PERSPECTIVES

OPINION

Two-way traffic: centrosomes and the cell cycle Greenfield Sluder Abstract | The well recognized activities of the mammalian centrosome — microtubule nucleation, duplication, and organization of the primary cilium — are under the control of the cell cycle. However, the centrosome is more than just a follower of the cell cycle; it can also be essential for the cell to transit G1 and enter S phase. How the centrosome influences G1 progression is a mystery.

The centrosome is the primary microtubule organizing centre (MTOC) of the mammalian cell. It consists of a pair of centrioles associated with a cloud of amorphous pericentriolar material (PCM; BOX 1). During interphase, the centrosome nucleates and organizes the cytoplasmic microtubule array, which is involved in the determination of cell polarity, cell motility and organelle transport, towards or away from the cell centre. In preparation for mitosis, the interphase centrosome is precisely duplicated as the cell enters S phase, and in G2 phase the sister centrosomes start to move to opposite sides of the nucleus. During mitosis, centrosomes nucleate most spindle microtubules and thereby determine spindle polarity, spindle position/orientation and the plane of cleavage. However, even though centrosomes function in a dominant manner to determine spindle polarity, mammalian somatic cells can assemble a bipolar spindle, segregate chromosomes and divide successfully without centrosomes1,2. But cells dividing without centrosomes take a significantly longer and more variable amount of time to go through mitosis, they sometimes fail to assemble a bipolar spindle, almost half fail to complete cleavage and spindle positioning can be abnormal (E. Hinchcliffe, personal communication)3–6. So, although centrosomes are not absolutely required for mitosis, they are essential for the fidelity of cell division. The centrosome as cell-cycle ‘follower’

The traditional viewpoint that the animal centrosome is a downstream target of the

NATURE REVIEWS | MOLECULAR CELL BIOLOGY

mechanisms that control the cell cycle originated from two sorts of observations. First, centrosomes are not needed for cellcycle progression in sea urchin zygotes and Xenopus laevis egg extracts7,8. And mammalian somatic cells from which the interphase centrosome is microsurgically removed progress with approximately normal timing from S phase into mitosis 3 . Second, the well-recognized activities of the centrosome are temporally linked to, and dependent on, cell-cycle progression. For example, the number of microtubules that are nucleated by centrosomes is low in early interphase and increases as the cell approaches mitosis. This does not simply reflect changes in cytoplasmic conditions that alter microtubule assembly characteristics as the cell enters mitosis. The number of microtubules growing in vitro from centrosomes in lysed cell models increases the closer the cells are to the onset of mitosis9,10 owing to the phosphorylation of proteins in the PCM11–13 and increases in the amount of γ-tubulin at the centrosome14,15. Another important activity of the centrosome — its precise duplication — is directly under the control of the cell cycle. Under normal circumstances this event is initiated by the late G1 rise in the activities of cyclin-dependent kinase 2 (CDK2) coupled to cyclin E or cyclin A, the kinase complexes that drive the cell into S phase16 BOX 2. So, coordination of centrosome and DNA duplication is assured by a common control. The centrosome as cell-cycle ‘leader’

This view of the centrosome as just a faithful follower of the cell cycle, although true, is incomplete. Indications that the centrosome might carry out an active role in the control of cell-cycle progression first emerged from a study in which mammalian somatic cells were microsurgically cut at various points during interphase into a centrosome-free nucleated cell (karyoplast) and a centrosome-containing fragment of the cell (cytoplast)17 (FIG. 1). Intermittent observations over 10 days revealed that the karyoplasts assembled an interphase array of

microtubules but no karyoplasts were seen in mitosis. This was not the consequence of mechanical damage, because control amputations of large areas of cytoplasm did not alter the ability of the cells to divide several times3,17. However, when this experiment was repeated and karyoplasts were continuously monitored by time-lapse video microscopy, a more detailed picture of karyoplast behaviour emerged3. When the microsurgery was carried out around the time of S phase, most of the karyoplasts progressed into mitosis without delay, something not previously seen owing to the intermittent nature of the observations17. Spindle assembly occurred through an acentrosomal pathway and chromosomes appeared to be equally segregated in anaphase. However, the karyoplasts took a substantially greater and more variable amount of time than normal to complete mitosis, and 41% failed to complete cleavage. The real surprise came when the postmitotic karyoplasts were followed for another 50–70 hours (equivalent to 2–3 cell-cycle durations). Almost all karyoplasts, whether cleavage was complete or not, arrested before the onset of DNA synthesis. Although the karyoplasts did not reform centrioles, as observed by serial section electron microscopy, they all reassembled a functional MTOC, as evidenced by an extensive interphase array of microtubules that were associated with a single concentration of γ-tubulin and pericentrin (FIG. 2). The extent to which the microtubule array was organized in a radial fashion varied between karyoplasts. In some, the degree of organization was similar to that seen in the adjacent control cells; in others, the microtubule array did not appear to have a well defined focus (FIG. 2). Subsequent work indicated that postmitotic karyoplasts arrest with elevated levels of p21, an absence of the Ki-67 proliferation antigen and hypophosphorylated retinoblastoma protein (Rb) — implying a p53-mediated, early G1 arrest (J. Nordberg and G.S., unpublished observations). The tumour suppressor p53 is activated in

“This view of the centrosome as just a faithful follower of the cell cycle, although true, is incomplete.” VOLUME 6 | SEPTEMBER 2005 | 743

PERSPECTIVES

Box 1 | Centrosome structure a

b

Daughter centriole Distal appendages Mother centriole

Interconnecting fibres PCM

Subdistal appendages

Microtubule

The centrosome of the mammalian somatic cell is a small (~1 μm) body composed of a pair of centrioles associated with a cloud of amorphous material called the pericentriolar material (PCM). Part a of the figure is an electron micrograph of a mother-daughter centriole pair; the mother is shown in cross-section and the daughter is shown in longitudinal section. The arrowheads delimit the distribution of the PCM. Both centrioles consist of a barrel of nine triplet microtubules held together by linkers; within each barrel are a number of poorly understood luminal structures (not shown). One centriole (which is often called the mother centriole) is older and functions as the assembly site for the younger, or daughter, centriole during centrosome duplication in S phase of the previous cell cycle. The mother centriole differs from the daughter in that it has nine distal and nine subdistal appendages radiating from its exterior surface (see figure, part b)47,48. The primary cilium, when present during interphase, is organized by the mother centriole (not shown). The PCM surrounds primarily the mother centriole in many somatic cells and both centrioles in eggs47,49–52. The PCM binds γ-tubulin ring complexes that nucleate the interphase microtubule array and the astral microtubules during mitosis53,54. In addition, the PCM binds a variety of cell-cycle regulatory proteins24,55–57 (for a comprehensive list of proteins found in the isolated mammalian centrosome, see supplementary materials in REF. 58). Although the centrioles do not directly nucleate cytoplasmic microtubules, microtubules that are nucleated in the PCM become attached to the globular tips of the nine subdistal appendages of the mother centriole47,59,60, and function to anchor this centriole in the cytoplasm5,61. The centriole pair localizes the PCM into a focal body, and thereby determines the number of microtubule-organizing centres that are present in the cell under normal circumstances16,62. Part a of the figure is reproduced with permission from REF. 49 © (1982) Société Française de Microscopies. Part b is modified with permission from Nature Reviews Molecular Cell Biology REF. 55 © (2001) Macmillan Magazines Ltd.

response to various cellular stresses; it functions as a transcription factor to upregulate the expression of CDK inhibitors, such as p21, which arrest the cell cycle BOX 2. Khodjakov and Rieder4 obtained similar and complementary results using a laser to ablate centrosomes in mammalian somatic cells during mitosis. After ablation of one centrosome at metaphase, the cells completed a bipolar mitosis and cleaved without noticeable delay. The daughter cells inheriting a centrosome underwent DNA synthesis and some cells later entered mitosis. The daughters lacking a centrosome, by contrast, assembled a disorganized array of interphase microtubules

744 | SEPTEMBER 2005

| VOLUME 6

and arrested in G1 for at least 72 hours. So removal of the centrosome in mitosis leads to an immediate G1 arrest; the cell does not need to go through mitosis in its entirety without a centrosome for the G1 arrest to occur. G1 arrest, prolonged mitosis and tetraploidy. Karyoplasts spent longer than normal, on average, in mitosis; this raised the possibility that the G1 arrest was the result of a prolongation of mitosis and, possibly, exit from mitosis without full downregulation of the ‘spindle-assembly checkpoint’ signalling pathway that prolongs mitosis when one or more kinetochores are not attached to

microtubules. Indeed, nocodazole (a drug that inhibits microtubule assembly)-treated cells that exit mitosis without microtubules need prolonged activity of the spindleassembly checkpoint to later arrest in G1 in a p53-dependent fashion18. However, prolongation of mitosis is not the reason for the G1 arrest of karyoplasts, because some went through mitosis within the normal range of mitotic durations, yet still arrested in G1 REF. 3, and laser ablation of centrosomes at metaphase does not prolong mitosis4. Also, when control cells were treated with low doses of paclitaxel (Taxol; which stabilizes microtubules), mitosis was prolonged to the same extent (mean and range) as that observed for karyoplasts, and many progressed through interphase into another mitosis3. Putative chromosome segregation defects that could result in aneuploid daughter karyoplasts are not a probable cause for the G1 arrest because karyoplasts that failed cleavage also arrested in G1 even though they have a complete — albeit doubled — complement of chromosomes. Finally, the G1 arrest of post-mitotic karyoplasts cannot be a result of the failure of the cleavage process or a ‘tetraploidy checkpoint’19 because over half of the karyoplasts successfully completed cleavage, and this checkpoint does not exist in mammalian somatic cells20. Focusing on the centrosome. Together, these studies indicated that the G1 arrest caused by centrosome removal does not occur through the lack of MTOC activity, total microtubule content, or prolongation of mitosis. The logical conclusion at the time was that G1 progression depended on the presence of core centrosome components — centrioles or molecular complexes that require the presence of centrioles for their activity. However, the notion that the G1 arrest is caused uniquely by the removal of centrioles is too simple. Alterations in the expression levels of several PCM proteins can also cause G1 arrest despite the presence of a presumably normal complement of centrioles. For example, when PCM1, a protein of the PCM and pericentriolar satellite bodies (small electron-dense aggregates of amorphous material near the centrosome), is knocked down using short interfering (si)RNA, cultured cells arrest in G1 with an extensive, although disorganized, interphase array of microtubules21 (A. Merdes, personal communication). Similarly, the injection of a PCM1 antibody into mouse

www.nature.com/reviews/molcellbio

PERSPECTIVES

Box 2 | G1 cell-cycle controls During G1, the mammalian somatic cell is sensitive to many intracellular and extracellular stimuli, which are integrated to determine whether the cell will re-enter the cell cycle, wait for cellular damage to be repaired, differentiate, or trigger apoptosis. The figure shows a highly simplified representation of some of the interactions between the important players that control G1 progression. For more complete coverage of this complex subject, as well as the incompletely understood compensatory pathways that allow cell proliferation to occur in mouse models in which CDK2 and cyclin E1 and E2 have been knocked out, see REFS 6369.

Starting (green pathway) At the cell surface, mitogenic signals and integrin signalling at focal adhesions activate a pathway — through focal adhesion kinase (FAK), Ras and extracellular-signal-regulated kinase (ERK) — that leads to the accumulation of cyclin D.

Core pathway (blue) Cyclin D forms a complex with the cyclin-dependent kinases CDK4 and CDK6, which then phosphorylate retinoblastoma protein (Rb) bound to the transcription factor E2F. Multiple phosphorylations of Rb lead to the release of E2F, which promotes the synthesis of cyclins A and E. These associate with CDK2 to form active kinase complexes that drive entry into S phase. CDK2–cyclin E and CDK2–cyclin A also both phosphorylate Rb to form a positive feedback loop. P, phosphate.

Controlling entry into S phase (red) Various cellular defects (see figure, upper right) can each lead to the accumulation of p53, which can promote apoptosis (programmed cell death) or the accumulation of the CDK2 inhibitor p21. p21 activity is also influenced by the activity of the ERK1/2 kinases, the activities of which respond to cell-surface signalling (left). Other negative regulatory factors (lower right) can lead to the accumulation of p27, which inhibits the activity of CDK4/6–cyclin D as well as CDK2– cyclin A/E. Mitogenic signals Integrins FAK Apoptosis

Ras

Centrosome defects

Transient/strong ERK activity ERK1/2

p53

DNA damage

Sustained ERK activity Cyclin D

Rb–E2F

Low integrin signalling

p21

CDK4/6–cyclin D P

CDK4/6

P + E2F

Rb P

Cyclin E

CDK2–cyclin E

Cyclin A

CDK2–cyclin A

S phase CDK2

Contact inhibition Cytoskeletal defects

p27

Serum starvation Mitogenic signals

zygotes causes a G1 arrest in the first cycle after fertilization22. Also, knocking down centriolin, a protein found at the tips of the subdistal appendages of centrioles and involved at the midbody in cleavage completion, causes a G1 arrest23. In addition, expression of the centrosome-anchoring domain of large A-kinase anchoring protein AKAP450 displaces the endogenous fulllength protein from the centrosome and causes a G1 arrest24; the expression of the

NATURE REVIEWS | MOLECULAR CELL BIOLOGY

N-terminal centrosome-binding domain of poly (ADP-ribose) polymerase has been reported to block HeLa cells in G1 REF. 25; the expression of CSPP–EGFP (centrosome/spindle pole-associated protein coupled to enhanced green fluorescent protein) was reported to arrest HEK293T cells in G1 REF. 26; and the destabilization of centrosome-associated dynactin seems to delay entry into S phase27. Remarkably, the siRNA-mediated knockdown of any of 15

centrosome-associated proteins results in a p53-dependent G1 arrest28 (K. Mikule and S. Doxsey, personal communication). What causes the G1 arrest?

How centrosome defects lead to a G1 arrest resists simple explanation, because many seemingly disparate experimental perturbations of the centrosome produce the same result. We do not know if all these perturbations produce different defects or if they all lead to a common defect that is sensed by the cell. Below, I briefly outline some theoretical, and admittedly speculative, possibilities for this G1 arrest in the hope of stimulating thought and discussion. The centrosome as an integral part of the mechanisms that drive G1 progression. In principle, experimental perturbations of the centrosome could impair a pathway directly involved in G1 progression or entry into S phase. This notion is supported by the recent discovery that cyclin E contains a centrosome localization signal (CLS) motif 29. Importantly, expression of the CLS peptide not only prevented cyclin E from binding to the centrosome, but also blocked the initiation of S phase. This implies that binding of cyclin E to the centrosome is required for DNA synthesis. This intriguing possibility, however, must be considered in light of two sorts of observations. First, cyclin-E1- or cyclin-E2-knockout mice are viable, although the cyclin E2–/– males are sterile30. CyclinE1,E2-double knockout mice progress to day 10 of embryonic development before placental defects occur. So, unless maternal pools of cyclin E2 support development to day 10, the binding of cyclin E to the centrosome is not required for entry into S phase. Nevertheless, this does not preclude the possibility that the binding to the centrosome of cyclin A (which contains a putative CLS motif 29) or other proteins containing CLS-like motifs is important for entry into S phase. Second, unless the previously described knock-downs of various centrosomal proteins block the binding of CLS-containing proteins to the centrosome, there must be additional causes for the centrosome-based G1 arrest. In a different vein, the loss of centrioles or individual centrosome-associated proteins could loosen or disrupt centrosome structure in a functionally equivalent manner. As with a chain, a break in any link can profoundly compromise the integrity and function of the structure. If the centrosome functions as an intermediate that brings regulatory proteins, their upstream regulators, and substrates into close proximity to

VOLUME 6 | SEPTEMBER 2005 | 745

PERSPECTIVES G1 in a p53-dependent manner18,35,36. This is correlated with the displacement of active p53 from the centrosomes37,38. However, the notion that the G1 arrest is uniquely the consequence of disrupted microtubule anchorage at the centrosome is undercut by the report that the knock-down of centriolin produces cells that arrest in G1 without apparent loss of microtubule organization23.

a Experimental microsurgery Microneedle

Centrosome

Nucleus

Cytoplast

Karyoplast

b Control microsurgery

Figure 1 | Diagrammatic representation of the microsurgical operation. a | Cells in which the centrosome is slightly separated from the nucleus are selected and the microneedle is pressed down between the nucleus and the centrosome. After the operation, the enucleate cytoplast containing the centrosome is wiped off the coverslip and the nucleated karyoplast is viewed by continuous time-lapse video cinematography. The karyoplast assembles a microtubule organizing centre and grows in size. b | Control microsurgery. Removal of a large area of cytoplasm does not compromise the ability of the cell to continue proliferating. Modified with permission from REF. 70 © (1992) Elsevier Science.

interact31–33, loosening of the centrosomal fabric might prevent essential reactions from occurring efficiently. However, this line of reasoning must be tempered by observations that cell-cycle progression through G1 does not absolutely depend on the presence of centrioles. For example, HeLa cells, which have suppressed p53 levels, progress through G1 after centrosome ablation in mitosis34. Nevertheless, when the p53 pathway is intact, the cell might be able to sense perturbations in the centrosome structure. Signalling by the centrosome? Perhaps there is a signalling pathway of centrosomal origin that responds to defects in the structure or function of the centrosome by inducing a G1 arrest. In principle, such a response could arise from the lack of a positive signal that promotes G1 progression28 or the activation of a negative, or ‘stop’, signal that enforces a cell-cycle arrest. The fact that zygotes and HeLa cells progress through G1 without centrioles shows that a positive signal originating from an intact centrosome is not required for G1 progression. Should the G1 arrest be caused by a negative signal, such a signal cannot originate from the centrioles, because their removal does not relieve the G1 arrest in

746 | SEPTEMBER 2005

| VOLUME 6

normal cells3,4. The logical place to look for the origin of such a putative negative signal would be the PCM, or — in the case of karyoplasts — the PCM components that come together to function as a MTOC after mitosis. Alternatively, the defect sensed by the cell could be in the interaction between the centrosome and the microtubules that it nucleates. Post-mitotic karyoplasts and cells in which PCM1 has been knocked down assemble an extensive array of microtubules that are usually less well organized into a radial array near the centrally located MTOC when compared to control cells (FIG. 2). This implies that there is a reduction or alteration of microtubule minus-end anchorage at the cell centre3,4,21,22. If interactions between microtubule minus ends (the ends that are closest to the MTOC) and nucleating and/or anchoring complexes participate in a hypothetical signalling activity, then a reduction in microtubule anchoring could produce a defect that the cell could sense. The possible importance of microtubule presence and function at the end of mitosis for G1 progression is consistent with findings that cells coming out of prolonged mitosis without cytoplasmic microtubules, but with otherwise complete centrosomes, arrest in

Moving away from the centrosome. Attention so far has been centered on what might or might not be happening at the centrosome, but this is not the only place to look for signalling that could stop the cell cycle. Perhaps centrosomal perturbations compromise interactions that occur at the plus ends of microtubules that are nucleated by the G1 centrosome (the growing tips distal to the centrosome). The tips of cytoplasmic microtubules are intimately associated with focal adhesions, and repeatedly grow back and forth to these sites of substrate adhesion39–41. Loss of microtubule anchorage at the cell centre owing to the various experimental interventions described above might alter the interactions of microtubule tips with focal adhesions, or influence the functional properties of the actin cytoskeleton42. This, in turn, could alter signalling activities at focal adhesions that promote cell-cycle progression. The notion that altering microtubule-tip dynamics can influence cell-cycle progression is supported by a report that paclitaxel leads to G1 arrest in non-transformed cells, and this correlates with elevated p27Kip1 levels43. In addition, when microtubules are disrupted shortly after release from serum starvation, rat and human fibroblasts arrest or delay in G1 owing to the accumulation of transcriptionally active p53 REFS 43,44. This is thought to result from the upregulation of the integrin–Raf–extracellular-signalregulated-kinase (ERK) signalling pathway owing to the formation of more, and larger, focal adhesions44. Last, alterations in the spatial organization of the cytoskeleton, caused by growing cells on small areas of substrate to constrain cell spreading, leads to a G1 arrest 45,46 with elevated levels of p27Kip1, depressed cyclin D levels and hypophosphorylated Rb. This arrest is independent of the reduced contact area between the cell and its substrate, because growth of cells on an array of much smaller spots of substrate allows spreading of the cells and G1 progression, although the total area of cell–substrate contact is equivalent to those that are spatially constrained45,46. Together, these observations

www.nature.com/reviews/molcellbio

PERSPECTIVES effects this has on cytoskeletal organization. Conceivably, both possibilities are true. The discovery of the involvement of the centrosome in G1 cell-cycle progression opens a new chapter in our appreciation for the complexity of the interrelationship between centrosomes and the cell cycle. Although the fact that so many seemingly diverse centrosomal perturbations all cause a G1 arrest is perplexing, the new mysteries to be addressed continue our fascination with the centrosome and its key role in so many aspects of the cell cycle. Greenfield Sluder is at the Department of Cell Biology, University of Massachusetts Medical School, 377 Plantation Street, Worcester, Massachusetts 01605, USA. e-mail: [email protected] doi:10.1038/nrm1712 1. 2.

Figure 2 | Centrosomal proteins and microtubules in post-mitotic karyoplasts. Two post-mitotic karyoplasts, arrested in G1 phase, stained for microtubules and centrosomal proteins. a | This karyoplast was followed in vivo, then fixed, and immunostained for α-tubulin (green) and γ-tubulin (red) ~4 hours after mitosis (cleavage failed). The extensive interphase array of microtubules is centred on a single concentration of γ-tubulin. The microtubules, although not tightly focused where the γ-tubulin concentration is highest, have a radial organization that is within the range that is shown by control cells. b | Another post-mitotic karyoplast fixed and immunostained for α-tubulin and pericentrin (red) ~64 hours after mitosis (cleavage failed). The extensive array of cytoplasmic microtubules does not seem to be well organized or tightly associated with the pericentrin next to the two nuclei (blue). Scale bar, 10 µm. Image reproduced with permission from REF. 3 © (2001) American Assocation for the Advancement of Science.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

raise the possibility that centrosomal perturbations lead to a G1 arrest by altering the spatial organization and therefore the function of the cytoskeleton.

13.

14.

Concluding remarks

In seeking to explain how centrosome defects induce a G1 arrest, it is difficult to express a preference for one line of reasoning over all others, because disruption of centrosomal structure could produce different insults, any one of which could trigger a G1 arrest. Going forward, we would like to know if there is something special about the centrosome itself — for example, a putative signalling activity — that prevents the cell from activating p53, or whether a centrosomal defect is sensed only by the downstream

NATURE REVIEWS | MOLECULAR CELL BIOLOGY

15.

16. 17.

18.

19.

Karsenti, E. & Vernos, I. The mitotic spindle: a self-made machine. Science 294, 543–547 (2001). Wadsworth, P. & Khodjakov, A. E pluribus unum: towards a universal mechanism for spindle assembly. Trends Cell Biol. 14, 413–419 (2004). Hinchcliffe, E. H., Miller, F. J., Cham, M., Khodjakov, A. & Sluder, G. Requirement of a centrosomal activity for cell cycle progression through G1 into S phase. Science 291, 1547–1550 (2001). Khodjakov, A. & Rieder, C. L. Centrosomes enhance the fidelity of cytokinesis in vertebrates and are required for cell cycle progression. J. Cell Biol. 153, 237–242 (2001). Piel, M., Nordberg, J., Euteneuer, U. & Bornens, M. Centrosome-dependent exit of cytokinesis in animal cells. Science 291, 1550–1553 (2001). Rieder, C. L., Faruki, S. & Khodjakov, A. The centrosome in vertebrates: more than a microtubule-organizing center. Trends Cell Biol. 11, 413–419 (2001). Murray, A. W. & Kirschner, M. W. Cyclin synthesis drives the early embryonic cell cycle. Nature 339, 275–280 (1989). Sluder, G., Miller, F. J. & Rieder, C. L. Reproductive capacity of sea urchin centrosomes without centrioles. Cell Motil. Cytoskeleton 13, 264–273 (1989). Snyder, J. A. & McIntosh, J. R. Initiation and growth of microtubules from mitotic centers in lysed mammalian cells. J. Cell Biol. 67, 744–760 (1975). Kuriyama, R. & Borisy, G. G. Microtubule-nucleating activity of centrosomes in Chinese hamster ovary cells is independent of the centriole cycle but coupled to the mitotic cycle. J. Cell Biol. 91, 822–826 (1981). Vandre, D. D. & Borisy, G. G. in Mitosis: Molecules and Mechanisms (eds Hyams, J. S. & Brinkley, B. R.) 39–76 (Academic Press, New York, 1989). Centonze, V. E. & Borisy, G. G. Nucleation of microtubules from mitotic centrosomes is modulated by a phosphorylated epitope. J. Cell Sci. 95, 405–411 (1990). Vandre, D. D., Feng, Y. & Ding, M. Cell cycle-dependent phosphorylation of centrosomes: localization of phosphopeptide specific antibodies to the centrosome. Microsc. Res. Tech. 49, 458–466 (2000). Khodjakov, A. & Rieder, C. L. The sudden recruitment of γ-tubulin to the centrosome at the onset of mitosis and its dynamic exchange throughout the cell cycle, do not require microtubules. J. Cell Biol. 146, 585–596 (1999). Young, A., Dictenberg, J. B., Purohit, A., Tuft, R. & Doxsey, S. J. Cytoplasmic dynein-mediated assembly of pericentrin and γ-tubulin onto centrosomes. Mol. Biol. Cell 11, 2047–2056 (2000). Sluder, G. in Centrosomes in Development and Disease (ed. Nigg, E.) 167–189 (Wiley-VCH, Weinheim, 2004). Maniotis, A. & Schliwa, M. Microsurgical removal of centrosomes blocks cell reproduction and centriole generation in BSC-1 cells. Cell 67, 495–504 (1991). Vogel, C., Kienitz, A., Hofmann, I., Muller, R. & Bastians, H. Crosstalk of the mitotic spindle assembly checkpoint with p53 to prevent polyploidy. Oncogene 23, 6845–6853 (2004). Margolis, R. L., Lohez, O. D. & Andreassen, P. R. G1 tetraploidy checkpoint and the suppression of tumorigenesis. J. Cell. Biochem. 88, 673–683 (2003).

20. Uetake, Y. & Sluder, G. Cell cycle progression after cleavage failure: mammalian somatic cells do not possess a ‘tetraploidy checkpoint’. J. Cell Biol. 165, 609–615 (2004). 21. Dammermann, A. & Merdes, A. Assembly of centrosomal proteins and microtubule organization depends on PCM-1. J. Cell Biol. 159, 255–266 (2002). 22. Balczon, R., Simerly, C., Takahashi, D. & Schatten, G. Arrest of cell cycle progression during first interphase in murine zygotes microinjected with anti-PCM-1 antibodies. Cell Motil. Cytoskeleton 52, 183–192 (2002). 23. Gromley, A. et al. A novel human protein of the maternal centriole is required for the final stages of cytokinesis and entry into S phase. J. Cell Biol. 161, 535–545 (2003). 24. Keryer, G. et al. Dissociation the centrosomal matrix protein AKAP450 from centrioles impairs centriole duplication and cell cycle progression. Mol. Biol. Cell 14, 2436–2446 (2003). 25. Augustin, A. et al. PARP-3 localizes preferentially to the daughter centriole and interferes with the G1/S cell cycle progression. J. Cell Sci. 116, 1551–1562 (2003). 26. Patzke, S. et al. Identification of a novel centrosome/ microtubule-associated coiled-coil protein involved in cell-cycle progression and spindle organization. Oncogene 24, 1159–1173 (2005). 27. Quintyne, N. J. & Schroer, T. A. Distinct cell cycledependent roles for dynactin and dynein at centrosomes. J. Cell Biol. 159, 245–254 (2002). 28. Doxsey, S., Zimmerman, W. & Mikule, K. Centrosome control of the cell cycle. Trends Cell Biol. 15, 303–311 (2005). 29. Matsumoto, Y. & Maller, J. L. A centrosomal localization signal in cyclin E required for Cdk2-independent S phase entry. Science 306, 885–888 (2004). 30. Geng, Y. et al. Cyclin E ablation in the mouse. Cell. 114, 431–443 (2003). 31. Palazzo, R. E., Vogel, J. M., Schnackenberg, B. J., Hull, D. R. & Wu, L. in Curr. Top. Dev. Biol. (eds Palazzo, R. E. & Schatten, G.) 449–470 (Academic Press, San Diego, 2000). 32. Doxsey, S. J. Centrosomes as command centres for cellular control. Nature Cell Biol. 3, 105–108 (2001). 33. Fry, A. & Hames, R. in Centrosomes in Development and Disease (ed. Nigg, E.) 143–166 (Wiley-VCH, Weinheim, 2004). 34. La Terra, S. et al. The de novo centriole assembly pathway in HeLa cells: cell cycle progression and centriole assembly/maturation. J. Cell Biol. 168, 713–722 (2005). 35. Lanni, J. S. & Jacks, T. Characterization of the p53dependent postmitotic checkpoint following spindle disruption. Mol. Cell. Biol. 18, 1055–1064 (1998). 36. Casenghi, M. et al. p53-independent apoptosis and p53-dependent block of DNA rereplication following mitotic spindle inhibition in human cells. Exp. Cell Res. 250, 339–350 (1999). 37. Ciciarello, M. et al. p53 displacement from centrosomes and p53-mediated G1 arrest following transient inhibition of the mitotic spindle. J. Biol. Chem. 276, 19205–19213 (2001). 38. Tritarelli, A. et al. p53 localization at centrosomes during mitosis and postmitotic checkpoint are ATM-dependent and require serine 15 phosphorylation. Mol. Biol. Cell 15, 3751–3757 (2004). 39. Kaverina, I., Krylyshkina, O. & Small, J. V. Microtubule targeting of substrate contacts promotes their relaxation and dissociation. J. Cell Biol. 146, 1033–1044 (1999). 40. Kaverina, I., Rottner, K. & Small, J. V. Targeting, capture, and stabilization of microtubules at early focal adhesions. J. Cell Biol. 142, 181–190 (1998). 41. Krylyshkina, O. et al. Nanometer targeting of microtubules to focal adhesions. J. Cell Biol. 161, 853–859 (2003). 42. D’Addario, M., Arora, P. D., Ellen, R. P. & McCulloch, C. A. Regulation of tension-induced mechanotranscriptional signals by the microtubule network in fibroblasts. J. Biol. Chem. 278, 53090–53097 (2003). 43. Trielli, M. O., Andreassen, P. R., Lacroix, F. B. & Margolis, R. L. Differential taxol-dependent arrest of transformed and nontransformed cells in the G1 phase of the cell cycle, and specific-related mortality of transformed cells. J. Cell Biol. 135, 689–700 (1996). 44. Sablina, A. A., Chumakov, P. M., Levine, A. J. & Kopnin, B. P. p53 activation in response to microtubule disruption is mediated by integrin–Erk signaling. Oncogene 20, 899–909 (2001).

VOLUME 6 | SEPTEMBER 2005 | 747

PERSPECTIVES

45. Huang, S., Chen, C. S. & Ingber, D. E. Control of cyclin D1, p27Kip1, and cell cycle progression in human capillary endothelial cells by cell shape and cytoskeletal tension. Mol. Biol. Cell 9, 3179–3193 (1998). 46. Ingber, D. E. Tensegrity II: how structural networks influence cellular information processing networks. J. Cell Sci. 116, 1397–1408 (2003). 47. Wheatley, D. N. The Centriole: A Central Enigma of Cell Biology (Elsevier Biomedical Press, Amsterdam, 1982). 48. Paintrand, M., Moudjou, M., Delacroix, H. & Bornens, M. Centrosome organization and centriole architecture: their sensitivity to divalent cations. J. Struct. Biol. 108, 107–128 (1992). 49. Rieder, C. L. & Borisy, G. G. The centrosome cycle in PtK2 cells: Asymmetric distribution and structural changes in the pericentriolar material. Biol. Cell. 44, 117–132 (1982). 50. Sluder, G. & Rieder, C. L. Centriole number and the reproductive capacity of spindle poles. J. Cell Biol. 100, 887–896 (1985). 51. Vorobjev, I. A. & Nadezhdina, E. S. The centrosome and its role in the organization of microtubules. Int. Rev. Cytol. 106, 227–293 (1987). 52. Bornens, M., Paintrand, M., Berges, J., Marty, M. C. & Karsenti, E. Structural and chemical characterization of isolated centrosomes. Cell Motil. Cytoskeleton 8, 238–249 (1987). 53. Mogensen, M. in Centrosomes in Development and Disease (ed. Nigg, E.) 299–319 (Wiley-VCH, Weinheim, 2004). 54. Moritz, M., Rice, L. & Agard, D. in Centrosomes in Development and Disease (ed. Nigg, E.) 27–41 (WileyVCH, Weinheim, 2004). 55. Doxsey, S. Re-evaluating centrosome function. Nature Rev. Mol. Cell Biol. 2, 688–698 (2001).

748 | SEPTEMBER 2005

| VOLUME 6

56. Lange, B. M. Integration of the centrosome in cell cycle control, stress response and signal transduction pathways. Curr. Opin. Cell Biol. 14, 35–43 (2002). 57. Wilkinson, C., Andersen, J., Mann, M. & Nigg, E. in Centrosomes in Development and Disease (ed. Nigg, E.) 125–142 (Wiley-VCH, Weinheim, 2004). 58. Andersen, J. S. et al. Proteomic characterization of the human centrosome by protein correlation profiling. Nature 416, 570–574 (2003). 59. Preble, A. M., Giddings, T. M., Jr & Dutcher, S. K. Basal bodies and centrioles: their function and structure. Curr. Top. Dev. Biol. 49, 207–233 (2000). 60. Bornens, M. Centrosome composition and microtubule anchoring mechanisms. Curr. Opin. Cell Biol. 14, 25–34 (2002). 61. Piel, M., Meyer, P., Khodjakov, A., Rieder, C. L. & Bornens, M. The respective contributions of the mother and daughter centrioles to centrosome activity and behavior in vertebrate cells. J. Cell Biol. 149, 317–330 (2000). 62. Pelletier, L. et al. The Caenorhabditis elegans centrosomal protein SPD-2 is required for both pericentriolar material recruitment and centriole duplication. Curr. Biol. 14, 863–73 (2004). 63. Massague, J. G1 cell-cycle control and cancer. Nature 432, 298–306 (2004). 64. Kastan, M. B. & Bartek, J. Cell-cycle checkpoints and cancer. Nature 432, 316–323 (2004). 65. Ding, Q. et al. p27Kip1 and cyclin D1 are necessary for focal adhesion kinase (FAK) regulation of cell cycle progression in glioblastoma cells propagated in vitro and in vivo in the scid mouse brain. J. Biol. Chem. 280, 6802–6815 (2005). 66. Lowe, S. W., Cepero, E. & Evan, G. Intrinsic tumour suppression. Nature 432, 307–315 (2004).

67. Sherr, C. J. & McCormick, F. The RB and p53 pathways in cancer. Cancer Cell 2, 103–112 (2002). 68. Bode, A. M. & Dong, Z. Post-translational modification of p53 in tumorigenesis. Nature Rev. Cancer 4, 793–805 (2004). 69. Sherr, C. J. & Roberts, J. M. Living with or without cyclins and cyclin-dependent kinases. Genes Dev. 18, 2699–2711 (2004). 70. Sluder, G. Double or nothing. Curr. Biol. 2, 243–245 (1992).

Acknowledgements I thank S. Doxsey, A. Khodjakov, A. Krzywicka, J. Nordberg and Y. Uetake for helpful discussions. C. English provided invaluable help in the preparation of the figures. I apologize to those whose work could not be cited owing to space limitations.

Competing interests statement The author declares no competing financial interests.

Online links DATABASES The following terms in this article are linked online to: Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query. fcgi?db=gene cyclin E Swiss-Prot: http://www.expasy.ch/sprot CDK2 | p21 | p53 FURTHER INFORMATION Greenfield Sluder’s laboratory: http://www.umassmed.edu/ cellbio/faculty/sluder.cfm Access to this interactive links box is free online.

www.nature.com/reviews/molcellbio