Cytokinesis requires Rab11-dependent membrane trafficking and the Rab11 effector protein, FIP3
We have shown that membrane traffic from recycling endosomes (REs) into the midbody of dividing cells is required for abscission, and that this traffic is controlled by the small GTPase Rab11, and its effector protein FIP3. FIP3-positive vesicles traffic into the furrow and midbody of dividing cells (see below). Depletion of FIP3 or expression of a mutant unable to bind Rab11 results in defective abscission, and both these genes are over-expressed in many cancers. This prompted us to propose that RE-derived vesicles are loaded onto microtubules at centrosomes, and are then trafficked into the furrow and midbody where they function to deliver membrane and cargo for abscission.

We have shown that FIP3 can bind both Rab11 and Arf6 GTPases simultaneously, and FIP3, Rab11 and Arf6 can form a ternary complex (at least in vitro¬). This prompted us to suggest that multiple interactions between Rab11 and Arf6 with the Exocyst complex may serve to anchor FIP3-containing vesicles in the midbody prior to abscission, where they function as an organisation platform for the assembly of the abscission machinery. Such observations suggest that the regulation of FIP3 interactions may play a crucial role in abscission.

FIP3 localisation is dynamically regulated during the cell cycle
Using real-time imaging we found that FIP3 exhibits profound spatial and temporal dynamics during cell division. FIP3 redistributes from diffuse cytosolic staining onto membranes at the centrosome during early anaphase before rapidly moving to the furrow at the onset of cytokinesis, associated with Rab11-positive endosomes (Figure 1). After abscission FIP3 returns to the centrosome.

Figure 1:  GFP-FIP3 dynamics in HeLa cells.
Panel A. At time 0, GFP-FIP3 is diffusely cytosolic with some punctate staining. As cells enter late anaphase (around 28 min in this example), GFP-FIP3 localises to membranes near the centrosomes then traffics into the midbody.
In panel B, a later series of images are shown to reveal the penetration of FIP3-positive membrane vesicles within the midbody in late cytokinesis.

These observations support the following model for membrane traffic during cytokinesis. Firstly, during prophase/metaphase, endosomal traffic stops. During early telophase, endosomal trafficking resumes and a re-direction of REs into the furrow occurs. Rab11 recruits FIP3 onto RE-derived vesicles at the centrosome that then traffic to the furrow, and at later stages into the midbody (shown in Figure 2). A model is shown schematically in Figure 3.

Figure 2: FIP3 at the midbody of a dividing cell.
Shown is a HeLa cell at telophase, DNA is stained blue, microtubules green and FIP3 in red. FIP3 accumulates in the intercellular bridge.

FIP3 delivers Arf6 to the furrow/midbody where it may regulate actin dynamics and act together with Rab11 to tether FIP3-containing vesicles to the Exocyst machinery, resulting in an accumulation of vesicle in the midbody prior to abscission. We postulate that the arrival of these vesicles at the midbody delivers important cargo needed for the abscission step. A key gap in our knowledge of this process is how FIP3 is regulated in space and time. This is one of the key objectives of my lab.

Figure 3. A model for the function of membrane trafficking during early telophase.
A During telophase, vesicles (circles) traffic bi-directionally into and out of the intercellular bridge, as indicated by the red arrows. Interdigitating microtubules are shown as blue lines. Recycling endosomal vesicles, identified by the presence of Rab11 and its effector, Rab11-FIP3, traffic Exocyst components into the intercellular bridge. B The establishment of the centriolin ring, and/or microtubule buckling, may then demarcate a specialised plasma membrane domain (the secondary ingression zones), to which the endosomal vesicles within the intercellular bridge are directed to fuse, thus resulting in C, the thinning of the intercellular bridge and the formation of the midbody bulge (see text for details). It is proposed that the action of the ESCRT complex at these sites then mediates abscission (see Figure 2 and text for details). Model adapted from (Schiel et al., 2011). We suggest that during the establishment of the secondary ingression zones, the distribution of tethering complexes, such as Exocyst components, between the plasma membrane and the endosomal vesicles may assist in the directionality of exocytosis and thus the specialisation of the fusion machinery at these sites.

Intracellular traffic and abscission
Intracellular membrane trafficking is essential for abscission: secretory and endosomal membrane traffic is re-directed into the intercellular bridge during telophase, and in late telophase vesicles from these pathways accumulate in the intercellular bridge (Figure 4). These vesicles are thought to fuse with each other and with the plasma membrane, thus ‘thinning’ the intercellular bridge and facilitating subsequent abscission (see also Figure 1 above).

Defining the mechanism(s) by which membrane traffic is re-directed into the intercellular bridge and the mechanism(s) involved in the tethering and fusion of these vesicles are important research questions.  We have identified an important role for the SNARE protein Syntaxin 16, and its regulator, mVps45, in cytokinesis. We propose that these proteins constitute an important control node in regulating membrane traffic into the intercellular bridge, and are studying how these proteins regulate the abscission event.

Figure 4. Endosome platforms in cytokinesis
For clarity, we show endosomal traffic arriving at the midbody and the assembly of the Cep55/Alix/ESCRT complex from opposite sides of the midbody. It is likely such mechanisms exist on both ‘sides’ of the midbody ring.

(A) During late telophase, Rab11- and Rab35-positive endosomes traffic from the centrosome into the furrow and midbody region where they accumulate via an interaction with the vesicle tethering Exocyst complex, which is localised to the midbody ring by centriolin (not shown here for clarity) and perhaps by interaction with BRUCE (see text). BRUCE is delivered to the midbody (also on membrane vesicles) where it is thought it may also function to act as a diffusion barrier, preventing vesicle movement through the furrow. Collectively, these interactions serve to anchor vesicles in the midbody (shown on the left of A). (B) BRUCE is a ubiquitin ligase, and endosomal cargo may become ubiquitinated. Cep55 also recruits Alix and Tsg101 (components of ESCRT I) into the midbody. ESCRT II and III may also become localised here, perhaps via interactions with Alix or Tsg101, or also by interaction with ubiquitinated endosomal cargo. Interestingly, ubiquitin levels in the midbody are dramatically elevated during late telophase. The events represented in panel B may occur contemporaneously with vesicle docking (shown in A). (C) ESCRT proteins may form a lattice on endosomal membranes and so induce membrane deformation. ESCRT complexes may also facilitate sorting away from the midbody localised endosomes, perhaps removing a ‘fusion brake’. More recently, the microtubule severing protein Spastin has been shown to bind to the ESCRT III complex CHMP2b (which has also been implicated in cytokinesis). These events (lattice formation and recruitment of spastin) may take place on endosomal surfaces providing a degree of spatial resolution. In response to a signal, perhaps from an adjacent signalling complex, endosomes (and similarly localised secretory vesicles) undergo compound fusion. Cep55 in the midbody ring may also function to recruit signalling proteins such as Aurora B and Plk1 into this region. This, if spatially and temporally coupled to spastin cleavage of microtubules may constitute the basic machinery of abscission.

Adapted from Gould and Lippincott-Schwartz, 2009.