Meeting Report – Building a Centrosome

The Workshop covered most areas of centrosome biology, from centriole duplication and pericentriolar material (PCM) recruitment to the role of the centrosome in the nervous system, mitosis and cell migration. The structure of the primary cilia was also addressed, but its function was beyond the scope of the meeting. The limited number of participants (fewer than 30) combined with the isolation of the venue and the snowy weather produced ideal conditions for stimulating scientific discussions.

A prominent theme of the Workshop was the role of the Polo-like kinase Plk4 (and of its functional homolog ZYG-1 in Caenorhabditis elegans) in the process of centriole duplication. Plk4 is now recognised as a central player in this process by controlling both the timing of duplication as well as the number of newly formed centrioles. However, very few targets of Plk4 have been identified so far, and the mechanisms by which Plk4 activity is regulated are still not completely understood.

Another rapidly growing area of interest in the community is the study of primary microcephaly (MCPH), which is characterised in patients by a smaller, albeit structurally normal, brain and by intellectual disability. Interestingly, all causative genes identified so far encode centrosomal proteins, suggesting that centrosomal defects are the cause of the disease. Newly identified MCPH-associated genes were discussed during the meeting as well as data describing the consequences of knockdown or mutation of MCPH proteins on centrosomal function. Interestingly, a link seems to emerge between the centriole duplication pathway and microcephaly.

The first Workshop session was given by Erich Nigg (Biozentrum, University of Basel, Switzerland) with a talk that focused on key players in the initiation of centriole duplication. The role of mammalian Cep192 in centriole biogenesis has been controversial, but Nigg presented compelling data and argued that mammalian Cep192 is required for centriole duplication and that it cooperates with Cep152 in the recruitment of Plk4 to the centrioles.

As mentioned, Plk4 emerged as a key kinase triggering centriole duplication, and this function is likely to depend on its localisation, abundance and activity at the centrioles. Mónica Bettencourt-Dias (IGC, Oeiras, Portugal) has made considerable gains in understanding the cell cycle control of Plk4 protein levels using Drosophila melanogaster S2 cells as a model system. Extending previous work on Slimb ubiquitin-ligase-mediated Plk4 proteolysis, Bettencourt-Dias is now trying to characterise how Plk4 trans auto-phosphorylation promotes Slimb binding and acts as a mechanism by which Plk4 can regulate its own stability.

Pierre Gönczy (EPFL, Lausanne, Switzerland) continued with his presentation. His group unravelled the architecture of the basal body cartwheel in Trichonympha (flagellates that have exceptionally elongated basal bodies, with cartwheels larger than 4000 nm) by using cryo-electron tomography. Their stunning images reveal that the cartwheel is not a helical structure but comprises a stack of spindle assembly abnormal 6 (SAS-6) protein rings. Moreover, his group was able to identify extra density regions in the cartwheel that do not correspond to the SAS-6 protein and the identification of the proteins comprising these regions will be an exciting challenge. Gönczy ended his talk by announcing a ‘Centriole Screen Resource’. This database is based on a human-genome-wide RNAi screen that was carried out in his laboratory in order to identify new regulators of centrosome numbers. It is expected to be online in the near future.

Later on, Karen Oegema (UCSD, San Diego, CA) addressed the mechanism of centriole cartwheel formation in the C. elegans embryo. This talk was not without a little controversy as it presented evidence contradicting published data. The previous model suggests that kinase ZYG-1, the Plk4 homologue in C. elegans, drives cartwheel assembly by directly phosphorylating SAS-6, a main structural component of the cartwheel. Oegema's group dutifully identified the ZYG-1 phosphorylation sites on SAS-6 but, in contrast to embryos that express kinase-dead ZYG-1, embryos that express a mutant version of SAS-6 that lacks those phosphoprylaton sites do not display defects in centriole assembly. Thus, it seems that the crucial target of ZYG-1 for cartwheel formation in C. elegans remains to be discovered.

The final talk of the session was delivered by Ingrid Hoffmann (DKFZ, Heidelberg, Germany), who presented the functional analysis of human GCP6 (TUBGCP6; Dgrip163 in Drosophila), a core component of the γ-tubulin ring complex (γ-TuRC). GCP6 localises to the PCM but also to the centriole distal appendages and is required for Plk4-induced overduplication of centrioles. By using mass spectrometry, Hoffmann's laboratory identified ten Plk4-specific phosphorylation sites on GCP6 that are needed for centriole overduplication. Her group also identified the so far unknown centrosomal protein Cep78, which interacts with Plk4, as well as CP110 and Cep97 proteins.

The centrosome assembly session started with Tim Stearns (Stanford University, CA), whose laboratory focuses on the identification of new centrosome and cilia components by using transcriptomic and proteomic analysis. By examining the transcriptional profile of sperm and multi-ciliated mouse tracheal epithelial cells (MTECs), they identified many centrosome and cilia genes that are upregulated in these specialised cells. Stearns then introduced a new proximity labelling technique called BioID that had been developed by Kyle Roux (Sanford Research Center, Sioux Falls, CA). In brief, biotin protein ligase is fused to a targeting protein, which can transfer the biotin moiety to neighbouring and interacting proteins within the cell. The biotinylated proteins are then isolated by affinity capture and identified using mass spectrometry. Taking advantage of both techniques has enabled the identification of a number of new centrosomal proteins, including the distal appendage protein Ccdc146 and the Cep63 paralogue Ccdc67.

Jordan Raff (University of Oxford, UK) presented work that aims to identify the hierarchy of protein recruitment to the centrosome in Drosophila by using fluorescence recovery after photobleaching (FRAP). The intensity profiles of the studied centrosomal proteins showed gradients that concentrated around the centrioles. After FRAP, the fluorescence of most proteins recovered evenly throughout the gradient, indicating binding to a receptor that itself is distributed as a gradient. However, the fluorescence of Cnn (Cdk5Rap2), Asterless (Cep152) and Spd2 (Cep192) recovered first at the centriole and then diffused away, suggesting that those proteins act upstream in the formation of the PCM gradient. Raff's group is now coupling this approach to the knockout of individual centrosomal components in order to describe PCM recruitment and organisation.

Jeff Woodruff from Anthony Hyman's group (Max Planck Institute, Dresden, Germany) studies similar questions by using a different approach. Using purified PCM components from C. elegans, he presented work that aims to reconstitute PCM in vitro. The SPD-5 protein – which has no sequence-based homologue in flies and vertebrates, but is suspected to be the functional homologue of Cnn and Cdk5Rap2 – only formed small complexes when incubated on its own. However, adding PLK-1 and ATP strikingly induced the self-assembly of SPD-5 into large networks. These structures disassemble when a general phosphatase is added, arguing against non-specific aggregation of SPD-5. This approach should be a powerful one for understanding PCM formation as well as its structure.

For a centrosome workshop, it was surprising that only Ken Sawin (University of Edinburgh, UK) talked about the proteins that drive microtubule nucleation at the centrosome. He proposed that regulation through phosphorylation and dephosphorylation of the Mto1–Mto2 complex is central to cell cycle regulation of microtubule nucleation in fission yeast. Combining live-cell imaging of yeast cells that express the phosphorylation mutant Mto2 with the hydrodynamic analysis Mto1 and Mto2, his data suggest that phosphorylation of Mto2 inhibits microtubule nucleation by disrupting the formation of an active Mto1–Mto2 complex. Finally, Sawin introduced some work in progress, in which he attempts to reconstitute this complex in vitro and test its ability to enhance microtubule nucleation through the γ-tubulin complex.

The next day, the focus of the meeting was shifted slightly to concentrate on the centrosome with regard to mitosis. Proceedings were brought underway by Iain Hagan (Paterson Institute of Cancer Research, Manchester, UK). He presented work that showed how mitotic commitment is tied to events on the spindle pole body in fission yeast. Mitotic commitment is known to be dependent on the activation of the Cdk1–cyclin-B complex, and the data he presented delved deeper into this activation. More specifically, Hagan showed that Cut12, a spindle pole body protein, recruits the protein phosphatase PP1 and sets a threshold for the feedback-loop activity of Polo kinase Plo1 that maintains cells in interphase until the activation of the Cdk1–cyclin B complex.

Elmar Schiebel (University of Heidelberg, Germany) presented the second talk of the session and focused on the centrosome linker proteins. He presented unpublished work discussing the role of epidermal growth factor (EGF) in the regulation of centrosome separation through the Nek2–Mst2 pathway. His group had found that Mst2 kinase phosphorylates human Sav1, Nek2 and Cep250 and that phosphorylation of Nek2A by Mst2 kinase is important for centrosome splitting. These proteins cooperate with Eg5 for bipolar spindle assembly. He concluded with the surprising observation that EGF induces centrosome separation through the PI3K–Akt pathway, and that this effect can be suppressed by inhibiting both Eg5 and Kif15. These results raise the possibility that EGF inhibitors can improve the effect of Eg5 inhibitors in cancer therapy.

The recent discovery that the flatworm planaria do not possess centrosomes prompted Fanni Gergely to define some of the key cellular functions of the organelle. By engineering two DT40 chicken B lymphocyte lines with a knockout of CEP152 or STIL (two genes essential for centriole biogenesis), she generated cells that lack functional centrosomes. The characterisation of these cell lines revealed a new role for the centrosome in the organisation of the Golgi complex, and challenged the supposed link between centrosomes and the DNA damage response. Interestingly, Gergely noticed that cells without functional centrosomes had mitotic defects – including errors in chromosome segregation, which can lead to aneuploidy. Because CEP152 and STIL are often mutated in patients with primary microcephaly, these findings suggest that aneuploidy is a mechanism for the pathogenicity of this disorder.

Interesting findings about the formation of the mitotic spindle in early mouse embryos were provided by David Glover. Mouse embryos are naturally devoid of centrioles until the blastocyst stage, and spindle assembly occurs by a non-centrosomal pathway in the first few embryonic cleavage cycles. His team found that Plk4, a master regulator of centriole biogenesis that, therefore, influences centrosome-mediated spindle assembly, is essential for acentriolar, chromosome-mediated spindle assembly in the early mouse embryo. Moreover, Cep152, a Plk4-interacting partner, is also involved in acentriolar spindle assembly. Noticeably, although Plk4 is known to promote de novo centriole formation in unfertilised Drosophila eggs, Glover's team could not detect any centriole-like structures following overexpression of Plk4 in acentriolar early mouse embryos, pointing to a strict temporal regulation of centriole biogenesis in these early stages of mouse development.

The mitotic spindle was also the focus of the talk by Alexey Khodjakov (Wadsworth Center, Albany, NY), who used laser microsurgery to break the microtubule bundles between a kinetochore and the centrosome. By following subsequent chromosome oscillations with sophisticated live-imaging techniques, he made the exciting observation that a chromosome with a severed kinetochore fibre was still able to move to the pole during anaphase. These intriguing observations may modify the current models describing anaphase chromosomal movement.

The talk by Stephen Doxsey (University of Massachusetts, Worcester, MA) focused on the emerging role of centrosomes in the regulation of endocytosis and recycling of endosomes. Endosomal proteins have been observed at the centrosome, but the significance of this localisation is unclear. By using an in vitro method to reconstitute endosome protein complexes on isolated membrane-free centrosomes, his group demonstrated that the active GTP-bound form of Rab11 associates specifically with mother centriole appendages, and that Evi5 (Rab11 GAP) regulates Rab11 levels at the mother centriole. They also demonstrated that centriolar localisation of Evi5 depends on the mother centriole appendage protein centriolin (Cntrl), and that depletion of Cntrl or cenexin (Odf2) alters the organisation and function of recycling endosomes. Now, Doxsey's laboratory is investigating the contribution of Rab11 endosomes to the ciliary membrane.

In the afternoon, talks focused on the role of centrosomes in the nervous system. Cayetano Gonzalez (IRB, Barcelona, Spain) discussed the importance of centrosome asymmetry in Drosophila neuroblasts. In these cells, the daughter centrosome, which has high microtubule nucleation activity, is retained in the neuroblast at mitosis. Gonzalez showed that centrobin (CNB), which selectively localises to the daughter centrosome, has a key role in this behavior. When CNB is knocked out, both centrosomes behave as mother centrosomes, whereas when CNB is artificially targeted to the two centrosomes, they both behave as daughter centrosomes. Finally, he showed that this function for CNB requires its phosphorylation by Polo kinase.

Alexandre Baffet from Richard Vallee's group (Columbia University, New York, NY) delivered interesting insights into molecular mechanisms of interkinetic nuclear migration (INM). INM is an oscillatory nuclear movement process that occurs in synchrony with the cell cycle in neural progenitor cells and other pseudostratified epithelia. The Vallee laboratory studies the role of cytoplasmic dynein in nuclear migration during INM by using live-cell imaging of rat brain slices. This enabled the observation that dynein recruitment to the nuclear pores is essential for the basal-to-apical nuclear movement during G2 phase, and that the disruption of this recruitment results in the complete loss of mitotic divisions in neural progenitor cells. Another interesting observation was a 3–5 µm jump of the centrosome towards the nucleus prior to mitotic entry. The exact role of this process has yet to be elucidated.

The final talk of this session was given by Renata Basto (Institut Curie, Paris, France), who investigates the consequences of centrosomal amplification on brain development. Her laboratory has created a conditional Plk4-overexpressing mouse that displays centrosome amplification in the cortex. Strikingly, Plk4 overexpression also leads to a reduced brain size and, hence, to the creation of a mouse model for microcephaly development. The neuronal progenitors have a higher frequency of multipolar mitotic spindles and lagging chromosomes, suggesting that aneuploidy is a cause of microcephaly. Brains of mice that overexpress Plk4 also present a marked increase in apoptosis, which can be rescued in a p53 mutant background. The fact that these brains rapidly degenerate led to the interesting hypothesis that aneuploidy can also be a cause of neurodegeneration.

The final session kicked off with David Pellman's (Dana Farber Cancer Institute, Boston, MA) intriguing exploration of the involvement of centrosomes in carcinogenesis. Abnormal centrosome numbers are frequently observed in cancerous cells, but the relationship between these two phenomena is unclear. Pellman's group analysed the effects of centrosome amplification on cell invasiveness, a crucial feature of late-stage carcinogenesis, using in vitro three-dimensional tissue culture models. He showed that extra centrosomes interfere with normal cell–cell adhesion and induce the formation of invasive structures. These invasive cells display an increased number of microtubules, which, in turn, leads to hyperactivation of Rac1, a key regulator of cell adhesion. Moreover, applying a Rac1 inhibitor decreased cell invasiveness, a finding that will surely influence the future of cancer therapy.

The following talk, delivered by Susan Dutcher (Washington University, St Louis, MO), was an homage to the genetic and structural tractability of the unicellular algae Chlamydomonas reinhardtii. Dutcher's group specialises in basal body formation and function, and she presented data from a recent screen that links taxol super-sensitivity to defective basal body formation; basal bodies in the taxol-sensitive mutants were analysed by electron tomography. One key finding was that the formation of triplet microtubules is essential for the recruitment of the microtubule-severing protein katanin to the basal body. Further investigation will, hopefully, shed light on the regulation of microtubule formation and architecture within centrioles and basal bodies.

One of the final speakers was Michel Bornens (Institut Curie, Paris, France), whose pioneering work on the centrosome was cited as a source of inspiration by many of the participants. As one of the founders of the field, he first put the Workshop into perspective by discussing his thoughts about centrosome biology and evolution. He argued that the innovation of the centrosome contributed much to the evolution of individuality in Metazoa, as it is instrumental in the early sequestration of the germ plasm in development. Early specification of the germ line is a crucial condition of Weismann's doctrine at the basis of the modern synthetic theory of evolution. Michel Bornens then presented data on the Cap-Gly protein Cap350 and its role in the centrosome–cortex connection through its interaction with α-catenin.

The final talk of the day, and the meeting, was delivered by Gillian Griffiths from the University of Cambridge, UK. The work she presented focused on the centrosome and its relationship with the immunological synapse in cytotoxic T lymphocytes (CTLs). She began by showing that centrosome polarisation towards the synapse is very sensitive to T cell receptor (TCR) signalling, and then discussed the striking similarities between the immunological synapse and the primary cilium. She showed that Hedgehog signalling appears to be crucial for CTL function as its inhibition reduced centrosome docking as well as CTL killing efficiency. Finally, she presented beautiful electron microscopy images revealing that the centrosomes appeared to be anchored at the immunological synapse through their distal appendages. This interesting possibility still remains to be tested.

The ‘Building a Centrosome’ Workshop ended with concluding remarks from Fanni Gergely and David Glover. The Workshop gave a very nice overview of the state of the field of centrosome biology. Although impressive progress has been made in the past few years with, for example, the improved description of the centriole structure, many open questions remain, such as the mechanism of action of Plk4 in centriole duplication, the organisation of the PCM, the relationship between centrosomes and primary microcephaly and the consequences of aneuploidy among the most prominent ones. Obtaining more structural data about the centrosomes as well as studying their function in a physiological context (brain, tumours, etc.) should lead to many important breakthroughs in the near future.