Mitotic kinase cascades orchestrating timely disjunction and movement of centrosomes maintain chromosomal stability and prevent cancer

Abstract

Centrosomes are microtubule-organizing centers that duplicate in S phase to form bipolar spindles that separate duplicated chromosomes faithfully into two daughter cells during cell division. Recent studies show that proper timing of centrosome dynamics, the disjunction and movement of centrosomes, is tightly linked to spindle symmetry, correct microtubule-kinetochore attachment and chromosome segregation. Here we review mechanisms that regulate centrosome dynamics, with emphasis on the roles of key mitotic kinases in the proper timing of centrosome dynamics and how aberrancies in these processes may cause chromosomal instability and cancer.

INTRODUCTION

The centrosome is the major microtubule-organizing center of animal cells and is surrounded by pericentriolar material (PCM) containing many different proteins (Bornens, 2012; Nigg and Raff, 2009). The centrosome consists of an orthogonal arrangement of two centrioles, consisting of nine sets of microtubule triplets organized in a symmetric cartwheel, which are connected by fibers (Azimzadeh and Bornens, 2007; Bettencourt-Dias and Glover, 2007; Marshall and Rosenbaum, 1999). In non-cycling cells, centrosomes are located in close proximity to the nucleus and play a role in establishing cell shape, polarity, and proper positioning of subcellular organelles (Nigg, 2007). In cycling cells, centrosomes duplicate and then separate while nucleating microtubules to form the spindle apparatus that separates chromosomes between daughter cells during cell division (Bettencourt-Dias and Glover, 2007). Both centrosome duplication and separation are highly complex biological processes that occur only once per division and tightly coordinated with cell cycle progression (Meraldi and Nigg, 2002; Nigg and Stearns, 2011; Wang et al., 2014). Centrosome duplication overlaps with DNA replication, whereas centrosome separation occurs in late G2 phase and early mitosis and according to two steps: centrosome disjunction and movement. Centrosome disjunction involves the release of two linker proteins, c-Nap1 and rootletin, which hold sister centrosomes together after completion of duplication (Bahe et al., 2005; Fry et al., 1998; Mardin et al., 2011; Mayor et al., 2000). Centrosome movement is mediated by the coordinated actions of spindle microtubules and kinesin and dynein motor proteins, and concludes with the anchoring of centrosomes at opposite cell poles (Gayek and Ohi, 2014; Mardin and Schiebel, 2012; Tanenbaum et al., 2010; Tanenbaum et al., 2008; Toso et al., 2009; van Heesbeen et al., 2013).

Historically, several observations have portrayed centrosome separation as a noncritical process. First, in certain experimental settings mitotic spindles still form after removal of centrosomes by laser ablation or microsurgery (Hinchcliffe et al., 2001; Khodjakov et al., 2000). Second, initiation of centrosome separation at nuclear envelope breakdown (NEBD) at prometaphase onset instead of late G2 phase does not overtly compromise chromosome segregation (Rosenblatt et al., 2004; Toso et al., 2009). Third, Drosophila melanogaster mutants lacking centrosomes are viable and do not exhibit any overt abnormalities (Basto et al., 2006). Nevertheless, it has become increasingly clear over recent years that the timing of centrosome dynamics needs to be meticulously controlled to achieve proper bipolar spindle formation to avoid chromosome missegregation, aneuploidy and increased risk for neoplastic transformation. In this review, we will discuss current insights into the molecular mechanisms of normal centrosome disjunction and movement with emphasis on mitotic kinase cascades, and how deregulation of these pathways cause aneuploidy and cancer.

CENTROSOME DYNAMICS

Centrosome disjunction and movement are regulated by mitotic kinase cascades in which Cdk1 and Plk1 are key participants as detailed below (Bertran et al., 2011; Mardin et al., 2011; Mardin and Schiebel, 2012; Nam and van Deursen, 2014; Smith et al., 2011; Wang et al., 2014).

Centrosome disjunction

Duplicated centrosomes stay connected through centrosomal cohesion complexes consisting not only of c-Nap1 and rootletin, but also components such as Cep68, Cep215, conductin, and LRRC45 (Bahe et al., 2005; Fry et al., 1998; Graser et al., 2007; Hadjihannas et al., 2010; He et al., 2013). C-Nap1 is a large coiled-coil protein that localizes rootletin to the proximal end of centrioles and physically links duplicated centrosomes through fibrous polymers (Bahe et al., 2005; Fry et al., 1998; Mayor et al., 2000). Late in G2, c-Nap1 and rootletin are phosphorylated by the Never In Mitosis (NIMA)-related serine/threonine kinase Nek2A, triggering the dissociation of c-Nap1 and other cohesion complex proteins and resulting in centrosome disjunction (Fry et al., 1998; Hardy et al., 2014; Meraldi and Nigg, 2001). Nek2A accumulation at centrosomes is regulated by two components of the Hippo pathway, mammalian sterile 20-like kinase 2 (Mst2) and scaffold protein Salvador (Sav1) (Figure 1). Of these, Mst2 is a kinase that activates Nek2A through phosphorylation (Mardin et al., 2010). Plk1, in turn, serves as the upstream activating kinase of Mst2. Plk1 also promotes disjunction by preventing the association of Mst2-Nek2A with PP1γ, a phosphatase that counteracts Nek2A activity at centrosomes (Mardin et al., 2011). In addition, PP1γ is phosphorylated and inhibited by activated cyclin-Cdk1 (Dohadwala et al., 1994). The pre-mitotic and mitotic functions of the Cdk1, Aurora A, and Plk1 kinases are complex and multi-faceted and whether Cdk1 functions upstream or downstream of Aurora A remains controversial (Bruinsma et al., 2012; Lindqvist et al., 2009). However, our own studies in which cyclin B2 levels were markedly reduced or increased indicate that centrosome-associated cyclin B2-Cdk1 complexes indirectly activate Plk1 at centrosomes to induce disjunction via direct or indirect activation of Aurora A (Nam and van Deursen, 2014), showing that cyclin B2 is the activating Cdk1 partner in this context, and placing cyclin B2-Cdk1 at the top of the kinase cascade (Figure 1).

Figure 1. Mitotic kinases implicated in centrosome disjunction.

Centrosome-localized cyclin B2-Cdk1 triggers the phosphorylation of Aurora-A, which in turn activates Plk1 through phosphorylation. Subsequently, Plk1 phosphorylates Sav1-bound Mst2 to induce interaction with Nek2A. This allows for centrosome accumulation and activation of Nek2A and subsequent phosphorylation of c-Nap1 and rootletin, resulting in centrosome disjunction. Cdk1 also contributes to centrosome disjunction by inhibiting PP1γ, a phosphatase that antagonizes Nek2A-mediated phosphorylation of c-Nap1 and rootletin. The activating cyclins involved remain to be established. β-catenin is a substrate of Nek2A and phosphorylation of β-catenin has been observed at centrosomes. Cep85 localizes at proximal ends of centrioles and has been identified as a negative regulator of Nek2A. In addition, EGFR signaling at the plasma membrane leads to GRK2 kinase activation, which seemingly triggers centrosome separation by targeting the Mst2-Nek2A kinase module. The biological relevance of the EGFR-mediated control of centrosome movement is currently unclear.

The latter findings identified a unique role for cyclin B2 in mitosis and firmly established the cyclin B2/Cdk1-Aurora A-Plk1-Mst2-Nek2A kinase cascade for centrosome disjunction. However, several other proteins also seem to be involved in centrosome disjunction (Figure 1). First, β-catenin, a key effector of the Wnt signaling pathway, has been identified as a component of the centrosomal protein linker complex. β-catenin is a substrate of Nek2A and its depletion by RNAi leads to inhibition of centrosome separation (Bahmanyar et al., 2008; Kaplan et al., 2004). Furthermore, β-catenin phosphorylation at Ser33-Ser37-Thr41 has been observed at centrosomes and appears to be involved in spindle and microtubule dynamics (Chilov et al., 2011; Huang et al., 2007). However, which kinases phosphorylate these sites and how centrosomal β-catenin stability and function are regulated during mitosis remains to be established. Second, centrosome protein 85 KDa (Cep85) has been identified as a negative regulator of Nek2A (Chen et al., 2015). Cep85 interacts with Nek2A and localizes to the proximal ends of centrioles. Depletion of Cep85 results in premature centrosome separation, whereas its overexpression leads to centrosome disjunction failure and prometaphase arrest (Chen et al., 2015). Third, another member of the NIMA-related kinase family, Nek5, is also located at centrosomes, where it has been proposed to be important for maintenance of PCM integrity and the mitotic loss of centrosome cohesion (Prosser et al., 2015). This is based on the observation that Nek5 depletion led to loss of PCM components, inappropriately retained centrosome linker components, and exhibited delayed centrosome separation and defective chromosome segregation (Prosser et al., 2015). Fourth, epidermal growth factor receptor (EGFR) signaling has been determined to be associated with centrosome separation timing (Figure 1). Addition of EGF to the culture medium or overexpression of EGFR induces premature centrosome separation, presumably by activating the Mst2-Nek2A module in S phase (Mardin et al., 2013). G protein-coupled receptor kinase 2 (GRK2) has been found at centrosomes and may be involved in EGFR-mediated centrosome separation (So et al., 2013). Upon EGF addition, GRK2 phosphorylates and activates Mst2, leading to the initiation of centrosome disjunction. Fifth, transcriptional element-interacting factor (TEIF) co-localizes with c-Nap1 at the proximal ends of centrioles. Centriolar loading of TEIF is stimulated by the EGF-PI3K-Akt pathway, and known to displace c-Nap1 from centrosomes, thereby initiating centrosome disjunction (Zhao et al., 2014).

Centrosome movement

The main proteins involved in centrosome movement are the kinases Cdk1, Plk1, the NIMA-related kinases Nek6, Nek7, and Nek9, and the motor proteins Eg5 (also known as Kif11) and dynein (Figure 2). As mentioned above, Cdk1 and Plk1 play a central role in centrosome disjunction as well, designating these kinases as critical coordinators of centrosome dynamics during spindle formation. Here we will describe the various mechanisms by which protein kinases control Eg5 to push centrosomes apart, while cortical and nuclear envelope (NE)-associated dynein forces pull the centrosomes apart.

Figure 2. Control of Eg5 recruitment to centrosomes to induce movement.

Phosphorylation at Thr926 and Ser1033 is a requirement for Eg5 accumulation at centrosomes and subsequent localization to microtubules once the centrosomes start to move apart. Nek9 activation at prophase is a key timer of Ser1033 phosphorylation. 1) Nek9 is activated when Cdk1 phosphorylates Ser869 (activating cyclins unknown). 2) This allows for binding of Plk1, phosphorylation of Thr210 in the activation loop of Nek9 and subsequent accumulation of activated Nek9 at duplicated centrosomes. 3) As a binding partner of Nek9, LC8 acts to inhibit Nek9 by preventing Nek6 and Nek7 binding. Nek9 autophosphorylation at Ser944 releases LC8, allowing for Nek9 to activate Nek6 by phosphorylating Ser206. A similar mechanism is thought to mediate Nek7 phosphorylation at Ser195. Activated Nek6 and Nek7 are enriched at centrosomes, where they bind to and phosphorylate Eg5 at Ser1033. Besides an indirect role in Eg5 accumulation through regulation of Nek9, Cdk1 is directly implicated in Eg5 loading onto centrosomes through modification of Eg5 Thr926. Again, the activating cyclins remain to be established, with cyclin A2, cyclin B1 and B2 representing the obvious candidates. Additional question marks in the model indicate other unknown aspects of Eg5 loading control that warrant further investigation. For instance, is the pool of centrosome-associated Plk1 that regulates Mst2 activation responsible for activation of Nek9? If so, this would place cyclin B2-Cdk1 and Aurora A in control of both centrosome disjunction and movement. Furthermore, besides LC8, are other proteins implicated in the control of Nek6 and Nek7 activation? Are other Eg5 phosphorylation sites beyond T926 and S1033 implicated in Eg5 loading? Is Eg5 loading solely controlled by Eg5 phosphorylation or does it also involve associations with other proteins, including proteins associated with centrosomes during mitosis or proteins that interact with Eg5 in the mitotic cytosol prior to loading?

Nek9-Nek6/7 signaling cascade

A subset of NIMA-related kinases constitutes a kinase cascade that acts to modify Eg5 at Ser1033. This modification is a key requirement for the accumulation of Eg5 at disjointed sister centrosomes. Nek9 becomes engaged in this process early in mitosis when phosphorylation of Nek9 at Ser869 by Cdk1 allows for binding to Plk1 (Figure 2). Subsequent phosphorylation of Thr210 in the activation loop of Nek9 by Plk1 results in Nek9 accumulation at centrosomes and turns on its enzymatic activity (Bertran et al., 2011). As an alternative mechanism of Nek9 activation it has been proposed that Plk1 may target residues outside the Nek9 activation loop, thereby disrupting an auto-inhibitory conformation that prevents self-phosphorylation at Thr210 (Roig et al., 2005). Before Nek9 is fully active, it needs to be released from its inhibitory binding protein LC8, an essential component of the cytoplasmic dynein motor protein complex. Nek9 autophosphorylation on Ser944 has been documented to release LC8 and allows for binding and catalytic activation of Nek6 (Regue et al., 2011). Thus, activated Nek9 binds, phosphorylates, and activates Nek6 and Nek7 (Figure 2), two NIMA-related protein kinases that share about 80% homology and are known to be required for spindle assembly (Belham et al., 2003; Roig et al., 2005).

Nek9-mediated phosphorylation of Ser206 in the activation loop of Nek6 turns on kinase activity (Belham et al., 2003). A similar activation mechanism has been proposed for Nek7 (Belham et al., 2003; Roig et al., 2005; Roig et al., 2002). Both Nek6 and Nek7 are enriched at centrosomes, where they bind to and phosphorylate Eg5 at Ser1033, (Bertran et al., 2011; Rapley et al., 2008). In this context, Nek6 and Nek7 are not fully redundant as depletion of either protein impaired mitotic spindle formation, pole movement and spindle assembly, and delayed metaphase progression (Bertran et al., 2011; O’Regan and Fry, 2009; Yissachar et al., 2006). Interactome analysis and gene knockout studies in mice are consistent with the idea that Nek6 and Nek7 have independent roles (de Souza et al., 2014; Salem et al., 2010). For instance, Nek6^−/− mice die early in embryogenesis, whereas Nek7^−/− mice die late in development or shortly after birth. However, differential spatio-temporal tissue distribution, subcellular localization, and enzymatic control have been reported for Nek6 and Nek7 (Feige and Motro, 2002; Minoguchi et al., 2003; O’Regan and Fry, 2009), which could account for the phenotypic differences observed in knockout mice rather than functional divergence.

Cdk1

Cdk1 is a major conductor of chromosome segregation, controlling the activities of many components of the underlying processes, including centrosome disjunction and movement (Bruinsma et al., 2012). However, due to the apparent multi-faceted nature of its functions, it is inherently difficult to study individual Cdk1 roles in isolation, particularly with only limited tools available, such as Cdk1 inhibitors and RNA interference. Furthermore, studies on the mitotic functions of Cdk1 provide little or no information about the activating cyclin involved (cyclin A2, cyclin B1 or cyclin B2). In general, mitotic cyclins are largely viewed as redundant or partially redundant (Hochegger et al., 2008). For example, cyclin B1 may fully compensate for the loss of cyclin B2 in mice, as cyclin B2 knockout mice are viable and apparently normal (Brandeis et al., 1998). However, this view of redundancy was recently challenged when cyclin B2, but not cyclin B1, was found to be the unique activating binding partner of Cdk1 in the Cdk1-Aurora A-Plk1-Mst2-Nek2A signaling cascade that induces centrosome disjunction (Nam and van Deursen, 2014). Thus, while it is unlikely that centrosome disjunction is prevented in cyclin B2 knockout mice, the extent to which this process is disturbed remains to be determined by further studies on these animals. As mentioned above, Cdk1 also plays a role in the activation of Nek9 through phosphorylation of Ser869, which subsequently leads to Nek9 phosphorylation at Thr210 by Plk1. However, the activating cyclin in this case is not known. Catalytically active cyclin B2-Cdk1 resides at centrosomes as the upstream activating kinase complex of Plk1 (through Aurora A). Therefore, it is tempting to speculate that it is cyclin B2 that serves as the master activator of the Plk1-Nek9-Nek6/Nek7 signaling cascade that mediates Eg5 phosphorylation, accumulation at centrosomes and centrosome movement (Figure 2). It will be interesting to further explore this possibility. In addition to controlling Eg5 loading through Nek6/Nek7-mediated phosphorylation of Ser1033, Cdk1 also has a more direct role in the recruitment process by phosphorylating Eg5 at Ser926 (Blangy et al., 1995; Smith et al., 2011). Again the activating cyclin implicated in this process remains to be determined (Figure 2).

Motor proteins in centrosome movement

The main motor proteins involved in centrosome movement are Eg5 and dynein. Eg5 is a tetrameric plus-end-directed motor protein that generates an outward pushing force that drives the centrosomes apart (Tanenbaum and Medema, 2010). At the same time, cortical and NE-associated dynein move centrosomes apart through a minus-end-directed force (Kotak et al., 2012; van Heesbeen et al., 2013). Eg5, which belongs to the BimC-family of motor proteins (Le Guellec et al., 1991), accumulates at centrosomes and astral microtubules upon phosphorylation at Thr926 and Ser1033 (Figure 2), to drive the sliding of overlapping anti-parallel polar microtubules, thus pushing the poles apart (Kapitein et al., 2005). Cdk1 phosphorylates nearly all Eg5 at Thr926 during early mitosis, whereas Nek6 and Nek7 seemingly phosphorylate only about 3% of the Eg5 pool at Ser1033, primarily at the spindle poles (Rapley et al., 2008). However, phosphorylation by both proteins is required for proper formation of the mitotic spindle and mitotic progression (Rapley et al., 2008).

Dynein is a minus-end directed motor protein that localizes to many distinct subcellular compartments. It is loaded to the plus-ends of growing microtubules, the cell cortex, and the NE. Among these, NE-associated dynein might be more closely involved in centrosome movement, pulling centrosomes along the NE towards opposite sides of the nucleus (Gonczy et al., 1999; Robinson et al., 1999; Splinter et al., 2010; Tanenbaum et al., 2010). However, this pulling force could also be produced by dynein localized at the cell cortex. In addition, NE-associated dynein has been implicated in NEBD itself and interaction between the NE and dynein is responsible for efficient chromosome capturing and alignment after NEBD (Beaudouin et al., 2002; Tanenbaum et al., 2010).

Spatial and temporal functions of dynein are regulated by adaptor proteins such as dynactin, lissencephaly, NudE and BICD2 (Kardon and Vale, 2009). Of these, dynactin is the best-characterized interactor of dynein. Although dynactin contributes to dynein recruitment to the NE and kinetochores it is dispensable for dynein’s function in organizing spindle microtubules (Raaijmakers et al., 2013). Centrosome movement is further facilitated by distributing nuclear pore complex (NPC)-associated dynein forces tethered to lamins, with studies in lamin-deficient mice revealing that NPCs in the vicinity of centrosomes are subject to aggregation late in G2 and prophase (Guo and Zheng, 2015).

ABERRANT CENTROSOME DYNAMICS AND CANCER

Traditionally, investigations on the relationship between centrosome aberrations and neoplastic growth have primarily focused on supernumerary centrosomes resulting in dysfunctional spindles, chromosomal instability and aneuploidization, a hallmark of human cancers. However, data from recent studies raise the possibility that ill-timed centrosome disjunction or movement may represent another source of chromosomal instability with major implications for neoplastic transformation (Nam and van Deursen, 2014; Silkworth et al., 2012; Zhang et al., 2012). Silkworth and Cimini initially found that insufficient centrosome separation at the time of NEBD increases the risk for merotelic microtubule-kinetochore misattachments that produce lagging chromosomes (Silkworth et al., 2012). Consistent with this pioneering work, various mouse models for abnormal timing of centrosome separation are characterized by formation of asymmetric mitotic spindles that inaccurately segregate chromosomes and tumor predisposition.

One of these models involves the deubiquitinase Usp44 (Zhang et al., 2012). Mouse embryonic fibroblasts (MEFs) lacking Ups44 exhibit incomplete centrosome disjunction, abnormal spindle positioning, chromosome lagging and aneuploidy, while the corresponding mice are tumor-prone. How Usp44 controls centrosome disjunction remains unclear. A second mouse model, involving cyclin B2 overexpression, has the opposite problem in that centrosomes separate too early (Nam and van Deursen, 2014). Strikingly, the resulting mitotic defects were virtually identical to those of delayed centrosome separation, as was the cancer predisposition phenotype. These findings seem to suggest that both accelerated and delayed centrosome separation drive tumorigenesis in vivo, and imply that proper timing of centrosome separation is critical in tumor suppression. In contrast, in vitro studies in HeLa cells cultured in the presence of EGF suggest that EGFR signaling accelerates centrosome separation and mitosis, and promotes the accuracy of chromosome segregation, but this effect is only seen in genetically unstable cells (Mardin et al., 2013). For the in vivo models discussed above it remains possible that chromosome segregation is affected through another mechanism besides centrosome separation. For example, Cyclin B2/Cdk1 activated Plk1 is a multi-functional kinase that controls proper chromosome segregation through various mechanisms, including spindle positioning (Kiyomitsu and Cheeseman, 2012), regulation of microtubule-kinetochore attachment error correction (Suijkerbuijk et al., 2012) and stabilization of microtubule-kinetochore attachment (Liu et al., 2012). Furthermore, the identification of different targets of Usp44 will help to better understand by which mechanisms its inactivation affects chromosome segregation.

These observations raise the question whether abnormalities in the centrosome movement process produce mitotic and cancer phenotypes similar to those seen in models with defects in centrosome disjunction. A mouse model for aberrant centrosome movement involving Eg5 overexpression suggests that this indeed might be the case. Eg5 transgenic mice display abnormal spindle development, which leads to chromosomal missegregation and increased tumorigenesis (Castillo et al., 2007). It will be important to further address centrosome movement defects in relation to mitosis and cancer using other animal models, for instance mutant mice in which key Eg5 regulators such as NIMA-related kinases are targeted. As mentioned above, EGFR signaling has been shown to influence the timing of centrosome separation, with robust signaling obviating the need for Eg5 in spindle assembly in select cell types (Mardin et al., 2013). Given that the EGF pathway has been linked to tumor growth (Hanahan and Weinberg, 2011), these findings raise the possibility that one of the mechanisms by which aberrant EGFR signaling promotes tumorigenesis involves timing of centrosome separation and its potential impact on the accuracy of chromosome separation. The relation between cancer and dynein, another key regulator of centrosome movement, remains understudied mainly because genetic models that directly study the involvement of dynein in oncogenesis are lacking. Indeed, homozygous null mice for the cytoplasmic dynein heavy chain gene exhibit early embryonic lethality, whereas heterozygotes show no obvious abnormalities (Chen et al., 2007; Hafezparast et al., 2003).

CONCLUSIONS AND PERSPECTIVES

The centrosome is the microtubule-organizing center of animal cells. Besides its role in the control of cell shape, cell polarity, cell motility, and immune synapse formation, it plays an important role in the formation of the mitotic spindle, which is necessary for distributing duplicated chromosomes equally into two daughter cells. Recently, several studies have implicated centrosome dynamics as a source of chromosomal stability and tumor suppression. This led to the hypothesis that improper centrosome dynamics, including delayed and premature centrosome disjunction or movement, stimulates tumorigenesis by promoting chromosomal instability through the formation of mitotic spindles with asymmetrically positioned poles, thus increasing the incidence of merotelic attachments and lagging chromosomes. The observation that lagging chromosomes are hallmarks of cancer cells supports this hypothesis and warrants further attention to three areas of investigation. First, it will be important to expand our understanding of the cellular processes and mechanisms that directly or indirectly control centrosome dynamics. As documented here, important players have been identified but the true level of complexity is unclear. If the elaborate collection of proteins associated with duplicated centrosomes is an indication for the complexity of the centrosome dynamics control system, then much remains to be discovered in this area. Second, assuming that deregulation of centrosome dynamics is a significant driver of aneuploidization during tumorigenesis, it will be important to determine which components of the machinery are frequently targeted for deregulation in human cancers. As suggested by recent work on cyclin B2 (Nam and van Deursen, 2014), potential candidates are modulators of centrosome dynamics, whose deregulation is a characteristic of molecular signatures that predict tumor aggressiveness and poor prognosis in a broad spectrum of human cancers. Third, how could the insights gained be exploited in cancer treatment? In this context, it will be interesting to test whether tumor cells carrying genetic changes that delay centrosome movement may be hypersensitive to small molecule inhibitors of key motor proteins implicated in spindle assembly, such as Eg5. Overall, centrosome dynamics seems to be emerging as an exciting new area for mechanistic discovery and therapeutic opportunity.

ABBREVIATIONS

Cdk1

Cyclin-dependent kinase 1

Cep85

centrosome protein 85 KDa

EGF

Epidermal growth factor

EGFR

Epidermal growth factor receptor

GRK2

G protein-coupled receptor kinase 2

Sav1

Scaffold protein Salvador

KT

Kinetochore

MEF

Mouse Embryonic Fibroblast

Mst2

mammalian sterile 20-like kinase 2

NE

Nuclear Envelope

NEBD

Nuclear envelope breakdown

Nek2A

NIMA family kinase 2A

NIMA

Never in Mitosis A

NPC

Nuclear pore complex

PCM

Pericentriolar material

Plk1

Polo-like kinase 1

TEIF

transcriptional element-interacting factor