Cytokinesis, the final step of cell division

Participants :

Renata Basto, William (Bill) Bement, Julie C. Canman, Jeremy Carlton, Arnaud Echard (organiser), Ulrike (Riki) Eggert, Christine Field, Daniel Gerlich, Mariagrazia Giansanti, Michael Glotzer, Gilles Hickson, Carsten Janke, Péter Lénárt, Juan Martin-Serrano, Juliette Mathieu, Thomas Müller-Reichert, Karen Oegema, Rytis Prekeris, Aurélien Roux, Anne Royou
And Héloïse Dufour (Cercle FSER) for the “Scientific meeting open to the public”

Cytokinesis, the final step of cell division
by Arnaud Echard
2 – 4 April, 2018


The seminar « Cytokinesis, the final step of cell division » took place in April 2018 at the “Domaine des Treilles” and gathered Cell Biology experts. Twenty scientists (comprising 7 women and 2 junior participants) from France, Europe and USA presented unpublished work and had fruitful discussions about key future directions of this exciting field. 

Cytokinesis is one of the most fundamental processes in cell biology and has fascinated biologists since the invention of microscopes. Generations of zoologists and cell biologists have been working on how cells physically divide into independent daughter cells in a variety of organisms, from echinoderms to salamanders. Years of experimental embryology and cell micromanipulation, intimately linked to the names of Walther Flemming and Ray Rappaport, revealed how the cytokinetic furrow is positioned and demonstrated the crucial role of microtubules in activating contraction in animal cells. Cells in culture as well as cells in tissues adopt stereotyped cell shape changes culminating with cleavage furrow ingression driven by actin and myosin II cytoskeletons. Bridge stability then depends on the septin cytoskeleton. The final cut or abscission is ultimately driven by the constriction of ESCRT-III helices on the side of the Flemming body. Failure at any of these steps produces binucleated, unstable tetraploid cells that can turn into a starting point for tumorigenesis. The fact that almost 40% of tumors arise from tetraploid cells indicates that failure in cytokinesis is a fundamental cause of tumor initiation.

Key, fundamental questions were addressed during the meeting, using complementary experimental models ranging from human cultured cells, Xenopus, Drosophila, C. elegans and echinoderms. In particular, we discussed the latest progress in furrow contraction, intercellular bridge stability and abscission. A variety of approaches, ranging from local activation of thermosensitive mutants to biochemistry, in vitro reconstitution assays, genetics and mathematical modeling have brought fascinating new data to this expanding scientific field. The first connections between cytokinesis failure, the abscission checkpoint and disease were also revealed during this meeting.


Cell division, cytokinesis, cleavage furrow, abscission, cytoskeleton, actin, septins, ESCRT


Renata Basto (Institut Curie) presented findings related with the consequences of cytokinesis failure in vivo using Drosophila melanogaster as a model system. By manipulating the cytokinesis machinery, using either mutations or in vivo RNA interference (RNAi) to inhibit the last step of cell division, her lab has found tissue-specific responses. While cells from the wing disc epithelium do not seem to tolerate cytokinesis failure, other non-epithelia tissues appear to tolerate it and become highly polyploidy accumulating large numbers of chromosomes and centrosomes. Interesting in these cells, mitotic division was very affected contributing to generate highly instable karyotypes.

Bill Bement (U. of Wisconsin) discussed the notion that during cytokinesis, the cell cortex, rather than being a passive target of spindle-based signaling, in fact can interpret and amplify signals from the spindle. This idea was based on the observation that that prior to developing cytokinetic Rho zone and cytokinetic apparatus, the cortex displays propagating, undamped waves of Rho activity and F-actin assembly. Moreover, during wave propagation, while Rho elicits F-actin assembly, F-actin subsequently inactivates Rho. Experimental and modeling results showed that waves represent excitable dynamics of a reaction-diffusion system with Rho as the activator and F-actin the inhibitor. Ect2, a protein that activates Rho and is required for cytokinesis, stimulates cortical excitability, while spindle microtubules inhibit cortical excitability, thereby confining the waves to the equatorial cortex. It was thus concluded that the cell cortex is an excitable medium and that the excitability is carried not by ions and ion channels, as in other biological examples of excitability but rather by the cytoskeleton and its regulators. Finally, based on these results, Bill concluded that cortical excitability is harnessed by the cell to direct cytokinesis.

Julie C. Canman (Columbia University) spoke about work in her lab on cytokinetic diversity and new technology to probe the mechanisms of cell division. Her lab has recently identified mechanisms that promote robust contractile ring constriction during cytokinesis within multicellular C. elegans embryos. Using fast-acting temperature sensitive (ts) cytokinesis-defective mutants, they found that asymmetrically dividing cells are protected against cytokinesis failure when the contractile ring is weakened. This cytokinetic protection is dependent on both cell intrinsic mechanisms, such as the PAR cell polarity (PARtitioning defective) proteins; as well as cell extrinsic mechanisms, requiring direct cell-cell contact with and cell signaling from other cells within the embryo. She also presented initial results based on new microscopy-based technology development of FLIRT (Fast Local Infrared Thermogenetics), which uses an infrared laser to locally heat and locally inactivate ts mutant proteins in vivo, using non-ts-mutants as controls. In future studies, the lab will use this FLIRT technology to probe the spatiotemporal regulation of cytokinesis in diverse cell types.

Jeremy Carlton (The Crick Institute) reported new findings regarding the membrane biology that occurred between chromosome segregation and cytokinesis. Focussing on a newly described role for the ESCRT-III machinery in regenerating the nuclear envelope during exit from mitosis by closing small holes of identical topology to ESCRT-III mediated membrane fission events, Jeremy reported that the conserved ESCRT-II/ESCRT-III hybrid protein, CHMP7, functioned to recruit downstream ESCRT-III components to this organelle. CHMP7 contained a membrane interaction domain which was required to anchor this protein in the ER and allow subsequent enrichment at sites of hole closure in the reforming nuclear envelope. CHMP7 achieves its location to these sites through interaction with the LEM-domain containing protein, LEM2, on the inner nuclear membrane. Thus, as well as being required for cytokinetic abscission, the ESCRT-III machinery is essential for nuclear envelope regeneration during mammalian mitotic exit.

Arnaud Echard (Institut Pasteur) presented work from his lab indicating a critical role of actin oxidation during the terminal step of cytokinesis, providing the first connection between oxidation-reduction and cell division. He showed that the Rab35 GTPase recruits and activates the enzyme MICAL1 to the intercellular bridge before abscission, which is required for local actin depolymerization and correct ESCRT-III recruitment at he abscission site. He also described another post-translational modification, SUMOylation that controls the assembly of Septin filaments in human cells and is critical for successful cytokinesis. He finally presented efforts for identifying new proteins essential for abscission using original proteomic approaches.

Riki Eggert (King’s College) discussed work in progress on creating a toolbox of small molecules that inhibit different proteins and pathways in cytokinesis. It has been challenging to study cytokinesis by traditional methods because it is a very rapid and dynamic process that occupies only a small portion of the cell cycle. New approaches are needed to overcome these barriers to deeper understanding, one of which is to develop probes that act rapidly and with high temporal control. Riki focussed on her work to understand how membranes and membrane trafficking participate in cytokinesis. Although it is known that membranes are needed to seal daughter cells after severing, very little is known about whether (and how) specific lipids are involved in cytokinesis. Massive membrane rearrangements occur during cell division, suggesting that lipids play specific roles. The Eggert group used mass spectrometry to determine if the lipidome changes in dividing cells and at a division site (the midbody) and found that only very specific lipids with specific side chains accumulate. In parallel, they systematically used RNAi to knock down lipid biosynthetic enzymes and identified enzymes required for division, which highly correlated with lipids accumulated in dividing cells. Having determined the nature of lipids involved in cell division and their biosynthetic enzymes, the next steps are to understand their functions. To further investigate the lipids’ biological roles, the Eggert group is using chemical biology and cell biology approaches.

Christine Field (Harvard Medical School) reported on using Xenopus egg extract to study spindle to cortex communication in large frog eggs. A central question in cytokinesis is how spatial information is communicated from the metaphase spindle to the cortex to position the cleavage furrow. In frog eggs, this occurs over hundreds of microns. Using confocal microscopy of fixed Xenopus zygotes after normal and polyspermic fertilization her lab discovered a central role for the planar boundary between sister microtubule asters that grow from the poles of the spindle at anaphase. Proximity to anaphase chromatin triggers assembly of CPC and Centralspindlin onto stable microtubule bundles at the aster boundary. The resulting CPC-positive state of the aster boundary propagates outwards as the asters grow towards the cortex. Her lab is investigating initiation and propagation of CPC-positive aster boundaries using an egg extract system, and also the influence of these boundaries on other components of cytoplasm. Actin and keratin networks disassemble in the vicinity of CPC-positive microtubule bundles, dependent on AurkB kinase activity, which we interpret as a mechanism for softening the cytoplasm in preparation for furrow ingression. Using active CPC immobilized on beads, and actin disassembly as a readout, Chris estimate that kinase activity extends ~20 microns as a reaction-diffusion gradient. She hypothesize this kinase activity gradient plays a key role in propagating CPC recruitment outwards from chromatin, towards the cortex, via CPC auto-activation on microtubule bundles.

Daniel Gerlich (IMBA) presented recent insights into the interaction between membranes and chromosomes during cell division. Cells of higher eukaryotes completely disassemble the cell nucleus while they divide, and they reassemble the nucleus during mitotic exit. This requires a coordinated interaction between membranes and each set of segregated chromosomes to warrant that the entire genome is packaged into a single nucleus. If by mistake single chromosomes are packaged into small nuclei, they are prone to DNA damage, which imposes a severe risk to develop cancer phenotypes. Daniel Gerlich’s laboratory identified a protein that builds a transient network around each set of segregated anaphase chromosome and thereby guides membranes to form a single nucleus. The study reveals a mechanism by which cells form a single nucleus, a hallmark of almost all higher eukaryotic cells.

Maria Grazia Giansanti (Roma U.) discussed the role of Golgi phosphoprotein 3 (GOLPH3), in cytokinesis. GOLPH3, has been characterized as a Phosphatidylinositol 4-Phosphate [PI(4)P] effector at the Golgi, involved in vesicle trafficking and Golgi glycosylation. GOLPH3 is frequently amplified in several solid tumor types including breast cancer, melanoma and lung cancer. Moreover GOLPH3 overexpression is correlated with poor prognosis in multiple cancer types. The group of Dr. Giansanti has shown that GOLPH3 accumulates at the cleavage furrow and is essential for cytokinesis in Drosophila melanogaster. During cytokinesis GOLPH3 is essential for maintenance of centralspindlin and Rho1 at cell equator and stabilization of Myosin II and Septin rings. GOLPH3 function during cytokinesis is dependent on its binding to PI(4)P: mutations that affect interaction with PI(4)P, disrupt localization of GOLPH3 at both the Golgi stacks and the cleavage furrow. Telophase spermatocytes from mutants with defective GOLPH3-PI(4)P interaction, also fail to accumulate PI(4)P-and Rab11-associated secretory organelles at the cleavage site. They hypothesize that GOLPH3 might act as a key molecule in coupling vesicle trafficking with actomyosin ring assembly and stability during cytokinesis. Amongst the putative molecular interactors of GOLPH3, they have identified the Drosophila ortholog of human COG7 (Cog7), a subunit of the Conserved Oligomeric Golgi (COG) complex, which functions as a vesicle-tethering factor for intra-Golgi retrograde trafficking, playing a crucial role in maintaining the glycosylation enzymes across the Golgi cisternae. These data suggest that GOLPH3 protein might bind Cog7 and cooperates with the COG complex to regulate Golgi trafficking and glycosylation.

Mickael Glotzer (U. of Chicago) and his lab has developed a method for optogenetics which allows high resolution spatiotemporal control of protein localization in live cells using patterned light. Using this approach, they have demonstrated that a limited sized zone of active RhoA is sufficient to induce cleavage furrow formation. Strikingly, such zones of active RhoA can induce furrow formation irrespective of cell cycle stage or position of the cell cycle. However, adherent interphase cells with abundant stress fibers do not form furrows, but simply detaching these cells from the substrate, allowing them to round up allows them to furrow robustly in response to local RhoA activation.The second main topic concerned the mechanism by which cells natively generate zones of local RhoA activation. Mickael presented evidence from a variety of experimental systems that support the notion that a membrane bound pool of centralspindlin-ECT-2 is responsible for RhoA activation, despite the fact that this pool is not always readily detected. In particular, he discussed evidence indicating that Aurora B-induced centralspindlin oligomerization, is an important contributor to the membrane-bound pool of centralspindlin. Lastly, he discussed unpublished efforts to understand how astral microtubules negatively regulate contractility by inducing displacement of key activators of RhoA.

Gilles Hickson (St Justine Hospital) discussed efforts to understand how the contractile ring, a dynamic actomyosin-dependent structure, transitions into a stable midbody ring at the close of the furrowing phase of cytokinesis. Structure-function analyses of the multi-domain scaffold protein Anillin in Drosophila tissue culture cells have uncovered differential requirements for Anillin and its interacting partners in both the formation of the midbody ring and its anchoring to the plasma membrane. Furthermore, Anillin- and septin-dependent plasma membrane extrusion and shedding accompanies thinning of the late contractile ring/ nascent midbody ring. Synthesis of these observations led to the proposal that closure of the ring involves the controlled release of membrane-anchored Anillin-septin elements from the contracting and disassembling actomyosin elements within the ring. Counter-acting mechanisms that limit the release of Anillin-septins from the nascent midbody ring were also discussed; these ensure that the release is controlled and coordinated with ring maturation.

Carsten Janke (Institut Curie) discussed the role of tubulin diversity in microtubule functions. He showed unpublished results of his team that demonstrate how single-amino acid variations in tubulin molecules can actually affect the dynamic instability of microtubules. This has important implications for microtubule functions, and was so far overlooked because no appropriate approaches were available. Based on these results, he also discussed the evolutionary diversity of tubulin genes, and showed that for instance C. elegans, a model organism frequently used to study cell division, cytokinesis and abscission, has a much larger diversity among its tubulin genes as compared to mammals. This implies a potentially more important impact of tubulin diversity on cellular processes in C. elegans as compared to mammalian cells, and should be considered in interpreting data from this model organism. Carsten further discussed how posttranslational modifications of tubulin could interfere with the cell division process, and in particular with abscission. His team has preliminary data that show a role for polyglutamylation in abscission timing, but they have not yet been confirmed. The meeting allowed him to make new collaborative contacts to further pursue his research on these questions.

Peter Lenart (EMBL) and his laboratory at the European Molecular Biology Laboratory, Heidelberg, Germany study the specialized oocyte divisions that produce the fertilizable egg. Their main interest is to understand how the cell division machinery adapted to this specialized function by dividing the very large egg cell in a very asymmetric manner, and thereby retaining all the nutrients stored for the embryo in a single large egg. Peter showed several examples of such adaptations of the underlying molecular mechanisms including a novel, actin-driven mechanism required to break the exceptionally large oocyte nucleus. This involves a very transient (1-2 minutes) assembly of a massive actin ‘shell’ underlying the nuclear envelope, driven by the Arp2/3 nucleator complex. This actin shell then forces apart the nuclear membranes and the underlying lamina, a network of intermediate filaments that is normally required to stabilize the nuclear envelope. As another example, he showed how the capture of chromosomes scattered in the very large oocyte nucleus is coordinated in order to incorporate each and every one of them into the forming meiotic spindle. Surprisingly, this is again mediated by the actin cytoskeleton, which transports chromosomes near to microtubules and at the same time coordinates the timing of capture.

Juan Martin-Serrano (King’s College) discussed the molecular mechanisms that regulate cytokinetic abscission, the last step in cell division that facilitates the physical separation of the daughter cells. This discussion was focussed in the regulatory pathway known as the abscission checkpoint (also known as NoCut) that delays abscission until anaphase chromatin bridges are cleared from the midbody, the intercellular bridge that connects the dividing cells. It was also discussed how the abscission checkpoint is engaged in response to defective nucleopore assembly and high levels of midbody tension, and how this process is regulated by ULK3, a poorly characterized kinase. Critically, ULK3 phosphorylation of ESCRT-III is essential to sustain the abscission checkpoint in multiple physiological contexts. Despite this progress in the identification of the events involved in abscission regulation, how the genome is protected by this checkpoint remains less clear. Unpublished data was discussed to answer this question, showing that CHMP4C-deficient cells accumulate DNA damage, and these cells are genetically unstable under experimental conditions that increase replication stress and chromosome mis-segregation. This conclusion was strengthened by the structural and functional characterization of a cancer-associated polymorphism in CHMP4C, which disrupts the abscission checkpoint. Based on these observations, a novel oncogenic mechanism that involves the dysregulation of abscission was proposed.

Juliette Mathieu (Collège de France) reported unpublished work regarding the involvement of the deubiquitinating enzyme Ubpy during the regulation of cytokinesis in the drosophila female germline. Although drosophila germline stem cells perform complete abscission, their differentiating progeny divide 4 times synchronously with incomplete cytokinesis and form cysts of 16 cells interconnected by ring canals. How this conserved process is regulated is poorly understood so far. The analysis of Ubpy loss of function showed that this enzyme is essential to prevent complete cytokinesis in the germline cyst. ESCRT proteins, known to physically interact with Ubpy and to promote abscission in many cell types, become up-regulated at the ring canals of Ubpy mutant cysts. Moreover, reducing their levels is sufficient to rescue the ectopic abscission observed in the Ubpy mutant cysts. Ongoing experiments aim to establish if one or several ESCRT could be direct target of Ubpy during the regulation of abscission in the germline cysts.

Thomas Müller-Reichert (TU Dresden) reported on the ultrastructure of the abscission machinery in mammalian (Hela) cells and nematode early embryos of C. elegans. Abscission is the final step of cytokinesis, leading to a physical separation of post-mitotic daughter cells. Using these two model systems, the Müller-Reichert group has developed correlative microscopic approaches by combining light microscopy and serial-section electron tomography to study this last step of cell-pair separation. Similarities and differences in the three-dimensional ultrastructure of intercellular bridges in mammalian cells versus nematode embryos were shown and discussed. Particularly, the ultrastructure of ESCRT-dependent filaments as observed in cellular systems was compared to structural data obtained from in vitro experiments. Finally, Thomas Müller-Reichert reported on his vision to compare the ultrastructure of intercellular bridges in various animal phyla and analyze cytokinetic events within various tissues.

Rytis Prekeris (U of Colorado Denver) reported new unpublished data about the role of polyglutamylation of microtubules in regulating the recruitment of Spastin and the establishment of the abscission site during late telophase. While the requirement of Spastin-dependent microtubule severing and depolymerization during mitosis is now well-established, how Spastin is recruited to the abscission site remained unknown. This study suggests that TTLL12, a microtubule modifying enzyme of previously unknown function, regulates tubulin polyglutamilation at the central spindle, thus regulating the recruitment of Spastin and microtubule severing. In addition to defining a new mechanisms regulating the establishment of the abscission site, Rytis Prekeris also presented new unpublished work on the roles of post-mitotic midbodies in regulating proliferation and differentiation. Recently it has been proposed that post-mitotic midbodies may function as polarity cues as well as signaling platforms that regulate cell fate and stemness. However, most studies so far have been correlational and it remained to be determined how and if post-mitotic midbodies actually regulate cell functions. The new data that has been shown in the meeting demonstrate that cells can readily internalize post-mitotic midbodies. Importantly, these midbodies are retained within the cell as a novel signaling structure, called MBsome. Finally, these MBsomes appear to signal via integrin and EGF receptor dependent pathways to regulate the cell differentiation and proliferation, thus defining and entirely new mode of intracellular signaling.

Anne Royou (IBGC CNRS) presented her work on the mechanisms that coordinate chromosome segregation with cell cleavage. Using methods that increase the length of sister chromatid arms without affecting spindle integrity, and thus chromosome segregation in Drosophila neuroblasts, she reported a mechanism by which cells undergo myosin-mediated adaptive elongation to facilitate the clearance of trailing chromatid arms from the midzone before completion of cytokinesis She described how cells undergo two phases of adaptive elongation. The first phase relies on the assembly of a wide contractile ring at the onset of cytokinesis. The second phase requires outward flux of myosin from the ring toward the polar cortex during ring constriction. She detailed how, in normal cells, a pool of myosin flows from the ring and invades transiently the whole cortex of both nascent daughter cells. Myosin then dissociates from the cortex in concert with nuclear envelope reassembly and the ensuing nuclear sequestration of Rho-GEF. This myosin efflux is a novel feature of cytokinesis as it occurs systematically during each cell division and in all cell type examined. She, then, described, how the presence of trailing chromatids at the midzone induces a delay in nuclear envelope reassembly concomitant with an extended period of active cortical myosin, thus, providing forces for the second adaptive elongation. Attenuation of Rho-GEF prevents myosin efflux and adaptive elongation. Conversely, the retention of Rho-GEF in the cytoplasm is sufficient to prolong cortical myosin activity and to trigger both phases of adaptive elongation without trailing chromatids at the midzone. She also elaborated on the role of Anillin and the chromatin-associated Ran-GEF, RCC1, on the adaptive elongation mechanism.

Aurelien Roux (U. de Genève) reported new data showing how ESCRT filaments are remodelled at the site of abscission by Vps4 in order to constrict the membrane. Moreover, he presented unpublished data showing how ESCRT filaments made of Snf7/Vps2/Vps24 are able to assemble into various structures: spiralling helices or ribbons made of bundled filaments, or spiralling membrane tubes coated with bundles of filaments. These structures showed how the large spirals observed in vivo at the abscission site can be formed by assembly of various subunits of ESCRT-III. Furthermore, he showed how membrane tension at the level of endosomes can regulate the assembly of ESCRT-III structures. Using a newly developed membrane tension probe called FliptR (for Fluorescent Lipid Tension Reporter), the membrane tension in endosomes was directly visualised and decreased upon hypertonic shocks. This correlate with high recruitment of ESCRT-III markers, and was reconstituted in vitro using purified Chmp4b and liposomes, subjected to hypertonic shocks.

The beautiful location of Les Treilles and the amazing scientific atmosphere were much appreciated by all participants. It was also a pleasure to be part of the DECLIC program again, organized by Héloïse Dufour of the Cercle FSER (Fondation Schlumberger for Education and Research), which allowed us to meet scientific high school students and Biology teachers from Draguignan and share our passion for science with them.

Renata Basto Bill Bement Julie Canman Jeremy Carlton Arnaud Echard Riki Eggert Christine Field Daniel Gerlich Mariagrazia Giansanti Michael Glotzer Gilles Hickson Carsten Janke Peter Lénárt Juan Martin-Serrano Juliette Mathieu Thomas Müller-Reichert Karen Oegema Rytis Prekeris Aurélien Roux Anne Royou Cytocynèse, cytokinesis - Fondation des Treilles
This entry was posted in Activities, Reviews and tagged , , , , , , , . Bookmark the permalink.