2A), indicative of a defective spindle checkpoint. On the other hand, the S844D mutant, with aspartic acid in place of serine to mimic phosphorylation, was able to support the spindle
Current Biology 1766
Figure 3. S844A Affects Binding of the Checkpoint Proteins to Unattached Chromosomes, but Not the Kinase Activity of Mps1 Figure 2. S844A Mutation Abolishes the Spindle-Checkpoint Function of Mps1 (A) Sperm nuclei were incubated with mock-depleted extracts or Mps1-depleted extracts supplemented with wild-type (WT), kinasedead (KD), S844A, or S844D Mps1 proteins as indicated on the left. Nocodazole was added to samples as indicated on the right. Calcium chloride was then added to override the CSF activity. Cdc2 activities in the samples, as determined by in vitro histone H1 phosphorylation, were measured at the times indicated after calcium chloride addition. Autoradiographies of histone H1 phosphorylation are shown. An immunoblot of Mps1 in the samples at time 0 is shown in (B).
checkpoint (Figure 2A). These results suggest that both the kinase activity and phosphorylation at S844 are essential for the spindle-checkpoint function of Mps1. In addition, extracts containing the S844D mutant in the absence of nocodazole did not maintain Cdc2 activities after calcium addition (Figure 2A), indicating that S844D protein was unable to constitutively activate the spindle checkpoint without spindle disruption. Because the kinase activity of Mps1 is critical for the spindle checkpoint [6], we tested whether S844A mutation might affect the kinase activity of Mps1. Metaphase extracts depleted of endogenous Mps1 were supplemented with wild-type, kinase-dead, or S844A-mutant Mps1 to the endogenous levels. Mps1 proteins were immunoprecipitated for in vitro autophosphorylation
(A) Wild-type Mps1 (lane 1), kinase-dead (lane 2), or S844A (lane 3)mutant proteins were immunoprecipitated from metaphase egg extracts, and in vitro autophosphorylation reactions followed. An autoradiography of 32P incorporated into Mps1 is shown (upper panel). An immunoblot of Mps1 proteins in the extracts is shown in the lower panel. (B) Sperm nuclei and nocodazole were incubated with egg extracts that were mock depleted (lanes 1 and 6) or both depleted for the endogenous Mps1 and supplemented with mock (lanes 2 and 7), wildtype (lanes 3 and 8), kinase-dead (lanes 4 and 9), or S844A (lanes 5 and 10) Mps1 proteins as indicated. Chromosomes in the samples were isolated by centrifugation through a sucrose cushion. The extracts (lanes 1–5) and chromosomal fractions (lanes 6–10) were immunoblotted for proteins indicated on the left.
reactions. As expected, kinase-dead Mps1 had no autophosphorylation activity (Figure 3A, lane 2). Interestingly, the activity of the S844A mutant was comparable to that of the wild-type protein (Figure 3A, compare lanes 1 and 3), indicating that phosphorylation at S844 is not required for the kinase activity of Mps1. This result suggests that the checkpoint defect of the S844A mutant was not due to a lack of the kinase activity. Phosphorylation at S844 Is Required for Kinetochore Localization of the Spindle-Checkpoint Proteins Because Mps1 is required for kinetochore localization of other spindle-checkpoint proteins [6, 7], we tested the
MAPK-Dependent Kinetochore Targeting of Mps1 1767
Figure 4. S844 Phosphorylation Is Required for Kinetochore Localization of Mps1 (A) Sperm nuclei and nocodazole were incubated with mock-depleted extracts or Mps1-depleted extracts supplemented with mock or various Mps1 proteins as indicated on the left. The chromosomes were fixed and isolated for immunofluorescence staining with Mps1 and Mad1 antibodies. (B) Extracts were treated with DMSO or U0126 as indicated on the left before incubation with sperm nuclei and nocodazole for immunofluorescence staining of Mps1 and Mad1. Scale bars represent 10 mm.
possibility that S844A mutation might affect this function. We purified the chromosomes from nocodazoletreated egg extracts containing various forms of Mps1 and then analyzed the spindle-checkpoint proteins by Western blotting. The result showed that Mps1 depletion caused a great reduction in the levels of BubR1, Bub1, Bub3, Mad1, and Mad2 in the chromosomal fractions (Figure 3B). Addition of wild-type Mps1, but not kinase-dead or S844A mutants, to the depleted extracts restored the levels of the checkpoint proteins as well as hyperphosphorylation of Bub1 and BubR1 in the chromosomal fractions (Figure 3B). This result indicates that both the kinase activity of Mps1 and phosphorylation at S844 are required for efficient
recruitment of the spindle-checkpoint proteins to unattached chromosomes. We were unable to detect Mps1 in the chromosomal fractions by Western blotting, probably because of a low level of Mps1 at kinetochores or a very dynamic interaction of the protein with kinetochores. In fact, fluorescence recovery after photobleaching (FRAP) analysis of GFP-Mps1 reveals a fast turnover rate at unattached kinetochores with a half-life of approximately 10 s [14]. We thus examined Mps1 localization by immunofluorescence staining of isolated chromosomes. We found that Mps1 and Mad1 antibodies costained kinetochores of chromosomes prepared from mock-depleted extracts treated with nocodazole (Figure 4A, top panels). Both
Current Biology 1768
Mps1 and Mad1 disappeared from kinetochores in Mps1-depleted extracts (Figure 4A). Adding back wildtype Mps1 or the S844D mutant, but not the S844A mutant, restored both anti-Mps1 and anti-Mad1 staining at kinetochores (Figure 4A). Consistent with the notion that MAPK phosphorylates S844 in Mps1, U0126 treatment diminished both Mps1 and Mad1 at kinetochores (Figure 4B). These results suggest that kinetochore targeting of Mps1 requires phosphorylation at S844 and that the spindle-checkpoint defect of the S844A mutant is due to its inability to associate with kinetochores. In addition, kinase-dead Mps1 was able to localize to kinetochores, although the staining level was lower than that of wild-type and S844D proteins (Figure 4A). The result indicates that the kinase activity of Mps1 is not absolutely required for its kinetochore targeting but that phosphorylation of some kinetochore or spindle-checkpoint proteins, or both, by Mps1 might in turn enhance kinetochore association of Mps1. In this study, we demonstrate that Mps1 in Xenopus egg extracts is phosphorylated differentially during the cell cycle, and these results are similar to previous findings in human cells [10, 13]. We show that the increased phosphorylation at M phase is dependent on MAPK. MAPK is required for the spindle checkpoint, and phosphorylation of Cdc20 by MAPK is required for Cdc20 to associate with spindle-checkpoint proteins [15]. In addition, MAPK phosphorylates and activates Bub1 at kinetochores, and this activity facilitates the spindle checkpoint in egg extracts [16]. We herein identify Mps1 as another target for MAPK in the spindle checkpoint. Together, these findings suggest that MAPK controls several major steps in the spindle checkpoint. We map the major MAPK-dependent phosphorylation site to S844 and demonstrate that phosphorylation at this site is essential for the spindle checkpoint. The unphosphorylatable S844A mutant is defective in the spindle checkpoint, whereas the phosphorylation-mimicking S844D mutant is functional. The S844A mutant retains the normal kinase activity but is unable to localize to the kinetochore. These results suggest that active Mps1 must associate with kinetochores in order to exert its function in the spindle checkpoint. It is possible that Mps1 might phosphorylate some kinetochore components in order for spindle-checkpoint proteins to bind. Alternatively, Mps1 might efficiently phosphorylate its substrate when both Mps1 and the substrate are brought to close proximity by the unattached kinetochore. In budding yeast, it has been shown that overexpression of Mps1 constitutively activates the spindle checkpoint in the absence of any kinetochore or spindle defect [12, 17, 18]. Overexpression of Mps1 might allow efficient phosphorylation of its substrates and thus bypass kinetochore targeting. Furthermore, the S844A mutant cannot localize to the kinetochore, suggesting that phosphorylation of S844 by MAPK might create some phosphoepitope that allows Mps1 to interact with kinetochore components. It has also been shown that Aurora-B/INCENP complex is required for kinetochore localization of Mps1 [7]. It remains to be determined if there is any interplay between Aurora-B/INCENP and MAPK. Upon docking at the kinetochore, Mps1 then phosphorylates other kinetochore or spindle-checkpoint proteins, or both, to recruit other spindle-checkpoint
proteins to kinetochores. The nature of the kinetochore component that associates with phosphorylated Mps1 and the substrates for Mps1 remains to be determined. Supplemental Data Supplemental Data include Experimental Procedures and can be found with this article online at http://www.current-biology.com/ cgi/content/full/16/17/1764/DC1/. Acknowledgments This work was supported by grants from the National Institutes of Health (USA), the David and Lucile Packard Foundation (USA), the National Science Council (Taiwan), and Academia Sinica (Taiwan). The authors declare that there is no financial conflict of interest related to this work. Received: May 16, 2006 Revised: July 1, 2006 Accepted: July 13, 2006 Published: September 5, 2006
References 1. Taylor, S.S., Scott, M.I., and Holland, A.J. (2004). The spindle checkpoint: A quality control mechanism which ensures accurate chromosome segregation. Chromosome Res. 12, 599–616. 2. Minshull, J., Sun, H., Tonks, N.K., and Murray, A.W. (1994). MAPkinase dependent mitotic feedback arrest in Xenopus egg extracts. Cell 79, 475–486. 3. Wang, X.M., Zhai, Y., and Ferrell, J.E., Jr. (1997). A role for mitogen-activated protein kinase in the spindle assembly checkpoint in XTC cells. J. Cell Biol. 137, 433–443. 4. Shapiro, P.S., Vaisberg, E., Hunt, A.J., Tolwinski, N.S., Whalen, A.M., McIntosh, J.R., and Ahn, N.G. (1998). Activation of the MKK/ERK pathway during somatic cell mitosis: Direct interactions of active ERK with kinetochores and regulation of the mitotic 3F3/2 phosphoantigen. J. Cell Biol. 142, 1533–1545. 5. Zecevic, M., Catling, A.D., Eblen, S.T., Renzi, L., Hittle, J.C., Yen, T.J., Gorbsky, G.J., and Weber, M.J. (1998). Active MAP kinase in mitosis: Localization at kinetochores and association with the motor protein CENP-E. J. Cell Biol. 142, 1547–1558. 6. Abrieu, A., Magnaghi-Jaulin, L., Kahana, J.A., Peter, M., Castro, A., Vigneron, S., Lorca, T., Cleveland, D.W., and Labbe, J. (2001). Mps1 is a kinetochore-associated kinase essential for the vertebrate mitotic checkpoint. Cell 106, 83–93. 7. Vigneron, S., Prieto, S., Bernis, C., Labbe, J.C., Castro, A., and Lorca, T. (2004). Kinetochore localization of spindle checkpoint proteins: Who controls whom? Mol. Biol. Cell 15, 4584–4596. 8. Lauze, E., Stoelcker, B., Luca, F.C., Weiss, E., Schutz, A.R., and Winey, M. (1995). Yeast spindle pole body duplication gene MPS1 encodes an essential dual specificity protein kinase. EMBO J. 14, 1655–1663. 9. Weiss, E., and Winey, M. (1996). The S. cerevisiae SPB duplication gene MPS1 is part of a mitotic checkpoint. J. Cell Biol. 132, 111–123. 10. Stucke, V.M., Sillje, H.H., Arnaud, L., and Nigg, E.A. (2002). Human Mps1 kinase is required for the spindle assembly checkpoint but not for centrosome duplication. EMBO J. 21, 1723– 1732. 11. Fisk, H.A., Mattison, C.P., and Winey, M. (2003). Human Mps1 protein kinase is required for centrosome duplication and normal mitotic progression. Proc. Natl. Acad. Sci. USA 100, 14875–14880. 12. Hardwick, K.G., Weiss, E., Luca, F.C., Winey, M., and Murray, A.W. (1996). Activation of the budding yeast spindle assembly checkpoint without mitotic spindle disruption. Science 273, 953–956. 13. Liu, S.T., Chan, G.K., Hittle, J.C., Fujii, G., Lees, E., and Yen, T.J. (2003). Human MPS1 kinase is required for mitotic arrest induced by the loss of CENP-E from kinetochores. Mol. Biol. Cell 14, 1638–1651.
MAPK-Dependent Kinetochore Targeting of Mps1 1769
14. Howell, B.J., Moree, B., Farrar, E.M., Stewart, S., Fang, G., and Salmon, E.D. (2004). Spindle checkpoint protein dynamics at kinetochores in living cells. Curr. Biol. 14, 953–964. 15. Chung, E., and Chen, R.H. (2003). Phosphorylation of Cdc20 is required for its inhibition by the spindle checkpoint. Nat. Cell Biol. 5, 748–753. 16. Chen, R.H. (2004). Phosphorylation and activation of Bub1 on unattached chromosomes facilitate the spindle checkpoint. EMBO J. 23, 3113–3121. 17. Fraschini, R., Beretta, A., Lucchini, G., and Piatti, S. (2001). Role of the kinetochore protein Ndc10 in mitotic checkpoint activation in Saccharomyces cerevisiae. Mol. Genet. Genomics 266, 115–125. 18. Poddar, A., Daniel, J.A., Daum, J.R., and Burke, D.J. (2004). Differential kinetochore requirements for establishment and maintenance of the spindle checkpoint are dependent on the mechanism of checkpoint activation in Saccharomyces cerevisiae. Cell Cycle 3, 197–204.