“Life is a Highway”: Membrane Trafficking During Cytokinesis

Abstract

Cytokinesis, the final stage of the cell cycle, is an essential step toward the formation of two viable daughter cells. In recent years, membrane trafficking has been shown to be important for the completion of cytokinesis. Vesicles originating from both the endocytic and secretory pathways are known to be shuttled to the plasma membrane of the ingressing cleavage furrow, delivering membrane and proteins to this dynamic region. Advances in cell imaging have led to exciting new discoveries regarding vesicle movement in living cells. Recent work has revealed a significant role for membrane trafficking, as controlled by regulatory proteins, during cytokinesis in animal cells. The endocytic and secretory pathways as well as motor proteins are revealed to be essential in the delivery of vesicles to the cleavage furrow during cytokinesis.

Introduction

Cytokinesis is the final step of the cell cycle resulting in the formation of two daughter cells. Successful completion of cytokinesis is dependent on a number of processes including separation of the sister chromatids, formation and constriction of the actomyosin contractile ring, and abscission of the plasma membrane. Over the past two decades, membrane trafficking has also been discovered to play an essential role in the completion of cytokinesis. The goal of this review is to provide a general overview of the roles of membrane trafficking, and key regulatory proteins, in this last event of animal cell cytokinesis. The final act in the separation of the two daughter cells, abscission, is well addressed in a new comprehensive review (1).

Vesicle-mediated membrane trafficking to the plasma membrane occurs via two main pathways: the recycling endosome pathway and the secretory pathway (reviewed in detail elsewhere (2,3). The endocytic pathway begins with vesicle budding at the plasma membrane. Endocytosed vesicles are then transported to the early endosome where they are returned to the plasma membrane either by way of the recycling endosome (RE), or brought to the lysosome for degradation by the late endosome. The secretory pathway, on the other hand, begins at the endoplasmic reticulum (ER), where vesicles are transported to the Golgi apparatus for sorting before being shuttled to the plasma membrane. Membrane trafficking by way of both of these pathways has been shown to be pivotal in cytokinesis.

It has long been known that membrane trafficking plays a dominant role in cell division of plant cells. During division, these cells form a structure termed the phragmoplast at the division plane. The phragmoplast is made up of actin and microtubules, and forms a framework for vesicle delivery to the site of cell division. The vesicles then fuse at the center of the two daughter cells, forming a cell plate, which then grows outwards along the division plane until final fusion events at the plasma membrane complete cytokinesis (for a comprehensive Review see (4)).

The role of vesicle trafficking during cytokinesis in animal cells is not as well understood. Current research has suggested that, as in plants, both the endocytic and secretory pathways play an essential role in the successful completion of cytokinesis. A comprehensive review of this subject was presented in 2005 (5). Here we present an updated look at recent work in the field of membrane trafficking during cytokinesis and discuss questions remaining in this field.

The Scenic Route: Recycling Endosome Pathway

In 2004, Skop, et al.(6) performed a comprehensive proteomics-based analysis of proteins involved in cytokinesis. In order to investigate the final stage of mitosis, the researchers isolated midbodies from Chinese hamster ovary (CHO) cells, and analyzed the associated proteins by tandem liquid chromatography and tandem mass spectrometry. Interestingly, the results showed that the largest category of protein identified with the midbody were those involved in membrane trafficking or secretion; of the 160 proteins deemed as candidates for involvement in cytokinesis, 33% were of this category (6).

Among the proteins investigated, RAB GDI was found to localize to the midbody in HeLa cells using immunofluorescence. RAB GDI is a member of the GDP-dissociation inhibitor family of proteins that are involved in the maintenance of RAB in its inactive, GDP-bound state. This function leads to the recycling of RAB proteins from a membrane-bound state to the cytosol (7). RNAi experiments in C. elegans show that loss-of-function of RAB GDI leads to defects in both the early and late stages of cytokinesis. Loss of function in germline cytokinesis also leads to sterility and embryonic lethality (6). Studies on the localization and function of RAB GDI have illuminated a role for the involvement of recycling Rab proteins in the completion of cytokinesis.

Another protein identified in the proteomic screen was RACK-1, a widely conserved receptor for activated protein kinase C (8). Loss-of-function studies using RNAi for the C. elegans homologue of RACK-1 showed defects in the late stages of cytokinesis, embryonic lethality, and sterility (6). During further investigation of the role RACK-1 plays in cytokinesis, Ai, et al. (2009) found that defects seen in the membranes of rack-1(RNAi) worms were similar to the phenotype seen in mutant rab11, a small GTPase, and Nuclear-fallout (nuf), a homologue of the mammalian FIP3 Rab11 effector, in Drosophila embryos (9). Rab11 is a small GTPase that colocalizes with transferrin and is used as a marker of the recycling endosome (10). Following this finding, the number of Rab11 vesicles was found to be significantly decreased in rack-1(RNAi) C. elegans embryos (11). In addition, through the use of a yeast two-hybrid screen, it was discovered that RACK-1 interacts with the p50/dynamitin subunit of dynactin, and further analysis revealed that RACK-1 is necessary for proper localization of DNC-2 (Dynactin Complex component) (11). Collectively, this evidence supports a model where RACK-1 and its interaction with dynactin play a central role in the maintenance of recycling endosomes at the centrosomes. Upon disruption of RACK-1, recycling endosomes are not properly delivered to or retained at the cleavage furrow, leading to disruption in furrow ingressions and a failure of cytokinesis (11).

Studies have shown that Rab11 is necessary for proper furrow ingression during Drosophila cellularization and in the C. elegans embryo (6,12,13). The study of Rab11 and the role of the recycling endosome during cytokinesis were further explored in mammalian cells. In HeLa cells, Rab11 colocalizes with transferrin to the midbody as well as the ingressing furrow during telophase (14). Using fluorescence resonance energy transfer (FRET), FIP3, a Rab11 effector, was found to be in complex with Rab11 at the cleavage furrow. Interference with either Rab11 or FIP3 via siRNA led to an increase in binucleate cells as well as daughter cells that retained cytoplasmic bridges (14). This data suggests that in mammalian cells, as in cells from other model organisms, the recycling endosome plays an essential role in the successful completion of cytokinesis.

Further work by Giansanti, et al. (15), found that Drosophila spermatocyte Rab11 mutants exhibit not only defects in furrow ingression, but also in the actomyosin ring. In these mutants, the actomyosin contractile ring forms normally during early telophase, however, the ring shows defects in density and either fails to contract or disassembles during the later stages of telophase (15). Another component of the cytoskeleton, microtubules (MTs) have been shown to be necessary for the maintenance of Rab11 localization at the microtubule-organizing center (MTOC) during cellularization of the Drosophila embryos (16).

Are the defects seen in Rab11 mutants due to a failure in targeting or in fusion of vesicles at the cleavage furrow? Current studies seem to indicate that both of these processes are controlled by Rab11. The mislocalization of Rab11 from the cleavage furrow to the cytoplasm in Rab11 mutants is seen across a variety of different species, supporting the hypothesis that the delivery of vesicles to the furrow is Rab11 dependant. Additional data have demonstrated that vesicles accumulate as punctate structures in the absence of Rab11, revealing the inability of these vesicles to fuse with the plasma membrane (15). Further investigation using high-resolution fluorescence microscopy will help to elucidate the roles Rab11 may play at the site of furrow ingression.

In another striking observation, FIP3/FIP4 (Rab11 effectors) maintain their localization at the midbody independently of Rab11 (14). What is responsible for the delivery of FIP3/FIP4 to the midbody, and what role do they play there? A comprehensive study by Fielding, et al. (2005), found that FIP3 and FIP4 interact with ADP-ribosylation factor 6 (Arf6) independently of Rab11. Arf6 is a small GTPase known to regulate actin dynamics (for a Review see (17)). In this study, there is also evidence that Exo70p, a member of the exocyst complex, interacts with FIP3, FIP4 and Rab11 in CHO cells (18). Given these findings, the authors propose a model in which FIP3/FIP4 is delivered to RE vesicles in a Rab11-dependant manner; these vesicles are then targeted via microtubules (MTs) to the midbody and cleavage furrow during telophase (14,18). Arf6 and Exo70p then maintain the FIP3/FIP4 vesicles at the midbody until they can fuse with the plasma membrane, thus completing cytokinesis (18).

Arf6 was later shown to be essential for cytokinesis in the Drosophila male germ line (19). Arf6 was found to colocalize with both early endosome and recycling endosome markers Rab4 and Rab11 respectively. Early endosomes were found at the furrow during early cytokinesis, whereas recycling endosomes did not appear at the ingressing furrow until the later stages of cell division. In arf6¹ mutant spermatocytes, cells exhibit normal furrow initiation; however the furrows then regress (19). The failure of arf6¹ mutant spermatocytes to complete furrow ingression shows an essential role for recycling endosomes during late cytokinesis. This evidence suggests a model in where there are two populations of endocytic vesicles that play essential roles in different stages of cytokinesis (19).

The Highway: Secretory Pathway

The secretory pathway consists of the shuttling of proteins from the ER to the Golgi apparatus, where they are sorted and packaged for delivery to the plasma membrane, in some cases routing through the RE in the process. Early studies investigating the role of Golgi-derived vesicles in cytokinesis made use of Brefeldin A (BFA). BFA is a compound isolated from fungi that inhibits secretion in part by blocking non-clathrin-coated vesicle budding (20). An extensive study in C. elegans has shown that embryos treated with BFA are unable to successfully complete abscission, causing fully-formed cleavage furrows to regress, similar to the phenotype seen in rab-11 mutant embryos (21). Interestingly, defects in spindle orientation are also seen, as normal asymmetric divisions in the C. elegans embryo become symmetric prior to abscission failure (21).(21)(Skop, et al. 2001)(21)(20)(17) BFA treatment also effectively inhibits later stages of furrow ingression during cellularization in Drosophila embryos and in multiple mammalian cell types (22)(23,24). In contrast, Shuster and Burgess (25) demonstrated that the early divisions of sea urchin embryos are not susceptible to BFA treatment likely indicating the presence of a large store of fusion ready vesicles generated during oogenesis in this species. These studies raise interesting questions as to the variability of how cytokinesis is completed in different cell types.

Work in the late 1990s brought to light the importance of the secretory pathway in cytokinesis through the study of syntaxins. Syntaxins (also called t-soluble NSF attachment protein receptors, or t-SNAREs) form a complex with v-SNAREs and SNAP-25 for the successful docking and fusion of vesicles at their target membranes. Conner and Wessel (26) found that the inhibition of the syntaxin homologue by either Botulinum neurotoxin C1, a potent syntaxin inhibitor, or by antibody injection leads to inhibition of cytokinesis in the sea urchin Lytechinus variegatus. These defects seem to be due to a reduction of syntaxin found localized to the vesicles in the treated blastomeres of the sea urchin embryo (26). In the same year, Jantsch-Plunger and Glotzer (27) demonstrated that syntaxins were also involved in cytokinesis in C. elegans embryos. Worms injected with dsRNA for the syn-4 gene had reduced fertility and a high rate of lethality during early larval stages. Of the embryos that were laid, three phenotypes were observed: premature nuclear envelope reformation, re-entry of the polar body, and regression of the cleavage furrow, suggesting defects in the final stages of cytokinesis (27). A similar larval lethal phenotype has also been shown upon disruption of syntaxin 5 in Drosophila embryos (28). Surviving males of an incomplete penetrance allele exhibited defects in germ line cytokinesis and spermatid elongation (28). These results suggest that syntaxins play an evolutionarily conserved role in the successful completion of cytokinesis through their function of mediating the fusion of Golgi-derived vesicles at the plasma membrane.

Mammalian cells have also been shown to utilize syntaxins and other SNARE proteins for abscission. In Madin-Darby Canine Kidney (MDCK) cells, syntaxin 2 is necessary for the abscission of the midbody, but not mitosis or furrow ingression (29). In addition, inhibition of endobrevin/VAMP-8, a small v-SNARE protein that colocalizes with syntaxin 2, leads to similar abscission defects as syntaxin 2 disruption (29). These data suggest that syntaxins are not the only family of proteins involved in vesicle targeting, docking, and fusion that are necessary for the completion of cytokinesis. In further support, a recent paper by Song, et al. (30) demonstrated that knockdown of syntaxin 16 in HeLa cells also leads to similar defects in the later stages of cytokinesis. These studies highlight the existence of multiple v- and t-SNARE proteins, which may play varied roles in vesicle docking and fusion. Therefore, more investigation into the role these proteins play in varying organisms and cell types is necessary.

The exocyst is a protein complex involved in vesicle transport to and maintenance at the plasma membrane. This complex forms the link between vesicle trafficking and the role of SNAREs in the final stages cytokinesis. This was demonstrated by Gromley, et al. (23) who proposed a model where exocyst proteins along with endobrevin and syntaxin-2 localize to a “midbody ring” during the later stages of cytokinesis. Vesicles from the Golgi, marked by v-SNAREs arrive at the ingressing furrow and fuse with the plasma membrane to complete abscission. In HeLa cells, knockdown of proteins comprising the exocyst causes delays in cytokinesis or incomplete abscission leading to daughter cells going through continuous rounds of mitosis while still being connected by a cytoplasmic bridge (23). This is most likely due to the inability of v-SNARE-containing secretory vesicles to fuse with the plasma membrane at the cleavage furrow, since accumulation of secretory vesicles are seen in this region when sec5, a component of the exocyst, is depleted (23). These data implicate vesicle docking and fusion via SNARE complex proteins as important players in the successful completion of abscission.

Perturbations of other Golgi-associated proteins have been shown to exhibit cytokinesis defects as well. The conserved oligomeric Golgi (COG) complex is known to be essential for Golgi structure and function (31). In Drosophila, the Cog5 homologue four way stop (fws) is necessary for ingression of the cleavage furrow (32). Spermatocytes lacking functional fws showed defects in actomyosin ring constriction during mid-telophase. Interestingly, spermatid elongation was also inhibited in fws mutants (32). Another Golgi-associated protein, Lava Lamp (Lva), is necessary for cellularization in Drosophila embryos (22). This protein is localized to Golgi-derived vesicles that accumulate at the furrow region during the mid to late stages of ingression. Inhibition of Lva leads to a delay in furrow invagination similar to that seen in embryos treated with BFA (22).

Modes of Transportation: Molecular Motors

Recent studies have found both actin and microtubule-associated motors are involved in the transport of new membrane to the site of furrow ingression. The minus end microtubule motor dynein and its activator dynactin have been shown to be essential for the correct localization of proteins necessary for the completion of cytokinesis. In Drosophila embryos, Nuclear-fallout (Nuf) has been shown to associate with dynein and the proper localization of Nuf at the MTOC is dependent on microtubules and dynein activity (16). The interaction between Nuf and dynein may thus be necessary for the recruitment of actin to the ingressing furrow leading to successful cytokinesis.

As previously described, RACK-1 was discovered in a proteomic screen of the midbody of C.elegans and found to be necessary for the completion of cytokinesis (6). In addition to RACK-1’s role in the confinement of recycling endosomes to the centrosomes, Ai, et al. showed that RACK-1 interacts with dynactin in the C.elegans embryo (11).

The role of the dynein/dynactin complex in proper membrane trafficking during cytokinesis is conserved in mammalian cells as well. Recently, it has been shown that there are two populations of the cytoplasmic dynein1 in HeLa cells: one involved in trafficking from the ER to the Golgi, and the other involved in the endosomal pathway (33). Arf6 stabilizes the interaction between its effectors JIP3 and JIP4 with dynein. Arf6 has also been shown to compete with kinesin-1 for binding to JIP3 and JIP4 (34). These experiments make it clear that the dynein/dynactin complex is essential to the regulation and proper localization of trafficking components during the later stages of cytokinesis.

In addition to the dynein/dynactin complex, the plus-end directed kinesin motors have also been shown to be involved in membrane trafficking during late stages of cytokinesis. Kinesin-like proteins are (35) necessary for abscission in C.elegans (36) and Drosophila (37) embryos, as well as in mammalian cells (35). The kinesin-like protein KIF14 localizes to the midbody in the later stages of mitosis, and has recently been found to interact with the membrane protein supervillin during interphase (35,38), suggesting a possible role for KIF14 in membrane trafficking to the furrow during cytokinesis. Kinesin II is known to be involved in vesicle trafficking from the Golgi to the ER and the RE pathway, as well as the proper localization of syntaxin-containing vesicles to the midbody during cell division (39–41). It has been suggested that because of the defects in cytokinesis seen in the absence of these kinesins, and their known roles in vesicle trafficking, that these proteins not only act to stabilize the midbody, but are also necessary for membrane addition at the cleavage furrow (42). However, the connection between these two roles for kinesin remains unclear and direct evidence of kinesin-regulated membrane trafficking during cytokinesis remains to be discovered.

Myosin VI is an unusual myosin motor in that it moves toward the minus end of actin filaments. During mitosis, Myosin VI localizes to vesicles and migrates from the spindle poles to the plasma membrane at the furrow region, finally localizing to either side of the midbody during late cytokinesis (43). Lending evidence to its necessity during cytokinesis, siRNA depletion of myosin VI in HeLa cells leads to the cells showing defects in abscission. These cells were also delayed in metaphase, further supporting the hypothesis that myosin VI plays an essential role early in mitosis (43).

Conclusions

While it is evident that membrane trafficking plays a critical role in cytokinesis, many questions still remain. It is still unclear whether the main function of vesicle transport to the furrow is for plasma membrane insertion or to deliver proteins necessary for the continuation of ingression. Currently, it seems that both of these functions are carried out via membrane trafficking to the cleavage plane. Additional investigations are necessary to discern the relationship of new membrane addition and actomyosin ring contraction during cytokinesis. A few studies have found cursory evidence that the two processes are independent of one another (8,19,44). It also remains unclear whether the two pathways involved in vesicle trafficking act separately, or if Golgi-derived vesicles are routed through the recycling endosome before being directed to the cleavage furrow.

Although beyond the scope of this review, an increasingly interesting field of study is the function of the unique lipid content found at the cleavage furrow. It has been proposed that these lipids are involved in both the organization of the cytoskeleton and/or the addition of new membrane during cytokinesis(45,46). In addition, the formation of distinct apical and basolateral membranes in polarized cells, especially in light of evidence that most of the vesicles involved in cytokinesis are endocytosed from the apical regions of the cell raises many questions for further research (47). Advances in light microscopy and investigations into protein-protein interactions will undoubtedly shed more light onto the role of membrane trafficking during cytokinesis.