|Home | About | Journals | Submit | Contact Us | Français|
Cell polarity is important for a number of processes, from chemotaxis to embryogenesis. Recent studies suggest a new role for polarity in the orchestration of events during the final cell separation step of cell division called abscission. Abscission shares several features with cell polarization, including rearrangement of phosphatidylinositols, reorganization of microtubules, and trafficking of exocyst-associated membranes. Here we focus on how the canonical pathways for cell polarization and cell migration may play a role in spatiotemporal membrane trafficking events required for the final stages of cytokinesis.
Cell polarity is a common feature of eukaryotic cells. A polarized cell has a single axis of asymmetry, which can be thought of as “apical” and “basal” (illustrated in Figure 1A) or “front” and “rear” (Figure 1B). This asymmetry encompasses a variety of cellular morphologies that can differ among cell types in a single organism, as well as within a single-celled organism such as Saccharomyces cerevisiae. Cells within multicellular organisms also exhibit a distinct apical–basal axis of polarity. One example is epithelial initiation, which occurs during development and involves congression of mesenchymal cells into aggregates that differentiate to form polarized apical–basal cell monolayers. At the migratory stage, mesenchymal cells set up a front–rear axis of polarity (Figure 1B). These polarity principles contribute to the higher-order organization of cell-based systems, such as shaping the embryo, wiring the developing nervous system, maintaining and regenerating tissue, and developing the immune response (reviewed in Nelson, 2009 ). Perturbations in cell polarity contribute to a number of tissue pathologies. Particularly striking examples include neuronal migration disorders (NMDs) such as schizencephaly, porencephaly, and lissencephaly, which are caused by defects in the polarized migration of neurons (for a review on NMDs, see Valiente and Marin, 2010 ).
Specific sorting and maintenance of proteins to distinct membrane domains ensures polarity formation. The defined distribution of plasma membrane and cytoskeletal proteins in a polarized cell requires signaling networks and protein complexes comprising the Rho and Rab family GTPases, their downstream effectors, and polarity complexes (Crumbs, PAR, and Scribble). Effectors include cytoskeletal proteins and vesicle-trafficking pathways. Vesicle trafficking from the endocytic and/or exocytic membrane compartments along a polarized cytoskeleton network is required for the establishment of cellular polarity (reviewed in Nelson, 2009 ). This has been elegantly illustrated in the budding yeast, S. cerevisiae, in which bud growth is ensured by polarized secretion of Golgi apparatus–derived membrane vesicles to cortical actin patches under the regulation of Rho GTPases and polarity complexes at the bud tip (reviewed in Chant, 1999 ; Irazoqui and Lew, 2004 ). In multicellular eukaryotes, protein-sorting events occur at the endocytic pathway and Golgi apparatus to ensure cell polarization (Nelson, 2009 ). For example, plasma membrane–endocytic recycling is critical for maintaining polarized membrane protein residency to establish appropriate responses to stimuli such as nutrient internalization, junctional protein sorting (e.g., E-cadherin), and ion channel recycling (Lock and Stow, 2005 ; Ducharme et al., 2006 ). Of interest, recent evidence shows that vesicle trafficking is also required for the establishment of polarized domains during cytokinesis, the final stage of cell division (Figure 1C).
Recent studies suggest that principles of cell polarity are engaged during the process of cytokinesis. For instance, a migrating polarized cell requires constant membrane addition via secretion at the leading edge to maintain “front–rear” polarity (Nelson, 2009 ), just as abscission (the final stage of cytokinesis) requires membrane addition at the cytokinetic bridge via secretion and endocytic vesicle delivery (further discussed in the review by Prekeris and Gould, 2008 , and depicted in Figure 2). In both cases, membrane vesicles may act as a platform for delivering essential regulators ensuring cell polarization. Another similarity between cell polarity and cytokinesis occurs within S. cerevisiae. In G1/S an actin patch is focused at the bud tip where secretory vesicles are directed, and then the patch dissipates as the bud becomes larger and the cell enters cytokinesis. Prior to spindle disassembly and cell separation, polarized-actin patch proteins and secretory vesicles are redirected to the mother bud neck, a structure analogous to the cytokinetic bridge (VerPlank and Li, 2005 ). The apparent importance of polarized vesicle trafficking to polarity and the final stages of cytokinesis leads one to speculate about a conserved underlying mechanism between the two processes. This notion will be discussed and proposed throughout this essay.
Asymmetry in an epithelial, neuronal, or migrating cell is reflected in its structural, molecular, and functional polarity. For example, in a migrating cell the broad leading edge (front) of the cell defines the direction of movement, and the more-focused rear trails behind. When a migrating cell divides, it first rounds up, thus eliminating its preexisting polarity. This involves vesiculating and dispersing sorting compartments such as the Golgi apparatus and the endosomal system. During cytokinesis, furrow ingression gives rise to the intercellular bridge, a thin cytoplasmic connection between the two nascent daughter cells that is later resolved in the process of abscission. Each nascent daughter assumes an intracellular organization roughly similar to an interphase cell, where the Golgi apparatus and endocytic system have partially reorganized over the mother and daughter centrosomes (Figure 2). At this time, cytokinetic cells have started to reestablish front–rear polarity such that they need only to sever the bond between them to gain the freedom to migrate.
The morphological front–rear polarity observed between migration and cytokinesis is further defined by molecular reorganization—specifically, changes in phosphotidylinositol concentration within the plasma membrane (Figure 1). However, little is known about how lipid domain polarization occurs during polarization or cell division. Possibilities include 1) remodeling of existing lipid domains by membrane traffic (endocytosis, directed secretion); 2) cortical flow of specific phosphatidylinositols to and immobilization at a specific region; and 3) targeting of lipid-modifying enzymes to a specific region. Cytokinetic proteomic and RNA interference screens have led to the identification of intracellular transport genes required for cytokinesis (Echard et al., 2004 ; Skop et al., 2004 ) and subsequently for lipid domain polarization. These genes included Rab GTPase family members and inositol-modifying enzymes, such as phosphatidylinositol 3-kinases (PI3Ks), which are Rab effectors (e.g., Rab5; Vieira et al., 2003 ) that connect membrane traffic with lipid modification.
Cytokinesis is accompanied by polarized organization of phosphatidylinositols. For instance, phosphatidylinositol (3,4,5)-triphosphate (PI(3,4,5)P3) localizes to the plasma membrane near the spindle poles. On the other hand, phosphatidylinositol (4,5)-bisphosphate (PI(4,5)P2) predominantly localizes to the cleavage furrow and then to the intracellular bridge between mother and daughter cells analogous to the rear of the migrating cell (Figure 1C; Janetopoulos et al., 2005 ; Toyoshima et al., 2007 ). In a migrating cell, polarity is established by local PI(3,4,5)P3 accumulation at the cell's leading edge (Figure 1B; Parent and Devreotes, 1999 ), which is achieved through localization of PI3 kinases and the tumor suppressor PTEN (Kolsch et al., 2008 ). This reorganization facilitates pseudopodia extension (Wang et al., 2002 ) through actin polymerization induced by actin-binding proteins (WASP, profilin, cofilin, capping protein) binding to membranous PI(3,4,5)P3 and PI(4,5)P2 (Yin and Janmey, 2003 ) and subsequent membrane addition by vesicle transport. Mislocalization of phosphatidylinositols disorganizes polarity of the aforementioned actin-binding proteins (reviewed in Gassama-Diagne and Payrastre, 2009 ; Nelson, 2009 ) and also inhibits cytokinesis by a similar mechanism (Janetopoulos et al., 2005 ). For example, cells lacking PTEN and PI3K are unable to create a stable actin/myosin–based contractile ring, thus failing to form a cleavage furrow between dividing mother and daughter cells (Janetopoulos et al., 2005 ). This finding suggests that polarization of phosphatidylinositols at the midzone is required to initiate cytokinesis.
Secretion from the Golgi apparatus has been suggested as the primary trafficking route for both cell polarization and cytokinesis (Gromley et al., 2005 ; Nelson, 2009 ). However, recent evidence links the secretory and endocytic recycling pathways, making it difficult to dissect their individual functions. Originally, the exocyst vesicle-tethering complex was found at the apex of the lateral membrane domain in polarized epithelial cells (Figure 1A), where it specified basolateral secretory-vesicle delivery (Grindstaff et al., 1998 ). More recently, the exocyst was also found at the apical recycling endosome compartment (Oztan et al., 2007 ; Bryant et al., 2010 ) and at the basal bodies of the primary cilium (Babbey et al., 2010 ; Figure 1A). In addition, the exocyst subunit Sec15 was identified as a Rab GTPase-Rab11 effector that colocalizes at the recycling endosome in both Drosophila and mammals (Zhang et al., 2004 ; Wu et al., 2005 ). Thus the exocyst has distinct localizations that include the plasma membrane, secretory vesicles, the basal body, and recycling endosomes, which suggests not only roles in secretory vesicle fusion at the plasma membrane and during endocytic recycling, but perhaps various other functions.
The exocyst may be involved in regulating polarized-vectorial membrane traffic toward the leading edge of a migrating cell during interphase (Figure 1B) or in the opposite direction toward the midbody during cytokinesis (Figure 1C). A migratory cell requires de novo plasma membrane addition at the leading edge that is driven by membrane traffic (Lim et al., 2005 ). This occurs through use of the small-GTPase RalB to confine vesicle trafficking to the leading edge, achieving directional cell movement (Camonis and White, 2005 ; Rosse et al., 2006 ). These vesicles then tether at the leading edge via the RalB-recruited-exocyst complex (Rosse et al., 2006 ), suggesting that localized membrane addition is required to create a polarized migrating cell. Of interest, this same process seems to be similar, if not the same, for membrane addition at the cytokinetic bridge. As in cell migration, RalB recruits the exocyst to the cytokinetic bridge during abscission (Cascone et al., 2008 ). In addition, two independent groups showed that secretory vesicles dock via the exocyst and fuse within the cytokinetic bridge (Gromley et al., 2005 ; Goss and Toomre, 2008 ). Most important, these studies found that vesicle fusion near the midbody ring is required for abscission. One interesting point that these findings highlight is that vectorial membrane transport toward the midbody is opposite that of membrane transport in a migrating cell, suggesting that cytokinetic cells can reorganize the direction of polarized membrane traffic.
During both cell migration and cytokinesis, plasma membrane addition driven by the exocyst complex is likely coordinated with Rab-dependent and soluble N-ethylmaleimide–sensitive fusion protein attachment protein receptor (SNARE)–dependent machinery. One of the best examples of this possible coordination is with the mother-centriole-localized protein centriolin. During cytokinesis centriolin localizes to the midbody ring within the cytokinetic bridge (Figure 2). At the bridge, centriolin can interact with both the exocyst complex and a SNARE-associated protein, Snapin, which can mediate secretory vesicle fusion (Gromley et al., 2005 ). In a separate study, exocyst depletion not only caused inhibition of secretory vesicle fusion at the midbody, but also impaired Rab11 localization to the cytokinetic bridge (Fielding et al., 2005 ). These results suggest that the exocyst is required for both endosomal and secretory membrane-trafficking steps during cytokinesis (Fielding et al., 2005 ; Gromley et al., 2005 ), which is similar to the requirement of the exocyst complex for endocytic recycling and secretory traffic to the leading edge in polarized cells (Grindstaff et al., 1998 ; Zhang et al., 2004 ; Wu et al., 2005 ; Rosse et al., 2006 ; Oztan et al., 2007 ; Nelson, 2009 ).
Rab GTPases are key regulators of membrane traffic, with more than 60 members in mammalian cells that define particular routes within the secretory and endocytic pathways. Rab11 is a well-established participant in recycling endosomal trafficking. In polarized epithelial cells Rab11 is associated with membranes in the apical portion near the centrosome and beneath the apical plasma membrane (Casanova et al., 1999 ). This localization enables Rab11 to regulate efficient endosomal recycling to apical plasma membrane domains (Prekeris et al., 2000 ). Rab11 is required for cytokinesis in Caenorhabditis elegans and Drosophila and was the first Rab GTPase implicated in cytokinesis in human cells (reviewed in Strickland and Burgess, 2004 ).
Drosophila embryogenesis has provided a unique system for studying Rab11-dependent membrane trafficking (discussed in Strickland and Burgess, 2004 ). Molecular elements required for cellularization of the Drosophila embryo are often homologous to those that drive cytokinesis in mammalian cells. Of interest, Rab11 activity is essential for cellularization of the Drosophila embryo. Dominant-negative Rab11 caused defects in membrane addition and furrow morphology (Ullrich et al., 1996 ). These phenotypes were very similar to Drosophila Nuclear Fallout mutants (Rothwell et al., 1998 ), suggesting that both genes may be involved in common pathways. In fact, Nuclear Fallout is closely related to the mammalian Rab11-binding protein FIP3 (Hickson et al., 2003 ), which is required for mammalian cytokinesis (Fielding et al., 2005 ; Wilson et al., 2005 ).
The formation and targeting of Rab11 vesicles to membrane domains within the cytokinetic bridge may be analogous to the mechanism for targeting Rab11 vesicles to the apical plasma membrane. In fact, Rab11-positive recycling endosomes, together with the Rab11 effectors FIP3 and FIP4, are required for membrane delivery during cytokinesis and abscission, although the site of targeting is unknown (Wilson et al., 2005 ). Both FIP3 and FIP4 bind the small GTP-binding protein ARF6, which binds to the exocyst subunit Exo70 and is required for cytokinesis (Fielding et al., 2005 ). One proposed role for these complex interactions is that FIP3 and FIP4 couple Rab11-positive vesicle traffic from the recycling endosome to the midbody, where they are tethered via interactions with ARF6 and the Exocyst (Fielding et al., 2005 ). Of interest, polarized migrating cells rely on ARF6-regulated endosomal trafficking to concentrate active Cdc42 at the leading edge and recruit the Par6-aPKC polarity complex (Osmani et al., 2010 ).
Rab35 is a newly discovered GTPase required for cytokinesis. Like Rab11, Rab35 localizes to the endocytic recycling pathway and regulates cytokinesis. However, during interphase Rab35 does not colocalize exactly with the Rab11 compartment. Rab35 functions at an early (fast recycling) endosome, prior to the relatively slow recycling endosome step regulated by Rab11 (Kouranti et al., 2006 ). Whether a role exists for Rab35 in regulating either polarity in migratory cells or epithelial cells is not yet known. One argument that it may regulate polarity is that Rab35 is required during cytokinesis to concentrate PI(4,5)P2 at the cytokinetic bridge (Figure 1C; Kouranti et al., 2006 ). In addition, like Rab11, Rab35 is required during cytokinesis after furrow ingression to provide polarized delivery of membrane vesicles derived from recycling endosomes to the cytokinetic bridge. This membrane could be required during abscission for nascent daughter cell separation (Wilson et al., 2005 ). Because Rab11 and Rab35 are localized to different subcellular compartments and control distinct endocytic recycling pathways (Zerial and McBride, 2001 ), it is likely that there are multiple endocytic routes that are individually essential for cytokinesis and, hypothetically, polarity.
Crumbs proteins are important determinants of apical membrane identity (McCaffrey and Macara, 2009 ) and may be required for cytokinesis. A recent study found that the apical membrane is established during and after cytokinesis through the delivery of Crumbs3-positive membranes from a Rab11-regulated recycling endosome compartment at the spindle poles that move centripetally along microtubules toward the plasma membrane (Schluter et al., 2009 ). Live-cell imaging revealed Crumbs3 localization between the two daughter cells within the cytokinetic bridge (Schluter et al., 2009 ). Although these data strongly suggest that polarity is initiated during cytokinesis, a requirement for polarity complexes in cytokinesis has yet to be determined. Of interest, the studies of Schluter et al. (2009 ) and others (reviewed in Shivas et al., 2010 ) have shown that the endocytic pathway can play an important role in the localization of polarity proteins during cytokinesis and polarity formation; there is also evidence for a reciprocal role for polarity proteins in the regulation of endocytic machinery (Balklava et al., 2007 ). Taken together, these studies suggest a synergistic model in which polarity and endocytosis regulators may cross-regulate one another and mutually contribute to the completion of cytokinesis and the formation and function of polarized cells.
Establishment of polarity is important for cell migration and partitioning an organism into external and internal compartments. The discussion here suggests that polarity also functions in cell division. This is suggested by the organization of polarity complexes to cytokinetic sites and defects in cell division following loss or mutation of polarity complex proteins (Albertson and Doe, 2003 ; Humbert et al., 2006 ). Similarly, although the localization of polarity components such as Crumbs and phosphatidylinositols to mitotic structures, as well as directed vesicular traffic, highlight the initial stages of polarity during the final stages of cytokinesis, the functional relevance of this reorganization is unknown. Therefore comparing components that are involved in cytokinesis, such as Rab35, that are not yet known to be involved in cell polarity and investigating their possible conserved function between cytokinesis and polarity may shed light on novel molecular mechanisms. The same approach could be used in a reciprocal manner to address whether known polarity proteins not yet shown to be involved in cytokinesis have cytokinetic functions.
We conclude that there is a likely requirement for spatiotemporal orchestration of membrane trafficking and polarity-complex machinery to construct new polarized membranes during both cytokinesis and apical surface formation. However, the precise spatiotemporal relationships and functional relevance of these factors have yet to be determined. Future insights into how endocytic recycling, secretory, and polarity pathways act together during cytokinesis will require careful spatiotemporal analysis of these components at the cytokinetic bridge.
We thank David Lambright (University of Massachusetts Medical School), Charles Yeaman (University of Iowa, Iowa City, IA), Alison Bright (Doxsey lab, University of Massachusetts Medical School), and Benedicte Delaval (Doxsey lab, University of Massachusetts Medical School) for critical reading of the manuscript. The National Institutes of Health, the Ellison Medical Foundation, and the W. M. Keck Foundation supported the work in S.D.'s laboratory. H.H. is supported by a National Research Service Awards Postdoctoral Fellowship.