Centrosomal and Non-centrosomal Functions Emerged through Eliminating Centrosomes

Abstract

Centrosomes are highly conserved organelles that act as the major microtubule-organizing center (MTOC) in animal somatic cells. Through their MTOC activity, centrosomes play various roles throughout the cell cycle, such as supporting cell migration in interphase and spindle organization and positioning in mitosis. Various approaches for removing centrosomes from somatic cells have been developed and applied over the past few decades to understand the precise roles of centrosomes. Centrinone, a reversible and selective PLK4 (polo-like kinase 4) inhibitor, has recently emerged as an efficient approach to eliminate centrosomes. In this review, we describe the latest findings on centrosome function that have been revealed using various centrosome-eliminating approaches. In addition, we discuss our recent findings on the mechanism of centrosome-independent spindle bipolarization, discovered through the use of centrinone.

Centrosome cycle and its function

The centrosome is a highly conserved membrane-less organelle, which is composed of one or two centrioles surrounded by pericentriolar material (PCM) (Conduit et al., 2015; Nigg and Holland, 2018; Nigg and Raff, 2009). Centrosomes mainly function as the major microtubule-organizing center (MTOC) in somatic animal cells (Conduit et al., 2015), and undergo multiple events throughout the cell cycle (Fig. 1). Centrioles are duplicated once per cell cycle during S phase (Nigg and Holland, 2018). In this duplication process, only a single daughter centriole is formed next to each mother centriole (Ohta et al., 2014; Rattner and Phillips, 1973). Several factors critical for centriole duplication have been identified, such as PLK4, SAS6, STIL, CPAP, and CEP152 (Bettencourt-Dias et al., 2005; Cizmecioglu et al., 2010; Dammermann et al., 2004; Dzhindzhev et al., 2010; Habedanck et al., 2005; Hatch et al., 2010; Kleylein-Sohn et al., 2007; Leidel et al., 2005; Lin et al., 2013; Stevens et al., 2010; Tang et al., 2011). After centriole duplication, the process of centrosome maturation is initiated in G2 phase with an extreme expansion of PCM, allowing centrosomes to acquire robust MTOC activity (Conduit et al., 2014; Izquierdo et al., 2014; Paweletz et al., 1984). At the G2/M transition, the two centrosomes are separated from each other and move to the opposite sides of the mitotic cell (centrosome separation) (Raaijmakers et al., 2012; Rattner and Berns, 1976; Toso et al., 2009; Waters et al., 1993; Whitehead and Rattner, 1998). During mitosis, centrosomes localize at the mitotic spindle poles (Nigg and Holland, 2018). Following cell division, the connection between the mother and daughter centrioles is resolved (centriole disengagement) (Tsou et al., 2009; Tsou and Stearns, 2006; Vidwans et al., 1999), and the daughter centriole recruits PCM proteins and becomes a mother centriole (daughter-to-mother centriole conversion). After the conversion process, both centrioles obtain the ability to duplicate (Fu et al., 2016; Izquierdo et al., 2014; Tsou et al., 2009; Tsou and Stearns, 2006; Tsuchiya et al., 2016). Overall, centrosomes undergo many processes throughout the cell cycle and their number is strictly regulated in animal somatic cells.

Fig. 1.

The centrosome cycle. In mitosis, a bipolar spindle is formed and centrosomes function as the poles of the spindle. After cell division, the connection between the mother and daughter centrioles is resolved (centriole disengagement), and the daughter centriole recruits PCM proteins and becomes a mother centriole. When cells stop proliferating and enter G0 phase, centrosomes act as basal bodies for the growth of cilia (ciliogenesis). In S phase, the process of centriole duplication is promoted by several factors (e.g., PLK4, SAS6, STIL, CPAP, and CEP152). Starting at G2 phase, the two centrosomes experience extreme PCM expansions, which allow for stronger MTOC activity (centrosome maturation). At G2/M transition, the two centrosomes are separated from each other and move to the opposite sides of the mitotic cell (centrosome separation), followed by initiation of the next cell division process.

Centrosomes play various crucial roles in animal cells through their MTOC activity. During interphase, centrosomes regulate the microtubule network and govern cell motility (Werner et al., 2017). During mitosis, centrosomes function as the poles of the mitotic spindle to maintain spindle structure and spindle positioning for correct chromosome segregation (Conduit et al., 2015).

Centrosome-eliminating approaches developed over the past few decades

To investigate the necessity of centrosomes in animal somatic cells for proper mitotic spindle formation and chromosome segregation, several groups have eliminated centrosomes from these cells through different technical approaches. Traditionally, microsurgery or depletion of centriole duplication factors have been used to remove centrosomes (Table I). Khodjakov et al. succeeded in selectively removing centrosomes without damaging surrounding structures by laser microsurgery targeting GFP-tagged γ-tubulin (a centrosome marker) in mammalian cells (Khodjakov et al., 2000). They reported that bipolar mitotic spindles were formed without centrosomes. Basto et al. created DSas-4 (an ortholog of human CPAP) mutant flies, whose cells do not have centrosomes. They confirmed that centrosomes are not necessary for most aspects of Drosophila development (Basto et al., 2006). Hornick et al. performed microsurgery with a microneedle to obtain acentrosomal mammalian cells (Hornick et al., 2011). They found that the fidelity of bipolar mitotic spindle assembly is highly compromised in the absence of centrosomes. Sir et al. obtained chicken cells lacking centrosomes through knockout of CEP152 or STIL (Sir et al., 2013). These cells showed delays in bipolar spindle formation and high rates of chromosome segregation errors. In summary, the presence of centrosomes in animal somatic cells is important for proper mitotic progression to some extent, but, in many cases centrosomes are not essential. However, the detailed mechanisms of centrosome-independent bipolar spindle formation in somatic cells have not been extensively discussed.

Table I.

Examples of studies on acentrosomal somatic cells

Species/cell	Approach	Main finding	Reference
Monkey/CVG-2	Laser microsurgery	A bipolar mitotic spindle is formed in the absence of centrosomes.	Khodjakov et al., 2000
Fly	Mutant of DSas-4	Centrosomes are not necessary for the development of the fly.	Basto et al., 2006
Monkey/BSC-1	Needle microsurgery	The fidelity of the bipolar mitotic spindle is highly compromised in the absence of centrosomes.	Hornick et al., 2011
Chicken/DT40	Knockout of CEP152/STIL	Centrosome loss leads to delays in spindle formation and high rates of chromosomal instability.	Sir et al., 2013

Emergence of centrinone: a reversible and selective inhibitor of PLK4

All the above methods for the removal of centrosomes require special devices for microsurgery or efficient systems for gene knockdown or knockout. In contrast, centrinone, a reversible and selective PLK4 inhibitor developed in 2015 (Wong et al., 2015), has enabled the easy removal of centrosomes from living animal cells. Wong et al. selected VX-680, a pan-Aurora kinase inhibitor which also inhibits PLK4 (Harrington et al., 2004; Sloane et al., 2010), as a template to develop selective PLK4 inhibitors. Through the introduction of a methoxy substituent at the VX-680 C5 position and further optimization, they succeeded in synthesizing centrinone and centrinone-B. Both inhibitors exhibited >1,000-fold selectivity for PLK4 over Aurora kinases (Wong et al., 2015). After treatment with centrinone, they observed cell cycle arrest in G1 phase in normal human cells (RPE-1 cells) and confirmed that the arrest was not due to previously described stress responses to DNA damage, Hippo signaling, or chromosome segregation errors. On the other hand, they noted that cancer cell lines could proliferate in the absence of centrosomes, and the proliferation rates were not correlated with the basal frequency of centrosome amplification observed in those cell lines. These results suggest that numerous cancer cells have abolished the response system that arrests the cell cycle after centrosome loss.

Since the advent of centrinone, other various studies have used this agent (Table II). Meitinger et al. performed a genome-wide CRISPR/Cas9 screen in RPE-1 cells and identified a 53BP1-USP28 module which induces G1-phase cell cycle arrest after centrosome loss (Meitinger et al., 2016). Two other groups identified the same pathway by using another method to remove centrosomes from human somatic cells (Fong et al., 2016; Lambrus et al., 2016). Gheiratmand et al. performed BioID experiments targeting seven centriolar satellite components while treating cells with centrinone. BioID is a method to screen for protein interactions in living cells using biotin ligase. The ligase fused to the protein of interest biotinylates proximal proteins and enables their isolation and identification (Roux et al., 2012). Using this technique, they revealed that the interactions among centriolar satellite proteins are not affected by centrosome removal (Gheiratmand et al., 2019). Another group performed similar experiments using the knockout of STIL or CEP152 to remove centrosomes, and reached the same conclusion (Quarantotti et al., 2019). Collectively, these results suggest that treatment with centrinone is as effective as other existing methods. The Golgi apparatus also exhibits MTOC activity and localizes around centrosomes (Chabin-Brion et al., 2001; Efimov et al., 2007; Rivero et al., 2009; Ríos et al., 2004). Centrinone has been used in several studies to investigate the relationship and dependency between the two organelles (Guizzunti and Seemann, 2016; Wu et al., 2016). Moreover, Dudka et al. produced human cells which have only one centrosome during mitosis through treatment with centrinone (Dudka et al., 2019). They used these cells as a tool for studying k-fiber dynamics in cells with asymmetric bipolar spindles and found that the centrosome regulated k-fiber plus-ends in a HURP-dependent manner. Overall, owing to its ease of use, centrinone has made a huge impact on the latest developments in cell biology. However, a detailed study on the mechanisms of acentrosomal spindle formation using centrinone has not been performed.

Table II.

Examples of studies using centrinone and human cells

Cell	Approach	Main finding	Reference
BT549, Calu-6, HCT116, HeLa, MDA-MB-231, RPE-1, U2OS	Centrinone	Centrosome loss leads to G1-phase cell cycle arrest in normal cells, but not in cancer cells.	Wong et al., 2015
SV589 HeLa	Centrinone	The link between the Golgi apparatus and centrosomes must be dissolved to reach metaphase.	Guizzunti and Seemann, 2016
RPE-1	Centrinone	A single Golgi can be maintained in the absence of centrosomes.	Wu et al., 2016
RPE-1	Centrinone CRISPR/Cas9 screen	The 53BP1-USP28 module induces G1-phase cell cycle arrest after centrosome loss.	Meitinger et al., 2016
HEK293T	Centrinone BioID	The interactions of centriolar satellite proteins are not affected by centrosome loss.	Gheiratmand et al., 2019
RPE-1	Centrinone	Centrosomes indirectly regulate k-fiber plus-ends via spindle length-dependent accumulation of HURP.	Dudka et al., 2019
A549, DU145, GI-1, HCT116, HeLa, MCF-7, PANC-1, SKOV-3, U2OS	Centrinone	NuMA-mediated pathways promote spindle bipolarity independently of centrosomes.	Chinen et al., 2020

A centrosome-independent mechanism of spindle bipolarity establishment revealed using centrinone

Based on the aforementioned evidence, we used centrinone to investigate the machinery of bipolar spindle formation in acentrosomal human cells. The establishment of a bipolar spindle is crucial for equational chromosome segregation (Prosser and Pelletier, 2017). In somatic cells, spindle bipolarity is achieved through centrosome separation, which occurs ahead of or following nuclear envelope breakdown (NEBD) (Beaudouin et al., 2002; Kaseda et al., 2012; Raaijmakers et al., 2012; Rattner and Berns, 1976; Toso et al., 2009; Waters et al., 1993; Whitehead and Rattner, 1998). During centrosome separation, two centrosomes linked by antiparallel microtubules are pushed away from each other through microtubule crosslinking and the sliding activity of kinesin Eg5 (Bertran et al., 2011; Hata et al., 2019; Kapitein et al., 2005; Kaseda et al., 2012; Smith et al., 2011). During mitosis, these separated centrosomes function as the core of spindle poles and organize microtubules into a bipolar spindle.

On the other hand, meiotic spindles are formed without centrosomes in the oocytes of many species, including flies, frogs, mice, and humans (Calarco-Gillam et al., 1983; Heald et al., 1996; Holubcová et al., 2015; Matthies et al., 1996). In addition, the somatic cells of flies and vertebrates can organize functional bipolar spindles following the removal of centrosomes by lasers, microneedles, or mutations of centriole duplication factors (Basto et al., 2006; Bonaccorsi et al., 2000; Hornick et al., 2011; Khodjakov et al., 2000; Sir et al., 2013). Therefore, it appears that the establishment of spindle bipolarity is not completely dependent on centrosome separation, and an acentrosomal pathway may exist as a compensation mechanism. The in vitro reconstitution of meiotic spindles utilizing Xenopus egg extracts has been a useful model system for the study of acentrosomal spindle formation (Cross and Powers, 2009). In this system, microtubules are nucleated via chromatin- and Ran-based pathways (Carazo-Salas et al., 1999; Heald et al., 1996; Kalab et al., 1999; Karsenti et al., 1984; Wilde, 1999; Zhang et al., 1999) and are presumably rearranged into a bipolar state in a microtubule- and motor protein-dependent manner (Groen et al., 2008; Heald et al., 1996; Loughlin et al., 2010; Petry et al., 2011). In addition, in mouse oocytes, multiple MTOCs that contain PCM are formed de novo, and these MTOC foci are eventually clustered into a bipolar state in an Eg5-dependent manner (Schuh and Ellenberg, 2007). However, other factors involved in acentrosomal spindle bipolarization, besides microtubules, motor proteins, and PCM components have not been extensively investigated thus far.

NuMA (nuclear mitotic apparatus) protein, one of the spindle pole components, directly interacts with microtubules (Du et al., 2002; Haren and Merdes, 2002) and presumptively organizes centrosome-independent microtubule asters (Du et al., 2002; Gaglio et al., 1995; Haren and Merdes, 2002). This protein is necessary for focusing microtubules at spindle poles (spindle pole organization) (Merdes et al., 2000, 1996). In addition, cortical NuMA creates spindle-pulling forces in coordination with the dynein-dynactin complex, placing the spindle at an appropriate position within the cell (Du and Macara, 2004; Okumura et al., 2018). However, the role of NuMA in spindle bipolarization, rather than spindle pole organization and spindle positioning, is not completely understood.

We treated human cells with centrinone to create artificial acentrosomal cells (Chinen et al., 2020). We confirmed that these cells displayed prolonged mitotic duration and increased frequency of chromosome segregation errors, consistent with previous studies using centrinone (Meitinger et al., 2016; Wong et al., 2015). These results suggest that mitotic fidelity is compromised in the absence of centrosomes in human cells as well as other vertebrates (Sir et al., 2013). Subsequently, we investigated the role of NuMA in spindle bipolarization in acentrosomal cells. We found that in acentrosomal cells, shortly after NEBD, NuMA formed several aggregates that organized microtubule asters. Subsequently, these aggregates assembled into two NuMA structures (initial bipolarity establishment), in a manner dependent on microtubules, dynein, and the clustering activity of NuMA. These two structures, located close to each other, organized a radial array of microtubules around them. This radial array of microtubules incorporated Eg5. Eventually, these two NuMA structures were separated to form a bipolar spindle. The separation of the two NuMA structures is dependent on Eg5 motor activity and kinetochore-microtubule attachment. Disruption of Eg5 motor activity or kinetochore-microtubule attachment resulted in fusion of the two NuMA structures after the initial bipolarity was established, leaving cells in a monopolar-like state. Following the depletion of NuMA in acentrosomal cells by auxin-inducible degradation (Natsume et al., 2016; Nishimura et al., 2009; Okumura et al., 2018), cells failed to establish bipolar spindles.

Importantly, we found that NuMA promoted the initial steps of spindle bipolarization in early mitosis even in cells with centrosomes (Chinen et al., 2020). We observed that the time from NEBD to bipolarity establishment prolonged upon depletion of NuMA. The results of our study suggest that the canonical centrosomal pathway and the NuMA-mediated acentrosomal pathway complementally regulate bipolar spindle assembly in somatic cells, and that the latter becomes predominant in the absence of centrosomes (Fig. 2). Understanding the multiple pathways that ensure robust bipolar spindle formation will assist in the design of anticancer drugs that target spindle assembly (Henriques et al., 2019; Tischer and Gergely, 2019).

Fig. 2.

Models for two pathways that promote spindle bipolarization in human cells. (A) The canonical centrosomal pathway. At G2/M transition, the two centrosomes are pushed apart by the plus-end-directed motor activity of kinesin Eg5. (B) The NuMA-mediated pathway, which occurs independently of centrosomes. (1) At the onset of mitosis, NuMA aggregates and organizes microtubule asters. (2) Dynein activity and the clustering activity of NuMA assembles the NuMA aggregates into two poles; subsequently, Eg5 is loaded onto the antiparallel microtubules. (3) Spindle poles are separated through Eg5 motor activity and kinetochore-microtubule attachments.

Concluding remarks

Since its development in 2015, centrinone has allowed for great advances in our understanding of centrosome function and the consequences of centrosome loss. Its ease of use has led to the success of numerous large-scale screens that shed light on the interplay between centrosomes and other organelles or cellular systems. Moreover, centrinone has enabled a detailed analysis of the cell division machinery in acentrosomal cells. This analysis assists us in understanding the complemental mechanisms through which the mitotic spindle is regulated by both centrosome-dependent and centrosome-independent pathways.