Research

The regulation of centrosome number and function underlies bipolar mitotic spindle formation and genetic integrity. In normal cells, the centrosome is usually composed of a pair of barrel-shaped structures, the centrioles, which are embedded in an electron-dense amorphous matrix, the pericentriolar material. The latter provides the site for microtubule nucleation and therefore strongly influences microtubule numbers and organisation throughout the cell cycle. Malignancies exhibit a wide variety of centrosome abnormalities that range from numerical and structural to functional and positional aberrations. It is not yet understood how these abnormalities arise in cancer and how they contribute to tumourigenesis. We have two basic goals in the laboratory. First, we want to address what the consequences are of deregulation of centrosomal proteins on mitotic spindle formation and genome stability. Second, we want to study the molecular mechanisms and signals that govern centrosome behaviour in response to environmental cues. These goals aim to provide insight into basic biological processes whose deregulation is implicated in the development of cancer.

CDK5RAP2: linking centrosomes with mitotic spindle poles

To investigate the function of specific centrosomal proteins during cell division we have turned to the chicken B-cell line, DT40. These vertebrate cells have an exceptionally high ratio of targeted versus random integration of transfected DNA constructs, and are therefore genetically manipulable. DT40 cells have provided us with particularly exciting results regarding the function of the highly conserved centrosomal protein, CDK5RAP2. In search of a mitotic role for CDK5RAP2, we have deleted two evolutionarily highly conserved regions of the cdk5rap2 gene in DT40 cells. Both cdk5rap2 mutants exhibited poor cell viability and a decrease in the clonogenic potential of individual cells. However, in these mutants the most striking phenotype became apparent only during mitosis. As soon as cells began to assemble a bipolar mitotic spindle, the centrosomes detached from the mitotic spindle poles (Figure 1). This phenotype gradually worsened as mitosis progressed and, by anaphase, the centrosomes detached from the majority of the mitotic spindle poles. We therefore propose that CDK5RAP2 serves as a platform for microtubule anchoring within the mitotic centrosome, thereby connecting centrosomes and spindle poles (Barr et al., J Cell Biol 2010; 189: 23). During the course of these studies we identified at least two centrosome components that depended on CDK5RAP2 for their centrosomal localisation, however the molecular mechanism by which CDK5RAP2 links centrosomes and mitotic spindle poles is not fully understood and is the subject of our current investigations. Importantly, mitotic spindle positioning relies on astral microtubules, polymers that emanate from the centrosomes and interact with the cell cortex. Centrosome detachment from mitotic spindle poles, however, leads to mitotic spindles without astral microtubules. We therefore predict that cdk5rap2 mutant cells will be defective for spindle positioning. This is an exciting possibility, since the ability of a cell to position its mitotic spindle is central to asymmetric cell division, a process that generates daughter cells with different fates. Normal homeostasis of many of our tissues is dependent on controlled asymmetric cell division, and deregulation of this process occurs in malignancies.

Figure 1
CDK5RAP2 is required to connect centrosomes with mitotic spindle poles. Top panels show a wild-type DT40 cell in which both centrosomes are positioned within the mitotic spindle poles. Bottom panels show a cdk5rap2-mutant DT40 cell in which one centrosome is fully detached from the mitotic spindle pole. Yellow arrows indicate centrosomes that are located within spindle poles. The pink arrow shows a fully detached centrosome located near the cell cortex, while the pink asterisk highlights a spindle pole that does not contain a centrosome. The centrosome is marked with centrin antibody that recognises individual centrioles, whereas microtubules are stained with anti-alpha-tubulin antibodies. DNA is blue in merged image. Scale bar is 5μm.

CDK5RAP2 and DNA damage signalling

In order to maintain their genome stability, cells respond to DNA damage by activating a complex signalling cascade that triggers cell cycle checkpoints, repair of damaged lesions or apoptosis in cases of irreparable damage. Intriguingly, at least three centrosomal components have been reported to be required for cell cycle arrest in G2 in response to DNA damaging agents. We have now shown that CDK5RAP2 also belongs to this group of proteins, since following irradiation cdk5rap2 mutant cells exhibit a reduced capacity to maintain G2 arrest. Thus, in addition to its role during mitosis, CDK5RAP2 is also an integral part of a DNA damage-sensing signalling network at the centrosome. It is unclear how CDK5RAP2 promotes an effective G2 checkpoint, but our working hypothesis is that its role in the process involves the recruitment and/or maintenance of Chk1 at the centrosome. Indeed, levels of Chk1 kinase are reduced in the centrosomes of cdk5rap2 mutant cells. Chk1 kinase is central to the DNA damage checkpoint in cells, and its centrosomal accumulation prevents premature mitotic entry not only in the presence of DNA damage but also during unperturbed cell cycles. We now want to understand the molecular mechanism by which CDK5RAP2 regulates Chk1 levels at the centrosome and address the broader role of the centrosome in DNA damage sensing.

CEP63 controls centrosome duplication

Centrosomes duplicate once and only once per cell cycle. Centrosome duplication is under strict control in cells, like DNA replication, since abnormal centrosome number can lead to tumour formation, mitotic catastrophe and aneuploidy. Despite its importance, control of centrosome duplication is still poorly understood. Our most recent work reveals a new regulator of this process, the core centrosomal protein, CEP63. We generated cells that are deficient in CEP63 and found that these cells contained too few centrosomes. They grew more slowly than control cells and displayed monopolar instead of bipolar spindle formations indicative of abortive centrosome duplication cycles (Figure 2). We are currently investigating where CEP63 fits into the molecular network that controls centrosome number in cells.

Figure 2
CEP63 is required for centrosome duplication. Top panels show a wild-type DT40 cell in which two centrosomes are visible each within a pole of a bipolar mitotic spindle. In contrast, bottom panels show a cep63-mutant DT40 cell in which only one centrosome is present (yellow arrow) and therefore the mitotic spindle is monopolar. The centrosome is marked with an antibody that recognises CDK5RAP2, whereas microtubules are stained with anti-alpha-tubulin antibodies. DNA is blue in merged image. Scale bar is 5μm.