R698

Dispatch

Centrosome duplication: Three kinases come up a winner! Edward H. Hinchcliffe* and Greenfield Sluder

Despite over one hundred years of research, the duplication of the centrosome is a poorly understood process. Three recent papers — exploring three different kinases — may have provided the answer. Address: University of Massachusetts Medical School, Department of Cell Biology, 377 Plantation Street, Worcester, Massachusetts 01605, USA. *Present address: University of Notre Dame, Department of Biological Sciences, 107 Galvin Life Science Center, Notre Dame, Indiana 46556, USA. E-mail: [email protected] Current Biology 2001, 11:R698–R701 0960-9822/01/$ – see front matter © 2001 Elsevier Science Ltd. All rights reserved.

The ultimate purpose of mitosis is the generation of genetically identical daughter cells. All chromosomes are replicated during S phase, and the resulting pairs of sister chromatids must be segregated equally during mitosis. The fidelity of this process depends on the assembly of a strictly bipolar mitotic spindle. This essential ‘twoness’ within the mitotic cell is dependent upon the reproduction or duplication of the centrosome [1]. In higher animal cells, the centrosome is an organelle that consists of a pair of centrioles surrounded by a matrix of pericentriolar material which contains the γ-tubulin ring complexes that nucleate microtubules during interphase and mitosis (reviewed in [2]). The duplication of the interphase centrosome occurs only once per cell cycle (reviewed in [3]). If the centrosome fails to duplicate, the cell forms a monopolar spindle, does not divide and becomes polyploid. More than one duplication of the centrosome can lead to the assembly of a multipolar spindle, and the chromosomes may become unequally distributed to the daughter cells. In either case, mistakes in centrosome reproduction can lead to the genomic instability that is characteristic of the cells of many aggressive human cancers (reviewed in [4]). Spindle multipolarity is a particularly intractable problem because cells have neither a way to eliminate extra centrosomes nor a checkpoint that aborts mitosis in response to extra spindle poles [5]. In order to ensure that only one new centrosome assembles per cell cycle, the cell must tightly coordinate the events of centrosome duplication with the cycle of nuclear replication and division. Within the past two years a number of studies have revealed that DNA replication and centrosome duplication share a common regulatory pathway: the rise in the activity of cyclin-dependent kinase 2 (Cdk2) complexed with cyclin E [6–8], or cyclin A for some cultured

cells [9]. Cdk2 activity increases in late G1 and persists into the beginning of S phase. These are times when daughter centrioles are first assembled at the proximal end of each mature or parent centriole [10]. While these studies indicated that Cdk2 activity is required for centrosome reproduction, they did not reveal whether Cdk2 acts directly on centrosomes, or indirectly through other pathways. Two recent studies [11,12] have addressed this question: the answer appears to be both. The first study, by Okuda et al. [11], provided a significant advance in our understanding of how the cyclin dependent kinase 2–cyclin E complex (Cdk2-E) acts to control centrosome duplication, by identifying one of the centrosomal substrates for Cdk2-E. The authors began by isolating unreplicated centrosomes from G0 cells, and then subjected them to in vitro phosphorylation by purified Cdk2-E. A single phospho-polypeptide was isolated and analyzed by mass-spectroscopy. It was found to be a previously identified component of the nucleolus: nucleophosmin [13]. Immunofluorescence analysis of Swiss 3T3 cells revealed that nucleophosmin localizes to un-duplicated centrosomes but is not found on centrosomes after duplication. Later, when the cells are in mitosis, nucleophosmin is again observed at the centrosomes (also see [14]). Conceivably, phosphorylation of other sites on this protein by the mitotic kinase Cdk1–cyclin B [15] could cause nucleophosmin to re-bind to the centrosome during M phase. Importantly, Okuda et al. [11] found that expression of a non-phosphorylatable form of nucleophosmin blocked the duplication of centrosomes, as did the microinjection of antibodies to nucleophosmin. Electron microscopy of cells expressing the non-phosphorylatable nucleophosmin revealed that the centrioles had not split apart. Thus, it appears that nucleophosmin associates with centrosome during mitosis and thereby prevents centriole splitting and duplication until late G1, when Cdk2-E activity rises in preparation for S phase. This is followed by the assembly of procentrioles and hence reproduction of the centrosome. Thus, the Cdk2-E phosphorylation of nucleophosmin directly coordinates the start of centrosome duplication with DNA synthesis (Figure 1). In the second study, Fisk and Winey [12] demonstrate the importance of mMps-1p (also known as mouse Esk protein kinase) in the duplication of the mammalian centrosome. This is the mouse ortholog of the yeast kinase Mps-1p, originally identified as an essential protein kinase involved in the duplication of the spindle pole body, the yeast equivalent of the centrosome [16]. Mps-1p is also involved

Dispatch

R699

Figure 1

Cdk2–cyclin E Nucleophosmin to centrosomes (Cdk1–cyclin B activity?) Mp-1 kinase

Unreplicated centrosome G0

Splitting

G1

Reproduction

S-phase

Replicated centrosome

Spindle poles

Unreplicated centrosome

G2

Mitosis

G0 Current Biology

Schematic drawing of the cycle of centrosome reproduction in animal cells. The centrosome is represented as two blue rectangles (the centriole pair). In G0, the centrosome is unduplicated and is decorated with nucleophosmin (red stars). As the cell cycle proceeds through G1 into S phase, Cdk2–cyclin E phosphorylates nucleophosmin, causing its release from the centrosome and the splitting of the daughter

centrioles. In addition, Cdk2–cyclin E activity stabilizes Mps-1, which is required for the centrosome to reproduce. In G2, the pair of duplicated centrosomes disjoin and separate to fom the poles of the mitotic spindle. These spindle poles re-acquire nucleophosmin during M phase (red stars), perhaps in response to Cdk1–cyclin B activity. Finally, the cell exits mitosis and the centrosome cycle begins again.

in the function of the spindle assembly checkpoint in yeast, and Abrieu et al. [17] have recently shown that this kinase also has a role in the mammalian spindle assembly checkpoint. Fisk and Winey [12] found that, in mouse cells, antibodies to mMps-1p decorate centrosomes throughout the cell cycle and kinetochores during mitosis. That these localizations reflect the true in vivo distributions of the kinase was confirmed by examining stable cell lines expressing a GFP–mMps-1p fusion.

was lost. This stabilization of mMps-1p by Cdk2 activity is another, albeit indirect, way for the cell to ensure that centrosome duplication is coordinated with S phase (Figure 1).

The functional experiments carried by Fisk and Winey [12] have shown that mMps-1p kinase activity is required for centrosome duplication. By correlative light and serialsection electron microscopy they found that the overexpression of mMps-1p (with or without GFP) during S phase arrest caused the repeated duplication of centrosomes and their centrioles. Importantly, S phase arrest alone, without mMps-1p overexpression, does not cause centrosome reduplication. Furthermore, expression of a kinase-inactive point mutant form of mMps-1p blocked centrosome duplication, even though it still localized to the centrosomes. In a surprising twist, Fisk and Winey [12] went on to demonstrate that Cdk2 kinase activity is required for mMps-1p-dependent centrosome duplication by stabilizing mMps-1 protein levels. When Cdk2-E activity was blocked by drug treatments or by overexpression of the Cdk inhibitors p21 or p27, the cellular level of mMps-1p dramatically dropped and its localization to the centrosome

It is still too early to know how nucleophosmin phosphorylation and mMps-1p kinase activity fit together into the sequence of events involved in centrosome duplication. Fisk and Winey [12], though open to several possibilities, suggest that mMps-1p could act downstream of nucleophosmin phosphorylation and its release from the centrosome. Perhaps the dissociation of nucleophosmin from the centrosome uncovers sites for phosphorylation by mMps-1p that are needed for reproduction to continue. While these papers have focused on mammalian somatic cells, much of our knowledge about centrosome reproduction has come from live-cell studies of early cleavage stage zygotes. In a third paper, O’Connell et al. [18] have added modern genetics to this classical approach. Using the nematode Caenorhabditis elegans, they have characterized Zyg-1 — a novel kinase required for centrosome duplication. Previous analysis of zyg-1 mutant embryos revealed a defect in cell division characterized by the failure to form a bipolar spindle, even though cell-cycle progression appeared normal [19]. In the present study [18], the authors found that this was due to the assembly of a monopolar spindle at mitosis. To investigate the basis for this surprising result, O’Connell et al. [18] then conducted a series of crosses

R700

Current Biology Vol 11 No 17

Figure 2

zyg-1 maternal

Wild-type

zyg-1 paternal

(a)

(d)

(g)

(b)

(e)

(h)

(c)

(f)

(i)

Interphase

Mitosis

Interphase

Current Biology

using sperm or eggs homozygous for what are essentially null alleles of zyg-1. They then analyzed the resultant mutant embryos by time-lapse videomicroscopy and correlative thin-section electron microscopy. Paternal zyg-1 is required for the assembly of a centriole pair during spermatogenesis. In male mutants, this fails, and wild-type eggs fertilized with mutant sperm receive a single centriole (Figure 2). During the first cell cycle, this centriole duplicates — in response to maternal Zyg-1 activity — and forms a single centrosome with a pair of centrioles, and a monopolar spindle results. During the second cell cycle, this centrosome duplicates normally, and the subsequent division is bipolar. Mutant eggs fertilized with wild-type sperm receive a pair of centrioles (Figure 2). During the first cell cycle, these centrioles split apart, but fail to duplicate because of the lack of maternal Zyg-1 kinase activity. This gives rise to a bipolar spindle, with each pole containing a single centriole. At the two-cell stage, these single centrioles again cannot reproduce, and the subsequent mitotic spindles are both monopolar. Immunofluorescence microscopy revealed that the Zyg-1 kinase transiently localizes to the centrosome in C. elegans cells in late mitosis, but is totally absent from centrosomes during interphase. As centrosome duplication normally begins sometime in late G1 or early S phase in somatic cells, such localization of Zyg-1 only during mitosis suggests it may function to prepare the centrosome to duplicate prior to the actual onset of centrosome reproduction. It is also possible, however, that Zyg-1 activity and centrosome reproduction overlap, because S phase in the early

Schematic drawing of centrosome reproduction and spindle formation in early embryos of C. elegans. In wild-type, (a) the sperm provides a pair of centrioles that duplicate once. (b) The daughter centrosomes separate and form the poles of the spindle; each pole has a pair of centrioles. (c) Each daughter cell inherits a centrosome with a pair of centrioles. In zygotes with a maternal mutation in zyg-1, (d) the sperm provides a pair of centrioles. (e) These split, but do not duplicate in the absence of maternal zyg-1 activity. Consequently, though the first spindle is bipolar, each pole only has one centriole. (f) These single centrioles do not duplicate again. In zygotes with a paternal mutation in zyg-1, (g) the sperm centrioles do not reproduce during spermatogenesis, and the egg is fertilized with a single centriole. (h) This centriole duplicates once, giving a monopolar spindle – the zygote does not cleave. (i) The centrosome reproduces again, giving a zygote with two centrosomes each containing a pair of centrioles.

C. elegans zygote begins immediately after the completion of mitosis [20]. In this case, Zyg-1 kinase activity would be directly involved in centriole replication. Either way, it will be interesting to determine whether other organisms have functional homologs of Zyg-1. In summary, the combined results of the three recent studies discussed above [11,12,18] are clear: three kinases give centrosomes the winning combination. Acknowledgements We thank Mark Winey, Harold Fisk, Kevin O’Connell and John White for communicating results prior to publication. Work from the author’s laboratory is funded by a grant to GS from the National Institutes of Health (GM 30758). EHH is supported by a senior post-doctoral fellowship from the American Cancer Society, Massachusetts Division.

References 1. Mazia D: The chromosome cycle and the centrosome cycle in the mitotic cycle. Int Rev Cytol 1987, 100:49-92. 2. Doxsey SJ: Centrosomes as command centres for cellular control. Nat Cell Biol 2001, 3:E105-108. 3. Stearns T: Centrosome duplication. a centriolar pas de deux. Cell 2001, 105:417-420. 4. Brinkley BR: Managing the centrosome numbers game: from chaos to stability in cancer cell division. Trends Cell Biol 2001, 11:18-21. 5. Sluder G, Thompson EA, Miller FJ, Hayes J, Rieder CL: The checkpoint control for anaphase onset does not monitor excess numbers of spindle poles or bipolar spindle symmetry. J Cell Sci 1997, 110:421-429. 6. Hinchcliffe EH, Li C, Thompson EA, Maller JA, Sluder G: Requirement of Cdk2 - Cyclin E activity for repeated centrosome reproduction in Xenopus egg extracts. Science 1998, 283:851-854. 7. Lacey KR, Jackson PK, Stearns T: Cyclin-dependent kinase control of centrosome duplication. Proc Natl Acad Sci USA 1999, 96:2817-2822. 8. Matsumoto Y, Hayashi K, Nishida E: Cyclin-dependent kinase 2 (Cdk2) is required for centrosome duplication in mammalian cells. Curr Biol 1999, 9:429-432.

Dispatch

9. Meraldi P, Lukas J, Fry AM, Bartek J, Nigg EA: Centrosome duplication in mammalian somatic cells requires E2F and Cdk2Cyclin A. Nat Cell Biol 1999, 1:88-93. 10. Kuriyama R, Borisy GG: Centriole cycle in chinese hamster ovary cells as determined by whole-mount electron microscopy. J Cell Biol 1981, 91:814-821. 11. Okuda M, Horn HF, Tarapore P, Tokuyama Y, Smulian AG, Chan PK, Knudsen ES, Hofmann IA, Snyder JD, Bove KE, Fukasawa K: Nucleophosmin/B23 is a target of CDK2/cyclin E in centrosome duplication. Cell 2000, 103:127-140. 12. Fisk HA, Winey M: The mouse Mps1p-like kinase regulates centrosome duplication. Cell 2001, 106:95-104. 13. Schmidt-Zachmann MS, Hugle-Dorr B, Franke WW: A constitutive nucleolar protein identified as a member of the nucleoplasmin family. EMBO J 1987, 6:1881-1890. 14. Zatsepina OV, Rousselet A, Chan PK, Olson MO, Jordan EG, Bornens M: The nucleolar phosphoprotein B23 redistributes in part to the spindle poles during mitosis. J Cell Sci 1999, 112:455-466. 15. Peter M, Nakagawa J, Doree M, Labbe J-C, Nigg EA: Identification of major nucleolar proteins as candidate mitotic substrates of cdc2 kinase. Cell 1990, 60:791-801. 16. Winey M, Goetsch L, Baum P, Byers B: MPS-1 and MPS-2: novel yeast genes defining distinct steps of spindle pole body duplication. J Cell Biol 1991, 114:745-754. 17. Abrieu A, Magnaghi-Jaulin L, Kahana JA, Peter M, Castro A, Vigneron S, Cleveland DW, Lorcha T, Labbe JC: X-Mps-1, an essential component of the mitotic checkpoint, is necessary for Mad1, Mad2, and CENP-E association with kinetochores, in Xenopus egg extracts. Cell 2001, 106:83-93. 18. O’Connell KF, Caron C, Kopish KR, Hurd DD, Kemphues KJ, Li Y, White JG: The C. elegans zyg-1 gene encodes a regulator of centrosome duplication with distinct maternal and paternal roles in the embryo. Cell 2001, 105:547-558. 19. O’Connell KF, Leys CM, White JG: A genetic screen for temperature-sensitive cell-division mutants of Caenorhabditis elegans. Genetics 1998, 149:1303-1321. 20. Edgar LG, McGhee JD: DNA synthesis and the control of embryonic gene expression in C. elegans. Cell 1988, 53:589-599.

R701