Control of the centriole and centrosome cycles by ... - Nature

0 downloads 0 Views 270KB Size Report
Keywords: ubiquitin ligases; ubiquitination; centrosome duplication. Introduction. The role of the centriole in organizing the cell's cytoskeleton and its mechanism ...
Oncogene (2002) 21, 6209 – 6221 ª 2002 Nature Publishing Group All rights reserved 0950 – 9232/02 $25.00 www.nature.com/onc

Control of the centriole and centrosome cycles by ubiquitination enzymes David V Hansen1,2, Jerry Y Hsu1,2, Brett K Kaiser1,2, Peter K Jackson*,1,2 and Adam G Eldridge1,2 1

Programs in Chemical Biology and Cancer Biology, Stanford University School of Medicine, 300 Pasteur Drive, Stanford, California, CA 94305-5324, USA; 2Department of Pathology, Stanford University School of Medicine, 300 Pasteur Drive, Stanford, California, CA 94305-5324, USA

Oncogene (2002) 21, 6209 – 6221. doi:10.1038/sj.onc. 1205824 Keywords: ubiquitin ligases; ubiquitination; centrosome duplication

Introduction The role of the centriole in organizing the cell’s cytoskeleton and its mechanism of duplication have been long-standing puzzles for cell biologists. Whereas the semi-conservative replication of chromosomes was established by Meselson and Stahl (1958), the likely parallels for semi-conservative duplication of the centrioles remain fuzzy. Further considering this parallel, molecular studies of chromosomal replication have begun to uncover how cell cycle regulators including cyclin-dependent kinases and ubiquitin ligases ensure that chromosomes replicate once-and-only-once per cell cycle (Blow and Hodgson, 2002; Dutta and Bell, 1997). The obvious need to maintain accurate control of centrosome number and thereby ensure spindle bipolarity would suggest that a similar once-and-only once control restricts the centrosome cycle. Studies over the last decade on the budding and fission yeast spindle pole bodies (SPB) and the animal cell centrosome have defined a growing parts list of conserved components, as well as those specific to fungi or animals. The functional connection between the centrosome duplication cycle and the regulatory mechanisms controlling the chromosome duplication cycle suggested that semi-conservative replication for both centrosomes and chromosomes might be linked by these global timing mechanisms. In 1999, a series of studies demonstrated that cyclin-dependent kinases and both the SCF and APC ubiquitin ligases – cell cycle regulators already well established in control of the chromosome replication cycle – also had fundamental roles in controlling the centrosome cycle (Freed et al., 1999; Hinchcliffe et al., 1999; Lacey et al., 1999; Meraldi et al., 1999; Vidwans et al., 1999). Since then, a number of other cell cycle regulators have been directly implicated in the centrosome cycle,

*Correspondence: PK Jackson; E-mail: [email protected]

many of which are described in the accompanying reviews. Here we will focus on the role of ubiquitin ligases in controlling the centrosome cycle, considering both those known or postulated core centrosomal factors that are directly ubiquitinated, as well as ubiquitination of specific cell cycle regulators – including kinases and ubiquitin ligases themselves – that more globally control the centrosome cycle. First, we will review the biochemistry of ubiquitin ligases, and then return to the role of these enzymes in the various phases of the centrosome cycle. The biochemistry of ubiquitin ligases The addition of polymeric chains of ubiquitin to lysine side chain residues of specific proteins is a sufficient signal to target the proteins to the 26S proteasome for proteolytic destruction. Ubiquitination of a substrate requires a ubiquitin enzyme shuttle using E1, E2, and E3 enzymes to direct the formation of ubiquitin chains on specific substrates. Among these ubiquitination enzymes, the E3 ubiquitin ligases appear to provide the critical elements of specificity that direct the formation of polyubiquitin chains on protein targets (See Appendix). Ubiquitin-dependent proteolysis is a widely used mechanism for regulating protein function in processes ranging from cell cycle and developmental switches to homeostatic control of environmental sensing (see reviews: Deshaies, 1999; Jackson et al., 2000). Briefly, after activation by an E1 enzyme, the E2 ubiquitin-conjugating enzyme and E3 ubiquitin ligase cooperate to assemble the polyubiquitin chain on a protein substrate. In studies so far, the E3 ubiquitin ligase uses protein-protein interaction domains outside the catalytic domain to bind substrate (Figure 1). To date, there are two major structural classes of E3 ubiquitin ligases, although newer studies suggest additional domains may possess E3 activity (Hatakeyama et al., 2001; Jiang et al., 2001; Lu et al., 2002; Patterson, 2002). Of the two major classes, the HECT domain proteins are a modest sized family of proteins with no currently known connections to centrosome biology. They have been recently reviewed (Jackson et al., 2000) and will not be addressed further here. In the largest class of E3 ubiquitin ligases, the so-called RING (for ‘Really Interesting New Gene’) finger

Ubiquitination and the centrosome cycle DV Hansen et al

6210

Figure 1 A summary of RING finger ubiquitin ligases. Presented are representative examples of RING finger ubiquitin ligases discussed in the text

proteins use a zinc-binding function to stimulate catalysis of ubiquitin chain formation and may combine substrate binding and catalytic domains within a single polypeptide. Alternatively, a smaller RING finger polypeptide can be associated within the context of a multi-protein complex with separate subunits containing (i) the RING finger and associated catalytic polypeptides; and (ii) substrate binding (or ‘adapter’) functions. SCF (Skp1, Cullin, F-box) complexes The SCF class of ubiquitin ligases contains at least four proteins: Skp1, the cullin protein Cul1, the Ring Finger protein Rbx1, and an F-box protein, a substrate binding protein (see Figure 1). Other proteins associated with SCF ubiquitin ligases are still being defined, but biochemical reconstitution of these four components plus E1 and E2 enzymes is sufficient for ubiquitin chain formation on an appropriately modified substrate. SCF substrates are bound directly by the F-box proteins, which contain a *45 amino acid motif called an F-box and bind to substrates through proteinprotein interaction domains (Jackson et al., 2000). The F-box is required for binding to Skp1, which in turn associates with Cul1. The cullin protein acts as a scaffold, bringing the E2 ubiquitin-conjugating enzyme into close proximity with the rest of the complex Oncogene

(Zheng et al., 2002). There are at least six human cullins, including Cul1-5 and Apc2. Apc2 functions in a known E3 complex (discussed below). The remaining cullins are likely to organize ubiquitin ligase complexes similar to SCF and Cul1. Human Roc1/Rbx1/Hrt1, a protein containing a RING-H2 finger domain, appears to promote association of the cullin protein with the E2 enzyme and to enhance ubiquitin ligase activity (Kamura et al., 1999; Ohta et al., 1999; Skowyra et al., 1999; Tan et al., 1999). The RING-H2 is one of a class of RING finger proteins containing an octet of cysteine and histidine residues that participate in E2 binding and catalysis. The F-box adapter proteins direct ubiquitination of diverse substrates. There are 16 F-box proteins in S. cerevisiae, *100 in the complete C. elegans genome, and more than 50 described so far in vertebrates (Jackson et al., 2000). In the highly studied SCFCdc4, SCFSkp2 and SCFb7TrCP complexes, substrate phosphorylation is required for binding the F-box protein. In the budding yeast SCFCdc4 complex, a WD40repeat protein, Cdc4, binds each of the known substrates (Sic1p, Cdc6p, Gcn4p) in a phosphorylation-dependent fashion. The SCFb7TrCP recognizes a specific phosphoserine motif (DSGfXS) found in the regulatory proteins IkBa and b-catenin. Skp2, a leucine-rich repeat F-box protein, also recognizes only the phosphorylated form of the Cdk regulator p27Kip1. Although phosphorylation is clearly required for the

Ubiquitination and the centrosome cycle DV Hansen et al

6211

binding of specific substrates to their F-box adapters, it is not clear whether phosphorylation merely serves to improve the affinity of substrate to F-box, or has another role. In part, the insolubility of several F-box proteins has limited these types of biochemical studies. The anaphase promoting complex (APC) ubiquitin ligase The APC was the first multicomponent ubiquitin ligase described and is required for the degradation of substrates controlling the metaphase-to-anaphase transition and for the destruction of cyclin B to allow exit from mitosis. Similar to the SCF complex, the APC contains a cullin homolog, Apc2, and a RING-H2 finger protein similar to Rbx1, called Apc11. Reconstitution studies show that these two proteins form the catalytic core of the APC and are sufficient for ubiquitin chain assembly in vitro (Gmachl et al., 2000; Tang et al., 2001). The APC associates with two WD repeat-containing adapter proteins, Cdc20/Fizzy and Cdh1/Hct1/Fizzyrelated. Both bind the APC and have recently been shown to be important for substrate binding (Burton and Solomon, 2000; Hilioti et al., 2001; Pfleger et al., 2001a; Schwab et al., 2001), although the specific determinants of binding remain unclear. The APC in humans and yeast also contains other components (at least 11 polypeptides in total), although the role of these additional components and the possible heterogeneous composition of the APC are not understood. Two general destruction signals have been identified in substrates targeted for destruction by the APC, the destruction box (or D-box) and the KEN box (Pfleger and Kirschner, 2000). The 9 amino acid D-box is found in all the known APCCdc20 substrates, but also in some APCCdh1 substrates. In addition, the KEN box was described as a transposable, 7 amino acid motif that appears to target substrates specifically to APCCdh1. Thus, Cdh1 can target ubiquitination of both D-box and KEN box containing substrates, including Cdc20 itself. Although the D-box and KEN box motifs appear to be necessary for Cdc20and Cdh1-activated substrate ubiquitination, it is not clear whether Cdc20 or Cdh1 are sufficient for direct binding of these recognition sequences or whether additional sequences or factors can contribute to adapter binding. Ubiquitination of APC substrates is regulated both by phosphorylation of the APC core components and by phosphorylation of at least the Cdh1 adapter protein (Kramer et al., 2000). The APC is also controlled by inhibitory proteins that bind to the Cdc20 and Cdh1 adapter proteins. These include the spindle assembly checkpoint protein Mad2, which binds Cdc20 and inhibits APC activity following spindle damage (Shah and Cleveland, 2000), a closely related protein, called Mad2B, which binds and inhibits Cdh1 (Chen and Fang, 2001; Pfleger et al., 2001b), and the Emi1 protein (Hsu et al., 2002; Reimann et al., 2001a,b), which binds both Cdc20 and Cdh1 to inhibit their activity during interphase.

Single polypeptide RING finger E3 proteins A number of E3 ubiquitin ligases have been found to contain an N-terminal substrate binding domain and a C-terminal RING finger within the same polypeptide. In one of the better-studied examples, the p53 regulator Mdm2 uses an N-terminal p53-binding function and a C-terminal RING-H2 finger to direct the ubiquitination of the tumor suppressor p53 (Honda et al., 1997). An overview of the events of centriole and centrosome duplication throughout the cell cycle As cells exit mitosis and begin cell growth during G1 phase, each daughter cell has received one centrosome containing a paired centriole (or ‘duplosome’), which serves as the microtubule organizing center (MTOC) for the cell (Figure 2). The centriole appears to organize the pericentriolar material (PCM), including a ‘centromatrix’ and many copies of the gamma tubulin ring complex (g-TuRC), which nucleates microtubule outgrowth. In many animal cells, these centrioles split around the time of the G1/S transition, when DNA replication begins, but this is by no means universal (Vidwans et al., 1999). Upon S phase entry, the two separated centrioles (‘mother’ centrioles) each nucleate the outgrowth of a procentriole (‘daughter’), the elongation of which occurs during S and G2 phases and is completed in mitosis. Because the original centrosome was composed of a mother-daughter centriole pair, one of the split centrioles is an older mother and the other a younger or new mother. During G1 phase, the centriolar pair has MTOC activity, which is thought to be generated by the older mother centriole (Hinchcliffe and Sluder, 2001). After centriolar splitting and the outgrowth of procentrioles from the older and new mothers, the new mother must acquire MTOC activity as a prerequisite to forming the other pole of the mitotic spindle. This functional maturation late in S/G2 phase is correlated with the formation of the distal appendages, which emanate from the elongating mother centriole (Bornens, 2002; Marshall, 2001), and appear in immunoelectron micrographs as the site of microtubule outgrowth (Paintrand et al., 1992). Just before mitotic entry, the elongation of the procentrioles is nearly completed and the two centriolar pairs organize a more elongated PCM. The centriolar pairs appear to be tethered by a fibrous bridge (Paintrand et al., 1992), which also contains SCF components including the protein Skp1 (Freed et al., 1999). Near the time of mitotic entry, the centrosome splits – possibly by dissolution of this bridge – and the PCM organized around each centriole increases dramatically in size (Khodjakov and Rieder, 1999). This increase is not dependent on microtubules, but possibly on the activation of mitotic kinases. These coupled events provide the mechanism to build the bipolar mitotic spindle. Oncogene

Ubiquitination and the centrosome cycle DV Hansen et al

6212

Figure 2 A schematic of the centriole and centrosome cycles

Following spindle assembly and the triggering of chromosome segregation, a set of chromosomes and one centrosome are pulled toward each pole of the elongating cell and cytokinesis is triggered between the spindle poles to ensure that each daughter cell receives both a full complement of chromosomes and a single centrosome (Glotzer, 2001). Recent analysis from the Bornens’ lab suggests that the centrioles are considerably more dynamic than previously thought and that the mother centriole plays a critical role in triggering cytokinesis (Piel et al., 2000, 2001). Thus, our current model for the centriole and centrosome cycle may require some fundamental revision. The specific order of morphological events in the centrosome cycle suggests an orderly pathway. The close linkage of centrosomal events in step with cell cycle transitions has suggested the coupling of cell cycle regulators including ubiquitin ligases and cyclindependent kinases to the centrosome cycle. We will now examine the role of these ubiquitin ligases in specific transitions within the centrosome cycle. Ubiquitin-dependent proteolysis in centriole separation The first step of the centrosome duplication cycle begins with centriolar splitting at the G1/S transition, thereby allowing subsequent procentriolar outgrowth. The cyclin-dependent kinase Cdk2 has been shown to be important for centrosome duplication. Specifically, cyclin E/Cdk2 activity is necessary for centrosome duplication in Xenopus embryos and centriole separation in extracts from these embryos (Lacey et al., 1999; Oncogene

Hinchcliffe et al., 1999). Similar requirements for Cdk2 activity in centrosome duplication have been shown in somatic cells (Matsumoto et al., 1999), although cyclin A/Cdk2 instead of cyclin E/Cdk2 may be the cyclindependent kinase necessary for centrosomal duplication in somatic cells (Meraldi et al., 1999). The Cdk2 activity necessary for centriole separation may first play a positive role in phosphorylating some factor necessary for centriole splitting, similar to a role it is thought to play in triggering DNA replication (Dutta and Bell, 1997). Additionally, Cdk2 may phosphorylate and inactivate an inhibitory protein, possibly by triggering its destruction. By analogy, the Cdk2 kinase triggers destruction of SCF substrates like the cyclin E/Cdk2 inhibitor p27Kip1 to promote the G1/S transition. At this point, we have only one candidate for the cyclin E phosphorylation target at the centrosome, nucleophosmin (Okuda et al., 2000), but additional studies will be necessary to better understand the implications of this work. It does appear that there must be targets that are not simply Cdk2 inhibitors (discussed below). In an interesting twist, the kinase Mps1 is a positive regulator of centrosome duplication and an in vitro substrate of Cdk2. Inhibition of Cdk2-dependent phosphorylation results in proteasome-dependent degradation of murine Mps1 (Fisk and Winey, 2001). Thus, it is possible that the Mps1 dephosphorylation allows its recognition by a yet unidentified E3 ubiquitin ligase. A specific role for SCF ubiquitin ligases in the centrosome duplication cycle was suggested by the discovery that the proteins Skp1 and Cul1, two core components of the SCF ubiquitin ligase, localize to the

Ubiquitination and the centrosome cycle DV Hansen et al

6213

centrosome by indirect immunofluorescence and immunoelectron microscopy (Freed et al., 1999; Gstaiger et al., 1999). These proteins were also highly enriched in purified centrosomes and the centrosome-localized cullin Cul1 was quantitatively modified by Nedd8, a ubiquitin-like molecule necessary for SCF ubiquitination activity. This enrichment of Nedd8-modified Cul1 at the centrosome suggested that the centrosomally localized SCF complex was highly active. Centriole separation was inhibited by neutralizing antibodies to Skp1 or Cul1 in Xenopus extracts and could not be restored by addition of recombinant cyclin E/Cdk2 (Freed et al., 1999), suggesting that SCF-mediated proteolysis of a target other than a cyclin-dependent kinase inhibitor is required for centriole separation. Inhibition of the proteasome by addition of the proteasome inhibitor MG132 also resulted in inhibition of centriole separation in vitro and prevented centrosome duplication when injected into Xenopus embryos, further suggesting a role for proteolysis at this stage of the centrosome cycle. The Anaphase Promoting Complex (APC) also has a role in controlling events at the G1/S transition, notably by inhibiting the accumulation of cyclin A, which as mentioned before is important for triggering centrosome duplication in somatic cells (Meraldi et al., 1999). Thus, the ability of the APC to restrain cyclin A accumulation may be important in preventing premature centriole separation. Recent studies show that the human Emi1 protein can inhibit the APCCdh1 complex and thereby stabilize cyclin A to drive entry into S phase (Hsu et al., 2002). Perhaps Emi1 also participates in the regulation of centriole splitting at the G1/S transition. Limiting centrosome duplication to once per cell cycle The original cell fusion studies of Rao and Johnson (1970) provided evidence not only for S phase promoting factor and M-phase promoting factor, but also for a mechanism to limit rereplication of a G2 nucleus until cells had passed through mitosis. The mechanism for limiting DNA replication to once-andonly-once per cell cycle is controlled by at least two mechanisms. First, replication pre-inititiation complexes form only when Cdk activity is low in G1 phase, and second, the activation of DNA unwinding and polymerase functions occurs when Cdk activity is raised at the G1/S transition (Blow and Hodgson, 2002; Dutta and Bell, 1997). At the same time, the high Cdk activity required to trigger initiation inhibits the reformation or reuse of preinitiation complexes, thereby ensuring that triggering of replication initiation is linked to preventing the establishment of preinitiation structures. The process of reestablishing pre-initiation complexes is reset during mitosis by a mechanism that requires the destruction of the mitotic cyclins by the APC (Noton and Diffley, 2000). However, the block to rereplication can be lost by inactivating Cdk activity during S or G2 phases, which allows premature

assembly of replication preinitiation structures and a round of rereplication without an intervening mitosis. Given the parallels between the triggering of DNA replication and centrosome duplication by Cdks, it is tempting to speculate that a similar once-and-only-once mechanism regulates centriole duplication. A surprising aspect of the centrosome cycle is that treatments that arrest cells in S phase allow an eventual uncoupling of the centrosome cycle from the mitotic cycle, giving rise to ectopic accumulation of centrosomes. Brinkley and colleagues first demonstrated that arresting Chinese Hamster Ovary (CHO) cells in S phase with hydroxyurea blocked DNA replication, but resulted in overduplication of centrosomes with some cells showing 8, 16, or more centrosomes (Balczon et al., 1995). A similar experiment in Xenopus embryos demonstrated that arresting cells in interphase with the protein synthesis inhibitor cycloheximide blocked mitosis (by preventing accumulation of cyclin B), but allowed the eventual overduplication of centrosomes (Gard et al., 1990). Although it is not clear what allows the uncoupling of the centrosome cycle, the ability of the centrosome in the S phase-blocked cell to resume Cdk and SCFdependent duplication appears to be an adaption that occurs following a prolonged S phase block. In CHO cells arrested with hydroxyurea, the cells do not begin centrosome overduplication for a period longer than the time required for a full cell cycle. Thus, it may be that the normal mechanisms gating centrosome division are lost over time and ectopic centrosome duplication ensues. For example, if the activity of the APC normally caused the rapid destruction of some unstable inhibitor in mitosis, this could be the normal means to reset the centrosome and allow new duplication. However, if during a prolonged S phase arrest the slow destruction of this inhibitor occurred, then the process of centrosome duplication could be reinitiated. More detailed studies of the reduplication phenomenon will be necessary to establish its mechanisms. Centrosome reduplication would be deleterious to the genomic stability of the cell by causing multipolar spindles and thereby inducing chromosome segregation defects. Such defects could also arise through abortive cytokinesis following centrosome duplication, as has recently been emphasized by Nigg and colleagues (Meraldi et al., 2002). Here, the failure to segregate centrosomes into daughter cells would cause an increased centrosome number. The implications of these kinds of defects for carcinogenesis are discussed below. Control of early mitotic events by ubiquitin-dependent proteolysis Several important steps in the centrosome duplication cycle occur near the G2/M transition as mitotic cyclindependent kinase activity increases. These include the completion of procentriole elongation, splitting and Oncogene

Ubiquitination and the centrosome cycle DV Hansen et al

6214

separation of the duplicated centriole pairs, and functional maturation of the duplicated centrosomes. Functional maturation is observed as a dramatic increase in microtubule-nucleating capacity and concomitant recruitment of additional proteins to the PCM (Khodjakov and Rieder, 1999). Although ubiquitinmediated proteolysis has not been directly implicated in these processes, several key mediators of these events are recognized for ubiquitination by the anaphasepromoting complex (APC) through its substrate adapters, Cdc20 and Cdh1, and thus may be under control of the Mad2 and Emil APC regulators. Thus, controlled activation of the APC may directly or at least indirectly prevent mitotic centrosome misfunction. Centrosomal splitting In mammalian cells, the centrosomal kinase Nek2A has been shown to mediate the disjunction, or splitting, of the duplicated centriole pairs through phosphorylation of C-Nap1, which is believed to be an important component of a proteinacious structure tethering the proximal ends of the two parental centrioles (Fry et al., 1998a,b; Mayor et al., 2000). Nek2A and protein phosphatase type 1 (PP1) form a complex wherein each enzyme is thought to antagonize the other’s activity (Helps et al., 2000). The timing of centrosome splitting around the time of mitotic entry may be attributed to phosphorylation and inactivation of PP1 by cyclin B/ Cdc2 (Kwon et al., 1997), thus allowing phosphorylation of C-Nap1 by Nek2 to prevail over the opposing phosphatase activity. C-Nap1 disappears from the centrosome during prophase and reappears during late telophase. The mechanism of how parental centrioles disengage as a result of C-Nap1 phosphorylation is not understood, but given that cellular levels of C-Nap1 decrease during mitosis (Mayor et al., 2000), an intriguing possibility is that phosphorylation of CNap1 facilitates its recognition by a ubiquitin ligase and that proteolysis is involved in the disassembly of the tether. It is also possible that C-Nap1 phosphorylation is entirely sufficient for tether disassembly. Recently, Nek2A was shown to be a substrate for APCCdc20 with remarkable similarities to cyclin A – an extended destruction box sequence mediating recognition by Cdc20, accumulation during S and G2 phases, and proteolysis in early mitosis in the presence of an activated spindle assembly checkpoint (Hames et al., 2001). Nek2A has also been characterized as a substrate for APCCdh1 ubiquitination through its KEN-box motif (Pfleger and Kirschner, 2000). Given the similarities between Nek2A and cyclin A, we might expect that Nek2A accumulation would be regulated by Emil (Hsu et al., 2002). Centrosomal separation The Eg5 family of kinesin-like proteins is important for driving centrosome separation and properly maintaining spindle bipolarity (reviewed in Kashina et al., 1997). In vertebrate tissue culture cells, the localization Oncogene

of Eg5 to the mitotic spindle begins at the centrosomes during G2 or prophase just prior to centrosome separation and is dependent upon its phosphorylation by cyclin B/Cdc2 (Blangy et al., 1995; Giet et al., 1999a; Sawin and Mitchison, 1995). Also essential for centrosome separation in higher eukaryotes are kinases of the Aurora-A family, which, like Nek2A, are antagonized by PP1 activity (Giet et al., 1999b; Glover et al., 1995; Katayama et al., 2001). Aurora-A and Eg5 apparently act through the same pathway to drive centrosome separation – Aurora-A phosphorylates and co-immunoprecipitates Eg5, the two proteins colocalize from prophase through anaphase, and perturbing the function of either protein prevents the bipolarity of spindles formed in Xenopus extracts (Giet et al., 1999a; Roghi et al., 1998; Sawin et al., 1992). Both Aurora-A and Cin8p, the budding yeast homologue of Eg5, are degraded postmitotically. They have recently been characterized as substrates for ubiquitination by APCCdh1, and the amino acid motifs required for their ubiquitination are highly conserved among species (Arlot-Bonnemains et al., 2001; Castro et al., 2002; Hildebrandt and Hoyt, 2001; Taguchi et al., 2002). The vertebrate Eg5 has not been shown to be regulated at the level of abundance, but it is possible that critical pools of Eg5, such as those associated with the centrosome, may be controlled by destruction. Mitotic maturation of MTOC activity Recruitment of additional proteins, including the gtubulin ring complex (g-TuRC), to the PCM and enhancement of microtubule-nucleating capacity are hallmarks of the mitotic centrosome. The accumulation of the g-TuRC at the centrosome in mitosis is perturbed by disrupting the function of Polo-like kinase 1 (Plk1), Aurora-A, or Nek2B (Berdnik and Knoblich, 2002; do Carmo Avides et al., 2001; Fry et al., 2000; Hannak et al., 2001; Lane and Nigg, 1996; Uto and Sagata, 2000). Like Aurora-A, the Polo-like kinase Plk1 has been shown to be a substrate for post-mitotic APCCdh1mediated degradation in multiple species (Charles et al., 1998; Fang et al., 1998; Shirayama et al., 1998). Whether APCCdh1 also mediates the degradation of Nek2B is not known; however, Nek2B protein levels oscillate in a cell cycle-dependent manner reminiscent of other APCCdh1 substrates, being lowest in G1 and peaking in mitosis (Hames and Fry, 2002). A potential role for ubiquitin ligases in preventing mitotic centrosome events in response to stress The existence of checkpoints which arrest or delay mitotic progression in response to various types of stress improves the fidelity of chromosome segregation. Arrest in G2 prevents mitotic entry by inhibiting Cdc2 activity through modulation of critical regulators of Cdc2 (reviewed in OConnell et al., 2000), including the phosphatase Cdc25C. APCCdh1 has recently been reported to be an important component of the ionizing radiation-induced G2 checkpoint (Sudo et al., 2001).

Ubiquitination and the centrosome cycle DV Hansen et al

6215

Interestingly, the APCCdh1-mediated delay of mitotic entry does not involve degradation of mitotic cyclins, implying that the degradation of other APC targets prevents mitotic entry. Supporting this idea, the G2/M arrest induced by arsenite has been shown to include ubiquitin-mediated degradation of Cdc25C in a KENbox dependent manner, suggesting that its ubiquitination is performed by APCCdh1 (Chen et al., 2002). In Cdc25-null Drosophila embryos, which arrest in G2 of cycle 14, procentriole elongation is attenuated, though perhaps indirectly (Vidwans et al., 1999). Since the mitotic activation of Nek2A, Eg5, and Aurora-A all depend on Cdc2 activity, degradation of Cdc25C would be predicted to prevent centrosomal splitting, separation, and the acquisition of MTOC activity as well. Alternatively, the APC could directly prevent these processes by ubiquitinating Nek2A, Aurora-A, Plk1, Eg5, or other key effector molecules. Whether APC substrates other than Cdc25 are also degraded during APCCdh1-mediated arrest remains to be seen, but activating the APC in response to stress to directly inhibit centrosomal processes required for chromosome segregation is an appealing model for preserving genomic integrity. Chfr has recently been identified as an additional checkpoint protein capable of delaying the G2/M transition in response to microtubule stress (Scolnick and Halazonetis, 2000), and new evidence indicates that Chfr is a ubiquitin ligase that targets Plk1 for destruction, delaying Cdc2 activation and mitotic entry (Kang et al., 2002). The DNA damage response also results in the inhibition of Plk1 activity in an ATR- or ATM-dependent fashion (Smits et al., 2000; van Vugt et al., 2001). Because Plk1 performs essential functions in mitotic entry, centrosome maturation, spindle function, mitotic exit, and cytokinesis, its targeted destruction or inactivation represents an attractive mechanism for delaying mitotic progression in response to stress. Significantly, DNA damage not only delays the G2/M transition, but also blocks mitotic exit in tissue culture cells (Smits et al., 2000). In addition, DNA damage can cause loss of mitotic centrosome function accompanied by dissociation of g-TuRC in early Drosophila embryos, which cycle directly between S phase and mitosis without intervening gap phases (Sibon et al., 2000). It will be interesting to learn whether these mitotic blocks are accomplished via ubiquitin-dependent proteolysis of Plk1 or other proteins and exactly how centrosome maturation and segregation processes are affected. Cell cycle transitions: control by and of the centrosome The centrosome is emerging as an important regulator of cell cycle transitions, including the G1/S transition (Hinchcliffe et al., 2001; Khodjakov and Rieder, 2001), the metaphase-anaphase transition (Huang and Raff, 1999; Wakefield et al., 2000), and completion of cytokinesis (Khodjakov and Rieder, 2001; Piel et al., 2001). Although how the centrosome governs these

transitions is not well understood, one model is that the centrosome acts as a central organizing center, where enzymes and substrates are brought into close proximity, and the signals responsible for cell cycle progression are integrated and propagated throughout the cell. For instance, the proteasomal machinery has been reported to be concentrated at the centrosome (Fabunmi et al., 2000; Wigley et al., 1999). Proteolytic activity specifically associated with the centrosomes has yet to be shown to directly catalyze cell cycle transitions, but it is certainly tempting to speculate that such localized activity is crucial to these processes. Centrosome-mediated initiation of the metaphaseanaphase transition Provocative evidence that centrosomes are a crucial organizing center of ubiquitin-mediated proteolysis comes primarily from live-cell imaging studies of Drosophila embryos expressing GFP-cyclin B (Huang and Raff, 1999). In these cells, GFP-cyclin B decorates the entire mitotic spindle at metaphase. Disappearance of GFP-cyclin B, attributed to ubiquitination by APCCdc20, occurs in a wave that begins at the spindle poles and proceeds to the spindle equator, after which the chromosomes enter anaphase. This same research group has isolated a mutation called centrosomes fall off (cfo) with a remarkable phenotype in which the centrosomes somehow detach from the spindle in mitosis (Wakefield et al., 2000). In cfo mutants, GFPcyclin B disappears from the detached centrosomes but not from the centrosome-less spindle, which arrests in mitosis, indicating that destruction of cyclin B on the spindle not only begins at the centrosome but also requires an intact connection between the spindle and the centrosome. It is also possible, however, that a checkpoint mechanism on the spindle recognizes centrosome detachment and prevents APC activation and anaphase. Both papers suggest that the centrosome acts as an organizing center for initiating proteolytic activity at the metaphase to anaphase transition, at least in Drosophila embryos. Centrosome-mediated completion of cytokinesis Accumulating evidence also implicates the centrosome in directly triggering the completion of cytokinesis (for review see Ou and Rattner, 2002). These studies challenge the dogma that centriolar splitting is uniquely associated with the onset of centriolar duplication and that mother and daughter centrioles maintain their orthogonal, or at least juxtaposed, relationship from the time of the daughter centriole’s duplication in S phase until centriolar splitting at the following G1/S transition. Rather, it appears that centriole position and behavior is much more dynamic than previously thought. Live-cell imaging of HeLa and U2OS cells stably expressing GFPcentrin as a centriolar marker demonstrated that soon after anaphase, the centrioles in each postmitotic centrosome separate. The mother centriole Oncogene

Ubiquitination and the centrosome cycle DV Hansen et al

6216

maintains a fixed position in the cell center, whereas the daughter centriole wanders throughout the cytoplasm (Piel et al., 2000). Shortly before completion of cytokinesis (abscission), the mother centriole transiently migrates directly to the intercellular bridge, resulting in the release of microtubule bundles from the midbody and narrowing of the bridge. Abscission occurs shortly after the mother centriole moves away and rejoins the daughter centriole in the cell center (Piel et al., 2001). Studies in fixed cells a decade earlier had suggested a similar interaction of the centrosome with the intercellular bridge (Mack and Rattner, 1993). Though the nature of the centrosome’s actual function in this process is unknown, disassembly of microtubule structures at the midbody could certainly involve ubiquitin-mediated proteolysis. One possible model is that the centrosome initiates this terminal step of cell division through directing the activity of the APC, which localizes to the centrosome throughout the cell cycle (Tugendreich et al., 1995), toward relevant substrates at the intercellular bridge. Interestingly, misregulation of the Cdc14A phosphatase, a core centrosomal component and activator of APCCdh1 in human cells, causes a broad range of mitotic defects, including cytokinesis failure (Bembenek and Yu, 2001; Kaiser et al., 2002; Mailand et al., 2002). Of particular interest, when Cdc14A was down-regulated by siRNA duplexes, one of the specific defects observed was an abortive cytokinesis in which daughter cells attempted to separate but remained attached by the intercellular bridge for several hours before fusing to form a binucleate cell (Mailand et al., 2002). Although Cdc14A was also shown in this study to have a role in splitting the centrosome, it has been shown elsewhere that separation of mother and daughter centrioles is not requisite for centriole movement to and from the intercellular bridge (Ou and Rattner, 2002), implying that the failure to undergo abscission may be due to a specific lack of Cdc14A activity at the intercellular bridge rather than impairment of the mother centriole’s ability to migrate there. Other aspects of how proteolysis controls cytokinesis have been recently reviewed (Glotzer and Dechant, 2002). An important regulatory system termed the mitotic exit network (MEN) coordinates migration of the spindle pole with APCCdh1 activation and the completion of cell division in budding yeast (see McCollum and Gould, 2001 for review). At the top of this signaling cascade lies the small GTPase Tem1p, which associates with the spindle pole body (SPB), the yeast equivalent of the centrosome. Tem1p is maintained in its inactive, GDP-bound state until it becomes activated upon migration of the daughter SPB into the bud. The downstream effector of the MEN is the Cdc14p phosphatase, which promotes downregulation of mitotic Cdk activity through multiple mechanisms including the activation of APCCdh1. In addition to Tem1p, other members of the MEN also localize to the SPB, implicating the SPB as an organizing center to direct mitotic exit. Oncogene

The centrosome may thus organize epigenetic information that controls its own fate, overseeing the same cell cycle transitions that ultimately dictate its own duplication and segregation to daughter cells. Therefore, various defects in centrosome function can be expected to beget failure of the once-and-only-once centrosome duplication restraint or defects in cell division itself. In either case, the end result is a failure of the cell cycle as a means of generating two daughter cells each containing one centrosome and one complete genome. Proteolysis at the centrosome and cancer While aneuploidy has long been known to be one of the most common hallmarks of cancer, this trait has customarily been thought to be a late event in cancer progression resulting from the accumulation of other genetic lesions. A growing body of evidence, however, suggests that aneuploidy may occur at an early rather than late stage in tumorigenesis and may even be critical in the transition to malignancy (Lingle et al., 2002). Because the centrosome serves as the microtubule-organizing center (MTOC) of the cell and thus has central roles in the establishment of cytoplasmic organization and formation of the mitotic spindle, mutations resulting in defective or supernumerary centrosomes have the potential to result in damage to and/or missegregation of chromosomes through formation of multipolar or otherwise aberrant spindles (Brinkley, 2001; Doxsey, 1998; Duensing and Munger, 2001; Lingle and Salisbury, 2000; Marx, 2001). We have discussed several known and postulated mechanisms by which ubiquitin-mediated proteolysis helps direct multiple steps in the centrosome cycle, and thus disruption of these proteolytic events may be crucial in driving one possible pathway toward tumorigenesis: centrosomal aberrations followed by aneuploidy and finally malignancy. It should be noted, however, that many genetic alterations reported to cause centrosome overduplication have not been fully characterized, and that the phenotype of supernumerary centrosomes, or centrosome amplification, may often result more from failures in cytokinesis rather than reinitiation of the duplication process itself. The following paragraphs summarize the known examples of relationships between ubiquitin ligase misregulation, centrosomal abnormalities and cancer. The SCF, centrosomes and cancer In addition to the work described above regarding the requirement for the SCF ubiquitin ligase components Skp1 and Cul1 for centriolar splitting at the G1/S transition (Freed et al., 1999), other studies have identified particular F-box proteins whose misregulation affect the centrosome duplication cycle and genomic stability. Mice with a targeted disruption in the F-box protein Skp2 are viable, but many animals

Ubiquitination and the centrosome cycle DV Hansen et al

6217

show marked overduplication of their centrosomes and extensive polyploidy (up to 12 centrosomes and 16C DNA content) (Nakayama et al., 2000). Importantly, however, these mice appeared healthy up to 10 months of age and failed to display an increased incidence of cancer, suggesting that the causal link between centrosomal/chromosomal defects and tumorigenesis is still not clear. Studies in Drosophila have identified another F-box protein, Slimb/b-TrCP, that acts to negatively regulate both Hedgehog and Wnt/Wingless oncogenic signal transduction pathways (Jiang and Struhl, 1998; Theodosiou et al., 1998). Defects in the latter pathway, resulting in misregulation of b-catenin degradation, have already been found in a significant number of human cancers. Interestingly, Slimb has also been implicated in limiting centrosome duplication, as Slimb hypomorphs display increased numbers of centrosomes as well as polyploidy (Wojcik et al., 2000). Further investigation is required to discover the targets of ubiquitination by these SCF complexes, how they cause centrosome amplification, and whether they have a role in cancer progression. The APC, centrosomes and cancer As previously discussed, several substrates of the APC are also important players in the centrosome cycle. Of these, cyclin A and kinases of the Aurora-A and Polo families are relevant to both centrosome amplification and cancer. First, high cyclin A/Cdk2 activity is needed for the centrosome overduplication phenotype observed in conditions of prolonged hydroxyurea treatment (Balczon, 2001), and cyclin A overexpression is an indicator of poor prognosis in breast and colon cancers (Bukholm et al., 2001; Handa et al., 1999; Michalides et al., 2002). Second, overexpression of Aurora-A and Plk1 is frequently observed in various cancers (Bischoff et al., 1998; Holtrich et al., 1994; Sakakura et al., 2001; Tanaka et al., 1999; Tanner et al., 2000). Homologues of both Aurora-A (Drosophila) and Plk1 (Xenopus) have been demonstrated to regulate microtubule dynamics in mitosis (Budde et al., 2001; Giet et al., 2002), and the overexpression of either Aurora-A or Plk1 can be sufficient to transform cells and induce centrosome amplification and aneuploidy through defects in cell division (Meraldi et al., 2002; Smith et al., 1997; Zhou et al., 1998). Disruption of APC function, therefore, might be expected to contribute to tumorigenesis through stabilization of cyclin A, Aurora-A, Plk1, and possibly other APC substrates. The recently discovered Emi1 protein contributes to the inactivation of APCCdh1 and accumulation of cyclin A at the G1/S transition, and its overexpression causes defects in mitotic progression and cell division through APC inhibition (Hsu et al., 2002; Reimann et al., 2001a). Moreover, Emi1 was included in a recently published list of 231 genes, identified by microarray profiling in a screen of 25 000 genes, whose overexpression was prognostic of a poor clinical outcome in pre-metastatic breast cancer patients (van’t Veer et al., 2002).

BRCA1 and p53, centrosomes and cancer In addition to the SCF and APC ubiquitin ligases, members of other ubiquitin ligase families are also associated with the centrosome cycle, aneuploidy, and cancer. BRCA1 is a well-known tumor suppressor gene involved in hereditary cancers of the breast and ovary (Deng and Brodie, 2000; Deng and Scott, 2000). More recently, a heterodimeric complex of BRCA1 and BARD1 has been shown to possess ubiquitin ligase activity through the RING domains of each protein (reviewed in Baer and Ludwig, 2002), although no substrates have been isolated thus far other than the BRCA1/BARD1 heterodimer itself. Strikingly, cancer-predisposing mutations have been identified in the RING domain that abolish not only BRCA1’s ubiquitin ligase activity, but also its function in the G2/M checkpoint (Ruffner et al., 2001). BRCA1 associates with the centrosome during mitosis through a g-tubulin binding domain (Hsu et al., 2001; Hsu and White, 1998). Mouse embryonic fibroblasts (MEFs) harboring a deletion of exon 11 of BRCA1, which no longer includes the g-tubulin binding domain, exhibit a defective G2/M checkpoint, centrosome amplification, and aneuploidy (Brodie and Deng, 2001; Xu et al., 1999), hinting that the mitotic localization of BRCA1 to the centrosome is critical for accurate cell division and genomic stability. Exon 11 also includes a nuclear localization sequence, but NLS-deficient BRCA1 is still capable of nuclear import and formation of DNA damage-induced nuclear foci through association with BARD1 (Fabbro et al., 2002). Moreover, overexpression of the BRCA1 g-tubulin binding domain causes centrosome amplification and spindle abnormalities, presumably by preventing endogenous, full-length BRCA1 from associating with mitotic centrosomes (Hsu et al., 2001). Whether the ubiquitin ligase activity of BRCA1 is required at the centrosome for genomic stability is an important question that remains to be addressed. Finally, the tumor suppressor p53 also localizes to the mitotic centrosome (Ciciarello et al., 2001; Morris et al., 2000), and p53 loss of function results in supernumerary centrosomes and genomic instability (Carroll et al., 1999; Fukasawa et al., 1996; Mussman et al., 2000; Ouyang et al., 2001; Tarapore et al., 2001). In p53-null MEFs, multiple copies of functional centrosomes are generated in a single cell cycle, demonstrating that the centrosome amplification observed in these circumstances is at least in part a direct effect of excessive centriole duplication in S phase (Fukasawa et al., 1996). This can be explained by the failure of p53 to transactivate the expression of the Cdk2 inhibitor p21. However, p53 also participates in a G2 checkpoint (Bunz et al., 1998; Taylor and Stark, 2001) and monitors the fidelity of mitosis independent of its function in inhibiting centriole duplication (Cross et al., 1995; Lanni and Jacks, 1998; Meek, 2000; Notterman et al., 1998). When cells escape the spindle assembly checkpoint after prolonged Oncogene

Ubiquitination and the centrosome cycle DV Hansen et al

6218

treatment with spindle-damaging agents, they undergo a defective cell division in which chromosomes fail to segregate. p53 responds to this mitotic failure by inducing a G1 arrest and preventing the tetraploid cell from reentering the cell cycle. In the absence of p53 cell cycle reentry is allowed, and polyploidy and centrosome amplification ensues. Indeed, p53 absence worsened the tetraploidization and centrosome amplification phenotype observed upon overexpression of Aurora-A or Plk1 mentioned previously (Meraldi et al., 2002). In the presence of stress, p53 mutants can therefore be expected to proceed through mitosis in the presence of DNA or spindle damage, resulting in genomic instability. Interestingly, tumors in which centrosome amplification occurs in the presence of wild-type p53 tend to display increased levels of Mdm2, the E3 ubiquitin ligase responsible for degradation of p53 (Carroll et al., 1999; Setoguchi et al., 2001). Mdm2 overexpression depletes the cell of p53, resulting in all of the expected downstream effects including centrosome abnormalities and chromosomal instability. Given this fact, it seems likely that the widely quoted

frequency of mutation in the p53 gene in sporadic cancers (greater than 50%) represents an underestimate of the real incidence of p53 inactivation in human cancer. Conclusion Currently, the number of proteins at the animal cell centrosome is not known, but certainly hundreds of proteins will be involved. The number of potential ubiquitin ligases – at least based on the number of RING finger proteins in the human genome – may also number in the hundreds. Thus, our understanding of ubiquitin ligase control of the centrosome is likely in its infancy. In addition to the basic cell biology of the centrosome, the probable connection of centrosome biology to human disease, notably cancer, is just beginning to be explored. As this set of centrosome reviews attests, basic and cancer biologists will have much to discuss about this tiny organelle in the upcoming years.

References Arlot-Bonnemains Y, Klotzbucher A, Giet R, Uzbekov R, Bihan R and Prigent C. (2001). FEBS Lett., 508, 149 – 152. Baer R and Ludwig T. (2002). Curr. Opin. Genet. Dev., 12, 86 – 91. Balczon R, Bao L, Zimmer WE, Brown K, Zinkowski RP and Brinkley BR. (1995). J. Cell. Biol., 130, 105 – 115. Balczon RC. (2001). Chromosoma, 110, 381 – 392. Bembenek J and Yu H. (2001). J. Biol. Chem., 276, 48237 – 48242. Berdnik D and Knoblich JA. (2002). Curr. Biol., 12, 640 – 647. Bischoff JR, Anderson L, Zhu Y, Mossie K, Ng L, Souza B, Schryver B, Flanagan P, Clairvoyant F, Ginther C, Chan CS, Novotny M, Slamon DJ and Plowman GD. (1998). EMBO J., 17, 3052 – 3065. Blangy A, Lane HA, d’Herin P, Harper M, Kress M and Nigg EA. (1995). Cell, 83, 1159 – 1169. Blow JJ and Hodgson B. (2002). Trends Cell. Biol., 12, 72 – 78. Bornens M. (2002). Curr. Opin. Cell. Biol., 14, 25 – 34. Brinkley BR. (2001). Trends Cell. Biol., 11, 18 – 21. Brodie SG and Deng CX. (2001). Trends Genet., 17, S18 – S22. Budde PP, Kumagai A, Dunphy WG and Heald R. (2001). J. Cell. Biol., 153, 149 – 158. Bukholm IR, Bukholm G and Nesland JM. (2001). Int. J. Cancer, 93, 283 – 287. Bunz F, Dutriaux A, Lengauer C, Waldman T, Zhou S, Brown JP, Sedivy JM, Kinzler KW and Vogelstein B. (1998). Science, 282, 1497 – 1501. Burton JL and Solomon MJ. (2000). Mol. Cell. Biol., 20, 4614 – 4625. Carroll PE, Okuda M, Horn HF, Biddinger P, Stambrook PJ, Gleich LL, Li YQ, Tarapore P and Fukasawa K. (1999). Oncogene, 18, 1935 – 1944. Castro A, Arlot-Bonnemains Y, Vigneron S, Labbe JC, Prigent C and Lorca T. (2002). EMBO Rep., 3, 457 – 462. Charles JF, Jaspersen SL, Tinker-Kulberg RL, Hwang L, Szidon A and Morgan DO. (1998). Curr. Biol., 8, 497 – 507. Oncogene

Chen F, Zhang Z, Bower J, Lu Y, Leonard SS, Ding M, Castranova V, Piwnica-Worms H and Shi X. (2002). Proc. Natl. Acad. Sci. USA, 99, 1990 – 1995. Chen J and Fang G. (2001). Genes Dev., 15, 1765 – 1770. Ciciarello M, Mangiacasale R, Casenghi M, Zaira Limongi M, D’Angelo M, Soddu S, Lavia P and Cundari E. (2001). J. Biol. Chem., 276, 19205 – 19213. Cross SM, Sanchez CA, Morgan CA, Schimke MK, Ramel S, Idzerda RL, Raskind WH and Reid BJ. (1995). Science, 267, 1353 – 1356. Deng CX and Brodie SG. (2000). Bioessays, 22, 728 – 737. Deng CX and Scott F. (2000). Oncogene, 19, 1059 – 1064. Deshaies RJ. (1999). Ann. Rev. Cell. Dev. Biol., 15, 435 – 467. do Carmo Avides M, Tavares A and Glover DM. (2001). Nat. Cell. Biol., 3, 421 – 424. Doxsey S. (1998). Nat. Genet., 20, 104 – 106. Duensing S and Munger K. (2001). Biochim. Biophys. Acta, 2, M81 – M88. Dutta A and Bell SP. (1997). Ann. Rev. Cell. Dev. Biol., 13, 293 – 332. Fabbro M, Rodriguez JA, Baer R and Henderson BR. (2002). J. Biol. Chem., 277, 21315 – 21324. Fabunmi RP, Wigley WC, Thomas PJ and DeMartino GN. (2000). J. Biol. Chem., 275, 409 – 413. Fang G, Yu H and Kirschner MW. (1998). Mol. Cell, 2, 163 – 171. Fisk HA and Winey M. (2001). Cell, 106, 95 – 104. Freed E, Lacey KR, Huie P, Lyapina SA, Deshaies RJ, Stearns T and Jackson PK. (1999). Genes Dev., 13, 2242 – 2257. Fry AM, Descombes P, Twomey C, Bacchieri R and Nigg EA. (2000). J. Cell. Sci., 113, 1973 – 1984. Fry AM, Mayor T, Meraldi P, Stierhof YD, Tanaka K and Nigg EA. (1998a). J. Cell. Biol., 141, 1563 – 1574. Fry AM, Meraldi P and Nigg EA. (1998b). EMBO J., 17, 470 – 481. Fukasawa K, Choi T, Kuriyama R, Rulong S and Vande Woude GF. (1996). Science, 271, 1744 – 1747.

Ubiquitination and the centrosome cycle DV Hansen et al

6219

Gard DL, Hafezi S, Zhang T and Doxsey SJ. (1990). J. Cell. Biol., 110, 2033 – 2042. Giet R, McLean D, Descamps S, Lee MJ, Raff JW, Prigent C and Glover DM. (2002). J. Cell. Biol., 156, 437 – 451. Giet R, Uzbekov R, Cubizolles F, Le Guellec K and Prigent C. (1999a). J. Biol. Chem., 274, 15005 – 15013. Giet R, Uzbekov R, Kireev I and Prigent C. (1999b). Biol. Cell, 91, 461 – 470. Glotzer M. (2001). Annu. Rev. Cell. Dev. Biol., 17, 351 – 386. Glotzer M and Dechant R. (2002). Curr. Biol., 12, R344 – R346. Glover DM, Leibowitz MH, McLean DA and Parry H. (1995). Cell, 81, 95 – 105. Gmachl M, Gieffers C, Podtelejnikov AV, Mann M and Peters JM. (2000). Proc. Natl. Acad. Sci. USA, 97, 8973 – 8978. Gstaiger M, Marti A and Krek W. (1999). Exp. Cell. Res., 247, 554 – 562. Hames RS and Fry AM. (2002). Biochem. J., 361, 77 – 85. Hames RS, Wattam SL, Yamano H, Bacchieri R and Fry AM. (2001). EMBO J., 20, 7117 – 7127. Handa K, Yamakawa M, Takeda H, Kimura S and Takahashi T. (1999). Int. J. Cancer, 84, 225 – 233. Hannak E, Kirkham M, Hyman AA and Oegema K. (2001). J. Cell. Biol., 155, 1109 – 1116. Hatakeyama S, Yada M, Matsumoto M, Ishida N and Nakayama K. (2001). J. Biol. Chem., 276, 33111 – 33120. Helps NR, Luo X, Barker HM and Cohen PT. (2000). Biochem. J., 349, 509 – 518. Hildebrandt ER and Hoyt MA. (2001). Mol. Biol. Cell, 12, 3402 – 3416. Hilioti Z, Chung YS, Mochizuki Y, Hardy CF and CohenFix O. (2001). Curr. Biol., 11, 1347 – 1352. Hinchcliffe EH, Li C, Thompson EA, Maller JL and Sluder G. (1999). Science, 283, 851 – 854. Hinchcliffe EH, Miller FJ, Cham M, Khodjakov A and Sluder G. (2001). Science, 291, 1547 – 1550. Hinchcliffe EH and Sluder G. (2001). Genes Dev., 15, 1167 – 1181. Holtrich U, Wolf G, Brauninger A, Karn T, Bohme B, Rubsamen-Waigmann H and Strebhardt K. (1994). Proc. Natl. Acad. Sci. USA, 91, 1736 – 1740. Honda R, Tanaka H and Yasuda H. (1997). FEBS Lett., 420, 25 – 27. Hsu JY, Reimann JD, Sorensen CS, Lukas J and Jackson PK. (2002). Nat. Cell. Biol., 4, 358 – 366. Hsu LC, Doan TP and White RL. (2001). Cancer Res., 61, 7713 – 7718. Hsu LC and White RL. (1998). Proc. Natl. Acad. Sci. USA, 95, 12983 – 12988. Huang J and Raff JW. (1999). EMBO J., 18, 2184 – 2195. Jackson PK, Eldridge AG, Freed E, Furstenthal L, Hsu JY, Kaiser BK and Reimann JD. (2000). Trends Cell. Biol., 10, 429 – 439. Jiang J, Ballinger CA, Wu Y, Dai Q, Cyr DM, Hohfeld J and Patterson C. (2001). J. Biol. Chem., 276, 42938 – 42944. Jiang J and Struhl G. (1998). Nature, 391, 493 – 496. Kaiser BK, Zimmerman ZA, Charbonneau H and Jackson PK. (2002). Mol. Biol. Cell, 13, 2289 – 2300. Kamura T, Koepp DM, Conrad MN, Skowyra D, Moreland RJ, Iliopoulos O, Lane WS, Kaelin Jr WG, Elledge SJ, Conaway RC, Harper JW and Conaway JW. (1999). Science, 284, 657 – 661. Kang D, Chen J, Wong J and Fang G. (2002). J. Cell. Biol., 156, 249 – 259. Kashina AS, Rogers GC and Scholey JM. (1997). Biochim. Biophys. Acta., 1357, 257 – 271.

Katayama H, Zhou H, Li Q, Tatsuka M and Sen S. (2001). J. Biol. Chem., 276, 46219 – 46224. Khodjakov A and Rieder CL. (1999). J. Cell. Biol., 146, 585 – 596. Khodjakov A and Rieder CL. (2001). J. Cell. Biol., 153, 237 – 242. Kramer ER, Scheuringer N, Podtelejnikov AV, Mann M and Peters JM. (2000). Mol. Biol. Cell, 11, 1555 – 1569. Kwon YG, Lee SY, Choi Y, Greengard P and Nairn AC. (1997). Proc. Natl. Acad. Sci. USA, 94, 2168 – 2173. Lacey KR, Jackson PK and Stearns T. (1999). Proc. Natl. Acad. Sci. USA, 96, 2817 – 2822. Lane HA and Nigg EA. (1996). J. Cell. Biol., 135, 1701 – 1713. Lanni JS and Jacks T. (1998). Mol. Cell. Biol., 18, 1055 – 1064. Lingle WL, Barrett SL, Negron VC, D’Assoro AB, Boeneman K, Liu W, Whitehead CM, Reynolds C and Salisbury JL. (2002). Proc. Natl. Acad. Sci. USA, 99, 1978 – 1983. Lingle WL and Salisbury JL. (2000). Curr. Top. Dev. Biol., 49, 313 – 329. Lu Z, Xu S, Joazeiro CA, Cobb MH and Hunter T. (2002). Mol. Cell, 9, 945 – 956. Mack G and Rattner JB. (1993). Cell. Motil. Cytoskeleton, 26, 239 – 247. Mailand N, Lukas C, Kaiser BK, Jackson PK, Bartek J and Lukas J. (2002). Nat. Cell. Biol., 4, 318 – 322. Marshall WF. (2001). Curr. Biol., 11, R487 – R496. Marx J. (2001). Science, 292, 426 – 429. Matsumoto Y, Hayashi K and Nishida E. (1999). Curr. Biol., 9, 429 – 432. Mayor T, Stierhof YD, Tanaka K, Fry AM and Nigg EA. (2000). J. Cell. Biol., 151, 837 – 846. McCollum D and Gould KL. (2001). Trends Cell. Biol., 11, 89 – 95. Meek DW. (2000). Pathol. Biol. (Paris), 48, 246 – 254. Meraldi P, Honda R and Nigg EA. (2002). EMBO J., 21, 483 – 492. Meraldi P, Lukas J, Fry AM, Bartek J and Nigg EA. (1999). Nature Cell. Biol., 1, 88 – 93. Meselson M and Stahl FW. (1958). PNAS, 44, 671. Michalides R, van Tinteren H, Balkenende A, Vermorken JB, Benraadt J, Huldij J and van Diest P. (2002). Br. J. Cancer, 86, 402 – 408. Morris VB, Brammall J, Noble J and Reddel R. (2000). Exp. Cell. Res., 256, 122 – 130. Mussman JG, Horn HF, Carroll PE, Okuda M, Tarapore P, Donehower LA and Fukasawa K. (2000). Oncogene, 19, 1635 – 1646. Nakayama K, Nagahama H, Minamishima YA, Matsumoto M, Nakamichi I, Kitagawa K, Shirane M, Tsunematsu R, Tsukiyama T, Ishida N, Kitagawa M and Hatakeyama S. (2000). EMBO J., 19, 2069 – 2081. Noton E and Diffley JF. (2000). Mol. Cell, 5, 85 – 95. Notterman D, Young S, Wainger B and Levine AJ. (1998). Oncogene, 17, 2743 – 2751. O’Connell MJ, Walworth NC and Carr AM. (2000). Trends Cell. Biol., 10, 296 – 303. Ohta T, Michel JJ, Schottelius AJ and Xiong Y. (1999). Mol. Cell, 3, 535 – 541. Okuda M, Horn HF, Tarapore P, Tokuyama Y, Smulian AG, Chan PK, Knudsen ES, Hofmann IA, Snyder JD, Bove KE and Fukasawa K. (2000). Cell, 103, 127 – 140. Ou YY and Rattner JB. (2002). Cell. Motil. Cytoskeleton, 51, 123 – 132.

Oncogene

Ubiquitination and the centrosome cycle DV Hansen et al

6220

Ouyang X, Wang X, Xu K, Jin DY, Cheung AL, Tsao SW and Wong YC. (2001). Biochim. Biophys. Acta., 1541, 212 – 220. Paintrand M, Moudjou M, Delacroix H and Bornens M. (1992). J. Struct. Biol., 108, 107 – 128. Patterson C. (2002). Science’s STKE [online resource]: http:// stke.sciencemag.org/cgi/content/full/OC_sigtrans; 2002/ 116/pe4. Pfleger CM and Kirschner MW. (2000). Genes Dev., 14, 655 – 665. Pfleger CM, Lee E and Kirschner MW. (2001a). Genes Dev., 15, 2396 – 2407. Pfleger CM, Salic A, Lee E and Kirschner MW. (2001b). Genes Dev., 15, 1759 – 1764. Piel M, Meyer P, Khodjakov A, Rieder CL and Bornens M. (2000). J. Cell. Biol., 149, 317 – 330. Piel M, Nordberg J, Euteneuer U and Bornens M. (2001). Science, 291, 1550 – 1553. Rao PN and Johnson RT. (1970). Nature, 225, 159 – 164. Reimann JD, Freed E, Hsu JY, Kramer ER, Peters JM and Jackson PK. (2001a). Cell, 105, 645 – 655. Reimann JD, Gardner BE, Margottin-Goguet F and Jackson PK. (2001b). Genes Dev., 15, 3278 – 3285. Roghi C, Giet R, Uzbekov R, Morin N, Chartrain I, Le Guellec R, Couturier A, Doree M, Philippe M and Prigent C. (1998). J. Cell. Sci., 111, 557 – 572. Ruffner H, Joazeiro CA, Hemmati D, Hunter T and Verma IM. (2001). Proc. Natl. Acad. Sci. USA, 98, 5134 – 5139. Sakakura C, Hagiwara A, Yasuoka R, Fujita Y, Nakanishi M, Masuda K, Shimomura K, Nakamura Y, Inazawa J, Abe T and Yamagishi H. (2001). Br. J. Cancer, 84, 824 – 831. Sawin KE, LeGuellec K, Philippe M and Mitchison TJ. (1992). Nature, 359, 540 – 543. Sawin KE and Mitchison TJ. (1995). Proc. Natl. Acad. Sci. USA, 92, 4289 – 4293. Schwab M, Neutzner M, Mocker D and Seufert W. (2001). EMBO J., 20, 5165 – 5175. Scolnick DM and Halazonetis TD. (2000). Nature, 406, 430 – 435. Setoguchi A, Okuda M, Nishida E, Yazawa M, Ishizaka T, Hong SH, Hisasue M, Nishimura R, Sasaki N, Yoshikawa Y, Masuda K, Ohno K and Tsujimoto H. (2001). Am. J. Vet. Res., 62, 1134 – 1141. Shah JV and Cleveland DW. (2000). Cell, 103, 997 – 1000. Shirayama M, Zachariae W, Ciosk R and Nasmyth K. (1998). EMBO J., 17, 1336 – 1349. Sibon OC, Kelkar A, Lemstra W and Theurkauf WE. (2000). Nat. Cell. Biol., 2, 90 – 95. Skowyra D, Koepp DM, Kamura T, Conrad MN, Conaway RC, Conaway JW, Elledge SJ and Harper JW. (1999). Science, 284, 662 – 665.

Oncogene

Smith MR, Wilson ML, Hamanaka R, Chase D, Kung H, Longo DL and Ferris DK. (1997). Biochem. Biophys. Res. Commun., 234, 397 – 405. Smits VA, Klompmaker R, Arnaud L, Rijksen G, Nigg EA and Medema RH. (2000). Nat. Cell. Biol., 2, 672 – 676. Sudo T, Ota Y, Kotani S, Nakao M, Takami Y, Takeda S and Saya H. (2001). EMBO J., 20, 6499 – 6508. Taguchi S, Honda K, Sugiura K, Yamaguchi A, Furukawa K and Urano T. (2002). FEBS Lett., 519, 59 – 65. Tan P, Fuchs SY, Chen A, Wu K, Gomez C, Ronai Z and Pan ZQ. (1999). Mol. Cell, 3, 527 – 533. Tanaka T, Kimura M, Matsunaga K, Fukada D, Mori H and Okano Y. (1999). Cancer Res., 59, 2041 – 2044. Tang Z, Li B, Bharadwaj R, Zhu H, Ozkan E, Hakala K, Deisenhofer J and Yu H. (2001). Mol. Biol. Cell, 12, 3839 – 3851. Tanner MM, Grenman S, Koul A, Johannsson O, Meltzer P, Pejovic T, Borg A and Isola JJ. (2000). Clin. Cancer Res., 6, 1833 – 1839. Tarapore P, Horn HF, Tokuyama Y and Fukasawa K. (2001). Oncogene, 20, 3173 – 3184. Taylor WR and Stark GR. (2001). Oncogene, 20, 1803 – 1815. Theodosiou NA, Zhang S, Wang WY and Xu T. (1998). Development, 125, 3411 – 3416. Tugendreich S, Tomkiel J, Earnshaw W and Hieter P. (1995). Cell, 81, 261 – 268. Uto K and Sagata N. (2000). EMBO J., 19, 1816 – 1826. van’t Veer LJ, Dai H, van de Vijver MJ, He YD, Hart AA, Mao M, Peterse HL, van der Kooy K, Marton MJ, Witteveen AT, Schreiber GJ, Kerkhoven RM, Roberts C, Linsley PS, Bernards R and Friend SH. (2002). Nature, 415, 530 – 536. van Vugt MA, Smits VA, Klompmaker R and Medema RH. (2001). J. Biol. Chem., 276, 41656 – 41660. Vidwans SJ, Wong ML and OFarrell PH. (1999). J. Cell. Biol., 147, 1371 – 1378. Wakefield JG, Huang JY and Raff JW. (2000). Curr. Biol., 10, 1367 – 1370. Wigley WC, Fabunmi RP, Lee MG, Marino CR, Muallem S, DeMartino GN and Thomas PJ. (1999). J. Cell. Biol., 145, 481 – 490. Wojcik EJ, Glover DM and Hays TS. (2000). Curr. Biol., 10, 1131 – 1134. Xu X, Weaver Z, Linke SP, Li C, Gotay J, Wang XW, Harris CC, Ried T and Deng CX. (1999). Mol. Cell, 3, 389 – 395. Zheng N, Schulman BA, Song L, Miller JJ, Jeffrey PD, Wang P, Chu C, Koepp DM, Elledge SJ, Pagano M, Conaway RC, Conaway JW, Harper JW and Pavletich NP. (2002). Nature, 416, 703 – 709. Zhou H, Kuang J, Zhong L, Kuo WL, Gray JW, Sahin A, Brinkley BR and Sen S. (1998). Nat. Genet., 20, 189 – 193.

Ubiquitination and the centrosome cycle DV Hansen et al

6221

Appendix – A primer on ubiquitylation enzymes

E3, a ubiquitin ligase

Ubiquitin-dependent proteolysis occurs following the covalent addition of a polyubiquitin chain to a specific target protein. Polyubiquitylation targets proteins to the 26S proteasome, an ATP- and ubiquitin-dependent protease complex. The cascade of E1, E2 and E3 enzymes activate the ubiquitin and facilitate the assembly of the multiubiquitin chain. Generally, the E3s are the most diverse group.

The E3 ubiquitin ligases couple with the E2s to bind to substrate and assemble a multiubiquitin chain on the substrate. For the HECT domain E3s, the E3 itself forms a thioester with ubiquitin and presumably participates in transferring the ubiquitin directly to the substrate. For the E3 complexes containing cullins or RING finger proteins, no direct thioester between E3 and ubiquitin has been identified. In these E3 classes, the ubiquitin ligase might function to facilitate the interaction between substrate and E2 enzyme (see main text for further discussion).

Ubiquitin (Ub) A 76 amino acid protein with the C-terminal glycine (Gly76) residue capable of forming an isopeptide bond with a side chain lysine on a target protein or one of the side chain lysines of ubiquitin itself (notably lysine 48). There are a number of ubiquitinlike molecules (UBLs), such as SUMO-1 and NEDD8/Rub1, that are added as monomers to side-chain lysines.

E4, a multiubiquitin chain assembly factor The budding yeast UFD2 has been called an E4 protein. The protein binds to multiubiquitin chains and may facilitate part of the ubiquitin-dependent proteolysis pathway following ubiquitin chain assembly.

E1, a ubiquitin-activating enzyme

The 26S proteasome

Ubiquitin is activated by the E1 enzyme and ATP to form a thioester linkage between the C-terminal glycine 76 and a conserved active site cysteine. Uba1 is the major form of this enzyme in yeast and humans.

The proteasome is an abundant protease complex comprising a 20S core particle (CP) flanked by two 19S regulatory particles (RP). The CP is cylindrical, with narrow channels feeding a central cavity with multiple protease sites. The regulatory particles regulate substrate access on both ends of the core. The RP itself contains a base, which contains ATPase subunits thought to unfold substrates for access to the core, and a lid, which contains subunits for binding multiubiquitin chains and ubiquitin-deconjugating enzymes (isopeptidases).

E2, a ubiquitin-conjugating enzyme After activation by the E1, ubiquitin is transesterified to a conserved cysteine of an E2 enzyme. There are 13 E2s in yeast and approximately 30 – 50 in vertebrates. The genetic name for these enzymes includes the three letter code ‘Ubc’. Except for Ubc9 (a SUMO-conjugating enzyme) and Ubc12 (a NEDD8/Rub1 conjugating enzyme), the Ubcs have varying genetically defined functions in ubiquitylation, but some overlapping roles. Ubc3 is also known as Cdc34, a crucial E2 enzyme in the SCF ubiquitin ligase.

Ubiquitin hydrolases (isopeptidases) A diverse group of enzymes that hydrolyse isopeptide bonds in a multiubiquitin chain. These enzymes might provide regulatory and proofreading functions for assembly of ubiquitin chains and also allow recycling of ubiquitin monomers.

Oncogene